SECOND MINOR REVISION SUBMITTED TOTheAstrophysicalJournalON 23/12/2015. ORIGINAL VERSION: 7/8/2015 PreprinttypesetusingLATEXstyleemulateapjv.5/2/11 TOWARDSPRECISIONSUPERMASSIVEBLACKHOLEMASSESUSINGMEGAMASERDISKS REMCO C. E. VAN DEN BOSCH1, JENNY E. GREENE2, JAMES A. BRAATZ3, ANCA CONSTANTIN4, CHENG-YU KUO5 SecondminorrevisionsubmittedtoTheAstrophysicalJournalon23/12/2015.Originalversion:7/8/2015 Abstract Megamaserdisksprovidethemostpreciseandaccurateextragalacticsupermassiveblackholemasses. Here wedescribeasearchformegamasersinnearbygalaxiesusingtheGreenBankTelescope(GBT). Wefocuson galaxieswherewebelievethatwecanresolvethegravitationalsphereofinfluenceoftheblackholeandderive 6 astellarorgasdynamicalmeasurementwithopticalorNIRobservations. Sincethereareonlyahandfulof 1 supermassiveblackholes(SMBH)thathavedirectblackholemassmeasurementsfrommorethanonemethod, 0 evenasinglegalaxywithamegamaserdiskandastellardynamicalblackholemasswouldprovidenecessary 2 checksonthestellardynamicalmethods. Wetargeted87objectsfromtheHobby-EberlyTelescopeMassive n GalaxySurvey,anddetectednonewmaserdisks. Mostofthetargetedobjectsareellipticalgalaxieswithtypical u stellarvelocitydispersionsof250kms−1 anddistanceswithin130Mpc. Wediscusstheimplicationsofour J non-detections,whethertheyimplyathresholdX-rayluminosityrequiredformasing,orpossiblyreflectthe 4 difficultyofmaintainingamasingdiskaroundmuchmoremassive(>108M )blackholesatlowEddington ∼ (cid:12) ratio. Giventhepowerofmaserdisksatprobingblackholeaccretionanddemographics,wesuggestthatfuture ] masersearchesshouldendeavourtoremoveremainingsamplebiases,inordertosortouttheimportanceof A thesecovarianteffects. G . h 1. INTRODUCTION for cross validation of other black hole mass measurement p Watermegamasersat22GHzaredetectedin∼3%oflocal techniques. - Todate,thereareonlyahandfulofcross-checksondynam- Seyfert2andLINERgalaxiesthathavebeensearched. In o ical black hole mass measurements. The different methods r roughlyone-thirdofthese,characteristicred-andblue-shifted t components indicate thatthey arise from sub-pc scales in a oftendonotyieldconsistentmasses(e.g. Walshetal.2013, s 2012;Onkenetal.2014). Evenstellardynamicalmethodsdo a geometricallythinaccretiondiskaroundaweaklyaccreting notalwaysagree;see,forinstance,NGC3379(Shapiroetal. [ supermassiveblackhole(BH;Lo2005;Pesceetal.2015).The 2006;vandenBosch&deZeeuw2010),NGC1399(Houghton best-knownexampleisNGC4258(Miyoshietal.1995),but 2 etal.2006;Gebhardtetal.2007)andNGC4258(Siopisetal. therearenowmorethan15megamaserdiskgalaxiesknown v 2009; Drehmer et al. 2015). There is no single culprit for thatshowclean Keplerian rotation, and34totalmegamaser 5 thesediscrepancies,becausedifferentmethodsandpractices diskgalaxies(Zhuetal.2011;Pesceetal.2015). 4 areusedasgasandstarsprobethepotentialindifferentways. Despitetheirsmallnumbers,watermegamasershavedispro- 6 The inhomogeneityofthe mass measurements, andthe low portionatescientificimpact. Thousandsofgalaxieshavebeen 0 numberstatistics,makeitimpossibletoquantifythesystematic searched for maser activity, largely by the Megamaser Cos- 0 uncertainty. . mologyProject(MCP, Reidetal.2009;Braatzetal.2010). 1 By measuring the acceleration of the systemic water maser Sofar,theonlymegamaserdiskgalaxywitheitherastellar 0 orgas dynamical BH mass measurement is NGC4258. For features,itispossibletomeasuredirectgeometricdistancesto 6 this object, the state-of-the-artorbit-basedmodels basedon thesegalaxies(e.g.,Humphreysetal.2013). Themaingoalof 1 HST opticallong-slitspectroscopyfindablackholemassthat theMCPistogarnerapreciseandindependentmeasurement : is15%lowerthanthemegamaser-derivedmass(Siopisetal. v ofH (Reidetal.2013;Kuoetal.2013,2015). 0 2009), whereas simpler Jeans models with adaptive optics i Modelingofthenear-Keplerianrotationofthemaserspots X near-infrared integral field spectroscopy find a value that is alsoyieldsverypreciseandaccurateBHmasses,withuncer- r taintiesof∼10%thataredominatedbytheuncertaintyinthe 25%toohigh(Drehmeretal.2015). Furthermore,thesizes a ofthequoteduncertaintiesaresuchthatneitherofthesemea- galaxydistance. ThedeviationsfromKeplerianrotationareso surementsarewithin3σ ofthemaservalue. Thisistypical smallthatitispossibletoruleoutastrophysicalalternatives fortheothercross-checksalso. Itthusappearsthatsystematic tosupermassiveBHsinthesesystems(e.g.,Kuoetal.2011). unknownsdominatetheuncertainty. Morecomparisonsare Theseblackholemasseshavemuchsmalleruncertaintiesthan clearlyneededtoquantifyandunderstandthesediscrepancies. thoseobtainedbyothermethodsandarethusideallysuited Furthercross-checkswouldalsoprovideinformationonthe intrinsicscatterofBH-galaxyscalingrelations. Thewidely 1Max-PlanckInstitutfürAstronomie,Königstuhl17,D-69117Heidel- usedMBH−σ∗ relationcorrelatesthevelocitydispersionof berg,Germany,[email protected] thehostgalaxywiththeblackholemass. Thisrelation(and 2DepartmentofAstrophysics,PrincetonUniversity,Princeton,NJ08544, manyothers)haveanintrinsicscatterof∼0.4dex(ormore) USA,[email protected] 3NationalRadioAstronomyObservatory,520EdgemontRoad,Char- (e.g.,Beifiorietal.2012). Theintrinsicscatterasafunctionof lottesville,VA22903,USA galaxypropertiesshouldcontainimportantcluesontheorigin 4DepartmentofPhysicsandAstronomy,JamesMadisonUniversity,Har- ofthescalingrelations(e.g.,Robertsonetal.2006;Peng2007; risonburg,VA22807,USA Jahnke & Macciò 2011). However, we cannot measure the 5PhysicsDepartment,NationalSunYat-SenUniversity,No.70,Lien- HaiRd,KaosiungCity80424,Taiwan,R.O.C intrinsicscatterrobustlyuntilweunderstandtheunderlying 2 vandenBosch,Greene,etal. systematicuncertaintiesinthedynamicalBHmasses. Perhaps tomeasurethedynamicalimpactoftheBHonthesurrounding thesizeofthemeasuredscatterisartificiallyinflatedbythe stars or gas. We use the Hobby-Eberly Telescope Massive systematicuncertaintiesoftheblackholemassmeasurements? Galaxy Survey (HETMGS, van den Bosch et al. 2015) as Withcurrentinstrumentationitwillbeverydifficulttoat- ourparentsample. TheHETMGSusesseveraldifferenttarget temptanycross-calibrationusingthemegamaserblackhole selectionstofindallgalaxiesthataresuitableforstellar-and masses. Most of the maser galaxies are so distant that it is gas-dynamicalsupermassiveBHmassmeasurementsin the impossibletospatiallyresolvethestellarmotionsintheregion opticalornear-infrared. Thisisachievedbymaximizingthe ofthegalaxywherethegravityoftheBHdominates(known apparentsizeofthesphere-of-influenceoftheSMBHsinthe asthegravitationalsphereofinfluence, see§2). Thanksto targetgalaxies. TheHETMGSsurveycontains1022galaxies thesuperiorspatialresolutionoftheVLBIobservations,the andobservednearlyallpossibletargetsfordynamicalblack maserscanprobewellwithinthesphereofinfluenceforBHs holemassesmeasurementsintheNorth(−11<δ <73◦). It downto106M evenat100Mpc. Usingstellarorgasdynam- isthusanidealbasisforanearbymasersearch. (cid:12) icaltechniquesitisnotyetpossibletomeasurea106 M at (cid:12) distancesbeyond2.5Mpc. 2.1. SearchCriteria Theknownmaserswithspheresofinfluencelargeenough We apply two selection criteria. First, we use the stellar (>0.1(cid:48)(cid:48)) for a robust measurement from a dynamical tech- velocitydispersionmeasurementfromHETMGStoroughly niqueareCircinus,NGC4945andNGC4258. Thelatterwas estimate whether or not a dynamical BH measurement will discussedabove. Theothertwomasersshownon-diskcompo- bepossible. Specifically,thegravitationalsphereofinfluence nents,whichmakestheirinterpretationmoreambiguousthan oftheBH(θ ∝ GMBH; whereDisthedistance)needstobe the‘clean‘masers(Pesceetal.2015). Measuringtheirblack spatiallyresoIlvedD(θσ∗2>0.06(cid:48)(cid:48))inordertoproberegionsnear I holemasswithadynamicaltechniquewouldbeagoodtest the black hole. Since we have no measurement of the BH of the masermass. Howeverthese objects are not ideal for mass,weuseM ∝σ4fromGültekinetal.(2009)asaproxy6. BH ∗ understandingthesystematicuncertaintyofthedynamicalBH Thereare369suchgalaxiesintheHETMGS. masses. Second, we only select galaxies with optical signs of nu- Thus,theonlywaytoincreasethesampleofgalaxieswith clearactivity. Ideally,wewouldsearchall369galaxieswith directcomparisonsbetweenstellar/gasdynamicalandmega- θ >0.06(cid:48)(cid:48),butinordertoboosttheprobabilityoffindinga I maserdiskmassesistofindanewmegamaserdiskinagalaxy maserdisk,werestrictedourattentiontogalaxieswithopti- where we can resolve the gravitational sphere of influence. cal signatures of accretion. We use the HETMGS spectra7 Thatgoalmotivatedthemegamasersearchpresentedhere. We to measure strong emission lines ratios [N ] λ6583/H and II α focusongalaxieswiththelargestspheresofinfluenceonthe [O ]λ5007/H . Standard Baldwin, Phillips, & Terlevich III β sky, to boost the chances that we can get a megamaser and (BPT,1981)diagnosticdiagramsareusedtoisolatetheactive stellar/gasdynamicalmassinthesameobject. galacticnuclei(AGN;Fig. 1)fromthegalaxieswith(nuclear) Previoussurveyshavelookedformegamaserdisksinavari- star-formation. Selecting onlythe AGN furtherreduces the etyofgalaxytypes. SeeHenkeletal.(2005)foranicesum- targetlistto121objects.Forthissearch,wedidnotdistinguish mary. Themajorityofsearchestodatehavefocusedonknown betweenSeyfertsandlow-ionizationnuclearemissionregion obscuredactivegalaxies,tomaximizethechancetocatchan LINERs(Heckmanetal.1981),butseesection3.4. edge-ondisk. Theseactivegalaxieshavebeenselectedina Finally,wecheckedforoverlapbetweentheHETMGSand varietyofways. Themajorityofgalaxieswereobservedwith the GBT searches. After removing all galaxies previously thesensitiveRobertC. ByrdGreenBankTelescope(GBT), searchedformegamaseremission,93galaxiesremainedinour whichhasagainof∼1.9K/JyatK-band. TheMCPhascom- targetlist. ThislastcutremovedalmostallofthebrightAGN piledapubliclist(weblink)of∼3400galaxiesthathavebeen fromourtargetlist. SeeFig.1forthedistributionintheBPT searchedformegamaserwiththeGBT. Recentwork(Zhuetal. diagramoftheobservedgalaxies. 2011;Zhangetal.2010,2012;Constantin2012)investigates thedetectionfractionofmegamasergalaxiesasafunctionof 2.2. GBTObservationsandDataReduction variousgalaxyparameters. Becauseofourgoaltomaximize We conducted our survey using the K-Band Focal Plane the likely BH sphere of influence, our survey is dominated Array(KFPA)andtheGBTSpectrometer. Weused2beams by relatively massive elliptical galaxies relative to previous ofthe KFPA in the same mode usedby the MCP. We used searches,providingnewconstraintsonthephysicalproperties 2IF’sof200MHzbandwidthoffsetby180MHz,foratotal thatleadsomegalaxiestomase. coverageof380MHz. Eachwindowcovers2800kms−1and, Herewepresentourunsuccessfulattempttoidentifynew withoverlap, thetotalcoverageis5100kms−1. Wedonot masingdisksinnearbygalaxies. In§2wediscussthesample knowapriorithevelocityextentoftheputativemegamaser selection andobservationsofourmasersearch. Wediscuss disks in these more massive galaxies. However, if we had overall maser detection fractions in §3, and discuss the im- plications of our results for the physics of and detection of 6TheHETMGSonlymeasuredthevelocitydispersionσcinsideacentral megamaserdisksin§4,andconcludein§5. WeadoptaHub- ∼2(cid:48)(cid:48)aperture.HenceanMBH−σcispresentedinvandenBoschetal.(2015) bleconstantH of70kms−1. Thisvalueisconsistentwithall whichwouldbemoreappropriatetopredictblackholemasses.Thetraditional 0 thepublishedvaluesbasedongeometricdistancedetermina- MBH−σ∗usesthedispersionmeasuredwithinthegalaxyeffectiveradiusσe, whichistypicallymuchlargerthan2arcseconds. Likemanymorerecent tionsofmegamaserdiskgalaxies(e.g. Reidetal.2013;Kuo relations(McConnell&Ma2013;Kormendy&Ho2013),therelationbased etal.2013). onσcissteeperthantheMBH∝σ∗4fromGültekinetal.(2009)thatisadopted inthiswork.Therelationwithσcwouldyield631objectswithθI>0.06(cid:48)(cid:48), 2. OURSEARCHFORMEGAMASERDISKGALAXIES butthisfitwasnotavailablewhenweperformedoursearch. 7BecausetheHET’sopticalpathchangesduringeachoftheobservations, Ourgoalistofindnewmegamasersinnearbygalaxiesfor anabsolutefluxcalibrationwasnotperformedontheHETdata.Thespectra cross calibration of BH mass measurements. Thus we are werecorrectedin arelativesenseforthespectralresponse. Howeverno searchingformegamaserdisksingalaxiesthatarenearenough absolutelineluminositiesareavailablefortheHETMGSspectroscopy. NoMaserDisksinEllipticalGalaxies 3 1.5 withlikelyBHmassesinexcessof108M . Sincethesphere (cid:12) HETGBT of influence depends steeply on σ∗, our sample is heavily Megamasers skewedtowardshigh-dispersiongalaxies. AsshowninFigure HETMGS 1.0 2,ourprogrammorethandoublesthenumberofgalaxieswith σ >250kms−1thathavebeensurveyedformaseremission with the GBT. The observed galaxies have larger inferred ) β spheres-of-influencethanthegalaxiespreviouslysearchedas H 0.5 / well. AsshowninFigure2,thissurveydoublesthenumberof ] OIII galaxiessearchedwithθI>0.06(cid:48)(cid:48). Only 44 objects have a morphological type T in Hyper- [ g( 0.0 leda(Patureletal.2003). Almostalloftheseareellipticals: o l T = −3.2±1.9. The only known spiral in our sample is IC0356. Ingeneral,wedonothavedeepandreliableimaging -0.5 withwhichtodetermineHubbletypes. Thus,toputthemor- phologicaldistributionofoursampleincontext,weplotthe mass-sizerelationofallgalaxiessearchedformasersbythe -1.0 GBT. Ellipticalsobeyatightscalingbetweensizeandmass -1.0 -0.5 0.0 0.5 andarethesmallest(densest)galaxiesatagivenstellarmass. AsFigure3shows,theHETMGSgalaxiesthatwesearched log([NII]/H ) α preferentiallyprobedenseellipticalgalaxies,whileprevious Figure1. The BPT diagnostic diagram usedto identify active galaxies searcheswereoverwhelminglydominatedbyspiralgalaxies. forthissurvey. Shownaretheemissionlinesratios[NII]/Hα λ6583and Oursearchprobesthelocusofearly-typegalaxiesandadds [OIII]/Hβ λ5007fromtheHETMGSsurvey.GreencrossesrepresenttheHET- toapartofparameterspacethathasnotyetbeenexhaustively GBTsamplethatweobserved.Reddiamondsareknownmegamasersinthe searched. HETMGS.Forreference,theremainingHETMGSsourcesinwhichemission linesaredetectedareshownasblackcrosses.Theupperandlowerblacklines delineatetheempiricalandtheoreticaldivide(Kewleyetal.2006)between HETGBT star-formationandotheremission(AGN,shocks).Ourtargetsareselectedto Megamasers havelargespheresofinfluenceandnon-starformingnuclearemission,i.eto Non-detections fallontheupper-rightsideofthedivide. detectedonlysystemicmasersinanysystem,wewouldhave expandedthesearchtohighervelocities.Weusedatotal-power c) 10 p observingmode,noddingthetelescopetoalternatethetarget (k galaxybetweenthetwobeamsona2.5-minuteinterval. The us receiverpointedoffsourceisusedasthereferencebeam to di a measure the background. The reference beam spectrum is ht r smoothedwithakernelof16channels,toincreaseitssignal- g tmoi-nnuotiesse.peIrnggaloaoxdy,wweeaathcehrie(vTesdys∼∼44m0JKyr)masndpeirnctehgarnanteinlgaf1te0r Half-li Hanningsmoothing. Thepointingcorrections,doneroughly 1 eachhour,weretypically5(cid:48)(cid:48)orbetter,andthefluxcalibration is accurate to about 20%. The final velocity resolution and channelspacingis0.4and0.3kms−1. Thisobservingsetup issufficienttoidentifyanymegamasersthatcanbeimagedin follow-upobservationswithasingleVLBItrack. 109 1010 1011 We observed 87 galaxies using 21.5 hours during a visit totheNRAO’sGBT8 from6through15November2012as Stellar mass (MO •) program GBT/12B-052. See Table 1 with observations of Figure3. Size-mass diagram ofsearchedgalaxies, using half-lightradii this HETGBT sample. Apartfrom galaxies from ourtarget andstellarmassesfromtheNASASloanAtlasbyBlanton&Moustakas list,the87galaxiesincludefivefillerobjectsandthecontrol (2009)forallobjectswhereSDSSphotometryisavailable.Thereare2927 GBTnon-detections(bluedots),83masers(reddiamonds)and41HETGBT maserNGC6240. Thefillerobjectshavepropertiescloseto galaxiesthatwesearched(greencrosses)withSDSSphotometry.TheMCP ourselectioncriteriaandareincludedinourHETGBTsample. non-detectionscoverthemainpopulationoflate-typegalaxies,whereasour Atourmaximumdistanceof130Mpc,the3σ luminositylimit searchprobedthelocusofmassiveearly-typegalaxies,thatwerehithertonot is∼0.7L ,whiletheleastluminousknownH Omasingdisk sampled. (cid:12) 2 isinNGC2273,whichhasanisotropicluminosityof23L , (cid:12) sowearenotlimitedbysensitivity. Therewasalargesurveyofellipticalgalaxiescarriedout byHenkeletal.(1998),butitwasfocusedonluminousradio 2.3. SampleProperties galaxies,soselectedinadifferentwayfromthissearch.Indeed, Thesamplehasamediandistanceof74Mpcandamedian thereareonlyafewknownmegamasersinellipticalgalaxies: stellarvelocitydispersionof250kms−1. Themajorityofthe NGC1052 (Braatz et al. 1994; Tarchi et al. 2003), 3C403 sample galaxies are elliptical galaxies (see next paragraph), (Tarchi et al.2007), Centaurus A(Ottetal. 2013), possibly NGC2960 (Kormendy & Ho 2013). Object TXS2226-184 8TheNationalRadioAstronomyObservatoryisafacilityoftheNational is most likely not an early-type galaxy (Falcke et al. 2000). Science Foundation operatedundercooperative agreementbyAssociated AbouthalfofoursampleisdetectedinNVSS(NRAOVLA Universities,Inc. 4 vandenBosch,Greene,etal. Figure2. Left:Histogramofthestellarvelocitydispersions.WeshowourHETGBTsamplethatwesearchedformasersinblacksolid,theMCPnon-detections indotted,allMCPmasinggalaxiesinbluedashedandthemaserdiskgalaxiesinredlong-dashed.Thezoomedinregionhasthesamehorizontalspanbutshows thatwemorethandoublethenumberofσ∗>250kms−1galaxies.Right:Histogramofthespheres-of-influenceofknownmegamaserdisks(red),theHETGBT (black)andtheknownnon-detectionsfromtheMCP(dotted).Galaxieswithlargespheres-of-influenceareinherentlyrare. SkySurvey), witha median luminosityof10 mJy(Condon megamasers.Theserepresenttheluminositythatthemasersys- etal.1998). temwouldhaveifitwereisotropic. Thetrueluminositiesare veryuncertainastheyarehighlydependentonthe(generally 3. DETECTIONFRACTIONS unknown)beamingdistribution(Kuoetal.2011). Wedidnotdetectanynewmasersinthe87objectssurveyed. Weshouldalsonotethattheremaywellbesubtlebiasesin Theoveralldetectionrateofmasersis<3%forallgalaxies, thegalaxiesthathavebeentargetedspectroscopicallybythe hence our non-detections could just be due to low number SDSS. Thesebiasesarethenimprintedonthesubsetoftar- statistics. Assuminganaveragedetectionrateof3%andusing getedgalaxiespresentedhere. Forinstance,heavilyreddened the simple binomialprobabilitydistribution, the probability galaxiesmaybelesslikelytofallintotheSDSSmaingalaxy ofdetectingnomasers(7%)isonly1/3oftheprobabilityof sample(Straussetal.2002). Moremassivegalaxiesarealso detecting2or3masers(25%and22%respectively)giventhe lesslikelytobetargetedspectroscopicallyatverylowredshift samplesizeof87galaxies. (e.g.,Fukugitaetal.2007). Therefore,weshouldbewaryof Toputournon-detectionsintocontext,weshowmaser(and jumpingtoverystrongconclusionsuntilthesebiasesarealso maserdisk)detectionfractionsasafunctionofσ andL studiedandaccountedfor(Constantinetal. inpreparation). ∗ [OIII] inFigure4. Belowwecombineoursamplewithasubsetof thegalaxiessearchedwithGBTthathaveopticalspectroscopy 3.2. DetectionasaFunctionofVelocityDispersion fromtheliterature. BothZhuetal.(2011)andConstantin(2012)pointedout arisingfractionofmegamasergalaxiesasthegalaxystellar 3.1. Nondetections velocitydispersionrises. However,thereareveryfewgalax- As described above, the largest current megamaser disk iesintheirsampleswithσ∗>160kms−1. Nearlyallofour search with uniform sensitivity has been carried out by the galaxiesfallinthishigh-dispersionregime. InFigure4(left), GBT. We combine the full list of galaxies searched by the we show the maserfraction as a function ofσ∗ priorto our GBT with our smaller list to search for trends between de- survey (long-dashed lines) as well as the detection fraction tectionsandopticalpropertiesofthegalaxies. Tomaximize afteraddingallofournon-detections(solid). Unfortunately, thenumberofsearchedgalaxieswithliteratureopticalspec- justbasedonPoissonerrorsalone,westilldonothaveenough troscopy,werelyontheSloanDigitalSkySurvey. Constantin data in the high-dispersion bins to make a significant mea- etal. (inprep)startedwiththe3339masernon-detectionsas surementofthemaserfractionathighdispersion. However, ofJune2013,alongwith151galaxieswithmaserdetections. itisclearthatthecontinuedriseindetectionfractionabove Theycross-matchedthesegalaxieswiththeSloanDigitalSky σ∗≈150kms−1isnotreal. Ifwerestrictourattentiontothe Survey(SDSS Yorket al. 2000) andthe PalomarSurveyof megamaserdiskgalaxiesalone,thenwestillseearisetowards nearby galaxies (Ho et al. 1997a). There are spectroscopic σ∗≈150kms−1,butthemeasurementsatlowdispersionare matchesfor1330ofthenon-detections,and92ofthemaser alsohighlyuncertainduetosmallnumbers. galaxies,whichinclude15maserdisks. Fromthesematches, 3.3. DetectionasaFunctionofLuminosity wehavemeasurementsofstellarvelocitydispersion(σ ),and ∗ emission-lineproperties,includingtheBalmerdecrementand Apartfromthedependenceonvelocitydispersion,thetrend the[OIII]luminositycorrectedforextinction.Wewillalsodis- seenmostclearlyinZhuetal.(2011)andConstantin(2012)is cussmeasurementsofthemaserluminosities,whichtypically anincreaseddetectionfractionathigher[OIII]luminosity. Re- fallintherangeoftenstothousandsofsolarluminositiesfor callthatinSeyfertgalaxies,the[OIII]luminosityisanindirect NoMaserDisksinEllipticalGalaxies 5 Figure4. Left:Weshowthedetectionfractionasafunctionofstellarvelocitydispersion.ThefractionofmasersofallsortsintheMCPsample(bluedashed)is comparedtothedetectionfractionwhenthenon-detectionsfromthispaperareincluded(solidblue).Likewise,wecomparethemaserdiskdetectionfractionfrom Constantinetal.(redlong-dashedline)withthecorrecteddetectionfractionwhenoursampleisincluded(redsolid).Thesetwolinesarequitesimilar.Inboth panels,onlybinswithmorethan30objectsareshown.Itisclearthattheapparenttrendtowardsahigherdetectionfractionathigherdispersionisatleastpartially duetosmallnumberstatisticsinthesebins.Asweaddmorepoints,weseethatthedetectionfractionsdonotsignificantlyrisetowardshigherσ∗.Inthehighest dispersionbintheHETGBTmorethandoublesthenumberofsearchedobjects.Notethatthereareonly30masingdisksandthusthemaserdetectionratessuffer fromlownumberstatistics.Right:Similartoabove,weplotthemaserdetectionfractionasafunctionofthe[OIII]luminosity.Again,onlybinswithatleast30 galaxiesareshown.Inthiscasetherisingdetectionfractiontowardsmoreluminousactivegalaxiesappearstobereal.Again,onlybinswithatleast30galaxiesare shown. indicatorofthebolometricluminosityoftheAGN(e.g.,Yee 1980),andisoftenusedwhenthenon-thermalcontinuumcan- notbedirectlymeasured(e.g.,Zakamskaetal.2003;Heckman etal.2004;Liuetal.2009). InFigure4(right)weagainshowthemaserandmaserdisk detectionfractionsforthesampleofgalaxiessearchedbythe MCP. Thetrendtowardshigherdetectionfractionathigher [OIII]luminosityisquiteclearin thiscase, even excluding uncertain bins. Because we do not have accurate measure- mentsofthe[OIII]luminosityfortheHETMGSsample,we donotincludethem. ThetenHETGBTgalaxieswithSDSS spectra would fall at the faint end of the distribution, with L ≈5×1037 ergs−1. Fromthisfigurealone,onemight [OIII] concludethatournon-detectionsaredueentirelytothelumi- nositydistributionofthesourceswetargeted. Ifmegamaser diskluminosityandAGNbolometricluminosityarecorrelated (e.g.,Henkeletal.2005;Kondratkoetal.2006b)thenperhaps wesimplydidnothavethesensitivitytodetectthepossibly veryfaintmasersaroundtheseveryweakAGN. However,wesuggestthatthereismoretothestory. Exami- nationofFigure5revealsthatwhilethereisaweakcorrelation Figure5. Relationbetween22GHzH2Omaseremissionluminosity(as- betweenthemaserluminosityandL[OIII]whenlookingatall sumedisotropic)andtheobservedextinction-corrected[OIII]λ5007luminosity maser sources, this correlation vanishes for the maser disk forallknownwatermegamasersintheSDSS(greypoints).Megamaserdisks arecircled.Thebigredtriangleisthe3σupperlimitonthemaserluminosity sourcestakenalone. Instead,theyspanaverynarrowrange ofasourceatthemediansampledistanceof74Mpc,assumingourdetection inisotropicmaserluminositythatisvirtuallyindependentof limitof0.2L(cid:12). The[OIII]luminosityisthemedianforthesubsampleof L . Asimilartrendcanbeseenwhenlookingatafunction objectswithSDSSspectra.Themegamaserdiskgalaxiesspanaverynarrow [OIII] ofhardX-rayluminosityinKondratkoetal.(2006a);mostof rangeinisotropicmaserluminosity. Thisisexpected, giventheconstant surfacedensitypredictedbyNeufeldetal.(1994)andthesimilarsize(∼0.5 thecorrelationisdrivenbythenon-diskmasers. Atthesame pc)ofallthemasingdisks. time, ourtypicalL luminosity is wellbelow the typical [OIII] luminosityforknownmegamaserdisks(includingNGC4258). Perhapswearedetectingatruethresholdinluminosity,below luminosity of known maser disks (Kondratko et al. 2006a), which the temperature condition for masing (∼400 K e.g., is roughly consistent with the calculations of Neufeld et al. Neufeldetal. 1994)isnotmet. AnL ≈1038ergs−1cor- (1994). In §4, we discuss in more detail the physical con- [OIII] respondsroughlytoL ≈1040−1041ergs−1,dependingon nectionbetweenluminosity,BHmass,anddisksizethatmay X thebolometriccorrectionandtheassumeddustcorrectionfor causethedearthofmasersweobserve. L (Vasudevan&Fabian2007;Liuetal.2009;Shaoetal. [OIII] 2013). Thisthreshold,whichliesbelowtheobservedX-ray 3.4. LINERsandSeyferts 6 vandenBosch,Greene,etal. Wecanalsoaskwhetherthedetectionfractiondependson fromamoleculartoatomicdisk,toscalewiththeX-raylumi- the type of nuclear activity that we observe. In particular, nosity(Kondratkoetal.2006a). Suchascalingisobserved whilelineratiosassociatedwithhigh-ionizationSeyfertgalax- byWardle&Yusef-Zadeh(2012). Letusnowimaginethat iesaredifficulttoachievethroughanymechanismotherthan wetakeNGC4258asamodelforthemaserdisksthatwehad accretionontoanSMBH,LINERactivityhasawidearrayof hopedtofindinourellipticalgalaxysample. NGC4258isa causes, including shocks (see reviewin Ho 2008). Particu- goodanalogbecausebothitandtheellipticalsareaccretingat larlyinlarge-aperturedata,previousstudies(Sarzietal.2006; verylowfractionsoftheirEddingtonluminosity(Ho2008). In Singhetal.2013)haveshownthatthecentralLINERemission fact,giventhetypicalobservedluminosities,andinferredBH oftendoesnotarisedirectlyfromanactivegalacticnucleus. massesof∼108M ,wedoexpecttypicalEddingtonratiosof (cid:12) OurHETGBTsamplecontainsequalnumbersofLINERsand ∼10−4,asinNGC4258(Herrnsteinetal.2005) Seyferts and almost all are low luminosity, as indicated by We can imagine scaling the properties of the NGC4258 thelowAmplitude-over-Noiseoftheemissionlines(vanden maserdisktoinferthelikelypropertiesofcomparablemolec- Boschetal.2015). ulardisks around elliptical galaxies. If R ∝L (as shown cr X The majority of the megamaser detections are found in byKondratkoetal.2006a)then,providedthattheEddington Seyfert galaxies (see Figure 1), likely because many of the fraction,accretionefficiencyandtheX-ray-to-bolometriceffi- LINERsintheSDSSarenotactuallypoweredbynuclearac- ciencystayroughlyconstant(likelyavalidassumption,e.g., cretion. Ontheotherhand,NGC4258isaLINER,sowehad Vasudevan&Fabian2007),wefindthatR ∝M . Thatis, cr BH somehopethatasurveyfocusedonlow-luminositysystems the size of the accretion disk would scale linearly with the like NGC4258 wouldhave a higheryield(Ho et al. 1997b). BHmass. Inoursample,weexpecttheBHsareroughlyan Thatwasnotthecase. Weconcludethatitismoreimportantto orderofmagnitudemoremassivethantheBHinNGC4258. selectgalaxiesabovethepossibleX-rayluminositythreshold Therefore,ifthisscalingroughlyholds,theirmoleculardisks thantoworrydirectlyabouttheopticallineratios. wouldhavesizesof∼5−10pc. ThisscalingbetweenX-rayluminosity,BHmass,andEd- 4. DISCUSSION dington ratio, introduces a possible explanation for the low Our detected maser fraction of (cid:46)1% is nominally lower incidenceofmegamaserdisksinlocalellipticalgalaxies. Be- thanthatseenbythepreviousmasersearches. Whilealarger causeoftheuniformlylowEddingtonratioandhighM ,the BH sampleisneededtobesureofthislowerdetectionfraction,we accretiondisksizewillnaturallygrowbyatleastanorderof herediscussinmoredetailthemostlikelycausesofourlow magnitude. LikewisetheToomre(1964)Qparameterwould detectionfractionandimplicationsforfuturesearches. The decreasebyafullorderofmagnitudeatconstantdisksurface mostobvious differences between this sample andprevious density,asQdependslinearlyontheorbitalfrequency,andso samplesare: (1)lowerAGNluminositiesand(2)moremas- inverselyonR. Itisnotclearthataccretiondiskswouldbesta- sivehostgalaxieswithhigherstellarvelocitydispersions. Asa bleataradiusof10pc. Thus,onepossibleexplanationforthe result,thegalaxiesthatwetargetlikelyhavehigherBHmasses lackofmasersisthatathighM andlowEddingtonfraction, BH andverylowmassaccretionrates. Takingalloftheseproper- itisnolongerphysicallypossibletomaintainamoleculardisk tiesintoaccount,weinvestigatevariousexplanationsforthe withconditionsappropriateformasing. lackofmaserdetections. Firstwediscussthenon-detections Inactuality,theEddingtonfractionintheellipticalsmaybe of megamaser disks, then we discuss the non-detections of evenlowerthaninNGC4258. InthePalomarspectroscopic megamasersofanysort. survey of galaxies, Ho (2008) finds that LINERs radiate at typicalEddingtonfractionsof∼10−5,butspanalargerange 4.1. AGNLuminosityandMaserDiskSize from10−7to10−3. IftheEddingtonratiodropssystematically It is easy to imagine that the megamaser luminosity will towardsmoremassivesystems,thesituationismorecompli- correlatewiththeluminosityoftheAGNonaverage(Neufeld catedthanoutlinedinthepreviousparagraphandthedisksize 2000). ItistheX-raycoronathatmostlikelyheatsthemolec- wouldnotgrowasrapidlywithBHmass. ularaccretiondisk,thuscreatingametastablepopulationof 4.2. EnvironmentalEffects excitedwatermoleculesthatareabletomase(Herrnsteinetal. 2005). AweakcorrelationisseenbetweenhardX-raylumi- Anotherpossibilityisthattheellipticalgalaxyenvironment nosityandmegamaserluminositywhenallmegamaserAGN islessconducivetosupportingmoleculardisks,evenatafixed areconsidered(Kondratkoetal.2005),aswellasacorrelation Eddingtonfraction. HighermassgalaxiesdohostmorehotX- betweenmegamaserluminosityandFIRluminosity(Henkel raygas,whichcouldinhibittheformationoflargeamountsof et al. 2005). Furthermore, it is clearfrom Figure 4 that the densemoleculargas(Wiklind&Henkel2001;O’Sullivanetal. detectionfractionofmegamasersingeneral,andmaserdisks 2001;Sarzietal.2006).Albeitonmuchlargerscales,thereare inspecific,bothriseathigher[OIII](andthereforebolomet- documenteddifferencesinthepropertiesofmoleculardisksin ric)luminosity. However,asarguedabove,giventhelackof ellipticalandspiralgalaxies. Davisetal.(2014)findthatthe correlationbetweenmegamaserdiskluminosityandAGNlu- moleculardisksinearly-typegalaxieshavestar-formationrates minosity(Fig. 5),wefinditunlikelythatournon-detections farbelowthoseexpectedfromtheSchmidt-Kennicuttrelation canbeexplainedbythepresenceofveryfaintmasersaround (e.g.,Kennicutt1998). Theirexplanationisthatthehighlevels ourverylowluminosityAGN. Wefavorthepossibilitythat ofshearintheinnerdisksuppressesstarformation. Themaser thereisathresholdL ≈1040ergs−1,belowwhichthecondi- disksareevendeeperinthepotentialwell,andwespeculate X tionsformasingarenolongermet. Specifically,thesizeofthe thatthe gravitationalpotentialmayleadto a suppression of maserdiskrequiredatlowaccretionrateandhighBHmass masing. maysimplygrowtoolargetobestable. 4.3. JetMasers According to Neufeld et al. (1994), masing disks are ex- pectedtohaveaveryconstantsurfaceluminositydensity. We Evenifwedonotdetectanymasingdisks,givenoursample thusexpecttheouterradiusofthediskR ,setbythetransition size we might have expected to detect megamasers associ- cr NoMaserDisksinEllipticalGalaxies 7 atedwithjetactivity. Asiswell-documentedintheliterature, usingstellardynamics. IftheBHmasstogalaxymassratio jetsgrowmoreprevalentinmassiveellipticalgalaxies(e.g., issmallenough,thentherewillbeno“gravitationalsphereof Matthews et al. 1964; Best et al. 2005; Mandelbaum et al. influence”wheretheBHmassdominatesthestellardynamics. 2009)andahighfractionofellipticalgalaxynucleicontain TherewillbenokinematicsignatureoftheBHonthestars. lowlevelsofradioemission(e.g.,Sadleretal.1989;Wrobel Ontheotherhand,asub-pcscalemegamaserdiskcouldstill &Heeschen1991). Furthermore,ellipticalgalaxieswithradio bewithinthesphereofinfluenceoftheBH,andthusprobethe sources are also more likelyto contain dustlanes (e.g., van (hithertounexplored)regimeofverylowM /M . BH gal Dokkum&Franx1995)anddustisusuallyaccompaniedwith Inclosing,were-emphasizetheimportanceofmegamaser moleculargas,whichisneededtoformorfuelthemasingdisk. disksinrevealingBHdemographicsatlowmass. Atlowσ ∗ Giventherisingfrequencyofjetemissioninellipticalgalaxies, in particular, there are not many remaining galaxies in the andparticularlyinthemostluminousofellipticalgalaxies,our universewhereweshouldbeabletoresolvethegravitational non-detectionsareparticularlyinteresting. sphereofinfluencewithpresent-daytechnology(assumingthat Herewemightagaininvokeluminosity. Certainlyjetpower σ isavalidproxy;Batcheldor2010;Gültekinetal.2011;van ∗ isknowntocorrelatewith[OIII]luminosity(e.g.,Ho&Peng denBoschetal.2015, theirFig.10),andatlargerdispersion 2001). Another culprit is the lower gas mass in the ellipti- andluminosities,thesituationisonlyalittlebetter. Evenwith calgalaxy. Itispossiblethatbecausethereislessmolecular newfacilitieslikeJWST,ALMA(Davis2014),andELTs(Do gas along the path of the jet, there are fewer opportunities etal.2014),theimprovementinspatialresolutionwillonly formasing. Anotherfactoristhegeometryofthejet,asthe probeoneorderofmagnitudesmallerinBHmassatafixed Doppler-boostedforwardjetmaylieinfrontofthebulkofthe galaxyproperty,whichwillnotallowustoprobethefullrange moleculargas(seealsoHenkeletal.1998). inBHmassatσ <200kms−1. Hence,eveninthecoming ∗ decades,masingdisksandindirectAGNmethodswillremain 4.4. MegamaserDisksandBlackHoleDemographics criticalforfindinglow-massBHs(Reines&Volonteri2015). Asmentionedabove,oneofourprimarymotivationsinem- barking on this survey was to study the reliability of stellar 5. CONCLUSIONS dynamicalBHmasses(e.g.,Gebhardtetal.2003)viacompar- Using the GBT we have surveyed 87 galaxies with large isonwiththeBHmassderivedfromamegamaserdisk. Pre- dispersion(>250kms−1)insearchofwatermegamaseremis- ciousfewgalaxiestodayhaveBHmassmeasurementsbased sion at 22 GHz associated with AGN activity. The overall onmultipletechniques(e.g., Siopisetal.2009; Walshetal. detectionrateinprevioussurveys,whichcovermostlyspiral 2013),andevenonesuchcomparisoninamassiveelliptical galaxies, is 3%, while here we detected no masers in our wouldbeextremelyimportant. mostlyellipticalsample. Wediscussvariousexplanations,in- Megamaserdisksprovidesomeoftheonlyrobustandunbi- cludinglownumberstatistics, localchangesrelatedtoM BH asedmeasurementsofM inspiralgalaxies. Wehavethus andaccretionrateleadingtoverylargeunstabledisks,oren- BH usedthesesystemstomeasurescalingrelationsinlate-type vironmentaldifferencesduetothedeeppotentialorgas-free spiralgalaxiesthatcanbestudiedindetailwithstellardynam- natureofellipticalgalaxies. ics. WeseenocompellingcorrelationsbetweenBHmassand Thegoalofoursearchwastofindamasingdiskinagalaxy galaxypropertiesinthisregime(e.g.,Greeneetal.2010;Sun inwhichdualBHmassmeasurementwaspossibletocross- etal.2013). calibratedifferenttechniques. Continuingthissearchisworth- Thus far, known megamaser disks are found orbiting ac- while, but the yield could be low. Since taking the GBT tivegalacticnuclei(AGNs)withBHmassesthatareclustered observations reported here, more objects were added to the around∼107M (Herrnsteinetal.2005;Kuoetal.2011)and HETMGS,including35masersasspecialtargets(§2.5invan (cid:12) mostlyfoundinmassivespiralgalaxies(Greeneetal.2010). denBoschetal.2015). Intotal480ofthe1022galaxiesinthe Thisnarrowrangeinpropertiesislikelyanaturalbiproduct HETMGShavebeensearchedformasers. Howevermostof of the selection technique. In general, megamasersearches theremainingHETMGSgalaxiesaretoofarawayforastellar havefocusedonknownactivegalaxies(e.g.,Braatzetal.1997; dynamical black hole mass, or have no AGN-like emission Kondratkoetal.2006b;Greenhilletal.2009),predominantly lines. ThereremainonlyadozenobjectsinthefinalHETMGS selectedbasedontheirlocationinopticaldiagnosticdiagrams thatsatisfytheselectioncriteriaofthisworkandhaveyettobe (Baldwin et al. 1981; Veilleux & Osterbrock 1987). Since searched. Galaxieswithlargespheres-of-influenceareinher- thetypicalobscuredactivegalaxyinopticallyselectedsam- entlyrare. DroppingtheAGN-likeemission-linerequirement pleshasM ∼107M (Heckmanetal.2004;Greene&Ho increases the remaining candidates to 350 objects thathave BH (cid:12) 2007),itisnotsurprisingthatthemajorityofthediscovered notyetbeensearchedforamaser,dependingonthechoiceof megamaser disk galaxies also have similar BH masses. At blackholescalingrelation. Giventhevalueofsuchamaser higherBHmasses, theBHsarenolongerhighlyactiveand forcross calibration ofblackhole mass measurements, this thusarenotincludedintheSeyfertgalaxycatalogs. Atlower wouldbeaworthwhilesearch. BHmasses,thedearthofsourcesisaselectioneffectdueto(a) Alternatively,thedetectionratecanbemaximizedbytarget- thedifficultiesofisolatingaccretionsignaturesinthepresence tinggalaxieswithsimilarpropertiesasexistingmaserdisks. ofdustandstarformationand(b)thefactthatlow-massBHs The SDSS-DR7 spectroscopic galaxy sample is fairly close are faint even when radiating at their Eddington limit (e.g., to a random sampling of galaxies (York et al. 2000) and it Greene&Ho2007;Reinesetal.2013). contains – hopefully representative – 14 disk masers, all of Having failedto finda megamaserdiskaroundan M > whicharewithin156Mpc,have0.5<log[O ]/H <1.2,-0.5 BH III β 107 M SMBH, we wish here to make a slightly different <log[N ]/H <0.5,90<σ <190kms−1 andhaveintrin- (cid:12) II α ∗ pointaboutthepotentialbiasesinherentinstellardynamical sicL ≈1039−1043 ergs−1 (correctedbasedonBalmer [OIII] BHmassmeasurements(e.g.,Gültekinetal.2011). Inprin- decrements and a nominal λ−0.7 extinction law). This is a ciplemegamaserdisksinmassiveellipticalgalaxieshavethe verynarrowwindowofgalaxyproperties. Thereareonly65 potentialtorevealapopulationofBHsthatcannotbefound galaxies inside this box of properties in the SDSS spectro- 8 vandenBosch,Greene,etal. scopicsamplethathavealreadybeensearchedunsuccessfully Henkel,C.,Peck,A.B.,Tarchi,A.,etal.2005,A&A,436,75 withthe GBT formasers. Taking this numberatface value Henkel,C.,Wang,Y.P.,Falcke,H.,Wilson,A.S.,&Braatz,J.A.1998, providesaveryhigh,aposteriori,detectionrateof18%,given A&A,335,463 Herrnstein,J.R.,Moran,J.M.,Greenhill,L.J.,&Trotter,A.S.2005,ApJ, the14thatweredetected. Ifall14+65=79ofthesegalaxies 629,719 haveamasingdisk,thisdetectionfractionimpliesanopening Ho,L.C.2008,ARA&A,46,475 angleof10degrees,consistentwiththe8degreesobservedin Ho,L.C.,Filippenko,A.V.,&Sargent,W.L.W.1997a,ApJS,112,315 NGC4258(Braggetal.2000;Herrnsteinetal.2005). Theo- Ho,L.C.,Filippenko,A.V.,Sargent,W.L.W.,&Peng,C.Y.1997b,ApJS, reticalestimatesoftheopeninganglearemuchsmaller(0.1 112,391 Ho,L.C.,&Peng,C.Y.2001,ApJ,555,650 degrees,Lo2005),howevertheobservedbeamingisincreased Houghton,R.C.W.,Magorrian,J.,Sarzi,M.,etal.2006,MNRAS,367,2 by the warping that is often observed in these disks. There Humphreys,E.M.L.,Reid,M.J.,Moran,J.M.,Greenhill,L.J.,&Argon, remainlessthan430spectraintheSDSSwiththesameprop- A.L.2013,ApJ,775,13 erties,thathaveyettobesearched. Atthesamedetectionrate, Jahnke,K.,&Macciò,A.V.2011,ApJ,734,92 we would find 34 more. However, we caution that the box Kennicutt,Jr.,R.C.1998,ARA&A,36,189 Kewley,L.J.,Groves,B.,Kauffmann,G.,&Heckman,T.2006,MNRAS, wedrewinphysicalpropertieswasratherarbitraryandmay 372,961 notinfactincreaseourchancestoobserveadditionalmasers. Kondratko,P.T.,Greenhill,L.J.,&Moran,J.M.2005,ApJ,618,618 Furthermore,thesenewobjectswouldnothavelargespheres —.2006a,ApJ,652,136 ofinfluence,norwouldtheyalleviatethebiasesinthemaser Kondratko,P.T.,Greenhill,L.J.,Moran,J.M.,etal.2006b,ApJ,638,100 searches. Kormendy,J.,&Ho,L.C.2013,ARA&A,51,511 Kuo,C.Y.,Braatz,J.A.,Reid,M.J.,etal.2013,ApJ,767,155 Ultimatelyweneedbetterstatisticstodeterminedefinitively Kuo,C.Y.,etal.2011,ApJ,727,20 whyellipticalgalaxiesmayhostfewermegamasers. Itwould Kuo,C.Y.,Braatz,J.A.,Lo,K.Y.,etal.2015,ApJ,800,26 beinterestingtotargetsampleswithknownradioemission,to Lauer,T.R.,Faber,S.M.,Gebhardt,K.,etal.2005,AJ,129,2138 seeifwecanfindmasersassociatedwithjetactivity,although Liu,X.,Zakamska,N.L.,Greene,J.E.,etal.2009,ApJ,702,1098 Lo,K.Y.2005,ARA&A,43,625 previous searches done this way have not had a high yield Mandelbaum,R.,Li,C.,Kauffmann,G.,&White,S.D.M.2009,MNRAS, (Henkel et al. 1998). Likewise, it would be useful to target 393,377 sourceswithknowngasdiskson∼100pcscales(e.g.,Lauer Martini,P.,Dicken,D.,&Storchi-Bergmann,T.2013,ApJ,766,121 etal.2005;Martinietal.2013)toseeifthesemaybemore Matthews,T.A.,Morgan,W.W.,&Schmidt,M.1964,ApJ,140,35 likely to host masers. Many of these searches will require McConnell,N.J.,&Ma,C.-P.2013,ApJ,764,184 Miyoshi,M.,Moran,J.,Herrnstein,J.,etal.1995,Nature,373,127 largervolumesthansearchedhere,butrecentlymegamasers Neufeld,D.A.2000,ApJ,542,L99 diskshavebeendetectedwithdistancesaslargeas156Mpc Neufeld,D.A.,Maloney,P.R.,&Conger,S.1994,ApJ,436,L127 andsothereishope. Onken,C.A.,Valluri,M.,Brown,J.S.,etal.2014,ApJ,791,37 O’Sullivan,E.,Forbes,D.A.,&Ponman,T.J.2001,MNRAS,328,461 Ott,J.,Meier,D.S.,McCoy,M.,etal.2013,ApJ,771,L41 WewanttothankMartinBureauandTimDavisfororganis- Paturel,G.,Petit,C.,Prugniel,P.,etal.2003,A&A,412,45 ingtheSMBHWorkshopinOxfordinMarch2015,with- Peng,C.Y.2007,ApJ,671,1098 outwhichthismanuscriptwouldnothavebeenresurrected. Pesce,D.W.,Braatz,J.A.,Condon,J.J.,etal.2015,ApJ,810,65 JEGacknowledgessupportfromNSFgrantAST-1310405. Reid,M.J.,Braatz,J.A.,Condon,J.J.,etal.2009,ApJ,695,287 —.2013,ApJ,767,154 Reines,A.E.,Greene,J.E.,&Geha,M.2013,ApJ,775,116 REFERENCES Reines,A.E.,&Volonteri,M.2015,ApJ,813,82 Robertson,B.,Hernquist,L.,Cox,T.J.,etal.2006,ApJ,641,90 Baldwin,J.A.,Phillips,M.M.,&Terlevich,R.1981,PASP,93,5 Sadler,E.M.,Jenkins,C.R.,&Kotanyi,C.G.1989,MNRAS,240,591 Batcheldor,D.2010,ApJ,711,L108 Sarzi,M.,Falcón-Barroso,J.,Davies,R.L.,etal.2006,MNRAS,366,1151 Beifiori,A.,Courteau,S.,Corsini,E.M.,&Zhu,Y.2012,MNRAS,419, Shao,L.,Kauffmann,G.,Li,C.,Wang,J.,&Heckman,T.M.2013,MNRAS, 2497 436,3451 Best,P.N.,Kauffmann,G.,Heckman,T.M.,etal.2005,MNRAS,362,25 Shapiro,K.L.,Cappellari,M.,deZeeuw,T.,etal.2006,MNRAS,370,559 Blanton,M.R.,&Moustakas,J.2009,ARA&A,47,159 Singh,R.,vandeVen,G.,Jahnke,K.,etal.2013,A&A,558,A43 Braatz,J.A.,Reid,M.J.,Humphreys,E.M.L.,etal.2010,ApJ,718,657 Siopis,C.,Gebhardt,K.,Lauer,T.R.,etal.2009,ApJ,693,946 Braatz,J.A.,Wilson,A.S.,&Henkel,C.1994,ApJ,437,L99 Strauss,M.A.,Weinberg,D.H.,Lupton,R.H.,etal.2002,AJ,124,1810 —.1997,ApJS,110,321 Sun,A.-L.,Greene,J.E.,Impellizzeri,C.M.V.,etal.2013,ApJ,778,47 Bragg,A.E.,Greenhill,L.J.,Moran,J.M.,&Henkel,C.2000,ApJ,535,73 Tarchi,A.,Brunthaler,A.,Henkel,C.,etal.2007,A&A,475,497 Condon,J.J.,Cotton,W.D.,Greisen,E.W.,etal.1998,AJ,115,1693 Tarchi,A.,Henkel,C.,Chiaberge,M.,&Menten,K.M.2003,A&A,407, Constantin,A.2012,JournalofPhysicsConferenceSeries,372,012047 L33 Davis,T.A.2014,MNRAS,443,911 Toomre,A.1964,ApJ,139,1217 Davis,T.A.,Young,L.M.,Crocker,A.F.,etal.2014,MNRAS,444,3427 vandenBosch,R.C.E.,&deZeeuw,P.T.2010,MNRAS,401,1770 Do,T.,Wright,S.A.,Barth,A.J.,etal.2014,AJ,147,93 vandenBosch,R.C.E.,Gebhardt,K.,Gültekin,K.,Yıldırım,A.,&Walsh, Drehmer,D.A.,Storchi-Bergmann,T.,Ferrari,F.,Cappellari,M.,&Riffel, J.L.2015,ApJS,218,10 R.A.2015,MNRAS,450,128 vanDokkum,P.G.,&Franx,M.1995,AJ,110,2027 Falcke,H.,Wilson,A.S.,Henkel,C.,Brunthaler,A.,&Braatz,J.A.2000, Vasudevan,R.V.,&Fabian,A.C.2007,MNRAS,381,1235 ApJ,530,L13 Veilleux,S.,&Osterbrock,D.E.1987,ApJS,63,295 Fukugita,M.,Nakamura,O.,Okamura,S.,etal.2007,AJ,134,579 Walsh,J.L.,Barth,A.J.,Ho,L.C.,&Sarzi,M.2013,ApJ,770,86 Gebhardt,K.,Richstone,D.,Tremaine,S.,etal.2003,ApJ,583,92 Walsh,J.L.,vandenBosch,R.C.E.,Barth,A.J.,&Sarzi,M.2012,ApJ, Gebhardt,K.,Lauer,T.R.,Pinkney,J.,etal.2007,ApJ,671,1321 753,79 Greene,J.E.,&Ho,L.C.2007,ApJ,667,131 Wardle,M.,&Yusef-Zadeh,F.2012,ApJ,750,L38 Greene,J.E.,etal.2010,ApJ,721,26 Wiklind,T.,&Henkel,C.2001,A&A,375,797 Greenhill,L.J.,Kondratko,P.T.,Moran,J.M.,&Tilak,A.2009,ApJ,707, Wrobel,J.M.,&Heeschen,D.S.1991,AJ,101,148 787 Yee,H.K.C.1980,ApJ,241,894 Gültekin,K.,Tremaine,S.,Loeb,A.,&Richstone,D.O.2011,ApJ,738,17 York,D.G.,etal.2000,AJ,120,1579 Gültekin,K.,Richstone,D.O.,Gebhardt,K.,etal.2009,ApJ,698,198 Zakamska,N.L.,etal.2003,AJ,126,2125 Heckman,T.M.,Kauffmann,G.,Brinchmann,J.,etal.2004,ApJ,613,109 Zhang,J.S.,Henkel,C.,Guo,Q.,Wang,H.G.,&Fan,J.H.2010,ApJ,708, Heckman,T.M.,Miley,G.K.,vanBreugel,W.J.M.,&Butcher,H.R.1981, 1528 ApJ,247,403 Zhang,J.S.,Henkel,C.,Guo,Q.,&Wang,J.2012,A&A,538,A152 NoMaserDisksinEllipticalGalaxies 9 Zhu,G.,Zaw,I.,Blanton,M.R.,&Greenhill,L.J.2011,ApJ,742,73 10 vandenBosch,Greene,etal. Source RA Dec Vel Date Tsys Int. RMS kms−1 Kelvin min. mJy NGC0050 00:14:44.60 -07:20:42.0 5701 2012-11-09 45.8 10 4.6 NGC0093 00:22:03.20 +22:24:29.0 5380 2012-11-08 42.4 10 4.4 NGC0311 00:57:32.70 +30:16:51.0 5065 2012-11-08 44.5 10 4.8 NGC0384 01:07:25.00 +32:17:34.0 4233 2012-11-16 38.3 10 3.9 NGC0430 01:12:59.90 -00:15:09.1 5299 2012-11-16 44.6 10 4.6 NGC0533 01:25:31.42 +01:45:34.3 5549 2012-11-09 45.6 9 5.0 NGC0550 01:26:42.60 +02:01:20.9 5829 2012-11-09 46.2 10 5.1 NGC0584 01:31:20.81 -06:52:05.0 1802 2012-11-16 48.3 10 5.2 PGC006116 01:39:09.01 +48:16:56.9 5467 2012-11-06 32.9 10 3.5 NGC0898 02:23:20.40 +41:57:05.1 5495 2012-11-06 33.0 10 3.5 UGC01859 02:24:44.40 +42:37:22.9 5917 2012-11-06 32.6 10 3.2 NGC0982 02:35:24.89 +40:52:11.0 5737 2012-11-16 37.1 10 3.9 ARK090 02:42:29.00 +18:09:53.0 9508 2012-11-09 44.5 10 4.7 UGC02261 02:48:17.42 +50:48:00.8 4903 2012-11-06 32.5 10 3.2 NGC1153 02:58:10.30 +03:21:43.0 3126 2012-11-16 41.6 10 4.2 UGC02495 03:02:06.69 +41:35:37.1 9135 2012-11-16 36.4 10 3.7 NGC1208 03:06:11.90 -09:32:29.1 4356 2012-11-16 44.7 10 4.7 PGC138608 03:13:53.23 +62:32:58.9 3050 2012-11-10 87.3 9 9.6 UGC02755 03:29:23.91 +39:47:32.0 7326 2012-11-07 45.4 10 4.9 UGC02866 03:50:14.89 +70:05:41.0 1232 2012-11-06 34.8 10 3.4 UGC02881 03:52:16.91 +36:14:12.9 5764 2012-11-10 87.7 10 9.8 NGC1465 03:53:31.90 +32:29:34.0 4194 2012-11-16 38.0 10 3.9 NGC1469 04:00:27.71 +68:34:40.0 1102 2012-11-06 34.6 10 3.8 IC0356 04:07:46.90 +69:48:45.1 895 2012-11-06 34.4 10 3.6 IC0359 04:12:28.29 +27:42:07.1 4053 2012-11-07 44.9 9 4.9 UGC03024 04:22:26.61 +27:17:51.6 5236 2012-11-10 79.1 10 8.8 2M04310 04:31:05.21 +23:24:08.0 5105 2012-11-07 43.5 10 4.7 PGC165398 04:31:57.09 +59:25:47.0 4630 2012-11-07 40.0 10 4.1 NGC1653 04:45:47.40 -02:23:34.0 4331 2012-11-16 40.6 10 4.2 UGC03386 06:02:37.89 +65:22:16.0 4607 2012-11-07 37.1 10 3.7 PGC019864 06:55:27.70 +33:16:50.0 5302 2012-11-07 36.3 10 4.0 PGC020827 07:22:10.90 -05:55:47.1 1618 2012-11-08 43.1 10 4.7 UGC03855 07:28:13.04 +58:30:23.8 3167 2012-11-07 35.2 10 3.5 NGC2411 07:34:36.39 +18:16:52.9 5073 2012-11-08 38.2 10 3.9 NGC2522 08:06:13.52 +17:42:23.1 4705 2012-11-07 37.4 10 3.9 PGC023680 08:26:24.91 +59:53:42.8 7993 2012-11-07 33.1 10 3.3 MRK1216 08:28:47.10 -06:56:25.1 6394 2012-11-08 40.1 10 4.4 NGC2787 09:19:18.49 +69:12:12.1 696 2012-11-06 36.8 10 3.8 NGC3277 10:32:55.50 +28:30:42.1 1408 2012-11-08 36.6 10 3.8 IC0624 10:36:15.19 -08:20:02.1 5042 2012-11-08 40.2 10 4.1 NGC3348 10:47:09.98 +72:50:23.1 2837 2012-11-10 64.3 10 6.1 PGC032873 10:56:15.99 +42:19:58.9 7471 2012-11-10 56.3 10 5.7 PGC036650 11:45:27.69 +20:48:26.1 6935 2012-11-08 36.6 10 3.9 NGC3869 11:45:45.61 +10:49:29.1 3043 2012-11-08 38.7 10 3.9 NGC3894 11:48:50.42 +59:24:56.0 3223 2012-11-06 41.6 10 4.8 NGC3919 11:50:41.51 +20:00:53.9 6195 2012-11-14 37.5 10 4.1 NGC3992 11:57:36.01 +53:22:28.0 1048 2012-11-10 61.4 10 5.9 NGC4125 12:08:06.00 +65:10:27.1 1356 2012-11-06 41.8 10 4.6 NGC4256 12:18:43.01 +65:53:53.2 2528 2012-11-06 43.0 10 4.8 NGC4403 12:26:12.81 -07:41:06.0 5200 2012-11-10 67.8 10 7.7 NGC4646 12:42:52.19 +54:51:22.0 4647 2012-11-06 45.2 10 4.9 NGC4673 12:45:34.70 +27:03:39.3 6852 2012-11-10 59.5 10 5.9 NGC4786 12:54:32.42 -06:51:34.1 4647 2012-11-14 45.8 5 6.6 NGC4958 13:05:48.90 -08:01:13.0 1455 2012-11-10 71.7 9 8.0 PGC1021091 13:09:26.99 -07:18:45.0 6723 2012-11-10 67.4 10 6.9 NGC5133 13:24:52.90 -04:04:55.1 6132 2012-11-10 68.3 10 6.9 NGC5228 13:34:35.09 +34:46:41.0 7706 2012-11-10 60.9 10 6.2 IC0948 13:52:26.69 +14:05:28.1 6912 2012-11-14 35.8 10 3.2 NGC5400 14:00:37.23 -02:51:28.1 7437 2012-11-14 37.7 10 4.3 NGC5463 14:06:10.50 +09:21:12.1 7178 2012-11-14 35.7 10 3.8 NGC5623 14:27:08.71 +33:15:07.0 3356 2012-11-14 35.9 10 3.6 NGC5739 14:42:28.89 +41:50:32.1 5377 2012-11-10 65.7 10 6.2 UGC09602 14:55:55.20 +11:51:41.0 9652 2012-11-08 45.1 10 4.7 UGC09937 15:37:22.92 +20:32:58.7 4526 2012-11-08 43.1 10 4.2 UGC10097 15:55:43.30 +47:52:01.9 5962 2012-11-10 70.7 10 7.3 IC1153 15:57:03.02 +48:10:05.9 5919 2012-11-10 69.2 10 7.2 NGC6036 16:04:30.69 +03:52:07.1 5505 2012-11-08 44.9 10 4.6 NGC6146 16:25:10.31 +40:53:34.0 8820 2012-11-10 72.0 10 7.0 NGC6240 16:52:58.90 +02:24:03.0 7465 2012-11-08 43.1 9 5.0 PGC1347752 17:36:11.10 +08:28:56.0 814 2012-11-08 41.4 10 4.4 NGC6508 17:49:46.47 +72:01:16.0 7637 2012-11-10 75.1 10 7.3 UGC11082 18:00:05.50 +26:22:00.0 4739 2012-11-08 39.5 10 4.0 NGC6548 18:05:59.20 +18:35:14.0 2209 2012-11-08 41.0 20 3.1 NGC6619 18:18:55.51 +23:39:20.1 5038 2012-11-14 35.2 10 3.4 PGC062122 18:36:39.70 +19:43:45.0 4840 2012-11-14 35.8 10 3.5 NGC6688 18:40:40.11 +36:17:23.0 5462 2012-11-08 38.5 10 4.1 UGC11353 18:47:44.20 +23:20:49.9 4208 2012-11-14 35.8 10 3.6 NGC6921 20:28:28.80 +25:43:23.9 4337 2012-11-08 39.0 10 4.0 PGC066592 21:20:42.50 +44:23:58.9 3894 2012-11-16 48.5 10 5.5 UGC11920 22:08:27.40 +48:26:27.1 1103 2012-11-16 44.5 10 4.9 NGC7391 22:50:36.10 -01:32:41.0 3048 2012-11-09 43.5 10 4.5 NGC7426 22:56:02.80 +36:21:40.9 5325 2012-11-08 39.6 10 4.3 NGC7436 22:57:57.50 +26:09:00.0 7375 2012-11-08 38.8 10 4.1 IC5285 23:06:58.90 +22:56:10.9 6154 2012-11-16 43.6 10 4.8 NGC7671 23:27:19.30 +12:28:03.0 4128 2012-11-08 41.0 10 4.4 NGC7728 23:40:00.80 +27:08:01.0 9398 2012-11-08 39.5 10 3.9 Table1 Listoftheobserved87HETGBTgalaxiesobservedwiththeGBTtosearchformegamasers.Column(1)HETMGSname,(2,3)J2000position,(4)opticalLSRK velocityusedfortuningthespectralwindows,(5)Observationdate,(6)SystemtemperatureinKelvin,(7)Integrationtimeinminutes,(8)sensitivityinmJy.