Astronomy&Astrophysicsmanuscriptno.ms (cid:13)c ESO2009 January26,2009 LettertotheEditor A candidate tidal disruption event in the Galaxy cluster Abell 3571 N.Cappelluti1,2,M.Ajello3,4,P.Rebusco5,S.Komossa1,A.Bongiorno1,C.Clemens1,M.Salvato6,P.Esquej7,1,T. Aldcroft8,J.Greiner1,andH.Quintana9 9 0 1 Max-Planck-Institute fu¨r Extraterrestrische Physik, Postfach 1312, 85741, Garching bei Mu¨nchen, Germany e-mail: 0 [email protected] 2 2 UniversityofMaryland,BaltimoreCounty,1000HilltopCircle,Baltimore,MD21250. 3 SLACNationalAcceleratorLaboratory,2575SandHillRoad,MenloPark,CA94025,USA n 4 KIPAC,2575SandHillRoad,MenloPark,CA94025,USA a 5 KavliInstituteforAstrophysicsandSpaceResearch,MIT,Cambridge,MA02139,USA J 6 CaliforniaInstituteofTechnology,105-24Robinson,1200EastCaliforniaBoulevard,Pasadena,CA91125. 6 7 Dept.ofPhysicsandAstronomy,LeicesterUniversity,LeicesterLE17RH,U.K. 2 8 Harvard-SmithsonianCenterforAstrophysics,60GardenSt,Cambridge,MA02138 9 DepartmentofAstronomyandAstrophysics,PontificiaUniversidadCatolicadeChile.Casilla306,Santiago22,Santiago,Chile. ] O C ABSTRACT . h Context.Tidaldisruptioneventsarepossiblesourcesoftemporarynuclearactivityingalacticnucleiandcanbeconsideredasgood p indicatorsoftheexistenceofsupermassiveblackholesinthecentersofgalaxies. - Aims.AnewX-raysourcehasbeendetectedserendipitouslywithROSATinaPSPCpointedobservationofthegalaxyclusterA3571. o Giventhestrongfluxdecayoftheobjectinsubsequentdetections,thetidaldisruptionscenarioisinvestigatedasapossibleexplanation r t oftheevent. s Methods.WefollowedtheevolutionoftheX-raytransientwithROSAT,XMM-NewtonandChandraforatotalperiodof∼13years. a Wealsoobtained7-bandoptical/NIRphotometrywithGRONDattheESO/MPI2.2mtelescope. [ Results.WereportaverylargedecayoftheX-rayfluxoftheROSATsourceidentifiedwiththegalaxyLEDA095953,amemberof 2 theclusterAbell3571.Wemeasuredamaximum0.3–2.4keVluminosityLog(LX)=42.8ergs−1.Thehighstateofthesourcelasted v atleast150ks;afterwardsLXdeclinedas∼t−2.Thespectrumofthebrightestepochisconsistentwithablackbodywithtemperature 7 kT ∼0.12keV. 5 Conclusions. Thetotalenergyreleasedbythiseventin10yrwasestimatedtobe∆E>2×1050erg.Weinterpretthiseventasatidal 3 disruptionofasolartypestarbythecentralsupermassiveblackhole(i.e.∼107M⊙)ofthegalaxy. 3 Keywords. X-rays–Galaxies,X-rays:bursts,Galaxies:clusters:individual:A3571,Galaxies:Active,Galaxies:Nuclei– . 1 0 9 1. Introduction tidalradiuswouldtypicallyfallwithintheSchwarzschildradius 0 andthestarisswallowedasawhole.InthecaseofrotatingBHs, : There is growing evidence that most galaxies with a v the tidal disruption also can occur for higher BH masses if the bulge host a supermassive black hole (BH) in their cen- i star approaches from a favorable direction (Beloborodovetal. X ter (Kormendy&Richstone 1995; Magorrianetal. 1998). The 1992). However atmospherestripping of giant stars is possible r most recent theoretical and observational developments in the alsoforlargerBHmasses.Whenassumingablackbodyradiat- a field of galaxyevolutionsuggestthat duringtheir life, galaxies ingattheEddingtonluminosity,thetemperatureoftheemitting experienceoneormorephaseswheretheBHisactiveandpow- regioncalculatedatthe last stable orbit(3R , whereR is the eredbyaccretion.Oneoftheeffectsofthisphenomenonarethe Schwarzschildradius)is∼30-300eVfora10S6−7M blacSkhole. ⊙ socalledactivegalacticnucleiorquasarswhereaccretionoccurs After the disruption, the mass accretion rate, and therefore the ontimescaleslongerthan106−7yr.Itiswidelyacceptedthatdur- bolometricluminosity,maydecayapproximatelyast−5/3 (Rees ingthisperiodtheBHisfedbyanaccretiondisk.Ontheother 1990; Evans&Kochanek 1989), even though deviations from hand,duringthenon-activephaseofthegalaxylifetheBHscan this law are expected (see e.g. Lodatoetal. 2008). 7 candidate be briefly powered by stars tidally disrupted during close en- tidal disruption events have been detected in the X-ray by the counterswiththeBHs(e.g.Frank&Rees1976).Recentstudies ROSAT all-sky survey and in the XMM-Newton slew survey (seee.g.Wang&Merritt2004;Esquejetal.2008;Donleyetal. (seee.gKomossa&Bade1999;Greineretal.2000;Grupeetal. 2002)pointedoutthattheseeventshappenwitharateof∼10−4- 1999;Komossaetal.2004; Esquejetal. 2007).Candidatetidal 10−5 galaxy−1 year−1. These phenomena can generate strong disruption have been observed also in the Optical/UV (e.g. flaresin the UV-X-raybandas the resultof the onsetof an ac- Renzinietal.1995;Gezarietal.2008).TheX-rayeventsshow cretiondisk.TheX-ray/UVluminosityofthegalaxyistherefore strong X-ray variability (i.e. more than a factor of 100), 0.3– expectedtogrowtovaluesclosetothoseofAGNs(i.e.>1042erg 2.4keVluminositiestypicalofAGNandquasars(i.e.>1042erg s−1).IfthestarisSun-like,tidaldisruptioncanonlyoccurifthe s−1) declining as ∼t−5/3, ultrasoft(i.e. ∼0.1 keV) X-ray spectra massoftheBHsdoesnotexceed∼108M since,inthiscasethe ⊙ andtheabsenceofSeyfertactivityinground-basedopticalspec- tra(seeKomossa2002forareview).Inthisletterwereportthe Sendoffprintrequeststo:N.Cappelluti 2 N.Cappellutietal.:AcandidatetidaldisruptioneventintheGalaxyclusterAbell3571 5’ 30" Fig.1. LeftPanel:TheROSAT-PSPC0.3-2.4keVcolourcoded image of the field of A3571. The X-ray transient is indicated Fig.2. Left Panel: The ROSAT-PSPC unfolded spectrum of withthearrow.Right Panel:GROND imageintheK -bandof S the source TDXFJ134730.3-325451(Crosses) and the best fit the the region of the X-ray transient, the red circle represents blackbodymodelContinuousLine. theROSAT-PSPCerror-boxandtheredellipsoidistheChandra confidenceregion. withasplinemodel.Withthebackgroundmapdescribedabove serendipitousdetection of a very large decay in the X-ray flux andusingtheexposuremapwerunamaximunlikelihood,PSF- ofthegalaxyLEDA095953atz=0.0366(Quintana&deSouza fitting, source detection.The bright, eye-detected,source men- 1993) in the field of view of the galaxycluster Abell 3571 us- tioned in the previous section has been detected with a signif- ingROSAT,XMM-NewtonandChandraarchivalX-raydataand icance >10σ with coordinates of the centroid 13h47m29.8s - ′ ′′ GRONDoptical/NIRfollow-up. 32◦55 00 (J2000)withanerrorboxof∼25′′.Consideringthat the source is embedded in the X-ray emission of the galaxy cluster Abell 3571, the photometric properties have been eval- 2. Observationsanddataprocessing uatedusingaperturephotometryinsteadofusingtheestimateof thedetectionsoftware.We extractedthe sourcecountsin a cir- During the soft X-ray follow-up of the Swift-BAT, hard X- cular region with a radius correspondingto the 90% Encircled ray selected sample of galaxy clusters (Ajelloetal. 2008), we EnergyRadius(EER)aroundthecentroidestimatedbythemax- serendipitously discovered a bright source in a ROSAT-PSPC imumlikelyhoodfit.Backgroundcountshavebeenestimatedin imageofAbell3571(seeFig.1)thatapparentlydisappearedin aringaroundthesourcewithaninnerradius1.5timesthe90% subsequentobservationswithXMM-NewtonandChandra.The EERandouterradius2timesthe90%ERR.Asaresultweob- fieldofAbell3571hasbeenobservedinX-raysinpointedmode tained 305±27net source countscorrespondingto a count-rate with the ROSAT-PSPC/HRI detectors, XMM-Newton EPIC, of(5.2±0.4)×10−2s−1.PSFfittingestimatedaflux∼30%lower Chandra ACIS-S and during the ROSAT all-sky survey. The asaresultofalikelyoverestimateofthebackground.Inthesame ROSATdatawereprocessedusingthestandardMIDAS-EXSAS regionswealsoextractedsourceandbackgroundlightcurvesand software version 03SEPpl1,while XMM-Newtonand Chandra spectra.Alltheepochsbutonehavefluxconsistentwithamean data were processed using XMM-SAS version 7.0 and CIAO count-rateof∼5.2×10−2s−1.Notethattheobservationwassplit 4.0. We also performed an optical/NIR follow-up of the field intwopartsseparatedby∼150ks.Thisindicatesthatnostrong in the g’, r, i’, z’ bandsand J, H, K bands, by using GROND S variabilityispresentonsuchatimescale.Howeverinonetime (Greineretal.2008),mountedatthe2.2mESO/MPItelescope binthecount-ratereached(8.3±1.0)×10−2s−1.Thisfluxis∼3σ atLaSillaobservatory(Chile).InTable1welistthelogofthe abovethemeanandcouldindicateasmallflareduringtheevent. observationsofthefieldofAbell3571relevantforthiswork. The unfolded spectrum of the source is shown in Fig. 2 . We performed a χ2 fit to the spectrum using a black-body 2.1.Dataanalysis model plus cold absorption. We kept as free parameters the temperature kT, the columns density N and the normaliza- ROSAT-PSPC: Using the ROSAT-PSCP 0.3-2.4 keV energy H tion k. As a result we obtained a temperature kT=120±16 eV bandimageandexposuremapavailablein thearchive,wecre- and N =3.93+1.43×1020 cm−2 and χ2/ν=1.05/4.00. This corre- ated a background map, by first running a sliding cell source H −1.24 spondstoa0.3-2.4keVfluxof7.65±0.92×10−13 ergcm−2 s−1. detection with a signal-to-noise ratio threshold of 4, excising Assumingnospectralchangesduringtheobservation,themaxi- thedetectedsourcesfromtheimageandfittingtheresidualdata mum0.3-2.4keVfluxemittedbythesourceis1.16±0.13×10−12 ergcm−2 s−1. Notethatthe columndensityisconsistentwithin Instrument Date Band Exposure 1σ with the Galactic value predicted by Dickey&Lockman dd-mm-yyyy s (1990)(i.e.N =3.7×1020cm−2)suggestinganextragalacticori- H ROSAT-PSPC 12-08-1992 0.3–2.4keV 5800 gin of the source. We also tested other spectral models, like a ROSAT-HRI 05-08-1994 0.3–2.4keV 19200 power-lawplusabsorption,withoutimprovingthequalityofthe XMM-Newton-EPIC-PN 29-07-2002 0.3–2.4keV 24200 fitbyhavingΓ ∼6.5,andN ∼3×1021cm−2 asbestfitparame- Chandra-ACIS-S 31-07-2003 0.3–2.4keV 33900 H ESO/MPI-GROND 12-08-2008 g,r’,i’,z’ 458 tersandχ2/ν=5.30/4.00. ESO/MPI-GROND 12-08-2008 J,H,K 480 We also performedaperturephotometryonROSATallsky sur- Table 1. The log of the X-ray and optical/NIR observationof veyimagesatthesourcelocationandwedetermineda1σupper- thefieldofAbell3571usedforthiswork. limit(seenextSection)onthecount-rateof0.44s−1correspond- ingtoafluxof∼2.2×10−12ergcm−2s−1inthe0.3–2.4keVband N.Cappellutietal.:AcandidatetidaldisruptioneventintheGalaxyclusterAbell3571 3 assumingasaspectralmodelablackbodywithkT=0.12keV. ROSAT-HRI: The source is visually observable also in the HRI observationthoughmuch fainter than in the PSPC image. Considering the highest spatial resolution of the HRI detector weperformedasourcedetectiontoconstrainthesourceposition better than in the PSPC pointing.Unfortunatelybecause of the faintness of the source and the difficulty in modeling the high backgrounddue to the extended X-ray emission of the cluster, the source has been detected with a significance of ∼3-4σ, but asanextendedsource.Wethereforeperformedaperturephotom- etry around the ROSAT-PSPC centroid. By adopting the same procedure used for the PSPC data, we measured 38±9 source counts correspondingto a count rate of (1.97±0.45)×10−3 s−1. ByassumingnospectralchangebetweentheROSAT-PSPCand HRI observation(i.e. 2years) we estimated a 0.3–2.4keV flux of1.53±0.4×10−13ergcm−2s−1. Fig.3. TheSEDofthegalaxyLEDA095953intheg’,r’,i’,z’, XMM-Newton: A visual inspection of the EPIC-PN image H,JandK bandsobtainedwithGRONDoverthebestfitSED does not show any relevant signature of a X-ray source in the S modelofaS0galaxy(solidline). ROSATerrorbox.We runa maximumlikelihoodsourcedetec- tion in that region but no sources were detected with signifi- cance >3σ. We therefore estimated the 3σ upper-limit of the source count at the source location by using the prescriptions of Narsky (2000). Given M counts actually observed in a re- (Quintana&deSouza 1993). We determined the photomet- gionof 30′′ (Cappellutietal. 2009), andB backgroundcounts, ric properties of the source and obtained the following AB the 1σ upper limit is defined as the number of counts X that magnitudes:17.05±0.10,16.26±0.10,15.80±0.10,15.54±0.10, gives the probability to observe M (or fewer) counts equal 15.04±0.02, 14.75±0.07 and 14.87±0.04 in the g’ ,r’, i’, z’, to the formal 68.3% Gaussian probability: P(≤ M,X + B) = J, H, K bands, respectively. These magnitudes are calculated P (68.3%).AssumingPoissonian statistics this equationbe- S Gauss by using SA105-815and 2MASS stars as a reference and cor- comes: PGauss = e−(X+B)PiM=0 (X+i!B)i. By iteratively solving the rected for the expected foreground extinction of EB−V=0.054 previousequationin thecase of P =0.997,weobtainedthe Gauss (Schlegeletal. 1998). We examined the global spectral energy 3σupperlimitX=167counts.Thiscorrespondstoacount-rate distribution(SED)inFig.3ofthisobjectusingatoolbasedon of8.8×10−3s−1 andtoafluxof1.45×10−14ergcm−2s−11. a chi-squaredminimizationmethodthatallowsustofitthe ob- Chandra:Usingthetaskwavdetect,includedintheCIAOsoft- servedfluxeswithacombinationofAGNandgalaxyemission, ware, we performed a source detection on the ACIS-S4 Chip also allowingfor the possibility of extinctionof the AGN flux. where the source is observable. The analysis was performed From this analysis we found that the model fit that better de- in the 0.3–2.4 keV energy band. We ran wavdetect using an scribes the SED of this object is a normal S0 galaxy with no exposure map estimated at 1 keV with a threshold of 10−6, significantcontributionduetothepresenceofanAGN. correspondingto a maximum of 1 spurious detection expected per ACIS-chip. As a result, in the ROSAT PSPC error box we detected one source with 9.5±3.74 counts corresponding to a count-rate of 2.89×10−4 s−1. The new source coordinates are 3. Longtermlightcurve α=13h47m30.33s, δ=-32◦54′50.63′′. At the source position the The X-ray flux of TDXFJ134730.3-325451has been followed estimated σ positional error is ∼ 2.5′′. We also ran a source for13yearsanditslightcurveispresentedinFig.4.Remarkably, detection in the 2–7 keV band without detecting a source. We itshowsanamplitudeofthevariabilityofafactor>650.Weesti- also estimated the flux of the source by folding into XSPEC matedthemaximum0.3-2.4keVluminosity2asLog(L )=42.83 the Chandra response matrices and assuming as the spectral X erg s−1. By excluding the upper limits, we fitted the long term model the ROSAT-PSPC best fit and obtained a 0.3–2.4 keV lightcurve with a power-law model in the form: L (t) = k ∗ fluxof3.43±1.35×10−15ergcm−2s−1.Notethatthesourcecen- X troid lies about 5′′ from the chip edge, therefore the estimate (t−1tydrisr)−α, where k is the normalization, tdisr marks the time of ofthe backgroundcouldbeaffectedbyfeaturesnearthisedge. the star’s disruption and α is the slope of the decay. As a re- sultweobtainedk∼7×1042ergs−1,t ∼1991.6andα=2.2±0.5. According to its coordinate, the X-ray source has been named disr Because of the low numberof data points, the uncertaintieson TDXFJ134730.3-325451. thefitparametersarelarge.Thevalueoftheslopeαistherefore GROND: In Fig. 1 we show the GROND K-band image marginallyconsistentwiththepredictionsofvariantsoftidaldis- of the area covered by the ROSAT-PSPC error box. As one ruption models (i.e. α ∼5/3, see e.g. Evans&Kochanek 1989; can notice, the Chandra ellipsoid is placed right on the disk of a bright galaxy with an offset of ∼3 ′′ from the op- Rees 1990). Note that according to the lightcurve, the source at the maximum could have been at 1-2 orders of magnitude tical centroid of the galaxy. Note that the GROND images have an astrometric uncertainty of 0.5′′. We searched the brighter with respect to the time of the ROSAT-PSPC observa- tion. If the source continues its decay, it would be detectable NED catalog and the galaxy was previously known as LEDA 095953, a member of the cluster Abell 3571 at z=0.0366 with Chandra for all the year 2009 and 2010, with moderately deepexposures. 1 The flux was derived from the count-rate with the PIMMS software by assuming as the spectrum the ROSAT-PSPC best fit. 2 Theluminosity hasbeen computed assuming aΛ-dominated uni- http://heasarc.nasa.gov/Tools/w3pimms.html versewithΩ =0.7,Ω =0.3andH =72kms−1Mpc−1 Λ m 0 4 N.Cappellutietal.:AcandidatetidaldisruptioneventintheGalaxyclusterAbell3571 leaseoftheorderof2×1050erg.Ifǫ∼0.1isthetypicalefficiency ofmasstoenergyconversionforaccretionontoa BH, thenthe total mass accreted during a 10 year period is ∆M∼ ∆E /ǫc2. X Thisyieldsatotalmassdepositionof1×10−3M .Thisvalueis ⊙ actuallyalowerlimit,sincewestartedtheintegrationonlyatthe start of the observations(note that the fitting function diverges at the time of the disruption). Moreover it is possible that the totalemissionisdominatedbyanunobservedEUVcomponent (extrapolation of the single black body fit itself predicts a 7% higher luminosity), and the true accretion rate is much higher. Accordingto the results of Lietal. (2002), for a black hole of 107 M , the flare occurs∼0.07yr after the disruption.A 10 yr ⊙ integration of the lightcurve model from this epoch yields an estimate of the accreted mass of ∼0.03 M . A numerical sim- ⊙ ulation by Ayaletal. (2000) showed that for a 1 M star dis- ⊙ ruptedbya106M BH,onlyafractionoftheorderof∼10%of ⊙ Fig.4. The X-ray lightcurve of TDXFJ134730.3-325451.The theinitialmassisactuallyaccreted.Ourlimitsontheaccreated solid curve represents the result of the power-law fit while the mass are therefore consistent with the scenario of a disruption verticallinemarkstheexpectedbeginoftheflare. of a 1 M star by the BH. However a partial stellar disruption ⊙ oranexplosivedisruption(Brassart&Luminet2008)cannotbe ruled out, although disruption and accretion of a brown dwarf or of a planets is very unlikely. We note that since the tidal 4. Discussion disruption rates are likely of the order of 10−4-10−5 galaxy−1 We serendipitouslydetected a factor >650 decay in the flux of yr−1, galaxy clusters are ideal laboratories to explore this phe- anX-raysourceintherichgalaxyclusterA3571.Theflaremost nomenon:byconservativelyassuming100membergalaxiesper likelyoriginatedinthegalaxyLEDA095953atz=0.0366.The cluster (Reiprich&Bo¨hringer 2002), one should expect1 tidal most luminous epoch in the lightcurve reached LogL =42.83 disruptioneventperclusterevery10-100yrs.TheeROSITAall- X (0.3–2.4)andlastedformorethan150ks,althoughhigherlumi- skysurvey(Predehletal.2006)willdetectseveralthousandsof nosities at the time of the burst cannot be excluded. The spec- galaxyclustersandthereforealargenumberoftheseevents. trumoftheflareisconsistentwithablack-bodyoftemperature Acknowledgements. TheXMM-Newton,ChandraandROSATarchivalsystems kT∼0.12 keV with no significant hard components. The opti- andthefullGRONDteamareacknowledged.PRissupportedbythePappalardo cal/NIR SED of the host galaxy does not show any signature fellowship at MIT. HQ thanks partial support from the FONDAP Centro de ofAGN activity.Inadditiontothis, the decayof thelongterm Astrofisica. light curve is marginally consistent with t−5/3, and thanks to the stringent RASS upper limit, we likely detected the source References close to the burst. The X-ray luminosity and the variability of thesourcearethereforemuchhigherthanthoseofULX,which Ajello,M.,etal.2008,arXiv:0809.0006 haveL <1039−40ergs−1 andfluxvariationstypicallyofuptoa Ayal,S.,Livio,M.,&Piran,T.2000,ApJ,545,772 X factorof2-3(seee.g.Colbert&Ptak2002).Supernovaeexplo- Beloborodov,A.M.,etal.1992,MNRAS,259,209 Brassart,M.,&Luminet,J.-P.2008,A&A,481,259 sionscanalsoproduceX-rayflares,buttheseareeitherofvery Cappelluti,N.,etal.2009,arXiv:0901.2347 shortduration(hours),or else if longerlasting theyhave much Colbert,E.J.M.,&Ptak,A.F.2002,ApJS,143,25 harderX-rayspectraandtypicallyhaveL <1041ergs−1(seee.g. Dickey,J.M.,&Lockman,F.J.1990,ARA&A,28,215 X Immler&Lewin2003,forareview).Additionallynovaeandsu- Donley,J.L.,Brandt,W.N.,Eracleous,M.,&Boller,T.2002,AJ,124,1308 Esquej,P.,etal. 2007,A&A,462,L49 persoftsourcescanproduceburstsbutwith muchshorterdura- Esquej,P.,etal.2008,A&A,489,543 tion (Tanaka&Shibazaki 1996). According to the event prop- Evans,C.R.,&Kochanek,C.S.1989,ApJ,346,L13 erties, we can safely excludethese kindsofstellar phenomena. Frank,J.,&Rees,M.J.1976,MNRAS,176,633 However, such a strong variability could be observed in GRB Gezari,S.,etal.2008,ApJ,676,944 afterglows or Hypernovaebut with a much shorter duration of Greiner,J.,Schwarz,R.,Zharikov,S.,&Orio,M.2000,A&A,362,L25 Greiner,J.,etal.2008,PASP,120,405 theoutburst.Thecharacteristicsoftheflarearethereforeconsis- Grupe,D.,Thomas,H.-C.,&Leighly,K.M.1999,A&A,350,L31 tentwithpreviouscasesofcandidatetidaldisruptionevents(see Immler,S.,&Lewin,W.H.G.2003,SNeandGRBs,598,91 e.g.Komossaetal. 2004;Esquejetal. 2008)andwith theoreti- Kormendy,J.,&Richstone,D.1995,ARA&A,33,581 calpredictions(Rees1990;Lodatoetal.2008).Themassofthe Komossa,S.,&Bade,N.1999,A&A,343,775 BHshouldbenotmuchhigherthan∼108 M todisruptaSun- Komossa,S.2002,ReviewsinModernAstronomy,15,27 ⊙ Komossa,S.,etal.2004,ApJ,603,L17 likestaroutsidetheSchwarzschildradiusandtoallowtheonset Li,L.-X.,Narayan,R.,&Menou,K.2002,ApJ,576,753 of accretion. We then estimated the black hole mass from the Lodato,G.,King,A.R.,&Pringle,J.E.2008,arXiv:0810.1288 L -M relation determinedby Marconi&Hunt (2003) in the Magorrian,J.,etal.1998,AJ,115,2285 K BH localUniverse.IfL istheK-bandluminosityofthebulgeofthe Marconi,A.,&Hunt,L.K.2003,ApJ,589,L21 K Narsky,I.2000,Nucl.Inst.andMethodsinPhysicsResearchA,450,444 galaxyinsolarunitsthenlog(M /M )∼8.08+1.21(L /L - 10.9). Using an aperture radiusBHof ∼⊙4′′ (i.e. ∼2.5 kpcK,bulaget th⊙e PQrueidnetahnl,aP,.H,e.,t&al.d2e0S0o6u,zPar,oRc..S1P9I9E3,,6A2&66A,S,101,475 galaxydistance)weestimatedalog(L )∼10L/L whichcor- Rees,M.J.1990,Science,247,817 respondsto log(M /M )∼7. The totKalbulegenergyrele⊙ased during Reiprich,T.H.,Bo¨hringer,H.2002,ApJ,567,716 BH ⊙ ∞ Renzini,A.,etal.1995,Nature,378,39 the event can be estimated with ∆EX=Rt LX(t)dt, where LX(t) Tanaka,Y.,&Shibazaki,N.1996,ARA&A,34,607 isthebestfitto thelightcurve.Byintegratingoveraperiodof Schlegel,D.J.,Finkbeiner,D.P.,&Davis,M.1998,ApJ,500,525 ∼10 yr after our first data point we obtained a total energy re- Wang,J.,&Merritt,D.2004,ApJ,600,149