1: %\documentclass[12pt,preprint]{aastex}

2: \documentclass{emulateapj}

3: 

4: \newcommand{\Ms}{M_\star} \newcommand{\Rs}{R_\star}

5: \newcommand{\Mbh}{M_{\bullet}} \newcommand{\Mo}{M_{\odot}}

6: \newcommand{\Ro}{R_{\odot}} \newcommand{\Jdot}{\mbox{$\dot{J}$}}

7: \newcommand{\rdot}{\mbox{$\dot{r}$}}

8: \newcommand{\Rdot}{\mbox{$\dot{R}$}}

9: \newcommand{\mdot}{\mbox{$\dot{m}$}}

10: \newcommand{\Mdot}{\mbox{$\dot{M}$}}

11: \newcommand{\Lx}{\mbox{$L_{\rmx}$}} \newcommand{\LEdd}{\mbox{$L_{\rm

12: Edd}$}} \newcommand{\msun}{\mbox{${\rm M}_\odot$}}

13: 

14: 

15: \def\apgt{\ {\raise-.5ex\hbox{$\buildrel>\over\sim$}}\ } \def\aplt{\

16: {\raise-.5ex\hbox{$\buildrel<\over\sim$}}\ }

17: 

18: 

19: 

20: 

21: 

22: \shortauthors{HOPMAN, PORTEGIES ZWART \& ALEXANDER}

23: \shorttitle{ULTRALUMINOUS X-RAY SOURCES}

24: 

25: \bibpunct{\hspace{-4pt}}{}{}{a}{}{}

26: 

27: 

28: \begin{document}

29: \title{Ultraluminous X-ray sources as intermediate mass black holes

30: fed by tidally captured stars}

31: 

32: \author{Clovis Hopman\altaffilmark{1}, Simon F. Portegies

33: Zwart\altaffilmark{2, 3} and Tal Alexander\altaffilmark{1}}

34: 

35: \altaffiltext{1}{Faculty of Physics, Weizmann Institute of Science,

36: POB 26, Rehovot 76100, Israel; clovis.hopman@weizmann.ac.il,

37: tal.alexander@weizmann.ac.il} \altaffiltext{2}{Astronomical Institute

38: 'Anton Pannekoek', University of Amsterdam, Kruislaan 403, Netherlands}

39: \altaffiltext{3}{Institute for Computer Science, University of

40: Amsterdam, Kruislaan 403, Netherlands; spz@science.uva.nl}

41: 

42: \begin{abstract}

43: The nature of ultraluminous X-ray sources (ULXs) is presently unknown.

44: A possible explanation is that they are accreting intermediate mass

45: black holes (IBHs) that are fed by Roche lobe overflow from a tidally

46: captured stellar companion. We show that a star can circularize around

47: an IBH without being destroyed by tidal heating (in contrast to the

48: case of $M_\bullet\!>\!10^6 M_\odot$ massive black holes in galactic

49: centers, where survival is unlikely). We find that the capture and

50: circularization rate is $\sim 5\!\times\!10^{-8}\,{\rm yr}^{-1}$,

51: almost independently of the cluster's relaxation time. We follow the

52: luminosity evolution of the binary system during the main sequence

53: Roche lobe overflow phase and show it can maintain ULX-like

54: luminosities for $>10^7\,{\rm yr}$. In particular, we show that the

55: ULX in the young cluster MGG-11 in star-burst galaxy M82, which

56: possibly harbors an IBH, is well explained by this mechanism, and we

57: predict that $\gtrsim\!10\%$ of similar clusters with IBHs have a

58: tidally captured circularized star. The cluster can evaporate on a

59: time-scale shorter than the lifetime of the binary. This raises the

60: possibility of a ULX that outlives its host cluster, or even lights up

61: only after the cluster has evaporated, in agreement with observations

62: of host-less ULXs.

63: 

64: \end{abstract}

65: 

66: 

67: 

68: 

69: \keywords{black hole physics --- stellar dynamics --- galaxies: star

70: 	  clusters --- X-ray binaries}

71: 

72: \section{Introduction}

73: 

74: 

75: Black holes (BHs) have deep potential wells and can transform

76: gravitational energy very efficiently to radiation.  The energetic

77: central engines of quasars are thought to host massive black holes

78: (MBHs) of $M_\bullet>10^6M_\odot$. It is natural to extrapolate this

79: idea and invoke an intermediate mass black hole (IBH; $10^2\lesssim

80: M_\bullet/M_\odot\lesssim10^5$) to explain ultraluminous X-ray sources

81: (ULXs), which are considerably brighter than a stellar mass object

82: radiating at its Eddington luminosity.  For example, Kaaret et

83: al. ({\cite{Ka01}) have suggested that an IBH powers the ULX in MGG-11

84: in star-burst galaxy M82.

85: 

86: The origin of the gas that fuels the X-ray source is unclear, since

87: almost all the gas in young clusters is rapidly blown away by the

88: strong winds of massive stars. One possibility for providing the gas

89: is the tidal disruption of a main sequence (MS) star of mass $\Ms$ and

90: radius $\Rs$ with periapse $r_p<r_t$, where $r_t = (M_\bullet/\Ms

91: )^{1/3}\Rs $ is the tidal radius. However, direct disruptions lead to

92: a short flare ($t_{\rm flare}\lesssim\,\mathrm{yr}$; Rees \cite{RE88};

93: Ulmer \cite{U99}; Ayal, Livio, \& Piran \cite{ALP00}), which is

94: incompatible with the $\sim 20\,\mathrm{yr}$ observation period of the

95: X-ray source. In this {\it Letter} we investigate a more gradual

96: process for feeding the IBH, namely the tidal capture of a MS star and

97: the subsequent Roche lobe overflow.

98: 

99: 

100: 

101: 

102: \section{Tidal capture Rate}\label{sec:rate}

103: 

104: A BH in a cluster with velocity dispersion $\sigma$ dominates the

105: potential within its radius of influence $r_a=GM_\bullet/\sigma^2$;

106: inside $r_a$ orbits are approximately Keplerian, and stars are

107: distributed according to a power law $n_\star\propto r^{-\alpha}$,

108: with $\alpha \approx 3/2$ (Bahcall \& Wolf \cite{BW76}; Baumgardt,

109: Makino, \& Portegies Zwart \cite{BMPZ03}). The cusp is truncated inside

110: some radius $r_\mathrm{in}$, e.g. $r_{\rm in}\sim (M_\bullet/\Ms )\Rs

111: $ where the rate of destructive collisions exceeds the two-body

112: relaxation rate (Frank \& Rees \cite{FR76}).

113: 

114: Stars can reach an orbit with periapse of order of the tidal radius by

115: angular momentum diffusion. When the star passes at $r_p$, an energy

116: $\Delta E_t(r_p)$ is invested in raising tides, causing the star to

117: spiral in ($r_p\!<\!3r_t$ is typically required for an appreciable

118: effect). The evolution of a tidally heated star is not

119: well-understood. Two extreme models of ``squeezars'' (stars that are

120: continually powered by tidal squeezing) were studied by Alexander \&

121: Morris (\cite{AM03}). ``Hot squeezars'' are heated only in their outer

122: layers and radiate their excess energy efficiently; they hardly

123: expand. ``Cold squeezars'' dissipate the tidal

124: energy in their bulk and puff up to giant size.

125: 

126: Our analysis is based on the following assumptions.

127: 

128: (1) The stars are ``hot squeezars'' (in \S\ref{sec:disc} we discuss

129: some consequences of relaxing this assumption).

130: 

131: (2) As long as the eccentricity $e$ is high,

132: \begin{equation} 

133: \label{eq:xie}	

134: 1-e=r_p/a<\xi_e\sim0.1,

135: \end{equation}

136: where $a$ is the orbital semi-major axis, the stellar structure is not

137: significantly affected by the tidal heating, and the tidal energy

138: dissipated per orbit,

139: \begin{equation} 

140: \Delta E_t(b)=\frac{G\Ms^2}{\Rs }\frac{T(b)}{b^6}\,,

141: \end{equation} 

142: is constant (Alexander \& Morris \cite{AM03}); here $b=r_p/r_t$, and

143: $T(b)$ is the tidal coupling coefficient, which depends on the stellar

144: structure and is a strongly decreasing function of $b$ (e.g. Press \&

145: Teukolsky \cite{PT77}). When the orbit decays to the point where

146: $1-e>\xi_e$, the tidal heating drops off until eventually the star

147: circularizes at $a\gtrsim r_t$ and $\Delta E_t\!=\!0$ (Hut

148: \cite{H80}).

149: 

150: (3) The star can survive as long as its tidal luminosity does not

151: exceed, to within order unity, its Eddington luminosity $L_E=1.3\times

152: 10^{38}\,\mathrm{erg\,s^{-1}}\Ms /M_\odot$,

153: \begin{equation} 

154: \label{eq:xiL}	

155: \Delta E_t/P< \xi_L L_E ,

156: \end{equation} 

157: where $P$ is the orbital period and $\xi_L \approx1$.

158: 

159: 

160: The tidal heating rate is highest when $P$ is shortest, just before

161: tidal heating shuts off when $a=b r_t/\xi_e$

162: (Eq. \ref{eq:xie}). Therefore, the Eddington luminosity limit

163: (Eq. \ref{eq:xiL}) corresponds to a minimal periapse

164: $b_{\mathrm{min}}r_t$ that a star can have and still circularize

165: without being disrupted, which is given implicitly by

166: \begin{equation}

167: \label{eq:bm}

168:   \Delta E_t(b_{\mathrm{min}}) = \xi_L L_E \frac{2\pi}{\sqrt{G

169:            M_\bullet}}\left(\frac{b_\mathrm{min}

170:            r_t}{\xi_e}\right)^{3/2}\,.

171: \end{equation}

172: When $r_p<b_{\rm min}r_t$, the star is evaporated by its own tidally

173: powered luminosity during in-spiral.

174: 

175: Stars within the ``loss-cone'', a region in phase space where stars

176: have periapse smaller than $r_t$ (Frank \& Rees \cite{FR76}; Lightman

177: \& Shapiro \cite{LS77}), are disrupted by the BH. Two-body scattering

178: sustains a flow of stars in angular momentum space towards the

179: loss-cone. During in-spiral, two-body interactions change the periapse

180: of the star. The time $t_p$ over which the periapse of a star is

181: changed by order unity due to many small angle deflections is

182: (Alexander \& Hopman \cite{AH03})

183: \begin{equation}

184: \label{eq:tp}

185: t_p(b, a)= \frac{br_t}{a} t_r\,,

186: \end{equation}

187: where $t_r$ is the relaxation time. Note that $t_r$ does not depend on the distance from the BH for $\alpha=3/2$.

188: 

189: The in-spiral time is the time it takes until the semi-major axis of the star becomes formally zero; for a hot squeezar it is (Alexander \& Hopman

190: \cite{AH03})

191: \begin{equation}

192:   t_0(b, a) = \frac{2\pi \Ms \sqrt{G M_\bullet a}}{\Delta E_t(b) }\,.

193: \end{equation}

194: If deflections increase the periapse, the dissipation becomes much

195: less efficient, while if the periapse decreases, the star may cross

196: $r_t$ and be disrupted. Either way it fails to

197: circularize. Circularization can happen only if the in-spiral time

198: $t_0$ is shorter than the time-scale for deflections, $t_p$. The

199: widest orbit $a_c(b)$ from which a star can still spiral in for

200: periapse $b r_t$ is given by $t_0(b, a_c)=3t_p(b, a_c)$ (Alexander \&

201: Hopman \cite{AH03}, Eq. [11] for $\alpha = 3/2$). It then follows from

202: Eq. (\ref{eq:bm}) that the maximal distance $a_{\mathrm{max}}$ from

203: which a star can originate to reach the tidal radius without being

204: destroyed is

205: \begin{equation}\label{eq:amax}

206:     a_{\mathrm{max}} = \left[\frac{3\Delta

207: 	E_t(b_\mathrm{min})b_\mathrm{min}r_tt_r}{2\pi \Ms \sqrt{G

208: 	M_\bullet}}\right]^{2/3}.

209: \end{equation}

210: Within $r_\mathrm{in}$ the cusp flattens and relaxation is

211: inefficient, so there are hardly any stars on eccentric orbits.  Since

212: $r_{\rm in}$ grows more rapidly with $M_\bullet$ than $a_{\rm max}$,

213: there exists a maximal BH mass $M_\mathrm{max}$, such that for

214: $M_\bullet>M_\mathrm{max}$, $a_{\mathrm{max}}<r_\mathrm{in}$, and

215: tidal capture is strongly suppressed. Fig. (\ref{fig:radii}) shows

216: $a_\mathrm{max}$ and $r_\mathrm{in}$ as a function of $M_\bullet$; for

217: the calculation of $t_r$ we assumed that the $M_\bullet - \sigma$

218: relation

219: \begin{equation}\label{eq:Msigma} 

220:   M_\bullet = 1.3\times10^8 M_\odot\left

221:       ( \frac{\sigma}{200\,\mathrm{km\,s^{-1}}}\right)^4

222: \end{equation}

223: (Ferrarese \& Merritt \cite{FM00}; Gebhardt et al. \cite{Gea00};

224: Tremaine et al. \cite{Tr02}) can be extended to IBHs (see e.g.,

225: Portegies Zwart \& McMillan \cite{PZMcM02}). Circularization is only

226: possible for $M_\bullet<M_{\rm max}\approx10^5M_\odot$.

227: 

228: \begin{figure}

229: \plotone{f1.eps} 

230: \caption{Dependence of $a_\mathrm{max}$ (solid), $r_\mathrm{in}$ (dashed), and $r_t$

231: (dashed-dotted) on the mass of the BH, for a $10\Mo$ star. Circularization is only

232: possible provided that

233: $a_\mathrm{max}>r_\mathrm{in}$. \label{fig:radii}}

234: \end{figure}

235: 

236: 

237: 

238: 

239: The rate $\Gamma$ at which stars diffuse into orbits that allow

240: successful circularization is given by (Eq. [9] Syer \& Ulmer

241: \cite{SU99})

242: \begin{equation}\label{eq:rate}

243:   \Gamma = \frac{(a_\mathrm{max}/r_a)^{3 - \alpha}N_a}

244: 	        {t_r\mathrm{ln}(2\sqrt{a_\mathrm{max}/b_\mathrm{min}r_t})}

245: 	        \,\,\,\,\,\,\,\,\,\,\,\,(a_{\mathrm{max}}>r_\mathrm{in}),

246: \end{equation}

247: where the logarithmic term expresses the depletion of the stellar

248: density near the loss-cone; $N_a$ is the number of stars within the

249: radius of influence. The rate is essentially independent on $\Ms$ for a fixed stellar mass within $r_a$, and it decreases only logarithmically

250: with $t_r$ (cf. Eqs [\ref{eq:amax}, \ref{eq:rate}]): a larger $t_r$

251: increases the volume of stars that contributes to $\Gamma$, but

252: decreases the rate at which stars enter the loss-cone. The rate does

253: not depend very sensitively on our assumptions: roughly,

254: $\Gamma\propto\xi_L/\xi_e$.

255: 

256: \section{Roche lobe overflow on the main sequence}\label{sec:RLOF}

257: Orbital angular momentum conservation implies that the circularization

258: radius is $a_{\rm circ}\!=\!2 b_{\rm min} r_t$. Efficient in-spiral

259: and successful circularization require $b_{\rm min}\!\sim\!2-2.5$, so

260: that $a_{\rm circ}\!\sim\!(4-5)r_t$. The onset of mass transfer

261: through the Roche lobe occurs when the distance between the IBH and

262: the star is $a_{\rm circ}\!\sim\!2 r_t$ (assuming $\Ms\!=\!10M_\odot$,

263: $\Ms/\Mbh\sim0.01$, Eggleton \cite{Eg83}). This is roughly a factor of

264: two smaller than the typical value of $a_{\rm circ}$. However, as it

265: evolves on the MS, a $10M_\odot$ star expands by a factor of

266: $\sim\!2.7$ by the time it reaches the terminal age MS (TAMS). This

267: implies that Roche lobe overflow (RLOF) occurs at some point on

268: the MS, and continues for a time $t_{\rm x}$, which is shorter than

269: the MS lifetime $t_{\rm MS}$ (we assume here that the star expands

270: significantly only after it has circularized). In the following

271: analysis we assume for simplicity that RLOF holds over the entire

272: MS. This does not affect our conclusions significantly since the

273: observationally relevant phase of high X-ray luminosity occurs toward

274: the TAMS. For massive MS stars RLOF is preceded by a less 

275: luminous phase resulting from the accretion of strong stellar winds.

276: 

277: Mass transfer from a MS star to an IBH is driven by the thermal

278: expansion of the donor and the loss of angular momentum from the

279: binary system.  Mass transfer then implies that the donor fills its

280: Roche lobe ($R_{\star} = R_{\rm Rl}$) and continues to do so

281: ($\Rdot_{\star} = \Rdot_{\rm Rl}$). We assume that as long as the

282: Eddington limit is not exceeded, all the mass lost from the donor via

283: the Roche lobe is accreted by the IBH ($\Mdot_{\bullet} =

284: -\Mdot_{\star}$). Otherwise, the mass in excess of the Eddington limit

285: is lost from the binary system.

286: 

287: The expansion of the donor on the MS is calculated using fits from

288: Eggleton, Tout \& Fitchett (\cite{ETF89}) to detailed stellar

289: evolution calculations. We assume that the evolution of the donor

290: was not affected by mass loss.  Variations in the Roche radius of the

291: donor can be computed from the redistribution of mass and angular

292: momentum in the binary system. The radius of the Roche-lobe is

293: estimated with the fitting formula from Eggleton (\cite{Eg83}).

294: 

295: 

296: We stop following the binary evolution at the TAMS; the simple model

297: for calculating the amount of mass transfer may be inappropriate for

298: the post-MS evolution of the donor, as the star then rapidly expands.

299: However, at the end of the MS the donor still has a considerable

300: envelope and the star ascends the giant branch. The post-MS

301: evolution is likely to result in a short ($t_\mathrm{PMS}<0.1\, t_{\rm

302: MS}$) phase in which the luminosity increases by more than an order of magnitude.

303: 

304: In Fig. (\ref{fig:Mdon}) we plot the mass of the donor as a function

305: of time. It is assumed here that there is Roche lobe contact directly

306: after circularization. As discussed, this actually only happens

307: when the star has evolved towards a later stage of the MS

308: phase, which is when the mass loss in the plot starts to drop more

309: rapidly.

310: 

311: \begin{figure}

312: %\epsscale{0.75}

313: %\plotone{f2.eps}

314: \includegraphics[angle=-90, width =

315: 0.45\textwidth]{f2.eps}

316: \caption{Mass evolution of the donor star, assuming Roche-lobe

317: contact at the zero-age MS to a $1400 M_\odot$ IBH.  The solid, dashed

318: and dotted curve are for a $15 M_\odot$, $10 M_\odot$, and $5 M_\odot$

319: donor.\label{fig:Mdon}}

320: \end{figure}

321: 

322: 

323: We estimate the X-ray luminosity during mass transfer with the model

324: discussed by K\"ording, Falcke, \& Markoff (\cite{KFM02}). They argue

325: that the X-ray luminosity is generated by an accretion disk. The

326: binary is in the hard state if $\Mdot> \Mdot_{\rm crit}$, in which

327: case $L_{\rm x} = \epsilon \Mdot c^2$. At lower accretion rates

328: $L_{\rm x} = \epsilon \Mdot c^2 \Mdot/\Mdot_{\rm crit}$, in which case

329: the X-ray source becomes transient (i.e. short outbursts, separated by long states of quiescence, Kalogera et al.\, 2003).  For

330: $\Mdot_{\rm crit}$ we adopt the equation derived by Dubus et al.\,

331: (\cite{DU99}, see Eq. 32) and assume $\epsilon = 0.1$.  These choices are

332: comparable to $\Mdot_{\rm crit} \sim 10^{-7} M_\odot\, \mathrm{yr}^{-1}$ of K\"ording

333: et al.\, (\cite{KFM02}).  The resulting X-ray luminosity is presented

334: in Fig.\,(\ref{fig:Lx}).  Note that lower mass donor binaries

335: ($\Ms \aplt 5$\,\msun) live longer, are less luminous and tend to show transient behavior, where

336: high mass donor binaries are more luminous, shorter lived, steady sources.

337: 

338: \begin{figure}

339: %\epsscale{0.75} 

340: %\plotone{f3.eps} 

341: \includegraphics[angle=-90, width =

342: 0.45\textwidth]{f3.eps}

343: \caption{X-ray

344: luminosity for a $1400 M_\odot$ accreting IBH as a function of time.

345: Line styles as in Fig.\,\ref{fig:Mdon}.  To the right side of the

346: figure we added $\dot{M}$ in logarithmic units of

347: $M_\odot\,{\rm yr^{-1}}$ for the regime where $\Mdot > \Mdot_{\rm

348: crit}$. The radius of the star grows significantly towards the TAMS

349: (where the lines for 15\msun\, and 10\,\msun\, donors rise). It is probably only near that point at which

350: RLOF actually starts, so that the luminosity is higher than would be

351: estimated from RLOF of a zero age MS star.  The 5\,\msun\, donor does

352: not show this rise in $L_{\rm x}$ as $\dot{M}<\Mdot_{\rm crit}$ after $\sim 10$\,Myr. This also causes the  drop of $L_{\rm x}$ for the 10\msun\, donor star at the end of its evolution. When  $\dot{M}<\Mdot_{\rm crit}$, the source becomes transient.

353:   \label{fig:Lx} }

354: 

355: \end{figure}

356: \section{Cluster MGG-11}\label{sec:MGG-11} 

357: We apply our analysis to the young dense star cluster MGG-11 in the

358: irregular galaxy M82, at a distance of $\sim 4$ Mpc. This cluster

359: contains the variable X-ray source M82-X7 with $L_{\rm x} = (0.8-160)

360: \times 10^{39}\, {\rm erg s^{-1}}$ (Watson, Stanger \& Griffiths

361: \cite{WSG84}; Matsumoto \& Tsuru \cite{MT99}; Kaaret et

362: al. {\cite{Ka01}). The velocity dispersion in the cluster is

363: accurately measured, $\sigma=11.4\pm0.8 \,\mathrm{km\, s^{-1}}$

364: (McCrady, Gilbert, \& Graham \cite{MGG03}). We assume that this cluster contains an

365: IBH which is also the engine for the X-ray source. If this IBH obeys

366: the $M_\bullet - \sigma$ relation (\ref{eq:Msigma}), its mass is

367: $M_\bullet = 1.4\times10^3M_\odot$, which is consistent with the

368: recent calculations of Portegies Zwart et al. (\cite{PZ04}), who show that an IBH could have formed dynamically in MGG-11 by a runaway merger of MS stars. Within

369: its radius of influence $r_a=0.05 \,\mathrm{pc}$, the number of stars

370: is $N_a=2M_\bullet/\Ms $ (Merritt \cite{Me03}) and $t_r\sim10^5-10^6\,\mathrm{yr}$.

371: 

372: The age of the cluster is $t_{cl}= (7-12)\,\mathrm{Myr}$,

373: corresponding to a turn-off mass of $17-25 M_\odot$ (Eggleton et

374: al. \cite{ETF89}). The mean stellar mass of the cluster at birth

375: is $\langle M_\star \rangle=3\,M_\odot$, but as a result of

376: mass-segregation the average mass within $r_a$ is much higher. The

377: direct N-body calculations of Portegies Zwart et al. (\cite{PZ04}) show that at an age of 7

378: Myr, the mean mass of the single stars in the core of MGG-11 is

379: $\langle M_\star \rangle = 8\pm3 M_\odot$. For simplicity we assume

380: within $r_a$ a single mass population of stars with $\Ms =10M_\odot$,

381: and radius $\Rs = 5.4 R_\odot$ (Gorda \& Svechnikov \cite{GS98}).  The

382: results of Fig. (\ref{fig:Lx}) show that a MS donor of mass $\Ms

383: \gtrsim10\,M_\odot$ can account for the luminosity of the ULX in

384: MGG-11.

385: 

386: We assume $(\xi_e, \xi_L)=(0.1, 0.5)$, and take the numerical values

387: for the function $T(b)$ for parabolic orbits from Alexander \& Kumar

388: (\cite{AK01}).  With these parameters we find a capture rate of

389: $\Gamma=5\times10^{-8}\, \mathrm{yr^{-1}}$. This implies that a

390: fraction $\Gamma\,t_{cl}t_{\rm x}/t_{\rm MS}= (30-50)\%$ of clusters

391: harboring an IBH has formed a tidal binary and may be observed during

392: RLOF in the MS phase. A fraction $\Gamma t_{cl}t_{\rm PMS}/t_{\rm

393: MS}\sim 4 \%$ of clusters with an IBH should be observed during the

394: more luminous post-MS phase.

395: 

396: \section{Summary and discussion}

397: \label{sec:disc}

398: 

399: MS stars can spiral into an IBH as a result of tidal capture and

400: circularize close to the tidal radius. This process is unique to IBHs,

401: since stars cannot survive tidal in-spiral around a MBH in a galactic

402: center. After circularization, the star expands on the MS until

403: high luminosity RLOF accretion starts toward the end of the MS

404: phase. We analyzed RLOF during the MS phase in some detail and

405: calculated the X-ray luminosity. Post-MS RLOF is harder to model, but

406: the resulting luminosity is expected to be at least an order of

407: magnitude brighter and about an order of magnitude shorter in

408: duration. The X-ray luminosity is consistent with observed ULXs, such

409: as MGG-11.

410: 

411: MGG-11 is the only cluster out of hundreds in M82 with a ULX. Possibly

412: other clusters were not sufficiently dense to form IBHs (Matsushita et

413: al. \cite{Mea00}; Portegies Zwart et al. \cite{PZ04}). If a fraction

414: $f_\bullet$ of the $N_{cl}$ clusters in M82 harbors an IBH, the number

415: of ULXs is estimated by $N_{\rm x} = f_\bullet \Gamma t_{cl}t_{\rm

416: x}/t_{\rm MS}$. Thus $f_\bullet$ has to be of the order of a few

417: percent in order to account for one ULX in M82.

418: 

419: In order to circularize, a star has to dissipate $\sim(M_\bullet/\Ms

420: )^{2/3}$ times its binding energy. If a certain fraction

421: $\delta$ of the energy is invested in bulk heating (for ``hot

422: squeezars'' $\delta=0$ as assumed so far, for ``cold squeezars'' $\delta=1$), the star expands. An X-ray binary can form only if $\delta\!<\!(\Ms

423: /M_\bullet)^{2/3}$. Nevertheless, a shorter lived ULX-phase is

424: still possible even if $\delta> (\Ms /M_\bullet)^{2/3}$. When the star

425: expands to a radius $>\!b\Rs$ it is gradually peeled every

426: periapse passage and feeds the IBH for a period much longer than

427: $t_{\rm flare}$. However, the process is limited by the two-body

428: deflection time-scale $t_p$, which is $\lesssim\!10^3\,\mathrm{yr}$

429: for a $10^3M_\odot$ IBH. This translates to a detection probability of

430: only $\Gamma t_p\!\sim\!5\!\times\!10^{-5}$, and so it is unlikely

431: that the ULX in MGG-11 originates in this type of process. For similar

432: reasons, it is improbable to observe a very luminous tidally heated

433: star (squeezar) during the final stages of its in-spiral into an IBH

434: in a stellar cluster (this may be possible for squeezars near the MBH

435: in the Galactic Center, where the in-spiral time is longer and the

436: capture rate higher, Alexander \& Morris \cite{AM03}).

437: 

438: The lifetime of the host cluster is limited by the galactic tidal

439: field, and can be as short as 100\,Myr (Portegies Zwart et

440: al. \cite{PZ01}). This is much shorter than the RLOF phase of a low mass

441: donor (e.g. $\sim$ Gyr for a $2 M_\odot$ star). Thus, the X-ray

442: life-time of a low-luminosity binary can be much longer than the

443: life-time of the cluster. Quiescent orphaned IBHs can suddenly light

444: up when their companion ascends the giant branch and starts to

445: transfer mass to the IBH. Our scenario predicts the existence of

446: host-less ULXs, which are more likely to be transient and less luminous. Their exact fraction in the ULX population cannot be

447: reliably estimated at this time. However, it is interesting to note

448: that $3-10$ out of $14$ of the ULXs in the Antennae Galaxies are

449: coincident with a stellar cluster, while the others are not (Zezas et

450: al. \cite{ZFRM02}).

451: 

452: 

453: 

454: \acknowledgements

455: We thank the referee, Monica Colpi, for comments that improved the manuscript.

456: This work is supported by the Royal Netherlands Academy of Sciences

457: (KNAW), the Dutch organization of Science (NWO), the Netherlands

458: Research School for Astronomy (NOVA), ISF grant 295/02-1, Minerva

459: grant 8484, and a New Faculty grant by Sir H. Djangoly, CBE, of

460: London, UK.

461: 

462: 

463: 

464: \begin{thebibliography}{}

465: 

466: \bibitem[ 2001]{AK01} Alexander, T., \& Kumar, P., 2001, \apj, 549,

467: 948

468: 

469: \bibitem[ 2003]{AH03} Alexander, T., \& Hopman, C., 2003, \apj, 590,

470: L29

471: \bibitem[ 2003]{AM03} Alexander, T., \& Morris, M., 2003, \apj, 590,

472: L25

473: 

474: \bibitem[ 2000]{ALP00}Ayal, S., Livio, M., \& Piran, T., 2000, \apj,

475: 545, 772

476: 

477: \bibitem[ 1976]{BW76} Bahcall, J. N., \& Wolf, R. A., 1976, \apj, 209,

478: 214

479: 

480: 

481: \bibitem[ 2003]{BMPZ03} Baumgardt, H., Makino, J., \& Portegies Zwart,

482: S., 2003, to be published in {\it Scientific Highlights of the IAU

483: XXVth General Assembly}, eds.  D. Richstone, P. Hut

484: 

485: 

486: 

487: \bibitem[ 1999]{DU99} Dubus, G., Lasota, J-P, Hameury, J-M, \& Charles, Ph., 1999, \mnras, 303, 139D

488: 

489: \bibitem[ 1983]{Eg83} Eggleton, P. P., 1983, \apj, 268, 368

490: 

491: \bibitem[ 1989]{ETF89} Eggleton, P. P., Tout, C. A., \& Fitchett,

492: M. J., 1989, \apj, 347, 998

493: 

494: \bibitem[ 2000]{FM00} Ferrarese, L., \& Merritt, D., 2000, \apj, 539,

495: L9

496: 

497: \bibitem[ 1976]{FR76} Frank, J., \& Rees, M. J., 1976, \mnras, 176,

498: 633

499: 

500: \bibitem[ 2000]{Gea00} Gebhardt, K., et al., 2000, \apj, 539, L13

501: 

502: \bibitem[ 1998]{GS98} Gorda, S. Yu., \& Svechnikov, M. A., 1998,

503: Astron. Rep. 42, 793

504: 

505: \bibitem[ 1980]{H80} Hut, P. 1980, \aap, 92, 167

506: 

507: \bibitem[ 2001]{Ka01} Kaaret, P., Prestwich, A. H., Zezas, A., Murray,

508: S. S., Kim, D.-W., Kilgard, R. E., Schlegel, E. M., \& Ward,

509: M. J. 2001, MNRAS 321, L29

510: 

511: \bibitem[ 2003]{Kea03} Kalogera, V., Henniger, M., Ivanova, N., \&

512: King, A. R., 2003, preprint (astro-ph/0309035)

513: 

514: 

515: \bibitem[ 2002]{KFM02} K\"ording, E., Falcke, H., \& Markoff, S.,

516: 2002, \aap, 382, L13

517: 

518: \bibitem[ 1977]{LS77} Lightman, A. P., \& Shapiro, S. L., 1977, \apj,

519: 211, 244

520: 

521: \bibitem[ 1999]{MT99} Matsumoto, H., \& Tsuru, T. G., 1999 ,PASJ, 51,

522: 321M

523: 

524: \bibitem[ 2000]{Mea00} Matsushita, S., Kawabe, R., Matsumoto, H.,

525: Tsuru, T. G., Kohno, K., Morita, K., Okumura, S. K., Vila-Vilar��, B.,

526: 2000, \apj, 545, L107

527: 

528: 

529: \bibitem[ 2003]{MGG03} McCrady, N., Gilbert, A., M., \& Graham, J.,

530: R., 2003, preprint (astro-ph/0306373)

531: 

532: \bibitem[ 2003]{Me03} Merritt, D., 2003, preprint (astro-ph/0301257)

533: 

534: 

535: \bibitem[ 1977]{PT77} Press, W. H., \& Teukolsky, S. A., 1977, \apj,

536: 213, 183

537: 

538: \bibitem[ 2001]{PZ01} Portegies Zwart, S. F., Makino, J., McMillan,

539: S. L. W., \& Hut, P., 2001, \apj, 546, L101

540: 

541: \bibitem[ 2004]{PZ04} Portegies Zwart, S. F., Baumgardt, H., Hut, P.,  Makino,

542: J., \& McMillan, S. L., 2004, \nat, in press

543: 

544: \bibitem[ 2002]{PZMcM02} Portegies Zwart, S.F., McMillan, S.L., 2002

545: \apj, 576, 899

546: 

547: \bibitem[ 1988]{RE88} Rees, M. J. 1988, \nat, 333, 523

548: \bibitem[ 1999]{SU99} Syer, D., \& Ulmer, A., 1999, \mnras, 306, 35

549: 

550: \bibitem[ 2002]{Tr02} Tremaine, S., et al., 2002, \apj, 574, 740

551: \bibitem[ 1999]{U99} Ulmer, A., 1999, \apj, 514, 180

552: 

553: \bibitem[ 1984]{WSG84} Watson, M. G., Stanger, V., \& Griffiths, R.,

554: E., 1984, \apj, 286, 144

555: 

556: 

557: \bibitem[ 2002]{ZFRM02} Zezas, A., Fabbiano, G., Rots, A. H., \&

558: Murray, S. S., 2002, \apj, 577, 710

559: 

560: \end{thebibliography}

561: 

562: \end{document}

563: 

564: