article
stringlengths 0
745k
| abstract
stringlengths 0
92.7k
|
|---|---|
the discrepancy between the observed lifetimes of protostellar disks ( 23 myr ; haisch et al .
2001 ) and the length of time required for planet formation by core accretion ( @xmath0 myr ; pollack et al .
1996 ) has long presented a problem for planet formation theory . however , in 2005 , two new models of planet formation showed that the core accretion - gas capture process could form gas giants within 2.5 myr .
hubickyj et al .
( 2005 ) modeled jupiter s formation _ in situ _ at 5.2 au and found that jupiter could grow from a @xmath1 core to its present mass in 2.2 myr .
the core accretion models of alibert et al .
( 2005 ) allowed both jupiter and saturn to form concurrently within 2.5 myr in disk of mass @xmath2 .
each model employed a different approach in order to speed up giant planet formation .
alibert et al .
added type i migration to the core accretion model : the inward motion of the protoplanets allows them to receive a fresh supply of planetesimals and gas as they move into undepleted regions of the solar nebula .
the accretion rate is then no longer limited by the rate at which the protoplanet s hill sphere expands , so the giant planet formation timescale decreases by up to an order of magnitude .
these models require proto - jupiter to have an initial semimajor axis of @xmath3 au and proto - saturn to begin forming at 11.9 au , and migrate to their current positions on a @xmath4 myr timescale .
one notable feature of the @xcite core accretion model is that it successfully predicts the heavy metal content of jupiter and saturn s atmospheres according the clathrate hydrate trapping theory of lunine & stevenson ( 1985 ) .
however , the type i migration rate is a free parameter : alibert et al .
note that the analytical work of @xcite predicts migration rates far too large to be consistent with the observed frequency of extrasolar planets .
the authors get around this problem by scaling the @xcite by an arbitrary factor @xmath5 , where @xmath6 .
finally , there is one more basic assumption underlying the model , which is that the gas / solid ratio beyond the ice line is constant at @xmath7 .
the _ in situ _ planet formation models of hubickyj et al .
( 2005 ) decrease jupiter s formation time by requiring that grains quickly settle to the bottom of the protoplanetary envelope , where they are destroyed by sublimation ( podolak 2003 ) . assuming grain settling lowers envelope opacity due to grains to @xmath8% of the interstellar value
, the gas can contract efficiently and make way for new material entering the protoplanet s hill sphere .
the most important free parameter in the hubickyj et al .
models is the solid surface density of planetesimals in the planet s feeding zone : since the protoplanet does nt move through the disk , it requires a feeding zone with @xmath9 g @xmath10 in order to attain the @xmath11 core necessary for accreting a massive gaseous envelope ( papaloizou & nelson 2005 ) . the availability of new calculations of solid surface density as a function of heliocentric distance in the solar nebula ( robinson et al .
2008 ; hereafter paper 1 ) raises the possibility of extending _ in situ _ core accretion simulations to include saturn . since the nice model of tsiganis et al .
( 2005 ) predicts that proto - saturn migrated 1 au outward at most , we consider the _ in situ _ approximation reasonable when applied to saturn . by providing theoretically and observationally motivated values for solid surface density @xmath12 ,
the paper 1 results move core accretion simulations away from parameter studies and toward determinism . since the solar nebula had to be capable of forming both jupiter and saturn concurrently ( and of course uranus and neptune , the formation of which we will examine in future work )
, we first assess the ability of the paper 1 solar nebula model to produce jupiter . in paper 1 , the value for @xmath12 after @xmath13 yr of solar nebula evolution is 13.2 g @xmath10 . adopting the relationship between solid surface density and jupiter formation time at 5 au calculated by robinson et al .
( 2006 ) , @xmath14 we find that the paper 1 results allow jupiter to form in 2.5 myr .
built in to the @xcite scaling relation is the assumption that efficient grain settling leads to protoplanetary envelope opacities of @xmath15 those of interstellar grains . in this letter
, we will relax this assumption and also investigate the limiting case of 100% interstellar grain opacity with respect to saturn .
if the fiducial disk from paper 1 can form saturn within 23 myr , we will have successful core accretion models of the two gas giants forming near their present positions in a gravitationally stable disk ( see paper 1 for a discussion of the solar nebula dynamics ) . in [ methods ] , we describe our theoretical treatment of the core accretion process . in [ results ]
, we discuss the results of our simulations , with special emphasis on formation timescale , effect of atmospheric opacity , and core / atmosphere mass ratio .
we present our conclusions in [ caconclude ] .
we use the theoretical model of planet formation described by laughlin et al .
( 2004 ) to model the core accretion and gas capture of proto - saturn .
initially , a protoplanetary core of mass @xmath16 is embedded at saturn s heliocentric distance , 9.5 au , in a viscously evolving disk of age @xmath17 yr , surrounding a t - tauri star of mass @xmath18 .
we assume that by @xmath17 yr , the available dust has formed 100 km planetesimals that are invulnerable to gas drag ( weidenschilling 1977 ) : planetesimal orbits are modified only by interactions with proto - saturn .
gas temperature and density are regulated by viscous evolution of the solar nebula .
we use the time - evolving temperature and density at 9.5 au , beginning at @xmath19 yr , as calculated in paper 1 .
the contraction and buildup of protoplanetary cores and their gaseous envelopes embedded in our model evolving disk are computed with a henyey - type code ( henyey et al .
1964 ) . following the argument of @xcite
that grain settling in the protoplanetary envelope would reduce envelope opacity where grains exist , we adopt grain opacities of 2% of the interstellar values used in @xcite in our fiducial model , which we will call s0 . however , in order to assess the effect of envelope opacity on saturn s formation timescale , we present a second core - accretion simulation , s1 , using full interstellar grain opacity .
we use a core accretion rate of the form @xmath20 ( papaloizou & terquem 1999 ) , where @xmath12 is the surface density of solid material in the disk , @xmath21 is the orbital frequency at 9.5 au , @xmath22 is the effective capture radius of the protoplanet for solid particles , @xmath23^{1/3}$ ] is the tidal radius of the protoplanet ( where @xmath24 is the semimajor axis of the protoplanet s orbit ) , and @xmath25 is a constant near unity .
the outer boundary conditions for the protoplanet include the decrease with time in the background nebular density and temperature . during the late phase of planet growth ,
when planetesimals may be ablated by the massive envelope before reaching the core , we consider a planetesimal captured if it deposits 50% or more of its mass in the envelope . at this stage , we invoke the sinking approximation and assume the ablated planetesimal debris sinks rapidly to the planet core without leaving remnants in the envelope .
in both simulations ( s0 and s1 ) , we start the core accretion model with midplane temperature , gas density and solid surface density from the solar nebula model of paper 1 .
this model has two key features favoring planet formation that are missing from passive disk models : ( 1 ) viscous stresses drive the initial @xmath26 surface density profile toward uniformity , so that saturn s feeding zone gains mass during the first @xmath27 yr of disk evolution , and ( 2 ) the presence of hydrated ammonia ice at the snow line increases the solid surface density by 7% over the standard water ice rock refractory chon mixture .
we use a starting solid surface density of 8.6 g @xmath10 ( paper 1 ) which decreases with time as proto - saturn captures planetesimals . figure 1a shows the growth of saturn from a core of @xmath28 to its present - day mass of @xmath29 .
solid lines correspond to the s0 model ( in which grains quickly settle and sublimate , reducing their contribution to envelope opacity by an assumed factor of 50 ) , and dashed lines show the s1 model ( in which grains stay in the envelope and opacity due to grains takes on the full interstellar value ) . an important property of our model , and one of two key results of this letter , is saturn s formation time : the planet attains its current mass in only 3.37 myr for model s0 and 3.47 myr for model s1 .
these are the first _ in situ _ core accretion models of saturn with formation times within @xmath30 of observed protostellar disk lifetimes .
the second key result is that saturn s formation time is nearly independent of the assumed grain opacity in the envelope .
this surprising result occurs because saturn s core growth rate is limited by the keplerian speed in the feeding zone : a slow - moving core takes a long time to encounter and capture planetesimals ( see equation [ coregrowth ] ) , and may never reach isolation mass .
whereas during jupiter s formation early core buildup takes only 0.5 myr and the gas contraction phase dominates planet growth ( hubickyj et al .
2005 ) , in saturn s case , core growth lasts until the planet reaches its final mass . only in the last @xmath31
yr of growth does saturn possess a gas mass of @xmath32 .
given that the planet has a negligible envelope mass throughout most of its formation , it stands to reason that envelope opacity would not exert much influence on saturn s formation timescale .
the idea that solid growth rate controls saturn s formation is consistent with the planet s high core / atmosphere mass ratio : @xmath33 , as opposed to @xmath34 for jupiter ( saumon & guillot 2004 ) .
if all accreted planetesimals and their ablated debris end up in the solid core , as is assumed in our model , saturn s total core mass reaches @xmath35 in model s0 and @xmath36 in model s1 .
based on gravitational moment measurements and internal structure modeling , saumon & guillot ( 2004 ) place saturn s core at 922@xmath16 and total heavy element content at 1328@xmath16 .
our model saturn has a heavy element mass that is too high by at least a factor of 1.6 ( model s0 ) .
we note , however , that the carbon enrichment in saturn s atmosphere has recently been revised upward to @xmath37 ( flasar et al . 2005 ) , which may allow for a larger heavy element inventory than previously thought . one method of reducing the core mass , though not the total heavy element mass , is to account for planetesimal disruption not only by ablation of debris ( [ methods ] ) , but by sublimation of volatiles .
our planetesimals are 50% h@xmath38o by mass , 6% nh@xmath39 , and 1% other ices , such as hcn and h@xmath38s , for a total of 57% ice .
these ices could sublimate either during infall through the envelope , analogously to meteorites in earth s atmosphere , or upon impact with the solid core .
helled et al .
( 2008 ) found that sublimation of volatiles during the collapse of a giant planet formed by disk instability ( boss 2005 ) is efficient : refractory silicate grains sediment to form a core , while ices remain in the planet atmosphere .
tingle et al .
( 1991 ) tested the survival of volatiles experiencing hypervelocity impacts ( @xmath40 ) by shocking samples of the murchison meteorite with pressures up to 36 gpa .
they found that 70% of organic and organosulfuric material , including h@xmath38s , sublimates upon experiencing an impact with @xmath41 .
the volatiles in saturn - building planetesimals are not likely to survive a high - velocity impact and subsequent contact with the hot protoplanetary core , @xmath42 k , in solid form .
if we assume all accreted ices undergo a phase transition from solid to gas , saturn s core mass drops to @xmath43 ( s0 ) and @xmath44 ( s1 ) , which are near the range determined by saumon & guillot ( 2004 ) . however , the total heavy element / hydrogen mass ratio in the planet is still higher than observed : saturn is 5070% hydrogen by mass , whereas our model predicts a hydrogen mass fraction of only 42% ( s0 ) and 31% ( s1 ) .
another possible way to reduce saturn s heavy element mass without slowing the planet s growth is to cut off solid accretion midway through planet formation .
this approach simulates the effect of another embryo competing for planetesimals .
although our solid accretion rates are calculated for the monarchic growth paradigm , in which planetesimal dynamics are determined by proto - saturn only , this scenario is an approximation : kokubo & ida ( 1998 ) predict an oligarchic planet formation epoch with competing embryos spaced @xmath45 hill radii apart that lasts for @xmath46 myr .
hubickyj et al .
( 2005 ) tested the effect of a core accretion cutoff on jupiter s formation and found that as long as the core has mass @xmath47 before solid accretion ceases , the planet reaches hydrodynamic gas accretion even more quickly than when solid accretion continues unchecked : 0.78 myr vs. 2.22 myr .
continuous , late - stage planetesimal accretion slows planet formation by depositing kinetic energy in the protoplanetary envelope and inhibiting gas contraction .
the competing embryo scenario holds promise for bringing saturn s core / atmosphere mass ratio into agreement with observations while still retaining the quick formation time .
however , it is a double - edged sword : oligarchs with overlapping zones of gravitational influence can increase the rms planetesimal eccentricity , @xmath48 , far more efficiently than a single monarch core , decreasing the gravitational focusing ability of all embryos . @xcite and @xcite
find that significant planetesimal stirring can occur when embryos are between @xmath49 and @xmath50 .
indeed , @xcite replaced equation [ coregrowth ] with the analytical oligarchic growth rate of @xcite in _ in situ _ core accretion simulations of jupiter and found formation times of 10 - 20 myr .
thommes et al .
( 2008 ) created a self - consistent planet formation model that included gas disk evolution , planet - disk interactions ( including gap opening and gas accretion onto solid cores ) , and planet - planet interactions .
gas giants with jupiter - like core / atmosphere mass ratios consistently emerged in disks with mass @xmath51 .
even a modest amount of migration , as occurred in the solar system , appears to compensate for the inefficiency of _ in situ _ oligarchic growth .
the true saturn formation scenario probably involved several @xmath52 embryos spanning 9 - 12 au , some outward motion of proto - saturn , and possible competition with neptune , forming near 12 au ( tsiganis et al . 2005 ) .
one new feature of both models , s0 and s1 , is that saturn does not need to accrete all the planetesimals in its feeding zone to reach its current mass .
figure 1b shows the remaining solid surface density , not incorporated in saturn s core , as a function of time .
( the upturn near the end of the simulation results from the rapid expansion of the planet s feeding zone during hydrodynamic gas accretion and the assumption that planetesimals are always uniformly distributed , which is unphysical if the planet grows more quickly than planetesimal redistribution can occur . ) in both models , there are still 2 g @xmath10 of solids left in saturn s feeding zone at @xmath53 myr .
this means the core never reaches isolation mass . following lissauer ( 1993 ) , we calculate @xmath54 whereas saturn s core attains a mass of @xmath35 in model s0 and @xmath36 in model s1 .
both hubickyj et al .
( 2005 ) and pollack et al .
( 1996 ) found that giant planets pass through a lengthy plateau phase in which both solid and gas accretion rates are low and planet mass changes very little with time .
this phase begins when the planet core nears isolation mass , and its duration is regulated by gas contraction efficiency .
examining figure 1a , we see that saturn never experiences this mass plateau : both solid and gas accretion rates increase with time .
with extra solid and gas mass remaining in the disk after 3.4 myr , why would saturn not continue to grow ?
we arbitrarily stop the simulation once saturn reaches @xmath29 , but our disk could quickly form a more massive planet .
explaining saturn s low mass in the context of the disk models in paper 1 may require either another nearby embryo competing for both solids and gas perhaps neptune , forming near 12.5 au according to the nice model ( tsiganis et al .
2005)or a mechanism for disk dissipation that begins after a few myr .
photoevaporation ( alexander et al . 2006 ) is one possibility : the disk near 10 au begins to be disrupted by ionizing radiation at 23 myr .
we also require a mechanism to populate the oort cloud with @xmath55 of comets ( weissman 1996 ) .
saturn scattering the @xmath56 it does not accrete into the oort cloud would be a good start , and jupiter could contribute more material during the late stages of its formation and/or migration .
is the concept of isolation mass relevant for any giant planet except jupiter ?
from equation [ miso ] , we see that @xmath57 increases with radius if @xmath58 in the power - law surface density profile @xmath59 .
extrapolating from the simulation presented here and the near - flat solid surface density profile seen in paper 1 , we propose that saturn , uranus and neptune do not reach isolation mass .
due to their low keplerian orbital speed , planet cores in the outer solar nebula experience a lower planetesimal collision rate than proto - jupiter . instead of their formation timescale
being governed by gas contraction efficiency , the solid accretion rate , mediated by @xmath21 and @xmath60 , is the critical factor .
the hypothesis that the three outer planets never reach isolation mass is consistent with the planets bulk composition : all three have solid / gas ratios far higher than jupiter .
based on the solid surface densities calculated in paper 1 , we have created core accretion models of saturn with a formation timescale of 3.37 - 3.48 myr , which is compatible with observed protostellar disk lifetimes . unlike previous studies of jupiter ( e.g. hubickyj et al .
[ 2005 ] ) , we find that grain opacity in the protoplanetary envelope has virtually no effect on saturn s formation timescale . the same solar nebula model that was the basis for simulations s0 and
s1 is capable of forming jupiter in 2.5 myr , assuming reduced grain opacities . finally , this is the first _ in situ _ core accretion model of a gas giant that lacks a plateau phase , in which the planet s mass remains nearly constant at approximately the isolation mass .
we postulate that the low orbital speeds far from the sun prevented saturn , uranus and neptune from accreting solids efficiently enough to reach isolation mass .
these planets never experienced a gas - only accretion phase , as did jupiter , and consequently have much higher core / atmosphere mass ratios .
we thank geoff bryden and jonathan fortney for helpful conversations and the anonymous referee for suggesting model s1 .
this research was supported by a scholarship award from the achievement rewards for college scientists foundation to s.d.r . ;
by nsf career grant ast-0449986 and nasa planetary geology and geophysics program grant nng04gk19 g to g.l . ; and by nsf grant ast-0507424 and nasa origins grant nnx08ah82 g to p.b . | we present two new _ in situ _ core accretion simulations of saturn with planet formation timescales of 3.37 myr ( model s0 ) and 3.48 myr ( model s1 ) , consistent with observed protostellar disk lifetimes . in model s0 , we assume rapid grain settling reduces opacity due to grains from full interstellar values ( podolak 2003 ) . in model s1
, we do not invoke grain settling , instead assigning full interstellar opacities to grains in the envelope .
surprisingly , the two models produce nearly identical formation timescales and core / atmosphere mass ratios .
we therefore observe a new manifestation of core accretion theory : at large heliocentric distances , the solid core growth rate ( limited by keplerian orbital velocity ) controls the planet formation timescale .
we argue that this paradigm should apply to uranus and neptune as well . |
besides the numerous attempts to implement a large - scale quantum machine for local and centralized applications such as efficient factoring @xcite , there are other vast areas of research on interesting and unique distributed quantum algorithms with no classical variants @xcite . perceived advantages with respect to security , time complexity or communication complexity are sufficient to motivate researchers to focus on design , implementation and optimization of the algorithms .
quantum key distribution ( qkd ) protocols @xcite and quantum leader election algorithms @xcite are a subset of these algorithms@xcite .
however , although detailed , resource - aware analyses of monolithic algorithms are increasing , the literature for equivalent analysis of distributed algorithms remains sparse .
a quantum approach for solving the classical problem of _ byzantine agreement _ is another important distributed quantum algorithm , proposed by ben - or and hassidim @xcite .
their pure - theoretical distributed algorithm terminates in @xmath2 expected number of rounds in the presence of a computationally unbounded , full information and adaptive adversary .
there is no similar variant in classical deterministic or randomized solutions with these unique features at the same time .
distributed algorithms for solving the byzantine agreement problem are crucial for designing fault - tolerant systems in many domains .
these algorithms have broad applications in areas ranging from fault - tolerant real - time and online services @xcite , to secure and large - scale peer - to - peer services @xcite .
although quantum sharing - based byzantine agreement can be theoretically faster and more secure than the classical algorithms , a considerable gap has remained from the abstract layer to experimental layer . to fill this gap
, we must address challenges ranging from finding a practical architecture for quantum processing elements in each independent node to consideration of a quantum - based infrastructure for communication known as a _
quantum repeater network ( qrn ) _ , by using entanglement and teleportation@xcite .
analysis of faults in end - nodes and network imperfections in comparison to the ideal model of the original algorithm ( which assumes perfect computation in each node and a perfect point - to - point quantum communication link between each pair ) increase the complexity of the problem .
we face some questions about the quantum architecture of the abstract algorithm in addition to computational and communication resources required for running this algorithm . in this work
we have : * extracted the minimal architecture requirements for the quantum part of the qba protocol , * proposed two optimization techniques for the architecture and circuit for a minimum setup of qba with 5 nodes , and * estimated computation and communication costs for the minimum setup , including required fidelity .
the remainder of the paper is organized as follows : in section [ sec : preliminary ] , we review classical and quantum aided byzantine agreement and quantum repeater networks .
a high level analysis of the qba protocol is presented in section [ sec : highlevelanalysis ] . in section [ sec :
design ] , the overall architecture design is proposed .
we provide required optimization techniques in section [ sec : implementation ] .
the assessment results is shown in section [ sec : result ] .
finally we conclude the paper in section [ sec : conclusion ] .
in this section , we review essential concepts and related background . we start with the standard definition of classical byzantine agreement and history of byzantine - tolerant solutions .
we continue with the first scalable quantum sharing - based byzantine agreement proposed by ben - or and hassidim in @xcite .
a general review of the network requirements for execution of the algorithm concludes this section .
the history of the byzantine agreement problem goes back to a proposal by lamport et al . for defining a more sophisticated form of fault model with active and malicious behavior , now known as byzantine faults @xcite . tolerating this form of the fault which is stronger than fail - stop faults , requires more computation and communication resources in a distributed systems . to analyze the behavior of this type of fault , lamport et al .
employed the colorful metaphor of a distributed system as a group of byzantine generals arrayed around a city , trying to decide as a group whether to attack or retreat .
it must be assumed that less than third of these generals are traitors ( active fault as an internal adversary ) .
the mission of the distributed generals will be successful if all the loyal generals agree on a unique command ( attack or retreat ) in the presence of inconsistent messages sent by traitors ( byzantine faults ) , otherwise the protocol fails . in this problem ,
all communication is done by messenger . from the communication point of view , byzantine agreement is a form of reliable broadcast between multiple nodes in a network which supports point to point channels between nodes .
one of the first attempts to exploit quantum advantages in a weaker version of agreement ( known as detectable broadcast ) as a form of reliable broadcast is presented by fitzi et .
al . in @xcite
although the work does nt solve lamport s original problem , the solution is suitable for a small - scaled distributed systems ( with 3 nodes ) in a detectable broadcast application instead of byzantine agreement . &
@xcite & @xcite & @xcite + round & & & + complexity & @xmath3 & @xmath4 & @xmath2 + & & & + in this paper , the focus is on ben - or and hassidim s algorithm .
they proposed a scalable solution for creating a quantum aided byzantine agreement protocol by modification of feldman & micali s classical probabilistic algorithm @xcite to share and verify a known quantum state , instead of sharing and verifying classical random numbers@xcite .
ben - or and hassidim s algorithm is based on the following assumptions : * there exists a full - duplex ideal quantum channel between each pair of players ( end - nodes ) .
note that during execution of the algorithm we also need an ideal classical channel between each pair of the players .
* to tolerate an upper bound on the number of faulty players ( @xmath5 ) , the communication model need to be synchronous .
each round consists of two separate phases : a communication phase and a computation phase . for the asynchronous case ,
they prove the effectiveness of their algorithm with an upper bound of @xmath6 .
* the adversary can be adaptive , full information and computationally unbounded . at the end of algorithm execution ,
the agreement between all non - faulty players , the validity of the output between them and termination of the algorithm ( with the probability of 1 ) are guaranteed .
the algorithm can be analyzed from two points of view : 1 .
performance analysis under security assumptions : the algorithm requires a constant number of expected rounds in the presence of a full information adversary with upper bound of malicious players @xmath7 between @xmath8 players ( @xmath9 ) . table [ tab : roundcomplexity ] shows the upper bound of the round complexity in classical deterministic @xcite , classical randomized @xcite and quantum - aided algorithms @xcite .
security analysis under performance assumptions : as another view , the quantum algorithm is more secure than deterministic and randomized algorithm in the case of the lowest bound of round complexity . if we analyze the security of byzantine agreement protocols for fixed - round algorithms , we get the result similar to table [ tab : securityqba ] .
there is no fixed - round , deterministic algorithm with the strongest type of adversaries ( adaptive , computationally unbounded and full information ) @xcite . for the case of the randomized algorithms ,
the best available algorithm suffers from the assumption of communication security between each pair of non - faulty nodes @xcite . & & @xcite & @xcite + adaptive adv .
& & yes & yes + unbounded adv . &
no solution & yes & yes + full info . adv .
& & no & yes + we will describe the details of the algorithm in addition to its theoretical behavior in the next section .
execution of fundamentally distributed quantum algorithms requires a quantum - based solution for communication known as a _ quantum repeater network ( qrn ) _
@xcite ( fig .
[ fig : qrn-00 ] ) .
quantum repeater networks provide an efficient infrastructure for distributed systems by using entanglement , teleportation and some forms of error detection and correction . in these networks
, information could be represented by entangled states and a link between two quantum nodes creates entangled states supporting quantum teleportation .
purification and entanglement swapping repeaters @xcite , error correction - based repeaters @xcite and quasi - asynchronous @xcite repeaters have been proposed for qrns @xcite . despite many ongoing demonstrations of quantum key distribution ( qkd ) application without the benefits of entangled repeater networks ,
the research on architecture analysis and design of other promising distributed quantum applications is rare and narrow . to the best of our knowledge
, there is no detailed analysis and design for quantum - aided byzantine agreement algorithm .
therefore , there is a need for analysis of the minimum requirements in quantum repeater networks for complete execution of quantum byzantine agreement .
the modeling must be extended beyond that described by ben - or and hassidim @xcite .
in particular , the algorithm as proposed has been analyzed assuming only pure states , and without reference to the demands made on the repeater network . for the remaining sections , the focus will be on finding appropriate answers for the following questions : * what are the required quantum resources for qba protocol ? *
how resilient is it to network and gate error ? *
is it practical and attractive in the real world ?
* can we use qba as an early demonstration application of quantum repeater networks ? for exploration of an efficient solution , we need to analyze the relevant part of graded verifiable quantum secret sharing ( vqss ) @xcite and the qba protocol in detail , and determine the overall architecture of the quantum processing elements as will be presented in the next section .
in the previous section , we introduced the quantum aided byzantine agreement protocol and enumerated its features in qualitative terms . in this section
we begin the quantitative analysis of the algorithm . as shown in fig .
[ fig : img01 - 1 ] , to run the quantum - aided byzantine agreement ( qba ) , all the end - nodes concurrently and independently run the agreement protocol with 3 sub - protocols .
the overall protocol is based on the original randomized algorithm @xcite and has a constant expected number of rounds . as shown in fig .
[ fig : img01 - 1 ] , the sequential sub - protocols of quantum - aided byzantine agreement are @xmath10 , @xmath11 and @xmath12 . at the beginning , all nodes @xmath13 are supposed to start the protocol with an input value @xmath14 .
execution of the first sub - protocol ( @xmath10 ) advances toward an agreement for non - faulty and uncertain nodes .
the other sub - protocols @xmath11 and @xmath12 are executed sequentially after @xmath10 .
they are pure classical protocols for biasing the outcome of the coin flipping procedure into zero or one respectively .
the only quantum part of all of these sub - protocols is in the quantum aided oblivious common coin procedure of sub - protocol @xmath10 , labeled qocc .
qocc is a modification of the original oblivious common coin ( occ ) procedure for independent and random coin flipping . in the quantum version ,
the procedure requires sufficiently random numbers among the nodes to be generated .
this procedure can be successfully applied by sharing the following quantum state : @xmath15 in the above equation @xmath8 is the number of end - nodes in the distributed system and @xmath16 is a known quantum state for producing sufficiently random numbers between the nodes . in general , a secret sharing procedure can be effectively used to generate a common coin between the distributed nodes in the presence of a limited set of faulty nodes . for this procedure ,
one of the end - nodes can be selected as a dealer to share known quantum state @xmath16 . to avoid cheating from a faulty dealer , the verifiable version of secret
sharing has been employed .
the original state @xmath16 is encoded by the dealer into a multi - qubit register : @xmath17 in the above equation , @xmath18 is the encoded state of the original state by using a quantum error correction code such as quantum reed - solomon code @xcite .
as described in the next sections , the encoded state is shared by the dealer and if the dealer is honest , the shared state can verified by all the honest nodes . for this procedure ,
a quantum error correction scheme is employed .
the encoded qubits are shared between all nodes by using a variant of verifiable quantum secret sharing ( vqss ) described in the next section @xcite .
in the modified version of vqss ( known as graded vqss ) , the players independently run three phases : the sharing phase , the verification phase and the measurement phase @xcite . in general , after sharing and verification of the known state @xmath16 , the players measure their remaining qubits .
the results produce sufficient random numbers .
the only difference between the original vqss and graded vqss is replacement of ideal broadcast in @xcite with the gradecast procedure in @xcite .
the overall analysis of all phases of quantum aided byzantine agreement will be presented in the next section .
the protocol as defined calls for p - level quantum variables ( qupits ) with the minimum prime number @xmath19 which is larger than number of nodes @xmath8 ( @xmath20 ) . in our implementation we encode them in @xmath21 qubits , because bell pairs support only distribution of qubits , and error correction circuits for addition , and physical systems are all best developed for two - level systems .
there are two sub - phases to sharing , with an agreed upon security parameter @xmath22 .
this parameter is employed to improve the probability of capturing a dishonest dealer with probability on the order of @xmath23 .
in the first level , @xmath24 qupit registers are prepared by the dealer .
the sharing phase continues by encoding and sharing the received qubits between all nodes . in this phase
a two level tree is created and distributed among the nodes . in the first level ,
@xmath24 qupit registers are prepared by the dealer .
the first register is the encoded state of the state @xmath16 ( @xmath18 ) . in the second level ,
the prepared system is also encoded and distributed by other nodes using additional qupit systems . as shown in fig .
[ fig : vqss - phases-01 ] , the protocol starts with a dealer s action for sharing the qupits of encoded @xmath16 ( @xmath18 ) .
the dealer is considered to be one of the participant nodes in the distributed system and like any other player may be traitorous .
the sharing phase continues by encoding and sharing the received qubits between all nodes .
after sharing and encoding the known state by the dealer and then the others , all the non - faulty nodes need to be confident about the correctness of the shared quantum state .
a verification phase is executed to ensure that the dealer has shared a recoverable quantum state .
the verification phase is applied by each node separately . in this phase ,
all nodes apply local quantum operations on their shares .
each node measures some of the remaining qubits , step by step .
the results of the measurements will be distributed by using a classical gradecast protocol . in the measurement phase
, the remaining qubits will be measured by all nodes and distribute the sufficiently random numbers as the required routine in the oblivious common coin procedure .
to establish the node architecture , we extracted the requirements for a minimum setup for vqss .
we require 5 nodes and need to set the prime number @xmath19 to be 7 for this setup . the security parameter ( @xmath22 ) is selected to have the value two .
these parameters lead to a minimum classical - quantum distributed system with five nodes which tolerates one malicious node inside the system .
the communication design of the system for execution of qba protocol is based on the following assumptions : * both reliable classical and quantum channels exist between each pair of nodes . * for quantum communication ,
the communication resource is bell pairs .
we focus on bell states instead of w state or ghz states in consideration of the capabilities of quantum repeater networks . *
the bell pairs may have fidelity @xmath25 .
the above assumptions result in a complete graph topology .
although the protocol requires expected constant number of rounds to be successfully completed , each round is computationally intensive , in addition to the high rate quantum and classical communications .
this requires long - live connections between each pair of nodes .
since the encoding and sharing scheme for the data qubits is generally similar for the dealer and the other players , the general architecture is the same .
each player applies the encoding scheme as shown in fig .
[ fig : encoder - arch-01 ] . as describe in the previous section , for the sharing phase
we need two level encoding .
the first one is applied by the dealer and the last one is applied by other players . before the verification phase
, each player has distributed the encoded qubits and in this phase the non - faulty nodes can verify the correctness of the original share ( encoded by the dealer ) . the required architecture for each node is shown in fig .
[ fig : verificationarch-02_1 ] . each player is required to have a quantum verifier module for every other node in the network .
one of the main part of the quantum verifier module is modular multiplication and addition ( @xmath26 ) described as below : @xmath27 the other components of quantum verifier consist of the quantum fourier transform ( @xmath28 ) and its inverse ( @xmath29 ) . the required module for these operations is presented in fig .
[ fig : verificationarch-02_1 ] . in the last phase ,
the only internal operation is measurement of the remaining qubits .
since it is not required to use any specific quantum operation , it may be considered as the simplest part of the protocol . but maintaining the remaining qubits is considered to be an important challenge for this phase .
the result of the measurement will be gradecast by the corresponding node .
in this section we present some techniques for implementing and optimizing the quantum architecture level and quantum circuit level . for the first level , we apply resource sharing in a pipelined approach , and for the last one we propose an efficient low - cost modulo 7 classical - quantum multiplier . to reduce the number of required quantum operations , reusing the quantum modules should be considered . in this technique
, it is not required that all the modules be available at the same time . in the verification circuit , since multiplication and addition ( @xmath26 ) and @xmath28 ( @xmath29 ) are dependent on complete execution of the previous @xmath26 and @xmath28 ( @xmath29 ) circuits , a resource sharing scheme has been designed to reduce the number of quantum modules and active number of qubits . as shown in fig .
[ fig : parallel-01 ] , we divide the execution time into 7 stages and in each stage a subsection of the required quantum operations is bound to the designed quantum modules .
although the overall execution time is increased by one stage in comparison to full circuit architecture execution time , the number of required quantum modules decreased dramatically . by applying this technique , in each verification module and in the worst case
, we need to implement two @xmath26 and one @xmath28 and/or @xmath29 , instead of 8 active @xmath26 modules , 4 active @xmath28 and @xmath29 modules .
in addition to reduction of the number of quantum modules , this optimization technique can reduce the required number of active qubits in the architecture . as shown in the previous section ,
the most challenging computational part of the quantum modules is the circuit for computation of multiplication and addition modulo @xmath19 ( @xmath26 ) .
the concept for standard design of quantum circuit for computation of modular multiplication is based on modular addition . motivated by designing efficient modular multiplication for the integer factorization application ,
many researchers have proposed low cost quantum circuits @xcite .
the standard approach for computation of @xmath30 ( for classical integers @xmath8 and @xmath31 and quantum variable @xmath32 ) is using computation of modular addition @xmath33 .
although this problem is similar to the state of the art modular multiplication ( and modular addition ) , the concept of computation and the resulting circuit may be different from the state of the art computations and related circuits .
the first difference is related to modulo computation . for integer factorization ,
the modulo @xmath8 is a parameter and can be changed for each new session .
in contrast , the modulo @xmath19 depends on the upper bound of number of the players in qba protocol , and can be considered a fixed number for a designed quantum distributed system .
the next difference is related to the classical operand @xmath31 in integer factorization and classical random numbers @xmath14 in modular multiplication and addition circuit of the verification module .
for the first one the operand can be implicitly considered in the circuit design , but for the last one , the random number must be present as a parameter for the modular multiplication .
these differences cause a need for rethinking novel quantum circuits to be used in the qba protocol .
thus we have proposed a multiplier modulo 7 for the minimum setup .
we havent followed the standard approach in integer factorization and attempted to compute the multiplication modulo 7 directly . the proposed modular multiplier is shown in fig .
[ fig : proposedmultiplier-03 ] . note that @xmath34 in the protocol and the enable classical signals for 3 swap gates ( @xmath35 ) and not gates ( @xmath36 )
are based on the following boolean functions : @xmath37 @xmath38 @xmath39 @xmath40
in this section the required quantum resources have been analyzed and enumerated for the minimum setup described in section [ sec : design ] .
for the analysis the following criteria have been considered : * number of qubits per node * number of quantum operations ( including total quantum cost and circuit depth ) * number of bell pairs consumed by the application for analysis of quantum cost in the end - nodes , we consider the basic quantum circuits to be used in the architecture .
the most complex quantum operations are in the verification phase .
we used the basic quantum circuits for 3-qubit @xmath28 and @xmath29 . as we described in the previous section ,
the most challenging operation is related to modular multiplication and addition ( @xmath26 ) .
we consider five different designs for analysis of the quantum depth and total quantum cost in the verifier module : * _ vbe96 : _ use the adder , modulo adder and modulo multiplier proposed in @xcite . *
_ cdkm04 : _ replace the the basic adder of @xcite by the second adder proposed in @xcite . *
_ vi05 : _ use quantum addition and modulo addition circuits of cdkm04 .
the only difference is replacement of the modulo multiplier in @xcite by the multiplier proposed in @xcite . *
_ custom : _ use the proposed multiplier modulo 7 instead of the previous standard design . *
_ custom+pipelined : _ apply the proposed pipeline technique and using the circuit in the previous design
. to evaluate the total cost of the required quantum operations , we estimate the required quantum gates based on the cost of cnot as the basic gate . due to reduction of the number of ancilla by using the proposed multiplier modulo 7 ,
we gain atleast 20% improvement in this parameter .
fig [ fig : result - both-01 ] plots the number of qubits and total depth of the quantum circuit . in this paper ,
quantum error correction has not been not applied for the architecture because the target was to present a lightweight and simple design .
this architecture can be effectively optimized to be demonstrated in experiments . in this situation ,
the error threshold for local gates must be considered as an important criterion .
we evaluate the error threshold based on the estimation of @xmath41 parameter , where @xmath42 is the total depth of the circuit and @xmath43 is the number of qubits in each node .
the result is shown in fig .
[ fig : threshold-01 ] .
to estimate traffic cost , we divide the result into two sections : the first one is related to the quantum communication cost which is based on transmission of bell pairs between each pair of nodes in the network .
we also consider the cost for sending classical messages during the execution of the protocol .
the results are shown in table [ tab : networkcost ] .
destination & quantum cost & classical cost + & & comm type & node & node & ( bell pairs ) & ( bits ) + 1(a ) & sharing & unicast & dealer & all nodes & 108 & 0 + 1(b ) & sharing & unicast & node(i ) & node ( j ) & 540 & 0 + 2(a ) & verification & broadcast & ttp & all nodes & 0 & 30 + 2(b ) & verification & gradecast & all nodes & all nodes & 0 & 14160 + 2(c ) & verification & broadcast & ttp & all nodes & 0 & 30 + 2(d ) & verification & gradecast & all nodes & all nodes & 0 & 4720 + 3(b ) & measurement & gradecast & all nodes & all nodes & 0 & 2360 + total & - & - & - & - & 648 & 21300 +
in this paper , we described an optimized quantum architecture for end - nodes in quantum aided byzantine agreement protocol .
the node architecture in the protocol is not as complex as that required for e.g. factoring a large number but at minimum of @xmath44 qubits per node , the experimental demands are substantial compared to current capabilities . in addition , the number of required bell pair is in the order of the number required for quantum key distribution . during the design process we found that modular classical - quantum multiplication and fully modular quantum addition are the most critical part of the quantum computation in the qba protocol .
this requires novel circuit design that is fully depends on the minimum prime number @xmath19 satisfy the inequality ( @xmath20 ) .
the other requirement is related to long lived qubits to maintain the shared and encoded qubits .
this work is supported by jsps kakenhi grant number 25280034 .
the authors would like to thank shota nagayama , shigeru yamashita , shigeya suzuki , takaaki matsuo and takahiko satoh for valuable technical conversations .
16 p. w. shor , `` algorithms for quantum computation : discrete logarithms and factoring . '' _ in proc .
35th annual symposium on foundations of computer science , _ 1994 . | quantum aided byzantine agreement ( qba ) is an important distributed quantum algorithm with unique features in comparison to classical deterministic and randomized algorithms , requiring only a constant expected number of rounds in addition to giving higher level of security . in this paper , we analyze details of the high level multi - party algorithm , and propose elements of the design for the quantum architecture and circuits required at each node to run the algorithm on a quantum repeater network .
our optimization techniques have reduced the quantum circuit depth by 44% and the number of qubits in each node by 20% for a minimum five - node setup compared to the design based on the standard arithmetic circuits . these improvements lead to an architecture with @xmath0 per node and error threshold @xmath1 for the total nodes in the network .
the evaluation of the designed architecture shows that to execute the algorithm once on the minimum setup , we need to successfully distribute a total of 648 bell pairs across the network , spread evenly between all pairs of nodes .
this framework can be considered a starting point for establishing a road - map for light - weight demonstration of a distributed quantum application on quantum repeater networks .
january 2017 _ keywords _ : distributed quantum algorithms , byzantine agreement , quantum repeater network ( qrn ) |
the terminology `` c - free independence''(or c - freeness ) was first used in the 1990 s by m. bozejko , r. speicher and m.leinert ( see @xcite , @xcite ) to denote a relation similar to d .- v .
voiculescu s free independence , but in the framework of algebras endowed with two linear functionals ( see definition [ def:1 ] below ) .
the additive c - free convolution and the analytic characterization of the correspondent infinite divisibility was described in 1996 ( see @xcite ) ; appropriate instruments for dealing with the multiplicative c - free convolution appeared a decade later , in @xcite . there , for @xmath5 a non - commutative random variable , we define an analytic function @xmath6 , inspired by voiculescu s @xmath4-transform , such that if @xmath5 and @xmath7 are c - free , then @xmath8 .
alternate proofs of this result were given in @xcite and @xcite .
the present work discusses the multiplicative c - free convolution in the framework of @xcite , namely when one of the functionals is replaced by a linear map with values in a ( not necessarily commutative ) banach algebra .
in particular , we show that the combinatorial methods from @xcite can be adapted to this more general framework .
notably , in the case of free independence over some banach algebra , as showed in @xcite , @xcite , the analogue of voculescu s @xmath9-transform satisfies a `` twisted multiplicative relation '' , namely @xmath10 .
the main result of the present work , theorem [ mainthm ] , shows that the non - commutative @xmath11-transform satisfies the usual multiplicative relation : @xmath12 for @xmath13 c - free non - commutative random variables .
the paper is organized as follows .
section 2 presents some preliminary notions and results , mainly concerning the lattice of non - crossing linked partitions and its connection to free and c - free cumulants and to @xmath14- and @xmath15-coefficients .
the main result of the section is proposition [ vanishtcoeff ] , the characterization of c - freeness in terms of @xmath15-coefficients .
section 3 restates the results on planar trees used in @xcite and utilize them for the main result of the paper , theorem [ mainthm ] .
section 4 discusses some aspects of infinite divisibility for the multiplicative c - free convolution in operator - valued framework .
the results from the scalar case ( see @xcite ) can be easily extended to the framework of a commutative algebra of operators , but they are generally not valid in the non - commutative case .
in the first part of this paper we will consider @xmath16 and @xmath17 to be two unital banach algebras and @xmath18 , respectively @xmath19 to be two unital and linear maps .
if @xmath16 and @xmath17 are c@xmath20- or von neumann algebras , then we will require that @xmath3 , respectively @xmath1 to be positive , respectively completely positive .
[ def:1 ] @xmath21 and @xmath22 , two unital subalgebras of @xmath16 , are said to be _ c - free _ with respect to @xmath23 if for all @xmath24 and all @xmath25 such that @xmath26 and @xmath27 with @xmath28 and @xmath29 we have that 1 .
@xmath30 2 .
@xmath31 .
two elements , @xmath5 and @xmath32 of @xmath16 are said two be _ c - free _ with respect to @xmath33 if the unital subalgebras of @xmath34 generated by @xmath5 and @xmath7 are c - free , as above .
if @xmath16 is a c@xmath2- or a von neumann algebra , then we will require that the unital c@xmath2- , respectively von neumann subalgebras of @xmath16 generated by @xmath5 and @xmath7 are c - free .
if only condition ( i ) holds true , then the subalgebras @xmath35 ( respectively the elements @xmath13 ) are said to be _
free _ with respect to @xmath36 .
as shown in @xcite , @xcite , there is a convenient combinatorial characterization of c - freeness in terms of non - crossing partitions , that we will summarize below . a non - crossing partition @xmath37 of the ordered set @xmath38 is a collection @xmath39 of mutually disjoint subsets of @xmath38 , called blocks , such that their union is the entire set @xmath40 and there are no crossings , in the sense that there are no two blocks @xmath41 and @xmath42 such that @xmath43 and @xmath44 .
* example 1 * : below is represented graphically the non - crossing partition + @xmath45 : @xmath46 the set of all non - crossing partitions on the set @xmath40 will be denoted by @xmath47 .
it has a lattice structure with respect to the reversed refinement order , with the biggest , respectively smallest element @xmath48 , respectively @xmath49 .
for @xmath50 we will denote by @xmath51 their join ( smallest common upper bound ) . for @xmath52 , a block @xmath53 of @xmath37 will be called _ interior _ if there exists another block @xmath54 and @xmath55 such that @xmath56 .
a block will be called _ exterior _ if is not interior .
the set of all interior , respectively exterior blocks of @xmath37 will be denoted by @xmath57 , respectively @xmath58 .
the set @xmath59 is totally ordered by the value of the first element in each block . for @xmath60 , we define the free , respectively c - free , cumulants @xmath61 , respectively @xmath62 as the multilinear maps from @xmath63 to @xmath64 , respectively @xmath17 , given by the recurrences below : @xmath65 \cdot [ \prod_{\substack { d \in \text{int } ( \gamma ) \\
d = ( i_1 , \dots , i_s ) } } \kappa_s ( x _ { i_1 } , \dots , x _ { i_s } ) ] \end{aligned}\ ] ] with the convention that if @xmath66 is a totally ordered set and @xmath67 is a collection of elements from @xmath17 , then @xmath68 we will us the shorthand notations @xmath69 for @xmath70 and @xmath71 for @xmath72 . as shown in @xcite , @xcite , and in @xcite , if @xmath73 and @xmath74 are c - free , then @xmath75 where , for @xmath76 , we let @xmath77 and @xmath78 .
that is , the mixed free and c - free cumulants in @xmath73 and @xmath74 vanish .
we will prove next a result analogous to theorem 14.4 from @xcite , more precisely a lemma about c - free cumulants with products as entries , in fact an operator - valued version of lemma 3.2 from @xcite .
the kreweras complementary @xmath79 of @xmath80 is defined as follows ( see @xcite , @xcite ) . consider the symbols @xmath81 such that @xmath82 .
then @xmath83 is the biggest element of @xmath84 such that @xmath85 is an element of @xmath86 .
the total number of blocks in @xmath37 and @xmath87 is @xmath88 . for @xmath89 and @xmath90
, we will denote by @xmath91 $ ] the block of @xmath92 that contains @xmath93 .
[ lemma:1 ] suppose that @xmath13 are two c - free elements of @xmath16
. then 1 .
@xmath94 2 .
@xmath95 | } ( x ) \cdot { { } ^c\kappa } _ { | \text{kr } ( \gamma ) [ \overline{n } ] | } ( y ) \cdot \prod _ { \substack { b \in \gamma\\ b \neq \gamma [ 1 ] } } \kappa _ { | b | } ( x ) \cdot \prod_{\substack { d \in \text{kr}(\gamma ) \\
d \neq \text{kr } ( \gamma ) [ \overline{n } ] } } \kappa _ { | d | } ( y ) $ ] part ( i ) is shown in @xcite , theorem 14.4
. we will show part ( ii ) by induction on @xmath24 . for @xmath96 ,
the statement is trivial , since @xmath97 from the c - freeness of @xmath5 and @xmath7 , therefore @xmath98 for the inductive step , in order to simplify the writting , we will introduce several new notations .
let @xmath99 elements from the same block of @xmath100 have the same parity @xmath101 . for @xmath102 , denote @xmath103 , respectively @xmath104 the restriction of @xmath105 to the even , respectively odd , numbers and define @xmath106 also , we will need to consider the mappings @xmath107 the juxtaposition of partitions , and @xmath108 constructed by doubling the elements , that is if @xmath109 is a block of @xmath110 , then @xmath111 is a block of @xmath112 .
then , for @xmath113 , we define @xmath114 & = \prod_{\substack { c \in \pi \\ c = ( i_1 , \dots , i_l ) } } \kappa_l ( x _ { i_1 } , \dots , x _ { i_l } ) \\ { \mathcal{k}}_\pi [ x_1 , \dots , x_n ] & = \prod_{\substack { b \in \text{ext } ( \gamma ) \\
b = ( j_1 , \dots , j_l ) } } { { } ^c\kappa}_l ( x _ { j_1 } \cdots x _ { j_l } ) ] \cdot [ \prod_{\substack { d \in \text{int } ( \gamma ) \\
d = ( i_1 , \dots , i_s ) } } \kappa_s ( x _ { i_1 } , \dots , x _ { i_s } ) \end{aligned}\ ] ] remark that @xmath115 and @xmath116 .
also , if @xmath117 is an exterior block of @xmath118 , then @xmath119 so lemma [ lemma:1 ] is equivalent to @xmath120 & = \sum_{\substack{\sigma\in nc_{s}(2n)\\ \sigma\bigvee\widehat{0_{n}}=\widehat{\pi } } } \kappa_{\sigma}[x , y,\dots , x , y ] \\ { \mathcal{k}}_{\pi}[xy,\dots , xy ] & = \sum_{\substack{\sigma\in nc_{s}(2n)\\ \sigma\bigvee\widehat{0_{n } } = \widehat{\pi } } } { \mathcal{k}}_{\sigma}[x , y,\dots , x , y],\end{aligned}\ ] ] therefore@xmath121 .
$ ] on the other hand , @xmath122.\ ] ] since the mixed cumulants vanish , the equation above becomes @xmath123\\ & = & \sum_{\sigma\in nc_{0}(2n)}{\mathcal{k}}_{\sigma}[x , y,\dots , x , y]+\sum_{\substack{\sigma\in nc_{s}(2n)\\ \sigma\notin nc_{0}(2n ) } } { \mathcal{k}}_{\sigma}[x , y,\dots , x , y ] \end{aligned}\ ] ] but @xmath124 , and , for @xmath125 , one has that @xmath126 if and only if @xmath127 , therefore : @xmath128 henceforth @xmath129 & = & \sum_{\substack{\pi\in nc(n)\\ \pi\neq\mathbbm{1}_{n } } } \sum_{\substack{\sigma\in nc_{s}(2n)\\ \sigma\bigvee\widehat{0_{n}}=\widehat{\pi } } } { \mathcal{k}}_{\sigma}[x , y,\dots , x , y]\\ & = & \sum_{\substack{\pi\in nc(n)\\ \pi\neq\mathbbm{1}_{n } } } { \mathcal{k}}[xy,\dots , xy ] . \end{aligned}\ ] ] so the proof is now complete . @xmath130
+ by a non - crossing linked partition @xmath118 of the ordered set @xmath38 we will understand a collection @xmath131 of subsets of @xmath38 , called blocks , with the following properties : 1 .
@xmath132 2 .
@xmath133 are non - crossing .
3 . for any @xmath134 ,
the intersection @xmath135 is either void or contains only one element .
if @xmath136 , then @xmath137 and @xmath138 is the minimal element of only one of the blocks @xmath139 and @xmath140 . for @xmath100 as above
we define @xmath141 to be the set of all @xmath142 such that there are no blocks of @xmath118 whose minimal element is @xmath143 .
a block @xmath144 of @xmath118 will be called _ exterior _ if there is no other block @xmath145 of @xmath118 containing two elements @xmath146 such that @xmath147 .
the set of all non - crossing linked partitions on @xmath148 will be denoted by @xmath149 . * example 2 * :
below is represented graphically the non - crossing linked partition @xmath150 from @xmath151 .
its exterior blocks are @xmath152 and @xmath153 .
@xmath154 similarly to @xcite and @xcite , we define the @xmath155-coefficients , respectively @xmath156-coefficients , as follows .
take @xmath157 .
then , for @xmath24 a positive integer , the maps @xmath158 and @xmath159 are given by the following recurrences : @xmath160\\\ ] ] respectively @xmath161 to simplify the writing we will use the shorthand notations @xmath162 $ ] , respectively @xmath163 $ ] for the summing term of the right - hand side of ( [ eq:31 ] ) , respectively ( [ eq:41 ] ) ; also , if @xmath164 , we will use the shorter notations @xmath165 $ ] , @xmath166 $ ] respectively @xmath167 , @xmath168 .
note that all the factors in @xmath169 are @xmath155-coefficients , but the development of @xmath170 contains both @xmath156- and @xmath155-coefficients .
on the set @xmath149 we define a order relation by saying that @xmath171 if for any block @xmath172 of @xmath118 there exist @xmath173 blocks of @xmath105 such that @xmath174 .
with respect to the order relation @xmath175 , the set @xmath149 is a lattice that contains @xmath47 .
we say that @xmath176 and @xmath138 are _ connected _ in @xmath177 if there exist @xmath178 blocks of @xmath118 such that @xmath179 , @xmath180 and @xmath181 , @xmath182 . for every @xmath177 there exist a unique partition @xmath183 defined as follows : @xmath176 and @xmath138 are in the same block of @xmath184 if and only if they are connected in @xmath118 .
we will use the notation @xmath185=\{\sigma\in ncl(n ) : c(\sigma)=c(\pi)\}.\ ] ] in example 2 from above , we have @xmath186 for @xmath187 . for every @xmath188 , the set @xmath189 $ ] is a sublattice of @xmath149 with maximal element @xmath37 .
moreover , if @xmath37 has the blocks @xmath190 , each @xmath139 of cardinality @xmath191 , then we have the following ordered set isomorphism : @xmath192 \simeq [ \mathbbm{1}_{k_1}]\times\cdots\times[\mathbbm{1}_{k_s}]\ ] ] [ cum - tcoeff ] for any positive integer @xmath193 and any @xmath194 we have that @xmath195}t_\pi[x_1,\dots , x_n]\\ { { } ^c\kappa}_n(x_1,\dots , x_n ) & = \sum_{\pi\in[\mathbbm{1}_n]}{{}^ct}_\pi[x_1,\dots , x_n]\end{aligned}\ ] ] first relation is shown in proposition 1.4 from @xcite .
the second relation is trivial for @xmath96 . for @xmath196 , note that @xmath197 = \sum_{\gamma\in nc(n)}\sum_{\pi\in[\gamma ] } { { } ^ct}_\gamma[x_1,\dots , x_n]\ ] ] and the similar relation for @xmath169 .
suppose that @xmath198 and @xmath89 are such that @xmath199 $ ] .
from the definition of @xmath156- and @xmath155-coefficients we have that @xmath200 = \prod _ { \substack { b \in \text{ext } ( \pi ) \\ b = ( i_1 , \dots , i_s ) } } { { } ^ct } _ { \pi _ { | b } } [ x _ { i_1 } , \dots , x _ { i_s } ] \cdot \prod _ { \substack { d \in \text{int } ( \pi ) \\ d = ( j_1 , \dots , j_t ) } } t _ { \pi _ { | d } } [ x _ { j_1 } , \dots , x _ { j_t } ] \ ] ] and @xmath201 = \prod _ { \substack { b \in \pi ) \\ b = ( i_1 , \dots , i_s ) } } { { } ^ct } _ { \pi _ { | b } } [ x _ { i_1 } , \dots , x _ { i_s } ] .\ ] ] therefore , the equation ( [ eq:32 ] ) becomes @xmath202 } } { { } ^ct } _ { \pi _ { | b } } [ x _ { i_1 } , \dots , x _
{ i_s } ] ) \\ & \cdot \prod
_ { \substack { d \in \text{int}(\pi ) \\ d = ( j_1 , \dots , j_l ) } } ( \sum _ { \substack { \pi \in ncl(n ) \\ \pi \in [ \gamma ] } } t _ { \pi _ { | d } } [ x _ { j_1 } , \dots , x _ { j_l } ] ) ] \end{aligned}\ ] ] and the factorization ( [ factorization ] ) gives : @xmath203 } { { } ^ct } _ { \sigma } [ x _ { i_1 } , \dots , x _ { i_s } ] ) \\ & \cdot \prod _
{ \substack { d \in \text{int}(\pi ) \\ d = ( j_1 , \dots , j_l ) } } ( \sum _ { \sigma\in[\mathbbm{1}_l ] } t _
{ \sigma } [ x _ { j_1 } , \dots , x _ { j_l } ] ) ] \end{aligned}\ ] ] the conclusion follows now utilizing the moment - cumulant recurrence and induction on @xmath193 . for @xmath204 , we define the formal power series @xmath205 and @xmath206 . also , we consider the moment series of @xmath5 , namely @xmath207 , and @xmath208 . as
shown in @xcite , lemma 7.1(i ) , the recurrence [ eq:31 ] gives that @xmath209 the proposition below gives an analogous relation for the series @xmath210 ( i. e. a non - commutative analogue of lemma 7.1(ii ) from @xcite ) .
[ analytic - ct ] with the notations above , we have that @xmath211 let @xmath212 .
note that for each @xmath213 , there exists a unique triple @xmath214 , @xmath215 and @xmath216 such that @xmath217 .
indeed , taking @xmath93 to be the maximal element of the block of @xmath218 containing 1 , it follows that @xmath219 and that @xmath220 .
conversely , each triple @xmath221 as above determine a unique @xmath222 , hence @xmath223 therefore the recurrence [ eq:32 ] gives @xmath224 \nonumber \\ & = \sum _ { p \leq n } [ \sum _ { \pi \in ncl(p ) } { { } ^ct}_\pi(x ) \cdot \phi(x^ { n - p } ) ] \label{new:01}. \end{aligned}\ ] ] let @xmath225 .
fix @xmath226 and let @xmath227 be the exterior block of @xmath100 .
define @xmath228 such that , with the notations following the recurrences ( [ eq:31])([eq:32 ] ) , @xmath229 , as follows : 1 .
if @xmath230 is an interior block of @xmath100 , then @xmath231 is a block in @xmath232 ; 2 .
if @xmath233 and the only block of @xmath100 containing @xmath234 is exterior , then @xmath235 is a block of @xmath232 i.e. @xmath232 is obtaining by `` deleting '' the 1 and the exterior block of @xmath100 .
for each @xmath236 in the exterior block @xmath237 of @xmath100 we define @xmath238 to be the maxinal element connected to @xmath236 in @xmath100 , and let @xmath239 .
then each set @xmath240 is nonvoid and we have the decomposition @xmath241 * example 3 . *
if @xmath242 , then @xmath243 see the diagram below : @xmath244 using the equality @xmath229 , we obtain @xmath245 since @xmath246 , the equations ( [ new:01 ] ) and ( [ new:02 ] ) give that @xmath247 which is the relation of the left hand side and right hand side coefficients of @xmath248 in equation ( [ eq:39 ] ) .
we conclude this section with the following result .
[ vanishtcoeff]_(characterization of c - freeness in terms of @xmath15-coefficients ) _ + two elements @xmath249 from @xmath250 are c - free if and only if all their mixed @xmath155- and @xmath156-coefficients vanish , that is for all @xmath251 and all @xmath252 such that @xmath253 and @xmath254 for some @xmath255 we have that @xmath256 we will show by induction on @xmath24 the equivalence between vanishing of mixed free and c - free cumulants of order @xmath24 in @xmath5 and @xmath7 and vanishing of mixed @xmath155- and @xmath156-coefficients of order up to @xmath257 . for @xmath258 the result is trivial , since @xmath259 and @xmath260 .
for the inductive step suppose that @xmath261 are not all @xmath5 nor all @xmath7 .
proposition [ cum - tcoeff ] gives @xmath262\\ \pi\neq \mathbbm{1}_n } } t_\pi [ a_1,\dots , a_n ] \\ { { } ^c\kappa}_n ( a_1 , \dots , a_n ) = & { { } ^ct}_{n-1}(a_1,\dots , a_n)+ \sum_{\substack{\pi\in[\mathbbm{1}_n]\\ \pi\neq \mathbbm{1}_n } } { { } ^ct } _ { \pi } [ a_1 , \dots , a_n ] .\end{aligned}\ ] ] fix @xmath263 $ ] , @xmath264 .
since @xmath100 is connected , there is @xmath265 , a block of @xmath100 with @xmath266 such that @xmath267 are not all @xmath5 not all @xmath7 , so equations ( [ eq:24 ] ) and ( [ eq:25 ] ) and the induction hypothesis imply that @xmath268 , hence we have that @xmath269 and @xmath270 so q.e.d ..
in this section we will use the combinatorial arguments from @xcite to show that whenever @xmath272 and @xmath7 are two c - free elements from @xmath250 , we have that @xmath273 and @xmath274 where , for @xmath275 , we define @xmath276 , respectively @xmath277 @xmath130 we will start with a review of the notations and results from @xcite . by an _ elementary planar tree
_ we will denote a graph with @xmath278 vertices , @xmath279 , and @xmath280 ( possibly 0 ) edges , or branches , connecting the vertex @xmath281 ( that we will call _ root _ ) to the vertices @xmath282 ( that we will call _ offsprings _ ) . by a _
planar tree _ we will understand a graph consisting in a finite number of _ levels _ , such that : 1 .
first level consists in a single elementary planar tree , whose root will be considered the root of the planar tree ; 2 .
the @xmath143-th level will consist in a set of elementary planar trees such that their roots are offsprings from the @xmath283-th level .
below are represented graphically the elementary planar tree @xmath284 and the 2-level planar tree @xmath285 : @xmath286 the set of all planar trees with @xmath24 vertices will be denoted by @xmath287 . if @xmath288 is a planar tree , the set of elementary trees composing it will be denoted by @xmath289 , the elementary tree containing the root of @xmath290 will be denoted by @xmath291 , and we define @xmath292 . on the set of vertices of a planar tree we consider
the `` _ _ left depth first _ _ '' order from @xcite , given by :
1 . roots are less than their offsprings ; 2 .
offsprings of the same root are ordered from left to right ; 3 .
if @xmath293 is less that @xmath294 , then all the offsprings of @xmath293 are smaller than any offspring of @xmath294 .
( i.e. the order in which the vertices are passed by walking along the branches from the root to the right - most vertex , not counting vertices passed more than one time , see the example below ) . * example 4 * : @xmath295 next , consider , as in @xcite , the map @xmath296{\longrightarrow}\mathfrak{t}(n)$ ] such that @xmath297 is the unique planar tree with the property that if @xmath298 is a block of @xmath118 , then the vertices of @xmath299 form an elementary tree from @xmath300 . more precisely ,
if @xmath301 is the block of @xmath118 containing 1 , then the first level of @xmath297 is the elementary planar tree of root numbered 1 and @xmath302 offsprings numbered @xmath303 .
the second level of @xmath297 consists on the elementary trees representing the blocks ( if any ) having @xmath304 as first elements etc ( see example 5 below ) .
as shown in @xcite , the map @xmath305 is well - defined and bijective .
* example 5 * : for @xmath204 , define the maps @xmath306 , respectively @xmath307 from @xmath308 to @xmath64 , respectively to @xmath17 , as follows
. if @xmath288 is an elementary planar tree with @xmath309 vertices , then let @xmath310 and @xmath311 ; for @xmath312 , let @xmath313 as in @xcite , proposition [ cum - tcoeff ] and the bijectivity @xmath314 give that @xmath315 by a _
bicolor elementary planar tree _ we will understand an elementary tree together with a mapping from its offsprings to @xmath316 such that the offsprings whose image is 1 are smaller than ( with respect to the order relation considered above , i.e. at right of ) the offsprings with image 0 .
branches toward offsprings of color 0 , respectively 1 , will be also said to be of color 0 , respectively 1 .
the set of all bicolor planar trees with @xmath193 vertices will be denoted by @xmath317 .
following @xcite , branches of color 1 will be represented by solid lines and branches of color 0 by dashed lines .
* example 6 * : the graphical representation of @xmath318 : a planar tree whose constituent elementary trees are all bicolor will be called _ bicolor planar tree _ ; the set of all bicolor planar trees will be denoted by @xmath319 .
let @xmath320 elements from the same block of @xmath100 have the same parity @xmath101 .
we will say that the blocks of a partition from @xmath321 with odd elements are of color 1 and the ones with even elements are of color 0 ; blocks of color 1 will be graphically represented by solid lines and blocks of color 0 by dashed lines .
* example 7 * : representation of @xmath322 : as shown in @xcite , there exist a bijection @xmath323 , constructed as follows ( see also example 8 below ) : 1 . if @xmath324 and @xmath325 are the two exterior blocks of @xmath118 , where @xmath326 , then the first level of @xmath327 is the elementary tree with @xmath328 offsprings , the first @xmath329 of color 1 , representing @xmath330 , in this order , and the last @xmath331 of color 0 , representing @xmath332 , in this order .
2 . suppose that @xmath333 and @xmath334 are consecutive elements in a block @xmath172 of @xmath118 already represented in some tree of @xmath327 .
let @xmath335 and @xmath336 if @xmath337 is the minimal element of some block @xmath338 of @xmath100 other than @xmath17 and @xmath339 otherwise .
let @xmath340 be , in this order , the exterior blocks with more than one element of @xmath341 such that each @xmath342 has @xmath343 elements . finally ,
if @xmath344 is the minimal elements of some block @xmath345 of @xmath100 , then let @xmath346 ( otherwise let @xmath347 )
. then we represent @xmath348 by an elementary tree with root the vertex representing @xmath349 and of @xmath350 offsprings , where the first @xmath351 are representing , and keeping the color of , the vertices of @xmath352 except for the minimal one , the next @xmath353 representing elements of @xmath354 except for the minimal one etc .
* example 8 * : for @xmath355 , we define the maps @xmath356 and @xmath357 from @xmath358 to @xmath359 , respectively @xmath360 as follows .
if @xmath361 has @xmath362 offsprings of color 1 and @xmath363 offsprings of color 0 , then we define @xmath364 for @xmath365 , define @xmath366 remark that , for @xmath367 , the definitions of @xmath368 and @xmath369 , @xmath370 give @xmath371\kappa_{\pi_+}[y],\ ] ] respectively @xmath372 | } ( x ) \cdot { { } ^c\kappa } _ { | \pi_+ [ 2n ] | } ( y ) \cdot
\prod _ { \substack { b \in \pi _ { - } \\ b \neq \pi _ { - } [ 1 ] } } \kappa _ { | b | } ( x ) \cdot \prod_{\substack { d \in \pi _ { + } \\ d \neq \pi _ { + } [ 2n ] } } \kappa _ { | d | } ( y ) \ ] ] [ mainthm ] if @xmath374 are c - free elements from @xmath375 , then @xmath376 and @xmath377 .
we need to show that , for all @xmath378 @xmath379 if @xmath380 denotes the elementary planar tree with @xmath24 vertices , with the notations from the previous two sections , the equations ( [ final ] ) are equivalent to @xmath381 i.e. for example , @xmath382 we will prove ( [ finaltree ] ) by induction on @xmath24 . for @xmath383 ,
the result is trivial .
suppose ( [ finaltree ] ) true for @xmath384 .
relation ( [ eq:10 ] ) and lemma [ lemma:1 ] give @xmath385 | } ( x ) \cdot { { } ^c\kappa } _ { | \pi_+ [ 2n ] | } ( y ) \cdot \prod _ { \substack { b \in \pi _ { - } \\ b \neq \pi _ { - } [ 1 ] } } \kappa _ { | b | } ( x ) \cdot \prod_{\substack { d \in \pi _ { + } \\ d \neq \pi _ { + } [ 2n ] } } \kappa _ { | d | } ( y ) \end{aligned}\ ] ] and equation ( [ eq:12 ] ) and the bijectivity of @xmath368 give : @xmath386 similarly , we have that @xmath387 all non - elementary trees from @xmath388 are composed of elementary trees with less than @xmath24 vertices .
the relations ( [ eq1:7 ] ) and ( [ eq1:8 ] ) imply that the image under @xmath389 , respectively @xmath390 of any such tree is the sum of the images under @xmath391 , respectively under @xmath392 of its colored versions .
hence @xmath393 finally , ( [ eq1:7 ] ) , ( [ eq1:8 ] ) and the two equations above give that @xmath394 which imply ( [ finaltree ] ) .
fix a unital @xmath395-subalgebra @xmath172 of @xmath396 , the @xmath395-algebra of bounded linear operators on a hilbert space @xmath397 . in this section ,
we study the infinite divisibility relative to the c - freeness .
a natural framework for such a discussion is in a _
c - free probability space _ @xmath398 , that is , the algebras @xmath399 is also a concrete @xmath395-algebra acting on some hilbert space @xmath400 , the linear map @xmath401 is a unital completely positive map , and the expectation functional @xmath36 is a state on @xmath402 .
note that we have the norm @xmath403 .
the _ distribution _ of a unitary @xmath404 , written as the spectral integral @xmath405 is the pair @xmath406 , where @xmath407 is a positive borel probability measure on the circle @xmath408 , and @xmath409 is a linear map from @xmath410 $ ] , the ring of laurent polynomials , into the @xmath395-algebra @xmath172 such that @xmath411.\ ] ] of course , the positivity of @xmath412 and the stinespring theorem ( see @xcite , theorem 3.11 ) imply that the map @xmath409 extends to a completely positive map on @xmath413 , the @xmath395-algebra of continuous functions on @xmath414 .
more generally , given any sequence @xmath415 , it is known ( see @xcite ) that the _ operator - valued trigonometric moment sequence _
@xmath416 extends linearly to a completely positive map @xmath417 if and only if the operator - valued power series @xmath418 converges on the open unit disk @xmath419 and satisfies @xmath420 for @xmath421 .
in particular , for the unitary @xmath422 this implies that its moment generating series @xmath423 and @xmath424 satisfy the properties : @xmath425 for @xmath426 .
thus , the formula @xmath427 defines an analytic function from the disk @xmath419 to the algebra @xmath172 , with the norm @xmath428 for @xmath421 .
analogously , the function @xmath429 will be an analytic self - map of the disk @xmath419 .
notice that we have @xmath430 and @xmath431 .
conversely , suppose we are given two analytic maps @xmath432 and @xmath433 satisfying @xmath434 for @xmath426 .
then the maps @xmath435 and @xmath436 are well - defined in @xmath419 , and they can be written as the convergent power series : @xmath437 where the operators @xmath438 and the coefficients @xmath439 . since @xmath440 , we have @xmath441^ { * } \geq 0\ ] ] for every @xmath442 .
therefore , the solution of the operator - valued trigonometric moment sequence problem implies that the map @xmath443 extends linearly to a completely positive map from @xmath413 into @xmath172 . in the case of @xmath444
, we obtain a borel probability measure @xmath445 on @xmath414 satisfying .
the pair @xmath446 is uniquely determined by the analytic maps @xmath172 and @xmath447 .
it is now easy to construct a c - free probability space @xmath398 and a unitary random variable @xmath448 so that the distribution of @xmath422 is precisely the pair @xmath446 .
indeed , we simply let @xmath449 , whose members are viewed as the multiplication operators acting on the hilbert space @xmath450 , @xmath451 , and the variable @xmath422 can be defined as @xmath452 in summary , we have identified the distribution of @xmath422 with the pair @xmath453 of contractive analytic functions .
a unitary @xmath448 is said to be _ c - free infinitely divisible _ if for every positive integer @xmath193 , there exists identically distributed c - free unitaries @xmath454 in @xmath399 such that @xmath422 and the product @xmath455 have the same distribution .
it follows from the definition of the c - freeness that if a unitary @xmath456 , with the distribution @xmath446 , is c - free infinitely divisible , then the law @xmath445 must be infinitely divisible with respective to the _ free multiplicative convolution _
@xmath457 , that is , to each @xmath458 there exists a probability measure @xmath459 on @xmath414 such that @xmath460 the theory of @xmath457-infinite divisibility is well - understood , see [ 5 ] , and we shall focus on the c - free infinitely divisible distribution @xmath409 , or equivalently , on the function @xmath461 . from equation ( [ b - trans ] ) and
proposition [ analytic - ct ] , we have that the @xmath11-transform of @xmath422 satisfies @xmath462 therefore , theorem 2.1 yields immediately the following characterization of c - freely infinite divisibility . a unitary @xmath404 with distribution @xmath446 is c - free infinitely divisible if and only if @xmath445 is @xmath457-infinitely divisible and the function @xmath461 is infinitely divisible in the sense that to each @xmath458 , there exists an analytic map @xmath463 such that @xmath464^n,\quad z\in \mathbb{d}.$ ] it was proved in [ 5 ] that a @xmath457-infinitely divisible law @xmath445 is the haar measure @xmath465 on the circle group @xmath466\}$ ] if and only if @xmath445 has zero first moment .
we now show the c - free analogue of this result .
denote by @xmath446 the distribution of @xmath422 .
assume first that @xmath468 , hence the law @xmath445 equals @xmath465 .
the c - free infinitely divisibility of @xmath422 shows that there exist c - free and identically distributed unitaries @xmath471 and @xmath472 in @xmath399 such that @xmath473 and @xmath474 in distribution .
therefore , for @xmath475 , we have @xmath476 the case of @xmath477 follows from the identity @xmath478 . an interesting case for c - freely infinite divisibility arises from the commutative situation . to illustrate , suppose @xmath479 , the algebra of continuous complex - valued functions defined on a hausdorff compact set @xmath480 equipped with the usual supremum norm
denote by @xmath481 the family of all borel finite ( positive ) measures on @xmath414 , equipped with the weak*-topology from duality with continuous functions on @xmath414 .
then we shall have the following let @xmath445 be a @xmath457-infinitely divisible law on @xmath414 and @xmath482 .
a unitary @xmath404 is c - free infinitely divisible if and only if its @xmath11-transform admits the following lvy - hinin type representation : @xmath483 where the map @xmath484 is a continuous function from @xmath485 to the circle @xmath414 and the map @xmath486 is weak*-continuous from @xmath485 to @xmath481 .
the integral representation follows directly from the characterization of c - free infinite divisibility in the scalar - valued case @xcite .
to conclude , we need to show the continuity of the functions @xmath487 and @xmath488 . to this purpose , observe that @xmath489 in particular , we have @xmath490 thus , if @xmath491 is a net converging to a point @xmath492 , then the family @xmath493 is bounded , and for every @xmath442 we have @xmath494 as @xmath495 .
since bounded sets in @xmath481 are compact in the weak - star topology , the above convergence of poisson integrals determines uniquely the limit @xmath496 of the net @xmath497 .
we deduce that the measures @xmath497 converge weakly to @xmath487 , and therefore the function @xmath496 is continuous .
this and the identity @xmath498 imply further the continuity of the function @xmath488 .
we end the paper with the remark that if the algebra @xmath172 is non - commutative , a @xmath172-valued contractive map @xmath499 does not always have contractive analytic @xmath193-th roots for each @xmath500 , even when @xmath499 is invertible for every @xmath501 . to illustrate , consider the constant map @xmath502,\quad z \in \mathbb{d}.\ ] ]
then @xmath499 is an invertible contraction if the modulus @xmath503 is sufficiently small , and yet its square roots @xmath504\ ] ] have the same norm @xmath505 . | the paper discusses some results concerning the multiplication of non - commutative random variables that are c - free with respect to a pair @xmath0 , where @xmath1 is a linear map with values in some banach or c@xmath2-algebra and @xmath3 is scalar - valued . in particular , we construct a suitable analogue of the voiculescu s
@xmath4-transform for this framework . |
neutrinos can not be detected directly , because they do not ionize the materials they are passing through , and hence neutrino detectors are based on neutrino - nucleus interactions .
thus , a correct understanding of these interactions is crucial to minimize systematic uncertainties in neutrino oscillation experiments .
current and upcoming neutrino experiments ( sciboone , miniboone , t2k , miner@xmath0a , minos , lbne , microboone , ) to measure oscillation effects and neutrino interaction cross sections are / will be mostly sensitive to neutrino energies up to 10 gev . in this talk , we present results from a microscopic model @xcite limited to three - momentum and energy transfers below 1.2 gev .
double differential 2p2h cross section for neutrino - carbon interactions at energies of 1 , 3 , and 10 gev . the black contours show the location of the genuine qe events , while the white ones show lines of constant three - momentum transfer from 0.2 to 1.2 gev.,title="fig : " ] double differential 2p2h cross section for neutrino - carbon interactions at energies of 1 , 3 , and 10 gev .
the black contours show the location of the genuine qe events , while the white ones show lines of constant three - momentum transfer from 0.2 to 1.2 gev.,title="fig : " ] double differential 2p2h cross section for neutrino - carbon interactions at energies of 1 , 3 , and 10 gev .
the black contours show the location of the genuine qe events , while the white ones show lines of constant three - momentum transfer from 0.2 to 1.2 gev.,title="fig : " ]
thanks to the miniboone @xmath1 puzzle @xcite , the theoretical understanding of the so called ccqe - like reactions ( cc quasi - elastic neutrino - nucleus processes that do not produce a pion in the final state ) at intermediate neutrino energies ( @xmath2 gev ) has experienced an enormous boost in the recent years @xcite .
the absolute values of the ccqe cross section reported in @xcite were too large as compared to the consensus of theoretical predictions for the qe contribution @xcite . moreover ,
the cross section per nucleon on @xmath3c was clearly larger than for free nucleons , and a fit , using a relativistic fermi gas model , to the data led to an axial mass , @xmath4 gev @xcite much larger than the previous world average ( @xmath5 1.03 gev ) .
the inclusive cross section for the process @xmath6 is determined by the @xmath7 gauge boson selfenergy in the nuclear medium @xcite , and in particular for the different modes in which it can be absorbed . in the case of genuine qe events ,
the gauge boson @xmath7 is absorbed by just one nucleon , which together with a lepton is emitted ( see figure 1 in ref .
however , the qe - like sample includes also multinucleon events where the gauge boson is absorbed by two interacting nucleons ( in the many body language , this amounts to the excitation of a 2p2h nuclear component ) .
the consideration of the 2p2h contributions allows to describe @xcite the miniboone ccqe - like flux averaged double differential cross section @xmath8 @xcite with values of @xmath1 around 1 gev .
this can be seen in figure [ fig:1 ] , where we also show results for antineutrinos .
this is reassuring from the theoretical point of view and more satisfactory than the situation envisaged by some other works that described these ccqe - like data in terms of a larger value of @xmath1 of around 1.31.4 gev ( see the discussion in @xcite ) , difficult to accommodate with our current knowledge on the nucleon axial radius .
however , not only multinucleon mechanisms , but also rpa corrections turn out to be essential to obtain axial masses consistent with the world average .
this can be appreciated in the left panel of figure [ fig:2 ] , where we see that rpa strongly decreases the cross section at low energies , while multinucleon mechanisms accumulate their contribution at low muon energies and compensate for that depletion .
therefore , the final picture is that of a delicate balance between a dominant single nucleon scattering , corrected by collective effects , and other mechanisms that involve directly two or more nucleons .
both effects can be mimicked by using a large @xmath1 value ( red lines in the neutrino panels of figure [ fig:1 ] ) .
because of the multinucleon mechanisms , the neutrino energy reconstruction based on the qe kinematics is not totally reliable @xcite .
the energy of the neutrino that has originated a cc event is unknown , and it is common to define a reconstructed neutrino ( @xmath9 ) energy obtained from the measured angle and three - momentum of the outgoing charged lepton .
it corresponds to the energy of a neutrino that emits a lepton and a gauge boson that is being absorbed by a nucleon at rest .
each event contributing to the flux averaged cross section defines unambiguously a value of @xmath9 , however the actual energy of the neutrino that has produced the event will not be exactly @xmath10 .
thus , for each @xmath9 , there exists a distribution of true neutrino energies that could give rise to events whose outgoing charge lepton kinematics would lead to the given value of @xmath10 .
for genuine qe events , this distribution is sufficiently peaked around the true neutrino energy to make the used algorithm accurate enough to study the neutrino oscillation phenomenon @xcite or to extract neutrino flux unfolded ccqe cross sections from data @xcite . however , for 2p2h events included in the ccqe - like sample , there appears a long tail in the distribution of true energies associated to each @xmath9 that makes less reliable the qe based energy reconstruction procedure @xcite .
moreover , the unfolded ccqe - like cross section turns out not to be a very clean observable , since the unfolding procedure itself is model dependent and assumes that the events are purely qe . this is illustrated in the right panel of figure [ fig:2 ] , where different predictions from our model , together with the ccqe - like miniboone data are depicted . the unfolding procedure ( see ref .
@xcite ) does not appreciably distort the genuine qe events , however the situation is drastically different for the 2p2h contribution , and as result the miniboone unfolded cross section @xcite exhibits an excess ( deficit ) of low ( high ) energy neutrinos , which is an artifact of the unfolding process that ignores multinucleon mechanisms .
this systematic distortion of the energy spectrum will increase the uncertainty on the extracted oscillation signal .
we have extended to 10 gev the results from the microscopic model .
we find @xcite , limiting the calculation to three momentum transfers less than 1.2 gev , the 2p2h mechanisms produce a two dimensional distribution in momentum and energy transfer that is roughly constant as a function of energy ( figure [ fig:3 ] ) .
the 2p2h cross section scales approximately with the number of nucleons for isoscalar nuclei , and becomes around 25% ( 33% ) of the qe cross section for 3 gev neutrinos ( antineutrinos ) . the distortion of the energy and @xmath11 spectra using the qe kinematic reconstruction , large at 1 gev and below ,
steadily decreases as a function of the neutrino energy . when confronted with the miner@xmath0a data @xcite , the model has the qualitative features and magnitude to provide a reasonable agreement ( figure [ fig:4 ] ) .
neutrino ( left ) and antineutrino ( right ) differential @xmath11 distributions with 2p2h and qe with rpa effects calculated at 3 gev ( solid line ) and compared to miner@xmath0a data @xcite .
dot - dashed lines stand for results without rpa and without 2p2h effects.,title="fig : " ] neutrino ( left ) and antineutrino ( right ) differential @xmath11 distributions with 2p2h and qe with rpa effects calculated at 3 gev ( solid line ) and compared to miner@xmath0a data @xcite .
dot - dashed lines stand for results without rpa and without 2p2h effects.,title="fig : " ] this work has been produced with the support of the spanish ministerio de economa y competitividad and european feder funds under the contracts fis2011 - 28853-c02 - 01 , fis2011 - 28853-c02 - 02 , fpa2011 - 29823-c02 - 02 , csd2007 - 0042 and sev-2012 - 0234 , the generalitat valenciana under contract prometeo/2009/0090 , and the u.s .
national science foundation under grant nos . 0970111 and 1306944 .
99 nieves j , amaro j e and valverde m ( 2004 ) _ phys .
_ c * 70 * 055503 nieves j , ruiz simo i and vicente vacas m j ( 2011 ) _ phys .
_ c * 83 * 045501 aguilar - arevalo a , _ et al . _
( 2010 ) _ phys .
_ d * 81 * 092005 morfin j g , nieves j and sobczyk j t ( 2012 ) _ adv .
high energy phys .
_ 2012 , 934597 nieves j , snchez f , ruiz simo i and vicente vacas m j ( 2012 ) _ phys .
_ d * 85 * 113008 nieves j , ruiz simo i and vicente vacas m j ( 2012 ) _ phys .
_ b * 707 * 72 martini m , ericson m , and chanfray g ( 2011 ) _ phys .
_ c * 84 * 055502 | we discuss some nuclear effects , rpa correlations and 2p2h ( multinucleon ) mechanisms , on charged - current neutrino - nucleus reactions that do not produce a pion in the final state .
we study a wide range of neutrino energies , from few hundreds of mev up to 10 gev .
we also examine the influence of 2p2h mechanisms on the neutrino energy reconstruction . |
the deviation from scale invariance of the primordial scalar power spectrum is a critical prediction of inflation , and unlike other potential signatures such tensor modes or non - gaussianity , it is the only signature that is _ generic to all inflationary models_. it is therefore a vital test of the inflationary paradigm , and we address it with a minimally parametric approach .
briefly , the idea is as follows . choose a functional form which allows a great deal of freedom in the form of the deviation from scale invariance ( _ e.g. _ smoothing splines ) . naively fitting this to the data will lead one to fit the fluctuations due to cosmic variance and experimental noise , with arbitrary improvement in the chi - square . instead , one performs cross - validation : throw out some of the data ( the validation set ) , fit the rest ( the training set ) , and see how well it predicts the validation set .
a very good fit to the training set , which poorly predicts the validation set , indicates over - fitting of noisy data .
the final ingredient in the algorithm is a roughness penalty , a parameter that penalizes a high degree of structure in the functional form . by performing cross - validation as a function of this penalty
, one can judge when the amount of freedom in the smoothing spline is what the data require without fitting the noise . a minimally parametric power spectrum reconstruction combined with a roughness penalty set by cross - validation
thus provides a method of determining smooth departures from scale invariance which avoids two pitfalls .
firstly , a strong theory prior on the form of the power spectrum ( _ e.g. _ the commonly used power law prescription ) can lead to artificially tight constraints on even a spurious detection of a deviation from scale invariance , which is mostly due to the strength of the prior than that of the data . secondly
, simple binning techniques @xcite or direct inversion @xcite of the data to obtain the primordial power spectrum can lead one to fit noisy data with a large improvement in chi - square / d.o.f .
( easily plausible with current data ) , it would not be surprising to see an improvement of @xmath0 relative to a smooth power spectrum by `` fitting the noise '' with a power spectrum containing a high degree of structure . ] . a minimally parametric approach combined with cross - validation
avoids these issues , providing a way to actually determine the strength of the shape prior justified by the quality of the data .
cross - validation would also be helpful for alternative minimally - parametric methods @xcite _ e.g. _ in choosing the number of basis functions . in this work ,
we use the best available data over a wide range of scales corresponding to the longest `` lever arm '' of wavenumbers currently extant , to reconstruct the shape of the primordial power spectrum in a minimally - parametric way .
esa s planck satellite , which has already begun taking data , is expected to provide superior constraints @xcite on the shape of the primordial scalar power spectrum by 2012 .
our goal here is to establish a benchmark of what was known about the shape of the power spectrum before the planck analysis .
we perform a minimally - parametric reconstruction of the primordial power spectrum based on the method of ref . @xcite .
since the simplest inflationary models , which are consistent with the data , predict the primordial power spectrum to be a smooth function , we search for smooth deviations from scale invariance with a cubic smoothing spline technique ( for details , see refs .
@xcite which we only briefly summarize here ) . in this approach ,
one aims to recover a function @xmath1 from measurements @xmath2 at @xmath3 discrete points @xmath4 . consider a description of @xmath5 by a piecewise cubic spline @xmath6 .
it is uniquely defined by the values of @xmath7 at @xmath8 `` knots '' once we ask for continuity of @xmath6 and its first and second derivatives at the knots , and two boundary conditions : we require the second derivative to vanish at the exterior knots . in our application
, @xmath6 is the primordial power spectrum @xmath9 , and the data are : the angular power spectrum of the 5 year wilkinson microwave anisotropy probe ( wmap5 ) cosmic microwave background ( cmb ) temperature and polarization @xcite ; alone or in combination with higher resolution , ground - based cmb experiments ( quad @xcite and acbar @xcite ) ; or with large scale structure data : the sloan digital sky survey ( sdss ) data release 7 ( dr7 ) luminous red galaxy ( lrg ) power spectrum @xcite ; and the lyman - alpha forest ( ly@xmath10 ) power spectrum constraints from ref .
this work thus represents a significant advance over previous work @xcite , with a new wmap release ( two further years of data and a significant advance in the understanding of systematic errors ) plus substantial improvements in both ground - based cmb data and large - scale structure data .
we use 5 to 7 knots depending on the data set considered ( see table [ tab : specs ] for details ; the locations of the knots in @xmath11 space are indicated in figs .
[ fig : cvcmb ] and [ fig : cvlss ] ) . if the knot values were allowed infinite freedom and were set simply by minimizing the chi - square , in general the reconstruction would fit features created by the random noise present in the data .
it is therefore necessary to add a roughness penalty which we chose to be the integral of the second derivative of the spline function .
the roughness penalty is weighted by a smoothing parameter : by increasing the smoothing parameter the roughness penalty effectively reduces the degrees of freedom , disfavouring jagged functions that `` fit the noise '' .
.[tab : specs ] the cross - validation set - up and the adopted number of knots for each data set used in the analysis . [ cols="^,^,^,^",options="header " , ] in generic applications of smoothing splines , cross - validation is a rigorous statistical technique for choosing the optimal smoothing parameter .
cross - validation ( cv ) quantifies the notion that if the underlying function has been correctly recovered , it should accurately predict new , independent data . to make the problem computationally manageable
, we opt for a @xmath12fold cross - validation , where @xmath3 is the number of data points .
that is , the data set is split into two halves , say , @xmath13 and @xmath14 .
a markov chain monte carlo ( mcmc ) parameter estimation analysis ( for a given smoothing parameter ) is carried out on one half of the data , finding the best fit model .
then the log likelihood of the second half of the data given the best fit model for the first half , cv@xmath15 , is computed and stored .
this is repeated by switching the roles of the two halves , obtaining cv@xmath16 .
the sum , cv@xmath15+cv@xmath16 , gives the cv score " for that smoothing parameter .
finally , the smoothing parameter that best describes the entire data set is the one that minimizes the cv score .
table [ tab : specs ] gives details of the implementation .
note that , as in ref .
@xcite , the basic cosmological parameters ( @xmath17 , @xmath18 , @xmath19 , @xmath20 ) are varied in the mcmc as well as the values of the smoothing spline at the knots , which describe the primordial power spectrum . the mcmc is implemented with modified versions of the camb @xcite and cosmomc @xcite packages , with very stringent convergence criteria .
now we will describe our treatment of the data .
/mpc)@xmath21 , showing knot placement ( triangles , arbitrary normalization ) and cross - validation set - up .
cv@xmath22 is red and
cv@xmath23 is blue .
red points represent the lrg power spectrum from ref .
the lyman alpha measurement is represented by a filled box encompassing the constraints on the observed flux power spectrum from ref .
the light blue line is a concordance lcdm model . ]
_ cmb data : _ we use the @xmath24 version of the wmap5 likelihood function with standard options , with the temperature data divided into alternate ( roughly equal signal - to - noise ) @xmath25 bins for cv@xmath22 and cv@xmath23 respectively , exactly as in ref .
the polarization data are always used in both cv cases .
for cv@xmath22 we use the acbar bandpowers from ref .
@xcite between @xmath26 .
for cv@xmath23 we use the pipeline 1 quad bandpowers between @xmath27 from ref .
@xcite ( see fig . [
fig : cvcmb ] ) .
_ sdss dr7 lrg power spectrum : _ the lrg data are used in cv@xmath22 with wmap5 data .
the data spans the range of wavenumbers @xmath28 [ @xmath29/mpc ] @xmath30 .
the likelihood function we use is identical to that presented in ref .
@xcite ( see fig . [
fig : cvlss ] ) . _
lyman - alpha constraints : _ the ly@xmath10 data are used in cv@xmath23 with wmap5 data ( see fig .
[ fig : cvlss ] ) .
we use the publicly - released likelihood function by a. slozar @xcite to obtain lyman-@xmath10 forest constraints . for this likelihood to be valid , the model @xmath9 must be well described by a three parameter model of amplitude , spectral slope and running at the lyman - alpha forest scales _
i.e. _ @xmath31 @xmath32mpc@xmath33<3 $ ] . to check that this assumption holds in this @xmath11-range for our more general description of @xmath9 , we extrapolated the @xmath9 from the monte carlo markov chains of ref .
@xcite to the lyman-@xmath10 scales and found that in this @xmath11-range the resulting spline can be well approximated by the prescription of ref .
the residuals are at the percent level , well below the intrinsic lyman-@xmath10 errors . with the more recent data sets we consider here ,
the approximation is expected to be even better .
our main results are presented in fig . [ fig : fourpanel ] for several data sets with increasing range in @xmath11 : wmap5 only , wmap5 in combination with quad and acbar , and wmap5 in combination with sdss lrgs and ly@xmath10 .
we show the reconstructed @xmath34 for ease of comparison with the standard power law results . however , the quantity that was actually reconstructed using cross - validation to find the optimal penalty is the power spectrum . the optimal penalty for wmap5 @xmath35 is higher than what was found for wmap3 by a factor of @xmath36 , and is consistent with the optimal penalty for wmap3 in combination with cmb data at smaller scales @xcite .
the corresponding @xmath37 is shown in the top left panel of fig .
[ fig : fourpanel ] .
the same optimal penalty is found for wmap5 and wmap5@xmath38quad@xmath38acbar ( @xmath39 ) , and the latter reconstruction is shown in the top right panel of fig . [
fig : fourpanel ] . for wmap5@xmath38lrg@xmath38ly@xmath10
, we find that cv becomes less sensitive to the value of the penalty , and the cv score dependence on the penalty flattens out at @xmath40 .
while this may indicate a preference for a less smooth @xmath9 , the data can not distinguish between @xmath41 and a penalty an order of magnitude higher .
the reconstructed @xmath37 are shown in the left and right bottom panels of fig .
[ fig : fourpanel ] for penalties @xmath41 and @xmath42 , respectively .
the dark and light blue regions enclose the best ( ordered by likelihood ) 95% and 68% reconstructions .
the 95% constraints are not significantly broader than the 68% because the reconstructed spectra are simply more wiggly ; they are not allowed by the data to deviate more from the best fit , consistently across scales .
cross - validation is a useful tool to check for indications of unidentified systematic biases in the data .
for example , in ref .
@xcite we found that the 3 year wmap data ( wmap3 ) by itself favored a primordial power spectrum with a downward deviation from a power law at small scales ( see fig . 2 of ref .
however , this feature disappeared when combining wmap3 with other data sets ( see figs . 3 and 5 of ref .
@xcite ) which overlapped wmap3 on the scales corresponding to the feature an inconsistency suggestive of a small residual systematic effect in the high @xmath25 wmap3 data . ref .
@xcite argued ( based on considerations of frequency dependence ) that the unresolved residual point source contribution to be subtracted from the raw @xmath43 should have been smaller by @xmath44% and its uncertainty increased by 60% compared to the wmap3 official values . to judge
if smoothing spline cross - validation could give some insights on possible residual systematic errors , we investigated how the point source subtraction level should have been changed for the aforementioned downturn at small scales to disappear from the reconstructed power spectrum .
we obtained a point source amplitude @xmath45% lower than the wmap estimated value , which is tantalizingly close to the estimate of ref .
@xcite .
we find that wmap5 , cmb experiments at smaller scales , and the lrg power spectrum are all consistent with each other . with the addition of ly@xmath10 data ,
a lower penalty value is allowed .
this could be a tentative indication of possible tension between ly@xmath46 and the other data sets , but not a very significant one : there is a cancellation between the effect of penalty and the effect of the likelihood over a wide range of penalty values as shown in the bottom panels of fig .
[ fig : fourpanel ] ) . in addition , as lrg and ly@xmath10 scales do not overlap , we can not exclude the possibility of a low - significance local feature in the power spectrum . in fig .
[ fig : cmbdr7 ] we show the reconstructed @xmath37 for the cmb and lrg data , with optimal penalty @xmath35 .
the cv setup for wmap5 is the same as before , lrgs are added in cv@xmath22 , and quad+acbar are included together in cv@xmath23 .
we have excluded the ly@xmath10 data as it is the only non - overlapping data set . for comparison
, we also show the 95% and 68% @xmath47 constraints @xcite for wmap5@xmath38lrg data when a power - law spectral index is assumed to describe the shape of the primordial power spectrum .
we see no evidence that any @xmath11-dependence of @xmath47 is necessary to describe the data in the cv reconstruction . while @xmath48 is disfavoured , the significance of the departure from scale invariance is weaker than when the `` inflation - motivated '' power law spectral index prior is adopted .
this minimally - parametric reconstruction highlights how constraints relax when generic forms of @xmath9 are allowed .
while this reconstruction is in agreement with the inflationary prior , it illustrates that better data are needed to justify its adoption observationally .
forthcoming data from planck will significantly reduce the current reliance on priors in our understanding of the shape of the primordial power spectrum .
future large - scale structure data and planck will overlap over a decade in scale , offering extra consistency checks .
lyman alpha data , on the other hand , offer the potential to extend the lever arm by at least another decade .
we hope that the results presented here will form a basis to judge the robustness of our present knowledge when confronted with the precision measurements that are on the horizon . from wmap5 , acbar , quad and sdss dr7 lrg data with optimal penalty determined from cross - validation excluding ly@xmath10 .
the orange - red band shows the 95% and 68% @xmath47 constraints @xcite for wmap5@xmath38lrg data with a power - law prior . ] | we present a minimally - parametric reconstruction of the primordial power spectrum using the most recent cosmic microwave background and large scale structure data sets .
our goal is to constrain the shape of the power spectrum while simultaneously avoiding strong theoretical priors and over - fitting of the data .
we find no evidence for any departure from a power law spectral index .
we also find that an exact scale - invariant power spectrum is disfavored by the data , but this conclusion is weaker than the corresponding result assuming a theoretically - motivated power law spectral index prior .
the reconstruction shows that better data are crucial to justify the adoption of such a strong theoretical prior observationally .
these results can be used to determine the robustness of our present knowledge when compared with forthcoming precision data from planck . |
we thank a. strumia and f. bezrukov for useful discussions .
n. cabibbo , l. maiani , g. parisi and r. petronzio , nucl .
b * 158 * ( 1979 ) 295 .
p. q. hung , phys .
* 42 * ( 1979 ) 873 .
m. lindner , z. phys .
c * 31 * ( 1986 ) 295 .
m. lindner , m. sher and h. w. zaglauer , phys .
lett . b * 228 * ( 1989 ) 139 .
m. sher , phys .
* 179 * ( 1989 ) 273 .
b. schrempp and m. wimmer , prog .
* 37 * ( 1996 ) 1 [ hep - ph/9606386 ] .
j. a. casas , j. r. espinosa and m. quiros , phys .
b * 382 * ( 1996 ) 374 [ hep - ph/9603227 ] .
g. isidori , g. ridolfi and a. strumia , nucl .
b * 609 * ( 2001 ) 387 [ hep - ph/0104016 ] . n. arkani - hamed , s. dubovsky , l. senatore and g. villadoro , jhep * 0803 * ( 2008 ) 075 [ arxiv:0801.2399 [ hep - ph ] ] .
g. degrassi , s. di vita , j. elias - miro , j. r. espinosa , g. f. giudice , g. isidori and a. strumia , jhep * 1208 * , 098 ( 2012 ) [ arxiv:1205.6497 [ hep - ph ] ] .
d. buttazzo , g. degrassi , p. p. giardino , g. f. giudice , f. sala , a. salvio and a. strumia , jhep * 1312 * , 089 ( 2013 ) [ arxiv:1307.3536 ] .
f. bezrukov , m. y. kalmykov , b. a. kniehl and m. shaposhnikov , jhep * 1210 * , 140 ( 2012 ) [ arxiv:1205.2893 [ hep - ph ] ] . | in previous publications we have proposed that inflation can be realized in a second minimum of the standard model higgs potential at energy scales of about @xmath0 gev , if the minimum is not too deep and if a mechanism which allows a transition to the radiation dominated era can be found .
this is provided , _
e.g. _ , by scalar - tensor gravity models or hybrid models . using such ideas
we had predicted the higgs boson mass to be of about @xmath1 gev , which has been confirmed by the lhc , and that a possibly measurable amount of gravity waves should be produced . using more refined recent theoretical calculations of the rge
we show that such scenario has the right scale of inflation only for small higgs mass , lower than about 124 gev , otherwise gravity waves are overproduced .
the precise value is subject to some theoretical error and to experimental errors on the determination of the strong coupling constant .
such an upper bound corresponds also to the recent claimed measurement by bicep2 of the scale of inflation through primordial tensor modes .
finally we show that introducing a moderately large non - minimal coupling for the higgs field the bound can shift to larger values and be reconciled with the lhc measurements of the higgs mass . in
@xcite we have proposed that inflation in the early universe can be realized for a narrow band of values of the top quark and higgs boson masses , for which the standard model ( sm ) higgs potential develops a second local minimum @xcite at energy scales of about @xmath0 gev .
such a scenario is viable if a successful transition to a radiation - dominated era can be obtained , which we have shown to be possible in ref .
@xcite using an explicit model in the framework of a scalar - tensor theory of gravity developed in @xcite and in @xcite using a hybrid inflationary model with an additional weakly coupled scalar , in standard gravity .
such a generic scenario could be be realized only for a higgs mass @xmath2 in the range @xmath3 gev , the error being mainly due to the theoretical uncertainty of the 2-loops renormalization group equations ( rge ) used in that calculation .
such a prediction for the higgs mass range has turned out to be surprisingly compatible with the measurement of a higgs mass of @xmath4 gev by the lhc @xcite .
as it is well known , inflation can generate quantum mechanically tensor ( gravity wave ) modes , usually parameterized through @xmath5 , the ratio of tensor - to - scalar perturbation spectra at large scales . in @xcite we had predicted a relatively large amplitude for @xmath5 in our scenario , which is set by the overall height of the second minimum , which can be computed up to the precision of the rge calculations and of the experimental input for the standard model parameters : @xmath2 , the top mass @xmath6 and the strong coupling constant @xmath7 .
however the situation has recently changed , since more refined theoretical calculations of the standard model potential have been performed , therefore making more precise the correspondence between the input parameters and the scale of the second minimum .
moreover the tensor modes have now been claimed to be measured by the bicep2 collaboration @xcite with a large amplitude , which corresponds to a scale of about @xmath8 gev . in this paper
we update our results , using the results of @xcite for the higgs effective potential , and show that if the theoretical errors are under control the scale of inflation is generically larger than @xmath8 gev and can be reconciled with experiments only if @xmath9 gev ( plus a theoretical error , at present estimated to be of order @xmath10 gev ) . of course such conclusion can be changed if extra ingredients are added to the standard model .
we show the results that can be obtained by introducing a non - minimal coupling of the higgs to gravity .
of course less minimal modifications , such as a direct coupling of the higgs to an extra scalar , can also easily shift the scale of inflation . finally , provided the scale of inflation due to the higgs potential is low enough
, we also reanalyze the specific model based on scalar tensor theories @xcite , which realizes the exit to a radiation dominated era , showing that it could reproduce observations on @xmath5 and on the scalar spectral index @xmath11 . as in @xcite
we consider the potential for the higgs field @xmath12 in the sm of particle physics .
for very large values of the higgs field such potential can be written as @xmath13 where @xmath14 is some effective quartic coupling @xcite which is very close , but not identical , to the quartic coupling @xmath15 .
due to the running of couplings @xmath14 can become very small at high energies , while keeping always positive values ( stability regime ) .
in such conditions the potential can have an additional minimum at very high field values .
while this is at present disfavored by the current calculations and measurements in the sm it still not ruled out . in particular this
can still happen if the top quark mass has a low value compared to the present best fit @xcite .
if the higgs field starts in the false minimum at @xmath16 and dominates the energy density of the universe , the friedmann equation leads to a stage of inflationary expansion h^2 h_i^2 , a(t)e^h_i t [ eq - m ] where @xmath17 is the scale factor , @xmath18 is the hubble rate and @xmath19 is the reduced planck mass ( @xmath20 ) .
a nontrivial model - dependent ingredient is a mechanism to achieve a graceful exit from inflation , that is a transition to a radiation - dominated era , in a nearly flat universe at a sufficiently high - temperature . in
@xcite we have proposed that the higgs field can tunnel to the other side of the potential barrier by nucleating bubbles @xcite that can successfully collide and percolate , in the presence of a scalar tensor theory of gravity . alternatively in
@xcite we had proposed that a smooth transition may happen also in standard gravity , in the presence of an extra scalar field very weakly coupled to the higgs field .
the main point of both mechanism is that they do _ not _ affect appreciably the sm runnings of couplings , since in the first model there is only a gravitational coupling to the standard model , and in the second case the dimensionless coupling constant is extremely small , of order @xmath21 . for this reason precise connections with low energy parameters still hold even in the presence of such new fields .
of course it is possible that such a new scalar can also interact with the higgs field with a large direct coupling and in this case this could also affect the rge equations and we briefly discuss later such possibilities . subsequently the higgs field could roll down the potential , reheat the universe and finally relax in the present true vacuum with @xmath22 gev . in the scalar - tensor theory of gravity model an extra scalar field @xmath23
is introduced , the brans - dicke scalar or dilaton , which has an interaction term of the form @xmath24 , where @xmath25 is the ricci scalar and @xmath26 thus sets the value of the planck mass .
the presence of such field makes the planck mass time - dependent , and therefore also @xmath27 during inflation : this slows down the expansion sufficiently enough if @xmath28 grows faster than @xmath29 for large @xmath23 values , since a stage of quasi exponential expansion is followed by a stage of power - law ( even decelerated ) expansion @xcite . during the exponential phase the quantum fluctuations in @xmath23
lead to a nearly scale - invariant spectrum of perturbations . during the subsequent decelerated phase ,
@xmath27 decreases rapidly , and therefore the expansion is sufficiently slowed down , so that many bubbles can be nucleated in a hubble patch and subsequently collide and reheat the universe , leading to a spatially flat radiation dominated universe . as discussed in @xcite , after tunneling we require the field @xmath23 to relax to zero if a suitable potential @xmath30 is present , which allows us to identify the present reduced planck mass with the quantity @xmath19 and , at the same time , to satisfy constraints from fifth - force experiments and time - dependence of the newton constant @xmath31 @xcite . an alternative to scalar - tensor theories is given by models @xcite with a direct tiny coupling of the higgs field to an additional scalar field , which induces a time - dependence directly into @xmath32 by flattening the barrier in the potential .
it is crucial to notice that a graceful exit can be realized in the above models only if there is a very shallow false minimum , otherwise the tunneling rate would be negligibly small in the scalar - tensor model , since the probability is exponentially sensitive to the barrier @xcite .
such shallowness is required also in the hybrid model since a change in the barrier by a large amount would introduce a departure from an almost flat scalar spectral index .
so , the shape of the potential is very close to the case in which there is just an inflection point and thus we have a generic prediction for the scale of inflation and for @xmath5 , while the specific model only affects the prediction for the spectral index of cosmological density perturbations @xmath33 .
using the recent state - of - the - art theoretical calculations given in @xcite , we show the specific values of the top and higgs masses which allow for the presence of a false minimum .
such calculations are now so accurate that also the precise value of @xmath34 inside the present allowed experimental range @xmath35 @xcite is a relevant parameter . as an example , in fig.[fig - vmin ] we also show the higgs potential for very specific values of @xmath6 and @xmath2 , showing agreement with fig.7 of @xcite .
the extremely precise values shown in the caption are not to be taken sharply , because of a theoretical uncertainty , estimated to be of about @xmath10 gev on @xmath2 and about @xmath36 gev on @xmath6 . in units of @xmath37 gev .
we fixed @xmath38 .
the upper curve correspond to @xmath39 , which shows good agreement with the results of @xcite , fig . 7 .
the lower curve corresponds to @xmath40 , @xmath41 , and a value of @xmath42 . as mentioned in the text and discussed in @xcite , in order to have a sizable tunneling probability through the left side , the barrier must be very low , almost as in an inflection point.,width=321 ] .2 cm increasing ( decreasing ) @xmath6 , one has also to increase ( decrease ) @xmath2 in order to develop the shallow false minimum ; accordingly , the value of both @xmath43 and @xmath16 increase ( decrease ) .
the solid blue lines in fig.[fig - mtmh ] show the points in the @xmath44 plane where the shallow sm false minima exists , for different values of @xmath45 in the 3-sigma range .
the red dashed lines show the regions in ( @xmath46 ) which corresponds to a value of @xmath5 compatible with observations , directly given by the height of the potential via the equation @xmath47 and it turns out that these lines point to a too small value of @xmath2 compared to the present experimental values .
of course this is a valid conclusion under the assumption that the theoretical errors on @xmath2 are now estimated to be of about @xmath10 gev by @xcite .
this scenario could still be compatible with data only if the theoretical errors in the rge and the matching conditions were larger , say of about @xmath48 gev .
discarding such possibility we show that this could be cured by introducing a non - minimal coupling @xmath49 also for the higgs field .
this transforms the potential in the slow roll phase in the einstein frame in the following form ( see @xcite and citations therein ) : @xmath50 which suppresses the potential at large field values and we show in fig .
[ plotxi ] that a moderately large value of @xmath51 allows for the potential to fit cmb observations , also for larger @xmath2 , compatible with the lhc measurements .
however let us comment that this is based only on a tree level analysis and when having a value of the extra coupling constant @xmath51 larger than @xmath52 we have to be careful about the regime of validity of the theory , similarly to the models proposed in @xcite .
.22 cm .22 cm finally we also show the results for the @xmath53 plot for the scalar tensor model proposed in @xcite in figure [ nsr ] .
such a model is based on the introduction of an additional non - minimally coupled field @xmath23 with the following action : @xmath54 \label{azione}\ ] ] where @xmath55 is the standard model lagrangian and the potential @xmath56 is not specified and assumed to be relevant only after inflation to stabilize the field .
the coupling function is given by : @xmath57 with the requirement that @xmath58 , which is guaranteed for instance if the couplings @xmath59 are positive .
under these conditions @xmath60 starts almost constant and after a sufficient umber of efolds starts decreasing as a power law , allowing for a tunneling transition when @xmath61 becomes equal to the tunneling rate per unit volume @xmath32 . as a simple example we will study the case in which @xmath62 and all other couplings are vanishing .
if the higgs fields is at the minimum and it is not evolving during inflation the only time dependent quantity is @xmath23 and so we can go the einstein frame with a metric defined by @xmath63 and study the evolution of a canonically normalized field @xmath64 defined through @xmath65 , where @xmath66 in this frame the action becomes : s_e= 12d^4x [ m^2 |r-(|)^2- 2 |l_sm ] [ e - action-2 ] . where the bar represents quantities in the einstein frame .
the higgs potential contained in @xmath67 is now @xmath68 , so that the potential energy at the false higgs minimum @xmath16 gives rise to a potential term for @xmath64 s_e^_0=d^4x |v , |v .
[ eq - vb0 ] this acts as a hill - top potential for the @xmath64 field and we assume that @xmath64 rolls down the potential from small to high values . given the potential we define as usual the slow - roll parameters ( n ) = | @xmath64 under such a potential assuming that it starts as close to zero as possible ( _ i.e. _ with an initial value @xmath69 given by the quantum fluctuations @xmath70 ) together with the friedmann equation , which gives us a numerical solution for @xmath71 , where n is the number of efolds at some value of the scale factor @xmath72 , defined as @xmath73 , where @xmath74 is the scale factor at the end of the inflationary stage , defined by either @xmath75 or @xmath76 becoming of order 1 .
finally we evaluate the quantites @xmath77 where @xmath78 is the number of efolds that corresponds to a certain cosmological scale .
we take @xmath79 although the precise number depends on the history of the evolution subsequent to inflation .
the results are shown in fig.[nsr ] .
for the slow - roll parameters , which does not hold for very small @xmath80 , therefore leading to incorrect results for @xmath81 , which corresponds to a sizable @xmath5 in fig.[nsr ] . ] in conclusion we have shown that a mechanism in which the energy density for inflation is provided by the higgs field starting in a shallow false minimum can be compatible with observations only if @xmath9 gev ( and also @xmath6 has to be smaller than about @xmath82 gev ) .
this is significantly lower than experimental observations unless the theoretical error on the rge , of about @xmath10 gev on @xmath2 according to @xcite , is largely underestimated .
we have also shown that introducing a non - minimal coupling of the higgs of about @xmath83 can push this upper bound on larger values , closer to present observations .
of course such bounds can be also evaded by modifying the shape of the potential in a model - dependent way .
for instance a possibility would be to introduce direct couplings of the higgs with the extra scalar as in @xcite , which would change the potential either at tree level or in the rge , or perhaps to allow a faster change of the barrier in hybrid models during inflation . and tensor - to - scalar ratio @xmath5 for the scalar - tensor model of @xcite .
here the number of efolds is assumed to be @xmath79 .
only one nonvanishing parameter @xmath80 was assumed to be present in eq .
[ eq - ftot ] and the various points along the black lines correspond to different values of @xmath80 .
, width=340 ] .2 cm |
to date , the theory of anderson localization in one - dimensional disordered systems is developed in great detail . in particular , various analytical approaches to the models with continuous potentials
allow to derive all transport characteristics in dependence on the disorder strength and size of the samples ( see , for example , refs .
@xcite ) . on the other hand , for the tight - binding and kronig - penney models
the rigorous analysis is a difficult task due to the presence of resonances of the fabry - perot type .
the famous example is the standard tight - binding anderson model for which at the band center the correct expressions for the localization length @xmath0 and transmission coefficient @xmath1 can be obtained with specific methods only @xcite .
as was found , the band center corresponds to the lowest ( most important ) resonance resulting in a non - flat distribution @xmath2 of the phase @xmath3 of wave function , emerging when the wave propagates along a disordered sample .
the same situation occurs for kronig - penney models with weak disorder for which the fabry - perot resonances have to be taken into account , if the aim is to develop general expressions valid for any value of energy inside allowed energy bands .
although away from these resonances the analytical results for @xmath0 and @xmath1 can be obtained relatively easy , in the vicinity of the resonances the transport properties are mainly understood with the use of numerical simulations .
one of the open problems , in connection with the presence of these resonances , is how to relate global transmission characteristics to the localization length @xmath0 which near the resonances can be obtained with one of specific methods .
in contrast with continuous scattering potentials for which the so - called single parameter scaling ( sps ) holds , the question about the validity of the sps for tight - binding and kronig - penney models remains open .
in the theory of scattering for continuous one - dimensional models the sps is trivially valid since the distribution of @xmath1 depends on one single parameter only , which is the ratio between the localization length @xmath0 and size of the sample @xmath4 .
this means that the knowledge of the localization length ( defined in the limit @xmath5 ) gives a complete solution of the scattering problem .
as for the tight - binding and kronig - penney models , the relation between the localization length and transport properties in the regions close to resonances is typically unknown , the fact that makes the sps hypothesis questionable . therefore , for such systems the problem of the relevance of @xmath0 to transport properties is of great importance .
recently , the detailed study of the transmittance and reflectance in the vicinity of the lowest resonance has been performed in ref .
the authors where able to develop the theory and obtain the analytical results for a quite simple model for which the potential consists of barriers and wells of a fixed thickness @xmath6 , however , with a weak variation of their heights and depths .
the lowest resonance emerges when the phase shift @xmath7 of the wave equals @xmath8 after passing freely a barrier or well .
numerical simulation shows that for such a value of @xmath7 a clear dip occurs for the transmission coefficient .
it was shown how to describe the landauer resistance and transmittance in the ballistic regime , with the use of special technics based on the so - called building block " method . in our paper
we analyze the same model , however , paying main attention to the relevance of the localization length to transport characteristics , namely , to the transmission coefficient @xmath1 , its logarithm @xmath9 , and their variances . to do this
, we develop a new method consisting of few steps .
first , we show how to find an analytical expression for the phase distribution @xmath2 which is highly non - uniform near the resonances .
then , with this distribution we demonstrate how the localization length can be analytically obtained in the vicinity of any resonance @xmath10 with @xmath11 integer .
finally , we show that if to insert this localization length into the expression for the moments of @xmath1 obtained for the models with random continuous potentials , one gets a nice correspondence with numerical data in a whole energy region including the resonances .
the comparison with numerical data is excellent outside the ballistic regime , i.e. in the region of a strong and intermediate localization , when the obtained localization length @xmath0 is smaller than or of the order of the system size @xmath4 .
we also suggest how to improve the correspondence for the ballistic regime , indicating that the localization length found in a strong limit @xmath5 may have no physical sense since for a not large enough value of @xmath4 the phase distribution @xmath2 is still non - stationary . to overcame this problem we suggested to use the finite - length lyapunov exponent which can be computed numerically .
then , the data show much better agreement with this semi - analytical approach .
we study the localization and transport properties of a quantum particle with the mass @xmath12 propagating through an array of rectangular potential barriers and/or wells .
the height @xmath13 of the @xmath14-th scatterer randomly depends on index @xmath14 , however , all the barriers / wells are of the same thickness @xmath6 ( see fig . [ pre - fig01 ] ) .
our study is restricted to the case of the over - barrier scattering when the particle energy @xmath15 is much larger than the strength of the random potential , @xmath16 .
the stationary schrdinger equation for the wave function @xmath17 of the particle traveling over the @xmath14-th barrier / well , reads @xmath18 where @xmath19 is the wave number of the particle .
its general solution can be presented as a superposition of two standing waves , @xmath20\\ + k_{n}^{-1}\psi'_{n}(x_{n})\sin\left[k_{n}(x - x_{n})\right],\nonumber\end{aligned}\ ] ] for @xmath21 .
the @xmath22-axis is directed along the array with @xmath23 standing for the coordinate of the left - hand edge of the @xmath14-th barrier , see fig . [ pre - fig01 ] .
the prime implies the derivative with respect to @xmath22 .
note that the constant thickness of the unit @xmath14-th barrier is defined as @xmath24 ( color online ) set of random barriers and wells.,width=302,height=188 ] the general solution has to be complemented by two continuity conditions at the interfaces between neighboring barriers / wells , @xmath25 the combination of eqs . and
yields the recurrent relations describing the wave - function transfer through the @xmath14-th unit cell of the array , @xmath26 here @xmath27 and @xmath28 refer to the wave - function and its rescaled derivative , respectively , taken at left - hand edge of the @xmath14-th unit cell , @xmath29 the phase shift @xmath30 randomized by the compositional disorder , and its unperturbed counterpart @xmath7 are defined by @xmath31 here , for convenience , we have introduced the dimensionless random strength @xmath32 of the potential barriers / wells and the unperturbed particle wave number @xmath33 , @xmath34 remarkably , the recurrent relations can be regarded as the classical _ hamiltonian map _ describing the evolution of trajectories in the phase space @xmath35 with discrete time @xmath14 for a linear oscillator subjected to the time - dependent parametric force . in such a representation @xmath27 and @xmath28 can be treated as the classical coordinate and momentum , respectively @xcite .
thus , the problem of quantum localization can be formally reduced to the analysis of the energetic instability of a stochastic oscillator @xcite .
note that the hamiltonian map belongs to the class of area - preserving maps whose determinant equals unity . for the analytical study it is convenient to pass to polar coordinates , namely , to the radius @xmath36 and angle @xmath37 , @xmath38 according to eq .
, the hamiltonian map in the _ radius - angle presentation _ gets the form , @xmath39 \tan\theta_{n+1}&=&\frac{-(\varphi_n/\varphi)\sin\varphi_n+\cos\varphi_n\tan\theta_n } { \cos\varphi_n+(\varphi/\varphi_n)\sin\varphi_n\tan\theta_n}.\label{eq : map - theta}\end{aligned}\ ] ] as one can see , the _ linear _ two - dimensional map for @xmath27 and @xmath28 can be reduced to the _ non - linear _ one - dimensional map for the angle @xmath37 only .
it should be stressed that in this map the angle @xmath37 can be considered in the range @xmath40 $ ] .
note also that both maps are the time dependent ones , this fact makes the rigorous analysis problematic . in line with the concept of the hamiltonian map ,
the localization length @xmath0 is determined by the rate of exponential growth of the coordinate @xmath27 or momentum @xmath28 , once the initial conditions ( @xmath41 ) are specified . the conventional definition of the localization length @xmath0 is due to the inverse of the lyapunov exponent @xmath42 , and the latter can be defined as @xmath43 another definition which gives the same result , takes the form @xcite , @xmath44 here the averaging @xmath45 is performed along the discrete time " @xmath14 . the second relation in
can be derived directly from eqs .
it is useful for the analytical analysis since due to ergodicity the averaging over @xmath14 can be substituted by the averaging over different realizations of the disorder @xmath32 and random phase @xmath37 . in what follows
, the dimensionless variable @xmath32 imposing the disorder , is specified by a random sequence of white - noise type with the zero average and variance @xmath46 , @xmath47 in numerical analysis , when generating random sequence @xmath32 we use the entries of the uniform box probability distribution inside a finite interval @xmath48 $ ] with the variance @xmath49
. however , our analytical results are valid for any distribution of @xmath32 with correlation properties and finite small variance . in terms of the statistical characteristics for the random quantity @xmath32 the conditions of weak disorder ( @xmath50 , @xmath51 ) and the over - barrier scattering ( @xmath52 ) can be rewritten in the explicit form , @xmath53 these conditions allow us to develop a proper perturbation theory .
it is worthwhile to mention that the system under consideration is similar to an array of optic slabs with random and frequency dispersive refractive indices .
the features of optic wave localization in the non - dispersive array , where the refractive index is independent of the wave frequency , were analyzed in detail in refs .
as one can recognize below , the effect of energy / frequency dispersion drastically changes the localization properties of both quantum and optic disordered systems .
for a weak disorder , see , we expand the @xmath54-map up to the second order in perturbation @xmath32 . then , we substitute the result into eq . with the subsequent expansion of the logarithm , keeping the terms quadratic in disorder .
taking into account that we consider the case of a white - noise disorder , one can neglect high - order correlations between the disorder @xmath32 and phase @xmath37 @xcite .
this allows us to perform the statistical averaging over @xmath32 in accordance with the correlation properties . as a result
, we arrive at the following quadratic approximation for the lyapunov exponent , @xmath55.\nonumber\end{aligned}\ ] ] here we substituted the averaging of @xmath37 over @xmath14 by the statistical average over @xmath3 ( denoted by the bar ) assuming that the distribution @xmath2 exists .
the starting point for obtaining @xmath2 is the quadratic expansion of the @xmath3-map , @xmath56\qquad\\ + \frac{v_n^2}{8\varphi^4}\big[\varphi+\sin \varphi\cos(2\theta_n-\varphi)-2\varphi\cos(2\theta_n-2\varphi)\nonumber\\ -\sin^2\varphi\sin(4\theta_n-2\varphi)-2\sin\theta_n\sin\varphi\sin(\theta_n-\varphi)\big].\nonumber\end{aligned}\ ] ] by analyzing eq .
one can suggest that for non - zero values of @xmath7 only a small number of iterations are needed for @xmath37 to fill the whole interval [ @xmath57 .
therefore , in this case a uniform phase distribution can be expected in the lowest order of perturbation , @xmath58 the averaging of eq .
with the probability density is trivial and gives rise to the expression @xmath59 this result is in a complete correspondence with those previously obtained for discrete optic systems with randomized refractive index ( see , e.g. , refs .
. however , the flat distribution may not be valid for the resonant values of @xmath7 , namely , for @xmath60 with @xmath61 integers .
for such rational values ( with respect to @xmath62 ) the unperturbed trajectory @xmath37 is the periodic orbit with the period @xmath63 , therefore , @xmath64 is the periodic delta - function of the same period . by adding a weak disorder the phases @xmath37 begin to diffuse around each of the delta - peaks , and it is not clear whether the fingerprint of these periodic orbits disappears in the form of @xmath2 in the limit @xmath65 .
one can expect that the strongest resonances correspond to @xmath66 and @xmath67 .
below we restrict our study by these resonances only . as for high - order resonances with @xmath68 , it is quite naturally to expect that they give much less influence to the localization length , if any . indeed , as is displayed in fig .
[ pre - fig02 ] , the lyapunov exponent obtained from the numerical iteration of the hamiltonian map differs from eq . only very close to the points @xmath69 and @xmath70 ( @xmath11 is an integer ) , i.e. , when @xmath66 and @xmath67 .
note that for @xmath69 the unperturbed @xmath3-map is given by a single point , and for @xmath70 it results in two fixed points .
therefore , in the analytical approach one has to treat these two cases separately .
( color online ) lyapunov exponent @xmath42 vs phase shift @xmath7 : continuous curve depicts the analytical result , dashed curve shows the numerical computation of @xmath42 obtained with the use of eq . .
the intensity of disorder is @xmath71.,width=302,height=188 ] from the physical point of view , the origin of the above peculiarities is the resonance effect emerging due to the coherent interaction of the waves reflected from the boundaries of wells / barriers .
this effect may be compared with the well known fabry - perot resonances emerging due to multiple reflections of the wave from the interfaces in multi - layered photonic structures .
as was shown both theoretically and in experiments @xcite , in the non - dispersive systems these resonances strongly suppress the localization .
thus , the fabry - perot resonances are typically associated with the resonance _ enhancement _ of the transmission . in our case , as is seen in fig .
[ pre - fig02 ] , the resonances result in the _ suppression _ of the transmission for both odd and even values of @xmath63 .
such a peculiarity of the lowest resonance @xmath72 occurring in the model has been predicted in ref .
@xcite . below the analysis of the resonance effects will be performed in a way similar to that used in the study of peculiarities of the localization arising in the one - dimensional anderson model near the band center and band adges @xcite .
here we consider the energy region around the even resonances , @xmath73 were the unperturbed @xmath3-map has almost - periodic orbits of the period one .
therefore , the point @xmath74 is very close to @xmath37 provided a weak disorder is imposed .
the disordered @xmath3-map near the even resonances can be obtained from eq . by the corresponding first - order expansion with respect to a small resonance detuning @xmath75 . after omitting the term that is integer multiple of @xmath62 , the recurrence relation for the polar angle
@xmath3 reads @xmath76 here we have neglected the terms containing @xmath77 since they do not contribute to the associated fokker - plank equation derived within the linear approximation in the detuning @xmath75 and quadratic one in the disorder @xmath32 . in order to obtain the phase distribution
, one has to derive the stationary fokker - plank equation for @xmath2 .
this can be done in the same way as described , e.g. , in refs .
@xcite . specifically , we rewrite the map in the continuum limit and after replace the random variable @xmath78 with the wiener process @xmath79 in accordance with the definition , @xmath80 . as a result ,
we come to the so - called it stochastic differential equation , @xmath81dt.\ ] ] with this equation one can study the dynamics of the stochastic process @xmath82 once the initial condition @xmath83 is known . on the other hand ,
following the theory of stochastic differential equations @xcite , one can readily associate the it equation with the stationary fokker - plank equation for the probability density @xmath84 , @xmath85\rho_s=0.\ ] ] this equation should be complemented by the condition of periodicity and by the normalization condition , @xmath86 ( color online ) distribution @xmath84 for the even resonance ( @xmath87 ) and various values of @xmath88 : continuous curves show the analytical equation , while different symbols correspond to numerical data .
the intensity of disorder is @xmath71.,width=302,height=188 ] after the integration of eq .
we obtain the following linear first - order equation , @xmath89\rho_s(\theta)=c,\ ] ] with the constant @xmath90 which can be found from the periodicity condition . remarkably , in eq .
the resonance detuning turns out to be modified as follows , @xmath91 this means that the even resonances are shifted by disorder to the right , @xmath92 also , eq . manifests an emergence of the disorder - induced scale @xmath93 for the resonance detuning .
( color online ) rescaled lyapunov exponent vs phase shift @xmath7 in the vicinity of the even resonance with @xmath87 and @xmath71 .
continuous curve corresponds to the analytical result , circles represent the numerical data.,width=302,height=188 ] solving equation with the periodicity condition yields the nonuniform phase distribution in the vicinity of the even @xmath11-resonance , @xmath94,\qquad\\ \mu(\theta)&=&\left[b_j\theta + \frac{1}{2j\pi}\sin 2\theta\right].\nonumber\end{aligned}\ ] ] here the initial value @xmath95 is specified by the normalization condition from eqs . .
it is important to emphasize that the initial eq . for the lyapunov exponent
contains only zero , second and fourth harmonics with respect to the @xmath3-phase .
hence , in order to perform the averaging procedure only the corresponding zero , second and fourth harmonics of the probability density @xmath84 are needed .
all the others give zero result after the averaging . the simplest way to extract the important harmonics from eq .
is to replace the smooth factors in @xmath96 $ ] and @xmath97 $ ] with their approximate expansions , e.g. , @xmath98 then , the integrals in eq . can be taken explicitly that yields the truncated distribution function @xmath84 as a superposition of the uniform one with oscillating modulations , @xmath99 one can see that at the resonance @xmath100 ( @xmath101 ) the distribution profile has strong oscillations . however , when the phase shift @xmath7 moves away from the resonance ( @xmath102 ) , the oscillations are decreasing , with a smooth ( power - law ) convergence to the flat distribution .
the similar effect takes place as the order of the resonance @xmath11 increases : the higher the order @xmath11 the closer the @xmath3-distribution to the flat one .
therefore , the resonances with large @xmath11 are hardly observable . fig .
[ pre - fig03 ] displays the change of the @xmath3-distribution with the variation of the modified detuning parameter @xmath88 in the vicinity of the even resonance @xmath103 .
the data shown in fig . [ pre - fig03 ] demonstrate an excellent agreement between the analytical result and the corresponding numerical simulation .
now we are able to perform the averaging in eq . with the use of the probability density .
within the lowest approximation in disorder @xmath46 and in resonance detuning @xmath75 , the non - zero contribution comes from the first and second terms in both eqs . and .
as a result we get @xmath104 fig .
[ pre - fig04 ] compares this expression with the numerical data for the lyapunov exponent in the vicinity of the first ( @xmath87 ) even resonance .
the energy region of the odd resonances is defined by condition @xmath105 at odd resonances the unperturbed @xmath3-map for a fixed initial value @xmath106 has periodic orbits with period 2 and is presented by two points .
therefore , the phase @xmath107 coincides with @xmath37 . for a weak disorder , due to a small resonance detuning @xmath75 these points are no more fixed
, however , after two steps the distance between them is quite small .
for this reason it is convenient to treat the two - step recurrent @xmath3-relation between @xmath107 and @xmath37 .
this relation is readily obtained by the iteration of the general map . within the lowest order in the resonance detuning @xmath75 and with the use of statistical independence of the variables @xmath32 and @xmath108 ,
, one gets @xmath109 being written in the continuum limit in the terms of two independent wiener processes , @xmath110 and @xmath111 , the @xmath3-map takes the form of the it equation , @xmath112dt.\end{aligned}\ ] ] it is important that the random processes @xmath110 and @xmath111 have the same statistical properties . therefore , as in the previous section ,
we can apply the method described in ref .
@xcite and write down the corresponding fokker - plank equation for the stationary distribution function @xmath84 , @xmath113\rho_s=0.\ ] ] this equation is complemented by the conditions of periodicity and normalization .
( color online ) distribution @xmath84 for the first odd resonance ( @xmath87 ) and various @xmath88 : continuous curves show the analytical equation while different symbols stand for the numerical data .
the intensity of disorder is @xmath71.,width=302,height=188 ] it is clear that the solution of eqs . ,
has the form , however , with a new function @xmath114 , @xmath115,\ ] ] and with another resonance detuning @xmath116 , @xmath117\frac{8(2j-1)^2\pi^2}{\sigma^2}\\ & = & \left[\varphi-(2j-1)\pi-\frac{\sigma^2}{8(2j-1)^3\pi^3}\right]\frac{8(2j-1)^2\pi^2}{\sigma^2}.\nonumber\end{aligned}\ ] ] thus , the odd resonances are shifted by disorder exactly as even resonances , @xmath118 and have similar disorder - induced broadening of the order of @xmath119 .
the truncated distribution function @xmath84 containing only zero , second and fourth harmonics that contribute to the averaging of eq .
, can be extracted from the equations , , in the same manner as for even resonances . as a result
, we have @xmath120 one can see that this probability density has the structure similar to the truncated distribution obtained for even resonances .
specifically , at the resonance @xmath100 ( @xmath101 ) the distribution oscillates .
these oscillations decrease and smoothly disappear when the phase shift @xmath7 moves away from the resonance ( @xmath102 ) .
as the order of the odd resonance @xmath11 increases the oscillations begin to be smoother .
[ pre - fig05 ] shows the change of the @xmath3-distribution with the variation of the modified detuning @xmath88 in the vicinity of the first ( @xmath87 ) odd resonance @xmath121 .
results confirm the validity of the theoretically obtained equation in comparison with the corresponding numerical data .
after the averaging of eq . with the distribution function
we obtain the lyapunov exponent for odd resonances , @xmath122 as one can see , it is actually of the same form as eq .
derived for even resonances .
the only difference is in the definition of the modified resonance detuning : for odd resonances @xmath116 contains the quantity @xmath123 instead of @xmath124 for even resonances , compare eqs . and .
let us now compare expressions , and for the lyapunov exponent that are valid away from resonances and in the vicinity of odd / even resonances , respectively . from this comparison
one can easily conclude that eq . or , the same , eq . can be regarded as the general interpolation for the lyapunov exponent if to write the parameter @xmath116 in the generalized form . in order to realize this idea , in eqs . or we replace the quantities @xmath125 and @xmath126 with the phase shift @xmath7 , and generalize the definition of the resonance detuning @xmath75
. the explicit result reads @xmath127;\nonumber\\[6pt ] & & \epsilon=\left\{\begin{array}{ccc } \varphi-\left[\frac{\varphi}{\pi}\right]\pi&\mbox{for}\quad0\leqslant\frac{\varphi}{\pi}-\left[\frac{\varphi}{\pi}\right]\leqslant\frac{1}{2},\\[6pt ] \varphi-(\left[\frac{\varphi}{\pi}\right]+1)\pi&\mbox{for}\quad\frac{1}{2}<\frac{\varphi}{\pi}-\left[\frac{\varphi}{\pi}\right]<1,\end{array}\right.\end{aligned}\ ] ] where @xmath128 $ ] stands for the integer part . within the quadratic approximation in disorder , equation
adequately describes the lyapunov exponent @xmath42 inside a wide range of the phase shift @xmath129 .
the applicability of eq . is restricted only by the conditions of weak disorder and over - barrier scattering .
( color online ) resonant line - shape of the lyapunov exponent for first odd and even resonances , panels ( a ) and ( b ) , respectively .
continuous curve depicts analytical interpolation whereas the circles show corresponding numerical data .
the intensity of disorder is @xmath71.,title="fig:",width=302,height=188 ] ( color online ) resonant line - shape of the lyapunov exponent for first odd and even resonances , panels ( a ) and ( b ) , respectively .
continuous curve depicts analytical interpolation whereas the circles show corresponding numerical data .
the intensity of disorder is @xmath71.,title="fig:",width=302,height=188 ] out of the resonances , the detuning is of the order of unity , @xmath130 .
the second term in square brackets is negligibly small being of the order of @xmath131 . due to this fact , eq .
is equivalent to eq . in the region between the neighboring resonances .
the second ( resonant ) term in the square brackets has the lorentzian form . for both odd and even resonances we have , @xmath132 ,
therefore , this term is equals 1 and strongly prevails over the first term which is of the order of @xmath133 .
the half - width of the resonances is @xmath134 , thus , showing that the resonant line - shape is very sharp .
it is worthwhile to note that the higher the resonance order @xmath11 the sharper the resonance , however , the smaller its amplitude @xmath135 .
when the phase shift @xmath7 moves away from the resonance the main contribution turns in eq . from the second term to first one .
[ pre - fig06 ] shows that the interpolation provides a good agreement with the numerical data , apart from the transition regions where the lyapunov exponent is so small that the perturbation terms of higher order have to be taken into account .
now , in connection with the localization length we address the problem of transport properties in finite systems constructed by an array of @xmath136 unit cells ( random barriers and/or wells ) . in terms of trajectories of the hamiltonian map , the transmittance @xmath137 of a finite system of length @xmath138
can be expressed through the following relation @xcite , @xmath139 here @xmath140 and @xmath141 are the radii of two independent trajectories at the time @xmath142 which start , respectively , from the points @xmath143 and @xmath144 . in the present context
, one can use the famous definition of the inverse localization length @xmath145 via the transmittance , @xmath146 as is known from the theory of disordered 1d systems , this definition of the lyapunov exponent is equivalent to that considered above , see eq . .
in our numerical calculations we perform the averaging @xmath147 over @xmath148 different realizations of the disorder @xmath32 which allows one to reduce the fluctuations . with the knowledge of the localization length
, the mean value of @xmath149 can be obtained due to the famous relation , @xmath150 which can be rigorously derived for 1d weakly disordered models with continuous potentials , see for example , @xcite .
it should be stressed that for such models this relation is valid for any ratio between the localization length @xmath0 and the sample size @xmath4 , therefore , both in the ballistic regime ( for @xmath151 ) and in the localized regime ( for @xmath152 ) . in the case of finite one - dimensional continuous systems with weak random potential ,
the scattering problem was rigorously solved by various analytical methods .
our interest below is in the validity of the following rigorous expression for the moments of the transmittance @xmath137 ( see , e.g. , refs .
@xcite ) that can be readily adapted for a discrete model , @xmath153 here @xmath42 is the lyapunov exponent , or the same , dimensionless inverse localization length @xmath154 ( see details , e.g. , in ref .
@xcite ) .
we would like to note that , strictly speaking , expression can not be derived for the discrete models like our model with the wells and barriers .
the reason is that for continuous potentials for which the expression has been derived , the resonances similar to those we are discussing here , are absent .
it is known that for the standard tight - binding anderson model the existence of the resonances do not allow to develop general analytical approach valid for any value of the energy of incident waves .
the famous example is the band center for which the standard perturbation theory fails and one needs to use specific methods ( see discussion and references in ref .
@xcite ) .
however , recently the expression has been tested for the anderson model for non - resonant values of energy , and a perfect correspondence between the analytical predictions and numerical data has been manifested for two first moments of @xmath137 @xcite .
thus , our idea here is to explore the validity of the above expression for _ both _ non - resonant and resonant values of @xmath7 . in view of the results obtained in ref .
@xcite we also expect that away from the resonances , where the phase distribution is flat , the formula gives correct result . indeed , the data in figs .
[ pre - fig07 ] and [ pre - fig08 ] display an excellent agreement of the curves depicted with the use of eq . containing the lyapunov exponent , with the numerical results calculated via eq . in the regions of the phase shift @xmath7 where phase distribution is flat
on the other hand , our data have shown that in narrow regions of energy close to the resonances @xmath155 and @xmath156 , the expression completely ignores the presence of the resonances and gives incorrect results , provided the localization length is obtained by assuming the phase distribution is flat .
thus , our key point is to explore whether the same expression , however , with the correct lyapunov exponent can serve both for non - resonant and resonant values of @xmath7 .
the idea to combine the standard expression with the lyapunov exponent valid both in the non - resonant and resonant regions turns out to be very fruitful .
indeed , the data in figs .
[ pre - fig07](b , c ) and [ pre - fig08](b , c ) demonstrate an excellent agreement with our analytical predictions not only for the mean value of @xmath137 but also for the variance @xmath157 . note , however , that our approach does not work if the value of @xmath136 is not large enough .
this fact is clearly seen in figs .
[ pre - fig07](a ) and [ pre - fig08](a ) where @xmath158 .
the estimate of the lyapunov exponent @xmath42 for the chosen strength of disorder @xmath71 shows that for both resonances the localization length @xmath0 is much larger than the system size @xmath159 , this corresponds to the ballistic regime .
as one can see , a good correspondence between the data and our analytical approach occurs in a strongly localized regime , figs .
[ pre - fig07](c ) and [ pre - fig08](c ) , and in the intermediate regime where the localization length @xmath0 is of the order of @xmath159 .
again , we have to recall that the analytical expression works well for any ratio between the localization length and system size , provided the disorder is described by continuous potentials for which there are no resonance effects .
the failure of our approach in the ballistic regime is due to the non - stationarity of the phase distribution as we explain below . a closer inspection of the equation describing the evolution of phase @xmath3 shows that at the resonances the filling of the whole range @xmath40 $ ] by the phase is due to the terms containing the disorder , and not due to the constant drift due to non - resonant values of @xmath7 .
therefore , the length @xmath160 for the emergence of a stationary distribution for @xmath3 can be very large , in contrast with what happens out of resonances .
indeed , a rough estimate of this critical length @xmath160 gives @xmath161 .
thus , we have @xmath162 and @xmath163 for the resonances @xmath155 and @xmath156 , respectively .
these estimates explain the discrepancy which can be seen in figs .
[ pre - fig07](a , b ) and [ pre - fig08](a , b ) . according to eq .
the localization length at the resonances is @xmath164 . as one can see , the ratio between @xmath0 and @xmath165 is @xmath166 , the estimate which gives an additional information about the role of resonances in our model .
( color online ) average transmittance vs phase shift @xmath7 : ( a ) n=300 , ( b ) n=5000 , ( c ) n=50000 .
continuous curves correspond to the analytical expression complemented by eq . , circles represent the numerical simulation .
squares stand for the analytical eq . with @xmath42
numerically computed from eq . .
inset ( a1 ) is a zoom of the resonant region at @xmath155 , inset ( b1 ) is a zoom of the resonant region at @xmath156 .
the intensity of disorder is @xmath71.,title="fig:",width=302,height=188 ] ( color online ) average transmittance vs phase shift @xmath7 : ( a ) n=300 , ( b ) n=5000 , ( c ) n=50000 .
continuous curves correspond to the analytical expression complemented by eq . , circles represent the numerical simulation .
squares stand for the analytical eq .
with @xmath42 numerically computed from eq . .
inset ( a1 ) is a zoom of the resonant region at @xmath155 , inset ( b1 ) is a zoom of the resonant region at @xmath156 .
the intensity of disorder is @xmath71.,title="fig:",width=302,height=188 ] ( color online ) average transmittance vs phase shift @xmath7 : ( a ) n=300 , ( b ) n=5000 , ( c ) n=50000 .
continuous curves correspond to the analytical expression complemented by eq . , circles represent the numerical simulation .
squares stand for the analytical eq . with @xmath42 numerically computed from eq . .
inset ( a1 ) is a zoom of the resonant region at @xmath155 , inset ( b1 ) is a zoom of the resonant region at @xmath156 .
the intensity of disorder is @xmath71.,title="fig:",width=302,height=188 ] ( color online ) transmittance variance vs @xmath7 .
we use the same notations as in fig .
[ pre - fig07].,title="fig:",width=302,height=188 ] ( color online ) transmittance variance vs @xmath7 .
we use the same notations as in fig .
[ pre - fig07].,title="fig:",width=302,height=188 ] ( color online ) transmittance variance vs @xmath7 .
we use the same notations as in fig .
[ pre - fig07].,title="fig:",width=302,height=188 ] ( color online ) resonant size - dependent lyapunov exponent @xmath167 at @xmath155 vs system size @xmath136 .
circles show @xmath167 numerically calculated from eqs .
, , straight line presents @xmath168 defined by analytical eq . .
the system length @xmath136 is rescaled by @xmath169 .
the intensity of disorder is @xmath71.,width=302,height=188 ] as the next step towards a better agreement between the analytical description and numerical data , we can suggest to use the size - dependent lyapunov exponent @xmath167 defined as follows , @xmath170 then , one can try to use it in the integral formula , instead of the stationary lyapunov exponent @xmath42 .
it turns out that in this case one can get a quite good agreement with the numerical simulations , at least for the first two cumulants of the transmittance as figs .
[ pre - fig07](a ) and [ pre - fig08](a ) show . as one can see , the analysis of the ballistic regime ( @xmath171 ) in the vicinity of the resonances where @xmath172 requires two definitions , eqs . and , for the lyapunov exponent
the first ( standard ) definition is given for an infinite system , therefore , the stationary @xmath3-distribution is always achieved . on the other hand , eq .
is the prelimit counterpart of the first one , and , therefore , depends on the system size @xmath136 .
thus , eq . automatically takes into account an actual phase distribution and provides quite good results even in the resonant ballistic regime , where the phase distribution is non - stationary . fig .
[ pre - fig09 ] demonstrates that the size - dependent lyapunov exponent is saturated and becomes equivalent to the size - independent one at @xmath173 .
consequently , at this system length the stationarity of @xmath3-distribution is successfully reached , and the analytical expression is valid . indeed , the numerical simulations originated from eqs . , are in excellent agreement with analytical equations and .
this fact is clearly seen in figs .
[ pre - fig07](b , c ) and [ pre - fig08](b , c ) .
our results demonstrate that the formulas , provide quite good analytical description of the transport properties in comparison with numerical simulation even at the resonant energies .
this may suggest that the hypothesis of single parameter scaling ( sps ) is correct provided the value of the localization length takes into account the non - flat distribution of phases @xmath3 in the vicinity of resonances .
it should be however stressed that there are , at least , two different definitions of the sps .
the first one , which is a trivial consequence of the expression , is that all transport properties depend on the ratio between the localization length and the sample size .
another definition is originated from the analysis of tight - binding models of the anderson type and various kronig - penney models ( see discussion in ref .
@xcite ) . since for such models the resonances are unavoidable ,
the rigorous analysis in the general form is absent .
for this reason one uses another definition of the sps according to which all properties of the transport depend on the mean logarithm of transmittance , @xmath174 .
this is correct , e.g. , for the gaussian distribution of @xmath149 ( log - normal distribution ) if the ratio @xmath175 is assumed to be constant .
the latter occurs in strong localization regime and when random phase hypothesis holds true . under such circumstances @xmath176 equals @xmath177 , and
this result is often used as a proof or disproof of the sps ( see , e.g. , @xcite and references therein ) .
however , one has to bear in mind that random phase hypothesis is not a necessary condition for the validity of the sps @xcite . in view of this common approach ,
we have performed a careful numerical calculations of the parameter @xmath176 in a wide range of the sample size @xmath136 .
the results shown in fig
. [ pre - fig10 ] turn out to be quite unexpected .
namely , in spite of a good description of the first and second moments of @xmath137 by the expression with the correct lyapunov exponent , we have found the failure of the sps when exploring the variance of @xmath149 .
this is in a strong contrast with the result according to which if eq . for the moments of @xmath137 is valid ( and , therefore , the whole distribution of @xmath137 is known ) , the value of the parameter @xmath176 has to be 2 , as is predicted by the theory for the continuous systems ( see , e.g. , @xcite and references therein ) .
however , our data clearly demonstrate that @xmath178 at the resonant phase shift @xmath155 . whereas @xmath179 for the non - resonant phase shift @xmath180 in accordance with our expectation . due to these results
, one may conclude that the sps is not valid for resonant energies , however , more extensive studies , both the analytical and numerical ones , are required in order to resolve an apparent paradox demonstrated by the data .
( color online ) the parameter @xmath176 as a function of the system size rescaled by the inverse localization length @xmath42 for a disorder intensity of @xmath181 .
circles represent numerical data at the resonant phase shift @xmath155 and @xmath182 , while squares are the numerical result for the non - resonant value @xmath180 with @xmath183 .
the averaging @xmath147 was performed over @xmath184 different realizations of the disorder . to obtain the error
bars an ensemble of @xmath185 different random values of @xmath176 was used.,width=302,height=188 ] the authors are thankful to pier mello for fruitful discussions .
we acknowledge support form the sep - conacyt ( mxico ) under grant no .
cb-2011 - 01 - 166382 , viep - buap grant mebj - exc12-g , and pifca buap - ca-169 . | we study localization properties of the eigenstates and wave transport in one - dimensional system consisting of a set of barriers / wells of fixed thickness and random heights .
the inherent peculiarity of the system resulting in the enhanced anderson localization , is the presence of the resonances emerging due to the coherent interaction of the waves reflected from the interfaces between wells / barriers .
our theoretical approach allows to derive the localization length in infinite samples both out of the resonances and close to them .
we examine how the transport properties of finite samples can be described in terms of this length .
it is shown that the analytical expressions obtained by standard methods for continuous random potentials can be used in our discrete model , in spite of the presence of resonances that can not be described by conventional theories .
we also discuss whether the single parameter scaling is valid in view of the suggested modification of the theory .
all our results are illustrated with numerical data manifesting an excellent agreement with the theory . |
in this article we present a derivation of the thermal equations of state of ideal quantum gases by using solely dimensional analysis , thermodynamic considerations and simple physical arguments @xcite .
the thermodynamic relations of gases depend strongly on the energy - momentum relation of the gas particles and on their spin , i.e. , whether the particles are fermions or bosons .
the energy - momentum relation of free massive particles is given by @xmath0 @xmath1 denotes the rest mass of the particles , @xmath2 is the velocity of light .
here we have subtracted the rest energy @xmath3 from the particle energy which results in a redefinition of the chemical potential , i.e. @xmath4 has to be replaced by @xmath5 further on . in the _ non_-_relativistic _
limit @xmath6 , @xmath7 , whereas in the _
limit @xmath8 , @xmath9 .
both cases can be written in the single exponent dispersion relation @xmath10 where @xmath11 for relativistic and non - relativistic particles , respectively .
another group of particles to which our considerations apply are _
quasi_-_particles _ in condensed matter systems . in these systems deviations from the state of perfect order can be described by a superposition of quantized elementary excitations which behave as bosonic particles with an energy momentum relation also of the form ( [ eq : dispersion_relation ] ) .
the quasi - particles we are considering , as well as photons , have a vanishing chemical potential , i.e. their number is not fixed . at low temperatures this number is small and hence their interaction can be neglected .
quasi - particles are characterized by a fixed value of @xmath12 @xcite .
@xmath13 has to be expressed in terms of the quasi - particle parameters .
examples for the case @xmath14 are lattice vibration of solids with acoustic phonons as quasi - particles @xcite ( @xmath2 denotes now the sound velocity ) or spin waves in anti - ferromagnets with magnons as quasi - particles @xcite ( @xmath2 denotes here the spin wave velocity ) .
an example with @xmath15 are magnons in ferromagnets @xcite , where @xmath16 .
@xmath17 denotes the lattice constant and @xmath18 the curie temperature . for further examples of quasi - particles
see e.g. reference @xcite .
it is interesting to remark that our considerations can be extended to _
non_-_integer _ values of @xmath19 .
one example are capillary wave excitations of the surface of superfluid @xmath20 , the so - called ripplons @xcite which show a dispersion relation ( [ eq : dispersion_relation ] ) with @xmath21 and @xmath22 . here
@xmath23 and @xmath24 denote surface tension and the liquid density , respectively .
our results apply also to this case , but in the further discussion we will mainly focus on the cases of integer values of @xmath19 .
considering a homogeneous gas of @xmath25 point particles confined in a container of volume @xmath26 and coupled to a bath of temperature @xmath27 , the free energy can be written in the form @xmath28 @xmath29 is the free energy per particle .
the mean separation @xmath30 between particles is related to the volume per particle @xmath31 , where @xmath32 denotes the dimension of the system .
the pressure follows then from @xmath33 .
we begin with the consideration of a _ classical gas _ with fixed exponent @xmath19 of the dispersion relation ( [ eq : dispersion_relation ] ) . in this case
the pressure can only depend on the three parameters @xmath13 , @xmath30 and @xmath27 .
there is no further parameter .
the temperature @xmath27 appears in thermodynamics in the relation @xmath34 , where @xmath35 and @xmath36 denote the entropy and the energy , respectively .
entropy is determined only up to a multiplicative ( boltzmann ) constant @xcite . for the moment
we will only assume that this constant is dimensionless , i.e. we measure @xmath27 in energy units . denoting the dimension of mass , length and time by @xmath37 , @xmath38 and @xmath39 , respectively ,
the dimension of @xmath13 , @xmath30 , @xmath27 and @xmath40 are @xmath41 , @xmath38 , @xmath42 and @xmath43 , respectively .
since the three parameters have different dimensions , there is a unique combination of these which has the dimension of pressure @xcite @xmath44 as expected .
we fix now our temperature scale by choosing the constant of proportionality in ( [ eq : p_classical ] ) to be equal one , i.e. @xmath45 , which is the known result @xcite .
this suggest @xmath46 . clearly @xmath47 should occur in the free energy in combination with another volume to make the argument of the logarithm dimensionless .
this shows the limitations of classical physics . in the case of massive particles
there is only a dependence on @xmath1 _ or _
@xmath2 as long as we use the single power law relation ( [ eq : dispersion_relation ] ) with @xmath48 or @xmath49 fixed . however , for the description of the _ cross_-_over _ between the non - relativistic and the relativistic regime both @xmath1 _ and _ @xmath2 occur and hence a new dimensionless parameter @xmath50 appears . the relation ( [ eq : p_classical ] ) is formally correct even in the region @xmath51 .
however , one has to expect that in this region pair creation processes will take place and a grand canonical description is more appropriate , as we will consider in section [ grand_canonical_description ] . in the case of
_ charged _ particles the ideal gas properties are only preserved under certain conditions . if we consider for simplicity electrons ( on a neutralizing positively charged background ) the electron - electron interaction is of the order @xmath52 . to have still an ideal gas the coulomb energy has to be small compared with the thermal energy , i.e. @xmath53 or @xmath54 @xcite . in the quantum case
the planck constant of action @xmath55 appears as a new parameter .
the pressure is now no longer uniquely fixed by dimensional analysis since in order to determine the four exponents of @xmath56 and @xmath55 only three relations ( from the comparison of the powers of @xmath39 , @xmath38 and @xmath37 ) exist .
thus the equation of state will depend additionally on a _ dimensionsless parameter _
@xmath57 which includes @xmath55 as well as @xmath27 , @xmath30 and @xmath13 .
dimensional analysis shows that the only possible choice is @xcite @xmath58 @xmath59 here @xmath60 denotes the thermal de broglie wave length of the ( quasi- ) particles .
@xmath61 is the compton wave length in the case of massive particles .
we ignore all numerical pre - factors since they are beyond the accuracy of the dimensional analysis presented here .
since for @xmath62 , i.e. @xmath63 , we have to regain the classical result ( [ eq : p_classical ] ) , quantum behavior will be seen for @xmath64 , i.e. for @xmath65 .
the borderline @xmath66 defines a _ characteristic _ _ energy _
@xmath67 which turns out to be the fermi energy for fermions and the temperature of bose condensation for bosons , respectively .
this will be discussed in detail in the subsequent section .
in addition to the cross - over from the classical to the quantum regime , _ massive particles _ also show a cross - over from non - relativistic to relativistic behavior .
if the mean particle distance is given by @xmath30 , then for @xmath68 , the typical particle energy @xmath69 approaches the non - relativistic expression @xmath70 whereas for @xmath71 the relativistic expression @xmath72 applies .
from the mutual order of the three length scales @xmath73 and @xmath74 _ six _
regions can be distinguished @xcite . anticipating the result of the detailed discussions of the following sections these six regions exhibit very different physical properties : ( i ) @xmath75 : : : this is the region of classical behavior .
the only relevant length scale is the the inter - particle distance @xmath30 .
( ii ) @xmath76 : : : the region of non - relativistic quantum behavior .
fermions are degenerated , bosons condense . since @xmath77 relativistic effects do not play a role .
( iii ) @xmath78 : : : a region of quantum behavior where bosons behave non - relativistically ( as in ( iii ) ) whereas fermions exhibit relativistic effects .
the pauli principle makes the difference .
( iv ) @xmath79 : : : fermions show still low temperature behavior ( as in ( iii ) ) , bosons exhibit particle - antiparticle pair production .
( v ) @xmath80 : : : fermions and bosons are described by the same equation of state as photons .
( vi ) @xmath81 : : : same behavior as in ( v ) .
bose condensation as well as particle - antiparticle pair production will in general change the effective inter - particle distance from @xmath30 to @xmath82 and @xmath83 , respectively .
since pair production keeps some of the quantum numbers constant ( e.g. the total charge or the lepton or hadron number ) in these regions @xmath30 is a measure of the corresponding quantum number of the gas . for
_ charged _ fermions dimensional analysis gives a further intrinsic length scale , the bohr radius @xmath84 . for simplicity
we consider again a gas of electrons on a positively charged background . to have still an ideal gas the coulomb interaction @xmath85 has to be small compared to the kinetic energy . for non - relativistic electrons
this results in @xmath86 , i.e. , @xmath87 . for relativistic electrons
one obtains in the same way @xmath88 or @xmath89 , i.e. , the ideal gas condition is always fulfilled . for non - relativistic charged bosons outside the condensate
, @xmath30 has to be replaced by @xmath90 ( see below ) which leads to the condition @xmath91 .
the effect of coulomb interaction will be more pronounced in the condensate since the kinetic energy is zero .
indeed , it has been argued that there is _ no _ bose condensation in the presence of coulomb interaction @xcite .
the different regions of the parameter space can be visualized in the diagram depicted in figure [ fig : idal_gas1 ] where we plot @xmath92 versus @xmath93 in a double logarithmic representation .
the lines @xmath94 and @xmath95 correspond to the horizontal and vertical axis , respectively .
the line @xmath96 separates the classical region @xmath97 from the quantum region @xmath98 . in our double logarithmic plot
this line is straight with a slope equal to @xmath99 .
ordinary massive matter at temperatures realized on earth is restricted to the region below the horizontal axis . for electrons @xmath100 , and
hence for room temperatures @xmath101 this region is given by @xmath102 .
the bohr radius @xmath103 corresponds to @xmath104 .
so far we assumed that the particle number @xmath105 and hence @xmath30 was fixed .
in some cases it is more appropriate to consider the grand canonical description where instead of @xmath25 the _ chemical _ _ potential _
@xmath4 is given .
this description is in particular useful in those parameter regions where particle creation and annihilation processes ( like the inverse @xmath106-decay or electron - positron pair creation ) become important .
the corresponding thermodynamic potential @xmath107 is related to the pressure by the gibbs - duham relation @xmath108 since @xmath27 and @xmath4 have the same dimension there is _ no _ intrinsic length scale and hence no purely classical expression for the pressure . if we include @xmath55 as a further parameter @xmath109 appears as the _ only _ length scale of the problem .
@xmath110 is so far an arbitrary but dimensionless function of @xmath111 .
following our previous dimensional arguments the pressure can now be written as @xcite @xmath112 with @xmath113 likewise so far unknown . for particles with vanishing chemical potential
the pressure is _ always _ given by equation ( [ eq : p_grand_canonical ] ) with @xmath114 .
important examples are photons , magnons in anti - ferromagnets and acoustical phonons ( all @xmath14 ) , magnons in ferromagnets ( @xmath15 ) and ripplons ( @xmath115 ) . with ( [ eq : j ] ) we get @xmath116 and hence @xmath117 using the classical equation of state ( [ eq : p_classical ] ) this requires @xmath118 .
hence for a classical gas @xmath119 ( [ eq : p_grand_canonical ] ) and ( [ eq : omega_of_mu / t ] ) can be used to calculate the chemical potential of the classical gas @xcite @xmath120
since for @xmath121 the system has to approach the classical limit we make the following ansatz for the pressure @xmath122 with @xmath123 . here the subscript f or b stands for fermions or bosons .
the scaling function @xmath124 describes the cross - over from classical to quantum behavior of particles with a dispersion relation with fixed value of @xmath19 . for massive particles
this ansatz can be generalized to @xmath125 which describes in addition also the cross - over from non - relativistic to ultra - relativistic behavior .
the scaling function @xmath126 can be related to ( [ eq : p_scaling_ansatz ] ) in the limiting cases @xmath127 and @xmath128 , respectively . in the non - relativistic limit @xmath129 ,
i.e. @xmath127 , @xmath130 . in the ultra - relativistic case @xmath71 ,
i.e. @xmath128 , @xmath131 . in the following
we will mainly concentrate on the discussion of @xmath132 .
we consider first the case of _ weak _ _ quantum _ _ fluctuations_. assuming that the pressure can be expanded in powers of the particle density , i.e. in powers of @xmath133 , ( the classical result ( [ eq : p_classical ] ) is the lowest order term of this expansion ) @xmath132 can be written in the form @xmath134 here the plus and minus sign corresponds to fermions and bosons , respectively . indeed ,
for fermions the pauli principle will lead to an increase of the pressure with respect to the classical case . for bosons
the pressure is reduced since quantum mechanics increases the probability for the double occupancy of a state @xcite .
( [ eq : x ] ) , ( [ eq : p_scaling_ansatz ] ) and ( [ eq : p_weak_corrections ] ) describe indeed the lowest order quantum correction to the classical equation of state@xcite .
in the _ strong quantum limit _ @xmath135 , we have even more pronounced differences between fermions and bosons . for _ fermions _ the pauli - principle guarantees a non - zero pressure even for @xmath136 .
from ( [ eq : x ] ) and ( [ eq : p_scaling_ansatz ] ) this requires @xmath137 and hence @xmath138 on the r.h.s .
we have introduced the _ fermi _ _ energy _ @xmath139 which is the kinetic energy of the fermions with momentum @xmath140 .
apparently , in ( [ eq : p_fermions_general ] ) the fermi energy replaces the temperatures of the classical expression ( [ eq : p_classical ] ) .
in particular this gives for relativistic particles @xcite @xmath141 and for non - relativistic particles @xcite @xmath142 respectively .
both cases can be combined in the scaling formula @xmath143 the precise form of the scaling function @xmath144 remains unknown in this approach .
the cross - over between ( [ eq : p_fermions_non - relativistic ] ) and ( [ eq : p_fermions_relativistic ] ) takes place at @xmath145 , as follows from the general properties of @xmath144 discussed before .
as long as @xmath76 ( region ( ii ) in figure 1 ) the particles show non - relativistic quantum behavior .
this situation exists e.g. for electrons in solids ( @xmath146 , @xmath147 ) or neutrons in neutron stars ( @xmath148 , @xmath149 ) . on the other hand , for @xmath150
the kinetic energy is larger than @xmath3 and the particles behave relativistically .
one has to expect that this low temperature relativistic quantum behavior is seen as long as the temperature is small compared to the fermi energy @xmath151 , i.e. for @xmath152 ( corresponding to regions ( iii ) and ( iv ) of figure 1 ) .
relativistic electrons exist e.g. in white dwarfs ( @xmath148 ) . for _ very _ _
dense _ systems of atoms , where nuclei and electrons form a plasma with @xmath153 , electrons and protons may _ recombine _ to produce a neutron and a neutrino ( the inverse @xmath106-decay ) .
the charge of the resulting nucleus is then reduced by one . in this case
the electron density is no longer an independent variable but determined by the chemical potential @xmath154 of this process .
if @xmath155 and initially @xmath156 the system is in the ultra - relativistic region , thermal fluctuations can be neglected .
@xmath157 determines then unambigously a new length scale @xmath158 corresponding to @xmath159 in ( [ eq : lambda_grand_canonical ] ) which replaces @xmath30 in the region @xmath160 in eq.([eq : p_fermions_relativistic ] ) .
since thermal fluctuations are irrelevant and @xmath30 is not longer fixed , @xmath27 is replaced by @xmath161 in all expressions .
the pressure is therefore reduced to @xcite @xmath162 we come now to a discussion of the parameter region of ultra - high temperatures of fermionic systems where particle - antiparticle _ pair production _ becomes important ( regions ( v ) and ( vi ) of figure 1 ) . in this region
the particle number is not conserved and it is more adequate to describe the system by specifying a chemical potential , as we have done in subsection [ grand_canonical_description ] . for simplicity
we assume that only fermions of the same kind ( and their corresponding antiparticles ) are generated .
as discussed in subsection [ quantum_case ] the coulomb interaction can be neglected since the effective particle distance turns out to be much smaller than the bohr radius @xmath103 . for temperatures @xmath163 the pressure
is determined by ( [ eq : p_grand_canonical ] ) with @xmath164 and @xmath14 ( since we are in the relativistic region ) , i.e. @xcite @xmath165 which agrees with the result for photons ( compare ( [ eq : p_grand_canonical ] ) ) .
hence we will call ( [ eq : p_fermions_radiation ] ) _ radiation _ _ behavior_. @xmath83 has to be considered as the mean distance between the fermions which is here smaller than @xmath30 because of pair production .
if several kinds of fermions exist each of it gives the same contribution ( [ eq : p_fermions_radiation ] ) to the pressure .
finally we consider the cross - over between ( [ eq : p_fermions_radiation ] ) and the pressure formulas valid at lower temperature .
we begin with the region @xmath166 . lowering temperature
we expect a cross - over to the classical result ( [ eq : p_classical ] ) . to describe the cross - over we have to understand the contribution of pair production if we raise temperature starting in the classical region .
it is therefore convenient to use a grand canonical description also for the description of this cross - over . to make contact with the classical result ( [ eq : p_classical ] ) , we have to express the chemical potential of particles and anti - particles in terms of the parameters of the system .
particles and anti - particles can annihilate into photons . since the sum of the chemical potentials is conserved in this reaction @xcite and the chemical potential of photons is zero , we find that the chemical potentials of particle and anti - particles have opposite sign , say @xmath167 . from ( [ eq : p_grand_canonical ] ) and ( [ eq : omega_of_mu / t ] ) we conclude that in the classical region the density of particles and anti - particles , respectively , is given by @xmath168 here
we have included explicitly the term @xmath169 resulting from the shift of the energy scale discussed in [ introduction ] .
if the density of _ generated _ particles is small compared to the initial density @xmath170 such that @xmath171 we get after elementary manipulations for the density of the generated particles @xmath172 the cross - over to the radiation region sets in if the density ( [ eq : pair_density_fermions ] ) of generated particles is a finite fraction of the initial density @xmath170 .
this gives for the cross - over line @xmath173 ( compare figure 2 ) @xmath174 in the relativistic region @xmath156 essentially the same arguments apply but now we have to replace @xmath82 by @xmath83 in ( [ eq : particle_antiparticle_density ] ) .
the cross - over follows then at @xmath175 which corresponds for @xmath156 essentially to @xmath176 ( compare figure 2 ) , in agreement with our previous conjecture . for _ bosons _ the low temperature pressure can not depend on the volume since all bosons may accumulate in the state of lowest energy .
this requires in ( [ eq : p_scaling_ansatz ] ) @xmath177 from which we get for the pressure @xmath178 where on the r.h.s .
we introduced the temperature of _ bose condensation _
@xmath179 in particular , we get for relativistic particles @xcite @xmath180 and for non - relativistic particles @xcite @xmath181 again both the relativistic and the non - relativistic case can be combined into the scaling form @xmath182 the cross - over between ( [ eq : p_bosons_non - relativistic ] ) and ( [ eq : p_bosons_relativistic ] ) occurs at @xmath183 , i.e. at @xmath184 . since @xmath185 the pressure is _ reduced _ with respect to the classical result ( [ eq : p_classical ] ) .
this can be explained by assuming that a finite fraction of particles sits now in the state of zero energy and hence does not contribute to the pressure . comparing ( [ eq : p_classical ] ) and ( [ eq : p_bosons ] )
we conclude that @xmath186 has to be considered as the mean distance of particles _ above _ the condensate @xcite .
next we discuss the phase diagram of massive bose particles obeying the full dispersion relation ( [ eq : relativistic_dispersion_relation ] ) .
we begin with the classical region @xmath187 where the pressure is given by ( [ eq : p_classical ] ) . reducing @xmath27
we reach at @xmath188 the region where a macroscopic fraction of particles does not contribute to the pressure since it is in the state of zero energy .
as argued above , the number @xmath189 of the remaining particles of finite energy can be written as @xmath190 thence the number of particles in the condensate is @xmath191 which shows that @xmath192 is indeed the transition temperature of bose condensation @xcite . if one enters the condensed phase by lowering @xmath27 at _ fixed _
@xmath30 , the total number of bosons remains fixed whereas the mean distance of the particles above the condensate is equal to @xmath193 , which increases if we further decrease temperature . on the other hand , if one approaches the transition line by decreasing @xmath30 at _ fixed _
@xmath27 , the mean distance @xmath90 between particles above the condensate remains constant ( since @xmath27 is constant ) whereas the excess particles go to the ground state .
this behavior is seen also for @xmath150 since the effective distance of particles contributing to the pressure remains @xmath194 ( compare fig.3 ) .
thus , bose condensation as described by equations ( [ eq : p_bosons_non - relativistic ] ) and ( [ eq : condensate ] ) is seen both in regions ( ii ) and ( iii ) of figure 1 . at sufficiently high temperatures we have to expect that production of particle - antiparticle pairs sets in . for simplicity
we will again assume that only bosons of the same type can be created ( and their anti - particles ) .
at temperatures @xmath195 bosons behave relativistically and the pressure is given by ( [ eq : p_bosons_relativistic ] ) . the discussion of the cross - over to ( [ eq : p_classical ] ) if @xmath196 and to ( [ eq : p_bosons_non - relativistic ] ) if @xmath156 , respectively , follows essentially the same lines as in the case of fermions , with the important difference that in describing the cross - over in the region @xmath197 the effective particle distance is now @xmath82 ( instead of @xmath30 as in the case of fermions ) .
from ( [ eq : pair_density_fermions ] ) we therefore conclude that the cross - over occurs at @xmath198 .
note , that the cross - over is smooth , since at @xmath184 , @xmath199 if @xmath156 . since in the relativistic region ( iv ) , @xmath79 , the effective particle density @xmath200 is smaller than @xmath170 one is inclined to conclude that here both pair creation and bose condensation takes place . for @xmath196
the cross - over is the same as for fermions .
an important application of the various equations of state can be found in the history of _ stars _ which we present here in a cartoon like picture .
initially the star burns hydrogen to helium and other light elements and keeps in this way its temperature @xmath27 .
the ratio @xmath201 is of the order @xmath202 , where @xmath203 is the proton mass .
this is the present situation of our sun .
the pressure of a star can roughly be described by the classical equation ( [ eq : p_classical ] ) , which counteracts the gravitational pressure @xmath204 here @xmath205 and @xmath37 denote the radius and the mass of the star , respectively , and @xmath206 is the gravitational constant .
we do not present the derivation of ( [ eq : star ] ) , which is straightforward , but remark that @xmath207 is the gravitational energy of a sphere @xmath208 , as one obtains also from a dimensional argument .
the ratio of the mass to the radius of a star in this state is controlled by the thermal equation of state ( [ eq : p_classical ] ) .
indeed , from the equality of ( [ eq : p_classical ] ) and ( [ eq : star ] ) follows @xmath209 , which increases linearly with temperature .
once the nuclear fuel is exhausted , temperature can not be kept and the star collapses under the graviational force . as a result the density increases and the star matter crosses over to the region ( ii ) of figure 1 , where the pressure is given by ( [ eq : p_fermions_non - relativistic ] ) .
since the pressure is now inversely proportional to the mass , mainly the electrons contribute to it .
once under further contraction @xmath30 has reached @xmath210 ( corresponding to densities of the order @xmath211 ) ( [ eq : p_fermions_relativistic ] ) has to be applied for the pressure .
the star has reached the state of a _ white dwarf_(compare figure 2 ) . if the density is increased further ( i.e. @xmath212 ) ,
neutrons are generated via the inverse @xmath106-decay .
the pressure is determined then by ( [ eq : p_inverse_beta_decay ] ) until most of the nucleons became neutrons . for the pressure in _ neutron stars _
formulas ( [ eq : p_fermions_non - relativistic ] ) for @xmath196 and ( [ eq : p_fermions_relativistic ] ) for @xmath156 again apply , but the mass @xmath213 is now that of the neutrons ( compare figure 2 ) .
this is the fate of stars with a mass between 15 and 30 times the mass of the sun . in the ultra - relativistic regions of white dwarfs and neutrons
stars the equality between ( [ eq : star ] ) and ( [ eq : p_fermions_relativistic ] ) results in a relation for the mass @xmath214 this so - called _ chandrasekhar mass _ @xmath215 is the maximum mass of white dwarfs and neutron stars .
its value is determined by the thermal equation of state in region ( iii ) and is of the order of twice the mass of the sun . as a second illustration let us consider a _
contracting toy universe _ . for simplicity , this universe is supposed to have no gravity , i.e. there are no stars , clouds of cosmic dust etc .
, just photons and diluted massive particles . to make things even simpler
the only particles assumed to exist are hydrogen atoms ( a non - metallic universe in the language of astronomers ) .
initially the photon spectrum is assumed to correspond to black body radiation of about 3k . since for massless particles @xmath216 is the only existing length scale , the typical distance of photons is of the order @xmath217 .
the number of h - atoms per @xmath218 is assumed to be of order 1 .
hydrogen atoms are bosons and hence could be represented in figure [ fig : ideal_gas1_bosons ] .
the horizontal and vertical coordinates of the hydrogen gas in figure [ fig : ideal_gas1_bosons ] are given by @xmath219 and @xmath220 , respectively , i.e. they are outside of the margins of this figure . @xmath221 is the compton wave length of hydrogen .
next we switch on the cosmic contraction , i.e. we go backwards in time .
it is convenient to introduce the contraction factor @xmath222 , where @xmath223 denotes the distance at an earlier time @xmath224 when the final distance was @xmath225 .
since contraction of space reduces the wavelength of photons such that @xmath226 with @xmath227 , i.e. the photon gas heats up . without coupling to the photons ,
the hydrogen atoms form a classical non - relativistic gas ( region ( i ) of figure 1 ) which contracts adiabatically .
its temperature @xmath228 rises proportional to @xmath229 , _ provided _ there is enough interaction between the hydrogen atoms to reach thermal equilibrium .
the horizontal and vertical coordinates of the hydrogen gas in figure 3 thence change according to @xmath230 and @xmath231 , respectively .
when the temperature of the photons reaches about @xmath232 ( @xmath233 ) , the hydrogen atoms have a distance of about @xmath234 and are now partially ionized , i.e. we have free protons and electrons , both fermions , which appear in figure [ fig : ideal_gas1_fermions ] . since both particles are charged , they will emit and adsorb photons . from now on matter and radiation have approximately the same temperature @xmath235 .
the initial coordinates are @xmath236 , @xmath237 ( i.e. @xmath238 ) for protons and @xmath239 , @xmath240 ( i.e. @xmath241 ) for electrons , respectively .
thus their initial positions are in region ( i ) of figure 1 .
as long as @xmath242 , electrons and protons form initially an non - interacting non - relativistic classical plasma ( compare figure 2 ) . since @xmath243 and @xmath244 , the slope of the path under further contraction in the double logarithmic plot of figure 2 is @xmath245 and remains unchanged until we reach the relativistic region @xmath246 . when the contraction parameter reaches @xmath247 , i.e temperatures reach about @xmath248 ,
the mean particle distance is of the order of the bohr radius .
when temperatures reach the rest energy of the electrons , i.e. crossing the bisecting line @xmath249 of figure 2 , the electrons start to behave relativistically . for @xmath250 electron - positron pairs
are formed and leptons behave ultra - relativistically . for temperatures @xmath251 and @xmath252 hadrons decay into quarks and anti - quarks , which also behave ultra - relativistically . also their mean distance is now determined by pair production .
in addition gluons , i.e. the bosons which carry the interaction between quarks , have to be considered .
eventually the coordinates of all fundamental fermions and bosons reach the line @xmath176 .
the pressure in all cases is that or ultra - relativistic fermions ( [ eq : p_fermions_radiation ] ) or bosons ( [ eq : p_bosons_relativistic ] ) , respectively .
the total pressure is the sum of these contributions , i.e. it is proportional to the number of different fermions and bosons .
we can also use the results of the previous chapters to determine the scaling behavior of the _ free _ _ energy _
@xmath29 per particle which can be written as @xmath253 differentiation with respect to @xmath31 gives @xmath254 thus @xmath255 .
the entropy per particle @xmath256 follows from @xmath257 and hence we get for the energy per particle @xmath258 @xmath259 which is the well known relation between the energy @xmath258 and @xmath260 both for relativistic and non - relativistic particle in any dimension@xcite . for the chemical potential @xmath4 we obtain from the gibbs - duham relation @xmath261 with our anstze used above we find at low temperatures @xmath262 for fermions and @xmath263 for bosons .
these expressions can be used to calculate additional thermodynamic quantities .
specific _ _ heat _ follows from ( [ eq : dif_f ] ) as @xmath264 for _ fermions _ our ansatz @xmath265 , i.e. @xmath266 , gives @xmath267 instead of the correct result @xmath268 .
this is not surprising since we determined @xmath269 from the condition that the free energy should go to a constant ( the fermi energy ) for @xmath270 . to get the correct specific heat we have to add finite temperature corrections to @xmath258 .
indeed , the low temperature corrections to energy of the _ fermi gas _ result from the difference @xmath271 of the fermi distribution at non - zero and zero temperatures ( the latter is the step function ) .
these corrections are of the form @xcite @xmath272 where @xmath273 , and @xmath274 is the density of states which is smooth at @xmath275 . with @xmath276
it is clear that the low temperature expansion of @xmath258 includes only _ even _ powers in @xmath27 .
since @xmath277 is the only energy scale , the low temperature expansion of the energy of the ideal fermi gas is therefore of the form @xmath278 which is the sommerfeld expansion .
this gives for the specific heat @xmath279 @xcite .
on the other hand , calculating the compressibility we get with @xmath280 immediately the correct result @xmath281 antiferromagnets are magnets with an opposite magnetization on two inter - penetrating sublattices .
see e.g.l .
d. landau and e. m. lifshitz : statistical physics , part 2 , 7th ed . , akademie verlag berlin , chapter * 74*. to get this relation one starts from the ansatz for the pressure @xmath284 and compares the powers of the dimensions of @xmath285 on both sides of this relations , which determines the exponents @xmath286 .
considering the partition of two particles on two states @xmath293 , @xmath294 of equal energy , classically there are 4 possibilities , @xmath295 , @xmath296 , @xmath297 and @xmath298 .
the probability that two particles are in the same state , say @xmath293 , is 1/4 .
in quantum mechanics only the states @xmath295 , @xmath296 and @xmath299 exist , i.e. the probability for having two particles in the same state is 1/3 . | the thermodynamic properties of ideal quantum gases are derived solely from dimensional arguments , the pauli principle and thermodynamic relations , without resorting to statistical mechanics . |
the particle motion on a given curved surface is old but interesting problem in wide range of physics .
especially the diffusion process of particles on such a manifold is still an open problem , and related to various kinds of phenomena .
for example the motion of protein on cell membrane has great importance in biophysics .
there are several research papers discussing on this problem .
some of them are treating this problem by using usual diffusion equation with curved coordinate , and discuss the curvature ( gauss curvature ) dependence of its solution @xcite .
other of them use the langevin equation on curved surface and calculating the curvature dependence of diffusion coefficient @xcite .
the quantum mechanics of particle motion on such a curved manifold is also considered by many authors .
this problem is usually explained by the schroedinger equation with laplace - beltrami operator .
however , when we treat the curved surface as embedded one in 3 dimensional euclidean space , situation is changed and then we have a quantum potential term related to the curvature additional to the kinetic operator @xcite,@xcite,@xcite .
another example is in larger scale physics in which our consideration is devoted .
patterns of animal skins are well expressed by the reaction diffusion equation @xcite .
but the patterns are different for each parts even in one individual . for example , char fish , the side part has white spot pattern , but the back part has labyrinth pattern .
( for these two patterns , see for example @xcite . )
one of the reasons might come from the curvature difference between side part and back part .
if the diffusion is influenced by the curvature , this difference of patterns might be explained .
furthermore , the cross section of fish has form of ellipsoid and the surface can be approximated as the one of elliptic cylinder . in two dimensional space , we have only two kinds of curvature , one is gauss curvature and other is mean curvature .
both are constructed from second fundamental tensor by taking determinant or trace .
gauss curvature can also be constructed only by metric tensor and its derivatives , but this is not the case for the mean curvature .
the elliptic cylinder , in which we have much interest , has zero gauss curvature and non - zero mean curvature . therefore to explain the pattern change of char fish , solution of the diffusion equation should depend on mean curvature
this is impossible if we start from usual diffusion equation because it depends only on metric but not on second fundamental tensor .
therefore we need some new diffusion equation , which bring not only gauss curvature but also mean curvature . in this article
, we discuss how to construct such curvature dependent diffusion equation .
the simple extension of diffusion equation in euclidean space to riemannian space can be done by changing laplacian with cartesian coordinate to the one with curved coordinate , i.e. laplace beltrami operator .
this coordinate change is not enough for our purpose , however .
the way of construction of new diffusion equation in this paper is the followings .
we re - identify the two dimensional diffusion as the limiting process from three dimensional diffusion .
we place the curved surface @xmath2 in three dimensional euclidean space @xmath0 , and we put two similar copies of @xmath2 , called @xmath3 and @xmath4 at a small distance of @xmath5 .
our particles can only move between these two surfaces , and later we take a limit @xmath6 .
we look for the form of diffusion equation in this limit .
the coordinates we use hereafter is the followings .
( see fig.1 ) @xmath7 is the cartesian coordinate in @xmath0 .
@xmath8 is the cartesian coordinate which specifies only the points on @xmath2 .
@xmath9 is the curved coordinate on @xmath2 .
( small latin indices @xmath10 runs from 1 to 2 .
) @xmath11 is the coordinate in @xmath0 normal to @xmath2 .
further by using the normal unit vector @xmath12 on @xmath2 at point @xmath13 , we can identify any points between two surfaces @xmath4 and @xmath3 by the following thin - layer approximation @xcite .
@xmath14 where @xmath15 . from this relation
we can obtain the curvilinear coordinate system between two surfaces ( @xmath16 ) by the coordinate @xmath17 , and metric @xmath18 .
( hereafter greek indices @xmath19 runs from 0 to 2 . )
@xmath20 each part of @xmath18 is the following . @xmath21 where @xmath22 is the metric on @xmath2 .
hereafter indices @xmath23 are lowered or rised by @xmath24 and its inverse @xmath25 .
we also obtain @xmath26 we can proceed the calculation by using the new variables .
we first define the tangential vector to @xmath2 by @xmath27 note that @xmath28 .
then we obtain two relations .
gauss equation : @xmath29 weingarten equation : @xmath30 where @xmath31 @xmath32 is called euler - schauten tensor , or second fundamental tensor defined as @xmath33 the second fundamental tensor @xmath32 is the projection of @xmath34 into the surface .
furthermore , the mean curvature is given by @xmath35 and ricci scalar curvature @xmath36 is obtained by @xmath37 then we have the formula for metric of curvilinear coordinate in a neighborhood of @xmath2 .
@xmath38 under the inversion @xmath39 , we have @xmath40 as well as @xmath41 from @xmath42 .
therefore @xmath43 is invariant under @xmath39 .
+ now we have the total metric tensor such as , @xmath44
let us denote 3 dimensional diffusion field as @xmath45 , and laplacian as @xmath46 .
then we have the equation with normalization condition @xmath47 where @xmath48 is the diffusion constant , and @xmath49 .
our aim is to construct the effective two dimensional diffusion equation from 3d equation above . @xmath50 where @xmath51 is the two dimensional diffusion field , @xmath52 , and @xmath53 is unknown effective 2d diffusion operator which might not be equal to simple 2d laplace beltrami operator . from two normalization conditions ,
we obtain @xmath54 ~ \sqrt{g } ~ d^2 q , \nonumber\\ & = & \int \phi^{(2)}(q^1,q^2 ) \sqrt{g}~ d^2 q.\nonumber\end{aligned}\ ] ] therefore we obtain the relation , @xmath55 where @xmath56 we multiply @xmath57 to equation ( [ eq : diff ] ) and integrate by @xmath11 , then we obtain @xmath58 where @xmath59
next we analyze new operator @xmath60 . from the form of laplace beltrami operator @xmath61 we have @xmath62 where @xmath63 and @xmath64 then our diffusion equation has form @xmath65 the contribution from @xmath66 vanishes because @xmath67 \mid _ { -\epsilon/2}^{\epsilon/2 } = 0 .
\end{aligned}\ ] ] the last equality is the requirement that diffusion flow does not pass through the surface : @xmath4 and @xmath3 .
+ now we calculate r.h.s of ( [ eq : new ] ) up to @xmath68 .
since we have @xmath69 from ( [ eq : relation ] ) , we need to expand @xmath70 up to @xmath68 .
the following relations are useful @xmath71 where @xmath72 is used . then the operator @xmath70 can be expanded as follows @xmath73 where , @xmath74 @xmath75 then our two dimensional effective diffusion equation up to @xmath76 is , @xmath77 to proceed the @xmath11 integration , we suppose there is no diffusion flow in normal direction in layer , that is , @xmath78 solution is , @xmath79 each integration can be explicitly performed , and we obtain @xmath80 where we have used the definition @xmath81 we obtain the final form of equation up to @xmath82 as @xmath83 where @xmath84 .
+ we give two comments here .
first , @xmath76 term disappears .
since @xmath1 has the dimension of length , it always appears with curvature @xmath85 .
therefore the 1st order term , if it exists , it contains 1st order of curvature @xmath85 . but this curvature depends on unphysical choice of normal unit vector @xmath86 , and so it does not appear .
second , additional potential term disappears . in quantum mechanics ,
the similar embedding techniques leads to the appearance of additional potential term written by curvature .
but in our classical case we have no such terms .
because in diffusion equation , potential term breaks probability conservation law , i.e. @xmath87 @xmath88 the normal diffusion flow can be written in general coordinate , @xmath89 the anomalous diffusion flow is , @xmath90 the diffusion equation is written as @xmath91 where @xmath92 is the covariant derivative . by using a suitable boundary condition , we can prove @xmath93
the anomalous flow equals to zero for the flat surface .
the last term in equation ( [ eq : anom ] ) shows that curvature gradient generate the flow without particle density gradient . from the signature of this term
, this flow goes from the smaller ricci scalar point to the larger ricci scalar point .
ricci scalar can take the negative , zero , and positive values .
( ricci scalar @xmath36 is related to gauss curvature by @xmath94 $ ] . )
let us work with the coordinate which satisfies @xmath95,\ ] ] at point @xmath96 , where @xmath97 is the curvature radius along the @xmath9 coordinate and it takes positive or negative value for convex or concave .
( the metric can be diagonalized by choosing the two coordinates as to satisfy orthogonality , and it can be normalized by using the re - parametrization .
the second fundamental tensor is diagonalized by rotation of coordinate . )
then at point @xmath96 , we have @xmath98 and , * @xmath99 if the surface is convex or concave . * @xmath100 if the surface is essentially flat . *
@xmath101 if the surface is hyperbolic .
therefore the flow goes from hyperbolic or flat points to convex or concave points with positive larger ricci scalar value .
+ next we consider the first term in ( [ eq : anom ] ) .
we have positive or negative value for @xmath102 depending on the value of curvature . in our coordinate
, we can immediately write it in the simple form @xmath103 when the surface is hyperbolic ( @xmath104 ) , @xmath105 usual diffusion occurs ( see fig .
2 ) .
when the surface is convex or concave ( @xmath106 ) , @xmath107 @xmath108 in this case , we have three possibilities .
one possibility is that both are negative , if @xmath109 then we have no diffusion but concentration occurs ( see fig .
3 ) .
the second possibility is that one of two is positive and the other is negative , if @xmath110 then we have diffusion in one direction , but concentration in another direction ( see fig .
4 ) . the third possibility for @xmath106 is , @xmath111 this is critical point , where the flow stops for larger curvature direction and flow concentrates for smaller curvature direction
. + when the ricci scalar is zero ( @xmath112 ) , for example @xmath113 , @xmath114 and @xmath115 follows .
the diffusion occurs only in @xmath116 direction but not in another direction .
( see fig .
5 ) . in this way
, this anomalous diffusion flow has much varieties depending on the curvature .
let us consider one simple example where diffusion coefficient depends on curvature .
the surface of elliptic cylinder is the case of @xmath117 just as figure 5 , but the surface has non zero mean curvature .
ellipsoid is given by the equation @xmath118 then any points on cylinder are specified by curved coordinate @xmath119 and @xmath120 ; @xmath121 another choice of coordinate instead of @xmath119 is , @xmath122 where @xmath123 is defined as @xmath124 the length of @xmath125 is given by @xmath126 with @xmath127 @xmath128 is the elliptic integral of the second kind .
( see appendix ) + the total length of @xmath125 is given by @xmath129 we use the normalized value for @xmath125 hereafter such that , @xmath130 the normal unit vector @xmath86 is given as @xmath131 @xmath132 @xmath133 then we obtain the second fundamental tensor .
@xmath134 then we collect all the necessary quantities as follows @xmath135 then we obtain the total diffusion equation expressed by the parameters @xmath119 and @xmath120 .
@xmath136 where the effective diffusion coefficient depends on mean curvature .
@xmath137 where @xmath138 . under this equation , we obtain the following particle number conservation law . @xmath139 with suitable neumann boundary condition . by using the variable @xmath140 instead of
@xmath119 we have simple dimensionless equation , @xmath141 where @xmath142 . + by using the approximation of elliptic function given in appendix
, curvature dependent potential @xmath143 can be written as function of @xmath140 .
the simulation can be done as usual diffusion equation . for @xmath144 and @xmath145 using periodic boundary condition ,
since the length of @xmath146 is larger than one of @xmath140 , the diffusion in @xmath125 direction occurs fastly and then the diffusion in @xmath146 direction follows slowly just like @xmath147 .
the @xmath125-directional diffusion can not occur uniformly , because at @xmath148 and @xmath149 the diffusive coefficient is higher than other points .
therefore the slope of diffusion field is small especially at these two points during the diffusion process .
7 ) as the initial condition . at two points ( 0.25 and 0.75 ) ,
diffusion occurs quickly and its slope is smaller than other.,width=340 ] when @xmath150,width=151 ]
we have discussed on the diffusion equation on curved surface embedded in @xmath0 .
we obtained the new diffusion equation up to @xmath68 which includes anomalous diffusive flow additional to the usual one .
this anomalous flow depends on the second fundamental tensor , and it has not only diffusion but also concentration properties depending on the curvature of its manifold . at the point with negative ricci scalar @xmath104 , surface is hyperbolic , and diffusion in both direction occurs .
( fig.2 ) when ricci scalar is positive @xmath106 , we have three possibilities .
@xmath97 appearing below is curvature radius in each direction ( @xmath151 ) . * concentration in both direction ( fig.3 ) , when @xmath109 * concentration in smaller curvature direction , and diffusion in higher curvature direction ( fig.4 ) , when @xmath110 * concentration in smaller curvature direction , and no flow in higher curvature direction , when @xmath152 when ricci scalar is zero @xmath117 , surface is essentially flat , but we have finite curvature radius in one direction
. then we have diffusion only in this direction ( fig.5 ) . in the case of surface of elliptic cylinder
, we gave a concrete form of equation and we showed the curvature dependent diffusion coefficient .
@xmath153 where @xmath85 is the mean curvature . in this case curvature dependence
is simply included into diffusion coefficient .
however this is not true in general case , where situation is much more complicated , and this can be seen from the form of anomalous flow .
the application to the pattern formation by reaction diffusion using this obtained equation is not yet finished .
this will be done in further publication .
we approximate the elliptic integral of the second kind .
n. ogawa , k. fujii , and k. p. kobushkin , prog .
theor . phys .
* 83 * ( 1990 ) 894 ; n. ogawa , k. fujii , n. m. chepilko , and k. p. kobushkin , prog .
theor . phys .
* 85 * ( 1991 ) 1189 ; n. ogawa , prog .
* 87 * ( 1992 ) 513 .
a. m. turing , phil .
london * b 237 * ( 1952 ) 37 - 72 ; h. meinhardt , models of biological pattern formation . , academic press , london ( 1982 ) ; j. d. murray , mathematical biology .
springer , new york ( 1989 ) . | particle diffusion in a two dimensional curved surface embedded in @xmath0 is considered .
in addition to the usual diffusion flow , we find a new flow with an explicit curvature dependence .
new diffusion equation is obtained in @xmath1 ( thickness of surface ) expansion . as an example
, the surface of elliptic cylinder is considered , and curvature dependent diffusion coefficient is calculated . |
the state of matter at extremely hot or dense conditions is one of the most fundamental questions in physics .
terrestrial nuclei of heavy elements provide a glimpse at such extreme states , but there are far - more - extreme states of matter in our universe . a neutron star , which may be regarded as a huge nucleus , is one example of such an extreme state of matter .
the state of matter can be characterized by the equation of state ( eos ) . from the microscopic point of view the eos
is determined by how the constituent particles of the matter interact with each other .
therefore the state of matter at extreme conditions can be understood when we have enough knowledge about the properties of the constituent particles and how they interact in such conditions . in the early 90 s , brown and
rho proclaimed the scaling property of hadron masses in dense medium in terms of the scale invariance of the effective lagrangian @xcite .
they showed that the ratios of the in - medium masses of the nucleon , @xmath0- , @xmath1- and @xmath2-mesons to their masses in free space are approximately equal to each other at around the nuclear saturation density .
afterwards , changes in the hadron masses in medium were calculated in the framework of qcd - sum rule @xcite and quark - meson - coupling ( qmc ) models @xcite . in refs .
@xcite and @xcite , masses of @xmath2 and @xmath1 mesons were calculated at the nuclear saturation density , and similar amount of mass reduction from the free mass , @xmath3 was predicted .
experimentally , microscopic properties of hadrons as well as the states of matter at hot or dense environment can be probed in relativistic heavy - ion collisions , and possible meson mass reductions in nuclear matter were reported through the dileptonic decay of @xmath2 and @xmath1 mesons in the ceres / na45 @xcite and kek - ps e325 @xcite experiments . also , recent astronomical observations of binary systems yield the mass of compact objects as @xmath4 @xcite and @xmath5 @xcite , of which the former is believed to be a neutron star .
such a large mass of the neutron star requires a substantially stiff eos . motivated by these theoretical and experimental studies ,
in this work we take into account the reduction of meson masses in the nuclear models such as quantum hadrodynamics ( qhd ) and qmc , and apply these models to the neutron star matter . the nuclear models
are calibrated to the bulk properties of nuclear matter at the saturation density , such as saturation density @xmath6 (= 0.17 @xmath7 ) , binding energy @xmath8 (= 16.0 mev ) , symmetry energy @xmath9 (= 32.5 mev ) and compression modulus @xmath10 ( 200 @xmath11 300 mev ) .
we take into account the effect of meson mass changes in the following ways .
first , we use the meson masses in free space for the nuclear models considered here as is usually done in most works .
secondly , we adopt the brown - rho ( br ) scaling law @xcite , @xmath12 and assume a density - dependent scaling function for the meson masses @xcite . the nuclear models that employ the scaling function are denoted by the abbreviation s " .
thirdly , in using the qmc type models we may treat the heavy mesons ( @xmath2 and @xmath1 ) as meson bags ( denoted by mb " ) composed of a quark and an anti - quark @xcite . in such cases , heavy mesons are treated in consistent with the description of the nucleon in the framework of the qmc .
after calibrating different nuclear models for nuclear matter we apply them to the neutron star matter and see how the reduction of meson masses in matter may influence the eos of the neutron star matter and the properties of the neutron star . in sec .
2 , we briefly describe the nuclear models considered in this work and fix the parameters of the models and show the resultant properties of the nuclear matter at the saturation density . in sec . 3 , we apply the various nuclear models fixed in sec . 2 to the neutron star matter and discuss the results and implications for the properties of the neutron star .
a summary is given in sec .
in this work we will consider 5 different models .
we refer to them as qhd , qhd - s ; mqmc , mqmc - s and mqmc - mb .
all the models have three coupling constants @xmath13 , @xmath14 , and @xmath15 ( or @xmath16 , @xmath17 , and @xmath18 ) for the interaction of nucleons ( or quarks ) with @xmath0- , @xmath1- and @xmath2-mesons , respectively . @xmath0- and
@xmath1-coupling constants are fitted to reproduce the given values of @xmath6 and @xmath8 , and the coupling to @xmath2-meson is determined by @xmath9 .
details of each model are explained in the following subsections .
the lagrangian for qhd in the mean field approximation reads @xmath19
\psi_n \nonumber \\ & & - \frac{1}{2 } m^2_\sigma\ , \bar{\sigma}^2 - \frac{1}{3 } m_n\ , b\ , ( g_{\sigma n}\ , \bar{\sigma})^3 - \frac{1}{4 } c\ , ( g_{\sigma n}\ , \bar{\sigma})^4 \nonumber \\ & & + \frac{1}{2 } m^2_\omega\ , \bar{\omega}^2_0 + \frac{1}{2 } m^2_\rho\ , \bar{b}^2_{03}. \label{eq : qhdlagrangian}\end{aligned}\ ] ] the masses @xmath20 , @xmath21 , @xmath22 and @xmath23 refer to the values in free space , 939 , 550 , 783 and 770 mev , respectively . the effective mass of the nucleon is defined as @xmath24 the meson fields are determined from the equations of motion in the mean field approximation ; @xmath25 where @xmath26 is 1/2 ( -1/2 ) for the proton ( neutron ) .
@xmath27 is the scalar density and @xmath28 is the fermi momentum of the nucleon at a given density .
cubic and quartic self - interaction terms of the @xmath0-meson are included in eq .
( [ eq : qhdlagrangian ] ) to produce reasonable @xmath10 and @xmath29 values at the saturation density .
the coupling constants fixed to produce the saturation properties and the resulting @xmath29 and @xmath10 are listed in tab .
[ tab : qhd ] .
we now consider the meson mass changes in matter and adopt the model proposed in ref .
@xcite where the br - scaling law is incorporated in the original qhd .
the model lagrangian reads @xmath30
\psi_n \nonumber \\ & & - \frac{1}{2 } m^{*2}_\sigma\ , \bar{\sigma}^2 + \frac{1}{2 } m^ { * 2}_{\omega}\ , \bar{\omega}^2_0 + \frac{1}{2 } m^ { * 2}_\rho\ , \bar{b}^2_{03}.\end{aligned}\ ] ] the br - scaling law is parametrized as @xcite @xmath31 the effective mass of the nucleon is defined as @xmath32 ( note the use of @xmath33 in place of @xmath20 of eq .
( [ eq : meff - qhd ] ) . )
coupling constants may change with nuclear matter density .
such density dependences of coupling constants as well as masses have been considered in previous @xcite and very recent @xcite works within the framework of relativistic mean field theory . in ref .
@xcite , @xmath14 is assumed to vary with density to satisfy @xmath34 at around the saturation density .
@xmath35 scaling is expressed as @xcite @xmath36 the parameters for the qhd - s model and the resulting properties at the saturation are summarized in tab .
[ tab : qhd ] .
one can see that both qhd and qhd - s models give @xmath10 values in the acceptable range .
@xmath37 turns out to be 0.781 , which is close to the values obtained by previous studies @xcite .
@xmath38 from qhd - s is smaller than that from qhd because of the scaling in eq .
( [ eq : brscaling ] ) .
.parameters and the saturation properties for the qhd and qhd - s models .
the subscript @xmath39 for @xmath37 refers to @xmath0 , @xmath1 and @xmath2 mesons . [
cols="^,^,^,^,^,^,^,^,^,^,^",options="header " , ] the parameters that can produce the saturation at @xmath6 are listed in tab .
[ tab : mqmc ] .
mqmc and mqmc - mb give us acceptable results for @xmath10 and @xmath29 , but the @xmath10 value for mqmc - s is as large as twice of usually accepted ranges of values . to get a value of @xmath10 within a reasonable range ,
we have tried many numerical searches over a wide range of parameter space , but no solution has been found so far .
we now apply the five nuclear models described in sect .
ii to the neutron star matter .
two basic assumptions of the neutron star matter are the charge neutrality and the @xmath40equilibrium .
if we assume that only the nucleons and light leptons exist in the neutron star , charge neutrality is expressed as @xmath41 where @xmath42 is the number density of particle @xmath43 . under @xmath40equilibrium
, the processes @xmath44 occur at the same rate
. the condition can be satisfied when the chemical potentials before and after the decay are the same .
the chemical potential of each particle reads @xmath45 where @xmath46 is the fermi momentum of the lepton @xmath47 ( @xmath48 ) .
the chemical equilibrium condition is expressed as @xmath49 the eos tells us the pressure ( @xmath50 ) as a function of energy density ( @xmath51 ) , which are the diagonal elements of the energy - momentum tensor . the energy density and
the pressure in the qhd model read @xmath52 @xmath53 the energy density and the pressure from the qhd - s model can be obtained from eq .
( [ eq : eneqhd ] ) and eq .
( [ eq : prsqhd ] ) , respectively , by removing the cubic and quartic self - interaction terms of the @xmath0-meson and replacing the free meson masses with the scaled ones given in eq .
( [ eq : mscaling ] ) .
however the pressure from the qhd - s model has additional terms that stem from the density dependent meson masses and coupling constants .
thermodynamic consistency of these terms is discussed in ref .
@xcite , and the explicit form is given as @xmath54 the energy density and the pressure of the mqmc - type models can be obtained by using @xmath55 , @xmath56 and @xmath57 together with proper replacement of in - medium meson masses in eqs .
( [ eq : eneqhd ] ) and ( [ eq : pressure - br ] ) . the resulting eos curves for each model
are shown in fig .
[ fig : eos ] . if the pressure exerted by nuclear repulsion is strong , the matter becomes more incompressible , which corresponds to a large compression modulus and a stiff eos .
the eos s from the qhd , mqmc and mqmc - mb models whose @xmath10 s are close to each other ( 310.8 , 285.5 and 324.1 mev , respectively ) show similar behaviors at the energy densities considered here .
the eos from the scaling models ( qhd - s and mqmc - s ) turns out to be stiffer than the eos from other models .
this behavior can be understood by observing that the repulsion from the @xmath58meson is augmented at high densities ( see eq .
( [ eq : omega ] ) with @xmath22 replaced by @xmath59 ) while the attraction caused by the @xmath60meson is not strong enough to cancel the repulsion . to illustrate this argument clearly , we plot in fig .
[ fig : mene ] the magnitude of the contributions to the pressure @xmath50 from @xmath61 , @xmath0 , @xmath1 and @xmath2 .
each contribution @xmath62 ( @xmath63 ) is defined as @xmath64 the total pressure is approximately @xmath65 the remaining terms such as the cubic and quartic self - interaction terms in qhd , extra terms for thermodynamic consistency and lepton contributions can be neglected since they do not determine the overall behavior of eos . in the case of the qhd
( see the left panel of fig .
[ fig : mene ] ) , @xmath66 is non - negligible compared to @xmath67 in the density region considered here . since @xmath0-meson contributes to the pressure negatively , @xmath66 reduces the total pressure substantially , which in turn leads to a sizable softening of the eos . on
the other hand , @xmath67 from qhd - s is about 3 times larger than that of qhd , but @xmath66 from qhd - s is more or less similar to that from the qhd .
thus the softening of the eos due to @xmath66 in qhd - s is relatively weak as the density becomes high , and consequently the eos from qhd - s is stiffer than that from qhd .
here , we need to remark that in fig . [
fig : eos ] the eos curves for mqmc - s and mqmc - mb terminate at a certain @xmath51 .
it is due to a breakdown of the scc at a certain density . in appendix
, we show the details of why the scc does not have a solution at some density .
the composition of the neutron star matter is represented by the number of particles divided by the total baryon number .
the number of each species of particles is determined by the charge neutrality and the chemical equilibrium in the @xmath68-decay .
muons can be created when the chemical equilibration condition between the electron and the muon , @xmath69 can be fulfilled .
given the values of @xmath70 , @xmath71 and @xmath72 for each @xmath2 , charge neutrality of eq .
( [ eq : charge0 ] ) , chemical equilibrium of eq .
( [ eq : chemeq ] ) , @xmath69 and @xmath73 determine the number of particles unambiguously .
the results are shown in fig .
[ fig : composition ] .
+ the compositions from the qhd and mqmc models are very similar to each other , but the number of the protons for the qhd - s or mqmc - s model is larger than that for the qhd or the mqmc , respectively . the ratio of the proton number to the neutron number , @xmath74 is especially important in determining the cooling mechanism of the neutron star . two major mechanisms of the cooling are emission of the neutrinos from the interior and radiation of the photons near the surface of the neutron star .
the simplest neutrino emission mechanism is the so - called direct urca ( du ) process , which is nothing but the @xmath68-decay as in eq .
( [ eq : durca ] ) , and it is known as the most efficient neutrino - emitting process in the interior of the neutron star . in order for this process to occur , however , the ratio @xmath74 should be larger than a critical value .
the proton numbers from qhd - s and mqmc - s models increase faster than those from the qhd and mqmc models as the density increases .
this implies that @xmath74 from qhd - s or mqmc - s can reach the critical value at lower densities and the @xmath68-decay can take place over a wider region in the interior of the neutron star than the qhd or mqmc models predict , causing a rapid cooling of the neutron star .
however , a recent work on the cooling of the neutron star @xcite indicates that too low a threshold density for the du process can result in an unrealistic cooling scenario .
du process can happen when both energy and momentum conservation is satisfied in eq .
( [ eq : durca ] ) . since neutrinos have energy of the thermal fluctuation around the fermi momentum of the nucleon or electron , its momentum can be neglected .
then the momentum conservation condition for du can be written as @xmath75 , where @xmath76 being the fermi momentum of particle @xmath77 " .
the threshold density @xmath78 at which du starts to emit neutrinos is about @xmath79 for qhd and mqmc , and is about @xmath80 for qhd - s and mqmc - s .
these threshold densities are too low to explain the observation of the neutron star temperature @xcite . in a recent work @xcite ,
the density dependent meson masses and coupling constants are considered within the framework of relativistic mean field theory .
it is shown that they obtained lower value of @xmath78 than that of our qhd model .
therefore the nuclear models we have considered in this work may need further refinement and adjustment of parameters to produce threshold densities high enough for the cooling process to be reasonable .
the mass and the radius of the neutron star can be obtained by integrating oppenheimer - volkov equation with a given eos . in fig .
[ fig : neutronstar ] the neutron star mass with our eos is plotted as a function of the central density @xmath81 ( left panel ) and of the radius ( right panel ) in units of solar mass @xmath82 .
the solid line is the result from the qhd - s , the dashed line from the qhd , and the dotted line from the mqmc .
qhd and mqmc give us very similar maximum mass @xmath83 with @xmath84 km and @xmath85 . on the contrary , due to the stiff eos
the qhd - s gives us the maximum mass of about 2.9@xmath82 with the radius @xmath86 km and the central density @xmath87 .
we could not extract the maximum mass of the neutron star for mqmc - s and mqmc - mb models since , as noted in the previous section , these models encounter a problem in solving the scc at a density before reaching the maximum mass .
recent observations of the neutron star mass from the radio pulsars give us the mass range of the neutron stars as @xmath88 @xcite .
the maximum mass of the qhd - s is much larger than the upper limit of the observed values , so one may want to exclude the qhd - s among the models for the neutron stars .
we may note , however , that more considerations are needed as follows .
first , the qhd - s produces the saturation properties successfully .
secondly , we have considered only the nucleon degrees in this work , but other degrees of freedom can come in such as hyperons or the phases like meson condensation and quark matter . in general , when these exotic degrees of freedom appear , they reduce the fermi momentum of nucleons through the @xmath68-decay of the nucleons to exotic states .
this results in a smaller contribution of @xmath89 to the pressure , and thus the eos becomes softer with exotic degrees of freedom than otherwise .
then the maximum mass of the neutron star becomes smaller .
we have confirmed this behaviour by doing some more calculations including hyperons in the qhd and mqmc .
the maximum mass of @xmath90 is obtained for qhd , and @xmath91 for mqmc .
it will be interesting to investigate how these maximum masses will change when medium effects on meson masses and coupling constants are taken into consideration .
it is noted , however , that the inclusion of hyperons can give us too low a mass limit of the neutron star @xcite , which is incompatible with the observations . in ref .
@xcite , a phase transition to quark matter is considered , and the maximum mass of the neutron star is obtained to lie in the range @xmath92 . a transition to quark matter is one possibility , and at the same time transitions to other exotica are also possible .
these possibilities should be considered for a better understanding .
in this work , we have investigated the properties of the neutron star matter by including the effects of meson mass changes in medium .
we have explored various models at the hadron ( qhd and qhd - s ) and at the quark ( mqmc , mqmc - s and mqmc - mb ) levels .
we have incorporated in - medium meson masses in two ways : one by assuming a simple function of scaling in both qhd and mqmc models , and the other by treating heavy mesons as mit bags within the framework of mqmc model .
the eos and the particle fractions for the neutron star matter are calculated .
scaling models show appreciable difference from the models with constant meson mass or meson bag models .
the eos s from the scaling models are stiffer than those from other models .
the scaling models show higher fractions of the protons than other models .
stiff eos s from the scaling models lead to heavy and large neutron stars whose maximum mass is about 2.9@xmath82 which is larger than the observed values and the values predicted by other models by one solar mass .
we may , however , note that the present work is just a first step toward the understanding of the relationship between the in - medium properties of physical quantities ( mass , coupling constant and etc . ) and the neutron star , and further studies need to be done .
this work is supported by korea research foundation grant ( krf-2002 - 042-c00014 ) .
in this section , we show a breakdown mechanism that occurs in solving the scc of the qmc model . since the scc for mqmc - s and mqmc - mb are involved , we employ the qmc model for the purpose of a simple illustration .
the scc of the qmc model can be obtained by putting @xmath93 in eq .
( [ eq : sccmqmc ] ) , which leads to @xmath94 the breakdown takes place when @xmath95 changes the sign from positive to negative at a certain density .
the numerator in the square bracket of eq .
( [ eq : scc2 ] ) causes the sign change . rearranging the numerator of eq .
( [ eq : scc2 ] ) , we obtain @xmath96 where nu means the numerator . to show roughly how the sign of @xmath95 changes , we make the following approximations . first , all the terms in eq .
( [ eq : neglect ] ) can be neglected since @xmath97 and @xmath98 are very small compared to the terms in eq .
( [ eq : keep ] ) in the density region we consider .
secondly , we use @xmath99 instead of the actual value @xmath100 . with these approximations
, we have @xmath101 ( \omega_q -1 ) .
\label{eq : upapp}\end{aligned}\ ] ] since @xmath102 is always positive , the sign change of eq .
( [ eq : upapp ] ) occurs at @xmath103 where @xmath104 .
since @xmath105 is always greater than 1 , the solution to eq .
( [ eq : solution ] ) can exist for @xmath106 . by taking @xmath107 as an independent variable , we plot in fig .
[ fig : omega ] @xmath105 of eq .
( [ eq : solution ] ) with the solid line .
( [ eq : solution ] ) represents possible points at which the sign change occurs .
the values of @xmath105 obtained from the numerical solution of eq .
( [ eq : omega - q ] ) and the scc of the qmc is shown in fig .
[ fig : omega ] by the dots . in the actual numerical calculation with the qmc model ( not mqmc model ) ,
of which we have not shown anything in this paper , solutions can be obtained up to @xmath108 at which @xmath109 , @xmath110 and @xmath111 . in fig .
[ fig : omega ] , the two curves meet at around @xmath112 ( @xmath113 ) , which corresponds to the density @xmath114 . if the small terms in eq .
( [ eq : neglect ] ) are included , the two curves will meet at around @xmath108 , which is the breakdown point of the qmc in the actual numerical calculation .
similar breakdown behaviors are observed numerically for the mqmc - s and mqmc - mb models .
for this reason , we could not calculate the eos at higher densities , and consequently could not extract the mass and the radius of the neutron stars for these models .
99 g. e. brown and m. rho , phys .
* 66 * , 2720 ( 1991 ) .
t. hatsuda and s. h. lee , phys .
c * 46 * , r34 ( 1992 ) .
k. saito , k. tsushima and a. w. thomas , phys .
c * 55 * , 2637 ( 1997 )
. g. agakichiev _
et al . _ ,
* b 422 * , 405 ( 1998 ) .
et al . _ ,
lett . * 86 * , 5019 ( 2001 ) . i. h. stairs , science * 304 * , 547 ( 2004 ) .
h. quaintrell _
et al . _ ,
. astrophys . * 401 * , 303 ( 2003 ) . c. song , g. e. brown , d .-
min and m. rho , phys .
c * 56 * , 2244 ( 1997 ) .
k. saito and a. w. thomas , phys .
c * 51 * , 2757 ( 1995 ) .
n. k. glendenning , f. weber and s. a. moszkowski , phys .
c * 45 * , 844 ( 1992 ) . c. fuchs , h. lenske and h. h. wolter , phys .
c * 52 * , 3043 ( 1995 ) .
s. typel , h. h. wolter , nucl . phys . * a 656 * , 331 ( 1999 ) .
b. liu , h. guo , m. di toro and v. greco , nucl - th/0409014 .
e. e. kolomeitsev and d. n. voskresensky , nucl - th/0410063 .
p. a. m. guichon , phys .
* b 200 * , 235 ( 1988 ) .
x. jin and b. k. jennings , phys . rev .
c * 54 * , 1427 ( 1996 ) .
c. song , d .-
min and m. rho , phys .
b 424 * , 226 ( 1998 ) .
d. blaschke , g. grigorian and d. n. voskresensky , a&a * 424 * , 979 ( 2004 ) .
s. e. thorsett and d. chakrabarty , apj .
* 512 * , 288 ( 1999 ) .
g. f. burgio , m. baldo , p. k. sahu and h .- j .
schulze , phys .
c * 66 * , 025802 ( 2002 ) . | we investigate the properties of the neutron star with relativistic mean field models .
we incorporate in the quantum hadrodynamics and in the quark - meson coupling models a possible reduction of meson masses in nuclear matter .
the equation of state for neutron star matter is obtained and is employed in oppenheimer - volkov equation to extract the maximum mass of the stable neutron star .
we find that the equation of state , the composition and the properties of the neutron stars are sensitive to the values of the meson masses in medium . |
one of the main difficulties for the grounding of a platform for quantum technologies is the effect of noise or `` decoherence '' on quantum states .
no physical system is ever truly closed due to interactions with its environment . as a result of such interactions ,
the quantum state of the system will approach classicality , thus ceasing to be of interest for quantum - empowered protocols @xcite . before being able to create useful quantum technologies , we need to understand these processes , and eventually control them .
an intriguing aspect of decoherence is that , in specific cases ( such as quantum stochastic resonance @xcite , to throw an example ) , weak decoherence mechanisms give rise to sizeable advantages in , say , the performance of some quantum protocols or the transport of excitations across a quantum medium . such counterintuitive effects are tightly linked to quantum interference phenomena : decoherence changes the way the wave function of a given system evolves in time , thus affecting the occurrence of constructive and destructive interference .
accidental " constructive effects may be induced , without spoiling the working principle of a given quantum process , for sufficiently weak decoherence mechanisms .
all this is particularly important ( and evident ) in the quantum walk framework @xcite , whose advantage in terms of the spreading rate of the position of a walker on a given path " may be magnified by small degrees of phase noise . in this paper
, we build on the already well established body of research into the behaviour of quantum walks when affected by decoherence , reported for instance in refs .
@xcite and surveyed in ref .
@xcite .
after quickly revisiting the paradigm of quantum walks , we proceed to discuss the protocol under investigation , introducing phase noise and addressing the performance of the scheme for various strengths of such mechanism .
a quantum walk is best described as the quantum analogue of the classical random walk . however , unlike the classical random walk , the evolution of a quantum walk is entirely deterministic .
quantum walks of course allow for superposition states of the walker , enabling them to exhibit interesting behaviours not shown by their classical counterparts . a comprehensive survey of quantum walks , covering both continuous and discrete time variants , and detailing the behaviours of quantum walks on various structures , can be found in ref
. @xcite . in this paper
, we focus on discrete time quantum walks . a discrete time quantum walk operates within the hilbert space @xmath1 , where @xmath2 known as the position space describes the position of the walker on a well - defined structure ( here we shall refer to this structure as the walk s _ terrain _ ) , and @xmath3 known as the coin space describes an additional degree of freedom affecting the evolution of the walk : this degree of freedom determines the walker s behaviour in the next time step . for the evolution of the walk
, we define two operators : the shift operator , @xmath4 , and the coin operator , @xmath5 .
the shift operator will `` move '' the walker on to a new part of its terrain , depending on the coin state .
for example , if the terrain of a walk is a graph , and the walker is on some vertex of the graph , the shift operator will move it along one of the vertex s edges to another vertex .
the coin operator is analogous to the flipping of a coin in a classical random walk , it will act on the coin space which in turn affects how the walk shall evolve in its terrain when the shift operator is applied .
we move the walker along by one step by applying the coin operator followed by the shift operator ; the state of the walker , starting in some initial state @xmath6 , is thus described after @xmath7 total steps by @xmath8^t { \left\vert\psi(0)\right\rangle},\ ] ] where @xmath9 is the identity operator on the position space .
in other words , we apply the coin operator and the shift operator @xmath7 times to the initial walker state . in order to provide a complete view of the main features of the walk protocol
, we now give some concrete examples of quantum walks . to represent the walk terrain , a line
, we shall use the set of integers .
the walker can be anywhere on the line , so we give @xmath2 the basis @xmath10 .
as previously described , each step of the walk involves a coin flip and a shift . in the walk on the line ,
the walker has a `` choice '' of two directions , left and right . in the classical random walk , the decision of which direction to walk in at each step is reached by flipping a fair coin .
likewise , in our quantum walk we shall use a coin space of degree two , viz .
@xmath3 is given the basis @xmath11 .
we decide to use the hadamard operator for our coin , as in ref .
this has the effect of putting our walker into a superposition of coin states and will allow for interference to occur during the course of the evolution of the walk . with regards to the walker s behaviour on the terrain , the classical random walk will move one step to the left or one step to the right depending on the most recent coin flip .
the same idea applies for the quantum walk .
we define the shift operator as @xmath12 ambainis _ et al . _
have shown in ref .
@xcite that the quantum walk on the line spreads out quadratically faster than the classical random walk on the line . in general , a graph @xmath13 is specified by fixing a set of vertices @xmath14 along with a set of edges @xmath15 connecting them .
@xmath16-regular graphs are graphs with @xmath16 edges attached to each vertex .
we affix a label @xmath17 to each end of each edge , as illustrated in fig .
[ labelled_graph ] * ( a)*. the walker will traverse the graph s vertices , moving along the edges , so we define the position space @xmath2 as having the basis @xmath18 .
-0.5cm*(a)*5.25cm*(b ) * layers , @xmath19.,title="fig : " ] layers , @xmath19.,title="fig : " ] at each time step , the walker has a fan - out of @xmath16 vertices to move to and we thus have to use an iso - dimensional coin space .
we now give @xmath3 the basis @xmath20 .
we then introduce the grover coin @xmath21 which , as described in ref .
@xcite , generalises the hadamard coin to hilbert spaces of dimension larger than 2 .
in order for @xmath22 to be unitary , the conditions @xmath23 and @xmath24 have to hold ( with @xmath25 ) .
the values of @xmath26 and @xmath27 can be changed to vary the behaviour of the walk on the graph .
we shall use a grover coin later on to perform the simulations at the core of our work . as for the shift operator
, this must take the walker along the appropriate edge to a new vertex , depending on the coin state .
again , we state that this idea is a generalisation of the walk on the line in which we give the walker a choice of @xmath16 directions at each step rather than 2 .
we define our shift operator as @xmath28 where @xmath29 and is labelled @xmath30 on @xmath31 s end , and @xmath32 is the label assigned to the destination node s end of the edge .
the goal of the glued trees ( gt ) algorithm for quantum search is the following : beginning from the left - most vertex of a given gt graph , traverse the graph and reach the right - most vertex , referred to as the target vertex .
childs _ et al . _
@xcite use this algorithm to show quantum walk search to be fundamentally more effective than classical random walk search by presenting a class of graphs ( the gt graphs ) that force classical random walks to make exponentially many queries to an oracle encoding the structure of the graph , but that are traversable by quantum walks with a polynomial number of queries to such an oracle . in order to study the robustness of the algorithm to the detrimental effects of decoherence
, we shall determine how effectively it achieves its goal when subjected to an increasing degree of phase damping noise .
for this reason , we will focus on the probability that the walker is on the target vertex at the end of the walk .
we thus consider gt graphs such as the one illustrated in fig .
[ labelled_graph ] * ( b ) * , i.e. consisting of @xmath33 layers before the gluing stage , and thus labelled as @xmath34 .
the continuous time quantum walk exploited by childs _
et al . _
@xcite can be reformulated as a discrete time one by means of a shift operator analogous to the one that has been previously described , and a grover coin with @xmath35 and @xmath36 as described in ref .
more explicitly @xmath37 we shall use this discrete time reformulation of the protocol to study its behaviour when affected by phase damping decoherence .
the gt graphs described in ref .
@xcite can straightforwardly be converted to @xmath38-regular graphs by adding two self loops ( one each to the left - most and right - most vertices ) , thus allowing us to use the discrete time walk model described in sec .
[ discreteongraph ] . to model decoherence
we use the kraus operators for a phase damping channel acting on the @xmath39-dimensional system embodied by the coin only .
this modifies the initial density matrix @xmath40 of a system @xcite as @xmath41 = \sum_{k}e_k(\tau ) \rho(0)e^{\dagger}_{k}(\tau),\ ] ] where we have introduced the kraus operators for phase damping @xcite @xmath42 with @xmath43 $ ] the strength of the phase damping @xcite . by taking the parameterisation @xmath44 with @xmath45 the probability rate of a phase error
, we can say that the effect of the channel is weak ( strong ) for @xmath46 ( @xmath47 ) .
we have introduced the orthonormal basis @xmath48 spanning the coin space only .
we assume an initial position - coin state decomposed as @xmath49 where we have introduced the orthonormal basis @xmath50 spanning the position space , and the position - coin density matrix elements @xmath51 .
evolves under the action of the phase damping channel on the coin space as @xmath52=\sum_{x , y}\sum^{d-1}_{l , l'=0}\rho_{x , l , y , l'}\eta^{(l - l')^2}{|x\rangle\langle y|}_{p}\otimes{|l\rangle\langle l'|}_c.\ ] ] our investigation consists of simulating a process such that , for each walk , we select a value of @xmath53 , apply the channel to the density matrix of the system after each time step of the walk , and evaluate the probability of the walk reaching a given vertex of the graph . in this way
, we compare the behaviour of the walks affected by phase damping ( @xmath54 ) to that which is found for an ideal walk ( corresponding to @xmath55 ) .
it is important to note that the probability depends on the initial coin state . in our study
we have considered walks starting in the initial coin state that maximises the performance of the ideal protocol ( such state also depends on the way in which we label the graph ; in particular , for the labelling scheme that we have mainly used , this corresponds to @xmath56 , with @xmath57 and @xmath58 ) .
however , we have tested different ways of labelling the graph ( and therefore different initial coin states that maximise the walk s performances ) and we have obtained results similar to those presented in the remainder of the paper . in order to grasp the temporal behaviour of the walk
, we consider the change in the probability distribution of the walker s position on the graph at time @xmath7 . in fig .
[ nodeco ] we illustrate such a probability distribution for an ideal 25-step walk on the gt graph @xmath59 .
the target vertex ( vertex number 253 ) is reached , with the highest probability , on step 16 ( as shown by the large peak in the line corresponding to step 16 ) .
we will now focus on the probability of the walker being on particular vertices on the graph and the overall vertex probability distribution for specific time steps . in ref .
@xcite , the effect of decoherence on the walk on the hypercube was studied quantitatively .
starting from one corner , a decohered walker will reach the opposite corner in less time than in the ideal walk , hence highlighting a counterintuitive beneficial effect of such an incoherent process .
our goal here is to address similar questions for the walk on a gt graph , as well as to investigate the limits of validity of the claim made in ref .
@xcite , where a lingering effect of a decohered walk on the target vertex was suggested .
we will thus look for the possibility that a decohered walk reaches the target vertex in less steps than the ideal one , and also attempt to determine whether it `` lingers '' on the target vertex for a longer time than in the ideal walk .
here we present and discuss the results of our analysis , showing the behaviour of walks affected by decoherence of variable magnitude .
all the results reported in this section , unless otherwise specified , were generated by simulating the previously discussed discrete time reformulation of the algorithm by childs _ et al . _
@xcite on the gt graph @xmath59 . in fig .
[ target_over_time ] we plot how the probability of the walker being on the target vertex changes over time for various decoherence magnitudes , ranging from the ideal walk with @xmath60 to a decohered walk with @xmath61 .
we see from this plot that the ideal walk takes 13 steps to achieve its goal of reaching the target vertex , where we say that the walker has reached some vertex @xmath31 on some step @xmath7 of the walk when there is a non - negligible probability of the walker being on @xmath31 on the @xmath62 time step .
the ideal walker first appears on the target vertex on step 13 with probability @xmath63 , then the probability comes to a peak on step 16 with a value @xmath64 before steadily decreasing .
we henceforth refer to the probability that a walker is on some vertex @xmath31 at a time step @xmath7 as @xmath31 s _ vertex probability _ at time step @xmath7 .
we see in fig .
[ target_over_time ] that , as @xmath53 decreases , the target vertex probability steadily decreases .
the peak on step 16 decreases at a faster rate than the target vertex probabilities associated with steps 13 , 14 , 15 and 17 .
some interesting behaviour can be observed on steps 21 through 27 of the walk .
decoherence magnitude @xmath65 slightly increases the target vertex probability ; in other words , the ideal walker is less likely to be on the target vertex on steps 21 through 27 than the decohered walkers .
this behaviour was reported in ref .
@xcite and can be seen more clearly in fig .
[ target_over_mag ] , in which we plot the effect of decoherence on the target vertex probability on various steps of the walk . in fig .
[ target_over_mag ] , in order to present this behaviour more clearly , we show a curve for step 22 .
we see , for step 22 , that the target vertex probability is higher for walks affected by decoherence of magnitude @xmath65 , with the probability peaking at @xmath66 .
decoherence has spread out the range of time steps over which the target vertex probability is large compared to the classical value . in line with the claims in ref .
@xcite , we have indeed found that weak decoherence increases the target vertex probability for an extended number of time steps .
however , when we consider the difference between the @xmath67 and @xmath60 walks on step 22 , we see that the difference between the two target probabilities is very small . in our simulations
we have determined that there is a difference of approximately @xmath68 .
as previously stated , in our simulations we have confirmed that this `` lingering '' effect lasts until step 27 .
[ target_over_mag ] , as previously discussed , is a plot of the effect of decoherence on the target vertex probability on various time steps .
the curve representing step 16 illustrates the effect of decoherence very well , and confirms the results reported in ref .
@xcite : as the decoherence magnitude @xmath53 decreases , the target vertex probability decreases exponentially . in other words ,
the algorithm by childs _
et al . _
@xcite becomes exponentially less effective at achieving its goal as @xmath53 decreases .
we shall now investigate the extent of the `` damage '' that decoherence has on the algorithm s effectiveness .
we concede that as long as the target vertex probability is higher than the other vertex probabilities then the decoherence has not had a particularly damaging effect on the effectiveness of the scheme . on the other hand , if phase damping decreased the peak associated with the target vertex below the other probability peaks then the algorithm would end up in a state involving a more probable `` false - positive '' than a `` positive '' when affected by decoherence we would say that this is a serious blow to any scheme s usefulness . in fig .
[ step16zoom ] we have plotted the entire graph s vertex probabilities on step 16 ( the step on which , as previously mentioned , the probability of the ideal walker being on the target vertex was the highest ) of walks with various decoherence magnitudes .
observe that the probability of the target vertex ( corresponding to vertex number 253 ) is always higher than the non - target vertex probabilities , regardless of the value of @xmath53 . with fig .
[ step13tophits ] we can get a clearer picture on how the probability peaks change in the walks affected by decoherence . in order to improve the legibility of the plots while maintaining the number of shown vertex probabilities at an acceptable level , we present only the vertex probabilities @xmath69 where @xmath70 is the target vertex probability . from fig .
[ step16zoom ] and fig .
[ step13tophits ] we can see that , in steps 13 to 16 of the walk , the peak representing the target vertex probability never drops below any of the other vertex probability peaks .
we find this significant and can conclude from it that the algorithm by childs _
et al . _
@xcite is not affected to the extent previously described : the walker is never on a non - target vertex with greater probability than the target vertex when allowed to run for sufficient time and phase damping decoherence does not change this fact . the authors of ref .
@xcite investigate the evolution of a discrete time quantum walk on the hypercube using a grover coin .
they plot the probability that the walker is on the target vertex ( they begin the walk on a corner of the hypercube and take the target vertex to be the vertex in the opposite corner of the hypercube ) for a number of time steps , in the same way we have done in this paper .
they observe that decoherence lowers the probability peaks of their plots in the same way that it does in our plots regarding the walk on the gt graphs , but they also observe that the `` troughs '' in their plots ( sections of the curve representing vertices with probability much lower than others ) become raised when decoherence is applied . this same effect can be observed in the walks that we simulated on the gt graphs and can be seen in the middle plot on fig .
[ step16zoom ] , where the vertices have their probabilities boosted slightly by the decoherence , with the lowest decoherence magnitude we investigated , @xmath61 , raising the probability by the greatest amount . for the sake of completeness
, we finally extend our investigation to different sizes of the gt graph .
we show in fig .
[ plotvslayers ] the behaviour of the target vertex probability against the number @xmath33 of layers before the gluing stage in the gt graph ( i.e. we have generated gt graphs @xmath34 , with @xmath33 between 4 and 8) .
we focus our interest on the step on which the probability of the walker being on the target vertex is the highest . also in this case we consider a range of decoherence magnitudes @xmath71 .
the range of values for @xmath33 has been chosen as a reasonable trade - off between the computational power required by the simulation and the readability of the plot . of layers before the gluing stage in the gt graph , for range of decoherence magnitudes ( @xmath72 from the top to the bottom , respectively ) . ]
in this paper we discussed a discrete time reformulation of the continuous time quantum walk algorithm described by childs _
. @xcite .
we simulated this discrete time quantum walk on gt graphs and applied phase damping to the coin system , in order to analyse the algorithm s resilience to decoherence .
we did this by studying how effectively it achieved its goal when affected by decoherence of various magnitudes : we investigated how decreasing the decoherence magnitude @xmath53 ( making the decoherence `` stronger '' ) lowered the probability of the walker being on the target vertex at the end of the walk .
we first simulated the walk with no decoherence to find how many steps it took to reach the target vertex ( we say that the walk has reached a vertex when the probability that the walker is on that vertex is non - negligible ) . we then included a range of decoherence magnitudes @xmath73 to see the extent of the drop in the target vertex probability at the end of the walk .
we observed that the ideal walk found the target vertex at time step 13 , with a probability peak at step 16 ( for a gt graph @xmath59 ) .
we noted that strengthening the decoherence ( decreasing @xmath53 ) lowered the target vertex probability on time steps 13 to 17 , but that the target vertex always had a higher probability than any other vertex , regardless of decoherence magnitude .
we also observed that a decoherence magnitude of @xmath65 boosted the target vertex probability very slightly above the ideal walk s ( @xmath60 ) vertex probability on steps 21 through 27 .
finally , we found that vertices with very low vertex probability had this boosted slightly by decoherence .
our results have touched on some unexplored features of the algorithm by childs _
et al . _
@xcite on gt graphs , in turn opening up new questions to address .
the first of such behaviours is the rate at which the target vertex probability on time step 16 decreases with @xmath53 : out of the time steps 13 - 17 , step 16 s target vertex probability decreases at the highest rate .
the second behaviour that deserves a deeper investigation in the future is observed in steps 21 through 27 of the walks studied in this paper . when compared to the target vertex probability for the ideal walk , we see a slight increase in the target vertex probability for @xmath65 walks . to give a more specific example , we see an increase of @xmath74 in the target vertex probability for @xmath67 on step 22 when we compare it to ideal walk . finally , the boosting of the troughs in fig .
[ step16zoom ] by decoherence , which appears to be a similar phenomenon to the boosting of the target vertex probability in steps 21 through 27 .
our results add more weight to the claims made in ref .
@xcite on decohered walks : the `` lingering '' effect is shown to be present , but only marginally relevant because the boost in probability caused by the decoherence for the steps after the target vertex probability drop ( step 18 ) is very small .
on the other hand , we have observed that decoherence does not cause any upset to the notion that , at the end of the walk , the walker should be on the target vertex with a higher probability than any other vertex .
this work has been supported by the uk epsrc through a career acceleration fellowship , a grant under the
new directions for research leader " initiative ( ep / g004579/1 ) , and the equipment grant ( ep / k029371/1 ) . jl thanks the centre for theoretical atomic , molecular , and optical physics for hospitality during the early stages of this work . | we study the behaviour of the glued trees algorithm described by childs _
et al . _ in [ stoc 03 , proc .
35@xmath0 acm symposium on theory of computing ( 2004 ) 59 ] under decoherence .
we consider a discrete time reformulation of the continuous time quantum walk protocol and apply a phase damping channel to the coin state , investigating the effect of such a mechanism on the probability of the walker appearing on the target vertex of the graph .
we pay particular attention to any potential advantage coming from the use of weak decoherence for the spreading of the walk across the glued trees graph .
quantum walks , quantum algorithms , decoherence |
this work is based on experiments performed at the swiss spallation neutron source sinq , paul scherrer institute , villigen , switzerland , and was partially supported by fundacin j. b. saubern and anpcyt under grant pict no .
mre was supported by the u. s. department of energy , basic energy sciences under award no .
de - fg02 - 10er46783 .
we are grateful to g. nieva for providing the @xmath0 sample , m. kenzelmann for financial aid and e. r. de waard for assistance with the sans experiment .
we modeled rocking curves ( rcs ) as the convolution of an _ intrinsic _ intensity distribution with a smearing _
resolution function_. the former corresponds to the actual @xmath81-dependence of the structure factor , while the latter accounts for imperfect collimation and wavelength spread . for simplicity
, the resolution function was assumed to be a gaussian kernel with standard deviation @xmath82 . since
rocking curves are already integrated within the detector plane , only the longitudinal dependence of the structure factor is relevant . when spatial correlations of the vl decay strongly with certain characteristic length scale , the envelope of the correlation function can be modeled with an exponential function .
so , the structure factor @xmath83 ( the fourier transform of the correlation function ) must be lorentzian - shaped , i.e. @xmath84 where @xmath65 is the corresponding correlation length .
therefore , we took the intrinsic distribution to be the cauchy - lorentz distribution centred at the bragg angle @xmath85 , multiplied by a proportionality constant @xmath86 .
then , its half width at half maximum ( hwhm ) @xmath87 is directly related to @xmath65 by @xmath88 where @xmath89 is the magnitude of one reciprocal vector of the vl .
the resulting intensity profile is @xmath90 in the limiting cases of long correlation lengths @xmath91 , ( i.e. when @xmath65 is beyond our resolution ) the intrinsic distribution can be replaced by a dirac delta , so @xmath92 .
\label{longcorr}\ ] ] and the experimental rcs can be just modeled by the gaussian resolution distribution .
neutron count data were preprocessed by the software grasp @xcite .
detector counts were normalized by monitor counts .
then , after background subtraction , intensity integrated over a fixed sector of the detector was calculated for each image , and resulting rocking curves @xmath93 , @xmath94 , were exported , together with their error estimates @xmath95 .
statistical analysis was carried out using the software octave @xcite .
it consisted of three stages : 1 . preliminary analysis , 2 .
resolution estimation , 3 . intrinsic widths estimation
. non linear optimization often requires set of initial parameters close enough to the optimal in order to work properly .
the aim of this stage was to obtain suitable estimates of the parameters for each rocking curve individually , to be used in later stages .
we did this by fitting our convolution model individually to each rocking curve .
because of the convolution , the parameters @xmath58 and @xmath87 are expected to be highly correlated , making it difficult to fit both simultaneously . to overcome this issue , a rough prior theoretical estimate of @xmath58 , based on the experimental geometry @xcite , was calculated .
this resulted in @xmath96 .
a modified likelihood function @xmath97 was then constructed and parameters were obtained from minimizing @xmath98 the preliminary analysis showed that configurations shaken at low temperatures ( examples are shown with red and violet symbols in figure [ fig2](c ) ) were resolution limited ( @xmath99 ) .
therefore , we used the long - correlation - length approximation ( eq . [ longcorr ] ) to model these rocking curves . in turn , it enabled us to make an independent estimation of the resolution based solely on neutron data .
the resolution profile likelihood @xmath100 was constructed by maximizing the global likelihood function ( i. e. , including all rocking curves ) , constrained to @xmath58 taking a series of fixed values .
we obtained @xmath101 , which is consistent with our prior estimate .
this also revealed that the dependence of the estimated @xmath102 on the resolution width was common to all the curves .
examples of the resulting fitting curves are shown in full red and violet lines in figure [ fig2](c ) .
next , we determined the minimum intrinsic width @xmath102 which could be eventually distinguishable from the zero - width approximation .
this was done by generating randomized rocking curves of different widths with the convolution model , using the above estimation of @xmath58 .
similar number of points and relative errors as in our experimental data were considered . after fitting gaussian functions to generated data
, a threshold of @xmath103 out of 200 replicates resulting in non - overlapping error bars ( one standard deviation ) between @xmath104 and the fitted width was applied .
this resulted in @xmath105 , or @xmath106 @xmath107 ( vertical dashed line in figure [ fig4](b ) ) .
lastly , the profile likelihood of @xmath102 for each rocking curve was constructed by fitting only @xmath86 and @xmath85 , while letting @xmath108 be the estimated above ( fixed ) .
this procedure reduced marginal errors in @xmath102 by ignoring the covariance with @xmath108 . since a change in @xmath108 would only shift all estimated @xmath102 together ,
this approach is better suited for comparing different configurations instead of finding the exact correlation length .
the resulting likelihood ratios @xmath64 ( curves in figure [ fig4](b ) ) where then obtained through the relationship between @xmath65 and @xmath102 ( eq . [ corrl ] ) . as long as the elastic approximation is valid , the in - plane characteristic lengths @xmath109 in equilibrium configurations can be estimated from the elastic ratio @xcite @xmath110 where @xmath77 and @xmath76 are the shear and tilt elastic vl moduli respectively . in the local approximation @xcite ,
@xmath111 where @xmath112 is the london penetration depth and @xmath41 is the mean intervortex distance . in our case @xcite , @xmath113 k@xmath114 , so the corresponding in - plane characteristic lengths in the bg phase are expected to be @xmath115 . on the other hand ,
the fact that all the bps are resolution limited , indicates that all the measured vl configurations are not in the amorphous limit .
far from the amorphous limit , the elastic approximation is still roughly valid for equilibrium configurations with dislocations @xcite .
however , as mentioned in the main text , the partially disordered vls are metastable configurations at @xmath116 , originated from a dynamic reorganization in the vicinity of the odt , and therefore a quantitative calculation of @xmath117 from @xmath65 is not available .
a rough estimation using the elastic limit at @xmath118 ( where the local approximation for the elastic constant is valid ) gives @xmath119 and @xmath120 for configurations with intermediate disorder .
a better estimate can be made in the framework of a defective vl @xcite at high temperature ( where the dynamic reorganization occurs ) . in this model
each correlated volume @xmath121 is limited by dislocations spaced in @xmath8 and @xmath65 in the transverse and longitudinal directions respectively .
to calculate @xmath122 we follow eqs .
( 9 ) and ( 10 ) of ref .
@xcite ( where the notation is @xmath123 , @xmath124 ) and a non local approximation for the tilt modulus @xcite @xmath125 where the penetration depth @xmath126 .
the resulting @xmath127 are @xmath128-independent in this limit and the in - plane correlation lengths corresponding to the experimental @xmath129 result @xmath130 and @xmath131 for intermediate disordered configurations , slightly larger to that obtained in the elastic limits .
although these values are not quantitatively exact , these estimates could be useful to design future sans experiments .
equation ( [ intdist ] ) is properly normalized so that the parameter @xmath75 represents the total integrated intensity .
the maximum - likelihood estimate for @xmath75 and the corresponding variance were obtained along with the other parameters discussed above for each rc . in order to compare integrated intensities corresponding to different configurations , the dispersion in the intensities fitted to rcs measured after different histories were statistically assessed by means of @xmath132 difference tests of nested models .
the results of the tests are presented indicating the statistic @xmath132 , the degrees of freedom @xmath133 , and the corresponding @xmath134-value ( i. e. , the probability of getting a @xmath132 greater than the observed one solely due to randomness , under the assumption that a common mean is shared ) .
first we note that the set of all the values of @xmath75 had far more dispersion ( relative to their errors ) than expected for a single mean value model ( @xmath135 , @xmath136 , @xmath137 ) . in order to assess whether this excessive dispersion was connected to the different histories ,
the dispersion within sets of @xmath75 corresponding to each process was evaluated .
all intensities corresponding to fc configurations were indistinguishable ( @xmath138 , @xmath139 , @xmath140 ) and intensities corresponding to intermediate - order configurations were only marginally distinguishable ( @xmath141 , @xmath139 , @xmath142 ) . moreover , merging the sets of intensities for fc and intermediate - order configurations did not increase the dispersion enough to reject a single mean intensity for both ( @xmath143 , @xmath144 , @xmath145 ) . on the contrary , bg configurations presented a higher degree of variability , beyond their estimated variance ( @xmath146 , @xmath147 , @xmath148 ) and the set of intensities for bg configurations could not be merged with fc and intermediate - order configurations without increasing the dispersion significantly ( @xmath149 , @xmath144 , @xmath137 ) .
these results suggest that all rcs for fc and intermediate - order configurations share essentially the same value for the parameter @xmath75 .
instead , fits to rcs for bg configurations yielded statistically distinguishable values for the total integrated intensity , which averaged about @xmath150 below the mean @xmath75 for fc and intermediate - order configurations .
as mentioned in the main text , an underestimation of @xmath75 is likely to result from the long - tail decay of the structure factor in the bg phase , further supporting the observed structural differences .
( full symbols , example in panel ( b ) ) and @xmath151 ( open symbols , example in panel ( c ) ) .
below @xmath152 a shaking field orders the vl .
the hatched region indicates the transitional region , where we claim that the vl dynamically reorganizes in configurations with intermediate degree of disorder . above @xmath153
there is a narrow region where disordered vls are stable .
no history effects are observed beyond @xmath154 , which matches the spinodal line described in ref .
@xcite ( dark yellow line ) . ] in this supplementary section , the vortex phase diagram corresponding to our samples is presented and some features are compared with those reported in the literature . the magnetic field and temperature ranges used in our experiments are also included . as the sample used in the sans experiment ( in the following ,
sample a ) was too large to be measured in the 7-t mpms xl quantum design or in our cryostats , another crystal from the same source ( sample b , @xmath155 @xmath156 ) and a smaller sample obtained by cutting an edge of sample a ( sample c , @xmath157 @xmath156 ) were characterized instead . beyond the expected differences due to the larger area of sample b ,
results obtained from both samples are identical . in figure
[ fig : s1 ] , the phase diagram obtained from non - linear ac susceptibility curves @xmath151 ( open symbols ) and @xmath49 ( full symbols ) measured in sample b is shown in panel ( a ) , including the transitional region ( hatched area in fig .
[ fig : s1 ] ) .
the dark yellow line is the spinodal , lying above the transitional region , that demarcates the limit of stability with respect to fluctuations toward the thermodynamically stable state , extracted from ref .
@xcite and rescaled using the critical parameters of our sample @xmath158 k and @xmath159 koe ( estimated by extrapolating the line @xmath160 to @xmath161 ) .
this line matches the abrupt change in slope observed in both @xmath49 ( @xmath162 in panel ( b ) ) and @xmath163 ( @xmath164 in panel ( c ) ) , above which no history effects are observed .
the fact that our experiments are performed below the spinodal line indicates that fluctuations are not dominant in the underlying physics .
however , due to the close vicinity , some small influence can not be discarded .
the dotted horizontal line indicates the magnetic field applied in our sans experiment .
the full bulk penetration of the shaking fields , with the consequent vortex displacements in distances larger than @xmath41 , is essential to justify our claim of a bulk dynamic reordering .
the modification of the bulk vl structure factor observed in the sans experiment is the main direct evidence of such a reordering . in this supplementary section , we present some numerical estimations , together with additional experimental results obtained by magnetization and non - linear ac susceptibility measurements , that provide a consistent scenario regarding this crucial issue .
a simple calculation provides a rough estimation of the mean vortex displacement associated with the shaking field penetration . under full ac field penetration ,
the maximum variation of the vl parameter @xmath41 during a cycle of the shaking field is @xmath165 in our sans experiments , @xmath166 oe and @xmath167 koe , so @xmath168 nm . therefore , near the centre of the sample , vortices only need to move short distances @xmath169 to accommodate to the change in the density .
however , the propagation of the density change throughout the sample implies that the mean displacement increases with the distance @xmath170 from the centre as @xmath171 .
consequently , in our experiments we expect @xmath172 for @xmath173 @xmath174 m , reaching @xmath175 as large as several microns near the edge of the sample ( @xmath176 mm ) . in conclusion , during the shaking procedure , most vortices move long distances @xmath177 , allowing a dynamic reorganization . on the other hand , vortices that enter and exit the sample crossing its edge during the shaking procedure
are present only in a thin micrometric annulus near the sample boundary ( less than 2% of the sample area ) and may only marginally affect the average bulk structure .
the claim that the small ac field @xmath178 moe , applied during the linear ac susceptibility measurements , does not modify the vl configuration , i.e. the technique is non - invasive , has been extensively discussed in previous publications @xcite : the linear non - dissipative campbell regime @xcite holds when each vortex performs small harmonic oscillations around a minimum of the effective pinning potential well .
these small oscillations propagate inside the sample up to the campbell penetration depth , generally smaller than the sample dimension . the same reasons outlined above ( replacing @xmath166 oe by @xmath178 moe ) lead to the conclusion that during the ac measurement even outmost vortices move typically less than @xmath179 nm @xmath180 , where @xmath181 is the coherence length , not modifying the vl configuration .
[ [ linear - and - non - linear - ac - regimes - ac - field - penetration - and - dissipation ] ] linear and non - linear ac regimes : ac field penetration and dissipation ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ the full penetration of the shaking field was previously confirmed by measuring the non - linear ac response in sample c , and results are shown is figure [ fig : s2 ] ( color curves ) : below @xmath182 , the normalized ac susceptibility is amplitude dependent , with a maximum out of phase component @xmath183 , characteristic of the ac bean critical response @xcite . at @xmath184 k , an ac field with amplitude @xmath185 oe
( the maximum available in the mpms ) and frequency @xmath186 khz fully penetrates the sample ( @xmath187 ) ; an even higher penetration is expected for the shaking field @xmath166 oe .
notice that , despite the smaller volume of sample c , because samples a and c have exactly the same thickness @xmath70 , similar non linear ac field penetration will hold @xcite . in the same figure , the linear ( i.e. amplitude independent ) ac susceptibility , measured in sample a with the sans setup , is also included .
below @xmath182 , the dissipative out - of - phase component @xmath188 , as expected in the linear campbell regime @xcite . the conclusions of the above paragraph are also consistent with results obtained by magnetization measurements .
figure [ fig : s3 ] shows examples of magnetization curves at low fields @xmath189 measured at different temperatures .
the shape of the @xmath42 loops is characteristic of low pinning thin superconductors in a transverse geometry @xcite , where geometrical barriers inhibit the magnetic field entrance up to a field @xmath190 higher than the effective lower critical field @xmath191 ( @xmath192 in our experimental geometry ) . at fields @xmath193 , geometrical barriers do not play any significant role but a small irreversible magnetization @xmath194 , arising from weak bulk pinning , persists .
it can be observed in the inset of figure [ fig : s3 ] that @xmath195 decays quickly with increasing field . at @xmath167 koe ( the magnetic field applied in our sans experiments ) , the low irreversibility ( @xmath196 oe ) is consistent with the full bulk penetration of the shaking field @xmath166 oe . koe . the sharp zero - field transition is also included in light grey symbols . in the linear response ( @xmath178 moe ) @xmath197 below @xmath182 , consistently with a non - invasive campbell regime .
the highest - amplitude ac field fully penetrates the sample , for @xmath184 k. arrows in the linear curves show the direction of temperature variations .
non linear curves were recorded in cooling procedures . ] at low fields , for several temperatures .
the shape is characteristic of low pinning thin superconductors in a transverse geometry ; geometrical barriers inhibit the magnetic field entry up to @xmath198 ( the effective lower critical field ) .
inset : at @xmath193 , a small irreversible magnetization @xmath194 , arising from weak bulk pinning , persists . at @xmath167 k oe , @xmath199 oe , consistently with the full bulk penetration of @xmath166 oe . ]
p. l. gammel , u. yaron , a. p. ramirez , d. j. bishop , a. m. chang , r. ruel , l. n. pfeiffer , e. bucher , g. danna , d. a. huse , k. mortensen , m. r. eskildsen , and p. h. kes , phys .
* 80 * , 833 ( 1998 ) . | we report structural evidence of dynamic reorganization in vortex matter in clean @xmath0 by joint small angle neutron scattering and ac - susceptibility measurements . the application of oscillatory forces in a transitional region near the order - disorder transition results in robust bulk vortex lattice configurations with an _ intermediate _ degree of disorder .
these dynamically - originated configurations correlate with _ intermediate _ pinning responses previously observed , resolving a long standing debate regarding the origin of such responses . in a wide variety of complex systems ,
competing interactions promote an order - disorder transition ( odt ) .
ordered phases are characterized by spatial correlations decaying weakly over distances larger than the relevant system scale , whereas disordered configurations are characterized by correlation lengths @xmath1 of the order of the mean inter - particle distance @xmath2 .
configurations with intermediate degrees of disorder , strong enough to affect the system response , but still with @xmath3 , have received attention recently@xcite .
vortex matter in superconductors provides an ideal model system for the experimental study of the topic @xcite : vortex - vortex interactions favoring an ordered vortex lattice ( vl ) compete with both thermal fluctuations and pinning interactions that tend to disorder the system . in very low - pinning superconductors , such as clean @xmath0 single crystals ,
most of the vortex field - temperature phase diagram is properly described by an ordered dislocation - free bragg glass ( bg ) phase @xcite , which undergoes an odt near the normal - superconductor transition @xcite . in practice ,
when a superconductor is cooled from the normal state in an external magnetic field ( field - cooled / fc ) , energy barriers may trap the vl in highly disordered metastable configurations @xcite .
even so , high transport current densities @xcite or large oscillatory `` shaking '' magnetic fields @xcite may anneal the vl into the ordered low - temperature bg . once in the bg , as temperature or magnetic field is increased the vl softens and accommodates to the random pinning potential more efficiently . eventually , vortex entanglement and the proliferation of vl dislocations increase the effective pinning , producing a sudden rise of the critical current @xmath4 known as the peak effect ( pe ) @xcite , which is the fingerprint of the odt in vortex matter .
while both the high - pinning disordered phase and the low - pinning ordered phase , above and below the odt , are widely accepted in the literature , intermediate responses have been ascribed to surface contamination and partial reordering by the probing transport current @xcite . however , in the narrow transitional region adjacent to the pe , for which a multidomain phase has been theoretically proposed @xcite , non - invasive techniques have shown that pinning can be _ partially decreased _ or even _ increased _ by applying dc currents @xcite or ac magnetic fields @xcite .
the resulting intermediate responses are highly reproducible and independent of the previous history .
are these in - between responses originated from bulk vl configurations with an intermediate degree of disorder ?
to address this question , we performed an experiment in a clean @xmath0 single crystal , combining small angle neutron scattering ( sans ) with _ in - situ _ linear ac susceptibility measurements . in sans
imaging of a perfect triangular vl , first order bragg peaks ( bp ) are expected to appear as six symmetric sharp spots , matching the reciprocal lattice vectors .
this still holds for an ideal bg , where elastic disorder leads to slow decay of the correlations @xcite , although recent simulations @xcite have shown that elastic bg prefracturing would result in a slight broadening of the bps .
conversely , when disorder is dominated by dislocations , spatial correlations decay exponentially beyond a characteristic length @xmath1 associated with the scale at which dislocations appear @xcite . in such cases bps have a finite intrinsic width proportional to @xmath5 .
experimentally observed bps are further smeared by a convolution kernel due to finite resolution @xcite .
hence , the degree of disorder can be determined from the spread of the scattered intensity , as long as @xmath6 , a resolution - dependent bound . in a 3d vortex system ,
the anisotropy of the elastic constants gives rise to very different characteristic lengths along the direction of the flux lines ( @xmath7 ) and in the transverse plane ( @xmath8 ) .
because of the poorer in - plane resolution , in this experiment only the longitudinal @xmath7 was determined by measuring the integrated intensity on the plane of a single bp as the sample was rocked through the bragg condition ( rocking curve / rc ) .
the experiment was conducted at the sans - ii beam line of the paul scherrer institute s sinq facility , using the ma11 cryomagnet .
the geometry of the experiment is sketched in figure [ fig1 ] .
all sans measurements were performed at @xmath9 .
neutron wavelength was chosen to be @xmath10 ( @xmath11 ) and the applied dc field was @xmath12 .
the incident beam was collimated , resulting in a beam divergence @xmath13 .
the angular resolution along the rocking angle was estimated to be @xmath14 deg .
a statistical analysis showed that curves arising from @xmath15 were distinguishable from the resolution kernel .
rcs wider than our resolution were fitted by the convolution of the gaussian resolution with a lorentzian distribution corresponding to an intrinsic @xmath16 ( see appendix ) . a large ( @xmath17 )
clean @xmath0 single crystal was used for the experiment .
the sample is part of a batch grown in bell labs as described in ref .
@xcite , with @xmath18 ( @xmath19 of the linear ac susceptibility transition at @xmath20 ) .
the phase diagram for various crystals has been built using a 7-t mpms xl ( quantum design ) and compared with those published in the literature @xcite , showing non significant differences ( see appendix ) . and
the ac field @xmath21 are aligned with the @xmath22 axis of the single crystal , in the @xmath23 plane .
as the whole system is _ rocked _ around the @xmath24 axis , the distribution of the scattered intensity @xmath25 is measured by the position sensitive detector , which spans the @xmath26 plane , for each angle @xmath27 between the field direction and incoming neutron beam wave vector @xmath28 .
the maximum intensity is expected at @xmath29 , for the corresponding reciprocal lattice vector @xmath30 . ] since the goal was to explore the connection between structure and dynamics , a non - invasive measurement of pinning was required .
this was achieved by installing a sample holder with mutual inductance coils inside the sans cryomagnet . by adding a small ac magnetic field @xmath31 ( @xmath32 ) , vortices were forced to perform small oscillations inside their pinning wells , without modifying their spatial configuration .
these oscillations propagate through the sample due to the vortex - vortex repulsion , with a characteristic penetration depth @xmath33 that is related , in a mean field approximation @xcite , to the effective pinning potential @xcite .
the linearity of the response , as well as the low dissipation level characteristic of the linear campbell response @xcite had been previously verified .
then @xmath34 and thus the effective pinning can be assessed through the in - phase component of the linear ac susceptibility @xmath35 by means of a mutual inductance technique .
lower values of @xmath35 are associated with stronger pinning , saturating at perfect ac shielding ( normalized to @xmath36 ) , whereas @xmath35 vanishes under complete ac penetration when @xmath37 sample size . to probe oscillatory dynamic effects ,
a _ shaking _ procedure was applied at certain temperatures : here @xmath35 measurements were paused and a larger sinusoidal field with amplitude @xmath38 oe , and @xmath39 khz was applied for 1000 cycles before returning to the smaller @xmath40 to resume @xmath35 measurements .
the shaking field amplitude was sufficient to induce vortex displacements larger than @xmath41 throughout the sample .
the complete penetration of the shaking field into the sample had been previously confirmed by measuring the non - linear ac response and checking consistency with the irreversible dc magnetization determined from @xmath42 loops ( see appendix ) .
measured down to @xmath9 during a fc cooling procedure ( black squares ) and after shaking the vl at two different temperatures @xmath43 ( red circles and violet triangles ) below the onset of the pe ( @xmath44 ) .
shaking temperatures are indicated by vertical lines and arrows indicate temperature evolution .
reversibility is observed in the warming from @xmath45 of the vl shaken at @xmath46 ( dark yellow triangles ) .
the area enclosed by the rectangular box is magnified in fig . [ fig3](a ) .
( b ) rc after fc ; the black curve is the fit of a lorentzian peak convoluted by the resolution kernel .
( c ) rcs after shaking the vl at @xmath46 ( red circles ) and @xmath47 ( violet triangles ) ; the curves are gaussian fits with fwhm equal to the instrumental resolution @xmath48 deg . ]
figure [ fig2](a ) shows a comparison between @xmath49 measured during a fc procedure and after shaking the vl at two different temperatures @xmath43 below the onset of the pe at @xmath50 .
there is an unambiguous rise in @xmath34 after shaking , representing a reduction of the effective pinning .
the agreement between both cooling @xmath49 curves after shaking is consistent with an equilibration procedure below @xmath44 .
reversibility is observed in the warming curve , with @xmath49 increasing with @xmath51 up to @xmath44 , after which it drops due to the spontaneous pinning enhancement at the pe .
in contrast , the absence of pe in the linear response during the cooling @xcite is likely due to the vl remaining trapped in a metastable highly pinned state .
this explanation is consistent with the fc rc ( figure [ fig2](b ) ) , which is clearly wider than the resolution kernel , in agreement with other experiments @xcite .
conversely , the two rcs recorded after shaking the vl ( figure [ fig2](c ) ) show resolution - limited bps ( @xmath52 ) .
the fact that bps are well defined ( unlike the annulus observed for amorphous vls ) indicates that orientational order is preserved in all the cases .
the broadening of the fc configuration rc suggests a dislocated vl with broken positional order beyond a characteristic volume with longitudinal dimension @xmath53 . there is a clear correlation between the ordering of the vl and the decrease of the effective pinning . for different thermal and dynamic histories ( labeled ( b)-(e ) in panel ( a ) ) and the corresponding detector intensity distributions for one vl bp at the bragg condition ( panels ( b)-(e ) ) .
arrows in ( a ) indicate temperature evolution and the hatched area shows the transitional region .
data from figure [ fig2](a ) ( black and red lines ) is included for comparison .
the application of a shaking field at @xmath54 results in a unique response ( full blue and green symbols ) independent of whether the initial state was a disordered fc vl ( open blue symbols ) or an ordered vl ( open green symbols ) .
both intensity distributions ( c ) and ( d ) indicate partial disorder as compared to the fc intensity distribution ( b ) .
simply warming an ordered vl ( magenta squares ) up to @xmath55 shows moderate pinning enhancement during cooling , but the corresponding bp ( e ) is still sharp and intense .
the red boxes in ( b)-(e ) enclose the detector area included in the rcs in figure [ fig4 ] . ] while shaking the vl at any @xmath56 leads to an ordered configuration , the consequences are strikingly different when the shaking field is applied in the narrow transitional region above @xmath44 .
figure [ fig3](a ) shows the evolution of @xmath49 through a series of processes near the transitional region ( hatched area ) and figure [ fig4](a ) displays the corresponding rcs .
a strongly pinned fc vl was cooled to @xmath57 and shaken , partially reducing the effective pinning .
the vl was cooled to @xmath45 and a narrow rc , still wider than @xmath58 was measured , indicating a bulk _
partial ordering_. in contrast , applying the same shaking burst at the same @xmath59 ( @xmath60 ) to an initially ordered vl had the opposite effect : the shaking field partially increased the effective pinning to a similar level and _ disordered _ the vl .
the rc was broader than @xmath58 , but narrower than for the maximally disordered fc lattice . because the response after shaking is strongly dependent on the temperature when @xmath61 , the process was carefully repeated within @xmath62 , obtaining the same qualitative results . to rule out that the _ partial disordering _ was of thermal origin ,
a warming - cooling cycle up to @xmath59 but without applying the shaking burst was carried out for a replicated ordered vl . although the cooling @xmath49 branch indicates higher pinning than the warming branch , it is lower than for vls shaken above @xmath44 .
moreover , if there was any associated disorder , it was minor and below our sans resolution . therefore , the `` intermediate '' degree of disorder found before is only attributable to the shaking field applied at @xmath59 .
although the above results are qualitatively evident by simply comparing the corresponding intensity distribution on the position sensitive detector at @xmath63 ( figure [ fig3](b)-(e ) ) , we evaluated the statistical significance of the observed differences ( see appendix ) .
figure [ fig4](b ) shows the profile likelihood ratio @xmath64 resulting from the fit of each experimental rc to one expected from a configuration with an intrinsic width given by @xmath65 convoluted by the experimental resolution .
correlation lengths obtained after shaking the vl at @xmath59 were @xmath66 at the @xmath67 confidence level .
thus , we state that shaking the vl in the transitional region results in bulk intrinsic configurations with intermediate degree of disorder .
( b ) ) , after shaking at @xmath59 a fc ( blue ) and an ordered ( green ) vl , and after warming an ordered vl to @xmath68 without shaking at this temperature ( magenta ) .
curves are fits of lorentzian peaks convoluted by the resolution .
( b ) calculated profile likelihood ratio @xmath69 as a function of @xmath65 up to the sample thickness @xmath70 for each experimental rc .
symbols above the curves indicate the maximum ( @xmath71 ) and error bars , the @xmath72 confidence interval . for the ordered vl , @xmath73 and
is beyond our resolution .
the profile likelihood of replicate rcs of the fc ( black dashed line ) and the one shaken at @xmath74 ( green dashed line ) are reproducible within the error bars . ] the total integrated intensity @xmath75 ( the area under the rc ) for the different configurations should remain roughly constant .
even so , the long - tail decay of the structure factor in the bg phase may lead to an underestimation of @xmath75 @xcite .
indeed , for all fc configurations and for the configurations obtained after shaking the vl in the transitional region we observe a similar @xmath75 .
however , for the lattice shaken in the bg the integrated intensity is slightly , but statistically significant , smaller ( see appendix ) .
these features support a qualitative difference in the positional order decay in the elastic bg configurations and in those obtained after shaking the vl in the transitional region , where disorder is probably mediated by dislocations . in the elastic bg
, the in - plane correlation length @xmath8 can be derived from @xmath7 through the tilt and shear elastic constants @xmath76 and @xmath77 @xcite . from @xmath78 ,
a lower limit of @xmath79 can be estimated for the in plane size of the dislocation - free regions corresponding to the ordered configurations , which is consistent with results obtained in other works in the bg phase @xcite .
how @xmath8 relates to @xmath7 is not as straightforward for metastable vl configurations with dislocations . in theoretical works
@xcite , @xmath7 and @xmath8 are obtained for equilibrium vl configurations from the densities of screw and edge dislocations that minimize the free energy ; far enough from the amorphous limit ( i.e. , @xmath80 ) dislocation densities adapt in such a way as to satisfy the elastic ratio . moreover , numerical simulations @xcite and experimental results @xcite show that @xmath7 monotonically increases with the in - plane ordering from the plastic to the elastic regime . in our case ,
the partially disordered vls are metastable configurations ( at @xmath45 ) which originated from a dynamic reorganization in the vicinity of the odt . in this region
the elastic constants are dispersive and may also be affected by fluctuations .
therefore , a precise quantitative estimation of @xmath8 from @xmath7 is not available , but a monotonic correlation is still expected ( see appendix for a quantitative estimate ) .
such a correlation would be compatible with an intermediate density of dislocations in partially disordered vls .
our results show a clear connection between the linear ac response , related to the effective pinning , and the bulk spatial correlation of the vl .
configurations having correlation lengths much larger than the lattice parameter , but small enough to affect the global response , were observed after shaking the vl in the proposed transitional region .
these observations support the existence of such region and represent the first evidence of a dynamic reordering , driven by oscillatory forces , which results in robust bulk configurations with intermediate degrees of disorder .
a similar behavior was obtained in numerical simulations @xcite , where configurations with intermediate dislocation densities are accessible from stationary fluctuating dynamic states , as those proposed for colloids @xcite .
the interplay between these spatial configurations and their associated non - linear dynamics remains a fertile ground for further research . |
the purpose of this article is to summarize the experimental status of spin physics based on the material presented at the third circum - pan - pacific symposium on high energy spin physics .
a total of 12 experimental talks covering recent results and future plans at desy , jlab , slac , and rhic - spin were presented at this symposium .
these talks successfully convey the sense of excitement in this field through the presentation of many interesting recent experimental results , as well as exciting prospects for future experiments .
several excellent review articles on high - energy spin physics have been published recently@xcite .
many new experimental results were presented at this symposium for the first time , indicating that spin physics has become one of the most active areas of research in nuclear and particle physics .
the physics topics covered by the 12 experimental talks can be grouped into the following categories : * polarized structure functions + @xmath0 , @xmath1 , @xmath2 , their integrals , and the gdh sum rule .
* quark and gluon helicity distributions + @xmath3 and @xmath4 . *
transversity distributions + @xmath5 * generalized parton distributions + deeply virtual compton scattering ( dvcs ) . * other related topics + longitudinal spin transfer in @xmath6 production , exclusive meson productions , @xmath7 at large @xmath8 , etc . in the following sections
, we will discuss the recent progress in and the future prospects for these various areas of researches .
following the discovery of the spin - crisis " in the late 80s , extensive efforts have been devoted to accurate measurements of the spin - dependent structure functions @xmath10 and @xmath11 .
a series of experiments at slac ( e142 , e143 , e154 , e155 , e155x ) , at cern ( emc , smc ) , and at desy ( hermes ) , have measured @xmath12 and @xmath13 over a broad range of @xmath8 and @xmath14 .
scaling violation of @xmath12 is now clearly observed , as shown in fig .
[ fig : f1pscaling ] .
the @xmath14 evolution of @xmath10 is strikingly similar to that of the spin - averaged structure function @xmath15 .
in particular , the e155 collaboration recently found@xcite that the ratio @xmath16 of all existing data can be parameterized as @xmath17 the values of @xmath18 , @xmath19 for the proton ( neutron ) , are consistent with zero and indicate that @xmath20 and @xmath21 have very similar @xmath14 dependences .
the extensive data on @xmath0 allow accurate determinations of the integrals @xmath22 for the proton and the neutron , as well as @xmath23 .
table 1 lists the results of these integrals from recent nlo analysis of existing data by the e155@xcite and smc@xcite collaborations . while the values of @xmath24 and @xmath25 are different from the predictions of ellis and jaffe@xcite who assumed su(3 ) flavor symmetry and an unpolarized strange sea , the data are in good agreement with the prediction of the bjorken sum rule@xcite .
as the bjorken sum rule is now quite well tested , it is not surprising that the experimental activities on @xmath26 are winding down .
nevertheless , there are other interesting aspects of @xmath26 worthy of further studies .
first , the behavior of @xmath26 at low @xmath8 , @xmath27 , is not yet known .
perturbative qcd calculations , based on fits to existing data , give predictions for @xmath26 at low @xmath8 very different from regge and other models@xcite .
the largest uncertainty on the @xmath24 determination also comes from the unmeasured low-@xmath8 region .
future polarized @xmath28 collider is required for exploring the low-@xmath8 region@xcite .
another interesting topics is the behavior of @xmath26 integral at low @xmath14 , to be discussed next .
how does @xmath29 evolve as @xmath30 ?
this question is closely related to the gerasimov - drell - hearn ( gdh ) sum rule@xcite : @xmath31 \frac{d\nu}{\nu } = -\frac{2\pi^2\alpha}{m^2 } \kappa^2 .
\label{eq : gdh}\ ] ] the gdh sum rule , based on general physics principles ( causality , unitarity , lorentz and gauge invariances ) and dispersion relation , relates the total absorption cross sections of circularly polarized photons on longitudinally polarized nucleons to the static properties of the nucleons . in eq .
[ eq : gdh ] , @xmath32 and @xmath33 are the photo - nucleon absorption cross sections of total helicity of @xmath34 and @xmath35 , @xmath36 is the photon energy and @xmath37 is the pion production threshold , @xmath38 is the nucleon mass and @xmath39 is the nucleon anomalous magnetic moment .
the gdh sum rule predictions are -205 @xmath40b and -233 @xmath40b for the proton ( @xmath41 ) and neutron ( @xmath42 ) , respectively .
a first measurement@xcite of the helicity dependence of photoabsorption cross section on the proton was recently carried out at mami ( mainz ) and the contribution to the gdh sum was found@xcite to be @xmath43 @xmath40b for the photon energy range @xmath44 mev .
using the unitary isobar model@xcite and a regge model@xcite to estimate the contributions from unmeasured energy regions , the integral is found@xcite to be @xmath45b , consistent with the gdh sum rule .
measurements at higher energies are either underway or being prepared at elsa , jlab , and slac . the gdh integral in eq .
[ eq : gdh ] can be generalized from real photon absorption to virtual photon absorption with non - zero @xmath14 : @xmath46 \frac{d\nu}{\nu } \nonumber \\ & = & \frac{8\pi^2\alpha}{m } \int_{0}^{x_0 } \frac{g_1(x , q^2 ) - \gamma^2g_2(x , q^2)}{k } \frac{dx}{x } , \label{eq : ggdh}\end{aligned}\ ] ] where @xmath47 is the flux factor of the virtual photon , @xmath48 and @xmath49 .
the generalized gdh integral connects the helicity structures of the nucleons measured in high - energy electron dis to those in low - energy photo - absorption at the resonance region . assuming the validity of the burkhardt - cottingham sum rule@xcite , @xmath50
, it follows that for @xmath51 eq .
[ eq : ggdh ] becomes @xmath52 eq .
[ eq : ggdh1 ] shows that the @xmath14-dependence of the generalized gdh integral is directly related to the @xmath14-dependence of @xmath9 .
@xmath24 is known to be positive at high @xmath14 and the gdh sum rule ( eq . [ eq : gdh ] ) predicts @xmath53 at @xmath54 with a negative slope for @xmath55 , therefore , @xmath56 must become negative at low @xmath14 .
the gdh integrals at low @xmath14 have recently been measured in several jlab experiments . as reported by griffioen@xcite in this symposium , inclusive double spin asymmetry
have been measured at hall - b using the clas spectrometer with polarized electron beams of 2.5 and 4.2 gev scattered off longitudinally polarized @xmath57 and @xmath58 targets .
the kinematical region covered in this experiment corresponds to @xmath59 gev@xmath60 and @xmath61 gev .
preliminary results on @xmath56 extracted from this experiment are shown in fig .
[ fig : gammap ] .
these data indeed show that @xmath24 changes sign around @xmath62 gev@xmath60 .
the origin of the sign - change can be attributed to the competition between @xmath63 and higher nucleon resonances . at the lowest @xmath14
, the @xmath63 has a dominant negative contribution to @xmath24 .
however , at larger @xmath14 , higher mass nucleon resonances take over to have a net positive @xmath24 . as shown in fig .
[ fig : gammap ] , the strong @xmath14 dependence of @xmath56 is well reproduced by the calculation of burkert and ioffe@xcite . as reported by chen@xcite in this symposium
, an extensive spin physics program has been underway using the jlab hall - a spectrometers .
in particular , neutron spin - dependent structure functions , @xmath11 and @xmath64 , have been measured using an intense polarized electron beam on either longitudinally or transversely polarized @xmath65 target .
preliminary results@xcite on the generalized gdh integral for neutron and @xmath65 are shown in fig .
[ fig : ngdh ] .
in contrast to the proton case , the strong negative contribution to the gdh integral from the @xmath63 resonance now dominates the entire measured @xmath14 range .
the data appear to approach the gdh sum rule value at @xmath66 . however , the data are also consistent with the prediction of drechsel et al.@xcite for a rapid variation at very small @xmath14 and a departure from the gdh sum rule .
future experiment@xcite at hall - a will extend the @xmath65 measurement down to @xmath67 gev@xmath60 in order to map out the low @xmath14 behavior of the neutron generalized gdh integral .
the recent studies at jlab of the spin - averaged and spin - dependent structure functions at low @xmath14 region have shed new light on the subject of quark - hadron duality .
thirty years ago , bloom and gilman@xcite noticed that the structure functions obtained from deep - inelastic scattering experiments , where the substructures of the nucleon are probed , are very similar to the averaged structure functions measured at lower energy , where effects of nucleon resonances dominate .
this surprising similarity between the resonance electroproduction and the deep inelastic scattering suggests a common origin for these two phenomena , called local duality .
recently , high precision data@xcite from jlab have verified the quark - hadron duality for spin - averaged scattering on proton and deuteron targets . for @xmath14 as low as 0.5 gev@xmath60 , the resonance data are within 10% of the dis results . when the mean @xmath68 curve from the resonance data is plotted as a function of the nachtmann variable , @xmath69
, it resembles the @xmath70 structure function obtained in neutrino scattering experiments .
since @xmath70 is a measure of the valence quark distributions , the similarity between @xmath70 and the mean @xmath68 suggests that the @xmath68 structure function at low @xmath14 originates from valence quarks only .
it is of much interest to extend the study of quark - hadron duality to spin - dependent structure functions . in this symposium
, griffioen@xcite presented the preliminary result from clas showing that quark - hadron duality is also observed in @xmath12 .
a comparison of @xmath71 between the resonance data and the gehrmann and stirling@xcite parameterization of the dis data shows a good agreement for @xmath72 ( the valence quark region ) . at @xmath73 ,
the resonance data appear to deviate from the dis data .
this trend is reminiscent of what was observed for the spin - averaged structure functions@xcite .
additional information on the quark - hadron duality is also expected from the jlab hall - a measurement@xcite of @xmath74 for neutrons .
unlike the spin structure function @xmath20 , which has a clear interpretation in the quark - parton model ( qpm ) , the @xmath75 structure function is sensitive to higher - twist quark - gluon correlation effect and is not readily interpreted in qpm .
the @xmath75 structure function probes both the transverse and the longitudinal parton distributions in the nucleons . using the operator product expansion ( ope ) technique@xcite in qcd
, @xmath75 can be expressed in terms of three components@xcite : a leading twist-2 part @xmath76 originating from the same set of operators that contribute to @xmath20 , another twist-2 structure function @xmath77 depicting quark transverse polarization , and a twist-3 part from quark - gluon interactions @xmath78 , @xmath79 where @xmath80 and @xmath38 are quark and nucleon masses .
the twist-2 contribution , @xmath81 , is related to @xmath20 via@xcite @xmath82 the contribution from the transversity distribution @xmath77 is suppressed by the @xmath83 term and can be neglected .
hence , the difference between @xmath75 and @xmath81 , @xmath84 , will isolate the twist-3 contribution .
the moments of @xmath20 and @xmath75 can be derived from ope : @xmath85 where @xmath86 and @xmath87 are the twist-2 and twist-3 matrix elements of the renormalized operators , respectively . the twist-3 matrix elements @xmath87
can then be evaluated using @xmath88 a primary goal for measuring @xmath75 is therefore to determine the twist-3 contribution which reflects quark - gluon correlation effects .
a series of slac experiments , e142 , e143 , e154 , e155 and e155x , have measured @xmath1 for protons and neutrons . in this symposium ,
bosted@xcite reported the near - final results of @xmath89 and @xmath90 from e155x .
the @xmath8-dependence of all slac @xmath89 data is reasonably well described by the twist-2 component @xmath81 . however , the data also allow non - zero twist-3 contributions .
in particular , the twist-3 matrix elements , @xmath91 and @xmath92 , evaluated using eq .
[ eq : g2d ] , are shown to be small but non - zero@xcite .
meziani@xcite discussed in this symposium a proposal to measure @xmath93 with high accuracy at the proposed 12 gev cebaf upgrade using a polarized @xmath65 target and a large acceptance spectrometer .
this could provide a definitive result on the twist-3 content of the nucleon .
the new @xmath94 data from e155x also allow evaluations of the integrals @xmath95 in order to check the burkhardt - cottingham sum rule . in this symposium
, bosted@xcite reported that for @xmath96 at @xmath97 gev@xmath60 , the integral was found to be @xmath98 for proton and @xmath99 for deuteron .
the apparent disagreement between the proton result and the burkhardt - cottingham sum rule could be due to the contribution of the unmeasured small @xmath8 region .
another sum rule by efremov , leader and teryaev ( elt)@xcite , derived with the assumption of isospin symmetry of the sea - quark distributions , gives @xmath100 dx = 0 .
\label{eq : etd}\ ] ] the elt sum rule is much less sensitive to the small-@xmath8 uncertainty due to the factor @xmath8 in the integrand .
the most recent value of this integral for the proton is found@xcite to be @xmath101 at @xmath14 = 2.5 gev@xmath60 , consistent with the prediction of the sum rule . the recent jlab hall - a data@xcite on @xmath93 could provide a further check of this sum rule for the neutron .
the @xmath0 data obtained in inclusive dis have been analyzed@xcite in the framework of nlo qcd to extract information on @xmath102 , @xmath103 , and @xmath104 , where @xmath105 and @xmath106 correspond to the flavor - singlet and flavor - nonsinglet quark polarization , and @xmath107 is the gluon polarization .
the inclusive dis data , however , do not allow a detailed flavor decomposition of the nucleon spin . in particular , the contributions from valence and sea quarks are not separated . to overcome this limitation ,
the smc and hermes experiments have advocated the use of semi - inclusive deep inelastic scattering ( sidis ) , where the leading hadrons accompanying the dis process are also detected .
the flavor of the struck quark is expected to be reflected by the flavor of the produced hadrons .
hence the hadrons provide a tag " on the flavor of the struck quark .
this technique was first applied by the smc@xcite collaboration to determine @xmath108 , @xmath109 , and @xmath110 .
later , the hermes collaboration used this method to measure the @xmath111 flavor asymmetry of the unpolarized nucleon sea and obtained@xcite a result consistent with that from a completely different approach@xcite using the drell - yan process .
this agreement suggests that the sidis is indeed a valid tool for studying flavor decomposition .
extensive sidis data have been collected@xcite by the hermes collaboration on polarized gas targets of @xmath112 , @xmath113 , and @xmath65 .
shibata@xcite and bernreuther@xcite presented in this symposium the preliminary results of @xmath108 , @xmath109 , and @xmath114 based on an analysis of the hermes data collected up to spring 99 . as shown in fig [ fig : sidis1 ] , the hermes data have improved accuracy over the smc data .
both the smc and the hermes results are in very good agreement with the parameterization of gehrmann and stirling@xcite . in order to increase the statistical significance , the constraint @xmath115 has been imposed in the analysis shown in fig .
[ fig : sidis1 ] . with the large sample of polarized @xmath116 sidis data collected during 1999 and 2000 in conjunction with an operational rich detector for @xmath117 identification , the hermes collaboration is now analyzing their sidis data without the contraint of eq .
[ eq : sidis1 ] .
the identification of kaons with the rich detector would help isolating the @xmath118 component . the anticipated statistical accuracy for @xmath108 , @xmath109 , @xmath114 , @xmath119 , and @xmath120 in this 5-parameter analysis is shown in fig . [ fig : sidis2 ] .
the hermes analysis could lead to exciting first results on the flavor structure of the sea - quark polarizations .
several interesting issues , such as the large flavor asymmetry between the @xmath114 and @xmath119 predicted@xcite in the chiral - quark - soliton model as well as the negative strange - quark polarization expected from the analysis of @xmath20 , could be addressed .
the result of the five - parameter analysis is expected@xcite to be available soon .
another promising technique for measuring sea - quark polarization is @xmath121-boson production@xcite at rhic .
the longitudinal single - spin asymmetry for @xmath121 production in @xmath122 can be written in leading order as @xmath123 where @xmath124 are the bjorken-@xmath8 of the colliding quarks and antiquarks .
for @xmath125 , eq .
[ eq : wprod1 ] becomes @xmath126 and @xmath127 gives a direct measure of sea - quark polarization . for @xmath128 ,
one obtains @xmath129 and the valence quark polarization is probed . in this symposium
, kiryluk@xcite discussed the plan to measure @xmath130 at the star collaboration .
the phenix collaboration is capable of measuring the @xmath131 decays as well , as discussed by saito@xcite .
the rhic @xmath121-production and the phenix sidis measurements are clearly complementary tools for determining polarized quark and antiquark distributions .
analysis of existing @xmath20 data showed that only @xmath132 30% of the nucleon s spin is carried by quarks .
this suggests that gluons could have a large polarization .
a nlo analysis of @xmath0 by the smc collaboration@xcite showed that @xmath133 is positive , albeit with a large uncertainty .
global fits@xcite to existing spin - dependent structure functions can also accommodate very different parameterizations of @xmath133 , again showing that gluon polarization is poorly known .
a direct and precise determination of @xmath133 remains one of the most important goals of spin physics .
although inclusive dis does not directly probe the gluon distribution , certain semi - inclusive dis processes are sensitive to @xmath134 .
the hermes collaboration has reported a measurement@xcite of @xmath135 at @xmath136 using semi - inclusive hadron - pair production .
lepto- and photo- open - charm production has also been proposed at cern@xcite and at slac@xcite for @xmath135 measurements .
polarized proton - proton collision at rhic offers a great opportunity for studying gluon polarization . in this symposium , kiryluk@xcite discussed the proposed double longitudinal spin asymmetry , @xmath137 , measurement for @xmath138 at star and liu@xcite discussed the heavy - quark production at phenix .
other processes sensitive to @xmath134 are the inclusive prompt photon production , jet production , and high-@xmath139 hadron production .
although no new experimental results on @xmath133 were presented in this symposium , it is clear that a wealth of new results will be forthcoming .
in addition to the unpolarized and polarized quark distributions , @xmath140 and @xmath3 , a third quark distribution , called transversity , is the remaining twist-2 distribution yet to be measured .
this helicity - flip quark distribution , @xmath5 , can be described in qpm as the net transverse polarization of quarks in a transversely polarized nucleon .
the corresponding structure function is given by @xmath141 due to the chiral - odd nature of the transversity distribution , it can not be measured in inclusive dis experiments . in order to measure @xmath3 ,
an additional chiral - odd object is required .
for example , the double spin asymmetry , @xmath142 , for drell - yan cross section in transversely polarized @xmath143 collision , is sensitive to transversity since @xmath144 .
such a measurement could be carried out at rhic@xcite , although the anticipated effect is small , on the order of @xmath145% .
several other methods for measuring transversity have been proposed for semi - inclusive dis . in particular , collins suggested@xcite that a chiral - odd fragmentation function in conjunction with the chiral - odd transversity distribution would lead to an observable single - spin azimuthal asymmetry in semi - inclusive pion production .
an analysis of the jet structure in @xmath146 jets decay suggested that the collins function has a sizable magnitude@xcite .
as reported by schnell@xcite in this symposium , the hermes collaboration has recently measured@xcite single - spin azimuthal asymmetry for charged and neutral pion electroproduction .
using unpolarized positron beam on a longitudinally polarized hydrogen target , the cross section was found to have a sin@xmath147 dependence , where @xmath147 is the azimuthal angle between the pion and the @xmath148 scattering plane .
this single - spin - asymmetries ( ssa ) can be expressed as the analyzing power in the sin@xmath147 moment , and the result is shown in fig .
[ fig : trans1 ] for @xmath149 , @xmath150 , and @xmath151 as a function of the pion fractional energy @xmath152 , the bjorken @xmath8 , and the pion transverse momentum @xmath153 .
the sin@xmath147 moment for an unpolarized ( u ) positron scattered off a longitudinally ( l ) polarized target contains two main contributions @xmath154 where @xmath155 and @xmath156 are the longitudinal and transverse components of the target spin orientation with respect to the virtual photon direction . for the hermes experiment with a longitudinally polarized target ,
the transverse component is nonzero with a mean value of @xmath157 .
the observed azimuthal asymmetry could be a combined effect of the @xmath158 transversity and the twist-3 @xmath159 distribution .
figure [ fig : trans1 ] shows that a model calculation@xcite reproduces the @xmath152 , @xmath8 , and @xmath153 dependences of the @xmath151 asymmetry quite well .
the striking difference between the @xmath149 and @xmath150 analyzing power suggests that the collins fragmentation function is sizable only when the flavor of the truck quark is present in the final hadron .
if the azimuthal asymmetry observed by hermes is indeed caused by the @xmath158 transversity , a much larger asymmetry is expected for a transversely polarized target .
an earlier smc measurement had limited statistics and was inconclusive@xcite .
the hermes collaboration plans to measure@xcite the shape of @xmath160 ( and @xmath161 ) using a transversely polarized proton target in 2002 - 03 .
a proposal to measure @xmath162 using a transversely polarized deuterium target has also been discussed@xcite .
there has been intense theoretical and experimental activities in recent years on the subject of generalized parton distribution ( gpd ) . in the bjorken
scaling regime , exclusive leptoproduction reactions can be factorized into a hard - scattering part calculable in qcd , and a non - perturbative part parameterized by the gpds .
the gpd takes into account dynamical correlations between partons with different momenta .
in addition to the dependence on @xmath14 and @xmath8 , the gpd also depends on two more parameters , the skewedness @xmath163 and the momentum transfer to the baryon , @xmath164 .
of particular interest is the connection between gpd and the nucleon s orbital angular momentum@xcite .
the deeply virtual compton scattering ( dvcs ) , in which an energetic photon is produced in the reaction @xmath165 , is most suitable for studying gpd . unlike the exclusive meson productions
, dvcs avoids the complication associated with mesons in the final state and can be cleanly interpreted in terms of gpds .
an important experimental challenge , however , is to separate the relatively rare dvcs events from the abundant electromagnetic bethe - heitler ( bh ) background .
significant progress has been made recently , and several experiments at hera and jlab have reported observation of the dvcs events . from the collision of 800 gev protons with 27.5 gev positrons ,
both the zeus@xcite and the h1@xcite collaborations at desy observed an excess of @xmath166 events in a kinematic region where the bh cross section is largely suppressed .
the excess events were attributed to the dvcs process and the h1 collaboration further determined@xcite the dvcs cross section over the kinematic range @xmath167 gev@xmath60 , @xmath168 gev , and @xmath169 gev@xmath60 .
( the @xmath8 range covered is roughly @xmath170 for @xmath171 gev . ) at lower c.m .
energies , the hermes@xcite and the clas@xcite collaborations observed the interference between the dvcs and the bh processes , which manifests itself as a pronounced azimuthal asymmetry correlated with the beam helicity . in this symposium ,
bianchi@xcite reported the hermes measurement@xcite shown in fig .
[ fig : dvcs1 ] .
the hermes result is in nice agreement with the clas result@xcite , also shown in fig .
[ fig : dvcs1 ] .
note that there exists several differences between these two measurements .
first , the beam energy for the clas experiment ( 4.25 gev ) is lower than for the hermes experiment ( 27.6 gev ) .
second , the clas experiment detected the electron and proton , while the hermes experiment measured the positron and the photon in the final state .
finally , a polarized electron beam was used for clas instead of the polarized positron beam for hermes .
the qualitative agreement between these two experiments is reassuring .
it is interesting to note that the clas data showed an opposite sign for the azimuthal asymmetry relative to the hermes data ( in fig .
[ fig : dvcs1 ] the offsets of the azimuthal @xmath147 angle are different for the two experiments ) .
this is to be expected since the interference term is proportional to the sign of the lepton charge .
as discussed by bianchi@xcite in this symposium , another observable sensitive to the interference between the dvcs and the bh processes is the azimuthal asymmetry between unpolarized @xmath172 and @xmath173 beams .
in contrast to the beam spin asymmetry ( bsa ) which is sensitive to the imaginary part of the dvcs amplitudes , the beam charge asymmetry ( bca ) is probing the real part of the dvcs amplitudes@xcite .
analysis of the hermes @xmath173 data in 98 - 99 and the @xmath172 data in 99 - 00 is underway , and the first result on the bca is expected soon@xcite .
schnell@xcite described the recent progress at hermes on semi - inclusive @xmath6 production .
the hermes collaboration has measured the longitudinal spin transfer from the polarized electron beam to the @xmath6 particle in the final state , as well as the transverse @xmath6 polarization using unpolarized beam .
the main interest for the longitudinal spin transfer measurement is to deduce information on the polarized quark distribution in the @xmath6 . assuming @xmath174-quark dominance as well as helicity conservation in the @xmath175 fragmentation process
, the hermes measurement of the longitudinal spin transfer can reveal the role of @xmath174 quark for @xmath6 spin .
earlier measurements@xcite of @xmath6 polarization from @xmath176 decays are sensitive to @xmath177 quark polarization in the @xmath6 .
the nomad collaboration has also measured@xcite @xmath6 and @xmath178 polarization in neutrino dis .
the new hermes measurement@xcite covers a kinematic range ( roughly @xmath179 and @xmath180 ) much broader than previous dis experiments@xcite .
preliminary result on the average spin transfer in this kinematic region is @xmath181 .
the @xmath182 dependence is consistent with the previous nomad result@xcite .
the predictions of several model calculations@xcite for a rising spin transfer with increasing @xmath152 are not observed in the data , although the statistical accuracy is limited .
a new detector was installed at hermes to enhance the acceptance of @xmath6 detection for future measurements@xcite .
new results on exclusive meson production from hermes and clas collaborations have been presented in this symposium .
the current interest in hard exclusive processes is largely due to the fact that qcd factorization was proved to be valid for exclusive meson production with longitudinal virtual photons@xcite .
such factorization allowed new means to extract the unpolarized and polarized gpd .
in particular , unpolarized gpds can be measured with exclusive vector meson production , while polarized gdps can be probed via exclusive pseudoscalar meson production .
bianchi@xcite presented the preliminary hermes results on the longitudinal component of the exclusive @xmath183 and @xmath147 meson production cross sections .
the data are in good agreement with a calculation@xcite based on gpd . for exclusive pseudoscalar meson production ,
the hermes collaboration measured the single spin azimuthal asymmetry in the reaction @xmath184 .
the asymmetry is found@xcite to be very large ( @xmath185 ) and has a sign opposite to that of inclusive @xmath149 production @xcite .
the clas collaboration recently reported@xcite the first measurement of double spin asymmetry in the @xmath186 reaction .
this observable is sensitive to the contributions from various resonances , and the data indicate the dominance of the helicity-1/2 contribution in the second and third resonance regions .
a new measurement@xcite of the @xmath7 spin asymmetry has been carried out at hall - a in jlab using polarized @xmath65 target .
the value of @xmath7 as @xmath187 is sensitive to the underlying models describing valence quark dynamics in the nucleon .
isospin and su(6 ) symmetries predict @xmath188 as @xmath187 , while pqcd approaches@xcite predict @xmath189 as @xmath8 approaches 1 .
preliminary jlab result@xcite for @xmath7 in the @xmath190 range show that @xmath7 turns positive for large @xmath8 .
a definitive measurement of @xmath7 with high accuracy at very large @xmath8 has been considered for the 12 gev cebaf upgrade@xcite .
there has been an enormous progress in various areas of spin physics experiments since the last circum - pan - pacific spin symposium in 1999 .
an incomplete list of the major progress would include the following : * first jlab measurements on spin structure functions have provided new information on @xmath191 and the generalized gdh sum at the resonance region . *
observation of the azimuthal single spin asymmetry at hermes in semi - inclusive pion production holds a great promise for measuring transversity in the near future .
* observation of the dvcs process at hermes and jlab has generated much interest and could develop into an extensive program for measuring generalized parton distributions . * various spin observables in exclusive meson electroproductions
are being measured at hermes and jlab . *
commissioning of the rhic - spin has been successfully carried out@xcite .
many new results are anticipated in the near future : * first results from rhic - spin . * first results on @xmath192 and @xmath193 ) from the 5-parameter analysis of the hermes sidis data . *
first measurement of the transversity ( @xmath194 ) at hermes using transversely polarized targets . * precise data on dvcs from hermes and jlab .
* @xmath133 from slac , compass , hermes , and rhic - spin . * additional low-@xmath14 data from jlab . * and much more ....
i would like to thank professor bo - qiang ma for inviting me to attend this very interesting and successful symposium .
0 b. w. filippone and x. ji , hep - ph/0101224 .
g. bunce , n. saito , j. soffer , and w. vogelsang , _ ann .
nucl . part .
sci . _ * 50 * , 525 ( 2000 ) .
r. l. jaffe , hep - ph/0101224 .
m. anselmino , hep - ph/0107093 . n. bianchi and r. jakob , hep - ph/0108078 .
n. c. r. makins , talk presented at dis2000 .
the e155 collaboration , p. l. anthony _ et al .
. lett . _ * b493 * , 19 ( 2000 ) .
the smc collaboration , b. adeva _
_ , _ phys .
rev . _ * d58 * , 112001 ( 1998 ) .
the smc collaboration , b. adeva _
_ , _ phys .
rev . _ * d58 * , 112002 ( 1998 ) .
j. ellis and r. jaffe , _ phys .
_ * d9 * , 1444 ( 1974 ) ; * d10 * , 1669 ( 1974 ) .
j. d. bjorken , _ phys .
rev . _ * 148 * , 1467 ( 1966 ) ; * d1 * , 1376 ( 1970 ) . f. e. close and r. g. roberts , _ phys .
lett . _ * b336 * , 257 ( 1994 ) . j. kwiecinski and b. ziaja _ phys
. rev . _ * d60 * , 054004 ( 1999 ) . v. w. hughes and a. deshpande , _ nucl .
suppl . _ * 79 * , 579 ( 1999 ) , hep - ex/9906006 .
s. b. gerasimov , _ sov .
* 2 * , 430 ( 1966 ) .
s. d. drell and a. c. hearn , _ phys .
_ * 16 * , 430 ( 1966 ) .
the gdh and a2 collaborations j. ahrens _ et al .
_ , _ phys .
letts . _ * 87 * , 022003 ( 2001 ) , hep - ex/0105089 .
d. drechsel _ et al .
_ , _ nucl .
phys . _ * a645 * , 145 ( 1999 ) . n. bianchi and t. thomas , _ phys . lett . _ * b450 * , 439 ( 1999 ) .
h. burkhardt and w. n. cottingham , _ ann .
_ * 56 * , 453 ( 1970 ) .
k. griffioen , talk presented at this symposium ( 2001 ) .
j. soffer and o. v. teryaev , _ phys .
lett . _ * 70 * , 3371 ( 1993 ) .
v. burkert and b. ioffe , _ phys .
lett . _ * b296 * , 223 ( 1992 ) ; _ j. exp .
phys . _ * 78 * , 619 ( 1994 ) .
j. p. chen , talk presented at this symposium ( 2001 ) .
d. drechsel _ et al .
_ , _ phys .
rev . _ * d63 * , 114010 ( 2001 ) .
e. d. bloom and f. j. gilman , _ phys .
lett . _ * 25 * , 1140 ( 1970 ) ; _ phys .
rev . _ * d4 * , 2901 ( 1971 ) .
i. niculescu _
_ , _ phys .
_ * 85 * , 1182 , 1186 ( 2000 ) . t. gehrmann and w. j. stirling , _ phys
. rev . _ * d53 * , 6100 ( 1996 ) .
e. shuryak and a. vainshtein , _ nucl .
phys . _ * b201 * , 141 ( 1982 ) .
r. jaffe and x. ji , _ phys .
rev . _ * d43 * , 724 ( 1991 ) . j. l. cortes , b. pire and j. p.
ralston , _ z. phys . _ * c55 * , 409 ( 1992 ) . s. wandzura and f. wilczek , _ phys
. lett . _ * b72 * , 195 ( 1977 ) .
p. bosted , talk presented at this symposium ( 2001 ) .
z. meziani , talk presented at this symposium ( 2001 ) .
a. v. efremov , o. v. teryaev and e. leader , _ phys .
rev . _ * d55 * , 4307 ( 1997 ) .
the smc collaboration , b. adeva _
. lett . _ * b369 * , 93 ( 1996 ) .
the smc collaboration , b. adeva _
_ , _ phys .
_ * b420 * , 180 ( 1998 ) .
the hermes collaboration , k. ackerstaff _
lett . _ * 81 * , 5519 ( 1998 ) .
the e866 collaboration , e. a. hawker _
_ , _ phys .
lett . _ * 80 * , 3715 ( 1998 ) . the e866 collaboration , j. c. peng _ et al .
_ , _ phys .
_ * d58 * , 092004 ( 1998 ) . g. t. garvey and j. c. peng , _ prog .
_ * 47 * , 203 ( 2001 ) , nucl - ex/0109010 . the hermes collaboration , k. ackerstaff _ et al . _ , _ phys .
lett . _ * b464 * , 123 ( 1999 ) .
shibata , talk presented at this symposium ( 2001 ) .
s. bernreuther , talk presented at this symposium ( 2001 ) .
d. i. diakonov _
. rev . _ * d56 * , 4069 ( 1997 ) .
c. bourrely and j. soffer , _ phys .
_ * b314 * , 132 ( 1993 ) .
c. bourrely and j. soffer , _ nucl .
phys . _ * b423 * , 329 ( 1994 ) .
j. kiryluk , talk presented at this symposium ( 2001 ) .
n. saito , talk presented at this symposium ( 2001 ) .
the hermes collaboration , a. airapetian _
_ , _ phys .
lett . _ * 84 * , 2584 ( 2000 ) .
compass proposal , cern / spslc-96 - 14 ( march , 1996 ) .
m. liu , talk presented at this symposium ( 2001 ) .
j. collins , _ nucl .
_ * b396 * , 161 ( 1993 ) ; _ nucl .
phys . _ * b420 * , 565 ( 1994 ) .
a. v. efremov , o. g. smirnova and l. g. tkachev , _ nucl .
suppl . _ * 74 * , 49 ( 1999 ) .
g. schnell , talk presented at this symposium ( 2001 ) .
the hermes collaboration , a. airapetian _
_ , _ phys .
lett . _ * 84 * , 4047 ( 2000 ) .
the hermes collaboration , a. airapetian _
_ , _ phys .
rev . _ * d64 * , 097101 ( 2001 ) .
k. a. oganessian , h. r. avakian , n. bianchi and a. m. kotzinian , hep - ph/9808368 .
e. de sanctis , w .- d . nowak and k.
a. oganessian , _ phys .
lett . _ * b483 * , 69 ( 2000 ) .
a. bravar for the smc collaboration , _ nucl .
suppl . _ * 79 * , 520 ( 1999 ) .
v. a. korotkov , w. -d .
nowak and k. a. oganessian , _ eur .
phys . j. _ * c18 * , 639 ( 2001 ) .
x. ji , _ phys .
lett . _ * 78 * , 610 ( 1997 ) ; _ phys .
* d55 * , 7114 ( 1997 ) . the zeus collaboration , p. r. b. saull , proceedings of the international europhysics conference on high - energy physics , tampere , finland , 420 - 422 ( 1999 ) , hep - ex/0003030 . the h1 collaboration , c. adloff _
_ , _ phys .
lett . _ * b517 * , 47 ( 2001 ) .
the hermes collaboration , a. airapetian _
_ , _ phys . rev .
_ * 87 * , 182001 ( 2001 ) . the clas collaboration , s. stepanyan _
_ , _ phys .
lett . _ * 87 * , 182002 ( 2001 ) .
n. bianchi , talk presented at this symposium ( 2001 ) .
_ , _ phys . lett . _ * b411 * , 193 ( 1997 ) . the aleph collaboration , d. buskulic _ et al .
_ , _ phys .
_ * b374 * , 319 ( 1996 ) .
the opal collaboration , k. ackerstaff _
_ , _ eur .
phys . j. _ * c2 * , 49 ( 1998 ) .
the nomad collaboration , p. astier _
_ , _ nucl .
phys . _ * b588 * , 3 ( 2000 ) .
the nomad collaboration , p. astier _
_ , _ nucl .
_ * b605 * , 3 ( 2001 ) .
the e665 collaboration , m. r. adams _ et al .
_ , _ eur .
phys . j. _ * c17 * , 263 ( 2000 ) .
the hermes collaboration , a. airapetian _
_ , _ phys .
_ * d64 * , 112005 ( 2001 ) .
d. de florian , m. stratmann and w. vogelsang , _ phys .
* d57 * , 5811 ( 1998 ) .
b. q. ma , i. schmidt and j. j. yang , _ phys .
_ * b477 * , 107 ( 2000 ) .
j. c. collins _
_ , _ phys .
rev . _ * d56 * , 2982 ( 1997 ) . m. vanderhaeghen _
. rev . _ * d60 * , 094017 ( 1999 ) .
the hermes collaboration , a. airapetian _
et al . _ , hep - ex/0112022 .
the clas collaboration , r. de vita _ et al .
_ , _ phys .
lett . _ * 88 * , 082001 ( 2002 ) .
g. r. farrar and d. r. jackson , _ phys .
lett . _ * 35 * , 1416 ( 1975 ) .
s. brodsky , m. burkardt and i. schmidt , _ nucl .
phys . _ * b441 * , 197 ( 1995 ) .
g. bunce , talk presented at this symposium ( 2001 ) . | the current status of spin physics experiments , based on talks presented at the third circum - pan - pacific symposium on high energy spin physics held in beijing , 2001 , is summarized in this article .
highlights of recent experimental results at slac , jlab , and desy , as well as future plans at these facilities and at rhic - spin are discussed . |
open clusters ( ocs ) have long been used as important tools in the study of the galactic disc .
the young clusters have been used to determine spiral arm structure , and to map the rotation curve of the galaxy .
the old clusters are excellent tracers of the structure , kinematics , and chemistry of the galactic disc @xcite . in principle ,
basic parameters like distance , age , and metallicity can be determined for a cluster more accurately than for a field star .
therefore , ocs are better tracers of large scale properties of the galactic disc population than field stars @xcite .
about 20 years ago , the catalogue of open cluster data compiled by lyng @xcite was used to derive the radial gradients and other structural properties of the galactic disc @xcite . in recent years
, the available data on open clusters have increased very quickly and the basic parameters of these clusters have also considerably improved .
* hereafter daml ) compiled the new catalogue of optically visible open clusters and candidates including 1.5 times more clusters than the catalogue of @xcite .
* hereafter k05 ) presented a catalogue of astrophysical data for 520 ocs , with data derived from the all - sky compiled catalogue of 2.5 million stars . using these data ,
the properties of the galactic disc such as the scale height , the scale length , and the metallicity distribution of the disc , were derived @xcite .
@xcite used 106 ocs with available data on positions , distances , and radial velocities to analyze the local kinematics .
they found that the dispersion of the radial velocity increases with age : for the old clusters it is about twice that of young clusters .
* hereafter bko ) calculated the galactic orbits for 69 ocs based on carefully selected data on distances , absolute proper motions , and radial velocities .
@xcite calculated the orbits of ngc 188 and m67 .
in addition , @xcite calculated the orbits for ngc 188 , m67 , and ngc 2420 , ( * ? ? ?
* hereafter c94 ) expanded the sample with 5 classic clusters , and @xcite calculated the orbits for a total of 7 old clusters .
@xcite calculated the orbit for ngc 2355 , @xcite calculated the orbits for ngc 1912 and ngc 1907 , and @xcite calculated the orbit for m48 . in recent years , @xcite determined the rotation velocity of the spiral pattern of the galaxy by studying the birthplaces of ocs in the galactic disc as a function of their ages .
the birthplaces of these clusters were determined by assuming * that their orbits were circular .
* using 148 ocs within the projected distance onto the galactic plane @xmath9 kpc , @xcite derived the solar motion and the velocity ellipsoid of ocs .
they also calculated the galactic orbits of these clusters and presented the mean parameters of their orbits . based on the spatial and velocity distributions of ocs
, @xcite also identified the existence of four open cluster complexes of different ages @xcite , * which verifed the nature of clustering in ocs in the solar neighbourhood identified in previous studies@xcite .
* @xcite used 374 ocs taken from the 2.7 version of daml catalogue to measure the epicycle frequency @xmath10 in the galactic disc .
they also calculated the orbits of these ocs and discussed the distribution of initial velocities of these clusters .
the orbital motions of ocs are important not only for our understanding of the dynamical evolution of ocs in the galaxy @xcite , but also for investigating their effect on the time - evolution of the abundance gradient in the galactic disc ( c94 ) .
the main aim of this paper is to calculate the orbits of ocs with the improved data for an enlarged sample of 488 ocs and to discuss the kinematical properties of these clusters .
we present the collection of data in section 2 . in section 3 , we analyze the statistical properties of the sample , and especially their age - velocity dispersion relations .
we present the orbital parameters and their associated uncertainties calculated in a given galactic potential model in section 4 , followed by section 5 , which presents the differences of orbital parameters due to different galactic potential models . in section 6 , we compare our results with the orbital parameters calculated for field stars and globular clusters , and discuss the effect of the orbital motions of ocs for the radial abundance gradient derived for these clusters , we also compare the orbital parameters with those derived by previous studies .
our conclusions are given in section 7 .
we chose the daml catalogue as the main source of the fundamental parameters for the ocs .
this catalogue uses the webda database and previous catalogues of @xcite as a starting point .
kinematical , and metallicity data of the new objects when available are inserted @xcite .
they also made use of the simbad database and of the literature to find data on the clusters or on individual stars of the clusters , to obtain the averaged values of radial velocities and proper motions .
this catalogue is regularly updated , and the latest version is available from its website .
the present 2.9 version ( 13/apr/2008 ) of daml catalogue contains 1776 objects , of which 936 have published distances , ages , and reddening values , 890 have published proper motions and 447 have radial velocities .
869 clusters with distances and proper motions data are taken from the daml catalogue as the initial sample .
most of distances and ages listed in daml are taken from webda , and are updated with new data from literature @xcite .
the distances estimated by @xcite based on hipparcos parallaxes @xcite of member stars are also adopted by daml .
a comparison of the hipparcos parallaxes with photometric distances shows good agreement @xcite .
the distances listed in daml are taken from different authors using different observational techniques and reduction methods , and no errors for this parameter are listed in daml , the non - uniformity and uncertainties in the adopted distance data should be estimated .
@xcite studied the accuracy of available parameters such as the age , reddening , and distance for ocs by using the independently derived values published in the literature .
they used a sample of 395 clusters in their statistical analysis .
they found that , for about 80% clusters in their sample , the error of distance is less than 20% .
they compared their results with the data of daml catalogue , and pointed out that the distances listed in daml catalogue have the same error distributions . in our present study ,
20% relative errors in distances are adopted for ocs in our sample .
@xcite re - reduced the raw data of hipparcos mission , and the new reduction provides an improvement by a factor 2.2 compared to the catalogue published in 1997 @xcite .
we calculate the difference between the distances derived by @xcite using the re - reduced hipparcos data and that listed in daml catalogue for a sample of 17 ocs in common .
we get a difference of @xmath11 between these two data sets , which is less than our adopted @xmath12 error for distance parameters and is considered in the following orbit calculation .
the absolute proper motion data of ocs listed in daml are adopted in the present study .
the proper motion data are also compiled from different authors , but they are all based on the hipparcos system . in our final sample ,
proper motions of @xmath13 clusters are derived from the hipparcos catalogue @xcite , @xmath14 proper motions are derived from the tycho2 catalogue @xcite , @xmath15 proper motions are derived from the ascc-2.5 catalogue @xcite , and only @xmath16 proper motions are derived from the ucac2 catalogue @xcite .
the hipparcos catalogue is considered as the realization of the icrs at optical wavelengths .
the systematic error of proper motions in the hipparcos catalogue with respect to icrs is estimated to be @xmath17 @xcite .
@xcite searched the hipparcos catalogue and found 9 new ocs and derived the proper motions for these clusters .
@xcite determined mean proper motions of 205 ocs from their member stars found in the hipparcos catalogue . in our final sample ,
the proper motions derived from hipparcos catalogue are all taken from the results of @xcite and @xcite .
the tycho2 catalogue @xcite presents very precise proper motions with random errors between 1 and 3 @xmath18 in the hipparcos system .
there are no significant systematic differences between the proper motions of these two catalogues @xcite .
@xcite determined the mean absolute proper motions of 206 ocs from the data in tycho2 catalogue .
@xcite found 11 new ocs candidates in tycho2 catalogue and determined the mean proper motions for these clusters .
@xcite determined the mean proper motions for 167 ocs based on the kinematic and photometric data in tycho2 catalogue .
@xcite compared their results with those derived by @xcite based on hipparcos catalogue , and found that the mean difference in the proper motions is less than 1 mas yr@xmath6 .
@xcite also compared their results with those derived by @xcite and found that the difference is @xmath19 . in our final sample , the mean proper motions derived from tycho2 catalogue are taken from above mentioned studies . the ascc-2.5 catalogue is based on large , modern , high - precision catalogues of the hipparcos - tycho family , including the tycho2 catalogue , and provides the most complete all - sky catalogue of stars having uniform high precision astrometric and photometric data down to @xmath20 mag @xcite . in our final sample ,
the mean proper motions derived from ascc-2.5 catalogue by @xcite are adopted .
@xcite compared their results with those derived by @xcite and @xcite and found that their results agree quite well with previous studies .
the ucac2 catalogue presents proper motions in the hipparcos system with nominal errors of 1 to 3 @xmath18 for stars up to @xmath21 mag and about 4 to 7 @xmath18 for fainter stars up to @xmath22 mag .
the systematic errors of the proper motions in ucac2 are in the range 0.5 to 1.0 @xmath18 @xcite .
@xcite determined the mean proper motions for 428 ocs from the ucac2 catalogue .
they compared their results with those derived from hipparcos catalogue @xcite , tycho2 catalogue @xcite and ascc-2.5 catalogue @xcite , and found that there is no statistical distinction between the compared mean proper motions of ocs in these catalogues .
we also calculate the difference of the mean proper motions derived by @xcite using the re - reduced hipparcos data and that adopted in our final sample taken from the daml catalogue for the 17 ocs in common .
we get a difference of @xmath23 mas yr@xmath6 in @xmath24 and a difference of @xmath25 in @xmath26 .
the differences are within the range of the errors for the mean proper motions listed in the daml catalogue .
there are 431 ocs whose distances , proper motions , and radial velocities are available from the daml catalogue .
most radial velocities are taken from the catalogues compiled by @xcite . @xcite cross - identified the ascc-2.5 catalogue with the general catalogue of radial velocities of @xcite and derived radial velocities for 290 ocs based on their membership determination .
they found that the mean difference between their results and those in the published literature for common clusters is @xmath27 km s@xmath6 .
the 363 radial velocities of ocs compiled by @xcite are derived from the 2nd version of the catalogue of radial velocities of galactic stars with high precision astrometric data ( crvad-2 ) .
the crvad-2 is the result of updating and expanding the list of stars with known radial velocities and high precision astrometric and photometric data taken from the ascc-2.5 catalogue .
the mean difference of radial velocities for 177 clusters in common between the crvad-2 and the literature is @xmath28 km s@xmath6 .
@xcite presented high - precision radial velocities for 71 ocs obtained with multi - object spectrographs . for 25 clusters in their sample ,
the radial velocities are newly obtained and are not included in daml .
@xcite derived mean radial velocities for 166 ocs based on observations with the coravel spectrovelocimeters .
the radial velocities of 64 ocs derived by @xcite are not listed in daml .
we supplement the new radial velocities derived by @xcite and @xcite into our final sample .
we also update the known radial velocities in daml with those derived by @xcite due to their higher precision . in our final sample , 488 ocs with distances , proper motions , and radial velocities
are included .
our present sample is a subsample mainly taken from the daml catalogue .
we plot the volume density of 1082 ocs ( open circles ) with distance data available in daml as a function of heliocentric distance @xmath29 in fig .
[ incom ] . in order to estimate the completeness of the ocs in our present sample
, we also plot the volume density of 488 ocs ( filled circles ) in our present sample in fig .
[ incom ] . for clusters with heliocentric distance @xmath30 kpc ,
the volume density distributions for these two samples of ocs are completely consistent . for clusters with heliocentric distance @xmath31 kpc ,
the volume densities of ocs in our present sample are less than those for ocs in daml , but the distributions for these two samples are very similar .
[ incom ] indicates that our present ocs sample is a representative subsample of currently observed ocs in the galaxy . using a sample of 654 ocs with distance data available , @xcite simulated the effects of completeness in their ocs sample .
they found that a total number of @xmath32 ocs with heliocentric distance @xmath33 kpc should be observed . within the same distance range , there are 498 ocs in daml catalogue and 271 ocs in our present sample , the completeness can be estimated as @xmath34 and @xmath35 for daml catalogue and for our present sample , respectively .
a possible source for the remainder unobserved ocs comes from the very young clusters which are still embedded within giant molecular clouds , they are heavily obscured and are very difficult to identify @xcite .
table [ tb1 ] lists the adopted positions ( @xmath36 ) , heliocentric distances @xmath37 , radial velocities @xmath38 , and absolute proper motions ( @xmath24 , @xmath26 ) of 488 ocs in our final sample .
the adopted @xmath12 errors for distances , the errors of absolute proper motions and of radial velocities listed in daml , @xcite , and @xcite are also presented in table [ tb1 ] . in table
[ tb1 ] , we also list the ages and metallicities [ fe / h ] for each cluster if available in daml .
ages are currently available for 445 ocs , 109 of which have [ fe / h ] values also available .
the initial conditions for orbit calculation are the presently observed positions and velocities of ocs with respect to the galactocentric reference frame . adopting the solar motion ( @xmath39 , @xmath40 , @xmath41)@xmath42 @xmath43 ( 10.0 , 5.2 , 7.2 ) km s@xmath6 from @xcite , the local standard of rest ( lsr ) velocities of ocs
are determined from the data listed in table [ tb1 ] .
the lsr velocities are then corrected to the galactic standard of rest ( gsr ) by adopting the galactocentric distance of sun @xmath44 kpc @xcite and a rotation velocity of the lsr of @xmath45 km s@xmath6@xcite .
all space coordinates @xmath46 , @xmath47 , @xmath48 , and velocity components @xmath39 , @xmath40 , @xmath41 refer to a galactocentric righthanded cartesian coordinate system with the @xmath46 direction directed towards the galactic anticenter and the @xmath48 direction directed towards the galactic north pole @xcite .
the @xmath39 , @xmath40 , @xmath41 velocity components and their errors are calculated with the method of @xcite .
table [ tb2 ] lists the initial conditions used to calculate the orbital solution for each cluster in our sample .
the errors in velocity components include the errors in absolute proper motions , radial velocities , and distances of ocs listed in table [ tb1 ] .
[ cols="<,^,^,^,^,^,^,^,^ " , ] we use two different potential models to calculate the differences in the derived orbital parameters .
the first model is the one proposed by ( * ? ? ?
* hereafter p90 ) and then used by @xcite as their representative model to calculate the orbits of 38 globular clusters .
this model consists of axisymmetric potential with three components : bulge , disc , and dark halo .
the bulge is modeled as plummer potential @xcite .
the disc is the same as as91 in the form of @xcite with different coefficients .
the dark halo is modeled as logarithmic potential , which assures a flat rotation curve , but imply an infinite mass .
the second model is the one proposed by ( * ? ? ?
* hereafter fsc96 ) and used by n04 to calculate the orbits of nearby f and g dwarf stars .
this axisymmetric potential also consists of three components : central core , disc , and dark halo .
the potential of the central core is modeled by two spherical components , representing the bulge / stellar - halo and an inner core component .
the disc is modeled using a combination of three analytical discs of @xcite .
the potential of the dark halo is assumed to be logarithmic .
in figs . [ df90 ] and
[ df96 ] , we compare the orbital parameters derived from p90 and fsc96 models with those derived from as91 model .
[ df90 ] shows that the orbital parameters derived from p90 model are very consistent with those derived from as91 model , and no systematic differences can be found . for fsc96 model , fig .
[ df96 ] shows that @xmath49 and @xmath50 are consistent between fsc96 and as91 models , and there are no systematic differences in these two parameters .
but for @xmath51 kpc , the derived @xmath52 and @xmath53 for fsc96 model are systematically smaller than those derived from as91 model .
[ df96 ] also indicates that the derived @xmath54 and @xmath55 for fsc96 model are systematically bigger than those derived from as91 model . in figs .
[ obs90 ] and [ obs96 ] , for p90 and fsc96 models , we show the orbits for the same clusters as presented in fig . [ obs91 ] .
[ obs90 ] indicates that the orbital shapes for most of clusters , including the asymmetric meridional orbits of berkeley 33 and berkeley 20 , are very similar between p90 and as91 models .
[ obs96 ] shows that for most of clusters , the orbital shapes derived from fsc96 model are different from those derived from as91 model .
the meridional orbits of berkeley 33 and berkeley 20 are symmetric in fsc96 model . for each cluster ,
the relative difference in each orbital parameter from different models is calculated as @xmath56 , where @xmath57 , @xmath58 , and @xmath59 represent the derived orbital parameters by p90 , as91 , and fsc96 models respectively .
for each orbital parameter , the median of the relative difference between the given models and as91 model is listed in table [ tb5 ] .
table [ tb5 ] shows that the relative differences of the derived orbital parameters for p90 model are smaller than those for fsc96 model . especially for @xmath54 and @xmath55 ,
the results derived from fsc96 model are very different from those derived from as91 and p90 models .
the consistent orbital parameters derived from as91 and p90 models are due to the very similar mass distributions of these two models .
the main difference between as91 and p90 models is the bulge model , whereas bulge model dominates the orbit of a cluster only within 1 kpc from the galactic centre .
no clusters in our present sample can move to this range .
the fsc96 model is very different from as91 , the total mass interior to the same galactocentric distance @xmath60 in the fsc96 model is much larger than that in the as91 model when @xmath61 kpc
. the larger mass of fsc96 model can control the cluster to move within smaller apogalactic distance . on the other hand ,
the mass of disc in fsc96 model is smaller than that in as91 model , clusters can move to more larger vertical distances in fsc96 model .
table [ tb5 ] also shows that , for most of derived orbital parameters , the relative differences due to different potential models are smaller than those from observational errors .
but for fsc96 model , the relative differences in @xmath55 are much larger than the uncertainties due to observational errors of input data , which indicates the major effect of the disc model of fsc96 .
the left panels of fig . [ rzs ] show the @xmath52 vs @xmath62 , @xmath49 vs @xmath62 , and @xmath54 vs @xmath63 diagrams , where @xmath63 is the current observed vertical distance from the galactic disc .
the panel of @xmath52 vs @xmath62 diagram indicates that , for most of clusters in present sample , their current observed positions are very close to their apogalacticons .
whereas , the panel of @xmath49 vs @xmath62 diagram indicates that most of clusters are far away from their perigalacticons .
the panel of @xmath54 vs @xmath63 diagrams indicates that most of clusters do not arrive at their maximum distances from the disc and are crossing the galactic plane . in the right panels of fig .
[ rzs ] , we plot the histograms of the relative differences for corresponding parameters in the left panels .
the top panel shows the relative difference between @xmath52 and @xmath62 and indicates that , at present , about 70% clusters are moving within only about 5% distances from their apogalacticons .
the median of the relative differences between @xmath52 and @xmath62 is 2.0% .
the middle panel shows the relative difference between @xmath49 and @xmath62 , and indicates a significantly larger relative difference with respect to that between @xmath52 and @xmath62 .
the median of the relative differences between @xmath49 and @xmath62 is 8.1% which is larger than that between @xmath52 and @xmath62 .
the bottom panel shows the relative differences between @xmath54 and @xmath63 , and indicates very large differences between these two parameters .
the median of the relative differences is 68.8% .
the mean time @xmath55 for our present sample is very short , so the time for a cluster moving near its @xmath64 is short , which is the possible reason for the large differences between the observed @xmath48 and the orbital parameter @xmath64 . for two clusters
ngc 2682 and ngc 2420 whose orbital eccentricities are larger than 0.1 , c94 calculated the probability of finding the clusters at the given galactocentric distances and found that the detection probability is the largest at the cluster s apogalacticon .
following c94 , we also define a probability function @xmath65 to calculate the probability of a cluster to be observed at the galactocentric position @xmath60 during the orbital period @xmath53 , where @xmath66 is the cluster s velocity at @xmath60 .
[ prs ] shows the probability distributions @xmath67 for the same clusters as in fig .
[ obs91 ] .
the filled square in fig .
[ prs ] indicates the present observed position for each cluster .
[ prs ] shows that the detection probability @xmath67 increases with the galactocentric distance @xmath60 , the largest detection probability is at the cluster s apogalacticon .
the derived probabilities for other clusters also indicate that the detection probability for a cluster at its apogalacticon is the largest one during its orbital period , which is consistent with the result of c94 .
because the detection probability at the apogalacticon for a cluster is the largest during its orbital period , it is easier to find a cluster near its apogalacticon just as fig .
[ rzs ] indicates .
the ocs have long been used as tracers of radial metallicity gradients in the galactic disc . since the early work by @xcite , others have found general agreement in the existence and magnitude of the trend .
most investigators have found gradients of @xmath68 to @xmath69 dex kpc@xmath6 over a range of distances from 7 to 16 kpc from the galactic centre @xcite .
recently , @xcite presented metallicities for a sample of 39 intermediate age and old ocs based on an updated abundance calibration of spectroscopic indices .
they found a metallicity gradient of @xmath70 dex kpc@xmath6 over a range of galactocentric distances of 7 to 16 kpc .
@xcite compiled a ocs catalogue of 119 objects with ages , distances , and metallicities available , which led to a metallicity gradient of @xmath71 dex kpc@xmath6 , similar to the result derived by @xcite from a homogeneous sample .
the [ fe / h ] vs current observed galactocentric distances @xmath62 diagram is plotted in the top panel of fig .
109 clusters with [ fe / h ] data listed in our present sample are plotted as open circles , 48 clusters not listed in our present sample but with [ fe / h ] data are plotted as plus signs .
there are 12 clusters in the range of @xmath72 kpc , only 3 clusters are included in our present sample . considering the observational errors in [ fe / h ] and @xmath62 , we perform a linear least - square fitting to the clusters listed in our sample with @xmath73 kpc .
for each cluster , we assign the typical observational uncertainty of 0.15 dex to their [ fe / h ] data .
the errors of @xmath62 are calculated from the data listed in table [ tb2 ] .
we get a gradient of @xmath74 dex kpc@xmath6 .
we also perform the same fitting to all clusters in the range of @xmath75 kpc , we get a gradient of @xmath76 dex kpc@xmath6 .
the fitted straight - line for the clusters listed in our present sample within @xmath73 kpc is plotted in the same panel of fig .
[ fehr ] and the slope @xmath77 of this line is also labeled . in the * most * recent study ,
@xcite derived a radial metallicity gradient of @xmath78 dex kpc@xmath6 based on a sample of 144 ocs .
our derived radial metallicity gradient of @xmath79 is slightly smaller than the previous results @xcite .
based on 45 ocs with high - resolution spectroscopy , @xcite found a steep metallicity gradient for clusters with @xmath80 kpc and a plateau for clusters at larger galactocentric distances .
[ fehr ] indicates that the metallicity distribution of clusters with @xmath72 kpc is flat and there is no significant radial metallicity gradient within this distance range , which is consistent with the result of @xcite . in our present sample , only three ocs have galactocentric distances greater than 13.5 kpc ( the open circles in fig .
[ fehr ] ) . in order to diminish the effect of the small sample at large galactocentric distances , we only consider ocs with @xmath73 kpc in our present sample to derive the radial metallicity gradient .
the maximum galactocentric distances of the ocs in the previous studies extend to 17 kpc and some clusters within the flat metallicity distribution were included in their samples@xcite , which is the main reason * why * they obtained * larger * radial metallicity gradient .
if we use all clusters with @xmath81 kpc , we get a gradient of @xmath82 dex kpc@xmath6 which is consistent with the * most * recent result of @xcite .
the radial metallicity gradient of the ocs provides strong constraints on the formation and evolution of the galaxy .
detailed models of galactic chemical evolution have been improved over * the * last decades and most models can reproduce the presently observed radial metallicity distribution of the galaxy ( * ? ? ?
* and references therein ) .
more recently , @xcite considered various mechanisms including infall , star formation and delayed disk formation to find the effect of each mechanism on their derived galactic chemical evolution model .
they found that using the star formation rate(sfr ) of the modified kennicutt law , their model can properly predict both the current metallicity gradient and its time evolution .
but , their best model also predicts that the outer disc has a steeper gradient than the inner disc , which is contrary to the result derived from ocs .
@xcite adopted an inside - out formation model of the galactic disc to reproduce the radial metallicity gradient of the ocs . in their model , the infall of gas is represented by an exponential law combined with the distribution of gas in the halo .
the inner regions are rapidly evolving due to the higher infall and sfr , while the outer parts evolve more slowly .
their model can reproduce the main features of the metallicity gradient and the evolution of the ocs . in order to better reproduce the metallicity plateau at large galactocentric distances , an additional uniform inflow per unit disc area
should be considered , but it is difficult to reconcile with the present - day radial distribution of the sfr .
* a sequence of merging episodes * in the past history of the galaxy would explain the outer metallicity plateau of the ocs@xcite . in our previous discussion
, we have showed that the apogalacticon is the place where a cluster spends the largest fraction of its life . in the bottom panel of fig .
[ fehr ] , we plot the [ fe / h ] vs @xmath52 diagram for clusters listed in our sample within @xmath73 kpc .
we also perform a linear least - square fitting to the data considering the observational errors in [ fe / h ] and @xmath52 .
the errors of @xmath52 are taken from table [ tb3 ] .
we plot the fitted straight - line in the same panel and also label the slope @xmath8 dex kpc@xmath6 in the panel .
this result indicates that , for clusters within @xmath73 kpc , the observed metallicity gradient at present is similar to that derived from the most probable observed positions of the clusters , which is consistent with the result of c94 .
[ hcom ] shows the histograms of orbital eccentricities for different populations : globular clusters ( top left panel ) , disc giants ( top right panel ) , disc f and g dwarf stars ( lower left panel ) , and ocs ( lower right panel ) . *
each histogram is * normalized to its maximum value in the distribution .
the orbital parameters for 54 globular clusters calculated with as91 model are taken from a06 and a08 .
the orbital parameters of disc clump giants calculated with as91 model are taken from s08 .
the orbital parameters of disc f and g dwarf stars calculated with fsc96 model are taken from n04 .
globular clusters represent the halo population which is primarily a system with a large velocity dispersion , and a wide range of orbit characteristics is expected @xcite .
this is indeed what the top left panel of fig .
[ hcom ] shows . there is a large range in orbital eccentricities of globular clusters , which can get as low as 0.1 , but , in the mean , the eccentricities are high with an average of @xmath83 .
s08 assigns to each star its probability to belong to the thin disc , the thick disc , the hercules stream and the halo based on the basis of its ( @xmath39 , @xmath40 , @xmath41 ) velocity and their velocity ellipsoids . in the top right panel of fig .
[ hcom ] , we show the histograms of orbital eccentricities for stars with probabilities higher than 80% to belong to the thin disc ( solid lines ) and stars with probabilities higher than 80% to belong to the thick disc ( dot lines ) . fig .
[ hcom ] indicates that no thin disc giant has eccentricity greater than 0.3 and no thick disc giant has eccentricity less than 0.1 .
the orbital eccentricities of the thin and the thick disc giants are overlapped between 0.1 and 0.3 .
the average of orbital eccentricities for the thin disc giants is @xmath84 , and the one for the thick disc giants is @xmath85 , which is less than the average of orbital eccentricities for globular clusters .
the histogram of orbital eccentricities for disc f and g dwarf stars in the lower left panel of fig .
[ hcom ] shows that most of stars belong to the thin disc whereas a fraction of stars belong to the thick disc .
if we assume that stars with eccentricities greater than 0.3 belong to the thick disc , the thick - disc fraction is about 3.6% which is close to the value of 2.9% derived by @xcite from the same sample with different method . the lower right panel of fig .
[ hcom ] shows the histograms of orbital eccentricities for ocs in our present sample calculated with as91 model ( solid lines ) and fsc96 model ( dot lines ) .
the distributions of eccentricities in these two models are very similar .
the distribution of orbital eccentricities for ocs is similar to that of disc f and g dwarf stars which indicates that most of clusters belong to the thin disc and a fraction of thick - disc clusters exist in our sample .
we assume the same limited value of 0.3 for orbital eccentricities to distinguish thin - disc and thick - disc clusters , and find 3.7% clusters in our sample are probably thick - disc clusters . in our present sample ,
the @xmath54 values of berkeley 29 , berkeley 31 , and berkeley 33 are @xmath86 , @xmath87 , and @xmath88 kpc , which are much larger than those of the other clusters in the sample . the maximum of @xmath54 of the other clusters is less than 3 kpc and the observed maximum of @xmath48 is 2.1 kpc for berkeley 29 .
the very large errors of @xmath54 for these three clusters indicate that their orbits are very uncertain and more observations are needed to improve the precision of the input data for these clusters .
berkeley 20 has the maximum of @xmath89 in our present sample which indicates it is a halo cluster . but
this cluster has an age of 6.0 gyr and [ fe / h]@xmath90 indicating it is a thick disc cluster .
if we adopt the limited value @xmath91 with @xmath92 from the mean , the cluster can be identified as a member of the thick disc population only based on orbital eccentricity @xmath50 .
more observations are also needed to improve the orbital parameter for this cluster .
there are two studies which calculated the orbits for a large samples of ocs and the orbital parameters were also derived .
bko calculated the orbits for 69 ocs and the orbital parameters @xmath52 , @xmath49 , @xmath50 , and @xmath54 were also derived for those clusters . in fig .
[ comp ] , for common clusters , we compare the derived orbital parameters in as91 model with those derived by bko . fig . [ comp ] indicates that , for most clusters , the derived @xmath52 , @xmath49 , and @xmath50 are consistent with those derived by kbo .
the maximum difference in @xmath52 is for ngc 2420 .
the radial velocity for this cluster is @xmath93 km s@xmath6 in our present study , but the value of @xmath94 km s@xmath6 was adopted by bko .
the large difference in radial velocity make the large difference in @xmath52 for ngc 2420 .
the maximum difference in @xmath49 and @xmath50 is for ngc 7789 .
the radial velocity of @xmath95 km s@xmath6 is adopted for this cluster in our study , bko adopted the value of @xmath96 km s@xmath6 .
the large difference in @xmath49 and @xmath50 for ngc 7789 is also due to the large difference in radial velocity data . for most of ocs ,
the orbital parameter @xmath64 in our present study are less than those derived by bko .
the systematic difference in @xmath64 is due to the different disc models adopted by this study and by bko .
@xcite calculated the orbits for a sample of 148 ocs within @xmath97 kpc and the mean orbital parameters are listed in their table 2 : @xmath98 kpc , @xmath99 kpc , @xmath100 , and @xmath101 kpc . for the clusters in our present sample within the same distance range , the corresponding mean
orbital parameters calculated with as91 model are derived as : @xmath102 kpc , @xmath103 kpc , @xmath104 , and @xmath105 kpc . because the input data for most of clusters in this study are similar to those in the study of @xcite , the differences of the derived orbital parameters are mainly due to the adopted different galactic models .
@xcite used the galactic model of @xcite , the analytic solution of the orbital parameter can be derived from their model . in comparison ,
the much larger disc mass in as91 model is responsible for the large difference in the derived value of @xmath64 with respect to the other model .
many surveys have found that the structure of the galaxy is more complex than previously thought , and about 10 moving groups can be identified from nearby stars with heliocentric distances less than 100 pc @xcite .
using ocs with the projected distance onto the galactic plane @xmath97 kpc from the sun , and based on the surface density distribution and the tangential velocity distribution of those clusters , @xcite identified four open cluster complexes ( occs ) .
the numbers of kinematic member clusters for the four occs : occ 1 , occ 2 , occ 3 , and occ 4 are 23 , 27 , 8 , and 9 @xcite . in our present sample , for the corresponding occs , the numbers of member clusters identified by @xcite whose orbits can be determined are 20 , 13 , 3 , and 4 .
those occs were detected from their overdensity in the spatial distribution of ocs in the solar neighbourhood and the membership of the member clusters were determined only based on their tangential velocities@xcite . in fig .
[ occs ] , we use the @xmath106 @xmath107 and @xmath64 @xmath50 diagrams to * attempt to recover * those occs , where @xmath106 is the total galactocentric velocity of cluster . the crosses in fig .
[ occs ] represent clusters with heliocentric distance less than 1.3 kpc , the open circles represent member clusters in each occ . in the panels of occ 1 ,
the open circles represent member clusters with age less than 30 myr and triangles represent member clusters with age between 30 and 80 myr .
there are 18 member clusters in occ 1 whose age data are available .
the @xmath106 @xmath107 diagram of ooc 1 indicates that 6 clusters with age less 30 myr and 4 clusters with age great than 30 myr distribute in a small region with @xmath108 km s@xmath6 and @xmath109 kpc , those clusters can be considered as kinematic members of occ 1 .
but in the @xmath64 @xmath50 diagram of occ 1 , it is difficult to identify any clustering of those member clusters .
* we could only determine the orbits of 50% of the assumed members of occ 2 .
* most of member clusters of occ 2 have age between 200 and 400 myr .
the panels of occ 2 indicate that we can not detect any clustering of the member clusters in the @xmath106 @xmath107 and @xmath64 @xmath50 diagrams .
the occ 3 and 4 only include few of member clusters , the panels of those two occs indicate no clustering of the member clusters can be identified .
[ occs ] indicates that only young occs can be identified in the @xmath106 @xmath107 diagram , and it is difficult to detect clustering in ocs with the @xmath64 @xmath50 diagram .
more recently , using the radial velocity , proper motion , inclination , and galactic latitude of 341 ocs with age less than 100 myr and within 2.5 kpc from the sun , @xcite studied the clustering in those clusters .
most of member clusters in the closest occ detected by @xcite have age less than 30 myr , which is consistent with that we have found in the @xmath106 @xmath107 diagram of occ 1 .
@xcite pointed out that 20 myr is the characteristic timescale of an occ , after this timescale , the complex may no longer be recognizable in the space of the orbital elements as the majority of its member have evaporated ; only in the context of corotation resonances or mergers , the nongenetically related dynamical groups of old ocs can be detected in the galactic disc .
we have presented a sample of 488 ocs whose distances , radial velocities , and absolute proper motions are used to derive their orbits in the galaxy .
the kinematical and the orbital characteristics of this sample are analyzed .
the main results are listed as following : * for ocs with errors in the spatial velocities less than 20 km s@xmath6 in our present sample , the velocity ellipsoids are derived as ( @xmath0 , @xmath1 , @xmath110(28.7 , 15.8 , 11.0 ) km s @xmath6 .
the ages for most of clusters in our present sample are less than 500 myr , and this sample represents a young thin disc population .
the velocity dispersions of ocs in the three velocity components increase with the age of the cluster subsample , which indicates the continuous heating of the disc ( n04 ) . *
the orbits of ocs are calculated with three galactic gravitational potential models .
considering the intrinsic variation of orbital parameters and the effects of observational uncertainties , the errors for the derived orbital parameters are determined with a monte carlo fashion.the major errors come from the observational uncertainties of the input data and are mainly affected by the errors in the distance data .
the observational uncertainties mainly affect the derived orbital eccentricities @xmath50 and the maximum distances above the galactic disc @xmath54 .
* the total masses within given distance range for different galactic models can affect the derived orbital parameter @xmath52 and the orbital period @xmath53 .
the disc models for different galactic models mainly affect the vertical movement of clusters and change the derived @xmath54 and @xmath55 .
the uncertainties in the derived orbital parameters caused by different models are smaller than those caused by observational errors in the input data . *
the detection probability for a cluster at the given galactocentric distance is calculated and the largest detection probability is at the cluster s apogalacticon . for most of ocs in our present sample ,
their present observed positions are very close to their apogalacticons .
the mean of orbital period for ocs is about seven times longer than the time for clusters crossing the galactic plane . *
based on the presently observed galactocentric distances , the radial metallicity gradient for clusters with @xmath75 kpc is derived with a slope @xmath74 dex kpc@xmath6 , which is consistent with the previous studies @xcite .
the radial metallicity gradient derived based on apogalactic distances for the same sample is @xmath111 dex kpc@xmath6 .
* the orbital eccentricities @xmath50 for different populations : globular clusters , disc giants , disc f and g dwarf stars , and ocs are compared .
the orbital eccentricities for globular clusters occupy a large range and with a mean of @xmath83 .
the mean of orbital eccentricities for the thick disc giants is @xmath112 which is bigger than the mean of @xmath113 for the thin disc giants , f and g dwarfs , and ocs .
there are about 3.7% clusters in our sample belonging to the thick disc . * * using the @xmath106 @xmath107 diagram , only one occ could be identified and most of its members are younger than 30 myr .
we find it difficult to detect any occs in the @xmath64 @xmath50 diagram .
we are grateful to an anonymous referee for a particularly constructive and prompt report .
this work has been supported in part by the national natural science foundation of china , no.10633020 , 10603006 , 10778720 , 10873016 , and 10803007 ; and by national basic research program of china ( 973 program ) no .
w. acknowledges support from the knowledge innovation program of the chinese academy of sciences . this research has made use of the webda database , operated at the institute for astronomy of the university of vienna . alessi b. s. , moitinho a. , dias w. s. , 2003 , a&a , 410 , 565 allen c. , martos m. a. , 1988 , rev .
mexicana astron .
astrof . , 16 , 25 allen c. , santilln a. , 1991 , rev .
mexicana astron .
astrof . , 22 , 255 allen c. , moreno e. , pichardo b. , 2006 , apj , 652 , 1150 allen c. , moreno e. , pichardo b. , 2008 , apj , 674 , 237 barbier - brossat m. , figon p. , 2000 , a&as , 142 , 217 barkhatova k. a. , kutuzov s. a. , osipkov l. p. , 1987
, soviet astronomy , 31 , 501 barkhatova k. a. , osipkov l. p. , kutuzov s. a. , 1989 , soviet astronomy , 33 , 596 baumgardt h. , dettbarn c. , wielen r. , 2000 , a&as , 146 , 251 bergond g. , leon s. , guibert j. , 2001 , a&a , 377 , 462 bonatto c. , kerber l. o. , bica e. , santiago b. x. , 2006 , a&a , 446 , 121 bovy j. , hogg d. w. , roweis s. t. , 2009 , arxiv:0905.2980 [ astro - ph ] carraro g. , chiosi c. , 1994 , a&a , 288 , 751 chen l. , hou j. l. , wang j. j. , 2003 , aj , 125 , 1397 chen l. , hou j. l. , zhao j. l. , de grijs r. , 2007 , in jin w. j. , platais i. , perryman m. a. c. , eds , proc . iau symp .
248 , a giant step : from milli- to micro - arcsecond astrometry .
kluwer , dordrecht , p. 433 dehnen w. , binney j. , 1998 , mnras , 298 , 387 de la fuente marcos r. , de la fuente marcos c. , 2008 , apj , 672 , 342 de oliveira m. r. , fausti a. , bica e. , dottori h. , 2002 , a&a , 390 , 103 dias w. s. , lpine j. r. d. , alessi b. s. , 2001 , a&a , 376 , 441 dias w. s. , lpine j. r. d. , alessi b. s. , 2002 , a&a , 388 , 168 dias w. s. , alessi b. s. , moitinho a. , lpine j. r. d. , 2002 , a&a , 389 , 871 dias w. s. , lpine j. r. d. , 2005 , apj , 629 , 825 dias w. s. , assafin m. , flrio v. , alessi b. s. , lbero v. , 2006 , a&a , 446 , 949 dinescu d. i. , girard t. m. , van altena w. f. , 1999 , aj , 117 , 1792 efremov y. n. , 1978 , soviet astronomy letters , 4 , 66 igenson a. m. , yatsyk o. s. , 1988 , soviet astronomy , 32 , 168 esa , 1997 , the hipparcos and tycho catalogues , esa sp-1200 finlay j. , noriega - crespo a. , friel e. d. , cudworth k. , 1995 , bull . am .
soc . , 27 , 1437 flynn c. , sommer - larsen j. , christensen p. r. , 1996 , mnras , 281 , 1027 friel e. d. , 1995 , ara&a , 33 , 381 friel e. d. , 1999 , ap&ss , 265 , 271 friel e. d. , et al . , 2002 , aj , 124 , 2693 frinchaboy p. m. , majewski s. r. , 2008 , apj , 136 , 118 fu j. , hou j. l. , yin j. , chang r. x. , 2009 , apj , 696 , 668 holmberg j. , nordstrm b. , andersen j. , 2007 , a&a , 475 , 519 hg e. , et al . , 2000 ,
a&a , 355 , l27 janes k. a. , 1979 , apjs , 39 , 135 janes k. a. , tilley c. , lyng g. , 1988 , aj , 95 , 771 johnson d. r. h. , soderblom d. r. , 1987 , aj , 93 , 864 jrgensen b. r. , lindegren l. , 2005 , a&a , 436 , 127 keenan d. w. , innanen k. a. , house f. c. , 1973 , aj , 78 , 173 kerr f. j. , lynden - bell d. , 1986 , mnras , 221 , 1023 kharchenko n. v. , 2001 , kinematics and physics of celestial bodies , 17 , 409 kharchenko n. v. , pakulyak l. k. , piskunov a. e. , 2003 , astron . rep . , 47 , 263 kharchenko n. v. , piskunov a. e. , rser s. , schilbach e. , scholz r. -d . , 2005 , a&a , 438 , 1163 kharchenko n. v. , piskunov a. e. , 2006 , astron .
25 , 177 kharchenko n. v. , scholz r. -d . , piskunov a. e. , rser s. , schilbach e. , 2007 , astron .
notes , 328 , 889 kovalevsky j. , et al . , 1997 , a&a , 323 , 620 lada c. j. , lada e. a. , 2003 , ara&a , 41 , 57 lpine j. r. d. , dias w. s. , mishurov y. , 2008 , mnras , 386 , 2081 loktin a. v. , beshenov g. v. , 2003 , astron . rep . , 47 , 6 lyng g. , 1982 , a&a , 109 , 213 lyng g. , 1987 , catalogue of open cluster data , 5th edition , cds , strasbourg lyng g. , palou j. , 1987 , a&a , 188 , 35 magrini l. , sestito p. , randich s. , galli d. , 2009 , a&a , 494 , 95 mermilliod j. c. , mayor m. , udry s. , 2008 , a&a , 485 , 303 , miyamoto m. , nagai r. , 1975 , pasj , 27 , 533 nordstrm b. , et al .
, 2004 , a&a , 418 , 989 odenkirchen m. , brosche p. , 1992
, 313 , 69 odenkirchen m. , brosche p. , geffert m. , tucholke h. -j . , 1997 , new .
astron . , 2 , 477 paczyski b. , 1990 , apj , 348 , 485 paunzen e. , netopil m. , 2006 , mnras , 371 , 1641 piskunov a. e. , kharchenko n. v. , rser s. , schilbach e. , scholz r. -d . , 2006 , a&a , 445 , 545 platais i. , kozhurina - platais v. , van leeuwen f. , 1998 , aj , 116 , 2423 [ plummer h. c. , 1911 , mnras , 71 , 460 press w. h. , teukolsky s. a. , vetterling w. t. flannery b. p. , 1992
, numerical recipes in fortran : the art of scientific computing , 2rd edn .
cambridge univ . press , cambridge reid m. j. , 1993 , ara&a , 31,345 rser s. , kharchenko n. v. , piskunov a. e. , schilbach e. , scholz r .- d . , 2007 , in vesperini e. , giersz m. ,
sills a. , eds , proc .
246 , dynamical evolution of dense stellar systems .
kluwer , dordrecht , p. 115 saio h. , yoshii y. , 1979 , pasp , 91 , 553 seabroke g. m. , gilmore g. , 2007 , mnras , 380 , 1348 soubiran c. , odenkirchen m. , le campion j .- f . , 2000 ,
a&a , 357 , 484 soubiran c. , bienaym o. , mishenina t. v. , kovtyukh v. v. , 2008 , a&a , 480 , 91 twarog b. a. , ashma k. m. , anthony - twarog b. j. , 1997 , aj , 114 , 2556 urban s. e. , wycoff g. l. , makarov v. v. , 2000 , aj , 120 , 501 van leeuwen f. , 2008 , hipparcos , the new reduction of the raw data , assl , springer van leeuwen f. , 2007 , a&a , 474 , 653 wu z. y. , tian k. p. , balaguer - nez l. , jordi c. , zhao l. , guibert j. , 2002 , a&a , 381 , 464 zacharias n. , et al . , 2004 ,
aj , 127 , 3043 | we present and analyze kinematics and orbits for a sample of 488 open clusters in the galaxy . the velocity ellipsoid for our present sample
is derived as ( @xmath0 , @xmath1 , @xmath2=@xmath3 , @xmath4 , @xmath5 ) km s@xmath6 which represents a young thin disc population .
we also confirm that the velocity dispersions increase with the age of cluster subsample .
the orbits of open clusters are calculated with three galactic gravitational potential models .
the errors of orbital parameters are also calculated considering the intrinsic variation of the orbital parameters and the effects of observational uncertainties .
the observational uncertainties dominate the errors of derived orbital parameters .
the vertical motions of clusters calculated using different galactic disc models are rather different .
the observed radial metallicity gradient of clusters is derived with a slope of @xmath7 dex kpc@xmath6 .
the radial metallicity gradient of clusters based on their apogalactic distances is also derived with a slope of @xmath8 dex kpc@xmath6 .
the distribution of derived orbital eccentricities for open clusters is very similar to the one derived for the field population of dwarfs and giants in the thin disc .
galaxy : disc open cluster and associations : general galaxy : kinematics and dynamics |
understanding the relationship between the potential energy surface ( pes ) , or energy landscape , and the dynamics of a complex system is a major research effort in the chemical physics community .
one particular focus has been the dynamics of relaxation from an out - of equilibrium starting configuration down the energy landscape to the state of lowest free energy , which is often also the global minimum of the pes.@xcite the possible relaxation processes involved can be roughly divided into two types .
the first is relaxation from a high - energy disordered state to a low - energy ordered state , and examples include the folding of a protein from the denatured state to the native structure , and the formation of a crystal from the liquid . the second kind of relaxation process , which is the focus of the work here , is relaxation from a low - energy , but metastable , state to the most stable state .
this second situation often arises from the first ; the initial relaxation from a disordered state can lead to the population of a number of low - energy kinetically accessible configurations .
the time scales for this second relaxation process can be particularly long because of the large free energy barriers that can separate the states .
some proteins provide an instance of this second type of relaxation . often , as well as a rapid direct path from the denatured state to the native structure , there is also a slower path which passes through a low - energy kinetic trap.@xcite as this trapping is a potential problem for protein function , the cell has developed its own biochemical machinery to circumvent it .
for example , it has been suggested that the chaperonin , groel , aids protein folding by unfolding those protein molecules which get stuck in a trapped state.@xcite there are also a growing number of examples of this second type of relaxation involving clusters . for lennard - jones ( lj ) clusters
there are a number of sizes for which the global minimum is non - icosahedral.@xcite for example , for lj@xmath2 the global minimum is a face - centred - cubic truncated octahedron , but relaxation down the pes almost always leads to a low - energy icosahedral minimum .
similarly , for lj@xmath3 the icosahedral states act as a trap preventing relaxation to the marks decahedral@xcite global minimum .
a similar competition between face - centred - cubic or decahedral and icosahedral structures has recently been observed for metal clusters.@xcite for these clusters unbiased global optimization is difficult because the icosahedral states act as an effective trap .
more generally , kinetic traps are one of the major problems for a global optimization algorithm .
therefore , much research has focussed on decreasing the ` life - time ' of such traps .
for example , some methods simulate a non - boltzmann ensemble that involves increased fluctuations , thus making barrier crossing more likely.@xcite other algorithms transform the energy landscape in a way that increases the temperature range where the global minimum is populated , thus allowing one to choose conditions where the free energy barriers relative to the thermal energy are lower.@xcite recently , clear examples of trapping associated with structural transitions have emerged from experiments on nacl clusters .
these clusters have only one energetically favourable morphology : the magic numbers that appear in mass spectra correspond to cuboidal fragments of the bulk crystal ( rocksalt ) lattice,@xcite hence the term nanocrystals .
indirect structural information comes from the experiments of jarrold and coworkers which probe the mobility of size - selected cluster ions . for most ( nacl)@xmath4cl@xmath1 with @xmath5 ,
multiple isomers were detected which were assigned as nanocrystals with different cuboidal shapes.@xcite the populations in the different isomers were not initially equilibrated , but slowly evolved , allowing rates and activation energies for the structural transitions between the nanocrystals to be obtained.@xcite in a previous paper we identified the mechanisms of these structural transitions by extensively searching the low - energy regions of the pes of one of these clusters , ( nacl)@xmath0cl@xmath1 , in order to obtain paths linking the different cuboidal morphologies.@xcite the key process in these transitions is a highly cooperative rearrangement in which two parts of the nanocrystal slip past one another on a @xmath6 plane in a @xmath7110@xmath8 direction .
here we continue our examination of the structural transitions by investigating the dynamics of ( nacl)@xmath0cl@xmath1 . given the long time scales for which the clusters reside in metastable forms , it is not feasible to probe the transitions with conventional dynamics simulations
. instead , we use a master equation that describes the probability flow between the minima on the pes .
this method has the advantage that we can relate the dynamics to the topography of the pes . in this paper
we are particularly concerned with obtaining activation energies for the structural transitions : firstly , in order to compare with experiment , and secondly , to understand how the activation energy for a process that involves a series of rearrangements and a large number of possible paths is related to the features of the energy landscape . in section [
sect : methods ] we outline our methods , and then in section [ sect : results ] , after a brief examination of the topography of the pes and the thermodynamics , we present our results for the dynamics of the structural transitions .
the potential that we use to describe the sodium chloride clusters is the tosi - fumi parameterization of the coulomb plus born - mayer form@xcite : @xmath9 where @xmath10 is the charge on ion @xmath11 , @xmath12 is the distance between ions @xmath11 and @xmath13 and @xmath14 and @xmath15 are parameters@xcite .
although simple , this potential provides a reasonable description of the interactions .
for example , in a previous study we compared the global minima of ( nacl)@xmath16cl@xmath1 ( @xmath17 ) for this potential with those for a more complex potential derived by welch _
et al . _ which also includes terms due to polarization.@xcite most of the global minima
were the same for both potentials.@xcite given some of the other approximations we use in this study , the small advantages gained by using the welch potential do not warrant the considerable additional computational expense .
we should also note that a well - known problem associated with the above family of potentials for the alkali halides is that they never predict the cscl structure to be the most stable .
this problem arises because the potentials do not allow the properties of an ion to be dependent on the local ionic environment.@xcite this deficiency should not greatly affect the relative energies of the low - lying ( nacl)@xmath0cl@xmath1 minima because they all have the same rock - salt structure , but it may affect the barriers for rearrangements where some ions experience a different local environment at the transition state .
the samples of 3518 minima and 4893 transition states that we use here were obtained in our previous study on the mechanisms of the structural transitions for ( nacl)@xmath0cl@xmath1.@xcite this sampling was performed by repeatedly stepping across the pes from minimum to minimum via transition states , thus giving a connected set of stationary points .
we biased this search to probe the low - energy regions of the pes either by using a metropolis criterion@xcite to decide whether to accept a step to a new minimum@xcite or by systematically performing transition state searches from the lower energy minima in the sample.@xcite thus , although our samples of stationary points only constitute a tiny fraction of the total number on the pes , we have a good representation of the low - energy regions that are relevant to the structural transitions of ( nacl)@xmath0cl@xmath1 .
we used two methods to probe the thermodynamics : first , conventional monte carlo simulations and second , the superposition method .
the latter is a technique to obtain the thermodynamics of a system from a sample of minima.@xcite it is based on the idea that all of configuration space can be divided up into the basins of attraction surrounding each minimum.@xcite the density of states or partition function can then be written as a sum over all the minima on the pes , e.g. @xmath18 , where @xmath19 is the partition function of minimum @xmath11 .
the limitations of the superposition method are that the @xmath19 are not known exactly and that , for all but the smallest systems , the total number of minima on the pes is too large for us to characterize them all .
however , the harmonic expression for @xmath19 leads to a reasonable description of the thermodynamics.@xcite furthermore , anharmonic forms are available which allow the thermodynamics of larger clusters to be reproduced accurately.@xcite the incompleteness of the sample can be overcome by weighting the contributions from the minima in a representative sample.@xcite however , this approach is not necessary in the present study since we are interested in low temperature behaviour where the number of thermodynamically relevant minima is still relatively small .
furthermore , in this temperature regime the superposition method has the advantage that it is unaffected by large free energy barriers between low - energy minima which can hinder the determination of equilibrium thermodynamic properties by conventional simulation .
here we use the harmonic form of the superposition method , because we later use the harmonic approximation to derive rate constants ( reliable anharmonic expressions for the rate constants are not so readily available ) .
the partition function is then @xmath20 where @xmath21 , @xmath22 is the energy of minimum @xmath11 , @xmath23 is the geometric mean vibrational frequency of @xmath11 , @xmath24 is the number of vibrational degrees of freedom and @xmath25 is the number of permutational isomers of @xmath11 .
@xmath25 is given by @xmath26 , where @xmath27 is the order of the point group of @xmath11 . from this expression
thermodynamic quantities such as the heat capacity , @xmath28 , can be obtained by the application of standard thermodynamic formulae.@xcite the superposition method also allows us to examine the contributions of particular regions of configuration space to the the thermodynamics .
for example , the probability that the system is in region @xmath29 is given by @xmath30 where the sum runs over those minima that are part of region a. the master equation approach@xcite is increasingly being used to describe the inter - minimum dynamics on a multi - dimensional pes with applications to , for example , clusters,@xcite glasses,@xcite proteins@xcite and idealized model landscapes.@xcite the master equation is defined in terms of @xmath31 , the vector whose components are the ensemble - average probabilities that the system is associated with each minimum at time @xmath32 : @xmath33 , \label{eq : dpdt}\ ] ] where @xmath34 is the rate constant for transitions from minimum @xmath13 to minimum @xmath11 .
defining the matrix @xmath35 allows equation ( [ eq : dpdt ] ) to be written in matrix form : @xmath36 .
if the transition matrix @xmath37 can not be decomposed into block form , it has a single zero eigenvalue whose corresponding eigenvector is the equilibrium probability distribution , @xmath38 . as a physically reasonable definition for the rate constants must obey detailed balance at equilibrium , i.e. @xmath39 , the solution of the master equation can be expanded in terms of a complete set of eigenfunctions of the symmetric matrix , @xmath40 , defined by @xmath41 .
the solution is @xmath42 , \label{eq : msolution}\ ] ] where @xmath43 is component @xmath11 of the @xmath44 eigenvector of @xmath40 and @xmath45 is the @xmath44 eigenvalue .
the eigenvalues of @xmath37 and @xmath40 are identical and the eigenvectors are related by @xmath46 . except for the zero eigenvalue ,
all @xmath45 are negative .
therefore , as @xmath47 the contribution of these modes decays exponentially to zero and @xmath48 .
to apply equation ( [ eq : msolution ] ) we must first diagonalize @xmath40 . the computer time required for
this procedure scales as the cube of the size of the matrix and the memory requirements scale as the square .
therefore , it is advantageous for the matrix @xmath40 to be as small as possible .
for this reason we recursively removed those minima that are only connected to one other minimum ; these ` dead - end ' minima do not contribute directly to the probability flow between different regions of the pes .
after pruning our samples have 1624 minima and 2639 transition states . to test the effect of this pruning
, we performed some calculations using both the full and the pruned samples .
the effect on the dynamics of the structural transitions was negligible .
as the temperature of a system is decreased the spread of eigenvalues can increase rapidly . when the ratio of the largest to smallest eigenvalues is of the order of the machine precision of the computer , the accuracy of the extreme eigenvalues can become degraded by rounding errors .
we encountered these problems below 275k . without pruning the samples
these numerical difficulties are more pronounced.@xcite we model the rate constants , which are needed as input to equation ( [ eq : dpdt ] ) , using rrkm theory@xcite in the harmonic approximation .
therefore , the rate constant for a transition from minimum @xmath11 to minimum @xmath13 via a particular transition state ( denoted by @xmath49 ) is given by @xmath50
in our previous study of ( nacl)@xmath0cl@xmath1 we found that the low - energy minima all had rock - salt structures .
the different minima have four basic shapes : an incomplete @xmath51 cuboid , a @xmath52 cuboid with a single vacancy , a @xmath53 cuboid with a single vacancy and an incomplete @xmath54 cuboid .
the lowest - energy minimum for each of these forms is shown in figure [ fig : minima ] . in the experiments on ( nacl)@xmath0cl@xmath1 by jarrold and coworkers
the three peaks that were resolved in the arrival time distribution were assigned on the basis of calculated mobilities as @xmath51 , @xmath54 and @xmath53 nanocrystals.@xcite however , when the @xmath52 nanocrystal is also considered , better agreement between the calculated and observed mobilities can be obtained by assigning the three experimental peaks to the @xmath52 , @xmath51 and @xmath53 nanocrystals in order of increasing drift time.@xcite this reassignment is also in better agreement with the thermodynamics since the clusters convert to ( what is now assigned as ) the @xmath51 nanocrystal as time progresses,@xcite indicating that this structure has the lowest free energy . in our calculations
a @xmath51 isomer is the global potential energy minimum , and the @xmath51 nanocrystal is always more stable than the other nanocrystals ( see section [ sect : therm ] ) . -1.6
cm -2.5 cm disconnectivity graphs provide a way of visualizing an energy landscape that is particularly useful for obtaining insight into dynamics , and have previously been applied to a number of protein models@xcite and clusters.@xcite the graphs are constructed by performing a ` superbasin ' analysis at a series of energies .
this analysis involves grouping minima into disjoint sets , called superbasins , whose members are connected by pathways that never exceed the specified energy . at each energy a superbasin is represented by a node , and lines join nodes in one level to their daughter nodes in the level below .
every line terminates at a local minimum .
the graphs therefore present a visual representation of the hierarchy of barriers between minima .
the disconnectivity graph for ( nacl)@xmath0cl@xmath1 is shown in figure [ fig : disconnect ]
. the barriers between minima with the same cuboidal form are generally lower than those between minima that have different shapes .
therefore , the disconnectivity graph splits into funnels corresponding to each cuboidal morphology .
( a funnel is a set of downhill pathways that converge on a single low - energy minimum or a set of closely - related low - energy minima.@xcite in a disconnectivity graph an ideal funnel is represented by a single tall stem with lines branching directly from it , indicating the progressive exclusion of minima as the energy is decreased.@xcite ) the separation is least clear for the @xmath54 minima because of the large number of different ways that the nine vacant sites can be arranged .
for example , these vacancies are organized very differently in the two lowest - energy @xmath54 minima ( figure [ fig : minima ] ) , and in fact the barrier between minimum o and the low - energy @xmath52 isomers is lower than the barrier between o and minimum l. therefore , the minima associated with o form a sub - funnel that splits off from the @xmath52 funnel , rather than being directly connected to the main @xmath54 funnel .
the disconnectivity graph shows that the barriers between the @xmath51 , @xmath52 and @xmath54 nanocrystals are of similar magnitude , while the @xmath53 minima are separated from the rest by a considerably larger barrier .
the values of some of the barrier heights are given in table [ table : barriers ] .
0.1 cm the disconnectivity graph is also helpful for interpreting the ( nacl)@xmath0cl@xmath1 dynamics observed in experiments.@xcite in the formation process it is likely that a high - energy configuration is initially generated .
the cluster then relaxes to one of the low - energy nanocrystals .
simulations for potassium chloride clusters indicate that this relaxation is particularly rapid for alkali halides because of the low barriers ( relative to the energy difference between the minima ) for downhill pathways.@xcite however , there is a separation of time scales between this initial relaxation and the conversion of the metastable nanocrystals to the one with the lowest free energy .
the large barriers between the different cuboids make them efficient traps . 0.1 cm some thermodynamic properties of ( nacl)@xmath0cl@xmath1 are shown in figures [ fig : mc ] and [ fig : hsm ] .
the caloric curve shows a feature at @xmath55700k which indicates melting ( figure [ fig : mc]a ) ; the melting temperature is depressed relative to the bulk due to the cluster s finite size .
the effects of the barriers between nanocrystals are apparent in the mc simulations .
the radius of gyration , @xmath56 , provides a means of differentiating the nanocrystals .
it can be seen from the plot of @xmath56 for simulations started in the lowest - energy minima of the four cuboidal forms that each simulation is stuck in the starting structure ( figure [ fig : mc]b ) up to temperatures close to melting , implying that there are large free energy barriers between the nanocrystals .
0.1 cm these free energy barriers prevent an easy determination of the relative stabilities of the different nanocrystals by conventional simulations .
therefore , we use the superposition method to examine this question . first we assign the fifty lowest - energy minima to one of the cuboidal forms by visual inspection of each structure .
then , using these sets as definitions of the nanocrystals in equation ( [ eq : pa ] ) , we calculate the equilibrium probabilities of the cluster being in the different cuboidal morphologies as a function of temperature . it can be seen from figure [ fig : hsm]a that the @xmath51 nanocrystal is most stable up until melting .
the @xmath52 nanocrystal also has a significant probability of being occupied .
however , the probabilities for the @xmath53 and @xmath54 nanocrystals are always small .
the onset of the melting transition is indicated by the rise of @xmath57 in figure [ fig : hsm]a and by the peak in the heat capacity ( figure [ fig : hsm]b ) .
however , this transition is much too broad and the heat capacity peak occurs at too high a temperature because the incompleteness of our sample of minima leads to an underestimation of the partition function for the liquid - like minima . given these expected failings of the superposition method at high temperature when the partition functions of the minima are not reweighted , we restrict our dynamics calculations to temperatures below 600k .
although the probabilities of being in the different morphologies show little variation at low temperature , there are significant changes in the occupation probabilities of specific minima .
for example , the small low temperature peak in the heat capacity is a result of a redistribution of probability amongst the low - energy @xmath51 minima ; the third lowest - energy minimum becomes most populated .
it is also interesting to note that the second lowest - energy @xmath54 minimum ( o ) becomes more stable than the lowest - energy @xmath54 minimum ( l ) for temperatures above approximately 220k .
both these changes are driven by differences in vibrational entropy .
some examples of the interfunnel dynamics that we find on solution of the master equation are depicted in figure [ fig : flow ] .
the time scales involved are much longer than those accessible by conventional simulations .
the dynamics of relaxation to equilibrium depend significantly on the starting configuration .
when the lowest - energy @xmath52 minimum is the initial configuration there is a small transient population in @xmath54 minima before the system adopts a @xmath51 structure .
this is consistent with the lowest - energy pathway that was found between the two nanocrystals ; it passes through some intermediate @xmath54 minima.@xcite 0.1 cm 2 the relaxation to equilibrium is much slower when the cluster starts from the lowest - energy @xmath53 minimum .
this is a result of the large barrier to escape from this funnel ( figure [ fig : disconnect ] ) .
the probability flow out of the @xmath53 funnel leads to a simultaneous rise in the occupation probabilities of both the @xmath51 and @xmath52 nanocrystals towards their equilibrium values , even though the lowest - energy pathway out of the @xmath53 funnel directly connects it to the low - energy @xmath52 minima .
this occurs because the time scale for interconversion of these latter two nanocrystals is much shorter than that for escape from the @xmath53 funnel .
however , we do find evidence that the cluster first passes through the @xmath52 minima if we examine the probabilities on a log - scale . at times shorter than that required for local equilibrium between the @xmath51 and @xmath52 minima
the occupation probability for the @xmath52 minima is larger . as the two lowest - energy @xmath54 minima
are well - separated in configuration space we considered relaxation from both these minima . in both cases
, there is a large probability flow into the @xmath52 minima , which is then transferred to the @xmath51 funnel on the same time scale as when initiated in the @xmath52 funnel .
however , the time scale for the build - up of population in the @xmath52 minima depends on the initial configuration .
probability flows more rapidly and directly into the @xmath52 minima when initiated from minimum o , reflecting the low barriers between these minima ( figure [ fig : disconnect ] ) .
for relaxation from minimum l there are two active pathways , leading to an increase in the population of both the @xmath51 and the @xmath52 minima .
the direct path into the @xmath51 funnel has the lower barrier ( table [ table : barriers ] ) but is long ( 96.7 ) , and so has a smaller rate than the path into the @xmath52 funnel , which has a slightly higher barrier ( by 0.05@xmath58ev ) .
the small shoulder in the occupation probability of the @xmath51 minima occurs in the time range when the occupation probability of the @xmath54 minima has reached a value close to zero ( thus reducing the contribution from the direct path ) and when the probability flow out of the @xmath54 minima is only just beginning . the combination of our thermodynamics and dynamics results for ( nacl)@xmath0cl@xmath1 enable us to explain why a peak associated with the @xmath54 cuboids was not observed experimentally .
the @xmath54 minima have a shorter lifetime than the other cuboidal forms and have a low equilibrium occupation probability .
another way to analyse the dynamics is to examine how local equilibration progresses towards the point where global equilibrium has been obtained . to accomplish this we define two minima to be in local equilibrium at the time when @xmath59 is obeyed for all later times . in the present work we set @xmath60 ,
i.e. the two minima are within 1% of equilibrium . using this definition
we can construct equilibration graphs , in which nodes occur when two ( groups of ) minima come into local equilibrium.@xcite we show an example of an equilibration graph in figure [ fig : equil ] .
equilibration first occurs at the bottom of the @xmath52 and @xmath51 funnels between those minima that are connected by low barrier paths , then progresses to minima with the same cuboidal shape but which are separated by larger barriers , and finally occurs between the funnels .
the order of interfunnel equilibrium agrees with the time scales that we observe in the time evolution of the occupation probabilities of the nanocrystals ( figure [ fig : flow ] ) .
minimum o , then minimum l , come into equilibrium with the @xmath52 funnel .
then , the @xmath51 and @xmath52 funnels reach local equilibrium . finally , the @xmath53 funnel reaches equilibrium with the rest of the pes . as one of the major determinants of the time scale required for local equilibrium is the height of the barriers between minima ,
it is unsurprising that the equilibration graph reflects the structure of the disconnectivity graph ( figure [ fig : disconnect ] ) .
0.1 cm in the experiments on ( nacl)@xmath0cl@xmath1 rate constants and activation energies were obtained for the conversion of the @xmath52 and @xmath53 nanocrystals into the @xmath51 nanocrystal.@xcite it would , therefore , be useful if we could extract rate constants for the different interfunnel processes from the master equation dynamics . for a two - state system , where @xmath61 and @xmath62 and @xmath63 are forward and reverse rate constants , respectively , it can be shown that @xmath64 = -(k_+ + k_-)t , \label{eq : twostate}\ ] ] and the equivalent expression for b are obeyed.@xcite this is a standard result for a first - order reaction .
this expression will also hold for the rate of passage between two funnels in our multi - state system if the interfunnel dynamics are the only processes affecting the occupation probabilities of the relevant funnels , and if the interfunnel dynamics cause the occupation probabilities of the two funnels to converge to their equilibrium values . in figure [
fig : first ] we test the above expression by applying it to the interconversion of @xmath52 and @xmath51 nanocrystals .
the two lines in the graph converge to the same plateau value , before both falling off beyond 0.001s .
this plateau corresponds to the time range for which the interfunnel passage dominates the evolution of the probabilities for the two funnels . at shorter times ,
when the occupation probabilities for the two funnels are still close to their initial values , there are many other contributing processes . at longer times
the probabilities are both very close to their equilibrium values , and the slower equilibration with the @xmath53 funnel dominates the probability evolution . from the plateau in figure [ fig : first ]
we obtain @xmath65 .
the individual rate constants can be obtained by using the detailed balance relation : @xmath66 .
0.1 cm the application of equation ( [ eq : twostate ] ) to escape from the @xmath53 funnel also leads to a range of @xmath32 where there is a well - defined plateau .
however , this approach works less well for the interconversion of @xmath54 and @xmath52 minima ( figures [ fig : flow]a and b ) .
this is because the assumption that no other processes contribute to the probability evolution of the two funnels is obeyed less well , and because the occupation probabilities do not converge to their equilibrium values , but near to the equilibrium values that would obtain if the @xmath51 minima were excluded .
nevertheless , approximate values of @xmath67 and @xmath68 can be obtained .
diagonalization of the matrix @xmath40 produces a set of eigenvalues that give the time scales for a set of characteristic probability flows . the dynamical processes to which the eigenvalues correspond can be identified by examining the eigenvectors
flow occurs between those minima for which the corresponding components of the eigenvector have opposite sign ( equation ( [ eq : msolution ] ) ) .
this observation forms the basis for kunz and berry s net - flow index which quantifies the contribution of an eigenvector @xmath11 to flow out of a funnel @xmath29.@xcite the index is defined by @xmath69 the index allows the interfunnel modes to be identified ; the values of @xmath70 and @xmath71 for these modes will be large and of opposite sign . for example , at @xmath72=400k the mode with the most @xmath73 character has @xmath74 , @xmath75 and @xmath76 .
the eigenvalue is in good agreement with the sum of interfunnel rates obtained using equation ( [ eq : twostate ] ) .
the extraction of the interfunnel rate in this manner is hindered by the fact that the eigenvalues of @xmath40 can not cross as a function of temperature .
instead , there are avoided crossings and mixing of modes . for example
, the small difference between the two values for the sum of the @xmath77 interfunnel rate constants that we obtained above is probably due to mixing .
the eigenvector with which the interfunnel mode mixes gains some @xmath73 character ; it has @xmath78 , @xmath79 and @xmath80 . 0.1 cm by calculating and diagonalizing @xmath40 at a series of temperatures we can examine the temperature dependence of the interfunnel rate constants . for most of the interfunnel processes , the rate constants that we obtain fit well to the arrhenius form , @xmath81 , where @xmath82 is the activation energy and @xmath29 is the prefactor ( figure [ fig : barrier ] ) . in our previous study of the interconversion mechanisms in (
nacl)@xmath0cl@xmath1 we estimated the activation energies in order to compare with the experimental values . using the analogy to a simple one - step reaction , we equated the activation energy with the difference in energy between the highest - energy transition state on the lowest - energy path between the relevant nanocrystals and the lowest - energy minimum of the starting nanocrystal.@xcite however , it was not clear how well this analogy would work . in table
[ table : barriers ] we compare these estimates with the activation energies obtained from our master equation results .
there is good agreement , confirming the utility of the approximate approach .
similar agreement has also been found for the interfunnel dynamics of a 38-atom lennard - jones cluster.@xcite a simple explanation for this correspondence can be given .
we first label the minima along the lowest - barrier path between the two funnels in ascending order and define @xmath83 so that the highest - energy transition state on this path lies between minima @xmath84 and @xmath83 . if the minima behind the highest - energy transition state ( i.e. 1 to @xmath84 ) are in local equilibrium , then the occupation probability of the minimum @xmath84 is given by @xmath85 then
, if the rate of interfunnel flow , @xmath86 , is equated to the rate of passage between minima @xmath84 and @xmath83 , @xmath87 therefore , if the occupation probability for funnel a is dominated by the occupation probability for the lowest - energy minimum in the funnel , i.e. @xmath88 for all @xmath72 , then the activation energy is equal to the energy difference between the highest - energy transition state on the lowest - barrier path and the energy of the lowest - energy minimum in the starting funnel .
we should note that in the above derivation the interfunnel probability flow is assumed to all pass through a single transition state .
however , if there is competition between two paths , one with a low barrier and a small prefactor and one with a larger barrier and a large prefactor , we expect that the low - barrier path would dominate at low temperature and the high - barrier path at high temperature
. this behaviour would give rise to a interfunnel rate constant with a positive curvature in an arrhenius plot .
however , the lines are either straight or have a small amount of negative curvature .
the lack of positive curvature , and the agreement between the estimated and the observed activation energies , probably indicates that the interfunnel probability flow is dominated by paths which pass through the highest - energy transition state on the lowest barrier path .
it is interesting that , on a pes with so many minima and transition states , a single transition state can have such a large influence on the dynamics . at low enough temperature , @xmath89 , and
so the interfunnel barrier height can be measured with respect to the lowest - energy minimum in the starting funnel ( equation ( [ eq : barrier ] ) ) .
however , as the occupation probabilities of other minima in the funnel becomes significant relative to that for the lowest - energy minimum in the funnel , the ratio @xmath90 decreases , thus giving the lines in the arrhenius plot their slight negative curvature .
in other words , the apparent activation energy decreases with increasing temperature , because the barrier height should be measured with respect to some kind of average minimum energy for the funnel , perhaps @xmath91 . the negative curvature is most pronounced when the occupation probabilities in a funnel change considerably .
for example , the @xmath92 rate constant has the most curvature because minimum l has a particularly low vibrational entropy leading to population of other minima within that funnel ( figure [ fig : disconnect ] ) . in table
[ table : barriers ] we also compare our activation energies to the two experimental values .
our values are too large by 0.24 and 0.49@xmath58ev .
there are a number of possible sources of error .
firstly , the samples of minima and transition state provide only an incomplete characterization of the pes and so it is possible that the nanocrystals are connected by undiscovered lower - barrier paths .
however , we believe that our sample of minima is a good representation of the low - energy regions of the pes and consider it improbable that undiscovered pathways could account for all of the discrepancy . secondly , in our calculations the input rate constants @xmath93 were calculated on the basis of the harmonic approximation .
although this is likely to have a significant effect on the absolute values of the interfunnel rate constants and prefactors ( the latter are too large compared to the experiment values@xcite ) , it should not have such a significant effect on the activation energies .
instead , we consider the most likely source of the discrepancy between theory and experiment to be inaccuracies in the potential . when polarization is included by using the welch potential , the estimated barriers become closer to the experimental values ( the discrepancies are then 0.16 and 0.34@xmath58ev ) but significant differences still remain.@xcite for better agreement we may need a potential that allows the properties of the ions to depend on the local environment . unfortunately , although such potentials have been developed for a number of systems,@xcite one does not yet exist for sodium chloride .
the ( nacl)@xmath0cl@xmath1 potential energy surface has a multiple - funnel topography . structurally , the different funnels correspond to rocksalt nanocrystals with different cuboidal forms .
the large potential energy barriers between the funnels causes the time scales for escape from metastable nanocrystals to be far longer than those accessible by conventional dynamics simulations .
therefore , we examined the interfunnel dynamics by applying the master equation approach to a database of ( nacl)@xmath0cl@xmath1 minima and transition states . the slowest rate constant we obtained was @xmath94 at @xmath95k .
using a net flow index we were able to identify the eigenvalues of the transition matrix , @xmath96 , which correspond to interfunnel probability flow .
thus , we were able to obtain rate constants and activation energies for the interconversion of the different nanocrystals .
one particularly interesting finding is that the activation energies correspond fairly closely to the potential energy differences between the highest - energy transition state on the lowest - energy path between two nanocrystals and the lowest - energy minimum of the starting nanocrystal .
this is the result one might expect by a simple extrapolation from the dynamics of a simple molecular reaction .
however , it holds despite the multi - step , and potentially multi - path , nature of the interfunnel dynamics .
the question of whether this result is generally true for interfunnel dynamics involving large potential energy barriers or reflects some of the particulars of the ( nacl)@xmath0cl@xmath1 system is an interesting subject for further research .
we already know that this simplification holds for the 38-atom lennard - jones cluster.@xcite j.p.k.d . is the sir alan wilson research fellow at emmanuel college , cambridge .
is grateful to the royal society for financial support .
we would like to thank mark miller for helpful discussions .
.[table : barriers ] activation energies , @xmath82 , and prefactors , @xmath29 , for the interfunnel rate constants obtained from the fits in figure [ fig : barrier ] .
the activation energies are compared to @xmath97 , the difference in energy between the highest - energy transition state on the lowest - energy path directly between the two nanocrystals and the lowest - energy minimum of the starting nanocrystal,@xcite and experimental values for some of the processes.@xcite as probability flow out of minimum l leads to population of both the @xmath52 and @xmath51 nanocrystals we give @xmath97 for both pathways . [ cols="<,<,^,^,^,^ " , ] | in recent experiments on sodium chloride clusters structural transitions between nanocrystals with different cuboidal shapes were detected .
here we presents results for the thermodynamics and dynamics of one of these clusters , ( nacl)@xmath0cl@xmath1 . as the time scales for the structural transitions can be much longer than those accessible by conventional dynamics simulations , we use a master equation to describe the probability flow within a large sample of potential energy minima .
we characterize the processes contributing to probability flow between the different nanocrystals , and obtain rate constants and activation energies for comparison with the experimental values . 2 |
arp 220 is the nearest ( @xmath3 77 mpc ) example of an ultraluminous infrared galaxy ( ulirg ) that supports star formation at extreme levels .
it contains two nuclei separated by 350 pc , both surrounded by massive discs of dense molecular gas ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
radio detections of supernovae at a rate of 13 yr@xmath4 @xcite confirm that huge populations of massive stars are present with an implied star formation rate ( sfr ) of @xmath5 yr@xmath4 .
although arp 220 could contain active galactic nuclei ( agns ) , particularly in the western nucleus , the observed supernova rates indicate that star formation provides a substantial fraction of the power radiated by the nuclei .
the nuclei of arp 220 provide access to the high - intensity mode of star formation in dense molecular media that appears to have been more common in young galaxies .
these types of environments are of special interest from a range of perspectives , including the information they can provide regarding the role of galactic winds , cosmic rays , and magnetic fields in feedback processes that influence galaxy evolution .
previous investigations show that arp 220 is likely to be a hadronic cosmic ray calorimeter where all of the power in cosmic rays is absorbed within the nuclear starburst zones ( e.g. , * ? ? ?
both nuclei also contain extremely intense far - infrared ( fir ) radiation fields ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) , and the west nucleus is optically thick in the fir to wavelengths of @xmath6 @xcite .
the production of the observed radio synchrotron emission then requires magnetic fields of milligauss strength ( e.g. , * ? ? ?
* ; * ? ? ?
llllc physical parameters & east nucleus & west st & west cnd & references + distance & 77.0 mpc & 77.0 mpc & 77.0 mpc & + cmz radius & 70 pc & 90 pc & 30 pc & 1,2,3 + cmz disc scale height@xmath7 & 40 pc & 40 pc & 40 pc & 4 + molecular gas mass & @xmath8 @xmath9 & @xmath10 @xmath9 & @xmath8 @xmath9 & 2,5 + ionized gas mass@xmath11 & @xmath12 @xmath9 & @xmath13 @xmath9 & @xmath12 @xmath9 & + average ism density@xmath14 & @xmath157700 @xmath16 & @xmath153500 @xmath16 & @xmath1542 000 @xmath16 & + fir luminosity & @xmath17 @xmath18 & @xmath17 @xmath18 & @xmath19 @xmath18 & 2 + fir radiation field energy density@xmath20 & 40 000 ev @xmath16 & 27 000 ev @xmath16 & 440 000 ev @xmath16 & + dust temperature & 90 k & 50 k & 170 k & 2,6 + sn explosion rate ( @xmath21 ) & 0.7 yr@xmath4 & 0.7 yr@xmath4 & 1.3 yr@xmath4 & 7 + star formation rate ( sfr)@xmath20 & 65 @xmath9 yr@xmath4 & 65 @xmath9 yr@xmath4 & 120 @xmath9 yr@xmath4 & + sn explosion energy@xmath22 & 10@xmath23 erg & 10@xmath23 erg & 10@xmath23 erg & + sn energy in cosmic ray protons@xmath22 & 5 20% & 5 20% & 5 20% & + ratio of primary protons to electrons ( @xmath24/@xmath25 ) & 50 & 50 & 50 & + slope of primary cosmic ray source function & 2.1 2.3 & 2.1 2.3 & 2.1 2.3 & + + + + + + + + + + in this paper , we study cosmic ray interactions in the arp 220 starburst nuclear regions using an updated version of the @xcite models , hereafter yegz .
we develop a model with two spatial zones to accurately represent the inner and outer regions of the western nucleus as defined by its molecular gas properties @xcite .
we incorporate photopion energy losses and photon photon interactions to account for the extreme fir radiation field .
we calculate the hadronic calorimetry fraction for each nucleus for the best - fitting radio models , and we predict the total @xmath2-ray and neutrino fluxes . in section 2 , we review the physical parameters which we selected for the models .
section 3 details the basic assumptions of the models and our findings for the arp 220 starburst nuclei .
we present concluding remarks in section 4 .
due to its extreme properties , arp 220 has been extensively studied across the electromagnetic spectrum .
the nuclei of arp 220 are of particular interest as they contain more than half of the total bolometric infrared luminosity of the galaxy ( @xmath26 ; e.g. , * ? ? ?
* ; * ? ? ?
* and references therein ) .
as the nuclei are less than 100 pc in radius , the presence of an agn or a ` hot ' starburst is required to explain the extraordinarily large surface brightness in the western nucleus @xcite ; however , the existence of an agn has yet to be definitively established ( e.g. , * ? ? ?
further , the submillimetre observations suggest that whether or not agns are present , they are not the main heating source of the dust ( e.g. , * ? ? ?
estimates of the fir luminosities of the eastern and western nuclei range from @xmath27 to @xmath28 and from @xmath29 to @xmath30 , respectively ( e.g. , * ? ? ?
* ; * ? ? ?
the range on these luminosities is quite large due to uncertainty in the true sizes , inclinations , and opacities of the nuclei and their associated molecular disc . to keep our adopted fir luminosity in rough agreement with the observed supernova rate , we assume values of @xmath31 and @xmath32 for the eastern and western nuclei ( see table 1 ) . assuming similar ratios between the nuclei for the supernova rate and molecular gas content , we adopt values of @xmath33 yr@xmath4 , @xmath34 for the eastern nucleus and @xmath35 yr@xmath4 , @xmath36 for the western nucleus .
while our assumed molecular gas masses favour conservative estimates , other estimates of the gas content suggest the masses are as high as @xmath37 @xcite . [
cols="<,^,^,^,^,^,^,^ " , ] co observations of the western nucleus imply a temperature gradient increasing towards the centre and indicate significant differences in the physical conditions between the two nuclei @xcite . @xcite model the western nucleus as two distinct dust sources a cooler ( 50 k ) ring surrounding a hotter ( 170 k ) , dense dust core .
we use this two - zone model for the western nucleus and have adjusted our single - zone model to account for the differences in temperature and density between the two regions ( see section 3 ) .
for the eastern nucleus , we assume a single dust temperature of 90 k @xcite .
previously , we developed and tested a model for cosmic ray interactions in the central molecular zones ( cmzs ) of star - forming and starburst galaxies ( yegz ; * ? ? ?
* ; * ? ? ?
our single - zone model accounts for a variety of energy losses via interactions with the interstellar medium ( ism ) , magnetic fields , and radiation fields and for energy - independent advective escape via a galactic wind ( see fig .
1 ) . the resulting cosmic ray energy spectrum depends on both the total cosmic ray lifetime and a power - law injection spectrum which is directly proportional to the volume integrated supernova rate ( see yegz for further details ) .
accounting for the production of secondary cosmic rays , we use our calculations of the population of energetic particles to predict the radio , @xmath2-ray , and neutrino spectra . for the @xmath2-ray spectrum
, we include both leptonic ( bremsstrahlung , inverse compton ) and hadronic ( neutral pion decay ) emission mechanisms . for the radio spectrum
, we incorporate the effects of free free emission and absorption @xcite . as in our previous models , we assume that the ionized gas in the nuclei acts as a foreground screen that some fraction ( @xmath38 ) of the emitted synchrotron radiation passes through
. when the covering fraction is low ( @xmath39 ) , the radio spectrum flattens at low frequencies @xcite , and when the covering fraction is high ( @xmath40 ) , the radio spectrum turns down at low frequencies ( yegz ) .
as noted above , the western nucleus in arp 220 is best modelled with two separate regions : an inner circumnuclear disc ( cnd ) with a surrounding torus ( st ) .
we model the cosmic ray populations of the two regions independently , treating each region as a uniform slab .
however , the effects of absorption ( free free and @xmath2@xmath2 ) on the resulting radio and @xmath2-ray emission must be considered more carefully .
absorption occurs within each emission region , and in the case of the inner cnd , absorption also occurs as the emitted radiation moves through the external , st ( see the appendix for further details ) .
we perform @xmath41 tests following the approach described in yegz @xcite . comparing against radio observations for each nucleus , we vary magnetic field strength ( @xmath42 ) , wind speed ( @xmath43 ) , ionized gas density ( @xmath44 ) , and absorption fraction ( @xmath38 ) .
while magnetic field strength and wind speed both directly affect the total cosmic ray lifetimes , the ionized gas density and the absorption fraction only affect the emitted radio spectrum .
the free free emission and absorption coefficients are both directly proportional to the square of @xmath44 , and so , the frequency at which the radio spectrum flattens or turns down and the amount of free free emission at high frequencies both increase with @xmath44 .
observations in @xcite , @xcite and @xcite separate the integrated fluxes of the eastern and western nuclei from the total flux , allowing us to constrain parameters for each nucleus individually . as we do not have radio observations which are separable between the two regions of the western nucleus , we can not constrain the magnetic field strength in each region separately .
we therefore assume that the ratio between the magnetic field strength of the inner and outer regions of the western nucleus is equal to the square root of the ratio of the average gas densities , @xmath45 @xcite . thus , our magnetic field strength determination for the innermost western nucleus is an estimate that is guided by milky way observations .
when assuming the standard 10% cosmic ray acceleration efficiency , we find that agreement between the models and the observed radio data occurs only in a very narrow area of parameter space ( see fig .
the best - fitting models for the nuclei have magnetic field strengths limited to 1.0 mg for the western nucleus ( 3.5 mg in the cnd , estimated from scaling ) and 2.0 2.5 mg for the eastern nucleus ( see figs 2 and 3 and table 2 ) . as seen in fig .
3 , the total radio emission in both nuclei flattens at low frequencies , and in the eastern nucleus , the radio spectrum may be turning over completely .
this flattening of the radio spectra requires moderate to high absorption fractions of 50
100% in the eastern nucleus and low to moderate absorption fractions of 10 70% in the western nucleus .
in addition to moderate absorption fractions in each nucleus , we also find a high contribution from thermal emission to the total radio spectrum ( see fig .
3 ) , particularly in the western cnd where the majority of the radio emission is thermal above @xmath155 ghz . in part , this unusually high fraction of thermal emission is due to the inability of the model to effectively fit for free free absorption and free free emission simultaneously as seen in the eastern nucleus and in previous work ( see yegz ; * ? ? ?
the ability of the models to accurately fit the fraction of thermal emission is further strained by the complicated nature of the western nucleus and the lack of separable radio observations .
thus , in this particular case , the fractions of thermal emission in the best - fitting models have limited significance and do not necessarily contradict observations by @xcite which indicate more modest amounts of thermal emission ( @xmath46 ) .
the western nucleus arp 220 has a very complex structure which we greatly simplified .
while our two - zone density distribution reproduces the observed peak column density of @xmath47 @xmath48 , the mass is lower than that estimated by @xcite who derive the western nucleus gas mass from observations by the atacama large millimeter array ( alma ) of the submillimetre dust luminosity .
this yields a total gas mass of @xmath49 for the western nucleus or four times our adopted value .
we therefore explored the effect of increased gas mass on our model by tripling the mass in the western torus and ran a limited suite models with fixed parameters .
we set the spectral index to @xmath50 and ran @xmath41 tests over the entire range of magnetic field strengths and wind speeds but over a subset of the previously tested ionized gas densities and absorption fractions .
we ran tests on the western nucleus for acceleration efficiencies of 5% and 10% . in comparing the results of these models with a larger gas mass , we find that none of the tested models are within @xmath51 of the best - fitting model at the lower assumed gas mass .
further more , these results yield extremely short cosmic ray electron lifetimes such that the physical validity of the models are in question .
in addition to the higher gas mass estimates , @xcite also propose a geometry where the molecular gas in both nuclei is confined to a thin ( @xmath5210 pc ) disc .
the evolution of supernovae and cosmic ray interactions in this type of high molecular mass structure is beyond the scope of this study which is designed to estimate cosmic ray interaction rates in arp 220 but will need to be considered when arp 220 is detected in @xmath2-rays . in our earlier works
, we demonstrated that for a given ism , the yegz models are highly sensitive to the total flux of cosmic rays ( yegz ; * ? ? ? * ) .
this flux is primarily effected by the original energy input into cosmic rays and the advective time - scale , or escape fraction . the energy input into cosmic rays
is determined by the supernova rate and the assumed acceleration efficiency ( @xmath53 ) . within the uncertainty in the supernova rate , we vary acceleration efficiency from 5 to 20% . as shown above , for the standard 10% efficiency , the resulting best - fitting models are highly constrained in magnetic field strength , and we find that this is also true for an acceleration efficiency of 20% ( see table 2 ) .
however , for a lower acceleration efficiency of 5% , equivalent to a lower supernova rate , we find a much larger range of acceptable fits in the eastern nucleus with magnetic field strengths ranging from 4 to 7.5 mg and wind speeds spanning our entire tested range . as such , the best - fitting models for arp 220 are essentially independent of wind ( advection ) speed ( see figs 1 and 2 ) .
in contrast , a galactic wind was a vital component in modelling the cosmic ray populations of the starburst galaxies m82 and ngc 253 such that an extremely limited range of wind speeds resulted in fits within @xmath51 of the best - fitting models .
the wind speed determines the advective timescale for a galaxy and the fraction of cosmic rays which escape .
thus , wind speed is intrinsically tied to the proton calorimetry fraction for a galaxy which is closely related to the total radio and @xmath2-ray emission from a galaxy . other models for arp 220 have assumed fixed advection time - scales , thus ensuring proton calorimetry with the high gas densities in arp 220 @xcite . while our models agree with others in finding that the starburst regions of m82 and ngc 253 are only @xmath1540
60% proton calorimeters , we find that arp 220 s nuclei are 65 100% ( eastern ) and 90 100% ( western cnd ) proton calorimeters ( see fig .
lcccccc & supernova & average gas & cosmic ray & radiation field & magnetic field & magnetic field + & power & density & energy density & energy density & energy density & strength + & ( erg yr@xmath4 ) & ( @xmath16 ) & ( ev @xmath16 ) & ( ev @xmath16 ) & ( ev @xmath16 ) & ( @xmath54 g ) + milky way & @xmath55 & 1 & 1.4 & 0.3 & 0.9 & 6 + m82 & @xmath56 & 260 & 470 & 490 & 2200 & 300 + arp 220 east & @xmath57 & 7700 & 1100 & 40 000 & @xmath58 & 6500 + arp 220 west cnd & @xmath59 & 42 000 & 2500 & 440 000 & @xmath60 & 7000 + + despite the uncertainty in the calorimetry fraction and the total cosmic ray flux in the eastern nucleus , we can still use our best - fitting models to make a prediction on the emitted @xmath2-ray and neutrino fluxes from arp 220 . to calculate the possible @xmath2-ray flux ,
we apply the parameters of models within @xmath51 from our best - fitting radio model . combining each possible set of models from the eastern and western nucleus , we find that the resulting @xmath2-ray spectra peak around @xmath150.3 gev with a maximum flux of @xmath61 gev @xmath62 s@xmath4 ( see fig
. 4 ) . while this is roughly an order of magnitude lower than previous upper limits for arp 220 @xcite and _ fermi _ s differential sensitivity for four years of observations , it is only a factor of @xmath152 - 3 times smaller than the flux level of the recently detected ngc 2146 @xcite .
we also compared our @xmath2-ray flux with the differential sensitivity 50 h of observations with the future southern cta array ( see fig .
4 ) and find it to be only a factor of a few larger than our maximum flux .
arp 220 may still be detectable by _ fermi _ within the next several years and is a good target for cta , especially for energies near 1 tev .
in addition to making a prediction for the @xmath2-ray spectrum , we can use our same results from the radio emission to predict the neutrino flux from arp 220 .
proton interactions are responsible for the creation of secondary pions , both neutral and charged . while the neutral pions decay into @xmath2-rays , the charged pions decay into a neutrino and a muon which further decays into a secondary electron or positron and two more neutrinos .
the spectrum of the first neutrino from the decay of the charged pion is what we calculate here , as the calculation of the spectra of neutrinos produced during muon decay is more complex ( see * ? ? ?
the flux of our maximum model is roughly @xmath63 gev @xmath62 s@xmath4 at 0.1 pev , and at this energy , the range of possible models spans an order of magnitude in flux ( see fig .
current point source sensitivity limits for the northern sky for icecube are @xmath64 gev @xmath62 s@xmath4 , assuming a spectrum of @xmath65 @xcite .
thus , it seems unlikely that arp 220 will be detected as a point source during a similar time frame by icecube .
however , extreme ulirgs such as arp 220 should make a significant contribution to a diffuse neutrino background @xcite .
in addition to accounting for @xmath2-ray and neutrino emission in arp 220 , we have also take into account the effects of @xmath2@xmath2 absorption due to the intense radiation fields in the nuclei @xcite . at tev energies and above
, @xmath2-rays and infrared photons can interact to produce an electron / positron pair @xcite .
the resulting electrons will be of tev energies and most of their energy will be lost to emission of synchrotron x - rays @xcite .
beginning at @xmath152 5 tev , the opacity for @xmath2@xmath2 absorption in both nuclei is significantly greater than 1 .
this results in a steepening of the predicted @xmath2-ray spectrum at high energies ( see fig .
we find no such increase in slope in the neutrino flux as the steepening is an effect of interactions between the @xmath2-ray and the ambient radiation field and not the cosmic ray proton population .
therefore , in the case of arp 220 and other such ulirgs , the tev @xmath2-ray flux is an unreliable indicator of neutrino flux .
if the effects of spectral steepening by @xmath2@xmath2 absorption are accurate , then arp 220 is unlikely to be detected by cta or other ground based cherenkov telescopes above @xmath1510 tev .
in applying the yegz models to arp 220 , we find that the central starburst regions of arp 220 are moderate to complete cosmic ray proton calorimeters . as such ,
the leptonic cosmic ray population is dominated by secondary electrons and positrons .
the majority of these secondaries are produced at low energies ( e.g. , * ? ? ?
* ) and are likely a major contributor to heating of the ism via ionization @xcite .
based on our best - fitting models for the radio spectrum , we make predictions for both the @xmath2-ray and neutrino fluxes .
our maximum @xmath2-ray spectrum is a factor of a 2 5 lower than previous predictions by @xcite and less than a factor of 2 lower than those by @xcite . while the predicted @xmath2-ray flux will likely be detected by _
fermi _ in the future , under our model assumptions arp 220 is unlikely to be detected as a high energy neutrino point source with the current icecube observatory .
additionally , @xmath2@xmath2 absorption of the tev energy @xmath2-rays make the tev @xmath2-ray flux a poor indicator of the neutrino flux in ulirgs and other such systems with extremely intense infrared radiation fields .
in addition , we find that milligauss strength magnetic fields are still necessary to reproduce the observed radio fluxes from the starburst nuclei , even having assumed larger supernova rates than previous models by factors of 2 5 @xcite .
differences in assumed volume and radiation field energy density across the various models account for the similar best - fitting magnetic field strengths despite the range in assumed supernova rates .
the origins of milligauss strength magnetic fields in extreme starbursts and their impact on the evolution of these systems merit further examination . while the energy density in both magnetic and radiation fields is up from starbursts like m82 by two to three orders of magnitude
, the change in the ratio of their energy densities is up by less than an order of magnitude ( see table 3 ) .
conversely , we see a much larger change in the ratio of magnetic field energy density to cosmic ray energy density .
because the cosmic ray energy density depends on the particle energy loss rate , it does not increase at the same rate as the magnetic and radiation field energy densities ( yoast - hull , gallagher , zweibel , in preparation ) .
thus , the magnetic fields exceed energy equipartition with the cosmic rays by more than two orders of magnitude ( see table 3 ) .
this work was supported in part by nsf ast-0907837 , nsf phy-0821899 ( to the center for magnetic self - organization in laboratory and astrophysical plasmas ) , and nsf phy-0969061 ( to the icecube collaboration ) .
part of this research was carried out during jsg s appointment as a jubileumsprofessor at the chalmers university of technology .
we thank susanne aalto , kazushi sakamoto , dave sanders , nick scoville , and eskil varenius for conversations on arp 220 , justin vandenbroucke and reinhard schlickeiser for discussions regarding the modelling , and francis halzen for his help and support .
additionally , we thank the referee for their helpful comments .
our single - zone model uses a simple solution to the radiative transfer equation of ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) @xmath66 where @xmath67 is the radiative flux prior to absorption , @xmath68 is the radiative flux after absorption , and @xmath69 is the optical depth for either free - free absorption or @xmath2-@xmath2 absorption .
this is still the solution for the eastern nucleus and the surrounding torus in the western nucleus . in the western cnd ,
we must account for a standard emission and absorption region with an additional , external absorbing region .
this observed flux is given by @xmath70 where @xmath67 is still the radiative flux prior to absorption , @xmath71 is the optical depth for @xmath2-@xmath2 or free - free absorption in the emission region , and @xmath72 is the optical depth in the external , absorbing region .
aalto s. , wilner d. , spaans m. , wiedner m. c. , sakamoto k. , black j. h. , caldas m. , 2009 , a&a , 493 , 481 aartsen m. g. , et al .
, 2014a , phys .
lett . , 113 , 101101 aartsen m. g. , et al . , 2014b , apj , 796 , 109 anantharamaiah k. r. , viallefond f. , mohan n. r. , goss w. m. , zhao j. h. , 2000 , apj , 537 , 613 barcos - muoz l. , et al . , 2015 , apj , 799 , 10 bttcher m. , harris d. e. , krawczyski h. , 2012 , relativistic jets from active galactic nuclei .
wiley , weinheim crutcher r. m. , 2012 , ara&a , 50 , 29 dermer c. d. , menon g. , 2009 , high energy radiation from black holes .
princeton univ . press , princeton , nj downes d. , eckart a. , 2007 , a&a , 468 , l57 downes d. , solomon p. m. , 1998 , apj , 507 , 615 draine b. t. , 2011 , physics of the interstellar and intergalactic medium .
princeton univ . press , princeton , nj ghisellini g. , 2013 , vol .
873 : radiative processes in high energy astrophysics .
springer - verlag , berlin lacki b. c. , thompson t. a. , 2013 , apj , 762 , 29 lacki b. c. , thompson t. a. , quataert e. , 2010b , apj , 717 , 1 lacki b. c. , thompson t. a. , quataert e. , loeb a. , waxman e. , 2011 , apj , 734 , 107 lisenfeld u. , vlk h. j. , xu c. , 1996 , a&a , 314 , 745 lonsdale c. j. , diamond p. j. , thrall h. , smith h. e. , lonsdale c. j. , 2006 , apj , 647 , 185 mundell c. g. , ferruit p. , pedlar a. , 2001 , apj , 560 , 168 murase k. , ahlers m. , lacki b. c. , 2013 , phys .
d , 88 , 121301 papadopoulos p. p. , 2010 , apj , 720 , 226 papadopoulos p. p. , isaak k. , van der werf p. , 2010
, apj , 711 , 757 papadopoulos p. p. , thi w .- f . , miniati f. , viti s. , 2011 , mnras , 414 , 1705 parra r. , conway j. e. , diamond p. j. , thrall h. , lonsdale c. j. , lonsdale c. j. , smith h. e. , 2007 , apj , 659 , 314 rodrguez - rico c. a. , goss w. m. , viallefond f. , zhao j .- h . , gomez y. , anantharamaiah k. r. , 2005 , apj , 633 , 198 rovilos e. , diamond p. j. , lonsdale c. j. , smith h. e. , lonsdale c. j. , 2005 , mnras , 359 , 827 rybicki g. b. , lightman a. p. , 1979
, radiative processes in astrophysics .
wiley , new york sakamoto k. , et al .
, 2008 , apj , 684 , 957 scoville n. , et al . , 2015 , apj , 800 , 70 smith h. e. , lonsdale c. j. , lonsdale c. j. , diamond p. j. , 1998 , apj , 493 , l17 soifer b. t. , et al . , 2000 ,
aj , 119 , 509 tang q .- w . , wang x .- y . , tam p .- h .
t. , 2014 , apj , 794 , 26 torres d. f. , 2004 , apj , 617 , 966 tunnard r. , et al . , 2015 , apj , 800 , 25 wilson c. d. , rangwala n. , glenn j. , maloney p. r. , spinoglio l. , pereira - santaella m. , 2014 , apj , 789 , l36 yoast - hull t. m. , everett j. e. , gallagher j. s. , iii , zweibel e. g. , 2013 , apj , 768 , 53 ( yegz ) yoast - hull t. m. , gallagher j. s. , iii , zweibel e. g. , everett j. e. , 2014a , apj , 780 , 137 yoast - hull t. m. , gallagher j. s. , iii , zweibel e. g. , 2014b , apj , 790 , 86 | the cores of arp 220 , the closest ultraluminous infrared starburst galaxy , provide an opportunity to study interactions of cosmic rays under extreme conditions . in this paper , we model the populations of cosmic rays produced by supernovae in the central molecular zones of both starburst nuclei .
we find that @xmath0 of cosmic rays are absorbed in these regions due to their huge molecular gas contents , and thus , the nuclei of arp 220 nearly complete proton calorimeters .
as the cosmic ray protons collide with the interstellar medium , they produce secondary electrons that are also contained within the system and radiate synchrotron emission . using results from @xmath1 tests between the model and the observed radio spectral energy distribution
, we predict the emergent @xmath2-ray and high - energy neutrino spectra and find the magnetic field to be at milligauss levels . because of the extremely intense far - infrared radiation fields , the @xmath2-ray spectrum steepens significantly at tev energies due to @xmath2@xmath2 absorption .
neutrinos cosmic rays galaxies : individual : arp 220 galaxies : starburst gamma rays : galaxies radio continuum : galaxies |
the concepts of @xmath8-scaling @xcite and superscaling ( based on @xmath9-scaling variable ) @xcite have been used for extensive analyses of the vast amount of inclusive electron scattering world data ( see also @xcite ) .
these analyses showed the existence of high - momentum components in the nucleon momentum distribution ( md ) @xmath6 at momenta @xmath10 @xmath11 due to the presence of nucleon - nucleon ( nn ) correlations .
scaling of the first kind ( _ i.e. _ no dependence on the momentum transfer ) can be observed at excitation energies below the quasielastic ( qe ) peak .
scaling of second kind ( _ i.e. _ no dependence on the mass number ) is excellent in the same region .
when scaling of first and second type occur one says that superscaling takes place .
it was shown ( _ e.g. _ , in @xcite ) that the physical reason for the superscaling phenomena is the specific high - momentum tail of @xmath6 which is similar for all nuclei .
it has been shown @xcite that due to the contribution introduced by inelastic scattering together with the correlation contribution and meson exchange currents @xcite both scaling of the first and , to a lesser extent , of the second kind are violated at energies above the qe peak .
the theoretical concept of superscaling has been introduced in @xcite using the properties of the relativistic fermi gas ( rfg ) model .
the fermi momentum for the rfg was used as a physical scale to define the proper scaling variable @xmath12 for each nucleus . as emphasized in @xcite , however , the actual dynamical physics content of the superscaling phenomenon is more complex than that provided by the rfg model . in particular , the extension of the superscaling property to large negative values of @xmath12 ( @xmath1 )
is not predicted by the rfg model . the qe scaling function in the rfg model @xmath13 for @xmath14 , whereas the experimental scaling function @xmath15 has been observed for large negative values of @xmath12 up to @xmath16 in the data for @xmath17 processes .
thus , it has been necessary to consider the superscaling in theoretical approaches which go beyond the rfg model .
one of them is the coherent density fluctuation model ( cdfm ) ( _ e.g. _ , @xcite ) which gives a natural extension of the fermi gas case to realistic finite nuclear systems .
it was shown in @xcite that in the cdfm both basic quantities , density and momentum distributions , are responsible for the scaling and superscaling phenomena in nuclei .
the results of the cdfm for the qe scaling function @xmath0 agree with the available experimental data at different transferred momenta and energies below the qe peak position , showing superscaling for @xmath18 , including @xmath19 and going well beyond the rfg model . secondly ,
as pointed out in @xcite , the nucleon momentum distribution for various nuclei obtained in @xcite ( and with the modification in @xcite ) within a theoretical approach based on the light - front dynamics ( lfd ) method ( _ e.g. _ , @xcite ) can also be used to describe both @xmath8- and @xmath12-scaling data .
the superscaling analyses of inclusive electron scattering from nuclei ( for energies of several hundred mev to a few gev ) have been extended in @xcite to include not only qe processes but also the region where @xmath20 excitation dominates .
both qe scaling functions @xmath15 and @xmath21 in the delta region were deduced from phenomenological fits to electron - nuclei scattering data .
generally , the specific features of the scaling function should be accounted for by reliable microscopic calculations that take final - state interactions into account . in particular , the scaling function @xmath22 with asymmetric shape obtained in @xcite by using a relativistic mean field ( rmf ) for the final states agrees well with the experimental scaling function .
the features of the superscaling phenomenon in inelastic electron scattering have induced studies of neutrino scattering from nuclei on the same basis .
the neutrino - nucleus interactions have been studied using the superscaling analyses of few - gev inclusive electron scattering data in a method proposed in @xcite to predict the inclusive @xmath23 and @xmath24 cross sections for the case of @xmath25c in the nuclear resonance region .
various other theoretical considerations ( _ e.g. _ , @xcite ) have been devoted to studies of both neutral- ( _ e.g. _ , @xcite ) and charge - changing ( _ e.g. _ , @xcite ) neutrino - nucleus scattering .
the cdfm and the lfd method have been extended in @xcite from the qe to the delta - excitation region of the inclusive electron scattering and the qe scaling functions calculated in both methods have been used to calculate and to predict charge - changing neutrino - nucleus cross sections of the @xmath26 and @xmath27 reactions on @xmath25c at energies from 1 to 2 gev .
the asymmetry in the cdfm qe scaling function has been introduced in a phenomenological way .
these analyses make it possible to gain information about the nucleon correlation effects on both nucleon momentum and local density distributions .
it became clear that only the detailed knowledge of the behavior of @xmath6 at high momenta in realistic nuclear systems could lead to quantitative agreement with the experimental scaling function .
on the other hand , the behavior of the latter gives a valuable information about the nn correlation effects on the tail of the momentum distribution .
so , it was shown in @xcite within the cdfm that the @xmath8- and @xmath28-scaling data are informative for @xmath6 at momenta up to @xmath29@xmath30 @xmath11 and it was concluded that further experiments are necessary in studies of the high - momentum components of the nucleon momentum distributions .
the aim of the present work is to consider in more detail the connection between the nn forces in nuclear medium and their effect on the components of @xmath6 from one side and , from the other side , the role of @xmath6 on the behavior of the qe scaling function . for this purpose
we use firstly the md in a hard - sphere dilute fermi gas model ( hsdfg ) ( _ e.g. _ , @xcite ) to calculate the scaling function .
secondly , we make an attempt to throw light on the connection between the generally established high - momentum asymptotics of @xmath6 @xcite and the qe scaling function .
the latter makes it possible to establish ( at least approximately ) the particular form of the power - law decrease of @xmath6 at large values of @xmath4 .
this makes it possible to extract additional information about the nn forces from the description of the superscaling phenomenon . the theoretical scheme and
the results of calculations are given in section [ sect2ant ] .
the conclusions are summarized in section [ sect3ant ] .
in the first part of this section we consider the hard - sphere dilute fermi gas , _
i.e. _ the low - density fermi gas whose particles interact via a repulsive hard - core potential ( see , _
e.g. _ , @xcite ) , and use the nucleon momentum distribution in such a system to calculate the scaling function @xmath31 .
the quantities of interest in the hsdfg model can be expanded in powers of the parameter @xmath32 , where @xmath33 ( @xmath34 ) denotes the hard - core radius in the nn interactions or it is identified with the scattering length in free space , and @xmath35 is the fermi momentum . in @xcite
the value of @xmath32 is adopted to be equal to @xmath36 which corresponds to nn core radius of @xmath37 fm and the typical value of the fermi momentum @xmath38 @xmath11 . as was pointed out by migdal
@xcite , @xmath6 in the normal fermi gas is discontinuous at the fermi momentum .
the analytical expressions for the dimensionless @xmath6 in the hsdfg obtained in @xcite have the form : @xmath39 at @xmath40 : @xmath41,\end{aligned}\ ] ] where @xmath42 and @xmath43 @xcite is adopted . at @xmath44 : @xmath45\bigg\}.\end{aligned}\ ] ] at @xmath46 : @xmath47\bigg\}.\end{aligned}\ ] ] at @xmath48 : @xmath49\bigg\}.\end{aligned}\ ] ] the momentum distribution in the hsdfg model is presented in fig .
[ fig1ant ] for the value of @xmath50 .
the momentum distribution @xmath6 in hard - sphere dilute fermi gas @xcite as a function of @xmath51.,width=302 ] following the definition of the @xmath52-scaling function given by barbaro _
et al . _
@xcite , one can obtain for the case of the hsdfg : @xmath53 where @xmath54 , @xmath55 being the nucleon mass , and @xmath56\big[2+\psi'^2\big(\sqrt{1+\eta_f^2}-1\big)\big]\big\}^{1/2}.\ ] ] in eqs .
( [ eq6ant ] ) and ( [ eq7ant ] ) the dimensionless scaling variable @xmath57 ( in units of the fermi energy ) has the physical meaning of the smallest kinetic energy which one of the nucleons responding to an external probe can have @xcite . since @xmath58 we write eq .
( [ eq6ant ] ) as : @xmath59 it can be seen from eq .
( [ eq8ant ] ) that , as expected , the hsdfg system under consideration also @xmath60 @xmath61 .
( [ eq8ant ] ) has been obtained using the approximations following @xcite : * the fermi momentum distribution of the initial nucleon in the nucleus @xmath62 is replaced by the md ( with dimension ) , namely @xmath63 , where @xmath64 is the spectral function , and * for @xmath65 the step function @xmath66 is retained to take into account approximately pauli blocking for the final nucleon . in figs .
[ fig2ant ] and [ fig3ant ] we give the results for the hsdfg scaling function ( in logarithmic and linear scale , respectively ) calculated for different values of @xmath32 from @xmath36 to @xmath67 and compared to the result for the scaling function in the rfg model .
one can see that the hsdfg scaling function is extended for large negative values of @xmath12 in contrast to the case of the rfg scaling function , but there is not a good agreement with the experimental data .
one can see in the figures also the step behavior of the scaling function which reflects the discontinuity of @xmath6 at @xmath68 .
due to these reasons we consider in the second part of this section the relation between the asymptotic behavior of the md and the @xmath28-scaling function .
the scaling function @xmath0 in hsdfg calculated for different values of @xmath69 in comparison with the rfg model result .
the grey area shows experimental data taken from @xcite.,width=302 ] the same as in fig .
[ fig2ant ] but a linear scale for @xmath70.,width=302 ] a relationship between the nn force in nuclear medium @xmath71 and the asymptotic behavior of the momentum distribution @xmath6 was derived with great generality in @xcite . in these articles
amado and woloshyn showed that the asymptotics of @xmath6 at large values of @xmath4 is a power - law decrease : @xmath72{}\bigg[\frac{\widetilde{v}_\text{nn}(k)}{k^2}\bigg]^2,\ ] ] where @xmath73 is the fourier transform of the nn interaction @xmath74 . in the case of delta - forces
( _ i.e. _ in the hsdfg ) the asymptotics is @xmath75 @xcite .
it is still unknown if @xmath4 or @xmath76 must be large for eq .
( [ eq9ant ] ) to apply @xcite . in principle , it is shown in @xcite that @xmath6 in the hsdfg decreases faster than @xmath77 , typically like @xmath78 , where @xmath79 . in fig .
[ fig4ant ] we show @xmath80 as a function of @xmath81 .
it can be seen that the hsdfg @xmath6 decreases approximately like @xmath78 with a small value of @xmath82 , because the result for @xmath80 is almost a straight line which decreases slowly with the increase of @xmath83 . the momentum distribution in hsdfg @xmath84 multiplied by @xmath85.,width=302 ] further in this work we study the question about the general feature of the nn force @xmath71 that results in an @xmath6 with a power - law behavior that best agrees with the scaling function .
to this aim we assume an nn - potential @xmath74 different from a delta - function and calculate the scaling function using different asymptotics for @xmath6 in the dilute fermi gas at @xmath86 .
therefore we look for the proper value of @xmath82 . for @xmath87
we use @xmath6 [ eq . ( 2 ) ] from @xcite , but for @xmath86 we use : @xmath88 the value of @xmath89 is obtained by the total normalization of @xmath6 and is equal to @xmath90 the factor @xmath91 corresponds to the result for the part of the normalization ( for @xmath92 ) from the total normalization condition : @xmath93 finally , from eq .
( [ eq8ant ] ) one can obtain the following expression for the scaling function : @xmath94 in fig .
[ fig5ant ] we present the results for the scaling function [ eq .
( [ eq8ant ] ) ] for different values of @xmath82 , compared with the rfg model result .
one can see that agreement with the experimental qe scaling function is achieved when the value @xmath95 is used in eqs .
( [ eq10ant ] ) , ( [ eq11ant ] ) and ( [ eq13ant ] ) .
this means that the power - law decrease of @xmath6 which gives an optimal agreement with the data is @xmath96 the scaling function in a dilute fermi gas calculated using eq .
( [ eq13ant ] ) for different values of @xmath82 in the asymptotics of the momentum distribution @xmath97 given in comparison with the rfg result .
the grey area shows experimental data taken from @xcite.,width=302 ] we should note that this particular form of the power - law asymptotics is close to that obtained in the cdfm @xcite @xmath98 i.e. it corresponds to @xmath99 with @xmath100 .
the inverse fourier transform of @xmath71 for @xmath101 and @xmath102 gives @xmath103 and @xmath104 , respectively .
we would like to emphasize the consistency of both the optimal asymptotics of @xmath6 for the dilute fermi gas found in this work [ eq .
( [ eq14ant ] ) ] with that in the cdfm [ eq . ( [ eq15ant ] ) ]
as was shown in @xcite the calculated qe scaling function @xmath0 in the cdfm agrees well with the experimental scaling function .
this fact shows that the behavior of the qe scaling function depends mainly on the particular form of the power - law asymptotics of the nucleon momentum distribution .
this is proved on our work by the similarities of the result for the case of an interacting dilute fermi gas with that one obtained in the cdfm as a model accounting for nn correlations in realistic finite nuclear systems .
the results of the present work can be summarized as follows : i ) : : the superscaling observed in inclusive electron scattering from nuclei is considered within the model of dilute fermi gas with interactions between particles .
the latter gives an improvement in comparison with the results of the relativistic non - interacting fermi gas model allowing one to describe the qe scaling function for @xmath1 , whereas the rfg model gives @xmath2 in this region .
ii ) : : it is established that the hard - sphere ( with delta - forces between nucleons ) approximation for the dilute fermi gas is quite a rough one .
the use of more realistic nn forces leading to @xmath105 instead of @xmath106 ( for delta - force ) in the well - known power - law asymptotics of the momentum distribution @xmath107 at large @xmath4 leads to a good explanation of the data for the @xmath9-scaling function in inclusive electron scattering from a wide range of nuclei .
iii ) : : the asymptotics of @xmath108 found in the dilute fermi gas by optimal fit to the data for @xmath0 is similar to that in the cdfm ( @xmath109 ) @xcite which , being a theoretical correlation model , describes the superscaling in the quasielastic part of the electron - nucleus scattering .
thus , the momentum distribution in the dilute fermi - gas model with realistic nn forces can serve as an `` effective '' momentum distribution ( a step - like one with a discontinuity ) which gives a similar result for @xmath0 as the correlation methods for realistic finite nuclear systems .
it can be concluded that the momentum distribution with asymptotics from @xmath109 to @xmath110 is the proper one to explain the phenomenological shape of the scaling function obtained from inclusive qe electron scattering .
as already mentioned , the superscaling is due to the _ specific high - momentum tail _ of @xmath6 similar for all nuclei which is known to be caused by the short - range and tensor correlations related _ to peculiarities of the nn forces _ near their core .
the main result of the present work might be the observation that the values of @xmath0 for @xmath1 depend on _ the particular form _ of the power - law asymptotics of @xmath6 at large @xmath4 which is related to _ a corresponding particular behavior _ of the _ in - medium nn forces _ around the core .
namely , we point out that the power - law decrease of @xmath6 as @xmath111 with @xmath112 in the interacting dilute fermi gas is the proper one and it is close to that obtained in cdfm ( @xmath100 @xcite ) which describes the superscaling correctly as well .
the nn force for @xmath100 is expected to go as @xmath103 and for @xmath113 to go as @xmath114 .
hence , the present study allows one to conclude that the important property of the repulsive short - range core ( leading to nn correlations and high - momentum tail of @xmath6 ) is that it goes to infinity for @xmath115 as @xmath116 or softer .
the link between the asymptotic behavior of @xmath6 and nn force implies that inclusive qe electron scattering from nuclei provides important information on the nn forces in the nuclear medium .
this work was partly supported by the bulgarian science fund under contracts no .
@xmath117 - 1416 and @xmath117 - 1501 and by the ministerio
de educacin y ciencia ( spain ) under contract fis2005 - 00640 .
one of the authors ( m.k .
gaidarov ) is grateful for warm hospitality to the csic and for support during his stay there to the state secretariat of education and universities of spain ( n / ref .
sab20050012 ) .
d. faralli , c. ciofi degli atti , and g. b. west , in _ proceedings of 2nd international conference on perspectives in hadronic physics _ , ictp , trieste , italy , 1999 , edited by s. boffi , c. ciofi degli atti , and m. m. giannini ( world scientific , singapore , 2000 ) , p. 75 .
a. n. antonov , v. a. nikolaev , and i. zh .
petkov , bulg .
* 6 * , 151 ( 1979 ) ; z. phys .
a * 297 * , 257 ( 1980 ) ; _ ibid . _ * 304 * , 239 ( 1982 ) ; nuovo cimento a * 86 * , 23 ( 1985 ) ; nuovo cimento a * 102 * , 1701 ( 1989 ) ; a. n. antonov _
et al . _ ,
c * 50 * , 164 ( 1994 ) .
a. n. antonov , p. e. hodgson , and i. zh .
petkov , _ nucleon momentum and density distributions in nuclei _
( clarendon press , oxford , 1988 ) ; _ nucleon correlations in nuclei _ ( springer - verlag , berlin - heidelberg - new york , 1993 ) .
o. benhar , nucl .
b , proc . suppl . * 139 * , 15 ( 2005 ) ; nucl - th/0408045 ; o. benhar and n. farina , nucl .
b , proc . suppl . * 139 * , 230 ( 2005 ) ; nucl - th/0407106 ; o. benhar , n. farina , h. nakamura , m. sakuda , and r. seki , nucl . phys .
b , proc . suppl . * 155 * , 254 ( 2006 ) ; hep - ph/0510259 ; phys .
d * 72 * , 053005 ( 2005 ) . | the superscaling observed in inclusive electron scattering is described within the dilute fermi gas model with interaction between the particles .
the comparison with the relativistic fermi gas ( rfg ) model without interaction shows an improvement in the explanation of the scaling function @xmath0 in the region @xmath1 , where the rfg result is @xmath2 .
it is found that the behavior of @xmath0 for @xmath1 depends on the particular form of the general power - law asymptotics of the momentum distribution @xmath3 at large @xmath4 .
the best agreement with the empirical scaling function is found for @xmath5 in agreement with the asymptotics of @xmath6 in the coherent density fluctuation model where @xmath7 .
thus , superscaling gives information about the asymptotics of @xmath6 and the nn forces . |
in a previous paper of this symposium [ 14 ] we have discussed the induced representations of the euclidean group on the lattice via duality and fourier transform . in this paper
we apply the same method to the induced representation of poincar group on the lattice .
we explore the problem of irreducibility connected with the orbit condition , the assymptotic limit of the difference equation for the klein - gordon and dirac field . in section 2 we present the realization of the integral lorentz transformation , that can be factorized completely with the help of kac generators , both for the fundamental and spin representation . in section 3
we introduce several types of fourier transform that will be used to go from momentum to position space . in section 4
we reproduce the properties of covariant bispinors in discrete momentum space in similar fashion to the continuous one . in section 5 , we elaborate the induced representation of poincar group on the lattice following standard procedure and we compare the results derived from the different types of fourier transform .
a lorentz transformation is integral if leaves invariant the cubic lattice and at the same time keeps invariant the bilinear form @xmath2 according to coxeter [ 1 ] all integral lorentz transformations ( including reflections ) are obtained by combining the operations of permuting the spatial coordinates @xmath3 and changing the signs of any of the coordinates @xmath4 together with the operation of adding the quantity @xmath5 to each of the four coordinates of a point . these operations can be described geometrically by the reflections on the plans perpendicular to the vectors @xmath6 where @xmath7 is an orthonormal basis . in matritial form
these reflections are @xmath8 these reflections constitue a coxeter group , the dynkin diagram of which is the following kac has proved [ 2 ] that @xmath9 generate all the integral lorentz tranformations that keep invariant the upper half of the light cone . this result can be used to factorized any integral lorentz transformation that belong to the proper orthochronous group .
let @xmath10 be an integral matrix of determinant + 1 , satisfying @xmath11 and also @xmath12 . from @xmath13
it follows that only one of @xmath14 can be zero .
. then we apply @xmath16 to @xmath17 from the left until @xmath14 become non - positive integers . to the resulting matrix
we apply @xmath18 .
we get @xmath19 with @xmath20 .
therefore @xmath22 or @xmath23 by iteration of the same algorithm we get @xmath24 the last inequality is a consequence of the fact that @xmath17 and @xmath18 belong to the complete lorentz group . following this process
we get an integral matrix with @xmath25 which is a combination of @xmath16 , giving all the 24 elements of the cubic group on the lattice .
therefore a general integral lorentz transformation of the proper orthochronos type @xmath17 can be decomposed as @xmath26 where @xmath27 are the matrix which change sign of @xmath14 and @xmath28 a particular case of integral lorentz transformations are the boost or integral transformations of two inertial systems with relative velocity .
the general expresion for these transformations can be obtained with the help of cayley parameters [ 3 ] .
let us take @xmath29 and @xmath30 integer numbers .
we have two cases , corresponding to two diophantine equations : 1 .
@xmath31 @xmath32 2 .
@xmath33 @xmath34 the solutions of the diophantine equation i ) and ii ) are obtained by applying all the coxeter reflections to the vector @xmath35 in case i ) and to the vector @xmath36 in case ii ) from the inspection of ( 7 ) if we take the quotient of @xmath17 with respect the subgroup of all integral rotations or cubic group we are left with the coset representatives which are not exhausted by the pure lorentz transformations ( 8) or ( 9 ) , because these are always symmetric matrices .
therefor we have to add all the integral lorentz matrix that applied to the vector @xmath35 gives all the integral vectors of the type @xmath37 with @xmath31 .
this is equivalent to say that we apply to @xmath35 not only the matrix @xmath17 given by ( 8) but also its square root matrix , namely , @xmath38 in position space the space - time coordinates of the lattice @xmath39 are integer numbers .
they transform under integral lorentz transformations into integral coordinates .
the same is true for the increments @xmath40 . in momentum space
the components of the four - momentum are not integer numbers but they can be constructed with the help of integral coordinates , namely , @xmath41 if @xmath40 transform under integral lorentz transformations the new @xmath42 will be given in terms of integral @xmath43 .
using the homomorphism between the groups @xmath44 and @xmath45 we obtain the representation of integral lorentz transformations in 2-dimensional complex matrices . from the knowledge of the cayley parameters [ 5 ] we read off the matrix elements of @xmath46 @xmath47 @xmath48 for instance
, we calculate the 2-dimensional representation of the coxeter reflection @xmath49 multiplied by the parity operator @xmath50 in order to get an element of the proper lorentz group ) identifying its matrix elements with lorentz matrix written in terms of cayley parameters : easy calculations give the unique solutions : @xmath51 the integral lorentz transformations without rotations as given in ( 8) and ( 9 ) have a 2-dimensional representation making @xmath52 in ( 11 ) and the choice @xmath53 in ( 12 ) . in order to complete the picture we have to add the 2-dimensional representation of the matrix @xmath54 given in ( 10 ) which turns out to be @xmath55 we give also the @xmath56 matrix representation of the discrete momentum .
let @xmath57 a four momentum which is obtained by applying all the integral lorentz transformations , given by ( 7 ) divided by the cubic group , to the vector @xmath58 . in 2-dimensional matrix form
this is equivalent to apply the matrix @xmath59 given by ( 17 ) to the unit matrix multiply by @xmath60 : @xmath61 which is of the standard form if we identify the components of the 4-momentum as @xmath62 with @xmath63 integral numbers , satisfying @xmath31 .
in order to go from position space to momentum space on the lattice we can define several restrictions of the continuous variables of fourier transform to the discrete variables on the lattice . *
i type . discrete position and momentum variables of finite rank * we construct an orthonormal basis [ 6 ] @xmath64 @xmath65 that satisfy periodic boundary conditions : @xmath66 orthogonality relations : @xmath67 and completness relations : @xmath68 the finite fourier transform reads @xmath69 for some periodic function @xmath70 on the lattice @xmath71 if we write @xmath72 this transform coincide with standard finite fourier transform [ 7 ] .
* ii type . discrete position and continuous momentum
* when we restrict the position variables in continuous fourier transform to discrete values we obtain the fourier series .
@xmath73 fourier expansion : for a periodic function @xmath74 @xmath75 @xmath76 now we make the change of variable @xmath77 @xmath78 and the orthogonality relations become @xmath79 and the completness relation @xmath80 notice that @xmath81 may not be a periodic function .
* iii type : discrete position and discrete momentum of infinite rank * we construct an orthonormal basis @xmath82 satisfying @xmath83 ( proof : for @xmath84 we use the identity @xmath85 with @xmath86 ) .
the completness relation is now : @xmath87 where @xmath88 is a @xmath89 sequence satisfying @xmath90 ( proof:@xmath91 for all @xmath17 , as required ) the fourier transform becomes : @xmath92 where @xmath93 when @xmath94 and @xmath95 @xmath96 the fourier transform of type iii was introduced in [ 8 ] .
when @xmath97 @xmath98 the orthogonality relations converges @xmath99 the completness relations becomes @xmath100 and the fourier transform converges to : @xmath101
let @xmath102 be an element of the proper lorentz group corresponding to the element @xmath103 and @xmath104 the parity operator . if the components of four - momentum are written as @xmath105 matrix @xmath106 where @xmath107 and @xmath108 are the pauli matrices .
the transformations of @xmath109 under parity and @xmath110 are @xmath111 @xmath112 @xmath113 it follows @xmath114 therefore the matrix @xmath115 gives an other 2-dimensional representation of the lorentz group non - equivalent to @xmath103 . in order to enlarge the proper lorentz group with space reflection
we take both representations @xmath59 and @xmath116 .
let @xmath117 the space reflection group and @xmath118 , then the semidirect product @xmath119 with the multiplication law @xmath120 form a group .
this group has a 4-dimensional representation , a particular elements of which is @xmath121 that satisfy @xmath122 with respect to this representation a-4-component spinor in momentum space @xmath123 transform as follows @xmath124 using a similarity transformation we obtain an equivalent representation @xmath125 with @xmath126 in this representation @xmath127 the new four - spinor @xmath128 transform as @xmath129 the dirac wave equation can be considered as a consequence of the relativistic invariance [ 9 ] in the rest system we want a projection operator that selects one irreducible representation out of the dirac representatition .
this is achieve in the rest system by @xmath130 in order to get the projection operator in an arbitrary system we apply @xmath131 given by ( 49 ) and ( 17 ) @xmath132 where @xmath133 using the identities we find @xmath134 where @xmath135 are dirac matrices with the realization @xmath136 @xmath137 collecting these result
we obtain the dirac equation in momentum space @xmath138 or @xmath139 ( an equivalent method can be used applying to the projection operator the foldy - wouthuysen transformation [ 10 ] ) we apply the operator @xmath140 from the left to ( 59 ) and obtain @xmath141 the dirac equation is invariant under the group @xmath142 defined before . in other words ,
if @xmath143 is a solution of the dirac equation , so is @xmath144 . put @xmath145 .
then @xmath146 @xmath147 @xmath148 then @xmath149 as required . for the space reflection @xmath150 @xmath151 @xmath152 we get @xmath153 as required .
notice that all properties of dirac representation in continuous momentum space are carried out to the discrete momentum space without modification .
for the vector representation of the lorentz group we take an element @xmath154 such that takes the momentum in the rest system to an arbitrary system .
intherest system the projection operator is @xmath155 then @xmath156 and the wave equation in momentum space becomes @xmath157 or @xmath158 where @xmath159 is a 4-component vector function . in order to construct wave equation in configuration space for the dirac and vector representation we use the fourier transform given in section 3 .
we take the three types * type i*. boundary condition imposes finite values for the momentum @xmath160
@xmath161 define the difference operator [ 8 ] @xmath162 @xmath163 multiply the dirac equation ( 60 ) by @xmath164 with @xmath165 the plane waves on the lattice , and using the properties of theses functions @xmath166 we obtain @xmath167 adding for @xmath168 , we get the inverse fourier transform @xmath169 that satisfies the dirac equation on the lattice @xmath170 * type ii*. the momentum is continuous and has the range @xmath171 the plane waves appearing in the fourier transform are @xmath172 as before we multiply the dirac equation in momentum space by @xmath164 and using the property @xmath173 we get @xmath174 integrating for @xmath175 and puting @xmath176 we obtain the desired dirac equation in configuration space .
* type iii*. the momentum is discrete and was given in section 2 , namely @xmath177 the plane waves are @xmath178 notice that we do not impose boundary conditions therefore the parameters @xmath179 are discrete and infinite .
we multiply the dirac equation in momentum space by @xmath180 and using @xmath181 we get as before @xmath182 summing over all values of cayley parameter that define a four - momentum we get @xmath183 where @xmath184 notice that @xmath185 should satisfy @xmath186
let @xmath187 be the poincar group restricted to the integral lorentz transformations and discrete traslations on the lattice with the group composition @xmath188 in order to construct irreducible representations we follow standard method \(1 ) choose an @xmath189 of the translation group @xmath190 \(2 ) define a little group @xmath191 by the stability condition @xmath192 \(3 ) this condition leads to the following decomposition of the @xmath193 of @xmath194 @xmath195 \(4 ) choose coset generators @xmath196 of @xmath194 constructed from the group action @xmath197 \(5 ) then the induced representations is : @xmath198 this is an @xmath193 of @xmath199 in the case of the discrete pincar group for the massive case , @xmath200 and the little group is the cubic group that satisfies condition ( 81 ) . for the coset representative
@xmath201 we can choose the integral lorentz transformations @xmath202 that take @xmath203 into an arbitrary discrete momentum .
the dirac delta function in ( 84 ) is zero unless @xmath204 substituting in ( 84 ) and using ( 82 ) we get @xmath205 the spinor representation of the second factor is given with respect to the element @xmath206 that belongs to the little group , @xmath207 , and this group becomes irreducible when restricted to the spin 1/2 or spin 1 representation [ 12 ] .
the first factor can be written : @xmath208 where @xmath209 are all the points that defined the @xmath193 of translation group and belong to the orbit .
we want to characterized the discrete orbits by functions in the momentum space that vanish on the orbit points and only in these points . a natural way to construct these functions is through the dirac equation in momentum space ( 59 ) . multiplying this equation from the left by @xmath210 we get @xmath211 the @xmath160 have different physical meaning as we stressed in the realization of dirac equation in position space .
we have to check whether the constraints ( 88 ) satisfy the following conditions : \(1 ) they should be polynomials or infinite product of polynomials that vanish on the orbit points , \(2 ) the constraints should admit a periodic extension of the momentum space , \(3 ) the constraints should vanish on a orbit and they must be lorentz invariant , \(4 ) the constraints should vanish only on the points of the orbit , \(5 ) when the lattice spacing goes to zero , the difference equation should go to the continuous minkowski limit . in the fourier
transform of type i and ii the @xmath175 variable are related to the physical momentum @xmath212 by the expresion @xmath213 or @xmath214 due to the special trigonometric function this expression is periodic in momentum space .
nevertheless when lorentz transformations are applied to the components of physical momentum the new momentum do not satisfies the constraint equation and the contraints vanish at different points
. therefore conditions ( 3 ) and ( 4 ) are violated , but they can be recovered in the asymptotic limit when @xmath215 @xmath216 in the case of fourier transform of type iii the @xmath160 coincide whit the physical momentum whose discrete values are the points of the orbit given by @xmath217 and therefore condition ( 3 ) and ( 4 ) are fullfiled , although ( 2 ) is violated . in all the three cases the wave equation on position space , as described in section 4 gives in the asymptotic limit , the continuous dirac equation .
we have attempted a new program for introducing poincar symmetry on the lattice , that was considered broken , as many authors have claimed [ 13 ] . so far some points of this program have been achieved ; such as , realization of all the integral lorentz transformation and is representation in 2-dimensional space .
several @xmath193 of the translation group on the lattice , some versions of the fourier transform with discrete position and discrete or continuous momentum , the dirac equation on the lattice in position space .
as far as the induced representation of the poincar group , following similar technic as in the continuous case , leads to some inconsistensy : for the type i and ii of the @xmath193 of translation group the scheme is irreducible and invariant only in the asymptotic limit , but they satisfy all the requirement for the induced representation . for the type iii
the representation is lorentz invariant and irreducible , but the requirement of the orbit condition necessary for the induced representations is kept only in the asymptotic limit .
one of the authors want to expressed his gratitude to the director of the institut fr theoretische physik , t. universitt tubingen , where part of this work was done , for the hospitality .
this work has been partially supported by d.g.i.c.y.t .
contract # pb94 - 1438 ( spain ) . | following standard methods we explore the construction of the discrete poincar group , the semidirect product of discrete translations and integral lorentz transformations , using the wigner - mackey construction restricted to the momentum and position space on the lattice .
the orbit condition , irreducibility and assimptotic limit are discussed .
* induce representation of the poincare + group on the lattice : spin @xmath0 and @xmath1 case * + miguel lorente + peter kramer |
in @xcite , the minimality of semigroup actions ( equivalently , minimal set of iterated function systems , abbreviated as ifss ) on the interval are discussed and several examples of non - minimal actions are presented there . for these examples , we certainly know that the action is not minimal but we could not decide the shape " of the minimal set ( similar type of questions are extensively studied recently , see for example @xcite for complex dynamical systems case ) . in the case of group actions on the circle , there is a famous trichotomy of the minimal set ( see for example chapter 2 of @xcite ) : either it is a finite set , equal to the whole manifold , or homeomorphic to the cantor set .
thus once we know that the minimal set is niether the finite orbit nor the whole manifold , then we can immediately conclude that it is homeomorphic to the cantor set . on the other hand , in semigroup action case , because of the lack of homogeneity of the minimal set , this trichotomy is no longer valid .
the best thing we know in general is , as is stated in @xcite ( see theorem 5.2 ) , the trichotomy of the following type : the minimal set is either a finite set , homeomorphic to the cantor set or a closed set with non - empty interior .
this trichotomy is not enough to conclude the type of minimal set just from the non - minimality of the action . in this article
, we give an example of class of ifss with a minimal set homeomorphic to the cantor set .
they are produced by performing some modifications on the example presented in @xcite .
thus , still i do not know the type of minimal set for the original examples , see also remark [ r.open ] .
after the announcement of this example , masayuki asaoka ( kyoto university ) showed me another example of minimal cantor set using the measure theoretic argument . in the appendix we present that example . *
acknowledgements * this research is supported by the aihara project , the first program from jsps , initiated by cstp .
the author is thankful to masayuki asaoka , kengo shimomura , hiroki sumi and dmitry turaev for helpful conversations .
in this section , we prepare some definitions , give the axiomatic description of the example and the precise statement of our main result .
we consider ifss on the closed unit interval @xmath0 $ ] generated by two maps @xmath1 satisfying the following conditions : * @xmath2 are @xmath3-diffeomorphisms on their images . *
@xmath4 and @xmath5 . * @xmath6 for @xmath7 . * @xmath8 .
we denote the set of the pairs of diffeomorphisms satisfying these conditions by @xmath9 ( by @xmath10 we denote the set of @xmath3-maps from @xmath11 to itself which is diffeomorphism on its image ) .
we prepare some notations and basic definitions .
we denote the semigroup generated by @xmath12 and @xmath13 by @xmath14 .
it acts on @xmath11 in a natural way . for @xmath15 , the orbit of @xmath16 , denoted by @xmath17 ,
is defined to be the set @xmath18 .
a non - empty set @xmath19 is called a _
minimal set _
if for every @xmath20 , @xmath21 , where @xmath22 denotes the closure of @xmath23 .
finally , we say that the action of @xmath14 is minimal if @xmath11 is the minimal set , in other words , every point in @xmath11 has a dense orbit . for ifss generated by @xmath24 , we can prove the following ( see lemma 1 in @xcite ) : there exists a unique minimal set @xmath25 .
furthermore , we have @xmath26 . for @xmath27 , put @xmath28 $ ] and call it _ overlapping region _
( @xmath29 is the overlap between @xmath30 and @xmath31 ) .
note that by definition of @xmath32 , @xmath29 is an interval with non - empty interior .
we define two sequences of intervals @xmath33 and
@xmath34 as follows : * @xmath35 $ ] . *
@xmath36 $ ] .
let us consider the following property for @xmath37 : @xmath38 is contained in the interior of @xmath39 .
we call this property _ single overlapping property _ and denote it by @xmath40 . note that this condition is equivalent to @xmath41 and @xmath42 .
this implies that @xmath40 is a @xmath43-open property in @xmath32 .
suppose @xmath27 satisfies @xmath40 .
consider the following condition ( in the following , by @xmath44 we denote the set of the interior points of @xmath23 ) : there are closed intervals @xmath45 and @xmath46 with non - empty interior such that @xmath47 and @xmath48 . with @xmath40 and @xmath49 . ] we call this property _
hole property _ and denote it by @xmath49 ( see figure [ fig.aliint ] ) . if @xmath27 satisfies @xmath40 and @xmath49 , then we have the following ( see also example 2 in @xcite ) : [ p.tenganai ] if @xmath27 satisfies @xmath40 and @xmath49 , then the ( unique ) minimal set @xmath25 of @xmath14 does not coincides with the whole interval .
more precisely , we have @xmath50 .
suppose @xmath51 .
we consider the case @xmath52 ( the proof of the case @xmath53 is similar so we omit it ) .
then we know that there exists a point @xmath54 .
since @xmath55 , it means that there exists a point @xmath56 such that @xmath57 or @xmath58 holds .
however , since @xmath59 by definition , the second option can not happen .
hence we have @xmath60 and it implies @xmath61 . by repeating this process , we continue taking the backward image of @xmath16 contained in @xmath62 . at some moment
it must be equal to @xmath63 but this is a contradiction .
we define two maps @xmath64 and @xmath65 as follows : for @xmath66 , let @xmath67 be the least non - negative integer such that @xmath68 holds .
this defines a piecewise @xmath3 map @xmath64 by setting @xmath69 ( @xmath70 is just an induced map " from @xmath71 to @xmath72 of the inverse system @xmath73 ) .
we define @xmath74 similarly , exchanging the role of @xmath12 and @xmath13 .
then let us consider the following property : @xmath70 is uniformly expanding outside @xmath75 and @xmath74 is as such outside @xmath76 .
more precisely , there exists @xmath77 such that at each @xmath78 ( resp .
@xmath79 ) , @xmath80 ( resp .
@xmath81 ) where @xmath82 ranges over the points for which the term @xmath83 ( resp .
@xmath84 ) makes sense .
we call this property _ eventual expansion property _ and denote it by @xmath85 .
in general , the property @xmath85 may be violated by some small perturbation . however ,
if @xmath86 then we see that @xmath85 is a @xmath3-open property . finally , we define the castration property .
we put @xmath87 and @xmath88 .
we call them ruination region . note that by construction @xmath89 is the disjoint union of infinitely many intervals accumulating to @xmath90 and @xmath91 to @xmath92 .
then , the castration property is stated as follows : @xmath93 .
we denote this property by @xmath94 . by the structure of @xmath89 and @xmath91
, we observe the following : [ r.hairu ] if @xmath94 holds , then @xmath95 and @xmath96 .
a priori , the openness of the property @xmath94 is not clear , since it involves the conditions of infinitely many intervals .
however , by remark [ r.hairu ] , together with the structure of @xmath89 and @xmath91 , we see that it only involves the conditions of finitely many intervals . as a corollary
, we see that @xmath94 is a @xmath43-open property in @xmath32 .
we denote the set of pairs of diffeomorphisms in @xmath32 satisfying the conditions @xmath40 , @xmath49 , @xmath85 and @xmath94 by @xmath97 .
now , we state our main result .
if @xmath98 is in @xmath97 then the ( unique ) minimal set of @xmath14 is homeomorphic to the cantor set .
later we also prove the following : [ p.sonzai ] in @xmath9 , the set @xmath97 has non - empty interior ( with respect to the relative topology on @xmath32 induced by the @xmath3-topology on @xmath99 ) . as a corollary
, we see that our example can be taken @xmath100 or @xmath101 .
[ r.open ] for @xmath27 satisfying the condition @xmath40 and @xmath49 , we know that the action of @xmath14 is not minimal .
however , in general the author does not know if the minimal set is homeomorphic to the cantor set or not .
in this section , we give the proof of our theorem .
we fix @xmath102 and consider its unique minimal set @xmath25 .
we first remember the famous characterization of the cantor set : if a topological space is a compact , metrizable , perfect and totally disconnected , then it is homeomorphic to the cantor set ( see for example @xcite ) . in our setting , it is clear that @xmath25 satisfies the first and the second properties .
the perfectness of the minimal set can be seen easily by definition of the minimal set .
thus we only need to prove the totally disconnectedness of @xmath25 . for that , since we work on one dimensional setting , we only need to prove the emptiness of the interior .
we state it in more precise way .
[ r.reduce ] for the proof of the theorem , we only need to prove the following : for every @xmath20 and every non - empty open interval @xmath103 containing @xmath16 , there exists a non - empty interval @xmath104 such that @xmath105 .
first , we start from some simple observations .
[ l.aruyo ] let @xmath103 be a non - empty open interval . * if @xmath106 ( resp .
@xmath107 ) and @xmath108 , then @xmath109 ( resp .
@xmath110 ) .
* suppose @xmath111 and @xmath108 .
then , if @xmath112 ( resp .
@xmath113 ) , we have @xmath110 ( resp .
@xmath114 ) .
we start from the proof of the first item . by the symmetry
, we only consider the case @xmath106 ( and omit the case @xmath107 ) .
one important observation for the proof is the following : if @xmath115 belongs to @xmath116 , then @xmath117 . indeed , in general , for given @xmath118 we have either @xmath119 or @xmath120 ( both condition may hold simultaneously ) . since @xmath115
, the second option can not hold .
so we see that @xmath119 .
then , fix @xmath121 so that @xmath122 holds . by repeating the similar reasoning
, we can prove that @xmath123 for every @xmath124 with @xmath125 ( by the induction of @xmath124 ) .
in particular , we have @xmath126 , which is equal to @xmath127 by definition .
thus we see that @xmath128 contains a point @xmath129 , which implies @xmath109 .
let us consider the proof of the second item .
again by the symmetry between @xmath12 and @xmath13 , we only treat the case where @xmath112 and omit the case @xmath113 .
suppose @xmath112 and there exists a point @xmath130 . then , as is the previous discussion , one of the following options holds : @xmath131 or @xmath132 .
however , the condition @xmath133 implies @xmath134 and this , together with proposition [ p.tenganai ] prohibits the first option .
thus we have @xmath135 , which implies what we claimed . in the similar way
, we have the following ( we omit the proof of it ) . [ l.naiyo ] if @xmath136 , then @xmath137 .
we prepare one definition . for @xmath102
, we define @xmath138 where @xmath139 denotes the boundary of @xmath23 .
then we have the following . [ l.sukima ] if @xmath103 is an open interval with @xmath140 ( resp .
@xmath141 ) , then there exists a non - empty open interval @xmath142 with @xmath143 .
in other words , @xmath144 ( resp .
@xmath145 ) . take @xmath103 with @xmath140 .
we divide our situation into four cases : @xmath146 , @xmath147 , @xmath148 or @xmath149 . for the first case , remember proposition [ p.tenganai ] : it implies that every non - empty open set @xmath150 is disjoint from @xmath25 .
thus one of such @xmath151 gives us the desired set . in the second case , using the castration property ( that is , @xmath152 ) , we see that there exists a non - empty open set @xmath153 . by lemma [ l.naiyo ] , this @xmath151 gives the desired set .
the third case is reduced to the second case by remark [ r.hairu ] and the fact that @xmath89 accumulates to @xmath90 . in the last case ,
consider @xmath154 .
then it contains @xmath90 .
thus by the previous argument , we can take @xmath155 with @xmath156 .
then , by repeating the similar argument as lemma [ l.aruyo ] , we see that @xmath157 gives us the desired interval .
now we start the proof of the theorem .
remember that to prove the theorem we only need to prove the statement in remark [ r.reduce ] .
the idea of the proof is using lemma [ l.aruyo ] , [ l.naiyo ] and [ l.sukima ] inductively to examine the possible past behavior of @xmath158 .
we start the proof for @xmath158 satisfying @xmath159 .
the other case is easy to prove and will be discussed later . given @xmath160 with @xmath161
, we define a @xmath3-map @xmath162 which is a diffeomorphism on the image and a non - empty open interval @xmath163 . after that we will see that @xmath164 gives us the claimed interval in remark [ r.reduce ] . for that
, we define a finite sequence of maps @xmath165 and intervals @xmath166 inductively as follows .
we put @xmath167 and @xmath168 .
suppose that we have defined @xmath169 and @xmath170 ( @xmath171 is some non - negative integer ) .
then we proceed as follows : first , if @xmath172 , then take the interval @xmath151 in the conclusion of lemma [ l.sukima ] , put @xmath173 and finish the construction .
suppose not .
then we have five possibilities : @xmath174 , @xmath175 , @xmath176 , @xmath177 or @xmath178 .
we proceed as follows . * in the first ( resp .
second ) case , set @xmath179 ( resp .
@xmath180 ) , @xmath181 , ( resp .
@xmath182 ) and finish the construction . note that by proposition [ p.tenganai ] , we see that @xmath156 .
* suppose @xmath176 . in this case , we know that @xmath158 is contained in either @xmath183 or @xmath184 ( remember that we are under the assumption @xmath185 ) . * * if @xmath186 , then put @xmath187 , @xmath188 and continue the construction .
* * otherwise ( that is , if @xmath189 ) , then put @xmath190 , @xmath191 and continue the construction .
* in the fourth ( resp .
fifth ) case , set @xmath190 ( resp .
@xmath187 ) , @xmath191 , ( resp .
@xmath188 ) and continue the construction .
note that , by eventual expansion property , this process finishes in finite steps .
thus for given @xmath158 , we can take @xmath192 and @xmath193 .
note that , by construction , @xmath194 is a diffeomorphism on its image and @xmath195 satisfies @xmath196 . then consider the non - empty open interval @xmath197 .
we claim that @xmath198 . indeed , suppose @xmath199 .
this implies @xmath200 .
however , applying lemma [ l.aruyo ] or [ l.naiyo ] to each option of the construction of @xmath201 , we see that @xmath202 contains a point of @xmath116 . then by induction
we conclude that @xmath203 for every @xmath124 .
in particular , we see that @xmath204 contains some point of @xmath25 , which contradicts to the choice of @xmath151 .
we give the proof for the case where @xmath158 is outside @xmath39 , which completes the proof of the theorem . in this case , we know that @xmath158 is contained in @xmath205 or @xmath206 . by the symmetry , we only need to consider the first case . by shrinking @xmath158 if necessary , we can assume that there exists @xmath207 such that @xmath208 .
thus we have @xmath209 . applying the conclusion of the previous case , we take a non - empty open set @xmath210 such that @xmath156
. then , consider @xmath211 .
we claim that @xmath212 ( which implies what we want to prove ) . indeed ,
if not , there exists a point @xmath213 .
then , repeating the argument used in the proof of lemma [ l.aruyo ] , together with the single overlapping property @xmath40 , we see that @xmath214 for every @xmath215 ( by the induction of @xmath124 ) .
in particular it implies that @xmath143 , but this is a contradiction .
in this section , we prove proposition [ p.sonzai ] , that is , we describe how we construct the example .
we start from the system @xmath216 then , put @xmath217 and @xmath218 and take intervals @xmath219 , @xmath220 centered at @xmath221 , @xmath222 respectively , with sufficiently small diameter ( for example , @xmath223 is enough ) .
we modify @xmath224 and @xmath225 to get the example .
every modification is performed keeping the symmetry of the graph with respect to the diagonal : @xmath226 holds for every @xmath227 $ ] .
first we modify @xmath224 and @xmath225 to @xmath3-maps @xmath228 and @xmath229 respectively , keeping them in @xmath230 so that the following holds : * @xmath231 outside @xmath219 and @xmath232 outside @xmath220 .
* there are intervals @xmath233 , @xmath234 such that @xmath235 and @xmath236 .
such a modification can be done just by modifying local behavior around @xmath221 and @xmath222 expanding .
then we have the following : [ p.ana ] there exist intervals @xmath237 and @xmath238 such that @xmath239 , @xmath240 and @xmath241 , @xmath242 . by definition , we have @xmath243 .
so , @xmath244 is a closed interval contained in @xmath245 .
we also know that it can not be a point , since it contains @xmath246 .
furthermore , by definition we can see that @xmath247
. then @xmath248 and @xmath249 are the desired intervals . from @xmath250
, we construct a one - parameter family of ifss @xmath251 as follows .
we fix a small positive number @xmath171 ( @xmath252 is enough ) and modify @xmath253 , @xmath254 to @xmath251 , where @xmath255 is some non - negative small real number , keeping being continuous and so that the following holds : * on @xmath256 , @xmath257 is an affine map with slope @xmath258 , * on @xmath259 , @xmath260 is an affine map with slope @xmath258 , * and keep the dynamics outside intact .
note that there exists small @xmath261 so that for every @xmath262 the pair @xmath251 satisfies the condition @xmath32 , single overlapping property @xmath40 and hole property @xmath49 .
hence the notions such as @xmath263 , @xmath264 , @xmath29 , @xmath89 and @xmath91 make sense for @xmath265 .
these objects vary as @xmath255 varies . to clarify the dependence , we put @xmath255 to them .
for example , @xmath266 denotes the interval @xmath267 for @xmath251 .
note that if @xmath171 is chosen sufficiently small then @xmath249 and @xmath268 do not depend on @xmath255 .
thus we do not put @xmath255 to them .
we are interested in the shape of @xmath269 and @xmath270 in @xmath271 . to compare them of different @xmath255 , we introduce the following notion .
let @xmath272 be two different numbers .
then we define @xmath273 to be the unique orientation preserving affine homeomorphism from @xmath274 to @xmath275 . in @xmath276 ,
we define two sequences of intervals . for @xmath277
, we define @xmath278 as follows : @xmath279 similarly , we define @xmath280 as follows : @xmath281 for fixed @xmath255 , the interval @xmath278 , @xmath280 may be empty , but for large @xmath282 it is not empty . note that @xmath283 accumulates to @xmath284 , @xmath285 to @xmath286 .
we also have @xmath287 and @xmath288
. then we define @xmath289 as follows : @xmath290 in the definition of @xmath291 , the condition @xmath292 is in fact redundant , because of the symmetry .
now we prove the following ( later we will prove a stronger statement ) : @xmath291 is not empty .
consider the set @xmath293 .
this is a disjoint union of intervals converging to @xmath294 , and we can see that @xmath295 does not depend on @xmath255 .
the condition @xmath296 is equivalent to @xmath297 .
so we are interested in how @xmath298 varies when @xmath255 varies . by definition , we can see that for @xmath255 near @xmath63 , @xmath299 is monotone increasing as @xmath255 decreases and it converges to @xmath294 when @xmath300 .
thus , for sufficiently large @xmath282 , there is an interval @xmath301 such that for @xmath302 we have @xmath303 . in particular
, @xmath291 is not empty .
finally , note that because of the symmetry we do not need to pay attention to @xmath304 .
we fix @xmath305 .
then we claim the following : let @xmath305 .
there exists a sequence @xmath306 such that @xmath307 , @xmath308 , @xmath309 and @xmath310 for every @xmath311 and @xmath282 .
again consider the set @xmath293 . by definition ,
this set has a self - similarity in the following sense : @xmath312 . for @xmath277
, we define the number @xmath313 such that @xmath314 holds . for each @xmath277
such @xmath315 exists uniquely because of the monotonicity of @xmath316 with respect to @xmath255 . by construction
we can see that @xmath307 as @xmath317 .
then , for this sequence @xmath318 , consider the following diagram : @xmath319_{\phi_{\alpha_m , \alpha_n } } w^{\alpha_m } \ar[rr ] _
{ f^{-1}_{\alpha_m } \hspace{20pt } } & & ( f_{\alpha_m}^{-1 } \circ g_{\alpha_m}(0 ) , 1 ) \ar[d]_{g_{\alpha_0}^{(n - m ) } } & & \ar@{_{(}->}[ll ] h'_{p } \ar@{^{(}->}[lld ] \ar[d]^{g_{\alpha_0}^{(n - m ) } } \\ w^{\alpha_n } \ar[rr]^{f^{-1}_{\alpha_n } \hspace{20pt } } & & ( f_{\alpha_n}^{-1 } \circ g_{\alpha_m}(0 ) , 1 ) & & \ar@{_{(}->}[ll ] h'_{p}}$ ] together with the self similarity of @xmath295 and the equality @xmath320 for every @xmath277 , we see @xmath321 holds for every @xmath311 and @xmath282 . by the symmetry , we also have the equality for @xmath322 .
now we are ready for the proof .
let @xmath251 as above .
we fix the sequence @xmath318 .
then we take a @xmath3-diffeomorphism @xmath323 so that @xmath324 holds .
the existence of such @xmath325 is easy to see using the fact that @xmath326 .
then , for each @xmath282 we modify @xmath327 to @xmath328 so that for @xmath329 , @xmath330 holds , and keep intact outside .
then , this @xmath331 satisfies the single overlap , hole and castration property .
we claim that for sufficiently large @xmath282 , @xmath331 also satisfies the eventual expansion property @xmath85 .
we are interested in the differential of the first return map .
note that , by increasing @xmath282 , the number of the iteration for the first return increases , while the differential @xmath332 is intact .
the differetial of @xmath327 may vary , but it is a bounded change .
hence by taking sufficiently large @xmath282 , we get the uniform expansion of the first return map on @xmath333 .
the openness of the properties @xmath40 , @xmath94 and @xmath85 are easy to see .
the @xmath3-openness of @xmath49 comes from the fact that the hypothesis in proposition [ p.ana ] is @xmath3 ( indeed , @xmath43 ) open property .
after finding the previous example , masayuki asaoka showed me another type of example . in this appendix
we present it .
the construction of such pair of maps in @xmath32 is not such difficult ( see the graph in figure [ fig.aliappe ] ) , so we skip the detail .
note that these properties are open in @xmath32 ( the conditions @xmath346 are not open in @xmath347 , but in @xmath32 they are ) . for the proof , remember that we only need to prove the emptiness of the interior of the minimal set ( see section [ ss.setumei ] ) . for
that we prove that the lebesgue measure of @xmath25 is zero .
we denote the lebesgue measure on @xmath11 by @xmath348 . by the inclusion property of @xmath12 and @xmath13 , we know that @xmath349 is a nested sequence of compact sets .
thus @xmath352 is a non - empty compact set .
furthermore , by definition we see that @xmath353 .
thus we have @xmath354 ( indeed , we can prove the equality between @xmath25 and @xmath355 but since we do not need it we omit it ) .
the case @xmath361 is clear .
suppose the inequality is true for @xmath362 ( where @xmath171 is some non - negative integer ) .
first , by definition we have @xmath363 then , by uniform contracting property we have @xmath364 the first example and the second one are conceptually different examples .
the first example is constructed by investigating the past behavior of the dynamics , while the second one is constructed by investigating the future behavior . | we give some examples of ifss with overlap on the interval such that the semigroup action they give rise to has a minimal set homeomorphic to the cantor set .
* keywords : * ifs with overlap , cantor set , semigroup action , minimality . *
2010 mathematics subject classification : * 37b05 , 37e05 , 28a80 |
while classically clusters are presumed to be rather young ( gunn 1977 ) , and the present epoch is thought to be the era of cluster formation , recent results strongly indicate that clusters are ` old ' and thus are relics of the high - redshift universe . these indications are primarily due to rosat and asca results for @xmath1 clusters : the lack of evolution in the x - ray luminosity function of clusters at @xmath1 ( jones 1998 ; ebeling 1999 ; vikhlinin 1998 ) ; the relatively small change in the cluster temperature function out to @xmath2 ( henry 1997 ; donahue & voit 1999 ) ; the lack of evolution in the x - ray luminosity temperature relation to @xmath3 ( mushotzky & scharf 1997 ; donahue 1999 ) ; the lack of evolution in the metallicity in the cluster gas to @xmath2 ( fig . 1 ; mushotzky & loewenstein 1997 ; donahue 1999 ) ; and the lack of evolution in the galaxy velocity dispersion temperature relation ( tran 1999 ) .
in addition , the detection of massive clusters at @xmath2 ( bahcall & fan 1998 ) and their possible existence at @xmath4 ( dickinson 1999 ; stanford 1997 ) are highly unlikely in a high-@xmath5 universe ( however for a different opinion see blanchard 1999 ) .
other , independent , evidence is derived from hst and keck observations that indicate that the stars in cluster ellipticals are very old ( renzini 1999 )
. however it is not known , at present , if the galaxies themselves are old .
thus for the high - redshift universe the question is : what can we learn about the formation of large scale structure , galaxies and clusters from studies of these relics ?
clusters are , of course , the largest bound objects in the universe and are thought to be fair samples of the content of the universe ( baryons , dark matter and metals ) . as has been known for many years , x - ray observations of groups and clusters indicate that most of the observable baryons in these systems reside in the hot phase ( @xmath6k , fig . 2 ) .
if clusters and groups are ` representative ' samples , the stellar content of galaxies is , on average , @xmath7% of the gas mass ( see nevalainen 1999 for a recent detailed analysis ) .
the mean stellar ` metallicity ' ( fe ) is @xmath8 ` solar ' ( normalizing to the solar abundances of anders & grevesse 1989 and using the jorgensen 1999 data , averaging out over the metallicity gradients and the dependence of metallicity on luminosity ) . the metallicity of the gas is @xmath9% of solar ( fig .
thus with 80% of the mass in the gas and a similar metallicity for the gas and the stars , @xmath10 of the metals are in the gas phase .
since there are only subtle differences in the stellar spectra of ` bulges ' in groups and clusters from the field , these results should apply to most galaxies and therefore we expect that @xmath7% of the observable baryons and ` most ' of the total mass in stars is in bulges and that the rest of the baryons are in the hot gas phase .
it is amusing that this radical result is actually found in semi - analytic models ( somerville & primack 1999 ) .
what do these results imply for the formation and evolution of galaxies ?
* galaxies are open systems
much of their evolution involves outflow and inflow of gas .
* the presently - observable stars do not contain most of the metals .
* one can not study the history of metal creation by looking only at the ` relic ' stars . *
the relic gas contains most of the ` stuff ' where is this gas when it is not in groups and clusters ? * estimates of the total metals created by counting only stars are wrong , the controversy ( madau this symposium ) over the evolution of the star formation rate in the universe .
* there has been a lot more metal creation / star formation in the universe than indicated by normalizing to present day stellar data ( loewenstein & mushotzky 1996 ) because most of the metals are in the hot gas phase and not in the stars .
this is also essential for the semi - analytic models , which require a large contribution from supernova energy to explain the observed pattern of mass - to - light ratio and metallicity in galaxies ( see next section ) .
the gas in clusters and groups contains a history of the total energy deposited in it .
this energy originates from two general sources : ( 1 ) gravity ( infall shocks and mergers ) and ( 2 ) star formation ( supernova and stellar winds ) .
since massive star formation produces most of the metals , we can use the metallicity as an estimate of the total heat produced . for a fixed metallicity the fraction of the energy from star formation increases as the mass of the system is reduced ; in other words the available energy per particle from infall is reduced as the mass of the system is lowered while , for a fixed metallicity , the energy produced by stellar processes remains constant .
however it is important to know when and where the energy is produced .
it is believed that for massive clusters most of the gravitational energy is produced by the infall shock during `` cluster formation '' .
theoretical n - body hydrodynamical estimates of the creation of entropy ( heat ) via shocks ( eke , navarro & frenk 1998 ) are in good agreement with the data for _ massive _ clusters . however at low masses ( ponman , cannon & navarro 1999 ) there is excess entropy compared to these simulations .
this effect has also been noted by many in the context of the luminosity temperature relation and the distribution of entropy in clusters ( david 1996 ) .
the theoretical work predicts @xmath11 while the data show that @xmath12 ( markevitchl 1989 ) and this can also be explained by ` extra ' ( , non - gravitational ) heat input ( kaiser 1991 ) .
detailed calculations ( loewenstein 1999 ; tozzi & norman ( tn ) 1999 ; suginohara & ostriker 1998 ; balogh 1999 ) indicate that this energy can be provided by star formation but requires a fairly high efficiency of conversion of stellar energy into heat since the @xmath13kev / particle that is required is about the maximum that can be provided by the supenova that produce the metals .
since entropy @xmath14 ( at the time of energy injection ) for a given heat input one gets more entropy if the density is low at time of energy injection .
these calculations strongly constrain the total amount of energy and the epoch at which this energy is injected . following the scenarios of tn , where the entropy is produced at a single epoch before the collapse of cluster ,
the mass scale of the transition from a shock ( gravity dominated ) entropy to a adiabatic model at a given mass scale is a strong function of the total entropy and the redshift of energy injection .
the appearance of a floor in the entropy at a mass scale of @xmath15 ( @xmath16kev ) requires in their model a redshift of @xmath17 for the injection of the energy .
suginohara & ostriker ( 1998 ) , who use a rather different set of constraints based on the absence of thermal cooling runaway in clusters also require a similar redshift for energy injection .
while the situation is unclear it seems as if certain structure formation scenarios ( the semi - analytic models presented by valageas & silk 1999 and wu , fabian & nulsen 1999 ) are not consistent with the required amount of extra energy being provided solely by supernova .
this additional energy can also be checked ( loewenstein 1999 ; davis 1999 ) by : * calculating the cluster mass _ vs. _ @xmath18 relation ( horner 1999 ) some of the ` extra ' energy goes into ` temperature ' and at low masses the observed temperatures are `` too high '' for the the n - body hydrodynamical model masses . *
measuring the surface brightness distribution : the gas tends to get puffed - up and have a flatter surface brightness distribution than predicted by the n - body hydrodynamical models . * comparing the ratio of gas to total mass _
vs. _ temperature .
if there has been no energy input other than gravity this should be roughly constant from object to object ( white 1993 ) .
however recent analysis ( mohr 1999 ; fujita & takahara 1999 ) find that the ratio of gas mass to total mass rises as @xmath19 , that is , more massive clusters have a larger fraction of their virial mass in gas ( as originally suggested by david 1995 and as indicated in fig .
2).the obvious solution is that some of the gas gets expelled in low mass systems .
this is further supported by the metallicity found in groups and the wide range of other parameters found in groups ( see next section ) .
the observations that low redshift elliptical galaxies are old and have mostly stopped forming stars at @xmath20 , the lack of evolution in the cluster metallicity since @xmath2 ( fig . 1 ) and the predominance of ellipticals in rich clusters all indicate that most of the star formation in clusters , and thus the source of the additional entropy occurred at @xmath21 . if the heating occurred only prior to collapse ( ponman 1999 ) one expects that low mass systems are isentropic ( balogh 1999 ) and thus have a steep temperature gradient .
_ this has not been observed and the best data to date indicate that groups are roughly isothermal .
_ however , as noted above , the effects of heating are much more efficient if they occur at early times , when the gas was of lower density .
it thus seems likely that the heat was created at moderate to high redshift sometime near the collapse of the cluster
. clearly , detailed calculations are necessary to determine the critical observations which can determine the amount of energy injected , the epoch at which it is injected and the source of the energy .
while clusters may be ` rare ' high density perturbations , groups are the average place in the universe ( see next section ) .
the effects of heating are most pronounced in groups , as evidenced by their high specific entropy , their relatively low gas fractions and flat surface brightness profiles .
thus * most * galaxies have been subjected to this process . as indicated in the semi - analytic models heating has profound implications on galaxy formation and the epoch of formation
the lesson that one learns is that one can not understand galaxy formation by modeling dark matter halos alone .
there are many other implications of this ; for example : when overdense regions `` get hot '' due to gravitational collapse ( cluster and group formation ) , star formation stops because the cooling times increase drastically ( cen & ostriker 1999 ) this gives a time dependent bias such that overdense regions which collapse early ( rich clusters ) cease star formation at high redshifts and the regions in which star formation proceeds occur in lower and lower overdensities at lower redshifts .
the effects of heating due to star formation has a strong feedback loop ( menci & cavaliere1999 ) and strongly effects the timescale in semi - analytic models . if the true energy is @xmath13kev / particle , as indicated by the entropy models ( loewenstein & mushotzky ; loewenstein 1999 ) then the mean energy per particle of most of the baryons exceeds the binding energy of even the most massive galaxies and galactic winds were important for all galaxies ( not just dwarfs ) . _ the x - ray data for groups and clusters strongly argues that the galaxies are poor tracers of mass , baryons and metals and use of them as tracers may give a very poor ` picture ' of the universe . _
the fe abundances are strongly peaked near @xmath22 solar and do not vary with redshift out to at least @xmath23 and perhaps @xmath2 ( donahue 1999 ) .
this lack of evolution of gas metallicity combined with the absence of major star formation in the same clusters strongly constrains the epoch of major metal formation .
the absence of large number of a stars means that the period of massive star formation has ceased in these systems @xmath24gyrs ( the age of a stars ) before the age of the objects at @xmath25 . for @xmath26 , @xmath27 , @xmath28
this implies that the epoch of metal formation is these clusters is at @xmath4 . the highest-@xmath29 cluster with measured fe abundance is at @xmath30 ( hattori 1997 ) and if we adopt this as the highest redshift at which cluster metallicity is not evolving this would constrain the epoch of metal formation in clusters to @xmath31 .
the data are fairly robust with respect to fe and si ( fukazawa 1998 ) moderate for s but poor for mg , o , ca , ar , ne .
recent analysis of asca , xte and sax data ( dupke & white 1999 ) give a strong indication that ni maybe be very overabundant with respect to fe .
one needs to combine these patterns with supernova models to derive constraints ( gibson , loewenstein & mushotzky 1998 ) and the relative numbers of type i and type ii supernova .
unfortunately the uncertainties in yields and sn models make strong statements very uncertain .
the existence of abundance gradients in si / fe in about @xmath32 of all clusters and groups ( finoguenov & ponman 1999 ; ezawa 1997 ) seem to show that the centers of clusters are relatively enriched in type i products and the outer regions are almost ` pure ' type iis .
the bulk of all the heavy elements is in the outer regions .
as shown in lowenstein & mushotzky ( 1996 ) the toal mass of si seen in clusters indicates that type
ii s have dominated the total metal production , consistent with the early origin of the metals in bursts of star formation .
however the situation is not yet fully resolved ( , arimtoto 1997 ) and in order to determine the true ratio of type i to type ii will require better data for o , ni and mg the elements that most sensitive to the relative number of different types of supernova .
groups are the average place in the universe , over 70% of all the galaxies in the local volume of space occur in groupings and it is very rare indeed to find truly isolated galaxies ( burstein 1999 pc .
of the 2700 galaxies in the tully nearby galaxies atlas , only 35 are not in a group , and none of those 35 are more than 0.75mpc from another giant galaxy ) . thus the results for groups should apply to _ the universe as a whole if the x - ray emittings groups are representative .
_ recent x - ray results show that large x - ray halos are a common property of bulge dominated groups ( mulchaey 1998 ) but are very rare in spiral dominated systems .
the masses determined for the groups are determined via the same techniques as for clusters ( via use of hydrostatic equilibrium and spatially resolved temperatures ) .
the derived masses are ` typically ' @xmath33 at the viral radius ( but only two groups have been measured to the virial radius and if the systems are really dominated by non - gravitational heat the extrapolation to the virial radius is uncertain see loewenstein 1999 ) .
the space density of these systems is large and in rough agreement with that predicted for virialized systems in lcdm models .
thus use of an x - ray selected survey can derive the mass density of virialized systems over a mass range @xmath34 ( from the lowest mass groups to the richest clusters ) .
there is a very poor correlation between the optical richness of the group and the x - ray temperature or luminosity ( mulchaey 1996 ) .
however there is an excellent agreement between the velocity dispersion as derived from observations of the numerous small galaxies in x - ray luminous groups ( mulchaey & zabludoff 1998 ) and the x - ray temperature .
this essentially implies that in groups the optical light does not trace mass at all !
the x - ray emission is a much more direct tracer of mass than optical light and determination of the group mass and mass distribution will be a strong test of all structure formation models .
this is a serious warning about the use of optical light to trace mass in the universe since most galaxies are in poor groups .
the baryonic fraction , the gas to stellar mass and the @xmath35 ratio all show wide variations in groups , but the mean values are distributed around the cluster average ( fig . 4 ) the metallicity in groups has a much larger range than that found in clusters and is often less ( fig . 5 ) .
the `` iron mass to light ratio '' ( ratio of total mass in fe ( stars+gas ) to the total light renzini 1997 ) tends to be much lower than in clusters .
if we believe that the stars visible today in the groups are the tracers of the population that produced the metals then this indicates that the groups have lost a substantial fraction of their metals .
the abundance pattern is not well determined ( davis 1999 ) but the si / fe ratio is smaller than in clusters ( fukazawa 1998 ) indicative of either a larger ratio of type i to type ii supernova than in clusters , which is hard to understand , or a loss of type ii material ( davis 1999 ) . since the groups tend to have a higher specific entropy than clusters ( normalized to the expected shock value ) and have a wider range in baryon fraction , all these signatures are consistent with mass ejection via the energy input from stellar sources associated with massive star formation ( sn+winds ) on the group mass scale . the low iron mass to light ratios strongly constrain evolution scenarios and indicate ( 1 ) either these systems have lost much of their baryons and metals ( winds from galaxies and groups ) and/or ( 2 ) they have been much more strongly effected by infall of primordial material than clusters .
_ and thus even the most massive `` normal systems '' have been affected by non - gravitational processes . _
the picture of the high redshift universe that one obtains from x - ray observations of groups and clusters in the @xmath1 universe is radically different from what was the standard picture five years ago .
clusters are old , and because they can hold onto all their material ( and even accrete from the igm ) , they are the only systems that are representative of the universe as a whole
. there are strong indications from observations of groups and clusters that non - gravitational effects are extremely important in the formation of galaxies and groups and that gravity is not dominant on all scales . one finds that at the group scale that galaxies are very poor tracers of mass .
the combination of all these effects may help to explain the difficulty of comparing numerical n - body simulations of structure formation against the available galaxy data .
finally the importance of feed - back from non - gravitational energy sources , as indicated in the semi - analytic models , can not be underestimated .
we anticipate that the study of large scale structure with x - ray samples will radically change our understanding of the universe . in the next year , with the new data from chandra ( now operating ) , xmm and astro - e there will be a qualitative change in the signal to noise , angular resolution and spectral resolution of the data and many of the conclusions reached in this review will be strongly tested .
i would like to thank the organizers for the chance to present the `` hy - energy '' view of the universe and for an exciting and well organized meeting .
i would like to thank my collaborators : m loewenstein , u. hwang , j. mulchaey , d. davis , c. scharf and d. horner who have done so much of this work .
andrew bunker is thanked for his assistance in assembling this manuscript .
anders , e. and grevesse , n. 1989 gecoa .. 53 .. 197 arimoto , n , matsushita , k , ishimaru , y , ohashi , t. , renzini , a 1997 ap . j. 477 .. 128 bahcall , neta a.and fan , x 1998 ap .
j. 504 .1 balogh , m. babul , a. and patton , d. 1998 astro - ph/9809159 blanchard , a sadat , r. bartlett , j.g , le dour , m.1999 astro - ph/9908037 cen , r and ostriker , j , p. 1999
j. 519 l.109 david , l , jones , c , forman , w 1996 ap .
j .. 473 , p.692 david , l , jones , c , forman , w 1995 ap .
j. 445 , p. 578
davis , d. , mulchaey , j. , mushotzky , r. 1999 ap j 511 34 dickinson , m .
et all 1998 ap .
j. submitted donahue , m. and voit , m astro - ph/9907333 donahue , m. , voit , m. , scharf , c. , gioia , i. , mullis , c. , hughes , j. and stocke , j. 1999 astro - ph/9906295 dupke , r. and white iii , r.e.1999 astro - ph/9907343 ebeling , h. allen , s. w. , edge , a. c , crawford , c. s. 1999 head 31.1303 eke , v , navarro , j , and frenk , c 1998 ap . j. 503 .. 569 ezawa , h. , fukazawa , y. makishima , k. , ohashi , t. , takahara , f. , xu , h yamasaki , n. y. 1997 ap .
j. 490l .. 33 finoguenov , a. and ponman , t.j .
1999 mnras.305 .. 325 fujita , y and takahara , f 1999 ap .
j. 519 , l.55 fukazawa , y .
1996 phd thesis university of tokyo fukazawa , y .
et al 1998 pasj 50 , 187 gibson , b. k.,loewenstein , m , mushotzky , r. f. 1997 mnras.290 .. 623 hattori , m. , ikebe , y. asaoka , i. , takeshima , t. boehringer , h. , mihara , t , .neumann , d. m. , schindler , s. , tsuru , t. tamura , t. 1997 natur.388 .. 146 henry , j. patrick 1997 ap
. j. 489 l 1 horner , d , mushotzky , r , scharf , c 1999 ap j 520 78 jones , l. r. , scharf , c. ebeling , h. , perlman , e. wegner , g. , malkan , m. and horner , d. 1998 ap . j. 495 .. 100 jorgenson , i. 1999 mnras .
306 .. 607 kaiser , n. 1991 ap .
j. 383 .. 104 loewenstein , m. 1999 ap .
j. in press loewenstein , m and mushotzky , r. f. 1996 ap . j. 466 .. 695 madau , p. this symposium markevitch , m 1998 ap .
j. 504 27 menci n and cavaliere , a. 1999 astro - ph/9906103 mohr , j. , mathiesen , b. and evrard , g. 1999 astro - ph/9901281 mulchaey , j. and zabludoff , a. 1998 ap .
j. 496 73 mulchaey , j , davis , d. , s. , mushotzky , r. and burstein , d.1996 ap .
j. 456 80 mushotzky , r. f. , loewenstein , m. 1997 ap .
j. 481 l .. 63 mushotzky , r. f. , scharf , c. a. 1997 ap . j. 482 l .. 13 nevalainen j. m. markevitch
m , and forman , w. 1999 astro - ph/9906286 ponman , t. j. , cannon , d. b and navarro , j. f 1999 nature 397 , 135 renzini , a. 1999 astro - ph/9902108 schindler , s. 1999 astro - ph/9908130 somerville r. and primack , j. 1999 astro - ph/9811001 stanford , s.a . ,
elston , r , eisenhardt , p , spinrad , h , stern , d , and dey , a 1997 aj .114 . ,
2232 suginohara , t. and ostriker , j .
j. 507 16 tozzi , p. and norman , c. 1999 astro - ph/9905046 tran , k - v et al 1999 ap . j. 522 39 valageas , p. and silk , j. 1999 astro - ph/9907068 white , s.d.m . , navarro , j. f. , evrard , a.e . ,
frenk , c.s .
1993 nature .366 .. 429 wu , k.k.s . ,
fabian , a. and nulsen , p astro - ph/9907112 vikhlinin , a. , mcnamara , b. r , forman , w. , jones , c. , quintana , h. , hornstrup , a. 1998 ap .
j. 498 l .. 21 | i review the evidence from clusters and groups of galaxies for ` relic ' evidence of the high redshift universe .
contrary to the received wisdom , clusters are old .
their x - ray emitting intergalactic medium in massive clusters is a reservoir of metal and energy injection from the main epoch of star formation .
there are strong indications that non - gravitational energy was extremely important in the formation of low - mass clusters and groups , implying that gravity did not dominate the process on all scales below that of massive , rich clusters .
the data indicate that most of the metals in the universe are not in stars but in the hot gas and thus the use of stars to trace metal formation and evolution misses most of the processed material .
the group data shows that galaxies are very poor tracers of mass on the scale of @xmath0 and since groups are the ` average ' place in the universe this probably holds for most of the visible universe . |
a domain wall is a prototype collective excitation in a physical system , and it is also the paradigm example of a topological soliton @xcite .
a domain wall can appear whenever there is a global symmetry that becomes spontaneously broken .
it constitutes the boundary that separates two neighbouring domains , in which the order parameter that detects the symmetry breaking has different values . in the case of a one dimensional heisenberg spin chain
the order parameter is a three component unit length vector .
when one of the three vector components vanishes identically , the heisenberg spin chain reduces to the xy spin chain @xcite .
a domain wall along the spin chain is a localised excitation that interpolates between two different , ordered spin states in which the order parameter has different constant values .
two major types of domain walls are commonly identified along the heisenberg chain @xcite .
these are called the bloch wall and the nel wall , respectively . in the case of a bloch wall
, the heisenberg spin variable rotates through the plane of the wall and in the case of a nel wall the rotation takes place within the plane of the wall itself .
domain walls that are mixtures of these two , can also occur along the heisenberg spin chain , while along the xy spin chain , only domain walls of the bloch type can be present . in this article
we demonstrate that the formation of super - secondary structures , during folding of a protein @xcite , can be understood in terms of a bloch domain wall that forms along a heisenberg spin chain , or along a closely related xy spin chain .
we propose that the spin chain interpretation of a protein backbone provides a systematic framework for understanding and describing the process of protein folding .
for this we employ all - atom force fields @xcite to scrutinise protein folding dynamics at the level of individual atoms and their oscillations .
we analyse the folding pathway using a combination of topological techniques and global analytic tools .
we isolate the collective oscillations which are pertinent for the folding process , from the noisy background of thermal and random individual atom fluctuations .
in particular , we illustrate how the individual atom motions become organised and combined into a coherent structural excitation which we identify as the bloch wall . as a concrete example we consider an @xmath0-helical subunit of the hiv envelope glycoprotein gp41 @xcite , with protein data bank @xcite ( pdb ) code 1aik .
there are six @xmath0-helical subunits in the biological assembly , shown in figure 1 .
we consider in isolation the subunit , for which the first amino acid has number 628 in the pdb file . in isolation , the subunit is unstable and starts folding .
the transmembrane glycoprotein 41 is itself a subunit of the retrovirus envelope protein complex . in the case of the hiv ,
its structure has been studied extensively .
it is presumed to have substantial biological relevance to the initial viral infection .
accordingly , the gp41 protein is a popular target for the development of an anti - viral immune response , to prevent and cure hiv infection .
however , medical applications are beyond the direct scope of the present study . here
we shall solely address and identify the physical mechanism , why and how an individual , initially @xmath0-helical subunit of 1aik becomes unstable in isolation , and starts folding . for our all - atom molecular dynamics simulations ,
we utilise the gromacs 4.6.3 package @xcite .
we analyse the results using a variety of topological techniques and analytical tools .
our approach derives from the mathematical structure of heisenberg and xy spin chains , in combination with properties of a discrete nonlinear schrdinger ( dnls ) hamiltonian @xcite .
in particular , the dnls equation that describes the local extrema of the hamiltonian , enables us to analytically identify the profile of the domain wall , and to interpret it in terms of dnls soliton @xcite .
here we present results from the detailed investigation of a particular example .
however , we expect our observations and conclusions to be generic .
indeed , the present results are fully in line with the previous findings @xcite obtained by using the coarse grained unres energy function @xcite in the case of protein a. the similitude of results that are obtained by analysing the protein folding process using different tools , built and based on phenomena with very different characteristic time and length scales , demonstrates that we have correctly identified the relevant collective motions that command the folding process .
we have performed _ in silico _ experiments to fold one c - chain subunit of the core structure of gp41 @xcite .
the structure comes from the hiv envelope glycoprotein with pdb code 1aik .
the amino acid sequence is @xmath1 these amino acids are assigned the numbers 628 - 661 in the pdb entry of 1aik .
we have used the molecular dynamics package gromacs 4.6.3 @xcite .
we have analysed in detail a number of 80 ns long trajectories , with the crystallographic pdb conformation as the initial condition .
we have chosen the length of the trajectories by inspecting , when major structural deformations take place .
we have employed three different force fields , to eliminate force - field based artifacts .
these are the united - atom force field gromos53a6 , and the all - atom force fields charmm27 and opls / aa .
the 1aik subchain that we have investigated in detail , consists of 34 amino - acid residues , with pdb numbers 628 - 661 .
there are 16200 atoms in the entire system that we have simulated , including the solvent .
the simulation box has dimensions @xmath2 @xmath3 .
this ensures that there is a 2 nm minimal distance between the protein atoms and the box walls , with periodic boundary conditions .
we have described the solvent using the spc water model @xcite .
we have neutralised the system at a salt concentration of 0.15
we have used steepest - descent for initial energy minimisation .
the system was warmed up to 290 k by a simulated annealing in a 100 ps position - restraint simulation .
we have chosen this relatively low temperature value for a better control of random thermal noise but without forgoing the underlying physical phenomena . for temperature control
we have employed the berendsen - thermostat with a time constant 0.1 ps , and for pressure coupling the berendsen - barostat with a pressure set to 1 bar and a time constant 0.5 ps .
constraints on all bonds were imposed with the lincs algorithm @xcite .
we have used the particle mesh ewald ( pme ) method @xcite to compute the long - range electrostatic interactions , with van der waals and coulomb cutoff radii of 0.9 nm .
for the 80 ns production run with a time step of 2 fs , that we analyse here in detail , we have changed the thermostat to v - rescale and the barostat to parrinello - rahman , keeping the initial time constants , to ensure the generation of a proper canonical ensemble @xcite .
we have recorded the coordinates every 20 ps , which gives rise to 4000 frames that form the basis for our analysis .
we have introduced , employed and developed a number of topological tools and analytic techniques to analyse and interpret the results of our gromacs simulations .
we monitor the evolution of the protein geometry using frenet frames which are based on the backbone c@xmath0 atoms @xcite .
the framing depends _ only _ on the c@xmath0 atom coordinates @xmath4 , where @xmath5 labels the residues and @xmath6 in the case of 1aik . at a given @xmath4 the frame
consists of the unit backbone tangent ( @xmath7 ) , binormal ( @xmath8 ) and normal ( @xmath9 ) vectors , defined as follows , @xmath10 @xmath11 @xmath12 our aim is to identify and isolate the collective multi - atom motions that drive the protein folding process , from the background of the various random fluctuations .
we expect that such coherent motions and oscillations have characteristic time scales , which are much longer than the period of an individual atom covalent bond oscillation . in average , over the relevant time scales , the distance between two consecutive c@xmath0 atoms can then be taken to be nearly constant , and equal to @xmath13 thus , at relevant time scales , the backbone dynamics can be described entirely in terms of the virtual backbone bond and torsion angles @xmath14 and @xmath15 , as the _ complete structural order parameters _
these angles are defined as follows , @xmath16 where @xmath17 { \label}{eq : s}\ ] ] conversely , the frame vectors ( [ eq : t])-([eq : n ] ) can be expressed in terms of these two order parameters iteratively , using the _ discrete frenet equation _
@xcite @xmath18 and the c@xmath0 backbone is calculated from @xmath19 unlike the tangent vector @xmath7 , the normal and binormal vectors @xmath9 and @xmath20 do not appear in equation ( [ eq : dffe ] ) .
thus , if we rotate these two vectors simultaneously around the vector @xmath7 , the c@xmath0 geometry remains intact and only the way how it is framed changes .
in particular , rotation by @xmath21 constitutes the discrete @xmath22 gauge transformation @xmath23 this transformation has been previously used extensively , to analyse protein loop structure @xcite .
it will also be used in the sequel . according to ( [ eq : dffe ] ) the entire c@xmath0 backbone geometry
is determined by the tangent vectors @xmath24 .
thus , following @xcite we may visualise the backbone geometry in terms of these vectors : we take the base of @xmath7 to be at the location @xmath25 of the @xmath26 c@xmath0 atom .
we identify the tip of @xmath7 as a point on the surface of a unit two - sphere @xmath27 that is centered at the point @xmath25 .
we orient the coordinate system on the sphere so that the north - pole coincides with the tip of @xmath7 .
thus , the north - pole is always in the direction of the next c@xmath0 , which is at the site @xmath28 . we proceed to characterise the direction of the next tangent vector @xmath29 _ i.e. _ the direction from @xmath28 towards the c@xmath0 atom at site @xmath30 , in terms of the longitude and latitude angles of the @xmath26 two - sphere @xmath31 . for this
, we translate the center of @xmath27 from @xmath32 towards its north - pole , and all the way to the location @xmath28 of the @xmath33 c@xmath0 atom , without introducing any rotation of the sphere .
we then record the direction of @xmath29 as a point on the surface of the translated @xmath27 .
this defines the coordinate values ( @xmath34 ) , that determine how the backbone chain turns at site @xmath28 , to reach the @xmath35 central c@xmath0 atom at the point @xmath30 : the angle @xmath14 measures the latitude of @xmath29 on the translated two - sphere @xmath27 , from its north pole .
the angle @xmath15 measures the longitude of @xmath29 , starting from the great circle that passes both through the north pole and through the tip of the binormal vector @xmath8 .
when we repeat the above procedure for all c@xmath0 atoms , we obtain a ( @xmath36 ) distribution that characterises the overall geometry of a protein backbone . for a visualisation of this distribution we employ the geometry of a stereographically projected two - sphere
: we project the @xmath37 coordinates from the south pole to the tangent plane of the north pole , of the two - sphere . if ( @xmath38 ) are the coordinates of this tangent plane the projection is defined by @xmath39 when we perform the projection for all c@xmath0 atoms in all crystallographic protein structures in pdb that have been measured with resolution better than 2.0 , we arrive at the statistical angular distribution that we show in figure [ fig-2 ] .
it is the landscape for the shape of the protein backbones from the crystallographic data in pdb . by the way it is obtained
, the crystallographic protein structure should be very close to a stationary minimum of the ensuing gibbs free energy .
thus the figure [ fig-2 ] should be the collective landscape of stationary , minimum - energy protein structures .
we observe that the pdb data is concentrated in an annulus which is roughly between the circles @xmath40 and @xmath41 .
the exterior of the annulus @xmath42 is an excluded region , the ensuing conformations are subject to steric clashes .
the interior @xmath43 is sterically allowed but in practice excluded in pdb structures .
note that regular structures such as @xmath0-helices and @xmath44-strands are distinguished as highly localised regions in the figure [ fig-2 ] , with @xmath45 for @xmath0-helices and @xmath46 for @xmath44-strands .
different regions in figure [ fig-2 ] can be connected by loops , which can be considered as trajectories along the variables ( @xmath47 ) .
we have found that loops have the tendency to encircle the inner circle . in figure [ fig-3 ]
a ) we show , as an example , a generic loop that connects the right - handed @xmath0-helical region with the @xmath44-stranded region . to describe a backbone segment analytically , we combine its c@xmath0 s bond and torsion angles into the three component unit vectors @xmath48 we interpret these vectors as the local order parameters along an imaginary _ linear _ one dimensional heisenberg spin chain , labeled by the index @xmath49 .
this converts the c@xmath0 geometry into a configuration along a linear heisenberg chain in a one - to - one manner : in figure [ fig-3](b ) we have sketched how the ( generic ) trajectory shown in figure [ fig-3](a ) appears , figuratively , in terms of such a heisenberg spin chain configuration .
since the spin variable ( [ si ] ) takes values only in the annulus @xmath50 of the two - sphere @xmath51 , it is apparent that a loop can be _ de facto _ identified as a domain wall akin the bloch wall along a heisenberg chain .
the loop then interpolates between the two different regular secondary structures , denoted by state a and state b respectively , as shown in the figures 3 .
the amino acid side chains can be similarly interpreted in terms of a one dimensional linear spin chain .
in fact , there are several ways to identify the spin chain variable . here
we utilise the directional vector that points from the c@xmath0 atom at @xmath32 towards the ensuing c@xmath44 atom , located at @xmath52 .
this vector can be introduced for all amino acids except glycine ( g ) ; note that there is no glycine in ( [ lets ] ) .
we start with the unit vector @xmath53 we recall the c@xmath0 based discrete frenet framing with the coordinates ( @xmath34 ) and represent ( [ sideo3 ] ) as three component unit vectors in this coordinate system , @xmath54 here ( @xmath55 ) are the spherical coordinates of the @xmath26 c@xmath44 atom , on the surface of the c@xmath0 centered two - sphere @xmath27 . in figure [ fig-4 ]
we show the distribution of the vectors ( [ cbetai ] ) on the surface of the two - sphere , for all those crystallographic pdb structures that have been measured with resolution better than 1.0 .
note that the sphere is the same as in figures [ fig-2 ] and [ fig-3](a ) but now there is no stereographic projection .
we can interpret the distribution in figure [ fig-4 ] as the c@xmath44 landscape of stationary folded protein structures with minimum gibbs energy .
the highly localized character of the distribution shows that there is a very strong correlation between the c@xmath0 ( backbone ) geometry and the c@xmath44 ( side chain ) geometry .
accordingly , the ground state structures of the corresponding heisenberg spin chain hamiltonians must be very similar .
we proceed to introduce a set of o(2 ) spin variables for the side - chain c@xmath44 . for this
we define the projection of ( [ sideo3 ] ) onto the normal plane at the position of the @xmath26 c@xmath0 , @xmath56 for the next c@xmath44 along the chain , we introduce similarly the vector @xmath57 and compute its projection onto the _ same _ normal plane at the position of the @xmath26 c@xmath0 , @xmath58 we then define the relative angle @xmath59 , @xmath60 as shown in figure [ fig-5 ] , @xmath59 is the dihedral angle @xmath61 we note that the construction resembles that of newman projection in stereochemistry . in analogy with figure [ fig-3](b )
we identify the variable @xmath59 as an order parameter for a linear o(2 ) xy spin chain , @xmath62 like the heisenberg model , the xy model supports domain walls that interpolate between two configurations where the order parameter has different constant values .
the domain wall of the xy model is akin the bloch domain wall of a heisenberg model .
figure [ fig-4 ] shows that the ground state structure of the side chain xy model is closely related to that of the backbone heisenberg model , in the case of crystallographic pdb protein structures .
the formation , evolution and structure of a loop along a folding protein can be monitored in terms of topologically determined folding indices .
here we are interested in two particular examples of folding indices , one that relates to the backbone geometry and another one that relates to the side chain geometry . in the case of a heisenberg spin chain
, there is a topological index akin a winding number that characterises and classifies its bloch domain walls . for the c@xmath0 bloch wall shown in figure [ fig-3 ]
a ) , b ) this topological index counts the net number of times the corresponding trajectory encircles around the annulus in the figure _
i.e. _ around the north - pole of the two - sphere .
we remind that due to steric constraints , the heisenberg variable ( [ si ] ) takes values in the annulus shown in figures [ fig-2 ] , [ fig-3 ] a ) .
we also recall that for the first homotopy class of a circle @xmath63 which justifies the introduction of a topological concept .
analytically , we may assign to each loop , and more generally to a backbone segment , between residues @xmath64 and @xmath65 the following folding index @xmath66 @xcite , @xmath67 { \label}{eq : gamma}\ ] ] where @xmath68 here @xmath69 $ ] denotes the integer part of @xmath70 .
note that @xmath71 is the total rotation angle ( in radians ) that the projections of the c@xmath0 atoms of the consecutive loop ( segment ) residues make around the north pole .
the @xmath72 label the first and last residue of the loop .
commonly these are the last _ resp .
_ first residues in the preceding and in the following regular secondary structures .
the folding index is a positive integer when the rotation is counterclockwise , and a negative integer when the rotation is clockwise .
the folding index can be used to classify individual loop structures and backbone segments , even entire protein backbones @xcite .
note that the folding index is normalised so that it is equal to twice the number of times the vector in figure [ fig-3](b ) rotates around its axis , when the spin structure traverses a domain wall _
i.e. _ it assigns an even integer to the @xmath63 winding number , in the case of a closed trajectory .
for example , for the trajectory shown in figures [ fig-3 ] the folding index has the value -1 . for a loop connecting an @xmath0-helix and a @xmath44-strand ,
the folding index is generically an odd integer . for a loop connecting two @xmath0-helices , or two @xmath44-strands ,
the folding index is generically an even integer . in the case of the side chains ,
we utilise the c@xmath44 angular xy spin variable ( [ etai ] ) to define a similar topological folding index @xcite . for this
, we first choose a reference residue , _
e.g. _ the @xmath73 residue along the backbone . starting with this reference residue ,
we then evaluate the accumulated _ total _ angle @xmath74 over a segment with @xmath75 residues , @xmath76 and we define the ensuing index by @xmath77 { \label}{eq : etaprime2}\ ] ] again , the index acquires its topological justification from the fact that @xmath78 .
the dihedral @xmath79 , the accumulated total angle ( [ eq : etaprime ] ) and the ensuing index ( [ eq : etaprime2 ] ) can all be used to study and classify loop structures , protein segments , and entire proteins .
sect : landau a generic all - atom molecular dynamics simulation of a folding protein contains a wide range of intermediate conformations .
some of these are essential for the correct folding pathway , while some are merely random transients with no inherent relevance to the folding process _ per se_. in order to identify the relevant conformational processes , we need systematic methods that smooth over and weave out the irrelevant random fluctuations .
for this we recall the standard wilsonian universality arguments @xcite , to deduce the form of the landau free energy that emerges from the thermodynamical gibbs free energy . in the present context
, the derivation is based on the following two assumptions @xcite : @xmath80 we assume that the characteristic length scales that are associated with spatial variations and deformations along the protein backbone around its thermal equilibrium configuration , are large in comparison to the covalent bond lengths .
this presumes that there are no abrupt edges but only gradual slowly varying bends and twists along the backbone . from figures [ fig-2 ] and [ fig-4 ]
we conclude that the steric constraints between the backbone and the side chain atoms act as powerful inhibitors of sharp , edgy motions .
@xmath80 we also assume that the individual c@xmath0 virtual bond length oscillations have characteristic time scales , which are very short in comparison to the time scale which characterise a folding process .
the characteristic time scale of a random covalent bond oscillation is around ten femtoseconds while in our simulations we record the individual atomic coordinates every 20 picosecond .
we have tested that conformational changes which take place over shorter time scales , do not affect our conclusions .
accordingly , we may adopt ( [ eq : aveva ] ) as the ( time averaged ) value for all the nearest neighbour c@xmath0-c@xmath0 distances .
it has been shown that in the case of crystallographic pdb structures , the bond and torsion angles ( @xmath34 ) form a _
complete set of structural order parameters _
accordingly , in the vicinity of a gibbs free energy minimum we may expand the free energy in terms of these angles .
for this we consider the response of the interatomic distances to variations in these angles , with @xmath81 where @xmath82 is the distance between _ any _ two c@xmath0 atoms @xmath83 and @xmath84 along the backbone .
suppose that at a given local extremum of the free energy , the c@xmath0 bond and torsion angles have the equilibrium values @xmath85 we then consider a non - equilibrium conformation where the ( @xmath34 ) deviate from these extremum values .
we denote the deviations by @xmath86 when the deviations are slowly varying in space , _ i.e. _ ( [ devi ] ) are small , we may taylor expand the gibbs free energy around the extremum , @xmath87 \ = \ ] ] @xmath88 @xmath89 @xmath90 the first term in the expansion evaluates the free energy at the extremum . since ( @xmath91 ) correspond to the extremum , the second term vanishes .
thus we are left with the following expansion of the averaged free energy ,
@xmath92 @xmath93 @xmath94 when the characteristic length scale of spatial deformations around a minimum energy configuration is large in comparison to a covalent bond length , we may _ re - arrange _ the expansion ( [ fene ] ) in terms of the differences in the angles as follows : first come local terms , then come terms that connect the nearest neighbours , then come terms that connect the next - to - nearest neighbours , and so forth .
when we re - order the expansion ( [ fene ] ) in this manner and demand that the free energy is invariant under local rotations in the ( @xmath95)-plane , we conclude @xcite that to the leading order the expansion of the gibbs free energy _ necessarily _ coincides with the energy of the following discrete nonlinear schrdinger equation @xcite @xmath96 @xmath97 @xmath98 + \sum\limits_{i=1}^{n-1 } ( \kappa_{i+1 } - \kappa_{i})^2 { \label}{e2old}\ ] ] this functional form of the free energy is simply the most general landau free energy that one can write down using the available variables ( @xmath99 ) , in a manner which is consistent with the symmetry principle that a local rotation of the ( @xmath100 ) frames has no effect on the backbone geometry .
the corrections to ( [ e1old ] ) include next - to - nearest neighbours couplings and so forth , which are higher order terms from the point of view of our systematic expansion .
we note that the expansion ( [ e1old ] ) has the property that in continuum limit it yields the coleman - weinberg derivative ( low momentum ) expansion @xcite @xmath101 where , following @xcite , we have identified the bond angle with the complex scalar field @xmath102 and the torsion angle with the u(1 ) gauge field @xmath103 , in the continuum limit .
we _ emphasize _ that the approximation ( [ e1old ] ) is valid in the limit of slow spatial variations ( low momentum ) .
that is , as long as there are no abrupt , sharp edges but only gradual bends and twists along the backbone . in particular , long range interactions are accounted for as long as they do not cause any localised sharp buckles along the backbone , and the angular variations respect the steric constraints .
the wilsonian universality arguments are sufficient to conclude that in the limit of slowly varying backbone geometry _ any _ complete all - atom force field can be approximated by the energy function ( [ e1old ] ) .
the parameters @xmath104 , @xmath105 , @xmath106 , @xmath107 , and @xmath108 depend on the atomic level physical properties and the chemical microstructure of the protein and its environment . in principle
, these parameters can be computed from this knowledge .
but as always in the case of a landau free energy , it remains a challenge to compute these parameters from the all - atom level .
sect : solitons the free energy ( [ e1old ] ) relates to the dnls energy function @xcite .
the non - linear , quartic bond angle contribution is the familiar double - well potential that gives rise to a spontaneous breakdown of the @xmath22 symmetry @xmath109 the spontaneous breakdown of this discrete symmetry is pivotal for the emergence of a loop structure , in the case of proteins .
it gives rise to a bloch wave that interpolates between the two ground states @xmath110 .
more generally , the quartic potential admits a non - symmetric profile of the form @xmath111 here @xmath83 and @xmath84 are the positions of the minima of the quartic potential , and @xmath112 is a force constant . by carefully taking the continuum limit of the c@xmath0 lattice ,
_ i.e. _ the limit where ( [ eq : aveva ] ) becomes small , and by introducing a mass - scaled variable @xmath113 with @xmath108 representing the effective mass of a residue , the pertinent dnls equation becomes @xmath114 where @xmath115 is the arc length parameter along the backbone . with @xmath116 and @xmath117
the solution of equation ( [ eq : schr ] ) is @xmath118 { \label}{kappaprof}\ ] ] where @xmath119 is the position of the inflection point , _ a.k.a . _
the center of a kink . in terms of the original variables and parameters @xmath120 } { \label}{eq : cosh}\ ] ] this is known as the dark soliton solution of the nonlinear schrdinger equation .
it interpolates between the asymptotic values which correspond to the ( local ) minima of the potential , @xmath121 in the case of a protein , the soliton describes the bond angle profile of a super - secondary structure such as ( @xmath0-helix)-(loop)-(@xmath44-strand ) shown in figure [ fig-3 ] ; the parameters have the values @xmath122 and @xmath123 ( radians ) for the states @xmath124 and @xmath125 , shown in the figure . in the case of a protein chain , the arc length @xmath115 becomes replaced by a discrete variable which is equal to the position of the ensuing c@xmath0 in the sequence .
the variables @xmath14 and @xmath15 are also mutually interacting , according to ( [ e1old ] ) .
the soliton is constructed as the minimum of @xmath126 in equation ( [ e1old ] ) @xcite .
it is the solution of a system of @xmath127 nonlinear equations in @xmath127 unknowns , where @xmath128 is the number of residues . in order to obtain the solution
, we first solve for @xmath15 in terms of @xmath14 , @xmath129 = \frac{p}{r + q\,\kappa^2_{i } } \
\frac{u } { 1 + v \ , \kappa^2_{i } } { \label}{tauk}\ ] ] with @xmath130 and @xmath131 . by inserting equation ( [ tauk ] ) into equation ( [ e1old ] ) , the torsion angles @xmath132 are eliminated and we obtain a system of equations for the bond angles @xmath133 , @xmath134}{d\kappa_i^2 } \kappa_i \ \ \ \ \ ( i=1, ... ,n ) { \label}{nlse}\ ] ] where @xmath135 and @xmath136 = \frac{p}{r + q \ ,
\kappa^2 } + 2 ( 1- \lambda m^2 ) \kappa^2 + \lambda \ , \kappa^4 { \label}{u}\ ] ] here we recognize the discretised structure of equation ( [ eq : schr ] ) .
the difference is in the first term on the right - hand side in equation ( [ u ] ) .
however , it turns out that in the case of proteins , its effect is not that pronounced as the effect of the other terms ; it turns out that the first term is small in value when compared to the other two .
we can construct the profile of the dark soliton solution to equation ( [ nlse ] ) numerically , by following the iterative procedure introduced in reference @xcite ; the explicit form of the solution is until now unknown , in terms of elementary functions .
however , we obtain an _ excellent _ approximation @xcite by _ naively _ discretising the continuum dark nonlinear schrdinger equation soliton ( [ eq : cosh ] ) @xmath137 + \mu_{2 } \ , \exp\left [ - \sigma_{2 } ( i - s)\right ] } { \exp\left [ \sigma_{1 } ( i - s ) \right ] + \exp\left[- \sigma_{2 } ( i - s)\right ] } { \label}{eq : bond2}\ ] ] here @xmath138 \ { \rm mod}(2\pi)$ ] are parameters , which determine the amplitude of the variation of @xmath133 and the asymmetry of the inflection regions .
the parameters @xmath139 and @xmath140 are related to the inverse of the range of the inflection region .
we remark that in the case of proteins , the values of @xmath141 are determined entirely by the adjacent helices and strands .
furthermore , far away from the soliton center we have in analogy with ( [ asymp ] ) @xmath142 the corresponding torsion angles are evaluated in terms of the bond angles using equation ( [ tauk ] ) . note that in the case of proteins , the profile of equation ( [ eq : bond2 ] ) becomes monotonically increasing when we add multiples of @xmath143 to the experimental values . since the values of @xmath144 s are defined @xmath145 , this does not affect the backbone geometry .
the integer number of times the monotonically increasing variable @xmath14 covers its fundamental domain @xmath146 counts the number of solitons along the backbone .
recall that negative values of @xmath14 are related to positive values of @xmath14 by @xmath22 symmetry ( [ eq : dsgau ] ) .
finally , _ only _ the parameters @xmath139 and @xmath140 in ( [ eq : bond2 ] ) are intrinsically specific parameters for a given loop .
but they specify only the length of the loop , not its shape which is determined _ entirely _ by the functional form of equation ( [ eq : bond2 ] ) and , as in the case of @xmath147 and @xmath148 , they are combinations of the parameters in equation ( [ u ] ) . in the expression ( [ tauk ] ) of the torsion angles @xmath149 , there are only two independent parameters @xmath150 and @xmath151 .
consequently the profile of @xmath15 is determined entirely by @xmath14 , and by the structure of the adjacent regular secondary structures .
it has been shown @xcite that most crystallographic protein structures in pdb can be described with very high precision in terms of such soliton profiles as their modular building blocks .
moreover , it has been found that in the ensuing soliton profiles , the number of parameters is generically much smaller than the number of residues . thus , the energy function ( [ e1old ] ) has a very high predictive power , in describing folded proteins structures in pdb .
its predictions can be subjected to stringent experimental scrutiny , both in the case of static and dynamic proteins .
we now proceed to demonstrate , that all the concepts and structures we have identified are observed during an all - atom simulation of protein folding .
we start with individual atom level scrutiny , even though our goal is to identify and model those _ collective _ conformational deformations that cause a protein to fold .
we inquire how does self - organisation , in the case of a protein , relate to universal concepts such as formation of domain walls along spin chains .
we study how accurately can the dynamics and structure of the important collective deformations be modelled by soliton profiles such as the one described by the dnls equation .
we have subjected the single c - chain subunit of the core structure of gp41 @xcite with pdb code 1aik and residues 628 - 661 , to detailed all - atom simulations .
we have used the gromacs 4.6.3 .
package @xcite with three different force fields gromos53a6 , charmm27 and opls / aa having thus the protein described by 383 , 573 and 609 atoms , respectively .
the final production simulation models 80 ns of the protein evolution . we have concluded this to be sufficient , to identify and analyse the important structural deformations that can take place . in figure [ fig-6](a )
we show the secondary structure of the final conformations that we have obtained using the three force fields , in our 80 ns simulations .
we observe a clear deformation , in a segment that consists of the first 10 residues from the n - terminal ( upper part in the figure ) .
in figure [ fig-7 ] we display the weighted root - mean - square - deviation ( rmsd ) of the protein backbone in the three force fields , @xmath152^{1/2 } { \label}{rmsd}\ ] ] here @xmath153 is the position of the atom @xmath49 at time @xmath154 , and @xmath155 where @xmath156 are the individual atom masses , and @xmath157 is the total mass of the backbone .
the deformation is most intense in the gromos53a6 force field . with this force field ,
the initial @xmath0-helical structure begins collapsing within 45 ns . with the opls / aa force field
, we find that the deformation starts after around 40 ns . in the case of the charmm27 force field
, the helix tends to remain intact within the selected time range .
the deformation begins only after a substantially longer simulation .
apparently , this force field has a tendency to produce structures that have an overly @xmath0-helical content .
after extended comparisons of the three force fields , including different time steps and simulation lengths , we have chosen a 80 ns gromos53a6 force field trajectory with 2 fs time step , for the final production simulation that we analyse here .
qualitatively , the results that we present are independent of the force field and time step that we have chosen . in figure [ fig-8 ]
we show the results from a do.dssp @xcite secondary structure analysis , in the case of the gromos53a6 simulation .
the following _ qualitative _ observations can be made : @xmath80 after around 4 - 5 ns corresponding to frames 200 - 250 , there is an initial formation of a coil structure , according to do.dssp classification .
the coil becomes initially stabilised around residue number 29 which corresponds to amino acid number 656 in the pdb file .
the coil is connected to the c - terminal with a bend . at around 8 ns ( around frame 400 )
there are helical fluctuations in this structure towards n - terminal , and at around 22 - 24 ns ( frames 1100 - 1200 ) the coil structure moves back towards the c - terminal .
the motion takes place in two steps , at around frame 1200 and then again at around frame 2800 after which the coil disappears , by merging into the apparently random fluctuations of the c - terminal residues . at the level of do.dssp secondary structure analysis the coil which emerges near the c - terminal and propagates along the backbone , is putatively akin the propagating loop structure that has been previously identified and studied in coarse grained unres simulations of the protein g related albumin - binding domain with pdb code 1gab @xcite .
in particular the unres simulation @xcite identifies a displaced protein loop as a localised structure with a profile that can be described by the soliton solution of the discrete nonlinear schrdinger equation ( [ tauk ] ) , ( [ nlse ] ) .
the simulation demonstrated that when the loop - soliton moves along the protein lattice , with cells matching the residues , there are waves that are emitted in its wake as vibrations in the lattice structure .
these waves drain the kinetic energy of the soliton , and cause it to decelerate .
eventually the kinetic energy of the soliton becomes depleted , and it can no longer cross over the energy barriers between lattice cells and becomes localised around a particular set of lattice cells .
the energy barriers that prevent the soliton from translating along the backbone lattice were identified as peierls - nabarro barriers @xcite in @xcite . in the present case of the c - terminal coil ,
there is apparently a peierls - nabarro barrier that stops and prevents the coil that is supposedly modelled by a dnls , from propagating away from the c - terminal beyond the residues 28 - 29 . instead it becomes initially trapped , then moves towards the c - terminal and dissolves there .
the soliton moves step - wise , it s crossing - over the ensuing peierls - nabarro barriers is boosted by thermal fluctuations .
the soliton crosses a barrier whenever the amplitude of its thermal fluctuations exceeds the barrier - specific threshold value .
@xmath80 at around the same time when the c - terminal coil forms , we observe a turn deformation that forms and proceeds away from the n - terminal , and then fluctuates thermally between residues 5 and 10 . after around 20 ns ( frame 1000 ) of simulation time
, there is a rapid extended turn - like fluctuation that connects the n - terminal with a localised structure which is identified as a short coil by do.dssp .
this is an apparent dnls soliton , emerging at the end of the extended turn - like structure and stabilising around the residues 6 - 8 ( residues 634 - 636 in the pdb file ) .
we observe initially relatively strong fluctuations in the residues between the putative soliton and the n - terminal .
but the amplitudes of these fluctuations become damped and after around 50 ns ( frame 2500 ) there are only minor fluctuations in the soliton .
there is a bend between the soliton and the n - terminus which constitutes a peierls - nabarro barrier , high enough to prevent the soliton from moving towards the n - terminal , step - wise by thermal fluctuations .
we proceed to analyse the dynamics _
quantitatively _ , and we start with the backbone folding index ( [ eq : gamma ] ) . for this
we have divided the backbone into segments of varying length , and computed the folding index over the segments during the 80 ns time evolution .
examples of results are shown in figure [ fig-9 ] , where we plot the numerical values of the folding index density ( [ eq : foldind ] ) .
the first segment consists of the sites 4 - 11 corresponding to residues 632 - 639 in pdb .
this segment covers the n - terminal soliton structure ( see figure [ fig-8 ] ) .
the second segment consists of the sites 24 - 30 ( 652 - 658 in pdb ) .
this covers the segment where the c - terminal coil initially appears in figure [ fig-8 ] .
we observe the following : * figure [ fig-9](a ) : * initially , the folding index of the segment 4 - 11 vanishes .
but in the vicinity of frame 1000 , coinciding with the formation of the n - terminal coil / soliton in figure [ fig-8 ] , the folding index starts fluctuating between the values @xmath158 and @xmath159 .
we note how the pattern of the oscillations reflects the structural changes in the region between the coil and the n - terminal , shown in figure [ fig-8 ] : there are first fluctuations between turn and bend , during frames 1000 - 1500 .
when the oscillations in the values of the folding index diminish and vanish near the frame 1500 , we observe a formation of helical structure between the coil and the n - terminal in figure [ fig-8 ] .
the folding index then starts oscillating between the values @xmath160 , and this corresponds to a frame segment where the helix converts into a turn in the figure [ fig-8 ] . at around the frame
2500 , the folding index finally stabilises to the final value @xmath161 .
this stabilisation concurs with the formation of a bend between the coil and the n - terminal , in figure [ fig-8 ] .
* figure [ fig-9](b ) : * we observe the increase of folding index from @xmath162 to @xmath163 near frame 200 , and subsequent decrease back to @xmath162 near frame 1300 . according to figure [ fig-8 ] ,
these transitions coincide with the appearance of the c - terminal soliton , and its subsequent propagation towards the c - terminal , away from the segment 24 - 30 .
accordingly , we have found that the variations in the values of the folding index , in particular in the case of the n - terminal soliton , coincide with the structural deformations that take place along the backbone segment which is located between the soliton and the n - terminal . in particular
, the final stabilisation of the folding index concurs with the crossing over the peierls - nabarro barrier and subsequent stabilisation of the soliton , according to figure [ fig-8 ] .
moreover , the c - terminal soliton emerges with @xmath163 and remains stable until the peierls - nabarro barrier crossing takes place .
the evolution of the ensuing folding index is also fully in line with the results that we deduce from do.dssp .
we conclude that the behaviour in the backbone segment , surrounding the soliton , directly correlates with the topological character of the soliton , in both cases .
finally , in figure [ fig-10 ] we show the folding index density ( [ eq : foldind ] ) over the _ entire _ backbone and for all the 4000 frames .
we observe that : @xmath80 initially , the folding index vanishes .
this is consistent with the @xmath0-helical structure of the 1aik sub - chain .
but there is a sudden initial transition to the value @xmath163 , presumably reflecting the initial stages of c - terminal soliton formation .
@xmath80 up until the frame @xmath164 the folding index tends to vanish .
but there are fluctuations , mainly between values @xmath165 which reflect the various processed that take place near the terminals .
@xmath80 in the vicinity of frame 2700 there is a transition , and the value of the folding index starts stabilising toward the value @xmath163 .
this stabilisation concurs with the stabilisation of the n - terminal soliton , and the final departure of the c - terminal soliton .
the fluctuations also shift , oscillating between @xmath166 and @xmath167 and this shift is identified by the yellow arrows in the figure . in figure [ fig-11 ]
we show a close - up to the last 50 frames in figure [ fig-10 ] .
it confirms the stabilisation of the folding index towards the value @xmath163 , with occasional fluctuations where @xmath168 or @xmath169 . in figure [ fig-12 ] ,
following figure [ fig-3](a ) , we show the full trajectory for the entire final frame 4000 .
the trajectory starts from the n - terminal which is located in the @xmath44-stranded region of figure [ fig-2 ] .
it moves over to the @xmath0-helical region , then return to the @xmath44-stranded region to encircle the north pole .
finally , the trajectory merges and ends with the @xmath0-helical region .
the trajectory confirms that the final structure at frame 4000 indeed does support a twisting @xmath170 and that the twisting is furthermore located at the n - terminal soliton .
the stabilisation of the folding index to the value @xmath163 confirms the global character of the remaining n - teminal soliton structure : there is a total twisting by @xmath170 along the final backbone , in comparison to the initial configuration and including the terminal residues , and this twisting is localised on the n - terminal soliton .
moreover , we observe that the n - terminal residues are in a @xmath44-stranded position while the c - terminal residues are in the @xmath0-helical position .
figure [ fig-4 ] shows the landscape of the ground state ( crystallographic structure ) c@xmath44 atom directions in the c@xmath0 centered discrete frenet frames .
figure [ fig-13 ] shows how the directions of the c@xmath44 evolve during our entire gromos53a6 simulation .
we find it remarkable how similar the dynamical landscape of figure [ fig-13 ] is with the static ground state landscape shown in figure [ fig-4 ] : the direction of c@xmath44 nutates _ tightly _ around its static ground state landscape . clearly , there must be strong correlations between the backbone c@xmath0 and side chain c@xmath44 , during the entire dynamics .
accordingly , the information content in the angles @xmath59 in ( [ etai ] ) , ( [ etai2 ] ) should correlate strongly with the c@xmath0 geometry changes during the dynamical process .
more generally , we expect that the various backbone and side chain spin models that we have introduced , are all in the same dynamical universality class .
figure [ fig-14 ] shows the residue - wise accumulated distribution of all the individual angles @xmath59 in ( [ etai ] ) _ i.e. _ the ensuing landscape of the individual @xmath59 during the entire 80 ns gromos53a6 simulation .
the conclusions that can be deduced from this figure are in line with those in figure [ fig-8 ] .
in particular , we observe the presence of the n - terminal soliton , how it is centered around residues 5 - 6 .
we also observe that the residues between sites 6 - 27 are in a helical position during the entire time evolution .
we also observe the merging of the c - terminal soliton with the fluctuations of the c - terminal .
figure [ fig-15 ] shows the time resolved landscape of all the @xmath59 angles .
there is a remarkable similarity between this figure , and the figure obtained from the do.dssp backbone analysis shown in figure [ fig-8 ] . in particular , the formation and stabilisation of the n - terminal soliton around sites 5 - 6 is clearly visible .
the appearance of the c - terminal soliton and its subsequent evolution is similarly visible : we observe how this soliton is formed at around frame 200 , and then propagates towards the c - terminal in a stepwise manner , crossing over the various peierls - nabarro barriers and eventually merging with the c - terminal thermal fluctuations .
finally we note the apparent similarities between the structure of the landscape in figure [ fig-15 ] and the behaviour of the folding index density in figure [ fig-10 ] .
we proceed to analyse the detailed properties of soliton structure and formation
. our main focus will be on the n - terminal soliton structure .
we are particularly interested in the phenomena that take place when the soliton is formed , _
i.e. _ the vicinity of the frame @xmath1711000 , and when the soliton moves over to a peierls - nabarro barrier and stabilises , _
i.e. _ the vicinity of the frame @xmath1712500 .
we start with a closeup of the c - terminal part in figure [ fig-15 ] , shown in figure [ fig-16 ] .
we observe how , in terms of the side chain @xmath59 angles , the soliton structure which forms with center at residue 28 subsequently propagates back - and forth , in a step - wise manner , towards the c - terminal .
eventually it merges with the terminal , and dissolves into its fluctuations .
the soliton motion is fully in line with figure [ fig-8 ] , and the initial motion is consistent with the folding index analysis in figure [ fig-9](b ) . in particular , the step - wise propagation of the soliton is fully consistent and in line with the presence of peierls - nabarro barriers .
these barriers are high enough to trap the soliton momentarily , but low enough for the soliton to eventually cross over them when its thermal excitation energy fluctuates to high enough value .
we remind that the present simulations have been performed at 290 k. we have chosen this relatively low temperature value , from the _ in vivo _ perspective , in order to restrain the soliton mobility and to dampen noisy thermal fluctuations .
figure [ fig-17 ] shows a close - up of the n - terminal part in figure [ fig-15 ] , from two complementary perspectives .
the soliton appears in the vicinity of frame 1000 .
it subsequently translates one residue towards the n - terminal , in the vicinity of frame 2500 .
this is an apparent crossing over a peierls - nabarro barrier by thermal fluctuation , and it is followed by a stabilisation of the soliton at the final position . note the correlation between the soliton motion and the extent of the n - terminal fluctuations .
figure [ fig-18 ] shows the evolution of the individual @xmath59 angles in ( [ eq : etaprime ] ) , in the case of the n - terminal soliton structure .
the panels display the deviation of @xmath59 from the initial average value for residues @xmath172 , which we have found to be those of primary interest . for @xmath173 and for @xmath174 and larger , the deviations from the initial average value fluctuate around zero .
in figure [ fig-19 ] we scrutinise those segments of figure [ fig-18 ] , where the @xmath175 branch of the angle needs to be carefully resolved .
in each of the panels in figure [ fig-19 ] we display the data in figure [ fig-18 ] , over a subset of frames and now evaluated around the value @xmath176 : in figure [ fig-19](a ) we show the segment 1500 - 2500 of figure [ fig-18](a ) , in figure [ fig-19](b ) we show the segment 2000 - 3000 of figure [ fig-18](b ) and in figure [ fig-19](c ) we show the segment 500 - 1500 of figure [ fig-18](c ) . by combining figures [ fig-18 ] and [ fig-19 ]
we conclude that there are two major transitions , around frames 1000 and 2500 respectively .
these transitions are concurrent with the major transitions in figures [ fig-8 ] , [ fig-15 ] , [ fig-17 ] and in particular figure [ fig-9 ] .
the first transition corresponds to the creation of the n - terminal soliton , and the second one to its translation , by one site towards the n - terminal , and subsequent stabilisation .
we also observe the presence of an extended transition process , visible in figure [ fig-18](b ) between frames 1100 - 1500 . in summary , we conclude from these figures that : @xmath80 in the vicinity of the frame 1000 , when the n - terminal soliton structure forms , there is an initial twisting of the @xmath177 dihedral which is close to @xmath178 and a twisting of the @xmath179 dihedral by an approximatively equal amount but in the opposite direction .
thus , at this point the total twisting which is produced along the side chain segment vanishes .
we note that the presence of two twists by an equal amount but opposite in direction , is the hallmark of a bloch domain wall pair production .
but we recall that the backbone folding index detects only a single soliton , as shown in figure [ fig-9](a ) .
@xmath80 after the initial bloch wall pair formation , the @xmath177 dihedral becomes slowly twisted back to the original value between frames 1200 - 1600 , so that after frame 1600 we are left with a total of @xmath180 twist .
this process leaves us with a single soliton along the chain segment , with a total twisting around @xmath181 .
@xmath80 finally , in the vicinity of frames 2100 - 2200 , we observe a rapid twisting of the @xmath182 dihedral by an amount close to @xmath183 .
this is followed , in the vicinity of the frame 2400 - 2500 , by a twisting of the @xmath177 dihedral by an approximatively equal amount .
there is an accompanying twist of the @xmath179 dihedral by an amount somewhat less than @xmath184 , over the same frames .
these two twistings at @xmath185 accompany the peierls - nabarro barrier crossing , as can be deduced by comparison with figures [ fig-15 ] and [ fig-17 ] .
@xmath80 when the soliton stabilises , after frame 3000 , we conclude that there is a total twisting in the side chain structure which is close to @xmath186 , and carried by the final soliton configuration .
thus we assign to the final soliton the value @xmath187 of the index ( [ eq : etaprime2 ] ) .
note that there is also certain ( small ) spillage of the @xmath188-values which is distributed among the nearby residues . in summary
, the side chain analysis shows that at the level of the xy spin chain analysis , the n - terminal soliton structure forms by an initial rapid formation of a bloch domain wall _
i.e. _ soliton - antisoliton pair , followed by a slow twisting that apparently removes one of the two bloch wall solitons .
this is then followed by a rapid transition , in combination with a peierls - nabarro barrier crossing , that forms the final stable soliton structure .
accordingly , we may characterise the final soliton as a configuration with the c@xmath0 backbone folding index @xmath169 and the side chain xy folding index @xmath187 .
figure [ fig-13 ] revels the presence of strong correlations between backbone and side chain dynamics .
in particular , any formation of side chain soliton structure should correlate with the formation of corresponding backbone soliton .
accordingly , we proceed to construct explicitly the backbone soliton that accompanies the n - terminal side chain soliton .
we shall find that the backbone soliton can be modeled by a solution of the discrete nonlinear schrdinger equation ( [ tauk ] ) , ( [ nlse ] ) , with very high sub - atomic precision .
as an example , we consider the profile of the n - terminal soliton structure at two different frames . other frames , and the c - terminal soliton structure , can be analysed similarly .
we select the sites 4 - 11 ( pdb sites 632 - 639 ) for our analysis : the sites 0 - 3 are subject to fluctuations , and beyond the site 11 there is only a monotonic @xmath0-helix .
we start with the frame 2000 , which is located in the regime where the side chain structure of the soliton has stabilised according to figure [ fig-18](b ) and the backbone folding index of the segment shown in ffigure [ fig-9](a ) has the value @xmath163 ; the side chain index over this segment is @xmath189 , according to figures [ fig-18 ] , [ fig-19 ] . in figure [ fig-20 ]
we show the profile of the bond angle and the torsion angle over the segment 4 - 11 ( sites 632 - 639 in the pdb file ) in the frame 2000 , after we have implemented the @xmath22 gauge transformation ( [ eq : dsgau ] ) to identify the soliton profile .
note that in order to compute a single bond angle , we need to know three residues while the evaluation of a torsion angle consumes four residues .
thus , despite the smaller number of data points in the figure , the ensuing configuration engages 8 residues .
we observe that the bond angle has the profile of a single domain wall soliton of the dnls equation , approximated by ( [ kappaprof ] ) .
we use the software package propro that has been described at @xmath190 to numerically construct the ensuing soliton solution of the dnls equation . in figure [ fig-21 ] ( a ) we compare the profile of the bond and torsion angles in figure [ fig-20 ] with the profile of the soliton solution of ( [ tauk ] ) , ( [ nlse ] ) . in figure [ fig-22 ] ( top )
we compare the residue - wise distance between the all - atom configuration of frame 2000 , and the dnls solution .
the average c@xmath0 rmsd between the two configurations is less than 0.1 ngstrm , and at no residue is the distance between the c@xmath0-atoms more than 0.2 ngstrm : the difference is truly negligible .
the grey strip around the dnls soliton is a 0.2 ngstrm ( quantum mechanical ) fluctuation band @xcite .
the figure [ fig-22 ] ( bottom ) shows the 3d overlay of the ensuing c@xmath0 backbones , for all practical purposes they are the same . in figure [ fig-23 ]
we show the loop trajectories of the all - atom configuration of frame 2000 , and the corresponding dnls soliton , on the stereographically projected two - sphere of figure [ fig-2 ] and [ fig-3 ] .
note that for both of these two loop trajectories the folding index , as defined in ( [ eq : gamma ] ) is vanishing , in that the trajectory does not encircle the north - pole ( center of the disk ) . a very short change either in the position of the residue labeled @xmath125 or in the position of the residue labeled c in figure [ fig-23 ] a ) , can shift the trajectory
so that the line connecting them moves over to the other side of the north pole and the folding index becomes @xmath191 .
this is consistent with the result shown in figure [ fig-9 ] that the folding index fluctuates between the values @xmath162 and @xmath191 , around the frame 2000 , with the posture shown in figure [ fig-23 ] being the more stable one . in figure [ fig-24 ]
we show a close - up to the frame segment 2475 - 2525 around the soliton 2000 , in terms of the side chain angles @xmath188 .
the close - up reveals the presence of fluctuations between the soliton and the n - terminal , while the helix between the soliton and the c - terminal displays very small fluctuations .
thus , the fluctuations in the folding index around the frame 2000 are most likely due to shifts in the position of the residue labeled b in figure [ fig-23 ] ( a )
. figures [ fig-25]-[fig-28 ] show the same analyses for the conformation in frame 3500 .
we find that the dnls soliton describes the domain wall soliton that we have constructed by all - atom simulations , with a very high sub - atomic precision .
we note that the soliton in frame 3500 is a configuration that connects between the @xmath44-stranded region of figure [ fig-2 ] to the @xmath0-helical region , while in the case of the soliton at frame 2500 the initial residue is located in a sparsely populated region of the landscape in figure [ fig-2 ] .
the major qualitative difference between frames 2000 and 3500 is between figures [ fig-23 ] and [ fig-28 ] .
the soliton in frame 3500 is relatively stable . in particular , as shown in figure [ fig-9 ] , its folding index @xmath163 .
we can understand the stability of the folding index by comparing figure [ fig-23 ] ( a ) with figure [ fig-28 ] ( a ) . in the later , the residues have assumed positions where the connecting arrows are stabilised against small perturbations , they are protected from crossing over the north pole in a manner that causes fluctuations in the value of the folding index . finally , we compare the structure of the solitons at frames 2500 and 3500 . from figures
[ fig-20 ] , [ fig-25 ] we observe that in terms of the bond angles the soliton 3500 has indeed moved one site towards the n - terminal , from the position of soliton 2500 . in figure [ fig-29 ]
we overlay their 3d structures . for this
, we first translate the soliton in frame 3500 one residue away from the n - terminal , so that the two have the same location along the backbone .
the figure shows the ensuing 3d interlaced c@xmath0 backbones , in a relative position where the rmsd is minimal .
there is a visible difference , and the minimal rmsd is 2.0 ; the soliton has clearly become deformed .
molecular dynamics enables the scrutiny of protein folding , at the level of individual atoms and over very short time intervals .
however , it can leave us with the conceptual challenge to understand , how the individual atoms cooperate to produce the kind of large scale organisation that appears to be prevalent among crystallographic protein structures .
we have performed detailed molecular dynamics simulations , with the aim to find out how organised structure emerges when a protein folds .
we have first compared three different force fields using the gromacs 4.6.3 .
package , to select the proper tools .
we have chosen a c - chain subunit from hiv envelope glycoprotein with pdb code 1aik as a concrete example , partly due to its biomedical relevance even though this is an issue which has not been addressed by us .
we have introduced and further developed various tools of modern theoretical physics , to systematise and analyse the data .
these include topological tools , conceptual analogies drawn from the notion of spin chains , the notion of wilsonian universality , and methods based on the analytical structure of the discrete nonlinear schrdinger equation . in this manner
we have arrived at the conclusion that the protein folding is a process that relates intimately to the emergence and interactions of solitons . in particular , a configuration such as the bloch domain wall along a spin chain appears to be most useful in comprehending how structure emerges and self - organises when a protein folds . we have inspected both the static and dynamic properties of domain wall solitons and observed that concepts which are familiar from the study of lattice systems , such as the peierls - nabarro barrier , also appear along protein backbones lattices , and in fact assume a central role in dictating how the folding proceeds . we hope that our observations help to pave a way for the powerful analytical and topological tools and techniques that have been introduced and developed in the context of integrable spin chains and related solvable models , to become part of the arsenal used describe emergence of structure and organisation in the case of proteins and other biological macromolecules .
this research was supported in part by bulgarian science fund ( grant dnts - ch 01/9/2014 ) and china - bulgaria intergovernmental s&t cooperation project at ministry of science and technology of p.r .
china ( 2014 - 3 ) .
ajn acknowledges support from vetenskapsrdet , carl trygger s stiftelse fr vetenskaplig forskning and qian ren grant at bit , p.r .
china . as was supported by national science center poland ( maestro umo-2012/06/a / st4/00376 ) | we inquire how structure emerges during the process of protein folding . for this
we scrutinise collective many - atom motions during all - atom molecular dynamics simulations .
we introduce , develop and employ various topological techniques , in combination with analytic tools that we deduce from the concept of integrable models and structure of discrete nonlinear schrdinger equation .
the example we consider is an @xmath0-helical subunit of the hiv envelope glycoprotein gp41 .
the helical structure is stable when the subunit is part of the biological oligomer .
but in isolation the helix becomes unstable , and the monomer starts deforming .
we follow the process computationally .
we interpret the evolving structure both in terms of a backbone based heisenberg spin chain and in terms of a side chain based xy spin chain .
we find that in both cases the formation of protein super - secondary structure is akin the formation of a topological bloch domain wall along a spin chain . during the process
we identify three individual bloch walls and we show that each of them can be modelled with a very high precision in terms of a soliton solution to a discrete nonlinear schrdinger equation . |
experimental demonstration of the phenomenon of negative index of refraction first in the microwave @xcite and latter in the optical regime @xcite has stimulated growing interest in nonlinear properties of negative index materials @xcite .
this interest is motivated by specifics on the interaction of electromagnetic waves with negative index materials . in combination with a nonlinear response of the optical material to electromagnetic radiation
, this iteration leads to a new nonlinear optical phenomena .
study of these phenomena is of considerable importance both for better understanding of fundamentals of electrodynamics of negative index materials and their applications .
one of the most fundamental property of negative index material is an opposite directionality of the poynting vector , characterizing the energy flux , to the wave vector @xmath0 . on the other hand
, the negative index property can be realized only on particukar wavelength intervals .
these two features are offering a very unusual type of multi - wave interactions , if frequencies of interacting waves correspond to frequency intervals where optical material has different signs of refractive index .
multi - wave interaction must satisfy a phase matching condition , which is possible only when all wave vectors are pointed in the same direction @xcite .
therefore energy fluxes of the waves with frequencies corresponding to a negative sign of refractive index will propagate in opposite direction to those with frequencies corresponding to a positive sign of refraction index .
such effect was suggested for the first time in @xcite , which considered the particular case of three waves interaction - second harmonic generation .
a solution of the equation describing second harmonic generation in the case of exact phase matching was given in @xcite .
the feasibility of parametric amplification using three - wave interaction for compensation losses in negative index materials was studied in @xcite .
the dynamics of interacting wave packets propagating in negative index materials in the case of second harmonic generation was considered in @xcite .
it was shown that in contrast to a weak intensity of pump field , at high intensities a second harmonic pulse can be trapped by the pump pulse and forced to propagate in the same direction . in this paper
we investigate second harmonic generation in the presence of phase - mismatch @xmath1 .
this is an important case since phase - mismatch is more relevant to realistic experimental conditions .
additionally , it introduces two types of spatial distribution of second harmonic field intensity along the sample : monotonic and periodic 1 on the coordinate .
both cases are considered in this paper .
we also studied second harmonic generation near a critical phase - mismatch value , when the material becomes transparent for the pump wave .
the system of equations describing three wave interactions ( one dimensional case ) in a @xmath2 - medium for the slowly varying envelope and phase approximation can be written in the following form @xcite : @xmath3 where wave numbers @xmath4 are defined as follows @xmath5 and @xmath6 is the sign of the square root of @xmath7 and @xmath8 for the case of second harmonic generation eqs .
take the following form : @xmath9 where @xmath10 , @xmath11 is the fundamental wave with frequency @xmath12 , and a@xmath13 is the second harmonic generated in the material .
we consider the case , when the refractive index is negative at the fundamental frequency @xmath12 and is positive at the second - harmonic frequency @xmath14 .
the parameter @xmath15 plays an important role for the spatial distribution of the electromagnetic field along the sample . in the next section
we consider both cases @xmath16 and @xmath17 .
we consider second harmonic generation for continuos waves in a @xmath2 medium under ideal phase matching conditions @xmath16 . the length of the sample we assume to be @xmath18 .
using the symmetry properties of the susceptibility tensor @xmath2 with respect to permutations of @xmath12 and @xmath14 frequencies , the mathematical model of second harmonic generation can be formulated in the following way @xcite : @xmath19 where @xmath20 .
let us represent the complex functions @xmath11 and @xmath13 in terms of amplitudes @xmath21 and phases @xmath22 @xmath23 substitution of eqs . into and , and separation of real and imaginary parts lead to the following system of equations : @xmath24 with boundary conditions : @xmath25 here @xmath26 and
@xmath27 are defined as follows @xmath28 from the first two equations an integral of motion follows : @xmath29 this integral of motion corresponds to the modified manley - row relation . in case of second harmonic generation in conventional materials
the manley - row relation is equivalent to conservation of energy ( @xmath30 ) . in our case
, relation corresponds to conservation of total flux of the energy .
the second integral of motion for the system reads as : @xmath31 the integral of motion is consistent with boundary conditions only if @xmath32 .
taking into account that the pump wave energy decays in @xmath33 , we conclude that the phase difference is equal to @xmath34 , therefore the system of equations can be represented as follows : @xmath35 the solution of has the following form @xmath36 here @xmath37 and @xmath38 .
the solutions unknown parameter @xmath39 , is the value of the fundamental field at the end of the sample .
this parameter can be found from the boundary condition .
taking into account the manley - row relation , it leads to the transcendental equation for @xmath40 : @xmath41 this equation can be solved numerically .
the solution of together with determines the field distribution along the sample .
the dependence of intensities @xmath42 and @xmath43 on @xmath44 is represented in fig.[figevsz ] were the intensity boundary value @xmath45 is chosen to be @xmath46 , here @xmath47 and @xmath48 . the solution of transcendental equation for @xmath49 is shown in fig.[figmvse ] .
this plot illustrates the dependence of the output field intensity @xmath50 , as a function of @xmath51 ( the amplitude of the fundamental field pumped into the medium ) . as shown in fig.[figmvse ] , the formal solution of equation has multiple branches . however , only the lower branch presented by a solid curve has physical meaning .
upper brunches represented by dashed curves are originated from periodicity of the @xmath52 function in .
both @xmath53 and @xmath53 corresponding to these branches have singularities on the interval @xmath54 which is inconsistent with conservation of energy .
note that the lower
physical " branch shows saturation of output power of the electric field at the fundamental frequency @xmath55 with increase of input power @xmath56 .
this indicates that with the increase of input power @xmath56 above @xmath57 , all excessive energy of pump signal converts to energy of the second harmonic signal ( see fig .
[ fig : second : harm : convers ] ) .
let us consider the impact of phase mismatch @xmath15 ( eq . ) on second harmonic generation .
the system of equations describing the spatial distribution of field amplitudes @xmath59 and phase difference @xmath60 in the presence of phase mismatch reads : @xmath61 here @xmath62 and @xmath27 is defined in . by introducing variables @xmath37 and @xmath38 eqs .
can be represented in the following form : @xmath63 here @xmath64 .
the manley - row relation in this case remains unchanged : @xmath65 and a second integral of motion in presence of phase mismatch reads : @xmath66 taking into account boundary condition @xmath67 , we conclude that @xmath68 and therefore @xmath69 the function has an extremum at @xmath70 and @xmath71 at this value of @xmath72 gives @xmath73 .
since @xmath74 , then there is the critical value of mismatch @xmath75 such that @xmath76 .
notice that is defined for arbitrary values of @xmath77 if @xmath78 . if @xmath79 then there is a forbidden gap for values of @xmath77 : @xmath80 in this case @xmath81 if @xmath82 since the value of @xmath77 on the right side of the sample is ser to be zero ( @xmath67 ) , the branch of @xmath77 values is not accessible .
values of @xmath77 in this case remain within the branch . in this case the conversion efficiency of the pump wave to second harmonic is limited by the value @xmath83 .
the dependence of @xmath84 on @xmath72 for different values of mismatch is shown in fig .
[ figthetavse2 ] .
the bold curve on this figure corresponds to a critical value of the mismatch .
the forbidden gap for @xmath72 can be seen for two lowest curves , when curves are below @xmath85 .
the presence of a forbidden gap for @xmath77 suggests the existence of two types of solutions for @xmath77 .
the first type corresponds to mismatch values @xmath86 and in this case @xmath77 is not bounded from above .
this means that the conversion rate of fundamental harmonic to second harmonic , in principle , can be high ( close to @xmath87 - ideal conversion , similar to fig .
[ fig : second : harm : convers ] in the previous section ) .
the second type of solutions correspond to mismatch values @xmath88 ; in this case the amplitude @xmath77 is bounded from above @xmath89 .
this means the is a limitation of the output intensity of second harmonic field at the growing input intensity of fundamental harmonic . for further considerations it is more convenient to deal with field intensities rather then with amplitudes .
using expression for @xmath71 , the second equation of ( [ shgdelat2 ] ) can be represented as an equation for the intensity @xmath90 : @xmath91 where @xmath92 is a quartic polynomial @xmath93 with the following roots : @xmath94 notice that these roots define the forbidden gap @xmath95 $ ] for values of @xmath77 ( see fig . [ figthetavse2 ] and equations , ) . based on this qualitative analysis ,
we conclude that there are three regimes of second harmonic generation controlled by the absolute value of the phase mismatch . in the following subsection
we will analyse solutions describing spatial field distribution inside the sample .
the absolute value of the phase mismatch determine three different regimes of second harmonic generation : @xmath96 , @xmath97 and @xmath98 .
first we consider the case of subcritical mismatch : @xmath99 . in the case where @xmath100 , roots are complex - valued and the solution of can be expressed in terms of weierstrass function @xmath101 @xcite . by expanding @xmath102 into taylor
series and introducing a new variable : @xmath103 here derivatives of the polynomial are taken with respect to @xmath104 , then the solution of equation ( [ p2:delta ] ) can be represented in an implicit form : @xmath105 where @xmath106 and @xmath107 are invariants of weierstrass function .
@xmath108 finally , the amplitudes of second and fundamental harmonics have the following form : @xmath109 the parameters @xmath106 and @xmath107 are functions of @xmath1 and @xmath40 . to determine solutions of eqs .
( [ shgdelat2 ] ) we need to solve for the unknown value of the output pump wave @xmath39 .
the value of @xmath39 can be found taking into account the output boundary condition @xmath110 and the manley - row relation ( [ manley : row ] ) , which lead to the following transcendental equation for @xmath39 : @xmath111 to determine the unknown parameter @xmath112 , eq .
needs to be solved numerically .
the analysis of the @xmath55 dependence on @xmath51 and @xmath1 shows that with increasing phase mismatch from @xmath113 to @xmath114 , all branches ( physical and nonphysical ) shift upwards and nonphysical branches change their shapes .
[ fig : contour:3d ] shows the dependence of the output field amplitude @xmath55 at fundamental frequency on @xmath51 near the critical value of phase mismatch @xmath114 .
sheet , labeled as 1 " , corresponds to the physical branch , while sheet labeled as 2 " represents the first nonphysical branch .
other nonphysical sheets are located above nonphysical sheet 2 " shown on fig .
[ fig : contour:3d ] . the spatial distribution of @xmath53 and @xmath115 can be found by substitution of the solution of the eq .
( @xmath39 ) in eqs . and .
we found that all solutions @xmath53 and @xmath115 are monotonically decreasing in @xmath44 .
an example of @xmath115 at @xmath116 is shown in fig .
[ fige2vsd ] ) .
the conversion efficiency @xmath117 as a function of the input amplitude @xmath51 is presented in fig . [ fig : conver : eff : delta ] .
as one can observe , @xmath118 is approaching its asymptotic value @xmath119 in slower fashion for larger values of @xmath120 . when the value of phase mismatch is critical @xmath121 , the roots @xmath122 and the discriminant of the weierstrass function is zero . in this case
, the function @xmath123 can be represented in terms of hyperbolic functions .
thus the eqs . and
take the form : @xmath124 and the transcendental equation for @xmath40 reads as @xmath125 the numerical solution of eq .
is shown in fig.[fig : contour:3d ] ( line 3 " ) .
observe that at large values of @xmath51 , the solution of eq .
is proportional to @xmath51 ( @xmath126 ) .
therefore , at large values of @xmath51 the conversion efficiency @xmath127 is always less then one while in the subcritical regime @xmath128 ( see fig .
[ fig : conver : eff : delta ] ) . at large mismatch values , when @xmath129 , roots of are real .
in this case it is convenient to represent @xmath123 in terms of jacobi elliptic @xmath130 function @xcite .
eqs and can be represented as @xmath131 , \label{e2:above : solution}\end{aligned}\ ] ] here @xmath132 , and the equation for @xmath40 takes the following form : @xmath133 \label{eq : trans : overcrit}\end{aligned}\ ] ] the sheet corresponding to solutions of is labeled in fig .
[ fig : contour:3d ] as 4 " .
in contrast to the subcritical regime , all solutions in this case are represented by a single sheet .
this sheet has folds .
hence the intersection of this sheet with plane corresponding to @xmath134 gives multivalued dependance of @xmath55 on @xmath51 .
this dependance for two different values of @xmath1 is shown in fig .
[ fig : bistab ] . in the supercritical regime the second harmonic field experiences spatial periodic oscillations with period @xmath135 ( see fig . [ fige2vsd ] ) .
the distance between neighboring zeros @xmath136 of the amplitude of second harmonic is determined by the following formula : @xmath137 if the slab length satisfies the condition @xmath138 ( @xmath139 ) , then the amplitude of the second harmonic wave is zero at the both ends of the slab ( zero conversion efficiency ) .
therefore such slab is transparent for a pump wave .
a plot of the transmission coefficient @xmath140 as function of @xmath51 is shown in fig.[fig : trans : coeff ] .
the transmission coefficient is equal to @xmath87 at the points labeled as 1 " , 2 " , @xmath141 ( transmission resonances ) .
the spatial distribution of the fundamental and second harmonic fields corresponding to the transmission resonance at the point 1 " ( see fig.[fig : trans : coeff ] ) is shown in fig.[fig : point:1 ] .
we considered second harmonic generation in negative index materials .
specifics of this process is in the negative value of refractive index for the pump wave and the positive value for second harmonic .
this led to important features which are different from the case of second harmonic generation in conventional dielectrics .
the main difference is in the existence of nonzero critical values of the phase mismatch .
if the absolute value of phase mismatch is below critical , then the field intensities are monotonically decaying along the sample . when the absolute value of phase mismatch exceeds a critical value , monotonic decay of intensities transforms to a spatial periodic oscillations .
note , that in the conventional case the critical value of phase mismatch is zero .
another important feature is the dependance of conversion efficiency on the amplitude of the incident pump wave .
when the absolute value of phase mismatch is below critical value , then the conversion efficiency asymptotically approaches 100% at large values of the incident pump wave amplitude .
it should be stressed that in this case the asymptotic value of conversion efficiency does not depend on the phase mismatch value .
the phase mismatch affects only the rate of approaching of conversion efficiency to its asymptotic value
. when the phase mismatch is exactly equal to critical value , then the asymptotic value of conversion efficiency experiences a jump to a value which is less than 100% .
when the absolute value of phase mismatch is above critical value , the conversion efficiency becomes an oscillatory function of the incident pump wave amplitude .
finally , we found that the dependance of output amplitude of the pump wave on its input amplitude is single valued if the absolute value of phase mismatch is below critical and becomes multi - valued in the opposite case .
we would like to thank a. k. popov and v. m. shalaev for valuable discussions and a. aceves for help during preparation of this paper . a.i.m and zh.k .
appreciate support and hospitality of the university of arizona department of mathematics during the preparation on this manuscript .
this work was partially supported by nsf ( grant dms-0509589 ) , aro - muri award 50342-ph - mur and state of arizona ( proposition 301 ) , rfbr ( grant no .
09 - 02 - 00701-a ) and the federal goal - oriented program scientific and scientific - educational personnel of innovational russia " . | second harmonic generation in negative index metamaterials is considered . theoretical analysis of the corresponding model demonstrated significant difference of this phenomenon in conventional and negative index materials .
in contrast to conceptional materials there is nonzero critical phase mismatch .
the behavior of interacting weaves is dramatically different when phase mismatch is smaller or greater than critical value . |
individual chain polymers in good solvents are typically modelled by trajectories of random walks , or - in the continuum limit - by brownian paths .
such models by themselves however do not take into account that self - crossings of these paths should be suppressed , this `` the excluded volume '' effect will make the trajectories less curly and more extended .
fractional brownian paths have been suggested as a heuristic model for such swelling , or on the other hand for polymers in a collapsed state @xcite , but a more proper model would be based on self - avoiding random walks , or on a weight factor which penalizes self - crossings , such as in the continuum edwards @xcite @xcite @xcite @xcite west3 @xcite or the discrete domb - joyce @xcite model .
the ensuing swelling of the molecular conformations is given by the flory index @xcite @xcite which describes the scaling of the end - to - end distance as a function the number of monomers .
it has been extensively studied both in the ( chemical ) physics and the mathematics community .
the physics literature is characterized by structural intuition and far - reaching predictions , the mathematical results are less far - reaching but provide the high reliability characteristic of the mathematical approach .
both are too vast to be quoted here , we refer for this to recent reviews @xcite @xcite . in the present paper , after a few words on fractional brownian motion fbm , we shall see that one can extend to fbm the edwards model of brownian paths with exponentially suppressed self - intersections , @xmath0a mathematical existence proof has been established recently @xcite . in the third part of the paper
we generalize some by now classical arguments from the physics literature to explore what the flory index might be in the fbm case .
fractional brownian motion on @xmath1 , @xmath2 , with `` hurst parameter '' @xmath3 is a @xmath4-dimensional centered gaussian process @xmath5 with covariance function @xmath6 for @xmath7 it is ordinary @xmath4-dimensional brownian motion @xmath8 .
we refer to the recent monographs by biagini et al .
@xcite and by y. mishura mishura ; for self - intersection local times of fbm see hu and nualart hn .
self - repelling brownian paths for a time interval @xmath9 can be modelled via a `` gibbs factor '' to suppress self - intersections:@xmath10technically one defines this expression as a limit , using@xmath11if this quantity is well defined ; otherwise a renormalization is required , as , more generally , in theorem 2.2 below . recently ,
generalizing an argument of varadhan @xcite , this was extended in @xcite to @xmath12as follows .
the edwards model is well defined for all @xmath13 , with @xmath14 for @xmath15 and @xmath16 sufficiently small@xmath17with @xmath18is well - defined .
when the number @xmath19 of monomers of a polymer becomes large one expects its end - to - end length @xmath20 to scale @xcite@xmath21for ( fractional ) brownian motion the root - mean - square length @xmath22 is scaling with@xmath23 but the excluded volume effect makes the paths and polymers swell : the end - to - end length increases . for the brownian motion case
there is the famous flory formula @xmath24based originally on a mean field argument . since its proposal by flory fisher @xcite ,
numerous methods were invoked to put it on a more solid mathematical basis , a process which has up to now been fully successful in the case @xmath25 @xcite . to obtain what may be considered as a first guess of a similar formula for fbm we shall return to the modest beginnings , generalizing fisher s original argument @xcite @xcite ( see e.g. the review given in mckenzie @xcite ) to the case at hand . a partition function @xmath26 for a freely jointed chain of @xmath19 segments for which the end - to - end length has fixed modulus @xmath20
is given by @xmath27 and leads to a free energy@xmath28instead of such a chain a continuous model is that of a brownian trajectory from time zero to time @xmath19 , for which one computes@xmath29for the fbm case this formula generalizes to@xmath30from which we see that @xmath31 and hence we should [email protected].@xmath33for the repulsive excluded volume energy of fbm paths @xmath34 with @xmath35@xmath36dimensional considerations and mean field arguments @xcite suggest@xmath37maximizing@xmath38with regard to @xmath20 leads to@xmath39assuming that the 2nd term is negligible one [email protected].@xmath41with@xmath42 a polymer model with @xmath43 instead of @xmath44 would produce the same expression as in ( [ f ] ) , hence also the same flory index , but would not share the homogeneity implied by the stationary increments of @xmath45 the derivation of @xmath46 is evidently heuristic and needs validation . for this
it is worth noting that for brownian motion there is a critical dimension @xmath47 defined by the fact that for @xmath48 there is no excluded volume effect , so that @xmath20 scales like the unperturbed brownian motion : @xmath49i.e .
@xmath50we can ask for which dimension , more generally , the fbm flory index will show no excluded volume effect from self - crossings , i.e. @xmath51inserting our ansatz ( [ ( 1 ) ] ) one finds @xmath52and indeed it is known ( theorem 1.1 of talagrand @xcite ) that @xmath4-dimensional fbm has no double points iff@xmath53 in other words , our @xmath46 predicts @xmath54 correctly . as a consequence
, any flory formula should be considered only up to the critical dimension , i.e. as long as there are double points and an excluded volume effect .
similarly , any prediction of @xmath55 would be unphysical : the end - to - end distance can not grow faster than the number n of monomers . in the case at hand
this suggests for the one - dimensional case@xmath56note that for small @xmath57 the scaling exponent as predicted would be strictly less than one while for the brownian motion case @xmath58has been proven @xcite @xcite .
the infimum@xmath59happens to be the scaling exponent of the myopic random walk @xcite . in the attached figure 1 the two red lines correspond to @xmath60 and to the critical dimension as a function of the hurst index @xmath57 , respectively .
above these the flory index is unphysical . on the green lines @xmath46 is validated .
( the existence proof of the fbm edwards model in @xcite works below the dashed line . ) for fixed dimension @xmath4 , any extension @xmath61 of the flory formula to general hurst indices @xmath57 will have to obey@xmath62for the usual brownian motion ( flory - fisher ) , and @xmath63at the critical point ( talagrand ) .
note that our ansatz ( [ ( 1 ) ] ) @xmath64is just the unique linear interpolation between those two values . for @xmath7 kosmas and freed @xcite derive a recursion formula @xmath65(here and in the following we label formulas from - or analogous to those in - the paper @xcite by their numbers in that article , in double brackets .
) the derivation of this formula is specific to the brownian motion case and does not hold for general @xmath46 .
hence in what follows we shall generalize their arguments which led to ( [ ( 4.13 ) ] ) to first obtain a valid recursion formula and then check whether it is satisfied by @xmath66 as given in ( [ ( 1 ) ] ) .
we begin by considering@xmath67from the defining relation@xmath68we see that for @xmath69 the processes @xmath70 and @xmath71 obey the same law . making this substitution and a change of integration variables @xmath72
@xmath73 we obtain@xmath74 likewise we find for the mean - square end - to - end distance @xmath75 @xmath76 ( note the critical dimension @xmath77where @xmath78 . ) for large @xmath19 one expects a power law behavior for the unknown function @xmath79 , i.e.@xmath80with an exponent @xmath34 to be determined . as a next step
we restrict one coordinate of the positions , @xmath81 to the interval @xmath82 $ ] by inserting @xmath83}(b_{i}^{h})=\left\ { \begin{array}{cc } 1 & \text{if } b_{i}^{h}(t)\in \lbrack 0,d]\text { for all } t \\ & \\ 0 & \text { otherwise}% \end{array}% \right.\]]into ( [ ( 2 ) ] ) .
one obtains@xmath84}(b_{i}^{h})\left ( b^{h}(n)\right ) ^{2}\exp \left ( -g\int_{0}^{n}ds\int_{0}^{n}dt\delta \left ( b^{h}(s)-b^{h}(t)\right ) \right ) \right ) \notag \\ = \frac{a^{-2h}}{z_{d}\left (
g , n\right ) } \mathbb{e}\left ( 1_{[0,a^{h}d]}(b_{i}^{h})\left ( b^{h}(an)\right ) ^{2}\exp \left ( -a^{hd-2}% \underset{}{g\int_{0}^{an}}d\sigma \int_{0}^{an}d\tau \delta \left ( b^{h}(\sigma ) -b^{h}(\tau ) \right )
\right ) \right ) \notag \\ = a^{-2h}f(a^{h}d,\ a^{hd-2}g , an)=n^{2h}f(n^{-h}d,\ n^{2-hd}g,1 ) .
\tag{(4.3 ) } \label{(4.3)}\end{gathered}\ ] ] now assume that asymptotically there is a dimensionless correction factor @xmath85 for @xmath86it should grow as @xmath87 becomes small which suggests a power law behavior for the function @xmath85 : @xmath88with @xmath89 to be determined . as @xmath87
approaches a minimal value @xmath90 - approximately the extension of a monomer ( `` kuhn length '' ) - the polymer becomes effectively @xmath91-dimensional : @xmath92 this provides a relation between the end - to - end length for dimensions @xmath4 and @xmath93 . to obtain from this a recursion relation , recall equation ( ( 2.22)):@xmath94and introduce instead of @xmath34 the ( unknown)@xmath95so that @xmath96and@xmath0@xmath97on the other hand from ( [ ( 4.6 ) ] ) we have@xmath98comparing exponents in these two expressions we find @xmath99the first of these equations gives@xmath100with this the second one [email protected] .
this expression does not depend on the dimension @xmath4 so that all the @xmath102 are given in terms of e.g. @xmath103 @xmath104@xmath105 @xmath106 satisfies this recursion equation , with @xmath107@xmath108 the standard flory index ( [ ( 1 ) ] ) for @xmath7 obeys the recursion formula .
the recursion formula ( [ r ] ) implies the correct critical behavior , i.e. any solution will obey @xmath109 for @xmath110 , whatever the choice of @xmath111 . to see this explicitly , insert @xmath112 and find @xmath113
if @xmath111 turned out to be equal to one for all @xmath57 , the recursion formula would suggest@xmath114 an expression which then also produces the standard flory formula for @xmath115 as well as the critical dimension @xmath116
the edwards type model for self - repelling fbm now at hand will raise the question of how the end - to - end length of trajectories scales as a function of time ( or `` number of monomers '' ) . the original fisher argument , while criticized regarding its assumptions @xcite @xcite , provides a simple heuristic `` derivation '' of the flory formula which allows an extension to fbm .
the obtained scaling law needs further verification ; we note that it correctly predicts the critical dimension for which the excluded volume become negligible and obeys a recursion formula based on dimension reduction .
the latter would provide a useful constraint on any alternate scaling laws .
j. b. and l. s. would like to express their gratitude for the kind hospitality of the centro de matemtica e aplicaes fundamentais cmaf in lisbon which made our collaboration possible and pleasant .
this work is supported by ptdc / mat/100983/2008 , isfl-1 - 209 .
j. westwater : on edwards model for polymer chains , in trends and developments in the eighties , s. albeverio and p. blanchard , eds .
, bielefeld encounters in math .
phys . 4/5 ( world scientific , singapore , 1984 ) . | we present an extension of the edwards model for conformations of individual chain molecules in solvents in terms of fractional brownian motion , and discuss the excluded volume effect on the end - to - end length of such trajectories or molecules . |
the discrepancy between the number of observed dwarf satellite galaxies of the milky way and the anticipated number of dark matter halos in cosmological simulations has been heralded as the `` missing satellite problem '' @xcite and it still remains one of the greatest puzzles of the @xmath0cdm paradigm .
one class of suggested solutions to this puzzle involves modifications to the nature of dark matter ( e.g. , * ? ? ? * ) or to the initial conditions of cosmic structure ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) , while another class of solutions invokes astrophysical arguments such as inefficient cooling of cosmic gas and/or feedback from young stars ( e.g. * ? ? ?
* ; * ? ? ?
* ) . motivated by this latter class of solutions , ( * ?
* hereafter referred to as kgk04 ) developed a star formation model for the satellite galaxies , based on the mass assembly histories of individual dark matter halos in a @xmath0cdm cosmological simulation .
that model incorporated the accretion of gas in hierarchical mergers and the loss of gas caused by the extragalactic uv background , and included both a continuous mode and a starburst mode of star formation .
the model correctly reproduced the observed number of satellites of the milky way and m31 , the radial distribution of luminous satellites , and the morphological segregation of dwarf irregular ( dirr ) and dwarf spheroidal ( dsph ) galaxies .
recent @xmath1-body simulations of @xcite , with the updated wmap3 cosmological parameters , have confirmed the qualitative picture of kgk04 .
the amount of observational data for local group dwarfs has been increasing steadily in the last several years , shifting the focus from simply counting the number of dwarf galaxies to deriving their detailed star formation histories ( sfhs ) .
most local group dwarfs have now been observed with the _ hst _ , which has provided color - magnitude diagrams of resolved stellar populations in single or multiple fields @xcite .
we parametrize these data as the fractions of stars formed in the last 1 gyr , 2 gyr , 5 gyr , and 10 gyr , and present these fractions in our table [ tab : sfdata_lg ] .
the data show a great variety of star formation histories some continuous , some bursty , some truncated . in this paper
we extend the model of kgk04 to account for such varied histories of the local group dwarfs and argue that adding a stochastic threshold to the star formation law greatly improves the comparison between models and observations .
we also make predictions for the number of ultra - faint satellites of the milky way and m31 of which there have been many recent discoveries with the sloan digital sky survey ( sdss ) and the megacam survey @xcite .
our approach is complementary to running full hydrodynamic simulations of galaxy formation .
we attempt to model star formation with a phenomenological recipe , containing several but not too many free parameters .
we investigate the effects of varying each of these parameters and constrain them by comparing the model predictions for each parameter set with the observations . in this way
, we obtain empirical rules that govern star formation in dwarf galaxies , even though we may not yet fully understand their physical origin . in an alternative approach
, one uses cosmological hydrodynamic simulations with a detailed prescription for gas cooling and star formation , and thus obtains more physically sound results .
discrepancies between the simulation results and the data may , however , be difficult to interpret or computationally expensive to investigate . by combining the two approaches ,
our empirical rules will guide future developments of more sophisticated models of galaxy formation .
the paper is organized as follows .
we give a brief summary of the observational data in [ sec : obs ] .
we describe our model in [ sec : model ] and report the primary results from adding stochasticity to the star formation law in
[ sec : threshold ] . in
[ sec : other ] we describe the results of several variants of our star formation model , including late accretion of gas , steeper slopes of the schmidt law , an extended epoch of reionization , and rejection of galaxies with delayed star formation ; none of these changes significantly alters our results . in [ sec : lowmass ] we show our model projections for the new low - mass dwarfs , and in [ sec : conclusions ] we summarize our main results .
lcccrcllllr + & & & & & & & & & & + + m33 & ngc598 & sc & m31 & 203 & 9.9 & 0.093 & & & 0.52 & 8.4 + lmc & & irr & mw & 50 & 9.7 & 0.078 & 0.17 & 0.42 & 0.70 & 6.7 + smc & & irr & mw & 63 & 9.2 & 0.096 & 0.18 & 0.48 & 0.65 & 6.6 + m32 & ngc221 & de & m31 & 6 & 9.1 & 0.042 & & & 0.50 & 8.5 + ngc205 & m110 & de & m31 & 58 & 9.0 & 0.0049 & 0.0050 & 0.055 & 0.48 & 10.5 + ic10 & & dirr & m31 & 255 & 8.7 & 0.060 & 0.14 & 0.52 & 0.75 & 7.1 + ngc6822 & & dirr & mw & 500 & 8.7 & 0.087 & 0.16 & 0.57 & 0.68 & 6.9 + ngc3109 & & dirr & mw & 1360 & 8.7 & 0.054 & 0.065 & 0.084 & 0.12 & 11.1 + ngc185 & & dsph & m31 & 175 & 8.6 & 0.0053 & 0.0053 & 0.090 & 0.51 & 10.5 + ngc147 & & dsph & m31 & 101 & 8.4 & 0.032 & 0.036 & 0.050 & 0.17 & 12.4 + ic1613 & & dirr & m31 & 505 & 8.3 & 0.059 & 0.11 & 0.42 & 0.64 & 7.7 + wlm & ddo221 & dirr & m31 & 840 & 8.2 & 0.14 & 0.35 & 0.55 & 0.69 & 6.7 + sex b & ddo70 & dirr & mw & 1320 & 8.2 & 0.049 & 0.067 & 0.11 & 0.21 & 11.1 + sex a & ddo75 & dirr & mw & 1440 & 7.9 & 0.15 & 0.29 & 0.38 & 0.41 & 9.3 + sagittarius & & dsph & mw & 28 & 7.7 & 0.0008 & 0.0008 & 0.52 & 0.86 & 6.5 + fornax & & dsph & mw & 138 & 7.5 & 0.013 & 0.059 & 0.33 & 0.73 & 7.4 + ugc4879 & vv124 & dirr & mw & 1100 & 7.3 + pegasus & ddo216 & dirr & m31 & 410 & 7.2 & 0.057 & 0.095 & 0.40 & 0.64 & 7.4 + ugca92 & egb_0427 + 63 & dirr & mw & 1300 & 7.2 & & & & & + sag dig & eso594 - 4 & dirr & mw & 1060 & 7.1 & 0.11 & 0.17 & 0.20 & 0.20 & 11.6 + andvii & cassiopia dsph & dsph & m31 & 216 & 7.1 & 0.016 & 0.022 & 0.022 & 0.025 & 12.9 + andi & & dsph & m31 & 48 & 7.1 & 0.0038 & 0.0098 & 0.087 & 0.67 & 8.9 + andii & & dsph & m31 & 160 & 7.0 & 0.0049 & 0.0086 & 0.076 & 0.50 & 9.2 + andvi & pegasus dsph & dsph & m31 & 266 & 6.9 & 0.0023 & 0.023 & 0.19 & 0.60 & 9.0 + leo a & ddo69 & dirr & mw & 800 & 6.8 & 0.13 & 0.31 & 0.65 & 0.78 & 6.2 + antlia & & dsph & mw & 1330 & 6.8 & 0.043 & & & 0.43 & 9.0 + leoi & ddo74 & dsph & mw & 270 & 6.8 & 0.0099 & 0.18 & 0.50 & 0.76 & 6.4 + aquarius & ddo210 & dirr & mw & 950 & 6.7 & 0.037 & 0.083 & 0.12 & 0.12 & 12.0 + andiii & & dsph & m31 & 68 & 6.5 & 0.0022 & 0.0061 & 0.10 & 0.47 & 9.8 + cetus & & dsph & m31 & 680 & 6.4 & 0.0045 & 0.013 & 0.17 & 0.52 & 9.9 + lgs3 & pisces & dirr & m31 & 284 & 6.3 & 0.015 & 0.046 & 0.16 & 0.43 & 9.8 + leoii & ddo93 & dsph & mw & 205 & 6.3 & 0.0028 & 0.012 & 0.025 & 0.70 & 8.8 + phoenix & & dirr & mw & 405 & 6.3 & 0.027 & 0.071 & 0.23 & 0.42 & 10.3 + sculptor & & dsph & mw & 88 & 6.3 & 0.010 & 0.016 & 0.026 & 0.14 & 12.6 + tucana & & dsph & mw & 870 & 6.2 & 0.0048 & 0.011 & 0.014 & 0.30 & 11.6 + andxv & & dsph & m31 & 170 & 6.2 + andxvi & & dsph & m31 & 270 & 6.1 + sextans & & dsph & mw & 86 & 6.1 & 0.00 & 0.00 & 0.00 & 0.00 & 12.0 + andv & & dsph & m31 & 117 & 6.0 & 0.0045 & 0.048 & 0.066 & 0.35 & 10.8 + carina & & dsph & mw & 94 & 6.0 & 0.0065 & 0.0077 & 0.43 & 0.67 & 7.1 + draco & ddo208 & dsph & mw & 79 & 5.9 & 0.0004 & 0.010 & 0.025 & 0.49 & 10.9 + ursa minor & ddo199 & dsph & mw & 69 & 5.9 & 0.00 & 0.00 & 0.00 & 0.00 & 12.0 + andx & & dsph & m31 & 110 & 5.9 & & & & & + andxiv & & dsph & m31 & 162 & 5.8 + andxvii & & dsph & m31 & 44 & 5.8 & & & & & + andix & & dsph & m31 & 45 & 5.7 & & & & & + listed are all presently known satellite galaxies of the mw and m31 within 1 @xmath2 mpc of either host , with stellar mass @xmath3 .
star formation data are shown if available here
we parameterize the sfh in terms of the fraction of total stellar mass formed in the last 1 , 2 , 5 , and 10 gyr ( i.e. , @xmath4 , @xmath5 , @xmath6 , @xmath7 ) and the mean mass - weighted stellar age , @xmath8 . most of the star formation data are derived from the local group stellar photometry archive @xcite .
exceptions are m32 , m33 , antlia , sextans and ursa minor , which are taken from @xcite and @xmath8 is estimated from the reported @xmath7 value .
the lmc data are inferred from @xcite ; smc from @xcite .
distances from the host , @xmath9 , are directly taken or inferred from @xcite with some exceptions : andx @xcite , andxiv @xcite , andxv and andxvi @xcite , andxvii @xcite , ugc4879 @xcite , ugca 92 @xcite , and m33 @xcite .
stellar mass , @xmath10 , is from @xcite , except for m32 and sagittarius , whose values are estimated using @xmath11 , with @xmath12 from @xcite ; similarly we estimate @xmath10 for andix , andx , andxiv , andxv , andxvi , andxvii and ugc 4879 from the quoted @xmath12 or @xmath13 in the same references cited for @xmath14 .
the star formation data used in this study , and listed in table [ tab : sfdata_lg ] , come primarily from hubble space telescope observations with the wfpc2 camera . for a majority of the dwarfs ,
the sfhs were measured by the color - magnitude diagram fitting method , described in @xcite , using photometric data from the local group stellar photometry archive @xcite .
we condense the data into five parameters the fractions of stellar mass formed in the last 1 , 2 , 5 , and 10 gyr , respectively ( @xmath4 , @xmath5 , @xmath6 , @xmath7 ) , and the mass - weighted mean age of the stellar population ( @xmath8 ) .
the hst observations included in archive are far from homogeneous , and , therefore , the uncertainties in the derived star formation histories cover a broad range .
however , by concentrating on cumulative star formation fractions the uncertainties are greatly reduced , compared to specific star formation rates at specific times . thus , the entries in table [ tab : sfdata_lg ] do not have associated errors .
constraining comparisons between models and observations to cumulative distributions greatly reduces the sensitivity of these comparisons to uncertainties in individual galaxies .
note that the observed _ hst _ fields do not necessarily cover the full extent of each galaxy and , therefore , do not allow us to calculate the total stellar mass .
for the estimate of total stellar mass , with a few exceptions , we take the values quoted in @xcite .
the rest of the sfhs are taken from a variety of sources , as indicated in the notes to table [ tab : sfdata_lg ] . in the case of the lmc ,
the star formation data are inferred from @xcite .
since they give the star formation rates per unit area , in units of @xmath15 , separately for the disk and the bar of the lmc , we assume that the extent of the disk is 3.5 times larger than the extent of the bar in determining the star formation fractions and the mean age ( a. a. cole , personal communication ) .
the star formation histories for five dwarfs ( m32 , m33 , antlia , sextans , ursa minor ) are taken from table 1 of @xcite .
since that table does not include the fraction of a galaxy s star formation in the last @xmath16 gyr or @xmath17 gyr , our table is also missing these entries , except for sextans and ursa minor .
@xcite reported that sextans and ursa minor have formed no stars in the past 10 gyr ( i.e. , @xmath18 ) , which implies that the fractions of stars formed in the last 1 gyr , 2 gyr , and 5 gyr are also consistent with zero .
for the purpose of comparing the radial distributions of different morphological types with model predictions , we combine the sc , irr , and dirr types into one broadly - defined `` dirr '' group , and the de , dsph types into the `` dsph '' group . one object not included in table [ tab : sfdata_lg ] is the canis major dwarf , discovered by @xcite in the sdss field , which could be part of a larger monoceros tidal stream .
this object is the closest discovered to the galaxy , at a heliocentric distance of only 8 kpc .
whether this galaxy is still gravitationally self - bound is unclear at present @xcite , but in any case it is being strongly tidally disrupted .
an additional reason for not including it in our comparison is that such an object would have been completely disrupted in the @xmath1-body numerical simulation used to construct our models .
note also that table [ tab : sfdata_lg ] does not include any of the newly - discovered sdss and megacam ultrafaint dwarfs , since they are likely to have stellar masses below @xmath19 .
we list these new objects in table [ tab : lowmass_tbl ] and compare their estimated stellar mass function with the predictions of our fiducial model in [ sec : lowmass ] .
lccccc + & & & & & + + @xmath20 : @xmath21 & 3.1e-1 & 3.0e-1 & 2.5e-1 & 2.1e-1 & 2.2e-1 + @xmath20 : @xmath22 & 3.7e-1 & 3.6e-1 & 3.8e-1 & 3.6e-1 & 4.9e-1 + @xmath20 : @xmath23 & 8.7e-2 & 5.2e-2 & 4.2e-2 & 3.3e-2 & 3.9e-2 + @xmath20 : @xmath4 & 1.1e-19 & 2.0e-6 & 3.0e-3 & 3.1e-2 & 1.1e-1 + @xmath20 : @xmath5 & 2.0e-16 & 3.4e-2 & 4.3e-1 & 3.6e-1 & 1.0e-1 + @xmath20 : @xmath6 & 3.8e-13 & 2.5e-4 & 2.1e-2 & 7.8e-2 & 2.9e-1 + @xmath20 : @xmath7 & 1.3e-1 & 9.9e-2 & 4.4e-2 & 1.6e-2 & 2.5e-3 + @xmath20 : @xmath8 & 1.7e-3 & 1.3e-1 & 6.1e-1 & 3.8e-1 & 1.5e-1 + @xmath20 : @xmath10 & 6.2e-3 & 3.1e-3 & 3.4e-3 & 2.8e-3 & 1.2e-3 + dwarfs per halo & 20 & 21 & 22 & 24 & 26 + notes.other fixed parameters are @xmath24 , @xmath25 . the observed + number in the local group is 46/2 = 23 dwarfs per host halo .
the kgk04 model of star formation in dwarf galaxies is based on the mass assembly history of dark matter halos in a collisionless @xmath0cdm simulation of the local group - like environment .
the simulation volume contains three large host halos , with virial masses @xmath26 at @xmath27 , resolved with @xmath28 dark matter particles . the model incorporates the accretion of gas in hierarchical mergers and the loss of gas caused by the extragalactic uv background ( following the filtering mass approach of @xcite ) , and includes both a continuous mode and a starburst mode of star formation .
the model assumes that in the satellite halos the accreted gas dissipates its energy via radiative cooling and forms a disk .
the surface density of the gas follows an exponential profile , @xmath29 with the scale length @xmath30 set by the satellite halo s virial radius @xmath31 and angular spin parameter @xmath32 : @xmath33}. \label{eq : r_d}\ ] ] the last factor accounts for the less efficient gas dissipation in small halos with the virial temperature @xmath34 k , or equivalently , maximum circular velocity @xmath35 km s@xmath36 .
this factor is written in terms of @xmath37 km s@xmath36 , the virial velocity corresponding to the virial temperature @xmath38 k , and a constant , @xmath39 , chosen to reproduce the correct total number of dwarfs .
this important factor , coupled with the density threshold for star formation , suppresses star formation in most low - mass halos .
equation ( [ eq : r_d ] ) is consistent with results of recent hydrodynamic simulations of galaxy formation @xcite .
an alternative parametrization of dwarf gaseous disks with a temperature floor at @xmath40 k is given by @xcite . in their model
the disks are vertically puffed - up , which works to the same effect to reduce the gas density .
the observed stellar core radii in the local group dwarfs are a factor of @xmath41 smaller than those predicted by our model , which may favor the vertical expansion over the radial one assumed here . however , since it would not change the surface density @xmath42 , we can not fit it in our framework of the kennicutt - schmidt law of star formation .
the spin parameter @xmath32 is drawn randomly from a standard log - normal distribution with @xmath43 and @xmath44 @xcite .
this adds an intrinsic variance in the predictions of the star formation model . while the kgk04 model used only one set of the randomly - selected @xmath32 values to compare with the observations , in this paper we take 10 random realizations of each model , in order to account for this intrinsic variance .
the gaseous disk is modeled spatially by 50 radial zones . at each simulation output epoch
( about every @xmath45 yr ) , the newly accreted gas is added to these radial zones , according to the exponential profile of eq .
( [ eq : gas_radial ] ) . in each zone , the rate of star formation is determined by the kennicutt - schmidt law : @xmath46 wherever the gas density , @xmath42 , exceeds the threshold , @xmath47 @xcite .
standard parameters are @xmath48 , @xmath25 , and @xmath49 .
most of the variants of the model discussed in this paper employ this star formation law but with an important addition , described in [ sec : threshold ] below .
in addition to this quiescent mode of star formation , the model also allows a starburst mode prompted by strong tidal interactions with other halos ( see section 6.1 in kgk04 ) . in this mode
there is no density threshold and even galaxies with @xmath50 can form stars if a strong enough interaction occurs . finally , the effect of stellar evolution is taken into account following @xcite .
they find that some 40% of the initial stellar mass is lost to stellar wind and supernovae after @xmath51 gyr , given the assumptions of the initial stellar mass function from @xcite , stellar remnant masses as a function of initial stellar mass from @xcite , and main sequence lifetimes from @xcite . in the models investigated here
we simply assume that 40% of the initial stellar mass is lost between the time when the stars formed and the present day .
we do not `` recycle '' the liberated gas back into the ism , which would make it available to form more stars . as a result
our estimate of the stellar masses are the lower limit of the true masses a simple decrease of the total stellar mass , @xmath10 , for each dwarf by a factor of @xmath52 .
accordingly , the @xmath53g}$ ] and @xmath54 values are unchanged by stellar evolution in our models and since the _ hst _ photometry for the milky way satellites is typically good enough to measure the main sequence stars we simply compare the observed @xmath53g}$ ] and @xmath54 values ( table [ tab : sfdata_lg ] ) to the same quantities for the simulated dwarfs without any corrections .
we mark as luminous ( having corrected @xmath10 for stellar evolution as just described ) those satellite halos with a predicted stellar mass @xmath3 at @xmath27 .
such a cutoff agrees with the observational limit of all local group dwarfs ( table [ tab : sfdata_lg ] ) known until a few years ago , when the ultrafaint dwarfs were discovered .
we defer the discussion of the model predictions for these low - mass objects until [ sec : lowmass ] . as in kgk04 , we do a rough morphological classification of dwarf galaxies as dsph or dirr based on the ratio of the stellar rotation velocity to the velocity dispersion : @xmath55 for dsph and @xmath56 for dirr . in the model , the rotation velocity is calculated as the circular velocity at the outer - most stellar radius , as it would be measured in observation , while the velocity dispersion is estimated from the amount of external tidal heating in strong tidal interactions with other halos
. such classification does not take into account the recent star formation activity or the remaining gas content , and therefore , is only a crude indication of the observationally defined morphological type .
comparison of these model predictions to the data is , in fact , completely complementary to the comparison of the star formation histories . in the process of revising our models
we have discovered that the gas densities @xmath42 in @xcite were underestimated by a factor of 2 due to an error in the code .
when we quote the results for the kgk04 model here , we use the corrected values . despite significant successes in explaining the number and spatial distribution of the local group dwarfs , the kgk04 model did not predict a sufficient amount of recent star formation .
figure [ fig : sfplots ] shows that 95% of the dwarfs in that model have not formed any stars in the last 1 gyr , in serious disagreement with data . to quantify the level of this disagreement
, we use the kolmogorov - smirnov ( ks ) test for cumulative distribution functions of the following parameters : the distance to the host ( @xmath21 ) , the distances for the dsph and dirr galaxies separately ( @xmath57 and @xmath23 ) , the star formation fractions ( @xmath4 , @xmath5 , @xmath6 , @xmath7 ) , the mean age ( @xmath8 ) , and the stellar mass ( @xmath10 ) . table [ tab : sfdata ] shows that the recent star formation fractions in the kgk04 model ( see column @xmath58 ) have very low ks probabilities , below @xmath59 . on the other hand ,
the @xmath7 fraction , which is a more global measure of the overall star formation , is consistent with the data at the 13% level .
thus the star formation law used in the model is not necessarily at fault , but the apparent lack of recent star formation is a direct result of the fixed star formation threshold , as we will see below .
the mean mass - weighted age is also inconsistent with the data , but at a less significant level ( @xmath60 ) .
the mean age is a global , integral measure of the sfh , which combines the low - probability recent star formation with a higher - probability early star formation .
the stellar age is systematically overestimated in the model , by a few gyr .
the total stellar mass is also overestimated by a factor of several for most dwarfs .
the cumulative mass function is inconsistent with the data at a level similar to the age distribution , @xmath61 .
note also that kgk04 considered principally the satellite dwarfs located within the virial radius of the host galaxy at @xmath62 kpc . we extend our analysis to all dwarfs in the local group , even the potentially isolated ones lying outside the virial radius of either host .
figure [ fig : radial ] shows their cumulative radial distribution out to @xmath63 kpc .
the kgk04 model does well for the distribution of all dwarfs and the distribution of irr / dirr types separately .
the model slightly overestimates the number of dsph / de types outside @xmath64 kpc , but in all cases the ks probability is @xmath65% or higher , fully consistent with the observations .
our phenomenological model contains several parameters that allow for some freedom in the outcome .
two of the most important parameters in the model are the threshold density @xmath47 and the disk size parameter , @xmath66 ( eq . [ eq : r_d ] ) , both of which are difficult to know , _ a priori _ , from theory .
since these parameters can significantly change the predicted total number of dwarfs , the observed number of ( @xmath67 ) dwarfs provides a fairly stringent constraint on these values .
for example , with @xmath68 and @xmath49 , the predicted average number of dwarfs per host halo within 1 @xmath2 mpc is 20.5 .
if we take @xmath69 ( with @xmath49 ) this number increases to 40.7 , and if we take @xmath70 the number of dwarfs drops to 14.3 .
analogously , reducing the threshold density to 4 and @xmath71 ( with @xmath68 ) increases the average number to 21.7 and 22.3 , respectively .
none of these fixed threshold models , however , adequately reproduces the observed star formation histories .
therefore , we look for additional physical ingredients for our model . in our extension of the kgk04 model , we attempt to retain the correct radial distribution of the dwarf galaxies , while improving the predictions for their star formation histories .
we consider a number of modifications to the kgk04 model .
the most promising of the modifications is the introduction of a stochastic threshold to the star formation law . at each output epoch
through the course of the simulation , the threshold density for star formation is drawn from a log - normal distribution with a mean value @xmath72 and a small dispersion @xmath73 : @xmath74 } \ ; d\sigma_{\rm th}. \label{eq : sig_distrib}\ ] ] as a fiducial mean value we take @xmath75 , but we also vary that parameter in some runs
. what does this stochastic threshold mean ? and what , physically , is the threshold of star formation ?
we apply it to the azimuthally - averaged surface density of gas . in nearby star forming regions , which we can study directly with _
hst _ and _ spitzer _ , stars form in dense molecular clouds ( e.g. , * ? ? ? * ) .
locally the density is high , but the azimuthal average at a particular distance @xmath76 from the center may be either high ( if molecular clouds are common at @xmath76 ) or low ( if they are rare ) . thus , by invoking the threshold @xmath47 , we are effectively parameterizing the fraction of molecular gas available for star formation .
when we take the threshold to vary , we are thus accounting for stochastic star formation in isolated hii regions , such as those found by _
galex _ in nearby spirals @xcite .
nearby dwarf galaxies often show a very high hi gas fraction , i.e. the ratio of gas mass to baryon ( gas+stars ) mass , up to 90% @xcite .
the current star formation rates in those galaxies are low , and therefore most of the gas is inert .
hydrodynamic simulations of @xcite , which treat the formation of molecular hydrogen in detail , also suggest that dwarf disks may contain large reservoirs of diffuse atomic gas that is unable to condense in molecular clouds and participate in star formation .
this inefficiency of forming molecular clouds at low densities effectively results in the threshold density for star formation .
figure [ fig : sfplots ] shows that the stochasticity greatly improves the agreement of the recent star formation fractions ( @xmath77 , @xmath6 ) with the local group data .
the predicted distributions are much closer to the observed ones than in the kgk04 model , and are , in fact , statistically consistent with each other at @xmath78 level .
we quantify this effect by gradually increasing the amount of stochasticity , @xmath79 .
table [ tab : sfdata ] shows the ks test results for @xmath79 ranging from 0.05 to 0.2 , which approximately corresponds to 10% to 60% variation in the threshold density @xmath47 .
even a small amount of stochasticity , @xmath80 , leads to a dramatic improvement of the @xmath81 , and @xmath6 statistics .
the probabilities increase rapidly with @xmath79 to a good fraction of a percent or more .
the model with @xmath82 is already statistically consistent with the data . with the other parameters of the star formation law ( @xmath83 and @xmath84 ) being fixed
, increasing @xmath79 leads to better recent star formation parameters but worse early star formation parameters .
the probability of the @xmath7 distribution decreases from 13% to under 5% for @xmath82 and even below a percent for @xmath85 .
the mean age is most consistent with the data ( @xmath86 ) for @xmath82 .
the total number of dwarfs also increases systematically with @xmath79 while the stellar mass function and the radial distributions are not significantly affected by the variation of @xmath79 .
additionally , the @xmath87 and @xmath85 models begin to overpredict the total number of luminous dwarfs per host halo , implying that at these values the overall star formation becomes too efficient .
this overabundance can be seen in table [ tab : sfdata ] , or graphically in fig .
[ fig : radial ] .
therefore , we take the case with @xmath88 as our fiducial model striking the best balance for all star formation statistics .
we consider other variants of the stochastic model in
[ sec : other ] .
figures [ fig : sfr_gasprofile ] and [ fig : sfr_vs_sfr ] illustrate why the stochasticity is so successful .
they show star formation histories and corresponding gas density profiles for two representative model dwarfs .
figure [ fig : sfr_gasprofile ] is for the case of a fixed density threshold , @xmath89 . without the stochasticity , and in the absence of radial gas flows , the gas above the threshold is steadily converted into stars at early times ( 10 and 12 gyr ago ) at substantial rates , @xmath90 . at these epochs
the galaxies evolve effectively in isolation , while still growing by gas - rich hierarchical mergers . at later time ,
when these galaxies become satellites of the larger host galaxy , they are tidally truncated and no new gas is accreted .
after the high - density gas supply is exhausted , the rest of the gas hovers just under the threshold , unable to form new stars . in the first dwarf , shown in the top panels in figure [ fig : sfr_gasprofile ] , star formation almost completely halts 8 gyr ago
the other dwarf , shown in the bottom panels , experiences several distinct episodes of star formation , the last one finishing 3 gyr ago . in both cases there
is effectively no star formation in the last 2 gyr , @xmath91 .
if instead the threshold density varies in time , at later epochs the gas may find itself above the threshold and allow more recent star formation .
figure [ fig : sfr_vs_sfr ] shows the star formation histories for the same dwarfs but now with a variable threshold and using @xmath82 .
there is more star formation overall and more star formation at later times . as a result , the stellar mass is higher by 20% to 30% and mean stellar age is lower by about 1 gyr . more importantly , several percent of all stars are formed in the last 2 gyr .
this stochastic enhancement of the star formation rate allows the model to reproduce the star formation episodes at late times when gas - rich mergers ( which increase the available gas supply ) and strong interactions ( which can prompt starbursts ) are relatively less common .
we also show on fig .
[ fig : sfr_vs_sfr ] the observed star formation histories of two dwarf galaxies , ngc 3109 and ic 1613 .
we do not expect the model predictions to correspond in detail to the observed sfh features , as our modeling is necessarily statistical and is aimed at explaining not a specific galaxy s sfh but only an ensemble of sfhs of many local group dwarfs .
still , qualitatively , the agreement between the model and the data is good : in the case of ngc 3109 both show most stars being formed at early times with a small fraction in the last 4 gyr , while in the case of ic 1613 , both sfhs continue until the present in several distinct , extended episodes . despite the impressive improvements ,
there still remain discrepancies of the fiducial model with the observed data . first , there is still not quite enough very recent star formation , a problem which is quantified by the ks - test result for @xmath4 in table [ tab : sfdata ] ( @xmath92 ) and is apparent in fig .
[ fig : sfplots ] . the models with a higher amount of stochasticity , @xmath87 and @xmath85 ,
achieve better agreement with the observed @xmath4 distribution but they are disfavored for skewing all other star formation statistics .
thus , in the fiducial model ( @xmath82 ) over 25% of the dwarfs have less than @xmath93 of their stars formed in the last 1 gyr , compared with about 10% of such dwarfs in the observed sample .
note that the current observations are sensitive to sfhs with large fractions of star formation at later times .
the very deep _ hst _ acs imaging of leo a by @xcite have confirmed that majority of all star formation has occurred in the last half of the age of the universe .
second , the total stellar masses of the luminous dwarfs in the models tend to be significantly above the observed stellar masses of the local group dwarfs .
this result is apparent for any value of @xmath79 between 0 and 0.2 ( see bottom right panel of fig . [
fig : radial ] ) .
this excess stellar mass in our models can perhaps be traced to the fact that we do not include thermal and ionizing feedback from young massive stars ( e.g. , * ? ? ?
* ; * ? ? ?
* ) , which can disrupt molecular clouds and drive galactic outflows , thus reducing the available gas supply for star formation .
third , the model does not predict enough very _ early _ star formation .
this is a generic problem with any hierarchical model in which the presently - massive satellites form and accrete onto the host late and , at early times , therefore , do not contain significant amounts of gas above the threshold .
our model predicts that some 20% to 30% of the dwarf galaxies will have formed greater than 85% of their stellar mass in the last 10 gyr ( i.e. @xmath94 ) . in other words
, there exists in the model a number of dwarfs with anomalously young stellar populations , whereas , by contrast , the local group data in table [ tab : sfdata_lg ] does not show any dwarfs having formed more than 86% of its stellar mass in the last 10 gyr ( the saggitarius dsph has the highest fraction ) at least 14% of all stars in the observed dwarfs formed in the first 4 gyr after the big bang . some of the simulated dwarfs with very young stellar populations include objects which form the bulk of their stars in one , punctuated star formation event caused by a close tidal interaction , prompting a starburst which happened to occur in the last 10 gyr .
but not all of the galaxies in this category form stars through the starburst mode ; other dwarfs with very young stellar populations form all of their stars through the continuous mode with no starbursts at all .
lcc + & & + + @xmath20 : @xmath21 & 2.5e-1 & 2.4e-1 + @xmath20 : @xmath22 & 3.8e-1 & 4.1e-1 + @xmath20 : @xmath23 & 4.2e-2 & 3.1e-2 + @xmath20 : @xmath4 & 3.0e-3 & 9.1e-2 + @xmath20 : @xmath5 & 4.3e-1 & 8.1e-1 + @xmath20 : @xmath6 & 2.1e-2 & 1.9e-1 + @xmath20 : @xmath7 & 4.4e-2 & 5.8e-3 + @xmath20 : @xmath8 & 6.1e-1 & 3.8e-1 + @xmath20 : @xmath10 & 3.4e-3 & 6.3e-4 + dwarfs per halo & 22 & 23 + other parameters are @xmath82 , @xmath75 , @xmath25 .
with the discrepancies discussed in [ sec : discrepancies ] in mind , we have explored several variants of our star formation model in order to check if relaxing other model assumptions can improve the predicted star formation histories to the point where there is broad agreement with the observations .
one of these alternate models is motivated by the lack of enough recent star formation . in the kgk04 model and in the stochastic models shown in figs .
[ fig : sfplots ] and [ fig : radial ] , accretion of new gas onto the dwarf halos is shut off whenever the dwarf comes within the virial radius of the host halo .
this is based on the expectation that satellite halos near their host would be tidally truncated and unable to capture new gas even if they happen to increase their dark matter mass in mergers with other satellite halos .
however , becoming a satellite does not immediately stop the halo s star formation activity which can continue as long as the gas density remains above the threshold or if a starburst event is triggered . in the first of these alternative models
, we lift this assumption and allow the accretion of new gas even within the virial radius of the host , at @xmath95 . as a result ,
the agreement between the model and the data for the @xmath4 distribution improves as indicated by the ks - test results shown in table [ tab : late_acc ] from the rejection of the null hypothesis of the same - distribution at 0.3% significance level to rejection only at the 9% level , essentially statistically consistent .
qualitatively , the improvement is most dramatic for @xmath4 since allowing the additional accretion of gas is more important at late times , and because the virial radius of the host halo will be larger than at earlier epochs ; also the dwarf galaxies at late times are likely to be closer in .
the extra gas is converted into extra stars and so the @xmath4 value increases and the typical mass of the dwarfs increases .
generally , the extra accretion shifts the stellar age distribution to be younger .
however , another predicted feature of our models is a tail of the age distribution at @xmath96 gyr .
the extra accretion exaggerates this tail since it increases recent star formation .
in contrast , the youngest dwarf galaxy in the local group is leo a , with @xmath97 gyr ( though see the leo a sfh of @xcite , which has a significantly younger result for @xmath8 ) .
this is fundamentally the same problem as the existence of simulated dwarfs with @xmath94 , and the @xmath7 distribution similarly becomes less consistent with the data in this model .
lcccccc + & & & & + + @xmath84 & 1 & 0.5 & 0.14 & 0.012 + @xmath20 : @xmath21 & 2.5e-1 & 2.5e-1 & 2.6e-1 & 2.6e-1 + @xmath20 : @xmath22 & 3.8e-1 & 4.4e-1 & 4.2e-1 & 4.8e-1 + @xmath20 : @xmath23 & 4.2e-2 & 3.5e-2 & 3.3e-2 & 3.5e-2 + @xmath20 : @xmath4 & 3.0e-3 & 1.4e-2 & 2.5e-2 & 8.3e-2 + @xmath20 : @xmath5 & 4.3e-1 & 2.4e-1 & 2.4e-1 & 1.8e-1 + @xmath20 : @xmath6 & 2.1e-2 & 1.1e-1 & 9.4e-2 & 1.3e-1 + @xmath20 : @xmath7 & 4.4e-2 & 5.7e-3 & 7.7e-3 & 8.3e-3 + @xmath20 : @xmath8 & 6.1e-1 & 1.1e-1 & 4.5e-1 & 5.2e-1 + @xmath20 : @xmath10 & 3.4e-3 & 3.7e-3 & 3.1e-3 & 5.0e-3 + dwarfs per halo & 22 & 22 & 22 & 21 + other parameters are @xmath82 , @xmath75 . another important question to be addressed is whether the observed sfhs , given our model assumptions , favor steepening of the schmidt - kennicutt law ( @xmath98 ) .
there is some evidence for the exponents @xmath99 to become gradually larger at low gas densities in dwarf galaxies ( e.g. , @xcite ; see also discussion in @xcite ) .
since most of our dwarfs are close to the threshold , we investigate this effect simply by changing the exponent at all densities , setting it to @xmath100 , 2 , or 3 . for each of these values
, the normalization @xmath84 is allowed to float such that the overall star formation efficiency is lowered until the point where the median of the age distribution , @xmath101 , of the simulated dwarfs matches that of the local group galaxies , 9.25 gyr .
table [ tab : powerlaw ] shows the required values of @xmath84 and the results of ks - tests .
the rather high significance results from the ks tests applied to the @xmath8-distribution are not particularly surprising since the normalization of the star formation law has been fine - tuned for this comparison .
qualitatively this exercise keeps the overall star formation rate the same , i.e. the stellar mass function is basically unchanged . and
similarly the radial distributions stay roughly the same this is important since one of the primary successes of the kgk04 model was reproducing the observed radial distribution of the local group dwarfs , something that our models continue to do .
eventually the same amount of gas above the density threshold is converted into stars , so the rate at which star formation proceeds does not alter the total number or the radial distribution of the simulated dwarfs . by and large , all of the same discrepancies with the data listed in [ sec : discrepancies ] are apparent in these models with different @xmath99 too large stellar masses , not enough recent star formation , and too much overall star formation in the last 10 gyr .
arguably the \{@xmath102 , @xmath103 } model is a best fit to the data for @xmath4 and @xmath6 , however the fiducial \{@xmath100 , @xmath25 } model is still the best for @xmath5 and @xmath7 .
further , comparisons of the stellar mass and the @xmath54-distribution are inconclusive as well .
lcccccc + & & & & + + @xmath104 & 9 & 7 & 7 & 6 + @xmath105 & 10 & 10 & 8 & 10 + @xmath20 : @xmath21 & 2.3e-1 & 2.9e-1 & 2.5e-1 & 3.2e-1 + @xmath20 : @xmath22 & 5.7e-1 & 5.6e-1 & 3.8e-1 & 5.3e-1 + @xmath20 : @xmath23 & 4.2e-2 & 4.7e-2 & 4.2e-2 & 6.0e-2 + @xmath20 : @xmath4 & 1.6e-2 & 7.7e-3 & 3.0e-3 & 1.3e-2 + @xmath20 : @xmath5 & 7.7e-1 & 6.4e-1 & 4.3e-1 & 6.6e-1 + @xmath20 : @xmath6 & 1.2e-1 & 8.2e-2 & 2.1e-2 & 6.3e-2 + @xmath20 : @xmath7 & 4.8e-3 & 9.9e-3 & 4.4e-2 & 1.5e-2 + @xmath20 : @xmath8 & 6.3e-1 & 6.9e-1 & 6.1e-1 & 6.7e-1 + @xmath20 : @xmath10 & 1.6e-3 & 1.3e-3 & 3.4e-3 & 2.4e-3 + dwarfs per halo & 20 & 19 & 22 & 18 + @xmath104 refers to the redshift when reionization is completed , while @xmath105 refers to the redshift where cosmic hii regions begin to overlap ( see appendix b in @xcite for details ) .
other parameters : @xmath82 , @xmath75 , @xmath25 .
another interesting test is to vary the epoch of reionization .
appendix b of kgk04 presents an analytic fit to the numerical results of @xcite that parametrize the reionization epoch in terms of the redshift when cosmic hii regions begin to overlap , @xmath105 , and the redshift when the reionization is completed , @xmath104 .
varying these two parameters allows us to test the sensitivity of model predictions to the details of the evolution of the ionizing background radiation .
note that our parametrization of reionization here applies only to the local neighborhood of the galaxy and may differ from the global cosmic reionization .
reionization affects the amount of gas accreted onto dwarf halos .
increasing extragalactic uv flux during reionization photoionizes the gas inside and outside dark matter halos and prevents the halos with shallow potential wells from capturing new gas heated to @xmath40 k. in @xcite the effect of decreasing the gas fraction of halos was parametrized as @xmath106 where @xmath107 is the universal baryon fraction , @xmath108 is the halo mass , and @xmath109 is the cut - off mass parameter . in linear theory for baryon perturbations , this parameter can be related to the filtering mass as @xmath110 .
the filtering mass @xmath111 is an integral of a function of the intergalactic gas temperature over cosmic history and corresponds to the mass of a halo that loses 50% of its baryons as a result of external photoheating .
the values of the filtering mass from @xcite can be calculated for any @xmath105 and @xmath104 using eq .
( b1 ) in @xcite .
we consider four scenarios listed in table [ tab : reioniz ] : early \{@xmath112 , @xmath113 } , extended \{@xmath114 , @xmath113 } , and late \{@xmath115 , @xmath113 } reionizations , in addition to the standard scenario \{@xmath114 , @xmath116 } , which was used in kgk04 and in our fiducial model .
generally , the results are quite similar for all scenarios , likely because luminous satellites are hosted by the relatively massive halos , in which the virial temperatures are above @xmath117 k and the external heating of the gas does not affect its distribution significantly . early reionization is slightly preferred for the @xmath4 , @xmath5 , and @xmath6 distributions , but statistically all scenarios are consistent with each other .
the @xmath7 distribution is still reproduced best by the fiducial model .
recent hydrodynamic simulations indicate that the filtering mass may overestimate the mass of the halos that lose 50% of their baryons , especially at low redshift .
the most sophisticated art simulations of @xcite , including the effects of radiative transfer , give distributions of @xmath118 at several redshifts , @xmath119 ( their fig .
2 ) for the standard model of reionization .
we have fit these distributions with the form of equation ( [ eq : fgas ] ) and obtained best - fitting values of @xmath109 .
these values are similar to our old values at @xmath120 but deviate from them at lower redshift by as much as an order of magnitude .
we supplement these high - redshift fits with the @xmath27 results of sph simulations by @xcite and @xcite , which both give @xmath121 , about a factor of 5 smaller than our old value .
the combined sets at @xmath27 and @xmath122 can be fit by the following expression , accurate to better than 50% : @xmath123 ideally , we would like to derive the cutoff mass evolution from several simulations covering the whole range of redshifts , but at the moment it is the best we can assemble from the literature . we have run our model with the new expression ( [ eq : new_mc ] ) for @xmath109 and found that the corresponding changes in the predicted sfhs are not straightforwardly better or worse . the @xmath4 distribution improves a little , @xmath5 and @xmath6 are effectively unchanged , while the @xmath7 distribution is significantly worse off and the anomalously young dwarfs still persist .
the age distribution and stellar mass function are also more discrepant with the data than in the fiducial model . while the lack of improvement with the new cutoff mass prescription is unpleasant or , at any rate , suprising ,
until we have a more robust estimate of @xmath109 confirmed by several groups , we keep our fiducial model untouched .
lccc + & & & + & & & + + @xmath20 : @xmath21 & 3.1e-1 & 2.5e-1 & 3.0e-1 + @xmath20 : @xmath22 & 3.7e-1 & 3.8e-1 & 3.4e-1 + @xmath20 : @xmath23 & 8.7e-2 & 4.2e-2 & 3.7e-2 + @xmath20 : @xmath4 & 1.1e-19 & 3.0e-3 & 3.7e-2 + @xmath20 : @xmath5 & 2.0e-16 & 4.3e-1 & 7.1e-1 + @xmath20 : @xmath6 & 3.8e-13 & 2.1e-2 & 1.6e-2 + @xmath20 : @xmath7 & 1.3e-1 & 4.4e-2 & 2.7e-2 + @xmath20 : @xmath8 & 1.7e-3 & 6.1e-1 & 5.0e-1 + @xmath20 : @xmath10 & 6.2e-3 & 3.4e-3 & 2.0e-3 + dwarfs per halo & 20 & 22 & 23 + in addition to exploring the stochastically variable density threshold , we have also investigated a threshold @xmath47 that varies with redshift monotonically .
though , at the moment , we do not know of a convincing physical motivation for such a global systematic variation , we have investigated this possibility in exploring the full range of predictions of our model . here
we used a simple parameterization of the @xmath47 redshift dependence , @xmath124 where @xmath125 and @xmath126 , which results in @xmath127 .
interestingly , the results of this model are very similar to our fiducial model with the stochastic threshold .
the comparison is shown in table [ tab : mono ] .
the new model shares the same problems as the fiducial model : the stellar masses are still too large and the anomalously young dwarf problem is still present ( if not slightly worse ) .
we do not consider any other variants of the monotonic threshold but note that if a compelling theoretical or observational motivation for such a variation appears in the future , an evolving threshold may become a viable model .
lcccccc + & & & & + & & & & + + @xmath72 & 1.87 & 2.6 & 3.5 & 5.0 + @xmath20 : @xmath21 & 5.3e-1 & 4.0e-1 & 3.4e-1 & 2.5e-1 + @xmath20 : @xmath22 & 1.7e-1 & 3.7e-1 & 5.1e-1 & 3.8e-1 + @xmath20 : @xmath23 & 2.2e-1 & 9.6e-2 & 7.3e-2 & 4.2e-2 + @xmath20 : @xmath4 & 1.6e-3 & 3.3e-3 & 4.9e-3 & 3.0e-3 + @xmath20 : @xmath5 & 2.5e-1 & 5.0e-1 & 5.6e-1 & 4.3e-1 + @xmath20 : @xmath6 & 5.0e-3 & 7.3e-3 & 1.4e-2 & 2.1e-2 + @xmath20 : @xmath7 & 3.3e-1 & 3.5e-1 & 3.2e-1 & 4.4e-2 + @xmath20 : @xmath8 & 1.3e-1 & 3.1e-1 & 4.0e-1 & 6.1e-1 + @xmath20 : @xmath10 & 2.8e-4 & 2.7e-4 & 2.9e-4 & 3.4e-3 + dwarfs per halo & 22 & 22 & 23 & 22 + here we reject dwarfs that have not formed any stars by @xmath128 gyr , 1 gyr , or 2 gyr after the big bang ( redshifts @xmath129 , @xmath130 , and @xmath131 , respectively ) .
other parameters : @xmath82 , @xmath25 . finally , in directly addressing the problematic issue of the simulated dwarfs with too young stellar populations , we consider models that reject dwarfs with delayed star formation , i.e. the dwarfs that have not formed any stars within the first few gyr after the big bang . the reasoning for such an _ ad hoc _ cut is motivated by the uncertainty in the detailed effect of the early uv background on the gas content of low - mass halos .
the satellites without early star formation in our model must have acquired a significant gas reservoir only at late times . in the hierarchical paradigm , these satellites have been built by mergers of smaller objects , each of which carried an even smaller amount of gas . in order to become a galaxy in our model , at some time
the total gas density distribution in the satellite must reach above the density threshold . however , if for any reason we underestimated the effect of gas loss from small halos at early times , then at that time the combined amount of gas would be overestimated and star formation should not take place .
our treatment of the gas heating during reionization is very approximate , which makes it likely that we could either underestimate or overestimate the gas loss effect for particular dwarfs . in general ,
star formation at very high redshift could proceed even at lower densities than assumed in our models , such that all luminous dwarfs form at least a fraction of their stars before reionization @xcite . in the current concordance cosmology @xcite , the epoch of complete reionization can range from @xmath132 to @xmath133 , which corresponds to a range of times from 0.5 to 1 gyr after the big bang .
given this uncertainty , and uncertainties in the details of gas outflows , we consider three models that reject the dwarfs without star formation in the first 0.5 gyr , 1 gyr , or 2 gyr after the big bang .
the last model with @xmath134 gyr is the least restrictive and closest to the fiducial model ( no rejection ) , which can be formally written as @xmath135 . in each of these
`` reject late beginners '' models , the star formation density threshold @xmath72 is lowered in order to raise the number of dwarfs back to the number observed in the local group .
the details of these models are listed in table [ tab : reject_lb2 ] .
figure [ fig : sfplots_early2 ] shows the star formation parameters in the three models .
all of the distributions are similar to each other , but they all deviate from the fiducial model ( plotted in corresponding panels of fig .
[ fig : sfplots ] ) in one important aspect .
the @xmath7-distribution lacks objects with @xmath136 , i.e. with no star formation in the first 4 gyr after the big bang .
unfortunately , the time resolution of the observed sfhs does not allow us to discriminate among the three variants of the cut .
but all of them provide the needed fix : the @xmath7-distribution is fully consistent with the data ( @xmath137 ) .
table [ tab : reject_lb2 ] shows that the other probabilities remain roughly the same or even decrease relative to the fiducial model .
if we are to accept any of the `` reject late beginners '' variants , we would prefer the least restrictive @xmath134 gyr model .
ideally , of course , we prefer to develop a better understanding of early star formation that would make this cut unnecessary .
figure [ fig : noearly_t_vs_m ] shows the stellar age as a function of stellar mass in the fiducial model and its three variants . rejecting any of the `` late beginners '' increases the stellar age by less than 1 gyr , which is significantly smaller than the dispersion of the sample at all masses .
a general trend , also largely overcome by the dispersion , is for the mean age to increase with mass until @xmath138 and then to decrease at larger masses .
the local group data are consistent with the younger stellar ages , and extended sfhs , for more massive galaxies . at the lowest - mass end , however , the observations show a number of very old objects that are still not present in our models , even with the strictest age cut .
it is apparent from this and previous plots that our prescription for star formation in the first few gyr of cosmic time still needs improvement .
it is possible that the starburst mode of star formation is causing very young stellar ages in small galaxies .
we have checked , however , that the anomalously young dwarfs do not all have significant starbursts at late times and are hosted by dark matter halos with a wide range of masses .
in fact , galaxies with the highest fraction of stellar mass built in starbursts are typically old , with @xmath54 between 10 and 12 gyr .
starbursts are not very important overall
91% of the dwarfs have less than 10% of their stellar mass formed in the starburst mode .
also , we have varied the two parameters describing the starburst mode ( the minimum required tidal force and the fraction of gas converted into stars ) and found that it has little effect on the sfh distributions . it should also be mentioned that there does seem to be a trend , at @xmath139 , for the anomalously young dwarf halos to be more massive than the other , more typical , dwarf halos in the model observationally speaking , these dwarfs would have higher dynamical mass - to - light ratios .
interestingly , at the end of reionization these anomalous dwarfs are hosted by halos with a very wide range of masses , so that there is significant dynamical evolution from the epoch of reionization until the present day .
in other words , the younger dwarfs really do have rather `` tumultuous '' lives .
lcrccc + & & & & & + + leo t & mw & 417 & @xmath140 & @xmath141 & 1 + canes venatici i & mw & 218 & @xmath142 & @xmath143 & 2 + andxi & m31 & 103 & @xmath144 & @xmath145 & 3 + andxiii & m31 & 95 & @xmath146 & @xmath147 & 3 + andxii & m31 & 116 & @xmath148 & @xmath149 & 3 + hercules & mw & 138 & @xmath150 & @xmath151 & 4 + botesi & mw & 62 & @xmath152 & @xmath153 & 5 + ursa major i & mw & 106 & @xmath154 & @xmath155 & 6 + leoiv & mw & 158 & @xmath156 & @xmath157 & 4 + canes venatici ii & mw & 151 & @xmath158 & @xmath159 & 4 + sdssj100 + 5730 & mw & 83 & @xmath160 & @xmath161 & 7 + sdssj1329 + 2841 & mw & 76 & @xmath162 & @xmath163 & 7 + ursa major ii & mw & 32 & @xmath164 & @xmath165 & 8 + coma berencies & mw & 44 & @xmath166 & @xmath167 & 4 + botesii & mw & 60 & @xmath168 & @xmath169 & 9 + willmani & mw & 38 & @xmath170 & @xmath171 & 10 + (1 ) @xcite ; ( 2 ) @xcite , @xcite ; ( 3 ) @xcite ; ( 4 ) @xcite ; ( 5 ) @xcite , @xcite ; ( 6 ) @xcite ; ( 7 ) @xcite ; ( 8) @xcite ; ( 9 ) @xcite ; ( 10 ) @xcite , @xcite .
+ notes.@xmath10 is estimated assuming @xmath172 , with @xmath12 determined from @xmath13 quoted in the references .
satellites of mw are discovered by the sdss , satellites of m31 by the megacam survey .
satellite - to - host radii , @xmath9 , for andxi , andxii , & andxiii are derived using their projected distances from m31 , assuming @xmath173 kpc @xcite .
given the amazing rate of recent discoveries of the population of ultrafaint dwarfs in the sdss and megacam surveys , there stands a challenge to predict the yet - to - be - observed star formation properties of these objects .
our models should in principle be able to predict the mean age and stellar masses for the low - mass dwarfs , however current predictions are not satisfactory .
figure [ fig : lowmass_projections ] shows our fiducial model ( with the stochasticity parameter @xmath82 ) extended to masses as low as @xmath174 .
this model reproduces most closely the observed sfhs of the higher - mass dwarfs , but at @xmath175 , it predicts only a modest increase in the number of galaxies . the thick dashed line in fig .
[ fig : lowmass_projections ] shows the currently - known number of ultrafaint dwarfs , listed in table [ tab : lowmass_tbl ] .
the model predictions actually agree very well with the observed number and with the gentle slope of the mass function .
the problem , though , is that the sdss data release 5 , where the bulk of the new dwarfs have been discovered , covers only 20% of the sky .
therefore , we might expect the full sky to contain 5 times as many yet - to - be - discovered ultrafaint dwarfs @xcite . for the megacam survey ,
the incompleteness factor is even larger , @xmath176 @xcite .
extrapolation to all sky , by multiplying the observed number of dwarfs by these correction factors , is shown by the thin dashed line in fig .
[ fig : lowmass_projections ] .
our model falls well below this corrected mass function .
@xcite calculate the expected luminosity function of the faint dwarfs more accurately , by estimating the maximum accessible volume of the survey .
they conclude that the luminosity function should rise as @xmath177 , which for a fixed mass - to - light ratio results in @xmath178 .
this estimate is also plotted by a straight line in fig .
[ fig : lowmass_projections ] .
it lies below our first naive estimate but still significantly above the range of the model .
both incompleteness corrections predict over 50 satellites above @xmath179 , a factor of two larger than in the model .
plus , more dwarfs may remain undetected at larger distances than those probed by the current surveys ( @xmath180 kpc ) . the discrepancy at low mass persists for all variants of our model and we were unable to find a set of parameters which adequately reproduced the stellar mass function at @xmath181 while predicting appreciable numbers of @xmath182 objects .
this rather striking result is evidence that our models are not capturing some critical aspects of the formation of very low mass galaxies .
the most naive solutions of significantly reducing either the threshold @xmath47 or the disk structure factor @xmath66 in eq .
( [ eq : r_d ] ) , thereby allowing a larger portion of gas to participate in star formation , are not viable options since these modifications invariably overpredict the number of @xmath183 dwarfs . in other words
, lowering the density threshold produces so many dwarfs that the `` missing satellites '' are no longer missing and we are once again left with the expectation that we should see in the sky @xmath184 or more luminous dwarf galaxies around the milky way , whereas we only observe @xmath185 .
@xcite also reports a deficit of ultra - faint dwarfs from an independently - developed semi - analytical model of galaxy formation applied to a different collisionless @xmath1-body simulation and using the same cosmic reionization model employed here .
observationally , a new interesting puzzle for our understanding of dwarf galaxy formation is presented by @xcite . in that study
they observe the hi emission from leo t , the only ultrafaint dwarf with measurable gas content and recent star formation .
based on the density and temperature of hi gas and velocity dispersion of stars , they find that the gas is everywhere globally jeans - stable , whereas the observed pockets of blue , 200 myr - old stars indicate continuous star formation
. the observed peak of hi column density is a few @xmath186 , close to the star formation threshold .
leo t may thus present another example of stochastic star formation in a handful of isolated molecular clouds , surrounded by largely inert atomic gas .
another relatively massive dwarf , canes venatici i , also shows a small fraction of relatively young ( @xmath187 gyr ) , more metal - rich stars in addition to the predominantly old ( @xmath188 gyr ) , metal - poor population @xcite .
thus even the ultrafaint dwarfs may , in the future , reveal complex , extended star formation histories .
we have presented phenomenological models for star formation histories of dwarf galaxies in the local group , based on the mass assembly histories in cosmological simulations and the stochastic density threshold in kennicutt - schmidt law of star formation .
our main conclusions are as follows : @xmath189 models with a stochastic star formation threshold are much more successful than non - stochastic models , such as kgk04 , in reproducing the observed star formation histories of the local group dwarfs . while the kgk04 model predicted 95% of luminous dwarfs without any recent ( @xmath190 gyr ago ) star formation , our fiducial stochastic model correctly predicts that most dwarfs form a few percent of their stellar mass at late times , in agreement with the recent star formation fraction inferred for the local group dwarfs ( see table [ tab : sfdata_lg ] , @xmath4 column and fig .
[ fig : sfplots ] ) .
stochasticity allows star formation to proceed in isolated regions at late times if the threshold decreases .
@xmath189 despite significant improvements of the stochastic model , a some discrepancies with the data remain .
( 1 ) total stellar masses are typically too large by a factor of several .
this is a generic problem of both the fixed - threshold and stochastic models presented here .
( 2 ) about 10% of the dwarfs in both fixed - threshold and stochastic models have anomalously young stellar populations .
these objects build the bulk of their stellar mass in the last 10 gyr , whereas all of the observed dwarfs contain at least 15% of stars
older than 10 gyr .
these young stellar populations in the models are not created only by tidally - induced starbursts , but rather represent late mass assembly of some of the larger satellites .
@xmath189 relaxing several model assumptions does not significantly alter these predictions .
we have considered the following variants of the fiducial model : allowing late gas accretion within the virial radius of the host halo ; different slopes of the star formation law ; extended epoch of reionization ; and different prescriptions for the photoevaporation of gas from low - mass halos after reionization .
all of these variants predict statistically similar observables to the fiducial model .
@xmath189 a variant of the fiducial model that rejects dwarfs with no star formation in the first 1 or 2 gyr after the big bang significantly improves the @xmath7 stellar fraction .
even though we do not yet have an adequate justification for such a cut , this model predicts the @xmath7-distribution fully consistent with the local group data .
stellar masses , however , are still overestimated .
@xmath189 our fiducial model predicts only a modest population of dwarfs with @xmath191 , such as those recently discovered by sdss and the megacam survey . the predicted stellar mass function would be an underestimate if the observed numbers are extrapolated to all sky .
however , our mass function is still consistent with the presently known dwarfs .
our phenomenological model contains several free parameters , which allows significant freedom in the range of predicted properties of the satellite galaxies . as we discuss in [ sec : kgk ] , a combination of the parameters , @xmath66 and @xmath72 , is well constrained by the observed number of dwarfs .
other parameters of the model , apart from the stochasticity @xmath79 which strongly affects late - time star formation , lead only to small and relatively insignificant varations from the fiducial model .
this analytical prescription for star formation is applied to the mass assembly histories of the halos in the cosmological @xmath1-body simulation .
since we average over three host halos , our results should not depend significantly on a particular host - halo merger history .
also , by construction all the massive satellite halos that become galaxies in our model survive tidal disruption in the host halo , so that their total number is predicted robustly even if their mass after tidal stripping may depend on the particulars of the simulation . we should emphasize that our model is not a unique interpretation of the sfh data .
our inferences here neccesarily depend on the assumptions of a schmidt law of star formation and an exponential profile for the gas density distribution inside dwarf halos as well as the assumptions we have made with the minimum density threshold for star formation , @xmath47 . with this latter part of the model we have found that both stochasticity in this threshold over time ( [ sec : threshold ] ) and a monotonically decreasing function ( [ sec : mono_sec ] ) can lead to extended sfhs the feature missing from the kgk04 model .
we have assumed in our model that the gas clouds moving on circular orbits should generally remain at the same distance from the galaxy center .
however , during mergers and tidal interactions the angular momentum of the gas can be perturbed , leading to radial infall .
such inflow of gas may bring the central gas density above the threshold in some halos soon after the big bang and lead to more early star formation .
modeling this process would be very interesting , as it may provide a nice solution for the @xmath7 problem , but the complexity of such a process is beyond our simple model and requires a detailed hydrodynamic simulation . additionally , we have not considered any gas outflows due to the radiative and thermal feedback of young stars .
such processes could reduce the amount of available gas supply and the total stellar mass of the simulated dwarfs , which may lead to a closer agreement of the predicted and observed stellar mass function .
however , given the uncertainty in the evolution of the gas density profile , we can not conclude that such feedback is required to reconcile the stellar masses .
in fact , from figure [ fig : lowmass_projections ] we see that in very low mass halos , where feedback is expected to be stronger , our model needs a boost , rather than a reduction , of star formation .
the actual complex details of the condensation of molecular clouds in dwarf halos , which lead to the formation of stars , may turn out more important than the feedback of the stars after their formation . in spite of these uncertainties ,
the conclusion we would like to draw from our investigation is that the complex and extended sfhs of the local group dwarfs are generally consistent with the expected star formation in cold dark matter halos , and that this star formation is generally governed by the ( low ) efficiency of conversion of atomic gas into molecular clouds .
we thank andrew cole for useful comments on the lmc star formation data , and andrey kravtsov and todd thompson for helpful suggestions .
co thanks the ohio state university center for cosmology and astroparticle physics for its support .
og is supported by the nsf grant ast-0708087 . ,
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. 2008 , , submitted ( arxiv:0803.0547 ) , n. f. , coleman , m. g. , de jong , j. t. a. , rix , h .- w . ,
bell , e. f. , sand , d. j. , hill , j. m. , thompson , d. , burwitz , v. , giallongo , e. , ragazzoni , r. , diolaiti , e. , gasparo , f. , grazian , a. , pedichini , f. , & bechtold , j. 2008 , , 672 , l13 | the paucity of observed dwarf galaxies in the local group relative to the abundance of predicted dark matter halos remains one of the greatest puzzles of the @xmath0cdm paradigm . solving this puzzle
now requires not only matching the numbers of objects but also understanding the details of their star formation histories .
we present a summary of such histories derived from the hst data using the color - magnitude diagram fitting method . to reduce observational uncertainties ,
we condense the data into five cumulative parameters the fractions of stellar mass formed in the last 1 , 2 , 5 , and 10 gyr , and the mean stellar age .
we interpret the new data with a phenomenological model based on the mass assembly histories of dark matter halos and the schmidt law of star formation .
the model correctly predicts the radial distribution of the dwarfs and the fractions of stars formed in the last 5 and 10 gyr .
however , in order to be consistent with the observations , the model requires a significant amount of recent star formation in the last 2 gyr . within the framework of our model
, this prolonged star formation can be achieved by adding a stochastic variation in the density threshold of the star formation law .
the model results are not sensitive to late gas accretion , the slope of the schmidt law , or the details of cosmic reionization
. a few discrepancies still remain : our model typically predicts too large stellar masses , only a modest population of ultra - faint dwarfs , and a small number of dwarfs with anomalously young stellar populations .
nevertheless , the observed star formation histories of local group dwarfs are generally consistent the expected star formation in cold dark matter halos . |
the synthetic spin - orbit coupling ( soc ) in ultracold gas systems has attracted a lot of attention because of its importance in realizing a new tunable platform for nontrivial condensed matter phenomena @xcite .
for instance , the presence of soc is an essential part of topological insulators and superconductors @xcite and systems for quantum anomalous hall effect @xcite and topological quantum computation @xcite .
the realization of soc in ultracold atomic gases has rapidly progressed in recent years .
the equal mixing of rashba and dresselhaus soc has been realized firstly with bosons @xcite and then with fermions @xcite .
many proposals have been suggested for experimental realization of rashba - only soc @xcite .
very recently , the realization of two - dimensional soc which can be transformed into either rashba or dresselhaus soc has been reported @xcite . in attractively interacting fermi gases , spin - orbit coupling can largely affect the formation of fermion pairs in transition to superfluid ( for reviews , see , e.g. , @xcite ) .
in particular , tuning the strength of rashba soc can produce another type of crossover between a bardeen - cooper - schrieffer ( bcs ) superfluid and a bose - einstein condensate ( bec ) of tightly bound molecules . in the fermi gases without soc , the bcs - bec crossover is typically implemented by controlling interparticle scattering length throughout the feshbach resonance @xcite . in the presence of the soc , it has been suggested that a bec can be realized even at a fixed weak interaction where very strong soc can lead to the formation of the bosonic bound state called rashbon @xcite .
the bcs - bec crossover in spin - orbit - coupled fermi gases has been studied with various settings of soc , interaction , and dimensionality in the mean - field theory @xcite , also in the beyond - mean - field approaches for three @xcite and two dimensions @xcite , and very recently by using the quantum monte carlo ( qmc ) method for the ground state in two dimensions @xcite . in this paper
, we investigate a many - body correction to the estimates of superfluid transition temperature in two - dimensional fermi gases with rashba soc by employing the gorkov melik - barkhudarov ( gm ) approach @xcite .
the gm correction incorporates the induced interaction due to the second - order particle - hole processes in the fermi sea , providing a simple extension of the bcs mean - field framework . in two - component fermi gases with attractive @xmath0-wave interaction , without soc , the screening effect of the induced interaction reduces critical temperature , for instance , by a ratio of about @xmath1 in three dimensions @xcite and about @xmath2 in two dimensions @xcite .
the gm correction has been also extended to systems with mass imbalance @xcite , optical lattices @xcite , bec - bcs crossover @xcite , spin - density polarization @xcite , three - component gases @xcite , polar molecules @xcite , and spin - orbit coupling @xcite . here
we address two important properties of soc that make the calculations of the induced interaction essentially different from the conventional case without soc .
first , in the presence of soc , the spin components are mixed , forming the two helicity branches of energy dispersion .
these mixed spin states can mediate an antiscreening particle - hole polarization in addition to the conventional screening one , which is analogous to the three - component gases where the third component induces such effect @xcite .
second , the mixed - spin - state character of the fermi sea abruptly changes across the soc - driven bcs - bec crossover @xcite , which may substantially affect the medium contributions to the induced interaction .
however , these properties are not included in the earlier consideration of the induced interaction for a spin - orbit - coupled system @xcite . in the calculation of superfluid transition temperature in the weakly interacting limit , we find that the correction effect with the induced interaction exhibits a crossover behavior between the weak and strong soc regimes which is attributed to the change of the fermi sea character . at zero soc
, the transition temperature is confirmed to be reduced with the correction by a ratio of about @xmath3 . as the soc strength increases , this correction ratio decreases in the weak soc regime but then becomes insensitive to soc in the strong soc regime .
the ratio in the strong soc limit is estimated to be about @xmath4 for the mean - field transition temperature , however the correction effect turns out to be much more suppressed in the estimate of the berezinskii - kosterlitz - thouless transition temperature . in addition , we calculate the condensate fraction for comparison with the recent ground - state qmc results @xcite , showing significant improvement with the gm correction to the mean - field estimate . this paper is organized as follows . in sec .
[ sec : helicity ] , we describe the weak and strong soc regimes in association with the change in the noninteracting fermi sea character in the two - dimensional fermi gas with rashba soc . also in sec .
[ sec : helicity ] , the mean - field approximation to estimate the superfluid transition temperature is briefly reviewed . in sec .
[ sec : gm ] , we provide the detailed procedures of our induced interaction calculation and the correction to the mean - field equations . in sec .
[ sec : results ] , we present the correction to the estimates of superfluid transition temperature in the weakly interacting limit as a function of the soc strength .
the comparison with the qmc results for the condensate fraction is given in this section . in sec .
[ sec : conclusions ] , summary and conclusions are given .
the hamiltonian of the rashba spin - orbit - coupled fermi gases in two dimensions can be written as @xmath5 \\
+ g \sum_{\mathbf{k},\mathbf{k}^\prime } c^\dagger_{\mathbf{k},\uparrow } c^\dagger_{-\mathbf{k},\downarrow } c_{-\mathbf{k}^\prime,\downarrow } c_{\mathbf{k}^\prime,\uparrow}\end{gathered}\ ] ] where @xmath6 ( @xmath7 ) is the creation ( annihilation ) operator of a @xmath8-species fermion with momentum @xmath9 , and the dispersion @xmath10 for chemical potential @xmath11 .
the volume is set to be unity for simplicity .
we consider an attractive @xmath0-wave interaction between the @xmath12- and @xmath13-spin species , and thus a negative value is given to the interaction strength @xmath14 .
the rashba soc is given by the spin - off - diagonal term with @xmath15 where @xmath16 indicates the soc strength .
the noninteracting part of the hamiltonian is diagonalized in the @xmath17-helicity basis defined by the unitary transformation written as @xmath18 where @xmath19 .
the resulting noninteracting energy dispersions with soc are obtained as @xmath20 which are referred to as the @xmath21- and @xmath22-helicity branches , respectively . for the simplicity of calculations , we use the natural unit @xmath23 , and the boltzmann constant @xmath24
is set to be unity .
the particle density @xmath25 is fixed at @xmath26 throughout our calculations , which also sets the fermi momentum in the absence of soc ( @xmath27 ) to be unity as @xmath28 .
this is equivalent to setting the unit of momentum and energy to be @xmath29 and @xmath30 . the soc strength @xmath16 is accordingly expressed in the unit of @xmath31 .
( @xmath22 ) helicity branch of the energy dispersion .
the horizontal dashed lines indicate the fermi energies for the soc strengths @xmath32 ( a ) and @xmath33 ( b ) that belong to the weak and strong soc regimes , respectively.,scaledwidth=45.0% ] at a given particle density , the character of the noninteracting fermi sea abruptly changes as the soc strength @xmath16 increases from a small to large value as illustrated in fig . [
fig : fs ] . at a small @xmath16 , both of the @xmath21- and @xmath22-helicity branches can be occupied , and thus the fermi sea is composed of two different helicity branches .
in contrast , at a large @xmath16 the fermi sea transforms into a doughnutlike shape with two surfaces of the same @xmath21-helicity branch . for @xmath28
, one can easily verify that this transition occurs at @xmath34 .
for @xmath35 , the radii of the @xmath21- and @xmath22-helicity fermi seas are @xmath36 on the other hand , for @xmath37 , the fermi sea is surrounded by the two surfaces with radii @xmath38 and @xmath39 which are obtained as @xmath40 in order to distinguish these two different characters of the noninteracting fermi sea , we refer to the soc strengths of @xmath41 and @xmath42 as the weak and strong soc regimes , respectively .
the gm correction procedures are based on a perturbative approach that preserves the structure of the mean - field theory .
the many - body effect induced by the fermi sea is considered as an effective interaction . with the superfluid order parameter @xmath43 ,
the mean - field hamiltonian is written as @xmath44 where @xmath45 , and the matrix @xmath46 diagonalizing @xmath47 , one can obtain the quasiparticle energies as @xmath48 , and accordingly one can write down the thermodynamic potential as @xmath49 - \frac{|\delta|^2}{g}.\ ] ] from the saddle - point condition @xmath50 , the mean - field gap equation is derived as @xmath51 = 0,\ ] ] where @xmath52 . note that the bare interaction strength @xmath14 is replaced by the two - body binding energy @xmath53 through the relation for a two - dimensional gas @xmath54 the number equation @xmath55 is also written as @xmath56.\ ] ] for a given particle density , the superfluid transition temperature is determined by self - consistently solving the gap and number equations for the vanishing order parameter .
the modification by the gm correction is to be brought only into the mean - field gap equation by replacing the bare interaction @xmath14 with the effective interaction @xmath57 that includes the induced interaction correction .
- and @xmath13-spin components . the solid lines are propagators , and the single dotted line indicates the bare two - body interaction .
the diagram in ( a ) is conventional in usual two - component systems , and the one in ( b ) is mediated by the spin - component mixing due to soc .
( c ) the schematic diagram of the effective interaction ( double dotted line ) with the induced interaction correction.,scaledwidth=48.0% ] figure [ fig : diagrams ] illustrates the two relevant second - order diagrams of the induced interaction in the presence of soc . while diagram @xmath58 shown in fig .
[ fig : diagrams](a ) is typical in a two - component system with an @xmath0-wave interaction , diagram @xmath59 shown in fig .
[ fig : diagrams](b ) is allowed by the spin - offdiagonal propagators available in the mixed spin states due to the soc .
the particle - hole polarizations presented in diagrams @xmath58 and @xmath59 contribute to the induced interaction correction . in the gm correction scheme , the effective interaction @xmath60 is written with the correction up to the second - order of interaction as @xmath61 where the induced interaction @xmath62 is to be averaged over the fermi surfaces that are expected to dominantly contribute to the fermion pairing .
the contributing polarization functions @xmath63 and @xmath64 of diagrams 1 and 2 are expressed by using propagators in an imaginary - frequency domain as @xmath65 where the four - momentum @xmath66 is fermionic , and the noninteracting matsubara green s function @xmath67 .
the spin - diagonal and offdiagonal parts of @xmath68 can be evaluated through the helicity basis transformation as @xmath69 , \\
\mathcal{g}^{(0)}_{\sigma\bar{\sigma } } ( \tilde{k } ) & = & \frac{1}{2 } e^{s_{\sigma\bar{\sigma } } i\varphi_\mathbf{k } } \left [ \mathscr{g}_{0\uparrow } ( \tilde{k } ) - \mathscr{g}_{0\downarrow } ( \tilde{k } ) \right],\end{aligned}\ ] ] where @xmath70 , and @xmath71 for @xmath72 .
then , @xmath63 and @xmath64 can be decomposed with the inter- and intra - helicity - branch components as @xmath73_{ab } \label{eq : pi}\ ] ] where @xmath74_{ab } & = & \frac{1}{4 } \int\frac{d^2 q}{(2\pi)^2 } \chi_{ab } ( \mathbf{p}+\mathbf{k},\mathbf{q } ) , \label{eq : pi1 } \\ \left [ \pi_2 ( \mathbf{p},\mathbf{k } ) \right]_{ab } & = & -\frac{c_{ab}}{4 } \int\frac{d^2 q}{(2\pi)^2 } e^{i\delta\varphi } \chi_{ab } ( \mathbf{p}-\mathbf{k},\mathbf{q } ) \label{eq : pi2}.\end{aligned}\ ] ] the coefficient @xmath75 is @xmath76 for @xmath77 or @xmath78 for @xmath79 , and @xmath80 .
the function @xmath81 is calculated by performing imaginary - frequency summation as @xmath82 where @xmath83 , and the bosonic frequency @xmath84 is taken to be zero in the low temperature limit in the evaluation of the effective interaction .
our formulation of the induced interaction extends the previous approach @xcite where only the interbranch @xmath85 was considered within diagram @xmath58 .
later , we will show that the contribution of @xmath85 is rather small at strong soc . in the standard procedures of the gm correction ,
the induced interaction is evaluated as @xmath86 being averaged over the fermi surface momenta @xcite , which can be justified in the weakly interacting limit where the scattering processes for pairing would dominantly occur near the fermi surface .
indeed , the recent qmc calculations for two - dimensional fermi gases with rashba soc showed that the condensate wave functions are peaked at the two fermi surfaces in the weak and strong soc regimes @xcite .
therefore , we evaluate the induced interaction as @xmath87 \nonumber \\ & = & -g^2 \left [ \left\langle \pi_1 ( \mathbf{p},\mathbf{k } ) \right\rangle_{\mathrm{fs } } + \left\langle \pi_2 ( \mathbf{p},\mathbf{k } ) \right\rangle_{\mathrm{fs } } \right],\end{aligned}\ ] ] where @xmath88 , and @xmath89 denotes a set of the fermi surface momenta .
we treat the two fermi surfaces on an equal footing for averaging , and thus @xmath89 includes all momenta residing on the two surfaces composed of the one with @xmath90 ( @xmath91 ) and the other with @xmath92 ( @xmath93 ) in the weak ( strong ) soc regime .
all momentum integrations are done numerically for the evaluation of the induced interaction .
the gm correction to the mean - field gap equation is then readily done by replacing bare interaction @xmath14 with effective interaction @xmath60 . in the weakly interacting limit
, the inverse of the effective interaction becomes @xmath94 which rewrites the mean - field gap equation ( [ eq : gap_bcs ] ) as @xmath95 = \overline{\pi}_1 + \overline{\pi}_2 .
\label{eq : gap_gm}\ ] ] the applicability of the gm - corrected gap equation mainly depends on the interaction strength since the induced interaction is basically a second - order perturbation expansion for a weak interaction .
in addition , we have assumed that all intra- and inter - surface pairs of the fermi momenta equally participate in the fermi surface average for the induced interaction .
these approximations may need to be examined in more rigorous approaches .
we will examine the improvement with the correction by comparing with the recent ground - state qmc results for the condensate fraction @xcite in the following section .
figure [ fig : pi ] presents the induced interaction calculated from diagrams @xmath58 and @xmath59 as a function of the soc strength @xmath16 .
the total induced interaction @xmath96 is positive for all @xmath16 and thus reduces the bare attraction @xmath14 in the effective interaction @xmath97 .
this would lower the estimation of superfluid transition temperature .
however , we find that the induced interaction and its components exhibit a nontrivial soc - strength dependence .
this appears as a crossover from the weak to strong soc regime that is attributed to the abrupt change in the mixed - spin - state character of the noninteracting fermi sea . of two - dimensional fermi gases without soc .
( a ) total induced interaction @xmath98 , and the screening and antiscreening contributions @xmath99 and @xmath100 from diagrams @xmath58 and @xmath59 given in fig .
[ fig : diagrams ] , respectively . the inter- and intra - helicity - branch components @xmath101 are presented for diagrams @xmath58 ( b ) and @xmath59 ( c ) .
the vertical dotted lines indicate @xmath34 where the character of the fermi sea abruptly changes.,scaledwidth=48.0% ] in the weak soc regime , the induced interaction is mainly determined by the screening effect of @xmath99 from the contribution of diagram @xmath58 . in the limit of @xmath102 , all components of the polarization function , @xmath81 given in eq .
( [ eq : chi ] ) , become identical , leading to the complete cancellation in @xmath64 , while in turn @xmath63 recovers the case without soc .
the negative sign of @xmath99 can be easily understood from the form of @xmath81 which is similar to the conventional two - component systems without soc .
as @xmath16 increases , @xmath100 becomes finite with the opposite sign indicating an antiscreening effect , although it remains much smaller in magnitude than @xmath99 for small @xmath16 .
on the other hand , in the strong soc regime , the antiscreening contribution of @xmath100 is no longer small but considerably weakens the screening effect of @xmath99 in the total induced interaction . at @xmath103 ,
the noninteracting fermi sea is available only with the @xmath21-helicity branch for particles to reside , which leads to the dominant contribution of the @xmath104 component in both @xmath63 and @xmath64 while the other components are largely suppressed .
a kink was found at @xmath105 where the @xmath22-helicity branch disappears in the fermi sea .
the induced interaction is determined by a net effect of the opposite contributions of @xmath63 and @xmath64 , and their proper consideration becomes very important for the evaluation of the gm correction in spin - orbit - coupled systems . ) and
with the correction ( @xmath106 ) is compared between the weak and strong soc regimes at ( a ) @xmath107 and ( b ) @xmath108 , respectively.,scaledwidth=48.0% ] we calculate the correction to the estimates of superfluid transition temperature in the weakly interacting limit .
first , we present the effects of the induced interaction correction on the mean - field calculations which can be more relevant in a quasi - two - dimensional system supporting the long - range order . additionally , for a strictly two - dimensional system
where the long - range order is absent , we provide a rough estimate of the berezinskii - kosterlitz - thouless transition temperature under the assumption that the induced interaction effect is limited to the mean - field gap amplitude .
the influence of the induced interaction correction varies with the strength of soc , which we discuss below in terms of the ratio between the estimates of the transition temperature without and with the correction .
we first examine the interaction dependence of the gm correction to the superfluid transition temperature in the weak and strong soc regimes .
figure [ fig : tc ] shows the comparison between the mean - field transition temperature estimates @xmath109 without the correction and @xmath106 with the correction and the reduction ratio @xmath110 as a function of @xmath53 . in both of the values of @xmath16 that we have examined
, the reduction ratio goes below the known zero - soc value of about @xmath3 in the weakly interacting limit and also shows the weaker influence of the correction at the stronger soc .
the variation of the correction effect with the soc strength is more systematically shown in figs .
[ fig : main](a ) and [ fig : main](b ) at the fixed values of @xmath111 and @xmath112 that are close to the weakly interacting limit .
we find that the reduction ratio @xmath110 first decreases monotonically from the zero - soc value of about @xmath3 as @xmath16 increases within the weak soc regime , but then the ratio becomes insensitive to soc when @xmath16 goes into the strong soc regime .
a kink appears at @xmath34 where the noninteracting fermi sea abruptly changes . while the effect of the induced interaction decreases with increasing soc in terms of the reduction ratio
, it turns out that the ratio at a large @xmath16 approaches a constant value about @xmath4 which still gives a sizable reduction to @xmath109 . on the other hand , in strictly two dimensions , the true long - ranger order is absent because of strong quantum fluctuations .
although , it is well known that the superfluid transition at finite temperature is still possible by the berezinskii - kosterlitz - thouless ( bkt ) mechanism of the vortex - antivortex pairing @xcite . in the spin - orbit - coupled fermi gases , the bkt transition temperature and superfluid properties have been estimated with consideration of the phase fluctuations in the order parameter beyond the mean - field approximation @xcite . regarding the gm correction , the following question naturally arises for the bkt transition temperature . at weak interactions
, the previous result @xcite shows that the bkt transition temperature recovers the mean - field estimate when the rashba soc strength is small , while in the strong soc limit , the bkt temperature largely deviates from the mean - field estimate .
however , we find that for instance , at a very small @xmath16 , the estimate of the mean - field transition temperature with the gm correction is reduced by a factor of about @xmath3 , and therefore it happens that the mean - field estimate with the correction actually goes much below the bkt temperature if estimated without such correction . here
we suggest that in the weakly interacting limit , the gm correction may be applicable to the estimate of the bkt temperature by restricting the influence of the induced interaction within the mean - field or saddle - point gap equation .
we follow the description of the phase fluctuations given in ref . @xcite . in the procedures ,
when the amplitude fluctuations of the order parameter is neglected , the amplitude is still determined from the same mean - field gap equation .
thus , as a rough approximation in the weakly interacting limit , by ignoring any effects on the anomalous propagators , one may include the correction in the same way as for the mean - field calculations with the gm - corrected gap equation .
the bkt transition temperature ( @xmath113 ) can be determined through the universal nelson - kosterlitz ( nk ) relation @xcite which can be written as @xmath114 in our natural unit .
the superfluid density @xmath115 is evaluated as @xmath116 @xcite as a function of temperature @xmath117 , where @xmath118 and @xmath119 in our unit are given as @xmath120}{e_{k , s } } , \nonumber \\
\rho_2(t ) & = & \frac{\beta}{8\pi } \sum_{s=\pm }
\int_0^\infty k dk \left ( k + s \frac{v_r}{2 } \right)^2 \mathrm{sech}^{2 } [ \frac{\beta e_{k , s}}{2 } ] , \nonumber\end{aligned}\ ] ] for @xmath121 . with the gm - corrected gap equation employed for @xmath122
, one can find the corresponding correction in the estimate of the bkt transition temperature @xmath123 by solving the nk relation .
the correction in the bkt transition temperature is shown in figs .
[ fig : main](c ) and [ fig : main](d ) for the same choices of @xmath53 . with the correction
, it turns out that @xmath124 becomes very close to the mean - field estimate @xmath106 at small @xmath16 , recovering the expectation of ref .
@xcite . in the reduction ratio between @xmath125 , the one without the correction , and @xmath124 , the behavior of its soc dependence exhibits the similar crossover observed in the mean - field result :
the ratio decreases with @xmath16 at first and then becomes insensitive to @xmath16 in the strong soc regime .
interestingly , the asymptotic value of the ratio approaches unity at a large @xmath16 , implying that the effect of the induced interaction on the bkt transition temperature may diminish in the rashbon limit . finally , we demonstrate the applicability of the gm - corrected gap equation by the comparison with the recent ground - state qmc calculations @xcite for the condensate fraction . in the mean - field approximation
@xcite , the condensed density of fermion pairs @xmath126 is expressed as @xmath127 where the spin - singlet pairing field @xmath128 and the soc - induced spin - triplet pairing field @xmath129 at @xmath130 .
the total condensate fraction @xmath131 is then written as @xmath132 this only depends on the zero - temperature order parameter which is obtained from eq .
( [ eq : gap_gm ] ) when the gm correction is included . in figs .
[ fig : main](e ) and [ fig : main](f ) , we provide the condensate fraction calculated at the two chosen values of the soc strength for direct comparison with the qmc data in the weak and strong soc regimes .
the total condensate fractions indicate good agreement , showing a significant improvement with the gm correction to the mean - field estimate . the agreement is better in the stronger soc even at relatively strong interactions , which is notable since the gm correction is generally expected to be most reliable at weak interactions . in the spin - triplet component
, the perturbative nature of the gm correction indeed becomes apparent at strong interactions where the qmc data deviate from the gm - corrected one .
such systematic deviation with increasing interaction is not seen in the total condensate fraction at the stronger soc , implying that the underestimation of the singlet component compensates the difference .
we have investigated the gorkov melik - barkhudarov correction to superfluid transition temperature in attractively interacting fermi gases with rashba soc in two dimensions .
the main differences due to the presence of soc are summarized as follows .
first , the spin - component mixing allows another second - order process that additionally leads to the antiscreening effect that is prohibited in the usual two - component case without soc .
the induced interaction is thus determined by a net effect of the antiscreening and conventional screening contributions .
second , the induced interaction and corresponding correction to the superfluid transition temperature largely depends on the soc strength , which is attributed to the abrupt change of the fermi sea character across the soc - driven bcs - bec crossover . in the estimate of the mean - field transition temperature in the weakly interacting limit , the correction due to the induced interaction shows a crossover behavior across the weak and strong soc regimes . at strong soc
, it is found that the mean - field temperature is reduced by a ratio of about @xmath4 with the gm correction which is in contrast to the known value of about @xmath3 for the case without soc .
this ratio decreases with increasing soc in the weak soc regime , but then it becomes largely insensitive to soc in the strong soc regime .
the similar crossover is also found in our estimation of the bkt transition temperature at weak interactions .
however , our calculation indicates that the correction becomes quantitatively very small in the strong soc limit in the case of the bkt transition temperature .
the essence of the gm correction procedures is in the modification of the mean - field gap equation with the induced interaction . our comparison with the recent ground - state qmc result for the condensate fraction @xcite shows remarkable improvement to the use of the usual mean - field order parameter , demonstrating the importance of the gm correction in a spin - orbit - coupled system .
while exact numerics with soc for the transition temperature have not been available yet , our estimates with the gm correction would motivate applications of other beyond - mean - field approaches , for instance , with the functional renormalizaton group @xcite , particle - hole fluctuations with self - energy renormalization @xcite , non - gaussian fluctuation theory @xcite , and @xmath133-matrix approaches @xcite , to two - dimensional fermi gases with soc .
in addition , we expect that it is also possible to extend our formulation of the gm correction to systems in three dimensions or with other types of soc , which would provide more improved mean - field - based predictions of experimental relevance .
we thank peter rosenberg and shiwei zhang for providing their quantum monte carlo data and jildou baarsma for valuable discussions .
this work was supported from basic science research program through the national research foundation of korea funded by the ministry of science , ict & future planning ( grant no .
nrf-2014r1a1a1002682 ) . | we investigate the gorkov melik - barkhudarov ( gm ) correction to superfluid transition temperature in two - dimensional fermi gases with rashba spin - orbit coupling ( soc ) across the soc - driven bcs - bec crossover . in the calculation of the induced interaction
, we find that the spin - component mixing due to soc can induce both of the conventional screening and additional antiscreening contributions that interplay significantly in the strong soc regime .
while the gm correction generally lowers the estimate of transition temperature , it turns out that at a fixed weak interaction , the correction effect exhibits a crossover behavior where the ratio between the estimates without and with the correction first decreases with soc and then becomes insensitive to soc when it goes into the strong soc regime .
we demonstrate the applicability of the gm correction by comparing the zero - temperature condensate fraction with the recent quantum monte carlo results . |
we would like to thank m. prakash , m. rho and c.m .
ko for helpful discussions .
we also appreciate r. machleidt for sending us the code of the relativistic bonn potential and helpful discussions .
this work was supported by the u.s .
department of energy under grant no .
de - fg02 - 88er40388 .
the work of chl was supported in part by korea science and engineering foundation . | the momentum dependence of the empirical scalar and vector potentials needed for describing relativistic heavy - ion collisions at 1 gev / nucleon is compared with that derived from self - consistent dirac - brueckner calculations using the bonn - a potential .
our calculated scalar and vector potentials exhibit a weak momentum dependence and their magnitudes decrease with nucleon momentum , in close similarity with the momentum dependence required by the empirical potentials .
the effects of explicit momentum dependence on the properties of equilibrium nuclear matter are found to be small .
heavy - ion collisions at various energies open new frontiers and provide opportunities for nuclear physics
. theoretical description of heavy - ion collisions is usually based on transport models . for heavy - ion collisions at beam energies from a few hundred
mev / nucleon to a few gev / nucleon , the reaction dynamics is mainly governed by the stochastic two - body scattering and propagation in the mean - field potential .
transport models that include both effects , such as boltzmann - uehling - uhlenbeck ( buu ) @xcite , quantum molecular dynamics ( qmd ) @xcite , and the relativistic extensions , rbuu @xcite and rqmd @xcite , have been extensively used for the description of heavy - ion collisions in this energy region . in the buu or qmd calculations , the nucleon potential is usually parameterized in terms of skyrme - type effective interaction @xcite .
the importance of the momentum dependence of the nucleon potential was realized when analysing nucleon flow and pion production in heavy - ion collisions @xcite .
this has been incorporated into the buu and qmd approaches based on various phenomenological parameterizations that are constrained by the empirical information from proton - nucleus scattering experiments @xcite . on the other hand ,
the relativistic transport models @xcite are usually based on the walecka - type effective lagrangians @xcite . in this model , nucleons feel both attractive scalar and repulsive vector potentials .
the corresponding schrdinger equivalent potential , that is to be compared with the potentials used in buu and qmd calculations , is automatically and explicitly momentum dependent . in the mean field approximation both
the scalar and vector potentials are momentum independent .
the schrdinger equivalent potential then increases linearly with nucleon kinetic energy , which is in disagreement with what has been learned from the dirac phenomenology analysis of proton - nucleus scattering at intermediate energies @xcite . in a recent paper
, we have shown that the explicit momentum dependence in nucleon scalar and vector potential is important for the description of nucleon flow in heavy - ion collisions at 1 gev / nucleon region @xcite . in that work ,
the momentum dependence correction was introduced by scaling the momentum independent mean field potential with a momentum dependent function extracted from the dirac phenomenology .
there have been similar attempts to reproduce the correct momentum dependence of the schrdinger equivalent potential by including explicit momentum dependence in the nucleon scalar and vector potentials @xcite .
a main purpose of this paper is to study microscopically the origin of the momentum dependence in the scalar and vector potentials used in relativistic heavy - ion collisions @xcite .
the self - consitent dirac brueckner ( dbhf ) approach @xcite is employed .
it is well - known that the dbhf approach reproduces nicely the nuclear matter properties starting from realistic nucleon - nucleon interactions that are constrained by the two - nucleon data . in most dbhf calculations
the scalar and vector potentials are assumed to be momentum independent .
this is a good approximation for equilibrium nucleon matter , where the nucleon momenta involved are not very large .
the momentum dependence was studied by ter haar and malfliet @xcite , and they found a relatively weak momentum dependence for both the real and imaginary parts of the scalar and vector potentials .
recently , sehn _ et al . _
@xcite studied the momentum dependence of the nucleon scalar and vector potentials in the first - order approximation .
they first carried out the momentum independent dbhf calculation as in ref .
@xcite .
the scalar and vector potential so obtained were identified with a nucleon at fermi surface , namely @xmath0 .
the potentials for momenta below and above @xmath1 were then calculated using the formalism of horowitz and serot @xcite .
the self - consistency in the dbhf calculation has thus been neglected in ref .
@xcite .
we will compare our results to those of refs .
@xcite . in this work ,
we follow essentially the dbhf formalism of ref .
@xcite .
the basic quantity is the @xmath2 matrix that satisfies the in - medium thompson equation , @xmath3 where @xmath4 , @xmath5 is c.m .
momentum , @xmath6 and @xmath7 .
the single - particle potential can be calculated once the @xmath2 matrix is known , @xmath8 this potential is usually parameterized in terms of scalar and vector potential , @xmath9 in ref .
@xcite , both the scalar and vector potentials are assumed to be momentum independent .
thus for a given density , these potentials can be determined by choosing two momenta in eq .
( 3 ) , usually @xmath10 and @xmath1 . in the momentum
dependent case as we consider here , we have two unknowns ( @xmath11 and @xmath12 ) for each given momentum , which can not be uniquely determined without introducing further approximation .
the customary way to do this is to assume some functional dependence on momentum for these potentials .
for example , in ref .
@xcite , it was assumed that @xmath13 . since the real part of the single particle potential vanishes at very high momentum , we expect the following parameterization to be more appropriate @xmath14 actually , in ref .
@xcite , this kind of functional dependence was found to fit the dirac phenomenology rather well .
one of the purposes of this work is thus to study whether this kind of simple parameterization is supported by microscopic dbhf calculation .
numerically we find that it is rather difficult to determine the six parameters altogether in the minimization procedure .
we have checked the variation of @xmath15 around 0.5 , and find that @xmath16 provides a smooth self - consistent procedure , and the final results are not very sensitive to some small change in @xmath15 .
the remaining four parameters are then determined by a minimization procedure , @xmath17 our self - consistent procedure is the following . for a given density as specified by fermi momentum @xmath1 , we choose reasonable initial values for @xmath18 and @xmath19 .
these , together with eq .
( 4 ) provide @xmath20 for all momenta .
these potentials are then used as inputs in eq .
( 1 ) to determine the @xmath2 matrix , from which we obtain the single - particle potential @xmath21 using eq . ( 2 ) . our new set of @xmath18 and @xmath19 can then be determined by minimizing eq .
( [ eqchisq ] ) against @xmath21 .
this procedure is continued until desired self - consistency is achieved .
the quality of our self - consistency procedure can be seen in fig .
[ figmsp ] where we compare the single - particle potential from the dbhf calculation ( solid symbols ) and from eqs .
( 3 ) and ( 4 ) .
there are slight deviations near @xmath22 because in our definition of @xmath23 , eq . ( 5 ) , the contributions from small momenta are suppressed .
the self consistent parameters are summarized in tab .
[ tab1 ] . in fig .
[ figusv ] we show our results for the momentum dependence of the scalar and vector potentials at normal nuclear matter density .
both are seen to decrease in magnitude as nucleon momentum increases .
also shown in the figure are the results of ter haar and malfliet ( hm ) @xcite and those of sehn _ et al . _
( sff ) @xcite .
our results are very close to those of ter haar and malfliet , which supports our parametrization functional , eq .
( [ usvpar ] ) .
the results of sehn _ et al . _
show a much stronger momentum dependence than ours and those of ter haar and malfliet @xcite , because of the lack of the self - consistency .
we show in fig .
[ figusv2 ] the momentum dependence of the nucleon scalar and vector potentials at several different densities . from the scalar and vector potential we can calculated the schrdinger equivalent
potential given by @xmath24 where @xmath25 is the nucleon kinetic energy .
the results are shown in fig .
[ figsch ] .
the solid symbols are the schrdinger equivalent potential extracted from the dirac phenomenology analysis @xcite .
the solid and dashed curves are obtained with and without explicit momentum dependence in the scalar and vector potentials .
the dotted is from ref .
@xcite .
it is seen that the explicit momentum dependence in the scalar and vector potentials improves , to some extent , the agreement with the dirac phenomenology .
the schrdinger equivalent potential from our self - consistent calculation still show stronger kinetic energy dependence than that of the dirac phenomenology .
the onset of inelastic processes at high beam energies might explain part of this discrepancy .
it is also of interest to see how the nuclear equation of state are affected by the explicit momentum dependence .
@xmath26 in fig.[figbe ] , the binding energies are plotted as a function of density .
the solid and dashed curves are obtained with and without explicit momentum dependence in the scalar and vector potentials .
it is seen that the effects of the explicit momentum dependence on equilibriu nuclear matter is insignificant . the saturation density in both cases
are about @xmath27 .
the binding energy decreases by about 0.2 mev when explicit momentum dependence is included . in summary
, we studied the momentum dependence of nuclear scalar and vector potentials self - consistently in the formalism of dirac brueckner approach .
the magnitudes of both potentials have been found to decrease slowly with increasing nucleon momentum or kinetic energy .
this improves , to some extent , the agreement with the dirac phenomenology .
the effects of explicit momentum dependence on the properties of equilibrium nuclear matter were found to be small . |
intersubband polaritons are elementary excitations of microcavities embedding doped semiconductor quantum wells . in particular , they are the normal modes due to the strong coupling between a cavity photon mode and the transition between two conduction subbands in quantum wells containing a two - dimensional electron gas .
intersubband polaritons were experimentally demonstrated for the first time in 2003@xcite and their physics is the object of many interesting investigations , both fundamental @xcite and applied @xcite .
so far , theoretical investigations have been based mostly on a bosonized hamiltonian approach @xcite . in the dilute regime ,
i.e. , when the excitation density is much smaller than the electronic density in the doped quantum wells , intersubband polaritons behave as ( composite ) bosons . at higher excitation densities anyway , pauli saturation effects due to electron fermionic statistics and deviations from the bosonic behavior are expected @xcite .
an important aspect of the physics of such excitations is the role of the coulomb interaction . in the case of bare intersubband transitions ,
coulomb interaction is known not to modify dramatically the polaritonic lineshape @xcite , being responsible mainly for a simple shift of the excitation frequency , the so - called depolarization shift .
this is why , in many relevant situations , the presence of the coulomb interaction can be in first approximation neglected , by using a renormalized ( measured ) value for the frequency of the intersubband transition instead of the bare one , calculated from the single - electron subband structure .
very recently , the role of coulomb interaction in intersubband polariton systems has started to attract interest . in ref .
[ ] the part of the coulomb interaction responsible for the depolarization shift has been studied in the bosonic excitation formalism , allowing to calculate the renormalization from first principles .
other works have studied the same problem using different gauges @xcite and a green functions approach @xcite . as a matter of fact , up to now , only depolarization or static effects have been taken into account .
hence , the general problem of the many - body physics of intersubband polaritons is largely unexplored , in particular the physics of polariton - polariton interactions and polariton - polariton scattering . in this paper
we derive a microscopic many - body theory describing the intersubband polariton - polariton scattering due both to the coulomb interaction and to the non - perfect bosonicity of the intersubband excitations . as a first application of the present theoretical formalism
, we calculate how the depolarization shifts are modified and , most importantly , we determine the efficiency of the intersubband polariton - polariton scattering processes .
this article is organized as follows . in sec .
[ sec : general ] , we present the physical system , introduce the fermionic electron - photon hamiltonian including the different channels of the ( screened ) coulomb interaction . moreover , we write the intersubband excitations in second quantized formalism . in sec .
[ coboson ] , the composite boson commutator approach @xcite to calculate many - body matrix elements is presented for the case of intersubband excitations . in sec . [
effective ] , an effective bosonic hamiltonian ( up to the fourth order in the bosonic operators ) is derived in such a way that the two - excitation properties are the same as those predicted by the microscopic composite boson approach . in sec .
[ numerical ] numerical applications of the present theory to the calculation of the interaction energy between intersubband excitations are presented .
[ scattering ] is devoted to the calculation of polariton - polariton scattering processes .
finally , conclusions are drawn in sec . [ conclusions ] .
additional technical details are included in appendix [ sec : notation ] , [ sec : coeff ] and [ sec : fermi ] .
we will consider a photonic , planar microcavity embedding @xmath0 symmetric quantum wells , each doped with an areal electron density of @xmath1 ( see fig . [ sketch ] for a schematic representation of such a system ) .
as the quantum wells confine the electrons along the growth direction , the electronic bands will be split into multiple , almost parallel subbands .
the optically active transitions we are interested in will thus be the ones from the last occupied conduction subband to the first unoccupied one .
the photonic microcavity width will be chosen to have one of the photonic modes almost resonant with the electronic transition between these two electronic conduction subbands ( the first two for simplicity , even if more general situations are possible @xcite ) . in order to simplify the notation
, we will consider that the quantum wells are all identically coupled to the relevant microcavity field .
schematic representation of the system under investigation .
panel ( a ) : multiple doped quantum wells ( qws ) are embedded in a planar photonic microcavity .
the image above shows also the fundamental tm photonic mode in the case of metallic mirrors .
due to intersubband selection rule , the intersubband transitions only couple to the @xmath2-component of the electric field .
panel ( b ) : energy profile of the quantum well along the @xmath2-direction and its two bound states ( left ) .
electronic dispersion of the first two conduction subbands in each of the doped quantum wells , as a function of the in - plane electronic wavevector ( right ) .
the fermi wavevector , @xmath3 , the fermi energy , @xmath4 , and the intersubband gap , @xmath5 , are highlighted.,width=332 ] the hamiltonian describing the coupled electron - photon system can be written in the coulomb gauge as @xmath6 where @xmath7 includes only photonic degrees of freedom , while @xmath8 regroups all the terms presenting also electronic ones . introducing the annihilation operators @xmath9 for an electron in the subband @xmath10 with an in - plane wavevector @xmath11 , and @xmath12 for a microcavity photon with an in - plane wavevector @xmath13 ,
we can write the two hamiltonians explicitly as @xmath14 where @xmath15 are respectively the electronic kinetic energy , the light - matter coupling hamiltonian and the electron - electron coulomb interaction . coefficients @xmath16 and @xmath17 are the frequencies of a photon with in - plane wavevector @xmath13 and of an electron with in - plane wavevector @xmath11 in the subband @xmath10 , @xmath18 is the light - matter coupling frequency , @xmath19 the diamagnetic interaction coefficient and the @xmath20 term has to be excluded from the sum in @xmath21 in order to compensate for the homogeneous positive background of the jellium model @xcite . in order to simplify the notation both spin and quantum well indexes
have been omitted but one should keep in mind that there is an implicit sum over them .
the explicit notation is given in appendix [ sec : notation ] .
we also neglect to explicitly write the photon polarization because , due to the intersubband selection rule , only the tm modes are coupled with matter .
the coulomb matrix elements @xmath22 are given by @xcite @xmath23 where @xmath24 is the surface of the sample and @xmath25 is the real @xmath2-component wavefunction of an electron in the subband @xmath10 . due to the symmetry of the quantum wells , and thus of the wavefunctions , a certain number of matrix elements in eq .
( [ v ] ) can be seen to be zero , in particular all matrix elements with an odd number of @xmath26 and @xmath27 indices @xmath28 the other elements obey the relations @xmath29 so that there are only four distinct matrix elements , @xmath30,@xmath31,@xmath32 and @xmath33 . in fig . [
fig : coulomb ] we show a graphical representation of these four qualitatively different processes and the wavevector dependency of their respective coefficients for an infinite potential well . from such a figure
we can clearly see a qualitative difference in the long wavelength behavior between @xmath30 and the other three coefficients . as explained in greater detail in ref .
[ ] , they indeed refer to qualitatively different physical processes . while @xmath30 describes an intersubband , coherent coulomb scattering , that is responsible for the so - called depolarization shift , the others describe intrasubband scattering processes , in which electrons are scattered inside the same subband .
[ bc]@xmath34 [ bc]@xmath35 ( a.u . ) [ bc]@xmath36 ( a.u . ) [ bc]@xmath37 ( a.u . ) [ bc]@xmath38 ( a.u . ) [ cols= " > , > , > , > " , ] the dense two dimensional electron gas in the first electronic subband screens the coulomb interaction . in order to take it into account
we will thus replace the bare coulomb interaction @xmath39 defined by eq .
( [ v ] ) , with its static rpa - screened version @xmath40 . from refs .
[ ] , these coefficients obey the dyson equation @xmath41 where @xmath42 is the rpa polarization function for the @xmath43 transition .
the screened interactions thus take the form @xmath44},\nonumber \label{eq : screen22}\end{aligned}\ ] ] where @xmath45 .
screening of intrasubband and intersubband processes can thus be encoded respectively in the dielectric functions @xmath46 and @xmath47 , where @xmath48 and @xmath49 .
\label{eq : dielecdepol}\end{aligned}\ ] ] as we will see in the rest of this section , only the long - wavelength limit is relevant in the study of intersubband polaritons .
we will thus always consider the @xmath50 limit of the above experessions . in fig .
[ fig : coulomb ] the reader can find the screened and unscreened profiles of the different potentials as a function of the exchanged momentum normalized over the fermi wavevevector @xmath51 .
the theory in this paper will be developed using a zero temperature ( @xmath52 ) formalism . in order for this approximation to be valid
, we need to have a vanishing thermal electronic population in the excited subband , that is @xmath53 , where @xmath4 is the fermi energy and @xmath54 the boltzmann constant .
under such a condition the ground state of the uncoupled light - matter system is given by @xmath55 where @xmath3 is the fermi wavevector and @xmath56 is the vacuum state , with no electrons in both subbands and no photons in the microcavity . in the ground state
all electronic states up to the fermi energy are occupied and all the others are empty . if we neglect the incoherent intrasubband coulomb processes ( due to the coefficients @xmath57 , @xmath58 and @xmath59 ) in eq .
( [ eq : hfdetail ] ) , the full hamiltonian @xmath60 can be solved , in the dilute regime , in terms of bosonic coherent excitations named intersubband polaritons@xcite . in order to do this , we first notice that the microcavity photon mode is coupled through @xmath61 to only one particular linear superposition of electronic transitions @xmath62 where @xmath63 and @xmath64 is the heaviside function .
the choice of the notation with the index @xmath65 will be explained in the following . in eq .
( [ eq : isb0 ] ) , as well as in eq .
( [ eq : hfdetail ] ) , the sum is implicit over the spin and quantum well indexes to simplify notations .
this defines an intersubband collective excitation , which we will name bright intersubband excitation .
hamiltonian @xmath61 can thus be written exactly @xmath66 where @xmath67 is the rabi frequency .
( [ eq : hrabi ] ) makes it clear that bright excitations concentrate all the oscillator strength of the electron gas .
other similar collective excitations can be constructed by an orthonormalization procedure @xmath68 where index @xmath69 runs from @xmath26 to @xmath70 and the @xmath71 coefficients have support over the fermi sea and satisfy the orthonormality relation @xmath72 however , none of these new collective excitations is coupled to the microcavity photon field .
this is why we call them dark intersubband excitations .
we will consider them in sec .
[ coboson ] .
we then use the fact that the two electronic subbands are parallels , and we neglect the photonic exchanged momentum when compared to an electronic one , that is @xmath73 we can thus rewrite the hamiltonian as @xcite @xmath74 where @xmath75 is defined in eq .
( [ eq : defh ] ) and the last line comes from the intersubband part of the coulomb interaction only ( we have omitted the constant ground state energy ) .
if the @xmath76 operators are supposed to obey bosonic commutation rules @xmath77 then the hamiltonian in eq .
( [ hib ] ) is quadratic and can thus be solved by an hopfield - bogoliubov transformation @xcite .
the different approximations that led to eq .
( [ hib ] ) ( parabolicity , small exchanged momentum , perfect bosonicity , absence of intrasubband scattering ) have been thorougly tested in multiple experiments @xcite in the linear regime .
still , these simplifying approximations oblige us to neglect a certain number of phenomena , like the interplay between nonparabolicity and coulomb interaction @xcite , or the scattering toward electron - hole pairs at large wavevectors , that is known to be an important factor in the thermalization and dynamics of exciton polaritons @xcite .
moreover neglecting intrasubband coulomb scattering does not allow us to consider polariton - polariton coulomb interactions , that would appear as an higher order term in the hamiltonian of eq .
( [ hib ] ) . indeed , the bosonic approximation in eq .
( [ comm ] ) is satisfied only if its expectation value is taken between states corresponding to a weak excitation density .
otherwise , we will have a correction term in eq .
( [ comm ] ) , roughly proportional to the ratio between the number of excitations , @xmath78 , and of electrons , @xmath79 , in such a way that this deviation can become important at higher excitation densities @xcite . to go beyond these limitations
, we will derive an effective quartic bosonic hamiltonian , taking care of both nonbosonicity and coulomb scattering . in order to do so
, we will have to calculate the matrix elements of the full hamiltonian in eq .
( [ hf ] ) between states with multiple intersubband excitations .
the algebra necessary to calculate the coefficients of such an effective hamiltonian is rather heavy , and we will use the coboson formalism , originally developed for exciton polaritons @xcite . the reader can refer to the next section for a brief summary of such a formalism , highlighting the aspects relevant for the algebra of the present paper .
here we generalize the composite boson ( coboson ) approach , which was originally developed in refs .
[ ] and [ ] for the case of excitons .
such a formalism applied to intersubband excitations provides a systematic way to calculate many - body matrix elements . for simplicity
, we consider the case where @xmath80 .
the case with many quantum wells is more cumbersome , but relatively straightforward .
final results are obtained by replacing @xmath81 by @xmath82 in eq .
( [ eq : paulidef ] ) , ( [ eq : twoisb ] ) and ( [ eq : direct ] ) .
we start by writing hamiltonian @xmath8 in eq .
( [ hf ] ) in the appropriate form to use the composite boson commutator formalism .
in particular it must annihilate the ground state .
our main interest will be to calculate matrix elements describing scattering events between intersubband excitations .
we will thus limit ourselves to consider matrix elements of the form @xmath83 , where all intersubband excitation creation operators are on the right of the hamiltonian and annihilation operators on the left .
it is thus immediate to verify that terms in the hamiltonian containing operators @xmath84 or @xmath85 in normal order , with @xmath86 outside the fermi sea , do not contribute to such matrix elements , as they annihilate the state .
we can thus remove them by truncating sums over @xmath86 , relative to the first subband , to the fermi sea ( @xmath87 ) .
we then rewrite @xmath88 in eq .
( [ hf ] ) in the following form ( omitting the ground state energy and non - contributing terms ) @xmath89 where @xmath90 @xmath91 @xmath92 all sums over the first subband are again truncated to the fermi wavevector and @xmath93 in the previous equation we have thus explicitly separated the intersubband and intrasubband terms of the coulomb interaction , putting them respectively into @xmath94 and @xmath95 . the first term will give the depolarization shift @xcite , while the second one contains the electron - hole interaction which lowers the energy of bare intersubband excitations @xcite and is also responsible for the scattering between intersubband excitations .
moreover , we commuted the electronic operators in such a way that now @xmath96 and @xmath95 annihilate the ground state @xmath97 . writing the hamiltonian in such a form
makes it also apparent the renormalization effect of the coulomb interaction over the electronic dispersions , that is the same as the one that would be given by an hartree - fock approach@xcite .
the starting point of the composite boson formalism is the exact commutator between intersubband transitions defined in eq .
( [ isbt ] ) @xmath98 } } & = & \delta_{m , i } \delta_{\mathbf{q } , \mathbf{q ' } } - d_{m\mathbf{q } , i\mathbf{q } ' } , \end{aligned}\ ] ] where @xmath99 is the deviation from bosonicity operator defined by @xmath100 to calculate many - body matrix elements involving intersubband excitations , we need to evaluate the following commutator @xmath101 } } & = & \frac{1}{{\ensuremath { n_{\text{el } } } } } \sum_{n } \left\ { { \ensuremath { \lambda_{m , i}^{n , j } ( \mathbf{q - q ' } ) } } \right . \\ & & + \left .
{ \ensuremath { \lambda_{m , j}^{n , i } ( \mathbf{q - q '' } ) } } \right\ } { b^{\dagger}_{n,\mathbf{q'+q''-q } } } , \nonumber\end{aligned}\ ] ] where @xmath102 the term appearing in eq .
( [ eq : pauli ] ) is analogous to an effective scattering due to the nonbosonicity of the intersubband excitations ( pauli scattering ) : two intersubband excitations can exchange their electrons to create two excitations with different wavevectors , the total wavevector being conserved . with the identities above mentioned , it is possible to compute the scalar products of many - excitation states as @xmath103 from eqs .
( [ eq : hrabi ] ) and ( [ eq : hdep ] ) we see that matrix elements of @xmath61 and @xmath94 can be written as scalar products of many - excitation states and can thus be calculated with the previous formula . in order to evaluate the many - body matrix elements of the kinetic part of the hamiltonian @xmath96
, we need the commutator @xmath104 } } = \hbar \tilde{\omega}_{12 } { b^{\dagger}_{i,\mathbf{q}}},\ ] ] where we have approximated , for simplicity , @xmath105 with @xmath106 .
many body matrix elements of @xmath96 can thus be calculated by commuting it all the way to the right , leaving behind only terms in the form of scalar products between states with many intersubband excitations . to complete the framework
, we need to consider the matrix elements of @xmath95 .
we start again by calculating its commutator with an intersubband excitation creation operator @xmath107 } } & = & - \sum_{j } \gamma_{i , j } { b^{\dagger}_{j,\mathbf{q } } } + v_{i , \mathbf{q } } , \label{eq : defcreation}\end{aligned}\ ] ] where @xmath108 is a coupling between bright and dark excitations @xmath109 and @xmath110 is called the creation potential operator , @xmath111 the coefficient @xmath112 in eq .
( [ eq : defcreation ] ) is a renormalization of the intersubband transition energy due to the electron - hole interaction .
notice that the creation potential operator , @xmath110 , as well as @xmath96 and @xmath95 , annihilate the ground state @xmath97 , a property which will be important when calculating matrix elements .
the last commutator we need is the one between a creation potential operator and an intersubband excitation creation operator , that is @xmath113 } } & = & \frac{1}{{\ensuremath { n_{\text{el } } } } } \frac{e^{2 } { \ensuremath { n_{\text{el } } } } } { 2\epsilon_{0}\epsilon_{r } } \\ & & \sum_{m , n,\mathbf{q } } \frac { { \ensuremath { \xi_{m , i}^{n , j } } } ( \mathbf{q } ) } { q \epsilon({\ensuremath{\mathbf { q } } } ) } \ , { b^{\dagger}_{m,\mathbf{q+q } } } { b^{\dagger}_{n,\mathbf{q'-q } } } , \nonumber \qquad\end{aligned}\ ] ] where @xmath114 is the so - called direct scattering . in order to calculate an arbitrary matrix element concerning @xmath95 between many - excitations states ,
we can commute it all the way to the right , leaving behind terms that will have the form of a creation potential operator , @xmath115 , sandwiched between multiple intersubband excitation operators .
we can then commute creation potential operators all the way to the right , leaving at the end only many - excitations states scalar products , that we can calculate using eq .
( [ eq : twoisb ] ) .
our aim is to derive a bosonic hamiltonian capable of describing intersubband polariton - polariton interactions .
we will thus introduce the bosonic ground state @xmath116 and @xmath117 , the annihilation operator for a bright bosonic intersubband excitation , with in - plane wavevector @xmath13 satisfying the usual bosonic commutation rules @xmath118 } } & = & \delta_{\mathbf{q},\mathbf{q ' } } .\end{aligned}\ ] ] the index has been omitted as we are only interested in the dynamics of bright excitations .
we will then calculate the coefficients of the bosonic hamiltonian @xmath119 , imposing that it has the same expectation values that @xmath8 , for states with up to two excitations . while the most general hamiltonian , up to @xmath120-body interactions , can be calculated with the same procedure , here we will limit ourselves to consider only resonant two - body interactions involving bright intersubband excitations .
the actual calculation of the coefficients requires some rather complex operatorial algebra . for sake of simplicity
, we have thus moved the bulk of the calculations in the appendices , presenting here only the final form of the hamiltonian and the parameter dependency of its coefficients . in sec .
[ coboson ] , we have provided the essential ingredients of the composite boson commutator approach @xcite , that we then use extensively in appendix [ sec : coeff ] to calculate the actual coefficients of the bosonic hamiltonian .
the bosonic hamiltonian , describing the bright excitations , and using the quantities defined in sec . [ coboson ] and in appendix [ sec : coeff ] , has thus the form @xmath121 \nonumber \\ & & + { \sum_{\mathbf{q } } } \hbar\omega_{q } \ , ( { b_{\mathbf{q}}}+{b^{\dagger}_{\mathbf{-q}}})({a_{\mathbf{-q}}}+{a^{\dagger}_{\mathbf{q } } } ) \nonumber \\ & & - \frac{1}{{\ensuremath { n_{\text{qw } } } } { \ensuremath { n_{\text{el } } } } } { \sum_{\mathbf{q , q',p } } } \frac{\hbar\omega_{q}}{2 } \left [ 1 + \left ( 1-\delta_{\mathbf{p,0 } } \right ) \left ( 1-\delta_{\mathbf{p , q'-q } } \right ) \right ] \ , { b^{\dagger}_{\mathbf{q+p}}}{b^{\dagger}_{\mathbf{q'-p}}}{b_{\mathbf{q'}}}{a_{\mathbf{q}}}+ \text{h.c . } \nonumber \\ & & - \frac{1}{{\ensuremath { n_{\text{qw } } } } { \ensuremath { n_{\text{el } } } } } \frac{1}{2 } { \sum_{\mathbf{q , q',p } } } u \ , { b^{\dagger}_{\mathbf{q+p}}}{b^{\dagger}_{\mathbf{q'-p}}}{b_{\mathbf{q'}}}{b_{\mathbf{q } } } , \end{aligned}\ ] ] where the last nonlinear coefficient , describing the scattering of pairs of intersubband excitations , is @xmath122 in the previous expressions @xmath1 denotes the electron density in each quantum well . a comparison with eq .
( [ hib ] ) shows that taking into account the nonbosonicity of intersubband excitations and the intrasubband terms of the coulomb interaction resulted in the appearance of quartic terms in the hamiltonian , that makes it possible to observe nonlinear polariton - polariton processes . in the first and fourth line in eq .
( [ hb ] ) and in eq .
( [ eq : u ] ) coefficients have no dependence over the wavevectors of the intersubband excitations .
this is justified by the fact that these coefficients evolve significantly on a scale of the order of the fermi wavevector @xmath123 , which is much larger than the photonic wavevectors of interest .
there is also no spin dependence as one should expect from a naive comparison with excitons @xcite . as explained in sec .
[ sec : isb ] and appendix [ sec : notation ] this is due to the fact that all intersubband of interest are made of two fermions with the same spin . the first quartic term in eq . ( [ hb ] ) describes the saturation of the light - matter interaction .
its physical origin lies in the non - perfect bosonicity of the intersubband excitations and accounts for a renormalization of the rabi frequency at high excitation densities .
its presence is easy to understand if one remembers @xcite that the light - matter interaction is proportional to the square root of the total electronic population in the first subband minus the one in the second subband .
as the creation of one intersubband excitation corresponds , in average , to the promotion of one electron from the first to the second subband , we do expect that an increase in the intersubband excitations population will reduce the light - matter coupling , consistently with the minus sign in front of such a coefficient .
the second quartic term describes instead the scattering between pairs of intersubband excitations .
it has three contributions as seen in eq .
( [ eq : u ] ) .
the first one comes from the intersubband coulomb interaction .
the second and third contributions instead are due to the intrasubband coulomb interaction .
it is important to notice that the two nonlinear coefficients behave differently with respect to the number of quantum wells in the cavity .
while @xmath124 is proportional to the square root of the total number of electrons in the cavity , @xmath125 depends only on the number of electrons in each well .
this is due the fact that the photonic mode couples to all electrons in the cavity while the coulomb interaction does not couple one well with another .
as we can see , coulomb interaction and the non - perfect bosonicity also give a renormalization effect visible in the quadratic part of the hamiltonian .
the intersubband energy is renormalized both by hartree - fock terms and by the coulomb electron - hole interaction
. moreover , the non - resonant terms of the depolarization shift are renormalized by a @xmath126 coefficient as a consequence of the nonbosonicity . again
, the expressions of the coefficients can be found in appendix [ sec : coeff ] .
the previous hamiltonian is adapted to describe optically pumped experiments , in which the populations of the dark states remain negligible .
while we will not treat this case in the present work , we notice that , in the case of electrical pumping , the dark states population can become significant .
interaction between bright and dark modes is described by the hamiltonian @xmath127 where the index @xmath69 means that the mode is dark . in a mean field approach
this hamiltonian can be seen to describe energy shifts and saturation effects on the bright modes as a function of the dark state population .
coefficients in eq .
( [ hd ] ) can be calculated from a generalization of the formula given in sec .
[ coboson ] and in appendix [ sec : coeff ] . passing from the fermionic description @xmath88 to the bosonic one @xmath128 , we have kept only quartic terms in @xmath128 , discarding all higher order terms .
this is effectively equivalent to limit ourselves to the lowest order in the excitation fraction @xmath129 @xcite .
we thus expect that our theory will correctly describe the polariton - polariton interactions in the regime of moderate excitation densities@xcite . in order to clarify this point
we give here an explicit example not involving coulomb interaction .
the reader can refer to appendix [ sec : fermi ] for more complex and detailed examples .
we consider a process in which a state with one photon and @xmath78 intersubband exitations is converted into a state with no photon and @xmath130 intersubband excitations .
the matrix element of the coulomb part of the hamiltonian vanishes in this case , as it conserves the number of photons , and we can thus perform the calculation exactly , both in the fermionic and in the bosonic formalism .
we obtain respectively @xmath131 which are indeed equal to first order in @xmath129 .
in this section we provide the numerical values of the relevant coefficients that appear in the bosonic hamiltonian in eq .
( [ hb ] ) , highlighting the dependence of the main parameters .
we start by analyzing the renormalization effect due to the coulomb interaction on the quadratic part of the effective hamiltonian . performing a bogoliubov transformation on the first line of eq .
( [ hb ] ) we obtain the renormalized energy of the intersubband excitation ( that we should call , more properly , intersubband plasmon @xcite ) @xmath132 in fig .
( [ fig : plasmon ] ) we plot the dispersion of the intersubband transition energy considering a gaas quantum well of length @xmath133 ( left panel ) and @xmath134 ( right panel ) , corresponding to bare transitions of @xmath135 @xcite and @xmath136 @xcite respectively .
the solid line depicts the result for the intersubband transition including all contributions in eq .
( [ hb ] ) .
the dashed line represents the result obtained while neglecting the coulomb intrasubband interactions and the nonbosonicity .
the dash - dotted line is the result obtained as the dashed line , but neglecting also the screening of the intersubband coulomb interaction .
notice that the renormalized intersubband energy , @xmath137 , in eq .
( [ eq : eisbt ] ) , converges to the bare transition energy @xmath5 for vanishing doping .
we see that the main effect is due to the screening of the intersubband coulomb interaction .
it appears clear that these renormalization effects are relevant in the thz regime ( right panel ) , but rather moderate for mid - infrared transitions ( left panel ) .
[ tc]@xmath138 [ bc]@xmath139 [ tc]@xmath138 [ bc]@xmath139 of the intersubband transition for a bare energy @xmath140mev ( left panel ) and @xmath141mev ( right panel ) , as a function of the electron density .
solid line : result including all the contributions in eq .
( [ eq : eisbt ] ) , namely coulomb interaction and nonbosonicity .
dashed line : result neglecting intrasubband coulomb interactions and the non - bosonicity .
dash - dotted line : same as the dashed line , but with no screening of the intersubband coulomb interaction.,width=340 ] [ tc]@xmath138 [ bc]@xmath142 [ tc]@xmath138 [ bc]@xmath142 between intersubband excitations including all the contributions in eq .
( [ eq : u ] ) for @xmath140mev ( left panel ) and @xmath143mev ( right panel ) as a function of the electron density .
dashed - line : the first term in eq .
( [ eq : u ] ) corresponding to the intersubband coulomb interaction . thin red line : the absolute value of the second term in eq .
( [ eq : u ] ) , namely the direct coulomb interaction .
note that this term is negative , thus producing a red - shifted contribution .
dash - dotted line : the third term in eq .
( [ eq : u ] ) , which is due to the exchange coulomb interaction.,width=340 ] now , we consider the quartic part of the hamiltonian @xmath128 , responsible for the coulomb scattering of pairs of intersubband excitations .
the interaction energy is given by @xmath144 . in fig .
[ fig : fourcoef ] we plot the energy @xmath125 ( thick solid line ) for a mid - infrared transition ( left panel ) and a thz transition ( right panel ) .
the other lines depict the individual contributions of the three terms in eq .
( [ eq : u ] ) ( see caption for details ) . for the considered realistic parameters
, the interaction energy grows with increasing electron doping density almost linearly for the mid - infrared case , while it saturates at high densities for the case of a thz transitions .
notice that in the case of excitons@xcite , the direct scattering vanishes in the limit of small wavevectors , whereas it is not the case for intersubband excitations when the screened coulomb interactions are considered .
this is due the fact that electron - electron and hole - hole interactions are screened differently by the fermi sea@xcite .
note that for a doping density in the range of a few @xmath145 , the energy @xmath125 is of the order of a few mev both for the cases of thz and mid - infrared transitions .
the interaction energy can thus reach values close to the mev when the density of excitations in the system becomes significant .
this is rather promising , since as shown in the case of exciton - polaritons@xcite , very interesting nonlinear polariton physics occurs when the interaction energy becomes comparable to the linewidth of the polariton modes . for thz polaritons , state - of - the - art samples@xcite exhibits polariton width in the mev range .
in this section we consider the interactions in the polariton basis . for simplicity
we limit ourselves to the first mode of the cavity and we assume that the number of electrons is such that the system is in the strong coupling regime . the diamagnetic interaction in @xmath7 as well as all non - resonant terms in @xmath128 are thus neglected .
a bogoliubov transformation of the quadratic parts of @xmath7 and @xmath128 gives the expression of the polaritonic operators @xmath146 where @xmath147 and @xmath148 are polaritonic operators of the upper ( up ) and lower branch ( lp ) respectively , while @xmath149 and @xmath150 are the hopfield coefficients .
note that results can be generalized to the ultrastrong coupling regime ( vacuum rabi frequency comparable to intersubband transitions ) by considering the bogoliubov transformation involving the antiresonant light - matter interaction terms @xcite . here ,
for simplicity , we consider the case of moderate values of the vacuum rabi frequency . because of
the quartic terms in hamiltonian @xmath128 polaritons interact with each other through their matter part and can scatter . in the following ,
we will focus on the lower branch . replacing operators @xmath12 and @xmath117 by their expression ,
the two - body interaction between lp polaritons is found to be @xmath151 where @xmath152 as an example , we now consider the case where the lower branch is pumped at a wavevector @xmath153 so that the system is in the state @xmath154 .
this hamiltonian allows us to describe single - mode ( kerr ) and multimode ( parametric ) coherent non - linearities@xcite . for the parametric case
, one has to consider interaction channel @xmath155 where pairs of polaritons scatter from the pumped mode into signal - idler pairs . as for the case of exciton - polaritons
we expect that the maximum efficiency of this parametric processes is achieved when the energy conservation condition is fulfilled@xcite .
a mean - field approach of the problem@xcite shows that the matrix element between the initial and the final states is the relevant quantity to consider and has to be compared with the lifetime of the excitations .
for high pump intensity i.e. @xmath156 this matrix element is @xmath157 as discussed in sec .
[ interaction ] and shown in fig .
[ fig : fourcoef ] , for thz polaritons nonlinear interaction energies of the order of a few mev ( thus comparable to thz polariton linewidths ) can be achieved , thus paving the way to a very interesting coherent nonlinear physics for this kind of composite excitations .
in this paper , we have presented comprehensively a microscopic theory for the manybody physics of intersubband cavity polaritons . using a composite boson commutator approach ,
we have derived the manybody interaction matrix elements involving ( screened ) coulomb interaction processes and pauli saturation effects .
we have also derived an effective bosonic hamiltonian reproducing the matrix elements in the two - excitation manyfold calculated by the composite boson approach . as a first application
, we have calculated the renormalization of the intersubband transition frequency at the mean - field level and determined the strength of the interactions between intersubband excitations .
we have also derived the renormalization of the light - matter coupling due to the pauli saturation effects and considered the interaction physics in the polariton basis .
our results show that significant intersubband polariton - polariton interactions occur , especially for transitions in the thz regime .
our work paves the way to promising future studies of nonlinear quantum optics with intersubband cavity polaritons .
we acknowledge partial support from the french anr project thinqe - pinqe .
mb is also supported by kakenhi 24 - 632 .
we wish to thank c. sirtori , y. todorov , a. vasanelli for discussions .
in this appendix we give the explicit notations for eq .
( [ eq : hfdetail ] ) and eq .
( [ eq : isb0 ] ) with quantum well and spin indexes .
electronic wavefunctions are localized in quantum wells and we neglect electronic tunneling from one well to another .
we thus neglect coulomb interaction between electrons in different wells , which are sufficiently apart .
( [ eq : hfdetail ] ) and ( [ eq : isb0 ] ) can then be written with all indexes @xmath158 and @xmath159 where @xmath160 for all @xmath161 , @xmath11 is a wavevector such that @xmath162 , @xmath163 and @xmath164 .
bright intersubband excitations are thus linear superposition of pairs of fermions with the same spin
. we could generalize eq .
( [ eq : isb0index ] ) by allowing the two fermions to have different spins but the resulting collective excitation would be dark and thus not relevant if we consider only polariton .
in order to determine the coefficients of the effective bosonic hamiltonian in eq .
( [ hb ] ) , we impose that it has the same matrix elements as the fermionic one in eq .
( [ hf ] ) over the hilbert space spanned by one and two excitation states .
that is we impose that @xmath165 where @xmath166 , @xmath167 , @xmath168 is the projector over the subspace orthogonal to the ket@xcite and @xmath169 and @xmath170 are the normalizations of the orthogonalized states . the constraints in eq .
( [ constr ] ) , together with the fact that we decided to consider , beyond quadratic terms , only the quartic ones , are enough to uniquely define @xmath128 .
the calculation of the coefficients of @xmath128 thus reduces to the calculation of matrix elements of the fermionic hamiltonian @xmath8 , that we can perform exploiting the commutator approach outlined in sec .
[ coboson ] .
we thus start from a general bosonic hamiltonian @xmath171 we also introduce the following five matrix elements ( each one implicitly dependent on all the relevant wavevectors ) @xmath172 the relations between the coefficients of @xmath119 in eq .
( [ hbtest ] ) and the matrix elements in eq .
( [ dcoeff ] ) can be found with some algebra to be @xmath173 @xmath174 and @xmath175 where , for consistency , terms under the square root have also to be developed to the first order in @xmath82 .
we have thus reduced the determination of the coefficients of @xmath128 to the calculation of the five matrix elements in eq .
( [ dcoeff ] ) . without loss of generality , in the following calculations
, we will assume that the excitations wavevectors are small compared to the fermi wavevector , which implies that the pauli scattering involving two bright excitations , defined in eq .
( [ eq : paulidef ] ) , can be approximated by a kroneker delta @xmath176 the only contributing terms are the kinetic and coulomb part of the hamiltonian @xmath88 .
the matrix element is @xmath178 only the light - matter part of @xmath88 contributes and we obtain @xmath180 this term comes from the intersubband coulomb interaction and is given by @xmath182 where @xmath183 the coefficient @xmath184 is then @xmath185 .
as the number of electrons is large we can neglect the correction due to the normalization .
as for the coefficient @xmath187 , only the light - matter part of @xmath88 contributes , giving @xmath188 the last term , at the origin of the first quartic term in eq .
( [ hb ] ) , is due to the nonbosonicity .
the expression of the coefficient is given by the following relations ( to first order in @xmath190 ) ) @xmath191 @xmath192 where @xmath193 in eq .
( [ eq : matrixgeneral ] ) the long wavelength limit has been taken because the variation of the coefficients is significant only on a scale of the order of the fermi wavevector @xmath123 .
in this appendix we calculate the transitions rates for scattering of pairs of intersubband excitations using both the effective bosonic hamiltonian approach and the fermionic formalism developed by m. combescot and coworkers @xcite , showing that the two approaches give exactly the same results in this paragraph we use the composite boson commutator formalism to calculate the transition rate between an initial state @xmath194 and a final state @xmath195 , as well as the lifetime of the former , using a first - order , time - dependent perturbation theory .
the calculation is equivalent to a fermi golden rule as explained in ref .
the particular event we want to describe is the scattering of an initial pump beam of arbitrary intensity into a signal and an idler mode .
we will thus consider initial and final states to be proportional respectively to the kets @xmath196 and @xmath197 the transition rate can then be written as @xcite @xmath198 where @xmath199 as in appendix [ sec : coeff ] , @xmath168 is the projector over the subspace orthogonal to @xmath194 and @xmath200 verifies @xmath201 where @xmath202 converges to the dirac delta function for long times .
the orthogonal projection in eq .
( [ eq : psit ] ) can be calculated as the previous expression is a first order in the pertubations .
we can thus consider only the zero - th order in @xmath200 , _
i.e _ we do as if intersubband excitations were exact eigenstates of @xmath88 , obtaining with the same notations , the contribution of the many - body physics to the lifetime of the initial state is@xcite @xmath206 the second term can be calculated using the same method as for the transition rate and is found to contribute only to higher orders in the perturbation . after some rather cumbersome algebra
, the first term is found to be of the form @xmath207 where it is implicitly assumed that the summation is restricted to small wavevectors . to obtain this result we made @xmath200 act at zero - th order and we used the fact that bright and dark intersubband excitations are not resonant .
the counterintuitive @xmath208 factor comes from the overcompleteness of the composite boson basis . in this paragraph
we calculate the same quantities as in the previous one but we assume that intersubband excitations are bosons and that their dynamics is described by hamiltonian @xmath128
. we can therefore use a traditional fermi golden rule .
the initial and final states in this case are @xmath209 and @xmath210 the transition rate is thus given by the formula @xmath211 which is the same as in eq .
( [ eq : rate ] ) up to third order in @xmath129 .
as all bright intersubband excitations are resonant the lifetime of the initial state is the sum of the transition rates @xmath212 where the summation is again restricted to small wavevectors . a comparison with eq .
( [ eq : lifetime ] ) shows that this method underestimates the true lifetime by a factor two .
this is coherent with the results in ref .
[ ] , that show how an effective hamiltonian approach that correctly calculates transition rates needs to take into account an _ ad hoc _
factor @xmath208 when calculating lifetimes , due to the overcompleteness of the composite boson basis .
this is of course simply implies a renormalization of the composite boson density of states .
d. dini , r. kohler , a. tredicucci , g. biasiol and l. sorba , phys .
* 90 * , 116401 ( 2003 ) . c. ciuti , g. bastard and i. carusotto , phys .
b * 72 * , 115303 ( 2005 ) . s. de liberato , c. ciuti and i. carusotto , phys .
lett . * 98 * , 103602 ( 2007 ) .
e. dupont , j. a. gupta , and h. c. liu , phys .
b * 75 * , 205325 ( 2007 ) . m. f. pereira , phys . rev .
b * 75 * , 195301 ( 2007 ) . l. sapienza , a. vasanelli , r. colombelli , c. ciuti , y. chassagneux , c. manquest , u. gennser and c. sirtori , phys .
* 100 * , 136806 ( 2008 ) . g. gnter , a. a. anappara , j. hees , a. sell , g. biasiol , l. sorba , s. de liberato , c. ciuti , a. tredicucci , a. leitenstorfer and r. huber , nature ( london ) * 458 * , 178 ( 2009 ) . y. todorov , a. m. andrews , i. sagnes , r. colombelli , p. klang , g. strasser , and c. sirtori , phys . rev .
* 102 * , 186402 ( 2009 ) .
a. a. anappara , s. de liberato , a. tredicucci , c. ciuti , g. biasiol , l. sorba , and f. beltram , phys .
b * 79 * , 201303 ( 2009 ) .
m. zauny , w. zietkowski , phys .
b * 80 * , 245301 ( 2009 ) .
p. jouy , a. vasanelli , y. todorov , l. sapienza , r. colombelli , u. gennser , and c. sirtori , phys .
b * 82 * , 045322 ( 2010 ) .
a. auer and g. burkard , phys .
b * 85 * , 235140 ( 2012 ) .
r colombelli , c ciuti , y chassagneux , c sirtori , semiconductor science and technology * 20 * , 985 ( 2005 ) .
a. a. anappara , a. tredicucci , g. biasiol , l. sorba , applied physics letters * 87 * , 051105 ( 2005 ) .
a. a. anappara , a. tredicucci , f. beltram , g. biasiol , l. sorba , applied physics letters * 89 * , 171109 ( 2006 ) .
j. plumridge , e. clarke , r. murray , c. phillips , solid state communications * 146 * , 406 ( 2008 ) .
y. todorov , p. jouy , a. vasanelli , l. sapienza , r. colombelli , u. gennser , c. sirtori , applied physics letters * 93 * , 171105 ( 2008 ) .
m. geiser , c. walther , g. scalari , m. beck , m. fischer , l. nevou , j. faist , applied physics letters * 97 * , 191107 ( 2010 ) .
s. de liberato , c. ciuti and c. c. phillips , arxiv:1204.4053 m. porer , j .-
mnard , a. leitenstorfer , r. huber , r. deglinnocenti , s. zanotto , g. biasiol , l. sorba and a. tredicucci , phys .
b * 85 * , 081302 ( 2012 ) .
a. delteil , a. vasanelli , p. jouy , d. barate , j. c. moreno , r. teissier , a. n. baranov and c. sirtori , phys .
b * 83 * , 081404 ( 2011 ) .
s. de liberato and c. ciuti , phys .
lett . * 102 * , 136403 ( 2009 ) .
d. e. nikonov , a. imamolu , l. v. butov and h. schmidt , phys .
* 79 * , 4633 ( 1997 ) .
s. de liberato and c. ciuti , phys .
b * 85*,125302 ( 2012 ) .
y. todorov , a. m. andrews , r. colombelli , s. de liberato , c. ciuti , p. klang , g. strasser and c. sirtori , phys .
lett . * 105 * , 196402 ( 2010 ) .
y. todorov and c. sirtori , phys .
b * 85 * , 045304 ( 2012 ) .
o. kyriienko and i. a. shelykh , arxiv:1112.0240v1 ( 2012 ) m. combescot , o. betbeder - matibet and f. dubin , phys . rep . *
463 * , 215 ( 2008 ) .
a. anappara , a. tredicucci , f. beltram , g. biasiol , l. sorba , s. de liberato and c. ciuti , app .
. lett . * 91 * , 231118 ( 2007 ) . g. giuliani and g. vignale , _ quantum theory of the electron liquid _ , cambridge university press ( 2005 ) .
c . lee and i. galbraith , phys .
b * 59 * , 15796 ( 1999 ) .
c . lee and i. galbraith , phys .
b * 62 * , 15327 ( 2000 ) .
j. j. hopfield , phys .
rev . * 112 * , 1555 ( 1958 ) .
l. sapienza , a. vasanelli , c. ciuti , c. manquest , c. sirtori , r. colombelli and u. gennser , app .
. lett . * 90 * , 201101 ( 2007 ) .
d. porras , c. ciuti , j. j. baumberg and c. tejedor , phys .
b * 66 * , 085304 ( 2002 ) . c. ciuti , phys . rev .
b * 69 * , 245304 ( 2004 ) .
m. combescot , o. betbeder - matibet and r. combescot , phys .
b 75 , 174305 ( 2007 ) . c. ciuti ,
v. savona , c. piermarocchi , a. quattropani , phys revb * 98 * , 7926 ( 1998 ) . m. vladimirova , s. cronenberger , d. scalbert , k. v. kavokin , a. miard , a. lematre , j. bloch , d. solnyshkov , g. malpuech and a. v. kavokin , phys . rev .
b * 82 * , 075301 ( 2010 ) .
s. de liberato and c. ciuti , phys .
b * 77 * , 155321 ( 2008 )
. m. m. glazov , h. ouerdane , l. pilozzi , g. malpuech , a. v. kavokin , and a. dandrea , phys .
b * 80 * , 155306 ( 2009 ) .
i. carusotto , c. ciuti , rev .
( in press , 2012 ) ; arxiv:1205.6500v3 p. g. savvidis , j. j. baumberg , r. m. stevension , m. s. skolnick , d.m . whittaker and j.s
roberts , phys .
lett . * 84 * , 1547 ( 2000 ) c. ciuti , p. schwendimann and a. quattropani , semiconductor science and technology * 18*(10 ) , s279 ( 2003 ) . | intersubband polaritons are light - matter excitations originating from the strong coupling between an intersubband quantum well electronic transition and a microcavity photon mode . in this paper
we study how the coulomb electron - electron interaction and the pauli saturation of the electronic transitions affect the physics of intersubband polaritons .
we develop a microscopic many - body theory for the physics of such composite bosonic excitations in a microcavity - embedded two - dimensional electron gas . as a first application
, we calculate the modification of the depolarization shifts and the efficiency of intersubband polariton - polariton scattering processes . |
surface magnetic fields on low mass stars give rise to a variety of interesting astrophysical phenomena , including a persistent hot outer atmosphere ( which may be contained within magnetic loop structures ) and transient events such as starspots and flares .
non - radiative heating processes transfer energy from the magnetic field into the outer stellar atmosphere to power these phenomena , but a physical description of the heating processes remains elusive even on the sun .
observationally , the heating produces chromospheric ( t@xmath0k ) , transition region ( t@xmath1k ) and coronal ( t@xmath2k ) emission , primarily in emission lines of the hydrogen lyman and balmer series , and in ionized species of abundant elements such as helium , carbon , oxygen , calcium , magnesium , and iron .
recently , there has been renewed interest in the atmospheric properties of the lowest mass stars and brown dwarfs .
chromospheric observations of h@xmath3 emission @xcite indicate that both the frequency of active stars at a given spectral type , and the activity level ( measured by l@xmath4 ) decrease at spectral types later than approximately m7 .
recent investigations of magnetic activity in these very late type dwarfs have led to a new , yet already widely accepted , paradigm : that hot gas in the atmospheres of very low mass main sequence stars and brown dwarfs is present only during flares @xcite . here , hot gas refers to gas at transition region and coronal temperatures .
the m9 brown dwarf , lp944 - 20 , observed with chandra by @xcite does show persistent chromospheric ( h@xmath3 ) emission during quiescence ( i.e. outside of flares ) , as do many of its counterparts .
this new paradigm is based on the following evidence .
@xcite observed the m8 dwarf vb 10 with hst / ghrs and found no quiescent transition region emission , but did observe emission during a flare .
@xcite re - analyzed rosat data on vb10 and found that the observed x - ray emission came only from a flare event , and no x - rays were detected during six hours of quiescent observation .
the x - ray flare luminosity was substantially larger than expected based on the upper limit of the quiescent x - ray luminosity , when compared to flares on earlier type ( m0 - m6 ) stars .
an h@xmath3 flare on an m9.5 2mass dwarf showed a similarly substantial flare luminosity compared to quiescence @xcite .
finally , the afore - mentioned chandra investigation of the m9 brown dwarf lp944 - 20 detected no x - rays in a quiescent observation of nine hour duration @xcite .
by contrast , @xcite found both persistent and flaring radio emission from lp944 - 20 , while @xcite recently reported radio observations of both flaring and quiescent emission for 3 additional low mass dwarfs of spectral types m8.5 , m9.5 and l3.5 .
these authors speculate that perhaps the radio emission mechanism differs in the low mass dwarfs compared to earlier type active stars . on the theoretical side ,
both @xcite and @xcite have proposed mechanisms related to the low ionization state of very cool atmospheres which would prevent the occurrence of magnetic activity in very cool ( hence late spectral type ) dwarfs , in order to explain the various observations described above .
recent x - ray studies have probed into the brown dwarf regime in young clusters @xcite and find both quiescent and flaring emission .
however , the extreme youth of these objects means that most have early - mid m spectral types .
further , fossil magnetic fields may play a significant role in their activity @xcite , making them unsuitable as a testbed for theories of the evolution of dynamo - generated magnetic fields in older , main sequence dwarfs .
extensive observations of earlier m dwarfs ( type m0-m6 ) have led to scaling relations between the persistent h@xmath3 ( chromospheric ) and soft x - ray ( thermal coronal ) emission outside of flares @xcite .
we have used h@xmath3 observations @xcite and ( transition region ) measurements @xcite for the dme stars ad leo ( m3 ) , ev lac ( m3.5 ) , and yz cmi ( m4 ) to derive empirical to h@xmath3 line flux ratios of 0.06 , 0.07 , and 0.13 respectively .
there may be a tendency for this ratio to increase toward later spectral types .
a conservative estimate suggests that the ratio is typically @xmath5 0.1 . in terms of the observed fluxes ,
the relations between chromospheric , transition region , and coronal diagnostics for active early - mid m dwarfs are therefore : @xmath6 @xmath7 if these relations are applied to the observed h@xmath3 fluxes for vb 10 and lp944 - 20 , the upper limits for the reported non - detections in and x - rays are near or even above the predicted values .
for example , @xcite found an upper limit for vb 10 of @xmath8 ergs @xmath9 s@xmath10 while the scaling relation gives a predicted line flux of @xmath11 ergs @xmath9 s@xmath10 based on the observed h@xmath3 line flux from @xcite .
recently , @xcite reobserved lp944 - 20 with xmm - newton , again failing to detect quiescent emission ; however , this observation is only a factor of 3 more sensitive than the previous chandra result of @xcite .
thus , only slight deviations from the above relations would explain the current non - detections , without requiring a major paradigm shift with regard to magnetic activity on late m dwarfs compared to earlier m dwarfs .
the case for the absence of a quiescent transition region and corona in the lowest mass stars and brown dwarfs thus rests on tenuous observational evidence .
the goal of this investigation is to place stringent limits on the presence of transition region emission using the sensitive hst / stis spectrograph with the fuv - mama detector .
these observations provide the first rigorous test of the presence of persistent quiescent emission on very late type main sequence dwarfs at temperatures well above chromospheric levels , and by inference the ubiquitous existence of a hot outer atmosphere .
our target list comprises the brightest active stars known at spectral types m7 ( vb 8) , m8 ( vb 10 ) and m9 ( lhs 2065 ) .
these targets were chosen to give us the best possible detection limits for persistent quiescent emission .
the total exposure time on each object corresponds to a detection limit ten times fainter than the expected line strength scaled from earlier type active m dwarfs ( equation 1 ) .
the hubble space telescope was used with the stis spectrograph and the fuv - mama detector to obtain 15 consecutive exposures of each target .
all observations were made using the 52@xmath12 long by 0.2@xmath12 wide slit and the g140l grating to give a spectral resolution @xmath13 at 1500 .
for vb 8 , spectral observations started at 12:38 ut on 2001 october 6 .
each g140l spectrum was integrated for approximately 5 minutes , and the total exposure sequence spanned two orbits . for lhs 2065 , spectral observations started at 5:23 ut on 2002 april 27 .
each g140l spectrum was integrated for approximately 10 minutes , and the total exposure sequence spanned four orbits .
the observing sequence for vb 10 was identical to that of lhs 2065 with g140l spectral observations starting at 5:03 ut on 2002 september 7 .
table 1 gives a detailed log of the stis observations .
the data were downloaded from the hst archive , and data reduction was performed with the calstis package @xcite utilizing the stis pipeline .
however , due to the relative faintness of our targets , the pipeline package was unable to properly locate the spectrum of the star on the detector and extract it .
therefore , we performed the spectral extraction from the flatfielded science image manually . to accomplish this
, we coadded all 15 exposures for each star and examined the resulting image by eye , where it was easy to identify bright emission lines such as and .
we then used the order trace from the stis pipeline , applying an offset such that the trace went through the emission lines from our object . examining each of the 15 exposures individually
, we verified that the order location did not change substantially ( more than a pixel ) during any one exposure .
we then extracted the spectrum and background around this offset order location using the same extraction width ( 11 pixels ) used by the pipeline to determine a count rate spectrum .
the count rate spectrum was converted to a flux spectrum via the transformation provided in the calstis pipeline reduction .
we verified our procedure by extracting spectra around the nominal order locations provided by the calstis pipeline for the 15 exposures of each source , recovering the same ( noise ) spectrum produced by the stis pipeline to within the ( somewhat large ) photon statistics .
the spectra extracted for each target from the sum of all 15 exposures are shown in figure 1 .
the emission lines we have identified are typical of those found in active late - type stars @xcite , and are very similar in relative strength to earlier type m dwarf transition region spectra , as shown by comparison to the recent stis / fuv - mama spectrum of the dm1e star au mic @xcite .
the doublet at 1550 is the strongest feature typically observed from low mass stars in the wavelength range of study here and is also the strongest line observed in these 3 stars . for presentation purposes
, we have applied gaussian smoothing to the spectra with a fwhm of 3 .
this smoothing is not used in the flux measurements given in table 1 .
the 1-@xmath14 uncertainty limits are also indicated for the smoothed spectra in figure 1 .
table 1 gives the observed doublet line flux for each exposure in all three stars .
the fluxes were determined by integrating the line over 6.4 centered on the doublet ; the integration limits were set by visual examination of the individual spectra .
the significance of these detections can be estimated from the off - source background in the images , as discussed by @xcite ; see also @xcite .
the image of each stis spectrum shows a clear variation in the background along the dispersion direction , likely due to scattered light from the strong ly-@xmath3 airglow present .
the background shows no variation in the spatial axis at the position of the line . to estimate the significance of the detection in each exposure ,
we step above and below the position of the line on the spectral image and extract the number of background counts in a box the same size as that used to extract the actual line .
this produces approximately 100 estimates of the background , to which we fit a poisson distribution .
we then compare the net number of counts detected in the line to this distribution in order to determine the significance of the detection . in all but one exposure on lhs 2065 ,
the line is detected above the 3@xmath14 level .
for vb 8 , the weakest detection is at the 3.1@xmath14 level , the strongest at the 15.0@xmath14 level , with a mean significance of 7.7@xmath14 . for vb 10 ,
these three values are 3.3@xmath14 , 314@xmath14 , and 39.1@xmath14 respectively . for lhs 2065 ,
the corresponding detections are 2.9@xmath14 , 17.6@xmath14 , and 10.1@xmath14 .
the continuum adjacent to the line is not detected in the same sized integration box at a significant level in any exposure .
although these estimates show that the line is detected in all exposures , they do not represent the uncertainties in the measured flux . using the bayesian method described by @xcite , we have estimated the 68.3% confidence limits for a gaussian distribution .
we have not assumed a gaussian distribution here , but these confidence limits represent a typical benchmark . ] .
the flux confidence limits for each spectrum are reported in table 1 .
time sequences of the measurements , together with the uncertainties , for each target are shown in figure 2 .
there is no definite sign of long - lived flaring ( rise and decay ) in the time sequences observed .
the second orbit on vb 10 shows a strong flare at the end of the orbit ( exposure 7 ) , so the duration of the event can not be measured .
the flux does appear to be variable between exposures on all three stars .
quiescent emission does not typically occur at a constant flux level on active m dwarfs ; similar variability is commonly seen in the quiescent h@xmath3 emission fluxes @xcite .
this may be a sign of small scale flaring as a cause of the quiescent emission .
our data suggest a common origin for the chromospheric and transition region emission , in conflict with the hypothesis that quiescent chromospheric emission is present but that transition region emission is absent except during strong flares .
the line strengths we measure are : @xmath15 ergs s@xmath10 @xmath9 for vb 8 ; @xmath16 ergs s@xmath10 @xmath9 for vb 10 ; and @xmath17 ergs s@xmath10 @xmath9 for lhs 2065 .
the values in parentheses for vb 10 represent the quiescent flux and uncertainty measured from the total of all spectra excluding the strong flare observed in exposure 7 ( see figure 2 ) .
note that when integrated over the entire observation for each target , the line is detected with sufficient counts that uncertainties can be estimated assuming a gaussian distribution for the source and the background @xcite .
the scaling relation based on the h@xmath3 emission in these stars , which assumes that the to h@xmath3 ratio is the same as for earlier type m dwarfs ( equation 1 ) , predicted values of 2 @xmath18 ( vb 8) , 9 @xmath19 ( vb 10 ) , and 1 @xmath18 ( lhs 2065 ) ergs s@xmath10 @xmath9 .
these are close to the observed values , with vb 8 and lhs 2065 having slightly lower observed fluxes , while the observed vb 10 ( quiescent ) flux is somewhat higher than predicted .
the vb 10 ( quiescent ) flux is also a factor of two higher than the upper limit quoted in @xcite .
our conclusion is that active , very low - mass , main - sequence stars ( spectral types m7-m9 ) can maintain a quiescent chromosphere and transition region that is similar to those observed in active , earlier type m dwarfs . flaring may still be ultimately responsible for the outer atmospheric heating in these stars , even during `` quiescence '' ; the role of flares ( particularly micro - flares ) in coronal and transition region heating is still not understood for the sun or for active stars ( * ? ? ?
* ; * ? ? ?
* for example ) .
however , theories involving changes in atmospheric structure , dynamo heating and other mechanisms proposed to explain the lack of quiescent activity in very low mass stars and brown dwarfs @xcite must accommodate the persistent quiescent activity that we have observed at least through spectral type m9 . babu , g.j . & feigelson , e.d .
1996 , astrostatistics , ( cambridge : cambridge university press ) berger , e. et al .
2001 , nature , 410 , 338 berger , e. 2002 , , 572 , 503 burgasser , a. j. , liebert , j. , kirkpatrick , j. d. , & gizis , j. e. 2002 , , 123 , 2744 comer ' on , f. , neuh " auser , r. , & kaas , a. a. 2000 , , 359 , 269 feigelson , e. d. , broos , p. , gaffney , j. a. , garmire , g. , hillenbrand , l. a. , pravdo , s. h. , townsley , l. , & tsuboi , y. 2002 , , 574 , 258 fleming , t.a . , giampapa , m.s . , & schmitt , j.h.m.m .
2000 , , 533 , 372 gizis , j. e. , reid , i. n. , & hawley , s. l. 2002 , , 123 , 3356 gizis , j. e. , monet , d. g. , reid , i. n. , kirkpatrick , j. d. , liebert , j. , & williams , r. j. 2000 , , 120 , 1085 g " udel , m. , audard , m. , kashyap , v. l. , drake , j. j. , & guinan , e. f. 2003 , , 582 , 423 hawley , s.l . ,
gizis , j.e . , & reid , i.n .
1996 , , 112 , 2799 johns
krull , c.m . ,
valenti , j.a .
, & linsky , j.l .
2000 , , 539 , 815 kashyap , v. l. , drake , j. j. , g " udel , m. , & audard , m. 2002 , , 580 , 1118 kraft , r.p . ,
burrows , d.n . , & nousek , j.a .
1991 , , 374 , 344 liebert , j. , kirkpatrick , j.d . , reid , i.n .
, fisher , m.d .
1999 , , 519 , 345 lindler , d. 1999 , calstis reference guide ( greenbelt : nasa / lasp ) linsky , j.l . ,
wood , b.e . , brown , a. , giampapa , m.s . , &
ambruster , c. 1995 , , 455 , 670 martn , e. l. & bouy , h. 2002 , new astronomy , 7 , 595 mohanty , s. , basri , g. , shu , f. , allard , f. , & chabrier , g. 2002 , , 571 , 469 pagano , i. , linsky , j. l. , carkner , l. , robinson , r. d. , woodgate , b. , & timothy , g. 2000 , , 532 , 497 reid , i.n . , hawley , s.l . , & gizis , j.e .
1995 , , 110 , 1838 rutledge , r.e . ,
basri , g. , martn , e.l .
, bildsten , l. 2000 , , 538 , l141 cccccccccccc 1&12:38:04&300&1.92 - 3.13 & & 05:03:19&720&0.48 - 0.88 & & 05:23:22&679&0.46 - 0.87 + 2&12:45:21&300&0.76 - 1.62 & & 05:17:36&720&1.68 - 2.37 & & 05:36:58&720&0.42 - 0.80 + 3&12:50:43&300&0.61 - 1.42 & & 05:29:58&720&0.49 - 0.93 & & 05:49:20&720&0.64 - 1.09 + 4&12:56:05&300&0.72 - 1.58 & & 06:28:18&600&0.08 - 0.34 & & 06:48:47&600&0.53 - 1.01 + 5&13:01:27&300&1.31 - 2.35 & & 06:42:25&600&0.95 - 1.54 & & 07:02:54&600&0.79 - 1.34 + 6&13:06:49&300&1.92 - 3.13 & & 06:52:47&600&0.69 - 1.21 & & 07:13:16&600&0.80 - 1.36 + 7&13:12:11&291&1.45 - 2.57 & & 07:03:09&720&18.4 - 20.5 & & 07:23:38&720&0.39 - 0.77 + 8&14:04:10&300&0.26 - 0.92 & & 08:04:27&600&1.61 - 2.36 & & 08:24:58&600&0.41 - 0.85 + 9&14:13:17&300&0.68 - 1.53 & & 08:18:34&600&3.17 - 4.18 & & 08:39:05&600&0.72 - 1.26 + 10&14:18:39&300&0.52 - 1.28 & & 08:28:56&600&1.25 - 1.93 & & 08:49:27&600&0.09 - 0.37 + 11&14:24:01&300&1.18 - 2.18 & & 08:39:18&720&0.89 - 1.42 & & 08:59:49&720&0.16 - 0.47 + 12&14:29:23&300&0.25 - 0.90 & & 09:40:37&600&1.69 - 2.46 & & 10:01:09&600&0.52 - 1.01 + 13&14:34:45&300&1.41 - 2.48 & & 09:54:44&600&1.53 - 2.28 & & 10:15:16&600&0.29 - 0.69 + 14&14:40:07&300&0.26 - 0.92 & & 10:05:06&600&1.74 - 2.52 & & 10:25:38&600&0.39 - 0.84 + 15&14:45:29&360&0.26 - 0.85 & & 10:15:28&676&0.53 - 1.00 & & 10:36:00&676&0.46 - 0.90 + | we present results from our cycle 10 hst program to search for transition region emission in very low - mass , main - sequence stars in the spectral range m7-m9 .
our program is aimed at 1 ) detecting emission ; and 2 ) distinguishing between flaring and quiescent origin for the emission .
we have obtained hst / stis time series observations of three active , very low mass stars ( vb 8 , vb 10 , and lhs 2065 ) which show persistent activity in transition region lines including , , and .
emission in transition region lines appears to be variable between exposures , but is always observed .
a strong flare was also observed in one 10 minute exposure on vb 10 .
our results indicate that active , very low - mass stars maintain a persistent quiescent chromosphere and transition region that is similar to those observed in active , earlier type m dwarfs , in contrast to suggestions that these low - mass , main - sequence stars exhibit only relatively strong flares and no quiescent emission . |
until the next generation of large radio telescopes systematically detects the 21-cm hyperfine line of neutral , atomic hydrogen ( hi ) in emission at redshifts beyond @xmath3 , our knowledge of the hi at high redshift will remain solely dependent on the study of absorption - line spectra of high redshift quasars .
these spectra are dominated by ly@xmath0 absorption lines originating in intervening hi clouds , and show a large range in hi column density .
the systems with the largest column densities , termed damped ly@xmath0 absorbers ( dlas ; * ? ? ?
* ) , are characterized by @xmath4 and are known to hold most of the hi atoms in the universe , at all redshifts ( cf . * ? ? ?
* ; * ? ? ?
a popular explanation is that the dlas are the progenitors of gas disks in present - day galaxies @xcite .
at @xmath1 it is well established that most of the hi atoms are locked up in the gas disks of @xmath5-type galaxies @xcite .
these @xmath1 hi disks show an frequency distribution of hi column densities [ ] that is virtually identical in shape to that of the dlas @xcite .
furthermore , at redshifts @xmath6 there are approximately 20 identifications of galaxies giving rise to dla absorption in spectra of background quasars @xcite .
the properties of these galaxies ( such as luminosities and impact parameters ) are fully consistent with the idea that the dlas arise in galaxies that are typical in the local universe @xcite .
also the metallicities seen in low redshift dlas are fully consistent with what is expected when gas disks of @xmath1 galaxies are intercepted randomly .
therefore , dlas at low redshifts ( @xmath6 ) probably arise in gas disks of galaxies like those seen in the local universe . at higher redshifts ( @xmath7 )
the situation becomes more complicated .
direct detection of dla host galaxies has proven to be extremely difficult ( see e.g. , * ? ? ?
* ; * ? ? ?
* and references therein ) . this may not be surprising if at these epochs the dlas also arise in gas disks of normal spiral galaxies : the cross - section selection ensures that a large fraction of the dla galaxies is fainter than @xmath5 , and thus difficult to detect @xcite .
the shape of the column density distribution function , , at these redshifts is still consistent with what is seen at @xmath1 .
although this is consistent with the galaxy disk model , it is inconclusive evidence for this model .
it is conceivable that the shape of is determined by rather simple physics , for example ionization at low column densities and h@xmath8 formation at high densities .
this might produce the same at all redshifts , regardless of whether this gas is distributed in ordered disks or not .
other observations of high-@xmath9 dlas are also inconclusive ( see * ? ? ?
* for a comprehensive review ) .
the inferred cosmic star formation rate density from dlas is consistent with that measured from lyman break galaxies @xcite .
however , metallicities of the dlas are typically low ( e.g. , * ? ? ?
* ) , as are the molecular fractions ( e.g. , * ? ? ?
* ) , and analyses of elemental abundances show star formation histories typical of those of quiescent galaxies @xcite .
perhaps one of the most conclusive tests of whether dlas at high redshift arise in gas disks or not , can be made by studying their kinematics .
the first work on this topic was presented by @xcite , who compiled high resolution , high signal - to - noise spectra of low - ionisation metal lines in dlas , and compared them with the velocity profiles derived from simulations using model ensembles of artifical clouds with a range of velocities .
they concluded that rapidly rotating thick disks are the most likely hosts of dlas .
this conclusion was contested by @xcite , who showed that irregular protogalactic clumps can reproduce the absorption line properties of dlas equally well .
a direct comparison with 21-cm data of local galaxies was first made by @xcite , who found that low mass galaxies such as the large magellanic cloud ( lmc ) display much lower velocity widths than normally seen in dlas .
a fair comparison with dla kinematics can only be made by studying a representative sample of local galaxies for which high quality 21-cm data cubes exist , and which have high spatial and frequency resolution and low hi column density limits . in this paper
we make use of the things ( the hi nearby galaxy survey ; * ? ? ?
* ) data , which are ideal for this purpose .
it is important to note that the ly@xmath0 lines that define the dla carry little information on the kinematics of the absorbing systems themselves .
the velocity width that would correspond to the total extent of the damping wings is much larger than the physical velocity width corresponding to the turbulence or large scale movements of the gas ; any structure in the velocity distribution of the hi gas is completely washed out by the damping wings .
instead , workers in this field rely on low - ionization metal lines , such as fe@xmath10 , si@xmath10 , ni@xmath10 , which are believed to be good tracers of the neutral gas ( e.g. , * ? ? ?
in the same system each of these transitions gives consistent velocity widths , as long as the lines are unsaturated . in principle , many small clouds in the outskirts of galaxy halos could contribute to the total velocity profile measured in these low - ionization lines .
however , the parameter that we use to characterize the velocity width , ( the velocity width that encompasses the central 90% of the total optical depth seen in the line ) ensures that these low column density clouds do not contribute to the total width . in section [ approach.sec ]
we describe our strategy and data , in section [ results.sec ] we present the results , and in section [ interpretation.sec ] we discuss these results .
finally , we present the conclusions section [ conclusions.sec ] . throughout this paper
we use a value for the hubble constant of @xmath11 , although the distances to some of the things galaxies are derived using direct distance measurements .
for a quantitative comparison between the kinematics of dlas and local galaxies , high quality 21-cm maps are required that are characterized by high signal to noise and high spatial and velocity resolution .
a low noise level is required to sample all regions of the gas disks above the dla column density limit of @xmath12 .
a high spatial resolution is needed to avoid severe overestimation of the velocity profiles due to spatial smoothing ( see * ? ? ?
* ) . since for aperture synthesis interferometry the noise level ( measured as the minimal detectable hi column density per beam ) scales with the inverse of the beam area , these two requirements are very difficult to meet .
the things sample is currently the only sample with sufficient spatial resolution and depth , and contains a large enough number of galaxies such that the local universe galaxy population can be sampled representatively .
we base our analysis on vla observations of a sample of nearby galaxies drawn from things : the hi nearby galaxy survey @xcite .
things provides high angular and velocity resolution data of uniform quality for 34 galaxies spanning a range of hubble types , from dwarf irregular galaxies to massive spirals .
the sample selection is mainly drawn from the spitzer infrared nearby galaxy survey ( sings , * ? ? ?
* ) but excludes targets with distances @xmath1315 mpc , and elliptical galaxies .
in addition , edge on galaxies have been excluded .
the observations and data reduction are discussed in detail in @xcite . for our analysis
, we use the robust - weighted data cubes , which produce typical beam sizes of @xmath14 ( or 200pc for the mean distance of 6.5 mpc to our sample galaxies ) .
we choose the robust - weighted data cubes in order to achieve the highest possible spatial resolution , which is required for making a comparison to dla velocity profiles , which are measured along very narrow sight - lines to background quasars .
the velocity resolution in all cases is equal or better than 5.2 ( which we will smooth further , see below ) .
typical column density sensitivities are @xmath15 in the integrated hi maps / profiles , i.e. , much below the dla column density limit ( @xmath16 ) discussed in this paper .
the different distances among the galaxies in the sample cause the beam to trace different physical scales within galaxies . as discussed in section [ beam.sec ]
, beam smearing causes some increase in the measured velocity widths .
therefore , for the most distant galaxies the velocity width measurements are increased somewhat more than for the nearby ones . however , as we will see , this does not affect our main conclusions .
the things sample is composed of galaxies from optical catalogues , whereas dla systems are purely hi - selected .
( see * ? ? ?
* for a discussion on the selection of things targets ) .
however , in @xcite it was argued that this optical selection does not lead to any biases for this application .
the main argument is the following : @xcite has shown that the properties of hi - selected galaxies are not different from those of optically selected galaxies if compared at the same luminosity . therefore ,
if the weighting of the sample is properly taken into account ( using the luminosity function or hi mass function ) no strong biases should be introduced by using optically selected galaxies .
this fact is also corroborated by the calculation of @xcite , who showed that the redshift number density @xmath17 calculated from hi - selected nearby galaxies is consistent with the values presented in @xcite and in the present paper . another consideration is the distribution of inclinations @xmath18 of the things sample . ideally , the galaxies should be randomly oriented , that is , the distribution of @xmath19 should be consistent with a flat distribution . in practice
, however , given the selection discussed in section [ things.sec ] , the sample has a minor deficiency of edge - on galaxies .
a kolmogorov - smirnov test shows that the null - hypothesis that the inclination distribution is consistent with being random can only be rejected at the 19% level .
because of projection effects , the most inclined galaxies have hi column densities and measurements at the high end of the total distribution of these values in the things sample . on the other hand ,
the cross - section in highly inclined galaxies is very small , meaning that the fractional contribution of these galaxies to the total count of high column densities and high values is minor .
if we assume that things misses all galaxies with inclinations @xmath18 greater than @xmath20 , this would lead to a total incompleteness in the number of galaxies of @xmath21% ( assuming infinitely thin disks ) .
however , since the area @xmath22 of a disk scales as @xmath23 , the total area missed in these highly inclined galaxies is only @xmath24 3% .
we conclude that the contribution from the missing highly inclined galaxies to the dla cross - section is not a significant concern for this analysis .
if the kinematics of the hi change on spatial scales smaller than the beam size , measurement of the gas properties are subject to systematic uncertainties .
this effect is known as beam smearing .
@xcite demonstrate that decreasing the spatial resolution of 21-cm data systematically increases the measured median .
this result was established by convolving a high resolution hi data cube of the large magellanic cloud ( lmc ) with gaussian beams of various widths to simulate data with different spatial resolutions .
the increase in with beam width can be understood in terms of including adjacent pockets of gas that are offset in velocity .
the resulting velocity width measurement of the total area will be greater than the values of the individual hi clouds .
the lmc data has a physical resolution of 15 pc beam@xmath25 and a velocity resolution of 1.6 . in comparison ,
the mean physical resolution of things is 200 pc . given this factor of @xmath26 difference in physical resolution , and employing the same technique used by @xcite
, we estimate that the things velocity width measurements could be overstated by @xmath27 40% . despite this expectation of an increase ,
the velocity dispersion and measurements from the lmc and things data are similar . specifically , the distribution of velocity dispersion values in the lmc peaks at 6 , wheras values derived from the things galaxies range from 6 to 10 , i.e. not larger in all lines - of - sight .
thus , we conclude that beam smearing may have a mild effect on the things data , but is unlikely to cause a major systematic shift in the measured values . of course
the physical scales sampled by the highest resolution 21-cm measurements are still larger than those probed by qso lines - of - sight . in any case , as we will see in section [ results.sec ] , if we have overestimated the velocity widths of galaxies at @xmath1 using the things data this would imply an even stronger discrepancy between the true low redshift velocity widths and the high redshift values of in dlas . we have argued that things provides the most suitable sample of 21-cm data cubes of nearby galaxies to perform the kinematic analysis . however , the number of galaxies in this sample is an order of magnitude smaller than the whisp sample used in @xcite , which provides a very accurate measurement of the local hi column density distribution function . to test whether the things sample contains biases in the hi column densities it samples ,
we first calculate the hi column density distribution function .
this function is defined such that @xmath28 is the number of systems with column density between @xmath29 and @xmath30 over a distance interval @xmath31 .
we use the method described by @xcite , where is calculated as follows : @xmath32 here , @xmath33 is the space density of galaxy @xmath18 , measured via the hi mass function .
for the hi mass function we adopt the most recent hipass measurement as published by @xcite .
the function @xmath34 is a weighting function that takes into account the varying number of galaxies across the full range of @xmath35 , and is calculated by taking the reciprocal of the number of galaxies in the range @xmath36 to @xmath37 , where @xmath38 is taken in this case to be 0.35 dex .
@xmath39 ) is the area function that describes for galaxy @xmath18 the area in @xmath40 corresponding to a column density in the range @xmath41 to @xmath42 . in practice
, this area function is simply calculated by summing for each galaxy the number of pixels in a certain @xmath43 range multiplied by the physical area of a pixel .
finally , @xmath44 converts the number of systems per mpc to that per unit redshift .
figure [ fnthings.fig ] shows the resulting column density distribution function measured from the things data ( solid points ) , compared to the one published by @xcite based on the wsrt maps of the whisp sample ( open points ) .
given the different sample definitions of things and whisp and the different resolutions achieved in both surveys , there is a remarkable agreement between the two data sets , reaffirming our earlier conclusion that selection effects are not an important issue for the things sample , and thus that this sample is well suited for this kind of analysis .
note that we only calculated the points above the dla column density limit of @xmath45 ( corresponding to @xmath46 ) , since this is the limit we apply for our kinematic analyses . in principle , the things hi column density sensitivity would allow us to calculate down to much lower values . some differences between the two samples can be seen at high column densities , above @xmath47 .
this is due to the dearth of high inclination galaxies in things .
although the effect shows up clearly in the logarithmic representation of , we showed in section [ bias.sec ] that the effect of missing edge - on galaxies on the total hi cross section and on the calculation is not important for this analysis . from 21-cm emission - line observations .
the open symbols represent the calculation of @xcite based on the whisp sample , whereas the solid circles represent the calculation presented in this paper based on things data .
error bars include uncertainties in the hi mass function as well as counting statistics and indicate 1@xmath48 uncertainties .
[ fnthings.fig],width=294 ] the zeroth moment of gives the redshift number density @xmath17 of gas above the dla hi column density threshold in the @xmath1 universe .
this quantity simply represents the number of encounters with hi gas above the dla limit that a hypothetical observer would have if they were traveling through the local universe for a total path length of @xmath49 . from the things sample
we find @xmath50 , compared to @xmath51 from the whisp analysis , which again is in excellent agreement .
the standard metric used in dla studies to characterize the profile width of metal absorption lines is , the velocity width that encompasses the central 90% of the total optical depth seen in the line .
this definition assures that weak velocity components that are not good tracers of the large - scale velocity field , as well as statistical fluctuations , are not taken into account . in practice ,
is calculated by taking the cumulative optical depth distribution , and finding the velocities at which this distribution crosses the 5% and 95% levels . in dla studies ,
the data used for this analysis are normally limited to those with high signal to noise , where the peak optical depth exceeds @xmath52 .
these spectra are mostly products of the high resolution echelle spectrometer ( hires ) on the keck telescope and the ultraviolet and visual echelle spectrograph ( uves ) on the very large telescope ( vlt ) .
this signal - to - noise level is calculated on smoothed spectra with typical velocity resolutions of 15 to 20 . in order to make an equitable comparison , we first smooth the things data cubes in velocity with a boxcar filter , to a velocity resolution of approximately 15 to 20 , varying slightly from cube to cube , depending on the original channel width .
next , we calculate normalized cumulative velocity profiles at every spatial pixel in the cube .
using the same strategy as defined above ( just cutting the distribution at the 5% and 95% levels ) does not give good results , because for the lower column densities the hi data typically possess lower s / n than the optical spectra ( of which only the highest s / n examples are regarded ) .
therefore , the cumulative distribution can cross these levels several times , leaving us with no unique definition of . to avoid this problem ,
we first identify the 50% level and travel outward in both directions to find the first instance where the 5% and 95% levels are crossed .
the difference between the two velocities at which this happens is then defined as . in principle , this method could cause an underestimation of , but tests with synthetic spectra show that this effect is minimal .
figure [ mosaic.fig ] shows four examples of the calculations for the things galaxies .
we choose to show the nearly face - on barred spiral ngc925 , the m81-group member ngc3077 know for its extended tidal hi , the isolated nearly edge - on early - type spiral ngc2841 , and the barred nuclear starburst galaxy ngc2903 .
the top row of panels shows the hi column density distribution for these galaxies .
the bottom row of panels shows the distribution of over the regions of these galaxies where the hi column densities exceed the dla threshold of @xmath2 .
the black regions are those where exceeds 100 .
the black spots that can be seen for example in the northern and southern outskirts of ngc2903 are due to noisy spectra , where the calculation is unreliable . in ngc2841 ,
the extended dark regions in the map are associated with high regions and indicate true high values of .
examples of velocity profiles through things galaxies are shown in figure [ profiles.fig ] .
we chose 25 random profiles , with values between 31 and 165 .
this figure gives an idea of quality of the things data and the accuracy of the measurements .
the shaded areas correspond to the inner 90% of the optical depth and define the measurement of .
the number in the upper left corner of each panel shows the value of for each profile . to indicate the reliability of our method for determining by using the first crossing of the 5% and 95% levels from the center
, we also show in parenthesis the values of based on the second crossing of these levels . for most profiles , the cumulative distribution crosses the 5% and 95% levels only once , in which case the definition of is unique . in the few cases where the profiles are more noisy and a second crossing occurs
, it is clear that the measurements based on the first crossings are superior ; measurements based on the second crossing clearly overestimate .
these example profiles show that our method of determining is accurate and not introducing a significant bias . before presenting the results
, we would like to come back to the statement made in the introduction that low ionization metal lines such as fe@xmath10 , si@xmath10 , ni@xmath10 trace the neutral gas only .
the fundamental premise of this paper is that these lines trace the kinematics of the high column density neutral gas , and can therefore be compared directly to local 21-cm measurements .
it is well established that approximately 25% of a spiral galaxy s hydrogen mass is in a layer of ionized gas , known as the diffuse ionized gas ( dig , e.g. * ? ? ?
the scale height of this gas is 0.9 kpc in the milky way galaxy , much larger than that of the neutral gas seen in 21-cm .
if a large fraction of the optical depth seen in low - ionization lines observed in dlas would arise in such a medium , our results might be biased because this medium does not contribute to the profile of the hi 21-cm line .
@xcite present model calculations of ionization corrections in the dig of the milky way , and conclude that elemental abundances in ionized gases might be biased without significant ionization corrections .
however , much more relevant in the context of the present paper are studies directed at dlas specifically since our concern is whether the ionized gas influences the low - ionization metal line measurements of dla profiles .
a main motivation for such studies of ionization effects in dlas is the observation that al@xmath53 is present in dlas with the same radial velocity distribution as that of low ionization lines @xcite .
however , the results of e.g. , @xcite , @xcite and @xcite show conclusively that ionization effect corrections for low - ionization lines observed in dlas are negligible . the direct implication of this is that the velocity width measured in low - ionization lines is not affected significantly by an ionized medium . for systems below the dla
hi column density limit , the situation might be very different @xcite , but these are not considered in the present analysis .
we first use the things data to calculate the probability distribution function of the velocity width of galaxies responsible for high column density hi absorption at @xmath1 .
this function defines the number of systems above a certain column density limit that would be encountered along a random 1 mpc path through the @xmath1 universe , as a function of .
the function is calculated by multiplying the cross sectional area of hi as a function of by the space density of galaxies .
if dlas at @xmath1 arise in galaxies similar to those observed in the local universe , this function would describe the expected distribution of values for dlas . for the details of calculating the probability distribution function
we refer to @xcite .
figure [ w90dens.fig ] shows the result .
the lines show the distributions above column density limits of @xmath54 20.3 , 20.6 , 20.9 , 21.2 , and 21.5 , from top to bottom , respectively . for all column densities above the dla limit of @xmath46 the distribution peaks at @xmath55 , and shows a long tail toward higher values . as one progresses to higher column density cut - offs , the peak of the distribution shifts toward larger values , but the upper tail remains .
we will discuss this further in section [ dv-nhi.sec ] .
immediately obvious from this distribution is that the peak in is much lower than what is typically seen in dlas at higher redshifts .
this is illustrated directly in figure [ histv90.fig ] , where we show on a linear scale the total probability distribution function of for @xmath1 gas above the dla limit ( as a dashed line ) , together with the measured distribution of in dlas at redshifts between @xmath56 and @xmath57 from @xcite .
note that the dashed line is the same as the top solid line in figure [ w90dens.fig ] , but now on a linear scale .
the shape of the two distributions is very similar a sharp peak with a high tail but the location of the peak is approximately a factor two higher for the dlas than for the local galaxies based on things .
this result reaffirms the conclusions of @xcite , which showed that the gas kinematics in local galaxies are inconsistent with the majority of dla sightlines .
@xcite concentrated on the gas kinematics of the large magellenic cloud and we have shown that the same conclusion holds if a more representative sample of local galaxies is considered . whereas the kinematics of hi gas above the dla limit at @xmath1 is dominated by ordered rotation in cold disks , the velocity fields of dla systems at high redshifts must be influenced by other physical processes such as outflows , superwinds , and interactions .
we discuss this in more detail in section [ conclusions.sec ] . .
the galaxy data are weighted such that they represent cross - section selected gas .
the dashed line is the same as the top solid line in figure [ w90dens.fig ] , but now on a linear scale .
[ histv90.fig],width=219 ] it is interesting to investigate how the velocity spread of the hi gas depends on the hi column density range considered .
the contours in figure [ w-n.fig ] show the probability distribution of hi cross section in the - plane .
the things data were used to calculate the cross sectional area contributed by each element @xmath58 on a fine grid in the @xmath29-@xmath59 plane .
the hi mass function has again been used to assign weights to individual galaxies .
the weighting scheme is similar to that used for the calculation of , which is expressed in equation [ fn.eq ] .
the points in this figure are the dla data from @xcite , over the redshift range @xmath56 to @xmath57 .
distribution is at @xmath60 , and increases slowly toward larger hi column densities ; 2 ) large values of occur mostly in regions of galaxies with relatively low hi column densities .
[ w-n.fig],width=294 ] there are two notable effects visible in this diagram . first , the peak of the @xmath1 distribution occurs at values of around 30 , and increases slowly toward larger hi column densities , as already seen from figure [ w90dens.fig ] . for gaussian profiles ,
the value of = 30 translates to a velocity dispersion of @xmath61 , which is a typical value for the gas disks in spiral galaxies .
observationally the line - of - sight velocity dispersion ranges from @xmath62 in the outer parts to @xmath63 in the center ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
from the things data it is found that the dispersions in the regions away from obvious hi shells is @xmath64 , averaged over the galaxy disks @xcite .
the increase of toward larger values is consistent with the observation that the hi velocity dispersion in gas disks typically decreases with galactocentric radius ( e.g. , * ? ? ?
* ; * ? ? ?
* ) , and the fact that radial hi column density profiles usually show a decline toward larger radii @xcite .
the other notable feature in figure [ w-n.fig ] is that the larger values of are mostly associated with the lower column densities . to understand this
, we need to investigate further the origin of the high - end tail in the distribution .
there are several components contributing to this tail .
one of the most obvious contributions is that of pixels near the centers of galaxies , where the gradient in the rotation curve is large . a line of sight through these regions is likely to encounter gas over a large range in velocity .
the steepest gradients are seen in earlier type spiral galaxies ( e.g. , * ? ? ?
* ) , where typically the hi densities are depressed in the center because of the conversion of hi to molecular hydrogen ( e.g. , * ? ? ?
these factors combined cause a large contribution of low , high points . to illustrate that the larger values of are more likely to occur near the centers of galaxies we constructed figure [ w-r.fig ] , which shows the probability distribution of hi cross section in the @xmath65- plane , where @xmath65 is the distance from the center of the galaxy .
it is important to note that this diagram is also made in such a way that it represents probabilities for @xmath65- pairs to occur for random sight - lines through the local universe .
therefore , @xmath65 can be regarded as an
impact parameter " , similar to that defined in qso absorption line studies as the projected distance between the position of a background quasar and that of the center of a galaxy giving rise to an absorption line .
clearly , there is an upper envelope in the distribution : the maximum seen at any value of the impact parameter @xmath65 decreases toward larger @xmath65 . ,
the position from the center of the galaxy .
contours are calculated from the things data , and are drawn at probability levels of 1 , 5 , 15 , 30 , 60 , and 90% .
larger values of are more likely to occur at small distances from the center of a galaxy
. please note that the small scale structure in the 1% contour is due to noise .
[ w-r.fig],width=294 ] the simple fact that the lower pixels show noisier spectra causes another important contribution of high points . in
lower signal - to - noise spectra , noise peaks more easily contribute to the central 90% flux contained in the spectra .
this fact can be seen in figure [ profiles.fig ] , where we showed random velocity profiles through things galaxies . for the lowest values ,
the profile is clearly dominated by a single component . for @xmath66 ,
the profile typically shows two components , or one strong component with a lower flux tail .
this continues to be true for even higher profiles , but here irregular profiles seem to get increasingly important . comparing the contours and the points in figure [ w-n.fig ] , we note that there is some indication for the dla data to show the same trend as the things local galaxies data : higher values of are more likely to occur at lower hicolumn densities .
so far , our analysis has concentrated on the parameter , but it is interesting also to consider the shape of the profiles in more detail . in the previous section we noted that for sightlines where @xmath67 , the profile typically shows two components .
in fact , it appears that the individual line components have widths within the main peak of the distribution in figure [ w90dens.fig ] , and the high - tail is mostly due to line splitting and multiple components .
interestingly , this same behavior is seen in the metal absorption line profiles of dlas .
for example , fig 1 of @xcite shows 17 examples of velocity profiles of low - ionization transitions from dlas .
the narrow profiles are characterized by a single component with @xmath68 , very similar to what we observe in the local universe for most of the sightlines through galaxy disks ( cf .
figure [ histv90.fig ] ) . for dla profiles with larger widths ,
a shoulder or tail becomes progressively more important , and a pronounced second peak starts to separate from the main component for values exceeding @xmath69 .
apparently , the second peak in figure[histv90.fig ] beyond @xmath70 is mostly due to profiles clearly showing multiple components , whereas the main peak corresponds mostly to profiles displaying a single component with a weaker tail .
the profiles with multiple components are seen at @xmath1 as well ( see figure[profiles.fig ] ) , but are very rare .
when we consider just the profiles that correspond to the main peaks in figure[histv90.fig ] , we see that at @xmath1 they are mainly symmetric , single component profiles , whereas the dlas at higher redshift typically show a tail .
this characteristic is important for understanding the mechanisms repsonsible for shaping the profiles and will be discussed more in section [ interpretation.sec ] .
there is accumulating evidence for the existence of a relation between dla ( and sub - dla ) velocity spread and metallicity at high redshifts @xcite .
the interpretation of this correlation , assuming that is correlated with mass , is that the well - known luminosity - metallicity relation observed in local galaxies ( e.g. , * ? ? ?
* ) was already in place at redshifts @xmath57 all the way down to @xmath71 . as a proxy for luminosity ( or galaxy mass ) , @xcite use measurements of .
the authors argue that measured in low ionization lines is dominated by motions governed by gravity , irrespective of whether the gas is following the rotation of the galaxy or related to infall or outflow of gas or to merging of galaxy sub - clumps . in principle , a large collection of sightlines through the gaseous regions of galaxies should show a relation between depth of the potential well and .
only at @xmath1 are we in the situation that we can test whether the assumption that is a statistical measurement of galaxy mass or luminosity is valid .
the result is shown in figure [ condprobvm.fig ] .
here we show the conditional probability distribution of as a function of absolute @xmath72-band magnitude .
this figure a created by first calculating the bivariate distribution of cross - section in the @xmath73 , plane , and then normalizing at each @xmath73 the distribution function of cross - section as a function of .
obviously , the likelihood of finding a high value goes up only marginally toward brighter galaxies : the median does not rise significantly from low luminosity to high luminosity , but for more luminous galaxies the high tail becomes more pronounced .
therefore , in a statistical sense , is a very weak proxy for host galaxy mass . at any absolute magnitude , the median value of is @xmath74 , which for a gaussian profile translates to a velocity dispersion of 9 . in conclusion , at @xmath1
the velocity profile measured along a thin line of sight at a random position through a galaxy gas disk is a very weak indicator of the galaxy mass or luminosity .
although we have presently no means of testing this indicator at high redshift , the local relation does not lend much support to the suggestion that the observed metallicity relation at high redshift can be interpreted as a mass metallicity relation .
-band magnitude , @xmath75 .
the thick solid line represents the median of the velocity width distribution at each value of @xmath75 .
the thin solid lines show the 25th and 75th percentile distribution , the dotted lines show the 10th en 90th percentile , and the dashed line shows the 99th percentile distribution .
the curves indicate global trends , small scale structure in the curves is due to small number statistics .
the median remains constant around @xmath76 .
[ condprobvm.fig],width=294 ]
modeling of dla velocity profiles already suggested that high redshift dlas can not simply arise in rotationally supported cold gas disks such as those observed in local galaxies @xcite . " in this paper we have demonstrated empirically that the typical velocity profiles of @xmath1 hi gas above the dla limit have lower widths than those seen in dlas .
this result is fully consistent with that of @xcite , based on the analysis of a single low mass galaxy . how can these large velocity widths in dlas be interpreted ?
it is well established that most actively star - forming high - redshift galaxies show large - scale outflows , with velocities of several 100 up to 1000 .
several authors have suggested that these outflows could be related to a large fraction of the dla cross - section at these redshifts @xcite .
@xcite argued that the starburst accompanying the formation of dwarf galaxies blows out most of the available gas , which results in a weakly collimated wind .
once the starburst has subsided , the ( largely ionized ) ejected gas recombines and shows hi column densities typical of dlas .
@xcite built on these results and showed that the incidence rate of dlas is in agreement with the wind scenario .
he concluded that both disk gas and winds contribute to the high redshift dlas cross section . at @xmath1 ,
less extreme examples of star - formation driven blow - out can be found in hi shells , which are ubiquitous in high resolution 21-cm maps of nearby galaxies ( see e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
radiative and kinematic pressure of stellar winds and supernovae from young massive stars in ob associations are thought to lie at the origin of these expanding shells .
@xcite looked specifically at the issue of whether measurements are affected by these shells in the 30 doradus region in the lmc .
it was found that velocity widths of sight - lines through the hi bubbles were only enhanced by approximately 15 with respect to the average lmc value .
this result may not come as a surprise , because the typical expansion velocities of the shells range from 10 to 30 in the lmc @xcite .
similar numbers are found for other nearby galaxies ( e.g. , * ? ? ?
not only is the velocity width small in shell regions , the area filling factor of hi shells is also small .
@xcite quote a covering factor of @xmath77% for regions above the dla limit .
if the physics of supernova feedback at high redshifts is similar to what is seen in local galaxies , the above results imply that hi shells are unlikely to explain the high velocity widths observed in dlas . in order to explain the kinematics in high-@xmath9 dlas with outflows
, one has to invoke more extreme outflow scenarios , such as that seen in star - burst galaxies . to investigate this further , we focus on m82 , the prototypical starburst galaxy in the nearby universe .
this galaxy displays a well - documented example of the superwind phenomenon , driven by the molecular driven central starburst , which is very prominent in x - ray and h@xmath0 emission ( e.g. , * ? ? ?
* ; * ? ? ?
m82 is also a member of the m81 group of galaxies and experiences a strong tidal interaction with its neighbors m81 and ngc 3077 @xcite .
this interaction results in several hi tidal features in the close vicinity of m82 . at a distance of only @xmath78 mpc
, the galaxy can be observed in great detail , and therefore has been the object of studies at wavelengths ranging from the x - rays to the radio ( see * ? ? ? * for references ) .
here , we investigate the hi distribution in m82 , making use of the original c - array and d - array very large array ( vla ) data of @xcite , augmented with b - array data from the vla archive .
the final hi column density map is very similar to the one published in @xcite . and
@xmath79 are indicated in the north - east tail and in the central regions , respectively .
[ m82.fig],width=328 ] we use the same procedure as described for the things galaxies to calculate values from the m82 data for the regions of the hi gas above the dla column density limit .
figure [ m82.fig ] shows the distribution of in m82 .
we find that the median from these data is @xmath80 .
very high values of @xmath79 are seen in the central region over an area of several kpc@xmath81 . along the minor axis of m82
these high velocity widths are attributed to interactions of m82 s gaseous medium with the superwind @xcite .
interestingly , similarly high velocity widths along the minor axis are observed in co @xcite .
the tidal features in m82 also show high velocity widths .
the south - west tail that connects up to m81 is not well seen in our map , but the north - east tail shows values around 100 . obviously , m82 s high inclination to the line - of - sight ( @xmath82 ) will contribute to the high median observed in this galaxy .
however , the gaseous regions in this galaxy that are clearly identified as either tidal features or those related to the superwind display similarly high velocity spreads .
as we investigated only a single starburst galaxy it is difficult to draw quantitative conclusions on the basis of these data .
however , we argue that m82 clearly demonstrates that the superwind phenomenon and tidal interactions , which are both more frequent at higher redshifts , could go a long way in explaining the high values observed in dlas . both winds and tidal features could contribute to the main peak of the dla distribution in figure [ histv90.fig ] , as well as to the high bump . in this context
it is interesting to mention the results of @xcite , who find that the equivalent width of mgii absorbers is anti - correlated with the mass of the halo that hosts them .
these authors interpret this as evidence that the mgii absorbers are not virialized in the gaseous haloes of the galaxies , but , instead , are produced by superwinds . a fair fraction of these mgii systems are dlas .
@xcite showed that 35% of the mgii systems with ew ( mgii @xmath83 , feii @xmath84 ) @xmath13 0.5 have hi column densities above the dla threshold of @xmath16 .
additional evidence for a contribution to the dla cross - section from tidal gas in galaxy groups like m81 , comes from the recent observations of a much smaller sample of strong mgii systems by @xcite .
these authors find that these very strong intermediate redshift mgii absorption systems often arise in pairs or small groups of galaxies .
an estimated 60% of these systems are dlas .
@xcite argues that the likely origin of the high equivalent width mgii absorption is kinematically disturbed gas around interacting galaxies .
the different peaks in the distribution of velocity width in the low and high redshift samples in figure [ histv90.fig ] could be interpreted as being due to an increase in the number of interacting systems with redshift .
@xcite use n - boby simulations to show that the number of interacting systems ( i.e. major mergers ) scales as @xmath85 .
thus , we would expect a significant increase in lines - of - sight with large velocity widths at higher redshifts .
@xcite have argued that irregular protogalactic clumps could ( partly ) produce the large velocity widths of dlas at high redshift .
perhaps the most closely related examples of such clouds in the nearby universe are galactic high velocity clouds ( hvcs , * ? ? ?
the hi column densities in the clouds themselves very rarely exceed the dla limit , but it is conceivable that along a sight - line through the gas disk , hvcs could contribute sufficiently to the integrated optical depth to influence the measurements .
several authors have noted the similarity between the properties of galactic hvcs and those of mgii absorption systems ( e.g. * ? ? ?
* ; * ? ? ?
* ) , of which dlas are a subset . because of the unknown distances to the vast majority of the galactic hvcs , their physical properties are difficult to ascertain . in this respect ,
the recent identification of hvcs around m31 is particularly interesting @xcite .
synthesis folow - up observations by @xcite show that these clouds have typical sizes of 1 kpc and column densities in the range @xmath86 to @xmath87 , i.e. , below the dla threshold .
if these hvcs really are the relics of galaxy formation , it is likely that they present a much higher cross - section at higher redshifts when galaxy formation was more actively taking place . as such
, they could have a significant effect on the measurements in high-@xmath9 dlas .
we note that this component is closely related to ( and perhaps partly a variant of ) the tidal streams and winds discussed above . also interesting in this context
are the recent very deep 21-cm observations of ngc 2403 @xcite and ngc 891 @xcite .
these observations reveal low column density extended hi haloes , which typically contain 10 to 30% of the total hi mass of the galaxies are unlikely to add much to the velocity widths measured in the low - ionization lines , for reasons explained in section [ lowion.sec ] ] .
again , the hi column densities in the haloes are too low to cause dla absorption by themselves , but along sight - lines through the gas disks , the haloes can have a significant effect on measurements .
these haloes show structures such as clouds and filaments , which resemble the properties of galactic hvcs .
detailed modeling of the hi haloes @xcite shows that their origin lies partly in a fountain mechanism @xcite , as demonstrated by the close link between the extra - planar hi and h@xmath0 emission ( * ? ? ?
* and references therein ) .
the driving force behind this fountain is star formation in the disk .
the cosmic star formation rate density was much higher in the past , up to a factor 20 higher at @xmath88 compared to the present epoch @xcite .
the increase in the cosmic neutral hydrogen mass density is much less dramatic , approximately a factor 2 over same redshifts range @xcite .
these two results combined imply that the energy input into the cold gaseous medium per unit gas mass was much higher in the past .
this may lead to a relatively high contribution of fountain gas to the kinematics of cold gas at these redshifts .
this high velocity gas would probably be characterized by profiles very similar to the ones observed in dlas displaying a main peak and a weaker tail . in passing ,
we note that the higher energy input into the gas disk would also lead to an increase in the gas turbulence , which also causes larger velocity widths . in order to fully explain the properties of the gaseous haloes , @xcite speculate that accretion from the surrounding intergalactic medium
is also required .
this cold mode " of gas accretion is increasingly more important at higher redshifts in the smoothed particle hydrodynamics simulations of @xcite , indicating that this halo gas could play an important role in the gas kinematics of dlas .
in this paper we have tested whether the kinematics of damped ly@xmath0 absorbers at high redshift are consistent with the idea that these systems arise in the gas disks of galaxies such as those in the local universe .
we have examined the gas kinematics of local galaxies using the high quality hi 21-cm data from the hi nearby galaxies survey ( things ) . to characterize the velocity widths
, we adopt the parameter , the width that encompasses the central 90% of the total flux seen along any sightline through the galaxies .
the same parameter is used for dlas to measure the width of the low - ionization lines , which are believed to trace the kinematics of the neutral gas .
we can use the @xmath1 things data to calculate in detail how depends on hi column density , and we find that larger values are more likely to occur at low hi column densities . a similar ( but weak )
trend is seen for the dlas , albeit at at higher values of .
we also investigate how depends on impact parameter from the center of a galaxy and on galaxy luminosity . however , at present it is difficult to make a quantitative comparison between these calculations and the dla data because of the very few identifications of dla galaxies in the literature . for a proper understanding of the kinematics and other properties of dlas ,
it is important that more dla galaxies be identified at low and high redshifts .
the most important result is that the median at @xmath1 is approximately 30 , a factor of two lower than what is seen in dlas at redshifts @xmath89 .
we discuss several processes that could increase the velocity widths in dlas , and conclude that tidal tails , superwinds , accretion , as well as fountain gas are all likely to contribute to the kinematics of dlas .
each of these processes are enhanced at higher redshifts , where galaxy interactions are more common , superwinds are more frequently observed , ` cold mode ' accretion is more efficient , and the star formation rate density is higher .
we thank n. bouch and c. proux for their comments and the anonymous referee for a constructive report , which helped improving the presentation of this paper .
j. ott and a. weiss are thanked for their help with the reduction of the m82 vla data used in this analysis .
eb gratefully acknowledges financial support through an eu marie curie international reintegration grant ( contract no .
mirg - ct-6 - 2005 - 013556 .
the national radio astronomy observatory is a facility of the national science foundation operated under cooperative agreement by associated universities , inc . | we demonstrate in this paper that the velocity widths of the neutral gas in damped ly@xmath0 ( dla ) systems are inconsistent with these systems originating in gas disks of galaxies similar to those seen in the local universe .
we examine the gas kinematics of local galaxies using the high quality hi 21-cm data from the hi nearby galaxies survey ( things ) and make a comparison with the velocity profiles measured in the low - ionization metal lines observed in dlas at high redshifts .
the median velocity width of @xmath1 hi gas above the dla column density limit of @xmath2 is approximately 30 , whereas the typical value in dlas is a factor of two higher .
we argue that the gas kinematics at higher redshifts are increasingly influenced by gas that is not participating in ordered rotation in cold disks , but is more likely associated with tidal gas related to galaxy interactions or processes such as superwinds and outflows .
an analysis of the hi in the local interacting star - burst galaxy m82 shows that the velocity widths in this galaxy are indeed similar to what is seen in dlas . |
gene regulatory networks describe the effective interactions between genes .
the activity of a gene , i.e. , its current rate of being transcribed into rna molecules , can have effects on the activity levels of other genes , which will as a result become up- or down- regulated .
the sum of all up- and down- regulation relations in the whole genome is the gene regulatory network .
the complete knowledge of the gene network would reveal a large portion of an understanding of life . however , this goal is far from being achieved . with present dna - chip technology
it is possible to measure the transcription rates at a given point in time of an entire genome , but even these technologies only allow a glimpse on the structure of the underlying network , due to the underdeterminedness of the problem @xcite .
this situation got the physics community interested , to statistically characterize the available data and to ( crudely ) estimate the structure of the complex networks governing gene dynamics .
a step toward an identification of potential gene interaction networks is to identify and quantify meaningful _ statistical _ indicators of gene cooperative behavior , which is the main purpose of the present work .
the idea is that fluctuations of gene expressions over time , e.g. , during a cell - cycle , can be considered as an output of an interacting gene collective forming a structured network .
the hope is that a network structure estimate can be inferred from statistical properties .
at least it should be possible to statistically characterize the types of potential candidate networks .
we consider the time - course expression data @xmath1 for the genome of yeast _ s. cerevisiae _ @xcite .
we determine some statistical indicators of collective dynamical behavior of genes , such as the @xmath0-exponential fit of the cumulative distribution , a ranking distribution and a mean - variance analysis of differential gene expressions .
we construct and estimate the expression - correlation network from time increments of expression data and analyse clusters and spanning trees .
we identify biological function of genes with use of yeast database @xcite .
we find that the resulting , correlation based clusters match considerably well with specific biological functions of genes in the cell .
the genome - wide gene expression data in @xcite are given in the form of a matrix @xmath1 in which every row represents one of @xmath2 yeast genes and each column contains the time evolution of gene expression of that gene @xmath3 .
gene expressions are measured at 17 time points , taken every 10 minutes which covers approximately two full cell - cycles .
we first properly normalize the gene expressions for each of the 17 measurements separately by dividing each gene expression value by the average value of gene expression for that corresponding column . in order to avoid systematic trends in the time series we use _ differential _ expression data defined as @xmath4 for each gene @xmath3 .
we determine the cumulative distribution @xmath5 for each time - interval separately and also all measurements ( all entries in matrix ) .
the results are given in fig .
1 a. this distribution can be fitted to a @xmath0-exponential form @xcite , @xmath6^\frac{1}{1-q } \ ; \ \ q \neq 1 \quad , \label{pdf}\ ] ] where @xmath0 represents the non - extensivity parameter .
the fitted values of @xmath0 for the various time - intervals are in the range @xmath7 .
the average over all times yields @xmath8 , potentially indicating a non - trivial collective behavior of genes along the cell - cycle . [
cols="^,^ " , ] for @xmath9 the functional clusters and branches remain for a while before they gradually disappear in the noise .
this is in agreement with the observed character of the correlation distributions in fig . 2 , where smaller deviations from a gaussian distribution are found for @xmath10 .
in conclusion , we made several observations about the statistical nature of gene expression data which seem to suggest that at least a significant fraction of genes is up / down regulated in a highly collective manner .
indicators pointing in this direction are : ( _ i _ ) the cumulative distribution of differential gene expressions can be fitted to @xmath0-exponentials , with a non - trivial @xmath11 ; ( _ ii _ ) an approximate zipf s law holds in the ordering distribution of differential expressions ; ( _ iii _ ) an almost linear mean variance dependence with @xmath12 signals tightly driven dynamics ; ( _ iv _ ) the correlation matrix element distributions are non - gaussian and non - poisson and finally , ( _ v _ ) even crude correlation coefficient network displays the emergence of clusters and functional branches in minimum spanning trees , which seem to be biologically relevant .
s. tavazoie , j.d .
hughes , m.j .
campbell , r.j .
cho and g.m .
church , nature genet . * 22 * , 281 ( 1999 ) ; w. wang , j.m .
cherry , d. botstein and h. li , proc .
natl . acad .
sci . * 99 * , 16893 ( 2002 ) ; k. rho , h. jeong and b. kahng , cond - mat/0301110 ; f. li , t. long , y. lu , q. ouyang and c. tang , proc . natl .
* 101 * , 4781 ( 2004 ) ; d. balcan and a. erzan , eur .
j. b * 38 * , 253 ( 2004 ) . in the available data - bases of yeast ,
apart from genes with unknown functions , often multiple functions can be assigned to a single gene . to keep one color of each gene we select
the first listed function in the database .
we also have grouped several fuctions related to protein synthesis to a single one in order to keep the number of colors visually distinguishable . | we analyze gene expression time - series data of yeast ( _ s . cerevisiae _ ) measured along two full cell - cycles @xcite .
we quantify these data by using @xmath0-exponentials , gene expression ranking and a temporal mean - variance analysis .
we construct gene interaction networks based on correlation coefficients and study the formation of the corresponding giant components and minimum spanning trees . by coloring genes according to their cell function
we find functional clusters in the correlation networks and functional branches in the associated trees .
our results suggest that a percolation point of functional clusters can be identified on these gene expression correlation networks .
+ pacs numbers : 87.10.+e , 89.75.-k , 89.75.hc + keywords : spanning trees , functional clustering , q - statistics , ranking distribution |
supersymmetry ( susy ) provides an elegant solution to the gauge hierarchy problem , and grand unified theories ( guts ) give us a simple understanding of the quantum numbers of the standard model ( sm ) fermions .
in addition , the success of gauge coupling unification in the minimal supersymmetric standard model ( mssm ) strongly supports the possibility of the susy gut .
other appealing features include that the electroweak symmetry breaking is induced by radiative corrections due to the large top quark yukawa coupling , and that tiny neutrino masses can be naturally explained by the see - saw mechanism
. therefore , susy gut is one of the most promising candidates that describe the known fundamental interactions in nature except gravity .
however , there are problems in the 4-dimensional susy gut : the doublet - triplet splitting problem , the proton decay problem , the fermion mass problem , and the gut gauge symmetry breaking mechanism . during the last few years
, orbifold guts have been studied extensively @xcite .
the main point is that the supersymmetric gut models exist in 5 or 6 dimensional space - time , and they are broken down to 4-dimensional @xmath0 supersymmetric sm like models for the zero modes through compactification on various orbifolds .
this is because the orbifold parity projects out the zero modes of some components in vector multiplet and hypermultiplets .
the gut gauge symmetry breaking problem , the doublet - triplet splitting problem and the fermion mass problem have been solved elegantly by orbifold projections .
the proton decay problem can be solved because we can define a continuous @xmath9 symmetry .
other interesting phenomenology , like flavour symmetry from the @xmath10 symmetry , gauge - fermion unification , gauge - higgs unification , and gauge - yukawa unification , have also been studied @xcite .
on the other hand , deconstruction was proposed about three year ago to latticize the gauge theories in higher dimensions @xcite .
the idea of deconstruction is interesting because it gives a uv completion to the higher dimensional theories .
applying this idea to orbifold susy guts , we are able to construct interesting 4-dimensional susy guts where the problems in the usual 4-dimensional models can be solved , in other words , the merits of orbifold guts can be preserved @xcite .
in addition , inspired by deconstruction of orbifold susy guts , we can construct new models which can not be obtained from orbifold guts while still retain the nice properties in orbifold models .
deconstruction of the orbifold @xmath11 models were discussed in ref .
@xcite . in these models
the doublet - triplet splitting problem and the proton decay problem can be solved .
deconstruction may also provide insight into fermion masses and mixings @xcite .
since the number of fields in these models is finite , the corrections to gauge couplings can be reliably calculated , and there exist threshold corrections to the differential runnings of the gauge couplings @xcite .
elements of deconstruction can be found in earlier papers @xcite . in this paper
, we construct 4-dimensional @xmath0 supersymmetric @xmath1 models inspired by deconstruction of the 5-dimensional @xmath0 supersymmetric orbifold @xmath1 models and high dimensional non - supersymmetric @xmath1 models with wilson line gauge symmetry breaking .
we study @xmath2 models with bi - fundamental link fields where the gauge symmetry can be broken down to the pati - salam ( ps ) , @xmath3 , flipped @xmath4 or the sm like gauge symmetry .
however , we need to fine - tune the superpotential in the models , in order to have the doublet - triplet splitting . in addition , we propose an @xmath12 model where the gauge symmetry can be broken down to ps gauge symmetry , and an @xmath6 model with bi - spinor link fields ( @xmath13 ) and ( @xmath14 ) , in which the gauge symmetry is broken down to the flipped @xmath15 gauge symmetry . in these models
, the gauge symmetry can be further broken down to the sm gauge symmetry , and the doublet - triplet splitting is naturally realized through the missing partner mechanism .
the proton decay due to dimension-5 operators is thus negligible and the proton lifetime due to dimension-6 operators is well above the current experimental bound because the gut scale is at least a few times @xmath16 gev .
therefore , there is no proton decay problem .
we also discuss the gauge coupling unification in these two models .
furthermore , we briefly comment on the interesting variation models with gauge groups @xmath7 and @xmath17 where the proton decay problem can be solved , and the @xmath6 model with bi - spinor link fields in which the gauge symmetry is broken down to the @xmath3 gauge symmetry .
we first give a brief review of the orbifold susy guts and the non - susy guts with wilson line gauge symmetry breaking , and deconstruction of both types of models in section [ sec : review ] .
we then discuss the models where the gauge symmetry can be broken down to the diagonal ps gauge symmetry in section [ sec : patsal ] .
we comment on the models inspired by deconstructions of non - susy guts with wilson line gauge symmetry breaking in section [ sec : noncom ] . in section [ sec : bispinor ] , we consider gauge symmetry breaking with bi - spinor link fields .
we summarize our results in section [ sec : conclu ] .
in the 5-dimensional orbifold susy guts , the 5-dimensional manifold is factorized into the product of ordinary 4-dimensional minkowski space - time @xmath18 and the orbifold @xmath19 .
the corresponding coordinates are @xmath20 ( @xmath21 ) and @xmath22 .
the radius for the fifth dimension is @xmath10 .
the orbifold @xmath19 is obtained by @xmath23 moduloing the equivalent class @xmath24 where @xmath25 .
there are two fixed points , @xmath26 and @xmath27 .
the @xmath0 supersymmetric theory in 5-dimension have 8 real supercharges , corresponding to @xmath28 supersymmetry in 4-dimension . in terms of the physical degrees of freedom ,
the vector multiplet contains a vector boson @xmath29 with @xmath30 , two weyl gauginos @xmath31 , and a real scalar @xmath32 . in the 4-dimensional @xmath0 supersymmetry language
, it contains a vector multiplet @xmath33 and a chiral multiplet @xmath34 which transform in the adjoint representation of group @xmath35 .
the 5-dimensional hypermultiplet consists of two complex scalars @xmath36 and @xmath37 , and a dirac fermion @xmath38 .
it can be decomposed into two chiral mupltiplets @xmath39 and @xmath40 , which are in the conjugate representations of each other under the gauge group .
the general action for the group @xmath35 gauge fields and their couplings to the bulk hypermultiplet @xmath41 is @xcite @xmath42 \nonumber\\ & & + \int{d^5x } \left [ \int{d^4\theta } \left ( { \phi}^c e^v { \bar \phi}^c + { \bar \phi } e^{-v } \phi
\right ) \right.\nonumber\\&&\left .
+ \int{d^2\theta } \left ( { \phi}^c ( \partial_5 -{1\over { \sqrt 2 } } \sigma ) \phi + { \rm h. c. } \right)\right]~.~\,\end{aligned}\ ] ] under the parity operator @xmath43 , the vector multiplet transforms as @xmath44 @xmath45 for the hypermultiplet @xmath41 and @xmath46 , we have @xmath47 @xmath48 where @xmath49 is @xmath50 , @xmath51 and @xmath52 are respectively the numbers of the fundamental index and anti - fundamental index for the bulk multiplet @xmath41 under the bulk gauge group @xmath35 . for example , if @xmath35 is an @xmath53 group , for fundamental representation , @xmath54 , @xmath55 , and for adjoint representation , @xmath54 , @xmath56 . moreover , the transformation properties for the vector multiplet and hypermultiplets under @xmath57 are the same as those under @xmath43 .
for @xmath58 , to break the @xmath11 gauge symmetry , we choose the following @xmath59 matrix representations for the parity operators @xmath43 and @xmath57 @xmath60 under the @xmath57 parity , the gauge generators @xmath61 ( @xmath62 , @xmath63 , ... , @xmath64 ) for
@xmath11 are separated into two sets : @xmath65 are the generators for the sm gauge group , and @xmath66 are the generators for the broken gauge group @xmath67 @xmath68 the zero modes of the @xmath69 gauge bosons are projected out , thus , the 5-dimensional @xmath0 supersymmetric @xmath11 gauge symmetry is broken down to the 4-dimensional @xmath0 supersymmetric sm gauge symmetry for the zero modes .
for the zero modes and kk modes , the 4-dimensional @xmath0 supersymmetry is preserved on the 3-branes at the fixed points , and only the sm gauge symmetry is preserved on the 3-brane at @xmath27 . for @xmath70
, the generators @xmath61 of @xmath1 are imaginary antisymmetric @xmath71 matrices . in terms of the @xmath72 identity matrix @xmath73 and
the pauli matrices @xmath74 , they can be written as tensor products of @xmath72 and @xmath75 matrices , @xmath76 and @xmath77 as a complete set , where @xmath78 and @xmath79 are the @xmath80 real anti - symmetric and symmetric matrices .
the generators of the @xmath81 are & & _ 0 a_3 , _ 0 a_2 , _ 0 a_x + & & _ 2 s_3 , _ 2 s_2 , _ 2 s_x , the generators for flipped @xmath15 are & & _ 0 a_3 , _ 0 a_2 , _ 1 a_x + & & _ 2 s_3 , _ 2 s_2 , _
3 a_x , and the generators for @xmath82 are @xmath83 where @xmath84 and @xmath85 are respectively the diagonal blocks of @xmath78 and @xmath79 that have indices 1 , 2 , and 3 , while the diagonal blocks @xmath86 and @xmath87 have indices 4 and 5 . @xmath88 and @xmath89 are the off diagonal blocks of @xmath78 and @xmath79 .
we choose the @xmath90 matrix for @xmath43 as @xmath91 to break the @xmath1 down to @xmath3 , we choose @xmath92 to break the @xmath1 down to flipped @xmath15 , we choose @xmath93 and to break the @xmath1 down to the ps gauge symmetry , we choose @xmath94 for the zero modes , the 5-dimensional @xmath0 supersymmetric @xmath1 gauge symmetry is broken down to the 4-dimensional @xmath0 supersymmetric @xmath3 , flipped @xmath15 and the ps gauge symmetries . including the kk modes , the 3-branes at the fixed points
preserve the 4-dimensional @xmath0 supersymmetry , and the gauge symmetry on the 3-brane at @xmath27 is @xmath3 , flipped @xmath15 and the ps gauge symmetries , for different choices of @xmath57 .
we can also break the @xmath1 down to the sm like ( @xmath95 ) gauge symmetry , by choosing the following matrix representations for @xmath43 and @xmath57 @xmath96 or @xmath97 or @xmath98 other than the orbifold guts , wilson line gauge symmetry breaking can be applied to break high dimensional gauge symmetries @xcite . because supersymmetry can not be broken with this mechanism , only non - supersymmetric models are considered .
first , let us consider the 5-dimensional space - time @xmath99 with coordinates @xmath20 , and @xmath22 .
the radius for the fifth dimension is @xmath10 .
the gauge fields are denoted as @xmath100 where @xmath30 . because @xmath101 belongs to the fundamental group of @xmath23
, we can define a @xmath101 parity operator @xmath102 for a generic bulk multiplet @xmath103 @xmath104 where @xmath105 and @xmath106 . for a @xmath11 model , if we choose @xmath102 to be equal to @xmath57 in eq .
( [ eq : pppsu5 ] ) , the @xmath11 gauge symmetry is broken down to the sm gauge symmetry for the zero modes . for a @xmath1 model , if we choose @xmath102 the same as @xmath57 in eqs .
( [ eq : psu5 ] ) , ( [ eq : pfsu5 ] ) or ( [ eq : pps ] ) , the @xmath1 gauge symmetry is broken down to the @xmath3 , flipped @xmath4 , or the ps gauge symmetry respectively , for the zero modes .
similarly , we can consider the 6-dimensional space - time @xmath107 , with coordinates @xmath20 , @xmath22 , and @xmath108 .
the radii for the fifth and sixth dimensions are @xmath109 and @xmath110 respectively .
we can define two @xmath101 parity operators @xmath102 and @xmath111 for a generic bulk multiplet @xmath112 @xmath113 @xmath114 where @xmath115 , @xmath116 , @xmath106 , and @xmath117 . by choosing the following @xmath102 and @xmath111 exactly the same as @xmath43 and @xmath57 in eqs .
( [ eq : psmx1 ] ) , ( [ eq : psmx2 ] ) or ( [ eq : psmx3 ] ) , we can break the @xmath1 down to the sm like gauge symmetry .
inspired by the high dimensional non - supersymmetric guts with wilson line gauge symmetry breaking , we can construct the 4-dimensional @xmath0 supersymmetric @xmath118 models where the gauge symmetry @xmath118 can be broken down to a diagonal subgroup of @xmath35 or the sm like gauge symmetry .
susy gut models with wilson line gauge symmetry breaking have the problem that the 5-dimensional @xmath0 supersymmetry or the 6-dimensional @xmath28 supersymmetry can not be broken down to the 4-dimensional @xmath0 supersymmetry .
however , we can start from the 4-dimensional @xmath0 supersymmetry , which is exactly the advantage of the 4-dimensional supersymmetric gut models inspired by the deconstructions . for the deconstruction of the orbifold susy guts , we break the @xmath119 to a diagonal @xmath120 as the following .
suppose there are @xmath121 nodes , and the gauge symmetry on the first @xmath122 nodes is @xmath35 while the gauge symmetry on the last node is @xmath123 .
the nodes are connected with @xmath122 bi - fundamental fields @xmath124 , and there is no link field between the first and last nodes .
when the @xmath124 s acquire uniform vevs , @xmath125 , for all @xmath126 s .
the gauge bosons of the group @xmath123 have a @xmath127 mass matrix , which has a zero eigenvalue , while the @xmath128 gauge bosons have a @xmath129 mass matrix , which does not have a zero eigenvalue @xcite .
as the @xmath128 gauge bosons become heavy , @xmath119 is effectively broken down to @xmath123 for the massless fields .
additional fields will be needed on the first and @xmath121-th nodes to cancel anomalies .
alternatively , @xmath121 bi - fundamental fields @xmath124 connect the @xmath121 nodes to form a loop .
with the same uniform vevs , the @xmath123 gauge bosons have @xmath127 mass matrices , while the @xmath128 gauge bosons have @xmath129 mass matrices .
only one set of @xmath123 gauge bosons remain massless .
the anomalies are always canceled in this setup , thus no additional field is required . from now on we always discuss the simplified case of two copies of the initial @xmath35 group @xmath130 where @xmath131 , with @xmath132 and @xmath133 being the gauge fields and generators of @xmath134 , and @xmath135 and @xmath136 the gauge fields and generators of @xmath137 .
we assume that the gauge couplings are equal at the breaking scale for simplicity .
the generalization to more copies of the @xmath35 group is straightforward . with @xmath138 and @xmath139
being two bi - fundamental fields , the covariant derivatives are [ eq : covderi ] d_u_1 = _ u_1 - i a_^ t_u_1 + i b_^ t_u_1 , + d_u_2 = _
u_2 + i a_^ t_u_2 - i b_^ t_u_2 , and the effective action for the scalar fields is [ eq : action ] s&=&d^4x \{-1 tr f_1 ^ 2 + tr[(d _
u_1)^ d^ u_1 ] - 1 tr f_2 ^ 2 + .
u_2)^ d^ u_2]+ ... } . for simplicity , we only write down the effective action for the scalar fields . inspired by the high dimensional non - supersymmetric guts with wilson line gauge symmetry breaking , we can use the vevs of the bi - fundamental fields that are not commutating with a subset of the generators to break the original group @xcite .
let us suppose the initial gauge group is @xmath140 and @xmath123 is a subgroup of @xmath35 .
the term @xmath141 + tr[(d_{\mu } u_2)^{\dagger } d^{\mu } u_2]$ ] in eq .
( [ eq : action ] ) produces a mass matrix of the form [ eq : gmmatrx ] ( cc a_^&b_^ ) ( cc tr[u_1^t_^ t_ u_1+t_^ u_2^ u_2 t_ ] & -tr[u_1^ t_^ u_1 t_ + t_^ u_2^ t_ u_2 ] + -tr[t_^ u_1^ t_ u_1 + u_2^ t_^ u_2 t_ ] & tr[t_^ u_1^ u_1 t_ + u_2^ t_^ t_ u_2 ] ) ( c a_^ + b_^ ) .
after @xmath138 and @xmath139 acquire vevs , the @xmath124 s in the mass matrix will be replaced by @xmath142 .
the vev @xmath142 has the property that @xmath143 is always proportional to the identity matrix .
the normalized generators satisfy the relation @xmath144 for @xmath53 or @xmath145 for @xmath146 , thus the diagonal terms in matrix in eq .
( [ eq : gmmatrx ] ) are always proportional to @xmath147 , while the off - diagonal terms depend on the commutation relations between the vevs of the bi - fundamental fields and the generators .
if the vev of @xmath138 commutes with all the generators of @xmath35 , while the vev of the @xmath139 commutes with the generators of @xmath123 group , @xmath148 , and anti - commutes with those of @xmath128 , @xmath149 ^ t_a < u_1 > = t_a ^ t_a < u_1 > = t_a , ^ t_a < u_2 > = t_a ^ t_a < u_2 > = - t_a , the mass matrices become proportional to ( cc a_^a & b_^b ) ( cc 1 & -1 + -1 & 1 ) ( c a_^a + b_^b ) ( cc a_^a & b_^b ) ( cc 1 & 0 + 0 & 1 ) ( c a_^a + b_^b ) , respectively .
while the former matrix has a @xmath150 eigenvalue , the latter does not .
thus the gauge symmetry @xmath140 is broken down to the diagonal subgroup @xmath123 . before presenting our models
, we would like to emphasize that we consider the 4-dimensional @xmath0 supersymmetry in all the models discussed in this paper .
one way to realize the deconstruction of orbifold supersymmetric @xmath1 models is to start with the initial gauge group @xmath151 .
it is broken down to the diagonal ps gauge symmetry by bifundamental link fields .
the doublet - triplet splitting can be achieved by the missing partner mechanism , and the proton decay problem is solved because the gauge coupling unification scale is a few times @xmath16 gev .
a simple variation of this model is the @xmath7 model .
in addition , we can consider the @xmath6 model , where the gauge symmetry can be broken down to the diagonal ps gauge symmetry , which is inspired by the deconstruction of 5-dimensional wilson line gauge symmetry breaking .
however , the doublet - triplet splitting problem can not be solved without fine - tuning in superpotential .
to break @xmath152 to the diagonal ps gauge symmetry , we rely on two isomorphisms , @xmath153 and @xmath154 .
we introduce the bi - fundamental fields @xmath138 and @xmath139 with the following quantum numbers @xmath155 suppose , for example , we have the superpotential @xmath156 where @xmath157 and @xmath87 are singlets , and the link fields acquire vevs < u_1 > = v ( c i_66 + 0_46 + ) , < u_2 > = v ( c 0_64 + i_44 + ) , where @xmath158 is the @xmath159 identity matrix , and @xmath160 is a @xmath161 matrix where all the entries are zero . the @xmath162 mass matrix for the ps gauge bosons is @xmath163 with eigenvalues @xmath164 .
the masses of the @xmath165 gauge bosons are simply @xmath166 .
if generalized to a model with gauge group @xmath167 , the ps gauge bosons have a @xmath127 mass matrix which has a zero eigenvalue , whereas the non - ps gauge bosons have a @xmath129 mass matrix , for which the zero eigenvalue is absent .
three families of the sm fermions form three @xmath168 spinor representations of @xmath1 , and we introduce a 10-dimensional higgs field @xmath169 .
to break the ps gauge symmetry and at the same time require a left - right symmetry , , the coupling @xmath170 of the @xmath171 equals to @xmath172 of the @xmath173 above the ps gauge symmetry unification scale ( @xmath174 ) , we introduce the fields @xmath175 , @xmath176 , @xmath177 , @xmath178 , @xmath179 , @xmath180 , @xmath181 and @xmath182 .
their quantum numbers under the gauge symmetry @xmath183 are given in the following @xmath184 note that @xmath175 , @xmath176 , @xmath177 , @xmath185 can form one pair of @xmath168 and @xmath186 under @xmath1 .
the ps gauge symmetry is broken down to the sm gauge symmetry when the right - handed neutrino scalar component in @xmath175 and its charge conjugation in @xmath187 obtain vevs at the @xmath174 scale .
the superpotential is @xmath188 where @xmath189 s are yukawa couplings , and @xmath190 s are the usual yukawa couplings for the sm fermions .
moreover , similarly to the doublet - triplet splitting via the missing partner mechanism , the color triplets in @xmath169 and @xmath179 obtain the gut scale masses while the doublets are massless because of the superpotential term @xmath191 .
subsequently , the proton decay due to the dimension-5 operator is negligible because the mixing of the color triplets through the @xmath192 term is very small .
limits on the contributions from the dimension-6 proton decay operators require that the gauge coupling unification scale be larger than @xmath193 gev @xcite , which is obviously satisfied in our model where the gut scale is at least a few times @xmath16 gev .
furthermore , the @xmath192 term for the higgs doublets can be generated from the superpotential @xmath194 , as is similar to the next to the minimal supersymmetric standard model @xcite .
we also assume that the masses for @xmath180 , @xmath175 and @xmath187 are around @xmath174 , and the masses for @xmath179 and higgs triplets in @xmath169 are about @xmath195 .
we do not discuss the fermion masses and mixings , as they are out of the scope of this paper . for simplicity , we assume the universal masses @xmath196 gev for the supersymmetric particles . from @xmath197 to @xmath198 , the gauge couplings evolve the same way as those in the sm , and the beta functions are @xmath199 , where @xmath200 , @xmath201 and @xmath202 are for the gauge symmetries @xmath203 , @xmath171 and @xmath204 , respectively . from @xmath198 to the ps gauge symmetry unification scale @xmath174 , the beta functions are just those of the mssm , _
i.e. _ , @xmath205 . at the ps gauge symmetry unification scale @xmath174 ,
the gauge couplings for @xmath203 and @xmath204 are _
1 ^ -1(m_ps ) = _
1 ^ -1(m_z ) - ( ) - ( ) , _ 3 ^ -1(m_ps ) = _ 3 ^ -1(m_z ) - ( ) - ( ) .
the @xmath173 gauge coupling @xmath172 is related to @xmath206 and @xmath207 by _
r^-1(m_ps ) = _ 1 ^ -1(m_ps ) - _ 3 ^ -1(m_ps ) . above the ps scale , including the contributions from fermions , ps gauge multiplets , two higgs doublets @xmath208 , @xmath175 @xmath187 , @xmath209 , @xmath210 , and @xmath180 , the beta functions @xmath211 become b^ii & = & b(m ) + b^ps(v ) + b(h^d ) + b ( ) + b(c_2 ) + & = & ( 6,6,6 ) + ( -6,-6,-12 ) + ( 1,1,0 ) + ( 4 , 4 , 4 ) + ( 0,0,1 ) = ( 5,5,-1 ) , where @xmath212 , @xmath213 and @xmath214 are the beta functions for @xmath173 , @xmath171 and @xmath215 , respectively . in the region
above @xmath216 , one set of @xmath217 gauge multiplets , @xmath179 and higgs triplets @xmath218 in @xmath169 appear , then the beta functions are b^iii & = & b^ii + b^so(10)/ps(v ) + b(c_1 ) + b(h^t ) + & = & ( 5,5,-1 ) + ( -18,-18,-12 ) + ( 0 , 0 , 1 ) + ( 0 , 0 , 1 ) = ( -13,-13,-11 ) . we show the runnings of the gauge couplings near the unification scale in the left panel of fig . [
fig : uni_ps1 ] . note that the runnings of @xmath170 and @xmath172 coincide above @xmath174 because we have required the left - right symmetry .
the gauge coupling @xmath219 of @xmath220 unify with @xmath172 and @xmath170 at the scale @xmath221 gev and the ps scale occurs at @xmath222 gev .
thus , there is no proton decay problem .
right : the gauge coupling unification for the left - right non - symmetric ps model near @xmath223 gev . [ fig : uni_ps1],title="fig:",width=302 ] gev .
right : the gauge coupling unification for the left - right non - symmetric ps model near @xmath223 gev .
[ fig : uni_ps1],title="fig:",width=302 ] in the above construction , we maintain the left - right symmetry above the @xmath174 scale .
if we no longer require the left - right symmetry , we can drop the additional fields @xmath209 and @xmath224 from the table in eq .
( [ so10epf ] ) . in this setup
, the discussion of the doublet - triplet splitting is unchanged from that of the above symmetric case .
the rge runnings of the gauge couplings are different due to the simpler field content .
more specifically , the beta functions are , @xmath225 and @xmath226 . above @xmath227 gev , @xmath172 and
@xmath170 have different rge runnings , as is evident in fig .
[ fig : uni_ps1 ] right panel . with a gut scale of @xmath228 gev ,
the proton lifetime is safely above the current experimental bound .
if we want to solve the proton decay problem , we can also consider the model with gauge group @xmath7 , which can not be obtained but is inspired by the orbifold gut . to break @xmath229 to ps gauge symmetry , we introduce bi - fundamental field @xmath138 as @xmath230 the link fields acquire the following vevs < u_1 > = v ( c i_66 + 0_46 + ) . the @xmath7 gauge symmetry is broken down to the ps gauge symmetry . the discussions on the proton decay , ps gauge symmetry breaking and the gauge coupling unification are similar to those in the above subsection , so
, we will not repeat them here .
however , we would like to emphasize that the initial gauge group is smaller than that in the above subsection . inspired by the deconstruction of 5-dimensional @xmath231 model with wilson line gauge symmetry breaking , we consider the model with the simplest gauge group @xmath6 .
let us introduce bi - vector fields @xmath138 and @xmath139 with the following quantum numbers @xmath232 the effective action is the same as that in eq .
( [ eq : action ] ) .
if the @xmath138 and @xmath139 fields acquire vevs < u_1 > = v _ 0 ( 1 , 1 , 1,1,1 ) , < u_2 > = v _ 0 ( -1,-1,-1,1,1 ) , the @xmath162 mass matrix for the ps gauge bosons is @xmath233 with @xmath234 defined in eq .
( [ eq : sm32 ] ) .
the squared masses for the ps group gauge bosons are either @xmath150 or @xmath235 .
the mass matrix of the non - ps ( @xmath217 ) gauge bosons is @xmath236 it has degenerate eigenstates with squared masses equal to @xmath237 .
three families of the sm fermions form three @xmath168 spinor representations under the first @xmath1 while they are singlets under the second @xmath1 .
we also introduce a higgs field @xmath169 . to give masses to the color triplets in the higgs field @xmath169 ,
we introduce a 10-dimensional field @xmath238 , and to break the ps gauge symmetry , we introduce the fields @xmath239 , @xmath240 , and two singlets @xmath181 and @xmath182 .
the quantum numbers for these particles are @xmath241 the superpotential is @xmath242 where @xmath243 is the @xmath1 unification scale .
the doublet - triplet splitting can be obtained through the fine - tunning superpotential @xmath244 , as in the usual @xmath11 model . because it is less interesting than the usual @xmath1 model from the phenomenological point of view ,
we do not study it in detail .
for the @xmath247 and @xmath248 models from the deconstruction of 5-dimensional supersymmetric orbifold @xmath1 models , the discussions on how to break the gauge symmetry down to the diagonal flipped @xmath249 or @xmath250 gauge symmetry are identical to those of section [ sec2.1 ] . two link fields with simple diagonal vevs will leave the diagonal subgroup gauge bosons massless and give the other gauge bosons masses of order @xmath251 .
we will consider the models inspired by deconstruction of the 5-dimensional @xmath1 models with wilson line gauge symmetry breaking , where the @xmath1 gauge symmetry can be broken down to the flipped @xmath246 and @xmath250 , and the 6-dimensional @xmath1 model where the @xmath1 gauge symmetry can be broken down to the sm like gauge symmetry .
starting with the @xmath252 , we use two bi - vectors @xmath138 and @xmath139 with quantum numbers @xmath253 . in order to break the gauge symmetry down to the diagonal @xmath81 , the vevs of the link fields are assumed to be < u_1 > & = & v _ 0 ( 1,1,1,1,1 ) , + < u_2 > & = & v _ 2 ( 1,1,1,1,1 ) , and to obtain the flipped @xmath15 , @xmath139 vev is replaced with < u_2 > = v _ 2 ( 1,1,1,-1,-1 ) . note the similarity between the vevs and the parity operators in eqs .
( [ eq : psu5 ] ) and ( [ eq : pfsu5 ] ) . as @xmath254 commutes with @xmath148 and @xmath149 , while @xmath255 commutes with @xmath148 and anti - commutes with @xmath149 , the gauge bosons of the unbroken subgroup @xmath256 or flipped @xmath15 have a mass matrix identical to the right hand side of eq .
( [ eq : so102 ] ) while the @xmath128 gauge bosons have a mass matrix the same as the right hand side of eq .
( [ eq : so2nops ] ) .
one set of the @xmath123 gauge bosons remain massless and the @xmath128 gauge bosons acquire masses of @xmath257 .
moreover , the @xmath258 can be broken to the intersection of two maximal subgroups , @xmath259 . if we introduce four bi - vector link fields with vevs & = & v _ 0 ( 1,1,1,1,1 ) , + < u_2 > & = & v _ 2 ( 1,1,1,1,1 ) , + < u_3 > & = & v _ 0 ( 1,1,1,1,1 ) , + < u_4 > & = & v _ 0 ( 1,1,1,-1,-1 ) , where all @xmath124 s are @xmath260 under the original group @xmath2 .
it is easy to see that @xmath138 and @xmath139 are the same as the link fields we used for breaking @xmath261 to @xmath81 , and @xmath262 and @xmath263 are the same as those for breaking @xmath2 to the ps group . to clarify the notation
, we assign @xmath264 for the mass matrices of the gauge bosons in the intersection of the @xmath3 and ps gauge symmetry , which is @xmath265 , @xmath266 for those in @xmath81 but not in ps gauge group , @xmath267 for those in the ps gauge group but not in @xmath81 , and @xmath268 is reserved for the mass matrices of the rest gauge bosons belong to neither @xmath3 or ps gauge group .
these mass matrices are given by m_smx^2 = 4 g^2 v^2 ( cc 2&-2 + -2&2 + ) , m_5 ^ 2 = m_ps^2 = 4 g^2 v^2 ( cc 2&-1 + -1&2 + ) , m_r^2 = 4 g^2 v^2 ( cc 2&0 + 0&2 + ) .
it is obvious that only @xmath264 has 0 eigenvalue , and the others do not . with the combination of these two sets of link fields ,
the @xmath269 gauge symmetry is broken to the @xmath270 gauge symmetry .
similarly , we can break @xmath269 down to @xmath95 by choosing & = & v _ 0 ( 1,1,1,1,1 ) , + < u_2 > & = & v _ 2 ( 1,1,1,-1,-1 ) , + < u_3 > & = & v _ 0 ( 1,1,1,1,1 ) , + < u_4 > & = & v _ 0 ( 1,1,1,-1,-1 ) , or < u_1 > & = & v _ 0 ( 1,1,1,1,1 ) , + < u_2 > & = & v _ 2 ( 1,1,1,1,1 ) , + < u_3 > & = & v _ 0 ( 1,1,1,1,1 ) , + < u_4 > & = & v _ 2 ( 1,1,1,-1,-1 ) . in these models ,
the doublet - triplet splitting is still a problem without fine - tuning in the superpotential , thus it is difficult to solve the proton decay problem induced by the dimension-5 proton decay operators .
in this section , we demonstrate that , by using bi - spinor link fields ( @xmath13 ) and ( @xmath271 ) , we are able to break the gauge symmetry @xmath6 down to the diagonal @xmath3 and flipped @xmath15 gauge symmetry , and at the same time solve the doublet - triplet splitting problem via the missing partner mechanism in the model with the flipped @xmath15 breaking chain . in this approach , we do not rely on the commutation relations between the link field vevs and the generators .
we introduce two bi - fundamental fields @xmath138 and @xmath139 with following quantum numbers under @xmath272 @xmath273 then , the covariant derivative can be rewritten as , using @xmath274 defined in appendix [ apdx ] , d_u_1 = _ u_1 - u_1 + u_1 , + d_u_2 = _ u_2 + u_2 - u_2 .
the effective action for scalar components is the same as that in eq .
( [ eq : action ] ) , and the bi - spinors @xmath138 and @xmath139 acquire the following vevs = < u_2 > = v diag(0,0,0,1,0,0,0,1,1,1,1,0,0,0,0,1 ) .
note that the assignment of the @xmath275 s are consistent with the diagonal ( @xmath276 ) and ( @xmath277 , @xmath277 ) in the ( @xmath278 ) , see apppendix [ apdx ] .
the gauge fields in eq .
( [ eq : gauge16 ] ) have the mass matrices of the following forms m_^2 = m_v^2= m_w_l^2= m_y^2 = g^2 v^2 x_1 , m_a^2 = m_w_r^2 = m_x^2 = 2 g^2 v^2 x_2 . thus , the massless fields are the 5 states in the @xmath279 fields , 6 @xmath280 , 2 @xmath281 and 12 @xmath282 states , total of 25 independent states .
they are the gauge bosons of the diagonal @xmath81 group . for the flipped @xmath15 , the @xmath283 and @xmath277 consist of different fermion fields .
the vevs of @xmath138 and @xmath139 are chosen as = < u_2 > = v diag(0,0,0,1,0,0,0,1,0,0,0,1,1,1,1,0 ) .
the mass matrices become m_^2 = m_v^2= m_w_l^2= m_a^2 = g^2 v^2 x_1 , m_y^2 = m_w_r^2 = m_x^2 = 2 g^2 v^2 x_2 , and the massless states are 12 @xmath284 , 2 @xmath281 , 6 @xmath280 and the 5 states in @xmath279 s .
these 25 massless states are the gauge bosons of the diagonal flipped @xmath15 .
three families of the sm fermions form three @xmath168 spinor representations under the first @xmath1 while they are singlets under the second @xmath1 .
we also introduce a 10-dimensional higgs field @xmath169 , and one pair of higgs @xmath239 and @xmath240 in the spinor representation .
moreover , we introduce one pair of the fields @xmath285 and @xmath286 in the spinor representation , two singlet fields @xmath181 and @xmath182 .
the quantum numbers for these particles are @xmath287 the superpotential is @xmath288 where @xmath289 is the unification scale for the flipped @xmath15 , and @xmath290 is the planck scale .
the ( @xmath291 ) multiplets in @xmath239 and @xmath285 , and the ( @xmath292 ) multiplets in @xmath293 and @xmath286 are assumed to have the @xmath1 unification scale masses from the first two terms in the superpotential , while the @xmath294 multiplet @xmath175 in @xmath239 , the @xmath295 multiplet @xmath296 in @xmath285 , the @xmath297 multiplet @xmath176 in @xmath293 , and the @xmath298 multiplet @xmath299 in @xmath300 are still massless at @xmath301 . for simplicity
, we assume that the masses @xmath302 for @xmath296 and @xmath299 are about @xmath303 gev from the last term in superpotential . under flipped @xmath15 , @xmath169
is decomposed into one pair of 5-plets @xmath304 and @xmath305 . to be explicit ,
we denote @xmath175 , @xmath187 , @xmath304 and @xmath305 as @xmath306 @xmath307 from the third , fifth and sixth terms in eq .
( [ superpotential33 ] ) , we obtain the superpotential below @xmath301 @xmath308 there is only one f - flat and d - flat direction , which can always be rotated along the @xmath309 and @xmath310 directions .
hence we obtain that @xmath311 .
in addition , the superfields in @xmath175 and @xmath176 are eaten or acquire large masses via the supersymmetric higgs mechanism , except for @xmath312 and @xmath313 .
the superpotential @xmath314 and @xmath315 couple the @xmath312 and @xmath316 with the @xmath317 and @xmath318 respectively to form the massive eigenstates with masses @xmath319 and @xmath320 .
we naturally have the doublet - triplet splitting due to the missing partner mechanism @xcite . because the triplets in @xmath304 and @xmath321 only have small mixing through the @xmath192 term , the higgsino - exchange mediated proton decay are negligible , _
i.e. _ , we do not have the dimension-5 proton decay problem .
proton decay via the dimension-6 operators is well above the current experimental bounds .
the gauge coupling unification for the flipped @xmath246 is realized by first unifying @xmath322 and @xmath207 at scale @xmath323 , then the gauge couplings of @xmath11 and @xmath324 further unify at @xmath301 . from @xmath197 to @xmath198 , the beta functions are @xmath325 , and from @xmath198 to @xmath326 gev , the beta functions are @xmath205 . from @xmath302 to the @xmath322 and @xmath207 unification scale @xmath323 ,
the beta functions are @xmath327 , because @xmath328 and @xmath299 contribute to the gauge coupling rge runnings .
unification of @xmath322 and @xmath207 at @xmath323 gives the condition @xmath329 which can be solved for @xmath323 .
the coupling @xmath330 of @xmath324 is related to @xmath206 and @xmath331 at @xmath323 by _
1^-1(m_23 ) = _ 1 ^ -1(m_23 ) - _ 5 ^ -1(m_23 ) . between @xmath323 and @xmath301 , besides the 3 families of the standard model fermions @xmath332 , there are the @xmath15 gauge bosons in @xmath333 , @xmath169 , @xmath334 , @xmath335 , @xmath328 and @xmath299 , thus the beta functions for @xmath324 and @xmath11 are @xmath336 . in fig .
[ fig : flipuni ] , we show the runnings of the gauge couplings near the unification scale @xmath301 .
the unification scale @xmath323 for @xmath171 and @xmath204 is about @xmath337 gev , and the @xmath6 unification scale @xmath301 is about @xmath338 gev .
if the masses @xmath302 for @xmath328 and @xmath339 are larger than @xmath323 , we will have @xmath340 at @xmath323 . because above @xmath323 the beta function of @xmath324 is positive while that of @xmath11 is negative
, we can not achieve the unification of these two gauge couplings .
in short , the intermediate mass scale @xmath302 is important for the gauge coupling unification @xcite .
gev for flipped @xmath15 model .
[ fig : flipuni],width=302 ] we introduce the link fields @xmath138 , @xmath139 , @xmath262 and @xmath263 with following quantum numbers @xmath342 we choose the following vevs for the link fields @xmath343 and @xmath344 , where @xmath345 is a @xmath346 matrix with elements at @xmath347 , @xmath348 , @xmath349 , @xmath350 , and @xmath351 being @xmath352 and all other elements @xmath150 , and @xmath353 is a column vector with @xmath352 as the @xmath354 element and all other elements @xmath150 . similarly
to above subsection , we can break the @xmath355 gauge symmetry down to a diagonal flipped @xmath15 gauge symmetry .
three families of the sm fermions form three @xmath168 spinor representations under the first @xmath1 while they are singlets and neutral under the @xmath356 .
we also introduce a 10-dimensional higgs field @xmath169 , and one pair of higgs @xmath239 and @xmath293 in the spinor representation .
moreover , we introduce one pair of the 5-plet fields @xmath357 and @xmath358 , one pair of the 10-plet fields @xmath328 and @xmath299 , one pair of the @xmath324 charged singlets @xmath359 and @xmath360 , and two singlet fields @xmath181 and @xmath182 .
the quantum numbers for these particles are @xmath361 the superpotential is @xmath362 the discussions for the proton decay and the gauge coupling unification are similar to these in the previous subsection .
we want to emphasize that the doublet - triplet splitting problem can be solved by the missing partner mechanism , and the solution of the proton decay problem follows .
we have considered the 4-dimensional @xmath0 supersymmetric @xmath1 models inspired by the deconstruction of 5-dimensional supersymmetric orbifold @xmath1 models and high dimensional non - supersymmetric @xmath1 models with wilson line gauge symmetry breaking .
we studied the @xmath2 models with bi - fundamental link fields where the @xmath2 gauge symmetry can be broken down to the ps , @xmath3 , flipped @xmath4 and the sm like gauge symmetry .
however , we need to fine - tune the superpotential to obtain the doublet - triplet splitting .
we proposed two interesting models : the @xmath363 model where the gauge symmetry can be broken down to ps gauge symmetry , and the @xmath6 model with bi - spinor link fields in which the gauge symmetry can be broken down to the flipped @xmath15 gauge symmetry .
these intermediate gauge symmetries can be further broken down to the sm gauge symmetry . in these models ,
the missing partner mechanism can be implemented to solve the doublet - triplet splitting problem , then the higgsino - exchange mediated proton decay is negligible , and consequently the proton decay is mainly induced by the dimension-6 operators . with the gut scale being at least a few times
@xmath16 gev , the proton lifetime is well above the current experimental bound .
in addition , we discussed the gauge coupling unification in these two models : in the @xmath12 models , the gauge couplings unify at @xmath364 gev for a left - right symmetric model and at @xmath365 gev for a left - right non - symmetric model ; and in the @xmath6 model , the gauge couplings unify at about @xmath366 gev .
furthermore , we briefly commented on the interesting variation models with gauge groups @xmath7 and @xmath367 where the proton decay problem can be solved in the similar manner , and the @xmath6 model with bi - spinor link fields in which the gauge symmetry can be broken down to the @xmath3 gauge symmetry .
j. jiang thanks csaba balazs and cosmas zachos for helpful discussions .
the research of c .- s .
huang was supported in part by the natural science foundation of china , the research of j. jiang was supported by the u.s .
department of energy under grant no .
w-31 - 109-eng-38 , and the research of t. li was supported by the national science foundation under grant no .
phy-0070928 .
the generators and the assignment of the fermions in the * 16 * can be found in ref .
we copy the @xmath368 , and rename it @xmath369
. the @xmath370 matrix can be re - written into four @xmath371 matrices , [ eq : gauge16 ] / = ( cccccccc _
1 1&v_12&v_13&x_1 ^ 0&w_l^- & & & + v_12^*&_2 2&v_23&x_2 ^ - & & w_l^- & & + v_13^*&v_23^*&_3 3&x_3 ^ - & & & w_l^- & + _ 1 ^ 0&x_2^+&x_3^+&_4 4 & & & & w_l^- + w_l^+ & & & & _ 5 5&v_12&v_13&x_1
^ 0 + & w_l^+ & & & v_12^*&_6 6&v_23&x_2 ^ - + & & w_l^+ & & v_13^*&v_23^*&_7 7&x_3 ^ - + & & & w_l^+&_1 ^ 0&x_2^+&x_3^+&_8 8 cccccccc 0&a_6 ^ 0&-a_5 ^ 0&-y_1^+&0&-y_6 ^ -&y_5 ^ -&-_1 ^ 0 + -a_6 ^ 0&0&a_4 ^ -&-_2 ^ 0&y_6 ^ -&0&-y_4^&-a_2 ^ - + a_5 ^
0&-a_4 ^ -&0&-_3 ^ 0&-y_5 ^ -&y_4^&0&-a_3 ^ - + y_1^+&_2 ^ 0&_3 ^ 0&0&_1 ^ 0 & a_2 ^ -&a_3 ^ -&0 + 0&-a_3^+&a_2^+&-y_4^++&0&y_3 ^ 0&-y_2 ^ 0&-a_4^+ + a_3^+&0&a_1 ^ 0&-y_5^+&y_3 ^ 0&0&-y_1 ^ -&-_5 ^ 0 + -a_2^+&a_1 ^ 0&0&-y_6^+&y_2 ^ 0&-y_1 ^ -&0&-_6 ^ 0 + y_4^++&y_5^+&y_6^+&0&a_4^+&_5 ^ 0&_6 ^ 0&0 cccccccc 0&-_6 ^ 0&_5 ^ 0&y_1 ^ -&0&a_3 ^ -&-a_2 ^ -&y_4^ + _ 6 ^ 0&0&-a_4^+&y_2 ^ 0&-a_3 ^ -&0&_1
^ 0&y_5 ^ - + -_5 ^ 0&a_4^+&0&y_3 ^ 0&a_2 ^ -&-_1 ^ 0&0&y_6 ^ - + -y_1 ^ -&-y_2 ^ 0&-y_3 ^ 0&0&-y_4 ^ -&-y_5 ^ -&-y_6 ^ -&0 + 0&y_6^+&-y_5^+&a_1 ^ 0&0&-_3 ^ 0&_2 ^ 0&a_4 ^ - + -y_6^+&0&y_4^++&a_2^+&_3 ^ 0&0&-y_1^+&a_5 ^ 0 + y_5^+&-y_4^++&0&a_3^+&-_2 ^ 0&y_1^+&0&a_6 ^ 0 + -a_1 ^ 0&-a_2^+&-a_3^+&0&-a_4 ^ -&-a_5 ^ 0&-a_6 ^ 0&0 cccccccc _ 9 9&-v_12^*&-v_13^*&-_1 ^ 0&w_r^- & & & + -v_12&_10 10&-v_23^*&-x_2^+ & & w_r^- & & + -v_13&-v_23&_11 11&-x_3^+ & & & w_r^- & + -x_1 ^ 0&-x_2
^ -&-x_3 ^ -&_12 12 & & & & w_r^- + w_r^+ & & & & _ 13 13&-v_12^*&-v_13^*&-_1 ^ 0 + & w_r^+ & & & -v_12&_14 14&-v_23^*&-x_2^+ + & & w_r^+ & & -v_13&-v_23&_15 15&-x_3^+ + & & & w_r^+&-x_1 ^ 0&-x_2 ^ -&-x_3 ^ 3&_16 16 the 45 gauge bosons consist of 12 @xmath284 , 6 @xmath372 , 6 @xmath280 , 12 @xmath282 , 2 charge @xmath281 , 2 charge @xmath373 and 16 @xmath279 which can be rewritten as 5 independent fields , @xmath374 , @xmath375 , @xmath376 , @xmath377 and @xmath378 .
the first family of the sm fermions forms a @xmath168 , _
1= ( u_r , u_g , u_b,_e , d_r , d_g , d_b , e^-,d_r^c , d_g^c , d_b^c , e^+,-u_r^c ,-
u_b^c,-_e^c)^t , similarly for the second and third families . as the @xmath1 is broken down to @xmath250 or flipped @xmath246 , the spinor representation @xmath168 is decomposed as + * |5 * + * 1 * where = ( q , u^c , e^+ ) , = ( d^c , l ) , = ^c for breaking to @xmath250 , and = ( q , d^c,^c ) , = ( u^c , l ) , = e^+ for breaking to flipped @xmath246 .
l. j. hall , h. murayama and y. nomura , nucl .
b * 645 * , 85 ( 2002 ) [ arxiv : hep - th/0107245 ] .
n. haba , t. kondo , y. shimizu , t. suzuki and k. ukai , prog . theor .
phys . * 106 * , 1247 ( 2001 ) [ arxiv : hep - ph/0108003 ] .
t. asaka , w. buchmuller and l. covi , phys .
b * 523 * , 199 ( 2001 ) ; [ arxiv : hep - ph/0108021 ] .
t. li , nucl .
b * 619 * , 75 ( 2001 ) [ arxiv : hep - ph/0108120 ] .
r. dermisek and a. mafi , phys .
d * 65 * , 055002 ( 2002 ) [ arxiv : hep - ph/0108139 ] . c. s. huang , j. jiang , t. li and w. liao , phys .
b * 530 * , 218 ( 2002 ) [ arxiv : hep - th/0112046 ] .
t. li , nucl .
b * 633 * , 83 ( 2002 ) [ arxiv : hep - th/0112255 ] .
b. kyae and q. shafi , phys .
d * 69 * , 046004 ( 2004 ) [ arxiv : hep - ph/0212331 ] .
h. d. kim and s. raby , jhep * 0301 * , 056 ( 2003 ) [ arxiv : hep - ph/0212348 ] .
i. gogoladze , y. mimura and s. nandi , phys .
b * 562 * , 307 ( 2003 ) [ arxiv : hep - ph/0302176 ] ; phys .
* 91 * , 141801 ( 2003 ) [ arxiv : hep - ph/0304118 ] .
t. li , jhep * 0403 * , 040 ( 2004 ) [ arxiv : hep - ph/0309199 ] .
n. arkani - hamed , a. g. cohen and h. georgi , phys . rev .
lett . * 86 * , 4757 ( 2001 ) [ arxiv : hep - th/0104005 ] .
c. t. hill , s. pokorski and j. wang , phys .
d * 64 * , 105005 ( 2001 ) [ arxiv : hep - th/0104035 ] . c. csaki , g. d. kribs and j. terning , phys .
d * 65 * , 015004 ( 2002 ) [ arxiv : hep - ph/0107266 ] .
h. c. cheng , k. t. matchev and j. wang , phys .
b * 521 * , 308 ( 2001 ) [ arxiv : hep - ph/0107268 ] .
t. li and t. liu , eur .
j. c * 28 * ( 2003 ) 545 [ arxiv : hep - th/0204128 ] .
see , , m. b. halpern and w. siegel , phys .
d * 11 * , 2967 ( 1975 ) ; v. d. barger , w. y. keung and e. ma , phys .
b * 94 * , 377 ( 1980 ) .
n. arkani - hamed , t. gregoire and j. wacker , jhep * 0203 * , 055 ( 2002 ) . y. hosotani , phys .
lett . b*126 * ( 1983 ) 309 ; phys .
lett . b*129 * ( 1983 ) 193 ; e. witten , nucl .
b*258 * ( 1985 ) 75 ; j. d. breit , b. a. ovrut and g. c. segre , phys . lett . b*158 * ( 1985 ) 33 .
g. f. giudice and a. romanino , arxiv : hep - ph/0406088 .
j. ellis , j. f. gunion , h. e. haber , l. roszkowski , and f. zwirner , phys .
d * 39 * , 844 ( 1989 ) ; and references therein .
recent references may be found in u. ellwanger , j. f. gunion , and c. hugonie , hep - ph/0111179 . c. s. huang , t. li and w. liao , nucl .
b * 673 * , 331 ( 2003 ) [ arxiv : hep - ph/0304130 ] .
j. l. lopez , d. v. nanopoulos and a. zichichi , arxiv : hep - ph/9307211 , and references therein .
s. rajpoot , phys .
d * 22 * , 2244 ( 1980 ) . | we consider 4-dimensional @xmath0 supersymmetric @xmath1 models inspired by deconstruction of 5-dimensional @xmath0 supersymmetric orbifold @xmath1 models and high dimensional non - supersymmetric @xmath1 models with wilson line gauge symmetry breaking .
we discuss the @xmath2 models with bi - fundamental link fields where the gauge symmetry can be broken down to the pati - salam , @xmath3 , flipped @xmath4 or the standard model like gauge symmetry .
we also propose an @xmath5 model with bi - fundamental link fields where the gauge symmetry is broken down to the pati - salam gauge symmetry , and an @xmath6 model with bi - spinor link fields where the gauge symmetry is broken down to the flipped @xmath4 gauge symmetry . in these two models ,
the pati - salam and flipped @xmath4 gauge symmetry can be further broken down to the standard model gauge symmetry , the doublet - triplet splittings can be obtained by the missing partner mechanism , and the proton decay problem can be solved .
we also study the gauge coupling unification .
we briefly comment on the interesting variation models with gauge groups @xmath7 and @xmath8 in which the proton decay problem can be solved .
anl - hep - pr-04 - 66 + hep - ph/0407180 + |
early extragalactic source ( e.g.r.s ) rotation measure ( rm ) probes of the milky way magnetic field by gardner & davies , ( 1966 ) , and the first pulsar rm analyses by manchester ( 1972 , 1974 ) revealed a prevailing magnetic field in the local spiral arm that is clock - wise as viewed from the north galactic pole ( ngp ) .
further study with a 555-source extragalactic rm dataset produced the discovery of a large scale field reversal in the sagittarius - carina arm ( simard - normandin & kronberg , 1979 ) , indicating a global bisymmetric field geometry .
subsequent numerical analyses by simard - normandin & kronberg ( 1980 ) confirmed a prevailing local field pointing to @xmath4 , a more complex pattern in the north galactic hemisphere ( ngh ) , an anomalously high off - plane , high rm zone ( `` region a '' ) , a @xmath5 angular autocorrelation scale of rm s off the plane , and a galactic magneto - ionic scale height of @xmath6 1.8 kpc .
confirmation of a large scale ( bisymmetric ) pattern has since come from several e.g.r.s and pulsar rm studies rand & lyne , 1994 ; han _ et al .
_ , 1999 ; frick _ et al .
_ , 2001 & weisberg _ et al .
variants of this model have been proposed by han _
_ , 2006 @xmath7 a bisymmetric model with field reversals at each arm / interarm interface , valle ( 2008 ) @xmath7 a concentric ring model having one field reversal , and sun _ et al . _
( 2008 ) a 3d - model combined with observations , concluding that a combined axisymmetric and toroidal model best describes the galactic magnetic field .
et al . _ , ( 2008 ) , using pulsar rm s , concluded that neither a concentric ring model nor the bi - symmetric or axi - symmetric field models alone , will suffice when attempting to describe the galactic magnetic field . the high density of e.g.r.s .
rm s in the @xmath8 canadian galactic plane survey ( cgps ) ( brown , taylor , & jackel 2003 ) , and in the @xmath9 southern galactic plane survey ( sgps ) ( brown _ et al . _ , 2007 ) were recently used to probe both small - scale ( haverkorn _ et al .
_ , 2006 ) , and large - scale field structures ( brown _ et al .
_ , 2007 ) .
the best fit model of the latter has a clockwise field everywhere except for two reversals , one in the scutum - crux arm and the other in the molecular ring at a smaller galactocentric radius of @xmath10 3.7 kpc .
this contrasts with a pulsar - rm analysis by han _
et al . _ , ( 2006 ) , which concluded multiple field reversal zones in the galactic disk . in the current paper
we present an all - sky plot investigation of the milky way s smoothed rm pattern .
we use a new 2257-rm compilation of @xmath11 1500 revised and more accurate e.g.r.s .
rm s of our own , combined with the recently published cgps / sgps rm s , and two smaller published lists from ( klein _ et al .
_ 2003 ( 108 rm s ) , and mao _ et al .
_ 2008 ( 67 sources ) .
combination with the cgps and sgps rm surveys enables us to explore the match between the `` in - plane '' and `` off - plane '' rm sky . to an extent
not seen before , we find better defined , and remarkably coherent patterns of rm sign over the inner @xmath12 of galactic longitude of the milky way disk .
figure 1 shows an equal area , all - sky projection of the smoothed rm s . the rm smoothing used the method of simard - normandin & kronberg ( 1980 ) , in which an iterative calculation of the mean rm was performed at each ( @xmath1,@xmath3 ) , using all neighbor source rms within a 15@xmath13 radius .
the rm error was @xmath14 rad m@xmath15 for @xmath16 of our off - plane rm s , and @xmath17 rad m@xmath15 for @xmath18 of the same sample . at each iteration at a given source ( @xmath1,@xmath3 ) , rm values @xmath19 1.3@xmath20 were eliminated as `` outliers '' , and a minimum of 4 rm s was required to define a smoothed rm value at that ( @xmath1,@xmath3 ) .
we tested varying `` cutoff @xmath20 s '' for the optimum iterative removal of rm outliers .
the criterion was to minimize , at high @xmath21 , the distribution width of galaxy - corrected rms around 0 rad m@xmath15 .
we re - confirmed the 1.3@xmath20 value determined in simard - normandin & kronberg ( 1980 ) . in the galactic plane from @xmath22 to @xmath23
( figure 1 ) only one major sign reversal occurs , at @xmath24 310@xmath13 . an apparent contradiction to additional rm reversals on smaller scales by brown _
( 2007 ) is purely an effect of our smoothing half - width , @xmath25 , which averages over ( real ) smaller scale rm reversals .
this also applies to other zones near @xmath3 = 0@xmath13 , and at some higher latitudes ( figure 1 ) . in doing so , it more faithfully reveals major , _ large _ scale rm features near the milky way disk , which is the subject of this letter .
publication of the entire rm set and comparisons with higher latitude features is in preparation ( newton - mcgee , _ et al .
_ , in preparation ( 2010 ) ) .
the following analysis focuses on @xmath26 , and uses the rm data at our smoothing resolution .
some large scale patterns and symmetries not previously apparent become clearer at this rm resolution , as we show in the following section .
the smoothed , latitude - averaged rm s at @xmath27 are shown in figure 2(a ) .
the rm s in the two @xmath1-semicircles are first folded about @xmath28 .
when we then reverse the rm signs in one semi - circle ( figure 2(b ) ) a clear symmetrical / anti - symmetrical pattern emerges in the rm s . specifically , the _ forms _ of the smoothed / averaged rm s are strikingly similar but they have opposite signs on opposite sides of the galactic center .
now , having folded about @xmath1 = 0@xmath13 it is obvious that @xmath1 = 0@xmath13 is not the correct symmetry axis : an 11@xmath13 relative shift of the folded datasets needs to be applied @xmath7_i.e .
_ fold , invert sign , _ and _ shift . for optimum determination of this shift ,
we restricted the rm s to within @xmath29 to @xmath30 of the post - shift folding center .
figure 2(c ) shows the remarkably good overlap that results .
this shift corresponds to a pointing of the prevailing local b - field to @xmath31 , _
i.e. _ inward from the @xmath32 tangential .
independently , the same @xmath33 rotation applies to rm s at nearly all latitudes in the outer galaxy from @xmath3 @xmath34 to @xmath3 = @xmath35 ( figure 1 ) .
it agrees within errors to the less precise starlight polarization@xmath7derived field orientation ( heiles(1996 ) and references therein ) .
more interestingly it agrees with the general spiral arm orientation , making the milky way similar in this respect to well - studied nearby galaxies such as m51 , m83 , and ng6946 .
this , combined with the symmetries / antisymmetries seen in figure 2 indicates that an extragalactic observer would see a remarkably organized spiral structure in the milky way disk , with a pitch angle of 11@xmath13 in our region of the galaxy , and a prevailing disk field that is close to the general spiral arm direction .
this resolves longstanding disagreements on the magnetic vs. general arm comparison .
we confirm that the milky way behaves like other grand - design spiral galaxies in this respect .
outward of @xmath1 @xmath36 ( @xmath37 ) and of @xmath3860 @xmath13 ( @xmath39 ) , the galactic magnetic field has the same prevailing sign above , in , and below the galactic plane .
this includes the `` in - plane '' rm regions of the cgps and sgps ( @xmath8 and @xmath40 , respectively ) . in
virtually _ all _ @xmath1 ranges below the plane , the rm s keep the same prevailing sign right down to the south galactic pole .
these results confirm that the `` _ outer galaxy _ '' rm s taken by themselves are consistent with a very clear axisymmetric global field pattern .
_ inward _ of the above longitude boundaries ( @xmath6 280@xmath13 to @xmath6 60@xmath13 ) the prevailing rm signs in these same inner zones are opposite on either side of @xmath1 = 0@xmath13 ( figure 1 ) .
there are three further noteworthy features : first , as @xmath3 approaches @xmath41 , the rm(@xmath3 ) reverses sign between @xmath42 and 100@xmath13 and between 300@xmath13 and @xmath43 , a slightly wider @xmath1 range ( about @xmath44 ) than for the reversal at @xmath45 .
second , as @xmath21 increases above @xmath46 the sharp features and symmetries shown in fig .
2 gradually become less distinct .
third , when we repeat the procedure in fig . 2
but accept only the rm s at very low latitudes ( @xmath8 ) , the best fit prevailing field direction is exactly the same as in fig .
2 ( @xmath31 ) .
this shows that the large scale rm(@xmath1 ) variations in the plane are preserved up to at least @xmath47 .
as @xmath21 increases above @xmath46 , as mentioned above , the sharpness and clear shape of the patterns in figure 2 begin to `` dissolve '' .
we have discovered new and striking symmetries in the regular component of the galactic disk magnetic field using an expanded and revised compilation of 2257 e.g.r.s . rotation measures .
this new clarity results from 3 factors : ( 1 ) the increased numbers of rm s in zones immediately in and around the galactic plane , ( 2 ) the smoothing scale that we have chosen filters out smaller scale ( though not insignificant ) sign reversals and other smaller scale galactic magnetic features .
our analysis shows , in effect , that these have previously tended to `` mask '' the galaxy s true _ global _ galactic magnetic geometry . near the disk .
finally , ( 3 ) our latest rm determinations have many more accurate rm s than were previously available .
the rm variations in a 20@xmath13 band from @xmath48 to @xmath49 about @xmath1 = 0@xmath13 around the galactic plane show new clear and sharp reversals at our 20@xmath13 smoothing resolution .
this rm smoothing favors , on average , longer galactic pathlengths .
there is striking mirror symmetry about a reference direction 11@xmath50 from @xmath51 .
this gives a precisely determined , @xmath1 = 79@xmath50 mean magnetic field direction near the sun @xmath7 similar , but more precisely determined than several earlier observations beginning with simard - normandin & kronberg s ( 1980 ) estimate of @xmath4 . in a 3-d representation
, this spiral tilt appears to apply over the entire @xmath52 kpc @xmath53height of the milky way disk , at least in our broad galactic vicinity .
similarly , the clear galactic magnetic symmetries we have presented(fig .
2 ) appear to apply within the galactic disk scale height of @xmath54 kpc a consensus of several analyses over the past 3 decades .
the gradual `` softening '' of the rm(@xmath1 ) patterns at @xmath55 further suggests that these clear patterns are confined to within 1 - 2 kpc of the milky way s disk .
other patterns emerge as the galactic @xmath56 height increases . the improved definition of rm reversal zones are of particular interest for the propagation of charged ultra - high energy cosmic ray nuclei ( uhecr ) through our galaxy ( the pierre auger collaboration , 2007 ) . some of our field sign reversal boundaries may represent galactic zones where incoming ( extragalactic ) uhecr s could be minimally deflected in the galaxy s magnetic field .
the large scale disk rm patterns described here are a key basis for defining the galactic 3-d magnetic structure . in turn
, this will be crucial for interpreting the observed patterns of uhecr arrivals as a function of energy , charge and direction .
this work is in progress .
ppk acknowledges support from an nserc ( canada ) discovery grant , the u.s .
department of energy ( lanl ) , the hospitality of csiro , epping and prof .
b. gaensler at the university of sydney .
ppk and knm gratefully acknowledge the support from the australian research council federation fellowships of prof .
bryan gaensler ( ff0561298 ) and prof .
ron ekers ( ff0345330 ) . vs. @xmath1 .
panel ( b ) shows the grm s with reversed sign , _ and _ `` flipped '' about @xmath1 = 0@xmath57 .
panel ( c ) is similar to ( b ) except that instead of the 180@xmath57 rotation ( flip ) about @xmath58 , it is rotated about @xmath1 = 349@xmath57(-11@xmath57).,width=302 ] | we report new , remarkably coherent patterns of faraday rotation ( rm ) at @xmath0 in the inner galactic plane , using an expanded extragalactic source rm compilation of unprecedented average accuracy .
the patterns , relevant to rm pathlengths toward the inner galactic disk , clearly indicate a global disk magnetic field structure , with remarkable reflection symmetry in rm .
sharply defined rm(@xmath1 ) features replicate with the opposite sign on opposite sides of the galactic center , confirming the bi - symmetric magnetic field pattern _ inward _ of the sagittarius - carina arm originally found by simard - normandin and kronberg ( 1979 ) .
the prevailing magnetic field points to @xmath2 , very close to the general spiral arm direction .
additional sharp rm sign reversals in @xmath3 , just below the galactic plane , are consistent with an a0 galactic dynamo configuration .
at all `` outer '' longitudes , there are no large scale rm sign reversals either _ at _ the galactic plane or near to it .
the outer galactic field pattern is thus globally axisymmetric , consistent with that seen in many external spiral galaxies .
our galaxy thus contains a mix of an axisymmetric ( outer ) and bisymmetric - like ( inner ) field configurations and would appear highly `` patterned '' to an extragalactic observer .
this fact is of particular significance for interpretation of the `` energy / species / arrival directions '' of ultra high energy cosmic rays ( uhecr ) . |
wondering what may be learned about active galactic nuclei ( agn ) in the coming years requires that one thinks about what is currently known and about the type of knowledge that is relevant .
indeed , agn seem to be complex systems that need not be described at a level at which the general physical processes are hidden behind local and/or statistical fluctuations .
the learning process may in some sense be similar to understanding the physics of the terrestrial atmosphere by doing meteorological observations .
the aim in this case is not to describe the local weather but to extract from the complex descriptions of the observations the guiding principles .
clearly , what knowledge or understanding one wishes to obtain on any given subject is a matter of personal taste . indeed
many people , even among astronomers , might have only a limited interest in agn .
the following account is therefore a personal approach to the subject . among the facts that are well established now and documented in a number of books and reviews in recent years
one finds : the source of energy : : is the gravitational energy of matter falling into a massive black hole . + with this general principle one deduces two characteristics of agn , namely an estimate of the mass of the central black hole through the eddington luminosity ( more appropriately , this estimate gives a lower boundary to these masses ) .
this mass is found to be up to @xmath0 solar masses . also from the luminosity of the objects and from reasonable assumptions on the efficiency with which accreted matter can be made available to be radiated
one deduces the rate at which matter must be accreted into the central black hole .
rates of up to 100 solar masses per year for the brightest nuclei are found .
+ the role of the rotation of the central black hole in the global energetics still needs be understood .
the emission processes : : through which energy is radiated are now mostly identified .
they include synchrotron emission , emission of dust heated by ultraviolet and x - ray radiation fields , compton processes ( be they thermal or due to relativistic electron distributions ) and emission from optically thick media ( black body radiation ) .
it is possible from the identification of the emission processes and the shape of the spectral energy distributions to deduce ( some of ) the physical conditions of the emitting material .
the state of the gas : : is known for the gas that emits the lines charateristics of most agn types .
temperature , density and filling factors as well as the origin of the ionisation can be read from the spectra and provide a picture in which photoionisation plays a dominant role .
aberrations and geometry : : play an important role through doppler boosting of components related to relativistic jets and through absorption of some components that happen to lie behind dense clouds of dust and gas .
these effects considerably modify the appearance of agn and probably lead to our agn classification scheme .
agn sit in the center of galaxies : : whith which they must have intense relations , be it only through material that is falling into the galactic potential well to the agn .
the evolution : : of agn is well documented . the agn number density and/or luminosity have greatly changed over the age of the universe .
agn have a very rich phenomenology : : for which we have large amounts of data .
we thus have descriptions of many spectral energy distributions , of many types of variations in different objects and at different wavelengths and we have a vast zoology of objects in which populations differ greatly in one or several characteristic parameters
. we are not always succesful in understanding this rich phenomenology in terms of underlying physical processes .
we know that reprocessing of radiation by a medium takes place in agn .
examples are the emission of heated dust ( reprocessed uv radiation ) , compton processes in which soft photons aquire additional energy from hot electrons and the presence of a fluorescence fe line among others .
what still needs be quantitatively assessed is the relative importance of the so - called compton reflection hump that was suggested by @xcite to be important between 10 and some tens of kev .
this hump is due to the reflection of a primary x - ray component ( taken as a power law ) by cold material . that this causes a hump
can be qualitatively understood by noting that the primary component is absorbed by the medium at low energies ( the medium is assumed optically thick ) while at high energies the scattering cross section decreases due to klein - nishina effects ; see @xcite for a review .
there will therefore be a region in the spectrum where the reflection is maximal , thus producing a broad hump on top of the primary component .
it is suggested in models of the ultraviolet emission of agn ( see next section ) that the uv emission is due to a disk heated by the primary x - ray component rather than by gravitational energy released within the ( optically thick ) accretion disk . in this scenario
the reflection hump observed in the medium energy x - rays , the fluorescence line and the uv emission are all due to the same reprocessing scheme .
all three elements should be consistently observed and variations in these components must be closely correlated and related ( albeit possibly in a complex way ) to the variations of the primary x - ray component .
observations spanning the ultraviolet to hard x - ray domain can provide measurements that will allow us to separate the primary component from the reprocessed ones , to measure the flux in all components as a function of time and to seek some geometrical information by looking at the relative intensity of the components in different classes and as a function of time .
these observations will be possible using xmm and integral in a coordinated manner over the coming decade on a variety of objects .
the peak in emission in the blue - ultraviolet region of the spectrum that is observed in seyfert galaxies and quasars has been associated with the accretion process since 1978 @xcite .
this suggestion has given rise to a large activity in the subsequent years to provide a good match beteen observations and accretion disk theory . despite 2 decades of work
the concordance between the observations and theory is still not satisfactory .
the continuum spectral energy distribution , spectral features , polarisation , relationship between continuum shape and luminosity and , possibly most importantly , the variability properties of the blue bump do not match those expected from an optically thick medium inside which gravitational energy is released to be radiated at the surface of the medium ; see @xcite and the review on agn accretion disk related problems in @xcite .
the blue bump flux is observed to vary almost simultaneously at all wavelengths from the visible to the ultraviolet range in 3c 273 as well as in seyfert galaxies @xcite .
this is in contradiction with the expectations that disturbances that propagate through the accretion disk ( or whatever other structure ) do so on viscous timescales that are many order of magnitude longer than the observed lags or upper limits thereof
. this has lead to the development of models in which an x - ray source shines onto the disk .
the latter is thus heated from the outside and the illuminated regions respond to variations of the primary source on timescales that are given by light propagation effects in good agreement with measurements of lags @xcite .
one of the most elaborate models of this type is that of @xcite in which a multiple corona shines onto the disk .
it may be worth mentioning that the observations require only that the signal responsible for the perturbances propagate with the speed of light , not necessarily that the energy be distributed with the speed of light . in a detailed study of the visible and ultraviolet emission of 3c 273 @xcite
have shown that the short wavelength emission of the blue bump in 3c 273 can be well modeled by a series of independent events occuring at a rate of about 3 per year and having each an energy of some @xmath1ergs .
if this approach is physically sound ( a good mathematical description of the data is not sufficient to ascertain the physical validity of the assumptions ) than several questions must be answered : what is the origin of the events ( see next subsection for a proposal ) ? what about the long wavelength emission of the blue bump ? can the analysis be generalised to other objects ?
be it in the frame of the accretion disk models with or without a corona or in that of the individual events , there are still fundamental questions to be solved to explain the nature of the blue bump emission .
accretion disks @xcite are a natural gas accretion structure in that they provide a means of expelling angular momentum while the material is being accreted provided that the nature of the viscosity is understood . the difficulties of the accretion disk models to explain the observed properties of the blue bump ,
the description of the blue bump in 3c 273 in terms of individual events and the presence of stars in the center of galaxies has lead us to study a model in which the accreted material is not a diffuse gas but rather in the form of stars @xcite , @xcite .
stars orbiting a supermassive black hole at some 100 gravitational radii have velocities close to 0.1c . when two such stars collide their kinetic energy amounts to some @xmath1ergs for stars of the mass of the sun .
this is in good agreement with the energy contained in one event as discussed in the previous section . in order to have few collisions per year ( to provide the event rate described above ) and to explain the average luminosity of the quasars there
must be approximately the same mass in the central black hole and in the stellar population within some 100 s of gravitational radii . in this volume
the star density must be therefore very high , of the order of @xmath2 stars per cubic parsec .
accretion of matter in the form of stars and dissipation of the kinetic energy through stellar collisions rather than in a disk has the property that the dependence of the variability on the object luminosity discussed in @xcite based on iue observations can be explained .
this dependence is much flatter than expected based on independent events that are poissonian distributed in time , suggesting that the event properties are a function of the average luminosity of the object .
the star accretion scenario provides this naturally , because the average star collision energies depend on the distribution of stars with distance to the black hole , which also determines the average luminosity of the object @xcite .
much work is still needed to assess whether this accretion process has a good prospect to contribute to the luminosity of agn in a significant way .
the emission processes occuring during and after a collision must be calculated and compared with observations of spectral energy densities and variability , the dynamics and evolution of very dense clusters in the vicinity of a massive black hole must be studied .
an alternative line of work may be to prove that this model is irrelevant by showing that the accretion material is organised in a single plane and rotating all in the same direction .
starburst activity is observed in relation with agn ( the luminosity of which comes from accretion processes ) .
this type of link can intuitively be expected .
as the material falls in the central regions of a galaxy it will dissipate angular momentum .
this will stir the interstellar material and provide conditions that are propitious for star forming activity .
starburst activity with a luminosity that is comparable to that of the agn seems to be common among absorbed seyfert galaxies ( seyfert ii galaxies ) . both @xcite and @xcite report that at least close to half of the seyfert ii galaxies that they observed have important starburst activity . on the other hand ,
@xcite report that none of the 8 unobscured seyfert galaxies ( seyfert i galaxies ) they have observed showed starburst properties . that the difference between seyfert i and seyfert ii galaxies lies in the presence of absorbing material that obscures the nuclei of the seyfert ii galaxies is in no doubt based on the observation of broad lines in the polarised light of seyfert ii galaxies @xcite and the clearly absorbed x - ray emission of the same .
the observations described above , suggest further that the presence of the obscuring material is related to the star forming activity .
this link may indeed be expected as star formation requires the presence of large quantities of gas and dust forming dense clouds that are also associated with higher probabilities of obscuration .
the observations discussed here are in marked contrast with the expectations of the unified models @xcite which postulate that seyfert i and seyfert ii galaxies are differentiated by their orientation with respect to the line of sight .
the model suggested by the different level of starburst activity in both types of galaxies is indeed that the nuclei are similar but that the presence of obscuring material is a genuine difference in the two classes of objects .
the role of orientation effects is then probably secondary and needs be studied with the help of complete samples .
agn and quasars come in a great variety of disguises that led to a complex classification .
this is due in great part to the fact that some properties are enhanced when associated with a relativistic jet pointed towards us ( and correspondingly weakened when the jet is pointed in other directions ) while some other aspects may be hidden by obscuring material as noted in the previous subsection .
it is worth noting that some of the differences are certainly intrinsic to the objects and not simply related to geometrical and/or doppler boosting and/or absorption . among those intrinsic differences is the presence of the blue bump and the existence of the jet in the first place . in order to understand what are the intrinsic differences
as opposed to viewing or orientation effects and in order to measure the effects related to viewing and orientation it is mandatory to obtain complete samples of agn .
this task is proving difficult because of the difficulty to find parameters that are indeed independent of orientation and absorption and that are unique signatures of agn activity .
the presence of hard x - ray radiation is for example a clear signature of agn activity and , above the energy at which absorption is important , is rather insensitive to the presence of obscuration material .
the x - ray emission of radio loud quasars is , however , more important than that of radio quiet quasars @xcite .
this may be due to the fact that a fraction of the x - ray emission is linked to relativistic jets .
thus samples based on hard x - ray flux measurements of agn whilst unbiased towards obscured or unobscured agn will contain biases related to the presence or not of jets and to their orientation with respect to the line of sight .
@xcite have computed the continuum spectral energy distribution of a small sample of quasars with very different radio properties for which far infrared data from iso have been measured .
this analysis which is based on a sample that is not complete suggests that the far infrared emission is the emission component that is least affected by biases .
this may indeed be expected since dust emission occurs on large scales , is isotropic and is unrelated to the jet .
nonetheless if confirmed this implies that the dust properties of quasars are independent of their class , a non trivial result .
based on this preliminary result we suggest that , at least for quasars , complete samples should be based on far infrared measurements of objects that are identified by their x - ray emission to be agn rather than starburst galaxies .
this approach may not work in the case of bl lac objects in which the far infrared emission is dominated by synchrotron emission .
several authors ( also in this volume ) have noted the similar shapes of the star formation history in the universe and of the qso or agn evolution .
this similarity is then taken as a strong suggestion that star formation and agn activity are intimately linked in the development of galaxies .
@xcite has compared the the radio luminosity of quasars as a function of redshift with the star formation history in the universe and has come to the conclusion that there are @xmath3 solar masses of stars formed per solar mass accreted into a black hole in a radio loud quasar . as far as energetics is concerned a 10 solar mass star is representative of a population and generates about @xmath4 ergs over its lifetime ( maeder private communication ) .
this means that for @xmath3 solar masses some @xmath5 ergs are generated .
the energy liberated by the accretion of one solar mass assuming an efficiency of 10% is @xmath6 ergs .
it follows that the ratio of accreted energy generation onto massive black holes in radio loud quasars to that of nucleosynthesis in the universe is @xmath7 . in order to convert this ratio into a global estimate of the total rate of energy production by accretion onto all massive black holes to that due to nucleosynthesis
it would be necessary to know the density ratio of agn to the radio loud qso s used in dunlop ( 1998 ) and their respective average luminosities .
@xcite have used the hard x - ray background and generic agn spectral energy distributions to estimate the total contribution of agn to the photon flux and compared this with the photon flux of star formation .
they conclude that the total power emitted by accretion is about 1/5 that due to stars in the universe .
a further comparison between star formation history and quasar evolution is presented in @xcite and can also be used to derive that the ratio of accretion to star formation power is of the order of @xmath8 .
the close ties between the star formation and agn luminosity in the universe as it evolved provides a fascinating example of the interconnection of phenomena which many of us would have thought independent . a small fraction of agn are radio loud .
it would indeed be interesting to know this fraction with some precision as a function of the luminosity and redshift of the agn .
the orgin of this difference can not be sought in absorption as radio waves are undisturbed by neutral gas and dust , nor can it be sought in orientation and doppler boosting effects , because radio galaxies show that a significant fraction of the radio luminosity is emitted isotropically .
the presence or absence ( or at least extreme weakness ) of radio emission is therefore intrinsic to the objects .
radio emission is related to the presence of jets in the agn .
the origin of the radio emission is therefore related to the presence of jets .
the question may therefore be re - phrased to asking why a small fraction of agn have jets and others not .
the presence of magnetic fields and/or the rotation of the black hole certainly play a role in these question .
evidence for black holes in galaxies has now become strong , in particular in the center of our galaxy @xcite .
the luminosity of the galactic center , and that of the nuclei of other galaxies in which black holes are probable , is , however , orders of magnitude less than the eddington luminosities inferred from the masses of the black holes .
the presence of the black hole is , therefore , not a sufficient condition for the presence of a bright agn in a galaxy .
there must be in addition to the black hole a sufficient quantity of matter that can be accreted in order to turn the central black hole of a galaxy into an active agn .
this matter can be in the form of diffuse gas accreted through a disk or in the form of a sufficiently dense cluster of stars surrounding the black hole .
@xcite have recently shown by observing with high resolution and high signal to noise the broad lines of 3c 273 that the number of clouds is in excess of @xmath9 .
this number is larger than the number of stars that have been suggested as a source of confinement for the clouds @xcite , @xcite , @xcite .
it is therefore probable that models in which the broad line region is seen as a continous hydrodynamic medium will have to be preferred to those in which a collection of well defined clouds exist for a significant time .
see @xcite for a list of references to such models .
all the questions discussed above have to be addressed with observations in several wavebands .
this is even true for the question of the broad line cloud confinement in which a possible confining medium could have been detected through x - ray observations
. we will be in a very priviliged position in the next years with respect to multi - waveband coordinated observations of agn with the simultaneous presence in orbit of high energy observing satellites that will cover the energy domain from 10s of gev ( agile , ams ) to mev and 100s of kev ( integral ) , to soft x - rays and ultraviolet ( xmm , chandra , spektrum x - g ) .
these high energy instruments are complemented with a large array of very performing ground based telescopes in the visible , near infrared and radio domains of the spectrum and of the satellite sirtf in the far infrared domain .
we should thus be in a position to address many of the questions above in a very comprehensive way .
one can reasonably expect that well planned observations using all these facilities should allow us to solve the question of re - processing , that of the relationship between starburst and agn activity and to obtain well defined complete samples in a convincing way
. we should then be able to make very significant advances on many of the other topics mentionned here and in particular on the way in which material is accreted and the nature of the blue bump . to be able to , and to have to , formulate the questions mentionned here
is both a tribute to those who have pioneered the field of agn , lo woltjer being prominent among them , and a sign that the subject is difficult and has been progressing only slowly once the bases were laid .
it is our hope that the terms in which the questions are expressed and the tools that are now available will allow us to progress .
we are well aware , though , that we are not the first to express these hopes and that observations of agn have often provided new questions rather than answers to those they were expected to solve . | we discuss how several of the questions that remain unclear on the physics of active galactic nuclei may find elements of answers when using in the coming years the extraordinary set of instruments that will be flying simultaneously to observe in all bands of the electromagnetic spectrum .
the choice of questions mentioned here is personal and not exhaustive . |
short gamma - ray bursts ( sgrbs ) are historically identified based on the duration of their gamma - ray emission ( @xmath6s ) and their hard spectrum ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
recently , based on a statistical approach , several attempts have been made to improve this classification @xcite .
short - grbs also differ from the `` long '' grbs class for their redshift distribution and , likely , their progenitors ( see , for example , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* and references therein ) .
the optical afterglows of sgrbs are on average two orders of magnitude less optically luminous than their long duration counterparts @xcite , making broadband follow - up , and optical spectroscopy in particular , quite challenging .
nevertheless , the _ swift _ satellite @xcite has enabled the localization of a modest sample in x - rays and a smaller set have been detected in optical / near - ir passbands ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
in contrast , long grbs ( @xmath7s ) present brighter afterglows allowing accurate localization and spectroscopic follow - up hours after the events occur .
robotic facilities and target of opportunity ( too ) programs have provided a plethora of photometric and spectroscopic data in support of theoretical models of long grb progenitors and the host galaxies they live in .
these data first established conclusively the extragalactic nature of the events @xcite and , eventually , analysis of the forest for high-@xmath8 events ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) and fine - structure transitions at lower redshift provided unambiguous physical associations to their hosts .
optically bright long grbs seem exclusively hosted by star - forming galaxies with high specific star - formation and sub - l@xmath9 luminosities , indicating massive stars as likely progenitors of long grbs ( see * ? ? ?
* and reference therein ) , while `` dark '' grbs @xcite seem to be harbored , on average , in more massive and highly star - forming galaxies ( @xmath10m@xmath11 at @xmath12 , * ? ? ?
thanks to a concerted community effort , of the @xmath13 short grbs detected by @xmath14 have been localized to within a few arcseconds accuracy , and , with a _ a posteriori _ probabilistic arguments , have been securely associated with nearby galaxies ( see , e.g. , * ? ? ?
* ; * ? ? ?
the number of these events , however , is still in the few dozens @xcite .
the few well - observed sgrbs have been associated primarily to a population of galaxies very similar to field galaxies at similar redshifts with moderate to negligible sfr , lending credence to the idea that at least some sgrbs explode with delay times @xmath15 gyr consistent with the progenitor model of compact mergers ( e.g. , neutron star binaries , or neutron star - black hole * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
recently , using deep observations from space and from the ground it has been possible to quantify the relative fraction short - duration grb hosts : @xmath16 early- and @xmath17 late - type galaxies @xcite . from the afterglow perspective
, there has yet to be a _ bona fide _
short duration grb afterglow for which we have measured a redshift from absorption features in the optical spectrum . among these it is important to note the long grb090426 , for which the duration of the prompt emission ( @xmath18 = 1.2s ) and the properties of the host made its classification uncertain @xcite .
another debated short grb for which an afterglow+host galaxy spectrum has been obtained is grb 051221a @xcite : also in this case the non - collapsar nature of this event has been recently questioned based on probabilistic arguments and an accurate analysis of the instrumental biases which may lead one to mistakenly associate collapsar events like this one to short - grbs @xcite .
finally , grb100816a , despite having @xmath19s , has been associated with the sgrb class based on lag analysis @xcite and its afterglow ( or a combination of afterglow and host ) has been spectroscopically observed @xcite .
the lack of a large sample of sgrb afterglow spectra has made it impossible to conduct analogous studies of the host galaxy ism and circumburst environment , as are routinely achieved for long - duration events .
finally , the lower rate of space - based localization ( compared to the long grb class ) and their faintness demand a very rapid response by large area telescopes to reach the quickly fading afterglows and obtain similar high - quality data for a larger sample of sgrbs .
such a collection will serve two main goals : 1 ) provide unambiguously , based , e.g. , on absorption - line diagnostics like fine - structure transition , the redshift of these events and their association with an underlying host galaxy ; 2 ) directly allow the characterization of the interstellar and circumstellar environment from 200 - 300pc up to the host halo @xcite and , then , provide clues about the progenitor itself .
in this letter we present our prompt optical spectroscopic and photometric observations of the short , obtained at the gemini south telescope . based on the optical and near infrared emission at late time ( @xmath20 days post - burst ) derived by deep _ hst _
observations which allowed to clearly resolve the grb - host complex , it has been proposed that this event might resemble the expected emission due to the decay of radioactive species produced and initially ejected during the merging process of a neutron star s binary system , referred to as a `` kilonova '' @xcite .
thanks to our too program , we observed this event within the first day , when the afterglow dominated the host flux despite their small projected separation .
this provides constraints on the event redshift and the properties of the grb explosion site , in particular , in comparison to the overall sgrb host galaxy population .
the paper is divided as follows : in [ sec : data ] we present our observing campaign ; in sec .
[ sec : analys ] we present our spectral analysis , and in [ sec : results ] we discuss our results and compare them with previous studies on sgrbs and their host galaxies .
finally , in [ sec : conclusions ] we present some implications on the nature of and the possibilities offered by rapid response facilities for sgrb studies . throughout the paper
, we will use the standard cosmological parameters , @xmath21 km s@xmath22 mpc@xmath22 , @xmath23 , and @xmath24 .
on june 3 , 2013 , at @xmath25=15:49:14 ut , the satellite @xcite triggered on @xcite .
the on - board burst alert telescope ( bat ; * ? ? ?
* ) detected a single bright peak with a duration of 0.4 seconds , placing this event unambiguously in the short - grb category @xcite .
after slewing to the source location , the telescope ( xrt ; * ? ? ?
* ) , began observing 59s after the trigger , and detected a fading counterpart at @xmath2628@[email protected] , @[email protected] ( with an uncertainty of 2.7 ; * ? ? ? * ) .
no optical counterpart was found in the uv / optical telescope ( uvot ; * ? ? ?
* ) prompt data @xcite , while a counterpart was detected when more data became available @xcite . at @xmath31 hours
, using the william herschel telescope , @xcite identified a point - like source inside the xrt error circle , which they determined to be the optical counterpart of the short .
this position lies in the outskirts of a galaxy present in the archival sloan digital sky survey dr9 @xcite .
other subsequent follow - up observations were performed on yielding a spectroscopic redshift for the afterglow and/or the host galaxy @xcite .
the afterglow was also later detected at radio wavelengths @xcite .
unfortunately , no observations to date have been providing detections of fine - structure lines , which are connected to the grb radiation itself ( leaving still some uncertainty on the actual grb - host association ) .
using our too program ( gs-2013a - q-31 ; pi cucchiara ) we performed a series of photometric observations with the gmos camera @xcite on gemini south in the , , and filters for a total of 8x180s exposures per band from @xmath32h to @xmath33h .
the data were analyzed using the standard gemini / gmos data analysis packages within the iraf environment .
the afterglow is detected at a projected distance of @xmath34 from the center of a bright , neighboring galaxy ( see the false - color image in fig . [
fig : finder ] ) . at the galaxy s redshift ( see next section ) , this corresponds to 4.8 ( @xmath35)kpc in projected distance .
subsequently , we obtained a spectroscopic sequence with the same instrument : we obtained 2x900s spectra , using the r400 grism with the 1 slit ( resolution of about 5.5 ) centered at 6000 , covering wavelengths @xmath36 .
we reduced the spectroscopic data with standard techniques , performing flat - fielding , wavelength calibration with cuar lamp spectra , and cosmic ray rejection using the lacos_spec package @xcite .
a sky region close in the spatial direction , but unaffected by the spectral trace , was used for sky subtraction .
the two - dimensional spectra were then coadded .
figure [ fig : twod ] presents the processed data which reveal the spectrum of the extended galaxy exhibiting a faint continuum and a series of emission lines .
clearly visible superimposed on the galaxy light is an unresolved ( spatially ) trace that coincides with the expected position of the afterglow based on our imaging data . using the iraf / apall routine
, we extracted a spectrum corresponding to the entire detected trace , therefore including both the galaxy and the grb afterglow signal .
we then extracted a second spectrum with the aperture restricted to the spatial location of the grb afterglow .
variance spectra were extracted in both cases evaluating sky contribution in two regions unaffected by the galaxy or the grb light ( plotted in grey in fig .
[ fig : comparis ] and with dash lines in fig .
[ fig : twod ] ) .
flux calibration was performed using an observation of the spectrophotometric standard star feige 110 taken with the same instrument configuration , although no correction has been made for slit losses .
an air - to - vaccum correction was applied to the 1d wavelength solution . using the measured optical brightness of the grb+host we also corrected for a small slit - loss ( @xmath37 ) . on the following day ,
starting at @xmath38 days after the burst , we observed again the field with gemini south obtaining @xmath39s imaging exposures with the gmos camera in the , , bands and a single 900s spectrum with the same configuration used the first night .
the afterglow had faded considerably in comparison with the host galaxy .
the reduced spectrum revealed no trace from the grb , but only a faint continuum from the host , with identical emission lines superimposed .
we also obtained an optical spectrum of the host galaxy of on 2013 june 6 ut with the deep imaging multi - object spectrograph ( deimos ; * ? ? ? * ) mounted on the 10 m keck ii telescope .
the instrument was configured with the 600 lines mm@xmath22 grating , providing spectral coverage over the region @xmath409500 with a spectral resolution of 3.5 .
the spectra were optimally extracted @xcite , and the rectification and sky subtraction were performed following the procedure described by @xcite .
the slit was oriented at an angle such to include the host nucleus and the grb location .
flux calibration was performed relative to the spectrophotometric standard star bd+262606 .
finally , on june 16 ut , we imaged the field with the gemini - south telescope in , , , and bands to estimate the galaxy contribution at a time when the afterglow was expected to have faded well beyond our detectability .
figure [ fig : comparis ] presents extractions associated with the combined afterglow and galaxy and for an aperture restricted to the afterglow location . in the combined spectrum , we identify a series of nebular emission lines ( e.g. @xmath41 , @xmath424959,5007 , and ) , at a common redshift @xmath43 .
we associate these lines to regions near the center of the galaxy , spatially offset from the afterglow location ( fig .
[ fig : twod ] ) . the only significant absorption features present in the spectrum are h+k features , despite the fact that the flux and s / n are lower ( see sec . [
sec : ot ] ) the narrow aperture centered on the afterglow location shows a largely featureless continuum ( lower panel of figure [ fig : comparis ] ) . at wavelengths
@xmath44 the two - dimensional spectrum and our imaging indicate the afterglow dominates the flux .
we measure a spectral slope in the optical of @xmath45 with @xmath46 . examining the grb afterglow spectrum at the wavelengths of the galaxy s nebular lines we detect little or no emission ( see zoom - in regions in fig . [
fig : comparis ] ) .
upper limits on the line fluxes are given in table [ tab : host ] .
figure [ fig : hk ] shows a section of the two - dimensional spectrum centered on the location of the doublet ( dashed lines represent the variance spectrum ) . in the outset we highlight this region in the corresponding extracted spectra for the afterglow only and the afterglow+host : at the location of the redshifted h&k we see a decrement in flux in the two - dimensional spectrum and in both extracted spectra . using the variance spectra
we can determine the pixel - by - pixel flux error .
at the expected location we detect a redshifted k - line at @xmath47 significance level in the afterglow+galaxy spectrum ( @xmath48 in the afterglow only ) , while the h - line only at @xmath49 ( and @xmath50 .
this result , in combination with the emission lines detected , places strong constraints on the grb redshift ( @xmath51 ) and suggests a likely association of the grb with the galaxy .
our second and third imaging epochs do not show any clear sign of the afterglow either in the single exposures nor in coadded ones . assuming that the afterglow completely faded below our detection limit at @xmath52 days we subtract from our first epoch of imaging this `` reference '' to measure the afterglow flux at the time of our initial observation . using the hotpants code and nearby point sources from sdss for calibration , we measure the following , extinction corrected , ab magnitudes : @xmath53 , and @xmath54 .
we derive , at @xmath55 days , an afterglow brightness of @xmath56 , after correcting for galactic extinction ( @xmath57(b - v ) = 0.02 ) : this value is similar to other optically - detected short grbs observed around the same epoch @xcite .
moreover , using the data from june 16 as a `` reference '' does not change these results , while a similar procedure operated between our second and third epochs , allows us to place stringent upper limits on the afterglow brightness at @xmath58 days ( @xmath59 and @xmath60 ) , confirming our assumption that indeed the optical afterglow was undetectable by gemini .
interestingly , almost at similar time of our latest gemini observation , a near - ir counterpart was still visible in the @xmath61 data as reported by @xcite and @xcite .
the peak of the galaxy emission and the centroid of the afterglow emission derived from the subtracted image ( see last panel on the right in fig .
[ fig : finder ] ) are separated by @xmath62 , corresponding to a projected distance of 4.8kpc ( @xmath35 ) at @xmath3 . at this redshift
the probability of a random association between the galaxy and the grb at such distance @xmath63 is very small ( @xmath64 ; see also * ? ? ?
based on our @xmath65 photometry we derive an optical spectral index @xmath66 ( assuming @xmath67 ) consistent with analysis of our optical spectrum and similar to other sgrbs @xcite . at @xmath3 the band magnitude samples the rest - frame b - band luminosity of the host .
therefore , using our second epoch observations we derive = 20.76 @xmath68mag for the grb host and we estimate a rest - frame , @xmath69-corrected , absolute @xmath70-band magnitude ( ab ) of @xmath71 . despite the brightness of this galaxy ( @xmath72 ; * ? ? ?
* ) , the derived luminosity is not unusual among the short - grb hosts population ( figure [ fig : sfr ] , * ? ? ?
* ) . using the late - time multiband photometry and the idl package _ kcorrect _ @xcite in a similar fashion to @xcite we estimate the mass of the host galaxy as @xmath73 . the gmos spectra from the first two nights and
the deimos data cover key nebular emission lines .
we measured the fluxes of these lines using the latter data , but for the ones also detected in the gmos spectra we obtain similar results ( table [ tab : host ] ) .
we apply standard emission - line analysis to derive intrinsic galaxy properties .
we have estimated the optical extinction using the balmer lines decrement and assuming case - b recombination @xcite : we derived e(b - v ) = 0.43 ( @xmath74 mag , assuming a milky - way extinction curve ) . using the calibration of @xcite ,
the [ oii ] line luminosity gives a sfr([oii ] ) @xmath75 yr@xmath22 .
similarly , from the luminosity we derive sfr ( ) @xmath76 @xmath5 . together with the @xmath70-band lumiosity , we derive a specific sfr : ssfr = sfr@xmath77 yr@xmath78 . as in other many short grb hosts ,
this value is consistent with a star - forming galaxy ( see fig.[fig : sfr ] , berger 2009 ) .
also , collisionally - excited oxygen and the and balmer series recombination lines provide an estimate of the gas - phase oxygen abundances of the host galaxy .
we adopt the @xmath79)/@xmath80 metallicity indicator @xcite , which depends on both the metallicity and the ionization state of the gas .
to help disentangle a degeneracy in the values , we use the @xmath81 indicator ( @xmath82/@xmath83 ) .
however , our measured fluxes still allow for two solutions : @xmath84 for the upper @xmath85 branch and @xmath86 for the lower branch .
the typical error in this measurement , due to systematic uncertainty in the calibration of these metallicity relations is typically @xmath87 dex .
no field galaxy at @xmath88 with a similar brightness to exhibits metallicities consistent with the lower - branch of our @xmath85 analysis .
furthermore , based on the detection of and @xmath89 lines in the deimos spectrum we place a lower limit on the metallicity of 12 + log(o / h ) @xmath90 .
so we can conclude the galaxy has approximately solar metallicity @xcite .
grb130603b marks one of the best cases in which , for a short grb , an afterglow - dominated spectrum has been obtained .
overall , the extracted grb afterglow spectrum presents a smooth , almost featureless continuum ( figure [ fig : comparis ] lower panels ) . despite the absence of strong spectral features
, we can still constrain the grb redshift from these data alone : 1 ) the lack of a forest requires that the grb exploded at @xmath91 ; 2 ) the appearance of weak h+k absorption lines ( see figure [ fig : hk ] ) as also seen by @xcite and @xcite , gives a strong indication that the grb redshift is at least a the the emission line redshift ( @xmath92 ) ; 3 ) if one assumes the afterglow spectrum will exhibit significant ( @xmath93 ) absorption from the surrounding interstellar or circumgalactic medium then we set an upper limit @xmath94 based on its non - detection .
all of these constraints are consistent with having occurred within the coincident galaxy .
ultimately , an indisputable measurement of @xmath95 from an afterglow spectrum may require the detection of fine - structure transitions excited by the grb itself @xcite and/or the observed termination of the forest .
unfortunately , our wavelength coverage and the small number of this kind of features that would have been redshifted in our observed window ( and therefore detected ) prevent us to obtain such a secure measurement .
furthermore , analysis of these transitions would reveal physical properties of the progenitor environment ( e.g. metallicity , size ) and may offer new insight into the progenitors of sgrbs . associating to the neighboring galaxy , we may place an upper limit to the star - formation rate at the position of the event .
no strong emission lines are identified , in contrast to the putative host spectrum : the striking difference can be seen at the location of the and @xmath424959,5007 lines ( insets in fig .
[ fig : comparis ] ) .
this indicates that , unlike the majority of long - duration grbs , exploded in a region of very minimal , if not negligible , star - formation .
based on the integrated flux at the location of the line , we place an upper limit sfr ( ) @xmath96 yr@xmath22 .
this provides further support for models unassociated with massive star formation .
we present our gemini rapid follow - up campaign on the afterglow of the short . despite the intrinsic faintness of these events ,
the low - energy emission of , in particular in the optical afterglow , was still detectable several hours after the explosion .
we triggered our approved too program at the gemini south telescope and obtained a series of images of the grb field as well as spectroscopic observations with the gmos camera starting 8 hours after the burst .
we repeated similar observations the following night . based on our absorption lines analysis and the very small probability of a random association with such a bright galaxy ( @xmath97=0.00064 )
we conclude that this is the host of .
our optical images provide firm evidence that this event occurred in the outskirt of a star - forming galaxy ( @xmath98 ) , around 4.8kpc from its center ( at the grb redshift of @xmath99 ) . while this offset is not unusual for long duration grbs ( e.g. , * ? ? ?
* ) , the lack of blue light at the location is very unlike long - duration grb locations @xcite .
deep @xmath61 observations also confirm this findings @xcite .
the early - time spectra present a bright afterglow continuum superimposed upon a fainter galaxy .
this represents one of the few cases where an afterglow - dominated spectrum for a short grb has been recorded .
the host is an @xmath72 galaxy , and the spectrum shows strong nebular emission lines as well as recombination features .
we measure a dust - corrected sfr @xmath76@xmath5 and solar metallicity .
nevertheless , we were able to extract the grb spectrum and constrain the properties of the host galaxy at its location : the grb spectrum present a smooth continuum showing almost no sign of emission lines throughout the wavelength coverage .
also , at the same redshift we identify absorption lines , which place a strong constraint to the grb redshift , despite the lack of fine - structure transitions .
therefore , at the grb location , the star - formation activity is almost negligible ( sfr ( ) @xmath96 yr@xmath22 ) .
this last result , once more , emphasizes the importance of rapid follow - up observations with large aperture facilities in order to firmly identify the faint afterglows of these short - lived events . only with a larger sample of data similar to those presented in this work
we will be able not only to identify a larger number of short grbs and measure their redshifts , but also to characterize , via absorption spectroscopy , their explosion site environment and move forward in our understanding of the nature of their progenitors .
ac thanks the anonymous referee for the valuable comments and suggestions which have helped significantly improving the manuscript .
gemini results are based on observations obtained at the gemini observatory , which is operated by the association of universities for research in astronomy , inc .
, under a cooperative agreement with the nsf on behalf of the gemini partnership : the national science foundation ( united states ) , the national research council ( canada ) , conicyt ( chile ) , the australian research council ( australia ) , ministrio da cincia , tecnologia e inovao ( brazil ) and ministerio de ciencia , tecnologa e innovacin productiva ( argentina ) .
partial support for this work was provided by nasa through hubble fellowship grant hst - hf-51296.01-a awarded by the space telescope science institute , which is operated by the association of universities for research in astronomy , inc .
, for nasa , under contract nas 5 - 26555 .
ll @xmath100 & 11:28:48.227 + @xmath101 & + 17:04:18.42 + @xmath102 & 11:28:48.168 + @xmath103 & + 17:04:18.06 + @xmath8 & @xmath104 + separation@xmath105 ( @xmath63 ) & 0.85 w , 0.36 s + p(@xmath106 ) & 0.00064@xmath107 + @xmath108&@xmath109 + @xmath110&@xmath111 + @xmath80&@xmath112 + @xmath113&@xmath114 + @xmath115&@xmath116 + @xmath117&@xmath118 + @xmath119&@xmath120 + @xmath121&@xmath122 + @xmath123&@xmath124 + @xmath125&@xmath126 + @xmath127 & -20.96 @xmath128 + sfr@xmath129 & @xmath5 +
12 + log(o / h ) & 8.7 @xmath130 + @xmath131 & @xmath132 m@xmath11 + @xmath133 & 1.3 mag + , m. , grevesse , n. , & sauval , a. j. 2005 , in astronomical society of the pacific conference series , vol .
336 , cosmic abundances as records of stellar evolution and nucleosynthesis , ed .
t. g. barnes , iii & f. n. bash , 25 , s. m. et al .
2003 , in society of photo - optical instrumentation engineers ( spie ) conference series , vol . 4841 , society of photo - optical instrumentation engineers ( spie ) conference series , ed .
m. iye & a. f. m. moorwood , 16571669 | we present early optical photometry and spectroscopy of the afterglow and host galaxy of the bright short - duration gamma - ray burst discovered by the satellite . using our target of opportunity program on the gemini south telescope , our prompt optical spectra reveal a strong trace from the afterglow superimposed on continuum and emission lines from the @xmath0 host galaxy .
the combination of a relatively bright optical afterglow ( @xmath1 at @xmath2hr ) , together with an observed offset of 09 from the host nucleus ( 4.8kpc projected distance at @xmath3 ) , allow us to extract a relatively clean spectrum dominated by afterglow light .
furthermore , the spatially resolved spectrum allows us to constrain the properties of the explosion site directly , and compare these with the host galaxy nucleus , as well as other short - duration grb host galaxies .
we find that while the host is a relatively luminous ( @xmath4 ) , star - forming ( sfr = @xmath5 ) galaxy with almost solar metallicity , the spectrum of the afterglow exhibits weak absorption features but negligible emission features .
the explosion site therefore lacks evidence of recent star formation , consistent with the relatively long delay time distribution expected in a compact binary merger scenario .
the star formation rate ( both in an absolute sense and normalized to the luminosity ) and metallicity of the host are both consistent with the known sample of short - duration grb hosts and with recent results which suggest grb130603b emission to be the product of the decay of radioactive species produced during the merging process of a ns - ns binary ( `` kilonova '' ) .
ultimately , the discovery of more events similar to and their rapid follow - up from 8-m class telescopes will open new opportunities for our understanding of the final stages of compact - objects binary systems and provide crucial information ( redshift , metallicity and chemical content of their explosion site ) to characterize the environment of one of the most promising gravitational wave sources . |
as galaxies evolve , they often go through periods when their star formation rates increase substantially .
this is usually ascribed to large - scale perturbations in their structure , such as the formation of bars triggered by interactions with companions , leading to the accumulation of large quantities of gas in their nuclei .
a small number of those so - called ` starburst ' galaxies ( e.g. * ? ? ?
* ) have been identified in the local universe and have been studied extensively at all wavelengths . at high redshifts ,
obscured starburst galaxies are much more common @xcite , and are believed to give rise to the bulk of the cosmic infrared background radiation .
reviews of their properties can be found in @xcite and @xcite .
understanding the physics of starburst galaxies in the local universe is of great importance , as they are the closest laboratories we have for studying the starburst phenomenon , and because they may represent ` scaled - down ' versions of the starbursts seen at high redshift .
the obscured nature of ( most ) starburst galaxies , regardless of their redshifts , however means that probing their nuclear regions in the optical is challenging .
this task is better done in the infrared , where dust opacity is lower . the advent of major infrared space facilities such as _ iras _ @xcite , _ iso _ @xcite and _ spitzer _
@xcite , and in particular the accessibility to high quality spectroscopy , has allowed great progress in this field , including detailed studies of the mid - infrared ( mir ) region of the spectrum in galaxies .
it was discovered that the mid - infrared ( mir ) is rich in emission and absorption features of gas and dust .
emission from polycyclic aromatic hydrocarbons ( pahs , * ? ? ?
* ) was shown to be a reliable star formation tracer ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) , though variations with metallicity have also been observed @xcite .
the use of infrared atomic lines with different degrees of ionization has proven invaluable in studying the properties of the ionization field in starbursts ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) . in ultra - luminous infrared galaxies ( ulirgs ) , diagnostic diagrams based on line ratios such as [ nev@xmath2m]/[neii@xmath3 m ] , [ oiv@xmath4m]/[neii@xmath3 m ] ,
[ siii@xmath5m]/[siii@xmath6 m ] and combinations of those with pah emission features have been developed to reveal their nature ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
additional diagnostics using the silicate absorption features @xcite or the strength of the rotational molecular hydrogen lines ( e.g. * ? ? ?
* ; * ? ? ?
* ) have also been proposed .
llrrrrrrcc ngc253 & sb & 00@[email protected]@xmath9 & -25@[email protected]@xmath9 & 0.0008 & 2.5 & 10.23 & 9072640 & 12 @xmath11 6 & 12 @xmath11 6 + ngc520 & sb & 01@[email protected]@xmath9 & + 03@[email protected]@xmath9 & 0.0076 & 30.2 & 10.91 & 9073408 & 4 @xmath11 30 & 3 @xmath11 60 + ngc660 & sb+agn & 01@[email protected]@xmath9 & + 13@[email protected]@xmath9 & 0.00283 & 12.3 & 10.49 & 9070848 & 4 @xmath11 30 & 8 @xmath11 14 + ngc1097 & sb+agn & 02@[email protected]@xmath9 & -30@[email protected]@xmath9 & 0.00456 & 16.8 & 10.71 & 3758080 & 4 @xmath11 30 & 2 @xmath11 60 + ngc1222 & sb & 03@[email protected]@xmath9 & -02@[email protected]@xmath9 & 0.00818 & 32.3 & 10.60 & 9071872 & 4 @xmath11 30 & 2 @xmath11
60 + ngc1365 & sb+agn & 03@[email protected]@xmath9 & -36@[email protected]@xmath9 & 0.00545 & 17.9 & 11.00 & 8767232 & 4 @xmath11 30 & 8 @xmath11 14 + ic342 & sb & 03@[email protected]@xmath9 & + 68@[email protected]@xmath9 & 0.00010 & 4.6 & 10.17 & 9072128 & 4 @xmath11 30 & 8 @xmath11 14 + ngc1614 & sb & 04@[email protected]@xmath9 & -08@[email protected]@xmath9 & 0.016 & 62.6 & 11.60 & 3757056 & 4 @xmath11 30 & 8 @xmath11 14 + ngc2146 & sb & 06@[email protected]@xmath9 & + 78@[email protected]@xmath9 & 0.003 & 16.5 & 11.07 & 9074432 & 1 @xmath11 6 & 1 @xmath11 60 + ngc2623 & sb+agn & 08@[email protected]@xmath9 & + 25@[email protected]@xmath9 & 0.01846 & 77.4 & 11.54 & 9072896 & 4 @xmath11 30 & 2 @xmath11 60 + ngc3079 & sb+agn & 10@[email protected]@xmath9 & + 55@[email protected]@xmath9 & 0.00372 & 17.6 & 10.52 & 3755520 & 4 @xmath11 30 & 2 @xmath11 60 + ngc3256 & sb & 10@[email protected]@xmath9 & -43@[email protected]@xmath9 & 0.00913 & 35.4 & 11.56 & 9073920 & 1 @xmath11 6 & 1 @xmath11 60 + ngc3310 & sb & 10@[email protected]@xmath9 & + 53@[email protected]@xmath9 & 0.0033 & 19.8 & 10.61 & 9071616 & 4 @xmath11 30 & 2 @xmath11 60 + ngc3556 & sb & 11@[email protected]@xmath9 & + 55@[email protected]@xmath9 & 0.00233 & 13.9 & 10.37 & 9070592 & 4 @xmath11 30 & 2 @xmath11 60 + ngc3628 & sb+agn & 11@[email protected]@xmath9 & + 13@[email protected]@xmath9 & 0.00281 & 10.0 & 10.25 & 9070080 & 4 @xmath11 30 & 2 @xmath11 60 + ngc4088 & sb & 12@[email protected]@xmath9 & + 50@[email protected]@xmath9 & 0.00252 & 13.4 & 10.25 & 9070336 & 4 @xmath11 30 & 2 @xmath11 60 + ngc4194 & sb+agn & 12@[email protected]@xmath9 & + 54@[email protected]@xmath9 & 0.00835 & 40.3 & 11.06 & 3757824 & 4 @xmath11 30 & 2 @xmath11 60 + mrk52 & sb & 12@[email protected]@xmath9 & + 00@[email protected]@xmath9 & 0.0071 & 30.1 & 10.14 & 3753216 & 4 @xmath11 30 & 2 @xmath11 60 + ngc4676 & sb & 12@[email protected]@xmath9 & + 30@[email protected]@xmath9 & 0.022 & 94.0 & 10.88 & 9073152 & 8 @xmath11 30 & 6 @xmath11 60 + ngc4818 & sb & 12@[email protected]@xmath9 & -08@[email protected]@xmath9 & 0.00355 & 9.4 & 09.75 & 9071104 & 4 @xmath11 30 & 2 @xmath11 60 + ngc4945 & sb+agn & 13@[email protected]@xmath9 & -49@[email protected]@xmath9 & 0.00186 & 3.9 & 10.48 & 8769280 & 1 @xmath11 6 & 1 @xmath11 60 + mrk266 & sb+agn & 13@[email protected]@xmath9 & + 48@[email protected]@xmath9 & 0.02786 & 115.8 & 11.49 & 3755264 & 4 @xmath11 30 & 2 @xmath11 60 + ngc7252 & sb & 22@[email protected]@xmath9 & -24@[email protected]@xmath9 & 0.0156 & 66.4 & 10.75 & 9074688 & 4 @xmath11 30 & 3 @xmath11 60 + ngc7714 & sb & 23@[email protected]@xmath9 & + 02@[email protected]@xmath9 & 0.00933 & 38.2 & 10.72 & 3756800 & 4 @xmath11 30 & 2 @xmath11 60 + as part of our continuing efforts to provide a comprehensive picture of the starburst phenomenon at low redshifts , we here present an atlas of high resolution ( r@xmath0600 ) 9.7 - 37@xmath1 m spectra of well known nearby starburst galaxies , taken with the infrared spectrograph @xcite on - board spitzer
. this follows on from our earlier work on the same sample using the low resolution irs modules @xcite .
we have created a high s / n template from these spectra .
our goal is to make this set of high quality spectra available to the community , and to report on their characteristic features .
observations are described in [ obs ] , and data reduction and analysis is described in [ danalysis ] .
we present results in [ results ] and discussion in [ discussion ] . finally , we summarize our conclusions in [ conclusion ] .
[ allspectra ]
our sample is presented in table 1 where information on the galaxy optical classifications , distances , magnitudes and other basic information on the observations is given . in nine objects
there is evidence of an active galactic nucleus ( agn ) , even though it does not dominate the bolometric luminosity of the system .
the sample was observed as part of the irs guaranteed time program ( pid 14 ) and consists of 24 well known nearby starburst galaxies , with an average distance of 33 mpc and an average infrared ( ir ) ( 8 - 1000@xmath1 m ) luminosity of 5@xmath12 l@xmath13 .
the observations were made using both the high- and low - resolution modules of the irs .
here we present the high - resolution spectra .
the low - resolution spectra can be found in the companion paper by @xcite .
the two high - resolution modules , short - high ( sh ) and long - high ( lh ) , cover the wavelength region between 10 to 37 @xmath1 m at a resolution of 600 and have slit sizes of [email protected] and [email protected] respectively .
the data were taken using the staring mode observing template which produces spectra in two _ nod _ positions .
the separation between the _ nod _ positions is 3.8 for sh and 7.4 for lh .
the observations were centered on the nucleus of the galaxies ( see fig .
1 ) at the coordinates given in table 1 .
however , as can be seen in figure 1 , the observations of ngc520 and mrk52 are slightly mispointed with the sh slit measuring part of the nucleus . in ngc3310
the mispointing is more severe and affects both modules and therefore any information we give about this galaxy in the paper refers to the position we measured and not the nucleus .
mrk266 is also mispointed but the lh module is large enough to contain the whole galaxy and for the sh module we made use of an observation of another spitzer program in which the source is well centered ( pid@xmath143237 , aorkey@xmath1410510592 ) .
the area covered by the sh slit is equivalent to a spatial scale of about 0.085 kpc to 3.2 kpc , depending on the distance to the objects in our sample .
no dedicated background observations were made .
this means that there will be a small contribution of the zodiacal light to the spectra in figures 2 , 3 and 4 .
this has no consequence for the line measurements .
foreground emission of pahs from our mw is also possible but this is much fainter than the targets themselves . for information of the mid - infrared continuum properties the reader
is referred to the paper on the low resolution spectra @xcite where the background contribution to the continuum has been subtracted from the spectra .
the data were processed using version s15.3 of the _ spitzer _ science center s pipeline which is maintained at cornell and using a script version of _ smart _ @xcite .
the reduction begins from the _ droop _ images which are equivalent to the more commonly used _
bcd _ data but without flatfield and stray - cross - light corrections .
unstable sensitive pixels ( usually referred to as ` rogue ' pixels ) were replaced using the ssc tool _ irsclean _ , which uses campaign - based pixel masks to identify the bad pixels .
at the location of a pixel identified as bad , the empirical profile in the cross dispersion direction is determined using the average profile in the neighboring rows .
this profile is then scaled to the profile of the row containing the bad pixel using the remaining good pixels of that row .
the rogue pixels are then replaced by the value given by the empirical profile .
this method of replacing bad pixels is equivalent to interpolation when the bad pixel is located on the wings of the spatial profile but it has a definite advantage over interpolation when the bad pixel is located at the maximum of the spatial profile .
l c c c c c c c c c c c!h ngc253 & @xmath1510.50 & 64.32 & @xmath1527.45 & 2832.33 & @xmath1520.50 & 25.57 & 204.64 & 67.24 & 22.89 & 25.24 & 666.37 + & .... & 6.78 & .... & 64.20 & .... & 5.92 & 9.59 & 2.35 & 6.01 & 5.97 & 14.93 + ngc520 & @xmath150.47 & 3.49 & @xmath150.65 & 44.62 & @xmath150.64 & 0.89 & 7.53 & 10.82 & @xmath150.51 & @xmath150.48 & 9.10 + & .... & 0.22 & .... & 1.25 & .... & 0.15 & 0.21 & 0.35 & .... & .... & 0.59 + ngc660 & 2.79 & 12.53 & @xmath153.68 & 353.01 & @xmath152.26 & 4.05 & 36.96 & 23.52 & 4.89 & 2.52 & 90.02 + & 0.47 & 1.01 & .... & 7.55 & .... & 0.72 & 0.49 & 0.90 & 1.00 & 0.36 & 1.71 + ngc1097 & @xmath150.63 & 6.38 & 0.57 & 37.24 & @xmath150.43 & @xmath150.47 & 6.33 & 13.22 & @xmath150.99 & @xmath150.51 & 15.88 + & .... & 0.50 & 0.11 & 9.27 & .... & .... & 0.18 & 0.36 & .... & .... & 0.40 + ngc1222 & 22.24 & 2.93 & 2.12 & 80.57 & @xmath150.57 & @xmath150.88 & 89.41 & 9.53 & 1.49 & 0.83 & 64.80 + & 0.45 & 0.16 & 0.20 & 1.30 & .... & .... & 1.74 & 0.31 & 0.21 & 0.26 & 1.13 + ngc1365 & 23.37 & 9.00 & 3.18 & 139.50 & 18.83 & 2.05 & 59.53 & 18.42 & 2.22 & 1.72 & 53.72 + & 1.42 & 0.91 & 0.39 & 3.27 & 0.51 & 0.37 & 0.92 & 0.63 & 0.38 & 0.38 & 3.58 + ic342 & 4.76 & 10.32 & 6.22 & 615.46 & @xmath152.41 & 7.99 & 37.20 & 11.69 & 8.27 & 6.55 & 320.03 + & 0.66 & 0.96 & 0.79 & 10.52 & .... & 0.77 & 0.90 & 0.76 & 0.79 & 0.84 & 5.96 + ngc1614 & 6.89 & 5.14 & 1.77 & 249.00 & @xmath150.99 & 1.15 & 63.32 & 9.42 & 5.14 & 2.13 & 83.03 + & 0.54 & 0.51 & 0.37 & 7.00 & .... & 0.32 & 1.59 & 0.58 & 0.58 & 0.51 & 2.64 + ngc2146 & 6.30 & 12.79 & 6.81 & 625.00 & @xmath152.81 & 5.48 & 91.16 & 26.12 & 7.71 & 3.24 & 190.12 + & 0.44 & 1.35 & 1.53 & 15.35 & .... & 0.66 & 0.95 & 1.64 & 0.42 & 0.63 & 4.15 + ngc2623 & 1.20 & 3.15 & @xmath150.49 & 55.55 & 2.71 & 0.69 & 15.08 & 7.27 & @xmath150.52 & @xmath150.52 & 8.20 + & 0.15 & 0.19 & .... & 0.66 & 0.15 & 0.13 & 0.24 & 0.23 & .... & .... & 0.36 + ngc3079 & @xmath150.96 & 23.81 & @xmath152.09 & 98.96 & 1.31 & 1.57 & 23.17 & 40.63 & @xmath150.52 & 0.90 & 11.83 + & .... & 1.28 & .... & 2.03 & 0.21 & 0.21 & 0.46 & 1.04 & .... & 0.13 & 0.16 + ngc3256 & 5.25 & 15.04 & 6.19 & 514.19 & @xmath152.35 & 5.99 & 64.42 & 27.99 & 6.29 & 3.69 & 171.83 + & 0.57 & 1.07 & 1.38 & 9.34 & .... & 0.85 & 0.86 & 2.84 & 0.73 & 0.53 & 2.07 + ngc3310 & 4.46 & @xmath150.56 & 0.58 & 27.57 & @xmath150.25 & 0.29 & 28.35 & 1.88 & @xmath150.61 & @xmath150.59 & 20.50 + & 0.41 & .... & 0.13 & 0.61 & .... & 0.10 & 0.45 & 0.11 & .... & .... & 0.31 + ngc3556 & @xmath150.56 & 1.06 & 0.71 & 21.47 & @xmath150.37 & @xmath150.40 & 3.23 & 2.90 & @xmath150.45 & @xmath150.34 & 12.36 + & .... & 0.16 & 0.18 & 0.42 & .... & .... & 0.09 & 0.12 & .... & .... & 0.22 + ngc3628 & @xmath150.90 & 6.53 & @xmath151.68 & 125.63 & 0.90 & 2.35 & 10.05 & 16.24 & @xmath150.79 & @xmath150.96 & 21.32 + & .... & 0.43 & .... & 2.37 & 0.35 & 0.18 & 0.29 & 0.60 & .... & .... & 0.77 + ngc4088 & @xmath150.67 & 2.70 & 0.95 & 37.03 & @xmath150.39 & 0.51 & 2.45 & 4.72 & @xmath150.40 & @xmath150.37 & 12.50 + & .... & 0.32 & 0.11 & 1.16 & .... & 0.16 & 0.14 & 0.16 & .... & .... & 0.26 + ngc4194 & 11.46 & 5.03 & 2.12 & 165.47 & 2.96 & 2.49 & 53.99 & 9.25 & 1.70 & 1.90 & 70.92 + & 0.75 & 0.51 & 0.58 & 3.90 & 0.65 & 0.38 & 0.78 & 0.66 & 0.32 & 0.38 & 1.44 + mrk52 & 1.02 & 1.50 & @xmath150.61 & 29.38 & @xmath150.57 & @xmath150.43 & 3.82 & 1.80 & @xmath150.45 & @xmath150.39 & 28.45 + & 0.10 & 0.22 & .... & 0.53 & .... & .... & 0.29 & 0.19 & .... & .... & 1.63 + ngc4676 & 0.56 & 2.03 & 0.48 & 27.78 & @xmath150.25 & @xmath150.31 & 4.68 & 5.19 & @xmath150.36 & @xmath150.36 & 11.32 + & 0.06 & 0.14 & 0.09 & 0.96 & .... & .... & 0.16 & 0.32 & .... & .... & 0.37 + ngc4818 & 2.21 & 5.87 & 1.71 & 184.93 & @xmath151.21 & 2.27 & 13.73 & 16.83 & 2.61 & 1.41 & 71.96 + & 0.16 & 0.35 & 0.50 & 5.60 & .... & 0.35 & 0.45 & 0.57 & 0.35 & 0.18 & 1.05 + ngc4945 & 0.72 & 33.93 & @xmath153.66 & 583.84 & 3.84 & 6.51 & 68.98 & 54.15 & 1.77 & 2.85 & 30.53 + & 0.18 & 0.85 & .... & 12.50 & 0.34 & 0.27 & 14.25 & 4.22 & 0.25 & 0.24 & 0.58 + mrk266 & 9.00 & 3.72 & 0.69 & 57.04 & 7.96 & 0.46 & 27.95 & 8.40 & 0.67 & 0.48 & 24.33 + & 0.38 & 0.32 & 0.06 & 1.64 & 0.17 & 0.07 & 0.76 & 0.97 & 0.15 & 0.14 & 0.69 + ngc7252 & @xmath150.56 & 2.70 & @xmath150.57 & 41.68 & @xmath150.33 & 0.56 & 3.72 & 5.10 & @xmath150.54 & @xmath150.57 & 11.96 + & .... & 0.29 & .... & 1.94 & .... & 0.10 & 0.14 & 0.18 & .... & .... & 0.31 + ngc7714 & 14.76 & 2.85 & 2.18 & 102.55 & @xmath151.00 & @xmath150.68 & 77.42 & 4.43 & 2.01 & 1.58 & 81.50 + & 0.63 & 0.41 & 0.36 & 1.98 & .... & .... & 0.92 & 0.35 & 0.25 & 0.27 & 1.10 + template & 14.16 & 13.86 & 5.56 & 452.83 & .... & 4.58 & 100.00 & 30.83 & 5.74 & 3.30 & 187.97 + for each module the individual observations were averaged together , and spectra extracted using full aperture extraction .
flux calibration was performed by dividing the averaged spectra by the spectrum of the calibration star @xmath16 dra ( extracted in the same way as the target ) and multiplying by its theoretical template ( * ? ? ?
* sloan et al . in prep ) .
fringes , when present , were removed using the irs de - fringing tool _
irsfringe_. finally , the spectra for each _ nod _ position were combined .
it is worth mentioning that in six of the starbursts ( ic 342 , ngc520 , ngc1097 , ngc1614 , ngc3310 , and ngc3628 ) the continua differed between _ nod _ positions .
these sources are extended ( see figure 1 ) , and this difference may reflect the different regions the _ nod _ positions are sampling .
since we are interested in the overall properties of the emission lines and pah bands , and differences in these quantities lie within the errors in each nod position , we combined the nod positions in these starbursts as it was done for the rest of the sample .
l c c c c c c c c ngc253 & 120.84 & @xmath1573.36 & 154.74 & 250.56 & @xmath1577.55 & 1538.03 & 2412.03 & -243.86 + & 18.17 & .... & 26.95 & 24.43 & .... & 30.06 & 48.02 & .... + ngc520 & 5.95 & @xmath151.42 & 8.10 & 7.56 & 8.23 & 89.44 & 190.71 & @xmath156.90 + & 0.67 & .... & 0.88 & 0.78 & 0.60 & 8.60 & 3.41 & .... + ngc660 & 12.33 & @xmath156.59 & 18.80 & 22.92 & 10.58 & 246.07 & 441.96 & @xmath158.96 + & 1.64 & .... & 2.57 & 1.82 & 2.67 & 3.21 & 3.56 & .... + ngc1097 & 4.20 & @xmath151.07 & 4.41 & 11.48 & 6.59 & 100.05 & 225.19 & @xmath155.29 + & 0.18 & .... & 1.21 & 1.07 & 0.69 & 3.95 & 7.00 & .... + ngc1222 & 7.04 & @xmath150.78 & 9.92 & 4.74 & 3.25 & 132.52 & 112.05 & 12.72 + & 0.38 & .... & 0.48 & 0.38 & 0.34 & 2.30 & 1.70 & 1.30 + ngc1365 & 9.24 & 35.96 & 141.57 & 22.20 & 15.91 & 246.89 & 500.30 & @xmath1514.78 + & 1.20 & 1.18 & 2.06 & 2.33 & 1.51 & 12.17 & 17.78 & .... + ic342 & 36.43 & @xmath154.95 & @xmath157.70 & 51.87 & @xmath159.10 & 672.46 & 985.73 & @xmath1529.96 + & 2.69 & .... & .... & 1.89 & .... & 7.28 & 10.29 & ....
+ ngc1614 & 13.14 & @xmath153.27 & 8.68 & 12.99 & @xmath158.98 & 101.06 & 148.60 & 12.15 + & 2.10 & .... & 0.85 & 1.15 & .... & 2.11 & 4.11 & 3.44 + ngc2146 & 23.05 & @xmath153.70 & 19.33 & 52.69 & 16.66 & 848.02 & 1209.35 & @xmath1560.59 + & 2.58 & .... & 4.48 & 4.59 & 2.33 & 36.93 & 21.70 & ....
+ ngc2623 & @xmath151.31 & 2.21 & 9.62 & 2.23 & @xmath152.08 & 13.78 & 28.74 & @xmath159.41 + & .... & 0.14 & 0.72 & 0.51 & .... & 1.41 & 1.89 & ....
+ ngc3079 & 1.54 & @xmath150.91 & 8.45 & 14.57 & 11.02 & 56.88 & 173.51 & @xmath157.05 + & 0.21 & .... & 0.84 & 0.62 & 0.75 & 2.70 & 5.83 & ....
+ ngc3256 & 31.55 & @xmath153.85 & 12.23 & 33.56 & 12.70 & 484.64 & 623.37 & @xmath1533.15 + & 2.77 & .... & 3.02 & 1.82 & 1.32 & 13.79 & 9.81 & .... + ngc3310 & 5.66 & @xmath151.00 & 4.03 & 8.78 & @xmath152.32 & 106.02 & 143.54 & 8.01 + & 0.59 & .... & 0.64 & 0.74 & .... & 1.17 & 1.58 & 1.19 + ngc3556 & 2.24 & @xmath151.18 & @xmath151.50 & 3.14 & 4.55 & 96.37 & 96.44 & @xmath153.00 + & 0.34 & .... & .... & 0.47 & 0.47 & 1.18 & 0.80 & .... + ngc3628 & 3.69 & @xmath151.03 & @xmath152.37 & 11.96 & 11.57 & 149.93 & 277.55 & @xmath1510.09 + & 0.60 & .... & .... & 1.07 & 0.42 & 2.20 & 7.75 & .... + ngc4088 & 0.98 & @xmath150.50 & 0.73 & 2.15 & 3.38 & 35.02 & 63.33 & @xmath151.33 + & 0.09 & .... & 0.21 & 0.27 & 0.23 & 0.41 & 0.73 & .... + ngc4194 & 9.96 & 4.04 & 27.45 & 8.48 & @xmath152.66 & 151.01 & 185.82 & @xmath1510.83 + & 1.00 & 0.23 & 1.63 & 0.96 & .... & 1.65 & 3.31 & .... + mrk52 & 4.53 & @xmath150.83 & 1.31 & 1.90 & @xmath151.35 & 60.43 & 38.02 & @xmath152.44 + & 0.61 & .... & 0.23 & 0.31 & .... & 1.07 & 0.60 & ....
+ ngc4676 & @xmath150.58 & @xmath150.33 & 1.37 & 1.78 & 1.89 & 23.95 & 41.27 & @xmath152.49 + & .... & .... & 0.23 & 0.22 & 0.25 & 1.51 & 1.42 & ....
+ ngc4818 & 7.96 & @xmath151.96 & @xmath151.59 & 10.83 & 4.68 & 80.54 & 128.66 & @xmath155.61 + & 0.50 & .... & .... & 0.88 & 0.51 & 2.44 & 1.97 & .... + ngc4945 & 8.69 & @xmath157.95 & 47.85 & 55.21 & 44.56 & 338.91 & 846.23 & -160.25 + & 2.09 & .... & 6.95 & 5.35 & 8.72 & 70.93 & 50.98 & .... + mrk266 & 2.15 & 11.12 & 52.94 & 4.35 & @xmath151.60 & 50.93 & 87.02 & .... + & 0.29 & 0.59 & 5.66 & 0.70 & .... & 2.91 & 4.44 & .... + ngc7252 & 1.36 & @xmath150.51 & 1.33 & 3.12 & 1.10 & 24.75 & 70.53 & @xmath152.20 + & 0.26 & .... & 0.25 & 0.20 & 0.33 & 0.47 & 0.43 & .... + ngc7714 & 7.88 & @xmath151.77 & 5.51 & 5.93 & @xmath151.84 & 115.90 & 102.97 & 7.45 + & 0.70 & .... & 0.94 & 0.54 & .... & 2.45 & 1.63 & 1.29 + template & 11.25 & .... & 7.85 & 16.07 & 9.51 & 267.16 & 364.23 & 8.10 + lccccccccccc ngc253 & .... & 3.39 & 19.00 & 405.27 & 18.67 & 257.50 & 19.85 & 13.78 & 23.67 & .... & 15.86 + & .... & 0.35 & 1.24 & 20.26 & 2.30 & 12.88 & 2.16 & 3.30 & 4.98 & .... & 6.98 + ngc520 & 0.23 & 0.08 & 0.91 & 17.21 & 1.11 & 9.66 & 0.34 & 0.32 & 0.86 & 0.03 & 1.36 + & 0.02 & 0.02 & 0.05 & 0.86 & 0.06 & 0.48 & 0.06 & 0.02 & 0.13 & 0.01 & 0.07 + ngc660 & 0.52 & 0.77 & 6.07 & 83.40 & 4.60 & 43.45 & 1.05 & 2.09 & 5.64 & .... & 4.06 + & 0.21 & 0.10 & 0.30 & 4.17 & 0.44 & 2.17 & 0.26 & 0.22 & 0.32 & .... & 0.38 + ngc1097 & 0.30 & .... & 1.17 & 18.84 & 0.99 & 6.32 & 0.29 & 0.32 & 1.06 & .... & 0.99 + & 0.02 & .... & 0.06 & 0.94 & 0.05 & 0.32 & 0.07 & 0.02 & 0.23 & .... & 0.05 + ngc1222 & 0.66 & 0.26 & 2.72 & 27.56 & 1.09 & 9.69 & 0.38 & 0.27 & 1.54 & 0.33 & 1.53 + & 0.12 & 0.08 & 0.14 & 1.38 & 0.18 & 0.48 & 0.10 & 0.09 & 0.17 & 0.13 & 0.22 + ngc1365 & 0.70 & 0.51 & 3.64 & 51.38 & 2.45 & 20.04 & 1.14 & 0.65 & 2.64 & 0.31 & 3.20 + & 0.50 & 0.28 & 0.46 & 2.57 & 0.49 & 1.08 & 0.23 & 0.27 & 0.32 & 0.23 & 0.46 + ic342 & 1.41 & 0.64 & 9.70 & 119.19 & 6.38 & 47.77 & 3.04 & 3.26 & 6.21 & .... & 4.67 + & 0.16 & 0.11 & 0.49 & 5.96 & 0.32 & 2.39 & 0.15 & 0.16 & 0.45 & .... & 0.23 + ngc1614 & 1.18 & 0.12 & 3.82 & 47.29 & 2.46 & 22.80 & 0.30 & 1.98 & 3.35 & 0.34 & 3.93 + & 0.06 & 0.03 & 0.19 & 2.36 & 0.12 & 1.14 & 0.11 & 0.10 & 0.60 & 0.03 & 0.20 + ngc2146 & 1.38 & 1.36 & 14.80 & 168.24 & 7.03 & 84.87 & 1.58 & 4.69 & 9.94 & 1.19 & 12.16 + & 0.21 & 0.13 & 0.74 & 8.41 & 0.41 & 4.24 & 0.31 & 0.36 & 0.50 & 0.36 & 0.75 + ngc2623 & 0.29 & 0.15 & 0.91 & 12.61 & 0.63 & 7.37 & 0.43 & 0.40 & 1.13 & .... & 0.74 + & 0.04 & 0.02 & 0.05 & 0.63 & 0.06 & 0.37 & 0.03 & 0.03 & 0.06 & .... & 0.07 + ngc3079 & .... & .... & 3.22 & 56.68 & 3.13 & 35.39 & 1.58 & 1.59 & 5.07 & 0.24 & 3.41 + & .... & .... & 0.16 & 2.83 & 0.16 & 1.77 & 0.08 & 0.08 & 0.25 & 0.02 & 0.17 + ngc3256 & 1.31 & 0.86 & 8.65 & 108.19 & 5.41 & 50.18 & 1.50 & 1.69 & 7.77 & 0.99 & 6.27 + & 0.31 & 0.17 & 0.43 & 5.41 & 0.30 & 2.51 & 0.12 & 0.13 & 0.43 & 0.29 & 0.55 + ngc3310 & .... & .... & 1.21 & 14.79 & 0.48 & 5.03 & 0.44 & 0.31 & 0.78 & 0.08 & 0.74 + & .... & .... & 0.06 & 0.74 & 0.02 & 0.25 & 0.02 & 0.02 & 0.10 & 0.01 & 0.04 + ngc3556 & .... & .... & 0.84 & 8.46 & 0.46 & 3.17 & 0.31 & 0.29 & 0.44 & .... & 0.44 + & .... & .... & 0.04 & 0.42 & 0.02 & 0.16 & 0.05 & 0.01 & 0.11 & .... & 0.03 + ngc3628 & 0.02 & 0.23 & 1.81 & 29.98 & 1.48 & 19.88 & 0.01 & 0.63 & 2.04 & 0.19 & 2.38 + & 0.02 & 0.01 & 0.09 & 1.50 & 0.07 & 0.99 & 0.14 & 0.03 & 0.27 & 0.01 & 0.12 + ngc4088 & .... & .... & 1.21 & 12.64 & 0.66 & 5.09 & 0.17 & 0.22 & 0.93 & .... & 0.68 + & .... & .... & 0.06 & 0.63 & 0.03 & 0.25 & 0.01 & 0.02 & 0.09 & .... & 0.03 + ngc4194 & 0.88 & 0.27 & 4.84 & 49.18 & 1.55 & 20.15 & .... & 1.00 & 2.83 & 0.31 & 3.19 + & 0.14 & 0.09 & 0.24 & 2.46 & 0.21 & 1.01 & .... & 0.16 & 0.19 & 0.15 & 0.27 + mrk52 & .... & .... & 0.64 & 5.88 & 0.43 & 2.49 & .... & 0.17 & 0.35 & .... & 0.31 + & .... & .... & 0.07 & 0.29 & 0.08 & 0.12 & .... & 0.03 & 0.04 & .... & 0.05 + ngc4676 & 0.23 & .... & 0.79 & 9.41 & 0.35 & 4.18 & 0.09 & 0.43 & 0.68 & 0.05 & 0.68 + & 0.02 & .... & 0.04 & 0.47 & 0.03 & 0.21 & 0.02 & 0.02 & 0.03 & 0.02 & 0.06 + ngc4818 & 0.59 & 0.48 & 3.72 & 43.53 & 2.07 & 18.95 & 0.55 & 1.18 & 3.14 & 0.41 & 1.87 + & 0.11 & 0.03 & 0.19 & 2.18 & 0.12 & 0.95 & 0.09 & 0.06 & 0.16 & 0.10 & 0.17 + ngc4945 & .... & .... & 1.36 & 54.64 & 3.64 & 83.84 & 4.85 & 2.64 & 6.20 & 0.17 & 3.49 + & .... & .... & 0.13 & 2.73 & 0.57 & 4.19 & 0.24 & 0.13 & 0.31 & 0.06 & 0.17 + mrk266 & .... & .... & 0.27 & 3.71 & 0.27 & 1.07 & .... & 0.19 & 0.23 & .... & 0.34 + & .... & .... & 0.06 & 0.19 & 0.11 & 0.22 & .... & 0.06 & 0.06 & .... & 0.13 + ngc7252 & 0.14 & .... & 2.05 & 23.54 & 1.12 & 7.55 & .... & 0.50 & 1.31 & .... & 1.54 + & 0.04 & .... & 0.10 & 1.18 & 0.06 & 0.38 & .... & 0.05 & 0.07 & .... & 0.08 + ngc7714 & 0.52 & 0.27 & 3.18 & 28.21 & 1.53 & 11.34 & 0.22 & 0.50 & 1.83 & 0.21 & 1.81 + & 0.14 & 0.05 & 0.16 & 1.41 & 0.15 & 0.57 & 0.07 & 0.09 & 0.11 & 0.09 & 0.16 + template & 1.24 & 0.59 & 8.57 & 100.00 & 4.65 & 41.26 & 0.61 & 1.31 & 6.58 & 0.47 & 5.04 + fine - structure line fluxes are derived by fitting a gaussian profile and integrating the flux above a local continuum .
a polynomial of order one was used to fit the local continuum for a given line , except for the line which sits at the sharp flank of a pah feature ( see fig .
3 ) . for this line ,
a third order polynomial was used to fit the pah underneath the line .
this is done for each _ nod _ separately and then the flux in both measurements is averaged .
errors bars were estimated using the difference between the two _ nod _ positions or the standard deviation of the residuals of the fit , whichever is the larger .
upper - limits are derived by measuring the flux of a gaussian with a height three times the local _ rms _ , and with a _ fwhm _
equal to the instrumental resolution .
note that , except for the line measurements of the template , tables 2 and 3 quote the line fluxes without any normalization applied to correct for the difference in aperture between the sh and lh modules .
following @xcite;and @xcite , we have measured the strength of the pah emission features by integrating the flux of the pah bands above an adopted spline interpolated local continuum . for a given band
the anchor points used to define the continuum ( see table 4 ) are the same for all sources , as are the integration limits .
spectral lines superimposed on the pah bands were removed before integration .
our spline method for measuring pah fluxes differs from that of _ pahfit _ @xcite and others ( e.g. * ? ? ?
* ; * ? ? ?
* ) by adopting a continuum ( which ignores the plateau ) and integrating under the bands , rather than assuming a pah band profile and performing a simultaneous fit to all the bands .
@xcite concluded that both methods reveal similar trends in the data .
the uncertainties in the pah fluxes ( table 4 ) are based on the differences in the fluxes measured from the two individual _ nod _ spectra .
however , this does not take into account all the systematic uncertainties ( especially flux calibration ) .
we estimate these to be 5% and choose it as our minimum uncertainty .
[ composite ] [ compositezoom ] .
the spectra are shown in figure 2 .
the files used to generate this plot are available in electronic format ( tab1-24.txt files ) . measurements of fine - structure lines and pah features that fall within the wavelength range of the irs sh and lh modules are given in tables [ lines_hr_s ] , [ lines_hr_t ] , and 4 respectively .
the spectra contain a variety of emission features .
all of the spectra contain the prominent fine - structure lines of ( 12.81@xmath1 m ) , ( 15.55@xmath1 m ) , and ( 18.71 and 33.48@xmath1 m ) .
higher ionization lines such as ( 25.98@xmath1 m ) and ( 10.51@xmath1 m ) are seen in most of the sample , but are weak .
the line at 14.32@xmath1 m is detected in seven sources .
the line at 13.1@xmath1 m is only detected in mrk266 .
in addition , many objects show emission of ( 25.99@xmath1 m ) and ( 22.92@xmath1 m ) .
the fine - structure lines in the irs spectra are unresolved ( even those with agn ) , except for mrk266 and ngc3079 where there is evidence of broadening in the and lines ( spoon et al . in prep . ) .
the recombination line hi ( 6 - 7 ) at 12.37@xmath1 m is detected in 14 objects .
rotational lines of h@xmath17 , with the strongest transition 0 - 0 s(1 ) emitting at 17.03@xmath1 m , are present in most of the spectra .
a number of pah features are seen , the most prominent being the neutral c - h out - of - plane bending bands at 11.2@xmath1 m and the 12.7@xmath1 m . the 16.4@xmath1 m feature and 17@xmath1 m complex
are also detected throughout the sample .
the high s / n allow us to detect in the spectra other weaker pah features and a few ` unidentified ' bands .
these are best seen in figures 3 and 4 , and are discussed in the next section .
we have constructed a high - resolution starburst template spectrum from the 15 spectra in our sample with no signs of agn activity ( see table 1 ) .
the template is available in electronic format ( tab25.txt ) .
this is equivalent to the starburst template derived by @xcite but using the irs high - resolution modules . to create this template ,
the sh spectrum of each object was first scaled up to the lh spectrum using the overlapping continuum around 19@xmath1 m . based on a spline - defined continuum ( with anchor points given in caption of figure 4 )
each spectrum were separated into a continuum - subtracted spectrum ( feature spectrum ) and the associated continuum spectrum ( featureless spectrum ) .
the continuum template was generated by averaging every individual continuum spectrum after normalization to the 14@xmath1 m flux .
likewise , the feature template spectrum was created by computing the noise - weighted mean of every individual continuum - subtracted spectrum , after normalization to the total flux of the emission features .
finally , the starburst template spectrum was assembled by scaling the continuum spectrum to the average ratio between the sum of the pahs and line fluxes over the 14@xmath1 m continuum before adding it to the feature template spectrum .
the template is shown in figure 3 .
up until now , the 2 - 40@xmath1 m _
iso_-sws spectrum of m82 has been usually adopted in the literature as a starburst high - resolution template .
however , m82 is a dwarf starburst with a low luminosity ( 3x10@xmath18 l@xmath19 ) while our template is built from a sample of 15 galaxies spanning a range in ir color and luminosity typical of starburst galaxies .
in fact , seven of them are lirgs ( @xmath2010@xmath21 l@xmath19 ) , which in terms of galaxy evolution are much more important as they play a major role in the total energy production of the universe at z@xmath01 @xcite .
the template we created may hence be more representative of the average physical parameters of local starburst galaxies , and the average line ratios and relative pah strengths may serve as comparison data for studies of starburst galaxies at higher redshifts .
figure 4 shows a zoom into several interesting parts of the template spectrum .
a _ shows that , in addition to the line at 10.51@xmath1 m , the spectrum contains the rarely detected 10.60@xmath1 m pah feature and another broad feature at 10.74@xmath1 m .
b _ shows details of the 11.3@xmath1 m pah feature including the 11.0@xmath1 m component .
@xcite noted that this feature ( 11.0@xmath1 m ) indicates ionized pahs based on spatially resolved spectra of the reflection nebula ngc1333 svs3 .
another pah feature at 12.0@xmath1 m is seen followed by the h@xmath22 0 - 0s(2 ) line and the hi hu@xmath23 recombination line at 12.37@xmath1 m .
the line ( 12.81@xmath1 m ) sits in the flank of the 12.7@xmath1 m pah feature . in panel _
c _ there is a pah feature at 13.53@xmath1 m followed by an unidentified absorption feature ( 13.65 - 14.0@xmath1 m ) which has , to our knowledge , not been previously detected .
this absorption is detected in most of the starbursts but the strongest absorptions are found in those sources also harboring an agn .
this is shown in figure 5 , where the objects have been ordered according to the strength of the absorption feature , and where out of the bottom seven objects ( those with stronger absorption ) , six have an agn .
its detection in ngc4945 is also confirmed from a map study in the region where the dip is seen to change in strength at different positions in the central region of the galaxy ( spoon et al .
, in prep ) . while not reported , this dip is also present in the low - resolution starburst template ( made from the same sample ) in @xcite . the weak 14.22@xmath1 m pah feature is present and the line at 14.36@xmath1 m corresponds to .
d _ displays more pah features at 16.4@xmath1 m , 16.7@xmath1 m , and 17.0@xmath1 m .
the h@xmath17 0 - 0 s(1 ) transition at 17.0 @xmath1 m is strong and the line at 17.98@xmath1 m , while weak , is still detected . another iron line ( at 22.93@xmath1 m ) is shown in panel _
e_. the last panel shows a blend of lines which correspond to ( 25.89@xmath1 m ) and ( 25.98@xmath1 m ) , the stronger of which is the iron line .
[ weirdabs ] m. the sources have been ordered by decreasing strength of the feature .
the vertical lines indicate the position of the and fine - structure lines.,width=302 ]
the presence of individual lines can set constraints on the ionization conditions in the hot gas in our sample .
the most notable of these is .
the has a high ionization potential ( ip ) of 97.1 ev and has never been detected in purely stellar environments , but is routinely detected in spectra of galaxies hosting an agn @xcite and in about half of the 53 ulirgs analyzed by @xcite using _ spitzer _ data .
this line is also seen in young sne remnants @xcite but in starbursts , where the overall shape of the mid - ir continuum is not flat , the presence of [ nev ] would imply a weak agn rather than sne emission ( the lifetime of the latter is too small to allow for a substantial contribution to the spectrum ) . from the objects listed in table 2 previously known to have an agn component in the literature , the line is present in mrk266 , ngc1365 , ngc2623 , ngc3079 , ngc3628 , and ngc4945 ( see also * ? ? ? * ) , but not in ngc660 and ngc1097 .
in these two galaxies the 14.3@xmath1m/12.8@xmath1 m line ratio is @xmath150.07 hinting at an agn contribution of less than a few percent using the diagram by @xcite ( their fig .
ngc1097 is a particularly interesting case ; it also does not show or , hence the agn is likely very weak in the mid - ir .
it either has a complex extinction pattern around it or it is overwhelmed by the starburst .
we further detect in ngc4194 , which has no previous agn classification .
m/18.71@xmath1 m ) and neon ( 15.55@xmath1m/12.81@xmath1 m ) line ratios.,width=245 ] /[neii])=-0.38(\pm0.12)+0.68(\pm0.08)\times\log([siv]/[neii])$].,width=245 ] m ) .
the dashed line indicates the best fit to the data : @xmath24=-5.3(\pm2.1)+1.18(\pm0.20)\times\log[l(ir)/l_\odot]$].,width=245 ] the strongest detections of relative to in our sample are found in mrk266 and ngc1365 , at ratios of [email protected] and [email protected] , roughly a factor 10 above the rest of the sample . adopting the agn - starburst mixing scenarios of sturm et al .
( 2002 ) and farrah et al .
( 2007 ) , this would imply the agn contribution to be 10 times larger in these two sources compared to the rest of the sample .
two other individual lines are worth mentioning ; the line , and the 18@xmath1 m and 26@xmath1 m lines .
the line is present in most of the objects but is always weak .
the line has an ip of 34.8 ev , below that of the line , and therefore should be easily excited .
however , the photo - ionization cross - section of s@xmath26 is lower than ne@xmath27 , and one needs many photons to produce the line .
the ip of is even lower , at 7.87 ev .
it could therefore ionize within the photo - dissociation region ( pdr ) . @xcite
have shown that this ion correlates well with the line which is a good tracer of the ionized gas near the pdr . !
ht 10@xmath28 , which is significantly higher that the range of values in the starburst sample ( top panel ) and is therefore not plotted.,width=321 ] !
t l c c c c c c c c c c!h ngc1097 & 3.22 & @xmath150.22 & 0.84 & 0.68 & 1.18 & 0.56 & 1.09 & 3.54 & 0.91 & 1.90 + & 0.79 & 0.00 & 0.14 & 0.11 & 0.19 & 0.16 & & 0.79 & 0.15 & 0.35 + ngc1222 & 3.53 & 2.09 & 0.49 & 2.58 & 0.40 & 0.78 & 1.58 & 4.09 & 0.75 & 1.87 + & 0.87 & 0.37 & 0.08 & 0.43 & 0.07 & 0.22 & & 0.91 & 0.12 & 0.29 + ngc1365 & 1.96 & 1.47 & 0.56 & 1.15 & 0.56 & 0.31 & 1.17 & 2.30 & 0.68 & 1.02 + & 0.48 & 0.25 & 0.10 & 0.19 & 0.09 & 0.08 & & 0.50 & 0.12 & 0.17 + ic342 & 5.95 & 0.15 & 1.27 & 0.37 & 1.47 & 1.34 & 1.07 & 6.56 & 1.31 & 3.01 + & 1.46 & 0.03 & 0.21 & 0.06 & 0.24 & 0.38 & & 1.46 & 0.22 & 0.62 + ngc1614 & 5.42 & 0.78 & 1.80 & 2.19 & 2.23 & 2.94 & 1.20 & 6.04 & 2.02 & 6.21 + & 1.33 & 0.14 & 0.31 & 0.36 & 0.36 & 0.82 & & 1.34 & 0.35 & 1.18 + ngc2146 & 3.23 & 0.18 & 1.18 & 0.82 & 0.64 & 0.37 & 1.09 & 3.57 & 1.26 & 1.10 + & 0.79 & 0.03 & 0.20 & 0.14 & 0.10 & 0.10 & & 0.79 & 0.21 & 0.21 + ngc3256 & 3.21 & 0.17 & 1.07 & 0.64 & 0.71 & 0.87 & 1.13 & 3.55 & 1.13 & 1.78 + & 0.79 & 0.03 & 0.18 & 0.11 & 0.12 & 0.25 & & 0.79 & 0.19 & 0.37 + ngc3310 & 4.07 & 1.53 & 0.61 & 2.98 & 1.08 & 0.90 & 1.38 & 4.63 & 0.91 & 2.73 + & 1.00 & 0.27 & 0.10 & 0.50 & 0.18 & 0.25 & & 1.03 & 0.15 & 0.43 + ngc3556 & 2.01 & @xmath150.16 & 0.39 & 0.28 & 0.21 & 0.19 & 1.13 & 2.21 & 0.42 & 0.45 + & 0.50 & 0.00 & 0.07 & 0.05 & 0.04 & 0.06 & & 0.50 & 0.07 & 0.09 + ngc4088 & 1.53 & @xmath150.14 & 0.50 & 0.16 & 0.30 & 0.18 & 1.06 & 1.68 & 0.52 & 0.51 + & 0.37 & 0.00 & 0.09 & 0.03 & 0.05 & 0.05 & & 0.37 & 0.09 & 0.10 + ngc4194 & 3.86 & 1.08 & 1.00 & 1.56 & 0.69 & 1.06 & 1.25 & 4.35 & 1.16 & 2.19 + & 0.95 & 0.18 & 0.17 & 0.26 & 0.11 & 0.30 & & 0.97 & 0.20 & 0.41 + ngc4676 & 2.74 & 0.23 & 0.75 & 0.60 & 0.65 & 0.00 & 1.00 & 3.04 & 0.81 & 0.65 + & 0.68 & 0.04 & 0.13 & 0.10 & 0.11 & 0.00 & & 0.68 & 0.14 & 0.11 + ngc4818 & 4.87 & 0.26 & 1.39 & 0.49 & 2.09 & 1.99 & 1.08 & 5.38 & 1.44 & 4.43 + & 1.20 & 0.04 & 0.23 & 0.08 & 0.35 & 0.58 & & 1.21 & 0.24 & 0.92 + ngc7714 & 4.32 & 1.35 & 0.60 & 2.17 & 0.71 & 1.22 & 1.43 & 4.89 & 0.82 & 2.76 + & 1.06 & 0.24 & 0.10 & 0.37 & 0.12 & 0.34 & & 1.08 & 0.14 & 0.46 + solar & .... & .... & .... & .... & .... & .... & .... & 1.4 - 2.1 & 0.7 - 1.2 & 32.4 + fine structure lines of the same species but from different ionization stages can be used to trace the properties ( hardness and ionization parameter ) of the ionizing radiation field .
the diamond symbols in figure 6 shows the ratio of the 10.5@xmath1 m ( [email protected] ev ) and 18.7@xmath1 m ( [email protected] ev ) lines plotted against the ratio of the 15.5@xmath1 m ( [email protected] ev ) and 12.8@xmath1 m ( [email protected] ev ) lines for the starburst sample . as expected both ratios trace each other .
exceptions are mrk266 and ngc1365 .
these two outliers both have a strong agn contribution to the neon lines , as measured by their @xmath2910 times elevated / ratio compared to the other sources . excluding these two sources , the slope for the correlation between / and / is [email protected] .
our values values fall within the range of values found by @xcite in their _ iso _ study of starburst galaxies ( including ngc1365 ) .
figure 6 also includes representative subsamples of blue compact dwarf galaxies ( bcds * ? ? ?
* ) , ulirgs @xcite , and galaxies from the spitzer infrared nearby galaxy survey ( sings * ? ? ?
the ratios in the sings galaxies studied by @xcite and in several star forming regions in irregular galaxies analyzed by @xcite follow the trend of the starburst sample .
the ratios in the starburst galaxies are lower than the bcd sample studied by @xcite , and in fact figure 6 shows only the lowest bcd ratios from their sample .
@xcite compare in detail our starburst sample with similarly analyzed samples of ulirgs , bcds , and agns .
they find that the starburst galaxies have the lowest excitation and are located in the lower branch of this diagram . compared to the starburst and ulirg samples of @xcite and @xcite the excitation level in our sample is on the low and middle end , with the highest ratios being /@xmath00.4 and /@xmath01 , and the @xcite and @xcite samples reaching values as high as 10 for both ratios .
this lower ionization in our sample may indicate that there are fewer massive stars or that they are older .
interestingly , the three sources with the highest / ratios in our sample ( @xmath290.6 ) are not those with an agn component , but pure starbursts .
ngc1222 , ngc3310 and ngc7714 all are found on the ` bcd branch ' of the / versus / diagnostic diagram of @xcite .
@xcite studied the relation between the well known mid - infrared ionization diagnostic with the ground accessible ratio of /.
they found a good correlation between the ratios in their sample , which included starburst galaxies , bcds , ulirgs , agns , planetary nebulae , and galactic and extra - galactic hii regions .
we confirm this relation ( figure 7 ) , and find a slope value of [email protected] ( excluding ngc4945 ) , in good agreement with the slope found by @xcite ( 0.65 for the starburst sub - sample and 0.81 for their full sample ) .
ngc4945 has the strongest silicate absorption feature of the sample @xcite and this attenuates the line flux relative to the line which is exactly what is seen in figure 7 .
@xcite illustrated how the neon lines in the mid - infrared spectra can be a powerful indicator to derive the star formation rate in star - forming galaxies .
they showed that the sum of the luminosities of both lines correlate well with the infrared luminosity in over 5 orders of magnitude in luminosity , and proposed a calibration between the luminosity of the lines and the star formation rate .
figure 8 shows the relation between the neon lines and infrared luminosity .
our data covers a narrow range of infrared luminosities , but it can be seen that a correlation is present .
the slope in the figure has a value of [email protected] , agreeing within the errors with that reported by @xcite ( [email protected] ) .
infrared lines offer many advantages over optical or ultraviolet ( uv ) lines for measuring elemental abundances , the most important of which is their small dependence on the electron temperature ( e.g. * ? ? ?
* ; * ? ? ?
this reduces the uncertainty in the abundance estimates when there are electron temperature variations , which is the case in integrated spectra of galaxies , or when the electron temperature can not be measured .
ir lines are also less affected by extinction .
optical and uv studies also rely on the often rather uncertain , ionization correction factors ( icfs ) to account for unobserved stages of ionization .
the direct measure of important stages of ionization of some elements in the infrared greatly reduces , or completely avoids , the use of icfs .
the hi ( 6 - 7 ) line at 12.37@xmath1 m is detected in 14 of the stabursts .
this allows us to measure the abundances of neon , sulfur , and ( tentatively ) iron relative to hydrogen .
the abundances are derived adopting an electron temperature ( t@xmath30 ) of 10000 k and a gas density of 100 @xmath31 .
these are values typical of hii regions .
we derived abundances for a range of t@xmath30 ( 7500 - 12500 k ) , and used the differences to obtain the uncertainties in the measurements .
these uncertainties amount to around 20% in the total abundances . using these parameters and solving the equation of statistical equilibrium for the levels of the atoms ,
the ionic abundance can be derived ( see eq . 1 * ? ? ?
these values are reported in table [ lines_abun_t ] .
note that we can not derive the density from the pair of ( 18.7@xmath1m/33.4@xmath1 m ) or ( 15.5@xmath1m/36.0@xmath1 m ) lines because the dependence of these ratios on the density starts above 3000 @xmath31 for the sulfur ratio , and higher for the neon ratio .
compared to our reported values , an electron density of 1000 @xmath31 ( on the high side for hii regions ) will produce abundances that are no more than 1.7% higher for neon , 13% higher for sulfur , and about 80% lower for iron .
the latter is due to the electronic configuration of iron . !
h in starburst galaxies the most important stages of ionization of neon for abundance determination are ne@xmath32 and ne@xmath33 which are measured in the ir . thus , no icfs are needed . among the 14 objects with a hi detection ,
ne@xmath34 is present in ngc1365 and ngc4194 ( table [ lines_abun_t ] ) , but this likely originates from the agn given the hard radiation field needed to break the high ionization potential ( ip ) barrier of 97.1 ev to produce this ion .
this applies as well to the ne@xmath26 ( ip of 63.5 ev , also above the double ionized helium cutoff ) .
in addition , the contribution of ne@xmath34 to the total neon abundance is less than 1% .
the 14.3@xmath1m/12.81@xmath1 m ratio in ngc1365 and ngc4194 is 0.13 and 0.018 respectively .
these ratios correspond in the @xcite diagrams to an agn contribution of @xmath010% and @xmath151% respectively .
this implies that the abundances we derive for ngc4194 is indeed representative of the starburst while in ngc1365 our abundances may be @xmath010% too high .
the total sulfur abundance is derived adding the contribution from s@xmath33 and s@xmath26 .
no s@xmath32 is measured in the infrared but its contribution to the total sulfur abundance is usually small .
@xcite estimated a contribution of about 15% of s@xmath32 in the galactic ultra - compact hii regions they studied .
@xcite analyzed a sample of magellanic cloud hii regions using ir data , and combined optical measurements of lines .
they derived an ionic abundance of s@xmath32 between 2 to 20% of the total sulfur abundance , with @xmath08% being the average . using the photo - ionization code of @xcite
we have created a grid of models at typical conditions of hii regions ; stellar temperatures considered between 37500 to 55000 k , and electron density of 10 @xmath31 , and a filling factor ranging from 0.2 to 1.0 .
these models predict the contribution of s@xmath32 to be between 7 and 15% of that of s@xmath33 ( which is the dominant contributor to the total sulfur abundance ) . we have thus adopted a 10% contribution of s@xmath32 to the sulfur abundance ( [email protected] ) to account for the contribution of this ion in our sample . while we have measured fe@xmath32 and fe@xmath33 from our _ spitzer _ data , fe@xmath26 may be an important contributor ( up to half ) to the total iron .
its contribution has been estimated using the photo - ionization grid models by @xcite , and taking advantage of the similarity in ip between the / and / line ratios to scale the contribution of this ion in each object . at
can be seen from table 5 the icf varies from a few percent to a factor 2 . given the importance of this ion to the total iron budget in those objects with a high icf , the iron abundances we report should be interpreted with care .
the uncertainty for this element in table 5 only takes into account the uncertainty in the fe@xmath32 and fe@xmath33 ionic abundance . in figure 9 ( top ) , the neon abundance of our starburst sample correlates reasonably well with the sulfur abundance .
this is expected because they are both @xmath23-elements and are being produced in the same type of stars .
this finding was also pointed out by @xcite in their bcd sample and it contradicts earlier findings @xcite who based on _ iso _ data did not find such a correlation between the abundance of these two elements for the starbursts they studied . in our sample
the neon abundances range from slightly supersolar ( ic342 , ngc1614 , ngc4818 ) to several times lower than the solar value ( see table 5 ) .
note that the solar abundances of certain elements have been subject to many changes during the last decade @xcite , and thus , we resort to compare to the range of values found in the literature ( 0.69 - 1.2@xmath1110@xmath35 ) .
there is however increasing evidence that the higher solar neon abundance reported agrees better with the findings of studies in galactic hii regions and planetary nebulae .
the sulfur abundance is also lower than solar ( more so than neon , with solar sulfur values ranging from 1.4 - 2.1@xmath1110@xmath28 ) , which is in agreement with many previous studies .
available optical - derived abundances of oxygen by @xcite in ic342 , ngc3310 , and ngc7714 ( 7.1 , 1.5,1.8@xmath1110@xmath35 respectively ) confirm the trend of our ir - derived neon abundances , with ic342 supersolar , and the other two galaxies having a 2 - 3 times lower abundance than the sun ( adopting 4.9@xmath1110@xmath35 for the solar oxygen abundance ) .
the bottom panel in figure 9 displays the ne / s ratio versus the neon abundance . the range of solar values is indicated by the grey box in the figure . the ne / s ratio varies from 18 to 40 .
this range of values is typical for other infrared derived abundances in starburst galaxies , hii regions and even planetary nebulae @xcite .
it is overall higher than the ratio found by @xcite in their abundance analysis of bcds , where they find values from 10 - 20 .
the ne / s ratios are higher than the solar ne / s ratio , revealing a discrepancy between the neon and sulfur solar values , and which has now been widely reported ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* and references therein ) . in figure 10
the neon abundance is plotted versus the hardness of the radiation field .
the sample is small but seems to indicate that a hard radiation field is found only for sources with a low metallicity which ties well with the common knowledge that at low stellar metallicity , line blanketing is diminished resulting in harder stellar radiation fields . in figure 11 the pah luminosity for the different bands
is plotted .
we chose to compare the pah features in the spectra to the integrated emission of the [email protected]@xmath1 m complex because it is the strongest in this wavelength region ( panels ` a ' through ` j ' ) . panels e , h , and j show a very good correlation of the 11.3@xmath1 m complex with the relatively strong pah bands at 12.7 , 16.4 , and the 17.0@xmath1 m pah complex .
the 17.0@xmath1 m complex was discovered by @xcite as a weak feature in _ iso _ spectra and attributed to pahs by laboratory analysis @xcite .
@xcite corroborate this using spectra from the sings galaxies and showing good correlation between the strength of the 17.0@xmath1 m and the 11.3@xmath1 m pah complexes .
our sample supports this finding ( panel ` j ' ) .
other bands also correlate well ( e.g 11.0 , 12.0 , 16.7@xmath1 m ) .
the correlation of the weaker bands ( 10.6 , 13.5 , 14.2@xmath1 m ) is good but shows more scatter .
the 16.7@xmath1 m feature is listed by @xcite as part of the 17.0@xmath1 m complex but in our higher resolution spectra we can separate it from this complex . @xcite also report a pah feature at 15.9@xmath1 m .
this feature is not present in our high s / n template spectrum but may be present in ngc3079 , ngc3628 and ngc4676 , albeit weakly . the unidentified feature at 10.75@xmath1 m ( panel ` b ' ) shows a good correlation with the 11.3@xmath1 m complex . in the bottom panels ( ` l ' and ` m ' ) we plot this feature with respect to other weak pahs .
although there is some scatter , they do seem to correlate .
it is therefore likely that the 10.75@xmath1 m feature is either a pah or at least its carrier is carbonaceous in nature .
panel ` n ' shows a very good correlation between the satellite features in the 11.0 and 12.0@xmath1 m bands , and the last panel displays the equally good correlation between the 12@xmath1 m pahs . in summary , the strengths of the pah features scale with each other .
bottom panel : the 11.3@xmath1 m pah equivalent width with respect to the hardness of the radiation field as given by the 15.5@xmath1m/12.8@xmath1 m ratio.,width=340 ] l c c c ngc253 & 523.7 & 49.9 & 0.2 + ngc520 & 282.5 & 22.7 & 10.1 + ngc660 & 358.9 & 29.1 & 2.2 + ngc1097 & 340.6 & 18.1 & 2.5 + ngc1222 & 277.1 & 21.0 & 10.8 + ngc1365 & 342.8 & 24.9 & 3.9 + ic342 & 492.5 & 9.2 & 0.1 + ngc1614 & 363.4 & 11.4 & 21.9 + ngc2146 & 343.2 & 35.2 & 4.7 + ngc2623 & 322.7 & 11.1 & 32.7 + ngc3079 & 378.2 & 45.9 & 7.0 + ngc3256 & 360.5 & 34.4 & 21.2 + ngc3556 & 298.6 & 5.3 & 0.5 + ngc3628 & 311.6 & 26.8 & 1.3 + ngc4088 & 372.6 & 5.5 & 0.5 + ngc4194 & 362.8 & 11.2 & 9.0 + mrk52 & 472.9 & 1.5 & 0.7 + ngc4676 & 307.6 & 8.8 & 38.3 + ngc4818 & 292.2 & 32.4 & 1.4 + ngc4945 & 393.1 & 57.4 & 0.4 + mrk266 & 326.6 & 12.5 & 82.4 + ngc7252 & 357.7 & 6.4 & 13.8 + ngc7714 & 399.5 & 4.6 & 3.3 + figure 12 displays the relation between the pah features and the emission lines . in the top panel ,
the sum of the pah luminosities is plotted against the sum of the h@xmath17 lines .
both are produced in the pdr , albeit in different regions , and their fluxes correlate well . in the bottom panel
the equivalent width ( ew ) of the 11.3 @xmath1 m pahm components .
this is valid as both bands correlate well with each other ( figure 11 ) . ]
is plotted against the over line ratio .
there is no correlation , indicating that there is no dependence of the pah with the hardness of the radiation field .
in fact , the ew of the 11.3@xmath1 m shows little variation in figure 12 . in agns ,
the pah ew is seen to decrease with harder radiation field as the pah are further ionized or destroyed , but the pah ew is flat for soft radiation fields as those in starbursts ( lebouteiller et al .
, in prep . ) .
another characteristic of the pahs is their profile , which has been found to change between and within sources .
this led @xcite to establish a classification based on the peak position of the different bands .
figure 13 shows the continuum subtracted spectra for all the objects in the sample in the regions of interest ( where the emission lines have been removed ) .
the grey area in the figure represents the template spectrum .
as opposed to the template shown in figures 3 and 4 , this template has been normalized to the sum of the pahs ( and not the sum of the pahs plus lines ) , but the are both very similar .
this pah continuum subtracted spectrum of the template starburst is also available in electronic format ( tab26.txt ) .
the fact that the spectrum of each object ( thin black line in the figure ) follows that of the template ( grey area ) indicates there is hardly any variations in the profiles .
the relative strengths of the features is also the same for most of the object although there are a few variations .
the most striking is 12.7@xmath1 m pah in ngc4945 which is stronger than the 11.3@xmath1 m pah .
the reason for this is the stronger silicate dust extinction on the 11.3@xmath1 m pah than the 12.7@xmath1 m pah .
this band is also strong ( relative to the 11.3@xmath1 m pah ) in those starburst with agn signature ( e.g. ngc660 , ngc3079 , ngc3628 ) but also in a few objects with no agn ( e.g. ngc253 ) .
the 17.0@xmath1 m complex does not always follow that trend . despite these differences
the profiles are near akin . as we saw from previous sections the metallicities are similar ( within a few factor of each other ) , and
the radiation fields are low in most of them .
therefore , it is possible that the invariability in the pah spectra of the starburts in our sample are probably due to the similar conditions in these galaxies .
we detect three pure - rotational transitions 0 - 0 s(2 ) , s(1 ) , and s(0 ) of h@xmath17 at 12.28 , 17.03 , and 28.21 @xmath1 m respectively .
the s(1 ) transition at 17.03 @xmath1 m is the strongest of the three .
figure 14 shows that , while there is some scatter , there is a good correlation of the h@xmath17 lines with the line .
this is expected since both originate in the pdr .
in fact , is an important fine - structure cooling line of the pdr .
we note that in order to combine the three lines we used a scaling factor between the sh and lh modules determined by matching the continuum in the overlap region in both modules around 19.5 @xmath1 m .
this is not ideal since no off observations were subtracted and it contributes to the scatter present in the figure .
the ratio of the sum of the h@xmath17 lines over the line is 0.22 , with values ranging from 0.04 to 0.5 .
this is very similar to the values found by @xcite in their starburst ( sings ) sample .
the molecular hydrogen lines can be used to give information on the conditions of the warm component in the pdr .
there is a relation between the density and temperature assuming local thermal equilibrium and using the boltzman equation @xcite . using this relation and the lines in the sh module ( to avoid scaling effects )
, we have derived the excitation temperature ( t@xmath37 ) , column density n(h@xmath22 ) , and warm molecular mass for the sample .
these are given in table 6 .
the average excitation temperature and column density are 360@xmath2564 k and 2@xmath1110@xmath38 @xmath39 respectively , and the warm molecular mass ranges from several 10@xmath40 to 8@xmath1110@xmath41m@xmath13 , with an average mass of 1.1@xmath1110@xmath42m@xmath13 .
we stress that these numbers correspond only to the region covered by the sh slit and are based on the fluxes for just the s(1 ) and s(2 ) lines .
rotational lines measured in the irs spectrum . for this plot the fluxes of the h@xmath17 s(2 ) and s(1 ) lines have been scaled based on the sh to lh continuum ( see text for further details).,width=245 ] recently , @xcite studied the warm molecular hydrogen emission in the nuclear region of a sample of sings galaxies .
the range in areas covered in their observations ( 60 pc to 3.8 kpc ) is very similar to the range covered in our study ( see 2 ) .
their column densities compare well to those found for our starburst galaxies , but their excitation temperatures are on average cooler by @xmath0150 k , and their masses significantly larger ( 10@xmath43 - 10@xmath44 m@xmath13 ) .
this is expected because the sings galaxies have a much lower star formation rate per unit volume compared to our starbursts , and since they are less luminous , a larger fraction of their molecular hydrogen is colder .
` cold ' molecular hydrogen is traced by the s(0 ) line which is not used in our analysis as it falls in the lh slit , but it is used in the sings sample where aperture size differences do not exist .
hence , our computed @xmath45 masses are lower than those found for the sings sample , but this is likely an artifact of the absence of the s(0 ) line in our calculations , rather than a real difference .
we find no relation between the excitation temperature and sources with and without agns .
@xcite analyzed _ spitzer _ data of a sample of ulirgs .
their average excitation temperature ( 336@xmath2515 k ) agrees with that of our starburst sample , but again their warm molecular masses are several hundred s times larger ( 2@xmath1110@xmath46m@xmath13 ) .
it is important to note that most ulirgs fit in the sh aperture while in the starbursts we are only measuring the central region .
still , this difference may point as well to a larger fraction of mass being colder in the starburst sample .
1 . we have presented high resolution spectra ( r@xmath0600 ) of a sample ( 24 ) of nearby starburst galaxies .
the spectra are dominated by emission from forbidden fine - structure lines and polycyclic aromatic hydrocarbons features .
is detected in six of the eight starbursts with a previously known agn component .
this line is also seen in ngc4194 .
several pah features have been measured in the spectra , some of which are weak and are rarely seen in the spectra of starburst galaxies and hii regions .
an emission feature around 10.7@xmath1 m and an absorption feature at 13.7@xmath1 m are detected for the first time .
their identification is unknown , but the strength of the feature at 10.7@xmath1 m correlates with that other pah features .
the average conditions of the warm gas as traced by the s(1 ) and s(2 ) pure rotational molecular hydrogen lines indicates an average excitation temperature of 355@xmath2560 k and a density of 2@xmath1110@xmath38 @xmath39 .
the warm molecular mass ranges from several 10@xmath40 to 4@xmath1110@xmath41 m@xmath13 .
we provide the combined average spectrum of the 15 ` pure ' starburst galaxies of our sample , which is likely the highest signal to noise mid - infrared spectrum of a @xmath47 l@xmath13 starburst galaxy available from spitzer .
this template is representative of a large range in luminosity and ir color , and provides a good indication of the spectral properties of starburst galaxies at high redshift . | we present an atlas of _
spitzer_/irs high resolution ( r@xmath0600 ) 10 - 37@xmath1 m spectra for 24 well known starburst galaxies .
the spectra are dominated by fine - structure lines , molecular hydrogen lines , and emission bands of polycyclic aromatic hydrocarbons .
six out of the eight objects with a known agn component show emission of the high excitation line .
this line is also seen in one other object ( ngc4194 ) with , a priori , no known agn component .
in addition to strong polycyclic aromatic hydrocarbon emission features in this wavelength range ( 11.3 , 12.7 , 16.4 @xmath1 m ) , the spectra reveal other weak hydrocarbon features at 10.6 , 13.5 , 14.2 @xmath1 m , and a previously unreported emission feature at 10.75@xmath1 m .
an unidentified absorption feature at 13.7@xmath1 m is detected in many of the starbursts .
we use the fine - structure lines to derive the abundance of neon and sulfur for 14 objects where the hi 7 - 6 line is detected .
we further use the molecular hydrogen lines to sample the properties of the warm molecular gas .
several basic diagrams characterizing the properties of the sample are also shown .
we have combined the spectra of all the pure starburst objects to create a high s / n template , which is available to the community . |
tests of gravity theories within the solar system are usually analyzed in the framework of the so - called _ parametrized post - newtonian framework _ which enables the comparison of several theories through the estimation of the value of a limited number of parameters . among these parameters ,
@xmath0 and @xmath1 are of particular importance for astrometry since they are connected with the classical astrometric phenomena of the light deflection and of the excess of perihelion precession in the orbits of massive objects .
the same parameters are of capital importance in fundamental physics , for the problem of characterizing the best gravity theory , and for the dark energy / dark matter @xcite . moreover , precise astrometric measurements are also important in other tests of fundamental physics since , e.g. , they have the potentiality to improve on the ephemeris of the solar system bodies .
these are the reasons why solar system astrometric experiments like gaia and other projects presently under study @xcite have received a particular attention from the community of fundamental physicists and cosmologists . such kind of experiments , however , call for a reliable model applicable to the involved astrometric measurements which has not only to include a correct relativistic treatment of the propagation of light , but a relativistic treatment of the observer and of the measures as well . at the same time , the large amount of data to be processed , and the complexity of the problem to be solved , call for the use of high - performance computing ( hpc ) environments in the data reduction .
the development of an astrometric model based on a relativistic framework can be dated back to at least 25 years ago @xcite . in their seminal work of 1992 klioner and
kopeikin @xcite described a relativistic astrometric model accurate to the @xmath2as level foreseen for the next generation astrometric missions .
this model is built in the framework of the post - newtonian ( pn ) approximation of gr , where the finite dimensions and angular momentum of the bodies of the solar system are included and linked to the motion of the observer in order to consider the effects of parallax , aberration , and proper motion , and the light path is solved using a matching technique that links the perturbed internal solution inside the near zone of the solar system with the assumed flat external one .
the light trajectory is solved in a perturbative way , as a straight line plus integrals containing the perturbations which represent , i.e. , the effects of the aberrational terms , of the light deflection , etc .
an extension of this model accurate to @xmath3as called grem ( gaia relativistic model ) was published in 2003 @xcite .
this has been adopted as one of the two model for the gaia data reduction , and it is formulated according to the ppn ( parametrized post - newtonian ) formalism in order to include the estimation of the @xmath0 parameter .
a similar approach was followed by kopeikin , schfer , and mashhoon @xcite in the post - minkowskian approximation . in this case , however , the authors used a linard - wiechert representation of the metric tensor to describe a retarded type solution of the gravitational field equations and to avoid the use of matching techniques to solve the geodesic equations .
ramod ( relativistic astrometric model ) is another family of models , whose development started in 1995 . in this approach
the definition of the observable according to the theory of measure @xcite and the immediate application to the problem of the astrometric sphere reconstruction was privileged . as a consequence , it started as a simplified model @xcite based on a plain schwarzschild metric .
further enhancement brought to the first realistic estimation of the performances of gaia for the determination of the ppn @xmath0 parameter @xcite , and to the development of a fully accurate n - body model of the light propagation and of an observer suitable for application to space missions @xcite . since the so - called ramod3 @xcite , this model was built on a complete pm background , and the light propagation was described with the equation of motion of measurable quantities varying all along the geodesic connecting the starting point to the observer .
this approach brought to a specific form of the geodesic equations as a set of coupled nonlinear differential equations which could be solved only by numerical integration .
this represented a problem for an extensive application of this model to practical astrometric problems , which has been solved only recently for ramod3 by crosta @xcite who applied a re - parametrization of these equations of motion to demonstrate their equivalence to the model in @xcite , thus opening the road to an analytical solution of ramod3 @xcite and to its full application to astrometry problems .
finally , another class of models based on the time transfer function ( ttf ) technique , has been developed since 2004 @xcite .
the ttf formalism stands as a development of the synge world function @xcite which , contrary to all the method described so far , is an integral approach based on the principle of least action . in this models
one does not solve the system of differential equations of the geodesic equations , and thus does not retrieve the solution of the equations of motion of the photons , but it concentrates on obtaining some essential information about the propagation of these particles between two points at finite distance ; the coordinate time of flight , the direction triple of the light ray at either the point of emission ( @xmath4 ) and of reception ( @xmath5 ) , and the ratio @xmath6 of the temporal components of the tangent four - covector @xmath7 , which is related to the frequency shift of a signal between two points .
all these models are conceived to be used at least at the @xmath2as level , suitable for the accuracy foreseen by future astrometric experiments like gaia and game @xcite .
nonetheless it has to be considered that , because of the unprecedented level of accuracy which is going to be reached , both the astrometric models and the data processing software will be applied for the first time to a real case .
moreover , in the case of gaia this problem is even more delicate since here the satellite is self - calibrating and will perform _ absolute measurements _ which is equivalent to the definition of a unit of measure .
these are some of the reasons why extensive analytical and numerical comparisons among the different models are being conducted . from the theoretical and analytical point of view ,
a first comparison was conducted in showing that grem and ramod3 have an equivalent treatment of the aberration .
later the equivalence of ramod3 and the model in @xcite at the level of the ( differential ) equations of motion has then been shown in @xcite , while the explicit formulae for the light deflection and the flight time of grem , ramod3 , and ttf was compared in @xcite where it is demonstrated the equivalence of ttf and grem at 1pn in a time - dependent gravitational field and that of ttf and ramod in the static case .
numerical comparisons between the grem and the pm models showed that they give the same results at the sub-@xmath2as level . on the other side ,
grem has been compared with a low - accuracy ( @xmath8 ) version of ramod proving that even a relatively unsophisticated modeling of the planetary contributions can take into account of the light deflection up to the @xmath2as level almost everywhere in the sky .
this means that , in principle , some experiments like the reconstruction of the global astrometric sphere of gaia could initially be done by @xmath8 models .
both the analytical and the numerical comparison , however , showed that the correct computation of the retarded distance of the ( moving ) perturbing bodies is fundamental to achieve the required accuracy .
the reduction of the data coming from astrometric missions bring to the attention of the scientific community another kind of new problems , i.e. those connected to the need of reducing a huge amount of astrometric data in ways that were never experienced before .
a significant example is given by the problem of the reconstruction of the global astrometric sphere in the gaia mission . from a mathematical point of view , the satellite observations translate into a large number of equations , linearized with respect to the unknown parameters around known initial values , which constitute an overdetermined and sparse equations system that is solved in the least - squares sense to obtain the astrometric catalog with its errors . in the gaia mission
these tasks are done by the astrometric global iterative solution ( agis ) but the international consortium which is in charge of the reduction of the gaia data decided to produce also an independent sphere reconstruction named avu - gsr .
this was motivated by the absolute character of these results , and by uniqueness of the problem which comes from several factors , the main being represented by the dimensions of the system which are of the order of @xmath9 .
a brute - force solution of such system would require about @xmath10 flops , a requirement which can not be decreased at acceptable levels even considering that the sparsity rate of the reduced normal matrix is of the order of @xmath11 .
it is therefore necessary to resort to iterative algorithms .
agis uses additional hypotheses on the correlations among the unknowns which are reflected on the convergence properties of the system and permit a separate adjustment of the astrometric , attitude , instrument calibration , and global parameters , allowing the use of an embarrassingly parallel algorithm .
the starting hypotheses , however , can hardly be proved rigorously , and have only be verified `` a posteriori '' by comparing the results with simulated true values , a situation which can not hold in the operational phase with real data . moreover ,
this method by definition prevents the estimation of the correlations between the different types of unknown parameters , which constitute the other unique characteristic of this problem .
these considerations about the agis module lead to the solution followed by avu - gsr , which uses a modified lsqr algorithm @xcite ) to solve the system of equations which , however , can not be solved without resorting to hpc parallel programming techniques as explained in @xcite .
the increasing precision in the modern astrometric measurements from space makes high - accuracy tests of the dm / de vs. gravity theory debate a target accessible to future space - born astrometric missions . to this aim ,
viable relativistic astrometric models are needed , and three classes of models have been developed during the last two decades .
work is still on - going to cross - check their mutual compatibility at their full extent , but what has been done so far demonstrated that they are equivalent at least at the level of accuracy required for the gaia measurements . at the same time
these missions put new challenges to the efforts of data reduction .
we have briefly shown how the problem was faced in gaia , in the limited contest of the reconstruction of the global astrometric sphere , where an additional constraint is put by the absolute character of its main product .
this work has been partially funded by asi under contract to inaf i/058/10/0 ( gaia mission - the italian participation to dpac ) .
10 url # 1#1urlprefix[2][]#2 t , piazza f and veneziano g 2002 _ phys .
lett . _ * 89 * 81601 s g , shao
m and nordtvedt jr k 2004 _ astronomische nachrichten _ * 325 * 267277 m , vecchiato a , ligori s , sozzetti a and lattanzi m g 2012 _ experimental astronomy _ * 34 * 165180 m h 1989 _ relativity in astrometry , celestial mechanics and geodesy _ astronomy and astrophysics library ( berlin heidelberg new york : springer - verlag ) bandieramonte m , becciani u , vecchiato a , lattanzi m and bucciarelli b 2012 _ 2012 ieee 21st international workshop on enabling technologies : infrastructure for collaborative enterprises _ * 0 * 167172 issn 1524 - 4547 | we review the mathematical models available for relativistic astrometry , discussing the different approaches and their accuracies in the context of the modern experiments from space like gaia and game , and we show how these models can be applied to the real world , and their consequences from the mathematical and numerical point of view , with specific reference to the case of gaia , whose launch is due before the end of the year . |
understanding surface effects in magnetic nanoparticles is a big challenge from both the experimental and theoretical points of view .
experimentally , this challenge resides mainly in the fact that the current experimental techniques do not allow for probing the microscopic structure of the surface in round geometries , as is the case for @xmath0 magnetism @xcite .
in particular , we so far have no clear idea as to what kind of surface anisotropy we have in a nanoparticle .
the corresponding constant has been estimated by many authors in an indirect way from neutron scattering experiments @xcite , magnetic measurements @xcite , and fmr experiments @xcite , by fitting the results assuming an effective anisotropy constant .
this is a rough estimation that is only valid when deviations from collinearity are very small . on the other hand , the easy direction of the magnetization on the surface ,
let aside that of the atomic magnetic moments , is still unknown .
in fact , even the anisotropy in the core of the particle is assumed to be that of the bulk material . as such , to investigate surface effects on the properties of nanoparticles theorists can only try " some models .
the most often used model is what we call here _ transverse surface anisotropy _ ( tsa ) model which attributes an easy axis along the transverse ( radial ) direction to each spin on the boundary [ see the textbook @xcite and references therein ] .
another more physically plausible model is what we call _ nel surface anisotropy _ ( nsa ) model , developed by nel @xcite .
this model is more realistic because the anisotropy at a given lattice site only occurs if the site s environment presents some defects , e.g. , in the coordination number . in @xcite
( see also @xcite ) we used the tsa model and studied the details of the switching process in a spherical particle when the surface anisotropy constant @xmath1 is varied .
we solved the coupled landau - lifshitz equations with local and global rotation constraints .
in particular , we found that there is a characteristic value of @xmath1 separating two different regimes : for small @xmath1 , as compared to the exchange coupling , the switching is basically stoner - wohlfarth like , i.e. , coherent , whereas for large values the magnetization switching can no longer be regarded as coherent , as it operates cluster - wise .
in @xcite we used the green s function technique in the continuum limit , in the absence of magnetic field , and computed for the nsa model the contribution of the surface to the anisotropy energy of a spherical particle with simple cubic lattice structure .
we found that this contribution is second order in the surface anisotropy constant @xmath1 [ see eq .
( [ en_greenfunct ] ) below ] , scales with the particle s volume @xmath2 and has cubic symmetry with preferred directions @xmath3 $ ] .
the present work follows on the previous ones @xcite , @xcite and is an attempt to give some hints for answering the question addressed in the title , namely which one of these models is most adequate to describe surface effects in a small magnetic system .
we have no pretention here to give a definite answer to this question as we believe that only with the help of experiments , e.g. , measuring the hysteresis cycle of an isolated nanoparticle , could one do so .
however , today technical difficulties prevent such measurements [ see the review @xcite ] .
meanwhile , it is important to understand the effects of surface anisotropy within the framework of the current models , and eventually make some predictions for future investigations .
in addition , we do not try to be exhaustive since a detailed study of the magnetization switching within the tsa model has already been presented in @xcite .
instead , we only investigate the main difference between the two models with regard to the hysteresis loop and magnetic structure so as to gain some insights into their relevance to the study of surface effects in nanomagnets .
more precisely , we compute the hysteresis loop and magnetic structures of a spherical particle with simple cubic lattice structure and uniaxial anisotropy in the core with easy axis @xmath4 .
we consider both cases of vanishing and non - vanishing exchange interaction between spins inside the particle . the unrealistic non - interacting case is considered only because it allows for a comparison of the tsa and nsa in pure form .
our model hamiltonian is given by the ( classical ) anisotropic dirac - heisenberg model @xcite @xmath5 where @xmath6 is the unit spin vector on site @xmath7 , @xmath8 the uniform magnetic field , @xmath2 the total number of spins ( core and surface ) , and @xmath9 the nearest - neighbor exchange coupling .
@xmath10 can be the uniaxial single - site anisotropy energy @xmath11 with easy axis @xmath12 and constant @xmath13 . if the spin at site @xmath7 is in the core , the anisotropy axis @xmath14 is taken along the reference @xmath4 axis and @xmath15 .
otherwise , for surface spins , this axis is along the radial ( or transverse ) direction and @xmath16 . in this case , the model in ( [ uaa ] ) for surface spins is the tsa model .
the second model that will be studied and compared with the previous one , is the more physically appealing model of surface anisotropy that was introduced by nel @xcite , @xmath17 where @xmath18 is the coordination number of site @xmath7 and @xmath19 a unit vector connecting the site @xmath7 to its nearest neighbors @xmath20 .
this model is more realistic because the anisotropy at a given site occurs only when the latter loses some of its neighbors , i.e. , when it is located on the boundary .
this is the nsa model . in this work
we ignore dipolar interactions between spins inside the particle , since in our case of small spherical particles , the volume term is negligible @xcite and the shape anisotropy yields an irrelevant constant . in addition , here we restrict ourselves to the investigation of pure " surface anisotropy within the two models . in the sequel , we will use the following notation @xmath21 for field , with the magneto - crystalline ainsotropy constant @xmath22 in the core taken as the energy scale , @xmath23 for the exchange coupling , which will be taken in our calculations the same everywhere inside the particle , @xmath24 , and @xmath25 denotes the particle s diameter .
we study hysteretic properties of a nanoparticle by solving the ( local ) landau - lifshitz equations at zero temperature .
the method used to obtain the hysteresis loop and other characteristics were given in great detail in @xcite and will not be repeated here .
we compare the implications of the two models , tsa and nsa , on the hysteretic properties of a nanoparticle , such as hysteresis loop and coercive field as a function of the surface anisotropy strength @xmath26 . in all these calculations
the anisotropy in the core is taken as uniaxial along the reference axis @xmath4 , and the field is applied at an angle of @xmath27 with respect to the latter in the @xmath28 plane .
in order to compare the tsa and nsa models in pure form , we first consider the case of a spherical particle of non - interacting spins ( @xmath29 ) , and the core - anisotropy constant equal to that on the surface . in fig .
[ t_n_j0_d10_ks1 ] we plot the hysteresis loop for both tsa and nsa models . first , in both cases the core switches at the same field of @xmath30 since
the anisotropy axis in the core is along @xmath4 and makes the angle @xmath27 with respect to the field direction . in the case of tsa , at zero field
the surface spins are obviously directed along the radial direction [ see fig . [ struct_t_d10_j0_k1 ] @xmath31 . at non zero field , as was discussed in detail in @xcite , because of the radial direction of the surface anisotropy easy axes , the latter make different angles with respect to the field direction , and hence switch at different values of the applied field .
this leads to a cluster - wise progressive switching of surface spins , as can be seen in the structures of fig .
[ struct_t_d10_j0_k1 ] at the respective field values @xmath32 at which no spins , 2 spins , 3 spins , and all spins have switched . + + for the surface anisotropy constant used in fig .
[ struct_t_d10_j0_k1 ] the core switches before the full switching of the surface . in the case of nsa ,
the situation is quite different .
the hysteresis loop exhibits a first jump at a very small field , and more importantly this jump corresponds to the switching of the entire surface .
this can be seen in fig .
[ struct_n_d10_j0_k1 ] by examining the structures at the field values @xmath33 and @xmath34 . for smaller particles with more spins on the boundary than in the core
, the first jump would also correspond to the switching of the particle s magnetization .
+ + moreover , in fig .
[ t_n_j0_d10_ks1 ] there appear two jumps corresponding to surface spins with two different local environments : the spins on the faces and those located near the corners . however , this difference in the switching field is just a numerical artefact , since as will be shown shortly these two groups of spins jump at the same field value , that is zero .
indeed , a spin on the face has 5 nearest neighbors , so that the local energy given by ( [ nsa ] ) yields , @xmath35.\end{aligned}\ ] ] where @xmath36 are the spherical coordinates of the spin @xmath37 .
the arrow in eq .
( [ energy_face ] ) indicates that the irrelevant constant @xmath38 is discarded .
so , in zero field the energy ( [ energy_face ] ) has a minimum in the @xmath39 plane .
similarly , a spin near the corner has three nearest neighbors ( two on the same plane and one on the adjacent plane ) , so that its energy reads @xmath40.\ ] ] whose minimum , in zero field , is along the diagonal of the cube [ see eqs .
( [ en_greenfunct ] , [ en_difference ] ) below and ref .
@xcite ] . as noted above , while the smallest switching field for a tsa spin is @xmath30 , here we see that the switching field of an nsa spin may assume very small values .
indeed , for spins on the faces and near the corners , it can be shown from ( [ energy_face ] ) and ( [ energy_corner ] ) , respectively , that the switching field is zero .
for example , the magnetization direction which yields the lowest energy of an nsa spin on the face is given by @xmath41 and @xmath42 , \nonumber\\ & & a \equiv a - \sqrt{1+a^2},\quad a\equiv \frac{h}{\sqrt{2}},\end{aligned}\ ] ] which when inserted in the determinant of the energy hessian yields a zero switching field .
this is confirmed by the result of the numerical calculations in fig .
[ t_n_j0_d10_ks1 ] , up to numerical precision .
in fact , the reason for such small switching fields for surface spins in the nsa model is quite simple , and one actually does not need the above calculations to convince onself .
indeed , for spins on the face there is a rotation symmetry which makes all directions in , e.g. , the @xmath39 plane degenerate , see eq .
( [ energy_face ] ) .
so , when the field increases from the negative saturation value towards zero , these spins select the direction given by the projection of the field direction @xmath43 on the @xmath39 plane .
as soon as the field points into the opposite direction , even a very small value thereof is sufficient to make the spins on the face reverse their direction .
the same arguments hold for spins near the corners , since these too have full symmetry in the plane containing their easy direction . in order to investigate the surface influence on the core and vice - versa
, we consider the same particle as before but now with realistic non - vanishing exchange coupling everywhere in the particle .
the results of the hysteresis loop and some of the corresponding spatial distributions of the magnetization are shown in figs .
[ t_n_j1_d10_ks015 ] and [ mag_struct_jnotzero ] , respectively .
+ + in this case of fully interacting spins , we see that as the field goes from @xmath44 to @xmath45 , the magnetization of a nsa particle experiences its first jump , while that of the tsa particle only decreases ( in absolute value ) in a progressive manner .
this is due to the fact that the angle of rotation between the direction at @xmath33 and that at @xmath46 , and thereby the difference in the projection of magnetization along the field , is much larger for the nsa particle than for the tsa one .
the reason is that at zero field , in the case of the tsa model , all core spins and those surface spins located on the @xmath39 faces are directed along the easy axis @xmath4 , and hence make up a strong effective field that forces the spins on the @xmath47 faces into nearly the same direction as in the core . as to the nsa model
, there are near - corner spins directed along the diagonal of the cube , spins on the @xmath39 faces which lie in these faces , and finally the more numerous core spins directed along their easy axis @xmath4 .
this results in a competition which ends up with a net magnetization in a direction inside the cube with its largest component along @xmath4 .
these results suggest that the surface has a stronger effect in the case of nsa model as compared to the tsa model . indeed , in fig .
[ mag_struct_jnotzero ] ( nsa ) , for instance , we see that at zero field the core spins are deviated from their easy direction .
moreover , in the non - interacting case , in the nsa model , the surface jumps in a coherent way at a much lower field than the core which also switches coherently . in the interacting case ,
the surface imposes , via exchange coupling , its switching at low fields on the core , and this results in an abrupt partial coherent rotation of all spins , core and surface .
the core then proceeds to switch and drives the whole bunch of spins at a higher field . in @xcite
we studied the nsa model using both the numerical solution of landau - lifshitz equation at zero temperature and the analytical green s function technique in the continuum limit , though preserving the discreteness of the lattice by introducing a spin density function .
we showed that the contribution of the nsa to the energy of a spherical particle , in the absence of core anisotropy and applied field , reads , @xmath48 where @xmath49 , are the components of the net magnetization of the particle , @xmath50 is the fourier transform of the exchange coupling @xmath51 , @xmath52 the coordination number in the core , and @xmath53 a surface integral . computing
the energy difference between two major orientations of @xmath54 leads to @xmath55 this indeed confirms what we said above , namely that nsa has a stronger effect than tsa . in fig .
[ nsa_j1_d10_vsks ] we plot the hysteresis loop as obtained from the nsa model for different values of the surface anisotropy constant @xmath26 .
these results demonstrate that actually the amplitude of the abrupt coherent rotation and the field at which it occurs increase with @xmath26 .
moreover , the length of the subsequent shoulder " corresponding to a smooth rotation of the whole spins and that precedes the magnetization switching decreases with @xmath26 and vanishes at some higher value of @xmath26 ( @xmath56 here ) .
this means that as @xmath26 becomes large , the amplitude ( or angle ) of the abrupt coherent rotation , induced by surface anisotropy , becomes so large that this rotation coincides with the switching of the net magnetization and thus no other steps are observed . in summary ,
surface anisotropy in the tsa model induces several jumps in the hysteresis loop because of the cluster - wise switching of spins @xcite .
this is due to the single - site nature of the anisotropy in this model , and to the radial direction of surface anisotropy axes , and thence to the distribution of angles that these axes make with the field direction .
in the nsa model , the existence of several jumps in the hysteresis loop does not require high values of the surface anisotropy constant [ see fig .
[ t_n_j1_d10_ks015 ] ] , and is due to successive coherent partial rotations of all spins in the particle . for large values of the surface
anisotropy constant , there is only one big coherent rotation which coincides with the net magnetization switching . as in ref .
@xcite where the switching field was also studied as a function of the surface anisotropy constant @xmath26 , we have performed the same calculation here for the nsa model , for the sake of comaprison .
the results are presented in fig .
[ hcvsks_t_n ] . for the tsa model
, there is a sudden increase in @xmath57 occurring at a value of @xmath58 of order one .
this was interpreted in @xcite as a critical value marking a departure from the stoner - wohlfarth model , i.e , a transition from the coherent - rotation regime , where the hysteresis loop exhibits only one jump , into a regime of cluster - wise magnetization switching , where several jumps appear in the hysteresis loop . in the case of nsa
, @xmath59 assumes much lower values than in the tsa model as already discussed earlier , and increases in a more progressive way . on the other hand ,
it is apparent from the inset in fig .
[ hcvsks_t_n ] that there is in fact some kind of transition between different regimes , though with a lower critical value " for @xmath58 than for tsa model .
we have compared the effects of transverse surface anisotropy ( tsa ) and nel s surface anisotropy ( nsa ) on the hysteretic properties of a spherical particle . in the non - interacting case ,
i.e. , with zero exchange coupling everywhere inside the particle , the nsa induces more jumps than the tsa in the hysteresis loop the first of which occurs at very low fields , and corresponds to the switching of the entire surface .
this is due to the rotational symmetry which renders all directions in , e.g. , the @xmath39 plane , degenerate .
moreover , in the nsa model , the spin switching within the surface and within the core is coherent . in the interacting case with the same exchange coupling in the whole particle , we still have qualitatively the same hysteresis loop as before , but now the first jump in the nsa model occurs at a field slightly larger than before , and more importantly corresponds to a coherent partial rotation of all spins ( core and surface ) .
more precisely , in this case the surface imposes its switching on the core , and thereby the two jumps in the hysteresis loop , in fact , correspond to two successive coherent rotations of the whole bunch of spins inside the particle .
this shows that , for the same constant , the surface anisotropy in the nsa model has a stronger effect than in the tsa model .
moreover , spin switching in tsa is cluster - wise , while that for nsa is coherent but operates by successive coherent partial rotations .
experimental results on individual cobalt clusters [ see the review @xcite ] on the stoner - wohlfarth astroid , i.e. , the angular dependence of the switching field , together with the switching field distribution and the non - switching probability , have confirmed the validity of the stoner - wohlfarth and nel - brown models describing the magnetization reversal by uniform rotation @xcite at zero and non - zero temperature , respectively . on the other hand , even though relatively large surface anisotropy constants are obtained for cobalt particles @xcite , the results are still well fitted by the aforementioned models that are based on coherent rotation of the net magnetization .
this is at variance with what one expects from the tsa model with large anisotropy constant , as was demonstrated in @xcite and reviewed here .
therefore , in addition to the fact the nsa model is intuitively more realistic , these experimental indications together with the present calculations give a further hint that the surface anisotropy in nanomagnets is most probably of the nsa kind .
however , a more convincing check would only come from direct measurements of the hysteresis loop of an individual nanoparticle , were they to become possible .
h. mahboub thanks the laboratoire de magntisme et doptique for financial support and for the hospitality extended to her during her training 1 december 2002 - 31 july 2003 .
j. l. dormann , f. dorazio , f. lucari , e. tronc , p. pren , j. p. jolivet , d. fiorani , r. cherkaoui , and m. nogus , phys .
b53 ( 1996 ) 14291 ; j .- l .
dormann , d. fiorani , and e. tronc , adv .
98 ( 1997 ) 283 ; r.h .
kodama , a.e .
berkovitz , e.j .
mcniff , and s. foner , phys .
77 ( 1996 ) 394 ; r.h .
kodama and a.e .
berkovitz , phys .
b59 ( 1999 ) 6321 .
v. shilov , _ effects of surface anisotropies of the ferromagnetic resonance in ferrite nanoparticles _ , ph.d .
thesis , universit paris vii , 1999 ; f. gazeau , j.c .
bacri , f. gendron , r. perzynski , yu.l .
raikher , v.i .
stepanov , and e. dubois , j. magn .
mater . 186
( 1998 ) 175 . | through the hysteresis loop and magnetization spatial distribution we study and compare two models for surface anisotropy in nanomagnets : a model with transverse anisotropy axes and nel s model .
while surface anisotropy in the transverse model induces several jumps in the hysteresis loop because of the cluster - wise switching of spins , in the nel model the jumps correspond to successive _ coherent partial rotations _ of the whole bunch of spins .
these calculations , together with some hints from available experimental results , suggest that nel s model for surface anisotropy is more appropriate .
_ nanomagnetism , magnetization switching , hysteresis _ : 75.50.tt - 75.10.hk |
since the observations of ngc 1068 in polarized light by antonucci & miller ( 1985 ) , showing the presence of broad permitted lines in emission , the idea of the unification scheme of radio quiet seyfert galaxies and quasars emerged ( for reviews see antonucci 1993 ; robson 1996 ; peterson 1997 ; wills 1999 ; krolik 1999 ) . the simplest version of the scheme assumes the presence of a dusty `` torus '' surrounding the central regions of the active galactic nucleus ( agn ) intercepting a fraction of the illuminating accretion disc radiation and re emitting it in the infrared .
if the absorption is due to dust , there is a natural temperature scale in the system , since dust sublimates for temperatures greater than @xmath1 k , corresponding to a peak in the corresponding black body spectrum at @xmath2 hz ( or @xmath3 m ; the 3.93 factor is appropriate for the peak in the @xmath4 spectrum ) .
the torus origin , stability , structure ( see e.g. krolik & begelman 1988 ) and its very presence in both highly luminous radio quiet quasars and in low luminosity radio loud sources is under debate .
the amount of reprocessed ir radiation seems to become smaller for larger optical luminosity in radio quiet objects ( i.e. `` receding torus '' , lawrence 1991 ) , while , for radio sources , the absence of broad emission lines in low power fr i radio galaxies ( and bl lacs ) could be intrinsic , and not due to an obscuring torus ( chiaberge , capetti & celotti 1999 ) .
along the years , the idea of a simple and uniform `` doughnut '' around the accretion disc has been replaced by a clumped material , possibly outflowing ( or inflowing ) , as envisaged and modeled by many authors ( see e.g. elvis 2000 ; risaliti , elvis & nicastro 2002 ; elitzur & shlosman 2006 ; nenkova 2008 ) .
the existence of the unifying scenario based upon intrinsically equal but observationally different agn is also at the base of synthesis models of the x ray background ( setti & woltjer 1989 ; madau , ghisellini & fabian 1994 ; comastri et al . 1995 ; gilli , comastri & hasinger 2007 ) , since also the x
rays are partly absorbed , and partly ( compton ) reflected by the torus ( ghisellini , haardt & matt 1994 ) : for large viewing angles , the observed x ray emission becomes very hard , as required to fit the x ray background .
the covering factor of the absorbing material forming the `` torus '' is not well known .
estimates come from direct observations of optical and ir agn , as well from statistical considerations concerning the number of type 1 and type 2 agn . in the first case ,
the studies were hampered up to now by the relatively small samples of objects ( especially in the ir ) suitable for a combined study ( see e.g. landt et al .
2011 for a sample of 23 objects observed spectroscopically in the optical and in the ir , down to @xmath5 m ) . in order to study the accretion disc torus connection in agn , we need to collect the largest group of radio quiet agn with reliable detections of the ir luminosity and an optical spectrum to characterize the accretion disc features .
the sloan digital sky survey ( sdss ; york et al .
2000 ) and the wide field infrared survey explorer ( wise ; wright et al .
2010 ) are the catalogs with the widest number of objects in these two bands , hence they are the most appropriate for our study .
wise provided photometric observations in 4 ir bands ( 3.4 , 4.6 , 12 and 22 @xmath6 m ) for half a billion sources ( all sky ) with fluxes larger than 0.08 , 0.11 , 1 and 6 mjy in unconfused regions on the ecliptic in the four bands .
the sensitivity improves toward the ecliptic poles due to denser coverage and lower zodiacal background .
we adopt a flat cosmology with @xmath7 km s@xmath8 mpc@xmath8 and @xmath9 .
we consider the fifth edition of the sdss quasar catalog ( schneider et al .
2010 ) , containing 105,783 quasars with magnitude smaller than @xmath10 ( i.e. @xmath11@xmath12 ) , at least an emission line with fwhm@xmath13 and a reliable spectroscopical redshift .
continuum and line luminosities ( as well as many other spectral properties ) in sdss spectra have been measured by shen et al .
( 2011 ; hereafter s11 ) .
we will use these data to estimate optical bolometric luminosity of the sources , following two independent methods .
the first one relies on the 3000 continuum luminosity : @xmath14 ) ( e.g. elvis et al . 1994 ; richards et al .
2006 ) . the second one will be used for a consistency check of our results , and relies on h@xmath15 and mgii line luminosities ( [ sec - lines ] ) . in the following
we will assume that the bolometric luminosity equals the accretion disc luminosity .
the superscript `` iso '' reminds that they are derived under the assumption of isotropic emission . the requirement that all sources are observed in the rest frame 25005500 range ( to comprise both the h@xmath15 and mgii lines and the continuum at 3000 ) sets our first selection criterion .
given the wavelength coverage of the sdss , we require a corresponding redshift range : @xmath16 . the s11 catalog has also been cross correlated with the faint images of the radio sky at twenty centimeter survey ( first ; becker et al .
1995 ) and hence s11 include in their sample the radio fluxes .
the flux limit of the first sample is @xmath01 mjy at 1.4 ghz .
therefore , we can select the radio quiet quasars as those objects observed by the first without a detectable radio flux . the radio quiet requirement ensures the absence of a contamination from the jet in the wavelength intervals of interest . after the radio quietness and the redshift selections ,
we are left with 5122 sources .
we have cross correlated this sample with the wise all sky source catalog requiring that the optical and ir positions are closer than 2 arcsec ( 5082 sources ) , and selecting only those objects with detections in all the four wise ir bands , to have the most complete ir luminosity information .
this last selection leaves us with a sample of 3965 wise detected , radio quiet type 1 agn in a redshift range @xmath17=0.560.73 .
for all the 3965 sources in our sample we computed the ir flux in the four wise bands by first transforming the observed ( vega ) magnitudes in the ab systems setting @xmath18 , with @xmath19 given in tab .
[ tab - wiseflux ] . in the ab system ,
magnitude relation is simply : @xmath20 where the flux density is measured in erg s@xmath8 @xmath21 hz@xmath8 . .
center wavelengths and frequencies of the four wise bands , and value of @xmath19 needed to transform the magnitudes given in the vega system to the ab one .
see cutri et al .
( 2012 ) in http://wise2.ipac.scaltech.edu/docs/release/allsky . [
cols="^,^,^,^",options="header " , ] the estimates given above do not take into account that the observed optical continuum can include different components , besides the disc emission . as a consistency check ,
we use an alternative method to derive the disc ( bolometric ) luminosity , by using the luminosities of the h@xmath15 and the mgii broad lines , always present in the sdss spectra in our redshift selection . for
radiatively efficient discs , indeed , the overall luminosity of the broad line region ( blr ) , @xmath22 , is a proxy of the disc luminosity @xmath23 , since on average @xmath24 , where the factor @xmath25 is directly connected to the blr covering factor ( see e.g. baldwin & netzer 1978 ; smith et al . 1981 ) . in turn , @xmath26 should be equal to @xmath23 ( real , not isotropically equivalent ) .
estimates of @xmath25 lie in quite large ranges , historically between 0.002 and 0.35 ( according to baldwin & netzer ) , but preferentially @xmath27 ( smith et al . 1981 ) .
typically , an average value @xmath28 is assumed .
@xmath29 can be calculated from individual broad line luminosities , as in celotti et al .
specifically , setting the ly@xmath30 flux contribution to 100 , the relative weights of the h@xmath30 , h@xmath15 , mgii and civ broad lines are 77 , 22 , 34 and 63 , respectively ( francis et al .
the total broad line flux is fixed at 555.8 . since
all our sources have measured h@xmath15 and mgii line luminosity , we average for each object the estimates of @xmath29 : @xmath31\ ] ] a good agreement between the continuum based and the blr based bolometric luminosities is obtained using @xmath32 . this value is found for sources in the highest luminosity subsample ( c ) , for which we do not have spurious contributions from components other than the disc ( e.g. host galaxy ) . fig . [
fig - histo - covering ] shows the histogram of the ratio @xmath33 for all sources ( green solid line ) .
-0.3 cm , as computed using integrated ir luminosity @xmath34 and both the continuum based ( red solid line ) and blr based ( green solid line ) bolometric luminosities . dashed lines show the cumulative fraction ( values on right axis).,title="fig:",width=340 ] -0.3 cm this histogram not only has the same average of the distribution of @xmath33 based on the 3000 luminosity ( which is expected , given our assumptions ) , but the two distributions are similar for all @xmath33 , implying that broad lines are a good proxy to compute the total disc luminosity , and that our estimates of @xmath33 are reliable .
consider the simplest case of a doughnut shaped torus with opening angle @xmath35 , as measured from the symmetry axis .
the covering factor @xmath36 is defined as : @xmath37 we must relate @xmath36 to the observed ratio @xmath33 , accounting for the anisotropy of disc and torus emission . since the emission of geometrically thin discs follows a @xmath38 pattern , for a given viewing angle @xmath39 ( calculated from the disc axis ) the ratio between the real disc luminosity @xmath23 and the isotropic estimate @xmath40 is : @xmath41 the ratio is smaller than unity for @xmath42 , thus for type 1 agn we likely have @xmath43 .
we are not able to determine @xmath44 for each source , but we can safely assert that @xmath45 , since we are dealing with type 1 agn . therefore a reasonable estimate is :
@xmath46 a relation similar to eq .
[ eq - anisotropy - disc ] for the torus luminosity ( @xmath47 ) is currently unknown .
however , we can reasonably state that @xmath48 the lower limit corresponds to a thin disc shaped emitting torus , the upper limit to an isotropic emitting torus .
both limits are rather unrealistic : the torus is expected to show a lower degree of anisotropy than the disc since we are able to detect radiation emitted from the side ( i.e. type 2 agn ) ; also , the torus is hardly an isotropic emitter since ir signatures are different in type 1 and 2 agn ( calderone et al . , in prep . ) .
the above limits should then bracket the real case .
the amount of disc radiation intercepted ( and re processed ) by the torus is : @xmath49 rearranging the previous equations , we find a relation between the observable parameter @xmath50 and the covering factor @xmath36 : @xmath51 this relation can be inverted to find the allowed range of @xmath36 and @xmath35 , given a value of the observable parameter @xmath33 .
finally , the covering factor @xmath36 can be used to estimate the count ratio between type 1 and type 2 agn : @xmath52 the last three columns of tab .
[ tab - rescov ] report the value of @xmath36 , @xmath35 and # 2/#1 corresponding to the observed values of @xmath33 in all discussed samples .
the main result of our work is the determination of the average covering factor of the torus using a very large data set .
the _ observed _ fraction of ir to bolometric , isotropically equivalent optical luminosity is about 30% .
this implies that the obscuring torus covering factor is in the range 0.50.7 and that the opening angle @xmath35 is 40@xmath53 60@xmath53 . on average ,
our sources emit in the ir a similar fraction of their bolometric luminosities ( @xmath54 ) . for each type
1 agn , there should be between 1 and 3 type 2 sources .
if there is a broad distribution of covering factors ( as suggested by elitzur , 2012 ) our type 1 sample may be drawn preferentially from the lower end of the distribution . in this case our estimate of # 2/#1 ratio is a lower limit .
the very basic prediction of the unified model that the torus re processes a given amount of disc luminosity is verified ( fig .
[ fig - whole ] and fig .
[ fig - histo - covering ] ) .
the dispersion of this fraction is remarkably small , being at most a factor of 2 .
the broad band spectral energy distribution ( sed ) from ir to near uv are expected to be quite similar among type 1 agn .
a hint of the `` receding torus '' hypothesis is found in fig .
[ fig - whole ] , with @xmath55 . in fig .
[ fig - wise - sdss ] we show both data and model for a prototypical broad band sed , in the three luminosity classes considered above ( coded with colors ) . for each subsample
we also compute a composite spectrum using sdss spectra .
-0.3 cm at ir wavelengths the torus emission dominates .
spectral indices between the four wise bands are very similar for different overall luminosities ( tab .
[ tab - reswise ] ) . despite the rather poor coverage
, it appears that the ir emission is structured with at least two broad bumps .
such features are easily modeled by the superposition of two black bodies with temperatures of @xmath0300 k and @xmath01500 k respectively .
a nave interpretation is to consider the hotter one as emitted from the hot part of the torus facing the disc , at the dust sublimation temperature .
the colder one would come from the cooler outer side of the torus .
this should be the region visible also in type 2 agn .
the underlying optical continua are well described by a standard shakura & sunyaev ( 1973 ) accretion disc spectrum .
the dashed lines in fig . [ fig - wise - sdss ] are the models of three accretion discs having the same bolometric luminosity as the spectra in the subsample , and masses @xmath56 , @xmath57 and @xmath58 m@xmath59 respectively , grossly in agreement with the ( virial ) masses calculated in s11 .
the wise data points ( in @xmath4 ) lie a factor @xmath03 below the disc peaks ( at log(@xmath60/hz)@xmath015.5 ) .
this factor corresponds to the value @xmath01/3 found in tab .
[ tab - rescov ] and fig .
[ fig - histo - covering ] .
the composite spectra follow closely the accretion disc continuum in all but the lowest luminosity subsample , in which some other component is present at frequencies below log(@xmath60/hz)@xmath6114.9 .
this further component may be the starlight contribution from host galaxy ( vanden berk et al .
2011 ) , as shown by the yellow line which is the sum of the accretion disc spectrum and an appropriately scaled template for an elliptical ( quiescent ) galaxy from mannucci et al .
( 2001 ) . at higher
luminosity subsamples , the contribution from galaxy becomes relatively less important .
this publication makes use of data products from the wide field infrared survey explorer , which is a joint project of the university of california , los angeles , and the jet propulsion laboratory / california institute of technology , funded by the national aeronautics and space administration . | we selected all radio quiet agn in the latest release of the sloan digital sky survey quasar catalog , with redshift in the range 0.560.73 .
about 4000 ( @xmath080% ) of these have been detected in all four ir bands of wise ( wide field infrared survey explorer ) .
this is the largest sample suitable to study the disc
torus connection .
we find that the torus reprocesses on average @xmath01/31/2 of the accretion disc luminosity .
[ firstpage ] galaxies : active quasars : general infrared : general |
gravitational weak lensing by large scale structures is known to be a promising aproach to probe the mass distribution in the universe .
gravitational lensing by cluster of galaxies already gave new insights about the cluster mass and their 2d mass distribution whereas observation of lensing by large scale structure is only at its beginning ( @xcite ) .
its potential scientific impact on the knowledge of the cosmological parameters and the mass power spectrum pionnered by @xcite , @xcite , @xcite , @xcite , has been developped by many authors ( see @xcite , @xcite , @xcite , @xcite ) .
one may also be interested in comparing the mass maps reconstructed from weak lensing and the galaxy catalogues ( i.e. the mass and the light distributions ) in order to probe the so - called bias relation .
galaxy catalogues ( via high order statistic ) and velocity fields being the standard ways to probe the bias , a method able to _ see _ directly the mass , like gravitational lensing , may play a role . in the following some clues about the measurement of bias using gravitational lensing
will be exposed , most of them are detailed in @xcite .
biasing theory has known recently a raise of interest because of the growing evidences that it is not trivial . however , in order to give some insights on the method , local linear biasing relation is used as a basis along this paper , and the extension to non trivial biasing is let for a forthcoming paper @xcite . the basic idea is to probe the bias by measuring the correlation @xmath0 between the projected density contrast of a set of foreground galaxies and the shear of background galaxies .
the difficulty with the gravitational lensing is that it gives the projected mass along the line of sight , whereas the 3d mass field is required in order to constrain the bias , in particular if we wish to investigate its redshift evolution .
fortunately the @xmath1 statistic introduced by @xcite gives the possibility to deproject the observed mass distribution . to understand how it is possible ,
let us write the correlation function @xmath0 explicitly @xcite , @xmath2 @xmath3 is the convergence smoothed with a compensated filter of size @xmath4 , and @xmath5 is the projected density contrast of the foreground galaxies smoothed with the same filter .
@xmath6 is the linear bias parameter , @xmath7 is the comoving distance to a redshift @xmath8 , @xmath9 is the comoving angular diameter distance , and @xmath10 is the time - evolving 3-d mass power spectrum .
@xmath11 is the density parameter , and @xmath12 is the cosmic expansion factor .
the function @xmath13 depends on the redshift distribution of the sources @xmath14 and @xmath15 is the redshift distribution of the foreground galaxies .
@xmath16 is the fourier transform of the compensated filter , which is very localised in the fourier space around the angular scale @xmath4 .
if in addition a narrow redshift distribution is used for the foreground galaxies then the integrals in eq .
( [ mn_def ] ) are almost reduced to a point in the @xmath17 plane , which is equivalent to a deprojection ( a single angular scale @xmath18 is observed at a given redshift @xmath19 ) .
a useful estimator to probe the bias , independent on the power spectrum normalisation is , @xmath20 where @xmath21 is the dispersion of the smoothed projected density contrast of the foreground galaxies , @xmath22 in the case of linear theory , for a power law power spectrum and for a narrow foreground redshift distribution ( located around @xmath19 ) it is easy to show from eq.([mn_def ] ) and eq.([n2_def ] ) that eq.([r_def ] ) can be approximated by , @xmath23 which is a number depending only on the cosmological parameters , the biasing factor @xmath6 and the foreground and background redshift distributions .
it does not depend on the power spectrum index and the smoothing angle @xmath4 .
it turns out that this is still true for a general power spectrum even in the non linear regime .
the reason for this is that if the redshift integration in eq.([mn_def ] ) and eq.([n2_def ] ) are narrow enough , a general power spectrum can be approximated _ locally _ as a power law because of the narrow function @xmath16 . _
locally _ here means in the @xmath17 space .
an illustration of this is given if figure [ script_r.ps ] where @xmath24 and @xmath25 are plotted versus the smoothing angle @xmath4 for different cosmological models , and with the narrow foreground ] redshift distribution @xmath26 well localised around @xmath27 .
the left plot shows the value of @xmath28 and the right plot the ratio @xmath29 .
different linestyle correspond to different cosmological models as indicated on the left plot , thin lines are for the linear regime and thick lines for the non - linear regime .
cdm - like power spectrum has been used .
the power spectrum evolution has been extended into the non - linear regime via the peacock & dodds formula in @xcite .
figure [ script_r.ps ] shows that @xmath24 is has a constant value to a percent accuracy over the scale range @xmath30 $ ] which only depends on the cosmological parameters , the bias ( at the redshift of the foregrounds ) and the redshift distributions whatever the power spectrum .
calculations with various power spectra shows that @xmath24 changes by less than @xmath31 .
the @xmath24 values fitted to the @xmath11 dependence ( assuming a zero cosmological constant ) gives , @xmath32 where the proportionality constant and the exponent both depend on the redshift distributions according to eq.([r_approx ] ) .
conjointly used with the velocity fields or any independent determination of @xmath11 , eq.([r_fit ] ) allows to determine the bias at the redshift of the foreground galaxies ( here roughly @xmath33 ) and at the scale @xmath4 from weak lensing data provided that the redshift distributions are known .
preliminary investigations show that only the mean and the variance of these distributions are important even if they are skewed .
the quantity @xmath34 defined above gives a direct estimate of the scale dependence of the bias , independent on the cosmology , the power spectrum and the redshift distributions . strictly speaking it gives the ratio of the fourier components of the bias at two different scales provided that the bias can be expressed as a convolution , @xmath35 ( linear non - local non - stochastic biasing sheme ) .
a signal to noise analysis shows that assuming that @xmath36 is perfectly known ( from the future big galaxy surveys for instance ) , a lensing survey of @xmath37 degrees ( with 4 hours exposure depth ) should be able to detect a bias scale variation of @xmath38 or more in the @xmath39 $ ] scale range .
gravitational lensing can probe the bias relation between mass and light distributions .
the simultaneous use of a narrow foreground redshift distribution and a compensated filter permits to _ deproject _ the observed mass distribution and to recover the redshift and scale dependence of the bias using eq.([r_fit ] ) and the estimator @xmath34 .
redshift distorsion effects are not a problem in this method since the foreground redshift range is large enough do dilute these effects , and on the other hand , it is small enough to assume that a given angular scale corresponds to the physical scale at which the bias is probed . of course , the statistical bias properties are assumed to remain unchanged within the foreground galaxies redshift interval . obviously the biasing sheme adopted in this paper is too simple and should be extended to non - linear , stochastic and non - local biasing .
one can hope to use weak lensing jointly with galaxy catalogues and velocity fields to separate these features and try to build a consistent bias picture .
99 bernardeau f. , van waerbeke l. , mellier y. , 1997 , , 322 , 1 blandford r. d. , saust a. b. , brainerd t. g. , villumsen j. v. , 1991 , , 251 , 600 gunn , j.e .
, 1967 , , 150 , 737 jain , b. , seljak , u. , 1997 , , 484 , 560 kaiser n. , 1992 , , 388 , 272 kaiser n. , 1998 , apj , 498 , 26 mellier , y. , fort , b. , 1998 , to appear in araa miralda - escud , j. , 1991 , 380 , 1 peacock , j.a . , dodds , s.j . , 1996 , , 280 , l19 schneider , p. , 1998
, , 498 , 43 schneider , p. , van waerbeke , l. , jain , b. , kruse , g. , 1998 , , 296 , 873 van waerbeke , l. , 1998 , 334 , 1 van waerbeke , l. , 1998 , in preparation villumsen , j.v . , 1996 , , 281 , 369 | gravitational weak lensing by large scale structures is view as a tool to probe the bias relation between the mass and the light distributions .
it is explained how a particular statistic can be used to deproject the 2d mass distribution observed by weak lensing , in order to probe the bias at a given redshift and a given scale .
dependence with the cosmology is written , and observational issues are pointed out , like the importance of the redshift distribution of the galaxies .
a signal to noise analysis shows that the scale dependence of the bias can be measured with a rather modest lensing survey size . # |
consumer spending is an essential component of the economy , accounting for 71% of the total us gross domestic product ( gdp ) in 2013 @xcite . in spite of representing just over 10% of all consumer spending @xcite
, online shopping is rapidly growing as people are becoming more comfortable with the online payment systems , security and delivery of the purchased goods . over the last three years
, online sales grew over @xmath3 @xcite and are showing signs of exponential growth .
one of the largest and fastest - growing online markets is apple s ios market , where people can make digital purchases from several different categories .
apple s revenue from digital purchases surpassed $ 20 billion in 2015 @xcite and more than doubled in just three years @xcite .
mobile digital markets offer a wide variety of digital goods , including apps , songs , movies , and digital books .
they also include items that users can buy within an app , called _ in - app purchases _ , such as virtual currencies , bonuses , extra game lives and levels , etc . despite the popularity of the ios market
, there has not been a large - scale study characterizing users spending on different types of content and apps .
for example , it is not known how much is spent on in - app purchases compared to songs . learning how people spend their money in this context
has direct practical implications on the business of several stakeholders , including app developers and managers of online app stores , but it also has important theoretical implications for the understanding of consumer behavior in an emerging market whose dynamics are still poorly known . we study a longitudinal data set extracted from hundreds of millions of email receipts for digital purchases on iphones , ipads , and other ios devices ( which we refer to as `` iphone purchases '' for brevity ) . besides its large scale , our data has two unprecedented advantages over other data collected in the past .
first , besides recording the purchase history , it also includes _ demographic information _ such as people s age , gender , and country of residence .
this enables us not only to characterize the consumer population , but also study how spending differs by income and location .
for example , we find that the average spending is not correlated with income but strongly depends on the country of residence . and second , in contrast to the application - centric view of previous studies , our data is user - centric and allows for the observation of user spending behavior across _
multiple apps_. this allows us to study how users abandon an app and start using a new one , and to compare the behavior of users who make purchases from a single app with users who make purchases from multiple apps simultaneously .
we analyze how users spend money across different categories and show that a small fraction of users are responsible for the majority of spending .
we call these users _ big spenders_. moreover , our analysis informs predictive models of digital spending behavior .
we first show that the time between purchases is best described by a pareto distribution .
we then build a supervised classifier that accurately predicts whether a user will make a purchase from a new app or one with previous purchases .
finally , based on the outcome of the previous step , we predict the exact app from which the user will purchase . in summary ,
our main findings and contributions are as follows : * in - app purchases account for 61% of all money spent on digital purchases on iphones , followed by songs ( 23% ) and app purchases ( 7% ) . * the spending is highly heterogeneous : the top 1% of spenders account for 59% of all money spent on in - app purchases . *
big spenders tend to be 3 - 8 years older , 22% more likely to be a man , and 31% less likely to be from the us , compared to the typical spender .
interestingly , income is comparable between the big spenders and the rest of the users . *
big spenders become slower to repurchase from an app as time passes , but their rate of spending within an app initially increases , then decreases . *
from the perspective of app developers and ad networks , big spenders are the most valuable targeting segment . even if they abandon an app that they are frequently buying from , they are 4.5x more likely to be a big spender in a new app compared to a random user .
* we model the purchasing process in several steps : modeling time between purchases , predicting purchase from a new app or an app from which the user has already purchased , and predicting the exact app of an in - app purchase .
both consumers and producers in the mobile app market might benefit from this study .
our results can inform the deployment of better app recommendation systems so that people download apps they are more likely to enjoy , thus creating higher revenue opportunities for app developers .
shortly after each digital purchase on an iphone ( or any other ios device ) , the user receives a confirmation email with details of the purchase .
this email contains information about the purchase , including the amount of money spent and the type of purchase .
the email has a specific format , making it easy to parse automatically .
we obtained information about digital purchases of yahoo mail users , using an automated pipeline that hashes the names and ids of users to preserve anonymity .
all the analyses were performed in aggregate and on unidentifiable data .
we gathered data covering 15 months , from march 2014 to june 2015 .
our data set includes 26 m users , who together made more than 776 m purchases totaling $ 4.6b .
there are six main categories of iphone purchases : applications ( apps ) , songs , movies , tv shows , books , and in - app purchases ( purchases within an app , e.g. , bonuses or coins in games ) .
these categories differ vastly in numbers of purchasers : 16 m people purchased at least one song , but only 671k people purchased a tv show . the number of purchases by category varies greatly as well : there are 430 m song and 255 m in - app purchases , while movies , books , and tv shows have fewer than 40 m purchases all together .
the total money spent in each category varies even more : in - app purchases account for $ 2.8b , or 61% , of all money spent on digital purchases ; 23% of the money is spent on songs , 7% on app purchases ( purchasing apps themselves , not the purchases within apps ) , 6% on movies , 2% on books , and only 0.7% on tv shows .
even though there are considerably fewer total in - app purchases compared to songs ( 60% fewer ) , the money spent on in - app purchases is 2.7 times more than on songs , showing that on average an in - app purchase is much more expensive than a song purchase .
figure [ fig : bar_all ] shows the number of users , purchases , and the money spent in each of these six categories . our data set also includes user age , gender , and zip code as provided by the users at the time of sign up .
spending varies significantly based on demographics .
figure [ fig : spending_gender ] shows the cumulative distribution function ( cdf ) of the spending for men and women .
men spend more money on purchases than women : the median spending for women is $ 31.1 , and for men it is $ 36.2 , which is 17% higher .
age also affects spending : figure [ fig : spending_age ] shows that the peak age for iphone spenders is the mid-30s , and after that the spending level decreases quickly . for us residents we use the median income of the zip code they declared as their residence as an estimate of their income .
this estimate has shown meaningful trends in earlier work @xcite .
surprisingly , income only affects the spending of people with less than $ 40k annual income ( figure [ fig : spending_income ] ) .
this is in contrast with online shopping , where users with higher income tend to spend more money shopping online @xcite . [
cols="^,^ " , ] & accuracy + app embeddings & 4.7% + nmf & 4.1% + top apps & 2.2% + lda & 1.7% + [ table : baselines_knn ] in the case that the model from section [ subsec : novelty ] predicts that a re - purchase is most likely to happen , we use the frequency and recency of the user s previous app consumption to predict the app from which the re - purchase will occur .
this may appear to be an easy prediction problem , as one might think that users almost always purchase from the last app they purchased from , or the app from which they made the majority of purchases .
however , in case of in - app re - purchases , only 46.5% of re - purchases come from the latest app they purchased from , and only 45.3% of them come from the app from which the user made most of the purchases . this justifies the need for a more involved re - purchase model .
we follow a similar approach to the one from @xcite , and use both recency and popularity of previous apps to predict from which app the user s next in - app purchase will come .
we use a weight function and a time function that maps the frequency of the usage and time since previous usage to the learnt values .
this repeat consumption model could be used for the @xmath4th consumption : @xmath5 in this equation , function @xmath6 represents the frequency of the purchase from the app , and function t represents the time between the purchases .
these functions are optimized jointly by calculating the negative log - likelihood over the equation .
the negative log - likelihood is not convex in @xmath6 and @xmath7 , but is convex in each function when the other one is fixed .
thus , we use a standard gradient descent to maximize the likelihood with respect to @xmath6 and @xmath7 , separately . after learning the weight functions , we are able to predict the correct app from which the user is going to make a purchase with 54.8% accuracy , which is considerably higher than the baselines mentioned above , i.e. 46.5% and 45.3% accuracy by always predicting the latest or the most consumed app , respectively .
online shopping is becoming more popular as people learn to trust online payment systems , which was not the case in the past @xcite .
multiple studies , aimed at profiling online shoppers , found that online shoppers tend to be younger , wealthier , and more educated compared to the average internet user @xcite .
a more recent work showed that while women are more likely to be online shoppers , men spend more money per purchase and make more purchases overall @xcite . in our work , we focus on a particular subset of online purchases , iphone digital purchases .
there are considerable differences in characteristics of iphone purchases and purchases of physical goods .
one of the main differences is that people are much more likely to purchase the same item multiple times .
similar to online shopping , spending on mobile digital goods is increasing , and people have spent more than $ 20 billion dollars in the apple app store in 2015 @xcite , which is four times more per user than in android app stores . this might be due to the different demographics of iphone users .
given this high level of spending , understanding the market would help us to more effectively target apps toward users who are likely to become regular users and frequent spenders . despite the popularity of the iphone digital market , there has not been any large - scale study of how people are spending money on this platform . in this work ,
we show that most of the money is spent on in - app purchases , and we present a demographic and prediction analysis of spending .
usage and purchases from apps have been the subject of a few studies .
sifa et al . studied the purchase decisions in free - to - play mobile games @xcite .
they built a classifier that predicts whether a user is going to make any purchase in the future and also built a regression model to estimate the amount of money that will be spent by each user .
the models are moderately accurate .
schoger studied the monetization of popular apps in the global market , identifying growing markets and that in - app purchases are increasingly accounting for a larger fraction of total purchases @xcite .
our study , unlike those studies , includes the full history of iphone purchases by the users and considers that many users make purchases from multiple apps .
moreover , the large scale of the data set allows us have enough big spenders to analyze their behavior accurately .
we also study changes in user purchases over time , how users becomes frequent buyers in a particular app , and how their purchases evolve over time .
the abandonment of a service is called _ consumer attrition _ or _
churn_. the importance of consumer attrition analysis is driven by the fact that retaining an existing consumer is much less expensive than acquiring a new consumer @xcite .
thus , prediction of consumer churn is of great interest for companies , and has been studied extensively . for example .
, ritcher et al .
exploit the information from users social networks to predict consumer churn in mobile networks @xcite .
braun and schweidel focus on the causes of churn rather than when churn will occur @xcite .
they find that a considerable fraction of churn in the service they studied happens due to reasons outside the companies control , e.g. , the consumer moving to another state . in the context of mobile games , runge et al .
study user churn for two mobile games and predict it using various machine learning algorithms @xcite . they also implement an a / b test and offer players bonuses before the predicted churn .
they find that the bonuses do not result in longer usage or spending by the users .
kloumann et al .
study the usage of apps by users who use a facebook login and model the lifetime of apps using the popularity and sociality of apps , showing that both of these affect the lifetime of the app @xcite .
baeza - yates et al . addressed the problem of predicting the next app the user is going to open through a supervised learning approach @xcite . in our work , we model the whole sequence of purchases that users make , including adoption , churn , and prediction of the next app .
our work is the first work that studies the details of all iphone purchases made by a large number of users .
this allows us to better understand the interplay between usage of multiple apps that are competing for the same users , their attention and their purchasing power .
mobile devices have grown wildly in popularity and people are spending more money purchasing digital products on their devices . to better understand this digital marketplace , we studied a large data set of more than 776 m purchases made on iphones , including songs , apps , and in - app purchases .
we find that , surprisingly , 61% of all the money spent is on in - app purchases , and a small group of users are responsible for most of this spending : the top 1% of users are responsible for 59% of all spending on in - app purchases .
we characterize these users , showing that they are more likely to be men , older , and less likely to be from the us .
then , we focus on how these big spenders start and stop making purchases from apps , finding that as users gradually lose interest , the delay between purchases increases .
the amount of money spent per day on purchases initially increases , then decreases , with a sharp drop before abandonment .
nevertheless , from the perspective of app developers these big spenders are a valuable user segment as they are 4.5x more likely to be a big spender in a new app than a random app user . in the last part of our study
, we model the purchasing behavior of users by breaking it down into three different steps .
first , we model the time between purchases by testing a variety of different distributions , and we find the pareto distribution fits the data most accurately .
second , we take a supervised learning approach to predict whether a user is going to make purchase from a new app .
finally , if the purchase is from a new app , we use a novel approach to predict the new app based on the previous in - app purchases . if the purchase is from an app that the user purchased from in the past , we combine the earlier frequency of the purchases and the time between the purchases to predict from which app the re - purchase will come .
the models proposed in our study can be leveraged by app developers , app stores and ad networks to better target the apps to users . | with mobile shopping surging in popularity , people are spending ever more money on digital purchases through their mobile devices and phones .
however , few large - scale studies of mobile shopping exist . in this paper
we analyze a large data set consisting of more than 776 m digital purchases made on apple mobile devices that include songs , apps , and in - app purchases .
we find that 61% of all the spending is on in - app purchases and that the top 1% of users are responsible for 59% of all the spending .
these big spenders are more likely to be male and older , and less likely to be from the us .
we study how they adopt and abandon individual app , and find that , after an initial phase of increased daily spending , users gradually lose interest : the delay between their purchases increases and the spending decreases with a sharp drop toward the end .
finally , we model the in - app purchasing behavior in multiple steps : @xmath0 we model the time between purchases ; @xmath1 we train a classifier to predict whether the user will make a purchase from a new app or continue purchasing from the existing app ; and @xmath2 based on the outcome of the previous step , we attempt to predict the exact app , new or existing , from which the next purchase will come .
the results yield new insights into spending habits in the mobile digital marketplace . |
with the installation of a 112 node hitachi sr8000-f1 at the leibniz - rechenzentrum in munich the qcdsf collaboration has started simulating qcd with two degenerate flavors of fermions . in a cooperative program with the ukqcd collaboration ( see talk by irving @xcite ) we share configurations giving both collaborations the possibility to do measurements at a larger set of bare parameters .
this contribution has two parts . in the first part we report on performance of qcd calculations on the hitachi sr8000-f1 .
in the second part we present results on the light hadron mass spectrum for dynamical , non - perturbatively @xmath0 improved wilson fermions .
we focus on comparison with our quenched results @xcite .
the sr8000 in munich is a `` cluster '' of 112 symmetric multiprocessor nodes .
each node has 8 cpus and ( at least ) 8 gbyte of shared memory .
each cpu has a a 128 kbyte 4-way associative cache , 160 registers and has a pseudo - vectorization facility .
the peak performance of a cpu is 1.5 gflop / s giving a peak performance of 1344 gflop / s for the whole machine .
it order to get good performance it is essential to employ pseudo - vectorization . at the hardware level pseudo - vectorization is implemented by efficient pre - fetch and pre - load mechanisms . at the programming level the compiler tries to vectorize most inner loops .
these loops are vectorized under the usual conditions , i.e. , if there are no data access conflicts , no function / subroutine calls and no if - blocks in the loop body . on the sr8000
one can basically use the same code as on a standard risc processor and more important one can have the same data layout as for a risc cpu .
for example our data layout is ( in fortran 90 notation ) .... complex(8 ) u(3 , 3 , vol/2 , even : odd , dim ) complex(8 ) a(4 , 3 , vol/2 ) complex(8 ) c(18 , 2 , vol/2 , even : odd ) .... for the gauge , pseudo - fermion and clover field .
in contrast to a real vector computer the vector index does not have to be shifted to the first place . to fix the notation let @xmath1 be the fermion matrix , @xmath2 the wilson hopping term and @xmath3 the clover term : @xmath4 @xmath5 performance figures for the multiplication of @xmath2 , @xmath3 , @xmath6 and @xmath1 with a vector are given in table 1 .
these figures were obtained with fortran 90 code using 64 bit arithmetic on an @xmath7 lattice .
the code performs well in comparison with risc processors where one typically sustains 1020 % of the peak performance .
.performance of the multiplication with a vector on a single cpu .
[ cols="^ , > , > " , ] our code is parallelized with mpi .
in addition we have experimented with a hybrid programming model using openmp for distributing the work over the cpus of a single node and mpi for the communication between nodes .
figure 1 shows the communication loss for the fermion matrix multiplication .
the solid lines ( being identical in both plots ) show performance measurements for a lattice volume of @xmath7 per cpu for the pure mpi code . in the left plot these
are compared with measurements on volumes of @xmath8 per cpu . in the right plot comparison
is done with the hybrid programming model again on a cpu local volume of @xmath7 . as can be seen from figure 1
the communication loss is quite substantial .
the observations on the pure mpi code can be understood qualitatively in the following way . using more and more cpus
the communication pattern becomes more and more complex . within one node when using 2 , 4 or 8 cpus there are 2 , 4 or 6 boundaries which have to be communicated and the surface / volume ratio is 2/8 , 4/8 or 6/8 respectively . between nodes
communication is along 1 , 2 or 3 dimensions ( in our program the lattice can be decomposed maximally in three dimensions ) when using 2 , 4 , or 8 ( and more ) nodes leading again to more and more complicated communication patterns . on a single node parallelization with openmp
is more efficient than with mpi .
this is to be expected because there is no communication needed in that case .
but surprisingly this gain does not persist when switching on inter node communication . in both cases
the same amount of data has to be transferred between the nodes .
but in the pure mpi version more calls are needed to copy data between nodes and in addition data has to be copied within the nodes .
both should lead to larger communication overhead . according to the hardware performance
monitor we are running at about 400450 mflop / s per cpu in our production runs .
this is in accordance with the 433 mflop / s given in figure 1 .
we investigate full qcd with the standard wilson action for the plaquette and the fermions plus the @xmath0 improvement term . for the improvement coefficient
@xmath9 we use the non - perturbative values determined by the alpha collaboration @xcite .
simulation parameters and statistics are listed in table 2 . in these simulations
the @xmath10 ratio lies between 0.60 and 0.83 .
the sommer scale @xcite @xmath11 varies between 4.7 and 5.5 @xcite .
cccr @xmath12 & @xmath13 & @xmath14 & @xmath15 + + 5.20 & 0.1350 & @xmath16 & 6000 + & 0.1355 & @xmath16 & 8000 + 5.26 & 0.1345 & @xmath16 & 4000 + 5.29 & 0.1340 & @xmath16 & 4000 + + 5.29 & 0.1350 & @xmath16 & 2270 + & 0.1355 & @xmath17 & 605 + figure 2 shows the integrated autocorrelation time of the plaquette @xmath18 ( @xmath19 is the normalized autocorrelation function ) for the qcdsf run on the @xmath16 lattice . from the plot we read off that in this case @xmath20 .
figure 3 shows an ape plot . for clarity error bars for the quenched results ( 0.13.5 % )
are not shown .
one sees that for currently reachable sea quark masses unquenching has , if any , only small effects on the light hadron spectrum and that with dynamical fermions it is a long way to @xmath10 ratios as small as in the quenched case . in the following we compare our dynamical results with chiral extrapolations done in the quenched case at @xmath21 where @xmath22
this is within the range of @xmath11 of the dynamical runs studied .
let us first recall the chiral extrapolation of @xmath23 .
our quenched results for @xmath23 are very well described by a phenomenological fit @xmath24 where @xmath25 and @xmath26@xcite .
the lightest three points were excluded from the fit .
the deviation of these points is believed to be a quenching artifact @xcite .
a quenching artifact appearing in quenched chiral perturbation theory for the vector meson mass is a term @xmath27 @xcite .
in figure 4 we plotted our quenched results together with the dynamical ones .
the line represents the fit ( 3 ) . the fit range is much larger than the range of this plot which only includes the six smallest masses ( open symbols ) . towards the chiral limit
we see that the dynamical data ( filled symbols ) lie systematically below the quenched data .
the dynamical results are consistent with the quenched fit .
the corresponding picture for @xmath28 is shown in figure 5 .
the open symbols represent quenched results and the line is a phenomenological fit like ( 3 ) where the curvature @xmath29 is again small .
the filled symbols are dynamical results .
the agreement between quenched and dynamical results is more pronounced than for @xmath23 .
we observe that the results with non - perturbatively @xmath0 improved dynamical fermions scale well .
although we are not able to do a continuum extrapolation yet , we see no large discretization errors . in the considered range of sea quark masses we found no clear evidence for unquenching effects in comparison with previous quenched results
however , our results seem to indicate that the vector meson mass @xmath23 approaches the chiral limit in a different way . this may be another hint , that quenched artifacts are visible in the quenched data .
the computations were done on the hitachi sr8000-f1 at lrz ( munich ) , on the cray t3es at epcc ( edinburgh ) , nic ( jlich ) , ( zib ) berlin and on the ape / quadrics at desy zeuthen .
we thank these institutions for support .
ukqcd acknowledges pparc grants gr / l22744 and ppa / g / s/1998/000777 . 9 a.c .
irving , these proceedings .
qcdsf collaboration , nucl .
b ( proc . suppl . ) 8384 ( 2000 ) 173 .
d. pleiter , ph.d .
thesis , freie universitt berlin , 1999 .
k. jansen and r. sommer , nucl .
phys . b530
( 1998 ) 185 .
r. sommer , nucl .
b411 ( 1994 ) 839 . m. booth , g. chiladze and a.f .
falk , phys .
d55 ( 1997 ) 3092 . | we present results on light hadron masses from simulations of full qcd and report on experiences in running such simulations on a hitachi sr8000-f1 supercomputer . |
there is an ongoing interest for macroscopic quantum processes , especially in the perspective of quantum memories and several groups are currently working on performing storage and retrieval of a quantum state of light into an atomic system .
most quantum memory processes rely on the storage of a quantum signal into an atomic spin coherence , which is free from spontaneous emission .
the optical excitation carrying the information resonantly excites an atomic transition .
a control pulse then converts the optical coherence into a hyperfine long - lived coherence .
another control pulse can eventually change the spin coherence back into an optical coherence and allow the retrieval of the stored signal .
this quantum memory scheme can be achieved in a three - level @xmath0 system , where two spin states are connected to a common upper level by optical transitions ( cf fig .
[ fig : lambda ] ) .
system coupled by two lasers . @xmath2 and @xmath3 are the rabi frequencies that characterize the atom - laser interaction . ]
the mapping of a quantum state of light over a macroscopic atomic ensemble has been demonstrated in various materials such as atomic vapors @xcite or cold atom clouds @xcite .
the full quantum memory scheme has recently been achieved in cold atomic clouds of rubidium @xcite .
however , rare - earth ion - doped crystals also appear as promising candidates for quantum storage applications .
they offer properties similar to atomic vapors with the advantage of no atomic diffusion . at low temperature ( @xmath4 k )
, the optical coherence lifetime can reach several ms , and a hyperfine coherence lifetime has been extended to tens of seconds in pr - doped @xmath5 ( yso ) @xcite . given the absence of atomic motion , extremely long population lifetime can be observed .
electromagnetically induced transparency and `` slow light '' @xcite have been demonstrated in rare - earth ion - doped crystals , along with storage times greater than a second @xcite . among all the rare - earth ions ( lanthanides ) , non - kramers ions with an even number of
@xmath6 electrons exhibit the longest coherence lifetime in the absence of external magnetic field @xcite . with the additional condition on the hyperfine splitting to be of the order of a few tens of mhz in the electronic ground state ,
the rare earth ions are restricted to praseodymium and europium .
both pr@xmath1 and eu@xmath1 ions have been extensively studied over the past 20 years @xcite . however , only dye lasers are available at operation wavelengths in pr@xmath1 and eu@xmath1 doped crystals . for instance , @xmath7 and @xmath8
are respectively operated at 606 nm and 580 nm . alternatively
, other non - kramers rare - earth ions such as thulium fall within reach of more tractable lasers .
@xmath9 ( yag ) has been widely studied in the field of coherent transient - based signal - processing schemes @xcite and lasers @xcite .
it is especially attractive because the @xmath10 transition at 793 nm is in the range of semi - conductor lasers which can be stabilized to better than 1 khz @xcite .
although the only isotope of thulium @xmath11tm ions has a @xmath12 nuclear spin , it does not exhibit hyperfine structure at zero magnetic field because of @xmath13 quenching @xcite .
lifting the nuclear spin degeneracy with an external magnetic field offers a way to build a @xmath0-system provided the magnetic field induces different spin - state mixing in ground and excited electronic states . in an earlier paper @xcite we showed that this was the case for a specific magnetic field orientation with respect to the crystalline axes .
we determined the field direction that gives maximum branching ratio between the two optical transition probabilities in the @xmath0 system thus created .
in addition , a lower bound to the branching ratio was derived from the experimental measurement of the gyromagnetic factors
. further characterization of this @xmath0 system includes direct experimental measurement of the optimal branching ratio .
this is the purpose of this study .
the paper is arranged as follows . in sec .
[ lambda ] we summarize the previous results concerning the building of a @xmath0 system in @xmath14 . in sec .
[ exp ] the experimental setup is described . in sec .
[ shb ] spectral hole - burning is presented and discussed in the case of @xmath14 , and we measure the population lifetime of the ground sublevels . in sec . [ branching ] we show that each transition can be individually excited and we measure their relative strengths by optical nutation and two - pulse photon echoes . from these experiments
we derive an experimental value of the branching ratio .
this section is a brief reminder of the discussion presented in a previous paper @xcite .
let us consider the optical transition at 793 nm connecting the fundamental levels of the stark multiplets @xmath15 and @xmath16 of tm@xmath1 in yag .
application of an external magnetic field lifts the nuclear spin degeneracy by splitting the electronic levels . as shown in fig .
[ fig : rule ] , a three - level @xmath0 system would involve the two hyperfine sublevels of the ground state @xmath17 and one hyperfine sublevel of the excited state @xmath18 .
if the interaction with the magnetic field is restricted to the zeeman effect , the electronic levels split into @xmath19 and @xmath20 spin levels .
one of the two optical transitions of the @xmath0 is forbidden .
indeed the selection rule @xmath21 forbids any electronic excitation to flip the nuclear spin .
fortunately , the coupling effect of the zeeman electronic and the hyperfine interactions mixes the nuclear spin states and may lead to comparable transition strengths for the two legs of the @xmath0 if the magnetic field is suitably oriented . as a consequence , spin - flipping transitions @xmath22 and @xmath23
are no longer forbidden .
the atomic states can be represented as the product of an electronic state and a nuclear spin state .
the ground level nuclear spin states are derived from the excited level nuclear spin states by a unitary transformation . therefore the 4 transition dipole moments satisfy : @xmath24 the weak ( respectively , strong ) transition dipole moments are referred to as @xmath25 ( respectively , @xmath26 ) .
the branching ratio @xmath27 is the ratio of the transition probabilities along the two legs of the @xmath0 : @xmath28 in the yag crystal , @xmath29 ions substitute for yttrium ions in 6 crystallographically equivalent but orientationally inequivalent sites depicted in fig .
[ fig : sites ] . because of the site symmetry ( @xmath30 ) , the optical transition dipole moment is oriented along the @xmath31 local axis of each site .
let the light beam propagate along @xmath32 $ ] .
two linear polarization directions are considered .
is represented by an oblong and is directed along the local @xmath31 axis . here
the local frame @xmath33 is given for site 1 .
the direction of the magnetic field @xmath34 is defined by angle @xmath35.,width=302 ] when it is directed along @xmath36 $ ] , the optical electric field @xmath37 is perpendicular to the electric dipoles @xmath38 of sites 2 , 4 and 6 .
the only excited sites are then 1 , 3 and 5 .
they interact with the light beam with the same rabi frequency @xmath39 and therefore contribute the same amount to the crystal optical density
. alternatively , when the light beam is polarized along direction @xmath40 $ ] , the only excited sites are 1 , 4 and 6 .
their rabi frequencies are equal .
for a given magnetic field orientation , the field coordinates are specific to each site local frame .
since the gyromagnetic tensor is strongly anisotropic in both electronic levels , the 6 sites exhibit different hyperfine splittings : @xmath41 where @xmath42 and @xmath43 denote the ground and excited electronic states , and where the gyromagnetic tensor @xmath44 and the magnetic field @xmath34 coordinates are given in the local frame .
let @xmath34 be applied in the bisector plane defined by crystalline axes @xmath45 $ ] and @xmath46 $ ] .
the magnetic field direction is defined by angle @xmath35 formed with direction @xmath45 $ ] , as illustrated in fig .
[ fig : sites ] . for any magnetic field orientation within the bisector plane , ions in sites 3 and 5 experience the same field .
hence they have identical zeeman splittings and equal branching ratios .
similarly , ions in sites 4 and 6 experience the same magnetic field and exhibit equal branching ratio .
the branching ratio for ions in sites 3 and 5 was calculated from theoretical gyromagnetic factors for any magnetic field orientation @xcite .
when the magnetic field is in the bisector plane , fig .
[ fig : br ] shows that the theoretically predicted branching ratio goes through a maximum @xmath47 . besides , fig .
[ fig : br ] shows that the accurate orientation of the magnetic field is critical around the maximum branching ratio orientation . the experimental measurement of the gyromagnetic tensor coefficients @xcite made it possible to determine a lower boundary for the maximum branching ratio : @xmath48 these measurements also yielded the magnetic field orientation that gives maximum branching ratio for ions in sites 3 and 5 : @xmath49 .
this is rather close to direction @xmath40 $ ] ( corresponding to @xmath50 ) .
the system is excited and probed with an extended cavity semiconductor laser operating at 793 nm .
it is stabilized on a high - finesse fabry - perot cavity through a pound - drever - hall servo - loop @xcite .
it is amplified with a toptica boosta laser amplifier .
a single - mode fiber spatially filters the beam . at the end of the fiber one
measures up to 100 mw output power , depending on the amplifier adjustment .
the @xmath51 at.% @xmath14 crystal is 5 mm thick .
it is cut perpendicular to direction @xmath32 $ ] along which the laser beam propagates ( cf fig .
[ fig : sites ] ) .
we use a commercial oxford 6 t spectromag sm4 cryostat .
the adjustable magnetic field is generated by superconducting coils immersed in liquid helium .
the sample is cooled down to @xmath52 k. fig .
[ fig : setup ] shows the optical setup after the fiber .
pulse sequences are shaped by an arbitrary waveform generator ( tektronix awg 520 ) that monitors two external acousto - optic shifters ( ao 1 and ao 2 ) .
the acousto - optic shifters are imaged on the sample ( s ) so that the beam does not move across the sample when the frequency is varied .
the spot size on the crystal is adjusted to 100 @xmath53 m , except in the experiments described in sec .
[ sub ] , where the spot size is @xmath54 m .
the transmitted light is then collected on an avalanche photodiode ( apd ) hamamatsu c5460 .
an acousto - optic modulator ( ao 3 ) protects the detector from possible strong burning pulses .
in thulium , the optical transition at 793 nm exhibits a very large inhomogeneous broadening @xmath55 ghz .
application of a magnetic field splits each electronic level into a pair of hyperfine sublevels with a zeeman splitting of the order of @xmath56 to @xmath57 mhz / t , depending on the magnetic field orientation with respect to the local site axes .
a field lower than 1 tesla does not modify the absorption spectrum since the inhomogeneous linewidth is much larger than the zeeman splitting .
because of inhomogeneous broadening , a laser excitation at given frequency @xmath58 simultaneously excites 4 different classes of ions , as illustrated in fig .
[ fig:4niv ] .
each class refers to a specific ground and excited state sublevels that are connected by the incoming field .
initially , in each ion class both ground sublevels are equally populated .
this is even true at working temperature ( @xmath59 k ) , since the level splitting is much smaller than @xmath60 ghz . when the sample is exposed to sustained incident light at @xmath58 , it experiences optical pumping that tends to unbalance the ground level state distribution .
ions in the resonantly excited sublevel are transferred to the off - resonant sublevel ( cf fig .
[ fig:4niv ] ) .
this gives rise to a hole in the absorption spectrum at frequency @xmath58 .
besides , since we consider 4-level systems , additional features appear .
two side holes are observed at @xmath61 , where the probe field excites transitions from the depleted ground sublevel to the other excited state sublevel .
conversely , transitions from an overpopulated ground sublevel to either excited sublevel give rise to antiholes in the absorption spectrum .
such antiholes are observed at @xmath62 and @xmath63 .
depending on the sublevels they connect , the various transitions excited by the probe beam can be sorted as weak or strong ones .
it should be noted that at least one strong transition contributes to any hole and antihole , except for the antiholes located at @xmath64 .
these antiholes arise from a probing process along a weak transition in the ion classes ( ii ) and ( iv ) of fig .
[ fig:4niv ] . as a result
, the observation of an antihole at this position shall bear evidence that the nuclear spin flip selection rule is actually relaxed . .
solid lines : strong transitions .
dashed lines : weak transitions .
ions are depicted by open circles . ]
we perform spectral hole - burning in tm@xmath1:yag .
a 0.19 t magnetic field is oriented in the direction that is expected to give maximum branching ratio for ions in sites 3 and 5 ( @xmath65 , cf sec .
[ lambda ] ) .
the laser beam is linearly polarized in direction @xmath36 $ ] so that only sites 1 , 3 and 5 are excited .
the burning step consists in a series of ten @xmath66s pulses every 10 ms .
then the sample is probed with a weak 1 ms long pulse chirped over a @xmath67 mhz interval .
the burning and probing sequence is repeated every 150 ms .
transmitted probe intensity is collected on a photodiode , averaged over 8 shots and plotted in fig .
[ fig : spectre ] .
the holes and antiholes are ascribed to sites 3 and 5 .
we measured the ground and excited splittings for these two sites in this field direction : @xmath68 where @xmath69 is the magnetic field amplitude .
this agrees with the calculated splittings derived from the experimental gyromagnetic tensor components @xcite : @xmath70 mhz / t and @xmath71 mhz / t .
the ground level gyromagnetic tensor values are close to those measured by schmidt in tmag by nuclear magnetic resonance @xcite .
one can see in fig .
[ fig : spectre ] that the antiholes are generally wider than the holes .
this suggests that the inhomogeneous broadening of the zeeman splitting is larger in the ground state than in the excited state . under a @xmath72 t magnetic field
the spectral holes and antiholes correspond to site 3 and 5 ions .
the features corresponding to site 1 are out of the probing spectral window.,width=325 ] the hole - burning features associated with site 1 can not be seen in fig .
[ fig : spectre ] because the zeeman splitting for site 1 is over 60 mhz / t for this magnetic field orientation .
the hole - burning spectrum is erased by return to thermal equilibrium . to measure the corresponding decay rate , we measure the central hole depth as a function of delay between pumping and probing steps .
the sample is cooled down to @xmath73 k. a @xmath74 t magnetic field is applied in direction @xmath75 , very close to @xmath40 $ ] .
it should be noted that this is not the optimal direction discussed in sec .
[ lambda ] .
the light beam is polarized in direction @xmath40 $ ] so that only sites 1 , 4 and 6 are excited , with identical rabi frequency .
in addition , since they are magnetically equivalent , their zeeman splittings are equal ( @xmath76 mhz and @xmath77 mhz ) and they exhibit the same population lifetime .
[ fig : pop ] shows that the central hole does not decay for 10 seconds following excitation . in other words ,
the population lifetime relative to sites 1 , 4 and 6 for this magnetic field orientation and strength is estimated to a few minutes .
the light beam is then polarized along @xmath36 $ ] so as to excite sites 1 , 3 and 5 with identical rabi frequency .
since we are interested in measuring the sublevel population lifetime for sites 3 and 5 only , we isolate ions in sites 3 and 5 by means of a bleaching process : a strong beam chirped over a 25 mhz interval optically pumps a large spectrum range of ions in site 1 to the non - resonant very long lifetime ground sublevel .
site 3 and 5 ions also undergo optical pumping but the chirp amplitude is much larger than the ground sublevel splitting @xmath78 mhz . as a consequence
, a large fraction of ions in sites 3 and 5 is pumped back to the initial sublevel .
three seconds after this bleaching process , we perform ordinary spectral hole - burning and observe the decay of the central hole depth by varying the delay between burning and probing ( cf fig .
[ fig : pop ] ) .
an exponential fit gives a population lifetime of @xmath79 s for sites 3 and 5 .
$ ] .,width=325 ]
therefore in crystallographically equivalent sites 1,3 and 5 , the sublevel population lifetime apparently increases with the ground state splitting . at present
we can not explain this behavior . in contradiction with our measurements , n. ohlsson _
et al . _ observed longer lifetimes for smaller splittings in a given site @xcite . since the spin level thermalization does not involve optical transitions , we expect the corresponding decay rate to be insensitive to a small angle rotation of the magnetic field
. indeed , when @xmath34 orientation is varied from @xmath75 to @xmath65 , the branching ratio of the optical transitions strongly increases ( cf fig .
[ fig : br ] ) , but the level splitting @xmath80 varies little , as shown in table [ tab : deltag ] .
the decay rate is expected to be mainly sensitive to this level splitting .
.calculated values for normalized ground splitting @xmath81 in different sites , for two specific field orientations .
[ cols="^,^,^",options="header " , ] before sites 3 and 5 can be coherently manipulated for quantum storage , it is necessary to prepare site 1 ions so that they do not interfere .
their very long population lifetime guarantees that a bleaching process can store them for a long time in a non - resonant ground sublevel . as for sites 3 and 5 ,
a spin level lifetime of a few seconds is long enough for coherent manipulation .
in order to precisely compare the optical transitions along the two legs of the @xmath0 , we resort to coherent processes along with frequency selective optical pumping .
this enables us to isolate the weak optical transition contribution . a two - level system driven by a resonant monochromatic field @xmath82 oscillates between its ground and excited states at rabi frequency @xmath83 .
along with this population oscillation , the system emits an induced radiation that modulates the driving field at frequency @xmath84 .
this gives a direct and robust way to measure the rabi frequency of a transition even when the inhomogeneous broadening is infinite and the beam spatial profile is not uniform : for an infinitely inhomogeneously broadened medium driven by a stepfunction gaussian laser beam at @xmath85 , the transmitted signal takes the form of damped oscillations @xcite : @xmath86 where @xmath87 is the incoming beam intensity , @xmath88 is the optical density of the sample , @xmath84 is the rabi frequency of the ion - laser interaction in the center of the beam and @xmath89 is the bessel function of order 1 .
this expression is valid if the optical density of the sample is small ( @xmath90 ) .
according to eq .
[ eq : nut ] , the first maximum of @xmath91 is located at @xmath92 , therefore the rabi frequency can be directly obtained from the nutation signal .
the optical density @xmath88 can be assessed by comparing the transmission of the material at @xmath85 with its transmission after a long exposure to laser light : @xmath93 this way , one can measure the optical density of the material with a single nutation temporal profile .
besides , the optical density is proportional to @xmath94 and to the number of active atoms . if optical nutation is performed on each transition , then the optical density measurement gives direct access to the relative dipole moments of the weak and strong transitions .
we first observe optical nutation on the two - level system formed by @xmath14 at zero magnetic field .
the transmitted signal is displayed in fig .
[ fig : nut0 ] .
the sample is cooled down to @xmath95 k and is excited with a @xmath96s resonant light pulse polarized along @xmath36 $ ] . for this light polarization , sites 1 , 3 and 5 participate in absorption with equal rabi frequency and each one of them contributes one third to the total optical density @xmath97 . .
the signal is fitted according to eq .
[ eq : nut ] with @xmath98 mhz and @xmath99 ( solid line).,width=325 ] let us apply a magnetic field in the direction that is expected to give maximum branching ratio for sites 3 and 5 ( cf sec .
[ lambda ] ) .
application of a magnetic field splits the electronic levels , the ions being equally distributed in both ground sublevels . as mentioned in sec .
[ hb ] , this does not change the absorption profile of the sample , so the nutation signal is not modified .
we now apply @xmath100s pulses .
the sample temperature is decreased gradually from 5 k to 2 k. as long as thermalization processes dominate , the system returns to thermal equilibrium between successive pulses .
as the temperature goes down , thermalization gets slower .
ultimately , resonant sublevel depletion by optical pumping supersedes repopulation by relaxation processes ( see fig .
[ fig:4niv ] ) and the nutation signal disappears .
the repopulation process can be assisted by optical repumping .
if one tunes the repumping laser to @xmath101 so as to counterbalance optical pumping at @xmath58 , only ions in sites 3 and 5 from classes ( i ) and ( ii ) are pumped back into the initial ground sublevel [ see fig .
[ fig : repump](a ) ] .
all ions in site 1 , as well as ions in sites 3 and 5 with excitation schemes ( iii ) and ( iv ) , stay off resonance with the repumping beam . in a given class of ions ,
the steady - state fraction of ions in the resonant sublevel reads as @xmath102 where @xmath103 , @xmath104 and @xmath105 respectively stand for the pumping rate , the repumping rate and the temperature dependent relaxation rate . in the absence of optical excitation
this fraction reduces to @xmath106 , which means that the ground state population is evenly shared between the two sublevels . when @xmath107 and @xmath108 , all ions are pumped to the off - resonance sublevel and @xmath109 .
when @xmath110 the relative population reduces to @xmath111 . optical pumping consists in repeated excitation and relaxation processes .
excitation is carried out by a series of 10 @xmath53s pulses .
within this 10 @xmath53s pulse duration the transition at @xmath58 is bleached , resulting in evenly populated ground and upper levels . as will be shown later in this section ,
this is true even when the excitation takes place along the weak transition .
therefore the excitation process brings the system in the same final state , whatever excitation path is followed .
relaxation to the ground state strongly differs from spontaneous emission decay . indeed , non - radiative processes dominate , proceeding through the intermediate states @xmath112 and @xmath113 .
we assume here that these stepwise relaxation processes relax the nuclear spin selection rule . to sum up
, neither excitation nor relaxation depends on the transition strengths , and the same value of @xmath114 is used for classes ( i ) and ( ii ) .
the electronic transition dipole moment does not depart from its zero magnetic field value . since only 2 sites out of 3 participate in the absorption , the optical density at stake here reads as @xmath115 where @xmath116 is the optical density at zero magnetic field .
similarly , if the repumping laser is tuned to @xmath117 [ cf fig .
[ fig : repump](b ) ] , only ions from system ( iv ) are repumped to the resonant sublevel : the nutation signal is ascribed to ions in sites 3 and 5 resonating on a weak transition . the optical density is expected to be much smaller than @xmath116 and reads as @xmath118 finally , if @xmath119 , only ions from system ( iii ) are repumped to the resonant sublevel : the nutation signal is ascribed to ions in sites 3 and 5 resonating on a strong transition .
the optical density reads as @xmath120 .
the repumping frequencies are displayed as thick solid ( respectively , dashed ) arrows along strong ( respectively , weak ) transitions .
( a ) @xmath101 : only systems ( i ) and ( ii ) participate in the nutation signal , along a strong transition for system ( i ) and along a weak transition for system ( ii ) .
( b ) @xmath117 : only ions in system ( iv ) participate in the nutation signal , along a weak transition . ]
the combination of optical pumping and repumping processes makes it possible to isolate a nutation signal related either to a weak or to a strong transition . the transmitted intensity for various repumping frequency values
is displayed in fig .
[ fig : nut ] .
the repetition rate is 50 s@xmath121 , the sample temperature is @xmath59 k and the incident intensity is the same whatever the repumping frequency value .
the optical density of the sample under zero magnetic field is @xmath122 according to eq .
[ eq : contrast ] applied to fig .
[ fig : nut](a ) .
when a 0.3 t magnetic field is applied , the sample becomes transparent : the damped oscillations disappear and the transmitted intensity is almost constant [ cf fig .
[ fig : nut](d ) ] .
the repumping process consists in a series of ten @xmath123s chirps over 1 mhz around the repumping frequency @xmath124 .
it is applied between successive excitation pulses . in fig .
[ fig : nut](b ) we plotted the transmitted intensity when a repumping beam is applied at frequency @xmath125 .
the optical density is measured on the nutation signal : @xmath126 .
[ eq : dsw ] gives the relative population of the resonant sublevel in systems ( i ) and ( ii ) : @xmath127 . besides , in both cases ( a ) and ( b ) of fig .
[ fig : nut ] , the nutation signal is associated with exciting strong transitions as well as weak transitions . this is why the rabi frequency and hence the position of the first maximum in the nutation signal do not change . finally , in fig .
[ fig : nut](c ) we plotted the transmitted intensity for a repumping beam at @xmath128 .
the optical density of the sample excited only along a weak transition is very small : @xmath129 .
no damped oscillations are visible , but the transmitted intensity slowly increases with time during excitation .
the experimental data clearly show that the position of the first maximum in nutation signal ( c ) is different from cases ( a ) and ( b ) , even though it can not be accurately located .
this proves that signal ( c ) does not come from a residual signal emitted along a strong transition .
, c : with magnetic field and repumping at frequency @xmath130 , d : with magnetic field but no repumping beam .
the dashed traces are theoretical curves given by eq .
[ eq : nut ] and fitted to the experimental data .
the graphs are vertically offset for clarity.,width=325 ] according to eqs .
[ eq : dsw ] and [ eq : dw ] , the ratio of @xmath131 and @xmath132 finally leads to : @xmath133 which is consistent with the lower boundary given in eq .
[ eq : lower ] derived from gyromagnetic tensor measurements @xcite .
the position of the first maximum of nutation profile for the weak transition is @xmath134 times the corresponding position for the strong transition .
according to the measured @xmath27 value , this maximum occurs at @xmath135s .
therefore , as previously assumed , saturation is reached within the @xmath100s pulse duration even for the weak transition . in this section ,
two - pulse photon echoes are achieved separately on each leg of the @xmath0 .
application of a magnetic field splits the electronic levels , and repeated excitation achieves optical pumping between the two ground sublevels . therefore the echo is present as long as the temperature is high enough so that relaxation compensates for optical pumping .
[ fig : echo ] shows a two - pulse photon echo sequence in tm : yag under 0.3 t magnetic field , with a 50 s@xmath121 repetition rate .
the two pulses are gaussian and have equal intensity . at @xmath136 k
, thermalization still dominates the optical pumping caused by repeated excitation . in each two - pulse sequence ,
the first pulse partly bleaches the optical transition , so the second pulse is less absorbed than the first one . as the sample
is cooled down , optical pumping becomes dominant . at @xmath137
k , the resonant level is totally empty and the photon echo signal disappears .
the excitation pulses propagate through a transparent medium where there is no bleaching process , which is why their transmitted intensities are equal . under magnetic field with two gaussian pulses at @xmath136
k ( solid line ) and @xmath137 k ( dashed line ) . the detection channel is gated by an acousto - optic modulator ( ao 3 in fig .
[ fig : setup ] ) which attenuates the excitation pulses ( grey area ) .
this aom opens a few @xmath53s after the second pulse to let the photon echo through .
the echo disappears when optical pumping dominates thermalization.,width=325 ] we need to perform the same repumping process as described in sec .
[ nutation ] in order to recover a signal and ascribe it to a weak or strong transition .
when the repumping frequency is tuned to @xmath138 , ions in system ( i ) or ( iii ) in fig .
[ fig : repump ] are repumped and the photon echo is emitted along the strong transition at frequency @xmath58 .
when the repumping frequency is tuned to @xmath64 , the photon echo is emitted along the weak transition at frequency @xmath58 in ions in system ( ii ) or ( iv ) . to make sure that the detected signal is imputable to the weak transition ,
we slightly detune the repumping beam : the echo intensity goes down , which proves that the off - resonant repumping is less efficient and that the signal does not come from a residual signal emitted along a strong transition . the photon echo amplitude can be expressed as a product of the transition dipole moment and a function of the rabi frequency .
this applies to both transitions , either strong or weak . since the rabi frequency is proportional to the transition dipole
moment and to the applied field , the photon echo intensity along the strong transition reads as : @xmath139 where @xmath140 represents the laser intensity and where the shape of function @xmath42 need not be detailed any further .
conversely , when the weak transition is excited with the same intensity @xmath140 , the photon echo intensity reads as @xmath141 where @xmath27 is the branching ratio . the photon echo intensity @xmath142 on the strong transition is measured for different excitation intensities .
a weighted echo intensity is defined as : @xmath143 this function is interpolated with a polynomial function as shown in fig .
[ fig : is ] . if we measure @xmath144 on the weak transition for a given intensity @xmath145 , we are able to derive the branching ratio value by solving the following equation : @xmath146 as shown in fig . [
fig : is ] and according to eqs .
[ eq : w ] and [ eq : implicit ] , the graphical solution of this equation is given by the intersection point of @xmath147 and @xmath148 .
the @xmath149-coordinate of this intersection point is the branching ratio @xmath27 : @xmath150 this is consistent too with the expectations derived from gyromagnetic factor measurements @xcite . .
a graphical solution is given by the intersection point of @xmath151 and @xmath148 .
the value is directly read out as @xmath152 at the intersection.,width=313 ] with optical nutation and photon echo experiments we measured two branching ratio values that are in agreement with one another .
as far as theoretical gyromagnetic tensors are concerned , the @xmath153 anisotropy is much larger in the ground state than in the excited state @xcite .
this difference , known as _ anisotropy disparity _ , is at the origin of the high value for the maximum theoretical branching ratio @xmath47 .
the experimental value of anisotropy disparity appeared to be smaller than predicted by theory @xcite .
as a result , a branching ratio value close to its lower boundary @xmath154 was expected .
the direct measurements presented in this paper confirm this result .
these measurements have been achieved for the magnetic field orientation ( @xmath155 ) given by de seze _ et al . _
we did not study the dependence of @xmath27 as a function of @xmath35 .
however , the branching ratio is almost constant over a @xmath156 wide interval @xcite
. therefore , even with a @xmath157 mechanical precision on the angle , one can reasonably consider that the measured branching ratio is close to its maximum . finally , in the experiments presented in sec .
[ nutation ] , the nutation signal along the strong transition goes through its first maximum at @xmath158s .
this means that a @xmath159-pulse along the strong transition would take @xmath160s .
according to the present branching ratio measurements , a @xmath159-pulse along the weak transition would then take @xmath161s
. therefore , we are able to excite each leg of the @xmath0 system with a @xmath159-pulse , which is a capital asset in the perspective of coherent processes dynamics .
we directly measure the transition probability ratio along the two legs of the @xmath0 through coherent processes along with frequency selective optical pumping .
this is performed in the optimum magnetic field orientation derived from a previous experiment @xcite .
the direct measurement of the branching ratio is consistent with the value derived from the gyromagnetic tensor components and confirms that a @xmath0-system can operate efficiently in tm : yag .
this way we have cross - checked and supported the theoretical description , complementing prior experimental data @xcite with a new set of data different in nature .
in addition , this provides us with the branching ratio value we have to know in order to efficiently build and control nuclear magnetic superposition states with optical excitation .
side investigation on the sublevel lifetime leads to unexpected results .
indeed , broadly spaced sublevels appear to be more stable than closely spaced ones , in contradiction with previously reported experiments @xcite .
this has to be clarified .
still , the measured lifetime is consistent with long - time coherent manipulation of spin states .
the demonstrated existence of the @xmath0 system promises interesting spin state properties in either ground or excited level .
next step will deal with the investigation of these coherent spin features .
pap1 b. julsgaard , a. kozhekin and e. s. polzik , nature ( london ) * 413 * , 400 ( 2001 ) .
a. kuzmich , w. p. bowen , a. d. boozer , a. boca , c. w. chou , l .- m .
duan and h. j. kimble , nature ( london ) * 423 * , 731 ( 2003 ) .
t. chanelire , d. n. matsukevich , s. d. jenkins , s .- y .
lan , t. a. b. kennedy and a. kuzmich , nature ( london ) * 438 * , 833 ( 2005 ) .
e. fraval , m. j. sellars and j. j. longdell , phys .
lett . * 92 * , 077601 ( 2004 ) .
a. v. turukhin , v. s. sudarshanam , m. s. shahriar , j. a. musser , b. s. ham and p. r. hemmer , phys .
88 * , 023602 ( 2002 ) .
j. j. longdell , e. fraval , m. j. sellars and n. b. manson , phys . rev . lett . * 95 * , 063601 ( 2005 ) .
r. m. macfarlane , j. lumin .
* 100 * , 1 ( 2002 ) .
j. mlynek , n. c. wong , r. g. devoe , e. s. kintzer and r. g. brewer , phys .
* 50 * , 993 ( 1983 ) .
l. e. erickson , phys .
b * 39 * , 6342 ( 1989 ) .
j. j. longdell , m. j. sellars and n. b. manson , phys . rev .
b * 66 * , 035101 ( 2002 ) .
r. klieber , a. michalowski , r. neuhaus and d. suter , phys .
b * 67 * , 184103 ( 2003 )
. w. r. babbitt , a. lezama and t. w. mossberg , phys .
b * 39 * , 1987 ( 1989 )
. s. krll , e. y. xu , m. k. kim , m. mitsunaga and r. kachru , phys .
b * 41 * , 11568 ( 1990 ) .
r. w. equall , y. sun , r. l. cone and r. m. macfarlane , phys .
. lett . * 72 * , 2179 ( 1994 ) . k. d. merkel and w. r. babbitt , opt . lett . * 23 * , 528 ( 1998 ) .
h. lin , t. wang and t. w. mossberg , opt .
lett . * 20 * , 1658 ( 1995 ) .
v. lavielle , i. lorger , j .- l .
le gout , s. tonda and d. dolfi , opt . lett .
* 28 * , 384 ( 2003 ) .
g. gorju , v. crozatier , i. lorger , j .- l .
le gout and f. bretenaker , ieee photon . tech
* 17 * , 2385 ( 2005 ) .
v. crozatier , f. de seze , l. haals , f. bretenaker , i. lorger and j .- l
. le gout , opt . comm . * 241 * , 203 ( 2004 ) . a. abragam and b. bleaney , _
electronic paramagnetic resonance of transition ions _ ( clarendon press , oxford , 1970 ) , chap . 15 . f. de seze , a. louchet , v. crozatier , i. lorger , f. bretenaker , j .- l .
le gout , o. guillot - nol and ph .
goldner , phys .
b * 73 * , 085112 ( 2006 ) .
o. guillot - nol , ph .
goldner , e. antic - fidancev and j .- l .
le gout , phys .
b * 71 * , 174409 ( 2005 ) . v.
h. schmidt and e. d. jones , phys .
b * 1 * , 1978 ( 1970 ) .
n. ohlsson , m. nilsson , s. krll and r. krishna mohan , opt . lett . * 28 * , 450 ( 2003 ) .
y. sun , g. m. wang , r. l. cone , r. w. equall and m. j. m. leask , phys .
b * 62 * , 15443 ( 2000 ) . e. l. hahn , phys . rev . * 80 * , 580 ( 1950 ) . | a three - level @xmath0 system in tm@xmath1 doped yag crystal is experimentally investigated in the prospect of quantum information processing .
zeeman effect is used to lift the nuclear spin degeneracy of this ion . in a previous paper [ de seze _ et al .
_ phys .
rev .
b , * 73 * , 85112 ( 2006 ) ] we measured the gyromagnetic tensor components and concluded that adequate magnetic field orientation could optimize the optical connection of both ground state sublevels to each one of the excited state sublevels , thus generating @xmath0 systems . here
we report on the direct measurement of the transition probability ratio along the two legs of the lambda .
measurement techniques combine frequency selective optical pumping with optical nutation or photon echo processes . |
calculating the energy states for the motion of a particle in a ` simple ' one dimensional infinite square well finds its way into all standard quantum mechanics texts .
but because the potential representing the walls is sharp and infinite the model has subtleties . for instance
a consistent definition of a momentum operator for it is challenging @xcite .
another intriguing property of the infinite well is that for certain initial states the time development of the spatial probability density can show fractal behavior @xcite .
there is also the topic of wave packet revival in such a well @xcite . and
the time dependence of a particle s wave function in a suddenly expanded well can show interesting features @xcite .
clearly , when sharp edges and infinities are involved , care is important . in this paper
we compound the singularity by adding within the well a fixed delta function potential of arbitrary sign , strength and location . for this model , due to the containing walls , all solutions to the schrdinger equation are bound states , but the model does have its niceties .
we find for instance that although , as expected for the bound states of any one - dimensional system , all of its energy eigen - states are non - degenerate , should the delta function be located at a rational fraction of the well s width , then a subset of these energies approach _ in the limit _ of strong attraction or repulsion , double degeneracy .
in addition to examining several other properties and limits of the model we have computed for both attraction and repulsion the ground state energy of the system as a function of the location of the delta function . for the case of attraction
this energy has a minimum when the force center is located at the center of the well . for a wide box
this minimum is nearly flat , but it becomes sharper when the well width is reduced to sizes of order of the spatial decay length characterizing the ` molecular ' size of the bound state of an attractive delta function in free space .
the model has not , of course , escaped attention .
patil @xcite and atkinson and crater @xcite give analyses for the case that the delta function potential is located precisely at the center of the well .
bera and his co - workers @xcite place it anywhere within the well and utilize perturbation theory .
joglekar @xcite computes the energy eigenvalues when it is located at irrational multiples of the well s width and considers weak and strong coupling limits .
epstein @xcite uses the well with the delta function located at the center in order to gain insight into the interesting question posed by wigner @xcite as to whether the energy levels of a hydrogen atom should be derivable from second order perturbation theory in view of the fact that it is known to be proportional to the fourth power of the electron charge and therefore to the square of the coulomb interaction .
we were motivated to consider the model by the long standing problem of the effects of containment on molecule states and energies dating back to work by michels @xcite , sommerfeld @xcite and their co - workers , in the early heyday of quantum mechanics , applying approximations to study what might be said to be the canonical problem that of a hydrogen atom with its nucleus fixed at the center of an impenetrable sphere .
other more recent papers discuss this model with various methods of approximation .
see for instance references @xcite and , especially @xcite which includes a short review of work on this model up to 1987 .
more recently changa and co - workers @xcite have applied perturbation theory to estimate the energies when the hydrogen nucleus is _ shifted _ from the center of the the spherical box .
the particular advantage of the model we consider here is that no approximations are necessary to discuss the energy levels and their dependence either on the well width or on the location of the force center
. it would be gratifying to be able to do the same with other model interactions .
for instance one might try a ` one dimensional coulomb ' repulsion or attraction , @xmath0 , @xmath1 .
but it is a sobering fact that even with no walls present , so that @xmath2 is any real number bar zero , there seems to be no final agreement for that model as to the available energy levels and corresponding eigenstates @xcite .
the paper is arranged as follows .
we give the general solution in sec .
[ solution ] .
its nature depends upon whether the ratio @xmath3 is a rational or irrational number , where @xmath4 is the delta function s position within a well of width @xmath5 . when that ratio is irrational all solutions are straightforward ` ordinary ' solutions , but when it is rational a complete set of solutions includes both the ordinary ones together with those we choose to call ` nodal ' solutions . for both cases we confirm in sec .
[ orthog ] mutual orthogonality between all solutions corresponding to different energies . in secs .
[ weak ] and [ strong ] we consider the limits of weak and strong coupling for any location of the delta function potential within the well . in taking these limits ,
when @xmath3 is rational care must be exercised to include both the nodal and ordinary solutions . in sec .
[ 1/2 ] we discuss solutions for the special symmetric case that the delta function is located precisely in the middle of the well , for which @xmath3 is the rational number @xmath6 . in sec .
[ gs ] we consider in some detail the ground state of the model , especially with respect to its energy dependence on the relative location @xmath3 , for various signs and strengths of the delta function potential .
this is plotted in fig .
[ single ] .
a particle of mass @xmath7 free to move within a one - dimensional well with sides at @xmath8 and @xmath9 has energy eigenstates @xmath10 and corresponding energies @xmath11 for odd values of @xmath12 these states are symmetric about the middle ( @xmath13 ) while those for even @xmath12 are antisymmetric .
a particle constrained by no walls but acted upon by a one - dimensional delta function potential located at @xmath14 , namely @xmath15 has the associated schrdinger equation @xmath16 where , if @xmath17 and @xmath18 are solutions to the left ( @xmath19 ) and right ( @xmath20 ) of the delta function , the conditions to be satisfied at the potential are @xmath21 for an attractive interaction ( @xmath22 ) the single bound state is represented by @xmath23 where @xmath24 is a measure of the ` size ' of the bound state and the energy is @xmath25 whichever the sign of @xmath26 , solutions to ( [ schrod ] ) also include a continuum of positive energy unbound states , the scattering states . now consider putting the potential ( [ deltaf ] ) inside a box of finite side @xmath5
. then all states of the system are bound under the combined action of the walls and delta function . for this model
, solutions must satisfy eq .
( [ schrod ] ) , must vanish at the walls and must obey conditions ( [ bc-1 ] ) . for @xmath27 solutions of eqs .
( [ deltaf ] ) and ( [ schrod ] ) are linear combinations of @xmath28 , with energy @xmath29 . to be continuous at @xmath30 and
vanish at @xmath31 and @xmath32 solutions will have the form @xmath33 where @xmath34 is a constant which may depend on @xmath35 , @xmath5 and @xmath4 .
applying the first of boundary conditions ( [ bc-1 ] ) gives the energy dispersion relation .
it is @xmath36 where @xmath37 is given by ( [ biglam ] ) .
one set of solutions to ( [ dispeq ] ) is given by @xmath38 where any one of these equations implies the other two . here
@xmath35 must meet the two - fold condition @xmath39 where @xmath40 and @xmath41 are positive integers .
we shall assume that the ratio @xmath42 has been reduced to its primitive form that is to say with no common factors other than unity so that , because @xmath43 , we can write all of these _ nodal _ solutions @xmath44 , and their energies , for any such given value of @xmath3 as @xmath45 with wave numbers given by @xmath46 the particular feature of these solutions is that they have _ nodes _ at the location of the delta function potential .
they are that subset of the standing wave solutions ( [ well-1 ] ) for a free particle in a well with this property . because they vanish at @xmath14 they satisfy both conditions ( [ bc-1 ] ) . as an example ,
when @xmath3 equals @xmath47 ( or @xmath48 ) the states are @xmath49 with energies @xmath50 where @xmath51 .
when @xmath52 does not vanish eq .
( [ dispeq ] ) can be re - expressed as @xmath53 or , equivalently , as @xmath54 and @xmath4 varies from @xmath55 to @xmath5 as @xmath56 ranges from @xmath57 to @xmath58 .
we choose to call solutions for @xmath35 to either of these versions of the dispersion relation ( together with their associated wave functions ) _ ordinary _ solutions . should @xmath3 be a rational number , complete information about the nodal states and their energies ( all positive ) are given by eqs .
( [ sol-3 ] ) and ( [ sol-31 ] ) .
the ordinary states exist for all values of @xmath4 between @xmath55 and @xmath5 , with energies @xmath59 . for positive energies
the allowed values of @xmath35 are real - valued . in that case ,
solutions for @xmath60 to eq .
( [ disp-21 ] ) occur at intersections with the expression on the right - hand side of a straight line passing through the origin with slope @xmath61 .
similar comments apply to eq .
( [ lb-1 ] ) in terms of @xmath62 and slope @xmath63 . for negative energies , should there be any , the wave number is @xmath64 , where @xmath65 is real . in that case
the energy is @xmath66 and , letting @xmath67 in the pair ( [ disp-21 ] ) and ( [ lb-1 ] ) and in ( [ sol-1 ] ) , one must solve either of the equations @xmath68 with corresponding wave functions @xmath69
we choose to confirm by direct calculation that for this one - dimensional problem all solutions corresponding to different energies are , as expected , mutually orthogonal .
note first that , because the nodal solutions are free particle states in a simple well , it follows at once that those for different energies are mutually orthogonal .
then too these nodal solutions , eq . ( [ sol-3 ] ) , are orthogonal to the ordinary states . to see this ,
write the nodal states in condensed notation as @xmath70 now , for @xmath71 and @xmath72 , @xmath73 and @xmath74 then using definition ( [ sol-1 ] ) gives @xmath75 orthogonality also holds between any @xmath76 and any ordinary solution @xmath77 with negative energies @xmath78 , where @xmath65 is a solution to equation ( [ disp-3 ] ) .
finally , the ordinary solutions ( [ sol-1 ] ) , where @xmath35 satisfies the dispersion relation ( [ disp-21 ] ) , are orthogonal to each other for different values of @xmath35 .
for instance , including subscripts to indicate the @xmath35-dependence of wave functions , @xmath79 and normalization constants @xmath80 , we find by direct calculation , that when @xmath81 , @xmath82 where , doing the integrals and rearranging , @xmath83\nonumber\\ \mbox{and}\quad b & = & \frac{1}{(k+k ' ) } \big[\sin(k(l-\ell))\sin(k'(l-\ell))\sin((k+k')\ell ) \nonumber\\ & + & \sin(k \ell ) \sin(k'\ell ) \sin((k+k')(l-\ell ) ) \big].\nonumber\end{aligned}\ ] ] then , by using the eigenvalue equation ( [ disp-21 ] ) and trigonometric identities one finds that @xmath84 so that , when @xmath81 , @xmath85
the need to include both the nodal and ordinary solutions to make up a complete set can be seen by considering the limit of weak coupling for which @xmath86 is small and @xmath87 large . in that limit
one would expect , whatever the value of @xmath88 , that solutions ( [ sol-1 ] ) and their energies would approach those of a free particle in a well , eqs .
( [ well-1 ] ) and ( [ energy-1 ] ) , no more and no less .
consider solutions with positive energies .
for arbitrary @xmath88 , in addition to the nodal states which exist only when @xmath3 is rational we have the ordinary states . for them
@xmath89 is not an integer and the dispersion relation can be written , say , as ( [ disp-21 ] ) .
solutions for @xmath60 to either equation are intersections of a straight line passing through the origin with slope @xmath90 , large positive for attraction or large negative for repulsion .
these intersections occur only near any divergences of the right - hand side ( rhs ) of eq .
( [ disp-21 ] ) , i.e. near the zeroes of @xmath52 . then , setting @xmath91 , where @xmath92 is small but of either sign , shows that near these zeroes we have , approximately , @xmath93
\,\,\,\,\,(\epsilon/2)(1- \mu^2 ) + b(\epsilon)\quad(\mbox{for}\,\,n = 0,\,\,\epsilon < \pi ) , \end{cases}\ ] ] where @xmath94 is monotonically increasing as a function of @xmath92 in the interval @xmath95 and @xmath96 .
suppose first that @xmath97 is irrational .
in that case the function @xmath98 does nt vanish for any non - zero integer @xmath12 .
then , using the first of these two equations in ( [ disp-21 ] ) , with @xmath99 , gives ( for @xmath100 ) , @xmath101 using the second of equations ( [ lb-2 ] ) in ( [ disp-21 ] ) verifies that there is no solution in this limit for @xmath102 .
to summarize : when @xmath3 is irrational , solutions for @xmath103 to ( [ disp-21])or , what is equivalent , ( [ lb-1])will , for ever weaker coupling ( @xmath104 ) , approach @xmath105 for attraction or repulsion , where @xmath106 , so making up the full set of energies for a free particle in a well . also , it is easy to see , in the limit @xmath107 with @xmath108 , that solutions ( [ sol-1 ] ) approach those of the free particle in a well , eq .
( [ well-1 ] ) .
if instead @xmath3 should take the rational value @xmath42 , which we assume has been reduced by eliminating all common factors , then @xmath109 will vanish when and only when @xmath110 , @xmath51 .
there are in the limit , therefore , no ordinary solutions for these values of @xmath12 , but we do have the nodal states , eqs .
( [ sol-3 ] ) and ( [ sol-31 ] ) .
in addition to these nodal states , as we know from the argument just given for irrational @xmath3 , for large @xmath87 there exist ordinary solutions for all other integral values of @xmath12 again making up the full set of solutions for a free particle in a one - dimensional well . for negative energies ,
similar arguments show that when @xmath87 is large there are no solutions to either of eqs .
( [ disp-3 ] ) .
now consider , for any @xmath88 , the strong coupling limit , for which the magnitude of @xmath90 is small .
graphical argument shows that in this case , and only for attraction ( @xmath111 ) , there is a single state of negative energy .
this is easily verified by considering eq .
( [ disp-3 ] ) for small positive @xmath90 . for any value of @xmath56 between @xmath57 and @xmath112 the right - hand side of that equation is positive and monotonically increasing as a function of positive @xmath113 , from zero asymptotically to the value + 1 .
thus , the single solution approaches the value @xmath114 , giving the limiting energy value @xmath115 . except for the single bound state when the potential is attractive , for strong coupling the behavior of the oscillatory positive energy solutions to eq .
( [ schrod ] ) is controlled to some extent by the product @xmath116 in the potential energy term , for which one expects that whatever its sign , as the magnitude of @xmath26 increases any ordinary solution evaluated at the potential , @xmath117 , will decrease in compensation .
of course when @xmath3 is rational , the corresponding nodal solutions strictly vanish at the delta function and so are independent of the magnitude of @xmath26 , so long as it is non - zero . for the positive energies , when @xmath87 is small , we can say , from equation ( [ dispeq ] ) , that @xmath118 is also small , so that , whatever the sign of @xmath26 , @xmath119 where @xmath120 and @xmath121 are any positive integers and @xmath122 and @xmath123 are small and vanish as @xmath124 . note that , in the limits of vanishing @xmath122 and @xmath123 , cases ( a ) and ( b ) can not both hold , for then , according to ( [ sol-1 ] ) , the wave function would strictly vanish .
suppose first that @xmath3 is irrational .
then the two cases given in ( [ biglambda ] ) generate , in the limit , distinct energies and do not overlap for any values of @xmath120 and @xmath121 : if this were not so , then for some integers @xmath120 and @xmath121 we would have @xmath125 and @xmath126 for the same value of @xmath35 so that @xmath3 would be rational . cases ( a ) and ( b ) of ( [ biglambda ] ) give , in the limit of vanishing @xmath90 , nothing other than the energy levels of a free particle in separate wells of width @xmath4 and @xmath127 .
we can underline this by deriving the wave functions in the strong coupling limit . for case
( a ) , using in eq .
( [ sol-1 ] ) the first of expressions ( [ biglambda ] ) and taking the limit @xmath128 , gives @xmath129 and , similarly , for case ( b ) : in the limit @xmath130 these states approach the ordinary well states constrained to the interval @xmath131 $ ] with quantum number @xmath121 .
that the energies approach those of free particle motion in separate wells of widths @xmath4 and @xmath132 as @xmath133 ( when @xmath3 is an irrational number ) has been pointed out by joglekar @xcite .
there are subtleties when @xmath3 is rational , with @xmath134 , where we assume as usual that all common factors have been eliminated from numerator and denominator . in that case , for all values of @xmath37 , we have the nodal solutions , eqs .
( [ sol-3 ] ) and ( [ sol-31 ] ) , with wave numbers given by @xmath135 , where @xmath136 and , as before , in addition to those we have the ordinary solutions for which @xmath137 is small . for these , from eq .
( [ biglambda ] ) , @xmath138 where @xmath139 and @xmath140 are small . in this case possibilities ( a ) and ( b )
are both satisfied whenever @xmath141 but that instance gives as we have said a vanishing wave function and so must be ruled out as a solution .
another property is that when @xmath142 or @xmath143 , where @xmath144 , they are converging to nodal state energies . but note that , however small @xmath87 may be , so long as it is non - zero these ordinary states are distinct from nodal states and orthogonal to them . in summary : when @xmath145 ( reduced ) then the set of limiting values for @xmath60 as @xmath87 approaches zero comprises the nodal values together with all distinct ( non - duplicating ) limits in eq . ( [ ratb ] ) . as an example we have considered the case @xmath146 .
its nodal states have wave vectors given by @xmath147 , @xmath148 .
in addition to these we have the ordinary states . in the limit @xmath124
these are given by the zeroes of the right - hand side ( rhs ) of eq .
( [ disp-21 ] ) ( and ( [ lb-1 ] ) ) with @xmath146 .
this is shown as the dashed lines in fig .
[ lowlambda ] as a function of @xmath89 for values up to @xmath149 .
the solid curves show the same function plotted for the nearby value @xmath150 .
solutions for @xmath89 in equation ( [ disp-21 ] ) occur at intersections of the curves with a straight line through the origin with slope @xmath151 .
thus the figure shows that in the limit @xmath152 the degeneracy between the ordinary state and the lowest nodal energy at @xmath153 is _ lifted _ by moving the potential .
it also shows that , with @xmath146 for example , the double degeneracy at @xmath153 ( that occurs _ in the limit _
@xmath124 ) is split when @xmath90 is nonzero and of either sign .
the zeroes of the functions shown in fig .
[ lowlambda ] can easily be computed from eq .
( [ ratb ] ) . to explore the rational case
@xmath146 a bit further we have calculated the lowest energy eigenfunction that is doubly degenerate in the limit @xmath124 .
the lowest energy nodal state is , from ( [ sol-3 ] ) , given by @xmath154 . to get the ordinary solution at this energy we let @xmath155 in eq .
( [ sol-1 ] ) and take the limit to find @xmath156 where @xmath157 is a normalization constant . in the limit this ordinary solution vanishes at the delta function @xmath158 , suffers a discontinuity of slope there , and is orthogonal to the nodal solution @xmath159 .
the simplest choice is to locate the delta potential precisely in the middle so that @xmath3 is the rational number @xmath6 and @xmath161 . for this symmetric case
it turns out that the ordinary and nodal states are precisely equal in number in the sense that their energies interleave .
that is not the case for all other choices of @xmath3 , where the symmetry is lost , but it is still true that both types are countably infinite in number and that their inclusion is necessary to make up a complete set .
the nodal wave functions and their energies , @xmath162 and @xmath163 , are given by ( [ sol-3 ] ) with wave vectors @xmath164 , @xmath165 , eq . ( [ sol-31 ] ) .
these energies @xmath163 are , of course , all positive . for the ordinary solutions any negative energy eigenvalue would have @xmath166 and @xmath167 where , putting @xmath168 in ( [ disp-3 ] ) , @xmath65 would satisfy the equation @xmath169 . a sketch of this equation with respect to the variable @xmath170 shows that there are no solutions for @xmath65 when the interaction is a repulsion , none for relatively weak attraction ( @xmath171 ) , and but a single solution for relatively strong attraction ( @xmath172 ) .
in particular in the limit of very strong binding , for which @xmath63 approaches zero from above , @xmath173 , and the energy approaches @xmath174 . for the ordinary solutions at positive energies , acceptable values for @xmath103 must , from ( [ lb-1 ] ) , obey the equation @xmath175 .
this will include all the ordinary solutions except when the potential is relatively strongly attractive , for which @xmath176 and the lowest energy moves to negative values .
when the interaction is repulsive or relatively weakly attractive all ordinary states have positive energies , and a sketch reveals that , for this symmetric case , these solutions for @xmath60 interleave between those for the nodal solutions . in this sense , for the special case @xmath177 the nodal and ordinary states are equal in number . for a strong attraction ( @xmath176 )
there is but one negative energy state ( [ sol-11 ] ) .
if the potential is repulsive , whatever its strength , then all states have positive energies .
in addition to the nodal states @xmath162 we have the ordinary states . of these , should there be attraction , there will be , as @xmath178 , the single negative energy state with energy @xmath179 and wave function approaching @xmath180 , where ( see eq .
( [ bc-11 ] ) ) @xmath181 for strong repulsion ( and for strong attraction apart from the ground state @xmath180 ) all states have positive energies @xmath182 . for these states
@xmath87 is small and solutions for @xmath60 are in the neighborhood of the zeroes of @xmath183 , namely @xmath184 , where @xmath185 approaches zero as @xmath124 and @xmath51 .
thus , as @xmath186 , the energies of these ordinary states converge towards a double degeneracy with each and every one of the the nodal state values @xmath163 , eq . ( [ sol-3 ] ) , except , for attraction , the lowest state .
it is easy to find the positive energy wave functions for strong interaction . from ( [ sol-3 ] )
the nodal states are @xmath187 .
using @xmath188 in eq .
( [ sol-1 ] ) , taking the limit @xmath189 and normalizing , gives the following expression for the ordinary states in the limit of strong interaction : @xmath190 where @xmath191 is the step function @xmath192 these ordinary states have a discontinuous slope at the delta function , are even valued with respect to reflection about @xmath193 and are manifestly orthogonal to the nodal states which are odd valued about the middle . in this limit
the energy of each ordinary state @xmath194 approaches that of the corresponding nodal state @xmath162 : in this case of high symmetry all states approach doubly degeneracy in the limit of strong interaction
. however we note that for all finite strengths of the contact potential all energies of this one - dimensional system are , of course , non - degenerate .
we considered in subsection [ 1/2 ] solutions to eqs .
( [ lb-1 ] ) and ( [ disp-3 ] ) with @xmath168 , i.e. the case @xmath195 . in that instance
the ground state energy is positive for all strengths of repulsion , @xmath196 , and also for relatively weak attraction , @xmath197 .
but for relatively strong binding , for which the inequality @xmath198 holds , the ground state energy is negative . and , in particular , in the limit of strong binding , when @xmath199 , it approaches @xmath200 , the bound state energy of the free system , eq .
( [ ebee ] ) .
when the force center is not in the middle , then @xmath56 does not vanish and a re - analysis is required . for that case , whatever the details may be ,
we can say from the start that for either sign and any magnitude of @xmath37 , because eqs .
( [ lb-1 ] ) and ( [ disp-3 ] ) are symmetric with respect to @xmath56 about @xmath201 , as a function of @xmath4 , any ordinary state energy eigenvalue must be symmetric about @xmath195 .
it is easy to see from ( [ sol-3 ] ) that this symmetry also applies to the nodal states .
consider , in particular , configurations with the delta function located near to the left wall , say , with @xmath202 , where @xmath92 is small and positive .
as @xmath92 decreases , the nodal solution energy eigenvalues see eq .
( [ sol-3])will rapidly increase to large positive values . to see this , note that with @xmath203 , @xmath40 must increase without bound as @xmath204 , taking with it the accompanying energies @xmath205 , @xmath206 . as for the ordinary solutions , with @xmath207 and @xmath208
, we can say that as the delta function approaches a wall there are no negative energies . to see this ,
expand eq .
( [ disp-3 ] ) for small @xmath92 to get @xmath209 . in the limit @xmath210
this has no solutions for any finite value of @xmath65 .
so all energy eigenvalues are non - negative when the delta function potential is near to either wall . for arbitrary @xmath211 ,
a sketch allows us to predict the existence of ordinary solutions for small positive and negative energies , the transition between which occurs at @xmath212 .
the second of eqs .
( [ lb-2 ] ) ( which applies when @xmath60 is small ) and the graphical picture suggests the following : that for small @xmath60 and an attractive potential ( @xmath213 ) there is a solution for small real valued @xmath60 provided the slope @xmath90 is greater than @xmath214 ; that @xmath35 vanishes when @xmath215 ; that there is a negative energy solution with @xmath216 when @xmath90 is less than @xmath217 . from eq .
( [ lb-1 ] ) @xmath218 , so we can say that as a function of @xmath3 , the ground state energy passes through zero when @xmath219/2 = 2 ( 1 -\ell / l ) \ell / l$ ] , or @xmath220 note that , because @xmath3 is restricted to the interval @xmath221 , there can only be zeroes of energy provided @xmath222 , i.e. for relatively strong attraction . finally , consider the ordinary solutions for values of @xmath4 near , say , the left wall so that @xmath207 , where @xmath92 is small . expanding eq .
( [ disp-21 ] ) to lowest order and collecting terms gives @xmath223 , so that in the limit @xmath224 , @xmath225 @xmath226 , or @xmath227
. thus , in this limit the energies approach those of a free particle in a well , eq .
( [ energy-1 ] ) . to see how this limit is approached as @xmath228 we can set @xmath229 , where @xmath230 is small . then @xmath231 , so that @xmath232 or @xmath233 .
the corresponding energy is @xmath234 this result holds for either sign of @xmath235 .
the conclusion , then , is that when the delta function is near either wall , as a function of distance away , the energies are parabolic in form with zero slope at the wall , and they are a local maximum ( minimum ) for an attractive ( repulsive ) interaction .
the lowest , ground state , energy corresponds to the choice @xmath236 in eq .
( [ wall ] ) .
numerical calculation gives the dependence of energy eigenvalues on @xmath4 as it is varied from @xmath55 to @xmath5 .
all energies above the ground state are positive so for them one must solve either of the pair ( [ disp-21 ] ) or ( [ lb-1 ] ) . for the ground state ,
when @xmath237 , these equations pertain , but when @xmath238 eq . ( [ disp-3 ] ) holds .
their solutions join smoothly at zero energy , for which @xmath35 vanishes .
this happens at values of @xmath4 given by eq .
( [ kvan ] ) .
results for the ground state are shown in fig .
[ single ] where the ratio @xmath239 versus @xmath240 is plotted for the three values @xmath241 for the parameter @xmath242 .
we have done this for attractive and repulsive delta function potentials having the same strength @xmath243 . for the former ( latter ) case @xmath244 ( @xmath245 ) and
the the energy eigenvalues have their minimum ( maximum ) value at @xmath246 .
generally , for the case of attraction , should the wall separation @xmath5 increase to values large with respect to the characteristic length @xmath37 one expects that the ground state energy would approach @xmath179 , eq .
( [ ebee ] ) . indeed in the limit
@xmath247 eq .
( [ disp-3 ] ) becomes @xmath248 , provided @xmath249 so that @xmath250 and @xmath251 .
note , however , that however large @xmath5 may be , according to eq .
( [ wall ] ) with @xmath252 , in the limits @xmath228 and @xmath253 the wall dominates so that the energy approaches @xmath254 , the lowest energy of a free particle in the well .
this is consistent with fig .
[ single ] . in the other extreme of small @xmath5 , the walls dominate , whatever the value of @xmath4 . in this limit
@xmath87 is large , which is equivalent to weak coupling . for this limit the analysis around the first of eqs .
( [ lb-2 ] ) shows that @xmath255 as @xmath256 , so that in the limit @xmath257 we find that @xmath258 , the lowest energy of a free particle in a well .
the model analyzed here is simple and exactly solvable , but we are unaware of any complete published analysis .
if we think of an attractive delta function as a very simple model of the nucleus of a relatively massive atom , then along the lines of the born - oppenheimer approximation we might suppose that its position @xmath88 can be varied as a parameter .
thus the model suggests , reasonably , that for an attractive interaction the configuration of lowest energy is for the ` nucleus ' to avoid the walls .
this agrees with the perturbation calculations of changa et al @xcite for hydrogen in a hard spherical cavity .
note that as the well width is decreased the uncertainty principle demands that eventually the kinetic energy term of the schrdinger equation is the dominant contribution to the energy . as we pointed out in sec .
[ gs ] this also applies for a well of any width should the fixed potential be located sufficiently close to either wall .
we have shown that the energy levels and wave functions for choice @xmath195 follow as the limit of a ` top hat ' model potential having height @xmath259 and width @xmath260 where the ` area ' @xmath261 is held constant and @xmath262 , and it is reasonable to conjecture that the same result applies for off - center top hat potentials and , more generally , to any sequence of model potentials which approaches the delta function as a limit .
as a limit of sequences of non - singular potentials , the choice @xmath263 does require care in its handling .
in particular the behavior of the nodal energies , eqs .
( [ sol-3 ] ) and ( [ sol-31 ] ) , is somewhat chaotic as @xmath4 is varied .
for instance , for any given ` rational ' position @xmath134 , the lowest nodal energy is proportional to @xmath264 , so that the lowest such energies occur , in increasing order , at relative position @xmath265 and then at @xmath266 to the left of the midpoint or , symmetrically , at points @xmath267 to its right .
but should @xmath3 be moved through nodal points _ arbitrarily close _ to any one of these points then the lowest associated energy sweeps through chaotically large , even infinite , values .
nevertheless , we emphasize that , for rational values of @xmath3 a complete set of solutions requires both the nodal and ordinary ones .
by contrast , the ordinary solutions have energies which are solutions to eqs .
( [ disp-21 ] ) or ( [ lb-1 ] ) .
they are smooth continuous functions of position @xmath88 .
for instance this is true for the ground state discussed in sect .
[ gs ] . to conclude the paper we should like to make a few remarks of a purely mathematical nature concerning the model ,
most of which remains to be done .
first we prove that the hamiltonian , properly defined mathematically , is self adjoint .
as the reader will know , self adjointness is that subtle property of a symmetric ( hermitian ) operator that guarantees that the operator can be exponentiated . for the hamiltonian ,
this condition is necessary and sufficient to guarantee a unique time translation group . for our formal hamiltonian , which we here denote @xmath268
the process begins by defining a domain of definition for its interpretation as a sesquilinear _
form_. we choose as domain @xmath269 the set of functions @xmath270 in the hilbert space which belong to the domain of the kinetic operator and , in addition ,
have a first derivative @xmath271 which satisfies the discontinuity boundary condition at @xmath14 .
this implies that @xmath272 .
we can then consider the form @xmath273\hspace{-0.17em}\textnormal ] } } } } } : = \int_0^l { \settodepth{\pm@dpth}{$\underline{a}$}\settoheight{\pm@hght}{$g(t)$}\addtolength{\pm@hght}{\pm@dpth}\settowidth{\pm@wdth}{$g(t)$}\makebox[0pt][l]{\rule[\pm@hght]{\pm@wdth}{\fboxrule}}g(t)}\ , [ sf](t)\,dt\ ] ] for all @xmath274 in the domain @xmath269 . by a simple calculation , involving no more than integration by parts
, it follows that the associated quadratic form is lower bounded , @xmath275\hspace{-0.17em}\textnormal ] } } } } } = \frac{\hbar^2}{2 m } { \lvertf^\prime\rvert}^2-\lambda { \lvertf(\ell)\rvert}^2 \,\ge -\lambda\,\sup_{0\le x\le l}|f(x)|^2\ , > -\infty\,.\ ] ] it is now standard that the form can be extended to a closed form which defines a self adjoint operator , the friedrichs extension of @xmath276 .
it is this extension that we take as the mathematically proper hamiltonian @xmath259 . for details of the extension
we recommend davies @xcite and 124 of riesz and sz .-
nagy @xcite .
this leaves the following to prove : the spectrum of @xmath259 consists only of eigenvalues , indeed of those eigenvalues we have obtained earlier and only those .
we firmly believe these to be true , and that , in consequence , for a given @xmath4 and @xmath26 the eigenfunctions we obtained constitute an orthonormal basis for @xmath277 .
we remark that were we to independently prove that the eigenfunctions were complete then the eigenvalues we found would comprise the spectrum of @xmath259 . 5 d. a. dubin , j. kiukas , and j - p pellonp `` momentum observables for a bounded interval '' ( article in preparation ) ( 2010 ) .
p. l. garcia de leon , j. p. gazeau , and j. queva `` infinite quantum well : a coherent state approach , '' phys . lett .
a * 372 * , 35973607 ( 2008 ) . m. v. berry , `` quantum fractals in boxes , '' j. phys . a : math
* 29 * , 66176629 ( 1996 ) .
r. w. robinett , `` quantum wave packet revivals , '' physics reports * 392 * , 1119 ( 2004 ) .
d. f. styer , `` quantum revivals versus classical periodicity in the infinite square well , '' am .
* 69 * , 5662 ( 2001 ) . c. aslangul , `` surprises in the suddenly - expanded infinite well , '' j. phys .
a : math . theor . * 41 * , 075301 - 1 - 23 ( 2008 ) .
s. h. patil , `` completeness of the energy eigenfunctions for the one - dimensional @xmath278-function potential , '' am .
* 68 * 712 - 714 ( 2000 ) .
d. a. atkinson and h. w. crater , `` an exact treatment of the dirac delta function potential in the schrdinger equation , '' am .
* 43 * 301 - 304 ( 1975 ) .
n. bera , k. bhattacharyya and j. k. bhattacharjee , `` perturbative and nonperturbative studies with the delta function potential , '' am .
* 76 * 250 - 257 ( 2008 ) .
y. n. joglekar , `` particle in a box with a @xmath278-function potential : strong and weak coupling limits , '' am
. j. phys . * 77 * 734 - 736 ( 2009 ) .
s. t. epstein , `` application of the rayleigh - schrdinger perturbation theory to the delta function potential , '' am .
* 28 * 495 - 496 ( 1960 ) .
e. p. wigner , `` application of the rayleigh - schrdinger perturbation theory to the hydrogen atom , '' phys
. rev . * 94 * 77 - 78 ( 1954 ) .
a. michels , j. de boer , and a. bijl , `` remarks concerning molecular interaction and their influence on polarizability , '' physica * 4 * 981 - 994 ( 1937 ) .
a. sommerfeld and h. welker , `` knstliche grenzbedingungen beim keplerproblem , '' annalen der physik * 32 * 56 - 65 ( 1938 ) .
y. p. varshni , `` accurate wavefunctions for the confined hydrogen atom at high pressures , '' j. phys .
b : at . mol .
. phys . * 30 * l589593 ( 1997 ) .
y. p. varshni , `` critical cage radii for the confined hydrogen atom , '' j. phys .
b : at . mol .
31 * 2849 - 2856 ( 1998 ) .
s. yngve , `` the energy levels and the corresponding normalized wave functions for a model of a compressed atom .
ii , '' j. math
* 29 * 931 - 936 ( 1988 ) .
s. goldman and c. joslin , `` spectroscopic properties of an istropically compressed hydrogen atom , '' j. chem
* 96 * 6021 - 6027 ( 1992 ) . p. l. goodfriend , `` on the use of linear variation basis functions that do not satisfy the boundary conditions '' j. phys .
b : at . mol .
* 23 * 1373 - 1379 ( 1990 ) . p.
o. frman , s. yngve and n. frman , `` the energy levels and the corresponding normalized wave functions for a model of a compressed atom , '' j. math .
* 28 * 1813 - 1826 ( 1987 ) .
m. e. changa , a. v. scherbinin , and v. i. pupyshev , `` perturbation theory for the hydrogen atom in a spherical cavity with off - centre nucleus '' j. phys .
b : at . mol . opt . phys . * 33 * 421 - 432 ( 2000 ) .
r. loudon , `` one - dimensional hydrogen atom , '' am .
* 27 * 649 - 655 ( 1959 ) .
m. andrews , `` ground state of the one - dimensional hydrogen atom , '' am . j. phys .
* 34 * 1194 - 1195 ( 1966 ) .
dai xianxi , jixin dai and jiqiong dai , `` orthogonality criteria for singular states and the nonexistence of stationary states with even parity for the one - dimensional hydrogen atom , '' phys .
a * 55 * 2617 - 2624 ( 1997 ) .
s. nouri , `` generalized coherent states for the coulomb problem in one dimension , '' phys .
a * 65 * 062108 - 1 - 5 ( 2002 )
. g. abramovici and y. avishai , `` the one - dimensional coulomb problem , '' j. phys . a : math .
* 42 * 285302 - 1 - 29 ( 2009 ) . e. b. davies , _ spectral theory and differential operators_. cambridge university press , 1995 . f. riesz and b. sz .- nagy , _ functional analysis _ , f. ungar , new york 1955 .
this shows the right - hand side ( rhs ) of eq .
( [ disp-21 ] ) as a function of @xmath60 in units of @xmath279 for @xmath280 ( dashed ) and @xmath281 ( solid ) .
solutions for the ordinary states are intersections of the curves with a straight line from the origin having slope @xmath282 . when @xmath283 the lowest nodal state has a wave number given by @xmath153 , which is shared with an ordinary state ( dashed line ) in the limit @xmath284 .
that the dashed curve moves from the value @xmath285 shows that this limiting degeneracy is lifted as @xmath286 moves from zero . on the other hand , note that even when @xmath287 this double degeneracy is lifted by moving the potential away from the value @xmath288 .
the fact that @xmath289 is also rational does not effect this basic splitting : its lowest lying nodal state occurs well to the right , at @xmath290 . ]
the result of numerical calculations ( for values @xmath241 of the parameter @xmath291 ) for the dependence of the ` relative energy ' @xmath239 of the lowest energy eigenvalue versus @xmath292 locating the delta function within a box of side @xmath5 .
the curves with minima are for an attractive potential @xmath293 .
their partner curves , with maxima , correspond to a repulsive interaction of the same strength .
the energy of the single bound state of an attractive potential of strength @xmath294 is @xmath295 , eq .
( [ ebee ] ) , and @xmath37 is the spatial decay length for that state , eq .
( [ biglam ] ) . ] | we analyze here the energy states and associated wave functions available to a particle acted upon by a delta function potential of arbitrary strength and sign and fixed anywhere within a one - dimensional infinite well .
we consider how the allowed energies vary with the well s width and with the location of the delta function within it .
the model subtly distinguishes between whether the delta function is located at rational or irrational fractions of the well s width : in the former case all possible energy eigenvalues are solutions to a straightforward dispersion relation , but in the later case , to make up a complete set these ` ordinary ' solutions must be augmented by the addition of ` nodal ' states which vanish at the delta function and so do not ` see ' it .
thus , although the model is a simple one , due to its singular nature it needs a little careful analysis .
the model , of course , can be thought of as a limit of more physical smooth potentials which , though readily succumbing to straightforward numerical computation , would give little generic information .
+ pacs numbers : 03.65.-w , 73.21.fg , 01.40.-d |
one of the most important driving forces in condensed matter physics in the last thirty years has been the controlled growth of layered structures so thin that interface effects dominate bulk properties and quantum size effects can be observed . in doped semiconductors ,
the large fermi wavelength of mobile charge carriers made it possible to observe finite size effects for layer thicknesses on a micron scale .
much thinner layers must be used in order to make such observations in metals because fermi wavelengths are typically of the order of an interatomic spacing .
nevertheless , following rapidly on the heels of a number of important discoveries in semiconductor heterostructures , interface - dominated effects such as interface magnetic anisotropy , oscillatory exchange coupling and giant magnetoresistance ( gmr ) were found in artificially layered transition metal materials . reflecting the shorter fermi wavelength ,
the characteristic length scale is of order of nanometers .
our main purpose in this paper is to give details of a scheme we have developed which is suitable for studying mesoscopic transport in inhomogeneous , mainly layered , transition metal magnetic materials . in the context of a large number of schemes designed to study transport either from first - principles@xcite or based upon electronic structures calculated from first - principles@xcite we will require our computational scheme to be ( i ) physically transparent , ( ii ) first - principles , requiring no free parameters , ( iii ) capable of handling complex electronic structures characteristic of transition metal elements and ( iv ) very efficient in order to be able to handle lateral supercells to study layered systems with different lattice parameters and to model disorder very flexibly .
a tight - binding ( tb ) muffin - tin - orbital ( mto ) implementation of the landauer - bttiker formulation of transport theory within the local - spin - density approximation ( lsda ) of density - functional - theory ( dft ) will satisfy these requirements . because wave transport through interfaces is naturally described in terms of transmission and reflection , the landauer - bttiker ( lb ) transmission matrix formulation of electron transport gained rapid acceptance as a powerful tool in the field of mesoscopic physics,@xcite once the controversies surrounding the circumstances under which different expressions should be used had been resolved .
@xcite the two - terminal conductance of a piece of material is measured by attaching leads on either side , passing a current through these leads and measuring the potential drop across the scattering region . in the lb formulation of transport theory ,
the conductance @xmath5 is expressed in terms of a transmission matrix @xmath6 @xmath7where the element @xmath8 is the probability amplitude that a state @xmath9 in the left - hand lead incident on the scattering region from the left ( see fig . [ fig : lsr ] ) is scattered into a state @xmath10 in the right - hand lead .
the trace simply sums over all incident and transmitted `` channels '' @xmath11 and @xmath12 and @xmath13 is the fundamental unit of conductance .
in much current work on first - principles transport the conductance is calculated directly from green s functions expressed in some convenient localized orbital representation.@xcite explicit calculation of the scattering states is avoided by making use of the invariance properties of a trace . because we want to make contact with a large body of theoretical literature@xcite on mesoscopic physics and address a wider range of problems in the field of spin - dependent transport , we will calculate the microscopic transmission and reflection matrices @xmath14 and @xmath15 . by using a real energy
, we will avoid the problems encountered in distinguishing propagating and evanescent states when a small but finite imaginary part of the energy is used .
the landauer - bttiker formalism satisfies our first requirement of physical transparency . )
is sandwiched by left- ( @xmath16 ) and right - hand ( @xmath17 ) leads which have translational symmetry and are partitioned into principal layers perpendicular to the transport direction .
the scattering region contains @xmath0 principal layers but the structure and chemical composition are in principle arbitrary . ] in developing a scheme for studying transport in transition metal multilayers , a fundamental difference between semiconductors and transition metals must be recognized .
transition metal atoms have two types of electrons with different orbital character .
electrons are spatially quite extended and , in solids , form broad bands with low effective masses ; they conduct easily .
electrons are much more localized in space , form narrow bands with large effective masses and are responsible for the magnetism of transition metal elements .
the `` magnetic '' electrons , however , being itinerant do contribute to electrical transport .
the appropriate framework for describing metallic magnetism , even for the late 3@xmath18 transition metal elements , is band theory.@xcite an extremely successful framework exists for treating itinerant electron systems from first - principles and this is the local density approximation ( lda ) of density functional theory ( dft ) . for band magnetism ,
the appropriate extension to spin - polarized systems , the local spin - density approximation ( lsda ) satisfies our second requirement of requiring no free parameters.@xcite oscillatory exchange coupling in layered magnetic structures was discussed by bruno in terms of generalized reflection and transmission matrices bruno : jmmm93 which were calculated by stiles @xcite for realistic electronic structures using a scheme @xcite based on linearized augmented plane waves ( lapws ) . at an interface between a non - magnetic and a magnetic metal ,
the different electronic structures of the majority and minority spin electrons in the magnetic material give rise to strongly spin - dependent reflection .
@xcite schep used transmission and reflection matrices calculated from first - principles with an embedding surface green s function method @xcite to calculate spin - dependent interface resistances for specular co / cu interfaces embedded in diffusive bulk material.@xcite the resulting good agreement with experiment indicated that interface disorder is less important than the spin - dependent reflection and transmission from a perfect interface .
calculations of domain wall resistances as a function of the domain wall thickness illustrated the usefulness of calculating the full scattering matrix .
@xcite however , the lapw basis set used by stiles and schep was computationally too expensive to allow repeated lateral supercells to be used to model interfaces between materials with very different , incommensurate lattice parameters or to model disorder .
this is true of all plane - wave based basis sets which typically require of order 100 plane waves per atom in order to describe transition metal atom electronic structures reasonably well .
muffin - tin orbitals ( mto ) form a flexible , minimal basis set leading to highly efficient computational schemes for solving the kohn - sham equations of dft.@xcite for the close packed structures adopted by the magnetic materials fe , co , ni and their alloys , a basis set of 9 functions ( _ s , p _ , and _
d _ orbitals ) per atom in combination with the atomic sphere approximation ( asa ) for the potential leads to errors in describing the electronic structure which are comparable to the absolute errors incurred by using the local density approximation .
this should be compared to typically 100 basis functions per atom required by the more accurate lapw method .
mtos thus satisfy our third and fourth requirements of being able to treat complex electronic structures efficiently .
the tight - binding linearized muffin tin orbital ( tb - lmto ) surface green s function ( sgf ) method has been developed to study the electronic structure of interfaces and other layered systems . when combined with the coherent - potential approximation ( cpa )
, it allows the electronic structure , charge and spin densities of layered materials with substitutional disorder to be calculated self - consistently very efficiently.@xcite in this paper we describe how we have combined a method for calculating transmission and reflection matrices based on wave - function matching ( wfm ) , in a form given by ando@xcite for an empirical tight - binding hamiltonian , with a first - principles tb - mto basis .
@xcite applications of the method to a number of problems of current interest in spin - transport have already been given in a number of short publications : to the calculation of spin - dependent interface resistances where interface disorder was modelled by means of large lateral supercells ; @xcite to the first principles calculation of the so - called mixing conductance parameter entering theories of current - induced magnetization reversal@xcite and gilbert - damping enhancement via spin - pumping;@xcite to a generalized scattering formulation of the suppression of andreev scattering at a ferromagnetic / superconducting interface ; @xcite to the problem of how spin - dependent interface resistances influence spin injection from a metallic ferromagnet into a iii
v semiconductor.zwierzycki:prb03,bauer:prl04 these examples amply demonstrate that the fourth requirement is well satisfied . in sec .
[ sec : theory ] , we give technical details of the formalism and illustrate it in sec . [ sec : calc ] where we calculate the transmission matrices for clean and disordered co / cu interfaces , document a number of convergence and accuracy issues and give a detailed `` channel - decomposition '' analysis of the scattering in the presence of disorder .
a comparison with other methods is made in sec .
[ sec : discussion ] .
central to the wave - function matching method for calculating the transmission and reflection matrices is the equation of motion ( eom ) for electrons with energy @xmath19 , relating the vectors of coefficients @xmath20 for layers @xmath21 , @xmath22 , and @xmath23 : @xmath24here , @xmath25 describes the wavefunction amplitude in terms of some localized orbital basis @xmath26 of dimension @xmath27 where @xmath28 labels the atomic orbital and atom site .
[ for the muffin - tin orbitals to be outlined in sec .
[ ssec : mtos ] , @xmath28 will be a combined index @xmath29 , where @xmath30 and @xmath31 are the azimuthal and magnetic quantum numbers , respectively , of the mto defined for an atomic - spheres - approximation ( asa ) potential on the site @xmath32 . ]
the eom does not restrict us to only considering nearest neighbour interactions since atoms can always be grouped into layers defined as to be so thick that the interactions between layers @xmath22 and @xmath33 are negligible ( see fig . [ fig : lsr ] ) .
such layers are called _
principal layers_. their thickness depends on the range of the interactions which in turn partly depends on the spatial extent of the orbital basis .
it will be minimized by using the highly localized tight - binding mto representation .
consider the situation sketched in fig .
[ fig : lsr ] where the scattering region @xmath34 is contacted with left ( @xmath35 and right ( @xmath36 leads which have perfect lattice periodicity and support well - defined scattering states .
we assume that the ground state charge and spin densities and the corresponding kohn - sham independent electron potential have already been calculated self - consistently .
the calculation of the scattering matrix can now be split into two distinct parts . in the first stage , to be discussed in sec .
[ ssec : leads ] , the eigenmodes of the leads @xmath37 for the @xmath38-th mode ) , of which there are @xmath39 , are calculated using an eom appropriate to mtos and making use of the lattice periodicity . by calculating their * k * vectors ( which are in general complex ) and velocities @xmath40 ,
the eigenstates can be classified as being either left - going @xmath41 or right - going @xmath42 .
they form a basis in which to expand any left- and right - going waves and have the convenient property that their transformation under a lattice translation in the leads is easily calculated using bloch s theorem ( with * k * complex ) .
we use the small roman letters _ i , j _ to label the non - orthogonal basis and the small greek letters @xmath43 to label the lead eigenmodes . in the second stage
discussed in sec .
[ ssec : scatt ] , a scattering region @xmath34 is introduced in the layers @xmath44 which mixes left- and right - going lead eigenmodes .
the scattering region can be a single interface , a complex multilayer or a tunnelling junction , and the scattering can be introduced by disorder or simply by discontinuities in the electronic structure at interfaces .
the @xmath45 element of the reflection matrix , @xmath46 is defined in terms of the ratio of the amplitudes of left - going and right - going solutions in the left lead ( in layer 0 for example ) projected onto the @xmath47 right - going and @xmath48 left - going propagating states ( * k * vector real ) renormalized with the velocities so as to have unit flux .
the scattering problem is solved by direct numerical inversion of a matrix with the leads included as a boundary condition so as to make finite the matrix which has to be inverted .
muffin - tin orbitals @xcite ( mto ) are defined for potentials in which space is divided into non - overlapping atom - centred `` muffin - tin '' spheres inside which the potential is spherically symmetric and the remaining `` interstitial '' region where the potential is taken to be constant .
the atomic spheres approximation ( asa ) is obtained ( i ) by taking the kinetic energy in the interstitial region to be zero and ( ii ) by expanding the muffin - tin spheres so that they fill all space whereby the volume of the interstitial region vanishes ; for monoatomic solids such spheres are called atomic wigner - seitz ( ws ) spheres . inside a ws ( or mt ) sphere at @xmath32 , the solution of the radial schrdinger equation regular at @xmath32 , @xmath49 can be determined numerically for energy @xmath19 and angular momentumum @xmath30 resulting in the partial wave @xmath50where @xmath51 and @xmath52 .
a continuous and differentiable orbital is constructed by attaching to the partial wave at the sphere boundary @xmath53 a `` tail '' consisting of an appropriate linear combination of the solutions of the laplace equation , @xmath54^{-1}y_{l}(% \mathbf{\hat{r}}_{r } ) \label{eq : j0}\]]and @xmath55which are respectively , regular at @xmath32 and at infinity .
@xmath56 is the average ws radius if the structure contains different atoms . in terms of the logarithmic derivative of @xmath57 at @xmath58 @xmath59(@xmath60
is the radial derivative ) , the radial solutions are matched if for @xmath61 , @xmath62\end{aligned}\]]where we drop the explicit @xmath32-dependence when this does not give rise to ambiguity , or in terms of the potential function @xmath63and normalization @xmath64 @xmath65 by subtracting from the partial wave , both inside and outside the mt sphere , the @xmath66 component which is irregular at infinity , a function is formed which is continuous , differentiable and regular in all space , an energy - dependent muffin tin orbital @xmath67 : @xmath68the tail @xmath69 has the desirable property that closed forms exist for expanding it around a different site @xmath70 in terms of the regular solutions @xmath71@xmath72the expansion coefficients @xmath73form a so - called canonical structure constant matrix : they do not depend on the lattice constant , on the mt ( or as ) potentials or on energy . because of the augmentation with @xmath66 , the resulting mto is no longer a solution of the schrdinger equation ( se ) inside its own sphere @xmath32 .
when , however , a solution of the se is sought in the form of a linear combination of mtos centred on different sites , @xmath74then the partial wave solution is recovered if the augmenting term @xmath75 on site @xmath32 is cancelled by the tails of mtos centred on all other sites @xmath76 , expanded about @xmath77 the condition for this to occur is the `` tail - cancellation '' condition : @xmath78 c_{r^{\prime } l^{\prime } } ^{0}=0 .
\label{eq : pms}\]]all of the information about the structural geometry of the system under investigation is contained in the structure constant matrix @xmath79 while all of the information about the atomic species on site @xmath32 needed to calculate the electronic structure ( eigenvalues and eigenvectors ) is contained in the potential functions @xmath80 .
these are determined by solving the radial schrdinger equation for the corresponding spherically symmetrical atomic sphere potential for energy @xmath81 and angular momentum @xmath30 .
a disadvantage of these `` conventional '' mtos is their infinite range .
however , there is a remarkably simple generalization of the mtos which allows their range to be modified by introducing a set of `` screening '' constants @xmath82 ( not to be confused with the lead eigenmode index ) while the `` tail - cancellation '' condition remains essentially unchanged : @xmath83 c_{r^{\prime } l^{\prime } } ^{\alpha } = 0 .
\label{eq : tbpms}\ ] ] @xmath84 is a diagonal matrix related to @xmath85 by @xmath86and@xmath87for any set of @xmath82 , the energy - dependent mtos with the normalization @xmath88form a complete set for the mt ( as ) potential used in their construction . here , @xmath89 denotes an energy derivative and ( [ eq : norm ] ) follows from the relation @xmath90^{1/2}$ ] .
sets of parameters @xmath91 have been found for which the `` screened '' structure constants @xmath92 have very short range , decaying exponentially with the interatomic separation .
@xcite the set of parameters , @xmath93 , which yields the shortest range mtos is called the `` tight - binding '' ( tb ) representation . @xcite for close - packed structures
, the range of @xmath94 is in practice limited to first- and second - nearest neighbours .
this tb set with @xmath95 is what we will use from now , unless stated otherwise , since it will allow us to define principal layers with a minimal thickness . for the determination of energy bands @xmath96 the tail - cancellation or kkr equations
are inconvenient because the energy - dependence of the potential function makes it necessary to solve ( eq : pms ) or ( [ eq : tbpms ] ) by searching for the roots of a determinant , which is very time consuming .
much more efficient methods have been developed based on energy - independent mtos .
however , to study transport we only need to know @xmath97 for a fixed energy , usually the fermi energy .
we assume that the kohn - sham equations have already been solved self - consistently ( using for example a linearized method ) so we have the potentials from which to calculate the potential functions .
although ( [ eq : tbpms ] ) can be brought into hamiltonian form by linearizing the energy dependent potential function ( see appendix [ sec : velocities ] ) , we will work directly with the more exact kkr equation .
we will assume that there exists two - dimensional translational symmetry in the plane perpendicular to the transport direction so that states can be characterized by a lateral wave vector @xmath98 in the corresponding two - dimensional brillouin zone .
the screened kkr equation @xcite in the mixed representation of @xmath98 and real space layer index @xmath22 ( see fig . [
fig : lsr ] ) is @xmath99where @xmath100 is a @xmath101 @xmath102 dimensional vector describing the amplitudes of the @xmath22-th layer with @xmath1 sites and @xmath103 orbitals per site . @xmath104 and @xmath105 are @xmath106 matrices .
@xmath104 is a diagonal matrix of potential functions characterizing the as potentials of layer @xmath22 and @xmath107where @xmath108 denotes the set of vectors that connect one lattice site in the @xmath22-th layer with all lattice sites in the @xmath109-th layer . by analogy with ( [ eq : eom ] ) , equation ( [ eq : pseom ] ) is the equation of motion we will use to calculate the amplitudes of right- and left - going waves which determine the scattering matrix
. we will solve it for a fixed value of @xmath19 ( usually @xmath110 ) , and some @xmath111 to find @xmath112 the component of the bloch vector in the transport direction . to keep the notation simple
, explicit reference to the @xmath98 and @xmath19 dependence will be omitted from now on .
the formalism to be described in the following can be applied to any electronic structure code based on the kkr equation ( [ eq : pseom ] ) , such as third - generation tb - lmto.@xcite let us first consider the bloch states in the ideal lead . to obtain linearly independent solutions , we set @xmath113 , since in a periodic potential the wave function should satisfy bloch s theorem .
the potential function matrix is the same for all unit cells .
the structure constant matrix depends only on the relative positions and , because that is how they are defined , there is only coupling between adjacent principal layers so the equation of motion becomes @xmath114the eigenvalue @xmath115 can be written in the form @xmath116 with @xmath117 connecting equivalent sites in adjacent prinicpal layers .
the wave vector @xmath118 can be decomposed into @xmath98 and a remainder which is in general not real , @xmath119 .
equation ( [ eq : leads ] ) has @xmath39 eigenvalues and @xmath39 eigenvectors , corresponding to @xmath27 right - going and @xmath27 left - going waves . by calculating the wavevectors and velocities ( see eq . and appendix [ sec : velocities ] ) of the lead eigenmodes ,
the propagating and evanescent states can be identified and sorted into right - going or left - going modes .
letting @xmath120 denote the left - going solutions @xmath121 corresponding to eigenvalues @xmath122 and @xmath123 the right - going solutions corresponding to eigenvalues @xmath124 , the matrix @xmath125 is defined as @xmath126and the matrix @xmath127 as the diagonal matrix with elements @xmath128 . following ando
, we next expand any left- or right - going wave , at @xmath129 for example , as @xmath130note that @xmath121 is a vector whose elements are labelled @xmath28 while the elements of the vector @xmath131 are labelled @xmath12.@xmath132 is the matrix of bloch factors ( including evanescent states ) transformed onto the basis @xmath133 and plays a central role in the following . knowing it makes it possible to translate a state expressed in the basis @xmath133 from layer @xmath109 of the lead to layer @xmath22 by @xmath134 the scattering region @xmath34 , divided into @xmath0 principal layers numbered @xmath135 to @xmath0 , is now inserted between the left and right leads .
the resulting ( scattering region + leads ) problem is infinite dimensional in the real space mto representation but by making use of their translational symmetry , the leads can be incorporated as boundary conditions and the scattering problem can be reduced to a finite problem whose dimension is determined by the size of the scattering region ( number of sites @xmath136 number of orbitals per site ) .
we set about decoupling the scattering region from the leads , first on the left - hand side , then on the right .
the amplitude in the @xmath137-th layer is first separated into right- and left- going components @xmath138 . because there is no additional scattering in the leads
, the right- and left - going components can be translated to the left by one ( principal layer ) lattice spacing using the generalized bloch factors ( [ eq : trans ] ) so the amplitude in layer @xmath139 can be related to that in layer @xmath137 as @xmath140 \mathbf{c}% _ { 0}(+)+f_{\mathcal{l}}^{-1}(-)\mathbf{c}_{0}. \label{eq : c-1}\end{gathered}\]]allowing us to express @xmath141 in terms of @xmath121 and @xmath142 and so eliminate it from the equation of motion for the @xmath143-th layer @xmath144 which becomes @xmath145 \mathbf{c}_{0}(+).\end{aligned}\ ] ] here @xmath16 denotes the left lead and @xmath146 .
@xmath147 is the `` embedding potential '' for the left lead and the net result is that the equations of motion have been truncated at layer @xmath148 on the right - hand side of the scattering region , we are interested in the situation where only right - going waves can exist in the @xmath149-th layer , so @xmath150 allowing us to eliminate @xmath151 * * * * from the eom for @xmath152 @xmath153where @xmath154 and @xmath155 is the embedding potential for the right lead .
making use of the lead boundary conditions , the tail cancellation condition for the scattering problem in real space is given by the set of inhomogeneous linear equations @xmath156
\mathbf{% c}_{0}(+ ) \\ 0 \\ 0 \\ \vdots \\ 0 \\ 0% \end{array } \right ) & & \label{eq : green}\end{aligned}\]]which can be solved in terms of @xmath157 @xmath158 \mathbf{% c}_{0}(+ ) \\ 0 \\ 0 \\ \vdots \\ 0 \\ 0% \end{array}% \right)\ ] ] this treatment is very similar to the widely used surface green function method.@xcite the boundary conditions in ( [ eq : green ] ) are explicitly defined by considering the bloch wave coming from the left - hand side while for conventional retarded or advanced green functions the boundary conditions are specified by an infinitesimal imaginary part of the energy parameter @xmath19 .
we are now in a position where we can relate the outgoing wave amplitude in the right electrode to the incoming wave in the left electrode through the green function by @xmath159 \mathbf{c}_{0}(+).\]]using the transformation between the eigenstates and the localized basis functions @xmath160 , we obtain the transmission and reflection matrix elements@xcite @xmath161 u_{\mathcal{l}}(+)\right\}_{\mu \nu } \label{trale}\ ] ] @xmath162 - 1\right\rangle u_{\mathcal{l}}(+)\right\}_{\mu \nu } , \label{refle}\ ] ] where @xmath38 and @xmath11 label bloch states and @xmath163 , @xmath163 are the components of the corresponding group velocities in the transport direction .
similarly , an incident wave from the right side is transmitted or reflected as @xmath164 u_{\mathcal{r}}(- ) \right\}_{\mu\nu } \label{trari}\ ] ] @xmath165 -1 \right\rangle u_{\mathcal{r}}(- ) \right\ } _ { \mu \nu } .
\label{refri}\ ] ] the group velocities in ( [ trale])-([refri ] ) are determined using the expression @xmath166
\label{eq : velocity}\ ] ] which is derived in appendix a. here , @xmath22 and @xmath23 denote neighbouring principal layers in either left or right lead , @xmath167 is the distance between equivalent monolayers in adjacent principal layers and @xmath168 is a unit vector in the transport direction
. interfaces between materials with different lattice parameters @xcite and disordered interfaces@xcite can be modelled very flexibly using lateral supercells .
this approach allows us to study the effect of various types of disorder on transport properties , ranging from homogeneous interdiffusion ( alloying ) to islands , steps etc .
the supercell description of disorder becomes formally exact in the limit of infinitely large supercells . in practice ,
satisfactory convergence is achieved for supercells of quite moderate size ( see sec .
[ ssec : intdis ] ) .
the factor limiting the `` size '' of scattering problem which can be handled in practice is the rank of the blocks of the block - tridiagonal scattering matrix in , which is proportional to the number of atoms in the supercell . if performed straightforwardly in the manner outlined in sec .
[ ssec : leads ] , the solution of the lead equation involves solving a non - hermitian eigenvalue problem whose rank is twice as large .
unless use is made of the greater translational symmetry present in the leads , this can become the limiting step in the whole calculation .
doing so makes it possible to reduce the dimension of the lead state calculation to a size determined by the dimension of a primitive unit cell which is usually negligible .
supercell ( rhs ) .
bottom panel : a single k - point in the bz ( rhs ) corresponding to the @xmath169 real - space supercell is equivalent to @xmath169 k - points in the bz ( lhs ) corresponding to the real - space primitive unit cell . ]
we consider an @xmath170 lateral supercell defined by the real - space lattice vectors @xmath171 where @xmath172 and @xmath173 are the lattice vectors describing the in - plane periodicity of a primitive unit cell ( fig .
[ fig : supercells ] ) .
the cells contained within the supercell are generated by the set of translations @xmath174 where @xmath175 is a convenient cell index . in reciprocal space the supercell brillouin zone is defined by the reduced vectors @xmath176 where @xmath177 and @xmath178 are the reciprocal lattice vectors corresponding to the real space primitive unit cell . as a result the brillouin zone ( bz ) is folded down , as shown schematically in fig .
[ fig : supercells ] ( bottom rhs ) , and the single @xmath179 point ( @xmath180 is used to label supercell quantities ) in the supercell bz corresponds to the set of @xmath170 @xmath181 points in the original unfolded bz @xmath182 with @xmath183 .
solutions associated with different @xmath184 in the primitive unit cell representaton become different `` bands '' at the single @xmath179 in the supercell representaton . the indices @xmath185 and @xmath186 provide a natural means of describing the supercell - related matrices @xmath187 and @xmath188 and their inverses in terms of @xmath189 sub - blocks with dimensions defined by the primitive unit cell .
thus @xmath190 is the block containing the amplitudes of the modes associated with @xmath184 in the @xmath185-th real - space cell .
solving the single unit cell problem for the set of @xmath191-points belonging to @xmath192 ( lhs of fig . [ fig : supercells ] ) and using the bloch symmetry of the eigenmodes , we get trivially @xmath193 where @xmath194 is the matrix of modes for a primitive unit cell for @xmath184 and the @xmath195 qualifier has been dropped for simplicity . defining the matrix of phase factors @xmath196 with @xmath197 , and its inverse @xmath198
, we can straightforwardly determine @xmath199^{-1}_{\mathcal{k } \mathcal{t } } = u^{-1}({\mathbf{k}_{\parallel}}_{\mathcal{k } } ) y_{\mathcal{k } \mathcal{t } } \label{eq : super_uinv}\ ] ] and @xmath200 the procedure outlined above for determining the matrices describing the lead modes scales linearly with the size of the supercell _ i.e. _ , as ( @xmath170 ) rather than as @xmath201 which is the scaling typical for matrix operations .
another advantage is that it enables us to analyse the scattering . by keeping track of the relation between supercell `` bands '' and equivalent eigenmodes at different @xmath184 ( fig .
[ fig : supercells ] ) we can straightforwardly obtain from - @xmath202 and other scattering coefficients . in other words
the `` interband '' specular scattering in the supercell picture translates , in the presence of disorder in the scattering region , into the `` diffuse '' scattering between the @xmath191 vectors belonging to the @xmath192 set .
even though the theoretical scheme outlined above contains no adjustable parameters , its practical implementation does involve numerous approximations , some physical , others numerical , which need to be evaluated . at present
, any workable scheme must be based upon an independent particle approximation .
the results of a transport calculation will be limited by the extent to which the single particle electronic structures used are consistent with the corresponding fermi surfaces determined experimentally using methods such as de haas - van alphen measurements or the occupied and unoccupied electronic states close to the fermi energy determined by , for example , photoelectron spectroscopy . in this section
we examine how various approximations affect our end results .
we begin with the calculation of the scattering states in bulk cu and bulk co ( [ ssec : calcsleads ] ) .
these are then used to study specular scattering from an ideal ordered cu / co(111 ) interface ( [ ssec : ordint ] ) after which we describe how we model disordered interfaces ( [ ssec : intdis ] ) and how the results can be analysed . for a crystalline conductor with bloch translational symmetry ,
each state at the fermi energy can move unhindered through the solid so that the transmission matrix is diagonal with @xmath203 . in this
_ ballistic _ regime , ( [ eq : lb ] ) reduces to @xmath204 and calculation of the so - called _ sharvin _ conductance becomes a matter of counting the number of modes ( channels ) propagating in the transport direction @xmath168 , denoted in as @xmath205 . to solve in practice , the orbital angular momentum expansions in ( [ eq : tailexp ] ) and ( [ eq : lcmto ] ) , which are in - principle infinite , must be truncated by introducing some cutoff in @xmath30 , denoted @xmath206 .
usually , a value of @xmath207 or @xmath208 is used , corresponding to _ spd_- or _ spdf_-bases .
( in units of @xmath209 ) for bulk _ fcc _ cu and co ( majority and minority spin ) plotted as a function of the normalized area element used in the brillouin zone summation , @xmath210 .
@xmath211 , the number of intervals along the reciprocal lattice vector is indicated at the top of the figure .
squares represent the series ( @xmath212 ) least - squares fitted by the dashed line ; diamonds , the series ( @xmath213 ) least - squares fitted by the dash - dotted line .
the part of the curve for the co minority spin case to the left of the vertical dotted line is shown on an expanded scale in the inset .
lattice constant of @xmath214 and _ spd _ basis were used together with von barth - hedin s exchange - correlation potential . ]
the @xmath111 summation is carried out by sampling , on a regular mesh , the 2d brillouin zone ( bz ) defined by the ( lateral ) translational periodicity perpendicular to @xmath168 .
the results of carrying out this bz summation are shown in fig .
[ fig : lead_conv ] where @xmath215 is plotted as a function of @xmath216 , the normalized area element per @xmath111-point for bulk _ fcc _ cu and for the majority and minority spins of bulk _ fcc _ co. when the 2d - bz reciprocal lattice vectors are each divided into @xmath211 intervals , then @xmath217 .
it can be seen that the sharvin conductance is converged to about 1% if @xmath218 points are used in the complete 2d - bz and to about 0.2% for @xmath219 sampling points .
the worst case is for the minority spin of co which has a complex multi - sheeted fermi surface . to see if there are any simple underlying trends in the convergence , we repeatedly bisect the intervals used in the bz summation starting with @xmath220 and @xmath221 , shown in the figure as squares and diamonds , respectively and least - squares fitted with the dashed and dash - dotted lines .
the convergence is fairly uniform but not very systematic indicating that the summation is limited by fine structure in the integrand at the smallest length scale studied which can only be resolved by increasingly fine sampling .
thus there is nothing to be gained by developing more sophisticated interpolation schemes and when we introduce disorder in section [ ssec : intdis ] , this will be even more so .
however , in the following we will see that the level of convergence we can achieve with discrete sampling is quite adequate and not a limiting step in the whole procedure .
the calculations shown in the figure were performed using an @xmath222-basis , for an _ fcc _ lattice constant @xmath223 corresponding to the experimental volume of bulk ( _ fcc _ ) cu and using the exchange - correlation potential calculated and parameterized by von barth and hedin.@xcite for convenience , and to avoid repetition , we will refer to this in the following as a `` standard '' configuration .
the converged values are given ( underlined ) in table [ tab : a ] together with values calculated using an _ fcc _ lattice constant @xmath224 corresponding to the volume of bulk _ hcp _ co.@xcite because we shall be studying cu / co interfaces where the volume per atom is not known very precisely from experiment , we will want to estimate the variation that can be expected when different but equally reasonable lattice constants are used .
the increase of 3.4% ( from 0.558 to @xmath225 ) observed for cu can be attributed to the increased areal density of cu atoms , @xmath226 corresponding to @xmath227 . the table also contains the corresponding results obtained with an _ spdf_-basis . to the numerical accuracy shown ,
there is no difference between the _ spd _ and _ spdf _ case for cu . for co majority spin states ,
there is a 4% decrease in the conductance on going from an _ spd _ to an _ spdf _ basis . for a lattice constant @xmath228 , the magnetization is 1.684@xmath229/atom for an _ spd_- and 1.648@xmath229/atom for an _ spdf _ basis corresponding , respectively , to @xmath230 5.342 and 5.324 electrons in the majority spin bands .
since all five ( nominal ) majority - spin @xmath18 bands are full there are 0.342 and 0.324 electrons in the free - electron - like @xmath231 band . in a free electron picture
the ratio of the projection of the spherical fermi surfaces is @xmath232 , thus explaining the observed numerical result .
lcllll & & & & + & @xmath233( ) & basis & @xmath234 & present calc . & schep + copper & 3.549 & @xmath222 & 5.5 & 0.577(0.577,0.577 ) & 0.57 + & 3.549 & @xmath235 & 5.5 & 0.577(0.577 ) & - + & 3.614 & @xmath222 & 5.5 & ( 0.559 ) & 0.55 + & 3.614 & @xmath235 & 5.5 & 0.558(0.558 ) & 0.55 + + cobalt & 3.549 & @xmath222 & 5.323 & 0.469(0.459,0.467 ) & 0.45 + majority & 3.549 & @xmath235 & 5.304 & 0.449(0.440 ) & 0.43 + & 3.614 & @xmath222 & 5.342 & ( 0.457 ) & 0.45 + & 3.614 & @xmath235 & 5.324 & 0.448(0.439 ) & - + +
cobalt & 3.549 & @xmath222 & 3.677 & 1.082(1.081,1.082 ) & 1.10 + minority & 3.549 & @xmath235 & 3.696 & 1.120(1.125 ) & 1.13 + & 3.614 & @xmath222 & 3.658 & ( 1.047 ) & 1.06 + & 3.614 & @xmath235 & 3.676 & 1.074(1.079 ) & - + the co majority - spin conductance scarcely changes with changing lattice constant , however .
the origin of this behaviour lies in the volume dependence of the magnetic moment .
when the lattice constant is decreased , the @xmath18 bands broaden and the magnetic moment decreases from 1.684 to 1.646@xmath229/atom in the @xmath222 case with a corresponding decrease of the occupancy of the @xmath231 band from 0.342 to 0.323 majority - spin electrons .
the corresponding 4% decrease in conductance is almost perfectly compensated by the increased areal density of atoms so there is no net change . for the minority - spin conductance ,
the same factors play a role but now the @xmath18 bands are only partly filled .
this results in complex fermi surfaces for which simple estimates can not be made . in this case recourse
must be made to full band structure calculations .
we return to this in sect .
[ ssec : ordint ] .
the calculations presented so far were carried out using the exchange - correlation potential calculated and parameterized by von barth and hedin.@xcite this is only one of a number of potentials we could have used , none of which is clearly better than the others in describing the ground state properties of magnetic materials . to gauge the uncertainty arising from this arbitrary choice ,
a number of calculations were carried out using the potentials given by perdew - zunger@xcite and vosko - wilk - nusair@xcite and the results are given in brackets in the table .
using different exchange - correlation potentials leads to variation in the conductances of the order of 1 or 2 percent .
a different ( but equivalent ) approach was adopted by schep _ _ et al.__@xcite to the determination of the sharvin conductances for the same systems using conventional first - principles lmto - asa bulk electronic band structures , i.e. using @xmath236 rather than @xmath237 as used here .
he expressed the sharvin conductance as a projection of the fermi surface onto a plane perpendicular to the transport direction and calculated the areas using a suitably modified 3d - bz integration scheme .
his results are also given in table [ tab : a ] and are as consistent with our present values as can be expected when using two entirely different computer codes .
in determining the conductance of the leads , the bz summation does not present a problem .
the uncertainties arising from small variations in the atomic volumes , from incompleteness of the basis and from the choice of lda parameterization are of comparable size .
the mto - as approximation can be systematically improved but only at substantial computational cost . since there is currently no way to systematically improve upon the lda we identify it and the lack of knowledge of the atomic structure as limiting factors in studying transport from first principles . though the atomic structures could be determined theoretically by total energy minimization , the lda again
presents a barrier since it systematically underestimates lattice constants of transition metals in particular of the 3@xmath18 series .
gradient corrections sometimes yield improvements but unfortunately not systematically so .
we conclude that our knowledge of and ability to calculate from first principles fermi surfaces for bulk magnetic materials such as fe or co does not at present justify using a more accurate but substantially more expensive computational scheme than the present one .
cu and co have slightly different atomic volumes .
the equilibrium lattice constant of cu is @xmath238 and of co @xmath239 , assuming an _ fcc _ structure . even in the absence of interface disorder
, the lattice spacing will not be homogeneous and will depend on the lattice constant of the substrate on which the sample was grown , on the global and local concentrations of cu and co , and on other details of how the structure was prepared . in principle
we could calculate all of this by energy minimization .
however , we judge that the additional effort needed is not justified by current experimental knowledge . instead , we content ourselves with estimating the uncertainty which results from plausible variations in the ( interface ) structure by considering two limiting cases and one intermediate case . in each case
an _ fcc _ structure is assumed , with lattice constants corresponding to ( i ) the atomic volume of cu , ( ii ) the atomic volume of co , ( iii ) an intermediate case with arithmetic mean of cu and co atomic volumes .
d .. 5 d .. 5 d .. 5 & & & + & & & & + @xmath240(bulk ) & 0.000 & 0.000 & 0.000 & 0.000 + @xmath240(int-4 ) & 0.001(1 ) & 0.001 & 0.001 & 0.001 + @xmath240(int-3 ) & -0.001(0 ) & 0.000 & 0.000 & 0.000 + @xmath240(int-2 ) & -0.005(5 ) & -0.005(4,5 ) & -0.005(4 ) & -0.005 + @xmath240(int-1 ) & 0.002(4 ) & 0.004(6,4 ) & 0.003(4 ) & 0.001(2 )
+ @xmath241(int+1 ) & 1.526(490 ) & 1.578(45,73 ) & 1.605(573 ) & 1.636(01 ) + @xmath241(int+2 ) & 1.621(597 ) & 1.656(35,53 ) & 1.673(53 ) &
1.690(70 ) + @xmath241(int+3 ) & 1.602(576 ) & 1.645(21,41 ) & 1.662(39 ) & 1.680(59 ) + @xmath241(int+4 ) & 1.610(587 ) & 1.649(27,45 ) & 1.665(45 ) &
1.683(62 ) + @xmath241(bulk ) & 1.609(590 ) & 1.646(22,42 ) & 1.667(45 ) & 1.684(62 ) + @xmath242 & 0.409(399 ) & 0.431(21,29 ) & 0.433(22 ) & 0.434 ( 24 ) + @xmath243 & 0.378(379 ) & 0.378(80,79 ) & 0.371(73 ) & 0.364 ( 67 ) our starting point is a self - consistent tb - lmto sgf calculation@xcite for the interface embedded between semi - infinite cu and co leads whose potentials and spin - densities were determined self - consistently in separate `` bulk '' calculations .
the charge and spin - densities are allowed to vary in @xmath244 layers of cu and @xmath245 layers of co bounding the interface .
the results of these calculations for cu / co(111 ) interfaces and the three different lattice constants detailed above are given in table [ tab : b ] for @xmath244=4 , @xmath245=4 . in the cu layers ,
only tiny moments are induced .
only four layers away from the interface on the co side , the magnetic moments are seen to be very close to the bulk values . at the interface , where the @xmath18-bandwidth is reduced as a result of the lower coordination number , the moments are _ suppressed _ rather than enhanced .
this occurs because the majority - spin @xmath18 bands are full and their number can not increase .
the width of the free - electron like @xmath231 band is less sensitive to the change in coordination and its exchange splitting also changes less . as a result
, there is little change in the @xmath231 moment .
when the @xmath18-bandwidth is reduced , there is conversion of minority- and majority - spin @xmath231 electrons , without loss of the @xmath231 moment , to the minority - spin @xmath18 band with loss of @xmath18 moment .
this picture is supported by the full calculations .
earlier we saw that an @xmath246 change in lattice constant changed the bulk magnetic moment of _ fcc _ co by @xmath247 .
the effect of changing the basis , from _ spd _ to _ spdf _ , was similar . from table [
tab : b ] , the interface moments are seen to behave in a comparable fashion .
the magnetic moment of the interface co atoms decreases by @xmath248 , from 1.636@xmath229/atom for @xmath228 to 1.578@xmath229/atom for @xmath224 for an @xmath222 basis and decreases from 1.578@xmath229/atom to 1.526@xmath229/atom for an @xmath235 basis for @xmath224 , a change of @xmath249 .
thus the @xmath231 to @xmath250 conversion is enhanced at the interface by the reduced @xmath18-bandwidth .
( in units of @xmath209 ) for an _ fcc _ cu / co(111 ) interface for majority and minority spins plotted as a function of the normalized area element used in the brillouin zone summation , @xmath210 .
@xmath211 , the number of intervals along the reciprocal lattice vector is indicated at the top of the figure .
squares represent the series ( @xmath251 ) least - squares fitted by the dashed line ; diamonds , the series ( @xmath213 ) least - squares fitted by the dash - dotted line .
the part of the curve for the co minority spin case to the left of the vertical dotted line is shown on an expanded scale in the inset .
a `` standard '' configuration was used.@xcite ] once the interface potential has been obtained , the transmission matrix can be calculated and the bz summation carried out .
the convergence of this summation , shown in fig .
[ fig : cl_int_conv ] for a lattice constant of @xmath228 and an _ spd _ basis closely parallels that seen in fig .
[ fig : lead_conv ] and therefore the k - summation does not represent a limitation in practice . converged conductances @xmath252 are given in the last two rows of table [ tab : b ] .
though we will not concern ourselves in this publication with the application of the formalism we have been developing to a detailed interpretation of experimental observations , it should be noted that even a modest spin - dependence of `` bare '' interface conductances ( @xmath253 ) can lead to spin - dependent interface resistances differing by a factor of @xmath254 once account is taken of the finiteness of the conductance of the perfect leads using a formula derived by schep _
et al._@xcite @xmath255 \label{eq : r_schep}\ ] ] where @xmath256 and @xmath257 are the sharvin conductances ( see eq . ) of the materials a and b forming the interface , in units of @xmath258 .
the majority - spin case can be readily understood in terms of the geometry of the fermi surfaces of cu and co so we begin by discussing this simple case before examining the more complex minority - spin channel . in the absence of disorder , crystal momentum parallel to the interface
is conserved .
if , for a given value of @xmath259 , there is a propagating state in cu incident on the interface but none in co , then an electron in such a state is completely reflected at the interface .
conversely , @xmath259
s for which there is a propagating state in co but none in cu also can not contribute to the conductance . to determine the existence of such states , it is sufficient to inspect projections of the fermi surfaces of _ fcc _ cu and majority - spin co onto a plane perpendicular to the transport direction @xmath260 , shown in fig .
[ fig : cuco111_maj ] for @xmath261 .
the first feature to note in the figure ( left - hand and middle panels ) is that per @xmath259 there is only a single channel with positive group velocity so that the transmission matrix in is a complex number whose modulus squared is a transmission probability with values between 0 and 1 .
it is plotted in the right - hand panel and can be interpreted simply .
regions which are depicted blue correspond to @xmath259 s for which there are propagating states in cu but none in co. these states have transmission probability 0 and are totally reflected . for values of @xmath259
for which there are propagating states in both cu and co , the transmission probability is very close to one , depicted red .
these states are essentially free electron - like states which have the same symmetry in both materials and see the interface effectively as a very low potential step .
close to the centre of the figure there is an annular region where there are propagating states in co but none in cu so they do not contribute to the conductance .
performing the sum in , we arrive at an interface conductance of @xmath262 to be compared to the sharvin conductances given in table [ tab : a ] for cu and co ; for @xmath228 and an @xmath222 basis these are , respectively , 0.558 and 0.466 in the same units .
the interface conductance of 0.434 is seen to be essentially the sharvin conductance of the majority states of co reduced because the states closest to the @xmath263-axis ( corresponding to the symmetry axis of the figures , the @xmath264 line in reciprocal space ) do not contribute .
the explanation of the 5% decrease found on going from an @xmath222 to an @xmath235 basis , ( 0.431 to 0.409 ) , parallels that given for the corresponding change in the sharvin conductance of bulk co ( 0.469 to 0.449 in table [ tab : a ] ) . for an _ fcc _ cu / co(111 ) interface .
a `` standard '' configuration was used.@xcite ] the minority - spin case is considerably more complex because the co minority - spin @xmath18 bands are only partly filled , resulting in multiple sheets of fermi surface .
these sheets are shown in fig .
[ fig : cuco111_min ] together with their projections onto a plane perpendicular to the ( 111 ) transport direction . compared to fig .
[ fig : cuco111_maj ] , one difference we immediately notice is that even single fermi surface ( fs ) sheets are not single valued : for a given @xmath259 there can be more than one mode with positive group velocity .
the areas depicted green in the projections of the fs sheets from the fourth and fifth bands are examples where this occurs .
an electron incident on the interface from the cu side , with transverse crystal momentum @xmath259 , is transmitted into a linear combination of all propagating states with the same @xmath259 in co ; the transmission matrix @xmath265 is in general not square but rectangular .
the transmission probabilities @xmath266 are shown in the bottom row of fig .
[ fig : cuco111_min ] . because there is only a single incident state for all @xmath259 ,
the maximum transmission probability is one .
comparison of the total minority - spin transmission probability @xmath267 ( fig .
[ fig : cuco111_min ] , bottom right - hand panel ) with the corresponding majority - spin quantity ( right - hand panel of fig .
[ fig : cuco111_maj ] ) strikingly illustrates the spin - dependence of the interface scattering , much more so than the integrated quantities might have led us to expect ; the interface conductances , 0.364 and @xmath262 from table [ tab : b ] , differ by only @xmath268 .
three factors contribute to the large @xmath259-dependence of the transmission probability : first and foremost , the complexity of the fermi surface of both materials but especially of the minority spin of co ; secondly and inextricably linked with the first because of the relationship @xmath269 , the mismatch of the fermi velocities of the states on either side of the interface .
thirdly , the orbital character of the states @xmath38 and @xmath270 which varies strongly over the fermi surface and gives rise to large matrix element effects . the great complexity of transition metal fermi surfaces , clear from the figure and well - documented in standard textbooks , is not amenable to simple analytical treatment and has more often than not been neglected in theoretical transport studies .
nevertheless , as illustrated particularly well by the ballistic limit,@xcite spin - dependent band structure effects have been shown to lead to magnetoresistance ratios comparable to what are observed experimentally in the current - perpendicular - to - plane ( cpp ) measuring configuration and can not be simply ignored in any quantitative discussion .
most attempts to take into account contributions of the @xmath18 states to electronic transport do so by mapping the five @xmath18 bands onto a single tight - binding or free - electron band with a large effective mass .
fermi surface topology alone can not explain all aspects of the tranmission coefficients seen in fig .
[ fig : cuco111_min ] . for example
, there are values of @xmath271 , such as that labelled @xmath272 in the figure , for which propagating solutions exist on both sides of the interface yet the transmission probability is zero .
this can be understood as follows . at @xmath273 , the propagating states in cu
have \{@xmath274 } character ( assuming the choice of in - plane axes as illustrated in the top righthand panel of fig .
[ fig : cuco111_min ] ) and are even with respect to reflection in the plane formed by the @xmath275-axis and the transport direction perpendicular to the ( 111 ) plane which we choose to be the @xmath276-axis . for this @xmath271
the only propagating state in co is in the fourth band .
it has \{@xmath277 } character which is odd with respect to reflection in the @xmath278 plane .
consequently , the corresponding hopping matrix elements in the hamiltonian ( and in the green s function ) vanish and the transmission is zero . along the @xmath279 axis
the symmetry of the states in cu and those in the fourth band of co remain the same and the transmission is seen to vanish for all values of @xmath279 .
however , at points further away from @xmath263 , we encounter states in the third band of co which have even character whose matrix elements do not vanish by symmetry and we see substantial transmission probabilities .
similarly , for points closer to @xmath263 , there are states in the fifth band of co with even character whose matrix elements also do not vanish and again the transmission probability is substantial . because it is obtained by superposition of transmission probabilities from cu into the third , fourth and fifth sheets of the co fs , the end result , though it may appear very complicated , can be straightforwardly analysed in this manner k - point by k - point . though the underlying lattice symmetry is only threefold , the fermi surface projections shown in fig .
[ fig : cuco111_min ] have six - fold rotational symmetry about the line @xmath263 because the bulk @xmath280 structure has inversion symmetry ( and time - reversal symmetry ) .
the interface breaks the inversion symmetry so @xmath281 has only threefold rotation symmetry for the individual fs sheets .
however , in - plane inversion symmetry is recovered for the total transmission probability @xmath282 which has full sixfold symmetry .
this follows from the time - reversal symmetry and is proven in appendix [ sec : symmetry ] .
instructive though the study of perfect interfaces may be in gaining an understanding of the role electronic structure mismatch may play in determining giant magnetoresistive effects , all measurements are made on devices which contain disorder , mostly in the diffusive regime .
because there is little information available from experiment about the nature of this disorder , it is very important to be able to model it in a flexible manner , introducing a minimum of free parameters . to model interfaces between materials with different lattice constants and disorder
, we use the lateral supercells described in section [ ssec : supercells ] .
since this approach is formally only valid if sufficiently large supercells are used , we begin by studying how the interface conductance depends on the lateral supercell size . ) for a disordered cu / co ( 111 ) interface modelled as 2ml of 50%-50% alloy in a @xmath283 lateral supercell as a function of @xmath284 .
the results are given for different randomly generated configurations of disorder ( 15 for minority spin , 5 for majority spin ) .
results are for a `` standard '' configuration.@xcite ] to perform fully self - consistent calculations for a number of large lateral supercells and for different configurations of disorder would be prohibitively expensive . fortunately , the coherent potential approximation ( cpa ) is a very efficient way of calculating charge and spin densities for a substitutional disordered @xmath285 alloy with an expense comparable to that required for an ordered system with a minimal unit cell.@xcite the output from such a calculation are atomic sphere potentials for the two sites , @xmath286 and @xmath287 .
the layer cpa approximation generalizes this to allow the concentration to vary from one layer to the next .
@xcite once @xmath286 and @xmath287 have been calculated for some concentration @xmath288 , an @xmath289 lateral supercell is constructed in which the potentials are distributed at random , maintaining the concentration for which they were self - consistently calculated .
the conductances calculated for @xmath290 are shown in fig .
[ fig : config_av ] for a cu / cu(111 ) interface in which the cu and the co layers forming the interface are totally mixed to give two layers of 50%-50% interface alloy . for each value of @xmath1 ,
the results for a number of different randomly generated disorder configurations are shown ( 20 for minority , 5 for minority spin ) .
the sample to sample variation is largest for the minority spin case , ranging from @xmath291 for a modest @xmath169 unit cell and decreasing to less than @xmath292 for a @xmath293 unit cell . for @xmath294 , the spread in minority spin conductances is @xmath295 which is comparable to the typical uncertainty we associated with the lda error , the uncertainty in lattice constants or the error incurred by using the asa .
comparing now the conductances without and with disorder , we see that interface disorder has virtually no effect on the majority spin channel ( 0.434 versus @xmath296 ) which is a consequence of the great similarity of the cu and co majority spin potentials and electronic structures . however , in the minority - spin channel the effect ( 0.364 versus @xmath297 ) is much larger .
as noted in the context of , a relatively small change in the interface transmission can lead to a large change in the interface resistance when account is taken of the finite conductance of the leads .
we will return to the consequences for the spin - dependent interface resistance after completing the study of the interface transmission on which it is based .
when disorder is modelled in lateral supercells , the transmission probabilities can be classified as being _
specular _ or _ diffuse _ depending upon whether or not transverse momentum is conserved.@xcite in the presence of interface disorder , the conductance per unit area can be expressed as @xmath298where @xmath299 and @xmath300 belong to the two dimensional brillouin zone for ( @xmath301 ) translational periodicity and @xmath302 .
the transmission matrix elements between two bloch states with the same @xmath98 are defined to be specular , those between scattering states with different @xmath98 as being diffuse . in the absence of interface disorder
, there is by definition only a specular component . the results in fig .
[ fig : config_av ] were obtained for a structural model of the co / cu(111 ) interface consisting of two monolayers ( 2ml ) of 50%-50% alloy that was derived from x - ray@xcite nmr@xcite and magnetic exafs@xcite studies .
though the most plausible model there is at present , it contains large uncertainties .
this makes it important to explore the consequences of varying the parameters defining the model . to do so
, we calculate the conductance using 20x20 lateral supercells as a function of alloy concentration for models in which the disorder is confined to one , two or four monolayers .
the three models are defined in fig .
[ fig : disord ] . from the results shown in fig .
[ fig : spec_diff ] , it can be seen that the interface transmission for majority - spin electrons depends only very weakly on alloy concentration and its spatial distribution .
the results for the 1ml , 2ml and 4ml models can not be distinguished on the scale of the figure .
when the conductance is decomposed using , the diffuse component is found to be very small .
therefore , only the results for the minority - spin case need be examined in any detail .
co@xmath303 alloy between cu and co leads , denoted as cu@xmath304cu@xmath305co@xmath306$]co .
middle ( 2ml ) : disorder modelled in two mls as cu@xmath304cu@xmath305co@xmath307cu@xmath308co@xmath309$]co .
bottom ( 4ml ) : starting from the 2 ml disorder case , 1/3 of the concentration @xmath288 of impurity atoms is transferred to the next layer resulting in disorder in four mls : cu@xmath304cu@xmath310co@xmath311cu@xmath312co@xmath313cu@xmath314co@xmath315cu@xmath316co@xmath317$]co . ] we start by varying the alloy concentration over the full concentration range ( 0 - 100% ) for a disordered monolayer .
the magnitude of resulting variation in transmission is limited ( @xmath318 ) but exceeds the statistical spread between different configurations of disorder , which according to fig .
[ fig : config_av ] is less than @xmath292
. on adding co to a layer of cu , the transmission decreases , reaches a minimum for 10% co , then increases monotonically up to 70 - 90% region where the transmission is _ higher _ than for a clean interface.@xcite 100% co represents a clean interface again , so this limit must yield the same transmission as 0% co. the variation can be examined in terms of the specular and diffuse components defined in . from fig .
[ fig : spec_diff ] , it can be seen that , for the minority spin channel , the diffuse scattering by co impurity atoms in cu is stronger than that by cu impurity atoms in co. however , the specular scattering is also more strongly reduced by cu in co than by co in cu .
the two effects largely cancel resulting in the observed undulatory total transmission as a function of the alloy concentration seen in the figure .
the diffuse scattering has a maximum close to a 50%-50% alloy concentration where its contribution to the conductance is almost twice as large as from the specular scattering .
while the conductance as such is scarcely effected , the strong diffuse scattering will play an important role in destroying the phase coherence of the electrons .
ultimately , this will be the physics underlying the so - called `` two current series resistor '' ( 2csr ) model.@xcite lateral supercell as a function of the alloy concentration @xmath288 .
results are shown for the three different models described in fig .
[ fig : disord ] with disorder in 1 , 2 or 4 mls . only a single disorder configuration was used and the size of the symbols corresponds to the spread in values found for this supercell size in fig .
[ fig : config_av ] . for 1ml , the total conductance is resolved into specular and diffuse components ( see the legend ) .
results are for a `` standard '' configuration.@xcite ] if the disorder extends over more than a monolayer , then modelling the interface as several layers of homogeneous alloy is not obviously realistic .
instead , one might expect the layers closest to the interface to be most strongly mixed , the amount of mixing decreasing with the separation from the interface . a simple way to model
this is to take two interface layers , one cu and one co , and to mix them in varying degrees .
denoting this cu@xmath319co interface as cu@xmath304cu@xmath305co@xmath307cu@xmath308co@xmath309$]co we consider @xmath320 _ i.e. _ , the cu concentration decreases monotonically from left to right .
the calculated interface transmission is seen to essentially interpolate linearly the results obtained previously for the clean ( @xmath321 ) and disordered ( @xmath322 ) cases . a slightly more elaborate model can be constructed from the 2ml model by distributing the @xmath288 impurity atoms so that @xmath323 are in the interface layer while @xmath324 are to be found further from the original interface , in the following layer .
this results in the concentration profile cu@xmath304cu@xmath310co@xmath311cu@xmath312co@xmath313cu@xmath314co@xmath315cu@xmath316co@xmath317$]co .
@xmath321 corresponds to a completely ordered interface while the maximum value @xmath288 can have so that the concentration decreases from left to right monotonically is 75% .
this relatively small redistribution of intermixed atoms is seen to reduce the transmission by 15% for @xmath322 .
even for very good metals , relatively opaque interfaces can result when the electronic structure on either side have different characters .
in such situations , disorder can influence the transmission strongly even reducing the interface resistance very substantially.@xcite a detailed analysis of the different contributions to the interface scattering in the 2ml and 4ml cases will be given in a separate publication .
lateral supercell .
the dark red point in the cu fermi surface projection corresponds to the point @xmath272 in the top righthand panel of fig .
[ fig : cuco111_min ] .
@xmath325 is shown in ( c ) , and in ( d ) magnified by a factor 500 where the ballistic component @xmath326 is indicated by a white point because its value goes off the scale .
results are for a `` standard '' configuration@xcite and averaged over 5 different configurations of disorder . ] the scattering induced by two layers of 50%-50% alloy is illustrated in fig .
[ fig : dis_scatt_maj ] and fig .
[ fig : dis_scatt_min ] for the majority and minority spins , respectively , of a cu / co(111 ) interface .
calculations were performed for the single @xmath327 point , @xmath328 , and a @xmath329 lateral supercell equivalent to using a @xmath330 interface cell and k - space sampling with @xmath293 points in the corresponding bz .
disorder averaging was carried out using 5 ( for majority spin ) or 20 ( for minority spin ) disorder configurations generated randomly .
[ fig : dis_scatt_maj](a ) and [ fig : dis_scatt_maj](b ) show the majority - spin fermi surface projections of _ fcc _ cu and co , respectively , obtained from `` unfolding '' the supercell calculation .
the coarse @xmath293 grid is seen to yield a good representation of the detailed fermi surface projections shown in fig .
[ fig : cuco111_maj ] .
@xmath331 is shown in fig .
[ fig : dis_scatt_maj](c ) for @xmath332 on the @xmath279 axis in fig .
[ fig : cuco111_min ] .
specular scattering dominates with @xmath333 .
the diffuse scattering is so weak that nothing can be seen on a scale of _ t _ from 0 to 1 . to render it visible , a magnification by a factor 500 is needed , fig .
[ fig : dis_scatt_maj](d ) . the total diffuse scattering
, @xmath334 can be seen from the figure to be made up of contributions of @xmath335 from roughly a quarter of the bz ( 100 @xmath336 points ) centred on @xmath337 .
the total transmission , @xmath338 , compared to a transmission of 0.99 in the absence of disorder .
similar results were obtained for other @xmath339 points . in the majority case , there is thus a strong specular peak surrounded by a weak diffuse background .
lateral supercell .
the dark red point in the cu fermi surface projection corresponds to the point @xmath340 in the top righthand panel of fig .
[ fig : cuco111_min ] .
( c ) @xmath341 and ( d ) @xmath342 calculated using 20 different disorder configurations ; the ballistic component @xmath343 is indicated by a white point because its value goes off scale .
results are for a `` standard '' configuration,@xcite 20 different configurations of disorder were used . ] the minority - spin fermi surface projections of _ fcc _ cu and co are shown in figs .
[ fig : dis_scatt_min](a ) and [ fig : dis_scatt_min](b ) , respectively . compared to the corresponding panels in fig .
[ fig : cuco111_min ] , the @xmath293 point representation is seen to be sufficient to resolve the individual fermi surface sheets of co. to study the effect of interface disorder , we consider scattering out of two different @xmath98s in cu ( figs .
[ fig : dis_scatt_min](c ) and ( d ) ) .
the first thing to note is the similarity of both transmission plots to the projected fs of co , fig .
[ fig : dis_scatt_min](b ) , suggesting very strong diffusive scattering proportional to the density of available final states .
the first case we consider is where @xmath344 for which the transmission was zero as a result of the symmetry of the states along the @xmath279 axis in the absence of disorder .
@xmath345 is shown in fig .
[ fig : dis_scatt_maj](c ) .
by contrast with the majority - spin case just examined , there is now scattering to all other k - points in the 2d bz , @xmath346 while @xmath347 has only increased from 0.00 in the clean case , to 0.01 in the presence of disorder .
the effect of disorder is to increase the total transmission , @xmath348 from 0.00 to @xmath349 ; for states which were originally strongly reflected , disorder _ increases _ the transmission .
the second case we consider is that of a k - point slightly further away from the origin @xmath263 along the @xmath279 axis which had a high transmission , @xmath350 , in the absence of disorder . for this k - point , @xmath351 , shown in fig .
[ fig : dis_scatt_maj](d ) , looks very similar to fig .
[ fig : dis_scatt_maj](c ) .
there is strong diffuse scattering with @xmath352 while @xmath353 has been drastically decreased from 0.98 in the clean case , to 0.06 as a result of disorder .
the total transmission , @xmath354 , is almost identical to what was found for the @xmath272 point .
the effect of disorder has been to _ decrease _ the transmission for states which were originally weakly reflected .
the strong k - dependence of the transmission found in the specular case is largely destroyed by a small amount of disorder in the minority - spin channel . the contribution from specular component ( integrated over 2d bz )
is reduced to 15% of the total transmission . to the best of our knowledge
, spin - dependent interface transmissions have not yet been measured directly .
what is usually done @xcite is to measure total resistances for a whole series of magnetic multilayers in which the total number of interfaces and/or the thicknesses of the individual layers is varied .
the measured results are interpreted in terms of volume resistivities and interface resistances . by applying an external magnetic field , the magnetizations of neighbouring layers which are oriented antiparallel ( ap )
can be forced to line up in parallel ( p ) . by measuring the resistances in both cases , spin - dependent volume resistivities and interface resistances
can be extracted using the two current series resistor model.@xcite if we take expression which relates the interface transmission to the interface resistance occurring in the 2csr model as given,@xcite we can study how typical uncertainties in interface transmission , arising from arbitrary assumptions about the interface disorder , lattice constant or basis set translate into uncertainty in predicted interface resistances . using the transmission probabilities from fig .
[ fig : spec_diff ] in results in the curves shown in fig .
[ fig : int_res ] . for comparison ,
a range of literature values for the spin - dependent interface resistances derived from experiments on sputtered and mbe ( molecular beam exitaxy ) grown multilayers@xcite is included in the figure .
.interface resistances , in units of f@xmath355 , for ordered interfaces , calculated using expression and the data from tables [ tab : a ] and [ tab : b ] .
the values given here for a lattice constant of @xmath214 differ slightly from those reported in ref . which were performed using energy - independent muffin - tin orbitals linearized about the centers of gravity of the occupied conduction states and not at the fermi energy . the current implementation@xcite uses energy - dependent , ( non - linearized ) mto s , calculated exactly at the fermi energy which improves the accuracy at no additional cost .
[ cols="<,^,^,^ " , ] for the minority - spin case , experimental values ( in units of f@xmath355 ) range from 1.30 - 1.80 compared to calculated values of 1.29 for cu@xmath304cu@xmath356co@xmath357$]co , through 1.37 for a disorder - free interface , to a value of 2.25 for the 4ml model with @xmath322 , cu@xmath304cu@xmath358co@xmath359cu@xmath360co@xmath361cu@xmath362co@xmath363cu@xmath364co@xmath365$]co .
the influence of lattice constant and basis set on the clean interface resistance values is small ( see table [ tab : c ] ) .
the present modelling of interface alloying shows that the interface resistance is more strongly dependent on the detailed spatial distribution of disorder than was previously found@xcite where only the concentration range @xmath366 of the 2ml interface alloy model extracted from experiment@xcite was explored . .
the experimental values for sputtered and mbe grown multilayers cited in table i of ref .
span a range of values which is indicated by the shaded regions . ] for the majority - spin case , the spread in values of the interface resistance extracted from experiment ( for the same samples as for the minority - spin case ) is quite small , 0.22 - 0.25 , and does not overlap with the values of @xmath367 found for a lattice constant of @xmath214 . unlike the minority - spin case , changing the lattice constant or using an @xmath235 basis leads to substantially _ larger _ values ( table [ tab : c ] ) . because the majority - spin transmission does not depend on the details of the interface disorder , this can not be the origin of the discrepancy .
motivated by the weak scattering in this case , we examine the validity @xcite of the 2csr model by calculating the resistance of a magnetic multilayer containing a large number of disordered interfaces and plot the resistance added by each additional interface in fig .
[ fig : diff_res ] . compared to similar calculations in ref .
, the number of interfaces , size of lateral supercell ( @xmath368 ) and disorder configurations averaged over are increased substantially . while the calculations are in very good agreement with ohm s law for the strongly scattering minority - spin case , it can be seen that this is not the case for the majority - spin electrons . for a small number of interfaces
there is a clear breakdown of ohm s law and thus of the 2csr model .
the interface resistance eventually saturates at a value much lower than those extracted from experiment . while inclusion of bulk scattering will modify this picture somewhat , exploratory calculations@xcite
indicate that the type of `` bulk '' impurities which may be reasonably expected to be found in sputtered or mbe grown multilayers affect the minority spin electrons much more than the majority spins .
agreement for the latter can only be achieved at the expense of ruining good agreement for the former .
lateral supercell was used and the interface was modelled as two layers of 50%-50% alloy ( 2ml model ) .
the results represent an average over 5 disorder configurations and were obtained for a `` standard '' configuration.@xcite the range of experimental values@xcite is indicated by the shaded regions . ]
details of a muffin tin orbital - based method suitable for calculating from first - principles scattering matrices involving layered magnetic materials have been given . in a wide range of applications,@xcite
it has been shown to be much more efficient and transparent than a previously used lapw - based method.@xcite various other schemes have been developed for calculating the transmission of electrons through an interface ( or a more extended scattering region ) both from first principles , or using as input electronic structures which were calculated from first principles .
most are based upon a formulation for the conductance in terms of non - equilibrium green s functions @xcite ( negf ) which reduces in the appropriate limit to the well known fisher - lee ( fl ) linear - response form@xcite for the conductance of a finite disordered wire embedded between crystalline leads .
most implementations of the negf or fl schemes have two disadvantages .
( i ) the transmission is calculated for a complex energy which leads to difficulties in studying for example , tunneling magnetoresistance , where the finite imaginary part can give rise to an exponential decay which obscures the interesting physical decay of the transmission as a function of the barrier thickness .
( ii ) for a given value of transverse crystal momentum , the transmission is expressed as a trace over the basis set in terms of which the green s function and self - energy are expressed.@xcite while this has the advantage that the total transmission can be calculated without explicitly determining the scattering states and can be computationally efficient , summation of the contributions from multiple scattering states can obscure real physical effects , for example , the role of the symmetries of individual scattering states seen in fig .
[ fig : cuco111_min ] . explicit determination of the scattering states not only makes a detailed analysis of the scattering possible .
the full scattering matrix , expressed in terms of the scattering states , can be used to bridge@xcite the gap between first - principles electronic structure calculations and phenomenological models of transport used to analyse complex situations where a full first - principles treatment is not practical .
we have instead made use of an alternative technique , suitable for hamiltonians that can be represented in tight - binding form , that was formulated by ando@xcite and is based upon direct matching of the scattering - region wave function to the bloch modes of the leads .
the relationship between the wave function matching @xcite and green function @xcite approaches is not immediately obvious .
it was suggested recently that wfm was incomplete @xcite but the equivalence of the two approaches could be proven.@xcite schemes similar in spirit to our own , but based upon empirical tight - binding hamiltonians have been presented by sanvito _
et al._@xcite and by velev.@xcite in contrast to these schemes , our tb - mto formalism is a parameter - free approach that has all of the advantages derived from self - consistent determination of potentials and spin densities for systems for which these are not known from experiment .
judging from the size of systems to which it has been applied , it would seem that our implementation is nevertheless substantially more efficient than these empirical schemes .
the scattering regions treated in figs .
[ fig : config_av],[fig : spec_diff],[fig : dis_scatt_maj ] and [ fig : dis_scatt_min ] contained as many as 3200 atoms ( @xmath293 lateral supercell @xmath136 8 principal layers where the potential was allowed to deviate from its bulk values ) or , in the case of fig .
[ fig : diff_res ] , @xmath369 atoms ( @xmath368 lateral supercell @xmath370 150 principal layers ) . our wfm scheme should not be confused@xcite with a recently developed transport formalim@xcite also based upon tb - lmtos but which makes use of the caroli nefg expression for the conductance in terms of a trace and a complex energy .
khomyakov and brocks@xcite have developed a scheme analogous to ours but based upon pseudopotentials and a real space grid which make it more suitable for studying quantum wires or the type of open structures studied in molecular electronics , but is computationally much more expensive . a third approach based upon
`` embedding''@xcite has been combined with full - potential linearized augmented plane wave method to yield what is probably the most accurate scheme to date @xcite but like the real space grid wfm method,@xcite these methods are numerically very demanding .
details of a wave - function matching method suitable for calculating the scattering matrices in magnetic metallic hybrid structures based upon first - principles tight - binding muffin tin orbitals have been given and illustrated with calculations for a variety of co / cu(111 ) interface - related problems .
the minimal basis of localized orbitals is very efficient , allowing large lateral supercells to be handled .
this allow us to model materials with large lattice mismatch or to study transport in the diffusive regime .
because the scattering states are calculated explicitly , the effect of various types of scattering can be analyzed in detail .
this work is part of the research program for the `` stichting voor fundamenteel onderzoek der materie '' ( fom ) and the use of supercomputer facilities was sponsored by the `` stichting nationale computer faciliteiten '' ( ncf ) , both financially supported by the `` nederlandse organisatie voor wetenschappelijk onderzoek '' ( nwo ) .
it was also supported by the european commission s research training network _ computational magnetoelectronics _
( contract no .
hprn - ct-2000 - 00143 ) as well as by the nedo international joint research program _
nano - scale magnetoelectronics_. mz wishes also to acknowledge support from kbn grant no .
pbz - kbn-044/p03 - 2001 .
we are grateful to : ilya turek for his tb - lmto - sgf layer code which we used to generate self - consistent potentials and for numerous discussions about the method ; anton starikov for permission to use his version of the tb - mto code based upon sparse matrix techniques to perform some of the calculations .
expressions for the velocities of the propagating modes in the leads ( sect .
[ ssec : leads ] ) are more easily derived using an energy independent hamiltonian than the energy dependent tail - cancellation condition of section [ ssec : mtos ] . to do so , we make use of the close relationship between the kkr tail - cancellation equation ( [ eq : tbpms ] ) and the linearized mto ( lmto ) hamiltonian , both of which can be expressed in terms of the hermitian matrix@xcite @xmath371^{-1/2 } \left(p^\alpha(\varepsilon)-s^\alpha \right ) [ \dot{p}^\alpha(\varepsilon)]^{-1/2 } \nonumber \\
= -p^\alpha(\varepsilon ) [ \dot{p}^\alpha(\varepsilon)]^{-1 } + [ \dot{p}^\alpha(\varepsilon)]^{-1/2 } s^\alpha [ \dot{p}^\alpha(\varepsilon)]^{-1/2}. \label{eq : halpha}\end{gathered}\ ] ] fixing the energy at @xmath372 and defining the potential parameters@xcite [ eq : potpar ] @xmath373^{-1/2 } \label{subeq:1}\ ] ] @xmath374 ( [ eq : halpha ] ) can be written as @xmath375 equation ( [ eq : tbham ] ) has the form of a two - center tight binding hamiltonian whose energy is given relative to @xmath376 .
it provides the lowest order approximation@xcite to the full lmto hamiltonian and yields eigenvalues correct to first order in @xmath377 . for eigenvalues equal to @xmath376 ,
it yields eigenvectors which are equal to those determined by the tail - cancellation condition ( [ eq : tbpms ] ) , up to a scaling factor @xmath378 . to calculate the group velocities of states precisely at the linearization energy , in the present case at the fermi energy , the first - order hamiltonian ( [ eq : tbham ] )
can be used since any error vanishes identically for @xmath379 . using the translational symmetry of the leads , the hamiltonian ( [ eq : tbham ] ) for bloch vector @xmath380 is @xmath381 where @xmath382 labels the sites and orbitals within the unit cell and @xmath383 runs over lattice vectors .
the energy eigenvalues @xmath384 are the expectation values @xmath385 where the eigenvectors @xmath386 are indexed by @xmath382 and we assumed normalization @xmath387 .
it is now straightforward to calculate the group velocity of the propagating mode @xmath388 in the mixed representation @xmath389 defined in section [ ssec : leads ] gives for the velocity in the stacking direction @xmath390 \label{eq : layered_velociy}\ ] ] where @xmath18 is the distance between equivalent monolayers in adjacent principal layers ( pl ) , the hopping is assumed ( as in section [ ssec : leads ] ) to extend only between neighbouring pls and @xmath391 with @xmath392 connecting equivalent sites in the neighbouring pls . using the definition of @xmath393 and recalling that the solutions @xmath394 of the tail - cancellation equation take implicitly into account the scaling factor @xmath378 we arrive at equation .
if we look closely at the transmission probabilities in fig . [ fig : cuco111_min ] , we see that the sheet resolved transmissions exhibit the geometrical symmetry of the underlying lattice ( _ i.e. _ the three - fold rotational axis ) . the total transmission probability on the other hand possesses an extra inversion symmetry , @xmath395 , which results in plots with a six - fold rotational axis .
this higher symmetry is the manifestation of the fundamental time - reversal symmetry obeyed in the absence of spin - orbit coupling and a magnetic field . in the case of the bulk system
time - reversal symmetry grants that for every eigenstate @xmath396 there exists the counterpart with the same energy and opposite wave vector ( _ i.e. _ @xmath397 ) and the wave functions are related by the complex conjugate .
the situation is more complicated in the case of the scattering state .
consider a state incoming from the left lead and scattered in the middle region .
the wave function consists then of the incoming and reflected states in the left lead @xmath398 and of the transmitted states in the right lead @xmath399 the time reversal operation transforms the above `` retarded '' state into the `` advanced '' one in which a number of incoming states ( from the left and the right ) combine to produce a single outgoing state on the left , _
i.e. _ @xmath400 and @xmath401 equations ( [ eq : left_adv ] ) and ( [ eq : right_adv ] ) impose a set of conditions on the values of scattering coefficients for the states with @xmath402 . combined with the analogous conditions derived for the states with the incoming state in the right lead , they are compactly expressed as
@xmath403 the scattering matrix @xmath404 is defined as @xmath405 where @xmath406 and @xmath407 are matrices in the space of the lead modes and the primed coefficients describe scattering of the states incoming from the right .
more specifically we have : @xmath408 equation gives @xmath409 in addition , for any two - terminal device , the hermiticity of the scattering matrix guarantees that @xmath410 ( see ref . ) which finally proves the in - plane inversion symmetry mentioned at the beginning . the last step can not however be taken for the partial ( fs resolved ) transmission probabilities .
these quantities thus possess only the geometrical symmetry of the system . | details are presented of an efficient formalism for calculating transmission and reflection matrices from first principles in layered materials . within the framework of spin density functional theory and using tight - binding muffin - tin orbitals ,
scattering matrices are determined by matching the wave - functions at the boundaries between leads which support well - defined scattering states and the scattering region .
the calculation scales linearly with the number of principal layers @xmath0 in the scattering region and as the cube of the number of atoms @xmath1 in the lateral supercell . for metallic systems for which the required brillouin zone sampling decreases as @xmath1 increases ,
the final scaling goes as @xmath2 . in practice ,
the efficient basis set allows scattering regions for which @xmath3 to be handled .
the method is illustrated for co / cu multilayers and single interfaces using large lateral supercells ( up to @xmath4 ) to model interface disorder . because the scattering states are explicitly found , `` channel decomposition '' of the interface scattering for clean and disordered interfaces can be performed . |
the newly discovered fe - based superconductors ( fescs ) share many similarities to the high @xmath15 cuprates,@xcite e.g. , both showing a relatively high superconducting transition temperature @xmath15 and possessing a layered crystal structure .
it is , therefore , natural to compare these two classes of superconductors , which might help unravel the puzzles of high @xmath15 superconductivity .
however , significantly distinct properties have been demonstrated in the fescs,@xcite including that ( i ) most of the parent compounds of fescs are typical ` bad metal ' instead of a mott insulator as found in the cuprates ; ( ii ) the fescs are a multi - band system , which seems to favor a @xmath16 pairing state rather than a _ d_-wave state ; ( iii ) both the fescs and the cuprates possess a very large upper critical field , but the fescs show nearly isotropic @xmath17 at low temperatures despite of their layered crystal structures .
clarification of the electronic coupling strength in fescs is the basis for establishing a pertinent theory of superconductivity .
various approaches , either based on the fermi surface nesting@xcite or started from the proximity to a mott - insulator@xcite were initially proposed to reveal the physics of iron pnictides , but no consensus has been reached .
recently , dual characters of localized and itinerant 3d - electrons were theoretically proposed@xcite and experimentally shown in some iron pnictides.@xcite to reveal the nature of magnetism and superconductivity in fescs and compare it with the high @xmath15 cuprates , it remains highly desired to search for fescs nearby a mott insulator .
very recently a new class of fescs , afe@xmath18se@xmath4 ( a = k,@xcite cs,@xcite rb,@xcite ( tl@xmath19k@xmath20)@xcite and ( tl@xmath19rb@xmath20)@xcite ) , were discovered with @xmath15 up to @xmath533k . remarkably different from the iron pnictides ,
superconductivity in iron selenides seems to develop from an antiferromagnetic mott insulator with a rather high nel temperature.@xcite in these compounds , one may tune the interplay of superconductivity and magnetism by changing the fe - vacancy order.@xcite furthermore , the reported arpes experiments on iron selenides showed that an isotropic superconducting gap emerges around the electron pocket at m point but the hole band centered at @xmath21 point sinks below the fermi level.@xcite this is in sharp contrast to that of the iron pnictide superconductors , in which both hole- and electron - pockets , connected with a nesting wave vector , were experimentally observed.@xcite it is , therefore , of great interest to find out whether the iron selenide superconductors represent a new type of fescs ( e.g. , similar to the high @xmath15 cuprates ) or remain similar to other iron pnictides . in any case , the iron selenide superconductors may provide an alternative example for studying the pairing mechanisms of high @xmath15 superconductivity , in particular for the fescs . to elucidate the above issues , it is highly important to compare the main superconducting parameters of the iron selenides with those of the iron pnictides , and also among the iron selenide series . in this article ,
we report measurements of the electrical resistivity in both a pulsed magnetic field and a dc magnetic field for the single - crystalline tl@xmath1rb@xmath2fe@xmath3se@xmath4 .
it is found that tl@xmath1rb@xmath2fe@xmath3se@xmath4 shows a very large upper critical field ( @xmath22 t , @xmath23@xmath24 t ) with a moderate anisotropic parameter @xmath25 ( @xmath26 ) , remarkably resembling those of iron pnictide superconductors.@xcite on the other hand , the superconducting transition of tl@xmath1rb@xmath2fe@xmath3se@xmath4 is substantially broadened in a magnetic field , indicating significant contributions of thermally activated flux flow in the vortex state .
single crystals of tl@xmath1rb@xmath2fe@xmath3se@xmath4 were synthesized by using a bridgeman method.@xcite the x - ray diffraction ( xrd ) identified the derived samples as a single phase with a tetragonal thcr@xmath4si@xmath4 crystal structure .
the actual composition of the crystals was determined by energy dispersive x - ray spectrometer ( edxs ) .
magnetic field dependence of the electrical resistivity , @xmath27 , was measured up to 57.5 t using a typical 4-probe method in a capacitor - bank - driven pulsed magnet .
the experimental data were recorded on a digitizer using a custom designed high - resolution , low - noise synchronous lock - in technique . in order to minimize the eddy - current heating caused by the pulsed magnetic field ,
very small crystals were cleaved off along the _
ab_-plane from the as - grown samples .
the electrical resistivity in a dc magnetic field ( 0 - 14 t ) was measured in an oxford instruments heliox vl system using a lakeshore ac resistance bridge and the angular dependence of the electrical resistivity was performed in a quantum design ( qd ) physical properties measurement system ( 9 t ppms ) .
angular linear ( @xmath28 ) transport measurements were carried out using the maximum lorentz force configuration ( * * j**@xmath12*h * ) , with * h * applied at an angle @xmath29 from the _ c _ axis of the crystal ( see the inset of fig .
[ figangle ] ) .
at @xmath30 for tl@xmath1rb@xmath2fe@xmath3se@xmath4(#1 ) .
the inset enlarges the section at the superconducting transition , where @xmath31k and @xmath32k are determined from the onset and the end - point of the superconducting transition , respectively.,scaledwidth=45.0% ] in fig .
[ fig1 ] , we show the temperature dependence of the electrical resistivity at zero field for tl@xmath1rb@xmath2fe@xmath3se@xmath4 ( # 1 ) .
one can see that the resistivity @xmath33 shows a hump around 154k , changing from semiconducting to metallic behavior upon cooling down from room temperature .
such a hump in @xmath33 has been widely observed in the iron selenides@xcite and its position can be tuned either by doping@xcite or pressure.@xcite the origin of the hump and its relation with superconductivity remain unclear .
a very sharp superconducting transition shows up at @xmath3433.5k , indicating a high quality of the sample .
note that we have totally measured four samples cut from the same batch in this context and their @xmath15 only varies slightly from 32.9k to 33.5k , indicating a good reproducibility of the superconducting properties in these samples . for tl@xmath1rb@xmath2fe@xmath3se@xmath4(#1 and #
2 ) measured in dc fields up to 14 t : ( a ) @xmath7@xmath12@xmath9 ; ( b ) @xmath7@xmath8@xmath9.,scaledwidth=45.0% ] at various temperatures for tl@xmath1rb@xmath2fe@xmath3se@xmath4(#3 ) : ( a ) @xmath7@xmath12@xmath9 ; ( b ) @xmath7@xmath8@xmath9.,scaledwidth=45.0% ] fig .
[ fig2dc ] and fig .
[ fig2 ] show the temperature- and the magnetic - field - dependence of the electrical resistivity for tl@xmath1rb@xmath2fe@xmath3se@xmath4 , respectively . in order to study the anisotropic behavior ,
the electrical resistivity was measured with field perpendicular ( a ) and parallel ( b ) to the _ c_-axis .
the magnetic field is applied up to 14 t for the dc fields ( fig .
[ fig2dc ] ) and up to 58 t for the pulsed magnetic fields ( fig .
[ fig2 ] ) . obviously , the superconducting transition eventually shifts to lower temperature upon applying a magnetic field . however , superconductivity is remarkably robust against the magnetic field in tl@xmath1rb@xmath2fe@xmath3se@xmath4 and it is not yet completely suppressed at our maximum field of 58 t .
furthermore , one can see that the superconducting transition is significantly broadened upon applying a magnetic field , showing a tail structure at low temperature .
for example , the width of the superconducting transition , defined from the onset temperature to the end point of the superconducting transition ( see the inset of fig.1 ) , is as small as 0.5k at zero field but increases to 2k and 3.6k for @xmath7@xmath12@xmath9 and @xmath7@xmath8@xmath9 at 14 t , respectively .
similar features were also observed in some 1111-type iron pnictides.@xcite we will argue later that such behavior might be attributed to the thermally activated flux flow in the vortex state . in the normal state , tl@xmath1rb@xmath2fe@xmath3se@xmath4 shows significant positive magnetoresistance for both @xmath7@xmath8@xmath9 and @xmath7@xmath12@xmath9 .
it is noted that , in iron pnictides , the magnetoresistance becomes very large while entering the magnetic state , but is negligible in the non - magnetic state.@xcite one possibility for the occurrence of such a large magnetoresistance in tl@xmath1rb@xmath2fe@xmath3se@xmath4 might be related to its magnetic ordering at high temperature.@xcite for tl@xmath1rb@xmath2fe@xmath3se@xmath4 .
( a ) the values of @xmath0 are determined from the superconducting onsets as described in fig . 1 and fig .
symbols of the open circle ( @xmath35 ) , filled circle ( @xmath36 ) and cross ( + ) represent the data obtained in a pulsed magnetic field and the triangles ( @xmath37 and @xmath38 ) denote those measured in a dc magnetic field .
note that samples # 3 and # 4 were measured in a pulsed field , but only sample # 3 was successfully measured for both @xmath7@xmath12@xmath9 and @xmath7@xmath8@xmath9 .
( b ) the values of @xmath0 are determined from the onset ( @xmath35 ) , the min - point ( @xmath39 ) and the end - point ( @xmath37 ) of the resistive superconducting transitions ( sample # 3 ) , respectively .
the inset shows the corresponding temperature dependence of the anisotropic parameter @xmath40.,scaledwidth=45.0% ] the upper critical field @xmath0 of tl@xmath1rb@xmath2fe@xmath3se@xmath4 is shown in fig .
[ fig3 ] , in which various symbols represent either different field orientations for the same sample ( # 3 ) or different samples as marked in the figure . in fig .
4(a ) , we determine the critical temperatures @xmath41 ( in the case of dc field ) or the critical fields @xmath6 ( in the case of pulsed field ) from the superconducting onsets as described in the inset of fig . 1 and also in fig.3 , i.e. , the intersection point of the resistive curves in the normal state and the superconducting transition .
such a determination of @xmath6 ( or @xmath41 ) is appropriate for the in - field measurements and is particularly useful in presence of the magnetoresistance @xmath42 . in this case
, one can extrapolate @xmath42 to lower temperatures to determine @xmath6 at the lowest temperatures since @xmath42 in the normal state hardly depend on temperature .
it is noted that similar field - induced broadening of the resistive superconducting transition was also observed in the high @xmath43 cuprates in which the onset temperature as we described here was shown to be close to that determined by other bulk measurements , e.g. , the magnetization.@xcite from fig .
4(a ) , one can see that the derived @xmath0 demonstrates quantitatively same behavior for all the investigated samples , independent of the detailed experimental methods .
the upper critical field @xmath44 linearly increases with decreasing temperature , reaching a value of @xmath10 t . on the other hand
, @xmath45 shows a convex curvature with a much larger value at low temperatures . such behavior of @xmath0 is not changed by the field - induced broadening of the superconducting transition . taking sample # 3 as an example , in fig .
4(b ) we plot the upper critical fields @xmath0 determined at the onset , mid- and end - point of the resistive transitions ( see fig .
3 ) , which follow remarkably similar temperature dependence .
we note that our results are consistent with those of k@xmath46fe@xmath47se@xmath4@xcite and k@xmath48fe@xmath49se@xmath50s@xmath51;@xcite the former was measured using a tunnel - diode resonator ( tdr ) technique in a pulsed magnetic field and the latter was measured only up to 9 t .
it was also shown that the values of @xmath6 determined from the end - point of the resistive transitions and the tdr technique are consistent , @xcite further confirming the validity of our methods in the determination of upper critical fields . in a superconductor ,
the cooper pairs can be destroyed by the following two mechanisms in a magnetic field : ( i ) the orbital pair breaking due to the lorentz force acting via the charge on the momenta of the paired electrons ( orbital limit ) ; ( ii ) the zeeman effect aligning the spins of the two electrons with the applied field ( pauli paramagnetic limit ) . according to the whh method , the orbital - limiting upper critical field @xmath52(0k ) for a single band bcs superconductor
is determined by the initial slope of @xmath0 at @xmath15 , i.e.@xcite @xmath53 which value may depend on the field orientations .
the pauli paramagnetic limiting field for weakly coupled bcs superconductors is given by@xcite @xmath54=1.86\it{t_c}\rm[k].\label{eq:2}\ ] ] while the upper critical field is usually restricted by the orbital limit in conventional superconductors , the spin paramagnetic effect may play an important role in pair breaking in some unconventional superconductors .
tl@xmath1rb@xmath2fe@xmath3se@xmath4 reveals a relatively large and anisotropic initial slope of @xmath0 near @xmath15 , which reaches a value of -12t / k and -2t / k ( from the superconducting onsets ) for @xmath7@xmath12@xmath9 and @xmath7@xmath8@xmath9 , respectively .
following eq .
[ eq:1 ] , one can derive the orbitally limited upper critical field , which gives @xmath55 t for @xmath7@xmath12@xmath9 and 45 t for @xmath7@xmath8@xmath9 .
as shown in fig . [ fig3 ] , @xmath56 considerably exceeds the experimental value of @xmath11 for @xmath7@xmath12@xmath9 , but lightly falls below the corresponding @xmath11 for @xmath7@xmath8@xmath9 . on the other hand , eq .
[ eq:2 ] gives a pauli paramagnetic limiting field of @xmath57 t in terms of the bcs theory .
thus , it is likely that the upper critical field is limited by orbital effect for @xmath7@xmath8@xmath9 , but by spin paramagnetic effect for @xmath7@xmath12@xmath9 . in order to further look into this point
, we fitted the experimental data of @xmath0 by the whh model,@xcite in which the effects of both orbital- and spin - pair breaking are considered ( see fig .
4a ) . in this model , @xmath58 and @xmath59
are the fitting parameters ; @xmath58 is the maki parameter , which represents the relative strength of spin and orbital pair - breaking , and @xmath59 is the spin - orbit scattering constant . as shown in fig.4(a ) , the upper critical field @xmath0 for both @xmath7@xmath8@xmath9 and @xmath7@xmath12@xmath9 are not well described by the whh method while ignoring the spin effect ( see the dot dashed lines with @xmath58=0 and @xmath59=0 ) .
the enhancement of @xmath44 at low temperature is likely attributed to its multi - band electronic structure as discussed in other fescs.@xcite indeed , @xmath45 can be well fitted by the whh model after considering the spin effect ( see the solid line in fig.4a ) , which gives @xmath58=5.6 and @xmath59=0.3 .
such a large value of @xmath58 indicates that the spin paramagnetism may play an important role on suppressing superconductivity for @xmath7@xmath12@xmath9 .
similar analysis applies to the data of @xmath0 derived at various resistive drops of the broad superconducting transition , showing generally consistent behavior . in the case of a cylinder - like fermi surface , the open electron orbits along the _ c_-axis make the orbital limiting upper critical field unlikely .
existence of a pauli limiting @xmath60 for @xmath61 seems to agree with the enhanced anisotropy in tl@xmath1rb@xmath2fe@xmath3se@xmath4 ( see below ) . on the other hand
, its multi - band electronic structure may complicate the analysis of @xmath0 .
nevertheless , the upper critical field of tl@xmath1rb@xmath2fe@xmath3se@xmath4 shows similar behavior to that of other iron pnictide superconductors,@xcite indicating a uniform scenario of @xmath60 in the fescs .
the anisotropic parameter @xmath40 of tl@xmath1rb@xmath2fe@xmath3se@xmath4 ( sample # 3 ) , derived at variant points of the superconducting transition , is shown in the inset of fig.4(b ) .
one can see that all these curves follow exactly same temperature dependence but with a small deviation on their absolute values of @xmath62 .
for example , the anisotropic parameter @xmath25 , determined from the superconducting onsets , is as high as 8 near @xmath15 but reaches 2 at @xmath63 k. similar values of @xmath40 were also obtained for other samples as a result of their consistent behavior of the upper critical field ( see fig .
the anisotropy of the upper critical field near @xmath15 can be associated with its electronic band structure .
observation of @xmath64 near @xmath15 might suggest a quasi-2d electronic structure for tl@xmath1rb@xmath2fe@xmath3se@xmath4 , which is compatible with the large resistive anisotropy in the normal state ( @xmath65)@xcite . an anisotropy of @xmath25@xmath58 near @xmath15 in tl@xmath1rb@xmath2fe@xmath3se@xmath4 is relatively large among the fescs , @xcite but it is close to that of the 1111-compounds .
@xcite nevertheless , in all these systems the superconducting properties tend to be more isotropic at low temperatures , which would be compatible with the isotropic - like superconducting energy gaps observed in the arpes experiments around 15k .
@xcite in order to further characterize the nature of the anisotropy in tl@xmath1rb@xmath2fe@xmath3se@xmath4 , we have measured the angular dependence of the electrical resistivity @xmath33 at various magnetic fields .
[ figangle ] plots the angular dependence of @xmath66 , where @xmath29 is the angle between the magnetic field and the _ c_-axis of the sample as marked in the figure . at magnetic fields of @xmath67=3 t , 6 t and 9 t .
the solid lines show the fittings to eq .
[ eq:4 ] . here
@xmath66 is determined from the superconducting onsets.,scaledwidth=45.0% ] according to the single band anisotropic ginzburg - landau ( g - l ) theory,@xcite the angular dependence of the upper critical field can be scaled by : @xmath68 where @xmath25=(@xmath69)@xmath70=@xmath71 .
here @xmath72 and @xmath73 are the effective masses of electrons for the in - plane and out - of - plane motion , respectively . in the case
that @xmath17 is a linear function of temperature , the angular dependence of @xmath74 can be converted to that of @xmath75 by:@xcite @xmath76 where @xmath77 is the zero field superconducting transition temperature and _ h _ is the applied magnetic field . in our case ,
the upper critical field near @xmath15 indeed shows nearly linear temperature dependence for both @xmath7@xmath12@xmath9 and @xmath7@xmath8@xmath9 .
therefore , one can estimate the anisotropic parameter @xmath25 from the angular dependence of @xmath66 .
it was shown that a single band anisotropic model can properly describe the angular dependence of @xmath6 in a multi - band system at temperatures near @xmath15.@xcite indeed , @xmath66 can be nicely fitted by eq .
[ eq:4 ] ( see the solid lines in fig.[figangle ] ) , indicating that , at least in the low field region , @xmath66 can be described by the g - l theory and the anisotropic upper critical field is attributed to the effective mass anisotropy in strongly coupled layered superconductors .
furthermore , the above fittings give @xmath25=8.5 , 7.3 and 6.3 for @xmath67=3 t , 6 t and 9 t , respectively .
these values of @xmath78 are very close to those shown in the inset of fig . 4 if we convert the magnetic fields to temperatures following the relation of @xmath0 . in comparison ,
the iron selenides show intrinsically similar properties of the upper critical field to the iron pnictide superconductors , @xcite but with a slightly enhanced anisotropy .
these findings suggest that all these fescs might share the same characters of superconductivity .
this is surprising because the electronic structure and the normal state of iron selenides seem to be very unique among the fescs .
for example , both hole pockets and electron pockets are observed in iron pnictides,@xcite but the hole pocket seems to be absent in iron selenides.@xcite the nesting between the hole pockets and the electron pockets was regarded as a prerequisite factor for the forming of @xmath79 paring state,@xcite a widely accepted proposal for the iron pnictide superconductors .
our findings of the universal behavior of @xmath0 in iron pnictides and selenides , therefore , urge to check whether the missing of hole pockets in iron selenides is intrinsic or masked by other experimental factors , e.g. , the phase separation or sample non - stiochiometry .
if it is intrinsic , one probably needs to reconsider the order parameters and the pairing mechanism of fescs in a unified picture .
rb@xmath2fe@xmath3se@xmath4 at various magnetic fields : ( a ) @xmath7@xmath12@xmath9 ; ( b ) @xmath7@xmath8@xmath9 .
the inset shows the thermally activated energy , @xmath80 , obtained from the slope of the arrehnius plot .
the solid lines are fitted to @xmath81 with [email protected] and [email protected] for @xmath7@xmath12@xmath9 and @xmath7@xmath8@xmath9 , respectively.,scaledwidth=45.0% ] as already mentioned above , the superconducting transition of tl@xmath1rb@xmath2fe@xmath3se@xmath4 is significantly broadened upon applying a magnetic field ( see fig .
[ fig2dc ] and fig .
[ fig2 ] ) .
similar features were previously observed in other layered superconductors , including the cuprates@xcite and the 1111-type iron pnictides,@xcite which were interpreted in terms of the energy dissipation caused by vortices motion . in general ,
both thermally activated flux flow ( taff ) and superconducting critical fluctuations may broaden the resistive superconducting transition in a magnetic field .
the importance of thermal fluctuations is measured by ginzburg number , @xmath83(@xmath25@xmath15/@xmath84@xmath85)@xmath86 , where @xmath87 is the thermal dynamic critical field and @xmath88 is the coherence length in the ab - plane .
in tl@xmath1rb@xmath2fe@xmath3se@xmath4 , the relatively large anisotropic parameter @xmath25 near @xmath15 and the short coherence length ( @xmath89 nm ) yield a large ginzburg number @xmath90 and a soft vortex matter . as a result
, thermal fluctuations may become important enough to overcome the elastic energy of the vortex lattice in a large part of the magnetic field -temperature phase diagram , melting the vortex lattice into a liquid . on the other hand
, the critical fluctuations may play a less dominant role here .
therefore , in the following we will try to interpret the resistive broadening in terms of the thermally activated flux flows which actually gives a rather good description to our experimental data .
further measurements are also under way in order to elucidate the possible effect of superconducting fluctuations . according to the taff model,@xcite
the resistivity @xmath91 can be expressed as : @xmath92e^{-j_{c0}hvl / t } , \label{eq:5}\ ] ] where @xmath93 is an attempt frequency for a flux bundle hopping , _
l _ is the hopping distance , _ j _ is the applied current density , _
@xmath94 _ is the critical current density in the absence of flux creep and _ v _ is the bundle volume . in the limit of @xmath95 ,
[ eq:5 ] can be simplified as @xmath96 where @xmath97 is the thermally activated energy and @xmath98=@xmath99 .
assuming that @xmath100 is a temperature independent constant , noted as @xmath101 , and @xmath102 , then eq .
[ eq:6 ] can be simplified to the arrhenius relation : @xmath103 thus , the apparent activation energy , @xmath80 , could be extracted from the slopes of the arrhenius plots , i.e. , the plot of ln@xmath28 _ v.s .
_ @xmath104 . in order to study such a possible vortex motion in tl@xmath1rb@xmath2fe@xmath3se@xmath4
, we plot the electrical resistivity @xmath28 as a function of @xmath104 in a semi - log scale at various magnetic fields ( see fig .
[ fig4 ] ) .
it is clear that the arrhenius relation holds over a wide temperature range for both @xmath7@xmath12@xmath9 and @xmath7@xmath8@xmath9 , suggesting that the taff model may nicely describe the field - induced resistive broadening in tl@xmath1rb@xmath2fe@xmath3se@xmath4 .
following @xmath80=-d@xmath105 , the apparent activation energy , @xmath106 , can then be determined from the slope of the linear parts in fig .
for example , this yields @xmath107@xmath1084900k for @xmath7@xmath12@xmath9 and 3607k for @xmath7@xmath8@xmath9 at 2 t .
the derived values of @xmath106 are plotted as a function of field in the insets of fig .
[ fig4 ] : ( a ) @xmath7@xmath12@xmath9 and ( b ) @xmath7@xmath8@xmath9 , which follow a power law of @xmath106@xmath5@xmath109 .
the fittings give _ _
[email protected] for @xmath7@xmath12@xmath9 and [email protected] for @xmath7@xmath8@xmath9 , indicating that the pinning force may have a weak orientation dependence .
the derived values of @xmath106 are slightly larger than those of some 1111-type iron pnictides , e.g. ndfeaso@xmath110f@xmath111@xcite and cefeaso@xmath112f@xmath113,@xcite but smaller than those of smfeaso@xmath114@xcite and ybco@xcite , indicating a moderate pinning force among the cuprates and the fescs .
note that in the 122-type iron pnictides where @xmath25@xmath142 near @xmath15 , the range of vortex liquid state is very narrow and no significant broadening of the resistive superconducting transition has been observed in a magnetic field .@xcite
the upper critical field @xmath0 , its anisotropy @xmath40 and the vortex motion of the newly discovered tl@xmath1rb@xmath2fe@xmath3se@xmath4 have been studied by measuring the field- , temperature- and angular- dependence of the electrical resistivity .
we found that this compound shows a large upper critical field ( @xmath22 t , @xmath115 t ) with a moderate anisotropy near @xmath15 .
the anisotropic parameter decreases with decreasing temperature , reaching @xmath25(20k)@xmath52 .
analysis based on the whh model indicates that the upper critical field is orbitally limited for @xmath7@xmath8@xmath9 , but is likely limited by the spin paramagnetic effect for @xmath7@xmath12@xmath9 .
these properties remarkably resemble those of iron pnictide superconductors , suggesting that all the fe - based superconductors may bear similar characters and , therefore , provide restrictions on the theoretical model .
similar to the cuprates and the 1111-series of iron pnictides , thermally activated flux flow might be responsible for the tail structure of the resistive transition below @xmath15 and the derived thermal activation energies are compatible with those of the 1111-type iron pnictides .
work at zju was supported by the national science foundation of china ( grant nos : 10874146 , 10934005 ) , the national basic research program of china ( 973 program ) ( 2009cb929104 , 2011cba00103 ) , the pcsirt of the ministry of education of china , zhejiang provincial natural science foundation of china , and the fundamental research funds for the central universities .
work at lanl was performed under the auspices of the national science foundation , the department of energy , and the state of florida .
bm was supported by the u.s .
department of energy , basic energy sciences , materials sciences and engineering division .
10 for reviews , see , e.g. , z. a. ren , and z. x. zhao , adv . mater . * 21 * , 4584 ( 2009 ) ; d. c. johnston , adv .
phys . * 59 * , 803 ( 2010 ) ; j. paglione , and r. l. greene , nat . phys .
* 6 * , 645 ( 2010 ) ; p. m. aswathy , j. b. anooja , p. m. sarun , and u. syamaprasad , supercond .
* 23 * , 073001 ( 2010 ) .
j. dong , h. j. zhang , g. xu , z. li , g. li , w. z. hu , d. wu , g. f. chen , x. dai , j. l. luo , z. fang , and n. l. wang , europhys .
lett . * 83 * , 27006 ( 2008 ) ; d. j. singh , and m. h. du , phys .
lett . * 100 * , 237003 ( 2008 ) .
q. si , and e. abrahams , phys .
lett . * 101 * , 076401 ( 2008 ) ; k. seo , b. a. bernevig , and j. p.
hu , phys .
* 101 * , 506404 ( 2008 ) ; j. h. dai , q. m. si , j. x. zhu , and e. abrahams , proc .
sci . * 106 * , 4118 ( 2009 ) .
s. p. kou , t. li , and z. y. weng , europhys . lett . * 88 * , 17010 ( 2009 ) ; m. d. johannes , and i. i. mazin , phys .
b * 79 * , 220510(r ) ( 2009 ) ; l. demedici , s. r. hassan , m. capone , and x. dai , phys .
102 * , 126401 ( 2009 ) . h. q. yuan , l. jiao , f. f. balakirev , j. singleton , c. setty , j. p. hu , t. shang , l. j. li , g. h. cao , z. a. xu , b. shen , and h. h. wen , arxiv : cond - mat/1102.5476 ( 2011 ) . s. j. moon , j. h. shin , d. parker , w. s. choi , i. i. mazin , y. s. lee , j. y. kim , n. h. sung , b. k. cho , s. h. khim , j. s. kim , k. h. kim , and t. w. noh , phys . rev .
b * 81 * , 205114 ( 2010 ) .
j. g. guo , s. f. jin , g. wang , s. c. wang , k. x. zhu , t. t. zhou , m. he , and x. l. chen , phys
b * 82 * , 180520 ( 2010 ) .
a. krzton - maziopa , z. shermadini , e. pomjakushina , v. pomjakushin , m. bendele , a. amato , r. khasanov , h. luetkens , and k. conder , j. phys . : condens .
matter * 23 * , 052203 ( 2011 ) .
a. f. wang , j. j. ying , y. j. yan , r. h. liu , x. g. luo , z. y. li , x. f. wang , m. zhang , g. j. ye , p. cheng , z. j. xiang , and x. h. chen phys .
b * 83 * , 060512 ( 2011 ) .
m. h. fang , h. d. wang , c . h. dong , z. j. li , c. m. feng , j. chen , and h. q. yuan , europhys . lett . *
94 * , 27009 ( 2011 ) .
h. d. wang , c. h. dong , z. j. li , q. h. mao , s. s. zhu , c. m. feng , h. q. yuan , and m. h. fang , europhys .
lett . * 93 * , 47004 ( 2011 ) .
r. h. liu , x. g. luo , m. zhang , a. f. wang , j. j. ying , x. f. wang , y. j. yan , z. j. xiang , p. cheng , g. j. ye , z. y. li , and x. h. chen , europhys . lett . * 94 * , 27008 ( 2011 ) .
z. shermadini , a. krzton - maziopa , m. bendele1 , r. khasanov , h. luetkens , k. conder , e. pomjakushina , s. weyeneth , v. pomjakushin , o. bossen , and a. amato , phys .
lett . * 106 * , 117602 ( 2011 ) .
w. bao , g. n. li , q. huang , g. f. chen , j. b. he , m. a. green , y. qiu , d. m. wang , and j. l. luo , arxiv : cond - mat/1102.3674 ( 2011 ) ; w. bao , q. huang , g. f. chen , m. a. green , d. m. wang , j. b. he , x. q. wang , and y. qiu , chinese phys . lett . * 28 * , 086104 ( 2011 ) .
y. j , yan , m. zhang , a. f. wang , j. j. ying , z. y. li , w. qin , x. g. luo , j. q. li , j. p. hu , and x. h. chen , arxiv : cond - mat/1104.4941 ( 2011 ) .
y. zhang , l. x. yang , m. xu , z. r. ye , f. chen , c. he , h. c. xu , j. jiang , b. p. xie , j. j. ying , x. f. wang , x. h. chen , j. p. hu ,
m. matsunami , s. kimura , and d. l. feng , nat . mat . * 10 * , 273 ( 2011 ) .
x. p. wang , t. qian , p. richard , p. zhang , j. dong , h. d. wang , c .- h .
dong , m. h. fang , and h. ding , europhys .
lett . * 93 * , 57001 ( 2011 ) .
d. x. mou , s. y. liu , x. w. jia , j. f. he , y. y. peng , l. zhao , l. yu , g. d. liu , s. l. he , x. l. dong , j. zhang , h. d. wang , c. h. dong , m. h. fang , x. y. wang , q. j. peng , z. m. wang , s. j. zhang , f. yang , z. y. xu , c. t. chen , and x. j. zhou , phys .
* 106 * , 107001 ( 2011 ) .
h. ding , p. richard , k. nakayama , k. sugawara , t. arakane , y. sekiba , a. takayama , s. souma , t. sato , t. takahashi , z. wang , x. dai , z. fang , g. f. chen , j. l. luo , and n. l. wang , europhys . lett .
* 83 * , 47001 ( 2008 ) ; v. brouet , m. marsi , b. mansart , a. nicolaou , a. taleb - ibrahimi , p. le fvre , f. bertran , f. rullier - albenque , a. forget , and d. colson , phys .
b * 80 * , 165115 ( 2009 ) . h. q. yuan , j. singleton , f. f. balakirev , s. a. baily , g. f. chen , j. l. luo , and n. l. wang , nature * 457 * , 565 ( 2009 ) .
s. a. baily , y. kohama , h. hiramatsu , b. maiorov , f. f. balakirev , m. hirano , and h. hosono , phys .
lett . * 102 * , 117004 ( 2009 ) .
m. h. fang , j. h. yang , f. f. balakirev , y. kohama , j. singleton , b. qian , z. q. mao , h. d. wang , and h. q. yuan , phys . rev .
b**81 * * , 020509 ( 2010 ) .
n. kurita , k. kitagawa , k. matsubayashi , a. kismarahardja , e. s. choi , j. s. brooks , y. uwatoko , s. uji , and t. terashima , j. phys .
. jpn . * 80 * , 013706 ( 2011 ) ; k. cho , h. kim , m. a. tanatar , y. j. song , y. s. kwon , w. a. coniglio , c. c. agosta , a. gurevich , and r. prozorov , phys .
b * 83 * , 060502 ( 2011 ) ; j. l. zhang , l. jiao , f. f. balakirev , x. c. wang , c. q. jin , and h. q. yuan , phys . rev .
b * 83 * , 004500 ( 2011 ) . h. s. lee , m. bartkowiak , j. h. park , j. y. lee , j. y. kim , n. h. sung , b. k. cho , c. u. jung , j. s. kim , and h. j. lee , phys .
b * 80 * , 144512 ( 2009 ) .
v. a. gasparov , l. drigo , a. audouard , d. l. sun , c. t. lin , s. l. bud@xmath116ko , p. c. canfield , f. wolff - fabris , and j. wosnitza , arxiv : cond - mat/1104.5619 ( 2011 ) .
j. guo , l. l. sun , c. zhang , j. g. guo , x. l. chen , q. wu , d. c. gu , p. w. gao , x. dai and z. x. zhao , arxiv : cond - mat/1101.0092(2011 ) . j. jaroszynski , f. hunte , l. balicas , y. jo , i. raicevic , a. gurevich , d. c. larbalestier , f. f. balakirev , l. fang , p. cheng , y. jia , and h. h. wen , phys . rev .
b * 78 * , 174523 ( 2008 ) .
m. shahbazi , x. l. wang , c. shekhar , o. n. srivastava , and s. x. dou , supercond .
* 23 * , 10500 ( 2010 ) .
y. z. zhang , z. a. ren , and z. x. zhao , supercond .
* 22 * , 065012 ( 2009 ) ; h. lee , m. bartkowiak , j. s. kim , and h. lee , phys .
b * 82 * , 104523 ( 2010 ) .
u. welp , w. k. kwok , g. w. crabtree , k. g. vandervoort , and j. z. liu , phys .
62 * , 1908 ( 1989 ) ; m. s. ospfsky , r. j. spulen , jr .
, s. a. wolf , j. m. brptp , h. rakptp , j. c. ousset , g. coffe , s .
askenazy , p. pari , i. bozovic , j. n. eckstein , and g. f. virshup , phys .
lett . * 71 * , 2315 ( 1993 ) .
e. d. mun , m. m. altarawneh , c. h. mielke , v. s. zapf , r. hu , s. l. budko and p. c. canfield , phys .
b * 83 * , 100514 ( 2011 ) . h. c. lei , and c. petrovic , europhys . lett . * 95 * , 57006 ( 2011 ) .
n. r. werthamer , e. helfand , and p. c. hohenberg , phys . rev . * 147 * , 295 ( 1966 ) .
a. m. clogston , phys .
lett . * 9 * , 266 ( 1962 ) ; b. s. chandrasekhar , j. appl .
. lett . * 1 * , 7 ( 1962 ) .
d. r. tilley , proc .
london * 86 * , 289 ( 1965 ) .
g. blatter , v. b. geshkenbein , and a. i. larkin , phys . rev
* 68 * , 875 ( 1992 ) .
u. welp , w. k. kwok , g. w. crabtree , k. g. vandervoort , and j. z. liu , phys .
b * 40 * , 5263 ( 1989 ) .
a. gurevich , phys .
b * 67 * , 184515 ( 2003 )
. i. i. mazin , and j. schmalian , phys .
c * 469 * , 614 ( 2009 ) .
w. k. kwok , s. fleshler , u. welp , v. m. vinokur , j. downey , g. w. crabtree , and m. m. miller , phys .
. lett . * 69 * , 3370 ( 1992 ) .
t. t. m. palstra , b. batlogg , r. b. van dover , l. f. schneemeyer , and j. v. waszczak , phys .
b * 41 * , 6621 ( 1990 ) .
b. maiorov , t. katase , s.a .
baily , h. hiramatsu , t.g .
holesinger , h. hosono and l. civale supercond .
* 24 * , 055007 ( 2011 ) .
g. blatter , m. v. feigelman , v. b. geshkenbein , a. i. larkin , and v. m. vinokur , rev .
phys . * 66 * , 1125 ( 1994 ) . | the upper critical field @xmath0 of tl@xmath1rb@xmath2fe@xmath3se@xmath4 single crystals has been determined by means of measuring the electrical resistivity in both a pulsed magnetic field ( @xmath558 t ) and a dc magnetic field ( @xmath514 t ) .
it is found that @xmath6 linearly increases with decreasing temperature for @xmath7@xmath8@xmath9 , reaching @xmath10 t. on the other hand , a larger @xmath11 with a strong convex curvature is observed for @xmath7@xmath12@xmath9 ( @xmath13(18k)@xmath1460 t ) .
this compound shows a moderate anisotropy of the upper critical field around @xmath15 , which decreases with decreasing temperature .
analysis of the upper critical field based on the werthamer - helfand - hohenberg ( whh ) method indicates that @xmath11 is orbitally limited for @xmath7@xmath8@xmath9 , but the effect of spin paramagnetism may play an important role on the pair breaking for @xmath7@xmath12@xmath9 .
all these experimental observations remarkably resemble those of the iron pnictide superconductors , suggesting a universal scenario for the iron - based superconductors .
moreover , the superconducting transition is significantly broadened upon applying a magnetic field , indicating strong thermal fluctuation effects in the superconducting state of tl@xmath1rb@xmath2fe@xmath3se@xmath4 .
the derived thermal activation energy for vortex motion is compatible with those of the 1111-type iron pnictides . |
dynamical spin excitations of impurities is a subject that douglas l. mills tackled 50 years ago , while in france as a post - doctoral fellow in the group of jacques friedel .
it was one of his first experiences in the realm of magnetism , after completing a phd thesis supervised by charles kittel .
since then , his several contributions in this field , going from magnetic bulk materials down to magnetic surfaces or adatoms , were tremendous and important .
his work is notorious for making significant predictions often years or decades before experimental techniques advanced to the point where his findings can be verified .
the seminal paper on spin - excitations @xcite , written with pascal lederer , is a beautiful example . within a framework conceptually similar to that used in the present contribution
, he predicted that , if an experimental tool would be able to probe the local magnetic response of isolated impurities embedded in a paramagnetic metallic host , and if a static external magnetic field is applied , the measured spectrum would display a shift in the position of the excitation signature from what is expected with a heisenberg model together with a very substantial broadening ( decrease of the lifetime ) of the excitational signature induced by the decay of the coherent spin precession to particle - hole pairs .
that tool turned out to be inelastic scanning tunneling spectroscopy ( ists ) , applied to the measurement of spin - excitations of magnetic adatoms adsorbed on nonmagnetic substrates ( see _ e.g. _ refs .
@xcite ) , and with which many of the predictions of douglas l. mills were verified .
understanding the excitation and the dynamical behavior of such nano - magnetic systems is naturally a topic of prime importance , actively studied recently both to provide insight into fundamental aspects of magnetism and as possible elements for future information technology .
access to magnetic excitation spectra is of high value in nanospintronics , due to the role of spin excitations as a dynamical route to control magnetic elements and ( spin ) currents , the building blocks of spintronics .
the principle of ists is depicted schematically in fig .
[ concept_ists ] .
electrons can tunnel , for example , from the tip to the available states of the substrate , leading to a tunneling current @xmath0 ( fig . [ concept_ists]a ) . if the bias voltage , @xmath1 , between the tip and substrate is large enough , and if a spin - excitational mode is possible in the adsorbate , the electron can exchange energy and possibly spin angular momentum with the latter , thereby triggering the excitation ( fig . [
concept_ists]b ) .
this creates an additional tunneling channel , translating to an increase in @xmath0 , also seen as a change in the slope of the @xmath0 vs @xmath1 curve ( fig .
[ concept_ists]c ) .
the first derivative of the current with respect to the bias voltage , @xmath2 , leads to a step - like function located at the characteristic frequency of the excitational mode ( fig .
[ concept_ists]d ) , while the second derivative @xmath3 displays a resonance ( fig .
[ concept_ists]e ) with a given linewidth . ,
to trigger the excitation mode , an additional tunneling channel is created ( b ) .
the slope of the tunneling current @xmath0 versus @xmath1 curve ( c ) changes at a bias voltage corresponding to the frequency of the excitation mode .
taking the first and second derivatives of this curve leads to a step - like function ( d ) or to a resonance ( e ) at @xmath4 . ]
several parameters determine the properties of these excitations and the related spin - dynamics , _ e.g. _
( i ) the type of substrate , ( ii ) the details of the hybridization between the electronic states of the adsorbate and those of the substrate , ( iii ) and the symmetry , shape and size of the adsorbate .
one theoretical approach to the interpretation of ists experiments is based on the heisenberg hamiltonian , describing an atomic - like localized moment with integer or half - integer spin .
such a model is useful mainly for systems where the substrate interacts weakly with the adsorbate @xcite ; it fails qualitatively to describe cases with strong coupling to the substrate electrons , where hybridization leads to moments far from integer and half integer values , and adsorbate electronic levels with widths in the range of @xmath5 ev .
interestingly , a heisenberg model predicts the existence of spin - excitations in all cases , however inelastic features are commonly not observed experimentally .
moreover a heisenberg model can not describe the main characteristics of spin - excitations , such as their lifetimes . a theoretical description of spin - excitations able to describe hybridization effects
is thus required .
employing a method based on first - principles electronic structure calculations is therefore a good choice but highly non - trivial , since access to dynamical response functions is required not only for bulk materials but also for surfaces and for nanostructures deposited on surfaces , which also present a high computational burden .
the transverse dynamical magnetic susceptibility , @xmath6 , describes , in linear response , the amplitude of the transverse spin motion produced by an external magnetic field @xmath7 of frequency @xmath8 .
it is a quantity providing a theoretical inelastic spectrum comparable to the experimental spectra .
indeed , the imaginary part of @xmath6 gives access to the local density of spin - excitations . to calculate the dynamical magnetic susceptibility
, one could employ either time - dependent density functional theory ( td - dft ) @xcite or many - body perturbation theory ( mbpt ) based on dft @xcite .
since these types of calculations for magnetic materials are computationally expensive , very few dft - based calculations , even for bulk systems , were performed @xcite , and it is only recently that a few groups have undertaken the task of calculating dynamical transverse magnetic response functions for bulk materials @xcite , thin films @xcite and adatoms or nanostructures on surfaces @xcite .
several studies based on empirical tight binding ( etb ) theory and mbpt were published @xcite which , although relying on the accuracy of fitting parameters , have advanced the understanding of many effects accompanying spin - excitations .
our first - principles scheme uses td - dft ( see refs .
@xcite ) and is designed for the investigation of spin - excitations in a real - space fashion , while applicable for systems with extended dimensions .
the focus on a real - space approach stems from our primary goal of investigating magnetic excitations in adatoms and small nanostructures measured with ists . in all schemes , a similar master equation , a dyson - like equation , must be tackled .
its solution may be be written in schematic notation , = ^0(1 - u^0)^-1 .
[ dyson ] in td - dft @xcite , @xmath9 is the kohn - sham ( ks ) susceptibility and @xmath6 , the total susceptibility , is in principle exact , if the full exchange and correlation kernel @xmath10 is known . in practice
, one often invokes the adiabatic local spin density approximation ( alda ) ; @xmath10 can also be viewed as a parameter whose value is in the range of @xmath11 for 3@xmath12 transition elements @xcite . in the etb or mbpt
, @xmath10 is the effective coulomb interaction corresponding to the local exchange splitting divided by the magnetic moment .
it turns out that in dft based methods , the goldstone theorem is violated in numerical studies : for instance , zero wavevector spin waves have finite frequency even with spin - orbit coupling set aside .
one then adjusts @xmath10 in an ad hoc manner making @xmath6 compatible with the goldstone theorem . in refs .
@xcite we introduced a scheme constraining @xmath10 obviating the need for ad - hoc adjustments .
the goal of this article is to review our recent work on the dynamical spin - excitations as described within linear response theory based on td - dft . after a description of the theoretical formalism ,
several applications are presented and discussed .
our focus is on the spin - dynamics of magnetic adatoms and dimers on different non - magnetic metallic substrates .
in refs . @xcite we presented a computationally attractive method that allows us to address magnetic excitations from first - principles .
we use the korringa - kohn - rostoker ( kkr ) single particle green function ( gf ) @xcite which contains an ab - initio description of the electronic structure .
such a method is required to describe spin - excitations in metallic systems where hybridization of electronic states plays a crucial role and a heisenberg model of the type used in ref .
@xcite would fail qualitatively . as mentioned above ,
our scheme is readily applied to bulk materials , to surfaces with adsorbed films , and it is a real space formalism ideal for diverse small nanostructures . to begin , we assume we have in hand a magnetic system for which a self - consistent groundstate calculation has already been performed .
we choose the coordinates so that the magnetization , @xmath13 , points along the @xmath14-direction ( collinear magnetic state and no spin - orbit coupling ) . perturbing the system with a small time - dependent external transverse magnetic field @xmath15 , an induced transverse
magnetization arises , @xmath16 , in the @xmath17-plane perpendicular to the @xmath14-direction . to describe the induced magnetization
, we require the frequency - dependent ks transverse susceptibility , or @xmath9 , which may be expressed in the form @xmath18
\label{chi0}\end{aligned}\ ] ] where @xmath19 is the fermi distribution function , @xmath20 and @xmath21 represent the retarded and advanced one - particle gfs connecting atomic sites @xmath22 and @xmath23 and @xmath24 .
a comment on the notation is in order .
the point @xmath25 is in the cell surrounding atom @xmath22 , and @xmath26 is in the cell around @xmath23 .
these vectors are measured from the center of their respective cells .
thus , if we wish to describe these two points with respect to a master origin @xmath27 , we write @xmath28 and @xmath29 , respectively , where @xmath30 is a vector from @xmath27 to the center point of cell @xmath22 ( likewise for @xmath23 ) . with this convention in mind ,
the single particle gf , often presented as @xmath31 , is abbreviated as @xmath32 , a notation that is very convenient when the kkr scheme we employ is utilized . by the kkr - gf method
@xcite , @xmath33 , where @xmath34 is the structural gf . here
the regular @xmath35 and irregular @xmath36 solutions of the schrdinger equation for cell @xmath22 are energy - dependent , thus making the calculation of @xmath9 in eq .
[ chi0 ] tedious and lengthy .
our ansatz expresses the gfs in terms of energy - independent wavefunctions @xmath37 such that @xmath38 , with @xmath39 generated from @xmath40 .
since the terms in the denominator are normalization factors instead of working with @xmath41 we introduce @xmath42 , choosing @xmath43 , _
i.e. _ , the @xmath12-regular solutions of the schrdinger equation at the fermi energy .
this is appropriate for the calculation of the @xmath12-block of the susceptibility of the several 3@xmath12 adatoms we studied .
the connection between the full susceptibility @xmath6 and the ks susceptibility @xmath9 is provided by the exchange - correlation kernel , @xmath44 , via eq .
[ dyson ] .
for the transverse response the kernel simplifies within the alda @xcite to @xmath45 ; @xmath46 is the magnetic part of the effective ks potential ( @xmath47 ) .
therefore @xmath10 is just the exchange splitting divided by the magnetization . combining our simplified form of the ks gf with the atomic sphere approximation ( asa ) , _ i.e. _ , spherical response to spherical perturbation
, the ks response simplifies to @xmath48 , a single number for each pair of atoms . the same procedure is repeated for the magnetic response function @xmath49 and the kernel @xmath10 by transforming eq .
[ dyson ] in the same way , so that we obtain a form that resembles the approach of lowde and windsor @xcite , which is very often used in the tight - binding simulations of magnetic excitations @xcite : |_ij ( ) = |_ij^0 ( ) + _ kl it is worth noticing that there is another way to determine the correct value of @xmath10 .
indeed , the following sum rule is a consequence of goldstone s theorem ( @xmath50 and @xmath51 ) @xcite : _ j _ij^0(,;0)b_j^(;0 ) = m_i(;0 ) [ sumrule ] this expression has been obtained after multiplying both sides of eq .
[ chi0 ] by @xmath52 , integrating over @xmath26 , summing up over all sites @xmath23 and using the following dyson equation : @xmath53 .
. [ sumrule ] can be rewritten using the alda : _
j _ij^0(,;0)u_j(;0)m_j(;0 ) = m_i(;0 ) and with our simplifications , calculating @xmath10 consistent with the goldstone theorem .
the right - hand - side requires knowledge of the groundstate magnetization and the static ks susceptibility .
stated otherwise , the correct @xmath10 is the one with the lowest eigenvalue of the denominator of eq .
[ dyson ] associated with the magnetic moments as components of the eigenvectors .
this has the advantage of applying also to more complex systems , involving non - equivalent atoms .
@xmath54 can be understood as a stoner parameter and gives once more a justification for the approach used by lowde and windsor @xcite : the effective intra - atomic coulomb interaction is expressed by only one parameter .
as an initial application , we explore the spin dynamics of single adatoms and dimers on the cu(001 ) surface .
the experimental lattice parameter of cu was considered and geometrical relaxations of the adatoms and dimers did not change the conclusions presented in this section . fig .
[ compilation](a ) shows our calculations of the resonant response of the local moments for four adatoms on this surface @xcite . if an external static magnetic field is applied along the initial direction of the magnetic moments , we expect from a heisenberg model a delta function in the excitation spectrum located at the larmor frequency , @xmath55 . for the field we have applied , a @xmath56-value of 2 would provide a resonance at 13.6 mev . as can be seen in fig .
[ compilation](a ) , instead of delta functions we obtain resonances of different widths that are shifted differently from the ideal larmor frequency depending on the chemical nature of the adatom .
cr and mn are characterized by sharper resonances ( smaller linewidths ) compared to fe and co , meaning that the former adatoms would be reasonably described by a heisenberg model where the magnetic exchange interactions are evaluated within an adiabatic approach .
the observed damping is induced by stoner ( electron - hole ) excitations described by the ks susceptibility @xmath9 and is related to the local density of states ( ldos ) @xcite , thus being influenced by the position of the @xmath12-states of each adatom relative to the fermi energy .
consequently , the co and fe resonances are quite broad , since their minority spin levels intersect the fermi level , whereas those for mn and cr are much sharper since for these adatoms the fermi level lies between the majority and minority states .
is plotted for every adatom when applying an additional magnetic field along the @xmath14-direction corresponding to a larmor frequency of 13.6 mev ( dashed line ) . in ( b ) @xmath57
is shown as calculated for the four dimers .
the optical modes , estimated for mn and fe from a heisenberg model , are indicated with dashed lines . ]
the resonant ( larmor ) frequency scales linearly with the applied magnetic field , as does the damping of the resonances for small frequencies , which is the regime of interest for ists .
this is observed in fig .
[ mn_susc_bfield ] for the case of the mn adatom . for the fields chosen
are represented by the black circles .
thus the @xmath56 shift is negative for this example . ]
we have also explored next nearest neighbor dimers of the same 3@xmath12 adatoms deposited on cu(001 ) surface @xcite , starting from a ferromagnetic configuration ( fig .
[ compilation](b ) ) . for the dimers
there are two resonances , a low frequency acoustical mode and a high frequency optical mode that is damped by decay into stoner excitations .
the position of the optical mode provides information on the stability of the assumed groundstate . for cr and co
, we find the optical mode at negative frequency , which informs us that the ferromagnetic state is unstable , whereas for fe and mn dimers the modes reside at positive frequencies , so ferromagnetism is stable in accordance with our static dft groundstate calculations . through adiabatic rotation of the moments @xcite , we extract an effective exchange magnetic interaction , @xmath58 , by fitting the energy change to the heisenberg form , @xmath59 where @xmath60 and @xmath61 are unit vectors .
the equation of motion for this model gives optical mode frequencies at @xmath62 where @xmath63 and @xmath64 are the magnetic moments of the two adatoms . within the adiabatic approximation
, we obtain the following resonant optical frequencies for cr , mn , fe and co dimers , respectively : 19.7 mev , 15.4 mev , 39.2 mev and 33.1 mev ( shown as dashed lines in fig . [ compilation](b ) ) . the same arguments used to describe the response of the single adatoms apply for the optical modes of dimers : compared to the optical resonance of the mn dimer , the one of fe is substantially broader and more shifted from the mode expected in the adiabatic approximation .
this demonstrates that due to stoner excitations , the effective magnetic exchange interaction obtained from the dynamical scheme can be very different from the one obtained in the adiabatic approach . in the case of dimers made of different magnetic adatoms , _
i.e. _ cofe dimer and mnfe dimer , we find the optical modes to occur at distinctly different frequencies , evident in fig .
[ dimer_different ] , when the spectral density is projected on either adatom .
this behavior is at variance with the heisenberg description of the excitation spectrum of two well - defined localized spins , where the optical mode is predicted to be unique for both adatoms .
the difference comes from the fact that the local density of stoner modes depends strongly on the kind of adatom , rendering the atom projected spectral densitites considerably different . indeed , mapping to the heisenberg model
, the optical modes are expected at 31.6 mev for mnfe and 21.7 mev for cofe as depicted by the vertical ines in fig .
[ dimer_different ] , irrespective of where the excitation is probed . for dimers with mixed adatoms are shown ( a ) for the cofe dimer and ( b ) for the mnfe dimer .
the vertical lines mark the energies of the optical model predicted by the heisenberg model ( see text ) .
it is interesting to note the presence of resonances at positive frequencies expressing a ferromagnetic ground state for both dimers . within each dimer ,
the atom - projected peaks are not located at the same position since the @xmath56-shift depends on the nature of the adatom . ]
the shift in the peak positions evident in fig .
[ dimer_different ] is a consequence of the itinerant nature of the magnetic moments . as each moment
precesses , the motion is heavily damped by its coupling to stoner excitations of the paramagnetic host . in the case of the mnfe dimer ,
the motions of the fe spin are damped far more heavily that those of the mn spin , as may be appreciated by contrast with fig .
[ compilation](a ) .
this has the consequence that the peak in @xmath65 is dragged down to a frequency somewhat lower than that in @xmath66 . to see this we constructed a toy model that consists of two coupled heisenberg spins , each connected to a reservoir that produces damping @xmath67 of the form encountered in the landau - lifshitz - gilbert equation@xcite .
the linearized equations of motion for this system reproduce the offset in the peaks evident in fig .
[ dimer_different](b ) .
we illustrate this in fig .
[ dimer_model ] , where @xmath68 and @xmath69 mimic the imaginary parts of @xmath70 and @xmath71 , respectively . by increasing the strength of the damping parameter @xmath72 compared to @xmath73
, we observe a shift to lower energies of the optical mode in @xmath69 ( _ i.e. _ @xmath65 ) .
the shape of the optical mode of mn - spin is completely different just by modifying a neighbor . indeed ,
close inspection of the optical mode observed in @xmath65 reveals that it is also more heavily damped in the mixed dimer mnfe ( fig .
[ dimer_different](b ) ) than in the pure mnmn dimer ( fig .
[ compilation](b ) ) . and
@xmath74 for two spins of unit length coupled by an exchange interaction of strength @xmath75 . here
, we mimic fe and mn by considering each spin coupled to a reservoir that provides a damping parameter @xmath76 ( 1 for mn and 2 for fe ) whose values are given in the inset . ]
the physical reason behind this intriguing behavior in the femn configuration is that the mn - spin during its precession feels the magnetic force of the more heavily damped fe - spin , which provides more damping for the mn atom .
it would be of great interest to employ stm - based spectroscopy to explore the response of the two spins in dissimilar dimers such as those just discussed .
recently we proceeded to a theoretical and experimental investigation of the spin - excitations spectra of fe adatoms deposited on the cu(111 ) surface @xcite . in fig .
[ comparison_cu111 ] a comparison is shown of the numerical derivative of the experimental @xmath2 to the imaginary part of @xmath6 for different magnetic fields .
at zero magnetic field ( see fig . [ comparison_cu111](a ) ) , two resonances are found in the experimental spectra at @xmath77 mev .
this excitation energy is related to the gap created by spin - orbit coupling and is closely related to the magnetic anisotropy energy of the fe adatoms . in ref .
@xcite , the imaginary part of the transverse dynamical susceptibility is presented as computed from a tight - binding scheme taking full - account of the spin - orbit interaction with parameters fitted to _ ab - initio _ calculations . in fig .
[ comparison_cu111 ] , however , the theoretical spectra with td - dft without spin - orbit coupling are shown .
an initial magnetic field is applied in order to fit the position of the experimentally obtained resonance .
the magnetic field dependence of the peak energy in the excitation spectrum agrees nicely with the measured data .
td - dft gives @xmath56-value around 2.1 in excellent agreement with the experiment .
the lifetime of the excitation , @xmath78 , defined by @xmath79 where @xmath80 is the full - width at half maximum ( fwhm ) of the resonance is found experimentally to be remarkably small ( 200 fs ) at zero magnetic field , while the theoretical values are approximately twice as large .
experimental spectra to the imaginary part of the transverse dynamical magnetic susceptibility ( from td - dft ) , for different values of the external magnetic field .
an offset field is added to the calculations , such that the position of the experimental peak with @xmath81 t is reproduced.@xcite ]
surprisingly , the spin - excitation spectra of fe adatoms on ag(111 ) surface are markedly different from those obtained when the substrate is cu(111 ) @xcite .
ists experiments and our _ ab - initio _ calculations demonstrate that the effective @xmath56-values is larger than 3 ( see fig . [ comparison_ag111 ] ) .
we find a @xmath56-factor of 3.3 , that is in very good agreement with the measured data ( @xmath82 ) .
thus our calculations confirm a large @xmath56-shift in the resonances of the excitation density of states .
the theoretical linewidth is about twice the corresponding experimental value and increases more in an applied field than the average measured linewidth .
our calculations identify the fe adatom @xmath56-factor as a special case , since cr , mn and co adatoms are characterized with ` regular ' @xmath56-values of 2.1 , 1.9 and 1.8 .
spectrum while in ( b ) the computed density of magnetization excitations ( im@xmath6 ) is depicted . ] at first sight , one would think that , since ag and cu are noble metals with similar characteristics , the spin - excitations of the adatoms placed on those substrates should be similar too . in order to understand this intriguing behavior
, we analyzed the differences between both substrates .
the ag(111 ) surface is peculiar since it has a surface state with an onset located much closer to the fermi energy ( @xmath83 ) compared to the one found on cu(111 ) @xcite .
our calculations indicate a value of 50 mev ( see fig .
[ fig_theory3](a ) and ( b ) ) .
moreover , it is well established that an adatom could induce a bound state that is split - off from the bottom of the surface state band if the potential of the adatom acts as attractive @xcite .
both effects could be large and affect @xmath56 .
as on the cu(111 ) surface , cr , mn , fe and co adatoms show the split - off state that is , as a side remark , spin - polarized ( fig .
[ fig_theory3](c ) ) .
if this bound state is responsible for the large @xmath56-shift observed for the fe adatom , it should also induce a large @xmath56-shift for cr , mn and co adatoms .
this is not the case and therefore we can rule out a strong effect arising from the bound state .
furthermore , we considered the impurities as inatoms ( embedded in the first substrate layer of the surface ) and impurity atoms in bulk , where no bound - state is expected as predicted for cu(111 ) @xcite . among all additional impurities investigated , only the mn inatom is characterized by a large @xmath56-factor of 2.9 .
this shows that even without a bound - state a large @xmath56-value can be obtained .
-point , where a is the lattice parameter of ag .
( b ) the s - ldos in the vacuum at 4.7 above the surface experiences a step - like behavior at the threshold of the surface state .
( c ) a spin - splitted bound state is created in case of an fe adatom around @xmath84 ev below @xmath83 .
( d ) the real part of @xmath6 for different adatoms is linear with respect to energy .
fe adatom is characterized by the smallest slope thereby producing its large @xmath56-shift . ] finally , to investigate the influence of the surface state , we shifted its location by reducing the thickness of the slab used to simulate ag(111 ) surface from 24 to 5 monolayers .
with such a theoretical trick , the two surface states at both surfaces of the film interact so strongly that they split into bonding and anti - bonding states , and move away from the fermi energy . on this prototype system ,
the @xmath56-factor of the fe adatom is similar to the one obtained on cu(111 ) , _ i.e. _ , @xmath85 .
also , regular @xmath56-values were obtained for the other adatoms .
this indicates that the surface state and its correct position play an important role in the large fe @xmath56-shift .
thus the presence and properties of a surface state can be crucial factors in determining the spin - dynamics of some adatoms .
while there is no simple picture of how the interplay of the surface - state electronic structure combined with that of the fe atoms produces the observed @xmath56-shift , a specific trend can be seen in the low - frequency properties of the susceptibility , which may explain the observed @xmath56-shift .
since we probe frequencies far below the electronic scale , we can expand @xmath86 in powers of @xmath8 if desired .
for small frequencies , it can be shown that the real part @xmath87 and the imaginary part @xmath88 are linear functions of @xmath8 , _
i.e. _ , @xmath89 and @xmath90 . as mentioned earlier , @xmath88 describes the density of stoner modes , while @xmath67 and @xmath91 are the slopes defining the linear frequency dependence .
analytically , @xmath92 is the product of the spin - dependent adatom density of states at @xmath83 , while @xmath67 is more involved but can be expressed in terms of single - particle gfs evaluated at @xmath83 . in fig .
[ fig_theory3](d ) , the real part of @xmath9 for cr , mn , fe and co adatoms with no applied field is plotted against frequency , and indeed we observe a linear behavior that is not modified with an applied field .
interestingly , @xmath67 is the smallest for fe adatom , which has a major contribution to its large @xmath56-shift .
in fact , by plugging the linear behavior of @xmath93 into eq . [ dyson ] and evaluating the imaginary part of @xmath6 we obtain an equation similar to what is given in @xcite : @xmath94 ^ 2 + ( u\alpha\omega)^2}\label{eq : th2}\ ] ] that has a resonance at @xmath95 . if no external magnetic field is applied along the @xmath14-direction , @xmath96 @xcite and the resonance moves to zero frequency , the goldstone mode ( see eq .
[ sumrule2 ] ) .
this is an important result : the position of the resonance , and thus the @xmath56-shift , depends equally on the slope of @xmath87 , the slope of @xmath88 as well as the change of @xmath87 at zero frequency .
thus , the right combination of properties must be satisfied in order to observe a large @xmath56-shift as obtained for the fe adatom .
magnetic properties of adatoms on the pt(111 ) surface are the subject of tremendous interest since the discovery by gambardella and coworkers @xcite of the large magnetic anisotropy energy ( mae ) of co adatoms .
a large mae is of utmost importance for the ultimate goal of stabilizing a single magnetic adatom and for the possibility of storing magnetic information down to the single - atom level .
recently , ists measurements by khajetoorians _
et al . _
@xcite of fe adatoms on pt(111 ) led to spin - excitation spectra different from those previously measured by balashov _
_ @xcite . indeed , the experiments reported in ref .
@xcite show that fe adatoms on pt(111 ) exhibit a very low mae and long precessional lifetimes , and moreover that these properties are strongly dependent on which hollow site an adatom occupies ( fcc or hcp , according to how they stack with respect to the sub - surface pt layer ) .
in contrast , those reported in ref .
@xcite suggest maes of the order of 10 mev for co adatoms and 6 mev for fe adatoms without distinguishing between the two possible hollow sites .
it is interesting to note that , contrary to balashov and coworkers , the experiments of khajetoorians _
involved the application of an external magnetic field which verified the spin - nature of the observed resonances .
to unravel the origin of the site - dependent mae for fe adatoms , we analyzed the connection between the mae and the binding site using the full - potential version of the kkr method @xcite .
pt(111 ) is a notoriously challenging substrate since its high magnetic polarizability @xcite surrounds the magnetic adatom with an extended spin polarization cloud , as seen for pd @xcite ; in this light we carefully checked all calculations .
considering the fe adatom relaxed by 20% of the bulk vertical interlayer distance towards the surface , we obtained fig .
[ mae ] depicting the dependence of the mae on the number of substrate pt atoms included in the self - consistent simulations .
mae versus number of atoms in cluster from kkr - gf calculations , for two different surface positions , fcc and hcp ( see text ) .
positive for preferred in - plane orientation and negative for out - of - plane .
the change of easy axis with increasing number of substrate pt atoms is only present for the hcp site . ]
the computed spin moments are , for the fcc site : @xmath97 ( @xmath98 ) ; hcp site : @xmath99 ( @xmath100 ) for the adatom ( whole cluster 220 pt atoms ) , respectively .
the orbital moments are for fcc site @xmath101 ( @xmath102 ) ; hcp site @xmath103 ( @xmath102 ) .
the mae yields @xmath104 mev ( out - of - plane ) and @xmath105 mev ( in - plane ) . here
, it was crucial to include a large number of substrate atoms , in order to converge the calculation and to reproduce the observed differences between the fcc site and the hcp site . for a cluster with 1012 pt atoms , calculations of both hcp- and
fcc - sites yield an out of plane easy axis , with convergence of the calculation occurring only when including more pt atoms .
the small value of the anisotropy for hcp - site is cautionary but mostly in line with those based on a supercell kkr - method @xcite , although the reported easy axis was found to be out - of - plane for both binding sites . for the fcc site
, the mae favors an out - of - plane configuration , and so the behavior of the spin - excitation spectra is expected to be similar to what was observed for fe adatoms on cu(111 ) and ag(111 ) substrates , _ i.e. _ , the resonance energy increases linearly with the magnetic field . however , in the hcp site the magnetic moments lies in - plane , and so , after applying a magnetic field normal to the surface , a gradual reorientation of the magnetic moment out of the surface plane will occur for increasing values of the magnetic field , leading to a plateau - like behavior of the resonance energy .
it is only when the magnetic field is large enough to completely reorient the magnetic moment out - of - plane that one recovers the usual linear behavior of the resonance energy versus magnetic field ( see ref .
@xcite ) .
density of spin excitations for ( a ) fe@xmath106 and ( b ) fe@xmath107 from td - dft kkr - gf calculations .
the curve marked @xmath81 t shows the effect of the offset field which attempts to mimic the mae .
the legend lists what @xmath108 value was added in each calculation to this initial offset field .
the dashed lines correspond to the larmor frequency ( @xmath109 ) obtained for the case of @xmath110.,title="fig : " ] density of spin excitations for ( a ) fe@xmath106 and ( b ) fe@xmath107 from td - dft kkr - gf calculations .
the curve marked @xmath81 t shows the effect of the offset field which attempts to mimic the mae .
the legend lists what @xmath108 value was added in each calculation to this initial offset field .
the dashed lines correspond to the larmor frequency ( @xmath109 ) obtained for the case of @xmath110.,title="fig : " ] from the imaginary part of @xmath6 ( see fig . [ susc_pt ] ) , which gives the density of states for spin excitations , we extract @xmath56 and @xmath78
. we obtain @xmath111 and @xmath112 , which , although not as dissimilar as the measured values ( @xmath113 and @xmath114 ) , are still bracketed by them . inputting the experimental @xmath115 for both cases ,
the calculated @xmath78 is found to be larger for hcp - site ( 4.8 ps ) than for fcc - site ( 1.2 ps ) , as experimentally obtained ( @xmath116 ps and @xmath117 ps ) . as spin - orbit coupling
was not included in these calculations , it is possible that it can modify the computed values of the @xmath56-factor and the lifetime ; however , given that the trends are reproduced , no qualitative impact is expected .
once more the shift in @xmath56 and the reduction of the lifetime for increasing magnetic field result from spin - dependent scattering by conduction electrons ( stoner excitations ) which damp the spin precession . unlike fe adatoms on both cu(111 ) and ag(111 ) , those on the pt(111 ) surface
show comparatively larger precessional lifetimes ( due to the lower excitation energies ) , which interestingly decrease more weakly ( @xmath118 ) in a magnetic field than in those other systems .
using our recently developed method , we investigated several 3@xmath12 magnetic adatoms and dimers deposited on noble substrates . our formalism , based on td - dft invoking the adiabatic approximation for the exchange and correlation kernel , demonstrated the origin of the lifetime and of the energy - shifts of the spin - excitation signature .
indeed , these main characteristics of the spin - excitations are determined by the electronic structure and can be extracted from the dynamical transverse ks susceptibility of the investigated materials .
the latter quantity describes electron - hole excitations within td - dft .
for instance , its imaginary part provides the density of electron - hole excitations .
we have demonstrated that the properties of the excitations spectra depend dramatically on the chemical nature of the adsorbates , their neighborhood and the substrates . in general ,
good agreement is found with the experimental measurements performed by ists . in the near future ,
we plan to extend our formalism by including spin - orbit coupling in a self - consistent manner as well as by evaluating the interactions between electrons and the spin - excitations which would allow to evaluate the impact on the electronic structure .
we would like to acknowledge the large contribution of late d. l. mills .
he was a mentor to us and a precursor of the realistic description of spin - dynamics of single impurities .
part of the presented work was done in collaboration with the experimental team involving t. schlenk , b. chilian , a. a. khajetoorians , j. wiebe and r. wiesendanger . we thank them for the several discussions and common works on this topic .
we also acknowledge the support of the hgf - yig programme vh - ng-717 ( functional nanoscale structure and probe simulation laboratory
funsilab ) .
a. j. heinrich , j. a. gupta , c. p. lutz , d. m. eigler , science * 306 * , 466 ( 2004 ) ; c. f. hirjibehedin , c .- y .
lin , a. f. otte , m. ternes , c. p. lutz , b. a. jones , a. j. heinrich , science * 317 * , 1199 ( 2007 ) .
t. balashov , t. schuh , a. f. takacs , a. ernst , s. ostanin , j. henk , i. mertig , p. bruno , t. miyamachi , s. suga , and w. wulfhekel , phys .
. lett . * 102 * , 257203 ( 2009 ) .
s. y. savrasov , phys .
lett . * 81 * , 2570 ( 1998 ) .
j. b. staunton , j. poulter , b. ginatempo , e. bruno , and d. d. johnson , phys .
lett . * 82 * , 3340 ( 1999 ) p. buczek , a. ernst , p. bruno , and l. m. sandratskii , phys .
lett . * 102 * , 247206 ( 2009 ) .
r. b. muniz , d. l. mills , phys .
b * 68 * , 224414 ( 2003);_ibid . _ * 66 * , 174417 ( 2002 ) a. t. costa , r. b. muniz , d. l. mills , phys . rev .
b * 70 * , 054406 ( 2004 ) ; _ ibid . _ * 73 * , 054426 ( 2006 ) ; a. t. costa , r. b. muniz , s. lounis , a. b. klautau , d. l. mills , _ ibid . _ * 82 * , 014428 ( 2010 ) ; a. t. costa , r. b. muniz , d. l. mills , phys .
lett . * 94 * , 137203 ( 2005 ) .
g. j. nieuwenhuys , adv . phys . * 24 * , 515 ( 1975 )
. a. oswald , r. zeller , and p. h. dederichs , phys .
lett . * 56 * , 1419 ( 1986 ) .
p. blonski , a. lehnert , s. dennler , s. rusponi , m. etzkorn , g. moulas , p. bencok , p. gambardella , h. brune , and j. hafner , phys .
b * 81 * , 104426 ( 2010 ) . | we review our recent work on the simulation , description and prediction of spin - excitations in adatoms and dimers deposited on metallic surfaces .
this work done together with douglas l. mills , is an extension of his seminal contribution ( with pascal lederer ) published 50 years ago on the spin - dynamics of transition metal impurities embedded in transition metal hosts [ p. lederer , d.l .
mills , phys . rev . * 160 * , 590 ( 1967 ) ] .
the main predictions of his model were verified experimentally with state of the art inelastic scanning tunneling spectroscopy on adatoms .
our formalism , presented in this review , is based on time - dependent density functional theory , combined with the korringa - kohn - rostoker green function method .
comparison to experiments is shown and discussed in detail .
our scheme enables the description and prediction of the main characteristics of these excitations , _
i.e. _ their resonance frequency , their lifetime and their behavior upon application of external perturbations such as a magnetic field . |
graph analytics has drawn much attention from research and industry communities , due to the wide applications of graph data in different domains .
one of the major issues in graph analytics is identifying cohesive subgraphs .
there are lots of indexes to depict the cohesiveness of a graph , such as cliques , k - truss , k - core , f - groups , n - clans and so on @xcite , among which @xmath0-core is recognized as one of the most efficient and helpful one . given a graph @xmath1 ,
the @xmath0-core is the largest subgraph in @xmath1 , such that the minimum degree of the subgraph is at least @xmath0 .
the core number of a vertex @xmath2 is defined as the largest @xmath0 such that there exists a @xmath0-core containing @xmath2 . in static graphs ,
the computation of the core number of each vertex is known as the @xmath0-core decomposition problem . besides the analysis of cohesive subgroup ,
@xmath0-core decomposition are widely used in a large number of applications to analyze the structure and function of a network .
for example , the k - core decomposition can be used to analyze the topological structure of internet @xcite , @xcite , to identify influential spreader in complex networks @xcite @xcite , to analyze the structure of large - scale software systems @xcite@xcite@xcite@xcite , to predict the function of biology network @xcite , and to visualize large networks @xcite@xcite and so on . in static graphs ,
the @xmath0-core decomposition problem has been well studied .
the algorithm presented in @xcite can compute the core number of each vertex in @xmath3 time , where @xmath4 is the number of edges in the graph .
however , in many real - world applications , graphs are subject to continuous changes like insertion or deletion of vertices and edges . in such dynamic graphs ,
many applications require to maintain the core number for every vertex online , given the network changes over time .
but it would be expensive to recompute the core numbers of vertices after every change of the graph , though the computation time is linear , as the size of the graph can be very large .
furthermore , the graph change may only affect the core numbers of a small part of vertices .
hence , the _ core maintenance _ problem @xcite is recommended , which is to identify the vertices whose core numbers will be definitely changed and then update the core numbers of these vertices .
there are two categories of core maintenance , _ incremental _ and _ decremental _ , which handle edge / vertex insertion and deletion respectively .
previous works focus on maintaining the core numbers of vertices in the scenario that a single edge is inserted or deleted from the graph . for multiple edge
/ vertex insertions / deletions , the inserted / deleted edges are processed sequentially .
the sequential processing approach , on the one hand , incurs extra overheads when multiple edges / vertices are inserted / deleted , as shown in fig .
[ fig1 ] , and on the hand , it does not fully make use of the computation power provided by multicore and distributed systems .
therefore , it is necessary to investigate the parallelism in the edge / vertex processing procedure and devise parallel algorithm that suits to implement in multicore and distributed systems .
but to the best of our knowledge , there are no known parallel algorithms proposed for the core maintenance problem . in the core maintenance problem ,
the insertions / deletions of vertices can be handled by implementing an edge insertion / deletion algorithm .
specifically , inserting a vertex is equivalent to the following process : first inserting the vertex into the graph by setting its core number as 0 , and then inserting the edges connected to the new vertex .
similarly , the deletion of a vertex is equivalent to the process that deleting the edges connected to the vertex and finally deleting the vertex .
hence , in this paper , we only consider the edge insertions and deletions .
it is a very difficult task to design parallel algorithms for core maintenance in dynamic graphs .
different from the single edge insertion / deletion case , where the core number of each vertex changes by at most 1 , it is hard to identify the change of a vertex s core number in the multiple edge insertion / deletion scenario , as the change of a vertex core number may be affected by several inserted edges . an intuitive manner is to split the inserted / deleted edges into sets that affect disjoint sets of vertices in the original graph .
however , the parallelism of this manner is poor . in this work ,
we take a more efficient approach that exhibits better parallelism .
specifically , we propose a structure called _ superior edge set_. the inserted / deleted edges can be split into multiple superior edges sets , and for each vertex connected to inserted / deleted edges , a superior edge set contains at least one inserted edge connected to it .
it is shown that the insertion or deletion of edges in a superior edge set can change the core number of every vertex by at most 1 .
hence , the core numbers of vertices when inserting or deleting a superior edge set can be maintained using a parallel procedure : first identifying the vertices whose core numbers will change due to the insertion or deletion of every edge in parallel , and then updating the core number of these vertices by 1 .
a parallel algorithm can then be obtained by iteratively handling the insertions / deletions of split superior edge sets using the above parallel procedure . in summary ,
our contributions are summarized as follows .
* we propose a structure called _ superior edge set _ , and show that if the edges of a superior edge set is inserted into / deleted from a graph , the core number of each vertex can change by at most 1 .
it implies that the insertion / deletion of these edges can be processed in parallel .
we also give sufficient conditions for identifying the vertices whose core numbers will change , when inserting / deleting a superior edge set . *
we then present parallel algorithms for incremental and decremental core maintenance respectively .
comparing with sequential algorithms , our algorithms reduce the number of iterations for processing @xmath5 inserted / deleted edges from @xmath5 to the maximum number of edges inserted to each vertex . in large - scale graphs , the acceleration is significant , since each vertex can connect to only a few inserted or deleted edges .
for example , as shown in the experiments , even if inserting @xmath6 edges to the livejournal graph ( refer to table [ table_graph ] in section [ sec : experiment ] ) , the number of iterations is just 3 in our parallel algorithms , in contrast with @xmath6 ones in sequential processing algorithms .
we also conduct extensive experiments over both real - world and synthetic graphs , to evaluate the efficiency , stability and scalability of our algorithms .
the results show that comparing with sequential processing algorithms , our algorithms significantly speed up core maintenance , especially in cases of large - scale graphs and large amounts of edge insertions / deletions .
[ fig1 ] and @xmath7 are inserted . *
traversal * algorithm in @xcite processes them one by one .
first for edge @xmath8 , it will visit vertices @xmath9 and update their core numbers . and
then when inserting @xmath7 , it will visit @xmath10 , and update core numbers of @xmath11 .
however in our parallel algorithm , edges @xmath8 and @xmath7 are handled in parallel using two processes . in the process handling @xmath8 ,
the algorithm execution will visit and update @xmath9 , and in another process for @xmath7 , the algorithm will visit and update @xmath11 .
hence , the parallel algorithm avoids duplicate visiting of @xmath12.,title="fig:",width=192 ] the rest of this paper is organized as follows . in section [ sec : relate ] , we briefly review closely related works . in section [ sec : problem ] , the problem definitions are given .
theoretical results supporting the algorithm design are presented in section [ sec : basis ] .
the incremental and decremental parallel algorithms are proposed in section [ sec : in ] and section [ sec : de ] respectively . in section
[ sec : experiment ] , the experiment results are illustrated and analyzed . the whole paper is concluded in section [ sec : conclusion ] .
in static graphs , the core decomposition problem has been extensively studied .
the state - of - the art algorithm was given in @xcite , the runtime of which is linear in the number of edges . in @xcite ,
an external - memory algorithm was proposed when the graph is too large to hold in memory .
core decomposition in the distributed setting was studied in @xcite .
the above three algorithms were compared in @xcite under the graphchi and webgraph models .
parallel core decomposition was studied in @xcite .
core maintenance in dynamic graphs has also been widely studied .
however , all previous works focus on the case of single edge insertion / deletion , and sequentially handle multiple edge insertions / deletions .
efficient algorithms were proposed in @xcite . in @xcite ,
an algorithm was proposed to improve the i / o efficiency .
furthermore , @xcite and @xcite solved the core maintenance problem in the distributed environment .
we consider an undirected , unweighted simple graph @xmath13 , where @xmath14 is the set of vertices and @xmath15 is the set of edges .
let @xmath16 and @xmath17 .
for a node @xmath18 , the set of its neighbors in @xmath1 is denoted as @xmath19 , i.e. , @xmath20 .
the number of @xmath21 s neighbors in @xmath1 is called the degree of @xmath21 , denoted as @xmath22 .
so @xmath23 .
the maximum and minimum degree of nodes in @xmath1 is denoted as @xmath24 and @xmath25 respectively .
we next give formal definitions for the _ core number _ of a vertex and other related concepts .
given a graph @xmath26 and an integer @xmath0 , the @xmath0-core is a maximal connected subgraph @xmath27 of @xmath1 , where each vertex has at least @xmath0 neighbors in @xmath27 , i.e. , @xmath28 . given a graph @xmath26 , the core number of a vertex @xmath29 , denoted by @xmath30 , is the the largest @xmath0 , such that there exists a @xmath0-core containing @xmath21 . for simplicity
, we use @xmath31 to denote @xmath30 when the context is clear .
the max - k - core associated with a vertex @xmath21 , denoted by @xmath32 , is the @xmath0-core with @xmath33 . in this work
, we aim at maintaining the core numbers of vertices in dynamic graphs .
specifically , we define two categories of graph changes : _ incremental _ , where a set of edges @xmath34 are inserted to the original graph , and _ decremental _ , where a set of edges are deleted . based on the above classification , we distinguish the core maintenance problem into two scenarios , as defined below . given a graph @xmath26 , the incremental core maintenance problem is to update the core numbers of vertices after an incremental change to @xmath1 . given a graph @xmath26
, the decremental core maintenance problem is to update the core numbers of vertices after a decremental change to @xmath1 . in this paper , we proposed efficient parallel algorithms for both the incremental and decremental core maintenance problem , which are named the * superiorinsert * and * superiordelete * algorithm .
the main idea of our algorithms is to find the definitely updated vertex set when the graph changes , which was first proposed in @xcite . besides
, our algorithms can deal with both edge and vertex insertions / deletions , this to say , we can deal with the insertion and deletion of an arbitrary new graph @xmath35 .
in this section , we give some theoretical lemmas that constitute the theoretical basis of our algorithms . at first , we introduce some definitions . given a graph @xmath26 , an edge @xmath36
is called a superior edge for @xmath21 if @xmath37 .
notice that in the definition , we do not require @xmath38 , i.e. , @xmath39 may be an edge that is about to insert to graph @xmath1 .
furthermore , we define the core number of an edge as the smaller one of its endpoints , i.e. , @xmath40 . an edge set @xmath41 is called an @xmath0-superior edge set , if for each edge @xmath42 , it satisfies : \(i ) @xmath43 is a superior edge with core number @xmath0 .
( ii ) if @xmath43 and @xmath44 have an common endpoint @xmath45 , @xmath46 . in other words , in a @xmath0-superior edge set @xmath47 , each edge is a superior edge for a vertex with core number @xmath0 , and in @xmath47 , each vertex connects to at most one superior edge for it .
the union of several @xmath0-superior edge sets with distinct @xmath0 values is called a _
superior edge set_. it can be known that in a superior edge set , each vertex can still connect to at most one superior edge for it . in the following , we will first show that when inserting / deleting a superior edge set ( lemma [ them : superioredgesetinsert ] and lemma [ them : superioredgesetdelete ] ) , the core number of every vertex can change by at most 1 , and then give a sufficient condition for identifying vertices whose core numbers change ( lemma [ le : exkpaths ] , lemma [ corollary : csd ] and lemma [ corollary : sd ] ) .
we first prove a result on the core number increase of every vertex when inserting a @xmath0-supeior edge set . for simplicity
, we use @xmath31 to denote @xmath30 when the context is clear .
[ lem : k - superioredgesetinsert ] given a graph @xmath26 , if a @xmath0-superior edge set @xmath48 is inserted to @xmath1 , where @xmath49 , for each node @xmath2 , it holds that : + @xmath50 if @xmath51 , @xmath52 can increase by at most 1 ; + @xmath53 if @xmath54 , @xmath52 will not change . for @xmath50 , we need to show that for a vetex @xmath2 with @xmath55 , @xmath52 can increase by at most 1 .
otherwise , assume @xmath52 increases by @xmath56 to @xmath57 , where @xmath58 .
let @xmath59 and @xmath60 be the max-@xmath0-core of @xmath2 before edge insertion and the max-@xmath61-core of @xmath2 after edge insertion respectively .
then , @xmath62 .
it can be concluded that one of inserted edges must belong to @xmath60 , as otherwise @xmath63 before insertion as well .
let @xmath64 . for a vertex @xmath65 ,
if @xmath66 , the degree of @xmath21 does not change when deleting the edges in @xmath47 , so @xmath67 . if @xmath68 , @xmath21 can lose at most one neighbor that is connected by a superior edge for it in @xmath47 , so @xmath69 . if @xmath70 , @xmath21 must have at least @xmath0 + 1 neighbors whose core numbers are not smaller than @xmath0 + 1 in @xmath1 . we add the vertices whose @xmath71 is larger than @xmath0 back to @xmath72 , and denote the induced graph as @xmath73 . it can be obtained that @xmath74 . and
from @xmath1 to @xmath73 , @xmath21 does not lose any neighbor whose @xmath71 is not smaller than @xmath0 + 1 .
hence , in @xmath73 , @xmath75 . then it can be seen that each vertex in @xmath73 has a degree at least @xmath0 + 1 , i.e. , @xmath76 .
this means that @xmath77 , which contradicts with @xmath78 .
hence , @xmath52 can increase by at most 1 . for @xmath53 , we need to show that for a vertex @xmath2 if @xmath54 , @xmath52 can not change .
we consider two cases : @xmath79 and @xmath80 .
assume @xmath52 = @xmath81 increases by @xmath56 to @xmath82 , where @xmath83 .
let @xmath59 and @xmath60 be the max-@xmath81-core of @xmath2 before edge insertion and the max-@xmath84-core after edge insertion respectively .
then we have @xmath85 , @xmath86 .
we first consider the @xmath87k@xmath88 case .
there must be at least one of the edges @xmath43 in @xmath47 belonging to @xmath60 , as otherwise @xmath89 before edge insertion .
consider the edge @xmath43 .
at least one of its endpoints has a core number @xmath0 , since @xmath47 is a @xmath0-superior edge set .
denote by @xmath45 the endpoint of @xmath43 with core number @xmath0 .
as shown before , @xmath90 can increase by at most 1 .
hence , after the edge insertion , @xmath91 .
this means that @xmath45 is not in @xmath60 , which is a contradiction .
therefore , if @xmath92 , @xmath52 will not change after the edge insertion .
we next consider the @xmath93 case .
similar as beofore , it can be shown that at least one of the edges @xmath94 in @xmath47 that is contained in @xmath60 .
let @xmath64 .
let @xmath21 be a vertex in @xmath72 , we consider three cases . if @xmath68 and as proved before , it can be obtained that @xmath95 .
if @xmath96 , as shown before , the core number of @xmath21 will not be affected by the edge insertions .
if @xmath66 , because @xmath86 , and @xmath21 does not connect to edges in @xmath47 , we can get that @xmath97 .
let @xmath98 .
based on above , it can be obtained that @xmath72 is a @xmath5-core and @xmath99 . but
this contradicts with the fact that @xmath100
. then we can get that the core number of @xmath2 does not change after inserting @xmath47 . combining all above together
, the lemma is proved .
using a similar argument as that for proving lemma [ lem : k - superioredgesetinsert ] , we can get that the core number changes of vertices after deleting a @xmath0-superior edge set from graph @xmath1 , as given in the following lemma .
[ lem : k - superioredgesetdelete ] given a graph @xmath26 , if a @xmath0-superior edge set @xmath48 is deleted from @xmath1 , where @xmath49 , for each vertex @xmath2 , it holds that : + @xmath50 if @xmath51 , @xmath52 can decrease by at most 1 ; + @xmath53 if @xmath54 , @xmath52 will not change . from the above lemma [ lem : k - superioredgesetinsert ] and lemma [ lem : k - superioredgesetdelete ] , we have known that for a graph @xmath26 , after a @xmath0-superior edge set @xmath48 is inserted into or deleted from @xmath1 , only vertices with core numbers @xmath0 may increase / decrease , and the change is at most 1 .
this implies that if a @xmath0-superior edge set is inserted / deleted , it will be enough to only visit vertices whose core numbers are @xmath0 and check if their core numbers will be updated . and because the core numbers of these vertices can change by at most 1 , we can handle these edge insertions in parallel : first we find the update set of vertices that will change core numbers because of the insertion of each particular edge in parallel , and the union of these update sets is just the set of vertices whose core numbers will change by one . in fact , we can get even better results , which are given in the following lemma [ them : superioredgesetinsert ] and lemma [ them : superioredgesetdelete ] .
[ them : superioredgesetinsert ] given a graph @xmath26 and a superior edge set @xmath101 , where @xmath102 for @xmath103 is a @xmath104-superior edge set and @xmath105 if @xmath106 , it holds that after inserting @xmath107 into @xmath1 , the core number of each vertex @xmath21 can increase by at most 1 .
it can be seen that inserting edges in @xmath107 into @xmath1 all together has the same result with inserting @xmath102 one by one .
we next assume @xmath102 are inserted one by one . to prove the lemma
, we need to prove that if inserting @xmath102 makes a vertex increase its core number from @xmath104 to @xmath104 + 1 , its core number can not change any more when inserting @xmath108 for @xmath109 .
clearly , we only need to prove the above result for @xmath110 .
there are two cases we need to consider .
if @xmath111 , by lemma [ lem : k - superioredgesetinsert ] , the core number of @xmath21 will not increase any more when inserting @xmath110 , since only vertices with core numbers of @xmath112 may increase their core numbers .
we next consider the case of @xmath113 .
we claim that if there is a vertex increasing its core number from @xmath104 to @xmath114 after the insertions of @xmath102 and @xmath110 , the vertex must have a neighbor which increases the core number from @xmath104 to @xmath114 as well during the insertions .
let @xmath21 be a vertex whose core number is increased from @xmath104 to @xmath114 after inserting @xmath102 and @xmath110 .
notice that @xmath21 does not connect to edges in @xmath110 .
hence , the degree of @xmath21 does not change when inserting @xmath110 .
furthermore , by lemma [ lem : k - superioredgesetinsert ] , the core number of each neighbor of @xmath21 can be increased by at most 1 .
so @xmath21 has at least @xmath114 neighbors whose core numbers are not smaller than @xmath115 and some of these neighbors have a core number of @xmath115 .
denote by @xmath116 the vertices in @xmath19 whose core numbers are @xmath115 before inserting @xmath110 .
it can be obtained that there must be a vertex @xmath117 whose core number is @xmath104 before inserting @xmath102 , as otherwise , the core number of @xmath21 is @xmath115 before inserting @xmath102 , which contradicts with our assumption .
let @xmath118 denote the set of vertices whose core numbers change from @xmath104 to @xmath114 after the insertions of @xmath102 and @xmath110 .
because inserting @xmath110 does not change the degrees of vertices in @xmath118 , there must be a vertex @xmath119 whose core number change is caused because of the core number change of vertices in @xmath120 , as otherwise no vertex in @xmath118 can change the core number .
let @xmath121 be a vertex in @xmath120 whose core number change causes the core number change of @xmath122 .
then @xmath123 is @xmath115 before inserting @xmath102 and is increased to @xmath114 after inserting @xmath110 . to make @xmath122 increase its core number to @xmath114 after inserting @xmath110
, there must be at least @xmath114 neighbors in @xmath120 whose core numbers are initially not smaller than @xmath115 before inserting @xmath102 and @xmath110 .
it concludes that @xmath124 before inserting @xmath102 and @xmath110 .
however , this contradicts with the fact that @xmath125 is @xmath104 before insertions .
the contradiction shows that if the core number of a vertex is changed when inserting @xmath102 , its core number will not change any more when inserting @xmath110 . combining all above together ,
the lemma is prove .
similarly , for the case of a superior edge set deletion , we have the following result .
[ them : superioredgesetdelete ] given a graph @xmath26 and a superior edge set @xmath126 , where @xmath102 for @xmath103 is a @xmath104-superior edge set and @xmath105 if @xmath106 , it holds that after deleting @xmath107 from @xmath1 , the core number of each vertex @xmath21 can decrease by at most 1 . in above
, we have shown that when inserting or deleting a superior edge set from a graph , the core numbers of vertices can change by at most 1 .
this implies that the core updates of inserting / deleting edges in a superior edge set can be processed in parallel by distributing distinct @xmath0-superior edge sets to distinct processes .
furthermore , we have also shown which set of vertices may change due to the insertion or deletion of a @xmath0-superior edge set .
in the subsequent section , we give more accurate conditions for a vertex to change its core number when inserting / deleting a superior edge set .
we first introduce some notations .
[ de : sd ] for a vertex @xmath21 in a graph @xmath1 , @xmath2 is a _ superior neighbor _ of @xmath21 if the edge @xmath127 is a _ superior edge _ of @xmath21 .
the number of @xmath21 s superior neighbors is called the _ superior degree _ of @xmath21 , denoted as @xmath128 .
it can be known that only superior neighbors of a vertex may affect the change of its core number .
the _ constraint superior degree _
@xmath129 of a vertex @xmath21 is the number of @xmath21 s neighbors @xmath122 that satisfies @xmath130 or @xmath131 . for a vertex @xmath21 ,
its constraint superior degree @xmath129 is the number of @xmath21 neighbors @xmath122 , that has a larger core number than @xmath21 or has the same core number but has enough neighbors to support itself to increase core number .
[ def : kpt_u ] for a vertex @xmath21 with a core number @xmath31 , the _ @xmath132-path - tree _ of @xmath21 is a dfs tree rooted at @xmath21 and each vertex @xmath122 in the tree satisfies @xmath133 . for simplicity we use @xmath134 to represent k - path - tree of @xmath21 .
the * @xmath134 * includes all vertices @xmath122 with @xmath133 that are reachable from @xmath21 via paths that consists of vertices with core numbers equal to @xmath31 .
when a superior edge of @xmath21 is inserted or deleted , as shown in lemma [ lem : k - superioredgesetinsert ] , only vertices in @xmath134 may change their core numbers . and for the insertion case , a more accurate condition was given in @xcite for identifying the set of vertices that may change core numbers , as shown below .
[ them : kpath ] given a graph @xmath26 , if an edge @xmath127 is inserted and @xmath135 , then only vertices @xmath122 in the * @xmath134 * of @xmath21 and @xmath136 may have their core numbers increased , and the increase is no more than 1 .
however , the above lemma [ them : kpath ] is just suitable for the one edge insertion scenario .
we next generalize the above result to the scenario of inserting a @xmath0-superior edge set , as shown in lemma [ them : exk - path ] below , which will help find the set of vertices with core number changes when inserting multiple edges . before giving the result
, we need to generalize the concept of @xmath132-path - tree to @xmath137-path - tree .
[ def : exk - path - tree ] for a @xmath0-superior edge set @xmath47 = \{@xmath138 } , w.l.o.g . ,
assume that for each @xmath43 = @xmath139 , @xmath140 .
the union of @xmath141 for every @xmath142 is called the _ exk - path - tree _ of @xmath47 . for simplicity
we use @xmath143 to represent exk - path - tree of @xmath47 . by lemma
[ them : kpath ] , we can get that when inserting a @xmath0-superior edge set @xmath47 , only vertices @xmath122 in the @xmath143 satisfying @xmath144 may have their core numbers change , and lemma [ lem : k - superioredgesetinsert ] ensures that these vertices can change their core numbers by at most 1 .
this result is summarized in the following lemma .
[ them : exk - path ] given a graph @xmath26 , if a @xmath0-superior edge set @xmath47 is inserted , then only vertices @xmath122 in the @xmath143 satisfying @xmath144 may have their core numbers increased , and the core change is at most 1 .
we only prove the insertion case , and the deletion case can be proved similarly .
notice that the core number changes of vertices are the same in scenarios of inserting all edges in @xmath47 together and inserting the edges in @xmath47 one by one .
so we can assume that the edges in @xmath47 are inserted one by one .
by lemma [ ] first we should know that , when a @xmath0-superior edge set @xmath47 = \{@xmath145 , @xmath146 , ... , @xmath147}is inserted / deleted , the vertices who update cores is the same as when inserting / deleting edges in @xmath47 one by one .
assume @xmath148 and core(@xmath149 ) @xmath150 core(@xmath142 ) = @xmath0 , after inserting / deleting @xmath43 into the graph , vertices who update cores construct a vertex set @xmath151 , after inserting / deleting @xmath47 the update vertex set is @xmath152 .
so we can have @xmath152 = @xmath153 and @xmath154 , where @xmath155 , as otherwise a vertex may update its core more than once , which is a contradiction to theorem [ lem : k - superioredgeset ] . and according to theorem [ them : kpath ] , vertices @xmath122 in @xmath151 are all in the * @xmath141 * of @xmath142 and sd(@xmath122 ) @xmath156 core(@xmath21 ) = @xmath0 .
so all vertices @xmath122 in @xmath152 are in the @xmath143 and satisfies sd(@xmath122 ) @xmath157 .
the proof is completed .
the theorem [ them : exk - path ] identifies vertices who may change cores after inserting a @xmath0-superior edge set @xmath47 .
further , for a superior edge set we can similarly have the following theorem .
[ def : superior - tree ] for a superior edge set @xmath158 = @xmath159 , @xmath160 is a @xmath0-superior edge set where @xmath161 .
we denote the union of @xmath143 for all @xmath160 as the * superior - tree * of @xmath158 .
the vertices in the * superior - tree * is denoted as * @xmath162 * , and all neighbors of all vertices in * @xmath162 * is defined as * @xmath163*. [ them : superior - tree ] given a graph @xmath26 , if a superior edge set @xmath158 is inserted / removed then only vertices @xmath122 in the * superior - tree * and sd(@xmath122 ) @xmath157 where @xmath0 is the @xmath0-superior edge set @xmath122 belongs to , may have their core numbers increased / decreased .
the proof is similarly as theorem [ them : exk - path ] .
since the insertion / deletion of a @xmath0-superior edge set only affects vertices whose core is @xmath0 as proved before , we can insert / delete a @xmath0-superior edge set at a time , the vertices that update core numbers will be exactly the same as inserting / deleting them all at a time .
we denote all vertices update cores upon the insertion / deletion of @xmath158 as @xmath164 . besides as theorem [ them :
superioredgeset ] and theorem [ them : exk - path ] proves , all vertices can only update core number once and all update vertices @xmath122 are in an @xmath143 and satisfies sd(@xmath122 ) @xmath157 the above lemma [ them : exk - path ] implies that after an edge in a @xmath0-superior edge set @xmath47 is inserted , the vertices whose core numbers change during the insertion will not change any more when inserting other edges in @xmath47 . based on the above result and lemma [ them : superioredgesetinsert ] and lemma [ them : superioredgesetdelete ]
, we can get the set of vertices whose core number change when inserting a superior edge set .
[ le : exkpaths ] given a graph @xmath26 , if a superior edge set @xmath165 is inserted , then only vertices @xmath122 in every * * @xmath166**s of every @xmath102 for @xmath103 satisfying @xmath144 may have their core numbers increased , and the core number change can be at most 1 . by the definition of @xmath167
, we have the following result .
[ corollary : csd ] given a graph @xmath26 and a vertex @xmath21 with core number @xmath0 . after inserting a superior edge set into @xmath1 ,
if @xmath168 , then @xmath21 can not be in a @xmath169-core .
lemma [ le : exkpaths ] and lemma [ corollary : csd ] give accurate conditions to determine the set of vertices that will change the core numbers , after inserting a superior edge set . for deletion case , we have the following result , which can be obtained directly from definition [ de : sd ] .
[ corollary : sd ] after deleting an edge set from a graph @xmath26 , for @xmath18 , if @xmath31 = @xmath0 and @xmath170 , then @xmath21 will decrease its core number . in this section ,
we have given accurate conditions for the core number changes of vertices after inserting / deleting a superior edge set .
in the subsequent section [ sec : in ] and section [ sec : de ] , we will show how to utilize these theoretical results to design parallel algorithms for incremental and decremental core maintenance respectively .
in this section , we present how to find a superior edge set given an arbitrary edge set .
the main structure we proposed is called * superior edge set * , in this structure , each vertex has only one neighbor that has a core number no smaller than itself .
as theorem[them : superioredgeset ] proved , when such a structure is inserted into or deleted from a graph , all vertices can change its core number by at most one , which is easy to update . in this subsection ,
we present how to find such a * superior edge set*. since the hierarchy of k - core , the finding process can be executed in parallel .
each child process searches for the * superior edge set * of a given core number @xmath0 and return the * k - superior edge set * to the main process .
[ htb][alg : findksuperioredges ] * input * + the graph , @xmath26 ; + the update edge set , @xmath34 ; + the vertices set connected to edges in @xmath34 , @xmath171 a core number @xmath0 ; + let @xmath160 be an empty edge set return @xmath160 the algorithm [ alg : findksuperioredges ] finds a superior edge structure for a given @xmath132 in an arbitrary edge set .
the judgement in line 3 ensures that there can be only one superior edge for each vertex in the final result @xmath160 .
in this section , we present the algorithm for incremental core maintenance , whose pseudo - code is given in algorithm [ alg : superioredgeinsert ] .
we consider the core number update of vertices after inserting a set of edges @xmath34 to graph @xmath26 .
let @xmath171 denote the set of vertices connecting to edges in @xmath34 .
the set of core numbers of vertices in @xmath171 is denoted as @xmath172 .
[ htb][alg : superioredgeinsert ] * input * + the graph , @xmath26 ; + the inserted edge set , @xmath34 ; + the set of vertices @xmath171 connected to edges in @xmath34 the core number @xmath52 of each vertex in @xmath14 ; + [ htb][alg : findksuperioredges ] * input * + the graph , @xmath26 ; + the update edge set , @xmath34 ; + the set of vertices @xmath171 connected to edges in @xmath34 a core number @xmath0 ; + @xmath173 * return * @xmath47 [ htb][alg : insertk ] * input * + the graph , @xmath26 ; + the @xmath0-superior edge set , @xmath47 ; + the core number @xmath52 of each vertex in @xmath14 * initially * , @xmath174 empty stack for each vertex @xmath175 , @xmath176\gets{false } , removed[v]\gets{false } , cd[v]\gets{0}$ ] compute @xmath177 for each vertex @xmath2 in @xmath143 of @xmath47 * return * @xmath178 ; [ alg : insertremove ] @xmath179 @xmath180 @xmath181 \gets{true}$ ] the algorithm is executed in iterations . basically , the algorithm split the inserted edges into multiple superior edge sets , and process the insertion of one superior edge set in one iteration . in each iteration , it first uses a parallel algorithm to find a suporior edge set from the inserted edges that have not been processed so far ( line 3 ) .
then a parallel algorithm is executed for each edge in parallel to identify the set of vertices whose core numbers change , and increase the core numbers of these vertices by 1 ( line 6 - 7 ) .
it deserves to point out that we do not use directly algorithms handling single edge insertion / deletion as subroutine .
instead , we make the edges inserted with the same core number processed together , as we find that this can efficiently avioding duplicate visiting of vertices , which further accelerates our parallel processing procedure .
we next introduce the two parts in each iteration respectively .
because the superior edges of vertices with different core numbers are disjoint , the @xmath0-superior edge sets @xmath182 for different core numbers @xmath183 can be computed in parallel using algorithm [ alg : findksuperioredges ] . then the computed superior edge set is inserted into the graph and deleted from @xmath34 .
the set of vertices with core number changes is also computed in parallel .
specifically , for each @xmath184 , a child process is assigned to find the vertices whose core number changes are caused by the insertion of the computed @xmath0-superior edge set , using algorithm [ alg : insertk ] .
algorithm [ alg : insertk ] first computes @xmath185 values for each vertex in @xmath143 of @xmath47 , and then for each edge @xmath186 in a @xmath0-superior edge set , finds the set of vertices whose core numbers change due to the insertion of @xmath43 . for @xmath43 , a _ positive _ depth - first - search ( dfs )
is conducted on vertices in @xmath187 from the root vertex @xmath188 , which is one of @xmath142 or @xmath149 that has a core number @xmath0 and @xmath149 have a core number equal to @xmath0 , then @xmath188 can be either @xmath142 or @xmath149 . ] , to explore the set of vertices whose core numbers potentially change . in the algorithm ,
the @xmath189 value of each vertex @xmath2 is used to evaluate the potential of a vertex to increase its core number , which records the dynamic changes of @xmath167 value .
the intial value of @xmath190 is set as @xmath191 . for a vertex @xmath2 , if @xmath192\le k$ ] , its core number can not increase .
if a vertex @xmath2 with @xmath192\le k$ ] is traversed in the positive dfs procedure , a _ negative _ dfs procedure initiated from @xmath2 will be started , to remove @xmath2 and update the @xmath189 values of other vertices with core number @xmath0 .
after all vertices in @xmath187 are traversed , the vertices that are visited but not removed will increase the core numbers by 1 .
* performance analysis .
* we next analyze the correctness and efficiency of the proposed incremental algorithm . at first ,
some notations are defined , which will be used in measuring the time complexity of the algorithm . for graph @xmath26 , the inserted edge set @xmath34 and a subset @xmath193 of @xmath34 ,
let @xmath194 and @xmath195 be the set of core numbers of vertices in @xmath196 .
for @xmath196 , let @xmath197 .
as shown later , @xmath198 is the max times a vertex @xmath21 can be visited by negative dfs procedures in the algorithm execution . for @xmath199 , let @xmath200 be the set of vertices with core number @xmath0 , and @xmath201 be the neighbors of vertices in @xmath200 .
. denoted by @xmath203 $ ] the set of edges in @xmath196 that are connected to vertices in @xmath204 .
then we define @xmath205 as follows , which represents the max number of edges travelled when computing @xmath185 in the case of inserting edges to @xmath196 .
@xmath206|\}.\ ] ] furthermore , we define the _ maximum inserted degree _ as the maximum number of edges inserted to each vertex in @xmath14 , denoted as @xmath207 .
[ insertcorrectness ] algorithm [ alg : superioredgeinsert ] can update the core numbers of vertices after inserting an edge set @xmath34 in @xmath208 time . the algorithm is executed in iterations , and each iteration includes two parts .
the first part computes the superior edge set from unprocessed edges in @xmath34 by executing algorithm [ alg : findksuperioredges ] in parallel , and then inserts the computed superior edge set into graph @xmath1 . by lemma
[ them : superioredgesetinsert ] , after inserting a superior edge set into the graph , each vertex can increase its core by at most 1 .
after that , in the second part , we identifiy vertices that will increase core numbers by executing algorithm [ alg : insertk ] in parallel .
different processes deal with distinct @xmath0-superior edge sets , and visit vertices with distinct core numbers . for each inserted @xmath0-superior edge set @xmath47 , we conduct two kinds of operations : ( 1 ) a positive dfs that visits vertices in the exkpt , and ( 2 ) a negative dfs that will remove vertices that are confirmed not to increase core numbers . by lemma
[ le : exkpaths ] , visiting vertices in the exkpt is enough to find all vertices whose core numbers potentially increase caused by the insertion of @xmath47 . and by lemma [ corollary : csd ] , if a vertex @xmath122 satisfies @xmath209 \le k$ ] , @xmath122 will not increase its core number , and it will affect the potential of its neighbors to increase their core numbers . notice that this influence procedure should be spread across vertices in * @xmath210 * , which is done by the negative dfs procedure .
when all eges in @xmath47 are handled , the potential vertices are visited and the ones that can not increase core numbers are removed .
all above ensures the corretness of the algorithm . as for the time complexity , because in each iteration , for each vertex , at least one inserted edge connected to it can be selected into the superior edge set and processed , there are at most @xmath211 iterations in the algorithm execution .
we next consider the time used in each iteration .
now consider an iteration @xmath212 , and we denote the graph obtained after iteration @xmath213 is @xmath214 .
denote by @xmath193 the superior edge set computed in iteration @xmath212 .
the computation of @xmath167 values for vertices in exkpt of @xmath193 takes @xmath215 time .
the positive dfs visits each vertex in exkpt for one time .
hence the positive dfs procedure takes @xmath216 time . for the negative dfs procedures ,
notice that after each dfs procedure , if a vertex @xmath2 is visited , @xmath190 is decreased by 1 .
hence , each vertex can be visited by at most @xmath198 times , since a vertex will be removed if its @xmath189 values is decreased to its core number .
combining together , the total time for an iteration is @xmath217 . by above
, it can be got the time complexity of the algorithm as stated in the theorem .
the decremental algorithm is showed in algorithm [ alg : superioredgedelete ] .
similar with the incremental algorithm , we deal with deleted edges in iterations . in each iteration
, a superior edge set is found using a parallel approach .
after that , the graph is updated by deleting the computed superior edge set and the @xmath0-superior edge sets are assigned to child processes . in each child process , the edges in a @xmath0-superior edge set is handled one by one similarly .
the main difference is that we use @xmath185 values to evaluate if a vertex will decrease its core number , and only execute the negative dfs to remove vertices that can not be in the current @xmath0-core . when deleting an edge @xmath127 with core(@xmath21 )
@xmath218 core(@xmath2 ) , it is checked if @xmath21 still has enough superior neighbors that can help it keep the core number .
if @xmath31 is decreased , algorithm [ alg : deleteremove ] is executed to remove it and disseminate the influence .
[ htb][alg : superioredgedelete ] * input * + the graph , @xmath26 ; + the deleted edge set , @xmath34 ; + the set of vertices @xmath171 connected to edges in @xmath34 the core number @xmath52 of each vertex in @xmath14 [ htb][alg : deletek ] * input * + the graph , @xmath26 ; + the @xmath0-superior edge set , @xmath47 ; + the core number @xmath52 of each vertex in @xmath14 * initially * , @xmath174 empty stack for each vertex @xmath175 , @xmath176\gets{false } , removed[v]\gets{false } , cd[v]\gets{0}$ ] * return * @xmath219 ; [ alg : deleteremove ] @xmath180 @xmath181 \gets{true}$ ] * performance analysis .
* we next analyze the correctness and efficiency of the proposed decremental algorithm . at first , some notations are defined , which will be used in measuring the time complexity of the algorithm . for graph @xmath26 , the deleted edge set @xmath34 and a subset @xmath220 of @xmath34 ,
let @xmath221 and @xmath222 be the set of core numbers of vertices in @xmath223 . for @xmath223 ,
let @xmath224 . for @xmath225 ,
let @xmath226 be the set of vertices with core number @xmath0 and @xmath227 .
denote by @xmath228 the set of edges connected to vertices in @xmath226 .
we then define @xmath229 as follows , @xmath230|\}. m_r=\max_{k\in k(g_r)}\{|e(v_r(k))|\}.\ ] ] @xmath231 , @xmath232 and @xmath229 will depict the time used in each iteration in the algorithm execution .
furthermore , we define the _ maximum deleted degree _ as the maximum number of edges deleted from each vertex in @xmath14 , denoted as @xmath233 . using a similar argument as that for analyzing the incremental algorithm
, we can get the following result , which states the correctness and efficiency of the decremental algorithm . the detailed proof is put in appendix .
[ deletecorrectness ] algorithm [ alg : superioredgedelete ] can update the core numbers of vertices after inserting an edge set @xmath34 in @xmath234 time .
the proof is similarly as the insertion case , the only difference lies in the second part in each iteration .
we only conduct the negative dfs to remove vertices that will decrease core numbers according to corollary [ corollary : sd ] . for each edge @xmath127 , if core(@xmath21 ) @xmath235 core(@xmath2 ) and @xmath236 , we start the negative dfs rooted as @xmath188 to travel the edges in * @xmath187 * where @xmath188 is @xmath21 or @xmath2 that has a smaller core number . if core(@xmath21 ) = core(@xmath2 ) and @xmath237 , we will travel * @xmath134 * to remove @xmath21 and disseminate the influence first , then if @xmath237 and @xmath2 is not removed yet , remve @xmath2 and disseminate the influence through the negative dfs rooted as @xmath2 . during the dfs
a vertex @xmath122 can be visited by at most @xmath238 times , since a visit will decrease @xmath209 $ ] by 1 and when @xmath209 < k$ ] , @xmath122 will be marked as removed and wont be visited again .
after all edges are handled , the vertices that are marked as visited and removed will decrease core numbers .
algorithm [ alg : deletek ] first compute @xmath185s for vertices in exkpt of @xmath47 and the time complexity is @xmath239 . and the time needed for the negative dfs is @xmath240 , since we travel each vertex in @xmath241 at most @xmath242 times to remove it .
so in one iteration the time complexity of algorithm [ alg : superioredgedelete ] is @xmath243 .
then similarly as the insertion cese , the number of iterations is bounded by the max deleted degree @xmath233 .
so the time complexity of the whole algorithm is @xmath244 .
in this section , we conduct empirical studies to evaluate the performances of our proposed algorithms . the experiments use three synthetic datasets and seven real - world graphs , as shown in table [ table_graph ] .
there are two main variations in our experiments , the original graph and the inserted / deleted edge set .
we first evaluate the efficiency of our algorithms on real - world graphs , by changing the size and core number distribution of inserted / deleted edges . then we evaluate the scalability of our algorithms using synthetic graphs , by keeping the inserted / deleted edge set stable and changing the sizes of synthetic graphs . at last ,
we compare our algorithms with the state - of - the - art core maintenance algorithms for single edge insertion / deletion , traversal algorithms given in @xcite , to evaluate the acceleration ratio of our parallel algorithms .
the comparison experiments are conducted on four typical real - world datasets .
all experiments are conducted on a linux machine with intel xeon cpu e5 - [email protected] and 64 gb main memory , implemented in c++ and compiled by g++ compiler .
* datasets .
* we use seven real - world graphs and random graphs generated by three models .
the seven real - world graphs can be downloaded from snap @xcite , including social network graphs ( livejournal , youtube , soc - slashdot ) , collaboration network graphs ( dblp , ca - astroph ) , communication network graphs ( wikitalk ) and web graphs ( web - berkstan ) .
the synthetic graphs are generated by the snap system using the following three models : the erds - r**@xmath245**nyi ( er ) graph model @xcite , which generates a random graph ; the barabasi - albert ( ba ) preferential attachment model @xcite , in which each node creates @xmath0 preferentially attached edges ; and the r - mat ( rm ) graph model @xcite , which can generate large - scale realistic graphs similar to social networks . for all generated graphs , the average degree is fixed to 8 , such that when the number of vertices in the generated graphs is the same , the number of edges is the same as well .
[ core1 ] and fig .
[ core2 ] show the core number distributions of the seven real - world graphs and the generated graphs with @xmath246 vertices . from fig .
[ core1 ] , it can be seen that in real - world graphs , more than 60 percent of vertices have core numbers smaller than 10 .
especially , in wt ( wiki - talk ) , more than 70% of vertices have core number 1 . for the core distributions of generated graphs , as shown in fig .
[ core2 ] , in the ba graphs , all vertices have a core number of 8 . in the er graph ,
the core numbers of vertices are small and the max core number of vertices is 10 , but almost all vertices have core numbers close to the max one .
the rm graph are more close to real - world graphs , where most vertices have small core numbers and as the core number @xmath0 increases , the percentage of vertices with core number @xmath0 decreases .
as shown later , the core distribution of a graph will affect the performances of our algorithms .
the _ core number _ of an edge is defined as the smaller core number of its two endpoints .
we use the _ average processing time per edge _ as the efficiency measurement of the algorithms , such that the efficiency of the algorithms can be compared in different cases .
.real - world graph datasets [ cols="^,^,^,^,^",options="header " , ] we evaluate the impacts of three factors on the algorithm performance : the size of inserted / deleted edges , the core number distribution of edges inserted / deleted , and the original graph size .
the first factor affects the iterations needed to process the inserted / deleted edges , and the last two factors affect the processing time in each iteration .
the first two evaluations are conducted on real - world graphs , and the third one is on synthetic graphs .
we first evaluate the impact of the number of inserted/ deleted edges on the performances of our algorithms .
the results for the incremental and decremental maintenance algorithm are illustrated in fig .
[ size_ins ] and fig .
[ size_del ] respectively . in the experiments ,
we randomly insert / delete @xmath247% edges with respect to the original graph , where @xmath248 for @xmath249 . in fig .
[ size_ins ] and fig .
[ size_del ] , the x - axis represents the datasets , and the y - axis represents the average processing time per edge . it can be seen that the processing time per edge is less than @xmath250 in all cases , and except for wt and lj , the processing time is much smaller than @xmath250 .
the figures show that the processing time decreases as the number of inserted / deleted edges increases , which demonstrates that our algorithms are suitble for handling large amount of edge insertions / deletions . in this case
, more edges can be selected into the superior edge set in each iteration , and hence our algorithms achieve better parallelism .
furthermore , fig .
[ size_ins ] and fig .
[ size_del ] also illustrate that it needs a larger average processing time when the size of original graphs increases .
the only exception is the wt graph .
though the graph has a smaller size than bs and lj graphs , the average processing time is larger .
this is because the core distribution of wt is rather unbalancing , as showed in figure [ core1 ] , where most vertices possess the same core number . in this extremal case ,
on the one hand , each iteration in the algorithm takes more time in processing the inserted edges , as more vertices need to be traversed , and on the other hand , the parallelism of the algorithm is very limited , as most edges are inserted to vertices with the same core .
we then evaluate the impact of the core number distribution of inserted / deleted edges on the algorithm performance .
the results are illustrated in fig .
[ corechange ] .
in particular , by the core distributions showed in fig .
[ core1 ] , we choose five typical core numbers \{@xmath251 } in an increasing order for each of the seven graphs . for each core number , 20% edges of that core number
are selected randomly as the update edge set . from fig .
[ corechange ] , it can be seen that larger core number induces a larger average processing time .
this is because , when inserting / deleting edges to vertices with larger core numbers , the degree of these vertices generated by these inserted edges is larger . in our algorithm , only one superior edge can be handled for each vertex in each iteration .
hence , it takes more iterations to process the inserted / deleted edges .
but on the other hand , it can be also seen that the processing time per edge does not vary significantly .
we finally evaluate scalability of our algorithms in synthetic graphs , by letting the number of vertices scale from @xmath252 to @xmath246 and keeping the average degree fixed as 8 .
the results are shown in fig .
[ graphchange ] . in the experiments
, for each graph , we randomly select 10000 edges as the update set . in fig .
[ graphchange ] , the x - axis represents the number of vertices in the graph , and the y - axis represents the average processing time per edge .
[ graphchange ] shows that though the graph size increases exponentially , the average processing time increases linearly .
it demonstrates that our algorithms can work well in graphs with extremely large size . from the figures
, it can be also seen that the processing time in the ba graph is larger than those of the other two graphs .
this is because all vertices in the ba graph have the same core number 8 .
this means that in our algorithm , all edges are initially handled in one process , and hence the parallelism is poor in this extreme case .
this can be seen as the worst case for our algorithms . however , as shown in fig .
[ core1 ] and fig .
[ core2 ] , real - word graphs exhibit much better balance in core number distribution . in this section ,
we evaluate the acceleration ratio of our parallel algorithms , comparing with algorithms sequentially handling edge insertions / deletions .
we compare with the state - of - the - art sequential algorithm , * traversal * algorithms given in @xcite .
the comparison is conducted on four typical real - world graphs , db , wt , yt and lj in table [ table_graph ] . for each graph
, we randomly select 5k-20k edges as the update set .
the evaluation results are illustrated in fig .
[ cmp_ins ] and fig .
[ cmp_del ] respectively . in the figures ,
the x - axis and y - axis represent the number of inserted / deleted edges and the acceleration ratio , respectively . from fig .
[ cmp_ins ] and fig .
[ cmp_del ] , it shows that in almost all cases , our algorithms achieves an acceleration ratio as large as @xmath253 times in both incremental and decremental core maintenance .
the acceleration ratio increases as the number of edges inserted / deleted increases , which illustrates that our algorithms have better parallelism in scenarios of large amounts of graph changes .
furthermore , it is also shown that our algorithms achieve larger acceleration ratios as the graph size increases .
all evaluation results show that our algorithms exhibit good parallelism in core maintenance of dynamic graphs , comparing with sequential algorithms .
the experiments illustrate that our algorithms are suitable for handling large amounts of edge insertions / deletions in large - scale graphs , which is desirable in realistic implementations .
in this paper , we present the first known parallel algorithms for core maintenance in dynamic algorithms .
our algorithms have significant accelerations comparing with sequential processing algorithms that handle inserted / deleted edges sequentially , and reduce the number of iterations for handling @xmath5 inserted / deleted edges from @xmath5 to the maximum number of edges inserted to / deleted from a vertex .
experiments on real - world and synthetic graphs illustrate that our algorithms implement well in reality , especially in scenarios of large - scale graphs and large amounts of edge insertions / deletions . for the future work
, it deserves more efforts to discovering structures other than superior edge set that can help design parallel core maintenance algorithms .
furthermore , it is also meaningful to design parallel algorithms for maintaining other fundamental vertex parameters , such as betweenness centrality @xcite .
the authors would like to thank ... * proof of theorem [ deletecorrectness ] . * + the deletion algorithm is executed in iterations , and each iteration includes two parts . the first part is similarly as the insertion case , which computes the superior edge set in parallel , and then deletes the computed superior edge set from graph @xmath1 . by lemma [ them : superioredgesetdelete ] , after deleting such a superior edge set @xmath47 from the graph , each vertex can decrease its core by at most 1 . then in the second part , we identifiy vertices that will decrease core numbers by executing algorithm [ alg : deletek ] in parallel . in each child process , it is sufficient to visit vertices in the exkpt of @xmath47 to find all vertices whose core numbers may decrease according to lemma [ le : exkpaths ] . for each edge ,
we start a negative dfs to remove vertices that are confirmed to decrease core numbers . and by lemma [ corollary : sd ] , for a vertex @xmath2 , if @xmath254 \le k$ ] , @xmath2 will decrease its core number , and this will affect the @xmath185 value of its neighbors .
so we use a variable value @xmath189 to represent the dynamic changes of @xmath185 value .
after all edges are handled , vertices in exkpt are visited and the ones that can not be in the current @xmath0-core are marked as removed .
all above ensures the corretness of the algorithm .
now consider an iteration @xmath212 , denote the superior edge set computed in current iteration @xmath212 as @xmath220 .
the computation of @xmath185 values for vertices in exkpt of @xmath220 takes @xmath239 time . for the negative dfs procedures ,
if a vertex @xmath2 is visited , @xmath190 is decreased by 1 .
hence , each vertex can be visited by at most @xmath231 times , since a vertex will be removed if its @xmath189 values is decreased below to its core number .
combining together , the total time for an iteration is @xmath256 . | this paper initiates the studies of parallel algorithms for core maintenance in dynamic graphs .
the core number is a fundamental index reflecting the cohesiveness of a graph , which are widely used in large - scale graph analytics .
the core maintenance problem requires to update the core numbers of vertices after a set of edges and vertices are inserted into or deleted from the graph .
we investigate the parallelism in the core update process when multiple edges and vertices are inserted or deleted .
specifically , we discover a structure called _ superior edge set _ , the insertion or deletion of edges in which can be processed in parallel .
based on the structure of superior edge set , efficient parallel algorithms are then devised for incremental and decremental core maintenance respectively . to the best of our knowledge ,
the proposed algorithms are the first parallel ones for the fundamental core maintenance problem .
the algorithms show a significant speedup in the processing time compared with previous results that sequentially handle edge and vertex insertions / deletions .
finally , extensive experiments are conducted on different types of real - world and synthetic datasets , and the results illustrate the efficiency , stability and scalability of the proposed algorithms . |
quantum chromodynamics ( qcd ) is the underlying theory of strong interactions , with quarks and gluons as fundamental degrees of freedom .
however , in the non - perturbative region of low and medium energy physics , mesons and baryons definitely keep their importance as efficient , collective degrees of freedom for a wide range of nuclear phenomena .
there is a widespread view that a successful effective field theory in terms of hadronic degrees of freedom should have the same symmetries as qcd .
chiral symmetry , in the limit of vanishing quark masses , is an exact qcd symmetry ; chiral perturbation theory@xcite is an effective field theory in terms of the light pseudoscalar mesons ( @xmath3,@xmath4,@xmath5 ) , the goldstone bosons of spontaneously broken chiral symmetry .
it is important to realize that such an effective field theory is mathematically eqivalent to qcd . at sufficiently low energies , smaller than the qcd mass scale of about 1gev
, it leads to a systematic expansion of effective lagrangians and scattering amplitudes in powers of momenta and quark masses ( the latter arise from the symmetry breaking terms ) .
of course , the effectiveness ( as a fastly converging scheme ) of chiral perturbation theory is by definition restricted to quite low energies .
moreover , a perturbational method can only work if the interactions are weak , which is the case for reactions involving the goldstone bosons as external particles .
given these constraints , chiral perturbation theory is remarkably successful ; for a review , see e.g. the recent article by meissner@xcite .
on the other hand , for baryon - baryon and baryon - antibaryon interactions strong correlation effects appear already at threshold ; it is hard to get such effects from an expansion which , for practical reasons , has to be limited to low orders .
furthermore , higher - mass mesons , especially vector mesons @xmath1 and @xmath2 , play an important role . in my opinion
, a successful description requires to treat both dynamical mechanisms explicitly , as will be discussed below .
let me start by looking at present chiral treatments of the nucleon - nucleon ( @xmath0 ) system .
recently , quite a lot of papers have appeared dealing with this issue . whereas refs . concentrated on the chiral structure of a specific contribution to the @xmath0 interaction , namely @xmath6-exchange , ref .
@xcite presented a `` complete '' @xmath0 potential strictly obeying chiral symmetry .
it is based on the most general effective chiral lagrangian involving low momentum pions , nonrelativistic nucleons and @xmath7-isobars , and is considered up to third order ( @xmath8 , with @xmath9 a typical external momentum and @xmath10 the qcd mass scale ) in the chiral expansion , thus containing tree as well as one - loop diagrams .
08.5 cm in diagrammatic terms , the resulting @xmath0 potential has the structure shown in fig .
it consists of one - pion - exchange ( ope ) and @xmath6-exchange terms with @xmath11 and @xmath7 intermediate states .
there are additional @xmath6-exchange processes in the chiral approach ( ref .
@xcite ) involving the @xmath12 vertex ( see fig .
[ figi ] ) , whose presence is dictated by chiral symmetry .
however , due to partial cancellations ( refs .
@xcite ) they are actually quite small , much smaller e.g. than `` @xmath13''- exchange representing correlated @xmath6-exchange in the scalar - isoscalar channel , which provides the main part ( about @xmath14 , see ref .
@xcite ) of the @xmath0 intermediate range attraction .
the chiral approach absorbs this contribution together with other important pieces involving vector ( @xmath15 ) mesons into the contact terms , i.e. terms of zero range ( last diagram of fig .
[ figi ] ) . in order to solve the corresponding schroedinger equation
a gaussian cutoff function , @xmath16 , with @xmath17 , has been used .
results for some low ( @xmath18 and @xmath19 ) @xmath0 partial wave phase shifts are shown in fig .
[ figii ] ( taken from ref .
@xcite ) , in comparison to those obtained from the ( full ) bonn potential@xcite . in addition
, table 1 shows some deuteron parameters .
l|cccc & experiment & chiral model & bonn & modified bonn & & ( ref .
@xcite ) & ( ref .
@xcite ) & binding energy ( mev ) & 2.2245754 & 2.18 & 2.2247 & 2.2246 quadrupole moment ( @xmath20 ) & 0.2859@xmath21 & 0.231 & 0.2807 & 0.2791 asymptotic d / s ratio & 0.0256@xmath22 & 0.0239 & 0.0267 & 0.0266 & ( 0.0271@xmath23 ) & & & 16.75 cm in this first confrontation of a chiral @xmath0 potential la weinberg to the very precise @xmath0 data rough agreement is reached in the low - energy domain . note that this model has 26 parameters available ( mainly the coefficients in the contact terms ) compared to about 10 meson - baryon coupling constants and formfactor parameters in the bonn potential
. a combined look at the @xmath24-parameter and deuteron parameters ( @xmath9 , @xmath25 ) is instructive since both measure the size of the @xmath0 tensor force . obviously , in the present chiral model
, the energy dependence of this piece is not realistic since the deuteron values are too low whereas the @xmath24-values grow too fast with increasing energy .
there are two possible sources for this discrepancy : first , the chiral expansion introduces a polynomial momentum dependence , with adjustable parameters , at the @xmath26 vertex ; it should be checked by the authors of ref .
@xcite whether their parametrization chosen is in reasonable agreement with the known monopole structure of the @xmath26 formfactor , see section 4 .
furthermore , @xmath1-exchange provides an important contribution to the tensor force ( see section 3 ) reaching out well into intermediate @xmath0 distances of @xmath27 .
it is highly questionable whether this piece ( as well as `` @xmath13 '' and @xmath2 exchange ) can be sufficiently represented by zero - range contact terms . in my opinion , in present chiral approaches to the @xmath0 system , the price you pay ( having to treat important contributions in a very rough way ) is too high for what you get ( a strict chirally symmetric @xmath28 derived from a systematic expansion ) .
it is well known that the physics of `` @xmath13 '' and @xmath1 exchange , so important in the @xmath0 system , is mainly built up by correlated @xmath6-exchange .
( for a recent dynamical model suitable for the discussion of medium modifications , see the paper by kim et al.@xcite . ) therefore , in principle , the chiral approach can take such effects , at least to some extent , into account by going to still higher orders in the chiral expansion , at the expense of getting additional open parameters .
the importance of @xmath1-exchange for the dynamics of the @xmath0 system derives from the following fact : it provides a sizable tensor force , which has opposite sign to the tensor force generated by one - pion - exchange , see fig .
thus there is a strong cancellation , over a relatively broad range of energies and distances , between @xmath3 and @xmath1 exchange in the tensor channel .
a similar cancelation occurs in the strange sector , between @xmath4 and @xmath29 exchange , e.g. in the hyperon - nucleon interaction .
therefore , in the @xmath0 ( and in the baryon - baryon system in general ) it is strongly suggested to always group @xmath3 and @xmath1 ( as well as @xmath4 and @xmath29 ) together in order to reach sufficient convergence in the expansion of the irreducible kernel ( potential ) . to single out @xmath3-exchange as done in the chiral approach might be more systematic from a formal point of view ; on the other hand , by putting @xmath1-exchange into the contact terms , one essentially loses the strong convergence - generating mechanism at intermediate distances provided by @xmath1-exchange .
thus , from a physical standpoint , it appears mandatory to treat @xmath3 and @xmath1 exchange on an equal footing .
4.5 cm in fact , this procedure has been an essential guideline when constructing the bonn potential .
unfortunately ( though for understandable reasons , see below ) it was not followed to a sufficient degree : whereas , in second - order diagrams ( cp.fig .
[ figviii ] ) @xmath30 as well as @xmath31 exchange have been included for uncorrelated processes ( with @xmath11 and @xmath7 intermediate states ) this has not been done for correlated processes : correlated @xmath6-exchange processes have been included ( in terms of sharp - mass @xmath32 and @xmath1 exchange ) but correlated @xmath31 processes have been left out .
the reason is quite simple : the evaluation of this missing piece is technically quite complicated , much more involved ( due to the spin of the @xmath1 ) compared to correlated @xmath6-exchange .
more importantly , a dynamical model for the interaction between a @xmath3 and a @xmath1 meson was not available . at the time of the construction of the bonn potential ( about 10 years ago )
the omission of this piece did not seem to be serious .
@xmath0 scattering data in a broad energy range as well as the deuteron data are described quantitatively with the full bonn model .
so why care about a missing piece whose practical relevance was not at all obvious ?
soon after the publication of the bonn potential in 1987 it became however clear that omission of correlated @xmath31 exchange demanded a high price , and this has to do with the structure of the @xmath26 vertex .
basic ingredients of the bonn meson exchange @xmath0 model@xcite are the meson - baryon couplings or vertex functions visualized in fig .
[ figix ] .
these vertex functions are the `` elementary '' building blocks of an effective and consistent meson exchange description based on qcd .
4.50 cm the analytic structure of these couplings is ( essentially ) determined by the quantum numbers of the particles involved at the vertex .
the strength is parametrized by coupling constants @xmath33 ; in addition , formfactors with cutoff masses @xmath34 as additional parameters are included , which parametrize the effect of the hadron extension and therefore represent a truly physical concept . in general
, the formfactor depends on all four - momenta of the particles involved at the vertex , i.e. @xmath35 and is normalized to 1 when all particles are on the mass shell .
while ultimately these formfactors have to be derived from qcd they are for the moment parametrized , usually in monopole form , keeping the dependence on the momentum of the exchanged particle only . with the additional suppression of the dependence on @xmath36 , the formfactors relevant in the @xmath0 system
can then be written as @xmath37 the cutoff masses @xmath34 together with the coupling constants @xmath33 represent the only parameters in the bonn potential ; they have been adjusted to the @xmath0 data . for the @xmath26 cutoff mass , @xmath38 , the resulting value turned out to be rather large ( @xmath39 ) leading to a mild suppression of opep in the inner region ( @xmath40 ) only , and thus to a hard formfactor .
this is necessary in order to have sufficiently strong tensor force to reproduce the deuteron properties , especially the asymptotic @xmath41- to @xmath18-wave ratio and the quadrupole moment@xcite . however , there is a long - standing discrepancy between this rather large value required in present - day potential models and information from other sources ; the latter consistently point to a much smaller value for @xmath38 around @xmath42@xcite .
in fact , a recent lattice calculation@xcite of the @xmath43 formfactor confirms this result , finding a monopole mass of @xmath44 .
such a soft formfactor leads to a strong suppression of the ope tensor force already at intermediate distances , in apparent disagreement with the deuteron data .
it is sometimes argued that a reduction ( or even elimination ) of @xmath1 exchange in @xmath0 models would remedy the situation easily since the tensor force of @xmath1 exchange has opposite sign compared to @xmath3-exchange .
this can certainly be done if one looks at the deuteron channel only .
however the sizable strength of @xmath1 exchange is tightly constrained by informations from @xmath45 scattering via dispersion theory ( see ref .
@xcite ) and in fact required by the precisely measured triplet @xmath19-wave @xmath0 phase shifts ; it therefore can not be arbitrarily changed .
4.5 cm consequently we have to find mechanisms so far not included in meson exchange @xmath0 models which provide additional ( short ranged ) tensor force .
two years ago it was pointed out by ueda@xcite that the contribution arising from the exchange of a correlated three - pion state between the nucleons , with the quantum numbers of a pion , should enhance the short - range tensor part of the @xmath0 interaction .
correlated @xmath31 exchange addressed in the foregoing chapter is a good part of it , and there is obviously no way around this contribution .
the explicit evaluation of the diagram in figure [ figx ] requires as input a realistic @xmath31 @xmath46-matrix . in the absence of elastic @xmath31 scattering data ( due to the fast decay of the @xmath1 meson into 2 pions ) we have to rely completely on a dynamical model .
3.75 cm we@xcite have recently constructed a corresponding potential model , in complete analogy to the @xmath30 case ( ref .
@xcite ) , with driving terms shown in fig .
[ figxi ] .
open parameters are adjusted mainly to empirical information in the @xmath47-channel ( the @xmath47 decaying into @xmath3 and @xmath1 ) , namely the resonance parameters ( pole position ) of the @xmath47-particle@xcite ( @xmath48 , @xmath49 ) obtained from @xmath50-meson decay , and the mass spectrum obtained from charge exchange production@xcite , @xmath51 , see fig .
[ figxii ] .
note that the shift of the maximum of the mass spectrum away from the true pole position is due to the sizable non - pole amplitude @xmath52 , obtained by iterating @xmath53 ( the first two diagrams of fig .
[ figxi ] . ) .
this non - pole contribution now acts in all other channels , e.g. in the pionic channel of interest here , with a definite strength distribution , which is characteristic of the underlying model .
this @xmath31 @xmath46-matrix is now inserted into the correlated @xmath31 exchange diagram of fig .
the evaluation@xcite proceeds via the same dispersion - theoretic treatment as used for the @xmath30 case in ref .
the result can be represented as integral over various spectral functions , the number of which depend on the considered channel . for the pionic channel we have one spectral function @xmath54 : @xmath55 @xmath54
is shown in fig .
[ figxiii ] .
obviously the correlated part provides a sizable contribution , with a peak around @xmath56 , somewhat smaller than the mass ( @xmath57 ) of the phenomenological @xmath58 introduced in some recent @xmath0 models to accommodate a soft @xmath26 formfactor@xcite .
indeed , as shown in fig .
[ figxiv ] and table 1 , the resulting interaction due to correlated @xmath31 exchange in the pionic channel is able to counterbalance the substantial suppression induced in opep when going from a cutoff mass @xmath38 of @xmath59 , phenomenologically required in the ( full ) bonn potential , to a value of @xmath60 .
further tensor contributions , with the same sign as pion exchange , arise from correlated @xmath31 exchange in the @xmath47-channel .
if these will be included , too , @xmath38 can be decreased further , into the region of @xmath42 .
thus the inclusion of correlated @xmath31 exchange in a meson exchange @xmath0 model should allow to use the correct , soft @xmath26 formfactor , and thus resolve a long - standing puzzle .
in this talk , i have tried to convince you that , in low energy baryon - baryon interactions , chiral symmetry does not appear to be the dominant symmetry : a meson exchange model like the bonn potential , which violates ( to some extent ) chiral symmetry , is able to describe the data quantitatively . on the other hand , a model strictly obeying chiral symmetry but restricting itself to pions as mesonic degrees of freedom and one - loop contributions shows noticeable deficiencies in the description of the empirical situation , already at very low energies . a minimal requirement for
improvement would be the inclusion of two - loop contributions , i.e. to go to still higher order in the chiral expansion , in order to take at least part of the @xmath6correlation effects ( `` @xmath13 '' and @xmath1 exchange ) into account anyhow , there are a lot of baryonic processes whose physics is dominated by quite high orders .
for example , we have recently shown@xcite that the @xmath61 reaction goes predominantly through the @xmath62 intermediate state , with strong initial ( @xmath63 ) , intermediate ( @xmath62 ) , and final ( @xmath64 ) state interactions .
each interaction itself has sizable higher order contributions , e.g. annihilation processes into 2 ( or more ) mesons .
a treatment of these coupled channels effects in the chiral approach , including systematically all diagrams at the required order , is surely beyond our capabilities . in this area of medium energy physics to be studied with our cosy accelerator , it appears more efficient not to single out the goldstone bosons as relevant degrees of freedom but to include from the beginning higher mass mesons as mediators of baryonic interactions , in this way summing up an important class of higher order correlation effects , as done already for the @xmath7-isobar in ref .
further selected summations of processes to arbitrarily high order , in a potential - type coupled channel framework , have to be performed .
necessarily the full and exact implications of chiral symmetry are then lost .
clearly , as always in physics , the actual treatment of a given theory ( kind of approximations ) depends on the phenomena under study and on the questions addressed .
if , for example , the main issue is to study modifications of hadronic interactions in the medium , chiral symmetry gains much more importance due to chiral restoration in matter .
thus , in order to study such effects seriously , a strictly chiral model is probably required .
still , one should always be aware about what has been ( necessarily ? ) sacrificed and about the consequences .
99 # 1(#2)#3*_phys .
_ * c # 1 * ( # 2 ) # 3 # 1(#2)#3*_phys .
* d # 1 * ( # 2 ) # 3 # 1(#2)#3*_phys .
lett . _ * # 1 * ( # 2 ) # 3 # 1(#2)#3*_phys .
* b # 1 * ( # 2 ) # 3 # 1(#2)#3*_nucl .
phys . _ * a # 1 * ( # 2 ) # 3 # 1(#2)#3*_nucl .
* b # 1 * ( # 2 ) # 3 | the present status of the chiral approach to the @xmath0 interaction as proposed by weinberg is discussed . the important role of correlation effects and explicit vector meson ( @xmath1 , @xmath2 ) exchange in the dynamics of baryon - baryon interactions is demonstrated . as an example , the inclusion of the exchange of a correlated pair of @xmath3 and @xmath1 meson between two nucleons appears to be mandatory in order to resolve a long - standing puzzle concerning the formfactor at the pion - nucleon - nucleon vertex . |
when calculating the ground state energy of a quantum field ( the casimir energy ) the main problem is to single out a finite part of the vacuum energy which is initially divergent .
usually for this purpose a subtraction procedure is used with preliminary regularization of the divergent expressions ( for example , by introducing ultraviolet cutoff ) .
however in quantum field theory treated with allowance for nontrivial boundary conditions or in the space - time with curvature , a complete renormalization procedure is not formulated explicitly . therefore , for any specific problem
the subtraction procedure should be invented anew . as a result ,
one succeeds in calculating the casimir energy only in the problems with known spectra or at least implicitly known spectra .
practically it implies the boundary conditions of high symmetry @xcite ( parallel plates , sphere , cylinder ) . in studies of the casimir energy
it is wildly used the zeta function technique @xcite which is also referred to as the zeta regularization or zeta renormalization .
in fact , the use of the zeta functions , as well as other regularizations , gives only regularized quantities for ground state energy , for effective potential and so on .
the necessity to renormalize the expressions obtained in this way certainly remains .
however , in some problems the zeta technique gives at once a finite result .
usually the latter is considered to be a renormalized physical answer though generally it is not the case.@xcite when using the zeta regularization in one or another problem , it is desirable to know beforehand whether the finite result can be obtained in this way . in order to answer this question the general analysis of the divergences in the problem at hand should be accomplished .
this can be done by calculating the heat kernel coefficients@xcite depending on the geometry of the manifold under consideration . for a large class of situations
these coefficients have been obtained.@xcite however there is a number of problems ( for example , boundaries with edges or corners ) for which no general results regarding the heat trace are known .
in this situation it is undoubtedly worth carrying out , in the framework of the zeta function technique , the calculations of the casimir energy for new configurations , both the cases being interesting with finite result and with pole contributions left in the final expression for the vacuum energy . in the present paper
we address the calculation of the casimir energy for boundaries with edges , more precisely , the vacuum energy of electromagnetic field will be calculated for a semi - circular cylindrical shell by making use of the relevant zeta functions .
this shell is obtained by crossing an infinite circular cylindrical shell by a plane passing through the symmetry axes of the cylinder .
all the surfaces , including the infinite cutting plane , are assumed to be perfectly conducting .
obviously it is sufficient to consider only a half of this configuration ( left or right ) which we shall refer to as a semi - circular cylindrical shell or , for sake of shortening , as a semi - circular cylinder .
the internal boundary value problem for this configuration is nothing else as a semi - cylindrical waveguide . in the theory of
waveguides@xcite it is well known that a semi - circular waveguide has the same eigenfrequencies as the cylindrical one but without degeneracy ( without doubling ) and safe for one frequency series ( see below ) . notwithstanding the very close spectra , the zeta function technique does not give a finite result for a semi - circular cylinder unlike for a circular one .
first the casimir energy of an infinite perfectly conducting cylindrical shell has been calculated in ref . by introducing ultraviolet cutoff and recently this result
was derived by zeta function technique@xcite ( see also refs . ) . as far as we know the asymmetric boundaries such as a semi - circular cylinder have not been considered in the casimir problem . the paper is organized as follows . in sec .
ii the electromagnetic spectra are considered in details for cylindrical and semi - cylindrical shells .
the general solution of the maxwell equations for boundary conditions chosen is expressed in terms of two scalar functions , longitudinal components of the electric and magnetic hertz vectors .
these scalar functions are the eigenfunctions of the two - dimensional transverse laplace operator and obey the dirichlet and neumann boundary conditions on the conducting surfaces . in sec .
iii the spectral zeta function is constructed for the dirichlet boundary value problem . to this end , the technique is used which has been elaborated before for representing the spectral zeta function , with given eigenfrequency equations , in terms of contour integral . when carrying out the analytic continuation of the zeta function into the physical region , the uniform asymptotic expansion for the modified bessel functions is used . in the same way , in sec .
iv the zeta function is constructed for a scalar field obeying the neumann boundary conditions given on the surface of a semi - circular cylindrical shell .
the section v is concerned with the complete zeta function for electromagnetic field with boundary conditions on the semi - circular cylinder .
transition to the relevant two - dimensional problem is also considered here . in the conclusion ( sec .
vi ) the results obtained are summarized , and the origin of the pole singularities of the zeta functions at hand and their relation to the respective boundary value problem are briefly discussed .
the construction of the solutions to the maxwell equations with boundary conditions given on closed surfaces proves to be nontrivial problem .
mainly it is due to the vector character of the electromagnetic field @xcite . in the case of cylindrical symmetry
the electric @xmath0 and magnetic @xmath1 fields are expressed in terms of the electric ( @xmath2 ) and magnetic ( @xmath3 ) hertz vectors having only one non - zero component @xmath4 here the cylindrical coordinate system @xmath5 is used with @xmath6 axes directed along the cylinder axes . the common time
dependent factor @xmath7 is dropped .
the scalar functions @xmath8 and @xmath9 are the eigenfunctions of the two dimensional transverse laplace operator and meet , respectively , the dirichlet and neumann conditions on the boundary @xmath10 @xmath11 @xmath12 where @xmath13 is the transverse part of the laplace operator @xmath14 and @xmath15 first we consider a cylindrical shell . in this case the functions @xmath8 and @xmath9 should be @xmath16-periodic in angular variable @xmath17 . as a result the dirichlet boundary value problem ( [ d ] ) has the following unnormalized eigenfunctions ( @xmath18-modes ) @xmath19 where @xmath20 is the cylinder radius , @xmath21 are the bessel functions , @xmath22 are the hankel functions of the first kind , and @xmath23 , @xmath24 stand for the roots of the frequency equations @xmath25 for the neumann boundary value problem ( [ n ] ) we have the @xmath26-modes @xmath27 where @xmath28 and @xmath29 are the roots of the equations @xmath30 as usual , it is assumed that for @xmath31 the eigenfunctions should satisfy the radiation condition .
it is important to note that each root @xmath32 is doubly degenerate since , according to eqs .
( [ efd ] ) , ( [ efn ] ) , there are two eigenfunctions which are proportional to either @xmath33 or @xmath34 . the frequencies with @xmath35 @xmath36 are independent on @xmath17 , and
the degeneracy disappears . for given hertz vectors
@xmath37 and @xmath38 the electric and magnetic fields are constructed by the formulas @xmath39 it has been proved @xcite that the superposition of these modes gives the general solution to the maxwell equations in the problem under consideration .
an essential merit of using the hertz polarization vectors is that in this approach the necessity to satisfy the gauge conditions does not arise .
let us consider a waveguide which is obtained by cutting the infinite cylindrical shell by a plane passing through the symmetry axes of the cylinder ( see fig . 1 ) .
all the surfaces are assumed to be perfectly conducting . in this case
the boundary value problems ( [ d ] ) and ( [ n ] ) for the hertz electric ( @xmath37 ) and magnetic ( @xmath38 ) vectors have the following eigenfunctions @xmath40 and @xmath41 the frequencies @xmath42 , @xmath43 , @xmath44 , and @xmath29 are determined by the same equations ( [ fed ] ) and ( [ fen ] ) .
however the new spectral problem has two essential distinctions : i ) the frequencies ( [ dr ] ) are now nondegenerate , and ii ) two series of eigenfrequencies @xmath45 are absent . at first sight
one could expect that such a change of the spectrum can not influence drastically on the ultraviolet behaviour of the relevant spectral density .
however , as it will be shown below , the zeta function for a semi - circular cylinder , unlike for a circular one , does not provide a finite answer for the casimir energy in the problem in question . in view of all
above - mentioned the zeta function for electromagnetic field obeying the boundary conditions on the surface of the semi - circular cylinder is the sum of two zeta functions for scalar massless fields satisfying the dirichlet and neumann conditions on the lateral of this cylinder .
first we consider the dirichlet boundary conditions
. we shall proceed from the following representation for the zeta function in terms of a contour integral for given frequency equations ( [ fed ] ) with @xmath46 @xmath47 the contour @xmath48 consists of the imaginary axis @xmath49 and a semi - circle of an infinite radius in the right half - plane of a complex variable @xmath50 .
the details of obtaining this integral representation can be found in refs . .
contribution into eq.([contour ] ) of integration along a semi - circle of infinite radius vanishes .
therefore upon integration over @xmath51 this formula acquires the form @xmath52\ ] ] with @xmath53 in order to accomplish the analytic continuation of ( [ 3.2 ] ) into the physical region including the point @xmath54 , we shall use the uniform asymptotic expansion for the modified bessel functions@xcite @xmath55&= & \ln(y t)+\frac{t^2}{8 n^2}(1 - 6t^2 + 5t^4)\nonumber\\ & & + \frac{t^4}{64 n^4}(13 - 284 t^2 + 1062 t^4 - 1356 t^6 + 565 t^8 ) + o(n^{-6})\,{,}\end{aligned}\ ] ] where @xmath56 . following the usual procedure applied in the analogous calculations,@xcite
we add and subtract in the integrand in eq .
( [ 3.2 ] ) the first two terms of the asymptotic expansion ( [ 3.3 ] ) .
after that we combine all the terms there in the following way @xmath57 , \label{3.5}\\ z_1(s)&=&\frac{1}{2}\sum\limits_{n=1}^{\infty}n^{1-s}\int_{0}^{\infty } dy\ , y^{1-s}\frac{d}{dy}\ln \left(\frac{y^2}{1+y^2}\right ) , \label{3.6}\\ z_2(s)&=&\frac{1}{8}\sum\limits_{n=1}^{\infty}n^{-1-s } \int_{0}^{\infty } dy\ , y^{1-s}\frac{d}{dy}\left[t^2(1 - 6\ , t^2 + 5\ , t^4)\right ] , \label{3.7}\\ z_3(s)&=&\sum_{n=1}^{\infty}n^{1-s } \int_0^{\infty } dy\ , y^{1-s}\frac{d}{dy } \biggl [ \ln(2yn \,i_n(yn ) k_n(ny ) ) \nonumber \\ & & -\left .\ln\frac{y}{\sqrt{1+y^2}}-\frac{t^2(1 - 6 t^2 + 5 t^4)}{8 n^2}\right]{. } \label{3.8}\end{aligned}\ ] ] analytic continuation of the function @xmath58 into vicinity of the point @xmath54 can be accomplished in the same way as it has been done in ref . .
therefore we write here only the final result of this continuation @xmath59 the integral in eq .
( [ 3.6 ] ) converges when @xmath60 , and the sum over @xmath61 is finite for re@xmath62 .
thus , the regions , where the integral and the sum exist , overlap , and this formula can be used for constructing the analytic continuation needed . for this aim
we substitute the sum by the riemann zeta function @xmath63 and define the integral as an analytic function by making use of the formula@xcite @xmath64 in view of the poles of the gamma functions on the right - hand side of this relation , the integral on the left - hand side of it is well defined , as a function of the complex variable @xmath65 , only in the region indicated in eq .
( [ 3.11 ] ) . doing the analytic continuation of this integral we define it outside this region also by this equation , keeping in mind that the gamma functions involved should be treated as the analytic functions over all the plane of the complex variable @xmath65 safe for the known poles .
this gives @xmath66.\ ] ] in order to investigate the convergence of the integral entering in eq .
( [ 3.7 ] ) it makes sense to substitute in the integrand the logarithmic function by expansion ( [ 3.3 ] ) .
after that it is easy to be convinced that the integral under consideration converges when @xmath67 .
the sum over @xmath61 in this formula is finite for @xmath68 .
hence , the function @xmath69 is an analytic function without singularities in the domain @xmath70 .
it is quiet enough for our purpose , and the analytic continuation is unnecessary . summarizing we conclude that eqs .
( [ 3.3 ] ) , ( [ 3.5 ] ) , ( [ 3.8 ] ) , ( [ 3.9 ] ) , and ( [ 3.12 ] ) afford the analytic continuation needed and define the zeta function @xmath71 as an analytic function in the region including the point @xmath54 .
now we are able to calculate the value of the zeta function @xmath72 at the point @xmath54 . for the coefficient @xmath73 in eq .
( [ 3.3 ] ) we have @xmath74 from eq .
( [ 3.9 ] ) it follows that @xmath75.\ ] ] with allowance for the relations @xmath76 where @xmath77 is the euler constant , @xmath78 , one derives @xmath79\left[\frac{2}{s+1}-\gamma+o(s+1)\right]\nonumber \\ & = & \zeta'(-2)=-0.030448\,{. } \label{3.16}\end{aligned}\ ] ] using the values of the riemann zeta function and its derivative at the origin @xmath80 and taking into account the behaviour of the gamma function near zero ( see eq .
( [ 3.15 ] ) ) we deduce from eq . (
[ 3.12 ] ) @xmath81 \nonumber \\ & & \times \left[\frac{2}{s+1}-\gamma+{\cal o}\left ( s+1 \right ) \right]\cdot
\left[-1+\frac{7}{4}(s+1)\right ] \nonumber\\ & = & -\frac{7}{32}-\frac{\gamma}{16}+\frac{1}{8}\ln(2\pi)+\left .
\frac{1}{8}\frac{1}{s+1}\right|_{s\to -1}{. } \label{3.17}\end{aligned}\ ] ] when calculating @xmath82 we shall use eq .
( [ 3.8 ] ) for sevral first values of @xmath83 and for @xmath84 we substitute the asymptotic expansion ( [ 3.4 ] ) into ( [ 3.8 ] ) with the result @xmath85 \nonumber \\ & = & \frac{1}{64}\left(\sum\limits_{n = n_0 + 1}^{\infty}n^{-3-s}\right ) \gamma\left(\frac{3-s}{2}\right ) \left[-13\ , \gamma\left(\frac{3+s}{2}\right)+ 142\ , \gamma\left(\frac{5+s}{2}\right)-\frac{532}{3}\ , \gamma\left(\frac{7+s}{2}\right)\right.\nonumber\\ & & \left.+\frac{113}{2}\ , \gamma\left(\frac{9+s}{2}\right)-\frac{113}{24}\ , \gamma\left(\frac{11+s}{2}\right)\right ] .
\label{z3as}\end{aligned}\ ] ] the value of @xmath86 should be chosen so as to provide the accuracy needed .
this algorithm with @xmath87 gives for @xmath82 @xmath88 summing up eqs .
( [ 3.16 ] ) , ( [ 3.17 ] ) , and ( [ 3.18 ] ) we obtain @xmath89 thus the zeta function @xmath72 has a pole at the point @xmath54 , therefore it does not give the finite ( renormalized ) value for the respective casimir energy @xmath90 it implies that further renormalization is required .
when constructing the zeta function for the boundary value problem ( [ n ] ) with @xmath91 being a semi - circular infinite cylinder , we shall again proceed from the frequency equations ( now from eq .
( [ fen ] ) .
it should be taken into account that all these roots are not degenerate .
therefore we can write analogously to eq .
( [ contour ] ) @xmath92 the contour @xmath48 is the same as in eq .
( [ contour ] ) and the prime on the bessel and hankel functions denotes differentiation with respect to the entire argument .
the product of the derivatives of the modified bessel functions @xmath93 has the following asymptotics when @xmath61 is fixed and @xmath94 is large @xcite @xmath95\,.\ ] ] taking this into account in calculation of the denominator in eq.([4.1 ] ) , we obtain for @xmath96 upon integration over @xmath51 @xmath97\ ] ] with the same function @xmath73 as in eq .
( [ 3.3 ] ) .
further we shall use the uniform asymptotic expansion for the derivatives of the bessel functions@xcite @xmath98&= & -\ln(y t)+\frac{t^2}{8 n^2 } ( -3 + 10t^2 - 7t^4)+\frac{t^4}{n^4}\left(-\frac{27}{64 } \right.\nonumber \\ & & \left
. + \frac{109}{16 } t^2-\frac{733}{32 } t^4+\frac{441}{16 } t^6-\frac{707}{64 } t^8\right ) + { \cal o}(n^{-6})\,{. } \label{4.4}\end{aligned}\ ] ] in order to render the integral in the term with @xmath35 in eq .
( [ 4.3 ] ) convergent we add and subtract the second term from the asymptotics ( [ 4.4 ] ) . for @xmath99 in eq.([4.3 ] ) we add and subtract in respective intagrands the first two terms of the asymptotic expansion ( [ 4.4 ] ) .
after that we combine all the terms in the following way @xmath100 , \label{4.5 } \\
v_0(s)&=&\int_0^{\infty } dy\ , y^{1-s}\frac{d}{dy } \left\{\ln[-2y i_0'(y ) k_0'(y)]-\frac{t^2}{8}(-3 + 10 t^2 - 7t^4)\right\},\label{4.6 } \\
v_1(s)&=&-\frac{1}{2}\sum\limits_{n=1}^{\infty}n^{1-s}\int_{0}^{\infty } dy\ , y^{1-s}\frac{d}{dy}\ln \left(\frac{y^2}{1+y^2}\right ) = -z_1(s)\ , { , } \label{4.7 } \\ v_2(s)&=&\frac{1}{8}\left(\sum\limits_{n=1}^{\infty}n^{-1-s}+ 1 \right)\int_{0}^{\infty } dy\ , y^{1-s}\frac{d}{dy}\left[t^2(-3 + 10\ , t^2 - 7\ , t^4)\right ] , \label{4.8 } \\
v_3(s)&=&\sum\limits_{n=1}^{\infty}n^{1-s } \int_{0}^{\infty } dy\ , y^{1-s}\frac{d}{dy}\biggl \{\ln \left [ -2yni_n'(ny)k_n'(ny ) \right ] \nonumber \\ & & \left .
+ \ln ( yt ) -\frac{t^2}{8n^2}(-3 + 10t^2 - 7t^4 ) \right \}{. } \label{4.9}\end{aligned}\ ] ] taking into account the behaviour of the product @xmath101 at the origin and at infinity @xmath102 @xmath103 it is easy to show that eq .
( [ 4.6 ] ) defines @xmath104 as an analytic function in the region @xmath105 . under this condition
the integration by parts can be done here @xmath106-\frac{t^2}{8}(-3 + 10 t^2 - 7t^4)\right\}{.}\ ] ] the function @xmath107 differs only in sign of the function @xmath58 from the proceeding section .
the integral in eq.([4.7 ] ) is convergent when @xmath108 .
the sum over @xmath61 in this formula is finite when re @xmath109 .
thus the regions , where the integral and the sum exist , overlap and this formula can be used for constructing the analytic continuation needed by making use of the substitutions ( [ 3.10 ] ) and ( [ 3.11 ] ) . substituting the sum in eq .
( [ 4.8 ] ) by the riemann zeta function and doing the integration according to eq .
( [ 3.11 ] ) one obtains @xmath110\gamma\left(\displaystyle{\frac{3-s}{2 } } \right )
\gamma\left(\displaystyle{\frac{1+s}{2}}\right)\left[3 - 5(1+s ) + \frac{7}{8}(1+s)(3+s ) \right]\,.\ ] ] the convergence of the integral in eq .
( [ 4.9 ] ) can be determined in the same line as it has been done for the function @xmath69 in the preceding section .
this integral converges when @xmath111 , and the sum encountered here is finite for @xmath112 .
hence there is no need to do analytic continuation for @xmath113 .
finally the zeta function @xmath96 for the massless scalar field obeying the neumann boundary conditions on a semi - circular cylinder is determined explicitly by eqs .
( [ 4.5 ] ) , ( [ 4.7 ] ) , ( [ 4.11 ] ) , and ( [ 4.12 ] ) in a finite domain of the complex plane @xmath65 containing the closed interval of the real axis @xmath114 .
now we turn to the calculation of the value of the function @xmath96 at the point @xmath54 .
integration in eq .
( [ 4.11 ] ) gives @xmath115+\frac{3}{8}t^2\right\}+\frac{13}{16 } \nonumber \\ & = & 2\cdot 0.475215 + 0.8123=1.76393\,{. } \label{v01}\end{aligned}\ ] ] from eqs .
( [ 4.7 ] ) and ( [ 3.16 ] ) it follows that @xmath116 developing the functions @xmath117 and @xmath118 in eq .
( [ 4.12 ] ) near the point @xmath54 one obtains @xmath119\cdot \left[\frac{2}{1+s}-\gamma+o(s+1)\right ] \nonumber \\ & & \times \left[3-\frac{13}{4}\ , ( s+1)\right ] = -\frac{13}{32}-\frac{3}{16}\gamma + \frac{3}{4}\,\zeta'(0)+ \left . \frac{3}{8}\frac{1}{s+1}\right|_{s\to
-1 } { . } \label{4.15}\end{aligned}\ ] ] when calculating @xmath113 for @xmath54 numerically we can not use the method applied in the preceding section because it requires now to take into account the next terms in the uniform asymptotic expansion ( [ 4.4 ] ) . instead of this
we calculate numerically the first 30 terms in the sum ( [ 4.9 ] ) with the result@xcite @xmath120 substituting in eq .
( [ 4.9 ] ) the logarithm by its uniform asymptotic expansion ( [ 4.4 ] ) we derive a rough estimation for @xmath113 without numerical integration @xmath121 \nonumber \\ & = & \zeta(3+s ) \gamma\left(\displaystyle{\frac{3-s}{2}}\right ) \left[\frac{27}{64}\,\gamma\left(\displaystyle{\frac{3+s}{2}}\right)- \frac{109}{32}\,\gamma\left(\displaystyle{\frac{5+s}{2}}\right)+ \frac{733}{192}\,\gamma\left(\displaystyle{\frac{7+s}{2}}\right)\right .
\nonumber \\ & & \left .
-\frac{441}{384}\,\gamma\left(\displaystyle{\frac{9+s}{2}}\right)+ \frac{707}{7680}\,\gamma\left(\displaystyle{\frac{11+s}{2}}\right)\right]{. } \label{4.17}\end{aligned}\ ] ] for @xmath54 it gives @xmath122 that is very far from eq .
( [ 4.16 ] ) having only the right sign .
summing up @xmath123 , @xmath124 with allowance for eq.([3.13 ] ) we arrive at the final result @xmath125 \nonumber \\ & = & \frac{1}{a^2}\left(-0.04345 - 0.0298\left .
\frac{1}{s+1}\right|_{s\to -1}\right)\ , . \label{4.19}\end{aligned}\ ] ] thus both the zeta functions for dirichlet and neumann boundary conditions have the pole at the point @xmath54 .
hence an additional renormalization is needed in order for a finite physical value of the relevant casimir energies to be obtained .
analysis of the spectral problem for the electromagnetic field with boundary conditions on a semi - circular cylinder ( see sec .
ii ) implies that the zeta function for this field is the sum of two zeta functions calculated in the preceding sections @xmath126 substitution of eqs .
( [ 3.19 ] ) and ( [ 4.17 ] ) into eq . ( [ 5.1 ] ) gives @xmath127 \nonumber \\ & = & \frac{1}{a^2}\left ( -0.04401 - 0.03978 \,\left.\frac{1}{s+1}\right|_{s\to -1}\right)\ , .
\label{5.2}\end{aligned}\ ] ] in both the zeta functions @xmath72 and @xmath96 the pole terms have the same sign . as a result
the pole contribution in the sum ( [ 5.1 ] ) retains .
thus , the situation here proves to be analogous to that when calculating , in the framework of zeta technique , the vacuum energy for spheres in spaces of even dimensions.@xcite as was noted above , we have derived the exact expressions for the zeta functions in question which determine these functions as analytic functions of the complex variable @xmath65 in a finite region of the plane @xmath65 containing the closed interval of the real axis @xmath128 .
it enables one to construct in a straightforward way the spectral zeta functions for relevant boundary value problem on the plane by making use of the relation @xcite @xmath129 where @xmath130 is the dirichlet or the neumann zeta function for a semi - circle , and @xmath131 is the respective zeta function for semi - circular cylinder .
we shall use this relation for calculating the values @xmath132 and @xmath133 which determine the vacuum energy of the massless scalar fields defined on the half - plane and obeying , respectively , the dirichlet or neumann boundary conditions on a semi - circle ( see fig . 1 ) . for @xmath132
we get from eqs .
( [ 5.3 ] ) .
( [ 3.4 ] ) , and ( [ 3.2 ] ) @xmath134 when @xmath135 integration in eq .
( [ 3.6 ] ) can be done explicitly with the result @xmath136 from eq .
( [ 3.12 ] ) it follows that @xmath137 numerical integration in eq .
( [ 3.8 ] ) with @xmath135 gives @xmath138 summing up eqs .
( [ 5.5 ] ) , ( [ 5.6 ] ) , and ( [ 5.7 ] ) we arrive at the result @xmath139 following the same way one can write @xmath140 using eq .
( [ 4.11 ] ) one gets @xmath141+\frac{3}{8}t^2\right\ } -\frac{\pi}{64 } \nonumber \\ & = & 0.475175-\frac{\pi}{64}\ , .
\label{5.10}\end{aligned}\ ] ] from eqs .
( [ 4.7 ] ) and ( [ 5.5 ] ) it follows that @xmath142 equation ( [ 4.12 ] ) gives @xmath143 for @xmath144 numerical integration in eq .
( [ 4.9 ] ) with @xmath135 gives @xmath145 finally , we have @xmath146\nonumber \\ & = & \frac{1}{a } \left(-0.237103 - 0.0124\left.\frac{1}{s}\right|_{s \to 0}\right)\ , . \label{5.14}\end{aligned}\ ] ] both the functions @xmath147 and @xmath148 have the pole at the point @xmath54 with the coefficients of the same ( negative ) sign .
for electromagnetic field defined on a plane the boundary conditions reduce to the neumann conditions .
hence the relevant zeta function is @xmath149 .
in the paper the spectral zeta functions are constructed for massless scalar fields obeying the dirichlet and neumann boundary conditions on a semi - circular infinite cylinder . proceeding from this ,
the zeta function for electromagnetic field is also derived for such a configuration . in all three cases ,
the final expressions for the relevant casimir energy contains the pole contribution .
hence for obtaining the physical result an additional renormalization is needed .
it is essential that for the zeta functions @xmath150 the exact formulas are derived which determine these functions in a finite region of the complex variable @xmath65 but not at the vicinity of one point @xmath54 .
this allowed one to get in a straightforward way the zeta functions for the two dimensional ( plane ) version of the boundary value problem at hand , i.e. the zeta functions for scalar fields defined on a half - plane and obeying the dirichlet and neumann boundary conditions on a semi - circle . in this case
the final expression for the vacuum energy contains the pole contributions also .
notwithstanding the spectrum of a semi - circular cylinder is very close to the spectrum of circular one , the zeta function technique does not give a finite value for vacuum energy in the first case and does for the second configuration .
in a recent paper@xcite the divergences found in our consideration are attributed to the existence of edges or corners in the boundaries under investigation .
closing , it is worth noting that , as far as we know , such boundary conditions with asymmetric geometry ( semi - circular cylinder ) has been considered in the casimir problem for the first time .
research is supported by the fund murst ex 40% and 60% , art .
this work has been accomplished during the visit of v.v.n . to the salerno university .
it is a pleasure for him to thank professor g. scarpetta , drs .
g. lambiase and a. feoli for warm hospitality .
the financial supports of iiass and istc ( project # 840 ) are acknowledged .
thanks the ue fellowship , p.o.m . 1994/1999 , for financial support .
g. plunien , b. muller , w. greiner , phys . rep . * 134 * , 87 ( 1986 ) .
v. m. mostepanenko and n. n. trunov , _ the casimir effect and its applications _
( clarendon press , oxford , 1997 ) .
e. elizalde , s. d. odintsov , a. romeo , a. a. bytsenko , and s. zerbini , _ zeta regularization technique with applications _ ( world scientific , singapore , 1994 ) .
e. elizalde , _ ten physical applications of spectral zeta functions _ ( springer , berlin , 1995 ) .
m. bordag , e. elizalde , k. kirsten , s. leseduarte , phys .
d * 56 * , 4896 ( 1997 ) .
m. bordag , k. kirsten , and d. vassilevich , phys .
d * 59 * , 085011 ( 1999 ) .
t. p. branson , p. b. gilkey , k. kirsten , d. v. vassilevich , nucl .
b * 563 * [ pm ] , 603 ( 1999 ) .
e. borgnis and c. h. papas , `` electromagnetic waveguides and resonators , '' in _ encyclopedia of physics _ , ed .
s. flge ( springer verlag , berlin , 1958 ) v. xvi , pp .
. l. l. deraad , jr . and k. milton , ann . phys .
( n.y . ) * 136 * , 229 ( 1981 )
. k. a. milton , a. v.nesterenko , and v. v. nesterenko , phys .
d * 59 * , 105009 ( 1999 ) .
g. lambiase , v. v.nesterenko , and m. bordag , j. math .
phys . * 40 * , 6254 ( 1999 ) . v.
v. nesterenko and i. g. pirozhenko , phys .
d * 60 * , 125007 ( 1999 ) .
p. gosdzinski and a. romeo , phys .
b * 441 * , 265 ( 1998 ) .
j. a. stratton , _ electromagnetic theory _ ( mcgraw - hill , new york , 1941 ) .
m. phillips , `` classical electrodynamics , '' in _ encyclopedia of physics _ , ed .
s. flge ( springer verlag , berlin , 1962 ) v. iv , pp . 1108 .
e. heyn , math .
nachrichten , * 13 * , 25 ( 1955 ) .
m. bordag , e. elizalde , and k. kirsten , j. math . phys . * 37 * , 895 ( 1996 ) .
v. v. nesterenko and i. g. pirozhenko , j. math . phys . * 41 * , 4521 ( 2000 ) . m. abramowitz and i. stegun , _ handbook of mathematical functions _
( dover publications , inc . , new york , 1972 ) .
k. a. milton , `` the casimir effect : physical manifestations of zero - point energy , '' in _ applied field theory , proc .
17th symposium on theoretical physics _ , eds . c. lee , h. min , and q .- h .
park ( chungburn , seoul , 1999 ) . c. m. bender and k. a. milton , phys .
d * 50 * , 6547 ( 1994 ) ; k. a. milton , phys .
d * 55 * , 4940 ( 1997 ) .
s. leseduarte and a. romeo , ann .
( n.y . ) * 250 * , 448 ( 1996 ) .
g. cognola , e. elizalde , and k. kirsten , `` casimir energies for spherically symmetric cavities , '' hep - th/9906228 . i. s. gradshteyn and i. m. ryzhik , _ table of integrals , series and products , fifth edition _
( academic press , san diego , ca , 1994 ) . the fortran subroutine has been used that approximates the bessel functions by chebyshev s polynomials .
j. s. dowker , `` divergences in the casimir energy , '' report - no : mutp/2000/2 ; hep - th/0006138 . | the casimir energy of a semi - circular cylindrical shell is calculated by making use of the zeta function technique .
this shell is obtained by crossing an infinite circular cylindrical shell by a plane passing through the symmetry axes of the cylinder and by considering only a half of this configuration .
all the surfaces , including the cutting plane , are assumed to be perfectly conducting .
the zeta functions for scalar massless fields obeying the dirichlet and neumann boundary conditions on the semi - circular cylinder are constructed exactly .
the sum of these zeta functions gives the zeta function for electromagnetic field in question .
the relevant plane problem is considered also . in all the cases
the final expressions for the corresponding casimir energies contain the pole contributions which are the consequence of the edges or corners in the boundaries .
this implies that further renormalization is needed in order for the finite physical values for vacuum energy to be obtained for given boundary conditions . |
rare earth pyrochlores @xmath14o@xmath1 ( @xmath15 = rare earth and @xmath16 = transition metal ) are known to present a wide range of interesting and unusual magnetic properties @xcite .
the three - dimensional network of corner - sharing tetrahedra of @xmath15 ions in these cubic symmetry ( space group @xmath17 ) pyrochlores cause highly frustrated magnetism making them a model material among the geometrically frustrated spin systems .
an interesting situation is encountered when the crystal electric field ( cef ) forces the magnetic spin to point in a particular direction .
the observation of spin - ice behavior in dy@xmath0ti@xmath0o@xmath1 and ho@xmath0ti@xmath0o@xmath1 is an exciting manifestation of such a cef forced geometrically frustrated magnetism @xcite . in these spin - ice systems
cef anisotropy forces the magnetic spins at the corners of tetrahedra to align parallel to the tetrahedron axes ( either towards the center or away from the center of tetrahedra ) .
this causes an exotic magnetic state where even the ferromagnetic interaction between the dy@xmath11 ( ho@xmath11 ) magnetic moments becomes frustrated @xcite .
the energy minimum state of magnetic exchange and dipolar interactions is achieved by two spins pointing inward and two spins outward on each tetrahedron , referred as the
2-in & 2-out " rule or ice rule " @xcite .
the constraint condition on magnetic spins is analogous to the disordered protons in hexagonal water ice , accordingly the magnetic state of dy@xmath0ti@xmath0o@xmath1 and ho@xmath0ti@xmath0o@xmath1 is named as spin - ice " .
one of the major consequences of the ice rule is that a residual entropy remains at the lowest temperatures which is almost the same as the pauling value for water ice @xcite .
another striking consequence is the realization of magnetic monopoles in spin - ice materials which interact through the magnetic coulomb interaction @xcite .
the diffuse neutron scattering study on dy@xmath0ti@xmath0o@xmath1 shows evidence of dirac strings and magnetic monopoles @xcite . polarized neutron scattering show evidence for a magnetic coulomb phase in ho@xmath0ti@xmath0o@xmath1 @xcite .
all these indicate a new state of magnetism in these materials which led to intense research activities on @xmath14o@xmath1 in recent years .
apart from the spin - ice behavior below @xmath18 k , the ac susceptibility measurements on dy@xmath0ti@xmath0o@xmath1 also show a frequency - dependent anomaly near @xmath19 k reflecting cooperative spin - freezing @xcite , however the dc susceptibility or heat capacity shows no corresponding anomaly .
this spin - freezing feature near 16 k is not seen in spin - ice ho@xmath0ti@xmath0o@xmath1 which distinguishes the spin - dynamics in dy@xmath0ti@xmath0o@xmath1 and ho@xmath0ti@xmath0o@xmath1 .
the spin - freezing behavior near @xmath20 in dy@xmath0ti@xmath0o@xmath1 is distinct from the spin - freezing in spin - glasses . in a spin - glass system
the application of a magnetic field is found to quench the spin - freezing .
in contrast , in dy@xmath0ti@xmath0o@xmath1 the application of magnetic field tends to enhance the spin - freezing @xcite , thus indicating unusual dynamics of spin - freezing in dy@xmath0ti@xmath0o@xmath1 .
we focus on this unusual glassiness in dy@xmath0ti@xmath0o@xmath1 and study the effect on spin - freezing behavior near @xmath20 upon diluting the magnetic site and doping with holes , i.e. substituting nonmagnetic ca@xmath10 for magnetic dy@xmath11 .
the effect of dilution on spin - freezing behavior near 16 k has been previously investigated for nonmagnetic y@xmath11 and lu@xmath11 substitutions @xcite .
the investigations on dy@xmath12y@xmath13ti@xmath8o@xmath9 by snyder _
et al_. @xcite revealed that the 16 k freezing is initially suppressed for low dilution concentrations ( @xmath21 ) , however the freezing behavior was found to re - emerge for @xmath22 , clearly reflecting the influence of local environment on the freezing process .
the 16 k freezing in dy@xmath0ti@xmath0o@xmath1 is considered to be an orbach - like relaxation process which involves an excited cef level for spin - lattice relaxations @xcite .
the spin dynamics near @xmath20 is controlled by transitions to / from the first excited doublet .
therefore the effect of dilution at dy site seems to result from a change in cef .
here we report the physical properties of dy@xmath6ca@xmath7ti@xmath8o@xmath9 based on ac magnetic susceptibility @xmath2 , dc magnetic susceptibility @xmath3 , isothermal magnetization @xmath4 and heat capacity @xmath5 measurements .
the results are compared with those of dy@xmath0ti@xmath0o@xmath1 and the effect of partial nonmagnetic ca@xmath10 substitution for magnetic dy@xmath11 on spin - freezing behavior near 16 k is addressed .
the @xmath2 data show the effect to be similar to that of nonmagnetic y@xmath11 substitution , however the suppression of spin - freezing behavior is much stronger in the case of ca substitution .
polycrystalline samples of dy@xmath12ca@xmath13ti@xmath8o@xmath9 ( @xmath23 , 0.2 ) were synthesized by solid state reaction method .
dy@xmath0ti@xmath0o@xmath1 was synthesized using dy@xmath0o@xmath24 ( 99.9% , sigma aldrich ) and tio@xmath0 ( 99.995% , alfa aesar ) in 1:2 molar ratio .
the stoichiometric amounts were mixed and ground for half an hour using an agate mortar and pestle .
the mixture was placed in a zirconium oxide crucible and fired thrice at 1350 @xmath25c in air with an intermediate grinding every 50 h , after which the finely ground mixture was pressed into a pellet and fired at 1350 @xmath25c for 50 h. for dy@xmath6ca@xmath7ti@xmath8o@xmath9 synthesis dy@xmath0o@xmath24 , tio@xmath0 and caco@xmath24 ( 99.95% , alfa aesar ) were taken in desired stoichiometry , mixed and ground thoroughly , and fired at 950 @xmath25c for 50 h in air .
the mixture was reground and fired at 1250 @xmath25c for 50 h with two more subsequent intermediate grindings and firings at 1350 @xmath25c for 50 h each and then the sample was pelletized and fired again at 1350 @xmath25c for 50 h. a single phase sample of 10% ca substituted dy@xmath6ca@xmath7ti@xmath8o@xmath9 could be obtained this way . however
, similar synthesis did not yield single phase sample for 20% or 30% ca substituted dy@xmath12ca@xmath13ti@xmath8o@xmath9 or 10% na substituted dy@xmath6na@xmath7ti@xmath8o@xmath9 .
a zirconium oxide crucible was used in each case .
crystal structure and quality of sample were determined by powder x - ray diffraction ( xrd ) .
bulk properties were measured using facilities at mag lab , helmholtz zentrum berlin .
the dc magnetic measurements were carried out using a vibrating sample magnetometer superconducting quantum interference device ( vsm - squid , quantum design ) and the ac susceptibility was measured using the acms option of a physical properties measurement system ( ppms , quantum design ) .
heat capacity was measured by the adiabatic relaxation method using the heat capacity option of ppms .
ca@xmath7ti@xmath8o@xmath9 recorded at room temperature .
the solid line through the experimental points is the rietveld refinement profile calculated for the @xmath26-type face centered cubic ( space group @xmath17 ) pyrochlore structure .
the short vertical bars mark the bragg peak positions .
the lowermost curve represents the difference between the experimental and calculated intensities.,width=307 ] lcc & dy@xmath0ti@xmath0o@xmath1 & dy@xmath6ca@xmath7ti@xmath8o@xmath9 + + @xmath27 ( ) & 10.1231(1 ) & 10.1311(1 ) + @xmath28 ( @xmath29 ) & 1037.396(4 ) & 1039.843(6 ) + + @xmath30 & 0.4243(5 ) & 0.4254(5 ) + + @xmath31 & 1.86 & 1.24 + @xmath32 ( % ) & 2.11 & 2.32 + @xmath33 ( % ) & 2.90 & 2.96 + the powder xrd data were analysed by structural rietveld refinement using the software fullprof @xcite .
the room temperature xrd pattern and rietveld refinement profile for dy@xmath6ca@xmath7ti@xmath8o@xmath9 are shown in figure [ fig : xrd ] .
the almost single phase nature of the sample is evident from the refinement of xrd data .
the structural refinement shows that no stuctural instability is caused by 10% ca substitution for dy , and like dy@xmath0ti@xmath0o@xmath1 , dy@xmath6ca@xmath7ti@xmath8o@xmath9 also has @xmath26-type face centered cubic ( space group @xmath17 ) pyrochlore structure .
the crystallographic parameters of dy@xmath6ca@xmath7ti@xmath8o@xmath9 are listed in table [ tab : xrd ] and compared with those of dy@xmath0ti@xmath0o@xmath1 .
the crystallographic parameters are in good agreement with the values reported for dy@xmath0ti@xmath0o@xmath1 @xcite .
dependence of ( a ) real @xmath34 and ( b ) imaginary @xmath35 parts of the ac magnetic susceptibility @xmath36 of dy@xmath6ca@xmath7ti@xmath8o@xmath9 measured at different frequencies from 100 hz to 10 khz in an applied ac magnetic field of 1.5 mt.,title="fig:",width=288 ] dependence of ( a ) real @xmath34 and ( b ) imaginary @xmath35 parts of the ac magnetic susceptibility @xmath36 of dy@xmath6ca@xmath7ti@xmath8o@xmath9 measured at different frequencies from 100 hz to 10 khz in an applied ac magnetic field of 1.5 mt.,title="fig:",width=288 ] the temperature @xmath37 dependence of ac magnetic susceptibility @xmath36 of dy@xmath6ca@xmath7ti@xmath8o@xmath9 measured in an ac field of 1.5 mt at different frequencies between 100 hz to 10 khz is shown in figure [ fig : chiac1 ] .
it is seen from figure [ fig : chiac1 ] that both the real part @xmath34 and imaginary part @xmath35 of @xmath36 show frequency dependent behavior . while at low frequencies there is no anomaly in @xmath38 or @xmath39 ,
clear anomalies related to spin freezing are seen at higher frequencies . at 10 khz an anomaly in @xmath38
is seen at around 19 k [ figure [ fig : chiac1](a ) ] and that in @xmath39 near 17 k [ figure [ fig : chiac1](b ) ] . this kind of frequency dependent behavior of @xmath36 , i.e. appearance of anomalies in @xmath38 and @xmath39 at higher frequencies have also been observed in dy@xmath8ti@xmath8o@xmath9 @xcite .
however , the frequency dependence is much weaker in dy@xmath6ca@xmath7ti@xmath8o@xmath9 than that in dy@xmath8ti@xmath8o@xmath9 .
while at 1 khz a well pronounced anomaly can be seen in the case of dy@xmath8ti@xmath8o@xmath9 @xcite , in the case of dy@xmath6ca@xmath7ti@xmath8o@xmath9 no detectable anomaly is seen at 1 khz in @xmath38 and anomaly in @xmath39 is very weak . the @xmath38 and @xmath39 data of dy@xmath6ca@xmath7ti@xmath8o@xmath9 and dy@xmath0ti@xmath0o@xmath1 at 10 khz are compared in figure [ fig : chiac2 ] . even though the anomaly temperatures of @xmath38 and @xmath39 are nearly same in both cases , the relative changes in magnitudes at these temperatures are much weaker in dy@xmath6ca@xmath7ti@xmath8o@xmath9 .
the frequency dependence of @xmath2 for dy@xmath6ca@xmath7ti@xmath8o@xmath9 is substantially weaker than the case of y substituted dy@xmath12y@xmath13ti@xmath8o@xmath9 @xcite .
thus ca@xmath10 substitution seems more effective in suppressing the spin - freezing behavior in dy@xmath8ti@xmath8o@xmath9 than y@xmath11 substitution .
dependence of real @xmath34 and imaginary @xmath35 parts of the ac magnetic susceptibility @xmath36 of dy@xmath6ca@xmath7ti@xmath8o@xmath9 and dy@xmath0ti@xmath0o@xmath1 measured at 10 khz in an applied ac magnetic field of 1.5 mt.,width=288 ] versus real part @xmath34 ) of the ac magnetic susceptibility for dy@xmath6ca@xmath7ti@xmath8o@xmath9 and dy@xmath0ti@xmath0o@xmath1 at different frequencies at 16 k.,width=288 ] the cole - cole plot of @xmath35 versus @xmath34 for dy@xmath6ca@xmath7ti@xmath8o@xmath9 is shown in figure [ fig : chiac3 ] for @xmath40 k. the cole - cole plot depicts the spin - dynamics at a given temperature and accesses the distribution of relaxation times of spin - freezing processes .
while a semicircular character of plot indicates a single relaxation time or a narrow distribution of relaxation times , the semicircle becomes flattened and/or distorted when there is a distribution of relaxation times . from figure [ fig : chiac3 ] we see that the cole - cole plot for dy@xmath6ca@xmath7ti@xmath8o@xmath9 has a semicircular character similar to that of dy@xmath8ti@xmath8o@xmath9 , though the center of the circle is not same for the two . within the cole - cole formalism
, the @xmath41 data can be fitted to the equation of semicircular arc given by @xcite @xmath42 } \nonumber \\ & + \sqrt{(\chi'-\chi_{\rm s})(\chi_0-\chi')+\frac{(\chi_0 - \chi_{\rm s})^2}{4 \tan^2[(1-\alpha)\frac{\pi}{2}]}}. \label{eq : cole - cole}\end{aligned}\ ] ] here @xmath43 is the isothermal susceptibility ( in the limit of low frequency ) and @xmath44 the adiabatic susceptibility ( in the limit of high frequency ) .
the parameter @xmath45 represents the width of the distribution of relaxation times , @xmath46 implies a single relaxation time and @xmath47 an infinitely wide distribution .
the fits of @xmath41 by eq .
( [ eq : cole - cole ] ) are shown by solid curves in figure [ fig : chiac3 ] . from the fits we obtained @xmath48
emu / mol , @xmath49 emu / mol and @xmath50 for dy@xmath6ca@xmath7ti@xmath8o@xmath9 , and @xmath51 emu / mol , @xmath52 emu / mol and @xmath53 for dy@xmath8ti@xmath8o@xmath9 .
accordingly we obtain @xmath54 which is very close to 0 for both samples .
the different values of @xmath43 and @xmath44 reflect different center and radius of the semicircular arc for dy@xmath6ca@xmath7ti@xmath8o@xmath9 and dy@xmath8ti@xmath8o@xmath9 .
however , @xmath45 is identical for both .
the semicircular nature of cole - cole plot and the extremely small values of @xmath45 suggest a single relaxation mode in dy@xmath6ca@xmath7ti@xmath8o@xmath9 as well as in dy@xmath8ti@xmath8o@xmath9 @xcite .
of dy@xmath6ca@xmath7ti@xmath8o@xmath9 as a function of temperature @xmath37 for @xmath55 k measured in magnetic fields @xmath56 t and 0.05 t. inset : inverse magnetic susceptibility @xmath57 for @xmath58 k in @xmath59 t. the solid red line is the fit of the @xmath57 data by the curie - weiss law in @xmath60 k and the dashed line is an extrapolation.,width=288 ] the zero - field - cooled dc magnetic susceptibility @xmath3 of dy@xmath6ca@xmath7ti@xmath8o@xmath9 measured in @xmath61 t and 0.05 t are shown in figure [ fig : mt ] .
the @xmath3 data do not show any anomaly near @xmath20 and paramagnetic behavior is evident from the @xmath3 data at @xmath62 k. the magnitude of @xmath63 at low-@xmath37 is quite large .
a similar @xmath3 behavior has been observed for dy@xmath0ti@xmath0o@xmath1 @xcite . from figure
[ fig : mt ] it is evident that the @xmath3 measured at 0.05 t is higher in magnitude than that at 0.01 t. this reflects the ferromagnetic nature of magnetic exchange as in the dy@xmath0ti@xmath0o@xmath1 . at high temperature ,
the @xmath3 data of dy@xmath6ca@xmath7ti@xmath8o@xmath9 follow curie - weiss behavior , @xmath64
. a linear fit of inverse magnetic susceptibility @xmath57 data ( measured in @xmath59 t ) over 150 k @xmath65 300 k yields the curie constant @xmath66 emuk / mol and weiss temperature @xmath67 k. the @xmath57 data when fitted over 20 k @xmath65 50
k yield @xmath68 k and @xmath69 emuk / mol .
the linear fit of the @xmath70 t @xmath57 data in 20 k @xmath65 50 k is shown by solid red line in the inset of figure [ fig : mt ] . like dy@xmath6ca@xmath7ti@xmath8o@xmath9 two different values of @xmath71 have also been found in the case of dy@xmath0ti@xmath0o@xmath1 : @xmath72 k and @xmath73 k from high-@xmath37 and low-@xmath37 fit ranges , respectively @xcite .
a positive @xmath71 has been reported for dy@xmath0ti@xmath0o@xmath1 from the analysis of @xmath3 data below 40 k @xcite .
demagnetizing effect is argued to cause the @xmath71 to be negative and when corrected for this the @xmath71 would be positive such that effectively there is a weak ferromagnetic interaction in dy@xmath0ti@xmath0o@xmath1 @xcite .
a similar situation should be applicable to present data . our sample for dc magnetic susceptibility measurements consisted of several small pieces of arbitrary shape and sizes
therefore we do not have a good estimate of demagnetization factor to correct the @xmath3 data for the demagnetizing field .
a crude approximation assuming spherical shape with demagnetization factor @xmath74 yields @xmath75 k on fitting the corrected @xmath3 data by curie - weiss law in 20 k @xmath76 k with @xmath77 emuk / mol .
thus the demagnetization correction gives a positive @xmath71 and hence a ferromagnetic exchange .
we estimate an effective moment @xmath78/dy from the value of @xmath79 emuk / mol which is very close to the theoretically expected value of @xmath80 for @xmath81g@xmath82 ground state of dy@xmath11 ions .
of dy@xmath6ca@xmath7ti@xmath8o@xmath9 as a function of applied magnetic field @xmath83 for @xmath84 t measured at the indicated temperatures.,width=288 ] the isothermal @xmath4 data of dy@xmath6ca@xmath7ti@xmath8o@xmath9 measured at selected temperatures of 2 k , 20 k , and 50 k are shown in figure [ fig : mh ] . at 2
k the @xmath4 isotherm exhibits a rapid increase below 0.5 t and tends to saturate thereafter with a saturation value of magnetization @xmath85/dy at 7 t. this value of @xmath86 is close to the value expected for [ 100 ] direction @xmath87/dy on account of fm interaction and ising anisotropy @xcite .
however , the @xmath86 value is much lower than the theoretical value of @xmath88 for free dy@xmath11 ions ( @xmath89 and @xmath90 ) , in line with the presence of strong ising anisotropy in this compound .
the saturation behavior of @xmath4 isotherms depends on the temperature : @xmath86 decreases with increasing temperature .
the overall @xmath4 behavior of dy@xmath6ca@xmath7ti@xmath8o@xmath9 is similar to that of dy@xmath0ti@xmath0o@xmath1 @xcite .
for a polycrystalline dy@xmath0ti@xmath0o@xmath1 , @xmath91/dy at 1.8 k @xcite .
thus we see an increase of about 12% in @xmath86 value at 2 k with 10% ca substitution for dy .
this effect of ca substitution is quite distinct from y substitution where dilution has negligible effect on saturation moment [ 15 ] . in dy@xmath0ti@xmath0o@xmath1
there is a strong ising anisotropy along the local @xmath92 direction that determines the ice rule and hence the magnetic behavior of this compound .
the @xmath4 data of dy@xmath0ti@xmath0o@xmath1 obey ice rule ( two - in , two - out configuration ) below 0.5 t and at fields higher than this zeeman energy exceeds the effective nearest neighbor ferromagnetic interaction which causes a breaking of ice rule along the [ 111 ] direction @xcite .
the [ 111 ] direction @xmath4 data for @xmath93 t with @xmath91/dy at 1.8 k ( much higher than the ice rule expected value of 3.33@xmath94/dy ) reveal three - in , one - out spin configuration , clearly reflecting the breaking of ice rule at high fields @xcite . for a spin ice system with two - in , two - out configuration
the predicted values of @xmath86 for the [ 100 ] , [ 110 ] and [ 111 ] directions are @xmath95 ( @xmath96/dy ) , @xmath97 ( @xmath98/dy ) and @xmath99 ( @xmath100/dy ) , respectively , where @xmath101 is the full saturation moment @xcite .
therefore for two - in , two - out configuration the weighted average @xmath86 value for the polycrystalline sample should be around @xmath102/dy according to @xmath103 } + 12\ , m_{[110 ] } + 8\ , m_{[111]})/26 $ ] , where the weight factors 6 , 12 and 8 are the number of equivalent directions for [ 100 ] , [ 110 ] and [ 111 ] , respectively .
our value of @xmath104 for dy@xmath6ca@xmath7ti@xmath8o@xmath9 is @xmath105/dy at 0.5 t and 2 k which is compatible with the theoretically expected ice rule value [ @xmath102/dy ] and is very close to the measured weighted average value ( over [ 100 ] , [ 110 ] and [ 111 ] directions ) of @xmath106/dy for single crystal dy@xmath0ti@xmath0o@xmath1 at 0.5 t and 1.8 k @xcite .
however , the observed @xmath104 at 0.5 t is about 14% higher than the @xmath107/dy reported for polycrystalline dy@xmath0ti@xmath0o@xmath1 at 0.5 t and 1.8 k in ref .
thus , the saturation behavior of @xmath4 of dy@xmath6ca@xmath7ti@xmath8o@xmath9 is substantially different from both undoped as well as y - doped dy@xmath0ti@xmath0o@xmath1 @xcite .
while we do not have a clear explanation for the increased @xmath108 for dy@xmath6ca@xmath7ti@xmath8o@xmath9 we suspect that a change in crystal field effect or a change in exchange interaction by ca substitution is the origin for the observed behavior .
the charge / valence consideration would suggest three possible situations : 1 ) ( @xmath109 i.e. with a deficiency of positive charge resulting in a hole doping .
2 ) ( @xmath110 i.e. a deficiency of oxygen concentration to balance the charge such that the composition is dy@xmath6ca@xmath7ti@xmath8o@xmath111 instead of dy@xmath6ca@xmath7ti@xmath8o@xmath9 .
3 ) ( @xmath112 i.e. a change in valence state of ti , however this would require an increase in valence of ti to compensate for loss in positive charge at dy / ca site if there is no change in oxygen concentration .
we would like to point out that ti is not known to have a valence state higher than 4 + , therefore this seems very unlikely . on the other hand from the view point of charge ( valence ) balancing
an oxygen deficiency is possible .
however , an oxygen deficiency in dy@xmath0ti@xmath0o@xmath1 has been found to reduce the saturation magnetization @xcite .
contrary to such an expectation , an increase in saturation magnetization in dy@xmath6ca@xmath7ti@xmath8o@xmath9 does not support the oxygen deficiency in this compound .
an increase in saturation magnetization can be brought by oxygen deficiency if the deficiency concentration is more than what is required to balance the loss in positive charge upon ca substitution such that a fraction of ti becomes magnetic , i.e. ti@xmath11 and the increased magnetization comes from ti .
another possible origin of excess magnetization can be canting of dy@xmath11 moments from the easy @xmath92 direction upon the partial removal of dy@xmath11 spins from the vertices of a tetrahedron with two - in , two - out spin configuration as a result of ca@xmath10 substitution .
ca@xmath10 is larger than dy@xmath11 in size and may distort the crystal field , changing slightly the dy@xmath11 wavefunction and resulting anisotropy .
if one of four dy is replaced by ca then such tetrahedra will have either two - in , one - out or one - in , two - out configuration . thus assuming uniform distribution of ca , about 10% of the tetrahedra in dy@xmath6ca@xmath7ti@xmath8o@xmath9 will no longer fulfil the two - in , two - out spin configuration and
when these tetrahedra are subjected to high magnetic fields they may behave differently , canting of spins could be one such situation .
( color online ) heat capacity @xmath113 as a function of temperature @xmath37 for 1.8 k @xmath114 300 k measured in zero field for ( a ) dy@xmath6ca@xmath7ti@xmath8o@xmath9 and ( b ) dy@xmath8ti@xmath8o@xmath9 .
the red solid curves are the sum of the contributions from the debye+einstein models and the crystal electric field according to eq .
[ eq : hc - fit ] .
insets : low-@xmath37 @xmath5 for @xmath115 k. ( c ) @xmath116 versus @xmath37 plot of @xmath5 data of dy@xmath6ca@xmath7ti@xmath8o@xmath9 and dy@xmath8ti@xmath8o@xmath9 in @xmath117 k. inset : @xmath116 versus @xmath118 plot for dy@xmath6ca@xmath7ti@xmath8o@xmath9 for @xmath119 k on an expanded @xmath120-scale .
the solid red line is the fit to @xmath121 in @xmath122 k and the dashed lines are extrapolations.,title="fig:",width=288 ] ( color online ) heat capacity @xmath113 as a function of temperature @xmath37 for 1.8 k @xmath114 300 k measured in zero field for ( a ) dy@xmath6ca@xmath7ti@xmath8o@xmath9 and ( b ) dy@xmath8ti@xmath8o@xmath9 .
the red solid curves are the sum of the contributions from the debye+einstein models and the crystal electric field according to eq .
[ eq : hc - fit ] .
insets : low-@xmath37 @xmath5 for @xmath115 k. ( c ) @xmath116 versus @xmath37 plot of @xmath5 data of dy@xmath6ca@xmath7ti@xmath8o@xmath9 and dy@xmath8ti@xmath8o@xmath9 in @xmath117 k. inset : @xmath116 versus @xmath118 plot for dy@xmath6ca@xmath7ti@xmath8o@xmath9 for @xmath119 k on an expanded @xmath120-scale .
the solid red line is the fit to @xmath121 in @xmath122 k and the dashed lines are extrapolations.,title="fig:",width=288 ] ( color online ) heat capacity @xmath113 as a function of temperature @xmath37 for 1.8 k @xmath114 300 k measured in zero field for ( a ) dy@xmath6ca@xmath7ti@xmath8o@xmath9 and ( b ) dy@xmath8ti@xmath8o@xmath9 .
the red solid curves are the sum of the contributions from the debye+einstein models and the crystal electric field according to eq .
[ eq : hc - fit ] .
insets : low-@xmath37 @xmath5 for @xmath115 k. ( c ) @xmath116 versus @xmath37 plot of @xmath5 data of dy@xmath6ca@xmath7ti@xmath8o@xmath9 and dy@xmath8ti@xmath8o@xmath9 in @xmath117 k. inset : @xmath116 versus @xmath118 plot for dy@xmath6ca@xmath7ti@xmath8o@xmath9 for @xmath119 k on an expanded @xmath120-scale .
the solid red line is the fit to @xmath121 in @xmath122 k and the dashed lines are extrapolations.,title="fig:",width=288 ] the @xmath5 data of dy@xmath6ca@xmath7ti@xmath8o@xmath9 are shown in figure [ fig : hc](a ) . it is seen from figure [ fig : hc](a ) that the @xmath5 data at @xmath123 k is much lower than the classical dulong - petit value @xmath124 j / molk , where @xmath15 is the molar gas constant and @xmath125 is the number of atoms per formula unit @xcite .
the @xmath5 data of dy@xmath8ti@xmath8o@xmath9 shown in figure [ fig : hc](b ) also fails to reach the high-@xmath37 dulong - petit limit by 300 k. the lower value of @xmath113 at @xmath123 k than its high-@xmath37 limit is consistent with the high debye temperature ( @xmath126 much greater than 300 k ) of pyrochlore 227 compounds . like @xmath3
, the @xmath5 data do not show any anomaly near @xmath20 . at low-@xmath37 an increase
is observed in the @xmath113 of dy@xmath6ca@xmath7ti@xmath8o@xmath9 with decreasing @xmath37 at @xmath127 k [ inset of figure [ fig : hc](a ) ] .
this behavior of @xmath5 is similar to that of dy@xmath8ti@xmath8o@xmath9 [ inset of figure [ fig : hc](b ) ] .
a comparison of @xmath116 versus @xmath37 plot for dy@xmath6ca@xmath7ti@xmath8o@xmath9 and dy@xmath8ti@xmath8o@xmath9 in figure [ fig : hc](c ) shows that there is almost no change in upturn feature of @xmath5 by 10% ca substitution . at low-@xmath37 heat capacity
is usually expressed by @xmath128 , where @xmath129 and @xmath130 are coefficients to the electronic and lattice contributions to heat capacity , respectively .
the sommerfeld coefficient @xmath129 is nonzero for a metallic sample .
however , for an insulating material @xmath131 , therefore @xmath5 can be described by @xmath132 term only .
we fitted the @xmath5 data of dy@xmath6ca@xmath7ti@xmath8o@xmath9 by @xmath128 and found that @xmath129 is indeed zero which is consistent with the insulating characteristic of dy@xmath6ca@xmath7ti@xmath8o@xmath9 .
the fit of @xmath116 versus @xmath118 by @xmath121 in @xmath122 k is shown by solid red line in the inset of figure [ fig : hc](c ) which gives @xmath133 j / molk@xmath134 .
a similar analysis of @xmath5 of dy@xmath0ti@xmath0o@xmath1 gives @xmath135 j / molk@xmath134 in very good agreement with the reported value of @xmath136 j / molk@xmath134 @xcite .
an estimate of debye temperature @xmath126 follows from @xmath130 , as @xmath137 , accordingly from the low-@xmath37 @xmath5 data we get @xmath138 k for dy@xmath0ti@xmath0o@xmath1 and @xmath139 k for dy@xmath6ca@xmath7ti@xmath8o@xmath9 .
the value of @xmath126 for dy@xmath0ti@xmath0o@xmath1 is in very good agreement with the reported value of @xmath140 k obtained from the @xmath5 data below 30 k @xcite .
the @xmath126 values obtained above failed to fit the high-@xmath37 @xmath5 data using the debye model lattice heat capacity due to acoustic phonons therefore we analyzed the @xmath5 data using a combination of the debye model and einstein model of phonon heat capacity , thus accounting for both acoustic and optic modes of lattice vibration .
the @xmath5 data were fitted by @xmath141 where @xmath142 represents the debye lattice heat capacity , @xmath143 represents the einstein lattice heat capacity and @xmath144 represents the crystal electric field contribution to heat capacity .
the phonon contributions @xmath142 and @xmath143 are given by @xcite @xmath145 and @xmath146 where @xmath147 is the einstein temperature .
for a two - level system @xmath144 is given by @xcite @xmath148 ^ 2 } , \label{eq : hc - cef_twolevel}\ ] ] where @xmath149 is the degeneracy of the ground state , @xmath150 is the degeneracy of first excited state , and @xmath151 is cef splitting energy between the ground state and the first excited state . under the action of cef the ( @xmath152)-fold degenerate ground state of dy@xmath11 ( @xmath90 ) splits into eight doublets , therefore @xmath153
. there are few estimates of cef splitting energy @xmath151 for dy@xmath0ti@xmath0o@xmath1 , however the cef parameters have not been determined precisely by inelastic neutron scattering .
the reported values of @xmath151 are different ranging from 100 @xmath154 to 287 @xmath154 @xcite .
we use the value @xmath155
@xmath154 determined from the cef analysis of magnetic susceptibility data @xcite .
the cef contribution is only a few percent of the total @xmath113 therefore uncertainty in estimate of @xmath144 is not expected to have significant effect on the analysis of the phonon contribution using debye+einstein models .
the @xmath144 was calculated using eq .
[ eq : hc - cef_twolevel ] for @xmath153 and @xmath155 @xmath154 and subtracted off from the measured data and then the [ @xmath156 data of dy@xmath0ti@xmath0o@xmath1 were fitted by debye+einstein models according to eqs .
( [ eq : hc - fit])([eq : eistein_hc ] ) .
a reasonable fit is obtained for @xmath157 i.e. with 71% weight to debye term and 29% to einstein term .
the sum of @xmath144 and debye+einstein models fit for dy@xmath0ti@xmath0o@xmath1 is shown by solid red curve in figure [ fig : hc](b ) . while fitting the data we used analytic pad approximant for @xmath158 as developed in ref .
the @xmath159 k and @xmath160 k were obtained from the fit . for y@xmath0ti@xmath0o@xmath1
a value of @xmath161 k has been estimated @xcite using the information from the first principles calculation of elastic properties of y@xmath0ti@xmath0o@xmath1 , which when scaled according to @xmath162 gives @xmath163 k for dy@xmath0ti@xmath0o@xmath1 , where @xmath104 and @xmath164 are the mass and volume per formula unit , respectively and the subscripts @xmath165 and @xmath16 label the materials @xmath165 and @xmath16 , respectively .
thus the value of @xmath159 k determined from debye+einstein models fit for dy@xmath0ti@xmath0o@xmath1 is consistent with the expected large value of scaled @xmath126 compared to the value of @xmath126 determined from coefficient @xmath130 above . since not much effect is seen on the @xmath5 data by 10% ca substitution for dy , to analyse the @xmath5 data of dy@xmath6ca@xmath7ti@xmath8o@xmath9 by debye+einstein model we used the same @xmath144 that we estimated for dy@xmath0ti@xmath0o@xmath1 .
the sum of @xmath144 and debye+einstein models fit for dy@xmath6ca@xmath7ti@xmath8o@xmath9 is shown by the solid red curve in figure [ fig : hc](a ) . in this case
we obtained @xmath166 , however the @xmath167 k and @xmath168 k obtained from the fit are slightly different from those of dy@xmath0ti@xmath0o@xmath1 .
the increase in @xmath126 and @xmath169 with ca substitution is consistent with their formula mass dependence according to eq .
( [ eq : debye_t ] ) as the formula mass is lowered by ca substitution .
the effect of partial nonmagnetic ca@xmath10 substitution for magnetic dy@xmath11 in spin - ice system dy@xmath0ti@xmath0o@xmath1 has been investigated by @xmath2 , @xmath3 , @xmath4 and @xmath5 measurements on dy@xmath6ca@xmath7ti@xmath8o@xmath9 .
the high - frequency @xmath2 data of dy@xmath6ca@xmath7ti@xmath8o@xmath9 show evidence for spin - freezing type anomaly near 16 k , however no corresponding anomaly is observed in @xmath3 and @xmath5 data .
the @xmath3 and @xmath5 data of dy@xmath0ti@xmath0o@xmath1 also do not show any anomaly near 16 k. the low-@xmath37 @xmath5 data give sommerfeld coefficient @xmath131 which reflects an insulating ground state for dy@xmath6ca@xmath7ti@xmath8o@xmath9 similar to dy@xmath0ti@xmath0o@xmath1 .
thus the holes produced by 10% ca@xmath10 substitution for dy@xmath11 ions is not sufficient to change the electrical properties of insulating dy@xmath0ti@xmath0o@xmath1 .
the high-@xmath37 @xmath5 data were analyzed by the debye+einstein models of lattice heat capacity along with a cef contribution which yielded @xmath167 k and @xmath168 k for dy@xmath6ca@xmath7ti@xmath8o@xmath9 and @xmath159 k and @xmath160 k for dy@xmath0ti@xmath0o@xmath1 .
the spin - glass type freezing behavior near 16 k in dy@xmath0ti@xmath0o@xmath1 is strongly influenced by ca substitution .
the frequency dependent @xmath2 data of dy@xmath6ca@xmath7ti@xmath8o@xmath9 show that the suppression of spin - freezing behavior by ca@xmath10 is substantially increased compared to that by the nonmagnetic isovalent y@xmath11 in dy@xmath6y@xmath7ti@xmath8o@xmath9 @xcite .
an analysis of @xmath2 within the cole - cole formalism suggests that no change in semicircular nature of cole - cole plot of @xmath41 results from ca substitution and the spin dynamics of dy@xmath6ca@xmath7ti@xmath8o@xmath9 is governed by a single relaxation time like dy@xmath8ti@xmath8o@xmath9 @xcite . the ca@xmath10 substitution for dy from the viewpoint of charge ( valence ) balancing can result in a hole doping or an oxygen deficiency .
since an oxygen deficiency in dy@xmath0ti@xmath0o@xmath1 has been found to reduce the saturation magnetization @xcite , the increased saturation magnetization in dy@xmath6ca@xmath7ti@xmath8o@xmath9 is counter - intuitive for oxygen deficiency in this compound .
as such the hole doping scenario seems more appropriate , however the holes produced by 10% ca doping apparently fails to bring any noticeable change in band structure to reveal metallic nonzero @xmath129 in heat capacity measurements .
it should be noted though that we can not completely rule out the possibility of o deficiency and the compound may have a composition dy@xmath6ca@xmath7ti@xmath8o@xmath170 instead of dy@xmath6ca@xmath7ti@xmath8o@xmath9 . in pure dy@xmath0ti@xmath0o@xmath1
the spin relaxation has been suggested to exhibit an unusual double crossover , the spin relaxation is thermally activated at temperatures above a crossover temperature @xmath171 k below which the relaxation process is dominated by quantum tunneling down to @xmath172 k and then again a thermally activated behavior dominates below @xmath173 as the spins become strongly correlated below this temperature @xcite .
et al_. @xcite made an attempt to explain the difference between the 16 k freezing in dy@xmath0ti@xmath0o@xmath1 ( presence ) and ho@xmath0ti@xmath0o@xmath1 ( absence ) based on the competition between the thermally activated and quantum spin tunneling in these compounds .
considering that the spin dynamics near 16 k can be characterized by an arrhenius type behavior with an activation energy @xmath174 k which is close to the cef splitting energy ( 240340 k ) , the spin freezing is considered to involve transitions to and from the first excited doublet @xcite . since dy@xmath11 ( @xmath90 ) is a kramers ion , a direct relaxation process
is not allowed unless the degeneracy of ground state doublet is removed ( e.g. by the application of magnetic field ) .
therefore an orbach process is considered as an alternative mechanism of relaxation in dy@xmath0ti@xmath0o@xmath1 which is a two stage process in which a spin first absorbs a phonon of the same magnitude as the energy of the excited crystal field state and then relaxes to a state lower in energy than its initial state by emitting a phonon of different energy than the absorbed one .
ho@xmath11 ( @xmath175 ) on the other hand being non - kramers ion , no orbach - like relaxation is observed near 16 k in ho@xmath0ti@xmath0o@xmath1 .
et al_. suggested that ho@xmath0ti@xmath0o@xmath1 can reveal 16 k feature under the application of magnetic field as confirmed by the neutron spin echo experiments @xcite .
considering the effect of ca substitution on spin - freezing within the orbach - like relaxation formalism that involves both the lattice degrees of freedom and the excited crystal field levels for spin - lattice relaxation , it seems that a significant change in the crystal field scheme / energy is brought by the ca substitution . with a coordination number of 8 ,
the y@xmath11 ( 101.9 pm ) has similar ionic radius as dy@xmath11 ( 102.7 pm ) , whereas ca@xmath10 with an ionic radius of 112 pm is larger than dy in size @xcite .
the larger size of ca@xmath10 might be at the origin for its effectiveness in changing the magnetic exchange interaction and/or crystal field , and could be the reason for stronger smearing of spin - freezing anomaly in ac susceptibility than by y@xmath11 . a comparative study of crystal field level schemes of pure dy@xmath0ti@xmath0o@xmath1 as well as ca- and y - substituted ones would be informative in understanding the magnetic behaviors of these compounds .
in a recent study of effect of dilution on spin - ice behavior scharffe et al .
@xcite found evidence for suppression of pauling s residual entropy in dy@xmath12y@xmath176ti@xmath0o@xmath1 for @xmath177 with a non - degenerate ground state and accordingly a rapid crossover from spin - ice behavior in dy@xmath0ti@xmath0o@xmath1 to a weakly interacting single - ion physics is inferred for intermediate dilution in dy@xmath12y@xmath176ti@xmath0o@xmath1 .
it would be of interest to see how ca substitution for dy influences the spin - ice behavior and associated monopole dynamics .
further investigations in this direction would be enlightening .
we thank b. klemeke and k. siemensmeyer for their experimental assistance and helpful discussions .
we acknowledge helmholtz gemeinschaft for funding via the helmholtz virtual institute ( project no .
vh - vi-521 ) . | physical properties of partially ca substituted hole - doped dy@xmath0ti@xmath0o@xmath1 have been investigated by ac magnetic susceptibility @xmath2 , dc magnetic susceptibility @xmath3 , isothermal magnetization @xmath4 and heat capacity @xmath5 measurements on dy@xmath6ca@xmath7ti@xmath8o@xmath9 .
the spin - ice system dy@xmath0ti@xmath0o@xmath1 exhibits a spin - glass type freezing behavior near 16 k. our frequency dependent @xmath2 data of dy@xmath6ca@xmath7ti@xmath8o@xmath9 show that the spin - freezing behavior is significantly influenced by ca substitution .
the effect of partial nonmagnetic ca@xmath10 substitution for magnetic dy@xmath11 is similar to the previous study on nonmagnetic isovalent y@xmath11 substituted dy@xmath12y@xmath13ti@xmath8o@xmath9 ( for low levels of dilution ) , however the suppression of spin - freezing behavior is substantially stronger for ca than y. the cole - cole plot analysis reveals semicircular character and a single relaxation mode in dy@xmath6ca@xmath7ti@xmath8o@xmath9 as for dy@xmath8ti@xmath8o@xmath9 .
no noticeable change in the insulating behavior of dy@xmath0ti@xmath0o@xmath1 results from the holes produced by 10% ca@xmath10 substitution for dy@xmath11 ions . |
the organizational principles driving the evolution and development of natural and social large - scale systems , including populations of bacteria , ant colonies , herds of predators and human societies , rely on the cooperation of a large population of unrelated agents @xcite .
even if cooperation seems to be a ubiquitous property of social systems , its spontaneous emergence is still a puzzle for scientists since cooperative behaviors are constantly threatened by the natural tendency of individuals towards self - preservation and the never - ceasing competition among agents for resources and success . the preference of selfishness over cooperation is also due to the higher short - term benefits that a single ( defector ) agent obtains by taking advantage of the efforts of cooperating agents .
obviously , the imitation of such a selfish ( but rational ) conduct drives the system towards a state in which the higher benefits associated to cooperation are no longer achievable , with dramatic consequences for the whole population .
consequently , the relevant question to address is why cooperative behavior is so common , and which are the circumstances and the mechanisms that allow it to emerge and persist . in the last decades
, the study of the elementary mechanisms fostering the emergence of cooperation in populations subjected to evolutionary dynamics has attracted a lot of interest in ecology , biology and social sciences @xcite .
the problem has been tackled through the formulation of simple games that neglect the microscopic differences among distinct social and natural systems , thus providing a general framework for the analysis of evolutionary dynamics @xcite .
most of the classical models studied within this framework made the simplifying assumption that social systems are characterized by homogeneous structures , in which the interaction probability is the same for any pair of agents and constant over time @xcite .
however , this assumption has been proven false for real systems , as the theory of complex networks has revealed that most natural and social networks exhibit large heterogeneity and non trivial interconnection topologies @xcite .
it has been also shown that the structure of a network has dramatic effects on the dynamical processes taking place on it , so that complex networks analysis has become a fundamental tool in epidemiology , computer science , neuroscience and social sciences @xcite .
the study of evolutionary games on complex topologies has allowed a new way out for cooperation to survive in some paradigmatic cases such as the prisoner s dilemma @xcite or the public goods games @xcite .
in particular , it has been pointed out that the complex patterns of interactions among the agents found in real social networks , such as scale - free distributions of the number of contacts per individual or the presence of tightly - knit social groups , tend to favor the emergence and persistence of cooperation .
this line of research , which brings together the tools and methods from the statistical mechanics of complex networks and the classical models of evolutionary game dynamics , has effectively became a new discipline , known as evolutionary graph theory @xcite .
recently , the availability of longitudinal spatio - temporal information about human interactions and social relationships @xcite has revealed that social systems are not static objects at all : contacts among individuals are usually volatile and fluctuate over time @xcite , face - to - face interactions are bursty and intermittent @xcite , agents motion exhibits long spatio - temporal correlations @xcite . consequently , static networks , constructed by aggregating in a single graph all the interactions observed among a group of individuals across a given period , can be only considered as simplified models of real networked systems . for this reason , time - varying graphs have been lately introduced as a more realistic framework to encode time - dependent relationships @xcite . in particular
, a time - varying graph is an ordered sequence of graphs defined over a fixed number of nodes , where each graph in the sequence aggregates all the edges observed among the nodes within a certain temporal interval .
the introduction of time as a new dimension of the graph gives rise to a richer structure .
therefore , new metrics specifically designed to characterize the temporal properties of graph sequences have been proposed , and most of the classical metrics defined for static graphs have been extended to the time - varying case @xcite . lately , the study of dynamical processes taking place on time - evolving graphs has shown that temporal correlations and contact recurrence play a fundamental role in diverse settings such as random walks dynamics @xcite , the spreading of information and diseases @xcite and synchronization @xcite . here
we study how the level of cooperation is affected by taking into account the more realistic picture of social system provided by time - varying graphs instead of the classical ( static ) network representation of interactions .
we consider a family of social dilemmas , including the hawk - dove , the stag hunt and the prisoner s dilemma games , played by agents connected through a time - evolving topology obtained from real traces of human interactions .
we analyze the effect of temporal resolution and correlations on the emergence of cooperation in two paradigmatic data sets of human proximity , namely the mit reality mining @xcite and the infocom06 @xcite co - location traces .
we find that the level of cooperation achievable on time - varying graphs crucially depends on the interplay between the speed at which the network changes and the typical time - scale at which agents update their strategy .
in particular , cooperation is facilitated when agents keep playing the same strategy for longer intervals , while too frequent strategy updates tend to favor defectors .
our results also suggest that the presence of temporal correlations in the creation and maintenance of interactions hinders cooperation , so that synthetic time - varying networks in which link persistence is broken usually exhibit a considerably higher level of cooperation .
finally , we show that both the average size of the giant component and the weighted temporal clustering calculated across different consecutive time - windows are indeed good predictors of the level of cooperation attainable on time - varying graphs .
we focus on the emergence of cooperation in systems whose individuals face a social dilemma between two possible strategies : _ cooperation _ ( c ) and _ defection _ ( d ) . a large class of social dilemmas can be formulated as in @xcite via a two - parameter game described by the payoff matrix : @xmath0 where @xmath1 , @xmath2 , @xmath3 and @xmath4 represent the payoffs corresponding to the various possible encounters between two players .
namely , when the two players choose to cooperate they both receive a payoff @xmath5 ( for _ reward _ ) , while if they both decide to defect they get @xmath6 ( for _ punishment _ ) . when a cooperator faces a defector it gets the payoff @xmath2 ( for _ sucker _ ) while the defector gets @xmath3 ( for _ temptation _ ) . in this version of the game
the payoffs @xmath2 and @xmath3 are the only two free parameters of the model , and their respective values induce an ordering of the four payoffs which determines the type of social dilemma .
we have in fact three different scenarios .
when @xmath7 and @xmath8 , defecting against a cooperator provides the largest payoff , and this corresponds to the _ hawk - dove _ game . for @xmath9 and @xmath10 , cooperating with a defector is the worst case , and we have the _ stag hunt _ game . finally , for @xmath7 and @xmath10
, when a defector plays with a cooperator we have at the same time the largest ( for the defector ) and the smallest ( for the cooperator ) payoffs , and the game corresponds to the _ prisoner s dilemma_. in this work we consider the three types of games by exploring the parameter region @xmath11 $ ] and @xmath12 $ ] . in real social systems ,
each individual has more than one social contact at the same time .
this situation is usually represented @xcite by associating each player @xmath13 to a node of a _ static _
network , with adjacency matrix @xmath14 , whose edges indicate pairs of individuals playing the game . in this framework ,
a player @xmath15 selects a strategy , plays a number of games equal to the number of her neighbors @xmath16 and accumulates the payoffs associated to each of these interactions .
obviously , the outcome of playing with a neighbor depends both on the strategiy selected by node @xmath15 and that of the neighbor , according to the payoff matrix in eq .
[ payoffs ] .
when all the individuals have played with all their neighbors in the network , they update their strategies as a result of an evolutionary process , i.e. , according to the total collected payoff .
namely , each individual @xmath15 compares her cumulated payoff @xmath17 with that of one of her neighbors , say @xmath18 , chosen at random .
the probability @xmath19 that agent @xmath15 adopts the strategy of her neighbor @xmath18 increases with the difference @xmath20 ( see methods for details ) .
the games defined by the payoff matrix in eq . [ payoffs ] and using a payoff - based strategy update rule have been thoroughly investigated in static networks with different topologies .
the main result is that , when the network is fixed and agent strategies are allowed to evolve over time , the level of cooperation increases with the heterogeneity of the degree distribution of the network , being scale - free networks the most paradigmatic promoters of cooperation @xcite .
however , in most cases human contacts and social interactions are intrinsically dynamic and varying in time , a feature which has profound consequences on any process taking place over a social network .
we explore here the role of time on the emergence of cooperation in time - varying networks .
we consider two data sets describing the temporal patterns of human interactions at two different time scales .
the first data set has been collected during the mit reality mining experiment @xcite , and includes information about spatial proximity of a group of students , staff , and faculty members at the massachusetts institute of technology , over a period of six months .
the resulting time - dependent network has @xmath21 nodes and consists of a time - ordered sequence @xmath22 of @xmath23 graphs ( snapshots ) , each graph representing proximity interactions during a time interval of @xmath24 minutes .
remember that each graph @xmath25 accounts for all the instantaneous interactions taking place in the temporal interval @xmath26 $ ] . the second data set describes co - location patterns , over a period of four days , between the participants of the infocom06 conference @xcite .
in this case , the resulting time - dependent network has @xmath27 nodes , and contains a sequence of @xmath28 graphs obtained by detecting users co - location every @xmath29 minutes .
the frequency of social contacts is illustrated in fig .
[ fig1 ] ( panels a and b ) , where we report the number of active links at time @xmath30 , @xmath31 , as a function of time . in the mit reality mining data set ,
social activity exhibits daily and weekly periodicities , respectively due to home - work and working days - weekends cycles .
in addition to these rhythms , we notice a non - stationary behavior which is clearly visible when we plot the activity averaged over a 1-month moving window ( red line in panel a ) . in the infocom06 data set we observe a daily periodicity and a non - stationary trend which is due , in this case , to a decreasing social activity in the last days of the conference as seen by aggregating activity over 24 hours ( red line in panel b ) .
we also report in fig .
[ fig1 ] ( panels c and d ) the distributions @xmath32 of contact duration , @xmath33 , and of inter - contact time , @xmath34 ( i.e. the interval between two consecutive appearances of an edge ) .
as it is often the case for human dynamics @xcite , the distributions of contact duration and inter - contact time are heterogeneous .
for the mit data set , an active edge can persist up to an entire day , while inactive intervals can last over multiple days and weeks ; similar patterns are observed in the infocom06 data set , where some edges remain active up to one entire day and inter - contact times span almost the whole observation interval .
edge activity exhibits significant correlations over long periods of time .
in particular , the autocorrelation function of the time series of edge activity shows a slow decay , up to lags of @xmath35 hours for the mit data set , and of @xmath36 hours for infocom06 , after which the daily periodicity becomes dominant ( figure not reported ) . to simulate the game on a time - varying topology @xmath37 , we start from a random distribution of strategies , so that each individual initially behaves either as a cooperator or as a defector , with equal probability .
the simulation proceeds in _ rounds _ , where each round consists of a _ playing _ stage followed by a _ strategy update_. in the first stage , each agent plays with all her neighbors on the first graph of the sequence , namely on @xmath38 , and accumulates the payoff according to the matrix in eq . [ payoffs ] .
then the graph changes , and the agents employ the same strategies to play with all their neighbors in the second graph of the sequence , @xmath39 .
the new payoffs are summed to those obtained in the previous iteration .
the same procedure is then repeated @xmath40 times with @xmath40 such that @xmath41 is equal to a chosen interval @xmath42 , which is the _ strategy update interval_. at this point , the playing stage terminates and agent strategies are updated .
namely , each agent compares the net payoff accumulated during the previous @xmath40 time steps with that of one of her neighbors chosen at random . in particular , we adopt here the so - called _ fermi rule _
@xcite with a parameter @xmath43 ( see methods ) .
we have checked that the results are quite robust , and we obtain qualitatively similar outcomes for a wide range of @xmath44 .
after the agents have updated their strategy , their payoff is reset to @xmath45 and they start another round , during the subsequent time interval of length @xmath46 , as described above . to evaluate the degree of cooperation obtained for a given value of the strategy update interval @xmath42 and a pair of values @xmath47 , we compute the average fraction of cooperators @xmath48 : @xmath49 where @xmath50 is the number of cooperators found at time @xmath51 and @xmath52 is the total number of rounds played .
in general , we set @xmath52 large enough to guarantee that the system reaches a stationary state .
we have simulated the system using different values of @xmath53 .
notice that for smaller value of @xmath42 the time - scale of strategy update is comparable with that of the graph evolution , while when @xmath42 is equal to the entire observation period @xmath54 the game is effectively played on a static topology , namely the weighted aggregated graph corresponding to the whole observation interval .
we focus here on the top panels of fig .
[ fig2 ] and fig .
[ fig3 ] , where we show how the average fraction of cooperators depends on the parameters @xmath2 and @xmath3 and on the length @xmath42 of the strategy update interval .
we considered six values of @xmath42 for the mit reality mining data set , from @xmath55 hour up to the whole observation interval , and eight values for infocom06 , ranging from minutes up to the aggregate network . at first glance
, we notice that the rightmost diagrams in both figures , which correspond to @xmath56 , are in perfect agreement with the results about evolutionary games played on static topologies reported in the literature ( see _ e.g. _
@xcite ) . if we look at the cooperation diagrams obtained by increasing the value of @xmath42 in the original sequences of graphs ( top panels of fig .
[ fig2 ] and fig .
[ fig3 ] ) , we notice that , for any pair @xmath47 , a larger update interval corresponds to a higher fraction of cooperators . in particular , for mit reality mining ( fig .
[ fig2 ] ) the fraction of cooperators increases up until @xmath57 months , after which the cooperation diagram is practically indistinguishable from that obtained on the static aggregated graph . for infocom06 , instead , a strategy update interval larger than @xmath58 hours already produces a cooperation diagram similar to that obtained in the aggregated graph .
these results indicate that defectors actually take advantage from the volatility of edges , and that cooperation can emerge only if the strategy update interval is large enough .
we will check this issue later in section [ secd ] . as we pointed out above ,
edge activation patterns show non - trivial correlations . to highlight the effects of temporal correlations and of periodicity in the appearance of links in the real data sets , we have simulated the games also on randomized time - varying graphs and on synthetic networks generated through the activity - driven model @xcite .
the results for randomized graphs and activity - driven graphs are reported , respectively , in the middle and in the bottom panels of fig .
[ fig2 ] and fig .
[ fig3 ] .
randomized time - varying graphs are obtained by uniformly reshuffling the original sequences of snapshots . in this case the frequency of each pairwise contact is preserved equal to that of the original data set .
however , the temporal correlations of these contacts , namely the persistence of an edge during consecutive time snapshots , are completely washed out . as expected , for @xmath56 the cooperation diagrams obtained on the reshuffled sequences ( middle rightmost panels of fig .
[ fig2 ] and fig .
[ fig3 ] ) are identical to those obtained on the corresponding original data sets ( top rightmost panels ) .
in fact , when @xmath56 each agent plays with all the contacts she has seen in the whole observation interval , with the corresponding weights , before updating her strategy . since , in this particular case the frequencies of contacts are the only ingredients responsible for the emergence of cooperation , and such frequencies are the same both in the original data set and in the reshuffled version , the results of the original and randomized graphs have to be identical .
conversely , for smaller values of @xmath42 , the importance of the temporal correlations of each pairwise contact becomes clear since the cooperation diagrams for randomized and original networks are very different in both mit reality mining and infocom06 .
in fact for the randomized graphs , the cooperation levels at @xmath55 week and @xmath55 hour , respectively , become comparable to those obtained for @xmath56 .
this points out that destroying the temporal correlations of each pairwise contact enhances cooperation .
* cooperation level and size of the giant component*. overall cooperation level @xmath59 and average size of the giant component @xmath60 as a function of the aggregation interval @xmath42 for mit reality mining ( top panels ) and infocom06 ( bottom panels ) .
blue circles correspond to the original data , black squares to the reshuffled networks and red triangles to the activity - driven model .
the shades indicate the standard deviation of @xmath60 across the sequence of graphs for each value of @xmath42 .
notice that the typical size of the giant component at time - scale @xmath42 correlates quite well with the observed cooperation level at the same time - scale.,title="fig : " ] + * cooperation level and size of the giant component*. overall cooperation level @xmath59 and average size of the giant component @xmath60 as a function of the aggregation interval @xmath42 for mit reality mining ( top panels ) and infocom06 ( bottom panels ) .
blue circles correspond to the original data , black squares to the reshuffled networks and red triangles to the activity - driven model .
the shades indicate the standard deviation of @xmath60 across the sequence of graphs for each value of @xmath42 .
notice that the typical size of the giant component at time - scale @xmath42 correlates quite well with the observed cooperation level at the same time - scale.,title="fig : " ] little differences are observed ( compared to the case of randomized graphs ) when using activity - driven synthetic networks ( results shown in the bottom panels of fig . [ fig2 ] and fig . [ fig3 ] ) . in this case not only temporal correlations are washed out , but also the microscopic structure of each snapshot is replaced by a graph having a similar density of links .
this rewiring distributes links more heterogeneously than in the original and the randomized sequences ( see methods for details ) .
the cooperation diagrams of activity - driven networks show a further increase of the cooperation levels for even smaller values of the strategy update interval , @xmath42 , than in the case of random graphs .
namely , for @xmath55 day in reality mining ( fig . [ fig2 ] ) and for @xmath61 minutes in infocom ( fig .
[ fig3 ] ) we already recover the cooperation levels of @xmath62 . the results reported in fig . [ fig2 ] and fig .
[ fig3 ] suggest that the ordering , persistence and distribution of edges over consecutive time - window are all fundamental ingredients for the success of cooperation . in general
, a small value of @xmath42 in the original data sets corresponds to playing the game on a sparse graph , possibly comprising a number of small components , in which agents are connected to a small neighborhood that persists rather unaltered over consecutive time windows .
the small size of the isolated clusters and the persistence of the connections within them allow defectors to spread their strategy efficiently . in the following
, we will analyze this hypothesis by characterizing the structural patterns of the original time varying graph , its randomized version and the activity - driven synthetic network . as a function of the strategy update interval @xmath42 on real data sets ( blue dots ) .
top ( bottom ) panel refers to reality mining ( infocom ) .
black squares correspond to randomly reshuffled sequences and red triangles to activity - driven synthetic networks . in both datasets
we notice that , for small values of @xmath42 , real data display an higher value of clustering ( persistence ) than synthetic cases followed by a transition value of @xmath42 above which we observe a rapid increase in the clustering of synthetic cases such that the previous situation is inverted.,title="fig : " ] + as a function of the strategy update interval @xmath42 on real data sets ( blue dots ) .
top ( bottom ) panel refers to reality mining ( infocom ) .
black squares correspond to randomly reshuffled sequences and red triangles to activity - driven synthetic networks . in both datasets
we notice that , for small values of @xmath42 , real data display an higher value of clustering ( persistence ) than synthetic cases followed by a transition value of @xmath42 above which we observe a rapid increase in the clustering of synthetic cases such that the previous situation is inverted.,title="fig : " ] + in order to understand the dependence of cooperation on the strategy update interval @xmath42 , we plot in fig .
[ fig : giant ] the average fraction , @xmath60 , of the nodes found in the giant component of the graphs as a function of @xmath42 , for the original data sets and for the reshuffled and synthetic sequences of snapshots . in general , for a given value of @xmath42 , the giant component of graphs corresponding to randomized sequences or to the activity - driven model is larger than that of the graphs in the original ordering .
the break down of temporal correlations between consecutive time snapshots in randomized and activity - driven networks produces an increase in the number of ties between different agents of the population even for small values of @xmath42 .
in addition , the more homogeneous distribution of links within the snapshots of the activity - driven network further increases the mixing of the agents and thus enlarges the size of the giant component compared to the randomized graphs . in fig .
[ fig : giant ] we also plot the overall level of cooperation at a given aggregation scale @xmath42 , @xmath63 , defined as : @xmath64 notice that @xmath59 is divided by the value @xmath65 corresponding to the whole observation interval , so that @xmath66 $ ] .
the value of @xmath42 at which @xmath60 is comparable with the number of nodes @xmath67 , i.e. when @xmath68 , coincides with the value of @xmath42 at which the cooperation diagram becomes indistinguishable from that obtained for the aggregate network , @xmath69 , for both the original and the reshuffled sequences of snapshots .
this result confirms that the size of the giant connected component , of the graph corresponding to a given aggregation interval , plays a central role in determining the level of cooperation sustainable by the system , in agreement with the experiments discussed in @xcite for the case of static complex networks .
it is also interesting to investigate the role of edge correlations on the observed cooperation level . to this aim
, we analyze the temporal clustering ( see methods ) , which captures the average tendency of edges to persist over time . in fig .
[ fig5 ] we plot the evolution of the temporal clustering as a function of the strategy update interval @xmath42 .
the results reveal clearly that , for small values @xmath53 , the persistence of ties in the two original data sets is larger than in the randomized and the activity - driven graphs . in this regime
the average giant component is small in both the real and randomized cases , thus pointing out that the temporally connected components are composed of small clusters .
however , the larger temporal clustering observed for small @xmath42 in the original data implies that the node composition of these small components changes very slowly compared to the faster mixing observed in the random data sets .
these are then the two ingredients depressing the cooperation levels in the original data as compared to the random cases : the size of the giant component and how much such components change even at fixed size . as further confirmation
, we notice that link persistence grows in a similar way as @xmath42 increases in randomized and activity - driven networks .
this growth points out that the randomization of snapshots in one null model and the redistribution of links in the other one make the ties more stable as @xmath42 increases .
instead , the results found on the original data sets suggest ( in particular for the case of infocom ) that ties are rather volatile , being active for a number of consecutive snapshots and then inactive for a large time interval . in randomized and activity - driven graphs the stabilization of ties together with the fast increase in the size of the giant component
make the resulting time - varying graph much more similar to static networks , thus improving the survival of cooperation , as compared to the volatile and strongly fragmented scenario of the real time - varying graphs .
although the impact of network topology on the onset and persistence of cooperation has been extensively studied in the last years , the recent availability of data sets with time - resolved information about social interactions allows a deeper investigation of the impact on evolutionary dynamics of time - evolving social structures .
here we have addressed two crucial questions : does the interplay between graph evolution and strategy update affect the classical results about the enhancement of cooperation driven by network reciprocity ? and what is the role of the time - correlations of temporal networks in the evolution of cooperation ?
the results of the simulations confirm that , for all the four social dilemmas studied in this work , cooperation is seriously hindered when ( i ) agent strategy is updated too frequently with respect to the typical time - scale of agent interaction and ( ii ) real - world timecorrelations are present .
this phenomenon is a consequence of the relatively small size of the giant component of the graphs obtained at small aggregation intervals .
however , when the temporal sequence of social contacts is replaced by time - varying networks preserving the original activity attributes of links or nodes but breaking the original temporal correlations , the structural patterns of the network at a given time - scale of strategy update changes dramatically from those observed in real data . as a consequence , the effects of temporal resolution over cooperation are smoothed and , by breaking the real temporal correlations of social contacts , cooperation can emerge and persist also for moderately small strategy update frequencies .
this result highlights that both the interplay of strategy update and graph evolution and the presence of temporal correlations , such as edge persistence and recurrence , seem to have fundamental effects on the emergence of cooperation .
our findings suggest that the frequency at which the connectivity of a given system are sampled has to be carefully chosen , according with the typical time - scale of the social interaction dynamics .
for instance , as stock brokers might decide to change strategy after just a couple of interactions , other processes like trust formation in business or collaboration networks are likely to be better described as the result of multiple subsequent interactions .
these conclusions are also supported by the results of a recent paper of ribeiro _ et al . _
@xcite in which the effects of temporal aggregation interval @xmath42 in the behavior of random walks are studied .
also , the fundamental role played by the real - data time correlations in dynamical processes on the graph calls for more models of temporal networks and for a better understanding of their nature . in a nutshell ,
the arguments indicating network reciprocity as the social promoter of cooperator have to be revisited when considering time - varying graphs . in particular
, one should always bear in mind that both the over- and the under - sampling of time - evolving social graph and the use of the finest / coarsest temporal resolution could substantially bias the results of a game - theoretic model played on the corresponding network .
these results pave the way to a more detailed investigation of social dilemmas in systems where not only structural but also temporal correlations are incorporated in the interaction maps .
the data set describes proximity interactions collected through the use of bluetooth - enabled phones @xcite .
the phones were distributed to a group of 100 users , composed by 75 mit media laboratory students and 25 faculty members .
each device had a unique tag and was able to detect the presence and identity of other devices within a range of 5 - 10 meters .
the interactions , intended as proximity of devices , were recorded over a period of about six months .
in addition to the interaction data , the original dataset included also information regarding call logs , other bluetooth devices within detection range , the cell tower to which the phones was connected and information about phone usage and status . here , we consider only the contact network data , ignoring any other contextual metadata .
the resulting time - varying network is an ordered sequence of 41291 graphs , each having n=100 nodes .
each graph corresponds to a proximity scan taken every 5 minutes .
an edge between two nodes indicates that the two corresponding devices were within detection range of each other during that interval .
we refer to such links as _
active_. during the entire recorded period ,
2114 different edges have been detected as active , at least once .
this corresponds to the aggregate graph having a large average node degree @xmath70 .
however , this is an artefact of the aggregation ; the single snapshots tend to be very sparse , usually containing between 100 and 200 active edges . the data set consists of proximity measurements collected during the ieee infocom06 conference held in a hotel in barcelona in 2006 @xcite .
a sample of @xmath71 participants from a range of different companies and institutions were chosen and equipped with a portable bluetooth device , intel imote , able to detect similar devices nearby .
area `` inquiries '' were performed by the devices every 2 minutes , with a random delay or anticipation of 20 seconds .
the delay / anticipation mechanism was implemented in order to avoid synchronous measurements , because , while actively sweeping the area , devices could not be detected by other devices . a total number of 2730 distinct edges
were recorded as active at least once in the observation interval , while the number of edges active at a given time is significantly lower , varying between 0 and 200 , depending on the time of the day .
the fermi rule consists in the following updating strategy .
a player @xmath15 chooses one of her neighbors @xmath18 at random and copies the strategy of @xmath18 with a probability : @xmath72 where @xmath20 is the difference between the payoffs of the two players , and @xmath44 is a parameter controlling the smoothness of the transition from @xmath73 for small values of @xmath20 , to @xmath74 for large values of @xmath20 .
notice that for @xmath75 we obtain @xmath76 regardless of the value of @xmath20 , which effectively corresponds to a random strategy update . on the other hand , when @xmath77 then @xmath78 , being @xmath79 the heaviside step function . in this limit
, the strategy update is driven only by the ordering of the payoff values . the activity - driven model , introduced in ref .
@xcite , is a simple model to generate time - varying graphs starting from the empirical observation of the activity of each node , in terms of number of contacts established per unit time .
given a characteristic time - window @xmath42 , one measures the activity potential @xmath80 of each agent @xmath15 , defined as the total number of interactions ( edges ) established by @xmath15 in a time - window of length @xmath42 divided by the total number of interactions established on average by all agents in the same time interval .
then , each agent is assigned an activity @xmath81 , which is the probability per unit time to create a new connection or contact with any another agent @xmath18 .
the coefficient @xmath82 is a rescaling factor , whose value is appropriately set in order to ensure that the total number of active nodes per unit time in the system is equal to @xmath83 , where @xmath67 is the total number of agents .
notice that @xmath82 effectively determines the average number of connections in a temporal snapshot whose length corresponds to the resolution of the original data set .
the model works as follows . at each time
@xmath30 the graph @xmath84 starts with @xmath67 disconnected nodes .
then , each node @xmath15 becomes active with probability @xmath85 and connects to @xmath86 other randomly selected nodes . at the following time - step ,
all the connections in @xmath84 are deleted , and a new snapshot is sampled .
notice that time - varying graphs constructed through the activity - driven model preserve the average degree of nodes in each snapshot , but impose that connections have , on average , a duration equal to @xmath42 , effectively washing out any temporal correlation among edges .
several metrics have been lately proposed to measure the tendency of the edges of a time - varying graph to persist over time .
one of the most widely used is the unweighted temporal clustering , introduced in ref .
@xcite , which for a node @xmath15 of a time - varying graph is defined as : @xmath87 where @xmath88 are the elements the adjacency matrix of the time - varying graph at snapshot @xmath30 , @xmath89 is the total number of edges incident on node @xmath15 at snapshot @xmath30 and @xmath3 is the duration of the whole observation interval .
notice that @xmath90 takes values in @xmath91 $ ] . in general ,
a higher value of @xmath90 is obtained when the interactions of node @xmath15 persist longer in time , while @xmath90 tends to zero if the interactions of @xmath15 are highly volatile .
if each snapshot of the time - varying graph is a weighted network , where the weight @xmath92 represents the strength if the interaction between node @xmath15 and node @xmath18 at time @xmath30 , we can define a weighted version of the temporal clustering coefficient as follows : finally , if we focus more on the persistence of interaction strength across subsequent network snapshots , we can define the extremal temporal clustering as : @xmath94 where by considering the minimum between @xmath95 and @xmath96 one can distinguish between persistent interactions having constant strength over time and those interactions having more volatile strength . as in our case social interactions are seen to be highly volatile in real data sets , the extremal version of the temporal clustering seems to be the best choice to unveil the persistence of social ties at short time scales .
this work was supported by the eu lasagne project , contract no.318132 ( strep ) , by the eu multiplex project , contract no.317532 ( strep ) , by the spanish mineco under projects mtm2009 - 13848 and fis2011 - 25167 ( co - financed by feder funds ) , by the comunidad de aragn ( grupo fenol ) and by the italian to61 infn project .
is supported by spanish mineco through the ramn y cajal program .
is supported by the fet project
topdrim " ( ist-318121 ) .
is supported by the james s. mcdonnell foundation .
l. isella , m. romano , a. barrat , c. cattuto , v. colizza v , _
_ , close encounters in a pediatric ward : measuring face - to - face proximity and mixing patterns with wearable sensors . "
plos one * 6 * , e17144 ( 2011 ) .
j. tang , m. musolesi , c. mascolo , v. latora , and v. nicosia , analysing information flows and key mediators through temporal centrality metrics " proceedings of the 3rd acm workshop on social network systems ( sns10 ) , ( 2010 ) . | cooperation among unrelated individuals is frequently observed in social groups when their members join efforts and resources to obtain a shared benefit which is unachievable by single ones . however , understanding why cooperation arises despite the natural tendency of individuals towards selfish behavior is still an open problem and represents one of the most fascinating challenges in evolutionary dynamics .
very recently , the structural characterization of the networks upon which social interactions take place has shed some light on the mechanisms by which cooperative behavior emerges and eventually overcome the individual temptation to defect .
in particular , it has been found that the heterogeneity in the number of social ties and the presence of tightly - knit communities lead to a significant increase of cooperation as compared with the unstructured and homogeneous connection patterns considered in classical evolutionary dynamics .
here we investigate the role of social ties dynamics for the emergence of cooperation in a family of social dilemmas .
social interactions are in fact intrinsically dynamic , fluctuating and intermittent over time , and can be represented by time - varying networks , that is graphs where connections between nodes appear , disappear , or are rewired over time . by considering two experimental data sets of human interactions with detailed time information , we show that the temporal dynamics of social ties has a dramatic impact on the evolution of cooperation : the dynamics of pairwise interactions favor selfish behavior .
_ popular abstract.- _ * why do animals ( including humans ) cooperate even when selfish actions provide larger benefits ?
this question has challenged the evolutionary theory during decades .
the success of cooperation is essential to humankind and ubiquitous , no matter the cultural and religious traits of particular populations .
scientists have pointed out in the past a series of possible mechanisms that could favor cooperation between humans , as for example the peculiar way of establishing social relations which assumes the form of a complex network . basically , the structural attributes of these networks , such as the presence of a few individuals that have a large number of social ties , promote cooperation and discourage the imitation of free - riders . however , social interactions are inherently dynamic and changing over time , an issue which has been usually disregarded in the study of cooperation on networks .
we study evolutionary models in time - varying graphs and show that the volatility of social relations tends to decrease cooperation in respect to static graphs .
our results thus point out that the time - varying nature of social ties can not be neglected and that the relative speed of graph evolution and strategy update is a crucial ingredient governing the evolutionary dynamics of social networks , having as much influence as the structural organization . * |
the flow properties of granular materials puzzle engineers and physicists alike . despite numerous experimental studies of heap flows , avalanches , and chute flows ,
there still does not exist a general consensus on the nature of _ typical _ granular flow .
early systematic experiments in inclined plane geometries were concerned with flow rates of granular material down chutes @xcite , usually with smooth bases .
large regions of plug flow were observed , i.e. the whole bulk of the material slid down the plane at a velocity determined by the input mass rate and inclination angle of the chute .
kinetic theories for the inclined plane problem can explain some aspects of this flow behaviour for nearly elastic particles on smooth chutes @xcite .
other studies have determined structural features in the flow in two dimensional ( 2d ) experiments @xcite and simulations @xcite .
still , there has not been the same success in predicting flows down rough inclines for inelastic , frictional particles .
detailed measuring techniques have been developed recently to capture various flow properties of granular material down rough inclined planes @xcite and in heap flow geometries @xcite . accompanying these more thorough experiments ,
further theoretical treatments have also emerged @xcite , together with full three dimensional ( 3d ) computer simulations @xcite . despite this growth in literature , there appears to be a disturbing feature of these studies : reports on quantities such as velocity profiles differ quite dramatically . while some reports are agreement with the pioneering results of bagnold @xcite , others are not . in 1954 , bagnold @xcite proposed a momentum transfer picture to describe his experimental results on inertial granular flow
. this simple 2d model describes the collisional process between particle layers during flow .
the resulting relations predict a shear stress @xmath11 that is proportional to the square of the strain - rate , @xmath12 where the strain - rate is defined as the derivative of the velocity in the direction of flow ( @xmath13 ) with respect to depth ( @xmath14 ) , @xmath15 .
typically , in local , newtonian , hydrodynamic treatments of granular dynamics @xcite , the relationship @xmath16 is used , combined with the kinetic term @xmath17 , where @xmath18 is a viscosity and @xmath19 is the granular temperature ( to be discussed in section [ depth ] ) . since the strain - rate is also proportional to the square root of the granular temperature @xmath20 , eq .
[ chute3_eq1 ] is then recovered .
bagnold s argument , applied to the case of bulk granular flow , is predicated on a constant density depth - profile with no - slip condition at the base .
this leads to a velocity depth - profile of the form , @xmath21 , \label{chute3_eq2}\ ] ] for a pile of total height @xmath8 , where @xmath14 is measured from the bottom of the pile . because bagnold scaling comes naturally from dimensional analysis in the absence of intrinsic time scales
, one might expect that most studies would observe this kind of rheology .
however , results depend strongly on the experimental procedure ; profiles of this form have only been reported in a few instances @xcite .
one method to study gravity - driven flows is to construct a wedge - shaped static pile that evolves to produce an inclined surface close to the angle of repose of the material @xcite , fig .
[ chute3_fig1](a ) .
we term these _ heap flows_. here , the inclination of the surface is determined by material properties . to study flow ,
material is supplied via a hopper - feeder mechanism at the high end of the pile and flow is induced at the free surface . in this case , although the flow rate may be varied by changing the discharge rate of the hopper , the system typically flows only in a thin surface layer , indicated by the shaded region in fig .
[ chute3_fig1](a ) , which is of order 10 particles diameters in depth .
in one such experiment , lemieux and durian @xcite observed avalanche behavior for low particle flux , characterized by sporadic downstream movement at the free surface . on increasing the particle flux , a velocity profile with depth
was observed to be approximately exponential in the thin surface layer .
more recently , utilizing a similar experimental geometry , komatsu et al .
@xcite found that the ` stationary ' bulk material beneath the surface flow actually undergoes very slow , ` creeping ' motion with a well - defined exponential velocity profile at longer times .
another method uses inclined planes to study flows and avalanches , typically with rough beds designed to reduce slip at the base , thereby avoiding plug or unstable flow @xcite .
we term these _ chute flows_. roughness is introduced onto the chute by gluing particles to the chute base @xcite , or by attaching a frictional material to the base , such as the velvet cloth used by daerr and douady @xcite .
chute flow studies of two and three dimensional systems with rough bases report depth - averaged velocities with both linear @xcite and non - linear dependence @xcite on height . the geometry for these experiments is shown in fig .
[ chute3_fig1](b ) .
the material is contained in a reservoir at one end of a plane inclined at a variable angle @xmath0 with respect to the horizontal .
after the gate to the reservoir is opened , material flows down the incline .
steady flow occurs only at high enough angle ; while flows come to rest quickly on the rough bed for small inclination angles @xcite . increasing the gate opening leads to thicker flows down the incline :
there are no dead zones along the incline and all the material flows down the chute .
this allows one to independently control the inclination angle and material flux of the pile , in contrast to the heap experiments . using a range of inclination angles and flow depths ,
pouliquen @xcite measured the depth - averaged flow velocity of material for piles of different thickness .
the velocity in the direction of flow scaled with height of the pile as @xmath22 , which has a different form from the heap flow results reported by lemieux and durian @xcite and azanza et al .
however , ancey s comprehensive experimental investigation @xcite demonstrated that during the course of one experiment , a range of flow behaviors is obtained that depends , among other parameters , on the flow rate of the material .
ancey s results suggest that the different flow profiles observed are not necessarily in contradiction , but may be a consequence of the conditions under which the measurements are made . in our earlier studies
@xcite , we simulated the rheology of dense , cohesionless , frictional spheres in inclined plane geometries .
employing periodic boundary conditions in the flow and vorticity directions , our simulations can be considered equivalent to the experimental set up in fig .
[ chute3_fig1](b ) , for an infinitely long chute far away from the material reservoir and side walls , with a continuous supply of material .
our previous study was concerned with the flow of thick piles on rough bases over a range of tilt angles where we found bagnold scaling in stable steady flow , for a range of parameters ( e.g. particle friction and inelasticity , and inclination angles ) . to investigate the discrepancies between the various experimental results previously mentioned ,
we have extended our previous computational studies of granular chute flows to cover a wide range of pile thickness , from very thin to very thick piles , both near and far from the angle of repose @xmath1 .
we show here that even thick piles experience a crossover in behavior from bagnold rheology to linear velocity profiles , and also avalanche - like flow near the angle of repose .
we therefore refine our definition of stable , steady state flow : we report on three distinct steady state flows observed in our simulations which match the experimentally observed flows reported by lemieux and durian @xcite ( avalanches ) , ancey @xcite ( linear velocity profiles ) , and pouliquen @xcite ( bagnold scaling ) .
our main conclusion is that the observed behavior is strongly dependent on how close the system is to jamming or how close @xmath0 is to @xmath1 . in the next section
we briefly describe the simulation procedure . in section [ chute3_results ]
we present the results of our simulations , in particular defining the phase space of our study and describing transitions between the flow properties of thick and thin piles both near and far from the angle of repose .
we also present results for our analyses on the initiation and cessation of flow in section [ chute3_initiate ] . in section [ chute3_conc ]
we summarize the main conclusions of this work .
here we briefly introduce the simulation model ; a thorough description of the technique is available elsewhere @xcite .
we model @xmath23 cohesionless , frictional , and inelastic spheres of diameter @xmath24 and mass @xmath25 flowing on a fixed , rough bed of area @xmath26 with a free top surface .
although most of our results are for monodisperse systems , we have verified that our results hold for systems with small polydispersity @xcite .
we employed a 3d contact force model that captures the major features of granular particles : interactions between spheres are modeled as hertzian contacts with _
static friction_. the implementation of the contact forces , both the normal forces and the shear ( friction ) tangential forces , is essentially a reduced version of that employed by walton and braun @xcite , developed earlier by cundall and strack @xcite .
normal and tangential forces acting on particle @xmath27 due to contact with particle @xmath28 are given by @xmath29 where @xmath30 is the particle - particle overlap , @xmath31 , is the unit normal along the line of centers , for particle @xmath27 situated at @xmath32 and @xmath28 at @xmath33 .
@xmath34 are relative velocities in the normal and tangential directions respectively , and @xmath35 is the elastic tangential displacement that is set to zero at the initiation of a contact .
@xmath36 is the effective mass and @xmath37 and @xmath38 are elastic and viscoelastic constants respectively .
thus in a gravitational field * g * , the force on particle @xmath27 arises from newton s second law for the translational and rotational degrees of freedom , @xmath39 and the force on particle @xmath28 is determined from newton s third law . as shown in fig .
[ chute3_fig2 ] , we define flow to be parallel to the @xmath13-direction , @xmath14 the height normal to the base , and @xmath40 , the vorticity axis , is normal to the @xmath41-plane .
values for the friction coefficient @xmath42 , and the elastic and viscoelastic coefficients ( which are related to the moduli of the material ) are chosen to be characteristic of real materials .
for all simulations in this study we set @xmath43 , @xmath44 , @xmath45 , @xmath46 , and @xmath47 @xcite . the time - step @xmath48 , where @xmath49 . for a millimeter
sized particle @xmath50 .
all simulations were run for at least @xmath51 before any quantities were measured .
figure [ chute3_fig2 ] shows snapshots of flowing systems where the area of the chute base @xmath52 , and we vary the total number of particles @xmath23 . the fixed chute bed is made from a random close packing of particles with the same diameter as those in the bulk , mimicking gluing particles onto a chute base as in ref .
@xcite , for example . in section [ depth ]
we kept @xmath53 fixed . in section [ chute3_initiate ]
, we used a base with dimensions @xmath54 .
we have shown @xcite that results for thick piles are robust with respect to the dimensions of the chute base .
the height of a static pile @xmath55 depends on the density of the static packing @xmath56 , or equivalently the volume fraction @xmath57 @xcite ) which is defined in the hard sphere limit ( where the subscripts denote the static values ) . in the flowing state ,
the actual height of the pile increases with inclination angle due to dilation effects of the flow .
however , the dependence of @xmath8 , and therefore @xmath58 , on @xmath0 is weak .
we find that the computed bulk volume fractions vary over a small range , @xmath59 , for steady state flows for the particular parameters chosen in this paper @xcite .
thus for the simulations @xmath60 : for @xmath61 , @xmath62 . , that is inclined at an angle @xmath63 to the horizontal , where ( a ) @xmath64 , ( b ) @xmath65 , ( c ) @xmath66 , and ( d ) @xmath67 .
the fixed rough base is denoted by the black particles , the flowing particles are grey , and the frame is a guide to the eye fixed at the height @xmath68 of the @xmath64 system to provide a comparison between the different flow heights.,title="fig:",width=226 ] , that is inclined at an angle @xmath63 to the horizontal , where ( a ) @xmath64 , ( b ) @xmath65 , ( c ) @xmath66 , and ( d ) @xmath67 .
the fixed rough base is denoted by the black particles , the flowing particles are grey , and the frame is a guide to the eye fixed at the height @xmath68 of the @xmath64 system to provide a comparison between the different flow heights.,title="fig:",width=226 ] + ( a)(b ) + , that is inclined at an angle @xmath63 to the horizontal , where ( a ) @xmath64 , ( b ) @xmath65 , ( c ) @xmath66 , and ( d ) @xmath67 .
the fixed rough base is denoted by the black particles , the flowing particles are grey , and the frame is a guide to the eye fixed at the height @xmath68 of the @xmath64 system to provide a comparison between the different flow heights.,title="fig:",width=226 ] , that is inclined at an angle @xmath63 to the horizontal , where ( a ) @xmath64 , ( b ) @xmath65 , ( c ) @xmath66 , and ( d ) @xmath67 .
the fixed rough base is denoted by the black particles , the flowing particles are grey , and the frame is a guide to the eye fixed at the height @xmath68 of the @xmath64 system to provide a comparison between the different flow heights.,title="fig:",width=226 ] + ( c)(d ) +
in our previous studies @xcite , we roughly determined the range of parameters where steady state flow existed for chute flows , focusing mainly on steady state flow for thick piles away from @xmath1 .
we found that @xmath1 is approximately independent of height for piles of sufficient depth . in this work
we perform a more thorough analysis of the entire region of the @xmath69 phase space .
we determine several different flow regimes within the stable , steady state region which are qualitatively identical to several experimental observations in inclined plane geometries , on avalanche dynamics @xcite and rheology @xcite . although the exact location of the boundaries of the phase picture shown in fig .
[ chute3_fig3 ] are dependent on specific particle parameters , such as the friction and damping coefficients , the features we report here are general to systems that exhibit steady state flow .
the overall phase behaviour is determined by the height of the pile @xmath8 and the angle of inclination @xmath0 , as characterised in 2d by pouliquen and renault @xcite .
a prominent feature of that study is the existence of a _ flow line _ , later observed in 3d experiments by pouliquen @xcite and also in studies of avalanches @xcite .
the flow line defines the angle of repose as a function of height of material flowing @xmath70 .
one can equivalently refer to this in terms of the conjugate variable @xmath71 @xcite : @xmath70 is the angle at which a pile of thickness @xmath8 will cease to flow , and conversely , @xmath71 gives the maximum thickness of material that will cease to flow at an inclination angle @xmath0 .
various theoretical treatments have also successfully predicted the existence of the flow line @xcite .
experimental @xcite and numerical @xcite studies have also identified distinct sub - regions within the stable flow regime of which more will be discussed below . in order to quantify this boundary with precision for our system
, we performed a number of simulations to generate the @xmath69 phase diagram shown in fig .
[ chute3_fig3 ] .
initially , a static packing with a volume fraction @xmath72 , at @xmath73 @xcite , is tilted until flow is observed . the _ maximum angle of stability _
@xmath2 is defined as the angle at which flow is initiated and largely depends on construction history of the initial static packing , such as the roughness of the supporting bed and the packing fraction of the initial configuration @xcite .
the angle of repose @xmath1 is defined as the inclination angle at which the system subsequently jams and flow ceases upon reducing @xmath0 back down .
phase diagram for granular chute flow , where @xmath8 is the height of the pile .
the filled squares are the simulation data for @xmath70 and we indicate the _ flow line _ as a fit to eq .
[ chute3_eq7 ] .
the open squares are @xmath74 .
solid circles are @xmath75 , the angle at which we observe the onset of unstable flow .
route a corresponds to a path taken by keeping the angle fixed and reducing the system size @xmath23 , whereas route b has @xmath23 fixed and decreasing inclination .
these routes are discussed in section [ depth].,title="fig:",width=302 ] + our initial packing was generated by taking a flowing system and gradually reducing the inclination down to @xmath76 and then allowing the system to relax until it became completely static .
experimental studies on avalanches have shown that @xmath2 can fluctuate over a range of @xmath77 for different packings with the same height because of construction history effects @xcite . we have also observed avalanche - like dynamics in our simulations .
we characterise avalanches as distinct from regular or continuous flow by observing the time evolution of bulk quantities of the system such as bulk kinetic energy or stress . for continuous flow , the stress or bulk energy will fluctuate about a well - defined mean value , with fluctuations typically less than @xmath78 of the mean . in the case of avalanches ,
the stress and energy fluctuate about a mean value with o(1 ) fluctuations , indicative of sporadic flow that mainly occurs at the free surface .
we also checked any possible system size dependence . to compare with our smaller systems , we studied a system with @xmath79 and @xmath80 ( i.e. approximately @xmath81 high on a base with an area four times our standard system size ) .
we again observed avalanche - like behaviour for the big system , although we can not rule out that in the limit of infinite system size , avalanches might not be observed . from our extensive study
, we were able to determine the following general features of flow down inclined planes : there are three principal regions , corresponding to _ no flow _ , _ stable flow _ , and _
unstable flow_. for a pile of given thickness @xmath8 , and fixed microscopic interaction parameters , these three regions are separated by two angles : @xmath1 , the angle of repose , and @xmath3 , the maximum angle for which we observe stable flow . no flow : : @xmath82 : the system does not flow .
starting from @xmath73 , an inclination of @xmath83 is required to initiate flow . as @xmath2 is approached , a few particles at the free surface rearrange and tumble short distances before coming to rest , but this does not lead to bulk motion or flow @xcite .
@xmath1 is the angle at which flow ceases on reducing the inclination from a flowing state - it is the lowest angle at which flow is observed and depends on material properties ( friction and inelasticity ) @xcite , as well as the roughness of the supporting bed .
stable flow : : @xmath4 : bulk flow is observed ( after flow is initiated ) .
the bulk averaged kinetic energy of the system is constant with time and the entire pile is seen to be in motion while no slip is observed at the bottom with these rough boundary conditions .
the stable flow region can be divided into the following three regimes : + continuous flow i ; ; @xmath5 : away from the flow line , the rheology is well approximated by bagnold scaling , eqs .
( [ chute3_eq1 ] ) and ( [ chute3_eq2 ] ) .
continuous flow ii ; ; @xmath84 : close to the flow line , the rheology is not bagnold - like , rather the velocity profile is better approximated by a linear relationship with depth @xmath85 .
the upper boundary @xmath2 is also seen in experimental studies of avalanches and differentiates between different flow mechanisms , depending on whether one is above or below @xmath2 @xcite . avalanching flow ; ; @xmath86 : very close to the angle of repose , but just above it , regular continuous flow is not observed , rather in this regime , steady state involves slow , underlying creep motion of the entire pile combined with sporadic free surface motion down the incline . in the vicinity of @xmath1 , extremely large energy fluctuations are observed @xcite .
unstable flow : : @xmath87 : the development of a shear thinning layer at the bottom of the pile results in lift - off and unstable acceleration of the entire pile .
@xmath3 becomes approximately constant for @xmath88 .
the flow descriptions above capture the principal observations of several different experiments of gravity - driven flows @xcite . in particular , the transition from no flow to avalanche flow at @xmath89 is similar to the observations reported by lemieux and durian @xcite for low material flux .
the transition from avalanches to our continuous flow ii regime appears to resemble the observations reported by daerr and douady @xcite of avalanche studies for @xmath90 .
similar observations have been reported for rotating cylinder experiments @xcite .
pouliquen @xcite , later verified by daerr and douady @xcite , characterized the flow line - the dividing line between no flow and flow . using an empirical argument , pouliquen suggested the following relation : @xmath91\exp\left(-\frac{h}{l}\right ) , \label{chute3_eq7}\ ] ] where @xmath92 is the angle at which @xmath1 becomes independent of depth , @xmath93 is the largest value of @xmath1 , and @xmath94 is a characteristic length scale . in fig .
[ chute3_fig3 ] we fit eq .
[ chute3_eq7 ] to our simulation points and obtained the following fitting parameters ; @xmath95 , @xmath96 , and @xmath97 , where the values in parentheses are the data points from the simulations .
our simulations not only predict the phase behavior of chute flows , but also capture the subtle features emerging in recent experiments . in the following sections we present various depth profiles of density , velocity , and stress to better quantify the transitions between thin and thick piles .
we computed depth profiles as a function of @xmath14 for the packing fraction @xmath98 , velocity in the direction of flow @xmath6 , strain - rate @xmath99 , shear stress @xmath100 , and granular temperature @xmath19 ( see definition below ) , for a range of system sizes and inclination angles .
since our simulations employ periodic boundary conditions in the plane perpendicular to height , our simulations are not influenced by the effects of side walls and are therefore equivalent to measurements made along the centre line of a real chute flow , away from the edges .
it has been recognised that profiles taken at the side walls are significantly different from those made along the centre line @xcite .
because of the dependence of @xmath1 on @xmath8 , we approached @xmath1 in two ways : fixing the inclination angle of the chute and observing flow in systems of different heights until flow ceased for a pile of height @xmath71 ( route a ) .
alternatively , we held @xmath23 fixed and reduced the inclination angle of a flowing state down to @xmath70 ( route b ) .
* route a : * figures [ chute3_fig7]-[chute3_fig10 ] show various bulk profiles for route a , that is , systems at a fixed inclination angle of @xmath63 but of varying sizes , @xmath101 .
the density profiles in fig .
[ chute3_fig7 ] exhibit a constant density region throughout the bulk of the material which is a familiar observation for thicker systems @xcite .
, for systems of various size @xmath23 , varying in heights from @xmath102 to @xmath103 for @xmath63 : route a in fig .
[ chute3_fig3].,width=245 ] the flow velocity @xmath6 and strain - rate @xmath99 profiles shown in fig .
[ chute3_fig8 ] are for the same systems reported in fig .
[ chute3_fig7 ] . for moderate to thick piles , @xmath104 , the velocities and corresponding strain - rates scale with height .
however , for @xmath105 , there are significant differences .
the velocity profiles , top panel in fig .
[ chute3_fig8 ] , for the thinner piles at first become more linear with depth , then ultimately have a curvature opposite to that of the thicker piles .
this gradient transformation is more apparent from the strain - rate curves ( bottom panel fig .
[ chute3_fig8 ] ) : for thicker piles @xmath99 decreases from the chute base toward the free surface of the flowing pile , whereas in thinner piles there is a range of system sizes where @xmath106 throughout the pile , leading to a linear velocity profile .
for thinner systems still , @xmath99 increases near the free surface . .
note the depth is normalized by the respective height for clarity .
legend as in fig .
[ chute3_fig7].,width=245 ] in this work , we define the `` granular temperature '' @xmath19 to be the averaged velocity fluctuations about the mean bulk flow velocity at that height , @xmath107 .
the temperature in the flow direction is defined as @xmath108 .
figure [ chute3_fig9 ] shows the change in the profiles of @xmath109 from thick to thin piles , which is similar to the behavior of @xmath99 . from the simulations we find that for _ all sizes _ , @xmath110 although the data for thinner piles is considerably noisier .
, in the direction of flow as a function of @xmath14 for various system sizes for @xmath63 .
legend as in fig .
[ chute3_fig7].,width=245 ] similarly , the rheology plot in fig .
[ chute3_fig10 ] shows that thicker piles display bagnold rheology : @xmath111 , but there is a transition away from this as the piles become thinner . for thin piles , @xmath112 ,
the strain rate is approximately constant away from the free surface and the chute base . for the thinnest piles @xmath113 , the strain rate decreases with shear stress
this may come about from competing length scales in the system , i.e. the thickness of the flow determines the separation between the top free surface and the chute base .
some theoretical treatments based on these considerations have been used to explain the differences between observed velocity profiles in flowing piles @xcite .
: @xmath99 vs. @xmath100 .
legend as in fig .
[ chute3_fig7].,width=245 ] * route b : * figures [ chute3_fig11]-[chute3_fig14 ] show depth profiles for a system taken along route b in fig .
[ chute3_fig3 ] .
these are for a system of fixed size @xmath65 , where we vary the inclination angle down toward @xmath1 . in fig .
[ chute3_fig11 ] the density profiles for the system @xmath65 are shown for a range of inclination angles @xmath114 .
we see the constant density region , where the bulk density decreases with increasing angle due to dilatancy effects .
however , there are no qualitative differences between the behaviour at different angles . , for a system with a fixed number of particles , @xmath65 , over a range of inclination angles , route b in fig .
[ chute3_fig3].,width=245 ] for the velocity and strain - rate profiles shown in fig .
[ chute3_fig12 ] , differences do emerge , and these differences mimic the behavior observed along route a. away from the flow line , the velocity profiles scale as eq . [ chute3_eq2 ] , whereas for systems at lower angles the velocity profiles change to linear . @xmath99
changes likewise as does the granular temperature ( shown in fig .
[ chute3_fig13 ] ) .
( top panel ) and the corresponding strain - rate @xmath99 ( bottom panel ) , each normalized by their respective maximum values , for a system with a fixed number of particles , @xmath65 , over a range of inclination angles inclination angle .
the height is normalized by the system size .
legend as in fig .
[ chute3_fig11].,width=245 ] for a system with a fixed number of particles , @xmath65 , over a range of inclination angles .
legend as in fig .
[ chute3_fig11].,width=245 ] , over a range of inclination angles : @xmath99 normalized by @xmath115 vs. @xmath100 .
legend as in fig .
[ chute3_fig11].,width=245 ] in fig .
[ chute3_fig14 ] we plot the rheology curves for the different angles .
we have rescaled @xmath99 by @xmath116 , for convenience .
we find that for systems with @xmath117 , the rheology plots scale well .
however , for the system that is closest to @xmath1 , the flow behavior is very different , though similar to the differences observed along route a. therefore the two routes a and b indicated in fig .
[ chute3_fig3 ] offer complementary paths to similar states . because of the very nature of the dependence of @xmath1 on @xmath8 ,
the flow properties of systems at different angles and different heights show similarities provided they are taken equivalently close to @xmath1 .
pouliquen @xcite showed that flows of thick and thin piles approximately obey a simple scaling relation .
non - dimensionalizing @xmath118 , the velocity measured along the centre line at the top of the flowing pile by @xmath119 , where g is the acceleration due to gravity and @xmath8 is the measured height of the pile , one obtains the _ froude _ number @xmath120 . by plotting the froude number against @xmath121 , where @xmath71 is the height of a pile at which a system flowing at an angle @xmath0 would stop
, pouliquen was able to collapse his full data set of systems with different particle size and properties .
pouliquen proposed the simple scaling relation , @xmath122 where the coefficient @xmath123 was fit to the experimental data , resulting in an apparently universal value , @xmath124 .
the quantity @xmath125 encodes details of the particle roughness , the base roughness and all other intrinsic properties of a particular system .
accordingly , one need only measure @xmath125 of the system in question to determine the flow properties . in a similar vein , in fig .
[ chute3_fig15 ] we plot the flow velocity , taken from the top of the flowing pile , non - dimensionalised by @xmath119 for each height @xmath8 against @xmath121 .
the data agrees well with a linear relation , giving a coefficient @xmath126 , which is in good agreement with the experimental result determined by pouliquen @xcite , @xmath124 . as a function of @xmath121 for different inclinations .
a measure of the slope gives @xmath127 , compared with pouliquen s experimental determination of @xmath128.,width=302 ] the scaling relation eq .
[ chute3_eq9 ] although simple , appears to be quite robust .
pouliquen @xcite and also andreotti et al .
@xcite , have taken eqs .
[ chute3_eq9 ] and [ chute3_eq7 ] to predict phase behaviour and flow properties .
their resulting analyses of granular fronts captures many of the main features of granular flows .
however , as yet , there is no underlying explanation for these seemingly simple relationships .
we have also studied the initiation and cessation of flow , something difficult to probe experimentally . for these studies we used large systems , @xmath129 and @xmath130 , providing better resolution for these processes .
in addition , any size effects due to thin samples do not complicate our observations .
we have previously studied initiation of flow from a crystalline static state in 2d @xcite .
after such a system is tilted , a disordering front propagates upwards from the bottom of the pile , melting the ordered structure , until the entire pile is disordered and flows in its steady state form .
although the starting state in @xmath131 need not be ordered , we find that flow initiation in @xmath131 also undergoes a similar bottom - up process .
* initiation * was studied for a disordered , static packing of our larger system that was tilted from @xmath73 to @xmath132 .
see fig .
[ chute3_fig5 ] . at the instant of flow initiation , the system slips at the bottom boundary and begins to flow as a plug as indicated in fig .
[ chute3_fig5](a ) .
particles closest to the base undergo collisional motion with the fixed bed leading to the creation of a granular temperature front , as shown in fig .
[ chute3_fig5](b ) . this _
front propagates up the pile , disrupting the plug region . after the original initiation slippage
, the slip region disappears and the velocity profile appears approximately linear with height , with a crossover to the plug region . after the plug phase melts ,
the system dilates slightly , leading to a velocity profile of the bagnold steady state form . ) , initially at @xmath73 , that is tilted to @xmath132 .
initial failure occurs near the base , the pile then `` heats up '' from the bottom of the pile , a thermalization front propagates up the pile with time , disrupting the plug flow above it .
the lowest curve starts at @xmath133 after the system is first tilted , and we plot curves for @xmath134 , and @xmath135 respectively.,title="fig:",width=245 ] ) , initially at @xmath73 , that is tilted to @xmath132 .
initial failure occurs near the base , the pile then `` heats up '' from the bottom of the pile , a thermalization front propagates up the pile with time , disrupting the plug flow above it .
the lowest curve starts at @xmath133 after the system is first tilted , and we plot curves for @xmath134 , and @xmath135 respectively.,title="fig:",width=245 ] + * cessation * was observed for a system that was initially in steady state flow at @xmath136 , as the inclination angle was reduced to @xmath137 , which is below @xmath1 .
see fig .
[ chute3_fig6 ] .
although it may seem intuitive to expect an exact reversal of the initiation process , this is not observed .
after @xmath0 is reduced to below the angle of repose , particles at the bottom of the pile , which have the smallest velocities , come to rest first becoming trapped by the rough base .
the frozen layer then serves to trap the slowly moving particles above it .
this _ trapping _ front propagates up the pile towards the free surface , as indicated in fig .
[ chute3_fig6](a ) .
a consequence of this is that the granular temperature cools from the chute base as the whole system gradually comes to rest .
we note the emergence of a thin saltating region at the free surface during the cooling process . .
initially in steady state flow with @xmath63 , the inclination angle is then reduced to @xmath138 leading to the gradual stopping of flow into a jammed state .
particles moving slowest at the base stop flowing first , leading to a freezing front that propagates up the pile . the top curve is taken at a time @xmath139 after the angle is first reduced to below the angle of repose , and subsequent curves are separated by @xmath140 later.,title="fig:",width=245 ] . initially in steady state flow with @xmath63 , the inclination angle is then reduced to @xmath138 leading to the gradual stopping of flow into a jammed state .
particles moving slowest at the base stop flowing first , leading to a freezing front that propagates up the pile . the top curve is taken at a time @xmath139 after the angle is first reduced to below the angle of repose , and subsequent curves are separated by @xmath140 later.,title="fig:",width=245 ] + during this trapping process the pile exhibits two distinct behaviors separated by the trapping front .
below the trapping front , the particles are frozen into a pack , whereas above the trapping front , particles continue to flow . as the trapping front moves up the pile , the depth of the flowing section continuously shrinks with time . as shown in fig .
[ chute3_fig6](a ) , this leads to a bulk velocity profile that consequently transforms from bagnold - like flow to linear with depth until flow ceases . essentially , flow cessation is a transient version of route b in fig .
[ chute3_fig3 ] for flow at one angle over a range of flow heights .
another feature of flow transitions is in understanding the stress response of the system upon initiation and cessation of flow .
classical treatments of granular media @xcite demand that failure will only occur if the ratio of the shear stress to the normal stress equals the tangent of the angle of repose , @xmath141 however , eq . [ chute3_eq10 ] needs to be modified to take into account the history dependence of the starting point of flow for initiation at @xmath2 , and the general stopping point of flow for cessation at @xmath1 . in fig .
[ chute3_fig16 ] the stress ratio @xmath142 is plotted as a function of inclination angle @xmath0 for initiation of flow ( circles ) and cessation of flow ( squares ) .
we see that @xmath142 is linear with @xmath143 . still , to observe flow from initiation
, the stress ratio must exceed @xmath144 , or for flow to cease its value must be reduced to below @xmath145 . in other words ,
the shear stress can have a non - zero value and remain static as long as the stress ratio is smaller than the tangent of @xmath1 or @xmath2 .
the particular stress state of the _ static _ system is very much history dependent and comes about from the imposed conditions at the instant the system became static @xcite .
as a function of the tangent of the angle of inclination @xmath0 for initiation of flow ( squares ) and cessation of flow ( circles ) .
the open symbols correspond to static states and the filled symbols are flowing .
the lines indicate the positions of @xmath146 .
the stress ratio is linear with @xmath147 , but flow will not occur unless the stress ratio remains above @xmath145 in the case of a system already flowing , or exceeds @xmath144 for initiation of flow.,width=302 ]
we have performed a systematic simulation study of inclined plane , gravity - driven flow of granular materials , elucidating the reasons behind the discrepancies that exist between several different experimental procedures .
these arise as a consequence of how far the system is from jamming , or how close the inclination angle of the pile is from the angle of repose .
we have accurately determined the flow phase diagram for a system with a particular interaction parameter set , indicating the height dependence of the angle of repose and the maximum angle of stability .
we have also determined the boundary between steady state and unstable flow and find that the maximum angle for steady state flow also exhibits height dependence .
this suggests the existence of an intrinsic length scale in granular flows and forms part of our work in progress . for thick enough piles flowing on a rough inclined plane ,
say @xmath148 , the velocity profiles and rheology follow bagnold scaling . as one reduces the height of the flowing pile , a continuous transition from bagnold rheology to linear velocity profiles through to avalanche - like dynamics occurs , until finally , one reaches the angle of repose @xmath1 , and flow ceases . as a result
, one can obtain various velocity profiles and flow behaviors not only through changing the height of the flowing pile , but also by varying the inclination angle of the chute , such that there is an overlap region where one can obtain similar flow properties through either procedure .
these features can be used in surface flow treatments @xcite to better predict the evolution of an avalanche surface . our investigation on the initiation and cessation of flow shows that the initiation process starts from failure at the base that results in a `` heat '' front that propagates up the pile .
upon cessation of flow , an effective freezing front propagates up the pile with particles becoming trapped in the frozen layer until the whole system comes to rest .
in some cases , flow initiation or cessation only occurred after many time units ( approximately equivalent to 20 seconds for a 1 mm sized particle ) .
a consequence of these findings is that experimental reports of various flow behaviour in flowing piles may also be due to the time window of observation . in general , we have captured all the major observations from experiment , including quantitative agreement of the scaling parameter @xmath123 , eq .
[ chute3_eq9 ] , from pouliquen s empirical theory @xcite .
we thank deniz ertas and thomas halsey for helpful discussions and a critical reading of the manuscript .
les would also like to thank m. a. aguirre for insightful discussions .
this work was supported by the division of materials science and engineering , office of science , u.s .
department of energy .
sandia is a multiprogram laboratory operated by sandia corporation , a lockheed martin company , for the united states department of energy under contract de - ac04 - 94al85000 .
we note that in this study we are modelling hertzian contacts with static friction and tangenetial velocity damping . in our previous study
most of the 3d results were for linear spring contacts with static friction only . in the case of hertzian contacts , because the coefficient of restitution is velocity dependent and goes to @xmath151 as @xmath152 , the rolling damping term @xmath153 ensures that the energy of a static pile goes to zero . | the rheology of granular particles in an inclined plane geometry is studied using molecular dynamics simulations .
the flow no - flow boundary is determined for piles of varying heights over a range of inclination angles @xmath0 .
three angles determine the phase diagram : @xmath1 , the angle of repose , is the angle at which a flowing system comes to rest ; @xmath2 , the maximum angle of stability , is the inclination required to induce flow in a static system ; and @xmath3 is the maximum angle for which stable , steady state flow is observed . in the stable flow region @xmath4
, three flow regimes can be distinguished that depend on how close @xmath0 is to @xmath1 : i ) @xmath5 : bagnold rheology , characterized by a mean particle velocity @xmath6 in the direction of flow that scales as @xmath7 , for a pile of height @xmath8 , ii ) @xmath9 : the slow flow regime , characterized by a linear velocity profile with depth , and iii ) @xmath10 : avalanche flow characterized by a slow underlying creep motion combined with occasional free surface events and large energy fluctuations .
we also probe the physics of the initiation and cessation of flow .
the results are compared to several recent experimental studies on chute flows and suggest that differences between measured velocity profiles in these experiments may simply be a consequence of how far the system is from jamming . |
modelling nonparametric time series has received increasing interest among scholars for a few decades , see , for example , @xcite . in classical time series analysis , the stationarity of time series is a fundamental assumption .
yet , it may be violated on some occasions in such the fields as finance , sound analysis and neuroscience , especially when the time span of observations tends to infinity .
so , it is necessary to generalize the stationary process to the nonstationary process .
priestley ( 1965 ) @xcite first introduced a stochastic process with evolutionary spectra , which locally displays an approximately stationary behavior .
but in his framework , it is impossible to establish an asymptotic statistical inference .
dahlhaus ( 1997 ) proposed a new generalization of stationarity , called locally stationary process , and investigated its statistical inference .
more details can refer to @xcite .
in essence , the locally stationary process is locally close to a stationary process over short periods of time , but its second order characteristic is gradually changing as time evolves . a formal description of locally stationary process can refer to assumption ( a1 ) in the appendix . in parametric context , the statistical inference of locally stationary process has been studied extensively by @xcite . in nonparametric context , vogt ( 2012 )
@xcite considered the time - varying nonlinear autoregressive ( tvnar ) models including its general form and estimated the time - varying multivariate regression function using the kernel - type method .
however , it still suffers the `` curse of dimensionality '' problem when the dimension of covariates is high . in order to solve the aforementioned problem ,
a familiar way is to adopt the additive nonparametric regression model suggested by @xcite .
it not only remedies the `` curse of dimensionality '' , but also has an independent interest in practical applications due to its flexibility and interpretability .
there exists abound research findings about the additive regression model in the literature . in the case of iid observations ,
the additive nonparametric component functions can be estimated by kernel - based methods : the classic backfitting estimators of @xcite , the marginal integration estimators of @xcite , the smoothing backfitting estimators of @xcite , and two - stage estimators of @xcite . in the stationary time series context , there are kernel estimators via marginal integration of @xcite , spline estimators of @xcite , and the spline - backfitted kernel ( sbk ) estimators which borrow the strength of both kernel estimation and spline estimation , see @xcite .
vogt @xcite considered the locally stationary additive model and proposed smooth backfitting method to estimate bivariate additive component functions . on the other hand ,
the varying - coefficient model is a natural extension of linear model which allows the coefficients to change over certain common covariates instead of being invariant .
this model succeeds to relax the parameter limitation of linear model and may have practical as well as theoretical significance . for this model ,
there are three types of estimation methods : local polynomial smoothing method @xcite , polynomial spline estimation method @xcite and smoothing spline method @xcite .
zhang and wang @xcite proposed a so - called varying - coefficient additive model to catch the evolutionary nature of time - varying regression function in the analysis of functional data .
their model assumes the evolutionary regression function has the form @xmath2 which is more flexible in the sense that it covers both varying - coefficient model and additive model as special cases . specifically speaking
, it reduces to an additive model when @xmath3 are all constants , and a varying - coefficient model if @xmath4 are all linear functions . extracting the special meaning of time in functional data analysis
, one can generalize time to some other common covariates . in this paper , we model locally stationary time series . to concreteness ,
let @xmath5 be a length-@xmath6 realization of @xmath7 dimension locally stationary time series , and assume that the data is generated by varying - coefficient additive model as follows @xmath8 where @xmath9 s are i.i.d , @xmath10 is the varying - coefficient component function , @xmath11 is the additive component function and @xmath12 is a bivariate nonparametric function , which allows the heteroscedasticity case . without loss of generality , we require @xmath13^{p}},$ ] where the superscript ` @xmath14 ' means transposition of vector or matrix . in order to identify these multiplied component functions , we require that @xmath15 where @xmath16 is the @xmath1 norm of any function @xmath17 defined on @xmath18 $ ] such that @xmath19 for functional data , zhang and wang @xcite proposed a two - step spline estimation procedure . in the first step , sorting the data within each subject in ascending order of time and averaging the response for each subject using trapezoidal rule to fit an additive model , then , in the second step , fitting a varying - coefficient model by substituting the estimated additive function into varying - coefficient additive model .
his estimation methodology works since there are dense observation for every subject and covariates is independent of observation time within subject
. however , for some other practical problems , such as longitudinal data with finite observertion time , time series data , such an assumption fails . to circumvent this problem , under mild assumptions ,
we derive an initial estimation of additive component by employing a segmentation technique
. then we can fit a varying - coefficient model and an additive model , respectively , to get spline estimators of varying - coefficient function and additive function . as expected , we show that the proposed estimators of @xmath10 and @xmath11 are consistent and present the corresponding @xmath1 rate of convergence . on the other hand
, the product term in may simply reduce to a varying - coefficient term or an additive term in the case of @xmath11 being linear function or @xmath10 being constant .
so , in the parsimony sense , identifying additive terms and varying - coefficient terms in are of interest . to this end
, we propose a two - stage penalized least squares estimator based on scad penalty function , and , furthermore , show that our model identification strategy is consistent , i.e. , the additive term and the varying - coefficient term are correctly selected with probability approaching to 1 .
meantime , @xmath1 rate of convergence of penalized spline estimator of each component function achieves the rate of the spline estimator of univariate nonparametric function .
the rest of this paper is organized as follows .
we propose a three - step spline estimation method in section 2 and a two - stage model identification procedure in section 3 .
section 4 describes the smoothing parameter selection strategies .
section 5 establishes the asymptotic properties of the proposed model estimation and identification methods .
simulation studies are illustrated in section 6 .
the main technical proofs are presented in the appendix .
lemmas and other similar proofs are given in the supplementary .
in this section , we propose a three - step spline estimation method for the proposed locally stationary varying - coefficient additive model .
* step i : segment the rescaled time @xmath20 into several groups , and approximate each varying - coefficient function @xmath10 within the same group by a local constant . thus , model can be approximated by an additive model , and a scaled - version of the additive component functions @xmath11 can be obtained using spline - based method . * step ii : substitute the initial estimates of the scaled additive component functions into model to yield an approximated varying - coefficient model , and then obtain spline estimators of varying - coefficient component functions @xmath21 * step iii : plug - in spline estimators of varying - coefficient component functions @xmath10 into model to yield an approximated additive model , and then update the spline estimation of additive component functions @xmath11 .
we first present with some notations before detailing our proposed estimation method .
let @xmath22 be @xmath23 order b - spline basis with @xmath24 interior knots and @xmath25 is the number of b - spline functions estimating additive component function @xmath26 similarly , we denote @xmath27 as @xmath28 order b - spline basis with @xmath29 interior knots , and @xmath30 is the number of b - spline functions estimating varying - coefficient component function @xmath21 here , ` @xmath31 ' and ` @xmath32 ' in the subscript of b - spline functions and knots number mean that is for the additive component function and varying - coefficient function , respectively . denote @xmath33 and @xmath34 the nice properties of scaled b - spline basis are listed in the appendix . +
* step i : initial estimators of scaled additive component functions * we segment the sample @xmath35 in ascending order of time into @xmath36 groups with @xmath37 observations in each group , where @xmath38 hinges on the sample size @xmath6 and @xmath39 . then approximate @xmath40 in the @xmath41th group , i.e. @xmath42 by a constant @xmath43 , where @xmath43 is some constant dependent on @xmath44 and @xmath41 such that @xmath45 for the sake of convenient presentation , we suppress the triangular array index in locally stationary time series , and represent time index in the @xmath41th group as @xmath46 for given @xmath47 then one can approximate model as @xmath48 where @xmath49 if @xmath50 are all known , one can easily construct the spline estimator of @xmath26 suppose that @xmath51 minimizes @xmath52 ^ 2,\ ] ] then @xmath53 where @xmath54 however , @xmath43 s are unknown
. we instead rewrite as an additive model , @xmath55 where @xmath56 for each given @xmath57 let @xmath58 minimize @xmath59 ^ 2.\ ] ] by and , it is easy to see that @xmath60 which implies @xmath61 in a word , although the additive component function in can not be estimated directly , the scaled additive component function @xmath62 with @xmath63 is estimable using the proposed segmentation techniques . + * step ii : spline estimators of varying - coefficient component functions * define @xmath64 and @xmath65 , @xmath66 . by , substituting @xmath67 into yields @xmath68 where @xmath69.$ ] model can be viewed as a varying - coefficient model , and the spline estimators of varying - coefficient functions @xmath70 are easily obtained .
suppose that @xmath71 minimizes @xmath72 ^ 2.\ ] ] then , @xmath73 by the definition of @xmath74 and identifiability conditions for @xmath75 we have the spline estimators of varying - coefficient functions @xmath76 s in model as @xmath77 * step iii : spline estimators of additive component functions * substituting into model yields @xmath78 where @xmath79\beta_{k}(x_{t , t}^{\left(k\right)}).$ ] model can be viewed as a varying - coefficient model .
suppose that @xmath80 minimizes @xmath81 ^ 2,\ ] ] then , spline estimators of additive component functions in are given by @xmath82 * remark 1 : * the spline estimators @xmath83 and @xmath84 can be updated by iterating step ii and step iii .
however , one step estimation is enough and there is no great improvement through iteration procedure .
* remark 2 : * one may employ different b - spline basis functions in step i and step iii for estimating the additive component functions @xmath85 .
yet , we do nt distinguish them in symbols for the sake of simplicity .
the proposed varying - coefficient additive model is more general and flexible than either varying - coefficient model or additive model , and covers them as special cases .
but , in practice , a parsimonious model is always one s preference when there exist several potential options .
hence , it is of great interest to explore whether the varying - coefficient component function is truly varying and whether the additive component function degenerates to simply linear function . in this paper
, we decompose varying - coefficient additive terms into additive terms @xmath11 and varying - coefficient terms @xmath10 , and , motivated by @xcite , propose a two - stage penalized least squares ( pls ) model identification procedure to identify the term that @xmath11 is constant ( @xmath86 ) or / and @xmath10 is linear ( @xmath87 ) .
* stage i : plug - in the spline estimators of additive component functions @xmath11 obtained in the estimation stage into model , and penalize @xmath88 to identify linear additive terms . *
stage ii : given the penalized spline estimators of additive component functions @xmath11 obtained in stage i of the model identification process , penalize @xmath89 to select constant varying - coefficient terms .
we first introduce some notations .
let @xmath90 and @xmath91 denote @xmath92 and @xmath93 * stage i : identifying linear additive terms * by substituting the additive component functions @xmath94 by their spline estimates @xmath95 obtained in the estimation stage , model becomes @xmath96 where @xmath97.$ ] let @xmath98 with @xmath99 , and assume @xmath100 is determined by @xmath101 ^ 2\\ & + t\sum_{k=1}^{p}p_{\lambda_{t}}\left(k_{c}^{-3/2}\parallel\alpha'_{k}\parallel_{l_{2}}\right ) , \end{aligned}\ ] ] where @xmath102 and @xmath103 is a penalty function with a tuning parameter @xmath104 then the penalized spline estimators of @xmath76 are given by @xmath105 here the superscript ` @xmath106 ' denotes the penalized spline estimation . +
* stage ii : identifying constant varying - coefficient terms * by replacing the varying - coefficient function @xmath76 with their penalized spline estimates @xmath107 obtained in stage i , model becomes @xmath108 where @xmath109\beta_{k}\left(x_{t , t}^{\left(k\right)}\right).$ ] let @xmath110 with @xmath111 , and assume @xmath112 is given by @xmath113 ^ 2\\ & + t\sum_{k=1}^{p}p_{\mu_{t}}\left(k_{a}^{-3/2}\parallel\beta''_{k}\parallel_{l_{2}}\right ) , \end{aligned}\ ] ] where @xmath114 and @xmath115 is a penalty function with a tuning parameter @xmath116 therefore , the penalized spline estimators of @xmath117 are given by @xmath118,\ k=1,\cdots , p.\ ] ]
in this section , we discuss various implementation issues for the proposed model estimation and identification procedures .
we predetermine the degree of polynomial spline .
usual options are 0 , 1 or 2 , that is to choose linear , quadratic or cubic spline functions .
it is known that , when sufficient number of knots is used , the spline approximation method is quite stable .
therefore , we suggest to use the same number of interior knots @xmath119 for all component functions and @xmath120th order b - spline basis functions in the @xmath121th step estimation to facilitate the computation . by experience
, it is reasonable to choose @xmath122 in addition , in order to solve the least squares problem in each group in step i estimation , we require @xmath123 under this constraint , we choose the optimal @xmath119 and @xmath37 by bic @xmath124 where @xmath125 ^ 2 $ ] and @xmath126 is the number of b - spline basis functions used in step iii estimation . various penalty functions @xcite can be used in practice .
we choose the scad penalty function proposed by @xcite , which is defined by its first derivative @xmath127 for some @xmath128 and @xmath129 where symbol @xmath130 it is well - known that the scad penalty function has nice properties such as unbiasedness , sparsity and continuous .
meantime , it is singular at the origin , and have no continuous second order derivatives .
yet it can be locally approximated by a quadratic function .
specifically speaking , given an initial estimate @xmath131 or equivalently @xmath132 if @xmath133 then one can locally approximate @xmath134 by @xmath135 this implies that the objective function in , denoted by @xmath136 can be approximated , up to a constant , by @xmath137 where @xmath138 @xmath139 with @xmath140 and @xmath141 and @xmath142 therefore , we can find the solution of by iteratively computing the following ridge regression estimator @xmath143 until convergence . in the same vein , we can iteratively solve the optimization problem .
let @xmath144 @xmath145 with @xmath146 and @xmath147 and @xmath148 therefore , we can iteratively compute the following ridge regression estimator @xmath149 until convergence .
based on the optimal segment length @xmath150 and the optimal number of interior knots @xmath151 , we then select tuning parameters @xmath152 and @xmath153 for the proposed two - stage model identification procedure .
following @xcite , we take @xmath154 and find optimal tuning parameters @xmath152 and @xmath153 by bic in two steps .
first , to select optimal @xmath152 , we define @xmath155 where @xmath156^{2}$ ] , @xmath157 is the number of b - spline basis functions adopted in the second step estimation and @xmath158 is the number of linear additive terms , i.e. , @xmath159 is sufficiently small , say , no larger than @xmath160 second , to select optimal @xmath153 , we define @xmath161 where @xmath162^{2}$ ] , and @xmath163 is the number of constant varying - coefficient terms , i.e. , @xmath164 is sufficiently small .
thus , we select the optimal tuning parameters @xmath165
+ in this section , we demonstrate that the proposed three - step spline estimation method is consistent under regularity conditions and show that the proposed two stage model identification procedure can correctly select additive terms and varying - coefficient terms with probability approaching one .
furthermore , we conclude that @xmath1 rate of convergence of each component function achieves the optimal rate of the spline estimator of univariate nonparametric function stated in @xcite .
the regularity conditions and assumptions are given in the appendix .
+ let @xmath166 and @xmath167 .
we introduce @xmath168}|\beta_{k}\left(x\right)-\mu_{k}\left(x\right)|,\ \ \rho_{c}=\sum_{k=1}^{p}\inf_{\nu_{k}\in h_{k}}\sup_{x\in[0,1]}|\alpha_{k}\left(x\right)-\nu_{k}\left(x\right)|\ ] ] to measure the degree of spline approximation of varying - coefficient component function and additive component function .
proposition 1 establishes @xmath1 rate of convergence of initial estimators of scaled additive component functions @xmath169 [ prop1 ] under assumptions ( a1 ) , ( a2 ) , ( a4 ) , ( a5 ) , ( a7 ) , ( a8 ) and ( a9 ) , if @xmath170 as @xmath171 @xmath172 @xmath173^{2 } = o_p\left(\frac{\rho_{a}^{2}}{n_{t}}+\frac{k_{a}}{t}\right),\ ] ] * remark 3 : * in comparison with the convergence of the spline estimation of univariate nonparametric function , we notice that the bias term in @xmath1 rate of convergence of initial estimators is smaller when the number of groups @xmath36 is larger than 1 .
the larger number of segmentation groups , the smaller the bias , given the number of observations in each group is at least larger than the number of parameters in spline approximation of @xmath174 based on the result of proposition 1 , one can construct @xmath1 rate of convergence of the spline estimation of varying - coefficient component function @xmath10 as follows .
[ alpha ] under assumptions ( a1 ) ( a10 ) , if @xmath175 as @xmath171 @xmath176 @xmath177^{2 } = o_p\left(\frac{\rho_{a}^{2}}{n_{t}}+\rho_{c}^{2}+\frac{k_{a}\vee k_{c}}{t}\right),\ ] ] where ` @xmath178 ' denotes the maximum of @xmath179 and @xmath180 * remark 4 : * theorem [ alpha ] shows that there exists an additional bias term @xmath181 in comparison with the convergence of the spline estimation of univariate nonparametric function .
this term happens to be the rate of convergence obtained in proposition 1 and reflects the error of the initial estimator of scaled additive function @xmath174 next theorem presents @xmath1 rate of convergence of the spline estimation of additive component function @xmath26 [ beta ] under the assumptions of theorem 1 , @xmath182 * remark 5 : * similarly , in comparison with the rate of convergence of spline estimation for univariate nonparametric function , theorem [ beta ] also has an additional bias term @xmath183 .
the reason this term exists is because the estimation of additive component function @xmath11 in step iii is based on the spline estimates of varying - coefficient function @xmath10 obtained in step ii .
as expected , the convergence of the updated spline estimation of @xmath11 doe not depend on the number of segmentation groups in step i of the initial estimation of rescaled additive function @xmath174 we , here , respectively , demonstrate the consistency of selecting additive terms and varying - coefficient terms , and present the @xmath1 rate of convergence of penalized spline estimators of @xmath10 and @xmath26 [ cadditive ] suppose that @xmath184 given @xmath185 @xmath186 and @xmath187 with @xmath188 , then , under assumptions ( a1 ) ( a10 ) , as @xmath171 * with probability approaching to 1 , @xmath189 is constant a function a.s . for @xmath190
* @xmath1 rate of convergence for penalized spline estimator of @xmath10 is given by @xmath191 for @xmath192 where @xmath193 and @xmath194 [ cvarying ] suppose that @xmath195 given @xmath196 @xmath197 and @xmath198 with @xmath199 then , under assumptions ( a1 ) ( a10 ) , as @xmath171 * with probability approaching to 1 , @xmath200 is a linear function a.s . for @xmath201
* @xmath1 rate of convergence for penalized spline estimator of @xmath11 is given by @xmath202 for @xmath203 where @xmath204 and @xmath205 * remark 6 : * theorems [ cadditive ] and [ cvarying ] show that the penalized spline estimators of varying - coefficient component function @xmath10 and additive component function @xmath11 both have the same @xmath1 rate of convergence as that of the spline estimator of univariate nonparametric function .
we consider two simulation examples to illustrate the finite sample performance of the proposed three - step spline estimation method and two - stage model selection procedure , respectively .
the data are generated from the varying - coefficient additive model as follows @xmath206 where @xmath207 @xmath208 and @xmath209 are iid standard normal variables and @xmath210 to appraise the performance of the proposed three - step spline estimators , we use the mean integrated squared error ( mise ) based on @xmath211 monte carlo replications , that is @xmath212^{2}\mathrm{d}u,\ \ k=0,1,2,\ ] ] @xmath213^{2}\mathrm{d}x_{k},\ \ k=1,2,\ ] ] where @xmath214 and @xmath215 are estimators of @xmath76 and @xmath216 respectively , in the @xmath217-th monto carlo sample .
the univariate nonparametric functions are approximated by b - spline of order of three or the quadratic splines .
we consider the sample size @xmath218 the number of interior knots @xmath219 and the segmentation length @xmath220 in step i estimation .
we run the simulation for 500 times , and find out the mise of three - step estimators decreases as the sample size increases , regardless the values of @xmath119 and @xmath221 .
table [ tag:1 ] only gives the results of @xmath222 for different combinations of @xmath221 and @xmath6 .
in addition , we list the mise of oracle estimators , which refer to the spline estimator of @xmath10 given all additive component functions are known in advance , or correspondingly , the spline estimator of @xmath11 given all varying - coefficient component functions are known in advance .
as expected , mise of oracle estimators for varying - coefficient components and additive components are better than those of the proposed three - step spline estimators .
the last two columns in table [ tag:1 ] depict the mise of spline estimators for nonparametric component functions in misspecified varying - coefficient model and misspecified additive model .
we note that they are discernibly larger than three - step spline estimators in varying - coefficient additive model .
.comparison of mise of different estimators in example 1 ( @xmath222 ) [ cols="^,^,^,^,^,^,^",options="header " , ]
let @xmath223 be the euclidean norm of a real valued vector @xmath224 denote the space of @xmath225-order smooth functions defined on @xmath18 $ ] as @xmath226=\{m|m^{\left(l\right)}\in c[0,1]\}$ ] and the class of lipschitz continuous functions for some fixed constant @xmath227 as @xmath228,c\right)=\{m||m\left(x\right)-m\left(x'\right)|\leq c|x - x'|,\forall x , x'\in[0,1]\}.$ ] + the necessary conditions to prove asymptotic properties are listed as below .
* the process @xmath229 is stationary locally in time , that is , for each rescaled time point @xmath230,$ ] there exists a strictly stationary process @xmath231 such that @xmath232 a.s . with @xmath233\leq c$ ] for some @xmath234 and @xmath235 independent of @xmath236 and @xmath237 * at each rescaled time point @xmath230,$ ] the joint density function @xmath238 of the stationary approximation process @xmath239 is bounded below and above uniformly on @xmath230:$ ] @xmath240^{p}}f_{u}\left(\mathbf{x}\right)\leq\sup_{\mathbf{x}\in[0,1]^{p}}f_{u}\left(\mathbf{x}\right)\leq c_{f } \
\text{uniformly on}\ u\in[0,1].\ ] ] meantime , @xmath241 has density function with respective to certain measure .
* @xmath242 iid , and @xmath243 @xmath244 given @xmath245 @xmath246 * there exists positive constants @xmath247 and @xmath248 such that @xmath249 for all @xmath250 where @xmath251 is the @xmath252-mixing coefficients for process @xmath253 and defined as latexmath:[\[\alpha\left(k\right)=\sup_{b\in\sigma\{\mathcal{z}_s , s\leq t\ } , c\in \sigma\{\mathcal{z}_s , s\ge t+k\ } } * the conditional standard deviation function @xmath12 is bounded below and above uniformly on @xmath230,$ ] i.e. , @xmath255^{p}}\sigma\left(u,\mathbf{x}\right ) \leq \sup_{\mathbf{x}\in[0,1]^{p}}\sigma\left(u,\mathbf{x}\right)\leq c_{\sigma}\ \text{uniformly on}\ u\in[0,1]\ ] ] for some positive constants @xmath256 and @xmath257 * @xmath258 $ ] and @xmath259,c_{1,k})$ ] where @xmath260 is an integer such that @xmath261 * @xmath262 $ ] and @xmath263,c_{2,k})$ ] where @xmath264 is an integer such that @xmath265 * the knots for @xmath266 @xmath267 and the knots for @xmath268 @xmath269 has bounded mesh ratio : @xmath270 @xmath271 * @xmath272 , @xmath273 @xmath274 and @xmath275 * @xmath276 and @xmath277 * remark 7 : * assumption ( a1 ) specifies a data generating process ( gdp ) . assumptions ( a2 ) - ( a5 ) are standard in time - series context , see @xcite .
assumptions ( a6 ) - ( a8 ) are common in typical spline approximation literature , for instance , @xcite .
firstly , we need some lemmas before proving main theorems .
let @xmath278 and @xmath279 be any two - vector valued function .
define empirical inner product @xmath280 and theoretical inner product @xmath281\mathrm{d}u,\ ] ] denote the induced norm by @xmath282 and @xmath283 as @xmath284 and @xmath285 respectively .
in addition to , @xmath286 given sequences of positive numbers @xmath287 and @xmath288 @xmath289 means @xmath290 is bounded and @xmath291 means @xmath289 and @xmath292 hold .
[ inducenorm ] let @xmath293 and @xmath294 for @xmath295 denote @xmath296 and @xmath297 then @xmath298 holds under assumption ( a2 ) .
in combination with assumption ( a2 ) and the property ( iv ) of @xmath299 we have @xmath300\mathrm{d}u = \sum_{k=1}^{p}\int_{0}^{1}e\big[\sum_{l=1}^{j_{k , a}}\gamma_{kl}\psi_{kl}(x_{t}^{\left(k\right)}\left(u\right))\big]^{2}\mathrm{d}u\\ \asymp&\sum_{k=1}^{p}\int_{0}^{1}\parallel g_{k}\parallel^{2}_{l_{2}}\mathrm{d}u=\parallel\mathbf{g}\parallel_{l_{2}}^{2 } \asymp\sum_{k=1}^{p}|\gamma_{k}|^{2}=|\mathbf{\gamma}|^{2 } \end{aligned}\ ] ] [ inner ] let @xmath301 be the collection of vector valued functions @xmath278 with @xmath302 such that @xmath303 for @xmath304 where @xmath305 is defined in section 5.1 .
then under assumption ( a1 ) , ( a2 ) , ( a4 ) and ( a9 ) , as @xmath171 @xmath306 for any @xmath307 there exists coefficients @xmath308 and @xmath309 @xmath310 such that @xmath311 for @xmath295 it is not difficult to see that @xmath312 for any given @xmath304 let @xmath313 if the intersection of the supports of @xmath314 and @xmath315 contains an open interval .
that is , @xmath316 if @xmath317 moreover , it is known @xmath318 for some constant @xmath32 and all @xmath319 moveover , for any given @xmath320 and @xmath321 we have @xmath322\mathrm{d}u|\\ \leq & t^{-1}\sum_{t=1}^{t}|\psi_{kl}(x_{t , t}^{\left(k\right)})\psi_{kl'}(x_{t , t}^{\left(k\right ) } ) -\psi_{kl}(x_{t}^{\left(k\right)}(t / t))\psi_{kl'}(x_{t}^{\left(k\right)}(t / t))|\\ + & t^{-1}\sum_{t=1}^{t}|\psi_{kl}(x_{t}^{\left(k\right)}(t / t))\psi_{kl'}(x_{t}^{\left(k\right)}(t / t ) ) -e[\psi_{kl}(x_{t}^{\left(k\right)}(t / t))\psi_{kl'}(x_{t}^{\left(k\right)}(t / t))]|\\ + & |t^{-1}\sum_{t=1}^{t}e[\psi_{kl}(x_{t}^{\left(k\right)}\left(t / t\right))\psi_{kl'}(x_{t}^{\left(k\right)}\left(t / t\right ) ) ] -\int_{0}^{1}e[\psi_{kl}(x_{t}^{\left(k\right)}\left(u\right))\psi_{kl'}(x_{t}^{\left(k\right)}\left(u\right))]\mathrm{d}u| .
\end{aligned}\ ] ] by assumption ( a1 ) and the boundness of b - spline , the first term above is bounded by @xmath323 employing berstein s inequality , the second term is bounded by @xmath324 the last term is bounded by @xmath325 from the integral theory .
therefore , using cauchy - schwartz inequality and assumption ( a9 ) , we obtain that @xmath326 where @xmath327 and @xmath328 denote the vectors with entries @xmath308 and @xmath309 respectively . by lemma [ inducenorm ]
, we see @xmath329 which completes the proof .
[ eigenvalue ] under assumption ( a1 ) , ( a2 ) ( a4 ) and ( a9 ) , as @xmath171 * for each @xmath330 @xmath331 has eigenvalues bounded away from 0 and @xmath332 with probability tending to one ; * @xmath333 has eigenvalues bounded away from 0 and @xmath334 with probability tending to one .
we only show ( ii ) , the proof of ( i ) is similar . for
any given vector @xmath296 with @xmath335 @xmath336 let @xmath293 and @xmath337 where @xmath338 denote @xmath339 then by lemma [ inducenorm ] and [ inner ] , @xmath340 which implies @xmath341 has eigenvalues bounded away from 0 and @xmath342 therefore , @xmath343 also has eigenvalues bounded away from 0 and @xmath342 * proof for proposition 1 .
* + let @xmath344 then @xmath345 define @xmath346 @xmath347 and @xmath348 denote @xmath349 and @xmath350 @xmath351 note that @xmath352 by cauchy - schwartz inequality , we have @xmath353 it suffices to deal with the approximation error terms @xmath354 and stochastic error terms @xmath355 * approximate error terms : * we will show the rate of approximation error @xmath356 we have @xmath357\parallel^2_{l_{2}}\\ & \preceq \frac{1}{n_{t}^2}\sum_{s=1}^{n_{t } } \parallel\tilde{\beta}_{k}^{\left(s\right)}\left(x_{k}\right)-\beta_{k}^{\left(s\right)}\left(x_{k}\right)\parallel^2_{l_{2}}. \end{aligned}\ ] ] by the definition of @xmath358 there exists @xmath359 and @xmath360 such that @xmath361}\big|\breve{\beta}_{k}^{\left(s\right)}\left(x_{k}\right)-\beta_{k}^{\left(s\right)}\left(x_{k}\right)\big|=o\left(\rho_{a}\right).$ ] therefore , @xmath362 let @xmath363 then @xmath364 on the one hand , @xmath365 ^ 2 \asymp|\mathbf{\tilde{\alpha}}^{\left(s\right)}-\mathbf{\breve{\alpha}}^{\left(s\right)}|^2 $ ] by lemma [ eigenvalue ] . on the other hand
, @xmath366 ensures that @xmath367^{2 } \leq \frac{1}{i_{t}}\sum_{j=1}^{i_{t}}(\omega_{s , j}-v_{sj}^{\tau}\mathbf{\breve{\alpha}}^{\left(s\right)})^2\\ = & \frac{1}{i_{t}}\sum_{j=1}^{i_{t}}\big\{\sum_{k=1}^{p}\big[\beta_{k}^{\left(s\right)}(x_{t_{sj},t}^{\left(k\right ) } ) -\breve{\beta}_{k}^{\left(s\right)}(x_{t_{sj},t}^{\left(k\right)})\big]\big\}^{2}=o\left(\rho_{a}^{2}\right ) .
\end{aligned}\ ] ] thus , @xmath368 which means @xmath369 * stochastic error terms : * we next show the rate of stochastic errors : @xmath370 note that @xmath371 under assumption ( a5 ) , we obtain that @xmath372 \varepsilon_{t_{sj}}^{2}\big\}\\ & + c_{\sigma}e\big\{\sum_{\substack{j=1\\j\neq j'}}^{i_{t}}\big[1+\sum_{k=1}^{p}\sum_{l=1}^{j_{k , a}}\psi_{kl}(x_{t_{sj},t}^{\left(k\right ) } ) \psi_{kl}(x_{t_{sj'},t}^{\left(k\right)})\big]\varepsilon_{t_{sj}}\varepsilon_{t_{sj'}}\big\}. \end{aligned}\ ] ] assumption ( a3 ) makes the second term be zero , and the first term is bounded by @xmath373^{2}.$ ] however , @xmath374^{2 } \preceq & \sum_{k=1}^{p}\sum_{l=1}^{j_{k , a}}e\big[\psi_{kl}(x_{t_{sj},t}^{\left(k\right ) } ) -\psi_{kl}(x_{t_{sj}}^{\left(k\right)}\left(t_{sj}/t\right))\big]^{2}\\ & + \sum_{k=1}^{p}\sum_{l=1}^{j_{k , a}}e[\psi_{kl}(x_{t_{sj}}^{\left(k\right)}\left(t_{sj}/t\right))]^{2}. \end{aligned}\ ] ] by assumption ( a2 ) and the properties of b - spline , @xmath375 ^ 2 \asymp\sum_{k=1}^{p}j_{k , a}=o\left(k_{a}\right ) . \ ] ] on the other hand , @xmath376^{2 } = j_{k , a}e[b_{kl , a}(x_{t , t}^{\left(k\right)})-b_{kl , a}(x_{t}^{\left(k\right)}\left(t / t\right))]^{2}\\ \leq & cj_{k , a}e[|\mathbf{x}_{t , t}-\mathbf{x}_{t}\left(t / t\right)|^2 ] \leq cj_{k , a}\frac{1}{t^{2}}e[u_{t , t}^{2}\left(t / t\right ) ] .
\end{aligned}\ ] ] therefore , @xmath377^{2 } = o\left(k_{a}i_{t}\right)+\frac{i_{t}}{t^{2}}\sum_{k=1}^{p}j_{k , a}^{2}. \end{aligned}\ ] ] and in turn @xmath378 which completes the proof of and hence the first half of theorem 1 .
the rest is direct from lemma [ eigenvalue ] . + * proof for theorem [ alpha ] .
* let @xmath379 @xmath380 where @xmath381 denote @xmath382 @xmath383 denote @xmath384 where @xmath385 @xmath386 and @xmath387 then @xmath388 furthermore , assuming that @xmath389 with @xmath390 is given by @xmath391 and @xmath392 for @xmath393 by cauchy - schwartz inequality and identifiable condition @xmath394 we get @xmath395 * approximate error term : * we show the rate of approximate error term as follows @xmath396 by the definition of @xmath397 there exists @xmath398 such that @xmath399 satisfying @xmath400}|\breve{\alpha}_{k}\left(u\right)-\alpha_{k}\left(u\right)|=o\left(\rho_{c}\right)\ ] ] for @xmath393 thus @xmath401 note that @xmath402 and the normal equation @xmath403 yields @xmath404 according to proposition 1 and boundness of @xmath405 @xmath406^{2 } = o_{p}\left(\frac{\rho_{a}^{2}}{n_{t}}+\frac{k_{a}}{t}\right ) .
\end{aligned}\ ] ] on the other hand , from assumption ( a7 ) @xmath407^{2 } + \frac{1}{t}\sum_{t=1}^{t}\sum_{k=1}^{p}\hat{\gamma}_{k}^{2}(x_{t , t}^{\left(k\right ) } ) [ \delta_{k}\left(t / t\right)-\breve{\delta}_{k}\left(t / t\right)]^{2}\\ \preceq&\rho_{c}^{2}+\rho_{c}^{2}\frac{1}{t}\sum_{k=1}^{p}\sum_{t=1}^{t}\hat{\gamma}_{k}^{2}(x_{t , t}^{\left(k\right)})\\ = & o\left(\rho_{c}^{2}\right ) .
\end{aligned}\ ] ] * stochastic error terms : * we next show the following rate of stochastic error term @xmath408 it is easy to see @xmath409 based on assumption ( a5 ) , it is sufficient to bound @xmath410 let @xmath411 and @xmath412 then @xmath413 obviously , under assumption ( a3 ) , @xmath414 .
$ ] from assumption ( a1 ) and assumption ( a7 ) , @xmath415^{2 } \preceq & e[\gamma_{k}(x_{t , t}^{\left(k\right)})-\gamma_{k}(x_{t}^{\left(k\right)}(t / t))]^{2 } + e[\gamma_{k}(x_{t}^{\left(k\right)}(t / t))]^{2}\\ \preceq & e|\mathbf{x}_{t , t}-\mathbf{x}_{t}\left(t / t\right)|^{2}+o\left(1\right)\\ = & o\left(t^{-2}\right)+o\left(1\right)=o\left(1\right ) , \end{aligned}\ ] ] where @xmath416 is the @xmath44-th component of stationary approximation process @xmath231 of locally stationary process @xmath229 at rescaled time @xmath417 in combination with @xmath418 we have @xmath419\preceq & \sum_{l=1}^{j_{0,c}}\varphi_{0l}^{2}\left(t / t\right)+\sum_{k=1}^{p}\sum_{l=1}^{j_{k , c}}\varphi_{kl}^{2}\left(t / t\right ) = o\left(k_{c}\right ) , \end{aligned}\ ] ] which means @xmath420 meantime , @xmath421^{2}. \end{aligned}\ ] ] by cauchy - schwartz inequality , @xmath422^{2 } \leq 3 e[\hat{\gamma}_{k}(x_{t , t}^{\left(k\right)})-\hat{\gamma}_{k}(x_{t}^{\left(k\right)}\left(t / t\right))]^{2}\\ & + 3e[\hat{\gamma}_{k}(x_{t}^{\left(k\right)}\left(t / t\right))-\gamma_{k}(x_{t}^{\left(k\right)}\left(t / t\right))]^{2 } + 3e[\gamma_{k}(x_{t}^{\left(k\right)}\left(t / t\right))-\gamma_{k}\big(x_{t , t}^{\left(k\right)})]^{2}. \end{aligned}\ ] ] the third term is bounded by @xmath423 from assumption ( a1 ) and ( a7 ) .
assumption ( a2 ) and proposition 1 ensure that the second term is bounded by @xmath424 for the first term , we note that @xmath425^{2}\\ = & e\big[n_{t}^{-1}\sum_{s=1}^{n_{t}}\sum_{l=1}^{j_{k , a}}\hat{h}^{\left(s\right)}_{kl}j_{k , a}^{1/2 } \{b_{kl , a}(x_{t , t}^{\left(k\right)})-b_{kl , a}(x_{t}^{\left(k\right)}\left(t / t\right))\}\big]^{2}\\ \preceq & j_{k , a}n_{t}^{-2}\sum_{s=1}^{n_{t}}\sum_{l=1}^{j_{k , a}}(\hat{h}^{\left(s\right)}_{kl})^{2 } = & o_p\left(\frac{k_{a}^{2}}{t^{2}n_{t}}\right ) , \end{aligned}\ ] ] and thus @xmath426^{2 } o_{p}\left(\frac{\rho_{a}^{2}}{n_{t}}+\frac{k_{a}}{t}\right ) = o_p\left(\frac{k_{c}}{t}\big\{\frac{\rho_{a}^{2}}{n_{t}}+\frac{k_{a}}{t}\big\}\right ) , \end{aligned}\ ] ] which shows . + * proof for theorem [ beta ] .
* let @xmath427 @xmath428 and @xmath429 @xmath430 define @xmath431 where @xmath432 @xmath433 and @xmath434 for @xmath435 suppose @xmath436 is given by @xmath437 and @xmath438 analogously , represent @xmath439 with @xmath440 obviously , @xmath441 * approximation error term : * the rate of approximation error term is given by @xmath442 on the one hand , by the definition of @xmath358 there exists @xmath443 and @xmath444 such that @xmath445}|\beta^*\left(x_{k}\right)-\beta_{k}\left(x_{k}\right)|=o\left(\rho_{a}\right),\ ] ] which means @xmath446 on the other hand , @xmath447 furthermore , @xmath448 since @xmath449 according to theorem 1 , @xmath450^{2}+ \frac{1}{t}\sum_{k=1}^{p}[\alpha_{k}\left(t / t\right)-\hat{\alpha}_{k}\left(t / t\right)]^{2}\beta_{k}^{2}(x_{t , t}^{\left(k\right)})\\ = & o_{p}\left(\frac{\rho_{a}^{2}}{n_{t}}+\rho_{c}^{2}+\frac{k_{a}\vee k_{c } } { t}\right ) .
\end{aligned}\ ] ] finally , we note that @xmath451\big|^{2}\\ \preceq&\frac{1}{t}\sum_{t=1}^{t}\sum_{k=1}^{p } \hat{\alpha}_{k}^{2}\left(t / t\right)[\beta_{k}(x_{t , t}^{\left(k\right ) } ) -\beta^*_{k}(x_{t , t}^{\left(k\right)})]^{2 } = o\left(\rho_{a}^{2}\right ) \end{aligned}\ ] ] since for each @xmath304 @xmath452^{2 } + \frac{1}{t}\sum_{t=1}^{t}\alpha_{k}^{2}\left(t / t\right)\\ = & \int_{0}^{1}\alpha_{k}^{2}\left(u\right)\mathrm{d}u+o\left(1\right)=o\left(1\right ) .
\end{aligned}\ ] ] therefore , holds .
+ * stochastic error term : * we will show the rate of stochastic error term : @xmath453 firstly , @xmath454 however , @xmath455 which implies @xmath456 because of assumption ( a5 ) .
similar the counterpart in the proof of theorem 1 , we have @xmath457\\ = & \sum_{k=1}^p[\hat{\alpha}_{k}\left(t
/ t\right)-\alpha_{k}\left(t / t\right)]^{2}\sum_{l=1}^{j_{k , a}}e[\psi_{kl}(x_{t , t}^{\left(k\right)})]^2 . \end{aligned}\ ] ] furthermore , @xmath458
^ 2\preceq \sum_{l=1}^{j_{k , a}}e[\psi_{kl}(x_{t , t}^{\left(k\right ) } ) -\psi_{kl}(x_{t}^{\left(k\right)}\left(t / t\right))]^{2 } + \sum_{l=1}^{j_{k , a}}e[\psi_{kl}(x_{t}^{\left(k\right)}\left(t / t\right))]^2 . \end{aligned}\ ] ] by assumption ( a1 ) , @xmath459^{2 } = & j_{k , a}e[b_{kl , a}(x_{t , t}^{\left(k\right)})-b_{kl , a}(x_{t}^{\left(k\right)}\left(t / t\right))]^{2}\\ \preceq & j_{k , a}e|x_{t ,
t}^{\left(k\right)}-x_{t}^{\left(k\right)}\left(t / t\right)|^{2}=o\left(j_{k , a}/t^{2}\right ) \end{aligned}\ ] ] assumption ( a2 ) leads to @xmath460 ^ 2 \asymp \sum_{l=1}^{j_{k , a}}\int\psi_{kl}^2\left(z\right)\mathrm{d}z \leq \int\big[\sum_{l=1}^{j_{k , a}}\psi_{kl}\left(z\right)\big]^{2}\mathrm{d}z = j_{k , a}. \end{aligned}\ ] ] therefore , @xmath461^{2 } = o\left(j_{k , a}+j_{k , a}^{2}/t^{2}\right)=o\left(k_{a}+k_{a}^{2}/t^{2}\right),$ ] which yields @xmath462 similarly , @xmath463\\ = & \frac{1}{t^{2}}\sum_{t=1}^{t}\delta\left(t / t\right)^{\tau } e[\mathbf{\psi}\left(\mathbf{x}_{t , t}\right)\mathbf{\psi}\left(\mathbf{x}_{t , t}\right)^{\tau}]\delta\left(t / t\right ) .
\end{aligned}\ ] ] note that @xmath464 = & e\big[\sum_{l=1}^{j_{k , a}}\psi_{kl}^{2}(x_{t , t}^{\left(k\right)})\big]\leq\int \big\{\sum_{l=1}^{j_{k , a}}\psi_{kl}\left(z\right)\big\}^{2}f_{x_{t , t}}^{\left(k\right ) } \left(z\right)\mathrm{d}z\\=&j_{k , a } , \end{aligned}\ ] ] where @xmath465 is the marginal density of @xmath44-th component of @xmath466
so , @xmath467=diag\left(j_{k , a}\right)_{k=1}^{p}=k_{a}i_{p}\ ] ] and @xmath468 which completes the proof of . + * proof for theorem [ cadditive ] : * * without loss of generality , we assume the true model is @xmath469 let @xmath470 and @xmath471 as the collection of all functions having form @xmath472 it is sufficient to show @xmath473 for any @xmath474 such that @xmath475 and for any @xmath476 where @xmath477 such that @xmath478 + for the sake of convenient presentation , we also denote @xmath479 as @xmath480 if @xmath481 let @xmath482 and @xmath483 then @xmath484 furthermore , @xmath485 where @xmath486 is the inner product of vector @xmath179 and @xmath180 + let @xmath487 we have @xmath488 and @xmath489 where the last step holds since assumption ( a7 ) and theorem [ beta ] .
+ therefore , @xmath490 in which @xmath491 lies between 0 and @xmath492 and @xmath493 the proof of part(i ) is completed since @xmath494 @xmath495 and @xmath496 * according to theorem 6 ( p149 ) of @xcite , under assumption ( a6 ) , there exists @xmath497 and @xmath498 such that @xmath499 let @xmath500 and @xmath501 with @xmath502 next , we will show that for any given @xmath503 there is a sufficiently large @xmath32 such that @xmath504 according to lemma 3 of @xcite , @xmath505 has eigenvalues bounded away from 0 and @xmath332 with probability tending to one as @xmath506 therefore , @xmath507 where we use the fact that @xmath508 and @xmath509 for @xmath510 notice that the first term @xmath511 we also may choose the sufficiently large @xmath32 such that the third term can be dominated by the first term uniformly on @xmath512 finally , we observe that the @xmath513-th element of @xmath514 is given by @xmath515 which is bounded by @xmath516 thus , the second is bounded by @xmath517 which is also dominated by the first term . in combination with the nonnegativity of the first term , we show , which implies with probability at least @xmath518 that there exists a local minimizer in the ball @xmath519 i.e. , @xmath520 again by the property of b - spline , we have that @xmath521 the proof is finished in combination with @xmath499 the proofs for theorem [ cvarying ] is very similar to theorem [ cadditive ] , and thus omitted here .
9 cai , z. , fan , j. and yao , q. ( 2000 ) .
functional - coefficient regression models for nonlinear time series .
_ j. amer .
assoc . _ * 95 * , 941956 .
cai , z. and xu , x. ( 2008 ) .
nonparametric quantile estimations for dynamic smooth coefficient models .
_ j. amer .
assoc . _ * 103 * , 15951607 .
chiang , c .- t . , rice , j. a. andwu , c. o. ( 2001 ) .
smoothing spline estimation for varying coefficient models with repeatedly measured dependent variables .
_ j. amer .
assoc . _ * 96 * , 605619 .
mr1946428 dahlhaus , r. ( 1996a ) .
asymptotic statistical inference for nonstationary processes with evolutionary spectra , _ in : athens conference on applied probability and time series analysis , springer .
_ 145259 .
dahlhaus , r. ( 1996b ) . on the kullback - leibler information divergence of locally stationary processes .
_ stochastic process .
appl . _ * 62 * , 139168 .
dahlhaus , r. ( 1997 ) .
fitting time series models to nonstationary processes .
statist . _
* 25 * , 137 .
dahlhaus , r. , neumann , m. h. , and von sachs .
nonlinear wavelet estimation of time - varying autoregressive processes . _ bernoulli . _ * 5 * , 873906 .
dahlhaus , r. and rao , s. s. ( 2006 ) .
statistical inference for time - varying arch processes .
statist . _
* 34 * , 10751114 .
de boor , c. ( 1978 ) .
_ a practical guide to splines_. springer - verlag , new york .
fan , j. , hrdle , w. and mammen , e. ( 1998 ) .
direct estimation of low dimensional components in additive models .
statist . _
* 26 * , 943971 .
fan , j. and li , r. ( 2001 ) .
variable selection via nonconcave penalized likelihood and its oracle properties .
_ j. amer .
assoc . _ * 96 * , 13481360 .
fan , j. , yao , q. , and cai , z. ( 2002 ) .
adaptive varying - coefficient linear models .
_ j. r. stat .
methodol . _ * 65 * , 5780
. fan , j. and zhang , w. ( 1999 ) .
statistical estimation in varying coefficient models . _ ann .
statist . _
* 27 * 14911518 .
mr1742497 fryzlewicz , p. , sapatinas , t. and rao , s. s. ( 2008 ) .
normalized least - squares estimation in time - varying arch models._ann .
statist . _
* 36 * , 742786 .
fu , w. ( 1998 ) . penalized regression : the bridge versus the lasso .
_ j. comput . graph .
statist . _
* 7 * , 397416 .
hafner , c. m. and linton , o. ( 2010 ) .
efficient estimation of a multivariate multiplicative volatility model
. _ j. econometrics . _ * 159 * , 5573 .
hastie , t. j. and tibshirani , r. j. ( 1990 ) .
_ generalized additive models_. crc press .
hastie , t. j. and tibshirani , r. j. ( 1993 ) varying - coefficient models .
b. _ * 55 * , 757796 .
mr1229881 hoover , d. r. , rice , j. a. , wu , c. o. and yang , l .- p .
nonparametric smoothing estimates of time - varying coefficient models with longitudinal data .
_ biometrika . _
* 85 * , 809822 .
mr1666699 horowitz , j. , klemel , j. and mammen , e. ( 2006 ) .
optimal estimation in additive regression models .
_ bernoulli .
_ * 12 * , 271298 .
huang , j. z. ( 1998 ) .
projection estimation in multiple regression with application to functional anova models
_ * 26 * , 242272 .
huang , j. z. , wu , c. o. and zhou , l. ( 2002 ) .
varying - coefficient models and basis function approximations for the analysis of repeated measurements .
_ biometrika . _ * 89 * , 111128 .
mr1888349 huang , j. z. and shen , h. ( 2004 ) .
functional coefficient regression models for nonlinear time series : a polynomial spline approach .
j. stat . _ * 31 * , 515534 .
mr2101537 huang , j. z. and yang , l. ( 2004 ) .
identification of non - linear additive autoregressive models .
_ j. r. stat .
methodol . _ * 66 * , 463477 .
huang , j. z.,wu , c. o. and zhou , l. ( 2004 ) .
polynomial spline estimation and inference for varying coefficient models with longitudinal data . _ statist .
* 14 * 763788 .
mr2087972 huang , j. , horowitz , j. l. and wei f. ( 2010 ) .
variable selection in nonparametric additive models .
statist . _
* 38 * , 22822313 .
kim , w. , linton , o. b. , and hengartner , n. w. ( 1999 ) .
a computationally efficient oracle estimator for additive nonparametric regression with bootstrap confidence intervals .
_ j. comput . graph .
statist . _
* 8 * , 278297 .
linton , o. and nielsen , j. p. ( 1995 ) .
a kernel method of estimating structured nonparametric regression based on marginal integration .
_ biometrika .
linton , o. ( 1997 ) .
efficient estimation of additive nonparametric regression models
. _ biometrika . _ * 84 * , 469473 .
r. , yang .
l. and hrdle , w. k. ( 2013 ) .
oraclly efficient two - step estimation of generalized aditive model .
_ j. amer .
assoc . _ * 108 * , 619631 .
koo , b. and linton , o. ( 2012 ) .
semiparametric estimation of locally stationary diffusion models
. _ j. econometrics . _ * 170 * , 210233 .
mammen , e. , linton , o. and nielsen , j. ( 1999 ) . the existence and asymptotic properties of a backfitting projection algorithm under weak conditions . _
statist . _
* 27 * , 14431490 .
priestley , m. b. ( 1965 ) .
evolutionary spectra and non - stationary process .
_ j. r. stat .
methodol . _
* 27 * , 204237 .
stone , c. j. ( 1982 ) .
optimal global rates of convergence for nonparametric regression .
statist . _
* 10 * , 10401053 .
mr0673642 stone , c. j. ( 1985 ) .
additive regression and other nonparametric models .
statist . _
* 13 * , 689705 .
stone , c. j. ( 1994 ) .
the use of polynomial splines and their tensor products in multivariate function estimation .
statist . _
* 22 * , 118171 .
tibshirani , r. ( 1996 ) .
regression shrinkage and selection vis the lasso . _ j. r. stat .
soc . ser .
methodol . _ * 58 * , 267288 .
tjstheim , d. and auestad , b. h. ( 1994 ) .
nonparametric identification of nonlinear time series : projections .
vogt , m. ( 2012 ) .
nonparametric regression for locally stationary time series .
_ * 40 * , 26012633 .
wang , l. and yang , l. ( 2007 ) .
spline - backfitted kernel smoothing of nonlinear additive autoregression model .
_ * 35 * , 24742503 .
wang , j. and yang , l. ( 2009 ) .
efficient and fast spline - backfitted kernel smoothing of additive regression model .
_ * 61 * , 663690 .
wu , c. o. , chiang , c. t. and hoover , d. r. ( 1998 ) .
asymptotic confidence regions for kernel smoothing of a varying - coefficient model with longitudinal data .
_ j. amer .
_ * 93 * 13881402 .
mr1666635 xue , l. and yang , l.(2006 ) .
estimation of semi - parametric additive coefficient model
. _ j. statist .
plann . inference .
_ * 136 * , 25062534 .
xue , lan .
consistent variable selection in additive models .
_ _ statist .
sinica.__**19 * * , 12811296 .
yang , l. , hardle , w. and nielsen , j. ( 1999 ) .
nonparametric autoregression with multiplicative volatility and additive mean .
_ j. time series .
_ * 20 * , 579604 .
yuan , m. and lin , y. ( 2006 ) .
model selection and estimation in regression with grouped variables .
_ j. r. stat .
. b. stat .
_ * 68 * , 4967 .
zhang , x. k. , park , b.u . and wang , j. l. time - varying additive models for longitudinal data .
_ j. amer .
* 108 * , 983998 .
zhang , x. k , and wang , j. l. ( 2013 ) .
varying - coefficient additive models for functional data .
_ biometrika .
_ * 102 * , 1532 .
zou , h. ( 2006 ) .
the adaptive lasso and its oracle properties .
_ j. amer .
* 101 * , 14181429 . | nonparametric regression models with locally stationary covariates have received increasing interest in recent years . as a nice relief of `` curse of dimensionality '' induced by large dimension of covariates , additive regression model is commonly used . however , in locally stationary context , to catch the dynamic nature of regression function , we adopt a flexible varying - coefficient additive model where the regression function has the form @xmath0 for this model , we propose a three - step spline estimation method for each univariate nonparametric function , and show its consistency and @xmath1 rate of convergence . furthermore , based upon the three - step estimators , we develop a two - stage penalty procedure to identify pure additive terms and varying - coefficient terms in varying - coefficient additive model .
as expected , we demonstrate that the proposed identification procedure is consistent , and the penalized estimators achieve the same @xmath1 rate of convergence as the polynomial spline estimators .
simulation studies are presented to illustrate the finite sample performance of the proposed three - step spline estimation method and two - stage model selection procedure
. locally stationary process , varying - coefficient additive regression model , b - spline , scad , penalized least squares |
nicmos parallel observations , taken in parallel with one of the other science instruments on hst , has provided us for the first time a wealth of data at near infrared wavelengths with hst resolution .
small background at wavelengths of 0.8@xmath19 and 1.6@xmath19 and hst high angular resolution make nicmos a very efficient instrument in studying the faint galaxy population at high redshifts .
the nicmos parallel imaging and grism observations were both made with camera 3 with a field of view @xmath20
52@xmath21 52@xmath22 .
the imaging data were taken with broad band filters f110w and f160w at 1.1@xmath19 ( j band ) and 1.6@xmath19 ( h band ) .
the grism data has a spectral resolution of 200 per pixel and covers wavelength regions from 0.8@xmath19 to 1.2@xmath19 ( g096 ) and 1.1@xmath19 to 1.9@xmath19 ( g141 ) .
we have reduced and analysed the nic3 parallel imaging data covering @xmath20150 sq .
arcminutes and the grism data in g141 grism @xmath2065 sq . arcminutes .
the recent detections of dust enshrouded galaxies at @xmath23 at sub - millimeter wavelengths ( smail , ivison & blain 1997 ; hughes et al . 1998 ; barger et al .
1998 ; lilly et al .
1998 ) suggest that significant amounts of star formation activity at high redshifts may be obscured .
observations in the rest - frame uv wavelength suffer large uncertainties in extinction corrections .
furthermore , little is known about the properties of normal galaxies in the region between @xmath24 , where neither the 4000 break nor the ly continuum break are easily accessible .
the near - ir offers one means of accessing both redshift indicators and measures of star formation within this critical redshift range .
we have reduced and analysed the nic3 parallel grism g141 data , covering @xmath20 65 sq .
the details of the data reduction can be found in mccarthy et al .
we found a total of 33 emission line galaxies over an effective co - moving volume of @xmath25 mpc@xmath26 for @xmath27 .
the implied co - moving number density of emission line galaxies in the range @xmath28 is @xmath3 mpc@xmath4 , very similar to that of the bright lyman break objects at @xmath5 .
these objects have a median h@xmath1 luminosity of @xmath29 erg sec@xmath7 .
the most , if not all , of the emission lines detected are either h@xmath1 or unresolved blend of h@xmath1+[nii]6583/6548 .
this identification is mostly based on h - band apparent magnitudes , the emission line equivalent widths and the lack of other detected lines within the g141 bandpass .
the median h - band apparent magnitude of @xmath2020.5 ( which corresponds to a l@xmath30 galaxy at @xmath31 ) implies that the possibility of the line being [ oii ] or h@xmath32 is very small .
the redshifts of 6 galaxies in our sample have been confirmed by detection of [ oii]3727 emission in the optical spectra using lris on the keck 10 m telescope ( teplitz et al .
1999 ; malkan et al .
the fraction of agn contamination in our sample is around 10% ; for details , see mccarthy et al .
figure 1 shows the spectra for a subset of galaxies in our sample .
we compute the h@xmath1 luminosity function ( lf ) based on our sample of emission - line galaxies .
we have corrected the incompleteness in the original source catalog using simulations .
the final correction is significant only in the faintest luminosity bin and our main result does not sensitively depend on that .
all of the detailed results are in yan et al .
figure 2 shows our derived h@xmath1 lf at @xmath9 and the local h@xmath1 lf as measured by gallego et al .
this plot shows strong evolution in the h@xmath1 luminosity density from @xmath33 to @xmath10 .
this is no surprise given the evolution in the ultraviolet luminosity density , but our result provides an independent measure of evolution for h@xmath1 emission alone .
the lf is well fit by a schechter function over the range @xmath34 l@xmath35 ) @xmath36 erg sec@xmath7 with l@xmath37 erg sec@xmath7 and @xmath38 mpc@xmath4 for h@xmath8 km s@xmath7 mpc@xmath7 and q@xmath39 .
the integrated h@xmath1 luminosity density at @xmath10 ( our median @xmath40 ) is @xmath11 h@xmath12 erg s@xmath7 mpc@xmath4 , @xmath2014 times greater than the local value reported by gallego et al .
( 1995 ) .
we converted the integrated h@xmath1 luminosity density to a star formation rate ( sfr ) using the relation from kennicutt ( 1999 ) : @xmath41 .
this assumes case b recombination at @xmath42 k and a salpeter imf ( @xmath43 ) .
this conversion factor is about 10% smaller than the value listed in kennicutt ( 1983 ) , the difference reflecting updated evolutionary tracks .
while different choices of stellar tracks introduce modest uncertainties in the conversion of uv and h@xmath1 luminosities to star formation rates , the choice of different imfs lead to rather large differences . to make consistent comparisons between our results and those in the literature
derived from 1500 and 2800 uv continuum luminosity densities , we adopt the relation from kennicutt ( 1999 ) : @xmath44 .
this relation is appropriate for the salpeter imf used to derive the h@xmath1 conversion factor . in figure 3
, we plot uncorrected published measurements of the volume - averaged global star formation rate at various epochs .
our result is shown as a filled circle .
the star formation rates shown in figure 3 are calculated from the luminosity densities integrated over the entire luminosity functions , for both h@xmath1 and the uv continuum .
lilly et al . and connolly et al
. assumed a faint end slope of @xmath45 for the uv continuum luminosity functions at @xmath46 .
the 1500 continuum luminosity function at @xmath47 measured from lyman break galaxies by steidel et al .
( 1998 ) has a faint end slope of @xmath48 .
the clear trend for the longer wavelength determinations of the star formation rate to exceed those based on uv continua is one of the pieces of evidence for significant extinction at intermediate and high redshifts .
the amplitude of the extinction correction is quite uncertain our measurement spans @xmath49 , overlapping with the connolly et al .
photometric redshift sample and allowing a direct comparison between the observed 2800 luminosity density and that inferred from h@xmath1 .
our h@xmath1-based star formation rate is three times larger than the average of the three redshift bins measured by connolly et al .
( 1997 ) .
the star formation rates derived from line or continuum luminosities depend strongly on the choice of imf , evolutionary tracks , and stellar atmospheres that are input into a specific spectral evolution model .
the relevant issue for the present discussion is the ratio of the star formation rates derived from h@xmath1 and the 2800 continuum .
this ratio differs significantly for the scalo and salpeter imfs and is a function of metallicity ( glazebrook et al .
our choice of the salpeter imf comes close to minimizing the difference between the published uv- and our h@xmath1-derived star formation rates .
the use of a scalo imf and solar metallicity would increase the apparent discrepancy by a factor of @xmath50 .
the only model considered by glazebrook et al . that further reduces the h@xmath1/2800
star formation ratio is the salpeter imf with the gunn & stryker ( 1983 ) spectral energy distributions , and this model still leaves us with a factor of @xmath50 enhancement in apparent star formation activity measured at h@xmath1 .
if we attribute the entire difference to reddening , the total extinction corrections at 2800 and h@xmath1 are large and model - dependent .
the calculation is sensitive to the relative geometry of the stars , gas and dust , as well as the adopted reddening curve . in the extreme case of a homogeneous foreground screen and a mw or lmc reddening curve ,
we derive a@xmath51 magnitudes . in local starburst galaxies , differential extinction between the nebular gas , and stellar continuum , and scattering
produce an effective reddening curve that is significantly grayer than the mw or lmc curves ( calzetti , kinney & storchi - bergmann 1994 ; 1996 ; calzetti 1997 ) .
the calzetti reddening law ( calzetti 1997 ) is appropriate for geometries in which the stars , gas and dust are well mixed . in this model ,
our estimate of the dust extinction at 2800 is one to two magnitudes larger than in the simple screen case , and is an uncomfortably large correction compared to results from other methods .
several groups have discovered a population of galaxies with extremely red colors r@xmath52k @xmath53 or 6 .
however , the statistics of eros is still very poor and the nature of these objects remains unclear .
the central issue is to understand whether eros are intrinsically red stellar systems formed at high redshifts in a monolithic collapse or highly reddened starburst galaxies at low to moderate redshifts .
the detection of strong sub - mm continuum from ero hr 10 ( dey et al .
1999 ; cimmati et al . 1998 ) provides conclusive evidence that some of eros , if not all , are indeed dusty , starburst galaxies at a star formation rate of @xmath54 at moderate redshifts ( @xmath55 ) .
we have obtained deep ground - based optical images of 27 nic3 fields , yielding @xmath20 20 square arcminutes . among these fields ,
we have identified about a dozen of eros with r@xmath52h @xmath56 and h brighter than 20.6 .
the surface density of eros with h @xmath57 20.6 and r@xmath52h @xmath56 is roughly 0.6/sq
. arcminutes .
we also found some evidence that eros are highly clustered .
among 27 nic3 fields , we found 2 clusters of eros .
figure 4 shows a cluster of eros in a single nic3 field ( 0.75@xmath58 ) where we also have k - band magnitudes .
we thank the staff of the space telescope science institute for their efforts in making this parallel program possible .
this research was supported , in part , by grants from the space telescope science institute , go-7498.01 - 96a and p423101 .
barger , a.j .
, cowie , l.l . ,
sanders , d.b . , et al .
1998 , nature , 394 , 248 connolly , a. , szalay , a. s. , dickinson , m. , et al .
1997 , apj , 487 , l13 calzetti , d. , kinney , a.l . & storchi - bergmann , t. 1994 , apj , 429 , 582 calzetti , d. 1997 , aj , 113 , 162 gallego , j. , zamorano , j. , aragon - salamanca , a. , rego , m. 1995 , apjl , 459 , 1 glazebrook , k. , blake , c. , economou , f. , lilly , s. et al .
1998 , astro - ph/9808276 hughes , d. , et al .
1998 , nature , 394 , 241 .
kennicutt , r.c . 1999 ,
araa , in press .
astro - ph/9807187 kennicutt , r.c .
1983 , apj , 272 , 54 lilly , s.j . ,
le fevre , o. , hammer , f. & crampton , d. 1996 , apj , 460 , l1 lilly , s.j . ,
eales , s.a . , et al .
1999 , , in press .
astro - ph/9901047 mccarthy , p. , yan , l. , 1999 , , august issue , in press , astro - ph/9902347 steidel c. , dickinson , m. , et al .
1998 , apj , in press .
astro - ph/9811399 teplitz , h. , et al .
1999 , in prep .
tresse , l. , maddox , s. 1998 , , 495 , 691 yan , l. mccarthy , p. , 1999
, , july issue , in press , astro - ph/9904427 | we present the results of a survey which utilizes the nicmos camera 3 parallel grism and imaging observations of random fields .
we have identified @xmath0 h@xmath1 emission - line galaxies at @xmath2 .
the inferred co - moving number density of these objects is @xmath3 mpc@xmath4 , very similar to that of the bright lyman break objects at @xmath5
. the mean star formation rate of these galaxies 21 m@xmath6 yr@xmath7 for h@xmath8km / s / mpc . using this sample , we derived the h@xmath1 luminosity function ( lf ) at @xmath9 .
the integrated h@xmath1 luminosity density at @xmath10 is @xmath11 h@xmath12 erg s@xmath7 mpc@xmath4 , approximately 14 times greater than the local value reported by gallego et al .
( 1995 ) .
the volume averaged star formation rate at @xmath13 is 0.13 m@xmath6 yr@xmath7 mpc@xmath4 without correction for extinction .
the sfr derived at @xmath14 is a factor of @xmath15 higher than that deduced from 2800 continua .
we believe that this difference is largely due to dust extinction .
the implied total extinction at 2800 is in the range of @xmath16 magnitude .
however , the precise determination of the total extinction is sensitive to the model assumptions .
deep ground - based vri images of the nicmos fields have revealed roughly a dozen of extremely red objects ( eros ) with @xmath17 and @xmath18 brighter than 20.6 .
the surface density of these objects is around 0.6 per square arcminutes . |
bellantoni and cook@xcite introduced a class @xmath0 of functions on finite binary strings .
arguments of each function @xmath1 in the class @xmath0 are divided to _ normal _
arguments @xmath2 and _ safe _
arguments @xmath3 , and denoted @xmath4 .
let @xmath5 denote the empty string , and @xmath6 the concatenated string obtained from the binary string @xmath7 and @xmath8 .
the class @xmath0 is generated from initial functions ( projections , zero , binary successors @xmath9 , the predecessor @xmath10 , @xmath11 , the conditional(parity test ) @xmath12 if @xmath13 , @xmath14 otherwise ) by operating safe composition @xmath15 and predicative recursion on notation @xmath16 and @xmath17 for @xmath8 .
it is shown in @xcite that the polynomial time computable functions are exactly those functions in @xmath0 having no safe arguments .
it seems to me that the class @xmath0 not only characterize the class of the polynomial time computable functions , but also is of foundational importance since each function in @xmath0 is computable _
predicatively_. by computability we mean that each object reaches to a canonical form by some computations
. in a predicatively justifiable computation we can not assume a computation to be a completed process in advance since it involves infinite searches or at least the notion of finite computations ( completed processes ) in general .
for example a substitution of @xmath18 in a normal argument , @xmath19 is hard to justify predicatively since it assumes a hypothetical computation of @xmath18 to be completed . on the other side
, we see that a computation process of each function @xmath4 in @xmath0 can be obtained by imitating the generating process of normal arguments @xmath2 . in the computation process
the safe arguments @xmath3 act only as _ names_. in other words we do nt need to know the values ( canonical forms ) of @xmath3 , but need the values of normal arguments @xmath2 from which we know how the arguments are generated from @xmath5 by rules @xmath20 . in this sense
the predicative recursion on notation is justifiable predicatively .
this observation was implicit in our joint work@xcite with g. moser to design a path order pop for computations in @xmath0 .
we now ask how to define predicatively justifiable computations on _ sets _ ?
contrary to binary strings , there seem no canonical forms of sets even for hereditarily finite sets unless we assume , e.g. , the axiom of constructibility .
let us approach modestly .
first pick some functions on safe arguments to generate sets such as pairing and unions . then applying safe composition and a safe set recursion @xmath21 to get a class of functions on sets .
each set is inductively generated , i.e. , the epsilon relation @xmath22 is well founded .
safe set recursion is close to the idea of predicatively computable functions since we do nt need to know the values of intermediate terms @xmath23 to continue the computations of @xmath24 .
thus a class pcsf of predicatively computable set functions is obtained in section [ sec : pcsf ] .
the class pcsf is a subclass of the class srsf of safe recursive set functions due to a. beckmann , s. buss and s. friedman @xcite .
their joint work motivates ours , and is reported in section [ sec : bbf ] . in section [ sec : string ]
it is shown that each polynomial time computable function on finite binary strings is in the class pcsf , cf .
lemma [ lem : ptimepcsf ] . in section [ sec : hf ] the size of pcsf function @xmath4 is seen to be bounded by a polynomial in the sizes of normal arguments @xmath2 , and to depend linearly on the safe arguments @xmath3 , cf .
theorem [ th : size ] . from this
we see readily that each pcsf function @xmath25 on finite binary strings is polynomial time computable , cf .
corollary [ th : main ] .
a. beckmann , s. buss and s. friedman @xcite introduced a class srsf of _ safe recursive set functions_. the class srsf is obtained from gandy - jensen rudimentary set functions on _ safe arguments _ by safe composition scheme and predicative set ( primitive ) recursion scheme a l bellantoni - cook . (
projection ) : : @xmath26 ( difference ) : : @xmath27 ( pair ) : : @xmath28 ( bounded union ) : : @xmath29 ( safe composition ) : : @xmath30 ( predicative set recursion ) : : @xmath31 they investigate definability and complexity of safe recursive functions . 1
. for each @xmath32 there exists a polynomial function @xmath33 on ordinals such that @xmath34 .
2 . a set - theoretic function @xmath25 on infinite ranks
@xmath2 is in srsf iff it is @xmath35-definable on @xmath36 for an @xmath37 , where for ordinals @xmath38 and sets @xmath39 @xmath40 denotes the @xmath41-hierarchy relativized to @xmath39 , and @xmath42 the transitive closure of @xmath39 .
3 . for each @xmath32 there exists a polynomial function @xmath43 such that @xmath44 , where @xmath45 denotes the cardinality of sets @xmath39 .
4 . under a natural encoding of finite binary strings ,
@xmath32 on finite strings are exactly the functions computed by alternating turing machines running in exponential time with polynomially many alternations .
it seems to me that it is hard to justify the class srsf predicatively .
the problem lies in * ( bounded union ) * since it requires us to know _ all _ of the elements @xmath46 in the set @xmath47 in safe argument .
however we do nt know its _ value _ , but only know its _ name _ of @xmath47 . therefore collecting all the elements of sets in safe argument
might not be in the idea of predicatively justifiable computations .
let me propose a subclass pcsf of _ predicatively _ computable set functions .
first a subclass @xmath48 of pcsf is introduced .
each function @xmath1 in the subclass @xmath48 has no normal arguments @xmath49 .
initial functions in @xmath48 are * ( projection ) * on safe arguments , @xmath50 , * ( pair ) * , * ( null ) * , * ( union ) * , and * ( conditional@xmath51)*. ( null ) : : @xmath52 ( union ) : : @xmath53 ( conditional@xmath51 ) : : @xmath54 the class @xmath48 is closed under composition @xmath55 , and * ( safe separation)*. ( safe separation ) : : @xmath56 the class pcsf is then obtained from @xmath48 and * ( projection ) * @xmath57 by operating * ( safe composition ) * and * ( predicative set recursion)*. a relation @xmath58 is in pcsf if its characteristic function @xmath59 is in the class .
( @xmath60 if @xmath58 , @xmath61 otherwise . )
+ * remark*. it is open , but unlikely the case that the class pcsf is closed under the following safe separation scheme . @xmath62 recall that a function @xmath1 is said to be _ simple _ iff @xmath63 is @xmath64 for any @xmath64-relations @xmath65 . as in @xcite
we see the following proposition .
[ prp : simple ] each @xmath66 is a simple function .
hence @xmath1 is a @xmath64-function in the sense that its graph is @xmath64 . as in @xcite
we see the following proposition .
proposition [ prp : rathjen22].[prp : rathjen22.2a ] tells us that a relation is in @xmath48 iff it is rudimentary . as in set - theoretic literature , @xmath67 , which is the unique element @xmath68 such that @xmath69 if such a @xmath68 exists , and @xmath70 .
[ prp : rathjen22 ] 1 .
[ prp : rathjen22.-1 ] @xmath71 is in @xmath48 .
[ prp : rathjen22.0 ] if @xmath72 is in pcsf , then so is @xmath1 , where @xmath73 .
[ prp : rathjen22.1 ] if @xmath74 are in pcsf , then so is @xmath1 , where if @xmath58 , then @xmath75 else @xmath76 .
[ prp : rathjen22.2 ] the class of relations in pcsf is closed under boolean operations .
[ prp : rathjen22.2a ] + a relation @xmath77 is @xmath64 iff its characteristic function @xmath78 is in @xmath48 .
[ prp : rathjen22.2b ] @xmath79 is in @xmath48 for the @xmath64-relation @xmath80 where @xmath81 .
[ prp : rathjen22.4 ] if @xmath82 is in pcsf , then so is @xmath83
[ prp : rathjen22.3](cf . *
( bounded union)*. ) + if @xmath82 is in pcsf , then so is @xmath1 , where @xmath84 .
[ prp : rathjen22.5 ] if @xmath85 are in pcsf , then so are @xmath86 , where @xmath87 and @xmath88 . 10 .
[ prp : rathjen22.2c ] @xmath89 and @xmath90 are in @xmath91
[ prp : rathjen22.7 ] the transitive closure @xmath92 and the rank @xmath93 are in pcsf . 12 .
[ prp : rathjen22.4a ] if @xmath82 is in pcsf , then so is + ( predicative function recursion ) : : @xmath94 + where @xmath95 .
+ conversely any pcsf - function is generated from @xmath48-functions and * ( projection ) * by * ( safe composition ) * and * ( predicative function recursion)*. 13 .
[ prp : rathjen22.10 ] let @xmath65 be a @xmath64-relation .
assume that @xmath96 $ ] .
let @xmath97 iff @xmath98 .
then @xmath1 is in pcsf . _
proof_. [ prp : rathjen22].[prp : rathjen22.-1 ] . @xmath99 by * ( safe separation)*. + [ prp : rathjen22].[prp : rathjen22.2a ] . if @xmath100 , then @xmath101 is a @xmath64-relation by proposition [ prp : simple ] . conversely consider a relation @xmath102 with a @xmath64-relation @xmath103 .
then @xmath104 is in @xmath48 .
hence so is @xmath105 . for disjunctions @xmath106 use the finite union @xmath107 , and for negations @xmath77 use the conditional @xmath108 .
+ [ prp : rathjen22].[prp : rathjen22.3 ] .
let @xmath109 if @xmath22 .
@xmath110 otherwise , where @xmath22 is in pcsf by * ( conditional@xmath51 ) * and proposition [ prp : rathjen22].[prp : rathjen22.0 ] .
let @xmath111 .
then @xmath112 [ prp : rathjen22].[prp : rathjen22.5 ] . by proposition
[ prp : rathjen22].[prp : rathjen22.3 ] @xmath113 is in pcsf .
then so is @xmath114 .
+ [ prp : rathjen22].[prp : rathjen22.2c ] . by propositions
[ prp : rathjen22].[prp : rathjen22.2b ] and [ prp : rathjen22].[prp : rathjen22.5 ] both @xmath89 and @xmath115 are in @xmath91 .
+ [ prp : rathjen22].[prp : rathjen22.7 ] .
let @xmath116 for @xmath117 .
then @xmath118 and @xmath119 since @xmath120 is transitive , i.e. , @xmath121 . + [ prp : rathjen22].[prp : rathjen22.4a ] . let @xmath122 where @xmath123 .
then @xmath124 is in pcsf by propostions [ prp : rathjen22].[prp : rathjen22.4 ] , [ prp : rathjen22].[prp : rathjen22.5 ] and [ prp : rathjen22].[prp : rathjen22.2c ] .
suppose @xmath125 then we have for @xmath126 and @xmath22 , @xmath127 .
hence @xmath128 is in pcsf .
it remains to show ( [ eq : rathjen22.4a ] ) by induction on @xmath39 . by ih(=induction hypothesis ) we have @xmath129 .
hence by the definition of @xmath1 we have @xmath130 .
this shows ( [ eq : rathjen22.4a ] ) , and pcsf is closed under * ( predicative function recursion)*. conversely if @xmath1 is defined from @xmath82 by * ( predicative set recursion ) * , @xmath21 , then @xmath131 .
@xmath132 is in pcsf by proposition [ prp : rathjen22].[prp : rathjen22.2c ] @xmath1 is defined from @xmath133 by * ( predicative function recursion)*. + [ prp : rathjen22].[prp : rathjen22.10 ] . by propositions [ prp : rathjen22].[prp : rathjen22.2 ] and [ prp : rathjen22].[prp :
rathjen22.2a ] , @xmath64-relation @xmath134 defines a relation @xmath135 in pcsf .
so is @xmath136 by proposition [ prp : rathjen22].[prp : rathjen22.5 ] .
let @xmath138 denote the set of all _ hereditarily finite sets_. let us encode finite ( binary ) strings by hereditarily finite sets , @xmath139 slightly modified from @xcite .
@xmath140 ( @xmath5 is the empty string . ) @xmath141 .
@xmath142 . for example , @xmath143 .
[ lem : ptimepcsf ] for each polynomial time computable function @xmath144 there exists a function @xmath145 in pcsf such that for any finite strings @xmath146 @xmath147 _ proof_. let @xmath0 denote the class of safe recursive functions on binary finite strings in @xcite .
we show inductively for each @xmath148 there exists a function @xmath145 in pcsf such that for any finite strings @xmath149 , @xmath150 . for the binary successor @xmath9 , @xmath151 does the job . for the predecessor @xmath10 , @xmath11
, first let by propositions [ prp : rathjen22].[prp : rathjen22.1 ] and [ prp : rathjen22].[prp : rathjen22.2a ] @xmath152 \\ 0 & \mbox{{\rm otherwise } } \end{array } \right.\ ] ] namely @xmath153
. then let @xmath154 by proposition [ prp : rathjen22].[prp : rathjen22.-1 ] .
we have @xmath155 since @xmath156 .
next consider conditional(parity test ) @xmath12 if @xmath13 , @xmath14 otherwise .
since @xmath157 and @xmath158 , @xmath159 enjoys @xmath160 .
the case when @xmath161 is defined from @xmath162 by predicative composition is seen from ih .
finally consider predicative recursion on notation . @xmath16 and @xmath17 for @xmath8 .
let @xmath163 and @xmath164 be functions in pcsf for @xmath165 and @xmath166 , resp .
define @xmath145 as follows .
@xmath167 . let @xmath8 .
suppose @xmath169 , @xmath170 , and @xmath171 .
then let + @xmath172 .
@xmath173 for @xmath174 .
@xmath175 . otherwise @xmath176 .
then we compute for @xmath8 , if @xmath177 , + @xmath178 , and @xmath179 .
@xmath137 + * remark*. lemma [ lem : ptimepcsf ] holds also for a subclass @xmath180 .
the initial functions in the subclass are projections @xmath57 , @xmath181 , @xmath182 , @xmath183 in the proof of lemma [ lem : ptimepcsf ] , @xmath184 and @xmath185 .
the class @xmath180 is closed under * ( safe composition ) * and the scheme @xmath83 , cf .
proposition [ prp : rathjen22].[prp : rathjen22.4 ] .
moreover * ( safe separation ) * is needed only in defining @xmath186 , @xmath187 ( proposition [ prp : rathjen22].[prp : rathjen22.2b ] ) and @xmath183 for lemma [ lem : ptimepcsf ] .
namely the separation @xmath188 , @xmath189 , @xmath190\}$ ] and @xmath191 .
let us restrict our attention to hereditarily finite sets @xmath138 .
@xmath192 denote hereditarily finite sets .
each function @xmath1 in pcsf is a function on @xmath138 when it is restricted to @xmath138 .
the size of @xmath4 is seen to be bounded by a polynomial in the sizes of normal arguments @xmath2 , and depend _ linearly _ on the safe arguments @xmath3 , cf .
theorem [ th : size ] .
this readily yields the converse of lemma [ lem : ptimepcsf ] , cf .
corollary [ th : main ] .
let us introduce some abbreviations to state and shorten the proof of the following lemma .
for hereditarily finite sets @xmath213 , let us denote @xmath196 , @xmath214 , @xmath215 , @xmath216 , and @xmath217 .
let @xmath218 , and @xmath219 . for sequences
@xmath220 , @xmath221 of natural numbers @xmath222 , let @xmath223 .
let @xmath224 be sequences for lists @xmath225 of functions @xmath226 .
each @xmath227 is a sequence of natural numbers in length @xmath228 .
also let @xmath229 be lists of hereditarily finite sets .
+ then let @xmath230 and @xmath231 . let @xmath232 furthermore for lists @xmath233 of ( definitions of ) functions @xmath234 , let @xmath235 for any list @xmath214 of hereditarily finite sets @xmath238 , any list of functions @xmath239 of @xmath240 , any list @xmath201 of @xmath241 , any hereditarily finite set @xmath0 , any sequences @xmath224 for functions @xmath242 and any sequences @xmath243 for functions @xmath244 the following holds : lemma [ prp : coverp ] yields theorem [ th : size ] as follows . for @xmath246 with the list @xmath247 , @xmath248 , i.e. , the projection @xmath249 , and @xmath250
, we have a polynomial @xmath251 such that for any lists of hereditarily finite sets @xmath252 , @xmath201 , @xmath253 .
. then @xmath255 and @xmath256 for @xmath257 .
+ _ proof _ of lemma [ prp : coverp ] .
let us define a natural number @xmath258 for each function @xmath246 as follows .
first @xmath259 if @xmath1 is one of null , pair , projections @xmath57 , union , @xmath260 and functions defined by * ( safe separation)*. second @xmath261 if @xmath1 is defined by * ( safe composition ) * from @xmath262 , @xmath263 .
third @xmath264 if @xmath1 is defined by * ( predicative set recursion ) * from @xmath82 . finally for lists @xmath233 of functions , let @xmath265 .
let @xmath270 and @xmath271 .
by ih we have a polynomial @xmath272 for the list @xmath273 . + * ( null ) * if @xmath1 is @xmath274 , then @xmath275 .
+ * ( projection ) * + if @xmath1 is a projection @xmath276 , then @xmath277 is one of @xmath278 and @xmath279 . in the former case @xmath280 , while in the latter case @xmath275 . + * ( pair ) * + if @xmath1 is the pair @xmath281 , then @xmath282 .
+ * ( union ) * + if @xmath1 is the union @xmath283 , then @xmath275 .
+ * ( conditional@xmath51 ) * + if @xmath1 is the conditional @xmath284 , then @xmath275 .
+ * ( safe separation ) * + if @xmath1 is defined from @xmath82 by * ( safe separation ) * @xmath285 , then @xmath275 . + * ( safe composition ) * + consider the case when @xmath1 is defined from @xmath82 , @xmath286 and @xmath287 by * ( safe composition ) * , + @xmath288 , where each @xmath289 is a pcsf - function .
let @xmath299 for @xmath300 , and @xmath301 , where @xmath302 is substituted for each variable @xmath303 in the list @xmath304 , @xmath305\leq \prod_{i } ct(z_{i})$ ] , and @xmath306 .
then @xmath307 * ( predicative set recursion ) * + consider the case when @xmath1 is defined from @xmath82 by * ( predicative set recursion ) * , @xmath308 . by proposition
[ prp : rathjen22].[prp : rathjen22.5 ] there exists a pcsf - function @xmath309 .
now we show that any function @xmath246 is polynomial time computable when we restrict @xmath1 to @xmath138 . to be specific ,
let us encode hereditarily finite sets first by dags(directed acyclic graphs ) , and then encode dag s by natural numbers . in what follows a dag with root
is simply said to be a dag .
@xmath339 designates that there is an edge from @xmath340 to @xmath47 . from the condition ( [ df : dag2 ] ) in definition [ df : dag ] we see that @xmath163 is acyclic . for a dag @xmath331
we write @xmath341 , @xmath342 and @xmath343 . for nodes
@xmath344 , @xmath345 denotes a dag @xmath346 defined by @xmath347 , and for @xmath348 , @xmath349 iff there exists a path from @xmath340 to @xmath47 in @xmath163 , i.e. , there is a sequence @xmath350 such that @xmath351 , @xmath352 and @xmath353 .
the _ rank _ @xmath354 of nodes @xmath340 in @xmath163 is defined by @xmath355 , where @xmath356 .
then the rank of @xmath163 is defined by @xmath357 . while the _ length _ @xmath358 of the longest path from @xmath335 to @xmath340 is defined by @xmath359 . 1 .
each node @xmath344 encodes a hereditarily finite set @xmath361 defined by recursion on ranks @xmath354 : @xmath362 dag @xmath163 encodes a hereditarily finite set @xmath363 .
@xmath344 and @xmath364 are _ bisimilar _ ( with respect to @xmath365 ) , denoted @xmath366 or simply @xmath367 iff @xmath368 .
+ @xmath163 and @xmath369 are _ _ bisimilar _ _ , denoted @xmath370 iff @xmath371 , i.e. , @xmath372 .
@xmath163 is _ fully collapsed _
iff for any nodes @xmath373 in @xmath163 , if @xmath374 then @xmath375 .
we assume a feasible encoding of finite sequences of natural numbers .
@xmath377 denotes the code of sequence @xmath378 of natural numbers @xmath379 .
@xmath380 denotes the code of dag @xmath331 .
specifically @xmath381 , where for nodes @xmath382 , its code @xmath383 , and for edges @xmath384 , @xmath385 , where @xmath386 and @xmath387 .
it is plain to see that to be a code of a dag is polynomial time decidable , and ranks @xmath354 and lengths @xmath358 of nodes @xmath340 in @xmath163 are polynomial time computable from @xmath388 and @xmath340 .
moreover given a code @xmath389 of dag and a node @xmath390 , one can compute the code @xmath391 in polynomial time .
therefore let us identify dag @xmath163 with its code @xmath389 , and , e.g. , say that @xmath345 is polynomial time computable . 1 .
[ prp : bisimilarfullcollapsed1 ] bisimilarity in dag s is polynomial time decidable .
[ prp : bisimilarfullcollapsed15 ] there is a polynomial time function @xmath65 such that for any given dag @xmath163 , @xmath398 and @xmath163 are bisimilar and @xmath398 is balanced with @xmath399 .
moreover if @xmath163 is fully collapsed , then so is @xmath398 .
[ prp : bisimilarfullcollapsed2 ] there is a polynomial time function @xmath46 such that for any given dag s @xmath400 , @xmath401 is a fully collapsed dag such that @xmath402 _ proof_. [ prp : bisimilarfullcollapsed].[prp : bisimilarfullcollapsed1 ] .
let @xmath403 iff there exists an edge @xmath404 .
then @xmath370 iff @xmath405 .
a bisimilarity test is performed at most @xmath406 times .
+ [ prp : bisimilarfullcollapsed].[prp : bisimilarfullcollapsed2 ] .
we can assume that sets @xmath407 are disjoint , for otherwise replace @xmath408 by @xmath409 , where @xmath410 and @xmath411 for the bijective pairing @xmath412 .
note that @xmath413 .
let @xmath414 , and @xmath163 be the joined dag .
@xmath415 , @xmath416 and @xmath417 .
clearly @xmath418 . by recursion on ranks
define dag s @xmath419 so that each @xmath420 and any bisimilar pair @xmath421 has ranks larger than @xmath422 , @xmath423 , as follows . let @xmath424 .
assume that @xmath425 has been defined .
consider @xmath426 of rank @xmath422 and its bisimilar class @xmath427 , and let us share nodes in @xmath428 .
note that for @xmath429 and any @xmath68 , @xmath430 by the construction .
let @xmath431 .
delete every nodes in @xmath428 except @xmath379 , and each edge @xmath432 for @xmath433 is switched to a new edge @xmath434 , where @xmath435 .
the switchings are performed for each @xmath426 of rank @xmath422 .
the resulting dag @xmath164 is bisimilar to @xmath425 , and @xmath436 .
[ lem : hfptime ] for each @xmath246 , there is a polynomial time computable function @xmath145 such that for any balanced and fully collapsed dag s @xmath438 , @xmath439 , @xmath440 is a code @xmath441 of a balanced and fully collapsed dag @xmath442 such that @xmath443 . * ( union ) * for dag @xmath163 , a dag @xmath369 such that @xmath444 is obtained by @xmath445 , @xmath446 and for @xmath447 , @xmath448 iff either @xmath404 or there is a @xmath449 such that @xmath450 and @xmath451 . next consider * ( safe composition ) * @xmath452 if all of @xmath82 , @xmath286 and @xmath287 are polynomial time computable on dag s , then so is @xmath1 . finally consider * ( predicative set recursion ) * @xmath453 assume that @xmath454 are hereditarily finite sets @xmath455 for fully collapsed dag s @xmath456 .
let us describe informally a polynomial time computation of a fully collapsed dag @xmath41 such that @xmath457 .
by recursion on ranks @xmath354 of nodes @xmath340 in ` circuit ' @xmath163 , assign a dag @xmath458 such that @xmath459 to @xmath340 as follows .
if @xmath340 is the leaf , i.e. , the node of outdegree zero , then @xmath458 is a fully collapsed dag such that @xmath460 .
next consider the case when @xmath340 is not a leaf , and let @xmath461 be the sons of @xmath340 in @xmath163 : @xmath462 .
assume that for each son @xmath463 a fully collapsed dag @xmath464 is attached to @xmath463 so that @xmath465 .
then by proposition [ prp : bisimilarfullcollapsed].[prp : bisimilarfullcollapsed2 ] compute a fully collapsed dag @xmath466 , and then let @xmath458 be a fully collapsed dag such that @xmath467 .
let us estimate roughly the number of computation steps .
the number of computations of the function @xmath82 is @xmath468 . by theorem [ th : size ]
we have a polynomial @xmath43 such that @xmath469 since all dag s are balanced and fully collapsed , we have for a polynomial @xmath470 @xmath471 hence each computation of @xmath82 is performed in the number of steps bounded by a polynomial of @xmath397 , @xmath472 and @xmath473 .
moreover the number of computations of @xmath466 is @xmath396 , and each computation of @xmath474 is also performed polynomially in @xmath397 , @xmath472 and @xmath473 . hence the number of computation steps for @xmath41 is bounded by a polynomial of @xmath397 , @xmath472 and @xmath473 .
@xmath137 [ th : main ] suppose a set theoretic function @xmath475 is a function on binary finite strings when we restrict to finite strings : @xmath476 . if @xmath477 , then the function @xmath478 is polynomial time computable . _
proof_. assume @xmath479 , and let @xmath480 .
then @xmath145 is a polynomial time function on @xmath138 in the sense of theorem [ lem : hfptime ] .
since the function @xmath481 and its inverse @xmath482 are polynomial time computable , so is @xmath1 . @xmath137 + * remark*. let @xmath145 be a polynomial time computable function for @xmath246 in theorem [ lem : hfptime ]
. then @xmath145 has to be an ` extensional ' function on dag s .
this means that for any balanced and fully collapsed dag s @xmath438 , @xmath439 @xmath483 1
. it seems to us that the converse holds .
namely let @xmath145 be a polynomial time computable function such that @xmath484 is a code of balanced and fully collapsed dag for any balanced and fully collapsed dag s @xmath438 , and @xmath145 is extensional in the above sense .
then the set - theoretic function @xmath1 on @xmath138 is defined by @xmath485 where @xmath486 and @xmath487 for some ( any ) balanced and fully collapsed dag s @xmath438 and @xmath369 .
+ an affirmative answer to the following problem would show a stronger statement than lemma [ lem : ptimepcsf ] since there are polynomial computable functions mapping binary strings @xmath7 to dag s ( balanced and fully collapsed ) representing @xmath488 , and vice versa . + * problem*. show that the @xmath1 is a restriction of a function in the class pcsf on @xmath138 .
2 . let @xmath489 be a choice function which chooses an element @xmath490 from non - empty sets @xmath340 .
let us set @xmath491 .
it is unlikely the case that there is such a @xmath46 in the class pcsf , nor @xmath46 on @xmath138 is ( extensionally ) polynomial time computable in the sense of theorem [ lem : hfptime ] .
obviously there exists an intensional function @xmath474 which depends on codes .
given dag s @xmath163 , if @xmath492 , then let @xmath493 .
then @xmath494 , and @xmath495 is polynomial time computable , and so is the function @xmath496 .
however @xmath474 is not extensional .
arai , t and moser , g. : proofs of termination of rewrite systems for polytime functions . in : ramanujam ,
r. sen , sandeep(eds . )
fsttcs 2005 : foundations of software technology and theoretical computer science , pp.529 - 540 .
lectute notes in compt .
3821 , springer , berlin ( 2005 ) | inspired from a joint work by a. beckmann , s. buss and s. friedman , we propose a class of set - theoretic functions , predicatively computable set functions .
each function in this class is polynomial time computable when we restrict to finite binary strings .
[ section ] [ theorem]definition [ theorem]proposition [ theorem]lemma [ theorem]corollary |
lorentz and cpt violation has recently received substantial attention as a potential signature for underlying physics , possibly arising from the planck scale .
the most simple lorentz- and cpt - breaking field theory consists of a three dimensional chern - simons term embedded in maxwell s four dimensional classical electrodynamics @xcite .
such model , usually known as carroll - field - jackiw , is part of the standard - model extension ( sme ) , the general field - theory framework for lorentz and cpt tests , which contains the standard model of particle physics and general relativity as limiting cases @xcite .
recent studies in the context of the cpt - odd sector of sme , have been carried out in different areas such as radiative corrections @xcite , nontrivial spacetime topology @xcite , causality @xcite , supersymmetry @xcite , the cosmic microwave background @xcite , general relativity 11 and topological defects , specifically monopole structures was presented in @xcite and , the first investigation about bps vortex solutions in the presence of cpt - even lorentz - violating terms of the sme was performed in refs .
@xcite .
on the other hand , it is well known that the higgs models with the maxwell term , support topologically stable vortex solutions @xcite . with a specific choice of coupling constants ,
minimum energy vortex configurations satisfy first order differential equations @xcite .
when this occur the model presents another interesting feature , which says us that the model is the bosonic sector of a supersymmetric theory @xcite .
the two dimensional matter field interacting with gauge fields whose dynamics is governed by a chern - simons term support soliton solutions jk . when self - interactions are suitably chosen vortex configurations satisfy the bogomolnyi - type equations with a specific sixth order potential @xcite .
another important feature of the chern - simons gauge field is that inherits its dynamics from the matter fields to which it is coupled , so it may be either relativistic @xcite or non - relativistic @xcite .
in addition the soliton solutions are of topological and non - topological nature @xcite .
the relativistic chern - simons - higgs model described in ref.jw , was generalized to include a maxwell term for the gauge field in ref.@xcite .
there , the authors , in order to maintain a notion of self - duality , in which a lower bound for the energy is saturated by solutions to a set of self - duality equations , introduce an additional neutral scalar field .
in this paper we are interested in studying bps vortices in a context of a cpt - odd and lorentz - violating abelian higgs system , involving both the maxwell and chern - simons terms , which is obtained from carroll - field - jackiw theory by dimensional reduction . in particular
, we will show that this model support self - dual vortex solutions , which are different from those obtained in ref .
the difference lies in the fact that in our model the neutral scalar field appears naturally from the dimensional reduction , so that the model is self - dual without the requirement to introduce an additional neutral field such as in ref .
in addition we show that the vortex solution are identical in form to the nielsen - olesen vortices , although , in our model , the vortex possess electric charge .
let us start by considering the lorentz and cpt - violating maxwell - chern - simons theory , proposed by carroll , field , and jackiw cfj @xmath0 where @xmath1 is a four - vector couples to electromagnetic field , which determines a preferred direction in spacetime violating lorentz as well as cpt symmetry , and @xmath2 is the stress tensor defined by @xmath3 . here , the term @xmath4 represents the coupling between the gauge field and an external current .
the metric tensor is @xmath5 and @xmath6 is the totally antisymmetric levi - civita tensor such that @xmath7 .
the lagrangian ( [ ac1 ] ) leads to the following equations of motion @xmath8 where @xmath9 is the dual electromagnetic tensor . in the following
we consider a model composed by the gauge field in ( ac1 ) coupled to a higgs field @xmath10 , \notag\end{aligned}\ ] ] where greek indexes run on @xmath11 , @xmath12 is the covariant derivative and @xmath13 is a self - interacting potential to be determined below .
specifically , we are interested in exploring the solitonic structure of the model obtained from ( [ ac2 ] ) via dimensional reduction . in particular we are motivated by searching of a model with maxwell - chern - simons self - dual vortex solutions different from those present in ref .
@xcite . in order to analyze the @xmath14-dimensional problem it is natural to consider a dimensional reduction of the action by assuming that the fields do not depend on one of the spatial coordinates ,
say @xmath15 . renaming @xmath16 as @xmath17 , leads to an action that can be written as @xcite @xmath18 & & \left .
\frac{{}}{{}}+|d_{\mu } \phi |^{2}-e^{2}n^{2}|\phi |^{2}-v(|\phi |)% \right ] , \notag\end{aligned}\ ] ] where now greek s indexes are @xmath19 and the coefficient @xmath20 play the role of chern - simons s parameter .
the gauss law is @xmath21 where @xmath22 $ ] is the conserved matter current . integrating this equation , over the entire plane
, we obtain the important consequence that any object with charge @xmath23 also carries magnetic flux @xmath24 @xcite : @xmath25 here , we are interested in time - independent soliton solutions that ensure the finiteness of the action ( [ ac3 ] ) .
these are the stationary points of the energy which for the static field configuration reads @xmath26 \notag\end{aligned}\ ] ] since we are motivated by the desire to find self - dual soliton solution , we will choose @xmath27 . for this particular choice , the gauss law ( eqm4 ) takes the simple form @xmath28 this is just the gauss law of the model proposed in ref.@xcite . after integration by parts and using the gauss law , the energy functional
can be rewritten as@xmath29^{2}+|d_{i}\phi
|^{2}+|d_{0}\phi \mp ie\phi n|^{2}\right . \notag \\% [ -0.2 cm ] & & \\[-0.2 cm ] & & \left .
+ \frac{1}{2}(\partial _ { 0}n)^{2}+\frac{1}{2}f_{12}^{2}+n(p_{0}\pm p_{3})f_{12}+v(|\phi |)\right\ } \notag\end{aligned}\ ] ] here , the form of the potential @xmath30 that we choose is motivated by the desire to find self - dual soliton solution and coincides with the symmetry breaking potential of the higgs model @xmath31 to proceed , we need a fundamental identity @xmath32 using this identity and choosing @xmath33 the energy may rewritten as @xmath34^{2}+|d_{0}\phi \mp ie\phi n|^{2}\right .
\notag \\ & & + |(d_{1}\pm id_{2})\phi |^{2}+\frac{1}{2}[f_{12}\pm e(|\phi & & \left .
\pm ev^{2}f_{12}+(\frac{\lambda } { 4}-\frac{e^{2}}{2})(|\phi when the symmetry breaking coupling constant @xmath35 is such that @xmath36 i.e. when the self - dual point of the abelian higgs model is satisfied , the energy ( [ ejp2 ] ) reduce to a sum of square terms which are bounded below by a multiple of the magnitude of the magnetic flux : @xmath37 here , the magnetic flux is determined by the requirement of finite energy .
this implies that the covariant derivative must vanish asymptotically , which fixes the behavior of the gauge field @xmath38 .
then we have @xmath39where @xmath40 is a topological invariant which takes only integer values .
the bound is saturated by the fields satisfying the gauss law and @xmath41 .
this way , the first - order self - duality equations : @xmath42 f_{12}\pm e(|\phi |^{2}-v^{2 } ) & = & 0 , \label{bps2 } \\[0.2 cm ] d_{0}\mp ie\phi n & = & 0 , \label{bps3 } \\[0.2 cm ] ( d_{1}\pm id_{2})\phi & = & 0 , \label{bps4}\end{aligned}\ ] ] for the static field configuration this set of equations reduce to @xmath43 f_{12}\pm e(|\phi |^{2}-v^{2 } ) & = & 0 , \label{bpss2}\\[0.2 cm ] ( d_{1}\pm id_{2})\phi & = & 0 , \label{bpss3}\end{aligned}\ ] ] this set of equations ( [ bps1])([bps4 ] ) is similar to the bogomolnyi equations of the ref.@xcite . in ref.@xcite ,
the magnetic field @xmath44 is related to the fields @xmath45 and @xmath17 in the form @xmath46this is not the situation in our eq .
( [ bps2 ] ) , where the magnetic field only has an explicit dependence on the higgs field .
such difference is also presents in the potential of both models , whereas in ref.@xcite it takes the form @xmath47here the potential defining bps configurations takes the simple form ( pot2 ) with coupling constant in the self - dual point value @xmath48 .
thus , our model only possess topological solitons solutions whereas the maxwell - chern - simons solitons studied in ref.@xcite are both topological and nontopological .
another interesting aspect of our model refers to the set of equations ( [ bpss1])([bpss3 ] ) .
we see that the equations ( [ bpss2 ] ) and ( [ bpss3 ] ) are the bogomolnyi equations of the abelian higgs model @xcite , whereas the equation ( [ bpss1 ] ) is decoupled from ( [ bpss2 ] ) and ( [ bpss3 ] ) and only related the neutral scalar field with the gauge field @xmath49 .
thus , we have vortex solutions which are identical in form to the nielsen - olesen vortices , with the novelty that the gauge field @xmath49 is not required to be gauged to zero , as in the maxwell - higgs model .
specifically , we look for radially symmetric solutions using the standard static vortex ansatz @xmath50 where @xmath40 is the winding number of the vortex configuration . the scalar functions @xmath51 @xmath52 and @xmath53 are regular at @xmath54 : @xmath55where @xmath56 , and satisfy appropriated boundary conditions when @xmath57:@xmath58all boundary conditions will be explicitly established in subsection bbcc . as usual , the magnetic field
is expressed as @xmath59 so the remaining bps equations ( [ bps2 ] ) , ( [ bps4 ] ) , and gauss s law ( [ eqm5 ] ) are rewritten as@xmath60 & \displaystyle{\omega ^{\prime \prime } + \frac{\omega ^{\prime } } { r}-p_{3}b -2e^{2}v^{2}g^{2}\omega = 0\ , , } & \label{bq3}\end{aligned}\]]where the upper sign corresponds to @xmath61 and the lower signal to @xmath62 .
we observe that for fixed @xmath63 , if we have the solutions for @xmath61 , the correspondent solutions for @xmath62 are attained doing @xmath64 .
we can also observe that for fixed @xmath40 , under the change @xmath65 , the solutions go as @xmath66 .
the structure of bps equations and gauss s law given in eqs .
( [ bq1])([bq3 ] ) allows to affirm that the profiles of the higgs field , vector potential and magnetic field are exactly the same as those correspondent to the uncharged vortices of the maxwell - higgs model but now our vortices are electrically charged .
this affirmation is verified by the numerical analysis shown in figs .
[ s_bps ] and [ b_bps ] .
we obtain the behavior of the solutions of eqs .
( [ bq1]-[bq3 ] ) in the neighborhood of @xmath67 using power series method , @xmath68 a\left ( r\right ) & = & n-\frac{e^{2}v^{2}}{2}r^{2}+{\mathcal{\ldots } } \label{bc0_a } \\
\omega \left ( r\right ) & = & \omega _ { 0}+\frac{p_{3}ev^{2}}{4}r^{2}+{\mathcal{% \ldots } } \label{bc0_w}\end{aligned}\ ] ] the asymptotic behavior when @xmath69 is performed by setting @xmath70 , @xmath71 , @xmath72 , with @xmath73 , @xmath74 , @xmath75 infinitesimal functions to be computed . after substituting such forms in eqs .
( [ bq1]-[bq3 ] ) , and solving the linearized set of differential equations , we obtain @xmath76where @xmath77 , a positive real number , is given by @xmath78 from eqs .
( [ bc0_g1 ] ) , ( [ bc0_a ] ) and ( [ inf_1 ] ) , we can see that the behavior , in @xmath54 and @xmath79 , of the fields @xmath80 and consequently of the magnetic field , is exactly the same that of the maxwell - higgs self - dual solution .
we now introduce the dimensionless variable @xmath81 and implement the following changes : @xmath82 & \displaystyle{p}_{{3}}\rightarrow ev{\kappa , ~b\rightarrow ev^{2}\bar{b}% \left ( \rho \right ) , ~~\mathcal{e}\rightarrow v^{2}\bar{\mathcal{e}}\left ( \rho \right ) .
} & \end{aligned}\ ] ] thereby , the equations([bq1])-([bq3 ] ) are written in a dimensionless form as@xmath83 & \displaystyle{\bar{b}=-\frac{\bar{a}^{\prime } } { \rho } = \pm \left ( 1-g^{2}\right ) \
, , } & \label{eqt3 } \\[0.15 cm ] & \displaystyle{\bar{\omega}^{\prime \prime } + \frac{\bar{\omega}^{\prime } } { % \rho } -\kappa \bar{b}-{{2\bar{g}^{2}\bar{\omega}=0}}\ , . } & \label{eqt4}\end{aligned}\ ] ] we have performed the numerical analysis of the eqs .
( [ eqt2][eqt4 ] ) and the resultant profiles are depicted in figs .
[ s_bps][el_bps ] .
there are shown the topological solutions with winding numbers @xmath84 when the chern - simons - like parameter is fixed to be @xmath85 .
the profiles for our model are depicted by blue lines whereas the plots for model of ref.@xcite ( mcsh model ) are presented with red lines .
the winding numbers are represented in the following way : solid lines for @xmath86 , dash - dotted lines to @xmath87 and dotted lines do for @xmath88 .
all legends are summarized in fig .
[ s_bps ] .
[ 1 ] ( blue lines also correspond to maxwell - higgs model).,title="fig:",width=255 ] the figs .
[ s_bps ] and [ a_bps ] depict the profiles of the higgs and vector field , respectively . for @xmath89 ,
the plots for both models are similar , although the difference is more evident for large values of the winding number . in general , for fixed @xmath40 , the profiles for our model saturate more quickly than those from mcsh which are wider .
[ 1 ] ( blue lines also correspond to maxwell - higgs model).,title="fig:",width=255 ] fig .
[ b_bps ] depicts the magnetic field behavior .
the profiles for @xmath89 in both models are lumps centered at origin , however the magnetic field for the mcsh model has a smaller amplitude . on the other hand , for @xmath90 1 , in our model ( blue lines )
, the magnetic field keeps the same amplitude in the origin and develops a plateau around it which becomes greater as @xmath91 is increased .
such behavior is very different from the magnetic field of mcsh model ( red lines ) that form rings whose maximums are located away from the origin .
[ 1 ] ( blue lines also correspond to maxwell - higgs model).,title="fig:",width=255 ] fig .
[ w_bps ] depicts the scalar potential profiles which for @xmath92 always take negative values whenever @xmath40 positive . for @xmath89 , the models present very similar profiles , however , for @xmath93 there are notable differences . in our model ,
the profiles of the scalar potential are lumps centered in the origin whose amplitude is given by @xmath94 .
these are very localized compared with the plots of mcsh model .
it is worthwhile to note that the profiles of mcsh model , for @xmath95 , saturate its amplitude in the value @xmath96 generating a large plateau . [ 1].,title="fig:",width=283 ] fig .
[ el_bps ] shows the electric field behavior . for fixed @xmath97 ,
our profiles develop ring structures near to the origin whose amplitude increases with @xmath40 but far from the origin , the electric field decreases to zero satisfying the boundary conditions .
this behavior differs from the respective mcsh profiles which form rings , centered long away from the origin , whose amplitude not increase with @xmath40 . in general , the amplitude of the electric field are larger than those corresponding to the mcsh model . [ 1].,title="fig:",width=283 ]
in this paper , we have considered a self - dual system with maxwell and chern - simons terms , obtained by dimensional reduction of the carroll - field - jackiw model coupled to a higgs field .
thus , the neutral scalar field @xmath17 appears as a natural consequence of the dimensional reduction process , and the model presents self - duality equations without the necessity to introduce an additional field in the theory .
the main difference with the self - dual model obtained in ref .
@xcite is that in our model the magnetic vortex solution do not present explicit dependence on the neutral scalar field .
another interesting aspect is that the vortex solutions are identical in form to the nielsen - olesen vortices .
the only deference between the two solutions lies in the fact that our vortex solutions have electric charge .
finally we were able to resolve numerically the bogomolnyi equations of the model .
in addition we analyze the results by comparing our solutions with the solutions for the self - dual system studied in @xcite , obtained as the main result an increase in the intensity of the magnetic and electric fields .
it is worthwhile to observe that numerical analysis shows that the profiles correspondents to the higgs field and the magnetic field in our model are exactly the same as those correspondent to the maxwell - higgs model .
d. colladay , v.a .
kostelecky , phys .
d * 55 * , 6760 ( 1997 ) ; d. colladay , v.a .
kostelecky , phys .
d * 58 * , 116002 ( 1998 ) ; v.a . kostelecky , r. lehnert , phys .
d * 63 * , 065008 ( 2001 ) ; v.a .
kostelecky , phys .
d * 69 * , 105009 ( 2004 ) . | self - dual abelian higgs system , involving both the maxwell and chern - simons terms are obtained from carroll - field - jackiw theory by dimensional reduction .
bogomolnyi - type equations are studied from theoretical and numerical point of view . in particular
we show that the solutions of these equations are nielsen - olesen vortices with electric charge . |
self - assembled quantum dots are three - dimensional inclusions of one material in another one ( a matrix ) .
usually there is a lattice mismatch between materials of an inclusion and a matrix .
the lattice mismatch gives rise to a built - in inhomogeneous elastic strain which in turn produces significant changes in the electronic band structure .
@xcite therefore knowledge of the strain distribution is of crucial importance for electronic structure studies .
nearly all papers concerning electronic structure calculations of quantum dots start with evaluation of elastic strain . especially important
is the strain distribution for type - ii quantum dots where the confining potential for one type of carriers is mainly due to the strain inhomogeneity .
@xcite there are a lot of theoretical works on the strain distribution in quantum dot structures ( for a review , see refs . ) .
in addition to numerical calculations ( using finite difference , @xcite finite element , @xcite valence force field , @xcite and molecular dynamics @xcite methods ) , some analytical techniques have been proposed .
most of them are based on the usage of green s functions , either in the real space @xcite or in the reciprocal space .
@xcite some authors break the inclusion into infinitely small `` bricks '' @xcite or into infinitely thin cuboids @xcite and then apply the superposition principle . for ellipsoidal inclusions ,
eshelby s approach @xcite has proved to be effective .
also a number of results obtained in thermoelasticity theory may be applied to lattice - mismatched heterostructures , as pointed out in ref . .
different methods have their own merits and restrictions . in our opinion ,
an ideal solution of the elastic inclusion problem has to be analytical , to be expressed in terms of elementary functions and written in closed form , to be applicable to a broad range of inclusion shapes , and to take into account elastic anisotropy and atomistic corrections .
analytical closed - form solutions have been found for few cases of inclusion shapes : an ellipsoid , @xcite a cuboid , @xcite a pyramid , @xcite and a variety of quantum - wire - like structures .
@xcite nozaki and taya @xcite have presented a general solution for an arbitrary polyhedron , but it is extremely complicated .
all these solutions imply elastic isotropy and ( except the case of ellipsoidal inclusion ) equal elastic constants of the two media .
the aim of our paper is to develop a novel approach to constructing the solutions for the general case of a polyhedral inclusion , and to propose a new insight into the structure of a solution .
we stress that the solution should have a clear physical or geometrical meaning . without having a clear structure of a solution ,
it is hardly possible to develop its generalization to anisotropic media and/or to inclusions with elastic constants different from ones of the matrix .
this paper considers the following problem .
there is an infinite elastically isotropic medium ( a matrix ) with a finite polyhedron - shaped inclusion .
the crystal lattice of the inclusion matches the lattice of the matrix without any defects .
elastic moduli of the inclusion are assumed to be equal to ones of the matrix , but the matrix and the inclusion have different lattice constants .
this produces an elastic strain in both the inclusion and the matrix , and the task is to determine the strain tensor as a function of coordinates , @xmath0 .
we neglect atomistic and nonlinearity effects , assuming that the lattice mismatch is small , and lattice constants are small in comparison with the inclusion size .
it is important to note that the strain distribution produced by an inclusion in a _ semi - infinite _ matrix may easily be calculated , provided that the corresponding strain field in an _
infinite _ matrix is known .
@xcite we will discuss it in section [ sec : general ] .
the rest of the paper is organized in the following way . in section
[ sec : method ] , a new approach to evaluation the strain distribution , based on an analogy between electrostatic and elastic problems , is described .
the solution for an arbitrary polyhedron - shaped inclusion in an infinite matrix is presented and discussed in section [ sec : general ] .
then , in section [ sec : pyr_and_hut ] , this solution is applied to pyramidal , truncated pyramidal , and `` hut - cluster '' inclusions .
section [ sec : anizo ] shows the possibility of generalization of our method to anisotropic media .
section [ sec : concl ] contains the summary of the paper .
the appendix is devoted to evaluation of solid angles that is important for calculation of the strain .
the starting point of our investigation is a well - known analogy between the elastic inclusion problem and the electrostatic problem ( poisson equation ) .
@xcite namely , the _ displacement vector _
@xmath1 induced by the inclusion is proportional to the _ electric field _ @xmath2 that would appear if the inclusion were uniformly charged : @xmath3 where @xmath4 is the lattice mismatch ( @xmath5 , @xmath6 being the lattice constant )
, @xmath7 is the poisson ratio , @xmath8 denotes volume of the inclusion . for simplicity ,
in our auxiliary electrostatic problem we take the charge density and the dielectric constant equal to unity .
zero displacements correspond to positions of atoms exactly in sites of the ideal lattice of the matrix .
strain tensor is defined as @xmath9 where @xmath10 is @xmath11-th component of the position vector @xmath12 , @xmath13 is the kroneker delta , and @xmath14 is equal to 1 inside the inclusion and to 0 outside it . introducing an electrostatic potential @xmath15 we can express the field @xmath16 as @xmath17 combination of eqs .
( [ eq : u_ot_r ] ) , ( [ eq : eps_def ] ) and ( [ eq : field ] ) produces @xmath18 where @xmath19 . now our aim is to evaluate second derivatives of the potential @xmath20 . for this purpose ,
we introduce three additional functions : 1 ) @xmath21 an electrostatic potential of the uniformly charged ( with unit surface density ) @xmath22-th face of the inclusion surface ; 2 ) @xmath23 an electrostatic potential of the uniformly charged ( with unit linear density ) @xmath24-th edge of the inclusion surface ; 3 ) @xmath25 an electrostatic potential of the dipole layer uniformly spread over the @xmath22-th face with surface density of dipole moment equal to @xmath26 the outward normal to the face .
the potential @xmath27 of an uniformly charged edge is expressed as an integral @xmath28 , where @xmath29 is a linear element of the edge , and @xmath30 is a position vector of this linear element .
evaluation of this integral gives : @xmath31 where @xmath32 and @xmath33 are distances from the point @xmath12 to ends of the edge , and @xmath34 is the edge length .
the potential @xmath35 of a flat uniform dipole layer is known @xcite to be equal to the solid angle at which this layer is seen from the point @xmath12 , taken with positive sign if the positively charged side of the layer is seen from @xmath12 , and with negative sign otherwise .
thus below we will refer to the quantity @xmath25 as to a solid angle subtended by the @xmath22-th face from the point @xmath12 .
a sketch of modification of the `` charge distribution '' by taking derivatives of the potential : ( a ) transformation of a volume charge into a surface charge , ( b ) transformation of a surface charge into a linear charge and a dipole layer.,width=288 ] in order to find _ second _ derivatives @xmath36 , we note that the _ first _ derivative can be expressed as a sum over faces of the inclusion surface : @xmath37 ( see fig .
indeed , this derivative can be rewritten as @xmath38 the derivative @xmath39 plays the role of a `` charge density '' in eq .
( [ eq : dfiint ] ) .
it vanishes everywhere except the surface of the inclusion . near the @xmath22-th face of the surface
, @xmath14 is equal to @xmath40 , where @xmath41 is any point of this face , @xmath42 is the heaviside function ; consequently @xmath43 , that corresponds to the `` surface charge density '' @xmath44 at the @xmath22-th face .
therefore , a contribution of the @xmath22-th face to @xmath45 is equal to @xmath46 , according to eq .
( [ eq : dfi ] ) .
the next step is finding of derivatives of @xmath21 .
let , for simplicity , the @xmath22-th face lie in the plane @xmath47 , and its outward normal vector @xmath48 be directed along the axis @xmath49 . to find derivatives @xmath50 and @xmath51 , one can follow the same line of argumentation as at deriving eq .
( [ eq : dfi ] ) .
the result is : @xmath52 where @xmath24 runs over edges surrounding the @xmath22-th face ; @xmath53 is a unit vector which is parallel to the @xmath22-th face , directed out of this face , and perpendicular to the @xmath24-th edge ( see fig .
the last derivative , @xmath54 , transforms an uniformly charhed @xmath22-th face into a dipole layer with surface density of dipole momentum equal to @xmath44 , as shown in fig .
consequently , @xmath55 eqs .
( [ eq : dfxy ] ) and ( [ eq : dfz ] ) can be written together in a vector form , which is independent on orientation of a face with respect to co - ordinate axes : @xmath56 using eqs .
( [ eq : dfi ] ) and ( [ eq : dfvector ] ) , one can express second derivatives of the potential @xmath20 via solid angles @xmath25 and potentials of charged edges @xmath23 : @xmath57 where for each edge @xmath24 there are four unit vectors @xmath58 , @xmath59 , @xmath60 , @xmath61 , related to the two faces intersecting at this edge . with given normals @xmath58 and @xmath60 , the vectors @xmath59 and @xmath61 can be found in the following way : @xmath62 where @xmath63 is a unit vector directed along edge @xmath24 . from two possible directions of @xmath63 , one should choose the one going clockwise with respect to face @xmath64 and , correspondingly , counter - clockwise with respect to @xmath65 , when seeing from the outside of the inclusion .
in the previous section , it was shown that the strain tensor @xmath66 can be expressed ( via eq .
( [ eq : eps_ot_fi ] ) ) in terms of second derivatives of some auxiliary `` electrostatic potential '' @xmath67 . in turn , these second derivatives break down into contributions of all faces and edges of the inclusion surface ( eq . ( [ eq : d2firesult ] ) ) .
combining equations ( [ eq : eps_ot_fi ] ) and ( [ eq : d2firesult ] ) , we obtain the following expression for the strain tensor : @xmath68 where @xmath22 runs over faces of the inclusion surface , and @xmath24 runs over its edges . in eq .
( [ eq : main ] ) , @xmath69 is a solid angle subtended by the @xmath22-th face from the point @xmath12 ( positive if the outer side of the face is seen from the point @xmath12 , and negative otherwise ) ; @xmath70 is the electrostatic potential of an uniformly charged @xmath24-th edge ( with unit linear charge density ) at the point @xmath12 ; @xmath71 is equal to 1 inside the inclusion and to 0 outside it ; @xmath4 is the relative lattice mismatch between the inclusion and the matrix ; the constant @xmath72 is equal to @xmath73 , where @xmath7 is the poisson ratio ; @xmath13 is the kroneker delta ; @xmath26 is a normal unit vector to the @xmath22-th face , directed outside the inclusion ; and a constant tensor @xmath74 is equal to @xmath75 in eq .
( [ eq : gamma_def ] ) , @xmath58 and @xmath60 are normal unit vectors ( directed outside the inclusion ) to the two faces which intersect at the @xmath24-th edge ; @xmath59 is a unit vector perpendicular to the @xmath24-th edge and to @xmath58 , and directed out of the @xmath64-th face ; analogously , @xmath61 is a unit vector perpendicular to the @xmath24-th edge and to @xmath60 , directed out of the @xmath65-th face ( see eq .
( [ eq : b_def ] ) ) .
the tensor @xmath74 is symmetrical , and it can be written in an equivalent form @xmath76 where @xmath77 and @xmath42 is the internal dihedral angle between faces @xmath64 and @xmath65 . there is a simple expression ( [ eq : phi_k_def ] ) for potentials @xmath70 contributing into eq .
( [ eq : main ] ) . some closed - form expressions for solid angles
@xmath69 are presented in the appendix .
equation ( [ eq : main ] ) is the main result of the present paper .
it gives a closed - form analytical expression for strain distribution in and around a polyhedral inclusion buried into infinite isotropic elastic medium . with a known strain tensor
, one can easily obtain the stress tensor @xmath78 via hooke s law : @xcite @xmath79 where @xmath80 is the young modulus .
on the basis of eq .
( [ eq : main ] ) , the program code is written that can easily calculate the strain distribution produced by a lattice - mismatched inclusion in an infinite , isotropic matrix .
inclusion shape can be an arbitrary polyhedron .
this program , named `` easystrain '' , is freely available at http://easystrain.narod.ru . if the inclusion has the form of cuboid with faces perpendicular to the direction of the axes @xmath81 , @xmath82 and @xmath83 ( fig .
[ fig : cuboid ] ) , then separate components of eq .
( [ eq : main ] ) are simplified to @xmath84 and all the other components of strain tensor have a similar form .
so , the diagonal components @xmath85 , @xmath86 and @xmath87 depend only on solid angles @xmath88 , whereas off - diagonal components @xmath89 , @xmath90 and @xmath91 depend only on edge contributions @xmath92 .
this is in agreement with results of downes , faux and oreilly .
@xcite these authors pointed out that , in the case of cuboidal inclusion , solid angles subtended by faces contribute into the stress tensor ( and hence into the strain tensor too ) .
our paper generalizes this observation to the case of any polyhedral inclusion .
cuboidal inclusion .
numbers in circles refer to faces , italic numbers refer to edges.,width=192 ] now we consider some simple properties of the solution ( [ eq : main ] ) .
these properties can be regarded as tests of validity of the solution .
first , let us calculate the hydrostatic component of strain ( that is , the trace @xmath93 of the strain tensor ) . taking into account that @xmath94 , @xmath95 , and @xmath96 , we readily get from eq .
( [ eq : main ] ) @xmath97 the sum of solid angles , @xmath98 , vanishes for any point @xmath99 outside the inclusion .
indeed , all faces can be divided into two groups with regard to the point @xmath99 : 1 ) the ones whose outer sides are seen from the point @xmath99 , 2 ) the ones whose inner sides are seen from @xmath99 .
net solid angles subtended by the two groups are the same , but they contribute to the sum @xmath98 with opposite signs and therefore cancel each other .
if the point @xmath99 is inside the inclusion , all the faces belong to the second group and the solid angle subtended by them together are the full solid angle , @xmath100 .
so , @xmath101 . combining both cases ( @xmath99 outside and inside the inclusion ) we get @xmath102 and consequently @xmath103 we have come to the well - known result that the hydrostatic strain is zero outside the inclusion and is constant inside it .
@xcite then , it is easy to examine the behavior of the strain at the inclusion surface , starting from eq .
( [ eq : main ] ) . when the point @xmath12 , moving from outside toward inside , crosses a face of the inclusion , the strain changes stepwise .
the discontinuity of the strain , @xmath104 , can be written as @xmath105 there @xmath106 and @xmath107 denote discontinuities of @xmath88 and @xmath92 . in fact , edge contributions @xmath92 have no discontinuities at the face , and only one of @xmath88 has a discontinuity namely , the @xmath88 related to the face under consideration . for this face ,
@xmath108 , because the solid angle @xmath109 reaches @xmath110 at the face , and the value @xmath88 changes the sign from positive to negative .
so , @xmath111 with a unit vector @xmath112 perpendicular to the face .
it is convenient to consider the strain tensor with respect to the coordinate axes @xmath113 ,
@xmath114 , @xmath115 connected to the face : axes @xmath113 and @xmath114 parallel to the face , and the axis @xmath115 perpendicular to it .
therefore @xmath116 , @xmath117 , and @xmath118 takes the following form : @xmath119 off - diagonal components @xmath120 ,
@xmath121 , @xmath122 are zero . according to hooke s law ( [ eq : hookes_law ] ) ,
discontinuities of the stress tensor , @xmath123 , are @xmath124 and again off - diagonal components are zero . equations ( [ eq : delta_eps ] ) and ( [ eq : delta_sigma ] ) are in accordance to the boundary conditions at the interface : @xmath125 , @xmath126 ( continuity of displacement field @xmath127 ) , @xmath128 ( balance of elastic forces at the interface ) .
quantum wire inclusion .
point @xmath129 is the point where the strain is to be obtained .
@xmath130 and @xmath131 are the angle and distance contributing to eq .
( [ eq : main_2d]).,width=268 ] next , we consider the strain distribution in a quantum - wire - like inclusion and its surrounding ( fig .
[ fig : wire ] ) .
such an inclusion is a prism , the top and bottom of which go to infinity . for simplicity ,
let all side faces and edges be parallel to the axis @xmath83 .
so the strain is independent on @xmath83 . to obtain the strain distribution
, one may start from eq .
( [ eq : main ] ) for a prism of finite height , and then go to the limit of infinitely large vertical dimension . in this limit , contributions of base and bottom faces , as well as of edges adjoining to these faces ,
for each side face , the solid angle @xmath88 reduces to a doubled plane angle @xmath130 ( fig .
[ fig : wire ] ) subtended by the cross - section of this face by a plane parallel to axes @xmath81 ,
@xmath82 : @xmath132 .
edge contributions @xmath92 reduce to simple logarithmic expressions : @xmath133 where @xmath131 is the distance to the @xmath24-th edge ( fig .
[ fig : wire ] ) .
the constants in these expressions are infinitely large , but they cancel each other being substituted into eq .
( [ eq : main ] ) . as a result , we come to the following expression for the strain in a quantum - wire - like inclusion : @xmath134 there the indices @xmath22 and @xmath24 run over all side faces and edges , correspondingly ; @xmath130 is the plane angle subtended by the cross section of @xmath22-th face by the plane passing through the point @xmath12 parallel to the axes @xmath81 and @xmath82 ( positive if the outer side of the face is seen from the point @xmath12 , and negative otherwise ) ; @xmath131 is the distance from the point @xmath12 to @xmath24-th edge .
all the rest notations are the same as in eq .
( [ eq : main ] ) .
as the @xmath135- , @xmath136- and @xmath137-components of the tensors @xmath138 and @xmath139 are zero , the corresponding components of strain tensor are independent on @xmath130 and @xmath131 : @xmath140 equation ( [ eq : main_2d ] ) is an equivalent , but more simple and compact , form of the solution obtained by faux , downes and oreilly .
finally we discuss the strain distribution in a semi - infinite matrix .
davies @xcite proposed a method of reducing the elastic inclusion problem in a _ semi - infinite _ matrix to the corresponding problem in an _ infinite _ matrix . for convenience , we reproduce there the results of davies s work .
@xcite let an inclusion be buried in a semi - infinite matrix that fill a half - space @xmath141 , or @xmath142 , with a free surface in the plane @xmath143 .
isotropic linear elasticity is assumed , and elastic moduli of the matrix and the inclusion are the same . to calculate the strain distribution @xmath144 in this system
, one can previously find an analogous strain distribution @xmath66 in a system consisting of the same inclusion in an infinite matrix .
it can be found by eq .
( [ eq : main ] ) , for example .
then , components of @xmath144 are expressed via components of @xmath66 , @xmath145 and their derivatives @xmath146 , where the point @xmath147 is a `` mirror image '' of the point @xmath99 with respect to the surface : @xmath148 if the inclusion is a polyhedron , eq .
( [ eq : main ] ) provides an analytical expression for the strain tensor @xmath149 .
therefore its derivative @xmath150 can be evaluated analytically as a combination of derivatives of solid angles @xmath88 and values @xmath92 .
it is important to note that derivatives of @xmath88 can be expressed via derivatives of @xmath92 : @xmath151 where summation is over all the edges adjoining to the @xmath22-th face ; and @xmath152 is the one of unit vectors @xmath153 , @xmath154 , which is perpendicular to @xmath48 .
( [ eq : domega_via_dphi ] ) may be useful since analytical expressions for solid angles are rather complicated in comparison with the expression ( [ eq : phi_k_def ] ) for values @xmath92 .
among all polyhedrons , the three ones appear most often as geometrical models of quantum dots .
these are square - based pyramid ( fig . [
fig : pyr_hat]a ) , truncated square - based pyramid ( fig . [ fig : pyr_hat]b ) , and so - called `` hut - cluster '' ( fig .
[ fig : pyr_hat]c ) . in this section ,
we apply the general expression ( [ eq : main ] ) to the specific cases of pyramidal and hut - cluster inclusions .
the case of truncated pyramid does not demand a special consideration , because it is easy to obtain solution for truncated pyramid , provided that the solution for pyramid has yet been obtained ( see below ) .
inclusions of most common shapes : ( a ) a pyramid with square base ; ( b ) a truncated pyramid with square base ; ( c ) a `` hut - cluster '' . numbers in circles refer to faces ( the base has the number 0 ) , italic numbers refer to edges , and bold numbers to vertices.,width=288 ] with the numbering scheme of fig .
[ fig : pyr_hat]a , we get the following expressions for tensors @xmath138 and @xmath139 : @xmath155 where @xmath156 , @xmath157 , and @xmath158 is a dihedral angle between the pyramid base and any of its side face .
the tensor components are listed in braces in the following order : @xmath159 , @xmath160 , @xmath137 , @xmath161 , @xmath135 , @xmath136 .
it is worth to note the following property of the set of tensors @xmath139 : @xmath162 where @xmath163 is the length of the @xmath24-th edge .
this property comes from a requirement that all terms proportional to @xmath164 in eq .
( [ eq : main ] ) must cancel each other at @xmath165 .
it may serve as a useful test of correctness of the results .
these values of @xmath138 and @xmath139 , together with eq .
( [ eq : main ] ) , give the expression for the strain distribution in a pyramidal inclusion and its surrounding : @xmath166 there we use a shorthand notation : @xmath167 , and so on . note that , using eq .
( [ eq : sum_omega ] ) , one can simplify the expression for @xmath87 to the following one : @xmath168 eq .
( [ eq : solid_pyramid ] ) in the appendix provides analytical expressions for solid angles @xmath169 in the pyramid . to get the solution for the _ truncated _ pyramid , @xmath170 ,
the easiest way is to start from the solution for a pyramid , @xmath171 , and apply the superposition principle . the full pyramid , @xmath172 , consists of a truncated one , @xmath173 , and a smaller one , @xmath174 ( fig .
[ fig : pyr_hat]b ) . according to the superposition principle , @xmath175 where @xmath171 , @xmath170 and @xmath176 refer to figures @xmath172 , @xmath173 and @xmath174 ,
then , it is well known that , in the framework of the continual elasticity theory , similar inclusions produce similar strain fields . as pyramids @xmath172 and @xmath174
are similar , there is a relation between @xmath171 and @xmath176 : @xmath177 where @xmath178 is a position vector of the apex @xmath179 , @xmath180 is a truncation parameter ( a ratio of sizes of the two pyramids , see fig . [
fig : pyr_hat]b ) .
so @xmath170 can be expressed in terms of @xmath171 : @xmath181 this solution was compared numerically with the solution published in ref . .
we reproduce all strain profiles presented in that paper , except the component @xmath90 in fig .
11 of ref . , where the absolute value coincides with our results , but the sign was opposite .
we believe that the sign of @xmath90 in ref .
is erroneous , because it leads to an incorrect behavior of the strain at @xmath165 .
indeed , the multipole expansion , being applied to eq .
( [ eq : u_ot_r ] ) , gives for large @xmath182 @xmath183 where @xmath8 is a volume of the inclusion , and @xmath184 is a position vector of its center of mass . for a pyramid , @xmath185 .
this expression shows that , at fixed positive @xmath81 and @xmath82 , @xmath90 must be negative when @xmath186 and positive when @xmath187 .
this predicted behavior of @xmath90 disagrees with fig .
11 of ref . , but agrees with our calculations . finally , we consider the hut - cluster .
the hut - cluster is a figure that consists of the base ( a parallelogram ) and four side faces .
slope angles of all the side faces are the same
. therefore orientations of faces and edges of the hut - cluster are the same as of pyramid , except the top ( 9th ) edge .
so , the solution for the hut - cluster is very similar to the one for the pyramid .
the only difference is the addition of the contribution of 9th edge . of course ,
the values of solid angles @xmath169 and of edge contributions @xmath188 in the hut - cluster are not the same as in the pyramid .
analytical expressions for solid angles in the hut - cluster are given by eq .
( [ eq : solid_hut ] ) in the appendix . to get the solution for the hut - cluster from eq .
( [ eq : pyramid_solution ] ) , it is sufficient to add the term @xmath189 to @xmath86 , and to add the term @xmath190 to @xmath87 .
this demonstrates the flexibility of the general solution ( [ eq : main ] ) .
this is a property that is not inherent in previous particular solutions .
@xcite an analytical solution for a hut - cluster was first obtained by glas@xcite as a special case of a more general answer for a truncated pyramid with rectangular bottom and top faces .
our method provides a much more simple solution . if the point @xmath12 lies at the four - fold axis of symmetry of the pyramid , expressions ( [ eq : pyramid_solution ] )
are simplified greatly , because all values @xmath191 are the same , values @xmath192 are the same , and @xmath193 are also the same .
moreover , it is sufficient to evaluate only the @xmath137-component of the strain , because non - diagonal components @xmath89 , @xmath90 , @xmath91 are zero , and other diagonal components @xmath85 and @xmath86 can be expressed via @xmath87 using eq .
( [ eq : eps_alpha_alpha ] ) : @xmath194 it is convenient to use eq .
( [ eq : pyramid_epszz_simplified ] ) for evaluation the component @xmath87 .
let us put the origin of coordinate system to the center of the pyramid base .
so , coordinates of a point at the axis of symmetry are @xmath195 .
let @xmath196 be a pyramid height , @xmath6 be a length of the pyramid base , @xmath197 be a length of a side edge , and @xmath182 be a distance between the point @xmath195 and any vertex of the pyramid base : @xmath198 according to eq .
( [ eq : solid_pyramid ] ) , the solid angle @xmath199 is equal to @xmath200 then , @xmath201 substituting all these quantities into eq .
( [ eq : pyramid_epszz_simplified ] ) , we get @xmath202 here we expressed quantities @xmath203 and @xmath204 in terms of @xmath6 , @xmath196 , and @xmath205 ; @xmath19 ; @xmath206 if @xmath207 and @xmath208 otherwise ; the constant @xmath209 is the coefficient at @xmath210 in eq .
( [ eq : pyramid_epszz_simplified ] ) : @xmath211 for the truncated pyramid , eq . ( [ eq : truncated_solution ] ) together with eq .
( [ eq : profile_pyramid ] ) give @xmath212 in eq .
( [ eq : profile_truncated ] ) , @xmath206 if @xmath213 , and @xmath208 otherwise ; @xmath214 ; @xmath215 is a distance between the point @xmath216 and any vertex of the base ; @xmath197 . in a similar manner
, one can get the strain profile along the axis of symmetry of the hut - cluster . as there is only two - fold axis in the hut - cluster , the components @xmath85 and @xmath86 are no longer the same .
therefore we can not use eq .
( [ eq : epsxx_from_epszz ] ) to extract @xmath85 and @xmath86 from @xmath87 . instead
, we should find @xmath87 and @xmath85 independently , and then extract @xmath86 by means of eq .
( [ eq : eps_alpha_alpha ] ) : @xmath217 we chose the center of the cluster base as an origin of the coordinate system .
let @xmath196 be a cluster height , @xmath6 and @xmath218 be the smaller and the bigger edge lengths of the base , correspondingly .
then , values of @xmath87 and @xmath85 at the axis of symmetry are @xmath219 @xmath220r_1\!+\!(b^2\!+\!4z^2)r_5 } -\!\frac12\log\frac{2r_1\!+\!a}{2r_1\!-\!a } + \!\frac{a}{2l}\log\frac{r_5\!+\!r_1\!+\!l}{r_5\!+\!r_1\!-\!l } \right ) -\varepsilon_0\chi,\ ] ] where @xmath221 is a distance from the point @xmath195 to the first vertex of the hut - cluster ; @xmath222 is a distance from the point @xmath195 to the fifth vertex ; @xmath197 is a length of each side edge ; @xmath206 if @xmath207 and @xmath208 otherwise . the last log term in eq .
( [ eq : profile_hut_epszz ] ) is a contribution of the 9th edge , and the arctangent term in eq .
( [ eq : profile_hut_epsxx ] ) comes from the first and the third faces .
all the rest terms are similar to that of eq .
( [ eq : profile_pyramid ] ) . the strain component , @xmath87 , plotted along the @xmath83 axis for a pyramid ( eq .
( [ eq : profile_pyramid ] ) , solid line ) , a truncated pyramid ( eq . ( [ eq : profile_truncated ] ) , dash line ) , and a hut - cluster ( eq . ( [ eq : profile_hut_epszz ] ) , dash - dot line).,width=288 ] as an illustration , in fig .
[ fig : profile ] we plotted profiles of the strain component , @xmath87 , calculated by equations ( [ eq : profile_pyramid][eq : profile_hut_epszz ] ) for some particular cases of pyramidal , truncated pyramidal , and hut - cluster inclusions .
parameters of the structures are chosen to be the same as ones in ref .
: @xmath223 ; @xmath224 ; @xmath225 ; @xmath226 for the pyramid and the hut - cluster ; @xmath227 and @xmath228 for the truncated pyramid ; @xmath229 for the hut - cluster .
presented curves for the pyramid and the truncated pyramid are identical to ones of ref .
( curves d and b in fig . 5 , correspondingly ) , that confirms the correctness of our formulas .
a detailed discussion of these profiles is beyond the scope of the present paper .
we only note that @xmath87 diverges logarithmically at @xmath230 in the pyramid and the hut - cluster .
this divergence is a common feature of a strain distribution in a vicinity of a vertex or an edge of any polyhedral inclusion .
the above consideration was based on an assumption of elastically isotropic inclusion and matrix . for applications to semiconductor heterostructures , this assumption may be a source of considerable error .
for example , the young modulus of silicon in @xmath231 direction is 1.44 times greater than that in @xmath232 direction .
therefore , taking the elastic anisotropy into account is an actual problem . in this section
, we argue that our method can be expanded to anisotropic media .
we start from expression of strain tensor via green s tensor @xmath233 by faux and pearson @xcite : @xmath234 where @xmath8 is the inclusion volume , and @xmath4 is lattice mismatch .
these authors found a series expansion for green s tensor , assuming cubic anisotropy : @xmath235 where expansion coefficient @xmath236 is a measure of anisotropy ( @xmath237 for typical semiconductors ) , @xmath238 , @xmath239 and @xmath240 are elastic moduli .
each term of this expansion can be presented as a combination of partial derivatives of expressions like @xmath241 , @xmath242 , etc .
for example , isotropic term is @xmath243 @xmath159-component of first - order correction , @xmath244 , is a linear combination of the following terms : @xmath245 as a result , strain tensor @xmath0 expresses as a combination of derivatives @xmath246 where @xmath247 with a constraint @xmath248 .
each term in the expansion ( [ eq : expansion ] ) is a sum of a _ finite _ number of derivatives ( [ eq : integrals ] ) taken with proper constant coefficients . according to our method , the integrals in eq .
( [ eq : integrals ] ) can be regarded as electrostatic potentials induced by a _ non - uniformly _ charged inclusion . taking the derivatives in eq .
( [ eq : integrals ] ) , one proceeds from `` volume charge '' to `` surface dipoles '' on faces of the inclusion surface , and to `` linear charges '' and `` multipoles '' on its edges .
thus , our method allows to split the strain tensor into contributions of faces and edges of inclusion surface ( assuming that inclusion shape is a polyhedron ) : @xmath249 where indices @xmath22 and @xmath24 run over all faces and edges , correspondingly .
explicit formulas for contributions @xmath250 and @xmath251 are beyond the scope of the present paper and are the subject of a separate publication .
there we only note that each face contribution @xmath252 is proportional to a solid angle @xmath69 with a coefficient depending on the orientation of this face and on elastic constants .
edge contributions @xmath253 can be expressed in a closed form for each term of series expansion ( [ eq : expansion ] ) .
in summary , we propose a new , more simple and flexible expression for strain field in and around an inclusion buried in an infinite or semi - infinite isotropic medium .
this expression was also implemented as a computer program.@xcite we show that the strain field can be presented as a sum of contributions of the faces and edges .
this is the main point of our method ; it gives a possibility to construct expressions for strain distribution in inclusions of complicated shapes .
the general solution is applied to important particular cases of pyramidal and hut - cluster inclusions .
our solution for the pyramid reproduces previous solutions , but in a simpler and intuitively understandable form .
we believe that it paves the way for further simplifications and generalizations of the solution , for example , to the case of anisotropic elasticity .
this work was supported by rfbr ( grant 06 - 02 - 16988 ) , the dynasty foundation , and the president s program for young scientists ( grant mk-4655.2006.2 ) .
here we present some explicit formulas expressing solid angles as functions of coordinates .
a solid angle @xmath35 , that a surface @xmath254 subtends at a point @xmath12 , may be defined as an integral over the surface : @xmath255 where @xmath256 is a surface element , @xmath112 is a unit vector directed normally to this surface element , and @xmath30 is a position vector of the surface element .
this integral is easily evaluated if the surface @xmath254 is a rectangle . for simplicity , let this rectangle lie in the plane @xmath257 , and its edges be oriented along the axes @xmath81 and @xmath82 ( fig .
[ fig : solid]a ) .
let @xmath258 and @xmath259 be @xmath81-coordinates of edges directed along the axis @xmath82 @xmath260 ; @xmath261 and @xmath262 be @xmath82-coordinates of the rest two edges of the rectangle @xmath263 .
then the integral ( [ eq : solid_def ] ) is expressed as follows : @xmath264 or @xmath265 here @xmath266 are distances from the point @xmath12 to the corners of the rectangle : @xmath267 it is assumed that values of arctangents fall into the range @xmath268 . to find a solid angle subtended by a triangle ( fig .
[ fig : solid]b ) , one can use the relation by oosterom and strackee : @xcite @xmath269\,\mathbf{c } } { abc\!+\!(\mathbf{a}\mathbf{b})c\!+\!(\mathbf{a}\mathbf{c})b\!+\!(\mathbf{b}\mathbf{c})a}.\ ] ] vectors @xmath270 , @xmath271 , @xmath272 join the point @xmath179 , at which the solid angle is subtended , to the vertices @xmath129 , @xmath273 , @xmath274 of the triangle . to satisfy the condition that the solid angle @xmath275 is positive
if it is looked at from outside , and is negative otherwise , one should choose the proper order of following of the vertices @xmath129 , @xmath273 , @xmath274 .
namely , the closed contour @xmath276 should follow in a clockwise direction , seeing from outside the inclusion ( as shown by dashed arrows in fig .
[ fig : solid]b ) .
so the triple scalar product @xmath277\,\mathbf{c}$ ] is positive ( negative ) if the outer ( inner ) side of the triangular face is seen from the point @xmath179 .
care must be taken while resolving eq .
( [ eq : solid_oosterom ] ) with respect to @xmath275 . for simplicity
, we will refer to the right side of eq .
( [ eq : solid_oosterom ] ) as to @xmath180 .
the sign of @xmath275 is the same as the sign of the product @xmath277\,\mathbf{c}$ ] , but may differ from the sign of @xmath180 .
so we can not `` naively '' resolve eq .
( [ eq : solid_oosterom ] ) as @xmath278 . instead , we should write down @xmath279 if @xmath277\,\mathbf{c}>0 $ ] , and @xmath280 otherwise . joining together both cases , we obtain @xmath281\,\mathbf{c } } { abc\!+\!(\mathbf{a}\mathbf{b}\!)c\!+\!(\mathbf{a}\mathbf{c}\!)b\!+\!(\mathbf{b}\mathbf{c}\!)a } \,\mathrm{mod}\,\pi\right)\\ & -&2\pi\left(1-\theta([\mathbf{a}\times\mathbf{b}]\,\mathbf{c})\right),\end{aligned}\ ] ] where @xmath282 is the heaviside function ( 1 for positive @xmath81 , 0 for negative @xmath81 ) .
note that the easiest way to implement eq .
( [ eq : solid_triangle ] ) in a computer program is to use c math library function atan2 : @xmath283 , where @xmath284 and @xmath285 are numerator and denominator of the right part of eq .
( [ eq : solid_oosterom ] ) . now let us apply the expressions ( [ eq : solid_rect ] ) and ( [ eq : solid_triangle ] ) to faces of the pyramid and the hut - cluster considered in section [ sec : pyr_and_hut ] .
first , we choose a reference frame with the origin at the center of the pyramid base , the axes @xmath81 and @xmath82 along edges of the base , and the axis @xmath83 directed toward the apex of the pyramid .
so , position vectors of the vertices ( see fig . [
fig : pyr_hat]a ) are @xmath286 where @xmath6 is a base edge length , @xmath196 is a height of the pyramid .
note that a dihedral angle @xmath287 .
solid angles contributing into eq .
( [ eq : pyramid_solution ] ) are @xmath288 the minus sign at @xmath289 in eq .
( [ eq : solid_pyramid ] ) reflects the fact that the pyramid base is directed downwards . for the hut - cluster ,
position vectors of vertices are @xmath290 here @xmath6 and @xmath218 are the smaller and the bigger edge lengths of the base .
( again , the dihedral angle is @xmath287 . )
we introduce also two points @xmath291 and @xmath292 where side edges cross ( see fig . [
fig : pyr_hat]b ) : @xmath293 the solid angle @xmath294 of the trapezoidal 2nd face is the difference of two solid angles subtended by triangles @xmath295 and @xmath296 .
the angle @xmath297 is evaluated in the same way . as a result ,
solid angles subtended by faces of the hut - cluster are @xmath298 these expressions can be substituted into a modified version of eq .
( [ eq : pyramid_solution ] ) , as described in section [ sec : pyr_and_hut ] . | an analytical expression of the strain distribution due to lattice mismatch is obtained in an infinite isotropic elastic medium ( a matrix ) with a three - dimensional polyhedron - shaped inclusion ( a quantum dot ) .
the expression was obtained utilizing the analogy between electrostatic and elastic theory problems .
the main idea lies in similarity of behavior of point charge electric field and the strain field induced by point inclusion in the matrix .
this opens a way to simplify the structure of the expression for the strain tensor . in the solution ,
the strain distribution consists of contributions related to faces and edges of the inclusion .
a contribution of each face is proportional to the solid angle at which the face is seen from the point where the strain is calculated .
a contribution of an edge is proportional to the electrostatic potential which would be induced by this edge if it is charged with a constant linear charge density .
the solution is valid for the case of inclusion having the same elastic constants as the matrix .
our method can be applied also to the case of semi - infinite matrix with a free surface .
three particular cases of the general solution are considered for inclusions of pyramidal , truncated pyramidal , and `` hut - cluster '' shape . in these cases
considerable simplification was achieved in comparison with previously published solutions .
a generalization of the obtained solution to the case of anisotropic media is discussed . |
during the last decade , magnetic multilayers have received a lot of attention due to their interesting properties like oscillating interlayer exchange coupling ( iec ) between antiferromagnetic ( af ) and ferromagnetic ( fm ) ordered layers@xcite and giant magnetoresistance ( gmr ) apart from their industrial applications .
such oscillations in interlayer exchange coupling and the saturation magnetoresistance were reported by parkin @xmath2 @xmath3 @xcite with a period 15 - 20 in fe / cr , co / cr and co / ru multilayers .
@xmath3 @xcite showed that the coupling oscillates with a smaller period of 3.0 - 4.0 in fe / cr / fe sandwich structure because of the high degree of perfection of fe whisker substrate and the sharpness of fe / cr interfaces .
the first theoretical explanation was provided by bruno @xmath2 @xmath3 @xcite , who explained the oscillations as a function of spacer layer thickness due to the ruderman - kittel - kasuya - yosida ( rkky ) interaction between the magnetic layers .
they expressed the period of oscillations in terms of nesting vectors of the spacer s fermi surface .
however , the coupling strength j was described by an adjustable parameter in their work .
there are several other explanations of this phenonmenon based on electronic fabry - perot - like interference effects @xcite or quantum - well ( qw ) theories@xcite . in order to understand this phenonmenon ,
first - principles techniques have also been employed which are more transparent and parameter free but require considerable computational effort .
the first - principles calculations of the exchange coupling j have been reported for various systems such as fe / cr , fe / cu , fe / mo , co / cu @xcite . recently , the combined theoretical - experimental work on fe / au supercells , with only 2 monolayers of fe @xcite , showed that the exchange coupling j exhibits oscillatory behaviour . however ,
this was found to be ferromagnetic for all layers .
some results have also been reported for the oscillatory behaviour of magnetic moments as a function of spacer layer thickness @xcite .
fe / nb is an interesting system as it provides a way of exploring the coexistence of ferromagnetism and superconductivity .
this is one of the main interests of experimental studies of fe / nb / fe multilayers@xcite .
m@xmath4hge @xmath5 @xcite studied the magnetism and superconducting properties of fe / nb trilayer system .
they have found a strong dependence of the superconducting transition temperature @xmath6 on fe layer thickness .
they also studied the magnetic properties using the magneto - optical kerr effect , ferromagnetic resonance and magnetization measurements by squid magnetometer .
although a large number of studies exist for fe / cr , fe / au , co / cr @xmath7 , there are only a few theoretical studies for fe / nb system .
the difficulty one faces with fe / nb is that the lattice parameter of fe and nb differ by @xmath8 and thus it presents an additional complication for theoretical as well as experimental studies .
the fe / nb multilayers show strong interlayer exchange coupling @xcite which changes in continuous and reversible way by introducing hydrogen into the sample @xcite .
this system has not been explored much theoretically and therefore needs a detailed study .
the experimental results for fe / nb multilayers @xcite show oscillatory interlayer exchange coupling as a function of nb spacer layer thickness within a period of 9.0 to 12.8 .
@xmath3 @xcite showed the oscillating exchange coupling in fe / nb multilayers as a function of nb spacer layer thickness using the augmented spherical wave ( asw ) method @xcite . to study the exchange coupling in fe / nb multilayers , we have used a self - consistent full - potential linearized augmented plane - wave ( flapw ) method @xcite that treats crystal potential accurately and makes no shape approximation for the crystal potential . in this paper
we report the energy functional dependence of exchange coupling and magnetic properties of fe / nb multilayers .
some preliminary results of our work have been reported earlier @xcite .
our results show that the interlayer exchange coupling for both the local spin density approximation ( lsda ) and the generalized gradient approximation ( gga ) oscillates with a period of 6.0 which is close to the asw result of 4.5 @xcite .
our calculated period is close to the experimental result in the preasymptotic region .
we find that the gga favors fm coupling for all nb monolayers while the lsda gives fm as well as af coupling .
we have also studied the change in magnetic moment of fe and nb as a function of nb spacer layer thickness within the lsda and the gga .
we find that nb develops induced magnetic moments which oscillate as a function of nb spacer layer .
the magnetic moment on fe gets reduced appreciably from its bulk value and depends on nb spacer layer thickness .
the fe magnetic moment shows strong oscillations as a function of nb thickness in agreement with the experimental results @xcite .
the organization of the paper is as follows . in sec .
ii we provide the details of our calculation . in sec .
iii we present our results and discuss them .
finally , we give our conclusions in sec .
all calculations reported in this paper have been carried out using the self - consistent full - potential linearized augmented plane - wave ( flapw ) method @xcite in a scalar relativistic version without spin - orbit coupling .
we have used the lsda as well as gga for our calculations .
the muffin - tin radii are 1.33 for fe and nb .
the maximum value in the radial sphere expansion is @xmath9 , and the largest @xmath10 value for the non - spherical part of the hamiltonian matrix is @xmath11 .
the cutoff parameters are @xmath12 for the plane wave and @xmath13 for the charge density .
the number of plane waves ranges from 1400 to 5475 , depending on the number of nb layers .
an improved tetrahedron method has been used for the brillouin - zone integration @xcite . to perform the calculations we have constructed tetragonal supercells consisting of bcc fe and nb monolayers and in each supercell the number of nb monolayers ( ml ) range from 1ml to 7ml .
the schematic geometry of @xmath14 multilayers is shown in fig .
1 . the lattice constants of bulk fe and nb are very different with @xmath8 lattice mismatch . to reduce the lattice mismatch we started with a multilayer whose lattice constant is average of bulk fe and nb and then relaxed the lattice parameter .
the relaxed parameter was found to be 3.067 .
all calculations are performed with this relaxed lattice constant .
the lattice parameter along ( 001 ) direction changes from 3.067 to 12.269 in supercells .
the atomic positions have not been relaxed .
the total energy is calculated for ferro and antiferromagnetic ordering of fe atoms for each nb spacer layer thickness . to obtain reliable energy differences between ferromagnetic and antiferromagetic ordering , the unit cell of af structure is also used for fm structure .
the exchange coupling , j , is calculated by taking the energy difference @xmath16 where @xmath17 is the thickness of the spacer layer and @xmath18 and @xmath19 are the total energies of the system in antiferromagnetic and ferromagnetic arrangements . for the calculation of the iec ,
a sufficiently large number of @xmath20 points are needed .
the convergence of the iec as a function of @xmath20 points in the irreducible brillouin zone ( ibz ) is shown in fig . 2 for the lsda and gga potentials .
we find that the convergence is reached if we use about 200 @xmath20 points in the ibz and the iec converges faster than the total energy as a function of @xmath20 points in the ibz .
the total energies of the fe / nb multilayers have been calculated for para , ferro and antiferromagnetic ordering of fe atoms for each nb spacer layer thickness within the lsda and the gga exchange correlation potentials .
it is found that the total energy for paramagnetic multilayers is always higher than the total energy of the other two magnetic configurations in both the lsda and gga potentials . the interlayer exchange coupling of fe / nb multilayer is shown in fig . 3 for the lsda and gga potentials , as a function of nb spacer layer thickness .
the iec flips from ferromagnetic to antiferromagnetic at around 2.5 and 8.0 and then to ferromagnetic coupling at around 4.0 and 10.5 for the lsda .
the period of oscillation is found to be 6.0 which is in good agreement with the experimental value in the preasymptotic region . in the experiment ,
the preasymptotic value is about 6.5 while the asymptotic period is found to be 9.0 @xcite .
the iec has similar trends as in the asw work @xcite except the period of oscillation is slightly larger than the asw work . we have taken an optimized lattice constant which is smaller than the asw work as well as we have considered only 1ml of fe in our calculation while sticht @xmath2 @xmath3 @xcite have taken 2ml of fe .
this may be a possible reason for getting a different period from the asw work @xcite .
the iec also shows the oscillations for the gga potential but it does not change to antiferromagnetic coupling and the period of oscillation is found to be the same i.e. 6.0 .
this indicates that the gga favors ferromagnetic ordering in this system and it is in good agreement with the experimental results @xcite where the iec is found to be ferromagnetic for nb thickness less than 14.0 .
our calculations are restricted upto 11.0 of nb thickness .
similar behaviour has been seen in fe / au system where it was found that 2ml of fe showed fm coupling only @xcite while for higher number of fe ml , fm as well as af coupling @xcite was observed .
this is consistent with the fact that in pure fe , gga gives the ferromagnetic ground state while the lsda gives the paramagnetic ground state .
the magnitude of iec is 4.45 mj / m@xmath21 for the lsda and 22.64 mj / m@xmath21 for the gga potential , which is very large compared to the experimental value of 0.034 mj / m@xmath21 .
similar trend has been reported earlier in @xmath0 @xmath1 works @xcite on fe / cr , fe / au , co / cu multilayer systems where the theoretical value of the iec has been found to be 2 to 3 orders of magnitude higher than the experimental value .
the gga gives higher value of the iec than the lsda which is related to the fact that the gga favors ferromagnetic ordering in this system .
this is further supported by our calculation of energy difference between ferromagnetic and antiferromagnetic states in bcc fe using the gga and lsda . using the gga we get the energy difference of 0.49 ev while in the lsda
the energy difference is 0.39 ev .
the higher magnitude of the iec as compared to the experimental results can be attributed to the following reasons : ( 1 ) the roughness at interface has been neglected in our calculations .
@xmath3 @xcite studied the effect of interfacial dislocations , interfacial roughness and lattice strain on the period and the amplitude of the iec within the rkky approach .
they found that the interfacial dislocations reduce amplitude of the iec .
the reduction was more in ( 111 ) direction than ( 001 ) direction in cu / co and au / co cases .
the interface roughness causes fluctuations in the spacer layer thickness and also breaks in - plane translational symmetry . as a result
the iec is strongly reduced and the period is increased . in fe
/ nb multilayers , which have about @xmath8 lattice mismatch at the interfaces , this effect is probably larger compared to the multilayers , which have matched interfaces such as fe / cr , co / cu , fe / au etc .
the interdiffusion and intermixing at the interface causes formation of fenb alloy at the interface @xcite . this will further contribute to roughness and suppress the iec .
( 2 ) the magnitude of the iec in experimental study @xcite has been reported for higher thickness of nb layer but our calculations are limited to 11.0 nb layer thickness .
we find that the fe layers induce the magnetic moments on nb atoms .
the induced moment on nb atoms at interface layer is opposite to that of fe atoms . in fig .
4 , we show the induced magnetic moments per atom on nb spacer layers using the gga when fe atoms are ferromagnetically ordered in @xmath22 multilayers system .
the induced magnetic moments show oscillations although the amplitude of the oscillations away from the interface is quite small .
the induced magnetic moments at the interface is 0.17 @xmath23 while at the middle layer it is 0.01 @xmath24 in @xmath22 multilayer .
we notice that the amplitude of the oscillations increases with the increase in number of fe layers .
this is similar to the nature of induced magnetic moment in the spacer layer seen in other systems @xcite .
it is also seen that the nb atoms have higher magnetic moments on the interfacial nb spacer layer .
these induced magnetic moments are reduced in the lsda calculations .
the magnetic moment at the interface changes to 0.09 @xmath23 and zero magnetic moment at the middle layer of nb in the lsda .
the explanation is given in next section for the difference of magnetic moments of nb for both the lsda and gga functionals .
in fig . 5 , we plot the change in the magnetic moment per atom in fe layers as a function of nb spacer layer and find that it also shows oscillations in the lsda as well as gga . also shown in the inset
are the experimental results of fe / nb multilayers @xcite .
the magnetic moment of fe is reduced at the interface as compared to the bulk value , in agreement with the experimental results . in case of the gga potential ,
the magnetic moment of fe varies from 1.32 @xmath23 to 0.90 @xmath23 while in lsda , it varies from 1.12 @xmath23 to 0.46 @xmath23 depending on the nb layers in the system . in both cases
the magnetic moment of fe is smaller than the magnetic moment in bulk fe ( 2.24 @xmath23 ) .
our gga calculations show that the magnetic moment of fe at the interface decreases rapidly with nb thickness till 5.0 and is almost constant between 5.0 to 10.0 and then shows a sharp increase around 10.0 .
we note that the gga results show much better agreement with the experiment . in particular , the sharp rise of the magnetic moment around 10.0 is well reproduced by the gga results .
we also find the magnetic moment of fe , in case of @xmath25 , varies from 2.16 @xmath23 to 2.00 @xmath23 and the induced magnetic moment on nb layer also changes from 0.54 @xmath23 to 0.22 @xmath23 .
thus the reduction in magnetic moment of fe atoms at interface depends on nb spacer thickness and magnetic moments of nb spacer layer also depends on fe layer thickness . to understand the behaviour of the magnetic moments of fe and nb , we have examined the d - band density of states ( d - dos ) of fe and nb in @xmath26 multilayer system .
the ferromagnetic d - dos for fe in bulk and @xmath27 multilayers are shown in fig .
this figure shows that the difference between up and down electrons in bulk fe is more compared to the corresponding difference in @xmath27 multilayers .
thus the magnetic moment of fe in @xmath27 multilayers is smaller compared to the bulk value .
also we note from the figure that the down spin d - dos in @xmath27 is shifted to lower energy compared to the corresponding d - dos in bulk fe .
this shows that the exchange splitting in @xmath27 is smaller than that in bulk fe .
the figure also shows a comparison between the lsda and gga d - dos .
we see from the figure that exchange splitting in the gga is larger compared to the lsda . in fig . 7 , we have shown the d - dos of nb layers at the interface and at the middle layer for ferromagnetic @xmath27 multilayers using the gga .
the bulk d - dos of nb is also shown .
the layer at interface is denoted by nb1 and the middle layer by nb2 .
the interface and middle layers have different d - dos compared to the bulk d - dos of nb .
the d - dos at the interface layer has peaks between -3.0 ev to 2.0 ev while the middle layer has much smaller d - dos in this energy range .
this shows some hybridization of d - states of nb and fe at the interface layer while this hybridization is almost absent in the middle layer .
this is consistent with the result in fig .
4 which shows that the nb interface layer near the fe layer has higher magnetic moment as compared to the layers far from fe layers .
thus the induced magnetic moment on nb layer is a consequence of the hybridization between fe and nb d - states .
nb magnetic moment shows oscillations as shown in fig .
4 because fe layer induces a spin polarization in nb layers which oscillates due to friedel oscillations @xcite .
this is consistent with the earlier result for fe / mo multilayers where magnetic moment on mo also shows friedel oscillations @xcite .
@xmath3 @xcite have also found the reduction in the fe magnetic moments which is same in our case .
the mechanism of the oscillation of fe magnetic moments is not clear and open for further work .
several models @xcite have been presented to explain the period of oscillations of the iec .
the rkky and qw models express the period in terms of the nesting vectors at the fermi surface .
the rkky and the qw models give the same period because of the dependence on the dimensions of the fermi surface of the space material .
we shall now discuss the results of these models and compare them with our results .
in the rkky theory , one magnetic layer polarizes the spacer layer and the other magnetic layer interacts with this . since the spin polarization is oscillatory as a function of spacer layer thickness , the iec also oscillates .
the period of oscillation is given by @xmath28 , where * k * is the spanning vector @xcite across the fermi surface of the spacer layer . therefore , to calculate the period of oscillation of the iec we have calculated the fermi surface of nb which has many critical spanning vectors corresponding to many oscillation periods .
the calculated values of spanning vectors in [ 100 ] plane are 0.79 , 0.63 , 0.53 , 0.46 , 0.35 , 0.20 and 0.15 in unit of @xmath29 , where d is the distance between the lattice planes along ( 001 ) direction .
the corresponding oscillation periods for these spanning vectors are 4.20 , 5.43 , 6.20 , 7.24 , 9.31 , 16.29 and 21.73 which are in good agreement with earlier calculations @xcite .
the oscillation period of one of these spanning vector is 6.20 , which is in good agreement with our flapw calculation .
because our flapw calculations are restricted up to 7ml of nb spacer layer thickness , we could find only one period of oscillation .
we have also calculated the period of oscillation of the iec using the quantum well ( qw ) model @xcite even though within first order perturbation theory the rkky and qw models are equivalent . in this model
the fermi energy oscillates as a function of the well thickness .
the period of the oscillation is given by the wave vector at the fermi energy in the well .
we consider the free electron model where the standing wave form in qw if @xmath30 , where @xmath31 is the wavelength , @xmath32 is the thickness of the well and @xmath33 specifies the energy level .
@xmath32 is defined as @xmath34 , where @xmath35 is the number of spacer ml and @xmath17 is the interlayer spacing . thus using the above relations and @xmath36
, we get the condition @xmath37 . for a free electron in the well
, the energy can be written as @xmath38 as shown in ref .
@xcite , energy @xmath39 shows oscillation as a function of @xmath35 .
since the density of states ( dos ) for this case varies as @xmath40 , it will also show oscillation as a function of n at fermi level .
the period of oscillations can be predicted by analyzing the peak positions of qw states in the electronic density of states @xcite . in fig .
8 , we have plotted the variation in s and p dos in the nb spacer layer at fermi energy as a function of nb spacer layer thickness , which shows a peak at 3.0 and at 9.0 .
this gives period of oscillation to be 6.0 which is in good agreement with the result obtained from our flapw calculation and the rkky model .
since both of these models gives the same results for the period of the iec , the comparison of our results with the model results gives no clue regarding the mechanism of iec .
this question has been addressed by schilfgaarde and harrison @xcite and mirbt @xmath2 @xmath3 @xcite in connection with fe / cr / fe and fe / mo / fe layers . in the rkky theory , the amplitude of the coupling is bilinear in the magnetic moments on two sides while in qw model the amplitude does not depend on the magnetic moments . to analyze this , the calculated iec was fitted to the asymptotic limit of the rkky expression and the amplitudes were calculated . for the fit ,
results for spacer layer thickness less than 5ml were ignored as the expression used for the fit corresponded to the asymptotic limit . since the electronic structures of nb and mo are similar @xcite , fe / nb system may show a behaviour similar to fe / mo system . for the fe / mo system , mirbt @xmath2 @xmath3 @xcite fitted their results up to 20ml and found that the amplitude is proportional to the magnetic moments squared of fe for 3ml , 4ml and 10 ml caliper while the iec has no dependence on magnetic moments for 2ml caliper .
thus their analysis does not favor either the rkky or the qw model . in our case
the maximum thickness is 7ml and therefore this kind of fit to the asymptotic expression is not possible at present and will be taken up in future work .
we have calculated the interlayer exchange coupling between fe layers when separated by nb spacer layers using the self - consistent flapw method with the lsda and gga exchange correlation potentials .
we observe an oscillating exchange coupling as a function of nb spacer layer thickness with a period of 6.0 for the lsda as well as gga in good agreement with the preasymptotic value observed in the experiment .
it is close to earlier calculated period 4.5 by the asw method .
our calculated period is also in good agreement with the period calculated using the rkky and qw models .
the lsda calculations show an af coupling at 2.5 and 8.0 thickness of nb space layer but the gga calculations favor ferromagnetic coupling for all nb spacer layers .
we find that the induced magnetic moments on nb atoms show oscillatory behaviour .
the fe magnetic moment also shows strong oscillations as a function of nb spacer layer thickness .
the gga gives higher magnetic moments for fe and nb compared to the lsda .
the gga results show better agreement with the experiment than the lsda results .
the reduction in magnetic moments of fe and the induced magnetic moment on nb atoms is found to be a consequence of the hybridization between fe and nb d - bands .
we thank profs .
m. k. harbola , s. auluck , r. c. budhani and a. k. majumdar for helpful discussions .
this work was supported by the department of science and technology , new delhi , via project no .
sp / s2/m-51/96 .
99 p. gr@xmath4nberg , r. schreiber , y. pang , m. b. brodsky , and h. sowers , phys .
lett . * 57 * , 2442 ( 1986 ) .
parkin , n. more , and k.p .
roche , phys .
64 * , 2304 ( 1990 ) .
j. unguris , r.j .
celotta , and d.t .
pierce , phys .
lett . * 67 * , 140 ( 1991 ) .
purcell , w. folkerts , m.t .
johnson , n.m.e .
mcgee , k. jager , j. aan de stegge , w.b .
zeper , w. hoving , and p. gr@xmath4nberg , phys .
lett . * 67 * , 903 ( 1991 ) .
p. bruno and c. chappert , phys .
. lett . * 67 * , 1602 ( 1991 ) ; p. bruno and c. chappert , phys . rev . *
b 46 * , 261 ( 1992 ) .
p. bruno , phys . rev .
* b 52 * , 411 ( 1995 ) .
ortega , f.j .
himpsel , g.j .
mankey and r.f .
willis , phys . rev . *
b 47 * , 1540 ( 1993 ) .
p. lang , l. nordstr@xmath41 m , r. zeller , and p.h .
dederichs , phys .
lett . * 71 * , 1927 ( 1993 ) .
s. mirbt , o. eriksson , b. johansson , and h.l .
skriver , phys . rev . * b 52 * , 15070 ( 1995 ) .
s. mirbt , a.m.n .
niklasson , b. johansson , and h.l .
skriver , phys . rev .
* b 54 * , 6382 ( 1996 ) .
niklasson , s. mirbt , h.l .
skriver , and b. johansson , phys . rev . *
b 53 * , 8509 ( 1996 ) .
a. yoshihara , j.t .
wang , k. takanashi , k. himi , y. kawazoe , h. fujimori , and p. gr@xmath4nberg , phys . rev . *
b 63 * , 100405(r ) ( 2001 ) .
mattson , c.h .
sowers , a. berger , and s.d .
bader , phys .
* 68 * , 3252 ( 1992 ) . f. klose , ch .
rehm , d. nagengast , h. maletta and a. weidinger , phys .
lett . * 78 * , 1150 ( 1997 ) . th .
m@xmath4hge , n.n .
garifyanov , yu . v. goryunov , g.g .
khaliullin , l.r .
tagirov , k. westerholt , i.a .
garifullin and h. zabel , phys .
lett . * 77 * , 1857 ( 1996 ) .
rehm , f. klose , d. nagengast , h. maletta and a. weidinger , physica * b * , 234 - 236 , 480 - 482 ( 1997 ) . j. sticht , f. herman , and j. k@xmath4bler , in @xmath42 @xmath43 @xmath44 @xmath45 , edited by p.m. oppener , j. k@xmath4bler ( world scientific , singapore 1993 ) , vol-1 , p456 - 9 .
williams , j. k@xmath4bler , c.d .
gelatt jr . , phys . rev . *
b 19 * , 6094 ( 1979 ) .
p. blaha , k. schwarz , and j. luitz , * wien97 * , a full potential linearized augmented plane wave package for calculating crystal properties ( kalheinz schwarz , techn .
universitat wien , austria ) , 199.isbn 3 - 9501031 - 0 - 4 .
nitya nath shukla and r. prasad , bull .
* 26 * , 143 ( 2003 ) .
bl@xmath41chl , o. jepsen , and o.k .
anderson , phys . rev . *
b 49 * , 16223 ( 1994 ) . j. unguris , r.j .
celotta , and d.t .
pierce , phys .
* 79 * , 2734 ( 1997 ) d. stoeffler , k. ounadjela , j. sticht , and f. gautier , phys .
lett . * 49 * , 299 ( 1994 ) .
j. friedel philos mag .
* 43 * , 153 ( 1952 ) .
himpsel , phys . rev . *
b 44 * , 5966 ( 1991 ) . mark van schilfgaarde and walter a. harrison phys . rev .
lett . * 71 * , 3870 ( 1993 ) .
stiles , phys .
b 48 * , 7238 ( 1993 ) .
moruzzi , j.f .
janak , and a.r .
williams , @xmath46 @xmath47 @xmath48 @xmath43 @xmath45 ( pergamon , new york , 1978 ) .
jani , n.e .
brener , and j. callaway , phys . rev . *
b 38 * , 9425 ( 1988 ) .
2 . interlayer exchange coupling , j(mev ) for @xmath51 multilayers as a function of @xmath20-points in the ibz for different number of nb layers for ( a ) lsda and ( b ) gga .
the calculated results are shown by filled squares , circles and stars .
5 . magnetic moments on fe atoms in @xmath51 multilayers as a function of nb spacer layers .
the dots correspond to the gga and filled squares to the lsda results .
the experimental results of mattson @xmath2 @xmath3 @xcite are shown in the inset .
the top panel shows the partial density of states of spin - up and spin - down d - bands for ferromagnetic bulk fe using the gga .
the bottom panel shows the partial density of states of spin - up and spin - down d - bands for fe in ferromagnetic @xmath27 system using the lsda and gga .
the top panel shows the partial density of states of spin - up and spin - down d - bands for nb in ferromagnetic @xmath27 system .
nb1 denotes the d - dos at the interface while nb2 denotes the d - dos at the middle layer .
the bottom panel shows the d - dos for bulk nb using the gga . | we present an @xmath0 @xmath1 calculation of the exchange coupling for fe / nb multilayers using the self - consistent full - potential linearized augmented - plane wave ( flapw ) method .
the exchange correlation potential has been treated in the local spin density approximation ( lsda ) as well as generalized gradient approximation ( gga ) .
we find that for the lsda as well as the gga the exchange coupling oscillates with a period of 6.0 of nb spacer thickness which is close to the experimental value in the preasymptotic region .
this is also close to the earlier calculated period ( i.e. 4.5 ) by augmented spherical wave ( asw ) method .
the lsda shows antiferromagnetic coupling for 2 and 5 nb monolayers ( ml ) but the gga shows the ferromagnetic coupling for all nb spacer layers .
the period of oscillation is found to be in good agreement with the period calculated using the ruderman - kittel - kasuya - yosida ( rkky ) and quantum well ( qw ) models .
the magnetic moment of fe is found to be higher in the gga than the lsda .
fe magnetic moment also shows strong oscillations as a function of the spacer layer thickness , in agreement with the experimental results .
we find that the gga results show better agreement with the experiment than the lsda results . |
the environment of a galaxy has long been known to influence its properties , such as morphology , color , and star formation .
numerous investigations showed that galaxy populations in rich cluster environments were dominated by red elliptical and s0 galaxies @xcite .
@xcite , in particular , postulated that more evolved clusters hosted a larger proportion of early - type galaxies , suggesting a link between galaxy properties and the buildup of clusters .
@xcite subsequently studied a diverse set of clusters and made the key discovery that at a fixed local density , relaxed and un - relaxed clusters host a similar fraction of early - type galaxies and that this fraction increased at higher local densities .
dressler s morphology - density relation ( mdr ) challenged the long - standing role of ram - pressure stripping @xcite in transforming field spirals into cluster early - types galaxies , and instead pointed to the importance of local processes . in recent years
, studies have probed into the lower density environments that are characteristic of groups @xcite and the field galaxy population @xcite , and found the local environment to be an important factor there as well . in general , these studies found the _
lowest _ density environments to be comprised of lower mass , blue , late - type galaxies , a stark contrast to the types of galaxies found at higher densities .
studies of galaxy properties and their dependencies on local environment are key to understanding the mechanisms that are responsible for galaxy evolution .
for example , in the high density cluster environment processes like ram pressure stripping @xcite and harassment @xcite may suppress star formation and consequently evolve galaxies towards the red - sequence . at intermediate densities that are characteristic of the group environment ,
galaxy - galaxy interactions and mergers are more probable , given the favorable relative velocities of galaxies @xcite .
early hst observations of distant clusters revealed tidal interactions of blue disk galaxies , implying a strong role for galaxy - galaxy interactions in forming present day cluster ellipticals @xcite .
an analysis of galaxy properties in different environments can therefore shed light on the processes responsible for the observed buildup of the red - sequence between @xmath13 and today @xcite .
for example , given that cluster galaxies comprise a small fraction of their @xmath13 sample , @xcite argue against processes such as ram pressure stripping and harassment as the sole origin of the observed color - density trend found at low redshift . instead , they suggest that processes that form the high early - type fraction in @xmath14 groups of galaxies dominate red - sequence formation at higher redshifts .
the outskirts of _ massive _ clusters present an ideal location to study the impact of environment on galaxy properties over a broad range of local densities : from the high density cluster core to the intermediate density groups in the outskirts and the low density surrounding field .
furthermore , the findings of @xcite and @xcite suggest that the morphological makeup of galaxies changes dramatically at the intermediate densities found in the outskirts of clusters . these mass - limited studies
found that while in the cores of clusters the fraction of early - type galaxies is high and does not evolve significantly between @xmath13 and @xmath15 , the early - type fraction ( etf ) for galaxies in the low density field is lower and also does not evolve .
because clusters grow by accreting galaxies from the field , and massive clusters are expected to grow by a factor of @xmath16 in mass between @xmath13 and @xmath15 @xcite , the large reservoir of infalling galaxies from the field must transform in the outskirts of clusters in order to preserve the high etf found in the high density core .
thus , the outskirts of massive clusters provide an opportunity to assemble a large sample of galaxies in environments where we expect transformations to be occurring .
studying the properties of galaxies in the environments of clusters at intermediate redshifts has several advantages . at such redshifts , around @xmath17 , at a lookback time of @xmath18 gyr ,
a number of clusters exist where the formation of a massive cluster out of the large scale structure can be witnessed .
this is an epoch of rapid buildup in the cluster population .
observations of the evolution of the cluster mass function indicate that the _ number _ of massive clusters increases by a factor of a few since such redshifts @xcite . in addition , as noted above , simulations suggest that individual clusters grow in _ mass _ by a factor of @xmath16 over the same time period @xcite .
this changing landscape therefore allows one to study the properties of galaxy populations that will transform , evolve , and assemble into the present day cluster population . on much smaller physical scales ,
the star forming activity in galaxies was also different at these intermediate redshifts .
in general , galaxies formed stars at a higher rate in the past @xcite , particularly higher mass galaxies that will have ceased star formation by the present epoch @xcite .
_ at a redshift of @xmath17 , the contrast between star - forming and non - star - forming galaxies is accentuated , allowing us to better highlight those environments that influence star formation . _
we focus in this paper on a mass - limited sample for this important epoch of cluster mass growth .
luminosity - limited samples selected in the rest - frame uv include many bright , low mass , star forming galaxies as a result of the large scatter in their mass - to - light ratios at a fixed galaxy stellar mass .
selecting by stellar mass thus provides a more well - defined and homogeneous sample . in this paper
, we study the colors of galaxies in the @xmath0 cluster rx j0152.7 - 1357 ( hereafter rx j0152 - 13 ) and its outskirts .
this is a very rich cluster with extensive imaging and spectroscopic datasets that now cover a very wide area ( out to @xmath19 mpc in radius ) . as such it is an ideal target for a study of galaxy evolution and transformation over a wide range of densities , from the core to the field .
our initial spectroscopic sample of members is currently @xmath20 objects .
these galaxies are the focus of our first paper on this cluster .
such samples are becoming large enough to provide robust constraints on cluster galaxy properties ( e.g. mass functions , star formation rates , morphologies , environmental trends , etc ) .
the use of spectroscopic redshifts , as opposed to photometric ones , ensures that our membership list , and therefore results , do not suffer from the systematics that trouble photometric surveys .
ultimately we expect to reach 2000 - 3000 spectroscopic members between two clusters at @xmath17 .
the remarkable feat employs the new low - dispersion prism ( ldp ) in the inamori magellan areal camera and spectrograph ( imacs ) on the baade 6.5 m telescope at magellan ( see below for more details ) , further enhancing our insight into galaxy evolution in a variety of environments at @xmath17 . in
[ observations ] we describe the data employed in this wide - field observing campaign . in [ analysis ] the derived properties of galaxies in our sample are discussed .
galaxy colors and their trends with clustercentric radius and local density are presented in
[ results ] .
the implications of these results are discussed in [ discussion ] .
we summarize our findings in [ summary ] . throughout this paper ,
magnitudes from sloan filters will be on the ab system , and johnson filters on the vega system .
the use of both systems is a consequence of adhering to convention and making comparisons to previous work more manageable .
we assume a @xmath21cdm cosmology with @xmath22 km s@xmath23 mpc@xmath23 , @xmath24 , and @xmath25 , resulting in a look - back time of 7.0 gyr and an angular scale of 0.46 mpc arcmin@xmath23 at the cluster redshift , @xmath0 .
rx j0152 - 13 was discovered as part of the warps survey with _
@xcite and was also independently identified in the rdcs @xcite .
kinematic @xcite and x - ray @xcite studies suggest significant substructure in rx j0152 - 13 , most notably the presence of two distinct cores separated spatially by 1.5 and in redshift by @xmath26 ( @xmath27 km s@xmath23 ) . a color image of the cluster created with hst acs imaging using the f625w , f775w , and f850lp filters is shown in figure [ rxjcolorim ] ( see * ? ? ? * for a description of these data ) .
mass estimates for the combined system differ based on the assumptions made and typically range from @xmath28 @xmath7 @xcite to @xmath29 @xmath7 @xcite .
@xcite identified several groups in the outskirts of the cluster , some of which will fall in by @xmath15 .
those galaxies in such high density regions in the cluster outskirts will be discussed in later sections . @xcite
measured the mean redshift of all early - type galaxies in the two cores of rx j0152 - 13 to be @xmath0 .
we assume this value to represent the redshift of the cluster .
we used archival subaru suprime - cam imaging @xcite of rx j0152 - 13 @xcite in @xmath30@xmath31@xmath4@xmath32 for photometry and spectroscopic target selection .
the suprime - cam imaging was made up of 10 pointings covering a @xmath3329 @xmath34 39 field of view .
the data were reduced using standard techniques for image reduction , including steps for overscan subtraction , flat - fielding , and stacking with cosmic - ray rejection .
the photometry was calibrated using frames obtained from the archive of standard stars taken on the same night .
@xcite measured a 5@xmath35 limiting @xmath4-band depth for a @xmath36 aperture to be 26.1 mag , which is significantly deeper than our spectroscopic selection ( [ spectroscopy ] ) .
we used sextractor @xcite to identify objects and daophot ii @xcite to perform circular aperture photometry because of its superior estimate for the sky background .
colors were defined with a fixed circular aperture size of @xmath372 , which was twice the seeing fwhm .
we also used a fixed aperture size of @xmath38 for total magnitudes .
given the 0.6 seeing , the @xmath38 aperture encompassed most of the light ( @xmath39 ) for the typical galaxy at the cluster redshift in the suprime - cam imaging .
for the brightest galaxies this may slightly underestimate luminosities and therefore stellar masses , but not enough to exclude them from our mass - limited sample ( [ masses ] ) .
the sample of galaxies discussed in this paper incorporates cluster members with spectroscopic redshifts from the catalogs of @xcite and @xcite .
these observations focused on the core of the cluster and groups in the outskirts .
additional members with redshifts from keck and magellan spectroscopy were also included . galaxies with redshifts from these high - resolution spectroscopy catalogs in the range @xmath40
are shown as blue circles in figure [ map0152 ] . to expand the spectroscopic sample of galaxies in the vicinity of rx j0152 - 13
, we employed a unique instrument that allowed us to gather @xmath41 redshifts per pointing over a @xmath3327 arcmin diameter field of view ( @xmath3312 mpc at the cluster redshift ) : a low - dispersion prism ( ldp ; built for the primus collaboration by s. burles , private communication ) in the inamori magellan areal camera and spectrograph ( imacs ; * ? ? ?
* ; * ? ? ?
* ) on the baade 6.5 m telescope at magellan . in this paper we present results from the first set of observations , which have already allowed us to build one of the largest
_ spectroscopic _ samples of galaxies in a cluster and its outskirts ( now 754 members ) at intermediate redshift . over subsequent seasons
, we expect to obtain 2000 - 3000 members in and around two clusters at @xmath17 . with the ldp ,
imacs provided a spectral resolving power of @xmath42 over a corresponding wavelength range of 4500 @xmath43 9000 .
the data were taken using nod & shuffle for sky subtraction @xcite with the slits being just long enough to accommodate two object spectra ( a and b ) with each nod .
wavelength calibration was performed using helium lamp exposures taken between science frames , providing nearly a dozen features from 3888 to 10,830 .
the positions of the lines were solved for simultaneously in order to account for blending .
we then constructed mappings of wavelength to position on the ccd using trivariate polynomials of slit mask position and wavelength .
one - dimensional spectra were constructed by simultaneously extracting the a and b pairs using an optimal extraction technique similar to @xcite .
the spectra were flux calibrated .
we used a total of 5 slitmasks , each with @xmath332000 slits .
the first 3 masks selected galaxies from our suprime - cam photometry catalog with @xmath4 @xmath44 23 mag , while the last 2 went deeper to @xmath4 @xmath44 23.75 mag .
roughly @xmath45 of the objects had @xmath46 hrs of exposure and @xmath47 had @xmath48 hrs of exposure .
the cumulative time of observing all 5 masks would have been equivalent to about 3 nights on the telescope .
the efficiency of this setup is clear and allows us to obtain an order of magnitude more redshifts per night compared to traditional multi - object spectroscopy .
we stress that _ no priors , such as red - sequence membership or photometric redshifts , were implemented in our spectroscopic selection , thereby minimizing biases in our sample_. observations with the ldp are ongoing and we therefore plan to discuss a larger and more complete sample of cluster galaxies in a future paper .
until now , surveys of clusters at high redshift have relied , to some degree , on photometric redshifts ( photo - z s ) to build samples large enough to conduct studies with adequate statistics @xcite .
such studies have suggested that selection effects associated with photo - z s must be considered when interpreting results .
for instance , @xcite show that photo - z s for blue galaxies are systematically lower than their spectroscopic values , and in the case of rx j0152 - 13 , results in a deficit of such galaxies in the cluster . with the ldp , we are able to avoid both the systematics and the catastrophic redshift failures that complicate photometric studies . a detailed discussion of the methodology and errors of ldp redshift determinations will be presented in a future paper ( patel et al .
2009 , in preparation ) but is summarized here .
the ldp spectra are low resolution grism spectra taken through short slitlets .
the advantage of this observing technique is that spectra for very large samples of galaxies can be obtained quickly .
the resolution varies from @xmath49 in the blue to @xmath50 in the red ( for fwhm @xmath51 pixels ) .
such low resolution requires non - standard data reduction and processing in order to recover accurate redshifts .
however , as shown below through comparison with a substantial sample of galaxies observed spectroscopically at higher resolution , this technique has proven to return large numbers of spectra quickly and reliably , and has great potential for future studies of intermediate and high redshift galaxies .
redshifts for the ldp sample of galaxies were determined by simultaneously fitting the suprime - cam photometry and ldp spectrophotometry with a grid of template spectra generated using ( * ? ? ? * bc03 ) .
the templates used constant star formation models with the onset of star formation occurring at @xmath52 .
the grid of template spectra was constructed as a function of 3 parameters : redshift , metallicity , and truncation time .
the truncation times span a wide range in order to mimic galaxies that are passively evolving to constantly star - forming .
the fitting included a non - negative gaussian component at the position of [ oii ] 3727 .
at each grid point , a non - negative least squares algorithm was used to fit the stellar and [ oii ] 3727 components to the observed spectral energy distribution ( sed ) of a given galaxy .
the quality of the fit at each grid point was determined by @xmath53 .
the template spectra were convolved with the prism s wavelength - dependent resolution and dispersion prior to fitting the observations .
we note that more recent trials of sed fitting have tested the inclusion of varying amounts of extinction due to dust but no discernible difference in redshifts have been found .
we plan to discuss [ oii ] 3727 measurements from the sed fitting in a future paper .
we use a likelihood analysis to identify members of the rx j0152 - 13 superstructure . the redshift likelihood function for each galaxy
is defined such that the probability density at a given redshift , @xmath54 , is proportional to @xmath53 of the template fit to the sed in the following way : @xmath55 , and the likelihood for a galaxy to be in the redshift range @xmath56 is given by @xmath57 .
the redshift likelihood function is then used to find all galaxies with @xmath5850% of their likelihood within the redshift interval @xmath40 .
this redshift interval has been used in previous studies to discuss galaxies in rx j0152 - 13 and its outskirts @xcite .
ldp galaxies with redshifts in the range @xmath40 are shown as red circles in figure [ map0152 ] .
we found 649 galaxies in rx j0152 - 13 and its outskirts with the ldp out of 7659 extracted spectra from the 5 masks ( roughly @xmath59 of these 7659 had well constrained redshifts ) .
the median value for the full width of the 95% redshift confidence interval for the cluster members is @xmath60 , corresponding to a 1-@xmath35 uncertainty of @xmath61 . comparing 174 ldp redshifts , @xmath62 , to those derived from higher resolution spectra , @xmath63 , we find a scatter of @xmath64 , implying that our error estimates for ldp redshifts are robust .
the histogram of @xmath65 is shown in figure [ redshifts_hist ] .
note that there is a small , but systematic offset from @xmath66 of @xmath67 , which is accounted for in assigning membership .
the scatter for red - sequence galaxies , @xmath61 , is slightly smaller than galaxies with bluer colors , @xmath68 .
of the 174 with measurements of both @xmath62 and @xmath63 , 4 ( 2.3% ) have @xmath63 outside of the cluster redshift interval but @xmath62 inside of it .
however , none of these 4 galaxies deviate from the boundaries of the cluster interval by more than @xmath69 .
thus , the catastrophic redshift failures known to trouble photometric redshift samples are substantially reduced for ldp - selected members .
in addition , we expect the agreement between @xmath62 and @xmath63 to improve in the future as we refine the reduction and fitting of ldp spectra and track down systematic sources of uncertainty .
we incorporate 105 additional members , from existing spectroscopic catalogs , that have not yet been observed with the ldp , to build a sample of 754 galaxies at the redshift of the cluster with @xmath4@xmath70 mag .
to correct for the non - uniform depth of the selection between the first and second sets of spectroscopic data , we constructed the spectroscopic completeness as a 2-d function of @xmath4 and @xmath71 .
this is shown in figure [ fig_completeness ] .
galaxies are binned in @xmath4 in increments of 0.2 mag and in @xmath71 by 0.1 mag .
the completeness fraction for a particular color and magnitude bin is given by @xmath72 , where @xmath73 is the number of galaxies with sufficient constraints on their redshifts to assign membership and @xmath74 is the number of galaxies in the object catalog with @xmath4@xmath75 mag .
this magnitude limit for the spectroscopic target selection is much brighter than the photometric limit of the suprime - cam catalog . in our analysis below we weigh each galaxy with the inverse of the completeness fraction in its color - magnitude bin .
we transform colors and magnitudes to the rest - frame using the same method as @xcite and @xcite . using a range of @xmath76-model spectra from bc03
, we calculate magnitudes in the rest - frame for the relevant filters used in this paper ( @xmath77@xmath78@xmath79 ) and in the observed redshifted frame using suprime - cam @xmath30@xmath31@xmath4@xmath32 .
while these models are different from the constant star forming models used to fit seds and obtain redshifts , their use makes our rest - frame colors and magnitudes consistent with previous work .
also , our results are not significantly affected by this choice .
the ldp seds themselves are not used because they are lower s / n than the suprime - cam photometry and contain systematic uncertainties in their flux calibration . in creating our @xmath76-models ,
we used metallicities of 0.4 , 1.0 , and 2.5 @xmath80 , @xmath81-folding timescales @xmath76 ranging from 0.1 to 5 gyr , and elapsed time since onset of star formation ranging from 0.5 to 6 gyr .
these models span the colors seen in our observations for cluster members .
suprime - cam @xmath30@xmath31@xmath4@xmath32 filter transmissions , multiplied by the quantum efficiency of the detector , were used for the observed transmission curves and buser s @xmath82 and @xmath30 @xcite and sdss @xmath77@xmath78 transmission curves @xcite supplied with bc03 were used for the rest - frame .
the best - fit transformations from observed to rest - frame magnitudes and colors are given below .
we follow the convention where rest - frame magnitudes and colors are subscripted with the letter ` @xmath83 ' .
@xmath84 these transformations have a 1-@xmath35 scatter of 0.007 , 0.008 , 0.012 , and 0.016 mag respectively .
note that in the conversion to @xmath85 , the @xmath31 @xmath86 @xmath32 color represents a vega magnitude minus an ab magnitude . to aid the reader ,
the conversion between the two systems for the suprime - cam @xmath31 filter is @xmath87 .
the last term in each transformation , given in parentheses , corrects the derived rest - frame magnitudes for galactic extinction using the dust maps of @xcite .
we use @xmath88 and @xmath89 to compute stellar masses in [ masses ] and
@xmath90 for selecting galaxies to compute local densities in [ localdensities ] .
the @xmath85 color is used to study stellar populations since this combination of filters straddles the 4000 break and is therefore sensitive to recent or ongoing star formation .
our rest - frame @xmath85 colors show a scatter of @xmath330.04 mag when compared to those derived from @xcite who use hst acs data .
although different apertures were used in the two works to compute colors , the quoted scatter should provide a reasonable _ conservative _ estimate of the error in colors , particularly since @xcite found that color gradients were very small at these redshifts .
we therefore adopt @xmath330.04 mag as the uncertainty in @xmath85 colors .
all quantities in this paper have been calculated assuming the redshift for all members is @xmath0 . however , the actual redshifts range over @xmath40
. therefore properties such as rest - frame color and stellar mass ( [ masses ] ) will deviate slightly from those values derived using the transformations given above ( also see * ? ? ?
we find _ maximum _ deviations , resulting from the range in redshifts , in rest - frame @xmath85 color of @xmath91 mag and stellar mass of @xmath91 dex .
@xcite found stellar mass - to - light ratios and rest - frame colors to be well correlated .
@xcite explored the utility of this correlation for estimating the stellar masses of large samples of galaxies with a more formalized framework in @xcite .
for this paper we therefore compute stellar masses for galaxies using the @xcite relation between rest - frame @xmath92 color and @xmath93 : @xmath94 . @xcite use a `` diet '' salpeter imf in computing @xmath95 .
this imf has the typical salpeter slope ( @xmath96 ) between @xmath97 but is flat ( @xmath98 ) below @xmath99 @xmath7 .
this results in the same luminosity output at 70% of the mass when compared to a traditional salpeter imf .
@xcite provide conversions for @xmath95 derived from a diet salpeter imf to various other imfs . assuming a distance modulus for rx j0152 - 13 of 43.61 mag , and @xmath100 mag , the mass estimates are given by the formula below .
@xmath101 according to @xcite , @xmath95 uncertainties are on the order of 0.1 - 0.2 dex .
@xcite find a scatter of 0.25 dex about dynamical mass estimates for cluster galaxies at @xmath17 .
because our masses were computed in a similar way , we assume this value represents our uncertainty in stellar masses .
the top panel of figure [ colormass ] shows the rest - frame @xmath85 color plotted against our derived masses .
the distribution of masses is given in the bottom panel .
for galaxies with mass @xmath6 @xmath7 ( @xmath810.6 ) , the sample is representative of the full object catalog with a uniform completeness of @xmath5850% .
this value for our mass limit represents @xmath102 ( * ? ? ?
* @xmath103 @xmath7 , ) . for the remainder of this paper ,
we focus on galaxies above this stellar mass limit , @xmath6 @xmath7 ( @xmath810.6 ) .
finally , we would like to note that in this work quantities such as rest - frame color and mass - to - light ratio were not computed from the best - fitting truncation model template used to determine redshifts . instead , these quantities were computed from model star formation histories ( sfhs ) with exponentially declining star formation rates ( i.e. tau - models ) .
this allowed comparison to previous work to be more accessible .
for this paper , we viewed the redshift determination to be an independent task and therefore chose a sfh for the model templates that : ( 1 ) provided accurate redshifts and ( 2 ) optimized computing time given the large amount of data that needed to be processed ( truncated sfhs use one fewer free parameter than tau - models ) . in future work , we may explore integrating the entire process of redshift determination and extraction of rest - frame quantities . the colors of galaxies provide insight into the ages of their stellar populations .
red galaxies are typically associated with old stellar populations , like those found in ellipticals and s0s .
galaxies that actively form stars have blue colors and are typically comprised of spirals and irregular systems .
these late - type galaxies however may also appear red .
@xcite find @xmath104 of field red - sequence galaxies at @xmath17 with mass @xmath6 @xmath7 ( @xmath810.6 ) are late - types . the presence of dust and the viewing angle often play a hand in the appearance of these reddened late - type galaxies , which might otherwise appear blue . galaxies with intermediate colors between the blue - red color bimodality have been linked to galaxies in transition from blue to red after star formation is quenched @xcite but they are also known to be dust - obscured star forming galaxies @xcite .
we define whether a galaxy is on the red - sequence based on its deviation from the color - mass relation ( cmr ) shown in figure [ colormass ] ( red line ) .
this cmr was derived by iteratively fitting a line to galaxies with mass @xmath6 @xmath7 ( @xmath810.6 ) and rejecting 2@xmath35 outliers with each iteration .
the 2@xmath35 scatter of non - rejected galaxies about the converged cmr was used to distinguish red and blue galaxies , with blue galaxies having @xmath85 colors @xmath105 below the cmr ( i.e. approximately @xmath106 mag or more below the ridge line ) .
this division between red and blue galaxies is shown in figure [ colormass ] with a dashed red line .
`` blue '' galaxies therefore include the blue cloud as well as any intermediate color objects .
red galaxies account for @xmath107 of the mass - limited sample .
the local environment , or local galaxy density , has been shown to play a role in the evolution of galaxies @xcite .
the effectiveness of different mechanisms responsible for such evolution often depends on the local environment . to characterize the local environment for galaxies in our sample we computed the local projected galaxy density , @xmath108 , at each galaxy location .
we use an @xmath109-th nearest neighbor scheme , where @xmath110 , to compute @xmath108 . the local density at a given galaxy location is therefore given by @xmath111 where @xmath112 is the projected distance to the 7-th nearest neighbor .
note that this is a physical density ( as opposed to a comoving density ) .
given the potential for large spatial variations in the completeness of our spectroscopic survey , we used a multi - color - selected list of galaxies compiled from the suprime - cam photometry to compute @xmath112 as follows : ( 1 ) the @xmath113 vs. @xmath71 and @xmath114 vs. @xmath115 colors for cluster members were used to derive the locus of galaxy colors at the redshift of the cluster by fitting a 2nd order polynomial to the members in each color - color diagram .
( 2 ) galaxies in the suprime - cam photometry catalog with colors within the 3@xmath35 scatter about these loci were used to construct the multi - color - selected list of galaxies used for computing local densities .
this multi - color selection served to efficiently subtract contaminating background galaxies .
( 3 ) we excluded galaxies that were fainter than @xmath116 .
this magnitude limit corresponds to @xmath117 , where @xmath118 .
this is the same depth in the luminosity function used by previous authors in computing local densities , see @xcite and references therein .
we followed @xcite in using @xmath119 , and in implementing a redshift dependent correction to account for passive evolution in the characteristic magnitude of the luminosity function between @xmath0 and @xmath15 .
the magnitude limit is also very close to @xmath120 for a red - sequence galaxy at our mass limit .
local density contours are shown in gray in figure [ map0152 ] for @xmath121 10 , 20 , 30 , 60 , and 100 mpc@xmath122 .
the multi - color - selected list represents a combination of galaxies with redshifts that are in the cluster redshift interval and contaminants that are not .
a subset of galaxies in the multi - color - selected list have redshifts measured from high - resolution spectroscopy ( @xmath63 ) .
an analysis of this subset reveals that contaminants with redshifts outside of the interval , @xmath40 , account for @xmath123 of the list . after assigning a redshift ( @xmath63 or @xmath62 )
, when available , to objects in the multi - color - selected list , we find that 18% of galaxies contain at least 1 contaminant within their 7-th nearest neighbor circles .
furthermore , only @xmath332% of galaxies have more than 50% of the objects within their 7-th nearest neighbor circles identified as contaminants .
we conclude that for the vast majority of galaxies in our study , local density measurements are not significantly impacted by contaminants that exist in the multi - color - selected list , therefore we do not remove the small number of known contaminants since the high - resolution spectroscopic data do not uniformly cover the entire field .
in this work , we study the properties of galaxies inside and outside of the two cores .
the boundaries are given by their virial radii , @xmath124 , the radius within which the average density is 200 times the critical density of the universe at @xmath0 .
we use the velocity dispersions of the two cores to determine @xmath124 with the relation given in @xcite : @xmath125 here , @xmath35 is the velocity dispersion , @xmath126 is the hubble constant at the cluster redshift and @xmath124 is the virial radius one would measure given three spatial dimensions of information .
the observations however rely on 2-d information .
when projected onto a 2-d plane , the average radius of a ray with length @xmath127 emanating from the center of a sphere to its surface is equivalent to @xmath128 @xcite . as a result
, we define the observed virial radius as : @xmath129 we use the established convention where lowercase letters denote physical 3-d quantities while capitalized ones represent projected values on a 2-d surface .
the superscripted ` @xmath130 ' in the expression above further implies that the spherical projection factor of ( @xmath131 ) has been applied . note that this definition for the virial radius is different from most other works .
it is used to minimize contamination in the two cores from non - virialized portions of the halos and field interlopers , which make up an increasing portion of the sample at larger clustercentric radii .
the velocity dispersions of the northern and southern cores are @xmath132768 km s@xmath23 and @xmath132408 km s@xmath23 respectively @xcite , which imply projected virial radii of @xmath133 = 0.76 mpc and @xmath134 = 0.40 mpc .
these radii are represented by the two overlapping inner cyan - colored circles in figure [ map0152 ] .
we define @xmath135 to mark the boundary given by the virial radii of the two cores and therefore when discussing galaxies at clustercentric radii @xmath136 we are referring to those galaxies within both cores . for calculating distances from the centers of these two cores , we use their chandra x
- ray centroids : ( @xmath137 ) = ( 1:52:44.18 , -13:57:15.84 ) for the northern clump and ( @xmath137 ) = ( 1:52:39.89 , -13:58:27.48 ) for the southern clump @xcite .
the central position of the cluster as a whole lies between the two cores and is taken to be ( @xmath137 ) = ( 1:52:41.669 , -13:57:58.32 ) @xcite . the clustercentric radius for galaxies , @xmath31 ,
is defined from this position .
all ( @xmath137 ) are in j2000.0 coordinates .
in the following sections we examine the colors of galaxies in different radial and density regimes of rx j0152 - 13 . the two known agn detected in x - ray observations
@xcite are not included in our analysis because their optical colors are most likely contaminated . in figure [ crr ]
we show @xmath85 color plotted against projected clustercentric radius for galaxies in three mass bins above @xmath6 @xmath7 ( @xmath810.6 ) . the colored lines represent smoothed median colors for galaxies in these three mass bins .
high mass galaxies lie on the red - sequence at all radii , while lower mass galaxies show bluer colors at larger radii where the local density is lower .
this suggests that as galaxies fall into the high density cluster environment , where most galaxies above our mass limit are red , any evolution that takes place occurs in the lower mass galaxies in our sample . in figure [ redseqfrac ]
we show the fraction of galaxies with red or blue colors as a function of projected clustercentric radius for the mass - limited sample .
the radial bin sizes are indicated by the horizontal error bars and were chosen to provide roughly uniform signal - to - noise in each bin .
the projected radius of each data point represents the median for a particular bin .
error bars were computed using a binomial distribution .
note that the apparent overlap between the two smallest radial bins is a consequence of assigning all galaxies within the two virialized cores , which extend to different _ clustercentric radii _
, to a single data point ( _ open circle _ ) .
galaxies are not being double counted in these two radial bins .
the colored lines represent the color fraction at more highly sampled radial intervals .
they are computed using a sample of the 30 nearest galaxies at a given radius .
the uncertainty in the color fractions for these highly sampled radial intervals is @xmath339% .
note that given our definition of galaxy color , blue galaxies will exhibit a complementary trend to red galaxies and are plotted here for illustrative purposes .
@xcite find a red - sequence fraction for the goods field sample at @xmath17 of @xmath138 when adjusted to rest - frame @xmath139 color ( as opposed to @xmath140 in their work ) . the shaded red band in figure [ redseqfrac
] represents their result . within the two virialized cores ,
the red galaxy fraction is @xmath141 .
moving to larger clustercentric radii , the fraction of red galaxies decreases until @xmath142 mpc . beyond @xmath12
mpc the red fraction is roughly constant at @xmath10 .
this relatively high fraction highlights the importance of stellar mass in galaxy evolution at @xmath17 ; above our mass limit , the majority of galaxies that will fuel the future growth in mass of the cluster are already red .
the value for the red galaxy fraction at @xmath12 mpc is consistent with the field measurement of @xcite , implying that a field population is reached at @xmath143 times the virial radius of the more massive northern clump , @xmath133 .
@xcite analyzed a sample of clusters at low redshift and also found that the red galaxy fraction reaches a constant value at similar distances .
the red line representing the red fraction at highly sampled radial intervals in figure [ redseqfrac ] reveals localized regions of high or low red galaxy fractions .
this can also be seen by eye in figure [ crr ] .
the peaks in the red galaxy fraction at @xmath12 mpc occur at @xmath144 mpc and @xmath145 mpc .
these are the clustercentric radii of overdense regions identified as groups by @xcite .
panel ( @xmath146 ) of figure [ redseqfrac ] , which shows local galaxy density vs. clustercentric radius , confirms the presence of over - densities at these radii . in the next section
we argue that the elevated red galaxy fraction at these radii is a consequence of the local environment . in this section
we analyze the red and blue galaxy fractions as a function of local density . in doing so
, we correct the color fractions for contamination from field interlopers , i.e. those with redshifts within our chosen window ( @xmath40 ) , but which lie in the line - of - sight far from the cluster . in the previous section , we found that the fraction of red galaxies reached the value found in field surveys at @xmath11 mpc .
we therefore use those galaxies with @xmath12 mpc and densities of @xmath147 to form the field interloper sample ( see appendix for details ) .
the low density criterion is necessary in order to assess the level of contamination in intermediate and high density enhancements at @xmath12 mpc .
the `` field '' value for the red galaxy fraction , @xmath148% , is subsequently used in the discussion , figures and table below . in figure
[ colordensityall]@xmath149 , we show @xmath85 color vs. local density for galaxies with mass @xmath6 @xmath7 ( @xmath810.6 ) .
as expected from the red points in figure [ crr ] , the highest mass galaxies lie on the red - sequence at all densities , while we see more clearly here the result that was hinted at in figure [ crr ] , namely that galaxies of lower mass exhibit a broader range in colors at lower densities . also shown in figure [ colordensityall]@xmath146 are the red and blue galaxy fractions vs. local density .
both the observed color fractions ( _ solid circles _ ) and the field interloper - corrected fractions ( _ open circles _ ) are shown .
the red galaxy fractions are given in table [ table_colordensity ] . a clear trend of galaxy colors exists with local density as expected from the correlation between clustercentric radius and local density at @xmath150 mpc where a large number of galaxies in our sample lie . at low density there is a roughly equal mix of red and blue galaxies , while at the highest densities , red galaxies are dominant .
this trend of an increasing fraction of red galaxies with density is consistent with what has been seen in previous studies of color trends with density at low redshift @xcite .
the color - density relation in figure [ colordensityall]@xmath146 is an analog of the mass - limited morphology - density relation @xcite ; which is related to the @xcite luminosity - based mdr . as is generally recognized , galaxy colors and morphology track one another ( see e.g. [ colorfractions ] ) . @xcite computed the mass - limited mdr for clusters spanning redshifts of @xmath151 for the same stellar mass - limit as in this work .
the mass - limited mdrs in that work showed weak correlations between early - type fraction and density .
our survey reaches to larger clustercentric radii and so we are able to probe galaxy properties to much lower local densities .
interestingly , we find a stronger dependence of red galaxy fraction on local density when including densities that are lower than what could be investigated in @xcite .
this is consistent with the mass - limited mdr of @xcite , which shows a large range in early - type fraction with density when combining their low density field sample with the high density cluster sample of @xcite .
the large range in red galaxy fractions seen in our uniformly selected sample therefore confirms the morphological trends seen in the work of @xcite .
lcccc all & @xmath152 ( 34 ) & @xmath153 ( 163 ) & @xmath154 ( 77 ) & @xmath155 ( 26 ) + ( corr ) & & @xmath156 & @xmath157 & @xmath158 + @xmath136 & & @xmath159 ( 5 ) & @xmath160 ( 36 ) & @xmath161 ( 23 ) + ( corr ) & & @xmath162 & @xmath163 & @xmath164 + @xmath12 mpc & @xmath165 ( 27 ) & @xmath166 ( 100 ) & @xmath167 ( 15 ) & + ( corr ) & & @xmath168 & @xmath169 & + in order to gauge the role of local environment on galaxy evolution , we separate galaxies into two groups , as shown in figures [ colorfracdensity1 ] and [ colorfracdensity45 ] .
the first sample includes galaxies in the two cores ( @xmath136 ) while the second is those at larger clustercentric radii ( @xmath12 mpc ) .
the sample at @xmath12 mpc is far enough away from the region where density and clustercentric radius are correlated ( @xmath150 mpc , see figure [ redseqfrac]@xmath146 ) that local environmental effects ( e.g. galaxy - galaxy interactions , etc . )
can potentially be untangled from more global processes associated with clusters ( e.g. ram - pressure , harassment , etc . ) .
this is a particularly important benefit and feature of studying galaxy evolution at large radii , yet still in the cluster environment .
galaxies in the outskirts will , with high likelihood , be those that ultimately populate the core and will therefore be progenitors of typical cluster red - sequence galaxies .
the red galaxy fractions for both radial regimes are summarized in table [ table_colordensity ] . in the two cores ,
the red galaxy fraction is near unity for all densities ( figure [ colorfracdensity1]@xmath146 ) .
this is true regardless of the stellar mass of the galaxy ( figure [ colorfracdensity1]@xmath149 ) . at larger clustercentric radii ( @xmath12 mpc )
, galaxies obviously populate lower density environments ( figure [ colorfracdensity45 ] ) .
it is a fascinating result that high mass galaxies in this radial regime are predominantly red regardless of density , while a large fraction of lower mass ( but still massive ) galaxies are found to have blue colors in the two lowest density bins ( figure [ colorfracdensity45]@xmath149 ) . when considering all galaxies above our mass limit at @xmath12 mpc , we find an increasing fraction of red galaxies towards higher densities . for galaxies that occupy the highest density regions ( @xmath170 ) ,
the red galaxy fraction is similar to the value found in the two cores for the same density range .
this suggests that galaxies in high density environments in the outskirts of clusters have already evolved by @xmath17 to be similar , and to occur in similar proportions , to galaxies that occupy the two cores .
the presence of an increasing red galaxy fraction trend with local density implies that certain environments nurture or drive the transition from blue to red at a level above what is found in the field .
the red galaxy fraction in the density range @xmath171 at @xmath12 mpc is between the value found in adjacent density bins : in the higher density bin , most galaxies are already evolved to the canonical core fraction while in the lower density bin , we have reached the field population level where the red fraction is low ( at @xmath172 for our sample ) .
the density range @xmath171 therefore represents a regime where one expects the local environment to influence evolution above a level that is typical for the field and where star formation begins to shut off in galaxies , causing them to move towards redder colors .
given the large number of galaxies at these densities , a potentially large sample exists of galaxies that are beginning to transition from blue to red in these particular environments .
this sample will be a major resource for assessing the physical processes that drive galaxies from field - like galaxy properties to cluster - like galaxy properties .
additional observations to complement galaxy colors , such as those that allow morphological identification , estimates of star formation rates , structural measurements and assessments of merger events would be useful in characterizing such galaxies .
the simulations of @xcite suggest that massive clusters like rx j0152 - 13 will grow by more than a factor of two in mass between @xmath17 and today .
such massive clusters formed out of the highest density perturbations from the primordial fluctuations in the early universe . as a consequence , these high density perturbations were surrounded by intermediate density perturbations that have collapsed to form structures on the group mass scale that are subsequently accreted into the cluster .
therefore , much of the mass buildup in massive clusters like rx j0152 - 13 occurs through the accretion of nearby infalling groups .
the n - body simulations of @xcite however , suggest that clusters grow primarily through the accretion of individual galaxies from the field
. however , they also note that for more massive clusters , such as rx j0152 - 13 , a larger proportion of galaxies are accreted as part of group sized halos .
most of the galaxies above our mass limit at @xmath12 mpc and at high densities ( @xmath170 ) lie in two groups identified by @xcite .
as in the two cluster cores , the fraction of red galaxies in these groups is @xmath173 .
one group is located at a projected separation of @xmath174 ( @xmath51 mpc ) south of the cluster center and the other @xmath175 ( @xmath19 mpc ) southwest of the cluster center ( see figure [ map0152 ] ) .
@xcite refer to these systems as f4 and f6 respectively and measure velocity dispersions of @xmath176 km s@xmath23 and @xmath177 km s@xmath23 .
their offsets in redshift from the cluster ( @xmath178 and @xmath179 ) imply relative velocities ( @xmath180 ) of @xmath181 km s@xmath23 and @xmath182 km s@xmath23 , which means that both are encompassed by the @xmath183 velocity dispersion of the northern core .
if their redshift offsets instead are due to expansion with the hubble flow , the line of sight distance ( @xmath184 ) from the cluster would be @xmath185 mpc and @xmath186 mpc respectively .
the f4 group is therefore very likely to be falling into the cluster and given that it has a projected distance from the cluster center of @xmath51 mpc , it would nt be unreasonable for the f6 group , with a projected distance of @xmath19 mpc , to also be falling into the cluster , though this outcome is less certain .
evidence that the source of the constituent galaxies of massive clusters ( like rx j0152 - 13 ) are groups in the outskirts has also been seen in previous work .
@xcite found that assembling clusters at intermediate redshift had a high fraction of elliptical galaxies and was similar to the fraction found in virialized clusters , suggesting that many ellipticals were in place prior to virialization of massive clusters and probably formed earlier , in the group phase . at @xmath187
, @xcite found that at least four gravitationally bound groups were in the process of collapsing to form a larger cluster . a higher proportion of galaxies in these groups were found to be passive when compared to the field , similar to our findings .
others have also found groups or overdense regions in the outskirts of clusters to support galaxies with properties that are similar to those found among galaxies in cluster cores .
for instance , @xcite found passive spirals in the outskirts of a @xmath188 cluster to be preferentially located in overdense regions , perhaps in transition to becoming s0s .
meanwhile , @xcite found the morphological mix of overdense regions in the outskirts of a @xmath189 cluster to be very similar to the central regions .
regardless of their proximity to the two main cores of rx j0152 - 13 , the groups in the redshift interval @xmath40 , contain an enhanced fraction of red galaxies .
the velocity dispersions of the f4 and f6 groups imply virial radii , @xmath124 ( i.e. unprojected ) , of @xmath190 mpc and @xmath191 mpc respectively .
the median local density of galaxies in these groups is @xmath192 mpc@xmath122 .
recall that these densities are computed for galaxies above a luminosity - limit ( @xmath193 ) . converting this surface density into a true space density with the virial radii computed above
, we find @xmath194 mpc@xmath195 . given that the relative velocities of the galaxies in these groups are much closer to the galaxy interval velocities , and that the galaxies are in regions of high density ,
the frequency of galaxy - galaxy interactions and merging is enhanced .
the merger of disk galaxies can lead to the formation of ellipticals @xcite and processes associated with merger events , such as starbursts @xcite and agn feedback @xcite among others ( see * ? ? ? * and references therein ) , can quench star formation and move galaxies towards the red - sequence .
thus , mechanisms in these group environments exist to develop and enhance a high red galaxy fraction . in intermediate density enhancements (
@xmath171 : somewhat lower in density compared to the groups f4 and f6 discussed above ) at large clustercentric radii , @xmath12 mpc , the red galaxy fraction is also elevated relative to the value in the lowest density bin .
these intermediate densities therefore represent a regime in which galaxies begin to transition towards the red - sequence in larger proportions than what is found at the lowest densities , where the local environment is not effective in influencing galaxy evolution .
galaxies at these intermediate densities are not all obviously identified with large groups in the outskirts such as the f4 and f6 groups , though many are found near them .
instead , they appear to trace the filamentary structure surrounding the cluster , and so may indicate the sites of the earliest stages of transformation to red , quiescent early - type galaxies .
the local environment of a galaxy can be thought of as a tracer for the underlying mass of the dark matter halo in which it resides .
@xcite explored abundance matching and found that a galaxy s _ halo mass _ determines the amount of _ stellar mass _ the galaxy will have at @xmath13 as well as its star formation rate .
their work focuses primarily on the central galaxies in dark matter halos .
the satellite galaxies that occupy subhalos , which are part of more massive group and cluster sized halos , are not specifically addressed in their work .
the ubiquitous red colors of the highest mass galaxies in our sample could reflect the lack of star formation in high mass galaxies found at @xmath13 in this model .
galaxies of lower mass in the model are expected to have higher star formation rates at this epoch .
the low mass galaxies in our sample that reside outside of the two cluster cores and infalling groups support this view given their blue colors .
however , it is plausible , even likely , that these lower mass galaxies will eventually fall into larger mass halos , where a number of mechanisms are capable of changing their structure and halting star formation , some of which were discussed above .
we studied the environmental dependence of galaxy colors in the @xmath0 galaxy cluster rx j0152 - 13 in order to investigate where galaxies transition from blue , late - type systems into red , early - type ones that dominate the central regions .
the cluster is a massive x - ray luminous system that is comprised of two cores .
the northern and southern cores have velocity dispersions @xmath132768 km s@xmath23 and @xmath132408 km s@xmath23 , respectively , and overlapping projected virial radii of @xmath133 = 0.76 mpc and @xmath134 = 0.40 mpc and are separated in velocity along the line of sight by @xmath27 km s@xmath23 . with a low - dispersion prism ( ldp ) on imacs ,
the multi - object spectrograph on the 6.5 m baade ( magellan i ) telescope , we have built one of the largest samples of _ spectroscopically _ confirmed members for a cluster at intermediate redshift .
this was accomplished by targeting galaxies with @xmath4 @xmath5 mag within a @xmath196 ( @xmath197 mpc at @xmath0 ) fov encompassing the two cores . prior to this work
, @xmath33250 galaxies were known to be members of rx j0152 - 13 and its outskirts with redshifts in the range @xmath40 .
we have added 475 new members .
the redshifts derived from the ldp show a scatter of @xmath64 , or 1% in @xmath198 , about those derived from higher resolution spectra ( [ redshifts ] ) , allowing us to localize galaxies within the superstructure of the cluster .
stellar masses for galaxies were computed using the @xcite relation between rest - frame @xmath92 color and @xmath199 ( [ masses ] ) .
we used a mass - limited sample of 300 galaxies with mass @xmath6 @xmath7 ( @xmath810.6 ) to conduct our study .
this mass threshold corresponds to the @xmath4-band magnitude along the red - sequence at which the spectroscopic completeness is @xmath47 ( @xmath4 @xmath200 ) .
because the spectroscopic completeness varies for galaxies of different mass , we applied completeness corrections when necessary ( [ completeness ] ) .
we assigned galaxies to be blue or red based on their deviation from the red - sequence , with blue galaxies having @xmath85 colors @xmath106 mag or more below the red - sequence .
this initial sample provides an opportunity to analyze cluster galaxy properties at intermediate redshift with substantial statistical significance , and is only a glimpse of what will be possible when our survey is complete and we are likely to have over @xmath41 members .
our conclusions are the following : 1 . the most massive galaxies ( @xmath811.3 ) lie on the red - sequence at all clustercentric radii and all local densities , while less massive galaxies have a broad range in colors farther from the two cluster cores and in lower density environments . in this massive cluster at @xmath17
most of the evolution due to environmental effects is occurring in the lower mass galaxies . 2 .
the fraction of galaxies that are on the red - sequence is at a maximum in the two cores of the cluster at @xmath9 .
this red galaxy fraction declines at larger clustercentric radii until @xmath201 mpc , beyond which it is roughly constant at @xmath10 .
our lowest density sample at @xmath202 mpc ( @xmath147 ) has a red fraction of @xmath172 .
these values are broadly consistent with the red galaxy fraction for the field at @xmath17 .
3 . at @xmath12
mpc , where the correlation between clustercentric radius and density is nonexistent , we find that the red galaxy fraction increases with local density .
this implies that the local environment is a factor in galaxy evolution in the outskirts of the cluster .
4 . the high density environments ( @xmath170 ) in the outskirts ( @xmath12 mpc ) have a red galaxy fraction that is similar to what is found in the two cores .
these galaxies are primarily located in two previously identified groups .
a large number of galaxies at @xmath12 mpc are found in our intermediate density bin ( @xmath171 ) , where the red galaxy fraction is elevated relative to the value found for the lowest density bin ( @xmath147 ) .
these intermediate densities therefore represent a regime where galaxies above our mass - limit first begin to undergo transformations from star forming to quiescent as a result of the local environment .
most of these galaxies are not associated with either of the large groups in the outskirts , but instead trace the surrounding filamentary structure .
exploring the outer parts of clusters provides opportunities to identify the regions where galaxies of different masses are growing in mass and/or transforming from blue , star - forming , late - type galaxies to red , quiescent , early - type galaxies .
the uniformity of the colors of the highest mass galaxies at @xmath17 suggests that the vast majority of their star formation has taken place already , even at this epoch , while the diverse colors of the somewhat lower mass sample indicate that the transformation process is still underway in the intermediate density regions in the outer parts of rx j0152 - 13 .
studies of galaxies in such regions of transformation promise to provide great insight into the physical processes that drive galaxy evolution and build up the red - sequence population .
we are in the process of expanding the present sample of galaxies in rx j0152 - 13 with the ldp and we are also targeting another cluster at @xmath17 , ms 1054.4 - 0321 .
as we move forward , the ldp reductions and sed fitting are being refined to provide better constraints on redshifts and other galaxy properties .
we are also in the process of obtaining @xmath203-band imaging , which will allow us to improve our stellar mass estimates . when completed
, this dataset will yield a wealth of information concerning galaxies in overdense environments in and around clusters at intermediate redshift . between the two targeted fields
, we expect to find 2000 - 3000 galaxies in the clusters and their outskirts , allowing us to study the environmental dependence of mass functions with @xmath204 and @xmath205 to significantly higher precision than at present , possibly as high as @xmath206 dex and @xmath207 respectively .
we will also explore stellar population parameters from sed fitting , an analysis of blue galaxies down to low masses , and many other exciting aspects of cluster galaxies at an epoch some 7 - 8 gyr ago when clusters were rapidly building up .
this research was supported by nasa grant nag5 - 7697 and spitzer grant jpl 1277397 .
we wish to acknowledge those who have contributed to the construction and deployment of imacs as well as scott burles for developing the low - dispersion prism , and the primus collaboration for allowing us to investigate clusters with their hardware .
the combination of these two components resulted in the impressive dataset presented in this work .
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 .
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 width of the cluster redshift window , @xmath40 , was chosen to accommodate redshift uncertainties for galaxies determined to be members from imacs spectroscopy .
this window is large enough to allow field interlopers to fall into our sample .
these galaxies have redshifts within the cluster redshift window but are far from the cluster superstructure .
their deviation in redshift from the cluster s reflects expansion with the hubble flow rather than an induced peculiar velocity from the nearby massive structure . to correct the red galaxy fractions for contamination from field interlopers we employ a statistical background subtraction . for a given surface area of size @xmath208 ,
the number of galaxies with density @xmath108 falling into this patch of sky is given by @xmath209 and the number of field interlopers is given by @xmath210 .
the contamination fraction at a particular density @xmath108 is therefore @xmath211 .
we use the median density in a particular density bin to represent @xmath108 when computing @xmath212 . the corrected red galaxy fraction , @xmath213 , and its uncertainty , @xmath214 ,
can then be written in terms of the observed red fraction at a particular density , @xmath215 , the red fraction for field interlopers , @xmath216 , and their respective uncertainties as follows : we define a field interloper sample using those galaxies with @xmath6 @xmath7 ( @xmath810.6 ) at clustercentric radii @xmath12 mpc and densities of @xmath147 .
this sample represents galaxies in the lowest density bin in figure [ colorfracdensity45]@xmath146 where the red fraction is @xmath218 and the uncertainty is @xmath219 .
the median density for this sample is @xmath220 mpc@xmath122 . at @xmath12 mpc ,
the observed red fraction in figure [ redseqfrac ] is roughly constant , implying that a field population has been reached at these clustercentric radii .
we further select galaxies in the lowest density bin to represent field interlopers because we are interested in determining the level of contamination in intermediate and high density clumps at @xmath12 mpc .
in some cases , the corrected value for the red fraction exceeds unity , especially when the _ observed _ red fraction is already close to it .
this effectively results from over - subtraction of blue galaxies when the observed number of blue galaxies is small . under these circumstances ,
we assign the corrected red fraction to be @xmath221 . ,
b. c. & dressler , a. m. 2003 , in instrument design and performance for optical / infrared ground - based telescopes .
edited by iye , masanori ; moorwood , alan f. m. proceedings of the spie , volume 4841 , pp .
1727 - 1738 ( 2003 ) . ,
m. iye & a. f. m. moorwood , 17271738 , a. , hare , t. , bigelow , b. c. , & osip , d. j. 2006 , in ground - based and airborne instrumentation for astronomy . edited by mclean , ian s. ; iye , masanori .
proceedings of the spie , volume 6269 , pp .
62690f ( 2006 ) . | we present the first results from the largest spectroscopic survey to date of an intermediate redshift galaxy cluster , the @xmath0 cluster rx j0152.7 - 1357 .
we use the colors of galaxies , assembled from a @xmath1 mpc region centered on the cluster , to investigate the properties of the red - sequence as a function of density and clustercentric radius .
our wide - field multi - slit survey with a low - dispersion prism ( ldp ) in the imacs spectrograph at the 6.5 m baade telescope allowed us to identify 475 new members of the cluster and its surrounding large - scale structure with a redshift accuracy of @xmath2 and a contamination rate of @xmath3 for galaxies with @xmath4 @xmath5 mag .
we combine these new members with the 279 previously known spectroscopic members to give a total of 754 galaxies from which we obtain a mass - limited sample of 300 galaxies with stellar masses @xmath6 @xmath7 ( @xmath810.6 ) .
we find that the red galaxy fraction is @xmath9 in the two merging cores of the cluster and declines to a level of @xmath10 at projected clustercentric radii @xmath11 mpc . at these large projected distances ,
the correlation between clustercentric radius and local density is nonexistent .
this allows an assessment of the influence of the local environment on galaxy evolution , as opposed to mechanisms that operate on cluster scales ( e.g. harassment , ram - pressure stripping ) . even beyond @xmath12
mpc we find an increasing fraction of red galaxies with increasing local density .
the red galaxy fraction at the highest local densities in two large groups at @xmath12 mpc matches the red galaxy fraction found in the two cores .
strikingly , galaxies at intermediate densities at @xmath12 mpc , that are not obvious members of groups , also show signs of an enhanced red galaxy fraction .
our results point to such intermediate density regions and the groups in the outskirts of the cluster , as sites where the local environment influences the transition of galaxies onto the red - sequence . |
the purpose of this work is to study the influence of a thin material obstacle on the behavior of two - dimensional incompressible ideal flows .
more precisely , we consider a family of obstacles @xmath0 which are smooth , bounded , open , connected , simply connected subsets of the plane , contracting to a smooth curve @xmath1 as @xmath2 .
given the geometry of the exterior domain @xmath3 , a velocity field ( divergence free and tangent to the boundary ) on this domain is uniquely determined by the two following ( independent ) quantities : vorticity and circulation of velocity on the boundary of the obstacle . throughout this paper
we assume that initial vorticity @xmath4 is independent of @xmath5 , smooth , compactly supported outside the obstacles @xmath0 and that @xmath6 , the circulation of the initial velocity on the boundary , is independent of @xmath5 . from the work of k. kikuchi @xcite , we know that there exists @xmath7 a unique global solution to the euler equation in the exterior domain @xmath3 associated to the initial data described above . our aim is to determine the limit of @xmath8 as @xmath9 . as a consequence
, we also obtain the existence of a solution of the euler equations in the exterior of the curve @xmath1 .
the study of incompressible fluid flows in presence of small obstacles was initiated by iftimie , lopes filho and nussenzveig lopes @xcite .
the paper @xcite treats the same problem as above but with obstacles that shrink homothetically to a point @xmath10 , instead of a curve .
the case of navier - stokes is considered in @xcite . in the inviscid case
, these authors prove that if the circulation @xmath6 vanishes , then the limit velocity verifies the euler equation in @xmath11 ( with the same initial vorticity ) .
if the circulation is non - zero , then the limit equation involves a new term that looks that a fixed dirac mass in the point @xmath10 of strength @xmath6 ; the initial vorticity also acquires a dirac mass in @xmath10 . in the case of navier - stokes ,
the limit equation is always navier - stokes but the initial vorticity of the limit equation still has an additional dirac mass in @xmath10 . here
we will show that , in the inviscid case , the limit equation is the euler equation in @xmath12 . the initial velocity for the limit equation is a velocity field which is divergence free in @xmath11 , tangent to @xmath1 such that the curl computed in @xmath13 is @xmath4 and the curl computed in @xmath11 is @xmath14 where @xmath15 is a density given explicitly in terms of @xmath4 and @xmath6 .
alternatively , @xmath15 is the jump of the tangential velocity across @xmath1 .
more precisely , let @xmath16 be a cut - off function in a small @xmath5-neighborhood of the boundary ( the precise definition of @xmath16 is given in subsection [ cutoff ] ) and set @xmath17 .
our main theorem may be stated as follows : [ 1.1 ] there exists a subsequence @xmath18 such that * @xmath19 strongly in @xmath20 ; * @xmath21 weak@xmath22 in @xmath23 ; * @xmath24 is related to @xmath25 by means of relation * @xmath24 and @xmath25 are weak solutions of @xmath26 in @xmath27 .
the limit velocity @xmath24 is explicitly given in terms of @xmath25 and @xmath6 ( see theorem [ 5.6 ] ) and can be viewes as the divergence free vector field which is tangent to @xmath1 , vanishing at infinity , with curl in @xmath13 equal to @xmath25 and with circulation around the curve @xmath1 equal to @xmath6 .
this velocity is blowing up at the endpoints of the curve @xmath1 as the inverse of the square root of the distance . and has a jump across @xmath1 .
moreover , we have @xmath28 in @xmath29 , where @xmath30 is the dirac function of the curve @xmath1 , and the @xmath31 which is defined in lemma [ 5.9 ] depends on @xmath25 and the circulation @xmath6 .
the function @xmath31 is continuous on @xmath1 and blows up at the endpoints of the curve @xmath1 as the inverse of the square root of the distance .
one can also characterize @xmath31 as the jump of the tangential velocity across @xmath1 .
the presence of the additional term @xmath31 in the expression of @xmath32 , compared of the euler equation in the full plane , is compulsory to obtain a vector field tangent to the curve , with a circulation @xmath6 around the curve .
there is a sharp contrast between the behavior of ideal flows around a small and thin obstacle . in @xcite , the authors studied the vanishing obstacle problem when the obstacle tends homothetically to a point @xmath10 .
their main result is that the limit vorticity satisfies a modified vorticity equation of the form @xmath26 , with @xmath33 and @xmath34 . in other words , for small obstacles
the correction due to the vanishing obstacle appears as time - independent additional convection centered at @xmath10 , whereas in the thin obstacle case , the correction term depends on the time .
although treating a related problem , the present work requires a different approach .
indeed , in @xcite , the proofs are simplified by the fact that the obstacles are homothetic to a fixed domain .
indeed , an easy change of variables @xmath35 allows in that case to return to a fixed obstacle and to deduce the required estimates .
this argument clearly does not work here and a considerable amount of work is needed to characterize the conformal mapping that sends the exterior of a small obstacle into the exterior of the unit disk .
moreover , in @xcite the authors use the div - curl lemma to obtain strong convergence for velocity .
this is made possible by the validity of some bounds on the divergence and the curl of the velocity .
a consequence of our work is that these estimates are no longer valid in our case , so this approach can not work .
we will be able to prove directly strong convergence for the velocity through several applications of the lebesgue dominated convergence theorem .
we finally observe that , in contrast to the case of @xcite , the vanishing of the circulation @xmath6 plays no role in our result .
the limit velocity will always verify the same type of equation .
we also mention that lopes filho treated in @xcite the case of several obstacles with one of the obstacles tending to a point , but the author had to work on a bounded domain . in this case
, we do not have explicit formulas anymore , and the conformal mapping technique is replaced by qualitative analysis using elliptic techniques , including variational methods and the maximum principle . the remainder of this work is organized in five sections .
we introduce in section 2 a family of conformal mappings between the exterior of @xmath0 and the exterior of the unit disk , allowing the use of explicit formulas for basic harmonic fields and the biot - savart law , which will be really helpful to obtain sharp estimations . in the third part
, we precisely formulate the flow problem in the exterior of a vanishing obstacle . in section 4
, we collect _ a priori _ estimates in order to find the equation limit in the section 5 of this article .
the last subsection concerns an existence result of the euler equations on the exterior of a curve . for the sake of clarity ,
the main notations are listed in an appendix at the end of the paper .
let @xmath36 and @xmath37 . in what follows we identify @xmath11 with the complex plane @xmath38 .
we begin this section by recalling some basic definitions on the curve .
we call a _
jordan arc _ a curve @xmath39 given by a parametric representation @xmath40 , @xmath41 with @xmath42 an injective ( @xmath43 one - to - one ) function , continuous on @xmath44 $ ] . an _
open jordan arc _ has a parametrization @xmath40 , @xmath45 with @xmath42 continuous and injective on @xmath46 .
we call a _ jordan curve _ a curve @xmath39 given by a parametric representation @xmath47 , @xmath48 , @xmath49-periodic , with @xmath50 an injective function on @xmath51 , continuous on @xmath52 .
thus a jordan curve is closed ( @xmath53 ) whereas a jordan arc has distinct endpoints .
if @xmath54 is a jordan curve in @xmath38 , then the jordan curve theorem states that @xmath55 has exactly two components @xmath56 and @xmath57 , and these satisfy @xmath58 .
the jordan arc ( or curve ) is of class @xmath59 ( @xmath60 ) if its parametrization @xmath42 is @xmath61 times continuously differentiable , satisfying @xmath62 for all @xmath63 , and if @xmath64 for all @xmath65 and @xmath66 .
let @xmath67 be a jordan arc .
then the subset @xmath13 is connected and we will denote it by @xmath68 .
the purpose of this part is to obtain some properties of a biholomorphism @xmath69 .
[ 2.2 ] if @xmath1 is a @xmath70 jordan arc , such that the intersection with the segment @xmath71 $ ] is a finite union of segments and points , then there exists a biholomorphism @xmath72 which verifies the following properties : * @xmath73 and @xmath74 extend continuously up to the boundary , and @xmath73 maps @xmath75 to @xmath1 , * @xmath74 is bounded , * @xmath76 and @xmath77 extend continuously up to @xmath1 with different values on each side of @xmath1 , except at the endpoints of the curve where @xmath76 behaves like the square root of the distance and @xmath77 behaves like the inverse of the square root of the distance , * @xmath77 is bounded in the exterior of the disk @xmath78 , with @xmath79 , * @xmath77 is bounded in @xmath80 for all @xmath81 and @xmath82 .
we first study the case where the arc is the segment @xmath71 $ ] .
we can have here an explicit formula for @xmath76 .
indeed , the joukowski function @xmath83 is a biholomorphism between the exterior of the disk and the exterior of the segment .
it maps the circle @xmath84 on the ellipse parametrized by @xmath85 with @xmath86 , and it maps the unit circle on the segment . remarking that @xmath87
we can conclude that @xmath88 is also a biholomorphism between the interior of the disk minus @xmath89 and the exterior of the segment .
therefore , any @xmath90 $ ] has one antecedent of @xmath88 in @xmath91 and another one in @xmath92 . for @xmath93
$ ] the antecedents are @xmath94 .
therefore , there are exactly two antecedents except for @xmath95 and @xmath49 . in fact
, we have considered the double covering @xmath88 from @xmath96 to @xmath38 , which is precisely ramified over @xmath95 and @xmath49 .
let @xmath97 be the biholomorphism between the exterior of the segment and @xmath92 , such that @xmath98 , and let @xmath99 be the biholomorphism between the exterior of the segment and @xmath100 , such that @xmath101 .
to find an explicit formula of @xmath97 , we have to solve an equation of degree two .
we find two solutions : @xmath102 we consider that the function square - root is defined by @xmath103 with @xmath104 , the argument of @xmath105 , verifying @xmath106 .
it is easy to observe that @xmath107 on @xmath108 and @xmath109 on @xmath110 . despite this
, @xmath97 is @xmath111,{{\rm int}\,}d^c)$ ] because @xmath112 .
therefore in the segment case , @xmath113 and the first two points are straightforward .
an obvious calculation allows us to find an explicit formula for @xmath114 : @xmath115 with the choice of sign as above .
this form shows us that @xmath116 blows up at the endpoints of the segment as the inverse of the square root of the distance , which is bounded in @xmath117 for @xmath81 .
moreover , a mere verification can be done to find that for every @xmath118 , we have @xmath119 even if @xmath120 , and @xmath121 in the same way , @xmath77 extends continuously up to each side of @xmath1 , which concludes the lemma in the segment case .
now , we come back to our problem , with any curve @xmath1 . after applying a homothetic transformation , a rotation and a translation , we can suppose that the endpoints of the curve are @xmath122 and @xmath123 .
next , we consider the curve @xmath124 .
we now show that @xmath125 is a @xmath126 jordan curve .
we consider first the case where @xmath1 does not intersect the segment @xmath127 .
then @xmath128 and @xmath129 are @xmath130 open jordan arcs , with the endpoints on @xmath95 and @xmath49 ( see picture 1 ) .
so @xmath131 and we can observe that @xmath132 is a jordan curve .
@xmath1 @xmath133 * picture 1 * : @xmath1 does not intersect @xmath71 $ ] @xmath134 we wrote _ open _ jordan curve because the problem with @xmath95 and @xmath49 is that @xmath135 is not defined
. however , if we use the arclength coordinates @xmath136 which are well - defined and bounded because @xmath137 is bounded in @xmath138 , then we have @xmath139 .
so to prove the derivative continuity , we should show that @xmath140 and @xmath141 exist and are opposite .
for that , we will use the following lemma : [ 2.3 ] if there exists a neighborhood of @xmath89 where @xmath142 does not intersect the segment @xmath127 , then @xmath143 has a limit as @xmath144 . first , since @xmath145 in a neighborhood of @xmath95 , we compute @xmath146 the second fraction tends to 1 as @xmath144 .
we have @xmath147 , so we can write @xmath148 for @xmath149 . if @xmath150 then for @xmath63 small enough @xmath151 and @xmath152 for @xmath153 , as @xmath154 , then we have @xmath155 and the curve is tangent to the segment @xmath71 $ ] .
we have here two cases : * if @xmath1 is over the segment on the neighborhood , then @xmath156 and @xmath157 * if @xmath1 is below the segment on the neighborhood , then @xmath158 and @xmath159 @xmath160 let us continue the proof of proposition [ 2.2 ] .
lemma [ 2.3 ] allows us to observe that @xmath132 is @xmath130 in @xmath95 , because @xmath161 because @xmath99 . to prove that @xmath162 is lipschitz
, we will show that @xmath163 and @xmath164 are @xmath130 with the arclength parametrization denoted by @xmath165 and defined in ( [ curviligne ] ) ( @xmath63 denotes the variable of the original parametrization ) .
let @xmath166 , where the primes denote derivatives with respect to @xmath63 , then we need to prove that @xmath167 has a limit when @xmath168 .
we have @xmath169 we compute @xmath170 with @xmath171 we do some taylor expansions in a neighborhood of zero : @xmath172 the last expansion holds in any case , except when @xmath1 is tangent to the segment ( @xmath173 ) and over the segment on a neighborhood of @xmath95 . in this last case , we should replace @xmath174 by @xmath175 instead of @xmath176 .
then @xmath177 and @xmath178 now , we can evaluate @xmath179 : @xmath180 we can easily see that @xmath181 , so @xmath182 and @xmath183 , which means that @xmath184 has a limit as @xmath168 .
this argument holds for @xmath185 , doing the calculations with @xmath186 .
so @xmath187 and @xmath188 are @xmath130 on @xmath44 $ ] and we see that @xmath189 is lipschitz because @xmath190 . finally , if @xmath1 intersects @xmath71 $ ] at one point @xmath191 , then @xmath192 is the union of two jordan curves with a jump : @xmath193 ( see picture 2 ) . in this case , @xmath194 is also the union of two jordan arcs which extend @xmath192 , indeed @xmath195 @xmath1 @xmath192 * picture 2 * : @xmath1 intersects @xmath71 $ ] at one point @xmath194 @xmath196 to show the continuity of @xmath162 on @xmath197 , we consider for example that @xmath198 and that @xmath199 and @xmath200 .
we can compute @xmath201 to check that @xmath202 , which allows us to conclude that @xmath203 we leave to the reader the other cases .
let us do just another case : @xmath204 then @xmath205 and as @xmath206 we have the continuity of @xmath189 . as @xmath207
is bounded , @xmath189 is lipschitz , so the curve @xmath125 is @xmath126 and closed .
we have just studied the case of one or zero intersection of @xmath1 with the segment @xmath127 but this argument works in the general case because we have a finite number of intersection .
for example , if @xmath208 $ ] in a neighborhood of @xmath95 , then @xmath209 , so @xmath210 is obviously @xmath126 in this neighborhood .
we denote by @xmath211 the unbounded connected component of @xmath212 . we claim that we can construct @xmath213 , a biholomorphism between @xmath68 and @xmath211 , such that @xmath214 . indeed ,
if we introduce @xmath215 and @xmath216 ( see picture 1 ) , we observe that @xmath217 , because @xmath218 and @xmath219 . as @xmath220
, @xmath88 is bijective on @xmath221 and @xmath222 then @xmath88 is bijective on @xmath223 and @xmath224 . therefore , we have an function @xmath213 mapping the exterior of the jordan arc to the exterior of the inner domain of a @xmath126 jordan curve , such that @xmath225 .
next , we just have to use the riemann mapping theorem and we find a conformal mapping @xmath226 between @xmath223 and @xmath227 , such that @xmath228 maps @xmath68 to @xmath227 and @xmath229 .
[ f ] to finish the proof , we use the kellogg - warschawski theorem ( see theorem 3.6 of @xcite , which can be applied for the exterior problems ) , to observe that @xmath226 and @xmath230 have a continuous extension up to the boundary . therefore , adding the fact that @xmath231 and @xmath232 are bounded at infinity ( see remark [ 2.5 ] ) , we find the same properties as in the segment case , in particular that @xmath77 blows up at the endpoints of the curve like the inverse of the square root of the distance ( see ( [ t ] ) ) . [ 2.4 ] if @xmath1 intersects the segment @xmath71 $ ] infinitely many times , the curve @xmath132 may not be even @xmath130 . for example a curve which starts as @xmath233 $ ] has two sequences @xmath234 and @xmath235 such @xmath236 tends to @xmath175 and @xmath176 . [
2.5]if we have a biholomorphism @xmath237 between the exterior of an open connected and simply connected domain @xmath179 and @xmath227 , such that @xmath238 , then there exists a nonzero real number @xmath239 and a holomorphic function @xmath240 such that : @xmath241 with @xmath242 this property can be applied for the @xmath226 above to see that @xmath231 and @xmath232 are bounded .
indeed , after a translation we can suppose that @xmath243 , and we consider @xmath244 .
the function @xmath245 is holomorphic in a neighborhood of @xmath89 and can be written as @xmath246 .
we have @xmath247 so @xmath248 .
now we want to prove that @xmath249 thanks to the univalence .
indeed , if @xmath250 , we consider the first non - zero @xmath251 , and we observe that there exists @xmath252 such that @xmath253 in @xmath78 .
next , we denote by @xmath254 . on @xmath255 , @xmath256 .
then we can apply the rouch theorem to conclude that @xmath245 and @xmath257 have the same number of zeros in @xmath78 , which is a contradiction with the fact that @xmath245 is a biholomorphism and @xmath257 not .
therefore @xmath249 and @xmath258 , which ends the proof . multiplying @xmath237 by @xmath259
, we can assume that @xmath260 is real number .
@xmath160 let @xmath261 be a bounded , open , connected , simply connected subset of the plane , the boundary of which , denoted by @xmath262 , is a @xmath263 jordan curve .
we will denote by @xmath264 the unbounded connected component of @xmath265 , so that @xmath266 .
we denote by @xmath267 the green s function , whose the value is : @xmath268 writing @xmath269 .
the green s function is the unique function which verifies : @xmath270 then the kernel of the biot - savart law is @xmath271 . with @xmath272 , the explicit formula of @xmath273
is given by @xmath274 we require information on far - field behavior of @xmath273 .
we will use several times the following general relation : @xmath275 which can be easily checked by squaring both sides .
we now find the following inequality : @xmath276 for @xmath277 , we introduce the notation @xmath278=k_{\pi_0}[f](x)\equiv\int_{\pi_0 } k_{\pi_0}(x , y)f(y)dy.\ ] ] it is easy to see , for large @xmath279 , that @xmath280|(x)\leq \dfrac{c_1}{|x|^2}$}$ ] where @xmath281 depends on the size of the support of @xmath282 .
we used here the explicit formulas for the biholomorphism @xmath283 ( remark [ 2.5 ] ) .
[ 2.6 ] the vector field @xmath284 $ ] is a solution of the elliptic system : @xmath285 the proof of this lemma is straightforward
. we will denote by @xmath286 the unit normal exterior to @xmath264 at @xmath262 . in what follows
all contour integrals are taken in the counter - clockwise sense , so that @xmath287 . [ 2.7 ] there exists a unique classical solution @xmath288 of the problem : @xmath289 moreover , @xmath290 as @xmath291 to prove this , one can check that @xmath292 is the unique solution .
the details can be found in @xcite .
let us formulate precisely here the small obstacle limit .
let @xmath293 be the velocity of an incompressible , ideal flow in @xmath294 .
we assume that @xmath24 is tangent to @xmath262 and @xmath295 as @xmath296 .
the evolution of such a flow is governed by the euler equations : @xmath297 where @xmath298 is the pressure .
an important quantity for the study of this problem is the vorticity : @xmath299 the velocity and the vorticity are coupled by the elliptic system : @xmath300 lemma [ 2.6 ] and proposition [ 2.7 ] assure us that the general solution of this system is given by @xmath301(x)+{\alpha}h_{\pi_0}(x)$ ] for a function @xmath302 .
however , using the fact that the circulation @xmath6 of @xmath24 around @xmath1 is conserved , we prove that @xmath303 does not depend on the time , and @xmath304 ( see @xcite ) . therefore ,
if we give the circulation , then we have the uniqueness of the solution of the previous system . finally , we can now write the vorticity formulation of this problem as : @xmath305+{\alpha}h_{\pi_0}&\text { in } { \pi_0}\times[0,\infty ) \\ & { \omega}(x,0)={{\rm curl}\,}u_0(x ) & \text { in } { \pi_0 } \end{aligned}\right.\ ] ] we will formulate in this subsection a family of problems , parametrized by the size of the obstacle . therefore ,
we fix @xmath4 such that its support is compact and does not intersect @xmath1 .
we will consider a family of domain @xmath0 , containing @xmath1 , with @xmath5 small enough , such that the support of @xmath4 does not intersect @xmath0 .
if we denote by @xmath306 the biholomorphism between @xmath307 and @xmath227 , then we suppose the following properties : [ 3.1 ] the biholomorphism family @xmath308 verifies * @xmath309 for any @xmath82 , * @xmath310 is bounded in @xmath311 independently of @xmath5 , * @xmath312 for any @xmath82 , * there exist @xmath82 and @xmath313 such that @xmath314 on @xmath315 .
[ 3.2 ] we can observe that point ( iii ) implies that for any @xmath316 , @xmath317 is bounded in @xmath318 independently of @xmath5 , for @xmath319 . just before going on
, we give here one example of an obstacle family .
we consider @xmath320 . in this case , @xmath321 verifies the previous assumption .
if @xmath1 is a segment , then @xmath0 is the interior of an ellipse around the segment .
the problem of this example is that the shape of the obstacle is the same of @xmath1 .
we naturally denote by @xmath322 and @xmath323 .
we denote also by @xmath324,@xmath325 and @xmath326 the previous functions corresponding at @xmath311 .
consider also the following problem : @xmath327+{\alpha}h^{\varepsilon}&\text { in } \pi_{\varepsilon}\times[0,\infty ) \\ & { \omega}^{\varepsilon}(x,0)={\omega}_0(x ) & \text { in } \pi_{\varepsilon}\end{aligned}\right.\ ] ] it follows from the work of kikuchi @xcite that , for any @xmath328 , if @xmath4 is sufficiently smooth then this system has a unique solution .
we now write the explicit formulas for @xmath329 and @xmath330 : @xmath331 and @xmath332 we introduce in the same way , @xmath333 and @xmath237 , replacing @xmath306 by @xmath76 .
the regularity of @xmath76 implies that @xmath330 is bounded in @xmath334 , which is really better than the punctual limit for the obstacle ( see @xcite ) where @xmath330 is just @xmath138 . in our case ,
the limit is easier to see when @xmath335 , and this extra regularity will allow us the passing to the limit .
these estimates are important to conclude on the asymptotic behavior of the sequences @xmath336 and @xmath337 .
the transport nature of ( [ e_e ] ) gives us the classical estimates for the vorticity : @xmath338 and for @xmath339 $ ] . in this article
, we suppose that @xmath4 is @xmath340 and compactly supported .
moreover we choose @xmath5 small enough , so that the support of @xmath4 does not intersect @xmath311 .
we begin by recalling a result found in @xcite .
[ 4.1 ] let @xmath341 , @xmath342 and @xmath343 be a function in @xmath344 . then @xmath345 the goal of this subsection is to find a velocity estimate thanks to the explicit formula of @xmath8 in function of @xmath346 and @xmath6 ( subsection [ evanescent ] ) : @xmath347 with @xmath348 and @xmath349 we begin by estimating @xmath350 and @xmath351 . [ i_est ] let @xmath341 and @xmath352 be a function in @xmath353 .
we introduce @xmath354 @xmath355 there exists a constant @xmath313 depending only on the shape of @xmath1 , such that @xmath356 [ remark4.3 ] it will be clear from the proof that similar estimates hold true with @xmath306 replaced by @xmath76 . we start with the @xmath357 estimate .
let @xmath358 and @xmath359 .
making also the change of variables @xmath360 , we find @xmath361 , with @xmath362 the characteristic function of the set @xmath363 . changing variables back , we get @xmath364 the second point of assumption [ 3.1 ] allows us to write @xmath365 so we apply the previous lemma for @xmath282 and we finally find latexmath:[\[\label{1_a } this concludes the estimate for @xmath357 .
let us estimate @xmath367 : @xmath368 we use , as before , the notations @xmath54 , @xmath105 and the change of variables @xmath369 @xmath370 , to obtain : @xmath371 first we estimate @xmath372 .
as @xmath359 , one has that @xmath373 , and if @xmath374 then @xmath375 .
hence @xmath376 finally , we estimate @xmath377 .
let @xmath378 .
we have @xmath379 as above , we deduce by changing variables back that @xmath380 it is also trivial to see that @xmath381 by lemma [ 4.1 ] @xmath382 since @xmath383 , one can easily see from ( [ h ] ) that @xmath384 .
moreover , applying the previous lemma with @xmath385 and @xmath386 , we get that @xmath387 thanks to the explicit formula ( [ u_e ] ) , we can deduce directly the following theorem : [ 4.2 ] @xmath8 is bounded in @xmath388 independently of @xmath5 .
more precisely , there exists a constant @xmath313 depending only on the shape of @xmath1 and the initial conditions @xmath389 , @xmath390 , such that @xmath391 \text{\ and for any subset } s \text{\ of } \pi_{\varepsilon}.\ ] ] the difference with @xcite is that we have an estimate @xmath117 instead of @xmath340 , but in our case , this estimate concerns all the velocity @xmath8 .
it is one of the reason of the use of a different method to the velocity convergence .
if we want to compare the different velocity and vorticity , the issue is that @xmath8 and @xmath346 are defined on an @xmath5-dependent domain .
for this reason we extend the velocity and vorticity on @xmath11 by multiplying by an @xmath5-dependent cutoff function for a neighborhood of @xmath0 .
let @xmath392 be a non - decreasing function such that @xmath393 , @xmath394 if @xmath395 and @xmath396 if @xmath397 .
then we introduce @xmath398 clearly @xmath399 is @xmath400 vanishing in a neighborhood of @xmath401 .
we require some properties of @xmath402 which we collect in the following lemma . [ 4.4 ] the function @xmath399 defined above has the following properties : * @xmath403 in @xmath311 , * there exists a constant @xmath313 such that the lebesgue measure of the support of @xmath404 is bounded by @xmath405 .
first , we remark that @xmath406 and @xmath407 what gives us the first point .
finally , the support of @xmath404 is contained in the subset @xmath408 .
the lebesgue measure can be estimated as follows : @xmath409 for @xmath5 small enough .
we introduced the cutoff function @xmath399 in order to extend the velocity and the vorticity to @xmath410 .
one needs to make sure that the limit velocity and vorticity are not affected by the way the extension is constructed .
we observe that our method of extension does not produce an error in the limit velocity and vorticity . indeed , we denote by @xmath411 , respectively @xmath412 , the extension of @xmath8 , respectively @xmath346 , by 0 inside the obstacle and we prove that @xmath413 and @xmath414 in @xmath415 . indeed
, using theorem [ 4.2 ] and remark [ 3.2 ] , point ( b ) of the previous lemma allows us to state that @xmath416 and @xmath417 for all @xmath418 . in the case where the limit is a point ( @xcite ) , the lebesgue measure of the support of @xmath402
is bounded by @xmath419 , which implies that the norm @xmath420 of this gradient tends to @xmath89 .
moreover , the authors use a part of velocity @xmath421 bounded independently of @xmath5 , so they can compute the limit of @xmath422 and @xmath423 which is necessary for the calculation of the curl and div .
finally they conclude thanks to the div - curl lemma .
in our case , for @xmath424 we have @xmath425 , and we can not compute the limit of @xmath426 . for this reason we can not use a similar proof as in @xcite .
however , the following lemma gives us a piece of information about the limit behavior .
[ 4.5 ] @xmath427 strongly in @xmath428 and uniformly in time , when @xmath429 . using the explicit formulas ( [ u_e ] ) , ( [ i_1 ] ) ,
( [ i_2 ] ) and ( [ fi ] ) , we write @xmath430 but @xmath306 is holomorphic , so @xmath317 is of the form @xmath431 and we can check that @xmath432 , so @xmath433 we compute the @xmath420 norm , next we change variables twice @xmath434 and @xmath435 , to have @xmath436 where @xmath437 . thanks to lemma [ 4.4 ] , we know that @xmath438 .
so it is sufficient to prove that @xmath439 as @xmath429 , uniformly in time .
let @xmath440 we compute @xmath441 we now use that @xmath373 , to write @xmath442 moreover , @xmath443 allows to have @xmath444 we can now estimate @xmath179 by : @xmath445 with @xmath446 , to be chosen later .
we remark also that @xmath447 and the cauchy - schwarz inequality gives @xmath448 we now use the fact that @xmath449 , to estimate ( [ tronc ] ) : @xmath450 in the same way we passed from ( [ i_1d ] ) to ( [ 1_a ] ) , we obtain for @xmath451 : @xmath452 moreover , as we passed from ( [ i_2d ] ) to ( [ i_21 ] ) and ( [ i_22 ] ) , we obtain for @xmath453 : @xmath454 we choose @xmath455 , @xmath456 , and using the hlder inequality we finish the proof .
for example if we fix @xmath457 , @xmath458 and @xmath459 , we obtain @xmath460 which tends to zero when @xmath5 tends to zero .
if the proof is a little bit technical , the idea is natural . on the boundary ,
the velocity @xmath8 is tangent to @xmath461 , whereas @xmath402 is normal . to see that
, we can check that @xmath462 when @xmath463 ( which means that @xmath464 ) . before
going to the last section , we derive directly from the pde a temporal estimate for the vorticity .
if we fix @xmath465 , we remark that there exists @xmath466 such that the support of @xmath467 is contained in @xmath468 for all @xmath469 . to see that , let @xmath470 be such that @xmath471 contains the support of @xmath4 .
equation @xmath472 means that @xmath346 is transported by the velocity field @xmath8 and the trajectory of a particle moving with the flow verifies @xmath473 moreover , theorem [ 4.2 ] states @xmath474 and the last point of assumption [ 3.1 ] states that there exist @xmath82 and @xmath475 such that @xmath317 is bounded by @xmath476 outside @xmath78 . if a material particle reaches the region @xmath477 , its velocity is uniformly bounded by @xmath478 , and we obtain the following inequality : @xmath479 that holds true in such a region .
applying gronwall lemma , we observe that the trajectory of a material particle is bounded independently of @xmath5 ( up to the fixed time @xmath480 ) .
[ 4.6 ] there exists a constant @xmath39 , which does not depend on @xmath481 $ ] and @xmath5 such that @xmath482 we write the equation verified by @xmath483 : @xmath484 which is bounded in @xmath485 for the following reason . note that @xmath399 and @xmath346 are uniformly bounded , @xmath8 is bounded in @xmath486 thanks to theorem [ 4.2 ] , remark [ 3.2 ] and the previous remark .
moreover @xmath487 in @xmath420 according to the previous lemma .
we finally conclude , bearing in mind that @xmath420 and @xmath488 are embedded into @xmath485 .
fix @xmath465 .
we will need the following lemmas to the passing to the strong limit @xmath489\times { { \bbb r}}^2)$ ] of the sequence @xmath490 . as in the previous subsection ,
let @xmath466 be such that the support of @xmath467 is contained in @xmath468 for all @xmath469 and @xmath491 .
[ 5.1 ] for all fixed @xmath492 , there exists @xmath493 such that @xmath494 for all @xmath495 .
the two following functions @xmath496 @xmath497 are bounded in @xmath498 independently of @xmath495 ( but not necessarily independent of @xmath499 ) . moreover , @xmath500 and @xmath501 are bounded in @xmath502 .
bearing in mind the properties of @xmath76 and that @xmath335 uniformly in @xmath503 , we know that @xmath504 , for some @xmath505 independent of @xmath5 . to bound @xmath506 , we change variables @xmath434 and denote by @xmath435 : @xmath507 for the second function , we begin in the same way , next we change again variables with @xmath508 : @xmath509 with @xmath510 not depending on @xmath5 . replacing @xmath76 by @xmath306 , one can obtain the bounds for @xmath511 and @xmath512 .
we now consider the limit of @xmath483 . as @xmath513 and @xmath514
, we can extract a subsequence such that @xmath21 weak-@xmath22 in @xmath515 with @xmath516 and @xmath517 .
moreover , we want that @xmath518 weak @xmath519 for all @xmath63 . as @xmath520 , and
thanks to alaoglu s theorem , for all @xmath63 we can extract a subsequence which verifies @xmath518 weak @xmath519 .
the problem is that the subsequence depends on the time .
let us look for a common subsequence for all @xmath63 .
we observe that @xmath521 is defined for all @xmath522 . indeed , @xmath517 , so @xmath523 .
moreover , since @xmath516 we also have that @xmath524 for all @xmath522 and @xmath525 for all @xmath526 .
[ 5.2 ] there exists a subsequence of @xmath483 ( again denoted by @xmath483 ) such that @xmath518 weak @xmath527 for all @xmath63 . we can choose a common subsequence for all rational times , by the diagonal extraction , because @xmath528 is countable .
we now prove that this subsequence converges for any @xmath63 .
let @xmath529 , and @xmath530 then @xmath531 for all @xmath532 .
moreover @xmath533 which allows us to state that the family @xmath534 is equicontinuous , using the temporal estimate ( lemma [ 4.6 ] ) and that @xmath535 . therefore , we have an equicontinuous family which tends to @xmath89 on a dense subset , so it tends to @xmath89 for all @xmath63 .
to finish , let @xmath536 .
the set @xmath537 being dense into @xmath538 , there exists a sequence @xmath539 which converges to @xmath42 in @xmath538 .
we introduce @xmath540 in the same way as @xmath541 , replacing @xmath42 by @xmath542 .
let @xmath63 be fixed , we have by the first part @xmath543 moreover @xmath544 tends to @xmath89 uniformly in @xmath5 , which according to ( [ fn ] ) allows to conclude that @xmath545 .
this completes the proof . [
5.3 ] the two following functions : @xmath546 and @xmath547 tend to @xmath89 as @xmath429 in @xmath548,l^6_{{\operatorname{{loc}}}}({{\bbb r}}^2))$ ] .
let @xmath333 be a compact set of @xmath11 .
firstly , we fix @xmath481 $ ] and we prove that the norm @xmath549 of @xmath541 tends to @xmath89 . for all @xmath550 , @xmath551 with @xmath511 given in lemma [ 5.1 ] .
furthermore , lemma [ 5.1 ] states that @xmath511 is bounded in @xmath538 and as @xmath21 weak @xmath519 , we obtain that for fixed @xmath499 , @xmath552 moreover , we can apply lemma [ i_est ] ( estimate of @xmath357 ) to @xmath541 , with @xmath553 and we obtain a bound for @xmath541 independently of @xmath499 , @xmath63 and @xmath5 . then , @xmath554 almost everywhere as @xmath429 , and @xmath555 is uniformly bounded .
we can apply the dominated convergence theorem to conclude that for fixed @xmath63 and @xmath333 a bounded set @xmath556 we now let @xmath63 vary and we apply again the dominated convergence theorem to obtain the result on @xmath541 . using the estimate of @xmath367 in lemma [ i_est ]
, we proceed in the same manner to prove the result for @xmath557 . moreover , we need two last lemmas which are a consequence of the convergence of @xmath306 to @xmath76 ( assumption [ 3.1 ] ) . [ 5.4 ] for all fixed @xmath492 , @xmath558 in @xmath559 ( where the norm is taken with respect to @xmath560 ) .
let @xmath499 be fixed . using the relation ( [ frac ] )
we have @xmath561 by assumption [ 3.1 ] , we know that @xmath562 pointwise as @xmath429 . moreover , if @xmath563 then @xmath564 , and as @xmath565 we can write latexmath:[\ ] ] which proves ( [ equiv ] ) .
now , we write a formulation for the velocity @xmath24 , by replacing @xmath669 by @xmath736 to obtain in @xmath11 @xmath737 however , since @xmath738 , we can remark that @xmath739=0 $ ] in @xmath603 .
we can obtain directly an equation for @xmath24 on @xmath603 by passing to the limit @xmath429 .
we multiply the velocity equation ( [ e_e ] ) by some divergence - free test vector field @xmath740 and assume that @xmath5 is small enough such that the support of @xmath42 is contained in @xmath311 and do not intersected the support of @xmath402 .
after integration , @xmath741 which easily pass to the limit since @xmath742 strongly in @xmath334 by theorem [ 5.6 ] . indeed , we can prove easily that @xmath743 in @xmath744 thanks to the proof of theorem [ 5.6 ] . therefore , the above relation
holds true with @xmath24 instead of @xmath8 and this is the formulation in the sense of distributions of the euler equation in @xmath603 : @xmath745 where @xmath226 is the formula from theorem [ 5.6 ] expressing explicitly the velocity in terms of vorticity and circulation .
i want to thank some researchers for the help offered with the subject of complex analysis , which allows the generalization to any curve @xmath1 . in particular etienne ghys and alexei glutsyuk who searched with me a proof using quasi - conformal mappings for the straightening up of the curve by a holomorphism , @xmath263 up to the boundary .
i would like to give many thanks to xavier buff for suggesting the use of @xmath746 .
i also want to thank m.c .
lopes filho and h.j .
nussenzveig lopes for several interesting and helpful discussions
. 99 iftimie d. , lopes filho m.c . and nussenzveig lopes h.j . , _ two dimensional incompressible ideal flow around a small obstacle _ , comm .
partial diff .
( 2003 ) , no . 1@xmath7472 , 349 - 379 . beel s. and krantz s.g . ,
_ smoothness to the boundary of conformal maps _ , rocky mt j math 1987 ; 17(1):23 - 40 .
pommerenke c. , _ univalent functions _ , vandenhoeck @xmath747 ruprecht , 1975 .
pommerenke c. , _ boundary behaviour of conformal maps _
, berlinnew york : springer - verlag , 1992 .
iftimie d. , _ evolution de tourbillon support compact _ , actes du colloque de saint - jean - de - monts , 1999 .
kikuchi k. , _ exterior problem for the two - dimensional euler equation _ , j fac sci univ tokyo sect 1a math 1983 ; 30(1):63 - 92 .
lopes filho m.c .
, _ vortex dynamics in a two dimensional domain with holes and the small obstacle limit _ , to appear , siam journal on mathematical analysis , 2006 .
iftimie d. , lopes filho m.c . and nussenzveig lopes h.j . ,
_ two dimensional incompressible viscous flow around a small obstacle _ , math .
ann . , * 336 * ( 2006 ) , 449 - 489 . | in this work we study the asymptotic behavior of solutions of the incompressible two - dimensional euler equations in the exterior of a single smooth obstacle when the obstacle becomes very thin tending to a curve .
we extend results by iftimie , lopes filho and nussenzveig lopes , obtained in the context of an obstacle tending to a point , see [ comm .
pde , * 28 * ( 2003 ) , 349 - 379 ] . |
in this paper , we mainly consider the linear system @xmath7 where the matrix @xmath4 is a complex symmetric matrix .
this kind of linear systems ( [ eq:1.1 ] ) arise from different physical applications , for example , in modeling electromagnetic waves under the assumption of time - harmonic variation in the electromagnetic fields , helmholtz s equations with a complex shift ( see , e.g. , @xcite ) etc .
moreover , the complex - valued linear system ( [ eq:1.1 ] ) can be directly generated in the field of lattice quantum chromo dynamics ( qcd ) , where a model of the interactions of fermions ( or quarks ) on a lattice is given in terms of a complex - valued gauge field that directly leads to the linear system ( [ eq:1.1 ] ) ( see , @xcite ) .
in addition , this complex symmetric linear system also arises in centered difference discretization of the @xmath8-pad approximations in the time integration of parabolic partial differential equations ( @xcite ) and in direct frequency domain analysis of an @xmath9-degree - of - freedom ( @xmath9-dof ) linear system ( @xcite ) .
there are some examples of scientific applications in @xcite .
therefore , researches on numerical solutions of the linear system ( [ eq:1.1 ] ) are greatly needed .
next , for convenience , let @xmath10 denote the set of @xmath11 complex matrices and @xmath4 be a nonsingular matrix in @xmath10 .
@xmath12 and @xmath13 are the largest and smallest eigenvalues of @xmath14 , respectively .
recently , a _ complex symmetric positive definite _ ( cspd ) matrix arising from the linear system ( [ eq:1.1 ] ) in gaussian elimination without pivoting was firstly studied by higham in @xcite , which is called by _ * higham matrices * _ in @xcite .
subsequently , the paper @xcite gave a broader class of complex matrices_*generalized higham matrices*_(sometimes they are also called accretive - dissipative matrices ( see @xcite ) ) , i.e. , for any @xmath15 , if , its hermitian decomposition ( see , @xcite ) @xmath16 satisfies that @xmath17 where @xmath18 is the conjugate transpose of @xmath1 , then the matrix @xmath4 is said to be a generalized higham matrix and denoted by @xmath19 .
here , the sign @xmath20 is usually called the loewner partial order of hermitian matrices ; i.e. , we write @xmath21 if the matrix @xmath22 is positive semidefinite , similarly , @xmath23 means that @xmath22 is positive definite .
in addition , a related class of matrices defined by @xmath24 will be denoted by @xmath25 as in @xcite .
this paper is a continuation of @xcite , and they both originate from the higham s paper @xcite . as it is well known , for the complex symmetric linear system ( [ eq:1.1 ] ) , the growth factor @xmath26 in gaussian elimination for @xmath4 is defined by @xmath27 where @xmath28 , @xmath29 and @xmath30 is the matrix obtained through the application of the first @xmath31 steps of gaussian elimination to a ( see , e.g.,@xcite ) . in particular
, @xmath32 is the upper triangular matrix resulting from the @xmath33 factorization of @xmath4 . obviously , for the matrix @xmath4 in the linear system ( [ eq:1.1 ] ) , if one is able to prove a satisfactory priori bound for @xmath26 , then it is safe not to pivot in computing the @xmath33 factorization of the matrix @xmath4 ( or to choose diagonal pivots based on other considerations such as sparsity preservation ) ( see , @xcite ) .
for any higham matrix @xmath4 , the growth factor in gaussian elimination @xmath34 was firstly conjectured by higham ( see , @xcite , p.210 ) .
an incorrect proof was given in @xcite , but ikramov et al .
@xcite subsequently showed that @xmath35 for any higham matrix @xmath4 .
in addition , if the higham matrix is extended by allowing @xmath1 and @xmath2 to be arbitrary hermitian positive definite matrices , then @xmath36 moreover , ikaramov et al noted that the above bound ( [ eq:1.6 ] ) remains true when @xmath1 or @xmath2 or both are negative ( rather than positive ) definite ( see , @xcite ) . for a very restricted subset of higham matrices ,
i.e. , when @xmath37 and @xmath2 is real , symmetric and positive definite , the authors in @xcite proved the better bound @xmath38 in addition , a. george and k. d. ikramov ( @xcite ) assumed that @xmath1 and @xmath2 with being positive definite , satisfy the inequality @xmath39 and they established a bound for the growth factor @xmath40 that has the limit @xmath41 as @xmath42 .
recently , lin @xcite proved that if @xmath4 is a generalized higham matrices , then the growth factor for such @xmath4 in gaussian elimination is less than 4 .
specially when @xmath4 is a higham matrix , then the growth factor is less than @xmath43 .
however , as authors in @xcite pointed out , in no case have they observed the growth bigger than higham s guess of @xmath6 from extensive numerical experiments with higham matrices .
therefore , they believed that the bound ( [ eq:1.4 ] ) is correct and took the proof of this bound for an open problem ( see also problem 10.12 in @xcite ) . in this work , we continue studying this open problem and then give a new result @xmath44 for the generalized higham matrix @xmath4 , where @xmath45 is the maximum of the condition numbers of @xmath1 and @xmath2 .
this directly leads to the higham s result ( [ eq:1.4 ] ) for any higham matrix @xmath4 , which proves the open problem . here , for a nonsingular matrix @xmath4
, its condition number is denoted by @xmath46 , i.e. , the ratio of the largest and smallest singular value of @xmath4 .
this paper is organized as follows . in section 2 , we show some new bounds on the growth factor , based on the condition number . in section 3 ,
some figures and numerical examples are given to illustrate our results .
in this section , let @xmath15 be partitioned as @xmath47 where @xmath4 is an @xmath11 nonsingular matrix .
if @xmath48 is invertible , then the schur complement of @xmath48 in @xmath4 is denoted by @xmath49 ( see , @xcite ) .
( @xcite).[lem:2.1 ] let @xmath4 be a cspd matrix , then @xmath4 is nonsingular , and any principal submatrix of @xmath4 and any schur complement in @xmath4 are also cspd matrices .
obviously , lemma [ lem:2.1 ] shows that , being a cspd matrix is an hereditary property of active submatrices in gaussian elimination .
( @xcite)[lem:2.2 ] .
the largest element of a cspd matrix @xmath4 lies on its main diagonal .
thus , for any cspd matrix @xmath4 , the definition ( [ eq:1.2 ] ) can be replaced by @xmath50 which greatly simplifies the analysis on bounding the growth factor for a cspd matrix @xmath4 .
( @xcite).[lem:31 ] if @xmath1 is a nonzero @xmath11 positive definite matrix having eigenvalues @xmath51 , then for all orthogonal vectors @xmath52 , and @xmath53 denotes the conjugate transpose of @xmath54 , the following equality holds , @xmath55 ( @xcite).[lem:3.1 ] let @xmath1 be as in lemma [ lem:31 ] , then for any @xmath56 matrix @xmath57 satisfying @xmath58 , where @xmath59 means the conjugate transpose of the matrix @xmath57 , we have that @xmath60 by lemma [ lem:3.1 ] , it is easy to obtain the following lemma .
( @xcite).[lem:3.1 ] let @xmath61 be an @xmath11 hermitian positive definite matrix , where @xmath62 is any @xmath63 principal submatrix of @xmath1 ( @xmath64 ) , then @xmath65 where @xmath66 is the condition number of @xmath67 .
( @xcite).[th:3.2 ] let @xmath1 be a hermitian positive definite matrix , then @xmath68 , ( @xmath69 ) , where @xmath70 is the @xmath71 principal submatrix of @xmath1 .
( @xcite).[cor:3.2 ] let @xmath1 be a hermitian positive definite matrix , and partitioned as in lemma [ lem:3.1 ] , then @xmath72 .
( @xcite ) .
let @xmath0 , @xmath73 , @xmath74 , be partitioned as in ( [ eq:1.3 ] ) , if @xmath75 , @xmath76 are invertible , then @xmath77 where @xmath78 .
( @xcite)[cor:3.3 ] .
let @xmath0 , @xmath73 , @xmath74 be a generalized higham matrix and be partitioned as in ( [ eq:1.3 ] ) , if @xmath79 is its hermitian decomposition , then @xmath80 , @xmath81 next , we give our main result .
[ th:3.3 ] let @xmath4 be a generalized higham matrix , then @xmath82 where @xmath83 is the maximum of the condition numbers of @xmath1 and @xmath2 .
* proof*. fix the number @xmath84 and @xmath85 , where @xmath86 .
denote @xmath87 by the leading principal submatrix of order @xmath31 in @xmath4 .
we consider the @xmath88 matrix @xmath89 where @xmath90 @xmath91 note that @xmath92 , @xmath93 and @xmath94 are principal order @xmath95 submatries of @xmath4 , @xmath1 and @xmath2 , respectively .
it is easy to see that @xmath96 can be obtained by performing block gaussian elimination in @xmath92 ; i.e. , @xmath97 setting @xmath98 . since both @xmath93 and @xmath94 are hermitian positive definite , according to the result of the lemma [ lem:3.1 ] , we have @xmath99 next , by the corollary [ cor:3.3 ] and @xmath100 is decreasing in @xmath101 , we get @xmath102 where @xmath103 .
since @xmath104 , @xmath105 and @xmath106 , by the corollary [ cor:3.3 ] , we have @xmath107 noting that @xmath108 is increasing in @xmath101 , one has @xmath109 and @xmath110 equivalently , @xmath111 thus , we can obtain @xmath112 according to the above inequalities ( [ eq:3.6 ] ) and ( [ eq:3.7 ] ) , the following inequalities is obvious that @xmath113 note that @xmath114 is decreasing in @xmath115 $ ] , and @xmath116 is increasing in @xmath115 $ ] ( see , fig .
1 ) , by corollary [ cor:3.2 ] , we have that @xmath117 the proof is completed .
@xmath118 [ fig1 ] with @xmath83 decreasing , right : the variation curve of @xmath119 with @xmath83 increasing.,title="fig:",width=235,height=259 ] with @xmath83 decreasing , right : the variation curve of @xmath119 with @xmath83 increasing.,title="fig:",width=235,height=259 ] [ cor:3:4 ]
if @xmath4 is an @xmath11 generalized higham matrix , then @xmath120 * proof*. since @xmath121 , we calculate simultaneously the limit of both sides of the inequality , we have @xmath122 therefore , the result ( [ eq:38 ] ) holds .
* remark 2.1*. obviously , the above result ( [ eq:38 ] ) holds also for any higham matrix , and hence the higham s conjecture ( see ( [ eq:1.4 ] ) ) is correct , which solves this open problem .
in addition , since @xmath123 , then @xmath124 , i.e. , @xmath125 .
so , generally speaking , @xmath126 , see the following numerical experiment .
in this section , a numerical example will be described .
the goal of the experiment is to examine the effectiveness of our result .
we consider the complex symmetric system of linear equation ( [ eq:1.1 ] ) arises in the centered difference discretizations of the @xmath8-pad approximations in the time integration of parabolic partial differential equations , further details refer to @xcite . for convenience
, the complex coefficient symmetric matrix ( see , @xcite ) may be written as @xmath127 where @xmath128 is the identity matrix , @xmath129 is the time step - size and @xmath130 is the five - point centered difference matrix approximating the negative laplacian operator @xmath131 with homogeneous dirichlet boundary conditions , on a uniform mesh in the unit square @xmath132\times[0,1]$ ] with the mesh - size @xmath133 . denote @xmath139 and apply matlab 2011b functions to compute the condition number of @xmath1 and @xmath2 : @xmath140 let @xmath141 , @xmath142 , respectively .
the numerical results as follows ( see , table 1 ) . 12.1cmp45ptp70ptp70ptp70ptp55pt + + size ( @xmath143)&@xmath144&@xmath145&@xmath83&@xmath146 + 700 & 4.4239e+003 & 1.1861e+003 & 4.4239e+003 & 1.9991 + 800 & 5.0548e+003 & 1.3552e+003 & 5.0548e+003 & 1.9992 + 900 & 5.6858e+003 & 1.5242e+003 & 5.6858e+003 & 1.9993 + 1000 & 6.3167e+003 & 1.6933e+003 & 6.3167e+003 & 1.9994 + 1100 & 6.9477e+003 & 1.8623e+003 & 6.9477e+003 & 1.9994 + 1200 & 7.5786e+003 & 2.0314e+003 & 7.5786e+003 & 1.9995 + 1300 & 8.2095e+003 & 2.2005e+003 & 8.2095e+003 & 1.9995 + 1400 & 8.8405e+003 & 2.3695e+003 & 8.8405e+003 & 1.9995 + obviously , the results on this experiment conform with our theoretical analysis . + * acknowledgements*. _ the authors sincerely thank prof . n.j .
higham for bringing @xcite to our attention and some friendly discussion on this topic , and dr .
lin ( university of waterloo ) offered some useful materials on the manuscript of this paper , which led to a substantial improvement on this paper .
_ 99 a. george , kh.d .
ikramov , a.b .
kucherov , on the growth factor in gaussian elimination for generalized higham matrices , numer .
, 9 ( 2002 ) 107114 .
n.j . higham , factorizing complex symmetric matrices with positive real and imaginary parts , math .
comput . , 67 ( 1998 ) 15911599 .
d. lahaye , h.d .
gersem , s. vandewalle , k. hameyer , algebraic multigrid for complex symmetric systems , ieee trans .
magnetics , 36 ( 2000 ) 15351538 .
erlangga , c.w .
oosterlee , c. vuik , a novel multigrid based preconditioner for heterogeneous helmholtz problems , siam j. sci . comput . ,
27 ( 2006 ) 14711492 .
adams , algebraic multigrid methods for direct frequency response analyses in solid mechanics , comput .
mech . , 39 ( 2007 ) 497 - 507 .
j. smit , introduction to quantum fields on a lattice , cambridge lecture notes phys .
15 , cambridge university press , cambridge , 2002 .
j. brannick , m. brezina , d. keyes , o. livne , i. livshits , s. maclachlan , t. manteuffel , s. mccormick , j. ruge , and l. zikatanov , adaptive smoothed aggregation in lattice qcd , in domain decomposition methods in science and engineering xvi , lecture notes in comput .
55 , springer , new york , 2007 , 505512 .
o. axelsson , a. kucherov , real valued iterative methods for solving complex symmetric linear systems .
7 ( 2000 ) 197 - 218 .
a. feriani , f. perotti , v. simoncini , iterative system solvers for the frequency analysis of linear mechanical systems .
methods appl .
mech . eng . , 190 ( 2000 ) 17191739 .
bai , m. benzi , f. chen , modified hss iteration methods for a class of complex symmetric linear systems , comput . , 87 ( 2010 ) 93 - 111 .
stewart , introduction to matrix computations , academic press , new york , usa , 1973 .
higham , accuracy and stablility of numerical algorithms ( 2nd ed . ) , society for industrial and applied mathematics(siam ) press , philadelphia , pa , 2002 .
mr 97a : 65047 .
ikramov , a.b .
kucherov , bounding the growth factor in gaussian elimination for buckley s class of complex symmetric matrices , numer .
7 ( 2000 ) 269274 .
a. george , k. d. ikramov , on the growth factor in gaussian elimination for matrices with sharp angular field of values , calcolo , 41 ( 2004 ) 2736 .
horn , c.r .
johnson , topics in matrix analysis , cambridge university press , cambridge , 1991 .
zhang , equivalence of the wielandt inequality and the kantorovich inequality , lin .
alg . , 48 ( 2001 ) 275279 .
wang , f .- z .
zhang , schur complements and matrix inequalities of hadamard products , lin .
alg . , 43 ( 1997 ) 315326 . s .-
wang , m .- x .
jia , matrix inequalities(in chinese ) , 2nd ed .
science press , beijing , 2006 .
m. lin , reversed determinantal inequalities for accretive - dissipative matrices , math .
, 12 ( 2012 ) 955958 .
m. lin , a note on the growth factor in gaussian elimination for accretive - dissipative matrices , calcolo , 2013 , doi 10.1007/s10092 - 013 - 0089 - 1 . | the higham matrix is a complex symmetric matrix @xmath0 , where both @xmath1 and @xmath2 are real , symmetric and positive definite and @xmath3 is the imaginary unit . for any higham matrix @xmath4
, ikramov et al showed that the growth factor in gaussian elimination is less than @xmath5 . in this paper , based on the previous results , a new bound of the growth factor
is obtained by using the maximum of the condition numbers of matrixes @xmath1 and @xmath2 for the generalized higham matrix @xmath4 , which strengthens this bound to @xmath6 and proves the higham s conjecture .
[ multiblock footnote omitted ] [ multiblock footnote omitted ] |
numerical simulations of qcd , formulated on the lattice , make use of a variety of composite operators , made out of quark fields . in particular , matrix elements and correlation functions of such operators , which include local and extended bilinears , as well as four - fermion operators ,
are computed in order to study hadronic properties in this context .
a proper renormalization of these operators is essential for the extraction of physical results from the dimensionless quantities measured in numerical simulations .
the present paper is the second in a series of papers regarding the calculation of renormalization functions of fermion bilinear operators to two loops in lattice perturbation theory .
the calculation of the scalar and pseudoscalar cases was carried out in ref .
@xcite . in this work
we study the renormalization function @xmath5 of fermion bilinears @xmath6 on the lattice , where @xmath7 ( @xmath8 $ ] ) .
we consider both flavor singlet and nonsinglet operators .
we employ the standard wilson action for gluons and clover - improved wilson fermions . the number of quark flavors @xmath9 , the number of colors @xmath10 and the clover coefficient @xmath11 are kept as free parameters .
one necessary ingredient for the renormalization of fermion bilinears is the 2-loop quark field renormalization , @xmath12 , calculated in @xcite .
the one - loop expression for the renormalization function @xmath13 of the coupling constant is also necessary for expressing the results in terms of both the bare and the renormalized coupling constant .
our two - loop calculations have been performed in the bare and in the renormalized feynman gauge . for the latter ,
we need the 1-loop renormalization functions @xmath14 and @xmath15 of the gauge parameter and gluon field respectively , as well as the one - loop expressions for @xmath5 with an arbitrary value of the gauge parameter .
the main results presented in this work are the following 2-loop bare green s functions ( amputated , one - particle irreducible ( 1pi ) ) : * 2-pt function of the vector operator @xmath16 @xmath17 * 2-pt function of the axial - vector operator @xmath18 @xmath19 * 2-pt function of the tensor operator @xmath20 @xmath21 ( @xmath22 lattice spacing , @xmath23 external momentum ) . in general
, one can use bare green s functions to construct @xmath24 , the renormalization function for operator @xmath25 , computed within a regularization @xmath26 ( @xmath27 : lattice regularization ; @xmath28 : dimensional regularization ) and renormalized in a scheme @xmath29 .
we employ two widely used schemes to compute the various two - loop renormalization functions : * the @xmath30 scheme : @xmath31 , @xmath32 , @xmath33 * the @xmath4 scheme : @xmath34 , @xmath35 , @xmath36 for convenience , the results for @xmath24 are given in terms of both the bare coupling constant @xmath37 and the renormalized one : @xmath38 , @xmath39 .
finally , as one of several checks on our results , we construct the 2-loop renormalized green s functions in @xmath30 : @xmath40 ( @xmath41 ) , as well as their counterparts in @xmath4 : @xmath42 .
the values of all these functions , computed on the lattice , coincide with values computed in dimensional regularization ( we derive the latter from the results of ref .
@xcite ) .
the present work , along with @xcite , is the first two - loop computation of the renormalization of fermion bilinears on the lattice .
one - loop computations of the same quantities exist for quite some time now ( see , e.g. , @xcite , @xcite , @xcite and references therein ) .
there have been made several attempts to estimate @xmath43 non - perturbatively ; recent results can be found in refs .
a series of results have also been obtained using stochastic perturbation theory @xcite .
a related computation , regarding the fermion mass renormalization @xmath44 with staggered fermions can be found in @xcite .
the paper is organized as follows : section [ formulation ] provides a formulation of the problem , as well as all necessary definitions of renormalization schemes and of the quantities to compute .
section [ results ] describes our computational methods and results .
finally , in section [ discussion ] we discuss some salient features of our calculation , and comment on future extensions to the present work .
recently , there has been some interest in gauge theories with fermions in different representations @xcite of the gauge group .
such theories are being studied in various contexts @xcite , e.g. , supersymmetry @xcite , phase transitions @xcite , and the ads / qcd correspondence .
it is relatively straightforward to generalize our results to an arbitrary representation ; this is presented in appendix a. some special features of 2- , 3- and 4-index superficially divergent integrals are described in appendix b. finally , a detailed presentation of our calculation results on a per diagram basis , is provided in appendix c.
in the present work we employ the wilson formulation of the qcd action on the lattice , with the addition of the clover ( sw ) @xcite term for fermions . in standard notation
, it reads : @xmath45}\nonumber \\ & + & { i\over 4}\,c_{\rm sw}\,\sum_{f}\sum_{x,\,\mu,\,\nu } \bar{\psi}_{f}(x ) \,\sigma_{\mu\nu } \,{\hat f}_{\mu\nu}(x ) \,\psi_f(x ) , \label{latact}\end{aligned}\ ] ] @xmath46 @xmath47 is the standard pure gluon action , made out of @xmath48 plaquettes . the clover coefficient @xmath49 is treated here as a free parameter ; @xmath50 is the wilson parameter ( set to @xmath51 henceforth ) ; @xmath52 is a flavor index ; @xmath53 $ ] .
powers of the lattice spacing @xmath54 have been omitted and may be directly reinserted by dimensional counting .
the `` lagrangian mass '' @xmath55 is a free parameter here .
however , since we will be using mass independent renormalization schemes , all renormalization functions which we will be calculating , must be evaluated at vanishing renormalized mass , that is , when @xmath55 is set equal to the critical value @xmath56 : @xmath57 . as a prerequisite to our programme
, we will use the renormalization functions , @xmath15 , @xmath58 , @xmath59 , @xmath13 and @xmath14 , for the gluon , ghost and fermion fields ( @xmath60 ) , and for the coupling constant @xmath61 and gauge parameter @xmath62 , respectively ( for definitions of these quantities , see ref .
@xcite ) ; we will also need the fermion mass counterterm @xmath56 .
these quantities are all needed to one loop , except for @xmath59 which is required to two loops .
the value of each @xmath43 depends both on the regularization @xmath26 and on the renormalization scheme @xmath29 employed , and thus should properly be denoted as @xmath63 .
our one - loop results for the vector and axial - vector operators , even though performed in a generic gauge , turn out to be independent of the gauge parameter .
these results along with the one - loop expression for the tensor operator , are in agreement with results found in the literature ( see , e.g. , ref .
@xcite ) .
as mentioned before , we employ the @xmath0 renormalization scheme @xcite , which is more immediate for a lattice regularized theory .
it is defined by imposing a set of normalization conditions on matrix elements at a scale @xmath64 , where ( just as in the @xmath4 scheme ) @xcite : @xmath65 where @xmath66 is the euler constant and @xmath67 is the scale entering the bare coupling constant @xmath68 when regularizing in @xmath69 dimensions . for easier comparison with calculations coming from the continuum ,
we need to express our results in the @xmath4 scheme .
each renormalization function on the lattice , @xmath71 , may be expressed as a power series in the renormalized coupling constant @xmath38 .
for the purposes of our work the conversion of @xmath38 to @xmath4 is trivial since : @xmath72 the conversion of the gauge parameter @xmath73 to the @xmath4 scheme is given by @xcite : @xmath74 where the conversion factor @xmath75 may be calculated more easily in dimensional regularization ( dr ) @xcite , since the ratio of @xmath76 s appearing in eq .
( [ alphaconversion ] ) is necessarily _
regularization independent_. to one loop ,
the conversion factor @xmath75 equals : @xmath77 \label{ca}\ ] ] ( here , and throughout the rest of this work , both @xmath61 and @xmath62 are in the @xmath4 scheme , unless specified otherwise . )
once we have computed the renormalized green s functions in the @xmath30 scheme , we can construct their @xmath4 counterparts using the quark field conversion factor which , up to the required perturbative order , is given by : @xmath78 \label{cpsi}\end{aligned}\ ] ] where @xmath79 is the quadratic casimir operator in the fundamental representation of the color group ; @xmath80 is riemann s zeta function .
the lattice operators @xmath81 must , in general , be renormalized in order to have finite matrix elements .
we define renormalized operators by @xmath82 the flavor singlet axial - vector green s function receives additional contributions as compared to the nonsinglet case , while for the rest of the operators under study , singlet and nonsinglet green s functions coincide . for the vector ( v ) ,
axial - vector ( av ) and tensor ( t ) operators , the renormalization functions @xmath83 can be extracted through the corresponding bare 2-point functions @xmath84 ( amputated , 1pi ) on the lattice .
let us first express these bare green s functions in the following way : @xmath85 it is worth noting here that terms which break lorentz invariance ( but are compatible with hypercubic invariance ) , such as @xmath86 , turn out to be absent from all bare green s functions ; thus , the latter have the same lorentz structure as in the continuum .
let us also point out that the presence of the @xmath87 matrix in the tensor operator definition does not affect the bare green s function on the lattice , in the @xmath88 scheme .
we have performed the calculation both with and without the inclusion of the @xmath87 matrix , and we ended up with identical 2-point functions .
thus , for the purpose of converting our results to the @xmath4 scheme , we employed the conversion factors given in ref .
@xcite , where the definition of the tensor operator does not contain the @xmath87 matrix .
furthermore , we expect that @xmath89 must vanish , since this is the case for the corresponding quantity coming from the continuum .
indeed , after performing the calculation on the lattice , it turns out that all contributions of this type vanish .
once all necessary feynman diagrams contributing to the bare green s functions presented above are evaluated , one can obtain the renormalization functions for the three operators through the following conditions : @xmath90_{q^2=\bar{\mu}^2 } & = & \gamma_\mu \label{zxrulesa } \\
\lim_{a_{_{\rm l}}\rightarrow 0}\left[z_{\psi}^{l , ri^{\prime}}\,z_{av}^{l , ri^{\prime}}\,\sigma^{(1),\,l}_{av } ( q a_{_{\rm l}})\right]_{q^2=\bar{\mu}^2 } & = & \gamma_5\,\gamma_\mu \label{zxrulesb } \\
\lim_{a_{_{\rm l}}\rightarrow 0}\left[z_{\psi}^{l , ri^{\prime}}\,z_t^{l , ri^{\prime}}\,\sigma^{(1),\,l}_t ( q a_{_{\rm l}})\right]_{q^2=\bar{\mu}^2 } & = & \gamma_5\,\sigma_{\mu\,\nu } \label{zxrulesc}\end{aligned}\ ] ] where : @xmath91 the conversion of the quantities @xmath92 to the @xmath4 scheme is a straightforward procedure . in the case of the vector and tensor operators , the renormalization functions , @xmath34 and @xmath36 ,
can be obtained by : @xmath93 where @xmath94 are _ regularization independent _ conversion factors ( @xmath95 ) .
these conversion factors have been calculated in dimensional regularization @xcite : @xmath96 \label{ct}\end{aligned}\ ] ] unlike the tensor operator , where the presence of the @xmath87 matrix is irrelevant , the axial - vector bilinear ( @xmath97 ) requires special attention also in the @xmath4 scheme , due to the non - unique generalization of @xmath87 to d dimensions . a practical definition of @xmath87 for multiloop calculations , which is most commonly employed in dimensional regularization and does not suffer from inconsistencies is @xcite : @xmath98 of course , @xmath87 as defined in eq . ( [ gamma5 ] )
does not anticommute ( in @xmath99 dimensions ) with @xmath100 , for @xmath101 ; an ultimate consequence of this fact is that ward identities involving the axial - vector and pseudoscalar operators , renormalized in this way , are violated . to obtain a correctly normalized axial - vector operator @xcite , @xmath102 , one must introduce an extra _ finite _ factor , @xmath103 , in addition to the usual renormalization function @xmath104 ( the latter only contains poles in @xmath105 ) .
we set : @xmath106 for the definition of @xmath103 we must express the @xmath4 renormalized green s functions @xmath107 , @xmath108 as well as the renormalized green s function @xmath109 ( corresponding to @xmath110 ) , in a form similar to eq .
( [ brgrnfunctn ] ) : @xmath111 @xmath103 is then defined by the requirement that the renormalized green s functions @xmath112 and @xmath113 coincide : @xmath114 eq.([axialcurrent ] ) is valid for both the singlet and nonsinglet currents , provided of course , the appropriate choice for @xmath103 is used .
thus , we have two different expressions , @xmath115 and @xmath116 corresponding to the singlet and nonsinglet axial - vector operator , respectively .
they are gauge independent and differ only in the @xmath117 term ; this is expected considering the fact that the additional feynman diagrams contributing to the singlet axial operator have the insertion within the closed fermion loop .
both @xmath115 and @xmath116 were evaluated in dr @xcite and up to two loops they read : @xmath118 @xmath35 can now be obtained by : @xmath119 where @xmath103 stands for @xmath120 or @xmath121 ( eqs .
( [ z5s]-[z5ns ] ) ) , for the singlet or nonsinglet cases , respectively .
similarly , one can convert the @xmath30 renormalized green s functions , @xmath122 , to their @xmath4 counterparts , through : @xmath123 ( in eqs.([axialconversion]-[renormgreen ] ) it is understood that powers of @xmath124 , implicit in @xmath30 quantities , must also be converted to @xmath125 , respectively , using eqs.([gconversion]-[alphaconversion ] ) ) .
note that the combination @xmath126 appearing above yields the value of @xmath127 .
the feynman diagrams contributing to the bare green s functions for the vector , axial - vector and tensor operators , @xmath128 , at 1- and 2-loop level , are shown in figs .
[ zvat1loopdiagrams ] and [ zvat2loopdiagrams ] , respectively . for flavor
singlet bilinears , there are 4 extra diagrams , shown in fig .
[ singletvat ] , which contain the operator insertion inside a closed fermion loop .
these diagrams give a nonzero contribution only in the axial - vector case . , @xmath129 and @xmath130 .
a wavy ( solid ) line represents gluons ( fermions ) .
a cross denotes the dirac matrices @xmath131 ( vector ) , @xmath132 ( axial vector ) and @xmath133 ( tensor ) . [
zvat1loopdiagrams ] ] the evaluation and algebraic manipulation of feynman diagrams , leading to a code for numerical loop integration , is performed automatically using our software for lattice perturbation theory , written in mathematica .
the most laborious aspect of the procedure is the extraction of the dependence on the external momentum @xmath134 .
this is a delicate task at two loops ; for this purpose , we cast algebraic expressions ( typically involving thousands of summands ) into terms which can be naively taylor expanded in @xmath134 to the required order , plus a smaller set of terms containing superficial divergences and/or subdivergences .
the latter can be evaluated by an extension of the method of ref .
@xcite to 2 loops ; this entails analytical continuation to @xmath135 dimensions , and splitting each expression into a uv - finite part ( which can thus be calculated in the continuum , using the methods of ref .
@xcite ) , and a part which is polynomial in @xmath134 . a primitive set of divergent lattice integrals involving gluon propagators , which can be obtained in this manner , can be found in ref .
@xcite . due to the presence of at least one free lorentz index in the definition of the operators ( for the case of the tensor bilinear
there are two such indices ) , it is possible to end up dealing with superficially divergent integrals with two , three or even four free lorentz indices . in appendix b , we provide a brief description of the manipulations performed to resolve such terms , based on the method described above . some of the diagrams contributing to @xmath136 are infrared divergent when considered separately , and thus must be grouped together in order to give finite results .
such groups are formed by diagrams ( 3 - 7 ) , ( 8 - 9 ) , ( 10 - 11,19 ) in fig .
[ zvat2loopdiagrams ] and diagrams ( 1 - 2 ) , ( 3 - 4 ) in fig .
[ singletvat ] . in figs .
[ zvat1loopdiagrams ] to [ singletvat ] , `` mirror '' diagrams ( those in which the direction of the external fermion line is reversed ) should also be included . in most cases ,
these coincide trivially with the original diagrams ; even in the remaining cases , they can be seen to give equal contribution , by invariance under charge conjugation . as mentioned before
, all calculations should be performed at vanishing renormalized mass ; this can be achieved by working with massless fermion propagators , provided an appropriate fermion mass counterterm is introduced ( diagram 11 in fig .
[ zvat2loopdiagrams ] ) .
all two - loop diagrams have been calculated in the bare feynman gauge ( @xmath137 ) .
one - loop diagrams have been calculated for generic values of @xmath138 ; this allows us to convert our two - loop results to the renormalized feynman gauge ( @xmath139 or @xmath140 ) . after performing the calculation for the cases of the vector and axial - vector operator , we see that one - loop expressions for the renormalization functions do not depend on the gauge parameter . especially for the case of the vector operator , having in mind eqs .
( [ gconversion]-[alphaconversion ] ) and eq .
( [ cv ] ) , this fact causes the lattice results in the @xmath88 and in the @xmath4 scheme to coincide .
numerical loop integration was carried out by our `` integrator '' program , a _ metacode _ written in mathematica , for converting lengthy integrands into efficient fortran code .
two - loop numerical integrals were evaluated as sums over finite lattices , of size up to @xmath141 ; the results were then extrapolated to @xmath142 .
extrapolation is the only source of systematic error ; this error can be estimated quite accurately ( see , e.g. ref .
@xcite ) , given that @xmath143-dependence of results can only span a restricted set of functional forms .
1-loop results for @xmath92 are presented below in a generic gauge .
as it turns out , only the tensor renormalization function depends on the gauge parameter , while for all other operators , one - loop expressions that emerge are gauge independent .
the errors appearing in the expression for @xmath33 , result from the @xmath144 extrapolation .
@xmath145 } \label{zt1loopri}\end{aligned}\ ] ] the corresponding expressions for @xmath146 , @xmath147 can be read off from eqs .
( [ zv2loopri]-[za2loopri ] ) below . in all cases , one - loop results in the @xmath4 scheme ( @xmath34 , @xmath35 , @xmath36 ) present no dependence on the gauge parameter .
in order to derive the expressions for the bare green s functions @xmath148 , one must evaluate all feynman diagrams presented in figs .
[ zvat1loopdiagrams]-[singletvat ] .
the extraction of @xmath149 , @xmath150 and @xmath151 is then straightforward via eqs .
( [ zxrulesa]-[zxrulesc ] ) ( for @xmath152 ) , and via eqs . ( [ vtconversion]-[ct ] , [ axialconversion ] ) ( for @xmath153 ) .
to this end , we need the following one - loop expression for @xmath154 ( note that @xmath155 to this order ) : @xmath156 + { \cal o}\left(g_{\rm o}^4\right ) \label{za1loop}\end{aligned}\ ] ] to express our results in terms of the renormalized coupling constant , we also need the one - loop expression for @xmath157 : @xmath158 + { \cal o}\left(g_{\rm o}^4\right ) \label{zg1loop}\end{aligned}\ ] ] eqs .
( [ za1loop ] , [ zg1loop ] ) are in agreement with older references ( see , e.g. , ref .
@xcite ) .
a final necessary ingredient is the two - loop expression for @xmath159 , as required by eqs.([zxrulesa]-[zxrulesc ] ) ; this was calculated in ref .
@xcite , in the renormalized feynman gauge @xmath139 , and is included here for completeness : @xmath160 } \nonumber \\ & + & \frac{g_\circ^4}{(16\pi^2)^2}\,c_f \bigg { [ } \ln^2(a_{_{\rm l}}^2 \bar{\mu}^2 ) \left ( \frac{1}{2 } c_f + \frac{2}{3 } n_f -\frac{8}{3 } n_c
\right ) \nonumber \\ & & \quad\quad\,\ , + \ln(a_{_{\rm
l}}^2 \bar{\mu}^2 ) \big{(}-6.36317446(8)\,n_f + 0.79694523(2)\,n_f\,c_{{\rm sw } } \nonumber \\ & & \hskip 3.2 cm - 4.712691443(4)\,n_f\,c_{{\rm sw}}^2 \nonumber \\ & & \hskip 3.2 cm + 49.83082185(5)\,c_f - 2.24886861(7)\,c_f\,c_{{\rm sw } } \nonumber \\ & & \hskip 3.2 cm - 1.39726705(1)\,c_f\,c_{{\rm sw}}^2 + 29.03029398(4)\,n_c\big { ) } \nonumber \\ & & \quad\quad\,\ , + n_f\,\big{(}-7.838(2 ) + 1.153(1)\,c_{{\rm sw } } + 3.202(3)\,c_{{\rm sw}}^2 \nonumber \\ & & \hskip 2.2 cm + 6.2477(6)\,c_{{\rm sw}}^3 + 4.0232(6)\,c_{{\rm sw}}^4 \big { ) } \nonumber \\ & & \quad\quad\,\ , + c_f\,\big{(}505.39(1 ) - 58.210(9)\,c_{{\rm sw } } + 20.405(5)\,c_{{\rm sw}}^2 \nonumber \\ & & \hskip 2.2 cm + 18.8431(8)\,c_{{\rm sw}}^3 + 4.2793(2)\,c_{{\rm sw}}^4 \big { ) } \nonumber \\ & & \quad\quad\,\ , + n_c\,\big{(}-20.59(1 ) - 3.190(5)\,c_{{\rm sw } } - 23.107(6)\,c_{{\rm sw}}^2 \nonumber \\ & & \hskip 2.2 cm -5.7234(5)\,c_{{\rm sw}}^3 - 0.7938(1)\,c_{{\rm sw}}^4 \big { ) } \bigg { ] } \label{zpsi2loopri}\end{aligned}\ ] ] we present below @xmath31 , @xmath32 and @xmath33 to two loops in the renormalized feynman gauge @xmath139 ; we also present the @xmath4 analogues @xmath35 and @xmath36 in the gauge @xmath140 ( as already mentioned , @xmath161 ) .
the bare green s functions are relegated to appendix c , where a _ per diagram _ breakdown of the results is provided .
it is a straightforward exercise to recover the total bare green s functions from the corresponding @xmath76 s and the renormalized green s functions . @xmath162 } \nonumber \\ & + & \frac{g_\circ^4}{(16\pi^2)^2}\,c_f \bigg { [ } n_f\,\big{(}25.610(3 ) -11.058(1)\,c_{{\rm sw } } + 33.937(3)\,c_{{\rm sw}}^2 \nonumber \\ & & \hskip 3 cm -13.5286(6)\,c_{{\rm sw}}^3 -1.2914(6)\,c_{{\rm sw}}^4 \big { ) } \nonumber \\ & & \qquad\qquad\quad + c_f\,\big{(}-539.78(1 ) -223.57(2)\,c_{{\rm sw } } -104.116(5)\,c_{{\rm sw}}^2 \nonumber \\ & & \hskip 3.2 cm -32.2623(8)\,c_{{\rm sw}}^3 + 4.5575(3)\,c_{{\rm sw}}^4 \big { ) } \nonumber \\ & & \qquad\qquad\quad + n_c\,\big{(}-51.59(1 ) + 18.543(5)\,c_{{\rm sw } } + 20.960(6)\,c_{{\rm sw}}^2 \nonumber \\ & & \hskip 3.2 cm + 2.5121(5)\,c_{{\rm sw}}^3 + 0.1765(1)\,c_{{\rm sw}}^4 \big { ) } \bigg { ] } \label{zv2loopri}\end{aligned}\ ] ] @xmath163 } \nonumber \\ & + & \frac{g_\circ^4}{(16\pi^2)^2}\,c_f \bigg { [ } n_f\,\big{(}18.497(3 ) -1.285(1)\,c_{{\rm sw } } + 19.071(3)\,c_{{\rm sw}}^2 \nonumber \\ & & \hskip 3 cm + 1.0333(6)\,c_{{\rm sw}}^3 -6.7549(6)\,c_{{\rm sw}}^4 \big { ) } \nonumber \\ & & \hskip 2 cm + c_f\,\big{(}-184.01(1 ) -389.86(1)\,c_{{\rm sw } } -166.738(6)\,c_{{\rm sw}}^2 \nonumber \\ & & \hskip 3 cm + 7.894(1)\,c_{{\rm sw}}^3 + 4.3201(3)\,c_{{\rm sw}}^4 \big { ) } \nonumber \\ & & \hskip 2 cm + n_c\,\big{(}-21.62(1 ) -33.652(5)\,c_{{\rm sw } } + 26.636(6)\,c_{{\rm sw}}^2 \nonumber \\ & & \hskip 3 cm + 10.2186(5)\,c_{{\rm sw}}^3 + 1.4893(1)\,c_{{\rm sw}}^4 \big { ) } \bigg { ] } \label{za2loopri}\\[2.0ex ] z_{av}^{l,\overline{ms } } = 1 & + & \frac{g_\circ^2}{16\pi^2}\,c_f\,\bigg{[}-11.796283066(5 ) -0.247827627(3)\,c_{{\rm sw } } + 2.251366176(5)\,c_{{\rm sw}}^2 \bigg { ] } \nonumber \\ & + & \frac{g_\circ^4}{(16\pi^2)^2}\,c_f\,\bigg{[}\ln(a_{_{\rm l}}^2 \bar{\mu}^2 ) \left(\frac{8}{3}\,n_f - \frac{44}{3}\,n_c \right ) \nonumber \\ & & \hskip 2 cm + n_f\,\big{(}14.045(3 ) + 1.903(1)\,c_{{\rm sw}}+0.220(3)\,c_{{\rm sw}}^2
\nonumber \\ & & \hskip 3 cm + 1.0333(6)\,c_{{\rm sw}}^3 -6.7549(6)\,c_{{\rm sw}}^4 \big { ) } \nonumber \\ & & \hskip 2 cm + c_f\,\big(-95.28(1)-390.85(1)\,c_{{\rm sw } } -157.733(6)\,c_{{\rm sw}}^2 \nonumber \\ & & \hskip 3 cm + 7.894(1)\,c_{{\rm sw}}^3 + 4.3201(3)\,c_{{\rm sw}}^4\big ) \nonumber \\ & & \hskip 2 cm + n_c\,\big(18.67(1)-33.652(5)\,c_{{\rm sw}}+26.636(6)\,c_{{\rm sw}}^2 \nonumber \\ & & \hskip 3 cm + 10.2186(5)\,c_{{\rm sw}}^3 + 1.4893(1)\,c_{{\rm sw}}^4 \big ) \bigg { ] } \label{za2loopms}\end{aligned}\ ] ] @xmath164 & & \hskip 1.7 cm + 3.91333261(4)\,c_{{\rm sw } } + 1.972295300(5)\,c_{{\rm sw}}^2 \bigg { ] } \nonumber \\ & + & \frac{g_\circ^4}{(16\pi^2)^2}\,c_f \bigg { [ } \ln^2(a_{_{\rm l}}^2 \bar{\mu}^2 ) \left ( \frac{1}{2 } c_f -\frac{1}{3 } n_f -\frac{11}{6 } n_c
\right ) \nonumber \\ & & \hskip 2 cm + \ln(a_{_{\rm l}}^2 \bar{\mu}^2 ) \big{(}3.1685002(6)\,n_f -0.79694524(6)\,n_f\,c_{{\rm sw } } \nonumber \\[-0.44ex ] & & \hskip 4 cm + 4.71269143(3)\,n_f\,c_{{\rm sw}}^2 \nonumber \\[-0.44ex ] & & \hskip 4 cm -13.96033835(7)\,c_f -3.9133325(1)\,c_f\,c_{{\rm sw } } \nonumber \\ [ -0.44ex ] & & \hskip 4 cm -1.97229535(2)\,c_f\,c_{{\rm sw}}^2 -25.04361149(6)\,n_c\big { ) } \nonumber \\[-0.44ex ] & & \hskip 2 cm + n_f\,\big{(}16.923(6 ) -8.399(2)\,c_{{\rm sw } } + 18.711(3)\,c_{{\rm sw}}^2 \nonumber \\[-0.44ex ] & & \hskip 3 cm -10.8351(8)\,c_{{\rm sw}}^3 -5.1253(6)\,c_{{\rm sw}}^4 \big { ) } \nonumber \\[-0.44ex ] & & \hskip 2 cm + c_f\,\big{(}-868.0(1 ) + 551.6(2)\,c_{{\rm sw } } + 63.9(1)\,c_{{\rm sw}}^2 \nonumber \\[-0.44ex ] & & \hskip 3 cm -79.49(1)\,c_{{\rm sw}}^3 -12.586(1)\,c_{{\rm sw}}^4 \big { ) } \nonumber \\[-0.44ex ] & & \hskip 2 cm + n_c\,\big{(}-15.76(8 ) + 27.6(1)\,c_{{\rm sw } } + 38.2(1)\,c_{{\rm sw}}^2 \nonumber \\[-0.44ex ] & & \hskip 3 cm + 7.021(8)\,c_{{\rm sw}}^3 + 1.6653(9)\,c_{{\rm sw}}^4 \big { ) } \bigg { ] } \label{zt2loopri}\\[1.0ex ] z_t^{l,\overline{ms } } = 1 & + & \frac{g_\circ^2}{16\pi^2}\,c_f\,\bigg { [ } -\ln(a_{_{\rm l}}^2 \bar{\mu}^2 ) -17.018079209(7 ) \nonumber \\[-0.44ex ] & & \hskip 1.7 cm + 3.91333261(4)\,c_{{\rm sw } } + 1.972295300(5)\,c_{{\rm sw}}^2 \bigg { ] } \nonumber \\ & + & \frac{g_\circ^4}{(16\pi^2)^2}\,c_f\,\bigg { [ } \ln^2(a_{_{\rm l}}^2 \bar{\mu}^2 ) \left(-\frac{1}{3}\,n_f + \frac{1}{2}\,c_f + \frac{11}{6}\,n_c \right ) \nonumber \\ & & \hskip 2 cm + \ln(a_{_{\rm l}}^2 \bar{\mu}^2 ) \big{(}2.5018336(6)\,n_f -0.79694524(6)\,n_f\,c_{{\rm sw } } \nonumber \\[-0.44ex ] & & \hskip 4 cm + 4.71269143(3)\,n_f\,c_{{\rm sw}}^2 \nonumber \\[-0.44ex ] & & \hskip 4 cm -12.96033835(7)\,c_f -3.9133325(1)\,c_f\,c_{{\rm sw}}\nonumber \\[-0.44ex ] & & \hskip 4 cm -1.97229535(2)\,c_f\,c_{{\rm sw}}^2 -21.37694482(6)\,n_c \big { ) } \nonumber \\[-0.44ex ] & & \hskip 2 cm + n_f\,\big{(}21.989(6)-9.196(2)\,c_{{\rm sw}}+23.424(3)\,c_{{\rm sw}}^2 \nonumber \\[-0.44ex ] & & \hskip 3 cm -10.8351(8)\,c_{{\rm sw}}^3 -5.1253(6)\,c_{{\rm sw}}^4 \big { ) } \nonumber \\[-0.44ex ] & & \hskip 2 cm + c_f\,\big(-893.2(1)+547.7(2)\,c_{{\rm sw } } + 61.9(1)\,c_{{\rm sw}}^2 \nonumber \\[-0.44ex ] & & \hskip 3 cm -79.49(1)\,c_{{\rm sw}}^3 -12.586(1)\,c_{{\rm sw}}^4\big ) \nonumber \\[-0.44ex ] & & \hskip 2 cm + n_c\,\big(-41.44(8)+27.6(1)\,c_{{\rm sw}}+38.2(1)\,c_{{\rm sw}}^2 \nonumber \\[-0.44ex ] & & \hskip 3 cm + 7.021(8)\,c_{{\rm sw}}^3 + 1.6653(9)\,c_{{\rm sw}}^4 \big ) \bigg { ] } \label{zt2loopms}\end{aligned}\ ] ] all expressions reported thus far for @xmath165 , @xmath129 and @xmath130 refer to flavor nonsinglet operators . in the case of @xmath165 and @xmath130 ,
all diagrams of fig .
[ singletvat ] vanish , so that singlet and nonsinglet results coincide , just as in dimensional regularization . for @xmath129 on the other hand , the above diagrams give an additional contribution : @xmath166 the same extra contribution applies also to the @xmath4 scheme . for the sake of completeness , and as an additional check on our results
, we compute the renormalized green s functions ( for _ vanishing _ renormalized mass ) . since the bare green s functions have two contributions of different structure ( as defined in eq .
( [ 2ptfunct ] ) , see also eq .
( [ rngrnfunctn ] ) ) , we derive the renormalized expressions for these contributions separately : @xmath167 where @xmath168 . similarly for @xmath4 , taking into account eq .
( [ renormgreen ] ) .
since these functions are regularization independent , they can be calculated also using , e.g. , dimensional regularization .
we have computed @xmath169 , @xmath170 and @xmath171 in both ways : either starting from our eqs.([zt1loopri]-[zt2loopms ] ) or using renormalization functions from dimensional regularization @xcite . in all cases the two ways are in complete agreement .
we obtain : @xmath172 \label{vector1renormgreenfnri } \\
g^{(2)\,ri^{\prime}}_v(q)&= & \frac{g_{ri^{\prime}}^2}{16\pi^2}\,c_f\left(-2 \ln(\bar{\mu}^2/q^2)\right ) \nonumber \\ & + & \frac{g_{ri^{\prime}}^4}{(16\pi^2)^2}\,c_f\bigg[\ln(\bar{\mu}^2/q^2)\left(-2\,c_f-4\,n_c\right ) + \frac{38}{9}\,n_f + 3\,c_f-\frac{251}{9}\,n_c\bigg ] \label{vector2renormgreenfnri } \end{aligned}\ ] ] the vector renormalized green s function in the @xmath88 scheme coincides with the corresponding axial - vector expression , and thus eqs.([vector1renormgreenfnri]-[vector2renormgreenfnri ] ) also hold for the axial - vector case : @xmath173 , @xmath174 . of course , even though the @xmath4 expression for the vector renormalization function , @xmath34 , coincides with the @xmath88 expression , that is not the case for the renormalized @xmath4 green s function , due to @xmath175 appearing in eq .
( [ renormgreen ] ) .
this factor results in the following quantities : @xmath176 \label{vector1renormgreenfnms } \\
g^{(2)\,\overline{ms}}_v(q ) & = & -\frac{g_{\overline{ms}}^2}{16\pi^2}\,2\,c_f \nonumber \\ & + & \frac{g_{\overline{ms}}^4}{(16\pi^2)^2}\,c_f\bigg[\ln(\bar{\mu}^2/q^2)\left(-2\,c_f-4\,n_c\right ) + \frac{38}{9}\,n_f + c_f-\frac{251}{9}\,n_c\bigg ] \label{vector2renormgreenfnms } \end{aligned}\ ] ] furthermore , the axial - vector renormalized 2-point functions in the @xmath4 scheme differ from eqs .
( [ vector1renormgreenfnms]-[vector2renormgreenfnms ] ) , due to the finite conversion factor @xmath116 ; they read : @xmath177 \label{axial1renormgreenfnms } \\
g^{(2)\,\overline{ms}}_{av}(q)&= & -\frac{g_{\overline{ms}}^2}{16\pi^2}\,2\,c_f \nonumber \\ & + & \frac{g_{\overline{ms}}^4}{(16\pi^2)^2}\,c_f\bigg[\ln(\bar{\mu}^2/q^2)\left(-2\,c_f-4\,n_c\right ) + \frac{38}{9}\,n_f - 7\,c_f-\frac{251}{9}\,n_c\bigg ] \label{axial2renormgreenfnms } \end{aligned}\ ] ] if one considers the singlet axial - vector current , then there exists an extra contribution to the expressions above : @xmath178 for the @xmath88 scheme , similar relations hold , the only difference being the absence of the factors @xmath115 , @xmath116 ; we obtain : @xmath179 finally , for the tensor renormalized green s function , we obtain : @xmath180 \label{tensor1renormgreenfnri}\end{aligned}\ ] ] just as was expected from dimensional regularization , @xmath181 .
the corresponding quantity in the @xmath4 scheme reads : @xmath182 \label{tensor1renormgreenfnms } \end{aligned}\ ] ] in figs .
[ zvriplot ] , ( [ zariplot],[zamsplot ] ) , ( [ ztriplot],[ztmsplot ] ) we plot @xmath31 , ( @xmath32 , @xmath35 ) and ( @xmath33 , @xmath36 ) , respectively , as a function of @xmath49 .
values of the clover parameter used in simulations lie within the typical range @xmath183 . for definiteness ,
we have set @xmath184 , @xmath185 and @xmath186 .
our results up to two loops for each @xmath76 are shown for both @xmath187 and @xmath188 , and compared to the corresponding one - loop results .
furthermore , in the axial - vector case , we also present the two - loop result for the flavor singlet operator , for @xmath188 . in fig .
[ zvatriplot ] we present , on the same plot , the values of @xmath31 , @xmath32 , @xmath189 and @xmath33 up to 2 loops , versus @xmath49 .
we have chosen @xmath184 , @xmath185 , @xmath188 and @xmath190 .
the corresponding results in the @xmath4 scheme are plotted in fig .
[ zvatmsplot ] .
in this paper we have reported results regarding the vector , axial - vector and tensor fermion bilinear operators .
this work , along with a previously published paper @xcite regarding the scalar and pseudoscalar operators , provide a complete two loop calculation for the renormalization functions for local fermion bilinears , considering both the singlet and nonsinglet cases .
the two - loop wave function renormalization constant , @xmath59 , which is a prerequisite for our calculation , was presented in ref .
@xcite ( the reader should also refer to this paper for any necessary notation not included in the present sequel paper ) .
it is clear from figs .
[ zvriplot ] to [ ztmsplot ] that the two - loop renormalization functions differ significantly from 1-loop values ; this difference must then be properly taken into account in reducing systematic error in mc simulations . at the same time ,
2-loop contributions are typically smaller than 1-loop contributions , especially for @xmath191 , indicating that the ( asymptotic ) perturbative series are under control .
the results are presented as a function of the clover parameter , where the values of @xmath3 lie within the standard range @xmath192 .
optimal values for @xmath3 , which have been estimated both non - perturbatively @xcite and perturbatively ( to 1-loop ) @xcite , lie within this range . a breakdown of our results on a _ per diagram _ basis is presented in appendix c , for completeness .
as already mentioned , we take into account both singlet and nonsinglet operators . after evaluating all feynman diagrams
involved , we found that , for the vector and tensor operators , singlet renormalization functions coincide with nonsinglet ones . on the other hand
, the axial - vector operator receives an additional contribution in the flavor singlet case .
the numerical integrations over loop momenta were executed on a pentium iv cluster ; they required the equivalent of 60 months on a single cpu .
a possible extension to the present calculation is the renormalization of more extended operators , with the same continuum limit as we have considered here . a standard basis of higher dimension operators , with the same quantum numbers as the local bilinears which we have considered , can be found e.g. in ref . @xcite .
such operators are frequently used to reduce @xmath193 effects .
a number of additional feynman diagrams must be introduced , since the vertices coming from these operators may also contain gluon lines . however , the additional integrals resulting after the contractions will be free of superficial divergences , leading to a less cumbersome computation , despite an increase in the size of the integrals .
further directions regard higher dimensional operators , such as @xmath194 , which enter structure function calculations , and 4-fermion operators .
finally , our computation can be easily extended to improved lattice actions . with regard to improved fermion actions , such as those containing twisted mass terms @xcite or sterwalder - seiler terms @xcite , our results remain unchanged , since they pertain to mass - independent schemes .
improving the gluon action , on the other hand , is more cpu consuming , but conceptually straightforward : splitting ( in iterative fashion ) the symanzik propagator into a wilson gluon propagator plus the remainder , leads to the same bare green s functions as the ones presented in this paper , with the addition of superficially convergent terms , which can be more easily manipulated .
based on our experience with other similar calculations , the algebraic expressions for the integrands will grow roughly by a factor of 5 ; furthermore , the gluon propagator must now be inverted numerically for each value of the momentum , leading to an additional factor of @xmath1952 in cpu time . finally , if one wishes to employ more than one set of values for the symanzik coefficients , cpu time for numerical integration will increase almost proportionately .
the results presented up to this point regarded renormalization constants of various fermion bilinear operators constructed with fermions in the fundamental representation of the gauge group .
our results were expressed in terms of the clover parameter , @xmath49 , the number of fermions , @xmath9 , the number of colors , @xmath10 , and the quadratic casimir operator in the fundamental representation , @xmath196 .
recently there has been interest in theories with fermions in other representations ; some preliminary non - perturbative calculations have also appeared ( see e.g. @xcite , @xcite ) . in this appendix
we describe the conventions we use in our work , regarding the generators of the algebra , and we then express our findings in an arbitrary representation
. our results for @xmath165 , @xmath129 , @xmath130 , eqs.([zv2loopri ] , [ za2loopri ] , [ zt2loopri ] ) , can be easily generalized to an action with wilson / clover fermions in an arbitrary representation @xmath197 , of dimensionality @xmath198 . in this case
, the gluon part of the action remains the same , while all link variables appearing in the fermion part of the action assume the form : @xmath199 using standard notation and conventions , the generators @xmath200 in the fundamental representation satisfy : @xmath201 = i\,f^{abc}\,t^c,\quad \sum_at^at^a \equiv \openone\cdot c_f = \openone\cdot { n_c^2 - 1\over 2n_c}\ , , \quad { \rm tr}(t^at^b)\equiv\delta^{ab}\,t_f= \delta^{ab}\,{1\over2}\ ] ] in the representation @xmath197 we have : @xmath202 = i\,f^{abc}\,t^c_r,\quad \sum_at^a_rt^a_r \equiv \openone\cdot c_r,\quad { \rm tr}(t^a_rt^b_r)\equiv\delta^{ab}\,t_r\ ] ] where : @xmath203 .
for the 1-loop quantities , eqs .
( [ za1loop ] , [ zg1loop ] ) , converting to the representation @xmath197 is a straightforward substitution : @xmath204 and , in addition , for eq .
( [ zt1loopri ] ) : @xmath205 aside from these changes , all algebraic expressions ( and the numerical coefficients resulting from loop integrations ) remain the same .
a similar reasoning applies to the 2-loop quantities in eqs .
( [ zv2loopri ] , [ za2loopri ] , [ zt2loopri ] ) :
for most diagrams , once their value is expressed as a linear combination of @xmath206 , @xmath207 and @xmath208 , it suffices to apply substitutions ( [ tr ] ) and ( [ cr ] ) .
the only exceptions are diagrams containing a gluon tadpole [ diagram 3 of fig .
[ zvat2loopdiagrams ] ; 1-loop diagrams , when expressed in terms of @xmath209 , @xmath210 by means of @xmath13 , @xmath15 ] : in these cases , only one power of @xmath196 should be changed to @xmath212 ; a possible additional power of @xmath196 originates from the gluon tadpole and should stay as is .
this peculiarity implies that , in order to perform the substitutions as described above , one must start from the _ per diagram _ breakdown of 2-loop results . to avoid a lengthy presentation , we apply , instead , substitutions ( [ tr ] ) and ( [ cr ] ) indiscriminately on eqs .
( [ zv2loopri ] , [ za2loopri ] , [ zt2loopri ] ) ; consequently , we must then add a correction term , as follows : @xmath213 & & \hskip 4.5 cm -21.4434142(2)\,c_{{\rm
sw}}^2 \big ] \\
z_{av}^{l , ri^{\prime}}\big|_r & = & z_{av}^{l , ri^{\prime}}\big|_{c_f\to c_r\,,\ n_f\to 2 n_f\,t_r } \nonumber \\ & & + \frac{g_\circ^4}{(16\pi^2)^2}\,c_r\,(c_r - c_f ) \cdot \big[623.6122595(2 ) + 9.7838425(1)\,c_{{\rm sw}}\nonumber \\[-0.5ex ] & & \hskip 4.5 cm -88.8803741(2)\,c_{{\rm sw}}^2 \big ] \\
z_t^{l , ri^{\prime}}\big|_r & = & z_t^{l , ri^{\prime}}\big|_{c_f\to c_r\,,\ n_f\to 2 n_f\,t_r } \nonumber\\ & & + \frac{g_\circ^4}{(16\pi^2)^2}\,c_r\,(c_r - c_f ) \cdot \big [ 4 \pi^2 \ln(a_{_{\rm l}}^2 \bar{\mu}^2 ) + 632.3684202(3)\nonumber\\[-0.5ex ] & & \hskip 4.5 cm -154.492179(2)\,c_{{\rm sw } } -77.8630975(2)\,c_{{\rm sw}}^2 \big]\end{aligned}\ ] ] actually , the reader could arrive at these results without knowledge of the per diagram breakdown , by virtue of the following fact : all `` exceptional '' powers of @xmath196 cancel out of @xmath31 , @xmath32 , @xmath33 , if these are expressed in terms of the renormalized coupling constant @xmath209 .
thus , one may : * express eqs .
( [ zv2loopri ] , [ za2loopri ] , [ zt2loopri ] ) in terms of @xmath38 by means of @xmath214 , with @xmath215 in the fundamental representation ( eq . ( [ zg1loop ] ) )
* apply substitutions ( [ tr ] ) , ( [ cr ] ) throughout * if desired , reexpress everything in terms of @xmath37 ( using @xmath216 from eq .
( [ zg1loop ] ) , with @xmath217 and @xmath196 as is ) no correction terms are necessary in this procedure .
in the case of the scalar and pseudoscalar bilinears @xcite all superficially divergent terms involved at most one free lorentz index , but when one considers other bilinear operators different structures may arise .
for example , in the case of the vector and axial - vector operators , two - index integrals may arise and , of course , when working with the tensor operator three- or even four - index integrals appear during the manipulation of superficially divergent terms .
the most laborious aspect of such an evaluation is to extract the explicit dependence of the bare matrix element on the external momentum , by expressing the superficially divergent parts in terms of known primitive divergent integrals .
we will focus on an arbitrary four - index integrand emerging , for example , from a `` diamond''-like diagram with a @xmath218 insertion .
taking into account the symmetries of this object , we can deduce all possible tensor structures which may appear , as a linear combination , in the result for the corresponding integral .
such structures are certainly tensors under the hypercubic group , but not necessarily so under the full @xmath219 euclidean rotation group : terms such as @xmath220 or @xmath221 might be present , @xmath222 ; @xmath223 , otherwise ; @xmath224 . ]
( @xmath134 : external momentum ) , and if so they might spoil the renormalizability and/or the lorentz invariance of the theory .
we must show that in all cases , such terms are absent .
let us begin by taking as an example an algebraic expression which contains both superficial and sub divergencies ; this example serves as a prototype for all the cases we have encountered .
such an expression may arise from a `` diamond''-like diagram with the insertion @xmath225 : @xmath226 where @xmath134 is the external momentum and @xmath227 no summation is implied over the indices @xmath67 , @xmath228 , @xmath229 , @xmath230 . from simple ultraviolet power counting on the term above
, one can realize that the superficial degree of divergence is @xmath231 and the degree of divergence in each of the two loops is @xmath232 and @xmath233 .
thus , this term is not only superficially divergent but also contains a subdivergence in the right loop .
all divergences are resolved by using a bphz procedure .
the potential ir divergences , which may arise in intermediate steps , necessitate working in @xmath135 dimensions as in @xcite . performing a bphz subtraction for the right loop
, we split the integral into two parts : @xmath234 i_{\rm sub}^{\mu\,\nu\,\rho\,\sigma}(q)&\equiv & \bigg[\int \frac{d^d p\,d^d k}{(2 \pi)^{2d } } \frac{\kcirc^{\mu}\,\kcirc^{\sigma}\,\pcirc^{\nu}\,\pcirc^{\rho}}{(\hat{p}^2)^2\ \widehat{p+q}^2\ \widehat{k - p}^2\ \hat{k}^2\ \widehat{k+q}^2 } \nonumber \\ & & \hskip 1.5 cm - \int \frac{d^d p}{(2 \pi)^d}\frac{\pcirc^{\nu}\,\pcirc^{\rho}}{(\hat{p}^2)^2\ \widehat{p+q}^2 } \int \frac{d^d k}{(2\pi)^d}\frac{\kcirc^{\mu}\,\kcirc^{\sigma}}{(\hat{k}^2)^3 } \bigg ] \label{firstsubb}\end{aligned}\ ] ] the last term in eq .
( [ firstsuba ] ) is a separable integral .
the integral over momentum @xmath235 is a standard primitively divergent integral ( see , e.g. , @xcite ) , whose value contains only lorentz invariant structures .
the integral over momentum @xmath236 does not depend on the external momentum @xmath134 , and gives nonzero result only when the indices @xmath67 and @xmath230 are in the same direction .
thus , this term assumes the following structural form : @xmath237 in the remaining part of the original expression , we must still perform an extra subtraction , to cure the superficial divergence : @xmath238 + i_{\rm sub}(0 ) \label{secondsub}\ ] ] according to the bphz procedure , the quantity @xmath239 $ ] is now uv - finite , and thus it equals the corresponding continuum expression . consequently , once again
, only lorentz invariant structures arise : @xmath240 & \longrightarrow & a^\prime\,\frac{q^\mu\,q^\nu\,q^\rho\,q^\sigma}{(q^2)^2 } + \frac{b^\prime}{q^2}\,\left(\delta^{\mu\,\nu}\,q^\rho\,q^\sigma + \delta^{\sigma\,\nu}\,q^\rho\,q^\mu + \delta^{\mu\,\rho}\,q^\nu\,q^\sigma + \delta^{\sigma\,\rho}\,q^\mu\,q^\nu \right ) \nonumber \\ & & \hskip - 1 cm + c^\prime\,\delta^{\mu\,\sigma}\,\frac{q^\nu\,q^\rho}{q^2 } + d^\prime\,\delta^{\nu\,\rho}\,\frac{q^\mu\,q^\sigma}{q^2 } + e^\prime\,\left(\delta^{\mu\,\nu}\,\delta^{\rho\,\sigma } + \delta^{\sigma\,\nu}\,\delta^{\mu\,\rho}\right ) + f^\prime\,\delta^{\mu\,\sigma}\,\delta^{\nu\,\rho } \label{contintegr}\end{aligned}\ ] ] the last part of the integral @xmath241 ( last term in eq .
( [ secondsub ] ) ) equals : @xmath242 this @xmath134-independent integral could give rise to a structural form of the type @xmath220 , which would spoil lorentz invariance ; however , this problem is avoided since the indices @xmath243 in @xmath244 actually originate from the insertion @xmath245 . as a consequence
, only the combination @xmath246 appears in the feynman diagram , and no @xmath220 contribution survives .
thus , we are led to : @xmath247 \longrightarrow a\,\left(\delta^{\mu\,\rho}\,\delta^{\nu\,\sigma } - \delta^{\nu\,\rho}\,\delta^{\mu\,\sigma } \right ) \label{lastpart2}\ ] ] having completed the whole procedure for the integral shown in eq .
( [ fourindxobj ] ) , we conclude that the only functional form that a four - index object ( with the symmetries described above ) can have reads : @xmath248 & + & \frac{b\left(q^2\right)}{q^2}\,\big(\delta^{\mu\,\rho}\,q^\nu\,q^\sigma -\delta^{\nu\,\rho}\,q^\mu\,q^\sigma \big ) \nonumber \\[2.0ex ] & + & \frac{c\left(q^2\right)}{q^2}\,\big(\delta^{\mu\,\sigma}\,q^\nu\,q^\rho -\delta^{\nu\,\sigma}\,q^\mu\,q^\rho\big ) \label{structform4indx}\end{aligned}\ ] ] we emphasize again that , even though the above expression would be obvious in a continuum regularization , it is not so on the lattice , where one could have ended up with terms breaking lorentz invariance . using similar considerations , we can prove that the two- and three - index expressions which appear in our calculation , will take the same structural form as in the continuum ; i.e. , they will be free of lorentz non invariant contributions , which could be present a priori , such as @xmath249 , @xmath250 , etc . once we establish the structural form of the two- , three- and four - index integrals , we must compute the coefficients multiplying each tensor structure , such as the coefficients @xmath251 , @xmath252 , @xmath253 of eq .
( [ structform4indx ] ) .
we illustrate the procedure by taking as an example the following two - index integral : @xmath254 along the same lines of reasoning as above , we conclude that the lattice integral @xmath255 is of the same form as its continuum counterpart : @xmath256 the problem is now reduced to evaluating the coefficients @xmath257 and @xmath258 . upon contracting the integral shown in eq .
( [ twoindexterm2 ] ) with @xmath259 and @xmath260 , we get : @xmath261 where @xmath99 is the number of dimensions ( on the lattice , @xmath262 ) .
once we evaluate the integrals @xmath263 and @xmath264 , we are able to determine the quantities @xmath257 and @xmath258 through the following relations : @xmath265 let us proceed with the evaluation of the integral @xmath263 by contracting @xmath259 with eq .
( [ twoindexterm ] ) , keeping only terms of order @xmath266 . at first
, we aim to reduce the number of propagators appearing in the denominator by employing the following property , which is valid on the lattice : @xmath267 \label{dotprod}\ ] ] omitting some intermediate steps , the resulting expression can be written as follows : @xmath268 where @xmath269 i_b & = & -\frac{1}{2}\sum_\rho \int\frac{d^4k}{(2\pi)^4}\,\frac{\kcirc^\rho}{\hat{k}^2\ \widehat{k+q}^2}\cdot \int\frac{d^4p}{(2\pi)^4}\,\frac{\pcirc^\rho}{(\hat{p}^2)^2\ \widehat{p+q}^2 } \label{ib } \\[2.0ex ] i_c & = & \frac{1}{2 } \int \frac{d^4 p\,d^4 k}{(2 \pi)^8}\,\frac{\kcirc\cdot\pcirc}{\hat{k}^2\ \widehat{k+q}^2\ \hat{p}^2\ \widehat{p+q}^2\ \widehat{k - p}^2 }
\ \ \label{ic}\\[2.0ex ] i_d & = & \int \frac{d^4 p\,d^4 k}{(2\pi)^8}\,\frac{\kcirc\cdot\pcirc}{(\hat{p}^2)^2\ \widehat{p+q}^2\ \widehat{k+q}^2\ \widehat{k - p}^2 } = \frac{1}{2}\int \frac{d^4 p\,d^4 k}{(2\pi)^8 } \,\frac{(k\ , \minuscirc\ , q)\cdot(p\ , \minuscirc\ , q)}{\hat{p}^2\ \big(\widehat{p - q}^2\big)^2\ \hat{k}^2\ \widehat{k - p}^2}\ \ \ \label{id}\end{aligned}\ ] ] as can be seen from the expressions above , we have managed to reduce diamond - like expressions into simpler integrals .
integral @xmath270 is ir convergent : one can set @xmath271 and carry out the integration numerically .
integral @xmath272 , shown in eq .
( [ ib ] ) , is the product of two 1-loop integrals : the first one is well known and tabulated in @xcite , whereas the second , being uv - finite , equals its continuum analogue and can be solved by using the following formula found in ref .
@xcite : @xmath273_{p = q } \label{chetyrkinint}\ ] ] where @xmath274 , @xmath275 , and @xmath276 @xmath277 is the gamma function and @xmath278 is the beta function .
@xmath279 is an arbitrary homogeneous polynomial in @xmath235 : @xmath280 .
all uv - convergent integrals in our calculation can be treated using eq .
( [ chetyrkinint ] ) , for various values of @xmath62 , @xmath281 , @xmath279 ( and also eq .
( 3.4 ) of @xcite for diamond diagrams ) .
integral @xmath282 of eq .
( [ ic ] ) can be treated with one further application of eq .
( [ dotprod ] ) , leading to either ir convergent contributions or tabulated integrals ( such as the scalar `` eye '' integral , eq .
( c.5 ) of @xcite ) . regarding the integral @xmath283 , employing integration by parts ,
we find : @xmath284 simple trigonometry on the first term in eq .
( [ partialintegration ] ) : @xmath285 leads to the scalar eye diagram plus a simple , separable integral .
the second term in eq . ( [ partialintegration ] ) is the derivative of a vector eye diagram , and can be resolved in two steps : ( i ) the integrand , @xmath286 , after simple trigonometry and use of the symmetry @xmath287 , may be expressed in terms of superficially convergent and/or known divergent integrals , plus an integrand of the form @xmath288 ; ( ii ) the latter , upon contraction with @xmath289 and use of eq .
( [ dotprod ] ) , is expressed completely in terms of known integrals .
some superficially convergent integrals , which contain subdivergences , often appear in various stages of our calculation .
a simple prototype example is : @xmath290 which is logarithmically divergent for @xmath291 . in such cases , a subtraction of the form : @xmath292 leads to known separable integrals , plus terms in which one can set @xmath271 without appearance of divergences . in conclusion , using the steps which we outlined above , we have managed to evaluate the integral @xmath263 of eq .
( [ contract1 ] ) , by reducing diamond - like structures into simpler ones , leading to expressions containing uv - finite integrals and standard primitively divergent integrals whose values are known . using similar considerations
, one can also evaluate the integral @xmath264 of eq .
( [ contract2 ] ) , which is needed up to order @xmath293 . with the evaluation of these two integrals
, we can fully determine the coefficients @xmath257 and @xmath258 of eq .
( [ abvalues ] ) , leading to the calculation of the original two - index integral .
let us point out that , throughout the whole procedure , the necessity to work in @xmath294 dimensions does not emerge ( it was only necessary in order to carry out demonstrations , such as eqs .
( [ firstsuba]-[lastpart2 ] ) , leading to the conclusion : eq .
( [ structform4indx ] ) ) .
in this appendix we present our perturbative results for the bare green s functions , @xmath295 ( where @xmath296 ) , on a per diagram basis .
our results are expressed in terms of the bare coupling constant , @xmath297 , the lattice spacing @xmath54 , the external momentum , @xmath134 and the clover parameter @xmath49 .
for the sake of simplicity we have set @xmath298 in two - loop expressions ; at one - loop level the number of colors is left unspecified and the bare gauge parameter , @xmath138 , may take arbitrary values . in all cases ,
the number of flavors , @xmath9 , can take arbitrary values .
only one feynman diagram , shown in fig .
[ zvat1loopdiagrams ] , contributes to one - loop expressions .
our corresponding results for the three operators read : @xmath299 \label{perdiagramv1loop } \end{aligned}\ ] ] @xmath300 \label{perdiagramav1loop}\end{aligned}\ ] ] @xmath301 \label{perdiagramt1loop}\end{aligned}\ ] ] where the index @xmath302 runs over all contributing two - loop diagrams .
the dependence on @xmath49 is polynomial of degree up to 4 ( @xmath306 ) .
the number of flavors , @xmath9 , is raised to the power @xmath236 where , of course , @xmath307 only for diagrams 8 and 9 of fig . [ zvat2loopdiagrams ] , since they are the only diagrams containing a closed fermion loop ; the remaining diagrams have @xmath308 .
the coefficients @xmath309 , @xmath310 and @xmath311 are numerical constants obtained upon evaluating each two - loop feynman diagram . in tables
[ tabva]-[tabvb ] , [ tabava]-[tabavb ] and [ tabta]-[tabtb ] , we present our results for @xmath312 , @xmath310 and @xmath311 , respectively , with accuracy up to 10 decimal places . in the case of the singlet axial - vector operator , it turns out that the only additional diagram contributing to @xmath313 is diagram 4 of fig .
[ singletvat ] .
it is a straight forward exercise to recover the bare matrix element ; starting from @xmath189 ( eq . ( [ scalarsingletri ] ) and @xmath159 ( eq . ( [ zpsi2loopri ] ) , one can employ the @xmath0 renormalization condition , eq .
[ zxrulesb ] , to extract the corresponding matrix element . | we compute the two - loop renormalization functions , in the @xmath0 scheme , of local bilinear quark operators @xmath1 , where @xmath2 corresponds to the vector , axial - vector and tensor dirac operators , in the lattice formulation of qcd .
we consider both the flavor nonsinglet and singlet operators .
we use the clover action for fermions and the wilson action for gluons .
our results are given as a polynomial in @xmath3 , in terms of both the renormalized and bare coupling constant , in the renormalized feynman gauge .
finally , we present our results in the @xmath4 scheme , for easier comparison with calculations in the continuum . the corresponding results , for fermions in an _ arbitrary _ representation , together with some special features of superficially divergent integrals , are included in the appendices . |
kinematic data for galaxy clusters offer a unique possibility to study their mass distribution and the orbits of the galaxies within them . depending on the quality of data sample and
the purpose of the analysis a number of methods of data modelling has been proposed , from the mass estimate based on the virial theorem @xcite , the jeans analysis with the two even velocity moments @xcite and the caustic technique @xcite to the recently presented kinematic deprojection with the mass inversion @xcite and the projected phase - space analysis with a fully anisotropic model of the distribution function @xcite .
there is no doubt that the component which is best constrained with all methods of data modelling is the mass profile .
the analysis of galaxy kinematics in nearby galaxy clusters revealed that the mass distribution is consistent with the universal nfw @xcite profile ( see e.g. * ? ? ? * ; * ? ? ? * ; * ? ? ? * ) , which was confirmed in other studies based on x - ray observations ( see e.g. * ? ? ? * ) or lensing data ( e.g. * ? ? ? * ) . on the other hand ,
constraints on the parameters of the universal density profile of individual clusters from galaxy kinematics are available for a very limited number of objects ( see * ? ? ?
* ; * ? ? ?
one of the main aims of this work is to provide robust estimates of these parameters for a statistically significant sample of nearby ( @xmath1 ) galaxy clusters .
constraining the anisotropy of galaxy orbits from kinematic data is a difficult task .
the main problem arises from the so - called mass - anisotropy degeneracy which occurs in the jeans analysis of the velocity dispersion profile @xcite . in order to break this degeneracy a few solutions have been proposed .
@xcite derived the velocity anisotropy profile by combining the jeans standard analysis with an independent constraint on the mass profile .
this approach was adopted by several authors who used the estimate of the mass profile from the kinematics of ellipticals whose orbits appear to be distributed isotropically @xcite , lensing observations @xcite or x - ray data @xcite .
another method to solve the problem of the mass - anisotropy degeneracy relies on modelling the degree of non - gaussianity in the velocity distribution @xcite .
a way to account for this effect is to use two even velocity moments @xcite .
apart from this , in a few attempts a full model of the distribution function has been used , but in all cases a flat anisotropy profile was assumed ( see e.g. * ? ? ?
* ; * ? ? ?
* ) . nevertheless , this method is still promising .
as shown by @xcite in a series of tests on mock kinematic data , the projected phase - space analysis with a suitable model of the distribution function proposed by @xcite allows to place robust constraints on the spatial variation of the orbital anisotropy within the virial sphere as well as on the total mass profile . in this paper , we follow the method described in @xcite to constrain the orbital anisotropy at the cluster centres and around the virial radius .
most measurements of the orbital anisotropy in galaxy clusters have been obtained under the assumption of a flat profile .
all results are consistent within errors with isotropic orbits ( see e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
several studies were devoted to constraining the spatial variation of the orbital anisotropy .
the analysis of galaxy kinematics of individual clusters points to rather isotropic orbits , although the anisotropy exhibits a significant scatter between clusters ( see e.g. * ? ? ? * ; * ? ? ? * ; * ? ? ?
more robust constraints on the anisotropy profile were obtained in the analysis of the composite clusters resulting from combining a number of kinematic data sets into one .
it was demonstrated that the orbits of early type galaxies tend to be isotropic within the virial sphere , whereas late type galaxies are on predominantly radial orbits at the virial radius @xcite .
this property is referred to as the orbital segregation and is commonly attributed to the orbital evolution from radially dominated orbits for spirals to isotropic ones for ellipticals .
we note , however , that the orbital structure of both morphological types of galaxies can hardly be reconciled with the result of global isotropy .
this problem is also addressed in this paper .
the paper is organized as follows . in section 2
, we describe our cluster sample as well as kinematic data and the method of interloper removal . in section 3 ,
we outline the details of our method of the projected phase - space analysis
. the results of the analysis are presented in section 4 . in the same section ,
we discuss the constraints on the mass - concentration relation and study the mass scaling relations .
section 5 is the devoted to comparison between our constraints on the parameters of the mass profile and those found in the literature and based on other methods . in section 6
, we present two tests of consistency between the data and some assumptions underlying our analysis .
the summary follows in section 7 . in this work
, we adopted a flat @xmath11cdm cosmology with density parameter @xmath12 and the hubble constant @xmath13 km s@xmath14 mpc@xmath14 .
the primary criterion for cluster selection is the number of galaxies with spectroscopic redshifts per cluster .
an optimum choice of a threshold for this number should result from the consideration of two contradictory effects .
on one hand , the quality of statistical reasoning requires the number of redshifts to be as large as possible , but on the other we wish to build a sample with a statistically significant number of clusters which means that we also have to include clusters with a rather low number of spectroscopic redshifts . trying to find a compromise we checked the number of available redshifts per cluster , as provided by nasa / ipac extragalactic database ( ned ) .
we found that the optimal choice for the purpose of our analysis is to consider clusters with at least @xmath15 galaxies within the aperture of @xmath16 mpc and with velocities from the range @xmath17 km s@xmath14 in the rest frame of the cluster .
the radius of the aperture equals to the virial radius of massive galaxy clusters and its role is to truncate all clusters to the approximate size of the virial sphere ( in projection ) .
the velocity range @xmath18 km s@xmath14 is a commonly used velocity cut - off which allows to perform a preliminary separation of cluster members from the interlopers ( galaxies of background or foreground ) . the second criterion for cluster selection concerns their dynamical state .
the reason for considering this property is the fact that the basic assumption underlying our method of dynamical analysis is the state of equilibrium .
it means that our sample should be possibly free of dynamically disturbed clusters , especially major mergers . in order to identify such clusters we studied x - ray images of clusters obtained with asca @xcite , xmm - newton @xcite and rosat @xcite satellites .
we rejected all clusters whose images exhibit signatures of merger activity in the form of clearly visible distortions ( e.g. a3667 ) or the presence of massive subclusters ( e.g. a3558 , the most massive cluster in the shapley concentration ) .
another approach to assess the dynamical state of clusters was to study the general properties of galaxy kinematics .
this method was particularly important in the case of clusters for which x - ray data were not available . following this approach , we excluded all clusters with clearly asymmetric or bimodal velocity distributions which are good signatures of a merger or at least an accidental alignment of some structures .
let us note that the analysis of galaxy kinematics was carried out after the proper removal of interlopers ( see subsection 2.4 ) . as an independent test
, we also checked for the presence of a cool core which appears to correlate with the dynamical state of clusters @xcite and may be regarded as an indicator of how relaxed the cluster is .
we note , however , that the presence of a cool core was a secondary criterion of our selection .
we used this criterion only in cases when the conclusions drawn from the kinematics or the x - ray image seemed to be ambiguous or contradictory .
finally , we also excluded clusters which appear to be accompanied by other clusters at comparable redshifts and relative positions on the sky smaller than @xmath19 mpc which is around twice the expected virial radius for a typical cluster
. the majority of them are binary clusters at the initial state of merging , e.g. a399 and a401 .
the main reason for rejecting these systems is the fact that their velocity diagrams are severely contaminated by galaxies from neighbouring clusters . due to the proximity of these clusters both on the sky and in redshift space ,
velocity diagrams overlap and the selection of cluster members becomes impossible or at least highly ambiguous . in order to find all nearby ( @xmath1 ) galaxy clusters which satisfy selection criteria
described above we searched ned and the recently published , but not yet included in ned , spectroscopic data from wings ( the wide field nearby galaxy cluster survey ) which is a photometric and spectroscopic survey of a few tens of nearby galaxy clusters @xcite .
the final sample resulting from our data mining consists of @xmath0 galaxy clusters , out of which @xmath20 are rich clusters from the abell catalogue @xcite and one ( mkw 4 ) is a poor cluster containing a cd galaxy @xcite .
table [ tab_prop ] provides the list of all clusters as well as their basic observational properties , such as the heliocentric redshift @xmath21 , the number of spectroscopic redshifts within the aperture of @xmath16 mpc and the velocity range @xmath17 km s@xmath14 around the mean @xmath22 and after the removal of interlopers ( see subsection 2.4 ) @xmath23 .
.observational properties of galaxy clusters selected for the analysis .
the table includes the cluster name , heliocentric redshift ( @xmath21 ) , number of galaxy redshifts within the aperture @xmath16 mpc and the velocity range @xmath17 km s@xmath14 ( @xmath22 ) , number of redshifts after interloper removal ( @xmath23 ) , bautz - morgan morphological type ( b - m type ) and the comment on a cool core judged on the basis of : @xmath24@xcite , @xmath25@xcite , @xmath26@xcite , @xmath27@xcite .
superscript @xmath28 marks clusters from the wings survey . [ cols="<,^,>,>,>,<",options="header " , ] we see that there is no trend in the mean value of the residuals .
one may notice a negative bias in the outer asymptotic value of the anisotropy profile for @xmath29 mpc .
this is , however , a consequence of the fact that using a smaller aperture we probe the inner part of the clusters for which the orbits are more isotropic .
we also note that the logarithmic dispersions which describe the scatter of the relative residuals are smaller or comparable to the relative errors associated with the @xmath30 credibility ranges listed in table [ tab_para ] . this property together with the stability of the mean values means that our results are not sensitive to the change of the truncation radius of velocity diagrams .
this excludes the possibility that the data may be subject to a significant incompleteness effect . as a second test of self - consistency between the assumptions of our analysis and the data we show in fig .
[ sigma_test ] the surface number density of galaxies in all clusters combined into one .
the surface density was evaluated in radial bins equally spaced in logarithmic scale , where the universal scaling of radii by @xmath31 was adopted ( @xmath32 is the number of galaxies within the @xmath33 aperture ) .
the points above @xmath34 were corrected for the incompleteness of velocity diagrams which occurs due to the fact that the value of @xmath35 is not universal in the cluster sample . as a line of reference we plotted the projected nfw profile ( black solid line ) . apart from radii above @xmath36 ,
where some deviations between the model and the data occur mostly due to the problem of incompleteness , galaxy density is well approximated by the nfw profile .
our data for small radii do not reveal a signature of a core in galaxy distribution , as advocated by @xcite .
although our analysis concerns the projected densities , we may conclude from fig .
[ sigma_test ] that the assumption of our phase - space density model about a constant mass - to - number density is also satisfied .
this confirms the observational fact that has already been discussed in the literature ( see e.g. * ? ? ?
* ; * ? ? ?
in this paper we presented the analysis of @xmath0 nearby ( @xmath1 ) relaxed galaxy clusters . using the method based on the modelling of velocity diagrams in terms of the projected phase - space density , as described in section 3 , we constrained parameters of the total mass and anisotropy profiles in these clusters .
parameter estimation was carried out in the framework of bayesian data analysis with the use of the mcmc technique .
we found that the concentration parameter is weakly correlated with the morphological type of the clusters as well as with the presence of a cool core .
we interpret this fact as a signature of cluster evolution from the state with a flatter density profile ( low concentration parameter ) to one with a steeper density profile ( higher concentration parameter ) .
our explanation is consistent with the recent results of cosmological simulations which predict steeper density profiles for more dynamically evolved clusters @xcite .
the virial mass of the clusters correlates very well with the velocity dispersion , the x - ray temperature and the x - ray luminosity of the clusters .
the slopes of the mass scaling relations agree with the estimates found in the literature and in the case of the velocity dispersion and x - ray temperature are consistent with the prediction from the virial theorem .
we used constraints on the virial mass and the concentration parameter to test the most up - to - date estimates of the @xmath5 parameter against the observational relation between the virial mass and the concentration parameter . comparing the normalization of the mass - concentration relation for our cluster sample with the calibration from cosmological simulations we found that our data favours a rather high value , @xmath37 . like @xcite
, we found an excellent agreement with @xmath5 from the 1st data release of the wmap satellite .
our model of the projected phase - space density allowed us to constrain the profile of the orbital anisotropy in clusters .
we showed at high significance level that galaxy orbits are isotropic in the cluster centres and radially anisotropic at the virial sphere , with @xmath38 ( @xmath39 ) .
this finding is consistent with the results obtained by @xcite for late type galaxies in nearby galaxy clusters and generalizes a number of estimates of the global anisotropy of galaxy orbits in clusters ( see e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
our constraints on the asymptotic values of the anisotropy parameter implies a growth of the anisotropy from @xmath40 at the cluster centre to @xmath41 at the scale radius .
interestingly , this result agrees fairly well with recent estimates of the anisotropy for dark matter in galaxy clusters @xcite .
the orbital anisotropy at the virial sphere of the low concentration clusters tends to be more radially biased than for the clusters of high concentration parameter .
we suggest that this tendency is a consequence of a change in the galaxy orbits from radially biased at the early stage of cluster evolution to more isotropic for more dynamically evolved clusters .
this scenario is supported by the studies of the orbital anisotropy for different galaxy populations which reveals a similar dichotomy between early and late type galaxies representing more and less relaxed galaxy populations in the clusters ( see * ? ? ?
rw wishes to thank g. mamon , s. gottlber , a. schwope and g. lamer for discussions .
rw is grateful for the hospitality of institut dastrophysique de paris and astrophysikalisches institut potsdam where parts of this work were done .
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 .
this work was partially supported by the polish ministry of science and higher education under grant nn203025333 as well as by the polish - french collaboration program of lea astro - pf .
rw acknowledges support from the start fellowship for young researchers granted by the foundation for polish science .
abell g. o. , 1958 , apjs , 3 , 211 abell g. o. , corwin h. g. , olowin r. p. , 1989 ,
apjs , 70 , 1 adami c. , biviano a. , mazure a. , 1998 , a&a , 331 , 439 adelman - mccarthy j. k. , ageros m. a. , allam s. s. et al . , 2008 , apjs , 175 , 297 an j. h. , evans n. w. , 2006 , apj , 642 , 752 arnaud m. , pointecouteau e. , pratt g. w. , 2005 , a&a , 441 , 893 bautz l. p. , morgan w. w. , 1970 , apj , 162 , 149 benatov l. , rines k. , natarajan p. , kravtsov a. , nagai d. , 2006 , mnras , 370 , 427 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 borgani s. , murante g. , springel v. et al . , 2004 , mnras , 348 , 1078 bhringer h. , voges w. , huchra j. p. et al . , 2000 , apjs , 129 , 435 bhringer h. , schuecker p. , guzzo l. et al . , 2004 , a&a , 425 , 367 bryan g. l. , norman m. l. , 1998 , apj , 495 , 80 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 buote d. a. , gastaldello f. , humphrey p. j. , zappacosta l. , bullock j. s. , brighenti f. , mathews w. g. , 2007 , apj , 664 , 123 burns j. o. , hallman e. j. , gantner b. , motl p. m. , norman m. l. , 2008 , apj , 675 , 1125 carlberg r. g. , yee h. k. c. , ellingson e. , 1997 , apj , 485 , 13 cava a. , bettoni d. , poggianti b. m. et al . , 2009 , a&a , 495 , 707 chen y. , reiprich t. h. , bhringer h. , ikebe y. , zhang y. y. , 2007 , a&a , 466 , 805 colless m. , dalton g. , maddox s. et al.,2001 , mnras , 328 , 1039 comerford j. m. , meneghetti m. , bartelmann m. , schirmer m. , 2006 , apj , 642 , 39 comerford j. m. , natarajan p. , 2007 ,
mnras , 379 , 190 den hartog r. , katgert p. , 1996
, mnras , 279 , 349 diaferio a. , geller m. j. , 1997 . , apj , 481 , 633 dolag k. , borgani s. , murante g. , springel v.,2009 , mnras , 399 , 497 duffy a. r. , schaye j. , kay s. t. , dalla vecchia c. , 2008 , mnras , 390 , l64 duffy a. r. , schaye j. , kay s. t. , vecchia c. d. , battye r. a. , booth c. m. , 2010 , mnras , in press durret f. , lima n. g. b. , forman w. , 2005 , a&a , 432 , 809 ettori s. , de grandi s. , molendi s. , 2002 , a&a , 391 , 841 gastaldello f. , buote d. a. , humphrey p. j. , zappacosta l. , bullock j. s. , brighenti f. , mathews w. g. , 2007 , apj , 669 , 158 gelman a. , carlin j. b. , stern h. s. , rubin d. b. , 2004 , bayesian data analysis . chapman & hall / crc , boca raton .
girardi m. , giuricin g. , mardirossian f. , mezzetti m. , boschin w. , 1998 , apj , 505 , 74 gnedin o. y. , kravtsov a. v. , klypin a. a. , nagai d. , 2004 , apj , 616 , 16 gregory p. c. , 2005 , bayesian logical data analaysis for the physical science hansen s. h. , piffaretti r. , 2007 , a&a , 476 , 37 heisler j. , tremaine s. , bahcall j. n. , 1985 , apj , 298 , 8 horner d.j . , 2001 , phd thesis , univ .
maryland college park host o. , hansen s. h. , piffaretti r. , morandi a. , ettori s. , kay s. t. , valdarnini r. , 2009 , apj , 690 , 358 hudson d. s. , mittal r. , reiprich t. h. , nulsen p. e. j. , andernach h. , sarazin c. l. , 2010 , a&a , in press hwang h. s. , lee m. g. , 2008 , apj , 676 , 218 isobe t. , feigelson e. d. , akritas m. g. , babu g. j. , 1990 , apj , 364 , 104 komatsu e. , dunkley j. , nolta m. r. , bennett c. l. , gold b. , hinshaw g. , jarosik n. , larson d. , 2009 , apjs , 180 , 330 ledlow m. j. , voges w. , owen f. n. , burns j. o. , 2003 , aj , 126 , 2740 lin y .- t . , mohr j. j. , stanford s. a. , 2004 , apj , 610 , 745 lopes p. a. a. , de carvalho r. r. , kohl - moreira j. l. , jones c. , 2009 , mnras , 399 , 2201 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. , 2003 , mnras , 343 , 401 okas e. l. , wojtak r. , gottlber s. , mamon g. a. , prada f. , 2006 , mnras , 367 , 1463 macci a. v. , dutton a. a. , van den bosch f. c. , 2008 , mnras , 391 , 1940 mahdavi a. , geller m. j. , 2004 , apj , 607 , 202 mamon g. a. , bou g. , 2010 , mnras , 401 , 2433 merrifield m. r. , kent s. m. , 1990 , aj , 99 , 1548 merritt d. , 1987 , apj , 313 , 121 morgan w. w. , kayser s. , white r. a. , 1975 , apj , 199 , 545 natarajan p. , kneib j. p. , 1996 , mnras , 283 , 1031 navarro j. f. , frenk c. s. , white s. d. m. , 1997 , apj , 490 , 493 neto a. f. , gao l. , bett p. , 2007 ,
mnras , 381 , 1450 pedersen k. , dahle h. , 2007 , apj , 667 , 26 pointecouteau e. , arnaud m. , pratt g. w. , 2005 , a&a , 435 , 1 popesso p. , biviano a. , bhringer h. , romaniello m. , voges w. , 2005 , a&a , 433 , 431 popesso p. , biviano a. , bhringer h. , romaniello m. , 2007 , a&a , 464 , 451 reiprich t. h. , bhringer h. , 2002 , apj , 567 , 716 rines k. , geller m. j. , kurtz m. j. , diaferio a. , 2003 , aj , 126 , 2152 rines k. , diaferio a. , 2006 , aj , 132 , 1275 sanderson a. j. r. , ponman t. j. , finoguenov a. , lloyd - davies e. j. , markevitch m. , 2003 , mnras , 340 , 989 sanderson a. j. r. , edge a. c. , smith g. p. , 2009 ,
mnras , 398 , 1698 sarazin c. l. , 1988 , x - ray emission from clusters of galaxies .
cambridge university press , cambridge schuecker p. , bhringer h. , reiprich t. h. , feretti l. , 2001 , a&a , 378 , 408 smith r. j. , hudson m. j. , nelan j. e. et al . , 2004 ,
aj , 128 , 1558 snowden s. l. , mushotzky r. f. , kuntz k. d. , davis d. s. , 2008 , a&a , 478 , 615 spergel d. n. , verde l. , peiris h. v. et al . , 2003 , apjs , 148 , 175 spergel d. n. , bean r. , dor o. et al . , 2007 , apjs , 170 , 377 van der marel r. p. , magorrian j. , carlberg r. g. , yee h. k. c. , ellingson e. , 2000 , aj , 119 , 2038 vikhlinin a. , kravtsov a. , forman w. , jones c. , markevitch m. , murray s. s. , van speybroeck l. , 2006 , apj , 640 , 691 white d. a. , 2000 , mnras , 312 , 663 white d. a. , jones c. , forman w. , 1997 , mnras , 292 , 419 wojtak r. , okas e. l. , 2007 , mnras , 377 , 843 wojtak r. , okas e. l. , mamon g. a. , gottlber s. , prada f. , moles m. , 2007 , a&a , 466 , 437 wojtak r. , okas e. l. , mamon g. a. , gottlber s. , klypin a. , hoffman y. , 2008 , mnras , 388 , 815 wojtak r. , okas e. l. , mamon g. a. , gottlber s. , 2009 , mnras , 399 , 812 xu h. , jin g. , wu x .- p . , 2001 , apj , 553 , 78 | we analyze kinematic data of @xmath0 nearby ( @xmath1 ) relaxed galaxy clusters in terms of the projected phase - space density using a phenomenological , fully anisotropic model of the distribution function .
we apply the markov chain monte carlo approach to place constraints on total mass distribution approximated by the universal nfw profile and the profile of the anisotropy of galaxy orbits .
we find the normalization of the mean mass - concentration relation is @xmath2 at the virial mass @xmath3 . by comparison with the calibration from cosmological @xmath4-body simulations it is demonstrated that this result is fully consistent with @xmath5 from wmap1 data release and agrees at @xmath6 level with that from wmap5 . assuming a one - to - one correspondence between @xmath5 and the normalization of the mass - concentration relation in the framework of the concordance model we estimate the normalization of the linear power spectrum to be @xmath7 .
our constraints on the parameters of the mass profile are compared with estimates from x - ray observations and other methods based on galaxy kinematics .
we also study correlations between the virial mass and different mass proxies including the velocity dispersion , the x - ray temperature and the x - ray luminosity .
we demonstrate that the mass scaling relations with the velocity dispersion and the x - ray temperature are fully consistent with the predictions of the virial theorem .
we show that galaxy orbits are isotropic at the cluster centres ( with the mean ratio of the radial - to - tangential velocity dispersions @xmath8 ) and radially anisotropic at the virial sphere ( with the mean ratio @xmath9 ) .
although the value of the central anisotropy appears to be universal , the anisotropy at the virial radius differs between clusters within the range @xmath10 . utilizing the bautz - morgan morphological classification and information on the prominence of a cool core we select two subsamples of galaxy clusters corresponding to less and more advanced evolutionary states .
it is demonstrated that less evolved clusters have shallower mass profiles and their galaxy orbits are more radially biased at the virial sphere .
this property is consistent with the expected evolution of the mass profiles as well as with the observed orbital segregation of late and early type galaxies .
galaxies : clusters : general galaxies : kinematics and dynamics cosmology : dark matter |
the scattering matrix @xmath0 is a useful tool to describe the multiple scattering that occur when planes waves enter to a system which in general is of complex nature , like atomic nucleus , chaotic and/or disordered systems , etc .
@xcite although it is well known for particle waves in quantum mechanics @xcite it can also be applied to any kind of plane waves . in the electromagnetic context
the transfer matrix @xcite is known instead of @xmath0 @xcite but they are equivalent and related .
@xcite by definition , @xmath0 relates the outgoing plane waves amplitudes to the incoming ones to the system , from which the _ reflection _ @xmath2 and _ transmission _
@xmath3 coefficients are obtained ; they are called _ reflectance _ and _ transmitance _ in electromagnetism .
@xcite in the absence of dissipation , as happens in quantum electronic and some electromagnetic systems , @xmath0 becomes a unitary matrix , in particular @xmath4 .
however , if the system contains a dissipative medium , @xmath0 is no longer unitary , in fact it is a sub - unitary matrix and @xmath5 , where the lack of unity is called _ absorbance _ in the electromagentic subject .
@xcite we are concerned with the electromagnetic case . there
, a natural definition of @xmath0 through one of its properties arises from the poynting theorem .
this theorem is a energy balance equation which in the simplest form , i. e. for linear and non dispersive media , is given by @xcite @xmath6 where @xmath7 is the poynting vector which gives the energy flux per unit area per unit time , and @xmath8 is the electromagnetic energy density @xmath9.\ ] ] here , we have assumed that the electric @xmath10 , electric displacement @xmath11 , magnetic induction @xmath12 , and magnetic @xmath13 fields , as well as the current density @xmath14 , are complex vectors whose real part only has physical meaning .
we have also written explicitely the dependence on the position @xmath15 and time @xmath16 .
the term in the right hand side of eq .
( [ ec : cenergia ] ) is the negative of the work done by the fields per unit volume and represents the conversion of electromagnetic energy to thermal ( or mechanical ) energy .
the version of poynting s theorem for dispersive media will not be touched here .
our purpose in this paper is to apply the poynting theorem to the simplest scattering system , an interface between two different media , to illustrate the relation of poynting s theorem and one property of @xmath0 .
first , we will consider the absence of dissipation for the dielectric - dielectric interface for which the defined @xmath0 matrix is unitary .
second , the dissipative case is considered in the dielectric - conductor interface for which the scattering matrix , called @xmath1 , is sub - unitary .
the last system is the simplest example to explain quantitatively a model , that we call the `` parasitic channels '' model , introduced in contemporary physics @xcite to describe complex systems with losses .
@xcite also , @xmath17 could represents the scattering of an `` absorbing patch '' used to describe surface absorption . @xcite the paper is organized as follows . in the next section
we write the time averaged poynting s theorem and calculate the corresponding poynting s vector for planes waves and its flux through an open surface in a dielectric medium , as to be used in the sections that follow sect .
[ sec : poynting ] .
the dielectric - dielectric interface is considered in sect .
[ sec : dielectric ] while sect .
[ sec : conductor ] is devoted to the dielectric - conductor interface . finally , we conclude in sect . [
sec : conclusions ]
in what follows we will consider monocromatic high frequency oscillating fields such that @xmath18 where have written the precise dependence on the spatial and temporal variables to avoid confusion due to the abuse of the notation . here
, the temporal average is of importance . using the definition @xmath19 for the average of a time dependent function @xmath20 over a period @xmath21 , the average of eq .
( [ ec : cenergia ] ) can be written as @xmath22 where we have used that , with help of eqs .
( [ ec : e ] ) and ( [ ec : h ] ) , @xmath23 also , by substitution of eqs . ( [ ec : e ] ) and ( [ ec : h ] ) into eq .
( [ ec : sp ] ) , it is easy to see that @xmath24 with ( we write again the precise dependence ) @xmath25 being the time averaged poynting s vector . in equivalent way , using eqs .
( [ ec : e ] ) and ( [ ec : j ] ) , we get @xmath26.\ ] ] finally , the time averaged energy flux conservation law is given by @xmath27 where only the real part has physical meaning . if we integrate over a volume @xmath28 enclosed by a close surface @xmath29
, it says that the net flux @xmath30 of @xmath31 through @xmath29 is the negative of the work done by the fields if there are dissipative components in @xmath28 .
i. e. , @xmath32 where , from one side , @xmath33 and , for the other side , using the divergence theorem @xmath30 can be written as a surface integral over @xmath29 as @xmath34 for a monocromatic plane wave with linear polarization in the @xmath36 axis , propagating along the positive @xmath37-direction , the spatial component is @xmath38 being the wave number , @xmath39 the index of refraction of the medium and @xmath40 the speed of light in vacuum ; @xmath41 is a unit vector pointing in the positive @xmath36-axis , and the complex number @xmath42 is the amplitude of @xmath43 .
the magnetic field @xmath44 , calculated from @xmath43 using one of the maxwell equations ( faraday s law of induction ) , is @xmath45 where @xmath46 is a unit vector pointing in the positive @xmath47-direction and @xmath48 is the permeability of the free space ( we have assumed a non - magnetic medium ) .
direction traveling in the @xmath37 direction cross the open surface of area @xmath35 .
the long arrows represent the corresponding time averaged poynting s vector . ] sustituting eqs .
( [ ec : eplana ] ) and ( [ ec : hplana ] ) into eq .
( [ ec : sprom2 ] ) we get @xmath49 where @xmath50 is a unit vector pointing in the positive @xmath37-axis .
this equation means that the time averaged energy flux is constant along the propagation .
if we consider an open surface of area @xmath35 as in fig .
[ fig : flux ] , the flux of @xmath51 through it is ( see eq .
( [ eq : flux ] ) ) @xmath52 where we have taken the normal unit vector of @xmath35 as @xmath53 .
let us consider an interface between two dielectrics , with refractive indices @xmath39 and @xmath54 , in the @xmath55 plane as shown in fig . [ fig : energia ] .
for the shake of simplicity , we consider plane waves with linear polarization in the @xmath36 axis with normal incidence on both sides of the interface . therefore ,
the spatial part of the electric fields , @xmath43 on the left and @xmath56 on the right , are of the form @xmath57 where @xmath58 and @xmath59 ( @xmath60 and @xmath61 ) are the complex amplitudes of the incoming ( outgoing ) planes waves . and @xmath54 . the amplitudes for the incoming ( @xmath58 and @xmath59 ) and outgoing ( @xmath60 and @xmath61 ) planes waves are shown with the corresponding poynting s vectors represented by long arrows .
short arrows denote the unit vectors @xmath62 and @xmath63 of the covers of the cylindrical surface @xmath29 .
the cross section of the cylinder is @xmath35 . ] for this case , the right hand side of eqs .
( [ ec : cenergia2 ] ) and ( [ eq : balance ] ) is zero ( @xmath64 ) because there is not dissipation .
hence , @xmath65 . for convenience ,
we take a cylinder of cross section @xmath35 , including the covers , to be the closed surface @xmath29 ( see fig . [
fig : energia ] ) . due to normal incidence
only the flux through the covers of section @xmath35 contribute to @xmath30 .
then , eq . ( [ eq : balance ] ) gives @xmath66 where @xmath67 ( @xmath68 ) is the flux through one of the covers with normal unit vector @xmath69 ( @xmath70 , 2 ) due to the @xmath71-th plane wave ; it is given by @xmath72 with the convention that @xmath69 points outwards of @xmath29 , @xmath73 and @xmath74 are positive quantities while @xmath75 and @xmath76 are implicitely negative .
using the result of eq .
( [ eq : fluxplane ] ) for a plane wave , eq .
( [ ec : cflujo ] ) can be written as @xmath77 which can be arranged in a matricial form that we will use later , namely @xmath78 by definition the fresnel coefficients relate the outgoing to the incoming plane waves amplitudes as @xcite @xmath79 where @xmath80 is the @xmath81 matrix @xmath82 the fresnel coefficients of reflection @xmath83 , @xmath84 and transmission @xmath85 , @xmath86 are @xcite @xmath87 although for this particular case they are real numbers , in general they are complex , in which case @xmath80 is a complex matrix . in order to be more general , in what follows
we will assume that @xmath80 is complex .
by substitution of eq .
( [ ec : fresnel ] ) , eq .
( [ ec : cflujo4 ] ) can be written as @xmath88 from which we see that @xmath89 or , equivalently , @xmath90 where @xmath91 is the @xmath81 identity matrix and @xmath92 @xmath0 is known in the literature as the _ scattering matrix _ which has the following general structure @xcite @xmath93 here , @xmath94 and @xmath16 ( @xmath95 and @xmath96 ) are the reflection and transmission amplitudes for incidence on the left ( right ) . from eqs .
( [ ec : fresnel2 ] ) , ( [ ec : cfresnel1 ] ) , ( [ ec : cfresnel2 ] ) and ( [ ec : sdef1 ] ) it is easy to see that @xmath97 two remarks are worthy of mention .
@xcite the first one is that @xmath98 because the system described by @xmath0 is invariant under inversion of time , but in a more general context that is not necessary the case .
@xcite the second is that @xmath99 because the optical path is not the same for the reflected trajectory when incidence is from right or left . of course , eq .
( [ ec : sflujo ] ) is satisfied in the case we are considering as can be easily checked using eqs .
( [ eq : rrp ] ) and ( [ eq : ttp ] ) . in particular , @xmath100 where @xmath101 and @xmath102 are the reflection and transmission coefficients . eqs .
( [ ec : fresnel ] ) and ( [ ec : sdef1 ] ) implies that by definition @xmath0 relates the outgoing to incoming plane waves amplitudes but normalized with the index of refraction .
i. e. , @xmath103 where @xmath104 ( @xmath105 ) , and the electric field on both sides is given by @xmath106 we recall that eq .
( [ ec : sdef3 ] ) is a definition of @xmath0 that has arised in a natural way from the restriction imposed by flux conservation , eq .
( [ ec : sflujo ] ) , and its structure , given by eq .
( [ ec : sdef2 ] ) , reflects the symmetries present in the problem , one of them being the _ time reversal invariance_.
when one of the two media is a conductor with electric conductivity @xmath107 , the one on the right in fig .
[ fig : energia2 ] let say , the treatment is equivalent as in sect .
[ sec : dielectric ] but with a complex index of refraction : @xmath108 , where @xmath54 and @xmath109 are the optical constants .
@xcite also , the corresponding wave number becomes complex : we replace @xmath110 with @xmath111 then , on the conductor side the electric field is an evanescent wave while on the dielectric side we have incoming and outgoing plane waves .
they are @xmath112 from eq .
( [ ec : fresnel ] ) , or eq . ( [ ec : sdef3 ] ) , with @xmath113 we see that @xmath114 where @xmath94 and @xmath16 are obtained from eqs .
( [ eq : rrp ] ) and ( [ eq : ttp ] ) but replacing @xmath54 by @xmath115 , namely @xmath116 .
the refraction index , as well as the wave vector , becomes complex such that only evanescent waves are present with an amplitude @xmath61 decaying exponentially as @xmath37 increases . ] in this case , the scattering matrix of the system is @xmath117 and we denote it by @xmath1 . by definition ( it is not necessary to normalize with respect to the index @xmath39 ) @xmath118 comparing with eq .
( [ eq : rf1x1 ] ) we see that @xmath119 where , from eqs .
( [ eq : rrpcomplex ] ) , @xmath120 as for @xmath0 , the poynting theorem impose a restriction to @xmath17 . to apply eq .
( [ eq : balance ] ) we consider that the closed surface @xmath29 in fig .
[ fig : energia2 ] extends to infinity on the right of the interface , such that eq .
( [ eq : w ] ) gives @xmath121 where we used that @xmath122 .
then , eq . ( [ eq : balance ] ) can be written as @xmath123 where the area @xmath35 has been cancelled . using eqs .
( [ eq : tf1x1 ] ) and ( [ eq : tildes ] ) , after an arrangement , eq .
( [ eq : psub1 ] ) gives latexmath:[\[\label{eq : psub2 } @xmath117 subunitary matrix with @xmath125 the lack of unitarity ( _ strength of absorption _ or absorbance ) , @xmath126 the last equality is valid for metals in the upper infrared part of the spectrum and for metals at microwave and lower frequencies .
@xcite of course @xmath127 as can be easily verified using eqs .
( [ eq : rtheta ] ) .
( [ eq : psub2 ] ) can be seen as resulting of the unitarity condition for an @xmath0-matrix that satisfy flux conservation ( compare with eq .
( [ eq : unit1 ] ) ) .
this @xmath0-matrix should has the structure ( see eq .
( [ ec : sdef2 ] ) ) @xmath128,\ ] ] where the phase @xmath129 can be taken as the phase of @xmath130 ( see eqs .
( [ eq : tf1x1 ] ) and ( [ eq : tpdef ] ) ) ; using eq .
( [ eq : ttpcomplex ] ) @xmath131 this form of @xmath0 says that the losses because of the conductor can be interpreted as due to a single mode of absorption whose `` coupling '' to the interface is @xmath125 .
this is the simplest version of what we call the `` parasitic channels '' model @xcite appeared in the literature of contemporary physics in recent years to describe power losses in more complex systems ( see ref . and references there in ) .
that model consists in simulate losses with @xmath132 equivalent absorbing modes , each one having an imperfect coupling @xmath125 to the system .
the total absorption is quantified by @xmath133 but @xmath132 and @xmath125 can not be determined separately .
our result not only explain this abstract model but quantify in an exact way the coupling @xmath125 of each absorbing mode as well the scattering matrix of a single absorbing patch in the surface absoprtion model .
@xcite our treatment presented here can also be used to construct artificially a system with multiple absorbing modes . @xcite
we reduced the time averaged poynting theorem to a property of the scattering matrix . for
that we applied this balance equation to a simplest scattering system , consisting of normally incident planes waves at an interface between two media .
the simplest version of this theorem was used such that dispersive media and the corresponding dissipation were ignored .
two kind of interfaces were considered . in the first one , a dielectric - dielectric interface
, the energy flux conservation leads to a natural definition of the scattering matrix @xmath0 which is restricted to be a unitary matrix .
we recalled that the structure of @xmath0 reflects the symmetries present on the problem in other contexts . in the second one , the dielectric - conductor interface ,
the definiton of @xmath0-matrix was used to describe the scattering taking into account the losses on the conductor side via the poynting s theorem .
this allowed us introduce the parasitic channels model used in contemporary physics to describe the scattering with losses in more complex systems .
we should were able to quantify the coupling of the single parasitic mode of absorption in our simple system .
a system with multiple absorbing modes can be constructed and the results will be published elsewhere . finally , the same treatement can be generalized for oblique incidence . | we apply the poynting theorem to the scattering of monochromatic electromagnetic planes waves with normal incidence to the interface of two different media .
we write this energy conservation theorem to introduce a natural definition of the scattering matrix @xmath0 . for the dielectric - dielectric interface
the balance equation lead us to the energy flux conservation which express one of the properties of @xmath0 : it is a unitary matrix . for the dielectric - conductor interface the scattering matrix , that we denote by @xmath1 , is no longer unitary due to the presence of losses at the conductor .
however , the dissipative term appearing in the poynting theorem can be interpreted as a single absorbing mode at the conductor such that a whole @xmath0 , satisfying flux conservation and containing @xmath1 and this absorbing mode , can be defined .
this is a simplest version of a model introduced in the current literature to describe losses in more complex systems . |
it is imperative to know the frequency response of an optical device ( e.g. a cavity or high - speed modulator ) for it to be useful in optics or communication applications .
for example , the frequency spacing between resonances ( the free spectral range , fsr ) of a cavity must be known before being used to calibrate a laser s frequency @xcite , perform absolute length measurements @xcite , generate non - classical light @xcite , or produce large cluster states using frequency - entangled photons @xcite .
dense wavelength division multiplexed systems often use fabry - perot etalons to select and stabilize the wavelength of a tunable diode laser @xcite .
a precise measurement of the etalon s fsr is necessary for matching its transmission channel with the international telecommunication union grid .
similarly , it is necessary to calibrate a high - speed amplitude modulator before use in long - haul optical fiber transmission systems @xcite , or as a wideband optical signal source for characterizing the frequency response of photo - detectors or optical fibers @xcite .
several methods exist for measuring the frequency response of an optical cavity or modulator .
the most - straightforward method is direct detection with a calibrated fast receiver .
however , this technique is highly dependent on the bandwidth of the optical receiver , and the ability to calibrate it .
most characterization techniques involve a combination of either a calibrated wide bandwidth optical receiver , a commercial optical spectrum analyzer or a calibrated high - speed modulator .
for example , techniques based on frequency modulation are effective in determining a cavity s fsr @xcite .
however , these methods require both a calibrated modulator and a fast photo - detector with a known frequency response .
several techniques exist that can determine the response of a modulator that do not require a fast photo - detector , such as optical heterodyning @xcite or swept - frequency techniques @xcite .
unfortunately , these methods still rely on fast electronics or specialized laser sources that only operate at particular wavelengths . as optical device bandwidths continue to increase
, it is necessary to develop a measurement technique that determines their frequency response without depending on the precise calibration of a high - speed photo - detector , modulator , or optical cavity . in this paper
, we use a recursive method to obtain the frequency response of an optical system .
this system consists of both an optical cavity and a high - speed ( uncalibrated ) amplitude modulator , which can be simultaneously characterized from the same measurement configuration .
our approach extends a technique demonstrated by locke _
@xcite , and involves a fiber - coupled amplitude ( intensity ) modulator , and a low - frequency ( dc ) photo - detector , both of which are commercial laboratory equipment . the photo - detector is used to measure the dc cavity transmitted or reflected power as a function of the modulation frequency , while the carrier remains frequency locked to a resonance . instead of characterizing different cavity resonances by shifting the center laser wavelength @xcite
, we characterize the full bandwidth up to 15.5@xmath0ghz from the optical carrier ( 28 resonances ) by locking the center wavelength , and only adjusting the modulation frequency . by measuring the entire cavity response ( both on and off - resonance ) over this frequency range , we are able to estimate the frequency - dependent modulation depth , and use this knowledge to more precisely determine the cavity characteristics ( fsr and linewidth ) . in addition , we explore the precision of our technique by quantifying the effect modulation harmonics have on the measured linewidth , and determining the different resonant frequencies of orthogonal polarization cavity modes caused by the presence of a birefringent material ( non - linear crystal ) . once the modulator has been calibrated , this measurement technique can be used to characterize the frequency response of any passive optical element , and is not limited to cavities .
in the next three subsections , we will present our proposed method , theories related to characterizing optical cavities , and how to estimate the frequency response of an optical cavity and amplitude modulator . in this section
, we will briefly explain our proposed method , which is shown in fig .
[ fig : theorysetup ] .
the setup consists of a laser , a high - frequency amplitude modulator , an optical element to be measured , and low - frequency ( dc ) photo - detectors .
we first use a cavity as the optical element under investigation as we can use the cavity s frequency response to calibrate the modulator s response .
once the modulator has been calibrated , our method can be applied to characterize any optical element to a fine resolution ( e.g. , kilohertz ) over tens of gigahertz .
first we modulate the laser beam with an amplitude modulator . the electric field of the output of the amplitude modulator , @xmath1 ,
can be written as @xmath2 } ) , \label{eq : em } \end{aligned}\ ] ] where @xmath3 is the amplitude of the input to the amplitude modulator , @xmath4 is the laser carrier angular frequency , @xmath5 and @xmath6 are phase shifts caused by dc and ac voltages applied to the amplitude modulator , respectively .
note that here we assume that the amplitude modulator consists of a mach zehnder interferometer with a phase modulator in one of two paths , but our derivation will also be valid for other amplitude modulator configurations @xcite .
the ac signal applied to the modulator is a sine wave with an angular frequency @xmath7 , as shown in fig .
[ fig : theorysetup ] . the phase shift caused by this ac signal
is given as @xmath8 , where @xmath9 is the modulation depth .
the modulation depth is usually frequency dependent , as we will later show is the case in our experiment .
the electric field @xmath1 is now written as @xmath10 where @xmath11 are @xmath12-order bessel functions of the first kind .
the modulated beam is then injected into the optical element under investigation , and the output ( either transmission or reflection ) is measured with a low - frequency ( dc ) photo - detector , as depicted in fig .
[ fig : theorysetup ] . in order to calculate the output signal
, we apply a fourier transform to the modulator output @xmath1 , which gives @xmath13 after passing through the optical element , whose transmission / reflection coefficient ( transfer function ) is given as @xmath14 , the output electric field is written as @xmath15 note that we ignored the pick - off beam splitter in fig .
[ fig : theorysetup ] as it only changes the power scaling of the detected signal . by applying the inverse fourier transform ,
we obtain @xmath16 the output electric field is measured by a dc photo - detector whose cut - off frequency is much lower than the modulation frequency , @xmath7 .
the detector output is given by @xmath17 where @xmath18 is the dc gain of the detector , @xmath19 , and we let @xmath20 and @xmath21 . this is the general expression of our method , where the modulation depth , @xmath9 , and the coefficient @xmath14 are frequency dependent .
we also note that the laser power ( @xmath22 ) and the dc offset ( @xmath5 ) may fluctuate during experiments .
in particular , internal heating due to conductive losses in the modulator may change the dc offset .
since this loss is frequency dependent , the dc offset can drift with changing modulation frequency .
our procedure makes it possible to precisely determine both @xmath9 and @xmath14 in the presence of these fluctuations . before detailing our procedure
, we will briefly present a mathematical description of an optical cavity in the next section . in this section
, we will briefly explain the theories related to the characterization of optical cavities .
firstly , the transmission @xmath23 and reflection @xmath24 of ideal cavities are written as @xcite @xmath25 where @xmath26 and @xmath27 are intensity reflection coefficients for the input and output mirrors , respectively , and @xmath28 is the intra - cavity loss for a round trip ( @xmath29 is the loss for a half trip , @xmath30 ) .
@xmath31 is the phase shift for a round trip of the cavity , which is given as @xmath32 , where @xmath33 is a speed of light , and @xmath34 is the effective length of the cavity .
we assume the laser carrier is locked to the cavity ( i.e. , @xmath35 ) .
the cavity is on resonance when @xmath36 , where @xmath37 is the resonant angular frequency , which satisfies @xmath38 , and @xmath39 is an integer . near a cavity resonance ( i.e. , @xmath40 )
, @xmath41 and @xmath42 can be approximated to a lorentzian function as @xmath43^{1/4 } } f_{fsr } , \\
f_{fsr } = & \ : \frac{c}{l } , \end{aligned}\ ] ] where @xmath44 and @xmath45 are the ( constant ) intensity transmission and reflection coefficients at the resonance , respectively , @xmath46 is the free spectral range ( fsr ) , and @xmath47 is the linewidth ( full width at half maximum , fwhm ) .
figure [ fig : theoryplot ] shows the theoretical detector outputs for a cavity under investigation using our method , which is derived by substituting eq . [ eq : tandr ] ( or eq . [ eq : tandr2 ] ) into eq .
[ eqn : idc ] . in this case , the transmission @xmath23 and reflection @xmath24 are symmetric , that is , @xmath48 and @xmath49 .
we chose similar model parameters to our experimental values .
the intensity reflection coefficients for the input and output mirrors are @xmath50 and @xmath51 , respectively .
the total intra - cavity loss is @xmath52 , and the effective path length is @xmath53 m . we also assume that the modulation depth , @xmath9 , is frequency dependent as @xmath54 , and that the laser power and dc offset of the modulator are constant for simplicity .
the input optical power to the amplitude modulator is distributed between the power at the carrier frequency and at the sideband frequencies .
if the input power is constant during the measurement , then the sum of the optical powers at these frequencies will also be constant .
as a result of the modulator s frequency response , if the modulation signal strength is held constant while the frequency is swept , the optical power in the sidebands will decrease for higher modulation frequencies .
thus the optical power at the carrier frequency will increase .
the overall upward ( downward ) trend of the off - resonance data in the transmitted ( reflected ) data , located inbetween the resonances , is due to the increasing power at the carrier frequency as a function of the modulator s frequency response .
also , evidence of the second harmonic modulation sidebands from the amplitude modulator coupling into the cavity is visible as smaller peaks located halfway between the main resonances . in the next section , we will present the mathematical description of our measurement procedure . in this section
, we will derive how to estimate the frequency response of an optical cavity and the frequency - dependent modulation depth , @xmath9 , of a high - speed amplitude modulator .
first we scan the modulation frequency over tens of gigahertz to determine the responses of the cavity and modulator . during the frequency scan ,
the laser power may fluctuate , as well as the dc offset of the waveguide modulator may change due to internal mechanisms ( e.g. internal heat effects via pyroelectric effect @xcite , or changing environmental conditions @xcite ) .
we assume that the dc offset is set to @xmath55 at the start of the scan ( i.e. , at low frequency ) , and that the modulator has no significant acoustic resonances over the measurement bandwidth .
we also assume that the modulation depth is approximately constant over the cavity resonance ( which is @xmath56mhz ) . under these assumptions
, we can first estimate the resonant frequencies and linewidth of cavity from the data , and then determine the modulator response by using the cavity as a reference .
finally , we determine the cavity characteristics more precisely using the information about the modulator response .
first we derive an expression for the modulator s output , which is monitored using a dc photo - detector , as depicted in fig .
[ fig : theorysetup ] .
the output of the modulator is derived by letting @xmath57 in eq .
[ eqn : idc ] , which gives @xmath58 where @xmath59 is proportional to the incident laser power , @xmath22 ( i.e. @xmath60 ) .
next we will look at the transmission or reflection of a cavity . in order to eliminate the effects from laser power fluctuations
, we normalize eq .
[ eqn : idc ] with eq .
[ eqn : idc0 ] , and define @xmath61 , which is constant .
the normalized dc detector output becomes @xmath62 once we measure @xmath63 over a wide frequency range , we can determine the fsr of the cavity from the resonance peaks visible in fig . [
fig : theoryplot ] . to determine the linewidth
, we may assume that @xmath9 and @xmath5 are approximately constant over the linewidth , which gives @xmath64 where @xmath65 and @xmath66 are constants .
we can use the derived expressions of @xmath23 and @xmath24 to approximately fit the data and obtain the linewidth .
if the modulation depth , @xmath9 , is unknown at this point , we may use @xmath9 as a fitting parameter , and neglect @xmath67 to realize a reasonable fit .
next we will extract @xmath9 over the measured frequency range , and then revisit the fitting to improve our precision in determining the fsr and linewidth .
we can extract the on - resonance and off - resonance data from @xmath63 to determine the modulation depth . if the modulation sidebands are on resonance with the cavity , then @xmath68 .
the normalized on - resonance detector output is @xmath69 this expression indicates that the on - resonance detector response is constant regardless of the modulation depth , @xmath9 , and the dc offset , @xmath5 .
we can estimate @xmath70 from the multiple resonance peaks visible in the cavity s frequency response , as illustrated in fig . [
fig : theoryplot ] . note that @xmath70 is constant over the entire measurement bandwidth . if the sidebands are off resonance , and we measure the transmitted power , then @xmath71 and we find , @xmath72 this off - resonance data can be obtained from @xmath63 by eliminating data around the resonant frequencies , and then interpolating . by interpolating the data ( assuming the modulation depth and the dc offset are approximately constant within the resonance linewidth ) , we can obtain continuous off - resonance data . combining eq .
[ eqn : idcon ] and [ eqn : idcoff ] , we get @xmath73 since we can assume that @xmath74 at @xmath75 , the dc detector output of the modulator s response at @xmath75 is @xmath76 . therefore , the normalized dc detector output of the modulator s response without the presence of a cavity is @xmath77 combining eq .
[ eqn : idcratio ] and [ eqn : idcmod ] , we obtain @xmath78 using this expression , we can determine the optical modulator response , @xmath9 , over the entire measurement bandwidth .
once we obtain @xmath9 , we can revisit our derivation of the linewidth and fsr frequency to more precisely determine these values . from eq .
[ eq : t2 ] , [ eq : r2 ] and [ eqn : idcprime ] , we can derive an expression for the lineshape of a cavity resonance as @xmath79 note here we use @xmath7 rather than @xmath80 to clarify the modulation frequency .
this function takes the form of a constant plus two lorentzian functions .
the first function , scaled by @xmath81 , represents the first - order modulation sidebands being on resonance with the cavity , and has a bandwidth equal to the linewidth , @xmath47 . while the second function , scaled by @xmath82 , represents the second harmonic sidebands being on resonance , and has a bandwidth of @xmath83 .
since the modulation depth is not expected to be strong enough to excite the third harmonic , we can ignore modulation sidebands of order @xmath84 .
therefore , our technique can characterize an optical system that consists of a cavity and a modulator with a single measurement configuration .
the on and off resonance data from the cavity measurements is used to learn about the frequency - dependent modulation depth , @xmath9 .
then we use @xmath9 to fit the lorentzian lineshape described by eq .
[ eqn : lineshape ] to the on - resonance data to give a more precise fsr and linewidth .
a clear advantage of this method is that we only require a dc photo - detector .
we do not need an expensive or elaborate high - frequency detector , or to calibrate the frequency response of the detector or modulator beforehand . in the next section
, we will present experimental results from applying this technique to characterize an optical system , consisting of a waveguide modulator and an optical cavity with a non - linear crystal .
our frequency response measurement technique involves a recursive approach that allows us to simultaneously characterize both the optical element under investigation ( cavity ) and the measurement device ( amplitude modulator ) .
thus both the cavity and modulator are characterized from a single setup , and the technique is not restricted by , or dependent on , the frequency response of the photo - detector .
we determine the frequency response of our optical system using the experimental set - up shown in fig .
[ fig : exptdiagram ] .
the output from a 1550 nm fiber laser is first sent through a phase modulator , followed by an amplitude modulator .
both devices are fiber - coupled broadband low - loss lithium niobate electro - optical modulators ( eospace ) .
the amplitude modulation frequency , @xmath7 , is varied while the light before and after the cavity is monitored using low - frequency ( dc ) photo - detectors .
the cavity has a birefringent non - linear crystal inside , which has different refractive indices for horizontally and vertically - polarized light .
this results in non - degenerate polarization modes that have distinguishable resonant frequencies .
since we use a polarization beam splitter to send linearly - polarized light to the cavity , we can measure either horizontal or vertical polarization modes by simply rotating a half - wave plate before the cavity .
the incident power of the laser is quite low ( a few milliwatts ) to avoid exciting second harmonic generation from the non - linear crystal .
the cavity is locked on resonance with the laser optical carrier frequency during the frequency measurements using a phase modulator and the pound - drever - hall locking technique @xcite .
it must remain locked on resonance while the amplitude modulation frequency is swept to ensure that the cavity is not disturbed by any environmental noise while the data is captured .
the cavity s response to the amplitude modulation as a function of cavity path length is illustrated in fig .
[ fig : exptdiagram ] .
this graph depicts the energy ratio between the sidebands and carrier during our measurements .
we can control this energy ratio in two ways : the ac modulation signal strength , and the dc voltage applied to a mach - zehnder interferometer inside the amplitude modulator .
we operate the modulator near the dc quadrature point ( @xmath74 ) to ensure a linear response .
as mentioned in the theory section , we found that the dc offset changes during the frequency sweep .
our calibration method allows us to eliminate the effect of the drifting dc offset .
one limitation of this technique is it will not work if both the measurement setup ( amplitude modulator ) and the optical element ( cavity ) are completely unknown .
we have to assume that the modulator s response is relatively smooth with no acoustic resonances around the cavity resonant frequencies , and that the variation in modulation depth over the linewidth of the cavity is small .
fortunately , the s21 response data provided by the manufacturer with the modulator can be used to determine whether these assumptions are valid ( they are valid for our modulator ) .
we also need to assume that the cavity has a lorentzian - type response with periodic resonances .
this can be easily determined by using the dc photo - detectors to monitor the cavity s transmitted or reflected light as a function of cavity path length , as illustrated in fig .
[ fig : exptdiagram ] . once the modulator has been characterize using the cavity s response , this technique can be used to determine the frequency response of any passive optical element including , but not limited to , cavities .
we will present the results from our frequency response measurements in the next section .
in the next three subsections , we will present wide bandwidth measurements of the frequency response of our cavity and amplitude modulator , and how we use this information to more precisely determine the cavity s fsr and linewidth .
we begin our method by capturing the cavity s transmitted and reflected light with dc photo - detectors as a function of the amplitude modulation frequency over a wide frequency range ( @xmath85ghz ) . based on the measured wide bandwidth response of the cavity shown in fig .
[ fig : refltx16ghz ] , we can estimate the cavity s resonant frequencies and linewidth .
this data is the normalized dc detector output , which corresponds to @xmath63 in eq .
[ eqn : idcnorm ] . the input polarization to the cavity was set to horizontally - polarized light for this wide bandwidth measurement , and the cavity was locked to the carrier frequency of this polarization .
note how the measured transmission increases to a maximum as the modulation frequency approaches the cavity s first resonance at @xmath86mhz , while the reflected intensity decreases to a minimum .
the modulation frequency at the center of each peak ( or dip ) is equal to a multiple of the cavity fsr , while the full width at half maximum of the on - resonance data gives the linewidth .
thus from this data , we estimate the cavity s fsr to be @xmath86mhz and the linewidth to be @xmath56mhz .
the overall upward ( downward ) trend of the off - resonance data in the transmitted ( reflected ) data is in good agreement with the theoretical prediction shown in fig .
[ fig : theoryplot ] .
since we are using dc photo - detectors , the bandwidth of this measurement is not dependent on , or limited by , the photo - detector s response . despite the modulator
having a @xmath56ghz bandwidth , we can clearly distinguish resonances at modulation frequencies up to @xmath87ghz .
we normalize the cavity data with the modulator s output measured before the cavity to correct for any laser power drifts .
we also note that the on - resonance detector levels are approximately constant , as predicted by eq .
[ eqn : idcon ] . in the next section
, we will discuss how the modulator s response can be extracted from this cavity measurement .
once the cavity characteristics have been estimated , we can extract the modulator s response using the on and off - resonance data in the cavity s transmission response .
the on - resonance data , corresponding to @xmath70 , is shown as a relatively constant level in fig .
[ fig : refltx16ghz]b given by the 28 distinguishable peaks .
the data inbetween these peaks corresponds to just the carrier being on resonance with the cavity while the sidebands are off resonance .
since the fsr frequency and cavity linewidth have been estimated , the on - resonance data can be removed from the transmission response to reveal just the off - resonance data , corresponding to @xmath88 .
we interpolate the off - resonance data with a high - order polynomial to extract the modulation depth for the entire frequency range , including at the on - resonance frequencies .
the modulator s frequency - dependent modulation depth , @xmath9 , calculated from the on and off resonance cavity data is shown in fig .
[ fig : beta ] .
note that @xmath9 varies from 1.6 to almost 0 over the @xmath85ghz frequency range .
a @xmath9 of 1.6 corresponds to a measurable portion of optical power present in the second harmonic modulation sidebands .
evidence of the second harmonic coupling into the cavity is visible in fig .
[ fig : refltx16ghz]a and [ fig : refltx16ghz]b as smaller resonances located halfway between the main resonances . in the next section
, we will explore how the presence of second harmonic modulation sidebands in the modulator s output affects the measured cavity linewidth , and how knowing @xmath9 allows us to correct for this effect .
we use the measured modulation s response to more precisely determine the cavity s characteristics by fitting a lorentzian function to the on - resonance data . in order to improve the accuracy of the fitting
, we used an average of five data sets taken around each resonance , and normalized to the modulator s transmitted power .
the amplitude modulation settings and dc bias voltage are kept constant while the frequency is swept in this narrow range .
a least squares algorithm was used to fit the lineshape function ( eq . [ eqn : lineshape ] ) to the on - resonance data to determine the fsr and linewidth .
the resulting linewidths of the first 26 cavity resonances are shown in fig .
[ fig : fwhm_refl ] , which correspond to the first 26 resonances visible in fig .
[ fig : refltx16ghz ] . the linewidth results from the reflected and transmitted cavity data are quite similar
so only the reflected data results are shown for clarity . the data in fig .
[ fig : fwhm_refl ] shown in black are from fitting the on - resonance data with just a single lorentzian function ( first term in eq .
[ eqn : lineshape ] ) , which only models the first - order sidebands .
the data shown in red corresponds to fitting with the complete lineshape function . fitting the on - resonance data with these two methods illustrates the importance of including the second harmonic in modelling the lineshape . since our technique measures the dc power , if modulation harmonics are present , then the detector output will consist of the sum of responses from all the sidebands .
if harmonics are present but not accounted for , then the fitted linewidth will be narrower compared to the intrinsic cavity linewidth .
we have quantified this effect from the frequency - dependent modulation depth , and determined the true cavity linewidth to be ( @xmath89mhz ( fwhm ) averaged over the 26 resonances .
the error bars are calculated by individually fitting each of the five data sets , and then taking the standard deviation of the resulting five linewidths .
the two fitting methods agree within error bars for frequencies above 9@xmath0ghz ( 19th fsr ) where the first - order sidebands clearly dominate .
thus , assuming the variation in @xmath9 over the cavity s linewidth is small , knowing the frequency - dependent modulator response allows us to precisely determine the characteristics of our cavity .
the presence of second harmonic modulation sidebands does not corrupt the fsr measurement as all sidebands are simultaneously on resonance with the cavity .
the precision and repeatability of our technique is demonstrated by our ability to measure the different resonant frequencies of orthogonal polarization modes caused by the presence of a non - linear crystal ( @xmath90 mgo - doped linb@xmath91 ) inside the cavity .
the crystal has different refractive indexes along the ordinary and extra - ordinary axes , which causes horizontally - polarized light to experience a slightly different cavity path length compared to vertically - polarized light .
we used a half - wave plate located before the cavity to rotate the polarization of the incident field , and characterize the distinguishable cavity modes .
the two polarization modes are non - degenerate , as shown in fig .
[ fig : hvfsrs ] , so we can lock the cavity to the carrier frequency of either polarization .
measurements of the reflected light captured around the first resonance for horizontally and vertically - polarized light are shown in fig .
[ fig : hvfsrs ] , as well as the corresponding fit from eq .
[ eqn : lineshape ] .
this data corresponds to the first visible resonance in the wide bandwidth measurements presented in fig .
[ fig : refltx16ghz ] .
we have taken the normalized average of five data sets for each polarization mode . the cavity mode due to horizontally - polarized light
has a measured fsr of ( @xmath92mhz , whereas the vertically - polarized mode has a slightly lower fsr of ( @xmath93mhz . to illustrate the effectiveness of our measurement technique
, we can use the known refractive indices of the non - linear crystal to predict the difference in resonant frequencies of the two polarization modes .
the refractive indices for the ordinary and extraordinary axes of a @xmath90 mgo - doped linb@xmath91 can be calculated using the temperature - independent sellmeier equation @xcite , @xmath94 where @xmath95 is the refractive index of the ordinary axis , @xmath96 is the refractive index of the extraordinary axis , and @xmath97 is the wavelength of the laser in microns ( 1.55032@xmath98 ) .
the resonant frequency of a cavity can be calculated based on the total path length as @xmath99 where @xmath33 is the speed of light in a vacuum , @xmath100 is the index of refraction of air ( 1.000273 ) , @xmath101 is the cavity path length in air , @xmath102 is the index of refraction of the non - linear crystal , and @xmath103 is the crystal length ( 10.18@xmath0 mm ) .
first we can determine @xmath101 from the measured fsr for horizontally - polarized light .
then we can predict the fsr for vertically - polarized light based on this value , and the refractive index of the ordinary axis , @xmath95 . using this method
, we predict an fsr of 514.384@xmath0mhz , which is very close to the measured value of ( @xmath93mhz ( relative difference of only @xmath104 ) .
thus , our technique is sensitive enough to measure the effect of a path length difference of 785@xmath98 ( @xmath105 of the total path length ) between two distinct cavity modes .
the effectiveness of our measurement technique is highlighted by the fact that such an inappreciable path length difference due to the birefringent material can be precisely measured .
we have described a simple yet powerful calibration technique that can determine the frequency response of an optical system consisting of both an optical cavity and a high - speed amplitude modulator .
we characterized both the cavity and the modulator by measuring the cavity s response over a wide frequency range with a dc photo - detector .
our method allowed us to extract the frequency - dependent modulator depth of our amplitude modulator , and characterize an optical cavity , without needing a calibrated broadband photo - detector or optical spectrum analyzer .
we used the on and off - resonance cavity data to precisely identify the intrinsic cavity linewidth , which would otherwise be corrupted by the presence of higher - order modulation harmonics from the amplitude modulator .
in addition , we demonstrated the precision and repeatability of our technique by measuring the different resonant frequencies of orthogonal polarization cavity modes .
once the modulator has been characterized , our method can be applied to characterize any passive optical element including , by not limited to , cavities .
this work was supported financially by the australian research council centres of excellence scheme number ce110001027 , the office of naval research ( onr ) , and industry canada .
the authors would like to thank greg milford for lending us the rf signal generator , and darryl budarick , shane brandon , and mitchell sinclair for much appreciated technical support .
we would also like to thank david moilanen for fruitful discussions . c. gamache , m. ttu , c. latrasse , n. cyr , m. a. duguay , and b. villeneuve , `` an optical frequency scale in exact multiples of 100 ghz for standardization of multifrequency communications , '' ieee photon .
. lett . * 8*(2 ) , 290292 ( 1996 ) .
h. haitjema , p. h. j. schellekens , and s. f. c. l. wetzels , `` calibration of displacement sensors up to @xmath106 m with nanometre accuracy and direct traceability to a primary standard of length , '' metrologia * 37*(1 ) , 2533 ( 2000 ) .
r. j. senior , g. n. milford , j. janousek , a. e. dunlop , k. wagner , h - a .
bachor , t. c. ralph , e. h. huntington , and c. c. harb , `` observation of a comb of optical squeezing over many gigahertz of bandwidth , '' opt . express * 15*(9 ) , 53105317 ( 2007 ) .
r. medeiros de arajo , j. roslund , y. cai , g. ferrini , c. fabre , and n. treps , `` full characterization of a highly multimode entangled state embedded in an optical frequency comb using pulse shaping , '' phys . rev . a * 89
* , 053828 ( 2014 ) .
s. k. korotky , a. h. gnauck , b. l. kasper , j. c. campbell , j. j. veselka , j. r. talman , and a. r. mccormick , `` 8-gbit / s transmission experiment over @xmath108 km of optical fiber using a ti : linbo3 external modulator , '' ieee j. lightw . technol . *
lt-5*(10 ) , 15051509 ( 1987 ) .
t. okiyama , h. nishimoto , i. yokota , and t. touge , `` evaluation of 4-gbit / s optical fiber transmission distance with direct and external modulation , '' ieee j. lightw . technol . *
6*(11 ) , 16861692 ( 1988 ) .
e. l. wooten , k. m. kissa , a. yi - yan , e. j. murphy , d. a. lafaw , p. f. hallemeier , d. maack , d. v. attanasio , d. j. fritz , g. j. mcbrien , and d. e. bossi , `` a review of lithium niobate modulators for fiber - optic communications systems , '' ieee j. sel .
topics quantum electron . , *
6*(1 ) , 6982 ( 2000 ) .
i. ozdur , s. ozharar , f. quinlan , s. gee , and p. j. delfyett , `` modified pound - drever - hall scheme for high - precision free spectral range measurement of fabry - perot etalon , '' electron . lett . *
44*(15 ) , 927928 ( 2008 ) .
m. aketagawa , s. kimura , t. yashiki , h. iwata , t. q. banh , and k. hirata , `` measurement of a free spectral range of a fabry - perot cavity using frequency modulation and null method under off - resonance conditions , '' meas .
* 22 * , 025302 ( 2011 ) . t. s. tan , r. l. jungerman , and s. s. elliott , `` optical receiver and modulator frequency response measurement with a nd : yag ring laser heterodyne technique , '' ieee trans .
theory techn . * 37*(8 ) , 12171222 ( 1989 ) .
r. t. hawkins ii , m. d. jones , s. h. pepper , and j. h. goll , `` comparison of fast photodetector response measurement by optical heterodyne and pulse response techniques , '' ieee j. lightw .
technol . * 9*(10 ) , 12891294 ( 1991 ) .
a. k. m. lam , m. fairburn , and n. a. f. jaeger , `` wide - band electrooptic intensity modulator frequency response measurement using an optical heterodyne down - conversion technique , '' ieee trans .
theory techn . * 54*(1 ) , 240246 ( 2006 ) .
a. a. chtcherbakov , r. j. kisch , j. d. bull , and n. a. f. jaeger , `` optical heterodyne method for amplitude and phase response mmeasurement for ultrawideband electrooptic modulators , '' ieee photon .
. lett . * 19*(1 ) , 1820 ( 2007 ) .
r. l. jungerman , c. johnsen , d. j. mcquate , k. salomaa , m. p. zurakowski , r. c. bray , g. conrad , d. cropper , and p. hernday , `` high - speed optical modulator for application in instrumentation , '' ieee j. lightw
. technol . * 8*(9 ) , 13631370 ( 1990 ) .
r. w. p. drever , j. l. hall , f. v. kowalski , j. hough , g. m. ford , a. j. munley , and h. ward , `` laser phase and frequency stabilization using an optical resonator , '' appl .
b * 31*(2 ) , 97105 ( 1983 ) . | precise knowledge of an optical device s frequency response is crucial for it to be useful in most applications .
traditional methods for determining the frequency response of an optical system ( e.g. optical cavity or waveguide modulator ) usually rely on calibrated broadband photo - detectors or complicated rf mixdown operations . as the bandwidths of these devices continue to increase , there is a growing need for a characterization method that does not have bandwidth limitations , or require a previously calibrated device . we demonstrate a new calibration technique on an optical system ( consisting of an optical cavity and a high - speed waveguide modulator ) that is free from limitations imposed by detector bandwidth , and does not require a calibrated photo - detector or modulator .
we use a low - frequency ( dc ) photo - detector to monitor the cavity s optical response as a function of modulation frequency , which is also used to determine the modulator s frequency response .
knowledge of the frequency - dependent modulation depth allows us to more precisely determine the cavity s characteristics ( free spectral range and linewidth ) .
the precision and repeatability of our technique is demonstrated by measuring the different resonant frequencies of orthogonal polarization cavity modes caused by the presence of a non - linear crystal . once the modulator has been characterized using this simple method , the frequency response of any passive optical element can be determined . |
the @xmath0c+@xmath0c reaction has attracted enormous interest over the years and considerable effort has been devoted to the theoretical and experimental studies of this system .
there is a large body of experimental data measured so far @xcite which have been attempted to be explained theoretically by using both phenomenological and microscopic potentials ( see @xcite for the details and the references therein ) . in the last two decades ,
the scattering observables of this reaction have been described by using the optical model and coupled - channels formalisms .
significant progress has been achieved in explaining the interaction potential between two nuclei , in particular , in the energy region of 6 mev per nucleon and over .
their angular location and cross - section have led to the determination of the gross features of the local optical potentials and ambiguities have been clarified in many cases regarding the depths of the real parts of the nuclear potentials @xcite . a good understanding of the theoretical basis of their features has been provided . as pointed out in reference @xcite ,
the resulting phenomenological potentials are strongly attractive , with relatively weak absorption , and they depend upon the bombarding energy .
however , a simultaneous description of the angular distributions , excitation functions , resonances and total reaction cross - section data have not been provided in a systematic way so far , for the energies around 6 mev per nucleon and under . for this energy region ,
three different types of theoretical calculations have been conducted : the first type of calculations focuses on the observed resonances @xcite , which are one of the outstanding features of the light - heavy - ion collisions .
it has been speculated @xcite that these broad resonances may be traced to either rotational structures @xcite , coupling of the elastic scattering channel to the inelastic channel via the crossing of aligned molecular bands ( the band crossing model ) @xcite , or interference effects resulting from reflected and refracted waves within the nuclear medium @xcite .
the second type of calculations concentrates on the excitation functions and reaction cross - section data @xcite . among these studies ,
brandan _ et al _ @xcite have extended their potentials which fit the high experimental data ( @xmath2 ) to lower energies . from these analyses ,
mainly two types of deep potentials have been obtained , called unam potentials .
unam potentials explain the high energy experimental data in a systematic way and they also predict the overall features of the 50@xmath1 to 90@xmath1 elastic scattering excitation functions and reaction cross - section data . a slightly modified version of these potentials has also been used in the excitation function analysis at low energies @xcite .
these two potentials have linear and quadratic energy - dependence ( equation 6 and 7 ) in reference @xcite .
although these potentials provide good agreement with the experimental data for energies @xmath2 , the analyses fail for the excitation function and reaction cross - section data at low energies : first of all , in the excitation function analysis , the linearly - energy dependent potential ( equation 6 ) gives the correct period and phase for the gross oscillations , but the strength is too weak to reproduce the excitation functions at 60@xmath1 and 80@xmath1 ( ref .
@xcite , figure [ totalxsec ] ) .
the quadratic one ( equation 7 ) gives a much worse prediction as shown in reference @xcite , the calculation exactly out of phase with respect to the data for the 90@xmath1 excitation function .
an analysis similar to the work of brandan _ et al _ @xcite has also been conducted by kondo _
et al _ @xcite .
they have modified the unam potentials of brandan _ et al _ @xcite even further in order to explain 50@xmath1 to 90@xmath1 excitation function and reaction cross - section data . although there are still problems with the strength and phases of the oscillation , an overall explanation of the excitation functions and reaction cross - section data have been provided by this work .
neither brandan _ et al _ @xcite nor kondo _ et al _ @xcite shows in their paper how good their predictions for the elastic scattering angular distributions are .
thirdly , coupled - channels calculations have been conducted for this reaction .
boztosun and rae @xcite have recently analyzed this reaction over a wide energy range from 32.0 to 126.7 mev using a new coupling potential .
they have added to the usual first - derivative coupling potential with a second - derivative one by considering the orientation of two interacting nuclei and have attempted to solve particularly the magnitude problem of the inelastic mutual-2@xmath3 state cross - section at high energies .
their approach has provided a good explanation of the elastic and inelastic angular distributions as well as of the excitation functions and has solved the magnitude problem for the inelastic mutual-2@xmath3 state data , but there are still justification problems regarding the use of such a coupling potential . in sum , the survey of the literature clearly shows that there is not a potential model that can explain simultaneously the measured elastic scattering angular distribution , the average behavior of the 50@xmath1 to 90@xmath1 excitation functions and reaction cross - section data in the low energy region . in the light of these studies , we have examined the experimental data of @xmath0c+@xmath0c elastic scattering in low and intermediate energy regions .
the examination has been conducted in a systematic way in order to find a potential family that simultaneously fits the angular distributions and excitation functions data and to address the problems of this reaction within the framework of the optical and coupled - channels models .
18 individual angular distribution , 5 excitation functions and reaction cross - section data between 32.0 and 70.7 mev in the laboratory system have been studied , both within the framework of the optical model and the coupled - channels formalism .
the experimental data analyzed in this paper is taken from @xcite . in the following section ,
we have introduced our optical model as well as the potential parameters . in this section , we have first presented the results of the theoretical calculations for the individual angular distributions , 50@xmath1 to 90@xmath1 single - angle elastic scattering excitation functions and the reaction cross - section data by using the optical model and have compared them with the experimental data .
we have then provided the volume integrals of the real and imaginary potentials and have discussed their properties in terms of the dispersion relation . in section [ ccmodel ] , we have also introduced our coupled - channels model and the results of this analysis are shown for the individual angular distributions , 50@xmath1 to 90@xmath1 single - angle elastic scattering excitation functions and the reaction cross - section data .
the optical model and coupled - channels calculations have been compared and the effect of including the inelastic channels on the scattering for low energies has been clearly demonstrated .
section [ conc ] is devoted to our summary and conclusion .
in our optical model analysis , we have used phenomenological complex potentials : @xmath4^{2 } } + i \frac{-w}{1+e^{\frac{r - r_{w}}{a_{w } } } } \label{imagx}\ ] ] here , @xmath5=@xmath6 $ ] ( @xmath7 or @xmath8 ) where @xmath9 and @xmath10 are the masses of projectile and target nuclei and @xmath11 and @xmath12 are the radius parameters of the real and imaginary parts of the nuclear potential respectively .
the real part of the nuclear potential has the square of the woods - saxon shape and the depth ( @xmath13=280.0 mev ) has been fixed to reproduce the experimental data over the whole energy range considered .
the experimental data of the @xmath0c+@xmath0c reaction in low energy region has a very oscillatory structure changing very rapidly from energy to energy .
therefore , the radius ( @xmath11 ) and diffuseness ( @xmath14 ) of the real potential have been varied on a grid , respectively from 0.6 to 0.9 mev , with steps of 0.01 mev , and from 1.1 to 1.5 fm with steps of 0.01 fm in order to obtain the best fit to the data .
the results of this systematic search are shown in figure [ chiall ] which is a three - dimensional plot of the @xmath11 , @xmath14 and 1/@xmath15 , where @xmath15 has the usual definition and measures the quality of the fit . in figure
[ chiall ] , the best fit parameters , producing oscillating cross - sections with reasonable phase and period , correspond to low @xmath15 values and peaks in the 1/@xmath15 surface . for the four incident energies , the figures present discrete peaks ( or hills ) for correlated @xmath11 and @xmath14 values , which are best fit real potential families and
indicate that the @xmath11 or @xmath14 parameters can not be varied continuously and still find equally satisfying fits . for the radius ( @xmath11 )
, the lowest @xmath15 values are generally obtained between 0.72 and 0.78 and for the diffuseness @xmath14 , it is between 1.30 and 1.39 .
therefore , the radius ( @xmath11 ) and diffuseness ( @xmath14 ) of the real potential are energy - dependent and the parameters are shown in table [ param ] .
the imaginary part of the nuclear potential given by equation [ imagx ] has been taken as the woods - saxon volume form and its depth ( w ) is given in table [ param ] .
the other parameters of the imaginary potential have also been fixed in the calculations as @xmath12=1.1 fm and @xmath16=0.55 fm .
the real and imaginary parts of the nuclear potential are displayed in figure [ realpot ] for various values of the orbital angular momentum .
the coulomb potential is derived from a uniformly charged sphere with a radius of 5.5 fm .
we have analyzed 18 angular distributions data measured by references @xcite between 32.0 and 70.7 mev within the above - described optical potential .
the parameters of the real and imaginary potentials as well as @xmath15 values are given in table [ param ] .
the volume integrals of the potentials are also shown in table [ param ] and they are displayed in figure [ volint ] in comparison with the dispersion relation curve between the real and imaginary parts of the potential .
the results of our optical model calculations ( solid lines ) are shown in figures [ ground1 ] and [ ground2 ] in comparison with the experimental angular distributions of @xcite ( circles ) . as it can be seen from these figures ,
the places of the maxima and minima in experimental data have been correctly reproduced and there is no magnitude problem between our theoretical predictions and the experimental data .
good agreement between the theoretical calculations and the experimental data has been obtained within the framework of the optical model .
we have also compared our optical model results with the results of the brandan _ et al _
@xcite for 5 energies between 50 and 60 mev in figure [ ground2 ] . in this figure ,
the dashed lines are the results of the unam potential by brandan _
et al _ @xcite which is used to explain the excitation functions data at low energies .
although unam potential gives the gross structure of the 90@xmath1 excitation function , it fails to give a good account of the individual angular distribution at low energies . in the forward angle region up to almost 70@xmath1 for some energies , the theoretical cross - sections are structureless whereas the experimental cross - section shows oscillatory structure .
it is pointed out in the introduction that the theoretical calculations so far have had limited success in describing the experimental data of this reaction .
the potential used to describe the individual angular distributions could not predict the overall behavior of the excitation functions or _
vice versa_. this may be seen from figure [ ground2 ] for the elastic scattering angular distribution where the dashed - lines are the predictions of unam potentials used to explain the excitation functions and reaction cross - section data by brandan @xcite and kondo _ et al _ @xcite .
therefore , we have used our potential which fits the individual angular distribution data in order to examine 50@xmath1 to 90@xmath1 excitation functions as well as reaction cross - section data .
the experimental data of references @xcite for the 50@xmath1 , 60@xmath1 , 70@xmath1 , 80@xmath1 and 90@xmath1
excitation functions have been analyzed using the parameters of table [ param ] . in the analysis of individual angular distributions ,
we have used only 3 free parameters ( @xmath11 , @xmath14 and w ) within the optical model and we use now the linear interpolation of the @xmath11 , @xmath14 and w for the excitation function calculations .
the reason why we interpolate these parameters is to prevent artificial peaks that might be created by the change of parameters from energy to energy .
the linear interpolation of the radius , diffuseness and imaginary potential parameters are @xmath17 the results of our optical model calculations with these interpolations are shown in figure [ exc-cc.ps ] for the 90@xmath1 excitation function and in figure [ excnew.ps ] for the other angles .
it may be perceived from these figures and the @xmath15 values in table [ chisqaure ] that the same optical potentials fitting the individual angular distributions could predict the magnitude of the cross - section correctly and reproduce the period of the gross - structure reasonably well . in the same energy region , we have also examined the reaction cross - section data of this reaction by using the same potential parameters _
i.e. _ the parameters in table [ param ] and equation [ interp ] for the @xmath11 , @xmath14 and w interpolation . in figure
[ totalxsec ] , our reaction cross - section result is shown in comparison with the experimental data as well as with other theoretical calculations conducted so far . we have observed from this figure that our results are in very good agreement with the experimental data ; the magnitude and the phases of the oscillations have been correctly provided by our theoretical calculations up to @xmath18 e@xmath19=60 mev .
it should be pointed out that our calculations are in better agreement with the experimental data in comparison with the other theoretical calculations conducted so far for the same experimental measurement . as a result , we have shown in this paper that a potential family could explain 18 individual angular distributions data , 5 excitation functions and reaction cross - sections data simultaneously in a systematic way by using the optical model .
this outcome reveals that our potential belongs to a different family other than the potentials that have been previously used to analyze the experimental data in the higher energy region @xcite . in figure [ potcomp ]
, we compare our real potential with the unam potentials of brandan _ et al _ @xcite and kondo _ et al _ @xcite which are successful in describing the high energy data but fail to explain the low energy data .
it may be perceived in this figure that our real potential is more diffusive and shallower than their potentials which is a phenomenon distinguishing it from other potentials in the surface region where it is almost half as big as other potentials .
similar differences may be asserted for the imaginary potentials .
the volume integrals and the dispersion relation between the real and imaginary potentials are shown in figure [ volint ] in comparison with brandan _ et al _ s potential @xcite .
the volume integrals of the real and imaginary potentials and the dispersion relation @xcite between them have been calculated by using the following formula : @xmath20 @xmath21 \label{disp}\ ] ] here @xmath22 with @xmath23 as @xmath24 and @xmath25 respectively .
the parameters are @xmath26 mev , @xmath27 mev , @xmath28 mev and @xmath29 mev
. we can derive from figure [ volint ] that the real potential does not obey the dispersion relation at resonance energies observed by cormier _
@xcite for single and mutual-2@xmath3 states for j=14 , 16 and 18 spin values at e@xmath19=@xmath30 and @xmath31 mev respectively . in this resonance region , the volume integral of the real potential oscillates remarkably , but the imaginary potential does not accompany the variation of the real one .
one explanation of this may be due to the rapid variation of the experimental data , which can not be described by a smoothly varying parameter set and thus , the parameters oscillate .
this is also clearly seen from our @xmath15 search results in figure [ chiall ] .
for example , the lowest @xmath15 values for e@xmath32=50 and 52 mev in this figure require two very different @xmath11 and @xmath14 values although there are only @xmath33 mev energy difference between two data .
it is also possible to interpret that the oscillations in the real potential are manifestations of the coupling to the 2@xmath3 state , due to the strongly deformed structure of the @xmath0c nucleus . in a previous report @xcite ,
the strongly deformed structure of the @xmath0c nucleus has been examined and it has been shown that coupled - channels calculations with a 20@xmath34 decreased imaginary potential of the optical model calculations have provided an equally good fit to the experimental data .
this outcome has shown that the inclusion of the inelastic channels mainly affects the absorption and they have almost no effect in the real part of the nuclear potential in the high energy region for the @xmath0c+@xmath0c system .
however , other previous works such as sakuragi and kamimura @xcite and references therein have investigated the breakup effect and have shown that the inclusion of the inelastic channels induces a large repulsive real potential .
the energies we have studied in this paper are lower than the analysis of ref .
@xcite and the nature of the experimental data is very different than the nature of those in higher energies .
therefore , it should be questioned how the inclusion of the inelastic channels affects the scattering observables of this reaction . for this purpose ,
test calculations have been performed at e@xmath19=35.0 mev to see the coupling effect for low energies . in these calculations
, we have included the single-2@xmath3 ( 4.44 mev ) and mutual-2@xmath3 ( 8.88 mev ) states of the strongly deformed @xmath0c nucleus .
the results of test calculations obtained by using the optical model parameters as shown in table [ param ] for e@xmath19=35.0 mev are displayed in figure [ ccresults1 ] .
test runs with and without coupling ( _ i.e. _ optical model ) at e@xmath19=35.0 mev have , however , shown striking differences .
as shown in figure [ ccresults1 ] , the inclusion of the excited states at e@xmath19=35.0 mev has very large effects not only in backward , but also in forward - angle regions ( @xmath35 ) .
the coupling changes the magnitudes and phases of the oscillations in the cross - section .
we have infered from this effect that the coupling not only affects the absorptive or imaginary part , but also the real part of the nuclear potential . as a result
, it modifies the interference between incoming and outgoing waves , which creates the oscillatory structure in the cross - section .
our findings are in agreement with the observation of sakuragi and kamimura @xcite .
therefore , according to this outcome , the claim of ref .
@xcite that is the inclusion of the inelastic channels mainly affects the absorption and they have almost no effect in the real part of the nuclear potential in the high energies is not valid for the low energy region the @xmath0c+@xmath0c system .
having seen the coupling effect , we conduct coupled - channels calculation for the same energies considered in optical model calculations . in the coupled - channels calculations , the interaction between the @xmath0c nuclei is described by a deformed optical potential . the real potential has the square of a woods - saxon shape as in equation [ imagx ] with a depth of 280 mev . the other parameters ,
as shown in table [ paramcc ] , have been fitted to reproduce the elastic scattering data .
the imaginary potential has the standard woods - saxon volume shape as in equation [ imagx ] and the parameters of its depth are given in table [ paramcc ] .
the radius and diffuseness of the imaginary potential have been fixed in the calculations as @xmath12=1.1 fm and @xmath16=0.55 fm .
it has been assumed that the target nucleus @xmath0c has a static quadrupole deformation and this assumption has been taken into account by deforming the real potential in the following way : @xmath36+r_{0}a_{t}^{1/3}[1+\beta_{2 } y_{20}(\theta,\phi ) ] \label{beta}\ ] ] where the first and second terms account for the projectile and target excitations respectively . in equation [ beta ] , @xmath37=-0.6 is the deformation parameter of the @xmath0c nucleus .
this value is derived from its known b(e2 ) value , which is 42 e@xmath38 @xcite .
we have noticed that deformation of the imaginary part of the nuclear potential does not have any significant effect .
thus , for computational simplicity , we have not deformed it in the present coupled - channels calculations . an extensively modified version of the code chuck and the code fresco has been used both in the optical model and coupled - channels calculations @xcite . using this coupled - channels model , we have analyzed the same experimental data as in the case of the optical model .
the parameters of the real and imaginary potentials as well as their volume integrals are given in table [ paramcc ] .
the volume integrals are also displayed in figure [ volint ] in comparison with the volume integrals of the optical model potentials and the dispersion relation curve obtained by using equation [ disp ] .
the results of our coupled - channels calculations ( solid lines ) are shown in figures [ ground1cc ] and [ ground2cc ] in comparison with the experimental data ( circles ) .
as it can be seen from these figures and the @xmath15 values in table [ paramcc ] , reasonable agreement between the theoretical calculations and the experimental data has been obtained . the phases and magnitudes of the oscillations
have been predicted correctly for the energies considered .
the effect of the inclusion of the single-2@xmath3 state of the @xmath0c nucleus in coupled - channels calculations should be underlined here .
the effect of the coupled - channels calculations has been very large on the scattering and its effect has been clearly observed in our theoretical calculations for the cross - section at forward , intermediate and large angles .
it has changed the phases and magnitudes of the oscillations at all angles . in order to obtain the agreement between theoretical calculations and experimental data
, we have had to change not only the imaginary part , but also the real part of the nuclear potential .
as seen in figure [ volint ] , the inclusion of the single-2@xmath3 state has reduced the strength of the imaginary potential at all energies in comparison with the imaginary potential of the optical model .
this reduction of the imaginary potential was expected since the coupled - channels calculations take into account the effect of the eliminated channels of the optical model .
however , the coupling has had a big effect on the real potential and has changed the strength of the real potential very much .
this remarkable change of the real potential parameters has been due to the coupling between ground and single-2@xmath3 states and in contrast with the observations at high energies . at high energies ,
the coupling does not change the phase of the oscillations and therefore , reducing the imaginary potential would be enough to obtain optical model results .
we should state that this strong coupling effect should be taken into account in the analysis in order to gain a better description of the nuclear potential in the interaction of such strongly deformed two nuclei .
we have also analyzed the 50@xmath1 to 90@xmath1 excitation functions as well as reaction cross - section data . we have used linear interpolation of 3 free parameters given in table [ paramcc ] , _ i.e. _ @xmath11 , @xmath14 and w , as follows : @xmath39 the results for the 90@xmath1 excitation function are shown in figure [ exc-cc.ps ] , for the other angles in figure [ excnew.ps ] , and for the reaction cross - section data in figure [ totalxsec ] in comparison with the optical model results and experimental data . again reasonable agreement with the experimental data both for the excitation functions and for the reaction cross - section data has been obtained within the framework of the coupled - channels method .
however , we should point out that the optical model results for the excitation function calculations are in better agreement with the experimental data as it can be seen from the @xmath15 values in table [ chisqaure ] . for the reaction cross - section ,
the coupled - channels results are better than the optical model results at low energies , but towards high energies our coupled - channels calculations have over - estimated the data .
it is higher than the experimental data and optical model results .
this discrepancy arises due to the interpolation formulae given in equation [ interp2 ] for @xmath11 , @xmath14 and w. second or third - order interpolations of the parameters in table [ paramcc ] have provided a better result for the excitation functions and reaction cross - section data . in this section , we give the prediction of our potential for the well - known resonances measured by cormier _
@xcite for the single-2@xmath3 state .
these resonances observed at low energies could not be predicted by a potential , which also fits the angular distributions and the excitation functions .
we show the predictions of our potential in figure [ smat ] . in the upper part of this figure ,
we display the real versus the imaginary part of the s - matrix for the resonance spin values j=14 , j=16 and j=18 and in the lower part of the figure we show the magnitudes of the s - matrix ( @xmath40 ) for the elastic ( @xmath41 ) and single-@xmath42 ( @xmath43 ) channels against the center of mass energy for the same spin values .
our potential predict the places of the resonances at the correct energies with reasonable widths .
we provide a consistent description of the elastic scattering of the @xmath0c+@xmath0c system from 32.0 to 70.7 mev in the laboratory system by using a phenomenological , strongly attractive woods - saxon squared nuclear potential , with a relatively weak absorption , with both the optical model and coupled - channels formalism .
this reaction has been one of the most extensively studied reaction over the last forty years , in particular , in the low energy region where an oscillatory structure in the excitation functions and resonances observed has been pronounced .
unfortunately , no global model has been set forth so far that consistently explains the measured experimental data over a wide energy range . in the introduction
, we present the problems that this reaction manifests , one of the most important of which is the simultaneous description of the elastic scattering angular distributions and the single - angle excitation functions data .
the theoretical calculations in the past have reported that a potential family that fits the individual angular distributions is unable to reproduce the excitation functions data of the same reaction in a systematic way . by considering these problems ,
we analyze this reaction with available experimental data in the energy range we considered and show that it is possible to improve the agreement for the individual angular distributions , excitation functions and the reaction cross - section data simultaneously with a deep real and shallow imaginary potential . from our analysis , we observe that away from the resonance energies , the variation of the potential parameters has a systematic energy - dependence as in the case of the high energy region .
however , when there is a rapid variation of the experimental data , it can not be described by a smoothly varying parameter set and thus , the parameters oscillate . in our analysis
, we observe this effect .
it should be underlined here that the remarkable success of our potential in explaining the individual angular distributions , excitation functions and the reaction cross - section data simultaneously is achieved by using only 3 free parameters ( @xmath44 , @xmath45 and @xmath8 ) in the calculations .
we should also point out that the coupling between ground and excited states has a large effect on the scattering in the low energy region .
the inclusion of the excited states of the @xmath0c nucleus at e@xmath19=35.0 mev alters the places and magnitudes of oscillations at both forward and backward angles around 90@xmath1 in the theoretical results .
that means , the coupling not only changes the absorptive part of the nuclear potential , but also the real part of the nuclear potential which is in agreement with the microscopic calculations of sakuragi and kamimura @xcite .
our results show that at low energies , the coupled - channels effect should be explicitly taken into account for strongly deformed nuclei such as @xmath0c . at high energies , we observe that the inclusion of the excited states does not change the phases of the oscillation but affects the magnitude of the cross - section at large angles
. we do not observe effects at forward angles .
no effect on the phases of the oscillations means that the inclusion of the excited states has almost no effect on the real part of the nuclear potential .
the effect at large angles can also be compensated by decreasing the absorptive part of the nuclear potential .
this outcome is in agreement with previous claims @xcite , but valid only at high energies .
the deep real potential used in our optical and coupled - channels analysis has a different radial shape , separating it from other potentials used in the analysis of the higher energy data ( @xmath46 mev ) .
both optical and coupled - channels potentials in our analysis provide very good fits to the elastic scattering angular distributions and reasonable improvement to the 50@xmath1 to 90@xmath1 excitation functions and reaction cross - section data .
it also predicts the resonances at the correct energies with reasonable widths . to our knowledge
, this has not been achieved for the @xmath0c+ @xmath0c reaction at low energies so far .
bromley , j.a .
kuehner and e. almquist , phys .
* 4 * ( 1960 ) 365 .
e. almquist , d.a .
bromley and j.a .
kuehner and , phys .
rev . lett . * 4 * ( 1960 ) 515 .
erb and d.a .
bromley , phys .
c * 23 * ( 1981 ) 2781 .
cormier , j. applegate , g.m .
berkowitz , p. braun - munzinger , p.m. cormier , j.w .
harris , c.m .
jachcinski , l. lee , j. barrette and h.e .
wegner , phys .
* 38 * ( 1977 ) 940 .
cormier , c.m .
jachinski , g.m .
berkowitz , p. braun - munzinger , p.m. cormier , m. gai , j.w .
harris , j. barrette and h.e .
wegner , phys .
* 40 * ( 1978 ) 924 .
fulton , t.m .
cormier and b.j .
herman , phys . rev . * c21 * ( 1980 ) 198 .
cosman , t.m .
cormier , k. van bibber , a. sperduto , g. young , j. erskine , l.r . greenwood and o. hansen , phys
lett * 35 * ( 1975 ) 265 .
cosman , r. ledoux , and a.j .
lazzarini , phys .
c * 21 * ( 1980 ) 2111 .
stokstad , r.m .
wieland , c.b .
fulmer , d.c .
hensley , s. raman , a.h .
snell and p.h .
stelson , oak ridge national laboratory , report no .
ornl / tm-5935 , 1977 ( unpublished ) .
stokstad , r.m .
wieland , g.r .
satchler , c.b .
fulmer , d.c .
hensley , s. raman , l.d .
rickertsen , a.h . snell and p.h .
stelson , phys .
c * 20 * ( 1979 ) 655 .
a. morsad , f. haas , c. beck , and r.m .
freeman , z. phys .
* a338 * ( 1979 ) 61 .
j.j.kolata , r.m .
freeman , f. haas , b. heusch and a. gallman , phys .
c 21 ( 1980 ) 579 .
erb and d.a .
bromley , in treatise on heavy - ion science , vol .
3 , ed . d.a .
bromley ( plenum , new york , 1985 ) .
w. greiner , j.y .
park , w. scheid , nuclear molecules ( world - scientific , singapore 1995 ) .
satchler , direct nuclear reactions ( oxford uni . press , oxford 1983 ) .
+ h. feshbach , theoretical nuclear physics : nuclear reactions ( wiley , newyork 1992 ) . m.e .
brandan and g.r .
satchler , phys .
285 ( 1997 ) 143 .
m. freer , m.p .
nicoli , s.m .
singer , c.a .
bremner , s.p.g .
chappell , w.d.m .
rae , i. boztosun , b.r .
fulton , d.l .
watson , b.j .
greenhalgh , g.k .
dillon , r.l .
cowin and d.c .
weisser , phys .
c**70**(2004 ) 064311 .
bremner , s.p.g .
chappell , w.d.m .
rae , i. boztosun , m. freer , m.p .
nicoli , s.m .
singer , b. fulton , d.l .
watson , b.j .
greenhalgh , g.k .
dillon , r.l .
cowin , phys .
rev . c**66 * * ( 2002 ) 034605 .
s. marsh , and w.d.m .
rae , phys .
* 180b * ( 1986 ) 185 .
h. flocard , p.h .
heenen , s.j .
krieger , and m.s .
weiss , prog .
72*(1984 ) 1000 .
ordoez , r.j .
ledoux and e.r .
cosman , phys . lett .
* 173b * ( 1986 ) 39 .
y. kondo , y. abe and t. matsuse , phys .
c * 19 * ( 1979 ) 1356 .
y.abe , y. kondo and t. matsuse , suppl .
* 68 * ( 1980 ) 303 .
t. matsuse , y. abe , y. kondo , prog .
* 59 * ( 1978 ) 1904 .
a. morsad , f. haas , c. beck , and r.m .
freeman , z. phys * 338 * ( 1991 ) 61 .
n. rowley , h. doubre , c. marty , phys .
69b * ( 1997 ) 147 .
brandan , m. rodriguez - villafuerte and a. ayala , phys .
c * 41 * ( 1990 ) 1520 .
y. kond , m.e .
brandan and g.r .
satchler , nucl . phys .
* a637 * ( 1998 ) 175 .
mcvoy and m.e .
brandan , nucl . phys . * a542 * ( 1992 ) 295 .
brandan , m.s .
hussein , k.w .
mcvoy and g.r .
satchler , comments nucl .
( 1996 ) 77 .
i. boztosun and w.d.m rae , phys .
c * 63 * ( 2001 ) 054607 .
+ i. boztosun and w.d.m .
rae , phys .
* 518b * ( 2001)229 .
f michel and s. ohkubo , eur .
j. * a19 * ( 2004 ) 333 .
satchler , phys .
199 ( 1991 ) 147 .
y. sakuragi and m. kamimura , phys . lett . *
149b * ( 1984 ) 307 .
+ y. sakuragi , m. yahiro and m. kamimura , prog .
* 70 * ( 1983 ) 1047 .
+ m. ito , y. sakuragi and y. hirabayashi , phys .
c * 63 * , 064303 ( 2001 ) .
+ y. sakuragi , m. ito , m. katsuma , m. takashina , y. kudo , y. hirabayashi , s. okabe , and y. abe , in _ proceedings of the 7th international conference on clustering aspects of nuclear structure and dynamics _ , edited by m. korolija , z. basrak , and r. caplar ( world scientific , singapore , 2000 ) , p. 138 .
stelson and l. grodzins , nucl .
data * 1a * ( 1966 ) 21 .
+ f. ajzenberg - selove , nucl . phys . *
a248 * ( 1975 ) 1 .
kunz , chuck , a coupled - channels code , unpublished .
i.j . thompson , fresco , a coupled - channels code , unpublished
thompson , computer physics reports * 7 * ( 1988 ) 167 .
.the parameters and the volume integrals of the real and imaginary potentials as well as the @xmath15 values of the optical model calculations .
coulomb radius , @xmath47=1.2 fm .
[ cols="<,^,^,^,^,^,^,^ " , ] c and @xmath0c is plotted against the separation r for various values of the orbital angular momentum quantum number , @xmath48 .
the inserted figure shows our imaginary potential at @xmath49=32.0 mev and 70.7 mev .
the parameters are given in the text.,title="fig:",scaledwidth=100.0%]+0.0 cm c+@xmath0c system ( _ continued from figure [ ground1 ] _ ) .
the dashed lines are the angular distributions of the unam potential by brandan _
et al _ @xcite , which is used to explain the excitation functions data at low energies .
the experimental data is taken from @xcite .
, title="fig:",scaledwidth=100.0% ] -2.0 cm , 60@xmath1 , 70@xmath1 and 80@xmath1 elastic scattering excitation function results with the experimental data by using the optical model ( solid lines ) and coupled - channels model ( dashed lines ) .
the experimental data is taken from @xcite.,title="fig:",scaledwidth=100.0% ] + 0.0 cm and mutual-2@xmath3 excited states of the @xmath0c nucleus at e@xmath19=35.0 mev .
the optical model parameters of table [ param ] are used in the calculations . see the text for the discussion.,title="fig:",scaledwidth=100.0% ] | we examine the @xmath0c+@xmath0c elastic scattering over a wide energy range from 32.0 to 70.7 mev in the laboratory system within the framework of the optical model and the coupled - channels formalism .
the @xmath0c+@xmath0c system has been extensively studied within and over this energy range in the past .
these efforts have been futile in determining the shape of the nuclear potential in the low energy region and in describing the individual angular distributions , single - angle 50@xmath1 to 90@xmath1 excitation functions and reaction cross - section data simultaneously . in order to address these problems systematically ,
we propose a potential that belongs to a family other than the one used to describe higher energy experimental data and show that it is possible to use it over this wide energy range .
this potential also predicts the resonances at correct energies with reasonable widths .
optical model , coupled - channels calculations , elastic and inelastic scattering , dispersion relation , resonance , excitation functions , reaction / absorption cross - section , @xmath0c+@xmath0c reaction . 24.10.ht ; 24.10.eq ; 24.50.+g |
weakly interacting massive particles ( wimps ) are popular candidates to account for dark matter ( dm ) in the universe .
there are different ways to explore the landscape of dm models from a phenomenological perspective .
one is the use of effective operators describing the interactions between the standard model ( sm ) and the wimp in the framework of effective field theory ( eft ) .
another approach is to use simplified models containing a minimal set of new particles that allow to describe the phenomenology of dm via renormalizable interactions . although the eft framework has been successfully used for the description of dm interactions at rather low energies like the interpretation of direct detection experiments
it has been pointed out that eft may break down when probing dark matter production at the lhc @xcite . in this article
we consider a model that extends the sm by a majorana fermion dm , @xmath0 , and a vector mediator , @xmath1 , which couples to the dm and the sm quarks with axial - vector interactions , with couplings @xmath2 and @xmath3 , respectively .
for such a model , lhc searches are expected to be more sensitive than direct detection experiments as the model does not provide any contribution to spin - independent wimp - nucleon scattering .
we present limits on the parameters space of this model from mono - jet searches at the lhc @xcite and compare them to the respective limits obtained in the eft approximation . for realistic values of the couplings , @xmath4
, the lhc provides limits on the messenger mass @xmath5 .
as these are accessible energies at lhc collisions , contributions from on - shell messenger production can be large .
hence , limits from the simplified model and the eft can differ significantly as we will discuss in sec .
[ sec : lhc ] . as a complementary constraint on the parameter space we consider limits on the spin - dependent wimp - proton scattering from dm annihilation in the sun .
to this end we re - interpret limits from icecube @xcite within our model , where annihilation into top quark pairs and pairs of mediators are important .
these limits are particularly constraining for large dm masses where the lhc looses its sensitivity .
we discuss them in sec .
[ sec : icecube ] . in sec .
[ sec : concl ] we conclude .
in this work we interpret two searches for mono - jet plus missing transverse momentum ( met ) signatures performed by atlas @xcite and cms @xcite at the @xmath6tev lhc . to this end
we performed a monte carlo simulation of the signal and imposed the search cuts detailed in @xcite .
based on the background analysis provided in these references we are thus able to set 95% cl exclusion limits on the parameters of the model . for details
we refer to @xcite .
the considered model has four independent parameters .
the dm mass , @xmath7 , the mediator mass , @xmath8 , and the couplings of the mediator to the dm , @xmath2 , and the sm quarks @xmath9 .
we assume universal couplings to all sm quarks and neglect couplings to leptons .
we show our results for various slices of the parameter space where we fix the product of the couplings , @xmath10 and the mediator width , @xmath11 .
we choose this parametrization as the cross section for dm production directly depends on these parameters .
however , not all values of @xmath11 and @xmath12 are consistent within this model as we will show below . for comparison
we also derive limits in the eft approximation of the considered model . for this
we integrate out the messenger to obtain a 4-fermion contact interaction with an effective coupling @xmath13 .
hence , the parameter space reduces to two parameters , @xmath7 and @xmath14 . in fig .
[ fig : mainres ] we show the exclusion limits for the eft ( dashed lines ) and the simplified model ( solid lines ) for four slices of the considered parameter space . whilst the eft limit extents to very high dm masses above a tev
the limit from simplified models goes down very drastically for @xmath15 . in this region
the eft approximation significantly over - estimates the sensitivity .
further , also for @xmath16 we find significant deviations in the resulting limit on @xmath8 .
this is due to the fact that the limit on @xmath8 placed for @xmath17 lies in the region of reachable lhc energies .
this can lead to both an under- or over - estimation of the sensitivity .
on the one hand , contribution from on - shell mediator production could greatly enhance the cross section .
this is the dominant effect for the parameter slices with @xmath18 ( left panels of fig .
[ fig : mainres ] ) .
the effect becomes more pronounced for the smaller coupling , @xmath19 , ( see lower panels ) as the limits are placed at lower @xmath8 where the contribution from on - shell mediator production is even larger . in this region
eft under - estimates the sensitivity . on the other hand , for very small @xmath8 ( below the energy scale of the required met in the considered searches )
the cross section is reduced and the eft approximation again over - estimates the limit .
this can be seen in the case @xmath19 , @xmath20 ( lower right panel ) where the cms limit for the simplified model completely vanishes whilst the eft would exclude @xmath21 .
as mentioned above not all combinations of @xmath7 , @xmath8 , @xmath22 and @xmath11 are consistent within the model . in fig .
[ fig : mainres ] we marked in blue the regions where no such solution exist .
note that the region @xmath23 the region where the eft shows its best agreement is barely accessible for a reasonably small width of @xmath18 .
( 1.,0.87 ) ( -0.01,0.48 ) ( 0.08,0.025)-@xmath24 plane in four slices of the considered parameter space regarding @xmath22 and @xmath11 @xcite .
we show the 95% cl lower exclusion limits from mono - jet searches at atlas ( blue lines ) and cms ( red lines ) for the simplified model ( solid lines ) and the eft approximation ( dashed lines ) .
further , we show 90% cl lower exclusion limits from the icecube neutrino observatory ( green lines ) .
the dark grey shaded band denotes the region where the thermal relic density matches the dm density within @xmath25 . in the light - grey shaded region above it dm is over - produced . in the blue shaded region in the left panels
no solution exist for the individual couplings @xmath12 requiring @xmath26 and @xmath11 .
the orange shaded regions are excluded by searches for resonances in di - jet signatures taken from ref .
@xcite.,title="fig:",scaledwidth=40.0% ] ( 0.123,0.415)@xmath27 , @xmath18 ( 0.26,0.0)@xmath28 $ ] ( 0.05,0.16 ) ( 0.131,0.15 ) ( 0.232,0.06 ) ( 0.414,0.3 ) ( 0.255,0.337 ) ( 0.15,0.28 ) ( 0.433,0.131 ) ( 0.353,0.184 ) ( 0.227,0.175 ) ( 0.48,0.48 ) ( 0.08,0.025)-@xmath24 plane in four slices of the considered parameter space regarding @xmath22 and @xmath11 @xcite . we show the 95% cl lower exclusion limits from mono - jet searches at atlas ( blue lines ) and cms ( red lines ) for the simplified model ( solid lines ) and the eft approximation ( dashed lines ) .
further , we show 90% cl lower exclusion limits from the icecube neutrino observatory ( green lines ) .
the dark grey shaded band denotes the region where the thermal relic density matches the dm density within @xmath25 . in the light - grey shaded region above it dm is over - produced . in the blue shaded region in the left panels
no solution exist for the individual couplings @xmath12 requiring @xmath26 and @xmath11 .
the orange shaded regions are excluded by searches for resonances in di - jet signatures taken from ref .
@xcite.,title="fig:",scaledwidth=40.0% ] ( 0.123,0.415)@xmath27 , @xmath20 ( 0.26,0.0)@xmath28 $ ] ( 0.05,0.16 ) ( 0.131,0.15 ) ( 0.414,0.3 ) ( 0.02,0.09 ) ( 0.39,0.188 ) ( 0.386,0.17 ) ( 0.39,0.152 ) ( 0.22,0.325 ) ( 0.16,0.283 ) ( 0.433,0.131 ) ( 0.353,0.184 ) ( 0.227,0.175 ) ( -0.01,0.0 ) ( 0.08,0.025)-@xmath24 plane in four slices of the considered parameter space regarding @xmath22 and @xmath11 @xcite .
we show the 95% cl lower exclusion limits from mono - jet searches at atlas ( blue lines ) and cms ( red lines ) for the simplified model ( solid lines ) and the eft approximation ( dashed lines ) .
further , we show 90% cl lower exclusion limits from the icecube neutrino observatory ( green lines ) .
the dark grey shaded band denotes the region where the thermal relic density matches the dm density within @xmath25 . in the light - grey shaded region above it dm is over - produced . in the blue shaded region in the left panels
no solution exist for the individual couplings @xmath12 requiring @xmath26 and @xmath11 .
the orange shaded regions are excluded by searches for resonances in di - jet signatures taken from ref .
@xcite.,title="fig:",scaledwidth=40.0% ] ( 0.123,0.415)@xmath19 , @xmath18 ( 0.26,0.0)@xmath28 $ ] ( 0.05,0.16 ) ( 0.14,0.062 ) ( 0.343,0.09 ) ( 0.19,0.29 ) ( 0.17,0.24 ) ( 0.355,0.28 ) ( 0.413,0.295 ) ( 0.48,0.0 ) ( 0.08,0.025)-@xmath24 plane in four slices of the considered parameter space regarding @xmath22 and @xmath11 @xcite .
we show the 95% cl lower exclusion limits from mono - jet searches at atlas ( blue lines ) and cms ( red lines ) for the simplified model ( solid lines ) and the eft approximation ( dashed lines ) .
further , we show 90% cl lower exclusion limits from the icecube neutrino observatory ( green lines ) . the dark grey shaded band denotes the region where the thermal relic density matches the dm density within @xmath25 . in the light - grey shaded region above it dm is over - produced . in the blue shaded region in the left panels
no solution exist for the individual couplings @xmath12 requiring @xmath26 and @xmath11 .
the orange shaded regions are excluded by searches for resonances in di - jet signatures taken from ref .
@xcite.,title="fig:",scaledwidth=40.0% ] ( 0.123,0.415)@xmath19 , @xmath20 ( 0.26,0.0)@xmath28 $ ] ( 0.05,0.16 ) ( 0.342,0.09 ) ( 0.2,0.224 ) ( 0.15,0.175 ) ( 0.14,0.062 ) ( 0.413,0.295 ) ( 0.02,0.08 ) ( 0.39,0.188 ) ( 0.386,0.17 ) ( 0.39,0.152 )
if wimps scatter in heavy objects like the sun , they can loose enough energy to become gravitationally trapped and accumulate inside the sun .
this leads to a locally enhanced wimp density providing significant dm annihilation .
neutrinos that are produced as primary or secondary products of such annihilations can escape the sun and be detected on earth . on large time - scales an equilibrium between the capturing and annihilation will be be reached . in this case , a limit on the flux of neutrinos produced in wimp annihilations can be translated into a limit on the scattering cross section of wimps inside the sun . as the sun contains large amounts of hydrogen it provides sensitivity to spin - dependent wimp - proton scattering .
we use data from the icecube neutrino observatory @xcite which is interpreted in two benchmark scenarios according to dm annihilating 100% into @xmath29 or @xmath30 . in most of the parameter space of our model ,
annihilation into @xmath29 , @xmath31 or @xmath32 dominates .
therefore we re - interpret the limits on the spin - dependent wimp - proton cross section from ref .
@xcite and estimate a limit for 100% annihilation into @xmath31 and @xmath32 as a function of @xmath7 ( and @xmath8 ) @xcite .
we then apply the limit to our model parameter space taking into account the respective contribution to annihilation of the channels @xmath29 , @xmath31 and @xmath32 .
the resulting limits are shown in fig .
[ fig : mainres ] ( green lines ) . in the region of large @xmath7
where lhc searches loose sensitivity , the limits from icecube are able to exclude mediator masses up to @xmath33 for @xmath34 .
although the capture of wimps in the sun is well described by eft , the annihilation process is in general not . as a consequence of the different solutions for the individual couplings @xmath2 and @xmath3 for different mediator widths chosen in our parameter slices the relative contributions of the annihilation channels @xmath35 ,
@xmath36 and @xmath32 can differ drastically . since limits for annihilation into @xmath32 are much weaker than for @xmath36 , a large @xmath32 contribution can significantly weaken the icecube limit .
this is the case for @xmath27 , @xmath18 and @xmath37tev , where annihilation into @xmath32 is particularly important .
we have considered a model with a vanishing spin - independent wimp - nucleon cross section and set new limits on the model parameter space from lhc mono - jet searches as well as icecube . using lhc limits for @xmath38
we can exclude mediator masses up to around 3tev for @xmath39 while for @xmath40 we exclude @xmath8 in the range of 500gev to 1.5tev with a strong dependence on the mediator width .
we have compared these limits to the ones obtained in the eft and found that those can both over- or under - estimate the sensitivity depending on the corner in parameter space .
limits from icecube are complementary probing the region of large @xmath7 where the lhc is not sensitive at all reaching up to @xmath41 for @xmath38 .
o. buchmueller , m.j .
dolan , c. mccabe , `` beyond effective field theory for dark matter searches at the lhc '' , http://dx.doi.org/10.1007/jhep01(2014)025[jhep * 1401 * ( 2014 ) 025 ] , http://arxiv.org/abs/1308.6799[[arxiv:1308.6799 [ hep - ph ] ] ] . g. busoni , a. de simone , j. gramling , e. morgante , a. riotto , `` on the validity of the effective field theory for dark matter searches at the lhc , part ii : complete analysis for the s - channel '' , http://dx.doi.org/10.1088/1475-7516/2014/06/060[jcap * 1406 * ( 2014 ) 060 ] , http://arxiv.org/abs/1402.1275[[arxiv:1402.1275 [ hep - ph ] ] ] .
o. buchmueller , m.j .
dolan , s.a .
malik , c. mccabe , `` characterising dark matter searches at colliders and direct detection experiments : vector mediators '' , http://dx.doi.org/10.1007/jhep01(2015)037[jhep * 1501 * ( 2015 ) 037 ] , http://arxiv.org/abs/1407.8257[[arxiv:1407.8257 [ hep - ph ] ] ] .
icecube collaboration , m. aartsen , _ et al .
_ , `` search for dark matter annihilations in the sun with the 79-string icecube detector '' , http://dx.doi.org/10.1103/physrevlett.110.131302[phys.rev.lett .
* 110 * ( 2013 ) no . 13 , 131302 ] , http://arxiv.org/abs/1212.4097 [ [ arxiv:1212.4097 [ astro-ph.he ] ] ] .
atlas collaboration , g. aad , _ et al .
_ , `` search for new phenomena in final states with an energetic jet and large missing transverse momentum in pp collisions at @xmath42 = 8 tev with the atlas detector '' , http://dx.doi.org/10.1140/epjc/s10052-015-3517-3[eur.phys.j . * c75 * ( 2015 ) no . 7 , 299 ] , [ erratum : eur.phys.j
. * c75 * ( 2015 ) no . 9 , 408 ] , http://arxiv.org/abs/1502.01518[[arxiv:1502.01518 [ hep - ex ] ] ] .
cms collaboration , v. khachatryan , _ et al .
_ , `` search for dark matter , extra dimensions , and unparticles in monojet events in proton - proton collisions at @xmath42 = 8 tev '' , http://dx.doi.org/10.1140/epjc/s10052-015-3451-4[eur.phys.j .
* c75 * ( 2015 ) no . 5 , 235 ] , http://arxiv.org/abs/1408.3583[[arxiv:1408.3583 [ hep - ex
j. heisig , m. krmer , m. pellen , c. wiebusch , `` constraints on majorana dark matter from the lhc and icecube '' , http://arxiv.org/abs/1509.07867[[arxiv:1509.07867 [ hep - ph ] ] ] .
m. chala , f. kahlhoefer , m. mccullough , g. nardini , k. schmidt - hoberg , `` constraining dark sectors with monojets and dijets '' , http://dx.doi.org/10.1007/jhep07(2015)089[jhep * 1507 * ( 2015 ) 089 ] , http://arxiv.org/abs/1503.05916[[arxiv:1503.05916 [ hep - ph ] ] ] . | we present an interpretation of searches for dark matter in a simplified model approach . considering majorana fermion dark matter and a neutral vector mediator with axial - vector interactions we explore mono - jet searches at the lhc and searches for neutrinos from dark matter annihilation in the sun at icecube and place new limits on model parameter space .
further , we compare the simplified model with its effective field theory approximation and discuss the validity of the latter one . |
chaotic scattering is a phenomenon which appears in many scientific fields , such as astronomy , electromagnetism , statistical mechanics , chemistry , quantum mechanics , just to mention a few .
two objects that are separated initially by some distance , come closer , interact with each other for a while and then they separate again .
more specifically , in astronomy , chaotic scattering can be found in many situations , such as , for example , in the formation and evolution of the outer solar system ( e.g. [ 1 ] ) , in planetary rings ( e.g. [ 11 ] ) , in exosolar planetary systems ( e.g. [ 4 ] ) , in the dynamical evolution of globular clusters ( e.g. [ 3 ] ) . regarding the last situation ,
stars are expected to escape from a globular cluster due to a variety of dynamical processes , such as two body relaxation or an external tidal field .
this theoretical prediction has also been confirmed by observations ( e.g. [ 9 ] ) .
it is important to know whether a star will escape from a cluster and also the timescale on which that will happen as it affects the scaling of the n - body simulation results ( normally , the number n of stars used by a simulation is less than the number of stars of the cluster under study ) and reduce the computational effort required to perform a simulation . also , the timescale of escape is important when one uses monte carlo models in order to study the dynamical evolution of globular clusters .
further reasons regarding the importance of escape in globular clusters are discussed in [ 6 ] .
a star can escape from a cluster when its energy exceeds some critical threshold but escape may not occur immediately .
numerical simulations show that it may take a long time for a star to escape or it may not even escape at all ( escape on a timescale comparable with the age of the universe falls into this category too ) .
the assumption of rapid escape referred to in [ 5 ] and [ 2 ] is therefore a significant source of error for theoretical predictions regarding the evolution of globular clusters .
interest has therefore been generated in the dynamics underpinning the escape mechanism .
petit and h@xmath0non [ 10 ] investigated numerically the motion of two satellites around a planet . however , certain difficulties arose in their study and in order to overcome those difficulties , h@xmath0non [ 8 ] devised a model problem which demonstrated similar behaviour to the satellite problem but it was easier to study .
that model consisted of an inclined billiard , i.e. a point particle bouncing on two fixed disks and a two dimensional mapping was used to describe the dynamics of the model . in the present paper , we extend the results obtained in [ 8 ] by deriving an analytical expression for the escape distribution of a particular class of orbits , the so called h - orbits .
the structure of the paper is roughly as follows : first , we give a description of the model developed by h@xmath0non ( sec .
then we obtain an analytical expression for the escape distribution of the h - orbits and the formula is compared with results from numerical simulations ( sec .
next , there is some discussion about the billiard model and hill s problem , which originally inspired h@xmath0non to devise the inclined billiard model ( sec .
finally , the last section of the paper gives a brief summary .
h@xmath0non s model is described as follows : a point particle moves in the ( x , y ) plane and bounces elastically on two fixed disks of radius @xmath1 whose centres are located at @xmath2 and @xmath3 .
in addition , the particle is subject to a constant acceleration @xmath4 in the negative @xmath5 direction . for reasons of simplicity , @xmath1 is considered to be large , an assumption which means that the two disks overlap . a one parameter family of orbits , the so called h - orbits , is defined by assuming that the particle is dropped from rest at @xmath6 .
@xmath7 is a positive constant which fixes the energy and @xmath8 is a variable . for those orbits ,
there are intervals of @xmath8 in which the orbit changes continuously and there are also critical values of @xmath8 at which a transition in behaviour is observed .
for @xmath9 the particle bounces ad infinitum on the right disk , while for @xmath10 the particle bounces ad infinitum on the left disk .
if the particle is dropped either to the left of the left disk or to the right of the right disk , then the particle escapes and never returns .
if the particle is dropped either to the right of the left disk or to the left of the right disk , then the particle exhibits a more complex behaviour , bouncing from one disk to another .
hence , the values @xmath11 are associated with a transition . in general , transitions are present at values of @xmath8 which lead to solutions which approach asymptotically one of the @xmath11 orbits .
a convenient way of dealing with the problem is to introduce a two dimensional explicit mapping , which is equivalent to studying a hamiltonian system with two degrees of freedom .
details about the mapping and its properties can be found in [ 8 ] . if @xmath12 is the position of the particle on the @xmath13th rebound , then the two dimensional mapping , keeping the notation of [ 8 ] , is the following : @xmath14 @xmath15 is defined by @xmath16 where @xmath17 is the transverse velocity , @xmath18 is the total energy and @xmath19 is a dimensionless parameter defined by @xmath20 finally , @xmath21 .
if the mapping is applied repeatedly , we obtain : @xmath22}}{\cosh{\frac{\phi}{2}}}-2\tanh{\frac{\phi}{2 } } \sum^{n-1}_{j=1}s_{j}\sinh{[(n - j)\phi]},\\ w_{n } & = & \frac{h\sinh{[(n-\frac{1}{2})\phi]}}{\cosh{\frac{\phi}{2}}}-2\tanh{\frac{\phi}{2 } } \sum^{n-1}_{j=1}s_{j}\sinh{[(n - j)\phi]}-s_{n}\tanh{\frac{\phi}{2}}\nonumber.\end{aligned}\ ] ] a given h - orbit is associated to a sequence of binary digits ( 0 or 1 depending on whether the particle bounces on the left or right disk respectively ) and the corresponding to that sequence number a , with @xmath23 .
an h - orbit changes continuously within an interval of continuity . in that interval ,
the sequence of rebounds remains the same throughout the whole of this interval .
this means that the value of a is constant .
if a is plotted as a function of h ( for a fixed value of @xmath19 ) a fractal picture is generated .
this has the appearance of the devil s staircase , which consinsts of an infinite number of horizontal bars .
each bar corresponds to an interval of continuity , meaning that when a particle starts within that interval , it will follow the same orbit in terms of which disk it bounces on .
such an example is given in fig.1 .
after the brief introduction to h@xmath0non s model , we are going to obtain an expression for the number of bounces it takes for an h - orbit to escape from the system .
the derivation will apply for @xmath24 or @xmath25 . when @xmath26 , the devil s staircase is not continuous any longer and gaps appear between the horizontal bars , i.e. there exist values of @xmath27 to which no h - orbits correspond .
also , many of the assumptions made about the h - orbits do not hold .
more details about that situation can be found in [ 8 ] , although , for reasons not known to us , h@xmath0non only refers to properties regarding @xmath28 .
the mapping given by eqs ( [ eq1 ] ) and ( [ eq2 ] ) has two fixed points ( @xmath29 and @xmath30 ) , each one with eigenvalues @xmath31 and @xmath32 , which suggests that when @xmath33 , we get at the same time @xmath34 . for @xmath35 ,
all the points of the mapping are fixed points . however , as stated in [ 8 ] , it is not clear whether the peculiarities of the @xmath26 case have general relevance to the problem of chaotic scattering .
the time to escape , for a given value of h , is the smallest integer @xmath36 such that @xmath37 for all @xmath38 . in particular
, we would like to know which values of h , i.e. which sub - intervals of @xmath39 , correspond to orbits which escape after at least @xmath36 bounces within the constrained region ( @xmath40 ) . by summing up the lengths of these sub - intervals it is possible to produce an escape distribution . according to [ 8 ] , at each end of every horizontal bar
there are left and right asymptotic orbits corresponding to @xmath41 and @xmath42 respectively , where @xmath43 is associated with each horizontal bar ( p - bar ) , indicating the number of digits which occurred in the binary sequence before the repetition of either zeros or ones , i.e. before the particle escapes over the left or right disk .
those orbits take infinitely many bounces to escape .
if we move away from those extremes and consider @xmath44 or @xmath45 , we find orbits which escape after a finite number of bounces .
the smaller @xmath46 is , the longer it takes for an orbit to escape ( more precisely , @xmath47 ) . for values of h in the interval @xmath48 , for a given p - bar
, the number of bounces never falls below p. in order to find those intervals of @xmath8 which correspond to orbits that remain bound for at least @xmath36 bounces , we require to sum together the full lengths of the horizontal bars for which @xmath49 .
only partial contributions will be required from the bars with @xmath50 .
this is because some values of @xmath8 near the centre of these bars lead to orbits which escape too rapidly ( under @xmath36 bounces ) .
the whole calculation may be performed in the following three steps : + i ) calculating the partial contribution from a bar with @xmath50 by finding an expression for the time to escape , @xmath36 , as a function of @xmath46 .
this expression may be inverted to yield @xmath46 as a function of @xmath36 .
we refer to @xmath51 or @xmath52 depending on which end is being considered . the total partial contribution from each bar
is given by @xmath53 .
+ ii ) summing together the partial contributions ( of which there are @xmath54 ) from all the p - bars for each @xmath50 .
+ iii ) considering the contributions from all bars with @xmath49 ( the length of a bar is @xmath55 ) .
we start with the first step , calculating an expression for @xmath56 by considering the values of h which lead to left and right asymptotic orbits .
substituting eq .
( [ eq4 ] ) into eq .
( [ eq3 ] ) , we obtain @xmath57 \frac{\cosh{[(n-\frac{1}{2})\phi]}}{\cosh{\frac{\phi}{2}}}-\nonumber\\ & - & 2\tanh{\frac{\phi}{2}}\sum^{n-1}_{j=1}s_{j}\sinh{[(n - j)\phi]}\nonumber.\end{aligned}\ ] ] if we expand the above equation ( keep in mind that @xmath58 and @xmath59 for all @xmath60 since we deal with a left escaping orbit ) , we end up with @xmath61 where @xmath62.\ ] ] now suppose that @xmath63 and that the @xmath64 s are the same as in eq .
( [ eq4 ] ) , i.e we are looking at the same bar . then , from ( [ eq3 ] ) , we get @xmath65}}{\cosh { \frac{\phi}{2}}}-2\tanh{\frac{\phi}{2}}\sum^{n-1}_{j=1}s_{j}\sinh{[(n - j)\phi]}=\nonumber\\ & = & x_{n}-\delta_{l}\frac{\cosh{[(n-\frac{1}{2})\phi]}}{\cosh{\frac{\phi}{2}}}= -1+e^{-n\phi}l_{l}-\delta_{l}\frac{\cosh{[(n-\frac{1}{2})\phi]}}{\cosh{\frac{\phi}{2}}}. \label{eq5}\end{aligned}\ ] ] as we are interested in the number of bounces @xmath36 before the particle escapes over the left disk , eq .
( [ eq5 ] ) yields ( @xmath66 ) @xmath67 following the above approach , we can obtain a similar expression for right asymptotic orbits . in this case , we find that @xmath68 where @xmath69\ ] ] and eventually , @xmath70 now , we continue with the second step , i.e. summing together all the partial contributions . in order to do that , we are going to use two subscripts .
the first subscript is @xmath43 , which was defined earlier , while the second subscript @xmath71 distinguishes between bars of the same p - value . in the case of bars with @xmath50
, we must evaluate @xmath72,\ ] ] where @xmath73 @xmath74 @xmath75,\ ] ] and @xmath76.\ ] ] when we substitute the above equations into eq .
( [ eq6 ] ) , the following quantity will appear : @xmath77+\nonumber\\ & + & \sum^{2^{p-1}}_{q=1}\sum^{p-1}_{j=1}(e^{\phi}e^{-j\phi}+e^{j\phi})s_{qj}\tanh{\frac{\phi}{2}}.\nonumber\end{aligned}\ ] ] the same expression arises for @xmath78 , except a minus sign before the double sum .
the double summation term is zero because , for a given @xmath13 , @xmath79 in equal numbers of bars .
hence : @xmath80=\frac{e^{\phi}+1}{e^{2k\phi}+e^{\phi } } \sum^{k-1}_{p=1}\sum^{2^{p-1}}_{q=1}[(l_{l})_{pq}+(l_{r})_{pq}]=\frac{2^{k}[e^{(k-1)\phi}-e^{-(k-1)\phi}]}{1 + e^{(2k-1)\phi}}. \label{eq7}\ ] ] finally , the contribution from all bars with @xmath49 is @xmath81 thus , @xmath82 , the fraction of h - orbits in the interval @xmath39 which escape after at least k bounces , is found by adding eq . ( [ eq7 ] ) and eq . ( [ eq8 ] ) and dividing by 2 : @xmath83}{1 + e^{(2k-1)\phi}}+\frac{2^{k}e^{-k\phi}(e^{\phi}-2)}{1 - 2e^{-\phi}}\right\}=2^{k-1}\frac{e^{k \phi}(1+e^{\phi})}{e^{\phi}+e^{2k \phi}}. \label{eq9}\ ] ] in order to test our analytical result , we advanced eq .
( [ eq1 ] ) and ( [ eq2 ] ) numerically .
the interval @xmath39 was divided into n subintervals , each having a width of @xmath84 .
the initial conditions were @xmath85 with the initial values @xmath86 distributed uniformly in the interval @xmath39 . using the iterative scheme given by eq .
( [ eq1 ] ) and ( [ eq2 ] ) , @xmath87-orbits were evolved forward in time and for each @xmath87-orbit , we found the minimum number of iterations @xmath88 required to satisfy @xmath89 .
an escape distribution is built up by determining what fraction of the @xmath87-orbits escaped on the first bounce , on the second bounce and so on .
the simulations were done for different values of @xmath19 and @xmath90 .
the value of @xmath91 was even increased to @xmath92 to ensure that the results were independent of the number of subdivisions .
fig.2 shows the escape distribution for different values of @xmath19 , both analytical [ eq .
( [ eq9 ] ) ] and numerical .
the numerical results are in excellent agreement with those obtained from eq .
( [ eq9 ] ) .
there is only a small range of @xmath31 for which there is some discrepancy between the numerical and the analytical results for some p - bars .
every horizontal bar has a point @xmath93 , which , according to [ 8 ] , is given by @xmath94}}\sum^{n-1}_{j=1}s_{j}\sinh{[(n - j)\phi]}\ ] ] and for which @xmath95 .
when @xmath96 or @xmath97 , our analytical result needs some correction , i.e. the difference in bar length @xmath98 or @xmath99 .
for example , also visible in fig.2 , for @xmath28 and @xmath100 our analytical result is around @xmath101 larger than the numerical one and that happens because our calculation overestimates the contribution of the left side of the @xmath102 bar ( for @xmath103 ; for @xmath104 is the right side of the @xmath102 bar that exhibits some problem ) . for @xmath105 ,
the error for @xmath100 reduces to around @xmath106 .
if the extra bar length is subtracted from our analytical calculation , we obtain the correct percentage . for the @xmath102 bar ,
the correction is : @xmath107
as it was stated in the introduction , h@xmath0non s billiard model was inspired by [ 10 ] , which dealt with the interaction of two satellites around a planet .
that study was done in the context of the so called hill s problem , a special case of the restricted three body problem where the massless particle moves in the neighbourhood of the secondary body .
originally intended as a model for the motion of the moon around the earth with perturbations by the sun , with some modification it can also serve as a simplified model of the dynamical behaviour of escaping stars in globular clusters [ 7 ] . in that case , the centre of the galaxy and the globular cluster play the role of the two main bodies , while the star is treated as the massless particle .
let us assume that the cluster moves on a circular orbit of radius @xmath108 around the centre of the galaxy with constant angular velocity , the mass of the cluster with respect to the galaxy is assumed to be small , the mass of the star is considered to be negligible compared to that of the cluster and all three bodies lie in the same plane .
if we consider a coordinate system that rotates with the angular velocity of the cluster around the galaxy and with its origin being at the cluster , hill s equations can be written as follows [ 13 ] : @xmath109 where @xmath110 and @xmath111 are the @xmath112 and @xmath113 coordinates of the star and @xmath114 .
note that the above equations have been scaled in terms of time and distance . in the above mentioned coordinate system ,
the two lagrangian points of interest occur at @xmath115 and @xmath116 .
one recalls that in our inclined billiard model , the particle was dropped with zero horizontal velocity onto the inclined surface from a fixed height , i.e. with constant @xmath19 for various @xmath8 in the interval ( -1,1 ) . in hill
s problem , the h - orbits can be defined to be those with initial conditions given by @xmath117 at a fixed value of the initial hamiltonian @xmath118 ( analogous to the condition of constant @xmath19 in h@xmath0non s model ) , @xmath119 is calculated from @xmath120 each time the orbit intersects the surface of section @xmath121 ( with @xmath122 when @xmath123 and with @xmath124 when @xmath125 ) , the value of @xmath110 is recorded . if @xmath126 on the @xmath13th intersection with the surface of section , then the orbit may be represented symbolically by a sequence @xmath127 , where @xmath128 if the orbit escapes through the right lagrangian point after @xmath36 intersections with the surface of section , then @xmath129 .
similarly , if the orbit escapes through the left lagrangian point after @xmath36 intersections with the surface of section , then @xmath130 . hence a real number can be attached to a given orbit in a similar way as it was done for the inclined billiard model .
fig.3 is an example of an h - orbit in hill s problem , exhibiting similar features to those found in fig . 1 .
h@xmath0non [ 8 ] used an inclined billiard in order to investigate the phenomenon of chaotic scattering . a point particle , initially at rest , fell from a certain height and bounced elastically off the surface of two disks .
when the particle bounced beyond the top of either disk , it was considered to have escaped . by assigning a number @xmath27 to the orbit of the particle
, we were able to plot @xmath131 , where @xmath8 ( @xmath132 ) was the initial value of the x - coordinate of the particle .
for certain values of the parameter @xmath19 , which was related to the particle energy , the graphical representation of @xmath131 had the form of the so - called devil s staircase .
based on certain properties of the devil s staircase , we were able to derive an analytical expression for the number of particles that escaped after at least @xmath36 bounces .
it may be possible to approach the issue of stars escaping from a cluster or other similar situations by using the billiard model as a simplification of hill s problem .
therefore , our future aim is to investigate the possibility of connecting the billiard model and hill s problem in a way that can be used to describe a more complex situation such as for example the dynamical evolution of a star within a globular cluster .
the latter is far more complicated as a system than the inclined billiard we have studied , but many times , it is possible to get a good approximation of a complicated system by following a more simple approach to the problem .
the authors would like to thank prof .
heggie for valuable comments regarding certain aspects of this work .
alan roy should also like to thank prof .
jan sykulski for supporting his research visitor status in ecs at the university of southampton .
astakhov , s.a . , lee , e.a . , farrelly , d. , formation of kuiper - belt binaries through multiple chaotic scattering encounters with low - mass intruders , mnras , 2005 , vol .
401 - 415 .
heggie , d.c . , mass loss from globular clusters , _ dynamics of star clusters and the milky way _ ,
s. deiters , b. fuchs , a. just , r. spurzem , r. wielen ( eds . ) , san francisco : astron .
soc . pac .
, 2001 , asp conf .
29 - 41 .
heggie , d.c . , escape in hill s problem , _ the restless universe , proceedings of the 54th scottish univ .
summer school in physics _ ,
steves , a.j .
maciejewski ( eds . ) , bristol : scottish universities summer school in physics and iop publishing , 2001 , pp . 109 - 128 .
bounces against the number of bounces @xmath36 . from right to left , the first two curves correspond to @xmath28 ( the left curve comes from the simulations , while the right one is based on eq.([eq9 ] ) ) , the third one corresponds to @xmath134 and the fourth one corresponds to @xmath135 .
note that each of the third and fourth curves are actually two curves ( one on top of the other one ) , as the numerical and analytical results are almost identical .
also note the small discrepancy between the numerical and analytical results for @xmath28 . ] against @xmath136 in hill s problem , where @xmath137 .
the initial value of the hamiltonian is @xmath138 .
the results for hill s problem were obtained by integrating equations ( [ eq10 ] ) and ( [ eq11 ] ) numerically , using a burilsch - stoer integrator with a variable time step [ 12 ] .
the time of integration was @xmath139 which , in our system of units , is twice the orbital period of the cluster around the galaxy . ] | h@xmath0non [ 8 ] used an inclined billiard to investigate aspects of chaotic scattering which occur in satellite encounters and in other situations .
his model consisted of a piecewise mapping which described the motion of a point particle bouncing elastically on two disks .
a one parameter family of orbits , named h - orbits , was obtained by starting the particle at rest from a given height .
we obtain an analytical expression for the escape distribution of the h - orbits , which is also compared with results from numerical simulations .
finally , some discussion is made about possible applications of the h - orbits in connection with hill s problem .
+ + chaotic scattering , inclined billiards , hill s problem + |
among different double - logarithmic ( dl ) contributions , there are the contributions that depend on the infrared cut - off
. results of resuming them to all orders in the coupling(s ) can be often written in an exponential form , with a negative exponent , so that it falls when the total energy grows .
this is known as the sudakov suppression @xcite .
it means the suppression of cross - sections of non - radiative processes compared to cross sections of the processes where bremsstrahlung is allowed .
such a suppression in qed ( see e.g review @xcite ) and in qcd@xcite was studied in details to all orders in the couplings .
exponentiation of the first - loop electroweak ( ew ) contributions was obtained in ref .
@xcite and then was confirmed in refs .
@xcite-@xcite by the fixed orders calculations .
the complete list of publications on this subject can been found in the recent refs .
@xcite,@xcite .
however , resumming dl contributions of the same infrared origin sometimes leads to different results .
in particular , it was found in ref .
@xcite that one - loop dl contribution to the @xmath2 has the opposite sign compared to the one for @xmath3 and @xmath0 -decays .
it suggests that the total resummation of the dl corrections to the amplitude of the @xmath2 may be an exponential with a positive exponent . in the present paper we consider amplitudes for the @xmath4 -decay .
we show that the result of accounting for dl contributions to all orders in the ew couplings has an exponential form with the positive exponent . in other words
, there is the enhancement instead of the suppression for such amplitudes .
we consider the decays of the electroweak bosons into the left fermions @xmath5 and the right antifermions @xmath6 though drop the subscripts @xmath7 and @xmath8 through the paper .
the paper is organised as follows : in sect .
ii we briefly review the sudakov suppression for amplitude of @xmath9 , comparing the use the feynman and the coulomb gauges . in sect .
iii we discuss the decay of @xmath4 -bosons into fermions , again with using both the feynman and the coulomb gauges , and arrive at the enhancement .
the origin of the enhancement is considered in in sect .
finally , sect .
v is for concluding remarks .
the sudakov suppression@xcite is well - known , so we discuss it very briefly .
let us consider amplitude @xmath10 of the decay @xmath11 where @xmath5 stands for any ( charged ) fermion from the left doublets of the standard model . as we assume all external particles to be on - shell , @xmath10 is gauge - independent . among the often used gauges for propagators @xmath12 of the ew bosons , there are the feynman gauge and the coulomb ( temporal ) gauge . in the first - loop approximation and
when the feynman gauge is used , the dl contribution to @xmath10 comes from the graph in fig . 1 .
the vertical waved line in fig .
1 denotes the external ew boson .
the horizontal waved line may correspond to any virtual electroweak boson though the dl contribution to @xmath10 appears only when this boson is a photon .
when the coulomb ( temporal ) gauge is used , dl contributions can come from the self - energy graphs depicted in fig . 2 .
again , dl contributions come only when the virtual bosons in fig .
2 are photons . using the feynman gauge
, it is easy to arrive at the well - known one - loop dl contribution @xmath13 to @xmath10 : @xmath14 where @xmath15 is the electric charge of the fermion @xmath5 , @xmath16 is the mass of the @xmath17 -boson and @xmath18 is the infrared cut - off . with dl accuracy , one can neglect the difference between masses of @xmath17 and @xmath4 -bosons . the double logarithm in eq .
( [ rz ] ) is written in the simplest way for the particular case when @xmath19 , with @xmath20 being the mass of the fermion @xmath5 .
it also easy to arrive at eq .
( [ rz ] ) when the coulomb gauge is used . in this case , dl contributions come from the self - energy graphs depicted in fi . 2 .
the dl contributions @xmath21 from the fermion self - energy graphs ( graphs ( b ) and ( c ) in fig . 2 )
are @xmath22 . in order to get a dl contribution from graph ( a ) in fig . 2 , at least one of the virtual bosons in fig .
2a has to be a photon . however , there is not such a vertex in the lagrangian of the standard model .
it leaves us with two contributions @xmath23 and we are back to eq .
( [ rz ] ) . using the coulomb gauge for the sudakov logarithms
was discussed in details in refs .
the dl asymptotics of @xmath10 accounting for contributions both @xmath24 and @xmath25 can be obtained by direct graph - by - graph calculations like it was done in ref .
the complete expression for the dl asymptotics for for @xmath10 is @xmath26 e^{- q^2_f ( \alpha/4\pi ) \ln^2\rho } \ ] ] where we have omitted the spinors of the fermions and the @xmath17-boson polarization vector , @xmath27 is the sum of the fermion momenta @xmath28 ; we have chosen in eq .
( [ zdecay ] ) @xmath29 and denoted @xmath30 .
the minus sign in the exponential of eq .
( [ zdecay ] ) means the sudakov suppression .
it suppresses the non - radiative decays of @xmath17 -bosons .
now let us consider dl contributions to amplitude @xmath31 of the decay @xmath32 where @xmath5 and @xmath33 @xmath34 belong the same doublet of the standard model .
when the feynman gauge is used , dl contribution comes from the graph in fig .
1 where the horizontal waved line denotes again the virtual photon .
this contribution is @xmath35 where @xmath36 is the electric charge of the fermion @xmath37 . invoking results of ref .
@xcite leads immediately to the dl asymptotics for @xmath31 : @xmath38 e^{- q_fq_{f ' } ( \alpha/4\pi ) \ln^2\rho } .\ ] ] the term @xmath39 in the expansion of eq .
( [ wdecay ] ) into series in @xmath40 was obtained in ref .
@xcite .
let us notice that when the fermions @xmath5 and @xmath33 in eq .
( [ wdecay ] ) are quarks , they have opposite signs .
it means that the exponent in eq .
( [ wdecay ] ) is positive and therefore there is the dl enhancement in eq .
( [ wdecay ] ) instead of the dl suppress of eq .
( [ zdecay ] ) . for reactions
@xmath41 and @xmath42 , the dl amplitude of the process is @xmath43 notation @xmath44 in eq .
( [ wdecayq ] ) stands for the expression in the squared brackets in eq .
( [ wdecay ] ) . at the same time , eq .
( [ wdecay ] ) reads that there are no dl contributions to the lepton decays @xmath45 . although amplitudes @xmath31 are gauge - invariant , it is useful to demonstrate how the first - loop contribution @xmath46 of eq .
( [ woneloop ] ) can be easily obtained when the coulomb gauge is used . in this case , dl contributions come from the self - energy graphs .
contributions of graphs ( b ) and ( c ) in fig
. 2 are @xmath22 and @xmath47 respectively .
contrary to the case of the @xmath17 -boson decay , graph ( a ) in fig .
1 now yields a dl contribution @xmath48 and therefore @xmath49= \nonumber \\
a_w^{born } [ q^2_w - q^2_f - q^2_{f ' } ] l/2 ~. \end{aligned}\ ] ] the electric charge conservation states that @xmath50 .
it allows to rewrite eq .
( [ quarks ] ) in the form of eq .
( [ woneloop ] ) .
both the sudakov suppression of eq .
( [ zdecay ] ) and the enhancement of eq .
( [ wdecay ] ) have a very simple origin in terms of the algebra of the gauge group generators .
we explain it , first using the feynman gauge and then proceed to the case of the coulomb gauge . to begin with ,
let us remind that all electroweak boson fields @xmath51 are linear combinations of the gauge isoscalar field @xmath52 and isovector fields @xmath53 : @xmath54 where @xmath55 is the weinberg angle .
therefore , propagator @xmath56 of the isocalar field @xmath52 and propagator @xmath57 of the isovector field @xmath58 can be expressed in terms of the photon propagator @xmath59 and the @xmath17 -boson propagator @xmath60 . as dl contributions at the energies
@xmath61 come from the photon propagators only , @xmath62 the @xmath63 -parts of the vertexes of the interaction of fields @xmath64 and @xmath52 with fermions are @xmath65 and @xmath66 , with @xmath67 being the @xmath68 -generators @xmath69 , @xmath70 being the fermion hypercharge and @xmath71 being the electroweak couplings ( @xmath72 ) .
now let us choose the feynman gauge for the photon propagator and consider again the one - loop graph in fig . 1 .
we can regard both the external and virtual electroweak bosons in this fig . as consisting of the gauge fields @xmath73 . in order to discuss both the @xmath17 and the @xmath4 -decays at the same time ,
let us suppose that the external field @xmath74 consists of the isoscalar and isovector fields : @xmath75 where @xmath76 ; @xmath77 and @xmath27 are some scalar factors .
then the born amplitude can be written as @xmath78 and the graph in fig .
1 yields the following dl contribution : @xmath79 ( 1/16 \pi^2 ) \ln^2(-m^2/\mu^2 ) ~.\end{aligned}\ ] ] therefore , @xmath80 where @xmath81 eq .
( [ t12 ] ) leads to the suppression when @xmath82 and to the enhancement when @xmath83 provided @xmath84 .
the exponentiation of dl contributions in terms of the gauge fields @xmath85 can be proved easily , e.g. by composing the infrared evolution equations for amplitudes @xmath86 in the same way as it was done in ref .
@xcite for calculating the dl asymptotics of the form factors of electron ( abelian gauge group ) and quark ( non - abelian gauge group ) . eq .
( [ t12 ] ) reads that the first - loop dl contributions from exchange of the virtual isoscalar boson ( the contributions @xmath87 ) always lead to the suppression whereas the exchanges of the isovector bosons lead either to the suppression or to the enhancement due to non - commutativity of the generators of the gauge group . in order to obtain this result when the coulomb gauge is used , it is enough to notice that the only difference between dl contributions of the graphs in fig . 2 compared to eqs . ( [ sigmaw],[quarks ] )
is in @xmath68 -group factors .
after accounting for them , one reproduces eq .
( [ t12 ] ) .
we have considered the decays of the massive electroweak bosons into fermions in the dl approximation . we have shown that when the ew bosons are charged , accounting for the dl radiative corrections leads to exponentiation of the first - loop contribution , however with the positive exponent , i.e. instead of the sudakov suppression there is the enhancement for such decays . on the other hand ,
accounting for the soft photon emision accompanying such decays leads to the falling exponenentials . therefore
, these dl contributions cancel each other in expressions for inclusive cross - sections .
the results we obtained can be useful for analyses of @xmath88 processes in theories with spontaneously broken non - abelian gauge groups which involve different mass scales for charged massive bosons .
the work is supported by grant cern / fis/43652/2001 .
carazone , e.c .
poggio and h.r . quinn .
d11(1975)2286 ; j.m .
cornwall and g. tiktopolous .
35(1975)338 ; phys .
d13(1976)3370 ; j. frenkel and j.c.taylor .
b116(1976)185 ; v.v .
belokurov and n.i .
b94(1980)251 ; theor . math . phys .
44(1980)657 ; ibid 45(1980)957 . | accounting for the double - logarithmic ( dl ) contributions to amplitude of @xmath0 leads to the sudakov suppression .
in contrast , resumming dl contributions to amplitudes of @xmath1 results into an exponential with a positive exponent , i.e. into the enhancement . this effect is intrinsic property of the theories with non - abelian gauge groups . |
in the recent years there is a remarkable progress in the technology of the nanofabrication of well defined magnetic materials .
the material nanoscience based on the epitaxial and lithographic techniques @xcite allows the fabrication of the regular arrays of the magnetic particles - dots of well controlled and interesting shape @xcite , lattice geometry and composition .
the increasing technological flexibility calls for further physical ideas , which should be incorporated into design of the artificial nanoscale magnetic systems .
the uniformity of the polarization is the general basic aspect discussed in the connection with the small magnetic particles .
the concept of uniformly polarized particle is justified only for the particles of an intermediate size @xcite . in the theory
@xcite the magnetostatic coupling was derived for the homogeneously polarized and saturated cylindrical dots on a rectangular lattice .
more restrictive are conditions of the simulation @xcite , where each dot of array is substituted by a single dipolar moment .
this approximation can be used only for monodomain dots separated by a sufficiently large distances . when a dot array is represented by a system of the interacting dipoles , the search for the ground state configuration leads to the formulation typical for the classical dipolar lattices @xcite .
the violation of the intradot homogeneity stems from the competition between the magnetostatic , anisotropy and exchange energy terms .
the analytical model of the dot array going towards the non - uniformity was proposed in @xcite . in this model
the interactions of dots were described by the quadrupolar terms .
the problem of the calculation of the magnetization field of a dot array can be in principal formulated in the terms of classical micromagnetic theory @xcite .
due to complexity of the problem , the important role in its treatment will play the numerical simulations .
they require the implementation of the sufficiently dense discretization within the each ferromagnetic dot . as usual ,
the magnetic part of the system can be subdivided into interacting dipoles or grains @xcite , small ferromagnetic cubes @xcite or finite elements @xcite .
then the optimum spacing of the mesh nodes is determined by a minimum magnetic length scale ( exchange , wall ) of the system . for the majority of ferromagnetic materials ,
the comprehensive micromagnetic description is attained when the size of discretization elements decreases into the nanometer regime @xcite .
thus , the simulation of a single ultrathin dot of the micrometer size requires about @xmath2 nodes , although , the qualitative simulations can be realized even for @xmath3 nodes @xcite . from this
we can conclude that both detailed and truncated micromagnetic description of many - dot array represents rather demanding computational task .
we summarize , that principal difficulties of micromagnetic dot array analysis come from : ( i ) the interplay of the phenomena on the intradot ( exchange , domain wall ) and external geometric length scales ; ( ii ) complexity of the magnetic structure of the non - uniformly polarized dots ; ( iii ) long - range magnetostatic interdot interactions . to make the problem of the magnetic ordering of dot arrays tractable by a moderate computer facilities , we have developed method , which works on a much coarser mesh than usual discretization schemes allow ( except the adaptive and multigrid methods ) .
its general idea is the simultaneous simulation of the intradot - micromagnetic and multidot scales .
this idea was strongly inspired by the multiscale approach @xcite . at the present stage of the project ,
the multidot part of simulation has been developed separately and the behaviour of the small - intradot scales has been treated only phenomenologically .
the approach allows a remarkable increase of the simulation speed , indeed , the price payed to the scale separation is the appearance of additional parameters .
the completing of the project needs support of the algorithms of the parameter estimation developed on the basis of the standard micromagnetic simulations .
let us to note that similar problems were solved in a cellular automaton version of the molecular dynamics @xcite .
in this paper we presented the results of the simulation of magnetic properties of quasi - two - dimensional cluster of ultra - thin magnetic dots on the square lattice .
the phenomenological aspect of our model is a variable intradot inhomogeneity .
the formalism we developed for this aim has been adopted from the models of the neural networks . from the point of view of the information theory ,
the neural networks are continuous , unique mappings constructed from the system of known activation functions .
the synaptic weights of these activation functions are adjusted by the training process .
the standard problem , which can be effectively solved by the neural networks is the association of the input patterns ( in our case inputs are effective magnetic moments ) with the desired outputs ( magnetic self - energy of dot ) .
there are many applications , where neural networks can be implemented .
they allow interpolation of the data generated by the simulations or experiment .
the example of the physical application is @xcite , where neural network was used to fit a complicated analytic potential to the set of ab initio data . in @xcite ,
the hopfield type of the interaction matrix was suggested to simulate the dynamics of the complex protein molecule .
the specific magnetic application represents the solution of the magnetic inverse problem @xcite . for the purpose to model the variable magnetic intradot inhomogeneity we adopted the theory of the radial basis function networks ( rbfn ) @xcite .
the rbfn variant of the neural network was chosen , because its ingredient is a straightforward and explicit estimation of the synaptic weights , which allows more transparent analysis of the physical symmetries .
the aim of the paper is to make general presentation of the model and present some numerical results .
the paper is organized as follows : in section [ modell ] the model of the dot array energy functional is introduced .
this section consists of two parts : in subsection [ model0 ] we introduced the general phenomenological concept of the nonuniform magnetization , which utilizes self - energy interpolation by means of rbfn approach . in subsection [ model1 ] the interaction between dots is introduced . in section [ square ]
our method is applied to the ultrathin square dots , where tendency of the formation of the vortex intradot phase prevails .
section [ simull ] provides some details about the implementation of the simulated annealing algorithm to the problem of the total energy minimization .
finally , in section [ numsimull ] we bring examples of the numerical simulations .
the microstate of the system of @xmath4 magnetic dots is described by @xmath5 effective magnetic moments @xmath6 which are associated with the magnetization field of dot @xmath7 via the volume averages @xmath8 the formal integration is performed here over @xmath9th volume element of @xmath10th dot labeled as @xmath11 .
the saturation magnetic moment @xmath12 , where @xmath13 is the saturation magnetization of the dot , plays role of the normalization factor in eq.([eqqq1 ] ) .
thus for the effective magnetic moments we have the bounding @xmath14 where @xmath15 @xmath16 subscripts refer to the cartesian components of the effective magnetic moments @xmath17 , where @xmath18 and @xmath19 are the cartesian unit vectors .
the `` softness '' of @xmath20 expressed by eq.([bounds ] ) is the important model aspect , which differs from the fundamental brown s postulate @xcite .
the reason for this modification is that our model is formulated for sufficiently larger elements than classical micromagnetic approach .
the total magnetization per dot per interaction center is given by @xmath21 the effective moments characterizing @xmath10th dot inside @xmath4-cluster are distributed around the dot center @xmath22 and located at @xmath23 positions @xmath24 where @xmath25 are some relative coordinates of the interaction centers . from the assumption that identical dots are arranged into array it follows that system of @xmath25 vectors is independent of the dot position inside the cluster . in a quasi - two dimensional systems , where @xmath20 is confined to @xmath26 plane , reduced information about @xmath10th dot microstate
is involved in @xmath27 dimensional row vector @xmath28\ , \label{moments } \\ & = & \left[\ , \underbrace{m_{i1,x } , m_{i 1,y}}_{{\rm center}\,\ , { \bf x}_{i 1 } } , \underbrace{m_{i2,x } , m_{i 2,y}}_{{\rm center}\,\ , { \bf x}_{i 2 } } , \ldots\ , , \underbrace{m_{i n_{\rm c},x } , m_{i n_{\rm c},y}}_{{\rm center}\,\ , { \bf x}_{i n_{\rm c}}}\ , \right]\ , . \label{moments01}\end{aligned}\ ] ] we continue with the construction of relations associating the effective dot moments from eq.([moments ] ) with the corresponding self - energies . here
the intradot self - energy is understood as a part of the total energy , which includes only anisotropy , exchange and intradot magnetostatic energy contributions .
the zeeman term and interdot magnetostatic terms , which do not contribute to the self - energy are defined independently of the neural network part of the model .
the construction of the self - energy formula @xmath29 is based on a proper choice of the set of special @xmath30-dimensional memorized vectors ( the input patterns of neural networks ) of the type eq.([moments01 ] ) .
the memorized vectors are constructed by putting into row @xmath23 two - dimensional vectors @xmath31 @xmath32\ , .
\label{moments1}\ ] ] here , the superscript @xmath33 identifies so called feature @xcite . in our case
it is an integer from the set @xmath34 . in the analogy with eq.([moments ] ) , the subscripts of @xmath35 run over the interaction centers @xmath25 . to measure the differences between the configurations we have introduced the euclidean norm @xmath36
written here for some magnetic moment @xmath37 [ again encoded via the rule from eq.([moments01 ] ) ] .
the superscript @xmath38 from eq.([metrix ] ) denotes the vector transposition .
we start the construction of @xmath29 by assuming that self - energy is known for @xmath39 memorized vectors @xmath40 where @xmath41 are free parameters of our model .
the quality of the interpolation via rbfn depends on the choice of the basis functions and corresponding weights .
most convenient for our preliminary purposes seems to be the use of nadaraya - watson regression estimator @xcite . according to this ,
the self - energy input - output relation can be written @xmath42 where @xmath43 are radial basis functions @xmath44 , @xmath45 satisfying the normalization conditions @xmath46 the dot index of @xmath37 was omitted whenever the distinctions between the individual dots is unimportant .
the choice of the dispersion @xmath47 in eq.([eqnw ] ) , where @xmath48 is consistent with the recommendation @xcite . speaking in terms of the neural networks @xmath49 is the activation function , which determines @xmath50th neuron s response to a given input @xmath37 and self - energy parameter @xmath51 is the optimized weight of the link between the input and output layer of the network .
having established the formula for the calculation of the self - energy , it is easy to prove that if the systems of the memorized vectors is composed from the conjugate vector pairs @xmath52 , @xmath53 with the same self - energy parameters @xmath54 rbfn self - energy form including the typical combination of terms @xmath55 posses the reflection symmetry @xmath56 to analyze the configuration snapshots generated during the simulation process we have introduced so called feature map @xmath57 .
it associates any pattern vector @xmath37 with the feature @xmath58 @xcite .
the feature @xmath59 identifies the index of a nearest memorized vector @xmath60 : @xmath61 the classification of the features performed for the whole cluster gives rise to the @xmath62component vector @xmath63 .
the information about this vector can be concentrated to the form of the sample averages @xmath64 where @xmath65 is the usual kronecker symbol .
the standard assumption about the interdot interactions is that they are essentially magnetostatic @xcite . by using the concept of the effective moments and interaction centers one can construct easily the interdot interaction potential . for this aim
we expressed the energy contribution @xmath66 consisting of the dipole - dipole interactions of the effective moment @xmath20 [ located at the @xmath9th center of @xmath10th dot at the position @xmath67 with @xmath23 effective moments of @xmath68th dot ( @xmath69 ) @xmath70 where @xmath71 because @xmath72 , @xmath73 are scaled by the lattice spacing @xmath74 , the energy dimension is absorbed into dipolar constant @xmath75 where @xmath76 is the relative permeability of the matrix .
for the interdot magnetostatic energy of dot pair @xmath77 we obtain the expression @xmath78 the key remaining contribution is the zeeman energy . for @xmath10th dot interacting with the external magnetic field @xmath79
we obtain @xmath80 in further , to characterize the external field , we have used the reduced undimensional field @xmath81 , its components @xmath82 @xmath83 , @xmath84@xmath85 and polar angle @xmath86 .
the final form of total energy functional is then given by @xmath87 the schematic view on its structure is displayed in fig.[fig3 ] .
according to the scheme of the computation , the interacting dots have their counterpart in the interacting rbfn blocks .
this aspect makes our formulation close to the concept of the interacting neural networks @xcite .
in the concrete case we have modelled the ultrathin ( quasi - two dimensional ) rectangle magnetic dots of the square profile @xmath88 and very small height @xmath89 @xmath90 .
each dot is replaced by the four moments @xmath91 at the positions
@xmath92 @xmath93 @xmath94\,$ ] ( see fig.2(a ) ) .
the magnetic moment of @xmath11 element [ see fig.[fig2](b ) ] maintains the saturation value @xmath95 .
after this eq.([lambdaeq ] ) modifies to the form @xmath96 . .
the list of eleven memorized configurations of the magnetic moments @xmath52 with the corresponding self - energy parameters .
the minus sign in @xmath52 substitutes @xmath97 . [ cols=">,<,^ " , ] in the case with the non - zero self - energy parameters interact magnetostatically , the choice of the initial conditions of the simulated annealing becomes to be more complicated .
several final configurations obtained by the annealing process are displayed in figs.[fig8](b)-(i ) .
the preliminary runs evolving from the initial random state get stuck in the local minimum @xmath98 .
the configuration of this metastable state is displayed in fig.[fig8](b ) .
the application of the feature map [ see fig.[fig8](c ) ] reveals that the clustering of the dot states resembles the formation of the homogeneous domains in @xmath99state potts model @xcite .
the lowest energy @xmath100 was obtained for the system initialized from the vortex state .
no essential differences between the pure vortex and mixed vortex - countervortex initial conditions were observed , contrary to our expectation evoked by the study of dipolar system @xcite .
the question of the helicity will require more detailed investigation .
fig.[fig8](d ) shows that magnetostatic deformation of vortices is rather pronounced feature . after the deformation , four - moment vortices
acquire the nonzero magnetic moments and resemble the fans or vortices with the non - central nel - type core .
their analysis via @xmath57 map shows the mixing of vortex features @xmath101 [ @xmath102 @xmath103 @xmath104 @xmath105 and separation of the clusters with different vorticity [ see fig.[fig8](d ) .
this aposteriory finding confirms that parameters of eq.([casea ] ) are sufficient for the stabilization of the vortex ground state for @xmath106 .
two configurations displayed in figs.[fig8](e),(f ) were obtained for the external magnetic field .
their energies are listed in table 3 . for @xmath107
the field - deformed vortices resemble the fans ordered into the large - scale wave - like structures .
the waves are better visible from fig.[fig8](h ) showing the detailed snapshot , where intradot magnetic moments are averaged for each dot separatelly .
the chaining of field - oriented phase is also visible . for @xmath108 fig.[fig8](i ) demonstrates the formation of the clusters with @xmath109 . to characterize the anisotropy ,
we have studied the angular dependence of the magnetization for @xmath110 and varying @xmath111 . in the simulations we have distinguished between the clockwise and counter - clockwise field rotation directions .
after the annealing starting from the purely vortex state @xmath112 , @xmath113 ( @xmath114 ) we performed the series of quasistatic remagnetization steps at zero temperature .
these simulations have revealed the hard axes @xmath18 , @xmath19 and easy axes @xmath115 [ see fig.[fig7 ] ] .
in addition , the model system exhibits the angular hysteresis .
these results demonstrate how the outputs of rbfn mimic the biaxial anisotropy and how the anisotropy is reinforced by the magnetostatic couplings .
we believe that very general method we have introduced in this paper will be stimulating for the people working in the field of the micromagnetic simulations of the nanoscale systems . from the point of view of magnetostatics ,
direct model improvement is possible in many ways : a ) near dot interactions can be taken into account more accuratelly by including the rectangle - rectangle magnetostatic terms ; b ) to speed - up the computations and to extend the system size one can use the hierarchical summation @xcite .
we thing that more realistic simulations will be possible after the finding of a closer relationship between the neural networks and outputs of the standard micromagnetic approaches . * acknowledgement * : this work was supported by the grant no.1/6020/99 and by the polish - slovak international grant .
chou , proceedings of the ieee 85 ( 1997 ) 652 . te - ho
wu , j.c .
wu , bing - mau chen , han - ping d. shieh , j.magn.magn.mater .
193 ( 1999 ) 155 .
a. aharoni , j.appl.phys .
68 ( 1990 ) 2892 .
a. aharoni , ieee trans.magn . 27
( 1991 ) 4775 .
gushlienko , appl.phys.lett . 75 ( 1999 ) 394 .
stamps and r.e .
camley , j.magn.magn.mater . 177
( 1998 ) 813 .
luttinger and l. tisza , phys.review 70 ( 1946 ) 954 .
brankov and d.m .
danchev , physica a 144 ( 1987 ) 128 .
s. prakash and c.l .
henley , phys.rev .
b 42 ( 1990 ) 6574 . k.yu .
gushlienko , `` magnetic anisotropy in dot arrays induced by magnetostatic interdot coupling '' , to be published in j.magn.magn.mat .
brown , `` magnetostatic principles in ferromagnetism '' , north - holland publishing co. - amsterdam , interscience publishers inc . - new york , 1962 .
berkov and n.l .
gorn , phys.rev .
b 57 ( 1998 ) 14332 . j.j .
miles , and b.k .
middleton , j.magn.magn.mater .
95 ( 1991 ) 99 .
schabes and h.n .
bertram , j.appl.phys .
64 ( 1988 ) 1347 .
t. schrefl , j.magn.magn.mater . 207
( 1999 ) 66 .
popkov , l.l .
savchenko and n.v .
vorotnikova , proceedings of the european conference `` physics of magnetism 99 '' , poznan 1999 , 499 . s. labb and p .-
bertin , j.magn.magn.mater .
206 ( 1999 ) 93 .
abraham , j.q .
broughton , n. berstein and e. kaxiras , computers in physics 12 ( 1998 ) 538 .
a. lejeune , j. perdang and j. richert , phys.rev.e 60 ( 1999 ) 2601 .
brown , m.n .
gibbs and d.c .
clary , j.chem.phys .
105 ( 1996 ) 17 .
liebovitch , n.d .
arnold and l. selector , journal of biological systems 2 ( 1994 ) 193 .
jones , a.w.g .
duller , r.w .
chantrell , a. hoare and p.r .
bissel , j.magn.magn.mater . 193
( 1999 ) 416 .
t. kohonen , `` self - organizing maps '' , springer , 1995 . s. haykin , `` neural networks '' , by prentice hall , inc . simon and schuster / a viacom company upper saddle river ,
new jersey 07458 , 1999 .
r. metzler , w. kinzel and i. kanter , phys.rev.e 62 ( 2000 ) 2555 .
vadmedenko , a. ghazali and j .- c.s .
lvy , phys.rev .
b 59 ( 1999 ) 3329 .
t. shinjo , t. okuno , r. hassdorf , k. shigeto and t. ono , science 289 ( 2000 ) 930 .
a. fernandez , m.r .
gibbons , m.a . wall , c.j .
cerjan , j.magn.magn.mater .
190 ( 1998 ) 71 . m.r .
scheinfein , j.l .
blue , j.appl.phys . 89
( 1991 ) 7740 .
s. kirkpatrick , j.stat.phys .
24 ( 1983 ) 975 .
k. binder and d.w .
heerman , `` monte carlo simulation in statistical physics '' , springer - verlag , berlin , new york , 1992 . n. metropolis , a.w .
rosenbluth , m.n .
rosenbluth , a.h .
teller , and e. teller , j.chem.phys . 21
( 1953 ) 1087 .
c. jeppesen , h. flyvbjerg and o.g .
mouritsen , phys.rev.b 40 ( 1989 ) 9070 .
th dot from the input vector @xmath116 ; c ) the interaction of dots ( rbfn ) mediated by the effective magnetic moments .
, title="fig:",width=377,height=415 ] [ fig3 ] th dot is located at the position @xmath22 ; the magnetic structure of the dot is represented by the four interaction centers and four effective magnetic moments @xmath20 ; ( b ) two examples of the dots showing how the effective magnetic moment arises from the integration of the magnetization field over the @xmath118 element ; ( c ) the example of the memorized intradot configuration consisting of the four magnetic moments .
the example of the encoding of intradot configuration by means of the eight - dimensional vector @xmath119.,width=302,height=529 ] .
the configuration shows only the moments obtained for each dot separatelly .
the annealing provides @xmath123 .
the results obtained in cases ( b)-(i ) belong to the parameters from eq.([casea ] ) .
part ( b ) displays disordered - quenched metastable microstate of the energy @xmath98 ( for @xmath106 ) obtained for the random initial condition and forbiden moves * 3.2 ( a),(b),(c)*. the metadomain structure was analyzed by @xmath57 map [ case ( c ) ] .
the annealing from the initial vortex state stabilizes at the lowest energy @xmath100 [ cases ( d),(g ) ] .
part ( g ) shows only the dots with @xmath124 , the remaining part with @xmath125 was removed for the clearness .
for the non - zero external fields we constructed snapshots ( e ) , ( h ) @xmath126 , @xmath127 ( wave - like ordered structures ) ; ( f ) , ( i ) @xmath128 , @xmath127 ; part ( h ) shows averaged magnetic moments per each dot .
it forms wave - like structure . for
( i ) the formation of @xmath109 phase clusters is visible ( here labels of @xmath101 features are removed ) .
, title="fig:",width=491,height=597 ] [ fig8 ] | two dimensional square lattice general model of the magnetic dot array is introduced . in this model
the intradot self - energy is predicted via the neural network and interdot magnetostatic coupling is approximated by the collection of several dipolar terms .
the model has been applied to disk - shaped cluster involving @xmath0 ultrathin dots and @xmath1 interaction centers . in this case among the intradot magnetic structures retrieved by neural networks the important role play single - vortex magnetization modes .
several aspects of the model have been understood numerically by means of the simulated annealing method . |
in the last decade or two we have made remarkable progress in measuring cosmological parameters to unprecedented accuracy .
parameters such as the hubble constant , @xmath0 @xcite , the temperature of the cosmic background radiation ( cbr ) , @xmath1 @xcite , and many other parameters are now known to impressive precision .
of course , we do nt invest so much time and effort to determine cosmological parameters because of an interest in numerology , but rather , we do it because the parameters are necessary input for the task of constructing a standard cosmological model . the precision cosmological measurements have lead to the latest cosmological model , usually called the _ standard cosmological model , _ or @xmath2cdm , where @xmath2 indicates einstein s cosmological constant ( or more generally , dark energy ) , and cdm stands for cold dark matter . aspects of this cosmological standard model relevant for the purposes of these lectures are indicated in figure [ pie ] .
the most remarkable feature of the standard cosmological model is that it seems capable of accounting for all cosmological observations ; i.e. , it seems to work !
there are several interesting features of the @xmath2cdm model illustrated in fig .
first , while the early universe was radiation dominated , at present the radiation energy density is only @xmath3 of the total .
the figure also illustrates why chemistry is not very important : the chemical elements ( by which i mean elements other than the simplest elements of hydrogen and helium ) are only about @xmath4 of the total .
neutrinos contribute a much larger fraction of the total mass - energy density , about @xmath5 .
the neutrino contribution is about as large as the contribution from stars .
most of the baryons in the universe are not found in stars , but rather they are in the form of a hot intracluster gas of hydrogen and helium .
this completes what we `` see '' ( although there is only indirect evidence for the cosmic neutrino background ) .
the most striking feature of the present composition of the universe in the @xmath2cdm model is that @xmath6 of the present universe is dark ! of the dark components ,
roughly @xmath7 of the total mass - energy density is dark matter , associated with galaxies , clusters of galaxies , and other bound structures .
the bulk of the mass - energy of the universe , about @xmath8 in the standard @xmath2cdm model , is dark energy , which drives an accelerated expansion of the universe . in the first lecture
i will discuss issues associated with dark energy . in the second lecture
, i will discuss possibilities for dark matter .
also in the second lecture i will briefly discuss some aspects of cosmic inflation , in particular what we might learn about inflation from present observations .
for the purposes of this school , let me begin by recalling einstein s comment to arnold sommerfeld on december 9 , 1915 , `` how helpful to us is astronomy s pedantic accuracy , which i used to secretly ridicule . ''
while einstein was referring to precision measurements of the advance of the perihelion of mercury , now we find precision measurements helpful to us because dark energy , dark matter , inflation , and baryo / leptogenesis seem to point to physics beyond the standard model of particle physics .
cdm cosmological model .
illustrated by the pie is the fraction of the mass - energy of the universe in various components . also noted are some other necessary parts of the standard model , namely seed perturbations and baryo / leptogenesis.,width=415 ]
the most important point regarding dark energy , which i will stress repeatedly , is that all evidence for dark energy is _ indirect . _
the only effect of dark energy is on the expansion history of the universe .
the expansion rate of the universe is perhaps the most fundamental quantity in cosmology . in the standard cosmological model the expansion rate
is determined from the @xmath9-component of the einstein equations , @xmath10 . in the friedmann - robertson - walker - lematre ( flrw )
model , @xmath11 , where @xmath12 is the expansion rate ( here , @xmath13 is the robertson - walker scale factor and @xmath14 or @xmath15 depending on the geometry ) . with the assumption of a perfect - fluid stress tensor , @xmath16 , where @xmath17 is the mass density .
the stress tensor is assumed to consist of several components with different equations of state . for any component @xmath18 ,
the equation of state is defined as @xmath19 .
the energy density in component @xmath18 evolves with redshift @xmath20 as @xmath21 , where @xmath22 is the present value of the scale factor . in the event that @xmath23 is a function of @xmath13 ,
then @xmath24 should be replaced by @xmath25\right\}$ ]
. then the expansion rate as a function of redshift can be expressed as @xmath26 .
\label{hsquared}\ ] ] here , @xmath27 is the ratio of the present energy density in component @xmath18 compared to the critical density @xmath28 , and @xmath29 .
the subscript `` @xmath30 '' indicates a matter component ( @xmath31 ) while the subscript `` @xmath32 '' indicates a radiation component ( @xmath33 ) .
the value of @xmath34 is well determined by wmap to be @xmath35 at the @xmath36 confidence level @xcite .
the value of @xmath37 is also well determined by cbr measurements : it is of order @xmath38 .
finally , @xmath39 is determined by measuring the expansion history of the universe , and determining the values of @xmath39 and @xmath23 necessary to fit the data .
many cosmological observables depend on the expansion history of the universe .
they often depend on the expansion history through the coordinate distance @xmath40 of a source of redshift @xmath41 . recall that the robertson - walker metric can be written in the form @xmath42 , \label{rw}\ ] ] where here @xmath43 is the angular differential and @xmath40 is the comoving `` radial '' coordinate .
the radial coordinate of a source at redshift @xmath41 is determined by integrating the null geodesic ( @xmath44 ) to obtain @xmath45 , \label{rofz}\ ] ] where @xmath46 , @xmath47 , @xmath48 obtains for @xmath49 , respectively .
now let us turn to the observables defined in table [ obs ] .
one can see that they all depend on the time evolution of the expansion rate , hence to the properties of the stress tensor , hence to a new fluid characterized by @xmath23 .
.[obs ] the relationship of observables to the time evolution of the expansion rate of the universe .
the coordinate distance @xmath50 depends on the time evolution of @xmath51 through eq .
( [ rofz ] ) . [ cols="<,^,^,>",options="header " , ] 1 .
the inflaton potential must be `` flat '' in the sense that its derivative must be unusually small .
2 . while it is always easy to write a potential that is flat , it must remain flat when one includes radiative corrections .
this presumably suggests some ( possibly approximate ) symmetry .
although it has often been said that susy can come to the rescue , in practice it is not so easy to implement @xcite .
many attractive susy models give @xmath52 hybrid models .
there are no general predictions , but in many models we find : 1 .
ratio of tensor to scalar amplitudes : @xmath53 .
scalar spectral index : @xmath54 .
3 . tensor spectral index : @xmath55 .
running of @xmath56 : @xmath57 .
5 . running of @xmath58 : @xmath59
we need experimental guidance . for phenomenological guidance , it is now traditional to present the observational results on @xmath56 and the ratio of tensor to scalar perturbations on the same graph @xcite .
the utility of this approach can be seen in fig .
[ rn1 ] . to a first approximation ,
small - field , large - field , and hybrid models populate different regions of the @xmath56@xmath40 plane .
the experimental situation is shown in fig .
[ rn2 ] . if one analyzes only the wmap three - year data set , and takes a prior that there is no running of the spectral index , the harrison
zeldovich spectrum ( @xmath63 , @xmath64 ) is ruled out at approximately the @xmath65 confidence level .
the result is little changed if one adds additional data sets .
if one allows running of the scalar spectral index , then a slightly blue ( @xmath66 ) spectrum is preferred .
my conclusion is that one can not make a strong statement about whether the harrison
zeldovich spectrum is ruled out or not .
it is also premature to draw any conclusion regarding phenomenology of inflation models .
now let us return to the possibility of producing non - thermal dark matter during inflation .
particles of all types ( at least those that are not conformally coupled ) are produced during inflation .
if there is a massive stable particle species , they will also be produced during inflation and be around today .
two groups @xcite proposed this could be the source of dark matter .
detailed calculations show that that if there is a stable particle with mass comparable to the inflaton mass , they would be produced in the proper abundance to be the dark matter .
since the mass of the inflaton is expected to be @xmath67 or @xmath68 gev , the dark matter particle would be supermassive much more massive than a wimp can be hence , it is known as a wimpzilla . unlike thermal relics , where the contribution to @xmath69 depends on the annihilation cross section and independent of the mass ,
the wimpzilla contribution to @xmath69 is independent of the interactions and depends only on the mass | in these two lectures i will discuss some outstanding problems in the standard model of cosmology , concentrating on the physics that might be related to the title of this school , `` searching for the totally unexpected in the lhc era . '' in particular , i will concentrate on dark energy , dark matter , and inflation . |
we consider only simple undirected graphs ( no loops , no multiple edges ) . a _
vertex numbering _ of a graph @xmath4 is a bijective map @xmath5 $ ] from the vertex set @xmath6 of @xmath7 to the set of the first @xmath2 positive integers @xmath8=\{1,2,\dots , k\}$ ] , where @xmath9 .
the absolute value of the difference between the numbers at the two endpoints of an edge @xmath10 in @xmath11 is denoted by @xmath12 and called the _ length _ of @xmath10 induced by @xmath13 .
thus any vertex numbering of @xmath7 induces an edge labelling @xmath14 of @xmath7 by positive integers .
the largest length of an edge in @xmath11 ( i.e. the largest label used by @xmath15 ) is called the _ bandwidth of the numbering _ @xmath13 .
if the edge set is empty the bandwidth is 0 by definition .
the smallest possible bandwidth , taken over all possible vertex numberings of @xmath7 , is called the _ bandwidth _ of the graph @xmath7 .
sometimes the adjective `` linear '' is used due to the following physical interpretation .
every vertex numbering @xmath13 corresponds to a _ linear arrangement _ of the graph @xmath7 in which the vertex @xmath16 in @xmath6 is placed at the number @xmath17 on the real line .
the ( linear ) bandwidth of a linear arrangement is then just the length of the longest wire needed to assemble the graph @xmath7 and the minimal such bandwidth , taken over all possible linear arrangements of @xmath7 , is the ( linear ) bandwidth of @xmath7 . the two - dimensional _ rectangular grid _
@xmath18 , for @xmath19 , is the graph whose vertices are the points in the set @xmath20\times[n]=\{(i , j)|1 \leq i \leq m , 1 \leq j \leq n\}$ ] with an edge between two vertices if and only if the euclidean distance between them is 1 ( think of the natural embedding of the vertex set in the cartesian plane ) .
we write @xmath21 for the square grid @xmath22 .
it is well known that the bandwidth of the rectangular grid @xmath18 is @xmath23 , unless @xmath24 .
[ thm : chvatalova ] let @xmath25 . if @xmath26 , then the bandwidth of the two - dimensional rectangular grid @xmath18 is @xmath23 .
the bandwidth is , in some sense , a measure of how difficult it is to embed the graph on a line .
the following result provides some refinement of this measure for the rectangular grids .
[ thm:2(n-1 ) ] let @xmath27 .
every vertex numbering of @xmath18 of bandwidth @xmath23 induces at least @xmath28 edges of length @xmath23 .
moreover , the _ down diagonal lexicographic linear arrangement _ induces exactly @xmath29 edges of length @xmath23 .
as an example , the down diagonal lexicographical linear arrangement on the square grid @xmath21 is given by @xmath30 where @xmath31 .
this numbering has bandwidth @xmath23 and induces exactly @xmath32 edges of length @xmath23 ( all horizontal edges incident to the main diagonal vertices ) .
note that the down diagonal lexicographical vertex numbering of @xmath21 that we just displayed indicates that we sometimes think of @xmath18 as a rectangular board of dimension @xmath0 ( meaning @xmath23 rows and @xmath33 columns ) in which the squares represent the vertices and any pair of squares that have common side are considered to be neighbors . by the result of chvtalov , we know that in order to make a linear arrangement of the square grid @xmath21 we must be ready to use pieces of wire of length at least @xmath23 . moreover , by the result of fishburn and wright , in order to make an arrangement that does not use any pieces longer than @xmath23 , we must be ready to use at least @xmath32 pieces of length @xmath23 .
but what if we do not have such long pieces ? in such a case the most practical thing one can do is to try to assemble as large part of the graph as possible , which amounts to deletion of some of the edges .
the bandwidth of the path @xmath34 of length @xmath35 is 1 , of the cycle @xmath36 of length @xmath37 is 2 , and of the complete graph @xmath38 on @xmath39 vertices is @xmath40 .
the deletion of any edge in the cycle @xmath36 produces the path @xmath34 and thus reduces the bandwidth by 1 .
similarly , the deletion of any edge in @xmath38 also reduces the bandwidth by 1 .
we are interested in the minimal number of edges that need to be deleted in an arbitrary graph in order to reduce its bandwidth .
the _ bandwidth reduction number _ of a graph @xmath7 of bandwidth @xmath41 , @xmath42 , is the minimal number of edges that need to be deleted from @xmath7 in order to obtain a subgraph of bandwidth no greater than @xmath43 .
for example , cycles and complete graphs have bandwidth reduction number 1 , while paths have bandwidth reduction number equal to their length . as a more interesting example , consider the wheel @xmath44 , @xmath45 , a graph on @xmath33 vertices that consists of a cycle of length @xmath46 together with an `` center '' vertex connected to every vertex in the cycle by an edge .
the following table provides the bandwidth and bandwidth reduction number of the wheel graphs .
we do not provide the easy proofs , but we note the exceptional case of @xmath47 , which is the only wheel whose bandwidth reduction number is 3 .
@xmath48 another relatively easy example is provided by the complete bipartite graphs @xmath49 , for @xmath50 . @xmath51
the notion of bandwidth reduction can be further extended as follows .
let @xmath7 be a graph of bandwidth @xmath41 . for @xmath52 ,
the @xmath2-th bandwidth reduction number of @xmath7 , denoted by @xmath53 , is the minimal number of edges that need to be deleted from @xmath7 in order to obtain a subgraph of bandwidth no greater than @xmath54 .
we note that the @xmath2-th bandwidth reduction number of a graph @xmath7 of bandwidth @xmath41 is the minimal number of edges @xmath10 of induced length @xmath12 greater than @xmath54 , taken over all possible numberings @xmath13 of @xmath7 , i.e. , @xmath55 for example , in the case of the complete graph @xmath38 we have @xmath56 for @xmath57 , since the bandwidth of @xmath38 is @xmath40 and every numbering of @xmath38 induces @xmath58 edges of length @xmath59 , for @xmath60 .
this implies that if a graph on @xmath23 vertices has more than @xmath61 edges , then its bandwidth is at least @xmath62 , for @xmath57 .
before we move on , we observe that the bandwidth of a graph can be reduced by more than 1 by deletion of a single edge .
this is why we require the newly obtained graph to have bandwidth no greater than @xmath54 rather than exactly @xmath54 in the definition of the @xmath2-th bandwidth reduction number .
consider the graph @xmath7 that consists of two copies of the wheel @xmath47 with an additional edge between the wheel centers ( two wheels with an axis between them ) .
since an arbitrary graph on @xmath16 vertices with diameter @xmath63 and bandwidth @xmath41 satisfies @xmath64 the bandwidth of the graph @xmath7 is at least 5 ( @xmath16=14 and @xmath65 ) .
however the deletion of the edge between the centers of the two wheels leads to a graph of bandwidth 3 .
thus @xmath66 for this graph .
we give now an upper bound on the @xmath2-th bandwidth reduction number in rectangular grids , for small @xmath2 .
[ reduction ] let @xmath67 .
the @xmath2-th bandwidth reduction number of the rectangular grid @xmath18 satisfies @xmath68 we construct an example of a numbering @xmath13 of @xmath18 in which only @xmath3 edges are longer than @xmath62 .
the example is a modification of the down diagonal lexicographical linear arrangement .
think of @xmath18 as of rectangular board of dimension @xmath69 .
first modify the board by cutting out the lower right part of the board of dimension @xmath70 and flipping it over as indicated in figure [ fig : reduction ] .
with @xmath3 edges longer than @xmath62 ] enumerate all the squares in the modified board in the down diagonal lexicographical fashion .
consider a `` typical '' vertex @xmath72 that lives on a diagonal of length @xmath73 .
the vertical edges incident to @xmath72 have length @xmath73 and the horizontal ones have length @xmath74 .
a `` typical '' vertex @xmath75 that lives on a diagonal of length @xmath62 is incident to horizontal edges of length @xmath62 and vertical edges of length @xmath76 .
a `` typical '' vertex @xmath77 that lives on a diagonal of length @xmath2 is incident to vertical edges of length @xmath2 and horizontal edges of length @xmath78 , where the inequality comes from the assumption that @xmath79 . in all `` atypical '' cases , near the bottom left or the upper right corner(s ) of the modified board ,
the diagonals are even shorter , which implies that the incident edges are also shorter .
thus the bandwidth of the given numbering of the modified board is @xmath62 .
after we flip back the modified part of the board to its original position the newly created @xmath2 horizontal edges ( between the non - shaded and the shaded part of the board ) will have length longer than @xmath62 and all but one of the newly created @xmath80 vertical edges will have length longer than @xmath62 . indeed
the induced length of the rightmost vertical edge between the shaded and the non - shaded part is exactly @xmath62 , since its endpoints ( denoted by @xmath81 and @xmath11 in figure [ fig : reduction ] ) were already neighbors in the modified board .
we provide one example to illustrate the preceding result .
assume that we want to reduce the bandwidth of @xmath82 to 6 .
thus @xmath83 , @xmath84 and the numbering of the modified board is given by @xmath85 another way to state theorem [ reduction ] is to say that the incidence matrix of @xmath18 can be written in such a way that all non - zero entries , except for @xmath3 symmetric pairs , are within distance @xmath62 from the main diagonal .
for example , this means that the incidence matrix of the square grid @xmath21 can be written in such a way that all non - zero entries , except for 2 symmetric pairs , are within distance @xmath40 from the main diagonal .
this contrasts nicely with theorem [ thm:2(n-1 ) ] , which implies that if all non - zero entries of the incidence matrix of @xmath21 are within distance @xmath23 from the main diagonal then there are at least @xmath32 symmetric pairs of non - zero entries at distance @xmath23 from the main diagonal .
we show now that the bandwidth of square grids can not be reduced if only 1 edge is deleted .
the exceptional cases are @xmath86 and @xmath87 .
indeed , @xmath88 has bandwidth 0 which can not be reduced , while @xmath89 is the cycle @xmath90 and its bandwidth reduction number is 1 .
[ thm:2 ] let @xmath91 .
the bandwidth reduction number of @xmath21 is 2 , i.e. , the minimal number of edges that needs to be deleted from @xmath21 in order to obtain a graph of bandwidth @xmath40 is 2 .
moreover , the edges that need to be deleted may , but do not have to , share a common endpoint . for an arbitrary numbering @xmath13 of @xmath21 , call an edge _
long _ if its induced length is at least @xmath23 and call it _ short _ otherwise .
an example of a numbering of @xmath21 , for @xmath91 , with only two long edges is already provided in the proof of theorem [ reduction ] .
we note that the two long edges in that example share a common vertex , namely the vertex in row @xmath23 column @xmath40 .
the lengths of the long edges are @xmath92 and @xmath93 . rather than providing a general example of a numbering with only two long edges that are not adjacent
, we provide an example for the case @xmath94 , from which one can easily extract the general pattern .
the two long edges , which are the two leftmost vertical edges between the top two rows , have lengths @xmath95 and @xmath96 .
@xmath97 it remains to show that every numbering of @xmath21 , @xmath91 , has at least two long edges .
let @xmath13 be an arbitrary labelling of @xmath21 .
choose the smallest @xmath2 such that all rows or all columns of @xmath21 have a label in @xmath8 $ ] .
without loss of generality we may assume that all rows have a label in @xmath8 $ ] .
define the partial numbering @xmath98 of @xmath21 to be the numbering of the vertices obtained from the numbering @xmath13 by deletion of the labels greater than @xmath2 ( thus only @xmath2 vertices have a label ) .
* claim 1 . *
the label @xmath2 is alone in its row in the partial numbering @xmath98 .
otherwise each row would have a label from @xmath99 $ ] which contradicts the minimality in the choice of @xmath2 .
* claim 2 . *
no row is completely numbered by @xmath98 .
otherwise each column would have a label from @xmath99 $ ] which contradicts the minimality in the choice of @xmath2 .
call an edge between vertex numbered by an element in @xmath8 $ ] and an element not in @xmath8 $ ] a _ boundary edge_. if there are at least @xmath100 boundary edges then at least two of them would have to be long .
a direct corollary of claim 2 is that each row has at least one horizontal boundary edge .
this already means that there must be at least one long edge ( apropos , this proves theorem [ thm : chvatalova ] in the case of square grids ) .
thus , we assume in the sequel that there are exactly @xmath23 horizontal boundary edges and only one of them is long .
moreover , we assume that this is the only long edge in @xmath13 and we seek a contradiction .
* claim 3 . * the endpoints of the @xmath40 short horizontal boundary edges are numbered by the pairs @xmath101 while the endpoints of the long edge come from a pair @xmath102 , where @xmath103 is a `` small '' number less than or equal to @xmath104 and @xmath105 is a `` large '' number greater or equal to @xmath106 . indeed ,
any other numbering of the endpoints of the boundary edges would result in at least two long horizontal edges .
* claim 4 . * for every row @xmath58 in @xmath21 , the vertices labelled by the partial numbering @xmath98 form a horizontal path @xmath107 that ends at the leftmost or the rightmost column .
this follows from the fact that every row has exactly one horizontal boundary edge .
thus each row contains a unique maximal path that consists of vertices numbered by @xmath98 .
these @xmath23 paths will be called @xmath98-paths .
* claim 5 . * the @xmath98-path in any row next to @xmath2 consist of a single vertex in the column of @xmath2 .
indeed , if the neighboring row has a vertex labelled by @xmath98 that is not in the same column as @xmath2 , this vertex would be involved in a vertical edge that would have to be long ( all vertices in the row of @xmath2 are labelled by numbers greater or equal to @xmath108 ) . without loss of generality
we may assume that the vertex @xmath2 ( together with its vertical neighbors from @xmath8 $ ] ) is in the leftmost column in @xmath21 .
* claim 6 . *
if the @xmath98-path @xmath109 is no longer than @xmath110 , then the @xmath98-path @xmath111 in any neighboring row occupies the same columns as @xmath109 or differs in exactly one column .
again , in any other case there would be a long vertical edge involving a number from @xmath8 $ ] and a number outside of @xmath8 $ ] .
note that we required that the length of @xmath109 be no longer than @xmath110 , since if the length is @xmath40 the neighboring @xmath98-path @xmath111 may also have length @xmath40 and start at the opposite end without creating additional long vertical edge(s ) .
however , we will see that this does not happen .
* claim 7 . * no horizontal @xmath98-path is longer than @xmath110 and they all start at the left end .
the length of the @xmath98-path @xmath2 is 1 and so is the length of the neighboring @xmath98-path(s ) .
going further away , the length of the @xmath98-paths can only go up by 1 from one row to another , so the only way we can have a @xmath98-path of length @xmath40 is if @xmath2 is in the bottom or in the top row , the next @xmath98-path has length 1 and the length of the remaining @xmath98-paths grows by 1 as we move away from the row of @xmath2 .
without loss of generality assume that @xmath2 is in the bottom left corner .
consider now the @xmath98-path @xmath112 , corresponding to the row with a long horizontal edge , together with the horizontal path @xmath111 in the row below .
@xmath113 where the stars represent arbitrary numbers in @xmath8 $ ] .
clearly , if @xmath114 is not in @xmath8 $ ] we obtain a long vertical edge with endpoints @xmath114 and @xmath103 .
thus @xmath114 is in @xmath8 $ ] and the lengths of @xmath112 and @xmath111 are equal .
since the lengths of the @xmath98-paths start at 1 and repeat at the row of @xmath2 and at the row of @xmath103 , the longest @xmath98-path can only reach the length @xmath110 .
* claim 8 . *
the @xmath98-path @xmath112 corresponding to the row with the long horizontal edge @xmath115 has the same length as the @xmath98-paths in the rows next to it .
we already proved that the @xmath98-path below @xmath112 can not be shorter than @xmath112 .
it can not be longer as well , since the configuration @xmath116 with @xmath114 in @xmath8 $ ] , indicates that @xmath105 and @xmath114 are the endpoints of a long vertical edge . by symmetry
, the @xmath98-path above @xmath112 also has length equal to the length of @xmath112 .
this finishes the proof of claim 8 .
we now consider the partial numbering @xmath117 that corresponds to the set of numbers @xmath118 $ ] .
as observed before , the numbers @xmath119 are placed in different rows , one in each row except for the row of the long horizontal edge @xmath120 .
thus @xmath40 horizontal boundary edges are now formed between the vertices numbered by @xmath118 $ ] and those not numbered by @xmath118 $ ] . in addition , there exist at least one vertical boundary edge with endpoints in an element of @xmath118 $ ] and @xmath105 .
these @xmath23 boundary edges of the partial numbering @xmath117 must contain a long edge , a contradiction .
we conjecture that in order to reduce the bandwidth of @xmath18 by @xmath2 , for @xmath121 , we need to delete at least @xmath3 edges , i.e. , we conjecture that the upper bound provided by theorem [ reduction ] is the actual value of the bandwidth reduction number for all sufficiently small values of @xmath2 .
theorem [ thm:2 ] shows that this conjecture is true when @xmath122 and @xmath123 .
the proof of theorem [ thm:2 ] indicates an easy way to see that at least @xmath2 edges need to be deleted in order to reduce the bandwidth of @xmath18 by @xmath2 .
namely , the first time some partial numbering has a representative in each row ( or each column ) there would be at least @xmath23 horizontal ( or vertical ) boundary edges at least @xmath2 of which must have length at least @xmath124 .
for a numbering @xmath13 of a graph @xmath4 one can define the _ vertex - isoperimetric _ number of @xmath13 to be the maximal number of vertices in @xmath7 numbered by @xmath8 $ ] that have neighbors that are not numbered by @xmath8 $ ] , where the maximum is taken over all @xmath125 .
the smallest possible vertex - isoperimetric number , taken over all numberings of @xmath7 , is the _ vertex - isoperimetric number _ of the graph @xmath7 .
if @xmath126 is the bandwidth of @xmath7 and @xmath127 is its vertex - isoperimetric number then @xmath128 it is well known that the vertex - isoperimetric number of the square grid @xmath21 is @xmath23 .
the down diagonal lexicographical linear arrangement provides an example of a numbering with vertex - isoperimetric number @xmath23 and the proof of theorem [ thm:2 ] essentially contains a proof that this number can not be less than @xmath23 .
thus in order to reduce the bandwidth of @xmath21 we need to delete enough edges so that at least the vertex - isoperimetric number is reduced .
however , one can reduce the vertex - isoperimetric number of @xmath21 by deleting only one edge . for example , consider the case of @xmath129 and the numbering given by @xmath130 if the edge with endpoints labelled by 7 and 13 is deleted , the newly obtained graph has vertex - isoperimetric number 3 .
this indicates that , in general , the problem of bandwidth reduction is more difficult than the problem of vertex - isoperimetric number reduction . it also explains why , in the course of the proof of theorem [ thm:2 ] , it was relatively easy to show that at least one edge needed to be deleted , but one had to work harder for the second edge . finally , we note that the @xmath2-th bandwidth reduction number of the square grid @xmath21 is bounded linearly by @xmath131 , for @xmath121 , i.e , one has to delete relatively small number of edges to reduce the bandwidth substantially .
however , it is clear that for @xmath132 the bandwidth reduction gets more difficult .
a linear bound on the bandwidth reduction number simply can not be found for all @xmath2 .
after all , for @xmath133 , one needs to remove all @xmath134 edges to get to bandwidth 0 .
thanks to kiran kedlaya for his input and interest . | we show that the bandwidth of a square two - dimensional grid of arbitrary size can be reduced if two ( but not less than two ) edges are deleted .
the two deleted edges may not be chosen arbitrarily , but they may be chosen to share a common endpoint or to be non - adjacent .
we also show that the bandwidth of the rectangular @xmath0 ( @xmath1 ) grid can be reduced by @xmath2 , for all @xmath2 that are sufficiently small , if @xmath3 edges are deleted . |
the deep inelastic scattering ( dis ) experiments done with neutrino beams can now provide data with an accuracy comparable to the analogous processes studied with charged leptons @xcite . due to the low cross sections , for neutrino interaction with matter , these data are collected mostly from experiments done on heavy nuclei like iron @xcite . since the discovery of the emc effect @xcite , the deviation of the nuclear structure function @xmath2 with respect to the deuteron , many theoretical attempts were made to explain it ( for a review see @xcite ) . in the conventional nuclear physics models ,
the usual effects considered are fermi motion @xcite and binding of the nucleons in the nuclei @xcite and the contribution of the virtual meson cloud @xcite . for the neutrino processes ,
the study of possible nuclear effects has received much less attention , probably because of the lack of accurate measurements in the past . in present analyses of neutrino experiments ,
the high @xmath3 data ( @xmath4 ) , where fermi motion effects are expected to be important @xcite , are generally avoided , especially in the case of nuclear data @xcite .
however , it has been shown that even in the case of the deuteron , the fermi motion effects could be important @xcite .
it is interesting to note that the older fits of diemoz et al .
@xcite , show that even for lower @xmath3 ( @xmath5 ) , the nuclear effects could be important .
most of the analyses done for the neutrino data assume the same nuclear corrections as applicable to the charged lepton experiments in their global fits of the parton distributions @xcite , while allowing for an overall normalisation of the data .
this is not appropriate for the parity violating structure function @xmath6 , in the neutrino scattering data , which has no counterpart in the charged lepton case . even in the case of the structure function @xmath7 , which is related to the @xmath8 , of the emc effect
, the nuclear corrections would not be the same due to the different contribution of the sea quarks @xcite .
recently , there have been some calculations of the nuclear effects in the @xmath6 structure function @xcite .
it is found by sidorov and tokarev ( st ) @xcite that the nuclear effects in @xmath9 are quite small in the case of the deuteron , a result similar to the one found in charged lepton scattering in @xmath10 @xcite .
keeping in view the result that the nuclear effects in @xmath11 are quite different for the deuteron and heavy nuclei , the assumption that they are the same for @xmath9 , as done by st is therefore not justified . in the calculation of kulagin @xcite , the iron structure function
is evaluated and sizeable deviations from the deuteron one were found . the effect on the gross - llewellyn smith ( gls ) sum rule
@xcite is on the contrary very small .
this is so because in kulagin s approach , the spinors are normalized in order not to affect this sum rule .
it is similar to the prescription introduced by means of the so called flux factor @xcite in order to take into account relativistic effects in non relativistic formalisms .
it is interesting to note that in the fit that the ccfr collaboration made to the @xmath12 and @xmath13 neutrino data @xcite , the integral of the valence quark distributions is not 3 , @xcite .
this apparent lack of normalization of the @xmath14 in the nuclear data is of crucial importance in order to extract the strong coupling constant , @xmath15 , from the analysis of the gls sum rule , as done in @xcite .
recently , a new calculation of the nuclear effects in the case of the charged lepton scattering was made @xcite . in this calculation a relativistic many - body approach for the nucleon contribution was developed in order to avoid the use of the flux factor mentioned before @xcite .
the meson cloud contribution was found to give relatively important contributions .
good agreement was found with the experimental data .
for the structure function @xmath14 , the virtual meson cloud gives no contribution , since it is related to the difference of the quark and antiquark distributions . in the present paper
we calculate the nuclear effects to @xmath14 in the same formalism of @xcite .
the structure of the paper is as follows .
we briefly present the formalism for deep inelastic neutrino scattering in section ii . in section iii
we show how nuclear effects are calculated . in section
iv we present the results and state our conclusions .
the invariant matrix element for the feynman diagram shown in fig .
1 corresponding to the inelastic neutrino - nucleon scattering is written as @xmath16 the cross section for the neutrino scattering , @xmath17 , is then given by @xmath18 where @xmath19 stands for fermions and @xmath20 for bosons in the final state @xmath21 .
the index @xmath22 is split in @xmath23 for fermions and bosons respectively . by inserting expression ( [ tinv ] ) in eq .
( [ cross ] ) , summing over the final spins and averaging over the initial spins , the cross section can be expressed in the rest frame of the nucleon , with @xmath24 referring to the outgoing muon , as @xmath25 \ ; w_{\alpha \beta}^{\nu}\,,\ ] ] where the lepton tensors , @xmath26 and @xmath27 are given by @xmath28 @xmath29 and the hadronic tensor is defined as @xmath30 where @xmath31 is the momentum of the virtual @xmath32 , @xmath33 the spin of the nucleon and @xmath34 the spin of the fermions in @xmath21 . in the case of antineutrino scattering ,
the invariant @xmath35 matrix is given by @xmath36 and the cross section reads @xmath37 \ ; w_{\alpha \beta}^{\bar{\nu}}\,.\ ] ] the usual convention is to express the hadronic tensor as @xcite @xmath38 where @xmath39 are the structure functions , which depend on the scalars @xmath40 and @xmath41 .
the terms depending on @xmath42 , @xmath43 and @xmath44 in eq .
( [ had_ten ] ) do not contribute to the cross section in the dis @xcite . defining the variables
@xmath45 we can write in the bjorken limit @xmath46 where the @xmath47 ( @xmath48 ) sign stands for the neutrino ( antineutrino ) cross section , and the @xmath49 are dimensionless structure functions defined as @xmath50 in the quark parton model , the structure functions are expressed in terms of the quark distributions , @xmath51 and @xmath52 . using callan - gross relation ,
these structure functions are given as @xmath53 @xmath54 @xmath55
@xmath56=q(x ) -\bar{q}(x)= u_v(x)+d_v(x)\,.\ ] ] here @xmath57 the neutrino structure function @xmath58 is related with the analogous one for charged lepton scattering @xmath59 by @xmath60 in the deep inelastic region , quark distributions satisfy sum rules in order to give the correct charge , strangeness and charm of the proton and the neutron .
this implies the constraint @xmath61 known as the gross - llewellyn smith ( gsl ) sum rule @xcite .
qcd corrections modify the expression of the structure functions in terms of the quark distributions , eqs .
( [ struc ] ) . at leading order , eqs .
( [ struc ] ) retain the same structure but the quark distributions evolve with @xmath1 according to the dglap ( dokshitzer - gribov - lipatov - altarelli - parisi ) evolution equations @xcite .
when higher orders are considered , eqs .
( [ struc ] ) are modified including explicitly the strong coupling constant , @xmath62 .
the gls sum rule is modified and in the nlo approximation is given by @xmath63 further higher order corrections up to order @xmath64 can be found in @xcite .
in order to calculate the neutrino - nucleus cross section we first evaluate the related neutrino self - energy in the medium , fig . 2 .
the self - energy is given by @xmath65 where @xmath66 is the @xmath32 self - energy in the medium .
( [ neu_self1 ] ) can be rewritten in the form @xmath67 the probability per unit time for the neutrino to collide with nucleons when travelling through nuclear matter is @xcite @xmath68 and the cross section for an element of volume @xmath69 in the nucleus is @xmath70 @xmath71 can be easily evaluated from eq .
( [ neu_self2 ] ) by means of the cutkosky rules @xcite @xmath72 and we get @xmath73 \ ; \int d^3 r \ ; \hbox{im } \ ; \pi_{\alpha \beta } ( q ) \,.\ ] ] comparing eq .
( [ dif_cross ] ) and eq .
( [ cross_nuclear ] ) we see that @xmath74 next , we evaluate the @xmath32 self - energy in the medium @xmath75 where we have a minus sign because of the necessary fermion loop . in the antineutrino case
the expressions obtained are very similar .
@xmath27 appears , as in eq .
( [ dif_cross2 ] ) , with a minus sign in front and in the @xmath32 self - energy , eq .
( [ selfw ] ) , we have @xmath76 , instead of @xmath77
. from now on we will always speak of the average of neutrino and antineutrino structure functions and will omit the superscripts @xmath78 and @xmath79 . for the nucleon propagator in the medium , @xmath80 , we take a relativistic version @xcite , which can be written as @xmath81\,,\end{aligned}\ ] ] @xmath82 and @xmath83 being the hole and particle spectral functions respectively , which are taken from the work of @xcite .
we use the local density approximation in which the spectral functions depend on the density of the point of the nucleus at which they are evaluated . in our formalism
we use spectral functions for symmetric nuclear matter .
the normalization of the hole spectral function is given by @xmath84 where @xmath85^{1/3}$ ] is the local fermi momentum at the point @xmath86 .
all our calculations are done for @xmath87fe . the density for this nucleus
is expressed as a two fermi parameter distribution given in @xcite .
we now calculate @xmath88 by using cutkosky rules , eq .
( [ cutkosky ] ) , the expression of eq .
( [ medium_prop ] ) for the nucleon propagator in the medium , free propagators for particles in the final state and by means of eq .
( [ w_a ] ) we have the hadronic tensor @xmath89 in order to evaluate @xmath90 , we calculate the components @xmath91 on both sides of eq .
( [ conv_wa ] ) .
we have , using @xmath92 , and taking @xmath93 along the @xmath94 axis , as usual , @xmath95 and for the right hand side we will have for the moving nucleon @xmath96\,.\ ] ] since we have @xmath97 @xmath98 with @xmath3 as defined in eq .
( [ bj_var ] ) and @xmath99 is the bjorken variable expressed in terms of the nucleon variables , @xmath100 , in the nucleus @xmath101 we obtain the expression for @xmath102 in the bjorken limit @xmath103f_{3n}(x_n)\,,\ ] ] where the contribution of @xmath104 appearing in eq .
( [ w_n ] ) vanishes after momentum integration . defining @xmath105 as @xmath106 we get @xmath107
we begin by showing the results for @xmath108 and comparing them with the corresponding ratio @xmath109 in fig . 3 .
for @xmath110 and @xmath111 we have taken the parametrization of @xcite .
for the meson structure functions that contribute to @xmath112 we have used the parametrization given in @xcite .
we can see some similarities in the region around @xmath113 where the ratio is smaller than unity and which is mostly due to nuclear binding effects , as discussed in detail in @xcite .
similarities in the region of @xmath114 , where the ratio shows a fast increase , are also apparent , and they are mostly due to the effect of fermi motion @xcite . the differences between the neutrino and charged lepton ratios are more important at values of @xmath115 .
these differences are mostly due to the lack of meson renormalization effects in the neutrino structure function , as can be seen in the figure , where we also show the results for @xmath112 without the meson renormalization effects
. we should note that the nuclear effects are sizeable , with values of @xmath116 around 0.8 in the region of @xmath117 and around 0.9 for low values of @xmath3 .
these nuclear corrections are considerably larger than those found for the deuteron @xcite , as it was also the case for @xmath112 in the charged lepton case . in fig .
4 we show results for @xmath116 for different values of @xmath1 .
we observe that for low values of @xmath3 the results for @xmath116 are rather independent of @xmath1 , but this is not the case at large values of @xmath3 , where there are substantial differences .
given the fact that the most important contribution to the gls sum rule comes from the values of @xmath118 , the results of the figure indicate that nuclear effects can reduce the gls sum by about 10% .
since one of the purposes of this work is to facilitate the task of experimentalists in analyses of the gls sum rule and other qcd predictions for @xmath110 , we provide here an easy parametrization of @xmath116 which can serve to induce the results for @xmath110 from the measured nuclear data of @xmath90 . in table
i we give the parameters of a fit of the ratio @xmath119 for the @xmath1 values of 5 , 30 and 50 gev@xmath120 .
the form used is @xmath121 . for larger values of @xmath1 results
are similar to those for 50 gev@xmath120 .
we also want to stress that the factor @xmath122 appearing in eq .
( [ finalf3 ] ) which goes to unity in the bjorken limit ( see eq .
( [ gamma ] ) ) , is not negligible for values of @xmath123 and produces changes of around 10% in @xmath116 for values of @xmath124 at @xmath125 gev@xmath120 . in fig .
5 we compare our results with those of kulagin @xcite and st @xcite .
it is quite clear that both for the case of kulagin and the present results the differences with the deuteron results of st are significant .
our results and those of kulagin are qualitatively similar around @xmath126 , but they divert from each other both at large @xmath3 and small @xmath3 .
the discrepancies at large @xmath3 are not surprising since it was shown in @xcite that the results for @xmath127 at large @xmath3 were very sensitive to the parametrization used for @xmath128 , which is different in @xcite and the present case .
these differences , however , should not be very relevant for tests of the gls sum rule , since the contribution to the sum rule from the region of @xmath114 is very small .
the discrepancies for values of @xmath118 are more relevant in connection with this sum rule since 10% effects , typical differences at low @xmath3 between both approaches , are of the same order of magnitude as the qcd corrections to the gls . the main reason for the discrepancies in the two results should be seen from the different approaches followed . in @xcite
the spinors are normalized in a way to force baryon number conservation in a non relativistic formalism , in line of the idea of the flux factor of @xcite to introduce relativistic corrections in a nonrelativistic formalism .
the advantage of the present work is that it uses a relativistic formalism from the beginning , in which baryon number conservation is automatically fulfilled .
the differences of 10% in the region of small @xmath3 between the two aproaches is about the same one as found between our approach and those using the flux factor in studies of @xmath129 @xcite . in summary , we present here results for @xmath116 based on a relativistic approach and the use of an accurate nuclear spectral function which reproduces nuclear bindings , momentum distributions , etc . within this formalism
we were able to reproduce the emc effect for different nuclei and the experimental results of @xmath127 in the region of @xmath130 .
we have used here the same formalism to evaluate @xmath90 .
there are some differences with respect to @xmath127 in the nuclear effects , mostly due to the absence of meson renormalization corrections in @xmath90 .
the results obtained are parametrized in an easy way to facilitate future analyses of experimentalists in order to deduce the elementary @xmath110 structure function from the nuclear measurements and be able to test qcd corrections and obtain reliable values of @xmath15 .
we would like to acknowledge w. seligman for providing us with a copy of his ph .
d. thesis and interesting comments .
one of us , s.k.s . wishes to acknowledge financial support from ministerio de educacin y cultura in his sabbatical stay in valencia .
e.m . wishes to acknowledge financial support from the ministerio de educacin y cultura .
this work is partially supported by dgicyt , contract no .
pb96 - 0753 .
permanent address : physics department , aligarh muslim university , aligarh , india 202002 .
m. conrad , m.h .
shaevitz and t. bolton , hep - ex/9707015 .
w.g . seligman et al . , phys .
lett . * 79 * , 1213 ( 1997 ) .
aubert et al . , phys .
* b123*,275 ( 1983 ) .
bickerstaff and a.w .
thomas , j. phys .
g * 15 * , 1523 ( 1989 ) .
m. arneodo , phys .
rept . * 240 * , 301 ( 1994 ) .
geesaman , k. saito and a.w .
thomas , annu .
nucl . part .
45 * , 337 ( 1995 ) .
a. bodek and j.l .
ritchie , phys .
d * 23 * , 1070 ( 1981 ) . s.v .
akulinichev , s.a .
kulagin and g.m .
pis ma zh . eksp .
fiz * 42 * , 105 ( 1985 ) ; jetp lett . *
42 * , 127 ( 1985 ) ; phys
158b * , 485 ( 1985 ) .
m. ericson and a. w. thomas , phys . lett . *
128b * , 112 ( 1983 ) .
e. marco , e. oset and p. fernndez de crdoba , nucl . phys . * a611 * , 484 ( 1996 ) .
jones et al . , z. phys .
c * 62 * , 575 ( 1994 ) .
m. diemoz , f. ferroni , e. longo and g. martinelli , z. phys .
c * 39 * , 21 ( 1988 ) .
martin , w.j .
stirling and r.g .
roberts , phys .
d * 47 * , 867 ( 1993 ) .
lai , j. botts , j. huston , j.g .
morfin , j.f .
owens , j.w .
qiu , w.k .
tung and h. weerts , phy .
d * 51 * , 4763 ( 1995 ) .
roberts , _ the structure of the proton _ ( cambridge university press , cambridge , 1990 ) , p. 41 .
sidorov and m.v tokarev , phys .
* b358 * , 353 ( 1995 ) .
kulagin , nucl - th/9801039 . c. ciofi degli atti and s. liuti , phys .
c * 41 * , 1100 ( 1990 ) .
v.v . burov and a.v .
molochkov , nucl - th/9703008 .
gross and c.h .
llewellyn smith , nucl .
phys . * b14 * , 337 ( 1969 ) .
frankfurt and m.i .
strikman , phys .
* b183 * , 254 ( 1987 ) .
seligman ph.d .
thesis , nevis report 292 , 1997 .
harris , hep - ex/9506010 .
p. fernndez de crdoba , e. marco , h. mther , e. oset and a. faessler , nucl
* a611 * , 514 ( 1996 ) .
bilenky , _ introduction to feynman diagrams in electroweak interactions physics _
( editions frontiers , gif - sur - yvette , 1994 ) . g. altarelli and g. parisi , nucl .
* b126*,298 ( 1977 ) ; v.n .
gribov and l.n .
lipatov , sov .
. phys . * 15*,438 ( 1972 ) ; yu .
l. dokshitzer , sov .
jetp * 46 * , 641 ( 1977 ) .
s.a . larin and j.a.m .
vermaseren , phys . lett . *
b259 * , 345 ( 1991 ) . c. itzykson and j.b .
zuber , _ quantum field theory _
( mcgraw - hill , new york , 1980 ) .
p. fernndez de crdoba and e. oset , phys .
c * 46 * , 1697 ( 1992 ) . h. de vries , c.w . de jager and c. de vries , at .
data nucl .
data tables * 36 * , 583 ( 1971 ) .
martin , w.j . stirling and r.g .
roberts , phys .
d * 51 * , 4756 ( 1995 ) .
m. glck , e. reya and a. vogt , z. phys .
c * 53 * , 651 ( 1992 ) .
e. marco and e. oset , nucl .
* a618 * , 427 ( 1997 ) . | by using a relativistic framework and accurate nuclear spectral functions we evaluate the ratio @xmath0 of deep inelastic neutrino scattering .
parametrizations of this ratio for different values of @xmath1 are provided .
these results should be useful for taking into account the nuclear effects in analyses of experimental data in neutrino reactions in nuclear targets , and test qcd predictions for the nucleon structure functions .
in particular , the size of the nuclear corrections is of the same order of magnitude as the size of the qcd corrections to the gross - llewellyn smith sum rule . |
in a rayleigh - bnard experiment , a horizontal viscous fluid layer is heated from below .
when the temperature difference between upper and lower sides is small , heat transfer solely occurs through thermal conduction .
yet once beyond a critical temperature difference , a regular pattern of convection cells or rolls emerges ( bnard , 1901 ) . this sudden shift from conduction to convection
is referred to as the rayleigh - bnard ( rb ) instability , and is often quoted as an archetypal example of self - organization in non - equilibrium systems ( nicolis and prigogine , 1989 ; prigogine , 1967 ) . intuitively , it makes sense to try to apply the concept of self - organization in non - equilibrium systems to ecologicaly systems , as there are similarities between ecological and physical systems . like the rayleigh - bnard set - up
, ecological systems are open systems that receive a throughput of energy and/or mass via coupling to two environments ( morowitz , 1968 ; schrdinger , 1944 ) .
these two environments are typically large reservoirs and they drive the system from equilibrium .
consider the example of a laboratory chemostat ecosystem ( smith and waltman , 1995 ) .
this is a prime example of a chemotrophic ecosystem whereby a resource of energetic high quality chemical substrate is pumped from a reservoir into the system . in the ecosystem
this resource is degraded into low quality waste products which are emitted to the waste reservoir .
when there is low feeding of resource , no biota can survive , and the resource is degraded by abiotic processes only .
but when the feeding is above a critical threshold , biota can survive by consuming the resource .
there is a sudden shift from a lifeless to a living state . in other words , the energetic quality difference between incoming and outgoing chemical substrates
is exploited by various abiotic and biotic processes .
the latter biotic processes contain the biomass synthesis and turnover of consumer micro - organisms feeding on the resource . so it is tempting to look for a deeper connection .
can one compare biological processing with convection ?
both mechanisms involve self - organizing structures , biological cells or convection cells , that can only survive after a critical threshold .
both energetic pathways , biotic resource conversion and thermal convection , degrade energy from high quality to low quality form . and these energetic pathways are additional to the abiotic conversion or thermal conduction processes of the background . here
, our ambition is to examine the link between ecological processes and convective fluid motions in a quantitative way .
the first part of this article contains a highly intriguing result : the mathematical expressions of the resource - consumer chemostat ecosystem dynamics are exactly the same as the dynamics that describes the basics of the rayleigh - bnard system .
furthermore , not only are the mathematical equations identical , also the physical / ecological interpretations give appealing results .
particularly , by looking at the energetic pathways of the ecosystem , the ecological quantities can be mapped to the quantities used in the fluid system and vice versa .
the second part tries to extend the correspondence between fluid convection and ecosystem functioning to include new processes .
we will study two extensions .
first , one can look at ecological competition and translate the notion of competitive fitness to the fluid system .
the convection cells are in darwinian competition with each other and the fittest ones will survive
. one can generalize the lorenz model to include this fluid competition . as the size of a convection cell will depend on the fitness measure
, we will demonstrate that the mathematical identity of the ecological and the fluid dynamics predicts the experimentally correct size of the cells at the onset of convection .
second , one can look at ecological predation . translating this notion to the fluid system leads to a new conjecture to extend the lorenz model in order to describe more complex convection patterns .
these new patterns only appear when the system is driven beyond a second critical value for the temperature gradient .
the predatory behavior in the convective fluid system leads us to a conjecture which we will not prove , but will be successfully tested by looking at the energy dissipation .
let us start by deriving the dynamics that describes the rayleigh - bnard ( rb ) convective fluid system , named after bnard ( 1901 ) and rayleigh ( 1916 ) who were the first to study this system experimentally and theoretically .
a full mathematical treatment of thermal convection requires the combined solution of the heat transport , navier - stokes and incompressibility equations , resulting in a set of five coupled non - linear partial differential equations ( chandrasekhar , 1961 ; rayleigh , 1916 ) . rather than solving this full set , we employ the approximation adopted by lorenz ( 1963 ) , which became famous as it gave an impulse to the development of chaos theory .
the model describes the lowest modes of an expansion of the temperature and velocity fields for a rb system with free - free boundary conditions ( see e.g. getling 1998 ) . in appendix
[ appendix xyz ] , the derivation of the lorenz system is given in a way that will suit our further discussion . a non - linear set of three ordinary differential equations
is obtained , with three variables ( see fig . [ yzandxfig ] ) : @xmath0 measures the rotational rate of the rolls and represents the maximal velocity at the bottom of the rolls .
@xmath1 and @xmath2 are temperature deviations , where the linear profile of the conduction state is taken as a reference . ,
@xmath1 and @xmath2 .
the temperature ( @xmath3 ) and horizontal velocity ( @xmath4 ) profiles at three vertical sections ( dashed lines ) are shown .
these vertical sections are parallel with the axes of the convection rolls , where the fluid is moving up , moving horizontal or moving down .
the thin linear profiles correspond with the conduction state , the thick profiles with the convection state .
as indicated , @xmath1 and @xmath2 are temperature deviations and @xmath0 is the velocity at the bottom of a roll . ] with these three variables , the xyz lorenz system is rich enough to describe the rayleigh - bnard instability , the sudden shift from conduction to convection .
but there is an even simpler model , the xz system with only two variables , that is rich enough as well .
it is this xz model that allows us to make the correspondence . roughly speaking
, we will perform an averaging over the horizontal directions , such that only the average vertical profile remains . as @xmath1 is the temperature deviation in horizontal direction ,
it is this variable that will disappear after the averaging .
specifically , this is done by making the pseudo steady state assumption @xmath5 for the variable @xmath1 .
the latter becomes a constant and the dynamics turns into : @xmath6 with @xmath7 the height of the fluid layer , @xmath8 a geometric factor such that @xmath9 is the width of the straight convection rolls , @xmath10 the thermal expansion coefficient , @xmath11 the gravitational acceleration , @xmath12 the heat conduction coefficient , @xmath13 the kinematic viscosity and @xmath14 the temperature gradient .
this important quantity is the thermodynamic gradient that drives the system out of equilibrium .
@xmath15 is the high temperature of the heat reservoir below the fluid layer and @xmath16 is the low temperature of the heat reservoir above the layer . for further reference
, we will also need measures for the temperatures at the middle and the lower halve of the fluid layer .
define @xmath17 as the horizontally average temperature at height @xmath18 , and @xmath19 fig .
[ thfig ] shows the interpretation of @xmath15 as a temperature measure for a linearized temperature profile in the lower half of the fluid layer .
( due to symmetry in the approximation leading to the lorenz system , we will not have to include the upper half of the fluid layer . ) .
the lower half of the fluid layer is shown , with the vertical ( horizontally averaged ) temperature profiles in the conduction state ( thin line ) , the convection state ( thicker line ) and the linearized convection profile ( thickest line ) .
the variables @xmath2 and @xmath15 are at height @xmath20 . ]
next , we discuss the ecosystem model , which is in essence a simple chemotrophic resource - consumer food web model , one of the mainstay models of ecology ( e.g. yodzis and innes , 1992 ) .
consumer organisms are feeding on some food resource ( r ) , which is partly converted to consumer biomass ( c ) and partly to waste product ( w ) . for reference
, one can think of a chemostat set - up where a chemical reactor tank contains a monoculture of micro - organisms that are feeding on a chemical substrate like methane or glucose , while respiring @xmath21 . .
the color denotes the energetic quality of the substances , from high ( red ) to low ( light yellow ) . ]
figure [ rcwmodelfig ] gives a schematic overview of the ecosystem coupled with the two environments ( denoted with the superscripts @xmath22 ) .
there are two environmental compartments , the resource at constant concentration @xmath23 and the waste at constant concentration @xmath24 ) and three ecosystem compartments , with variable concentrations @xmath25 , @xmath26 and @xmath27 for the resource , the consumer biomass and the waste respectively . in appendix
[ appendix rcw ] , the complete dynamics of the resource - consumer - waste ( rcw ) ecosystem is given , explaining the fluxes @xmath28 between the compartments .
however , as we will see , the correspondence only works in a limiting case , whereby roughly speaking we will average over the waste concentrations of the system and the environment .
specifically , this can be done by taking a very small relaxation time for the exchange of the waste between the ecosystem and the reservoir .
this means that by studying the ecosystem at longer time scales than this relaxation time , the dynamics for w is forced to be in a pseudo steady state condition .
hence , w is no longer a variable and we end up with the resource - consumer ( rc ) model , with two dynamical equations for two variables : @xmath29 with , @xmath30 the resource exchange rate parameter , @xmath31 the abiotic conversion ( from r to w ) rate parameter , @xmath32 the consumer growth rate parameter , @xmath33 the yield factor for the consumer growth , @xmath34 the consumer decay ( biomass turnover ) rate parameter , and @xmath35 the equilibrium constant for the chemical reaction ( oxidation ) from r to w which always slowly proceeds at the background .
this is the well - known chemostat dynamics ( smith and waltman , 1995 ) , which is extended in two ways : first , abiotic conversion is included in terms of chemical oxidation with parameters @xmath31 and @xmath36 .
second , instead of the classical dependence of the growth on the resource @xmath25 , the growth is now made dependent on @xmath37 .
this is done for thermodynamic consistency : at chemical equilibrium , biomass synthesis should also cease .
.two corresponding models with analogous mechanisms for gradient degradation [ cols="^,^,^",options="header " , ] the abiotic conversion is a chemical reaction with equilibrium constant @xmath35 and a constant abiotic conversion rate parameter @xmath31 . the latter abiotic conversion rate is increased due to a parallel biotic conversion , described by a simple linear functional response with parameter @xmath32 .
this biotic conversion has two parts : a fraction of the resource is used for consumer growth , the other part of the resource turns immediately into waste . from a thermodynamic perspective
, the latter resource turnover is necessary to drive the growth process .
this fractioning is described by the yield parameter @xmath38 : this is the growth efficiency which denotes the amount of resource required to build up one unit of biomass .
the third metabolic transformation is the biotic decay ( biomass turnover ) , represented by the rate constant @xmath34 .
when the resource is turned into waste , the latter is emitted into the waste reservoir from the environment .
the latter has a constant waste concentartion @xmath24 and the exchange flux can be described as @xmath39 putting the two exchange fluxes and the three metabolic fluxes together , the complete dynamics for the resource concentration @xmath40 , the consumer biomass concentration @xmath41 and the waste concentration @xmath42 now look like @xmath43 this is the rcw model .
next , we have to simplify this model to the rc model , by assuming @xmath44 to be very large .
this means that the relaxation time of the waste exchange is negligibly small , and we get the condition that @xmath45 , resulting into ( [ dyn c_r]-[dyn c_c ] ) .
in this appendix , we will give all approximations and a schematic derivation in order to arrive at the lorenz system for the rayleigh - bnard convective fluid ( see berge and pomeau , 1984 or lorenz , 1963 ) .
* there are no pressure terms in the energy balance equation .
* the heat conduction coefficient @xmath12 and the kinetic viscosity @xmath13 are constants . *
the local density field @xmath46 depends on the temperature as @xmath47 with @xmath48 and @xmath49 the constant reference density and temperature , @xmath3 the local temperature field , and @xmath10 the constant thermal expansion coefficient . * the above dependence of the density on the temperature is taken into account only in the gravitational force term in the momentum balance equation . at other places in the equations
, we will write the density as @xmath48 . * the fluid is incompressible ( except in the thermal expansion term ) : @xmath50 , which results in an equality between heat capacities at constant pressure and volume : @xmath51 , or it can be written in terms of the velocity field @xmath52 as : @xmath53 * the local internal energy differential is @xmath54 . with these approximations ,
the heat transport equation can be derived from an energy balance equation , and looks like @xmath55 the first term on the right hand side is the advective heat transport term , and the second is the heat conduction term .
the equation for the velocity field is derived from the momentum balance , and results into the navier - stokes equation . in the boussinesq approximation
, this leads to @xmath56 with @xmath57 the pressure field , @xmath11 the gravitational acceleration and @xmath58 the unit vector in the vertical z - direction . on the right hand side
we see respectively the advection term , the pressure gradient term , the external gravitational force term and the viscous diffusion term . as a final step , in order to fully describe our system , we need boundary conditions .
the boundary condition for the temperature is simply @xmath59 for the velocity , we have @xmath60 because there is no fluid flowing out of the layer .
this is not enough , and we need another condition on the velocity .
we will take free - free boundary conditions to make the description of the solutions easier .
this gives @xmath61 in summary , we have five partial differential equations : three from the three velocity components , one from the incompressibility condition and one from the temperature .
our five local variables are the velocity , pressure and temperature fields .
lorenz made some further assumptions in order to turn these five p.d.e.s into three o.d.e.s with only three global variables .
due to ( [ div v ] ) , one can write the velocity field as @xmath62 , with @xmath63 the streamfunction . we know from experiment that at the onset of convection ( near the critical gradient ) , a convection roll pattern will arise ( getling , 1998 ) .
suppose that the axis of the rolls are along the horizontal y - direction .
hence , there will be no @xmath64 component . the simplest way to obtain
this is by assuming @xmath65 . as a final step
, we will expand the temperature and @xmath67 fields in fourier modes , taking the boundary conditions into account , and we will retain only three of these modes : @xmath68 with the width of the convection cell equal to @xmath9 . in fig .
[ yzandxfig ] a physical interpretation is given to the variables @xmath0 , @xmath1 and @xmath2 .
plugging these expressions into the above partial differential equations ( [ dyn t ] ) and ( [ curlpdev ] ) , and collecting the factors with the same spatial dependence , gives : @xmath69 as can be seen , the system does not close because there is a @xmath70 term .
a final approximation consists of taking this cosine equal to one .
we finally arrive at the lorenz equations , which we will call the xyz model .
next , we have to simplify this xyz model to the xz model , by assuming the pseudo steady state condition for @xmath1 ( i.e. taking @xmath5 ) , resulting into ( [ dyn x]-[dyn z ] ) .
we conclude this appendix with an important remark .
there are two important approximations for the xz model .
the first is the cancelation of the cosine factor in ( [ dyn y lorenz with cos ] ) .
therefore , solutions of the lorenz system are not exact solutions of the complete fluid system in the boussinesq approximation . in this sense , the result of section [ competitive exclusion ] is not trivial , because we arrived at the correct answer whereas the underlying dynamics does not give exact solutions .
our second approximation is the pseudo steady state restriction .
this means that the steady states of the xz system are also steady states of the xyz model ( but as mentioned above , not necessarily of the complete fluid system ) . in this article
we mostly restricted the discussion to the steady states of the xz model , but one should be cautious to use this model to try to find correct transient solutions for the xyz or the complete fluid systems . as an example , the xyz system has chaotic solutions with unstable steady states , whereas these chaotic solutions are absent in the xz model .
the latter has always stable steady states . | both ecological systems and convective fluid systems are examples of open systems which operate far - from - equilibrium .
this article demonstrates that there is a correspondence between a resource - consumer chemostat ecosystem and the rayleigh - bnard ( rb ) convective fluid system .
the lorenz dynamics of the rb system can be translated into an ecosystem dynamics . not only is there a correspondence between the dynamical equations ,
also the physical interpretations show interesting analogies . by using this fluid - ecosystem analogy , we are able to derive the correct value of the size of convection rolls by competitive fitness arguments borrowed from ecology .
we finally conjecture that the lorenz dynamics can be extended to describe more complex convection patterns that resemble ecological predation . * stijn bruers * + instituut voor theoretische fysica , + katholieke universiteit leuven , + celestijnenlaan 200d , b-3001 leuven , belgium * filip meysman * + centre for estuarine and marine ecology , + netherlands institute of ecology ( nioo - knaw ) , + korringaweg 7 , 4401 nt yerseke , the netherlands + _ keywords : _ rayleigh - bnard system ; lorenz model ; resource - consumer chemostat ; ecosystem metabolism ; thermodynamics |
there are twenty six molibdenum isotopes on the segr chart , studied by means of @xmath6-ray spectroscopy @xcite .
they span a wide region , where a variety of shapes are observed , from the spherical @xmath7mo to one of the most deformed nuclei in the @xmath8 mass region .
of particular interest are the @xmath9mo nuclei , which are placed in a region where an offset of the quadrupole deformation takes place @xcite .
also , the two nuclei are placed in the third oscillator shell in protons and at the beginning of the fourth oscillator shell in neutrons , where high-@xmath10 and low-@xmath10 single particle orbits with @xmath11 enhance the effects of octupole collectivity @xcite .
being on the edge between the single - particle and collective modes , the two nuclei represent an excellent laboratory , where different approaches can be tested . in this respect , of particular interest
are the nuclear lifetimes of excited states which are directly related to transition matrix elements and hence are sensitive to the underlying structure .
fast - timing measurements were performed in @xmath9mo .
the nuclei were produced in fusion - evaporation reactions , performed at the nipne tandem accelerator .
a beam of @xmath12o was accelerated up to 62 mev and focused on a @xmath13se target with a thickness of 5 mg/@xmath14 on a 2 mg/@xmath14 thick au backing .
the cross section for the @xmath13se(@xmath12o , 5n@xmath6)@xmath2mo channel was estimated as 400 mb , while that for the @xmath13se(@xmath12o , 4n@xmath6)@xmath5mo channel as about 100 mb .
the beam intensity was of the order of 20 pna .
[ 0.31]mo.,title="fig : " ] the gamma - ray detector system consisted of eight labr@xmath15:ce detectors and eight hpge detectors , working in coincidence@xcite .
the labr@xmath16:ce detectors issue positive dynode and negative anode signals .
the dynode signals were used for energy measurements while the anode signals were used for timing .
the time signals were shaped by constant fraction discriminators ( cfds ) and sent to time - to - amplitude - converters ( tacs ) operating in a common stop mode .
the energy and time signals were digitized and the data were stored in event files , collected for approximately two hours each .
the system was triggered by coincidences between two labr@xmath16:ce and one hpge detector .
the data were analyzed with gaspware and radware @xcite packages .
three dimensional matrices were sorted with labr@xmath15:ce energy on two of the axes and a relative time , defined as @xmath17 , on the third axis .
here , @xmath18 is the moment of interaction of the preceding in the time @xmath6-ray and @xmath19 is the moment at which the delayed @xmath6-ray was detected .
@xmath20 is an arbitrary offset .
if the two @xmath6-rays feed and de - excite a particular nuclear level and @xmath21 denotes the energy of the feeding transition , while @xmath22 is the energy of the delayed transition , then for each event two matrix elements @xmath23 and @xmath24 were incremented .
then , from the time projected @xmath25 and @xmath26 two - dimensional gates , two time spectra , symmetric with respect to the arbitrary offset @xmath20 , were obtained .
if the lifetime of the level of interest is of the order of the detectors time resolution , the centroids @xmath27 of the time distributions will be displaced by 2@xmath28 one from the other , where @xmath28 is the lifetime of the level .
this procedure is known as the centroid shift method @xcite .
it should be noted that the two centroids will overlap in the cases where @xmath21 and @xmath22 feed and de - excite nuclear level with a lifetime shorter than the binning of the converter , which was of 6 ps / channel . for lifetimes longer than the detector time resolution
the slope method was used to determine level lifetimes . to reduce the background in the labr@xmath16:ce spectra
, the @xmath29 matrices were incremented after an @xmath30 gate was implied on any of the hpge detectors .
sample energy spectra , obtained with hpge and labr@xmath16:ce detectors , are presented in fig .
fig.[ene](a ) shows a total energy projection for all hpge detectors .
peaks corresponding to transitions in @xmath2mo are denoted with the transition energies .
fig [ ene](b ) presents the total energy projection obtained with all labr@xmath16:ce detectors . a labr@xmath16:ce spectrum , gated on the 692-kev transition in @xmath2mo in the hpge detectors is shown on fig . [ ene](c ) .
the partial level scheme of @xmath2mo presented on fig .
[ shf ] is based on the coincidence measurements performed in the present study and is consistent with the level scheme presented in ref .
nuclear level spins and parities j@xmath31 , level energies @xmath32 , @xmath6-ray energies @xmath33 , intensities @xmath34 and multipolarities @xmath35 m along with the half - lives @xmath36 of the respective levels in @xmath2mo are listed in table [ tab ] .
gamma - ray intensities for the strongest yrast transitions were deduced from a spectrum gated by the 949-kev transition and normalized to the 594-kev transition . in the cases where a particular level decays by more transitions , spectra gated on the feeding transition were used to normalize the intensity of the weaker transition to the intensity of the strongest transition .
energy and intensity balance was performed . the lowest lying levels in @xmath2mo ( fig .
[ shf ] ) have half - lives shorter than the electronics time binning of 6 ps / channel .
the centroids of the symetric time distributions constructed for the @xmath37 state ( fig . [ time95]a ) overlap within the time binning , which is consistent with the half - life of 2.58(11 ) ps quoted in nndc @xcite .
.[tab]energy levels and transitions in @xmath2mo . [ cols="^,^,^,^,^,^,^,^",options="header " , ] @xmath38from nndc @xcite , unless otherwise noted ; @xmath39from a least - squares fit to e@xmath40 ; @xmath41uncertainty of 1.0 kev ; @xmath42from the @xmath43 difference ; @xmath44no enough statistics in the present experiment .
[ time95](b ) shows the time distributions , sorted for the @xmath45 state in @xmath2mo .
the time spectrum is obtained by gating on 152-kev and 386-kev transitions with the labr@xmath16:ce detectors .
in addition , to reduce the background , gates on 949-kev , 594-kev , 692-kev or 766-kev transitions were applied with the hpge detectors . a half - life of 3.4(4 )
ns was obtained from the slope of the time distribution . according to nndc ,
the @xmath45 state decays via two branches with energies of 38 kev and 386 kev .
the 38-kev transition is highly converted , and its energy is outside of the detectors range of sensitivity .
hence , the transition was not observed experimentally in the present experiment .
therefore , the intensity of the 38-kev transition was estimated from the intensity balance performed for the 2582-kev level and after evaluation of the levels side feeding . using the branching ratios obtained from the present data , @xmath46 w.u .
was obtained .
[ time95](c ) shows symmetric time spectra for the @xmath47 state in @xmath2mo .
the spectrum is gated by the 903-kev and 152-kev transitions in the labr@xmath16:ce detectors . to reduce the background ,
the same hpge gates as for the @xmath48 state were imposed . a half - life of 138(18 )
ps was obtained from the centroid shift method .
a weaker reaction channel leads to the even - even @xmath5mo nucleus .
an analysis similar to the one performed for @xmath2mo was carried out .
two levels with picosecond half - lives were observed .
[ time96](a ) shows the time distributions obtained for 8@xmath49 state in @xmath5mo .
the time spectrum is gated on 809-kev and 538-kev transitions with labr@xmath15:ce detectors and on 812-kev transition with hpge detectors . a half - life of 0.19(4 )
ns was obtained with the centroid shift method .
hence , the @xmath50 is 2.3(5 ) w.u .
time spectra for the 10@xmath51 state in @xmath5mo are presented on fig .
[ time96](b ) .
the same gate as in the case of the @xmath3 state was set on the hpge detectors .
gates on the 1346-kev and 809-kev transitions imposed on labr@xmath15:ce detectors were used to increment the symmetric time spectra .
a half - life of 0.11(5 ) ns was obtained using the centroid shift method . the transition strength for the 809-kev transition results as @xmath520.6(3 ) w.u .
the ground state of @xmath2mo is a @xmath53 state , which is consistent with the ordering of the neutron single - particle levels in the @xmath54 mass region @xcite where @xmath55 single - particle orbit appears at the beginning of the fourth oscillator shell .
next to the @xmath55 single - particle orbit is @xmath56 .
indeed , the @xmath57 level , placed at 767 kev in @xmath2mo , decays via a @xmath58 transition to the ground state with @xmath59 w.u @xcite .
the next yrast @xmath60 level in @xmath2mo has an energy of 948 kev , which is close to the @xmath61 level energy of the neighboring even - even nuclei .
also , the @xmath62 state decays via a strong @xmath63 transition with @xmath64 w.u @xcite .
such a state can be interpreted as a @xmath55 single particle state fully aligned with the first @xmath61 phonon of an even - even core @xcite . in this respect
it is interesting to compare the matrix elements , obtained for the @xmath63 transitions in the even - even @xmath5mo , with the @xmath63 matrix elements , obtained from the yrast transitions in the even - odd @xmath2mo nucleus .
the matrix elements , calculated from @xmath65@xmath66 for the yrast transitions in @xmath9mo , are shown on fig .
[ matrixel ] .
the experimental matrix elements are also compared to the harmonic vibrator matrix elements , which were parameterized from the first excited state in @xmath5mo .
the @xmath67 matrix element , calculated for @xmath5mo @xcite , remarkably coincides with the harmonic vibrator @xmath67 matrix element .
also , the @xmath68 matrix element , calculated from the @xmath2mo data , is consistent with its interpretation as a fully aligned particle - core coupled state .
further , with the increase of the angular momentum , the experimental matrix elements deviate from the harmonic vibrator description .
in fact , the experimental matrix elements are even smaller than the @xmath69 matrix element . in this spin range ,
the odd- and even- mass molibdenum matrix elements are again in the same range . in order to account for
the high-@xmath70 value , these core excited states most probably involve high-@xmath10 single - particle excitations .
this work is supported by the bulgarian science fund - contracts dmu02/1 , drnf02/5 , did-02/16 , romanian uefiscdi - contract idei 115/2011 and by a bulgarian - romanian partnership contract , numbers dnts-02/21 and 460/pnii module iii .
9 nndc data base ( www.nndc.bnl.gov ) | half - lives of the @xmath0 and @xmath1 states in @xmath2mo and of the @xmath3 and @xmath4 states in @xmath5mo were measured .
matrix elements for yrast transitions in @xmath2mo and @xmath5mo are discussed . |
dust properties govern several physical and chemical phenomena in the interstellar medium ( ism ) and act as a tracer of local environmental conditions .
when starlight passes through various components of interstellar dust grains , which are aligned to the galactic magnetic field , the radiation becomes partially plane - polarized , typically at the level of a few percent .
the nature of this polarization reveals important information regarding the shape , size and composition of interstellar dust .
the detailed process(es ) by which grains get aligned with the magnetic fields has ( have ) long been , and is still actively under study ( davis & greenstein 1951 ; dolginov 1990 ; cho & lazarian 2005 ) . in general the dust grains tend to align with their long axis perpendicular to the magnetic field ( purcell 1979 ) .
the clue to the interstellar origin of the polarization came from the observed correlation between the degree of polarization ( @xmath15 ) and the colour excess , @xmath16 .
the values of @xmath15 ( in visual wavelengths ) for the stars with large @xmath16 are found to be in the range between zero and a maximum value given by @xmath17 ( aannestad & purcell 1973 ) .
the relation between @xmath9 and the colour excess , and the variation of @xmath15 with wavelength , are interpreted in terms of the grain properties and the efficiency of the grain alignment .
therefore , polarimetry is a useful technique to investigate the properties like maximum polarization @xmath18 , the wavelength @xmath10 corresponding to @xmath18 and the orientation of the magnetic field in various galactic locations .
the photometric and spectroscopic information already available for stars in open clusters are specifically important to make a meaningful study of the dust grains located in the foreground and in the intracluster regions ( e.g. , trumpler 27 , feinstein et al .
2000 ; ngc 6231 , feinstein et al . 2003
; stock 16 , feinstein et al .
2003 ; hogg 22 & ngc 6204 , martnez et al . 2004 ; ngc 5606 , orsatti et al . 2007
; ngc 5749 , vergne et al . 2007 ;
ic 1805 , medhi et al . 2007 ;
ngc 6250 , feinstein et al . 2008 ; ngc 654 , medhi et al . 2008 ; ngc 6124 , vergne et al . 2010 ; ngc 6823 , medhi et al .
2010 ) . in this paper ,
we present multi - wavelength polarimetry of stars towards ngc 1893 .
we also present @xmath0-band polarimetry of stars towards four additional clusters , @xmath19 ngc 2281 , ngc 1664 , ngc 1960 and stock 8 , to get information about the ism foreground to the cluster ngc 1893 .
the main aim of this study is to investigate the dust properties as a function of distance towards the anticenter direction of the galaxy ( @xmath1 @xmath20 to @xmath7 ) using stars in open clusters .
the basic parameters of the observed clusters are given in table [ basic_para ] .
using these five open clusters , we have made an attempt to study the dust properties of the ism distributed between @xmath21 pc ( ngc 2281 ) and @xmath22 kpc ( ngc 1893 ) .
the paper is organized in the following manner . in section [ instu_obs_data ] , we present a brief discussion on the observations and data reduction .
we present our results in section [ results ] , and analysis and discussion in section [ discuss ] .
the dust components responsible for the observed polarization are discussed in section [ dust_components ] .
we conclude our results in section [ conclude ] . | we present multi - wavelength linear polarimetric observations for 44 stars of the ngc 1893 young open cluster region along with @xmath0-band polarimetric observations of stars of other four open clusters located between @xmath1 @xmath2 160@xmath3 to 175@xmath3 .
we found evidence for the presence of two dust layers located at a distance of @xmath2 170 pc and @xmath2 360 pc .
the dust layers produce a polarization @xmath4 @xmath2 @xmath5 .
it is evident from the clusters studied in the present work that , in the galactic longitude range @xmath1 @xmath2 @xmath6 to @xmath7 and within the galactic plane ( @xmath8 ) , the polarization angles remain almost constant , with a mean @xmath2 163@xmath3 and a dispersion of 6@xmath3 .
the small dispersion in polarization angle could be due to the presence of uniform dust layer beyond 1 kpc .
present observations reveal that in case of ngc 1893 , the foreground two dust layers , in addition to the intracluster medium , seems to be responsible for the polarization effects .
it is also found that towards the direction of ngc 1893 , the dust layer that exists between 2 - 3 kpc has a negligible contribution towards the total observed polarization .
the weighted mean for percentage of polarization ( @xmath9 ) and the wavelength at maximum polarization ( @xmath10 ) are found to be @xmath11 and @xmath12 @xmath13 m respectively .
the estimated mean value of @xmath10 indicates that the average size of the dust grains within the cluster is similar to that in the general interstellar medium .
the spatial variation of the polarization is found to decrease towards the outer region of the cluster . in the present work ,
we support the notion , as already has been shown in previous studies , that polarimetry , in combination with @xmath14 colour - colour diagram , is a useful tool for identifying non - members in a cluster .
[ firstpage ] polarization - dust , extinction - open clusters and associations : individual : ( ngc 2281 , ngc 1664 , ngc 1960 , stock 8 and ngc 1893 ) . |
the neutrino oscillation interpretations of the solar @xcite and atmospheric @xcite neutrino data and the lsnd experiment @xcite require three neutrino mass - squared differences of different orders of magnitude .
since it is not possible to obtain this within the standard model framework of three active neutrinos it has been proposed to introduce a light sterile neutrino @xcite to reconcile all the experimental hints for neutrino oscillations . here
we present an analysis of the global neutrino oscillation data in terms of four - neutrino mass schemes , including data from solar and atmospheric neutrino experiments , the lsnd experiment , as well as data from short - baseline ( sbl ) experiments @xcite and long - baseline reactor experiments @xcite reporting no evidence for oscillations .
this analysis updates the work presented at the meeting and is based on the data of summer 2002 .
we find that for all possible types of four - neutrino schemes different sub - sets of the data are in serious disagreement and hence , four - neutrino oscillations _ do not _ provide a satisfactory description of the global oscillation data including lsnd .
the details of our calculations can be found in refs .
@xcite .
four - neutrino mass schemes are usually divided into the two classes ( 3 + 1 ) and ( 2 + 2 ) , as illustrated in fig .
[ fig:4spectra ] .
we note that ( 3 + 1 ) mass spectra include the three - active neutrino scenario as limiting case . in this case
solar and atmospheric neutrino oscillations are explained by active neutrino oscillations , with mass - squared differences @xmath0 and @xmath1 , and the fourth neutrino state gets completely decoupled .
we will refer to such limiting scenario as ( 3 + 0 ) .
in contrast , the ( 2 + 2 ) spectrum is intrinsically different , as there must be a significant contribution of the sterile neutrino either in solar or in atmospheric neutrino oscillations or in both .
the six types of four - neutrino mass spectra , divided into the classes ( 3 + 1 ) and ( 2 + 2 ) . ]
neglecting cp violation , in general neutrino oscillations in four - neutrino schemes are described by 9 parameters : 3 mass - squared differences and 6 mixing angles in the unitary lepton mixing matrix . here
we use a parameterization introduced in ref .
@xcite , which is based on physically relevant quantities : the 6 parameters @xmath0 , @xmath2 , @xmath1 , @xmath3 , @xmath4 , @xmath5 are similar to the two - neutrino mass - squared differences and mixing angles and are directly related to the oscillations in solar , atmospheric and the lsnd experiments . for the remaining 3 parameters we use @xmath6 and @xmath7 . here , @xmath8 is the fraction of @xmath9 participating in solar oscillations , and ( @xmath10 ) is the fraction of @xmath11 participating in oscillations with @xmath1 ( for exact definitions see ref .
@xcite ) . for the analysis we adopt the following approximations : * we make use of the hierarchy @xmath12
this means that for each data set we consider only one mass - squared difference , the other two are set either to zero or to infinity . * in the analyses of solar and atmospheric data ( but not for sbl data ) we set @xmath13 , which is justified because of strong constraints from reactor experiments @xcite .
parameter dependence of the different data sets in our parameterization . ] due to these approximations the parameter structure of the four - neutrino analysis gets rather simple .
the parameter dependence of the different data sets solar , atmospheric and sbl is illustrated in fig .
[ fig : diagram ] .
we see that only @xmath14 links solar and atmospheric data and @xmath7 links atmospheric and sbl data .
all the other parameters are `` private '' to one data set .
the strong preference of oscillations into active neutrinos in solar and atmospheric oscillations leads to a direct conflict in ( 2 + 2 ) oscillation schemes .
we will now show that thanks to the new sno solar neutrino data @xcite and the improved sk statistic on atmospheric neutrinos @xcite the tension in the data has become so strong that ( 2 + 2 ) oscillation schemes are essentially ruled out .
indeed , latest solar neutrino data lead to the bound @xmath15 at 99% cl , where @xmath14 is the parameter describing the fraction of the sterile neutrino participating in solar neutrino oscillations . in contrast , in ( 2 + 2 ) schemes atmospheric data imply @xmath16 at 99% cl , in clear disagreement with the bound from solar data . , @xmath17 and @xmath18 as a function of @xmath14 in ( 2 + 2 ) oscillation schemes . ] in fig .
[ fig : sol+atm ] we show the @xmath19 for solar data and for atmospheric combined with sbl data as a function of @xmath14 .
furthermore , we show the @xmath19 of the global data defined by @xmath20 from the figure we find that only if we take both data sets at the 99.95% cl a value of @xmath14 exists , which is contained in the allowed regions of both sets .
this follows from the @xmath19-value @xmath21 shown in the figure . in ref .
@xcite we have proposed a statistical method to evaluate the disagreement of different data sets in global analyses .
the _ parameter goodness of fit _
( pg ) makes use of the @xmath22 defined in eq .
( [ chi2solatm ] ) .
this criterion evaluates the gof of the _ combination _ of data sets , without being diluted by a large number of data points , as it happens for the usual gof criterion ( for details see ref .
we find @xmath23 , leading to the marginal pg of @xmath24 .
we conclude that ( 2 + 2 ) oscillation schemes are highly disfavoured by the disagreement between the latest solar and atmospheric neutrino data .
this is a very robust result , independent of whether lsnd is confirmed or disproved .
from sbl and atmospheric neutrino data in ( 3 + 1 ) schemes @xcite compared to the lsnd allowed region @xcite . ] it is known for a long time @xcite that ( 3 + 1 ) mass schemes are disfavoured by the comparison of sbl disappearance data @xcite with the lsnd result . in ref .
@xcite we have calculated an upper bound on the lsnd oscillation amplitude @xmath25 resulting from sbl and atmospheric neutrino data . from fig .
[ fig:3 + 1 ] we see that this bound is incompatible with the signal observed in lsnd at the 95% cl .
only marginal overlap regions exist between the bound and lsnd if both are taken at 99% cl .
an analysis in terms of the parameter goodness of fit @xcite shows that for most values of @xmath4 sbl and atmospheric data are compatible with lsnd only at more than @xmath26 , with one exception around @xmath27 ev@xmath28 , where the pg reaches 1% .
these results show that ( 3 + 1 ) schemes are strongly disfavoured by sbl disappearance data .
with the methods developed in ref .
@xcite we are able to perform a global fit to the oscillation data in the four - neutrino framework .
this approach allows to statistically compare the different hypotheses .
let us first evaluate the gof of ( 3 + 1 ) and ( 2 + 2 ) spectra with the help of the pg method described in ref .
we divide the global oscillation data in the 4 data sets sol , atm , lsnd and nev , where nev contains the experiments karmen @xcite , cdhs @xcite , bugey @xcite and chooz / palo verde @xcite , reporting no evidence for oscillations .
following ref .
@xcite we consider @xmath29 where @xmath30 ( @xmath31 = sol , atm , nev , lsnd ) . in tab .
[ tab : pg ] we show the contributions of the 4 data sets to @xmath32 for ( 3 + 1 ) and ( 2 + 2 ) oscillation schemes .
as expected we observe that in ( 3 + 1 ) schemes the main contribution comes from sbl data due to the tension between lsnd and nev data in these schemes . for ( 2 + 2 ) oscillation schemes
a large part of @xmath33 comes from solar and atmospheric data , however , also sbl data contributes significantly .
this comes mainly from the tension between lsnd and karmen @xcite , which does not depend on the mass scheme and , hence , also contributes in the case of ( 2 + 2 ) .
therefore , the values of @xmath33 in tab .
[ tab : pg ] for ( 2 + 2 ) schemes are higher than the one given in sec .
[ sec:2 + 2 ] , where the tension in sbl data is not included .
.[tab : pg]contributions of different data sets to @xmath33 in ( 3 + 1 ) and ( 2 + 2 ) schemes . [ cols="^,^,^,^,^,^",options="header " , ] the parameter goodness of fit is now obtained by evaluating @xmath33 for 4 dof @xcite .
this number of degrees of freedom corresponds to the 4 parameters @xmath34 describing the coupling of the different data sets ( see eq .
( [ chi2bar ] ) ) .
the best gof is obtained in the ( 3 + 1 ) case . however , even in this best case the pg is only 0.56% .
the pg of @xmath35 for ( 2 + 2 ) schemes shows that these mass schemes are essentially ruled out by the disagreement between the individual data sets .
although we have seen that none of the four - neutrino mass schemes can provide a reasonable good fit to the global oscillation data including lsnd , it might be interesting to consider the _ relative _ status of the three hypotheses ( 3 + 1 ) , ( 2 + 2 ) and the three - active neutrino scenario ( 3 + 0 ) .
this can be done by comparing the @xmath19 values of the best fit point ( which is in the ( 3 + 1 ) scheme ) with the one corresponding to ( 2 + 2 ) and ( 3 + 0 ) .
first we observe that ( 2 + 2 ) schemes are strongly disfavoured with respect to ( 3 + 1 ) with a @xmath36 .
the reason for the big change with respect to the value of @xmath37 found in ref .
@xcite is the improved sensitivity of solar ( sno nc ) and atmospheric ( sk 1489-days ) data to a sterile component . with this
new data now ( 3 + 1 ) schemes are clearly preferred over ( 2 + 2 ) : for 4 dof a @xmath38 implies that ( 2 + 2 ) is ruled out at 99.87% cl with respect to ( 3 + 1 ) .
further we find that ( 2 + 2 ) is only slightly better than ( 3 + 0 ) , which is disfavoured with a @xmath39 with respect to ( 3 + 1 ) .
using latest solar and atmospheric data , we re - analyze the four - neutrino description of current global neutrino oscillation data , including the lsnd evidence for oscillations .
the higher degree of rejection for non - active solar and atmospheric oscillation solutions implied by the sno neutral current result as well as by the latest 1489-day sk atmospheric neutrino data allows us to rule out ( 2 + 2 ) oscillation schemes . using an improved goodness of fit method especially sensitive to the combination of data sets we obtain a gof of only @xmath35 for ( 2 + 2 ) schemes .
further , we find that also ( 3 + 1 ) schemes are strongly disfavoured by the data .
the disagreement between negative short - baseline experiments and lsnd in ( 3 + 1 ) schemes imply a gof of @xmath40 .
this leads to the conclusion that all four - neutrino descriptions of the lsnd anomaly , both in ( 2 + 2 ) as well as ( 3 + 1 ) realizations , are highly disfavoured .
our analysis brings the lsnd hint to a more puzzling status , and the situation will become even more puzzling if lsnd should be confirmed by the up - coming miniboone experiment @xcite .
s. fukuda _ et al .
_ , super - kamiokande coll .
b * 539 * ( 2002 ) 179 ; b.t .
cleveland _ et al .
_ , astrophys .
j. * 496 * , 505 ( 1998 ) ; d.n .
abdurashitov _ et al .
_ , sage coll .
rev . * c60 * , 055801 ( 1999 ) ; astro - ph/0204245 ; w. hampel _ et al .
_ , gallex coll .
, phys . lett .
* b447 * , 127 ( 1999 ) ; c. cattadori , gno coll . , nucl .
b ( proc . suppl . )
* 110 * ( 2002 ) 311 .
j. t. peltoniemi , d. tommasini and j. w. f. valle , phys .
b * 298 * ( 1993 ) 383 ; j. t. peltoniemi and j. w. f. valle , nucl . phys .
b * 406 * , 409 ( 1993 ) ; d.o .
caldwell and r.n .
mohapatra , phys .
d * 48 * ( 1993 ) 3259 .
bilenky , c. giunti and w. grimus , in _ proceedings of neutrino 96 _ , helsinki , finland , 1319 june 1996 , p. 174
j. c * 1 * ( 1998 ) 247 ; n. okada and o. yasuda , int .
phys . a * 12 * ( 1997 ) 3669
; v. barger , _ et at .
d * 58 * ( 1998 ) 093016 ; s.m .
bilenky , c. giunti , w. grimus and t. schwetz , phys .
d * 60 * ( 1999 ) 073007 ; w. grimus and t. schwetz , eur .
j. c * 20 * ( 2001 ) 1 [ hep - ph/0102252 ] . | we present an analysis of the global neutrino oscillation data in terms of four - neutrino mass schemes . we find that the strong preference of oscillations into active neutrinos implied by the latest solar as well as atmospheric neutrino data allows to rule out ( 2 + 2 ) mass schemes , whereas ( 3 + 1 ) schemes are strongly disfavoured by short - baseline experiments .
our analysis shows that four - neutrino oscillations do not provide a satisfactory description of the global neutrino oscillation data including the lsnd result . |
about half of edge - on disk galaxies contain boxy or peanut - shaped bulges ( bpbs , * ? ? ?
their photometric and kinematic properties are consistent with predictions from disk galaxy simulations in which a bpb formed through bar and buckling instabilities @xcite . the galactic bulge shows many characteristics of a bpb formed in this way : a boxy shape in projection @xcite ; a triaxial density distribution @xcite with a dense inner , rounder component ( * ? ? ?
* ; * ? ? ?
* hereafter gmv12 ) ; an x - shaped structure @xcite ; cylindrical rotation @xcite ; and a transition to an outer planar bar ( * ? ? ?
* hereafter mvg11 ) , seen in star count observations as long bar @xcite .
the galactic bulge is also known to consist of predominantly old stars @xcite with a broad , asymmetric metallicity distribution function ( mdf ) @xcite .
the data clearly show a vertical metallicity gradient , such that the more metal - rich part of the mdf thins out towards high latitudes ( @xmath0 , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) . at latitudes lower than baade s window ( @xmath1 )
no vertical gradient is seen @xcite .
these properties are not obviously reconciled with a disk origin of the bulge . because bars tend to mix stars from different radii and so wash out preexisting metallicity gradients @xcite , the observed metallicity gradient in the galactic bulge
has long been taken as a signature for a classical bulge in the milky way .
indeed , @xcite found that they could not reproduce the observed gradient with a simulated boxy bulge which had evolved from a pure disk with an initial metallicity gradient . in order to obtain the lower mean metallicities at high @xmath2
, they needed to include an additional metal - poor thick disk in their initial model . here
we show with the help of a suitable boxy bulge / bar simulation , that contrary to widespread belief boxy bulges can have metallicity gradients similar to those observed in our galaxy and in other boxy bulges , if they evolve from disks with similarly steep radial metallicity gradients .
the simulation used in this work is that already analyzed in mvg11 and gmv12 , with similar characteristics as described in @xcite .
the model evolved from an exponential disk , with @xmath3 , initial radial scale length of @xmath4 kpc and vertical scale height of @xmath5 kpc , embedded in a dark matter halo . following bar and buckling instabilities it developed a prominent boxy bulge which is already relaxed at @xmath6 gyr .
as in gmv12 we consider the simulated galaxy at time @xmath7 gyr when the bar has resumed its growth through further angular momentum transfer to the halo , and we vertically symmetrize the particle distribution to optimize the resolution .
we showed previously that this model reproduces different milky way star count observations , both near the galactic plane for the long bar out to @xmath8 ( mvg11 ) , and in the inner boxy bulge for @xmath9 and @xmath10 ( gmv12 ) as well as for @xmath11 , where the predictions of the model were subsequently confirmed by @xcite . here
we show that the model s boxy bulge at that time also shows approximate cylindrical rotation , with the slope of the mean velocity curve along @xmath12 increasing slightly towards lower latitudes ; see fig .
[ fig : kinematics ] .
this is similar to the brava kinematic data @xcite and the simulation by @xcite .
the cylindrical rotation in boxy bulges is clearest in edge - on view , as shown in previous simulations ( e.g. , * ? ? ?
* ; * ? ? ?
* ) ; it is weakened by the nearly end - on orientation of the milky way bar and the perspective effects . as in our previous work ( mvg11 , gmv12 ) , we here take the solar galactocentric distance @xmath13 kpc , scale the bar length to @xmath14 kpc , and use a bar orientation of @xmath15 with respect to the sun - galactic center line .
we study the formation of bulge metallicity gradients in the unstable disk
boxy bulge scenario with a simple illustrative model .
we assume that the initial exponential disk follows a specified radial metallicity gradient , which we imagine is set up during disk build - up prior to bar formation .
we assign to each particle a metallicity depending on its initial radius ( e.g. , * ? ? ?
* ) , according to @xmath16(r)=[m / h]_0+\alpha_{r}\times r/{\rm kpc}$ ] . after some exploration
we choose @xmath16_0=0.6 $ ] and @xmath17 .
unlike @xcite , we do not link these metallicities to the present - day galactic disk near the sun , because the buckling instability in the milky way must have happened long ago when the outer disk may have been only incompletely assembled .
the bar and buckling instabilities leading to the formation of the boxy bulge strongly change the particle orbits . thereby the initial disk metallicity distribution is mapped into the bulge accordingly .
[ fig : rv_grad ] shows longitudinal and vertical metallicity profiles for the bulge at @xmath18 .
the vertical gradient for @xmath19 along the minor axis is @xmath20 dex / deg , compared to the observed slope in the milky way bulge @xmath21
dex / deg from @xcite and @xmath22 dex / deg from @xcite . near the galactic plane , @xmath23 , the vertical gradient becomes flatter and even positive .
the positive gradient is due to the contamination from foreground / background disk particles .
the flat part comes from the mixing during the buckling instability in the inner regions that evolve into the central near - spheroidal component ( gmv12 ) , and also from the greater fraction of low - latitude outer bar particles .
the absence of a clear vertical gradient near the galactic plane is consistent with recent results from @xcite .
however , their measured mean metallicities for m - giants ( [ fe / h]@xmath24 ) are slightly lower in baade s window than those for k - giants @xcite and for our simple model .
[ fig : vgrad ] shows the resulting mdfs in several minor axis fields in the boxy bulge at @xmath25 gyr .
note the clear gradient of mean metallicity with latitude , and that similarly to the observed histograms , it arises mostly from the decreasing fraction of higher metallicity stars away from the galactic plane .
the maximum metallicity in all histograms is @xmath26 , the central value in the initial disk , and the lowest value is @xmath27 .
2 gyr later , at @xmath28 gyr , the model metallicity gradient is nearly identical , while the mean metallicities in the minor axis fields have slightly decreased , due to the capture of lower - metallicity disk stars by the slowly growing bar .
had we assumed a shallower radial gradient in the initial disk , also the final vertical gradient in the bulge would be smaller .
so far we have shown that a radial metallicity gradient in the unstable initial disk can survive through the bar and buckling instabilities , and generate , for suitable parameters , a vertical metallicity gradient similar to that observed in the milky way bulge .
for comparison with upcoming survey results ( e.g. , vvv , argos , gaia - eso ) , we now provide a full metallicity map , which gives a different view of the predicted metallicities for this scenario . differently from fig .
[ fig : vgrad ] , we exclude foreground / background disk stars with a distance cut , using only particles with @xmath29 kpc from the position of the sun .
the assumed initial disk metallicity distribution and simulated galaxy snapshot are as before .
[ fig : map_grad ] shows the resulting average bulge metallicity map over the area on the sky extended by the galactic bulge .
it has several noteworthy properties : \(i ) the approximate outline of the boxy bulge can be seen together with low - metallicity indentations in the galactic plane .
the latter are due to foregroud / background stars in the planar bar , which in these positions dominate the bulge stars .
\(ii ) a metallicity gradient is present in all directions on the sky , both vertically and radially .
this is expected from the binding energy argument of sect .
[ sec : mechanism ] .
however , in the central few degrees the gradients become shallower .
\(iii ) iso - metallicity contours are more elongated vertically than horizontally , whereas the surface density contours are flattened to the plane ( mvg11 ) .
\(iv ) the asymmetry between @xmath30 and @xmath31 at intermediate latitudes is due to perspective effects and signifies the bar origin . in fig .
[ fig : map_grad ] high metallicities extend to larger @xmath32 on the @xmath30 side at @xmath33 , but for @xmath34 they extend to larger @xmath35 on the @xmath31 side .
this map has some remarkable similarities with that being constructed from the near - infrared photometric vista vvv survey @xcite .
for now we can compare quantitatively with a few more measured values in off - axis fields .
details will of course depend on the parametrisation of our simple model . in the longitudinal direction the gradient along @xmath36 is @xmath37 dex / deg ( see fig . [
fig : rv_grad ] ) .
@xcite found mean metallicities [ fe / h]@xmath38 in two fields at @xmath39 and @xmath40 ) for bulge stars selected with [ fe / h]@xmath41 .
the model values in these fields are -0.30 and -0.28 .
clearly , while the bar and buckling instabilities scramble the orbits of disk stars , they do not do so enough to completely erase the preexisting metallicity gradient .
high - metallicity stars tightly bound to the galactic center initially remain more tightly bound in the final bulge , and initially more loosely bound stars with lower metallicities end up at larger final radii , on average .
this can be quantified by considering the change in jacobi energy @xmath42 in the rotating frame of the boxy bulge and bar during the evolution .
here we use standard cylindrical coordinates , @xmath43 is total velocity , @xmath44 is the gravitational potential , and @xmath45 is the pattern speed at @xmath25 gyr . fig .
[ fig : energy ] shows that particles are scattered in jacobi energy by the two instabilities , but only over a fraction of the available total range in @xmath46 .
i.e. , they retain some memory of their initial values . it has also been shown that most bar particles inside the vertical inner lindblad resonance ( vilr ) mix during the buckling instability , but stay in the inner bar regions , while most bar particles around vilr are scattered to orbits which can visit larger heights @xcite . from these arguments
follows that the number of high - metallicity particles in the final boxy bulge will decrease with height above the plane where @xmath47 becomes less negative .
in the histograms of fig . [
fig : vgrad ] , therefore , the number of stars on the metal - rich part of the distribution decreases with latitude .
the maximum metallicity in all histograms , but that at @xmath48 , is still given by the assumed maximum metallicity at the center of the initial disk , here @xmath49 , because a small fraction of the tightly bound particles is scattered up even to @xmath50 .
the low - metallicity tail , on the other hand , is more prominent at high latitudes , because of the larger fraction of particles coming from the outer parts of the bar .
note that within the assumed model , a lower limit to the metallicity in the bulge fields is expected , which is set by the outer boundary of the part of the disk which participates in the instability .
from the histograms in fig .
[ fig : vgrad ] , this is @xmath51 , which is due to particles in the initial disk at @xmath52 , just inside the corotation radius of the bar before buckling , @xmath53 kpc .
these properties of the model histograms are surprisingly similar to those of the galactic bulge mdfs discussed in the literature .
if our scenario were the full explanation for the metallicity structure of the galactic bulge , the observed range in the metallicity histograms could be used directly to estimate the central metallicity and the radial gradient in the inner galactic disk prior to bar formation and buckling .
the observational histograms , however , also contain metal - poor stellar halo and thick disk stars , and possibly metal - rich disk stars formed after the buckling , which must be taken into account in this argument .
finally we note that a radial metallicity gradient in the precursor disk could have been accompanied by a stellar age gradient . in the context of the present model , depending on the time of buckling
, this could lead to a measurable vertical age gradient in the boxy bulge which would provide an independent test of the model .
vertical gradients have so far been published only for a small number of galaxies with boxy bulges .
@xcite found a vertical metallicity gradient in ngc 7332 comparable to that in the milky way .
@xcite determined vertical gradients in two other boxy bulges , ngc 1381 and ngc 3390 .
they integrated along slits parallel to the major axis to increase signal - to - noise , and obtained a single bulge measurement at several @xmath54 .
a similar averaging over @xmath55 $ ] in our model reduces the measured gradient from @xmath56 dex / kpc to @xmath57 dex / kpc , because radial and vertical gradients are comparable . still , in ngc 1381 the observations show a strong gradient of averaged @xmath58 dex / kpc , while only a weak gradient is seen in ngc 3390 , @xmath59 dex / kpc ( @xmath60 $ ] arcsec ) .
@xcite found negative metallicity gradients along the _ major axis _ for the majority of bulges in barred galaxies , while @xcite found a range of major axis metallicity gradients in boxy bulges , from negative to positive , with shallower sample average than for unbarred early - type galaxies . in summary ,
the limited data suggest a range of metallicity gradients in boxy bulges , with the metallicity gradient in the milky way on the steep side . in the framework discussed here , this suggests a range of initial disk metallicity gradients before buckling in these galaxies , perhaps depending on the time of bulge formation .
the results presented in the previous subsections are consistent with early n - body simulation work by @xcite .
he found that pre - existing _ vertical _ abundance gradients in the disk were quickly flattened both in the bar and disk regions but not entirely erased .
their models also included a shallow , outward _ radial _ gradient , which became much flatter in the disk region due to the mixing induced by the bar , but was approximately preserved in the bar region .
@xcite also stated that when no initial vertical gradient was present , a negative vertical gradient appeared at @xmath61 .
however , in his model the initial gradient was @xmath62 dex / kpc , while the model analyzed here has initial @xmath17 dex / kpc , for similar bar size . in our model the final gradient along the bar is @xmath63 dex / kpc and the final vertical gradient is @xmath56 dex / kpc ( for @xmath64 $ ] ) .
consistent with @xcite we also find that at larger @xmath12 the vertical gradient is shallower ( fig [ fig : rv_grad ] , _ right _ ) .
recent n - body numerical simulations by @xcite for a pure exponential disk initial model ( pds in their nomenclature ) showed only a very shallow vertical gradient in the final boxy bulge .
the difference to our results can be traced to the different initial disk metallicity profile .
while their inner disk parameters are similar to ours , due to linking the profile to the present metallicity near the sun , their profile becomes quite shallow beyond @xmath65 times the eventual final bar length .
thus the stars in the upper bulge which come from the outer parts of the bar are quite metal - rich .
the difference , and the reason for the steep vertical gradient in the bulge of our model is its steep initial radial gradient in the disk
. finally we note that if the secular evolution proceeds slowly , a radial gradient in the disk before the instability may be ( partially ) erased by migration processes .
thus a steep final bulge gradient is favoured by rapid secular evolution such as in the model investigated here .
we can summarize our conclusions as follows : \(i ) the vertical metallicity gradient observed in the galactic bulge can be reproduced with a secularly evolved barred galaxy model in which a boxy bulge formed after a buckling instability . in itself , a vertical gradient is therefore not a strong argument for the existence of a classical bulge in the milky way .
the success of our simple model in explaining milky way observations suggests that its underlying idea has some merit and should be pursued further .
future work also needs to consider chemical evolution and possible vertical metallicity gradients in the initial disk , star formation induced by the bar , the possible late build - up of a metal rich inner disk , and the contribution of halo stars and other components @xcite to the bulge mdfs .
another important next step for understanding the origin of the milky way bulge is to investigate the correlations between metallicities and kinematics in this model and in generalisations of it .
comparison with similar observations @xcite is likely to shed light on the possible multi - component nature of the galactic bulge .
the results will be important also for interpreting the data for bulges in external galaxies . | observations show a clear vertical metallicity gradient in the galactic bulge , which is often taken as a signature of dissipative processes in the formation of a classical bulge .
various evidence shows , however , that the milky way is a barred galaxy with a boxy bulge representing the inner three - dimensional part of the bar . here
we show with a secular evolution n - body model that a boxy bulge formed through bar and buckling instabilities can show vertical metallicity gradients similar to the observed gradient , if the initial axisymmetric disk had a comparable radial metallicity gradient . in this framework
the range of metallicities in bulge fields constrains the chemical structure of the galactic disk at early times , before bar formation .
our secular evolution model was previously shown to reproduce inner galaxy star counts and we show here that it also has cylindrical rotation .
we use it to predict a full mean metallicity map across the galactic bulge from a simple metallicity model for the initial disk .
this map shows a general outward gradient on the sky as well as longitudinal perspective asymmetries
. we also briefly comment on interpreting metallicity gradient observations in external boxy bulges .
[ firstpage ] |
the origin of planetary rings is still a debated question .
saturn s main rings are unique as they are made of 90 - 95% water ice @xcite with a mass of @xmath0 kg @xcite .
in contrast , the much less massive rings of uranus and neptune are dark and likely to have a higher rock content @xcite than saturn s rings . whereas dusty saturn s e and g rings are likely to be formed via the destruction or surface erosion of the nearby present satellites @xcite , saturn s main rings can not result from the same process as there is no obvious source of material to feed them today .
note , however , that a recent study shows that the origin of saturn s f ring and uranian @xmath6 ring could be a natural consequence of the collisional destruction between small satellites just outside the main rings @xcite that is formed by the spreading of ancient rings @xcite .
+ several ring formation scenarios have been proposed for massive rings , like those of saturn : ( 1 ) primordial satellite collisional destruction by passing comet @xcite , ( 2 ) tidal destruction of a primordial satellite at the roche limit after inward migration due to tidal interaction with the circum - saturn gas disk @xcite and ( 3 ) tidal disruption of passing objects @xcite
. the inward migration of a primordial titan - sized satellite and the removal of only its pure icy mantle could beautifully explain the silicate deficit of saturn s rings .
however , it requires some fine - tuning of the timing of the event ( likely at the end of the evolution of the circum - saturn circumplanetary disk ) so that the disk is still massive enough to allow inward migration of the satellite , but light enough in order to prevent the rapid infall of debris into the planet because of gas drag .
in addition , it would be difficult to directly form centimetre- to meter - sized particles that are currently seen in saturn s main rings by tidal destruction alone .
@xcite proposes collision between fragments can form small particles , but detailed studies are still required .
tidal disruption of a passing differentiated object could also potentially explain the high ice / rock fraction by capturing only the icy mantle of the incoming body and letting the remnant core escape from saturn s gravity field .
however , so far this scenario has scarcely been studied and only used a simplified analytical model of a homogeneous body @xcite .
thus , direct numerical simulation of the tidal splitting of a big differentiated body is necessary now to investigate this scenario in detail .
in addition , even though some mass capturing occurs , the fragments are expected to have highly eccentric orbits around the planet and the long - term evolution of such fragments remains unclear , in particular by which process a ring of cm - sized particles forms . in this work , we investigate , for the first time , the details of tidal disruption of a passing large differentiated object and the long - term fate of its debris by using direct simulations .
+ such an event may have occurred , with the most probability , either during the phase of planet formation where the giant - planets cores are expected to scatter the neighbouring planetesimals efficiently , or later during the late heavy bombardment ( lhb ) .
the well known `` nice model '' explains not only the lunar cataclysm or today s orbital architecture of giant planets @xcite , but also the implantation of jupiter s trojan asteroids as well as the irregular satellites around giant planets @xcite . during this instability phase
, giant planets could have experienced a significant number of close encounters with bodies scattered from the primordial kuiper belt that surrounded the giant planets .
this belt could have been significantly massive and may have contained a larger number of big objects than what is currently observed in the kuiper belt @xcite .
@xcite estimated an encounter rate of the primordial kuiper belt objects ( kbos ) with the giant planets during the lhb and investigated the captured mass around giant planets using a simplified analytical model assuming homogeneous small bodies like comets .
the flux of such undifferentiated small bodies is enormous and isotropic , and thus the average angular momentum of captured mass should be almost zero , resulting in no contribution to the formation of the rings .
however , only a single tidal disruption of a large object that deviates from the average could decide the story , and such large objects could be expected to be differentiated like pluto .
+ here we use two different direct simulations and investigate successive processes from tidal destruction of a passing object to the possible formation of planetary rings ( figure [ summary ] )
. our work will address ( 1 ) the captured mass as well as its ice / silicate fraction due to the tidal disruption of differentiated bodies at different planets , ( 2 ) the orbits of the captured fragments , and ( 3 ) the long - term orbital and collisional evolution of the captured fragments .
to do so , we first investigate the physics of tidal disruption of a differentiated object that is initially on a hyperbolic orbit about saturn or uranus , and calculate how much mass is gravitationally captured around these planets by using smooth particle hydrodynamics ( sph ) simulations . then we perform direct n - body simulations in order to see the longer - term evolution of such captured fragments around saturn , including the effect of oblateness potential of the planet . in section 2 , in light of our newly derived semi - analytical models that take into account spin and self - gravity of a differentiated object , we briefly review a previous analytical formula of the capture efficiency of tidal disruption . in section 3
, we explain our sph method and model . in section 4 ,
we show the results from sph simulations and discuss the mass capture efficiency as well as orbits of captured fragments . in section 5 , using the data obtained from sph simulations as initial conditions , we perform n - body simulations of subsequent long - term evolution of captured fragments .
then , using results of the simulations as well as analytic estimation , we discuss the fate of the captured fragments .
section 6 summarises our results and discusses the origin of planetary rings .
@xcite derived the mass that is captured on bound orbits around a planet during tidal disruption at close encounters . assuming uniform energy distribution across the body , the amplitude of energy variation is , @xmath7 where @xmath8 , @xmath9 , @xmath10 and @xmath11 are the gravitational constant , the mass of the planet , pericenter distance and radius of the body , respectively .
the mass fraction that is captured with @xmath12 is obtained assuming energy distribution is uniform between @xmath13 and @xmath14 as @xmath15 where @xmath16 is the velocity of the object at infinity and we take @xmath17 where @xmath18 is the planet s hill radius .
equation [ dones1991 ] ( hereafter we call this dones formula ) has been used to estimate mass captured around giant planets in @xcite .
however , it has not been well established if this formula is applicable in any case .
the above dones formula considers only ballistic orbits for all constituent particles of the body .
however , the spin state of the body as well as the body s self - gravity could play an important role in the capture efficiency , especially when the body is massive . here , we consider a spherical body located at pericenter during its close encounter with saturn ( mass @xmath19 ) .
we three - dimensionally divide a cubic box that contains the body into @xmath20 cells by using a cartesian grid , with @xmath21 along 1-dimension .
the box width is @xmath22 times the diameter of the spherical body and the object s center of mass is located at the center of the box .
each cell @xmath23 within the body has its mass @xmath24 , position @xmath25 and velocity @xmath26 relative to saturn s center .
then , we calculate the energy @xmath27 of each cell , and assume any cell that satisfies @xmath28 is gravitationally captured around saturn . the body is a differentiated spherical body of mass @xmath2 kg or @xmath3 kg and the mass fraction of the core is 0.5 .
the densities of the core and the mantle are @xmath29kg / m@xmath30 and @xmath31kg / m@xmath30 , respectively . using the above procedure
, we calculate the capture efficiency at different pericenter distances ranging from @xmath32 km to @xmath33 km .
figure [ modeled_body ] shows an example of our modelled passing body with pericenter distance @xmath34 km and velocity at infinity @xmath35km / s .
figure [ effect_of_spin ] shows the capture efficiency obtained by our semi - analytical model described above .
we include the spin velocity around the object s center of mass with spin period @xmath36h in the prograde or retrograde direction . here
, prograde rotation means that the body rotates in the same direction as its orbit .
closer to saturn , the energy decreases due to saturn s potential field .
therefore , inside the body , cells closer to saturn are more prone to having negative energy . however , with retrograde spin , the radial inner half of the body has a higher spin velocity in the same direction as its orbital velocity at pericenter , and thus achieves larger relative velocity at pericenter compared to the no spin case .
therefore , capture efficiency becomes lower in the case of retrograde spin .
in contrast , when a body has prograde spin , the radial inner half of the body has a smaller relative velocity with respect to saturn , resulting in a larger capture efficiency as seen in figure [ effect_of_spin ] . during the tidal disruption of a passing body on a parabolic orbit , dones formula assumes that the body is instantaneously destroyed and that the inner half is captured and the outer half escapes ( thus , the capture efficiency becomes 0.5 ) .
since a passing object on a hyperbolic orbit has a non - zero value of velocity at infinity , the capture efficiency should be smaller than 0.5 when there is no spin .
therefore , a larger amount of mass escapes from the planet s gravity field than is captured .
neither dones formula nor the above formula includes the effect of self - gravity between components of the body .
however , the escaping fragments could gravitationally attract other non - captured fragments , and thus , more fragments could escape .
the specific gravity of the passing object can be expressed as @xmath37 , where @xmath38 is the distance from the center of mass of the object .
therefore , the object s gravity becomes stronger as the mass of the object becomes larger , and thus , the effect of self - gravity could play a significant role in the capture efficiency at a larger body .
+ thus , we model the effect of self - gravity by considering what we call `` hill capture '' . we calculate the hill radius of the escaping mass @xmath39 , where @xmath40 is the distance between saturn and the center of mass of the escaping mass .
cells that satisfy @xmath28 but are within the hill sphere of escaping fragments are considered as `` hill captured '' by the escaping material .
therefore , the mass that is captured by saturn is the mass that is @xmath28 and lies outside the hill sphere of the escaping mass .
+ figure [ effect_of_grav ] shows the capture efficiency including hill capture .
compared to figure [ effect_of_spin ] , the self - gravity lowers the capture efficiency , especially at larger pericenter distance .
this is because , as the pericenter distance increases , the escaping mass as well as the hill radius increases , thus hill capture becomes more efficient .
compared to dones formula , in the case of @xmath41 kg , the captured mass is always lower than dones formula .
however , when @xmath2 kg , the capture efficiency could become larger than dones formula ( see the case of prograde spin in figure [ effect_of_grav ] left panel ) . in this case , it seems that the effect of spin dominates over the effect of self - gravity .
+ the capture efficiency could either be larger or lower than dones formula as seen in figure [ effect_of_grav ] , depending on spin state and size of the body .
therefore , it is crucial to consider the actual spin state of the body at pericenter . as we show in the next section , the spin at pericenter could be significantly different from the initial spin state at infinity due to the planet tidal torque .
in addition , the object is tidally deformed during the close encounter and may be far from spherical at closest approach ( see next section ) .
therefore , even though our above procedures predict better capture efficiency than dones formula , direct simulations are necessary in order to obtain a more accurate result .
note that the classical definition of the roche limit is the critical distance within which a self - gravitational object separates due to tides and this implicitly assumes an infinite duration of the body experiencing the tidal force .
however , since the time spent in the roche limit by a passing object on a hyperbolic orbit is limited , the tidal disruption and fragment capture around a planet only occurs well inside the roche limit ( see also figures [ capture_efficiency_saturn ] and [ capture_efficiency_uranus ] ) .
+ in this subsection , we estimate the number of objects that experience close encounters with giant planets within their roche limit during the lhb . following the procedure in @xcite ,
the total number of bodies with radius @xmath42 that pass between a giant planet s surface and its roche limit , @xmath43 , can be calculated as @xmath44 where , @xmath45 is the total number of bodies with radius @xmath42 in the primordial kuiper belt , @xmath46 is the intrinsic impact probability onto giant planets up to the age of the solar system , @xmath47 is the mean velocity at infinity before the impact , @xmath48 ) is the escape velocity at the roche limit for icy material . using values obtained by re - processing the lhb simulations presented in @xcite , we get @xmath49 km@xmath50 and @xmath51 km / s for jupiter , @xmath52 km@xmath50 and @xmath53 km / s for saturn ,
@xmath54 km@xmath50 and @xmath55 km / s for uranus , and @xmath56 km@xmath50 and @xmath57 km / s for neptune , respectively ( private communication ) also provides similar collision probabilities with slightly ( @xmath58 times ) higher values for jupiter and saturn . ] .
then , using these values , we obtain @xmath59 @xmath60 , @xmath61 , @xmath62 , and @xmath63 for jupiter , saturn , uranus and neptune , respectively .
on the other hand , a recent model of primordial kuiper belt formation implies the existence of 1000 to 4000 pluto - sized objects ( @xmath64 kg ) and @xmath65 1000 bodies with a mass twice as massive as pluto @xcite .
thus , it is expected that the giant planets experience at least @xmath65 8 , 4 , 2 and 2 close encounters with pluto - sized bodies within their roche limit at jupiter , saturn , uranus and neptune , respectively , and up to a maximum of 32 , 16 , 8 and 8 encounters . in conclusion ,
the passing of pluto - sized objects was not a rare event during the lhb and may have happened several times for each giant planet .
using the smoothed particle hydrodynamics ( sph ) method , we perform simulations of large differentiated bodies passing close to saturn or uranus in order to understand the detailed physical process of tidal destruction .
the sph method is a lagrangian method @xcite in which hydrodynamic equations are solved by considering averaged values of particles through kernel - weighted summation .
we applied the tillotson equation of state @xcite to calculate the pressure from the internal energy and density .
for the artificial viscosity , we use a von neumann - richtmyer - type viscosity with the standard parameter sets ( @xmath66 and @xmath67 ) .
our numerical code is the same as that used in @xcite and more details are described in their papers .
+ the passing body is assumed to be differentiated with 50w% silicate core represented by basalt and 50w% icy mantle represented by water ice .
we used the parameter sets of basalt and water ice for tillotson eos listed in @xcite .
the total mass of the body is set to be @xmath2 or @xmath3 kg .
we use the total number of @xmath68 sph particles . in our simulations , the gravity of the planet
is also taken into account .
saturn and uranus are represented by a point mass located at the origin of the coordinate system and their masses are @xmath69 kg and @xmath70 kg , respectively .
the initial spherical body for our encounter simulations is created by distributing the sph particles in a 3d lattice ( face - centered cubic ) and then by performing sph calculation in isolation until the velocity of particles is much smaller than the escape velocity of the sphere .
+ initial positions and velocities of all particles follow a hyperbolic orbit around the planet given analytically with initial distance to the planet set to @xmath71 km for saturn and @xmath72 km for uranus , which are about twice the roche limit of water ice material .
these initial distances are large enough to avoid the tidal effect of the planet ( hyodo & ohtsuki 2014 ) .
the hyperbolic orbit is entirely determined by the pericenter distance @xmath10 and the velocity at infinity @xmath16 . in our work
, we investigate the dependence on pericenter distance but assume a fixed velocity at infinity @xmath73km / s for saturn and @xmath74km / s for uranus , which are the expected values during the lhb @xcite .
we note that changing the value of velocity at infinity strongly affects the outcome .
therefore , the outcome is sensitive to the initial states and we will leave this matter for future work .
we also investigate the effect of initial spin state of the passing body .
the spin - axis of the initial spherical body is assumed to be perpendicular to the orbital plane with either prograde or retrograde spin to the direction of the hyperbolic orbit with the spin period @xmath36h .
such a rotation period is common in the trans - neptunian belt @xcite .
we stop our calculation when the mass evolution ( e.g. captured mass ) in the system achieves a steady - state .
first , we consider the simplest case where the passing body is homogeneous and made of water ice . we perform sph simulations of close encounters with saturn with @xmath35km / s .
then , we calculate mass that is gravitationally captured around the planet .
figure [ homogeneous ] shows the captured mass ( blue points ) against different pericenter distances of the passing object and comparison to dones formula ( equation [ dones1991 ] ) .
data obtained from sph simulations are in very good agreement with dones formula when the pericenter distance is below @xmath75 m but significant deviations are seen at larger distances .
figures [ homo_q7 - 0 ] and [ homo_q9 - 1 ] show snapshots of our simulations for pericenter distances @xmath76 m and @xmath77 m , respectively . when the pericenter distance is small enough ( @xmath78 m ) ,
the passing body experiences strong tides and is elongated homogeneously into a needle - like structure .
then , due to a small scale gravitational instability , the body splits into several similar - sized clumps ( figure [ homo_q7 - 0 ] ) .
however , for pericenter distances large enough ( @xmath79 m ) , tides are weaker and the body is no longer stretched , but split into fewer , larger clumps of different sizes ( figure [ homo_q9 - 1 ] ) , which differ from the description in @xcite that assumes a uniform distribution of the energy in the fragments .
therefore , it seems that dones formula is valid when the body experiences an encounter very close to the planet as long as the body is homogeneously elongated .
we now return to the case of a differentiated body .
+ figure [ capture_efficiency_saturn ] shows the capture efficiency around saturn obtained from our sph simulations ( dots ) for incoming differentiated bodies with masses @xmath80 kg ( left panels ) and @xmath3 kg ( right panels ) as well as dones formula ( dashed line ) .
the different coloured dots represent the silicate mass fraction of the total captured mass ( @xmath81 ) .
the light blue regions in the panels represent areas within the radius of the planet , and thus such encounters never actually occur ( here , in order to understand detailed physics of tidal disruption , we consider these cases too ) .
+ like in the case of a homogeneous body , numerical results deviate from dones formula .
the slight jump observed in the capture efficiency in the case of prograde and no - spin cases ( e.g. between @xmath82 m and @xmath83 m in the case of @xmath41 kg with no initial spin ) corresponds to the point where the core also starts to be tidally destroyed ( see also figures [ differ_t-8h_q5 - 6 ] and [ differ_tinf_q7 - 0 ] ) .
as discussed in the previous section , in the case of a light object ( case of @xmath2 kg , left panels ) , the prograde spin effects dominate over the self - gravity effect ( hill capture ) and thus more mass than that predicted by dones formula is captured .
conversely , in the case of a heavier object ( case of @xmath41 kg , right panels ) , the effect of self - gravity ( hill capture ) dominates and more mass escapes ( thus , the capture efficiency decreases ) .
+ on the other hand , in the case of retrograde spins ( both left and right bottom panels in figure [ capture_efficiency_saturn ] ) , neither dones formula nor our semi - analytical model is applicable .
in addition , at small pericenter distances ( below @xmath84 m ) , capture efficiency becomes constant at around 0.08 . during such close encounters , an object with an initial retrograde
is spun up in the prograde direction by tidal forces and experiences a major and continuous complex change of its shape .
we also confirm an increase in the internal energy of the constituent sph particles during the encounter .
thereby the body heats up and mostly ends up splitting into two large objects ( figure [ differ_t8h_q4 - 2 ] ) , but the increase of internal energy is not enough to melt the icy material .
these effects are not included either in dones formula or our semi - analytical model .
however , in the prograde and no - spin cases , even though our semi - analytical model can not reproduce numerical results , it is still better than dones formula as it includes spin effects and self - gravity ( hill capture ) .
+ figure [ capture_efficiency_uranus ] is the same as figure [ capture_efficiency_saturn ] but for the case of uranus .
general trends of the capture efficiency are similar to the case of saturn . however , since the density of uranus is larger than that of saturn , the passing object can physically pass deeper inside the potential field of the planet .
thus , the objects can be tidally destroyed more significantly than in the case of saturn without directly colliding with the planet ( outside the light blue region in the figure ) , resulting in a higher silicate fraction of the captured mass . in this section ,
we discuss the orbits of captured fragments . during the encounter
, the orbital kinematic energy is redistributed among the fragments and thus it changes their semi - major axes .
then , due to the conservation of the orbital angular momentum , captured fragments have smaller eccentricities and escaping fragments have larger eccentricities than the initial body .
the eccentricity distribution depends on the initial size of the body and so the distribution range is wider for initially larger bodies as they can be more elongated by tides . considering the conservation of the specific orbital angular momentum of a passing object @xmath85
, we can derive the relationship between semi - major axis @xmath86 and eccentricity @xmath87 as @xmath88 figure [ ea_fragments ] shows eccentricities of the captured fragments ( black dots ) as well as the relation derived from equation ( [ ea_capture_analytic ] ) ( case of @xmath89 kg ( left panel ) and @xmath3 kg ( right panel ) , prograde spin with @xmath36h , @xmath90 m and @xmath73km / s around saturn ) .
the slight increase of the eccentricities of sph data from the analytical estimation is due to the transfer of the orbital angular momentum to the spin angular momentum of the body during the encounter .
+ thus , captured fragments have large eccentricities around @xmath91 ( case of @xmath3 kg body ) and @xmath92 ( case of @xmath93 kg body ) .
our sph simulations are stopped before the captured fragments reach their apocenter .
however , if their apocenter distance is larger than the planet s hill radius , or their pericenter distance is smaller than the planet s radius , then they will either escape from or collide with the planet , so that they can not be incorporated into the rings .
figures [ pericenter_apocenter_saturn ] and [ pericenter_apocenter_uranus ] show pericenter distances and apocenter distances of captured fragments in our sph simulations ( note that , in the figures , different coloured dots represent initial different pericenter distances of the passing objects ) . in the panels , the light red area corresponds to the region where the fragment s apocenter is larger than the hill radius of the planet and the blue area is where their pericenter is smaller than the planet s radius . in figures [ capture_efficiency_saturn ] and [ capture_efficiency_uranus ] , the mass of the captured fragments outside the red region is summed up and shown .
+ now , we discuss how much mass can be incorporated into the rings .
first , we calculate equivalent circular orbital radius of the captured fragments using their center of mass position and velocities as @xmath94 where @xmath95 and @xmath96 are semi - major axis and eccentricity obtained from sph simulations .
then , any fragments that satisfy @xmath97 as well as @xmath98 where @xmath99 is the radius of the planet , @xmath100 is the pericenter distance , @xmath101 is the apocenter distance and @xmath18 is the hill radius of the planet , are counted as ring mass ; this includes all those in the white region in figures [ pericenter_apocenter_saturn ] and [ pericenter_apocenter_uranus ] .
+ figures [ ring_mass_saturn ] and [ ring_mass_uranus ] show the mass that can be incorporated into the rings .
as we described in section 4.2 , more mass can be captured by the planet with decreasing pericenter distance of the passing body . on the other hand , when the fragment s pericenter distance decreases , the fraction of the fragments that collide with the planet increases ( see figures [ pericenter_apocenter_saturn ] and [ pericenter_apocenter_uranus ] ) .
therefore a smaller ring mass is obtained in some cases as the pericenter distance of the passing object becomes smaller ( figures [ ring_mass_saturn ] and [ ring_mass_uranus ] ) . in both saturn and uranus cases ,
the ring mass is about 0.1 - 10% of the initial mass of the passing object .
however , since the density of uranus is higher than that of saturn , an object can penetrate deeper inside uranus roche limit than saturn s and experience more significant tidal destruction .
therefore , in the case of uranus , the resulting silicate fraction in the ring mass can be much higher than in the case of saturn .
this argument holds for neptune as well , as it is also denser than saturn .
this could explain the observation that uranus and neptune rings are darker than saturn s rings as they are likely to have a higher rock fraction @xcite .
the ratio of @xmath102 to the initial pericenter distance of captured fragments @xmath10 is @xmath103 .
since fragments eccentricities are mostly larger than @xmath91 after the tidal destruction , the ratio is about @xmath104 .
therefore , from the results shown in figures [ pericenter_apocenter_saturn ] and [ pericenter_apocenter_uranus ] , the radial locations of circular rings are expected to be between @xmath105 m ( @xmath106 ) in the case of saturn and @xmath107 m ( @xmath108 ) in the case of uranus , where @xmath109 is the planet s roche limit .
in the previous sections , by using sph simulations , we have shown that during a close encounter of a passing object with saturn or uranus , some fragments are captured .
however , they are still in the form of large clumps ( @xmath110 kg in the case of @xmath2 kg and @xmath111 kg in the case of @xmath41 kg ) and are on highly eccentric orbits ( @xmath112 ) .
therefore , it is necessary to investigate their longer - term evolution and see if they can settle into a collection of small particles on nearly circular orbits as seen in current ring systems . in this section ,
we examine the long - term evolutions of captured fragments by using n - body simulations as well as analytical estimates . as an example
, we use the case of sph data obtained using a passing body ( @xmath2 kg ) on a prograde spin ( @xmath36h ) , with @xmath113 m and @xmath35km / s around saturn . + using the data obtained from our sph simulations , we perform n - body simulations on the longer - term evolution of their successive orbits around the giant planets .
our n - body code is essentially the same as that used in @xcite , but we now include the effect of a planet s oblate potential up to the factor @xmath5 for saturn as given by ( e.g. * ? ? ?
* ) @xmath114 the equation of motions are @xmath115 where @xmath116 is the position vector of the @xmath23-th particle , and @xmath117 with @xmath118 + in our n - body model , particles are considered to be smooth spheres and collisions between them are solved by using a hard - sphere model .
when two particles collide , we compute the velocity change using a normal coefficient of restitution @xmath119 . any particles that hit the planet
are removed .
initial fragments are modelled as hexagonally close - packed spherical aggregates consisting of same - sized particles ( radius @xmath120 km ) with density @xmath121kg / m@xmath122 ( figure [ nbody_initial ] ) . here
we focus on the initial dynamical evolutions of particles but do not investigate the long - term flattening because of computer limitations .
the total number of particles is @xmath123 .
initial positions and velocities of the center of mass of the fragments are the same as those obtained from sph simulations with no initial spin ( figure [ nbody_initial ] ) . after running the n - body simulation we find that aggregates are progressively tidally destroyed after a few orbits ( @xmath65 few years ) because their pericenters are well inside the planet s roche limit ( see figures [ pericenter_apocenter_saturn ] and [ pericenter_apocenter_uranus ] ) , and thus form an eccentric ringlet - like structure with a large eccentricity ( @xmath124 ) . as particles have slightly different semi - major axes ( see purple dots in figure [ ea_fragments ] ) , when their separation is @xmath125 ( @xmath126 is a factor of the order of unity and @xmath127 is the particle radius ) from the closest particles , neglecting the effect of oblateness potential of the planet , their collision timescale can be estimated using their synodic period as @xmath128 where @xmath86 is the semi - major axis and @xmath129 is the orbital frequency .
using @xmath130 m and @xmath131 km which are similar to the values in the n - body simulation , we get @xmath132 years . in the next subsection
, we will compare this collision timescale with the precession timescale
. + the precession rate of the argument of pericenter @xmath133 and the longitude of ascending node @xmath129 due to the @xmath134 term can be described as @xcite @xmath135 @xmath136 where @xmath137 is orbital mean motion and @xmath23 is the inclination from the planet s equatorial plane .
the precession timescale of the argument of pericenter and of the longitude of ascending node are @xmath138 @xmath139 for saturn
( @xmath140 m ) with @xmath141 m , @xmath142 and @xmath143 and @xmath144 degrees , for example , we have @xmath145 and @xmath146 years , and @xmath147 and @xmath148 years , respectively .
therefore , @xmath149 and thus precession dominates over the collisional evolution after initial aggregates are tidally destroyed .
+ thus , neglecting collision , we perform an n - body simulation under the effect of the @xmath134 and @xmath5 terms assuming the radius of the constituent particles of aggregates @xmath150 km .
the initial orbital plane is assumed to be inclined by @xmath151 degrees with respect to the planet s equatorial plane .
figure [ nbody_no_collision ] shows the longitudes of ascending node ( left panel ) and the arguments of pericenter ( right panel ) at different times . at smaller semi - major axes
, they are rapidly randomised .
the timescale of the precession is around hundreds years consistent with the above discussion , and eventually the system will form a torus - like structure .
+ in this section , we discuss what kind of collision could occur as well as its timescale after particles form a torus - like structure . in the particle - in - a - box approximation , the collision timescale can be written as @xmath152 where @xmath137 is the number density of particles , @xmath153 is the collision cross section and @xmath154 is the relative velocity . here , particles still have large eccentricities ( @xmath155 ) with randomised arguments of pericenter and longitudes of ascending node . therefore , orbital crossing occurs and @xmath154 can be given by @xmath156 where @xmath157 is the keplerian velocity . since eccentricity @xmath158 and inclination , for example , @xmath159 degrees , the relative velocity becomes about keplerian velocity .
we have confirmed such collisions occur by using n - body simulations .
the collision cross section is written as @xmath160 here we neglect the gravitational focusing term because , when particles are closer or within the roche limit , gravitational attraction between particles becomes negligible @xcite .
the volume number density of particles is ( appendix ) : @xmath161 where @xmath162 and @xmath163 are the probability of finding a particle at radial distance @xmath38 and an angle @xmath164 from the equatorial plane of the planet and are written as @xmath165 and @xmath166 respectively . using @xmath167 as well as equations ( [ sigma ] ) and ( [ number_density ] )
, we can get the collision timescale ( figure [ tau_coll_seb ] ) .
particles are more prone to collide when closer to the planet and collision velocity could be larger than a few km / s
. such a high collision velocity would be destructive enough and could generate smaller fragments as seen in current ring systems @xcite or might vaporise them @xcite .
however , including such physical effects is beyond the scope of our present work and we will leave them to future works .
collisional destruction as well as inelastic bouncing between particles is expected to damp their eccentricities and inclinations , and thus they would eventually form thin equatorial circular rings ( see also * ? ? ?
* ) made of small particles . + in this subsection , we discuss the potential effect of a primordial satellite on the dynamical evolution of the captured mass .
saturn s largest satellite titan is thought to have formed by accretion in the protoplanetary gas / dust disk @xcite .
thus , at the time of lhb , titan would have already been in orbit around saturn .
since the orbits of captured mass are highly eccentric ( section 5.2 ) - apocenter ( or pericenter ) distance is larger ( or smaller ) than that of titan - the captured mass could be removed either by gravitational scattering or accretion by titan . here
, we consider only titan as a primordial satellite .
if dione or rhea was also already present , the removal could be enhanced .
+ here we perform additional simulations including titan , assuming its equatorial orbit has the current eccentricity of @xmath168 .
the captured fragments are represented by test particles using the sph data obtained for @xmath169 kg with a prograde spin ( @xmath170h ) , @xmath171 m and @xmath172km / s .
initial orbital plane of the captured fragments is set to be 0 , 25 or 45 degrees with respect to the equatorial plane .
any particles that collide with either titan or saturn , or that go beyond 10 times of saturn s hill radius are removed .
+ figure [ effect_of_titan ] shows the results of n - body simulations .
only when the initial orbital plane of captured fragments is the same as that of titan ( @xmath173 degrees ) , their orbits cross that of titan and the effect of accretion by titan is significant
. however , when orbits of captured fragments are inclined ( @xmath174 or @xmath151 degrees ) with respect to the equatorial plane , direct collision with titan rarely happens .
in contrast , collision with saturn as a result of gravitational scattering by titan is common in all the cases ( without titan , we confirm that accretion by saturn does not occur ) .
the mass that escapes from saturn s hill sphere is negligible ( few percent or less ) .
the timescale and efficiency of removal of captured mass due to the effect of titan depends on the initial inclination of captured mass such that those fragments on the more inclined orbits are harder to remove .
note that , in figure [ effect_of_titan ] , the total removed mass keeps increasing over time .
however , in the real system ( including collisional dumping ) , the collision would damp the system and further removal is expected to cease at some point .
+ in the case of 0 degrees , about 50% of the captured mass is removed after 100 years ( figure [ effect_of_titan ] ) .
the precession timescale and successive collision timescale are comparable to or larger than 100 years ( see figure [ tau_coll_seb ] ) .
thus , a large portion of captured fragments could be removed before they settle in thin rings in this case . however ,
when the initial orbital plane is inclined by 25 or 45 degrees , only about 40 or 20% is removed even after 1000 years .
the timescale of precession and successive collisional damping is shorter than or comparable to the above timescale for the mass removal ( see figure [ tau_coll_seb ] ) .
thus , a large portion of captured fragments is expected to survive even under the existence of titan , unless the fragments happened to be distributed near titan s orbit with very low inclinations .
+ now , one may wonder whether a single collision is enough to damp the eccentric orbit inside that of titan to avoid successive close encounters .
here we simply estimate the change of orbital elements after one collision around saturn .
assuming two particles with mass @xmath175 collide with each other , the orbit of colliding particles is changed while conserving energy and angular momentum as @xmath176 where @xmath177 , @xmath178 and @xmath179 , @xmath180 are semi - major axes and eccentricities of the colliding particles after and before collision .
@xmath181 is the dissipated energy during the collision with the normal coefficient of restitution @xmath182 and the reduced mass @xmath183 .
the impact velocity is about the keplerian velocity at the radial distance @xmath38 from saturn as @xmath184 because the orbit of particles crosses as a result of precession .
then , assuming the simplest case with @xmath185 , @xmath186 , @xmath187 and @xmath188 , @xmath189 , new orbital elements after collision are derived as a function of those of initial values as @xmath190 with the initial eccentricity and semi - major axis of two particles , for example , @xmath142 , and @xmath191 m , and assuming that the collision takes place at the roche limit ( @xmath192 km ) because the collision probability is higher closer to the planet ( figure [ tau_coll_seb ] ) , we get @xmath193 m and @xmath194 .
this means that the apocenter of the new orbit after the collision is well inside the orbit of titan ( @xmath195 m ) .
thus , damping due to a single collision can sufficiently shrink the initial orbit of captured fragments so that they settle inside the orbit of titan , avoiding the second encounter with titan after the collision .
of course , this is a rather simple estimate and neglects detailed collisional processes such as destruction , compaction or vaporisation .
thus , direct numerical simulations are clearly needed in future work .
the origin of rings around planets is still debated . in this work
, we investigate the possibility that a single close encounter of a large differentiated body may form rings around giant planets ( figure [ summary ] ) .
we perform two different direct simulations ( sph simulations and n - body simulations ) in order to understand the physics of the tidal disruption of a passing large differentiated primordial kbo and long - term evolution of the resultant captured fragments . at the time of lhb
, we assume that giant planets could suffer a significant number of encounters with primordial kbos ( see also * ? ? ?
* ) . in the previous works ,
such a process was investigated using a simplified analytical formula for the capture efficiency of mass around a planet @xcite . however , the physical model in these works only considered ballistic orbits of the constituent particles and neglected the self - gravity and spin state of the passing body .
recently , @xcite have argued that the primordial kuiper belt should have had @xmath196 to @xmath197 pluto - sized bodies ( @xmath64 kg ) in order to explain the observed abundance of objects on non - resonant orbits in the present kuiper belt .
computation of encounter probabilities ( see section 2.3 ) suggests that all giant planets may have experienced a few to several tens of encounters inside their roche limit with pluto - sized objects during the lhb .
thus , this big reservoir of impactors might have delivered significant mass to the giant planet system through tidal disruption . using semi - analytical arguments and sph simulations , we first investigate the detailed physical process of tidal disruption of a passing large differentiated object .
our semi - analytical model shows significant effects of the self - gravity on the captured mass and we find that dones ( 1991 ) formula for the capture efficiency is not always consistent with our simulation results .
when the object is large enough ( @xmath41 kg ) , the capture efficiency tends to be smaller than dones formula regardless of the spin state of the object .
in contrast , when the object is smaller ( @xmath2 kg ) , the capture efficiency can be larger and smaller than dones formula depending on the direction of the spin of the object . such deviation from dones formula becomes larger as the pericenter distance becomes larger .
we also investigated the effect of an initial spin of the passing body and found that destruction becomes more significant and capture efficiency becomes larger when the body has an initial prograde spin with respect to the direction of the encounter , while the capture efficiency becomes smaller when the body has an initial retrograde spin .
sph simulations show that , during a close encounter , tidal forces spin up the body in the prograde direction .
therefore , a body with an initial prograde spin is more easily and efficiently deformed ( smoothly elongated ) and destroyed than a body with an initial retrograde spin . if the body s pericenter is well inside the planet s roche limit and above the planet s radius , then the body is disrupted and about 0.1 - 10% of its mass is captured and expected to end up in a ring - structure . assuming an average capture probability of about 1% and encounter rates reported in section 2.3 , the total mass delivered to giant planets through the tidal break - up of objects with the mass of @xmath1 kg may range from 8 , 4 , 2 and @xmath198 kg ( for jupiter , saturn , uranus and neptune , respectively ) up to 32 , 16 , 8 and 8 @xmath199 kg .
close encounters of small bodies with planets is a natural consequence of the lhb .
as was also discussed in @xcite , tidal disruption of passing objects could form massive rings around all giant planets .
in addition , recent works have shown that the inner regular satellites of saturn , neptune and uranus could form by spreading of ancient massive rings @xcite , and narrow rings such as saturn s f ring and uranian @xmath6 ring with their shepherding satellites could be the natural consequence at the last stage of such ring spreading across the roche limit @xcite . during a close encounter of a @xmath169 kg passing object with saturn ,
our numerical simulations show that only enough mass to explain saturn s current ring ( @xmath200 kg ) can be embedded . on the other hand , when the object has the mass @xmath201 kg , enough mass not only for the current rings but also for the total mass of its regular satellites ( up to and including rhea ) can be embedded .
in addition , we find a small fraction of silicate material in the captured mass because the object s mantle is preferentially disrupted . in the case of uranus , the captured mass could explain only the mass of the current rings ( mass of @xmath202 kg @xcite ) , assuming an encounter with a body with mass @xmath169 kg .
in contrast , ring mass as well as the mass of all satellites up to oberon ( @xmath203 kg ) can be explained by @xmath201 kg body .
furthermore , in the case of uranus , due to its higher density , the width between the planet s surface and its roche limit is larger than in the case of saturn .
thus , a body can pass deeper potential field of the planet . as a result
, the tidal destruction could be significant enough to disrupt not only the body s icy mantle but also its silicate core , and thus silicate components can be more efficiently captured than in the case of saturn .
this would also be applicable to neptune since it is also denser than saturn .
therefore , this could explain the fact that the rings of uranus and that of neptune are darker than that of saturn @xcite + soon after the capture , the fragments are in the form of big chunks ( @xmath110 kg in the case of @xmath169 kg and @xmath111 kg in the case of @xmath201 kg ) and they have very large eccentricities ( @xmath112 ) .
therefore , their long - term evolution needs to be investigated to see if the system becomes circular equatorial rings as currently seen in the ring systems around planets .
thus , using the data obtained from sph simulations , we perform n - body simulations of the captured fragments including the oblate potential of the planets . since their pericenter distances are deep inside the planet s roche limit , after several keplerian orbits ( @xmath65 year ) , the fragments are further tidally destroyed and form a ringlet - like structure which still has a large eccentricity on the same plane . however , they would still be large particles ( @xmath65 km to 100 km ) assuming their physical tensile strength for pure ice at low temperatures @xcite .
+ due to orbital precession , these individual particles form a torus - like structure ( @xmath65 hundreds - to - thousand years ) . in this configuration ,
orbital crossing between particles is enhanced and particles could experience high velocity collisions ( @xmath204 few km / s ) . tidal disruption itself can only form km sized fragments . however , such highly energetic collisions would be catastrophic enough to grind them down to centimetre - to - meter sized particles that are currently seen in saturn s main rings or some vaporisation of particles might occur . eventually , destruction or inelastic collisions leads to energy dissipation of random velocities and the cloud of debris would form a thin equatorial ring inside the planet s roche limit . note that direct simulations are still necessary , including such destruction and vaporisation , in order to understand more details about the final state of the ring systems . however , such complicated physics is beyond the scope of this paper and we will leave it to future works . + the above scenario seems to be very consistent with the satellite formation picture depicted in @xcite , where an initial massive ring system spreads and gives birth to satellites at the roche limit .
the silicate content of saturn s satellites may come from big chunks of silicate initially implanted at random in the primordial ring system .
the tidal disruption scenario may naturally form such big chunks of silicate .
these chunks may also constitute the `` silicate shards '' of material thought to be embedded in the core of propeller structures observed in rings by cassini @xcite .
+ in conclusion , the tidal disruption of a passing large differentiated kbo could form massive rings around giant planets and could explain the compositional differences between the rings of saturn and those of uranus ( and possibly those of neptune too ) . however , the major caveat of this model is the question of angular momentum @xcite .
rings and satellite systems of the four giant planets rotate in the prograde direction
. however , encounters of kbos with giant planets happened from random directions , so that the net total angular momentum should be close to zero in the case of a very large number of encounters ( relevant for comet - sized objects ) . however , the number of encounters with pluto - sized objects is quite small .
so small number statistics may apply so that , by chance , only a couple of encounters may dominate the total angular momentum budget for a given planet . therefore , having a system rotating in the prograde direction may mostly be a matter of chance .
whereas this explanation may hold for one planet , it is hard to believe that the same encounter s history happened for all giant planets .
we think this problem might be solved by considering the effect of tides on retrograde objects .
all fragments on retrograde orbits will experience a stong negative tidal torque from the giant planet leading to their orbital decay and crash into the planet .
the elimination timescale depends strongly on the size of the fragments , but such information is not accurately provided by the current sph simulations .
however , if the tidal decay time is smaller than the collisional flattening time , this would imply that all retrograde material could be removed before the disk is formed in the equatorial plane .
this process is far beyond the scope of our paper and thus we leave this matter to future work .
one of the most striking features of giant planet ring systems is their diversity . whereas saturn s rings are massive , those of jupiter , uranus and neptune are much less massive and show different structures .
our model suggests that all giant planets could have initially had massive rings as in @xcite .
however , they could have different subsequent evolution : as noted in @xcite , saturn s ring is the only system with a synchronous orbit inside the roche limit , whereas the other planets have opposite relative locations .
this means that any satellite forming at the edge of the roche limit of jupiter , uranus or neptune should tidally decay inside the ring system and may push the rings inwards due to resonant interactions , leading to a faster removal of the ring system .
this might explain the diversity of today s ring system even though they have all started from massive rings .
however , the satellite might be tidally destroyed once it enters deep inside the roche limit and additional ring material could be supplied .
this process is also far beyond the scope of our paper and thus we leave this matter to future work .
we thank d. nesvorn@xmath205 for useful discussions .
is grateful to shigeru ida and masahiko arakawa for discussion .
this work was supported by jsps grants - in - aid for jsps fellows ( 15j02110 ) and by scientific research b ( 22340125 and 15h03716 ) .
part of the numerical simulations were performed using the grape system at the center for computational astrophysics of the national astronomical observatory of japan .
we acknowledge the financial support of the univearths labex programme at sorbonne paris cit ( anr-10-labx-0023 and anr-11-idex-0005 - 02 ) .
this work was also supported by universit paris diderot and by a campus spatial grant .
sbastien charnoz thanks the iuf ( institut universitaire de france ) for financial support .
arakawa , m. , 1999 .
collisional disruption of ice by high velocity impact .
icarus 142 , 34 - 45 .
burns j.a . ,
hamilton d.p .
, showalter m.r . , 2001 .
dusty rings and circumplanetary dust : observations and simple physics . in : grm ,
e. , gustafson , b.a.s .
, dermott , s.f .
, fechtig , h. ( eds . ) , interplanetary dust .
springr - verlag , berlin pp .
641 - 725 .
canup , r.m , & ward , w.r . , 2009 .
origin of europa and the galilean satellites . in europa ,
r. t. pappalardo , et al .
( tucson , az : univ .
arizona press ) , 59 canup , r.m . , 2010 .
origin of saturn s rings and inner moons by mass removal from a lost titan - sized satellite .
nature 468 , 943 - 946 .
charnoz , s. et al . , 2009 . did saturn s rings form during the late heavy bombardment ?
icarus 199 , 413 - 428 .
charnoz et al . , 2009b .
origin and evolution of saturn s ring system . in saturn from cassini - huygens , ed .
m. k. dougherty , l. w. esposito & s. m. krimigis ( dordrecht : springer science & business media ) , 537 charnoz , s. , salmon , j. , crida , a. , 2010 . the recent formation of saturn s moonlets from viscous spreading of the win rings .
nature 465 , 752 - 754 .
charnoz , s. et al . , 2011 .
accretion of saturn s mid - sized moons during the viscous spreading of young massive rings : solving the paradox of silicate - poor rings versus silicate - rich moons .
icarus 216 , 535 - 550 .
crida , a. , charnoz , s. , 2012 .
formation of regular satellites from ancient massive rings in the solar system .
science 338 , 1196 - 1199 .
colwell j.e . , 1994 .
the disruption of planetary satellites and the creation of planetary rings .
space sci .
42 , 1139 - 1149 .
cuzzi , j.n . ; estrada , p.r . , 1998 . compositional evolution of saturn s rings due to meteoroid bombardment .
icarus 132 , 1 - 35 .
dones l. , 1991 . a recent cometary origin for saturn s rings ?
icarus 92 , 194 - 203 .
esposito l.w .
, ocallaghan m. , & west r.a . , 1983 . the structure of saturn s rings : implications from the voyager stellar occultation .
icarus 56 , 439 - 452 .
esposito , l.w .
planetary rings : ever decreasing circles .
nature 354 , 107 .
esposito l.w . , 1993 .
understanding planetary rings .
earth planet .
21 , 487 - 523 .
estrada , p. r. , et al . , 2009 .
formation of jupiter and conditions for accretion of the galilean satellites . in europa ,
r. t. pappalardo et al .
( tucson , az : univ .
arizona press ) , 27 french , r. g. , p. d. nicholson , c. c. porco , and e. a. marouf 1991 .
dynamics and structure of the uranian rings . in uranus
( j. t. bergstralh , e. d. miner , and m. s. matthews , eds . ) , pp .
327409 . univ . of arizona press , tucson .
genda , h. , fujita , t. , kobayashi , h. , et al .
resolution dependence of disruptive collisions between planetesimals in the gravity regime .
icarus 262 , 58 - 66 .
genda , h. , kobayashi , h. , & kokubo , e. 2015 .
warm debris disks produced by giant impacts during terrestrial planet formation . astrophys .
j. 810 , 2 .
gomes r. , levison h.f .
, tsiganis k. , morbidelli a. , 2005 .
origin of the cataclysmic late heavy bombardment period of the terrestrial planets .
nature 435 , 466 - 469 .
hadjifotinou , k.g . , 1999 .
numerical integration of satellite orbits around an oblate planet .
a&a 354 , 328 - 333 .
harris a. , 1984 .
the origin and evolution of planetary rings . in planetary rings , r. greenberg , a. brahic , eds . , ( univ .
arizona press , 1984 ) pp .
hartmann , w.k .
planet formation : mechanism of early growth .
icarus 33 , 50 - 61 .
hedman m.m .
, burns j.a .
, tiscareno m.s . ,
porco c.c .
, jones g.h .
, roussos e. , krupp n. , paranicas c. , kempf s. , 2007 .
the source of saturn s g ring .
science 317 , 653 - 656 .
hyodo r. , & ohtsuki , k. , 2014 .
collisional disruption of gravitational aggregates in the tidal environment .
j. 787 , 56 .
hyodo r. , ohtsuki , k. & takeda , t. 2015 .
formation of multiple - satellite systems from low - mass circumplanetary particle disks .
j. 799 , 40 .
hyodo r. , & ohtsuki , k. 2015 .
saturn s f ring and shepherd satellites a natural outcome of satellite system formation . nature geo .
8 , 686 - 689 .
kaula , w.m . ,
1966 . theory of satellite geodesy .
blaisdell , waltham , ma .
kraus , r. g. , senft , l. e. , & stewart , s. t. , 2011 .
impacts onto h@xmath206o ice : scaling laws for melting , vaporization , excavation , and final crater size .
icarus , 212(2 ) , 724 - 738 .
levison , h.f .
et al . , 2008 .
origin of the structure of the kuiper belt during a dynamical instability in the orbits of uranus and neptune .
icarus 196 , 258 - 273 .
melosh , h.j . , 1989 .
impact cratering : a geologic process .
oxford university press , new york , p. 253
( oxford monographs on geology and geophysics , no .
monaghan , j.j .
, 1992 . smoothed particle hydrodynamics .
30 ( a93 - 25826 09 - 90 ) 30 , 543 - 574 .
morbidelli a. , levison h.h .
, tsiganis k. , gomes r. , 2005 .
chaotic capture of jupiter s trojan asteroids in the early solar system .
nature 435 , 462 - 465 .
morbidelli , a. , tsiganis , k. , batygin , k. , crida , a. , gomes , r. , 2012 . explaining why the uranian satellites have equatorial prograde orbits despite the large planetary obliquity .
icarus 219 ( june ) , 737 - 740 .
nesvorn@xmath205 , d. , vokrouhlicky , d. , morbidelli , a. , 2007 .
capture of irregular satellites during planetary encounters .
aj 133 , 1962 - 1976 .
nesvorn@xmath205 & vokrouhlicky 2016 .
neptune s orbital migration was grainy , not smooth .
nicholson p.d . ,
french r.g . ,
campbell d.b . , margot j - l , nolan m.c . , black g.j .
, salo h.j . ,
radar imaging of saturn s rings .
icarus 177 , 32 - 62 .
ohtsuki , k. , 1993 .
capture probability of colliding planetesimals : dynamical constraints on accretion of planets , satellites , and ring particles .
icarus 106 , 228 - 246 .
porco c. , and 21 co - authors 2006 .
cassini observes the active south pole of enceladus .
science 311 , 1393 - 1401 .
porco , c.c . ,
thomas , p.c . ,
weiss , j.w . ,
richardson , d.c . , 2007 .
saturn s small inner satellites : clues to their origins .
science 318 , 1602 .
pollack j.b . , 1975 .
the rings of saturn .
space sci .
18 , 3 - 93 .
pollack j.b .
, summers a. , baldwin b. , 1973 .
estimates of the sizes of the particles in the rings of saturn and their cosmogonic implications .
icarus 20 , 263 - 278 . poulet f. , cruikshank d.p .
, cuzzi j.n . , roush t.l . ,
french r.g .
composition of saturn s rings a , b , and c from high resolution near - infrared spectroscopic observations .
412 , 305 .
stewart , s.t .
, leinhardt , z.m . ,
2012 . collisions between gravity - dominated bodies .
ii . the diversity of impact outcomes during the end stage of planet formation . astrophys .
j. 751 , 32:1 - 32:17 .
thirouin , a. , k.s .
noll , j.l .
ortiz , and n. morales 2014 .
rotational properties of the binary and non - binary populations in the trans - neptunian belt .
astron . & astrophys .
569 , a3.1 - 20 .
tillotson , j.h .
metallic equations of state for hypervelocity impact .
general atomic rept .
ga-3216 , 1 - 142 .
tiscareno , m.s .
et al . , 2006 .
100-metre - diameter moonlets in saturn s a ring from observations of propeller structures .
nature 440 , 648 - 650 .
tiscareno , m. s. , hedman , m. m. , burns , j. a. , & castillo - rogez , j. , 2013 .
compositions and origins of outer planet systems : insights from the roche critical density .
apjl , 762 , l28 - 33 .
tsiganis k. , gomes r. , morbidelli a. , levison h.f . , 2005 .
origin of the orbital architecture of the giant planets of the solar system .
nature 435 , 459 - 461 .
here we estimate the probability of particle positions on their orbits around a planet .
the classical kepler s equation is @xmath207 where @xmath208 is the mean anomaly and @xmath209 is the eccentric anomaly .
the relationship between the radial distance @xmath38 from the planet and @xmath209 can be written as where @xmath212 and @xmath213 are the probabilities of finding a particle at @xmath209 and @xmath208 , respectively .
note that the factor @xmath104 comes from the symmetricity of the orbit on @xmath209 or @xmath208 to @xmath38 .
since @xmath214 , we get next , we try to estimate the probability of the vertical direction ( @xmath215 direction which is perpendicular to the equatorial plane of the planet ) .
we define an angle of @xmath164 from the equatorial plane as where @xmath133 is the argument of pericenter and @xmath129 is the longitude of ascending node . using the longitude of pericenter @xmath218 ,
we can write the probability of particle at the angle @xmath164 as where @xmath220 and @xmath221 are the probabilities of finding a particle at @xmath215 and @xmath222 , respectively , and the factor 2 comes from the symmetricity of the orbit on @xmath222 or @xmath164 on @xmath215 . since @xmath223 , we get | the origin of rings around giant planets remains elusive .
saturn s rings are massive and made of 90 - 95% of water ice with a mass of @xmath0 kg . in contrast , the much less massive rings of uranus and neptune are dark and likely to have higher rock fraction . according to the so - called `` nice model '' , at the time of the late heavy bombardment ,
giant planets could have experienced a significant number of close encounters with bodies scattered from the primordial kuiper belt .
this belt could have been massive in the past and may have contained a larger number of big objects ( @xmath1 kg ) than what is currently observed in the kuiper belt . here
we investigate , for the first time , the tidal disruption of a passing object , including the subsequent formation of planetary rings .
first , we perform sph simulations of the tidal destruction of big differentiated objects ( @xmath2 and @xmath3 kg ) that experience close encounters with saturn or uranus .
we find that about @xmath4% of the mass of the passing body is gravitationally captured around the planet .
however , these fragments are initially big chunks and have highly eccentric orbits around the planet . in order to see their long - term evolution ,
we perform n - body simulations including the planet s oblateness up to @xmath5 starting with data obtained from the sph simulations .
our n - body simulations show that the chunks are tidally destroyed during their next several orbits and become collections of smaller particles .
their individual orbits then start to precess incoherently around the planet s equator , which enhances their encounter velocities on longer - term evolution , resulting in more destructive impacts .
these collisions would damp their eccentricities resulting in a progressive collapse of the debris cloud into a thin equatorial and low - eccentricity ring .
these high energy impacts are expected to be catastrophic enough to produce small particles .
our numerical results also show that the mass of formed rings is large enough to explain current rings including inner regular satellites around saturn and uranus . in the case of uranus ,
a body can go deeper inside the planet s roche limit resulting in a more efficient capture of rocky material compared to saturn s case in which mostly ice is captured .
thus , our results can naturally explain the compositional difference between the rings of saturn , uranus and neptune . |
a del pezzo surface over a field @xmath1 is a smooth , projective , geometrically integral surface @xmath2 over @xmath1 with ample anticanonical divisor @xmath11 ; the degree of @xmath2 is defined to be the self - intersection number @xmath12 .
a del pezzo surface is minimal if and only if there is no birational morphism over its ground field to a del pezzo surface of higher degree .
every del pezzo surface of degree @xmath13 is geometrically isomorphic to @xmath14 blown up at @xmath15 points in general position , or to @xmath16 if @xmath17 .
conversely , every smooth , projective surface that is geometrically birationally equivalent to @xmath14 is birationally equivalent over the ground field to a del pezzo surface or a conic bundle ( see @xcite ) .
a surface @xmath2 over a field @xmath1 is unirational if there is a dominant rational map @xmath18 over @xmath1 .
segre and manin proved that every del pezzo surface @xmath2 of degree @xmath19 over a field @xmath1 with a @xmath1-rational point is unirational , at least if one assumes that the point is in general position in the case @xmath20 . for references , see @xcite for @xmath21 and @xmath22 , see (
* theorem 29.4 and 30.1 ) for @xmath20 and @xmath23 , as well as @xmath24 under the assumption that @xmath1 is large enough , and see ( * ? ? ?
* theorem 1.1 ) and ( * ? ? ?
* proposition 5.19 ) for @xmath21 and @xmath25 in general . on the other hand , even though del pezzo surfaces of degree @xmath3 always have a rational point , we do not know whether any minimal del pezzo surface of degree @xmath3 that is not birationally equivalent to a conic bundle is unirational over its ground field .
if @xmath1 is infinite , then unirationality of @xmath2 implies that the set @xmath0 of @xmath1-rational points on @xmath2 is zariski dense .
the following question asks whether this weaker property may hold for all del pezzo surfaces of degree @xmath3 .
[ q : alwaysdense ] if @xmath2 is a del pezzo surface of degree @xmath3 over an infinite field @xmath1 , is the set @xmath0 of @xmath1-rational points zariski dense in @xmath2 ? over number fields , a positive answer to this question
is implied by the conjecture by colliot - thlne and sansuc that the brauer manin obstruction to weak approximation is the only one for geometrically rational varieties ( * ? ? ?
* conjecture d ) , p. 319 ) .
this conjecture may in fact hold more generally for geometrically rationally connected varieties over global fields ( see @xcite for number fields ) .
the primary goal of this paper is to state conditions under which the answer to question [ q : alwaysdense ] is positive .
let @xmath1 be a field of characteristic not equal to @xmath4 or @xmath5 , and @xmath2 a del pezzo surface of degree @xmath3 over @xmath1 with a canonical divisor @xmath26 .
then the linear system @xmath27 induces an embedding of @xmath2 in the weighted projective space @xmath28 with coordinates @xmath29 .
more precisely , there are homogeneous polynomials @xmath30 $ ] of degrees @xmath31 and @xmath32 , respectively , such that @xmath2 is isomorphic to the smooth sextic in @xmath28 given by @xmath33 for some special families of del pezzo surfaces of degree @xmath3 it is known that the set of rational points is zariski dense .
examples that are minimal and have no conic bundle structure include those given by a. vrilly - alvarado .
he proves in ( * ? ? ?
* theorem 2.1 ) that if we have @xmath22 , while @xmath34 is zero and @xmath35 satisfies some technical conditions , then the set of @xmath9-rational points on the surface @xmath2 given by is zariski dense if one also assumes that tate shafarevich groups of elliptic curves over @xmath9 with @xmath36-invariant @xmath37 are finite ( cf .
example [ ex : tony ] ) .
these technical conditions are satisfied if @xmath38 for nonzero integers @xmath39 with @xmath40 not a square , or with @xmath41 and @xmath42 relatively prime and @xmath43 ( * ? ? ?
* theorem 1.1 ) .
m. ulas @xcite , as well as m. ulas and a. togb ( * ? ? ?
* theorem 2.1 ) , also give various conditions on the homogeneous polynomials @xmath44 $ ] for the set of rational points on the surface @xmath45 over @xmath9 given by to be zariski dense .
besides hypotheses that imply that @xmath2 is not smooth or not minimal , all their conditions imply that either ( i ) @xmath46 and @xmath47 is monic , or ( ii ) @xmath47 has degree at most @xmath31 , or ( iii ) @xmath46 and @xmath35 vanishes on a rational point of @xmath7 .
e. jabara generalizes ulas work on case ( iii ) in ( * ? ?
* theorems c and d ) and treats the case over @xmath9 with @xmath47 monic and the pair @xmath48 sufficiently general .
the techniques in this paper are a generalization of a geometric interpretation of ulas work on case ( iii ) ; they are independent of the work of jabara ( see remark [ genulas ] ) .
the projection @xmath49 onto the last two coordinates is a morphism on the complement @xmath50 of the line given by @xmath51 in @xmath28 . for any point @xmath52 not equal to @xmath53 , we let @xmath54 denote the family of sections of @xmath55 that meet @xmath2 at @xmath56 with multiplicity at least @xmath57 ; we will see that @xmath54 has the structure of an affine curve , the components of which have genus at most @xmath3 ( see paragraph containing ) .
the restriction @xmath58 corresponds to the linear system @xmath59 and has a unique base point @xmath60 .
this map @xmath61 induces an elliptic fibration @xmath62 of the blow - up @xmath63 of @xmath2 at @xmath64 .
the exceptional curve on @xmath63 above @xmath64 is a section , also denoted by @xmath64 . for any @xmath65 ,
the fiber @xmath66 is isomorphic to the intersection of @xmath2 with the plane @xmath67 given by @xmath68 ; the set @xmath69 of smooth @xmath1-points on @xmath66 naturally carries a group structure characterized by the property that three points in @xmath70 sum to the identity @xmath64 if and only if they are collinear .
our first main result is the following .
[ maincor ] let @xmath1 be an infinite field of characteristic not equal to @xmath4 or @xmath5 .
let @xmath45 be a del pezzo surface given by with @xmath30 $ ] , and @xmath62 the elliptic fibration induced by the anticanonical map @xmath71 .
let @xmath52 be a point that is not fixed by the automorphism of @xmath2 that changes the sign of @xmath72 .
let @xmath54 be the curve of those sections of the projection @xmath73 that meet @xmath2 at the point @xmath56 with multiplicity at least @xmath57 .
set @xmath74 .
suppose that the following statements hold .
* the order of @xmath56 in @xmath69 is at least @xmath5 . *
if the order of @xmath56 in @xmath69 is at least @xmath31 , then @xmath54 has infinitely many @xmath1-points . *
if the characteristic of @xmath1 equals @xmath57 , then the order of @xmath56 in @xmath69 is not @xmath57 . *
if the order of @xmath56 in @xmath69 is @xmath5 or @xmath57 , then @xmath56 does not lie on six @xmath75-curves of @xmath2 .
then the set @xmath0 of @xmath1-points on @xmath2 is zariski dense in @xmath2 .
note that all four assumptions of theorem [ maincor ] are hypotheses on the point @xmath56 .
given @xmath2 , we provide an explicit zero - dimensional scheme of which the points correspond to the @xmath75-curves of @xmath2 going through @xmath56 ( cf .
remark [ minonecurves ] ) , so the first and the last two conditions of theorem [ maincor ] are easy to check . if the set @xmath0 is indeed zariski dense in @xmath2 , then the subset of those points @xmath52 that satisfy these three conditions is also dense ; theorem [ maincor ] provides a proof of zariski density of @xmath0 as soon as @xmath76 is infinite for one of these points @xmath56 .
if the answer to question [ q : alwaysdense ] is positive , then it may be true that for _ every _ del pezzo surface @xmath2 of degree @xmath3 , there exists such a point .
theorem [ maincor ] is the first result that states sufficient conditions for the set of rational points on an arbitrary del pezzo surface of degree @xmath3 to be zariski dense .
moreover , if @xmath1 is an infinite field that is finitely generated over its ground field , then @xmath76 is infinite if and only if the curve @xmath54 has a component that is birationally equivalent to @xmath7 or a component of genus @xmath3 whose jacobian has a point of infinite order .
the fact that the order of a point on an elliptic curve over such a field @xmath1 is infinite is effectively verifiable by applying the theorem of nagell - lutz to sufficiently many multiples of the point .
this means that for such fields @xmath1 , independent of question [ q : alwaysdense ] , if @xmath0 contains a point @xmath56 satisfying the conditions of theorem [ maincor ] , then we can find such a point , thus reducing the verification of zariski density of @xmath0 to a finite computation .
note that if the order of @xmath56 in @xmath77 is @xmath5 and @xmath56 does not lie on six @xmath75-curves of @xmath2 , then the assumptions in theorem [ maincor ] are automatically satisfied without any further condition on @xmath54 .
besides verifying zariski density of rational points on explicit surfaces , theorem [ maincor ] also implies the following two results .
note that both show that our criterion is strong enough to prove zariski density of the set of rational points on a set of del pezzo surfaces of degree @xmath3 over @xmath9 that is dense in the real analytic topology on the moduli space of such surfaces .
[ denseinmodulispace ] let @xmath78 be such that the surface @xmath79 given by @xmath80 is smooth .
then for each @xmath81 , @xmath82 , and @xmath83 , there exist @xmath84 with @xmath85 and @xmath86 such that the surface @xmath87 given by with the two values @xmath88 and @xmath89 replaced by @xmath90 and @xmath91 , respectively , is smooth and the set @xmath92 is zariski dense in @xmath93 .
[ densewhennodal ] suppose @xmath1 is an infinite field of characteristic not equal to @xmath4 or @xmath5 .
if @xmath2 is a del pezzo surface of degree @xmath3 and the associated elliptic fibration @xmath94 has a nodal fiber over a rational point in @xmath7 , then @xmath0 is zariski dense in @xmath2 .
our strategy to prove theorem [ maincor ] is to exhibit a rational map @xmath95 such that its image has a component whose strict transform on @xmath63 is a multisection of @xmath96 of infinite order ( cf .
@xcite ) . in the next section
, we will construct @xmath97 . to show that the image @xmath98 always has a horizontal component under the conditions in theorem [ maincor ] , we first choose a completion @xmath99 of the affine curve @xmath54 and show that the added points correspond naturally to limits of the sections in @xmath54 , which allows us to show that @xmath97 extends to the extra points in @xmath100 , sending them to @xmath101 or @xmath102 on the fiber of @xmath96 containing @xmath56 ( section [ s : completion ] ) .
this allows us to characterize all cases where no component of @xmath99 has a horizontal image under @xmath97 in sections [ examples ] and [ s : multisection ] . in section [ s : torsionbachange ] , we show that the base change of @xmath103 by a curve of genus at most @xmath3 has no nonzero torsion sections .
finally , we apply this to a horizontal component of @xmath104 to prove all our main results in section [ proofs ] . for any explicit surface @xmath2 with a point @xmath56 , it is easy to check whether @xmath98 has a horizontal component , and if so , whether that component is a multisection of infinite order . since this is indeed the case for some specific examples
, we may already conclude that it is true for @xmath2 and @xmath56 sufficiently general .
while we consider only surfaces given by that are smooth , i.e. , del pezzo surfaces of degree @xmath3 , one could also consider _ generalized del pezzo surfaces _ of degree @xmath3 , which have a birational model given by that may have isolated rational double points .
as for del pezzo surfaces , there is a natural elliptic fibration on the blow - up of a generalized del pezzo surface at the point corresponding to @xmath64 .
all our results up to and including section [ s : multisection ] also hold for generalized del pezzo surfaces of degree @xmath3 , as long as we assume that the point @xmath56 does not lie on a reducible fiber , with the exception of proposition [ outofhell5q ] .
the proof of proposition [ outofhell5q ] shows that there is one more singular surface that we should add to the list of examples where @xmath104 is not horizontal .
one can actually generalize many of our results to the case that @xmath56 lies on a reducible fiber , but given the significant amount of additional computations required , this is not included in this paper . in the proof of theorem [ densewhennodal ] , we will view the family of the curves @xmath99 as @xmath56 runs through the points on the nodal fiber as an elliptic surface . we may also consider the family of
_ all _ curves @xmath99 as an elliptic threefold over @xmath2 , possibly adding some extra component in some fibers to achieve flatness .
this threefold has real points for any surface @xmath2 over @xmath8 ; it would be interesting to study the hasse principle and weak approximation for this elliptic threefold .
all computations were done using magma @xcite . we thank peter bruin , jean - louis colliot - thlne , bas edixhoven , marc hindry , christian liedtke , bjorn poonen , alexei skorobogatov , damiano testa , anthony vrilly - alvarado , and bianca viray for useful discussions .
we thank the anonymous referee for their useful comments and suggestions .
the first author thanks universiteit leiden and the max - planck - institut in bonn and the second author the centre interfacultaire bernoulli in lausanne for their hospitality and support .
by a variety over a field we mean a separated scheme of finite type over that field .
in particular , we do not assume that varieties are irreducible or reduced . by a component
we always mean an irreducible component .
curves are varieties whose components all have dimension @xmath3 and surfaces are varieties whose components all have dimension @xmath4 .
let @xmath1 be a field of characteristic not equal to @xmath4 or @xmath5 and let @xmath105 denote the weighted projective space @xmath28 over @xmath1 with coordinates @xmath29 .
let @xmath7 be the projective line over @xmath1 with coordinates @xmath106 .
the subset @xmath107 given by @xmath51 contains the two singular points @xmath108 and @xmath109 of @xmath105 , so the complement @xmath110 is nonsingular .
the projection @xmath111 onto the last two cooordinates is well defined on @xmath50 . for each field extension
@xmath112 of @xmath1 , let @xmath113 denote the family of all curves @xmath114 in @xmath115 , defined over @xmath112 , for which the restriction @xmath116 is an isomorphism , that is , @xmath113 is the family of sections of @xmath117 .
whenever convenient , we will freely switch between viewing the elements of @xmath113 as curves and viewing them as morphisms @xmath118 .
the following lemma shows that there is an algebraic variety whose @xmath112-points are naturally in bijection with the curves in @xmath113 .
[ cqa5 ] for every field extension @xmath112 of @xmath1 , there is a bijection @xmath119 sending the point @xmath120 to the curve defined by @xmath121 without loss of generality , we assume @xmath122 . clearly , the described map is well defined and injective . to show surjectivity ,
let @xmath123 be a section of @xmath124 .
if we set @xmath125 , then there are polynomials @xmath126 $ ] such that @xmath97 is given on @xmath127 by @xmath128.\ ] ] the fact that the image @xmath114 of @xmath97 is contained in @xmath50 implies that @xmath129 and @xmath130 are constant and the degrees of @xmath131 and @xmath132 bounded by @xmath4 and @xmath5 respectively .
this shows that indeed there are @xmath133 such that @xmath114 is given by
. let @xmath30 $ ] be homogeneous of degree @xmath31 and @xmath32 , respectively , and let @xmath134 be the surface given by .
the number of @xmath75-curves on @xmath2 is finite . over a separable closure @xmath135 of @xmath1
there are @xmath136 such curves ; those that are defined over @xmath1 are characterized by the following lemma .
[ minusone ] the curves in @xmath137 that are contained in @xmath2 are exactly the @xmath75-curves of @xmath2 that are defined over @xmath1 . the @xmath75-curves are defined over a separable extension of @xmath1 by ( * ? ? ?
* theorem 1 ) .
this shows that the assumption that @xmath1 be perfect is not necessary in ( * ? ? ?
1.2 ) , which therefore implies that the @xmath75-curves on @xmath138 are exactly the curves given by for some @xmath139 , which also follows from ( * ? ? ?
* lemma 10.9 ) .
the lemma follows from taking galois invariants .
[ csdeg6 ] for each curve @xmath140 that is not contained in @xmath2 , we have @xmath141 .
the equations show that @xmath114 has degree @xmath32 .
also , @xmath114 is contained in @xmath50 , so the intersection @xmath142 with the hypersurface @xmath2 of degree @xmath32 is contained in @xmath50 , which is smooth .
therefore , intersection multiplicities are defined as usual , and the weighted analogue of bzout s theorem gives @xmath143 , where @xmath144 is the product of the weights of @xmath105 .
the statement follows .
the intersection @xmath145 consists of the single point @xmath53 . for any point @xmath146 , and for @xmath147 , we let @xmath148 denote the subvariety of all points whose associated curve , through the bijection of lemma [ cqa5 ] , intersects @xmath2 at @xmath56 with multiplicity at least @xmath149 . note that for @xmath150 this coincides with the definition of @xmath54 in the introduction .
let @xmath151 .
after applying an automorphism of @xmath7 ( and the corresponding automorphism of @xmath105 ) , we assume without loss of generality that @xmath152 , say @xmath153 for some @xmath154 .
the variety @xmath155 consists of the points of @xmath156 whose first two coordinates equal @xmath157 and @xmath158 , respectively , so the projection onto the last five coordinates gives an isomorphism @xmath159 . from now on we will freely use this isomorphism to identify @xmath160 and @xmath161 with coordinates @xmath162 . as in the introduction , we let @xmath63 denote the blow - up of @xmath2 at @xmath64 and @xmath103 the elliptic fibration induced by the anticanonical map @xmath58 .
we will sometimes identify the fiber @xmath66 above @xmath163 with its isomorphic image on @xmath2 , equal to the intersection of @xmath2 with the hyperplane @xmath67 given by @xmath164 and denoted by @xmath165 .
the intersection @xmath70 is given by a weierstrass equation ; in particular , all fibers are irreducible , and therefore all singular fibers have type @xmath166 or @xmath167 .
whenever we speak of vertical or horizontal curves or of fibers on @xmath2 or @xmath63 , we refer to this fibration .
we write @xmath168 so the fiber @xmath169 above @xmath170 , containing @xmath56 , is given by @xmath171
. we can give equations for @xmath172 inside @xmath173 as follows .
note that @xmath174 is a local parameter at the point @xmath175 on @xmath7 .
hence , around @xmath56 , the curve associated to @xmath176 is parametrized by @xmath177 for @xmath178 , let @xmath179 $ ] be the coefficient of @xmath180 in @xmath181 with @xmath182 and @xmath72 as in
. then we have @xmath183 and the variety @xmath184 is given by the equations @xmath185 .
we define the polynomials @xmath186 in @xmath187 $ ] . for every integer @xmath36 with @xmath188 ,
the polynomial @xmath189 is the factor of the @xmath36-th division polynomial of the fiber @xmath169 that corresponds to the _ primitive _ @xmath36-torsion . in particular , the polynomials @xmath190 , @xmath191 , and @xmath192 are the @xmath36-th division polynomials for @xmath193 , respectively . for notational convenience , we set @xmath194 for all @xmath195 , as well as @xmath196 for @xmath197 , where we set @xmath198 .
[ cq4a2 ] if @xmath199 , then the projection of @xmath173 onto its last two coordinates restricts to an isomorphism @xmath200 .
the inverse is given by @xmath201 with @xmath202 since @xmath203 is linear in @xmath41 , the projection of @xmath173 along the @xmath41-axis induces an isomorphism @xmath204 from @xmath205 to @xmath206 with coordinates @xmath207 , of which the inverse is determined by the given expression for @xmath41 .
the image @xmath208 has a defining equation that is linear in @xmath42 , as @xmath209 is linear @xmath42 and @xmath203 is independent of @xmath42 .
therefore , the projection from @xmath210 along the @xmath42-axis restricts to an isomorphism @xmath211 from @xmath212 to @xmath213 with coordinates @xmath214 , of which the inverse is determined by the given expression for @xmath42 . finally , the defining equation of the image @xmath215 is linear in @xmath216 , as @xmath217 is linear in @xmath216 and @xmath203 and @xmath209 are independent of @xmath216 .
therefore , the projection of @xmath218 along the @xmath216-axis restricts to an isomorphism @xmath219 from @xmath220 to @xmath221 with coordinates @xmath222 , of which the inverse determined by the given expression for @xmath216 .
the composition @xmath223 is the isomorphism of the lemma .
from now on we will assume @xmath199 , or equivalently @xmath224 , and we identify @xmath225 and @xmath221 with coordinates @xmath222 through the isomorphism of lemma [ cq4a2 ] .
we may eliminate the variables @xmath226 from the equation @xmath227 ; after multiplying all coefficients by @xmath228 , we find that the variety @xmath229 is defined by @xmath230 with @xmath231 as we assumed that @xmath232 are nonzero and that the characteristic of @xmath1 is not @xmath4 or @xmath5 , the vanishing of @xmath233 and @xmath234 would imply that @xmath56 has both order @xmath5 and @xmath31 in @xmath77 , which is a contradiction , so the coefficients @xmath235 and @xmath236 do not both vanish , and @xmath54 is a curve , though not necessarily reduced or irreducible
. we will identify @xmath54 with its image in @xmath221 and we view the coordinates @xmath41 , @xmath42 , and @xmath216 as functions on @xmath225 or @xmath54 , as given in lemma [ cq4a2 ] .
[ f4f5f6polys ] the functions @xmath237 , and @xmath238 are regular on @xmath239 and can therefore be identified with polynomials in @xmath240 $ ] .
[ minonecurves ] the @xmath75-curves on @xmath2 going through @xmath56 correspond to the points of the subscheme in @xmath241 given by @xmath242 .
[ genulas ] a special case of theorem [ maincor ] is theorem 2.1(2 ) of @xcite ; indeed , when @xmath46 and @xmath35 vanishes at @xmath243 , and @xmath244 , then the curve @xmath54 is isomorphic to the curve given in theorem 2.1(2 ) of @xcite .
the generalizations of this theorem given in ( * ? ? ?
* theorems c and d ) are also a special case of our theorem [ maincor ] , where one uses @xmath245 , which has order @xmath5 in its fiber in the case of theorem c. the proofs of theorems c and d in @xcite are incomplete , but they do work for surfaces @xmath2 that are sufficiently general .
more precisely , it is not shown that the rational function @xmath246 in the proof of theorem c ( and its implicit equivalent for theorem d ) is always nonconstant . in our geometric interpretation , this is equivalent to @xmath98 having a horizontal component . also , there is no proof of the claim that @xmath247 is never contained in @xmath248 $ ] in the proof of theorem c ( and its implicit equivalent for theorem d ) , which is crucial for the argument that the point @xmath249 has infinite order on the elliptic curve @xmath250 .
every curve @xmath251 corresponding to a point @xmath252 and not contained in @xmath2,intersects @xmath2 with multiplicity at least @xmath57 at @xmath56 , so by proposition [ csdeg6 ] , there is a unique sixth point of intersection , which is also defined over @xmath1 .
we define a rational map @xmath253 by sending @xmath254 to the sixth intersection point of @xmath114 with @xmath2 .
the map @xmath97 is defined over @xmath1 . by proposition [ csdeg6 ] , it is well defined at each point @xmath255 whose corresponding curve is not contained in @xmath2 , and thus there are at most @xmath136 points @xmath255 where @xmath97 is not well defined ( see lemma [ minusone ] and the sentences before ) . every horizontal component of the image of @xmath97 , or its strict transform on @xmath63 , yields a multisection of the elliptic fibration @xmath62 .
we can describe the map @xmath97 very explicitly .
the curve @xmath114 corresponding to @xmath256 is parametrized by .
when we substitute the expressions of into equation , we obtain @xmath257 , so the sixth intersection point of @xmath142 is given by with @xmath258 .
we keep the notation of the previous section . in particular , the field @xmath1 , the weighted projective space @xmath259 over @xmath1 with coordinates @xmath29 , the projective line @xmath7 over @xmath1 with coordinates @xmath106 , the surface @xmath260 , and the points @xmath261 are as before , and so are the objects that depend on them , including the elliptic fibration @xmath62 , the elements @xmath262 , the curve @xmath263 , the coordinates @xmath264 of @xmath221 , the functions @xmath265 on @xmath54 , and the map @xmath95 .
we will see in theorem [ ellfibnotors ] that when the closure of the image @xmath266 contains a horizontal component with respect to the natural elliptic fibration @xmath103 , then we can use such a component to construct a base change of @xmath96 with a section of infinite order .
unfortunately , in some cases the image @xmath98 does not contain such a component . in order to investigate
when this happens , we extend the map @xmath95 to a projective completion @xmath99 of the affine curve @xmath54 and first determine the image of the limit points in @xmath267 ( see proposition [ imageinf ] ) . for every extension @xmath112 of @xmath1 ,
the points in @xmath268 correspond to elements of @xmath113 , which are curves in @xmath269 .
so the curve @xmath54 parametrizes a family of curves in @xmath270 .
the elements of @xmath271 correspond to the limit curves of this family .
viewing the elements of @xmath113 as sections @xmath272 of @xmath273 , we define the morphism @xmath274 by @xmath275 , where @xmath276 is the section of @xmath273 corresponding to @xmath277 . the morphism @xmath278 is defined over @xmath1 . in terms of the coordinates
@xmath222 on @xmath279 , the map @xmath278 sends @xmath280 to @xmath281 with @xmath182 and @xmath72 as in and @xmath226 as in lemma [ cq4a2 ] . for each point
@xmath282 with corresponding section @xmath276 , the image @xmath283 is the image under @xmath278 of the fiber of the trivial @xmath7-bundle @xmath284 over @xmath254 .
therefore , we may find an appropriate completion @xmath99 , as well as the limit curves corresponding to the elements in @xmath285 as follows .
start with an arbitrary completion @xmath286 of @xmath54 and the trivial @xmath7-bundle @xmath287 over it .
now @xmath278 is defined on an open subset of @xmath288 .
after an appropriate sequence of blow - ups and blow - downs , we obtain a surface @xmath289 that is birational to @xmath288 to which @xmath278 extends as a morphism , as well as a new completion @xmath99 such that the @xmath7-bundle structure @xmath290 extends to a conic bundle structure @xmath291 .
note that it is not necessary to require that @xmath99 be smooth .
the limit curves are then the images under @xmath278 of the fibers of @xmath291 over the points in @xmath292 . the problem with the process above , in which we construct @xmath99 and @xmath289 , is that we are not working with a single del pezzo surface of degree @xmath3 , but with all of them , and we have to distinguish several cases of monoidal transformations , based on the types of singularities at the points in @xmath293 .
therefore , instead of presenting this process here , we will immediately introduce the result : a completion @xmath99 together with a conic bundle @xmath291 that works in all cases , in the sense that @xmath278 extends to it .
let @xmath294 be the coordinates of the weighted projective space @xmath295 , and let @xmath296 be the blow - up at the singular point @xmath297 . since @xmath295 is isomorphic to a cone in @xmath298 , the surface @xmath299 is smooth
; it is in fact a hirzebruch surface . by sending @xmath222 to @xmath300 , we identify @xmath221 with an open subset of @xmath295 and hence with an open subset of @xmath299 . in doing so
, we also identify the function field @xmath301 of @xmath221 with that of @xmath299 .
let @xmath99 denote the completion of @xmath54 inside @xmath299 .
note that the completion of @xmath54 inside @xmath295 contains the singular point @xmath297 if and only if the coefficient @xmath235 of @xmath302 in vanishes , i.e. , if and only if @xmath56 has order @xmath5 in @xmath77 .
hence , if @xmath56 does not have order @xmath5 , we may identify @xmath99 with the completion of @xmath54 inside @xmath295 ; as @xmath235 , @xmath236 , and @xmath303 do not all vanish , this completion is given by @xmath304 we identify @xmath299 with the variety in @xmath305 given by @xmath306 . denoting the zeroset in @xmath299 of a doubly homogeneous polynomial @xmath307 in @xmath308[s , t]$ ] by @xmath309
, we define the open subsets @xmath310 of @xmath299 . in the function field of @xmath299 , we have @xmath311 and @xmath312 .
we define the functions @xmath313 in the function field @xmath301 of @xmath299 .
[ lem : affcoords ] for each @xmath314 , the functions @xmath315 are regular on @xmath316 and the map @xmath317 sending @xmath318 to @xmath319 is an isomorphism .
the sets @xmath320 , @xmath321 , @xmath322 , and @xmath323 cover @xmath299 .
suppose @xmath314 .
the fact that @xmath324 and @xmath325 are regular on @xmath316 and define an isomorphism to @xmath221 is a standard computation . so is the last statement . using lemma [ cq4a2 ]
, one can express @xmath326 as polynomials in @xmath324 and @xmath325 , which shows that @xmath326 are regular as well . for each @xmath314 ,
set @xmath327 .
the union of the four affine curves @xmath328 , with @xmath329 , is @xmath99 . note that @xmath330 and @xmath331 .
the affine curve @xmath332 coincides with the affine part with @xmath333 of the curve in @xmath295 given by ; the affine coordinates @xmath334 correspond with @xmath335 . by abuse of notation
, we will denote the restrictions of @xmath324 , @xmath325 , @xmath336 , @xmath337 , and @xmath338 to @xmath328 by the same symbol .
we define the conic bundles @xmath339 over @xmath320 , @xmath321 , @xmath322 , and @xmath323 , respectively .
we glue these conic bundles to a conic bundle @xmath340 over @xmath299 as follows .
we glue @xmath341 and @xmath342 above the intersection @xmath343 by identifying @xmath344 with @xmath345 .
we also glue @xmath341 and @xmath346 above the intersection @xmath347 by identifying @xmath344 with @xmath348 .
we glue @xmath346 and @xmath349 above @xmath350 by identifying @xmath351 with @xmath352 .
one easily checks that these identifications also induce an isomorphism between the parts of @xmath353 and @xmath354 above the intersection @xmath355 for the remaining pairs @xmath356 . the map @xmath357 extends to @xmath358 by setting @xmath359 where , for any field extension @xmath112 of @xmath1 , we have @xmath360 , and the section @xmath276 of @xmath273 corresponds to @xmath361 .
the extended map , also denoted by @xmath278 , sends @xmath362 to @xmath281 with @xmath182 and @xmath72 as in , with @xmath363 , and with @xmath226 as in lemma [ cq4a2 ] .
the following proposition shows that @xmath278 extends to a morphism @xmath364 .
[ propgammaext ] the map @xmath278 extends to a morphism @xmath364 that is given on @xmath342 by sending @xmath365 to @xmath366 on @xmath346 by sending @xmath367 to @xmath368 and on @xmath349 by sending @xmath369 to @xmath370 it is easy to check that the given maps coincide with @xmath278 wherever they are well defined .
hence , it suffices to show that they are well defined on the claimed subsets .
suppose the first map is not well defined at a point @xmath371 .
by lemma [ lem : affcoords ] ,
the functions @xmath372 are all regular at @xmath254 , so the fact that the map is not well defined at @xmath365 implies that the four given polynomials that are claimed to define the map on @xmath342 vanish .
this yields @xmath373 , so @xmath374 , and thus @xmath375 , and @xmath376 all vanish at @xmath254 . from lemma [ cq4a2 ] and @xmath377 , we obtain @xmath378 , so the vanishing of @xmath376 at @xmath254 gives @xmath379 . from @xmath377 and equation , we find @xmath380 , so we also have @xmath381 .
this is a contradiction as @xmath56 can not have both order @xmath5 and @xmath31 on @xmath77 .
hence , the first map is well defined on @xmath342 .
it is clear that the second map is well defined at any point @xmath382 with @xmath383 or @xmath384 . to see that it is also well defined at points with @xmath385
, we identify @xmath105 with its image under the closed immersion to @xmath386 corresponding to @xmath387 on @xmath105 . substituting the expressions for the second map into the @xmath388 monomials of weighted degree @xmath32 in the variables @xmath182 , @xmath72 , @xmath389 , and
@xmath390 gives @xmath388 polynomials of total degree @xmath32 in @xmath391 , which after replacing @xmath392 by @xmath393 ( the conic bundle @xmath346 is given by @xmath394 ) are all divisible by @xmath395 .
the composition @xmath396 is given by these @xmath388 polynomials , each divided by @xmath395 .
the coordinate corresponding to the monomial @xmath397 is given by @xmath398 , which does not vanish at @xmath399 , so this composition , and thus the map @xmath400 , is well defined .
the third map is well defined whenever @xmath401 . on the other hand , if @xmath402 , then also @xmath403 , and one uses the composition @xmath404 to check that the map @xmath405 is well defined at points with @xmath406 , as in the previous case .
let @xmath289 be the inverse image of @xmath99 under the map @xmath407 .
we denote the restriction @xmath291 of the conic bundle @xmath407 by @xmath408 . by abuse of notation ,
we denote the restriction @xmath409 of the map @xmath410 by @xmath278 as well .
set @xmath411 .
then the limit curves described in the beginning of this section are the images under @xmath278 of the fibers of @xmath408 above the points in @xmath271 .
these images are described in lemmas [ curvesinf ] , [ l : e0node ] , [ l : e0cusp ] , and [ imagetangent ] . recall that @xmath332 can be identified with the open subset of @xmath295 given by @xmath333 .
[ curvesinf ] each point @xmath412 corresponds to @xmath413 for some @xmath414 satisfying @xmath415 ; the map @xmath278 sends @xmath416 to @xmath281 with @xmath417 and the image of the fiber @xmath418 under @xmath278 is a curve in @xmath105 of degree @xmath32 that intersects @xmath2 at @xmath56 with multiplicity at least @xmath57 . let @xmath412 .
the first part of the statement is obvious .
we have @xmath419 and @xmath420 .
from lemma [ cq4a2 ] we deduce @xmath421 hence , according to proposition [ propgammaext ] , the map @xmath278 sends @xmath422 to @xmath423 with @xmath424 and @xmath425 . from @xmath426
, it follows that the latter equals the expression given for @xmath72 in the lemma .
the curve @xmath427 in @xmath105 lies inside the hyperplane given by @xmath428 , which is isomorphic to the weighted projective space @xmath429 .
the intersection of @xmath430 with the curve @xmath431 in this hyperplane given by @xmath432 yields three intersection points for general @xmath90 and @xmath91 .
bzout s theorem tells us that the product of the weights @xmath433 times this intersection number @xmath5 equals @xmath434 , so we find @xmath435 .
since the degree of @xmath430 is @xmath32 , it is a full limit of images under @xmath278 of fibers of @xmath436 , all of which intersect @xmath2 at @xmath56 with multiplicity at least @xmath57 , so @xmath430 does this as well .
this can also be checked computationally by substituting the parametrization given in the lemma into the polynomial @xmath437 and checking that the coefficients of @xmath438 vanish for @xmath439 .
recall that @xmath440 is the image of @xmath169 on @xmath2 , which is the intersection of @xmath2 with the plane given by @xmath428 .
the following two lemmas give more information about the image under @xmath278 of the fibers of @xmath408 above points in @xmath441 in the case that @xmath440 is singular .
in particular , they show that @xmath440 is one of the limit curves in this case .
[ l : e0node ] suppose @xmath169 has a node . then @xmath442 contains the point @xmath443 with @xmath444 the map @xmath278 restricts to a birational morphism from the fiber @xmath445 to @xmath440 . if @xmath446 , then @xmath447 is the only point in @xmath442 .
if @xmath448 , then @xmath442 contains a unique second point @xmath449 with @xmath450 the image under @xmath278 of the fiber @xmath451 is not contained in @xmath2 . since @xmath169 has a node , we have @xmath452 with @xmath453 , so for @xmath454 we have @xmath455 and @xmath456 .
the curve @xmath457 is given by @xmath458 , and we have @xmath459 if @xmath448 , then @xmath460 and the polynomial @xmath461 factors as @xmath462 with @xmath463 and @xmath464 , which equal the expressions given in the proposition .
if @xmath465 , then @xmath466 and @xmath467 , so the only root of @xmath461 is @xmath468 , which equals the expression for @xmath469 given in the proposition .
it follows from lemma [ curvesinf ] that the points in @xmath442 are as claimed .
it follows from lemma [ curvesinf ] and the identities above that the restriction of @xmath278 to @xmath470 factors as the composition of the isomorphism @xmath471 and the birational morphism @xmath472 this proves the second statement . for the last statement , we assume @xmath448 , take @xmath473 and substitute the corresponding parametrization of lemma [ curvesinf ] in the equation @xmath474 which defines @xmath2 .
the obtained equation in @xmath475 and @xmath476 , multiplied by @xmath477 is @xmath478 as the left - hand side does not vanish identically , the curve @xmath479 is not contained in @xmath2 .
[ l : e0cusp ] suppose that @xmath169 has a cusp .
then @xmath99 equals @xmath480 and @xmath442 contains exactly one point , namely @xmath481 . the map @xmath278 restricts to a birational morphism from the fiber @xmath482 to @xmath440 .
since @xmath169 , or equivalently @xmath440 , has a cusp , we have @xmath483 .
the cusp @xmath484 is the only point on @xmath440 with @xmath182-coordinate @xmath37 , so from @xmath199 we find @xmath485 .
the @xmath486 are as in the proof of lemma [ l : e0node ] with @xmath487 . from @xmath488
we get @xmath489 .
the polynomial @xmath461 factors as @xmath490 with the unique root @xmath491 , which implies by lemma [ curvesinf ] that @xmath492 is the only point in @xmath442 .
one checks by a computation that it also follows from lemma [ curvesinf ] that the restriction of @xmath278 to @xmath493 factors as the composition of the isomorphism @xmath494 and the birational morphism @xmath495 that sends @xmath496 to @xmath497 .
this proves the second statement .
the points of @xmath100 that are not handled by the previous lemmas are the points in @xmath498 , that is , the points above the singular point @xmath297 in @xmath295 .
the next lemma takes care of these points .
[ imagetangent ] for each point @xmath499 , the map @xmath278 sends the fiber @xmath482 to the curve in @xmath105 given by @xmath428 and @xmath500 ; this curve intersects @xmath2 with multiplicity @xmath5 at @xmath56 and nowhere else . by lemma [ lem : affcoords ] , the open subset @xmath322 has affine coordinates @xmath501 .
if @xmath254 lies in @xmath502 , then it corresponds with a point with @xmath503 for some @xmath504 and the fiber @xmath482 is given by @xmath505 in @xmath506 .
the map @xmath278 sends @xmath507 to @xmath508 by lemma [ propgammaext ] . from lemma [ cq4a2 ] , we find @xmath509 .
it follows that the image of the fiber is the claimed curve . in the affine plane given by @xmath428 and @xmath510 ,
this curve is a line going through @xmath56 with slope @xmath511 , so it is exactly the tangent line to the curve @xmath440 at @xmath56 . note that the existence of @xmath254 implies that @xmath56 has order @xmath5 on @xmath512 , so this tangent line intersects @xmath440 with multiplicity @xmath5 at @xmath56 and nowhere else . as the curve intersects the surface @xmath2 only in the curve @xmath440 , the lemma follows .
if @xmath254 lies in @xmath513 , then the argument is analogous .
[ nonreducedinf ] scheme theoretically , the image under @xmath278 of the fiber of @xmath408 above @xmath254 in lemma [ imagetangent ] is not reduced , but given by @xmath514 and @xmath500 .
this nonreduced curve is also a limit curve as mentioned in the beginning of the section , and it intersects @xmath2 with multiplicity @xmath32 at @xmath56 . [ tremark ]
let @xmath515 be the image of @xmath516 .
then @xmath515 is the union of all curves @xmath517 corresponding to points @xmath255 and the limit curves corresponding to points @xmath518 .
the closure of the image @xmath98 in @xmath2 is contained in the intersection @xmath519 .
this intersection also contains all @xmath75-curves on @xmath2 that go through @xmath56 .
see also remarks [ whatisthiscurve ] and [ finalremarkchapter5 ] .
the rational map @xmath95 from the end of section [ multisection ] factors as @xmath520 , where @xmath521 is a rational section of @xmath522 that sends @xmath523 to @xmath524 . here , for @xmath525 , we view @xmath526 as in as a function on @xmath99 .
we use this in proposition [ imageinf ] to show that @xmath97 extends to a map that is well defined at every point in @xmath411 .
@xmath527_(0.4){\tau } \ar[r]^\gamma & { \mathbb{p}}&s \ar@{_(->}[l ] \\ { \mathcal{c}}_q(5 )
\ar@{^(->}[u ] \ar@/_5.5mm/@{-->}[rrr]_{\sigma}\ar@/^5.5mm/@{-->}[r]^{\rho } & { \mathcal{c}}_q(5 ) \times { \mathbb{p}}^1 \ar@{^(->}[u]\ar[l ] \ar[r]^(0.63)\gamma & u \ar@{^(->}[u ] & s-\{{\mathcal{o}}\ } \ar@{_(->}[l]\ar@{^(->}[u ] } \ ] ] [ imageinf ] the rational map @xmath97 extends to a rational map @xmath528 that is well defined at the points in @xmath271 . for every @xmath518
, we have @xmath529 if @xmath440 has a cusp or @xmath440 has a node and @xmath530 as in lemma [ l : e0node ] , and we have @xmath531 otherwise .
let @xmath518 .
then by lemmas [ curvesinf ] , [ l : e0node ] , [ l : e0cusp ] , [ imagetangent ] , and remark [ nonreducedinf ] , the scheme - theoretic image of @xmath482 under @xmath278 is a curve of degree @xmath32 in the plane given by @xmath428 in @xmath105 .
we denote this curve by @xmath114 .
a parametrization of @xmath114 is given in lemma [ curvesinf ] if @xmath532 ; the curve @xmath114 is nonreduced if @xmath533 .
the intersection @xmath142 is the same as the intersection of @xmath114 with with @xmath534 , and @xmath114 intersects @xmath440 with multiplicity at least @xmath57 at @xmath56 .
if @xmath440 is smooth , then @xmath440 has genus @xmath3 , so @xmath114 has no components in common with @xmath440 .
the curves @xmath440 and @xmath114 also have no components in common if @xmath533 ( lemma [ imagetangent ] and remark [ nonreducedinf ] ) or @xmath535 as in lemma [ l : e0node ] .
hence , in all these cases there is a unique sixth intersection point in @xmath536 , and we can extend @xmath97 to @xmath254 by sending @xmath254 to this sixth intersection point , say @xmath318 ; the divisor @xmath537 on @xmath440 is a hypersurface section inside the plane given by @xmath428 , so it is linearly equivalent to a multiple of @xmath538 on @xmath440 , and we find @xmath539 in @xmath512 .
we are left with the case that @xmath440 has a cusp ( lemma [ l : e0cusp ] ) , or @xmath440 has a node and @xmath530 as in lemma [ l : e0node ] . in both cases , there is a @xmath540 such that @xmath455 and @xmath541 and , in terms of the coordinates @xmath542 on @xmath295 , we have @xmath543 with @xmath544 . by lemma [ lem : affcoords ] ,
the functions @xmath545 and @xmath546 are affine coordinates for @xmath321 , with @xmath254 corresponding to @xmath547 , and the functions @xmath548 , and @xmath376 are regular on @xmath321 . as before ,
we denote the restrictions of @xmath549 , @xmath550 , @xmath548 , and @xmath376 to @xmath332 by the same symbols . using and
, we can express , for each @xmath551 , the function @xmath552 on @xmath99 as a polynomial in terms of @xmath549 , @xmath550 , @xmath548 , and @xmath376 , which shows that @xmath553 is regular on @xmath332 .
in particular , we have @xmath554 recall from subsection [ extendinggamma ] that @xmath341 and @xmath342 are glued by setting @xmath555 . hence , on @xmath556 , the rational map @xmath557 is given by @xmath558 the functions @xmath549 and @xmath559 are local parameters for @xmath321 at @xmath254 , so their restrictions generate the maximal ideal @xmath560 of the local ring @xmath561 of @xmath332 at @xmath254 . from , we find that in @xmath561 we have @xmath562 modulo @xmath563 .
now suppose @xmath448 .
then @xmath460 and the left - hand side of factors as @xmath564 with @xmath565 .
in fact , @xmath566 and @xmath567 correspond to @xmath469 and @xmath568 of proposition [ l : e0node ] .
modulo @xmath569 , the left- and right - hand side of are congruent to @xmath570 and @xmath571 , respectively .
assume @xmath572 as well .
then @xmath573 , so we find that modulo @xmath569 we have @xmath574 with @xmath575 hence , @xmath560 is generated by @xmath549 and one checks , preferably with the help of a computer , that we have @xmath576 we claim that also holds when @xmath487 or @xmath465 . indeed ,
if @xmath465 , then one uses @xmath467 , while @xmath466 and @xmath577 , so yields @xmath578 ; it follows that @xmath549 generates @xmath560 , and one checks again by computer .
if @xmath487 , then @xmath560 may not be principal , so being congruent modulo @xmath579 is potentially stronger than being congruent modulo @xmath569 ; but using that modulo @xmath579 we have and @xmath580 , one can again check that holds .
hence , holds in all cases .
now @xmath581 is nonzero because the surface @xmath2 is smooth at the singular point of @xmath440 .
also , since @xmath56 is not the singular point of @xmath440 , we have @xmath582 and @xmath234 and @xmath583 do not both vanish .
we conclude that @xmath584 is well defined at @xmath254 , sending @xmath254 to @xmath585 . substituting this into the parametrization of lemma [ curvesinf ]
, we find @xmath586 , with @xmath587 it is easy to check that this point equals @xmath101 in the group @xmath512 , using the fact that the tangent line to @xmath440 at @xmath56 intersects @xmath440 also in @xmath588 , the tangent line to @xmath440 at @xmath588 intersects @xmath440 also in @xmath589 , and the inverse of a point is obtained by negating the @xmath72-coordinate .
[ cor310 ] the following statements hold .
the multiples of @xmath56 are taken in the group @xmath590 . 1 .
we have @xmath591 if and only if @xmath440 is smooth .
2 . if @xmath592 , then @xmath440 is nodal .
the converse holds if @xmath593 .
3 . if @xmath440 is cuspidal , then @xmath594 .
the converse holds if @xmath593 .
4 . if @xmath595 and @xmath596 , then @xmath597 and @xmath598 . the ` if'-part of ( 1 ) follows immediately from proposition [ imageinf ] .
for the ` only if'-part , note that if @xmath440 is singular , then by proposition [ imageinf ] there exists a @xmath518 with @xmath599 : when @xmath440 is cuspidal , this holds for any @xmath518 and when @xmath440 is nodal , we can take @xmath530 as in lemma [ l : e0node ] . the first part of ( 3 )
follows directly from proposition [ imageinf ] .
together with ( 1 ) , this also implies the first part of ( 2 ) . if @xmath593 and @xmath440 is nodal , then for the points @xmath447 and @xmath600 as in lemma [
l : e0node ] , we have @xmath601 and @xmath602 by proposition [ imageinf ] , which proves the second part of ( 2 ) .
the second part of ( 3 ) now follows from ( 1 ) and ( 2 ) .
statement ( 4 ) follows immediately from proposition [ imageinf ] .
the next two sections investigate the conditions under which @xmath603 sends an irreducible component of @xmath99 to a fiber of @xmath58 .
the following lemma shows that if this is the fiber that contains @xmath56 , then @xmath97 is the constant map to @xmath56 on the component .
[ qabove0 ] let @xmath604 be a component of @xmath99 for which @xmath605 .
then @xmath606 .
without loss of generality , we assume @xmath1 is algebraically closed .
let @xmath607 be such that the associated section @xmath608 is not entirely contained in @xmath2
. then @xmath97 is well defined at @xmath254 and @xmath609 is the unique sixth intersection point of @xmath114 with @xmath2 .
since @xmath114 is a section of @xmath610 , it intersects the fiber @xmath440 only once , namely in @xmath56 , and as this sixth intersection point lies in @xmath440 as well , we conclude @xmath611 .
thus all but finitely many points of @xmath604 map to @xmath56 under @xmath97 , so @xmath606 .
in this section , @xmath1 still denotes a field of characteristic not equal to @xmath4 or @xmath5 . we will give examples of surfaces @xmath134 over @xmath1 given by , together with a point @xmath612 for which the map @xmath603 sends at least one irreducible component of @xmath99 to a fiber of @xmath58 .
in the next section we will see that , at least outside characteristic @xmath57 , these examples include all cases where every component of @xmath99 is sent to a fiber on @xmath2 . in view of theorem [ maincor ] , it is important to note that in all examples , there are at least six @xmath75-curves on @xmath2 going through @xmath56 .
recall from remark [ minonecurves ] that these correspond to the points of @xmath613 with @xmath242 .
recall also , from the last paragraph of section [ multisection ] , that the map @xmath614 is given by @xmath615 $ ] .
[ ex : hell5q ] let @xmath616 and assume the characteristic of @xmath1 is not @xmath57 . set @xmath617 and let @xmath134 be the surface given by with @xmath618 and @xmath619 , and with point @xmath153 .
assume that @xmath2 is smooth , so that it is a del pezzo surface .
the curve @xmath440 is nonsingular if and only if @xmath620 . the point @xmath56 has order @xmath57 in @xmath512 .
generically , in particular over a field in which @xmath621 and @xmath622 are independent transcendentals , the surface @xmath2 is smooth and the fibration @xmath62 has @xmath623 nodal fibers ( type @xmath166 ) and one cuspidal fiber ( type @xmath167 ) above @xmath624 .
let @xmath566 be an element in a field extension of @xmath1 satisfying @xmath625
. then @xmath99 splits over @xmath626 into two components .
the function @xmath238 vanishes on @xmath99 and the map @xmath95 sends each component birationally to the cuspidal fiber .
the conic bundle @xmath289 splits into two components as well
. both components of the image @xmath515 of @xmath516 ( cf .
remark [ tremark ] ) intersect @xmath2 in the cuspidal fiber and , over an extension of @xmath626 of degree at most @xmath57 , five @xmath75-curves ; the surface @xmath515 intersects @xmath2 doubly in the cuspidal curve , as well as in ten @xmath75-curves going through @xmath56 , corresponding to the points on the affine part @xmath54 where @xmath627 vanishes . indeed ,
if @xmath628 in an extension of @xmath1 satisfy @xmath629 then we have a section over @xmath630 going through @xmath56 with @xmath631 [ ex : char5 ] let @xmath1 be a field of characteristic @xmath57 containing elements @xmath632 .
let @xmath134 be the surface given by with @xmath633 and @xmath634 , and with point @xmath635 .
assume that @xmath2 is smooth , so that it is a del pezzo surface .
generically , and in particular when @xmath566 and @xmath621 are independent transcendentals , this is the case , and the fibration @xmath62 has @xmath623 nodal fibers ( type @xmath166 ) and one cuspidal fiber ( type @xmath167 ) , namely @xmath440 .
the curve @xmath54 is given by @xmath636 and @xmath627 vanishes on @xmath54 .
by lemma [ qabove0 ] , the map @xmath97 is constant and sends @xmath54 to @xmath56 .
generically , the curve @xmath54 is geometrically irreducible .
there are at least ten @xmath75-curves going through @xmath56 .
[ ex : order3 ] for any @xmath637 , the point @xmath638 has order @xmath5 on the weierstrass curve given by @xmath639 with @xmath640 and @xmath641 ; this curve is nonsingular if and only if @xmath642 .
_ subexample ( i ) .
_ for any @xmath643 we consider the surface @xmath260 given by with @xmath644 and with @xmath645 , so that @xmath56 has order @xmath5 on @xmath512 .
assume @xmath2 is smooth , so that it is a del pezzo surface .
the affine part @xmath54 of the curve @xmath99 is given by @xmath646 the function @xmath647 vanishes on the component given by the vanishing of the second factor ; by lemma [ qabove0 ] , this component is contracted by the map @xmath603 , which sends it to @xmath56 .
there are at least six @xmath75-curves on @xmath2 going through @xmath56 .
_ subexample ( ii ) .
_ for any @xmath648 we consider the surface @xmath260 given by with @xmath649 and with @xmath645 .
assume @xmath2 is smooth , so that it is a del pezzo surface .
the affine part @xmath54 of the curve @xmath99 is given by @xmath650 again , the function @xmath651 vanishes on the component given by the vanishing of the second factor ; again by lemma [ qabove0 ] , this component is contracted by the map @xmath603 , which sends it to @xmath56 .
there are at least nine @xmath75-curves on @xmath2 going through @xmath56 .
_ subexample ( iii ) .
_ let @xmath2 be any smooth surface that fits in both families of these examples , i.e. , with @xmath652 , @xmath653 , @xmath654 , and @xmath655 .
writing @xmath656 and @xmath657 , we have @xmath658 generically , say over a field in which @xmath659 , and @xmath660 are independent transcendentals , the surface @xmath2 is smooth and the fibration @xmath62 has twelve nodal fibers .
suppose @xmath2 is indeed smooth
. then @xmath661 .
the affine part @xmath54 of the curve @xmath99 is given by @xmath662 so it consists of two components . the function
@xmath627 vanishes on both components , so by lemma [ qabove0 ] , they are contracted to @xmath56 by @xmath95
. there are at least nine @xmath75-curves on @xmath2 going through @xmath56 . [ ex : order3sec ] for any @xmath663 , the point @xmath664 has order @xmath5 on the elliptic curve given by @xmath665 . in the following three subexamples , we take @xmath666 for some @xmath667 and the point @xmath668 , which in all cases has order @xmath5 on @xmath440 .
_ subexample ( i ) .
_ let @xmath2 be the surface given by with @xmath669 for some @xmath670 and assume that @xmath2 is smooth .
the affine part @xmath54 of the curve @xmath99 is given by @xmath671 .
the function @xmath672 vanishes on the component given by @xmath673 ; by lemma [ qabove0 ] , this component is contracted by the map @xmath603 , which sends it to @xmath56 .
there are at least six @xmath75-curves on @xmath2 going through @xmath56 .
generically , there are twelve nodal fibers .
_ subexample ( ii ) .
_ let @xmath2 be the surface given by with @xmath674 for some @xmath670 and assume that @xmath2 is smooth .
the affine part @xmath54 of the curve @xmath99 is given by @xmath675 .
the function @xmath676 vanishes on one of the components ; by lemma [ qabove0 ] , this component is contracted by the map @xmath603 , which sends it to @xmath56 .
there are at least nine @xmath75-curves on @xmath2 going through @xmath56 .
generically , there are twelve nodal fibers
. _ subexample ( iii ) .
_ let @xmath2 be the surface given by with @xmath677 and assume that @xmath2 is smooth .
the affine part @xmath54 of the curve @xmath99 is given by @xmath678 .
the function @xmath627 vanishes on both components , so we have @xmath679 by lemma [ qabove0 ] .
the surface is isotrivial ; all fibers have @xmath36-invariant @xmath37 .
there are at least nine @xmath75-curves on @xmath2 going through @xmath56 , and there are six cuspidal fibers .
we continue the notation of sections [ multisection ] and [ s : completion ] . in particular ,
the field @xmath1 with characteristic not equal to @xmath4 or @xmath5 , the surface @xmath2 , and the point @xmath56 are fixed as before , as are all the objects that depend on them . as we have seen in the previous section , not every component of @xmath99 necessarily has its image under @xmath603 map dominantly to @xmath7 under the projection @xmath58 . proposition [ vertcomp ] states that this does hold for every component if the order of @xmath56 is larger than @xmath32
. moreover , proposition [ vertcomp ] is sharp in the sense that there are examples where the order of @xmath56 is @xmath32 and @xmath54 has a component that maps under @xmath97 to @xmath56 .
[ vertcomp ] suppose the order of @xmath56 in @xmath512 is larger than @xmath57 and @xmath99 has a component @xmath604 that maps under @xmath603 to a fiber of @xmath273 .
then @xmath56 has order @xmath32 and @xmath606 .
the curve @xmath99 has a unique second component , which is sent under @xmath97 to a horizontal curve on @xmath2 .
since @xmath604 is projective , it contains a point in @xmath411 , say @xmath318 . by corollary [ cor310 ] , part ( 4 ) , we have @xmath680 .
suppose @xmath97 does not send @xmath604 to a horizontal curve .
then the composition @xmath681 sends @xmath604 to @xmath175 . from lemma [ qabove0 ]
we find @xmath606 and we obtain @xmath682 or @xmath683 from proposition [ imageinf ] . as the order of @xmath56 is larger than @xmath57 , we find that the order is @xmath32 and @xmath684 .
we have @xmath685 . from equations we find
@xmath686 so the factor @xmath687 equals @xmath688 as @xmath689 ( together with @xmath199 ) implies @xmath690 , we get @xmath691 , which in turn , together with @xmath460 , implies that @xmath54 is reduced .
suppose that each component of @xmath99 maps under @xmath97 to a fiber of @xmath273 .
then as above , we find @xmath692 and as the composition @xmath681 is given by @xmath693 , we find that @xmath627 vanishes on @xmath54 ; as @xmath54 is reduced , this implies that if we view @xmath694 and @xmath627 as polynomials in @xmath240 $ ] ( cf .
remark [ f4f5f6polys ] ) , then @xmath627 is a multiple of @xmath694 .
viewing @xmath694 and @xmath627 as quadratic polynomials in @xmath695 over @xmath696 $ ] , and comparing the coefficients in @xmath696 $ ] of @xmath302 in @xmath697 we find @xmath698 comparing the coefficient of @xmath699 in this equality gives @xmath700 since @xmath701 by equations , we find from that the difference of the two sides in equals @xmath702 hence , the equality is equivalent to @xmath703 , so we obtain @xmath704 , a contradiction from which we conclude that not all components map to a vertical component .
it follows that there is a second component , which is unique as @xmath460 implies that there are at most two components .
this second component maps to a horizontal curve on @xmath2 .
we say that two pairs @xmath705 and @xmath706 of a variety with a point on it are isomorphic if there is an isomorphism from @xmath707 to @xmath708 that maps @xmath709 to @xmath710 .
for example , the involution @xmath711 that sends @xmath712 to @xmath713 fixes @xmath56 , so it induces an isomorphism , also denoted @xmath714 , from the pair @xmath715 to @xmath716 ; the surface @xmath717 is given by @xmath718 , with @xmath719 note that @xmath714 fixes the points in the fiber above @xmath175 and it switches the fibers above @xmath720 and @xmath721 .
it also fixes @xmath722 and @xmath723 and it replaces @xmath724 and @xmath725 by @xmath726 and @xmath727 , respectively .
the following lemma is well known ( see ( * ? ? ?
* proposition 8.2.8 . ) for @xmath150 , @xcite for @xmath728 , and ( * ? ? ?
* table 3 ) for characteristic @xmath37 ) .
the lemma is used in propositions [ outofhell5q ] and [ nohell3q ] .
[ explicitunivcurve ] let @xmath729 be an elliptic curve over @xmath1 and @xmath730 an integer .
let @xmath731 be a point of order @xmath149 . then there exist elements @xmath732 and @xmath733 such that the pair @xmath734 is isomorphic to the pair @xmath735 , with @xmath736 given by @xmath737 after choosing an initial weierstrass model for @xmath729
, we may apply a linear change of variables to obtain a model @xmath736 in which @xmath254 corresponds to @xmath738 .
given that the order of @xmath254 is not @xmath4 , we may also assume that the tangent line to the model @xmath736 at @xmath254 is given by @xmath739 .
then there are @xmath740 with @xmath741 such that @xmath736 is given by @xmath742 .
we have @xmath743 , so @xmath744 , or , equivalently , @xmath745 , holds if and only if @xmath746 . if @xmath728 , so @xmath744 , then either @xmath747 or @xmath748 , and in the latter case , we may scale @xmath182 and @xmath72 such that we have @xmath749 .
these two cases are exactly the claimed cases , with @xmath750 and @xmath751 . if @xmath150 , then we have @xmath752 and @xmath753 , so we may scale @xmath182 and @xmath72 such that we have @xmath754 .
then we have @xmath755 , so the property @xmath756 , or , equivalently , @xmath757 , yields @xmath758 , which yields the claimed case with @xmath759 .
[ outofhell5q ] suppose that the characteristic of @xmath1 is not @xmath57 , and @xmath760 in @xmath512 .
if no component of @xmath99 maps under @xmath97 to a horizontal curve on @xmath2 , then there exist @xmath761 such that the pair @xmath715 is isomorphic to the pair of example [ ex : hell5q ] .
if @xmath729 is an elliptic curve over @xmath1 with a point @xmath254 of order @xmath57 , then by lemma [ explicitunivcurve ] , there exists a @xmath732 such that @xmath729 is isomorphic to the elliptic curve given by @xmath762 , with @xmath254 corresponding to the point @xmath738 . a short weierstrass model for this curve
is given by @xmath763 , with the point @xmath738 corresponding to @xmath764 , where @xmath765 if @xmath440 is smooth , then , as @xmath56 has order @xmath57 on @xmath440 and isomorphisms between short weierstrass models are all given by appropriate scaling of the coordinates , there are @xmath766 such that @xmath767 another way to phrase this is that gives a parametrization of the quadruples @xmath768 with @xmath769 for which the associated fifth division polynomial @xmath770[x]$ ] vanishes at @xmath157 .
hence , also in the case that @xmath440 is singular , there exist @xmath771 for which holds . from @xmath199
, we get @xmath772 . without loss of generality
, we assume @xmath773 . the fiber @xmath440 is singular if and only if @xmath774 is zero , and in this case @xmath440 is nodal .
note that because @xmath56 has order @xmath57 , we have @xmath775 and @xmath776 .
we first state two claims , both with a computational proof .
* claim 1 * : _ if @xmath777 and @xmath694 divides @xmath778 , then @xmath2 is singular . _ _ proof_. since the main coefficient @xmath235 of @xmath694 as a polynomial in @xmath695 over @xmath696 $ ] is invertible , we can compute ( by computer , with @xmath779 independent transcendentals ) the remainder of @xmath778 upon division by @xmath694 , which is a polynomial @xmath780 , with @xmath781 $ ] of degree @xmath782 and @xmath783 , respectively .
our special values of @xmath784 already imply that the coefficients of @xmath785 and @xmath786 in @xmath787 specialize to @xmath37 , and the fact that @xmath694 divides @xmath778 implies that @xmath787 specializes to @xmath37 .
we consider two cases , based on the characteristic of @xmath1 .
* case 1 * : the characteristic of @xmath1 is not @xmath788 , or @xmath789 . in this case , the vanishing of the ( specialization of the ) coefficients of @xmath790 , @xmath791 , @xmath792 , and @xmath793 in @xmath787 determine , in that order , the values of @xmath794 , and @xmath795 in terms of @xmath796 .
the vanishing of the coefficient of @xmath797 then implies that we have one of two subcases : ( a ) @xmath798 or ( b ) @xmath799 for some specific constant @xmath90 .
assume we are in subcase ( a ) , i.e. , @xmath798 .
the vanishing of the coefficient of @xmath800 yields @xmath801 ; then the vanishing of the coefficients of @xmath802 and @xmath699 implies @xmath803 , and finally the vanishing of the coefficients of @xmath804 gives @xmath805 , which shows that the pair @xmath715 is isomorphic to the pair in example [ ex : hell5q ] , with @xmath806 , though @xmath238 vanishes on both components @xmath604 and @xmath807 , and @xmath2 is singular .
assume we are in subcase ( b ) , i.e. , @xmath799 .
we may assume @xmath808 , and the vanishing of the coefficients of @xmath800 , @xmath809 , @xmath802 , and finally @xmath699 , express @xmath810 , in that order , in terms of the remaining unknown coefficients of @xmath35 , which in the end yields a surface @xmath2 that is singular .
* case 2 * : the characteristic of @xmath1 is not @xmath811 , or @xmath812 . as in case 1
, we similarly solve for the parameters @xmath813 and @xmath814 , except that we start by expressing @xmath815 in terms of @xmath813 .
we conclude also in these characteristics that @xmath2 is singular , thus proving the claim .
* claim 2 * : _ if @xmath816 and @xmath54 is reduced and @xmath694 divides @xmath817 , then either @xmath2 is singular , or there exists a @xmath818 , such that the pair @xmath715 is isomorphic to the pair of example [ ex : hell5q ] . _
_ proof_. since @xmath54 is reduced , i.e. , @xmath694 has no multiple factors , the condition that @xmath694 divides @xmath817 is equivalent to all components of @xmath54 being sent under the map @xmath681 to @xmath720 or @xmath721 . as the isomorphism @xmath714 described before lemma [ explicitunivcurve ] switches the fibers above these two points , the hypotheses of this claim hold for the pair @xmath715 if and only if they hold for the pair @xmath716 . hence , without loss of generality we may apply @xmath714 at some point .
viewing @xmath819 as polynomials in @xmath695 over @xmath696 $ ] , we find that generically , say over a field in which @xmath820 are independent transcendentals , there are @xmath821 and @xmath822 , in terms of these transcendentals , such that @xmath823 the fact that @xmath56 has order @xmath57 implies that @xmath824 specialize to @xmath37 . in our case , the other coefficients @xmath825 specialize to @xmath37 as well .
we claim that from the fact that @xmath826 and @xmath827 specialize to @xmath37 , it follows that @xmath828 for some element @xmath829 with @xmath830 and with @xmath831 indeed , for any @xmath278 in an extension of @xmath1 with @xmath832 and @xmath833 , the linear combinations @xmath834 of @xmath827 and @xmath826 factor into two linear factors .
therefore , the vanishing of @xmath827 and @xmath826 implies the vanishing of one of the first two factors and one of the second two .
the four combinations give four systems of two linear equations in the two variables @xmath724 and @xmath725 .
for each combination , the determinant of the system is a nonzero multiple of @xmath430 and therefore nonzero itself .
the systems yield exactly the four claimed pairs for @xmath835 in .
note that as the isomorphism @xmath714 replaces @xmath724 and @xmath725 by @xmath726 and @xmath727 , respectively , it switches the first two cases in , as well as the last two cases given by the third pair for @xmath836 . therefore , after applying the isomorphism @xmath714 if necessary , we may assume we have only two subcases .
* case 1 * : we have @xmath837 . the equations @xmath838 determine a system of two linear equations in @xmath839 and @xmath840 , of which the determinant is a nonzero multiple of @xmath430 and therefore nonzero itself .
the unique solution is @xmath841 .
subsequently , the system @xmath842 gives @xmath843 and then the system @xmath844 yields @xmath845 . at this point ,
the coefficients @xmath846 and @xmath847 specialize to @xmath37 automatically , and the equation @xmath848 determines @xmath849 . with @xmath850 , we obtain exactly the surface of example [ ex : hell5q ] .
* case 2 * : we have @xmath851 . as in the previous subcase , the linear systems @xmath852 determine @xmath853 and @xmath854 inductively for @xmath855 .
again , the coefficients @xmath846 and @xmath847 then specialize to @xmath37 automatically .
finally , the system @xmath856 is linear in @xmath857 and @xmath858 and determines these two parameters uniquely .
however , this yields a surface @xmath2 that is singular .
more specifically , the associated minimal elliptic surface has two singular fibers of type @xmath859 .
this proves the claim .
we continue the proof of the proposition .
suppose no component of @xmath99 maps under @xmath97 to a horizontal curve on @xmath2 , so @xmath681 has finite image .
if we had @xmath860 , then we would have @xmath861 by lemma [ qabove0 ] , so by corollary [ cor310 ] , part ( 3 ) , the fiber @xmath440 would be cuspidal . from this contradiction
we conclude that there is a component @xmath807 with @xmath862 . without loss of generality
, we assume @xmath863 and we write @xmath864 . we will distinguish the following three cases .
* there is a component @xmath604 of @xmath99 with @xmath865 .
* there is a component @xmath604 of @xmath99 with @xmath866 .
* there is no component @xmath604 of @xmath99 with @xmath867 . since @xmath460
, the curve @xmath99 has at most two components and both are reduced if there are two , so in cases ( a ) and ( b ) , the components are @xmath604 and @xmath807 , and @xmath54 is reduced . we start with case ( a ) .
assume that there is a component @xmath604 of @xmath99 with @xmath605 .
from lemma [ qabove0 ] we find @xmath868 , and as @xmath604 contains points of @xmath271 , we conclude that @xmath440 is singular from corollary [ cor310 ] , part ( 1 ) , so @xmath869 . if we consider @xmath237 , and @xmath238 as polynomials in @xmath695 over @xmath696 $ ] ( cf .
remark [ f4f5f6polys ] ) , then @xmath627 and @xmath238 vanish on @xmath604 and @xmath807 , respectively , so @xmath694 divides @xmath778 .
claim 1 implies that @xmath2 is singular , a contradiction .
we continue with case ( b ) .
assume that there is a component @xmath604 of @xmath99 with @xmath870 .
after applying an automorphism of the base curve @xmath7 that fixes @xmath175 and @xmath721 , we may assume @xmath871 , so that @xmath872 and @xmath238 vanish on @xmath604 and @xmath807 , respectively , and the product @xmath817 is divisible by @xmath694 . since the points in @xmath411 map under @xmath97 to @xmath440 , and @xmath873 , the points in @xmath271 map under @xmath97 to @xmath874 ; it follows from corollary [ cor310 ] , part ( 1 ) , that @xmath440 is smooth , so we find @xmath816 .
hence , we are done by claim 2 .
we finish with case ( c ) . in that case
, we have @xmath875 , so @xmath238 vanishes on @xmath54 , and also @xmath876 . from proposition [ imageinf ]
we conclude @xmath877 ; it follows from corollary [ cor310 ] , part ( 1 ) , that @xmath440 is smooth , so we find @xmath816 . from
we obtain @xmath878 as @xmath775 ( together with @xmath199 ) implies @xmath690 , we get @xmath691 , which in turn , together with @xmath460 , implies that @xmath54 is reduced . therefore , @xmath238 is a multiple of @xmath694 , so we are done by claim 2 .
indeed , in characteristic @xmath57 , there are other examples than those mentioned in proposition [ outofhell5q ] where @xmath56 has order @xmath57 and no component of @xmath99 maps under @xmath97 to a horizontal curve on @xmath2 ( see example [ ex : char5 ] ) .
it takes less computational force to deal with the case that @xmath56 has order @xmath31 .
[ nohell4q ] suppose @xmath879 in @xmath512 .
then @xmath54 has a component that maps under @xmath95 to a horizontal curve on @xmath2 .
first note that the fiber @xmath440 does not have a cusp , as the additive reduction together with the identity @xmath879 would imply that the characteristic of @xmath1 is @xmath4 , which it is not by assumption .
therefore , by corollary [ cor310 ] , parts ( 1 ) and ( 2 ) , at least one of the points in @xmath271 maps to @xmath880 .
let @xmath318 be such a point and let @xmath604 be a component of @xmath99 that contains @xmath318 .
suppose that @xmath604 is sent by @xmath97 to a fiber on @xmath2 , so that @xmath881 is a point on @xmath7 .
from @xmath882 we conclude @xmath605 , and lemma [ qabove0 ] implies @xmath883 , which contradicts @xmath884 , so @xmath604 is sent to a horizontal curve on @xmath2 .
finally , we deal with the case that @xmath56 has order @xmath5 .
[ nohell3q ] suppose @xmath885 in @xmath512 .
then @xmath886 has a unique component that projects birationally to @xmath887 .
if this component maps under @xmath95 to a vertical curve on @xmath2 , then the pair @xmath715 is isomorphic to one of the pairs described in examples [ ex : order3 ] and [ ex : order3sec ] .
moreover , if another component of @xmath54 maps to a vertical curve on @xmath2 as well , then the pair @xmath715 is isomorphic to one of the pairs described in subexamples [ ex : order3](iii ) and [ ex : order3sec](iii ) . if @xmath729 is an elliptic curve over @xmath1 with a point @xmath888 of order @xmath5 , then by lemma [ explicitunivcurve ] , there exist elements @xmath889 and @xmath733 such that @xmath729 is isomorphic to the elliptic curve given by @xmath890 , with the point @xmath254 corresponding to @xmath738 .
associated short weierstrass models are given by @xmath891 , with the point @xmath738 corresponding to @xmath892 , where @xmath893 and where @xmath894 .
we also simplified these expressions using @xmath895 . since @xmath56 has order @xmath5 in @xmath896 , we find , as in the proof of proposition [ outofhell5q ] , that there are @xmath897 and @xmath898 , such that @xmath899 also in the case that @xmath440 is singular ; the property @xmath900 follows from the assumption @xmath199 . without loss of generality
we assume @xmath901 .
any component @xmath604 of @xmath99 contains a point @xmath902 , which satisfies @xmath903 by corollary [ cor310 ] , part ( 4 ) , so if @xmath904 is contained in a fiber of @xmath71 , then @xmath627 vanishes on @xmath604 .
as the order of @xmath56 is @xmath5 , we have @xmath465 and thus @xmath905 , so @xmath906 and @xmath577 , and the curve @xmath54 is given by @xmath907 with @xmath908 from @xmath577 , we find that @xmath909 is not identically @xmath37 , so indeed , there is a unique component of @xmath54 , say @xmath807 , that projects birationally to @xmath887 .
assume that @xmath910 is contained in a fiber on @xmath2 .
then @xmath627 vanishes on @xmath807 by the above .
if we write @xmath627 as @xmath911 , with @xmath912 $ ] of degree @xmath913 , then we find that @xmath914 vanishes . in the case @xmath915 ,
so @xmath916 , we claim that the pair @xmath715 is isomorphic to one of the pairs of example [ ex : order3 ] .
we sketch a sequence of computations that proves the claim . a priori ,
say over a field in which the elements @xmath779 are independent transcendentals , the polynomial @xmath917 $ ] has degree @xmath782 , but from @xmath465 , it already follows that the degree is at most @xmath918 .
we will use the vanishing of all coefficients to identify the pair @xmath715 .
the vanishing of the coefficient of @xmath797 in @xmath787 gives @xmath919 .
since @xmath722 and @xmath723 do not both vanish , there is a @xmath818 such that @xmath920 and @xmath921 . after applying an automorphism of @xmath7 given by @xmath922
, we may assume without loss of generality that @xmath923 , so @xmath924 .
then the vanishing of the coefficient of @xmath800 in @xmath787 shows that there is an @xmath925 such that @xmath926 and @xmath927 .
the coefficient of @xmath809 now vanishes automatically and the vanishing of the coefficient of @xmath802 yields @xmath928 .
subsequently , the vanishing of the coefficient of @xmath929 gives @xmath930 .
then the coefficient of @xmath804 vanishes automatically .
the vanishing of the coefficient of @xmath931 yields @xmath932 or @xmath933 , but in the latter case , the vanishing of the coefficient of @xmath934 yields the former , so we have @xmath935 in any case .
finally , the vanishing of the coefficient of @xmath934 gives @xmath933 or @xmath936 ; the former case yields example [ ex : order3](i ) with @xmath937 and @xmath938 , while the latter case yields example [ ex : order3](ii ) with @xmath939 , @xmath940 and @xmath941 .
this proves the claim .
we continue ( still ) with @xmath915 .
suppose we are in the case of example [ ex : order3](i ) .
if @xmath97 sends one of the components of @xmath54 given by @xmath942 to a fiber of @xmath273 , then the first argument of this proof shows that @xmath943 vanishes on this component , which implies @xmath936 , so the pair @xmath715 belongs to the family described in example [ ex : order3](iii ) .
now suppose we are in the case of example [ ex : order3](ii ) .
if @xmath97 sends the component of @xmath54 given by @xmath944 to a fiber of @xmath273 , then similarly @xmath651 vanishes on this component , which implies @xmath945 , so again the pair @xmath715 belongs to the family described in example [ ex : order3](iii ) .
this finishes the case @xmath915 .
we now consider the case @xmath946 , so @xmath947 .
we claim that the pair @xmath715 is isomorphic to one of the pairs of example [ ex : order3sec ] .
we sketch a sequence of computations that proves the claim . since @xmath948
, we may apply an automorphism of @xmath949 given by @xmath950 to reduce to the case @xmath798 . as in the case @xmath915
, we will use the vanishing of all coefficients in @xmath787 to identify the pair @xmath715 .
the vanishing of the coefficients of @xmath797 , @xmath800 , @xmath802 , and @xmath929 yields @xmath951 . then the vanishing of the coefficient of @xmath931 yields @xmath952 ; if @xmath953 , then the vanishing of the coefficient of @xmath934 in @xmath787 gives @xmath805 , so we have @xmath805 in any case . then the vanishing of the coefficient of @xmath934 gives @xmath954 or @xmath953 and these cases correspond to subexamples [ ex : order3sec](ii ) and [ ex : order3sec](i ) , respectively
this proves the claim .
suppose we are in the case of example [ ex : order3sec](i ) , so @xmath953 and @xmath955 .
if @xmath97 sends one of the components of @xmath54 given by @xmath956 to a fiber of @xmath273 , then the first argument of this proof shows that @xmath672 vanishes on this component , which implies @xmath957 , so the pair @xmath715 belongs to the family described in example [ ex : order3sec](iii ) .
now suppose we are in the case of example [ ex : order3sec](ii ) , so @xmath954 and @xmath958 .
if @xmath97 sends the component of @xmath54 given by @xmath944 to a fiber of @xmath273 , then similarly @xmath676 vanishes on this component , which implies @xmath957 , so again the pair @xmath715 belongs to the family described in example [ ex : order3sec](iii ) .
this finishes the case @xmath946 and thus the proof .
[ alwayshorizontal ] let @xmath1 , @xmath2 , and @xmath56 be as before .
assume that the pair @xmath715 is not isomorphic to the one of the pairs described in examples [ ex : hell5q ] , [ ex : order3](iii ) , and [ ex : order3sec](iii ) .
if the order of @xmath56 in @xmath512 is @xmath57 , then also assume that the characteristic of @xmath1 is not equal to @xmath57 .
then the rational map @xmath95 sends at least one component of @xmath54 to a horizontal curve on @xmath2 . from the assumption @xmath959
, it follows that the order of @xmath56 is at least @xmath5 .
the statement now follows immediately from propositions [ vertcomp ] , [ outofhell5q ] , [ nohell4q ] , and [ nohell3q ] .
[ whatisthiscurve ] in remark [ tremark ] , we have seen that the closure of the image @xmath104 in @xmath2 is contained in the intersection @xmath519 , where @xmath515 is the image of @xmath960 .
generically , this containment is in fact an equality , but if there are any @xmath75-curves of @xmath2 going through @xmath56 , then these are components of the intersection @xmath519 as well . in degenerate cases ,
the intersection @xmath519 may contain even more components ( see remark [ finalremarkchapter5 ] ) . when studying our del pezzo surfaces of degree one in families , it is more natural to look at the divisor @xmath519 on @xmath2 than at the closure of the image @xmath104
. we will now describe this intersection @xmath519 and its arithmetic genus in terms of the picard group of @xmath2 , at least in the generic case .
generically , the picard group @xmath961 of @xmath2 is generated by the divisor class of @xmath11 .
also generically , the surface @xmath515 has degree @xmath962 and the intersection @xmath519 is irreducible and reduced , so @xmath963 is linearly equivalent with @xmath964 .
the arithmetic genus of @xmath104 is @xmath965 in this case , and @xmath104 has multiplicity @xmath623 at the point @xmath56 , multiplicity @xmath4 at @xmath102 , and is also singular at @xmath966 more points , which agrees with the fact that the geometric genus of the normalization equals @xmath967 these @xmath968 singular points of @xmath104 are the intersection points of @xmath2 with the singular locus of @xmath515 , which is a curve with an embedded point at @xmath56 .
conversely , the family of intersections of @xmath2 with a hypersurface of degree @xmath962 has dimension @xmath969 , as can be seen from the fact that the space of polynomials in @xmath29 of weighted degree @xmath962 modulo the multiples of the defining equation of @xmath2 has dimension @xmath970 or from the fact that the linear system of curves in @xmath14 of degree @xmath971 having multiplicity at least @xmath13 at each of @xmath918 given points has dimension @xmath972 , which equals @xmath969 for @xmath973 .
hence , the subfamily of those intersections that have multiplicity @xmath623 at @xmath56 , multiplicity @xmath4 at @xmath102 , and @xmath966 more singularities , has dimension @xmath974 so there are only finitely many curves satisfying these conditions .
we now give yet another description of @xmath104 that narrows it down to one of only finitely many curves .
note that the projection @xmath975 from @xmath2 to the weighted projective space with coordinates @xmath976 , gives @xmath2 the structure of a double cover of a cone that is ramified at the singular point @xmath64 ( corresponding to the vertex of the cone ) , as well as over the curve given by @xmath977 , i.e. , the locus of nontrivial @xmath4-torsion points .
the involution induced by this double cover is multiplication by @xmath978 on the elliptic fibration .
if we let @xmath979 denote the image of @xmath104 under this involution , then @xmath104 and @xmath980 intersect each other in @xmath981 points on the ramification locus of @xmath982 , as well as @xmath983 points off the ramification locus . the image @xmath984 is a curve of degree @xmath985 that intersects the branch locus of @xmath982 at @xmath981 points , being tangent at each , that has multiplicity @xmath623 at @xmath986 , multiplicity @xmath4 at @xmath987 , and that is singular at @xmath988 more points ( namely the @xmath966 images under @xmath982 of the remaining singular points of @xmath104 , and the @xmath989 images of the intersection points of @xmath104 and @xmath980 ) .
these properties narrow down the @xmath990-dimensional family of curves in @xmath991 of degree @xmath985 to only finitely many curves .
[ finalremarkchapter5 ] given that for all pairs @xmath715 described in the examples in the previous section there are at least six @xmath75-curves on @xmath2 going through @xmath56 , whenever we want to exclude any of these examples , it suffices to assume that @xmath56 does not lie on six @xmath75-curves .
we will now explain in terms of the image @xmath515 of @xmath960 ( cf .
remarks [ tremark ] and [ whatisthiscurve ] ) why it is not surprising that the existence of many @xmath75-curves on @xmath2 through @xmath56 is related to the existence of a component of @xmath99 that maps under @xmath603 to a fiber of @xmath992 .
in example [ ex : hell5q ] , the scheme - theoretic intersection @xmath993 consists of the ten @xmath75-curves going through @xmath56 and the cuspidal fiber @xmath994 with multiplicity @xmath4 . as a divisor on @xmath2
, we have that @xmath431 is linearly equivalent to @xmath964 ( cf .
remark [ whatisthiscurve ] ) .
in general , the pull - back of @xmath519 under the blow - up @xmath995 is a divisor @xmath430 on @xmath63 that consists of * the components of the strict transform @xmath996 of the image @xmath104 , * the strict transforms of the @xmath75-curves on @xmath2 through @xmath56 , * the fiber @xmath169 , and * the zero section @xmath64 with certain multiplicities .
the degree of the restriction of @xmath62 to @xmath430 equals the intersection number of @xmath430 with any fiber of @xmath96 .
generically , this degree equals @xmath997 , the multiplicities of @xmath169 and @xmath64 are @xmath37 , and the multiplicities of the components in ( i ) and ( ii ) are @xmath3 ; the @xmath75-curves on @xmath2 intersect fibers with multiplicity @xmath3 , so the restriction of @xmath96 to @xmath996 has degree @xmath998 , where @xmath999 is the number of @xmath75-curves on @xmath2 through @xmath56 .
note that example [ ex : hell5q ] is not generic in the sense that the multiplicity of the zero section @xmath64 is @xmath4 , thus reducing the degree of the restriction of @xmath96 to @xmath996 to @xmath37 . in general , the multiplicities of the components in ( iii ) and ( iv ) seem to depend only on the order of @xmath56 in @xmath77 and the singularity type of @xmath169 , but
in any case we find that the more @xmath75-curves there are on @xmath2 that go through @xmath56 , the smaller the degree of the restriction of @xmath96 to @xmath996 , forcing all components of @xmath996 to be vertical in extreme cases .
in fact , a thorough investigation of the degree of @xmath515 ( which may itself be nonreduced ) as well as all multiplicities might yield another proof of corollary [ alwayshorizontal ] under the assumption that @xmath56 not lie on six @xmath75-curves of @xmath2 , but it is not clear that this will require less computational effort than the given proof , especially given that even in the generic case , the intersection @xmath519 does not appear to admit a very elegant description ( cf .
remark [ whatisthiscurve ] ) .
in this section , @xmath1 is still a field of characteristic not equal to @xmath4 or @xmath5 .
[ changefibers ] let @xmath1000 be a smooth curve over @xmath1 and @xmath1001 a minimal nonsingular elliptic fibration .
let @xmath114 be a smooth curve over @xmath1 and @xmath1002 a nonconstant morphism .
let @xmath1003 be the minimal nonsingular model of the base change @xmath1004 of @xmath96 by @xmath408 .
let @xmath1005 be a point and set @xmath1006 .
let @xmath1007 and @xmath1008 be the fibers of @xmath96 and @xmath1009 over @xmath42 and @xmath216 , respectively .
let @xmath1010 be the ramification index of @xmath408 at @xmath216 .
then the following statements hold . 1 .
if @xmath1007 has type @xmath1011 for some integer @xmath13 , then @xmath1012 has type @xmath1013 .
if @xmath1007 has type @xmath1014 for some integer @xmath13 , then @xmath1012 has type @xmath1013 for even @xmath1015 and type @xmath1016 for odd @xmath1015 .
if @xmath1007 has type @xmath1017 , then @xmath1012 has type @xmath1018 for @xmath1019 , respectively .
if @xmath1007 has type @xmath167 , then @xmath1012 has type @xmath1020 for @xmath1021 , respectively .
if @xmath1007 has type @xmath1022 , then @xmath1012 has type @xmath1023 for @xmath1024 , respectively .
this follows directly from tate s algorithm ( see @xcite and ( * ? ? ?
* iv.9.4 ) ) . see also ( * ? ? ?
* table vi.4.1 ) , which is stated for characteristic zero .
[ infomn ] suppose @xmath1 is algebraically closed .
let @xmath2 be a del pezzo surface of degree @xmath3 over @xmath1 and @xmath103 the associated elliptic fibration .
let @xmath1025 and @xmath1026 denote the number of singular fibers of @xmath96 of type @xmath166 and type @xmath167 , respectively .
then we have @xmath1027 .
if @xmath96 is not isotrivial , then we have @xmath1028 .
let @xmath30 $ ] be such that @xmath2 is given by .
the surface @xmath63 is a minimal nonsingular elliptic surface with fibers that are either nodal ( type @xmath166 ) or cuspidal ( type @xmath167 ) .
the discriminant @xmath1029 vanishes at points of @xmath7 corresponding to nodal and cuspidal fibers to order @xmath3 and @xmath4 , respectively , so we get @xmath1030 . for any @xmath1031 for which @xmath1032 has type ii ,
both @xmath340 and the @xmath36-invariant @xmath1033 vanish at @xmath1034 ( see ( * ? ? ?
* tate s algorithm , iv.9.4 ) ) , which implies that @xmath34 vanishes at @xmath1034 .
it follows that @xmath34 vanishes at at least @xmath1026 points , so if @xmath1035 , i.e , @xmath1036 , then @xmath46 , so @xmath96 is isotrivial . in characteristic zero , the identity @xmath1027 follows from the more general fact that the euler number of @xmath63 , which is @xmath962 , equals the sum of the local euler numbers , which are @xmath3 and @xmath4 for fibers of type @xmath166 and @xmath167 , respectively ( see ( * ? ? ?
* table iv.3.1 ) and ( * ? ? ?
* lemma iv.3.3 ) ) .
the inequality @xmath1037 implies @xmath1036 , which implies @xmath1038 by ( * ? ? ?
* lemma 1.2 ) , which in turn implies that @xmath96 is isotrivial . for @xmath1039 ,
let @xmath1040 be the universal elliptic curve over the usual modular curve @xmath1041 over @xmath1042 $ ] with a section @xmath254 that has order @xmath149 in every fiber .
then every elliptic curve @xmath729 over a scheme @xmath2 over @xmath1042 $ ] with nowhere vanishing @xmath36-invariant if @xmath728 with a section that has order @xmath149 in every fiber , is the base change of @xmath1043 by a unique morphism @xmath1044 . let @xmath1045 be the usual projective closure of @xmath1041 , and let @xmath1046 be the minimal nonsingular elliptic fibration over @xmath1045 associated to @xmath1043 . from ogg s description of the cusps of @xmath1045 in @xcite , we conclude that for each @xmath1047 and each divisor @xmath13 of @xmath149 , the number of fibers of @xmath1048 of type @xmath1011 is @xmath1049 ( see also @xcite , or ( * ? ? ?
* table 3 ) for explicit models for small @xmath149 , which also show the types of the singular fibers ) .
table [ singfibsx1 ] gives the genus @xmath1050 of @xmath1045 ( see @xcite ) and describes the singular fibers of @xmath1048 for several @xmath149 ( see ( * ? ? ?
* proposition 4.2 ) ) .
.singular fibers of @xmath1048 [ cols=">,^,<",options="header " , ] [ ellfibnotors ] let @xmath2 be a del pezzo surface of degree @xmath3 over @xmath1 and @xmath103 the associated elliptic fibration .
let @xmath114 be a smooth , connected curve over @xmath1 of genus at most @xmath3 , and @xmath1051 a nonconstant morphism .
then the base change @xmath1052 of @xmath96 by @xmath408 has no nonzero section of finite order . without loss of generality , we assume that @xmath114 is projective and that @xmath1 is algebraically closed . as the curve @xmath114 is smooth and connected , it is integral , so it has a unique generic point @xmath1053 that is dense in @xmath114 .
the curve @xmath1054 is an elliptic curve over the function field @xmath1055 of @xmath114 , which is an extension of the function field @xmath1056 of @xmath7 with @xmath1057 .
let @xmath1058 be the @xmath36-invariant of the generic fiber of @xmath96 .
assume that the elliptic fibration @xmath1052 has a nonzero section of finite order , say order @xmath1059 .
then the curve @xmath1054 has a point of order @xmath149 over @xmath1055 . without loss of generality , we assume that @xmath149 is prime .
let @xmath30 $ ] be homogeneous polynomials such that @xmath2 is isomorphic to the surface in @xmath105 given by .
let @xmath1025 and @xmath1026 denote the number of fibers of @xmath96 of type @xmath166 ( nodal ) and @xmath167 ( cuspidal ) , respectively .
then @xmath1060 by lemma [ infomn ] .
we will show that the genus of @xmath114 is at least @xmath4 by considering several cases , thus deriving the contradiction that proves the statement .
\i ) we first consider the case @xmath1061 .
note that @xmath96 itself has no section of order @xmath4 , for if it did , it would be given by @xmath739 and @xmath1062 for some homogeneous polynomial @xmath1063 $ ] of degree @xmath4 , and then @xmath2 would be singular at the point on this section in the fibers given by @xmath1064 .
since the locus @xmath1065 of the @xmath4-torsion points has degree @xmath5 over @xmath949 , it follows that @xmath787 is irreducible . to compute its genus , note that the map
@xmath1066 ramifies whenever @xmath1067 vanishes .
moreover , if @xmath430 vanishes to order @xmath3 at @xmath1068 , which happens if and only if the fiber @xmath66 of @xmath96 has type @xmath166 , then there are two points on @xmath787 above @xmath1034 with ramification indices @xmath3 and @xmath4 , while if @xmath430 vanishes to order @xmath4 , which happens if and only if the fiber @xmath66 has type @xmath167 , then there is a unique point on @xmath787 above @xmath1034 with ramification index @xmath5 .
it follows that the degree of the ramification divisor of @xmath90 equals @xmath1027 , so the riemann - hurwitz formula applied to @xmath90 shows that the genus of @xmath787 equals @xmath1069 . if the base change of @xmath96 by @xmath408 has a section of order @xmath4 , then this section would map nontrivially to @xmath787 , so we get @xmath1070 .
\ii ) we now consider the case that @xmath36 is constant , that is @xmath1071 , and may assume @xmath1072 .
then there are @xmath1073 and @xmath1074 , where @xmath1075 denotes an algebraic closure of @xmath1056 , such that @xmath1076 and @xmath1077 .
if @xmath1078 , then @xmath1079 is contained in @xmath1056 and one checks that @xmath2 is not smooth .
if @xmath1080 , then @xmath2 is not smooth either , so we find @xmath46 and again from smoothness of @xmath2 , we find that @xmath47 is separable and has degree @xmath57 or @xmath32 in @xmath1034 ( cf .
* proposition 3.1 ) ) .
suppose @xmath1081 is a point over @xmath1055 of order @xmath149 .
let @xmath1082 be an algebraic closure of @xmath1055 and @xmath1083 an element satisfying @xmath1084 .
then @xmath1085 is a point of order @xmath149 on the curve given by @xmath1086 , so there are @xmath1087 such that @xmath1088 and @xmath1089 . from @xmath1090 , we get @xmath1091 . if @xmath1092 , then @xmath1093 is contained in @xmath1055 ; if @xmath1094 , then @xmath1095 , so in any case , @xmath47 is a square in @xmath1055 , which implies that @xmath1055 contains a subfield of genus @xmath4 , so @xmath114 has at least genus @xmath4 itself .
\iii ) the case @xmath1096 and @xmath1097 .
if the characteristic of @xmath1 is not equal to @xmath149 , then we set @xmath1098 and @xmath1099 and @xmath1100 and @xmath1101 ; otherwise , we set @xmath1102 and @xmath1103 and @xmath1104 and @xmath1105 . in either case , there is a morphism @xmath1106 such that the elliptic curve @xmath1054 over @xmath1053 is the base change of @xmath1107 over @xmath1108 .
this morphism extends to a morphism @xmath1109 , which is nonconstant because @xmath36 is not constant .
the elliptic surfaces @xmath1110 and @xmath1111 have isomorphic generic fibers @xmath1112 , so their minimal nonsingular models are isomorphic as well by ( * ? ? ? * proposition ii.1.2 and corollary ii.1.3 ) .
let @xmath1003 be this minimal nonsingular elliptic fibration .
@xmath1113_{\pi } & { \mathcal{e}}\times_{{\mathbb{p}}^1 } c \ar[l ] \ar@{<-->}[r ] \ar[d ] & { \mathcal{e } } ' \ar@{<-->}[r ] \ar[dr]_{\pi ' } \ar[dl]^{\pi ' } & \mathbb{e } \times_{x } c \ar[r]\ar[d ] & \mathbb{e } \ar[d]^{\upsilon}\\ { \mathbb{p}}^1 & c \ar[l]^{\tau } \ar@{=}[rr ] & & c \ar[r]_{\chi } & x } \ ] ] set @xmath1114 and let @xmath1115 denote the ramification divisor of @xmath408
. then the degree of @xmath318 is at least @xmath1116 where @xmath1117 denotes the ramification index of @xmath408 at @xmath216 .
lemma [ changefibers ] relates the types of the singular fibers of @xmath1009 to those of @xmath96 and the ramification of @xmath408 on one hand , and to those of @xmath1118 and the ramification of @xmath1119 on the other hand .
the remainder of the proof consists of a largely combinatorial argument to give a lower bound for the degree of @xmath318 , which then , by the riemann - hurwitz formula , yields a lower bound for the genus of @xmath114 .
lemma [ changefibers ] implies that the points @xmath1120 for which the fiber @xmath1121 of @xmath96 above @xmath1122 has type @xmath166 are exactly the points for which the fiber @xmath1012 of @xmath1009 above @xmath216 has type @xmath1123 for some integer @xmath1124 , and exactly the points for which the fiber @xmath1125 of @xmath1118 above @xmath1126 has type @xmath1127 or @xmath1128 for some integer @xmath1129 ; for such points @xmath216 , and integers @xmath909 and @xmath36 , the quotient @xmath1130 is a positive integer and we have @xmath1131 and @xmath1132 . for each @xmath1133 , let @xmath1134 denote the number of fibers of @xmath1118 of type @xmath1127 or @xmath1128 ; for each @xmath1135 , let @xmath1136 denote the number of fibers of @xmath1009 of type @xmath1137 that lie above a point @xmath1120 for which the fiber of @xmath1118 above @xmath1126 has type @xmath1127 or @xmath1128 . for every @xmath1138 we have @xmath1139 .
summing over all @xmath1140 for which the fiber @xmath1141 has type @xmath1127 , we find @xmath1142 for all @xmath1129 . the same argument applied to @xmath408 yields @xmath1143 it follows that the first term of the right - hand side of equals @xmath1144 we consider two subcases .
\a ) the characteristic of @xmath1 is not equal to @xmath149 . from @xmath1145
we conclude @xmath1146 or @xmath1147 or @xmath1148 ( see @xcite ) , and since @xmath1039 is prime , we have @xmath1149 . from table [ singfibsx1 ] above , we find that the fraction in the right - most expression of is at least @xmath1150 .
since @xmath1101 has only fibers of type @xmath1017 , @xmath166 , and @xmath1151 , lemma [ changefibers ] implies that @xmath1009 does not have fibers of type @xmath167 , @xmath1152 , or @xmath1153 . again from lemma [ changefibers ] , this time viewing @xmath1009 as the minimal model of the base change of @xmath96 by @xmath408 , we find that for every @xmath1120 for which the fiber of @xmath96 above @xmath1122 has type @xmath167 , the ramification index @xmath1117 is even , so we have @xmath1154 .
therefore , the second term of the right - hand side of is at least @xmath1155 , so the degree of @xmath318 is at least @xmath1156 .
the riemann - hurwitz formula applied to @xmath408 then yields @xmath1157 , so @xmath1158 .
\b ) the characteristic of @xmath1 is equal to @xmath149 . from @xmath1159
we conclude @xmath1160 ( for a formula for the genus of @xmath1108 , see ( * ? ? ? * and 99 ) ) . from table
[ singfibsx1 ] above , we find that the fraction in the right - most expression of is at least @xmath1161 .
also , from the fact that @xmath96 is not isotrivial , we get @xmath1162 by lemma [ infomn ] , so the degree of @xmath318 is at least @xmath1163 . as before , the riemann - hurwitz formula yields @xmath1164 .
in this section , the field @xmath1 is still of characteristic different from @xmath4 and @xmath5 . [ main ] suppose @xmath1 is infinite .
let @xmath260 be a del pezzo surface of degree @xmath3 over @xmath1 , given by for some homogeneous @xmath30 $ ] of degree @xmath31 and @xmath32 , respectively .
let @xmath1165 be a rational point with @xmath199 .
suppose that the following statements hold
. * if the order of @xmath56 in @xmath512 is at least @xmath31 , then @xmath54 has infinitely many @xmath1-points . *
if the characteristic of @xmath1 equals @xmath57 , then the order of @xmath56 in @xmath512 is not @xmath57 . *
the pair @xmath715 is not isomorphic to a pair described in examples [ ex : hell5q ] , [ ex : order3](iii ) , or [ ex : order3sec](iii ) .
* if the pair @xmath715 is isomorphic to a pair described in subexamples [ ex : order3](i ) or [ ex : order3sec](i ) , then the set of @xmath1-points on @xmath54 is zariski dense in @xmath54 .
given @xmath2 and @xmath56 , we let the curve @xmath54 , its completion @xmath99 , the rational map @xmath603 , the elliptic fibration @xmath103 , and the element @xmath1166 be as in sections [ multisection ] and [ s : completion ] .
we claim that there exists an irreducible component @xmath604 of @xmath99 for which @xmath904 is a horizontal curve on @xmath2 and @xmath1167 is infinite . indeed ,
if the order of @xmath56 in @xmath512 is at least @xmath31 , then @xmath1168 , so @xmath54 is a double cover of @xmath887 , the curve @xmath54 has at most two irreducible components , and if there are two , then there is an involution that switches them , so the first assumption of the theorem implies that @xmath76 is zariski dense in @xmath54 ; thus , there exists an irreducible component @xmath604 of @xmath54 that satisfies the claim by corollary [ alwayshorizontal ] .
suppose , for the remainder of this paragraph and the proof of the claim , that the order of @xmath56 is @xmath5 .
then for any pair @xmath715 that is not isomorphic to one of the pairs described in examples [ ex : order3 ] and [ ex : order3sec ] , the unique component of @xmath54 that projects birationally to @xmath887 satisfies the claim by proposition [ nohell3q ] . for any pair @xmath715 that _ is _ isomorphic to one of the pairs described in those examples , the curve @xmath54 contains a component @xmath604 whose projection to @xmath887 is constant ; this component @xmath604 satisfies the claim , as its image is horizontal by proposition [ nohell3q ] ( by assumption we are not in the subexample [ ex : order3](iii ) or [ ex : order3sec](iii ) ) and density of @xmath1167 follows either automatically in the case of subexample ( ii ) or by assumption in the case of subexample ( i ) .
let @xmath604 be a component of @xmath99 as in the claim , and let @xmath1169 be a normalization of @xmath604 .
then the rational map @xmath603 induces a morphism @xmath1170 .
the composition @xmath1171 corresponds on an open subset to the rational map @xmath1172 , so it is surjective by the claim .
let @xmath1173 denote the section @xmath1174 of the elliptic surface @xmath1175 .
@xmath1176 \ar[rr ] & & { \mathcal{e}}\ar[dd]^{\pi } \ar[dr ] \\ & & & s \ar@{-->}[dl]^{\varphi } \\
\tilde{{\mathcal{c}}}_0 \ar[rr]_{\pi \circ \tilde{\sigma } } \ar@/^5.5mm/[uu]^{\theta } \ar[rruu]^{\tilde{\sigma } } & & { \mathbb{p}}^1 } \ ] ] the section @xmath1173 is not the zero section because @xmath603 sends points @xmath255 whose associated curve @xmath114 is not contained in @xmath2 to a point unequal to @xmath60 .
the genus of @xmath1169 is at most @xmath3 , so by theorem [ ellfibnotors ] , the section @xmath1173 has infinite order . since @xmath1177 is zariski dense in @xmath1169 , it follows that the rational points are dense on the images of all infinitely many multiples of @xmath1173 .
thus , the @xmath1-rational points are dense in the surface @xmath1178 and as this surface maps dominantly to @xmath2 , we conclude that @xmath0 is zariski dense in @xmath2 .
obviously , for any point @xmath1179 , we may apply an automorphism of @xmath949 to ensure that we have @xmath1180 , so the implicit assumption in theorem [ main ] and related statements that @xmath1181 is not a restriction . note that if the order of @xmath56 in @xmath512 is @xmath5 and the pair @xmath715 is not isomorphic to a pair described in examples [ ex : order3](i ) and [ ex : order3sec](i ) , then the hypotheses of theorem [ main ] are automatically satisfied , without further assumptions on @xmath54 .
[ smallsample ] we took a small sample of approximately a hundred randomly chosen del pezzo surfaces over @xmath9 given by with @xmath34 and @xmath35 having only coefficients @xmath37 , @xmath3 , and @xmath978 . for nearly half of the cases , a short point search revealed a rational point @xmath56 for which we could show that it satisfies all conditions of theorem [ main ] , thus proving the rational points are zariski dense .
for the remaining cases , we could still find points @xmath56 , but the coefficients of @xmath54 were too large to show that @xmath54 has infinitely many rational points .
[ ex : tony ] as mentioned in section [ intro ] , a. vrilly - alvarado proves in ( * ? ? ?
* theorem 2.1 ) that if we have @xmath22 and @xmath46 , and some technical conditions on @xmath35 , as well as a finiteness conjecture hold , then the set of rational points is zariski dense on the surface given by .
he also mentions the surface @xmath2 with @xmath46 and @xmath1182 as an example that would not succumb to his methods , so we took @xmath2 as a test example for our method .
unfortunately , the point @xmath1183 of order @xmath5 on @xmath1184 lies on nine @xmath75-curves ( cf .
example [ ex : order3sec](iii ) ) .
it is not hard to find more rational points on this surface , but we did not succeeded in finding any points on the curve @xmath54 associated to any of these points @xmath56 as the coefficients are rather large : for the second - smallest point @xmath1185 , the conductor of the jacobian of @xmath54 has @xmath1186 digits
. n. elkies did prove that the points on @xmath2 are dense with a different method @xcite .
the fact that @xmath56 is not fixed by the automorphism that changes the sign of @xmath72 implies @xmath1187 . without loss of generality , we assume @xmath1188 , say @xmath153 , with @xmath199
. hence , we may apply theorem [ main ] .
the last hypothesis of theorem [ maincor ] implies the last two of theorem [ main ] , which shows that @xmath0 is indeed zariski dense in @xmath2 .
note that any point @xmath1189 on an elliptic curve given by @xmath1190 has order @xmath5 if and only if @xmath1191 .
define the polynomials @xmath1192 and @xmath1193 .
suppose we have @xmath81 , @xmath82 , and @xmath83 .
since every elliptic curve over the real numbers @xmath8 has a nontrivial @xmath5-torsion point , we may choose a nonzero rational number @xmath1194 and a point @xmath1195 such that the fiber @xmath165 given by @xmath1196 is smooth , the point @xmath56 has order @xmath5 in @xmath1197 , and @xmath56 does not lie on six @xmath75-curves on @xmath2 .
set @xmath1198 , so that @xmath56 being @xmath5-torsion implies @xmath1199 .
choose @xmath1200 close to @xmath1201 and @xmath158 , respectively , and set @xmath1202 and @xmath1203 .
also set @xmath1204 so that @xmath1205 and @xmath1206 are the polynomials obtained from @xmath34 and @xmath35 after replacing @xmath88 and @xmath89 by @xmath90 and @xmath91 , respectively .
then we have @xmath1207 and @xmath1208 , so @xmath56 lies on the surface @xmath93 given by with the two values @xmath88 and @xmath89 replaced by @xmath90 and @xmath91 , respectively . if we choose @xmath1209 and @xmath1210 arbitrarily close to @xmath1201 and @xmath158 , then @xmath90 and @xmath91 will be arbitrarily close to @xmath88 and @xmath89 . by choosing them close enough
, we also guarantee that @xmath93 and @xmath1211 are smooth , and that @xmath1212 does not lie on six @xmath75-curves on @xmath93 . from the identity @xmath1213
we conclude that @xmath1212 has order @xmath5 in @xmath1214 , so we may apply theorem [ maincor ] , which yields that @xmath92 is zariski dense in @xmath93 .
[ specinforder ] let @xmath1 be an infinite field and @xmath1215 an elliptic fibration over @xmath1 with a nontorsion section .
then there are infinitely many points @xmath1216 for which the fiber @xmath1217 contains infinitely many @xmath1-rational points .
if @xmath1 is algebraic over a finite field , then this follows from the weil conjectures .
otherwise , we replace @xmath1 without loss of generality by an infinite subfield that is finitely generated over its prime subfield , over which everything is defined .
then @xmath1 is either a number field or a transcendental extension of its prime field and in all cases , @xmath1 is hilbertian ( see @xcite for number fields and ( * ? ? ?
* theorem 13.4.2 ) for a modern treatment of the general case ) .
the lemma now follows immediately from nron s specialization theorem ( * ? ? ?
* thorme iv.6 ) ( see also ( * ? ? ?
* theorem 7.2 ) and ( * ? ? ? * remark 3.7(1 ) ) ) . without loss of generality ,
we assume that the nodal fiber lies above @xmath175 . when @xmath56 runs over the nodal curve @xmath440 , the curves @xmath99 form a family of genus - one curves .
more precisely , the equation in describes a surface @xmath1218 , and if @xmath153 is a point on @xmath440 with @xmath199 and not of order @xmath5 in @xmath512 , then the fiber of the projection @xmath1219 above @xmath157 is isomorphic to @xmath99 .
the fibered product @xmath1220 is the family of curves @xmath99 , at least outside finitely many points @xmath1221 .
let @xmath1222 be such that @xmath455 and @xmath456 .
note that we have two rational maps @xmath1223 , for @xmath1224 , that are rational sections of @xmath91 , namely given by @xmath1225 with @xmath463 and @xmath464 as in lemma [ l : e0node ] .
these maps extend to morphisms and we choose @xmath1226 to be the zero section , making @xmath91 a jacobian elliptic fibration .
we claim that the section @xmath1227 has infinite order . the model
@xmath1228 is highly singular , so instead we consider the surface @xmath1229 that is the image of @xmath1108 under the birational map @xmath1230 with @xmath1231 where @xmath1232 depend on @xmath157 ( which is now variable instead of fixed ) as they did before .
note that @xmath581 is nonzero because @xmath2 is smooth . for @xmath1233 ,
the fibers of @xmath1234 and @xmath1235 are isomorphic .
the model @xmath1236 is given by @xmath1237 , where @xmath1238[{\overline{p}},{\overline{r}}]$ ] is homogeneous of degree @xmath31 .
the fiber @xmath1239 of @xmath1240 above @xmath1241 is given by @xmath1242 this fiber @xmath1239 is singular at the point @xmath1243 , and in fact so is @xmath1236 , but the fiber is smooth everywhere else . the sections @xmath1226 and @xmath1227 correspond to the sections @xmath1244 and @xmath1245 of @xmath1240 , respectively .
these sections intersect in the point @xmath1246 , which is smooth in its fiber .
therefore , in a minimal nonsingular projective model @xmath1247 of the fibration @xmath91 , the two sections intersect as well . hence , @xmath1248 is in the kernel of the reduction @xmath1249 , where @xmath1250 is the fiber of @xmath1251 above @xmath1252 .
this kernel is isomorphic to a subgroup of the formal group associated to @xmath1251 ( or @xmath1240 ) and the completion of @xmath1253 $ ] at the maximal ideal @xmath1254 , cf .
* proposition vii.2.2 ) . by ( *
* proposition iv.3.2(b ) ) , all torsion elements of the formal group have @xmath934-power order , where @xmath934 is the characteristic of @xmath1 .
this proves the claim for @xmath944 ( cf .
* theorem 1.1(a ) ) ) .
we now assume @xmath1255 and determine the kodaira type of the singular fiber @xmath1250 of @xmath1251 .
one checks that the discriminant of @xmath1256 equals @xmath1257 where @xmath430 is a polynomial of degree @xmath1258 satisfying @xmath1259 . hence , the valuation of @xmath340 at @xmath1252 equals @xmath5 ; the fiber @xmath1239 described in is nodal , so the reduction is multiplicative and we conclude from ( * ? ? ?
* tate s algorithm iv.9.4 ) that @xmath1250 has type @xmath1260 .
it follows that the @xmath36-invariant of @xmath91 is not constant .
suppose that @xmath1248 is torsion .
then @xmath1251 admits a section of order @xmath934 , so there is a surjective morphism @xmath1261 such that the generic fiber of @xmath1251 is isomorphic to the generic fiber of the base change of the fibration @xmath1262 by @xmath1263 ( cf .
part iii of the proof of theorem [ ellfibnotors ] ) .
this implies that the minimal nonsingular model of this base change is isomorphic to @xmath1251 .
however , the existence of @xmath1263 implies that @xmath1264 has genus @xmath37 , so @xmath1265 , and from lemma [ changefibers ] and table [ singfibsig ] we find that no base change of @xmath1266 has a minimal nonsingular model with fibers of type @xmath1260 .
this contradicts the fact that @xmath1250 has type @xmath1260 and the claim follows .
it follows that the section @xmath1267 induced by @xmath1227 also has infinite order on the elliptic fibration @xmath1268 with the section induced by @xmath1226 as zero section . after replacing @xmath440 by its normalization
we may apply lemma [ specinforder ] , which implies that the curve @xmath99 has infinitely many rational points for infinitely many @xmath1269 , in particular for some @xmath56 of order larger than @xmath57 in @xmath512 .
theorem [ maincor ] then shows that @xmath0 is zariski dense in @xmath2 .
couveignes , j .-
m . , and edixhoven , s. , , ed .
bas edixhoven & jean - marc couveignes , with robin de jong , franz merkl & johan bosman , ann . of math .
, no . * 176 * , princeton univ . press , princeton , nj , 2011 . | we state conditions under which the set @xmath0 of @xmath1-rational points on a del pezzo surface @xmath2 of degree @xmath3 over an infinite field @xmath1 of characteristic not equal to @xmath4 or @xmath5 is zariski dense .
for example , it suffices to require that the elliptic fibration @xmath6 induced by the anticanonical map has a nodal fiber over a @xmath1-rational point of @xmath7 .
it also suffices to require the existence of a point in @xmath0 that does not lie on six exceptional curves of @xmath2 and that has order @xmath5 on its fiber of the elliptic fibration .
this allows us to show that within a parameter space for del pezzo surfaces of degree @xmath3 over @xmath8 , the set of surfaces @xmath2 defined over @xmath9 for which the set @xmath10 is zariski dense , is dense with respect to the real analytic topology .
we also include conditions that may be satisfied for every del pezzo surface @xmath2 and that can be verified with a finite computation for any del pezzo surface @xmath2 that does satisfy them . |
gravitational lensing is one of the most promising tools in modern astrophysics for probing the matter content of the universe directly , and thereby constraining cosmological models
. applications of strong gravitational lensing include measurements of the hubble constant using lensed quasars with time - delays ( refsdal 1964 ; _
e.g. _ of time delay measurements : barkana 1997 ; cohen et al . 2000 ;
burud et al .
2000 ; burud et al .
2002 a , b ; fassnacht et al . 2002 ; hjorth et al .
2002 , ofek & maoz 2003 ; jakobsson et al . 2005 ; kochanek et al . 2006 ; fohlmeister et al . 2006 ; vuissoz et al . 2007 ) and the determination of galaxy and cluster mass distributions using multiple images ( _ e.g. _ guzik & seljak 2002 , mandelbaum et al .
2006 , read et al .
moreover , homogeneous samples of strong lenses allow for a statistical determination of the properties of lensing galaxies ( _ e.g. _ kochanek 1996 ; rusin & ma 2001 ; chae , mao & kang 2006 , chae 2007 ) and samples of strong lenses selected based on source properties are used to test cosmological models ( _ e.g. _ fukugita & turner 1991 ; wambsganss et al .
2005 ; moeller , kitzbichler & natarajan 2006 ; hilbert et al . 2007 ) .
the giant arcs that are observed around massive clusters of galaxies are among the most spectacular lensing phenomena ( _ e.g. _ gladders et al .
2002 searched giant arcs in the red - sequence cluster survey ( rcs ) ; scarpine et al .
2006 , hennawi et al . 2006 ,
estrada et al .
2007 in the sloan digital survey ( sdss ) ) .
these cases are relatively rare because each square degree of sky contains only about one cluster of sufficient mass to produce a giant arc .
much more common are the strong lensing events found around massive early - type galaxies .
the first studies of galaxy - galaxy strong lensing systems targeted distant , _ bright _ lensed sources , including quasars ( eso hamburg survey : wisotzki et al . 1996 ; cfa - arizona - st - lens - survey : peng et al .
1997 , muoz et al .
1998 ) and radio sources ( jodrell bank - vla astrometric survey : patnaik et al .
1992 , browne et al . 1998
; cosmic lens all sky survey : jackson et al . 1995 , myers et al .
the identification of multiple images in these systems is generally unambiguous and does not require very deep observations ; but requires imaging over a large fraction of the sky .
the advent of deep and wide optical imaging combined with large spectroscopic surveys has greatly broadened the opportunities and means by which strong galaxy - galaxy lensing events can be found . indeed , since we expect that @xmath00.1% of massive galaxies ( the luminous elliptical galaxies ) strongly distort and magnify a galaxy source at z@xmath11 ( miralda - escud & lhar 1992 ) , new wide field surveys are likely to discover many hundreds of fainter galaxy - galaxy strong lensing systems .
spectroscopic identification of two galaxies at different redshifts in the same sloan digital sky survey fiber has proved to be a successful method in the sloan lens acs survey ( slacs : bolton et al .
deep , wide - field imaging surveys ( especially those utilizing the exquisite image resolution available from space ) provide another avenue for locating strong galaxy - galaxy lenses .
this was first attempted in the _ hubble space telescope _
medium deep survey ( griffith et al . 1994 , ratnatunga et al .
1999 ) , and more recently in the great observatories origins deep survey field ( goods ; dickinson et al .
2001 , fassnacht et al .
2004 ) or in the aegis survey ( moustakas et al .
automated software for identifying strong lenses in _ hst _ image archives are currently under development ( marshall et al .
2005 , moustakas et al . 2006 ) and appear to be very promising .
searches for strong lens systems are also underway from ground - based wide - field surveys .
in particular the sl2s initiative ( cabanac et al .
2007 ) was able to find a few lenses per square degree in the cfht - ls wide survey .
however , ground based imaging surveys are necessarily limited to finding systems with large deflection angles because of the seeing limitation compared to space - based observations .
the 1.64 square degree _
hubble space telescope _
cosmos survey ( scoville et al . 2007 ) provides an excellent opportunity to locate and study a large number of strong galaxy - galaxy lensing systems .
it includes the largest contiguous high - resolution astronomical imaging survey ever performed from space .
although observations with the _ hst _ advanced camera for surveys ( acs ) have only been obtained in a single band ( f814w filter ) , multi - wavelength coverage of this equatorial field is provided in the optical band by deep multicolor imaging from suprime on the _ subaru _ telescope and megacam at _
cfht_. in this paper , we present the first systematic attempt to locate strong galaxy - galaxy lensing systems in the cosmos field . the data - set used for our investigation
is described in @xmath2 [ data ] and the techniques used to uncover strong galaxy - galaxy lensing candidates are discussed in @xmath2 [ seccriteria ] .
our findings are presented in @xmath2 [ candidates ] .
the lens potential modeling scheme is discussed in @xmath2 [ model ] and we compare our lens sample to other strong lens catalogs in @xmath2 [ seccomparison ] . finally , the results are discussed in @xmath2 [ secdiscussion ] . throughout this paper ,
galaxy magnitudes are quoted in the ab system and cosmological distances are calculated assuming a @xmath3cdm model with parameters h@xmath4 = 73 km s@xmath5 mpc@xmath5 , @xmath6=0.3 , and @xmath7=0.7 .
the galaxies of the cosmos field : v@xmath8-band absolute magnitude versus photometric redshift .
the grey solid line contours show isodensity numbers of all the galaxies in the entire photometric catalog ( levels : 1 , 100 , 300 , 500 , 1000 , 1200 , 1600 galaxies ) .
the black solid line contours draw the isodensity numbers of galaxies in the ` p'-catalog ( levels : 1 , 10 , 30 , 50 , 100 , 140 galaxies ) .
the red crosses point the lensing galaxy candidates described in sect .
[ candidates ] .
, width=302 ] to conduct our strong galaxy - galaxy lens search we take advantage of three key data - sets of the cosmos survey : the acs/_hst _ high resolution imaging , the suprime/_subaru _ imaging , and the megacam/_cfht _ multicolor imaging .
we have analyzed the full acs field of view of the cosmos field which corresponds to 1.64 square degree observed during cycle 12 and 13 .
we use data release v1.2 of the cosmos acs data which can be accessed through the dedicated database at irsa .
individual exposures were combined using the multidrizzle software ( koekemoer et al .
2007 ) onto an output grid of pixel size 0.03 .
the limiting magnitude is about i@xmath9 mag .
a detailed description of the data is available in scoville et al .
( 2007 ) , and a detailed description of the acs - selected galaxy catalog can be found in leauthaud et al .
( 2007 ) .
multicolor imaging of the cosmos field was obtained from the _ subaru _ telescope with the suprime camera ( miyazaki et al .
2002 ) in @xmath10 bands , and from the _ cfht _ with the megacam camera ( boulade et al .
2003 ) in the @xmath11 and @xmath12 bands . in these data ,
the seeing is generally better than 1 .
a photometric catalog was derived from a combination of these data , based on detections in the _ subaru _
@xmath13 band ( capak et al . 2007 ) .
the publicly available _ le phare _ photometric redshift estimation code ( ilbert et al .
2006 ) has been used to measure the redshifts of 278526 galaxies with @xmath14 mag .
details concerning the multi - wavelength photometry can be found in mobasher et al .
the @xmath14 mag limit ensures a good accuracy of both the photometric redshift estimation and the fit to the spectral energy distribution ( sed ) .
the ground - based photometric zero - points were calibrated ( and the sed templates adjusted accordingly ) using 1095 spectroscopic redshifts from the zcosmos survey ( lilly et al .
2006 ) . using 8 bands
, this method achieves a photometric redshift recovery accuracy of @xmath15 with @xmath16 of catastrophic errors , defined as @xmath17 ( further details can be found in mobasher et al . 2007 ) .
a four - step procedure is employed in order to identify strong galaxy - galaxy lens systems . in order
, the four steps are the following : 1 .
select a list of potential lenses from the photometric redshift catalog ( hereafter ` p ' catalog ) .
2 . visually inspect the acs images of all galaxies in the ` p ' catalog to produce a catalog of potential strong galaxy - galaxy lens systems ( hereafter ` e ' catalog .
only 302 galaxies are in this catalog , the 35 others are missing from our personal archives . ) .
3 . investigate the ` e ' catalog using multicolor images to check for color differences between the main galaxy and the potentially lensed object .
4 . subtract a galaxy surface brightness model of the foreground galaxy to determine the morphology and lensing configuration of the potentially lensed background galaxy . in our methodology , systems that pass steps
both 3 and 4 qualify as strong galaxy - galaxy lens candidates .
we will now describe each step in more detail .
examples of strong lens candidates ruled out by the pseudo - color image test .
north is to the top of the figure and east is to the left .
the left stamp images are the acs i - band images of the systems .
the right stamp images are the color images.,title="fig:",width=264 ] examples of strong lens candidates ruled out by the pseudo - color image test .
north is to the top of the figure and east is to the left .
the left stamp images are the acs i - band images of the systems .
the right stamp images are the color images.,title="fig:",width=264 ] the acs - based photometric redshift catalog limited to @xmath18 mag contains 278819 entries : it is unrealistic to inspect visually all galaxies in this catalog .
furthermore , this would be inefficient , because we know that a galaxy needs to be massive and not too distant to efficiently distort and strongly magnify a more distant galaxy .
we therefore limit our search to intrinsically bright galaxies at moderate redshifts .
furthermore , for this first investigation of the cosmos field , we only look at early - type galaxies ( classified spectroscopically by the photometric redshift code ) , which generally have a simple , smooth surface brightness distribution . once again , we emphasize that the main reason for applying these selection criteria to the photometric redshift catalog is to reduce the sample of galaxies that need to be visually inspected .
it is highly possible that additional strong lens systems will be found around galaxies that do not belong to our search sample . the number of strong lensing systems that we find is thus a lower limit to the total number of strong lensing systems in the cosmos field . to search for potential lenses , we select galaxies according to the following criteria : * their photometric redshift must be in the range @xmath19 .
the lower redshift bound of @xmath20 is chosen to ensure that galaxies will indeed produce lensing events and the upper bound is essentially motivated to ensure a good accuracy on the photometric redshift given that no deep infrared data are yet available for the cosmos field .
the galaxy is retained if the error bars at 68% confidence level overlap , even partially , the redshift range [ 0.2;1.0 ] .
* their luminosity must be @xmath21 mag , on the assumption that the brightest galaxies are also likely to be the most massive ones with the largest cross - section for lensing .
* we limit our search to early - type galaxies that have been spectrally classified as such by the sed fitting of our photometric redshift algorithm .
the motivation for this criterion is to focus on massive systems with a simple light distribution that is easy to fit and to subtract .
we obtain a final ` p ' catalog of 9452 bright elliptical galaxies ( see fig . [ mvz ] ) .
further strong lensing searches using automated software will certainly relax these criteria .
during the course of our visual inspection , we serendipitously found some additional strong lens systems in which the lensing galaxy did not comply with the above limits .
although these constitute an inhomogeneous sample , we add them to the list of strong lensing systems for completeness .
those systems are identified in tables [ photom ] , and [ photom1 ] .
we inspect visually a postage - stamp 10@xmath2210 acs image surrounding all galaxies in the ` p ' catalog , looking for lensed features ( tangential arcs or multiple images ) .
the search box size limits us to r@xmath23 .
consequently , this puts an upper limit on the mass scale for the lens , which excludes clusters as a possible main deflector .
this inspection was performed by 5 of the co - authors using the ds9 fits image visualization tool .
our strategy reduced the 9452 galaxies in the ` p ' catalog to an ` e ' catalog of 337 candidates , with intentionally loose selection criteria to avoid excluding any real lenses at this stage .
a color image is always useful to verify potential lens candidates . indeed ,
gravitational lensing is achromatic , thus a gravitational arc or multiple images systems should have a consistent color across the image(s ) .
furthermore , the color of a lensed galaxy will only rarely have the same color as the main lensing galaxy .
this color test hence provides a good way to discriminate false strong lensing cases .
however , with our particular combination of ground- and space - based data ( especially when the seeing approaches 1@xmath24 ) , the color test is only useful for arcs more distant than @xmath00.5 from the lensing galaxy , because of strong contamination from the wings of the lensing galaxy .
for the 337 systems in the ` e ' catalog , we produced pseudo - color images made from the `` best psf '' stacks of the @xmath25 , @xmath26 and @xmath27 _ subaru _ images , using the rgb feature in the ds9 software , as well as ground - based images sharpened with the acs images .
the color images use the acs f814w data as an illumination map and the _ subaru _ @xmath28 , @xmath29 , and @xmath30 images as a color map . to achieve this each _ subaru _
image is divided by the average of the three _ subaru _ images , then multiplied by the acs f814w image .
this preserves the flux ratio between images , while replacing the overall illumination pattern with the f814w data .
each image is then smoothed by a 1 pixel gaussian to reduce the noise and divided by @xmath31 to enhance the color difference between star forming and passive galaxies .
the processed @xmath28 , @xmath29 , and @xmath30 images are then assigned to the blue , green , and red channels respectively .
the resulting images have the high spatial resolution of the acs imaging , but color gradients at ground based resolution .
of the 337 candidates , 67 have plausible colors for strong lensing systems .
the color images are displayed in figs .
[ figure11 ] to [ double ] . for illustrative purposes ,
two objects that were rejected at this stage are shown in fig .
the system displayed in the top panel of fig .
[ bad ] shows two arc - like features in the acs , north - east of the main galaxy bulge . with the single - color acs data - set only , it would not be clear whether these features were part of the galaxy s spiral arms , or distorted images of a background object .
the pseudo - color image shows blue knots of star formation all around the disc of this galaxy , suggesting that the arc - like features are most probably genuine regions of the main galaxy .
fitting and extracting the foreground lens galaxy light from the images reveals the shape of the more distant , lensed object . for this work , a robust 2-dimensional fit to the galaxy surface brightness distribution
is derived using @xmath32 ( simard 1998 , marleau & simard 1998 ) .
we adopt a srsic bulge plus exponential disc parametrisation to describe the two - dimensional surface brightness distribution of the lensing galaxy light profile .
the srsic law is parametrised by means of the total flux in the bulge , the srsic index @xmath33 , the bulge ellipticity ( @xmath34 ) , the position angle of the bulge and the effective radius of the bulge @xmath35 .
the exponential profile depends on the photometric disk total flux , the disk scale - length , the disk position angle and the disk inclination .
the @xmath32 software finds the best fitting values for all of these parameters , and provides them as an output of the decomposition process .
in addition to the structural parameters for each object , we calculate a model image of each galaxy , which is convolved with a point spread function before comparison with real data .
an image of the residuals is obtained by the subtraction of the model from the science image .
we present the result of the galaxy fits in tables [ photom ] , [ photom1 ] and in figs . [ figure11 ] to [ figure7 ] . in most cases
, the arc candidates appear clearly after the galaxy model is subtracted ( panel 3 in those figures ) . in 10 cases
, the modeling and subtraction of the lens galaxy proves difficult .
the problems encountered include irregular galaxy morphology and compact galaxy size .
we display the acs and color images of these systems in fig .
[ double ] .
additionally , the system cosmos 5737 + 3424 was discovered serendipitously in a part of the ground - based data - set not covered by the acs images .
we present just its pseudo - color image in fig .
[ double ] .
the location of lens candidates in the absolute v - band magnitude versus redshift plane are indicated by crosses in fig .
[ mvz ] .
we divide the sample in two sub - samples : the multiple arcs and long curved image systems , which we refer to hereafter as the `` best systems '' , and less certain single arclet systems . given their distinctive shape , the multiple images and long curved arcs have a greater probability of being genuine strong lenses relative to the single arclet systems . throughout the rest
this paper we therefore restrict our study to the sample of `` best systems '' .
the stamp images of the best systems are shown in figs .
[ figure11 ] , [ figure22 ] , [ figure33 ] and [ double ] and discussed in sect .
images of the single arclet systems are displayed in figs .
[ toto1 ] , [ figure2 ] , [ figure3 ] , [ figure4 ] , [ figure5 ] , [ figure6 ] , [ figure7 ] and [ double ] and discussed in sect .
their characteristics are given in tables [ photom ] , and [ photom1 ] . in fig .
[ historarc ] we have reported the distribution of the arc radius of the cosmos strong lens candidates .
most systems have an arc radius@xmath36 2 . the spectroscopic redshift distribution of the lensing galaxies in the slacs survey ( hashed histogram ) and in the castles database ( solid line histogram ) , compared to the photometric redshift distribution of the best systems in the cosmos survey ( filled - in grey histogram ) . , width=264 ] in twenty cases , we have detected systems with a single long curved image or with two or more arcs of similar color symmetrically located around the lensing galaxy or with ring - like features .
their coordinates and characteristics are listed in table [ photom ] .
we comment on the most interesting or / and less trivial systems hereafter .
a bright arc - like feature is clearly visible in the acs image of the system cosmos 0012 + 2015 .
the counter - arc becomes visible after subtraction of the lensing galaxy profile , south - west of the center of the lens galaxy .
the system cosmos 0018 + 3845 is most probably a triply or possibly quadruply imaged background galaxy .
the north - west arc would be a single image of the galaxy while the south - east arc would be a composite of two or three images of the same galaxy .
the system cosmos 0038 + 4133 is a complete einstein ring candidate , formed by the distorted image of a single background galaxy .
four systems show ring - like structures around a central elliptical object ( cosmos 0049 + 5128 , cosmos 0124 + 5121 , cosmos 5941 + 3628 and cosmos 5947 + 4752 ) . in those cases ,
it is difficult to conclude without any spectroscopic data whether the ring is due to gravitational lensing or if it is a genuine part of the central galaxy ( like a ring - galaxy or a shell ) .
the main argument in favor of them being lens candidates is the strong color difference between the elliptical central objects and the ring features , suggesting they are not physically related .
there are probably four images of a background source in the candidate system cosmos 0050 + 4901 .
three images are combined in an arc south - west of the lensing galaxy , while a counterpart image is visible north - east of the galaxy .
the system cosmos 0211 + 1139 displays two giant arcs , each with different colors in the _ subaru _ image .
they are probably distorted images of two different background galaxies .
the pseudo - color image of cosmos0216 + 2955 shows a strong color difference between the galaxy and the potential arcs .
nevertheless , the galaxy subtraction shows features that are not fitted by the sersic profile .
they seem to be linked to the galaxy core .
hence , there are two possible explanations for this system : either it is actually a strong lens , and the lensing galaxy is a spiral galaxy , or the two arcs observed are part of the arms of a spiral galaxy .
the system cosmos0254 + 1430 has two lensing galaxies at very short distance .
we have not been able to make a clean galaxy light profile fit .
the two arclets located north - east and south - west have are clearly visible in the color image . in the system cosmos 5737 + 3424 ,
we observe at least three giant arc like features around the central galaxies of an identified xmm detected cluster ( finoguenov et al .
2007 ) .
cosmos 5921 + 0638 is a quadruply - imaged lensed quasar or very compact source .
an einstein ring is clearly visible in the acs / f814w band image .
l l l l l l l l r l l l 0012 + 2015 & 10:00:12.6 + 02:20:15 & 0.41@xmath37 & 0.90 & 0.53 & 8.1&19.28@xmath38 & -10.2 & 0.25 & 21.6 + 0013 + 2249 & 10:00:13.9 + 02:22:49 & 0.38@xmath39 & 1.65 & 0.90 & 3.9 & 18.89@xmath38 & 38.2 & 0.16 & 22.4 + 0018 + 3845@xmath40 & 10:00:18.4 + 02:38:45 & 0.71@xmath41 & 0.40 & 0.29 & 5.6&23.60@xmath38 & -21.8 & 0.23 & 22.7 + 0038 + 4133 & 10:00:38.2 + 02:41:33 & 0.89@xmath42 & 0.74 & 0.72 & 5.2&20.39@xmath43 & -2.9 & 0.26 & 20.5 + 0047 + 5023 & 10:00:47.6 + 01:50:23 & 0.85@xmath44 & 0.70 & 0.72 & 1.9 & 20.65@xmath38 & 33.4&0.05 & 22.8 + 0049 + 5128 & 10:00:49.2 + 01:51:28 & 0.33@xmath37 & 2.22 & 1.07 & 1.2 & 19.61@xmath38 & -24.6&0.22 & 23.3 + 0050 + 4901@xmath40 & 10:00:50.6 + 02:49:01 & 1.01@xmath44 & 1.90 & 0.36 & 2.9 & 21.72@xmath45 & 26.6 & 0.39 & 22.7 + 0056 + 1226 & 10:00:56.7 + 02:12:26 & 0.44@xmath46 & 1.20 & 0.96 & 4.6 & 18.70@xmath47 & 40.3 & 0.00 & 23.3 + 0124 + 5121 & 10:01:24.5 + 01:51:21 & 0.84@xmath46 & 0.84 & 0.24 & 2.4&22.43@xmath48 & -51.4 & 0.25 & 23.2 + 0211 + 1139 & 10:02:11.2 + 02:11:39 & 0.90@xmath49 & 3.20 & 0.74 & 4.1&21.09@xmath50 & 11.0 & 0.35 & 21.6 + 0216 + 2955 & 10:02:16.8 + 02:29:55 & 0.67@xmath42 & 1.96 & 0.90 & 6.1 & 19.98@xmath38 & -16.6 & 0.06 & 22.0 + 0227 + 0451@xmath40 & 10:02:27.5 + 02:04:51 & 0.89@xmath37 & 1.62 & 0.30 & 6.3 & 21.94@xmath51 & 80.9 & 0.33 & 22.3 + 0254 + 1430 & 10:02:54.0 + 02:14:30 & _ & 1.54 & _ & _ & _ & -56.5 & 0.40 & 22.1 + 5737 + 3424@xmath52 & 09:57:37.0 + 02:34:24 & _ & 12.0 & _ & _ & _ & _ & _ & _ + 5758 + 1525@xmath52 & 09:57:58.6 + 02:15:25 & _ & 0.35 & _ & _ & _ & _ & _ & 21.7 + 5857 + 5949 &
09:58:57.0 + 01:59:49 & 0.39@xmath37 & 2.15 & 0.61 & 1.4 & 20.05@xmath38 & -32.8 & 0.36 & 21.9 + 5914 + 1219&09:59:14.7 + 02:12:19 & 1.05@xmath53 & 1.86 & 0.27 & 3.9 & 23.25@xmath54 & 14.4 & 0.08 & 22.8 + 5921 + 0638 & 09:59:21.7 + 02:06:38 & 0.45@xmath37 & 0.80 & 0.49 & 6.8 & 20.34@xmath55 & 29.9 & 0.08 & 20.6 + 5941 + 3628 & 09:59:41.3 + 02:36:28 & 0.90@xmath39 & 1.21 & 0.76 & 1.0&20.91@xmath55 & -5.4 & 0.08 & 22.8 + 5947 + 4752 & 09:59:47.8 + 02:47:52 & 0.28@xmath46 & 2.55 & 0.51 & 1.6 & 19.83@xmath56 & 4.1 & 0.07 & 22.8 forty seven of the discovered lens candidates have a single arclet identified in both the acs and the color images . in these cases , no counter image for these arcs could be found in the residuals of the galaxy - subtracted image .
as observed on the _ subaru _ data , the arclets are generally much bluer than the lensing galaxies , and are likely to be only weakly distorted by the galaxy .
their coordinates and characteristics are give in table [ photom1 ] .
we comment below on a few interesting systems .
the system cosmos 0009 + 2455 shows two elongated features symmetrically distributed around the elliptical galaxy .
the two features have slightly different colors , therefore if the system is a genuine lens , the sources are probably two different objects .
the arclet in the system cosmos 0028 + 1919 is located west of the lensing galaxy , and appears in blue in the color image . in the system cosmos 5805 + 0413
, a giant blue arclet appears between two galaxies .
the arc is slightly curved in direction of the western galaxy but there are no counterpart images around any of the two galaxies .
the parameters of the early spectral type galaxy at the center of the stamp image are presented in the table [ photom1 ] .
l l l l l l l
r l l 0009 + 2455 & 10:00:09.7 + 02:24:55 & 0.39@xmath46&2.63 & 0.70 & 3.7 & 19.49@xmath38 & -2.5 & 0.31 & 22.6 + 0028 + 1919 & 10:00:28.6 + 02:19:19 & 0.61@xmath37&1.87 & 0.45 & 1.7 & 20.34@xmath57 & 34.9 & 0.13 & 22.6 + 0029 + 4018 & 10:00:29.6 + 02:40:18 & 0.83@xmath53&2.99 & 0.83 & 2.6 & 18.94@xmath55 & 47.4 & 0.00 & 22.7 + 0047 + 2931 & 10:00:47.1 + 02:29:31 & 0.32@xmath46&2.00 & 0.73 & 2.3 & 19.56@xmath38 & -38.4 & 0.29 & 21.6 + 0050 + 0357 & 10:00:50.8 + 02:03:57 & 0.84@xmath46&0.66 & 0.31 & 2.8 & 22.31@xmath55 & -44.7 & 0.40 & 22.5 + 0055 + 3821 & 10:00:55.7 + 01:38:21 & 0.66@xmath58&1.05 & 0.40 & 2.4 & 20.86@xmath38 & 10.3 & 0.14 & 21.8 + 0056 + 2106 & 10:00:56.7 + 02:21:06 & 0.65@xmath37&1.75 & 0.34 & 3.4 & 21.21@xmath46 & 84.7 & 0.02 & 23.3 + 0104 + 2046 & 10:01:04.9 + 02:20:46 & 0.94@xmath59 & 3.54 & 0.11 & 3.8 & 24.27@xmath60 & -0.7 & 0.31 & 21.9 + 0104 + 2501 & 10:01:04.6 + 02:25:01 & 0.22@xmath39 & 3.26 & _ & _ & _ & _ & _ & 21.8 + 0105 + 4531 & 10:01:05.3 + 02:45:31 & 0.44@xmath46 & 1.00 & 0.47 & 3.4 & 20.38@xmath38 & 15.9 & 0.00 & 22.3 + 0107 + 0533 & 10:01:07.9 + 02:05:33 & 0.90@xmath58 & 1.71 & 0.24 & 3.2 & 21.83@xmath61 & -70.4 & 0.04 & 22.7 + 0108 + 5606 & 10:01:08.0 + 01:56:06 & 0.97@xmath44 & 1.25 & 0.20 & 1.8 & 22.81@xmath55 & -61.5 & 0.24 & 23.4 + 0120 + 4551&10:01:20.2 + 01:45:51 & 0.82@xmath58 & 1.28 & _ & _ & _ & _ & _ & 22.9 + 0148 + 2325 & 10:01:48.1 + 02:23:25 & 0.40@xmath46 & 1.43 & 1.27 & 2.9 & 18.81@xmath38 & -35.1 & 0.00 & 22.0 + 0208 + 1422 & 10:02:08.5 + 02:14:22 & 0.41@xmath46 & 1.59 & 0.45 & 7.7 & 20.06@xmath44 & -52.0 & 0.00 & 22.0 + 0220 + 2335 & 10:02:20.2 + 02:23:35 & 0.65@xmath37 & 1.04 & 0.60 & 1.7 & 20.99@xmath62 & 87.2 & 0.00 & 21.8 + 0221 + 3440 & 10:02:21.1 + 02:34:40 & 0.41@xmath37 & 1.57 & 1.06 & 5.5 & 19.35@xmath38 & -22.7 & 0.04 & 22.3 + 0236 + 4807 & 10:02:36.0 + 02:48:07 & 0.34@xmath58 & 1.27 & 0.68 & 1.8 & 20.15@xmath38 & 22.0 & 0.11 & 23.2 + 0248 + 1422 & 10:02:48.4 + 02:14:22 & 0.60@xmath46 & 1.38 & 0.22 & 5.5 & 21.72@xmath63 & 18.8 & 0.24 & 22.8 + 5748 + 5524 & 09:57:48.0 + 01:55:24 & 0.93@xmath44 & 2.21 & 0.33 & 1.4 & 22.15@xmath55 & 89.9 & 0.00 & 21.6 + 5750 + 5619 & 09:57:50.7 + 01:56:19 & 0.28@xmath46 & 0.73 & _ & _ & _ & _ & _ & 21.8 + 5752 + 2057 & 09:57:52.1 + 02:20:57 & 0.44@xmath46 & 0.98 & 0.51 & 1.0 & 20.27@xmath38 & 3.8 & 0.14 & 21.6 + 5754 + 5952 & 09:57:54.1 + 01:59:52 & 0.90@xmath39 & 1.33 & 0.22 & 6.9 & 22.23@xmath60 & -16.5 & 0.43 & 22.6 + 5805 + 0413 & 09:58:05.6 + 02:04:13 & 0.88@xmath46 & 1.61 & 0.12 & 4.8 & 23.00@xmath64 & -81.6 & 0.16 & 22.8 + 5806 + 5809 & 09:58:06.8 + 01:58:09 & 0.32@xmath46 & 1.50 & 0.43 & 2.9 & 19.14@xmath65 & -13.6 & 0.00 & 21.7 + 5821 + 4437@xmath40 & 09:58:21.4 + 01:44:37 & 0.67@xmath42 & 1.40 & 0.21 & 3.5 & 21.84@xmath66 & 00.0 & 0.28 & 22.9 + 5829 + 3734@xmath40 & 09:58:29.9 + 01:37:34 & 0.51@xmath42 & 2.97 & 0.46 & 1.6 & 19.95@xmath67 & -0.1 & 1.15 & 22.3 + 5831 + 4332@xmath40 & 09:58:31.0 + 01:43:32 & _ & 1.17 & _ & _ & _ & _ & _ & 21.0 + 5841 + 4646 & 09:58:41.4 + 02:46:46 & 0.36@xmath46 & 0.93 & 0.47 & 0.9 & 19.65@xmath38 & 21.1 & 0.12 & 21.0 + 5851 + 1813 & 09:58:51.9 + 02:18:13 & 0.29@xmath37&2.57 & _ & _ & _ & _ & _ & 21.9 + 5856 + 4755 & 09:58:56.1 + 02:47:55 & 0.32@xmath46 & 2.97 & 0.17 & 1.1 & 19.54@xmath38 & 48.8 & 0.39 & 22.9 + 5906 + 4524 & 09:59:06.5 + 02:45:24 & 0.37@xmath37&1.50 & _ & _ & _ & _ & _ & 21.6 + 5919 + 3853 & 09:59:19.4 + 01:38:53 & 0.40@xmath46 & 2.35 & 0.26 & 1.9 & 20.59@xmath68 & -15.7 & 0.08 & 22.7 + 5924 + 0852 & 09:59:24.7 + 02:08:52 & 0.93@xmath44 & 0.97 & 0.17 & 1.6 & 23.09@xmath69 & -9.8 & 0.01 & 22.7 + 5925 + 3039 & 09:59:25.8 + 02:30:39 & 0.91@xmath44 & 0.98 & 0.15 & 1.2 & 23.16@xmath70 & 89.9 & 0.38 & 23.2 + 5929 + 1352 & 09:59:29.9 + 02:13:52 & 0.36@xmath46 & 2.46 & _ & _ & & _ & _ & 22.0 + 5929 + 4553 & 09:59:29.0 + 01:45:53 & 0.54@xmath39 & 1.56 & 0.67 & 2.1 & 19.99@xmath38&-23.2 & 0.06 & 22.3 + 5931 + 0229 & 09:59:31.1 + 02:02:29 & 0.32@xmath46 & 1.48 & 0.81 & 1.9 & 19.20@xmath71&89.9 & 0.01 & 21.6 + 5932 + 1018 & 09:59:32.0 + 02:10:18 & 0.92@xmath72 & 0.68 & 0.40 & 1.2 & 21.92@xmath73&19.9 & 0.22 & 22.6 + 5936 + 3621 & 09:59:36.7 + 02:36:21 & 0.25@xmath37 & 2.40 & 0.71 & 0.9 & 19.14@xmath74 & -42.4 & 0.15 & 22.1 + 5939 + 3044 & 09:59:39.1 + 02:30:44 & 0.73@xmath37 & 0.70 & 0.58 & 1.7 & 20.08@xmath75&76.9 & 0.00 & 22.9 + 5940 + 3253 & 09:59:40.4 + 02:32:53 & 0.41@xmath37 & 1.53 & 0.51 & 1.4 & 19.45@xmath38 & 58.4 & 0.40 & 22.5 + 5940 + 5012 & 09:59:40.0 + 02:50:12&0.36@xmath46&1.97 & _ & _ & _ & _ & _ & 21.8 + 5942 + 2829 & 09:59:42.7 + 02:28:29 & 0.72@xmath46 & 1.39 & 0.31 & 1.6 & 21.88@xmath37 & 34.0 & 0.04 & 22.8 + 5943 + 2816 & 09:59:43.1 + 02:28:16 & 0.39@xmath42 & 1.64 & 0.69 & 1.3 & 19.43@xmath38 & 79.9 & 0.01 & 22.4 + 5951 + 1236 & 09:59:51.0 + 02:12:36 & 0.46@xmath39 & 1.53 & 1.21 & 2.2&18.66@xmath38 & 20.9 & 0.29 & 21.9 + 5959 + 0348 &
09:59:59.7 + 02:03:48 & 0.82@xmath58 & 1.20 & 0.20 & 3.5 & 21.62@xmath46 & -49.2 & 0.20 & 22.8 six of the lens candidates have radio counterparts in the 1.4ghz vla - cosmos data ( schinnerer et al .
the radio data cover the full cosmos field to a depth of @xmath010.5@xmath76jy@xmath77 rms in the center of the field .
all six sources lie above a 5@xmath78 level in total flux , ranging from 77@xmath76jy to 1.24mjy in total 1.4ghz flux density .
the distance between optical and radio position is in all cases less than 0.5 .
one of the lens candidates , cosmos 0013 + 2249 , which is also a radio emitter , shows associated x - ray point source emission from the xmm / newton - cosmos survey ( brusa et al .
it has 0.52kev x - ray emission of 3.74@xmath2210@xmath79 @xmath80 0.34@xmath2210@xmath81 erg@xmath5@xmath82s@xmath5 , corresponding to an x - ray luminosity of 1.21@xmath2210@xmath83 erg@xmath5s@xmath5 at the spectroscopic redshift of the lensing galaxy ( @xmath84 , trump et al .
this x - ray emission as well as the optical spectrum from magellan ( brusa et al .
2006 ) are consistent with the lensing galaxy being a radio galaxy with a central low - luminosity agn . in table
[ radio ] we list the radio counterpart properties and show vla overlays on the acs images in figure [ radiofig ] . [ cols= " < , < , < , < , < " , ]
in this section we compare the physical parameters of the foreground in the cosmos strong lens sample with independent galaxy - galaxy strong lensing catalogs from the slacs survey ( treu et al .
2005 , bolton et al .
2006 ) and the castles database ( muoz et al .
although the slacs and castles samples were selected very differently from the cosmos strong lensing sample , they also have _ hst _ imaging .
thus a rough comparison in terms of the galaxy lensing properties is warranted , at least for the sub - sample of best systems .
the effective radii of strong lenses in the slacs survey ( hashed histogram ) and the best systems from the cosmos survey ( histogram filled in grey).,width=264 ] photometric redshifts have been measured by mobasher et al . ( 2007 ) for most of the cosmos strong lensing galaxy candidates ( see table [ photom1 ] , and [ photom ] ) .
we first compare the redshift distribution of the lensing galaxy candidates found in the cosmos field ( 17 best candidates with a photometric redshift ) to the lensing galaxies in the castles database ( 90 strong galaxy - galaxy lenses , lens redshift known for 59 of them ) and in the slacs survey ( 19 strong galaxy - galaxy lenses ) ( fig .
[ histoslacscastles ] ) .
the ( photometric ) redshift distribution of the cosmos sample of strong lenses extends to significantly higher redshifts than the spectroscopic redshift distribution of galaxies in the slacs sample , which is intrinsically limited due to the sdss depth , but is similar to the spectroscopic redshift distribution of galaxies in the castles sample .
the effective radii and arc radius of the lensing galaxies ( in kpc ) are simply obtained by converting the effective radii in angular units to physical distances , via the photometric redshift of the lensing galaxy .
the slacs sample includes lenses with larger effective lensing radii than the cosmos sample ( see fig .
[ historeff ] ) .
this is probably an observational bias , as ( for practical reasons ) we inspect visually the lens candidates in a fixed 10@xmath2210 , box , and therefore miss any galaxies that would generate arcs at larger radii .
the absolute magnitudes of the slacs lensing galaxies are available in treu et al .
the mean value of their v - band magnitude is @xmath85 .
the absolute v - band magnitudes of the lensing galaxy candidates of our `` best systems '' sample have a mean value of @xmath86 .
thus the cosmos sample reaches slightly fainter magnitudes than the slacs sample , despite the fact that it extends to higher lens redshifts .
the velocity dispersions of the lensing galaxies of the cosmos sample are derived from simple mass models presented in sect . [ model ] .
they are displayed in table [ modeltable ] .
we have compared the velocity dispersions of the lensing galaxy candidates of the best systems to the velocity dispersions of the lensing galaxies in the slacs and in the castles samples ( also measured using isothermal mass potentials , fig .
[ histosig ] ) .
the lensing galaxies in the cosmos sample have a slightly higher velocity dispersion than the lensing galaxies in slacs and in castles .
the difference with the slacs sample can be due to the fact that we are probing in mean larger arc radius ( therefore larger einstein radius ) .
in addition , we have estimated the source redshift from the lensing galaxy photometric redshifts , that have an uncertainty .
this can lead in some cases to wrong estimations of the velocity dispersion of the main deflector .
in particular , we notice that two systems have a high velocity dispersion ( @xmath1 400 km s@xmath5 ) : cosmos 0211 + 1139 and cosmos 5914 + 1219 . if real , these values suggest that these galaxies are in the center of a group , and that we are measuring here both the galaxy mass and the group mass .
the velocity dispersions of strong lenses in the slacs survey ( hashed histogram ) and in the castles database ( solid line ) compared to the velocity dispersions of the best systems in the cosmos sample ( histogram filled in grey ) .
, width=264 ] the arc radius versus the einstein radius of the best systems .
the black stars show the ring like candidates .
the solid line traces r@xmath87=r@xmath88.,width=264 ] the einstein radius is directly derived from the velocity dispersion of the lensing galaxy , it is therefore another mass estimator for the lens potential .
it has the advantage of being directly comparable to the observable arc radius .
the value of the einstein radii of the lensing galaxies r@xmath87 are given in table [ modeltable ] .
for the best systems , the einstein radius is expected to be similar to the arc radius , while an equality between these two values is expected in case of perfect alignment between the observer , the source and the lens ( einstein ring ) . the plot in fig .
[ rarcrein ] shows that the two radii are indeed similar for most of the observed systems .
nevertheless we observe some exceptions : some systems have a large einstein radius in comparison to their arc radius : cosmos 0018 + 3845 ( @xmath89=0.40 , @xmath90=1.44 ) , 0047 + 5023 ( @xmath89=0.70 , @xmath90=1.29 ) and 5914 + 1219 ( @xmath89=1.86 , @xmath90=2.82 ) .
for the three systems , there is a strong external shear added to the sie .
this may indicates the presence of a massive structure in the direction to the source or in its vicinity that could have stretched the angular separation between the images and/or biased strongly the estimation of the velocity dispersion , and hence of the estimation of the einstein radius of the lensing galaxy .
in this paper we present 67 new strong galaxy - galaxy lensing candidates found by visual inspection of the cosmos field . from a catalog of 278,526 galaxies with photometric redshifts ( capak et al .
2007 ) , we have selected a sub - catalog of 9452 bright galaxies ( m@xmath91 - 20 mag ) , at intermediate redshift ( @xmath92 ) and spectrally early type .
the selection criteria for this sub - catalog were mainly motivated by the purpose of reducing the sample of galaxies to be visually inspected . from the visual inspection of acs / i@xmath93-band stamp images of 10@xmath2210 ,
we have built a sample of 337 possible lenses with sizes up to r@xmath945 .
this number is reduced to 67 strong galaxy - galaxy lensing candidates after a final cut based on an inspection of the pseudo - color images , and the fit and subtraction of the lensing galaxy luminosity profile .
there are 47 candidates with a single arclet and 20 candidates with multiple images of a background source or with a large curved arc , called the `` best systems '' in this paper .
we have produced a simple mass model ( sie+shear ) for each of the best candidates . in a few cases ,
we have also been able to predict the existence and direction of an external shear contribution .
comparing the velocity dispersions to those of lenses in the the slacs and castles samples , it is apparent that we are probing a slightly higher range of velocity dispersion .
nevertheless , massive objects in the immediate vicinity of the cosmos lenses , and the rough assumption concerning the source redshift , can explain this result .
we have inferred the einstein radius for the lensing galaxies considering that they are isothermal spheres , and compared these values with the arc radius .
the derived einstein radii agree well with the observed arc radii , making us confident that the sample consists mainly of genuine lens systems .
the `` visual inspection '' approach is long and tedious , and does not ensure a full understanding of the completeness of the strong lenses sample .
nevertheless , our meticulously - obtained large sample of strong lenses , even if incomplete , provides a valuable catalog to help develop , test and improve automatic strong lens - finding algorithms for similar data - sets .
such work is currently in progress ( cabanac et al .
2007 , marshall et al .
2007 , seidel & bartlemann 2007 ) .
also , in future work ( faure et al . 2007 ) the direct study of the spatial correlation between large - scale structure in the cosmos field ( scoville et al . 2007 , massey et al . 2007 ) and
the strong lenses discovered here will allow us to measure the impact of the total , projected mass distribution along the line of sight to a given strong lens , as recently investigated for three strong lenses in the aegis survey ( moustakas et al .
a spectroscopic follow - up of the lens candidates is in progress at the vlt using the fors1-multi - object spectroscopy capability .
this will allow us to measure precisely the redshift of both the lens and the source , and to explore the close neighborhood of the lens , thus allowing a definite measurement of the mass of each strong lensing galaxy .
finally , this study gives a lower limit on the number of strong lenses expected in future deep space surveys such as the proposed jdem / snap multicolor weak - lensing survey .
indeed , we expect to find at least 10 strong lenses per square degrees .
thus for a 10 000 square degree survey , we should discover about 100 000 strong lenses thus allowing a high precision statistical analysis of the mass properties of galaxies and its evolution over time .
the sample of strong lenses is available at this address : http://cosmosstronglensing.uni-hd.de/ the hst cosmos treasury program was supported through nasa grant hst - go-09822 .
we wish to thank tony roman , denise taylor , and david soderblom for their assistance in planning and scheduling of the extensive cosmos observations .
we gratefully acknowledge the contributions of the entire cosmos collaboration consisting of more than 70 scientists . more information on the cosmos survey is available at http://www.astro.caltech.edu/cosmos . based on observations with the nasa / esa _ hubble space telescope _
, obtained at the space telescope science institute , which is operated by aura inc , under nasa contract nas 5 - 26555 ; also based on data collected at : the subaru telescope , which is operated by the national astronomical observatory of japan ; the european southern observatory under large program 175.a-0839 , chile ; kitt peak national observatory , cerro tololo inter - american observatory , and the national optical astronomy observatory , which are operated by the association of universities for research in astronomy , inc .
( aura ) under cooperative agreement with the national science foundation ; and the canada - france - hawaii telescope operated by the national research council of canada , the centre national de la recherche scientifique de france and the university of hawaii .
it is a pleasure to acknowledge the excellent services provided by the nasa ipac / irsa staff ( anastasia laity , anastasia alexov , bruce berriman and john good ) in providing online archive and server capabilities for the cosmos datasets .
cf thanks the support from the european community s sixth framework marie curie research training network programme , contract no .
mrtn - ct-2004 - 505183 `` angles '' .
jpk thanks support from cnrs and caltech .
gc is partially supported by the cordis erg proposal n. 029159 .
we acknowledge the referee for their useful remarks and j. duke for correcting the draft .
boulade , o. , et al.2003 , , 4841 , 72 browne , i. w. a. , wilkinson , p. n. , patnaik , a. r. , & wrobel , j. m. 1998 , , 293 , 257 brusa , m. , et al.2006 , arxiv astrophysics e - prints , arxiv : astro - ph/0612358 burud , i. , et al.2000 , , 544 , 117 capak , p. , et al.2007 , arxiv e - prints , 704 , arxiv:0704.2430 chae , k .- h . ,
mao , s. , & kang , x. 2006 , , 373 , 1369 chae , k .- h .
2007 , , 658 , l71 cohen , a. s. , hewitt , j. n. , moore , c. b. , & haarsma , d. b. 2000 , , 545 , 578 dickinson , m. , & goods legacy team 2001 , bulletin of the american astronomical society , 33 , 820 estrada , j. , et al.2007 , , 660 , 1176 fassnacht , c. d. , xanthopoulos , e. , koopmans , l. v. e. , & rusin , d. 2002 , , 581 , 823 fassnacht , c. d. , moustakas , l. a. , casertano , s. , ferguson , h. c. , lucas , r. a. , & park , y. 2004 , , 600 , l155 faure c. et al .
2007 , letter to apj , in prep .
fohlmeister , j. , et al .
2007 , , 662 , 62 & ponman , t. j. 1999 , , 309 , 623 fukugita , m. , & turner , e. l. 1991 , , 253 , 99 gladders , m. d. , yee , h. k. c. , mccarthy , p. j. , barrientos , l. f. , hoekstra , h. , hall , p. b. , & ellingson , e. 2002 , bulletin of the american astronomical society , 34 , 1209 griffiths , r. e. , et al .
1994 , bulletin of the american astronomical society , 26 , 877 guzik , j. , & seljak , u. 2002 , , 335 , 311 hilbert , s. , white , s. d. m. , hartlap , j. , & schneider , p. 2007
, arxiv astrophysics e - prints , arxiv : astro - ph/0703803 hjorth , j. , et al.2002 , , 572 , l11 ilbert , o. , et al.2006 , , 457 , 841 jackson , n. , et al.1995 , , 274 , l25 jakobsson , p. , hjorth , j. , burud , i. , letawe , g. , lidman , c. , & courbin , f. 2005 , , 431 , 103 kneib , j .- p . , ellis , r. s. , smail , i. , couch , w. j. , & sharples , r. m. 1996 , , 471 , 643 kochanek , c. s. 1996 , astrophysical applications of gravitational lensing , 173 , 177 kochanek , c. s. 2002 , the shapes of galaxies and their dark halos , proceedings of the yale cosmology workshop the shapes of galaxies and their dark matter halos
, new haven , connecticut , usa , 28 - 30 may 2001 . edited by priyamvada natarajan .
singapore : world scientific , 2002 , isbn 9810248482 , p.62 , 62 kochanek , c. s. , morgan , n. d. , falco , e. e. , mcleod , b. a. , winn , j. n. , dembicky , j. , & ketzeback , b. 2006 , , 640 , 47 koekemoer , a. m. , et al .
2007 , , 172 , 196 leauthaud , a. , et al .
2007 , arxiv astrophysics e - prints , arxiv : astro - ph/0702359 lilly , s. j. , et al.2006 , arxiv astrophysics e - prints , arxiv : astro - ph/0612291 2007 , , in press .
mandelbaum , r. , seljak , u. , kauffmann , g. , hirata , c. m. , & brinkmann , j. 2006 , , 368 , 715 marleau , f. r. , & simard , l. 1998 , , 507 , 585 marshall , p. j. , moustakas , l. a. , hogg , d. w. , bradac , m. , fassnacht , c. d. , & blandford , r. d. 2005 , bulletin of the american astronomical society , 37 , 1498 marshall , p. et al .
2007 , in prep .
massey , r. et al .
2007 , nature 445 , 286 .
miralda - escude , j. , & lehar , j. 1992 , , 259 , 31p miyazaki , s. , et al.2002 , , 54 , 833 mobasher , b. 2007 , , astro - ph/00612344v1 moeller , o. , kitzbichler , m. , & natarajan , p. 2006
, arxiv astrophysics e - prints , arxiv : astro - ph/0607032 moustakas l.a . , marshall p.j . , newman j.a et al .
2006 , apj letters , submitted .
part of the aegis apjl special issue .
4 , arxiv : astro - ph/0607239 muoz , j. a. , et al .
1998 , , 263 , 51 myers , s. t. , et al.1995 , , 447 , l5 ofek , e. o. , & maoz , d.2003 , , 594 , 101 | we present the first catalog of 67 strong galaxy - galaxy lens candidates discovered in the 1.64 square degree _
hubble space telescope _
cosmos survey .
twenty of these systems display multiple images or strongly curved large arcs .
our initial search is performed by visual inspection of the data and is restricted , for practical considerations , to massive early - type lens galaxies with arcs found at radii smaller than @xmath05 .
simple mass models are constructed for the best lens candidates and our results are compared to the strong lensing catalogs of the slacs survey and the castles database .
these new strong galaxy - galaxy lensing systems constitute a valuable sample to study the mass distribution of early - type galaxies and their associated dark matter halos .
we further expect this sample to play an important role in the testing of software algorithms designed to automatically search for strong gravitational lenses . from our analysis a robust lower limit is derived for the expected occurrence of strong galaxy - galaxy systems in current and future space - based wide - field imaging surveys .
we expect that such surveys should uncover a large number of strong lensing systems ( more than 10 systems per square degree ) , which will allow for a detailed statistical analysis of galaxy properties and will likely lead to constraints on models of gravitational structure formation and cosmology .
the sample of strong lenses is available at this address : http://cosmosstronglensing.uni-hd.de/ |
light echoes refer to light emitted by a bright source , such as a supernova ( sn ) , which is scattered by dust located off the direct line - of - sight .
the additional path length from the sn to the dust cloud and then to the earth induces the time delay of the echo . depending on the system considered and the brightness of the source
, light echoes can be seen several hundreds of light years away from the source , and subsequently several hundred years after the initial sn outburst observation . infrared ( ir ) echoes differ from optical light echoes in that the sn light is not scattered , but rather absorbed by dust , which will then re - radiate it at ir wavelengths .
@xcite considered this effect as the most plausible explanation for the ir excess seen in sn 1979c , a point which was later on confirmed by @xcite .
ir echoes spanning over an area of several square - degrees on - sky have been discovered around the youngest known galactic core - collapse sn remnant cassiopeia a ( cas a ) @xcite .
@xcite used their spectra to deduce that these must have been created by a uv flash .
the same echoes also enabled @xcite to reconstruct a 3d map of the ism surrounding cas a , a concept that was previously used with sn 1987a by @xcite .
the cas a supernova remnant is believed to be the remains of a stripped - envelope type iib core - collapse from a red supergiant progenitor with an initial mass in the range of 15 - 25 m@xmath3 which might have lost much of its hydrogen envelope due to a binary interaction @xcite , and is located at r.a .
: 23@[email protected]@xmath6 ; dec : + [email protected] " ( j2000 ) @xcite .
studies of the proper motion of ejecta with the _ hubble space telescope _ suggest that the explosion occurred in 1681@xmath819 @xcite .
sir john flamsteed ( 1646 - 1719 ) might have visually observed the supernova on 1680 august 16 , but it is not clear whether his mysterious _
3 cassiopeia _ observation was actually cas a or not @xcite . along with the infrared echoes
, several optical light echoes have been discovered around cas a @xcite .
@xcite showed that the spectrum of one of them is strikingly similar to the spectrum of sn 1993j , which enabled the precise spectroscopic classification that cas a was a type iib sn .
the knowledge of the exact type of the cas a supernova will play a central role in our analysis .
sn 1993j exploded in m81 in late march 1993 .
it has been extremely well monitored across the entire spectrum from its very first moments @xcite , including in the uv ( e.g. * ? ? ?
* ) , the wavelength range at which the irradiation spectrum has critical consequences on interstellar dust .
such an amount of data makes it a target of choice for supernova simulations , that are able to simulate the observed spectrum already from its early phases @xcite .
sn 1993j experienced an initial , short - lived ( few hours ) intense uv burst , later on followed by a longer ( several days ) less intense optical component @xcite .
hence , it is a very fortunate situation for our study that cas a was of the same type and similar to sn 1993j - one of the brightest sne in the northern sky for the last century , and for which a nearly unparalleled amount of data ( including uv ) exists .
infrared echoes can be used to study the structure and composition of the thermally radiating dust , as was suggested by @xcite for circumstellar dust shells .
infrared echoes especially allow to study several dust processing mechanisms , such as photodissociation , dehydrogenation and ionization , in - situ . due to
the fact that more than 320 years have passed since cas a exploded , the dust giving rise to the infrared echoes observed today is at least 160 light years away from cas a. these echoes therefore most likely originate from interstellar rather than circumstellar dust .
such an analysis of infrared echoes is similar to what @xcite suggested for optical light echoes , but using their infrared counterpart .
as the sn light is absorbed and re - radiated at longer wavelengths , the resulting ir spectrum is a complex signature of both the sn light and the dust composition and characteristics .
using optical light echoes to identify the sn type , one can create a template light curve . feeding this template in a suitable dust modelling program
, one can simulate the ir echo spectrum - and use the resulting simulations to understand how ism dust is behaving when illuminated by the sn light .
one of the main advantages of this method is that ir light echoes potentially give us access to the consequences of _ pristine _ interstellar material subject to a strong and short radiation burst . as such
, assuming an undisturbed initial ism , the dust signature detected in the ir light echo will be the consequence of the sn burst only , as no other mechanism , such as dust formation for example , can be acting on the very short timescale of a sn burst . in this article
, we explore the potential of this approach by analyzing an ir echo spectrum recorded with the spitzer space telescope around cas a. we describe our observations and the resulting recorded ir echo spectrum in sec . [ sec : obs ] .
we discuss how we simulate a light echo in sec .
[ sec : simu ] , introduce our dust model in sect .
[ sec : simu_dust ] , and present our cas a sn template burst spectra in sec . [ sec : snbursts ] .
our results are detailed in sec .
[ sec : results ] . in sec .
[ sec : discussion ] , we discuss how pahs destruction ( sec .
[ sec : depletion ] ) , ionization ( sec .
[ sec : ionization ] ) and dehydrogenation ( sec .
[ sec : dehydro ] ) can improve our fit to the recorded ir echo spectrum .
we compare our results with the work of @xcite in sec .
[ sec : dwek ] .
finally , we explore the validity of our dust destruction hypothesis in sec .
[ sec : uv - link ] , and present our conclusions in sec .
[ sec : conclusion ] .
the observations used in this study were obtained with instruments aboard the spitzer space telescope @xcite .
after the first discovery of infrared echoes around cas a by @xcite , regular monitoring observations were carried out with the multiband imaging photometer for spitzer ( mips ; @xcite ) ( pids : 233 , 20381 , 30571 ; pis : g.rieke , o.krause ) .
a roughly 3 @xmath9 3 square degrees field was observed every year and regions with especially interesting activity were observed at an interval of about half a year .
mips observations were reduced using the mips data analysis tool ( dat ; @xcite ) , starting at the raw data product level . using those monitoring observations ,
a particularly bright source was detected at the location r.a .
: 23@xmath421@xmath540@xmath6 ; dec : + 59@xmath7 34 25 .
[ fig : ma1 ] shows its evolution at 24@xmath0 m over a time period of 3 years .
it lies at an angular distance of 2857 arcsec away from the optical expansion center of cas a. although the source is already visible in the first epoch of observations , it changes its morphology , brightness and position drastically within just a few months time . in early 2006
, it appears that most emission comes from three point like sources , suggesting that the real echoing structure is still unresolved with spitzer .
the emission diminishes very quickly between february and october 2006 indicating that the sn light beam swept past a probably very well defined end of the echo region . on 2006
january 29 , a low - resolution spectrum of this bright echoing source was obtained with the infrared spectrograph ( irs ; @xcite ) as part of program ( pid ) 20381 ( p.i .
: o. krause ) .
a 24 @xmath0 m aerial image of the echoing source is shown in fig .
[ fig : ma3 ] , where we have overlaid the irs red peak - up insert ( rectangle ) as well as the spectroscopic slits position ; 1@xmath9 long - low ( ll ) and 3@xmath9 short - low ( sl ) .
the red peak - up functionality of irs was used to achieve the best possible centering of the slits on this fast moving echo - feature .
in addition to space - based infrared observations , we also obtained ground based near - infrared ( @xmath10=2.2@xmath0 m ) ks - band images for this echo region using omega 2000 on the calar alto ( spain ) 3.5 m telescope in december 2005 , which is shown in the bottom of fig .
[ fig : ma3 ] for comparison .
these images show the similar overall structure of the region , but reveal very small scale filamentary substructure of the echo region .
a full analysis of these observations is to be presented in besel et al .
( in preparation ) . it should be noted here that the ks - band as well as the j - band and h - band images ( which have similar morphology , and are not shown here ) show the scattered light echo , not the ir radiating dust echo .
the s18.7.0 version of the basic calibrated data ( bcd ) products from the spitzer science center ( ssc ) pipeline were fed into smart @xcite for final reduction .
depending on the used module , total exposure times ranged between 120s and 300s .
one spectra was extracted for each of the two brightest point - like sources .
they were averaged separately and combined later .
a background was estimated based on the non - echo - contaminated parts of either the off - source order or the individual slits ( contamination is easily evident on the mips 24 @xmath0 m image for one end of the ll slit ) .
as the sl and ll modules cover slightly different special regions of the echoing source the resulting spectra were scaled to match each other s continuum levels .
absolute calibration was verified by comparing the integrated irs flux and the measured 24 @xmath0 m photometry .
m aerial image of the echoing source with irs red - peakup insert ( dashed rectangle ) and footprints of the long - low ( ll , in green in the online version ) and short - low ( sl , in blue in the online version ) spectroscopic slits positions .
bottom : calar alto omega 2000 ks - band image of the same area .
visible are tiny structures within the infrared emitting area . ]
[ fig : ma4 ] shows the final combined extracted spectrum of this ir echo .
also plotted are mips 24 @xmath0 m and 70 @xmath0 m photometric measurements within a 22.2 arcsec aperture ( white squares ) .
note that their error bars are of the order of the symbol size .
black rombuses at 5.7 @xmath0 m , 7.9 @xmath0 m and 24 @xmath0 m correspond to the spitzer synthetic photometric points of the echo spectrum , reconstructed using the recipe 10 of the spitzer data analysis cookbook . as the irs slits
do not cover the entire echo region , the spectrum and synthetic photometric points were scaled to match the 24 @xmath0 m photometry of the entire region .
m and 70 @xmath0 m photometric measurements of the same echo region .
the spectrum and synthetic photometric points have been scaled to match the 24 @xmath0 m mips photometry of the entire region . ]
the spectrum and photometric points presented in fig .
[ fig : ma4 ] are the result of the interaction of the cas a sn burst and a region of the ism located away from the sn center . in that respect ,
the data potentially contain information on both the sn explosion and the echoing dust . a simple visual inspection of the spectrum already reveals a very weak 11 @xmath0 m complex of pah features as compared to the strength of the 7 @xmath0 m pah features , a point that will be discussed further in sec . [ sec : dehydro ] . here
, we adopt the following strategy to study and extract the underlaying information on the echoing dust contained in the measured echo : we implement a dust modelling program ( see sect .
[ sec : simu_dust ] ) to simulate the ir response of ism dust when heated by radiation with a certain spectral energy distribution ( sed ) . to obtain the dust thermal emission response
, its distribution along the line of sight has to be convolved with the corresponding irradiation spectrum from the supernova .
we assume that the effective radiation field can be decomposed by ( i ) the uv dominated sed during the first hours after the shock outbreak and ( ii ) the optically dominated sed at the epoch near maximum light ( see sect .
[ sec : snbursts ] ) .
we adopt those two template seds to account for the convolution of the input radiation spectrum with the echo region on length scales of at least a few light days . in fig .
[ fig : ma2 ] , we present a geometrical sketch of the echo principle . the position of the echo region is located , at any given time , on a ellipsoid with cas a and the earth at the foci @xcite .
the distance @xmath11 between the echo region and the center of cas a is : @xmath12 where @xmath13 is the time delay between the light of cas a s explosion reaching earth and the observation of the light echo , @xmath14 the distance to cas a and @xmath15 the on - sky angular separation between cas a and the echo region .
using a distance to cas a of 3.4 kpc @xcite and an explosion date of ad 1681 @xcite , the absolute distance of the echo region to the supernova remnant is @xmath16 light years .
a schematic of the echo region itself is shown in fig .
[ fig : drawing ] . because the uv burst and optical phase do not occur simultaneously
, they will heat a given area of the echo region sequentially . in fig .
[ fig : drawing ] , area ( a ) corresponds to the echo region not seen excited at the moment of the observation , while areas ( b ) and ( c ) denote respectively the schematic location of the uv burst and optical light curve excited slabs , moving away from cas a. it is important to note here that if we _ see _ those excited slabs simultaneously , they _ are not _ excited simultaneously , because of the finite light speed . the uv - burst excited slab ( a.k.a .
uv - heated slab ) is located farther away from cas a and the earth , and must therefore be excited earlier than the optically - excited slab ( a.k.a optically - heated slab ) to be seen at the same time on earth . in situ , the two seds will be seen by a dust grain with the same time interval as when they occurred after the shock breakout . as seen from the earth , the excited slabs are encompassed in between ellipsoids of different radii @xmath17 , @xmath18 , @xmath19 and @xmath20 defined by eq .
[ eq : r ] , with @xmath21 , @xmath22 , @xmath23 and @xmath24 the delay times between the observation and the shock breakout , the end of the uv burst , the start of the optical phase and the end of the optical phase , respectively .
let us define @xmath25 and @xmath26 the on - sky dimensions of the echo , in the direction of and perpendicular to cas a , respectively . because the uv - heated and optically - heated slabs thickness will only be of a few light days , the excited volume of the echo region _
visible at one given moment in time _ is much smaller than the total volume of the echo region . for simplification , because we do not resolve individual ism structures and because the echo region is small compared to the ellipsoid and its local curvature depicted in fig .
[ fig : ma2 ] , let us assume the excited slabs to be parallelepipeds , of which the sections are depicted by the light - grey ( b ) and dark - grey ( c ) regions in fig .
[ fig : drawing ] ( in orange and blue in the online version ) .
their height ( perpendicular to the drawing plane ) is @xmath26 .
their respective thickness @xmath27 and @xmath28 along the line - of - sight are : @xmath29 where , from fig .
[ fig : ma2 ] and the law of cosines : @xmath30 for the echo considered here and presented in sec .
[ sec : obs ] , @xmath31 pc .
we can now define an upper limit to the _ visible _ volume of the echo region , @xmath32 , as seen from the earth at any given time : @xmath33 where @xmath34 and @xmath35 are the respective areas of the regions ( b ) and ( c ) in fig .
[ fig : drawing ] . the shape of a fictional echo region ( dotted surface ) depicted in fig .
[ fig : drawing ] illustrate why @xmath32 is an upper limit on the visible volume of the echo region , as the dust does not , in principle , have to fill completely the areas ( b ) and ( c ) .
we delay the numerical computation of @xmath32 until sec .
[ sec : fit ] , after introducing our adopted template uv burst and optical light curve in sec .
[ sec : snbursts ] .
simulations provide an efficient tool to study the theoretical behaviour of interstellar dust in an attempt to unveil its composition and detailed characteristics .
various models , such as the silicate core - carbonaceous @xcite , the composite @xcite or the silicate - graphite - polycyclic aromatic hydrocarbons ( pahs ) model @xcite have been proposed to reproduce the behaviour of ism dust . to study the recorded ir echo spectrum shown in fig .
[ fig : ma4 ] , we have implemented our own modelisation program , strongly based on the dust model of @xcite .
our dust mix contains pahs , quantum - heated carbon and silicate grains , as well as thermal graphite and silicate grains .
similarly to @xcite , the pahs and quantum - heated carbon grains are mixed and form what is referred to as the quantum - heated carbonaceous dust .
the @xcite model is a well established and observationally tested model for the average dust composition of the general milky way interstellar medium . since the echo region considered in this study is located in the general ism at a distance of @xmath1199 light years from the supernova , using this model appears very appropriate .
we simulate particles from @xmath13.5 up to @xmath16000 , and have a transition from quantum - heated to thermal grains at @xmath1250 .
the dust mix is following the formula of @xcite multiplied by 0.92 , as suggested by @xcite , from whom we have implemented the updated drude profiles for the pahs opacities .
our simulation yet differs from the @xcite model in a few aspects .
we use enthalpies from @xcite , which are based on a formula by @xcite .
our silicate opacities are calculated via the mie theory using scattering and absorption cross section from @xcite .
finally , we also use enthalpies of graphite for the silicate grains for simplification .
quantum - heated grains are subject to stochastic heating by the incoming photons .
they do not behave as constant - temperature grains , but experience strong temperature variations , and a careful approach is required to compute their emission .
more precisely , their emissivity is obtained by calculating the temperature probability distribution for a certain grain size , and we refer the reader to the work of @xcite for detailed explanations regarding the computation of the emissivity of small grains
. large thermal grains , on the other hand , will experience much smaller temperature excursions , and can be approximated as being in thermal equilibrium . under this assumption ,
their emissivity is obtained by comparing the incoming radiation to the re - radiated energy @xcite : @xmath36 with @xmath37 the incoming radiation flux at wavelength @xmath10 , @xmath38 the equilibrium mean temperature of the grain , @xmath39 the planck function and @xmath40 the absorption cross - section of a grain of radius @xmath41 at the wavelength @xmath10 . in fig .
[ fig : li01-hgl ] , we present our simulation ( full line , in green in the online version ) of the ism dust ir emissivity when heated by the standard interstellar radiation field ( isrf ) of @xcite ( see fig . [
fig : bursts ] ) .
the contribution of each individual dust components is plotted in dashed ( quantum - heated grains ) and dotted ( thermal grains ) lines ; ( a ) - quantum - heated carbonaceous dust , ( b ) - quantum - heated silicate dust , ( c ) - thermal graphite grains , ( d ) - thermal silicates grains .
the simulation code returns the emissivity of the dust mix per h nucleon , @xmath42 , in units of erg s@xmath43 sr@xmath43 h@xmath43 hz@xmath43 , and in fig .
[ fig : li01-hgl ] , we plot the spectrum in the form of @xmath44 , in units of erg s@xmath43 sr@xmath43 h@xmath43 . we validate our code output against the results of the @xcite code , and the comparison is made in fig .
[ fig : li01-chi ] .
our hgl simulation ( full line , in green in the online version ) reproduces closely the one from @xcite ( dotted line ) , with the exception of the pah features , that have been updated to reproduce the simulations of @xcite .
note that the spectrum of @xcite has been scaled by 0.92 to allow for comparison .
times the isrf ( dashed line , in blue in the online version ) . the latter has been scaled down by 10@xmath45 .
the corresponding simulations of @xcite are shown in dotted lines . ]
the strategy to simulate the ir echo spectrum presented in sec .
[ sec : obs ] is to use our code with a template input spectrum , replacing the isrf of @xcite by a theoretical sn burst spectrum . in order to validate that our code is stable with a more intense heating radiation field ,
we simulate ism dust heated by a radiation field equal to 10@xmath46 times the isrf , and compare our result with the similar simulation of @xcite ( see fig . 13 in their article ) .
the comparison is shown in fig .
[ fig : li01-chi ] ( dashed line , in blue in the online version ) .
it is clear that we reproduce very closely the @xcite simulations for stronger radiation fields , with exception to , once again , the intrinsic differences due to the update of the drude profiles for the pahs , which match the simulations of @xcite . we conclude that our simulation code is stable to intensity variations of the incoming radiation field .
simulating the spectrum of an infrared echo requires the knowledge of the exact radiation sed heating the dust .
generally , this is not known with precision , as the sn light swept past earth many years before a light echo observation .
fortunately , as we mentioned in sect . [
sec : intro ] , sn 1993j has been found to be a very good template for cas a , and has been the subject of many observations and simulations over a large range of wavelengths .
hence , following the characteristics of sn 1993j , we make the following simplification mentioned previously : we consider the cas a sn explosion to be made of two distinct phases - a uv burst and an optical light phase .
each phase consists of a single sed , which we take to be representative of the actual spectrum variations over the phase duration . to create the template uv burst spectrum
, we use the simulations of @xcite . in this model ,
the uv burst is close from its maximum luminosity during a few hours .
the total uv burst lasts a few days , with its luminosity reaching a minimum after about @xmath110 days .
energetic uv photons have a strong effect on dust and especially on pahs , as we will discuss later on .
it is therefore critical to have a representative sed that contains those energetic uv photons .
the peak of the uv burst is however fairly brief , and one has to pay attention not to over - estimate the amount of very energetic photons . with these considerations in mind
, we adopt our template cas a uv - burst to be a black - body spectrum , with a temperature of @xmath47 k. in an attempt to provide a slightly more realistic light curve , we subsequently adapt the shape of the black - body curve to that of the .278 curve of @xcite in the 0.001 - 1 @xmath0 m range .
we assume a burst duration of 2.5 days , and a total luminosity of l@xmath48 erg s@xmath43 .
for comparison , the luminosity of the black body curve prior to the shape modification is l@xmath49=5.9@xmath910@xmath50 erg s@xmath43 . by modifying the burst shape rather than increasing the burst temperature ,
we ensure that we do not overestimate the number of very energetic photons , while still taking into account the fact that a significant amount of energy is radiated away during a short period of time . to account for the smaller , but yet non - negligible , burst luminosity beyond the 2-days mark , we adopt a total burst duration of 2.5 days , shorter than the @xmath51 days mark , to avoid an obvious over - estimate .
we concede that there exist many other ways to obtain a cas a template uv burst from the data of sn 1993j , but any solution , including our adopted template , would involve quite a number of assumptions .
we emphasize here that we do not claim to have implemented neither the best , nor the most accurate reproduction of cas a s uv - burst . but rather made an good educated guess . in section
[ sec : dwek ] , we will compare our uv burst template with previous estimates by @xcite and show that both uv burst estimates are consistent .
we use the work of @xcite to create the template optical light curve .
they find the optical phase to last some 20 - 30 days , with a luminosity depending on the color excess e(b - v ) on the line of sight .
they provide two estimates , a low one with e(b - v ) = 0.08 , and a moderate one , with e(b - v)=0.32 . based on the work of @xcite , that finds e(b - v)=0.25 @xmath8 0.05 , and using the discussion on the absorption on the line - of - sight towards sn 1993j of @xcite , that adopt in their work a value of e(b - v ) = 0.19 , we assume our optical light curve to have the characteristics of the @xcite estimates ( see table 15 in their paper ) adapted for a color excess e(b - v)@xmath520.20 .
our optical light curve is a black - body spectrum , with a temperature @xmath53 k ( between the @xmath54 and @xmath55 estimates of @xcite ) , a luminosity of @xmath56 erg s@xmath43 ( between the @xmath57 and @xmath58 estimates of @xcite ) and a duration of 25 days ( see fig .
10 in @xcite ) .
we stress again that we consider this optical light curve as a good educated guess rather than the perfect estimate .
the resulting uv burst and optical light curve fluxes @xmath59 shining on the dust in units of erg s@xmath43 @xmath60 are shown in fig .
[ fig : bursts ] .
the final uv burst spectrum used in our simulations , of which the shape has been updated based on the work of @xcite , is labelled uv - b ( dashed line ) , while the black body uv burst with t=1.5@xmath910@xmath61 k is labelled uv ( dot - dashed line ) and is plotted for comparison . using the idl ismtau.pro routine , we have taken into account some foreground absorption for the uv burst assuming an atomic h column density @xmath62 of @xmath63 @xmath60 , an h ii / h i ratio of 0.26 , an he i and he ii column densities of @xmath64 , and used the photoionization cross - section of @xcite .
note that we have not taken into account any intrinsic extinction by the echo region itself .
the various burst characteristics are summarized in table [ table : bursts ] . .
the uv light curve ( dot - dashed ) is a black body with t=1.5@xmath910@xmath61 k and is shown for comparison , along with the isrf curve ( full line ) from @xcite . ]
c c c c c c & & & & + burst type & t@xmath65 [ k ] & luminosity [ erg s@xmath43 ] & absorption & duration [ days ] & total energy [ erg ] + optical&@xmath67&@xmath68&no & 25 & @xmath69 + uv - b&-&@xmath70&yes & 2.5 & @xmath71 + uv&@xmath72&@xmath73&yes & - & - +
with the adopted uv burst and optical light curve characteristics defined in the previous section , we can now get the estimates for the size of the echo region . specifically , with @xmath74 year , @xmath75 days , @xmath76 days , @xmath77 days , @xmath78 arcsec and @xmath79 light years ( 22 " on - sky aperture , see sec .
[ sec : obs ] ) , we get , using eq .
[ eq:2 ] to eq .
[ eq : volume ] : @xmath80 because we do not resolve the echo region , we make the assumption that @xmath81 the recorded echo flux is equal to the sum of the uv - heated slab flux and the optically - heated slab flux @xmath82 and @xmath83 respectively : @xmath84 furthermore , @xmath82 and @xmath85 relate to the uv - heated and optically - heated dust emissivities @xmath86 and @xmath87 in the following way : @xmath88 where @xmath89 and @xmath90 are the echoing dust mass , in g , heated by the uv burst and the optical light curve respectively , @xmath91 is the dust mass per h nucleon in g h@xmath43 , and @xmath92 is the distance from the earth to the echo region in cm .
we assume @xmath93 g h@xmath43 @xcite . from our assumptions in sec .
[ sec : simu ] ( see also fig . [
fig : drawing ] ) , we can express @xmath89 and @xmath90 as a function of the cloud density @xmath94 , and the volume of the excited slabs ( see eq .
[ eq:2 ] and [ eq:3 ] ) : @xmath95 hence , using eq .
[ eq : flux ] to [ eq : dopt ] , it is possible to express the echo spectrum @xmath81 as a function of the total mass of echoing dust @xmath96 in the following way : @xmath97\times\frac{m_{\mathrm{dust}}}{m_{\mathrm{h , dust}}\cdot x^2}\ ] ] where @xmath98 is in g and @xmath99 is the ratio of the mass of the uv - slab dust to the total echoing dust mass , which , using eq .
[ eq : dopt0 ] and [ eq : dopt ] , can be written as : @xmath100 eq . [ eq : fit ] and [ eq : zeta ] reflect the fact that the uv - heated and optically - heated dust emissivities _ are mixed _ with a 9%/91% ratio ( similar to the uv vs optical phase duration ratio ) . for simplification and clarity , we will refer to this ratio as our _ 9%/91% e - mix _
( for emissivity mix ) in the rest of this article .
we emphasize however that this is not equal to the _ luminosity _ ratio of the uv - heated to optically - heated component of the simulated echo spectrum , which we will show in the subsequent sections is of the order of @xmath180%/20% in the 5 - 38 @xmath0 m range .
we also note that this 9%/91% ratio is a direct consequence of our assumptions to have a symmetric echo region .
rapid size variations would lead to an alteration of the mixing ratio , and potentially rapid variability of the light echo . with only one echo spectrum , we are not able to address these questions , which will require the monitoring of one given echo region over a relatively short period of time . ultimately , in eq .
[ eq : fit ] , @xmath98 is the only undefined parameter , and we use it as the fitting parameter between the recorded spectrum and our simulated one ( i.e. we use it to move the simulated spectrum vertically to match the intensity of the recorded echo spectrum ) . the resulting simulated spectrum ( labelled sim1 ) , using standard ism dust ( described in sec . [ sec : simu_dust ] ) and a 9%/91% e - mix ( described in sec . [ sec : fit ] ) , is shown in fig .
[ fig : fit-1 ] ( top panel ) . fitting the intensity of the simulated spectrum to that of the recorded one ( thick line ) using eq .
[ eq : fit ] leads to an echoing dust mass of @xmath101 g , or
@xmath102 m@xmath103 .
the dashed and dotted lines correspond to the individual contributions from uv - heated and optically - heated dust , respectively . in the 5 - 38 @xmath0 m range , the uv component of the simulated echo spectrum accounts for 80.0% of the total luminosity of the simulated echo .
quite interestingly , the optical component influence becomes more important than that of the uv one beyond 50 @xmath0 m .
the presence of an optical component especially appears essential in order to fit the 70@xmath0 m photometric measurement .
it will therefore be of critical importance to obtain such photometric measurements beyond 50@xmath0 m in future studies of infrared echoes if one intends to better characterize the optically - heated dust component of such echoes . given the total echoing dust volume found in eq .
[ eq : volume ] , the resulting echo region density is @xmath104 h @xmath2 or @xmath105 g @xmath2 , with a column density of @xmath106 @xmath60 . because the volume of echoing dust calculated in sec .
[ sec : fit ] is an upper limit of the actual volume of echoing dust ( see sec .
[ sec : simu ] ) , the value of @xmath107 can be considered as a lower limit on the actual echoing dust density . , and associated synthetic photometric points ( black rhombuses ) .
the white squares correspond to mips 24 @xmath0 m and 70 @xmath0 m photometric measurements .
the white rhombuses represent the spitzer synthetic photometric measurements for our simulation .
the dashed and dotted lines correspond to the individual emission from uv - heated and optically heated dust , respectively .
bottom : fit residuals relative to the recorded echo spectrum . ]
the bottom panel shows the fit residual ( relative to the recorded echo ) .
the general trend of the recorded spectrum is well reproduced above @xmath115 @xmath0 m , as we match both the recorded echo spectrum as well as the 70 @xmath0 m photometric point .
yet , there are some discrepancies in the pahs features , where our simulated spectrum is significantly too strong .
we especially fail , at this point , to account for the peculiarly small 11 @xmath0 m complex .
the discrepancies around the pahs features is reflected in the difference between the recorded and simulated photometric points shown in the bottom panel of fig .
[ fig : fit-1 ] , with a @xmath1250% overestimate of the flux in those regions . in the following section ,
we modify our assumptions on the echoing dust to improve the fit below 15 @xmath0 m .
as we base our simulations on the dust model of @xcite , it is clear that resulting dust masses and densities of the echoing region are a direct consequence of this model .
alternate dust models , such as the ones mention in sec .
[ sec : simu_dust ] , are based on different total abundances ( which can vary along different line - of - sights ) , dust characteristics , mixtures and dust - to - gas ratios . because the aim of this paper is _
not _ to compare the _ goodness _ of different dust models and their ability to reproduce our recorded echo spectrum , but rather to show the potential of performing such a fit and its subsequent scientific output , we deliberately choose to alter the original dust model of @xcite _ only _ to explore the potential of such modifications to improve our fit to the recorded echo spectrum , but _ not _ in an attempt to solely improve their dust model .
performing a similar analysis to the one presented in this article , but with a different dust model , has a great potential but is outside the scope of this paper .
pahs can be strongly affected by uv radiation , which can cause photoionization , dehydrogenation or photodissociation @xcite .
if those various processes are not directly implemented in our simulation code , we can nevertheless ( approximatively ) take them into account by varying the amount of various dust components in the code , so as to reproduce their effect on the dust composition . specifically , we modify the equation responsible for the mix in between pahs and quantum - heated carbon grains as follows ( see eq . 2 in @xcite ) :
@xmath108 where @xmath109 is the absorption cross section of the carbonaceous dust , @xmath110 and @xmath111 the absorption cross sections of the pahs and quantum - heated graphite grains respectively , @xmath112 is the mixing ratio as defined by @xcite , and @xmath113 is a function which parametrizes the pahs destruction ratio , and that we will define below .
it should be emphasized that this is not equivalent to modifying the grain sizes distribution function @xmath114 defined by @xcite which would then impact the carbonaceous dust as a whole .
pahs are much more likely to be processed by uv radiation than quantum - heated graphite grains @xcite , and we do not see any explicit reason to alter the latter category at first .
this approach also agrees with the work of @xcite who showed that the dust emission in the horsehead nebula photodissociation region appears to contain a smaller ratio of pah vs. quantum - heated carbon grains than usual . in fig .
[ fig : fit-2 ] , we present the result of the simulation labelled sim2 , with a 9%/91% e - mix , and for which we have taken the dust mix @xmath115 as follow : @xmath116 in other words , with @xmath117 @xcite , we assume that there are no pahs containing less than @xmath180 carbon atoms , and a decrease of 50% of the amount of pahs containing from @xmath180 to @xmath1300 carbon atoms , with respect to the initial ism dust mix . , but for simulation sim2 : 9%/91% e - mix with pahs depletion as described in eq . [
eq : xi - fit ] .
bottom : fit residual relative to the recorded echo spectrum ( full line ) and the sim1 residual ( dash - dotted line ) for comparison .
pahs depletion enables to slightly better fit the background level of the emission at small wavelengths without impacting the fit above @xmath115 @xmath0 m . ] as expected , the lack of small pahs has an effect at small wavelengths , more precisely below @xmath115
@xmath0 m .
the subsequent simulated ir spectrum has a trend that matches better the recorded echo spectrum .
the 9 @xmath0 m gap is better reproduced , and the 6.2 @xmath0 m and 7.6 @xmath0 m features are less intense than in sim1 . however , this update of the dust mix still can not explain the small strength of the pah features in the echo spectrum , a point which we will discuss in sec .
[ sec : dehydro ] .
the resulting total mass of echoing dust for this simulation is 7.0@xmath910@xmath118 g , equivalent to a echo region density of @xmath107=679 h @xmath2 or @xmath119 g @xmath2 , with a column density of @xmath120 @xmath60 .
the uv component contributes 83.2% to the total luminosity of the simulated echo in the 5 - 38 @xmath0 m range .
the chosen form of @xmath115 might appear somewhat artificial .
it is important to clarify here that in our code , the uv burst spectrum and the @xmath115 function are two different parameters treated separately in order to fit the recorded ir echo spectrum . a first justification for this particular @xmath115 function would be to assume that the echoing dust content locally differs from the assumed _ standard _ ism dust distribution described previously , and that this precise echo region should not be considered to be pristine and unaffected by previous interactions prior to being heated by the cas a sn light . for example
, @xcite showed that pahs with less than 50 carbon atoms can be destroyed if encountering shocks with velocities greater than 100 km s@xmath43 .
furthermore , given that the _ standard _ ism dust content is not clearly defined , this justification represents a potential solution . a second explanation for the @xmath115 function is to assume that the dust content is directly affected by the uv burst via photodissociation . in this case
, the form of @xmath115 is not independent but rather a direct consequence of the uv burst . in sec .
[ sec : uv - link ] , we will discuss this possibility , and show that indeed , it is reasonable to assume pah destruction leading to the form of @xmath115 showed in eq .
[ eq : xi - fit ] using a simple theoretical model of dust destruction . in our simulations ,
we alter the dust composition of both the uv - heated and the optically - heated dust . as we will discuss in sec .
[ sec : uv - link ] , the destruction of dust by the uv burst is not an equilibrium process counter balanced by dust production , but rather a instantaneous ( on an ism dust timescale ) event .
it appears rather unlikely that any dust formation process could replenish the pahs destroyed by the uv burst before the optical component arrival . in that sense
, it appears reasonable to assume that the dust composition for the optical slab has the same composition left - over from the uv burst .
the study of other ir echo spectra should be , in the future , able to shine some more light on the behaviour of the dust after being processed by the uv burst of the sn .
as they have an ionization potential of the order of 7 ev @xcite , the uv burst can ionize large amounts of pahs in the echo region .
ionized pahs are known @xcite to have weaker 3.3 @xmath0 m and 11 @xmath0 m features compared to neutral pahs
. a stronger ionization fraction than in the normal ism could be one explanation for the weak 11 @xmath0 m complex .
however , ionized pahs also have stronger emission from 5 - 9 @xmath0 m which would increase even more the discrepancies of the fit in this zone .
since we have adopted the ionization function of @xcite in our dust model ( @xmath121 , see their fig .
7 ) , we can simulate the consequences of different ionization fractions .
in sim3 , we have assumed all of the pahs to be ionized ( @xmath122 ) , and kept all the other parameters identical to sim2 .
the resulting fit to the recorded echo spectrum in shown in fig .
[ fig : fit-2b ] .
the resulting total mass of echoing dust is 7.0@xmath910@xmath118 g , equivalent to an echo region density of @xmath107=679 h @xmath2 or @xmath119 g @xmath2 , with a column density of @xmath120 @xmath60 . , but for simulation sim3 : 9%/91% e - mix with pahs depletion as described in eq .
[ eq : xi - fit ] and 100% of pahs ionized . ] as expected , increasing the ionization fraction increases the 6.2 + 7.7 @xmath0 m to 11.3@xmath0 m ratio of the pahs features . as a consequence ,
the increased pah ionization fraction improves the fit in the 11.3 @xmath0 m region by reducing the intensity of the pah features , but also significantly worsen the fit for the 6 - 8 @xmath0 m region .
it can also be noted than even with a 100% of pahs being ionized , the 11.3 @xmath0 m feature remains strong - too strong with respect to the recorded echo spectrum .
hence , if pahs are in principle likely to be strongly ionized by the passage of the uv burst , the extent to which ionization influence the shape of echo spectrum is not evident , and another mechanism appears to be required to explain the very weak pah features . in the remaining simulations of this paper
, we will assume the original ionization fraction @xmath121 because , from our simulations , the role played by pahs ionization is not clear at this stage .
the lack of any reasonably strong features in the 11 @xmath0 m region in the recorded ir echo spectrum is quite puzzling .
these features are usually associated with c - h out - of - plane bending modes @xcite .
dehydrogenation might provide an alternative explanation to ionization for this peculiar 11 @xmath0 m complex .
dehydrogenation occurs when a pah loses some h atoms , in our case by absorption of a uv photon .
looking closer to the recorded spectrum , one can note , along with a weak 11 @xmath0 m complex , very little emission around the 5.27 @xmath0 m and 5.7 @xmath0 m features ( c - h bend / stretch combination mode , see @xcite ) .
those characteristics of the recorded echo spectrum point towards a possible depletion of emission originating from c - h bonds , and suggest a strongly dehydrogenated state of the pahs . if the intensity of the 11 @xmath0 m feature suggest a rather high dehydrogenation ratio , the presence of the 8.6 @xmath0 m c - h in plane bending is an indicator that not all hydrogen atoms have disappeared . to take dehydrogenation of pahs into account in our simulations
, we define the number of hydrogen atoms in a dehydrogenated pah @xmath123 : @xmath124 with @xmath125 the dehydrogenation ratio , and @xmath126 the number of hydrogen atoms in the _ un - altered _ pah molecule containing @xmath127 carbon atoms . in fig .
[ fig : fit-3 ] , we show the simulated spectra sim4 where the amount of h atoms per pah is 40% ( top ) , 20% ( center ) and 0% of the original value ( i.e. @xmath128 ) ) . all the other parameters are identical to the simulation sim2 shown in fig .
[ fig : fit-2 ] .
for the three cases , the resulting total mass of echoing dust is 7.0@xmath910@xmath118 g , equivalent to an echo region density of @xmath107=679 h @xmath2 or @xmath119 g @xmath2 , with a column density of @xmath120 @xmath60 .
note that we have assumed that the 5.2 @xmath0 m and 5.7 @xmath0 m features are functions of the hydrogenation ratio h / c , thus departing on this aspect from @xcite .
it is not clear why they did not have their intensities as a function of the h / c ratio , but given our recorded ir echo spectrum , we believe this update to be reasonable . the uv component of the simulated echo spectrum accounts for 83.4% , 83.3% and 83.3% of the total luminosity of the simulated echo for the 40% , 20% and 0% hydrogenation simulations .
the respective parameters of all our simulations , presented in fig .
[ fig : fit-1 ] to fig .
[ fig : fit-3 ] , are shown in table [ table : param ] .
c c c c c p1.2 cm p1.2 cm p2 cm p3 cm & & & & + name & emissivity mix & dust mix & @xmath129 & @xmath125 & @xmath98 @xmath130 [ g ] & @xmath107 [ h @xmath2 ] & echo region column density @xmath132 [ h @xmath60 ] & uv / opt . contribution to the total luminosity + sim1&9% uv + 91% opt.&ism&@xmath134&0&@xmath135 & @xmath136 & 2.8 & 80.0% / 20.0% + sim2&9% uv + 91% opt.&ism @xmath137 & @xmath134&0&@xmath138 & 679 & 3.0 & 83.2% / 16.8% + sim3&9% uv + 91% opt.&ism@xmath137 & 100 % & 0&@xmath138 & 679 & 3.0 & 83.8% / 16.2% + sim4&9% uv + 91% opt.&ism@xmath137 & @xmath134&0.4/0.2/0.0&@xmath138 & 679 & 3.0 & 83.4% / 16.6% + , but for simulations
sim4 : 9%/91% e - mix with pahs depletion as described in eq .
[ eq : xi - fit ] and pahs dehydrogenated at 60% ( top ) , 80% ( center ) and 100% ( bottom ) .
dehydrogenation might be a key factor able to explain the very weak 11 @xmath0 m pah complex . ] , but for simulations sim4 : 9%/91% e - mix with pahs depletion as described in eq .
[ eq : xi - fit ] and pahs dehydrogenated at 60% ( top ) , 80% ( center ) and 100% ( bottom ) .
dehydrogenation might be a key factor able to explain the very weak 11 @xmath0 m pah complex . ] , but for simulations sim4 : 9%/91% e - mix with pahs depletion as described in eq .
[ eq : xi - fit ] and pahs dehydrogenated at 60% ( top ) , 80% ( center ) and 100% ( bottom ) .
dehydrogenation might be a key factor able to explain the very weak 11 @xmath0 m pah complex . ]
the 11 @xmath0 m region is best fitted by the 80% dehydrogenation curve .
our aim in sim4 was mainly to highlight the effect of dehydrogenation on the simulated echo spectrum .
we are here , quite clearly , pushing our simulations to their limits . in reality , ionization of dust might play along with dehydrogenation and reduce this value . it will likely be very interesting to study how universal these c - h features strengths are in various light - echo spectra , and if they display any spatial variations , change with the sn type or are function of the echo region - to - sn distance , for example .
we note here that in our simulation , dehydrogenation seems to be more important ( and plays a bigger role ) than pahs destruction - a rather intuitive fact .
one would expect pahs first to get dehydrogenated by loosing their surrounding h - atoms , and only then , if the radiation field is strong enough , to get destroyed via further photodissociation events inside their more robust carbon skeleton .
it can be noted that compared to the sim2 simulation , the 6 - 9 @xmath0 m pah features are slightly stronger in the sim3 simulation .
in fact , looking closely at the residual curves , the stronger the dehydrogenation value , the stronger the 6 - 9 @xmath0 m features get .
this fact most likely arises as we alter and artificially remove hydrogen atoms from pahs , which will re - radiate more of their internal energy using c - c bonds . at this point , we are then able to reproduce closely the recorded echo spectrum and photometric measurements from 5 to 70 @xmath0 m , with the exception of the 5 - 9 @xmath0 m region , in which our simulation has too strong pahs features .
we do not have direct explanation for this overestimate , however , it appears quite clear that pahs are strongly altered in this echo region . especially , a more careful treatment of pahs dehydrogenation than the one adopted here would be required to better understand the consequences of a strong and short - lived uv burst , and might potentially be able to explain the peculiar 5 - 9 @xmath0 m region . finally , we note that based on the classification of @xcite , our recorded echo spectrum falls into the so - called ab category , based on its 7.7 @xmath0 m complex .
this is a peculiar category containing only two objects , a non - isolated herbig ae be star and a post - agb star , among the 57 sources studied , for which the 7.7 @xmath0 m complex is actually composed of two features at 7.6 @xmath0 m and 7.8 @xmath0 m with roughly equal strengths .
this association of the unusual light - echo spectrum with the peculiar ab class might eventually turn out to be of interest when trying to understand the origin of the 7.8 @xmath0 m feature @xcite .
using a similar approach to the present work , @xcite presented simulations of several cas - a light echoes in the 10 - 40 @xmath0 m range .
they especially experimented with three different types of heating spectra : optical light curves , uv bursts , and ( so - called ) extreme uv bursts ( or euv ) . because they used several different bursts characteristics , spanning a range of luminosities and temperatures , and managed to successfully reproduce their echo spectra ,
it is interesting to compare our uv - b burst based on our knowledge of the cas a sn type to their models . in fig .
[ fig : dwek ] , we plot our uv - b burst against their uv18 and euv18 curves .
both are black - body spectrum with a temperature t@xmath139 k and t@xmath140 k. the number _ 18 _ refers to the assumed sn - echo column density , @xmath141 @xmath60 .
note that compared to their fig .
6 , both uv18 and euv18 curves have been scaled to account for the different distance to the light echo ( 160 ly versus 199 ly ) , as well as by 0.11 and 0.15 respectively to match the averaged luminosity over all their echo regions for the respective burst model ( see table 3 in @xcite ) .
l@xmath103 and @xmath142 l@xmath103 respectively , as mentioned by @xcite ( see their table 3 ) . ]
it is interesting to note that our respective burst intensities agree very well .
the total luminosities of l@xmath143 erg s@xmath43 and l@xmath144 erg s@xmath43 are slightly lower but nevertheless in good agreement with our uv - b luminosity of l@xmath145 erg s@xmath43 .
the difference is mostly a consequence of the shape modification of our uv - b burst .
the luminosity of our black - body uv burst prior to the modification is @xmath73 erg s@xmath43 , in perfect agreement with the work of @xcite . because they did not use data below 13 @xmath0 m ,
@xcite were not able to find a model fitting the data significantly better than the others , with exception to the optical ones that proved not to be able to reproduce the light echo shape .
our simulations confirm these conclusions , in that the 10 - 40 @xmath0 m range in a light echo spectrum is mostly influenced by the uv burst heating the dust . distinguishing between various uv burst characteristics only becomes possible at smaller wavelengths , below 10 @xmath0 m , where the behaviour of the various pah features , as well as the global shape of the spectrum contains unique signatures of the intensity and temperature of the uv burst . in that sense ,
our simulations suggest the euv models of @xcite to be more likely , especially as energetic photons are required to explain the weak 11 @xmath0 m pah complex with dehydrogenation . as for the optical light curve , accounting for a mere 17% in the 4 - 38 @xmath0 m range in our simulations , its influence really becomes critical above 50 @xmath0 m , explaining why @xcite , which did not use 70 @xmath0 m photometry data , disregarded its influence in their fit . in our various simulations
, we find an echo region density of 640 - 679 h @xmath2 , relatively comparable to the value of @xmath1385 h @xmath60 of @xcite .
note that these do not , in principle , need to agree , as different light echoes in different locations have been considered in both works . in terms of column density ,
we find a much higher value , roughly @xmath120 h @xmath60 , as compared to @xmath146 h @xmath60 .
this difference arises as we model our echo region as containing both a uv - slab and an optical slab . taking only the uv - component of our echo model into account to fit the recorded spectrum ,
we would find a dust mass of @xmath147 g. subsequently , only taking the uv- slab as defining the volume of the echo region , our simulations would lead to a column density of @xmath148 h @xmath60 , more consistent with the results of @xcite , especially their echo 6 ( see their table 4 ) .
in order to improve the fit of the recorded spectrum ( especially the zone below 15 @xmath0 m ) , we have updated the dust mix in our simulations using the @xmath115 factor ( see eq .
[ eq : xi ] ) .
the form of @xmath115 as defined in eq .
[ eq : xi - fit ] has been plugged in as an individual parameter of the simulation . however , if this lack of dust is linked to the uv burst , then @xmath115 is not independent .
the strength of the burst will determine the amount and sizes of pahs destroyed , while the duration of the burst will determine the total column density of the echo region in which the pahs are destroyed , and the relative mix of uv - heated and optically - heated grains .
this dependency of @xmath115 can be used as a consistency check for our pahs destruction hypothesis .
the fate of pahs in the ism , and the influence of photodissociation , ionization and collisional processes has been subject to several studies ( e.g. * ? ? ?
* ; * ? ? ?
, we consider dust destruction using the same approach as @xcite .
let us define f@xmath149(@xmath150 ) , the incoming radiation flux heating the dust ; c@xmath151(@xmath150 , n@xmath152 ) , the pah absorption cross section as a function of the frequency @xmath150 and the number of c - atoms n@xmath152 .
excited pahs will de - excite themselves by ir emission in a time t@xmath11 s , unless their internal energy is higher than a threshold value @xmath153 . in this case
, the pah molecule will be disrupted and loose one or more atoms . for a pah with @xmath154 carbon atoms ,
we take the threshold energy input within 1 s for pah destruction , @xmath153 , to be : @xmath155}=\frac{n_c}{2}\ ] ] this relation is based on the assumption that : @xmath156 with @xmath157 the minimum temperature for destruction , and @xmath158 ev the critical ( arrhenius ) energy of the atom being ripped off the molecule .
let us subsequently define @xmath92 the number of dissociated atoms in a pah of size @xmath154 receiving an energy @xmath159 within 1 s of time , and @xmath160 the kinetic energy of one ripped off atom . then ,
similarly to eq .
[ eq : elim ] : @xmath161 in words , as a pah is heated by an incoming energy @xmath162 higher than the threshold value @xmath153 , it will loose @xmath92 atoms until its internal temperature decreases down to @xmath157 . rearranging the terms and using eq .
[ eq : elim ] , it is possible to express @xmath92 as a function of the incoming energy @xmath162 and the pah size n@xmath152 : @xmath163}-\frac{n_c}{10}\ ] ] where @xmath162 can be reached by the absorption of one single energetic photon , or several less energetic photons .
we here strictly follow the work of @xcite , which we refer the reader to for more details . at the estimated distance of the echo region , the amount of time required to get hit by a photon ( triangle - symbols line in fig .
[ fig : tin ] ) , assuming the template uv incoming radiation field described in sect .
[ sec : snbursts ] , is more than one second for pahs with @xmath164 .
hence , those pah can only loose atoms if they are hit by one single photon with an energy @xmath165 .
pahs with sizes above 10@xmath61 @xmath154 will , in average , get hit by several photons every second , however , taking into account the mean energy of the absorbed photons , the total absorbed energy will not reach the limit value @xmath153 for the radiation field strengths considered in this work .
the optical light curve , as considered in this study , does not contain photons energetic enough to be able to destroy pahs . for pahs with @xmath166 ,
the absorbed _ destroying _ energy @xmath167 per pah per second as a function of the pah size is : @xmath168 this corresponds to a number of _ destroying _ photons @xmath169 absorbed per pah per s : @xmath170 which finally leads to the mean time @xmath171 necessary to absorb one _ destroying _ photon : @xmath172 and the mean energy of the destroying photons : @xmath173 in fig .
[ fig : tin ] , we show the time required by a pah to absorb one destructive photon , t@xmath174 , as a function of pah size @xmath154 ( filled - dots line ) , along with the average time for a pah to get hit by a photon of any energy ( triangle - symbols line ) .
we use our sn burst templates described in sec .
[ sec : snbursts ] and assume the distance of 199 light years ( see sec .
[ sec : obs ] ) from cas a to the echo region . under those assumptions
, we see that it takes @xmath12.5 days for a pah with @xmath154=450 c - atoms to be hit by one destructive photon with the uv burst described in sect .
[ sec : snbursts ] shining on the dust particles . in the pah .
the triangle - symbols line shows the time required to get hit by a photon of any energy , including those not able to rip off atoms of the pah structure . ]
we also plot in fig .
[ fig : tin ] a higher - end estimate of the time required for a pahs of a certain size to be destroyed via photodissociation ( empty - dots line ) , based on the conservative following approach : assuming an initial pah size @xmath175 , we find the mean time required to get hit by a destructive photon ( eq . [ eq : tin ] ) , calculate the number @xmath92 of atoms kicked off ( via eq .
[ eq : x ] ) given the destructive photon mean energy ( from eq.[eq : numean ] ) .
we then start again with the new , smaller , pah with @xmath176 atoms , and iterate the process until we reach @xmath177 , a value below which the pahs are photolitically unstable @xcite .
summing the time of the various steps of the loop gives us the mean destruction time @xmath178 for a pah of size @xmath175 .
this time can be considered as an over estimate , as a pah , after having lost some atoms , might not turn into a stable pahs of the new size .
the c - atoms will likely be ripped off at random position , and thus weaken the pah structure . in that sense , it may be possible that less impacts of destroying photons are required in order to achieve the destruction of the pah - this , however , is not certain ( see * ? ? ?
* ) . as seen in fig .
[ fig : tin ] , the destruction time increases rapidly with the pah size , and it already takes @xmath12.5 days in order to destroy pahs with an initial size of @xmath1300 c - atoms .
let us recall here that assuming all pahs below 80 carbon atoms and 50% of pahs from 80 to 300 carbon atoms to be destroyed lead to the sim2 simulation .
these very similar values suggest , under the assumptions of this simple pah destruction model , that the destruction of pahs by the uv burst appears plausible and logical .
it should be noted here that we have neglected any dust processing mechanisms other than photodissociation , such as recombination , hydrogenation or the coagulation of small molecules on bigger ones @xcite .
we do not expect these to have an impact as important as photodissociation in shaping the echo spectrum , as the reaction timescale are expected to be significantly larger than the duration of the uv - burst , typically of the order of a few years for hydrogenation and recombination ( in a standard ism environment , see for example * ? ? ?
* ; * ? ? ?
such a dust destruction scenario also appears of special interest when considering the fact that @xcite have found strong [ c i]@xmath109850 carbon emission in their optical spectrum of an echo region around cas a. the [ c i ] lines can be excited via two main mechanisms : radiative recombination and collisional excitation by electrons .
@xcite suggested the first mechanism to explain the [ c i ] emission lines from m 42 and ngc 2024 , requiring strong radiation ( 10@xmath179 to 10@xmath180 times the isrf ) heating gas with a density of 10@xmath181 h @xmath2 ; a radiation field intensity not so different from the one considered in this work . in our dust model , the c / h ratio in pahs is 16.1 ppm for n@xmath182100 , 26.1 ppm for n@xmath182200 and 36.1 ppm for n@xmath182500 ( see table 3 in @xcite ) .
dust destruction in the form of @xmath115 described in eq .
[ eq : xi ] will therefore , if affected pahs are entirely disrupted , increase the ism gas phase abundance of c atoms by @xmath124 ppm .
standard carbon abundances in the gas phase of low to moderate density ism have been measured to be of the order of 150 ppm ( e.g. * ? ? ?
* ; * ? ? ?
* ) , while @xcite found abundances of the order of 50 - 100 ppm in the cold neutral medium .
recently , using a different estimation method based on the strong transition of cii @xmath101334 , @xcite suggested that the carbon gas phase abundance might be lower by @xmath140% than previous values found using methods based on a weak intersystem absorption transition . in any case , our dust destruction scenario could increase the amount of c atoms in the gas phase by as much as 10%-20% , which could in turn potentially affect the [ c i ] line by the same factor .
if this suggest that the [ c i]]@xmath109850 detected by @xcite is not an unambiguous and direct confirmation of our scenario of pahs destruction , it can not be ruled out that destroyed pahs influence its strength .
clearly , the study of other optical light echoes is required to better understand the origin and the ubiquity ( or not ) of the [ c i]@xmath109850 line .
we have obtained a spitzer irs spectrum and mips photometric measurements of an ir echo region around the cas a supernova . using simulations of an ism dust mix containing both carbon and silicate quantum - heated and thermal grain as well as pahs ,
we can , to a large extent , reproduce the recorded ir echo spectrum . specifically : 1 .
we find that the infrared echo spectrum can be described as the thermal signature from an uv and an optical component .
those uv and optical component characteristics correspond to the values of the type iib sn 1993j which is considered to be representative of the cas a supernova based on its optical spectrum .
their respective contribution in our best fit to the echo total luminosity are @xmath183% and @xmath117% ( in the 5 - 38 @xmath0 m range ) .
2 . the influence of the optically - heated dust becomes more important than uv - heated dust above @xmath150@xmath0 m .
especially , the presence of the optical component in our heating spectrum is consistent with the 70@xmath0 m photometric measurement of the recorded echo .
the spectrum below 15 @xmath0 m is better reproduced when removing artificially small pahs with less than 80 carbon atoms , and by removing 50% of pahs with 80 to 300 carbon atoms .
the especially weak 11@xmath0 m complex of pah features in the recorded echo spectrum can be reproduced when introducing 80% of dehydrogenation for all pahs in our simulation .
our best fit simulation implies a lower limit for the density of this echo region of @xmath183679 h @xmath2 .
the removal of small pahs is consistent with a simple and rather conservative theoretical model of pahs destruction via photodissociation assuming the considered uv burst radiation field .
the strong dehydrogenation ratio of pahs , as compared to the amount of pahs destruction , is also consistent with this picture , where pahs tend to loose hydrogen atoms first . in the future , a more careful modelling of pahs destruction and dehydrogenation as compared to the one adopted here could lead to an even better fit to the recorded ir echo spectrum .
the relatively weak 5 - 9 @xmath0 m pah complex , where our simulations differ from the recorded echo spectrum by a factor of @xmath1250% , remains to be explained .
the approach we adopted in this work , in order to study dust processing in the ism , has a great potential .
our results aim at being tested towards other echo regions around cas a and other snr , and this preliminary study shows the wealth of potential outcomes .
the unique nature of ir echoes enables the study of otherwise cold and very faint ism structures being strongly processed over a very short period of time .
such ism clouds , potentially being initially pristine , undisturbed , and located away from any altering radiation sources , can provide unique insights in the ism physics .
furthermore , the large amount of echoes provides us with an extensive number of locations and environments to test the simulations and conclusions presented in this article .
it is currently rather difficult to obtain near ir spectra of echoing regions , now that the spitzer space telescope has reached the end of its duty cycle .
fortunately , there are some exciting prospects for observations at those wavelengths , such as the _ james webb space telescope _ ( jwst ) @xcite or the _ stratospheric observatory for infrared astronomy _
( sofia ) @xcite which has seen its first light in may 2010 .
these telescopes and their instruments will be ideally suited to record infrared spectra of echoing regions , and we expect that infrared echoes will provide many new and unique insights into the ism dust composition and chemistry , as well as into the sn symmetries and characterization .
we thank the anonymous referee for his / her constructive review .
most of this work has been carried out during two visits of f.v . at the max - planck - institut f@xmath184r astronomie in heidelberg .
f.v . thanks alexandra bohm for her help dealing with the logistic related to those visits .
this research has made use of nasa s astrophysics data system .
bakes , e.l.o . , tielens , a.g.g.m . , bauschlicher , jr . ,
c.w . , hudgins , d.m .
, allamandola , l.j . :
theoretical modeling of infrared emission from neutral and charged polycyclic aromatic hydrocarbons .
560 * , 261 - 271 ( 2001 ) becklin , e.e . ,
tielens , a.g.g.m . ,
gehrz , r.d . ,
callis , h.h.s . : stratospheric observatory for infrared astronomy ( sofia ) .
society of photo - optical instrumentation engineers ( spie ) conference series * 6678 * ( 2007 ) dwek , e. , arendt , r.g . ,
fixsen , d.j . ,
sodroski , t.j . ,
odegard , n. , weiland , j.l . ,
reach , w.t . , hauser , m.g . , kelsall , t. , moseley , s.h . , silverberg , r.f . , shafer , r.a . , ballester , j. , bazell , d. , isaacman , r. : detection and characterization of cold interstellar dust and polycyclic aromatic hydrocarbon emission , from _ cobe _ observations . * 475 * , 565 - 579 ( 1997 ) fesen , r.a .
, hammell , m.c .
, morse , j. , chevalier , r.a . , borkowski , k.j .
, dopita , m.a . ,
gerardy , c.l . ,
lawrence , s.s . ,
raymond , j.c .
, van den bergh , s. : the expansion asymmetry and age of the cassiopeia a supernova remnant . * 645 * , 283 - 292 ( 2006 ) gardner , j.p . ,
mather , j.c .
, clampin , m. , doyon , r. , greenhouse , m.a . , hammel , h.b . ,
hutchings , j.b .
, jakobsen , p. , lilly , s.j .
, long , k.s . ,
lunine , j.i . ,
mccaughrean , m.j . , mountain , m. , nella , j. , rieke , g.h . ,
rieke , m.j . ,
rix , h - w . ,
smith , e.p .
, sonneborn , g. , stiavelli , m. , stockman , h.s . , windhorst , r.a . , wright , g.s .
: the james webb space telescope .
space science reviews * 123 * , 485 - 606 ( 2006 ) gehrz , r.d . , becklin , e.e .
, de pater , i. , lester , d.f . , roellig , t.l . , woodward , c.e .
: a new window on the cosmos : the stratospheric observatory for infrared astronomy ( sofia ) .
advances in space research * 44 * , 413 - 432 ( 2009 ) gordon , k.d . , rieke , g.h . , engelbracht , c.w . , muzerolle , j. , stansberry , j.a . , misselt , k.a . , morrison , j.e . , cadien , j. , young , e.t . ,
dole , h. , kelly , d.m . , alonso - herrero , a. , egami , e. , su , k.y.l . ,
papovich , c. , smith , p.s . ,
hines , d.c . , rieke , m.j . ,
blaylock , m. , prez - gonzlez , p.g .
, le floch , e. , hinz , j.l . , latter , w.b . ,
hesselroth , t. , frayer , d.t .
, noriega - crespo , a. , masci , f.j . , padgett , d.l . , smylie , m.p . , haegel , n.m .
: reduction algorithms for the multiband imaging photometer for spitzer . *
117 * , 503 - 525 ( 2005 ) higdon , s.j.u . ,
devost , d. , higdon , j.l . , brandl , b.r . , houck , j.r . , hall , p. , barry , d. , charmandaris , v. , smith , j.d.t . , sloan , g.c . ,
green , j. : the smart data analysis package for the infrared spectrograph on the spitzer space telescope . * 116 * , 975 - 984 ( 2004 ) hines , d.c . ,
rieke , g.h . ,
gordon , k.d . , rho , j. , misselt , k.a . ,
woodward , c.e . ,
werner , m.w . ,
krause , o. , latter , w.b . , engelbracht , c.w . , egami , e. , kelly , d.m .
, muzerolle , j. , stansberry , j.a . , su , k.y.l . , morrison , j.e . ,
young , e.t . , noriega - crespo , a. padgett , d.l . ,
gehrz , r.d . ,
polomski , e. , beeman , j.w . ,
haller , e.e . : imaging of the supernova remnant cassiopeia a with the multiband imaging photometer for spitzer ( mips ) . *
154 * , 290 - 295 ( 2004 ) hony , s. , van kerckhoven , c. , peeters , e. , tielens , a.g.g.m . ,
hudgins , d.m .
, allamandola , l.j . : the ch out - of - plane bending modes of pah molecules in astrophysical environments . * 370 * , 1030 - 1043 ( 2001 ) houck , j.r . ,
roellig , t.l .
, van cleve , j. , forrest , w.j . , herter , t. , lawrence , c.r . , matthews , k. , reitsema , h.j . , soifer , b.t . , watson , d.m . , weedman , d. , huisjen , m. , troeltzsch , j. , barry , d.j .
, bernard - salas , j. , blacken , c.e . , brandl , b.r .
, charmandaris , v. , devost , d. , gull , g.e . , hall , p. , henderson , c.p . ,
higdon , s.j.u . ,
pirger , b.e . , schoenwald , j. , sloan , g.c . , uchida , k.i . , appleton , p.n . , armus , l. , burgdorf , m.j .
, fajardo - acosta , s.b .
, grillmair , c.j .
, ingalls , j.g . , morris , p.w . , teplitz , h.i .
: the infrared spectrograph ( irs ) on the spitzer space telescope . * 154 * , 18 - 24 ( 2004 ) hunter , d.a . ,
kaufman , m. , hollenbach , d.j . ,
rubin , r.h . ,
malhotra , s. , dale , d.a .
, brauher , j.r . ,
silbermann , n.a . , helou , g. , contursi , a. , lord , s.d . : the interstellar medium of star - forming irregular galaxies : the view with iso . * 553 * , 121 - 145 ( 2001 ) jeffery , d.j . , kirshner , r.p . , challis , p.m. , pun , c.s.j . ,
filippenko , a.v . , matheson , t. , branch , d. , chevalier , r.a . , fransson , c. , panagia , n. , wagoner , r.v . ,
wheeler , j.c . , clocchiatti , a. : a hubble space telescope ultraviolet spectrum of sn 1993j . *
421 * , l27-l30 ( 1994 ) krause , o. , rieke , g.h . , birkmann , s.m .
, le floch , e. , gordon , k.d . , egami , e. , bieging , j. , hughes , j.p . ,
young , e.t . , hinz , j.l . , quanz , s.p . , hines , d.c .
: infrared echoes near the supernova remnant cassiopeia a. science * 308 * , 1604 - 1606 ( 2005 ) rest , a. , welch , d.l . , suntzeff , n.b . ,
oaster , l. , lanning , h. , olsen , k. , smith , r.c . ,
becker , a.c . , bergmann , m. , challis , p. , clocchiatti , a. , cook , k.h . , damke , g. , garg , a. , huber , m.e . , matheson , t. , minniti , d. , prieto , j.l . , wood - vasey , w.m .
: scattered - light echoes from the historical galactic supernovae cassiopeia a and tycho ( sn 1572 ) . *
681 * , l81-l84 ( 2008 ) rieke , g.h . ,
young , e.t . ,
engelbracht , c.w .
, kelly , d.m .
, low , f.j . , haller , e.e . , beeman , j.w . , gordon , k.d . , stansberry , j.a . , misselt , k.a . , cadien , j. , morrison , j.e
, rivlis , g. , latter , w.b . , noriega - crespo , a. , padgett , d.l . ,
stapelfeldt , k.r . ,
hines , d.c . ,
egami , e. , muzerolle , j. , alonso - herrero , a. , blaylock , m. , dole , h. , hinz , j.l .
, le floch , e. , papovich , c. , prez - gonzlez , p.g . , smith , p.s .
, su , k.y.l . , bennett , l. , frayer , d.t . ,
henderson , d. , lu , n. , masci , f. , pesenson , m. , rebull , l. , rho , j. , keene , j. , stolovy , s. , wachter , s. , wheaton , w. , werner , m.w . , richards , p.l . : the multiband imaging photometer for spitzer ( mips ) . *
154 * , 25 - 29 ( 2004 ) werner , m.w . ,
roellig , t.l . , low , f.j . , rieke , g.h .
, rieke , m. , hoffmann , w.f . ,
young , e. , houck , j.r .
, brandl , b. , fazio , g.g . ,
hora , j.l . ,
gehrz , r.d . , helou , g. , soifer , b.t . ,
stauffer , j. , keene , j. , eisenhardt , p. , gallagher , d. , gautier , t.n . , irace , w. , lawrence , c.r . , simmons , l. , van cleve , j.e . , jura , m. , wright , e.l . ,
cruikshank , d.p . : the spitzer space telescope mission . *
154 * , 1 - 9 ( 2004 ) | we present the analysis of an irs 5 - 38 @xmath0 m spectrum and mips photometric measurements of an infrared echo near the supernova remnant observed with the _ spitzer space telescope_. we have modeled the recorded echo accounting for pahs , quantum - heated carbon and silicate grains , as well as thermal carbon and silicate particles . using the fact that optical light echo spectroscopy has established that cas a originated from a type iib supernova explosion showing an optical spectrum remarkably similar to the prototypical type iib
, we use the latter to construct template data input for our simulations .
we are then able to reproduce the recorded infrared echo spectrum by combining the emission of dust heated by the uv burst produced at the shock breakout after the core - collapse and dust heated by optical light emitted near the visual maximum of the supernova light curve , where the uv burst and optical light curve characteristics are based on sn 1993j .
we find a mean density of @xmath1680 h @xmath2 for the echo region , with a size of a few light years across .
we also find evidence of dust processing in the form of a lack of small pahs with less than @xmath1300 carbon atoms , consistent with a scenario of pahs destruction by the uv burst via photodissociation at the estimated distance of the echo region from cas a. furthermore , our simulations suggest that the weak 11 @xmath0 m features of our recorded infrared echo spectrum are consistent with a strong dehydrogenated state of the pahs .
this exploratory study highlights the potential of investigating dust processing in the interstellar medium through infrared echoes . |
since photons do not interact strongly , direct photons are a powerful probe to study the initial state of matter produced in relativistic heavy ion collisions .
they are emitted from all the states such as the initial state , the quark - gluon plasma ( qgp ) , and the final hadron - gas state .
in addition , it is predicted that the contribution of photons from jet - photon conversion in the dense medium can be as large as the photon yield from hard scatterings @xcite . in the 2004 run ,
phenix recorded a high - statistics au+au data set .
the new data set allows us to measure direct photons beyond the transverse momentum ( @xmath3 ) of 10 gev/@xmath1 and also more precisely than before at intermediate @xmath3 where thermal photons and photons from jet - plasma interactions are important .
the phenix experiment can measure photons with two types of highly segmented electromagnetic calorimeters ( emcal ) @xcite .
one is a lead scintillator sampling calorimeter ( pbsc ) , and another is a lead glass cherenkov calorimeter ( pbgl ) .
measurement of direct photons is challenging because there is a large amount of background from decay of neutral mesons such as @xmath4 and @xmath5 .
it is estimated that the @xmath6 would be @xmath7 10 % at the @xmath3 range of 2 - 4 gev/@xmath1 , where thermal photon is expected to be dominant @xcite . in order to extract direct photon signal ,
conventional subtraction method has been used . @xmath4 and @xmath5
mesons are reconstructed via their two - photon decay mode .
the @xmath3 spectra of direct photons are obtained by subtracting the spectra of decay photons estimated based on the measured @xmath4/@xmath5 from the @xmath3 spectra of inclusive photons .
in addition to the conventional subtraction method , the analysis was carried out using a new method where real direct photons are measured by their virtual counterparts @xcite .
the idea is that all sources of real photons also produce virtual photons that decay into @xmath2 pairs with small invariant masses .
the invariant mass distribution of @xmath2 pairs from virtual photons can be given by knoll - wada formula @xcite for each virtual photon sources , such as direct photon , @xmath4 dalitz and @xmath5 dalitz .
the ratio of direct photons was obtained using the ratio of photon yield in different @xmath2 invariant mass region .
compared to the conventional subtraction measurement , the new technique ( internal conversion method ) improves both the signal - to - background ratio and the energy resolution at intermediate @xmath3 .
an excess of the signal over background has been seen as the red points of left panel in fig .
[ fig : midpt ] , which is consistent with the result of conventional subtraction method within the error .
= 360 mev ) predictions.,title="fig:",width=211 ] = 360 mev ) predictions.,title="fig:",width=211 ] the right panel of fig .
[ fig : midpt ] shows the direct photon spectrum which is obtained from the excess ratio measured with internal conversion method and inclusive photon spectrum measured with emcal .
the spectrum shows the enhancement comparing with the next - to - leading - order ( nlo ) pqcd calculation scaled by the au+au nuclear overlapping function ( @xmath8 ) .
@xmath9 is defined as @xmath10 , where @xmath11 is the average number of binary nucleon - nucleon collisions at an impact parameter @xmath12 with an inelastic cross section @xmath13 . a hydrodynamical prediction which computes the thermal photon emission from a qgp with initial temperatures @xmath14 = 360 mev @xcite supports the observed enhancement of the spectrum .
however , the reference baseline in those calculations is the @xmath8-scaled nlo pqcd photon spectrum . to confirm the existence of a thermal enhancement
, it is necessary to directly measure the p+p direct photon spectrum in the @xmath3 = 1 4 gev/@xmath1 range .
in contrast to the measurement of mid-@xmath3 direct photon , the strong suppression of neutral hadrons by a jet - quenching effect in heavy ion collisions @xcite allows to extract direct photons at the momentum region of @xmath15 5 gev/@xmath1 @xcite .
the left panel of fig .
[ fig : highpt ] shows the direct photon excess ratio as a function of @xmath3 in most central events ( 0 - 10% ) .
the ratio is in good agreement with a nlo pqcd calculation @xcite scaled by @xmath8 within the error and theoretical uncertainty .
this suggests that the initial - hard - scattering probability is not reduced .
the direct photon excess ratio measured by pbsc - emcal slightly deviates from the nlo pqcd calculation above 14 gev/@xmath1 .
the deviation may support that the compensation of jet - quenching and additional photons by a nuclear effect , such as jet - photon conversion , is breaking up above 14 gev/@xmath1 . because the direct photon calculated with pqcd consists of the prompt photon produced from the hard scattering directly and the jet fragmentation photons coming from hard - scattered partons ,
the direct photon yield measured in au+au collisions is naively expected to be suppressed comparing with nlo pqcd calculation scaled by @xmath8 . the agreement with scaled nlo pqcd calculations
might just be a coincidence caused by mutually counterbalancing effects like energy loss and compton like scattering of jet partons @xcite . in order to conclude that the deviation attributes to the such nuclear effect , it is important to measure high-@xmath3 direct photon in p+p system and to observe the deviation with pbgl - emcal as well .
the right panel of fig .
[ fig : highpt ] shows the fully corrected direct photon invariant yield as a function of @xmath3 for each centrality .
the spectra are combined results of pbsc measurement and pbgl measurement .
the yields are in good agreement with @xmath8-scaled nlo pqcd calculation up to 14 gev/@xmath1 . measured with subtraction method .
three types of solid curves are theoretical curve : nlo pqcd scaled by @xmath8 , and a calculation which takes into account the jet - quenching of direct photons and photons from jet - plasma interaction . right : the direct photon spectra for different centrality classes and minimum bias .
the solid curves are nlo pqcd calculations scaled by @xmath8 . ] measured with subtraction method .
three types of solid curves are theoretical curve : nlo pqcd scaled by @xmath8 , and a calculation which takes into account the jet - quenching of direct photons and photons from jet - plasma interaction .
right : the direct photon spectra for different centrality classes and minimum bias .
the solid curves are nlo pqcd calculations scaled by @xmath8 . ]
the phenix experiment has measured direct photons in au+au collisions at @xmath0 = 200 gev .
the large amount of data taken in rhic run4 has made it possible to extend @xmath3 region up to 18 gev/@xmath1 .
the result is consistent with the nlo pqcd calculation scaled by the nuclear overlap function .
the measurement using internal conversion of photons into @xmath2 shows the enhancement of the yield comparing with nlo pqcd calculation .
fries b. mller and d.k .
srivastava , phys . rev . * c72 * , 041902 ( 2005 ) .
l. aphecetche _ et al .
instr . and meth .
* a499 * , 521 ( 2003 ) . s. turbide r. rapp and c. gale , phys .
* c69 * , 014903 ( 2004 ) .
s. bathe _ et al .
_ , nucl - ex/0511042 ( 2005 ) . n.m . kroll and w. wada ,
phys . rev . *
98 * , 1355 ( 1955 ) .
d. denterria and d. peressounko , eur .
j. c * 46 * , 451 ( 2006 ) .
adler _ et al .
* 91 * , 072301 ( 2003 ) .
adler _ et al .
* 94 * , 232301 ( 2005 ) .
gordon and w. vogelsang , phys .
* d48 * , 3136 ( 1993 ) .
s. turbide c. gale s. jeon and g.d .
moore , phys . rev . *
c72 * , 014906 ( 2005 ) . | direct photons have been measured with the phenix experiment in au+au collisions at @xmath0 = 200 gev .
the direct photon result obtained with phenix - emcal up to 18 gev/@xmath1 is consistent with the nlo pqcd calculation scaled by the nuclear overlap function .
the measurement using internal conversion of photons into @xmath2 shows the enhancement of the yield comparing with nlo pqcd calculation . for the phenix collaboration relativistic heavy ion collision , quark gluon plasma , direct photon 25.75.nq , 25.75.-q |
synchronous phenomena abound in nature and in our daily lives and have been studied from centuries past , right from huygen s observations of synchronizing clocks .
various kinds of synchronous phenomena occur and have been identified ( see for example @xcite ) ( other references on the subject may also be found in @xcite ) . among these complete synchronization ( cs )
is one of the most interesting since the phase , frequency and amplitude of a subsystem all coincide with those of the other subsystem it is coupled to .
it is seen therefore that in cs the trajectories of the coupled elements match exactly .
cs is known to occur in identical systems , and was first demonstrated in chaotic systems in @xcite .
neurons and neuronal networks have been a subject of frequent theoretical and experimental studies @xcite .
synchronization of neural activity has elicited a great deal of interest since it is believed that such phenomena enable cognitive tasks such as feature extraction and recognition to be performed .
hodgkin and huxley classified neuron excitability mechanisms broadly into two classes @xcite : in type - ii neurons the transition from a quiescent state to a periodically spiking state occurs through a hopf bifurcation with a finite nonzero oscillation frequency . in type - i neurons ,
oscillations emerge through a saddle - node bifurcation on an invariant circle . as the bifurcation parameter changes , the stable and the unstable fixed points coalesce and then disappear , leaving a large amplitude stable periodic orbit .
this is a global bifurcation and the frequency of the global loop can be arbitrarily small .
since axonal excitability patterns of mammalian neurons fall under the type - i class , it is but natural that this class has received special attention in the literature .
various observations have been made on type - i neurons , some prominent points of which are as follows .
equations for type - i neuronal dynamics can be reduced to the canonical normal form for a saddle - node bifurcation @xcite .
repetitive firing occurs in the parameter regime when the system is in the close proximity of a saddle - node bifurcation on an invariant circle .
et al _ @xcite and ermentrout have shown @xcite that such neurons coupled via a certain class of time - dependent synaptic conductances are difficult to synchronize . brgers and kopell made further investigations of such coupled systems .
in particular they discussed the effects of random connectivity on synchronization and the ping mechanism in networks of excitatory ( e ) and inhibitory ( i ) neurons both in the presence and in the absence of external noise @xcite . in this work ,
we present some computer studies of generic type - i neurons coupled via synaptic conductances such as those considered in @xcite , which are governed by ordinary differential equations and which depend upon the outputs of the presynaptic neurons , and are subject to weak additive gaussian white noise .
we consider both excitatory - excitatory ( ee ) and inhibitory - excitatory ( ie ) bidirectional couplings and show that in certain regimes of the coupling constants and inputs , the system of coupled neurons shows complete synchronization ( cs ) .
the issue of cs in type - i neurons was not discussed in @xcite . as discussed in section-2 ,
largest lyapunov exponents are shown to not adequately give information about cs in the system .
we make an observation on the inputs to the neurons ( and which are also modulated by the feedback in the system ) : we point out that the variation of the instantaneous frequency of the input received at each neuron with the instantaneous phase of the input it receives exactly coincides with that of the other neuron , in the event of complete synchronization of their outputs . it will be noted that in the presence of noise and feedback , this is not a trivial statement .
we discuss the utility of this result in practical situations .
+ in general , for ee synapses , our results indicate that when a common , externally applied constant input is used to perturb two bidirectionally coupled type - i neurons having identical coupling strength magnitudes and synaptic rise & decay times , weak noise induces them to exhibit cs upto a critical value of the coupling strength @xmath0 . for coupling strengths larger than @xmath0
, we find the system de - synchronizes through a power - law before locking on to a partially synchronized state for larger coupling strengths .
we obtain a functional dependence for the synchronization error for neuronal outputs on coupling and noise strengths , in the regimes leading to partial synchronization .
such functional dependencies have not been reported in the literature yet , to the best of our knowledge . in the noiseless case for identical ee neurons separated by different initial conditions ,
we observe that the antiphase states are stable in agreement with @xcite and become completely in - phase in the presence of noise . for just two neurons with ie coupling , noise does not induce complete synchrony . in an ensemble of 200 non - identical neurons
however , we find unexpectedly that noise - induced cs is possible with all - to - all bidirectional ie random couplings .
in a system of @xmath1 neurons , the activity of the @xmath2th neuron is described by a variable @xmath3 which can be related to the membrane conductance .
its dynamics is represented by @xmath4 where @xmath5 denotes an inverse time constant for the membrane potential , @xmath6 denotes its total input comprising of a constant external input @xmath7 and the contributions from the presynaptic neurons , with @xmath8 .
@xmath9 is the synaptic gating variable and represents the fraction of ion channels open in the @xmath10th presynaptic neuron .
@xmath11 is the measure of the strength of the synapse from neuron @xmath2 to neuron @xmath10 ; we have taken @xmath12 . when @xmath13 this equation has no fixed points .
any initial condition tends to infinity in a finite time . to avoid this blow - up of solutions , a nonlinear transformation to new variables
@xmath14 may be made @xcite : @xmath15 which maps the real line onto a circle .
eqn.(2 ) then becomes @xmath16 the point @xmath17 then gets mapped to the point @xmath18 and is interpreted as firing of a spike .
we set @xmath19 and work in the parameter regime in which @xmath20 so that the width of the spikes turn out to be in milliseconds as in real neurons @xcite in these units .
the @xmath21 evolve in time according to the differential equation which was considered in @xcite @xmath22 where @xmath23 denotes the synaptic decay time and @xmath24 the synaptic rise time .
the values of @xmath9 always lie in the range 0 to 1 , reaching the maximal value when the neuron spikes .
the synapse is an excitatory one if @xmath25 , and @xmath26 if it models an inhibitory synapse .
lest there be any confusion , we would like to clarify at the outset that when we refer to _ identical _ neurons , we mean neurons that have the same nature of synaptic coupling ( ee ) and have same coupling strengths @xmath27 , have the same value of @xmath24 , @xmath28 , and receive the same constant input @xmath29 , while their initial conditions differ by very little .
+ transmembrane voltage and neuronal firing can be affected by various sources of neuronal noise , but predominantly by synaptic noise @xcite .
the synaptic noise itself occurs due to several factors , but chief among them is the synaptic bombardment at the inputs through the large number of neuronal connections , with each input spike adding a random contribution .
we model this through an additive gaussian white noise added to the neuronal input @xmath6 .
we study the dynamics of a system of two such neurons coupled bidirectionally as depicted in fig.(1 ) and subject to gaussian white noise @xmath30 with the following properties : @xmath31 , @xmath32 , where the stochastic variables are taken to obey stratonovich calculus .
addition of gaussian white noise @xmath33 to eqn.(1 ) , manifests as multiplicative noise in eqn.(2 ) because of the change of variables to @xmath34 , so that the equations now take the form @xmath35 eqns.(3 ) for the @xmath36 define the feedback regulating the activity of the @xmath10th neuron since @xmath9 depends upon @xmath14 which in turn depends upon @xmath37 ( @xmath38 ) via @xmath39 .
the feedback increments or decrements the constant external input @xmath7 received by neuron @xmath2 .
thus the control parameter @xmath6 acquires time dependence through the dynamical variables . as in @xcite
, we consider the neuronal output to be described by the variable @xmath40 as its time evolution pattern resembles that of a membrane potential in real neurons .
this transformation maps the resting point @xmath41 corresponding to @xmath42 to @xmath43 , and the spiking point @xmath44 to @xmath45 via the relation @xmath46 .
we choose to work with these variables as we get some new and interesting insights upon the dynamics underlying the phenomenon of complete synchronization . in terms of these variables , eqns.(4 ) and ( 3 ) become : @xmath47 numerous studies in the literature have reported the phenomenon of complete synchronization in various coupled systems @xcite .
an adequately satisfying explanation of why and under what conditions cs can occur for systems with more complicated couplings , such as , for instance , that described by eqn.(5 ) , is however , still lacking in our opinion .
+ in particular , in @xcite the authors study noise induced cs in systems subjected to a common additive white noise and show that a necessary condition for cs is the existence of a significant contraction region in phase space .
the systems studied in @xcite were lorenz and rossler systems which are far more amenable to analytical treatment than the equations above in eqn.(5 ) . + since excitability in type - i neurons results from a saddle node on an invariant circle bifurcation , complete synchronization of two uncoupled neurons by common noise alone
can be expected because of the existence of a contraction region close to the stable manifold of the saddle . on the other hand , when two such neurons are coupled together as in eqn.(5 ) , the existence and nature of a contraction region would depend upon the eigenvalues of the jacobian at the fixed points of the coupled system .
however for eqns.(5 ) , an analytical study becomes difficult since the jacobian becomes singular at the fixed points .
we therefore perform some computer studies on the system to learn more about the underlying dynamics .
moreover , since as we show below , lyapunov exponents need not adequately give information about cs , we seek other explanations for occurrence of cs .
+ as in @xcite , we define cs between the activities of neurons 1 and 2 as a vanishing value for the quantity @xmath48 which is the synchronization error averaged over all iterations .
largest lyapunov exponents for the system in eqn.(5 ) in the presence of noise for both ee and ie couplings were calculated following @xcite and are shown in fig.(2 ) . to incorporate noise in the numerical calculations ,
the stochastic runge - kutta-4 method @xcite was used .
we note that in both ee and ie cases , the largest lyapunov exponent @xmath49 becomes more negative on the addition of noise .
we observe also that for some intermediate values of the coupling constants ( such as @xmath50 ) , the @xmath49 values could be larger , i.e. , less negative , than those for lower coupling strengths ( e.g. , @xmath51 ) . for ee coupling , @xmath49 is almost always less than or equal to zero .
in the case of ie coupling however , we find that for smaller noise - strengths and smaller couplings , for small @xmath52 , @xmath49 fluctuates between positive and negative values in the presence of noise .
this happens because of the oscillation of the bifurcation parameter ( total input ) between two regimes , depending upon the relative strengths of @xmath7 and @xmath53 , since @xmath54 for neuron @xmath10 .
hence calculation of lyapunov exponents may not adequately give information about cs or large windows of zero synchronization error , such as , for example , for the situation shown in fig.(3 ) , though they may certainly show the emergence of a definite order in the presence of noise and possible synchronization between the coupled units .
indeed , cs is expected to occur between identical systems and finding cs between non - identical oscillators would be unusual . as we describe later
, however , we do find noise - induced cs in a system of 200 non - identical oscillators with random all - to - all couplings .
equally intriguingly , we find that there are parameter regimes where identical oscillators do not show cs at all , though the lyapunov exponents shown in fig.(2a ) remain negative .
negative transverse lyapunov exponents are widely accepted as characterizing cs , but their calculation for the system of 200 neurons , with a nontrivial feedback mechanism for phase resetting , as in eqn.(5 ) is a difficult task and we have not attempted it here . in our case , even for @xmath55 , if we were to define new variables : @xmath56 and @xmath57 , expressing the synchronization error dynamics through @xmath58 in terms of @xmath59 and @xmath60 alone is not at all straightforward for the equations in ( 5 ) .
+ moreover when each of the subsystems has a saddle - node - on - an - invariant circle bifurcation in the uncoupled limit , one could expect windows of intermittent firing patterns . in the coupled system , this is indeed observed in some parameter regimes , interspersed with large windows showing zero synchronization error , even in identical oscillators ( ee case ) for lower noise strengths ( fig.(3 ) ) .
hence we believe that negative largest lyapunov exponents alone may not constitute conclusive proof for predicting noise - induced cs in coupled neurons .
+ we therefore looked for other indicators which could help in understanding the mechanism of cs better in systems with feedback , such as in eqns.(5 ) .
we found one such simple indicator for cs in the context of the model under study and which we will now describe .
the same methods and analysis should also hold for getting information on cs in any other system .
since synchronous activity is brought about by a common input or through mutual interactions and since these include components which are highly random , we study the _ instantaneous _ values of the sum total _ of the inputs _
received by each unit of a coupled system .
we first set up a framework for this purpose and then provide a physical motivation and explanation for understanding cs through this indicator .
we construct the analytical signal @xcite : @xmath61 for the inputs @xmath6 and similarly @xmath62 for @xmath63 using hilbert transforms .
the instantaneous amplitudes and phases evolve according to : @xmath64 where @xmath65 , and @xmath66 denotes the instantaneous difference between the phase of the output of the @xmath2th neuron and that of the part it sends as feedback to the presynaptic conductance of the @xmath10th neuron this feedback is a stochastic component of the input for neuron @xmath10 .
we have constructed the noise terms @xmath67 and @xmath68 from the analytical noise signal @xmath69 : @xmath70 @xmath71 and @xmath72 are periodically modulated by @xmath73 which evolves according to the differential equation : @xmath74-r_i\sin\phi_i[r_i\cos(\zeta_i-\psi_i)-b_ir_i\cos(\zeta_i-\psi_i+\rho_i ) + b_i\cos(\psi_i+\rho_i)+ \xi_{r_i}]\ } \nonumber\\ & = & \frac{1}{2\sqrt{(1+r_i^2 - 2r_i\cos\phi_i)}}\big\ { ( r^2 - 1){\sqrt r } { \dot\phi } - \sin\phi { \dot r } \big\}\end{aligned}\ ] ] the instantaneous phase @xmath75 therefore has no deterministic time scales and its drift and diffusion in time in the presence of noise is influenced by @xmath76 and @xmath77 , the instantaneous values of the amplitude and phase respectively of the input , and also by the instantaneous amplitude @xmath78 of the neuronal output . @xmath79 and @xmath80 evolve as follows : @xmath81 these show the effect of feedback on neuronal response .
it is seen that cs between the outputs of neurons 1 and 2 occurs when the changes in the instantaneous phases and amplitudes of the two neuronal _ inputs _ exactly match each other . in other words ,
cs in the inputs to the neurons is required for the outputs to synchronize in phase , amplitude and frequency .
this observation is in general not quite obvious since the system is nonlinear , and has a feedback mechanism which depends upon the outputs of the other neurons .
+ in fig.(4 ) we present the instantaneous - phase versus instantaneous - frequency plots of the inputs received by the two neurons . in all the numerous cases we studied for the coupled system for @xmath55 , we found that the signature of cs is the almost identical nature of these plots for the two systems that are in synchrony , be it with or without noise . on the other hand , the absence of cs gets reflected in the non - identical variation between instantaneous phase and instantaneous rates of phase - change of inputs to the two neurons , in the plots shown in fig.(5 ) .
this is again true for both the noiseless as well as noisy cases .
+ the system of equations ( eqn.(5 ) ) is of the form @xmath82 where @xmath83 and since feedback to each neuron through the synaptic coupling is oscillatory , in a fokker - planck description of the stochastic process , the probability distribution @xmath84 of the neuronal ensemble will not be stationary in the @xmath85 limit @xcite .
in certain regimes of the noise strengths and coupling constants , where competing contributions from the drift and diffusion terms would make the noise - averaged difference in outputs @xmath86 zero , cs occurs .
physically , cs is brought about through the following sequence of events .
addition of a small amount of noise increases the decay time of the synaptic conductances @xmath9 gradually , and eventually , lowers their minimum to zero .
this delays the onset of the next @xmath9 peaks , and hence the input @xmath6 to each neuron at any further instant of time . increasing the noise strength further increases the decay time of @xmath9 .
the periodically maximal values of the inputs thus take longer to arrive at the neurons and this becomes visible in the neuronal firing pattern as departures from the previous ( noiseless ) values of the phase differences between the neurons , and those of their output differences @xmath87 .
the instantaneous values of the phases and the rate at which they change in time , i.e. , the instantaneous frequencies of each neuron , is determined by the strengths @xmath88 of the synaptic inputs it receives and the noise strength for any given set of @xmath89 .
hence for given initial conditions and different amplitudes for @xmath90 & @xmath91 , it would be reasonable to expect cs to occur when the following is satisfied for the inputs to the neurons : the variation of instantaneous values of the frequencies with instantaneous phases of the _ input _ for neuron 1 matches with that for neuron 2 .
this results in the instantaneous values @xmath92 and @xmath93 of neuron 1 changing in step with @xmath94 and @xmath95 respectively of neuron 2 .
this forces the amplitudes of neurons 1 & 2 to become identical with each other , since otherwise both conditions @xmath96 and @xmath97 can not be simultaneously maintained .
thus cs results .
a striking feature of all these plots is their strange , flame - like structure .
the flame shape is reminiscent of canards that are typically associated with systems exhibiting relaxation oscillations .
+ indeed , from the common factor occurring in the inverse square root on the right hand side of eqn.(8 ) , it is apparent that @xmath98 does evolve on a time scale different from that for @xmath99 & @xmath100 in eqns.(6 ) & ( 9 ) , though it is a dynamically varying time scale , determined also by the stochasticity of the system .
the separation of time scales is in fact manifested in the time series for the neuronal inputs which show relaxation oscillations ( figs.(4 ) & ( 5 ) ) .
this gives the input instantaneous phase - frequency curve its characteristic shape whenever noise is introduced .
this is indicative of some order emerging in the phase .
in fact , noise - induced phase synchronization can be demonstrated even in the ie system by interpreting the instantaneous phase differences @xmath101 in a statistical sense as in @xcite . the distribution @xmath102 , of cyclic instantaneous phase differences , @xmath103 $ ] is shown in fig.(6 )
. preferred phase differences between the 2 neurons manifest as peaks in @xmath104 which become sharper and taller with increasing noise - strengths a clear indication of noise - induced phase coherence .
in fig.(7a ) we have plotted the synchronization error @xmath105 for two coupled identical ( ee ) neurons and for coupled non - identical ( ie ) neurons , as a function of the coupling strength for different noise - strengths .
we have considered here the special case of @xmath106 , and each neuron receives inputs only from the other neurons ; _ i.e. _ , we look at the effect of feedback .
results and analysis for non - zero @xmath7 are presented elsewhere @xcite .
we see that although cs is expected between identical oscillators , it does not happen for the noiseless ( deterministic ) case .
increasing noise - strength brings down @xmath107 and indeed takes it down to zero for certain ranges of the coupling constant . for the identical ee coupled neurons ,
when feedback constitutes the only input ( there being no other explicit input ) , as in figs.(7a,7b ) , then beyond a maximal critical coupling constant strength @xmath0 , the system gets de - synchronized , with @xmath108 . for @xmath109
, there then exists a regime in the de - synchronized system where noise strength @xmath110 still plays a role in determining the output . in this transition regime , for @xmath109 , we find that the difference in outputs @xmath111 depends upon @xmath112 through the expression @xmath113 where @xmath114 is a constant that depends upon the noise strength @xmath110 .
+ the critical coupling strength @xmath0 depends upon the noise strength as well and we find that it varies as @xmath115 where @xmath116 . at very high @xmath112 the system gets locked to a partially synchronized state , with @xmath117 approaching a constant value , wherein noise strength no longer influences the difference in the outputs of the neurons . in fig.(7a ) , this constant value approaches 0.5 . for the curves in fig.(7a , b ) , for the entire regime following the beginning of desynchronization
, we find a functional dependence on the coupling constant @xmath112 given by an equation of the form @xmath118 where @xmath119 and @xmath120 depend on the noise - strength through @xmath121 so that @xmath122 this expression is plotted on the numerical data points in fig.(7b ) .
a rigorous theoretical treatment of the system needs to be done in future studies to establish these relations for the synchronization and desynchronization transitions through a fokker - planck approach .
+ in the ie case ( fig.(7c ) ) , we see that noise - induced cs does not occur for two coupled neurons ; however , there is partial synchronization ( by this we mean that @xmath123 ) since @xmath124 gets locked to a finite , non - zero value ( @xmath125 ) for large @xmath112 .
further , increasing noise - strength increases rather than decreases @xmath126 , in the region before partial synchronization , in contrast to the observation for the ee case .
we were unable to achieve cs in the ie case for two coupled neurons , even in the presence of noise .
however , for an ensemble of 200 coupled theta neurons of which 150 neurons are excitatory and 50 inhibitory , we obtain very different results . in this simulation shown in fig.(8 ) , each neuron receives the same input @xmath127 and there is all to all random coupling with the coupling strengths varying between @xmath128 and @xmath129 and with different initial conditions .
the excitatory neurons are shown in red while the inhibitory ones are in green .
on introducing gaussian white noise of strength @xmath130 into the system we observe synchronous phenomena emerging between the excitatory and inhibitory neurons . the interesting thing to note here
is that not only do most of the inhibitory neurons fire in synchrony with other inhibitory neurons , but that also most of them are synchronized with the excitatory neurons .
this kind of noise - induced near - cs in coupled type - i neurons with random non - zero coupling strengths has not , to our knowledge , been reported previously in the literature .
a detailed study of this situation is under way to explain the observed synchrony for @xmath131 and will be reported elsewhere : it is beyond the scope of the present work .
interestingly , in the literature , spatiotemporal synchronization has been shown to occur in networks of coupled chaotic maps with varying degrees of randomness in the coupling connections @xcite .
we have studied the issue of noise - induced cs in coupled type - i neurons , a class of neurons that are especially important since mammalian neurons fall under this category .
we find that complete synchronization between any two neurons is signalled whenever the variations of the instantaneous input phases @xmath132 versus the instantaneous frequencies @xmath133 of the inputs of the neurons being studied are identical .
we point out that such identical @xmath132-@xmath133 plots of the neuronal inputs would be a signature of cs between the neurons .
this suggests the possibility of producing completely synchronized outputs of coupled systems having feedback mechanisms in the presence of noise , by ensuring that the plots of the instantaneous frequency versus instantaneous phase of the _ inputs _ to the subsystems are identical . that this is a significant point
will be appreciated when one recalls that this condition is required in the continued presence of noise and feedback . when monitoring or control of neurons is required to be done in a living organism in the case where synchronized neuronal output is required at another , inaccessible spot through a given external input
, this result becomes important .
though cs is expected between coupled , identical neurons , we find that cs occurs only upto a critical value @xmath0 of the coupling constant @xmath112 for a given noise strength beyond which the system desynchronizes again , and then for large @xmath112 gets locked to a partially - synchronized state .
we find that the critical coupling strength @xmath0 depends upon the noise - strength through a power law .
for @xmath112 greater than the critical value @xmath0 , from the transition regime through upto the onset of partial synchronization , we find a functional dependence of the noise - averaged output difference on @xmath112 and the noise strength @xmath110 , given by eqn.(14 ) .
for a larger ensemble of 200 neurons , we find unexpectedly that non - identical neurons can show near - complete synchronization .
since type - i neurons model axonal excitability patterns in mammals , the results presented here would be useful in the study of synchronous mechanisms underlying the neural code . as an immediate application
, we believe our results would be useful in explaining the experimental observations reported on cat & awake monkey visual cortex @xcite which show synchronization of neuronal activity with a single stimulus , and which disappears when activated by different , independent stimuli . since a single stimulus would correspond to neuronal inputs of identical amplitudes , instantaneous phase and frequency ,
this is actually the same scenario that we have found for cs to occur .
it is also likewise clear why cs was experimentally observed to vanish on the activation of different stimuli , since the necessary conditions of identical input amplitude , phase and frequency are no longer met . *
b * + * figure 4 . * instantaneous phase - frequency flame " plots of stochastic * input * @xmath140 to neuron @xmath2 , for ee coupling , for * ( a ) * @xmath141 ( top ) and * ( b ) * @xmath142 ( bottom ) . in these plots ,
@xmath143 , showing the effect of feedback .
the extended panels below each set of flame plots show the corresponding time - series for neuronal input ( at left ) and the difference ( @xmath144 ) in neuronal outputs ( at right ) .
parameter values are : @xmath51 , @xmath145 , @xmath146 .
+ * figure 5 . * more instantaneous input phase - frequency
flame " plots with corresponding time - series of neuronal input .
row 1 : ee , @xmath147 ; row 2 : ee , @xmath148 ; row 3 : ie , @xmath147 ; row 4 : ie , @xmath148 .
parameters used are:@xmath51 , @xmath145 , @xmath149 , @xmath146 .
+ * figure 7 . * noise - induced synchronization in coupled type - i neurons ( a ) ee case ; transition from synchronized to partially synchronized state when @xmath109 ( b ) ee case , as in ( a ) ; solid lines correspond to eqn.(16 ) .
( c ) ie case ; noise - induced cs is absent but there is partial synchronization as the system gets locked to @xmath151 for large @xmath112 .
for both ( a ) & ( b ) : @xmath136 , @xmath152 , @xmath153 , initial conditions @xmath154 , @xmath155 , @xmath156 + | for a system of type - i neurons bidirectionally coupled through a nonlinear feedback mechanism , we discuss the issue of noise - induced complete synchronization ( cs ) . for the inputs to the neurons ,
we point out that the rate of change of instantaneous frequency with the instantaneous phase of the stochastic inputs to each neuron matches exactly with that for the other in the event of cs of their outputs .
our observation can be exploited in practical situations to produce completely synchronized outputs in artificial devices . for excitatory - excitatory synaptic coupling , a functional dependence for the synchronization error on coupling and noise strengths
is obtained .
finally we report an observation of noise - induced cs between non - identical neurons coupled bidirectionally through random non - zero couplings in an all - to - all way in a large neuronal ensemble .
complete synchronization , noise , coupled type - i neurons + |
circinus x-1 is one of the most exotic x - ray binary systems .
the detection of type i x - ray bursts @xcite and twin khz quasi - periodic oscillations in the x - ray power spectra @xcite strongly suggests that the compact object in the system is a neutron star .
the nature of the companion star is more debatable : some evidence is pointing towards a low - mass object @xcite , while other towards a high - mass one @xcite .
if the companion star is a supergiant , as recent evidence seems to suggest @xcite , then this is at odds with the low magnetic field that must be present in the system ( as inferred from type i x - ray bursts observations , for instance ) .
thus circinus x-1 appears to be a very exotic system , harbouring a young neutron star with a low magnetic field .
neutron star x - ray binaries ( xrbs ) are classified according to their x - ray spectral and timing properties as z and atoll sources .
the z class comprises seven objects accreting near the eddington limit .
all of them have been detected in the radio band and show variable emission at cm wavelengths .
the atoll type form the largest class of neutron star xrbs and have x - ray properties similar to black hole xrbs . due to their systematic lower radio fluxes , one order of magnitude less radio loud than z sources , only five of them
have been detected in the radio band .
the behaviour of circinus x-1 is puzzling , exhibiting similarities with z sources ( e.g. @xcite ) , but at times showing characteristics reminiscent of atoll sources @xcite .
the orbit of the binary system probably has a relatively high eccentricity ( e @xmath4 0.4 - 0.9 ) that produces variations in the accretion rate on the compact object @xcite .
this offers an explanation for the periodic flares ( [email protected] d ) observed in x - ray @xcite , infrared @xcite and radio @xcite , which are interpreted as enhanced accretion near the periastron passage .
circinus x-1 is associated with an arcmin scale synchrotron nebula @xcite which is probably powered by the jet originating close to the binary system @xcite . the jet is observed in radio at arcmin as well as arcsec scales and was recently detected in x - ray band by chandra ( @xcite ; soleri et al . 2008 ) .
observations at cm wavelengths offered evidence for the presence of a relativistic outflow aligned very close to the line of sight @xcite . however , x - ray @xcite and optical @xcite spectroscopy seem to favour a significantly higher angle between the jet axis and the line of sight .
besides the flux modulation at the orbital period , circinus x-1 also exhibits ` secular ' changes , documented in x - ray and radio .
the asm / rxte ( all sky monitor / rossi x - ray timing explorer ) light curve ( 2 - 10 kev ) shows that after a quasi - constant x - ray output , the flux decreased starting from 2000 up to now by one order of magnitude ( fig . 1 , left , bottom panel ) .
the picture is even more dramatic in radio . in the 1970s-1980s
the radio flares reached up to 1 jy ( e.g. @xcite ) , then they were detected at mjy level only @xcite , until recently when a relatively increased radio activity from circinus x-1 was noticed @xcite .
this offered the opportunity to obtain new vlbi ( very large baseline interferometry ) observations of the system within the framework of the first southern e - vlbi ( electronic vlbi ) experiment @xcite , after almost 25 years since the last successful vlbi observations @xcite .
circinus x-1 complex is very close to the supernova remnant snr g321.9 - 0.3 on the plane of the sky .
optical observations @xcite ruled out the possibility that the two objects are physically associated as it was assumed for a long time @xcite .
we have observed the x - ray binary circinus x-1 during a time interval of almost 10 years , between 1996 and 2006 , simultaneously at 4.8 and 8.6 ghz with the australia telescope compact array ( atca ) . as primary calibrators we used pks j0825 - 5010 ( pks b0823 - 500 ) for epochs 15 and 37 ( see table 1 ) and pks j1939 - 6342 ( pks b1934 - 638 ) for the rest .
pmn j1524 - 5903 ( b1520 - 58 ) was used as secondary calibrator for the entire data set .
its positional uncertainty is , according to the atca calibrator catalog , between 100 and 250 mas . throughout the paper we assumed conservatively the above upper limit as the systematic error in our determinations of positions .
standard calibration techniques were applied using the software miriad @xcite .
the observations were carried out at different spatial resolutions , although very similar ( see the array configuration column in table 1 and the atca website ) . to aid an homogeneous analysis , in obtaining the final radio images we have used the same restoring beams ( size and orientation ) for the entire data set at each of the two observed frequencies , for each weighting scheme , natural and uniform .
the size of these beams is in general less than 5 percent different than the sizes of the default synthesized beams .
the exceptions are the observations from epoch 14 when the difference was up to 40 percents ( array configuration 1.5d ) and most of the observations from 1996 july for which the default synthesized beams were very elongated due to the short observational runs . because of this limitation the radio images from the 1996 july epochs 3 - 7 , 9 and 10 are very likely to contain artifacts .
however , we note that the calibration process was successful and the _ uv _ data should not be significantly affected .
the only special case is epoch 10 , for which the phase - referencing calibration was only partially successful , but we are including it in the analysis for completeness .
the flux density within a run varied sometimes by up to 30 percent . in general the variations were smooth over the duration of the observing sessions and tests made by selecting only parts of the data showed that the maps are not affected significantly by this behaviour .
however , when a radio flare happened during an observational run the flux density changed by almost a factor 2 on timescales of hours . in these few cases artifacts could be present in the radio maps ( see section 3 , paragraph 4 ) . as mentioned in the previous section
, circinus x-1 periodically flares at radio wavelengths .
we used the ephemeris derived from the onset times of 21 well observed such outbursts spanning the past 28 years @xcite to illustrate in fig .
2 the orbital phase coverage of our observations . during the 10 years of data
we have swept the entire orbit of the system , even though more often than not only once for a given orbital phase , and gaps in the coverage are present . | we present observations of the neutron star x - ray binary and relativistic jet source circinus x-1 made at 4.8 and 8.6 ghz with the australia telescope compact array during a time interval of almost 10 years .
the system shows significant variations in the morphology and brightness of the radio features on all timescales from days to years . using the time delay between the successive brightening of the different components of the radio emission we were able to provide further evidence for the relativistic nature of the arcsec scale outflow , with an apparent velocity @xmath0 12 .
no compelling evidence for an evolution of the orientation of the jet axis was found .
we also place an upper limit on the proper motion of the system which is consistent with previous optical studies .
besides the previously reported radio flares close to the orbital phase 0.0 ( interpreted as enhanced accretion at periastron passage ) , we also identified outbursts with similar properties near the orbital phase 0.5 .
the global spectral index revealed a preferentially steep spectrum over the entire period of monitoring with a mean value and standard deviation @xmath1=-0.9 @xmath2 0.6 ( @xmath3 ) , which became significantly flatter during the outbursts .
polarization was detected in one third of the epochs and in one case faraday rotation close to the core of the system was measured .
[ firstpage ] accretion , accretion discs stars : individual : circinus x-1 ism : jets and outflows radio continuum : stars radiation mechanisms : non - thermal x - rays : stars . |
in this paper we construct the abelian chern - simons topological quantum field theory directly from the theory of classical theta functions .
it has been known for years , in the context of abelian chern - simons theory , that classical theta functions are related to low dimensional topology @xcite , @xcite .
abelian chern - simons theory is considerably simpler than its non - abelian counterparts , and has been studied thoroughly ( see for example @xcite and @xcite ) . here
we do not start with abelian chern - simons theory , but instead give a direct construction of the associated topological quantum field theory based on the theory of classical theta functions and using skein modules .
we consider classical theta functions in the representation theoretic point of view introduced by andr weil @xcite . from this point of view
, the space of theta functions is endowed with an action of a finite heisenberg group ( the schrdinger representation ) , which induces , via a stone - von neumann theorem , the hermite - jacobi action of the modular group .
all this structure is what we shall mean by the theory of theta functions .
we show how the finite heisenberg group , or rather its group algebra and its schrdinger representation on the space of theta functions , lead to algebras of curves on surfaces and their actions on spaces of curves in handlebodies .
these notions are formalized using skein modules .
the hermite - jacobi representation of the modular group on theta functions is a discrete analogue of the metaplectic representation .
the modular group acts by automorphisms that are a particular case of the fourier - mukai transform ; in this paper we will refer to them as discrete fourier transforms .
we will show that discrete fourier transforms can be expressed as linear combinations of curves .
a careful analysis of their structure and of their relationship to the schrdinger representation yields the reshetikhin - turaev formula @xcite of invariants of 3-dimensional manifolds , for @xmath0 chern - simons theory . as a corollary of our point of view
we obtain an explanation of why the composition of discrete fourier transforms and the non - additivity of the signature of 4-dimensional manifolds obey the same formula .
the paper uses results and terminology from the theory of theta functions , quantum mechanics , and low dimensional topology . to make it accessible to different audiences we include a fair amount of detail .
in section 2 we review the theory of classical theta functions on the jacobian variety of a surface .
the action of the finite heisenberg group on theta functions can be defined by translations in the associated line bundle , as it is usually done in algebraic geometry , or via weyl quantization of the jacobian variety in the holomorphic polarization .
we follow the second approach , because in our opinion it explains better the combinatorial picture of this action .
in fact it has been discovered in recent years that chern - simons theory is related to weyl quantization @xcite , @xcite , and this was the starting point of our paper .
section 3 describes the weyl quantization of the jacobian in the real polarization , and compares it to that in the holomorphic polarization . building on this
, the next section exhibits the combinatorial picture of theta functions , the action of the finite heisenberg group , and the hermite - jacobi action . up to this point topology has not come into play , but in section 5 we show that the combinatorial picture of the theory of theta functions is topological in nature .
we reformulate it using algebras of curves on surfaces , together with their action on skeins of curves in handlebodies which are associated to the linking number . in section 6
we derive a formula for the discrete fourier transform as a skein .
this formula is interpreted in terms of surgery in the cylinder over the surface .
section 7 analizes the exact egorov identity which relates the hermite - jacobi action to the schrdinger representation .
this analysis shows that the topological operation of handle slides is allowed over the skeins that represent discrete fourier transforms , and this yields in the next section to the reshetikhin - turaev formula for invariants of 3-dimensional manifolds , and to a @xmath0-topological quantum field theory .
we point out that the above - mentioned formula was introduced in an ad - hoc manner by its authors in @xcite , however this paper shows how to arrive at this formula in a natural way , at least in the context of the @xmath0-theory .
section 9 shows how to associate to the discrete fourier transform a 4-dimensional manifold .
this will explain why the cocycle of the hermite - jacobi action is related to that governing the non - additivity of the signature of 4-dimensional manifolds @xcite .
section 10 should be taken as a conclusion ; it brings everything to the context of chern - simons theory .
we start with a closed genus @xmath1 riemann surface @xmath2 , and consider a canonical basis @xmath3 for @xmath4 , like the one shown in figure [ homology ] . to it
we associate a basis in the space of holomorphic differential @xmath5-forms @xmath6 , defined by the conditions @xmath7 , @xmath8 . the matrix @xmath9 with entries @xmath10 is symmetric with positive definite imaginary part .
this means that if @xmath11 , then @xmath12 , @xmath13 and @xmath14 .
the @xmath15 matrix @xmath16 is called the period matrix of @xmath2 , its columns @xmath17 , called periods , generate a lattice @xmath18 in @xmath19 .
the complex torus @xmath20 is the _
jacobian variety _ of @xmath2 .
the map @xmath21 induces a homeomorphism @xmath22 .
the complex coordinates @xmath23 on @xmath24 are those inherited from @xmath25 .
we introduce real coordinates @xmath26 by imposing @xmath27 .
a fundamental domain for the period lattice , in terms of the @xmath28 coordinates , is simply @xmath29^{2g}\}$ ] .
moreover , @xmath24 has a canonical symplectic form , which in the @xmath28-coordinates is given by @xmath30 the jacobian variety with the complex structure and symplectic form @xmath31 is a khler manifold .
the symplectic form induces a poisson bracket for smooth functions on the jacobian , given by @xmath32 , where @xmath33 denotes the hamiltonian vector field defined by @xmath34 . the _ classical theta functions _
show up when quantizing @xmath24 in the complex polarization , in the direction of this poisson bracket .
for the purpose of this paper , we perform this quantization in the case where planck s constant is the reciprocal of an _ even _ positive integer : @xmath35 where @xmath36 , @xmath37 .
the hilbert space of the quantization consists of the holomorphic sections of a line bundle obtained as the tensor product of a line bundle with curvature @xmath38 and the square root of the canonical line bundle .
the latter is trivial for the complex torus and we ignore it .
the line bundle with curvature @xmath38 is the tensor product of a flat line bundle and the line bundle defined by the cocycle @xmath39 , @xmath40 @xmath41 .
( see e.g. 4.1.2 of @xcite for a discussion of how this cocycle gives rise to a line bundle with curvature @xmath38 . )
the flat line bundle does not add anything to the discussion , it only complicates computations , thus we choose the trivial one . as such
, the hilbert space can be identified with the space of entire functions on @xmath25 satisfying the periodicity conditions @xmath42 we denote this space by @xmath43 ; its elements are called classical theta functions .
a basis of @xmath43 consists of the theta series @xmath44 } , \quad \mu\in \{0,1\ldots , n-1\}^g.\end{aligned}\ ] ] the definition of theta series will be extended for convenience to all @xmath45 , by @xmath46 for any @xmath47 .
hence the index @xmath48 can be taken in @xmath49 .
the inner product that makes the theta series into an orthonormal basis is @xmath50^{2 g } } f(x , y)\overline{g(x , y ) } e^{-2\pi n y^tyy}dxdy .\end{aligned}\ ] ] to define the operators of the quantization , we use the weyl quantization method .
this quantization method can be defined only on complex vector spaces , the jacobian variety is the quotient of such a space by a discrete group , and the quantization method goes through . as such , the operator @xmath51 associated to a function @xmath52 on @xmath24 is the toeplitz operator with symbol @xmath53 , @xcite , where @xmath54 is the laplacian on functions , @xmath55 on a general riemannian manifold this operator is given in local coordinates by the formula @xmath56 where @xmath57 is the metric and @xmath58 . in the khler case , if the khler form is given in holomorphic coordinates by @xmath59 then @xmath60 where @xmath61 .
in our situation , in the coordinates @xmath62 , @xmath41 , one computes that @xmath63 and therefore @xmath64 ( recall that @xmath65 is the imaginary part of the matrix @xmath9 )
. for weyl quantization one introduces a factor of @xmath66 in front of the operator . as such , the laplace ( or rather laplace - beltrami ) operator @xmath67 is equal to @xmath68_j \left[(i_g - iy^{-1}x)\nabla_x+iy^{-1}\nabla_y\right]_k.\ ] ] ( a word about the notation being used : @xmath69 represents the usual ( column ) vector of partial derivatives in the indicated variables , so that each object in the square brackets is a column vector of partial derivatives .
the subindices @xmath70 are the corresponding components of those vectors . ) a tedious calculation that we omit results in the following formula for the laplacian in the @xmath28 coordinates : @xmath71 we will only need to apply @xmath54 explicitly to exponentials , as part of the proof of the following basic proposition .
note that the exponential function @xmath72 defines a function on the jacobian provided @xmath73 .
[ weylquantization ] the weyl quantization of the exponentials is given by @xmath74 let us introduce some useful notation local to the proof .
note that @xmath75 and @xmath9 are fixed throughout .
1 . @xmath76 , 2 . for @xmath77 and @xmath78 , @xmath79 .
3 . @xmath80 4 .
. with these notations , in the @xmath82 coordinates @xmath83 we first compute the matrix coefficients of the toeplitz operator with symbol @xmath84 , that is @xmath85^{2 g } } e_{p , q}(x , y)\theta_\mu(x , y ) \overline{\theta_\nu(x , y)}\ , e^{-2\pi ny^tyy}\ , dxdy.\end{aligned}\ ] ] then a calculation shows that @xmath86.\end{aligned}\ ] ] the integral over @xmath87^g$ ] of the @xmath88 term will be non - zero iff @xmath89 in which case the integral will be equal to one .
therefore @xmath90 unless @xmath91 = [ \mu + p],\ ] ] where the brackets represent equivalence classes in @xmath92 .
this shows that the toeplitz operator with multiplier @xmath84 maps @xmath93 to a scalar times @xmath94 .
we now compute the scalar . taking @xmath48 in the fundamental domain @xmath95 for @xmath92
, there is a unique representative , @xmath96 , of @xmath97 $ ] in the same domain .
this @xmath96 is of the form @xmath98 for a unique @xmath99 .
with respect to the previous notation , @xmath100 .
it follows that @xmath101^{g } } e\left [ q(n_\mu ) - \overline{q(m_\nu ) } \right.\\ & & \quad + \left.\bigl(\frac{q^t}{n}+n_\mu^t\pi -m^t_\nu\overline{\pi}\bigr)y+iy^tyy\right]\,dy,\end{aligned}\ ] ] where @xmath102 in the @xmath103th term . using that @xmath104 , one gets : @xmath105 and @xmath106 and so we can write @xmath107 \sum_{n\in{\mathbb z}^g } \int_{[0,1]^{g}}\,dy \\ & e\bigl [ i n_\mu^t y n_\mu - \frac{1}{n}p^t \overline{\pi}n_\mu + \bigl(\frac{1}{n } q^t+2in_\mu^ty -\frac{1}{n } p^t \overline{\pi } \bigr ) y + iy^tyy \bigr ] .
\end{split}\ ] ] making the change of variables @xmath108 in the summand @xmath103 , the argument of the function @xmath109 can be seen to be equal to @xmath110 since @xmath111 and @xmath103 are integer vectors , @xmath112 and therefore @xmath113 & e^{-2\pi i q^t\mu
/ n } \int_{{\mathbb r}^g } e^{-2\pi nw^tyw + 2\pi i\bigl(q^t - p^t\overline{\pi}\bigr ) w}\ , dw .
\end{split}\ ] ] a calculation of the integral ] yields that it is equal to @xmath114 and so @xmath115 the exponent on the right - hand side is @xmath116 times @xmath117y^{-1}[q-(x - iy)p ] \bigr ) \\ & & = 2iq^t\mu + ip^t(x - iy)p + \frac{1}{2}\bigl ( [ q^ty^{-1}-p^txy^{-1}+ip^t][q - xp+iyp]\bigr)\\ & & = 2iq^t\mu + ip^t(x - iy)p + \frac{1}{2}\bigl(q^ty^{-1}q-2q^ty^{-1}xp + 2iq^tp \\ & & \quad + p^txy^{-1}xp - 2ip^txp - p^typ\bigr ) \\ & & = 2iq^t\mu + iq^tp+ \frac{1}{2}{\mathcal r}\end{aligned}\ ] ] where @xmath118 that is , @xmath119 on the other hand , it is easy to check that @xmath120 and therefore @xmath121
so that , by ( [ matcoeff ] ) @xmath122 as desired . let us focus on the group of quantized exponentials .
first note that the symplectic form @xmath31 induces a nondegenerate bilinear form on @xmath123 , which we denote also by @xmath31 , given by @xmath124 as a corollary of proposition [ weylquantization ] we obtain the following result .
[ multofops ] quantized exponentials satisfy the multiplication rule @xmath125 this prompts us to define the heisenberg group @xmath126 with multiplication @xmath127 this group is a @xmath128-extension of @xmath129 , with the standard inclusion of @xmath129 into it given by @xmath130 the map @xmath131 defines a representation of @xmath132 on theta functions . to make this representation faithful
, we factor it by its kernel .
[ kernel ] the set of elements in @xmath132 that act on theta functions as identity operators is the normal subgroup consisting of the @xmath75th powers of elements of the form @xmath133 with @xmath134 even .
the quotient group is isomorphic to a finite heisenberg group . by proposition [ weylquantization ] , @xmath135 for @xmath133 to act as the identity operator
, we should have @xmath136 for all @xmath137 .
consequently , @xmath138 should be in @xmath139 .
then @xmath140 is a multiple of @xmath75 , so the coefficient @xmath141 equals @xmath142 .
this coefficient should be equal to @xmath5 .
for @xmath143 this implies that @xmath144 should be an even multiple of @xmath75 .
but then by varying @xmath48 we conclude that @xmath111 is a multiple of @xmath75 .
because @xmath75 is even , it follows that @xmath140 is an even multiple of @xmath75 , and consequently @xmath134 is an even multiple of @xmath75 .
thus any element in the kernel of the representation must belong to @xmath145 .
it is easy to see that any element of this form is in the kernel .
these are precisely the elements of the form @xmath146 with @xmath134 even .
the quotient of @xmath132 by the kernel of the representation is a @xmath147-extension of the finite abelian group @xmath148 , thus is a finite heisenberg group .
this group is isomorphic to @xmath149 with the multiplication rule @xmath150 the isomorphism is induced by the map @xmath151 , @xmath152 we denote by @xmath153 this finite heisenberg group and by @xmath154 the image of @xmath133 in it .
the representation of @xmath153 on the space of theta functions is called the _
schrdinger representation_. it is an analogue , for the case of the @xmath155-dimensional torus , of the standard schrdinger representation of the heisenberg group with real entries on @xmath156 .
in particular we have @xmath157 [ svn ] ( stone - von neumann ) the schrdinger representation of @xmath153 is the _ unique _
irreducible unitary representation of this group with the property that @xmath158 acts as @xmath159 for all @xmath160 .
let @xmath161 , @xmath162 , @xmath41 , @xmath163
. then @xmath164 , @xmath165 if @xmath166 , @xmath167 , @xmath168 , @xmath169 , @xmath170 , for all @xmath171 , and @xmath172 for all @xmath173 .
because @xmath174 commute pairwise , they have a common eigenvector @xmath175 . and because @xmath176 for all @xmath173 , the eigenvalues @xmath177 of @xmath175 with respect to the @xmath174 are roots of unity .
the equalities @xmath178 show that by applying @xmath179 s repeatedly we can produce an eigenvector @xmath180 of the commuting system @xmath174 whose eigenvalues are all equal to @xmath5 .
the irreducible representation is spanned by the vectors @xmath181 , @xmath182 .
any such vector is an eigenvector of the system @xmath174 , with eigenvalues respectively @xmath183 @xmath184 @xmath185 . so these vectors are linearly independent and form a basis of the irreducible representation .
it is not hard to see that the action of @xmath153 on the vector space spanned by these vectors is the schrdinger representation .
[ allspace ] the operators @xmath186 , @xmath187 form a basis of the space of linear operators on @xmath43 . for simplicity
, we will show that the operators @xmath188 form a basis .
denote by @xmath189 the respective matrices of these operators in the basis @xmath190 . for a fixed @xmath138 ,
the nonzero entries of the matrices @xmath189 , @xmath191 are precisely those in the slots @xmath192 , with @xmath193 ( here @xmath194 is taken modulo @xmath75 ) .
if we vary @xmath195 and @xmath111 and arrange these nonzero entries in a matrix , we obtain the @xmath1th power of a vandermonde matrix , which is nonsingular . we conclude that for fixed @xmath138 , the matrices @xmath189 , @xmath191 form a basis for the vector space of matrices with nonzero entries in the slots of the form @xmath192 .
varying @xmath138 , we obtain the desired conclusion .
the algebra @xmath196 of linear operators on the space of theta functions is isomorphic to the algebra obtained by factoring @xmath197 $ ] by the relation @xmath198 .
let us now recall the action of the _ modular group _ on theta functions .
the modular group , known also as the _ mapping class group _ , of a simple closed surface @xmath2 is the quotient of the group of homemorphisms of @xmath2 by the subgroup of homeomorphisms that are isotopic to the identity map .
it is at this point where it is essential that @xmath75 is _
even_. the mapping class group acts on the jacobian in the following way .
an element @xmath199 of this group induces a linear automorphism @xmath200 of @xmath4 .
the matrix of @xmath200 has integer entries , determinant @xmath5 , and satisfies @xmath201 , where @xmath202 is the intersection form in @xmath203 .
as such , @xmath200 is a _
symplectic linear automorphism _ of @xmath203 , where the symplectic form is the intersection form .
identifying @xmath24 with @xmath204 , we see that @xmath200 induces a symplectomorphism @xmath205 of @xmath24 .
the map @xmath206 induces an action of the mapping class group of @xmath2 on the jacobian variety .
this action can be described explicitly as follows .
decompose @xmath200 into @xmath207 blocks as @xmath208 then @xmath205 maps the complex torus defined by the lattice @xmath16 and complex variable @xmath209 to the complex torus defined by the lattice @xmath210 and complex variable @xmath211 , where @xmath212 and @xmath213 .
this action of the mapping class group of the surface on the jacobian induces an action of the mapping class group on the finite heisenberg group by @xmath214.\end{aligned}\ ] ] the nature of this action is as follows : since @xmath199 induces a diffeomorphism on the jacobian , we can compose @xmath199 with an exponential and then quantize ; the resulting operator is as above .
we point out that if @xmath75 were not even , this action would be defined only for @xmath200 in the subgroup @xmath215 of the symplectic group ( this is because only for @xmath75 even is the kernel of the map @xmath216 defined in proposition [ kernel ] preserved under the action of @xmath200 ) . as a corollary of theorem [ svn ] , the representation of the finite heisenberg group on theta functions given by @xmath217 is equivalent to the schrdinger representation , hence there is an automorphism @xmath218 of @xmath43 that satisfies the _ exact egorov identity _ : @xmath219 ( compare with @xcite , theorem 2.15 , which is the analogous statement in quantum mechanics in euclidean space . ) moreover , by schur s lemma , @xmath218 is unique up to multiplication by a constant .
we thus have a projective representation of the mapping class group of the surface on the space of classical theta functions that statisfies with the action of the finite heisenberg group the exact egorov identity from ( [ egorov ] ) .
this is the finite dimensional counterpart of the metaplectic representation , called the _ hermite - jacobi action_. we emphasize that the action of the mapping class group of @xmath2 on theta functions factors through an action of the symplectic group @xmath220 . up to multiplication by a constant , @xmath221\theta_\mu^{\pi'}(z')\end{aligned}\ ] ]
( 5.6.3 ) in @xcite ) .
when the riemann surface is the complex torus obtained as the quotient of the complex plane by the integer lattice , and @xmath222 is the map induced by a @xmath223 rotation around the origin , then @xmath224 is the discrete fourier transform . in general
, @xmath218 is an instance of the general fourier - mukai transform , and like the metaplectic representation ( see @xcite ) , can be written as a composition of partial discrete fourier transforms .
for this reason , we will refer , throughout the paper , to @xmath218 as a _ discrete fourier transform_.
there is an abstract version of theta functions , the schrdinger representation , and the hermite - jacobi action , which can be introduced naturally using the weyl quantization of the jacobian variety in a real polarization .
we construct the vector space following the methods in @xcite and @xcite . like before we consider the line bundle of the quantization to be the tensor product of a line bundle with curvature @xmath225 and the square root of the canonical line bundle , the latter being trivial . for the line bundle with curvature
@xmath225 we choose the one from the previous section .
let @xmath69 be the connection on the line bundle of the quantization whose connection form @xmath226 satisfies the prequantization condition @xmath227 its pull - back @xmath228 under the covering map @xmath229 is given by @xmath230 let us consider the distribution @xmath216 spanned by the vector fields @xmath231 the hilbert space of the quantization consists of those sections of the line bundle that are covariantly constant with respect to @xmath69 along @xmath216 . if @xmath232 is such a section then the condition that @xmath232 is covariantly constant along @xmath216 translates to @xmath233 for this to give a well - defined section @xmath232 , the exponent must be periodic in @xmath234 , hence @xmath235 should be an integer .
this can only happen when @xmath236 , where @xmath134 is an integer .
we conclude that the sections that give the hilbert space @xmath237 are distributional sections which are covariantly constant and have the support in the bohr - sommerfeld variety @xmath238 the connected components of @xmath239 , called bohr - sommerfeld fibers , are the sets @xmath240 , @xmath241 .
the bohr - sommerfeld fibers are lagrangian submanifolds of the jacobian variety .
a basis of the hilbert space of the quantization consists of the distributional sections @xmath242 to quantize the smooth functions on the jacobian variety , we lift them to the plane and apply weyl quantization .
we also lift the polarization to the plane . with respect to this polarization
, the functions @xmath243 , @xmath244 and @xmath245 , @xmath41 , are quantized as the operators @xmath246 and @xmath247 of multiplication by the variable @xmath248 .
the only functions on the plane that factor to the torus are the ones that are periodic with respect to the lattice @xmath249 .
the space of smooth periodic functions in the plane has a dense subset spanned by the exponentials @xmath250 , @xmath251 . according to the weyl quantization scheme , such an exponential is quantized as the operator @xmath252 , where @xmath253 and @xmath254 .
this operator maps a state @xmath255 to @xmath256 .
if we now think of @xmath257 as a function on the jacobian variety , then its weyl quantization acts as @xmath258 we thus obtain the representation of the heisenberg group with integer entries given by @xmath259 which descends to the schrdinger representation of the finite heisenberg group .
the quantization in the complex polarization and that in the real polarization are related by a bargmann transform . to define it we extend the orthogonal projection @xmath260 from the space of square integrable sections ( with respect to the inner product ( [ innerproduct ] ) ) onto the space of holomorphic sections to distributional sections .
( proposition 3.2 in @xcite ) for every @xmath241 , @xmath261 , where @xmath262 is a non - zero constant . since @xcite states
this result without proof , we will sketch the proof below .
the reproducing kernel of the space of theta functions is @xmath263 the projection of the section @xmath264 on the space of theta function is therefore ( with the convention that @xmath265 and @xmath266 ) the projection of the section @xmath264 on the space of theta function is therefore @xmath267^g}\overline{\theta_\nu\left(x_w+\pi \frac{\mu}{n}\right)}e^{2\pi i\left(\frac{\mu}{n}\right)^tx_w}dx_w.\end{aligned}\ ] ] the integral equals @xmath268 } \int_{[0,1]^g}e^{2\pi in\left[\left(\frac{\mu-\nu}{n}-n\right)^tx_w \right]}dx_w.\end{aligned}\ ] ] this is equal to zero unless @xmath269 ( and @xmath270 ) , in which case we obtain @xmath271 the conclusion follows .
the projection @xmath260 can be composed on the left by a diagonal operator to obtain a unitary operator between the spaces @xmath237 and @xmath272 .
this operator intertwines the schrdinger representation in the real polarization with the schrdinger representation in the complex polarization .
the unitary operator of this corollary is the bargmann transform in this setting .
we see that if we represent the elements of the form @xmath273 of the finite heisenberg group as the points of the lattice @xmath274 on the torus @xmath275 , then the elements of the form @xmath276 translate the lagrangian manifolds , while those of the form @xmath277 keep them fixed , except that the maps act in the wrong direction .
to fix this we apply the reflection of the torus @xmath278 to transform the bohr - sommerfeld fibers into the lagrangian submanifolds @xmath279 a such , one can identify , for each @xmath48 , the lagrangian submanifold @xmath280 with the coset @xmath281 .
an easy check shows that the elements of the form @xmath276 and @xmath277 now act by the left regular action on these cosets .
the hilbert space consists of linear combinations of these cosets , which can be identified with functions on the set @xmath282 that are invariant under the right translation in the variable by elements of the form @xmath277 .
all that is left is to incorporate the constants @xmath158 , @xmath283 into the picture .
this will be done in the next section .
in this section we apply to the finite heisenberg group the standard construction that gives the schrdinger representation as left translations on the space of equivariant functions on the heisenberg group ( see for example @xcite ) .
we start with a lagrangian subspace of @xmath203 with respect to the intersection form , which for our purpose will be the space spanned by the elements @xmath284 of the canonical basis .
let @xmath285 be the intersection of this space with @xmath129 . under the standard inclusion @xmath286
, @xmath285 becomes an abelian subgroup of the heisenberg group with integer entries .
this factors to an abelian subgroup @xmath287 of @xmath153 .
let @xmath288 be the subgroup of @xmath153 containing both @xmath287 and the scalars @xmath289 .
then @xmath288 is a maximal abelian subgroup .
define the character @xmath290 , @xmath291 .
the hilbert space of the quantization is the space @xmath292 of functions @xmath293 satisfying the equivariance condition @xmath294 the norm of @xmath295 is @xmath296 in this formula , @xmath297 . because of finiteness , the set of functions on @xmath153 can be identified with the group algebra @xmath197 $ ] by the map @xmath298 the map
@xmath299\rightarrow { \mathbb c}[{{\bf h}}({\mathbb z}_n^g)]$ ] , @xmath300 is an orthogonal projector ( with respect to the counting measure ) onto the space @xmath301 .
the kernel of the map @xmath302 is spanned by all elements of the form @xmath303 with @xmath304 and @xmath305 . given
some @xmath306 $ ] , for an arbitrary @xmath307 the @xmath308-term in @xmath309 is @xmath310 if @xmath308 does not belong to @xmath311 , and is equal to @xmath312 if @xmath313 with @xmath314 .
the latter is also @xmath310 , which shows that the elements of the form @xmath303 are all in the kernel of @xmath260 . for the converse ,
note that modulo elements of the form @xmath303 , every @xmath304 is equivalent to @xmath315 , where @xmath316 is any element of @xmath317 .
summing up over all such @xmath316 and dividing by @xmath318 , we deduce that @xmath319 is equivalent to @xmath320 the latter is in the image of @xmath302 , and the conclusion follows .
because of this result , the space of theta functions can be identified with @xmath197 $ ] modulo the subspace spanned by all elements of the form @xmath321 where @xmath322 $ ] and @xmath323 .
theta functions are equivalence classes of the form @xmath324 where @xmath325 is in the group algebra of the finite heisenberg group .
there is a left action of finite heisenberg group on @xmath326 given by @xmath327 as an action on @xmath197 $ ] , this can be written as @xmath328 [ abstracttheta ] the map @xmath329 , @xmath241 defines a unitary map between the space of theta functions @xmath43 and @xmath292 , which intertwines the schrdinger representation and the left action of the finite heisenberg group .
it is not hard to see that @xmath43 and @xmath292 have the same dimension .
also , for @xmath330 , @xmath331 and @xmath332 are not equivalent modulo @xmath288 , hence the map from the statement is an isomorphism of finite dimensional spaces .
the norm of @xmath333 is one , hence this map is unitary .
we have @xmath334 and @xmath335 it follows that @xmath336 in agreement with the schrdinger representation ( [ schroedinger ] ) .
we rephrase the hermite - jacobi action in this setting . to this
end , fix an element @xmath199 of the mapping class group of the riemann surface @xmath2 .
let @xmath285 be the subgroup of @xmath129 associated to a canonical basis as explained in the beginning of this section , which determines the maximal abelian subgroup @xmath288 .
the image of the canonical basis through @xmath200 is also a canonical basis to which one associates @xmath337 and the maximal abelian subgroup @xmath338 .
the discrete fourier transform should map an element @xmath339 in @xmath340/\mbox{ker}(\pi_{\bf l})$ ] to @xmath341 in @xmath197 /\mbox{ker}(\pi_{h_*({\bf l})}).$ ] in this form the map is not well defined , since different representatives for the class of @xmath325 might yield different images .
the idea is to consider all possible liftings of @xmath325 and average them . for lifting the element
@xmath342 we use the section of @xmath302 defined as @xmath343 then , up to multiplication by a constant @xmath344 this formula identifies @xmath218 as a fourier - mukai transform ( see @xcite ) . that this map agrees with the one defined by ( [ hermitejacobi ] ) up to multiplication by a constant follows from schur s lemma , since both maps satisfy the exact egorov identity ( [ egorov ] ) .
the finite heisenberg group , the equivalence relation defined by the kernel of @xmath302 , and the schrdinger representation can be given topological interpretations , which we will explicate below . _
the heisenberg group_. the group @xmath345 is a @xmath128-extension of the abelian group @xmath129 .
the bilinear form @xmath31 from ( [ bilinform ] ) , which defines the cocycle of this extension , is the intersection form in @xmath129 .
cycles in @xmath129 can be represented by families of non - intersecting simple closed curves on the surface . as vector spaces
, we can identify @xmath346 $ ] with @xmath347h_1(\sigma_g , { \mathbb z})$ ] , where @xmath348 is an abstract variable whose exponent equals the last coordinate in the heisenberg group .
we start with an example on the torus .
here and throughout the paper we agree that @xmath349 denotes the curve of slope @xmath350 on the torus , oriented from the origin to the point @xmath349 when viewing the torus as a quotient of the plane by integer translations .
consider the multiplication @xmath351 shown graphically in figure [ torusmult ] .
the product curve @xmath352 can be obtained by cutting open the curves @xmath353 and @xmath354 at the crossing and joining the ends in such a way that the orientations agree .
this operation is called _ smoothing of the crossing_. it is easy to check that this works in general , for arbitrary surfaces , and so whenever we multiply two families of curves we introduce a coefficient of @xmath348 raised to the algebraic intersection number of the two families and we smoothen all crossings .
such algebras of curves , with multiplication related to polynomial invariants of knots , were first considered in @xcite .
the group @xmath153 is a quotient of @xmath355 , but can also be viewed as an extension of @xmath356 . as such , the elements of @xmath197 $ ] can be represented by families of non - intersecting simple closed curves on the surface with the convention that any @xmath75 parallel curves can be deleted .
the above observation applies to this case as well , provided that we set @xmath357 .
it follows that the space of linear operators @xmath196 can be represented as an algebra of simple closed curves on the surface with the convention that any @xmath75 parallel curves can be deleted .
the multiplication of two families of simple closed curves is defined by introducing a coefficient of @xmath358 raised to the algebraic intersection number of the two families and smoothing the crossings . _ theta functions .
_ next , we examine the space of theta functions , in its abstract framework from section [ sec:3 ] . to better understand the factorization modulo the kernel of @xmath302 , we look again at the torus .
if the canonical basis is @xmath353 and @xmath354 with @xmath359 , then an equivalence modulo @xmath360 is shown in figure [ equivmodl ] .
if we map the torus to the boundary of a solid torus in such a way that @xmath285 becomes null - homologous , then the first and last curves from figure [ equivmodl ] are homologous in the solid torus . to keep track of @xmath348
we apply a standard method in topology which consists of framing the curves .
a framed curve in a manifold is an embedding of an annulus .
one can think of the curve as being one of the boundary components of the annulus , and then the annulus itself keeps track of the number of ways that the curve twists around itself .
changing the framing by a full twist amounts to multiplying by @xmath348 or @xmath361 depending whether the twist is positive or negative .
then the equality from figure [ equivmodl ] holds in the solid torus .
it is not hard to check for a general surface @xmath2 the equivalence relation modulo @xmath360 is of this form in the handlebody bounded by @xmath2 in such a way that @xmath285 is null - homologous . _
the schrdinger representation .
_ one can frame the curves on @xmath2 by using the blackboard framing , namely by embedding the annulus in the surface .
as such , the schrdinger representation is the left action of an algebra of framed curves on a surface on the vector space of framed curves in the handlebody induced by the inclusion of the surface in the handlebody .
we will make this precise using the language of skein modules @xcite .
let @xmath362 be an orientable @xmath363-dimensional manifold , with a choice of orientation .
a framed link in @xmath362 is a smooth embedding of a disjoint union of finitely many annuli .
the embedded annuli are called link components .
we consider oriented framed links .
the orientation of a link component is an orientation of one of the circles that bound the annulus .
when @xmath362 is the cylinder over a surface , we represent framed links as oriented curves with the blackboard framing , meaning that the annulus giving the framing is always parallel to the surface .
let @xmath348 be a free variable , and consider the free @xmath347$]-module with basis the isotopy classes of framed oriented links in @xmath362 including the empty link @xmath364 .
let @xmath365 be the the submodule spanned by all elements of the form depicted in figure [ skeinrelations ] , where the two terms in each skein relation depict framed links that are identical except in an embedded ball , in which they look as shown .
the ball containing the crossing can be embedded in any possible way . to normalize , we add to @xmath365 the element consisting of the difference between the unknot in @xmath362 and the empty link @xmath364 .
recall that the unknot is an embedded circle that bounds an embedded disk in @xmath362 and whose framing annulus lies inside the disk .
the result of the factorization of the free @xmath347$]-module with basis the isotopy classes of framed oriented links by the submodule @xmath365 is called the _ linking number skein module _ of @xmath362 , and is denoted by @xmath366 .
the elements of @xmath366 are called _
skeins_. in other words , we are allowed to smoothen each crossing , to change the framing provided that we multiply by the appropriate power of @xmath348 , and to identify the unknot with the empty link .
the `` linking number '' in the name is motivated by the following : if @xmath362 is a @xmath363-dimensional sphere , then each link is , as an element of @xmath367 , equivalent to the empty link with the coefficient @xmath348 raised to the sum of the linking numbers of ordered pairs of components and the writhes of the components .
said differently , the skein relations from figure [ skeinrelations ] are used for computing the linking number .
these skein modules were first introduced by przytycki in @xcite as one - parameter deformations of the group algebra of @xmath368 .
przytycki computed them for all @xmath363-dimensional manifolds .
any trivial link component , namely any link component that bounds a disk disjoint from the rest of the link in such a way that the framing is an annulus inside the disk , can be deleted .
the proof of the lemma is given in figure [ triviallink ] .
if @xmath369 $ ] , the cylinder over a surface , then the identification @xmath370\cup\sigma_g\times [ 0,1]\approx \sigma\times [ 0,1]\end{aligned}\ ] ] obtained by gluing the boundary component @xmath371 in the first cylinder to the boundary component @xmath372 in the second cylinder by the identity map induces a multiplication on @xmath373)$ ] .
this turns @xmath373)$ ] into an algebra , called the _ linking number skein algebra_. as such , the product of two skeins is obtained by placing the first skein on top of the second .
the @xmath103th power of an oriented , framed , simple closed curve consists then of @xmath103 parallel copies of that curve .
we adopt the same terminology even if the manifold is not a cylinder , so @xmath374 stands for @xmath103 parallel copies of @xmath375 .
additionally , @xmath376 is obtained from @xmath377 by reversing orientation , and @xmath378 . for a fixed positive integer @xmath75
, we define the _ reduced linking number skein module _ of the manifold @xmath362 , denoted by @xmath379 , to be the quotient of @xmath366 obtained by imposing that @xmath380 for every oriented , framed , simple closed curve @xmath375 , and by setting @xmath381 . as a rule followed throughout the paper ,
whenever we talk about skein modules , @xmath348 is a free variable , while when we talk about reduced skein modules , @xmath348 is a root of unity .
moreover , the isomorphisms @xmath382 $ ] and @xmath383 allow us to identify the linking number skein module of @xmath384 with the set of laurent polynomials in @xmath348 and the reduced skein module with @xmath385 .
for a closed , oriented , genus @xmath1 surface @xmath2 , consider a canonical basis of its first homology @xmath3 ( see section [ sec:1 ] ) .
the basis elements are oriented simple closed curves on the surface , which we endow with the blackboard framing .
let @xmath386 be a genus @xmath1 handlebody and @xmath387 be a homeomorphism that maps @xmath388 to null homologous curves .
then @xmath389 is a basis of the first homology of the handlebody .
endow these curves in the handlebody with the framing they had on the surface .
[ skeinmodules ] ( a ) the linking number skein module @xmath373)$ ] is a free @xmath347$]-module with basis @xmath390 ( b ) the reduced linking number skein module @xmath391)$ ] is a finite dimensional vector space with basis @xmath392 ( c ) the linking number skein module @xmath393 is a free @xmath347$]-module with basis @xmath394 ( d ) the reduced linking number skein module @xmath395 is a finite dimensional vector space with basis @xmath396 parts ( a ) and ( c ) are consequences of a general result in @xcite ; we include their proof for sake of completeness .
+ ( a ) we bring all skeins in the blackboard framing ( of the surface ) .
a skein @xmath397 , where @xmath398 is an oriented framed link in the cylinder over the surface is equivalent modulo the skein relations to a skein @xmath399 where @xmath400 is an oriented framed link such that the projection of @xmath400 onto the surface has no crossings , and @xmath195 is the sum of the positive crossings of the projection of @xmath398 minus the sum of negative crossings . moreover , because any embedded ball can be isotoped to be a cylinder over a disk , any skein @xmath401 that is equivalent to @xmath397 and in which @xmath402 is a framed link with no crossings has the property that @xmath403 .
if @xmath398 is an oriented link with blackboard framing whose projection onto the surface has no crossings , and if it is null - homologous in @xmath404 , { \mathbb z})$ ] , then @xmath398 is equivalent modulo skein relations to the empty skein .
this follows from the computations in figure [ nullhomologous ] given the fact that the closed orientable surface @xmath2 can be decomposed into pairs of pants and annuli .
view @xmath2 as a sphere with @xmath1 punctured tori attached to it .
then @xmath398 is equivalent to a link @xmath400 that consists of simple closed curves on these tori , which therefore is of the form @xmath405 where @xmath406 denotes the curve of slope @xmath407 on the @xmath173th torus .
this last link is equivalent , modulo skein relations , to @xmath408 the conclusion follows once we notice that the homology class of a link in an arbitrary manifold does not change when we apply the skein relation .
\(b ) like in part ( a ) , every oriented framed link is equivalent , modulo skein relations , to a skein of the form @xmath409 , where @xmath400 is an oriented link with the blackboard framing and whose projection on the surface has no crossings , with the exponent @xmath195 uniquely defined .
using the aditional factorization relation we find that such a link is further equivalent to one of the vectors from the statement .
hence these vectors form a system of generators .
that they form a basis follows from the fact that the skein relation and the additional factorization relation preserve the homology class of the link in @xmath356 .
one should remark that when resolving the crossings of a strand by @xmath75 parallel strands one obtains a factor of @xmath410 , hence the factorization relation @xmath380 induces only relations of the form @xmath411 at the level of scalars .
parts ( c ) and ( d ) are similar if we view the genus @xmath1 handlebody as the cylinder over a disk with @xmath1 punctures .
now we are at the point where we can phrase the theory of classical theta functions in the language of skein modules .
[ linkingheisenberg ] the algebras @xmath373)$ ] and @xmath346 $ ] are isomorphic , with the isomorphism defined by the map @xmath412,k+\ # [ \gamma]).\end{aligned}\ ] ] where @xmath375 ranges over all skeins represented by oriented simple closed curves on @xmath2 ( with the blackboard framing ) , @xmath413 $ ] is the homology class of this curve in @xmath414 , and @xmath415 .
this isomorphism factors to an algebra isomorphism of @xmath416)$ ] and @xmath196 . that the specified map is a linear isomorphism follows from proposition [ skeinmodules ] ( a ) .
it is straightforward to check that the multiplication rule is the same . in the reduced case
, the conclusion follows from proposition [ skeinmodules ] ( b )
. said differently , the map @xmath417 for all @xmath418 , defines an isomorphism of the algebras @xmath419)$ ] and @xmath346 $ ] and the map @xmath420 for all @xmath421 defines an isomorphism of the algebras @xmath422)$ ] and @xmath196 .
we use the maps defined in theorem [ linkingheisenberg ] to identify @xmath419)$ ] with @xmath346 $ ] and @xmath423)$ ] with @xmath196 .
the linking number skein module of a @xmath363-manifold @xmath362 with boundary is a module over the skein algebra of a boundary component @xmath2 .
the module structure is induced by the identification @xmath370\cup m\approx m\end{aligned}\ ] ] where @xmath424 $ ] is glued to @xmath362 along @xmath371 by the identity map .
this means that the module structure is induced by identifying @xmath424 $ ] with a regular neighborhood of the boundary of @xmath362 .
the product of a skein in a regular neighborhood of the boundary and a skein in the interior is the union of the two skeins .
this module structure descends to relative skein modules . in particular @xmath373)$ ]
acts on the left on @xmath393 , with the action induced by the homeomorpism @xmath387 , and this action descends to relative skein modules . before we state the next result , let us point out that @xmath395 is obtained by factoring @xmath422)$ ] by isotopies in the handlebody . also , in view of section [ sec:3 ] , @xmath43 is the quotient of @xmath197 $ ] by the kernel of @xmath302 ( where @xmath425 ) , and hence @xmath43 is a quotient of @xmath196 .
additionally , we will view the handlebody @xmath386 as the cylinder over the 2-dimensional disk with @xmath1 holes @xmath426 .
[ thetaskein ] the isomorphism between @xmath391)$ ] and @xmath196 given in theorem [ linkingheisenberg ] factors to the isomorphism between @xmath395 and @xmath43 given by @xmath427},\end{aligned}\ ] ] where @xmath375 ranges among all oriented simple closed curves in @xmath426 with the blackboard framing and @xmath413 $ ] is the homology class of @xmath375 in @xmath428 .
this isomorphism intertwines the left action of @xmath422)$ ] on @xmath395 and the schrdinger representation .
the map is a vector space isomorphism by proposition [ skeinmodules ] .
one can check that the left action of the skein algebra of the cylinder over the surface on the skein module of the handlebody is the same as the one from propositions [ weylquantization ] and [ abstracttheta ] .
the isomorphism between the reduced skein module of the handlebody and the space of theta functions is given explicitly by @xmath429 in view of theorem [ thetaskein ] we endow @xmath395 with the hilbert space structure of the space of theta functions .
now we turn our attention to the discrete fourier transform , and translate in topological language formula ( [ fouriertransform ] ) .
let @xmath199 be an element of the mapping class group of the surface @xmath2 .
the action of the mapping class group on the finite heisenberg group described in section [ sec:2 ] becomes the action on skeins in the cylinder over the surface given by @xmath430 where @xmath431 is obtained by replacing each framed curve of the skein @xmath432 by its image through the homeomorphism @xmath199 .
consider @xmath433 and @xmath434 two homeomorphisms of @xmath2 onto the boundary of the handlebody @xmath386 such that @xmath435 .
these homeomorphisms extend to embeddings of @xmath436 $ ] into @xmath386 which we denote by @xmath433 and @xmath434 as well .
the homeomorphisms @xmath433 and @xmath434 define the action of the reduced skein module of the cylinder over the surface on the reduced skein module of the handlebody in two different ways , i.e. they give two different constructions of the schrdinger representations . by the stone - von neumann theorem , these are unitary equivalent ; they are related by the isomorphism @xmath218 .
we want to give @xmath218 a topological definition .
to this end , let us take a closer look at the lifting map @xmath437 defined in ( [ lift ] ) .
first , it is standard to remark that one should only average over @xmath438 , hence @xmath439 if @xmath440 , then , as a skein , @xmath319 is of the form @xmath441 where @xmath375 is a framed oriented curve on @xmath442 and @xmath134 is an integer .
the element @xmath443 is just this skein viewed as lying inside the handlebody ; it consists of @xmath134 parallel framed oriented curves in @xmath386 . on the other hand , as a skein
, @xmath308 is of the form @xmath444 , and as such , the product @xmath445 becomes after smoothing all crossings another lift of the skein @xmath446 to the boundary obtained by lifting @xmath375 to the boundary and then taking @xmath134 parallel copies .
such a lift is obtained by pushing @xmath446 inside a regular neighborhood of the boundary and then viewing it as an element in @xmath423)$ ] .
when @xmath308 ranges over all @xmath287 we obtain all possible lifts of @xmath446 to the boundary obtained by pushing @xmath377 to the boundary and then taking @xmath134 parallel copies .
[ fourier ] for a skein of the form @xmath441 in @xmath395 , where @xmath377 is a curve in @xmath386 and @xmath134 a positive integer , consider all possible liftings to @xmath447)$ ] using @xmath433 , obtained by pushing the curve @xmath375 to the boundary and then taking @xmath134 parallel copies . take the average of these liftings and map the average by @xmath434 to @xmath448 .
this defines a linear endomorphism of @xmath449 which is , up to multiplication by a constant , the discrete fourier transform @xmath218 .
the map defined this way intertwines the schrdinger representations defined by @xmath433 and @xmath434 , so the theorem is a consequence of the stone - von neumann theorem .
_ example : _ we will exemplify this by showing how the @xmath239-map on the torus acts on the theta series @xmath450}\end{aligned}\ ] ] ( in this case @xmath9 is a just a complex number with positive imaginary part ) . this theta series is represented in the solid torus by the curve shown in figure [ thetas ] . the @xmath75 linearly independent liftings of this curve to the boundary are shown in figure [ fouriers ] .
the @xmath239-map sends these to those in figure [ fourier2s ] , which , after being pushed inside the solid torus , become the skeins from figure [ fourier3s ] .
note that in each skein the arrow points the opposite way as for @xmath451 . using the identity @xmath380
, we can replace @xmath173 parallel strands by @xmath452 parallel strands with opposite orientation .
hence these skeins are @xmath453 , @xmath454 ( note also that @xmath455 ) . taking the average
we obtain @xmath456 which is , up to a multiplication by a constant , the standard discrete fourier transform of @xmath451 .
as a consequence of proposition [ allspace ] , @xmath218 can be represented as an element in @xmath197 $ ] .
furthermore , theorem [ thetaskein ] implies that @xmath218 can be represented as left multiplication by a skein @xmath457 in @xmath391)$ ] .
the skein @xmath457 is unique up to a multiplication by a constant .
we wish to find an explicit formula for it .
theorem [ thetaskein ] implies that the action of the group algebra of the finite heisenberg group can be represented as left multiplication by skeins .
using this fact , the exact egorov identity ( [ egorov ] ) translates to @xmath458 for all skeins @xmath459)$ ] . by the lickorish twist theorem ( chapter 9 in @xcite ) , every homeomorphism of the surface @xmath2 is isotopic to a product of dehn twists along the @xmath460 curves depicted in figure [ lickorish ] .
we recall that a dehn twist is the homemorphism obtained by cutting the surface along the curve , applying a full rotation on one side , then gluing back .
the curves from figure [ lickorish ] are nonseparating , and any two can be mapped into one another by a homeomorphism of the surface .
thus in order to understand @xmath457 in general , it suffices to consider the case @xmath461 , the positive dehn twist along the curve @xmath462 from figure [ homology ] .
the word positive means that after we cut the surface along @xmath462 we perform a full rotation of the part on the left in the direction of the arrow . because @xmath463 for all skeins that do not contain curves that intersect @xmath462 , it follows that @xmath464 commutes with all such skeins .
it also commutes with the multiples of @xmath462 ( viewed as a skein with the blackboard framing ) .
hence @xmath464 commutes with all operators of the form @xmath465 with @xmath466 , the first entry of @xmath138 , equal to @xmath310 .
this implies that @xmath467 to determine the coefficients @xmath468 , we write the exact egorov identity ( [ egorov ] ) for @xmath469 . since @xmath470 this identity reads @xmath471 we transform this further into @xmath472= \sum_{j=0}^{n-1}c_je^{-\frac{\pi i}{n}j}\exp(p_1+jq_1),\end{aligned}\ ] ] or , taking into account that @xmath473 , @xmath474 where @xmath475 .
it follows that @xmath476 for all @xmath173 .
normalizing so that @xmath464 is a unitary map and @xmath477 we obtain @xmath478 , and hence @xmath479 turning to the language of skein modules , and taking into account that any dehn twist is conjugate to the above twist by an element of the mapping class group , we conclude that if @xmath480 is a positive dehn twist along the simple closed curve @xmath377 on @xmath2 , then @xmath481 this is the same as the skein @xmath482 where @xmath483 is obtained by adding one full positive twist to the framing of @xmath377 ( the twist is positive in the sense that , as skeins , @xmath484 ) .
this skein has an interpretation in terms of surgery .
consider the curve @xmath485 $ ] with framing defined by the blackboard framing of @xmath483 on @xmath2 .
take a solid torus which is a regular neighborhood of the curve on whose boundary the framing determines two simple closed curves .
remove it from @xmath424 $ ] , then glue it back in by a homeomorphism that identifies its meridian ( the curve that is null - homologous ) to one of the curves determined by the framing .
this operation , called surgery , yields a manifold that is homeomorphic to @xmath424 $ ] , such that the restriction of the homeomorphism to @xmath371 is the identity map , and the restriction to @xmath372 is the dehn twist @xmath480 .
the reduced linking number skein module of the solid torus @xmath486 is , by proposition [ skeinmodules ] , an @xmath75-dimensional vector space with basis @xmath487 . alternately , it is the vector space of 1-dimensional theta functions with basis @xmath488 , where @xmath9 in this case is a complex number with positive imaginary part .
we introduce the element @xmath489 in @xmath490 . as a diagram , @xmath491 is the skein depicted in figure [ omegafigure ] multiplied by @xmath492 . if @xmath239 is the homemorphism on the torus induced by the @xmath223 rotation of the plane when viewing the torus as the quotient of the plane by the integer lattice , then @xmath493 .
in other words , @xmath491 is the ( standard ) discrete fourier transform of @xmath494 . for an arbitrary framed link @xmath398 ,
we denote by @xmath495 , the skein obtained by replacing each link component by @xmath491 . in other words , @xmath495 is the sum of framed links obtained from @xmath398 by replacing its components , in all possible ways , by @xmath496 parallel copies .
the skein @xmath491 is called the coloring of @xmath398 by @xmath491 . here
are two properties of @xmath491 that will be used in the sequel .
[ omegaproperties ] a ) the skein @xmath495 is independent of the orientations of the components of @xmath398 .
+ b ) the skein relation from figure [ omegalink ] holds , where the @xmath103 parallel strands point in the same direction .
\a ) the computation in figure [ nullhomologous ] implies that if we switch the orientation on the @xmath173 parallel curves that represent @xmath497 we obtain @xmath498 .
hence by changing the orientation on all curves that make up @xmath491 we obtain the skein @xmath499,\end{aligned}\ ] ] which is , again , @xmath491 .
\b ) when @xmath270 there is nothing to prove . if @xmath500 , then by resolving all crossings in the diagram we obtain @xmath103 vertical parallel strands with the coefficient @xmath501 where the signs in the exponents are either all positive , or all negative .
since @xmath502 is a primitive @xmath75th root of unity , this is equal to zero . hence the conclusion . up to this point
we have proved the following result : for a dehn twist @xmath480 , @xmath503 is the skein obtained by coloring the surgery framed curve @xmath483 of @xmath480 by @xmath491 . since by the lickorish twist theorem
every element @xmath199 of the mapping class group is a product of twists , it follows that @xmath199 can be represented by surgery on a framed link @xmath504 whose components are the surgery curves of the twists in the composition .
the mapping cylinder of @xmath199 is obtained as the surgery on @xmath424 $ ] along @xmath504 ; it is homeomorphic to @xmath505 $ ] , where the homeomorphism is the identity map on @xmath371 and @xmath199 on @xmath506 .
we obtain the following skein theoretic description of the discrete fourier transform induced by the map @xmath199 .
[ skeinfourier ] let @xmath199 be an element of the mapping class group of @xmath2 defined by surgery on the framed link @xmath504 in @xmath424 $ ] .
then the discrete fourier transform @xmath507 is given by @xmath508
next , we give the egorov identity a topological interpretation in terms of handle slides . for this
we look at its skein theoretical version ( [ skeinegorov ] ) .
we start again with an example on the torus .
_ example : _ for the positive twist @xmath480 and the operator represented by the curve @xmath353 the exact egorov identity reads @xmath509 which is described graphically in figure [ exegorov ] .
the diagram on the right is the same as the one from figure [ egorovslide ] . as such ,
the curve @xmath352 is obtained by sliding the curve @xmath353 along the surgery curve of the positive twist .
here is the detailed description of the operation of sliding a framed knot along another using a kirby band - sum move .
the _ slide _ of a framed knot @xmath510 along the framed knot @xmath511 , denoted by @xmath512 , is obtained as follows .
let @xmath513 be a copy of @xmath511 obtained by pushing @xmath511 in the direction of its framing .
take an embedded @xmath514 ^ 3 $ ] that is disjoint from @xmath515 , and @xmath513 except for the opposite faces @xmath516
^ 2\times \{i\}$ ] , @xmath517 and which are embedded in @xmath518 respectively @xmath519 .
@xmath520 is embedded in the annulus @xmath521 such that @xmath514\times \{j\}\times\{i\}$ ] is embedded in @xmath522 .
delete from @xmath523 the faces @xmath520 and add the faces @xmath524\times [ 0,1]$ ] .
the framed knot obtained this way is @xmath512 . saying it less rigorously but more intuitively , we cut the knots @xmath510 and @xmath513 and join together the two open strands by pulling them along the sides of an embedded rectangle ( band ) which does not intersect the knots .
figure [ slide ] shows the slide of a trefoil knot over a figure - eight knot , both with the blackboard framing .
when the knots are oriented , we perform the slide so that the orientations match .
one should point out that there are many ways in which one can slide one knot along the other , since the band that connects the two knots is not unique . for a closed curve @xmath525 in @xmath526 ,
the curve @xmath527 is obtained from @xmath525 by slides over the components of the surgery link of @xmath199 . indeed , if @xmath199 is the twist along the curve @xmath375 , with surgery curve @xmath483 , and if @xmath525 and @xmath375 intersect on @xmath2 at only one point , then @xmath528 . if the algebraic intersecton number of @xmath525 and @xmath375 is @xmath529 , then @xmath527 is obtained from @xmath525 by performing @xmath134 consecutive slides along @xmath483 .
the general case follows from the fact that @xmath199 is a product of twists .
it follows that the exact egorov identity is a particular case of slides of framed knots along components of the surgery link .
in fact , the exact egorov identity covers all cases of slides of one knot along another knot colored by @xmath491 , and we have [ handleslide ] let @xmath362 be a @xmath363-dimensional manifold , @xmath432 an arbitrary skein in @xmath379 and @xmath510 and @xmath511 two oriented framed knots in @xmath362 disjoint from @xmath432 . then , in @xmath379 , one has the equality @xmath530 however one does the band sum @xmath512 . the knots from the statement of the theorem should be understood as representing elements in @xmath379 .
isotope @xmath510 along the embedded @xmath514 ^ 3 $ ] that defines @xmath512 to a knot @xmath531 that intersects @xmath511 .
there is an embedded punctured torus @xmath532 in @xmath362 , disjoint from @xmath432 , which contains @xmath533 on its boundary , as shown in figure [ puncturedtorus ] a ) .
in fact , by looking at a neighborhood of this torus , we can find an embedded @xmath534 $ ] such that @xmath535 .
the boundary of this cylinder is a genus @xmath536 surface @xmath537 , and @xmath531 and @xmath513 lie in a punctured torus of this surface and intersect at exactly one point . by pushing off @xmath531 to a knot isotopic to @xmath510 ( which we identify with @xmath510 ) ,
we see that we can place @xmath510 and @xmath511 in an embedded @xmath538 $ ] such that @xmath539 and @xmath540 . by performing a twist in @xmath534 $ ]
we can change the framing of @xmath511 in such a way that @xmath510 and @xmath511 look inside @xmath538 $ ] like in figure [ puncturedtorus ] b ) . then @xmath510 is mapped to @xmath512 in @xmath541 by the dehn twist of @xmath537 with surgery diagram @xmath511 .
hence the equality @xmath542 in @xmath538 $ ] is just the exact egorov identity , which we know is true . by embedding @xmath538
$ ] in @xmath543 $ ] we conclude that this equality holds in @xmath534 $ ] . by applying the inverse of the twist , embedding @xmath534 $ ] in @xmath362 , and adding @xmath432 , we conclude that the identity from the statement holds as well .
the operation of sliding one knot along another is related to the surgery description of 3-dimensional manifolds ( for more details see @xcite ) .
let us recall the basic facts .
we use the standard notation @xmath544 for an @xmath103-dimensional ( unit ) ball and @xmath545 for the @xmath103-dimensional sphere .
every @xmath363-dimensional closed manifold is the boundary of a @xmath546-dimensional manifold obtained by adding @xmath536-handles @xmath547 to @xmath548 along the solid tori @xmath549 @xcite . on the boundary @xmath384 of @xmath548 , when adding a handle we remove a solid torus from @xmath384 ( the one which we identify with @xmath549 ) and glue back the solid torus @xmath550 .
the curve @xmath551 in the solid torus @xmath549 that was removed becomes the meridian ( i.e. null - homologous curve ) on the boundary of @xmath550 .
this procedure of constructing @xmath363-dimensional manifold is called dehn surgery with integer coefficients .
the curve @xmath551 together with the core of the solid torus @xmath549 bound an embedded annulus which defines a framed link component in @xmath384 .
as such , the information for dehn surgery with integer coefficients can be encoded in a framed link in @xmath384 .
if @xmath510 is a knot inside a 3-dimensional manifold @xmath362 obtained by surgery on @xmath384 and if the framed knot @xmath511 is a component of the surgery link , then @xmath512 is the slide of @xmath510 over the @xmath536-handle corresponding to @xmath511 . indeed , when we slide @xmath510 along the handle we push one arc close to @xmath511 , then move it to the other side of the handle by pushing it through the meridinal disk of the surgery solid torus .
the meridian of this solid torus is parallel to the knot @xmath511 ( when viewed in @xmath552 , so by sliding @xmath510 across the handle we obtain @xmath512 . in particular , the operation of sliding one 2-handle over another corresponds to sliding one link component of the surgery link along another . in conclusion , we can say that the egorov identity allows handle - slides along surgery link components colored by @xmath491 .
we will make use of this fact in the next section .
theorem [ handleslide ] has two direct consequences : * the definition of a topological invariant for closed @xmath363-dimensional manifolds , * the existence of an isomorphism between the reduced linking number skein modules of @xmath363-dimensional manifolds with homeomorphic boundaries .
we have seen in the previous section that handle - slides correspond to changing the presentation of a @xmath363-dimensional manifold as surgery on a framed link .
kirby s theorem @xcite states that two framed link diagrams represent the same @xmath363-dimensional manifold if they can be transformed into one another by a sequence of isotopies , handle slides and additions / deletions of the trivial link components @xmath553 and @xmath554 described in figure [ trivialhandle ] .
a trivial link component corresponds to adding a @xmath536-handle to @xmath548 in a trivial way , and on the boundary , to taking the connected sum of the original @xmath363-dimensional manifold and @xmath384 .
theorem [ handleslide ] implies that , given a framed link @xmath398 in @xmath384 , the element @xmath555 is an invariant of the @xmath363-dimensional manifold obtained by performing surgery on @xmath398 , modulo addition and subtraction of trivial @xmath536-handles .
this ambiguity can be removed by using the linking matrix of @xmath398 as follows .
recall that the linking matrix of an oriented framed link @xmath398 has the @xmath556 entry equal to the linking number of the @xmath557th and @xmath173th components for @xmath558 and the @xmath559 entry equal to the writhe of the @xmath557th component , namely to the linking number of the @xmath557th component with a push - out of this component in the direction of the framing .
the signature @xmath560 of the linking matrix does not depend on the orientations of the components of @xmath398 , and is equal to the signature @xmath561 of the 4-dimensional manifold @xmath562 obtained by adding 2-handles to @xmath548 as specified by @xmath398 .
we remind the reader that the signature of @xmath561 is the signature of the intersection form in @xmath563 .
when adding a trivial handle via @xmath553 respectively @xmath564 , the signature of the linking matrix , and hence of the 4-dimensional manifold , changes by @xmath565 respectively @xmath566 . [ omegatrivial ] in any @xmath363-dimensional manifold , the following equalities hold @xmath567 consequently @xmath568 .
we have @xmath569 because @xmath75 is even , @xmath570 hence @xmath571 the last expression is a gauss sum , which is equal to @xmath572 ( see @xcite page 87 ) .
this proves the first formula . on the other hand , @xmath573 which is the complex conjugate of @xmath574 .
hence , the second formula .
[ manifoldinvariant ] given a closed , oriented , @xmath363-dimensional manifold @xmath362 obtained as surgery on the framed link @xmath398 in @xmath384 , the number @xmath575 is a topological invariant of the manifold @xmath362 .
using proposition [ omegatrivial ] we can rewrite @xmath576 where @xmath577 and @xmath578 are the number of positive , respectively negative eigenvalues of the linking matrix .
this quantity is invariant under the addition of trivial handles , and also under handleslide because of theorem [ handleslide ] , so it is a topological invariant of @xmath362 .
the second application of the exact egorov identity is the construction of a sikora isomorphism , which identifies the reduced linking number skein modules of two manifolds with homeomorphic boundaries .
let us point out that such an isomorphism was constructed for reduced kauffman bracket skein modules in @xcite .
[ sikora ] let @xmath579 and @xmath580 be two @xmath363-dimensional manifold with homeomorphic boundaries . then @xmath581 because the manifolds @xmath579 and @xmath580 have homeomorphic boundaries , there is a framed link @xmath582 such that @xmath580 is obtained by performing surgery on @xmath583 in @xmath579 .
let @xmath584 be a regular neighborhood of @xmath583 in @xmath362 , which is the union of several solid tori , and let @xmath585 be the union of the surgery tori in @xmath580 .
the cores of these tori form a framed link @xmath586 , and @xmath579 is obtained by performing surgery on @xmath587 in @xmath580 . every skein in @xmath579 ,
respectively @xmath580 can be isotoped to one that misses @xmath584 , respectively @xmath585 .
the homeomorphism @xmath588 yields an isomorphism @xmath589 however , this does not induce a well defined map between @xmath590 and @xmath591 because a skein can be pushed through the @xmath592 s . to make this map well defined ,
the skein should not change when pushed through these regular neighborhoods . to this end
we use we use theorem [ handleslide ] and define @xmath593 by @xmath594 and @xmath595 by @xmath596 by proposition [ omegaproperties ] b ) we have @xmath597 since each of the components of @xmath598 is a meridian in the surgery torus , hence it surrounds exactly once the corresponding component in @xmath583 .
this implies that @xmath599 .
a similar argument shows that @xmath600 .
now it is easy to describe the reduced linking number skein module of any manifold .
[ vectorspace ] for every oriented @xmath363-dimensional manifold @xmath362 having the boundary components @xmath601 , @xmath602 , one has @xmath603 if @xmath362 has no boundary component then @xmath604 , and if @xmath362 is bounded by a sphere , then @xmath605 , where @xmath606 denotes the @xmath363-dimensional ball . if @xmath362 has one genus @xmath1 boundary component with @xmath607 , then @xmath608 by proposition [ skeinmodules ] b ) . to tackle the case of more boundary components we need the following result : given two oriented @xmath363-dimensional manifolds @xmath579 and @xmath580 , and let @xmath609 be their connected sum .
the map @xmath610 defined by @xmath611 is an isomorphism . in @xmath609 , the manifolds @xmath579 and @xmath580 are separated by a @xmath536-dimensional sphere @xmath612 .
every skein in @xmath609 can be written as @xmath613 , where each @xmath614 intersects @xmath612 in @xmath173 strands pointing in the same direction .
a trivial skein colored by @xmath491 is equal to the empty link .
but when we slide it over @xmath612 it turns @xmath613 into @xmath615 by proposition [ skeinmodules ] b ) .
this shows that the map from the statement is onto . on the other hand ,
the reduced linking number skein module of a regular neighborhood of @xmath612 is @xmath385 since every skein can be resolved to the empty link .
this means that , in @xmath609 , if a skein that lies entirely in @xmath579 can be isotoped to a skein that lies entirely in @xmath580 , then this skein is a scalar multiple of the empty skein .
this allows us to define an inverse of the map from the statement , hence the map is also one - to - one . the lemma is proved .
returning to the theorem , an oriented @xmath363-dimensional manifold with @xmath103 boundary components can be obtained as surgery on a connected sum of @xmath103 handlebodies .
the conclusion of the proposition follows by applying the lemma .
if @xmath362 is a @xmath363-dimensional manifold without boundary , then theorem [ sikora ] shows that @xmath616 if we describe @xmath362 as surgery on a framed link @xmath398 with signature zero , which is always possible by adding trivial link components with framing @xmath617 , then the sikora isomorphism maps the empty link in @xmath362 to the vector @xmath618 more generally , @xmath362 can be endowed with a framing defined by the signature of the @xmath546-dimensional manifold @xmath562 that it bounds constructed as explained before . if @xmath398 is the surgery link that gives rise to @xmath362 and @xmath562 , then to the framed manifold @xmath619 we can associate the invariant @xmath620 the sikora isomorphism associated to @xmath398 identifies this invariant with the empty link in @xmath362 .
all these can be generalized to manifolds with boundary .
a @xmath363-dimensional manifold @xmath362 with boundary can be obtained by performing surgery on a framed link @xmath398 in the complement @xmath75 of @xmath103 handlebodies embedded in @xmath384 , @xmath621 .
we can again endow @xmath362 with a framing by filling in the missing handlebodies in @xmath384 and constructing the @xmath546-dimensional manifold @xmath362 with the surgery instructions from @xmath398 .
to the manifold @xmath619 we can associate the skein @xmath622 .
a sikora isomorphism allows us to identify this vector with @xmath623 .
another sikora isomorphism allows us to identify @xmath624 with the skein module of the connected sum of handlebodies , hence , via proposition [ vectorspace ] , @xmath495 can be identified with a vector @xmath625 where @xmath601 are the boundary components of @xmath75 .
this construction fits atiyah s formalism of a topological quantum field theory ( tqft ) @xcite with anomaly @xcite . in this formalism
* to each surface @xmath626 we associate the vector space @xmath627 which is isomorphic to the reduced linking number skein module of any @xmath363-dimensional manifold that @xmath628 bounds . * to each framed @xmath363-dimensional manifold @xmath629 we associate the empty link in @xmath379 . as a vector in @xmath630 , this is @xmath631 .
atiyah s axioms are easy to check .
_ functoriality _ is obvious .
the fact that @xmath632 is _ involutory _ namely that @xmath633 where @xmath634 denotes @xmath628 with opposite orientation follows by gluing a manifold @xmath362 bounded by @xmath628 to a manifold @xmath635 bounded by @xmath634 and using the standard pairing @xmath636 let us check the _ multiplicativity _ of @xmath632 for disjoint union .
if @xmath628 and @xmath637 are two surfaces , we can consider disjoint @xmath363-dimensional manifolds @xmath362 and @xmath638 such that @xmath639 and @xmath640 .
then @xmath641 also @xmath642 equals @xmath643 . if we now endow @xmath362 and @xmath638 with framings @xmath195 , respectively @xmath644 , then @xmath645 since the sikora isomorphism acts separately on the skein modules of @xmath362 and @xmath638 .
what about _ multiplicativity _ for manifolds glued along a surface ?
let @xmath579 and @xmath580 be @xmath363-dimensional manifolds with @xmath646 and @xmath647 , and assume that @xmath579 is glued to @xmath580 along @xmath628 .
then the empty link in @xmath648 is obtained as the union of the empty link in @xmath362 with the empty link in @xmath638 .
it follows that @xmath649 where @xmath650 is the contraction @xmath651 and @xmath652 expresses the anomaly of the tqft and depends on how the signature of the surgery link changes under gluing ( or equivalently , on how the signatures of the @xmath546-dimensional manifolds bounded by the given @xmath363-dimensional manifolds change under the gluing , see section [ sec:8 ] ) .
finally , @xmath653 , because the only link in the void manifold is the empty link . also , if @xmath654 $ ] , then the empty link in @xmath362 is the surgery diagram of the identity homeomorphism of @xmath628 and hence @xmath655 can be viewed as the identity map in @xmath656 .
this tqft is _
hermitian_. indeed , @xmath630 , being a space of theta functions , has the inner product introduced in section [ sec:2 ] . on the other hand ,
if @xmath362 is a @xmath363-dimensional manifold and @xmath635 is the same manifold but with reversed orientation , then the surgery link @xmath657 of @xmath635 is the mirror image of the surgery link @xmath398 of @xmath362 .
the invariant of @xmath362 is computed by smoothing the crossings in @xmath398 while the invariant of @xmath635 is computed by smoothing the crossings in @xmath657 , whatever was a positive crossing in @xmath398 becomes a negative crossing in @xmath657 and vice - versa .
hence @xmath658 . also ,
because for @xmath357 one has @xmath659 , and hence @xmath660 .
it follows that @xmath661 as desired .
an interesting coincidence in mathematics is the fact that the segal - shale - weil cocycle of the metaplectic representation @xcite and the non - additivity of the signature of 4-dimensional manifolds under gluings @xcite are both described in terms of the maslov index .
we will explain this coincidence by showing how to resolve the projectivity of the hermite - jacobi action using 4-dimensional manifolds .
note that the theory of theta functions explains the coincidence of the cocycle of the metaplectic representation and the cocycle of the hermite - jacobi action .
each element @xmath199 of the mapping class group of @xmath2 can be represented by surgery on a link @xmath662 $ ] .
fix a lagrangian subspace @xmath285 of @xmath4 and consider the closed 3-dimensional manifold @xmath362 obtained by gluing to the surgery of @xmath424 $ ] along @xmath398 the handlebodies @xmath663 and @xmath664 such that @xmath665 , @xmath666 , and @xmath285 respectively @xmath337 are the kernels of the inclusion of @xmath2 into @xmath663 respectively @xmath664 .
the manifold @xmath362 is the boundary of a @xmath546-dimensional manifold @xmath562 obtained by adding @xmath536-handles to the @xmath546-dimensional ball as prescribed by @xmath398 .
the discrete fourier transform @xmath218 is a skein in @xmath391)$ ] which is uniquely determined once we fix the signature of @xmath562 .
hence to the pair @xmath667 where @xmath199 is an element of the mapping class group and @xmath103 is an integer , we can associate uniquely a skein @xmath668 , its discrete fourier transform .
we identify the pair @xmath667 with @xmath669 where @xmath562 is a @xmath546-dimensional manifold defined as above .
note that by adding trivial 2-handles we can enforce @xmath670 to be any integer .
let us consider the @xmath128-extension of the mapping class group defined by the multiplication rule @xmath671 where @xmath672 and @xmath562 are glued in such a way that @xmath673 is identified with @xmath664 in @xmath562 .
if @xmath674 is the lagrangian subspace of @xmath203 used for defining @xmath672 , then necessarily @xmath675 . recall wall s formula for the non - additivity of the signature of 4-dimensional manifolds
@xmath676 where @xmath677 is the maslov index . by using this formula
we obtain @xmath678 where for the second step we changed the signature of the 4-dimensional manifold associated to @xmath679 by adding trivial handles and used proposition [ omegatrivial ] . or equivalently ,
if we let @xmath680 be discrete the fourier transform associated to @xmath199 , normalized by the ( signature of ) the manifold @xmath562 , then @xmath681 this formula is standard in the theory of the fourier - mukai transform ; in it we recognize the cocycle of the segal - shale - weil representation @xmath682 used for resolving the projective ambiguity of the hermite - jacobi action , as well as for resolving the projective ambiguity of the metaplectic representation .
by jacobi s inversion theorem and abel s theorem @xcite , the jacobian of a surface @xmath2 parametrizes the set of divisors of degree zero modulo principal divisors .
this is the moduli space of stable line bundles , which is the same as the moduli space @xmath683 of flat @xmath684-connections on the surface ( in the trivial @xmath0-bundle ) .
the moduli space @xmath683 has a complex structure defined as follows ( see for example @xcite ) .
the tangent space to @xmath683 at an arbitrary point is @xmath685 , which , by hodge theory , can be identified with the space of real - valued harmonic @xmath5-forms on @xmath2 .
the complex structure is given by @xmath686 where @xmath525 is a harmonic form .
in local coordinates , if @xmath687 , then @xmath688 .
if we identify the space of real - valued harmonic @xmath5-forms with the space of holomorphic @xmath5-forms @xmath689 by the map @xmath690 given in local coordinates by @xmath691 , then the complex structure becomes multiplication by @xmath557 in @xmath689 .
the moduli space is a torus obtained by exponentiation @xmath692 if we choose a basis of the space of real - valued harmonic forms @xmath693 @xmath694 such that @xmath695 then the above @xmath249 is the period matrix of this basis . on the other hand , if @xmath6 are the holomorphic forms introduced in section [ sec:2 ] , and if @xmath696 and @xmath697 , @xmath41 then one can compute that @xmath698 the basis @xmath699 determines coordinates @xmath700 in the tangent space to @xmath701 .
if we consider the change of coordinates @xmath702 , then the moduli space is the quotient of @xmath25 by the integer lattice @xmath249 .
this is exactly what has been done in section [ sec:2 ] to obtain the jacobian variety .
this shows that the complex structure on the jacobian variety coincides with the standard complex structure on the moduli space of flat @xmath684-connections on the surface .
the moduli space @xmath703 has a symplectic structure defined by the atiyah - bott form @xcite .
this form is given by @xmath704 where @xmath705 are real valued harmonic @xmath5-forms , i.e. vectors in the tangent space to @xmath703 . if @xmath706 , @xmath41 are as in ( [ rightperiods ] ) , then @xmath707 and @xmath708 ( which can be seen by identifying the space of real - valued harmonic @xmath5-forms with @xmath709 and using the topological definition of the cup product ) .
this shows that the atiyah - bott form coincides with the symplectic form introduced in section [ sec:2 ] . for a @xmath684-connection @xmath710 and curve @xmath375 on the surface ,
we denote by @xmath711 the holonomy of @xmath710 along @xmath375 .
the map @xmath712 induces a function on the jacobian variety called wilson line , the holonomy is just a complex number of absolute value @xmath5 , and the trace is the number itself . ] . if @xmath413=(p , q)\in h_1(\sigma_g,{\mathbb z})$ ] , then the wilson line associated to @xmath375 is the function @xmath713 .
these are the functions on the jacobian variety of interest to us .
the goal is to quantize the moduli space of flat @xmath684-connections on the closed riemann surface @xmath2 endowed with the atiyah - bott symplectic form .
one procedure has been outlined in section [ sec:2 ] ; it is weyl quantization on the @xmath155-dimensional torus in the holomorphic polarization .
another quantization procedure has been introduced by witten in @xcite using feynman path integrals . in his approach ,
states and observables are defined by path integrals of the form @xmath714 where @xmath715 is the chern - simons lagrangian @xmath716 } \mbox{tr}\left(a\wedge
da+\frac{2}{3}a\wedge a\wedge a\right).\end{aligned}\ ] ] according to witten , states and observables should be representable as skeins in the skein modules of the linking number discussed in section [ sec:4 ] .
witten s quantization model is symmetric with respect to the action of the mapping class group of the surface , a property shared by weyl quantization in the guise of the exact egorov identity ( [ egorov ] ) .
as we have seen , the two quantization models coincide .
it was andersen @xcite who pointed out that the quantization of the jacobian that arises in chern - simons theory coincides with weyl quantization . for non - abelian chern - simons theory , this phenomenon
was first observed by the authors in @xcite . in the sequel of this paper , @xcite
, we will conclude that , for non - abelian chern - simons , the algebra of quantum group quantizations of wilson lines on a surface and the reshetikhin - turaev representation of the mapping class group of the surface are analogues of the group algebra of the finite heisenberg group and of the hermite - jacobi action .
we will show how the element @xmath491 corresponding to the group @xmath717 can be derived by studying the discrete sine transform .
przytycki , _ a q - analogue of the first homology group of a 3-manifold _ , in _ contemporary mathematics _ 214 , perspectives on quantization ( proceedings of the joint ams - ims - siam conference on quantization , mount holyoke college , 1996 ) , ed .
coburn , m.a .
rieffel , ams 1998 , 135144 . | abelian chern - simons theory relates classical theta functions to the topological quantum field theory of the linking number of knots . in this paper
we explain how to derive the constructs of abelian chern - simons theory directly from the theory of classical theta functions .
it turns out that the theory of classical theta functions , from the representation theoretic point of view of a. weil , is just an instance of chern - simons theory .
the group algebra of the finite heisenberg group is described as an algebra of curves on a surface , and its schrdinger representation is obtained as an action on curves in a handlebody .
a careful analysis of the discrete fourier transform yields the reshetikhin - turaev formula for invariants of 3-dimensional manifolds . in this context
, we give an explanation of why the composition of discrete fourier transforms and the non - additivity of the signature of 4-dimensional manifolds under gluings obey the same formula . |
the herschel space observatory ( hereafter hso or simply herschel ) is a far - infrared and submillimetre observatory of the european space agency ( esa ) , successfully launched on may 14 , 2009 @xcite .
herschel orbits the second lagrangian point of the sun earth system , that is apparently on the direction opposite to the sun .
the mean height of the orbit is approximately @xmath2million kms , while the semi - amplitude of the actual motion is about the half of this value .
hso is equipped with three on - board scientific instruments : the photodetector array camera and spectrometer ( pacs , * ? ? ?
* ) , the spectral and photometric imaging receiver ( spire , * ? ? ?
* ) and the herschel - heterodyne instrument for the far - infrared ( hifi , * ? ? ?
herschel observes a typical target from a dozen of minutes up to several hours .
the list of observed targets , the utilized instruments and other information can be found on the public web page of hso .
since the solar panels and other stuff mounted on the exterior of the spacecraft reflects the sunlight efficiently , one can expect a considerably good detectability of the spacecraft .
however , due to the sharp terminations of ( some of ) the external devices ( see any photographs of models of the observatory itself ) , the amount of reflected light starkly depends on the spatial orientation of the spacecraft . indeed , the apparent visual brightness of the spacecraft can vary in a range of several magnitudes , yielding all values from the complete non - detection up to @xmath3 .
in addition , as we highlight in this draft , the observed brightness is extremely sensitive to the orientation : even a slew of few degrees can reduce or increase the flux coming from the spacecraft by a factor of two or three .
here we summarize our astrometric and photometric observations of the spacecraft that confirms the expectations discussed above .
our main conclusion is that the observed brightnesses significantly correlates the orientation and these brightness variations due to the slew of the telescope are not `` monotonic '' .
the structure of this draft is as follows .
section 2 describes the observations , section 3 summarizes the methods used in the data reductions while section 4 discusses the results .
since apparently hso moves relatively fast , planning of the observations should be done carefully .
the observatory itself is relatively faint so longer exposure times are needed for a good astrometry and/or photometry , however , the apparent proper motion of the spacecraft limits the exposure times . actually , one can compute an `` optimal '' exposure time by comparing the instrumental fwhm with the proper motion and let this point source move by a certain fraction ( e.g. the half or little more ) of the fwhm .
we carried out the photometric and astrometric observations of hso using the the 60/90/180 cm schmidt telescope of the konkoly observatory , located at the piszkstet mountain station on the night of february 7 , 2011 .
the telescope is equipped with an apogee alta - u @xmath4 ccd camera , yielding a square - shaped field - of - view with a size of @xmath5degrees .
the typical effective fwhm for this setup is between @xmath6 arcseconds , depending on the current seeing .
according to the predictions made available by minor planet center ( mpc ) , the expected proper motion of herschel was between @xmath7 and @xmath8 arcseconds per minute , therefore we concluded to use an exposure time of 60 seconds .
the observations covered the time range between 18:43 ut and 01:21 ut ( on february 8) , with small gaps . all in all we gathered @xmath9 good quality scientific frames . in order to obtain the highest signal - to - noise ratio
, we did not use any specific filter for the observations .
this is a common practice for the astrometry ( and sometimes photometry ) of small bodies in the solar system and due to the spectral responsivity of the employed ccd detector , such data can be interpreted as some sort of wide r measurements .
the expected total proper motion of the object was less than 20 arcminutes , so due to the large field - of - view of the telescope and camera , no manual tracking was needed during the observations .
[ cols= " < " , ]
following the standard calibration procedures ( subtraction of averaged dark frames and corrections using flat field images ) , the reductions of the data were done as follows .
individual stellar profiles has been searched using the ` fistar ` utility of the software package described in @xcite .
these lists of detected stars has been used to perform both relative and absolute astrometry on the images .
the relative ( or differential ) astrometry was done with respect to a selected frame from the middle of the observation series ( namely , the 130st out of the 263 frames ) while the absolute astrometry was done using the usno - b1.0 catalogue ( that is also commonly used in the practice of the astrometry of minor bodies in the solar system ) .
the tasks of astrometry has been done by the ` grmatch ` and ` grtrans ` utilities .
both the relative and absolute astrometry were done using a third - order polynomial fit , that has been found to be sufficient ( since the unbiased fit residuals did not decrease if the orders were increased ) .
the images have then been registered ( using the task ` fitrans ` ) to the reference frame using the results yielded by the differential astrometry .
following the registration of the images , the flux levels of the images have also been scaled linearly to the same level .
this was done by selecting several comparison stars and performing aperture photometry on these ( involving the task ` fiphot ` ) .
the mean of the aperture backgrounds have been subtracted from the respective images ( yielding a zero mean background level ) while the instrumental fluxes has been scaled appropriately to the flux level of the reference image ( that was , in practice , the same as the one used as an astrometric reference ) .
a master reference image was then built ( employing the program ` ficombine ` ) , from every tenth of the individually registered and scaled images by taking the per - pixel median of these .
herschel itself was then searched by eye on @xmath10 frames ( i.e. 3 subsequent images in 7 groups , spreading almost homogeneously through the whole observation ) . using these coordinates as a hint for astrometry ,
the task ` fiphot ` was used to refine the pixel coordinates and then a third - order polynomial has been fitted to both the @xmath11 and @xmath12 coordinates as the function of the time . we found that fitting a third order polynomial was sufficient at a certain level (
i.e. the residuals were not larger than few tenths of a pixel ) and this polynomial was used to interpolate the current position of the hso between the frames were it was marginally or absolutely not detected .
this interpolated coordinates have been exploited as a hint for a subsequent centroid determination ( using the task ` fiphot ` ) .
these refined centroid coordinates were then used as a photometric centroid , on which the actual aperture photometry was performed . in order to avoid the effect of nearby stars ( that were crossed by herschel apparently ) , this process has been done on differential images ( i.e. the master reference image has been subtracted from each registered and scaled image before astrometry and photometry ) . to have an accurate photometry , all of the nicely detected stars on the master reference image has been measured similarly ( with the same aperture size as for herschel ) and using the cross - match lists from the absolute astrometry , we were able do determine the shift between the standard ( usno ) and the instrumental magnitudes .
although this step sounds a bit optimistic ( i.e. the flux levels derived from a ccd image without any filter has been adjusted to a photographic reference catalogue with a given filter and we neglect the effects of the color - dependent extinction ) , we found that the residual of this fit is not larger than @xmath13 . on
can see the resulted light curve on fig .
[ fig : hsolightcurve ] . image stamps showing herschel itself on some of its brightest phases can be seen on fig .
[ fig : hsobrightest ] , while the results of astrometry and photometry in the format of mpc are displayed on table [ table : sample ] .
in this draft some photometric and astrometric measurements for the herschel space observatory have been presented . as it was discussed here , taking accurate photometry of hso has some difficulties due to the fast apparent proper motion .
anyway , the presented photometric results confirm that the brightness variations of hso ( yielded by the telescope slews ) is in accordance with the observation logs , and and this is really reassuring .
the author would thank jnos kelemen for the discussions and gathering the observations and krisztin srneczky for useful advises .
the scientific work of the author related to the herschel mission has been supported by the esa grant pecs 98073 and in part by the jnos bolyai research scholarship of the hungarian academy of sciences . | in this draft photometry and astrometry is presented from the herschel space observatory ( hso ) .
this spacecraft orbits the second lagrangian point ( l2 ) of the sun earth system , yielding a mean distance of a million miles ( @xmath0million kms ) for hso . from such a distance
, hso is observable as a @xmath1 magnitude object moving relatively fast ( apparently several arcseconds in a minute ) and the actual observed brightness highly depends on the spatial orientation of the spacecraft .
this draft describes briefly how observations from this observatory and the subsequent data reductions have been carried out .
our conclusion is really reassuring , namely the brightness variations of hso are in accordance with the publicly available reported logs and target coordinates of this spacecraft .
* photometry of a photometer * andrs pl _ konkoly observatory of the hungarian academy of sciences , konkoly thege mikls t 15 - 17 , budapest , h-1121 , hungary _ _ department of astronomy , lornd etvs university , pzmny pter stny 1/a , budapest h-1117 , hungary _ 2 |
networks have been widely used to describe systems in a multitude of disciplines , such as genetic networks , protein networks or the internet . in ecology ,
networks are mainly used to visualize and describe food webs .
but not only trophic interactions are the focus of attention . in the last years
researchers show a growing interest in the study of other species interactions such as parasitism ( vzquez et al . , 2005 ) , scavenger species ( selva and fortuna , 2007 ) and mostly mutualism ( bascompte and jordano , 2007 and references therein ) . studies on mutualistic food webs focus on specific pairwise interactions between a plant and an animal and how they are shaped by a community context , either in a single locality , or geographically ( bascompte and jordano , 2007 ) .
pairwise interactions can be described in the form of a bipartite graph or an interaction matrix .
these webs are characterized by nodes that represent species or species groups and observed interactions are drawn as links that , when not binary , can render their intensity or frequency in graded thicknesses . in the interaction matrix , links are represented as nonzero cells on the intersection of a row and column . according to almeida - neto et al.,2007 , bipartite webs do in fact offer several advantages of their own : first , they are often fully resolved , without the problems of uneven resolution which haunt the analysis of complete webs .
second , all links are of a single kind of ecological interaction ( e.g. mutualism ) , which ensures structural integrity as well as similar ecological and evolutionary processes throughout the entire assemblage . the most studied structure within a bipartite graph is the nested pattern of species interactions , although other structures are also possible ( prado et al .
2006 ; almeida - neto et al . , 2007 ) .
in nested assemblages , plants with few interactions are related only with generalist animals ; conversely , specialized animals are found related to plants with many links , that is , with large associated faunas .
moreover , generalists in one species set tend to interact with generalists in the other , forming a dense core of interactions ( prado et al . , 2006 ; bascompte and jordano , 2007 ) .
a nested structure is very cohesive and stable .
the fact that few species are involved in many interactions ( functional redundancy ) , poses the community with the possibility for alternative routes if some interactions disappear ( bascompte and jordano , 2007 ) .
a nested structure is also quite robust : it is less prone to sampling bias than number of species and links ( nielsen and bascompte , 2007 ) and not generated by the random combination of sets of plants and animals solely in proportion to their different abundances as previously thought ( prado et al . , 2006 ) .
recently , a large series of mutualistic interaction assemblages have shown a significantly nested structure ( bascompte et al .
substantial effort has been done in developing various measures and forms of calculating nestedness ( atmar and patterson , 1993 ; guimaraes and guimaraes , 2006 ; rodrguez - girons and santamara , 2006 ; almeida - neto et al .
, submitted ) .
the most commonly used nestedness metric is the `` nestedness temperature calculator '' , or `` nestcalc '' , used to calculate nestedness in binary matrices ( atmar and patterson , 1993 ) .
the nestedness from this algorithm has a problem , though : the absolute value of the nestedness temperature is dependent on matrix size and fill .
some studies show that the nestedness temperature of randomly assembled matrices increases with network size and attains its maximum value for intermediate fills ( rodrguez - girons and santamara , 2006 ; almeida - neto et al . , submitted ) .
so , smaller networks need lower temperature than larger ones do to be significantly nested ( nielsen and bascompte , 2007 ) . from a mathematical point of view , the central object in the discussion about nestedness is a matrix of zeros and ones .
the ecologist in the field interpret this matrix as a table where she ( he ) marks a cross at the @xmath1 column and @xmath2 row each time a species of group one @xmath1 ( e.g. plant ) is related to group two @xmath2(e.g .
insect ) . in order to visualize nestedness in the studied ecological community
the ecologist has to rank rows and columns of the table .
in fact , each time one row ( or column ) is permuted to another row ( or to column ) , the interactions among species of groups one and two do not change . for a matrix of @xmath3 species in group one and @xmath4 species in group 2 , there are @xmath5 possibilities to represent the matrix following different permutations of rows and columns .
each one of these possibilities is just a different visualization of the same network structure .
ranking rows and columns is a very practical option to visualize nestedness in a interaction matrix .
when we rank the elements of a matrix we choose one of the @xmath5 possibilites , that one where the elements of the matrix are the most packed .
in other words , we choose the representations where the elements of the matrix are as closest as possible from the @xmath6 , @xmath7 corner . in the literature .
packing procedure is a previous step before the evaluation of a nestedness index ( atmar and patterson , 1993 ) . in this article
we also pack the matrix before the evaluation of our nestedness index .
we introduce here a new nestedness measure applied to digraphs originated from ecological data .
however , the method is more general than its predecessor and can be naturally applied to graphs ( networks ) other than digraphs . in section
@xmath8 we describe the formal objects used in this article : adjacency matrix , manhattan distance in a matrix , projection of a generic matrix into the unit square lattice , packing process , random matrix , maximum nested matrix and nestedness estimator . in section @xmath9 we apply the nestedness estimator to a set of insect - plant herbivory networks extracted from the community ecology literature . in section @xmath10 we summarize the article , point out potential applications of the method and give the final words .
in this section we introduce the concept of distance in a matrix to characterize the nestedness of digraphs . in order to fix the notation we call digraph an object @xmath11 formed by two sets of vertices @xmath12 and @xmath13 and a set of links between these two sets . the digraph is completely described by the adjacency matrix , @xmath14 , of size @xmath15 , where @xmath3 and @xmath4 are the number of elements of @xmath12 and @xmath13 , respectively . by definition @xmath16 if there is a link between vertices @xmath1 of @xmath12 and @xmath2 of @xmath13 and @xmath17 if @xmath1 and @xmath2 are not linked .
it is useful to visualize @xmath14 as a @xmath3 versus @xmath4 lattice with empty ( zero ) or full ( one ) sites .
moreover , the number of links of a vertex @xmath18 is @xmath19 and the distribution of links of @xmath12 and @xmath13 is @xmath20 and @xmath21 respectively . in ecology ,
the field data corresponding to the digraph is composed by two sets of species and the corresponding links ( interactions ) between them . as we pointed out in the introduction , the standard procedure in this area consists in packing the adjacency matrix of the data .
the packing is performed in the following way : the link distributions @xmath20 and @xmath21 are ordered such that the most connected species go to the first position of the matrix . in this way the matrix @xmath14 shows more ones close to @xmath6 and @xmath7 corner and zeros at the opposite corner , @xmath22 and @xmath23 . from the matrix point of view , the packing process consists in replacing lines and columns until @xmath20 and @xmath21 are ordered .
we emphasize that since the packing process do not change the links between species it does not alter the phenomenology underlying the network .
the idea behind packing the matrix @xmath14 is to better visualize network nestedness .
in addition , nestedness is related with the dispersion of ones and zeros after the packing process .
a very nested matrix is one that , after packing , has a minimal mixing of ones and zeros . using a lattice analogy
, a very nested lattice shows a minimum of holes . to introduce distance properties in the original matrix
@xmath14 we map it into a cartesian space . in order to avoid distortions
we map the @xmath15 matrix to the unit square .
to perform this task the cell elements @xmath24 assume the positions @xmath25 and @xmath26 that are done by : @xmath27 @xmath28 in this article we use the manhattan distance because it is broadly employed to measure matrix distances .
euclidean distance is used for estimating distance between elements apart in continuum space , which is not the case here .
in fact , in the context of abstract metric spaces ( courant , r. and hilbert , d. , 1937 ) set of distances @xmath29 that depends on the parameter @xmath30 , the case @xmath31 corresponds to the manhattan distance and the case @xmath32 to the euclidean distance .
we define the occupancy number @xmath33 as the fraction of occupied sites in the adjacency matrix . for @xmath34
the total number of ones in @xmath14 we have @xmath35 .
to quantify nestedness of a given matrix @xmath14 we use two matrix benchmarks with the same @xmath3 , @xmath4 and @xmath33 : the maximal nested matrix @xmath36 and the random matrix @xmath37 .
the maximal nested matrix is constructed in such a way that it has no holes and its elements are as close as possible to the @xmath38 corner .
we construct @xmath36 filling the elements along equidistant diagonals to @xmath38 .
in fact all elements along the same diagonal have the same distance to the @xmath38 corner .
the construction of @xmath36 is the following : the first element occupied is @xmath38 , after that it comes @xmath39 and @xmath40 , followed by @xmath41 , @xmath42 and @xmath43 , etc .
figure [ fig1 ] illustrates the optimal filling strategy to build @xmath36 .
in contrast , the random matrix is constructed in such a way that all its elements are uniformly occupied with the same probability @xmath44 .
the maximum packed matrix of the species interaction will be in - between these two .
we use the manhattan distance to evaluate the distances of the filled elements @xmath45 ( the distances are defined for the matrix elements projected into the unit square in the cartesian plane ) .
we call @xmath46 the sum over the distances of all the elements of the matrix projected into the unit square , that means , @xmath47 for @xmath48 . in order to define the nestedness estimator we introduce two additional distances : the total distance of the artificial matrices @xmath36 and @xmath37 .
we note that @xmath36 has the smallest total distance among all the lattices with the same @xmath33 and we call @xmath49 its total distance , while the total distance of @xmath37 is @xmath50 .
consider a sample of @xmath34 points ramdomly distributed along the unit square , the expected value of the distance to the origin , @xmath51 , is the manhatann distance from the origin to the center of the square of size @xmath52 , that means , @xmath53 .
therefore : @xmath54 to get an insight about distances in @xmath14 we start exploring the behavior of @xmath49 and @xmath50 against occupancy @xmath33 in figure [ fig2 ] .
we use in this picture @xmath55 .
as expected , the distances follow the relation @xmath56 .
the total distance @xmath46 for any matrix , after the packing process , shows the property : @xmath57 in fact , @xmath58 since @xmath49 is derived from an artificial matrix whose components , by construction , have the minimal distance to the origin .
otherwise , @xmath59 because @xmath46 is derived from a packed matrix , and in the packing process the matrix reduces the distances of their elements when compared with a similar random matrix .
the distance as defined above depends on the matrix size and the occupation .
in fact , the total distance observes the relation @xmath60 for a given @xmath33 and the relation @xmath61 for a constant @xmath62 .
this behavior can be visualized in figures [ fig2 ] and [ fig3 ] . in order to have a @xmath34 free nestedness index of the system we define the nestedness index @xmath63 as follows : @xmath64 we emphasize that @xmath49 and @xmath50 are computed over a artificial matrix with the same @xmath3 , @xmath4 and @xmath33 of the original system . in the next section
we test @xmath63 over a set of digraphs from the context of community ecology and discuss the results .
in this chapter we select a set of @xmath0 insect - plant herbivory networks in the literature and apply the nestedness index we develop in this article . in table
[ tab1 ] we enumerate the set of networks with its main properties : the occupancy @xmath33 , size @xmath3 and @xmath4 , the nestedness estimator @xmath63 , the temperature @xmath65
( according to atmar and patterson , 1993 ) and the reference of the network . a visual inspection of the table does not reveal correlation between @xmath63 and @xmath65 .
in fact , a linear correlation analysis between the two variables revels no significant correlation ( @xmath66 and @xmath67 ) . the range of values of our estimator is @xmath68 and the average value is @xmath69 . in contrast , the usual temperature estimator have the range @xmath70 and average value @xmath71 . in order to improve the visual intuition about the problem we plot in figure @xmath10 four lattices of insect - plant networks .
it is clear in the figure that ( a ) and ( b ) are highly nested , and that on the contrary , ( c ) and ( d ) are not nested at all .
this intuitive idea is corroborated by the @xmath63 estimator , but not by the temperature , the estimator of the two initial matrices are 0.18 and 0.23 ( low values ) , and of the last two 0.79 and 0.76 ( high values ) .
the temperature estimator , on the other hand , shows an intermediary value in case ( a ) , where our estimator shows a low value and a very high temperature in ( c ) where our estimator points for a very large value .
in fact , figures ( c ) and ( d ) have a large number of specialists , and in consequence the matrix can not be well nested .
our estimator corroborates this observation . as a final remark concerning this set of figures
, we point that all the matrices in this figure are well packed , that means , the number of elements of lines and columns are ordered . on the other hand ,
the nestedness calculator from atmar and patterson , 1993 , usually fails in packing well the matrices . from the observation of figure [ fig5 ]
we see that the variables are correlated , in fact , a linear correlation analysis reveals @xmath72 , @xmath73 and @xmath74 .
an exponential correlation regression results @xmath75 , @xmath76 and @xmath77 .
therefore the adjust of the data to the exponential curve is slightly better than the linear one .
the dependence of a nestedness index to occupancy was already pointed in the literature ( girons and santamara , 2006 ) . at first
, a relation between @xmath33 and @xmath63 seems intuitive : once the number of sites increase , in the average , they will be more nested after the packing process .
we let for a future work a more carefull analysis of this point .
in this work we develop a new nestedness estimator @xmath63 based on distances over the adjacency matrix of the network .
we think that this estimator will be useful in the methodological discussion involving nestedness in community ecology . to make the method clearer to the reader we summarize the algorithm to find @xmath63 in the following sequence of steps : 1 .
evaluate the link distributions @xmath78 and @xmath79 of the adjascency matrix of the network .
2 . pack the matrix , that means , permute lines and columns of the matrix in order that @xmath78 and @xmath79 are ranked .
this step defines a corner of nestedness .
3 . project the matrix into the unit square in order to avoid distortion due to the diferences between the sizes @xmath3 and @xmath4 of the matrix .
4 . find the manhatann distance @xmath80 of all elements @xmath25 @xmath26 of the matrix and sum to find the total distance of the elements of the matrix @xmath46 .
5 . determine analitically the distance of the associated random matrix with the same occupancy @xmath33 : @xmath81 , for @xmath34 the total number of occupied elements of the matrix .
determine computationally the distance of the asociated maximally nested matrix with the same @xmath33 : @xmath49 .
finally , calcule the estimator @xmath82 . as estimated above ,
@xmath83 . in the limit @xmath84 the network is completely nested and @xmath84 corresponds to the random limit .
we tested our estimator for a set of @xmath0 insect - plant networks and the data is summarized in table [ tab1 ] .
an interesting result of our estimator is that it depends on the occupancy number .
this result is in agreement with the intuitive idea that the matrix nestedness increases with its occupancy density .
the parameter @xmath65 is a very used nestedness estimator in community ecology .
this parameter , despite of its popularity , is not well defined and present several problems ( fischer and lindenmayer , 2002 ; rodrguez - girons and santamara , 2006 ; almeida - neto et al . ,
submitted ) .
we are perfectly aware that our estimator will be compared with the usual temperature estimator developed by atmar and patterson ( 1993 ) .
what we should do is to show the good points of our method and let the methodological discussion to the scientific community . in this way we stress the strong points of our method in the following : 1 .
our algorithm is based on plain geometry and metric statements . in this way
it is simple and can be calculated with help of a easy computer program .
we have two benchmarks clearly defined : the total distance of the random matrix and of the completely nested one .
3 . we do not use any ad hoc parameter in the equation that defines the nestedness estimator .
4 . our estimator gives a number between zero and one .
5 . the visual inspection criterion of nestedness of empirical matrices agrees with our estimator .
this paper opens a new perspective in the study of nestedness .
we develop an original index to measure nestedness that is based on direct metric analysis of the matrix . instead of considering the dispersion of elements around an artificial isocline
, we estimate directly the distances of all the matrix elements from the packing corner .
the nestedness of a matrix is a measure of how much the elements of the matrix are close to the corner where the matrix is packed .
we hope this paper to be usefull to improve the understanding of nestedness in the community ecology context .
the authors thank m. almeida - neto for helpful comments on the manuscript .
umberto kubota , graciela valadares and thomas lewinsohn , who made unpublished data available .
the authors gratefully acknowledge the financial support of fapesp and cnpq , brazil .
prado , p. i. and lewinsohn , t. m. 1994 .
genus tomoplagia ( diptera , tephritidae ) in the serra do cip , mg , brazil : host records and notes of taxonomic interest .
revista brasileira de entomologia 38 : 3 - 4 . | a recent problem in community ecology lies in defining structures behind matrices of species interactions .
the interest in this area is to quantify the nestedness degree of the matrix after its maximal packing . in this work
we evaluate nestedness using the sum of all distances of the occupied sites to the vertex of the matrix .
we calculate the distance for two artificial matrices with the same size and occupancy : a random matrix and a perfect nested one . using these two benchmarks
we develop a nestedness estimator .
the estimator is applied to a set of @xmath0 real networks of insect - plant interactions .
nestedness , networks , insect - plant interactions |
these notes are writeups of three lectures delivered at the theoretical advanced study institute in boulder , colorado , in june 2012 .
the aim of the lectures is to provide students who have little or no experience with jets with the basic concepts and tools needed to engage with the rapidly developing ideas concerning the use of jets in new physics searches at the lhc . a certain amount of familiarity with the structure of qcd , and in particular with qcd showers , is assumed .
lecture one introduces sequential jet algorithms , and develops several main tools in substructure analyses using the boosted higgs as an example .
lecture two delves further into jet grooming and jet shapes , and in lecture three we conclude with an overview of top tagging and bsm searches .
to understand jet substructure and its applications , we must first begin by understanding jets .
jets , together with parton distribution functions and factorization theorems , are the phenomenological tool that allow us to separate out the perturbatively describable hard interactions in proton - proton collisions , and thereby enable us to make quantitative predictions for events involving strongly interacting particles .
jet cross - sections necessarily depend on the algorithm used to define a jet .
there are many jet algorithms , each one with its own strengths and weaknesses .
the first jet algorithm was developed for @xmath0 hadron events by sterman and weinberg in 1977 @xcite . in this algorithm events
are declared to have two jets if all but a fraction @xmath1 of the total energy in the event can be contained within two cones of half - angle @xmath2 .
that is , radiation off of one of the initial partons must be sufficiently hard , [ eq : eps ] e_rad > and at sufficiently wide angles from either of the other jets , [ eq : de ] _ min > for the radiation to be resolved as a separate jet .
how many events have two jets and how many contain three or more obviously depends on the exact values chosen for @xmath3 and @xmath2 . for all
sufficiently large @xmath4 and @xmath2 , the partonic cross - section for radiation of an extra parton into the region of phase space defined by eqs .
[ eq : eps ] and [ eq : de ] is sufficiently isolated from the soft and collinear singular regions of phase space that rates and distributions can be calculated reliably in perturbation theory .
of course , this is a partonic calculation , and to fully match the partonic picture onto reconstructed sprays of hadrons requires some additional theoretical machinery to describe such effects as ( for example ) hadronization . for our purposes , however , a parton shower picture will suffice .
the sterman - weinberg algorithm is the ur - example of a _ cone _ algorithm .
while cone algorithms present a very intuitive picture of parton radiation , they can be somewhat clumsy in practice , particularly as the number of jets increases , and they are not in active use in most experiments today .
other algorithms can deal much more flexibly with high jet multiplicity .
one such flexible algorithm is the jade algorithm , developed by the jade collaboration in the late 1980s , also for @xmath5 hadrons @xcite . here
, jets are constructed by iteratively recombining final state particles .
define a metric to measure the separation between final state particles @xmath6 and @xmath7 , [ eq : yjade ] y_ij , where @xmath8 is the total energy of the event .
note that @xmath9 vanishes if either @xmath6 or @xmath7 is soft ( @xmath10 or @xmath11 ) , or if @xmath6 and @xmath7 are collinear ( @xmath12 ) .
we can now construct jets using the following recipe : * compute the interparticle distances @xmath9 for all particles in the final state , and find the pair @xmath13 with the minimum @xmath9 . * if this minimum @xmath14 for some fixed parameter @xmath15 , combine @xmath6 and @xmath16 into a new particle , and go back to the previous step . * if @xmath17 , declare all remaining particles to be jets .
since clustering of particles proceeds from smaller values of @xmath9 to larger values , this recipe preferentially clusters particles that are probing the regions of phase space dominated by the soft and collinear singularities . in a sense
, the algorithm is trying to combine hadrons into partons by making its best guess for the reconstructed parton shower .
the jade algorithm has only one parameter , the separation cutoff @xmath15 , and clearly can handle different jet multiplicities in an efficient way by varying @xmath15 .
it is the ur - example of a _ sequential recombination algorithm _ , and the ancestor of all jet algorithms in wide use at the lhc .
the most direct descendent of the jade algorithm is the _
@xmath18 algorithm _
@xcite , which replaces the particle energy factor @xmath19 in the jade metric , eq . [ eq : yjade ] , with the factor @xmath20 : y_ij = .
[ eq : ktee ] this still ensures that the metric goes to zero when either @xmath21 or @xmath22 are soft , but has the advantage that the relative softness of a particle depends only on its own energy , and not that of the other particle in the pair .
this fixes up a technical drawback to the jade algorithm , where @xmath23 allows two very soft particles to be combined even if they are at very wide angles from each other .
using @xmath24 means soft particles will get preferentially clustered with nearby harder particles instead . for small @xmath25 , the numerator of eq .
[ eq : ktee ] can be written as simply @xmath26 , the transverse momentum of the softer particle relative to the harder particle hence the name of the algorithm . in this form the metric
is directly related to qcd splitting functions . to create a version of the @xmath18 algorithm that can be used at hadron colliders , where the total energy @xmath27 is unknown , both the algorithm and the metric have to be adapted @xcite . in the metric , we simply use longitudinally boost - invariant quantities @xmath28 and @xmath29 instead of @xmath30 and @xmath31 , and let the metric become dimensionful , d_ij = . the angular parameter @xmath32 introduced here will replace @xmath15 as determining the cutoff for combining particles , as we will see .
we need in addition to define the quantities d_ib = p_t , i^2 for each particle @xmath6 , since we need to also consider splittings from the beam .
the recombination algorithm now works as follows : * compute @xmath33 and @xmath34 for all particles in the final state , and find the minimum value . *
if the minimum is a @xmath34 , declare particle @xmath6 a jet , remove it from the list , and go back to step one . *
if the minimum is a @xmath35 , combine particles @xmath6 and @xmath7 , and go back to step one .
* iterate until all particles have been declared jets .
this algorithm is usually what is meant by when the @xmath18 algorithm is referred to , but you may occasionally see it referred to as the _ inclusive _ @xmath18 algorithm , as there is a related ( _ `` exclusive '' _ ) variant @xcite .
note that the parameter @xmath32 functions as an angular cut - off : two particles separated by a distance @xmath36 will never be combined , regardless of the @xmath28 s of the particles ( this does not necessarily preclude both particles being clustered into the same jet later ) .
in fact , with this jet algorithm , arbitrarily soft particles can become jets .
therefore jets are customarily returned down to some finite @xmath28 cutoff , typically tens of gev . because the @xmath18 algorithm clusters particles beginning with soft particles and working its way up to harder particles , the algorithm tends to construct irregular jets which depend on the detailed distribution of soft particles in an event . for this reason ,
@xmath18 jets are not especially practical for hadron colliders : irregular jets are hard to calibrate , and the jets are quite sensitive to unrelated radiation in the event .
other sequential algorithms are obtained by using different metrics .
the _ cambridge - aachen _ or c - a algorithm is obtained by taking @xcite d_ij = , d_ib = 1 . this metric clusters particles based only on their angular separation , giving a nicely geometric interpretation of jets .
the c - a algorithm still reflects aspects of the qcd parton shower , in particular the _ angular ordering _ of emissions .
however , it is less directly related to the structure of qcd parton splitting functions than the @xmath18 algorithm is , and represents a compromise between reflecting the structure of the parton shower and maintaining some insensitivity to soft radiation .
algorithm entirely abandons the idea of mimicking the parton shower @xcite . here , the metric is d_ij = ( , ) , d_ib = . with this metric
, particles are clustered beginning with the _ hardest _ particles .
this means that the most energetic cores of jets are found first . as soft particles clustered later
have a minimal impact on the larger four - momentum of the jet core , the anti-@xmath18 algorithm tends to cluster particles out to distances @xmath32 from the core of a jet , yielding very regular jets .
anti-@xmath18 jets are therefore much easier to calibrate at experiments , and the anti-@xmath18 algorithm has become the default used at the lhc .
let us conclude this section by emphasizing that all sequential jet algorithms return not only a list of jets but a _ clustering sequence _ for the event . varying the radial parameter @xmath37 simply acts to move the resolution scale up and down the clustering sequence , making it very easy to study how jet distributions and multiplicities depend on the angular resolution @xmath32 . in particular , for the c - a algorithm , the cluster sequence regarded as a function of @xmath32 has a purely geometric interpretation as resolving the event on different angular scales .
all three sequential jet algorithms discussed here also share the same _ reach _ , that is , regardless of the chosen metric , a splitting @xmath38 will not be combined if the angular distance between the daughters exceeds the chosen jet radius , @xmath39 .
this means that , to leading order , perturbative computations of quantities such as jet rates are identical between all three algorithms . finally , the infrared and collinear safety of all three sequential jet algorithms can be easily checked by asking how the cluster sequence would change with the addition of a soft or collinear emission .
for the shower - sensitive @xmath18 and c - a metrics , infrared and collinear safety follows automatically .
the anti-@xmath18 metric is also manifestly ir- and collinear - safe , as can be seen with a little more thought : anti-@xmath18 recombinations are clearly collinear - safe , since collinear splittings are combined near the beginning of the sequence .
ir safety also follows , as soft radiation has negligible impact on the jet built out from the hard core .
the main subject of these lectures are the possibilities and uses of jets to discover physics at and beyond the electroweak scale , which means , for practical purposes , at the lhc .
it is important to remember that events at lhc are a busy hadronic environment .
in addition to the showering and hadronizing hard partons which we want to study , there are large amounts of soft , unassociated radiation from ( 1 ) the _ underlying event _ , that is , the remnants of the scattering protons ; ( 2 ) possible _ multiple interactions _ , that is , additional collisions of partons arising from the same @xmath40-@xmath40 collision as the hard interaction ; and ( 3 ) _ pile - up _ , additional @xmath40-@xmath40 collisions from other protons in the colliding bunches .
these additional sources of radiation contribute a potentially sizable and largely uniform backdrop of hadronic activity that , when clustered into jets , will partially obscure the features of the hard interaction that we would like to reconstruct .
the default jets used at the lhc are formed using the anti-@xmath18 algorithm , with cone sizes @xmath41 ( at atlas ) and @xmath42 ( at cms ) .
these specific choices of @xmath32 come from a compromise between ( 1 ) the desire to collect all the radiation from a single parton , and ( 2 ) the desire _ not _ to sweep up an excessive amount of unrelated radiation .
many advances have combined to make jets at the lhc a particularly fertile field .
* _ advances in experiment _ : the calorimeters at atlas and cms have much finer resolution than in previous experiments , allowing a much more finely grained picture of events . moreover , _ local _ calibration of jets allows jets to be considered on multiple scales . * _ advances in computation _ : the development of fast algorithms @xcite allows broad implementation of sequential recombination . * _ advances in energy _ :
the lhc center of mass energy is large enough that particles with weak scale masses ( i.e. , @xmath43 and @xmath44 ) will for the first time have an appreciable cross - section to be produced with enough of a boost to collimate the daughter partons . the simple picture that one parton corresponds to one jet breaks down badly in this case , and new tools are needed to separate out collimated perturbative decays from qcd showers .
there are several reasons to be interested in boosted particles . very often , there is theoretical motivation to focus on a particular slice of phase space where the daughter particles are necessarily boosted .
high mass resonances are the simplest such examples . for instance , a resonance @xmath45 with mass @xmath46 tev which decays to pairs of gauge bosons would yield highly boosted @xmath47 pairs . even in the absence of a resonance or other mechanism to preferentially populate boosted regions of phase space , looking for boosted signals can also be useful for improving the signal to background ratio . changing the reconstruction method changes what the experimental definition of the signal is , and therefore necessarily the backgrounds change as well . this can sometimes but not always!be enough of an advantage to make up for the reduction in signal rate that comes from selecting only the boosted region of phase space .
background reduction comes in two forms . in high multiplicity final states ,
combinatoric background is often prohibitive .
when some or all of the final state particles are boosted , the combinatoric background is greatly reduced .
but it is also possible to use boosted selection techniques to identify regions where the background from other physics processes is intrinsically reduced . to appreciate the need for new reconstruction techniques at the lhc ,
consider the production of top quarks at fixed center of mass energy @xmath48 .
choosing some angular scale @xmath49 , we can ask , what fraction of top quarks have all three , only two , or none of their partonic daughters isolated from the others at the scale @xmath49 ? this gives a zeroth order estimate of how well a jet algorithm with @xmath50 will be able to reconstruct the three partonic top daughters as separate jets .
the answer we get depends sensitively on both @xmath49 and @xmath48 : .resolved parton multiplicities in @xmath51 events [ cols="^,^,^,^,^",options="header " , ] clearly , tops produced in the very interesting super - tev regime @xmath52 tev straddle the borderlines between several different topologies
. it would be much more desirable to have a flexible reconstruction method that could handle semi - collimated tops in a unified way . to see how we can go about building such reconstruction techniques ,
let s start by considering one of the landmark jet substructure analyses : the case of a boosted higgs decaying into @xmath53 .
this analysis will introduce us to several ideas that will be important tools in our boosted analysis toolbox : fat jets , jet mass , jet grooming , and sequential de - clustering .
searching for the higgs in its decay to @xmath54 is very difficult at the lhc , due to overwhelming qcd backgrounds . even in associated production ,
@xmath55 , the background processes @xmath56 , @xmath57 , and even @xmath51 are overwhelming .
nonetheless , thanks to ref .
@xcite , @xmath58 is now an active search channel at the lhc . to be specific ,
let s consider the process @xmath59 , followed by @xmath60 , and @xmath61 .
the traditional approach to this signal would be to look for final states with a leptonic @xmath62 and 2 @xmath63-tagged jets , construct the invariant mass of the jets , and look for a peak in the distribution of @xmath64 .
the new approach is instead to focus on events where the higgs is produced with substantial @xmath28 , @xmath65 gev , and cluster these events with a large ( @xmath66 ) jet radius , such that all of the higgs decay products are swept up in a single fat jet .
the signal is now a leptonic @xmath62 + a fat `` higgs - like '' jet , and the background to this signal is now @xmath67 one fat jet rather than @xmath56 .
what we ll see is that jet substructure offers us enough quantitative precision in what we mean by a `` higgs - like '' jet to reduce the background by an extent that makes up for the acceptance price demanded by the high @xmath28 cut . for an unboosted search ,
the ultimate discriminator between signal and background is the @xmath63-@xmath68 invariant mass : to find a resonance , look for a bump in the @xmath63-@xmath68 mass spectrum .
now that we have boosted the higgs and collected it into a single fat jet , the higgs mass should be reflected in the invariant mass of the fat jet itself . to understand jet masses for the background , let s take a quick look at how jet masses are generated in qcd .
[ [ jet - mass ] ] jet mass + + + + + + + + .
partons are generally massless ( we will neglect the @xmath63 quark mass ) , but jets are not .
jet mass in qcd arises from emission during the parton shower , and as such we can calculate the leading contribution .
jet mass , like most perturbative jet properties in qcd , is dominated by the first emission .
let s consider for concreteness a quark emitting a gluon , and work in the collinear regime ( small @xmath32 ) . in this approximation , we can consider the jet in isolation from the rest of the event , neglecting interference and splash - in , and we can approximate the qcd splitting functions with the singular portions . doing so
, the amplitude to radiate an extra parton can be written as d_n + 1 d_n dz p ( z ) , where @xmath69 is the virtuality of the parent @xmath70 , @xmath71 is the fraction of the parent energy retained by the daughter quark , and the splitting function @xmath72 for @xmath73 is given by ( z ) = c_f .
the parent virtuality @xmath69 is of course the jet mass - squared . in the collinear limit , t = e_p^2 z ( 1-z ) ^2 = ( p_t , p ) ^2 z(1-z)^ 2 . integrating over rapidity
, we can approximate the average jet mass - squared as : m^2 p_t , p^2 _ 0^r^2 dz z ( 1-z ) ^2 ( z ) . note the limits on the @xmath74 integral :
this is where the choice of jet algorithm enters . as established above , for all sequential jet algorithms , only radiation at angles smaller than @xmath32
will be clustered into the jet .
strictly , we should use a running @xmath75 evaluated at a scale set by the relative transverse momentum of the splitting , but to get a quick estimate , let s perform the integral in the approximation that @xmath75 is constant .
we then obtain m^2 c_f p_t^2 r^2 .
the jet mass scales like @xmath28 , as it had to , and is suppressed by @xmath76 . to this order
the mass increases linearly with @xmath32 .
the exact value of the numerical coefficient will in general depend on the quark versus gluon content of the jet sample .
for instance , the major qcd background for a doubly @xmath63-tagged boosted higgs comes from the splittings @xmath77 , where the splitting function is ( z ) = c_a ( z^2+(1-z)^2 ) , giving , in the constant-@xmath75 approximation , m^2 c_a p_t^2 r^2 .
[ [ coming - back - to - the - higgs ] ] coming back to the higgs , + + + + + + + + + + + + + + + + + + + + + + + + + consider now a splitting @xmath38 .
we have @xmath78 .
in other words , just from kinematics we can express the opening angle in terms of the parent mass and @xmath28 : [ eq : rofboost ] r_ij ~ ~ .
now consider the @xmath18 metric evaluated on this splitting @xmath38 : y_ij = ( e_t , i ^ 2 , e_t , j ^ 2)r_ij^2 = p_t^2
z^2r_ij^2 m^ 2 . for jets with a fixed mass @xmath79 , cutting on the splitting scale @xmath9
then can separate qcd jets , which have a soft singularity @xmath80 , from boosted higgses , which have a flat distribution in @xmath81 .
however , @xmath82 is not flat in @xmath81 , and in particular is minimized at the symmetric value @xmath83 , so cutting on @xmath9 can still help suppress this background . ] moreover , a boosted higgs will go from a mass @xmath84 to massless daughters in one step , while qcd splittings prefer to shed virtuality gradually . to see this ,
consider the sudakov form factor , which exponentiates the splitting functions to obtain the probability of evolving from an initial virtuality @xmath85 to a final virtuality @xmath69 without branching : ( t ) = .
evaluating @xmath86 and using an ir cut - off to regulate the splitting functions , at large @xmath69 , one can work out that @xcite ( t ) ( ) ^p for an exponent @xmath87 , in other words , @xmath88 for large @xmath69 .
in other words , the probability of a qcd jet making a large jump in mass at a branching falls off as @xmath89 .. ] we have now identified two ways in which a higgs boson @xmath44 decaying perturbatively to @xmath54 will behave very differently from a qcd parton branching : the splitting will be symmetric , and show a sudden drop in parton mass . the search algorithm for finding a boosted higgs looks for a splitting inside the higgs jet that behaves like a perturbative decay , and works as follows : * cluster the event on a large angular scale ( ref .
@xcite uses @xmath90 ) , using the c - a algorithm .
large angular scales are necessary in order to get good acceptance for collecting both higgs decay products into a single fat jet : from eq .
[ eq : rofboost ] , we can see that the @xmath63-@xmath68 separation for a 125 gev higgs boson is @xmath91 for @xmath92 gev .
we choose the c - a algorithm because it is a good compromise between accurately reflecting the shower structure of qcd , and minimizing sensitivity to soft radiation in the event .
* now , given a hard fat jet , successively _
unwind _ the jet by undoing the cluster sequence one branching at a time . at each branching @xmath38 ,
check to see whether the splitting looks sufficiently non - qcd - like , by asking that the branching be both _ hard _ , ( m_i , m_j ) < m_p for some parameter @xmath93 , and _ symmetric _ , y_ij > y_cut for some choice of @xmath94 . *
if the splitting fails to be sufficiently hard and symmetric , discard the softer of @xmath6 and @xmath7 , and continue to unwind the harder . *
continue until either an interesting splitting has been found or you run out of jet .
this procedure , often referred to as the `` splitting '' or `` mass - drop '' procedure , identifies an interesting higgs - like splitting @xmath95 , which determines a characteristic angular scale @xmath96 for a particular event .
once this scale @xmath97 has been identified , we benefit greatly by using smaller scales to resolve the event , rather than the large @xmath90 scale we started with .
the reason is the following : starting with such a large jet , we are guaranteed to sweep up a large amount of unassociated radiation along with the higgs decay products .
the effect of this unassociated radiation is to smear out the mass resolution .
the invariant mass is especially vulnerable to distortion from even soft unassociated radiation , because evaluating @xmath98 depends on large cancellations .
the amount of distortion scales like _ soft p_t , j r^3 , in the approximation that unassociated radiation contributes a constant energy @xmath99 per unit rapidity : the jet area scales like @xmath100 , while the incremental contribution to the invariant mass from a soft particle at distance @xmath101 from the jet core contributes as @xmath102 @xcite .
so to recover mass resolution , it is vital to whittle down our initial fat jet to jets only as big as necessary to capture the radiation from the higgs decay products .
in fact , we have already started whittling .
the splitting procedure discards soft , wide - angle radiation clustered into the jet on its way towards finding the higgs - like splitting .
this by itself helps to clean up the mass resolution .
but we can do better : given the scale @xmath96 which is our best guess at the angular separation of the higgs daughter particles , we can resolve the fat jet at the _ filtering _
scale @xmath103 , and keep only the three hardest subjets .
we keep three , rather than two , subjets in order to capture final - state radiation off of one of the @xmath63 quarks . finally , demanding that the two hardest filtered subjets be @xmath63-tagged , ref .
@xcite finds that the higgs can be seen in this channel with @xmath104 significance in 30 fb@xmath105 ( at 14 tev , combining @xmath106 , @xmath107 , and @xmath108 ) , and signal - to - background of @xmath109 however , we emphasize that _ this and all other lhc phenomenological studies are based on expectations from monte carlo_. even very sophisticated monte carlos necessarily capture only an approximation to the full physics of qcd . for this reason ,
both validation in data on one hand and formal theoretical study on the other are critical .
let us then end this section by showing a couple of the most important early experimental results . in fig .
[ fig : atlasdata ] , we show two plots from ref . @xcite . .
note that after filtering , the jet mass has little to no dependence on @xmath110 . from ref .
@xcite.,title="fig:",width=264 ] .
note that after filtering , the jet mass has little to no dependence on @xmath110 . from ref .
@xcite.,title="fig:",width=264 ] on the left , we see that shower monte carlos do a reasonable job of predicting the spectrum of jet masses for the qcd background . on the right ,
the jet mass is plotted as a function of the number of primary vertices @xmath110 in an event , or in other words , the amount of pileup .
note that after filtering , the jet mass has little to no dependence on @xmath110 , indicating that filtering is successfully isolating the hard process .
note also that filtering is _ necessary _ : prior to filtering , the dependence of jet mass on @xmath110 is significant , and in the 2012 operating environment average pileup multiplicity is @xmath111 . heartened by this evidence that our theoretical techniques have a reasonable relationship with reality
, we will proceed in the next section to discuss more ideas for cleaning up pileup , and more jet properties which can discriminate signals from qcd backgrounds .
our last section concluded with a walk - through of the pioneering boosted higgs study , where we saw examples of two topics we will be discussing in this lecture , namely jet grooming and jet shapes . in the boosted higgs analysis discussed in the previous lecture
, we saw that jet mass resolution was badly degraded by the presence of unassociated radiation in the jet , and introduced the process of _ filtering _ to mitigate these contributions .
filtering is one of several _ jet grooming _
algorithms , all of which are designed to `` clean up '' jets by subtracting the contributions of unassociated radiation . [ [ trimming ] ] trimming + + + + + + + + @xcite , similarly to filtering , reclusters the constituents of a fat jet and retains a subset of the subjets , but has a different criterion for keeping subjets . for each jet of interest , the algorithm is : * recluster the constituents using some jet algorithm ( the original reference specifies @xmath18 ) , and resolve on a fixed small angular scale @xmath49 .
* keep each subjet @xmath6 that passes a @xmath28 threshhold , p_t , i > f _ hard for a cutoff parameter @xmath112 and a hard momentum scale @xmath113 . *
the final trimmed jet is the sum of the retained subjets .
the essential idea is that radiation we want to keep tends to be distributed in clusters , reflecting a parent parton emission , while unassociated radiation we do nt want to keep is more uniformly distributed .
asking that radiation cluster sufficiently on small scales then preferentially picks out the radiation which ultimately originated from a parent hard parton .
the @xmath18 algorithm was originally proposed here because it increases the chances that soft fsr will be kept : since clustering in the @xmath18 metric works from soft up , using @xmath18 increases the chance that a relatively soft parton emitted in the parton shower will be reconstructed and pass above the @xmath28 threshold .
but it is possible to imagine using other algorithms for the small - scale reclustering , and indeed implementations using c - a @xcite or even anti-@xmath18 @xcite have been seen to be effective .
the trimming algorithm is simple to state ; the detailed questions arise when we ask how the parameters should be chosen , and in any particular application parameter choices should be optimized for the specific process under consideration .
typical values for the small angular scale range between @xmath114 ; for @xmath49 much smaller than @xmath115 , the finite angular resolution of the calorimeter starts to introduce irregularities .
good choices for @xmath113 are either the total jet @xmath116 , for dijet events or other such events where all jets have similar @xmath28s , or the scalar sum transverse energy of the event , @xmath117 , if jets have some spread over a broader range of @xmath28s .
typical values for the cutoff parameter @xmath112 range between @xmath118 ( more typically for jet @xmath116 ) and @xmath119 ( for event @xmath117 ) : this tends to work out to keeping subjets down to a 5 to 10 gev threshold .
[ [ pruning ] ] pruning + + + + + + + @xcite builds on the observation that the mass - drop algorithm improves mass resolution on boosted hard decays even before the filtering step , by discarding soft wide - angle radiation clustered into the fat jet at the final stages . in the c - a algorithm , the typical
last clusterings in the fat jet are of stray soft radiation , usually unassociated with the parent particle , at wide angles to the jet core .
these late , wide - angle clusterings have a disproportionate effect on jet mass .
pruning adapts the splitting algorithm to specifically check for soft , wide - angle splittings , and throw them away .
the algorithm is : * given a jet @xmath120 , recluster its constituents with c - a , and then sequentially unwind the cluster sequence .
* at each splitting @xmath121 , check whether the splitting is both _ soft _ , @xmath122 and at _ wide angle _
, @xmath123 if so , then drop the softer of @xmath6 , @xmath7 , and continue unwinding the harder . *
stop when you find a sufficiently hard ( or collinear ) splitting .
again , this algorithm has parameters that must be optimized specifically for each process under consideration .
typical values of @xmath124 are @xmath125 , while the radial separation should be tuned to the expected opening angle for a hard process , @xmath126 .
[ [ grooming - in - action . ] ] grooming in action .
+ + + + + + + + + + + + + + + + + + + all three grooming techniques ( filtering , trimming , and pruning ) improve signal to background by both improving mass resolution for signal _ and _ suppressing qcd background .
qcd jets , whose jet masses are generated by relatively softer and less symmetric emissions , are more likely to have their masses shifted substantially downward by jet grooming than collimated perturbatively decaying particles are , thus depleting the background to high - mass searches .
both the sharp gain in signal mass resolution and the depletion of the high mass background can be seen in fig .
[ fig : grooming ] .
we can also see in fig .
[ fig : grooming ] that the different grooming techniques all act slightly differently on background massive qcd jets@xcite .
qcd jets with high masses dominantly have this mass generated by a relatively hard perturbative emission , which all algorithms are designed to retain , so performance between the different algorithms is similar .
however , the effects of the different grooming algorithms on qcd backgrounds are still sufficiently distinct that some benefit can be obtained in applying multiple grooming algorithms @xcite . at low masses ,
the differences between the grooming algorithms become more pronounced .
qcd jets at low masses are dominated by a hard core .
filtering keeps a fixed number @xmath127 of subjets , and therefore retains relatively soft radiation .
trimming , by contrast , will typically drop all radiation except that within @xmath128 of the jet core .
pruning will also typically drop all but the radiation in the core , but the resolution radius @xmath129 is set to scale like @xmath130 , and therefore @xmath131 as @xmath132 .
thus at small masses typically @xmath133 , so pruning acts more aggressively than trimming .
another feature of the boosted higgs analysis we saw in the previous lecture was the importance of _ jet mass _ , which allowed us to concentrate signal in a sharp peak on top of a falling background @xcite .
jet mass is an example of a _ jet shape _ : a function @xmath134 defined on a jet @xmath120 that quantifies the properties of the jet without the ( explicit ) use of any jet algorithm .
the approach is conceptually akin to _ event shapes _ , which allow quantitative study of qcd without requiring specific characterization of an event in terms of jets , and indeed many jet shapes are descendants of event shapes .
before discussing individual jet shapes , let us make two general comments .
first , as we saw for jet mass , jet shapes are vulnerable to the inclusion of unassociated radiation , particularly pile - up , into jets , to a greater or lesser extent depending on the particular jet shape , and the sensitivity of the jet shape to unassociated radiation can be important .
second , one should bear in mind that any reasonable jet shape needs to be both infrared- and collinear - safe
. any linear function of particles @xmath28 is automatically safe ; factorization theorems for other jet shapes can be proven @xcite . the probability of a showering parton to emit a daughter parton depends on the running coupling @xmath75 evaluated at the @xmath135 scale of the splitting .
jet shapes which measure the angular distribution of particles in an event are therefore measuring both the strength and the running of the strong coupling constant , and are classic probes of qcd .
these jet shapes are also sensitive to the color charge of the parent parton : since @xmath136 , an initial gluon will radiate more , and at wider angles , than an initial quark .
[ [ jet - broadening ] ] jet broadening + + + + + + + + + + + + + + is a classic @xmath137 observable .
given a thrust axis @xmath138 , we can partition the particles @xmath6 in an event into hemispheres according to @xmath139 , which for dijet - like events is equivalent to associating each particle to a jet .
hemisphere broadening is then defined as the momentum - weighted transverse spread of the particles , b_h = _
n | where the sum runs over all particles @xmath6 in a hemisphere @xmath140 . [
[ differential - and - integrated - jet - shapes ] ] differential and integrated jet shapes + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + are , thanks to a historic quirk of nomenclature , names for two specific jet shapes : the so - called _ differential jet shape _ @xmath141 and the _ integrated jet shape _ @xmath142 , which characterize the radial distribution of radiation inside a jet .
these jet shapes are also sometimes called the _ jet profile_. both of these shapes are defined on an ensemble of @xmath143 jets formed with radius @xmath32 .
then for @xmath144 , the integrated jet shape @xmath145 is the ensemble average of the fraction of a jet s @xmath28 which is contained within a radius @xmath146 from the jet axis . defining @xmath147 as the distance of a constituent @xmath6 from the jet axis , ( r ) = _ j _ ij . here
the second sum runs over all constituents @xmath6 of a jet @xmath120 .
the differential jet shape @xmath141 is then given by ( r ) = _ j _ ij .
these variables are often included in the suite of qcd precision measurements performed by experimental collaborations , as for instance in the atlas study @xcite , and are useful for validating parton shower models .
[ [ girth ] ] girth + + + + + is another jet shape which probes the radial distribution of radiation inside a jet .
let @xmath147 again be the distance between a constituent @xmath6 and the jet axis . then the girth of a jet @xmath148 is the linear radial moment of the jet , g_j = _ ij . in the collinear limit @xmath149
, girth becomes equivalent to jet broadening ( where the thrust axis is replaced by the jet axis ) .
girth has been shown to be particularly useful for distinguishing quark - initiated jets from gluon - initiated jets @xcite .
[ [ angularities ] ] angularities + + + + + + + + + + + + @xcite are a related family of jet shapes , defined as a function of the parameter @xmath150 : _ a = _ ij p_,i e^(a-1)_i . here
@xmath151 is the separation in rapidity only between particle @xmath6 and the jet axis , and @xmath152 the momentum transverse to the jet axis .
the radial distribution jet shapes discussed in the previous section are geared toward probing the characteristic shower structure of qcd . here
we will discuss several examples of jet shapes which target evidence of non - qcd - like substructure in jets .
[ [ planar - flow ] ] planar flow + + + + + + + + + + + @xcite considers the spread of the jet s radiation in the plane transverse to the jet axis ( see also the closely related _ jet transverse sphericity _ shape @xcite ) . since qcd showers are angular - ordered , radiation subsequent to the first emission @xmath38 tends to be concentrated between the clusters of energy defined by @xmath6 and @xmath7 , leading to a roughly linear distribution of energy in the jet .
by contrast , boosted three - body decays , such as boosted tops , have a more planar distribution of energy .
define the tensor i^ab = _ ij , where the indices @xmath153 span the plane perpendicular to the jet axis , and @xmath154 denotes the projection of particle @xmath6 s momentum into this plane . letting @xmath155 be the eigenvalues of @xmath156 , the planar flow of a jet is given by pf_j= = .
with this normalization , @xmath157 .
monte carlo studies have demonstrated that qcd events do indeed peak at low values of @xmath158 , while boosted top decays show a relatively flat distribution in @xmath158 , but preliminary results show some sensitivity to shower modeling @xcite and the utility of this shape in data is so far unclear .
note that neither @xmath156 nor its eigenvalues are invariant under longitudinal boosts . for fully reconstructible events
this is not a worry in theory , as all events can be considered in the reconstructed cm frame , but finite experimental resolution can become an issue in transforming from the lab frame into the cm frame . [ [ template - overlaps ] ] template overlaps + + + + + + + + + + + + + + + + + define jet shapes based on ( aspects of ) the matrix elements for boosted object decays @xcite .
for example , consider the three body top quark decay with intermediate on - shell @xmath159 .
the phase space for this decay is ( in the narrow - width approximation ) determined by four parameters , which can be parameterized as the solid angle governing the two - body decays of both the @xmath69 and its daughter @xmath159 .
note that ( 1 ) the azimuthal angle @xmath160 is meaningful , as the detector geometry is not invariant under rotations around the top direction of motion , and ( 2 ) this phase space has both @xmath161 and @xmath162 built in .
a series of templates describing this phase space can be generated by discretizing the four - dimensional space .
to use these templates on a jet , the method of template overlaps finds the template which has best overlap with the kinematic configuration of the jet constituents according to a chosen metric .
the ultimate variable is the numerical value of the best overlap , which distinguishes between qcd jets and boosted tops .
[ [ n - subjettiness ] ] @xmath143-subjettiness + + + + + + + + + + + + + + + + + + + + + + + + + + + + @xcite takes a different and more general approach to probing jet substructure via jet shapes .
given @xmath143 axes @xmath163 , we define @xmath143-subjettiness as _ n = where @xmath49 is the jet radius , and @xmath164 is the distance between particle @xmath6 and axis @xmath163 .
the smaller @xmath165 is , the more radiation is clustered around the chosen axes , or in other words , smaller values of @xmath165 indicate a better characterization of the jet @xmath120 as having @xmath143 ( or fewer ) subjets .
conversely , if @xmath165 is large , then a description in terms of @xmath166 subjets is better . however , as qcd alone will happily make jets with subjets , to differentiate boosted objects we need to probe not just the possible existence of subjets , but their structure .
the real distinguishing power of @xmath143-subjettiness occurs when looking at _ ratios_. for instance , a two - prong boosted particle such as a higgs or @xmath167 will have large @xmath168 and small @xmath169 .
qcd jets which have small @xmath169 will generically have smaller @xmath168 than for signal , as the qcd jets are more hierarchical ; conversely , qcd jets which have large @xmath168 are generally diffuse , and will have larger @xmath169 as well than for signal .
thus the best single discriminating variable is @xmath170 , or , more generally r_n = for a boosted @xmath143-prong particle .
the question of how to determine the input subjet axes @xmath163 is an interesting one .
one approach , which is fast and perfectly serviceable for most applications , is to use a jet algorithm , such as exclusive @xmath18 , to determine subjet axes . naturally , the results then retain some dependence on the choice of jet algorithm used to find the axes .
another approach is to marginalize over all possible choices of @xmath163 , and choose the set which minimizes @xmath165 @xcite .
while this choice is computationally more intensive , it removes the dependence on the jet algorithm choice , and additionally guarantees the nice property that _
n-1 > _ n , which holds only approximately if fixed subjet axes are used .
@xmath143-subjettiness is a conceptual descendent of the event shape @xmath143-jettiness @xcite , which classifies events as being @xmath143-jet - like without reference to jet algorithms
. beyond kinematics , boosted perturbative decays can also differ from qcd backgrounds in their color structure .
consider a color singlet such as a @xmath140 or @xmath167 boson decaying to a quark - antiquark pair .
the daugher quark jets form a color dipole : they are color - connected to each other , but not to the rest of the event . meanwhile ,
the backgrounds to these processes come from qcd dijets , which necessarily have different color connections , as we show in fig .
[ fig : colorflow ] , where the radiation patterns for a color - singlet signal are plotted on the left and for a typical background on the right , as computed in the eikonal ( soft ) approximation .
this observation has motivated work on variables which can add color flow to the suite of features which can discriminate signal from background .
[ [ jet - pull ] ] jet pull + + + + + + + + @xcite defines for each jet a transverse vector @xmath171 characterizing the net directional distribution of the soft radiation surrounding the jet core . defining @xmath172 as the ( transverse ) direction of particle @xmath6 from the jet axis , the pull vector is _
the direction of @xmath173 relative to other jets in the event then is sensitive to the color connection of the jet @xmath120 .
two jets which are color - connected to each other will have pull vectors pointing toward each other .
jets which are color - connected to the beam will have pull vectors pointing toward the beam .
once two interesting ( sub)jets have been identified , the discriminating variable is then @xmath174 , the angle between the pull vector and the line connecting the two ( sub)jet cores .
an initial experimental study of pull has been carried out at d0 , using the @xmath159 in top events @xcite .
[ [ dipolarity ] ] dipolarity + + + + + + + + + + @xcite is a jet shape which is designed to test for color dipole - like structure when the apparent particle is boosted and the two ( sub)jets of interest are geometrically nearby . since pull scales like @xmath175 , it can be unduly sensitive to the detailed assignment of particles between the two ( sub)jet cores to one or the other of the two ( sub)jets . dipolarity therefore uses as the relevant distance measure @xmath176 , the transverse distance of particle @xmath177 to the line segment connecting the ( sub)jet cores , d_j = _ ij p_t , i r_i^2 .
note that dipolarity requires input ( sub)jet axes .
the major application studied to date has been in boosted top tagging , where dipolarity can improve the identification of the boosted daughter @xmath159 .
[ [ keeping - the - right - soft - radiation . ] ] keeping the right soft radiation .
+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + we have emphasized the need for jet grooming tools in the busy , high luminosity environment of the lhc .
however , that grooming will groom away most if not all of the information about color flow . to use the information contained in an event s color flow ,
it is necessary to retain at least some of the soft radiation . exactly which soft radiation is included , and at which stage in the analysis , is a question which has to be addressed case - by - case . as an example , we will discuss how the dipolarity shape can be incorporated into a boosted top tagger @xcite .
top tagging will be discussed at length in the next section ; for the moment , it suffices to think of a top tagger as an algorithmic black box which acts on a fat jet to return candidate @xmath63 , @xmath178 , and @xmath179 subjets , and discards some radiation in the process .
the returned subjet axes define the characteristic opening scale , @xmath180 , and provide the input axes for the dipolarity jet shape . as the top - tagger has discarded some of the radiation associated with the top quark in identifying the candidate subjets , to evaluate dipolarity we will need to go back to the original fat jet and include a larger subset of particles . clearly , the radiation we d like to include when evaluating the dipolarity of the candidate @xmath159 daughters is only that associated with the two light quark jets ; including radiation originating from the @xmath63 would just skew the results .
let us consider only moderately boosted tops , such that the @xmath63 jet is not overlapping with the other two . from the angular - ordered property of qcd showers ,
we know that in top events , all radiation associated with either light quark must be at angular separations less than the opening angle of the dipole , @xmath181 .
thus , all radiation from the @xmath159 is contained in cones of radius @xmath180 around each light quark jet .
the authors of ref .
@xcite find that keeping all radiation within these two cones is casting too wide a net , however , and a smaller cone size of @xmath182 is a better tradeoff between keeping all the radiation from the @xmath159 and avoiding pollution from pileup , underlying event , and splash - in from the nearby @xmath63 .
color flow variables capture a genuine physical difference between signal and background .
they have been shown , in theoretical work , to make a sizeable impact in signal significance @xcite , and show great promise as tools to expand our understanding of sm and bsm physics .
it is important to bear in mind , however , that these `` proof of principle '' analyses have all been performed using shower monte carlos , which capture only leading approximations to the full qcd dynamics . just as the jet shapes discussed in section [ sec : raddist ] above have been and are still important tools for assessing the validity of the approximations made in the monte carlo generators , measuring and calibrating color flow variables in data is critical to understand the validity of the shower models and the performance of any color flow variable .
this experimental program is , as of yet , in its infancy . in the meantime
, theoretical studies should bear this uncertainty in mind . to estimate the uncertainties ,
it is useful ( as it is for any novel substructure variable ) to check results using more than one shower model .
in this section we will assemble the tools and techniques developed in the previous two sections and apply them to searches for physics beyond the standard model . by far the most universally motivated application of jet substructure techniques to bsm physics is in the hunt for tev - scale new states which decay to electroweak - scale sm particles .
the best reason for new physics to live anywhere near the weak scale is that it is partially responsible for the generation of the electroweak scale .
new physics that is related to ewsb will naturally couple most strongly to those particles in the sm which feel ewsb most strongly , in particular the top quark and the ew bosons ( @xmath44 , @xmath159 , and @xmath62 ) , and thus will decay preferentially to these heavy particles rather than to the light quarks and leptons which yield simpler final states .
moreover , we have compelling reasons to believe new physics will naturally decay to _ boosted _ sm particles .
even before the lhc turned on , the lack of deviations from sm predictions for flavor or precision electroweak observables already hinted that the likely scale for new physics was not @xmath183 as naturalness might have suggested , but rather @xmath184 few tev .
evidence for this `` little hierarchy '' problem has of course only gotten stronger as the lhc has directly explored physics at tev scales .
thus many models which address the stabilization of the ew scale will naturally give rise to final states rich in boosted tops , higgses , @xmath159 s and @xmath62 s . in this section
we will provide an introduction to top tagging at the lhc , followed by a few brief concluding comments on searching for more general bsm physics with jets . as we established in section 1 , top pair production at the lhc
covers a broad range of kinematic regimes interpolating between threshold ( @xmath185 ) , where tops are well described as a six - object final state , up to tev - scale energies , where the tops are highly collimated and are best described as a two - object final state .
top reconstruction must thus be able to flexibly cover a wide range of kinematic scenarios . in the interest of time
, we will restrict our attention here to top taggers which target the hadronic decay of the top quark , although the semi - leptonic decay mode also requires interesting techniques for identification and reconstruction @xcite .
as for jet algorithms , the `` best '' top tagger depends on the question being asked . in particular ,
different strategies are required at high @xmath28 ( @xmath186 ) versus moderate @xmath28 ( @xmath187 ) .
another question is : what signal efficiency is necessary ?
every tagging technique trades off signal efficiency against background mistag rate . depending on the search in question ,
the composition of the backgrounds will change , and therefore the necessary mistag rate will shift as well . for example , consider a top pair event with at least one boosted hadronic top . if the other top is also hadronic , then qcd dijets are by far the dominant background , and small qcd mistag rates are required .
but if the other top is leptonic , then @xmath188 jets becomes an important background , and if the top is produced in association with some new physics objects , such as @xmath189 , then the backgrounds may be substantially smaller , and mistag rates may be entirely unimportant .
the aim of this section is to provide an introduction to top tagging by discussing a representative variety of top taggers .
specifically , we will consider the top taggers currently used by both lhc experiments , which work best in the highly boosted regime ; the `` hep top tagger '' , which targets moderate @xmath28 ; and top tagging with @xmath143-subjettiness .
the hadronic top tagger used by cms @xcite is largely based on the `` hopkins '' top tagger @xcite .
it builds on the techniques of the boosted higgs `` splitting / filtering '' or `` mass drop '' analysis , which we discussed in section 1 .
thus , we again begin by clustering the event using the c - a algorithm , on large angular scales , capturing all of the top decay products in a single fat jet , which we will then unwind until we find interesting substructure .
compared to the higgs analysis , there are two important differences .
first , we are looking for at least three hard subjets , instead of two .
second , we take the fat jet radius to be noticeably smaller than we did for the higgs case : @xmath190 . using our rule of thumb ,
@xmath191 , this means we are targeting tops with @xmath192 gev : appropriate for production from a tev - scale resonance .
contrast this with the boosted higgs , which was targeting the high-@xmath28 tails of sm associated production , where requiring large @xmath28 imposed a significant price in signal acceptance . iteratively declustering the fat jet , we encounter splittings @xmath193 .
our criterion for an interesting splitting is simply that both daughter subjets must carry a sufficiently large fraction of the _ total _ fat jet momentum , p_t , j > _ p p_t , j for some parameter @xmath194 . if a splitting fails to meet this criterion , discard the softer of @xmath195 , and continue to unwind the harder .
the splitting is rejected if it is too collinear , @xmath196 , for another parameter @xmath197 .
this procedure stops when either both @xmath6 , @xmath7 are softer than @xmath198 , or only one particle is left .
if an interesting hard , non - collinear splitting @xmath199 is found , then the next step is to successively unwind both @xmath178 and @xmath179 according to the same algorithm , in search of further interesting splittings .
this procedure returns a set of 2 , 3 , or 4 subjets .
fat jets returning only 2 subjets do nt have enough substructure to be good top candidates , and are rejected .
jets which return 3 or 4 subjets do show enough substructure to be interesting , and the next step is to test whether or not they also have top - like kinematics .
as for the higgs , the single most important discriminator is the jet mass .
cms requires that the jet mass , as computed from the sum of the returned subjets , lie within a top mass window , @xmath200 gev .
the onshell decay of the @xmath159 inside the jet will also help us separate signal from background , but rather than trying to explicitly identify a pair of subjets which reconstruct a @xmath159a procedure highly vulnerable to the misassignment of particles in overlapping jets we will exploit the presence of the @xmath159 mass scale in a less direct way . ( blue ) and @xmath201 ( purple ) in unpolarized top decay.,width=312 ] the pairwise invariant masses of all possible combinations of the three daughter quarks are all governed by the mass scales in the top matrix element , @xmath162 and @xmath161 . the distribution of the invariant mass of the @xmath63 and the @xmath202-type quark ( equivalent to the charged lepton in leptonic top decay ) is shown in fig .
[ fig : mblmbnu ] .
the most likely value of @xmath203 is approximately 115 gev .
the invariant mass of the @xmath63 and the @xmath204-type quark ( equivalent to the neutrino ) is peaked at even larger values . by contrast ,
subjet masses from qcd background processes are hierarchically smaller than the total parent jet mass . thus instead of trying to reconstruct the @xmath159
, we simply require that the _ minimum _ of the invariant masses formed from pairs of the three hardest subjets be sufficiently large to reject backgrounds , ( m_12 , m_13 , m_23 ) > 50 . these cuts on masses , together with the substructure requirement ,
constitute the tagger .
note that no @xmath63-tagging information is used .
tagging @xmath63-jets is very difficult in this environment for two reasons .
first , the @xmath63 is embedded in a highly collimated top , so disentangling the tracks that are associated with the @xmath63 from the other tracks in the jet is challenging .
second , the @xmath63 itself is at very high @xmath28 , so the opening angles of its daughter products are small , and it is difficult to get sufficient resolution from the reconstructed tracks to reconstruct the displaced vertex .
note also that the tagger does nt require jet grooming .
this is partly because the iterative decomposition procedure is performing some of that function in its own right , as it discards soft wide - angle radiation in the process of finding hard subjets ( compare pruning ) .
the smaller geometric size of the fat jets also means that pollution is not as large an effect .
we turn next to the atlas top tagger . like cms tagger ,
it is optimized for high @xmath28 , and like cms tagger , it is based on iterative declustering of a sequential algorithm . however , the atlas tagger draws on a very different set of ideas , largely based on work by ref . @xcite and the `` y - splitter '' of ref .
@xcite .
the atlas top tagger begins by clustering events using the anti-@xmath18 algorithm with @xmath205 .
( the slightly larger jet radius means that this tagger works best at slightly lower @xmath28 than does the cms tagger . )
since the anti-@xmath18 algorithm knows nothing about the singularity structure of qcd , its use is simply to identify a nicely regular initial set of particles .
the next step is to take this set of particles and recluster them using the @xmath18 algorithm .
recall that the @xmath18 algorithm preferentially clusters soft splittings .
this means that the hardest splittings in the jet are the very last ones .
thus , there is no need to do any preliminary unwinding , and the existence of hard substructure is directly reflected in the hardness of the scales given by the @xmath18 metric evaluated on the last few splittings in the jet : d_ij = ( p_t , i^2,p_t , j^2 ) r_ij^2 .
large splitting scales mean the emissions are both hard and at wide angles .
the atlas tagger uses as inputs the splitting scales of the last three recombinations , @xmath206 , @xmath207 , and @xmath208 .
the first two splittings correspond ( usually ) to the identification of the three daughter partons , and the third to possible fsr from one of the partons . since for tops
the splitting @xmath208 is the first which comes from the qcd shower , its scale can still be relatively large ; on the other hand , for background qcd jets , the hierarchical nature of the shower means that generally @xmath209 .
thus cuts on @xmath208 maintain some discriminating power . however , instead of cutting directly on the massive splittings @xmath35 , it is advantageous to change variables to a set which are less correlated with the jet and subjet invariant masses @xcite .
we define the energy sharing variables z_ij= where in the last step we have taken the collinear limit ( and @xmath210 ) .
notice that by performing this change of variables we have removed sensitivity to the collinear singularity , so that @xmath211 is only capturing information about the soft singularity .
meanwhile , jet invariant masses still retain information about the relative angles between the jets , so the correlation between the variables has been reduced .
the final set of variables that make up the atlas top tagger is then : * the total jet mass , @xmath212 .
the tagger requires @xmath213 gev , and no upper bound : no grooming procedure is used , so the mass spectrum is distorted upwards . * the variable @xmath214 , defined as the minimum pair invariant mass of the three subjets identified at the splitting scale @xmath207 .
this is the equivalent to cutting on the minimum pair invariant mass in the cms tagger ; only the method of finding the subjets is different .
we require @xmath215 gev . *
all three energy sharing variables , @xmath216 , @xmath217 , and @xmath218 , which are subject to numerical cuts .
we turn now to the heidelberg - eugene - paris top tagger , which functions on tops with @xmath219 gev @xcite . in some sense
this algorithm is more of an event reconstruction strategy than a top tagger .
the algorithm begins by clustering the event using c - a on the extremely large angular scale @xmath220 , and requiring the fat jets thus formed to have @xmath221 gev .
the @xmath28 cut of 200 gev puts us in the regime where the top is sufficiently boosted that its decay products will frequently lie in a single hemisphere . looking at extremely fat jets is effectively identifying hemispheres in an event while avoiding the need to set any fixed angular scales for resolution within those hemispheres .
this is an effective strategy for tops in this intermediate kinematic regime , where events will straddle any fixed angular scale ; by unwinding c - a hemispheres , we allow the angular scales to be flexibly identified event by event . the next step is to unwind the fat jet looking for interesting hard structure .
this is done by employing a ( loose ) mass - drop criterion . for a splitting @xmath121 , with @xmath222 ,
the splitting is deemed sufficiently interesting if [ eq : hepsplit ] m_j > 0.2 m_p .
if the splitting passes this criterion , retain both @xmath6 and @xmath7 in the list of jets to unwind ; otherwise , discard @xmath7 and keep unwinding @xmath6 until @xmath223 gev , at which point the unwinding stops .
this unwinding procedure is performed on all subjets identified as interesting via eq .
[ eq : hepsplit ] .
the output of this step is a list of subjets @xmath224 resulting from this iterative declustering ; if there are at least 3 such subjets , then we have found enough substructure to continue . at the next stage , we _ filter _ the substructures to shrink the geometric area associated with the top daughters and thereby reduce sensitivity to pileup , etc .
unlike in the higgs case , where the mass drop criterion identified a unique angular scale @xmath96 associated with the sole hard splitting , we have a more complicated set of jets with more than one interesting splitting , and it is not immediately obvious which angular scale should be used to filter the event .
the hep top tagger determines the filter radius @xmath225 by brute force , as follows . for each possible
set of three subjets that can be drawn from the @xmath224 , filter them by resolving the constituents of those subjets with radius @xmath226 , and retain up to five subjets .
let @xmath227 be the invariant mass of these up - to - five filtered subjets , and select the set with @xmath212 closest to @xmath161 as the top candidate .
these up - to - five filtered subjets are then ( yet again ) reclustered into three subjets , which are the candidates for the partonic top daughters .
the next step is to test whether or not the reconstructed top daughters have top - like kinematics .
again , we will exploit the presence of both the top and @xmath159 mass scales .
we have already used @xmath161 to identify the best set of subjets . unlike in the previous taggers
, we will now demand evidence of the on - shell @xmath159 in a more complex way .
label the three subjets returned by the previous step as @xmath228 in descending order of @xmath28 .
of the three invariant masses @xmath229 , @xmath230 , and @xmath231 , only two are independent .
this means that the top kinematics is characterized by a specific distribution in the two - dimensional space determined by the pair invariant masses .
top jets are focused into a thin triangular annulus in this space , as two subjets reconstruct an on - shell @xmath159 ( the annulus is triangular since any of the @xmath232 may correspond to the @xmath159 ) .
background , by contrast , is concentrated in regions of small pairwise invariant masses .
the kinematic cuts imposed in the hep top tagger pick out this top - like triangular annulus by asking that events lie on one of the three branches of the annulus . of each other
( blue , dashed ) ; all top decay products clustered into @xmath233 c - a fat jets ( purple , dotted ) ; tagged by the hep top tagger ( red , solid ) ; tagged , but with reconstructed subjects not matching original partons ( green , dot - dashed ) .
data from ref .
@xcite.,width=408 ] to understand how well this procedure covers the interpolating kinematic region , we plot efficiencies for tops to pass through these steps in fig . [
fig : heptop ] . as is evident from the blue ( dashed ) curve , simply demanding that all decay products of the top lie in a single hemisphere imposes a non - negligible acceptance price for tops at the low end of the @xmath28 range , which drops quickly as the tops become more energetic .
further demanding that the top daughters all be clustered into the same fat jet results in an additional mild efficiency loss , seen in the purple ( dotted ) curve .
the purple curve is the fraction of tops giving rise to taggable jets ( neglecting the possibility of mistagged signal ) .
the red line denotes the final efficiency of the full hep top tagger , after the filtering and kinematic cuts . at low @xmath28 ,
the fraction of taggable jets which are in fact tagged is near unity , but as the tops become more collimated , the probability of a taggable jet passing the kinematic cuts falls off , in large part because collimation and jet - particle misassignment make the @xmath159 mass reconstruction less precise . at the upper end of the @xmath28 range shown in the figure , the high-@xmath28 top taggers are useful , and would take over .
let us also comment that there is a possibility for tops to pass the top tagger by accident , when the algorithm picks up the wrong set of jets ; this is shown in the green ( dot - dashed ) curve . as we saw in section
[ sec : shapesub ] , @xmath143-subjettiness offers an entirely complementary test of the existence of hard substructure .
a simple and highly effective top tagger can be constructed using as input variables just the jet mass and the ratio @xmath234 .
further refinement is possible with a multivariate analysis which uses in addition @xmath170 as well as @xmath235 and @xmath236 individually @xcite . from our experience with the previous taggers , we can guess that even further improvement would be possible if some information about the @xmath159 were also incorporated ; since the @xmath143-subjettiness jet shape also provides a method of determining subjet axes , it naturally suggests methods for defining three subjets and computing the analog of @xmath214 . to the best of the author s knowledge , no such study has been publicly performed .
let us now consider the performance of the top taggers which we have discussed .
this task is made easier by the work performed in the boost 2010 @xcite and boost 2011 @xcite workshops , which compared the performance of different top taggers on the same reference sets of event samples .
these event samples are publicly available online , so should you develop your own brilliant ideas about top tagging , you can cross - check the performance of your novel technique with the techniques already in the literature . in fig . [
fig : eff1 ] we show performance curves for the high-@xmath28 top taggers which we discussed above . in the boost 2010 reference samples for : atlas ( blue , dotted ) ;
cms ( red , dashed ) ; and @xmath143-subjettiness in both the simple ( green , solid ) and multivariate ( green , dot - dashed ) versions .
data from refs .
@xcite.,width=384 ] overall , these high-@xmath28 top taggers have efficiencies on the order of @xmath237 , at a ( qcd ) background mistag rate of @xmath238 .
( for comparison , lhc @xmath63-tagging algorithms achieve @xmath239 , with a fake rate @xmath240 . )
it is evident from the performance curves that the atlas tagger outperforms the cms tagger when high signal efficiency is required , while cms does better at lower signal efficiency .
even the simple two - variable @xmath143-subjettiness tagger outperforms both cms and atlas taggers by a notable margin , except at high signal efficiency , while adding the additional multivariate discrimination to the @xmath143-subjettiness tagger provides a significant improvement .
further updates in the boost 2011 workshop show that ( 1 ) being more precise about modeling qcd radiation at wide angles and ( 2 ) including the effects of finite detector resolution reduce typical efficiencies to @xmath241 , at a ( qcd ) background mistag rate in the range @xmath242 depending on the tagger @xcite . incorporating
finite detector resolution also tends to reduce ( but not erase ! ) the relative advantage of @xmath143-subjettiness over the sequential decomposition - based taggers .
new physics produces jets with substructure when the kinematics are governed by a nontrivial hierarchy of scales .
for the top examples we ve been discussing , this hierarchy arises from the separation between the scale characterizing new physics and the electroweak scale : _ np _ ew _ qcd .
the little hierarchy problem results in a very strong motivation for developing tagging techniques for boosted sm objects . besides the top tagging discussed in the previous subsection
, much effort has also gone into tagging boosted @xmath159 , @xmath62 , and @xmath44 bosons arising from the decay of new tev - scale particles@xcite .
this is fortunate for theorists , as , once these techniques are put into use at experiments in one context , the barrier is much lower for their adaptation in other contexts where the theoretical motivation may not be so universal .
what other kinds of bsm physics are amenable to substructure analyses ? to engender events with interesting substructure , some multi - tiered hierarchy of scales is required .
we will enumerate an illustrative but far from exhaustive set of examples .
supersymmetry is one example of a new physics sector which naturally can generate multiple scales .
for example , if supersymmetry is broken at very high scales , rg effects will drive the colored superpartners much heavier than those superpartners with only ew charges .
thus , at the weak scale one could naturally expect @xmath243 . in the presence of a large hierarchy between gluino and neutralino
, the decay products of the neutralino would be collimated .
let us further suppose that the neutralino decays via the @xmath32-parity violating @xmath244 superpotential operator , so that @xmath245
. then gluino pair production would appear as a six - jet final state , where two of the jets are actually boosted neutralinos , containing interesting substructure @xcite .
the very large particle content of the mssm can easily accommodate many possible hierarchies , with different theoretical origins ; see for instance ref .
@xcite for another of the many possibilities .
another way to generate a hierarchy in a bsm sector is if the new physics sector contains a broken global symmetry , so that the scale @xmath246 characterizing the lightest states is set by the magnitude of the global symmetry breaking , rather than by the overall scale of the new sector , @xmath247 .
thus consider , for example , a composite rho @xmath45 , decaying into two pseudo - nambu - goldstone bosons @xmath248 , which are stable within their own sector , and therefore must subsequently decay into sm objects @xcite .
hidden valley models also have this kind of multi - scale structure . here
the hierarchy is between the mass of the mediator which connects the visible and hidden sectors and the mass scale of the light states in the hidden sector , _ med _ np > _ sm .
the mediating particle might be a sm particle , in particular the @xmath44 or @xmath62 , or a novel field such as a @xmath249 @xcite or the sm lsp @xcite .
exotic higgs decays to light particles also fall under this umbrella @xcite .
more generally , thinking more broadly and flexibly about jets leads to new approaches to combinatorics and event reconstruction @xcite , and provides novel methods to distinguish qcd events from new physics . as challenging
high - multiplicity and all - hadronic final states become a larger component of the lhc program , flexible and creative jet techniques will be critical to our ability to discover and interpret the physics .
jet algorithms themselves are still an evolving field ! the anti-@xmath18 algorithm was introduced only a few years ago .
as the nature of the questions that we ask about jets evolves , so do the best jet algorithms to address these questions .
there is still a lot of room for new ideas !
references which were invaluable in the preparation of these lectures are the text _ qcd and collider physics _ , by ellis , stirling , and webber @xcite , and the lecture notes `` toward jetography '' by salam @xcite .
the proceedings of the boost 2010 and 2011 workshops @xcite are valuable resources for those looking for a quantitative survey of both theoretical and experimental progress in jet physics at the tevatron and the lhc .
it is a pleasure to thank m. schmaltz and the organizers for the opportunity to be part of such an excellent program .
i thank d. krohn for introducing me to jet substructure , and for many useful conversations during the course of our collaborations .
thanks to c. vermilion for providing figure [ fig : grooming ] , and m. freytsis and d. krohn for comments on the manuscript . finally , thanks to my collaborators , y. bai , a. falkowski , a. thallapillal , and l - t .
i am supported by doe grant de - fg02 - 92er40704 , nsf grant phy-1067976 , and the lhc theory initiative under grant nsf - phy-0969510 .
g. f. sterman and s. weinberg , phys .
lett . * 39 * , 1436 ( 1977 )
. w. bartel _ et al . _
[ jade collaboration ] , z. phys .
c * 33 * , 23 ( 1986 ) .
s. bethke _ et al . _
[ jade collaboration ] , phys .
b * 213 * , 235 ( 1988 ) .
s. catani , y. l. dokshitzer , m. olsson , g. turnock and b. r. webber , phys .
b * 269 * , 432 ( 1991 ) .
s. d. ellis and d. e. soper , phys .
d * 48 * , 3160 ( 1993 ) [ hep - ph/9305266 ] .
s. catani , y. l. dokshitzer , m. h. seymour and b. r. webber , nucl .
b * 406 * , 187 ( 1993 ) .
y. l. dokshitzer , g. d. leder , s. moretti and b. r. webber , jhep * 9708 * , 001 ( 1997 ) [ hep - ph/9707323 ] .
m. cacciari , g. p. salam and g. soyez , jhep * 0804 * , 063 ( 2008 ) [ arxiv:0802.1189 [ hep - ph ] ] .
m. cacciari and g. p. salam , phys .
b * 641 * , 57 ( 2006 ) [ hep - ph/0512210 ] .
j. m. butterworth , a. r. davison , m. rubin and g. p. salam , phys .
lett . * 100 * , 242001 ( 2008 ) [ arxiv:0802.2470 [ hep - ph ] ] .
m. dasgupta , l. magnea and g. p. salam , jhep * 0802 * , 055 ( 2008 ) [ arxiv:0712.3014 [ hep - ph ] ] .
g. aad _ et al .
_ [ atlas collaboration ] , jhep * 1205 * , 128 ( 2012 ) [ arxiv:1203.4606 [ hep - ex ] ] .
d. krohn , j. thaler and l. -t .
wang , jhep * 1002 * , 084 ( 2010 ) [ arxiv:0912.1342 [ hep - ph ] ] .
a. abdesselam , e. b. kuutmann , u. bitenc , g. brooijmans , j. butterworth , p. bruckman de renstrom , d. buarque franzosi and r. buckingham _ et al .
_ , eur .
j. c * 71 * , 1661 ( 2011 ) [ arxiv:1012.5412 [ hep - ph ] ] .
a. falkowski , d. krohn , l. -t .
wang , j. shelton and a. thalapillil , phys .
d * 84 * , 074022 ( 2011 ) [ arxiv:1006.1650 [ hep - ph ] ] .
s. d. ellis , c. k. vermilion and j. r. walsh , phys .
d * 80 * , 051501 ( 2009 ) [ arxiv:0903.5081 [ hep - ph ] ] .
s. d. ellis , c. k. vermilion and j. r. walsh , phys .
d * 81 * , 094023 ( 2010 ) [ arxiv:0912.0033 [ hep - ph ] ] .
d. e. soper and m. spannowsky , jhep * 1008 * , 029 ( 2010 ) [ arxiv:1005.0417 [ hep - ph ] ] .
w. skiba and d. tucker - smith , phys .
d * 75 * , 115010 ( 2007 ) [ hep - ph/0701247 ] .
s. d. ellis , c. k. vermilion , j. r. walsh , a. hornig and c. lee , jhep * 1011 * , 101 ( 2010 ) [ arxiv:1001.0014 [ hep - ph ] ]
. g. aad _ et al . _ [ atlas collaboration ] , phys .
d * 83 * , 052003 ( 2011 ) [ arxiv:1101.0070 [ hep - ex ] ] .
j. gallicchio and m. d. schwartz , phys .
lett . *
107 * , 172001 ( 2011 ) [ arxiv:1106.3076 [ hep - ph ] ] .
l. g. almeida , s. j. lee , g. perez , g. f. sterman , i. sung and j. virzi , phys .
d * 79 * , 074017 ( 2009 ) [ arxiv:0807.0234 [ hep - ph ] ] .
j. thaler and l. -t .
wang , jhep * 0807 * , 092 ( 2008 ) [ arxiv:0806.0023 [ hep - ph ] ] .
l. g. almeida , s. j. lee , g. perez , g. sterman and i. sung , phys .
d * 82 * , 054034 ( 2010 ) [ arxiv:1006.2035 [ hep - ph ] ] .
j. gallicchio and m. d. schwartz , phys .
lett . * 105 * , 022001 ( 2010 ) [ arxiv:1001.5027 [ hep - ph ] ] .
v. m. abazov _ et al . _
[ d0 collaboration ] , phys .
d * 83 * , 092002 ( 2011 ) [ arxiv:1101.0648 [ hep - ex ] ] .
a. hook , m. jankowiak and j. g. wacker , jhep * 1204 * , 007 ( 2012 ) [ arxiv:1102.1012 [ hep - ph ] ] .
d. curtin , r. essig and b. shuve , arxiv:1210.5523 [ hep - ph ] . k. rehermann and b. tweedie , jhep * 1103 * , 059 ( 2011 ) [ arxiv:1007.2221 [ hep - ph ] ] .
[ atlas collaboration ] , atl - phys - pub-2009 - 081 , atl - phys - pub-2010 - 008 .
[ cms collaboration ] , cms - pas - jme-09 - 001 .
d. e. kaplan , k. rehermann , m. d. schwartz and b. tweedie , phys .
* 101 * , 142001 ( 2008 ) [ arxiv:0806.0848 [ hep - ph ] ] .
g. brooijmans , atl - phys - conf-2008 - 008 , atl - com - phys-2008 - 001 , jan 2008 .
j. m. butterworth , b. e. cox and j. r. forshaw , phys .
d * 65 * , 096014 ( 2002 ) [ hep - ph/0201098 ] .
t. plehn , g. p. salam and m. spannowsky , phys .
lett . * 104 * , 111801 ( 2010 ) [ arxiv:0910.5472 [ hep - ph ] ] .
t. plehn , m. spannowsky , m. takeuchi and d. zerwas , jhep * 1010 * , 078 ( 2010 ) [ arxiv:1006.2833 [ hep - ph ] ] .
a. altheimer , s. arora , l. asquith , g. brooijmans , j. butterworth , m. campanelli , b. chapleau and a. e. cholakian _ et al .
_ , j. phys .
g * 39 * , 063001 ( 2012 ) [ arxiv:1201.0008 [ hep - ph ] ] .
g. d. kribs , a. martin , t. s. roy and m. spannowsky , phys .
d * 81 * , 111501 ( 2010 ) [ arxiv:0912.4731 [ hep - ph ] ] .
a. katz , m. son and b. tweedie , jhep * 1103 * , 011 ( 2011 ) [ arxiv:1010.5253 [ hep - ph ] ] .
y. cui , z. han and m. d. schwartz , phys . rev .
d * 83 * , 074023 ( 2011 ) [ arxiv:1012.2077 [ hep - ph ] ] .
m. son , c. spethmann and b. tweedie , jhep * 1208 * , 160 ( 2012 ) [ arxiv:1204.0525 [ hep - ph ] ] .
j. m. butterworth , j. r. ellis , a. r. raklev and g. p. salam , phys .
lett . * 103 * , 241803 ( 2009 ) [ arxiv:0906.0728 [ hep - ph ] ]
. j. fan , d. krohn , p. mosteiro , a. m. thalapillil and l. -t .
wang , jhep * 1103 * , 077 ( 2011 ) [ arxiv:1102.0302 [ hep - ph ] ] .
y. bai and j. shelton , jhep * 1207 * , 067 ( 2012 ) [ arxiv:1107.3563 [ hep - ph ] ] .
m. baumgart , c. cheung , j. t. ruderman , l. -t .
wang and i. yavin , jhep * 0904 * , 014 ( 2009 ) [ arxiv:0901.0283 [ hep - ph ] ] . c. cheung , j. t. ruderman , l. -t .
wang and i. yavin , jhep * 1004 * , 116 ( 2010 ) [ arxiv:0909.0290 [ hep - ph ] ] . c. -r .
chen , m. m. nojiri and w. sreethawong , jhep * 1011 * , 012 ( 2010 ) [ arxiv:1006.1151 [ hep - ph ] ] .
d. e. kaplan and m. mcevoy , phys .
d * 83 * , 115004 ( 2011 ) [ arxiv:1102.0704 [ hep - ph ] ] . c. englert , t. s. roy and m. spannowsky , phys .
d * 84 * , 075026 ( 2011 ) [ arxiv:1106.4545 [ hep - ph ] ] .
i. lewis and j. schmitthenner , jhep * 1206 * , 072 ( 2012 ) [ arxiv:1203.5174 [ hep - ph ] ] .
p. draper and d. mckeen , phys .
d * 85 * , 115023 ( 2012 ) [ arxiv:1204.1061 [ hep - ph ] ] .
a. hook , e. izaguirre , m. lisanti and j. g. wacker , phys .
d * 85 * , 055029 ( 2012 ) [ arxiv:1202.0558 [ hep - ph ] ] .
g. p. salam , eur . phys .
j. c * 67 * , 637 ( 2010 ) [ arxiv:0906.1833 [ hep - ph ] ] . | jet physics is a rich and rapidly evolving field , with many applications to physics in and beyond the standard model .
these notes , based on lectures delivered at the june 2012 theoretical advanced study institute , provide an introduction to jets at the large hadron collider .
topics covered include sequential jet algorithms , jet shapes , jet grooming , and boosted higgs and top tagging . |
chiral perturbation theory ( chpt ) allows to systematically investigate the consequences of the spontaneous and the explicit chiral symmetry breaking in qcd . with the advent of cw machines ,
pion production by real and virtual photons has become a major testing ground for predictions based on nucleon chiral dynamics . in particular , over the last years there has been considerable activity to precisely measure neutral pion electroproduction in the threshold region at various laboratories , in particular at nikhef ( amsterdam ) and mami ( mainz ) .
these measurements are performed at energies very close to threshold and typical photon momentum transfer squared of @xmath6 .
after the pioneering measurement of welch et al .
@xcite at nikhef , which concentrated on the s wave cross section , kinematically more complete differential cross sections for various photon polarizations @xmath7 are now available based on the data of van den brink et al .
@xcite ( nikhef ) and distler et al .
@xcite ( mami ) .
furthermore , data at lower @xmath8 have also been taken at mami @xcite . in the review
@xcite we already considered charged and neutral pion electroproduction off protons and neutrons in the framework of relativistic nucleon chpt @xcite to order @xmath9 ( here , @xmath10 denotes the small expansion parameter which can be an external momentum or meson mass and the order @xmath11 refers to the lagrangian ) .
since then , the theoretical framework has been considerably refined based on the heavy mass expansion proposed in @xcite and discussed in detail in the recent review @xcite . in this framework and calculating the s waves to order @xmath3 , we have already reexamined neutral pion photoproduction @xcite .
novel p wave low energy theorems ( lets ) have been found together with a good description of the new taps and sal data for @xmath12 @xcite @xcite . in particular , chiral loops are necessary to understand the small value of the electric dipole amplitude @xmath0 at the @xmath13 and @xmath14 thresholds .
furthermore , the p wave lets have been shown to hold within 10% or better . while @xmath15 can be inferred directly from the unpolarized differential cross section , for @xmath16 one has to measure polarization observables .
a measurement of the photon polarization asymmetry is underway e.g. at mami .
however , there are somewhat model dependent means to indirectly determine @xmath16 , see e.g. @xcite and @xcite . these let predictions extended to the case of virtual photons @xcite were also used in the analysis of the nikhef data @xcite .
the investigation presented here is thus a natural extension of the one for photoproduction presented in @xcite . to orders @xmath9 and @xmath3
, there appear two new counter terms with a priori undetermined low energy constants ( lecs ) .
one can be fixed from a recent dispersion theoretical fit to the nucleon electromagnetic form factors , i.e. from the radius of the proton dirac form factor @xcite .
the other one will be determined from a best fit to the differential cross sections at @xmath17 .
as we will show , due to a soft pion theorem , these fits are too constrained since they lead to a @xmath18dependence of the longitudinal s wave multipole @xmath1 only through born and one loop graphs .
we therefore are forced to include the leading corrections to this soft pion theorem in the counter terms which is formally of order @xmath4 in the lagrangian .
this in turn leads to a satisfactory fit of the existing data at @xmath19 gev@xmath20 and we then make predictions for smaller values of @xmath8 , where this dimension five operator is much less important . in addition , three @xmath18independent low
energy constants are taken from the photoproduction calculation to order @xmath3 @xcite .
the paper is organized as follows . in section 2
, we briefly summarize the pertinent kinematics and define the various cross sections and structure functions .
we also present the natural basis of s and p wave multipoles which appear in the transition matrix element . in section 3 ,
the chiral expansion of the s wave multipoles @xmath0 and @xmath1 is given to three orders in small momenta and the five combinations of p wave multipoles to two orders .
the corresponding low
energy theorems were already given in @xcite .
we also introduce an approximation which facilitates the calculation of the energy
dependence of the s wave multipoles . in section 4 , we give the explicit expression for the various low energy constants based on the resonance saturation principle . in section 5 ,
the best fits to the differential cross sections are presented together with a detailed analysis of the various multipoles at @xmath17 . in section 6 ,
we give predictions for the new data taken at @xmath21 and investigate in some detail the @xmath8 range of @xmath22 gev@xmath20 .
a short summary together with a discussion of possible improvements is given in section 7 .
the appendix contains some consideration about the s wave lecs at order @xmath3 .
in this section , we assemble all the necessary definitions for the cross sections , structure functions and multipoles . to be specific , consider the process @xmath23 where @xmath24 is the virtual photon with @xmath25 .
denote by @xmath26 the cm energy of the @xmath27 system and its threshold value by @xmath28 , with @xmath29 and @xmath30 the ( neutral ) pion and the proton mass , respectively .
the following quantities are used @xmath31 with @xmath10 the pion cm momentum , @xmath32 the photon energy , @xmath33 the photon polarization , @xmath34 the longitudinal photon polarization , @xmath35 the electron scattering angle , @xmath36 the pion energy in the @xmath27 cm system and @xmath37 gives the invariant energy above threshold . as explained in some detail in @xcite ,
we will account for the pion mass difference @xmath38 in the loops but keeping one nucleon mass .
the @xmath14 threshold is located at @xmath39 which will also be used as the charged pion mass .
we introduce the ratio @xmath40 and as a further dimensionless quantity of order one @xmath41 @xmath42 varies between 0.93 and 1.12 for @xmath43 mev
. the unpolarized pion electroproduction triple differential cross section reads @xmath44 with @xmath45 the conventional virtual photon flux factor , @xmath46 the fine structure constant and @xmath47 is the laboratory energy of the incoming / outgoing electron .
the differential cross section can be split into transverse ( @xmath48 ) , longitudinal ( @xmath49 ) , transverse longitudinal ( @xmath50 ) and transverse - transverse ( @xmath51 ) terms , @xmath52 where the @xmath53 ( @xmath54 ) are called the structure functions .
@xmath55 is the azimuthal angle between the scattering and the reaction plane , and the corresponding polar angle , spanned by the photon and pion directions , is called @xmath56 , @xmath57 one also uses the separated virtual photon cross sections @xmath58 this completes the necessary definitions of kinematical quantities and cross sections . in what follows
, we will consider the threshold region , i.e. small values of @xmath37 , typically @xmath59 and small photon virtualities , @xmath60 . in that case , the pion three momentum is small and it is therefore advantageous to perform a multipole decomposition .
we confine ourselves to the s and p waves in what follows .
consequently , the current matrix element takes the form @xmath61 in terms of the two s
waves @xmath0and @xmath1 and five p waves .
we choose the following combinations of the more commonly used p wave multipoles @xmath62 , @xmath63 and @xmath64 , where @xmath65 stands for electric , magnetic and longitudinal , respectively and the @xmath66 refers to the total angular momentum of the pion
nucleon system , @xmath67 ( with @xmath68 the pion angular momentum ) , @xmath69 these are the combinations which appear naturally in the transition matrix element and lead to the most compact formulae for the various cross sections .
all multipoles are , of course , functions of the pion energy and the photon four momentum squared , like e.g. @xmath70 . in
what follows , we will drop these obvious arguments .
the structure functions @xmath53 can be expressed in terms of the multipoles as follows ( in the s and p wave approximation ) @xmath71 \ , , \nonumber \\
r_{tt } & = & \frac{1}{2 } \sin^2 \theta\ , ( |p_2|^2 - |p_3|^2 ) \ , \ , .
\label{rmult}\end{aligned}\ ] ]
to perform the calculations , we make use of the effective goldstone boson baryon lagrangian .
our notation is identical to the one used in @xcite and we discuss here only some additional terms .
the effective lagrangian takes the form ( to one loop accuracy ) @xmath72 where the chiral dimension @xmath73 counts the number of derivatives and/or meson mass insertions .
we will work out the s wave multipoles to order @xmath9 and the p waves to order @xmath74 as explained in some detail in ref.@xcite . here
, the order refers to the various multipoles .
the corresponding lagrangians are one order higher since the photon polarization vector counts as order @xmath10 . for the p waves , we thus have to consider tree graphs with insertions from @xmath75 and one loop graphs with insertions solely from @xmath76 . for the s wave multipoles , we have in addition to consider one loop diagrams with exactly one insertion from @xmath77 and tree graphs from @xmath78 . in comparison to the photoproduction calculation
, we have two additional terms with undetermined low - energy constants .
one is from @xmath79 and is related to the dirac form factor of the proton , @xmath80 , and was already considered in the relativistic calculation in @xcite . in the heavy fermion approach
used here , the effect due to the finite charge radius of the proton manifests itself via @xmath81 + \delta r_{1p}(\lambda ) \ , \ , \ , , \label{f1v}\ ] ] with @xmath82 the scale of dimensional regularization , @xmath83 the axial vector coupling constant and @xmath84 the pion decay constant .
the terms in the square brackets in eq.([f1v ] ) stem from the loops .
the constant @xmath85 can be fixed from the recent dispersion theoretical fit to the nucleon electromagnetic form factors , @xmath86 @xcite .
the other counter term is from @xmath78 and appears in the chiral expansion of the s wave multipoles .
we do not need the explicit form of this term in the effective lagrangian in what follows and thus refrain from giving it here . since with the sole @xmath18dependent counter term from @xmath78 togther with the radius correction from @xmath79 one is not able to describe the existing data at rather large @xmath87 . ]
consider first the chiral expansion of the p wave multipoles to order @xmath9 .
this should be rather accurate for the large multipoles but is afflicted with some uncertainty for the small ones , compare e.g. the discussion by bergstrom @xcite for the photoproduction case .
he finds a good description of the large multipole @xmath88 but some sizeable deviation for the much smaller @xmath62 .
as we will show later on , the existing electroproduction data are not yet accurate enough to pin down the small multipoles with great precision and we thus stick to the @xmath9 approximation .
the chiral expansion of the p - wave multipoles thus takes the form @xmath89 with @xmath90 and @xmath91 for @xmath92 .
the born terms include the contribution from the proton anomalous magnetic moment @xmath93 , which to lowest order appears in the dimension two pion nucleon lagrangian .
for the large multipoles @xmath94 we have the following born ( and counter ) terms ( in case of @xmath95 ) , @xmath96 \biggr\ } \ , , \nonumber \\ & & \nonumber \label{p1born}\end{aligned}\ ] ] @xmath97 \biggr\ } \
, , \nonumber \\ & & \nonumber \label{p2born}\end{aligned}\ ] ] @xmath98 with @xmath99 the strong pion
nucleon coupling constant .
we notice that the p waves scale with the pion momentum @xmath10 .
the constant @xmath100 has been determined from a best fit to the new taps data for @xmath12 , @xmath101 , a value which is well understood in terms of @xmath102 and vector meson exchange . however , there are also new data from sal for the same process .
the analysis of these gives a somewhat higher total cross section above @xmath14 threshold @xcite . to account for this
, we also use the larger value of @xmath103 as determined in @xcite .
the small multipoles @xmath104 have the following born terms @xmath105 all these born terms are , of course , real .
the one loop contribution to the @xmath106 @xmath107 takes the form @xmath108 \ , \ , , \label{p1loop}\end{aligned}\ ] ] @xmath109 \ , \ , , \label{p2loop}\end{aligned}\ ] ] @xmath110 \ , \ , , \label{p4loop}\end{aligned}\ ] ] @xmath111 \ , \ , , \label{p5loop}\end{aligned}\ ] ] with @xmath112 loop effects which renormalize physical quantities like e.g. the anomalous magnetic moment @xmath93 or the pion nucleon coupling are properly taken care of . for @xmath113 , these loop contributions become complex .
this has to be accounted for by the substitutions @xmath114 and @xmath115 for the imaginary parts of the p wave multipoles , one can give a rather compact and accurate description based on the fermi
watson final state theorem .
this serves as a consistency check and gives a handy and rather accurate estimate about the size of the i m @xmath106 .
of course , in the full calculations to be performed later , we are not using these approximate forms .
consider the rescattering diagram shown in fig.1 .
the corresponding imaginary part can be cast into 1.5 in = 1.5 in the product of a charged pion production process followed by a pion
nucleon charge exchange ( cex ) reaction .
consequently @xmath116 where @xmath117 is a generic symbol for any one of the p wave multipoles @xmath62 , @xmath63 , @xmath64 and @xmath118 for the @xmath119 ( @xmath120 ) multipoles , @xmath121 denotes the corresponding charge
exchange ( cex ) scattering volume and @xmath122 .
to lowest order , these scattering volumina are given by @xmath123 furthermore , the p
waves for @xmath124 can be easily evaluated in born approximation from the pion pole diagram , @xmath125 \ , \ , \ , , \label{pi+}\ ] ] where the terms in the square brackets come from the expansion of the pion propagator .
evaluating the operator structure of eq.([pi+ ] ) fixes the @xmath126 . combining these with cex scattering volumina and the appropriate kinematical factors
, we have @xmath127 this shows that @xmath16 has the largest imaginary part .
these approximations become very accurate as @xmath128 approaches zero .
isospin breaking effects not taken into account in such a consideration are of the order @xmath129 .
we now turn to the s wave multipoles @xmath0 and @xmath1 .
let @xmath130 be a generic symbol for either one of them .
the chiral expansion carried out to order @xmath3 takes the form @xmath131 as already mentioned , @xmath132 . however , from our photoproduction study @xcite , we know that in the threshold region , @xmath133 mev , the @xmath36dependence is mostly due to the unitarity cusp which is fixed by the fermi watson final state theorem .
stated differently , this accounts for the dominant isospin breaking effect due to the sizeable pion mass difference , @xmath134 .
we will discuss the numerical implications of this approximation later on .
consequently @xmath135 and the chiral expansion , eq.([s0+gen ] ) , is applied to @xmath136 .
this means that we can calculate @xmath137 at threshold in the isospin limit .
the born terms are readily evaluated , @xmath138 \nonumber \\ & & \qquad \quad + \frac{m_\pi } { 8m^2 } \bigl [ ( 5 + 4\kappa_p)k^2 - 3(5 + 2\kappa_p)m_\pi^2\bigr ] \biggr\ } \\ l(k^2)^{\rm born } & = & \frac{e g_{\pi n } } { 8 \pi m^2 } \biggl\ { -m_\pi + \frac{3m_\pi^2-k^2 } { 2 m } + \frac{m_\pi } { 8 m^2 } \bigl [ ( 5 - 2\kappa_p)k^2 - 15m_\pi^2\bigr ] \biggr\ } \nonumber \ , \ , . \label{sborn}\end{aligned}\ ] ] the order @xmath9 one loop contributions are given by @xmath139 \ , \
, , \nonumber \\
l(k^2)^{\rm q^3-loop } & = & \frac{e g_{\pi n } \omega_c^2 } { 128 \pi^2 m f_\pi^2 } \biggl [ \frac{2}{1+\rho } + \frac{\rho}{(1+\rho)^{3/2 } } \arccos \frac{-\rho } { 2+\rho } \biggr ] \ , \ , .
\label{s3loop}\end{aligned}\ ] ] these are finite and can be checked against the result of the relativistic calculation presented in @xcite .
to get an idea about the uncertainties induced by the approximation on the energy dependence discussed above , we have also evaluated to this order the full @xmath140 .
the difference of the full result and the approximation takes the form @xmath141 - \frac{\rho}{1+\rho } - \frac{(2+\rho)^2}{2(1+\rho)^{3/2 } } \ , \arccos \frac{-\rho}{2+\rho } - 4 \sqrt{1-y } \biggr\ } \ , \ , \ , , \nonumber \\
\delta l_{0+}^{\rm q^3-loop } & = & \frac{e g_{\pi n } \omega_c^2}{128 \pi^2 m f_\pi^3 } \biggl\ { \frac{y^{3/2}}{y+\rho } \ , \biggl [ 2 - 2 \sqrt{1-y}+\frac{\rho}{\sqrt{y+\rho } } \ , h(y , \rho ) \biggr ] \nonumber \\ & - & \frac{2}{1+\rho } - \frac{\rho}{(1+\rho)^{3/2 } } \ , \arccos \frac{-\rho}{2+\rho } + \frac{4}{2+\rho } \sqrt{1-y } \biggr\ } \ , \ , \ , . \label{s3fumapp}\end{aligned}\ ] ] we will come back to this later on .
the order @xmath3 contributions to @xmath142 can be grouped into the graphs which are proportional to @xmath83 and the ones proportional to @xmath143 .
these are no longer finite and depend on the scale of dimensional regularization , @xmath82 .
the counter terms to be discussed below have an infinite piece to cancel these divergences and their scale dependence is such that it cancels the one from the @xmath3 loops
. the explicit form of these terms is ( we first give the result of the @xmath143 graphs ) @xmath144 @xmath145 similarly , one finds for the diagrams proportional to @xmath83 ( these form an independent subset of gauge invariant feynman diagrams ) @xmath146 @xmath147 notice that the last integrand contains an integrable singularity of the type @xmath148 for all @xmath149 . as a check , the last integral in eq.([l0+ga ] ) at @xmath150 gives @xmath151 .
the one loop corrections to order @xmath3 within the approximation on the energy
dependence are thus determined . finally , we turn to the counter terms at this order .
first , there are the two terms from @xmath78 which appeared already in the photoproduction case .
we have refitted the sum @xmath152 since in refs.@xcite the full @xmath36dependence of @xmath0 was considered . within the same approximation used here ,
the value of @xmath152 is @xmath153 as compared to @xmath154 from the full @xmath36dependence .
we will also use this latter value as a measure for the theoretical uncertainty of our calculations . for the s
waves , we have a priori two new counter terms at order @xmath3 , @xmath155 so that @xmath156 .
however , as proven in the appendix , in the soft pion limit one can show that @xmath157 and thus there is only one lec and furthermore a strong correlation between @xmath158 and @xmath159 at order @xmath3 .
the low energy constant @xmath160 will be treated as a free parameter and pinned down by a fit to the available differential cross section data at @xmath161 gev@xmath20 .
it turns out , however , that with a @xmath18independent @xmath162 , i.e. with the @xmath18dependence of @xmath163 coming solely from the born and loop graphs , one is not able to fit the existing data .
we therefore have to include the first corrections to the soft pion constraint eq.([a34con ] ) away from the chiral limit .
this induces terms of the type @xmath164 which are arising from terms in the lagrangian @xmath165 and are thus of higher order .
these are the minimal terms one has to take to be able to describe the data at @xmath19 gev@xmath20 .
of course , there are other counter terms at this order .
these , however , merely amount to quark mass renormalizations of the already considered @xmath18independent counter terms and will therefore be set to zero here . in the next section , we show how to estimate the pertinent low
energy constants from resonance exchange .
a fully consistent @xmath166 calculation would also include two loop graphs and one loop graphs with insertions from @xmath79 ( besides others ) .
we take here the pragmatic approach and subsume all these effects in the effective couplings entering the higher order low energy constants . in addition , the form factor correction due to the finite proton dirac radius , cf .
eq.([f1v ] ) , contributes as follows , @xmath167 and is completely fixed from the knowledge of @xmath168 @xcite , @xmath169 using @xmath83 as determined from the goldberger
treiman relation , @xmath170 .
this contribution is always treated separately .
we now turn to the estimate of the size of the resonance contributions to the @xmath171 ( @xmath172 ) and the p wave low energy constant @xmath100 .
in this section , we consider the resonance saturation hypothesis to estimate the numerical values of the various low energy constants .
this principle works very well in the meson sector , see e.g. refs.@xcite and is also a good tool to estimate the lecs in the presence of baryons , see e.g. @xcite and @xcite . in the baryon sector ,
resonance saturation proceeds in two steps .
first , the effective field theory contains mesonic ( @xmath173 ) and baryonic @xmath174 ) excitations chirally coupled to the goldstone bosons and the nucleons . considering these excitations as very heavy , but keeping the ratios of coupling constants to masses fixed , one produces a string of higher dimensional operators involving only pions and nucleons with coefficients given in terms of the resonance parameters .
second , one then performs the heavy mass expansion for the nucleons , i.e. @xmath175 \to \bar{{\cal l}}_{\rm eff } [ u , n ]
\to { \cal l}_{\rm eff } [ u , n_v ] \ , \ , \ , , \label{chain}\ ] ] where @xmath176 denotes the velocity dependent heavy nucleon field ( in general , we suppress the subscript @xmath177 ) .
first , consider t
channel vector meson exchange , in this case the coupling of the @xmath178 and the @xmath36 to the nucleon and the subsequent vector meson decay into the @xmath179 ( a general discussion concerning the coupling of spin1 fields to the pion
nucleon system has recently been given @xcite ) .
this leads to @xmath180 where the @xmath181 couplings can be determined from the radiative widths @xmath182 , @xmath183 .
the tensor to vector coupling ratio @xmath184 is known to be large for the @xmath128 , @xmath185 , small for the @xmath36 , @xmath186 .
the @xmath187 coupling constant is fairly well known whereas there is some sizeable uncertainty concerning @xmath188 . to avoid these uncertainties
, we use a simpler form based in part on the gauged wess
zumino action , using the ksfr relation and setting @xmath189 , @xmath190 ( for details , see @xcite ) , @xmath191 leading to @xmath192 and @xmath193 ( see also @xcite ) . from the s channel ( baryonic )
excitations , the dominant one is @xmath194 exchange .
we use the following @xmath195 and @xmath196 lagrangians @xmath197 with @xmath198 @xmath199 and @xmath200 denote the relativistic spin1/2 ( dirac ) and spin3/2 ( rarita schwinger ) field , respectively .
the third term in @xmath201 vanishes for real photons @xcite .
the off shell parameters @xmath202 , @xmath203 and @xmath204 are severely constrained from @xmath27 scattering , the nucleon polarizabilities and the fit to the new taps data for neutral pion photoproduction , for details see ref.@xcite . using eqs.([ldelta ] ) , one finds for the @xmath102 contribution to the @xmath171 ( to order @xmath3 in the lagrangian ) @xmath205 \nonumber \\ & & \quad + g_3 \bigl [ m(2x'+1)(2z-1 ) - 2m_\delta(1+x'+z+4x'z)\bigr]\biggr\ } \ , \ , \ , .
\label{aidelta}\end{aligned}\ ] ] as discussed in the previous section , we have to account for the first correction to the soft pion theorem @xmath206 .
this is entirely given by @xmath102exchange and contributes to @xmath1 as follows , is the one which really gives the first correction to the soft - pion theorem whereas the equivalent @xmath0 term is a quark mass renormalization of @xmath160 . ]
@xmath207 with @xmath208 \nonumber \\ & + & g_2\biggl [ \frac{x}{2}(4z-1 ) + \frac{3m_\delta^2 - 4m_\delta m + 3m^2}{4 m ( m_\delta - m ) } \,\biggr ] \nonumber \\ & & + g_3 \biggl[z - x ' - \frac{m_\delta}{m}(1+x'+z+4x'z)\biggr]\biggr\ } \ , \ , \ , , \label{aq5l}\end{aligned}\ ] ] and similarly for @xmath0 @xmath209\frac{1}{(m - m_\delta)^2 } \nonumber \\ + g_2 \biggl [ m ( 18yz&-&7y+9z-\frac{7}{2 } ) - m_\delta ( 30xz-4x+12z+\frac{1}{2 } ) \nonumber \\ & & \quad \qquad\quad \qquad + \frac{m^2_\delta}{m}(12xz+3x+3z + \frac{1}{2 } ) \biggr]\frac{4}{m - m_\delta } \nonumber \\
+ g_3 \biggl [ -8x'z&-&4z+4x'+2 + \frac{4m_\delta}{m } ( 1 + 4x ' z + x ' + z ) \biggr ] \biggr\ } \label{aq5e}\end{aligned}\ ] ] a good check on the lengthy expressions eqs.([aq5l],[aq5e ] ) ( together with the analogous terms of order @xmath210 not shown here ) is that they fulfill the constraint @xmath156 .
furthermore , @xmath211 is the same as in the photoproduction case @xcite , @xmath212 the couplings @xmath213 , @xmath214 and the off
shell parameter @xmath215 have been previously determined from neutral pion photoproduction , @xmath27 scattering and the nucleons electromagnetic polarizabilities .
nothing is known about the signs and magnitudes of @xmath216 and @xmath217 .
these only show up in reactions with virtual photons . in what follows
, we will use these two parameters to obtain a best fit to the mami and nikhef data at @xmath161 gev@xmath20 .
since we do not constrain the magnitudes of @xmath216 and @xmath217 , this effectively amounts to a free fit with @xmath218 .
we finally remark that @xmath216 and @xmath217 also enter virtual compton scattering and can eventually be pinned down within some broad ranges in the future .
in this section , we show the combined fit to the nikhef ( @xmath219 ) @xcite and the mami ( @xmath220 and @xmath221 ) @xcite data at @xmath161 gev@xmath20 and the resulting multipoles .
note that the mami data are @xmath55integrated differential cross sections involving only @xmath222 and @xmath223 . in figs.2a , b , the dashed lines give the best fit at order @xmath3 , i.e. with the soft pion constraint @xmath224 .
clearly , this does not describe the data . if one adds , however , the order @xmath4 counter terms for the s wave multipoles as discussed in the previous sections , one finds an acceptable fit ( solid lines ) with a @xmath225/dof of 2.296 . in the resonance exchange picture
, this amounts to @xmath226 clearly , such a large value of the unknown third @xmath227 coupling constant indicates that one subsumes in its value other effects like e.g. from two loop graphs .
alternatively , one can only take the @xmath3 counter terms and relax the constraint @xmath228 in the fit .
this leads to @xmath229 gev@xmath230 and @xmath231 gev@xmath230 , which are numbers of natural size .
we remark that in a combined fit to the nikhef and the mami data , it is not possible to correctly get the normalization of the two mami data sets for the two different values of the photon polarization @xmath33 , compare fig.2b .
.6 in = 5.5 in fig .
2a : nikhef data .
open squares ( circles ) : @xmath232 . 3.2 in = 3.2 in = 3.2 in fig .
2b : mami data .
left ( right ) panel : @xmath233 .
the corresponding multipoles are shown in fig.3 for @xmath234 mev , all in units of @xmath235 .
for the s waves , we give the real and the imaginary parts , indicated by the solid and dashed lines , respectively .
we remark that re @xmath0 has changed sign as compared to the photoproduction case , it shows the typical cusp effect at the opening of the @xmath236 threshold .
in contrast , re @xmath1 1.in = 3.8 in fig .
3 : multipoles at @xmath161 gev@xmath20 . see text for notations .
is essentially energy
independent with a very small cusp . for the p waves , we only show the real parts since the imaginary parts are very small .
the large p
waves are all of the same size as shown in the lower left panel ( @xmath237 : solid , dashed , dash dotted line , in order ) whereas @xmath238 and @xmath239 are more than one magnitude smaller , see the lower right panel in fig.3 ( @xmath240 : solid and dashed line , respectively ) . in fig.4 , we show the effect of the approximation on the @xmath36dependence . without refitting any parameters
, we have added the correction eq.([s3loop ] ) to the s wave multipoles .
the solid curves in fig.4 refer to the approximation and the dash dotted ones to the complete @xmath36dependence in the loops to order @xmath9 .
we see that the effect on @xmath0 is small and somewhat more pronounced in @xmath1 .
we remark that these curves should only be considered indicative since there is also a difference in the @xmath3loops , which we did not evaluate in detail .
furthermore , part of this effect would be absorbed in the values of the lecs which we used in the fit .
1.in fig .
4 : approximation on the @xmath36dependence at order @xmath9 .
furthermore , we have also performed some fits with the larger value of @xmath241 gev@xmath242 .
it leads to a somewhat larger ( in magnitude ) value of @xmath243 and @xmath244 , consistent with the value given in eq.([fitval ] ) .
the @xmath225/dof is 2.945 , i.e. worse than for the smaller value of @xmath245 gev@xmath242 . to improve the fits with the larger @xmath100 , one thus would have to readjust the parameters in the photoproduction case . clearly , this discrepancy deserves further experimental clarification . in
what follows , we will always use the smaller value of @xmath246 .
we would like to stress again , see also @xcite , that for a good test of chiral dynamics , one needs data at lower photon virtualities as witnessed by the large value we find for the coupling constant @xmath216 and the necessity of using a particular dimension five operator .
however , as @xmath8 decreases , the latter becomes less and less important .
finally , we note that the convergence in the s wave loops and counter terms is slow , similar to the photoproduction case @xcite , whereas the born terms at order @xmath4 can safely be neglected .
having fixed all parameters at @xmath247 gev@xmath20 , we can now make predictions for lower photon virtualities .
we will concentrate here on the range of @xmath8 between 0.04 and 0.06 gev@xmath20 since at the upper end some data have been taken at mami and it is conceivable that in the near future one will not be able to go below 0.04 gev@xmath20 since the pion has to leave the target . in figs . 5 and 6
, we show the real parts of the s wave and the large p wave multipoles as a function of @xmath37 for @xmath248 gev@xmath20 ( solid , dashed dotted and dashed lines , in order ) . at threshold ,
re @xmath0 passes zero for @xmath249 gev@xmath20 . 0.8 in = 2.2 in fig . 5 : predictions for the s wave mulitpoles .
for notations , see text .
= 2.3 in fig . 6 : predictions for the p wave mulitpoles . for notations , see text . for the new mami data
, we show the various differential cross sections at @xmath250 gev@xmath20 with @xmath220 and @xmath251 and @xmath252 mev as indicated by the solid and dashed dotted lines , respectively .
1.in = 3.5 in fig . 7 : predictions for the differential cross sections .
for notations , see text .
the s wave cross section @xmath253 , defined as @xmath254 is shown in fig.8 .
the data of welch et al .
@xcite ( open squares ) are taken at different values of @xmath33 .
this range is indicated by the dotted lines in fig.8 .
the solid line refers to @xmath219 , .8 in = 2.5 in fig . 8 : s
wave cross section . for notations , see text .
i.e. the kinematics of van den brink et al .
@xcite ( crosses ) . within the sizeable uncertainty ,
the s wave cross section @xmath253 as predicted by the calculation is consistent with the data and shows again the flattening with increasing @xmath8 as already found in the relativistic calculation @xcite .
in this paper , we have considered neutral pion electroproduction off protons in the framework of heavy baryon chiral perturbation theory .
the main results can be summarized as follows : 1 .
extending the successful photoproduction calculation @xcite @xcite to account for all @xmath18dependent counter terms at order @xmath3 does not give a decent fit to the existing electroproduction data at the rather large @xmath161 gev@xmath255 .
this can be traced back to a particular soft
pion theorem which severely constrains the possible local dimension four contact terms .
we are thus forced to take into account a particular dimension five operator which gives the first correction to the soft pion theorem . with that term
, the data can be fitted .
we have given predictions for the @xmath8 range between 0.04 and 0.06 gev@xmath20 , where new data have been and are going to be taken .
at threshold , the real part of @xmath0 changes sign at @xmath256 gev@xmath20 .
the calculated s wave cross section is in satisfactory agreement with existing determinations .
more detailed predictions pertaining to the relevant kinematics of future experiments are available from the authors .
further improvements of the calculations presented here are : 1 .
the p wave multipoles should be calculated to one order higher .
this would give a more accurate description of the small p waves and can further be used to tighten the predictions of the p wave lets @xcite .
2 . the full energy dependence of the loops should be taken into account when more accurate data will become available .
3 . a more systematic study of higher order effects than done here should be done to eventually pin down the @xmath257 parameters @xmath216 and @xmath217 .
clearly , when more and more accurate data at lower photon virtualities will be available , neutral pion electroprodcution offers a variety of tests of the chiral qcd dynamics .
we are grateful to j.c .
bergstrom , a.m. bernstein , h. blok , m. distler , d. drechsel , o. hanstein , l. tiator , h.b .
van den brink and th .
walcher for useful comments and communications .
[ appa ] here , we wish to proof eq.([a34con ] ) which states that there are no polynomial s
wave counter terms with @xmath258 up - to - and - including order @xmath3 .
for that , we follow appendix e of ref.@xcite and consider the soft pion limit for the process @xmath259 ,
i.e. for @xmath260 ( this includes the chiral limit @xmath261 ) . in this limit , only the nucleon pole graphs ( including the electromagnetic form factors and , in particular , the radius counter term ) remain , for all values of the photon virtuality @xmath18 . the corresponding non
pole amplitude ( denoted by an overbar ) must thus vanish , @xmath262 \
, u_1 = 0 \label{tbar}\end{aligned}\ ] ] where we use the notation of ref.@xcite , i.e. @xmath263 ( see section 2 of that review ) . in the soft pion limit
, the following conditions arise @xmath264 in fact , the @xmath265 , @xmath266 and @xmath267 are individually zero since they are odd under crossing , i.e. proportional to @xmath268 .
modulo irrelevant prefactors and setting @xmath269 , the s wave multipoles are related to the invariant amplitudes @xmath270 and @xmath271 as follows ( we drop the subscript @xmath272 and refer to @xcite for definitions ) @xmath273 \ , \ , , \label{elab}\end{aligned}\ ] ] with @xmath274 and @xmath275 . to order @xmath3 , the polynomial terms ( in @xmath18 ) contributing to @xmath49 , taking into account the crossing properties of the various amplitudes ,
are solely given by @xmath276 , with @xmath277 two cases are possible that could lead to @xmath278 , namely @xmath279 however , gauge invariance ( @xmath280 ( see eq.(2.1 ) of ref.@xcite ) implies that @xmath281 .
furthermore , the second condition from eqs.([abcon ] ) leads to @xmath282 ( because of 1 ) in eq.([a34pol ] ) ) .
clearly , @xmath283 and @xmath284 must be equal and both possibilities in eq.([cases ] ) are ruled out .
this gives the desired result , @xmath224 . to summarize the argument
, the soft pion theorem excludes a polynomial of the type @xmath285 which would give rise to @xmath258 for the non pole contributions .
only the case @xmath286 is allowed leading immediately to the constraint .
evidently , this constraint does not apply to the nucleon pole graphs , compare eq.([a3rad ] ) . | we investigate neutral pion electroproduction off protons in the framework of heavy baryon chiral perturbation theory .
the chiral expansion of the s wave multipoles @xmath0 and @xmath1 is carried out to three orders .
there appear several undetermined low energy constants .
three are taken from a recent study of the new taps threshold @xmath2 photoproduction data , one is fixed from the proton dirac form factor and the novel s
wave constants appearing at orders @xmath3 and @xmath4 are determined from a best fit ( constrained by resonance exchange ) to the recent nikhef and mami data at photon momentum transfer squared @xmath5 .
the inclusion of a particular set of dimension five operators is forced upon the fit by a soft pion theorem which severely constrains the momentum dependence of the longitudinal s wave multipole @xmath1 at order @xmath3 .
we give predictions for lower photon virtualities for the various multipoles and differential cross sections .
further improvements are briefly touched upon . , , |
the molybdenum purple bronze , @xmath5 , is a subject of intensive experimental studies already for more than two decades@xcite , but its unusual properties remain unclear .
several very different experimental probes have been used : angle resolved photo - emission ( arpes)@xcite , scanning tunnelling microscopy ( stm)@xcite , dc@xcite and magneto- resistivity@xcite , thermal conductivity @xcite , optical conductivity@xcite , nernst signal@xcite , muons spectroscopy@xcite , x - rays@xcite , thermal expansion@xcite , neutron scattering@xcite .
although the main effort of those investigations was focused on the nature of a mysterious phase transition at around 25k , interesting knowledge about higher energy phase was also gathered .
certain properties of the one dimensional ( 1d ) metal , the luttinger liquid ( tll ) , have been invoked to explain the measured data@xcite and nowadays the presence of the 1d physics is well established experimentally@xcite , at least in some energy range . on the theoretical side ,
band structure calculations have shown a quasi-1d character of molybdenum purple bronze .
a remarkably simple band structure emerges from rather complex crystal structure . at the fermi surface
there are only two bands , lying very close to each other , in the form of flat sheets dispersing well only along the b - axis .
this gives a hope that purple bronze can indeed be a rare realization of the 1d physics .
the key problem is that several possible mechanisms has been invoked to explain the observed properties , which made the subject quite unclear and controversial . in our opinion
the reason for this situation is that each of previous attempts was focused only on one out of many peculiar properties of @xmath5 and most of them searched for an explanation in the low energy regime ( below @xmath6 ) , where indeed the properties of @xmath5 are the most spectacular .
by now , not enough attention has been paid even to the parameters of the 1d state .
it is only agreed that it emerges at energies as high as 250mev .
the values of these parameters are the first issue one must determine before pursuing research towards low energy regimes .
the unusual physics observed in @xmath5 at the energy scales of order 10mev is obviously a motivation for revising the question of the 1d physics . in order to establish a proper low energy effective theory
one has to begin at highest energies and the step by step move towards the physics taking place around fermi energy .
the first step is to link the results of the dft calculations with the well defined field theory describing the experimental results at around 20mev .
it is this `` high energy '' regime which must be well understood first .
this is the main task of this paper .
the plan of this work is as follows . in sec.[sec : bands ] we begin with a brief introduction of the band structure . in the following section sec.[ssec : tight - bind ]
we propose a tight binding model which is able to approximate bands around the fermi energy @xmath7 , however in addition it contains also the strong correlation terms , beyond the lda - dft . in sec.[ssec : ll - intro ] we give basic notions of 1d physics , used in the rest of the paper .
section sec.[sec : intra ] is dedicated to intra - chain physics : we give values parameterizing strong correlations ( sec.[sec : strint ] ) , estimate tll parameters which this implies ( sec.[sec : ll - res ] ) and give energy scales for the spin sector ( sec.[ssec : spin ] ) . in section
sec.[sec : inter - phys ] we introduce the inter - chain physics .
once again , first we estimate the strength of these interactions ( sec.[ssec : inter - values ] ) and then ( sec.[sec : ladder - ham ] ) cast as many of them as possible into effective ll description , now within the ladder framework . later in sec.[sec : rg ] we study how the non - linear interaction terms will affect the luttinger liquid parameters and what instabilities they can potentially produce . finally in sec.[sec : discus ] we discuss our results for ll parameters in a context of the experimental findings ( sec.[ssec : disc - alpha ] ) , as well as the validity of the 1d approach itself in sec.[ssec : disc - tperp ] .
we also discuss in sec.[ssec : disc - disor ] the role of substitutional disorder in our model .
the conclusions in sec.[sec : concl ] close the paper .
the lattice space group and atomic positions within the structure are known from experiment done by onoda _
et al _ @xcite recently confirmed by the neutron diffraction experiment@xcite .
this rather complicated structure , which consist of well separated slabs parallel to b - c plane , is presented on fig.[fig : structure]a ) .
based on this crystallography knowledge the electronic structure of li@xmath8mo@xmath9o@xmath10 can be calculated using density functional theory ( dft ) _
e.g. _ in the local density approximation ( lda ) .
recent calculations @xcite lead to results which are globally consistent with several previous computations@xcite .
overall the band dispersions agree well with the measured results obtained by angle resolved photoemission spectroscopy ( arpes ) @xcite .
for example it shows a flattening of the two dispersive bands at about 0.4 - 0.5 ev below @xmath7 .
other bands are found at least 0.25 ev below @xmath7 .
the only visible discrepancies are the relative vertical shifts of the bands of order 0.1ev .
the computed ratio between the fermi velocities along b- and c - crystal axis , is about 40 , compatible with the reported anisotropic 1d - like resistivity @xcite .
the velocity along @xmath11 is even smaller . at low energies ( around fermi energy @xmath7 )
all calculations gave qualitatively similar results .
there are only two bands which cross the fermi energy and they are placed very close to each other .
they have a strong dispersion in the @xmath12 direction ( b - axis ) and barely no dispersion in the perpendicular directions .
the two bands originate from a pair of zig - zag chains build out of @xmath13 atoms inside @xmath14 octahedra .
the low energies ( close to @xmath7 ) spectral function is located mostly on @xmath15 ( precisely @xmath16 ) orbitals of @xmath17 and @xmath18 atoms , using notation from fig.[fig : structure]a ) .
these two corners of zig - zag are symmetrically inequivalent , thus dimerization is possible .
however their distinction comes more from out of chain , than in chain , environment .
the difference solely within a chain is hard to notice .
it is also hard to distinguish the corresponding gap on the edge of brillouin zone upon the analysis of the lda band structure ( in the following we this assume that it is not larger than @xmath19 ) .
the standard procedure ( see appendix [ app : tight - bind ] ) is to fit the dft band structure with an effective tight - binding model . along the b - axis ( along the chains )
the lda calculations were done in the reduced brillouin zone ( because of the above described presence of two un - equivalent mo sites ) .
then , in a tight binding approximation , we expect that the band will be back - folded with a gap at the boundary of a reduced brillouin zone which corresponds to the above mentioned strength of dimerization ( not larger than @xmath19 ) . in
li@xmath8mo@xmath9o@xmath10 case , while it is rather easy to distinguish the lower half of dispersion bands ( however there are some peculiarities at @xmath20 ) , the upper bands hybridizes strongly with other bands ( originating mostly from oxygen orbitals ) .
their proper identification is then difficult .
the band gap seems to be quite small , however the presence the hybridization makes the estimate quite difficult . in the following
we assume that it is @xmath21 . along the @xmath22axis
the dispersion relation is quite unusual ( there is a node at zero momenta along this axis ) and this peculiar feature appears within all independent dft calculations@xcite . here
we are also dealing with the reduced brillouin zone , but the presence of the node means that the standard back - folding does not apply at all .
one has to try possible combinations of cosines(see appendix [ app : tight - bind ] ) , keeping in mind two facts : the unequal distances for intra- and inter - ladder hopping and the large inter - ladder distance which implies that the next - nearest neighbor hopping must be very small ( only the light dashed lines on fig.[fig : structure]a ) could give non - zero contribution ) .
we were able to deduce ( see appendix [ app : tight - bind ] for details ) that the hopping in the perpendicular direction @xmath23 is @xmath24 ( certainly smaller than @xmath25 ( @xmath26 ) ) .
one should also notice a large frustration , change of sign , when hopping to next - nearest chain ( interladder hopping ) with the amplitude of this hopping @xmath27 .
this change of sign can be ascribed to a phase shift acquired when hopping between pairs of chains , between two @xmath28 atoms via @xmath29 octahedra ( see fig.[fig : structure]a ) . along the a - axis
we have well separated slabs , with a void between them filled by @xmath30 atoms and mo(6 ) , mo(3 ) tetrahedra .
electrons residing there ( if any ) stay on energies few evs away from @xmath7 .
this explains why lda give a very low dispersion in this direction .
significant deviations from lda were seen in arpes only below 0.2ev . then the two bands seems to merge .
this is also the energy scale revealed by optics@xcite : it shows a formation of a first plasmon edge for electrons along b - axis ( @xmath12 direction ) and a gap value in perpendicular direction .
thus it sounds reasonable to take it as a point where the 1d physics forms .
this is the starting point ( on the high energy side ) of the present study .
\a ) and @xmath23 hoppings as given in the first line of hamiltonian eq.[eq : ham - strong ] are shown , sites 1 and 4 lie within the ladder , site 2 is outside it but
@xmath23 hopping most likely goes through this site.,title="fig : " ] + b ) and @xmath23 hoppings as given in the first line of hamiltonian eq.[eq : ham - strong ] are shown , sites 1 and 4 lie within the ladder , site 2 is outside it but
@xmath23 hopping most likely goes through this site.,title="fig : " ] +
the lda - dft results presented in sec . [ sec : bands ] treat electron - electron interactions at the mean field level .
it is assumed that a single electron moves in an average electrostatic potential .
however it is known that the higher order terms in electron - electron interaction can bring very different physics , particularly in the case of reduced dimensionality .
thus we need to introduce them to our model .
the lda results shows that a majority of carriers density at the fermi energy is localized on zig - zag chains formed by the mo(1 ) and mo(4 ) octahedra ( see fig .
[ fig : structure]a ) ) .
we can safely assume that the low energy physics is described by the dynamics of these electrons .
then the low energy hamiltonian formally writes : @xmath31 where the vectors @xmath32 $ ] , @xmath33 $ ] , @xmath34 $ ] define the @xmath35 crystal lattice , summation runs over all ladder sites and depends on the parity of a given site index along directions @xmath36 and @xmath37 .
the @xmath38 are the electron - electron interactions not included in the ( mean - field ) dft calculation and the sum goes through positions of all carriers .
the @xmath39 are hopping parameters along each crystal axis , as estimated in the previous section .
as discussed there @xmath40 ( down to @xmath41 ) so we can neglect it for the `` high energy '' range we are interested in .
the electrons are moving exclusively within one slab ( and mostly along the zig - zag chains ) . in eq .
[ eq : ham - tight - bind ] we kept only nearest neighbor hopping , since they are expected to be the dominant ones .
the @xmath42 indicate a possibility of dimerization , namely different hoppings along _ i - th _
axis for even and odd bonds . because mo(1 ) and mo(4 ) are crystallographically different we included the possibility that we are dealing with a dimerised chain , half - filled in the reduced brillouin zone .
we denoted the two corresponding hopping by @xmath43 and @xmath44 .
as explained in the previous section we take @xmath45 ( with @xmath46 accuracy ) .
the @xmath47 describe intra- and inter - ladder hopping respectively ( see appendix [ app : tight - bind ] ) . to lighten the notations we will denote from now on @xmath48 by @xmath49 and @xmath50 by @xmath23 , to emphasize that they correspond respectively to hopping along and perpendicular to the chain direction . the resulting simplified tight - binding model is shown on fig . [ fig : structure]b ) .
we simplify the interactions in a similar way , by explicitly considering the intra - chain and inter - chain parts of the interactions .
the microscopic hamiltonian we take is given by @xmath51 where @xmath52 ( see table.[tab : tight - bind ] ) is hopping between nearest neighboring @xmath53 ( denoted @xmath54 ) molybdenum atoms along zig - zag chains ( b - axis ) and @xmath55 are a hoppings in a perpendicular direction , within a slab ( c - axis ) , between nearest - neighbors , @xmath56 is intra - ladder while @xmath57 is inter - ladder .
the values of @xmath55 are estimated in appendix [ app : tight - bind ] and summarized in tab.[tab : tight - bind ] . as already mentioned @xmath58 ,
so we can treat it as a perturbation on the top of the intra - chain physics .
the strong correlations ( last two lines in eq.[eq : ham - strong ] ) are usually parameterized by several quantities , which enter into the effective hamiltonian : @xmath59 is the local on - site interaction between charge densities of opposite spin @xmath60 ( hubbard term ) , @xmath61 is the interaction of two carriers placed inside the same chain at a distance @xmath62 , @xmath63 is the interaction of two carriers placed in two different chains in a distance @xmath64 ( both @xmath61 and @xmath63 are defined assuming an environment for which the lda screening is included ) .
we will discuss their strength in sec .
[ sec : strint ] . in the following sections , it will be shown that the interchain couplings ( both hopping and interactions ) are reasonably weak compared to the intra - chain ones .
we can thus anticipate the need to describe a strongly interacting one dimensional systems .
in such a case , a very convenient formalism to incorporate the strong intra - chain interactions from the start is provided by the luttinger liquid formalism@xcite .
this formalism allows to describe non - fermi liquid state ( tomonaga - luttinger liquid ( tll ) ) that occurs in one dimension as a result of the interactions .
the low energy dynamics of such a state is captured by collective , bosonic modes representing charge and spin density fluctuations .
these fluctuations are connected to the two fields @xmath65 , where @xmath66 for charge fluctuations and @xmath67 for spin fluctuations ( linked with the fluctuations of charge and spin density ) .
these fields have two canonically conjugate fields @xmath68 linked with the respective currents fluctuations . in terms of these collective variables the hamiltonian reads@xcite : @xmath69\ ] ] where @xmath70 .
all the intra - chain interactions conserving the momentum are now only fixing the precise values of the parameters @xmath71 ( the velocities of the corresponding charge or spin density modes ) or @xmath72 ( the luttinger parameters which control the decay of the various correlation functions ) .
in particular the local spectral function behaves as a power law @xmath73 with the following ( interaction dependent ) exponent : @xmath74 for spin - rotationally invariant case @xmath75 , thus interactions , leading to the anomalous behavior , affects only the charge sector
. we can thus take ( [ eq : ll - ham ] ) as our starting point for the chain physics , and study in the bosonized representation the effects of the various inter - chain coupling terms .
this will ensure that at least the upper energy scales of the `` high energy '' regime will be properly treated .
the first step to determine the luttinger parameters @xmath71 and @xmath72 is to estimate the strength of the various interactions , which we do in the following section ( sec.[sec : strint ] ) . at energies lower than the fermi energy @xmath7 only two scattering channels are allowed @xmath76 and @xmath77 ( plus eventually higher harmonics ) .
all density - density interactions , with small @xmath78 momentum exchange , can be incorporated into the hamiltonian given by eq.[eq : ll - ham ] .
they contribute to a highly non - trivial dependence of the tll parameter @xmath79 ( see sec.[sec : ll - res ] ) .
the remaining interaction terms produce non - solvable hamiltonians of sine - gordon type , which in general is a functional @xmath80 $ ] of cosine terms : @xmath81\ ] ] where @xmath82 indicate higher harmonics ( scattering with larger momenta exchange ) . due to these terms
the total problem allows only for approximate solution ( at least in terms of continuous fields theory ) .
the terms present in @xmath83 are derived from large momentum exchange interactions ( of both intra- and inter- chain origin ) .
this is why their presence is analyzed carefully in the further sections .
they are usually treated using renormalization group ( rg ) transformations , which allows to extract the terms that affect the most the tll physics .
this usually enables one to find the existence of gaps in the spectrum of bosonic modes , sometimes even to predict the ground state phase diagram .
this kind of approach will be applied in sec.[sec : rg ] .
let us first consider the physics taking place within a single chain .
we want to obtain ( in sec.[sec : strint ] ) the values of the various intra - chain interactions in ( [ eq : ham - strong ] ) and from there ( in sec.[sec : ll - res ] ) the values of the luttinger parameters for the hamiltonian ( [ eq : ll - ham ] ) .
a fully self - consistent calculation has been performed in @xcite and a value @xmath84ev was found .
this value seems at first sight surprisingly large ( for @xmath15 electrons ) , but a recent study using constrained - rpa @xcite allows to understand this result .
although for bulk mo @xmath85 , it was convincingly shown that the suppression of the pure coulomb value @xmath86 is mostly due to efficient screening in 3d of the d - electrons in our case ( as discussed in details below ) the system is underscreened , which entails also a significantly reduced plasmon frequency in comparison with the pure mo ( from @xmath87 to @xmath88 @xcite ) .
thus the value @xmath89 is justified or probably even modest .
then , if we take previously estimated hopping @xmath49 , we get @xmath90 .
note that this makes the local repulsion by far the largest energy scale in the problem .
in hamiltonian eq.[eq : ham - strong ] we defined intra - chain non - local interaction @xmath61 . in this subsection
we will discuss its strength both in real and in momentum space @xmath92 .
we are going to use shorthand notation @xmath93 , as this is usually the single parameter which enters into so called @xmath94 model for a 1d chain .
this parameter is much more difficult to estimate than @xmath95 ( see sec.[ssec : inter - values ] ) as it involves the dynamics of the 1d metal . rather than trying to estimate it directly from the interaction itself ,
we will simply adjust the value to be put in ( [ eq : ham - strong ] ) and ( [ eq : ll - ham ] ) in order to reproduce the experimental data on optical spectroscopy . the idea@xcite is based on two independent estimates of kinetic energy in the system .
the first one is related ( by the optical sum rule ) to the plasmon ( edge ) frequency @xmath96 and gives us the total possible kinetic energy available for the given number of carriers .
the same quantity can be computed as an integral @xmath97 of the optical conductivity @xmath98 ( taken only over the highest conducting band ) .
the point is that the second estimate gives us the real kinetic energy , renormalized by interactions .
fortunately the necessary data is available for @xmath99 .
the value of of the plasmon frequency can be read out from fig .
2 in ref.@xcite and gives @xmath100 ( @xmath101 extracted from the lda @xcite is even larger @xmath102 ) .
the sum rule integral @xmath97 was in fact already evaluated in fig . 3 of ref.@xcite .
determining the value of the first saturation from the plot has of course some error attached to it .
we take @xmath103 ( @xmath104 ) .
this value is reasonable because at higher energies one expects that the other bands start to intervene .
the rest of the procedure is straightforward , the ratio of bare and interaction suppressed kinetic energy @xmath105 equals to @xmath106 .
then the results of@xcite , together with the previously estimated value of @xmath59 allows to predict @xmath107 which means @xmath108 ( @xmath109 ) .
in addition to the interactions discussed up to now ( which fit within the @xmath94 model for a single chain ) there can also exist interactions between more distant neighbors ( see fig .
[ fig : interactions ] ) .
these interactions are non - zero because of the under - screening , a characteristic property of the 1d systems . when the separation is larger than that between nearest neighbor we can approximate the interaction by a continuum limit @xmath110 where @xmath111 is a lattice constant along the b - axis@xcite and @xmath112 is a real space dielectric constant . in 2d and 3d systems
the electrostatic potential is usually well screened thus the thomas - fermi approximation , already used in the dft calculation , is sufficient . in our problem
this means that for the very long length scales @xmath113 the potential is well screened , @xmath114 ( where @xmath115 is the largest lattice constant , along the a - axis , perpendicular to the slabs planes ) , because then the network of the zig - zag chains can be thought as a homogenous 3d bulk . for shorter distances between the interacting electrons @xmath116 and then the distance between the neighboring chains does matter .
the relation @xmath117 implies that there are a few non - zero interaction terms @xmath61 for @xmath118 .
the screened potential inside 1d wire behaves approximately like @xmath119 @xcite , same relation holds for the screening induced by the presence of other wires @xcite . in total
we approximate that @xmath120 will get renormalized by a factor @xmath121 .
this gives a reasonable estimate for @xmath120 , however one has to be aware that for the value of charge mode tll parameter the whole q structure has its importance .
the values of real and momentum space interactions are summarized in tables tab.[tab : str - coupl - r ] and tab.[tab : str - coupl - m ] .
clearly the density - density interactions caused by @xmath59 and @xmath122 are not small perturbations compared to the kinetic energy .
in fact they are the largest energy scale in the problem , larger than hopping along the chains .
thus they shall strongly affect the charge sector , they are strong enough to give rise to well pronounced non - fermi liquid properties . we thus need to estimate the tll parameters entering ( [ eq : ll - ham ] ) .
as we will prove in later sections the following hierarchy of energies holds @xmath123 ( where @xmath124 is an effective superexchange which determines energy scales for spin sector ) .
we thus start by solving the tll problem for a single chain and will consider later the interchain couplings . even when @xmath125 we stay @xcite in a gapless phase within the tll universality class , but computing the values of the luttinger liquid parameters @xmath79 beyond the weak coupling limit is usually a very difficult problem @xcite , since one can not compute them perturbatively . in the present case
we have to deal with a quarter filled system with a very small doping ( @xmath126 around 2% ) , and a small dimerization ( as we mentioned in sec.[ssec : tight - bind ] it can not be larger than 10% we take moderately 5% in any future calculations ) and several non - zero interaction terms @xmath127 .
as a first approximation we decided to analyze the physics of these zig - zag chains by looking at a quarter filled extended hubbard chain with a local interaction @xmath59 and a nearest neighbor one @xmath122 . if @xmath59 is the larges energy scale in the problem then the charge sector can be mapped onto a spinless chain of fermions ( or an xxz spin chain ) for which the exact tll parameters are known @xcite . for a certain value
@xmath91 we have @xmath128}\ ] ] where @xmath129 is the critical value of intra - chain nearest neighbor interaction for a given value @xmath59 . when @xmath130 the mapping is exact and the critical value is known @xmath131 . in this case we get @xmath132 which sets the lower limit for the tll parameter .
this is a rough estimate , but suggests that for further analysis we should use approaches which are valid for the gapless phase .
a more precise estimate can be obtained from a numerical study of the extended hubbard model , and an estimation of the tll parameters in the usual way via thermodynamic quantities @xcite . from the relevant plots one reads out that our model is located somewhere in the range @xmath133 .
one can try to give an even better estimate using eq.[eq : kxxz ] even when u is not infinite ( but still with u much larger that any other energy scale in the problem ) .
it is commonly believed that the value @xmath129 extracted from the numerics can give quite good approximation when substituted into ( [ eq : kxxz ] ) .
estimates of critical @xmath129 are usually done with higher precision , than for arbitrary @xmath134 . in our case
@xmath135 implies @xmath136 @xcite and @xmath137 from older work by mila and zotos@xcite .
this gives @xmath138 ( with @xmath139 ) or @xmath140 .
we have also followed the critical scaling analysis proposed in @xcite and found similar value @xmath141 .
the exact solutions e.g. bethe ansatz are available for a few special cases .
if we include a ( quite weak ) dimerization and ( very small ) doping then the problem is located far away from any integrable model .
however the influence of both these perturbations are known .
the dimerization is able to lower a bit value of @xmath142 , this effect can be present in particular in our case when we are not far from @xmath143 @xcite .
the doping has the opposite effect , however when one is not very close to critical point @xmath144 and close to commensurate case then the effect is negligible . to conclude : the knowledge about the @xmath145 we have collected above allowed us to give a several estimates based on complementary numerical calculations of @xmath94 models .
all these estimates gives @xmath146 ( we allowed for 10% error in the @xmath145 due to uncertainty of parameters ) .
this is also in agreement with a very recent study for a system with finite range interaction@xcite : @xmath147 when we take similar value of on - site term and account for the presence of @xmath61 interactions up to four - five mo sites .
however in this last paper interactions with an exponential character were studied , which means that they decay much faster than those in our problem .
note that this @xmath142 value is reasonably close to the one @xmath148 that would lead to a quarter - filled mott insulator in the presence of an infinitesimal @xmath91
. the system will be thus very susceptible to the precise value of @xmath91 . given the accuracy of our estimation we will take @xmath149 for further calculations .
this value corresponds to the case for which the @xmath150 charge fluctuations decay with the same exponent than the @xmath151 charge and spin density fluctuations@xcite . by now we have studied interactions in real space ( which we believe can be useful for numerical studies ) , while cosine terms beyond the tll general expression ( [ eq : ll - ham ] ) are defined in the reciprocal space representation .
these values of scattering for large momenta exchange will be used as an input for the discussion done in sec .
[ sec : rg ] .
the hubbard @xmath59 interaction has a form of a delta function in real space thus in momentum space it contributes equally to small and large momenta exchange scattering ( provided they are the intra - chain ones ) .
the previous considerations also imply that , the fourier transformed @xmath92 is a weakly decaying function ( slower than logarithm ) , so it affects the amplitude of scattering processes with large momenta exchange . as it was already expressed above the intrachain interactions
are much larger then @xmath49 thus it is not straightforward to obtain the value of backscattering they cause .
in fact the @xmath149 parameter given above , known from numerical studies of similar models , is a fixed point value @xmath152 which means that it contains a part of these contributions .
precisely it is the part included in the simplified @xmath94 model .
the longer range , slower decaying interaction leads to smaller value of @xmath152 .
this suggests that the proper value of @xmath152 in our problem which contains long range interaction can be even smaller . in our case
the charge sector is renormalized by the possible umklapp terms @xmath153 , so the last statement can be rephrased : the smaller final value of @xmath152 can be linked with larger value of initial , bare @xmath154 .
the precise estimation of the amplitude of @xmath154 is a very difficult task , the detail discussion of the bare _ large - q _ , intra - chain , term will be done in sec .
[ sec : rg ] .
in addition , the @xmath112 when computed using rpa ( beyond standard dft ) gives a well known phenomena , the friedel oscillations ( in real space ) .
the effect comes from the peculiar screening ( singular susceptibility ) at large momenta @xmath155 . in our case ,
due to the value of @xmath156 along chains , it gives an extra gain of energy for electrons located every second site .
this affects the large momentum exchange part of interaction @xmath157 .
the value of this gain can be estimated using the fact that , for 3d metal , the additional oscillating part @xmath158 thus we give an estimation @xmath159 .
it is not as large as u or v , but can be significant if we compare it with @xmath23 .
the u - v model that we have considered so far ( in particular in sec.[sec : ll - res ] ) is of course only an approximation of the intra - chain physics and interactions .
it should capture most of the effects at the energy scales we are considering , however it may slightly underestimate the strength of interactions .
all the interactions considered up to now were connected with the charge sector . in 1d systems , because of the spin - charge separation , electrons spin degree of freedom should be discussed separately . for a half filled chain the knowledge about @xmath49 and @xmath59 allows to estimate the spin - spin exchange ( superexchange ) constant @xmath160 .
this determines the energies at which spin sector starts to play a role .
the problem of purple bronze is more complex since the compound is quarter filled . for such cases the formula for superexchange interaction
can be still obtained from second order perturbation theory@xcite .
it reads : @xmath161 where we have taken into a account the fact that due to dimerization there are two slightly different alternating hopping @xmath162 and @xmath163 along the chain taking @xmath164 and using the values of @xmath59 and @xmath91 obtained in sec.[sec : strint ] we obtain : @xmath165 it is indeed significantly smaller than @xmath49 ( and also the charge sector interactions @xmath166 ) which implies that charge dynamics will dominate the @xmath167 energy range .
however it is a non - negligible value in the sense that even for energies as high as @xmath168 ( in the middle of the considered `` high energy '' regime ) the spin excitations are coherent and dispersion is linear thus a tll description with spin and charge modes is applicable .
the spin - incoherent tll@xcite is not an appropriate framework for our problem . for a spin - rotational , su(2 )
invariant model we have @xmath169 . for the interacting case
the spin velocity @xmath170 is smaller than @xmath171 , in particular @xmath172 .
this last value is comparable with the one observed in experiment@xcite .
the only candidate to break the su(2 ) symmetry would be spin - orbit coupling @xmath173 on the heaviest atom , molybdenum .
a series of experimental and lda studies allow to set its value in bulk bcc mo@xcite to @xmath174 .
this value was obtained for the @xmath175 point of a brillouin zone in the mo - bcc crystal . in our problem
it should be smaller because the active electrons have mostly @xmath16 character ( with a larger * * _ @xmath176-_**number value ) .
thus @xmath173 can be treated as a perturbation for @xmath177 , contrary to @xmath59 for @xmath49 in a charge sector .
this implies that @xmath178 deviates from the non - interacting value @xmath179 much less than @xmath79 .
to be precise from weak coupling theory ( applicable in this case ) we know@xcite : @xmath180 from which we predict @xmath181 which can increase the green s function exponent @xmath182 only by @xmath183 .
the spin sector thus can not be responsible for the experimentally observed values of the @xmath182 exponent ( defined in eq.[eq : alpha - chain ] ) which are of order @xmath184 .
however the influence of the spin - orbit coupling should be taken into account when one studies the physics taking place lower energy scales , which are beyond the scope of this paper .
after giving the description of a single chain physics we move beyond this model and study the strength ( and the role ) of interactions between carriers moving in different chains .
the density - density interactions can be estimated as a standard coulomb interaction between charges in a 3d dielectric , as it was done in ref.@xcite .
those authors gave estimate for a value of interaction from standard electrostatic coulomb law ( with the simplest static screening ) : @xmath185 which in fact gives the value for two electrons in two different 2d slabs .
we are more interested in interactions inside the slab which is obviously larger because ( the smallest ) interchain distance is then @xmath186 .
a more fundamental problem is the value of the dielectric permittivity @xmath187 . in the previous work @xcite the bulk @xmath188
was used , which is typical for bulk semiconductors with similar value of a gap for oxygen states(@xmath189 ) .
taking into account the very weak metallic character along the c - axis the semiconductor approximation is correct . for further neighbors , with many oxygen atoms in between
, one can take the bulk @xmath190 value .
however for the nearest chains , as there is only one raw of oxygen atoms in the space between them , such large @xmath187 value overestimates the screening provided by sparse environment ( a material with locally similar structure ) the measured value is @xmath191@xcite ] .
it is then convenient to take a function @xmath192 such that @xmath193 is reduced by @xmath194 in comparison with @xmath190 . for larger @xmath64 distances at
it saturates to bulk value @xmath190 . with this set of values
we estimate @xmath195 . for @xmath196
the metallic character along b - axis intervene and @xmath197 .
thus the large distance interaction is strongly suppressed .
in such a case the above estimated value for @xmath95 can be taken as the density - density interaction ( @xmath198 ) between the nearest chains .
the above estimate was done between two nearest chains which form a pair ( see fig . [
fig : interactions ] ) . on the other side of each chain
there is another neighbor which is placed two times further .
according to ( [ eq : voutcoulomb ] ) these interactions @xmath199 ( we defined the inter - ladder term in analogy with _
@xmath200 in eq.[eq : tight - bind - disp - def ] ) are at least two times smaller .
we then estimate ( keeping @xmath192 in mind ) @xmath201 .
( dashed lines within a black chain ) and @xmath95 ( dashed lines between the black and blue chains ) are shown .
the numbering of sites is like on fig.[fig : structure].,title="fig : " ] + the treatment of terms with _ large momentum exchange _ is more complex .
there is a @xmath202 term which locks the interchain density wave , the so called @xmath203 , which is shown on the fig .
[ fig : out - instab]a ) . from the figure it is clear that the distances @xmath204 between charges in such configuration are rather large ( around three times larger than the distance entering to previous calculation for @xmath205 ) thus the screening is quite efficient and
@xmath206 $ ] is definitely not reduced , but probably even enhanced with respect to the bulk value .
in fact not only further chains , but also further slabs can intervene ( because @xmath207 )
. there will be either a very efficient screening ( like in a metal ) , or coulomb potential approximation ( but with enhanced @xmath187 ) is still applicable .
we assume , optimistically , the second case and use again ( [ eq : voutcoulomb ] ) .
two ways of proceeding are possible when one is interested in the staggered component of the interaction between two chains . in real space
( direct application of ( [ eq : voutcoulomb ] ) ) one computes interactions with a linear set of dipoles . due to increase of @xmath192 the interaction ( with further dipoles ) decays rapidly , which makes this straightforward approach quite tedious . in the reciprocal space approach
one must take into account the fact that fourier transform of @xmath208 does decay in momentum space ( it will be a @xmath209 decay in a 2d case corresponding to a separate slab ) .
the second approach is simpler when one notice that the previously computed @xmath195 correspond to @xmath210 ( a distance between slabs sets the large distance cut - off ) .
then the @xmath209 scaling allows to estimate that @xmath202 term will be suppressed by an extra factor four .
overall we have : * a factor @xmath211 from the value @xmath212 * a factor @xmath213 from the distance @xmath214 * a factor @xmath215 because we compute the staggered ( large q ) component taking all these factors into account we estimate that @xmath151 component is reduced by at least one order of magnitude which means @xmath216 .
\a ) + @xmath217dw instability shown ( the light , dashed line correspond to the distance @xmath214 ) b ) the @xmath150 @xmath217dw instability shown.,title="fig : " ] + b ) + @xmath217dw instability shown ( the light , dashed line correspond to the distance @xmath214 ) b ) the @xmath150 @xmath217dw instability shown.,title="fig : " ] + there is also the @xmath218 term .
now the distance is smaller , thus @xmath187 is unaffected .
however this term is strongly suppressed due to above mentioned coulomb character of @xmath208 . at short distances corresponding to @xmath219 the interaction
may even have a 3d character ( with no screening ) thus @xmath220 . by a similar reasoning like above we get @xmath221 .
the smallness of inter - chain large momenta terms ( in comparison with density - density term ) implies that the local field corrections are small , so our mean field model of screening is self - consistent .
the influence of the inter - chain exchange interactions ( both @xmath151 and @xmath150 ) on 1d physics will be discussed in sec .
[ sec : rg ] .
to summarize the results for values of the strong coupling parameters found in the last two section we present all parameters in the tables table [ tab : str - coupl - r ] and table [ tab : str - coupl - m ] .
.the real space values of strong coupling parameters ( given in ev ) . [ cols="^,^,^,^,^,^",options="header " , ]
we begin with a fermionic for of considered scattering processes : 1 .
the intra - chain umklapp : @xmath222 \2 .
interchain umklapp scattering ( @xmath4 ) : @xmath223 where the following relation between momenta holds @xmath224 in order to preserve momentum conservation during scattering ; the low energy limit implies that for each momenta there exist the ultraviolet cut - off @xmath225 .
\3 . interchain exchange scattering ( @xmath4 ) : @xmath226 \4 .
interchain exchange scattering ( @xmath3 ) : @xmath227 where @xmath228 is the @xmath3 charge density in the _ i - th _ chain .
usually one is also interested in correlation functions , which have the general form : @xmath229 where the simplest intra - chain examples of possible operators @xmath14 are : * in peierls channel - the charge density wave ( cdw ) + @xmath230 * in cooper channel - the singles superconductivity ( ss ) + @xmath231 when we consider a single chain then a chiral fermion creation operator is related to bosonic spin and charge fields as follows ( in the continuum limit ) :
@xmath232 where the coefficient @xmath233 is the majorana fermion which does nt have any spatial dependance and it is introduced only in order to preserve anti - commutation for the fermion operators @xmath234 . usually they do nt play any role in the physical description of the system but they are able to change signs of some correlation functions , so one has to take care about them . in the case of the ladder one has a straightforward generalization : @xmath235\ ] ] the above given equations eq.[eq : ferm - to - boson ] and eq.[eq : ferm - to - bos - ladder ] allow to rewrite all fermionic terms in the hamiltonian like _
e.g. _ the one in eq.[eq : umklapp - ferm ] or in general any interesting operator into the language of bosonic fields . in our particular case
we are interested in the following interaction terms : \1 .
umklapp scattering ( at quarter filling ) in the _ i - th _
chain @xmath236 where @xmath126 indicates the doping away from a commensurate case , quarter filling in our case ( where @xmath237 is not oscillating ) .
interchain umklapp scattering ( @xmath4 ) : @xmath238 \3 .
interchain exchange scattering ( @xmath4 ) : @xmath239 \4 .
interchain exchange scattering ( @xmath3 ) : @xmath240 these are the terms for which rg equations are derived in sec.[sec : rg ] . in all the above we took a following convention : @xmath241 which allows to go from one basis to another .
estimating the value of @xmath153 at certain energy scale @xmath242 is an impossible task , so all our results can not be taken strictly quantitatively .
this is a generic problem of the rg method present for any model even a simple half filled chain .
the complexity of our system , makes the task even more tedious .
we propose the approach , based on the first order expansion of rg equations , to get a reasonable value of @xmath243 .
however one has to keep in mind that as we are are working with rather large couplings , of order 0.1 ( but always smaller than 0.25 ) so higher order terms can introduce non - negligible corrections even within single instability flow . thus the results has to be taken with caution , can be thought as only approximation . with this
remarks being said we can proceed .
one can in principle integrate out bkt equations to get a values of along the flow .
one gets the following result@xcite : @xmath244= \frac{a}{\sinh\left [ al+\tanh^{-1}(a / y_{\parallel}^0)\right]}\ ] ] @xmath245= \frac{a}{\tanh\left [ al+\tanh^{-1}(a / y_{\parallel}^0)\right]}\ ] ] where a is an invariant of the flow , in our case @xmath246 .
every point on an rg trajectory leads us to fixed point value @xmath247 . from numerics based on u - v model we found ( sec.[sec : ll - res ] ) @xmath248
and we know that for quarter - filled chain @xmath249 . thus in our case @xmath250 . our aim is to extract the value of @xmath251=k_{rho}^c+y_{\parallel}[l]$ ] at certain energy scale @xmath252 ( corresponding to @xmath253 ) which upon rg leads to known @xmath247 .
this procedure can be thought as moving against the direction of the rg flow .
the only missing quantity in eq.[eq : rg - itegr1 ] is @xmath254 $ ] which should be interpreted as the distance of an initial point of the flow @xmath255 $ ] ( formally in the infinite energy , i.e. somewhere around uv - cut - off of the model ) from @xmath256 .
although around the fixed point the flow is rather vertical , we decided to take the most conservative ( and giving the most modest value of @xmath243 ) assumption that @xmath255=1/2 \rightarrow y_{\parallel}[l=0]=1/4 $ ] .
it is because certainly at @xmath257 the quarter filled chain remains insensitive to mott localization . with this
we are able to estimate the bare @xmath258 to be around @xmath259 .
then @xmath260 is only 6% larger than corresponding @xmath261 , while the position of the separatrix of rg flow in the first order is set by @xmath262 .
the distance from the separatrix is of the same order like additional intra - chain terms described in sec.[ssec : intra - rest ] .
from the remarks above we know this is only an estimate , but it certainly implies that the umklapp , even in intra - chain u - v approximation , is an order of magnitude larger than any of @xmath263 .
although it seems to be irrelevant , but its bare amplitude is large and quite close to separatrix so the intra - chain umklapp has to be taken with care in the rg analysis below .
the fact that bare @xmath154 falls very close to separatrix can be confirmed by an independent reasoning .
the position of separatrix is frequently constrained by symmetry considerations .
taking into account the hierarchy of energy scales , we take simplistic approximation of spinless fermions in @xmath264 ( like in sec.[sec : ll - res ] ) .
this obviously overestimates @xmath153 , but numerics @xcite shows us that beyond a threshold @xmath265 , the dependence on u is weak , thus the corrections breaking ( particle - hole ) symmetry should be quite small .
the advantage is that now one can map charge sector on pseudo - spin model ( empty and occupied sites ) , for which the position of separatrix of rg flow ( phase transition ) in terms of bare parameters is known exactly thanks to su(2 ) symmetry .
it is located at @xmath266 where @xmath267 accounts for all unscreened interactions along the chain ( while @xmath268 , if one want to be more precise he would found it slightly reduced with respect to @xmath49 because of dimerization ) .
if we put values found before ( in table tab.[tab : str - coupl - m ] for @xmath267 ) we indeed find ourself very close to @xmath269 .
j. d. denlinger , g .- h .
gweon , j. w. allen , c. g. olson , j. marcus , c. schlenker , and l .- s .
non - fermi - liquid single particle line shape of the quasi - one - dimensional non - cdw metal @xmath270 : comparison to the luttinger liquid .
, 82:25402543 , mar 1999 .
xiaofeng xu , a. f. bangura , j. g. analytis , j. d. fletcher , m. m. j. french , n. shannon , j. he , s. zhang , d. mandrus , r. jin , and n. e. hussey .
directional field - induced metallization of quasi - one - dimensional @xmath272 . , 102:206602 ,
may 2009 .
nicholas wakeham , alimamy f. bangura , xiaofeng xu , jean - francois mercure , martha greenblatt , and nigel e. hussey . gross violation of the wiedemann - franz law in a quasi - one - dimensional conductor .
, 2:396 , 2011 .
j. choi , j. l. musfeldt , j. he , r. jin , j. r. thompson , d. mandrus , x. n. lin , v. a. bondarenko , and j. w. brill .
probing localization effects in @xmath271 purple bronze : an optical - properties investigation .
, 69:085120 , feb 2004 .
m. s. da luz , j. j. neumeier , c. a. m. dos santos , b. d. white , h. j. izario filho , j. b. leo , and q. huang .
neutron diffraction study of quasi - one - dimensional lithium purple bronze : possible mechanism for dimensional crossover .
, 84:014108 , jul 2011 .
gweon , j. w. allen , and j. d. denlinger .
generalized spectral signatures of electron fractionalization in quasi - one- and two - dimensional molybdenum bronzes and superconducting cuprates . , 68:195117 , nov 2003 .
feng wang , s .- k .
mo , j. w. allen , h .- d .
kim , j. he , r. jin , d. mandrus , a. sekiyama , m. tsunekawa , and s. suga .
case for bulk nature of spectroscopic luttinger liquid signatures observed in angle - resolved photoemission spectra of @xmath271 . , 74:113107 , sep 2006 .
gweon , s .- k .
mo , j. w. allen , j. he , r. jin , d. mandrus , and h. hchst .
luttinger liquid angle - resolved photoemission line shapes from samples of @xmath271 grown by the temperature - gradient - flux technique .
, 70:153103 , oct 2004 . c. a. m. dos santos , m. s. da luz , yi - kuo yu , j. j. neumeier , j. moreno , and b. d. white
. electrical transport in single - crystalline @xmath273 : a two - band luttinger liquid exhibiting bose metal behavior .
, 77:193106 , may 2008 .
feng wang , j. v. alvarez , j. w. allen , s .- k .
mo , j. he , r. jin , d. mandrus , and h. hchst .
quantum critical scaling in the single - particle spectrum of a novel anisotropic metal .
, 103:136401 , sep 2009 . | molybdenum purple bronze li@xmath0mo@xmath1o@xmath2 is an exceptional material known to exhibit one dimensional ( 1d ) properties for energies down to a few mev .
this fact seems to be well established both in experiments and in band structure theory .
we use the unusual , very 1-dimensional band dispersion obtained in _ ab - initio _ dft - lmto band calculations as our starting point to study the physics emerging below 300mev . a dispersion perpendicular to the main dispersive direction is obtained and investigated in detail . based on this , we derive an effective low energy theory within the tomonaga luttinger liquid ( tll ) framework .
we estimate the strength of the possible interactions and from this deduce the values of the tll parameters for charge modes .
finally we investigate possible instabilities of tll by deriving renormalization group ( rg ) equations which allow us to predict the size of potential gaps in the spectrum . while @xmath3 instabilities strongly suppress each other
, the @xmath4 instabilities cooperate , which paves the way for a possible cdw at the lowest energies .
the aim of this work is to understand the experimental findings , in particular the ones which are certainly lying within the 1d regime .
we discuss the validity of our 1d approach and further perspectives for the lower energy phases . |
hard x - ray surveys of the galaxy provide an opportunity to discover populations of extreme sources .
the promise of such surveys has been partially realized with the _ international gamma - ray astrophysics laboratory _ ( _ integral _ , * ? ? ?
* ) , which carried out a 20100kev survey of the entire galactic plane and has discovered hundreds of new sources @xcite , including new types of high mass x - ray binaries ( hmxbs ) , pulsar wind nebulae ( pwne ) , and magnetic cataclysmic variables ( cvs ) .
the _ nuclear spectroscopic telescope array _ ( _ nustar _ , * ? ? ? * ) , which launched in 2012 june and covers the 379kev bandpass , is the first focusing hard x - ray telescope in orbit .
while it has a much smaller field of view than _ integral _ , it has much lower background and greatly improved sensitivity .
thus , one of _ nustar _
s science goals is to extend our view deeper into the galactic plane to look for hidden hard x - ray populations .
during its first year of operation , _ nustar _ has initiated surveys of @xmath11 deg@xmath13 areas both in the galactic center ( e.g. , * ? ? ? * ; * ? ? ? * ) and in a region which samples the spiral arm population in order to probe potentially different environments where x - ray binaries are found .
a spiral arm region centered on galactic coordinates of @xmath14 and @xmath15 was chosen for having the highest known density of ob star associations @xcite and hmxbs @xcite .
this is part of the `` norma '' spiral arm region , which was identified early in the _ integral _ mission as having an unusually high density of hard x - ray sources @xcite .
the combination of active star formation and evidence that compact objects have already formed suggests that a survey by _ nustar _ may uncover compact objects associated with populations of massive stars such as magnetars or faint hmxbs that are early in their evolutionary process and may have neutron star or black hole accretors .
the full surveys of the galactic center and the norma region will be carried out over a period of @xmath12 years ; here we report on the discovery of a transient source made during the first part of the survey . in the following , 2 describes observations made with _
nustar _ , _ swift _ , and _ chandra _ , as well as the procedures we used to reduce the data .
the results are presented in 3 , and we discuss possibilities for the nature of the transient in 4 .
as the first part of the _ nustar _ survey of the norma region , nine @xmath120ks _ nustar _ observations were performed between ut 2013 february 20 and february 24 .
each @xmath16-by-@xmath16 field of view was partially overlapping with adjacent pointings , and the entire region covered was @xmath10.2 deg@xmath13 .
the results from all nine pointings will be reported in bodaghee et al .
( submitted ) . here
we focus on the observations that covered a new transient , nustar j1634334738.7 .
these observations are listed in table [ tab : obs ] , including two that were obtained during the survey and a follow - up _ nustar _
observation on 2013 march 23 that was coordinated with _
chandra_. in addition , @xmath12ks _ swift _ x
- ray telescope ( xrt ) observations of the region were carried out during 2013 february 21 - 24 , and four of these observations covered nustar j1634334738.7 .
table [ tab : obs ] lists these along with three other _ swift _ observations that were acquired after our survey as part of another observing program .
we also analyzed archival data covering the source , including two _ chandra _ observations ( obsids 12529 and 12532 with exposure times of 19.0ks and 19.5ks , respectively ) that were acquired in 2011 as part of a survey of the same region being covered by _
( fornasini et al . ,
submitted ) .
we reduced the _ nustar _ and _ swift _ data using heasoft v6.14 and the latest version of the calibration database ( caldb ) files as of 2013 august 30 .
we produced cleaned event lists for the _ nustar _ focal plane modules ( fpma and fpmb ) using nupipeline and for the _ swift_/xrt using xrtpipeline , and further analysis of the event lists is described below . for _ chandra _
, we processed the advanced ccd imaging spectrometer ( acis , * ? ? ? * ) data with the _ chandra _ interactive analysis of observations ( ciao ) software , using chandra_repro to make event lists .
we obtained near - ir observations covering the nustar j1634334738.7 error region .
this includes @xmath17 , @xmath18 and @xmath19 observations performed on 2011 july 19 with ctio / newfirm in the framework of near - ir mapping of the _ chandra _ survey field . a detailed description of the data and their reduction can be found in rahoui et al .
( in prep ) .
they were reduced with the dedicated iraf package nfextern following the standard procedure tailored for wide - field mosaics which consists of bad pixel removal , dark subtraction , linearity correction , flatfielding and median sky subtraction .
the resulting images were then flux - calibrated through relative photometry with the 2mass catalogue .
we also obtained @xmath19-band imaging with the ohio - state infra - red imager / spectrometer ( osiris ) at the southern astrophysical research ( soar ) 4.1 m telescope .
we used the f/7 camera , providing a @xmath20 field of view , centered at the nominal coordinates of nustar j163433 - 4738.7 .
we obtained 36@xmath2160s dithered exposures of the field on 2013 april 3 under good conditions with seeing of @xmath22 , and observed it for 9@xmath2160s again on 2013 april 5 , also under good conditions but with somewhat worse seeing of @xmath23 .
reductions were done using the xdimsum iraf package and photometric calibration was obtained by comparing to 2mass all - sky point source catalog sources in the field .
the new transient was discovered from an inspection of the image from the 22.6ks _ nustar _ observation that took place starting on 2013 february 23 , 14.52 h. as shown in figure [ fig : nuimages ] , the source was detected in fpma and fpmb . to determine the significance of the detection
, we extracted 310kev counts from a @xmath24-radius circle centered on the approximate position of the source .
we determined the background level using a nearby source - free circular region with a radius of @xmath25 .
after background subtraction , we obtained @xmath26 and @xmath27 counts in fpma and fpmb , respectively .
the combined significance is 8.0-@xmath0 , confirming the detection . and smoothed with a 6-pixel gaussian .
the scale on the bottom of the figure is in counts per pixel , and a logarithmic scaling is used .
the apparent elongation of the source is due to the distorted psf shape at large off - axis angles .
[ fig : nuimages ] ] to determine the source position , we extracted all the events from a @xmath28-by-@xmath28 square region and made histograms of the counts , binning in the r.a . and
decl . directions .
we performed @xmath5 fitting of the histograms with a model consisting of a constant ( accounting for the flat background ) and a gaussian for the source .
it should be noted that this is an approximation since the _ nustar _ point spread function ( psf ) is non - gaussian .
we determined the centroids separately for fpma and fpmb , and they are consistent with each other .
the weighted average of the two centroids is r.a.=@xmath29 , decl.=@xmath30 ( j2000.0 ) with 3-@xmath0 statistical uncertainties of @xmath31 and @xmath32 in r.a . and decl .
, respectively .
after considering that the systematic pointing uncertainty for _ nustar _ is @xmath1@xmath33 @xcite , the error region can be approximated with a @xmath34-radius circle .
we searched on - line catalogs ( e.g. , simbad ) for x - ray sources consistent with the position of nustar j1634334738.7 , but we did not find any likely candidates .
the catalog from the 2011 _ chandra _ survey ( fornasini et al . , submitted ) does not have any sources in the nustar j1634334738.7 error region . based on this and the analysis described below ( including a re - analysis of the 2011 _ chandra _ observations ) ,
we conclude that nustar j1634334738.7 is a previously undetected source and that it is very likely to be a transient given the sensitivity of the _ chandra _ observations
. confidence .
the time of a _ swift _ observation is indicated .
zero on the time axis corresponds to mjd 56,346.60000 .
[ fig : lc ] ] for the _ nustar _ and _ swift _ observations listed in table [ tab : obs ] , we estimated count rates or upper limits for nustar j1634334738.7 using @xmath24-radius source regions centered at the source position derived from obsid 40014007001 .
the xrt 90% encircled energy radius is approximately @xmath35 , but we use @xmath24 to account for the uncertainty in the source position as nustar j1634334738.7 is not bright enough in any of the _ swift _ observations to improve the measurement of the source position . as described above , we used a larger , circular region that does not contain any detected point sources for background . selecting the background regions for _ nustar
_ requires some care because of scattered light from a nearby bright source ( 4u 163047 ) .
the 310kev count rates or limits obtained for all observations are given in table [ tab : obs ] .
for _ nustar _ , the background rates are high enough to use gaussian statistics , and we use the poisson limits tabulated in @xcite for _ swift _ and _ chandra_. in all cases , we quote the 1-@xmath0 error bars if the minimum of the 1-@xmath0 error region is positive . otherwise , we give the 90% confidence upper limit . while the _ nustar _ observation taken on february 23 provides the only highly significant detection , the _ swift _ observation with the highest count rate is the one that occurred during this _ nustar _ observation . the only other possible evidence for activity from nustar j1634334738.7 occurred on february 22 , when _ nustar _ obtained a 2.7-@xmath0 detection in fpma ; however , this is not confirmed by the fpmb data . for _ chandra _ obsids 12529 and 12532 ( from 2011 ) and obsid 15625 ( from 2013 ) , we analyzed the acis data to search for a detection of nustar j1634334738.7 .
based on an inspection , no sources are apparent in the 0.310kev or 310kev images .
the source is @xmath36 and @xmath37 from the _ chandra _ aimpoint for the 2011 obsids . at these off - axis angles , the 90% encircled energy fraction ( eef ) radii ( for 4.5kev photons )
are @xmath38 and @xmath39 for obsids 12529 and 12532 , respectively .
obsid 15625 was a dedicated pointing with the target on - axis , and the 90% eef radius is @xmath40 . for obsid 12529 , the largest number of 310kev counts within any @xmath38-radius circle inside the _ nustar _ error region is three , and , after accounting for background , we calculate a 90% confidence upper limit of @xmath41@xmath42s@xmath12 on the count rate ( see table [ tab : obs ] ) . for obsid 12532 , the largest number of 310kev counts within any @xmath39-radius circle inside the _ nustar _ error region is two , and the count rate limit is @xmath41@xmath43s@xmath12 . for obsid 15625 ,
the largest number of 310kev counts within a @xmath40-radius circle is two , and the count rate limit is @xmath41@xmath44s@xmath12 . this is higher than for the 2011 obsids because of the lower exposure time .
we made _ nustar _ light curves for obsid 40014007001 with several different time binnings between 0.1s and 5500s ( the approximate satellite orbital period ) in the 310kev and 379kev bandpasses .
the lack of any apparent variability in the 0.110s light curves rules out flares or bursts that might prove the presence of a neutron star .
figure [ fig : lc ] shows the 310kev orbit - by - orbit light curve for fpma and fpmb combined . at the 5500s binning , a @xmath5 test shows that the 310kev and 379kev light curves are consistent with the source being constant over @xmath140ks .
next , we extracted _ nustar _ fpma and fpmb spectra for obsid 40014007001 and an xrt spectrum for obsid 00080508001 , which are the two observations with significant detections of nustar j1634334738.7 .
these were fitted jointly by minimizing the cash ( or @xmath45 ) statistic @xcite using the xspec software package .
the statistical quality of the spectrum is low , and it is well fit by a power - law , a blackbody , or a thermal bremsstrahlung model . in all three cases , we included absorption using @xcite abundances and @xcite cross - sections .
the parameters are given in table [ tab : spectra ] , and the errors quoted are 90% confidence for one parameter of interest , @xmath46@xmath45 = 2.7 .
although we used the @xmath45-statistic to determine the parameters , we also calculated the reduced-@xmath5 values , and they appear in table [ tab : spectra ] .
this quantity is slightly smaller for the power - law model ( @xmath47 for 8 degrees of freedom ) compared to the blackbody model ( @xmath48 for 8 dof ) and the bremsstrahlung model ( @xmath49 for 8 dof ) .
however , given the small number of dof , the difference in @xmath50 is not significant , and the steeply falling power - law index ( @xmath2 ) suggests that the emission probably has a thermal origin . the spectrum for the blackbody model is shown in figure [ fig : spectrum ] .
we used the absorbed blackbody model and the count rates for _ nustar _ , _ swift _ , and _
chandra _ given in table [ tab : obs ] to calculate measurements or upper limits on the absorbed 310kev flux , and these flux histories are shown in figure [ fig : lc_overview ] .
we also calculated the flux upper limits for the _ chandra _ observations from 2011 .
the upper limits on the absorbed 310kev fluxes are @xmath41@xmath51erg@xmath7s@xmath12 and @xmath41@xmath52erg@xmath7s@xmath12 for obsids 12529 and 12532 , respectively .
although the count rate limits are considerably lower for these obsids compared to _
obsid 15625 , the flux limits are similar because of the different effective areas for the acis - i and acis - s instruments .
the lowest _ chandra _ upper limit indicates that the flux from this source changed by at least a factor of @xmath53 , suggesting that it is a transient rather than simply being a highly variable x - ray source . confidence .
[ fig : spectrum ] ] the _ nustar _ error circle includes dozens of near - ir candidate counterparts brighter than @xmath54 ( vega magnitude system ) . in the newfirm images from 2011 , the brightest source in the error circle ( 2mass j163432884738393 ) has @xmath55 , @xmath56 , and @xmath57 , and the 2mass magnitudes are consistent , indicating a possible lack of variability on long time scales .
the next two brightest sources are 2mass j163433624738479 at @xmath58 and the _ spitzer_/glimpse source g336.7870 + 00.0111 at @xmath59 .
these may be more likely counterparts to nustar j1634334738.7 because they are relatively highly reddened ( @xmath60 and @xmath61 , respectively ) .
however , we do not find any evidence that these sources are variable .
for the two osiris @xmath19 images taken on april 3 and april 5 , we performed aperture photometry for all the sources within the _ nustar _ error circle , but did not find any variable sources .
the high time - resolution of the _ nustar _ data allows for a search for a coherent signal with periods @xmath62ms , covering the range expected for either an isolated rotation - powered pulsar , a binary , or a magnetar . given the paucity of source counts in the observations listed in table 1 , we concentrate our attention on the obsid 40014007001 data .
photon arrival times , adjusted for the _ nustar _ clock drift , were corrected to the solar system barycenter using jpl de200 ephemeris and the _ nustar _ derived source coordinates .
we extracted photons in the 310kev band from a 30@xmath63-radius aperture centered on the source to optimize the signal - to - noise ratio .
we searched for significant power from a coherent signal using an fft sampled at the nyquist frequency .
the observation span was too short to consider the smearing of the pulse profile by spin - down of even the most highly energetic pulsar or the typical binary orbit period .
the most significant signal found has a power of @xmath64 , corresponding to a probability of false detection of @xmath65 for @xmath66 fft search trials .
we conclude that no pulsed x - ray signal is detected in from nustar j1634334738.7 . after taking into account the local background , we place an upper limit on the pulse fraction at the 3-@xmath0 confidence level of @xmath67% for a blind search for a sinusoidal signal @xmath68ms .
@xmath7 and a temperature of 1.2kev .
the errors on the data points are 1-@xmath0 significance , and the upper limits are 90% confidence . [
fig : lc_overview ] ]
in considering the nature of nustar j1634334738.7 , it is useful to estimate the x - ray luminosity that the source reached during its outburst .
although the distance is highly uncertain , the strong absorption either indicates a large distance through a region of the galaxy with heavy extinction or absorption local to the source .
the former is a strong possibility since the transient is in the direction of a region of the galactic plane that is crowded with h ii / molecular cloud regions .
nustar j1634334738.7 is at @xmath69 , near a group of h ii regions ( 336.732 + 0.072 , 336.840 + 0.047 , 336.90.1 , 337.1470.181 , and 337.30.1 ) at a distance of @xmath70kpc @xcite .
if the absorption is interstellar , then nustar j1634334738.7 is very likely beyond or in these molecular clouds , indicating a 210kev unabsorbed luminosity limit of @xmath10@xmath71erg s@xmath12 for the blackbody spectrum and @xmath10@xmath72erg s@xmath12 for bremsstrahlung . in the following ,
we consider both possibilities : a distance of @xmath111kpc and a peak luminosity of @xmath1@xmath11erg s@xmath12 ; and a smaller distance with absorption local to the source and a lower luminosity . considering the first possibility , a luminosity as high as @xmath11erg s@xmath12 would be unprecedented for at least two common types of x - ray sources in the galaxy .
non - magnetic cvs have strong optical outbursts along with persistent and transient x - ray emission .
however , extensive studies have shown that their x - ray luminosities are in the @xmath73erg s@xmath12 range @xcite .
while magnetic cvs can reach higher luminosities @xcite , they do not typically show outbursts .
secondly , active binaries , including rs cvn systems and low - mass flare stars , produce x - ray flares , some of which can last for about a day .
superflares that peak at x - ray luminosities of @xmath74erg s@xmath12 have been seen @xcite , and there have been flares that have released @xmath75@xmath76erg @xcite . if nustar j1634334738.7 is at @xmath111kpc , then in addition to having a higher peak luminosity , the energy released is @xmath75@xmath77erg , which is based on the source being at its peak luminosity for @xmath7540ks ( see figure [ fig : lc ] ) . in summary , while we do not rule out the active binary possibility ( and see below for further discussion on this topic ) , at the 11kpc distance , the event seen by _
nustar _ would need to be extreme for this explanation to be correct .
a source type that might be a good match to the nustar j1634334738.7 properties is the class of highly magnetic isolated neutron stars : magnetars .
while these sources are best known for their very bright and brief ( @xmath10.1s ) x - ray and gamma - ray flares , they also have persistent but variable emission that can easily reach @xmath11erg s@xmath12 or higher @xcite .
the x - ray spectrum of their persistent emission is dominated by a @xmath10.51kev blackbody @xcite .
furthermore , as mentioned in 1 , magnetars are associated with regions where high - mass star formation is occurring , such as the norma region , and there is a known magnetar , sgr 162741 , that is @xmath10.2@xmath78 from nustar j1634334738.7 .
however , one possible counter - argument to the magnetar hypothesis is that magnetar periods of activity usually last for months . from figure
[ fig : lc_overview ] , the peak activity from nustar j1634334738.7 could not have lasted for more than @xmath11.5days based on the _ swift _ upper limits , and _
chandra _ and _ nustar _ place a tight upper limit on activity @xmath13 weeks after the _ nustar _ detection . while we do not detect pulsations or @xmath10.1s flares , which would prove that the source is a magnetar , the pulsation search is limited by the statistical quality of the data , and flaring episodes are relatively rare .
a blackbody spectrum could also be produced from an optically thick accretion disk in a black hole binary .
most often , such sources show temperatures of @xmath11kev in the inner parts of their accretion disks when they are at high luminosity @xmath75@xmath79erg s@xmath12 , which would require a very large distance of @xmath75100300kpc for nustar j1634334738.7 .
two sources ( grs 1758258 and 1e 1740.72942 ) have shown fainter soft states , but their blackbody spectra have been at temperatures of 0.4kev @xcite and 0.7kev @xcite , which are lower than the 1.2kev we see for nustar j1634334738.7 . also , transient black hole binaries typically have outbursts that last for several months , which is very different from nustar j1634334738.7 . considering the second possibility that the source is closer than @xmath111kpc , then the high column density must be due to material local to the source . in fact ,
_ integral _ has found a large number of obscured hmxbs that have column densities of @xmath80@xmath7 due to the wind from their supergiant companions ( e.g. * ? ? ?
* ) . while many of the sources in this class are persistent , the supergiant fast x - ray transients ( sfxts ) , which have some members that are also obscured hmxbs , have outbursts that can be as short as a few hours , and it would not be unusual for an sfxt to produce x - ray emission that lasts for half a day @xcite .
however , a possible problem for this interpretation is that the spectrum of nustar j1634334738.7 is softer than usually seen for active sfxts . during flares ,
igr j175442619 and igr j164794514 show a power - law spectrum with a photon index near @xmath81 @xcite .
somewhat softer spectra with @xmath822.4 can be seen during periods of weaker activity , but their spectra are only as soft as nustar j1634334738.7 when they are in quiescence . for cvs and active binaries , the challenge is to explain a column density as high as we measure with absorption local to the systems .
one possibility is that material expelled from the system ( either related to the evolutionary state of the stellar component or related to the x - ray flaring event ) could cause increased column density .
however , for one active binary ( ax ari ) , the column density was seen to increase to @xmath83@xmath7 during an event @xcite , which is 23 orders of magnitude less than is required for nustar j1634334738.7 .
variation in optical brightness of some stars has also been attributed to ejection of material from the stars .
an example of this is a change of 0.5 magnitudes in the optical seen for ku cyg @xcite .
however , this would translate into a column density near @xmath84@xmath7 , which is still far less than is required .
the near - ir information that we have provides only weak constraints on the source type .
given that we were not able to identify the counterpart , we can only say that the source must be fainter ( before and after the x - ray flare ) than the brightest source in the error circle ( 2mass j163432884738393 ) .
this corresponds to @xmath85 , @xmath86 , and @xmath87 , and these values are consistent with all the possibilities for the source type discussed above .
however , in the scenario where nustar j1634334738.7 is a nearby source with local absorption , these limits do provide some constraint .
for example , a very nearby and luminous obscured hmxb is ruled out . in summary ,
nustar j1634334738.7 is a fast x - ray transient , which has a thermal spectrum with relatively high temperature ( @xmath88kev for a blackbody or 3.0kev for bremsstrahlung ) .
if its high column density is due to interstellar material , the source is probably distant ( @xmath7511kpc ) , making the peak luminosity @xmath75@xmath11ergs@xmath12 .
we discuss the origin of the flare and suggest that the most likely possibilities if the source is distant are an unusually bright flare from an active binary or a short outburst from a magnetar .
we also consider the possibility that the source is closer and that the absorption is local to the source .
observations in the galactic plane will determine whether such transients are common and , hopefully , shed light on the nature of nustar j1634334738.7 .
this work was supported under nasa contract no .
nng08fd60c , and made use of data from the _ nustar _ mission , a project led by the california institute of technology , managed by the jet propulsion laboratory , and funded by the national aeronautics and space administration .
we thank the _ nustar _ operations , software and calibration teams for support with the execution and analysis of these observations .
this research has made use of the _ nustar _ data analysis software ( nustardas ) jointly developed by the asi science data center ( asdc , italy ) and the california institute of technology ( usa ) .
rja was supported by gemini - conicyt grant 32120009 .
feb was supported by basal - cata pfb-06/2007 and conicyt - chile ( through fondecyt 1101024 , gemini - conicyt 32120003 , and anillo act1101 ) .
ln wishes to acknowledge the italian space agency ( asi ) for financial support by asi / inaf grant i/037/12/0 - 011/13 .
we thank harvey tananbaum for providing _
director s discretionary time for this project .
this research has made use of the simbad database , operated at cds , strasbourg , france .
, g. p. , bautz , m. w. , ford , p. g. , nousek , j. a. , & ricker , g. r. , 2003 , in x - ray and gamma - ray telescopes and instruments for astronomy .
edited by joachim e. truemper , harvey d. tananbaum .
proceedings of the spie , 4851 , 28 ccccccc mission & instrument & obsid & start time ( ut ) & end time ( ut ) & exposure ( ks ) & count rate and the upper limits are 90% confidence . ]
+ + & acis - i & 12529 & jun 16 , 6.97 h & jun 16 , 12.53 h & 19.0 & @xmath410.19 + _ chandra _ & acis - i & 12532 & jun 16 , 23.85 h & jun 17 , 5.55
h & 19.5 & @xmath410.21 + + & xrt & 00080509001 & feb 21 , 21.70 h & feb 21 , 23.13 h & 1.79 & @xmath89 + _ nustar _ & fpma & 40014004001 & feb 22 , 7.77 h & feb 22 , 17.52 h & 19.4 & @xmath90 + & fpmb & & & & & @xmath411.3 + _ swift _ & xrt & 00080505001 & feb 22 , 11.88 h & feb 22 , 13.71 h & 1.92 & @xmath410.54 + _ swift _ & xrt & 00080506001 & feb 23 , 0.83 h & feb 23 , 2.62 h & 1.98 & @xmath411.9 + _ nustar _ & fpma & 40014007001 & feb 23 , 14.52 h & feb 24 , 1.77 h & 22.6 & @xmath91 + & fpmb & & & & & @xmath92 + _ swift _ & xrt & 00080508001 & feb 23 , 20.20 h & feb 23 , 21.70 h & 1.97 & @xmath93 + _ swift _ & xrt & 00080511001 & feb 24 , 13.72 h & feb 24 , 15.18 h & 1.75 & @xmath410.82 + _ swift _ & xrt & 00032728001 & feb 28 , 1.06 h & feb 28 , 6.24 h & 4.95 & @xmath410.45 + _ swift _ & xrt & 00032728002 & mar 3 , 9.00 h & mar 3 , 19.06 h & 4.85 & @xmath410.12 + _ swift _ & xrt & 00032728003 & mar 5 , 10.66 h & mar 5 , 20.75 h & 4.47 & @xmath411.1 + _ chandra _ & acis - s & 15625 & mar 23 , 8.30 h & mar 23 , 11.91 h & 9.84 & @xmath410.50 + _ nustar _ & fpma & 30001012002 & mar 23 , 8.52 h & mar 23 , 18.35 h & 16.6 & @xmath410.50 + & fpmb & & & & & @xmath412.0 + | during hard x - ray observations of the norma spiral arm region by the _ nuclear spectroscopic telescope array ( nustar ) _ in 2013 february , a new transient source , nustar j1634334738.7 , was detected at a significance level of 8-@xmath0 in the 310kev bandpass .
the source is consistent with having a constant _ nustar _ count rate over a period of 40ks and is also detected simultaneously by _ swift _ at lower significance .
the source is not significantly detected by _
nustar _ , _ swift _ , or _
chandra _ in the days before or weeks after the discovery of the transient , indicating that the strong x - ray activity lasted for between @xmath10.5 and 1.5 days .
near - ir imaging observations were carried out before and after the x - ray activity , but we are not able to identify the counterpart . the combined _ nustar _ and _ swift _ energy spectrum is consistent with a power - law with a photon index of @xmath2 ( 90% confidence errors ) , a blackbody with @xmath3kev , or a bremsstrahlung model with @xmath4kev .
the reduced-@xmath5 values for the three models are not significantly different , ranging from 1.23 to 1.44 for 8 degrees of freedom .
the spectrum is strongly absorbed with @xmath6@xmath7 , @xmath8@xmath7 , and @xmath9@xmath7 , for the power - law , blackbody , and bremsstrahlung models , respectively .
although the high column density could be due to material local to the source , it is consistent with absorption from interstellar material along the line of sight at a distance of 11kpc , which would indicate an x - ray luminosity @xmath10@xmath11erg s@xmath12 .
although we do not reach a definitive determination of the nature of nustar j1634334738.7 , we suggest that it may be an unusually bright active binary or a magnetar . |
the field of metamaterials @xcite is driven by fascinating and far - reaching theoretical visions such as , e.g. , perfect lenses @xcite , invisibility cloaking @xcite , and enhanced optical nonlinearities @xcite .
this emerging field has seen spectacular experimental progress in recent years @xcite . yet , losses are orders of magnitude too large for the envisioned applications . achieving such reduction by further design optimization appears to be out of reach .
thus , incorporation of active media ( gain ) might reduce the losses .
the procedure would be to simply inject an electrical current into the active medium , leading to gain and hence to compensation of the losses .
however , experiments on such intricate active nanostructures do need guidance by theory via self - consistent calculations ( using the semi - classical theory of lasing ) for realistic gain materials that can be incorporated into or close to dispersive media to reduce the losses at thz or optical frequencies .
the need for _ self - consistent _ calculations stems from the fact that increasing the gain in the metamaterial , the metamaterial properties change , in turn changes the coupling to the gain medium until a steady - state is reached .
a specific geometry to overcome the severe loss problem of optical metamaterials and to enable bulk metamaterials with negative magnetic and electric response and controllable dispersion at optical frequencies is to interleave active , optically pumped gain material layers with the passive metamaterial lattice . for reference , the best fabricated negative - index material operating at around @xmath0 wavelength @xcite has shown a figure of merit , @xmath1 , where @xmath2 is the effective refractive index .
this experimental result is equivalent to an absolute absorption coefficient of @xmath3 , which is even larger than the absorption of typical direct - gap semiconductors such as , e.g. , gaas ( where @xmath4 ) .
so it looks difficult to compensate the losses with this simple type of analysis , which assumes that the bulk gain coefficient is needed .
however , the effective gain coefficient , derived from self - consistent microscopic calculations , is a more appropriate measure of the combined system of metamaterial and gain . due to pronounced local - field enhancement effects in the spatial vicinity of the dispersive metamaterial
, the effective gain coefficient can be substantially larger than its bulk counterpart .
while early models @xcite using simplified gain - mechanisms such as explicitly forcing negative imaginary parts of the local gain material s response function produce unrealistic strictly linear gain , our self - consistent approach presented below allows for determining the range of parameters for which one can realistically expect linear amplification and linear loss compensation in the metamaterial @xcite . to fully understand the coupled metamaterial - gain system , we have to deal with time - dependent wave equations in metamaterial systems by coupling maxwell s equations with the rate equations of electron populations describing a multi - level gain system in semi - classical theory @xcite .
this paper aims to apply a detailed computational model to the problem of metamaterials with gain .
in , we present the semi - classical theory of lasing and describe in detail the computational approach . in , we verify that our code agrees well with simple soluble models ( gain material only ) .
in addition , our code is applied to 1d superlattice of gain and negative index layers .
next , a 2d problem is considered , which is a square lattice of lorentz dielectric cylinders with layers of gain material .
finally , a 2d split ring resonator ( srr ) with gain material inclusions is considered .
gain can compensate the losses and lasing ( spasing ) is achieved in our numerical simulations . in , we present our conclusions .
the gain atoms are embedded in host medium and described by a generic four - level system , as shown in .
all quantities including the fields and occupation numbers are tracked at each point in space and take into account energy exchange between gain atoms and fields , external pumping and non - radiative decays @xcite .
electrons are pumped by an external mechanism from the ground state level ( @xmath5 ) to the third level ( @xmath6 ) . after a short lifetime @xmath7
, they quickly relax into the metastable second level ( @xmath8 ) .
the second level ( @xmath8 ) and the first level ( @xmath9 ) are called as the upper and lower lasing levels , respectively .
electrons can transfer both radiatively ( spontaneous and stimulated emissions ) and non - radiatively from the upper to the lower lasing level . at last
, they transfer quickly and non - radiatively from the first level ( @xmath9 ) to the ground state level ( @xmath5 ) .
the energies of ground state and the third level are @xmath10 and @xmath11 . in optical pumping mechanism , electrons are raised from the ground state level ( @xmath5 ) to the third level ( @xmath6 ) by an external electromagnetic wave with the pumping frequency @xmath12 , which is chosen to be @xmath13 in our simulations .
the local intensity of the pumping em wave varies with the position and determines the pumping rate at each point .
the lifetimes and energies of the upper and lower lasing levels are @xmath14 and @xmath15 , respectively .
the center frequency of the radiation is @xmath16 , which is chosen to be @xmath17 .
the parameters @xmath7 , @xmath18 and @xmath19 are chosen @xmath20 , @xmath21 and @xmath20 , respectively . the total electron density , @xmath22 .
the time - dependent maxwell equations are given by @xmath23 where @xmath24 and @xmath25 is the electric polarization density of the gain material .
( @xmath26 is the induced electric polarization density on the atomic transition between the upper ( @xmath8 ) and lower ( @xmath9 ) lasing levels , and @xmath27 is between the ground state level ( @xmath5 ) and the third level ( @xmath6 ) . )
the induced electric polarizations behave as harmonic oscillators and couple to the local @xmath28 field , which is propagated by maxwell equations .
the polarization density @xmath29 obeys locally the following equation of motion @xcite @xmath30 where @xmath31 is the linewidth of the atomic transition @xmath32 , @xmath33 is the coupling strength of @xmath34 to the electric field , and @xmath35 and @xmath36 are the population inversions that drive the polarizations . in our simulations
, @xmath37 is chosen to be equal to @xmath38 or @xmath39 and @xmath40 is equal to @xmath41 .
the values for @xmath42 and @xmath43 are taken to be @xmath44 and @xmath45 , respectively . from , it can be easily derived @xcite that the atomic response of gain atoms has a lorentzian lineshape and is homogeneously broadened .
the occupation numbers at each spatial point vary according to the following rate equations , @xmath46 where @xmath47 ( @xmath48 ) is the induced radiation rate or excitation rate depending on its sign . instead of using an external em wave to optically pump electrons from the ground state level ( @xmath5 ) to the third level ( @xmath6 ) , we can simplify this process in and by pumping electrons with a homogeneous pumping rate @xmath49 , which is proportional to the optical pumping intensity in an experiment .
this simplification is valid only if the gain slab is thin and the gain of the laser is low , because the real pumping rate depends on the local optical intensity and should be a function of position . we ll discuss this in more detail in . based on this simplification , we can have the rate equations as follows , @xmath50 correspondingly , we only need to consider the electric polarization density @xmath51 on the atomic transition between @xmath8 and @xmath9 in and . in order to solve the behavior of the gain materials in the electromagnetic fields numerically , the finite - difference time - domain ( fdtd ) method
is utilized @xcite . at the left and right ends of the computational space , perfect matched layers ( pml )
are used to impose the absorbing boundary condition ( abc ) . in the fdtd calculations ,
both the space and time are discretized into small steps compared to the characteristic space and time periods . in our simulations presented below , the discrete time and space steps are chosen to be @xmath52 and @xmath53 in and , @xmath54 and @xmath55 in , and @xmath56 and @xmath57 in .
the initial condition is that all electrons are in the ground state and all electric , magnetic and polarization fields are zero .
then the electrons are pumped from @xmath5 to @xmath6 optically or with a homogeneous pumping rate @xmath49 .
the system begins to evolve according to the equations above .
to understand the lasing behavior of gain material , we first study a gain material slab surrounded by vacuum ( shown in ) .
we generate a continuous wave ( cw ) at the frequency @xmath58 ( @xmath59 ) and let it propagate through the gain slab , and then we calculate the reflected and transmitted waves and implement the fourier transforms to see if there is lasing and how much power is emitted around @xmath60 the emission frequency @xmath61 between @xmath9 and @xmath8 .
first we start with a very low input power @xmath62 for the incident cw wave , but no lasing happens , then we increase the input power till it reaches the lasing threshold , for which the system starts to have lasing and we can see a small peak at the emission frequency @xmath63 in the fourier transforms of reflected and transmitted waves , i.e. , low power emitted around the emission frequency @xmath61 . if we keep increasing the input power , the peak will get higher and the emitted power will get larger .
shows the transmitted waves and their corresponding fourier transforms for three different input powers at the gain slab @xmath64 .
we can see there is no lasing ( ) when the input power is low ( @xmath65 ) and there is only one peak for the pumping frequency in its fourier transform ( ) . when the input power @xmath66 , the system starts lasing ( ) and a small peak appears at the frequency @xmath60 ( ) . if we increase the input power to a higher value @xmath67 , the lasing gets stronger ( ) and the peak for the emission frequency gets higher ( ) , i.e. , more power emitted around the emission frequency @xmath61 .
we have calculated the emitted power at the emission frequency @xmath61 versus the input power at the pumping frequency @xmath58 for the same gain slab system . as shown in , we can see that there is a sharp rise in the emission around @xmath68 , which corresponds to the lasing threshold for this system . below the threshold , there is no lasing .
+ we also notice that the lasing time ( the time when the system starts lasing ) varies according to the input power .
shows the detailed results for the lasing time versus the input power with the slab width @xmath69 and @xmath70 .
we can see the lasing time decreases as the input power increases because the system pumps the electrons at a higher rate from the ground state level to higher levels and then reaches the population inversion between @xmath9 and @xmath8 in a shorter time . for a fixed input power
, one can see the lasing time decreases as the gain slab width gets larger .
this occurs because more input energy is absorbed and then converted to lasing by the wider gain slab . as the input wave propagates inside the gain slab
, it will decay due to the absorption from the gain material at the pumping frequency @xmath58 ( see ) .
thus the pumping rate , which is determined by the local input optical intensity , is inhomogeneous inside the gain slab .
but for a thin gain material layer , the electric field of the input wave can be approximately treated as homogeneous , thus we can simplify the pumping process between @xmath5 and @xmath6 by using a homogeneous pumping rate @xmath49 . for a very thin gain slab @xmath64 ,
simulations are done with a homogeneous pumping rate and the results for the power emitted around the emission frequency @xmath61 versus the pumping rate are plotted in . comparing with , where the electrons are optically pumped , we can see they are very similar . for a fixed output power such as @xmath71
, we can find the corresponding input power @xmath72 in and the corresponding pumping rate @xmath73 in .
then we do simulations for optical pumping case with the input power @xmath72 and for the homogeneous pumping rate case with the pumping rate @xmath73 .
the graphs of the occupation numbers as a function time are plotted in for both cases .
one can see that they are almost the same .
this verifies that the homogeneous pumping rate simplification is valid for a thin gain slab . in our following simulations
, we ll use this simplification because the gain slab widths in our structures are very thin ( @xmath74 ) . as the first simple model system
, we will discuss a one - dimensional metamaterial system which consists of layers of negative index material ( nim ) and gain material , as shown in , to see if we can compensate the losses of the metamaterials associated with the nims by the amplification provided by the gain material layers and how the system starts lasing .
we first let a narrow band gaussian pulse of a given amplitude go through the metamaterial without gain , and we calculate the transmitted signal emerging from the metamaterial system , which also has a gaussian profile but its amplitude is much smaller than that of the incident pulse due to the losses of nim layers .
then we introduce the gain into the system and start increasing the pumping rate .
the amplitude of the transmitted signal gets larger and we can find a critical pumping rate , for which the transmitted pulse is of the same amplitude as the incident one .
since the gain material itself is nonlinear , we increase the amplitude of the incident gaussian pulse for a fixed pumping rate to see how high we can go in the strength of the incident electric field and still have the full compensation of the losses , i.e. , the transmitted signal equals the incident signal , independent on the signal strength . in this region
we have compensated loss and still have linear response of the metamaterial .
the shape of the transmitted signal is only affected by the dispersion but not dependent on the signal strength . for a three - layer system ( nim - gain material - nim ) , we have calculated the transmission versus the strength of the electric field of the incident signal for several pumping rates close to the critical pumping rate @xmath75 , as shown in .
we found it has a linear response within a very broad range up to incident electric field of @xmath76 .
if we use 19 layers of , the critical pumping rate is @xmath77 , which is even smaller than half of the three - layer case , and the linear regime becomes narrower and drops faster than the 3-layer case for higher strength of incident electric field ( shown in ) . to include the nonlinearity of gain material for strong incident signal , it is necessary to do a self - consistent calculation using fdtd method .
, for different pumping rates @xmath49 .
@xmath49 is increased from @xmath78 ( lowest ) to @xmath79 ( highest ) in steps of @xmath80 .
the material parameters for nim are @xmath81 .
the metamaterial response is linear in a very wide range . when the loss - compensated transmission reaches exactly unity , the pumping rate @xmath82 , which is called the critical pumping rate . for incident fields stronger than @xmath83
the metamaterial behaves nonlinearly.,scaledwidth=40.0% ] as an example , we have also studied the three - layer system with different losses , to see how much @xmath49 we need to compensate the losses .
shows there exists a linear relation between the critical pumping rate and the imaginary part of the refractive index @xmath2 of nims . at the critical pumping rates @xmath84 .
the material parameters for nim are same as . for incident fields stronger than @xmath76
this metamaterial becomes non - linear.,scaledwidth=40.0% ] we have also numerically calculated the susceptibilities of the gain material to see if it really has a lorentzian lineshape .
we first let a gaussian pulse of a given amplitude ( @xmath85 ) propagate through the metamaterial system and calculate the time - domain electric polarization @xmath86 and the local electric field @xmath87 .
then we implement the fourier transforms to obtain the frequency - domain polarization and electric fields and calculate the frequency - dependent susceptibility by using the equation @xmath88 .
simulations are done for both 3 and 19 layers and results are compared with the analytic results calculated using the equations @xcite @xmath89 and @xmath90 with @xmath91 and @xmath92 , where @xmath93 . as shown in , we found that the numerical susceptibilities are the same as the analytic ones and they do have a lorentzian lineshape . of nims .
the structure is a three - layer system ( nim - gain material - nim ) and @xmath94 for nims.,scaledwidth=40.0% ] to understand the lasing behavior of the metamaterial system , we increase the pumping rate to provide more gain from the gain material .
we found the amplification of the incident signal gets larger and at last the system starts lasing till the pumping rate reaches a certain high value .
shows the lasing behavior at the pumping rate @xmath95 and there appears a peak at the emission frequency @xmath60 in the corresponding fourier transform ( ) .
if we use real metal layers instead of negative index materials , we can not compensate the losses of the metals .
the reason is that the permittivity @xmath96 for metals is large and negative and we ll have large reflections due to the impedance mismatch . in
, we simply force the permittivity and the permeability of the metamaterial to be negative to have an unrealistic negative index material . in this section ,
we consider a two - dimensional ( 2d ) metamaterial system ( shown in ) which consists of one layer of gain material and two layers of dielectric wires that have a lorentz - type resonant electric response to emulate the resonant elements in a realistic metamaterial , such as cut - wires . , where the resonance frequency @xmath97 and @xmath98 , and @xmath99 takes different values in the cases we have examined .
the dimensions are @xmath100 , @xmath101 , and @xmath102.,scaledwidth=40.0% ] + we first study the three - layer system of with the full width at half maximum ( fwhm ) @xmath103 for lorentz dielectric ( i.e. , @xmath104 ) and @xmath105 for gain .
the transmission , @xmath106 , the reflection , @xmath107 , and the absorption , @xmath108 , as a function of frequency for the system are obtained in the propagation direction . with the introduction of gain ,
the absorption at the resonance frequency of @xmath60 decreases and reaches 0 at a certain pumping rate .
so the gain compensates the losses .
if we continue increasing the gain , i.e. , the pumping rate , the system gets overcompensated and the absorption becomes negative . to see how the losses of the emulated resonators get compensated by the gain
, we exploit the usual retrieval procedure based on inverting the scattering amplitudes @xcite to obtain the effective permittivities @xmath96 without gain and with gain .
( a ) - 16(c ) show the retrieved results for the real and the imaginary parts of the effective permittivities @xmath96 of the system for the below compensation , loss - compensated , and overcompensated cases , respectively , together with the effective permittivity without gain .
the retrieved results for @xmath96 without gain have exactly the lorentzian shape but the amplitude of the real and the imaginary parts of @xmath96 is a factor of 4 less than the lorentz formula for the square cylinders .
this is due to the filling ratio of the square cylinders in the unit cell . due to the loss compensation from the gain material
, one can see the imaginary part of the effective permittivity gets lower as the gain increases .
below compensation , its value at the resonance frequency of @xmath60 is positive , while it s zero and negative for the loss - compensated and overcompensated cases , respectively .
notice that we can have @xmath109 with @xmath110 at @xmath111 and @xmath112 with @xmath110 at @xmath113 for the below compensation case ( ) . for the loss - compensated case ( ) , we have @xmath114 with @xmath110 at the resonance frequency @xmath60 , which is just same as vacuum and makes no sense for us .
for the overcompensated case ( ) , the imaginary part of the effective permittivity @xmath96 is negative within all the frequency range .
second , we study the three - layer system with the fwhm of gain smaller than lorentz dielectric , where the bandwidths for gain and lorentz dielectric are @xmath103 and @xmath105 ( i.e. , @xmath115 ) , respectively .
the introduction of gain develops a peak at the resonance frequency of @xmath60 for the transmission while the absorption has a dip .
the retrieved results for the real and the imaginary parts of the effective permittivities @xmath96 without gain and with gain are plotted in - 16(f ) for the three different cases discussed above .
similar to the first case we examined where the loss bandwidth is smaller than the gain , the imaginary part of the effective permittivity @xmath96 gets smaller due to the gain .
the difference is we get interesting results for the overcompensated case instead of the below compensation case , where we can have @xmath116 with @xmath117 at @xmath118 and @xmath119 with @xmath117 at @xmath120 .
so for both the two systems , one can obtain a lossless metamaterial with positive or negative @xmath121 , either below compensation or over compensation .
in we also have plotted the sum of @xmath122 without gain and the imaginary part of @xmath123 , the dielectric function of the gain material .
one can see the imaginary part of @xmath96 of our total system with gain is equal to the sum of @xmath122 and @xmath124 .
this is unexpected because there is no coupling between the lorentz dielectric and the gain .
this is indeed true for the 2d lorentz dielectric cylinders , because they have a continuous shape like a solenoid and the gain material slabs have zero depolarization field .
different from finite length wires [ hence a three - dimensional ( 3d ) problem ] where the dipole interactions between lorentz wires and gain material are dominated by the near field @xmath125 , the interaction for infinite length wires is only via the propagating field @xmath126 , and much weaker .
that s why the lorentz wires and the gain material are approximately independent in our 2d simulations .
so there is a need for a true 3d simulation to solve this problem and obtain different behaviors .
however , the 3d simulation is computationally excessively demanding . like the layered system in ,
if we keep increasing the pumping rate , i.e. , the gain , at last both of the two systems will have lasing .
for example , when the pumping rate reaches @xmath127 , the three - layer system with gain bandwidth @xmath103 and lorentz bandwidth @xmath105 starts lasing the system itself has a coherent self - sustained steady output . to avoid the decoupling problem in and still limit our simulations in a 2d model , we consider a 2d split ring resonator ( srr ) as a more realistic and also more relevant model , where the relevant polarization is across the finite srr gap and therefore the coupling to the gain material is dipolelike . in , we show the geometry for the unit cell of the srr system with the gain material embedded in the gap .
the dimensions of the srr are chosen such that the magnetic resonance frequency of the srr overlaps with the emission frequency ( @xmath60 ) of the gain material . due to the strong electric field inside the gap
, there will be strong coupling between the srr and the gain material .
we also want to see if the losses of the magnetic response can be compensated by the electric gain .
simulations are done for one layer of the square srr with gain bandwidth of @xmath105 .
shows the retrieved results for the real and the imaginary parts of the effective permeability @xmath128 , with gain ( pumping rate @xmath129 ) and without gain .
one can see that with the introduction of gain , the gain undamps the magnetic resonance of the srr and the weak and broad resonant effective permeability @xmath128 of the lossy srr becomes strong and narrow .
the fwhm with gain is @xmath130 , while the fwhm without gain is @xmath131 , which is more than twice larger than the former .
notice that in the off - resonance range in , we can obtain the effective permeability @xmath128 with a smaller imaginary part with the introduction of the gain , which means the magnetic loss is compensated by the electric gain .
shows the retrieved results for the real and the imaginary parts of the corresponding effective index of refraction @xmath2 , with and without gain .
note that for a lossless srr , @xmath2 is purely real away from the resonance except in a small band above the resonance where it s purely imaginary due to the negative @xmath128 . at the frequency of @xmath132 , slightly below the resonance ( ) ,
the imaginary parts of the index of refraction @xmath2 without and with gain are @xmath133 and @xmath134
. then we can find the effective extinction coefficient without gain is @xmath135 and the one with gain is @xmath136 . and
hence the effective amplification coefficient of the gain in the combined system is @xmath137 , which is much larger than the amplification @xmath138 for the bulk gain material @xcite at the given pumping rate @xmath129 .
this is due to the strong local electric field enhancement in the gap of the resonant srr .
while we have the incident electric field @xmath85 , the induced electric field in the gap is around @xmath139 . taking the observed field enhancement factor around 45 in the gap of srr , the energy produced by the gain in the gap is around @xmath140 times larger than by the homogeneous bulk gain material in the size of the unit cell , which agrees well with the factor around 15 between the simulated srr effective medium and the homogeneous gain medium . if we continue increasing the pumping rate , the magnetic resonance becomes narrower ( @xmath141 for pumping rate @xmath142 ) .
when the pumping rate reaches @xmath143 , the metamaterial system gets overcompensated and the imaginary part of the effective permeability @xmath128 at the resonance frequency gets flipped down and becomes negative .
if we increase the pumping rate even more ( around @xmath144 ) , the srr system starts lasing @xcite . instead of having the gain in the gap of srr
, we also have done simulations on the srr system where the gain is surrounding around the srr . the retrieved results for the real and the imaginary parts of the effective permeability @xmath128 without gain and with gain for a given pumping rate @xmath145 are plotted in . due to the larger gain
filling in the unit cell , the coupling between the gain material and the srr gets stronger and the losses of srr are easier to be compensated . compared with the case with the gain in the gap only , the effective permeability @xmath128 of the simulated srr system in can have stronger and narrower resonance ( fwhm = @xmath146 ) even with lower pumping rate .
similar to the previous case , the magnetic resonance becomes stronger and narrower ( @xmath147 at @xmath129 ) as we increase the pumping rate , and the system gets overcompensated so that the magnetic resonance peak gets flipped down when the pumping rate reaches @xmath148 .
it s also easier for this system to have lasing @xcite , which is observed when the pumping rate is @xmath149 .
we have performed numerical simulations on one dimensional gain material system by using fdtd method .
the system starts lasing when the input reaches over the lasing threshold . for more input power and wider gain slab ,
the lasing is faster .
comparisons were done for a gain material slab between the optical pump method and its homogeneous pumping rate simplification and results show that this simplification can be valid for a thin gain slab .
a self - consistent model incorporating the gain into dispersive metamaterial nanostructure was proposed and numerically solved .
we numerically show that the losses of dispersive metamaterials can be compensated by gain by investigating the transmission , reflection and absorption data as well as the retrieved effective parameters .
there is a relatively wide range of input signal amplitudes where the metamaterial - gain system behaves linearly .
when the amplitudes get higher , the system becomes nonlinear due to the nonlinearity of the gain material itself .
it s necessary to have self - consistent calculations to determine the signal range where we can expect linear response .
further , if we have strong signals so that we are in nonlinear regime or we want to study lasing , self - consistent calculation is needed . as examples ,
two srr systems with different gain inclusions were studied .
we have demonstrated that the magnetic losses of the srrs can be easily compensated by the electric gain .
the pumping rate needed to compensate the losses is much smaller than the bulk gain material .
the losses of the srr surrounded by gain can be easier to be compensated than the srr with gain in the gap only due to more coupling with the gain .
provided that the pumping rate is high enough , the metamaterial nanostructures can have lasing .
work at ames laboratory was supported by the department of energy ( basic energy sciences ) under contract no .
de - ac02 - 07ch11358 .
this work was partially supported by the office of naval research ( award no .
n00014 - 07 - 1 - 0359 ) , by the laboratory directed research and development program at sandia national laboratories and biotechnology group from afrl / rxb .
99 shalaev v m 2007 _ nature photon . _ * 1 * 41 | a self - consistent computational scheme is presented for one dimensional ( 1d ) and two dimensional ( 2d ) metamaterial systems with gain incorporated into the nanostructures .
the gain is described by a generic four - level system .
the loss compensation and the lasing behavior of the metamaterial system with gain are studied . a critical pumping rate exists for compensating the losses of the metamaterial .
there exists a wide range of input signals where the composite system behaves linearly .
nonlinearities arise for stronger signals due to gain depletion .
the retrieved effective parameters are presented for one layer of gain embedded in two layers of lorentz dielectric rods and split ring resonators with two different gain inclusions : ( 1 ) gain is embedded in the gaps only and ( 2 ) gain is surrounding the srr . when the pumping rate increases , there is a critical pumping rate that the metamaterial system starts lasing .
_ keywords _ : metamaterial , gain material , lasing , loss compensation |
the quantum group @xmath2 is a deformation of the universal enveloping algebra @xmath3 of a semisimple lie algebra . for lie algebras , a theorem by sophus
lie states that the borel subalgebras are the maximal solvable lie - subalgebras and that they are all conjugate .
+ in @xcite we ask for quantum groups the same question : which are the maximal right coideal subalgebras with the representation theoretic property , that all irreducible finite - dimensional representations are one - dimensional .
we call this a _ borel subalgebra _ of @xmath2 .
besides the standard borel subalgebra @xmath4 and its reflections , there is already for @xmath5 a new family of borel subalgebras , generated by elements @xmath6 which are isomorphic to the quantized weyl algebra .
note that they are not defined for @xmath7 and interpolate between @xmath8 and @xmath9 .
we have conjectured ( and proven for type @xmath0 ) a main structural result on the graded algebra associated to any right coideal subalgebra that has a triangular decomposition . as a second conjecture ,
this suggests an explicit description of all triangular borel subalgebras .
+ the goal of the present article is to directly construct and classify triangular borel subalgebras fulfilling an additional non - degeneracy condition ( [ nondegeneracy ] ) .
the algebras we construct consist of copies of the quantized weyl algebra , as well as a subset of the remaining standard borel subalgebra . in type @xmath0
we can prove that all triangular non - degenerate borel subalgebras are of this form .
we also discuss some degenerate examples for type @xmath0 . + our main reason for studying borel subalgebras is to gain more knowledge on the theory of arbitrary right coideal subalgebras .
another curious application is that for a given borel subalgebra one may look at induced modules from one - dimensional representations . for the new borel subalgebras of @xmath10 this
yields infinite - dimensional reprentations isomophic as a vector space to @xmath11 $ ] , on which @xmath12 acts by multiplication .
we could prove that most of these induced modules are irreducible , while a discrete family has as quotients the irreducible finite - dimensional representations of @xmath10 .
it would be interesting to study an analogue of category @xmath1 for these borel subalgebras .
+ after preliminaries on right coideal subalgebras in section 2 we proceed as follows : + in section [ sec_counterindicators ] we prepare a strategy to construct non - one - dimensional irreducible representations of right coideal subalgebras by restricting suitable @xmath2-modules and detecting non - one - dimensional composition factors .
in section [ sec_weyl ] we study a curious questions about weyl groups , namely about `` filling up '' two given weyl group elements to maximal weyl group elements that retain the same property .
this theorem later - on precisely tells us how far an interesting coideal subalgebra can be filled up until it is maximal . in this paper
we were only able to prove this weyl group assertion for type @xmath0 .
in section [ sec_nondegenerate ] we construct borel subalgebras that consist of several quantized weyl algebras associated to a set of pairwise orthogonal simple roots , together with a suitable subset of the remaining positive roots , see main theorem [ konstruktionborel ]
. then we can prove that under the additional nondegeneracy condition these are in fact all borel subalgebras for type @xmath0 .
finally we determine the induced representations of one - dimensional representations of the borel subalgebra ; these are infinite - dimensional representations with a non - diagonal action of the cartan part of @xmath2 . in section [ sec_degenerate ]
we also treat the next - difficult step , which are borel subalgebras , where the degeneracy height is @xmath13 in some sense .
the degeneracy causes extensions of the weyl algebras . in this case
we first derive necessary conditions for the weyl group elements and then again construct for type @xmath0 right coideal subalgebras where all irreducible representations are one - dimensional . in section [ sec_example ]
we discuss as examples all borel subalgebras for @xmath14 by - hand to compare this with our conjectures and results .
moreover we determine all triangular borel subalgebras for @xmath15 , which is the first case where a degeneracy of height @xmath16 appears .
this is worked out thoroughly by the second author in @xcite , using a generating system for an arbitrary right coideal subalgebra .
let @xmath17 be a finite - dimensional semisimple lie algebra of rank @xmath18 over the field of complex numbers @xmath19 .
we denote by @xmath20 a set of positive simple roots , by @xmath21 the root lattice , and by @xmath22 the set of all positive roots .
we denote by @xmath23 the symmetric bilinear form on @xmath24 with the cartan matrix @xmath25 .
our article is concerned with the quantum group @xmath2 where @xmath26 is not a root of unity .
there are algebra automorphisms due to lusztig @xmath27 for each weyl group element @xmath28 , and with these one constructs root vectors @xmath29 for all @xmath30 , see @xcite chapter 8 .
a subalgebra @xmath31 of a hopf algebra @xmath32 is called a right coideal subalgebra ( rcs ) if @xmath33 . three essential results in the theory of coideal subalgebras of quantum groups
are : + call a right coideal subalgebra @xmath34 homogeneous iff @xmath35 ( in particular @xmath31 is then homogeneous with respect to the @xmath21-grading ) . [ hs09 ] for every @xmath28 there is an rcs @xmath36u^0 $ ] , where @xmath36 $ ] is generated by the root vectors @xmath37 for all @xmath38 in the subset of roots @xmath39 in particular @xmath40 , and @xmath41 and for the longest element @xmath42 .
conversely , every homogeneous rcs @xmath43 is of this form for some @xmath44 .
the homogeneous rcs @xmath34 are of the form @xmath45u^0u^-[v]\ ] ] for a certain subset of pairs @xmath46 .
non - homogeneous rcs are only classified on @xmath47 ( or @xmath48 ) : [ hk11b ] for @xmath28 , let @xmath49\to { \mathbb{k}}$ ] be a character and define @xmath50 \text { with } \phi(x_\beta)\neq 0\}\ ] ] take any subgroup @xmath51 , then there exists a character - shifted rcs @xmath52_\phi:=\{\phi(x^{(1)})x^{(2)}\mid \forall x\in u_q^-[w]\}\ ] ] and an rcs @xmath53_\phi t_l$ ] with group ring @xmath54\subset u^0 $ ] .
conversely , every rcs @xmath55 if of this form to construct non - homogeneous rcs @xmath34 we shall in the following restrict our attention to : [ deftriang ] we call a right coideal subalgebra _ triangular _ , if each element splits into elements in @xmath56 and @xmath57 : @xmath58 we denote @xmath59 , @xmath60 and @xmath61 and @xmath62
. our main interest is we call an algebra _ ede _ iff every finite - dimensional irreducible representation is one - dimensional .
we call an rcs of @xmath2 a _ borel subalgebra _ iff it is ede and it is a maximal rcs with this property .
[ standardborel ] all homogeneous borel subalgebras are isomorphic ( via some @xmath27 ) to the _ standard borel subalgebra _ @xmath63 .
the fact that this is a borel subalgebra is not entirely trivial , see @xcite .
[ weyl algebra ] there is a family of non - homogeneous borel subalgebras of @xmath10 generated by @xmath64 different choices of @xmath65 are isomophic via some hopf automorphism @xmath66 . + as an algebra , this is isomorphic to a _ quantized weyl algebra _
we want to prove the ede property of the quantized weyl algebra : let @xmath68 be a finite dimensional irreducible representation .
consider the eigenvector @xmath69 of the element @xmath70 with eigenvalue @xmath71 .
one can easily see , that @xmath72 is an eigenvector with the eigenvalue @xmath73 : @xmath74 similarly , one can show that @xmath75 is an eigenvector of @xmath76 with eigenvalue @xmath77 : @xmath78 thus the eigenvectors of @xmath76 are a basis of @xmath68 , as @xmath68 is irreducible . on the other hand for each @xmath79
there are eigenvectors @xmath80 of @xmath76 .
as @xmath68 is finite dimensional , they can not have pairwise distinct eigenvalues .
if @xmath76 has eigenvalue @xmath81 , @xmath13 would thus be an eigenvalue too and there would be an infinite number of different eigenvalues , as @xmath26 is not a root of unity .
thus the eigenvalue @xmath81 is not possible , and so each two eigenvectors must be equal and thus each two eigenvalues must be equal
. then follows @xmath82 . with this @xmath71
we get : @xmath83 as @xmath76 has only the eigenvalue @xmath71 , it acts as a scalar on @xmath68 , the same is true for @xmath84 .
thus @xmath85 and @xmath86 commute on all of @xmath68 and then each finite dimensional irreducible representation is @xmath13- dimensional .
a main structure theory of general borel subalgebras stems from : [ vermutunghecken ] the map @xmath87_{\phi})\to u^{\leq0}$ ] , sending all elements to their leading terms is an injektive homomorphism of @xmath88-graded right coideal subalgebras .
the image @xmath89 of @xmath90 has the following form : * @xmath91 is the monoid ( ! ) @xmath92 * for the quotient group @xmath93 of @xmath94 we get @xmath95t_l$ ] for the following @xmath96 : @xmath97 as all elements in @xmath98 are pairwise orthogonal , @xmath99 is the element , which arises from @xmath44 by deleting all factors @xmath100 , for all @xmath79 with @xmath101 .
[ beweishecken][@xcite ] conjecture [ vermutunghecken ] holds for @xmath0 .
this conjecture has implications to the representation theory of the rcs : [ borelvermutung][@xcite ] let @xmath102_{\phi^-}t_l\psi(u^+[w^+])_{\phi^+}$ ] be a triangular right coideal subalgebra with @xmath103 and @xmath104 , where @xmath105 for all elements @xmath106 of degree @xmath107 , which is essentially the antipode .
then @xmath31 is a borel subalgebra , if and only if @xmath108 .
+ at least we could prove : [ annichtede][@xcite ] in the case @xmath0 holds : given @xmath44 and @xmath99 as in conjecture [ vermutunghecken ] and given a triangular right coideal subalgebra @xmath102_{\phi^-}t_l\psi(u^+[w^+])_{\phi^+}$ ] with @xmath109 . if @xmath110 then @xmath31 is not ede .
[ sl2rep ] for all one - dimensional representations @xmath68 of @xmath10 follows in particular @xmath111 and @xmath112 .
this can be seen directly from the relations : due to the relation @xmath113 we get that in any one dimensional representation @xmath12 has to act as @xmath114 . from @xmath115 and @xmath116 thus follows , that @xmath117 and @xmath118 have to act trivial .
this gives a first method to detect non - one - dimensional composition factors in representations restricted to rcs .
a direct application is that rcs containing the semisimple subalgebra @xmath10 are not ede as follows .
later we will deal with restrictions to non - semisimple subalgebras : [ k-1inb ] let @xmath119 be a borelsubalgebra . if for @xmath120 the element @xmath121 lies in @xmath122 , then the element @xmath123 lies in @xmath122 too .
assume @xmath124 , consider @xmath125 : this is an rcs as both @xmath122 and @xmath126 $ ] are rcs .
due to the assumed maximality of @xmath127 there exists an irreducible representation @xmath68 of @xmath128 with @xmath129 .
as @xmath122 is ede , the representation @xmath130 has a one dimensional subrepresentation @xmath131 and with @xmath132 we get @xmath133 for some @xmath134 .
but this already gives a one - dimensional subrepresentation of @xmath68 which contradicts the maximality of @xmath122 .
[ ef ] let @xmath128 resp .
@xmath135 be subalgebras of @xmath136 with @xmath137 resp .
@xmath138 for some @xmath139 .
then @xmath128 resp .
@xmath135 has a multidimensional irreducible representation .
let @xmath94 be a representation of @xmath128 .
consider first the restriction of @xmath128 on @xmath140 generated by @xmath141 and @xmath142 .
due to corollary [ sl2rep ] we know , that in each finite dimensional representation of @xmath10 the element @xmath143 acts as @xmath13 .
we know that restricted to the semisimple algebra @xmath10 the representation @xmath94 decomposes into irreducible representations .
assume each finite dimensional irreducible representation of @xmath128 is one dimensional , then @xmath144 decomposes in one - dimensional representations , thus @xmath143 acts as @xmath13 on @xmath94 .
it suffices to prove that there is a finite - dimensional representation on @xmath136 , for which @xmath143 does not act as @xmath145 . we know
, that for all dominant weights @xmath146 the @xmath136-module @xmath147 is one dimensional . with the fundamental dominant weights we can find for all @xmath148 a @xmath149 with @xmath150 , and for this @xmath149 the element @xmath143 does not act as @xmath145 on @xmath147 .
this proves the assertion .
+ in the second case : @xmath138 the proof is completely analogous : we would suspect that @xmath135-modules decomposing into their @xmath143-eigenspaces are semisimple with irreducible modules @xmath151 .
this is however not necessary for the proof : the restriction of an @xmath136-module @xmath152 decomposes in @xmath10 semisimple into @xmath153 .
apparently , the restriction of any irreducible @xmath10-module @xmath153 on @xmath135 is still irreducible .
@xmath154 still decomposes into its irreducible representations and as the element @xmath143 does not act as @xmath13 , by choice of @xmath149 , again we can find a multidimensional representation .
[ efohnek ] let @xmath155 be an rcs with @xmath156 for @xmath157 , then @xmath128 is not a borel subalgebra .
this follows from the previous theorem , but we need to get the @xmath158 inside @xmath128 : assume @xmath128 was a borelsubalgebra .
consider the q - commutator of @xmath159 with @xmath160 : @xmath161_{q_{\alpha}^2}=(1-k_{\alpha}^{-2})\frac{q_{\alpha}^2}{q_{\alpha}-q_{\alpha}^{-1}}\ ] ] thus @xmath162 and due to lemma [ k-1inb ] also @xmath143 lie in @xmath128 .
then applying theorem [ ef ] proves the claim .
[ traegerschnittpi ] let @xmath163)_{\phi^+}t_lu^-[w^-]_{\phi^-}$ ] be an arbitrary triangular rcs .
if there exists an @xmath164 with @xmath165 , then @xmath31 is not a borel subalgebra .
[ nurweyl ] if for an @xmath157 the character - shifted elements @xmath166 with @xmath167 lie in a subalgebra @xmath155 , then @xmath128 is not a borel subalgebra . if however for two elements @xmath168 holds @xmath169 they form a weyl algebra as in example [ weyl algebra ] , thus they can be part of a borel subalgebra .
considering the @xmath170-commutator of the two elements yields : @xmath171_{q_{\alpha}^2}=\frac{q_{\alpha}^2}{q_{\alpha}-q_{\alpha}^{-1}}(1-k_{\alpha}^{-2})+(1-q_{\alpha}^2)\lambda_{\alpha } \lambda_{\alpha } ' k_{\alpha}^{-2}\ ] ] we can see , that for @xmath167 again @xmath162 lies in @xmath128 .
+ assume @xmath128 was a borel subalgebra , then due to lemma [ k-1inb ] also @xmath172 .
then similar to corollary [ traegerschnittpi ] also @xmath137 follows , which is a contradiction by theorem [ ef ] . even without the ede property
follows quickly : [ lemma3.16 ] in a triangular right coideal subalgebra @xmath173)_{\phi^+}t_lu^-[w_2]_{\phi^-}$ ] with @xmath109 holds for all @xmath174 the relation @xmath175 from @xmath176 we can follow as in lemma [ nurweyl ] again @xmath177 which is a contradiction to @xmath109 .
we now consider the restriction of @xmath2-representations to @xmath31 ( which is in general nonsemisimple ) in order to construct non - one - dimensional irreducible @xmath31-representations as composition factors .
this works particularly smooth for minuscule representations , which is sufficient for type @xmath0 .
[ notationwurzelninan ] for the roots @xmath178 in @xmath0 we use the notation @xmath179 $ ] for @xmath180 if @xmath181 .
all roots in @xmath0 are of this form .
[ nurdiemittezaehlt ] let @xmath182 and @xmath183 , @xmath184 be two character - shifted root vectors in @xmath163)_{\phi^+}t_lu^-[w^-]_{\phi^-}\subset u_q({\mathfrak{g}})$ ] for a root @xmath179\in\phi^+(w^+)\cap\phi^+(w^-)$ ] such that for all @xmath185 with @xmath186 holds @xmath187 , then @xmath31 is not ede .
our general strategy for constructing irreducible representations of dimension greater @xmath13 is the following : we consider the action of the commutator @xmath188_1\in c$ ] on a suitable finite - dimensional @xmath2-representation @xmath189 , which we restrict to @xmath31 . the commutator acts trivial on any finite dimensional irreducible representation of @xmath31 and thus nilpotent on all representations of @xmath31 , whose series of compositions contains only one - dimensional representations . + in our case we get from @xmath187 for all @xmath190 and @xmath191 that @xmath192,[\mu_1,k']\notin ( { \mathtt{supp}}(\phi^+)\cap { \mathtt{supp}}(\phi^-))$ ] . due to this restriction for @xmath193 the element @xmath183 acts on the minuscule representation @xmath194 with highest weight @xmath195 equal to @xmath196 , and
@xmath184 acts equal to @xmath197 .
+ _ if _ for a given @xmath198 we can find a @xmath199 , such that @xmath200_1.v = cv$ ] with @xmath201 , the commutator can not act nilpotent and thus @xmath202 must have an irreducible composition factor of dimension greater @xmath13 , so @xmath31 is not ede . for example
this condition is for @xmath203 fulfilled for the highest weight vector @xmath69 , but in general we do nt know @xmath200_1 $ ] .
+ in the case of @xmath0 we can calculate the condition explicitly on the minuscule representation @xmath204 .
let @xmath179=\sum_{k=\mu_1}^{\mu_2}\alpha_k$ ] .
then the action of the simple root vectors is : @xmath205 and all weight spaces are one dimensional .
consider @xmath206 , then @xmath207 is trivial , @xmath208 is a non trivial multiplier of @xmath209 , because only the summand @xmath210 can act nontrivial , and with the same argumentation also @xmath211 is a nontrivial multiplier of @xmath69 .
thus @xmath69 is a nontrivial eigenvector of the commutator .
thus , according to the above explanation , the assertion follows .
[ folgerungnurdiemittezaehlt ] let @xmath182 and @xmath183 , @xmath184 be two character - shifted root vectors in @xmath163)_{\phi^+}t_lu^-[w^-]_{\phi^-}\subset u_q({\mathfrak{g}})$ ] for a root @xmath179\in { \mathtt{supp}}(\phi^+)\cap { \mathtt{supp}}(\phi^-)$ ] such that for all @xmath212 with @xmath186 holds @xmath187 , and @xmath213 , @xmath214 with @xmath215 , then @xmath31 is not ede . as in the proof of lemma [ nurdiemittezaehlt ] we can see that , on the minuscule representation with highest weight @xmath216 the element @xmath183 acts equal to @xmath217 and @xmath184 equal to @xmath218 .
thus the representation has a multidimensional irreducible subrepresentation .
[ dasssiesichnichtschneiden ] let @xmath219)_{\phi^+}t_lu^-[w]_{\phi^-}\subset u_q({\mathfrak{sl}}_n)$ ] such that @xmath220 and given two character - shifted root vectors @xmath221 for a root @xmath179\notin\pi$ ] with @xmath222 , but @xmath223 .
if there are two different reduced expressions @xmath224 and @xmath225 of @xmath44 with @xmath226 and @xmath227,such that @xmath228 for two elements @xmath229 and @xmath230 , then @xmath31 is not ede . if @xmath231 and @xmath232 then the situation is a special case of the preceding lemma [ nurdiemittezaehlt ] .
thus it is enough to consider the case @xmath233 ( resp .
@xmath234 ) : then @xmath183 acts on the minuscule representation with highest weight @xmath235 equal to @xmath196 .
thus the requirements of the proof of lemma [ nurdiemittezaehlt ] are fulfilled and we can find a multidimensional irreducible representation as factor of the decomposition of @xmath236 . on the other hand , a method of finding one - dimensional subrepresentations and thereby proving the ede property is : [ maxials0allgemein ] let @xmath128 be an algebra with a generating system @xmath237 and with the following property : there is an element @xmath238 such that for all generators @xmath239 the commutator @xmath240_c=0 $ ] for some constant @xmath241 , and there is an element @xmath242 with @xmath243_{c'}=0 $ ] for a constant @xmath244 , which is not a root of unity , and @xmath12 has a non - trivial eigenvector .
then @xmath245 for any finite dimensional irreducible representation @xmath68 of @xmath128 .
assume there is an eigenvector of @xmath85 with eigenvalue @xmath81 , then all elements @xmath246 with @xmath247 span a subrepresentation @xmath68 , because by assumption after acting with any generator @xmath86 : @xmath248 because @xmath68 is assumed irreducible , the existence of an eigenvalue @xmath81 thus implies @xmath249 for all @xmath250 .
+ assume now there is no eigenvector of @xmath85 with eigenvalue @xmath81 , then let us consider the assumed eigenvector @xmath251 of @xmath12 , with associated eigenvalue some @xmath149 . as @xmath252
we get @xmath253 is again an eigenvector of @xmath12 : @xmath254 as @xmath255 is assumed to be not a root of unity and as @xmath256 for all @xmath257 , the elements @xmath258 are eigenvectors of @xmath12 with pairwise distinct eigenvalues . as @xmath68 is finite dimensional ,
only a finite number of these values can exist , this yields a contradiction .
for @xmath28 consider as in theorem [ hs09 ] the subset of roots : @xmath259 the goal of this section is to prove , how to supplement two weyl group elements @xmath260 in a special way to two elements @xmath261 , such that @xmath262 .
[ weyltheorem ] fix @xmath260 and assume @xmath263 is not empty and consists of pairwise orthogonal roots . then : 1 .
there exist elements @xmath264 with @xmath265 , @xmath266 , such that the following relations hold : @xmath267 2 .
there exists an element @xmath268 such that @xmath269 and @xmath270 has a reduced expression @xmath271 with @xmath272 we can prove this theorem for type @xmath0 .
the rest of this section is devoted to prove theorem [ weyltheorem ] .
in fact most steps are always true , but a key property of root systems [ kinbvermutung ] we did only prove for @xmath0 and for rank @xmath16 . + we consider two orderings on the root lattice : for @xmath273 with @xmath274 let @xmath275 if and only if @xmath276 for all @xmath277 .
+ we define , depending on the choice of the reduced expression of @xmath44 , a total ordering @xmath278 on @xmath279 by @xmath280 this ordering is convex , i.e. for @xmath281 we get @xmath282 , if @xmath283 see therefore @xcite s.662 .
on @xmath284 we use the weak ordering . from the strong exchange property
follows : [ piamanfang ] let @xmath28 and @xmath285 , then there exists a reduced expression of @xmath44 of the form @xmath286 for a @xmath287 with @xmath288 .
@xmath289 [ charakterdurchstreichen ] in explicit examples of small rank the set @xmath290 is easy to calculate : if @xmath291 is a subset , such that the elements in @xmath292 are pairwise orthogonal , then @xmath293 belongs to @xmath290 if and only if after deleting all reflections @xmath100 for @xmath294 the resulting expression @xmath295 is still a reduced expression .
[ rangk ] each @xmath296 roots in @xmath297 can be reflected to a rank @xmath296 root system , i.e. for any set of roots @xmath298 there is a @xmath287 and a rank @xmath299 root system @xmath300 with @xmath301 for all @xmath79 .
a proof can be found e.g. in @xcite theorem 6.3 a ) .
[ immerinalpha ] for a @xmath302 not simple , there exists a @xmath303 such that @xmath304 . the proof is found in @xcite lemma 10.2 a. [ immerinalphafolg ] given a @xmath28 .
for all @xmath305 there exists a @xmath306 such that @xmath307 .
we prove this by induction on the height @xmath308 : + for @xmath309 follows @xmath203 and the claim is trivial .
let @xmath310 .
due to lemma [ immerinalpha ] there exists a @xmath303 such that @xmath304 . due to the convexity of the ordering @xmath278 on @xmath279
we get either @xmath306 or @xmath311 . in the first case
the claim is already proven . in the second case there
exists from the induction assumption an @xmath312 with @xmath313 , thus the claim is true .
let @xmath314 be orthogonal ( @xmath315 ) , that is @xmath316 .
we call @xmath178 and @xmath317 _ strongly orthogonal _ , if there exists a @xmath287 such that @xmath318
. using lemma [ rangk ] we can reduce to the rank @xmath16 situation .
then clearly [ starkortlinear ] two orthogonal roots @xmath314 are strongly orthogonal , if @xmath319 for all @xmath320 . in the following
we study for two weyl group elements @xmath260 the intersection @xmath321 more closely .
in particular we want to make some statements on possible supplements of @xmath224 and @xmath225 for @xmath322 consisting of orthogonal elements .
+ [ starkort ] assume all elements in @xmath322 are pairwise orthogonal .
then all the elements in @xmath322 are pairwise strongly orthogonal . given two arbitrary elements @xmath323 .
assume @xmath324 , we get from the convexity of @xmath278 on @xmath325 resp . @xmath326
already @xmath327 , so in particular @xmath328 .
that is a contradiction to the pairwise orthogonality of the elements in @xmath329 .
now we use a projection @xmath287 from @xmath178 and @xmath317 in a rank @xmath16 root system . from @xmath330 ,
i.e. @xmath331 the only options for the root systems are @xmath332 and @xmath333 , in the first case the claim is proven . in the second case
@xmath334 with a long root @xmath178 and a short root @xmath317 orthogonal to each other , also the roots between @xmath178 and @xmath317 with respect to the total ordering for an arbitrary reduced expression of @xmath335 lie in @xmath336 as is easily shown , these are however not orthogonal on @xmath178 , which leads us to a contradiction . [ kinbvermutung ]
assume all elements in @xmath322 are pairwise orthogonal .
then there exists either an element @xmath337 with @xmath338 or @xmath339 . in the small cases of rank 2 the assumption [ kinbvermutung ] is true , as we know from lemma [ starkort ] , that the elements in @xmath329 are even pairwise strongly orthogonal .
thus in the case of rank 2 only @xmath340 is possible and the claim follows directly . in the following
we will restrict our considerations to @xmath0 , but we conjecture , that the theorems are also true in general .
[ kinb ] in type @xmath0 conjecture [ kinbvermutung ] holds .
assume there exists no element @xmath341 with @xmath338 , i.e. : @xmath342 we want to show @xmath343 : + a ) for all @xmath344 we get from lemma [ immerinalphafolg ] a @xmath345 with @xmath346 . due to ( [ kinbannahme ] ) even @xmath347 holds .
+ b ) assume there is a @xmath348 with @xmath349 .
we look at the minimal root @xmath179 $ ] of this type with respect to the ordering @xmath350 .
let @xmath351 with @xmath352 be the elements @xmath353 in @xmath329 .
due to a ) there exists at least one such element and from the pairwise orthogonality of the elements in @xmath322 we even get @xmath354 .
we define the following types of roots for @xmath355 : @xmath356\ ] ] @xmath357\ ] ] here for all @xmath355 holds @xmath358 because @xmath359 and @xmath360 and because the elements in @xmath329 are pairwise orthogonal .
+ we want to show now for all @xmath355 holds @xmath361 . for this purpose
, we use the following considerations : 1 .
@xmath362 , as @xmath363 .
as well for @xmath225 .
+ @xmath364 , as @xmath363 .
as well for @xmath225 .
@xmath365 , this is due to the convexity of @xmath278 .
as well for @xmath225 and @xmath366 .
+ @xmath367 , this is due to the convexity of @xmath278 .
as well for @xmath225 and @xmath366 .
@xmath368 , as @xmath369 .
as well for @xmath366 .
+ @xmath370 , as @xmath371 .
as well for @xmath366 . with these considerations we can prove the claim inductively on @xmath355 : + @xmath372 lies in @xmath325 , because : @xmath373 let @xmath361 .
assume @xmath374 : @xmath375 the induction yields @xmath361 for all @xmath355 .
this is a contradiction to a ) for @xmath376 .
this proves the assertion .
[ charakterebeisammen ] let @xmath329 consist of pairwise orthogonal roots and let @xmath377 .
then @xmath224 has a reduced expression of the form @xmath378 and @xmath379 with @xmath380 . as well for @xmath225 .
we prove this claim with an induction on the length @xmath381 : + for @xmath382 we get @xmath383 and the claim is trivial .
let now @xmath44 be with @xmath384 .
assume @xmath385 then , due to corollary [ piamanfang ] and as the elements in @xmath322 are pairwise orthogonal , we can write all elements in @xmath322 in the beginning of @xmath44 .
then the claim is true for @xmath386 and @xmath387 . if an element @xmath348 with @xmath349 exists , then due to lemma [ kinb ] there is a @xmath345 with @xmath346 and @xmath388 .
so there is a reduced expression of @xmath224 with @xmath389 in the beginning , due to corollary [ piamanfang ] .
then we look at the two weyl group elements @xmath390 and @xmath391 . here
the claim is already proven , as @xmath392 and for @xmath393 we can bring the element @xmath394 in the desired form . with this reduced expression of @xmath395 , as @xmath396 also @xmath397 is in the right form .
now we want to prove our main theorem [ weyltheorem ] which gives us an supplement of @xmath398 to @xmath261 .
we will do this in several steps .
[ weylauffullen][alphaschnitt ] under the assumptions of theorem [ weyltheorem ] the two assertions 1 ) and 2 ) are equivalent .
`` @xmath399 '' let @xmath400 , fulfilling 1 ) be given i.e. @xmath401 and @xmath402 .
now we construct for @xmath403 a reduced expression of the required form . + for this we look at @xmath404 : @xmath405 because for @xmath406 already @xmath407 is true , so @xmath408 and from @xmath402 it follows , that @xmath409 . due to the relation [ redexpvonteilmenge ] we can choose now the following reduced expression of @xmath394 : @xmath410 for some @xmath287 with @xmath411 .
it only remains to prove : @xmath412 `` @xmath413 '' : this is true , because the elements in @xmath329 are exactly the elements which lie in both @xmath414 and @xmath415 .
the elements on the left hand side lie in @xmath415 by construction and for @xmath416 holds @xmath417 , so they also lie in @xmath418 .
thus we have found a suitable reduced expression and 2 ) holds .
+ `` @xmath419 '' : this is true , because @xmath420 , on the other hand for all elements @xmath421 , so in particular for all elements @xmath422 we have @xmath423 , i.e. @xmath424 . so we get @xmath425 , but @xmath424 , so @xmath426 .
+ `` @xmath427 '' assume there exists a @xmath428 with @xmath429 for an @xmath79 .
+ then we choose @xmath430 and @xmath431 and show that these elements fulfil 1 ) : we know @xmath432 as for @xmath433 we have @xmath434 , so @xmath409 .
moreover we know @xmath435 as for all @xmath436 is @xmath437 . to prove theorem [ weyltheorem ] we look at some special cases for @xmath322 and
will then combine the methods of the proofs .
[ weyltheoremmit1 ] for @xmath438 theorem [ weyltheorem ] is true . if @xmath439 , then the claim is trivial .
. then we construct elements @xmath394 and @xmath441 with @xmath265 , @xmath266 inductively as follows : we enlarge @xmath224 until for @xmath442 and @xmath225 either 1 ) or 2 ) in the theorem [ weyltheorem ] is fulfilled : let @xmath443 be the following sequence of elements in @xmath284 with associated reduced expressions , such that @xmath444 is true for all @xmath79 , that is @xmath445 for all @xmath79 : @xmath446 for elements @xmath447 with @xmath448 and @xmath449 .
these reduced expressions exist due to corollary [ piamanfang ] .
then in particular @xmath450 with @xmath451 .
as @xmath440 there exists at least a @xmath452 with @xmath453 , so with @xmath451 also @xmath454 is true .
now we distinguish two cases for @xmath455 : + _ 1 .
@xmath456 : then we define @xmath457 and we get @xmath458 . by induction
we can enlarge @xmath224 , until either @xmath459 , i.e. 1 ) is fulfilled , or for some @xmath79 the 2 .
case occurs .
@xmath460 : then we get from @xmath454 already @xmath461 . in this case
there is a reduced expression of @xmath443 of the form : @xmath462 and 2 ) is fulfilled for @xmath463 and @xmath464 , because @xmath465 , i.e. we know from lemma [ piamanfang ] applied to @xmath466 , that there exists a reduced expression of @xmath443 of the form @xmath467 . in this reduced expression
we can see , that in particular holds : @xmath468 , i.e. @xmath469 . thus the claim is proven .
[ weylinpi ] for @xmath385 the theorem [ weyltheorem ] is true .
we want to prove this corollary analogous to the proof of corollary [ weyltheoremmit1 ] . from @xmath470 and lemma [ piamanfang ] we know , that there exists a reduced expression of @xmath224 of the form : @xmath471 for a @xmath287 with @xmath472 , the same for @xmath225 . as above
we construct again elements @xmath261 inductively : let therefore @xmath443 be the following sequence of elements in @xmath284 with associated reduced expression , such that @xmath444 holds for all @xmath79 , so @xmath445 for all @xmath79 : @xmath473 for elements @xmath447 with @xmath474 and with @xmath475
. again we get @xmath450 with @xmath451 . as in the proof above we search for elements @xmath476 with @xmath477 .
i.e. with @xmath478 .
for these elements we get @xmath479 , from @xmath451 .
if we find such elements we define @xmath480 until @xmath481 or there exists an @xmath79 , for which @xmath482 fulfils the property 2 ) from the theorem [ weyltheorem ] .
this works as follows : + _ 1 . case _ : there exists a @xmath483 with @xmath484 and @xmath485 .
+ then the requirements for the induction are fulfilled , as from @xmath485 follows @xmath486 and therefore @xmath487 .
then we define again @xmath480 .
we get @xmath488 .
so we can continue the induction with this @xmath483 until either in the theorem [ weyltheorem ] assertion 1 ) is fulfilled or the second case occurs .
case _ : for all @xmath489 with @xmath490 holds @xmath491 .
+ then we take a closer look at the following elements : @xmath492 for these elements the requirements from lemma [ kinb ] are fulfilled , because : * @xmath493 .
this can be seen as follows : * * @xmath494 , because for all @xmath495 we get @xmath496 * * for all @xmath497 holds @xmath498 , because : + assume @xmath499 * @xmath500 consists of pairwise strongly orthogonal roots , as @xmath329 consists of pairwise strongly orthogonal roots * for all @xmath501 we know that @xmath489 lies in @xmath502 , because we get @xmath491 and @xmath503 , so @xmath504 and @xmath505 .
thus we can apply lemma [ kinb ] , this provides that there exist exactly @xmath299 elements in @xmath506 .
we call these elements @xmath507 , they are pairwise strongly orthogonal and we define : @xmath508 for all @xmath509 we get @xmath510 , because @xmath511 for some @xmath512 with @xmath513 , then @xmath514 .
+ due to @xmath515 and as the @xmath516 are pairwise orthogonal there exist by lemma [ piamanfang ] a reduced expression of @xmath466 beginning with @xmath517 .
the associated reduced expression of @xmath443 fulfils the requirements of theorem [ weyltheorem ] .
this follows analogously to corollary [ weyltheoremmit1 ] and the claim is proven .
the general case of arbitrary @xmath518 can be reduced to the special case in corollary [ weylinpi ] with corollary [ charakterebeisammen ] .
let therefore be a reduced expression of @xmath224 given as in corollary [ charakterebeisammen ] : @xmath519 and @xmath520 with @xmath521 .
then we look at the elements @xmath522 and @xmath523 . for these elements
@xmath524 , i.e. the requirements of corollary [ weylinpi ] are fulfilled and we can apply the theorem [ weyltheorem ] .
now we choose for @xmath224 and @xmath225 the following elements : @xmath525 and @xmath526 , we get for the intersection : @xmath527 because @xmath528 and further we get : @xmath529 moreover by induction on the length @xmath79 , we get for the union of @xmath415 and @xmath414 : @xmath432 thus the theorem [ weyltheorem ] is proven for an arbitrary @xmath322 .
in this section we construct a large family of borel subalgebras associated to a choice @xmath530 of an orthogonal subset of simple roots @xmath531 ( i.e. a coclique in a dynkin diagram ) and an associated character @xmath532 , which is uniquely given by a family of scalars @xmath533 .
these borel subalgebras look like a disjoint family of weyl algebras for each element in @xmath531 , filled up with a suitable maximal set of remaining positive roots .
+ the borel subalgebras we construct are by construction triangular and fulfil an additional non - degeneracy property : @xmath534 where for an arbitrary triangular rcs the relation @xmath535 holds by construction . in the next section
we will then prove that all triangular nondegenerate borel subalgebras are of the type constructed here ( assumed type @xmath0 ) .
+ consider now the right coideal subalgebra @xmath536)_{\phi^+}t_lu^-[x]_{\phi^-}\ ] ] where @xmath537 is the longest element of the weyl group , @xmath287 is a choice of a weyl group element such that @xmath538 consists of pairwise orthogonal roots ; i.e. @xmath539 .
furthermore let @xmath540 be characters with @xmath541 .
any such character is defined by : @xmath542 for @xmath543 and @xmath81 otherwise and @xmath544 and @xmath81 otherwise with the condition @xmath545 moreover let @xmath546 ; this choice is for maximality , there is an ede rcs for any choice @xmath547 .
+ using the commutation relations in @xmath31 we will show first , that @xmath31 is a triangular right coideal subalgebra , that means closed under multiplication .
afterwards we will show , that right coideal subalgebras of this form are ede , that means they have the property , that each finite dimensional irreducible representation is one dimensional , and if furthermore @xmath546 we show , that @xmath31 is a borel subalgebra .
[ konstruktionborel ] the right coideal subalgebra @xmath536)_{\phi^+}t_lu^-[x]_{\phi^-}\ ] ] with data @xmath548 as discussed above , is a borel subalgebra of @xmath2 .
in the rest of the section we will work out the proof of this theorem .
+ first we take a closer look at the commutation relations in @xmath31 . for this purpose
we first generalize known results on the description of first terms in the comupltiplication via technical maps @xmath549 , that are of utmost importance .
then we apply this knowledge to show that for all @xmath550 the character - shifted root vectors @xmath551_{\phi^-}$ ] q - commute with all character - shifted root vectors in @xmath552)_{\phi^+}$ ] except @xmath553 .
+ for calculating the commutators we utilize the maps @xmath554 and @xmath555 due to lusztig and the following definition and lemmata can be found in @xcite chapter 6 .
the aim of the subsection is to generalize lemma [ lemmaralpha1 ] , which describes the first nontrivial term ( containing a simple root vector ) in the coproduct of any element , stepwise until lemma [ ralphaallgemein ] , which gives the same information for certain non - simple root vectors .
last we apply lemma [ rist0 ] , which describes @xmath556)$ ] as zeroes of @xmath554 to the specific situation of a non - simple root vector in lemma [ rist0neu ] .
[ ralpha ] let @xmath557 . for all @xmath157 there exist elements @xmath558 and @xmath559 in @xmath560 such that : @xmath561 where ( rest ) contains terms in @xmath562 for @xmath563 and @xmath564 in resp . @xmath565 and @xmath566 in .
in particular @xmath567 and @xmath568 for all @xmath303 .
[ lemmaralpha1 ] for these @xmath554 and @xmath569 the following relations hold : + a ) for all @xmath557 and @xmath570 : @xmath571 b ) for all @xmath557 and @xmath572 : @xmath573 c ) we have @xmath574 for all @xmath557 with @xmath575 the cartan involution [ lemmaralpha2 ] let @xmath157 and @xmath576 .
then for all @xmath577 and @xmath557 : @xmath578 see @xcite chapter 6 .
[ rist0 ] for @xmath157 and with @xmath335 a longest element : @xmath579)=\{x\in u^+\mid r_{\alpha}(x)=0\}\\ u^+[s_{\alpha}w_0]=\{x\in u^+\mid r'_{\alpha}(x)=0\ } \end{aligned}\ ] ] where @xmath580 are the lusztig automorphisms .
first we prove some generalizations of the lemmata [ lemmaralpha1 ] and [ lemmaralpha2 ] : [ ralphaeigenschaft ] let @xmath139 and @xmath555 as in definition [ ralpha ] , then we get for any @xmath581 in degree @xmath582 and any @xmath583 : @xmath584 for the constant @xmath585_{\alpha}\in k$ ] .
we prove the claim by induction over @xmath79 . from lemma [ lemmaralpha1 ]
we get @xmath586 for @xmath587 and @xmath588 .
this proves the claim for @xmath386 . from the induction assertion
follows easily : @xmath589 that is , it is sufficient to show that : @xmath590 .
this holds if and only if @xmath591 .
so it only remains to prove : @xmath592_{\alpha}=q_{\alpha}^{1-(i-1)}[i-1]_{\alpha}q_{\alpha}^{-2}+1 $ ] which works as follows : @xmath593_{\alpha}-1-q_{\alpha}^{1-(i-1)}[i-1]_{\alpha}q_{\alpha}^{-2}\\ = & q_{\alpha}^{1-i}\frac{q_{\alpha}^{i}-q_{\alpha}^{-i}}{q_{\alpha}-q_{\alpha}^{-1}}-1-q_{\alpha}^{1-(i-1)}\frac{q_{\alpha}^{i-1}-q_{\alpha}^{1-i}}{q_{\alpha}-q_{\alpha}^{-1}}q_{\alpha}^{-2}\\ = & ( q_{\alpha}^{1-i}(q_{\alpha}^{i}-q_{\alpha}^{-i})-q_{\alpha}+q_{\alpha}^{-1}-q_{\alpha}^{-i}(q_{\alpha}^{i-1}-q_{\alpha}^{1-i}))\frac{1}{q_{\alpha}-q_{\alpha}^{-1}}\\ = & ( q_{\alpha}-q_{\alpha}^{1 - 2i}-q_{\alpha}+q_{\alpha}^{-1}-q_{\alpha}^{-1}+q_{\alpha}^{1 - 2i})\frac{1}{q_{\alpha}-q_{\alpha}^{-1}}\\ = & 0 \end{aligned}\ ] ] this proves the lemma . [ alphahochi ] for @xmath157 let @xmath555 be the element as in definition [ ralpha ] and @xmath581 in degree @xmath582 then : @xmath594 where @xmath595 contains terms in @xmath596 for any @xmath597 which is not a multiple of @xmath148 .
+ here @xmath598 for the constant @xmath599}\in k$ ] .
both sides are @xmath296-linear .
we prove the statement inductively over the height @xmath308 of @xmath600 .
we look at words in the generators @xmath601 . for @xmath309 , i.e. @xmath602 for an @xmath157
, the claim is trivial .
assume the claim is true for elements @xmath603 , then we prove the claim for @xmath604 for an arbitrary @xmath303 : @xmath605 where @xmath606 contains terms of @xmath595 and @xmath607 .
we want to prove , that : @xmath608 let us first consider the case @xmath609 , here @xmath610 it remains to prove that @xmath611 this holds if and only if @xmath612 . the latter ,
however , follows from the corresponding property of @xmath613 .
+ now we consider the case @xmath614 . here : @xmath615 .
so we can rewrite the left hand side of equation as : @xmath616 comparing this to the right hand side of equation , it remains to prove that for all @xmath79 : @xmath617 this in turn follows directly from the definition of @xmath618 and lemma [ ralphaeigenschaft ] .
[ ralphakommutieren ] for all @xmath555 and @xmath554 as in definition [ ralpha ] the following equations hold : + a ) for all @xmath619 : @xmath620 b ) for @xmath619 with either @xmath621 or @xmath622 : @xmath623 c ) for @xmath619 with either @xmath621 or @xmath622 : @xmath624 we prove the statements again by induction on @xmath308 for elements @xmath625 for @xmath582 : + a ) for @xmath203 both sides are @xmath81 .
consider now for an arbitrary @xmath626 , @xmath582 the product @xmath627 : @xmath628 on the other side : @xmath629 \b ) here for @xmath203 both sides are @xmath81 .
for @xmath630 and arbitrary @xmath619 we consider first @xmath631 : @xmath632 on the other side we get for @xmath633 : @xmath634 by the induction hypothesis , the two sides are equal , if : @xmath635 this is true if and only if @xmath621 or @xmath636 .
c ) the proof is analogous to b ) .
[ ralphabar ] we expand the definition of @xmath554 to @xmath637 for elements : @xmath638 with @xmath639 and pairwise orthogonal roots @xmath640 as follows : let @xmath641 be a homogeneous element in @xmath642 . then we define : @xmath643 due to lemma [ ralphakommutieren ] this is well defined .
moreover let @xmath644 .
now we get for arbitrary @xmath645 for @xmath646 and @xmath647 for @xmath626 : @xmath648 we prove the statement with induction on @xmath649 . for @xmath650
the claim follows from lemma [ ralphaeigenschaft ] .
consider a root as above @xmath651 , then for an arbitrary @xmath652 with @xmath653 in particular we get @xmath654 .
we call @xmath655 , so @xmath656 and for the root @xmath657 the claim follows from the induction hypothesis . @xmath658 thus the claim is proven .
now we can prove the strongest generalization of the technical lemma [ lemmaralpha1 ] .
[ ralphaallgemein ] let @xmath638 and @xmath659 as above and @xmath660 for the elements @xmath661 from lemma [ alphahochi ] .
then we get for @xmath603 for a root @xmath582 : @xmath662 where @xmath595 contains terms in @xmath663 , such that @xmath664 is not a linear combination of pairwise orthogonal roots @xmath665 .
the proof works similar to the proof of lemma [ alphahochi ] together with lemma [ ralphabar ] .
having finished this we close this technical subsection by applying lemma [ rist0 ] : [ rist0neu ] let @xmath537 be the longest element of the weyl group with a reduced expression @xmath666 .
let @xmath667 with respect to this reduced expression . for an @xmath157 let @xmath668 .
for the root vectors @xmath37 the following relations hold : + a ) for @xmath669 : @xmath670 b ) for @xmath671 : @xmath672 according to lemma [ rist0 ] we know : @xmath579)=\{x\in u^+\mid r_{\alpha}(x)=0\}\\ u^+[s_{\alpha}w_0]=\{x\in u^+\mid r'_{\alpha}(x)=0\ } \end{aligned}\ ] ] a ) for @xmath669 in particular @xmath37 lies in @xmath673 $ ] , thus @xmath674 .
+ b ) for @xmath671 in particular @xmath37 lies in @xmath556)$ ] , thus @xmath675 .
we use the results of the preceding subsection to calculate commutator relations of character - shifted root vectors .
[ charakteralssumme ] let @xmath532 be a character on a right coideal subalgebra @xmath128 of @xmath642 and @xmath676 be the character - shifted subalgebra .
we define the character - shifted elements of @xmath676 as follows : @xmath677 let @xmath532 be a character whose support contains only pairwise orthogonal roots in @xmath678 , i.e. @xmath679 with @xmath680 .
we define @xmath681 as above and @xmath682 .
then : @xmath683 from lemma [ ralphaallgemein ] this follows directl , because @xmath595 is zero .
[ kommutatorinc ] given @xmath684)_{\phi^+}t_lu^-[x]_{\phi^-}\subset u$ ] as from in the beginning , i.e. for @xmath287 a weyl group element such that @xmath685 consists of pairwise orthogonal roots .
further let @xmath540 be the characters with @xmath541 defined as follows : @xmath542 for @xmath543 and @xmath81 otherwise and @xmath544 and @xmath81 otherwise , with @xmath686 . +
then for any @xmath543 and @xmath687 for elements @xmath688)_{\phi^+})_{\mu}$ ] and @xmath689_{\phi^-})_{\alpha}$ ] the following q - commutator relation holds : @xmath690 consider an element @xmath691 . then : @xmath692 for @xmath693 .
that is : @xmath694=s_{\bar{\alpha}+\alpha}(x_{\mu})(q_{\alpha}^{-i-1}\frac{q_{\alpha}^{2i}-1}{q_{\alpha}-q_{\alpha}^{-1}})\ ] ] inserting [ charakteralssumme ] in , yields : @xmath695_{q^{-(\mu,\alpha)}}=\sum_{0\preccurlyeq \bar{\alpha}}\lambda_{\bar{\alpha}}([f_{\alpha},s_{\bar{\alpha}}(x_{\mu})k_{\mu}^{-1}]_{q^{-(\mu,\alpha)}}+ \lambda_{\alpha}'[k_{\alpha}^{-1},s_{\bar{\alpha}}(x_{\mu})k_{\mu}^{-1}]_{q^{-(\mu,\alpha ) } } ) \end{aligned}\ ] ] for it remains to prove , that : @xmath696_{q^{-(\mu,\alpha)}}=-\sum_{0\preccurlyeq \bar{\alpha}}\lambda_{\bar{\alpha}}\lambda_{\alpha}'[k_{\alpha}^{-1},s_{\bar{\alpha}}(x_{\mu})k_{\mu}^{-1}]_{q^{-(\mu,\alpha)}}\end{aligned}\ ] ] to see this we use lemma [ lemmaralpha2 ] , this yields for @xmath157 and @xmath557 : @xmath697 consider first the left hand side of ( [ zuzeigen ] ) : @xmath698_{q^{-(\mu,\alpha ) } } = & \sum_{0\preccurlyeq \bar{\alpha}}\lambda_{\bar{\alpha}}(f_{\alpha}s_{\bar{\alpha}}(x_{\mu})k_{\mu}^{-1}-q^{-(\mu,\alpha)}s_{\bar{\alpha}}(x_{\mu})k_{\mu}^{-1}f_{\alpha})\\ = & \sum_{0\preccurlyeq \bar{\alpha}}\lambda_{\bar{\alpha}}(f_{\alpha}s_{\bar{\alpha}}(x_{\mu})-s_{\bar{\alpha}}(x_{\mu})f_{\alpha})k_{\mu}^{-1}\\ \overset{(\ref{verwendelemma})}{=}&-\sum_{0\preccurlyeq \bar{\alpha}}\lambda_{\bar{\alpha}}(q_{\alpha}-q_{\alpha}^{-1})^{-1}(r_{\alpha}(s_{\bar{\alpha}}(x_{\mu}))k_{\alpha}-k_{\alpha}^{-1}r_{\alpha}'(s_{\bar{\alpha}}(x_{\mu})))k_{\mu}^{-1}\\ \overset{lm.\ref{rist0neu}}{=}&\sum_{0\preccurlyeq \bar{\alpha}}\lambda_{\bar{\alpha}}(q_{\alpha}-q_{\alpha}^{-1})^{-1}q^{-(\alpha,\mu-\bar{\alpha}-\alpha)}r_{\alpha}'(s_{\bar{\alpha}}(x_{\mu}))k_{\mu}^{-1}k_{\alpha}^{-1}\\ = & \sum_{0\preccurlyeq \bar{\alpha}}\lambda_{\bar{\alpha}}s_{\bar{\alpha}+\alpha}(x_{\mu})k_{\mu}^{-1}k_{\alpha}^{-1}q_{\alpha}^{-i-1}\frac{q_{\alpha}^{2i}-1}{q_{\alpha}-q_{\alpha}^{-1}}(q_{\alpha}-q_{\alpha}^{-1})^{-1}q^{-(\alpha,\mu)}q_{\alpha}^i\\ = & \sum_{\alpha\preccurlyeq \bar{\alpha}}\lambda_{\bar{\alpha}}\lambda_{\alpha}^{-1}s_{\bar{\alpha}}(x_{\mu})k_{\mu}^{-1}k_{\alpha}^{-1}q_{\alpha}^{-1}\frac{q_{\alpha}^{2i-2}-1}{q_{\alpha}-q_{\alpha}^{-1}}(q_{\alpha}-q_{\alpha}^{-1})^{-1}q^{-(\alpha,\mu ) } \end{aligned}\ ] ] on the right hand side of the equation ( [ zuzeigen ] ) we have : @xmath699_{q^{-(\mu,\alpha)}}=&-\sum_{0\preccurlyeq \bar{\alpha}}\lambda_{\bar{\alpha}}\lambda_{\alpha}'(k_{\alpha}^{-1}s_{\bar{\alpha}}(x_{\mu})k_{\mu}^{-1}-q^{-(\mu,\alpha)}s_{\bar{\alpha}}(x_{\mu})k_{\mu}^{-1}k_{\alpha}^{-1}\\ = & -\sum_{\alpha\preccurlyeq \bar{\alpha}}\lambda_{\bar{\alpha}}\lambda_{\alpha}'s_{\bar{\alpha}}(x_{\mu})k_{\mu}^{-1}k_{\alpha}^{-1}q^{-(\mu,\alpha)}(q_{\alpha}^{2i-2}-1)\end{aligned}\ ] ] comparing both sides we get : @xmath700 it remains to prove that : @xmath701 rearranging yields : @xmath545 thus the claim is proven .
[ cistalgebra ] given @xmath684)_{\phi^+}t_lu^-[x]_{\phi^-}\subset u$ ] as from the beginning , i.e. for @xmath287 a weyl group element such that @xmath685 consists of pairwise orthogonal roots .
further let @xmath540 be characters with @xmath541 defined as follows : @xmath542 for @xmath543 and @xmath81 otherwise and @xmath544 and @xmath81 otherwise , with @xmath686 and let @xmath702 .
+ then @xmath31 is a triangular right coideal subalgebra , i.e. closed under multiplication .
this follows directly from the commutator relation of lemma [ kommutatorinc ] : @xmath703=0 \end{aligned}\ ] ] for all @xmath543 , @xmath704 . on the other hand
we know : @xmath705=[\bar{e}_{\alpha},\bar{f}_{\beta}]=[\bar{f}_{\alpha},\bar{f}_{\beta}]=0,\\ [ \bar{e}_{\alpha},\bar{f}_{\alpha}]=1\frac{q_{\alpha}^2}{q_{\alpha}-q_{\alpha}^{-1 } } \end{aligned}\ ] ] for all @xmath706 and of course @xmath707\subset \psi(u^+[w_0])_{\phi^+}\qquad [ \bar{e}_{\mu},k_{\gamma}]=0\end{aligned}\ ] ] for all @xmath708 and @xmath709 with @xmath710 .
so the ordered product of subalgebras @xmath31 is closed under multiplication and therefore a subalgebra . as a product of right coideals , it is in particular also a right coideal subalgebra .
[ maxiwurzelvektor ] let @xmath711)_{\phi^+}t_lu^-[x]_{\phi^-}$ ] be the right coideal subalgebra as from the beginning .
if @xmath712 , then for a maximal root @xmath30 the root vector @xmath713)_{\phi^+}$ ] we obtained in lemma @xcite lemma 3.36 q - commutes with all root vectors in @xmath31 .
due to lemma [ kommutatorinc ] @xmath183 q - commutes with all root vectors in @xmath714_{\phi^-}$ ] . from lemma
@xcite lemma 3.36 we know , that @xmath715 q - commutes with all root vectors in @xmath642 . as @xmath716\cong u^+[w_0]_{\phi}$ ] for characters
@xmath532 on @xmath716 $ ] , we already know , that all root vectors in @xmath552)_{\phi^+}$ ] q - commute with @xmath183 and thus the claim is proven .
recall how we have proven in @xcite that the standard borel subalgebras are indeed ede : [ nilpotentede ] given a weyl group element @xmath28 and the corresponding right coideal subalgebra @xmath717 $ ] .
each finite dimensional irreducible representation @xmath68 , on which all elements in @xmath717\cap ker\epsilon$ ] act nilpotent is one dimensional .
+ in particular @xmath717 $ ] is always weak ede .
the same is true for @xmath36 $ ] .
[ u+ede ] given a weyl group element @xmath28 , @xmath93 a subgroup of @xmath21 and the corresponding right coideal subalgebra @xmath718 $ ] with the property , that for all @xmath719 there exists a @xmath720 such that @xmath721 , then @xmath31 is ede .
+ in particular @xmath722 $ ] is ede , as well as the right coideal subalgebra @xmath723 $ ] .
from now on let @xmath711)_{\phi^+}t_lu^-[x]_{\phi^-}$ ] be as in the previous subsections , i.e. let @xmath287 be the weyl group element such that @xmath685 consists of pairwise orthogonal roots . further let @xmath540 be characters with @xmath541 , and @xmath724 and @xmath725 .
[ cede ] for @xmath684)_{\phi^+}t_lu^-[x]_{\phi^-}$ ] any finite - dimensional irreducible representation is one - dimensional .
let @xmath68 be a finite irreducible representation of @xmath31 . as the roots in @xmath193 are pairwise orthogonal and as @xmath726 , then due to lemma [ kommutatorinc ] the following commutator relations hold : @xmath727=[\bar{e}_{\alpha},k_{\mu}]=[\bar{e}_{\alpha},\bar{f}_{\beta}]=[\bar{f}_{\alpha},\bar{f}_{\beta}]=0\ ] ] for all @xmath706 and @xmath30 with @xmath728 .
we consider the reduced expression of the longest element : @xmath729 let @xmath730 be the subalgebra generated by the elements @xmath37 for @xmath731 . from the maximal choice
@xmath732 we get for any @xmath733 an element @xmath734 with @xmath735 .
@xmath730 is as algebra isomorphic to @xmath736 $ ] , so we can apply corollary [ u+ede ] and show , that @xmath730 is ede and we can thus prove the existence of a vector @xmath250 which is annihilated by @xmath730 .
let @xmath737 be the annihilator of @xmath730 in @xmath68 , i.e. @xmath738 .
we claim that @xmath737 is a @xmath31-submodul of @xmath68 : we know from theorem [ kommutatorinc ] that all elements @xmath739 q - commute with all elements in @xmath714_{\phi^-}$ ] . for degree reasons we know that the q - commutator of the elements in @xmath730 and @xmath740)_{\phi^+}$ ] lies in @xmath730 , i.e. vanishes on the representation @xmath737 . thus with theorem [ maxials0allgemein ] @xmath737 is a @xmath31-subrepresentation , and as @xmath68 is irreducible we get @xmath741 .
thus we get for the irreducible representation @xmath68 of @xmath31 : @xmath742 so the ede property follows from the ede property of the components of the tensor product .
@xmath743 is abelian , and for any @xmath744 the subalgebras @xmath745 are quantised weyl algebras , which are ede by example [ weyl algebra ] . from lemma [ cede ] we already know ,
that @xmath31 is ede , it remains to prove the maximality of @xmath31 .
let @xmath730 be an ede right coideal subalgebra with @xmath746 .
+ due to the second authors generating system for any rcs in @xcite lemma 4.11 or @xcite we can choose a generating system of @xmath730 consisting of elements , whose e - leitterms lie in @xmath747.as we know that @xmath552)_{\phi^+}\subset c\subset c'$ ] , all elements with non - trivial e - leitterm lie in @xmath747 : that is , to each additional element @xmath85 in @xmath730 we find an element @xmath86 in @xmath748 , such that @xmath85 and @xmath86 have the same e - leitterms .
thus the difference has a smaller degree and can be generated as in @xcite lemma 4.7 by elements with e - leitterms in @xmath747 , which again have smaller degree as before . by induction it follows that all generators as in @xcite lie either in @xmath747 or @xmath749 and the right coideal subalgebra @xmath730 is triangular generated , i.e. of the form @xmath750)_{\phi^+_{neu}}t_{l'}u^-[xv]_{\phi^-_{neu}}$ ] for a @xmath751 with @xmath752 and characters @xmath753 and @xmath754 .
+ as @xmath755 , we can consider @xmath753 , @xmath754 such that all elements in @xmath730 are ad-@xmath743-stable .
as @xmath730 is ede we can conclude : @xmath756 . on the other hand theorem [ ef ] yields , that for all elements @xmath757 holds @xmath758 .
thus we get @xmath759 . as @xmath760 and of course
@xmath761 , so @xmath762 follows directly .
+ in the case @xmath763 the claim follows from lemma [ annichtede ] , in general we can prove it as follows : assume @xmath764 , then there exists @xmath765 .
is @xmath148 orthogonal to all elements in @xmath193 , then @xmath766 and corollary [ efohnek ] yields a contradiction to @xmath730 ede .
is @xmath148 not orthogonal on all elements in @xmath193 , then there exists a root @xmath767 and a term @xmath768 for a constant @xmath769 and terms @xmath770 with @xmath771 : this claim follows from the fact that the elements in @xmath193 are pairwise orthogonal and simple and in the relevant cases @xmath772 , @xmath773 resp .
@xmath774 resp .
@xmath775 one can calculate the coproduct of @xmath160 directly .
+ on the other hand we know @xmath776 . considering the action of the @xmath13-commutator on the lowest weight representation @xmath151 to a lowest weight @xmath149 with @xmath777
, we find : @xmath778.v_{\lambda}&=\bar{f}_{\mu}e_{\alpha}k_{\alpha}^{-1}.v_{\lambda}\\ & = cf_{\alpha}k_{\mu-\alpha}^{-1}e_{\alpha}k_{\alpha}^{-1}.v_{\lambda}+\sum_{\nu\succ \alpha}x_{\nu}e_{\alpha}k_{\alpha}^{-1}.v_{\lambda}\\ & = \frac{cq^{-(\mu-\alpha,\alpha)}}{q_{\alpha}-q_{\alpha}^{-1}}k_{\mu}^{-1}(k_{\alpha}-k_{\alpha}^{-1}).v_{\lambda}=c'v_{\lambda } \end{aligned}\ ] ] for a @xmath779 by choice of the smallest weight @xmath149 .
it follows that the commutator has a non - trivial eigenvalue in the finite - dimensional representation @xmath151 and thus not any finite - dimensional irreducible representation of @xmath730 can be one - dimensional .
so @xmath780 and @xmath781 , i.e. @xmath31 is maximal and thus a borel subalgebra .
a triangular right coideal subalgebra is always of the form : @xmath163)_{\phi^+}t_lu^-[w^-]_{\phi^-}$ ] , where @xmath782 are elements in the weyl group ; be advised that not any choice of @xmath783 conversely leads to a well defined algebra . by definition of the character - shift it makes only sense to consider @xmath784 ,
see theorem [ hk11b ] . in the previous subsection we have constructed borel subalgebras with an additional non - degeneracy property @xmath785 and @xmath103 .
now we conversely prove in the case of @xmath0 that we have found indeed all borel subalgebras with these properties .
we use in particular theorem [ weyltheorem ] on weyl groups , which tells us that an arbitrary triangular ede rcs is contained in a larger triangular ede rcs which is then isomorphic via some @xmath27 to one of the borel subalgebras we constructed .
it is this theorem on weyl groups which we generally conjecture , but here could only be proven for @xmath0 .
[ klassifikationtyp2 ] assume type @xmath0 .
every triangular borel subalgebra of @xmath2 @xmath786)_{\phi^+}t_lu^-[w^-]_{\phi^-}\ ] ] with the additional non - degeneracy property @xmath787 is isomorphic as algebra to a borel subalgebra of the form constructed in our main theorem [ konstruktionborel ] . +
that is : @xmath788})_{\phi+}t_lu^-[\prod_{\alpha\in c}s_{\alpha}]_{\phi^-}$]with @xmath789 , where @xmath531 is a set of pairwise orthogonal simple roots and @xmath790 , @xmath791 are characters on @xmath792 ) \text { resp . }
u^-[\prod_{\alpha\in c}s_{\alpha}]$ ] , given by an arbitrary choice of the values @xmath793 for @xmath794 with @xmath795 . in the rest of the section we will give a proof of this classification result .
[ triangerweitert ] let @xmath31 be a triangular ede right coideal subalgebra of the form @xmath173)_{\phi^+}t_lu^-[w_2]_{\phi^-}$ ] with @xmath796 .
then there are elements @xmath264 with @xmath265 , @xmath266 , such that @xmath797 and @xmath262 . for these elements @xmath261 and @xmath798 we get : is a product of coideals , but at this point not necessarily an algebra . in our specific situation
this follows from the next lemma ] @xmath799)_{\phi^+}t_{l'}u^-[w_2']_{\phi^-}=:c'\ ] ] the existence of such weyl group elements @xmath261 is the purpose of theorem [ weyltheorem ] .
of course we have @xmath800\subset u^+[w_1']$ ] and @xmath801\subset u^+[w_2']$ ] .
so it remains to prove , that @xmath790 is a character on @xmath802 $ ] . from theorem [ weyltheorem ] we already know , that @xmath394 has a reduced expression of the form @xmath803 , such that there is a @xmath804 with @xmath805 .
due to remark [ charakterdurchstreichen ] @xmath790 is thus a character on @xmath802 $ ] .
the same holds for @xmath806 $ ] .
thus the inclusion @xmath807)_{\phi^+}t_{l'}u^-[w_2']_{\phi^-}$ ] holds .
now we take a closer look at @xmath808)_{\phi^+}t_{l'}u^-[w_2']_{\phi^-}$ ] . in particular
we want to show , that it is a triangular right coideal subalgebra and ede . for this
in particular we use theorem [ weyltheorem ] and the lusztig automorphisms .
we already know that for @xmath441 there is a reduced expression @xmath809 , such that @xmath810 .
we define : @xmath811 and @xmath812 , such that @xmath813 and from the choice of @xmath394 we get @xmath814 , as @xmath262 , moreover @xmath815 consists of pairwise orthogonal simple roots , by definition of @xmath322 .
+ in the following lemma we now prove that in this situation we can apply lusztigs automorphism @xmath816 to the chracter - shifted rcs ( in general this does not give a character - shifted rcs ! ) and show that it has the intended result claimed in the classification : [ lustiganwenden ] let @xmath31 be a triangular ede right coideal subalgebra of the form @xmath173)_{\phi^+}t_lu^-[w_2]_{\phi^-}$ ] with @xmath796 . given the corresponding @xmath394 , @xmath813 with @xmath817 and @xmath641 as above , then the lusztigautomorphism @xmath816 maps the coideal @xmath730 constructed in the previous lemma : @xmath818)_{\phi^+}t_{l'}u^-[v^{-1}x]_{\phi^-}}_{c'})=\psi(u^+[vw_1'])_{t_v(\phi^+)}t_v(t_{l'})(u^-[x])_{t_v(\phi^- ) } \end{aligned}\ ] ] where @xmath819 is defined as @xmath820 .
let us consider the pbw generators of @xmath31 : as @xmath31 is triangular , all of them lie in @xmath821)_{\phi^+},u^-[v^{-1}x]_{\phi^-}$ ] and @xmath743 .
let s consider first @xmath822_{\phi^-}$ ] .
due to the choice of the reduced expression of @xmath823 , the basis elements of @xmath822_{\phi^-}$ ] have the following form : @xmath824 thus we get for the lusztig automorphism @xmath816 : @xmath825 the same is true for the basis elements of @xmath802_{\phi^+}$ ] .
+ with these considerations , we can argue analogously to non - character - shifted right coideal subalgebras in @xcite and obtain the assertion : as @xmath262 and @xmath797 we get @xmath826 , so : @xmath827)_{\phi^+}t_{l'}u^-[v^{-1}x]_{\phi^-})= & t_v(\psi(u^+[w_1'])_{\phi^+})t_v(t_{l'})t_v(u^-[v^{-1}]_{\phi^-}t_v^{-1}(u^-[x]_{t_v(\phi^-)}))\\ = & \psi(u^+[v])_{t_v({\phi^-})}t_v(\psi(u^+[w_1'])_{\phi^+}t_v(t_{l'})t_vu^-[x]_{t_v(\phi^-)}\\ = & \psi(u^+[vw_1'])_{t_v(\phi^+)}t_v(t_{l'})(u^-[x])_{t_v(\phi^-)}\\ = & \psi(u^+[w_0])_{t_v(\phi^+)}t_v(t_{l'})(u^-[x])_{t_v(\phi^-)}\end{aligned}\ ] ] the lusztig autmorphism @xmath27 is in general not an coalgebra homomorphism , but in this special case @xmath816 is an algebra homomorphism , sending a right coideal to a right coideal .
the image @xmath828 in the previous lemma is one of the borel subalgebras we constructed in our main theorem [ konstruktionborel ] ; regarding the characters : the relation between @xmath790 and @xmath791 has to be as asserted in order for @xmath730 and even @xmath31 to be ede by lemma [ lemma3.16 ] .
now also @xmath730 is an ede right coideal subalgebra , because @xmath816 is an algebra homomorphism , so in order for @xmath31 to be maximal we have at least @xmath829 .
+ the remaining problem is : a triangular ede right coideal subalgebra of the form @xmath173)_{\phi^+}t_{l'}u^-[w_2]_{\phi^-}$ ] which is via @xmath816 isomorphic to the above - constructed triangular borel subalgebra , a - priori itself does not have to be maximal , as @xmath31 could lie in a bigger ede right coideal subalgebra @xmath830 , whose reflections to @xmath831 is not a right coideal subalgebra any more .
+ however in @xmath0 we know from lemma [ annichtede ] and the maximality of @xmath93 , that @xmath31 can not lie in a bigger triangular ede right coideal subalgebra . +
this concludes the proof of the main classification theorem [ klassifikationtyp2 ] in the main theorem [ konstruktionborel ] we have constructed a large family of borel algebras @xmath832)_{\phi^+}t_lu^-[x]_{\phi^-}\ ] ] where @xmath539 such that @xmath538 consists of pairwise orthogonal simple roots , and @xmath833 , and the characters @xmath540 on @xmath834 are given as usual by suitable values for @xmath835 .
@xmath836 now let @xmath837 be a one - dimensional representation of @xmath122 . since there is a quotient algebra where all @xmath838 for all @xmath839 , the one - dimensional representations are in bijection to one - dimensional representations of the @xmath840-fold quantum weyl algebra generated by the character - shifted simple root vectors @xmath841 for all @xmath842 .
by example [ weyl algebra ] hence any one - dimensional representation @xmath837 is given though scalars @xmath843 for all @xmath835 with again @xmath844 .
it is then clear that : + let @xmath122 be a borel algebra and @xmath68 a one - dimensional representation as above .
then the induced representation @xmath845 is as a graded vectorspace isomorphic to @xmath846\right ) \otimes \left(\bigotimes_{\alpha\in{\mathtt{supp } } } { \mathbb{c}}[k_\alpha , k_\alpha^{-1}]\right)\ ] ] the first factor is the space of coinvariants under projection to the borel part @xmath847 where here @xmath848 is of type @xmath849 ; in fact this is a nichols algebra in the non - semisimple category of @xmath850-modules .
the second factor makes the induced module for @xmath851 non - diagonal in @xmath852 , which acts simply by left - multiplication .
similar to the @xmath853-case in @xcite we may ask for the decomposition behaviour of these modules . we expect that they are again largely irreducible up to discrete series in some @xmath843 . in particular if all @xmath843 are of this form , we expect again all finite - dimensional irreducible modules of @xmath2 as unique quotients .
now we construct and study borel subalgebras without the non - degeneracy property , so : @xmath854 these are not isomorphic as algebra to those of the non degenerated types .
these borel subalgebras can contain non trivial character - shifted root vectors @xmath183 , @xmath184 even though @xmath855 , if @xmath178 is not simple and there is a smaller @xmath856 with @xmath857 . the properly character - shifted root vectors
@xmath858 generate again a quantized weyl algebra and the entire borel subalgebra contains extensions of weyl algebras .
+ in this section we restrict to the case @xmath0 .
we want to give some preliminary classification results for general borel subalgebras @xmath163)_{\phi^+}t_lu^-[w^-]_{\phi^-}$ ] of the quantum group of type @xmath0 with @xmath859 . to prove the expected result we use two technical restrictions ,
of which we conjecture that they are true for general borel subalgebras : @xmath860 @xmath861 under these assumptions , we can show that every triangular borel subalgebra is of the following type .
[ w ] in @xmath0 we consider the following types of weyl group elements : * for @xmath862 a _ ladder _ @xmath863 is the following weyl group element together with a reduced expression : @xmath864 for the associated roots in @xmath865 we use the notation @xmath866 * for @xmath867 and @xmath868 , @xmath869 a @xmath870 is the following weyl group element together with a reduced expression : @xmath871 of course @xmath872 .
+ * for @xmath867 and @xmath873 a _ diamond _
@xmath874 is the following weyl group element together with a reduced expression : @xmath875 * for @xmath867 a _ palm _ @xmath876{90}{$\ll$}}}_i$ ] is the following weyl group element together with a reduced expression : @xmath877{90}{$\ll$}}}_i:={\mathtt{v}}_i^{l_1k_1}{\mathtt{v}}_i^{l_2k_2}\ldots\ ] ] with @xmath878 , @xmath879 and the property @xmath880 and @xmath881 .
+ [ wiewaussehenkann ] for @xmath163)_{\phi^+}t_lu^-[w]_{\phi^-}$ ] ede with @xmath882 and if @xmath220 only the following choices of @xmath44 are possible : 1 . in the case
@xmath883 : @xmath884 for some @xmath885 , as in definition [ w ] .
2 . in the more general case
@xmath886 : @xmath887{90}{$\ll$}}}_i:={\mathtt{v}}_i^{l_1k_1}{\mathtt{v}}_i^{l_2k_2}\ldots\ ] ] for @xmath867 with the property @xmath880 and @xmath881 . +
this @xmath44 is the inverse of a special element in the following sense + [ nebenklassendef][@xcite ] for a weyl group element @xmath28 and a root @xmath888 the following statements are equivalent : * the set @xmath889 contains exactly one simple root , which is @xmath890 * for all @xmath891 holds @xmath892 , but @xmath893 * in each reduced expression of @xmath44 the last factor is @xmath100 + let @xmath894 be the ( so called parabolic ) subgroup of @xmath284 , which is generated by the @xmath895 with @xmath652 different to @xmath79 . then in the case above @xmath44 is the unique representative of the left coset @xmath896 with minimal length .
+ if @xmath897 is the maximal diamond with @xmath898{90}{$\ll$}}}_i)$ ] , then : @xmath899 3
. in the general case @xmath900 ( pairwise orthogonal ) holds : @xmath44 is the special element with @xmath901 .
that is @xmath44 is the generalisation of the inverse of a special element in the sense of lemma [ nebenklassendef ] .
+ if @xmath31 is even a borel subalgebra , then we conjecture due to the maximality : @xmath902{90}{$\ll$}}}_i\ ] ] for special elements @xmath876{90}{$\ll$}}}_i$ ] , which commute .
the support of the character @xmath98 is the union of the support of the respective palms . as the elements in @xmath98
are pairwise orthogonal the claims follows directly from [ ef ] .
the form of the support @xmath98 in 2 . follows directly from corollary [ folgerungnurdiemittezaehlt ] .
+ if @xmath31 is a borel subalgebra we get from lemma [ dasssiesichnichtschneiden ] , that only those combinations of @xmath876{90}{$\ll$}}}_i$ ] are possible which are disjoint , i.e. which commute . in the following we want to construct non - degenerate ede right coideal subalgebras which correspond to palms of height @xmath13 in the previous subsection , @xmath903{90}{$\ll$}}}_i={\mathtt{v}}_i^{lk}\ ] ] for some @xmath904 . due to theorem [ ef ] the ede property of
@xmath31 implies @xmath905 for @xmath906 and @xmath81 otherwise and @xmath907 and @xmath81 otherwise .
[ konstruktion8palme1 ] for type @xmath0 the following is an ede right coideal subalgebra : let @xmath867 and @xmath868 , @xmath908 be arbitrarily chosen , then we consider an arbitrary palm of height 1 @xmath909 which means : @xmath910 as above @xmath911)_{\phi^+}t_lu^-[{\mathtt{v}}_i^{lk}]_{\phi^-}\ ] ] for @xmath912 and characters @xmath790 and @xmath791 with @xmath913 .
more precisely @xmath31 has the following relations between two character - shifted root vectors for roots @xmath914 : @xmath915=\frac{q^2}{q - q^{-1}}\\ [ \bar{e}_{\mu},\bar{e}_{\nu}]=0\quad [ \bar{f}_{\mu},\bar{f}_{\nu}]=0\\ [ \bar{e}_{\mu},\bar{f}_{\nu}]=0\quad \text{if } \mu\neq\nu\\ [ \bar{e}_{\mu},\bar{f}_{\mu}]=q^2[\bar{e}_{\mu'},\bar{f}_{\mu'}]_1\quad \text { if } \mu\neq\alpha_i \end{aligned}\ ] ] @xmath916 the commutator relations follow from the explicit calculations in @xcite chapter 3 .
+ now we prove the ede property : let @xmath68 be an arbitrary finite dimensional representation of @xmath31 .
we consider the restriction to the right coideal subalgebra @xmath917 , which is a quantized weyl algebra as usual by the choice of characters .
we know from example [ weyl algebra ] , that on any finite dimensional representation of the weyl algebra the commutator @xmath918_1 $ ] vanishes . in particular , each finite - dimensional representation of the weyl algebra factorizes to a representation of the commutative algebra @xmath919/ ( ef-\frac{q^2}{(q - q^{-1})(1-q^2)})$ ] .
+ we now consider the next - largest subalgebra , which is generated by the weyl algebra and all @xmath920 with @xmath921 . due to the commutator relation
@xmath922=[\bar{e}_{\mu'},\bar{f}_{\mu'}]_1 $ ] we get that @xmath923_{q^{(\mu,\mu)}}$ ] acts trivially on @xmath68 . due to theorem [ maxials0allgemein ] ( and as @xmath924 ) this implies , that the elements @xmath920 , which @xmath26-commute with all elements on @xmath68 , act trivial .
+ consider now inductively the next larger subalgebra with @xmath920 for @xmath925 .
from the relation @xmath922=[\bar{e}_{\mu'},\bar{f}_{\mu'}]_1 $ ] and the just proven trivial action of @xmath926 it follows that the @xmath26-commutator acts trivial on @xmath68 .
inductively we know that all @xmath920 with @xmath927 act trivial on @xmath68 . + thus , the category of finite - dimensional representations of @xmath31 is equivalent to the category of finite - dimensional representation of the commutative algebra @xmath919/(ef-\frac{q^2}{(q - q^{-1})(1-q^2)})$ ] and , in particular , all irreducible finite - dimensional representations of @xmath928 are one - dimensional .
we now present all borel subalgebras of the examples @xmath10 , @xmath929 and all triangular borel subalgebras of @xmath930 , as discussed in the second authors work @xcite chapter 9 - 11 or @xcite .
we do this for one to illustrate the results of our present article . on the other hand
it is interesting to know all ( potentially non - triangular ) examples with a different method , this is how we arrive at our conjecture that all borel subalgebras are triangular .
+ the strategy to find all maximal ede rcs requires first to get a hold on all rcs : this is the theorem providing a generating system for an arbitrary rcs in @xcite chapter 4 .
then we disprove the ede property using restrictions of suitable minuscule vermamodules as in section 3 .
the standard borel subalgebras @xmath747 and @xmath749 and @xmath931 are the only homogeneous right coideal subalgebras of @xmath10 .
furthermore there are the right coideal subalgebras @xmath932 in @xmath747 and @xmath933 in @xmath749 , and families of character - shifted right coideal subalgebras @xmath934 resp .
@xmath935 for characters on @xmath932 resp .
@xmath933 given by @xmath936 and @xmath937 .
they have the form @xmath938 and @xmath939 . + as in example [ weyl algebra ] the right coideal subalgebra @xmath940 is ede and hence a borel subalgebra , iff @xmath941 note , that different borel subalgebras @xmath122 for different choices of @xmath65 are mapped onto each other via the hopf - automorphism @xmath66 . + together with the standard borel subalgebras these are all borel subalgebras of @xmath10 . for @xmath929 we found that all borel subalgebras are triangular .
besides the standard borel subalgebra there are two isomorphism classes of non - degenerate borel subalgebras as discussed in section [ sec_nondegenerate ] and one degenerate of height @xmath13 as discussed in section [ sec_degenerate ] .
the borel subalgebras @xmath942)_{\phi^+}t_lu^-[w^-]_{\phi^-}$ ] with @xmath943 are due to theorem [ klassifikationtyp2 ] up to symmetry isomorphic as algebra to @xmath552)_{\phi^+}\langle k_{2\beta+\alpha},k_{2\beta+\alpha}^{-1}\rangle u^-[s_{\alpha}]_{\phi^-}$ ] with @xmath944 and @xmath81 otherwise , and @xmath945 , such that @xmath946 .
more precisely there are up to symmetry exactly two such borel subalgebras .
these are @xmath947)_{\phi^+}\langle k_{2\beta+\alpha},k_{2\beta+\alpha}^{-1}\rangle u^-[s_{\alpha}]_{\phi^-}\ ] ] with characters as above and @xmath948)_{\phi^+}\langle k_{\alpha-\beta } , k_{\alpha-\beta}^{-1}\rangle u^-[s_{\beta}s_{\alpha}]_{\phi^-}\ ] ] with @xmath949 and @xmath950 , such that @xmath951 holds .
[ [ typ3insl3 ] ] a degenerate borel subalgebra @xmath952)_{\phi^+}\langle k_{2\beta+\alpha},k_{2\beta+\alpha}^{-1}\rangle u^-[s_{\alpha}s_{\beta}]_{\phi^-}$ ] ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ the third type of triangular borel subalgebras in @xmath929 is of the form @xmath948)_{\phi^+}\langle k_{2\beta+\alpha},k_{2\beta+\alpha}^{-1}\rangle u^-[s_{\alpha}s_{\beta}]_{\phi^-}\ ] ] with @xmath944 and @xmath81 otherwise , and @xmath945 and @xmath81 otherwise , such that @xmath946 as above .
the generating elements of this right coideal subalgebra are the following : @xmath953 the product of the constants @xmath954 resp .
@xmath955 in the terms @xmath956 resp .
@xmath957 is : @xmath958 for these @xmath959 and @xmath960 consider the commutator of the both elements : @xmath961_{q^2}\\ = ( f_{\alpha}+\lambda_{\alpha}^-k_{\alpha}^{-1})(e_{\alpha}k_{\alpha}^{-1}+\lambda_{\alpha}^+k_{\alpha}^{-1})(q^4-q^2)+\frac{q^4}{q - q^{-1}}1\end{aligned}\ ] ] thus the commutators of the generating elements in @xmath122 are given by : @xmath962_1=[k,\bar{f}_{\alpha}]_1=[k,\bar{e}_{\alpha\beta}]_1=[k,\bar{f}_{\beta\alpha}]_1=0\\ [ \bar{e}_{\alpha},\bar{f}_{\alpha}]_{q^2}=\frac{q^2}{q - q^{-1}}1\quad [ \bar{e}_{\alpha},\bar{e}_{\alpha\beta}]_q=[\bar{e}_{\alpha},\bar{f}_{\beta\alpha}]_q=0\label{esuntereinander}\\ [ \bar{f}_{\alpha},\bar{e}_{\alpha\beta}]_{q^{-1}}=[\bar{f}_{\alpha},\bar{f}_{\beta\alpha}]_{q^{-1}}=0\\ [ \bar{e}_{\alpha\beta},\bar{f}_{\beta\alpha}]_{q^2}=\bar{f}_{\alpha}\bar{e}_{\alpha}(q^4-q^2)+\frac{q^4}{q - q^{-1}}1\end{aligned}\ ] ] consider now a representation @xmath68 of @xmath122 .
we already know , that @xmath745 is ede .
as in the general case of lemma [ konstruktion8palme1 ] @xmath963_1 $ ] acts trivial on every finite dimensional @xmath68 .
so the term @xmath964 acts as @xmath965 on @xmath68 . inserting this in the q - commutator of @xmath956 and @xmath957
we see that @xmath966_{q^2}$ ] acts as @xmath81 on all of @xmath68 .
+ on the other hand @xmath553 has an eigenvector @xmath69 to the eigenvalue @xmath967 and due to the coproduct the elements @xmath968 are eigenvectors of @xmath553 to either eigenvalue @xmath81 or pairwise distinct eigenvalues . as @xmath68 is finite ,
we can find a vector @xmath44 with @xmath969 .
as on @xmath68 all elements @xmath26-commute with @xmath956 we can apply theorem [ maxials0allgemein ] to show , that @xmath956 acts as @xmath81 on any irreducible representation .
the same we can show for @xmath957 with the same argument . then @xmath970 and @xmath122 is ede . +
the maximality of this borel subalgebra is proven in @xcite by considering any extension of this rcs .
thus we have found all borel subalgebras of @xmath929 .
we now give all possible triangular borel subalgebras of @xmath930 .
we do not consider non - triangular rcs so we can also not prove the maximality of the degenerate examples we give below .
however we do prove that these are ede rcs , which are maximal among all triangular rcs , and there exist no other triangular ede borel subalgebras .
we conjecture that these are in fact all borel subalgebras of @xmath930 .
* @xmath552)_{\phi^+}\langle k_{2\alpha_2+\alpha_1},k_{2\alpha_2+\alpha_1}^{-1}\rangle u^-[s_{\alpha_i}]_{\phi^-}$ ] for @xmath972 with @xmath973 and @xmath81 otherwise and @xmath974 , such that @xmath975 .
* @xmath552)_{\phi^+}\langle k_{2\alpha_2+\alpha_1},k_{2\alpha_2+\alpha_1}^{-1}\rangle u^-[s_{\alpha_1}s_{\alpha_3}]_{\phi^-}$ ] with @xmath976 , @xmath977 and @xmath81 otherwise , and @xmath978 , such that @xmath979 , and @xmath980 , such that @xmath981 .
we list all possibilities for @xmath982 from theorem [ wiewaussehenkann ] .
they all contain the degenerate borel subalgebra from @xmath929 in the previous subsection . due to lemma [ annichtede ] we know , that for any triangular ede right coideal subalgebra @xmath983 where @xmath99 is calculated via @xmath44 and @xmath193 .
+ we want to prove now that the possible non - degenerate triangular right coideal subalgebras in @xmath15 are in fact all ede , i.e. @xmath163)_{\phi^+}t_lu^-[w^-]_{\phi^-}$ ] with @xmath984 and with @xmath108 and @xmath732 .
there are up to isomorphism and symmetry three such right coideal subalgebras with @xmath985 and two with @xmath986 .
+ in each of these cases , we manually compute the q - commutator relations of the generators of @xmath987 and show , as the case may be , that an arbitrary finite - dimensional irreducible representation @xmath988 of the algebra @xmath987 is one - dimensional .
the strategy for this is as in lemma [ konstruktion8palme1 ] ( general palm of height 1 ) : we find a contained quantized weyl algebra @xmath989 , of which we already know that it acts on finite - dimensional representations commutative , i.e. the commutator @xmath990_1 $ ] acts trivial .
thus , by means of the respective explicit list of commutators , we find an element @xmath991 which q - commutes ( on @xmath992 ) with all generators of @xmath928 . by theorem [ maxials0allgemein ]
then @xmath993 must act on every irreducible representation @xmath988 trivial , which inductively induces further commutators to vanish until all elements except the weyl algebras act trivial .
the details of the calculation differ from case to case , but the argument is in any case the same . here ,
apart from isomorphism and symmetry , there is exactly one type of maximal triangular and right - coideal subalgebra given by @xmath995)_{\phi^+}\langle k_{\alpha_3 } , k_{\alpha_2}^2k_{\alpha_1},k_{\alpha_3}^{-1},k_{\alpha_2}^{-2}k_{\alpha_1}^{-1}\rangle u^-[s_{\alpha_1}s_{\alpha_2}]_{\phi^-}$ ] with @xmath996 otherwise @xmath81 and @xmath997 , otherwise @xmath81 , such that @xmath998 .
the first quantized weyl algebra is here @xmath1001 , whose commutator again disappears on each finite dimensional representation
. then the four elements of height @xmath16 q - commute with all generators and thus act trivially .
modulo these relations @xmath1002 is another quantized weyl algebra commuting with the first quantized weyl algebra . + * the standard borel subalgebra @xmath747 * a borel subalgebra of non - degenerate type : * * @xmath552)_{\phi^+}\langle k_{2\alpha_2+\alpha_1},k_{2\alpha_2+\alpha_1}^{-1 } , k_{\alpha_3},k_{\alpha_3}^{-1}\rangle u^-[s_{\alpha_1}]_{\phi^-}$ ] with @xmath1003 * * @xmath552)_{\phi^+}\langle k_{2\alpha_2+\alpha_1},k_{2\alpha_2+\alpha_1}^{-1}\rangle u^-[s_{\alpha_1}s_{\alpha_3}]_{\phi^-}$ ] with @xmath1004 * a borel subalgebra of degenerate type , i.e. : * * @xmath1005)_{\phi^+}\langle k_{\alpha_3 } , k_{\alpha_2}^2k_{\alpha_1},k_{\alpha_3}^{-1},k_{\alpha_2}^{-2}k_{\alpha_1}^{-1}\rangle u^-[s_{\alpha_1}s_{\alpha_2}]_{\phi^-}$ ] with @xmath1003 * * @xmath1006)_{\phi^+}\langle k_{\alpha_3 } , k_{\alpha_2}^2k_{\alpha_1},k_{\alpha_3}^{-1},k_{\alpha_2}^{-2}k_{\alpha_1}^{-1}\rangle u^-[s_{\alpha_1}s_{\alpha_2}s_{\alpha_3}]_{\phi^-}$ ] with @xmath1003 * * @xmath1006)_{\phi^+}\langle k_{\alpha_3}k_{\alpha_2}^2k_{\alpha_1},k_{\alpha_3}^{-1}k_{\alpha_2}^{-2}k_{\alpha_1}^{-1}\rangle u^-[s_{\alpha_1}s_{\alpha_2}s_{\alpha_3}s_{\alpha_1}]_{\phi^-}$ ] with @xmath1004 * * @xmath1007)_{\phi^+}\langle k_{\alpha_1}^{-1}k_{\alpha_3 } , k_{\alpha_1+\alpha_2+\alpha_3},k_{\alpha_1}k_{\alpha_3}^{-1 } , k_{\alpha_1+\alpha_2+\alpha_3}^{-1}\rangle u^-[s_{\alpha_2}s_{\alpha_1}s_{\alpha_3}]_{\phi^-}$ ] with @xmath1008 * * @xmath1009)_{\phi^+}\langle k_{\alpha_1}^{-1}k_{\alpha_3},k_{\alpha_1}k_{\alpha_3}^{-1}\rangle u^-[s_{\alpha_2}s_{\alpha_1}s_{\alpha_3}s_{\alpha_2}]_{\phi^-}$ ] with @xmath1010 both authors thank istvan heckenberger for answering questions and thank the rtg 1670 of the university of hamburg for support and hospitality , in particular for a microworkshop on quantum symmetric pairs ( hamburg , february 2017 ) , where most of this work was completed . | we construct explicit families of right coideal subalgebras of quantum groups , where all irreducible representations are one - dimensional and which are maximal with this property .
we have previously called such a right coideal subalgebra a borel subalgebra . conversely we can prove that any tringular borel subalgebra fulfilling a certain non - degeneracy property is of the form we construct ; this classification requires a key assertion about weyl groups which we could only prove in type @xmath0 .
borel subalgebras are interesting for structural reasons , but also because the induced representations give interesting unfamiliar analoga of category @xmath1 .
zmp - hh / 17 - 9 + hamburger beitrge zur mathematik nr .
649 + february 2017 + * constructing new borel subalgebras of quantum groups with a non - degeneracy property * simon lentner + algebra and number theory , university hamburg , + bundesstrae 55 , d-20146 hamburg + karolina vocke + phillips university marburg , + hans - meerwein - strae , d-35032 marburg |
in spontaneous parametric down - conversion ( spdc ) , photon pairs are generated with several degrees of freedom quantum correlated .
in particular , the down - converted photons are spatially entangled . due to energy and momentum conservation , the sum of the transverse momenta and the difference of the transverse positions of the photons
can be well defined , even though the position and momentum of each photon are undefined @xcite .
this type of einstein - podolsky - rosen ( epr ) correlation @xcite , has been used as a resource for fundamental studies of quantum mechanics @xcite , for quantum imaging @xcite , and experiments of quantum information @xcite .
the process of spdc has been studied extensively in the past @xcite , and much of the recent effort has been put in the quantification of the spatial entanglement for a given experimental geometry .
traditionally , this has been done through the technique of schmidt decomposition of the two - photon wave function , which gives the schmidt number , @xmath2 , and the schmidt modes allowed for each photon @xcite .
while this approach describes important properties of the spatial correlation , it basically gives no information of the form of the spatial joint distributions , therefore giving no information about the gaussianity of the spatial two - photon state of spdc .
moreover , it is well known that the state of the down - converted photons depends on the phase - matching conditions .
nevertheless , due to its complex structure , the phase - matching functions are usually approximated by gaussian functions @xcite .
while the approach being adopted seems to be reasonable , it has already been shown that fine ( new ) structures in the spatial distributions of these photons can be observed due to ( the manipulation of ) the phase - matching conditions @xcite . in this work
we study the non - gaussianity of the spatial two - photon state of spdc by properly taking into account the phase - matching conditions .
we start by showing how the variances of the near- and far - field conditional probability distributions are affected by the phase - matching functions .
then , we analyze the role of the epr - criterion @xcite regarding the non - gaussianity and entanglement detection of this state .
even though it has been demonstrated that higher order separability criteria can be used for the entanglement detection of spatial non - gaussian entangled states @xcite , we show that a proper consideration of the phase - matching function reveals , precisely , when the _ simpler _ epr - criterion can still be used for the spatial entanglement detection .
we also show that ( and when ) the epr - criterion can be used as a witness for the non - gaussianity of this state .
furthermore , we introduce a statistical measure , based on the _ negentropy _ @xcite of the near- and far - field joint distributions , which allows for the quantification of the non - gaussianity of the spatial two - photon state of spdc .
this measure does not correspond to a quantum mechanical generalization of the negentropy , such as the non - gaussianity measure based on the quantum relative entropy ( qre ) @xcite , and so does not require the knowledge of the full density matrix .
only the moments associated with the diagonal sub - matrices of the covariance matrix need to be measured .
thus , it is experimentally more accessible @xcite .
moreover , for most of the configurations used so far , we show that its value can be estimated from the ( easier to measure ) marginal and conditional distributions .
we also demonstrate that it has common properties with previous introduced measures of non - gaussianity @xcite .
the quantification of the non - gaussianity of a quantum state has important applications for quantum information @xcite , and thus the practicality our measure shall be relevant for new applications of the spatial correlations of spdc in this field .
we consider the process of quasi - monochromatic spdc , in the paraxial regime , for configurations with negligible poynting vector transverse walk - off , that can be obtained using non - critical phase - matching techniques @xcite . in the momentum representation ,
the two - photon state is given by @xcite @xmath3 represents a two - photon state in plane - wave modes whose transverse wave vectors are @xmath4 and @xmath5 .
@xmath6 is the crystal length , @xmath7 is the pump beam wave number , @xmath8 , and @xmath9 is the pump beam waist , which is located at @xmath10 .
this state may be rewritten in the coordinate space as @xcite @xmath11 as @xmath12 , @xmath13 being the sine integral function .
the functions @xmath14 and @xmath15 form a fourier transform pair .
the @xmath16 and @xmath17 functions arise from the phase - matching conditions , and due to the difficulty of dealing analytically with them , they are usually approximated by gaussian functions .
the approach that has been adopted consists in approximating the function @xmath18 by @xmath19 .
sometimes it is used that @xmath20 @xcite , and in other cases the value of @xmath21 is chosen such that both functions coincide at @xmath22 @xcite or at @xmath23 @xcite of their peak .
while this approximation seems to be reasonable for the entanglement quantification @xcite , there has been no investigation to determine how precise it is for describing the distribution of the momentum correlations . besides
, it should be noticed that once a gaussian approximation is adopted for the @xmath16 function , the corresponding approximation for the @xmath17 function is already defined by the fourier transform .
therefore , it is also not clear that such approximation is indeed good for describing the position distributions .
thus , it is not clear whether the spdc two - photon state can indeed be written as a two - mode gaussian state . and @xmath24 ) and
when some gaussian approximations are assumed ( @xmath25 and @xmath26 ) . in ( @xmath27)-(@xmath28 ) the curves @xmath29 and @xmath24 ( red solid lines ) are compared with @xmath30 and @xmath31 for @xmath32 ( green dashed line ) , @xmath33 ( blue dot - dashed line ) , and for some specific values of @xmath34 . in ( @xmath35 ) and ( @xmath36 ) the variances of the momentum and position conditional probabilities are plotted in terms of @xmath34 , respectively .
@xmath37.,scaledwidth=70.0% ] to investigate this point we study how the variances of the momentum ( far - field ) and position ( near - field ) conditional probability distributions are affected by the phase - matching function , and compare the obtained results with the cases where gaussian approximations are considered . due to the symmetry of the two - photon wave functions ,
there is no loss of generalization if we work in one dimension ( i.e. , @xmath38 and @xmath39 ) . to simplify our analysis
we define the following dimensionless variables : @xmath40 and @xmath41 , @xmath42 .
the probabilities for coincidence detection at the far- and near - field planes are @xmath43 where @xmath44 , @xmath45 , and @xmath46 is the diffraction length of the pump beam .
the joint probabilities are related with the conditional ( and marginal ) probabilities @xmath47 [ @xmath48 ) ] and @xmath49 [ @xmath50 ) ] , through the rule : @xmath51 . in the case of the gaussian approximations discussed above , the probabilities of coincidence detection are @xmath52 where different values of @xmath53 represent distinct gaussian approximations for the @xmath16 function . in figure [ fig1](a )
[ ( b ) ] we compare the curves @xmath29 and @xmath25 ( @xmath24 and @xmath26 ) considering @xmath54 , @xmath55 ( and the crystal centered at @xmath10 ) , for the case where the dimensionless parameter @xmath56 .
this parameter has been used in the study of the quantification of the spatial entanglement @xcite , and it brings universality to the theory since a certain value of @xmath34 can be reached in three different ways .
here we considered the values of @xmath6 and @xmath7 as fixed parameters such that @xmath34 varies with @xmath9 .
in this case we find @xmath32 ( @xmath57 ) for the case where the @xmath16 and gaussian functions coincide at @xmath23 ( @xmath22 ) . from figure [ fig1](a)-(b )
one can see that for a small value of @xmath34 , the gaussian approximation only describes properly the momentum conditional distribution .
the position conditional distribution is barely described by the approximation . in figure
[ fig1](c ) and figure [ fig1](d ) we have the same type of analysis but now for a larger value of @xmath34 . in this case
, the gaussian approximation is useful only for describing the position conditional distribution .
the overall behavior of the gaussian approximations is showed in figure [ fig1](e ) and figure [ fig1](f ) , where the normalized variances of the far- [ @xmath58 and near - field [ @xmath59 conditional distributions are plotted in terms of @xmath34 .
the near and far - field conditional probabilities can be used for implementing the epr - paradox @xcite .
this is done by observing the violation of the inequality @xmath60 , which certifies the quantum nature of the spatial correlations of the down - converted photons .
since we have determined how the phase - matching function affects the variances @xmath61 and @xmath62 , we can also look for its effect on the epr - criterion .
this is showed with the red ( solid ) line in figure [ fig2 ] , which was calculated for the values of @xmath63 and @xmath64 at the origin . for smaller values of @xmath34 ,
the conditional variances are independent of the @xmath63 and @xmath64 values @xcite .
whenever @xmath34 increases , the variances become dependent on @xmath63 and @xmath64 . nevertheless for smaller values of @xmath63 and @xmath64 , which are of most experimental relevance , the red ( solid )
curve shown in figure [ fig2 ] captures the overall behavior of the product of @xmath61 and @xmath62 for the state of eq ( [ estmon ] ) .
as we can see , for values of @xmath34 smaller than @xmath65 or greater than @xmath66 , the epr - criterion can safely be used for detecting the spatial entanglement of the two - photon state of spdc . besides of being useful as a entanglement witness , the epr - criterion can also be used as a witness for the non - gaussianity of the spatial state of spdc .
this emphasizes another application for this criterion , which has been related already with other quantum information tasks @xcite . to observe this
, we note that for a pure two - mode gaussian state it is possible to show that @xmath67 [ see appendix [ suplementoa ] ] , such that @xmath68 is an upper bound for the epr - criterion with these states . since it has been demonstrated in @xcite that the two - photon state of eq .
( [ estmon ] ) is always entangled , we can say whenever the product of variances is greater than @xmath68 , that it witnesses the non - gaussianity of the entangled spatial two - photon state of spdc . as one can see in figure [ fig2 ] , this happens for @xmath69 .
the schmidt decomposition of the state wave function , used together with the epr - criterion value , reveals the non - gaussianity of a two - mode entangled state .
such observation does not necessarily hold true when other second - order moments criteria are considered .
this is shown in appendix [ suplementob ] for the criterion of ref .
@xcite , and for the states considered in figure [ fig2 ] . .
the red ( solid ) curve corresponds to the epr values of the two - photon state generated in the spdc [ eq . ( [ estmon ] ) ] .
the green ( dotted ) , blue ( dot - dashed ) and black ( dotted ) curves describe the epr - criterion for the two - photon gaussian states defined in terms of @xmath70 , @xmath71 and @xmath72 , respectively.,scaledwidth=70.0% ]
from our previous analysis it is clear that the spatial state of spdc can not be seen as a two - mode gaussian entangled state , even when it is generated with a gaussian pump beam and in the case of perfect phase - matching .
we now proceed to quantifying the non - gaussianity of this state .
first , we introduce the concept of _ negentropy _ which is the base of our approach @xcite .
the negentropy of a probability density function @xmath73 is defined as @xmath74-h[p(\xi_1,\xi_2)],$ ] where @xmath75 is a gaussian distribution with the same expected values and covariance matrix of @xmath73 .
the function @xmath76 $ ] , called _ differential entropy _ , is defined as @xmath76\equiv-\!\int\ !
d\xi_1 d\xi_2\,p(\xi_1,\xi_2)\log_2p(\xi_1,\xi_2)$ ] @xcite .
the advantage of using negentropy is that it can be seen as the optimal estimator of non - gaussianity , as far as density probabilities are involved .
this is due to the properties that it is always non - negative , and that it is zero only for gaussian distributions . besides , it is invariant under invertible linear transformations @xcite . motivated by these properties we define the total non - gaussianity of the spatial two - photon state of spdc as @xmath77 + n[p_{nf}^s({\tilde{x}}_1,{\tilde{x}}_2)],\ ] ] where @xmath78 $ ] and @xmath79 $ ] are the negentropies of the far- and near - field joint distributions of eq .
( [ pff ] ) and eq .
( [ pnf ] ) , respectively . in appendix [ suplementoc ]
we give , explicitly , the calculations of @xmath78 $ ] and @xmath79 $ ] .
it is possible to observe that @xmath80 if and only if @xmath81 is a two - mode gaussian state [ see appendix [ suplementod ] ] .
otherwise @xmath82 .
we obtain that @xmath78 \approx 0.15 $ ] and that @xmath79 \approx 0.22 $ ] .
thus , the total non - gaussianity of the spatial two - photon state of spdc [ eq ( [ estmon ] ) ] is @xmath83 .
it is interesting to note that it does not depend on @xmath34 .
this was expected since the phase - matching functions do not change their functional form when @xmath34 varies .
the calculations performed for these negentropies can be adapted for different experimental geometries , or used for the proper determination of other quantum information quantities related with the differential entropy of the spatial joint distributions @xcite .
the fact that @xmath84 , emphasizes that spatial gaussian approximations should be taken carefully due to the _ extremality _ of gaussian states @xcite . for comparison purposes , we calculated in appendix [ suplementoe ] the value of @xmath85 , which is the measure of non - gaussianity based on the qre @xcite .
we obtain that @xmath86 and such result also highlight the non - gaussian character of the state of eq .
( [ estmon ] ) .
furthermore , it has the same behavior of @xmath87 , since it does not depend on the parameter @xmath34 . in analogy with eq .
( [ ngt ] ) , we define the non - gaussianity of the conditional and marginal distributions as : @xmath88+n[p^s_{nf}(x_i|x_j)]$ ] and @xmath89+n[p^s_{nf}(x_i)]$ ] with @xmath90 and @xmath91 . according to these definitions
, one can observe that the non - gaussianity of the spatial state of spdc decreases under partial trace , such that @xmath92 ; and that it is additive when the composite system is represented by a product state , i.e. , if @xmath93 is a product state , then @xmath94 [ see appendix [ suplementod ] ] .
these are common properties with the qre measure of @xcite .
in figure [ fig3](a ) we plot the negentropies @xmath95 $ ] and @xmath96 $ ] as a function of @xmath34 .
it is interesting to note that these curves quantify the idea already presented in figure [ fig1](e)-(f ) . as larger
the value of @xmath34 is , the less the conditional momentum distribution can be approximated by a gaussian function .
on the other hand , the conditional position probabilities tends to a normal distribution when @xmath34 increases . in figure [ fig3](b ) we plot the negentropies of the near- and far - field marginal probabilities .
one can see that they have a different dependence on @xmath34 in comparison with the conditional probabilities .
now , the near - field distribution tends to a gaussian function for larger values of @xmath34 , and the far - field one for smaller values of @xmath34 . in figure [ fig3](c ) and
figure [ fig3](d ) we have @xmath97 and @xmath98 plotted in terms of @xmath34 .
the insets of these figures show the corresponding near- and far - field distributions at the points of minimum , indicated with red circles . as it is shown in appendix [ suplementof ]
[ see also figure [ fig3](c ) and figure [ fig3](d ) ] , in the limit of @xmath99 , one can use the relation between the joint and conditional density probabilities to decompose @xmath87 as the sum of @xmath97 and @xmath98 : @xmath100 in a typical experimental configuration for spdc , where the pump beam spot size is around @xmath101 mm at the crystal plane , the value of @xmath34 can be smaller than @xmath102 . thus , in general , the total non - gaussianity of the spatial two - photon state of spdc can be estimated in terms of the ( easier to measure ) near- and far - field negentropies of the conditional and marginal distributions .
this simplify the measurement of @xmath87 , since there is no need to scan the whole transverse planes associated with the near- and far - field joint distributions .
we have investigated the spatial distributions of the entangled down - converted photons by proper considering the phase - matching conditions . by understanding how the near- and far - field conditional distributions are affected by the phase - matching function
, we could show that the epr - criterion @xcite can be used as a witness for the non - gaussianity of the spatial state of the spdc .
the work culminated in the quantification of the non - gaussianity of this state , which was based in a new and very experimentally accessible measure . as it has been discussed in @xcite ,
the quantification of the non - gaussianity of a quantum state has many applications in the area of continuous variables quantum information processing .
thus , we envisage the use of our result for new applications of the spatial correlations of spdc .
as we mentioned in the main paper , the momentum representation of the spatial two - photon state generated in the spontaneous parametric down - conversion ( spdc ) process , under paraxial approximation and for configurations with negligible poynting vector transverse walk - off , can be written as @xcite @xmath103 where the amplitude @xmath104 is given by @xmath105 with @xmath106 being the angular spectrum of the pump beam and @xmath107 representing the phase - matching conditions of the non - linear process .
if both functions are represented by gaussian functions of the form and @xcite , we have in momentum representation that @xmath108 and in position representation that @xmath109 the parameter @xmath110 can be adjusted in order to approximate the phase - matching function by distinct gaussian functions . for a pump laser beam with a gaussian transverse profile , @xmath111 , where @xmath112 is the radius of this pump at the plane of the non - linear crystal . based on eqs .
( [ gaussian - q ] ) and ( [ gaussian - x ] ) , we can obtain the probability density functions for the conditional position and momentum distributions of the down - converted photons .
the variances of these curves are , in general , independent of the value considered for @xmath113 and @xmath114 @xcite and here we use , for simplicity , @xmath115 and @xmath116 . in this case @xmath117 and @xmath118 are given by @xmath119,\ ] ] and @xmath120.\ ] ] here @xmath121 and @xmath122 are the widths of the respective gaussian functions @xcite .
the variances of the conditional distributions can be calculated directly from eqs .
( [ p - x ] ) and ( [ p - q ] ) , which results in @xmath123 and @xmath124 then , the product of them is given by @xmath125 let us now consider the schmidt number @xmath2 for gaussian states .
it was showed in ref . @xcite that it is given by the expression @xmath126 comparing eqs .
( [ proddxdq ] ) and ( [ schmidt ] ) , it follows immediately that the product of the conditional variances is a function of the schmidt number , such that @xmath127 this expression is valid for any gaussian approximation taken for the phase - matching function .
it is possible to see that the maximal value of @xmath128 is equal to @xmath68 and that it happens when the schmidt number @xmath129 ( i.e. , for product gaussian states ) .
another entanglement detection criterion , based on second - order moments , is the one introduced by mancini _
_ @xcite . for the case of spatial entanglement it reads @xcite @xmath130 ^ 2[\delta(\tilde{q}_1+\tilde{q}_2)]^2\geq 1.\ ]
] when this inequality is violated , the state is not separable .
if this condition is satisfied , no information can be drawn .
for the far and near - field joint spatial distributions [ eq . ( 3 ) and eq .
( 4 ) of main paper , respectively ] , the variances in ( [ mancini ] ) are given by @xmath131 ^ 2 \propto \int d\tilde{q}_+\ , \tilde{q}_+^2 e^{-\frac{1}{2}\tilde{q}_+^2}=1,\ ] ] @xmath132 ^ 2 \propto \int d\tilde{x}_-\ , \tilde{x}_-^2 { \ensuremath{\mbox{\hspace{1.3pt}sint}\,}}^2\frac{1}{4p}\tilde{x}_-^2=4\frac{a_2}{a_1}p^2,\ ] ] where @xmath133 and @xmath134 .
the constants @xmath135 and @xmath136 figure [ figmancini ] shows the dependence of the mancini - criterion with the parameter @xmath34 , while considering the state of spdc and when distinct gaussian approximations for this state are considered .
since there is no upper limit for this criterion with gaussian two - mode states , it is not possible to use the mancini - criterion for the detection of the non - gaussianity of the spatial state of spdc .
let @xmath73 be a probability density function .
the negentropy ( @xmath137 ) of @xmath73 is defined as @xcite @xmath138-h[p(\xi_1,\xi_2)],\ ] ] where @xmath75 is a gaussian distribution with the same expected value vector @xmath139 and same covariance matrix @xmath140 of @xmath73 .
the function @xmath76 $ ] , called _ differential entropy _ , is defined as @xcite @xmath141\equiv-\!\int\ !
d\xi_1\!\int\!d\xi_2\,p(\xi_1,\xi_2)\log_2(p(\xi_1,\xi_2)).\ ] ] the joint probability density function in the far - field plane is given by @xmath142 the covariance matrix of @xmath143 is given by @xmath144 the elements in the diagonal are the variances of the marginal distributions @xmath145 and @xmath146 . in order to calculate the terms of @xmath147 , we first note that the expected values @xmath148 are null .
so , we have that @xmath149 and @xmath150 let us consider the following gaussian approximation for describing the joint probability density distribution at the far - field plane : @xmath151 the elements of the covariance matrix @xmath152 of @xmath153 are given by @xmath154 and @xmath155 from eq .
( [ varff ] ) , ( [ covarff ] ) , ( [ varffgaus ] ) and ( [ covarffgaus ] ) we can observe that the condition @xmath156 implies that the joint distributions ( [ psff ] ) and ( [ pgff ] ) have the same covariance matrix . hereafter , the gaussian distribution with the same covariance matrix of ( [ psff ] ) is denoted by @xmath157 .
it is possible to show that the differential entropy of the gaussian distribution @xmath158 is @xmath159=\log_2\frac{\pi e\sqrt{3}}{p}.\ ] ] the differential entropy of @xmath143 is given by @xmath160&=&-\int\!\!\!\int d\tilde{q}_1 d\tilde{q}_2\,p^s_{ff}(\tilde{q}_1,\tilde{q}_2)\log_2(p^s_{ff}(\tilde{q}_1,\tilde{q}_2))\nonumber\\ & = & \log_2\frac{4\sqrt{2}\pi}{3p}-\frac{3p}{4\sqrt{2}\pi}\int\!\!\!\int d\tilde{q}_1 d\tilde{q}_2\,\left\lbrace e^{-\frac{(\tilde{q}_1+\tilde{q}_2)^2}{2}}\right.\nonumber\\ & & \times\left.\left[\log_2e^{-\frac{(\tilde{q}_1+\tilde{q}_2)^2}{2}}+\log_2{\ensuremath{\mbox{\hspace{1.3pt}sinc}\,}}^2\frac{p^2}{4}(\tilde{q}_1-\tilde{q}_2)^2\right]\right\rbrace.\end{aligned}\ ] ] making @xmath161 and @xmath162 we obtain that @xmath160&=&\log_2\frac{4\sqrt{2}\pi}{3p}\frac{1}{2\ln 2}\nonumber\\ & & \times\left[1-\frac{3}{\sqrt{\pi}}\int_0^\infty\!\ ! dv\,{\ensuremath{\mbox{\hspace{1.3pt}sinc}\,}}^2 v^2\ln\left({\ensuremath{\mbox{\hspace{1.3pt}sinc}\,}}^2
v^2\right)\right].\label{intnum}\end{aligned}\ ] ] the last integral in ( [ intnum ] ) must be calculated numerically and it gives that @xmath163 therefore , the differential entropy of @xmath143 is @xmath164\approx\log_2\frac{4\sqrt{2}\pi}{3p}+1.17.\ ] ] according to eq .
( [ negent ] ) , we have that @xmath165&=&h[p^{\tilde{g}}_{ff}(\tilde{q}_1,\tilde{q}_2)]-h[p^s_{ff}(\tilde{q}_1,\tilde{q}_2 ) ] \nonumber \\ & & \approx\log_2\frac{\pi e\sqrt{3}}{p}-\log_2\frac{4\sqrt{2}\pi}{3p}-1.17\nonumber\\ & & \approx\log_2 e\sqrt{\frac{27}{32}}-1.17\nonumber\\ & & \approx 0.15.\end{aligned}\ ] ] the joint probability density function in the near - field plane is given by @xmath166 where @xmath167 is a normalization constant given by @xmath168 the covariance matrix for the near - field joint distribution is @xmath169 where @xmath170 and @xmath171 let us now consider the following gaussian approximation for the joint probability distribution at the near - field plane : @xmath172 the covariance matrix @xmath173 of @xmath174 is determined by @xmath175 where @xmath176 and @xmath177 from eq .
( [ varnf ] ) , ( [ covarnf ] ) , ( [ varnfgaus ] ) and ( [ covarnfgaus ] ) we can observe that the condition @xmath178 must be satisfied for having @xmath179 . the gaussian distribution which satisfies this condition is denoted by @xmath180 .
the differential entropy of this gaussian distribution is given by @xmath181=\log_2\left(2\pi e\sqrt{\frac{\sigma a_2}{a_1}}p\right).\ ] ] the differential entropy for @xmath182 reads @xmath183&=&-\int\!\!\!\int d\tilde{x}_1 d\tilde{x}_2\,p^s_{nf}(\tilde{x}_1,\tilde{x}_2)\log_2(p^s_{nf}(\tilde{x}_1,\tilde{x}_2))\nonumber\\ & = & \log_2(\sqrt{2\pi\sigma}a_1p)-\frac{1}{\sqrt{2\pi\sigma}a_1p}\int\!\!\!\int d\tilde{x}_1 d\tilde{x}_2\,\left\lbrace e^{-\frac{(\tilde{x}_1+\tilde{x}_2)^2}{2\sigma}}\right.\nonumber\\ & & \times\left.{\ensuremath{\mbox{\hspace{1.3pt}sint}\,}}^2\frac{(\tilde{x}_1-\tilde{x}_2)^2}{4p^2}\left[\log_2e^{-\frac{(\tilde{x}_1+\tilde{x}_2)^2}{2\sigma}}\right.\right.\nonumber\\ & & \left.\left.+\log_2{\ensuremath{\mbox{\hspace{1.3pt}sint}\,}}^2\frac{(\tilde{x}_1-\tilde{x}_2)^2}{4p^2}\right]\right\rbrace.\end{aligned}\ ] ] if @xmath184 and @xmath185 , we find that @xmath183&=&\log_2(\sqrt{2\pi\sigma}a_1p)\nonumber\\ & & + \frac{1}{2\ln 2}\left[1-\frac{2}{a_1}\int ds\,{\ensuremath{\mbox{\hspace{1.3pt}sint}\,}}^2 s^2\ln{\ensuremath{\mbox{\hspace{1.3pt}sint}\,}}^2 s^2\right].\end{aligned}\ ] ] the integral @xmath186 must be calculated numerically .
it gives that @xmath187 .
then , the differential entropy of the near - field joint distribution is @xmath188\approx\log_2(\sqrt{2\pi\sigma}a_1p)+1.434.\ ] ] therefore , the negentropy of the near - field joint distribution is @xmath189&=&h[p^{\tilde{g}}_{nf}(\tilde{x}_1,\tilde{x}_2)]-h[p^s_{nf}(\tilde{x}_1,\tilde{x}_2)]\nonumber\\ & \approx & \log_2\left(2\pi e\sqrt{\frac{\sigma a_2}{a_1}}p\right)-\log_2(\sqrt{2\pi\sigma}a_1p)-1.434\nonumber\\ & \approx & \log_2\left(\sqrt{2\pi } e\sqrt{\frac{a_2}{a_1 ^ 3}}\right)-1.434\nonumber\\ & \approx & 0.22.\end{aligned}\ ] ] thus , the total non - gaussianity of spdc is @xmath190
@xmath80 iff the spatial two - photon state of spdc is a gaussian state . since @xmath191 and
that negentropy is always nonnegative , we have that @xmath192=n[p^s_{ff}(\tilde{q}_1,\tilde{q}_2)]=0 $ ] .
thus , for the joint probabilities we have that @xmath193=h[p^s_{ff}(\tilde{q}_1,\tilde{q}_2)]$ ] and @xmath194=h[p^s_{nf}(\tilde{x}_1,\tilde{x}_2)]$ ] , such that @xmath195 and @xmath196 since the differential entropy @xmath197 is a continuous and monotonic function , it holds that @xmath198 and @xmath199 then , necessarily we have that @xmath200 and @xmath201 thus , we conclude that the two - photon state is a gaussian state . in the opposite direction we consider that the two - photon state is a gaussian state .
then , the joint probability density functions of transverse position and momentum are gaussian functions .
so , from the definition of negentropy we must have that @xmath202=n[p^s_{ff}(\tilde{q}_1,\tilde{q}_2)]=0.\ ] ] so , the total non - gaussianity of the spatial two - photon state becomes @xmath203 if the spatial two - photon state is a product state then @xmath87 is additive , i.e. , @xmath204 . for a set of two random and independent variables ,
the joint probability density functions are given by the product of the probability density functions associated to each variable , i.e. , @xmath205 .
such type of probability density function has the covariance matrix @xmath206 defined by the elements @xmath207 and a expected value vector @xmath208 . the gaussian distribution @xmath209 with the same covariance matrix and expected value vector is given by @xmath210\nonumber\\ & = & \frac{e^{-\frac{(\xi_1-{\ensuremath{\langle\,{\xi_1}\,\rangle}})^2}{2\lambda_{11}}}}{\sqrt{2\pi\lambda_{11}}}\times\frac{e^{-\frac{(\xi_2-{\ensuremath{\langle\,{\xi_2}\,\rangle}})^2}{2\lambda_{22}}}}{\sqrt{2\pi\lambda_{22}}}\nonumber\\ & = & p^{\tilde g}(\xi_1)p^{\tilde g}(\xi_2),\end{aligned}\ ] ] where @xmath211 .
note that these marginal gaussian distributions have the same variance and expected value that @xmath212 and @xmath213 .
since the random variables are independent , it holds that @xmath214=h[p(\xi_1)]+h[p(\xi_2)],\ ] ] and consequently , we have that the negentropy of @xmath73 can be written as @xmath215&=&h[p^{\tilde g}(\xi_1,\xi_2)]-h[p(\xi_1,\xi_2)]\nonumber\\ & = & h[p^{\tilde g}(\xi_1)]\!-\!h[p(\xi_1)]\!+\!h[p^{\tilde g}(\xi_2)]\!-\!h[p(\xi_2)]\nonumber\\ & = & n[p(\xi_1)]+n[p(\xi_2)].\end{aligned}\ ] ] if the two - mode spatial state of spdc is a product state , the near- and far - field probability density functions are written as @xmath216 and @xmath217 . therefore , the properties mentioned above hold .
then , the total non - gaussianity of the spatial two - photon state of spdc will be given by @xmath218+n[p^s_{nf}(\tilde{x}_1,\tilde{x}_2)]\nonumber\\ & = & n[p^s_{ff}(\tilde{q}_1)]+n[p^s_{ff}(\tilde{q}_2)]+n[p^s_{nf}(\tilde{x}_1)]+n[p^s_{nf}(\tilde{x}_2)]\nonumber\\ & = & ng^{m_1}+ng^{m_2},\end{aligned}\ ] ] where we have used that @xmath219+n[p^s_{nf}(\tilde{x}_1)]$ ] and @xmath220+n[p^s_{nf}(\tilde{x}_2)]$ ] .
due to the symmetry present in the two - photon wave function [ see eq .
( 3 ) and eq . ( 4 ) of the main paper ] , we have that @xmath221 . then @xmath222 which proves our statement .
the non - gaussianity of the spatial state of spdc decreases under partial trace , such that @xmath92 . to demonstrate this we first study the negentropies of the far- and near - field marginal distributions , while considering two limiting cases : ( i ) @xmath99 and ( ii ) @xmath34 very large .
the joint distributions of the far- and near - field planes are given by eq .
( [ psff ] ) and eq .
( [ psnf ] ) , respectively .
thus , the marginal distributions are given by @xmath223 and @xmath224 doing the following change of variables : @xmath225 and @xmath226 , we obtain that @xmath227 and @xmath228 considering the limiting case ( i ) we may write the marginal distributions as @xmath229 and @xmath230 in such case , it is clear that the negentropy of @xmath231 is null [ @xmath232=0 $ ] ] .
the negentropy of the far - field marginal distribution will be the same of the function @xmath233 .
it gives that @xmath234\approx\log_2\left(\sqrt{\frac{27e}{32}}\right)-0.4447=0.154.\ ] ] then , for values of @xmath99 , @xmath235 such that @xmath92 . in the other limiting case ( ii )
, we have that @xmath236 and @xmath237 now , the negentropy of @xmath146 is null [ @xmath238=0 $ ] ] .
the negentropy of the near - field marginal distribution will be the same of the function @xmath239 .
it gives that @xmath240\approx\log_2\left(\sqrt{\frac{2\pi e a_2}{a_1 ^ 3}}\right)-0.7122 = 0.224\ ] ] thus , for larger values of @xmath34 , @xmath241 which is also smaller than @xmath87 .
intermediate cases , for values of @xmath242 , have been calculated numerically .
the values of the negentropies of the far- and near - field marginal distributions are shown in figure 3(b ) of the main paper . in figure 3(d ) , @xmath98 is plotted in terms of @xmath34 .
as one can see , in such intermediate cases , one always have that @xmath92 .
for @xmath243 the value of @xmath244 .
for values of @xmath245 , @xmath98 already tends to its maximal value which is @xmath246 ( see figure [ fig2sup ] below ) .
this proves that the non - gaussianity of the spatial state of spdc decreases under partial trace .
for comparison purposes , we calculated the value of the non - gaussianity measure based on the qre , @xmath85 @xcite .
this was done by assuming the case of perfect phase - matching to simplify the calculation of the mean values in the covariance matrix .
then , we considered that no correlations are present between the position of one photon and the momentum of the other photon .
therefore , the cross joint probabilities @xmath247 ( where @xmath90 and @xmath248 ) are described by a product of the marginal position and momentum distributions @xmath249 and @xmath250 . in such case
the covariance matrix can be written , considering the order @xmath251 , as @xmath252 the gaussian state associated to this covariance matrix has a purity of @xmath253 @xcite , and its von - neumann entropy reads @xmath254 .
since the state of eq ( [ estmon ] ) is pure , we have that @xmath86 , which does not depend on the value of @xmath34 .
here we discuss the behavior of @xmath87 when considering the limit of @xmath99 , and that one of the down - converted photon is detected in transverse points around the origin .
first we analyze the negentropy of far - field joint distribution : @xmath257=h[p^{\tilde g}_{ff}(\tilde{q}_1,\tilde{q}_2)]-h[p^s_{ff}(\tilde{q}_1,\tilde{q}_2)].\ ] ] in the limits considered , the differential entropies can be written as @xmath76=h[p(\xi_1|\xi_2)]+h[p(\xi_2)]$ ] , such that the negentropy of the far - field joint distribution may also be rewritten as @xmath258&=&h[p^{\tilde g}_{ff}(\tilde{q}_1|\tilde{q}_2)]+h[p^{\tilde g}_{ff}(\tilde{q}_2)]\nonumber\\ & & -h[p^s_{ff}(\tilde{q}_1|\tilde{q}_2)]-h[p^s_{ff}(\tilde{q}_2)]\nonumber\\ & = & h[p^{\tilde g}_{ff}(\tilde{q}_1|\tilde{q}_2)]-h[p^s_{ff}(\tilde{q}_1|\tilde{q}_2)]\nonumber\\ & & + n[p^s_{ff}(\tilde{q}_2)].\label{nffc}\end{aligned}\ ] ] note that @xmath259 does not have the same expected value and variance of the conditional probability density function @xmath260 .
thus , in general , we can not define @xmath261-h[p^s_{ff}(\tilde{q}_1|\tilde{q}_2)]$ ] as the negentropy of the conditional far - field distribution . considering the joint far - field distribution ( [ psff ] ) , the conditional distribution when @xmath262 will be given by @xmath263 when @xmath264 , the @xmath16 function is much larger than the gaussian function , such that we can approximate this distribution as @xmath265 therefore
, the differential entropy is @xmath266\approx\frac{1}{2}\log_2(2\pi e).\ ] ] note that eq .
( [ ffappr ] ) is a normal distribution with the expected value equal to zero and the variance equal to 1 .
now let s consider the gaussian distribution @xmath158 of eq .
( [ pgff ] ) .
it gives the conditional gaussian distribution @xmath259 at @xmath267 equal to @xmath268 in the case when @xmath264 , this gaussian distribution has also the expected value equal to zero and the variance equal to 1 .
thus , it corresponds to the gaussian distribution with the same expected value and variance of @xmath269 .
hence , _ in this limit _ , we can define the negentropy of the far - field conditional distribution in terms of the @xmath270 , such that @xmath257=n[p^s_{ff}(\tilde{q}_1|\tilde{q}_2)]+n[p^s_{ff}(\tilde{q}_2)].\ ] ] now we consider the negentropy of the near - field joint distribution ( [ psnf ] ) .
if @xmath264 , then the near - field conditional distribution is given by @xmath271 where we assumed @xmath55 for simplicity .
the variance of this function is @xmath272 where @xmath273 and @xmath274 are the constants defined above .
now we consider the near - field gaussian joint distribution @xmath275 .
the near - field conditional distribution associated with this gaussian approximation , when @xmath99 , is given by @xmath276 as one can note , @xmath277 has the same expected value and variance of @xmath278 . therefore , _ in this limit _ we can write the negentropy of the near - field joint distribution as @xmath202=n[p^s_{nf}(\tilde{x}_1|\tilde{x}_2)]+n[p^s_{nf}(\tilde{x}_2)],\ ] ] where @xmath279 $ ] is defined in terms of the differential entropy of @xmath280 .
thus , when we consider @xmath264 we can write the total non - gaussianity of the spatial two - photon state of spdc as @xmath281 where @xmath282+n[p^s_{ff}(\tilde{q}_1|\tilde{q}_2)]$ ] , and @xmath283+n[p^s_{nf}(\tilde{x}_j)]$ ] , with @xmath42 .
we thank m. martinelli , j. l. romero , c. saavedra and l. neves for discussions related with earlier drafts of this paper .
we acknowledge the support of grants fondecyt 1120067 , milenio p10 - 030-f and pfb 08024 . c. h. m. acknowledges cnpq and capes .
e. s. g. acknowledges the financial support of conicyt . | the state of spatially correlated down - converted photons is usually treated as a two - mode gaussian entangled state .
while intuitively this seems to be reasonable , it is known that new structures in the spatial distributions of these photons can be observed when the phase - matching conditions are properly taken into account . here ,
we study how the variances of the near- and far - field conditional probabilities are affected by the phase - matching functions , and we analyze the role of the epr - criterion regarding the non - gaussianity and entanglement detection of the spatial two - photon state of spontaneous parametric down - conversion ( spdc ) .
then we introduce a statistical measure , based on the _ negentropy _ of the joint distributions at the near- and far - field planes , which allows for the quantification of the non - gaussianity of this state .
this measure of non - gaussianity requires only the measurement of the diagonal covariance sub - matrices , and will be relevant for new applications of the spatial correlation of spdc in cv quantum information processing .
999 j. c. howell , r. s. bennink , s. j. bentley , and r. w. boyd , `` realization of the einstein - podolsky - rosen paradox using momentum- and position - entangled photons from spontaneous parametric down conversion , '' * 92 * , 210403 ( 2004 ) .
m. dangelo , y. h. kim , s. p. kulik , and y. shih , `` identifying entanglement using quantum ghost interference and imaging , '' * 92 * , 233601 ( 2004 ) .
a. einstein , b. podolsky , and n. rosen , `` can quantum - mechanical description of physical reality be considered complete ?
, '' phys
. rev . * 47 * , 777 ( 1935 ) .
e. j. s. fonseca , c. h. monken , and s. pdua , `` measurement of the de broglie wavelength of a multiphoton wave packet , '' * 82 * , 2868 ( 1999 ) .
t. yarnall , a. f. abouraddy , b. e. a. saleh , and m. c. teich , `` experimental violation of bell s inequality in spatial - parity space , '' * 99 * , 170408 ( 2007 ) .
s. p. walborn , c. h. monken , s. pdua , and p. h. s. ribeiro , `` spatial correlations in parametric down - conversion , '' physics reports * 495 * , 87 ( 2010 ) .
t. b. pittman , y. h. shih , d. v. strekalov , a. v. sergienko , `` optical imaging by means of two - photon quantum entanglement , '' * 52 * , r3429 ( 1995 ) .
a. gatti , e. brambilla , and l. a. lugiato , `` entangled imaging and wave - particle duality : from the microscopic to the macroscopic realm , '' * 90 * , 133603 ( 2003 ) .
l. neves , g. lima , j. g. aguirre gmez , c. h. monken , c. saavedra , and s. pdua , `` generation of entangled states of qudits using twin photons , '' * 94 * , 100501 ( 2005 ) .
m. n. osullivan - hale , i. ali khan , r. w. boyd , and j. c. howell , `` pixel entanglement : experimental realization of optically entangled @xmath0 and @xmath1 qudits , '' * 94 * , 220501 ( 2005 ) .
a. k. jha , j. leach , b. jack , s. franke - arnold , s. m. barnett , r. boyd , and m. j. padgett , `` angular two - photon interference and angular two - qubit states , '' * 104 * , 010501 ( 2010 ) . c. k. hong and l. mandel , `` theory of parametric frequency down conversion of light , '' * 31 * , 2409 ( 1985 ) .
m. h. rubin , d. n. klyshko , y. h. shih , and a. v. sergienko , `` theory of two - photon entanglement in type - ii optical parametric down - conversion , '' * 50 * 5122 ( 1994 ) . c. h. monken , p. h. s. ribeiro , and s. pdua , `` transfer of angular spectrum and image formation in spontaneous parametric down - conversion , '' * 57 * , 3123 ( 1998 ) .
c. k. law and j. h. eberly , `` analysis and interpretation of high transverse entanglement in optical parametric down conversion , '' * 92 * , 127903 ( 2004 ) . h. di lorenzo pires , c. h. monken , and m. p. van exter , `` direct measurement of transverse - mode entanglement in two - photon states , '' * 80 * , 022307 ( 2009 ) .
k. w. chan , j. p. torres , and j. h. eberly , `` transverse entanglement migration in hilbert space , '' * 75 * , 050101 ( 2007 ) .
s. s. straupe , d. p. ivanov , a. a. kalinkin , i. b. bobrov , and s. p. kulik , `` angular schmidt modes in spontaneous parametric down - conversion , '' * 83 * , 060302 ( 2011 ) .
s. p. walborn , d. s. ether , r. l. de matos filho , and n. zagury , `` quantum teleportation of the angular spectrum of a single - photon field , '' * 76 * , 033801 ( 2007 ) .
d. s. tasca , s. p.
walborn , p. h. s. ribeiro , and f. toscano , `` detection of transverse entanglement in phase space , '' * 78 * , 010304 ( 2008 ) .
d. s. tasca , s. p. walborn , p. h. s. ribeiro , f. toscano , and p. pellat - finet , `` propagation of transverse intensity correlations of a two - photon state , '' * 79 * , 033801 ( 2009 ) .
l. j. zhang , l. neves , j. s. lundeen , and i. a. walmsley , `` a characterization of the single - photon sensitivity of an electron multiplying charge - coupled device , '' j. phys .
b * 42 * , 114011 ( 2009 ) . h. di lorenzo pires and m. p.
van exter , `` observation of near - field correlations in spontaneous parametric down - conversion , '' * 79 * , 041801 ( 2009 ) .
m. d. reid , p. d. drummond , w. p. bowen , e. g. cavalcanti , p. k. lam , h. a. bachor , u. l. andersen , and g. leuchs , `` colloquium : the einstein - podolsky - rosen paradox : from concepts to applications , '' * 81 * , 1727 - 1751 ( 2009 ) .
r. m. gomes , a. salles , f. toscano , p. h. s. ribeiro , and s. p. walborn , `` quantum entanglement beyond gaussian criteria , '' pro .
nat .
aca .
sci . * 106 * , 21517 ( 2009 ) .
a. hyvrinen , j. karhunen , and e. oja , _ independent component analysis _
( wiley , 2001 ) .
m. g. genoni , m. g. a. paris , and k. banaszek , `` quantifying the non - gaussian character of a quantum state by quantum relative entropy , '' * 78 * , 060303(r ) ( 2008 ) .
m. g. genoni , and m. g. a. paris , `` quantifying non - gaussianity for quantum information , '' * 82 * , 052341 ( 2010 ) .
m. ostermeyer , d. korn , d. puhlmann , c. henkel , and j. eisert , `` two - dimensional characterization of spatially entangled photon pairs , '' j. mod . opt . * 56 * , 1829 - 1837 ( 2009 ) .
h. di lorenzo pires and m. p. van exter , `` near - field correlations in the two - photon field , '' * 80 * , 053820 ( 2009 ) .
s. mancini , v. giovannetti , d. vitali , and p. tombesi , `` entangling macroscopic oscillators exploiting radiation pressure , '' * 88 * , 120401 ( 2002 ) .
t. m. cover and j. a. thomas , _ elements of information theory _
( wiley , 1991 ) .
p. comon , `` independent component analysis , a new concept ? , '' signal processing * 36 * , 287 - 314 ( 1994 ) .
p. b. dixon , g. a. howland , j. schneeloch , and j. c. howell , `` quantum mutual information capacity for high dimensional entangled states , '' arxiv:1107.5245v1[quant - ph ] .
m. m. wolf , g. giedke , and j. i. cirac , `` extremality of gaussian quantum states , '' * 96 * , 080502 ( 2006 ) . |
it is now generally accepted that the universe is currently undergoing a period of accelerated expansion . the most popular approach to modeling
the current epoch is to postulate a new energy component of the universe , dark energy , which has negative pressure and hence drives the acceleration .
possible dark energy candidates include the cosmological constant @xcite , which has well known fine tuning issues , quintessence fields ( see for example @xcite and references therein ) and elastic dark energy @xcite .
an alternative approach is to modify the theory of gravity , and there have been a number of attempts to construct late time accelerating solutions to the gravitational field equations by modifying gravity at large distances / late times .
one such class of models are found by considering a gravitational action that contains an arbitrary function @xmath0 of the ricci scalar . the first modified gravity model to be considered introduced quadratic terms into the gravitational action @xcite .
such models can give rise to early - time inflationary solutions to the gravitational field equations , where the hubble parameter is initially @xmath1 , and slowly rolls towards a stable minkowski vacuum state .
modified gravity models which yield late - time acceleration have been considered recently .
an important example is the cddtt model @xcite , for which @xmath2 .
this function has a de - sitter vacuum solution to the field equations , which might be associated with the current epoch of the universe with a suitable tuning of the mass scale @xmath3 .
however , it has been shown that this model can not satisfy local gravity constraints @xcite , and also has an instability @xcite .
in addition to the cddtt model , a number of modified gravity functions have been proposed which give rise to late time acceleration @xcite , and some progress has been made in understanding general properties of these models , for example the behaviour of metric perturbations @xcite and the cosmological evolution of generic @xmath0 functions @xcite have been considered .
any @xmath0 model must satisfy certain conditions in order to exclude the possibility of ghost degrees of freedom and other instabilities .
these conditions are @xmath4 , which ensures that there are no ghost degrees of freedom in the model , and @xmath5 , which prevents instabilities from arising in the early universe @xcite .
it seems also sensible to impose @xmath6 , which ensures that there is no cosmological constant , and @xmath7 for large @xmath8 , since we want any modification to gravity to only become significant at late times .
in addition to these conditions , it has been shown that if a standard matter era ( for which @xmath9 ) is to be a fixed point then further conditions must be satisfied @xcite .
recently , a number of models have been studied in the literature which satisfy the above constraints , at least for @xmath10 .
specifically the following functions have been considered @xcite , @xmath11 , \end{aligned}\ ] ] where @xmath12 is the current vacuum curvature of the universe , @xmath13 are dimensionless constants , and @xmath14 . in ref.@xcite , the @xmath15
is defined differently to ( @xmath16 ) , and in ref.@xcite the precise formula was slightly different to ( @xmath16 ) . however , both models considered in @xcite have the same expansion in the limit @xmath17 , which will be the important point in this paper .
the model @xmath18 represents a modification to general relativity that is a power law in @xmath8 , whilst @xmath19 contains exponential corrections .
these models both have late - time accelerating epochs , and can satisfy local tests of gravity .
they can also satisfy the conditions discussed in ref.@xcite .
although the models ( @xmath16,@xmath20 ) possess many desirable features , an additional issue has arisen recently with regards to their suitability as viable gravitational models @xcite .
specifically , it has been suggested that if we look for solutions to the gravitational field equations that are perturbations around known general relativistic solutions , then the ricci scalar oscillates over very short timescales , and the amplitude of these oscillations will increase without bound to the past @xcite . if true , this presents us with a number of issues , amongst which are that the small perturbation will become much larger than the general relativistic solution at some time in the past , violating the perturbative assumption , and @xmath8 will become negative at some finite time .
such phenomena would most likely lead to significant problems in reproducing the standard cosmology .
the aim of this paper is to study the two models ( @xmath16,@xmath20 ) in detail .
we find that since @xmath8 oscillates over such small timescales , the linearized analysis presented in refs.@xcite will not necessarily yield a solution that is indicative of the full solution for all times .
we then solve the gravitational field equations numerically and semi - analytically and show that the ricci scalar undergoes asymmetric oscillations , and will drift away from the general relativistic limit .
we note that non - linear oscillations have been considered in @xmath0 models previously , see ref.@xcite .
in this section we consider the modified gravity action @xmath21 which yields the following field equations @xmath22 f'(r ) = { t_{\mu\nu } \over m_{\rm pl}^{2 } } , \ ] ] @xmath23 where primes denote differentiation with respect to @xmath8 , @xmath24 is the energy momentum tensor and @xmath25 is the reduced planck mass . the hss and ab models can be expanded as @xmath26 for @xmath27 , where @xmath12 acts as a small ` cosmological constant ' ( although there is no true constant in these models since globally @xmath6 ) .
@xmath28 , @xmath29 and @xmath30 are all small functions of @xmath8 , in the sense that the dimensionless parameters @xmath31 , @xmath29 and @xmath32 all satisfy @xmath33 , @xmath34 and @xmath35 for @xmath36 . for the hss and ab models , we have @xmath37 where @xmath12 is the current vacuum curvature of the universe , and @xmath38 and @xmath39 are parameters that are smaller than @xmath12
. these expansions are valid for @xmath40 and @xmath41 for the hss and ab models respectively . for the rest of this paper
, we will drop the subscripts @xmath42 and @xmath43 , and use @xmath44 or @xmath45 and @xmath46 or @xmath47 , which should be obvious in context .
we will explicitly state which model is being studied in each section .
finally , we note that we will frequently make use of the notation @xmath48 , which is the general relativistic solution to the field equations @xmath49 .
in particular during the matter era we have @xmath50 , and during the radiation era @xmath51 .
in this section , we review the perturbative analysis that has been used in refs.@xcite to derive approximate solutions to the gravitational field equations for the ab and hss models .
the full field equations are a set of non - linear , fourth order differential equations for the scale factor , and solving these equations directly is difficult ( although see for example ref.@xcite for an attempt to do so for the cddtt model ) . the approach taken in this section is to look for solutions to the field equations that are perturbations around known general relativistic solutions , that is we look for a solution to ( @xmath52 ) of the form @xmath53 , where @xmath54 is some small perturbation .
this approach involves linearizing the equation in @xmath55 . in doing so
, the following expansions are used @xmath56 in which case ( @xmath52 ) becomes @xmath57 to linear order in @xmath55 . unless otherwise stated , in ( @xmath58 ) and for the remainder of this paper @xmath59 is taken to be a function of @xmath48 only , @xmath60 .
the function @xmath61 is @xmath62 which is a function of @xmath48 only and is small .
the field equation ( @xmath58 ) is an inhomogeneous , second order differential equation for @xmath55 .
the solution is the linear sum of a particular solution to ( @xmath58 ) and the solution to the homogeneous equation @xmath63 which can be written approximately as @xmath64 where @xmath65 is the scale factor . in ref.@xcite , a solution to ( @xmath66 ) was derived for the hss model .
using @xmath46 and @xmath44 , the equation for @xmath55 reads @xmath67 a wkb solution can obtained for the matter era , @xmath68 , \ ] ] and for the radiation era @xmath69 , \ ] ] where @xmath70 are constants .
we have solved equation ( @xmath71 ) numerically , and @xmath72 is shown in fig.[fig:1 ] , taking a pure matter era as an example , so @xmath50 .
we have used time coordinates such that @xmath73 , and have chosen @xmath74 and @xmath75 .
we have evolved backwards in the time coordinate , over the range @xmath76 , using the initial conditions @xmath77 , @xmath78 , where @xmath79 . by solving equation ( @xmath71 ) numerically , we have found that @xmath55 oscillates symmetrically around zero , confirming the behaviour expected from the wkb solutions .
however , in fig.[fig:1 ] we have not exhibited the actual oscillations of @xmath8 explicitly , since they are of too high frequency to be resolved , and so have only presented @xmath48 and the upper and lower envelopes of the oscillations . from fig.[fig:1
] , we see that the lower envelope appears to have a turning point , indicating that in this approach after this time the oscillatory component @xmath55 will come to dominate over @xmath48 , and hence the ricci scalar will become negative at some point in the past . in refs .
@xcite , a similar analysis has been considered for the model @xmath80 .
if we use @xmath47 in ( @xmath66 ) , we find the following equation @xmath81 which has the wkb solution @xmath82 , \ ] ] during the matter era and @xmath83,\ ] ] for the radiation era , where @xmath84 are constants .
we have solved equation ( @xmath85 ) numerically , taking @xmath50 , using time coordinates with @xmath86 and choosing @xmath87 .
the resulting @xmath88 is shown in fig.[fig:2 ] , using the initial conditions @xmath89 , @xmath90 over the time range @xmath91 .
once again , we have shown only the envelopes of the oscillations of @xmath8 .
these solutions exhibit similar oscillatory and growing behaviour as found for the hss model . using this approach for both the hss and ab models ,
we find similar behaviour for @xmath8 .
specifically , @xmath55 undergoes rapid oscillations , and both the amplitude and frequency of these oscillations increases to the past .
since the amplitude of @xmath55 grows without bound as @xmath92 , it follows that @xmath55 will violate the condition @xmath93 at some point in the past , at which point the perturbative analysis will break down . beyond this
, we can not assume that @xmath88 is a solution to the gravitational field equations .
this oscillatory behaviour presents us with a number of problems .
the first is that since @xmath55 grows to the past , it will eventually satisfy @xmath94 . beyond this point
, @xmath8 will periodically be negative , and when @xmath8 is negative we can no longer use the expansion @xmath95 .
this is a problem particularly relevant for the hss model , since for @xmath96 there will be a point at which @xmath97 , which is a singular point in the field equations . the second problem associated with the oscillatory behaviour of @xmath55
is that the frequency grows to the past without bound .
it has been pointed out in ref.@xcite that this issue can be ameliorated by introducing an additional term @xmath98 into the action , where @xmath99 is a mass scale .
this will provide a cut off in the frequency growth at @xmath100 .
+ + +
in this section , we first show that the above linearized analysis is only valid over a very limited range of @xmath8 , and that the gravitational field equations are generically non - linear . to see this , we first write ( @xmath52 ) as @xmath101 where we have expanded the function @xmath95 , and neglected the term @xmath102 .
both of these assumptions are acceptable if we consider a regime where @xmath36 . in section [ sec : am1 ] , we linearized this equation in @xmath55 , and in doing so we neglected the term @xmath103 , which was assumed to be second order .
however , since @xmath55 oscillates with high frequency @xmath104 , it follows that the term @xmath105 will not remain small , since it will grow like @xmath106 to the past .
hence the accuracy of the above linearized analysis will begin to reduce when @xmath103 is of the same order as @xmath107 . beyond this point , we can no longer use the linearized equation for @xmath55 , and we must include these non - linear terms . for the ab model , we can explicitly show the range of @xmath55 over which the linearized analysis is valid . to see this , we consider the expansions ( @xmath108-@xmath109 ) . as an example , we take @xmath110 , which was expanded as @xmath111 . for the ab model
this reads @xmath112 .
we compare this to our actual function @xmath113 , given by @xmath114 , and we see that the expansion of @xmath115 is only valid in the very limited range @xmath116 . once this condition is violated , we can no longer use the linearized equation ( @xmath58 ) for @xmath55 . since in general we will have @xmath117 , then it follows that the linearized analysis will break down for the ab model long before @xmath118 .
the above reasoning suggests that although the linearized analysis successfully predicts the oscillatory behaviour of these models , we can only trust the solution for a limited range of @xmath8 , and beyond this the field equation for @xmath8 is inherently non - linear .
we conclude that we can not find a solution to ( @xmath52 ) for which @xmath119 for all @xmath8 . therefore in this section
, we do not consider solutions to the field equations that are perturbations around the general relativistic ricci scalar , but rather look for full numerical solutions to ( @xmath52 ) .
however , it is not just a second order differential equation for @xmath8 , but rather a fourth order non - linear equation for the scale factor @xmath65
. therefore we will treat ( @xmath52 ) , @xmath120 and @xmath121 as a set of coupled non - linear differential equations for @xmath8 , @xmath122 and @xmath123 .
we stress that we are not making the assumption that @xmath124 , and the only assumption that we will make is that during the matter era @xmath125 .
we do so because we wish to compare the full numerical results obtained here with the results of the linearized analysis of the section [ sec : am1 ] , where the vacuum curvature was neglected .
we have checked that introducing the vacuum curvature does not significantly effect our results , since we can absorb it into the energy momentum tensor .
we look for solutions to these equations for two sets of initial conditions .
first we perturb @xmath8 initially from its general relativistic limit , and compare the full numerical solution to the linearized analysis of the previous section .
we then set the initial conditions such that @xmath8 , @xmath123 and @xmath122 are initially exactly their general relativistic values , and observe how the hss and ab models evolve backwards through the matter era .
for the hss model , using the example of a pure matter era for which @xmath126 , we find @xmath127 together with the equations for @xmath122 and @xmath123 . the trace of the energy - momentum tensor @xmath128 is related to @xmath123 and @xmath122 through the conservation equation @xmath129 .
specifically , during the matter era we have @xmath130 . to derive ( @xmath131 ) from ( @xmath52 ) , we have used @xmath132 in ( @xmath133 ) .
the only assumptions made are that @xmath134 and that ignoring the @xmath12 term will not significantly effect our results , which is valid throughout our numerical calculation .
it is clear that @xmath135 is not a solution of these equations , and hence the hss model will not exactly mimic general relativity .
as we will see , even if we set the initial conditions such that @xmath8 , @xmath122 and @xmath123 are initially equal to their general relativistic values , if we evolve this model backwards in time then they will deviate from @xmath49 .
we have solved the coupled differential equations ( @xmath131 ) , @xmath120 and @xmath121 numerically over the range @xmath76 , taking @xmath75 , @xmath73 and @xmath136 as before .
we have found that we can only evolve the ricci scalar over a small range of @xmath137 . as we will see in the next section , this is because the ricci scalar will generically evolve to a singularity at some finite point in the past .
for now , we will ignore this singular behaviour and solve ( @xmath131 ) for @xmath8 over a small dynamical range .
the ricci scalar , obtained using a stiff differential equation solver , is shown in fig.[fig:3 ] .
here we have chosen the initial conditions @xmath138 , @xmath139 , @xmath140 , and @xmath141 , where @xmath142 , and evolved backwards in the time coordinate .
these initial conditions are the same as those chosen in the previous section , so that we can compare the full solution found here to the solution found using the linearized approximation @xmath88 . in fig.[fig:3 ] , we have exhibited the envelope of the oscillations of the ricci scalar , and the general relativistic solution @xmath48 . from the envelope functions , we see asymmetric oscillations of @xmath8 about @xmath48 .
we note that the solution obtained in this section is significantly different to the one obtained in the linearized analysis , shown in fig.[fig:1 ] .
specifically , in fig.[fig:3 ] we find that there is no turning point in the lower envelope in the time range considered .
conversely , we find that the upper envelope function increases at a faster rate than predicted in the linearized analysis .
we have also plotted @xmath143 for this model in fig.[fig:4 ] , which is the fractional difference between the hubble parameter @xmath122 for the hss model and the general relativistic hubble parameter during the matter era , @xmath144 .
we see that @xmath145 oscillates asymmetrically and not exactly around its general relativistic limit @xmath146 .
the amplitude of these asymmetric oscillations is highly suppressed but increasing to the past , and unless @xmath147 is chosen to be sufficiently small these oscillations may come to dominate @xmath122 . for the scale factor ,
we have again plotted the fractional difference @xmath148 in figs.[fig:5 ] and [ fig:6 ] , and we see that @xmath123 deviates from the general relativistic scale factor , but this deviation @xmath149 is highly suppressed .
however , as with the hubble parameter , @xmath149 grows to the past , and hence may become significant at some earlier time .
having solved the system of equations for @xmath8 , @xmath122 and @xmath123 by perturbing @xmath8 initially from the general relativistic limit , we now solve with no initial perturbation , that is we take the initial conditions @xmath150 , @xmath139 , @xmath140 , and @xmath141 , where @xmath151 . the solution to equation ( @xmath131 ) is shown in figs.3(a - d )
. again we have taken @xmath152 , @xmath73 and evolved over the time regime @xmath153 .
we see that @xmath8 oscillates , and @xmath122 and @xmath123 deviate from their general relativistic values .
however , once again @xmath149 and @xmath154 are suppressed , and this model closely mimics general relativity over this range of time . we can perform similar calculations for the ab model .
taking once again @xmath50 as a specific example , we find @xmath155 we obtain ( @xmath156 ) by using @xmath47 in equation ( @xmath133 ) . we have solved this equation numerically , along with the equations for @xmath122 and @xmath123 , using the same differential equation solver as for the hss model , taking @xmath157 and using the initial conditions @xmath158 , @xmath159 , @xmath140 and @xmath141 . as in section
[ sec : am1 ] , for the ab model we only solve equation ( @xmath156 ) over the very small time range @xmath91 , due to the presence of a singularity . by solving ( @xmath156 ) over this limited dynamical range , we find asymmetric oscillations of @xmath8 , @xmath122 and @xmath123 .
the envelope of the oscillations of @xmath8 are shown in fig.[fig:11 ] .
once again , we see no turning point in the lower envelope , suggesting that @xmath8 does not become negative for these models .
we also observe that the upper envelope grows faster than predicted by the linear analysis .
the hubble parameter and scale factor undergo asymmetric , suppressed oscillations which grow to the past , as seen in figs.4(b - d ) .
we also solve equation ( @xmath156 ) without perturbing @xmath8 , @xmath122 and @xmath123 from their general relativistic limits initially , and evolving the system backwards over the same time regime @xmath91 .
the results are shown in figs.5(a - d ) .
we find that @xmath8 oscillates , and @xmath143 and @xmath148 grow , indicating that @xmath123 and @xmath122 both diverge from their general relativistic limits .
however , this divergence is highly suppressed , as was found in the hss model .
so far , we have considered two approaches to solving the modified gravitational field equations .
it has been found that the linearized approach , in which the ansatz @xmath88 with @xmath93 is used in ( @xmath52 ) , will not give a solution that is indicative of the full solution , since non - linear terms can not be neglected due to the rapid oscillations of the ricci scalar .
the second approach considered was to solve the full gravitational field equations numerically .
in doing so , it was found that the ricci scalar undergoes non - linear oscillations , very closely ( but not exactly ) about its general relativistic limit .
however , we found that solving the gravitational field equations over significant timescales is impossible due to some kind of singularity at a finite time . in this section , we consider neither a linearized analysis nor a full numerical study of the field equations . instead , under sensible assumptions
, we show that the trace of the gravitational field equations for both the ab and hss models reduces to a non - linear wave equation , and hence that @xmath8 undergoes asymmetric oscillations about its general relativistic limit , in agreement with the results of section [ sec : s10 ] .
moreover we find that there is a singularity in @xmath8 at a finite time in both models .
since the approach taken is model specific , we consider the ab and hss models separately .
we begin by substituting @xmath47 into ( @xmath52 ) , giving @xmath160 next , we define @xmath161 , so @xmath162 and ( @xmath163 ) becomes @xmath164 where we have assumed that @xmath165 and hence neglected the vacuum energy .
next , by defining @xmath166 , we can write ( @xmath167 ) as @xmath168 to solve this equation , we write it as @xmath169 @xmath170 if we now introduce the fast time coordinate @xmath171 which is the timescale associated with the oscillations of the ricci scalar , then we can write ( @xmath172 ) as @xmath173 where primes denote derivatives with respect to @xmath174 , and we have defined @xmath175 .
if we solve equation ( @xmath176 ) for @xmath177 , then we can deduce @xmath8 from the relation @xmath178 note that ( @xmath179 ) is of the form @xmath180 , however we have not assumed that @xmath181 to deduce this expression .
if we now define @xmath182 , then we can write ( @xmath176 ) as @xmath183 + \ ] ] @xmath184 equation ( @xmath185 ) is a non - linear , second order inhomogeneous differential equation for @xmath186 , and the full solution is the sum of the particular solution to ( @xmath185 ) and the solution to the corresponding homogeneous equation given by setting the right hand side to zero .
@xmath187 is now the general relativistic limit , but it is clear that this is not an exact solution . we will find that @xmath186 undergoes asymmetric oscillations governed by the homogeneous equation , around the particular solution of ( @xmath185 ) .
we now calculate the oscillatory and drift components of @xmath186 .
we begin by calculating the drift of @xmath186 away from @xmath187 .
to do so , we make the assumption that the derivative terms @xmath188 and @xmath189 in ( @xmath185 ) are subdominant and hence can be neglected . with this assumption , ( @xmath185 ) reduces to an algebraic equation for @xmath186 , @xmath190 + \epsilon e^{t/\epsilon m_{\rm pl}^{2}}(1+x ) \log ( 1+x)\ ] ] @xmath191 if we assume that @xmath192 , then by expanding @xmath193 in ( @xmath194 ) we find the following expression for @xmath186 , @xmath195 \right ) .\ ] ] this solution is valid if we only consider terms linear in @xmath186 , that is we consider terms of order @xmath196 only .
taking the derivative of @xmath186 , we find that @xmath189 and @xmath188 are given by @xmath197 since @xmath189 and @xmath188 are much smaller than @xmath186 ( they are suppressed by a factor of @xmath198 and @xmath199 relative to @xmath186 respectively ) , then it follows that our original assumption that we can neglect derivative terms of @xmath186 in ( @xmath185 ) is valid , and ( @xmath200 ) is an approximate solution to ( @xmath185 )
. the solution to the homogeneous equation , obtained by setting the left hand side of ( @xmath185 ) to zero , will give the oscillatory component of @xmath186 and hence @xmath8 .
we will first show that the last two terms on the left hand side of ( @xmath200 ) , are highly suppressed , and hence can be neglected for @xmath201 .
to do so , we write @xmath202 , @xmath203 and @xmath204 in terms of @xmath137 , @xmath205 and the corresponding homogeneous equation to ( @xmath200 ) can be approximately written as @xmath206 + \epsilon e^{t/\epsilon m_{\rm pl}^{2}}\log ( 1+x ) = 0 .\ ] ] we see that the last two terms on the left hand side of ( @xmath207 ) have a common factor of @xmath199 and hence can be neglected .
we arrive at the following equation for @xmath186 , @xmath208 where @xmath209 is a small function of @xmath174 . in ( @xmath210 )
, we will assume that @xmath211 and @xmath212 , since the oscillations of @xmath122 and @xmath123 are suppressed by a factor of @xmath213 and can therefore be neglected . with this assumption , ( @xmath210 )
reduces to an equation purely in terms of @xmath186 and its derivatives .
equation ( @xmath210 ) describes an anharmonic , non - linear oscillator with a small negative damping term , and hence we have shown analytically that the ricci scalar will undergo asymmetric oscillations , in agreement with the results of section [ sec : s10 ] .
we henceforth describe the oscillator considered here as the logarithmic oscillator .
the potential @xmath214 for the logarithmic and simple harmonic oscillators are shown in figs.6(a , c ) .
the logarithmic potential will give rise to wave solutions that are not symmetric . before continuing
, we observe that if we expand @xmath8 as @xmath88 for @xmath215 , then @xmath186 can be expanded as @xmath216 , and ( @xmath210 ) becomes @xmath217 if we write ( @xmath218 ) in terms of @xmath137 , then it reduces to the linearized equation ( @xmath85 ) .
we have solved the non - linear equation ( @xmath210 ) numerically , using the initial conditions @xmath219 , @xmath220 , and evolved over the range @xmath221 .
this range of @xmath174 and choice of initial conditions are the same as were chosen in section [ sec : s10 ] , where the full gravitational field equations were solved numerically . in fig.@xmath222(a ) , we compare the ricci scalar @xmath223 obtained in this section to that found by solving the full gravitational field equations numerically in section [ sec : s10 ] .
it is clear that the results obtained in this section closely mimic the full numerical solution . in fig.@xmath222(b ) we compare @xmath224 to @xmath225 , where @xmath8 is the ricci scalar obtained by solving the full gravitational field equations numerically .
we see that the fractional difference between @xmath55 and @xmath226 remains small throughout the dynamical range considered . using this approach ,
we have found that we can only evolve @xmath186 and hence @xmath8 backwards in the time coordinate @xmath137 over the same limited dynamical range as in section [ sec : s10 ] .
however , by solving equation ( @xmath210 ) we can see why this is the case .
in fig.[fig : l26 ] , we have perturbed @xmath186 from @xmath187 and evolved backwards in the @xmath174 coordinate .
we see that the amplitude of the oscillations of @xmath186 increase due to the presence of the damping term in ( @xmath210 ) , and after a finite time @xmath227 .
since @xmath228 and @xmath229 , then it follows that both @xmath8 and @xmath230 will be singular at this point .
we conclude that wide ranges of initial conditions @xmath231 and @xmath232 will give rise to a singularity in the field equations as we evolve backwards in time .
that is not to say that all initial conditions will lead to this problem , and there exists an extremely restricted range of @xmath231 and @xmath232 at the beginning of the matter era that will not give rise to singular behaviour at some finite time in the past . instead of evolving backwards over the matter era ,
if we choose @xmath231 appropriately at the beginning of the matter era and evolve forwards in time , @xmath186 ( and hence @xmath8 ) will be well behaved and regular . to find the allowed initial conditions
, we can first use the fact that @xmath233 is the lower bound . to obtain an upper bound ,
we assume that the damping term in ( @xmath210 ) is negligible and consider @xmath234 this expression can be multiplied by @xmath189 and integrated to obtain the following expression , @xmath235 = e , \ ] ] where @xmath236 is a constant of integration .
the values of @xmath186 at the peaks and troughs of its wavetrain can be obtained from ( @xmath237 ) by setting @xmath238 and solving the algebraic equation @xmath239 .\ ] ] for a given @xmath236 , there will be two solutions for @xmath186 , corresponding to the maxima and minima of the wave .
we have deduced that @xmath240 is the minimum value that @xmath186 can take in order to avoid singularities , and substituting this into ( @xmath241 ) gives @xmath242 .
as we have stated , ( @xmath241 ) will yield two solutions for any particular @xmath236 , and so finding the second solution for @xmath242 will give us the maximum value of @xmath186 .
this solution is @xmath243 , and hence we conclude that in order to obtain a matter era free from singularities , we must impose the condition that @xmath244 at the beginning of the matter era . in terms of @xmath225
, this corresponds to the range @xmath245 .
since we will generically encounter a singularity if we evolve @xmath186 backwards in the time coordinate @xmath137 , we now solve equation ( @xmath210 ) by taking @xmath186 initially at some early time and evolving forwards in the @xmath174 coordinate , which corresponds to evolving forwards in the standard time coordinate @xmath137 .
in doing so , @xmath186 and hence @xmath8 will remain regular .
we take as initial conditions @xmath246 , @xmath247 and evolve over the range @xmath248 . by solving ( @xmath210 ) , we obtain @xmath249 , which is the oscillatory component of @xmath186 . the full solution to ( @xmath185 )
is approximately given by @xmath250 , where @xmath251 is the drift term given in ( @xmath200 ) .
in fig.[fig : fg9](a - d ) we compare @xmath226 to @xmath55 .
we see that there are two competing effects ; the non - linear oscillations , which are initially large but quickly decay , and the drift away from @xmath252 , which grows as we evolve forwards in the time coordinate .
the approximate solution obtained in this section closely mimics the full numerical solution .
however , the predicted drift begins to deviate from the full solution at late times .
however , this only occurs when the approximation @xmath201 is no longer valid . as an aside
, we conclude this section by considering a constant energy momentum tensor , that is we take @xmath253 . for this ricci scalar , we have @xmath254 , @xmath255 and @xmath256 .
using these in ( @xmath176 ) , we obtain the following equation for @xmath177 , @xmath257 the case of constant @xmath48 was discussed in ref.@xcite using the perturbative analysis of section [ sec : am1 ] , and it was found that the ricci scalar contained an exponentially growing component , suggesting an instability in this model . here , we solve ( @xmath258 ) for @xmath177 and obtain the ricci scalar from @xmath259 .
the solution is presented in figs.[fig : f7](a - b ) , where we have taken @xmath260 , @xmath73 , @xmath261 and evolved over the time regime @xmath262 .
we see that @xmath8 undergoes damped , rapid oscillations about its general relativistic limit .
we find no exponential growth , contrary to the claim made in ref.@xcite . in this section , we consider a similar calculation to that presented above , but now for the hss model .
we first re - write the trace of the gravitational field equations in terms of a dimensionless function and a fast time coordinate .
we then derive approximate expressions to describe the oscillations and drift of the ricci scalar away from its general relativistic limit .
we begin by substituting @xmath132 into ( @xmath52 ) , which gives @xmath263 where we have neglected the term @xmath12 . to transform equation ( @xmath264 ) into an oscillator equation
, we must make a series of field redefinitions . by first defining @xmath265
we write ( @xmath264 ) as @xmath266 + { n+1
\over 3n}{t \over m_{\rm pl}^{2}}\alpha
u^{2n/(2n+1)}\approx 0 .\ ] ] next , by introducing a fast time coordinate @xmath267 and defining @xmath268 , we can write ( @xmath269 ) in terms of @xmath270 and @xmath174 as @xmath271 @xmath272 where primes denote differentiation with respect to @xmath174 , and @xmath175 .
this is a non - linear , inhomogeneous second order differential equation for @xmath186 , and by solving for @xmath186 we obtain the ricci scalar from the relation @xmath273 .
as before , we can deduce both the oscillatory component of @xmath8 and the drift away from @xmath48 .
we now take as a specific example a pure matter era and set @xmath75 , in which case
we can write @xmath274 and @xmath275 . we also make the assumption that @xmath276 , which implies that @xmath277 where we have neglected the third term on the left hand side of ( @xmath278 ) , which is negligible . as in the ab model , an approximate expression for the drift can be found by solving the full inhomogeneous equation ( @xmath279 ) .
we take as an ansatz @xmath280 with this choice , @xmath189 and @xmath188 are of order @xmath281 and @xmath282 , and hence the derivative terms in ( @xmath279 ) can be neglected if we only consider terms of order @xmath283 . as a result
, it reduces to an algebraic expression for @xmath186 , which is solved by ( @xmath284 ) ( this can be verified by direct substitution ) .
having calculated the drift by solving ( @xmath279 ) , the oscillatory component of @xmath186 can be obtained by neglecting the last term on the right hand side of ( @xmath279 ) ( since it is of order @xmath285 ) , and solving the equation @xmath286 this equation describes the non - linear oscillations of @xmath186 around @xmath187 , that is around the general relativistic limit @xmath287 .
we have solved equation ( @xmath288 ) , using the initial conditions @xmath289 , @xmath247 , over the range @xmath290 and taking @xmath136 , and the results are exhibited in fig.(@xmath291 ) .
we compare the ricci scalar obtained in this section , given by @xmath292 , to the ricci scalar found in section [ sec : s10 ] , and find that they are in close agreement . as in the ab model , by using this method we find that we can only evolve @xmath186 backwards over a very limited dynamical range , and this is once again due to the presence of a singularity . in fig.@xmath293 , we see that as we evolve @xmath186 , after a finite time @xmath227 , and since @xmath292 , we find that as @xmath227 , @xmath294 .
this singularity arises due to the initial conditions that we have imposed . as in the ab model , to evade this singularity we will instead choose initial conditions at an early time and evolve forwards through the matter era . by doing so , @xmath186 will remain regular throughout . to find the initial conditions that @xmath186 may take in order to evade this singularity , we perform the same steps as in the ab model . we first observe that @xmath295 is a lower bound , and the upper bound can be obtained from the expression @xmath296 where
we have neglected the damping term and terms of order @xmath285 in ( @xmath288 ) . by substituting @xmath227 , we find that @xmath297 , and hence the upper bound for @xmath231 is found by solving ( @xmath298 ) for @xmath242 . doing so , we arrive at @xmath299 , which for @xmath75 is given by @xmath300 . in terms of @xmath292
, this corresponds to the choice @xmath301 at the beginning of the matter era . with this in mind
, we now solve ( @xmath288 ) by evolving forwards in the fast time coordinate @xmath174 , which corresponds to evolving forwards in the @xmath137 coordinate .
we solve ( @xmath288 ) over the range @xmath302 , with initial conditions @xmath303 and @xmath247 . by solving ( @xmath288 )
we obtain the oscillatory component @xmath249 of @xmath186 , and hence @xmath304 , where @xmath251 is given in ( @xmath284 ) . in fig.([fig : fg11 ] ) we compare @xmath186 obtained in this section to @xmath305 , where @xmath8 is the ricci scalar obtained solving the full field equations numerically . as in the ab model , there are two effects ; the non - linear oscillations of @xmath186 ( and hence @xmath8 ) which decay as we evolve forwards in time , and a drift in @xmath186 away from its general relativistic limit @xmath187 .
it is clear that the approximate solution obtained in this section closely mimics the full solution of section [ sec : s10 ] .
in section [ sec:2 ] , we have calculated the ricci scalar for the hss and ab models , and have found that it will undergo non - linear oscillations and will drift away from the general relativistic limit @xmath135 as one goes back in time . from the ricci scalar
we can now calculate the hubble parameter for the two models , using @xmath306 .
we look for a solution of the form @xmath307 , where @xmath308 .
a solution of this form must exist in order for these models be viable ; they must have a matter era for which @xmath309 for example , in order for normal structure formation to take place .
linearizing in @xmath154 , we find the following expression @xmath310 next , by using the time coordinate @xmath174 , where @xmath311 , we obtain @xmath312 where @xmath313 is the general relativistic scale factor .
this expression can be integrated to obtain @xmath154 , @xmath314 where @xmath315 is a constant of integration , and we have taken the slowly varying factor @xmath316 outside the integral , which is a good approximation for @xmath317 . we see that @xmath154 contains two terms ; one describing the oscillations of @xmath122 due to @xmath55 , and a term that goes like @xmath318 .
the oscillatory component is suppressed by a factor of @xmath319 , and hence the oscillations of @xmath8 will not have a significant impact on the hubble parameter .
we note in obtaining @xmath154 , we have not had to specify either @xmath55 or @xmath174 , and hence ( @xmath320 ) is valid for any model for which @xmath321 , including the hss and ab models . in a similar manner , we can also consider the scale factor . by writing @xmath322 and expanding as @xmath323
, we have at linear order @xmath324 next , by introducing the fast time coordinate @xmath174 , we write ( @xmath325 ) as @xmath326 and by using ( @xmath320 ) and integrating , we arrive at the following approximate expression for @xmath149 , @xmath327 where @xmath328 is a constant of integration .
@xmath149 possesses an oscillatory term due to the rapid oscillations of @xmath55 , which is suppressed by a factor of @xmath329 .
we conclude that the rapid oscillations of the ricci scalar has no significant impact on the scale factor for models in which @xmath321 .
in this paper we have solved the gravitational field equations for the hss and ab models numerically and by using an alternative perturbative analysis .
we have shown that the oscillations of the ricci scalar are inherently non - linear , although they can be modeled as linear waves in a certain regime . however , for the ab model this linear regime is particularly restricted .
rather than using the earlier perturbative analysis , we have re - written the trace of the gravitational field equations as a damped , driven , non - linear oscillator . by solving this new equation ,
we have found that the ricci scalar does indeed oscillate with high frequency , as predicted in ref.@xcite . as the amplitude and frequency of these waves grow to the past
, we see non - linear behaviour becoming increasingly important .
we have also found that the ricci scalar does not exactly oscillate around its general relativistic limit , but rather there is a highly suppressed drift .
in section [ sec : s10 ] , we solved the full gravitational field equations for both models numerically . by specifying the initial conditions at late times and evolving backwards
, we observed that in general the ricci scalar would evolve to a singularity after a finite time .
this is another important consequence of the non - linear terms in the field equations , since no such behaviour was observed when using the linearized approach . by using our oscillator equation
, we were able to explain why this singularity occurs , and found that it can be avoided by choosing initial conditions at some early time and evolving forwards .
we were able to derive the allowed initial conditions for both the ab and hss models .
we have found that although the singularity could potentially be evaded with a suitable choice of initial conditions , the resulting ricci scalar is likely to be unstable to perturbations away from a perfect matter era . to obtain a viable modified gravity model ,
it is clear that some method of regularizing this singularity is required .
finally , we have considered the effect of the oscillations of @xmath8 on the hubble parameter and scale factor .
we have found that the oscillatory components of @xmath122 and @xmath123 are suppressed by factors of @xmath319 and @xmath330 respectively , where @xmath331 .
we have concluded that @xmath154 and @xmath149 will remain small , in spite of the potentially large oscillations of @xmath8 .
this conclusion is generic to models which have @xmath332 .
although we studied specifically the hss and ab models in this paper , we believe that many of the results obtained are generic to @xmath0 theories of gravity which satisfy @xmath332 .
we expect that the procedure adopted in section [ sec:2 ] , that is writing the trace of the gravitational field equations as a non - linear oscillator equation , will generalize to all models for which @xmath332 , and if that is the case then the singularity encountered in this paper will be a common feature of @xmath0 models .
sa is supported by a stfc studentship .
we would like to thank a. starobinsky for helpful comments on an earlier draft of the paper . during the preparation of this manuscript we became aware that a. frolov has arrived at similar conclusions @xcite . | we study cosmological expansion in f(r ) gravity using the trace of the field equations .
high frequency oscillations in the ricci scalar , whose amplitude increases as one evolves backward in time , have been predicted in recent works .
we show that the approximations used to derive this result very quickly breakdown in any realistic model due to the non - linear nature of the underlying problem . using a combination of numerical and semi - analytic techniques
, we study a range of models which are otherwise devoid of known pathologies .
we find that high frequency asymmetric oscillations and a singularity at finite time appear to be present for a wide range of initial conditions .
we show that this singularity can be avoided with a certain range of initial conditions , which we find by evolving the models forwards in time .
in addition we show that the oscillations in the ricci scalar are highly suppressed in the hubble parameter and scale factor . |
in the 1960 s ralf hagedorn proposed @xcite that if there was a limiting temperature of the universe , now known as the hagedorn temperature , then the addition of increasingly more energy to a system would no longer increase the temperature but rather create more massive , highly degenerate resonances .
the consequence of this idea was an exponentially increasing mass spectrum @xmath0 summed over the degeneracy , @xmath1 , of the known hadrons . in 2004
@xcite and 2015 @xcite the experimentally measured hadrons from the particle data group @xcite confirmed the continually exponentially increasing mass spectrum , as hagedorn originally suggested .
meanwhile , high energy heavy - ion collisions at rhic and the lhc probed temperatures surpassing hagedorn s original limiting temperature , producing a deconfined state of matter known as the quark gluon plasma .
we now understand , thanks to first principle lattice qcd calculations , that there is a cross - over phase transition @xcite , not a limiting temperature .
in this framework we can understand the hagedorn temperature as roughly equivalent to the critical temperature and then expect the effect of an exponentially increasing mass spectrum to appear close to the phase transition .
indeed , including missing resonances close to the phase transition can affect dynamical chemical equilibrium @xcite , decrease the shear viscosity over entropy ratio @xcite , affect the elliptical flow @xcite , and improve thermal fits @xcite . recent comparisons to lattice quantum chromodynamic calculations
@xcite suggested that there may be missing strange hadrons as calculated from quark model states @xcite due to a mismatch in the strange chemical potential to baryon chemical potential in lattice qcd vs. the hadron resonance gas model from the known pdg spectrum @xcite .
further more , there were suggestions @xcite that missing resonances could account for the @xmath2 vs. strange hadron tension at lhc when it comes to the thermal fits @xcite .
however , in @xcite the decays of the quark model states were not considered , which are necessary for thermal fits .
[ all ] shows the exponentially increasing mass spectra that includes these quark model states , implying that these missing resonances are consistent with hagedorn s original postulate . using the known branching ratios from @xcite
, we extrapolated up the branching ratios of the quark model states taking all quantum numbers into account . following this
, we analyzed the net - proton and net - charge fluctuations ( @xmath3 ) as in @xcite , to extract the corresponding @xmath4 and @xmath5 across energies in the beam energy scan .
the results are shown in fig .
[ sup ] where we find that the addition of the quark model states slightly decreases the freeze - out temperature but overall only has a small affect .
this is consistent with previous studies analyzing the affect of additional resonances on the thermal fits @xcite .
comparison of charged strange mesons from lattice qcd via the partial pressure to the experimentally measured ( preliminary ) kaon fluctuations at star for @xmath6 gev . ]
comparison of charged strange mesons from lattice qcd via the partial pressure to the experimentally measured ( preliminary ) kaon fluctuations at star for @xmath6 gev . ]
an alternative picture was also suggested to resolve the tension between the light and strange hadrons at lhc . in lattice qcd
, the inflection point of susceptibilities can provide clues of about the temperature of hadronization . using the light susceptibility
one finds an inflection point around @xmath7 mev whereas the inflection point for the strangeness susceptibility is around @xmath8 mev @xcite .
considering there is @xmath9 mev difference between light and strange hadrons , a logical consequence of this may be that strange hadrons reach chemical equilibrium at higher temperatures than light hadrons .
if this is true , it would be consistent with the tension between the light and strange hadrons because it would increase the population of strange baryons , which are typically under - predicted using lower temperatures .
this idea is consistent with many dynamical models . in urqmd
there is no specified chemical freeze - out temperature such that each particle species reaches chemical equilibrium on a different time scale @xcite .
similarly , using multi - body hadronic interactions via rate equations , one can also reach chemical equilibrium on different time scales depending on the species .
however , if one can provide directly from first - principle lattice calculations that different chemical equilibration temperatures are needed then it gives significantly more weight to these dynamical models .
furthermore , it will then require the hadronization schemes to be updated uniformly to include different temperatures for light and strange hadrons . in order to extract the chemical freeze - out temperature from the lattice we use ratios of susceptibilities as discussed in @xcite .
in the experiment , the only strange multiplicity fluctuations ( and their corresponding moments of these distributions ) that can be currently measured at the beam energy scan are charged kaons @xcite . however , on the lattice all particles exist as well as their corresponding interactions . in order to extract only the charged kaon contribution
, we implement partial pressures for charged strange mesons as in @xcite @xmath10 where @xmath11 and @xmath12 are supplied from the wuppertal budapest collaboration .
comparison of net-@xmath13 from lattice qcd ( wuppertal - budapest ) to preliminary star net-@xmath13 for @xmath6 gev .
] in fig [ kaons ] a comparison between the lattice qcd kaon partial pressure is shown compared to the preliminary star data for @xmath14 . due to the large error bars on the kaons data , it is not yet possible to extract the strange chemical freeze - out temperature from lattice qcd . however ,
if the error bars are decreased in the future then we will be able to say decisively if there is a splitting between the light and strange chemical equilibration temperatures . in this proceedings ,
we introduce two approaches to study the possible differences in chemical equilibration temperatures between light and strange hadrons .
the first approach is to include additional states predicted from the quark model as well as modeling their decay channels and branching ratios to see if this fixes the tension between light and strange hadrons .
the second approach is a new method to extract the charged , strange susceptibilities from lattice qcd and extract the strange freeze - out temperature from first principles . at this point in time , the error bars on kaon fluctuations are too large to extract the strange freeze - out temperature .
however , if the error bars are decreased in the future we may be able to decidedly settle the tension between light and strange hadrons .
this material is based upon work supported by the national science foundation under grant no .
phy-1513864 and by the u.s .
department of energy , office of science , office of nuclear physics , within the framework of the beam energy scan theory ( best ) topical collaboration .
this work contains lattice qcd data provided by the wuppertal - budapest collaboration .
an award of computer time was provided by the incite program .
this research used resources of the argonne leadership computing facility , which is a doe office of science user facility supported under contract de - ac02 - 06ch11357 .
the work of j. g. and a. p. was supported by the dfg grant sfb / tr55 .
the authors gratefully acknowledge the gauss centre for supercomputing ( gcs ) for providing computing time for a gcs large - scale project on the gcs share of the supercomputer juqueen @xcite at jlich supercomputing centre ( jsc ) .
9 r. hagedorn , nuovo cim .
suppl . * 3 * , 147 ( 1965 ) .
w. broniowski , w. florkowski and l. y. glozman , phys .
d * 70 * , 117503 ( 2004 ) doi:10.1103/physrevd.70.117503 [ hep - ph/0407290 ] .
p. m. lo , m. marczenko , k. redlich and c. sasaki , phys .
c * 92 * , no . 5 , 055206 ( 2015 )
doi:10.1103/physrevc.92.055206 [ arxiv:1507.06398 [ nucl - th ] ] . k. a. olive _ et al . _
[ particle data group collaboration ] , chin .
c * 38 * , 090001 ( 2014 ) .
doi:10.1088/1674 - 1137/38/9/090001 y. aoki , g. endrodi , z. fodor , s. d. katz and k. k. szabo , nature * 443 * , 675 ( 2006 ) doi:10.1038/nature05120 [ hep - lat/0611014 ] .
j. noronha - hostler , c. greiner and i. a. shovkovy , phys .
lett . * 100 * , 252301 ( 2008 ) doi:10.1103/physrevlett.100.252301 [ arxiv:0711.0930 [ nucl - th ] ] . j. noronha - hostler , m. beitel , c. greiner and i. shovkovy , phys .
c * 81 * , 054909 ( 2010 ) doi:10.1103/physrevc.81.054909 [ arxiv:0909.2908 [ nucl - th ] ] .
m. beitel , k. gallmeister and c. greiner , phys . rev .
c * 90 * , no . 4 , 045203 ( 2014 ) doi:10.1103/physrevc.90.045203 [ arxiv:1402.1458 [ hep - ph ] ] .
m. beitel , c. greiner and h. stoecker , phys .
c * 94 * , no . 2 , 021902 ( 2016 ) doi:10.1103/physrevc.94.021902 [ arxiv:1601.02474 [ hep - ph ] ] .
j. noronha - hostler , j. noronha and c. greiner , phys .
lett . * 103 * , 172302 ( 2009 ) doi:10.1103/physrevlett.103.172302 [ arxiv:0811.1571 [ nucl - th ] ] . j. noronha - hostler , j. noronha and c. greiner , phys .
c * 86 * , 024913 ( 2012 ) doi:10.1103/physrevc.86.024913 [ arxiv:1206.5138 [ nucl - th ] ] . g. p. kadam and h. mishra , nucl .
a * 934 * , 133 ( 2014 ) doi:10.1016/j.nuclphysa.2014.12.004 [ arxiv:1408.6329 [ hep - ph ] ] . s. pal , phys . lett .
b * 684 * , 211 ( 2010 ) doi:10.1016/j.physletb.2010.01.017 [ arxiv:1001.1585 [ nucl - th ] ] .
j. noronha - hostler , j. noronha , g. s. denicol , r. p. g. andrade , f. grassi and c. greiner , phys .
c * 89 * , no . 5 , 054904 ( 2014 ) doi:10.1103/physrevc.89.054904 [ arxiv:1302.7038 [ nucl - th ] ] .
j. f. paquet , c. shen , g. s. denicol , m. luzum , b. schenke , s. jeon and c. gale , phys .
c * 93 * , no . 4 , 044906 ( 2016 ) doi:10.1103/physrevc.93.044906 [ arxiv:1509.06738 [ hep - ph ] ] .
s. capstick and n. isgur , phys .
d * 34 * , 2809 ( 1986 ) [ aip conf .
proc . * 132 * , 267 ( 1985 ) ] .
doi:10.1103/physrevd.34.2809 , 10.1063/1.35361 d. ebert , r. n. faustov and v. o. galkin , phys .
d * 79 * , 114029 ( 2009 ) doi:10.1103/physrevd.79.114029 [ arxiv:0903.5183 [ hep - ph ] ] .
a. bazavov _ et al .
_ , phys .
* 113 * , no .
7 , 072001 ( 2014 ) doi:10.1103/physrevlett.113.072001 [ arxiv:1404.6511 [ hep - lat ] ] .
j. noronha - hostler and c. greiner , arxiv:1405.7298 [ nucl - th ] .
j. noronha - hostler and c. greiner , nucl .
a * 931 * , 1108 ( 2014 ) doi:10.1016/j.nuclphysa.2014.08.101 [ arxiv:1408.0761 [ nucl - th ] ] .
r. bellwied , s. borsanyi , z. fodor , s. d. katz and c. ratti , phys .
lett . *
111 * , 202302 ( 2013 )
doi:10.1103/physrevlett.111.202302 [ arxiv:1305.6297 [ hep - lat ] ] . s. a. bass _ et al .
_ , phys .
c * 60 * , 021902 ( 1999 ) doi:10.1103/physrevc.60.021902 [ nucl - th/9902062 ] .
f. karsch , central eur .
j. phys .
* 10 * , 1234 ( 2012 ) doi:10.2478/s11534 - 012 - 0074 - 3 [ arxiv:1202.4173 [ hep - lat ] ] .
s. borsanyi , z. fodor , s. d. katz , s. krieg , c. ratti and k. k. szabo , phys .
lett . * 113 * , 052301 ( 2014 ) doi:10.1103/physrevlett.113.052301 [ arxiv:1403.4576 [ hep - lat ] ] .
j. xu for [ star ] , strangeness in quark matter 2016 proceedings juqueen : ibm blue gene / q supercomputer system at the jlich supercomputing centre juqueen : ibm blue gene / q supercomputer system at the jlich supercomputing centre , tech . rep . 1 a1 ( jlich supercomputing centre , http://dx.doi.org/10.17815/jlsrf-1-18 , 2015 ) . | the precision reached by recent lattice qcd results allows for the first time to investigate whether the measured hadronic spectrum is missing some additional strange states , which are predicted by the quark model but have not yet been detected .
this can be done by comparing some sensitive thermodynamic observables from lattice qcd to the predictions of the hadron resonance gas model ( with the inclusion of decays [ 3 ] ) .
we propose a set of specific observables , defined as linear combinations of conserved charge fluctuations , which allow to investigate this issue for baryons containing one or more strange quarks separately .
applications of these observables to isolate the multiplicity fluctuations of kaons from lattice qcd , and their comparison with the experimental results , are also discussed . |
although details about stellar magnetic field generation remain uncertain , it is accepted that coronal magnetic fields are generated by a magnetic dynamo operating within the convection zone . for main sequence stars ,
the convection zone narrows towards earlier spectral types , disappearing entirely around a spectral type of a5 v. at this point , coronae disappear as well @xcite , emphasizing the central role that the convection zone dynamo plays in generating stellar coronae .
one general empirical goal of stellar coronal x - ray observations is to study how coronae change character as the convection zone narrows .
any clear correlation between coronal properties and spectral type tells us something useful about how the dynamo and the fields it generates change with spectral type , thereby providing potentially crucial information for theoretical models of how stellar magnetic dynamos operate in general .
coronal abundances represent one coronal property that appears to be tightly correlated with spectral type , at least for main sequence stars . on the sun
, it has long been known that coronal abundances are different from photospheric abundances , elements with low first ionization potential ( fip ) having enhanced abundances in the solar corona and solar wind @xcite .
analyses of x - ray and euv spectra have found a similar `` fip effect '' for some stars of low to moderate activity @xcite .
however , for many other stars , particularly extremely active ones , the fip effect is either absent , or sometimes an inverse fip effect is observed , where low - fip elements have coronal abundances that are _ depleted _ relative to the high - fip elements @xcite .
currently the only detailed theoretical model that is capable of explaining both solar - like fip effects and inverse fip effects is one that attributes the element fractionation to ponderomotive forces induced by alfvn and other mhd waves passing through or reflecting from the chromosphere @xcite .
initially , it appeared that coronal abundances were primarily dependent on activity , as the most active stars clearly tend to have inverse fip effects instead of a solar - like fip effect .
this activity dependence of abundances for active stars is illustrated and discussed in detail by ( * ? ? ?
* see figure 37 ) .
the abundance variations are not simply a uniform variation of low - fip elements relative to high - fip ones .
the high - fip elements appear to fractionate with respect to each other as well , as illustrated by measurements of ne / o .
the vast majority of spectroscopically observed stellar coronae exhibit high ratios of @xmath3 @xcite , but relatively inactive stars ( e.g. , the sun , @xmath4 cen ab , and procyon ) show @xmath5 @xcite .
as more abundance measurements have been made , it has become apparent that abundances are also highly dependent on spectral type @xcite . in order to explore this further
we have experimented with a high - fip / low - fip abundance ratio , @xmath6 , which quantifies a corona s fip bias as the average abundance of four high - fip elements ( c , n , o , and ne ) relative to fe ; normalized to photospheric abundance ratios . for normal main sequence stars with coronal x - ray luminosities of @xmath7 ( in ergs s@xmath8 ) ,
there is a surprisingly tight spectral type dependence for @xmath6 @xcite . in this correlation
, early g stars all have a solar - like fip effect , which decreases towards late g and early k stars , reaching no fip effect at all at a spectral type of k5 v. later than k5 v , inverse fip effects are observed , with the magnitude of the effect increasing into the m dwarfs .
we will refer to this as the `` fip - bias / spectral - type '' ( fbst ) relation . at least for stars with @xmath7 ,
any activity dependence for @xmath6 is lost in the scatter of the measurements .
the fbst relation would predict that the solar - like fip effect should become even stronger as you move from solar - like early g stars to f stars .
more precisely , instead of a factor of 4 enhancement of low fip elements in the corona as for the sun @xcite , the fbst relation would predict a factor of 6 enhancement for mid - f stars @xcite .
however , observations of two mid - f dwarfs seem to contradict this expectation .
the first of these is procyon ( f5 iv - v ) , the nearest and most frequently observed f star , which has been observed spectroscopically by euve , _ chandra _ , and xmm .
all of these spectra demonstrate that procyon s corona exhibits no significant fip bias at all .
its coronal abundances are consistent with its photospheric abundances @xcite .
more recently , using xmm data @xcite have also found little difference in the abundances of low - fip and high - fip elements in the corona of @xmath2 boo ( f7 v ) , although they suggest that _ both _ are depleted in the corona by similar amounts .
the procyon and @xmath2 boo measurements would seem to suggest that the fbst relation falls apart for the f dwarfs , possibly an effect of the thinning convection zones of such stars .
this would be a very interesting result , but there are potential issues with both procyon and @xmath2 boo regarding whether either can be considered representative of mid - f dwarfs .
for procyon , the problem is simply that it is not a pure main sequence star , as its `` iv - v '' luminosity class suggests .
its radius of 2.05 r@xmath9 is certainly larger than that of a true mid - f main sequence star @xcite . for @xmath2
boo , the problem is the presence of two very different companions , one stellar and one planetary . the stellar companion , @xmath2 boo b , is an m2 v star that is unresolved from the primary in the xmm observation , which could in principle be contributing x - ray photons to the spectrum . given that m dwarfs are expected to have inverse fip effects @xcite , contamination from @xmath2 boo b would be expected to weaken any fip effect in the combined x - ray spectrum .
as for the planetary companion , it is a very well - studied close - in giant planet , @xmath2 boo b , with an orbital period of only 3.31 days .
there is already evidence that the presence of this massive planet is producing coronal activity on the star @xcite .
is it possible that the planet could somehow be suppressing the expected strong coronal fip effect as well ?
this is an exciting possibility , which makes it all the more important to know what the corona of a truly normal mid - f dwarf looks like . to that end
, we here present an analysis of recent _ chandra _ observations of the nearby f6 v star , @xmath0 ori , which lies at a distance of 8.1 pc @xcite .
the star has a rapid rotation rate of @xmath10 km s@xmath8 , corresponding to @xmath11 days @xcite .
not surprisingly , given the known connection between stellar rotation and activity , @xmath0 ori also has a relatively high x - ray luminosity of @xmath12 @xcite , just barely within the @xmath7 activity regime mentioned above where the fbst relation remains tight .
this is in fact one of the ten most luminous coronal x - ray sources within 10 pc , so it is curious that @xmath0 ori has never before been observed spectroscopically in x - rays .
the @xmath0 ori spectrum promises to resolve the issue of what the coronal characteristics of moderately active mid - f main sequence stars really are .
the _ chandra _ x - ray telescope observed @xmath0 ori twice in late 2010 , the first time on 2010 november 11 for 57.8 ksec , and a second time on 2010 november 23 for 20.1 ksec .
the observations were performed with _
chandra _ s letgs configuration , combining the low energy transmission grating ( letg ) and the component of the high resolution camera detector used for spectroscopy ( hrc - s ) .
the data were processed using version 4.3 of the _ chandra _ team s ciao software .
the initial hrc - s image produced by the data processing consists of a zeroth - order image of the star , and an x - ray spectrum dispersed on both sides of the image , i.e. , a plus and minus order spectrum .
further processing extracts one - dimensional spectra from the image , and then the plus and minus order spectra are coadded into a final spectrum covering the full @xmath13 wavelength region of letgs .
photons in the time - tagged zeroth - order image are useful for assessing source variability .
we find no significant variability at all for @xmath0 ori , and the x - ray fluxes appear to be identical at the two observing times , so we simply coadd the spectra taken at the two times into a final spectrum representing a combined exposure time of 77.9 ksec .
this spectrum is shown in figure 1 .
we use version 6.1 of the chianti atomic database to identify emission lines in our spectrum @xcite .
the identified lines are listed in table 1 , and identifications are also provided in figure 1 .
counts for the detected lines are measured by direct integration from the spectrum , and these raw counts are then converted to photon fluxes .
both the counts and fluxes of the lines are provided in table 1 , together with 1@xmath14 uncertainties .
the third column of the table lists line formation temperatures , based on ionization equilibrium computations from @xcite .
there are many lines in table 1 with large quoted uncertainties that represent questionable detections .
the emission measure ( em ) analysis described in section 2.3 is used to confirm _ a posteriori _ that these marginal detections are at least plausible .
there are many line blends in table 1 . the em analysis and the line strengths in the chianti database
are used to assess what lines are contributing significant flux to each emission feature .
the line blends involving different species are ultimately ignored in the final em analysis .
there are many density sensitive he - like triplets in the @xmath0 ori spectrum @xcite , but only the o vii triplet near 21.8 is sufficiently free of blends and with high enough signal - to - noise ( s / n ) for a decent density measurement .
unfortunately , even for o vii we find we can only quote an upper limit for the electron density . based on the collisional equilibrium models of @xcite , the @xmath15 line flux ratio corresponds to an electron density ( in @xmath16 ) of @xmath17 .
similarly low coronal densities have been found for many moderately active g and k dwarfs in the past @xcite , so there is no evidence that our moderately active f dwarf is any different from later type stars in terms of coronal density .
inferring coronal abundances and temperature distributions from line flux measurements requires an em analysis , which is a nontrivial inversion problem . for this calculation
, we use version 2.6 of the pintofale software package @xcite , which was developed to solve the em inversion problem using a markov chain monte carlo approach
. the analysis also makes use of chianti line emissivities @xcite .
our em analysis closely mirrors previous work , as described in particular detail by @xcite , so we refer the reader to that paper for details about this computation .
the analysis considers the effects of interstellar absorption on the observed line fluxes .
there is no measured h i column density towards @xmath0 ori itself , but there is a measurement of @xmath18 ( in @xmath19 ) towards @xmath20 ori @xcite , which is roughly at the same distance and in the same direction as @xmath0 ori , so this is the value we assume .
we note , however , that this value is low enough that interstellar absorption does not have a large effect on line fluxes .
the line fluxes by themselves allow the computation of the shape of the emission measure distribution , and the computation of relative coronal abundances .
for the relative abundances , the natural reference element to use is fe , considering the large number of fe lines available in the letgs spectrum . in the initial em analysis
we simply assume a solar photospheric abundance for our reference element , @xmath21}=-4.50 $ ] @xcite .
table 2 lists the coronal abundances relative to fe computed in the em analysis , and figure 2 shows the derived emission measure distribution .
the 90% confidence intervals shown for the emission measures and abundances in figure 2 and table 2 , respectively , are indicative of the uncertainties in solving the inversion problem , based on the magnitude of the uncertainties in the measured line fluxes .
properly normalizing the em distribution requires the measurement of an absolute coronal [ fe / h ] ratio , which also allows the relative abundances in table 2 to be converted to absolute abundances .
measurement of [ fe / h ] requires an assessment of the line - to - continuum ratio in the letgs spectrum , which can be done after the initial em analysis is completed .
emission measures scale inversely with the assumed [ fe / h ] , since higher metal abundances mean less emission measure is required to account for the line fluxes .
thus , increasing [ fe / h ] by a factor of two from @xmath21}=-4.50 $ ] to @xmath21}=-4.20 $ ] , for example , uniformly lowers the em distribution by a factor of two .
synthetic line - plus - continuum spectra can be computed from the em distribution for various values of [ fe / h ] and compared with the data , in order to see which leads to the best match to the observed continuum level .
this is illustrated in figure 3 , which shows the best fit for @xmath21}=-4.32 $ ] .
this synthetic spectrum is also compared with the data in figure 1 . in letgs data ,
higher order spectra are superimposed onto the first order spectrum , so the synthetic spectrum accounts for the higher order spectra as well , specifically orders 25 .
the higher order contributions are explicitly plotted in both figures 1 and 3 .
the em distribution in figure 2 is the final normalized distribution assuming @xmath21}=-4.32 $ ] .
the most striking feature of the @xmath0 ori em distribution in figure 2 is a sharp peak at @xmath1 .
this peak actually seems to be very common for moderately active stars like @xmath0 ori .
many examples are provided by @xcite , but the best of these are @xmath22 boo a ( g8 v ) and @xmath23 eri ( k2 v ) .
the em distributions for these 2 stars are also shown in figure 2 .
the @xmath22 boo a em distribution is a particularly good match for that of @xmath0 ori .
this is perhaps not surprising considering that @xmath22 boo a and @xmath0 ori have similar x - ray luminosities of @xmath24 and @xmath12 , respectively @xcite , although the spectral types of the two stars are quite different .
thus , there is no evidence that the thinner convection zone of @xmath0 ori has led to coronal temperatures any different from those of similarly active later type stars .
interestingly enough , the @xmath1 peaks are also commonly observed for solar active regions @xcite . in figure 2 ,
the three stellar em distributions are compared with a solar one , which represents the average of 15 active region em distributions from @xcite , who emphasized the prevalence of @xmath1 em peaks on the sun .
the shape of the solar active region em agrees beautifully with those of @xmath0 ori and @xmath22 boo a for @xmath25 , although the three stellar em distributions in figure 2 have much higher em at @xmath26 . @xcite
have previously noted the similarity between the em distributions of solar active regions and those of @xmath22 boo a and @xmath23 eri , and have suggested that it implies that moderately active stars like these are covered almost completely with solar - like active regions .
is it possible that these seemingly ubiquitous @xmath1 peaks are somehow a systematic artifact of the pintofale computations that produced them ?
not only was pintofale s markov chain monte carlo approach used to compute the stellar em distributions in figure 2 , but it was also used by @xcite to compute the solar ones .
note , however , that @xcite analyzed a completely different spectral region and a different set of emission lines than those used in the stellar analyses , suggesting that the @xmath1 peaks are not just a product of working with a particular set of lines .
we explore this issue further using simulations .
these simulations involve taking a known em distribution , computing line fluxes from that distribution for the set of lines in table 2 , and then plugging those line fluxes back into pintofale to see if the em inversion procedure can recover the original em distribution .
uncertainties in the line fluxes are assumed to be the same as the relative uncertainties measured for the @xmath0 ori lines in table 1 .
some results of these simulations are shown in figure 4 .
the dotted line in figure 4 s `` simulation 1 '' is simply the @xmath0 ori distribution from figure 2 .
the solid line shows the result when that is used as the input em distribution for the simulation . the computed em distribution agrees encouragingly well with the original distribution , being within the displayed error bars for 15 of the 17 temperature bins .
most importantly , the simulation demonstrates pintofale s ability to recover the sharp @xmath1 em peak . in simulations 2 and 3
we simply shift the peak to @xmath27 and @xmath28 , respectively , and in simulation 4 we remove the peak entirely , assuming equal em values in the @xmath29 range .
although the reconstructed em values are still generally within error bars , we do note that for simulations 24 the em is systematically overpredicted at @xmath1 and underpredicted at @xmath27 and @xmath28 .
we conclude , therefore , that there may indeed be a modest tendency for pintofale to overemphasize the height and narrowness of the @xmath1 peaks , though the magnitude of this effect is not necessarily inconsistent with the inferred 90% confidence intervals .
we can only speculate as to the source of this tendency .
it does so happen that @xmath1 is the central formation temperature of the strongest lines in the letgs spectrum , the fe xvii lines between 1517 , so perhaps the need to fit those line fluxes particularly well leads to the modest overemphasis of em at @xmath1 , to the detriment of em at @xmath27 and @xmath28 . in any case , despite evidence for this small systematic error , the simulations provide support for the existence of a distinct maximum of em near @xmath1 for moderately active stars and solar active regions .
the question then becomes , why is @xmath1 such a popular temperature for both a variety of solar active regions , and for moderately active stars of various spectral types ?
is there something special about this particular temperature ?
one notable thing about it is that there is a local minimum in the radiative cooling curve near @xmath1 ( e.g. , * ? ? ?
* ; * ? ? ?
the cooling curve relates how emissivity varies with temperature .
plasma is stable against radiatively driven thermal instabilities where the cooling curve has a shallower slope than the temperature dependence of the heating function .
the cooling curve has a steep negative slope for a variety of abundance sets in the range @xmath30 , which could imply thermal instability unless the heating curve has an even steeper negative slope at those temperatures .
thus , @xmath1 is possibly the coolest stable temperature above @xmath31 .
plasma initially between these two temperatures will most likely either heat up or cool down until a stable temperature is reached .
perhaps the ubiquity of the @xmath1 peak implies something about the commonality of active region loops on the sun and solar - like main sequence stars , consistent with the idea that coronal emission from moderately active stars like @xmath0 ori and @xmath22 boo a is dominated by solar - like active regions that largely cover the stellar surface @xcite . on the sun ,
such loops are often found to have densities of @xmath32 @xmath16 ( e.g. , * ? ? ?
such densities are consistent with those commonly measured in stellar coronae as well , at least at temperatures of @xmath33 @xcite .
an explanation in terms of thermal stability presupposes that coronal plasma exists in structures ( loops ) that can support a quasi - steady state .
no such explanation would be possible in a corona heated and supplied by more dynamic phenomena such as type ii spicules ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) . in simple static loop models
there is a relation between a coronal loop s length and its apex temperature and density @xcite .
thus , if coronal loops in various solar and stellar contexts tend to have similar densities and lengths , then presumably they would have similar temperatures as well .
this presumes , of course , that these simple static models apply , which is questionable ( e.g. , * ? ? ?
* ; * ? ? ?
as discussed in section 2.3 , the em analysis provides coronal abundances measured relative to fe for many elements .
all of these measurements , in the form @xmath34}$ ] , are listed in the second column of table 2 .
we are interested in comparing coronal abundances with those of the photosphere .
thus , in the third column of table 2 , we list stellar photospheric abundance ratios relative to fe , in the form @xmath34}_*$ ] .
the photospheric abundances of @xcite are measured using line - by - line comparisons of solar and stellar spectral features , so the abundance measurements are intrinsically relative to solar abundances .
thus , computing the stellar photospheric abundance ratios in table 2 requires the assumption of solar abundances .
the assumed solar abundances of @xcite are listed in the last column of table 2 . for n and s ,
there are no stellar photospheric measurements available .
we assume these elements behave like high - fip elements , such as o. thus , we assume @xmath35}\equiv \log{\rm [ s_*/s_{\odot}]}\equiv \log{\rm [ o_*/o_{\odot}]}=0.16 $ ] .
neon is a special case .
the fundamental problem with ne is that even for the sun there is no real photospheric abundance measurement , because there are no solar photospheric ne absorption lines .
reference solar abundances for ne , such as the one listed in table 2 , are in reality based on measurements from the solar corona and transition region ( e.g. , * ? ? ?
we know very well that abundances can be different from photospheric in the corona , but not necessarily in the transition region , where @xcite finds a low ne / o abundance , and no evidence for a fip effect in other elements .
nevertheless , we do know that high - fip elements can be fractionated in the solar atmosphere , as he is depleted by a factor of 2 in the slow solar wind @xcite .
this depletion of he ( and also possibly of ne ) is also a feature of models for the solar fip effect based on the ponderomotive force @xcite . complicating the issue
further is the knowledge that moderately active and very active stars seem to univerally have coronal ne / o ratios much higher than that of the sun and other similarly inactive stars @xcite , raising the possibility that these stellar coronal ne abundances may be a better measure of the true cosmic ne abundance than the solar coronal ne abundance usually referenced as such .
we here follow the same convention as in our past papers and assume that the average stellar coronal @xmath36 abundance ratio of @xcite is indicative of the true cosmic abundance of ne , and therefore applies to solar and stellar photospheres as well , including for @xmath0 ori .
thus , the @xmath37}_*$ ] ratio in table 2 is derived assuming @xmath36 , instead of the much lower solar ratio of @xmath38 .
the validity of this assumption is very much debatable , but it does lead to coronal ne abundances in better agreement with those of other high - fip elements ( see figure 5 ) . after subtracting the logarithmic photospheric abundance ratios , the logarithmic coronal abundance ratios in table 2 are in figure 5a plotted versus fip .
the low - fip elements are all roughly consistent with the reference low - fip element , fe .
the high - fip element abundances are lower by about @xmath39 dex , consistent with a fip effect nearly identical to that of the sun .
the average value of the four high - fip elements in figure 5a is plotted explicitly as a horizontal dot - dashed line in the figure .
this is the `` fip bias '' quantity , @xmath6 , that we have used in the past to reduce coronal abundances to a single number , which is then used to study the fbst relation described in section 1 @xcite . for @xmath0 ori , @xmath40 .
the @xmath6 quantity has no clear variation with activity for main sequence stars with @xmath7 ( see figure 4 in * ? ? ?
* ) , but it is worth noting that of the four high - fip elements that go into computing @xmath6 ( c , n , o , and ne ) , ne is once again an exception , as we know that the coronal ne / o ratio does vary within the @xmath7 sample , suggesting activity - dependent fractionation of ne even for less active stars @xcite . as alluded to above , the vast majority of spectroscopically observed @xmath7 stars , including @xmath0 ori , have @xmath3 .
only a few truly inactive stars ( the sun and @xmath4 cen ab ) have @xmath5 @xcite .
there are a couple characteristics of the low - fip elements that are worthy of note .
@xcite found that the coronal ni abundance of the mid - f dwarf @xmath2 boo was anomalously high compared to other low - fip elements , being a factor of @xmath41 higher .
we do not see this behavior for our mid - f dwarf , @xmath0 ori . the ni abundance may be a little high in figure 5 , but it is consistent with the other low - fip elements within the error bars . as for mg and si , in the past we had noted a tendency for mg abundances to be higher than si abundances in the coronae of many moderately active stars @xcite .
this is not the case for @xmath0 ori , though , which if anything shows the opposite behavior , consistent with what @xcite found for @xmath2 boo .
figure 5a shows the coronal abundance measurements relative to fe . but
using the @xmath21}=-4.32 $ ] measurement from the line - to - continuum analysis , we can compute absolute abundances as well , and in figure 5b the absolute abundances are plotted , relative to stellar photospheric abundances as before .
this figure casts doubt on whether the coronal abundance fractionation pattern of @xmath0 ori is truly solar - like or not .
based on figure 5a , the @xmath0 ori coronal abundances were described above as very solar - like , because the @xmath40 measurement implies a factor of 4 underabundance of high - fip elements relative to low - fip elements in the corona , similar to the situation in the solar corona .
but in the solar corona this is really due to an enhancement of low - fip elements , as opposed to a depletion of high - fip elements , meaning that for the sun the high - fip elements would be near 0.0 in figure 5b and the low - fip elements would be near 0.6 .
in contrast , figure 5b would imply that for @xmath0 ori , the @xmath40 result is due to fractionation of _ both _ low - fip and high - fip elements , and is in fact mostly due to a depletion of high - fip elements .
this is actually consistent with what we have found before for other moderately active stars @xcite .
the question is whether this means that the fractionation behavior of these stars really is significantly different from the sun , or whether there is a systematic error in the line - to - continuum analysis that leads to systematic underestimates of @xmath21}$ ] .
the history of solar coronal abundance measurements provides reason for caution with regards to the accuracy of the line - to - continuum analysis , as many such solar analyses have found the data more consistent with a high - fip depletion than a low - fip enhancement @xcite , contradicting other evidence , including direct solar wind measurements , that suggest low - fip enhancement is the primary effect @xcite .
one concern is the possible presence of weak metal lines not listed in atomic line databases .
these lines could collectively produce a significant pseudo - continuum in x - ray spectra , leading to overestimates of the strength of the real continuum , corresponding to an underestimate of @xmath21}$ ] .
there is hope that modern databases such as chianti with more extensive line lists have greatly reduced this problem , but it is difficult to quantify what the systematic uncertainties really are .
another potential issue is the effect of non - maxwellian distributions on the line - to - continuum analysis .
non - maxwellian distributions will affect the analysis of both the line and continuum fluxes , but in different ways @xcite .
all these issues provide justification for relying primarily on relative , rather than absolute abundances .
the latter introduces the uncertain systematic errors involved in the line - to - continuum analysis , while the former does not .
nevertheless , when using relative abundances , and quantities derived from them such as @xmath6 , it is important to keep in mind that the issue of absolute abundances has been left unresolved .
a negative @xmath6 value , for example , could indicate either a low - fip enhancement or a high - fip depletion , and two stars with identical @xmath6 values could in principle be different in that respect .
that being said , it is hard to imagine how the rather tight fbst relation shown in figure 6 ( and discussed in detail in the next section ) could exist without some consistency regarding how @xmath6 relates to absolute abundances .
for example , the fbst relation suggests m4 v stars all have @xmath42 @xcite .
it is hard to imagine why these stars would have different degrees of low - fip depletions and/or high - fip enhancements , but always calibrated so that @xmath43 .
as described in section 1 , one of the main goals of the @xmath0 ori analysis is to see whether the fbst relation extends to f spectral types .
the answer is provided in figure 6 .
a full list of the stars plotted in the figure is provided by @xcite , but we now add the @xmath0 ori data point .
based on the g , k , and m stars in the figure , @xmath0 ori would be expected to have @xmath44 , potentially the strongest solar - like fip effect ever observed .
instead it has @xmath40 , comparable to the sun and other early g stars .
there are three possible interpretations for the higher - than - expected @xmath6 measurement for @xmath0 ori : 1 .
simply taken at face value , the @xmath0 ori measurement indicates that the fbst relation flattens towards spectral types earlier than g. observations of early f dwarfs would be very useful to verify this flattening .
if real , the flattening of the fbst relation could be interpreted as an interesting manifestation of the narrowing convection zone .
although somewhat higher than expected , it could be argued that the @xmath0 ori measurement is still close enough to the expected @xmath44 value to be considered consistent with a near - linear fbst relation , considering the amount of scatter that one sees in figure 6 .
perhaps @xmath0 ori is actually too active to precisely follow the fbst relation .
it is known that stars with @xmath45 lie significantly above the fbst relation @xcite . with @xmath12 @xcite ,
@xmath0 ori is close to this limit .
if the @xmath46 limit is actually spectral type dependent , and lower for f dwarfs , @xmath0 ori would be expected to lie somewhat above the fbst curve . in section 1 , there were two f dwarfs discussed that are known from previously published work to be inconsistent with the fbst relation : procyon ( f5 iv - v ) and @xmath2 boo ( f7 v ) . based on measurements from @xcite and @xcite ,
we compute values of @xmath47 and @xmath48 for procyon and @xmath2 boo , respectively .
these values are plotted in figure 6 , explicitly showing just how inconsistent these two stars are with both @xmath0 ori and the fbst relation . as for procyon ,
the problem is that it is not a pure main sequence star , as its `` iv - v '' luminosity class and large 2.05 r@xmath9 radius demonstrate @xcite .
not only does procyon have coronal abundances very different from @xmath0 ori , but it has a much cooler corona as well , with an em peak near @xmath49 and no significant em above @xmath50 @xcite .
cool coronae are associated with less active stars , but procyon s x - ray luminosity is @xmath51 , only a factor of 3 lower than @xmath0 ori , and comparable to many of the stars in the figure 6 sample with hotter coronae .
for example , @xmath23 eri has a coronal x - ray luminosity of @xmath52 @xcite , lower than procyon , but it has an em distribution similar to @xmath0 ori ( see figure 2 ) .
we believe that procyon s inconsistency with @xmath0 ori emphasizes just how sensitive coronal properties can be to luminosity class .
procyon simply can not be considered to be a main sequence star , and it is dramatically inconsistent with the fbst relation as a consequence .
as for @xmath2 boo , the inconsistency of @xmath2 boo with @xmath0 ori strengthens the argument made in section 1 that @xmath2 boo s high @xmath6 value is due to one of its two companions . either the m2 v companion , @xmath2 boo b , is contributing substantial flux to the xmm exposure , leading to erroneous abundance measurements , or the close - in giant planet , @xmath2 boo b , is somehow affecting the coronal abundances .
the latter possibility is easily the more interesting of the two , as it would add a new twist to the current hot topic of star - planet interaction .
1 . spectral type .
this is the fbst relation illustrated in figure 6 .
there are numerous basic stellar , photospheric , and convection zone properties that are known to be tightly correlated with spectral type for main sequence stars ( stellar radius , temperature , mass , surface gravity , convection zone depth , etc . ) .
the existence of the fbst relation demonstrates that somehow one or more of these characteristics also determines coronal abundances .
it is worth noting that a spectral type dependence is also known to exist for t tauri stars and very active main sequence stars @xcite .
2 . activity and/or rapid rotation .
the fbst relation seems to work very well for the vast majority of main sequence stars .
however , there is clearly an activity dependence as well .
this is most evident for extremely active stars with @xmath45 , which all lie significantly above the fbst relation in figure 6 @xcite .
this activity dependence is also clearly seen in figure 37 of @xcite .
an indication that activity dependence of coronal abundances extends to less active stars is the difference in ne / o seen between moderately active stars and truly inactive stars like the sun @xcite .
for the activity extremes , it is possible that some aspect of high activity ( e.g. , high magnetic field strength ) is affecting coronal abundances .
but it is also possible that the underlying cause of the extreme activity , rapid rotation , is the true determining factor .
perhaps very rapid rotation affects photospheric and/or convection zone characteristics , thereby allowing coronal abundances to be altered in the same manner as spectral type affects abundances in the fbst relation .
3 . luminosity class .
the best example of this is the @xmath0 ori / procyon dichotomy discussed in section 3.1 . there are evolved stars other than procyon that would lie above the fbst relation in figure 6 , but most are part of active binaries @xcite .
for such stars , it is not clear whether the discrepancy from the fbst relation is truly due to luminosity class , or whether it is instead an effect of high activity or in some cases particularly close binarity .
is there a common theme that can be found among these abundance determinants ?
one avenue to explore concerns photospheric oscillation properties , which are known to depend on both spectral type and luminosity class ( # 1 and # 3 above ) , and may depend on rotation and activity as well ( # 2 ) .
if coronal abundance fractionation is controlled by alfvn and other mhd waves traveling through the chromosphere , as suggested by @xcite , the photospheric oscillation characteristics might be expected to be crucial , since these oscillations are likely sources of many of the chromospheric waves .
oscillation frequencies increase and amplitudes decrease towards later type stars on the main sequence @xcite , which have higher @xmath6 values .
this would imply a potential connection between high @xmath6 and oscillations with high frequency and/or low amplitude .
unfortunately , this hypothetical connection is not consistent with the @xmath0 ori / procyon dichotomy .
oscillation frequencies decrease and amplitudes increase as stars evolve off the main sequence @xcite , so the aforementioned main sequence oscillation/@xmath6 connection would predict _ lower _ @xmath6 for procyon compared with @xmath0 ori .
this is inconsistent with what we observe , so photospheric oscillation properties can not be the sole determinant of coronal abundances . [ [ a - possible - theoretical - framework - for - resolving - the - coronal - abundance - problem ] ] a possible theoretical framework for resolving the coronal abundance problem ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ turning to theoretical considerations , we outline a framework for trying to interpret the observed coronal abundance variations in solar - like stars . @xcite and
@xcite argue that the alfvn waves giving rise to the ponderomotive force that generates the solar fip effect most plausibly have a coronal origin .
this is because waves in resonance with a coronal loop having period @xmath53 , where @xmath54 is the coronal loop length and @xmath55 is the coronal alfvn speed , are required to produce the observed coronal depletion of he , and the only way to guarantee such a matching is if processes in the corona ( e.g. , nanoflares ) excite normal modes of the loop . @xcite
present a `` toy model '' of the inverse fip effect , where fast mode waves propagating up through the chromosphere from below are reflected back down again in the region of the chromosphere where the alfvn speed is increasing with height .
this wave population gives a ponderomotive force directed _ downwards _ , which competes with the _ upwards _ ponderomotive force induced by the coronal alfvn waves propagating downwards and reflecting back up .
we argue that at the f and g dwarf end of the fbst relation , the positive fip effect from coronal alfvn waves dominates , while at the m dwarf end the inverse fip effect due to initially upward propagating fast mode waves dominates , with a transition between the two somewhere in the k dwarfs .
why should this be so ?
we discuss first the coronal alfvn waves .
these are assumed to arise from coronal nanoflare reconnection events .
several authors ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) have argued that at least some , and maybe even most , of the magnetic energy released in reconnection should be converted to kinetic energy or waves in the surrounding plasma .
it is plausible that such wave generation explains why surveys to find localized hot plasma as evidence of nanoflare reconnection have generally been unsuccessful ( e.g. , * ? ? ?
instead , energy goes from magnetic field to waves , and is thus gradually dissipated as heat throughout the corona , and not quickly and locally as might have been expected .
@xcite studied the efficiency of electron acceleration in reconnection with the ambient plasma @xmath56 ( the electron plasma @xmath57 , to be more precise ) .
they found maximum energy input to electrons at @xmath58 , with reduced electron heating at higher @xmath59 , or lower magnetic field .
we suggest that at the left hand side of the fbst relation , coronal reconnection primarily generates aflvn waves that end up causing positive fip fractionation when they reflect from the chromosphere .
as one moves to the right , to later spectral type , coronal reconnection puts more energy into electrons , and less into waves .
consequently , the positive fip effect diminishes .
conversely , it is also possible that the amplitudes of fast mode waves penetrating the chromosphere from below and reflecting back down again may increase toward later spectral types , thus increasing the inverse fip effect . even though p - mode amplitudes decrease towards later type stars @xcite , it is possible that later type stars could be more efficient at converting the p - mode oscillations to fast mode mhd waves .
different regions of the sun are different in this respect .
it is known that sunspots are sinks of p - mode energy ( see * ? ? ?
* and references therein ) , most likely through mode conversion or resonant absorption to alfvn or fast mode waves .
there are many cases of sunspots , active regions , and flares observed to have reduced fip effects ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) , although it is useful to note that an inverse fip effect has never been reported in any solar observation .
an important factor that could act to restrict inverse fip effects to later spectral types is the expansion of the magnetic field within the chromosphere . in their `` toy model , ''
@xcite found that the density and magnetic field scale heights , @xmath60 and @xmath61 , were required to satisfy @xmath62 to yield an inverse fip effect .
although more realistic models are required to refine this relation , we note that @xmath60 should be smaller for later spectral types due to higher surface gravities , allowing the relation to be met more readily if @xmath61 does not also systematically decrease with spectral type .
it is also worth noting that this relation will be met more readily for extremely active stars with magnetic fields so crowded that the fields can not expand in the chromosphere , which effectively means an increase in @xmath61 .
this could potentially explain the tendency towards high @xmath6 for extremely active stars ( dependence # 2 in section 3.2 ) .
the @xmath0 ori / procyon dichotomy ( see section 3.1 ) remains difficult to explain within this framework .
procyon would have larger @xmath60 than @xmath0 ori , which would bias the star towards lower @xmath6 , not higher ( as observed ) , assuming magnetic field expansion is the central issue .
if @xmath61 increases by much more than @xmath60 as one moves from the main sequence to more evolved stars , that would reverse this behavior , but we know of no physical reason to expect this to be the case .
perhaps procyon has a fundamentally different coronal heating mechanism than the stars that follow the fbst relation .
we have argued in section 2.4 that the commonly observed emission measure spike at @xmath63 would be inconsistent with a type ii spicule model of coronal heating and mass supply , if interpreted in terms of a thermal stability argument .
but procyon does not possess this peak ( see section 3.1 ) , so a spicule - based coronal heating picture might work for procyon .
such a model could also be consistent with procyon s lack of a fip effect , because the spicule flow speed up through the chromosphere would conceivably be too fast to allow fractionation to occur . in this picture , the lower surface gravity of procyon would allow upward propagating spicules to reach higher altitude and presumably higher temperature than on main sequence stars , potentially allowing the spicules to dominate the rather cool coronal emission observed from the star .
drake et al .
( 1995 , their subsection 6.2 ) previously argued that the absence of fip effect in the corona of procyon is due to the effect of `` unresolved fine structures '' extending to higher temperatures there than they do on the sun .
these speculations provide a framework for trying to understand how coronal abundances depend on stellar parameters .
we hope this discussion can serve to focus attention on the critical observational and theoretical points required to subject it to further scrutiny .
we would like to thank harry warren for providing us with the solar active region em distribution in figure 2 , and the referee manuel gdel for helpful comments .
support for this work was provided by nasa through atp award nnh11aq23i and chandra award number go1 - 12012z issued by the chandra x - ray center ( cxc ) .
allende prieto , c. , barklem , p. s. , lambert , d. l. , & cunha , k. 2004 , a&a , 420 , 183 ammler - von eiff , m. , & reiners , a. 2012 , a&a , 542 , a116 arnaud , m. , & rothenflug , r. 1985 , a&as , 60 , 425 asplund , m. , grevesse , n. , sauval , a. j. , & scott , p. 2009
, ara&a , 47 , 481 .
audard , m. , behar , e. , gdel , m. , raassen , a. j. j. , porquet , d. , mewe , r. , foley , c. r. , & bromage , g. e. 2001 , a&a , 365 , l329 audard , m. , gdel , m. , sres , a. , raassen , a. j. j. , & mewe , r. 2003 , a&a , 398 , 1137 ball , b. , drake , j. j. , lin , l. , kashyap , v. , laming , j. m. , & garca - alvarez , d. 2005 , apj , 634 , 1336 braun , d. c. 1995 , apj , 451 , 859 brinkman , a. c. , et al .
2001 , a&a , 365 , l324 chaplin , w. j. , houdek , g. , appourchaux , t. , elsworth , y. , new , r. , & toutain , t. 2008 , a&a , 485 , 813 chaplin , w. j. , elsworth , y. , isaak , g. r. , miller , b. a. , & new , r. 2000 , mnras , 313 , 32 colgan , j. , abdullah , j. jr . , sherrill , m. e. , foster , m. , fontes , c. j. , & feldman , u. 2008 , apj , 689 , 585 de pontieu , b. , et al . 2011 , science , 33 , 55 dere , k. p. , landi , e. , mason , h. e. , monsignori fossi , b. c. , & young , p. r. 1997 , a&as , 125 , 149 dere , k. p. , landi , e. , young , p. r. , del zanna , g. , landini , m. , & mason , h. e. 2009 , a&a , 498 , 915 drake , j. f. , swisdak , m. , che , h. , & shay , m. a. 2006 , nature , 443 , 553 drake , j. j. , laming , j. m. , & widing , k. g. 1995 , apj , 443 , 393 drake , j. j. , laming , j. m. , & widing , k. g. 1997 , apj , 478 , 403 drake , j. j. , peres , g. , orlando , s. , laming , j. m. , & maggio , a. 2000 , apj , 545 , 1074 drake , j. j. , & testa , p. 2005 , nature , 436 , 525 dudk , j. , dzifkov , e. , karlick , m. , & kulinov , a. 2011 , a&a , 529 , a103 dudk , j. , kaparov , j. , dzifkov , e. , karlick , m. , & mackovjak , . 2012 , a&a , 539 , a107 feldman , u. , & laming , j. m. 2000 , phys .
, 61 , 222 feldman , u. , widing , k. g. , & lund , p. a. 1990 , apj , 364 , l21 fludra , a. , & schmelz , j. t. 1995 , apj , 447 , 936 gdel , m. 2004 , a&ar , 12 , 71 gdel , m. 2007 , living rev . sol .
phys . , 4 , 3 ( http://www.livingreviews.org/lrsp-2007-3 ) gdel , m. , et al .
2001 , a&a , 365 , l336 gdel , m. , skinner , s. l. , melnikov , s. y. , audard , m. , telleschi , a. , & briggs , k. r. 2007 , a&a , 468 , 529 huenemoerder , d. p. , canizares , c. r. , drake , j. j. , & sanz - forcada , j. 2003 , apj , 595 , 1131 huenemoerder , d. p. , canizares , c. r. , & schulz , n. s. 2001 , apj , 559 , 1135 jordan , c. 1980 , a&a , 86 , 355 judge , p. g. , de pontieu , b. , mcintosh , s. w. , & kosovare , o. 2012 , apj , 746 , 158 kashyap , v. & drake , j. j. 1998 , apj , 503 , 450 kashyap , v. , & drake , j. j. 2000 , bull .
india , 28 , 475 kervella , p. , thvenin , f. , morel , p. , berthomieu , g. , bord , p. , & provost , j. 2004 , a&a , 413 , 251 kigure , h. , takahashi , k. , shibata , k. , yokoyama , t. , & nozawa , s. 2010 , pasj , 62 , 993 kjeldsen , h. , & bedding , t. r. 1995 , a&a , 293 , 87 kjeldsen , h. , & bedding , t. r. 2011 , a&a , 529 , l8 laming , j. m. 2004 , apj , 614 , 1063 laming , j. m. 2009 , apj , 695 , 954 laming , j. m. 2012 , apj , 744 , 115 laming , j. m. , & drake , j. j. 1999 , apj , 516 , 324 laming , j. m. , drake , j. j. , & widing , k. g. 1996 , apj , 462 , 948 laming , j. m. , drake , j. j. , & widing , k. g. 1995 , apj , 443 , 416 lenz , d. d. , deluca , e. e. , golub , l. , rosner , r. , & bookbinder , j. a. 1999 , apj , 517 , l155 liefke , c. , ness , j. -u . ,
schmitt , j. h. m. m. , & maggio , a. 2008 , a&a , 491 , 859 longcope , d. w. , & tarr , l. 2012 , apj , 756 , 192 maggio , a. , sanz - forcada , j. , & scelsi , l. 2011 , a&a , 527 , a144 martnez - sykora , j. , hansteen , v. , & moreno - insertis , f. 2011 , apj , 736 , 9 mcintosh , s. w. , de pontieu , b. , carlsson , m. , hansteen , v. , boerner , p. , & goossens , m. 2011 , nature , 475 , 477 ness , j. -u . ,
schmitt , j. h. m. m. , burwitz , v. , mewe , r. , raassen , a. j. j. , van der meer , r. l. j. , predehl , p. , & brinkman , a. c. 2002 , a&a , 394 , 911 perryman , m. a. c. , et al .
1997 , a&a , 323 , l49 phillips , k. j. h. , chifor , c. , & dennis , b. r. 2006 , apj , 647 , 1480 phillips , k. j. h. , pike , c. d. , lang , j. , watanabe , t. , & takahashi , m. 1994 , apj , 435 , 888 porquet , d. , mewe , r. , dubau , j. , raassen , a. j. j. , & kaastra , j. s. 2001 , a&a , 376 , 1113 raassen , a. j. j. , et al .
2002 , a&a , 389 , 228 rakowski , c. e. , & laming , j. m. 2012 , apj , 754 , 65 reale , f. 2010 , living rev .
solar phys . ,
7 , 5 , url : http://www.livingreviews.org/lrsp-2010-5 robrade , j. , schmitt , j. h. m. m. , & favata , f. 2008 , a&a , 486 , 995 rosner , r. , tucker , w. h. , & vaiana , g. s. 1978 , apj , 220 , 643 sanz - forcada , j. , affer , l. , & micela , g. 2009 , a&a , 505 , 299 sanz - forcada , j. , maggio , a. , & micela , g. 2003 , a&a , 408 , 1087 schmelz , j. t. , nasraoui , k. , rightmire , l. a. , kimble , j. a. , del zanna , g. , cirtain , j. w. , deluca , e. e. , & mason , h. e. 2009 , apj , 691 , 503 schmelz , j. t. , nasraoui , k. , roames , j. k. , lippner , l. a. , & garst , j. w. 2005 , apj , 634 , l197 schmitt , j. h. m. m. , golub , l. , harnden , f. r. , jr . , maxson , c. w. , rosner , r. , & vaiana , g. s. 1985 , apj , 290 , 307 schmitt , j. h. m. m. , & liefke , c. 2004 , a&a , 417 , 651 shkolnik , e. , bohlender , d. a. , walker , g. a. h. , & collier cameron , a. 2008 , apj , 676 , 628 simoniello , r. , finsterle , w. , garca , r. a. , salabert , d. , jimnez , a. , elsworth , y. , & schunker , h. 2010 , a&a , 516 , a30 sturrock , p. a. 1999 , apj , 521 , 451 telleschi , a. , gdel , m. , briggs , k. , audard , m. , ness , j. -u . , & skinner , s. l. 2005 , apj , 622 , 653 telleschi , a. , gdel , m. , briggs , k. r. , audard , m. , & scelsi , l. 2007 , a&a , 468 , 443 veck , n. j. , & parkinson , j. h. 1981 , mnras , 197 , 41 vesecky , j. f. , antiochos , s. k. , & underwood , j. h. 1979 , apj , 233 , 987 von steiger , r. , wimmer - schweingruber , r. f. , geiss , j. , & gloeckler , g. 1995 , adv .
space res . , 15(7 ) , 3 walker , g. a. h. , et al .
2008 , a&a , 482 , 691 warren , h. p. , brooks , d. h. , & winebarger , a. r. 2011 , apj , 734 , 90 warren , h. p. , winebarger , a. r. , & brooks , d. h. 2012 , apj , 759 , 141 winebarger , a. r. , schmelz , j. t. , warren , h. p. , saar , s. h. , & kashyap , v. l. 2011 , apj , 740 , 2 wood , b. e. , laming , j. l. , & karovska , m. 2012 , apj , 753 , 76 wood , b. e. , & linsky , j. l. 2006 , apj , 643 , 444 wood , b. e. , & linsky , j. l. 2010 , apj , 717 , 1279 wood , b. e. , redfield , s. , linsky , j. l. , mller , h. -r .
, & zank , g. p. 2005 , apjs , 159 , 118 young , p. r. 2005a , a&a , 439 , 361 young , p. r. 2005b , a&a , 444 , l45 lcccc si xiii & 6.648 & 6.99 & @xmath64 & @xmath65 + si xiii & 6.688 & 6.99 & & + si xiii & 6.740 & 6.99 & & + mg xii & 8.419 & 7.11 & @xmath66 & @xmath67 + mg xii & 8.425 & 7.11 & & + mg xi & 9.169 & 6.80 & @xmath68 & @xmath69 + mg xi & 9.231 & 6.80 & & + mg xi & 9.314 & 6.79 & & + ne x & 12.132 & 6.87 & @xmath70 & @xmath71 + ne x & 12.138 & 6.87 & & + fe xvii & 12.264 & 6.62 & @xmath72 & @xmath73 + fe xxi & 12.285 & 6.98 & & + ni xix & 12.435 & 6.69 & @xmath74 & @xmath75 + ne ix & 13.447 & 6.58 & @xmath76 & @xmath77 + ne ix & 13.553 & 6.58 & & + ne ix & 13.699 & 6.58 & @xmath78 & @xmath79 + fe xvii & 13.823 & 6.60 & @xmath80 & @xmath81 + ni xix & 14.043 & 6.68 & @xmath82 & @xmath83 + ni xix & 14.077 & 6.68 & & + fe xviii & 14.203 & 6.74 & @xmath84 & @xmath85 + fe xviii & 14.208 & 6.74 & & + fe xvii & 15.015 & 6.59 & @xmath86 & @xmath87 + fe xvii & 15.262 & 6.59 & @xmath88 & @xmath89 + o viii & 15.176 & 6.65 & & + fe xix & 15.298 & 6.83 & & + o viii & 16.006 & 6.63 & @xmath90 & @xmath91 + fe xviii & 16.005 & 6.73 & & + fe xviii & 16.072 & 6.73 & & + fe xvii & 16.778 & 6.58 & @xmath92 & @xmath93 + fe xvii & 17.053 & 6.58 & @xmath94 & @xmath95 + fe xvii & 17.098 & 6.58 & & + o vii & 18.627 & 6.34 & @xmath96 & @xmath97 + o viii & 18.967 & 6.59 & @xmath98 & @xmath99 + o viii & 18.973 & 6.59 & & + o vii & 21.602 & 6.32 & @xmath100 & @xmath101 + o vii & 21.807 & 6.32 & @xmath102 & @xmath103 + o vii & 22.101 & 6.31 & @xmath104 & @xmath105 + n vii & 24.779 & 6.43 & @xmath106 & @xmath107 + n vii & 24.785 & 6.43 & & + c vi & 33.734 & 6.24 & @xmath108 & @xmath109 + c vi & 33.740 & 6.24 & & + s xiii & 35.667 & 6.43 & @xmath110 & @xmath111 + si xi & 43.763 & 6.25 & @xmath112 & @xmath113 + si xii & 44.019 & 6.44 & @xmath114 & @xmath115 + si xii & 44.165 & 6.44 & @xmath116 & @xmath117 + si xii & 45.521 & 6.44 & @xmath118 & @xmath119 + si xii & 45.691 & 6.44 & @xmath120 & @xmath121 + fe xvi & 46.661 & 6.43 & @xmath122 & @xmath123 + fe xvi & 46.718 & 6.43 & & + si xi & 49.222 & 6.24 & @xmath124 & @xmath125 + fe xvi & 50.361 & 6.43 & @xmath126 & @xmath127 + fe xvi & 50.565 & 6.43 & @xmath128 & @xmath129 + si x & 50.524 & 6.15 & & + si x & 50.691 & 6.15 & @xmath130 & @xmath131 + si xi & 52.298 & 6.24 & @xmath132 & @xmath133 + fe xvi & 54.127 & 6.43 & @xmath134 & @xmath135 + fe xvi & 54.710 & 6.43 & @xmath136 & @xmath137 + mg x & 57.876 & 6.22 & @xmath138 & @xmath139 + mg x & 57.920 & 6.22 & & + fe xv & 59.405 & 6.32 & @xmath140 & @xmath141 + fe xvi & 62.872 & 6.42 & @xmath142 & @xmath143 + mg x & 63.152 & 6.21 & @xmath144 & @xmath145 + mg x & 63.295 & 6.21 & @xmath146 & @xmath147 + fe xvi & 63.711 & 6.42 & @xmath148 & @xmath149 + fe xvi & 66.249 & 6.42 & @xmath150 & @xmath151 + fe xvi & 66.357 & 6.42 & & + fe xv & 69.682 & 6.32 & @xmath152 & @xmath153 +
fe xv & 69.941 & 6.32 & @xmath154 & @xmath155 + fe xv & 69.987 & 6.32 & & + fe xv & 70.054 & 6.32 & & + fe xv & 73.472 & 6.32 & @xmath156 & @xmath157 + fe xvi & 76.497 & 6.42 & @xmath158 & @xmath159 + ne viii & 88.082 & 5.96 & @xmath160 & @xmath161 + ne viii & 88.120 & 5.96 & & + fe xviii & 93.923 & 6.68 & @xmath162 & @xmath163 + fe xix & 101.550 & 6.80 & @xmath164 & @xmath165 + fe xviii & 103.937 & 6.68 & @xmath166 & @xmath167 + fe xix & 108.355 & 6.79 & @xmath168 & @xmath169 + fe xx & 121.845 & 6.88 & @xmath170 & @xmath171 + fe xxi & 128.752 & 6.95 & @xmath172 & @xmath173 + fe xx & 132.840 & 6.88 & @xmath174 & @xmath175 + fe xxiii & 132.906 & 7.12 & & + fe xxii & 135.755 & 7.03 & @xmath176 & @xmath177 + fe ix & 171.073 & 5.95 & @xmath178 & @xmath179 + fe x & 174.534 & 6.04 & @xmath180 & @xmath181 + cccc @xmath182 & 1.5 & 1.0 & 1.0 + @xmath21}$ ] & @xmath183 & @xmath184 & @xmath184 + @xmath185}$ ] & @xmath186 & 1.12 & 0.93 + @xmath187}$ ] & @xmath188 & ( 0.49 ) & 0.33 + @xmath189}$ ] & @xmath190 & 1.35 & 1.19 + @xmath37}$ ] & @xmath191 & ( 0.96 ) & 0.43 + @xmath192}$ ] & @xmath193 & 0.08 & 0.10 + @xmath194}$ ] & @xmath195 & 0.07 & 0.01 + @xmath196}$ ] & @xmath197 & ( @xmath198 ) & @xmath199 + @xmath200}$ ] & @xmath201 & @xmath202 & @xmath203 + | a _ chandra _ spectrum of the moderately active nearby f6 v star @xmath0 ori is used to study the coronal properties of mid - f dwarfs .
we find that @xmath0 ori s coronal emission measure distribution is very similar to those of moderately active g and k dwarfs , with an emission measure peak near @xmath1 seeming to be ubiquitous for such stars .
in contrast to coronal temperature , coronal abundances are known to depend on spectral type for main sequence stars . based on this previously known relation
, we expected @xmath0 ori s corona to exhibit an extremely strong `` fip effect , '' a phenomenon first identified on the sun where elements with low `` first ionization potential '' ( fip ) are enhanced in the corona .
we instead find that @xmath0 ori s corona exhibits a fip effect essentially identical to that of the sun and other early g dwarfs , perhaps indicating that the increase in fip bias towards earlier spectral types stops or at least slows for f stars .
we find that @xmath0 ori s coronal characteristics are significantly different from two previously studied mid - f stars , procyon ( f5 iv - v ) and @xmath2 boo ( f7 v ) .
we believe @xmath0 ori is more representative of the coronal characteristics of mid - f dwarfs , with procyon being different because of luminosity class , and @xmath2 boo being different because of the effects of one of two close companions , one stellar ( @xmath2 boo b : m2 v ) and one planetary . |
to apply photonic technologies to data processing has been taking place for several decades . yet , their implementation in practical systems remained limited and failed to penetrate practical processors .
many demonstrated optical devices are still oversized , were based on exotic materials and needed rather complicated interfacing with other electronic components .
this paper takes advantage of recent developments in silicon microphotonics and novel interfacing schemes in order to realize photonic components for fast communication and signal processing .
in addition to its practical potential , as presented , the ideas exposed herein present basic physical and mathematical challenges . + microring modulators were proposed and demonstrated for analog signal modulation and for simple digital modulation , such as on - off - keying ( ook ) . recently ,
microring modulators were investigated for advanced digital modulation formats such as pam @xcite @xcite @xcite , qpsk @xcite and even qam @xcite @xcite . + all the works mentioned above ( @xcite excluded ) use an analog voltage signal to drive the modulator .
the application of analog signals usually calls for mediating electronic circuitry , such as digital - to - analog conversion . in the current work
we promote the application of the , so - called , _ direct digital drive _ ( ddd ) @xcite method for use with microring resonators . ddd allows the utilization of only two voltage levels directly on the photonic device ; it makes the need for mediating devices , such as electrical digital - to - analog converter , unnecessary .
+ in the first part of this work we present a design of an n - bit digital - to - analog converter .
a digital - to - analog converter is a device that converts an n - bit digital word to a corresponding analog ( voltage ) representation .
a 4-bit dac produces 16 , equally spaced analog levels and can therefore be viewed as ( a digitally controlled ) 16-pam modulator . in the second part of this work we extended a previous work @xcite for generating m - qam signals with microring modulators by utilizing the ddd approach .
an all - digital m - qam modulator is thus presented .
as an example , consider the 4-bit optical dac based on a microring modulator , illustrated in figure [ fig : mring_da ] .
the device basic layout is similar to previously published microring resonators .
a cw light of wavelength @xmath0 and intensity @xmath1 is coupled from the waveguide to the microring .
the coupling coefficient between the waveguide and the microring is denoted by @xmath2 and the loss per round trip inside the microring is @xmath3 .
the light inside the microring is modulated by several , in this example @xmath4 , independent phase shifter segments ( electrodes ) .
the phase shifter can be implemented as reversed - bias pn - diode or as zig - zag pn diode , etc .
hereinafter , each phase shifter will be referred to as an _
electrode_. at its electrical input , the device accepts an @xmath5 bits digital word , denoted @xmath6 , where @xmath7 .
the input word is mapped onto each of the @xmath8 electrodes via the digital - to - digital converter ( ddc ) .
in essence , the ddc converts a @xmath9-bit input to a @xmath8-bit output , which , in turn , control the @xmath8 electrodes . note that each electrode is driven by one of two voltage levels , @xmath10 and @xmath11 , representing binary @xmath10 and @xmath12 .
described as such , the ddc is basically a lookup - table that can be realized by a ( high speed ) digital memory .
an optical modulator is typically characterized by the product @xmath13 where @xmath14 is the voltage whose application to an electrode of length @xmath15 , induces a phase shift of @xmath16 . without loss of generality , we set @xmath17 , i.e. as the circumference of the microring , where @xmath18 is the radius of the ring .
let @xmath19 , denote the index of the electrodes ( in this example @xmath20 ) .
assume that the length of each electrode is given by : @xmath21 .
note that @xmath22 . if a voltage @xmath11 is applied to electrode @xmath19 , the induced phase shift @xmath23 will be : @xmath24 where @xmath25 is the effective refractive index modulation due to the applied voltage @xmath11 on phase shifter @xmath19
, @xmath26 is an empirical constant that accounts for both the optical confinement and the coefficient of the charge density induced refractive index change .
the parameter @xmath27 is a binary quantity that indicates whether voltage @xmath11 was applied to phase shifter @xmath19 .
the dependance between @xmath25 and the applied voltage @xmath11 is known to be a nonlinear function , @xmath28 .
the exact relation depends on the phase shifter design .
the intensity transmission of the ddd multi - electrode microring structure can be written as : @xmath29 where a binary matrix @xmath30 of dimensions @xmath31 holds the mapping of the @xmath5-bit input word @xmath6 on the @xmath4 electrodes - evidently , a highly non - linear transmission . to maximize the output dynamic range ( dr )
, the microring resonator should be in critical coupling , @xmath32 .
this will allow the smallest possible output intensity level to approach @xmath10 .
figure [ fig : ringresoiout ] shows the output intensity of a microring resonator for phase shifts between @xmath10 and @xmath16 .
phase shifts greater than about @xmath33 $ ] contribute very little to the output dynamic range because of the high nonlinearity of the microring modulator .
hence , the voltage @xmath11 will be set so as to produce the intensity @xmath34 .
the quantity @xmath35 is chosen as a tradeoff between ( best achievable ) linearity and output dynamic range .
smaller values of @xmath35 can increase the output dynamic range , but also will reduce the linearity of the device , unless additional electrodes are added
. a dynamic range of about @xmath36 of the total available dr will keep the electrode count low and close to the input bit length . in this example @xmath11
was set to induce a phase shift that leads to an output dynamic range of about @xmath37 .
figure [ fig : ioutn=4m=5 ] shows the output intensity of a 4-bit dac ( 16 levels on a straight line ) , based on a microring modulator with 5 electrodes .
figure [ fig : sinewaven4m5 ] shows a sinewave generated with the proposed device .
the linearity of the dac can be quantified by standard figure of merits : its differential non - linearity ( dnl ) is 0.2 bits ; the integral non - linearity ( inl ) is 0.4bits and the effective number of bits ( enob ) is 3.74bits . and @xmath16.,width=264 ] [ fig : sinewaven4m5],width=264 ]
in the previous sections we presented the design of a dac , which is equivalent to a pam modulator , by using a single microring modulator with several electrodes driven by a digital signal . in this section
we use a two microring configuration to generate m - qam ( two - dimensional ) signal constellations .
quadrature amplitude modulation ( qam ) is a modulation scheme that conveys data by means of modulating both the amplitude and the phase of a sinusoid carrier thus providing spectral efficiencies in excess of 2bit / symbol . in previous work @xcite
, we introduced a method for generating optical m - qam signals by using two mutually decoupled microring modulators . in this configuration ,
one ring was used to generate the amplitude of the desired signal ( perturbed by some deterministic phase shift ) , while the second ring was used to complement the phase required to obtain the desired complex signal . the electronic input to this modulator was an analog voltage .
herein , in order to drive the two rings with digital ( two - level ) signals , we split the electrode into segments .
an m - qam modulator is schematically depicted in figure [ fig : mring_qam ] .
the design process is similar to the one described in section [ sec : mpamdesign ] except that we now have to generate complex signals rather than intensity levels only .
the role of the modulator is to generate a specific constellation of points consisting of @xmath38 distinct complex points ( also refereed to as _ signals _ ) ; the constellation points can be generally formulated as follows : @xmath39 the proposed modulator is described by means of an example .
figure [ fig : mring_qam ] depicts a 16qam optical modulator based on two micro - ring resonators equipped with multiple electrodes .
the electrodes in this example are divided between the two micro - rings : @xmath40 electrodes on each micro - ring . ,
is mapped by the ddc to a 14-bit output that drives the 14 segmented electrodes .
the electrode lengths ( per micro - ring ) follow a divide - by - two sequence .
the ddc , shown here for exposition purposes as two separate boxes , is actually a single memory device .
, width=302 ] as input , this qam modulator accepts a 4-bit digital word , denoted @xmath41 .
the 4-bit input word is mapped onto each of the @xmath42 electrodes via the ddc .
thus , each electrode is driven by one of two voltage levels , @xmath10 or @xmath43 , representing binary @xmath10 and @xmath12 , respectively .
more generally , let @xmath44 and @xmath45 be two vectors of dimensions @xmath46 and @xmath47 , whose elements correspond to the lengths of the electrodes of the first microring and the second microring , respectively
. let @xmath48 be a binary @xmath49 matrix .
row @xmath50 of @xmath48 , @xmath51 , holds the mapping from input @xmath6 onto the @xmath46 electrodes of the left ring .
likewise , binary matrix @xmath52 , of dimensions @xmath53 , holds the mappings from each of the input digital words to the @xmath47 electrodes of the right ring . with this nomenclature ,
the output of the modulator can be formulated as a function of the digital input @xmath6 : @xmath54\frac{\alpha_1-t_1 exp{-j\phi_1}}{1-t_1\alpha_1 exp\left(j\phi_1\right)}\cdot % \ ] ] @xmath55\frac{\alpha_2-t_2 exp{-j\phi_2}}{1-t_2\alpha_2 exp\left(j\phi_2\right ) } % \ ] ] and the phase shift are @xmath56 @xmath57 where @xmath58 denotes the amplitude of the optical field entering the modulator .
the geometrical structure and the loss and coupling parameters are set in a similar manner to the description in @xcite .
the first ring is designed to work in critical coupling for which @xmath59 in order to generate the largest span of amplitudes .
the second microring is designed to work in under - coupling regime , @xmath60 , to act as a phase shifter with minimum amplitude loss . by applying all possible combinations of the digital words @xmath61
, a finite pool of points , of maximum cardinality @xmath62 , can be generated .
as an example , a pool of ( green ) points is shown in figure [ fig:16qam_n12=7 ] for @xmath63 . from this large pool of possible points , one can choose a finite set of points that form a constellation for digital optical communications , @xmath64-qam in this example .
an ideal set of points that constitute a @xmath64-qam constellation , as suggested by eq .
[ eq : target_signals ] with @xmath65 , is portrayed by the red triangles in figure [ fig:16qam_n12=7 ] .
out of the green pool of @xmath66 signal points , we first select the @xmath64 signals that provide the best match for the ideal @xmath64 points .
the selected set of best points amounts to a mapping between @xmath64 @xmath6 digital input values and the corresponding @xmath61 s , and is executed by the ddc . .
the evm obtained is @xmath67.,width=302 ] as can be seen in the figure , the selected points are not identical to the ideal points , thus producing an error . to quantify this error
, we employ the error vector magnitude ( evm ) @xcite measure . for the above example
the evm is @xmath67 .
note that the evm can be further reduced by increasing the number of electrodes in each microring .
such an increase will allow generating a denser pool of points from which one can select a desired set of constellation points with higher accuracy .
table [ tbl : qam_evm ] presents the obtained evm for various constellations and varying number of electrodes .
it shows the evm for 16-qam with @xmath68 and @xmath69 electrodes .
the difference between these configurations is @xmath70db .
it can be seen that the inherent nonlineariy of the microring leads to a more involved implementation of m - qam modulators compared to m - pam modulators .
the table also presents results for 64-qam and 256qam .
while for 64qam we achieve evm better than -30db with @xmath69 electrode configuration , 256qam requires one additional electrode in the amplitude - related modulator ring to achieve such evm .
evm as a function of number of electrode [ cols="^,^,^,^",options="header " , ] the number of distinct electrode segments that can be effectively mounted on a single microring is obviously limited .
this , however , does not limit the proposed technology , as the original configuration can be augmented by additional rings with additional electrodes .
thus for example , rather than generating @xmath71-qam with @xmath40 and @xmath72 electrodes on each ring , one can utilize @xmath9 rings with @xmath8 electrodes on each ring : two of the rings will act as amplitude modulators while the other two will act as phase modulators , as depicted in figure [ fig : microring_qam_p=2 ] . for the same parameters as above , with the new configuration of @xmath9 microrings , evm of @xmath73 is achieved .
the practical implementation of the ddd modulators is not obvious .
next , we briefly discuss issues that are associated with the implementation of the above devices in silicon photonic technology .
silicon has the potential of full integration of optics and electronics either in a monolithic or a hyrbrid process .
the first problem that arises with silicon is that the plasma dispersion effect ( which is used widely to make phase shifters for silicon modulators ) induces a voltage dependent loss .
meaning , it is impossible to make a lossless phase shifter .
however , this problem can be alleviated by the ddd approach .
more specifically , the problem is that the vdl ( voltage dependent loss ) will reduce the average optical power and hence `` compress '' the signal pool .
the solution is simply to configure the ddc to map ( shrink ) the required constellation to a lower average optical power .
the second problem that might arise is that in order to generate a phase shifter to induce phases between @xmath10 and @xmath74 , a large voltage signal will be needed .
this problem can be solved by using the multiple microring configuration as discussed in the previous section .
for example , instead of using a single ring to induce the @xmath74 phase shift , we can use multiple modulators to induce the phase we are targeting .
we presented application of the direct digital drive approach to microring resonators .
we showed that this approach enables one to realize a digitally driven optical devices , that include compact m - pam modulator and dacs .
extending the approach to more then one microring , enables the generation of optical m - qam constellations .
we showed that it is possible to generate any constellation order either by using a large number of electrodes , or by using a small number of electrodes but employ additional microring modulators .
p. dong , c. xie , l. l. buhl , and y .- k .
chen , `` silicon microring modulators for advanced modulation formats , '' in _ optical fiber communication conference_.1em plus 0.5em minus 0.4emoptical society of america , 2013 , pp . ow4j2 . y. ehrlichman , o. amrani , and s. ruschin , `` improved digital - to - analog conversion using multi - electrode mach zehnder interferometer , '' _ j. lightw .
_ , vol . 26 , no . 21 , pp .
35673575 , nov.1 , 2008 . | the method of direct digital drive is applied to a microring resonator .
the microring resonator is thus controlled by a segmented set of electrodes each of which is driven by binary ( digital ) signal .
digital linearization is performed with the aid of digital memory lookup table .
the method is applied to a single microring modulator to provide an m - bit digital - to - analog converter ( dac ) , which may also be viewed as an m - level pulse amplitude modulator ( m - pam ) .
it is shown , by means of simulation , that a 4-bit dac can achieve an effective number of bits ( enob ) of 3.74bits . applying the same method for two rings ,
enables the generation of two - dimensional optical m - qam signals .
it is shown , by means of simulation , that a 16-qam modulator achieves an evm better than -30db . micoring modulators , optical digital - to - analog converter , optical m - pam , optical m - qam . |
one - dimensional ( 1d ) quantum well ( qw ) with vanishing at its edges wavefunction @xmath5 is an example that is included in every textbook on quantum mechanics @xcite .
influence on its properties of an electric field that is applied perpendicularly to the confining planes @xcite is of not only purely academic interest but also of the large technological applications . by creating the chosen direction
, the field lowers the overall symmetry of the structure what results in remarkable phenomena such as , for example , the change in the interband optical electroabsorption of a semiconductor qw as compared to the bulk samples .
this quantum - confined franz - keldysh effect @xcite is employed in the design of the devices of integrated optics , such as high - speed modulators @xcite , optically bistable switches @xcite and wavelength selective detectors @xcite .
the condition @xmath6 of the zeroing at the confining surface @xmath7 of the solution @xmath8 of the wave equation is not the only demand implemented in the analysis of the physical and chemical systems .
another frequently used type of the boundary condition ( bc ) is the requirement of zeroing at the interface of the normal derivative of @xmath8 : @xmath9 here , @xmath10 is an inward unit normal to the surface .
this condition is satisfied , for example , by the transverse electric waves in the electromagnetic waveguides @xcite or the order parameter of the superconductor bordering the vacuum or the insulator @xcite .
the planar structures with different bcs on the opposite confining walls @xcite are also encountered in nature ; for example , the long - distance dynamics between the earth and ionosphere of the very low frequency electromagnetic oscillations with wavelengths @xmath11 km is described , in the first approximation , as a propagation in the flat waveguide with one of its walls ( the ground ) being a perfect electric conductor with a reflection coefficient @xmath12 what corresponds to the bc from eq . , while the ionosphere behaves as an ideal magnetic conductor with a reflection coefficient @xmath13 with the edge requirement governed by eq .
similarly , the sound pressure in the acoustical waveguide formed by the air - water surface and the bottom of the lake is zero at the upper free boundary with its vanishing derivative at the pond s rigid floor @xcite .
the same configuration of the bcs is used , e. g. , in the description of the continental shelf waves along curved coasts @xcite .
a short review of other structures with mixed combination of the dirichlet and neumann bcs , which are called zaremba geometries @xcite , is given in ref .
@xcite . in the present research , exact calculations of the qw with all possible permutations of the bcs from eqs . and in the uniform transverse electric field are presented and analyzed with the emphasis on the interrelations between them .
while the case of the dirichlet , eq . , bcs at the both edges is quite well known @xcite , other distributions of the interface requirements with a voltage drop between them have received almost no attention @xcite .
meanwhile , they are indispensable , for example , in the analysis of the processes in the metallic waveguide with insulator or ferrite inner filling whose refractive index linearly changes between the confining walls .
in addition , if the electron concentration in the ionosphere is also a linear function of the height above the earth @xcite , then the corresponding phenomena will be described too by the rules presented below . however , our best intention here is not only to advice specific applications of the obtained results but , first of all , to build up a more general picture , which is also of the large didactic significance , by establishing interdependencies , drawing parallels and discovering similarities and differences between miscellaneous geometries of the bcs and their interplay with the field .
following purely dirichlet case @xcite , we , for definiteness , will speak about the quantum particle of the mass @xmath14 and charge @xmath15 ( with @xmath16 being an absolute value of the electron charge ) and the wavefunction @xmath8 that governs its motion through the solution of the corresponding steady state schrdinger equation .
a transformation to , e.g. , maxwell equations is elementary to do as both these relations belong to the same class that in mathematics is called a helmholtz equation .
excellent discussion of the analogies between the helmholtz and time - independent schrdinger equations @xcite shows that drawing such parallels becomes one of the main driving ideas in nanophotonics . upon calculating and analyzing energies and functions ,
a derivation is done of the polarizations of the states , which are a quantitative measure of a system response to the field . a fundamental relation that directly follows from the hellmann - feynman theorem and
links the polarization with the rate of the energy change with the field , is derived and analyzed .
consideration of the asymptotic limits of the weak and strong fields allows to arrive at simple analytical expressions .
quantum information entropies in position and momentum space are calculated and analyzed too ; in particular , an andvantage of the entropic uncertainty relation with respect to its heisenberg counterpart is shown for the non - dirichlet bcs .
for the zero field , it is proved that the position entropy is bc independent and for all states but the lowest neumann level equals to @xmath3 while the momentum entropy depends on the level and type of surface requirement .
the field varies position and momentum entropies in the opposite directions in such a way that the corresponding uncertainty relation is always satisfied .
some other results , such as the magnitude of the field at which the excited - state polarizations vanish , are discussed too both numerically and analytically .
summary of the properties of the eigenvalues and eigenfunctions of the laplace operator in flat arbitrary bounded euclidean domains with , in particular , dirichlet , eq . , or neumann , eq .
, bc is given in a recent review , ref .
so , in a sense , the discussion below is a first extension of the previous research for the simplest structure with different edge conditions biased by the electric field .
a companion paper @xcite uses the results obtained here for the calculation of the statistical properties of the same system .
the outline below is as follows . in sec .
[ sec_model ] our physical model is introduced and the necessary mathematical formalism for its analysis is developed .
[ sec_results ] is devoted to the presentation of the results with their detailed numerical and analytical investigation .
[ sec_conclusions ] wraps up the study by some conclusions .
a charged particle in the 1d infinitely deep qw of the width @xmath17 is subject to a uniform electric field @xmath0 that is applied perpendicularly to its interfaces , fig . [ profile ] .
on each of the walls either dirichlet ( d ) or neumann ( n ) bc is imposed . accordingly , below we will denote the structure by the two characters where the first ( second ) one refers to the bc at the left ( right ) wall . of course , nd configuration can be considered as the dn geometry with the electric field pointing in the opposite direction .
however , to keep the results consistent with the uniform bcs that are symmetric with respect to the sign change of the field , we discuss positive values of @xmath0 only and consider nd and dn cases separately .
the hamiltonian @xmath18 in the schrdinger equation @xmath19 consists of the field - free part @xmath20 and the electrostatic energy @xmath21 of the field @xmath22 with @xmath23 and @xmath24 qw potential @xmath25 reads : @xmath26 our primary goal is to find and analyze energies @xmath27 and wavefunctions @xmath5 for all possible permutations of the bcs and , based on them , to calculate other quantities describing the system ; in particular , polarization @xmath28 that is defined as @xcite @xmath29 with @xmath30 denoting a quantum mechanical expectation value : @xmath31 potential profile @xmath32 of the qw located between @xmath33 in the uniform electric field @xmath34 directed in the negative @xmath35 direction .
origin of the potential axis coincides with the middle of the potential drop across the well . at each wall
either dirichlet or neumann bc is satisfied . ]
it is convenient from the very beginning to use dimensionless units such that the distances are measured in units of the well width @xmath17 , and energies - in units of the ground state energy @xmath36 of the dirichlet qw .
then , the unit of the electric field @xmath37 is naturally determined from the voltage drop @xmath38 across the well , and unit of the polarization will be @xmath39 .
hence , the field - free part @xmath20 and electrostatic potential @xmath21 in the hamiltonian @xmath18 from eq . are given as @xmath40 @xmath41 with @xmath42 bcs are : for the dirichlet qw : [ bcnormalized1 ] @xmath43 with the prime denoting a derivative of the function with respect to its argument @xmath44 . solutions of the schrdinger equation form a countably infinite set with the wavefunctions @xmath45 for the @xmath46th level ( @xmath47 ) written as the linear combination of the airy functions @xmath48 and @xmath49 @xcite : for the pure dirichlet bcs : [ eigenfunctions1 ] @xmath50\\ \textnormal{for the nd case:}\nonumber\\ \label{eigenfunctions1_nd } \psi_n^{nd}\!\!=c_n^{nd}\!\!\left[{\rm bi}\!\left(z_{n-}^{nd}\right){\rm ai}\!\left(z_n^{nd}\right)-{\rm ai}\!\left(z_{n-}^{nd}\right){\rm bi}\!\left(z_n^{nd}\right)\right]\\ \textnormal{for the dn case:}\nonumber\\ \label{eigenfunctions1_dn } \psi_n^{dn}\!\!=c_n^{dn}\!\!\left[{\rm bi}'\!\left(z_{n-}^{dn}\right){\rm ai}\!\left(z_n^{dn}\right)-{\rm ai}'\!\left(z_{n-}^{dn}\right){\rm bi}\!\left(z_n^{dn}\right)\right]\\ \textnormal{for the pure neumann bcs:}\nonumber\\ \label{eigenfunctions1_nn } \psi_n^{nn}\!\!=c_n^{nn}\!\!\left[{\rm bi}'\!\left(z_{n-}^{nn}\right){\rm ai}\!\left(z_n^{nn}\right)-{\rm ai}'\!\left(z_{n-}^{nn}\right){\rm bi}\!\left(z_n^{nn}\right)\right],\end{aligned}\ ] ] where @xmath51 with superscripts @xmath52 and @xmath53 taking the values of @xmath54 and/or @xmath55 .
expressions for the normalization constants @xmath56 , which are determined from @xmath57 will be discussed below .
the form of the functions from eqs . automatically satisfies the bc at the right edge , @xmath58 , for all four configurations . applying the corresponding requirement to the the second confining surface ,
one gets transcendental equations for finding eigenenergies @xmath59 : for the pure dirichlet bcs : [ eigenenergy1 ] @xmath60 eqs . represent corresponding energies @xmath61 as implicit functions @xmath62 of the electric induction @xmath0 : @xmath63 calculation of the coefficients @xmath56 is greatly facilitated by the fact that an indefinite integral of the products of the two linear combinations of the airy functions @xmath64 and @xmath65 , @xmath66 , with , in general , different constants @xmath67 and @xmath68 is expressed through permutations of the products of the same combinations and their derivatives @xcite .
this , together with the bcs from and following from them eqs
. leads to quite simple expressions for @xmath56 ; for example , for the dirichlet qw it is : @xmath69^{1/2}},\ ] ] where also the value of the wronskian of the airy functions , which is equal to @xmath70 @xcite , has been used .
note that the sign in eq .
is chosen in such a way that the functions in the vicinity of the right edge do not become negative , @xmath71 , what is more convenient for our subsequent discussion .
this completes the calculation of the normalized wavefunctions @xmath72 , which correspond to the energies @xmath59 , and ushers in the first method of finding the polarization @xmath73 that in our dimensionless units reads : @xmath74 it relies on the direct calculation of the integrals in eq . .
first , one needs to find their values at zero electric field .
it is important to underline that the well known for this case expressions for the energies @xmath75 and functions [ functionszerofield1 ] @xmath76 can be derived also with the help of the asymptotic properties of the airy functions @xcite as the limit @xmath77 of the corresponding dependencies from eqs . and , respectively .
close enough to the right edge , the functions from eqs .
stay positive for any permutation of the bcs and the arbitrary @xmath46 , @xmath78 .
elementary calculation yields : [ polarizationszerofield1 ] @xmath79 these expected results are naturally explained by the symmetry of the flat qw with respect to its middle for the uniform bcs and the lack of it for the zaremba configurations . applied electric field destroys ( for the dd and nn structures ) this symmetry or changes ( for the mixed bcs ) the asymmetry pushing in this way the polarization from its zero value
. calculation of the corresponding integrals in eq . can again be performed analytically @xcite producing , after some simple algebra : [ normalizedpolarization2 ] @xmath80 ^ 2}{1+z_{n+}^{nd}\left[{\rm ai}\!\left(z_{n-}^{nd}\right)\!/{\rm ai}'\!\left(z_{n+}^{nd}\right)\right]^2}\\ & & p_n^{dn}(\mathscr{e})=-\frac{e_n^{dn}}{\mathscr{e}}-\frac{1}{2\pi^2(n+1/2)^2}\nonumber\\
\label{normalizedpolarization2dn } & & -\frac{1}{3}\left(\pi^2\mathscr{e}\right)^{-1/3}\frac{\left(z_{n-}^{dn}\right)^2+z_{n+}^{dn}\left[{\rm bi}'\!\left(z_{n-}^{dn}\right)\!/{\rm bi}\!\left(z_{n+}^{dn}\right)\right]^2}{z_{n-}^{dn}+\left[{\rm ai}'\!\left(z_{n-}^{dn}\right)\!/{\rm ai}\!\left(z_{n+}^{dn}\right)\right]^2}\\ & & p_n^{nn}(\mathscr{e})=-\frac{e_n^{nn}}{\mathscr{e}}\nonumber\\ \label{normalizedpolarization2nn } & & -\frac{1}{3}\left(\pi^2\mathscr{e}\right)^{-1/3}\frac{\left[z_{n+}^{nn}{\rm bi}'\!\left(z_{n-}^{nn}\right)\right]^2-\left[z_{n-}^{nn}{\rm bi}'\!\left(z_{n+}^{nn}\right)\right]^2}{z_{n+}^{nn}\left[{\rm bi}'\!\left(z_{n-}^{nn}\right)\right]^2-z_{n-}^{nn}\left[{\rm bi}'\!\left(z_{n+}^{nn}\right)\right]^2}.\end{aligned}\ ] ] another way of arriving at the same equations will be given in the next section . important characteristics of the physical and chemical systems is their quantum information entropy , which can be calculated in the position as well as momentum coordinates .
for the 1d qw under consideration , the position @xmath1 and momentum @xmath4 entropies are defined as @xmath81 where @xmath82 and @xmath83 are , respectively , position and momentum space densities : @xmath84 with momentum space wave function @xmath85 being a fourier transform of its position counterpart : @xmath86 paramount physical significance of these quantities is due to the fact that the total information entropy @xmath87 allows one to express the quantum uncertainty relations in the form @xmath88 instructive history of the derivation of the @xmath89-dimensional analogue of eq .
is described in ref .
@xcite ; namely , the first conjectures of these relations from the late 1950-ies @xcite were rigorously confirmed in 1975 by biaynicki - birula and mycielski @xcite and beckner @xcite .
it is known that the entropic uncertainty relation @xcite from eq . is stronger than its heisenberg counterpart that is written in the form @xmath90 with @xmath91 and @xmath92 being position and momentum standard deviations : @xmath93 where the momentum averaging is performed on the basis of the corresponding functions @xmath85 : @xmath94 shortcoming of eq . as compared to the entropy relation can be shown for our system as well ; consider , for example , the field - free lowest nn state @xcite that due to the constant unit magnitude position function @xmath95 exhibits the following momentum density : @xmath96 obviously , heisenberg inequality in this case is meaningless since the momentum deviation @xmath92 diverges due to the infinite integral for @xmath97 .
however , the corresponding entropy @xmath98 is finite and can be calculated numerically producing the result of @xmath99 , which , together with the zero position entropy @xmath100 , does satisfy inequality with its right - hand side equal to @xmath101 .
note that for eq .
the situation is aggravated even worse if , instead of taking a momentum expectation value on the basis of the functions @xmath85 ( where @xmath102 is a number ) , one employs the averaging on the basis of the position functions @xmath5 when the momentum is an operator , @xmath103 : @xmath104 for the configuration considered above this immediately leads to @xmath105 what obviously violates .
it was shown recently @xcite that this discrepancy is resolved if one recalls that the heisenberg inequality was derived for the infinite volume where the wave function vanishes at infinity .
for the non - dirichlet edge requirements of the finite 1d system a generalized bc - dependent uncertainty condition was derived @xcite that for the lowest field - free neumann state produces @xmath106 , which is indeed satisfied as an equality . for the dirichlet conditions ,
the heisenberg relation remains true since the function @xmath5 vanishes at the surface and , accordingly , at the infinity ; moreover , in this case the momentum deviation @xmath92 is the same either calculated over the position @xmath5 or momentum @xmath85 functions : @xmath107 momentum space densities @xmath108 at zero electric field for ( a ) the dirichlet , ( b ) neumann , and ( c ) mixed bcs as functions of the momentum @xmath109 where the solid lines are for the ground states ( @xmath110 ) , dashed curves - for the first excited levels ( @xmath111 ) , dotted lines - for the second excited states ( @xmath112 ) , and dash - dotted curves - for the levels with the quantum number @xmath113 . ]
arguments outlined in the previous paragraph dictate a necessity of studying quantum information entropies , especially for the structures with the non - dirichlet bcs .
miscellaneous aspects of the properties of the densities @xmath82 and @xmath83 and entropies @xmath1 and @xmath4 for different quantum systems were addressed in the literature @xcite . before discussing the electric field influence on the densities and entropies
, it does make sense to investigate these quantities for the flat qws .
relevant to the present discussion , let us mention that the position and momentum entropies of the dirichlet structure were calculated numerically @xcite and it was shown analytically @xcite that @xmath114 , independent of the quantum number @xmath46 . applying known integrals from the famous reference book @xcite ,
it is easy to show that at the zero fields the same is true for all other bcs but the ground neumann state , which , of course , is zero : @xmath115 on the other hand , for the momentum entropies at the zero voltage with the densities [ densityk_2 ] @xmath116}{\left[k^2-(n+1)^2\pi^2\right]^2}\\ \label{densityk_nn1 } \gamma_{n>0}^{nn}(k)&=&\frac{2k^2\left[1-(-1)^n\cos k\right]}{\pi\left(k^2-n^2\pi^2\right)^2}\\ \gamma_n^{nd}(k)&=&\gamma_n^{dn}(k)\nonumber\\ \label{densityk_nd1 } & = & \frac{2(-1)^{n+1}\left(n+\frac{1}{2}\right)\pi k\sin k+k^2+\left(n+\frac{1}{2}\right)^2\pi^2}{\pi\left[k^2-\left(n+\frac{1}{2}\right)^2\pi^2\right]^2}\end{aligned}\ ] ] and for any integrals from eqs .
, and at @xmath117 there are no analytical expressions in known to us literature @xcite ; accordingly , their direct numerical quadrature was performed in calculating the results presented below .
physically , zero entropy in eq . means that the uncertainty of determining the position of the particle is the largest one .
[ momentumdensitieszerofield ] exhibits zero - field momentum densities @xmath118 from eqs . and for all bcs and several quantum states @xmath46 .
previous analysis discussed these dependencies for the dirichlet qw only @xcite .
as @xmath118 is an even function of its argument , @xmath119 , we plot the parts with the non negative @xmath102 only .
qualitatively , this behavior is the same for any type of the surface requirements ; namely , for all excited states , @xmath120 , the momentum density is characterized by the two conspicuous peaks , which are located symmetrically with respect to @xmath121 and are flanked by the series of much smaller extrema .
the distance between the sharp maxima decreases for the smaller @xmath46 , and for the ground state they merge together forming one symmetric peak with its magnitude being the largest as compared to the excited levels .
quantitatively , neumann maximum @xmath122 is larger than the dirichlet or zaremba one , @xmath123 . as the momentum density is smaller than one , @xmath124 , the corresponding entropy @xmath4 from eq .
is always positive .
more detailed numerical analysis reveals that it is a growing function of the quantum number @xmath46 .
in contrast , the parts of the position density , which are larger than unity , in their contribution to the integral from eq . overweigh those with @xmath125 leading in this way to the negative entropies from eq . .
returning to the momentum entropies , we note that despite the fact that @xmath126 , the ground - state entropy @xmath127 is greater than its mentioned above neumann counterpart what apparently is explained by the far - reaching spread of the corresponding density : the wide sidelobe is clearly seen for the @xmath110 curve in panel ( c ) of fig .
[ momentumdensitieszerofield ] . in this way
, the dirichlet ground state with its smallest momentum entropy @xmath128 comes closest to the limit imposed by the entropic uncertainty relation : @xmath129 .
our calculations indicate that the dirichlet momentum and total entropy for any level @xmath46 remain the smallest ones among all possible bc combinations : @xmath130 .
energy spectrum @xmath131 as a function of the electric field @xmath132 for the pure dirichlet ( solid lines ) , neumann ( dotted curves ) , dirichlet - neumann ( dashed lines ) and neumann - dirichlet ( dash - dotted curves ) qw .
horizontal straight line denotes zero energy .
polarizations @xmath133 for ( a ) the ground , @xmath134 , ( b ) first , @xmath135 , and ( c ) second , @xmath136 , excited states as functions of the normalized electric field @xmath132
. solid ( dotted ) lines denote pure dirichlet ( neumann ) configuration while the dashed ( dash - dotted ) curves are for the dn ( nd ) geometry .
thin horizontal lines in parts ( b ) and ( c ) denote zero polarizations .
note different @xmath137 scales in each of the panels . ]
energy spectrum as a function of the field @xmath0 is shown in fig .
[ energies ] for all four types of the possible combinations of the bcs , and corresponding polarizations for the ground and two lowest excited states are depicted in fig .
[ polarizations ] .
it is seen that the energies , which at the zero field have the same values for the opposite zaremba geometries , eq .
, at the small voltage move in the opposite directions while at the large @xmath0 they coincide with their counterparts for the uniform bc distribution that is imposed at the right surface of the asymmetric edge requirements .
this means that the nd geometry is the only configuration at which the lowest energy at the small fields increases .
ground - state polarizations for all bcs monotonically increase with the field and approach at the strong voltages the asymptotic values that will be discussed below while for the excited levels the polarizations take negative values for the small and moderate fields , and only after reaching minimum they increase with the growing @xmath0 .
classically , this result , which first was predicted for the dirichlet qw @xcite , is unexpected since it means that the charged particle moves against the applied force .
however , in the quantum mechanical treatment the electron is described as a wave whose behavior is governed by the solution of the corresponding equation inside the domain with the appropriate bcs . for the small electric fields ,
the energy and polarization dependencies are conveniently described by the standard perturbation theory @xcite ; namely , considering in the total hamiltonian @xmath18 from eq .
the second term as the small disturbance of the field - free hamiltonian @xmath20 , one writes the perturbed energies and wavefunctions as @xmath138 these expressions are valid at @xmath139 . here
, the prime means that the term with @xmath140 is dropped out from the series , and @xmath141 are the matrix elements of the electric potential , eq . , which are calculated on the basis of the undisturbed functions @xmath142 from eqs .
: @xmath143 note that the condition of the applicability of the perturbation theory reduces for the low - lying states basically to @xmath144 that , for simplicity , will be used below ; however , one has to remember that for the high - lying levels , @xmath145 , the approximation produces very good results at the @xmath46-dependent applied voltage @xmath0 that actually can be much larger than unity .
integration immediately shows that for the symmetric bcs the energies for the quite small fields are quadratically proportional to @xmath0 while for the zaremba geometries they contain also a linear dependence on the voltage : [ energysmallfields1 ] @xmath146 ^ 2\!\!\frac{(m+1)^2}{\left[(n+1)^2-(m+1)^2\right]^5}\mathscr{e}^2\\ & & e_n^{_{dn}^{nd}}(\mathscr{e}\!\ll1)=\left(n+\frac{1}{2}\right)^2\pm\frac{1}{2\pi^2\left(n+1/2\right)^2}\mathscr{e}\nonumber\\ & + & \frac{1}{\pi^4}{\sum_{m=0}^\infty}'\!\!\left\{\frac{\left[(-1)^{n+m}-1\right]^2}{(n - m)^5(n+m+1)}\right.\nonumber\\ \label{energysmallfields1_dn } & + & \left.\!\frac{\left[(-1)^{n+m}+1\right]^2}{(n - m)(n+m+1)^5}\right\}\mathscr{e}^2\\ & & e_n^{nn}(\mathscr{e}\!\ll1)\!=n^2\!+\!\frac{2\!-\!\delta_{n0}}{\pi^4}\nonumber\\ \label{energysmallfields1_nn } & \times&{\sum_{m=0}^\infty}'\!\!(2\!-\!\delta_{m0})\left[(-1)^{n+m}\!-\!1\right]^2\frac{(n^2+m^2)^2}{\left(n^2-m^2\right)^5}\mathscr{e}^2,\end{aligned}\ ] ] where @xmath147 is the kronecker @xmath148 .
infinite series in eqs .
are easily calculated on the basis of the method proposed in ref .
@xcite for the dirichlet case . as an example , let us consider its application for the ground state , @xmath110 , of the zaremba geometry . after some simple algebra , the series from is transformed to @xmath149 ^ 3}-96\sum_{m=1}^\infty\frac{1}{\left[(2m)^2 - 1\right]^4}\nonumber\\ \label{{series1 } } & -&128\sum_{m=1}^\infty\frac{1}{\left[(2m)^2 - 1\right]^5}.\end{aligned}\ ] ] explicit expression for @xmath150 with the noninteger @xmath151 is provided in eq .
5.1.25.37 of ref .
@xcite with other series being missed there and in other sources ; however , for our purposes it is useful to point out that all of them can be readily calculated by observing that each differentiation with respect to @xmath151 of the identity @xcite @xmath152 produces one additional power of @xmath153 in the denominator of the series in its left - hand side . after that , putting @xmath151 equal to @xmath154 allows to find the expressions above .
in the same way , the series @xmath155^{-k}$ ] are found by @xmath156 differentiations of the equality @xcite @xmath157 in addition , the series @xmath158 appearing in the calculation of @xmath159 can be derived as a limiting case @xmath160 of the above mentioned summation , or one can use directly the same source @xcite to get its value as @xmath161 . for reference ,
the expressions for the three lowest levels at @xmath144 are provided below : [ energysmallfieldsdd ] @xmath162 ( eq . was derived before either with the help of the perturbation calculations @xcite or variational approach @xcite or hypervirial - perturbational treatment @xcite ) [ energysmallfieldsnd ] @xmath163 ( expressions for the dn case are given by eqs . with , however , the negative sign of the linear term ) [ energysmallfieldsnn ] @xmath164 note that in this regime the difference @xmath165 decreases for the nd geometry and grows with @xmath0 for all other bcs what causes different voltage dependence of the critical temperatures of the bose - einstein systems in the corresponding 1d box @xcite . invoking regular units , one can say that , for the energy separation between the two lowest states , the small electrostatic field leads to the decrease of the effective width for all but nd wells . for calculating the polarizations @xmath73
, one can insert the expressions for the wavefunctions from into the general definition , eq .
, and perform the integration with the subsequent evaluation of the series as it just was described above .
however , there is another applicable at the arbitrary fields elegant method of computing them that employs the hellmann - feynman theorem @xcite stating : @xmath166 in this equation , @xmath167 is some continuous parameter of the system . applying it to our configuration where the role of the variable @xmath167 is played by the electric field one
arrives at @xmath168 where the last equality follows straightforwardly from eqs . , and .
in this way , the following expression for the polarization , which is equivalent to that given above , eq . , is derived : @xmath169 this fundamental relation @xcite establishes an intimate connection between the physical observable quantity ( in our case , polarization ) and the speed of the energy change ( for our geometry , with the electric field ) .
the magnitude of the energy itself is a relative value that depends on the choice of the origin from which we measure it ; for example , for the pure dirichlet ground state it is a continuously increasing function of the field if zero energy coincides with the lowest point located at the right edge @xcite , see fig .
[ profile ] .
however , the difference between the two energies has an absolute meaning , as eq . vividly manifests .
applying it to eqs .
, one gets : [ polarizationsmallfields1 ] @xmath170 ^ 2\!\!\frac{(m+1)^2}{\left[(n\!+\!1)^2\!-\!(m\!+\!1)^2\right]^5}\mathscr{e}\\ p_n^{nd}(\mathscr{e})&=&p_n^{dn}(\mathscr{e})=-\frac{2}{\pi^4}\!\!\!\!{\sum_{m=0}^\infty}'\!\!\left\{\frac{\left[(-1)^{n+m}-1\right]^2}{(n - m)^5(n+m+1)}\right.\nonumber\\ \label{polarizationsmallfields1_nd_dn } & + & \left.\frac{\left[(-1)^{n+m}+1\right]^2}{(n - m)(n+m+1)^5}\right\}\mathscr{e}\\ p_n^{nn}(\mathscr{e})&=&-\frac{2\left(2\!-\!\delta_{n0}\right)}{\pi^4}\nonumber\\ \label{polarizationsmallfields1_nn } & \times & { \sum_{m=0}^\infty}'\!\!\!\!\!\left(2\!-\!\delta_{m0}\right)\left[(-1)^{n+m}-\!1\right]^2\frac{\left(n^2+m^2\right)^2}{\left(n^2-m^2\right)^5}\mathscr{e},\end{aligned}\ ] ] valid at @xmath144 .
it is important to note that for the uniform bcs the energies at the small fields depend quadratically on the voltage while the polarizations are linear functions of @xmath0 .
this is explained by the symmetry of the qw with respect to its middle , which causes that the first - order energy perturbation @xmath171 vanishes identically .
however , the non - diagonal matrix elements of @xmath21 , which enter into the expression for the perturbed function , eq . , determine that the latter is deformed and , accordingly , the polarization survives already in the first order of the expansion with respect to the small parameter @xmath0 . for the zaremba qws
, there is no such symmetry at the zero field ; accordingly , both energies as well as polarizations are linear functions of the voltage .
it is also notable from eq . that in this regime the polarizations of the opposite mixed bcs are equal to each other , as it is also seen from fig . [ polarizations ] at the small fields where they are flanked by the uniform edge requirements .
in addition , it is well known that the second - order corrections to the ground - state energy are always negative @xcite ; as a result , the corresponding polarizations that are defined from eq .
increase .
the easiest way to explain qualitatively ( and to the large degree of precision , quantitatively ) the negative value of the excited - state polarization at the small and moderate electric fields lies in the analysis of the first excited level for the pure neumann bcs . in general , the applied voltage mixes all zero - field states , as eq .
but for our consideration it suffices to take into account the field - induced admixture to @xmath172 produced by the ground level since it is the nearest lying state that , accordingly , exerts the largest influence : @xmath173,\quad\mathscr{e}\ll1.\end{aligned}\ ] ] this equation manifests that the small electric admixture of the ground state subtracts the @xmath44-independent term from @xmath174 .
if one recalls that , by our convention , this function is nonnegative at the right interface , @xmath58 , it means that the applied voltage decreases ( increases ) the amplitude of the right ( left ) extremum . as a result , the square of the wavefunction @xmath175 gets smaller ( bigger ) at the right ( left ) edge of the qw what classically means the paradoxical shift of the particle against the applied force . however , quantum mechanical treatment considers the electron as a wave and the negative polarization is a manifestation of its wave nature equivalent to the interference of the modes in the fabry - prot resonator @xcite with gradually changing inside it index of refraction .
the decrease of the nn function at the right wall for the small fields is clearly seen in the corresponding panel of fig .
[ function1 ] that shows @xmath176 vs. the electric field for all four distributions of the bcs .
polarization of the state elementary follows from its definition , eq . , and approximate expression for @xmath177 from : @xmath178 taking into account all higher lying states modifies insignificantly the above result , as an application of the hellmann - feyman theorem , eq .
, to reveals : @xmath179 note that the same two - state approximation applied to the ground level produces the following expressions for the function @xmath180 and polarization @xmath181 : @xmath182 @xmath183 we retained in the last result more digits than before in order to show its excellent coincidence with the one obtained from eq . , which counts the contributions from all states : @xmath184 comparison of eqs . and proves that for the ground level the dominant admixture comes from the nearest lying excited state while the contributions from the higher levels can be safely neglected .
eq . tells us that for the lowest level the application of the field leads to such deformation of the particle concentration inside the well that is consistent with the rules of the classical mechanics : probability of finding the electron at the right ( left ) wall increases ( decreases ) with the electric force .
dependence on the field of the ground state wavefunctions @xmath185 for all possible @xmath52 and @xmath53 is shown in fig .
[ function0 ] . returning to the interaction of the right neumann wall and the field ,
let us note that each of them separately causes the particle to accommodate closer to the interface : the surface attracts the electron by creating an extremum of its wavefunction while the appropriately directed electric force simply pushes it closer to the border .
however , for any excited state their combined influence on the particle in some range of the small voltages leads to the opposite result , i.e. , to its repulsion from the wall . we underline again that this is a manifestation of the wave nature of the quantum charge carriers .
[ function1 ] shows that for all other combinations of the bcs the closest to the right wall extremum undergoes the same suppression of its magnitude what results in the negative polarizations exhibited by panels ( b ) and ( c ) of fig .
[ polarizations ] .
physically , their explanation is absolutely identical to the one discussed above for the purely neumann qw ; namely , it is the field - induced admixture of the different states that causes their destructive interference .
above , we provided the detailed analysis of the nn structure only since it is the simplest from the mathematical point of view , transparent and very instructive one .
it is also clear why at the small fields the magnitude of the polarization for the ground or excited states is the largest for the purely neumann qw , as fig .
[ polarizations ] depicts ; namely , for the flat structure all extrema have the same amplitude equal to @xmath186 , see eqs .
, and their relative weight in the integral from is determined by its distance from the centre of the well : the further from the origin , the larger its contribution is . obviously , the field - induced change of the maxima or minima located at the neumann walls will impact the polarization stronger than their dirichlet counterparts , which are repelled from the boundary into the interior of the well . for the same reason ,
zaremba geometries with one attractive and one repulsive surfaces have their polarizations for @xmath144 lying in between those two for the uniform bcs .
further growth of the electric field makes the perturbation theory less and less applicable , and eqs . and
should be used for finding energies and polarizations .
it is noteworthy to show that the latter dependencies are derivable from the former ones with the help of the hellmann - feynman theorem that holds for the arbitrary @xmath0 .
namely , for finding a derivative that enters one considers eigenvalue eqs . as implicit functions of the energy @xmath187 on the electric field with its most general form provided by eq .
a theorem of differentiating implicit functions states @xcite : @xmath188 applying this rule to each of eqs .
, one arrives at the corresponding dependence from .
thus , there are two equivalent methods of calculating the polarization : one implements a direct integration of with the use of the analytical properties of the primitives of the airy functions , and the second one relies on the application of the hellmann - feynman theorem and the rule for finding a derivative of the implicit functions . at the very strong fields , it is reasonable to assume that the electron is so squeezed to the right surface that it does not feel the left boundary what leads to the problem of the triangular potential with the vertical dirichlet @xcite or neumann right wall .
then , the wavefunction degenerates to @xmath189 application of the right bc leads in this limit to the following energies : [ energyhighfields1 ] @xmath190 what results in the polarizations [ polarizationhighfields1 ] @xmath191 here , @xmath192 and @xmath193 are the @xmath46th solutions of equations @xcite @xmath194 respectively .
their values are provided , for example , in table 10.13 of ref .
also , function * airyaizeros * of computer algebra system maple , ref .
@xcite , calculates @xmath192 with any desired precision .
numerically , it is noteworthy to remark that the absolute value of the difference between the two consecutive zeros is smaller for the airy function in comparison with its derivative : @xmath195 what leads to the different critical temperatures of the corresponding bose - einstein condensates @xcite .
show that there are saturation values of the polarizations in the asymptotics of the very high fields : [ polarizationhighfields2 ] @xmath196 due to small power of @xmath0 in , this limit is approached quite slowly , as fig .
[ polarizations ] shows . on the contrary ,
the equal values of the energies for the qws with the same bc at the right surface , eq . , are achieved at relatively moderate voltages , which are level dependent : for the higher lying states this happens at the larger fields , see fig .
[ energies ] .
next , let us discuss the electric fields and corresponding energies at which the polarizations turn to zero . in general
, they should be found from the requirement of the vanishing right - hand - side of ; for example , for the purely dirichlet case , eq . yields the following transcendental equation for calculating them : @xmath197 where the superscript at @xmath0 means that the energy at this electric field reaches maximum . as a first approximation ,
the high - field limit from eq .
provides the value of this voltage as @xmath198 while the corresponding maximum energy at this point is @xmath199 where , for convenience , we have substituted the superscript denoting a type of the bc by the one denoting that this is an extremum value .
one gets from eq . :
@xmath200 and @xmath201 while the exact ones are : @xmath202 and @xmath203 .
the corresponding approximate energies read : @xmath204 and @xmath205 with the exact ones being @xmath206 and @xmath207 .
so , the precision of the approximation increases with the level number and is better for the energy than for the field .
it directly follows from eqs . and that @xmath208 what automatically zeroes the right - hand side of eq . , as it should be , since it contains a contribution from the left wall that is neglected in our high - field approach developed above
assuming that the expression from eq .
is a small nonzero number , @xmath209 and applying the corresponding taylor expansion of the left - hand side of eq . with respect to the small parameter @xmath148
, one can improve the approximations from eqs . and ; however , the obtained results are not very transparent and , accordingly , not written here . as a next example , it is instructive to introduce zero - energy electric field @xmath210 , i.e. , the field at which the energy of the @xmath46th level turns to zero .
its physical meaning is obvious from its definition and the transcendental equation for its determination follows from eqs . by putting @xmath211 . assuming that this happens at quite large fields
, one can invoke asymptotic properties of the airy functions @xcite to get @xmath212 with the obvious requirement @xmath213 .
even though these dependencies are derived for the high voltages , they produce reasonably good results even for the small @xmath46 ; for example , the lowest dn level reaches zero at @xmath214 while eq .
calculates it as @xmath215 .
ground state momentum density is shown in fig .
[ momentumdensities0 ] as a function of the momentum @xmath102 and voltage @xmath0 .
the general tendency of the electric field influence is the decrease of the maximal values of @xmath216 and smoothing out of the oscillations along the @xmath102 axis .
note that for the neumann - dirichlet well the largest value @xmath217 as a function of the applied voltage has a local maximum of @xmath218 at @xmath219 while for all other geometries it is a monotonically decreasing function of the field .
similar dependencies are characteristic for the higher lying states too
. peculiar features of the position and momentum densities exemplified by figs .
[ function1]-[momentumdensities0 ] determine the properties of the corresponding entropies @xmath1 and @xmath4 .
[ entropies ] shows them together with the sum @xmath220 as functions of the electric field for all four possible bcs and several low - lying levels .
characteristic feature is a nonmonotonic dependence of the position @xmath1 and momentum @xmath4 entropies on the small and moderate voltages for the structures with the left neumann wall ( except the ground nn level ) .
for example , the maximum of the lowest nd state position entropy is explained by the sharp decrease of the corresponding function @xmath221 at the left surface for @xmath222 ( see the left lower panel of fig .
[ function0 ] ) and accompanying decline of its negative contribution to @xmath1 while the corresponding density at the right edge stays almost unchanged .
simultaneously , mentioned above increase of the momentum density results in the contraction of @xmath4 and , as a consequence of these two facts , the lowering of the total entropy that , however , stays above the fundamental limit from the right - hand side of eq .
, as expected . for the higher voltages
, the accumulating near the right interface function @xmath223 pushes the position space entropy @xmath1 downward forming the extremum seen in the figure . on the contrary ,
the dirichlet left interface subdues the entropy swaying , as the corresponding panels of fig .
[ entropies ] demonstrate . at the large @xmath0 ,
the entropies are determined by the right - wall bc , as discussed above ; accordingly , they are almost equal for the dn and pure neumann qws and their growing with the field magnitudes are larger than their ( almost identical ) counterparts for the nd and pure dirichlet structures . similar to the field - free case
, the dirichlet ground state has the lowest total entropy among all other levels and all possible bc combinations .
note its quite small variation with the voltage : from its zero - field value of @xmath224 it saturates to approximately @xmath225 at the large @xmath0 .
growing with the field magnitude of the negative position entropy @xmath1 means smaller uncertainty in determining particle location while the larger values of the positive momentum entropy @xmath4 indicate that the uncertainty in calculating momentum @xmath102 increases .
exact solutions of the helmholtz equation inside the 1d interval with different bcs at its ends and subjected to the linear potential inside revealed similarities as well as differences of the miscellaneous permutations of the edge requirements .
similarities include : a continuous growth with the electric field @xmath0 of the ground - state polarization for any bc geometry and its slow approach to the saturation value that , however , is different for the uniform and zaremba distributions of the surface requirements ; a decrease of the excited - level polarization for the small voltages and reverse of its change for the stronger electric forces .
the negative polarization at the small @xmath0 is conveniently described by the perturbation theory ; in particular , it is shown that its reason lies in the field - induced mixing of the unperturbed levels .
mathematical analysis of the simple two - level approximation proved that only an admixture of the neighbouring states plays a dominant role in the change of the polarization .
in addition , the perturbation theory confirms mathematically a physical conclusion about different magnitudes of the polarization for the different bcs . in the opposite limit of the high voltages ,
physically the quantum particle inside the box does not see one of the surfaces what mathematically means that solutions of the schrdinger equation in this case depend only on the bc at the opposite wall to which it is pushed by the electric field .
application of the hellmann - feynman theorem , which long ago proved to be a powerful tool of studying quantum systems @xcite , in our case revealed a fundamental relation between the speed of the energy change with the field , on the one hand , and the polarization , on the other .
applied voltage also modifies position and momentum densities and entropies in such a way that the entropic uncertainty relation , eq .
, is always satisfied .
dirichlet , eq . , and neumann , eq .
, conditions are the limiting cases of the so called robin bc @xcite @xmath226 where the variation of the extrapolation length @xmath227 continuously changes the edge requirements from the dirichlet ( @xmath228 ) to the neumann ( @xmath229 ) configuration . at the zero voltage , @xmath230 , the vanishingly small negative robin length , @xmath231 , heavily pushes two levels to the qw interfaces with their energies unrestrictedly decreasing :
@xmath232{}-1/(\pi\lambda)^2 $ ] @xcite .
their almost full degeneracy in this regime might lead to the nontrivial dependence of the spectrum on the field while the strong localization at the qw walls might produce unusual polarization and entropy behavior . as a simplifying example , for the single wall with the negative robin length @xmath227 @xcite it is elementary to prove that at @xmath230 it possesses one bound state with the energy ( in regular units ) @xmath233 and the entropies [ singlerobinwallentropy ] @xmath234 where it is assumed that @xmath235 .
eqs . and
are exact for any negative @xmath227 .
it is seen from them that even though the position and momentum entropies diverge at the limiting values of @xmath236 and @xmath231 , the total entropy @xmath237 stays finite , does not depend on the extrapolation length ( due to the exact cancelling of the divergent terms in @xmath1 and @xmath4 ) and does satisfy inequality . at the weak fields the polarization for any type of the bcs has opposite signs for the ground and any excited states .
the natural question arises : is there warm " enough temperature or / and sufficient number of the charged fermions that are able to push the total statistically averaged polarization to the negative values ?
the following paper @xcite , among other issues , answers this question .
this project was supported by deanship of scientific research , college of science research center , king saud university .
100 l. d. landau and e. m. lifshitz , quantum mechanics ( non - relativistic theory ) ( pergamon , new york , 1977 ) .
a. rabinovitch and j. zak , phys . rev .
b * 4 * , 2358 ( 1971 ) .
t. lukes , g. a. ringwood , and b. suprapto , physica a * 84 * , 421 ( 1976 ) . f. m. fernndez and e. a. castro , physica a * 111 * , 334 ( 1982 ) . g. bastard , e. e. mendez , l. l. chang , and l. esaki , phys . rev .
b * 28 * , 3241 ( 1983 ) .
d. ahn and s. l. chuang , appl .
lett . * 49 * , 1450 ( 1986 ) .
d. nguyen and t. odagaki , am . j. phys .
* 55 * , 466 ( 1987 ) .
d. a. b. miller , d. s. chemla , t. c. damen , a. c. gossard , w. wiegmann , t. h. wood , and c. a. burrus , phys .
lett . * 53 * , 2173 ( 1984 ) .
d. a. b. miller , d. s. chemla , t. c. damen , a. c. gossard , w. wiegmann , t. h. wood , and c. a. burrus , phys .
b * 32 * , 1043 ( 1985 )
. j. s. weiner , d. a. b. miller , d. s. chemla , t. c. damen , c. a. burrus , t. h. wood , a. c. gossard , and w. wiegmann , appl .
* 47 * , 1148 ( 1985 ) .
d. a. b. miller , d. s. chemla , and s. schmitt - rink , phys .
b * 33 * , 6976 ( 1986 ) . a w. achtstein , a. v. prudnikau , m. v. ermolenko , l. i. gurinovich , s. v. gaponenko , u. woggon , a. v. baranov , m. y. leonov , i. d. rukhlenko , a. v. fedorov , and m. v. artemyev , acs nano * 8 * , 7678 ( 2014 ) .
t. h. wood , c. a. burrus , d. a. b. miller , d. s. chemla , t. c. damen , a. c. gossard , and w. wiegmann , ieee j. quantum electron .
* 21 * , 117 ( 1985 ) .
d. a. b. miller , d. s. chemla , t. c. damen , t. h. wood , c. a. burrus , a. c. gossard , and w. wiegmann , ieee j. quantum electron . * 21 * , 1462 ( 1985 ) .
t. h. wood , c. a. burrus , a. h. gnauck , j. m. wiesenfeld , d. a. b. miller , d. s. chemla , and t. c. damen , appl .
. lett . * 47 * , 190 ( 1985 ) .
j. d. jackson , classical electrodynamics , 3rd edn .
( wiley , new york , 1999 ) .
p. g. de gennes , superconductivity of metals and alloys ( benjamin , new york , 1966 ) . o. olendski and l. mikhailovska , phys .
e * 67 * , 056625 ( 2003 ) .
k. davies , ionospheric radio ( peter peregrinus , london , 1990 ) .
k. g. budden , the wave - guide mode theory of wave propagation ( prentice - hall , englewood cliffs , n. j. , 1961 ) .
e. r. johnson , m. levitin , and l. parnovski , siam j. math .
37 * , 1465 ( 2006 ) .
s. zaremba , j. math .
. 9 sr . * 6 * , 127 ( 1927 ) .
h. najar and o. olendski , j. phys .
a : math . theor . * 44 * , 305304 ( 2011 ) .
p. exner , j. phys .
a : math . gen .
* 28 * , 5323 ( 1995 ) .
k. g. budden , the propagation of radio waves ( cambridge , cambridge , 1985 ) .
s. v. gaponenko , introduction to nanophotonics ( cambridge , cambridge , 2010 ) .
d. s. grebenkov and b .-
nguyen , siam rev . * 55 * , 601 ( 2013 ) .
o. olendski , ann .
( berlin ) * 527 * , 296 ( 2015 ) .
m. abramowitz and i. a. stegun , handbook of mathematical functions ( dover , new york , 1964 ) .
o. valle and m. soares , airy functions and applications to physics , 2nd edn ( world scientific , hackensack , nj , 2010 )
. i. biaynicki - birula , phys .
a * 103 * , 253 ( 1984 ) .
h. everett , iii , in : the many - worlds interpretation of quantum mechanics , edited by b. s. dewitt and n. graham , princeton series in physics ( princeton university press , princeton , 1973 ) , chap . 1 .
hirschman , am . j. math . *
79 * , 152 ( 1957 ) .
i. biaynicki - birula and j. mycielski , commun .
44 * , 129 ( 1975 ) .
w. beckner , annals math .
* 102 * , 159 ( 1975 ) .
i. biaynicki - birula and .
rudnicki , in : statistical complexity : applications in electronic structure , edited by k. d. sen ( springer , dordrecht , 2011 ) , chap
s. wehner and a. winter , new j. phys .
* 12 * , 025009 ( 2010 ) . i. biaynicki - birula , aip conf . proc . *
889 * , 52 ( 2007 ) .
m. h. al - hashimi and u .- j .
wiese , ann .
( n. y. ) * 327 * , 1 ( 2012 ) .
s. r. gadre , s. b. sears , s. j. chakravorty , and r. d. bendale , phys .
a * 32 * , 2602 ( 1985 ) .
r. w. robinett , am .
* 63 * , 823 ( 1995 ) .
j. tao , g. li , and j. li , and j. li , j. chem .
phys . * 107 * , 1227 ( 1997 ) .
j. snchez - ruiz , phys .
a * 226 * , 7 ( 1997 ) .
v. majernk and l. richterek , j. phys .
a : math . gen . * 30 * , l49 ( 1997 ) .
v. majernk and l. richterek , eur . j. phys . * 18 * , 79 ( 1997 ) .
v. majernk , r. charvot , and e. majernkov , j. phys .
a : math . gen . *
32 * , 2207 ( 1999 ) .
j. s. dehesa , a. martnez - finkelshtein , and j. snchez - ruiz , j. comp .
math . * 133 * , 23 ( 2001 ) .
s. lpez - rosa , j. montero , p. snchez - moreno , j. venegas , and j. s. dehesa , j. math .
chem . * 49 * , 971 ( 2011 ) .
a. i. aptekarev , j. s. dehesa , p. snchez - moreno , and d. n. tulyakov , j. math . chem . * 50 * , 1079 ( 2012 ) . h.
g. laguna and r. p. sagar , ann .
( berlin ) * 526 * , 555 ( 2014 ) .
a. p. prudnikov , y. a. brychkov , and o. i. marichev , integrals and series vol .
1 ( gordon and breach , new york , 1986 ) , p. 540 , eq . 2.6.34.12 .
a. p. prudnikov , y. a. brychkov , and o. i. marichev , integrals and series vol . 2 ( gordon and breach , new york , 1990 ) .
a. p. prudnikov , y. a. brychkov , and o. i. marichev , integrals and series vol .
3 ( gordon and breach , new york , 1990 ) . y. a. brychkov , handbook of special functions : derivatives , integrals , series and other formulas ( chapman & hall , boca raton , 2008 ) .
i. s. gradshteyn and i. m. ryzhik , table of integrals , series , and products ( academic , new york , 2014 ) .
@xcite , p. 687 . ref .
@xcite , p. 685 , eq
. 5.1.25.4 . ref .
@xcite , p. 688
5.1.26.1 . ref .
@xcite , p. 653
r. p. feynman , phys .
* 56 * , 340 ( 1939 ) .
lwdin , j. mol . spectrosc . * 3 * , 46 ( 1959 ) .
h. e. montgomery , jr . and
v. i. pupyshev , eur .
j. h * 38 * , 519 ( 2013 ) .
a. lipson , s. g. lipson and h. lipson , optical physics , 4th edn .
( cambridge , cambridge , 2011 ) . w. rudin , principles of mathematical analysis , 3rd edn .
( mcgraw - hill , new - york , 1976 ) .
j. katriel and g. adam , physica * 43 * , 546 ( 1969 ) .
k. gustafson and t. abe , math .
intell . * 20*(_1 _ ) , 63 ( 1998 ) .
o. olendski and l. mikhailovska , phys .
e * 81 * , 036606 ( 2010 ) .
p. eba , lett .
* 10 * , 21 ( 1985 ) .
t. flp , t. cheon , and i. tsutsui , phys .
a * 66 * , 052102 ( 2002 ) .
b. belchev and m. a. walton , j. phys . a : math
. theor . * 43 * , 085301 ( 2010 ) .
o. georgiou , g. gligori , a. lazarides , d. f. m. oliveira , j. d. bodyfelt , and a. goussev , europhys . lett . * 100 * , 20005 ( 2012 ) . | analytical solutions of the schrdinger equation for the one - dimensional quantum well with all possible permutations of the dirichlet and neumann boundary conditions ( bcs ) in perpendicular to the interfaces uniform electric field @xmath0 are used for the comparative investigation of their interaction and its influence on the properties of the system . limiting cases of the weak and strong voltages
allow an easy mathematical treatment and its clear physical explanation ; in particular , for the small @xmath0 , the perturbation theory derives for all geometries a linear dependence of the polarization on the field with the bc - dependent proportionality coefficient being positive ( negative ) for the ground ( excited ) states .
simple two - level approximation elementary explains the negative polarizations as a result of the field - induced destructive interference of the unperturbed modes and shows that in this case the admixture of only the neighboring states plays a dominant role .
different magnitudes of the polarization for different bcs in this regime are explained physically and confirmed numerically .
hellmann - feynman theorem reveals a fundamental relation between the polarization and the speed of the energy change with the field .
it is proved that zero - voltage position entropies @xmath1 are bc independent and for all states but the ground neumann level ( which has @xmath2 ) are equal to @xmath3 while the momentum entropies @xmath4 depend on the edge requirements and the level . varying electric field changes position and momentum entropies in the opposite directions such that the entropic uncertainty relation is satisfied .
other physical quantities such as the bc - dependent zero - energy and zero - polarization fields are also studied both numerically and analytically .
applications to different branches of physics , such as ocean fluid dynamics and atmospheric and metallic waveguide electrodynamics , are discussed . |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.