Black Hole Mass Limits for Optically Dark X-ray Bright Sources in Elliptical Galaxies
V. Jithesh, K. Jeena, R. Misra, S. Ravindranath, G. C. Dewangan, C. D. Ravikumar, B. R. S. Babu
aa r X i v : . [ a s t r o - ph . H E ] J a n Black Hole Mass Limits for Optically Dark X-ray Bright Sources inElliptical Galaxies
V. Jithesh , K. Jeena , R. Misra , S. Ravindranath , G. C. Dewangan , C. D. Ravikumar , and B.R. S. Babu ABSTRACT
Estimation of the black hole mass in bright X-ray sources of nearby galaxies is cru-cial to the understanding of these systems and their formation. However, the presentallowed black hole mass range spans five order of magnitude (10 M ⊙ < M < M ⊙ )with the upper limit obtained from dynamical friction arguments. We show that the ab-sence of a detectable optical counterpart for some of these sources, can provide a muchmore stringent upper limit. The argument is based only on the assumption that theouter regions of their accretion disks is a standard one. Moreover, such optically darkX-ray sources cannot be foreground stars or background active galactic nuclei, andhence must be accreting systems residing within their host galaxies. As a demonstra-tion we search for candidates among the point-like X-ray sources detected with Chan-dra in thirteen nearby elliptical galaxies. We use a novel technique to search for faintoptical counterparts in the
HST images whereby we subtract the bright galaxy lightbased on isophotal modeling of the surface brightness. We show that for six sourceswith no detectable optical emission at the 3 − sigma level, their black hole masses M BH < M ⊙ . In particular, an ultra-luminous X-ray source (ULX) in NGC 4486has M BH < M ⊙ . We discuss the potential of this method to provide stringent con-straints on the black hole masses, and the implications on the physical nature of thesesources. Subject headings: accretion, accretion disks, galaxies:photometry, X-rays:galaxies
1. Introduction
Compact, off-nuclear X-ray point sources in nearby galaxies, with luminosities 10 − ergs s − are referred to as Ultra-Luminous X-ray sources (ULXs). Detected in the early 1980’s, with the Department of Physics, University of Calicut, Malappuram-673635, India; [email protected] Inter-University Center for Astronomy and Astrophysics, Post Bag 4, Ganeshkhind, Pune-411007, India;[email protected]
Einstein
X-ray satellite (Fabbiano 1989), these objects were further studied with
ROSAT (Colbert & Mushotzky1999) and
ASCA (Makishima et al. 2000). The
XMM-Newton and
Chandra
X-ray observatorieswith their significantly higher angular resolution, dramatically confirmed the presence of ULXs(Kaaret et al. M ⊙ black hole. SinceULX are off-nuclear sources, their masses must be < M ⊙ from dynamical friction arguments(Kaaret et al. M ⊙ < M < M ⊙ ) lies between that of stellar mass black holes and super-massive black holes observed in galaxy centers (Makishima et al. 2000). Alternatively, ULX maybe stellar mass black hole systems exhibiting super-Eddington accretions with their radiation geo-metrically beamed (Shakura & Sunyaev 1973; King 2008).X-ray spectroscopy has provided supporting evidence in favor of IMBHs of ∼ M ⊙ inULXs (Miller et al. ∼ M ⊙ black holes (Strohmayer & Mushotzky 2003; Dewangan et al. 2006; Mucciarelli et al.2006; Strohmayer & Mushotzky 2009). While indicative these results are not conclusive, sincethere are also several arguments against IMBHs in ULXs (see e.g., Mushotzky 2004; Roberts2007) and further investigations are required to reveal the true nature of these sources.Study of the host galaxy properties of ULXs reveals that their number and total X-ray lumi-nosity is related to recent star formation activity, suggesting that they originate in young short-livedsystems (Swartz et al. et al. ∼
44% level (Swartz et al. M ⊙ < M < M ⊙ ) and it is important to obtain tighter constrains. Here, we show 3 –that the absence of a detectable optical emission allows us to impose an upper limit on the blackhole mass for these accreting systems based on some standard assumptions. Moreoever, We arguethat these optically dark X-ray sources cannot be foreground stars or background AGN and henceare a true sample of sources located within the host galaxy. To demonstrate the technique, wesearch for candiatates among X-ray bright sources detected by
Chandra in archival
HST
ACS, andWFPC2 images.
2. Observations and Data Reduction
Swartz et al. (2004) analyzed
Chandra observations of 82 nearby galaxies and have identifiedbright X-ray sources in them. Of these we select 13 elliptical galaxies for which
HST observationsare available. Devi et al. (2007) analyzed a subset of thirty galaxies from the Swartz et al. (2004)sample, and fitted the X-ray points sources with both a power-law and a disk black body model,hence obtaining a more conservative and robust estimation of their X-ray luminosity. Nine outof the sub-sample analyzed by Devi et al. (2007) are ellipticals, and for these galaxies we use theX-ray luminosity and coordinates given by them for the present analysis. For the remaining fourgalaxies we use the values quoted by Swartz et al. (2004). The details of the sample galaxies aregiven in Table 1.The optical study was carried out using the images taken with the Advanced Camera forSurveys (ACS), and Wide Field and Planetary Camera (WFPC2) that are available in the
HST data archive. When observations from both cameras were available, the datasets with the longerexposure time, and multiple filters were preferred. For the same combination of camera and filters,various images available for different pointings of the same galaxy were analyzed when available,so as to maximize the number of X-ray sources because for different pointings, a different set ofX-ray sources would fall in the field of view.Most of the optical sources are too faint to be detected against the dominant galaxy light thatfills most of the
HST images. To enhance the contrast, and aid in the detection of point sourcesin the image, the galaxy light was modeled based on the isophotes obtained using the ellipse taskin
IRAF/STSDAS software. Bright objects, if any, were masked during the fitting. The residualimage was obtained by subtracting the model image from the observed galaxy image. The objectextraction was done on the residual image using
SEXTRACTOR with a threshold level of 3 sigma.On visual inspection, we find that many of the
Chandra
X-ray sources have counterparts in the
HST images, within a positional offset of less than a few arc-seconds. This constant offset was appliedfor a given galaxy, to match the
Chandra sources to the optical sources in the
SEXTRACTOR catalog. When images are available in multiple filters, one of the optical images is considered asthe reference image, and the necessary offset is applied to match it to the X-ray co-ordinates. The 4 –Table 1. Sample Galaxy PropertiesGalaxy Distance (Mpc) N x N d NGC 1399 18 . . . . . . . . . . . . . (a) NGC 4486 (b) NGC 4697 (c) NGC 4649(d) NGC 4374 (e) NGC 1399 (f) NGC 1316 Fig. 1.—: The HST images of the first six X-ray sources (marked X) listed in Table 2. Overlaidare 1 ′′ and 3 ′′ circles centered on the shifted Chandra positions. The plus signs mark the original
Chandra positions. Note that for four of the images, there are no optical sources even within 3 ′′ ofthe X-ray position. 6 –Table 2. The Upper limit of Black hole Mass of some of the Optically Dark X-ray SourcesGalaxy R . A . ( J ) Decl . ( J ) log L x HST Filter F n × − F X / F O M U ( M ⊙ ) NGC 4486 12 30 50.82 +12 25 02.66 39 . + . − . F475W 0.409 533 1244NGC 4697 12 48 33.20 -05 47 41.17 38 . + . − . F475W 0.752 243 2890NGC 4649 12 43 41.90 +11 34 33.83 38 . + . − . F475W 0.402 164 3073NGC 4374 12 25 01.54 +12 52 35.59 39 . + . − . F475W 0.441 347 3378NGC 1399 3 38 25.92 -35 27 42.37 38 . + . − . F606W 0.370 228 3927NGC 1316 3 22 36.46 -37 13 24.68 38 . + . − . F475W 0.449 179 4780NGC 1316 3 22 35.58 -37 13 14.10 38 . + . − . F555W 0.520 287 6366NGC 1399 3 38 32.33 -35 26 45.73 38 . + . − . F606W 0.377 186 7829NGC 1399 3 38 27.62 -35 26 48.76 39 . + . − . F606W 0.766 702 7829NGC 4649 12 43 34.17 +11 33 41.93 39 . + . − . F475W 0.912 97 8073Note. — (1) Host galaxy name; (2) Right Ascension of shifted position in hours, minutes andseconds; (3) Declination of shifted position in degrees, arcminutes and arcseconds; (4) log of X-rayluminosity in ergs / s ; (5) HST filter for which the upper limit on flux and the black hole mass limitis calculated; (6) Upper limit on Optical flux in ergs/s/cm /Hz; (7) Lower limit on ratio of X-ray tooptical flux; (8) M U , Upper limit on black hole mass. 7 –images in other optical filters are then aligned to the reference image using geomap and geotran tasks in IRAF. While a more detailed report on the nature of sources with optical counterparts willbe presented later, in this work, we concentrate on those sources for which no optical counterpartwas detected.X-ray sources which did not have an optical counterpart (at the 3 − sigma level) within 1 ′′ oftheir shifted positions, are termed as optically “dark” sources. Six examples of such sources areshown in Figure 1, where the 1 ′′ and 3 ′′ circles centered on the X-ray co-ordinates are overlaid onthe observed optical image. Note that for four of these images, there are no optical source within3 ′′ of the X-ray position. Having identified such optically ”dark” X-ray sources, we estimate theupper limit on their optical flux based on the 3 − sigma threshold at that position.
3. Optically dark X-ray sources
These optically dark sources are X-ray bright compared to their optical emission and henceare not foreground stars. This can be further quantified by estimating the X-ray-to-optical fluxratio log ( f X / f O ) where f X is the unabsorbed flux in the 0 . − f O is the flux inan optical band. This ratio ranges from 0 . f X is inthe 0 . − . R , of an accretion disk around ablack hole with mass M and accretion rate ˙ M , is given by s T ( R ) = p GM ˙ MR d ( R ) where d ( R ) = − ( GM / Rc ) / . The observed flux from the disk at a frequency n is then given by the integratedsum of the black body emission over all radii, F n = cos iD Z R out R in B n ( n , T ( R )) p RdR (1)where B n ( n , T ( R )) is the blackbody intensity, R out and R in = GM / c are outer and inner radii ofthe disk, i is the inclination angle of the disk and D is the distance to the source. Assuming thatmost of the contribution to F n arises from regions in the disk that are far away from the inner andouter radii, the expected observed flux can be written as F n ∼ × − ergs s − cm − Hz − ( l A ) − / ( h . ) − / × ( L x ergs s − ) / ( D ) − ( M M ⊙ ) / (2)where l is the wavelength and h = L x / ˙ Mc is the radiative efficiency of the accreting system andcos i is taken to be 0 . F n , max . Thus one can estimate an upper limit on the black hole mass as M U < M ⊙ ( F n , max × − ergs s − cm − Hz − ) / × ( l A ) / ( h . )( L x ergs s − ) − ( D ) (3)For each dark X-ray source in our sample, and for all available filters, we estimate this upper limiton the black hole mass. We use the integration (Eqn 1) to evaluate the upper limit, rather thanthe approximation Eqn. (3) i.e. we take into account the effect of the inner boundary condition( d ( R ) = − ( GM / Rc ) / ) on the temperature profile. The difference in the upper limit obtainedis marginal ( < h = . i = .
5. Toobtain a more conservative upper limit, the one-sigma lower value of the X-ray luminosity areused L x − D L x . In Table 2, we list the ten best cases in ascending order of black hole mass limit M U . For the other sources, M U > M ⊙ and hence is not a significant constraint. The best caseis for the ULX in NGC 4486, for which M U = M ⊙ .X-ray irradiation of the outer disk may increase the local temperature there and the disk mayemit a larger optical emission. We have estimated this effect using the formalism given in theappendix of Vrtilek et al. (1990) and find that X-ray irradiation is not important for the constraintobtained here.
4. Discussion
Optically dark X-ray sources cannot be foreground stars or background AGN, otherwise theiroptical emission would be significantly higher than what is detected. Hence, these are a cleansample of sources within the host galaxies, which are probably accreting black hole systems. Theoptical emission from a standard accretion disk scales as mass of the black hole M / and hencethe non-detection of optical emission imposes an upper limit on the black holes mass M U . Forten of the sources M U < M ⊙ . For a source in NGC 4486 with an X-ray luminosity clearlyexceeding 10 ergs/s (and therefore a bona-fide ULX by definition), the estimated black hole massis smaller than 1244 M ⊙ . This is two orders of magnitude smaller than the constraint obtained fromdynamical friction, which is 10 M ⊙ . 9 –These sources with black hole mass, M U < M ⊙ cannot be accreting systems with massiveblack holes residing in star clusters, or in the nuclei of merged satellite galaxies. For typical low-luminosity dwarf galaxies ( M B ∼ − .
0; Mateo 1998), such an optical counterpart would be easilydetected, given that our 3 − sigma limits on the HST images are much fainter. Even a compactnucleus of a merged dwarf galaxy ( M B ∼ − .
0; Lotz et al. 2004) would have been easily identi-fied. If they are binary systems, their companion cannot be a massive O star as such a star wouldhave been detected in the optical image. Assuming an O star, with M B ∼ − .
5, we find that thepossibility of such a companion can be ruled out in all cases for which M U < M ⊙ (Table 2).In all of the above arguments, we have ignored the effect of dust obscuration in the hostgalaxies, because we are only considering elliptical galaxies in the present work. Although someellipticals are known to have dust lanes, and nuclear rings in their centers, the X-ray sources weare considering here are distributed at fairly large radial distance from the center to be significantlyaffected by dust.Even for the best cases of optically dark X-ray sources presented in Table (2), the range ofblack hole mass allowed is still large as the source could be a < ∼ M ⊙ intermediate mass blackhole, or a few solar mass object emitting at super-Eddington luminosities. The brightest X-raysources in the sample have a luminosity of a few times 10 ergs/s. This is unfortunate, since adark source with luminosity > would have provided an order of magnitude better constrainton the black hole mass. Since the black hole mass upper limit M U (cid:181) D a bright X-ray sourcein a more nearby galaxy ( D ∼ M pc ) would have also provided significantly better constraints.A systematic search for such sources in very nearby galaxies may indeed prove fruitful. Anotherpoint to note is that bright X-ray sources in these galaxies are known to be variable in X-rays.Our analysis in this work, implicitly assumes that the X-ray luminosity observed through a single
Chandra observation, represents an average luminosity which is used to derive an average accretionrate ˙ M = L X / ( h c ) which in turn is used to estimate the upper limit on the black hole mass (Eqn3). This assumption is required because the expected optical emission arises from the outer part ofthe disk and the local accretion rate there may be different than the one in the inner region whichproduces the X-rays. Any accretion rate fluctuation in the outer disk will be transfered along thedisk on the viscous time-scale which could be significantly longer than a day. Thus, in principleone needs to ascertain the average X-ray luminosity of a source, and using a single very bright, butrare, X-ray observation of the source will not represent the average accretion rate.A systematic and comprehensive multi-wavelength study (using also other bands like infra-red and radio), along with X-ray variability studies, can shed further light on the nature of thesesources.VJ, KJ, CDR and BRSB would like to thank the IUCAA visitors program and and UGC 10 –Special assistance program. This work has been partially funded from the ISRO-RESPOND pro-gramme. The authors would like to thank Phil Charles for useful discussions. REFERENCES
Bonfini, P., Hatzidimitriou, D., Pietsch, W., & Reig, P. 2009, A&A, 507, 705Colbert, E. J. M., & Mushotzky, R. F. 1999, ApJ, 519, 89Cropper, M., Soria, R., Mushotzky, R. F., Wu, K., Markwardt, C. B., & Pakull, M. 2004, MNRAS,349, 39Devi, A. S., Misra, R., Agarwal, V. K., & Singh, K. Y. 2007, ApJ, 664, 458Devi, A. S., Misra, R., Shanthi, K., & Singh, K. Y. 2008, ApJ, 665, 455Dewangan, G. C., Miyaji, T., Griffiths, R. E., & Lehmann, I. 2004, ApJ, 608, L57Dewangan, G. C., Titarchuk, L., & Griffiths, R. E. 2006, ApJ, 637, L21Fabbiano, G. 1989, ARA&A, 27, 87Goad, M., Roberts, T., Knigge, C., & Lira, P. 2002, MNRAS, 335, L67.Gutierrez, C. 2006, ApJ, 640, L17Kaaret, P., et al.