APreprinttypesetusingLTEXstyleemulateapjv.4/9/03
AHARDX-RAYVIEWONSCORPIUSX-1WITHINTEGRAL:NON-THERMALEMISSION?
T.DiSalvo1,P.Goldoni2,3,L.Stella4,M.vanderKlis5,A.Bazzano6,L.Burderi7,R.Farinelli8,F.Frontera8,
G.L.Israel4,M´endez9,F.Mirabel10,N.R.Robba1,P.Sizun3,P.Ubertini6,W.H.G.Lewin11
DraftversionFebruary5,2008
arXiv:astro-ph/0608335v1 16 Aug 2006ABSTRACT
WepresentheresimultaneousINTEGRAL/RXTEobservationsofScoX-1,andinparticularastudyofthehardX-rayemissionofthesourceanditscorrelationwiththepositionintheZ-trackoftheX-raycolor-colordiagram.WefindthatthehardX-ray(aboveabout30keV)emissionofScoX-1isdominatedbyapower-lawcomponentwithaphotonindexof∼3.Thefluxinthepower-lawcomponentslightlydecreaseswhenthesourcemovesinthecolor-colordiagraminthesenseofincreasinginferredmassaccretionratefromthehorizontalbranchtothenormalbranch/flaringbranchvertex.Itbecomesnotsignificantlydetectableintheflaringbranch,whereitsfluxhasdecreasedbyaboutanorderofmagnitude.TheseresultspresentcloseanalogiestothebehaviorofGX17+2,oneofso-calledSco-likeZsources.Finally,thehardpowerlawinthespectrumofScoX-1doesnotshowanyevidenceofahighenergycutoffupto100−200keV,stronglysuggestinganon-thermaloriginofthiscomponent.
Subjectheadings:accretiondiscs–stars:individual:ScoX-1—stars:neutronstars—X-ray:stars
—X-ray:general—X-ray:binaries
1.introduction
HardX-rayemissioninthebrightestlow-massX-raybinaries(hereafterLMXBs),theso-calledZ-sources,wasoccasionallydetectedinthepast(seee.g.Peterson&Jacobson1966).Theseresultsreceivedrelativelylittleattention,mostlybecausethelackofabroad-bandspec-tralcoveragedidnotpermittoestablishwhetheranextracomponentwasindeedrequiredtofitthehardspectrumofthesesources.RenewedinterestinthehardX-rayemissionpropertiesofbrightLMXBswasmotivatedbyrecentbroad-bandstudiesmainlyperformedwithRXTE(2–200keV)andBeppoSAX(0.1–200keV).ThesehaveshownthatmostZ-sourcesdisplayvariable,hardpower-lawshapedcomponents,dominatingtheirspectraabove∼30keV(seeDiSalvo&Stella2000forareview).
Thehardcomponentdetectedinbright(otherwisesoft)LMXBscanbefittedbyapowerlaw,withphotonindexintherange1.9–3.3,contributingfrom1%to10%
DipartimentodiScienzeFisicheedAstronomiche,Uni-versit`adiPalermo,viaArchirafi36-90123Palermo,Italy;email:disalvo@fisica.unipa.it
2APC,LaboratoireAstroparticuleetCosmologie,UMR7164,11PlaceMarcelinBerthelot,75231ParisCedex05,France.3CEASaclay,DSM/DAPNIA/Serviced’Astrophysique,F91191,Gif-sur-YvetteFrance.
4OsservatorioAstronomicodiRoma,viaFrascati33,00040MonteporzioCatone(Roma),Italy.
5AstronomicalInstitute”AntonPannekoek,”UniversityofAmsterdamandCenterforHigh-EnergyAstrophysics,Kruislaan403,NL1098SJAmsterdam,theNetherlands.
6IstitutodiAstrofisicaSpazialeeFisicaCosmica,SezionediRoma,INAF,viaFossodelCavaliere100,I-00133Rome,Italy.7Universit`adegliStudidiCagliari,DipartimentodiFisica,SPMonserrato-Sestu,KM0.7,09042Monserrato,Italy.8DipartimentodiFisica,Universit´adiFerrara,ViaParadiso12,44100Ferrara,Italy.
9SRONNetherlandsInstituteforSpaceResearch,Sorbonnelaan2,3584CAUtrecht,Netherlands.
10EuropeanSouthernObservatory-Vitacura,Casilla19001,Santiago19Chile.
11CenterforSpaceResearch,MassachusettsInstituteofTech-nology,77MassachusettsAvenue,Cambridge,MA02139-4307,USA.
1
oftheobserved(0.1−200keV)sourceluminosity.ThepresenceofthesecomponentsinZsourcesseemssome-timestoberelatedtothesourcestateoritspositionintheX-raycolor-colordiagram(hereafterCD).Theclear-estexampletodateisintheBeppoSAXobservationofGX17+2,wherethehardcomponent(apower-lawwithphotonindexof∼2.7)showedthestrongestintensityinthehorizontalbranch(HB)ofitsCD(DiSalvoetal.2000).Afactorof20decreasewasobservedwhenthesourcemovedfromtheHBtothenormalbranch(NB),i.e.fromlowtohigh(inferred)massaccretionrate.AhardtailwasalsodetectedinalmostallthecurrentlyknownZsources(e.g.DiSalvoetal.2001;Iariaetal.2001;DiSalvoetal.2002;Asaietal.1994).Thefactthatasimilarhardcomponenthasbeenobservedinsev-eralZsourcesindicatesthatthisisprobablyacommonfeatureofthesesources.However,theoriginofthishardcomponentisstillpoorlyunderstood.Whileinmostcasesthehardcomponentbecomesweakerathigherac-cretionrates,HEXTEobservationsofScoX–1showedahardpower-lawtailin5outof16observations,with-outanyclearcorrelationwiththepositionintheCD(D’Amicoetal.2001).Thethermalvs.non-thermalna-tureofthiscomponentremainstobeaddressed,yieldingimportantinformationontheproductionmechanism.ScoX–1,thebrightestpersistentX-raysourceinthesky,isalsothebrightestradiosourceamongneutronstarLMXBs,withameanradiofluxabout10timeshigherthanthatoftheotherZsources(e.g.Fender&Hendry2000).AhardX-raypower-lawcomponenthasbeenobservedinRXTE/HEXTE(20–200keV)dataofthissource(D’Amicoetal.2001).Asalreadymentioned,contrarytothecaseofGX17+2,inScoX–1thefluxofthiscomponentwasobservedtovarywithoutanyclearcorrelationwiththepositionintheCD.Interestingly,Strickman&Barret(2000)reportthatthehardX-rayemissionpresentinOSSEdataofScoX–1maybecor-relatedwithperiodsofradioflaring.
TostudythehardX-rayemissioninScoX-1,the
2
brightestofthesesources,aswellasitscorrelationwithothersourceproperties(suchasradioemissionandfasttimingvariability),wehaveperformedacampaignofob-servationsofScoX–1withINTEGRALandRXTE.PartoftheseobservationswerealsodonesimultaneouslywithradioVLBIobservations(whichwillbediscussedelse-where).TheINTEGRALspectrumofScoX-1showswithhighstatisticalsignificancethepresenceofahard(power-law)component,withoutanyclearexponentialcutoffupto∼100−200keV.WealsofindclearevidencethattheintensityofthiscomponentiscorrelatedwiththepositionofthesourceintheX-rayCD.
2.observationsandanalysis
ScoX-1wasobservedduringtwocompleteINTE-GRALrevolutionson2003July30–August1and2003August11–13.TheINTEGRALpayloadconsistsoftwomainγ-rayinstruments,aspectrometer,SPI(Ve-drenneetal.2003)andanimager,IBIS,andoftwomon-itorinstruments,theX-raymonitorJEM-X(3–35keV,Lundetal.2003)andtheOpticalMonitoringCamera(Vband,500-600nm,Mas-Hesseetal.2003).TheIBISinstrument(Ubertinietal.2003)coverstheenergyrangebetween20keVand8MeVwithtwodetectors,ISGRI(Lebrunetal.2003)andPICsIT(DiCoccoetal.2003)andhasafieldofviewof29◦withapoint×29◦athalfsensitivity(9◦×9◦fullycoded)spreadfunctionof12′FWHM.TheSPIalsocoverstheenergyrangebetween20keVand8MeVwithaFOVof31◦diameter(16◦fullycoded)andanangularresolutionof2.5◦.
ObservationswereperformedintheusualINTEGRALformat,i.e.theobservationwassplitinseparateexpo-sures(“sciencewindows”,orscws),eachlasting∼3600sec,followedbya5minutesslew.Theexposureswerearrangedfollowinga5×5ditherpattern.TheIBISandSPIinstrumentswereoperatedinstandardmodedur-ingthewholeobservation,whileJEM-Xwasinanon-standardmode(SPEC)which,unfortunately,isnotyetcalibrated.WethereforediscardedtheJEM-Xdatafromouranalysis.Theeffectiveexposuretimewas372ksforIBISand358ksforSPI12.TheIBISdatawereana-lyzedusingthestandardanalysisproceduresofOSAver-sion5andthelatestresponsematrices(October2004)rebinnedto26channelsbetween15and800keVandthelatestspectralextractionroutines(Goldwurmetal.2003).Asystematicerrorof1%wasappliedtoalltheINTEGRAL/ISGRIspectra.
Duringalltheobservations,ScoX-1wastheonlysourcedetectedinthewideFOVoftheinstruments.ThesourcecoordinatesasderivedfromtheISGRImosaicim-ageinthe20–35keVenergybandare:RA=161954.9,DEC=–153834.4(uncertainty±10′′,1sigmaconfi-dencelevel)atabout10′′fromtheSIMBADposition(McNamaraetal.2003).
TostudythesourcespectralstateinthestandardX-rayband,wealsoanalyzeddatafromthePropor-tionalCounterArray(PCA;Zhangetal.1993)onboardRXTE,whichconsistsoffiveco-alignedProportionalCounterUnits(PCUs),withatotalcollectingareaof6250cm2andafieldofview,limitedbycollimators,of1◦FWHM,sensitiveintheenergyrange2−60keV.
12
concentrateDuetoheretheonmuchtheloweranalysisstatisticsofISGRIofdatatheSPIofScodata,X-1.
wewillTheRXTEobservation,performedsimultaneouslywiththeINTEGRALobservation,isdividedintotwopartsof175kseach,approximatelytwoweeksapartfromeachother.Weselectedintervalsforwhichtheelevationan-gleofthesourceabovetheEarthlimbwasgreaterthan10degrees.Onafewoccasions,someofthefivePCUswereoff;wethereforeusedonlydatafromPCUs2and3whichwhereonformostoftheobservation.TheX-rayCDofScoX-1duringtheINTEGRALobservations,ob-tainedfromthePCAdata,isshowninFigure1.DuringeachofthetwoobservationsthesourcedescribedafairlycompleteZ-trackintheCD.
Inordertocheckforthepresenceofthehardtailinourdata,wefirstanalysedtheINTEGRAL(IBIS/ISGRI,en-ergyband20keV–200keV)spectrumintegratedoverthewholeobservation(seeFigure2,leftpanel).SimilartowhathasbeendoneforotherLMXBsoftheZ-class,wefittedtheINTEGRALspectrumwiththeComp-tonizationmodelcompTT(whosedescriptionisgiveninTitarchuk1994).Compttisananalyticalmodelde-scribingthethermalComptonizationofsoftphotons(forwhichaWienspectrumisassumed)inverse-ComptonscatteredinahotelectroncloudwithopticaldepthτandwhosetemperaturecanrangefromafewkeVto500keV.Thismodelincludesrelativisticeffectsandworksforopticallythickandopticallythinregimes.Thegeom-etryoftheComptonizingcloudcanbeeithersphericalordisk-like.Inthispaperweassumeasphericalgeome-try(thisassumptiononlyaffectsthevalueoftheopticaldepthderivedfromthefit).
ThismodelgivesagoodfitofthesoftpartoftheScoX-1spectrumupto∼40−50keV.Abovethisenergy,ahardexcessisclearlyvisibleintheresiduals,independentlyoftheparticularComptonizationmodelusedtofitthesoftpartofthespectrum.ThefitissignificantlyimprovedbyaddingtothecompTTmodelapower-lawwithpho-tonindex∼3.1(thisgivesareductionoftheχ2/doffrom892/13to14.4/11).Wetestedthepresenceofathermalcutoffinthehardpower-law;substitutingthepowerlawwithacutoffpower-law(thatisapowerlawmultipliedbyanexponentialcutoff)doesnotimprovethefitsignificantly(thelattermodelgivesaχ2/dof=15.6/10),andthetemperatureoftheexponentialcutoffiskT>200keV(90%confidencelevel).Thebest-fitparametersfortheISGRI(20–200keV)spectrumarereportedinTable1;dataandresidualswithrespecttothebestfitmodelareshowninFigure2(leftpanel).Tolookforvariabilityinthehardcomponentwiththespectralstateofthesource,asmeasuredbyitspositionintheX-rayCD,wedividedtheZ-trackintheCDofScoX-1intofourpartscorrespondingtotheHB/upper-NB,theNB,theflaringbranch(FB),andtheNB/FBvertex,re-spectively.WethereforeextractedINTEGRAL/ISGRIspectraforeachofthetimeintervalsmentionedabove,resultinginfourCDresolvedspectra.UnfortunatelytherewasnosuperpositionbetweenRXTEandINTE-GRALdataintheFBduringthefirstpartoftheobser-vation.TheINTEGRAL/ISGRIexposuretimesforthefourintervalswere52.4ks,45.5ks,60.6ksand6ks.Wefittedeachofthisspectrawithcompttandacutoffpowerlaw.Thismodelgaveagoodfitofthefirstthreespec-tra,withlittlevariabilityinthespectralparameters(seeTable1).Forthreeofthesespectra,thehardpowerlawcomponentwasrequiredinordertofitthedata.Remark-
ablytheFBspectrumdidnotrequireapower-lawcom-ponentandcouldbefittedwithasimpleComptonizationmodel(seeFig.2,rightpanel).IncludingthepowerlawinthespectralfitoftheFBspectrum,withthephotonindexfixedat3.1,agoodfitrequiresadecreaseofthenormalizationofthepowerlawcomponentbyafactoratleast5withrespecttotheaveragespectrum.InTable1weshowthetotalX-rayfluxofScoX-1calculatedinthe20−40keVandinthe40−200keVenergyrange,Flux(20–40)andFlux(40–200),respectively.WhiletheFlux(20–40)decreasesfromtheHBtotheNB/FBvertexandthenincreaseswhenthesourcegoestotheFB,theFlux(40–200)alwaysdecreases,varyingbyaboutoneorderofmagnitudewhenthesourcemovesfromtheHBtotheFB.
AlthoughfromTable1thereseemstobeacleartrendofthe40−200keVfluxtodecreasefromtheHBtotheNBandFBvertex,theuncertaintiesonthehardX-rayfluxarestillquitelargetodrawafirmconclusion.Infact,inordertocalculatethefluxinagivenenergyrangeandtheassociateduncertainty,XSPECneedstousethetotalbest-fitmodel,whichincludesboththecompttandthecutoffpower-lawcomponent;eachofthesecompo-nentshasseveralparametersandtheuncertaintiesonalltheparametersaretakenintoaccountinthecalcu-lationofthesourcefluxanditsuncertainty.Ifweareinterestedinknowingthefluxofthehardpower-lawcomponentaloneinagivenenergyrange(toseehowthiscomponentevolveswhenthesourcemovesintheX-rayCD),itismoreconvenienttousethemodelnamedpegpwrlwinXSPEC.Thismodelallowstodirectlycalcu-latethepower-lawfluxinagivenenergyrange(20−200keVisourchoice)andtheassociateduncertainty.Wethereforesubstitutedthecutoffpower-lawmodelwiththepegpwrlwmodeltofittheINTEGRALspectraofScoX-1.TheresultsofthefitswiththecompttpluspegpwrlwmodelarealsoreportedinTable1(sincetheparametersofthecompttdidnotchangesignificantlytheseareshownonlyonceinthetable).Thedecreaseofthepower-law20−200keVfluxalongtheCDisevident.Forclarity,wehavealsoplottedthesepower-lawfluxesinFigure1(inset).
3.discussion
Wereportonaspectralanalysisofasimultaneous∼and300RXTE.ks-longWeobservationshowthattheofScoadditionX-1withofaINTEGRALhardpower-lawcomponentdominatingtheX-rayspectrumabove∼GRAL30keV(20proves−200keV)necessaryspectrum.foraThegoodpowerfitoflawtheisINTE-quitesteep,withaphotonindexofabout3.1,contributingupto12%oftheobserved20−200keVluminosity,anddoesnotshowanyevidenceofahighenergycutoffupto100−200keV.SimilarlytowhatwasobservedintheBeppoSAXobservationoftheZsourceGX17+2,thepresenceofthehardcomponentinScoX-1seemstoberelatedtothepositionofthesourceintheZtrack;infactweobserveacleartrendofthe20−200keVpower-lawfluxtodecreasewhentheinferredmassaccretionrateincreases(i.e.fromtheHBtotheNB/FBvertex).Atthehighestinferredmassaccretionrate(i.e.intheFBoftheX-rayCD),thehardX-rayemissionseemstodisap-
3
pearcompletely(thehardX-rayfluxdecreasesbyaboutoneorderofmagnitudewithrespecttothefluxmeasuredintheHB/upper-NB).Note,however,thatthebehaviorofthehardcomponentintheFBmaybevariable(c.f.D’Amicoet.al.2001).
OneofthemostinterestingresultsofthisanalysisisthatthehardpowerlawdetectedinScoX-1doesnotshowanyevidenceofahighenergycutoff,withlowerlimitsontheelectrontemperaturethatareinmostcasesabove100keV(thelowerlimitonathermalcutoffintheScoX-1averagedINTEGRALspectrum,wherethestatisticsisthehighest,isabout200keV).SuchhightemperaturesarenotexpectedinasystemlikeScoX-1,becauseofthestrongComptoncoolingduetothepri-marysoftspectrum,wheremostoftheenergyisemit-ted.Theseresultsthereforeindicatethat,inanalogywiththesteephardpower-lawemissiondetectedinsomesoftstates(e.g.intermediateandveryhigh)ofsystemscontainingablackholecandidates,whichdonotshowanyenergycutoffuptoafewhundredskeV,thehardpowerlawobservedinZsourcesmaybeofnon-thermalorigin.
ThemostprobableoriginofthesecomponentsarethereforeComptonizationinahybridthermal/non-thermalcorona(wheretheMaxwellianvelocitydistri-butionoftheelectronshaveanon-thermalhigh-velocitytail,e.g.Poutanen&Coppi1998;seeFarinellietal.2005andD’A´ıetal.2006forasuccessfulapplicationofthismodeltothecaseofGX17+2andScoX-1,respectively).Forinstance,itispossiblethatahardpopulationofelec-tronsisacceleratedininternalshocksatthebaseofajet.Otherwisethehardpowerlawmayoriginateinamildlyrelativistic(v/c0.5)bulkmotionofmatterclosetothecompactobject.Thereforenon-thermalComptonizationinoutflows(orjets)maybetheoriginofthesecompo-nents,withflatterpowerlawscorrespondingtohigheropticaldepthsofthescatteringmediumand/orhigherbulkelectronvelocities(e.g.Psaltis2001).
Ithasbeenproposedthatnon-thermal,highenergyelectrons,responsibleforthehardtailsobservedinZsources,mightbeacceleratedinajet(DiSalvoetal.2000).Thisisalsoconfirmedbytheobservedcorrela-tionbetweenthestrengthofthehardX-rayemissionandthepositionofthesourceintheX-rayCD.InfactallZsourcesaredetectedasvariableradiosourceswiththehighestradiofluxassociatedwiththeHBandthelowestwiththeFB(Migliari&Fender2006,andrefer-encestherein).Inotherwords,boththeradioemissionfromtheseobjects,whichisprobablyduetojets(Fender&Hendry2000),andthehardX-rayemissionareanti-correlatedwiththeinferredmassaccretionrate.Inthisrespect,itwillbeveryimportanttostudythedirectcor-relationbetweenthepresenceandstrengthofthehardX-rayemissionandtheradioactivityofthesource.Theradio(VLA)observationsofScoX-1,simultaneoustopartoftheINTEGRALobservations,isunderinvestiga-tionandwillbediscussedelsewhere.
ThisworkwaspartiallysupportedbytheMinisterodellaIstruzione,dellaUniversit`aedellaRicerca(MIUR),andbytheAgenziaSpazialeItaliana(ASI).
4
Table1
ResultsofthefittingoftheScoX-1INTEGRAL/ISGRI(20–200keV)spectra.
Comptt+CutoffPower-lawModel
Comptt+PegpwrlwModel
ThemodelconsistsofaComptonizedspectrummodeledbycomptt,andacutoffpowerlaworthePegpwrlwmodeltofitthehardcomponent.kT0isthetemperatureoftheseedphoton(Wien)spectrum,kTetheelectrontemperatureandτtheopticaldepthinasphericalgeometry.Fluxes(20–40)and(40–200)arethetotalfluxfromthesourceinunitsof10−9ergcm−2s−1inthe20–40keVenergyrangeandinunitsof10−10ergcm−2s−1inthe40–200keVenergyrange,respectively.Ontheotherhand,Flux(20–200)isthepower-lawfluxinthe20–200keVrangeinunitsof10−10ergcm−2s−1.∆χ2isthevariationofχ2fortheadditionofapowerlawtothemodel.Notethattheadditionofapowerlawdoesnotchangetheχ2fortheFBspectrum.χ2/d.o.f.isthefinalreducedχ2forthebest-fitmodel.Alltheuncertaintiesarecalculatedat90%confidencelevel,upperlimitsat95%confidencelevel.
REFERENCES
[]Asai,K.,Dotani,T.,Mitsuda,K.,etal.1994,PASJ,46,479[]D’A´ı,A.,etal.2006,ApJ,submitted
[]D’Amico,F.,Heindl,W.A.,Rothschild,R.E.,Gruber,D.E.2001,
ApJ547,L147
[]DiCocco,G.,Caroli,E.,Celesti,E.,etal.2003,A&A,411,L189[]DiSalvo,T.,Stella,L.,Robba,N.R.,etal.2000,ApJ544,L119[]DiSalvo,T.,&Stella,L.2000,ProceedingsoftheXXXVIIth
RencontresdeMoriond,Eds.A.Goldwurm,D.Neumann,J.TranThanhVan,p.67
[]DiSalvo,T.,Robba,N.R.,Iaria,R.,etal.2001,ApJ,554,49[]DiSalvo,T.,Farinelli,R.,Burderi,L.,etal.2002,A&A,386,535[]Farinelli,R.,etal.2005,A&A,434,25
[]Fender,R.P.,&HendryM.A.2000,MNRAS,317,1
[]Fomalont,E.B.,Geldzahler,B.J.,&Bradshaw,C.F.2001aApJ,
553,L27
[]Fomalont,E.B.,Geldzahler,B.J.,&Bradshaw,C.F.2001bApJ,
558,283
[]Goldwurm,A.,David,P.,Foschini,L.,etal.2003,A&A,411,L223[]Iaria,R.,Burderi,L.,DiSalvo,T.,etal.2001,ApJ,547,412
[]Lebrun,F.,Leray,J.P.,Lavocat,P.,etal.2003,A&A,411,L141
[]Lund,N.,Budtz-Joergensen,C.,Westergaard,N.L.,etal.2003,
A&A,411,L231
[]McNamara,B.J.,etal.2003,AJ,125,1437
[]Mas-Hesse,J.M.,Gimenez,A.,Culhane,J.L.,etal.2003,A&A,
411,L261
[]Migliari,S.,&Fender,R.P.,2006,MNRAS,inpress[]Peterson,L.E.,&Jacobson,A.S.,1966,ApJ,145,962[]Poutanen,J.,&Coppi,P.S.,1998,Phys.Scripta,T77,57[]Psaltis,D.,2001,ApJ,555,786
[]Strickman,M.,&Barret,D.2000,Proc.ofthefifthCompton
Symposium,Eds.M.L.McConnellandJ.M.Ryan,AIPConferenceProceedings,Vol.510,p.222[]Titarchuk,L.1994,ApJ,434,570
[]Ubertini,P.,Lebrun,F.,DiCocco,G.,etal.2003,A&A,411,L131[]Vedrenne,G.,Roques,J.-P.,Sch¨onfelder,V.,etal.2003,A&A,
411,L63
[]Zhang,W.,Giles,A.B.,Jahoda,K.,Soong,Y.,Swank,J.H.,
Morgan,E.H.1993,Proc.SPIE2006,324
5
Fig.1.—Color-colordiagramofScoX–1fromPCAdataduringthesimultaneousINTEGRAL/RXTEobservation.HereSoftColoristheratioofthecountrateintheenergybands[3.5−6keV]/[2−3.5keV]andHardColoristheratio[9.7−16keV]/[6−9.7keV],respectively.TotakeintoaccountpossiblesmallgainchangesduringtheRXTEobservations,theHardandSoftcolorsofScoX-1arenormalizedwithrespecttothecolors(calculatedinthesameenergyrange)obtainedfromRXTEobservationsofCrabacquiredclosetothedatesofourobservations.OnlydataformPCUs2and3,whichwereonformostoftheobservation,havebeenused.Thefourcolorsindicatethefourregions(alsoindicatedwithnumbersfrom1to4)inwhichtheCDhasbeendivided,fromwhichINTEGRALspectrawereextracted.Inset:Themeasuredpower-lawfluxinthe20−200keVrangeplottedforeachofthefourCD-resolvedspectra.
Fig.2.—Left:ISGRI(20–200keV)averagedspectrum(left)andFBspectrum(right)ofScoX-1togetherwiththebest-fitmodel(solidlineontopofthedata)composedoftheComptonizationmodel(comptt,dashedline)andacutoffpowerlaw(dottedline).Right:residualsinunitsofσwithrespecttothebestfitmodelsshowninTable1.
因篇幅问题不能全部显示，请点此查看更多更全内容