astro-ph0512642/ms.tex
1: \documentclass{emulateapj}
2: \newcommand{\Fewbody}{{\em Fewbody\/}}
3: 
4: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
5: \begin{document}
6: 
7: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
8: \title{Massive Black Hole Binaries from Collisional Runaways}
9: \shorttitle{Massive Black Hole Binaries from Collisional Runaways}
10: \submitted{To appear in \apjl}
11: \author{M. Atakan G\"urkan\altaffilmark{1}, John M. Fregeau\altaffilmark{2}, and Frederic A. Rasio\altaffilmark{3}}
12: \shortauthors{G\"urkan, et al.}
13: \affil{Department of Physics and Astronomy, Northwestern University, Evanston, IL 60208}
14: \altaffiltext{1}{{\tt ato.gurkan@gmail.com} Current address: Foundations 
15: Development, Sabanc\i\ University, 34956 \.Istanbul, Turkey.} 
16: \altaffiltext{2}{\tt fregeau@alum.mit.edu}
17: \altaffiltext{3}{\tt rasio@northwestern.edu}
18: 
19: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
20: \begin{abstract}
21: Recent theoretical work has solidified the viability of the collisional runaway
22: scenario in young dense star clusters for the formation of very massive stars (VMSs), 
23: which may be precursors to intermediate-mass black holes (IMBHs).  We present
24: first results from a numerical study of the collisional runaway process 
25: in dense star clusters containing primordial binaries.  Stellar collisions during
26: binary scattering encounters offer an alternate channel for runaway growth,
27: somewhat independent of direct collisions between single stars.  We find that
28: clusters with binary fractions $\mathpunct{\gtrsim}10\%$ yield {\em two} VMSs
29: via collisional runaways, presenting the exotic possibility of forming
30: IMBH--IMBH binaries in star clusters.  We discuss the implications
31: for gravitational wave observations, and the impact on cluster
32: structure.
33: \end{abstract}
34: 
35: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
36: \keywords{stellar dynamics --- globular clusters: general --- intermediate-mass black holes 
37:   --- methods: $n$-body simulations}
38: 
39: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
40: \section{Introduction}\label{sec:intro}
41: 
42: Observations hinting at the possible existence of intermediate-mass black holes (IMBHs)
43: have mounted in recent years.  Ultra-luminous X-ray sources---point X-ray
44: sources with inferred luminosities $\mathpunct{\gtrsim} 10^{39}\,{\rm erg}\,{\rm s}^{-1}$---may
45: be explained by sub-Eddington accretion onto BHs more massive than the maximum mass
46: of $\mathpunct{\sim}10M_\sun$ expected via core collapse in main sequence stars,
47: although viable alternative explanations exist \citep{2004IJMPD..13....1M}.
48: Similarly, the cuspy core velocity dispersion profiles of the globular clusters
49: M15 and G1 may also be explained by the dynamical influence of a central IMBH
50: \citep{2002AJ....124.3255V,2002AJ....124.3270G,2005ApJ...634.1093G}, although 
51: theoretical work suggests that the observations of M15
52: may be equally-well explained by a collection of compact stellar remnants in the 
53: cores of the clusters \citep{2003ApJ...582L..21B}.
54: 
55: At least three distinct IMBH formation mechanisms have been discussed in 
56: the literature.  The first, and possibly simplest,
57: is core collapse of a massive Pop III star \citep{2001ApJ...551L..27M}.
58: The very low metallicity of Pop III stars ($Z\lesssim 10^{-5}Z_\sun$) allows
59: much larger main-sequence stars to form, limits mass loss during stellar evolution, and increases
60: the fraction of mass retained in the final BH (for stars more massive than $\mathpunct{\sim}250M_\sun$)
61: \citep{2001ApJ...550..372F,2004IJMPD..13....1M}.  The second is the successive merging of 
62: stellar mass BHs via dynamical interactions, which may occur in star clusters that do 
63: not reach deep core collapse before $\mathpunct{\sim}3\,{\rm Myr}$, when the most massive cluster 
64: stars have become BHs \citep{1993Natur.364..421K,1993Natur.364..423S,2000ApJ...528L..17P,2002MNRAS.330..232M,oleary2006}.  The process is relatively 
65: inefficient---in terms of the amount of mass added to the growing BH per BH ejected
66: from the cluster---requiring BH seeds $\mathpunct{\gtrsim}500 M_\sun$ to create $10^3 M_\sun$ IMBHs 
67: \citep{2004ApJ...616..221G}, even when aided by the Kozai mechanism \citep{2002MNRAS.330..232M} 
68: or gravitational wave losses during close approaches \citep{gultekin2005}.
69: Introducing a mass spectrum for the BHs decreases the required seed mass,
70: although growth is still rare \citep{oleary2006}.
71: The third is the runaway merging of main-sequence stars via direct physical
72: collisions to form a very massive star (VMS), which may then collapse to form an IMBH
73: \citep{1999A&A...348..117P,2001ApJ...562L..19E,2002ApJ...576..899P,2004ApJ...604..632G}.  
74: Recent work shows that runaway growth of a VMS occurs generically in clusters with deep 
75: core collapse times shorter than $\mathpunct{\sim}3\,{\rm Myr}$ \citep{freitag2005b}.
76: 
77: With the exception of one simulation \citep{2004Natur.428..724P}, all
78: simulations of runaway collisional growth in clusters have ignored the effects of primordial
79: binaries, which are known to exist in clusters in dynamically significant
80: numbers \citep{1992PASP..104..981H}.  Indeed, some numerical results suggest that
81: the primordial binary fraction ($f_b$) may have to be nearly 100\% to explain the currently
82: observed binary fractions in cluster cores \citep{2005MNRAS.358..572I}.
83: Primordial binaries are an important piece of the runaway collisional growth
84: puzzle, since they introduce two effects which may strongly affect the process.
85: On the one hand, binaries generate energy via dynamical scattering interactions
86: in cluster cores, supporting the core against deep collapse and limiting
87: the maximum stellar density attainable, and hence limiting the direct stellar
88: collision rate \citep{2003gmbp.book.....H}.  On the other hand, stellar
89: collisions are much more likely in dynamical interactions of binaries, since
90: the interactions are typically resonant \citep{1996MNRAS.281..830B,2004MNRAS.352....1F}.
91: Since these two effects of primordial binaries act in opposite
92: senses with respect to the collision rate, it is not clear {\em a priori} how they
93: affect the collisional runaway scenario.
94: 
95: Before appealing to numerical methods, however, one can gain insight
96: into the effects of primordial binaries by considering the coagulation
97: equation \citep{1993ApJ...418..147L,2000Icar..143...74L,2001Icar..150..314M}---a 
98: simplification of which is presented in \citet{freitag2005a}---which describes
99: the evolution of a spectrum of masses due to mergers.  For growth to occur in a
100: runaway fashion, the coagulation equation requires that the cross section for 
101: collisions with the 
102: runaway object scales sufficiently rapidly with its mass: $S_{\rm coll} \propto M^\eta$, with
103: $\eta > 1$.  For single--single star collisions in star clusters in which the central
104: velocity dispersion is less than the escape speed from the surface of a typical star (so
105: that the cross section is dominated by gravitational focusing), this corresponds to 
106: the constraint $R \propto M^\alpha$ with $\alpha > 0$ on the main-sequence mass--radius 
107: relationship, which is satisfied by main-sequence stars of any mass or metallicity.
108: With some approximations, the coagulation equation analysis \citep{1975MNRAS.173..729H,1983ApJ...268..319H,1993ApJ...415..631S,gultekin2005} can be applied to collisions
109: occurring in binary scattering interactions.  From Fig.~4 of \citet{gultekin2005},
110: the cross section for close approach distances of $r_{\rm min}$ in binary--single
111: scattering encounters scales as $(S_{\rm coll}/\pi a^2) (v_\infty/v_c)^2 \propto (r_{\rm min}/a)^\gamma$, 
112: with $0.3 \lesssim \gamma \lesssim 1$, where $a$ is the binary semimajor axis, $v_\infty$
113: is the relative velocity between the binary and single star at infinity, and $v_c$ is the
114: critical velocity \citep[see {eq.~[12]} of][]{gultekin2005}.  Using the radius of the
115: runaway star for $r_{\rm min}$, $R \propto M^\beta$ with 
116: $0.5 \lesssim \beta \lesssim 1$ for the scaling of the mass--radius relation, and 
117: assuming that the binding energy of the binary is roughly preserved during the encounter
118: so that $a \propto M$, we find: 
119: \begin{equation}\label{eq:coag}
120:  S_{\rm coll} \propto M^{2 + \gamma (\beta - 1)}\,,
121: \end{equation}
122: with $0.3 \lesssim \gamma \lesssim 1$ and $0.5 \lesssim \beta \lesssim 1$.  The 
123: minimum of the exponent is $\mathpunct{\approx} 1.5$.  Thus according to the coagulation equation,
124: collisions induced in binary--single scattering interactions should yield runaway
125: growth of a VMS.  This result says nothing about the {\em rate} of growth of
126: the runaway object.  In other words, it is still not clear whether binary interactions
127: will limit the cluster core density such that the runaway timescale is longer
128: than the massive star main-sequence lifetime of $\mathpunct{\approx} 3\,{\rm Myr}$, in which
129: case the process would be halted.
130: 
131: Assuming that the cluster core density
132: reached is high enough for the runaway to proceed, it would appear that a
133: single binary is sufficient for a binary interaction-induced runaway to occur.
134: This is, of course, not the case, since binary scattering interactions
135: tend to destroy binaries.  Thus we expect that for sufficiently low $f_b$,
136: the runaway will be primarily mediated by single--single collisions.  For
137: sufficiently large $f_b$, the runaway will be primarily mediated by binary--binary
138: interactions (the analysis above assumes binary--single interactions---for
139: binary--binary the value of $\gamma$ will be smaller, still allowing a 
140: runaway by eq.~[\ref{eq:coag}]).  For intermediate $f_b$ it is possible that a binary interaction
141: induced-runaway could proceed until the core binary population is sufficiently
142: depleted that the cluster's core collapses.  If the first runaway is 
143: far enough from the center of the cluster when the core collapses, it 
144: is possible that a second runaway will be formed during core collapse,
145: mediated by single--single collisions.  The exotic possibility of forming 
146: two VMSs in a cluster, and thus two IMBHs, is a tantalizing one, with 
147: implications for gravitational wave observations and cluster dynamics.
148: 
149: In this letter we present first results from a study of the runaway collisional
150: scenario for the formation of VMSs in young dense clusters with primordial binaries.
151: We briefly describe our numerical method, present results showing the growth of
152: two runaways, and discuss the implications.
153: 
154: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
155: \section{Numerical Analysis}\label{sec:analysis}
156: 
157: 
158: %\subsection{Method and Initial Conditions}\label{subsec:method}
159: 
160: We use our Monte Carlo cluster code to simulate the evolution
161: of young star clusters with primordial binaries 
162: \citep{2000ApJ...540..969J,2003ApJ...593..772F,2004ApJ...604..632G}.
163: This code uses the H\'enon method for two-body relaxation,
164: and incorporates the \Fewbody\ $N$-body integrator to perform
165: dynamical scattering encounters of binaries.  Stellar collisions are handled
166: by assuming that stars whose surfaces touch merge with
167: no mass loss, an assumption that has been shown to be valid
168: for clusters with low velocity dispersion, such as globulars or
169: young dense clusters \citep{freitag2005a,freitag2005b}.  
170: Collisions are allowed to occur directly in single--single star encounters, 
171: and during binary interactions \citep{2004MNRAS.352....1F}.
172: 
173: \begin{figure}
174:   \begin{center}
175:     \includegraphics[width=\columnwidth]{f1.eps}
176:       \caption{Evolution of the cluster Lagrange radii (lower panel) in units of the 
177: 	initial cluster half-mass radius (left axis) and in parsecs (right axis),
178: 	and mass of the (single) runaway for a cluster with $10^6$ objects and 
179: 	$f_b=0.05$.  Time is shown both in Myr and in units of the central
180: 	relaxation time.\label{fig:lr605}}
181:   \end{center}
182: \end{figure}
183: 
184: For initial conditions we assume a $W_0=3$ King model density
185: profile for the cluster, with no initial mass segregation.  
186: All stars are main-sequence stars with masses in the range 
187: $0.2<M/M_\odot<120$, distributed
188: according to a Salpeter mass function.  Simulations show that
189: the primary condition required for a runaway is that
190: the core collapse time is shorter than the main-sequence
191: lifetime for the most massive stars, $\mathpunct{\sim}3\,{\rm Myr}$ \citep{freitag2005b};
192: and that for clusters of single stars
193: with a wide mass spectrum, the core collapse time is
194: always 
195: %\begin{equation}\label{eq:tcc}
196: $t_{\rm cc} \approx 0.15\, t_{\rm rc} (0)$,
197: %\end{equation}
198: (where $t_{\rm rc}(0)$ is the initial relaxation time in the core),
199: independent of cluster mass, size, or density profile
200: \citep{2004ApJ...604..632G}.  This allows us to set the central density
201: of the cluster such that the predicted core collapse time 
202: is either less or greater than 
203: $3\,{\rm Myr}$.  We perform simulations with binary 
204: fractions\footnote{The binary fraction is defined
205: such that $f_b$ is the fraction of {\em objects} in the cluster
206: that are binaries, an object being either a binary or a single
207: star.} up to $0.2$.  The binary population is created from a cluster
208: of single stars by adding secondary companions to randomly chosen
209: cluster stars, with the secondary mass chosen uniformly in the 
210: binary mass ratio.  The binary binding energy is distributed
211: uniformly in the logarithm, truncated at high energy so that
212: the binary members do not make contact at pericenter, and truncated
213: at low energy so that the orbital speed of the lightest member
214: in the binary is larger than the local stellar velocity dispersion
215: \citep[see][]{fregeau2005}.  The eccentricity is chosen according
216: to a thermal distribution, truncated at large $e$ so that the binary
217: members do not make contact at pericenter.  We use either
218: $N=5\times 10^5$ or $10^6$ total cluster objects, finding no
219: difference in the runaway results between the two.  We run all simulations
220: until $3\,{\rm Myr}$.  Our criterion for a runaway is that its final
221: mass is $\mathpunct{\gtrsim}500 M_\sun$.  In our runs the final masses of the VMSs 
222: are always at least twice this value.
223: 
224: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
225: %\subsection{Results}\label{subsec:results}
226: \begin{figure}
227:   \begin{center}
228:     \includegraphics[width=\columnwidth]{f2.eps}
229:     \caption{Merger trees for the three most massive stars at the end of the simulation
230:       for the model presented in Figure~\ref{fig:lr605}.  Red shows the most massive,
231:       intermediate green the second most, and blue the third most.  For each collision with the massive
232:       star, a symbol is plotted at its mass at the time of the 
233:       collision with lines drawn connecting to the symbols representing the star 
234:       participating in the merger.  A triangle represents a single--single collision, a 
235:       square a collision in a binary--single interaction, a pentagon a collision in
236:       a binary--binary interaction, and a circle a star that has not yet undergone a 
237:       collision.  As an example, the three intermediate green points at $t\approx 0.2\,{\rm Myr}$
238:       represent a collision in a binary--single interaction of a $\mathpunct{\approx} 100 M_\sun$
239:       star with a $\mathpunct{\approx} 0.3 M_\sun$ star, each having never undergone a collision
240:       previously.  For this model there is clearly just one VMS, created 
241:       primarily in single--single collisions.\label{fig:mt605}}
242:   \end{center}
243:   \vspace{-0.5cm}
244: \end{figure}
245: 
246: Figure~\ref{fig:lr605} shows the evolution of the cluster Lagrange
247: radii and the mass of the (single) runaway as a function of time
248: for a cluster with $t_{\rm cc} < 3\,{\rm Myr}$ %(as predicted by
249: %eq.~[\ref{eq:tcc}]) 
250: and $f_b=0.05$.
251: The evolution is typical for models with $f_b \lesssim 0.1$
252: and $t_{\rm cc} < 3\,{\rm Myr}$ in that: 1) the energy generated 
253: in binary interactions is not sufficient to postpone core collapse 
254: beyond $3\,{\rm Myr}$, and 2) there is a single runaway.
255: In the model shown in Figure~\ref{fig:lr605}, the runaway
256: grows to $\mathpunct{\approx} 2800 M_\sun$ before the cluster begins
257: to expand in response to the energy generated in collisions
258: with the runaway.  
259: 
260: We follow the membership of the collision products in binaries during
261: the collisions and binary interactions. This allows us to produce
262: merger trees to track the contributions to the formation of VMSs.
263: Figure~\ref{fig:mt605} shows the merger 
264: trees for the three most massive stars at the end of the simulation.
265: There is clearly only one runaway for this model, and it grows
266: primarily via direct, single--single collisions.
267: Figure~\ref{fig:mt610} shows the same as Figure~\ref{fig:mt605}
268: for the same model but with $f_b=0.1$.  In this model there are
269: two VMSs at the end of the simulation.  A binary interaction-induced 
270: collisional runaway begins at 
271: $t\approx 1.5\,{\rm Myr}$ and proceeds to $\mathpunct{\approx} 1.3 \times 10^3\,M_\sun$.  
272: A direct collision-induced runaway begins at $t\approx 2.3\,{\rm Myr}$, yielding
273: a second runaway, of mass $\mathpunct{\approx} 2.5\times 10^3\,M_\sun$.  The third most 
274: massive star at the end of the simulation has only grown to $\mathpunct{\approx} 400 M_\sun$.
275: The evolution is typical for models with $f_b \gtrsim 0.1$
276: and $t_{\rm cc} < 3\,{\rm Myr}$ in that there are two runaways: 
277: one via direct collisions, and one via collisions in binary interactions.
278: \begin{figure}
279:   \begin{center}
280:     \includegraphics[width=\columnwidth]{f3.eps}
281:     \caption{Same as Figure~\ref{fig:mt605} for the same model but with
282:       $f_b=0.1$.  A binary interaction-induced collisional runaway begins at 
283:       $t\approx 1.5\,{\rm Myr}$ and proceeds to $\mathpunct{\approx} 1.3 \times 10^3\,M_\sun$.  
284:       A direct collision-induced runaway begins at $t\approx 2.3\,{\rm Myr}$, yielding
285:       a second runaway, of mass $\mathpunct{\approx} 2.5\times 10^3\,M_\sun$.  The third most 
286:       massive star at the end of the simulation has only grown to $\mathpunct{\approx} 400 M_\sun$.
287:       \label{fig:mt610}}
288:   \end{center}
289:   \vspace{-0.5cm}
290: \end{figure}
291: 
292: Although we have not yet conducted a full parameter space survey,
293: we have mapped a significant slice in $f_b$--$t_{\rm rc}(0)$ space,
294: from $f_b=0.02$ to $0.2$ and from $t_{\rm cc}/(3\,{\rm Myr})=0.16$ to $28$.
295: Without exception, and independent of $f_b$, models with 
296: $t_{\rm cc}=0.15\,t_{\rm rc}(0) > 3\,{\rm Myr}$ show no runaways (7 models), while
297: those with $t_{\rm cc}=0.15\,t_{\rm rc}(0) < 3\,{\rm Myr}$ show either one or two (10 models).
298: This is in agreement with the results for clusters with no primordial
299: binaries, presumably since binary fractions smaller than 20\% are not
300: sufficient to postpone core collapse beyond $3\,{\rm Myr}$.
301: Of the models that show runaways, those with $f_b < 0.1$ always yield
302: only single runaways (4 models), while those with $f_b \geq 0.1$ always yield 
303: double runaways (6 models).  No models ever produce more than two 
304: runaways.  Due to the computational cost involved, we have not yet explored 
305: the $f_b>0.2$ region of parameter space.  
306: 
307: The numerical results presented here clearly agree with the
308: analytical arguments made above on the runaway nature of binary
309: interaction-induced collisions. Binary interaction-induced
310: runaways occur in the outskirts or just outside the core (typically
311: a few percents of a parsec for our models). This is a region where
312: densities, in particular the binary densities, are high enough to
313: start a runaway but the relaxation time is long enough that massive
314: objects do not rapidly sink to the center.  The rapid collapse near
315: the center leads to an expansion of these regions, further increasing
316: the relaxation time and prevents a prompt merger of the VMSs. 
317: 
318: Finally, we note the apparent disagreement between our results and those of 
319: \citet{2004Natur.428..724P}, who performed two direct $N$-body integrations
320: of models of the cluster MGG-11 with $f_b=0.1$ and found only single runaways.  
321: Our results predict that whenever $f_b \gtrsim 0.1$, and the cluster
322: central relaxation time is short enough so that a runaway is expected
323: in a corresponding cluster with only single stars, a double runaway
324: should result.  The models of \citet{2004Natur.428..724P} are
325: extremely centrally concentrated, with $W_0=12$, while our models
326: all use $W_0=3$.  From the discussion above, it is clear that if the 
327: density profile is sufficiently steep, the binary interaction-induced runaway will 
328: take place close enough to the center so that it will seed the subsequent
329: direct collision-induced runaway, yielding just one runaway.  
330: In addition, the condition $f_b \gtrsim 0.1$
331: for double runaways is only approximate.  Since the models of 
332: \citet{2004Natur.428..724P} are right on this boundary, it is possible
333: that a small difference in the initial conditions has prevented a double
334: runaway in their models.
335: 
336: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
337: \section{Discussion}\label{sec:discussion}
338: 
339: Although the process is somewhat uncertain, it is likely that 
340: the VMSs formed in young clusters via collisional runaways 
341: will undergo core-collapse supernovae
342: and become IMBHs on a timescale of $\mathpunct{\sim}1\,{\rm Myr}$ \citep{freitag2005b}.  
343: Our results show no evidence of 
344: the VMSs merging prior to becoming IMBHs.  After their separate formation,
345: the two IMBHs will quickly exchange into a common binary via dynamical
346: interactions.  The IMBH--IMBH binary will then shrink via dynamical friction
347: due to the cluster stars, to the point at which the stellar mass enclosed
348: in the binary is comparable to the binary mass.  This occurs on a timescale
349: $\mathpunct{\sim}t_r \langle m \rangle / M_{\rm IMBH}$, where $t_r$ is the
350: local relaxation time and $\langle m \rangle$ is the local average stellar
351: mass.  Since $\langle m \rangle / M_{\rm IMBH} \lesssim 10^{-2}$, this timescale
352: is likely to be $\mathpunct{\lesssim}10\,{\rm Myr}$.  The binary will then shrink
353: via dynamical encounters with cluster stars, the rate of which
354: is governed by loss-cone physics.  The timescale for the binary to shrink
355: to the point at which it merges quickly via gravitational
356: radiation energy loss is likely to be $\mathpunct{\lesssim}1\,{\rm Gyr}$ 
357: \citep{2003ApJ...599.1129Y,2005ApJ...618..426M}.  
358: 
359: Although IMBH--IMBH binaries do not merge in the LISA band---the gravitational wave frequency at
360: merger is $\mathpunct{\sim}1\,{\rm Hz}$---they do represent bright
361: sources that take at least $\mathpunct{\sim}10^6\,{\rm yr}$ to cross the LISA band.  
362: Their inspiral (chirp) signals
363: should be easily detectable by LISA out to a few tens of Mpc. Thus the
364: number of detectable IMBH binary sources may be quite large, since
365: most clusters are probably born with $f_b \gtrsim 0.1$, and
366: any cluster with mass $\mathpunct{\gtrsim}10^6 M_\sun$ and central relaxation time 
367: $\mathpunct{\lesssim}20\,{\rm Myr}$ will lead to a double runaway.
368: 
369: Significant core rotation (with rotational speed comparable to the local
370: velocity dispersion) is observed in the clusters M15, $\omega$ Cen, 47 Tuc, and G1
371: \citep[e.g.][]{2002AJ....124.3255V,2005ApJ...634.1093G}.
372: This rotation suggests the presence of a core angular momentum source, 
373: such as an IMBH binary \citep{2005MNRAS.364.1315M}.  
374: Similarly, observations of a millisecond pulsar in the halo of NGC 6752, and two
375: others in the core with high negative spin derivatives, hint at the existence
376: of an IMBH binary in the core \citep{2003ApJ...599.1260C}.
377: Since the IMBH--IMBH binaries formed via collisional runaways will merge 
378: within $\mathpunct{\sim}1\,{\rm Gyr}$ after formation, any angular momentum imparted to 
379: the cluster by the IMBH--IMBH binary will quickly diffuse out of the core on a 
380: core relaxation time.  An alternate mechanism must be at work in creating
381: the core rotation seen in some globular clusters today.
382: 
383: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
384: \acknowledgements
385: \vspace{-0.6cm}
386: We thank Marc Freitag, Shane Larson, and Cole Miller for many fruitful
387: discussions, and an anonymous referee for comments which improved the paper.
388: Some of the
389: simulations were performed on the Tungsten cluster at the National Center for 
390: Supercomputing Applications.  MAG thanks Sabanc\i{} University 
391: for their hospitality during the finalization of this letter.
392: This work was supported by NASA grant NNG04G176G.
393: 
394: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
395: \bibliographystyle{apj}
396: \begin{thebibliography}{33}
397: \expandafter\ifx\csname natexlab\endcsname\relax\def\natexlab#1{#1}\fi
398: 
399: \bibitem[{{Bacon} {et~al.}(1996){Bacon}, {Sigurdsson}, \&
400:   {Davies}}]{1996MNRAS.281..830B}
401: {Bacon}, D., {Sigurdsson}, S., \& {Davies}, M.~B. 1996, \mnras, 281, 830
402: 
403: \bibitem[{{Baumgardt} {et~al.}(2003){Baumgardt}, {Hut}, {Makino}, {McMillan},
404:   \& {Portegies Zwart}}]{2003ApJ...582L..21B}
405: {Baumgardt}, H., {Hut}, P., {Makino}, J., {McMillan}, S., \& {Portegies Zwart},
406:   S. 2003, \apjl, 582, L21
407: 
408: \bibitem[{{Colpi} {et~al.}(2003){Colpi}, {Mapelli}, \&
409:   {Possenti}}]{2003ApJ...599.1260C}
410: {Colpi}, M., {Mapelli}, M., \& {Possenti}, A. 2003, \apj, 599, 1260
411: 
412: \bibitem[{{Ebisuzaki} {et~al.}(2001){Ebisuzaki}, {Makino}, {Tsuru}, {Funato},
413:   {Portegies Zwart}, {Hut}, {McMillan}, {Matsushita}, {Matsumoto}, \&
414:   {Kawabe}}]{2001ApJ...562L..19E}
415: {Ebisuzaki}, T., {Makino}, J., {Tsuru}, T.~G., {Funato}, Y., {Portegies Zwart},
416:   S., {Hut}, P., {McMillan}, S., {Matsushita}, S., {Matsumoto}, H., \&
417:   {Kawabe}, R. 2001, \apjl, 562, L19
418: 
419: \bibitem[{{Fregeau} {et~al.}(2005){Fregeau}, {Chatterjee}, \&
420:   {Rasio}}]{fregeau2005}
421: {Fregeau}, J.~M., {Chatterjee}, S., \& {Rasio}, F.~A. 2005, \apj, in press
422:   (astro-ph/0510748)
423: 
424: \bibitem[{{Fregeau} {et~al.}(2004){Fregeau}, {Cheung}, {Portegies Zwart}, \&
425:   {Rasio}}]{2004MNRAS.352....1F}
426: {Fregeau}, J.~M., {Cheung}, P., {Portegies Zwart}, S.~F., \& {Rasio}, F.~A.
427:   2004, \mnras, 352, 1
428: 
429: \bibitem[{{Fregeau} {et~al.}(2003){Fregeau}, {G{\"u}rkan}, {Joshi}, \&
430:   {Rasio}}]{2003ApJ...593..772F}
431: {Fregeau}, J.~M., {G{\"u}rkan}, M.~A., {Joshi}, K.~J., \& {Rasio}, F.~A. 2003,
432:   \apj, 593, 772
433: 
434: \bibitem[{{Freitag} {et~al.}(2005{\natexlab{a}}){Freitag}, {G\"urkan}, \&
435:   {Rasio}}]{freitag2005b}
436: {Freitag}, M., {G\"urkan}, M.~A., \& {Rasio}, F.~A. 2005{\natexlab{a}}, \mnras,
437:   in press (astro-ph/0503130)
438: 
439: \bibitem[{{Freitag} {et~al.}(2005{\natexlab{b}}){Freitag}, {Rasio}, \&
440:   {Baumgardt}}]{freitag2005a}
441: {Freitag}, M., {Rasio}, F.~A., \& {Baumgardt}, H. 2005{\natexlab{b}}, \mnras,
442:   in press (astro-ph/0503129)
443: 
444: \bibitem[{{Fryer} {et~al.}(2001){Fryer}, {Woosley}, \&
445:   {Heger}}]{2001ApJ...550..372F}
446: {Fryer}, C.~L., {Woosley}, S.~E., \& {Heger}, A. 2001, \apj, 550, 372
447: 
448: \bibitem[{{Gebhardt} {et~al.}(2005){Gebhardt}, {Rich}, \&
449:   {Ho}}]{2005ApJ...634.1093G}
450: {Gebhardt}, K., {Rich}, R.~M., \& {Ho}, L.~C. 2005, \apj, 634, 1093
451: 
452: \bibitem[{{Gerssen} {et~al.}(2002){Gerssen}, {van der Marel}, {Gebhardt},
453:   {Guhathakurta}, {Peterson}, \& {Pryor}}]{2002AJ....124.3270G}
454: {Gerssen}, J., {van der Marel}, R.~P., {Gebhardt}, K., {Guhathakurta}, P.,
455:   {Peterson}, R.~C., \& {Pryor}, C. 2002, \aj, 124, 3270
456: 
457: \bibitem[{{G{\"u}ltekin} {et~al.}(2004){G{\"u}ltekin}, {Miller}, \&
458:   {Hamilton}}]{2004ApJ...616..221G}
459: {G{\"u}ltekin}, K., {Miller}, M.~C., \& {Hamilton}, D.~P. 2004, \apj, 616, 221
460: 
461: \bibitem[{{G\"ultekin} {et~al.}(2005){G\"ultekin}, {Miller}, \&
462:   {Hamilton}}]{gultekin2005}
463: {G\"ultekin}, K., {Miller}, M.~C., \& {Hamilton}, D.~P. 2005, \apj, in press
464:   (astro-ph/0509885)
465: 
466: \bibitem[{{G{\"u}rkan} {et~al.}(2004){G{\"u}rkan}, {Freitag}, \&
467:   {Rasio}}]{2004ApJ...604..632G}
468: {G{\"u}rkan}, M.~A., {Freitag}, M., \& {Rasio}, F.~A. 2004, \apj, 604, 632
469: 
470: \bibitem[Heggie(1975)]{1975MNRAS.173..729H} Heggie, D.~C.\ 1975, \mnras, 
471: 173, 729
472: 
473: \bibitem[{{Heggie} \& {Hut}(2003)}]{2003gmbp.book.....H}
474: {Heggie}, D. \& {Hut}, P. 2003, {The Gravitational Million-Body Problem} (Cambridge University Press)
475: 
476: \bibitem[{{Hut} {et~al.}(1992){Hut}, {McMillan}, {Goodman}, {Mateo}, {Phinney},
477:   {Pryor}, {Richer}, {Verbunt}, \& {Weinberg}}]{1992PASP..104..981H}
478: {Hut}, P., {McMillan}, S., {Goodman}, J., {Mateo}, M., {Phinney}, E.~S.,
479:   {Pryor}, C., {Richer}, H.~B., {Verbunt}, F., \& {Weinberg}, M. 1992, \pasp,
480:   104, 981
481: 
482: \bibitem[Hut \& Bahcall(1983)]{1983ApJ...268..319H} Hut, P., \& Bahcall, 
483: J.~N.\ 1983, \apj, 268, 319 
484: 
485: \bibitem[{{Ivanova} {et~al.}(2005){Ivanova}, {Belczynski}, {Fregeau}, \&
486:   {Rasio}}]{2005MNRAS.358..572I}
487: {Ivanova}, N., {Belczynski}, K., {Fregeau}, J.~M., \& {Rasio}, F.~A. 2005,
488:   \mnras, 358, 572
489: 
490: \bibitem[{{Joshi} {et~al.}(2000){Joshi}, {Rasio}, \& {Portegies
491:   Zwart}}]{2000ApJ...540..969J}\looseness-1
492: {Joshi}, K.~J., {Rasio}, F.~A., \& {Portegies Zwart}, S. 2000, \apj, 540, 969
493: 
494: \bibitem[Kulkarni et al.(1993)]{1993Natur.364..421K} Kulkarni, S.~R., Hut, 
495: P., \& McMillan, S.\ 1993, \nat, 364, 421 
496: 
497: \bibitem[{{Lee}(1993)}]{1993ApJ...418..147L}
498: {Lee}, M.~H. 1993, \apj, 418, 147
499: 
500: \bibitem[{{Lee}(2000)}]{2000Icar..143...74L}
501: ---. 2000, Icarus, 143, 74
502: 
503: \bibitem[{{Madau} \& {Rees}(2001)}]{2001ApJ...551L..27M}
504: {Madau}, P. \& {Rees}, M.~J. 2001, \apjl, 551, L27
505: 
506: \bibitem[{{Malyshkin} \& {Goodman}(2001)}]{2001Icar..150..314M}
507: {Malyshkin}, L. \& {Goodman}, J. 2001, Icarus, 150, 314
508: 
509: \bibitem[{{Mapelli} {et~al.}(2005){Mapelli}, {Colpi}, {Possenti}, \&
510:   {Sigurdsson}}]{2005MNRAS.364.1315M}
511: {Mapelli}, M., {Colpi}, M., {Possenti}, A., \& {Sigurdsson}, S. 2005, \mnras,
512:   364, 1315
513: 
514: \bibitem[{{Miller}(2005)}]{2005ApJ...618..426M}
515: {Miller}, M.~C. 2005, \apj, 618, 426
516: 
517: \bibitem[{{Miller} \& {Colbert}(2004)}]{2004IJMPD..13....1M}
518: {Miller}, M.~C. \& {Colbert}, E.~J.~M. 2004, Int. J. Mod.
519:   Phy. D, 13, 1
520: 
521: \bibitem[{{Miller} \& {Hamilton}(2002)}]{2002MNRAS.330..232M}
522: {Miller}, M.~C. \& {Hamilton}, D.~P. 2002, \mnras, 330, 232
523: 
524: %\bibitem[{{O'Leary} {et~al.}(2005){O'Leary}, {Rasio}, {Fregeau}, {Ivanova}, \&
525: %  {O'Shaughnessy}}]{oleary2005}
526: %{O'Leary}, R.~M., {Rasio}, F.~A., {Fregeau}, J.~M., {Ivanova}, N., \&
527: %  {O'Shaughnessy}, R. 2005, \apj, accepted (astro-ph/0508224)
528: \bibitem[O'Leary et al.(2006)]{oleary2006} O'Leary, R.~M., Rasio, 
529: F.~A., Fregeau, J.~M., Ivanova, N., \& O'Shaughnessy, R.\ 2006, \apj, 637, 
530: 937 
531: 
532: \bibitem[{{Portegies Zwart} {et~al.}(2004){Portegies Zwart}, {Baumgardt},
533:   {Hut}, {Makino}, \& {McMillan}}]{2004Natur.428..724P}
534: {Portegies Zwart}, S.~F., {Baumgardt}, H., {Hut}, P., {Makino}, J., \&
535:   {McMillan}, S.~L.~W. 2004, \nat, 428, 724
536: 
537: \bibitem[{{Portegies Zwart} {et~al.}(1999){Portegies Zwart}, {Makino},
538:   {McMillan}, \& {Hut}}]{1999A&A...348..117P}
539: {Portegies Zwart}, S.~F., {Makino}, J., {McMillan}, S.~L.~W., \& {Hut}, P.
540:   1999, \aap, 348, 117
541: 
542: \bibitem[Portegies Zwart \& McMillan(2000)]{2000ApJ...528L..17P} Portegies 
543: Zwart, S.~F., \& McMillan, S.~L.~W.\ 2000, \apjl, 528, L17 
544: 
545: \bibitem[{{Portegies Zwart} \& {McMillan}(2002)}]{2002ApJ...576..899P}
546: {Portegies Zwart}, S.~F. \& {McMillan}, S.~L.~W. 2002, \apj, 576, 899
547: 
548: \bibitem[Sigurdsson \& Hernquist(1993)]{1993Natur.364..423S} Sigurdsson, 
549: S., \& Hernquist, L.\ 1993, \nat, 364, 423
550: 
551: \bibitem[Sigurdsson \& Phinney(1993)]{1993ApJ...415..631S} Sigurdsson, S., 
552: \& Phinney, E.~S.\ 1993, \apj, 415, 631 
553: 
554: \bibitem[{{van der Marel} {et~al.}(2002){van der Marel}, {Gerssen},
555:   {Guhathakurta}, {Peterson}, \& {Gebhardt}}]{2002AJ....124.3255V}
556: {van der Marel}, R.~P., {Gerssen}, J., {Guhathakurta}, P., {Peterson}, R.~C.,
557:   \& {Gebhardt}, K. 2002, \aj, 124, 3255
558: 
559: \bibitem[{{Yu} \& {Tremaine}(2003)}]{2003ApJ...599.1129Y}
560: {Yu}, Q. \& {Tremaine}, S. 2003, \apj, 599, 1129
561: 
562: \end{thebibliography}
563: 
564: \clearpage
565: 
566: \clearpage
567: 
568: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
569: 
570: \end{document}
571: