ebook img

Probes of Diffusive Shock Acceleration using Gamma-Ray Burst Prompt Emission PDF

0.12 MB·English
Save to my drive
Quick download
Download
Most books are stored in the elastic cloud where traffic is expensive. For this reason, we have a limit on daily download.

Preview Probes of Diffusive Shock Acceleration using Gamma-Ray Burst Prompt Emission

ToappearinProc.ofthe6thHuntsvilleGamma-RayBurstSymposium(2009), eds.C.A.Meegan,N.Gehrels,&C.Kouveliotou(AIPConf.Proc.,NewYork). Probes of Diffusive Shock Acceleration using Gamma-Ray Burst Prompt Emission Matthew G. Baring DepartmentofPhysicsandAstronomy,MS-108,RiceUniversity,P.O.Box1892, 9 Houston,TX77251-1892,USA Email:[email protected] 0 0 2 Abstract. Theprincipalparadigmforgamma-raybursts(GRBs)suggeststhattheprompttransientgamma-raysignalarises from multiple shocks internal to the relativistic expansion. This paper explores how GRB prompt emission spectra can n constrainelectron(orion)accelerationpropertiesattherelativisticshocksthatpertaintoGRBmodels.Thearrayofpossible a high-energypower-lawindicesinacceleratedpopulationsishighlighted,focusingonhowspectraabove1MeVcanprobethe J fieldobliquityinGRBinternalshocks,andthecharacterofhydromagneticturbulenceintheirenvirons.Whenencompassing 6 theMeV-bandspectralbreak,fitstoBATSE/EGRETburstdataindicatethatthepreponderanceofelectronsresponsibleforthe 1 promptemissionresideinanintrinsicallynon-thermalpopulation.Thisdiffersmarkedlyfromtypicalpopulationsgenerated inaccelerationsimulations;potentialresolutionsofthisconflictsuchastheactionofself-absorptionarementioned.Spectral ] E modelingalsosuggeststhatthesynchrotronmechanismisfavoredoversynchrotronself-Comptonscenariosduetothelatter’s typicallybroadcurvaturenearthepeak.Suchdiagnosticswillbeenhancedbythebroadbandspectralcoverageofburstsbythe H FermiGamma-RaySpaceTelescope;theGBMwillprovidekeyinformationonthelowerenergyportionsofthenon-thermal . particlepopulation,whiletheLATwillconstrainthepower-lawregimeofparticleacceleration. h p Keywords: Gamma-raybursts,non-thermalemission;diffusiveshockacceleration;hydromagneticturbulence - PACS: 98.70.Rz;95.85.Pw;98.70.Sa;52.35.Ra;52.25.Xz;52.27.Ny;52.35.Tc;52.65.Pp o r t s INTRODUCTION a [ The gamma-rayburstparadigmhas evolveddramaticallyoverthe last decadesince the first redshiftdeterminations 1 providedunambiguousevidenceforacosmologicaldistancescale.Yetmuchisstilltobelearnedaboutassociations, v progenitors,andtheenvironmentofboththepromptandafterglowemissionregions,andthephysicalprocessesactive 5 therein.For promptburstemission, the measurementof the high energyspectralindexprovidesa key constrainton 3 5 theinterpretationoftheelectronaccelerationprocess.Presumingthepopularburstparadigmofradiativedissipation 2 at internalshocksthataccelerateparticles[1–3], it is ofgreatinterestto understandwhatphysicalconditionsin the . shockedenvironscanelicittheobservedindicesandthespectralstructurearoundtheMeV-bandpeak.Toassessthis, 1 0 oneturnstomodelsofdiffusiveshockaccelerationinrelativisticsystems.Thesecomeindifferentvarieties,including 9 semi-analyticsolutionsofthediffusion-convectionequation[4,5],MonteCarlosimulations[6–9],andparticle-in-cell 0 (PIC) full plasma simulations [10–13]. Each has its merits and limitations. Tractability of the analytic approaches : generallyrestrictssolutiontopower-lawregimesfortheparticlemomentumdistributions.PICcodesarerichintheir v i informationon turbulenceand shock-layerelectrodynamics.To interfacewith GRB data, a broaddynamicrangein X momentaisdesirable,andthisisthenaturalnicheofMonteCarlosimulationtechniques,thefocusofthispaper. r AkeypropertyofaccelerationattherelativisticshocksthatarepresumedtoseedpromptGRBradiativedissipation a is that the distribution functions f(p) are inherently anisotropic. This renders the power-law indices and other distribution characteristics sensitive to directional influences such as the magnetic field orientation, and the nature of MHD turbulence that often propagatesalong the field lines. Accordingly,familiar results from relativistic shock acceleration theory such as the so-called canonical s =2.23 power-lawindex [5], while providinguseful insights, are of fairly limited applicability. This is fortunate since the GRB database so far has exhibited a substantial range ofspectralcharacteristics,soanyuniversalsignatureintheaccelerationpredictionswouldbeunnecessarilylimiting. ThispaperexploressomeofthefeaturesofdiffusiveshockaccelerationusingresultsfromaMonteCarlosimulation, andaddressesprobesofthetheoreticalparameterspaceimposedbyextantGRBobservationsaroundandabovethe MeVspectralbreak.Palpablediagnostics[14]havealreadybeenenabledbydataprovidedbytheBATSEandEGRET instrumentsontheComptonGamma-RayObservatory(CGRO)forahandfulofbrightbursts.Theburstcommunity expects significant advances in the understanding of such constraints in the next few years, afforded by the broad spectralcoverageandsensitivityoftheGBMandLATexperimentsonNASA’sFermiGamma-RaySpaceTelescope. DIFFUSIVEACCELERATION ATRELATIVISTIC SHOCKS MostinternalGRBshocksthatareputativelyresponsibleforthepromptemissionaremildly-relativisticbecausethey areformedbythecollisionoftwoultra-relativisticshells.Hence,thisregimeformsthefocusofthisexposition.The discussionwilladdressthetheoreticalmethodologyemployed,andtheresultingcharacteristicsofthediffusiveaccel- erationprocess,beforemovingontoconstraintsontheoreticalparametersimposedbythepromptGRBobservations. ThesimulationusedheretocalculatediffusiveaccelerationinrelativisticplanarshocksisakinematicMonteCarlo technique that has been employed extensively in supernova remnant and heliospheric contexts, and is described in detailinnumerouspapers[6,7,15–18].ItisconceptuallysimilartoBell’s[19]testparticleapproachtodiffusiveshock acceleration.Particlesareinjectedupstreamandgyrateinalaminarelectromagneticfield,withtheirtrajectoriesbeing governedbyarelativisticLorentzforceequationintheframeoftheshock.Ingeneral,thefluidframemagneticfieldis inclinedatanangle Q Bf1 totheshocknormal.Becausetheshockismovingwithavelocityu(x)relativetotheplasma restframe,thereis,ingeneral,au×Belectricfieldinadditiontothebulkmagneticfield.Particleinteractionswith Alfvénwaveandotherhydromagneticturbulenceismodeledbyusingaphenomenologicalscatteringofthecharges intherestframeoftheplasma.Thescatteringprecipitatesspatialdiffusionofparticlesalongmagneticfieldlines,and to a varyingextent,acrossthem as well. Thescatteringsare also assumedto be quasi-elastic,an idealizationthat is usually valid because in most astrophysicalsystems the flow speed far exceedsthe Alfvénspeed, and contributions fromstochasticsecond-orderFermiaccelerationaresmall.Thediffusionpermitsaminorityofparticlestotransitthe shockplanenumeroustimes,gainingenergywitheachcrossingviatheshockdriftandfirst-orderFermiprocesses. A continuumof scattering angles,between large-angle(LAS) or small-anglecases, can be modeledby the simu- lation. In the local fluid frame, the time, d t , between scatterings is coupled [6] to the mean free path, l , and the f maximumscattering (i.e. momentumdeflection) angle, q scatt via d tf ≈lq s2catt/(6v) for particles of speed v≈c. Usually l isassumedtobeproportionaltoapoweroftheparticlemomentum p (see[6,20]formicrophysicaljus- tificationsforthischoice),andforsimplicityitispresumedtoscaleastheparticlegyroradius, r ,i.e. l =h r (cid:181) p. g g Theparameter h inthemodelisameasureofthelevelofturbulencepresentinthesystem,couplingdirectlytothe amount of cross-field diffusion, such that h =1 correspondsto the isotropic Bohm diffusion limit, where the field fluctuationssatisfy d B/B∼1.Inkinetictheory, h couplestheparallel(k =l v/3)andperpendicular(k )spatial k ⊥ diffusioncoefficientsviatherelation k /k =1/(1+h 2) [16,21].Inparallelshocks,wheretheBfieldisdirected ⊥ k alongthe shocknormal(Q Bf1=0), h hasonlylimitedimpactonthe resultingenergyspectrum,principallydeter- miningthe diffusivespatialscale normalto the shock.However,in obliquerelativistic shockswhere Q Bf1>0, the diffusivetransportofparticlesacrossthefield(andhenceacrosstheshock)becomescriticaltoretentionoftheminthe accelerationprocess.Accordingly,forsuchsystems,theinterplaybetweenthefieldangleandthevalueof h controls thespectralindexoftheparticledistribution[7,22],afeaturethatiscentraltotheinterpretationofGRBspectrabelow. Acceleration Properties atRelativisticShocks Representativeparticledifferentialdistributions dN/dp thatresultfromthesimulationofdiffusiveaccelerationat mildly-relativistic(internalGRB)shocksaredepictedintheleftpanelofFigure1(see[7,9]for G ≫1 simulation 1 results);thesedistributionsareequallyapplicabletoelectronsorions,andsothemassscaleisnotspecified.Astriking feature is that the slope and shape of the non-thermal particle distribution depends on the nature of the scattering. Theoftencitedasymptotic,ultrarelativisticindexof s =2.23 for dN/dp(cid:181) p−s isrealizedonlyforparallelshocks with Q Bf1 =0◦ in the mathematical limit of small (pitch) angle diffusion (PAD), where the particle momentum is stochasticallydeflectedonarbitrarilysmallangular(andthereforetemporal)scales.Inpractice,PADresultswhenthe maximum scattering angle q scatt is inferiorto the Lorentz cone angle 1/G 1 in the upstream region.In such cases, particles diffuse in the region upstream of the shock only until their velocity’s angle to the shock normal exceeds around 1/G ,afterwhichtheyarerapidlysweptdownstreamoftheshock.TheFigureindicatesclearlythatwhenthe 1 field obliquity increases, so also doesthe index s , with values greater than s ∼3 arising for Q Bf1 &50◦ for this mildly-relativisticscenario.Thisisaconsequenceofmoreprolificconvectiondownstreamawayfromtheshock. Figure1alsoshowsresultsforlargeanglescatteringscenarios(LAS,with 4/G 1.q scatt.p ),wherethedistribu- tionishighlystructuredandmuchflatteronaveragethan p−2.Thestructurebecomesmorepronouncedforlarge G 1 [7,9,22],andiskinematicinorigin,wherelargeangledeflectionsleadtofractionalenergygainsbetweenunityand G 2 insuccessiveshockcrossings.Eachstructuredbumporspectralsegmentcorrespondstoanincrementoftwointhe 1 numberofshocktransits.For p≫mc, theyasymptoticallyrelaxto a power-law,in thiscase with index s ≈1.62. FIGURE1. Leftpanel:ParticledistributionfunctionsdN/dpfrommildly-relativisticshocks(G 1b 1=3,i.e.b 1=u1/c=0.949) of upstream-to-downstream velocity compression ratio r=u1x/u2x ≈3.24. Simulation results can be divided into two groups: parallel shock runs (Q Bf1 =0◦, upper three histograms), and oblique, superluminal shock cases (Q Bf1 =20◦,40◦,60◦, lower three histograms). Scattering off hydromagnetic turbulence was modeled by randomly deflecting particle momenta by an angle withinacone,ofhalf-angle q scatt,whoseaxiscoincideswiththeparticlemomentumpriortoscattering;theratioofthediffusive mean freepath l tothegyroradius rg was fixedat h =l /rg=5. Theheavyweight lines(twouppermost histograms) arefor the large angle scattering cases (LAS: 1/G 1 ≪q scatt ≤p ). All other cases constitute pitch angle diffusion (PAD) runs, when q scatt.1/G 1 andthedistributionsbecomeindependentofthechoiceof q scatt.Alldistributionsasymptoticallyapproachpower- lawsdN/dp(cid:181) p−s athighenergies.Forthetwocasesbracketingtheresultsdepicted,thepower-lawsareindicatedbylightweight lines,withindicesof s =1.61 (Q Bf1=0◦, q scatt≤p )and s =3.31 (Q Bf1=60◦, q scatt≤10◦),respectively. Rightpanel:Power-lawindices s forsimulationrunsinthelimitofsmallanglescattering(pitchanglediffusion),foranalmost non-relativisticshockofupstreamflowspeed b 1x≡u1x/c=0.5,andanMHDvelocitycompressionratio r=4.Theindicesare displayedasfunctionsoftheeffectivedeHoffman-Tellerframeupstreamflowspeed b 1HT=b 1x/cosQ Bf1,withselectvaluesof thefluidframefieldobliquityQ Bf1 markedatthetopofthepanel.Obliquitiesforwhichb 1HT>1constitutesuperluminalshocks. Thedisplayedsimulationindexresultswereobtainedfordifferentdiffusivemeanfreepaths l paralleltothemeanfielddirection, namely l /rg=1 (blue squares), l /rg=10 (black triangles), l /rg=102 (green pentagons), l /rg=103 (red triangles) and l /rg=104 (magentahexagons),aslabelled.ThelightweightblackcurveatthebottomlabelledKH89definesthesemi-analytic resultfromKirk&Heavens’[4]solutiontothediffusion-convectionequation,correspondingto l /rg→¥ . Both panels: The short heavyweight magenta lines labelled GRB 930131 indicate the approximate spectral index s that is appropriateforthisEGRETburst,ifacooledsynchrotronemissionscenarioisoperable(seetextforadiscussion). IntermediatecasesarealsodepictedinFigure1,with q scatt∼4/G 1.Thespectrumissmooth,likeforthePADcase, buttheindexislowerthan2.23.Fromtheplasmaphysicsperspective,magneticturbulencecouldeasilybesufficient toeffectscatteringsonthisintermediateangularscale,acontentionthatbecomesevenmoregermaneforultrarelativis- tic shocks[9].Astrophysically,suchflatdistributionsmayprovedesirableinmodelsofGRB orblazarjetemission, particularly if efficient radiative cooling is invoked.Note that acceleration distributions are qualitatively similar for ultra-relativisticshocks, i.e. those perhapsmoreappropriateforGRB afterglowstudies, with a trend [5–7, 9, 22] of declining s forhigher G ,theconsequenceofanincreasedkinematicenergyboostingincollisionswithturbulence. 1 It is clear thatthere is a considerablerangeof indices s possible for non-thermalparticlesacceleratedin mildly relativistic shocks, a non-universality that can attractively mesh with GRB observations.. The two key parameters that dictate the array of possible spectral indices s are the shock field obliquity Q Bf1, and the ratio h =l /rgof thediffusivemeanfreepathtotheparticle’sgyroradius r .Aparametersurveyfordiffusiveaccelerationatatypical g mildly-relativistic shock is exhibited in the right panel of Fig. 1, where only the pitch angle diffusion limit was employed.Thepower-lawindex s isplottedasafunctionofthedeHoffman-Teller(HT[23])framedimensionless speed b 1HT=b 1x/cosQ Bf1,wherethesubscript x denotescomponentsalongtheshocknormal.Thisistheupstream flow speed in the shock rest frame where the flow is everywhere parallel to the magnetic field; it correspondsto a physical speed when b 1HT <1 (i.e. the upstream field obliquity satisfies cosQ Bf1 <b 1x, and the shock is said to besubluminal.Whenmathematically b 1HT>1,theshockistermedsuperluminal,andthedeHoffman-Tellerframe technicallydoesnotexist—noLorentzboostcanrendertheflowparalleltoB.Clearly,thex-axismapsvariationsin Q Bf1,withmorehighlyobliqueshockscorrespondingtotherightoftheplot. Afeatureofthisplotisthatthedependenceof s onfieldobliquityisnon-monotonic.When l /r ≫1,thevalueof g s atfirstdeclinesas Q Bf1 increasesabovezero.Thisleadstoveryflatspectra.As b 1HT approachesandeventually exceedsunity,thistrendreverses,and s thenrapidlyincreaseswithincreasingshockobliquity.Thisisthecharacter of near-luminal and superluminal shocks evinced in the left panel of Fig. 1, precipitated by inexorable convection ofparticlesawaydownstreamoftheshock,steepeningthedistributiondramatically.Theonlywaytoamelioratethis rapiddeclineintheaccelerationefficiencyistoreduce l /r tovaluesbelowaround 10.Physically,thisistantamount g toincreasingthehydromagneticturbulencetohighlevelsthatforcetheparticlediffusiontoapproachisotropy.Then, transportofchargesacrossthemeanfieldbecomessignificantongyrationaltimescales,andtheycanberetainednear theshockforsufficienttimestoaccelerateandgeneratesuitablyflatdistributionfunctions.Suchlowvaluesof l /r g render the field direction immaterial, and the shock behaves much like a parallel, subluminal shock in terms of its diffusivecharacter.Then, s isonlyweaklydependenton Q Bf1,thesecondkeypropertyillustratedintherightpanel. The origin of the extremely flat distributions with s ∼1 is in the coherent effect of shock drift acceleration at the shock discontinuity.This phenomenonhas been widely understood in non-relativistic shocks, and is due to the energy gain of charges when they repeatedly encounter u×B electric fields (in frames other than the HT frame) in gyrations straddling the shock discontinuity. The energy gain rate then scales as the gyrofrequency c/r , i.e. g is proportionalto the particle’s energy/momentum,so that the equilibrium distribution realized in pure shock drift accelerationis dN/dp(cid:181) p−1 [18].Reducing l /r andtherebyintroducingextremelymodestamountsofcross-field g diffusion disrupts this coherence, removes particles from the shock layer, and steepens the spectrum. It is noted in passingthatthechoiceofthesomewhatlarge(for b =0.5)compressionratio r=4 wastoafforddirectcomparison 1x with thesemi-analyticconvection-diffusionequationresultsofKirk &Heavens[4],whichphysicallycorrespondto l /r ≫1 caseswithminimalcrossfielddiffusion.Thedifferencesbetweenoursimulationindicesinthislimitand g thoseof[4]arisebecausetheadiabaticapproximationemployedin[4]todescribeshockcrossingsdoesnottakeinto accountthegyrationaltrajectoriesofchargesintheshocklayer.Thesesubtletiesarediscussedfurtherin[18]. CONSTRAINTS ONACCELERATIONTHEORY FROMGRBOBSERVATIONS Gleaningmeaningfulinsightsintotheconnectionbetweenaccelerationtheoryandpromptburstsignalshingesupon twoaspects:thehighenergyspectralindex a (=−b forBandmodel[24]index b ),whenthedifferentialspectrum h is dn/deg (cid:181) eg−a h,andgeneratingburstspectralshapeatandbelowthe n -Fn peak.Thesewillbeconsideredinturn. The a High-Energy Spectral Index h Thehigh-energyphotonspectralindices a fortheCGROEGRETandComptelburstsaretabulatedin[25].One h oftheseEGRETdetections,GRB930131,hadanindex a ≈2,whichcorrespondstotheindicationinbothpanelsof h Fig.1ifasynchrotronorinverseComptoncoolingmodelisadopted.ThisisontheflatendoftheEGRET a index h distribution[26], for whicha handfulof sourcesevincedphotondifferentialspectrumindicesscattered in the range 2.a .3.7 (thoughnotethatEGRETsparkchamberdetectionswereconcentratedintheindexrange a .2.8 [25]). h h TheCGRO BATSE a indexdistribution[27]issimilarlybroad,butwith 1.5.a .3.5 andfargreaterstatistics. h h The recentFermi detection of GRB 08016Cin both the GBM and LAT instrumentsoffereda high energyindex of a ∼2.2 initsmostluminousepoch,andasteeperspectrumatothertimes[28].Accordingly,observationally,shock h accelerationmodelsmustaccommodatearadiationspectralindexintherange 2.a .4 inordertobeviable. h Thephasespaceforachievabilityofthisdependsontheradiationmechanism,andalsouponwhetherornotcooling isinvoked.ThemostpopularmodelscenariofortheproductionofthepromptGRBsignalisquasi-isotropicelectron synchrotronemission[2,3].Inthiscase,aswithinverseComptonscattering,thephotondifferentialspectralindexis given by a h =(s +1)/2 abovethe n -Fn peak, if coolingis not efficient, and a h =(s +2)/2 if the particles are continuously cooled in a large volume before being re-injected into an acceleration process, for which the particle distributionindexissteepenedbyunity.FromtheFigure,itisevidentthatthisisreadilyachievableforGRB930131 forquasi-parallelshockswithperhapssomecomponentofmodestanglescattering,orformoderatelyobliqueshocks as long as l /r .3−10. The EGRET GRB index range 2.a .3.7 then correspondsto diffusive acceleration g h with 2.s .5.4 in the presenceof cooling,and 3.s .6.4 otherwise.In either case, subluminalshocks, either quasi-parallel(Q Bf1.20◦) or moderatelyoblique(20◦.Q Bf1.60◦) are requiredto explain flatter GRB spectra as in GRB 910503, GRB 930131, GRB 950425 or GRB 080916C. For bursts with somewhat larger a indices, h superluminal shocks can be viable, but only if the scattering is strong, i.e. l /r .3−10. Otherwise, convection g awayfromtheshockdominatestheacceleration,andtheelectrondistributionbecomestoosteep.Forthelesspopular uncooled hadronic models, the photon and particle indices trace each other, i.e. a =s , and a similar conclusion h thattheenvironmentisrestrictedtosubluminalormodestlysuperluminalobliqueshocksisderived.Notethatwhile Lorentz transformations of circumburst fields of arbitrary orientation render ultra-relativistic external GRB shocks quasi-perpendicular(Q Bf1≈90◦),moderatefieldobliquitiesareoftenattainedinmildly-relativisticinternalshocks. Discriminating between the s constraints in the presence or absence of synchrotron cooling mandates a brief discussion of the balance between acceleration and cooling. Diffusive shock acceleration is usually dominated by gyroresonantinteractionsbetweenchargesandhydromagneticturbulence[29].Ifthisisthecase,thenforachargeof Lorentzfactor g ,theaccelerationrateessentiallyscalesas g timestheparticle’sgyrofrequency,implying dg /dt(cid:181) g 0. In contrast, synchrotronand inverseComptoncoolingrates yield dg /dt (cid:181) −g 2. Accordingly,simultaneousand co- spatial accelerationand synchrotroncoolingyield a power-lawdistribution with the “raw” accelerationindex s up to a maximumenergywhere cooling beginsto dominate, and generatesan exponentialturnover.This is the widely consideredscenarioforX-ray/TeVsupernovaremnantstudies.ForexternalGRBshocksspawningafterglowsignals, there is plenty of time for accelerated charges to diffuse away from the shock and subsequently cool, so that the cumulativeeffectisemissionfromacooling-steepenedinjectedpower-law.Thesituationisverydifferentforinternal shocks.Theaccelerationmustpersistuptothemaximumenergyseenintheobservations,forexample1GeVinGRB 930131and13GeVinGRB080916C,andonatimescale D t commensuratewiththevariabilityattheseenergies,i.e., D t.1secormuchless.Coolingbreaks,byanindexof D a =1/2,arenotobservedabove1MeVinburstemission,so h ifthespectraspanning1MeV–10GeVcorrespondtoacooling-dominatedcontribution,theaccelerationmustoperate inshort,impulsiveperiods,followedbylongcoolingepochsthatpermittheelectronstodeclineinmomentumbya factorof 102 orso.Accordingly,thecoolingepochsshouldpossessdurationsofaround4ordersofmagnitudelonger than the impulsive acceleration epochs. It is not clear that the superposition of a host of these presumably pseudo- FRED-likeprofilescanmimictypicalbursttimehistorieswhilemaintainingthetime-averagedspectralbehavior. Constraints from the n -Fn Peak An additional requirement for the success of an acceleration model is that when convolved with a radiation process such as synchrotronradiation, it can generate the shape of the GRB spectrum around and below the n -Fn peak. This was the focusof the investigationby Baring & Braby [14]. Early work on broad-bandspectralfitting of burstswasprovidedbyTavani[30],whoobtainedimpressivefits totime-integratedspectraof severalbrightCGRO BATSE bursts using a phenomenologicalelectrondistribution and the synchrotronemission mechanism.While this succeededinthislimitedsample,providingadriverforthepopularityofthesynchrotronGRBemissionmodel,there are difficulties with fitting low energy (i.e. .100keV) spectra in about 1/3 of bursts [31]. Employing perspectives based on acceleration theory, this investigative effort was extended and refined by [14], who pursued a program of spectralfittingofGRBemissionusingalinearcombinationofthermalandnon-thermalelectronpopulations.These fitsdemandedthatthepreponderanceofelectronsthatareresponsibleforthepromptemissionresideinanintrinsically non-thermal population. Such a constraint, dictated by the narrowness of the n -Fn peak, is commensurate with electrondistributionsfrequentlypresumedin GRB spectralmodels, i.e. truncatedpower-laws.Yet, this requirement stronglycontrastsparticledistributionsobtainedfromaccelerationsimulations,exemplifiedbythosedepictedinFig.1, andfurtherillustratedinnumerouspapers[6,7,9,16,17].Moreover,particle-in-cell(PIC)plasmasimulations[10– 12]generallyexhibitlargelyMaxwelliandistributions,andintheisolatedrecentsuggestion[13]ofanon-thermaltail generatedbydiffusivetransport,thethermalpopulationstronglydominatesthehigh-energytail. Theconsequenceisobviouslyapotentialconflictforaccelerationmodels,sincethenon-thermalelectronsarealmost always drawn directly from a thermalgas. This does not necessarily mean that acceleration at relativistic shocks is notprecipitatingthepromptemission.Itispossiblethatsomehow,relativisticshockscansuppressthermalizationof electrons,thoughsuchaconjecturepresentlyhasnosimulationalevidencetosupportit.Or,radiativeefficienciesmight becomesignificantonlyathighlysuperthermalenergies:aconvenientresolutiontothisdilemmaisthatstrongradiative self-absorptioncouldbeacting,inwhichcase n -Fn peaks(andthereforeBATSEandFermiGBMspectralprobes)are notactuallysamplingthermalelectrons.Itisalsopossiblethatothermechanismssuchaspitch-anglesynchrotron,or jitterradiationmayprovemoredesirable.Acautionarynoteisthatthiscomplicationfortheinterpretationispredicated upon spectra integratedover the burst duration,which can differsignificantly from those obtainedin temporalsub- intervals(e.g.see[28]).Notwithstanding,cumulativespectraprovideaglobalindicationofanarrownessoftheMeV spectralbreak,andthismustbeacharacteristicofatleastthemostluminousportionsofthebursttimehistory.Wenote alsothat[14]concludedthatthesynchrotronself-ComptonprocesswasunlikelytobeabletogeneratetheMeVband emission in prominentEGRET bursts because of its intrinsic spectralbreadth,even with extremelynarrow electron distributionfunctions.Onlyaself-absorptioninthesynchrotronseedspectrumcanaiditsviability. CONCLUSION Thispaperexplorestheconnectionbetweendiffusiveshockaccelerationtheoryandpromptemissioninbursts.Sim- ulation results presented clearly highlight the non-universalityof the index of energetic, non-thermalelectrons and ions,spawnedbythevarietyofshockobliquitiesandthecharacterofhydromagneticturbulenceintheirenvirons.This non-universalityposesnoproblemformodelingGRBhigh-energypower-lawindices,thoughobservationsgenerally constrainthe parameterspace tosubluminalorhighly-turbulentsuperluminalshocksnearthe Bohmdiffusionlimit. Also, demanding a connection between the thermal electron population and the GRB n -Fn peak gives rise to an inconsistency in that acceleration simulations predicta dominanceof the thermalpopulation,whereas the radiation modelsdemandthatthecontributingparticlesareintrinsicallynon-thermal.Escapeclausesincludeinvokingradiative selfabsorption,orperhapsrelinquishingpreferredemissionmechanisms,i.e.,synchrotronorinverseComptonradia- tion.Theprospectofanumberofplatinumstandard,broad-bandGRBdetectionsbyFermithatpermittime-dependent spectroscopyshouldhoneourunderstandingoftheconnectionbetweenshockaccelerationandpromptemission. Acknowledgments:thisresearchwassupportedinpartbyNASAgrantNNG05GD42GandNSFgrant0758158. REFERENCES 1. Rees,M.J.&Mészáros,P.1992,M.N.R.A.S.,258,41P. 2. Piran,T.1999,Phys.Rep., 314,575. 3. Mészáros,P.2002,Ann.Rev.Astron.Astr., 40,137. 4. Kirk,J.G.&Heavens,A.F.1989,M.N.R.A.S.,239,995. 5. Kirk,J.G.,Guthmann,A.W.,Gallant,Y.A.,Achterberg,A.2000,ApJ,542,235. 6. Ellison,D.C.,Jones,F.C.&Reynolds,S.P.1990,ApJ,360,702. 7. Ellison,D.C.&Double,G.P.2004,AstroparticlePhys.,22,323. 8. Niemiec,J.,&Ostrowski,M.2004,ApJ,610,851. 9. Stecker,F.W.,Baring,M.G.&Summerlin,E.J.2007,ApJ,667,L29. 10. Hoshino,M.,Arons,J.,Gallant,Y.A.&Langdon,A.B.1992ApJ,390,454. 11. Nishikawa,K.-I.,etal.2005,ApJ,622,927. 12. Medvedev,M.V.,etal.2005,ApJ,618,L75. 13. Spitkovsky,A.2008,ApJ,682,L5. 14. Baring,M.G.&Braby,M.L.2004,ApJ,613,460. 15. Jones,F.C.&Ellison,D.C.1991,SpaceSci.Rev.,58,259. 16. Ellison,D.C.,Baring,M.G.&Jones,F.C.1995,ApJ,453,873. 17. Summerlin,E.J.&Baring,M.G.2006,Adv.SpaceRes.,38(7),1281. 18. Baring,M.G.&Summerlin,E.J.2009,Adv.SpaceRes.,inpreparation. 19. Bell,A.R.1978,M.N.R.A.S.182,147. 20. Giacalone,J.,Burgess,D.,&Schwartz,S.J.1992,inESA,StudyoftheSolar-TerrestrialSystem,65. 21. Forman,M.A.,Jokipii,J.R.&Owens,A.J.1974,ApJ,192,535. 22. Baring,M.G.2004,Nucl.Phys.B, 136C,198. 23. deHoffman,F.&Teller,E.1950,Phys.Rev.D,80,692. 24. Band,D.L.,etal.1993,ApJ,413,281. 25. Baring,M.G.&2006,ApJ,650,1004. 26. Dingus,B.L.1995,Astr.SpaceSci.,231,187. 27. Preece,R.D.,etal.2000,ApJSupp.,126,19. 28. Abdo,A.A.,etal.(fortheFermiCollaboration)2009,Science,submitted. 29. Drury,L.O’C.1983,Rep.Prog.Phys.,46,973. 30. Tavani,M.1996,Phys.Rev.Lett.,76,3478. 31. Preece,R.D.,Briggs,M.S.,Mallozzi,R.S.,etal.1998,ApJ,506,L23.

See more

The list of books you might like

Most books are stored in the elastic cloud where traffic is expensive. For this reason, we have a limit on daily download.