A&Amanuscriptno. ASTRONOMY (willbeinsertedbyhandlater) AND Yourthesauruscodesare: ASTROPHYSICS 01(11.01.2;11.02.2Mrk501;13.25.2) Markarian 501 in X-ray bright state – RXTE observations 8 G.Lamer1andS.J.Wagner2,3 9 1 Institutfu¨rAstronomieundAstrophysik,Abt.Astronomie,Universita¨tTu¨bingen,Waldha¨userstr.64,D-72076Tu¨bingen,Germany 9 1 e-mail:[email protected] 2 LandessternwarteHeidelberg-Ko¨nigstuhl,D-69117Heidelberg,Germany n 3 MountStromlo&SidingSpringObservatories,PrivateBag,WestonCreekPO,ACT2611,Australia a J Received;accepted 2 1 v Abstract. Mrk501hasbeeninastateofveryhighfluxinX- 7 raysandVHE γ-raysduring1997.In July 1997near its hith- 0 ertomaximumX-raybrightnessintensemultifrequencyobser- 0 vationsofMrk501havebeenperformedatradio,nearinfrared, 1 optical,X-ray,andVHEγ-rayfrequencies. 0 8 Here we report on Rossi X-ray Timing Explorer (RXTE) 9 observationscarriedoutin1997betweenJuly11andJuly16. h/ TheX-rayspectrumhasbeenwelldetectedupto100keVand p is bestdescribedby a brokenpowerlaw with spectralindices - α1 = 0.70andα2 = 0.94belowandabovethe breakenergy o r ofEbreak =5.8keV. t TheX-rayfluxfromMrk501declinedandflaredby∼30% s a withinabout3dayseach,showinganunusualanti-correlation : v betweenfluxandspectralhardeningofbothpowerlawcompo- i nents.Thebreakenergyremainedconstant. X Fig.1. RXTE ASM light curve of Mrk 501 during 1996 and TheobservedbroadbandandX-rayspectraconfirmrecent r 1997 with the epochs of pointed observations with SAX and a observations,thatthe synchrotroncomponentin Mrk501can RXTEmarkedbytheshadedbars. extend up to energies of 100 keV with the maximum power beingemittedathardX-rays.Onlongertime-scalesthecut-off frequencyof the synchrotronspectrum changesby more than twoordersofmagnitude. athardX-rays,havenotbeenmeasuredduringearlierobserva- Key words: Galaxies: active – BL Lacertae objects:individ- tions(1975-1990,Ciliegietal.1995).The1keVfluxdensities ual:Mrk501–X-rays:galaxies measuredintheseobservationsvarybyafactorof∼3. Table1. Observationlog 1. Introduction Markarian501isthesecondclosest(z=0.034)BLLacertaeob- ObsID start end exposure[sec] ject.FromitsquiescentstateradiotoX-rayenergydistribution [MJD-50000.0] PCA HEXTEa (αrx ≈ 0.62, fν ∝ ν−α) it can be described as intermediate 20421-01-01-01 640.9746 640.9916 1456 873 between X-ray bright BL Lacs (XBLs) and radio bright ob- 20421-01-01-00 641.1764 641.1916 1296 1709 jects(RBLs).TheobjectusuallydisplaysarelativelysteepX- 20421-01-02-01 641.9740 641.9918 1520 640 ray spectrum (e.g. αx = 1.77 observed by ROSAT in 1991) 20421-01-02-00 642.1519 642.1919 2688 1771 indicating that the energy loss dominated high energy part of 20421-01-03-01 642.9733 642.9918 1584 971 the synchrotronspectrum is observedin X-rays (Lamer et al. 20421-01-03-00 643.1520 643.1920 2880 1843 1996).X-rayspectrawithα significantlylessthanunity,asre- 20421-01-04-01 643.9526 643.9922 1616 1078 x quiredforthemaximumofthesynchrotronpoweroutputtobe 20421-01-04-00 644.1526 644.1923 3152 996 20421-01-05-01 644.9530 644.9922 1744 1129 Sendoffprintrequeststo:G.Lamer,DepartmentofPhysicsandAs- 20421-01-05-00 645.1530 645.1922 3376 2135 2 G.Lamer&S.J.Wagner:Markarian501inX-raybrightstate–RXTEobservations Table2. Resultsofbrokenpowerlawfitstotheindividualspectra MJD duration α1 ±(1σ) Ebreak ±(1σ) α2 ±(1σ) 1keVnorm ±(1σ) χ2(d.o.f) [days] (E <Ebreak) [keV] (E >Ebreak) [(cm2 skeV)−1] 50640.9831 0.0170 0.756 0.015 5.519 0.149 1.054 0.008 0.1702 0.0030 1.03(224) 50641.1840 0.0152 0.767 0.013 5.984 0.186 1.028 0.009 0.1666 0.0030 1.04(224) 50641.9829 0.0178 0.664 0.016 5.380 0.148 0.962 0.007 0.1400 0.0030 1.04(224) 50642.1719 0.0400 0.687 0.010 5.856 0.143 0.922 0.006 0.1421 0.0020 0.95(224) 50642.9826 0.0185 0.653 0.014 5.621 0.189 0.889 0.008 0.1306 0.0026 0.96(224) 50643.1720 0.0400 0.669 0.011 5.604 0.152 0.893 0.006 0.1227 0.0020 1.11(224) 50643.9724 0.0396 0.674 0.015 5.546 0.161 0.944 0.007 0.1381 0.0030 0.98(224) 50644.1725 0.0396 0.722 0.009 5.889 0.131 0.956 0.005 0.1555 0.0018 0.98(224) 50644.9726 0.0393 0.746 0.010 5.981 0.140 1.029 0.007 0.1803 0.0032 0.97(224) 50645.1726 0.0393 0.730 0.008 5.724 0.090 1.035 0.005 0.1780 0.0020 1.02(224) integralspectrum 0.702 0.019 5.576 0.208 0.960 0.005 0.1510 0.0040 0.98a(224) a1%systematicerrorallowed In1995Mrk501wasdetectedasasourceofTeVgamma- Afterapplyingtheaboveelevationandoffsetcriteria,HEX- raysbytheWhippleteam(Quinnetal.1996)forthefirsttime. TEspectraandlightcurveshavebeenbinnedandbackground Its flux level corresponded to 8% of that of the Crab Neb- subtractedusingtheoff-sourcelookingintervals. ula.ThedetectionwasconfirmedbytheHEGRAcollaboration Spectral fitting was performed with XSPEC 10.0 using (Bradburyet al. 1997). A recentcompilationof TeV observa- the latest detector response matrix from 26 August 1997 for tions of Mrk 501 (Protheroe et al. 1997) shows the source at PCA andtheresponsefilesfrom20March1997forHEXTE. very high level with flares up to 10 Crab throughout the en- ThePCAspectrainthePHAchannelrange3-60(2.5-27keV) tire observing season of 1997 well into the July observations andthespectrafrombothHEXTEclustersinthechannelrange presentedhere. 15-100(15-100keV)werecombinedforeachofthe10point- AsimilarbrighteningofMrk501during1997hasbeenob- ingsandthenfittedwithsingleandbrokenpowerlawspectra. served in X-rays by the RXTE All Sky Monitor (ASM), the TheuncertaintyintheabsolutefluxcalibrationoftheHEXTE peakbrightnessinJune1997being∼40mCrab(Fig.1).The instrument, mainly due to uncorrected dead-time effects (see TeVandASM(2-10keV)lightcurvesshowsimilarflaringac- Rothschild et al. 1997), was accounted for by allowing for a tivity during1997,the correlatedX-ray and VHE gamma-ray scaling factor between the PCA and HEXTE spectra. Setting variabilitywillbediscussedinaforthcomingpaper. PCAto1.0,themeanbestfitvaluefortheHEXTEnormaliza- tionwas0.65. AttheonsetofthishighactivityinApril1997observations The absorbingcolumndensity on the line of sightto Mrk withtheSAXsatellitewerecarriedout(Pianetal.1997).They 501wasfixedto2.87·1020cm−2 asderivedfromtheROSAT findMrk501withanunusuallyhardX-rayspectrum,extend- spectrum (Lamer et al. 1996). This value is larger than the inguptophotonenergiesof100keVwiththehardestspectrum galactic value as derived from 21 cm H I measurements observedon16April,whenthesourcewasbrightest. (1.73·1020cm−2,Starketal.1992).However,theeffectofthis difference on the spectral results in the RXTE energy range is negligible. Note that Pian et al. 1997 used the galactic H I 2. Observationsanddataanalysis column density NH =1.73·1020cm−2 for the spectral fits to theSAXdataandthattheuncertaintyintheNHvaluetowards ThebrightX-raystateofMrk501inJune/July1997triggered thesourcemayhaveasignificanteffectonthespectralindices intensemulti-frequencyobservationsfromradiotoVHEgam- measuredwiththeLECSinstrumentatthesoftestX-rayener- ma-rays. Here we report on RXTE observations between 11 gies. Julyand16July1997resultingin21300secondsofgooddata in10pointings(Tab.1). 3. Spectraandvariability Wehaveusedftools 4.0forthereductionofthePCA and HEXTE data. PCA good times have been selected from Single power law models with N fixed to any reasonable H thestandard2modedatasetsusingthefollowingcriteria:tar- value do not yield acceptable fits to the data of any pointing, getelevation> 10◦, pointingoffset< 0.01◦, and all 5 PCUs whereasbrokenpowerlawmodelswithbreakenergiesbetween switchedon.Fortheresultingintervalsspectraandlightcurves 5.5and6.0keVgiveexcellentfitstotheindividualspectra(see wereextractedusingallXenonlayers.ThecorrespondingPCA Tab. 2). Below the break point the model spectra are excep- G.Lamer&S.J.Wagner:Markarian501inX-raybrightstate–RXTEobservations 3 beenobservedinanyotherXBL(Ciliegietal.1995,Lamerat andamplitudeshavebeenfound.Inaninvestgationofthespec- al. 1996). The spectrum measured on 16 April 1997 by SAX tral variability of 6 BL Lac objectsGiommiet al. foundtight in aboutthe same energyband (2.14-10keV) is even slightly correlationsoffluxandspectralhardnessforanyoftheobjects, harder(α=0.59,Pianetal.1997).AsthePCAhasnosensitiv- includingMrk501.Thishighlightsthepeculiarityofthespec- itybelow2keV,wearenotabletoverifythebreakat∼2keV tralvariabilityreportedhere. intheSAXspectrum. 4. Conclusions OurRXTEobservationsfromJuly1997showthattheperiodof increasedX-raybrightnesscontinuedthroughout1997andwas notlimited to a shortflare in April.The objectclearly was in along-lastinghigh-statein1997ratherthanexhibitingashort, spectacularX-rayflareduringtheepochofthe SAX observa- tions.WeconfirmtheextraordinaryhardX-rayspectrumwhich extendsupto100keV.Thepeakofthesynchrotronemissionis observedinthehardX-rays,morethan2ordersofmagnitude higher than in earlier observations (e.g. by ROSAT in 1991, Lameretal.1996). Fig.2. Top panel: PCA and HEXTE countrate spectra of the DuringtheRXTEobservationsinJulythetotalfluxandthe entire observation fitted with a broken power law model (see spectral hardness show an anti-correlation.This is a very un- Tab. 2 for the modelparameters).Bottom panel: Fit residuals usualspectralbehaviourforanyBLLacobject,ithasnotbeen with1%systematicuncertaintyallowed. observed previously in Mrk 501 and is in marked contrast to thefactthattheSAXobservationsofPianetal.showedaflatter spectrumduringthebrighterstage(thereareonlytwodifferent AbovethebreakpointthePCAspectrumsteepensbyupto brightnessandspectrallevelsintheSAXdatasetratherthana 0.3 in energyindex. In some of the spectra α2 exceeds unity, well-establishedcorrelation).Theflux-spectralindexrelations indicatingthatthemaximumofthesynchrotronpoweroutputis should hence not be regarded as universal and viable models reachedinthehardX-rayrange.However,thereisnoevidence shouldexplainbothkindsofbehaviour. forfurthersteepeningofthespectrumupto100keV. Thebrokenpowerlaw(orgradualsteepeningofthe spec- Duringthe RXTE observationsthe PCA count-ratevaried trum) is consistent with synchrotroncoolingof a single com- by ∼ 30%, but in contrast to the results from the April 1997 ponentanddoesnotrequirethesuperpositionofdifferentpar- SAXobservationsananti-correlationoffluxandspectralhard- ticle distributions.A homogeneousjetandamagneticfieldof ness was observed both below and above the spectral break. 0.4G are capable of producing the integrated spectral signa- Such an anti-correlation has never been reported in Mrk 501 ture.Wehavefoundratherdramaticspectraldifferencesinthe andisquiteunusualforX-rayBLLacobjectsaltogether.Pian lightcurveswhichcannotbeexplainedbyasinglepopulation et al. found a spectral hardening by ∆α = 0.32 in the SAX butrequiresthecontributionsofatleasttwo,ifnotmorespec- MECS 2-10 keV band during an increase of the flux in this tralcomponents.Thisillustratesthegreatimportanceofwell- energybandbyafactorof2.4. sampledmonitoringwithinstrumentscoveringawidespectral Thelargedegreeofspectralvariabilityanditsanti-correlat- range. ion with total flux implies greatdifferencesbetween the light curvesintheindividualX-rayenergybands.Wehencederived Acknowledgements. WethankJ.Swankforacceptingtherequestfor lightcurvesforsevenandfourdifferentchannelrangesofthe RXTETOOobservationsofMrk501,andtheRXTEschedulersfor PCA and HEXTE detectors,respectively(Figs. 3 and 4). The the flexible observation planning. D. Gruber and W. Heindl are ac- light curves in the softer bands show a decline by 20% over knowledged for their assistance with the HEXTE data analysis. We thankR.Staubertforacarefulreadingofthemanuscript.RXTEASM 3 days with a well defined minimum at MJD 50643.2 and a dataareprovidedbytheASM/RXTEteamsatMITandGSFC.PCA subsequentincrease by 30% during the following 2 days. On andHEXTEdatawereobtainedthroughtheHEASARCOnlineSer- the other hand the hardest bands are dominated by a feature vice,providedbyNASA/GSFC.ThisworkwassupportedbyDARA with a broadmaximumaroundMJD 50642.A smoothtransi- undergrant50OR96031,andtheDFG(SFB328). tion with superpositions of both morphologies is observed in the intermediateenergybands.The independentvariabilityof softandhardX-raysrequirestheexistenceofatleasttwoemis- References sioncomponentswhichisnotevidentfromtheX-rayspectrum Bradbury,S.M.etal.,1997,A&A320,L5 alone. Ciliegi,P.,Bassani,L.,andCaroli,E.,1995,ApJ439,80 The sharpest feature in the 2-20 keV lightcurves is a de- Giommi,P.,Barr,P.,Garilli,B.,etal.,1990,ApJ356,432 4 G.Lamer&S.J.Wagner:Markarian501inX-raybrightstate–RXTEobservations Protheroe, R.J., Bhat, C.L., Fleury, P., et al., 1997, Proc. 25th. Int. cosmicRayConf.,Durban1997,inpress Rothschild,R.E.,Blanco,P.R.,Gruber,D.E.etal.,1997,ApJ,inpress Quinn,J.etal.1996,ApJ456,L83 Stark,A.A.,Gammie,C.F.,Wilson,R.W.,etal.,1992,ApJS79,77 Fig.3. Background subtracted PCA light curves in different energybands. G.Lamer&S.J.Wagner:Markarian501inX-raybrightstate–RXTEobservations 5 Fig.4. BackgroundsubtractedHEXTElightcurvesindifferent energy bands, the count-rates of clusters A and B have been added.NotethegoodagreementofthePCAandHEXTElight curvesinthe3energybandsbetween15and80keV.