(No) dynamical constraints on the mass of the black hole in two ULXs
T. P. Roberts, J. C. Gladstone, A. D. Goulding, A. M. Swinbank, M. J. Ward, M. R. Goad, A. J. Levan
aa r X i v : . [ a s t r o - ph . H E ] N ov Astron. Nachr. / AN , No. 88, 789 – 793 (2006) /
DOI please set DOI! (No) dynamical constraints on the mass of the black hole in two ULXs
T.P. Roberts ,⋆ , J.C. Gladstone , A.D. Goulding , A.M. Swinbank , M.J. Ward , M.R. Goad , and A.J.Levan Department of Physics, Durham University, South Road, Durham DH1 3LE, UK Department of Physics, University of Alberta, Edmonton, Alberta, T6G 2G7, Canada X-ray and Observational Astronomy Group, Dept. of Physics & Astronomy, University of Leicester, University Road,Leicester LE1 7RH, UK Department of Physics, University of Warwick, Coventry CV4 7AL, UKThe dates of receipt and acceptance should be inserted later
Key words accretion, accretion discs – X-rays: binaries – black hole physics – binaries: spectroscopicWe present the preliminary results of two Gemini campaigns to constrain the mass of the black hole in an ultraluminousX-ray source (ULX) via optical spectroscopy. Pilot studies of the optical counterparts of a number of ULXs revealed twocandidates for further detailed study, based on the presence of a broad He II 4686 ˚A emission line. A sequence of 10long-slit spectra were obtained for each object, and the velocity shift of the ULX counterpart measured. Although radialvelocity variations are observed, they are not sinusoidal, and no mass function is obtained. However, the broad He II lineis highly variable on timescales shorter than a day. If associated with the reprocessing of X-rays in the accretion disc, itsbreadth implies that the disc must be close to face-on. c (cid:13) The key uncertainty responsible for driving the study of ul-traluminous X-ray sources (ULXs) over the past decade isthe mass of the compact objects powering this extraordinaryphenomenon. This unknown has been addressed by variousingenious methods, drawing evidence from across the elec-tromagnetic spectrum, many of which are discussed else-where in these proceedings. (A separate discussion of manyof these methods, and their results, is presented in Zampieri& Roberts 2009). A consensus has emerged that ULXs arepowered by accretion onto a black hole; but the question ofthe mass of these black holes remains unanswered.The reason that ULX masses remain mired in contro-versy is that none of the currently utilised methods pro-vides a direct, unambiguous measurement of the black holemass. This is perhaps best exemplified by X-ray spectralanalyses. A great deal of progress has been made in recentyears on the basis of
Chandra and
XMM-Newton spectra,firstly identifying soft excesses consistent with the cool ac-cretion disc signature one would expect from a ∼ M ⊙ intermediate-mass black hole (e.g. Miller et al. 2003; Miller,Fabian & Miller 2004). Latterly this interpretation has beenstrongly challenged by the detection of a spectral break atenergies of a few keV, identified in the best quality XMM-Newton data for a wide range of ULXs. This implies thatULXs are operating in an unfamiliar spectral state, mostlikely associated with super-Eddington processes (Stobbart,Roberts & Wilms 2006; Roberts 2007; Gladstone, Roberts ⋆ Corresponding author: e-mail: [email protected] & Done 2009; also Gladstone, these proceedings). This ‘ul-traluminous state’ appears to display the characteristic im-print of a strong outflowing wind, as predicted for super-Eddington emission (e.g. Begelman, King & Pringle 2006;Poutanen et al. 2007), and so implies that ULXs harboursmall, ∼ < M ⊙ black holes.However, neither of the above examples provides a di-rect mass estimate and, worse still, for most ULX X-rayspectral data degeneracy is a problem as the quality is suf-ficiently poor that neither model can be rejected . Other in-direct methods suffer similarly - the QPOs detected in thepower density spectra of NGC 5408 X-1 can be used to in-fer the presence of an IMBH (e.g. Strohmayer & Mushotzky2009; also Strohmayer these proceedings), but this assumesboth a specific type of QPOs and a sub-Eddington accre-tion state. Neither may be true for this source (Middleton etal. 2010; also these proceedings). Similarly, different mod-els for the optical colours and magnitudes of various ULXcounterparts lead to a range of mass estimates (e.g. Cop-perwheat et al. 2006; Madhusudhan et al. 2009; Patruno &Zampieri 2010). It is clear therefore that we require a ‘clean’test of ULX mass, untainted by model assumptions.The obvious way forward is to perform similar exper-iments to those that have a near four-decade heritage forGalactic systems: dynamical studies based on the co-orbitalmotion of the accreting compact object and its companion,donor star (see e.g. Charles & Coe 2006; Casares 2007).In such experiments one commonly obtains optical (and/or Indeed, other spectral models may also be applied in some cases,for example slim disc spectra (e.g. Vierdayanti et al. 2006), or reflection-dominated spectra (Caballero-Garc´ıa & Fabian 2010), to name but two. c (cid:13)
90 T.P. Roberts et al.: (No) dynamical constraints on the mass of two ULXs
UV/IR) spectra at a series of different epochs, and mea-sures the semi-velocity amplitude K and period P for thesinusoidal orbital motions, as traced out by shifts in the ob-served emission and/or absorption line wavelengths. Mea-surements are usually taken from periods when the donorstar dominates the optical light, and used to infer a massfunction f ( M ) that places a lower limit on the black holemass, M X . However, it is also possible to use emission fea-tures originating in the accretion disc to produce a massfunction (e.g. Orosz et al. 1994; Soria et al. 1998) such that f ( M ) = M sin i ( M C + M X ) = P K πG , (1)where M C is the mass of the companion star, K X thesemi-velocity amplitude of the black hole, and i the incli-nation of the orbital plane to our line-of-sight, thus placinglimits on the mass of the black hole.We can take encouragement from the recent reports ofmass functions for two extragalactic BHBs, M33 X-8 andIC 10 X-1 (Orosz et al. 2007; Prestwich et al. 2007; Sil-verman & Filippenko 2008). However, most ULXs are atleast three times more distant than these objects, with m V ∼ . at best (Motch, these proceedings) and more typically m V > (Roberts, Levan & Goad 2008). Furthermore,many are located in complex fields, where their spectro-scopic signal could be confused with neighbouring nebu-losity and/or stars. Indeed, only a handful of ULXs mightbe accessible for mass function measurements with currentfacilities. Very few attempts have been made to date, andthese have been unsuccessful. Kaaret & Corbel (2009) ob-tained 6 VLT/FORS observations of NGC 5408 X-1 over 3days, but found no stellar absorption lines to base a massfunction on. Pakull, Gris´e & Motch (2006) obtained multi-epoch data for NGC 1313 X-2 and did find an interestingvelocity shift ( ∆ v ∼
380 km s − ) in the centroid of a broadHe II line, but were unable to constrain a mass function fromsubsequent follow-up data (Gris´e et al. 2009). Hence thefirst mass function measurement for a ULX remains a tan-talising goal. Here, we detail the preliminary results of anew attempt to obtain the mass functions of two ULXs, HoIX X-1 and NGC 1313 X-2, using the Gemini observatorytelescopes. We have been developing a programme for the past fewyears, with the sole aim of obtaining a dynamical mass mea-surement for a ULX. The programme has three main steps:(i) identify the optical counterparts to ULXs on the basis ofthe most accurate available X-ray positions from
Chandra ,and
HST imaging; (ii) obtain pilot optical spectroscopy toinvestigate the presence of useful spectral features; and (iii)undertake the radial velocity measurements campaign.
Fig. 1
Pilot optical long-slit spectra of the ULX coun-terparts. Data were taken by the GMOS instruments onGemini-N (NGC 5204 X-1 & Ho IX X-1) and Gemini-S(NGC 1313 X-2), using the B600 gratings. The optical mag-nitudes of the counterparts (extinction-corrected Vegamags,in
HST filters) and exposure times were: m = 23 . , 3 hr(NGC 1313 X-2); m = 22 . , 0.8 hr (NGC 5204 X-1); m = 22 . , 1.5 hr (Ho IX X-1). In the first step we surveyed nearby ( d < Mpc) ULXswith available
Chandra and
HST data, and selected rela-tively bright ( m V ∼ < . ) and isolated objects for furtherstudy. In Fig. 1 we show the pilot spectra, obtained usingthe GMOS instruments on the Gemini telescopes, for threeX-ray luminous ( L X > × erg s − ) ULX counter-parts. Each has previously been identified in the literature,with positions shown by e.g. Liu, Bregman & Seitzer (2004,NGC 5204 X-1) and Ramsey et al. (2006, NGC 1313 X-2& Ho IX X-1).Fig. 1 shows the pilot spectra are dominated by a rel-atively featureless continuum, and emission lines from thebubble nebulae known to surround each of these three ob-jects (Pakull & Mirioni 2002). However, one interesting fea-ture is seen in two of the spectra: a broad He II Ten follow-up Gemini observations were obtained for eachof NGC 1313 X-2 (2.5 hr per observation on Gemini-S) andHo IX X-1 (1.5 hr per observation on Gemini-N) in semester2009B. Nine of the observations of NGC 1313 X-2 wereperformed over a 13-day period in December 2009, with aview to sampling over the known ∼ day photometric pe-riod of this ULX (Liu, Bregman & McClintock 2009). Theobservations of Ho IX X-1 were split into two blocks of c (cid:13) stron. Nachr. / AN (2006) 791 Fig. 2
Observed radial velocity shifts of the ULX coun-terparts relative to the bubble nebulae, measured from thebroad He II Top panel:
Ho IX X-1.
Lower panel:
NGC 1313 X-2.5 observations, in late December 2009 and February 2010.Details of the data analysis will be presented by Gladstoneet al. (in prep.). In brief, spectra were extracted for both thecounterpart, and for the surrounding nebula. The redshift ofthe region containing the ULX was constrained from the(off-ULX) nebular emission line spectrum for each obser-vation, and this was then used as a fiducial marker to searchfor relative changes in the He II The results we present here are based on an initial analysisof the data, and focus primarily on NGC 1313 X-2. A fur-ther, more complete analysis is in preparation by Gladstoneet al. and Roberts et al.The preliminary results show that the campaign was suc-cessful on one count: we detected the anticipated shifts inthe He II ±
100 km s − for HoIX X-1, and up to ∼
200 km s − for NGC 1313 X-2. How-ever, the data was not consistent with sinusoidal variationsin either case. Simple sinusoid fitting to both datasets re-sulted in poor fits, with P = 1 . days, K = 77 km s − and χ ν ∼ for Ho IX X-1; and P = 3 . days, K = Fig. 3 He II ≥ σ ) line detections. The last twodetections have lower statistical significances, of 4 and σ respectively.
61 km s − and χ ν ∼ for NGC 1313 X-2. Given the lackof evidence for sinusoidal (i.e. orbital) variations in the data,we must therefore conclude that the data does not provide anew, solid constraint on the mass function for either object.Despite the lack of mass function measurements, thedata does reveal interesting behaviour from the ULX coun-terpart. One notable feature is the highly variable natureof the broad He II ∼
900 km s − down to the instrumental resolu-tion ( <
290 km s − ), with variations of factor 2 seen in 24hours. This dramatic variability implies at the minimum thatthe broad He II emission must be originating within 24 light-hours of the ULX.Such variations might come from X-ray reprocessing inthe outer regions of the accretion disc. If so, the broad lines c (cid:13)
92 T.P. Roberts et al.: (No) dynamical constraints on the mass of two ULXs
Fig. 4
Mass limits on the black hole in NGC 1313 X-2, assuming the He II σ upper limit of
182 km s − derived from the rms scatter of the velocity shift, companion stellar massestimates based on Patruno & Zampieri (2010), and calculate lower limits on the mass (for a range of inclinations, as perthe legend) based on Eq. 1.tell us that the disc cannot be perfectly face-on (or else wewould see no line-of-sight velocity variations). This meansthat there should be some information on the black hole or-bit hidden within the radial velocity data: we therefore usethe σ upper limit derived from the scatter in radial veloc-ity measurements for NGC 1313 X-2 to derive lower lim-its on its black hole mass for a range of orbital periods inFig 4. However, as Kaaret & Corbel pointed out for NGC5408 X-1, such small measured velocities are inconsistentwith the expected velocities of material in the accretion disc( ≫ − ). A possible solution is that only a smallcomponent of the accretion disc rotational velocity is in theline-of-sight, i.e. the disc is very close to face-on ( i ∼ ◦ ).We see in Fig 4 that even at i = 30 ◦ there is little or nolower limit on the black hole mass, other than at very shortperiods . Hence if the disc is close to face-on, little can besaid about the black hole mass. The initial analyses of our campaigns to determine massfunctions for two ULXs betray no strong evidence for pe-riodic velocity variations from either object; and so no massfunction is forthcoming. We do see some interesting phe-nomenology in the ULX counterparts, including strong vari-ability in the He II L X > × erg s − ), andmodels of super-Eddington discs predict that their observedflux will be highest close to this line-of-sight (Mineshige,these proceedings). This is perhaps more circumstantial ev-idence for models of ULXs as super-Eddington stellar-masssystems rather than IMBHs. In the meantime, we still awaitthe first determination of the mass function for a ULX. Although removing the last two, low significance radial velocitypoints, significantly reduces the rms scatter to km s − , and this canlead to some interesting constraints, e.g. for a 16 M ⊙ companion and ◦ inclination, M BH > M ⊙ over all periods up to 10 days. Acknowledgements.
The authors thank Manfred Pakull and PeterJonker for suggestions on improvements to the analysis presentedin this paper, the anonymous referee for useful comments, and theorganisers of the “Ultraluminous X-ray sources and middle-weightblack holes” workshop for a very stimulating meeting. This workis based on observations obtained at the Gemini Observatory.
References
Begelman M.C., King A.R., Pringle J.E.: 2006, MNRAS 370, 399Caballero-Garc´ıa M.D., Fabian A.C.: 2010, MNRAS 402, 2559Casares J.: 2007, in: “Black Holes: from Stars to Galaxies - acrossthe Range of Masses” (Procs. of IAU Symp. in press . ( arXiv:1009.2686 )Miller J.M., Fabbiano G., Miller M.C., Fabian A.C.: 2003, ApJ585, L37Miller J.M., Fabian A.C., Miller M.C.: 2004, ApJ 614, L117Orosz J., Bailyn C.D., Remillard R.A., McClintock J.E., FoltzC.B.: 1994, ApJ 436, 848Orosz J., et al.: 2007, Nature 449, 872Pakull M.W., Gris´e F., Motch C.: 2006, in: “Populations of HighEnergy Sources in Galaxies” (Procs. of IAU Symp. arXiv:astro-ph/0202488
Patruno A., Zampieri L.: 2010, MNRAS 403, L69Poutanen J., Lipunova G., Fabrika S., Butkevich A.G., AbolmasovP.: 2007, MNRAS 377, 1187Prestwich A.H., et al.: 2007, ApJ 669, L21Ramsey C.J., Williams R.M., Gruendl R.A., Chen C.-H.R., ChuY.-H., Wang Q.D.: 2006, ApJ 641, 241 c (cid:13) stron. Nachr. / AN (2006) 793Roberts T.P.: 2007, Ap&SS 311, 203Roberts T.P., Levan A.J., Goad M.R.: 2008, MNRAS 387, 73Silverman J.M., Filippenko A.V.: 2008, ApJ 678, L17Soria R., Wickramasinghe D.T., Hunstead R.W., Wu K.: 1998, ApJ495, L95Stobbart A.M., Roberts T.P., Wilms J.: 2006, MNRAS 368, 397Strohmayer T.E., Mushotzky R.F.: 2009, ApJ 703, 1386Vierdayanti K., Mineshige S., Ebisawa K., Kawaguchi T.: 2006,PASJ 58, 915Zampieri L., Roberts T.P.: 2009, MNRAS 400, 677 c (cid:13)(cid:13)