ApJ, 679, 570, 2008

17 Pages·2008·1.41 MB·English

TheAstrophysicalJournal,679:570Y586,2008May20 A
#2008.TheAmericanAstronomicalSociety.Allrightsreserved.PrintedinU.S.A.
GLOBAL PROPERTIES OF X-RAY FLASHES AND X-RAYYRICH
GAMMA-RAY BURSTS OBSERVED BY SWIFT
T. Sakamoto,1,2 D. Hullinger,3 G. Sato,1,4 R. Yamazaki,5 L. Barbier,1 S. D. Barthelmy,1 J. R. Cummings,1,6
E. E. Fenimore,7 N. Gehrels,1 H. A. Krimm,1,8 D. Q. Lamb,9 C. B. Markwardt,1,10 J. P. Osborne,11
D. M. Palmer,7 A. M. Parsons,1 M. Stamatikos,1,2 and J. Tueller1
Received2007July14;accepted2008January27
ABSTRACT
We describe and discuss the spectral and temporal characteristics of the prompt emission and X-ray afterglow
emissionofX-rayflashes(XRFs)andX-rayYrichgamma-raybursts(XRRs)detectedandobservedbySwiftbetween
2004Decemberand2006September.Wecomparethesecharacteristicstoasampleofconventionalclassicalgamma-
raybursts(C-GRBs)observedduringthesameperiod.WeconfirmthecorrelationbetweenEobs andfluencenoted
peak
byothersandfindfurtherevidencethatXRFs,XRRs,andC-GRBsformacontinuum.Wealsoconfirmthatour
knownredshiftsampleisconsistentwiththecorrelationbetweenthepeakenergyintheGRBrestframe(Esrc )and
peak
theisotropicradiatedenergy(E ),theso-calledEsrc -E relation.Thespectralpropertiesof X-rayafterglowsof
iso peak iso
XRFsandC-GRBsaresimilar,butthetemporalpropertiesof XRFsandC-GRBsarequitedifferent.Wefoundthat
thelightcurvesof C-GRBafterglowsshowabreaktosteeperindices(shallow-to-steepbreak)atmuchearliertimes
thandoXRFafterglows.Moreover,theoverallluminosityof XRFX-rayafterglowsissystematicallysmallerbya
factorof 2ormorecomparedtothatof C-GRBs.ThesedistinctdifferencesbetweentheX-rayafterglowsof XRFs
andC-GRBsmaybethekeytounderstandingnotonlythemysteriousshallow-to-steepbreakinX-rayafterglowlight
curves,butalsotheuniquenatureof XRFs.
Subject headingsg: gamma rays: bursts — X-rays: bursts
Online material: color figures<![%ONLINE; [toctitlegrptitleOnline Material/title/titlegrplabelColor figures/
labelentrytitlegrptitlefigref rid="fg1" place="NO"Figure 1/figref/title/titlegrp/
entryentrytitlegrptitlefigref rid="fg4" place="NO"Figure 4/figref/title/titlegrp/
1e.nItNrTyRenOtDryUtCitTlIeOgNrptitlefigref rid="fg5" pStlraochem=a"yeNrOet"aFl.i(1g9u9r8e)id5en/fitifigerde2f/2tbitulrest/stoitblseegrvrepd/byGinga
entryentrytitlegrptitlefigref rid="fg6"thaptloaccceu=rr"edNObe"twFeiegnur1e9876/Mfiagrcrhefa/ntidtl1e9/9t1itOlecgtorbpe/r and for
Despite the rich gamma-ray burst (GRB) sample provided
entryentrytitlegrptitlefigref rid="fg7"whpiclhacteh=e"sNpeOct"raFcigouulrdebe7r/efiligarbelyf/atnitalleyz/teidt.leAgbropu/t 36% of
by BATSE (e.g., Paciesas et al. 1999; Kaneko et al. 2006),
entryentrytitlegrptitlefigref rid="fg8"GRpBlascoeb=se"rNveOd"bFyiGgiungraeha8d/fivegrryesfo/fttistpleec/ttriat.lTehgerypn/otedthat
BeppoSAX(e.g.,Frontera2004),Konus-Wind(e.g.,Ulanovetal.
entryentrytitlegrptitlefigref rid="fg9"thepsleacbuer=s"tsNrOes"eFmibgleudreBA9T/SfiEgrleofn/gtiGtlReB/tsitilnegdrupra/tion and
2005), and HETE-2 (e.g., Barraud et al. 2003; Sakamoto et al.
entryentrytitlegrptitlefigref rid="fg10"gepnelaracles=p"ecNtrOal"sFhiagpuer,ebut1t0h/efipgeraekfe/tnietrlgei/etsitolfegthrep(cid:2)/F spec-
2005), the emission properties of GRBs are still far from be- (cid:2)
entryentrytitlegrptitlefigref rid="fg11"trupml,aEceob=s",NexOte"nFdeigdutorelow1e1r/vfiaglureesf/tthiatnlet/htoisteleogfrtph/eBATSE
ingwellunderstood.Inrecentyears,however,anotherphenom- peak
entryentrytitlegrptitlefigref rid="fg1"buprlsatsc(eP=re"eNceOe"tFali.g20u0r0e;K1a2ne/fikogerteafl/.t2i0tl0e6/)t.iHtleeisgerept/al.(2003)
enonthatresemblesGRBsinalmosteveryway,exceptthatthe
entryentrytitlegrptitlefigref rid="fg13"reppolratecdet=h"atNaOm"oFngigthueresou1rc3e/sfiimgraegfe/dtibtyleth/teitWleidgerpF/ieldCam-
fluxcomesmostlyfromX-raysinsteadof (cid:1)-rays,hasbeendis-
entryentrytitlegrptitlefigref rid="fg14"erapsl(aWceF=C"sN)oOn"bFoiagrdurBeepp1o4S/AfiXgwreafs/atictllaes/stiotflefagsrtpX/-raytran-
coveredandstudied.Thisnewclassof burstshasbeendubbed
entryentrytitlegrptitlefigref rid="fg15"siepnltsawceit=h"dNurOati"oFnisgoufrleess1th5a/nfi1g0r0e0fs/ttihtaltew/teirtelneogtr‘p‘t/riggered’’
‘‘X-rayflashes’’(XRFs;Heiseetal.2003;Barraudetal.2003;
entryentrytitlegrptitlefigref rid="fg16"(i.ep.,ladceete=c"teNdO) b"yFitghuerGeam1m6/afiRgraeyf/Btuitrlset/Mtitolneigtorrp(/GRBM).
Sakamoto et al. 2005), and there is strong evidence to suggest
entryentrytitlegrptitlefigref rid="fg17"Thpislabceec=am"Ne Oth"eiFriwguorrkeing17de/fifingirtieofn/toitfleX/RtiFtlse.gKripp/pen et al.
that‘‘classical’’GRBs(C-GRBs)andXRFsarecloselyrelated
entryentrytitlegrptitlefigref rid="fg18"(20p0l3a)csee=a"rcNheOd"fFoirgCu-rGeRB1s8/afindgrXeRf/Ftsi,tlwe/htiicthlewgerrpe/observed
phenomena. Understanding what physical processes lead to
entryentrytitlegrptitlefigref rid="fg19" place="NsiOm"uFlitganuereou1s9ly/fibgyreWf/tFitCle/atintldegBrAp/TeSnEtr.y/Ttohce]y]>found 36 C-GRBs
their differences could yield important insights into their na-
and17XRFsina3.8yearperiod.JointWFCandBATSEspec-
tureandorigin.
tralanalysiswasperformedforthesample,andtheyfoundthat
XRFshaveasignificantlylowerEobs comparedwithC-GRBs.
peak
They also found that there is no systematic difference between
1 NASAGoddardSpaceFlightCenter,Greenbelt,MD20771. XRFs and C-GRBs in their low-energy photon indices, high-
2 OakRidgeAssociatedUniversities,P.O.Box117,OakRidge,TN37831- energyphotonindices,ordurations.Thesystematicspectralanal-
0117. ysisof asampleof 45HETE-2GRBsconfirmedthesespectral
3 Moxtek,Inc.,452West1260North,Orem,UT84057.
and temporal characteristics of XRFs. It is worth noting that
4 InstituteofSpaceandAstronauticalScience,JAXA,Kanagawa229-8510,
nine out of 16 XRF samples of HETE-2 have Eobs <20 keV
Japan. peak
5 Department of Physics, Hiroshima University, Higashi-Hiroshima 739- (Barraudetal.2003;Sakamotoetal.2005).
8526,Japan. Although the XRF prompt emission properties have been
6 JointCenterforAstrophysics,UniversityofMaryland,BaltimoreCounty, studied, until the launch of Swift (Gehrels et al. 2004), only
1000HilltopCircle,Baltimore,MD21250.
ahandfulof X-rayafterglowsassociatedwithXRFswerere-
7 LosAlamosNationalLaboratory,P.O.Box1663,LosAlamos,NM87545.
8 Universities Space Research Association, 10211 Wincopin Circle, Suite ported. D’Alessio et al. (2006) studied the prompt and after-
500,Columbia,MD21044-3432. glowemissionofXRFsandX-rayYrichGRBs(XRRs)observed
9 Department of Astronomy and Astrophysics, University of Chicago, by BeppoSAX and HETE-2. They found that the XRF and
Chicago,IL60637.
XRR afterglow light curves seem to be similar to those of
10 DepartmentofPhysics,UniversityofMaryland,CollegePark,MD20742.
11 Department of Physics and Astronomy, University of Leicester, LE1, C-GRBs,includingthebreakfeatureinthelightcurves.They
7RH,UK. also investigated the off-axis viewing scenarios of XRFs
570
SWIFT X-RAY FLASHES 571
for the top-hatYshaped jet (Yamazaki et al. 2002, 2004), the obtainbetterconstraintsonEobs.WefocusonX-rayafterglow
peak
universal power-lawYshaped jet (Rossi et al. 2002; Zhang & properties observed by the Swift XRTin this study. In x 2, we
Me´sza´ros2002;Lambetal.2005),andtheGaussianjet(Zhang discussourclassificationof GRBs,theanalysismethodsof the
et al. 2004), and they concluded that these models might be BATandtheXRTdata,andtheselectioncriteriaof oursample.
consistent with the data. Their sample, however, only contains Inxx3and4,weshowtheresultsofthepromptemissionandthe
nine XRFs/XRRs with measured X-ray afterglows. Further- X-ray afterglow analysis, respectively. We found distinct dif-
more, the data points in the X-ray light curves were not well ferences between XRFs and C-GRBs in the shape and in the
sampled, so that there are large uncertainties in the decay in- overallluminosityof X-rayafterglows.Wediscusstheimpli-
dicesandtheoverallstructuresof thelightcurveinmostcases. cationsof ourresultsinx5.Ourconclusionsaresummarized
Moreover,sincetheX-rayafterglowobservationsbegan>104s in x 6. We used the cosmological parameters of (cid:1) ¼0:3,
m
afterthetrigger,theirsampleisabletosaylittleabouttheearly (cid:1) ¼0:7,andH ¼70kms(cid:2)1Mpc(cid:2)1.Thequotederrorsare
(cid:2) 0
afterglow properties, which contain rich information that can atthe90%confidencelevelunlessstatedotherwise.
constrainjetmodelsforXRFs.OtherXRFtheoreticalmodels
2.ANALYSIS
are the inhomogeneous jet model (Toma et al. 2005), the in-
ternal shock emission from high bulk Lorentz factor shells 2.1.Working Definition of Swift GRBs and XRFs
(Mochkovitch et al. 2003; Barraud et al. 2005), the external
Thepreciseworkingdefinitionsadoptedbyotherswhohave
shock emission from low bulk Lorentz factor shells (Dermer
studiedXRFshavetended(understandably)tobebasedonthe
etal.1999;Dermer&Mitman2003),andtheX-rayemission
characteristics and energy sensitivities of the instruments that
from the hot cocoon of the GRB jet if viewed from off-axis
collectedthedata(Gotthelfetal.1996;Strohmayeretal.1998;
(Me´sza´rosetal.2002;Woosleyetal.2003).
Heiseetal.2003;Sakamotoetal.2005).Theeffectiveareaofthe
Because of the sophisticated on-board localization capabil-
BATissufficientlydifferentfromtheseotherinstrumentsthat
ity of the Swift Burst Alert Telescope (BAT; Barthelmy et al.
none of the definitions previously adopted are quite suitable
2005)andthefastspacecraftpointingof Swift,morethan90%
(Band2003,2006).Wedesireadefinition,however,thatwill
of SwiftGRBshaveanX-rayafterglowobservationfromthe
correspondtopreviousdefinitionssothatwemayreliablycom-
SwiftX-RayTelescope(XRT;Burrowsetal.2005a)withina
parethecharacteristicsoftheBAT-detectedXRFpopulationwith
fewhundredsecondsafterthetrigger.DuetothefactthatBAT
thosefromothermissions.Sakamotoetal.(2005)definedXRFs
issensitivetorelativelylowenergies(15Y150keV)andalsoa in terms of the fluence ratio S (2Y30 keV)/S (30Y400 keV),
largeeffectivearea((cid:1)1000cm2at20keVforasourceon-axis), X (cid:1)
andC-GRBs,XRRs,andXRFswereclassifiedaccordingtothis
BAT is also detecting XRFs and XRRs. However, because of
fluenceratio.Sakamotoetal.(2005)notedastrongcorrelation
BAT’s lack of response below 15 keV, it is very challenging betweentheobservedspectralpeakenergyEobs andthefluence
todetectXRFswithEpoebaskofafewkeV,whichdominatedtheXRF ratio.TheyfoundthattheborderEobs betweepneaXk RFsandXRRs
samplesof BeppoSAXandHETE-2(e.g.,Kippenetal.2003; is(cid:3)30keV,andtheborderEobs bepetawkeenXRRsandC-GRBsis
Sakamotoetal.2005).Nonetheless,Swifthasauniquecapabil- peak
(cid:3)100keV.
ity for studying detailed X-ray afterglow properties just after IntheBATenergyrange,afluenceratioofS(25Y50keV)/S(50Y
theburstforXRFsandXRRswithEobs k20keVforthefirsttime.
peak 100keV)ismorenaturalandeasiertomeasurewithconfidence.
The systematic study of the X-ray emissions of GRBs ob-
Wethereforechoseourworkingdefinitionintermsof thisratio.
served by XRT reveals a very complex power-law decay be-
Inordertoensurethatourdefinitionisclosetothatadoptedby
haviorconsistingtypicallyofaninitialverysteepdecay(t(cid:3) with
Sakamotoetal.(2005)wecalculatedthefluenceratioof aburst
(cid:2)10P(cid:3)1P(cid:2)2) (e.g., O’Brien et al. 2006; Sakamoto et al. for which the parameters of the Band function13 (Band et al.
2007), followed by a shallow decay ((cid:2)1P(cid:3)2P0), followed 1993)are(cid:3) ¼(cid:2)1,(cid:3) ¼(cid:2)2:5,andEobs ¼30keV.Theseval-
by a steeper decay ((cid:2)2P(cid:3) P(cid:2)1; e.g., Nousek et al. 2006; 1 2 peak
3 ues of (cid:3) and of (cid:3) are typical of the distributions for XRFs,
O’Brien et al. 2006; Willingale et al. 2007), sometimes fol- 1 2
XRRs, and C-GRBs found by BATSE (Preece et al. 2000;
lowedbyamuchsteeperdecay((cid:3) P(cid:2)2;e.g.,Willingaleetal.
4 Kanekoetal.2006),BeppoSAX(Kippenetal.2003),andHETE-2
2007)and,insomecases(about50%),overlaidX-rayflares(e.g.,
(Sakamotoetal.2005).Theratiothusfoundis1.32.Welikewise
Burrows et al. 2005b; Chincarini et al. 2007; Kocevski et al.
calculatedthefluenceratioof aburstforwhich(cid:3) ¼(cid:2)1; (cid:3) ¼
1 2
2007).Althoughthereisincreasingevidencethattheinitialvery (cid:2)2:5, and Eobs ¼100 keV, which was found to be 0.72. Our
steepdecaycomponent(cid:3) isatailof theGRBpromptemission peak
1 workingdefinitionof XRFs,XRRs,andC-GRBsthusbecomes
(e.g.,Liangetal.2006;Sakamotoetal.2007),theoriginof the
phase from a shallow (cid:3)2 to a steeper decay (cid:3)3 (hereafter a S(25Y50 keV)=S(50Y100 keV)(cid:4)0:72 ðC-GRBÞ;
shallow-to-steep decay) is still a mystery. Moreover, not all
0:72<S(25Y50 keV)=S(50Y100 keV)(cid:4)1:32 ðXRRÞ; ð1Þ
GRBs have a shallow-to-steep decay phase in their X-ray af-
terglow light curves. Thus, it is very important to investigate S(25Y50 keV)=S(50Y100 keV)>1:32 ðXRFÞ:
theX-rayafterglowlightcurvesofburstsalongwiththeirprompt
emission properties to find a difference in their characteristics Tochecktheconsistencyof ourdefinition,wecalculatedS(25Y
betweenC-GRBsandXRFs. 50 keV) and S(50Y100 keV) of the HETE-2 sample using the
Inthispaper,wereportthesystematicstudyofthepromptand best-fittime-averagedspectralparametersreportedinSakamoto
afterglowemissionof 10XRFsand17XRRsobservedbySwift et al. (2005). The 90% error in the fluences is calculated by
from 2004 December through 2006 September. Although the scalingtheassociatederrorinthenormalizationof thebest-fit
data from Swift BATare the primary data set for investigation spectral model. As shown inFigure 1, ourdefinition isconsis-
ofthepromptemissionproperties,wealsouseinformationfrom tentwiththeHETE-2definitionof XRFs,XRRs,andC-GRBs
Konus-Wind and HETE-2 as reported on the Gamma-ray burst (Sakamotoetal.2005).
Coordinate Network12 or in the literature, when available, to
13 f(E)¼K1E(cid:3)1exp½(cid:2)E(2þ(cid:3)1)/Epeak(cid:5) if E<((cid:3)1(cid:2)(cid:3)2)Epeak/(2þ(cid:3)1)
12 Seehttp://gcn.gsfc.nasa.gov/gcn_main.html. and f(E)¼K2E(cid:3)2 ifE(cid:6)((cid:3)1(cid:2)(cid:3)2)Epeak/(2þ(cid:3)1).
572 SAKAMOTO ETAL. Vol. 679
Fig. 1.—S(2Y30 keV)/S(30Y400keV) and S(25Y50 keV)/S(50Y100 keV) Fig.2.—RelationshipbetweenEobs derivedbytheBandfunctionwithafixed
peak
fluenceratiosofHETE-2bursts.Thedashedanddash-dottedlinescorrespondto high-energyphotonindex(cid:3)2¼(cid:2)2:3andEpoebaskderivedbytheC-Bandfunction
thebordersbetweenC-GRBsandXRRs,andbetweenXRRsandXRFs,respec- oraCPLmodel.
tively.[SeetheelectroniceditionoftheJournalforacolorversionofthisfigure.]
with (cid:5) of S(15Y150 keV)¼100:59 ergs cm(cid:2)2. In addition, the
2.2. Swift BAT Data Analysis
best-fitlognormaldistributionisusedforthet durationcen-
100
All the event data from Swift BAT are available through tering on t100 ¼101:74 s with a (cid:5) of t100 ¼100:53 s. The BAT
HEASARC at NASA Goddard Space Flight Center. We used energy response matrix used in the simulation corresponds to
thestandardBATsoftware(HEADAS6.1.1)andthelatestcal- anincidentangleof 30(cid:7),whichistheaverageof theBATGRB
ibration database (CALDB: 2006-05-30). The burst pipeline samples. We found equal or higher improvements in (cid:4)2 in
script,batgrbproduct,wasusedtoprocesstheBATeventdata. 62 simulated spectra out of 10,000. Thus, the chance prob-
Thexspecspectralfittingtool(ver.11.3.2)wasusedtofiteach ability of having an equal or higher (cid:4)(cid:4)2 of 6 with a CPL
spectrum. modelwhentheparentdistributionisacaseof aPLmodelis
For the time-averaged spectral analysis, we use the time in- 0.62%.
tervalfrom0%to100%of thetotalburstfluence(t interval) Because of the narrow energy band of the BAT, most of
100
calculatedbybattblocks.SincetheBATenergyresponsegen- theEobs valuesmeasuredfromtheBATspectraldataarebased
peak
erator,batdrmgen,performsthecalculationforafixedsinglein- on a CPL fit, but not on the Band function fit. For XRFs, we
cidentangleof thesource,itwillbeaproblemifthepositionof apply a constrained Band (hereafter C-Band) functionmethod
the source is moving during the time interval selected for the (Sakamoto et al. 2004) to constrain Eobs. However, there is
peak
spectral analysis due to the spacecraft slew. In this situation, a systematic problem in the Epoebask values derived by different
wecreatedtheresponsematricesforeach5speriodduringthe spectralmodels.Inparticular,fortheburstsforwhichthetrue
time interval taking into account the position of the GRB in spectralshapeistheBandfunction,thereisaknowneffectthat
detectorcoordinates.Wethenweightedtheseresponsematrices Eobs derivedfromaCPLmodelfithasasystematicallyhigher
peak
bythe5scountratesandcreatedtheaveragedresponsematrices valuethanEpoebaskderivedfromaBandfunctionfit(e.g.,Kaneko
usingaddrmf.Sincethespacecraftslewsabout1(cid:7)persecondin etal.2006;Cabreraetal.2007).Toinvestigatethiseffect,we
responsetoaGRBtrigger,wechose5sintervalstocalculatethe fit all the BAT GRB spectra for which Eobs are derived only
peak
energyresponseforevery5(cid:7). fromtheBATdatawithaBandfunctionwiththehigh-energy
We fit each spectrum with a power-law (PL) model14 and a photonindexfixedat(cid:3)2to(cid:2)2.3.Figure2showsEpoebaskderivedby
cutoffpower-law(CPL)model.15Thebest-fitspectralmodelis the Band function fixing (cid:3)2 ¼(cid:2)2:3 and Epoebask derived by a
determined based on the difference in (cid:4)2 between a PL and a CPL or a C-Band function. The Eobs values derived by the
peak
CPL fit. If (cid:4)(cid:4)2 between a PL and a CPL fit is greater than 6 Band function with fixing (cid:3)2 ¼(cid:2)2:3 and by a CPL model
(b(cid:4)et(cid:4)te2rr(cid:8)ep(cid:4)reP2sLe(cid:2)nta(cid:4)tiC2vPeLs>pec6t)r,awlmeoddeteelrfmorintehethdaattaa.CTPoLqumaondtiefylitshea afugnrecetiownithalisnoerargorrese.MwiotshtEofpoebEaskpoebdaskerviavleudesbdyetrhiveeBdabnydafCun-Bctaionnd
significanceofthisimprovement,weperformed10,000spectral withfixing(cid:3)2 ¼(cid:2)2:3towithinerrors.Therefore,weconclude
simulations taking into account the distributions of the power- that the systematic error in Eobs derived by different spec-
peak
lawphotonindexinaPLfit,thefluenceinthe15Y150keVbandin tralmodelsisnegligiblecomparedtothatof thestatisticalerror
aPLfit,andthet100durationof theBATGRBs(e.g.,Sakamoto assigned to Epoebask derived from the BAT spectral data alone.
et al. 2008), and we determined in how many cases a CPL fit Note that the BATspectral data include the systematic errors,
gives(cid:4)2improvementsof (cid:6)6overaPLfit.Weusedthebest-fit which are introduced to reproduce the canonical spectrum of
normaldistributionforthepower-lawphotonindexcenteringon theCrabnebulaobservedatvariousincidentangles(Sakamoto
1.65with(cid:5)of 0.36.Thebest-fitlognormaldistributionisused etal.2008).ToperformthesystematicstudyusingtheBATdata,
forthefluencecenteringonS(15Y150 keV)¼10(cid:2)5:92ergscm(cid:2)2 we only selected bursts for which the full BATevent data are
available.16
14 f(E)¼K50(E/50keV)(cid:3),whereK50isthenormalizationat50keVinunits
ofphotonscm(cid:2)2s(cid:2)1keV(cid:2)1. 16 We exclude bursts such as GRB 050820A, GRB 051008, and GRB
15 f(E)¼K50(E/50keV)(cid:3)exp½(cid:2)E(2þ(cid:3))/Epeak(cid:5). 060218becauseofincompleteeventdata.
No. 1, 2008 SWIFT X-RAY FLASHES 573
2.3. Swift XRT Data
We constructed a pipeline script to perform the XRTanaly-
sisinasystematicway.Thispipelinescriptanalysisiscomposed
of four parts: (1) data download from the Swift Science Data
Center (SDC); (2) an image analysis to find the source (X-ray
afterglow) and background regions; (3) a temporal analysis to
constructandfitthelightcurve;and(4)aspectralanalysis.The
screenedeventdataof theWindowTiming(WT)modeandthe
PhotonCounting(PC)modearedownloadedfromtheSDCand
used in our pipeline process. For the WT mode, only the data
of thefirstsegmentnumber(001)areselected.AllavailablePC
modedataareapplied.Thestandardgrades,grades0Y2forthe
WTmodeand0Y12forthePCmode,areusedintheanalysis.
Theanalysisisperformedinthe0.3Y10keVband.Thedetection
of an X-ray afterglow is done automatically using ximage as-
suming that an afterglow is the brightest X-ray source located
within 40 from the BATon-board position. However, in cases
whereasteadycatalogedbrightX-raysourceismisidentifiedas Fig.3.—DistributionsofthefluenceratioS(25Y50keV)/S(50Y100keV)for
anafterglow,wespecifythecoordinatesof theX-rayafterglow BAT(top)andHETE-2(bottom).Thedashedlinescorrespondtotheborders
manually. The source region of the PC mode is selected as a betweenC-GRBsandXRRs,andbetweenXRRsandXRFs.
circleof4700radius.ThebackgroundregionofthePCmodeisan
annulus in an outer radius of 15000 and an inner radius of 7000 countrate,thescriptalwayseliminatesacentralareawithin700
radius for the PC data and a 1400 ;6:70 box region for the WT
excludingthebackgroundX-raysourcesdetectedbyximagein
circularregionsof 4700 radius.FortheWTdata,therectangular data. The count rate derived from the region excluding the
regionof 1:60;6:70 isselectedasaforegroundregionusingan centralpartiscorrectedbytakingintoaccounttheshapeof the
ARFatanaveragedphotonenergy.Thespectralanalysisisper-
afterglowpositionderivedfromthePCmodedataasthecenter
formedusingonlythedataof <0.6countss(cid:2)1forthePCmode
of theregion.Thebackgroundregionisselectedtobeasquare
regionof6.70onasideexcludinga2:30;6:70rectangularregion and<100countss(cid:2)1fortheWTmode.
TwoGRBsinoursample,GRB050713AandGRB060206
centered at the afterglow position. The light curve is binned
haveabackgroundX-raysource(cid:1)2500 and(cid:1)1000,respectively,
basedonthenumberof photonsrequiredtomeetatleast5(cid:5)for
fromthepositionof theafterglow.Sinceitisdifficulttoexclude
thePCmodeand10(cid:5)fortheWTmodeineachlightcurvebin.
the contamination from the very closely located background
The light curve fitting starts with a single power law. Then,
additionalpower-lawcomponentsareaddedtominimize(cid:4)2of source,we excludedthelast portion of thelightcurves,which
haveaflatteningthatisverylikelyduetothecontaminationfrom
thefit.ComplicatedstructuressuchasX-rayflaresarealsowell
thebackgroundsource.
fitted with this algorithm. Although our pipeline script fits the
XRT light curve automatically for every GRB trigger by this 2.4. Sample of GRBs
algorithm, we excluded the time intervals during the X-ray
Wecalculatedthefluenceratiobetweenthe25Y50keVandthe
flares from the light curve data by visual inspections before
50Y100keVbandsderivedfromaPLmodelusingtheBATtime-
doingthefitbyourmethodbecausetheunderstandingof the
averaged spectrum for all Swift bursts detected between 2004
overall shape of the light curve is the primary interest in our
Decemberand2006September.ThenweclassifiedtheseGRBs
study.Theancillaryresponsefunction(ARF)filesarecreated
usingthedefinitiondescribedinx2.1.Outofatotalof158long
by xrtmkarffor the WTand the PC mode data individually.
GRBs,weclassified10asXRFs,97asXRRs,and51asC-GRBs.
The spectral fitting is performed by xspec 11.3.2 using an
absorbedpower-lawmodel17forboththeWTandthePCmode ThedistributionofthefluenceratioS(25Y50keV)/S(50Y100keV)
forthe158longGRBsisshowninFigure3.SimilartotheHETE-2
data.Foranabsorptionmodel,wefixtheGalacticabsorption
results (Sakamoto et al. 2005), the figure clearly shows that
ofDickey&Lockman(1990)attheGRBlocationandthenadd
Swift’sXRFs,XRRs,andC-GRBsalsoformasinglebroaddis-
anadditionalabsorptiontothemodel.Weusethexspeczwabs
tribution.Thisfigurealsoclearlyshowsthattheratioof thenum-
modelforknownredshiftGRBsapplyingthemeasuredredshift
berofBATXRFstoBATXRRsissmallerthanthatoftheHETE-2
to calculate the absorption associated to the source frame of
XRFsamples.AsdiscussedinBand(2006)thenumbersof each
GRBs.Thespectraarebinnedtoatleast20countsineachspec-
GRB class strongly depend on the sensitivity of the instrument.
tralbinbygrppha.Theconversionfactor fromacountrateto
anunabsorbed0.3Y10keVenergyfluxisalsocalculatedbased Thisproblembecomesmoreseriousfortheinstrumentsthatdonot
coverawideenergyrange,suchastheBAT.Thus,wewillnot
ontheresultof thetime-integratedspectralanalysis.
discusstheabsolutenumbersof eachGRBclassinthispaper.
A ‘‘pile-up’’ correction (e.g., Romano et al. 2006; Nousek
Since the determination of Eobs is crucial for our study,
et al. 2006; Evans et al. 2007) is applied during our pipeline peak
weonlyselectGRBshavingvaluesforEobs thatcanbedeter-
process. It assumes a ‘‘pile-up’’ effect exists whenever the un- peak
minedfromtheBATdataaloneorfromusingthedatafromother
correctedcountrateintheprocessedlightcurveexceeds0.6and
100countss(cid:2)1forthePCandtheWTmodes,respectively.Only GRBinstruments(Konus-WindandHETE-2).Sincewecanuse
theC-Bandfunctionmethodfor XRFs toconstrainEobs ifthe
thetimeintervalsthatareaffectedbythe‘‘pile-up’’asdescribed peak
photonindex(cid:3)inaPLfitismuchsteeperthan(cid:2)2intheBAT
in our definition above have corrections applied. Although the
spectrum,we select allbursts thathave(cid:3)<(cid:2)2ata90%con-
area of the spectral region affected by pile-up depends on its
fidence level. We exclude GRB 041224 from our sample be-
causethereisnoXRTobservation.WealsoexcludeGRB060614
17 Thewabs(cid:9)wabs(cid:9)pegpwrlworwabs(cid:9)zwabs(cid:9)pegpwrlwmodelinxspec. because there was no report on the time-averaged spectral
46 5 5059605 2526
6656863544 00 02010001010000
c3 0.0.0.0.0.0.0.0.0.0. 0.0.50.0.0.0.0.0.0.0.0.0.0.0.0.0.
2 (cid:10)(cid:10)(cid:10)(cid:10)(cid:10)(cid:10)(cid:10)(cid:10)(cid:10)(cid:10) (cid:10)(cid:10)8(cid:10)(cid:10)(cid:10)(cid:10)(cid:10)(cid:10)(cid:10)(cid:10)(cid:10)(cid:10)(cid:10)(cid:10)(cid:10)(cid:10)
SR 3146645445 380.29338425482338
1.2.1.1.1.1.1.1.1.1. 0.70.7 0.70.0.90.70.70.80.80.80.70.0.70.80.70.7
b
keV) 0.170.41.10.50.50.80.80.40.60.3 52220.11.40.90.61.50.41.43.11.520.51.50.7
0 (cid:10)(cid:10)(cid:10)(cid:10)(cid:10)(cid:10)(cid:10)(cid:10)(cid:10)(cid:10) (cid:10)(cid:10)(cid:10)(cid:10)(cid:10)(cid:10)(cid:10)(cid:10)(cid:10)(cid:10)(cid:10)(cid:10)(cid:10)(cid:10)(cid:10)(cid:10)(cid:10)
5
1 9705732382 12318840137506623
Y15 0.73.6.3.2.4.8.2.4.2. 1641550.33.8.12.17.8.15.80.16.149.77.11.
S(
hMo/Inst C-BandC-BandC-BandC-BandC-BandC-Band...C-BandC-BandC-Band .........iAT/KW.....................jKW...kKW...lKW...
r B
e
h
t
O
1
fobsEpeak 7þ2912(cid:2)6þ1710(cid:2)7þ2718(cid:2)6þ2217(cid:2)<19<33...8þ2318(cid:2)<27.6<23 .........117þ2180(cid:2).....................36þ8424(cid:2)...122þ5758(cid:2)...33(cid:10)...
4 1 2 6
1
e2(cid:4) ..................59.1......... 69.061.317.9...62.146.049.750.445.855.360.6...60.5...53.7...57.5
1(cid:2))
V
e
k
1(cid:2)
s s
Burst gK502(cid:2)cm ..................210þ915(cid:2)......... 10þ98(cid:2)13þ48(cid:2)30þ427(cid:2)...1390þ0111(cid:2)4þ23(cid:2)7þ53(cid:2)8þ75(cid:2)3þ31(cid:2)6þ44(cid:2)4þ42(cid:2)...10þ94(cid:2)...9þ06(cid:2)...7þ24(cid:2)
Swift otons 1 3227 131 1 1 2 1
L h
1 P p
f4 C 3(cid:2)
o 0
1TABLEProperties obsEpeakkeV)(1 ..................5þ2312(cid:2)......... 35þ0816(cid:2)19þ749(cid:2)4þ823(cid:2)...9þ436(cid:2)17þ618(cid:2)45þ7213(cid:2)19þ7410(cid:2)36þ6311(cid:2)38þ7813(cid:2)18þ588(cid:2)...21þ6310(cid:2)...28þ7311(cid:2)...25þ7211(cid:2)
n ( 1
o
si
omptEmis (cid:3) ..................:17þ:08(cid:2):12(cid:2)......... 1.00.2(cid:10)0.80.4(cid:10)1.00.1(cid:10)...:19þ:09:14(cid:2)1.40.3(cid:10):08þ:06(cid:2):06(cid:2)0.90.3(cid:10):06þ:10(cid:2):05(cid:2)1.20.3(cid:10):06þ:09(cid:2):05(cid:2)...:07þ:06(cid:2):06(cid:2)...1.20.3(cid:10)...0.90.4(cid:10)
r (cid:2)(cid:2)(cid:2) (cid:2) (cid:2) (cid:2) (cid:2) (cid:2)
P
7831967195 65486590665051864
e2 2.8.5.7.9.9.6.6.6.8. 6.8.6.0.5.5.6.9.2.4.1.4.0.0.4.3.0.
(cid:4) 8545456355 87677556567573647
1
7 3
4132322584 74 16213131121412
L dK50 13(cid:10)98(cid:10)12(cid:10)7(cid:10)9(cid:10)6(cid:10)12(cid:10)24(cid:10)49(cid:10)25(cid:10) 224(cid:10)94(cid:10)135028(cid:10)32(cid:10)67(cid:10)16(cid:10)75(cid:10)13(cid:10)74(cid:10)13(cid:10)362(cid:10)25(cid:10)67(cid:10)103(cid:10)62(cid:10)52(cid:10)
P
58 8 647 8 7 3748
2334 23 2 00 02010101010000
:03þ:04(cid:2)0.(cid:10)0.(cid:10)0.(cid:10)0.(cid:10):03þ:04(cid:2)0.(cid:10)0.(cid:10):02þ:03(cid:2)0.(cid:10) 0.(cid:10)0.(cid:10)760.(cid:10)0.(cid:10)0.(cid:10)0.(cid:10)0.(cid:10)0.(cid:10)0.(cid:10)0.(cid:10)0.(cid:10)0.(cid:10)0.(cid:10)0.(cid:10)0.(cid:10)0.(cid:10)
(cid:3) :24(cid:2)3.1(cid:2)2.4(cid:2)2.7(cid:2)2.8(cid:2):26(cid:2)2.6(cid:2)2.5(cid:2):25(cid:2)2.5(cid:2) 1.541.651.(cid:2)1.541.8(cid:2)1.901.551.651.7(cid:2)1.711.8(cid:2)1.571.7(cid:2)1.531.721.551.65
(cid:2)(cid:2) (cid:2) (cid:2)(cid:2)(cid:2) (cid:2) (cid:2) (cid:2)(cid:2)(cid:2)(cid:2)
aBATT100 6.43.050.347.326.665.365.79.79.98.7 76.149.512.8190.73.256.759.618.2157.312.6143.223.774.4230.310.6132.524.7
GRB 06......................16A....................14B....................19......................24......................19......................28B....................12......................23B....................26...................... 19B...................10......................25A...................13A...................15......................15B...................21B...................11A...................15......................06......................11A...................10A...................07......................14......................25......................04A...................27......................
4478824599 24578901122578899
0000000000 00000010000000000
5555566666 55555556666666666
0000000000 00000000000000000
FFFFFFFFFF RRRRRRRRRRRRRRRRR
RRRRRRRRRR RRRRRRRRRRRRRRRRR
XXXXXXXXXX XXXXXXXXXXXXXXXXX
574
653243433 2534
00000000010000
c3 0.0.0.0.0.0.0.0.0.0.0.0.0.0.
2 (cid:10)(cid:10)(cid:10)(cid:10)(cid:10)(cid:10)(cid:10)(cid:10)(cid:10)(cid:10)(cid:10)(cid:10)(cid:10)(cid:10)
R
S 95296651573843
0.60.60.60.50.60.50.60.60.60.0.50.60.60.6
b
)
V 7366134857 841
ke 0.2.1.1.3.2.2.1.0.2.31.1.1.
0 (cid:10)(cid:10)(cid:10)(cid:10)(cid:10)(cid:10)(cid:10)(cid:10)(cid:10)(cid:10)(cid:10)(cid:10)(cid:10)(cid:10)
5
1 92162671206560
Y15 11.50.41.88.82.63.61.63.16.22.1729.54.28.
S(
p
hMo/Inst .........mKWnKWoKW...AT/KWqHETErKWsKW.........
r B
e
h
t
O
fobsEpeak .........124(cid:10)216(cid:10)928(cid:10)...1934þ01830(cid:2)51þ3127(cid:2)224þ6158(cid:2)25þ2420(cid:2).........
034 1114
213 2
e2(cid:4) 45.444.845.5.........39.4............38.943.544.2
1(cid:2))
V
e
k
1(cid:2)
s
gK502(cid:2)cm 23þ715(cid:2)14þ510(cid:2)15þ110(cid:2).........4þ33(cid:2)............2(cid:10)5þ24(cid:2)5þ53(cid:2)
ns 434 1 321
o
ot
nued CPL 3(cid:2)ph
nti 10
o (
C
—
1BLE obsEpeakkeV) 39þ9516(cid:2)46þ1319(cid:2)12þ928(cid:2).........61þ2324(cid:2)............75þ0021(cid:2)117þ6839(cid:2)184þ5041(cid:2)
A ( 1 1 111
T e.
d
0.40.30.3 0.3 0.40.20.3 piso
PL ade2BAT(cid:3)(cid:4)(cid:3)KGRBT50100 RB050124....................6.01.470.082131158.70.7(cid:2)(cid:10)(cid:10)(cid:2)(cid:10)RB050128....................30.41.370.07172759.30.7(cid:2)(cid:10)(cid:10)(cid:2)(cid:10)RB050219A.................35.11.310.061234103.20.1(cid:2)(cid:10)(cid:10)(cid:2)(cid:10)...RB050326....................41.01.250.04216442.1(cid:2)(cid:10)(cid:10)...RB050401....................36.81.400.07231937.1(cid:2)(cid:10)(cid:10)...RB050603....................21.41.160.062891071.1(cid:2)(cid:10)(cid:10)RB050716....................90.11.370.0672352.50.8(cid:2)(cid:10)(cid:10)(cid:2)(cid:10)...RB050717....................209.21.300.0531148.5(cid:2)(cid:10)(cid:10)...RB050922C.................6.81.370.06247844.9(cid:2)(cid:10)(cid:10)...RB051109A.................45.41.50.251663.7(cid:2)(cid:10)(cid:10)...RB060105....................87.61.070.04191432.5(cid:2)(cid:10)(cid:10)RB060204B.................195.01.440.0917147.00.8(cid:2)(cid:10)(cid:10)(cid:2)(cid:10)RB060813....................36.71.360.04155354.11.0(cid:2)(cid:10)(cid:10)(cid:2)(cid:10)RB060908....................28.51.350.06103350.71.0(cid:2)(cid:10)(cid:10)(cid:2)(cid:10) aInseconds.Yb721(cid:2)(cid:2)(cid:2)BAT15150keVenergyfluencein10ergscmswiththeBATbest-fitmodel.YYcAfluenceratioofS(2550keV)/S(50100keV)derivedfromaPLfit.d4211(cid:2)(cid:2)(cid:2)(cid:2)In10photonscmskeV.eThedegreesoffreedominaPLfitandaCPLfitare57and56,respectively.fInkeV.g3211(cid:2)(cid:2)(cid:2)(cid:2)In10photonscmskeV.hobsSpectralfittingmodelused/GRBinstrumentthatreportsE.peakobsiMorrisetal.2007;EderivedfromaCPLmodel.peakobsjGolenetskiietal.2006a,GCNCirc.5113.EderivedfromaCPLmodel.peakobskGolenetskiietal.2006c,GCNCirc.5460.EderivedfromaCPLmodel.peaklobsGolenetskiietal.2006d,GCNCirc.5518.EderivedfromaBandmodel.peakobsmGolenetskiietal.2005a,GCNCirc.3152.EderivedfromaBandmodel.peakobsnGolenetskiietal.2005b,GCNCirc.3179.EderivedfromaBandmodelforthefirstepeakoobsGolenetskiietal.2005c,GCNCirc.3518.EderivedfromaBandmodel.peakobspKrimmetal.2006;EderivedfromaCPLmodel.peakobsqCrewetal.2005,GCNCirc.4021.EderivedfromaCPLmodel.peakobsrGolenetskiietal.2005d,GCNCirc.4238.EderivedfromaCPLmodel.peaksobsTashiroetal.2007;EderivedfromaCPLmodel.peak
GGGGGGGGGGGGGG
575
576 SAKAMOTO ETAL. Vol. 679
Fig.4.—S(25Y50keV)/S(50Y100keV)fluenceratiosandEobs valuesofBAT-
peak
detected bursts. The dashedlineshows the fluence ratiosas afunctionof Eobs
peak
assuming(cid:3)1¼(cid:2)1and(cid:3)2¼(cid:2)2:5intheBandfunction.Thedash-dottedlines
indicatetheboundariesbetweenXRFs,XRRs,andC-GRBs(eq.[1]inthetext).
[SeetheelectroniceditionoftheJournalforacolorversionofthisfigure.]
parameters by Konus-Wind (Golenetskii et al. 2006b). Based
ontheseselectioncriteria,atotalof 41GRBsareselected,in-
cluding10XRFs,17XRRs,and14C-GRBs.
3. PROMPT EMISSION
The spectral properties of the prompt emission for our
41GRBsaresummarizedinTable1.Figure4showstheS(25Y
50keV)/S(50Y100keV)fluenceratioversusEobs.Asseeninthe
peak
figure,Eobs oftheBATGRBsrangesfromafewtensofkeVtoa
peak
fewhundredsofkeV.ThisbroadcontinuousdistributionofEobs is
peak
consistentwiththeBeppoSAX(Kippenetal.2003)andHETE-2
(Barraudetal.2003;Sakamotoetal.2005)results.TheBATGRBs
followwellonthecurvecalculatedassuming(cid:3) ¼(cid:2)1and(cid:3) ¼
(cid:2)2:5fortheBandfunction.ThegapintheS(215Y50keV)/S(520Y Fig. 6.—Plotofthe15Y150keVfluenceandpeakspectralenergyEpoebaskof
XRFs,XRRs,andC-GRBsdetectedbyBAT.Thedashedanddash-dottedlinesare
100keV)fluenceratiofrom0.8to1.2inoursampleislikelydue thebestfitstothedatawithandwithouttakingintoaccounttheerrors,andthey
to a selection effect. Essentially, we selected burstsbased on the are given by log(Eobs)¼ 3:87þ0:33þ(0:33(cid:10)0:13)log½S(15Y150keV)(cid:5) and
measurement of Eobs forXRRs andC-GRBs.Thiscriterionis log(Epoebask)¼(5:46(cid:10)pe0ak:80)þ(0:(cid:2)602:1(cid:10)6 0:14)log½S(15Y150keV)(cid:5). Those bursts
peak forwhichEobs isderivedfromaconstrainedBandfunction,aCPL,andtheBand
more or less equivalent to selecting the bursts based on their peak
functionaremarkedassquares,circles,andtriangles,respectively.[Seetheelec-
brightness.Ontheotherhand,mostof theXRFswereselected
troniceditionoftheJournalforacolorversionofthisfigure.]
based on the photon index value in a PL fit ((cid:3)<(cid:2)2). This is
equivalenttoselectingbythesoftnessof thebursts.Therefore,
thereisadifferentwaytodistinguishbetweenXRFs,andXRRs
andC-GRBs.Evidently,asshowninFigure3,thereisnosuch
gapinthehistogramof thefluenceratiosfortheBATGRBsif
the whole burst sample has been examined. Therefore, we be-
lievethatthegapinthefluenceratioat0.8Y1.2isduetotheway
inwhichweselectedthebursts.
InFigure5,wecompareEobs inaCPLfitandthelow-energy
peak
photon indices (cid:3) for the BAT, HETE-2, and BATSE samples.
For both the HETE-2 (Sakamoto et al. 2005) and the BATSE
(Kanekoetal.2006)samples,weonlyplottedGRBswithaCPL
model as the best representative model for the time-averaged
spectrumtoreducethesystematicdifferencesinboth(cid:3)andEobs
peak
due to the different choices of spectral models (Kaneko et al.
2006).Asseeninthefigure,therangeof (cid:3)-valuesderivedfrom
theBATdataaloneareconsistentwiththeHETE-2andBATSE
results.Inaddition,wehaveconfirmedthatthe(cid:3)-valuesforXRFs
Fig.5.—Distributionofthelow-energyphotonindex(cid:3)andEobs inaCPL
peak andXRRs (GRBs with Eobs <100 keV)coverthesame range
model.ThesamplesofBAT,HETE-2,andBATSEareshownascircles,squares, peak
andtriangles,respectively.[SeetheelectroniceditionoftheJournalforacolor as for C-GRBs (GRBs with Epoebask >100 keV; Sakamoto et al.
versionofthisfigure.] 2005).
No. 1, 2008 SWIFT X-RAY FLASHES 577
Fig.7.—DistributionofE fortheSwift/BAT,HETE-2,andBATSEsamples.Thewhite,gray,andblackhistogramsrepresentE derivedbytheconstrainedBand
peak peak
function,aCPLmodel,andtheBandfunction,respectively.Theleft-sidearrowsareE withupperlimits.[SeetheelectroniceditionoftheJournalforacolorversionof
peak
thisfigure.]
The top panel of Figure 6 shows Eobs and the 15Y150 keV bottompanelof Figure6showsthedistributionbetweenEobs
peak peak
fluence, S(15Y150 keV), for the BAT GRBs. We note a corre- andS(1Y1000keV).
lationbetweenEobs andS(15Y150keV).Forthepurposeof the To take into account the errors associated with Eobs and
peak peak
correlationstudy,weassignedthemedianofthe90%confidence S(1Y1000keV)inourcalculationof thecorrelationcoefficient,
intervaltobethebest-fitvalueof Eobs,sothattheerrorswould wegenerate10,000randomnumbersassumingaGaussiandis-
peak
be symmetric. For cases in which we only have upper limits tributioninEobs andS(1Y1000keV)ofthecentralvalueandthe
peak
forEobs,weassignedthebest-fitvaluesofEobs tobethemedian error for each GRB in the sample. For GRBs only having the
peak peak
of 0and90%upperlimit,andweassignedthesymmetricerror upper limits in Eobs and/or S(1Y1000 keV), we use a uniform
peak
tobehalfthatvalue.Thelinearcorrelationcoefficientbetween distributiontogeneratetherandomnumbers.Thenwecalculate
log½S(15Y150 keV)(cid:5) and log(Eobs) is +0.76 for a sample of thelinearcorrelationcoefficient for the10,000 burstsamplein
peak
41GRBsusingthebest-fitvalues. log½Eobs(cid:5)Ylog½S(1Y1000 keV)(cid:5)spaceandmakeahistogramof
peak
The best-fit functions with and without taking into account thecalculatedcorrelationcoefficient.Thehighestpeakand68%
the errors are log(Eobs)¼3:87þ0:33þ(0:33(cid:10)0:03)log½S(15Y points from the highest value of the histogram are assigned as
peak (cid:2)0:16
150 keV)(cid:5) and log(Eobs)¼(5:46(cid:10)0:80)þ(0:62(cid:10)0:14)log thecentralvalueand1(cid:5)intervalof thecorrelationcoefficient.
½S(15Y150 keV)(cid:5),resppeaekctively. Weinvestigatethecorrelationcoefficientfor(1)GRBswithEobs
peak
Since the fluence in the 15Y150 keV band is not a good from aCPLmodel (sample A;total32GRBs); (2)GRBswith
quantity to examine the correlation with Eobs because of its a constrained Eobs from a C-Band model and a CPL model
peak peak
narrowenergyrangeof integration,wealsoinvestigatethecor- (sampleB;total37GRBs);and(3)all41GRBs(sampleC)to
relationbetweenEobs andthefluenceinthe1Y1000keVband, evaluate the systematic effect of Eobs due to the different
peak peak
S(1Y1000keV).ForGRBsthathavethemeasurementof Eobs spectralmodels(C-Bandvs.CPL).Thecalculatedcorrelation
peak
fromBATdataalone,wecalculateS(1Y1000keV)directlyfrom coefficientsareþ0:50þ0:11; þ0:63þ0:08,andþ0:68þ0:07(allin
(cid:2)0:12 (cid:2)0:10 (cid:2)0:08
a spectral fitting process using the Band function. Therefore, 1(cid:5)error)forsamplesA,B,andC,respectively.Theprobabili-
uncertaintyinthespectralparametersintheBandfunction,es- tiesof suchacorrelationoccurringbychanceineachsample
pecially in the high-energy photon index (cid:3) is also taken into sizeare3:4;10(cid:2)2Y2:4;10(cid:2)4,5:8;10(cid:2)4Y5:3;10(cid:2)7,and4:1;
2
account in an error calculation of the fluence. For GRBs for 10(cid:2)5Y2:3;10(cid:2)8 in the 1 (cid:5) interval for samples A, B, and C,
which we use E from the literature, we calculated the flu- respectively. Thus, the correlation between Eobs and the flu-
peak peak
enceusingthespectralparameterspresented intheliterature, enceisstillsignificantevenifweusethefluenceinthe1Y1000keV
and the error associated in the normalization of the best-fit bandandalsotakeintoaccounttheEobs derivedbythedifferent
peak
spectral model is used to calculate an error of the fluence. If spectralmodels.
the reported best-fit model is a CPL for these GRBs, we use ThehistogramsofEobs fortheSwift/BAT,HETE-2(Sakamoto
peak
(cid:3) ¼(cid:2)2:3to calculate the fluence in the Band function. The etal.2005),andBATSE(Kanekoetal.2006)samplesareshownin
2
578 SAKAMOTO ETAL. Vol. 679
Fig.9.—Isotropicequivalentenergy,E vs.thepeakenergyintheGRBrest
iso
frame,Esrc fortheknownredshiftBATGRBsinthiswork(circles),pre-Swift
peak
GRBs(dots)andtheknownredshiftSwiftGRBsobservedbyKonus-Windor
HETE-2(triangles).Thedashedlineisthebest-fitcorrelationreportedbyAmati
(2006)(Epsercak¼95keV½Eiso/ð1052 ergsÞ(cid:5)0:49).[Seetheelectroniceditionofthe
Journalforacolorversionofthisfigure.]
samplesiscalculatedbyscalingtheerrorinthenormalizationof
the best-fit spectral model. As clearly seen in the figure, S(15Y
150 keV) and Eobs of the BAT GRBs are consistent with both
peak
theHETE-2andtheBATSEsamples.Thestrongcorrelationbe-
tweenEobs andS(15Y150keV)stillexistsbycombiningtheBAT
peak
andtheHETE-2samples.Thecorrelationcoefficientcombin-
ingtheBATandHETE-2GRBsis+0.685for83samples.The
probabilityofsuchacorrelationoccurringbychanceis<10(cid:2)11.
Thebest-fitcorrelationfunctionsbetweenEobs andS(15Y150keV)
peak
withandwithouttakingintoaccounttheerrorsarelog(Eobs)¼
peak
2:74þ1:51þ(0:15(cid:10)0:02)log½S(15Y150keV)(cid:5) and log(Eobs)¼
(cid:2)0:08 peak
(4:77(cid:10)0:63)þ(0:52(cid:10)0:11)log½S(15Y150keV)(cid:5), respectively.
Fig.8.—Top:Plotofthe15Y150keVfluenceandpeakspectralenergyEobs However, as clearly shown in both Figures 7 and 8, the BAT
forBATandHETE-2samples.Bottom:Plotofthe15Y150keVfluenceandppeeaakk XRFsarenotsofter(orweaker)thantheHETE-2sample.This
spectralenergyEobs forBAT,HETE-2,andBATSEsamples.Thedashedanddash- is because of the higher observed energy band of the BAT
peak
dottedline are thebest fitto the BATandtheHETE-2datawith and without compared to that of the HETE-2 Wide-field X-ray Monitor
(ta0k:1in5g(cid:10)in0to:02ac)cloogu½nSt(1th5eY1e5rr0orkse,Van)(cid:5)dtahnedy aloreg(gEivoebsn)b¼y(l4o:g7(7Ep(cid:10)oebask0):6¼3)2þ:74(0(cid:2)þ:005::01285(cid:10)þ (WXM;2Y25keV)(Shirasakietal.2003).Thus,cautionmight
0:11)log½S(15Y150keV)(cid:5).[SeetheelectronicepdeiatkionoftheJournalforacolor beneededwhencomparingtheBATandHETE-2XRFsamples.
versionofthisfigure.] ItisalsoclearfromthefiguresthattheEobs distributionof the
peak
TABLE2
Figure 7. We notice a difference in the distributions of
Esrc andE DerivedfromtheBATData
Eobs for the three GRB instruments, especially between the peak iso
peak
BAT(ortheHETE-2)andtheBATSEdistributions.Applying
Esrc E
the two-sample Kolmogorov-Smirnov (K-S) test to the Epoebask GRB z (kpeeVak) (1052iseorgs) Instrument
distributions for the BATand HETE-2 samples, the BATand
BATSE samples, and the HETE-2 and BATSE samples, we 0504011.................. 2.9 467(cid:10)110 41(cid:10)8 Konus-Wind
findK-Stestprobabilitiesof 0.44,2:3;10(cid:2)9,and4:1;10(cid:2)16, 050416A................. 0.6535 28þ(cid:2)69 0:096(cid:2)þ00::000191 BAT
respectively. Based on these tests, we may conclude that 050525A................. 0.606 131þ(cid:2)43 2:5þ(cid:2)00::45 BAT
the BATSE GRB samples have a systematically higher Eobs 050603a.................. 2.821 1333(cid:10)107 70(cid:10)5 Konus-Wind
peak 050824.................... 0.83 <35 0:13þ0:10 BAT
than the BATand the HETE-2 samples. This is probably be- (cid:2)0:03
050922Ca................ 2.198 415(cid:10)111 6.1(cid:10)2.0 HETE-2
cause not only is the BATSE energy range higher than those 051109Aa............... 2.346 539(cid:10)200 7.5(cid:10)0.8 Konus-Wind
other instruments, but also the current BATSE spectral cata- 060115.................... 3.53 285þ63 6:3þ4:1 BAT
(cid:2)34 (cid:2)0:9
log only contains the bright GRBs, therefore systematically 060206.................... 4.048 394þ82 4:3þ2:1 BAT
(cid:2)46 (cid:2)0:9
selecting higher Eobs GRBs in the catalog (Kaneko et al. 060707.................... 3.425 279þ43 5:4þ2:3 BAT
peak (cid:2)28 (cid:2)1:0
2006). 060908b.................. 2.43 514þ224 9:8þ1:6 BAT
(cid:2)102 (cid:2)0:9
Figure8showsEobs andS(15Y150keV)oftheBAT,HETE-2, 060926.................... 3.20 <96.6 1:1þ3:5 BAT
andBATSEsamplesp.eaTkhefluenceinthe15Y150keVbandforthe 060927.................... 5.6 475þ(cid:2)7477 14:1(cid:2)þ(cid:2)022:::130 BAT
HETE-2andBATSEsamplesiscalculatedusingthebest-fitspec-
Note.—Theuncertaintyis1(cid:5).
tralmodelreportedinthecatalog(Sakamotoetal.2005;Kaneko a Esrc andE valuesfromAmati(2006).
peak iso
etal.2006).TheerrorinthefluencefortheHETE-2andBATSE b Thehigh-energyphotonindex(cid:3) oftheBandfunctionisfixedat(cid:2)2.3.
2
No. 1, 2008 SWIFT X-RAY FLASHES 579
TABLE3
XRTX-RaySpectralPropertiesof41SwiftBursts
WT PC
t t N t t N
start stop H start stop H
GRB (s) (s) (1021cm(cid:2)2) (cid:3)a (cid:4)2/dof (s) (s) (1021cm(cid:2)2) (cid:3)a (cid:4)2/dof
XRF050406......... 92 1.5 ; 105 ... (cid:2)2.3 32.0/16 1.1 ;104 1.4 ; 106 3:5þ5:9 (cid:2)3:5þ1:1 6.3/8
(cid:2)2:0 (cid:2)2:3
XRF050416A....... 85 1.4 ; 105 <11 (cid:2)2:4þ0:8 9.9/9 184 6.4 ; 106 5:6þ1:0 (cid:2)2.1(cid:10)0.1 81.4/100
(cid:2)1:5 (cid:2)0:9
XRF050714B....... 157 219 7:2þ1:2 (cid:2)5:8þ0:5 34.2/27 257 9.5 ; 105 2:9þ1:0 (cid:2)2:8þ0:3 21.9/17
þ1:0 (cid:2)0:6 (cid:2)0:8 (cid:2)0:4
XRF050819......... 147 202 <0.4 (cid:2)2:3þ0:2 7.3/10 239 6.3 ; 105 <2 (cid:2)2:2þ0:3 17.7/11
(cid:2)0:3 (cid:2)0:4
XRF050824......... ... ... ... ... ... 6.2 ; 103 2.1 ;106 2:4þ1:0 (cid:2)2.2(cid:10)0.2 29.4/39
(cid:2)0:9
XRF060219......... 126 5.7 ; 104 3:2þ6:9 <(cid:2)2.6 8.3/9 146 6.9 ; 105 3:2þ1:1 (cid:2)3:1þ0:4 23.4/19
(cid:2)2:9 (cid:2)0:9 (cid:2)0:5
XRF060428B....... 212 418 0.3(cid:10)0.1 (cid:2)2.8(cid:10)0.1 126.9/121 622 1.0 ; 106 <0.2 (cid:2)1.9(cid:10)0.1 17.2/30
XRF060512......... 110 155 0:6þ0:7 (cid:2)4:4þ0:6 13.9/10 3.7 ; 103 3.5 ; 105 <0.3 (cid:2)1:9þ0:1 24.6/15
(cid:2)0:5 (cid:2)0:7 (cid:2)0:2
XRF060923B....... ... ... ... ... ... 139 6.0 ; 103 3þ2 (cid:2)2:0þ0:4 2.6/8
(cid:2)1 (cid:2)0:5
XRF060926......... 66 875 25þ(cid:2)2187 (cid:2)1:9þ(cid:2)00::23 11.4/15 4.3 ; 103 2.8 ; 105 <40 (cid:2)1:8þ(cid:2)00::24 10.0/7
XRR050219B...... 3.2 ; 103 1.2 ; 105 0.6(cid:10)0.2 (cid:2)1:81þ0:08 152.6/161 9.1 ;103 3.2 ; 106 1:0þ0:6 (cid:2)2.0(cid:10)0.2 26.6/22
(cid:2)0:09 (cid:2)0:5
XRR050410......... 1.9 ; 103 7.9 ; 104 13þ18 <(cid:2)3.3 28.7/26 1.9 ; 103 9.2 ; 105 <8 (cid:2)1:7þ0:5 23.1/13
(cid:2)9 (cid:2)1:0
XRR050525A...... ... ... ... ... ... 5.9 ; 103 3.0 ; 106 2(cid:10)1 (cid:2)2.1(cid:10)0.2 31.8/41
XRR050713A...... 80 1.2 ; 104 2.4(cid:10)0.3 (cid:2)2:41þ0:08 146.5/166 4.3 ; 103 1.7 ; 106 2.5(cid:10)0.5 (cid:2)2.1(cid:10)0.1 57.6/78
(cid:2)0:09
XRR050815......... ... ... ... ... ... 89 1.8 ; 105 <2 (cid:2)1:8þ0:3 9.7/11
(cid:2)0:4
XRR050915B...... 150 6.5 ; 104 <0.5 (cid:2)2:6þ0:1 53.7/53 288 9.6 ; 105 <1 (cid:2)2:2þ0:2 25.7/24
(cid:2)0:2 (cid:2)0:3
XRR051021B...... 86 115 <10 (cid:2)1:2þ0:5 1.6/2 258 5.2 ; 105 <4 (cid:2)2:0þ0:2 9.1/14
(cid:2)1:1 (cid:2)0:4
XRR060111A...... 74 517 1.7(cid:10)0.1 (cid:2)2:33þ0:04 367.6/300 3.8 ; 103 7.6 ; 105 1:4þ0:5 (cid:2)2.2(cid:10)0.2 33.7/39
(cid:2)0:05 (cid:2)0:4
XRR060115......... 121 5.4 ; 103 <10 (cid:2)1:84þ0:08 78.9/85 616 4.6 ; 105 <8 (cid:2)2:3þ0:1 21.6/26
(cid:2)0:09 (cid:2)0:2
XRR060206......... 64 3.7 ; 104 14þ8 (cid:2)2:4þ0:1 72.3/79 1.7 ; 103 3.7 ; 106 12þ11 (cid:2)2:0þ0:1 46.5/45
(cid:2)7 (cid:2)0:2 (cid:2)10 (cid:2)0:2
XRR060211A...... 172 379 0.6(cid:10)0.2 (cid:2)1.95(cid:10)0.07 162.1/172 662 5.7 ; 105 1:3þ0:8 (cid:2)2.1(cid:10)0.2 16.2/23
(cid:2)0:7
XRR060510A...... 98 143 ... (cid:2)3.7 25.5/8 2.4 ; 104 5.7 ; 105 <0.4 (cid:2)2:03þ0:06 121.1/100
(cid:2)0:10
XRR060707......... 127 160 <6 (cid:2)1:8þ0:2 6.6/5 488 2.8 ; 106 10(cid:10)7 (cid:2)2.1(cid:10)0.1 33.6/39
(cid:2)0:3
XRR060814......... 163 5.2 ; 104 2.6(cid:10)0.2 (cid:2)2.01(cid:10)0.05 363.1/280 1.1 ;103 1.4 ; 106 3.1(cid:10)0.3 (cid:2)2.33(cid:10)0.08 169.6/158
XRR060825......... 199 1.1 ;105 <8 (cid:2)1:6þ0:5 5.4/4 92 5.9 ; 105 3þ4 (cid:2)1.9(cid:10)0.5 8.9/10
(cid:2)1:3 (cid:2)2
XRR060904A...... 97 2.1 ;103 1:8þ0:2 (cid:2)2:61þ0:07 255.7/208 5.4 ; 104 1.3 ; 106 3þ2 (cid:2)2:9þ0:5 10.3/10
(cid:2)0:1 (cid:2)0:08 (cid:2)1 (cid:2)0:8
XRR060927......... ... ... ... ... ... 147 2.1 ;105 <37 (cid:2)1.8(cid:10)0.2 6.7/12
GRB050124......... ... ... ... ... ... 1.1 ;104 5.0 ; 106 <0.8 (cid:2)1:9þ0:2 13.0/14
(cid:2)0:3
GRB050128......... ... ... ... ... ... 4.5 ; 103 9.9 ; 104 0:7þ0:3 (cid:2)2.1(cid:10)0.1 83.6/82
(cid:2)0:2
GRB050219A...... 112 5.7 ; 103 1:8þ0:5 (cid:2)2.1(cid:10)0.2 50.7/55 456 3.2 ; 106 <8 (cid:2)1:8þ0:5 6.6/4
(cid:2)0:4 (cid:2)1:3
GRB050326......... 3.3 ; 103 9.9 ; 103 0:9þ0:7 (cid:2)2:0þ0:2 22.3/25 5.0 ; 103 5.3 ; 105 0:6þ0:6 (cid:2)2.0(cid:10)0.2 27.3/26
(cid:2)0:6 (cid:2)0:3 (cid:2)0:5
GRB050401......... 133 8.5 ; 103 14(cid:10)2 (cid:2)1.91(cid:10)0.04 277.1/266 8.1 ;103 1.1 ;106 21þ17 (cid:2)2.0(cid:10)0.2 22.9/25
(cid:2)11
GRB050603......... ... ... ... ... ... 3.4 ; 104 1.8 ; 106 6(cid:10)4 (cid:2)1:98þ0:12 29.0/49
(cid:2)0:06
GRB050716......... 105 7.6 ; 104 <0.1 (cid:2)1:34þ0:03 208.9/202 4.1 ;103 1.8 ; 106 0.6(cid:10)0.5 (cid:2)2.1(cid:10)0.2 43.3/36
(cid:2)0:05
GRB050717......... 91 2.7 ; 104 1:8þ0:7 (cid:2)1.5(cid:10)0.1 110.7/105 4000 6.0 ; 105 <2 (cid:2)1:5þ0:2 23.6/15
(cid:2)0:6 (cid:2)0:3
GRB050922C...... 116 6.2 ; 104 <2 (cid:2)2.02(cid:10)0.07 107.9/124 3998 5.9 ; 105 7(cid:10)3 (cid:2)2:53þ0:07 60.2/49
(cid:2)0:08
GRB051109A...... 128 1.7 ; 104 <4 (cid:2)2.0(cid:10)0.1 42.9/32 3.4 ; 103 1.5 ; 106 5(cid:10)3 (cid:2)2.08(cid:10)0.07 130.7/129
GRB060105......... 97 4.6 ; 103 1.6(cid:10)0.1 (cid:2)1.99(cid:10)0.03 527.6/496 1.0 ; 104 3.8 ; 105 1.7(cid:10)0.4 (cid:2)2.2(cid:10)0.1 84.8/94
GRB060204B...... 103 1.8 ; 104 1.9(cid:10)0.2 (cid:2)2:28þ0:08 122.5/129 4.0 ; 103 8.1 ;105 1.3(cid:10)0.3 (cid:2)2:3þ0:1 54.9/56
(cid:2)0:09 (cid:2)0:2
GRB060813......... 85 7.6 ; 104 1.1(cid:10)0.4 (cid:2)1.88(cid:10)0.08 167.1/163 4.1 ;103 2.6 ; 105 1.3(cid:10)0.4 (cid:2)2.0(cid:10)0.1 105.1/102
GRB060908......... 80 1.3 ; 104 <8 (cid:2)2.3(cid:10)0.2 18.9/26 1.2 ; 103 1.1 ;106 <11 (cid:2)2:0þ0:2 13.7/14
(cid:2)0:3
a Thedefinitionofthephotonindex,(cid:3),isbasedonthespectralmodel: f(E)¼KE(cid:3).
BATSEsampleissystematicallyhighercomparedwiththeGRB knownredshiftGRBsissmall,wehaveconfirmedtheexistence
samplesof theHETE-2andtheBATbecauseof lackingsensi- andtheextensionof theEsrc -E relationtoXRFsandXRRs
peak iso
tivitybelow20keVforBATSE. (GRBswithEsrc <100keV)fortheSwiftGRBs(Amatietal.
peak
Figure9showsthecorrelationbetweenthepeakenergyinthe 2002;Lambetal.2005;Sakamotoetal.2004,2006).
GRBrestframeEsrc ((cid:8)(1þz)Eobs)andtheisotropicradiated
peak peak
energy Eiso. We calculated Epsercak and Eiso for the nine known 4. X-RAYAFTERGLOW EMISSION
redshiftGRBs18inoursampleusingtheBATdata(Table2).For
The spectral and temporal properties of the 41 X-ray after-
theseGRBs,E isderiveddirectlyfromthespectralfittingus-
iso
glowsaresummarizedinTables3and4.
ingtheBandfunctionandintegratingfrom1keVto10MeVat
theGRBrestframe.Esrc iscalculatedfromEobs basedonaCPL Figure10isacompositeplotoftheX-rayafterglowlightcurves.
fit.TheEsrc andE vpeaalkuesfortheremainingpSeawkiftGRBsareex- Figures11,12,and13showthelightcurvesineachGRBclass.As
peak iso we subsequently discuss in detail, we find that C-GRBs in our
tracted from Amati (2006). The values for the pre-Swift GRBs
sampletendtohaveafterglowswithshallowdecayindicesatearly
are also extracted from Amati (2006). Although our sample of
timesfollowedbysteeperindicesatlatertimesandthatthebreaks
18 WeexcludeGRB060512becauseof alesssecuremeasurementof its betweenthesetwoindicesoccuratabout103Y104s.Ontheother
redshift. hand,XRFafterglowsshowafairlyshallowdecayindexuntilthe