Mon.Not.R.Astron.Soc.000,000–000(0000) Printed15July2009 (MNLATEXstylefilev2.2) Optical–to–X-ray emission in low-absorption AGN: Results from the Swift-BAT 9 month catalogue 1 2 3 1 R. V. Vasudevan R. F. Mushotzky , L. M. Winter and A.C. Fabian 1InstituteofAstronomy,MadingleyRoad,CambridgeCB30HA 9 2LaboratoryforHighEnergyAstrophysics,NASA/GSFC,Greenbelt,MD20771,USA 0 3CenterforAstrophysicsandSpaceAstronomy,UniversityofColoradoatBoulder,440UCB,Boulder,CO80309-0440,USA 0 2 l 15July2009 u J 4 ABSTRACT 1 Wepresentsimultaneousoptical–to–X-rayspectralenergydistributions(SEDs)fromSwift’s X-ray and UV–optical telescopes (XRT and UVOT) for a well-selected sample of 26 low- ] E redshift(z < 0.1)activegalacticnuclei(AGN)fromtheSwift/BurstAlertTelescope(BAT) H 9-monthcatalogue,thelargestwell-studied,hardX-rayselectedsurveyoflocalAGNtodate. . OursubsampleconsistsofAGNwithlowintrinsicX-rayabsorption(NH < 1022cm−2)and h minimalspectralcomplexity,tomoreaccuratelyrecovertheintrinsicaccretionluminosityin p thesesources.WeperformacorrectionforhostgalaxycontaminationinallavailableUVOT - filter images to recover the intrinsic AGN emission, and estimate intrinsic dust extinction o r from the resultant nuclear SEDs. Black hole mass estimates are determined from the host- t galaxy2MASSK-bandbulgeluminosity.AccretionratesdeterminedfromourSEDsareon s a averagelow(EddingtonratiosλEdd . 0.1)andhardX-raybolometriccorrectionsclusterat [ ∼10–20,incontrastwiththehighervaluesseenforquasars.AnaverageSEDforthe22low 1 accretionrate(λEdd < 0.1)objectsispresented,withandwithoutcorrectionforextinction. v SignificantdustreddeningisfoundinsomeobjectsdespitetheselectionoflowNH objects, 2 emphasisingthecomplexrelationshipbetweenthesetwotypesofabsorption.Wedonotfind 7 acorrelationofoptical–to–X-rayspectralindexwithEddingtonratio,regardlessoftheoptical 2 referencewavelengthchosenfordefiningthespectralindex.Ananti-correlationofbolomet- 2 ric correctionwith black holemass may reinforce‘cosmic downsizing’scenarios, since the . higher bolometric correctionsat low mass would boost accretion rates in local, lower mass 7 0 black holes. We also performa basic analysis of the UVOT-derivedhost galaxycoloursfor 9 our sample and find hosts cluster near the ‘green valley’ of the colour-magnitudediagram, 0 but better quality images are needed for a more definitive analysis. The low accretion rates : andbolometriccorrectionsfoundforthisrepresentativelow-redshiftsampleareofparticular v importanceforstudiesofAGNaccretionhistory. i X Keywords: blackholephysics–galaxies:active–galaxies:Seyfert r a 1 INTRODUCTION resultstoidentifytrendsbetweendifferentSEDparameters.How- ever,recoveringthetrueaccretionluminositycanbecomplex,and Active Galactic Nuclei (AGN) are known to emit radiation over previous studies have found a wide spread in the conversion fac- thewholerangeofavailableenergiesobservableusingcurrentde- tors(bolometriccorrections)betweentheX-rayluminosityandthe tectors. Characterising their spectral energy distribution (SED) is totalaccretionluminosity.Thisis,inalargepart,duetointrinsic therefore important for understanding the different physical pro- variationbetweenobjectsinthewholeAGNpopulation,butisalso cessesatwork.Theacceptedparadigmbehindtheradiationoutput complicated by numerous systematics and biases in the samples ofAGNisaccretionontoasupermassiveblackhole.Theemission used. directlyduetoaccretionemergesprimarilyintheoptical,UVand The pioneering study of Elvisetal. (1994) (hereafter E94) X-rayregimesbyacombinationofthermalemissionfromanac- presentedradio–to–X-raySEDsfor47quasarsandhasprovideda cretion disc and inverse-Compton scattering of UV disc photons useful template for determining bolometric corrections and iden- byacoronaabovethedisc.AdustytoruspartiallyabsorbstheUV tifying trends in the SED shapes across a range of AGN prop- photons from the accretion disc, re-emitting them more isotropi- erties. Recent studies have confirmed the usefulness of the av- callyintheinfrared.Inthisstudy,weaimtorecoverthetrueaccre- erage bolometric corrections and SED parameters from E94, but tionluminosityinarepresentativesampleofAGN,andusethese the sample of quasars used was predominantly X-ray bright and (cid:13)c 0000RAS 2 R.V. Vasudevan,R. F.Mushotzky,L. M. Winter& A.C. Fabian their bolometric luminosities included the re-processed infrared Filter Centralwavelength(A˚) emission. Richardsetal. (2006) present infrared–to–X-ray SEDs for 259 quasars selected by a combination of optical and mid- V 5468 infraredcolourselectioncriteria,andhighlightthespreadinSED B 4392 parameters withinthe sample. Thislarge spread in SED parame- U 3465 UVW1 2600 terswasknownfromE94,promptingrefinementssuchasthetem- UVM2 2246 plateSEDofMarconietal.(2004)whoemploytheobservedcor- UVW2 1928 relationbetweenX-ray–to–optical spectralindexandX-raylumi- nositytoconstructaluminosity-dependent SED,andimportantly, Table1.CentralfrequenciesoftheUVOTfilters. they exclude the reprocessed infrared emission to avoid double- counting of part of the accretion energy budget. A similar tem- plate is presented by Hopkinsetal. (2007) in their study on the shapes. Most importantly, they present values of absorbing col- bolometric quasar luminosity function. These SED templates al- umn density N for each source, determined from fits to the X- H lowedthediversityofSEDshapestobetakenintoaccount when raydatafromtheliteratureandnewfitsbytheauthorsthemselves, calculatingthesupermassiveblackholemassdensityfromtheX- andidentifythosesourcesinwhichsignificantspectralcomplexity ray background and AGN luminosity functions. More recently, ispresent. Inthiswork, wepreferentiallyselect thosewithlower Vasudevan&Fabian (2007) (VF07 hereafter) presented optical– absorption and minimal spectral complexity, facilitating a more to–X-ray SEDsfor a sample of AGN observed by the Far Ultra- straightforwardcalculationoftheaccretionluminosity.Weassume violet Spectroscopic Explorer (FUSE), which yielded interesting thatforsuchsources,theaccretionluminosityexcludestherepro- trendsbetweenSEDshapeandEddingtonratio,confirmedlaterby cessedIRemissionandisprincipallyseenintheoptical–to–X-ray Vasudevan&Fabian (2009) (VF09 hereafter) using simultaneous regime(asinVF07andVF09).Additionally,AGNwithmoreX- data and reverberation mapping mass estimates to improve accu- rayabsorptionareexpectedtodisplayhigherlevelsofoptical–UV racyof thebolometricluminositiesandaccretion rates.However, reddening,thecalculationofwhichdependsonthepreciseformof the sample of VF07 is by necessity UV-bright, as it is selected theextinctioncurveandisdifficulttoaccountforwhencalculating byFUSEandthesampleofVF09isrestrictedtothosewhichare thetotalluminosity.Theseconsiderationsmotivateourselectionof bright enough in the optical/UV for reverberation mapping to be objectswithlowN . H carriedout. VariabilityinAGNcanproducelargechangesinluminosity, These considerations emphasise thepressing need for arep- overtimescalesfromhourstoyears.Inordertocatchanaccurate resentativesampleofAGNfromwhichconclusionscanbedrawn snapshot ofthetotalenergybudget inanAGNatagiventime,it aboutaccretionpropertiesofthewiderAGNpopulation.Onema- isthereforehighlydesirabletousesimultaneousdata.Thepoten- jorfactorinfluencingtheselectionofAGNforsurveysisthenature tialinaccuraciesinSEDparametervaluesfromnon-simultaneous oftheirabsorption.IthasbeenknownforsometimethatX-rayand SED data are discussed in detail in VF07 and VF09, with the optical surveys detect different segments of the underlying AGN latterfollowing theapproach of Brocksoppetal. (2006) byusing population (Mushotzky 2004), with X-ray surveys able to probe contemporaneousoptical–to–X-raydatafromXMM-Newton’sPN tohigher absorbingcolumndensity.Thestandardphysical mech- and Optical Monitor instruments. In this work, we make use of anisms for the generation of X-ray emission in AGN produce a Swift’s own co-aligned UV–optical telescope (UVOT) and X-ray power-law,butoftenAGNexhibit X-rayspectradeviatingsignif- telescope (XRT) to determine simultaneous SEDs. Many of our icantly from this, in part due to absorption. The geometry of the targetsubsamplefromtheSwift/BATcataloguehavesimultaneous absorption inX-raysisthesubject of muchdebate, andthecom- UVOT and XRT observations available in the archives; typically plexspectrainsomesourcescanbeaccountedforwitheither‘par- theUVOTdataspansixfiltersbetween5468A˚ and1928A˚,provid- tialcovering’scenarios(Gierlin´ski&Done2004)orstrongreflec- ingawindowontoasubstantialfractionofthediscemission.The tionfromtheaccretiondiscduetolightbending(Miniutti&Fabian centralfrequenciesoftheUVOTfiltersaregiveninTable1. 2004),andwarmabsorberscanalsocomplicateourviewofthein- Thelargewavelengthcoverage,fieldofviewandspatialreso- trinsic emission (Reynoldsetal. 1997, Blustinetal. 2005). Many lutionoftheUVOTdataalsoallowacorrectionforthehostgalaxy ofthereverberationmapped AGNinVF09showed signsof such fluxto beestimated for each AGN.Weemploy GALFITfor this spectralcomplexity,andtheauthorsidentifiedthedifficultiesinre- purpose,aprogramforperforming2DfittingofPSFsandvarious gainingthetrueaccretionluminosityinthesecases. analytical profile forms to galaxy images (Pengetal. 2002). For The Burst Alert Telescope (BAT) on board the Swift satel- a detailed discussion of the requirements of GALFIT, the reader lite proves invaluable for addressing many of these considera- isdirected to peruse the accompanying documentation1. We also tions. Canonical levels of absorption (column densities NH < presentbasicresultsonthehostgalaxiesfromtheGALFITfitting 1024cm−2) imprint signatureson the0.1–10 keV X-rayband, so process.TheUVOTdatausedinthisstudyalsoclarifyhowthede- theveryhardX-raysensitivityoftheinstrument(14–195keV)pro- greeofhostgalaxycontaminationvarieswithwavelength(UVOT videsthecapabilitytoobserveAGNinabandpassrelativelyunaf- filter). fectedbysuchabsorption.The9-monthcatalogueofBAT-detected Understanding variationsintheAGNSEDisimportant,and AGN(hereaftertheSwift/BATcatalogue,Tuelleretal.2008)there- one of the main drivers of these variations appears to be the ac- foreprovidesanunprecedentedlevelofcompleteness(withrespect cretion rate (parameterised as the Eddington ratio, L /L bol Edd to absorption) when surveying the AGN population as it is unbi- for an object with bolometric luminosity L and Eddington lu- bol ased to all but the most heavily obscured sources. The study of minosity LEdd = 1.38 × 1038[MBH/M⊙]ergs−1 for a black Winteretal.(2009)presentsacomprehensive overview oftheX- holeofmassM ).TherecentPrincipalComponentAnalysisof BH ray spectral properties of the 153 sources in the Swift/BAT cata- logue and allows us to select an appropriate sample from which wecanbeconfidentofcalculatingaccurateluminositiesandSED 1 http://users.ociw.edu/peng/work/galfit/galfit.html (cid:13)c 0000RAS,MNRAS000,000–000 Optical–to–X-rayemissioninlow-absorptionAGN 3 Kuraszkiewiczetal.(2008)onasampleofred2MASSAGNSEDs isdiscussedatgreaterlengthin§3.4onblackholemassdetermi- identifytheEddingtonratioasthe‘eigenvector 1’fortheseAGN nation.Werequiredthecentroidsreportedforthe2MASSdatato (being responsible for most of the variation in the SED shapes). be within 2.0 arcsec of the NED position in the initial case. For BothVF09andVF07highlightsubstantialvariationsinthebolo- afewobjectswhichliejustoutsidethislimit,the2MASSimages metric output (in comparison to the X-ray output) across the ob- weredownloadedandthelocationsprovidedwerecheckedagainst served range of Eddington ratios, and numerous studies explore therefinedpositionsobtainedfromtheXRTandUVOTimages;if thesethemesfurther(Shemmeretal.2008,Kellyetal.2008).Ed- aconfident identificationof acounterpart tothe X-ray AGN was dington ratios require accurate estimates of the black hole mass made, the 2MASS data were used in further processing. At any in AGN, but since the most accurate estimates from reverbera- rate, objects with 2MASS centroids greater than 5.0 arcsec away tionmappingareonlyavailablefor35AGN,hereweemploythe fromtheirNEDpositionswererejectedaltogether.Threepotential M −L correlationtocalculateblackholemasses.Thecom- objectsareexcludedfromoursamplebytheirabsencefromthecat- BH bulge binationofrobustdeterminationsofthebolometricluminosityand alogues(IGR21277+5656,IGRJ07597-3842andLEDA138501). sensible estimates of the black hole mass provide great opportu- Lastly,wecheckforblazarsinthesample.TheSEDproper- nitiesfor insight into theaccretion process. Inthiswork, wedis- tiesofblazarsmakethemunsuitable forperformingabolometric cuss the analysis of the Swift data, the construction of SEDs,the luminositycalculationbasedonthe‘disk+power-law’modelem- derivationofSEDparameters,theidentificationofcorrelationsand ployedheresincetheirUVandX-rayluminosityisdominatedby trends,andfinallythehostgalaxyproperties. jetemission.Outofthe30objectswhichsatisfyalltheabovecri- teria,fourobjectsareexlcudedbasedontheirNEDclassifactionas blazars: Mrk 501, 2MASX J19595975+6508547, ESO362-G021 (PKS0521-36) and PKS 0548-322. The final selection of 26 ob- 2 SAMPLESELECTION jectsispresentedinTable2,alongwiththeirredshiftsanddetails We firstly apply a cut in intrinsic column density N(int), select- oftheSwiftobservationsused. H ing only those with log(N(int))< 22 from the list of objects in H the 9-month BAT catalogue as presented in Winteretal. (2009). Thisisthesuggestedcrossoverpointbetween‘absorbed’and‘un- 3 PROCESSINGOFMULTIBANDDATA absorbed’ classesofAGNintheirstudy,whichalsoclassifiesthe Wedownloadtheoptical–UVandX-raydataforourselectionfrom objects in the sample into two broad categories based on X-ray HEASARC. When multiple observations were available, those spectral properties: ‘simple’ (S) and ’complex’ (C). The S class with the maximum UVOT exposure time and maximum number haveX-rayspectrawhicharebestfitbyasimplepower lawwith of UVOT filters were preferentially selected. Pipeline-processed intrinsic and galactic absorption (sometimes with a soft excess), whereastheCclassdisplaymorecomplexX-rayspectrawhichare ‘level2’FITSfilesarereadilyavailablefromHEASARC.Thedata forthe26sources identifiedinsection2werethenprocessed ac- better fit by models such as double-power laws or partial cover- cordingtotheprocedureoutlinedinthefollowingsections. ing(thesemodelsareillustrativeinthattheyhighlightthespectral complexity;moredetailedfitsincludingreflectionmayalsobepos- sible).WealsoeliminateanyobjectsfromtheCclass,astheywill 3.1 Optical–UVphotometry:Correctingforthehostgalaxy presentdifficultieswhendeterminingthetrueaccretionluminosity (asdiscussedinVF09). We employ the custom-built software tools designed specifically Inorder tofacilitateprofilefitstooptical andUVimagesof forprocessingUVOTdatawherepossible.EachindividualUVOT thehostgalaxies,wealsoimposearedshiftcut,requiringz <0.1 filter data file in general contains a number of exposures which forourselection.Forz <0.1,theangularresolutionoftheUVOT weresummedusingthetoolUVOTIMSUM.TheUVOTSOURCEtool (∼ 1.7 arcsec) allows separation of the nucleus from the galaxy was then used to extract magnitudes from simple aperture pho- to a physical length scale of 3 kpc, thus allowing a reasonable tometry.Sourceandbackgroundregionswerecreatedforthispur- galaxy–AGN separation while still yielding 54 potential objects pose, with the position of the source region being obtained from forstudyfromthesample.WethenidentifythoseforwhichSwift theNASAExtragalacticDatabase(NED)inthefirstinstance,fol- XRT and UVOT data are available from the High Energy Astro- lowedbyfineadjustmentofthesourceregionpositionifnecessary. physics Science Archive Research Center (HEASARC2), yielding TherequiredsourceregionsizeforUVOTSOURCEis5arcsec,and 33 objects. For estimating black hole masses, we also obtain K- thebackgroundregioncanbeanysize(atthetimeofwriting):we bandmagnitudesfromtheTwo-micronAll-SkySurvey(2MASS) usedbackgroundregionswithradiifrom5arcsecto∼30arcsecde- catalogues.Thekeyrequirementforcalculatingblackholemasses pendingonthefrequencyofotherforegroundsourcesintheimage. isanestimateofthebulgeluminosity.TheK-bandisleastsubject The magnitudes from UVOTSOURCE provide a useful first-order tothe effectsof Galacticreddening and predominantly tracesthe estimate of the nuclear flux, and were saved for a later compari- older stars in the bulge over the stellar populations of the galaxy sonwiththemorecarefullydeterminednuclearfluxesusingpoint disc, motivating its selection over the J and H bands also avail- spreadfunction(PSF)fitting. able in the 2MASS catalogues. The 2MASS data were gathered TheimageswerethenpreparedforusewithinGALFIT.One withtwogroundbasedtelescopes:theWhippleObservatoryinAri- oftherequirementsfor accuratelydeterminingthenuclear fluxis zona,USAandtheCerroTololotelescopeatLaSerena,Chile.The tohavesuitablePSFsavailableforeachimage.Preliminarystud- limitationsintroducedbyground-basedobservations(mostimpor- ies of the UVOT PSF in the six filters show that the PSF shape tantly, the level of seeing for each observation) need to be taken may depend on a variety of factors including position on the de- intoaccountwhenattemptingtorecoverthebulgeluminosity.This tector, countrate of the source and the filter being used. Average PSFfull-width half-maxima range between 1.7–2.5 arcsec, when considering all six filters. As discussed by Kimetal. (2008a) in 2 http://heasarc.gsfc.nasa.gov/cgi-bin/W3Browse/swift.pl thecontext ofHubbleSpaceTelescope(HST)images, thechoice (cid:13)c 0000RAS,MNRAS000,000–000 4 R.V. Vasudevan,R. F.Mushotzky,L. M. Winter& A.C. Fabian AGN redshift SwiftobservationID TotalUVOTexposuretime(ks) XRTexposuretime(ks) UGC06728 0.006518 00035266001 2.1 6.33 MCG-06-30-15 0.007749 00035068003 3.19 8.72 NGC4593 0.009 00037587001 1.83 4.88 Mrk766 0.012929 00030846039 1.34 3.71 ESO548-G081 0.01448 00035250002 2.8 6.33 Mrk352 0.014864 00035243002 6.38 15.91 NGC7469 0.016317 00031245005 1.38 4.27 NGC5548 0.017175 00030022062 1.84 5.13 WKK1263 0.02443 00035268002 1.45 8.94 ESO490-G026 0.02485 00035256001 2.87 8.51 Mrk590 0.026385 00037590001 1.43 4.46 1RXSJ045205.00+493248 0.029 00035281002 0.65 1.99 SBS1301+540 0.0299 00035269001 2.3 7.85 Mrk279 0.030451 00037591001 1.75 5.22 MCG+04-22-042 0.032349 00035263001 3.0 9.11 Ark120 0.032713 00037593003 1.44 4.11 IRAS05589+2828 0.033 00035255001 2.68 5.91 3C120 0.03301 00036369001 2.1 6.37 Mrk509 0.034397 00035469003 2.5 6.8 Mrk841 0.036422 00035468002 3.33 8.4 Mrk1018 0.042436 00035166001 1.41 4.53 NGC985 0.043143 00036530005 3.16 8.65 3C390.3 0.0561 00037596001 2.25 6.46 IRAS09149-6206 0.0573 00035233002 1.87 5.04 SBS1136+594 0.0601 00035265001 3.22 9.17 2MASXJ21140128+8204483 0.084 00035624002 1.72 5.09 Table2.DetailsofUVOTobservationsusedfortheobjectsinoursample. of PSF can significantly affect PSF fitting results, since a poorly PSFswereusedinGALFITtomodelthecentralAGNinthegalaxy chosen PSFcan give profile parameters that deviate dramatically profiles. from their ‘true’ values (also see Simmons&Urry 2008 for an- When this approach was not possible due to lack of guide otherdiscussionofthefactorsaffectingAGN–host-galaxydecom- stardetectionswithinreasonablecount-ratelimits(orlackofpoint position).TomitigatetheseproblemsandminimizePSFmismatch, sourceswhichwerealsoidentifiedintheguidestarcatalogue),the weadopted thefollowingapproaches. Thefirstand preferred ap- filterimagewasviewedusingthe XIMAGEpackage anda‘King’ proachwastogenerateauniquePSFforeachfilterimagefroma profilemodelwasfittoanidentifiablepointsourcewhichlooked fewknownguidestarsintheimage,requiringthestarsselectedto similarinformandintensitytotheAGNnucleus.Theparameters be withinacertainrange of thecount rateof the source of inter- fromthefitwerethenusedasamodelforthePSFinGALFIT(via est (as determined from simple aperture photometry with UVOT- the‘Moffat’modeloption),inlieuofarealPSFimage.Inthevast SOURCE).Firstly,theUVOTutilityUVOTDETECTwasusedtopro- majorityofcases,arealPSFimagecouldbeextractedfromtheim- videalistofdetectablesourcesineachimage(UVOTSOURCEes- agesanda‘King/Moffat’modelPSFwasrarelyrequired.Inonly sentiallyusestheSEXTRACTORpackage forthispurpose).These onefilterimageforoneobject,aGaussianmodelwasfoundtofit were then cross-checked against published catalogues of known thenucleus betterthan other availablemodels, withtheGaussian guidestarpositions.The XBROWSEtoolinthe HEASOFTsuiteof widthfixedattheaverageFWHMoftheUVOTPSFinthatfilter utilitiesprovidesaccess tomany suchcatalogues, such astheUS asreportedintheCALDBdocumentation4. NavalObservatory(USNO)andHST guidestarcatalogues.Typi- ThePSForPSFmodelthusgeneratedwasthenfitalongwith cally,theUSNOcataloguewasusedtoobtainalistofsuchstars, other galaxy profile components in GALFIT.We performed tests requiring them to be within 13 arcmin of the AGN of interest to to ascertain the level of detail discernible within the UVOT im- ensuretheywerewithintheUVOTfieldof view.Theguide stars ages,startinginitiallywithafour-component modelconsistingof wereselectedsuchthatbetween3-20 starswereusedtogenerate a constant sky background, acentral nuclear point source, a Ser- eachPSFimage.ForthelowestenergyV-band,thistranslatedinto sicprofileforthegalaxybulgeandanouterexponentialdisk.Such a requirement that the stars were typically within 0.1–0.2 dex of testsoverwhelminglyindicatedthatfourcomponentsincorporated thetargetAGNcountrate;forthehighestenergyUVW2band,the toohighadegreeofdegeneracy,duetothelargeUVOTPSFsize. limitswere typically within1.5-2.0 dex of the target AGN count ThePSFoftheUVOTinstrumentistypicallythesamesizeasthe ratetoobtainasimilarnumberofguidestars.Finally,theregions galaxybulgewouldbeinhigherresolutionimagesofgalaxiesnear of the images containing these stars were then summed to form redshift ∼0.05, and indeed for some of the more distant sources, a final PSF image, using IDL code by A. A. Breeveld3 designed theUVOTPSFisaslargeasthehostgalaxyitself.OftentheSer- specificallyforgeneratingPSFsfromUVOTimages.Theresulting sicprofilecomponentwouldshrinktoasizecomparabletothatof 3 MullardSpaceScienceLaboratory-http://www.mssl.ucl.ac.uk/ 4 http://heasarc.nasa.gov/docs/heasarc/caldb/swift/docs/uvot/ (cid:13)c 0000RAS,MNRAS000,000–000 Optical–to–X-rayemissioninlow-absorptionAGN 5 thePSF,andwould‘absorb’someofthefluxwhichwouldother- itshouldthereforebemorereadilyidentifiableintheHSTimages. wisebeenreportedascomingfromthePSFcomponent.Therefore, WeperformsimpleaperturephotometryusingtheDS9packageon threecomponentswereadoptedassufficienttogivesensiblemodel imagesfromACSfortheF550MandF330Wfilters(centeredon reconstructionsoftheprofilesseenintheimages(sky+PSF+ex- 5580A˚ and 3353A˚ respectively). We extract counts from the nu- ponential disk).Toget asaccuratearepresentation aspossibleof cleus with a circular region withradius 0.2 arcsec, which should the nuclear flux, the following iterative algorthim was employed provideanupperlimitonthenuclearfluxandcontain>90percent usingGALFIT: ofthecountsfromthenucleus.Wealsodeterminethetotalcounts froma5arcsecregion,asusedwithUVOTSOURCEforcomparison. (i) Allbrightforegroundstarswereexcludedfromthefitbycre- In the F550M filter (closest to the UVOT V-band in wave- atingabadpixelmaskwhichcoveredtheseunwantedobjects. length),weidentifyadifferenceinfluxesofafactor∼5between (ii) Theskybackgroundwasfitindependentlytoablankregion thesmallerandlargeraperturesforHST.Incomparison,thediffer- oftheskyclosetotheAGNofinterest. enceinfluxesbetweenthe5.0arcsecUVOTSOURCEmagnitudeand (iii) The central point source and sky were fit together, keep- thatfromPSFfittingtotheUVOTimageisafactor∼2.4.Assum- ingtheskyparametersconstantfromthepreviousstep.Thecentral ingthe5.0arcsecfluxesfromthetwoinstrumentsshouldtheoreti- pointsourcemagnitudewasseededwiththatobtainedfromUVOT- callybeidentical,thisimpliesthatthefluxreportedtothenucleus SOURCEaperturephotometry,toprovideasensibleinitialguessof byPSFfittingtoUVOTimagesisafactor∼5.0/2.4toolarge.The thenuclearflux. problemislessacuteintheF330Wfilter(closesttotheUVOTU- (iv) The sky, central source and exponential disk components band),withanidenticalanalysisyieldingthatthenuclearfluxfrom werefittedtogether,keepingtheskyandcentralsourceparameters theUVOTimageisafactor∼ 1.8/1.2 ofthatfromtheHSTim- fixedatthevaluesfromthepreviousstep. age.ThereisalsotheissueofopticalAGNvariability,whichcould Afterthesesteps,theresultantmodelfitswerecomparedwith be artificially increasing these ratios (the continuum near 5100A˚ theoriginalimagesandtheresidualimagesobtainedfromsubtract- isknowntovarybyfactorsof∼ 7inthecaseofNGC5548;see ing themodel from the datawere inspected. In thefirst instance, Petersonetal.1999),butitisdifficulttodisentanglethisfromthe thepositionsofthedifferentcomponentswereleftfreetofit.Ifany errorintrinsictotheinstrument,andthefactorswecalculatehere obvioussignsofPSFmismatchorotherproblemswiththefitwere are therefore probably upper limits. This reflects a fundamental evident, these positions were frozen manually by inspection and limitation of the resolution of the UVOT images, but the subse- thefitwasre-evaluated.Insomecases,suchasforpoint-likeAGN quentanalysisshowsthatinmanyoftheobjectsseen,attempting atthehigherendofourredshiftrange,asimplermodelconsisting toremovethehostgalaxydoesproduceanoptical–UVSEDshape ofjustaPSFandaconstantbackground wasadopted,ifthefirst- closertothatexpectedforanAGNdespitepossibleoverestimates passattemptatfittinganexponentialdiskdidnotshowasignificant of the nuclear flux. A more detailed analysis using observations galaxy diskcomponent. WherethePSFimage displayed obvious fromtheKittPeakNationalObservatory(KPNO)2.1mtelescope PSFmismatch,a‘King’profilefitwasalsoattempted.Thebestfit shouldprovideamuchmoredetailedgalaxyprofiledecomposition obtained from all these approaches, determined by visual inspec- (Kossetal.inprep),whichwillmakeuseofthesmallerPSFavail- tionof theresidual image, waschosen toobtain thefinal nuclear ablewiththeKPNO(FWHM∼1.0–1.4arcsec). magnitudes.Anexampleofthesky-nucleus-diskdecompositionis showninFig.2.Thechoicesofprofilesusedforeachfilterineach 3.2 Optical–UVphotometry:Correctingforcoincidenceloss objectaregivenintable3. andGalacticextinction For AGN with very high count rates, the fluxes obtained from 3.1.1 ComparisonwithHST UVOT images are highly susceptible to coincidence losses and ThedetailedstudyofBentzetal.(2006)highlightstheneedtotake therefore often appear lower than their true values. Coincidence into account host galaxy contamination when calculating nuclear lossreferstothephenomenonwheremultiplephotonsarriveatthe luminositiesfromoptical HSTimages, specifically inthecontext same location on the detector during a single frame, and is dis- ofdeterminingtheradius–luminosityrelationshipforthebroadline cussedindetailinPooleetal.(2008).Theythetheoreticalcorrec- region(BLR)inAGN.Theyfindthatevenforanapertureof1arc- tiontobeappliedtothemeasuredcountrateforindividualpixels, sec,therecanbesignificanthostgalaxycontamination. However, but also point out the need to account for the spread of a point theHSTAdvancedCameraforSurveys(ACS)offersasignificant source over many pixels. They also provide an empirical correc- advantage withaPSFof FWHM0.0575 arcsecincontrast tothe tion which takes the latter effect into account. We correct all of typicalFWHMof∼2arcsecfortheUVOTPSFs.Thissomewhat ourmagnitudesfromGALFITforcoincidencelossmanuallyusing limitsourabilitytodeterminethehostgalaxycontamination,since theseexpressions; the UVOTSOURCEaperturephotometry magni- thestarsusedtomodelthePSFoftenareaswideinextentonthe tudesareautomaticallycorrectedforcoincidenceloss. sky as some of the more distant galaxy disks in our sample. We There is some scope for error in the coincidence loss cor- attempttogaugethepossibleunderestimateofhostgalaxycontri- rection when the PSF image is not quite a point source, as the butionwithUVOTusingpublicallyavailableimagesfromtheHST coincidence loss correction only holds good for point sources. ACSandWideFieldandPlanetaryCamera2(WFPC2)foroneof Thisproblemisespeciallypronounced forverybrightsources.In ourobjects,NGC4593.Thisobjectisknowntohaveaclearlyiden- thesesources,thecoincidence-correctedfluxesaresometimeseven tifiablespiralgalaxystructuresurroundingthecentralAGNsopro- greaterthanthoseobtainedfromUVOTSOURCE,whichshouldthe- videsagoodfoundationforcomparisonbetweenUVOTandHST oretically provide an upper limiting nuclear flux. In cases where imagequality.OurUVOTanalysisalsoshowsacleardiscrepancy this phenomenon is very pronounced, we do not use the magni- betweenthe5.0arcsecaperturephotometryresultsandthosefrom tudes from GALFIT and instead employ the 5.0 arcsec aperture PSFfittingintheoptical(V,BandU)bands.Sincesomelevelof photometrymagnitudesinanyfurtheranalysis. galaxy contamination is clearly identifiable using UVOT images, After calculation of magnitudes from both GALFIT and (cid:13)c 0000RAS,MNRAS000,000–000 6 R.V. Vasudevan,R. F.Mushotzky,L. M. Winter& A.C. Fabian 100 200 300 400 500600700800 100 200 300 400 500600700800 100 200 300 400 500600700800 100 200 300 400 500600700800900 100 200 300 400 500600700800900 100 200 300 400 500600700800900 100 200 300 400500 100 200 300 400500 100 200 300 400500 50 100 150 200 50 100 150 200 50 100 150 200 10 203040 10 203040 10 203040 20 4060 20 4060 20 4060 Figure1.ExampleGALFITprofilefittingresultsforMCG-06-30-15.Thefirstcolumnineachrowshowstheoriginaldataimage,thesecondshowsthemodel fitandthethirdshowstheresidualobtainedfromsubtractingthemodelfromthedata.Therows,fromtoptobottom,showtheresultsfortheV,B,U,UVW1, UVM2andUVW2bands,inthatorder.IntheVandBband,theresidualsimplythatabulgecomponentmayalsobediscernibleinthisobject,buttheUVOT datadidnotallowrobustfitsincludingabulgecomponentformanyofourobjects.AsimpleanalysisusingDS9regionsindicatethattheputativebulgemay accountforupto10percentofthetotalgalaxyfluxintheVandBbandimages. UVOTSOURCE,these magnitudes were corrected for Galactic ex- 3.3 X-rayspectrafromXRT tinctionusingthevaluesforE(B−V) fromtheNASA/IPAC Gal Infrared Science Archive5 and the Galactic extinction curve of Cardellietal.(1989).ThefinalmagnitudeswereturnedintoXSPEC PHAfilesusingtheFLX2XSPutility,tofacilitatefittingalongwith Thepipeline-processedeventfilesfromtheXRTdetectorwerepro- thecontemporaneousXRTdatausingtheXSPECanalysissoftware. cessedusingthe XSELECTpackage, asdirectedintheSwiftXRT user guide. Source regions of 50 arcsec were used, with larger accompanying background regions (averageradius∼150arcsec). Backgroundlightcurvesweredeterminedfromtheeventfilesand inspectedforflaring,butthiswasnotfoundtobeaprobleminany ofourobservations.Sourceandbackgroundspectrawereextracted, andthesourcespectraweregroupedwithaminimumof20counts 5 http://irsa.ipac.caltech.edu/applications/DUST/ perbin. (cid:13)c 0000RAS,MNRAS000,000–000 Optical–to–X-rayemissioninlow-absorptionAGN 7 AGN V B U UVW1 UVM2 UVW2 1RXSJ045205.00+493248 P P P P – – 2MASXJ21140128+8204483 P+E P P+E P P+E P+E 3C120 P+E P+E P+E P+E M+E P+E 3C390.3 P+E P+E P P P+E P+E Ark120 P+E P+E P+E P+E P+E P+E ESO490-G026 P+E P+E P+E P+E P+E P+E ESO548-G081* P+E P+E P+E P+E M+E M+E IRAS05589+2828 P P P P M Gaussian IRAS09149-6206 P+E M P+E P+E P+E P+E MCG+04-22-042 P+E P+E P+E P+E M+E P+E MCG-06-30-15 P+E P+E P+E P+E P+E P+E Mrk1018 P+E P+E P+E P+E M+E M+E Mrk279 P+E P+E P+E P+E P+E P+E Mrk352 P+E P+E P+E P+E P+E M+E Mrk509 P+E P+E P+E P+E P+E – Mrk590 P+E P+E P+E P+E P+E P+E Mrk766 P+E P+E P+E P+E P+E P+E Mrk841 P+E P+E P+E P+E M+E P+E NGC4593 P+E P+E P+E P+E P+E P+E NGC5548 P+E P+E P+E P+E P+E P+E NGC7469 P+E P+E P+E P+E P+E P+E NGC985 P+E P+E P+E P+E P+E P+E SBS1136+594 P P P+E P+E P+E P+E SBS1301+540 P+E P P+E P P P+E UGC06728 P+E P+E P+E P+E – P+E WKK1263 P+E P+E P P P P Table3.TableshowingthemodelcomponentsusedinGALFITwitheachfilterimageforeachAGN.‘P’signifiesaPSFgeneratedfromguidestarswithin appropriatecount-rateranges,fromeachimage,‘M’signifiesaMoffat(King)profilewithparametersestimatedfromasinglepointsourceintheimage,and ‘E’signifiesanexponentialdiskmodelcomponent..*ForESO548-G081,anextremelybrightforegroundstarwaslocatedwithin∼30arcsecoftheAGN, renderingthephotometryobtainedsusceptibletolargeerrors. 1000 2000 3000 Figure2.ComparisonbetweenimagesfromHST(left)andUVOT(right)forthesourceNGC4593.A5arcseccircularregionisshownforreferenceonboth images. 3.4 Blackholemassestimatesusingthe2MASSAll-Sky holemassfortheSwift-BATcatalogueAGNfrom2MASSK-band SurveyCatalogues totalsourcemagnitudes,whichinvolvessubtractingthecentralnu- clear component from the extended source flux to estimate the bulgeluminosity.Weattempttorefinethismethodherebyincorpo- TheM −L correlationforgalaxiesprovidesausefulway BH bulge ratinginformationontheexpectedangularsizeofthebulgeonthe ofestimatingtheircentralblackholemasses.Thekeychallengeis skyandtheresolutionlimitationsofthe2MASSdata.Thecrucial toobtainanaccurateestimateofthebulgeluminosity.Thestudyof factor is the degree of seeing affecting the 2MASS observations; Marconi&Hunt(2003)presentsadecompositionofimagesfrom typically the seeing takes a value of 2.6 arcsec. For bulges with 2MASSusingGALFITtoobtainthebulgeluminosityandcorre- typical size ∼ 0.5−1 kpc (Graham&Worley 2008), the bulge late it with the black hole mass from direct determination meth- becomes unresolvable at redshifts of about z ≈ 0.01 (assuming ods (stellar or gas kinematics, maser kinematics, etc.). However, H =71kms−1Mpc−1,Ω =0.27andaflatUniverse).There- the redshifts of the objects under scrutiny here are generally sig- 0 M fore,fortheoverwhelmingmajorityofobjectsinour sample,the nificantlyhigherthanthoseinMarconi&Hunt(2003),presenting bulge should be unresolved. Inspection of the atlas images from problemsforperformingafullnucleus-bulge-discdecomposition. 2MASSforthelowestredshiftobjectsinoursampleconfirmsthat Mushotzkyetal.(2008)presentamethodforcalculationofblack (cid:13)c 0000RAS,MNRAS000,000–000 8 R.V. Vasudevan,R. F.Mushotzky,L. M. Winter& A.C. Fabian thevisuallyidentifiablebulgecomponentisnotresolved.Thebulge 42.0 < log(L ) < 43.0, 43.0 < log(L ) < 44.0, 44.0 < X X light istherefore mixedinwiththe nuclear light inthesesources log(L ) < 45.0). The fractions are plotted in Fig. 3 against the X andidentifiedasa‘point source’.Thepoint sourcesidentifiedby central luminosity of each bin. We also plot a logarithmic inter- 2MASS have been collated, along with their magnitudes, in the polation between the points, to allow determination of the ratio 2MASSPointSourceCatalogue(PSC);similarlyextendedsources of LK,host/LK,total at ageneral value of LX, assuming theratio have been catalogued in the 2MASS Extended Source Catalogue variescontinuously. Forvalues of L greater thanthose spanned X (XSC).Basedonour assumptionthat thebulgeisunresolved, we bytheSilvaetal.(2004)study,wesimplyemploythefractioncal- initially download only the magnitudes from the PSC for further culated for L = 1045ergs−1, extrapolating from the direction X use,astheywillcontainthebulkofthebulgelightwewishtore- ofthetrendshownbetweenthebinscentred onlog(L ) = 43.5 X cover. and 44.5. For luminosities log(L ) < 41.5, we use the fraction X Bywayofcalibratingourattempttoobtainblackholemasses, evaluatedatlog(LX) = 41.5(extrapolatingfromthetrendwould we first turn to the reverberation mapping (RM) sample of AGN yieldfractionslargerthanUnity).Thehighluminosityextrapola- presented by Petersonetal. (2004), representing the most secure tionisclearlynotlikelytobeaccurateforextremelypowerfulob- massdeterminationsforAGNtodate(augmentedwiththerefined jects(logLX > 46), butnone ofthereverberation mappedAGN result from Denneyetal. 2006 for NGC 4593). There are uncer- ortheAGNinourSwift-BATsubsampleareinthisregime.Inthe taintiesintheRMmethod,connectedtoassumptionsaboutthege- absenceofinformationonIRSEDsoutsidethisluminosityrange ometry of the broad-line region and whether it is gravitationally wedonotattemptamorecomplexextrapolation. bound,butthe∼35AGNforwhichRMmassestimatesexistnev- Before using these fractions to calculate bulge luminosities erthelessconstitutethesampleofAGNwiththemostcarefullyde- fromthe2MASSPSCmagnitudes, weconsider potentialsources terminedblackholemasses.Initially,wedownloadthe2MASSat- of bias. SinceSilvaetal.(2004) adopt areprocessing scenario to lasimagesforthoseRMAGNbelow aredshiftof0.01,inwhich accountfortheK-bandcontinuum,itispossiblethatthemodelnu- thebulgeshouldbetheoreticallyresolvedandperformaGALFIT clearK-bandluminosityisafunctionoftheEddingtonratio(since 3-component decompositionusingagaussianPSFforthenuclear itisultimatelylinkedtotheUVaccretiondiscluminosity,thedom- pointsource,aSersicbulgeandanexponentialdisc.OurGALFIT inantcomponent ofthetotalaccretionluminosity).TheX-raylu- analysis yields that there is a significant degree of uncertainty in minositycanberecastastheproductoftheEddingtonratio,bolo- the identification of the bulge component in these AGN, and the metric correction and black hole mass (via the Eddington ratio), magnitudeascribedtothebulgeisheavilydependent ontheprior so the trend seen in Fig. 3 could be the result of two functions constraintsimposed.Weexploreanumberofdifferentconstraints ofEddington ratiobeingplottedagainst eachother. However, we onthemodelprofiles;forexample,theSersicindexnisoftencon- again note the low X-ray luminosities probed by the Swift/BAT strained toan upper limitto avoidthefittingroutine confusing it 9-monthcatalogue(uptoafewtimes1044ergs−1),andfindthat withthe nuclear point source, and often requires alower limitto the host-to-total fraction varies at most by a factor of two inthis avoidconfusionwiththedisccomponent.Despitesuchstrategies, regime.Therefore,thedependenceisnotstrongandwillnotintro- insomesourcesthebulgeanddiscprofilesacquireverysimilaror ducesignificantbiasescomparedtoothersourcesoferror,suchas identicalradiiand‘share’thediscluminosity,withtheSersicpro- the intrinsic spread in RM masses and the intrinsic dispersion of filepartiallytracingthediscinsteadofthebulge.Insuchsources, theMBH −LK,bulge relation.Regardlessoftheparticularmodel three-componentmodelfitsdonotyeildmeaningfulbulgeparame- used topredict thenuclear K-band continuum, itisclear that the ters.OurattemptsatconstrainingthebulgeusingGALFITresultin fraction of an unresolved bulge that isattributable to the nucleus bulgeluminositiesthatvarybyuptoanorderofmagnitudedepend- willdecreaseforlowerluminosityAGNandviceversa;thesimple ingonthechoiceofmodelandconstraintsimposed,andshowthat relationship inFig.3 provides an observationally-rooted estimate evenfortheclosebysources,thebulgesareprobablynotresolved. ofthiseffect. In light of this, we then employ the 2MASS PSC catalogue WeemploythevaluesofLXfromXMM-Newtonreportedin magnitudestoprovideasimpleestimateofthetotalbulgeandnu- Vasudevan&Fabian (2009) to calculate the K-band host-to-total clearflux.Inordertoestimatetherelativecontributionsofthenu- ratioforeachoftheRMobjects,andassumetheSilvaetal.(2004) cleus and bulge in the PSC magnitudes, we employ the infrared Seyfert1IRSEDtemplateforallobjectsinthesample.Intheab- SEDtemplatespresentedinSilvaetal.(2004).Intheirstudy,they senceofadetailed,uniformstudyontheabsorbingcolumnsofthe presentnuclearIRSEDtemplatesconstructedusingavailablenear- RMAGNsamplethisisareasonableassumption,sincetherever- to-midIRdataon33Seyferts,usingaradiativetransfermodelsfor berationmappingsampleisheavilybiasedtowardsSeyfert1AGN; dustheatingtointerpolatebetweenthewavelengthscoveredbythe noneoftheRMAGNareclassifiedasSeyfert2sinNED.Wescale data.TheyalsopresenthostgalaxySEDtemplatesbysubtracting thePSCmagnitude bythe appropriate fraction, usingtheexpres- thenucleartemplatecontributionfromthetotalphotometryateach sion availablewavelength. Theypresent anumber of different nuclear SEDtemplatesappropriatefordifferentlevelsofX-rayabsorption, andhost SEDtemplatesarepresented fordifferent intrinsic2–10 MK,bulge =MK,PSC−2.5log10(fbulge(LX)), (1) keVAGNluminosity(LX)regimes.Usingtheappropriatenuclear where the fraction fbulge = LK,host/LK,total is calculated as SED along with the corresponding host galaxy SED, it is there- above.TheblackholemassesarethencalculatedusingtheM − BH forepossibletoestimatethefractionofthetotalluminositywhich LK,bulge relation of Marconi&Hunt (2003). The resulting mass canbeaccountedforbythehostforanAGNwithaparticularL . estimatesarecomparedwithRMmassestimatesinFig.4. X Intheprocesswemaketheassumptionthat,inthecaseoftheK- Areasonableagreementbetweenthetwomethodsisseen,al- band,thehostfluxshouldbedominatedbythebulgeasdiscussed beit with some significant deviation at higher masses. The grey in §1. We calculate the ratio of host to total luminosity (nuclear shaded areahighlightstheuncertaintiesinRMmassesduetothe plushostluminosity)fromtheseSEDtemplatesintheK-bandfor lackof preciseknowledge of theBLRgeometry. Thelargestdis- the four different luminosity regimes (41.0 < log(L ) < 42.0, crepancyisfortheAGNwiththehighestmass,3C273,forwhich X (cid:13)c 0000RAS,MNRAS000,000–000 Optical–to–X-rayemissioninlow-absorptionAGN 9 11 1 0.9 des10 u nit 0.8 mag LK,total 0.7 ASS PSC 9 =L/K,host 00..56 m scaling 2M 8 fbulge 00..34 (bulge) froMK om 7 0.2 M frBH 0.1 6 40 41 42 43 44 45 46 6 7 8 9 10 11 log(LX) Reverberation mapping MBH Figure3.Variation ofLK,host/LK,total withX-rayluminosity, extrap- Figure 4. Comparison between masses from reverberation mapping and olated fromthehostandnuclear IRSEDtemplates ofSilvaetal.(2004) massesestimatedfromK-bandbulgeluminosities.Thesolidlinerepresents (black filledpoints -determined usingthefourfiducialluminosities pro- thedesired one-to-one correspondence between thetwomethods andthe videdintheirpaper;solidline-extrapolation between points asdetailed shadedareadepictstheintrinsicuncertaintiesinreverberationmapping.The inthetext).Thesefractions areemployedtocalculate thecontributionof dashed line shows thebest fitoftheform log(MBH,LK,bulge) = A+ thebulge tothe2MASS point sourcemagnitudes, fordetermining black log(MBH,revamp). holemasses.ThecurveforanAGNwithintrinsicX-rayabsorptionNH = 1023.5cm−2 isshown(dashedline)forcomparison(thecurvesforother absorptionsaboveNH = 1022cm−2areverysimilar;thesignificantdif- gestapossiblecontributortothisoffsetandpresentacorrectionfor ferenceisbetweenobscuredandunobscuredAGN). it. The Silvaetal. (2004) host galaxy SEDs have been calcu- lated using large-aperture and extended-source data (such as the theK-bandbulgeluminosity estimategives ablack holemassan 2MASS XSC catalogue magnitudes) which could include more order of magnitude too high. This is known to be a very power- thanthebulgelight.Ifthegalaxydiscissignificantinsomecases, fulobject,anditispossiblethatoursimpleapproachtoestimating theSilvaetal.(2004)host-to-totalratiosactuallyprovidethefrac- thebulgefractioninthepointsourcelighthasstillunderestimated tion of bulge plus galaxy disc to the total luminosity rather than thepowerful nucleus. Thereisalsolikelytobeasignificant syn- justthebulge,andthereforeusingthemtoscalethe‘nucleusplus chrotron (non-accretion) contribution to fluxes at all wavelengths bulge’luminositiesfromthe2MASSPSCisnotstrictlycorrect.It inthissource,whichfurtherincreasesuncertainties.ThisAGNis isthereforenecessarytogaugetheeffectofneglectingthegalaxy far more powerful than the majority of the Swift-BAT catalogue disc in thiscalculation. We do this by obtaining the 2MASS Ex- AGN, so such problems are not likely to affect the lower-power tended SourceCatalogue magnitudes (matching thecentroids ap- objectsweconsiderhere. propriatelyas before), using thesefluxesasestimatesof thetotal Marconietal.(2008)pointoutthatthepreviousreverberation nuclear, bulge and disc light (L = L +L +L ). total nuc bulge disc massestimateshavebeencalculatedwiththeassumptionthatthe Employing thePSCmagnitudes asthebulgeand nuclear light as BLRisgravitationallybound,andintroduceacorrectionbasedon before(L +L ),wethencanutilisethefractionofnuclear- bulge nuc anestimator of theaccretion rate(the5100A˚ monochromatic lu- to-total luminosity predicted from the Silvaetal. (2004) IR SED minosity)toproviderevisedmassestimatesfortheRMAGNcat- templates(f =L /[L +L +L ])toobtainathe- nuc nuc nuc bulge disc alogue.Wealsoprovideacomparisonwiththeirrevisedmasses- oreticallymoreaccurateestimateofthebulgemagnitude: timates, in Fig. 5. The correlation between the two mass estima- torsisstrengthenedsignificantlyifthesefirst-ordercorrectionsto thereverberation massesareapplied,andparticularlyreducesthe MK(2,)bulge=MK,PSC−2.5log10(1−fnuc10MK,PSC2−.5MK,XSC)), scatterseenforthoseobjectswithhighaccretionrates(forwhich (2) thecorrectionismaximalandworkstoincreasethereverberation Thisexpression isobtained by solving for L fromtheavail- bulge massestimate).Theoffsetisstillpresenthowever,andasimplefit able quantities. Put simply, it arises from using the nuclear frac- of the form log(MBH,LK,bulge) = A+log(MBH,revamp) yields tionfromtheIRSEDtemplatestoestimatethenuclearfluxfrom A ≈ 0.57inbothcases(usingreverberationmasseswithorwith- the XSC flux; this nuclear component is then subtracted from outcorrectionforradiationpressure),implyinganoffsetofafactor the PSC flux to obtain the bulge flux. This estimate is not avail- ofupto∼ 3.7inblackholemass,justoutsidethetypicalquoted able for all of the objects available previously, as some objects toleranceforreverberationmasses(afactorof∼ 3).Wenowsug- do not have 2MASS XSC magnitudes. For a couple of sources, (cid:13)c 0000RAS,MNRAS000,000–000 10 R.V. Vasudevan,R. F.Mushotzky,L. M.Winter&A. C. Fabian 11 11 S10 10 AS 003) from 2M 9 & Hunt 2003) 9 & Hunt 2 Marconi M from L (Marconi BHK,bulge 78 log(M) from L (BHK,bulge 78 6 6 6 7 8 9 10 11 6 7 8 9 10 11 MBH from reverberation mapping (corrected for radiation pressure effects) log(MBH) from reverberation mapping Figure5.ComparisonbetweenmassesfromRM(corrected forradiation Figure 6. Comparison between masses from reverberation mapping and pressureeffectsasdiscussedbyMarconietal.2008)andmassesestimated massesestimatedfromK-bandbulgeluminosities (withoffsetduetoun- fromK-bandbulgeluminosities.Theemptywhitecirclesshowtheresults derestimationofthegalaxydisccontributiontakenintoaccount).Keyasin using the reverberation masses before correcting for radiation pressure. Fig.5. OtherkeyconventionsasinFig.4 (2009),andtheN valuesinthesamepaperareusedtoselectbe- H the predicted fraction fnuc = Lnuc/(Lnuc + Lbulge + Ldisc) tweenthetwohost-to-totalratiocurvesforobscuredorunobscured fromtheSilvaetal.(2004)templateswaslargerthanthefraction AGN(Fig. 3) asrequired. Bywayof another check ontheaccu- (Lnuc+Lbulge)/(Lnuc+Lbulge+Ldisc)foundfromtheratioofthe racyofourmassestimates,therecentstudyofWangetal.(2009) PSCandXSCfluxes,leadingtoanimpossiblesolutionforLbulge. presentsvirialmassestimatesforasampleofSeyfertswhichpar- Thisindicates that the nuclear-to-total fraction obtained from the tiallyoverlapswithSwift-BAT9-monthcatalogueAGN,usingthe Silvaetal. (2004) templates is not appropriate for that object, or single-epoch Hβ linewidth-based black hole mass estimator (as thatthebulgeisnotcompletelycontainedinthePSCflux.These giveninGreene&Ho2005).Weplotacomparisonbetweentheir problemslimittheuseofthismethod,butforthoseobjectswherea valuesandoursinFig.7forthe24overlappingobjects.Whilethe calculationispossible,wefindthatthenewK-bandmassestimates relation shows some scatter, we caution that the linewidth-based are systematically shifted down by 0.14 dex (afactor 0.72) from black hole mass estimation methods also show significant scatter thevaluesfromthesimplerapproachofusingthePSCmagnitude themselveswhencomparedtoRM.Thetwoestimatorsstillappear alone. We therefore apply this correction to all masses obtained to trace each other (with a correlation coefficient of 0.80); addi- using the PSC alone, which brings the best-fit line for the new tionally, preliminary calculations of black hole masses from Kitt mass estimates within the tolerance of the reverberation masses. PeakNationalObservatory(KPNO)andSloanDigitalSkySurvey The strength of the correlation is good, asindicated by the Pear- (SDSS) spectra for the Swift/BAT 9-month catalogue AGN also soncorrelationcoefficientsof0.83and0.93obtained(correlating seemtomatchourestimateswell,withascatterofafactorof∼2 withreverberationmassesbeforeandafterradiationpressurecor- (Winter et al., Koss et al. in prep). We see some indications that rection).Thisanalysisalsoindicatesthatthegalaxydisccontribu- thelocalSwift-BAT9-monthcatalogueAGNhavepredominantly tioninthehost-to-totalratiosisoftheorder∼ 30percent,anda lowblackholemasses,asexpectedin‘anti-hierarchical’structure similarvalueappliestoallAGNintheRMsample.Therestillre- formation scenarios. The masses thus determined are used in all mainssomeoffsetabovereverberationandwithinthetolerancere- subsequentanalysesforcalculationofaccretionratesandforcon- gion,butanyfurtherattempttoaccountforthisissomewhatredun- strainingaccretiondiscmodelsinSEDconstruction. dantinlightoftheintrinsicuncertaintiesinreverberationmasses, sowedonotmakeanyothercorrectionstoourapproach.Theoff- setofvirialmassestimatesbelowtheM −L relationhas BH bulge 4 GENERATIONOFSPECTRALENERGY beennotedanddiscussedindetailbyKimetal.(2008b);theyfind DISTRIBUTIONS that it isacomplex function of avariety of parameters including theblackholeaccretionrateandhostgalaxymorphology. We fit the optical, UV and X-ray SED data points with a simple Wefinallycalculatethemassesusingthisapproachforallof multicolour disc DISKPN and absorbed power-law model combi- thez<0.1objectsintheSwift-BAT9-monthcatalogueforwhich nationusingXSPEC,followingtheapproachoutlinedinVF09.In 2MASS PSC magnitudes are available. The K-band host-to-total thiswork,weextendtheapproachofVF09byinclusionofamulti- ratiosarecalculatedusingtheLX valuesreportedinWinteretal. plicativeZDUSTcomponentontheDISKPNmodelusedfortheUV (cid:13)c 0000RAS,MNRAS000,000–000
Description: