MNRAS000,1–10(2016) Preprint31January2017 CompiledusingMNRASLATEXstylefilev3.0 Constraining black hole spins with low-frequency quasi-periodic oscillations in soft states Alessia Franchini1⋆, Sara Elisa Motta2 and Giuseppe Lodato 1 1DipartimentodiFisica,UniversitàDegliStudidiMilano,ViaCeloria,16,Milano,I-20133,Italy 2UniversityofOxford,DepartmentofPhysics,Astrophysics,DenisWilkinsonBuilding,KebleRoad,OX13RH,Oxford,UK 7 1 0 Lastupdated...;inoriginalform... 2 n ABSTRACT a J Black hole X-ray transients show a variety of state transitions during their outburst 0 3 phases,characterizedbychangesintheirspectralandtimingproperties.Inparticular,power densityspectra(PDS)showquasiperiodicoscillations(QPOs)thatcanberelatedtotheac- ] cretionregimeofthesource.Welookedfortype-CQPOsinthedisc-dominatedstate(i.e.the E highsoftstate)andintheultra-luminousstateintheRXTEarchivaldataof12transientblack H hole X-ray binaries known to show QPOs during their outbursts. We detected 6 significant . QPOsinthesoftstate thatcanbeclassified astype-CQPOs. Undertheassumptionthatthe h accretiondiscindisc-dominatedstatesextendsdownorclosetotheinnermoststablecircular p orbit (ISCO) and that type-C QPOs would arise at the inner edge of the accretion flow, we - o use the relativistic precessionmodel(RPM) to place constraintson the blackhole spin. We r wereabletoplacelowerlimitsonthespinvalueforallthe12sourcesofoursamplewhilewe t s couldplacealsoanupperlimitonthespinfor5sources. a [ Keywords: accretion,accretiondiscs–binaries:close–blackholephysics–X-rays:stars 2 v 0 6 1 INTRODUCTION (see,e.g.,Mottaetal.2012).Whilethehardstatesarecharacterized 7 byhighervariability(30−40%rmsintheLHSand5−20%rmsin 1 Blackhole(BH)low-massX-raybinaries(LMXBs)aredoublesys- theintermediatestates),thevariabilityintheULSandintheHSS 0 temsharboringastellarmassblackholethataccretesmatterfroma isverylow(aroundandbelow5%rms,respectively). . Sun-likecompanionstar.ThevastmajorityofBHLMXBsaretran- 1 0 sientsystems,i.e.theyspendmostoftheirlifeinadim,quiescent Thedifferentstatesaredefinedbasedonthespectralproper- 7 state(LX∼1030–1034erg/s),occasionallyinterruptedbybrightout- tiesofthesourceandontheinspectionofthepowerdensityspec- 1 burstsduringwhichtheyshowlargeincreasesinluminosity(reach- tra(PDS),wherechangesinthefasttimevariabilitycanbeclearly v: ing occasionally the Eddington limit) and remarkable variations observed. The most remarkable features detected in the PDS are i in their spectral and fast variability properties (Belloni&Motta narrowpeaksknownasquasi-periodicoscillations(QPOs),whose X 2016). These changes can be described through the Hardness- centroid frequency are thought to be associated with dynamical r IntensityDiagram(HID)(Homanetal.2001;Bellonietal.2011), time scales in the accretion flow. The origin of QPOs is still un- a wherehysteresisloopscanbeidentified(Miyamotoetal.1992)and derdebateandseveralmodelshavebeenproposedtoexplaintheir differentbranchesofaq-shapedtrackroughlycorrespondtodiffer- existence(seeMotta2016forareview).However,itiscommonly entspectral/timingstates.MostactiveBHLMXBsshowfivemain acceptedthatmostQPOsoriginateinthevicinityoftheblackhole, different states:theLow HardState(LHS),dominated byComp- thustheypotentiallycarryinformation onthepropertiesand mo- tonized emission; the High Soft State (HSS), where the thermal tionofmatterinastronggravitationalfield. emissionfromanaccretiondiscdominatesthespectrum;theHard IntermediateState(HIMS)andtheSoftIntermediateState(SIMS), QPOs are usually divided in two groups based on their fre- wherethespectrumshowsthecontributionsfromboththeaccretion quency: low frequency and high frequency (LFQPOs and HFQ- discandtheComptonizedemission,butwherethemostdramatic POs, respectively). LFQPOs have frequencies typically varying changesinthetimingpropertiesareobserved.Afewsourcehave fromafewmHzupto∼ 30Hzandarefurtherdividedintothree alsoshownananomalousorultra-luminousstate(ULS),whichcan types-type-A,type-Bandtype-C-basedontheirproperties(e.g. beseenasasoft-intermediateandhard-intermediatestateinwhich peakfrequency,width,energydependenceandphaselags)andon theluminosityissignificantlyhigherwithrespecttotheotherstates theassociatedbroadbandnoiseinthePDS(shapeandtotalvari- abilitylevel,Wijnandsetal.1999,Casellaetal.2005,Mottaetal. 2011).AccordingtotheclassificationofHoman&Belloni(2005), ⋆ [email protected] Type-A QPOs are commonly found in the SIMS and are charac- (cid:13)c 2016TheAuthors 2 Franchini,MottaandLodato terized by a weak and broad peak around 6−8 Hz. 1 Their ori- able to rigidly precess around the compact object spin axis. Ac- gin is still under debate. Type-B are usually seen in the SIMS cording to this model type-C QPOs are produced by the Lense- (whichisdefinedbytheirpresence,Belloni&Motta2016).They Thirringprecessionofaradiallyextendedportionofthehotinner arecharacterizedbyarelativelystrongandnarrowpeakaround6 accretion flow that produces a modulation of the X-ray flux, dif- Hz(Mottaetal.2011)andithasbeensuggestedthattheyareas- ferentlyfromtheRPMwhereatestparticleisresponsibleforthe sociated with relativistic jets during the transition from the hard QPO. Inthe first case, the QPO frequency isa weighted average tothe soft state(Fenderetal.2009).Type-C QPOsarefound es- ofthefreeprecessionfrequenciesatthevariousradii.However,for sentially in any spectra/timing state (including the ULS, Motta the disc to precess rigidly, its radial extent must be narrow (tens 2016). They are characterized by a strong (up to 20% rms), nar- ofR )(Lodato&Facchini2013;Franchinietal.2016;Nixonetal. g row(ν/∆ν > 3,butoftenmuchlarger)andvariablepeak(itsfre- 2016).Inparticular,intheHSS,wheretheaccretiondiscinnerra- quencyandstrengthvaryingbyfewordersofmagnitudealongan dius is expected to be close to or coincident with the innermost outburst;see,e.g.,Mottaetal.2015)superposedonaflat-topnoise stable circular orbit (ISCO, e.g. Dunnetal. 2010), the inner hot that steepens above a frequency comparable to the centroid fre- (geometrically thick) accretion flow is expected to be extremely quencyoftheQPO.Type-CQPOsappearatlowfrequencies(from narrow (itsouter radiusdetermined bytheinner thindisctrunca- afewmHzuptotenthsofHzMottaetal.2011)inthehardstate tionradius).Therefore,itsprecessionfrequencyisreasonablywell where the disc is thought to be truncated at a large radius (tens describedbythatofatestparticleorbitingattheISCO.Thus,inthe tohundredsofR ).Theyincreasetheirfrequencyduringthetran- HSSandinparticularclosetotheISCO,theRPMandtherigidpre- g sition from the hard to the soft state, reaching their maximum in cessionarepracticallycoincident andtheRPMrepresentsagood theHSS,wheretheyshowfrequenciesaround30Hz(Mottaetal. approximationoftherigidprecession(Ingram&Motta2014). 2014a,Mottaetal.2014b).HFQPOsarefairlyrareinBHbinaries Themeasurementoftheblackholeparametersisamajoris- (whiletheyarecommonlyobservedinNSsystems,wheretheyare sueinastrophysics.Whilethemasscanbeestimatedthroughoften referredtoaskHzQPOs,seevanderKlis2006)andhavebeende- complexdynamicalstudies,thespincanonlybeinferredinanin- tected only inafew sources (Bellonietal. 2012). HFQPOsseem directway.IntheHSSmeasurementsofthespinscanbeobtained to appear preferentially in the HSS and in the ULS, even though through X-ray spectroscopy by modelling the thermal disc emis- observational biases could be at play because of the low number sion in the spectrum (McClintocketal. 2011, 2014; Steineretal. ofdetectionscurrentlyavailable.Remillard&McClintock(2006), 2010, 2011, 2012). In the hard and intermediate states, the spin usingadifferentclassification(Table1inMcClintocketal.2009), isinstead measured bymodelling thediscreflectionspectrum, in showedthatmostHFQPOshavebeendetectedinobservationswith particularbymodellingtheeffectsofthespinontheIron-Kαline asubstantialpower-lawcontribution. (Miller2007;Reynolds2013).Bothmethodsrelyuponidentifying There are different models that attempt to describe the ori- theinnerradiusoftheaccretiondiscwiththeISCO.Inaddition,for thespectralcontinuummethod,onemustalsoknowthemassofthe ginofQPOsinLMXBs.Therelativisticprecessionmodel(RPM) blackhole,theinclinationoftheaccretiondisc(generallyassumed is one of those. This model has been originally proposed by equal to the inclination of the binary system) with respect to the Stella&Vietri(1998);Stellaetal.(1999) andwasrecentlyrevis- lineofsightandthedistancetothebinary.Forthesereasons,such itedbyMottaetal.(2014a,b);Ingram&Motta(2014),toexplain methodsareoftenaffectedbylargesystematicerrorsandinafew threetypesofQPOs:type-CQPOs,thelowerandtheupperHFQ- POs.IntheRPM,thecentroidfrequencyofthreedifferenttypesof casestheyhavelead toinconsistent values of theblackhole spin whenusedonthesamesource(Shafeeetal.2006).TheRPM,on QPOsisassociatedwiththefrequenciesofatestparticleorbitinga theotherhand,isbasedonlyonmeasurementsthatcanbeinferred spinningblackhole,aspredictedbythetheoryofGeneralRelativ- throughverypreciseX-raytimingmeasurements(i.e.theprecision ity.TheLense-Thirring(LT)frequencyisassociatedwiththetype- of the measurement of a QPO centroid frequency is only limited CQPO,theperiastronprecessionfrequencywiththelowerHFQPO bythetimeresolutionofthedataused,whichisusuallyextremely andtheorbitalmotioncorrespondstotheupperHFQPO.TheRPM high).Therefore,theRPMcanbeusedtoestimateBHspinsinde- predictsthattheexpectedvaluesofthesefrequenciesdependsolely pendentlyfromothermethods. onthefundamental parametersoftheBH,i.e.massandspin,and InthisworkweapplytheRPMtoQPOsdetectedintheHSS, on the radius at which the particle motion occurs (i.e. where the i.e.producedcloseorattheISCO,placinglimitsontheblackhole QPOsareproduced). Basedon thisfact, once atype-C QPO and spinofanumberofsourcesbymakingreasonableassumptionson two HFQPOsare observed simultaneously, one can obtain apre- themass,whenthisisnotknownapriori. cisemeasurementofthefundamentalparametersoftheblackhole, spinandmass(Mottaetal.2014a).Aspinmeasurementcanalsobe obtainedwhenonlytwosimultaneousQPOsoftherelevanttypes aredetectedandwhenanindependent massmeasurement,forin- 2 OBSERVATIONSANDDATAANALYSIS stancedetermineddynamically,isavailable(Mottaetal.2014b).If themass isunknown, theRPMstillallowstoplace limitson the Based on Mottaetal. (2015) we analyzed RXTE/Proportional massandspinoftheblackhole(seeIngram&Motta2014). Counter Array(PCA)archival observations of12transient BHBs Ingrametal. (2009) proposed a model based on relativis- showing QPOs. We selected those sources that show a standard tic precession that requires a geometrically thin accretion disc HSSamongthe14sourcesconsideredbyMottaetal.(2015).We (Shakura&Sunyaev1973)truncatedatsomeradius(theinnerdisc did also include observations in the ULS for those sources that truncationradius)filledinwithaninnerhotaccretionflowthatis showed thiskind of stateduring their outbursts. Inthiswork, we excludedSwiftJ1753.5-0127 sinceithasnotshownaHSSinthe RXTEera(Solerietal.2013),andXTEJ1720-318, whichdidnot 1 However, we note that more recent works (Bellonietal. 2011; Motta showQPOsduringtheoutburstobservedbyRXTE.Welookedfor 2016;Belloni&Motta2016)placedthistypeofQPOsintheHSS,since QPOsintheHSSandintheULS,wheretheyarethoughttobepro- theytypicallyappearattotalfractionalrmsbelow5%inBHBs. duced at or closeto theISCO (seeMottaetal. 2012, Mottaetal. MNRAS000,1–10(2016) Type-C QPOs insoftstate 3 2015). We analyzed all the observations in the RXTE catalogue where r is the radius (in units of the gravitational radius R = g in the HSS and ULS that have not been included in Mottaetal. GM/c2)atwhichthefrequenciesareproduced,aisthedimension- (2014a,b,2015).ThenweaddedalsotheQPOsfoundintheprevi- lessblackholespinparameter,and Mistheblackholemass.The ousworks. signs plus and minus refer to prograde and retrograde orbits, re- WecomputedthePDSusingacustomsoftwareunderIDLin spectively. the total and hard energy band (∼ 2−37 keV, absolute channels According to the RPM, type-C QPOs are associated with 0-64, and ∼ 6−37 keV, absolute channels 14-64, respectively). the Lense-Thirring precession frequency ν , the lower HFQPO LT SinceQPOstypicallyshowahardspectrum(Sobolewska&Z˙ycki totheperiastronprecessionfrequencyν andtheupper HFQPO per 2006),producingthePDSusingonlyhardphotonsmaximizesthe to the orbital frequency ν . Under the hypothesis that when ob- φ chancesofdetectingafaintsignal,whoseintrinsicvariabilitycould servedsimultaneously,thesefrequenciesareproducedatthesame bedilutedbythenon-variablephotonsfromtheaccretiondisc,that radius r, the equations above - linking the black hole parameters dominatesinthesoftstatesandemitsmostlybelow6keV.Forsim- (spin and mass) to the QPO centroid frequencies - form a sys- ilarreasons,wedidnotconsiderthephotonsabovechannel64:the temofthreeequationsinsixvariables:(a,M,r,ν ,ν ,ν ).There- φ per LT efficiencyof thePCAdrops dramaticallyabove∼30keV andthe fore, when a type-C QPO and two HFQPOs are detected simul- variabilitycomingabovethisenergyismostlyPoissonnoise.Inad- taneously, one can solve the RPM system of equations exactly dition,intheHSStheenergyspectrumistypicallyverysteep,with (Ingram&Motta 2014), obtaining precise values for a, M and r verylittleornosignalabove∼20keV.Inafewcaseswecouldnot (Mottaetal.2014b,a). producePDSinthehardbandsincethedatamodesdidnotallowit, Inthisworkweassumethatthehighestfrequencytype-CQPO thereforeweonlyanalyzedthetotalenergybandPDS(channel0- detected in the HSS or ULS for each source corresponds to the 64ifavailable,otherwisechannels0-249).Weused16s-longinter- Lense-ThirringprecessionfrequencyofatestparticleattheISCO. valsandaNyquistfrequencyof2048HztoproducethePDS,which TheradiusoftheISCO(r ),aspredictedbythetheoryofGen- ISCO werethennormalizedfollowingLeahyetal.(1983)andconverted eralRelativity,isonlyafunctionof theblackholedimensionless tosquarefractionalrms(Belloni&Hasinger1990).Wemeasured spinparameter(see,e.g.,Stella&Vietri1998).Therefore,substi- theintegratedfractionalrms(i.e.thermsfromtheentirePDS)in tutingthe expression of r into theequations above, allows us ISCO the2-15keVbandandbetween0.1and64Hz.Fromnowon,ifnot toremovethervariablefromtheRPMsystemandthustoobtain otherwisestated,byrmswerefertotheintegratedfractionalrms. equationsthatonlydependonManda.Thisimpliesthat,ifeither Then,weinspectedallthePDSselectingonlythosewherewe themassorthespinareknown,onecanobtainanestimateofthe detectedoneormorenarrowpeaksandfittedthemusingthestan- otherparameterdetectingjustoneQPOandusingtherelatedRPM dard XSPEC fitting package with a one-to-one energy-frequency equation. While BH spin measurements can be obtained only in- conversion and a unity response matrix. Following Bellonietal. directlyandareoftenaffectedbylargeuncertainties, BHmasses, (2002), we fitted the noise components with a number of broad instead, canbeinferredthroughdynamical studies. Theliterature Lorentzians and the hints of QPOswith narrow Lorentzians 2. A includestodatemassmeasurementsforafewtensofcompactob- constant component wasadded totake intoaccount thecontribu- jectinbinarysystems,includingseveralstellarmassBHs.Wecan tionofthePoissonnoise.Weestimatedthestatisticalsignificance thusmake reasonableassumptions onthemassrangespanned by ofthesepeaksandweexcludedallfeatureswithsingletrialsignif- theBHshostedinthesystemsofoursample.Thismeansthat,even icancebelow3σ.Thenweconsideredalsothenumberoftrialsin without a dynamically determined BH mass, we are now able to thecalculationof thesignificanceand wediscarded allthepeaks placelimitsontheblackholespinusingtheRPM.Inprinciple,this thatresultednotsignificantafterthisanalysis.Werejectedalsothe couldbedonewithanyQPOrelevantfortheRPM(type-CQPO, peakswithqualityfactor(i.e.theratiobetweencentroidfrequency upper or lower HFQPO).However, type-C QPOsare much more andFWHMoftheQPOpeak)Q<2.SinceintheHSSbothtype-C commonandeasiertodetectthanHFQPOsandcanbedetectedin QPOsandHFQPOscanbeseen,weclassifiedtheQPOswefound basicallyallspectral/timingstates(Motta2016),includingtheHSS followingMottaetal.(2011)andMottaetal.(2012). andULS. In order to place limits on the spin in those cases where a dynamical BH mass is not available, we choose a conserva- 3 THERELATIVISTICPRECESSIONMODEL: tive range of masses, based on the mass distribution of BHBs CONSTRAININGTHESPIN (Casares&Jonker2014):3−20M .Thesmallestdynamicallycon- ⊙ firmedstellarmassblackholeisGROJ1655-40,withM =5.4M Thefundamentalfrequenciesofaparticleorbitingaspinningblack ⊙ (Beer&Podsiadlowski 2002).Inasimilarway,theupper limitis holearegivenby(fromlowesttohighestfrequency): based on the fact that the heaviest dynamically confirmed stellar mass black hole in a Galactic binary system is the one in GRS νLT =νφ(cid:20)1−(cid:16)1∓4ar−3/2+3a2r−2(cid:17)1/2(cid:21) (1) J1915+105(Reidetal.2014),withablackholeof12.4M⊙.There isnoevidenceforheavierstellarmassblackholes(eventhougha binarywithtwo30M BHshasbeenrecentlydetectedAbbottetal. ⊙ ν =ν 1− 1−6r−1±8ar−3/2−3a2r−2 1/2 (2) 2016),thereforeweassumedaconservativeupperlimitonthemass per φ(cid:20) (cid:16) (cid:17) (cid:21) of20M . ⊙ We explore the range between 3 and ∼ 150 Hz for the LT c3 1 ν =± (3) frequency. The lower limit is chosen based on the observational φ 2πGMr3/2±a evidencethatallBHBsshowtype-CQPOsreachingatleast3Hz (Mottaetal. 2015). Even if type-C QPOs are typically observed 2 Type-CQPOssometimesshowsignificantharmoniccontentespeciallyin below∼30Hz,wedecidedtoconservativelyexploreawiderrange thehardstates.However,asingleLorentzianisalwaysenoughtofittype-C of frequencies. Indeed, it is worth noticing that a type-C QPO QPOsintheHSS. produced around an almost maximally spinning black hole (e.g. MNRAS000,1–10(2016) 4 Franchini,MottaandLodato 110 20 20 18 18 140 16 147 16 170 14 14 200 ]⊙ 12 138 ]⊙ 12 230 260 290 320 350 M 3 M M[ 10 129 M[ 10 8 120 8 111 6 93 102 6 4 12 21 30 39 48 57 66 75 84 4 0.0 0.2 0.4 0.6 0.8 0.2 0.4 0.6 0.8 a a Figure 1. Couples of mass-spin values that give the LT precession fre- Figure2.Couplesofmass-spinvaluesthatgivetheperiastronprecession quency.ThenumbersarethefrequenciesevaluatedinHz.Eachlineisob- frequency. The exactly sameplot is obtained considering the orbital fre- tainednumericallysolvingeq.1assumingafrequencyrangefrom3to147 quencyattheISCO.Indeed,attheISCOthereisnoradialoscillationinthe Hz. frequency,thusνper=νφ.ThenumbersarethefrequenciesevaluatedinHz. Eachlineisobtainednumericallysolvingeq.2assumingafrequencyrange from50to350Hz. a = 0.98 obtained for GRO J1655-40 using spectroscopic meth- ods) should have a frequency above 100 Hz. Sinceinprinciple a type-CQPOthesehighcouldbedetected,wechooseawiderrange types of QPOs, which in a rms vs frequency plot are known to offrequenciestoexplorethispossibilityandinordertoguarantee generallyformdifferent,welldefinedgroups(seee.g.Casellaetal. anunbiasedanalysis. 2005,Kalamkaretal.2011,Mottaetal.2012).Inparticular,type- GRSJ1915+105showsHFQPOsaround67Hz(Morganetal. C QPOs are known to correlate well with the rms, forming in a (1997)) while the other sources show this type of QPOs above frequencyversusrmsplaneacurvedtrackwithrmsdecreasingfor 150 Hz.Therefore, for periastronprecession frequencies (and or- increasing frequency, breaking at νBreak ∼10 Hz. The slope is al- bitalfrequency,coincidentwiththeperiastronprecessionfrequency ways steeper for the high-frequency half of the correlation track at the ISCO) we assume a lower limit of 50 Hz and an upper (see,e.g.,Mottaetal.2012forthecaseofGROJ1655-40).HFQ- limit of 350 Hz since this isthe highest observed HFQPO so far POs,instead,formadifferentgroupofpointsatsignificantlyhigher (Strohmayer 2001; Bellonietal. 2012). We stress that, according frequenciesandrmscomparabletothatoftype-CQPOsfoundin to the RPM, it is possible to distinguish type-C QPOs from the thesoftstate(Mottaetal.2016,inprep). HFQPOassociatedtoeitherorbitalorperiastronmotionsincethere InFig.3,weplottheintegratedfractionalrmsversusthecen- isnomass-spincombinationthatgivesthesamefrequencyforboth troidfrequencyoftheQPOsforeachsourceofoursample.Wesee motions. thatformostofthesourcestheQPOcentroidfrequencycorrelates InFigure1weshowtheLense-Thirringprecessionfrequency wellwiththerms,formingacurvedtrack.Intheplotwemarkwith ν evaluated at r for every mass and spin couple within the blackdotstheQPOsreportedinMottaetal.(2012,2014a,b,2015) LT ISCO ranges3M < M < 20M and0<a<1(i.e.weconsideredonly andwithredstarsthetype-CQPOsfoundinthiswork. ⊙ ⊙ progradeorbits).NotethatforclarityweonlyshowLense-Thirring frequenciesspacedby9Hz.Forcomparison,wealsoshowinFig. 4.2 EvidenceofQPOsinthesoftstate 2theperiastronprecessionfrequenciesatr withthesamespin ISCO andmassranges. Ithasbeenshownthatfortwosourcesinoursample,GROJ1655- 40 and XTE J1550-564 (see Saitoetal. 2007 for GRO J1655-40 Inordertoplacelimitsonthespinvalue,weselectedthehigh- andKubota&Done2004forXTEJ1550-564), theaccretiondisc est frequency type-C QPO for each source (considering both the inner radius remains constant for most of the HSS, providing a QPOswefoundinthisworkandtheQPOsreportedinMottaetal. strong evidence for the accretion disc having reached the ISCO 2012, 2014a,b, 2015) and then solved eq. (1) for the spin a after (seealsoSteineretal.2011,Mottaetal.2014a,Plantetal.2014). substitutingtheexpressionofrISCO.Weusedeitheraknownvalue ItisworthnoticingthattheQPOsintheHSSofthesetwosources forM(whenavailableintheliterature),orweassumedthemassto stand out in terms of very low rms and high Q factor (see Table lieintherangegivenabove. 1).Giventhebehavior of thesourcesof our sample(andof tran- sientBHX-raybinariesingeneral)intheHIDs,itisreasonableto assumethattheinnerdiscradiusdoesindeedreachtheISCOany 4 RESULTS timeasourceisobservedintheHSS,whichisbasicallythesame assumptionusedforthespinmeasurementsbasedonspectroscopy 4.1 QPOClassification (seee.g.McClintocketal.2014andreferencestherein).Sincethe Plottingthetotalintegratedfractionalrmsversusthecentroidfre- type-C QPOsdetectedintheULShavefrequencies verycloseto quency of a QPO is a useful method for distinguishing different thoseobservedintheHSS(seee.g.Mottaetal.2012),weextended MNRAS000,1–10(2016) Type-C QPOs insoftstate 5 100 4U1543-47 100 4U1630-47 100 GROJ1655-40 100 GX339-4 Mottaetal.(2015) Type-C(thiswork) 10−1 10−1 10−1 10−1 10−2 10−2 10−2 10−2 100 101 102 10−1 100 101 102 10−1 100 101 102 10−1 100 101 102 10−1 XTEJ1817-330 100 XTEJ1550-564 100 H1743-322 100 XTEJ1859+226 ms r al n o dfracti 10−1 10−1 10−1 e grat e int10−2100 101 102 10−210−1 100 101 102 10−210−1 100 101 10−2100 101 100 XTEJ1650-500 100 XTEJ1748-288 100 XTEJ1752-223 100 MAXI1543-564 10−1 10−1 10−2 10−1 10−1 100 101 101 102 100 101 100 101 centroidfrequency(Hz) Figure3.IntegratedfractionalrmsversusQPOcentroidfrequency.EachpointcorrespondstoasingleRXTEobservation.Eachpanelreferstoasinglesource ofoursample. thisassumptionalsototheULS.Howeverwedecidednottoplace obtained applying the RPMto our sample in Table 1. In the first upper limitsusing type-C QPOs in the ULS since they might be column are listed the source names and the second contains the underestimated.WedescribebelowinSection4.3theuncertainty Obs-IDofthehighestdetectedtype-CQPO.Thethird,fourthand inthespinestimateproducedbyusingthisassumption. fifthcolumnscontainthehighesttype-CQPOfrequencyandqual- If the highest type-C QPO is detected in the HSS, we can ityfactorandthetotalfractionalrmsrespectively.Thesixthcolumn thereforeassumethatitwasproducedattheISCOandthusplace indicateswhetherthehighest type-CQPOwasdetectedinthisor a lower and an upper limit on the spin assuming a lower and an inpreviousworks(Mottaetal.2012,2014a,b,2015;Homanetal. upper limit for themass. For those sources for which wedid not 2005).Thelasttwocolumnscontainthestateinwhichthehighest detect a type-C QPO in the soft state, we can still use the high- type-CQPOwasdetectedandthespinlimitsobtainedrespectively. estfrequencytype-CQPOfromMottaetal.(2012,2014a,b,2015) Wewouldliketostressthatweperformedasystematicsearch toplaceaconservativelowerlimitassumingalowerlimitforthe fortype-CQPOsintheHSSandULS.Forthosesourcesforwhich masssincetheseQPOsprobablycorrespondedtoradiilargerthan wedidnotdetectanytype-CQPOinthesestatesweusedthetype- theISCO.Inthesecasesanupperlimitonthespinwouldbemean- CQPOsreportedinMottaetal.(2015).Forcompletenesswewere ingless(andinfactnotcorrect)givenourassumptions. madeawareofthefactthattherearetwoQPOsatevenhigherfre- Fig.4containstheHIDsofthe5sourcesthatshowedtype-C quencies missing in Mottaetal. (2015) for XTE J1859+226 and QPOsinthe HSS.Thered dots represent theobservations of the XTE J1650-500. Thus our lower limitsfor these two sources are newly detected QPOs. By looking at these diagrams and the rms slightlyunderestimated.WedecidednottoaddtheseQPOsinour valuesintableA1itisclearthatthetype-CQPOsinGX339-4,4U worksincewedonotwanttobiastheanalysisperformed. 1543-47andXTEJ1817-330 weredetectedinsoftstates.Instead Formostofthesourcesofoursamplethereisnoindependent thetype-CQPOsinGROJ1655-40and4U1630-47havebeende- massmeasurement, thusweassumed themasstolieintherange tectedintheULS.Wecanreasonablyassumethatalsointhisstate 3−20M .Inafewcases,however,wehavebetterconstraintson theaccretiondiscinnerradiusreachestheISCOsincethefrequen- ⊙ the black hole mass and thus on the spin value. The lower limit cies of type-C QPOs in the ULS are similar to those detected in for the black hole mass in GX 339-4 is 6M (Hynesetal. 2003; theHSS.Inthe4U1543-47HIDwemarkedwithreddotsthetwo ⊙ Muñoz-Dariasetal.2008).Thuswecanplaceastricterlowerlimit differenttype-CQPOswedetectedthatslightlyoverlapwitheach onthespinvalue:a =0.16.SinceweareassumingBHmasses other. ThustheseHIDscontain allthe6new QPOsfound inthis lower up to20M , theupper limitfor thespin of GX 339-4 isa = work. ⊙ upper 0.38. ForbothGROJ1655-40(5.31±0.07M Mottaetal.2014a) ⊙ and XTE J1550-564 (9.1±0.6M Oroszetal. 2011) we have a 4.3 Constrainingthespin ⊙ goodindependentmassmeasurement.Usingthesedynamicallyes- WeplacedlowerlimitsontheBHspinforeachsourceofoursam- timated mass values, we can narrow significantly the spin range pleasdescribedinSection3.Forthesourceswiththehighesttype- ofvaluesforthetwosources(seeTable1).Thespinvaluesfound CQPOdetectedinthesoftstatewedidalsoplaceanupperlimit by Mottaetal. (2014a,b) (a = 0.290±0.003 for GRO J1655-40 (assuminganupper limitonthemass).Wesummarizetheresults anda=0.34±0.01forXTEJ1550-564)are,asexpected,ingood MNRAS000,1–10(2016) 6 Franchini,MottaandLodato 104 GX339-4 0.8 • 0.7 s) 103 cts/ 0.6 E( 102 0.5 T A R 101 a 0.4 T / N a U ∆ 0.3 O 100 C 0.2 10−1 0.0 0.2 0.4 0.6 0.8 1.0 1.2 0.1 HARDNESSRATIO 0.0 104 4U1630-47 1.0 1.2 1.4 1.6 1.8 2.0 rout/rin • s/s)103 (ct Figure5.Relative changeinthespinestimate assumingthattheQPOis E AT102 producedbyanextendedrigidlyprecessingstructure(Ingrametal.2009) R ratherthanatestparticleatISCO,asafunctionoftheradialextentofthe T UN101 precessingstructurerout/rin,whereroutistheouterradiusandrin =rISCO O istheinnerradiusofthestructure. C 100 0.0 0.2 0.4 0.6 0.8 1.0 1.2 HARDNESSRATIO agreementwiththerangeweinferredusingtheRPMwiththehigh- esttype-CQPOattheISCO. 104 4U1543-47 Asmentionedabove,theremightstillbethepossibilitythat, ••• eveninthesoftstate,thediscdoesnotextenddowntotheISCO, s/s) 103 butsomewhatclosetoit.Inthiscase,onecanstillrelatetheQPO E(ct 102 frequency to the spin of the black hole, because we expect that AT the hot inner flow would have a finite and small radial extent R 101 so that it would precess as a rigid body (Ingram&Done 2011; T N U Franchinietal.2016).Thus,wecanassumethatthehotinnerflow O 100 C extendsfromtheISCOtoanouter radiusR ,andthencompute out 10−1 thespinvaluethatwouldreproduceagivenQPOfrequencyusing, 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 forexampleeq.(1)inIngram&Done(2011).Inthisway,wecan HARDNESSRATIO thusestimatetheassociatedchangetothespinvalueobtainedfrom 104 XTEJ1817-330 theRPMmodel,asafunctionofRout.Forinstance,usingthehigh- esttype-CQPOdetectedinGROJ1655-40(i.e.27.51Hz)together s/s)103 • withthemassestimateforthissource,wecanevaluatetheuncer- ct taintywithrespecttothevalueobtainedusingtheRPM.Theresults ( E AT102 inthiscaseareshowninFigure5.The x-axisistheratiobetween R theouterradiusoftheinner accretionflowandtheISCOandthe T UN101 y-axisistherelativechangeinthespinvalue.Asimilarresultholds O C alsofordifferentQPOfrequencies. 100 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 HARDNESSRATIO 5 DISCUSSION 105 GROJ1655-40 We analyzed RXTE observations of 12 black hole transients that s)104 • showed a transition to the soft state. We considered observations cts/ in the HSS and ULS of these sources. We then assumed that the E(103 highestfrequencytype-CQPOhasbeenproducedattheISCOand T RA102 weappliedtheRPMinordertoplacelimitsonthespinvalue.This NT isdone either using aknown value of theblack hole mass or as- U O101 sumingthemassinareasonablerangeofvalues,decidedapriori C (3−20M inourcase). 100 ⊙ 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 This method is based on the assumption that the inner disc HARDNESSRATIO truncation radius is constant and coincident with ISCO when a BHB is observed in the HSS or ULS. Demonstrating the coinci- dence of the inner disc truncation radius with ISCO is problem- Figure4.HIDofthe5sourceswithtype-CQPOsintheHSS(GX339-4, atic and involves spectroscopic measurements that are generally 4U1630-47,4U1543-47,XTEJ1817-330,GROJ1655-40).Thereddots representtheobservations thatcontainthetype-CQPOswefoundinthis affected by largeuncertainties. Therefore, our assumption consti- work.NotethattheQPOsinGX339-4,4U1543-47andXTEJ1817-330 tutesarelevantcaveat.However,bothobservationsandsimulations weredetectedinsoftstates,whiletheonesinGROJ1655-40and4U1630- providestrongevidenceaboutthestabilityoftheinnerdiscradius 47weredetectedintheULS.IntheHIDof4U1543-47wemarkedthetwo overtimeindisc-dominatedstatesofBHBs(e.g.McClintocketal. differenttype-CQPOswefoundwithreddots. MNRAS000,1–10(2016) Type-C QPOs insoftstate 7 Table1.Spinlimitsinferredfromthehighestfrequencytype-CQPOatISCOaccordingtotheRPM.Bothlimitsareevaluatedassumingablackholemass intherange3−20M⊙.InthecaseofGX339-4thelowerlimitcorrespondsto6M⊙.ForXTEJ1550-564andGROJ1655-40thelimitscorrespondtothe measuredmasses9.1±0.6M⊙ and5.31±0.07M⊙ respectively(MuñozDariasetal.2010;Mottaetal.2014b,a).Thefirstcolumncontainsthenameofthe source,thesecondistheObs-IDofthehighesttype-CQPOdetected.Thethird,fourthandfifthcolumnscontainthepropertiesoftheQPO,i.e.frequencyand theQfactor,andthetotalfractionalrms.Thesixthcolumnindicateswhetherthehighesttype-CQPOhasbeendetectedinthisworkorthepreviousones.The seventhcolumnindicatesthestateinwhichtheQPOwasdetectedandthelastcolumncontainsthespinlimits. Target Obs-ID νmax(Hz) Q rms New? state spinlimits GX339-4 92085-01-02-03 10.59±0.18 3.46±0.50 5.05±0.03 yes HSS 0.16-0.38 4U1630-47 80117-01-07-01 14.80±0.28 2.17±0.27 9.5±0.4 no ULS >0.12 4U1543-47 70133-01-01-00 15.37±0.18 2.57±0.27 4.2±0.03 yes HSS 0.13-0.47 XTEJ1859+226 40124-01-14-00 8.56±0.06 2.76±0.18 12.4±0.4 no HIMS >0.07 XTEJ1650-500 60113-01-13-02 6.84±0.05 8.05±1.23 20.8±0.3 no HIMS >0.06 XTEJ1817-330 91110-02-32-00 9.6±0.5 2.93±0.82 5.7±0.1 yes HSS 0.08-0.36 XTEJ1748-288 30171-02-01-00 31.55±0.13 6.00±0.42 10.2±0.5 no HIMS >0.23 XTEJ1752-223 95360-01-11-00 6.46±0.13 3.6±1.1 23.4±1.2 no HIMS >0.06 XTEJ1550-564 40401-01-48-00 18.10±0.06 19±4 6.2±0.1 no HSS 0.31-0.34 MAXIJ1543-564 96371-02-02-01 5.72±0.04 10.2±1.4 27.3±1.2 no HIMS >0.05 H1743-322 80135-02-03-00 14.6±0.2 12.6±6.3 4.4±0.2 no ULS >0.12 GROJ1655-40 91702-01-17-01 27.51±0.13 44±21 2.7±0.1 no HSS 0.29-0.31 2014).Tanaka&Lewin(1995)showedthattheinnerradiusisre- frequencyforagivensourceresultsinanunderestimationofboth markablystablealsoifthethermalfluxofthesourcesteadilyde- upper and lower limit of the spin. In other words, when estimat- creasesontimescalesofmonthsbyoneortwoordersofmagnitude. ing the spin range from a type-C QPO that is not the highest in Similar evidence for a constant inner disc radius in the soft state frequency, wewould befollowingthe wrongtrackinFig.1,that hasbeendemonstratedformanysources(Doneetal.2007andref- wouldbefoundtotheleftofthetrackcorrespondingtotheactual erences therein,Plantetal. 2014,Plantetal. 2015).According to highestfrequencytype-CQPO.ThiseffectislargerforQPOsthat Tanaka&Lewin(1995),theconstancyoftheinnerradiussuggests arenotdetectedintheHSS,thereforeproducedataradiusthatis that it is related to the ISCO. More recently Steineretal. (2010) almostcertainlylargerthanISCO. foundanotherevidenceofanonvaryingdiscinnerradiusovertime We note that the different tracks in Fig. 1 tend to lay closer forthesourceLMCX-3,unaffectedbythehighvariabilityofthis to each other for high frequencies (i.e. above 80 Hz). Thus, the source.Nevertheless,sincewecannotbecertainofthecoincidence erroronthespinestimatemightbedominatedbythatonthemass oftheinnerdiscradiuswithISCOforallthesourcesofoursample, iftheQPOfrequencyisveryhigh.Ingeneral,thetwouncertainty excludedGROJ1655-40andXTEJ1550-564(forwhichthecoin- thatcontributetotheerroronthespinvaluearethatontheblack cidencehasbeendemonstratedexplicitly),weadvicethereaderto hole mass and the identification of the highest type-C QPO. For considertheinferredlimitswithcaution. instance,iftheactualhighesttype-CQPOwasproducedat,say,30 The rigid precession of the inner accretion flow as an ex- Hzinsteadofbeingdetectedat21Hz,theerrorthatwecommiton planation of type-C QPOs has been supported by several stud- thespinestimateisoftheorderofafewpercent. ies(Ingrametal.2009;Ingram&Motta2014;Ingrametal.2016). The rigid precession model tends asymptotically to the RPM for radii approaching the ISCO, therefore, given the stability of the 5.1 Comparisonwithothermethods discinner radiusinthesoftstate,theRPMrepresentsagoodap- Forsomesources of our sample thespinhasbeen alsomeasured proximationoftherigidprecessionfortruncationradii(hencehot throughX-rayspectroscopybyfittingthethermaldiscemissionor flowouterradii)veryclosetotheISCO(Ingrametal.2009). by modelling the disc reflection spectrum. Fig. 6 shows the spin The approach we used in this work differs from that of measurements found in the literature obtained from X-ray spec- Ingram&Motta (2014), which always requires that at least two troscopy versus the spin values we inferred from timing analysis QPOs are detected simultaneously (the type-C QPO and one in this work. The black line marks the position where the mea- HFQPO)inordertoplaceanupperorlowerlimitonspinandmass, surementsshouldfalliftherewasperfectagreementbetweenspec- dependingonthecaseconsidered.Hereweassumedthatthehigh- troscopybasedmeasurementsandtimingbasedmeasurements.In estfrequencytype-CQPOisproducedattheISCO.ThustheRPM twooutoffivesources(XTEJ1550-564and4U1543-47)thespin systemofequationsisreducedtojustoneequation(i.e.eq.(1))that measurements obtained with spectroscopy and timing do agree. giveseitheralowerlimit(ifthehighest type-C QPOwasnotde- In particular, Morningstar&Miller (2014) obtained a spin value tectedintheHSS)orarangeofspinvalues.Therangesarebroadif a = 0.43+0.22, in agreement with our values, for 4U 1543-47 us- −0.31 thereisnoindependent(e.g.dynamical)massmeasurementavail- ing both the spectral continuum fitting method and the iron line able,orsignificantlynarrowerifthemassisknown.Thereasonis fittingmethodsimultaneously.Steineretal.(2011)useddatafrom thattheuncertaintyonthespiniscompletelydominatedbytheer- theBHBXTEJ1550-564tomodelthecontinuumandtofittheiron ror on themass (much larger thanthe error on theQPO centroid lineobtainingtwodifferentspinmeasurements,thatcombinedre- frequency). sultedinaspinofa=0.49+0.13whichismarginallyconsistentwith −0.20 Inprinciplewecanneverbecompletelysurethatthetype-C thespinrangethatwereport(seeTable1). QPOweassumedtobethehighestisactuallythehighestproduced For4U1630-47Kingetal.(2014)measuredthespinbyfitting bythesourceinthesoftstate.Thereisalwaysthepossibilitythat thecontinuum spectrum of thesource, obtaining a = 0.985+0.005, −0.014 anevenhigherfrequencytype-CQPOhasbeenproducedandnot which is in principle compatible with our lower limit (i.e. 0.12). detected. Using a type-C QPO that is not at the highest possible For XTE J1752-223 and GX 339-4, spectroscopy returned high MNRAS000,1–10(2016) 8 Franchini,MottaandLodato spinvaluesusingbothcontinuum fittingandthereflectionmodel 1.0 (Reisetal. 2011; Milleretal. 2002; Reisetal. 2008). Reisetal. (2011)foundanintermediateblackholespina = 0.52±0.11for XTEJ1752-223whileforGX339-4analmostextremespinvalue 0.8 a=0.935±0.010hasbeeninferred.WhileforXTEJ1752-223we wereabletoplaceonlyalowerlimitonthespinthusourfindings arestillconsistentwiththeresultsobtainedusingspectroscopy,the 0.6 spectroscopic estimate of the spin in GX 339-4 is not consistent ec p withtherangeobtainedusingtheRPM. as 0.4 Mottaetal. (2014a) measured with unprecedented precision 4U1543 the mass and spin of the BH in GRO J1655-40 applying the XTEJ1550 RPM using the three QPOs that have been simultaneously ob- 0.2 GX339-4 served(Mottaetal.2012;Strohmayer2001).Theyinferredamass GROJ1655 M = 5.31±0.07M and aspin value a = 0.290±0.003 that is, ⊙ 0.0 asexpected,ingoodagreementwiththerangeobtainedinthispa- 0.0 0.2 0.4 0.6 0.8 1.0 per.Throughspectroscopicstudies,different,muchhigher,values a RPM ofthespinhavebeeninferred.Bymodellingthethermalspectral continuumShafeeetal.(2006)foundaspinrangea=0.65−0.75. Reisetal.(2010)placedalowerlimita=0.9byfittingthestrong reflectionfeaturesinthespectrumandMilleretal.(2009)obtained Figure6.Thehorizontallinesrepresentthedifferentrangesofspinvalues ahighlyspinningblackholewitha = 0.94−0.98usingthesame obtainedinthisworkforeachsource.Theblack,red,greenandcyanlines referto4U1543,XTEJ1550,GX339-4andGROJ1655respectively.On method. Both the spectroscopic methods rely on the assumption they-axisthedotsrepresentthevaluesobtainedusingspectroscopicmeth- that the inner edge of the disc corresponds to the ISCO, which odswiththecorrespondent errorbars.Theblacklinerepresents thecase might not be a correct assumption in all spectral states of active aspec=aRPM. black holes (see, e.g. Doneetal. (2007)). Furthermore, in order tousethespectralcontinuumfittingmethod,onemustalsoknow the black hole mass, the inclination of the (inner) accretion disc thenappliedtheRPM.Thischoiceisjustifiedbythefactthatinthe withrespecttothelineofsightandthedistancetothesource.All HSSthediscisexpectedtoextenddownorveryclosetotheISCO. thesefactorsintroducelargeuncertaintiesandpossiblysignificant Asaconsequence,theinneraccretionflowisextremelynarrowin biasesinthespinmeasurement. Forinstance,asalreadynotedby thisstateandweassumedthatitsLense-Thirringfrequencycanbe Mottaetal. (2014b) in the case of GRO J1655-40, the disagree- welldescribedbythatofatestparticleattheISCO.Accordingto mentbetweenspinmeasurementsinferredthroughdifferentmeth- the RPM, we associated the highest frequency type-C QPO with odsmightbeduetothehighinclinationoftheinneraccretiondisc the Lense-Thirring precession frequency of a test particle at the with respect to the outer disc (i.e. the orbital inclination) and to ISCO.Wetheninferredlowerspinlimitsforallthesourcesofour the line of sight. When such misalignment is not taken into ac- sampleandupperlimitsfortheonesthatshowedtype-CQPOsin count,thefinalspinmeasurementcanbelargelybiased.TheRPM theHSSorULS,assumingareasonablerangeofblackholemasses methoddoesnotdependonanyofthesevariables,andinthecase orusingtheactualmassvalueforthosesourcewhereadynamical where3QPOsaresimultaneouslydetected,itallowstoobtainself massmeasurementisavailable.WesuccessfullyappliedtheRPM consistentlymassandspinoftheBH.Theprecisionontheinferred toall the sources inour sample that showed type-C QPOsin the parametersisonlydrivenbytheprecisionintheQPOfrequencies HSSorULS. measurement. Our results are in good agreement withthose inferred using Morespectroscopy-basedandtiming-basedmeasurementsof other spectroscopy-based methods (continuum fittingmethodand thespinarenecessarytoconfirmthevalidityoftheRPMmethod theK-αlinemodelling)onlyintwocases.Nextgenerationsatel- and to test its consistency with other methods. New generation lites,characterizedbyahigher signal-to-noiseratiomighthelpin satellites(suchaseXTPZhangetal.(2016))willprovidemoreand solvingtheissueofspinmeasurements. betterdatatotestthisandothermethodsforspinmeasurements. Finally, as already noted by Mottaetal. (2014a), we would liketostressthattheRPMisasimplifiedmodelaimedatdescribing 7 ACKNOWLEDGMENTS rather complex physical conditions, such as those encountered in theinnermostdiscregionsaroundanaccretingcompactobject.The Wearegratefultothereferee,Dr.JeroenHoman,thatprovideduse- RPM does not include yet a production mechanism for the QPO fulcommentsandsuggestionsthathelpedinsignificantlyimprove signals(butseePsaltis&Norman2000 orSchnittmanetal.2006 ourwork. AFacknowledges thetransitiongrant oftheUniversità for possible mechanisms), which remains an open question to be degliStudidiMilanoandtheUniversityofOxfordforhospitality. investigatedinthefuture. SEMacknowledgestheUniversityofOxfordandtheVioletteand SamuelGlasstoneResearchFellowshipprogram. 6 SUMMARYANDCONCLUSIONS APPENDIXA: POWERDENSITYSPECTRAAND Inthisworkweplacedlimitsonthevalueoftheblackholespinfor PROPERTIESOFQPOS severalbinariesthroughthesoleuseofX-raytiming.Inparticular welookedfortype-CQPOsinthesoftstateforanumberofsources FigureA1showsaselectionofPDScontainingatype-CQPOde- selectedbasedonthepresenceoftype-CQPOsintheirPDS,and tected in the HSS and in the ULS. Fromtop to bottom, the PDS MNRAS000,1–10(2016) Type-C QPOs insoftstate 9 ObsID ν(Hz) Q rms GX339-4 10−2 GX339-4 92085-01-02-03 10.59 3.46 0.05 4U1630-47 70417-01-09-00 12.3 2.15 0.033 y 4U1543-47 sit n 70133-01-01-00 15.37 2.57 0.042 e d 70133-01-05-00 15.0 3.75 0.046 al XTEJ1817-330 ctr e 91110-02-32-00 9.6 2.93 0.057 p s GROJ1655-40 er 10255-01-05-00 10.7 2.7 0.011 w o P TableA1.PropertiesoftheQPOsfoundinthisworkdividedbysource.For eachQPOthefirstcolumncontainstheObsID,thesecondisthecentroid frequency,thethirdisthequalityfactorQandthelastoneistheintegrated fractionalrms. 10−1 100 101 102 103 Frequency(Hz) comefromGX339-4,4U1543-47andGROJ1655-40respectively. ThefirsttwoQPOsarecharacterizedbyastrongandnarrowpeak 4U1543-47 at frequencies of theorder of 10 Hz. Thethird isa small though significant(3.96σ,singletrial)peakat27.5Hz. WereportinTableA1thepropertiesofthenewQPOsdetected y inthiswork:theObsID,thecentroidfrequency,thequalityfactor sit n e andtheintegratedfractionalrms.Notethatthetotalfractionalrms d measuredfortheobservationwhereagivenQPOisdetectedisal- al ways65%,whichshowsthatalltheseQPOsweredetectedinsoft ctr e states. As noted elsewhere, this impliesthat the inner disc radius sp iscloseorcoincidenttotheISCOandtheRPMcanbeappliedto er w obtainaconstrainonthespinofthecentralblackhole. o P 10−3 REFERENCES 10−1 100 101 102 103 Frequency(Hz) AbbottB.P.,etal.,2016,PhysicalReviewLetters,116,061102 BeerM.E.,PodsiadlowskiP.,2002,MNRAS,331,351 BelloniT.,HasingerG.,1990,A&A,230,103 GROJ1655-40 BelloniT.M.,MottaS.E.,2016,inBambiC.,ed.,AstrophysicsandSpace ScienceLibraryVol.440,AstrophysicsofBlackHoles:FromFunda- mentalAspectstoLatestDevelopments.p.61(arXiv:1603.07872), doi:10.1007/978-3-662-52859-4_2 y BelloniT.,PsaltisD.,vanderKlisM.,2002,ApJ,572,392 nsit BelloniT.M.,MottaS.E.,Muñoz-DariasT.,2011,BulletinoftheAstro- de nomicalSocietyofIndia,39,409 al BelloniT.M.,SannaA.,MéndezM.,2012,MNRAS,426,1701 ctr e CasaresJ.,JonkerP.G.,2014,SpaceSci.Rev.,183,223 p s CasellaP.,BelloniT.,StellaL.,2005,ApJ,629,403 er DoneC.,GierlinskiM.,KubotaA.,2007,A&A,15,1 w o DunnR.J.H.,FenderR.P.,KördingE.G.,BelloniT.,CabanacC.,2010, P MNRAS,403,61 FenderR.P.,HomanJ.,BelloniT.M.,2009,MNRAS,396,1370 FranchiniA.,LodatoG.,FacchiniS.,2016,MNRAS,455,1946 100 101 102 103 HomanJ.,BelloniT.,2005,Ap&SS,300,107 Frequency(Hz) HomanJ.,WijnandsR.,vanderKlisM.,BelloniT.,vanParadijsJ.,Klein- WoltM.,FenderR.,MéndezM.,2001,ApJS,132,377 HomanJ.,MillerJ.M.,WijnandsR.,vanderKlisM.,BelloniT.,Steeghs D.,LewinW.H.G.,2005,ApJ,623,383 FigureA1.Upperpanel:PDSobtainedfromobservation92085-01-02-03 HynesR.I.,SteeghsD.,CasaresJ.,CharlesP.A.,O’BrienK.,2003,ApJ, ofthesourceGX339-4duringitsoutburstin2007.Thetype-CQPOfound 583,L95 at10.6Hzisclearlyvisible.Middlepanel:PDSobtainedfromobservation IngramA.,DoneC.,2011,MNRAS,415,2323 70133-01-01-00ofthesource4U1543-47.Thetype-CQPOfoundat15.4 IngramA.,MottaS.,2014,MNRAS,444,2065 Hzisclearlyvisible.Lowerpanel:PDSobtainedfromobservation91702- IngramA.,DoneC.,FragileP.C.,2009,MNRAS,397,L101 01-17-10ofthesourceGROJ1655-40.Thereisasmallbutsignificantpeak IngramA.,vanderKlisM.,MiddletonM.,DoneC.,AltamiranoD.,Heil atroughly30Hz.Thisisthehighesttype-CQPOdetectedforthissource. L.,UttleyP.,AxelssonM.,2016,preprint, (arXiv:1607.02866) KalamkarM.,HomanJ.,AltamiranoD.,vanderKlisM.,CasellaP.,Linares M.,2011,ApJ,731,L2+ MNRAS000,1–10(2016) 10 Franchini,MottaandLodato KingA.L.,etal.,2014,ApJ,784,L2 KubotaA.,DoneC.,2004,MNRAS,353,980 LeahyD.A.,ElsnerR.F.,WeisskopfM.C.,1983,ApJ,272,256 LodatoG.,FacchiniS.,2013,MNRAS,433,2157 McClintockJ.E.,RemillardR.A.,RupenM.P.,TorresM.A.P.,Steeghs D.,LevineA.M.,OroszJ.A.,2009,ApJ,698,1398 McClintockJ.E.,etal.,2011,ClassicalandQuantumGravity,28,114009 McClintockJ.E.,NarayanR.,SteinerJ.F.,2014,SpaceSci.Rev.,183,295 MillerJ.M.,2007,ARA&A,45,441 MillerJ.M.,FabianA.C.,in’tZandJ.J.M.,ReynoldsC.S.,WijnandsR., NowakM.A.,LewinW.H.G.,2002,ApJ,577,L15 MillerJ.M.,ReynoldsC.S.,FabianA.C.,MiniuttiG.,GalloL.C.,2009, ApJ,697,900 Miyamoto S., Kitamoto S., Iga S., Negoro H., Terada K., 1992, ApJ, 391,L21 MorganE.H.,RemillardR.A.,GreinerJ.,1997,ApJ,482,993 MorningstarW.R.,MillerJ.M.,2014,ApJ,793,L33 MottaS.E.,2016,AstronomischeNachrichten,337,398 Motta S., Muñoz-Darias T., Casella P., Belloni T., Homan J., 2011, MNRAS,418,2292 MottaS.,HomanJ.,MuñozDariasT.,CasellaP.,BelloniT.M.,Hiemstra B.,MéndezM.,2012,MNRAS,427,595 MottaS.E.,BelloniT.M.,StellaL.,Muñoz-DariasT.,FenderR.,2014a, MNRAS,437,2554 MottaS.E.,Muñoz-DariasT.,SannaA.,FenderR.,BelloniT.,StellaL., 2014b,MNRAS,439,L65 MottaS.E.,CasellaP.,HenzeM.,Muñoz-DariasT.,SannaA.,FenderR., BelloniT.,2015,MNRAS,447,2059 Muñoz-DariasT.,CasaresJ.,Martínez-PaisI.G.,2008,MNRAS,385,2205 MuñozDariasT.,StieleH.,BelloniT.,MottaS.,2010,TheAstronomer’s Telegram,2999,1 NixonC.,FranchiniA.,LodatoG.,KingA.,2016,inprep OroszJ.A.,SteinerJ.F.,McClintockJ.E.,TorresM.A.P.,RemillardR.A., BailynC.D.,MillerJ.M.,2011,ApJ,730,75 PlantD.S.,O’BrienK.,FenderR.P.,2014,preprint, (arXiv:1411.7411) PlantD.S.,FenderR.P.,PontiG.,Muñoz-DariasT.,CoriatM.,2015,A&A, 573,A120 PsaltisD.,NormanC.,2000,ArXivAstrophysicse-prints, ReidM.J.,McClintock J.E.,Steiner J.F.,Steeghs D.,Remillard R.A., DhawanV.,NarayanR.,2014,ApJ,796,2 ReisR.C.,FabianA.C.,RossR.R.,MiniuttiG.,MillerJ.M.,Reynolds C.,2008,MNRAS,387,1489 ReisR.C.,FabianA.C.,MillerJ.M.,2010,MNRAS,402,836 ReisR.C.,etal.,2011,MNRAS,410,2497 RemillardR.,McClintockJ.,2006,ARA&A,44,49 ReynoldsC.S.,2013,SpaceSci.Rev., SaitoK.,HomanJ.,YamaokaK.,FukuyamaM.,MiyakawaT.G.,Yoshida A.,2007,ArXivAstrophysicse-prints, SchnittmanJ.D.,HomanJ.,MillerJ.M.,2006,ApJ,642,420 ShafeeR.,McClintockJ.E.,NarayanR.,DavisS.W.,LiL.-X.,Remillard R.A.,2006,ApJ,636,L113 ShakuraN.I.,SunyaevR.A.,1973,A&A,24,337 SobolewskaM.A.,Z˙yckiP.T.,2006,MNRAS,370,405 SoleriP.,etal.,2013,MNRAS,429,1244 Steiner J. F., McClintock J. E., Remillard R. A., Gou L., Yamada S., NarayanR.,2010,ApJ,718,L117 SteinerJ.F.,etal.,2011,MNRAS,416,941 SteinerJ.F.,etal.,2012,MNRAS,427,2552 StellaL.,VietriM.,1998,ApJ,492,L59+ StellaL.,VietriM.,MorsinkS.M.,1999,ApJ,524,L63 StrohmayerT.E.,2001,ApJ,552,L49 TanakaY.,LewinW.H.G.,1995,pp126–174 WijnandsR.,HomanJ.,vanderKlisM.,1999,ApJ,526,L33 ZhangS.N.,etal.,2016,preprint, (arXiv:1607.08823) vanderKlisM.,2006,AdvancesinSpaceResearch,38,2675 ThispaperhasbeentypesetfromaTEX/LATEXfilepreparedbytheauthor. MNRAS000,1–10(2016)