0804.2299/ms.tex
1: \documentclass[12pt,preprint]{aastex}
2: 
3: %\usepackage{graphics}
4: 
5: %% manuscript produces a one-column, double-spaced document:
6: 
7: %\documentclass[manuscript]{aastex}
8: %% preprint2 produces a double-column, single-spaced document:
9: 
10: %\documentclass[preprint2]{aastex}
11: 
12: %% Sometimes a paper's abstract is too long to fit on the
13: %% title page in preprint2 mode. When that is the case,
14: %% use the longabstract style option.
15: 
16: %\documentclass[preprint2,longabstract]{aastex}
17: 
18: \newcommand{\be}{\begin{equation}}
19: 
20: \newcommand{\ee}{\end{equation}}
21: 
22: %% You can insert a short comment on the title page using the command below.
23: %\slugcomment{Not to appear in Nonlearned J., 45.}
24: 
25: %% If you wish, you may supply running head information, although
26: %% this information may be modified by the editorial offices.
27: %% The left head contains a list of authors,
28: %% usually a maximum of three (otherwise use et al.).  The right
29: %% head is a modified title of up to roughly 44 characters.
30: %% Running heads will not print in the manuscript style.
31: 
32: \shorttitle{Runaway Merging in Cluster with Mass Segregation}
33: 
34: \shortauthors{Ardi, Baumgardt and Mineshige}
35: 
36: %% This is the end of the preamble.  Indicate the beginning of the
37: %% paper itself with \begin{document}.
38: 
39: 
40: 
41: \begin{document}
42: 
43: %% LaTeX will automatically break titles if they run longer than
44: %% one line. However, you may use \\ to force a line break if
45: %% you desire.
46: 
47:   
48: 
49: \title{The influence of initial mass segregation on the runaway merging of stars}
50: 
51: %% Use \author, \affil, and the \and command to format
52: %% author and affiliation information.
53: %% Note that \email has replaced the old \authoremail command
54: %% from AASTeX v4.0. You can use \email to mark an email address
55: %% anywhere in the paper, not just in the front matter.
56: %% As in the title, use \\ to force line breaks.
57: 
58: 
59: \author{Eliani Ardi,\altaffilmark{1} Holger Baumgardt,\altaffilmark{2} and
60:   Shin Mineshige \altaffilmark{3}}
61: \altaffiltext{1}{Kyoto International University, 610-0311 Kyoto, Japan.}
62: 
63: \altaffiltext{2}{Argelander Institute for Astronomy, University of Bonn, D-53121 Bonn, Germany}
64: \altaffiltext{3}{Yukawa Institute for Theoretical Physics, Kyoto University, 606-8502 Kyoto, Japan.}
65: 
66: \begin{abstract}
67: We have investigated the effect of initial mass segregation on the runaway
68: merging of stars. The evolution of multi-mass, dense star clusters was 
69: followed by means of direct N-body simulations of up to 131.072 stars. 
70: All clusters started from King models with dimensionless central potentials 
71: of $3.0 \le W_0 \le 9.0$. Initial mass segregation was realized by varying 
72: the minimum mass of a certain fraction of stars whose either (1) distances 
73: were closest to the cluster center or (2) total energies were lowest. 
74: The second case is more favorable to promote the runaway merging of stars 
75: by creating a high-mass core of massive, low-energy stars. 
76: 
77: Initial mass segregation could decrease the central 
78: relaxation time and thus help the formation of a high-mass core.  However, 
79: we found that initial mass segregation does not help the runaway stellar 
80: merger to happen if the overall mass density profile is kept constant.
81: This is due to the fact that the collision rate of stars is not increased 
82: due to initial mass segregation. Our simulations show that initial mass 
83: segregation is not sufficient to allow runaway merging of stars to occur 
84: in clusters with central densities typical for star clusters in the Milky Way. 
85: 
86: 
87: %{\noindent \emph{Draft of March 24,2008}}
88: 
89: 
90: \end{abstract}
91: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
92: 
93: \keywords{ stellar dynamics --- globular clusters: general --- methods: n-body simulations}
94: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
95: 
96: \section{Introduction}
97: 
98: The discovery of point-like, ultra-luminous X-ray (ULX) sources with 
99: luminosities larger than $L_X > 10^{40}~\rm{ergs~s^{-1}}$ 
100: by the {\it Chandra} satellite \citep{mat01,kaa01}, corresponding to a few 
101: hundred $M_{\odot}$ black holes (BHs) if the sources are
102: not beamed and accrete at the Eddington rate, 
103: could be a first hint for the
104: existence of so called intermediate-mass black holes (IMBH). IMBHs would
105: bridge the gap between stellar-mass BHs which form as the end-product
106: of normal stellar evolution and the supermassive BHs observed at the
107: centers of galaxies. The connection between IMBHs and ULX is also supported by 
108: quasi-periodic oscillations in the X-ray spectrum found in some of the sources
109: \citep{str03,fio04}.
110: 
111: Several additional arguments have also suggested the presence of IMBHs in
112: globular clusters (see \citet{bau05} for a review), such as (1) the extrapolation
113: of the $M_{BH}-M_{bulge}$ relation found for supermassive black holes in 
114: galactic nuclei \citep{mag98} (2) the analysis of the central velocity dispersion in the globular clusters
115: \object{M15} \citep{ger02} and \object{G1} \citep{geb05} (3) N-body
116: simulations of runaway merging of stars in young star clusters in \object{M82} \citep{por04}.
117: 
118: How IMBHs can form is still an open question. \citet{ebi01} proposed a
119: scenario in which IMBHs form through successive merging of massive stars in
120: dense star clusters. In a dense enough cluster, mass segregation of massive
121: stars is faster than their stellar evolution and the massive stars sink into
122: the center of the cluster by dynamical friction and form a dense inner
123: core. In the inner core, massive stars undergo a runaway merging process and 
124: a very massive star forms with a mass exceeding several 100 solar masses. 
125: 
126: A recent study of collisionally merged massive stars by \citet{suz07} showed that the merger products
127: return to an equilibrium state on a Kelvin-Helmholtz timescale
128: and then evolve like single homogeneous stars with corresponding mass and
129: abundance. The final fate of the very massive stars will depend on the assumed
130: mass loss rate, but IMBH formation is one possible outcome \citep{bel07}. 
131: 
132: Direct N-body simulations of star clusters with up to 65536 stars by
133: \citet{por02} showed that runaway merging can cause the formation of a massive
134: star with up to 0.1\% of the total cluster mass before it turns into an
135: IMBH. Formation of very massive stars, as progenitors of IMBHs through runaway
136: collisions in young star clusters has also been studied recently by
137: \citet{fre06a,fre06b} using Monte Carlo simulations. Utilizing a large number of
138: particles ($10^6$ -- $10^8$ particles), they found that runaway
139:   collisions could lead to formation of very massive stars with masses $\ge$
140: 400 $M_{\odot}$.
141:  
142: Runaway merging of stars in the star cluster \object MGG-11 in the
143: starburst galaxy M82, whose position is consistent with a luminous X-ray
144:   source, has been intensively examined by \citet{por04}. They reported that
145: \object MGG-11 can host an IMBH if its initial dimensionless central potential
146: was high enough. A dimensionless central-potential $W_0 \ge 9.0$ was
147: required for runaway growth through collisions to form an IMBH. Unfortunately,
148: such a high dimensionless central potential leads to a central density $\rho_c
149: \ge 10^6 M_\odot/{\rm pc}^3$ which is rarely seen in present-day star
150: clusters, implying that the formation of IMBHs in star clusters is a very rare
151: event.
152: 
153: One possible way which would allow runaway collisions to occur in clusters
154: with lower central density is the assumption of initial mass
155: segregation. Initial mass segregation, which allows massive stars to start
156: their life in the cluster center, might be a way to lower the density
157: requirement for the onset of runaway collisions. The
158: tendency for massive stars to form preferentially near the cluster center is
159: expected as a result of star formation feedback in dense gas clouds
160: \citep{mur96} and from competitive gas accretion onto protostars and mutual
161: mergers between them \citep{bon02}. Observational evidence for initial
162: mass segregation in globular clusters as well as in open clusters has also
163: been reported \citep{bon98,deg04}.
164: 
165: Dynamical evolution of young dense star clusters with initial mass segregation
166: until the onset of core-collapse stage has been studied by \citet{gur04} by
167: using Monte Carlo simulations. Besides decreasing the core collapse time, they
168: found that initial mass segregation applied in clusters with $N = 1.25 \times 10^6$ stars which
169: followed a Plummer density profile initially, results  
170: in a total mass of the collapsed core of about 0.2 \% of the total cluster mass.
171: 
172: Motivated by results of \citet{por04}, that without initial mass segregation,
173: the dense star cluster MGG-11 could experience runaway merging only if the
174: central density was higher than $10^6 M_\odot/{\rm pc}^3$, in the present
175: study we want to explore whether or not initial mass segregation could lower
176: the density required for runaway collisions in \object MGG-11 like
177: clusters.  For this purpose, we perform $N$-body
178: simulations of \object MGG-11 like clusters starting from different initial
179: conditions which are described in detail in the next section. Results and
180: analysis of our simulations are shown in section~3 while the discussion and
181: conclusions are presented in section~4.
182: 
183: \section{Details of numerical simulations}\label{sec:ns}
184: 
185: We have conducted a number of $N$-body simulations, using the collisional
186: $N$-body code NBODY4 (Aarseth 1999) on the GRAPE-6 special purpose computers
187: provided by ADC - CfCA NAO Japan, to follow the evolution of multi-mass star
188: clusters. All simulations are run for a time span of 3 Myrs by which time 
189: we assume that the
190: runaway stars are turned into BHs and stop the simulations.
191: 
192: Our clusters contain 131.072 stars initially, distributed according to a
193: Salpeter IMF with minimum mass and maximum mass equal to 1.0 
194: $M_{\odot}$ and 100 $M_{\odot}$ respectively, which is chosen to fit the
195: \citet{mcc03} observations for MGG-11. Stellar evolution is modeled   
196: according to \citet{hur00}. Since we only follow the first 3 Myrs of cluster evolution, 
197: stellar evolution is important only for the most massive stars. Two stars are assumed to 
198: 'collide' if the distance between them becomes smaller than the sum of 
199: their radii. We assume that the total mass of both stars ends in the merger 
200: product and do not follow the stellar evolution of the runaway stars. We 
201: examine the evolution of \citet{kin66} models with central 
202: concentration $3.0 \le W_0 \le 9.0$. The initial half-mass 
203: radius and total cluster mass are chosen similar to what \citet{por04} chose 
204: to fit the observed parameters of MGG-11, namely $r_{h} = 1.3~\rm{pc}$ and $M
205: = 3.5 \times 10^5 M_{\odot}$. Details of the simulated
206: clusters without initial mass segregation are presented in Table~1.
207: 
208: In order to examine the effect of initial mass segregation, we study   
209: two scenarios. In the first scenario, we vary the minimum mass 
210: $m_{min}$ within the lagrangian radius containing 5\% of the total 
211: cluster mass ($R_{005}$). Increasing the minimum mass $m_{min}$ within 
212: $R_{005}$ (from 1 $M_\odot$ to $m_{min} > 1 M_{\odot}$ for clusters with 
213: initial mass segregation), while keeping the total cluster mass and energy 
214: constant, will consequently decrease the number of stars within this 
215: sphere. This scenario allows massive stars to start their life in the 
216: cluster center. It is proposed to meet observations which show that massive 
217: stars are preferentially formed near the cluster center \citep{bon98,deg04}. 
218: Details of runs where mass segregation 
219: is introduced inside a certain radius are given in Table~2.
220: 
221: In the second scenario, we choose a certain fraction of stars 
222: (whose total mass is 5 \% -- 20 \% of total mass of the cluster) with 
223: the lowest total energy 
224: and then vary the minimum mass of them, while keeping the total cluster 
225: mass and energy constant. The number of stars is again lower than in a normal 
226: cluster. Compared to the first scenario, the second scenario brings 
227: massive stars even closer to the center since stars located in the center at
228: time t=0 could still have high energies and spent most of their life 
229: outside the center. Hence support for runaway collisions should be stronger 
230: in the second scenario.
231: 
232: We also vary the half-mass radius of the clusters to see the effect of 
233: different central densities.
234: Table~3 reports details for clusters with initial mass segregation,  using
235:   the second scenario.
236: 
237: \section{Results and Analysis}\label{sec:sim}
238: 
239: \subsection{Clusters without initial mass segregation}
240: 
241: We run five cluster models without initial mass segregation as shown 
242: in Table~1. Each cluster contains 131.072 stars, but has different $W_0$. 
243: Four of them are set to have the same half-mass radius, which is 1.3 pc, 
244: to mimic MGG-11. In addition, we also examine a $W_0$ = 7.0 cluster 
245: with a smaller half-mass radius of $r_h$ = 0.5 pc. The central density of 
246: each cluster refers 
247: to the density within the core radius of the cluster, which is determined
248: with the method of \citet{cas85}.
249: For clusters with the same $r_h$, the central density is higher 
250: for clusters with higher dimensionless central potential $W_0$. 
251: 
252: We also calculate the central relaxation time of each cluster to study the 
253: influence of this parameter on the occurrence of runaway merging. 
254: The central relaxation time $T_{rel,c}$ is defined as \citep{spi87}:
255: \begin{equation}
256: T_{\rm rel,c}= \frac{\sigma^3_{3D}}{4.88 \pi G^2 \ln(0.11 N)n \langle m
257:   \rangle^2} ,
258: \end{equation}
259: where $\sigma_{3D}$, $n$ and $\langle m\rangle$ are the three-dimensional
260: velocity dispersion, number density and average stellar mass at the cluster
261: core. Here the cluster core refers to the region inside the core radius
262: $\rm r_{core}$. 
263: 
264: Our simulations of MGG-11 like clusters (with $r_h$ = 1.3 pc) show (see
265: Table~1)  that 
266: only the star cluster with the highest dimensionless central potential ($W_0$ = 9.0, 
267: corresponding to a central density of 3.24 $\times 10^6~M_{\odot}/{\rm{pc^3}}$),
268: experiences runaway merging. This result is in a good agreement with 
269: the one found by \citet{por04}. Our result again proves that high central 
270: density is required to allow runaway merging to occur. Collisions among
271: massive stars also occur in the lower density cluster but none of them 
272: experiences subsequent collisions leading to a super-massive star.
273: 
274: Fig.~1 depicts the evolution of lagrangian radii containing 1\% -- 20 \%
275:   of total mass of cluster models~1~--~3. Core radii ($\rm r_{core}$), which
276:   are marked by bold lines, are calculated according to \citet{cas85}. 
277: The inner shells of the $W_0= 9.0$ cluster (model~1) suffer strong contractions 
278: due to the high central density. Core collapse happens in this cluster at 
279: $t \approx$ 0.6 Myrs. The core collapse supports the runaway merging to
280: happen since runaway merging sets in at $t = $ 0.54 Myrs, about the same
281: time when core collapse happens (see the 10th column of Table~1.). Inner shells of $W_0$ = 7.0 cluster 
282: (model~2), on the other hand, contract very slowly. Even until 3 Myrs, the contraction is not 
283: strong enough to produce core collapse. Consequently, no runaway merging 
284: occurs in this cluster. Evolution of inner shells of $W_0$ = 7.0 cluster 
285: however looks different when we decrease $r_h$ to 0.5 pc (model~3). 
286: Mild contraction brings the cluster to collapse. Core collapse occurs at 
287: $t \approx $ 2.6 Myrs. At the same time, the first collision
288: leading to runaway merging happens ($t = $ 2.55 Myrs, see the 10th column of
289:   Table~1). 
290: Although the runaway merging started later than in the $W_0 =$ 9.0 
291: cluster, three collisions are enough to form a super-massive star  
292: with few hundred $M_\odot$ (see columns 9 and 11 of Table~1). 
293: 
294: The two clusters which experience runaway merging (models~1 and ~3), 
295: have very high central densities. The $W_0$ = 7.0 model has 
296: $\rho_c$ = 2.95 $\times 10^6~M_{\odot}/\rm{pc^3}$ while 
297: the $W_0$ = 9.0 model has $\rho_c$ = 3.24 $\times 10^6~M_{\odot}/\rm{pc^3}$. 
298: Runaway merging does not occur in clusters whose central densities are lower 
299: than $10^6~M_{\odot}/\rm{pc^3}$. 
300: Therefore the critical density which allows clusters without initial mass 
301: segregation to experience runaway stellar merging should be larger than 
302: $10^6 M_{\odot}/\rm{pc^3}$. This limit holds for globular-cluster-size 
303: objects with masses of $10^5 M_{\odot}$. 
304: 
305: Since in our runs the central density is varied, we find that
306: the central relaxation time (see column 6 of Table 1) mainly depends on the
307: number density of stars in the center, where $T_{rel,c} \propto n^{-1}$ (see
308: eq. 1). The central relaxation time is hardly affected by the change of
309: velocity dispersion $\sigma$ and average mass $\langle m \rangle$ (on average
310: $\sigma \approx$ 27.9 km/s and $\langle m \rangle \approx$ 2.64 $M_{\odot}$.) A high number density of stars in the cluster center seems to be required
311: to support runaway stellar merger in a cluster without initial mass
312: segregation.
313: 
314: Our result, that runaway merging does not occur in clusters with too low central
315:   density, is in good agreement with the one found by
316:   \citet{fre06a,fre06b}. Fig.~1 of \citet{fre06b} (which is essentially the
317:   same as Fig.~1 of \citet{fre06a}) shows that a cluster with mass 3 $\times$
318:   $10^5 M_{\odot}$ and dimensionless central potential $W_0 = 8.0$ experiences
319:   runaway collisions if its N-body length unit $(R_{NB})$ $\le$ 2 pc. This
320:   value corresponds to an initial half mass radius $R_h \le$  1.74 pc (see 
321:   sec.~2.1 of their paper where they show that $R_{NB} \simeq 1.15~R_h$ for $W_0 =
322:   8.0$). This value is not too far from the critical value we find, since our
323:   simulations show that a $W_0 = 7.0$ cluster without initial mass segregation experiences
324:   runaway collisions when its initial half-mass radius is somewhere between 1.3
325:   pc and 0.5 pc (see models~2 and 3 in Table~1), while a $W_0 = 9.0$ cluster
326:   with initial half-mass radius 1.3 pc experiences runaway collisions.
327: 
328: %In the following sections we will study if this limit is lowered if we
329: %introduce initial mass segregation.
330: 
331: \subsection{Clusters with initial mass segregation}
332: 
333: In models 6 -- 8, we introduce initial mass segregation by replacing stars
334: within the 
335: 5 \% lagrangian radius ($R_{005}$) with massive stars whose masses are higher 
336: or equal than the mass $m_{min}$ written in the 6th column of Table~2.
337: Replacing is done by randomly selecting new positions and velocities
338: for the massive stars from the positions and velocities of innermost stars. The
339: number of massive stars is chosen such that the overall mass density
340: profile remains constant.
341: 
342: As we keep the mass within the $R_{005}$ lagrangian radius constant, 
343: introducing initial mass segregation by increasing $m_{min}$ means to 
344: increase the average mass $\langle m\rangle$ of stars and lower the 
345: total number of stars (see the 3rd column of Table~2). The increase of
346: $\langle m\rangle$ in this region consequently decreases the central 
347: relaxation time $T_{rel,c}$. The central relaxation time of these clusters
348: should be lower than the one of a $W_0$ = 7.0 cluster without initial mass
349: segregation (see column 6 of model~2 in Table~1). As the central parts of these clusters
350: relax faster, the clusters may evolve faster and core collapse could
351: happen earlier. One may therefore expect that runaway merging should
352: now occur at lower central densities.
353: 
354: The top part of Fig.~2 shows that model~6 (with $m_{min} =
355:   30~M_{\odot}$) does not experience core collapse before 3 Myrs. 
356: Even increasing $m_{min}$ up to
357: $90~M_{\odot}$, as in model~8, does not
358: lead the cluster to experience core collapse before 3 Myrs either
359: (see bottom part of Fig.~2). Our simulations also show that no
360: runaway merger occurs in these clusters.
361: The reason why runaway merging does not happen is that
362: massive stars, which start their life in the region within $R_{005}$ do not 
363: constantly stay there. Some of these massive stars, whose initial
364: velocities are high enough, leave this region. Since the cluster is initially
365: mass segregated, this outward movement of massive stars is not balanced by a sufficiently
366: large number of massive stars moving inward, hence the average mass of stars 
367: decreases in the center. We note however that, since our clusters are started in
368: virial equilibrium, the expansion of the
369: high-mass stars is balanced by a corresponding number of low-mass stars moving further in, so that
370: the central density remains constant. 
371: 
372: The depletion of massive stars from the initial $R_{005}$ is shown for model~6 in 
373: Fig.~3. This figure depicts the evolution of lagrangian radii of 
374: massive stars whose masses are higher than 30 $M_{\odot}$ and that started 
375: their life inside $R_{005}$. Total mass fraction of these massive stars is
376: indicated by $M_{005}$. The lagrangian radii containing between 10 \% up to 
377: 100 \% of these stars are presented. 
378: The upper figure shows the change of lagrangian radii within the first 
379: 0.05 Myrs. We can see that within a few core crossing times ($
380:   t_{cross} \approx 8
381:   \times 10^3 \rm{yrs}$ ) some of these 
382: massive stars leave the initial $R_{005}$. At t=0.05 Myrs, total mass of massive stars 
383: which still reside inside this region is only 60 \% of the initial mass 
384: $M_{005}$. Bottom figure shows that 
385: up to t=3 Myrs, this region contains only about 30 \% of total mass 
386: of these massive stars.
387:     
388: 
389: Increasing
390: the minimum mass of stars whose distances are closest to
391: the cluster center does not succeed to produce high-mass
392: cores. In order to keep massive stars in the cluster core, we used a second 
393: scenario where initial mass
394: segregation is realized by varying the minimum mass of a certain fraction
395: of stars whose total energies are lowest. Since the second scenario is more
396: favorable to create a high-mass core of massive, low-energy stars, we will base
397: our results on this scenario.
398: 
399: In the second scenario, initial mass segregation was
400: introduced by replacing
401: stars which have the lowest total energy, up to 5 \% -- 20 \% of the total 
402: mass of the cluster (models 9 -- 15, see $M_{IMS}$ in column~5 of Table~3)
403: with massive
404: stars whose masses are higher than $m_{min}$. The coordinates and 
405: velocities of massive stars are randomly chosen from the stars with the lowest total energy
406: and their total number is again adjusted such to keep the overall
407: mass density profile constant and the cluster in virial equilibrium.
408: 
409: In order to show that clusters are in virial equilibrium, Figs. 4 and 5 depict the evolution 
410: of lagrangian radii of all stars and those of massive (M $\ge$ 30 $M_{\odot}$) and less massive 
411: stars (M $< 30~M_{\odot}$) of cluster model~11. As can be seen lagrangian radii of 
412: massive as well as less massive stars are nearly constant within the  
413: first few crossing times. This shows that the cluster is in a stable equilibrium
414: condition after mass segregation was introduced.
415: 
416: The central density and central relaxation time are measured for the
417: region inside the cluster core. Since massive stars
418: are not strongly concentrated toward the cluster center, the mean mass 
419: of stars within the cluster core is not very high (4.77 $M_{\odot}$ -- 11.82 $M_{\odot}$). 
420: Therefore the central relaxation time of clusters with $r_h$ = 1.3 pc (6.55 $\rm{Myrs}$ $\le T_{rel,c} \le$ 8.69 $\rm{Myrs}$) 
421: is not as low as that of the clusters in Table~2 
422: (3.54 $\rm{Myrs}$ $\le T_{rel,c} \le$ 3.84 $\rm{Myrs}$).
423: One may expect that the central relaxation time should be short enough that 
424: massive, low-energy stars spiral into the cluster core and create 
425: a high-mass core. Once in the cluster core, these massive stars could 
426: collide with each other and promote runaway merging.   
427: 
428: Nevertheless, our simulations do not show runaway merging (see models~9 -- 12
429: of Table~3). Reducing the half-mass radius $r_h$
430: from 1.3 pc to 0.5 pc in order to increase the central density 
431: (models~13 -- 15, see column 4 of Table~3) does not
432: help runaway merger to occur either.  
433: 
434: Model~15 actually has the same initial central density and half-mass radius as
435: model~3. Initial mass segregation is not introduced in model~3, but the
436: cluster experiences runaway merging through three collisions (see
437: Table~1). Fig.~6 depicts the evolution of lagrangian radii of inner shells of
438: these two models. Both clusters experience contractions of their
439: cores. While the contraction of model~3 is sufficiently strong to
440: let core-collapse occur at $t=2.6$ Myrs, the core of model~15 does not 
441: collapse until 3 Myrs and no runaway merging occurs.
442: 
443: By using the Monte Carlo method, \citet{gur04} studied core-collapse of star
444: clusters with initial mass segregation. A direct comparison of their results
445: with ours is again difficult due to differences in the adopted initial mass
446: spectrum, density profile, number of particles (up to $N = 10^7$ for
447: \citet{gur04}) and the method used in introducing initial mass
448: segregation. \citet{gur04} note in the caption of Fig.~13 that stellar
449: evolution can reverse core collapse. This agrees at least qualitatively with what
450: we see in our runs, since for example Fig.~6 shows that, despite of similar
451: size and density profile, model 3 goes into core collapse earlier than model
452: 15. This could be due to the fact that core collapse in model 15, whose core
453: contains many high-mass stars due to initial mass segregration, is delayed by
454: the stronger mass loss from the core due to stellar evolution.
455: 
456: Besides the effect of stellar evolution, the difference of the evolution of
457: models~3 and~15 may due to the difference of their collision rates. We examine the collision rate of
458: these models by calculating the the collision rate $N_{Coll}$ using equation (8-122) of \citet{bin87}
.
459: 
460: \begin{equation}
461: N_{Coll\star} = 4\sqrt{\pi} n \sigma {(2~R_{\star})}^2 + 4 \sqrt{\pi} G M_{\star} n (2~R_{\star})/\sigma.
462: \end{equation}
463: 
464: Here $N_{Coll\star}$ is the average number of collisions that a star suffers 
465: per unit time, $n$ indicates the number density of stars,
466: $\sigma$ is the velocity dispersion of stars, 
467: $R_{\star}$ and $M_{\star}$ denote radius and mass of colliding stars, 
468: and $G$ is the gravitational constant. The first term is derived from the 
469: kinetic theory for inelastic encounters and the second term represents the
470: enhancement in the collision rate by the gravitational attraction of the two
471: colliding stars.
472: 
473: Let us consider the region inside the core radius $\rm r_{core}$. The average number of collisions per unit time $N_{Coll}$ 
474: is obtained by multiplying $N_{Coll\star}$ with the number of stars inside 
475: the core radius $N_{\star core}$. Therefore 
476: \begin{equation}
477: N_{Coll} = N_{Coll\star}~N_{\star core}.
478: \end{equation}
479: The number density
480: of stars inside the core radius $n$ can be written as
481: \begin{equation}
482: n = N_{\star core} / V_{core}. 
483: \end{equation} 
484: where $V_{core} = \frac{4\pi}{3} \rm r_{core}^3$.
485: Thus
486: \begin{equation}
487: N_{Coll} = N_{Coll \star}~n~V_{core}.
488: \end{equation}
489: Substituting $N_{Coll\star}$ with the expression written in eq.~2, we see that
490: \begin{equation}
491: N_{Coll} \propto n^2.
492: \end{equation}
493: 
494: We use the theoretical prediction of the collision rate (eqs.~2 and~5) to 
495: follow the growth in the number of collisions per unit time in models~3 and 15. 
496: $N_{Coll\star}$ is calculated by considering the mass and radius of 
497: each star and then summing up over all stars within the region inside
498: the core to obtain $N_{Coll}$. Core parameters and collision rates
499: are calculated each time $N$-body data was stored and are then summed up
500: over all times.
501: 
502: The theoretical estimates are compared with the collision rate we
503: find in our simulations in Fig.~7. Both theoretical and simulation results 
504: (see column 7 of  Table~1 and column 9 of Table~3)
505: show that the collision rate of model~3, which
506: experiences runaway merging, is higher than the one in model~15.
507: The theoretical prediction of the collision rate overestimates the simulation 
508: results by a factor $\approx$ 2. This may be due to assumptions 
509: (i.e. mass and radius of colliding stars are the same) and idealizations 
510: (i.e. distribution function of velocity is Maxwellian) used in the 
511: derivation of eq.~2, while in the
512: simulations we use a mass spectrum and stellar radii according to
513: a certain mass-radius relation.
514: 
515: \section{Discussion and Conclusions}\label{sec:discon}
516: 
517: We have followed the evolution of multi-mass, dense star clusters with 
518: dimensionless central potentials of 3.0 $\le W_0 \le$ 9.0. Our simulations 
519: show a good agreement with the results of \citet{por04} that in MGG-11 type clusters
520: without initial mass segregation, dimensionless central potentials $W_0 \ge 9.0$
521: corresponding to  
522: central number densities larger than $10^6/\rm{pc}^3$ are required 
523: for runaway mergers to occur. Examining clusters with lower dimensionless central potential,
524: $W_0 \le 7.0$, confirm this limit for runaway mergering, as shown in Fig.~8. 
525: 
526: Initial mass segregation increases the average mass of stars within 
527: the cluster center and thus decreases the central relaxation time. It 
528: also allows to form a high-mass core. However, as long as the mass density
529: profile is kept constant, we find that 
530: initial mass 
531: segregation does not support runaway stellar merging to happen since
532: the collision rate is decreased.
533: 
534: In spite of the differences in adopted IMF, number of particles, treatment of
535: stellar evolution and stellar collisions, our results are in line with
536: \citet{fre06a,fre06b} (see sec.~3.1) and \citet{gur04} (see sec.~3.2).
537: 
538: The data of Milky Way globular cluster given by \citet{har96} provide
539: the central luminosity density (in $L_{\odot}/pc^3$) which can be converted
540: into a central mass density by assuming a mass-to-light ratio $M/L = 1$. 
541: Doing this, we find that about 67 \% of Milky Way
542: globular clusters have central densities $10^2~M_{\odot}/\rm{pc^3} < \rho_c <
543: 10^5~M_{\odot}/\rm{pc^3}$. Only 4 \% have central densities exceeding $\rho_c
544: = 4.3 \times 10^5~M_{\odot}/\rm{pc^3}$, while none has a central density
545: larger than $\rho_c = 10^6~M_{\odot}/\rm{pc^3}$, as depicted on Fig.~9. 
546: Studies of the evolution of clusters containing IMBHs by \citet{bau04a,bau04b} have 
547: shown that clusters with IMBHs expand due to energy generation in their cusp. 
548: \citet{bau04b} found that the 
549: cluster expansion can be strong enough that very concentrated clusters can end up 
550: among the least dense clusters. However, Milky Way globular clusters have 
551: half-mass radii very similar to the radii of clusters which form today, like 
552: galactic open clusters or super-star clusters in interacting galaxies 
553: (see i.e. \citet{sch06,tra07}), which speaks against strong expansion. 
554: If the current densities are representative 
555: of the densities with which the clusters formed, then runaway merging would not have 
556: happened in any of these clusters. In addition, the data of young star clusters in the LMC given by
557: \citet{mac03} (see Table 6 of their paper) also shows that nearly all LMC 
558: clusters, including very young ones, have central densities far below the
559: critical value needed for runaway merging. Other possibilities
560: of forming IMBHs like the merging of many stellar mass black holes 
561: \citep{mil02} also need extreme initial conditions like very massive 
562: clusters \citep{gul04,ras06}.
563: 
564: Hence it seems likely that most star clusters did not have 
565: sufficient high central densities to form IMBHs.
566: This indicates that the formation of IMBHs in star 
567: clusters must have been a rare event.   
568: 
569: \acknowledgments
570: 
571: We thank for Douglas Heggie for valuable discussions. This work was
572: supported in part by the Grants-in-Aid of the Ministry of Education, Culture,
573: Sports, Science and Technology, Japan,~(14079205; EA,SM). Numerical computations
574: were carried out on GRAPE system at Center for Computational Astrophysics of
575: National Astronomical Observatory of Japan. Data analysis were in part carried
576: out in YITP, Kyoto University.
577: 
578: 
579: \begin{thebibliography}{}
580: \bibitem[Aarseth (1999)]{aar99} 
581:   Aarseth, S.J. 1999, \pasp, 111, 1333
582: \bibitem[Baumgardt et al.(2003a)]{bau03a} 
583:   Baumgardt, H., Hut, P., Makino, J., McMillan, S. and Portegies Zwart S. 2003a, \apj, 582, L21  
584: \bibitem[Baumgardt et al.(2003b)]{bau03b}
585:   Baumgardt, H., Makino, J., Hut, P., McMillan, S. and Portegies Zwart S. 
586:   2003b, \apj, 589, L25
587: \bibitem[Baumgardt et al.(2004a)]{bau04a} 
588:   Baumgardt, H., Makino, J., and Ebisuzaki T. 2004a, \apj, 613, 1133
589: \bibitem[Baumgardt et al.(2004b)]{bau04b} 
590:   Baumgardt, H., Makino, J., and Ebisuzaki T. 2004b, \apj, 613, 1143
591: \bibitem[Baumgardt et al.(2005)]{bau05} 
592:   Baumgardt, H., Makino, J., and Hut, P. 2005, \apj, 620, 238
593: \bibitem[Belkus et al.(2007)]{bel07} 
594:   Belkus, H., Van Bever, J., and Vanbeveren, D. 2007, \apj, 659, 1576
595: \bibitem[Binney \& Tremaine (1987)]{bin87}
596:   Binney, J., \& Tremaine, S. 1987, in Galactic Dynamics, Princeton University
597:   Press
598: \bibitem[Bonnell \& Davies (1998)]{bon98} 
599:    Bonnell, I.A.,\& Davies, M.B. 1998, \mnras, 295, 691 
600: %\bibitem[Bonnell et al.(2001)]{bon01}
601: %  Bonnell, I.A., Bate, M.R., Clarke, C.J., and Pringle, J.E. 2001, \mnras, 323,
602: %785
603: \bibitem[Bonnell \& Bate (2002)]{bon02}
604:   Bonnell, I.A., \& Bate, M.R. 2002, \mnras, 336, 659 
605: \bibitem[Casertano \& Hut (1985)]{cas85}
606:   Casertano, S., \& Hut, P. 1985, \apj, 298, 80
607: \bibitem[de Grijs et al.(2004)]{deg04}
608:   de Grijs, R., Gilmore, G.F., Johnson, R.A. 2004, in STScI Symp., The Local
609: Group as an Astrophysical Laboratory, ed. M. Livio 
610: \bibitem[Ebisuzaki et.al.(2001)]{ebi01} 
611:   Ebisuzaki, T. et.al. 2001, \apj, 562, L19 
612: %\bibitem[Gebhardt et al.(2002)]{geb02}
613: %  Gebhardt, K., Rich, R.~M., and Ho, L.~C. 2002, \apj, 578, L41 OK
614: \bibitem[Fiorito \& Titarchuk (2004)]{fio04}
615:   Fiorito, R., \& Titarchuk, L. 2004, \apj, 614, L113
616: \bibitem[Freitag et al.(2006a)]{fre06a}
617:   Freitag, M., Rasio, F.A., and Baumgardt, H. 2006a, \mnras, 368, 121 
618: \bibitem[Freitag et al.(2006b)]{fre06b}
619:   Freitag, M., G\"urkan, M.A., Rasio, F.A. 2006b, \mnras, 368, 141 
 
620: %\bibitem[Gal-Yam et al.(2007)]{gal07}
621: %  Gal-Yam, A., Leonard, D.C., Fox, D.B., Cenko, S.B., Soderberg, A.M., Moon,
622: %  D.S., Sand, D.J. Li, W., Filippenko, V., Aldering, G and Copin, Y., 2007, \apj, 656, 372
623: \bibitem[Gebhardt et al.(2005)]{geb05}
624:   Gebhardt, K., Rich, R.M., and Ho, L.C. 2005, \apj, 634, 1093 
625: \bibitem[Gerssen et al.(2002)]{ger02}
626:   Gerssen, J., van der Marel, R.P., Gebhardt, K., Guhathakurta, P., Peterson
627: R.C., and Pryor, C., 2002, \aj, 124, 3270
628: %\bibitem[Gerssen et al.(2003)]{ger03}
629: %  Gerssen, J., van der Marel, R.P., Gebhardt, K., Guhathakurta, P., Peterson
630: %R.C., and Pryor, C., 2003, \aj, 125, 376
631: \bibitem[G\"ultekin et al.(2004)]{gul04}
632:   G\"ultekin, K., Miller, M.C., and Hamilton, D.P., 2004, \apj, 616, 221 
633: \bibitem[G\"urkan et al.(2004)]{gur04}
634:   G\"urkan, M.A., Freitag, M., and Rasio, F.A., 2004, \apj, 604, 632
635: \bibitem[Harris (1996)]{har96}
636:   Harris, W.E., 1996, \aj, 112, 1487
637: \bibitem[Hurley et al.(2000)]{hur00}
638: Hurley, J.R., Pols, O.R., Tout, C.A., 2000, \mnras, 315, 543
639: %\bibitem[Ishii et al.(1999)]{ish99}
640: %Ishii, M., Ueno, M., \& Kato, M. 1999, \pasj, 51, 417
641: \bibitem[Kaaret et.al.(2001)]{kaa01}
642:   Kaaret, P. et.al. 2001, \mnras, 321, L29 
643: \bibitem[King (1966)]{kin66} 
644:   King, I.R. 1966, \mnras, 71, 64 
645: \bibitem[Mackey \& Gilmore (2003)]{mac03} 
646:   Mackey, A.D. \& Gilmore, G.F., 2003, \mnras, 338, 85
647: \bibitem[Magorian et al.(1998)]{mag98}
648:   Magorrian, J. et.al. 1998, \aj, 115, 2285
649: \bibitem[Matsumoto et al.(2001)]{mat01}
650:   Matsumoto, H. et.al. 2001, \apj, 574, L25
651: \bibitem[McCrady et al.(2003)]{mcc03}
652:   McCrady, N., Gilbert, A.M., Graham, J.R., 2003, \aj, 596, 240
653: \bibitem[Miller \& Hamilton (2002)]{mil02} 
654:   Miller, M.C. \& Hamilton, D.P., 2002, \apj, 576, 894
655: \bibitem[Murray \& Lin (1996)]{mur96}
656:   Murray, S.D. \& Lin, D.N.C. 1996, \apj, 467, 728 
657: %\bibitem[Noyola \& Gebhardt (2004)]{noy04}
658: % Noyola, E. \& Gebhardt, K. 2004, \aj
659: %\bibitem[Peebles(1972)]{pee72} 
660: % Peebles, P.~J. 1972, \apj, 178, 371
661: %\bibitem[Portegies Zwart et al.(1999)]{por99} 
662: % Portegies Zwart, S.~F., Makino, J., McMillan, S.~L.~W., \& Hut, P. 1999, \aap, 348, 117
663: \bibitem[Portegies Zwart et al.(2004)]{por04} Portegies Zwart, S.F.,
664:   Baumgardt, H., Hut, P., Makino, J., and McMillan, S.L.W. 2004, \nat, 438,
665:   724 
666: \bibitem[Portegies Zwart \& McMillan (2002)]{por02} 
667: Portegies Zwart, S.F. \& McMillan S.L.W.,  2002, \apj, 576, 899 
668: %\bibitem[Raboud \& Mermilliod (1998)]{rab98} Raboud, D. \& Mermiliod,
669: %  J.C. 1998, \aap, 333, 897
670: \bibitem[Rasio et al.(2006)]{ras06}
671:   Rasio, F.A., Baumgardt, H., Corongiu, A., D'Antona, F.,Fabbiano, G.,
672:   Fregeau, J.M., Gebhardt, K., Heinke, C.O., Hut, P., Ivanova, N., Maccarone,
673:   T.J., Ransom, S.M., and Webb, N.A. 2006, astro-ph/0611615
674: \bibitem[Schilbach et al.(2006)]{sch06}
675:  Schilbach, E., Kharchenko, N.V., Piskunov, A.E., R\"{o}ser, S., and Scholz, R.D. 2006, \aap, 456, 523 
676: \bibitem[Spitzer (1987)]{spi87} Spitzer, L.  1987, Dynamical Evolution of
677:   Globular Clusters, Princenton University Press, Princenton 
678: \bibitem[Strohmayer \& Mushotzky (2003)]{str03}
679:   Strohmayer, T.E. \& Mushotzky, R.F. 2003, \apj, 586, 61
680: \bibitem[Suzuki et al.(2007)]{suz07}
681:   Suzuki, T.K., Nakasato, N., Baumgardt, H., Ibukiyama, A., Makino, J., and
682:   Ebisuzaki , T. 2007, \apj, 668, 435 
683: %\bibitem[Spitzer \& Hart (1971)]{spi71} Spitzer, L.Jr. \& Hart, M.H. 1971,
684: %  \apj, 164, 399 OK
685: %\bibitem[van den Marel (2002)]{van02} van den Marel et.al. 2002, \aj, 124,
686: %  3255 
687: %Hubble Space Telescope Evidence for an Intermediate-Mass Black Hole in the Globular Cluster M15. I. STIS Spectroscopy and WFPC2 Photometry
688: %\bibitem[Vesperini \& Chernoff(1994)]{ves94} Vesperini, E. \& Chernoff, D.~F.\
689: %  1994, \apj, 431, 231 
690: %\bibitem[Zezas \& Fabiano(2002)]{zez02a} Zezas, A. and Fabiano, G.2002 \apj,
691: %  577, 726
692: %\bibitem[Zhao \& Bailyn (2005)]{zha05} Zhao, B. \& Bailyn, C.~D. 2005, \aj,
693: %  129, 1934 
694: \bibitem[Trancho et al.(2007)]{tra07}
695:  Trancho, G., Bastian, N., Schweizer, F., and Miller, B.W. 2007, \apj, 658,
696:  993
697: \bibitem[van der Marel (2001)]{van01}
698:  van der Marel, R.P. 2001, in Black Holes in Binaries and Galactic Nuclei,
699:  ed. E.P.J. van den Heuvel \& P.A. Woudt (Garching:ESO), 246
700: \bibitem[Zezas et al.(2002)]{zez02}
701:   Zezas, A., Fabbiano, G., Rots, A.H., and Murray, S.S. 2002, \apj, 577, 710
702: 
703: \end{thebibliography}
704: %\onecolumn
705: \begin{figure}
706: %\epsscale{0.6}
707: %\vspace{5mm}
708: %\plotone{lagrangeW7m30.eps}
709: \includegraphics[scale=0.8]{f1.eps}
710: %\plotone{f1.eps}
711: %\includegraphics[scale=0.8]{lagrW9W7W7rh05_v2.ps}
712: \caption{Evolution of lagrangian radii of inner shells containing 1\%, 2\%,
713:   3\%, 5\%, 10\% and 20\% of the total cluster mass of models~1 --~3. Core
714:   radii $\rm{r_{core}}$ are marked by bold lines. Model~1 has short 
715: enough central relaxation time that core collapse and subsequent runaway 
716: merging of stars happen within a few Myrs. Model~2 has a higher $T_{rel,c}$
717: (see column 6 of Table~1) which prevents its core to collapse. Compared to model~1, 
718: model~3 has similar value of $\rho_c$ which allows mild contractions 
719: to bring the core to collapse before 3 Myrs. 
720:   \label{fig1}} 
721: \end{figure}
722: 
723: \begin{figure}
724: %\includegraphics[scale=0.8]{lagrange_W7m30-90.eps}
725: \includegraphics[scale=0.8]{f2.eps}
726: %\plotone{f2.eps}
727: %\vspace{-7cm}
728: \caption{Evolution of lagrangian radii of inner shells containing 1\% -- 10\% 
729: of the total cluster mass of model~6 (top) and model~8 (bottom). Filling the
730: region inside the 5 \% lagrangian radii $R_{005}$ with stars more massive than
731: 30 $M_{\odot}$ (model~6) does not support the core to collapse. Even
732: increasing the minimum mass of stars in this region to 90 $M_{\odot}$
733: (model~8) does not help core collapse to happen. The reason is that a large
734: fraction of massive stars, which start their life inside the $R_{005}$, move
735: out of this region on a crossing time-scale (see Fig.~3).
736:   \label{fig2}} 
737: \end{figure}
738: 
739: \begin{figure}
740: %\includegraphics[scale=0.8]{lagrange_all-005.eps}
741: \includegraphics[scale=0.8]{f3.eps}
742: %\plotone{f3.eps}
743: %\vspace{-7cm}
744: \caption{Evolution of lagrangian radii of massive stars ($m_{min} = 30
745:   M_{\odot}$) which start their life inside the 5 \% lagrangian radius of the
746:   cluster model~6 (a) up to the first 0.05 Myrs, (b) until 3 Myrs. Total mass
747:   fraction of these massive stars is indicated by $M_{005}$. Within a crossing
748:   time-scale, some massive stars leave the region within the initial 5\%
749:     lagrangian radius, which is 0.24 pc in this model,
750:   due to their high initial velocities. The escape of the
751: massive stars is balanced by low-mass stars moving in from larger radii
752: (which is not shown in these figures).
753:   \label{fig3}} 
754: \end{figure}
755: 
756: \begin{figure}
757: %\includegraphics[scale=0.8]{lagrange_k7m30_010.ps}
758: \plotone{f4.eps}
759: %\vspace{-7cm}
760: \caption{Evolution of lagrangian radii of inner shells containing 1\% -- 10\% 
761: of the total cluster mass of model~11. Replacing 10\% of the lowest total
762: energy stars with stars more massive than 30
763:   $M_{\odot}$ does not support the core to collapse. This happens because the
764:   high-mass core did not be form until 3 Myrs.
765:   \label{fig4}} 
766: \end{figure}
767: 
768: %\newpage
769: \begin{figure}
770: %\includegraphics[scale=0.8]{lagrange_massive-non010.ps}
771: \includegraphics[scale=0.8]{f5.eps}
772: %\plotone{f5.eps}
773: %\vspace{-5cm}
774: \caption{Evolution of lagrangian radii of massive stars whose masses are at
775:   least 30 $M_{\odot}$ (top) and low-mass stars whose masses are less
776:   than 30 $M_{\odot}$ (bottom) up to 10 Nbody unit. These figures depict stable
777:   evolution of shells containing 1\% -- 10\% 
778: of the total mass of these stars in the cluster model~11. 
779:   \label{fig5}} 
780: \end{figure}
781: 
782: \begin{figure}
783: %\includegraphics[scale=0.8]{model3-15.ps}
784: \includegraphics[scale=0.8]{f6.eps}
785: %\plotone{f6.eps}
786: %\vspace{-7cm}
787: \caption{Evolution of lagrangian radii containing 1\% -- 30\% of total
788:   mass of model~15 and model~3. Initial
789:   mass segregation is applied in model~15 by replacing 20 \% of stars with
790:   the lowest total energy by massive stars with $m_{min}$ = 30
791:   $M_{\odot}$. The inner shells experience contraction but no core collapse
792:   until 3 Myr. Therefore runaway merger does not occur in this cluster. The cluster model~3 has same initial density profile and
793:   same half-mass radius as model~15, but no initial mass
794:   segregation. However, mild contraction in the inner shells of model~3 is
795:   enough to let runaway mergings occur. This may happen since the number of
796:   collisions inside the inner shells of model~3 is higher than the one in
797:   model~15 (see Fig.~7). 
798:   \label{fig6}} 
799: \end{figure}
800: 
801: 
802: \begin{figure}
803: %\includegraphics[scale=0.8]{colrate.eps}
804: \plotone{f7.eps}
805: %\vspace{-7cm}
806: \caption{Collision rate inside inner shells of cluster models~3 and~15
807:   obtained from simulations, compared to the theoretical prediction of
808:   collision rate based on inelastic encounters. The collision rate of the
809:   model without mass segregation (model~3) is higher than the model with
810:   initial mass segregation (model~15) because there are more stars inside the
811:   cluster core. Therefore the possibility for a runaway merger to occur
812:   is also higher. 
813:   \label{fig7}} 
814: \end{figure}
815: 
816: \begin{figure}
817: %\includegraphics[scale=0.8]{f8.eps}
818: \plotone{f8.eps}
819: %\vspace{-7cm}
820: \caption{Plot of log central density vs. the log number density of stars 
821: for all calculated models. In order for runaway mergers to occur, a number
822: density of stars larger than $10^6/\rm{pc}^3$ in the core is necessary.
823:   \label{fig8}} 
824: \end{figure}
825: 
826: \begin{figure}
827: %\includegraphics[scale=0.8]{f9.eps}
828: \plotone{f9.eps}
829: %\vspace{-7cm}
830: \caption{Distribution of central densities of Milky Way globular clusters. The 
831: dashed lines mark the central densities of clusters in which runaway merging 
832: occured (models~1 and 3 of Table~1). Most galactic globular clusters have 
833: central densities far below this limit, meaning that runaway merging of stars 
834: was unlikely to have occured in them.
835:   \label{fig9}} 
836: \end{figure}
837: 
838: 
839: 
840: \onecolumn
841: \begin{table}
842: \begin{center}
843: \caption{Properties of simulated clusters without initial mass segregation \label{tab1}}
844: \begin{tabular}{cccccccccccc}
845: \tableline\tableline
846:  ${}_1$ & ${}_2$ & ${}_3$ & ${}_4$  & ${}_5$  & ${}_6$  & ${}_7$  & ${}_8$
847:  & ${}_9$ & ${}_{10}$   & ${}_{11}$ & ${}_{12}$ \\
848:  $Model$ & $W_0$ & $N_{Star}$ & $r_h$ & $log~\rho_c$ &
849: $T_{\rm rel,c}$ & $Col$ & $\langle T_{\rm col}\rangle$ & $Col_{\rm rm}$ & $
850:  T_{\rm rm}$ & $M_{RS}$ & $RM$ \\
851:        &    &         & $(pc)$  & $(M_{\odot}/pc^3)$ &
852: $(Myr)$         &     & $(Myr)$   &     &  $(Myr)$     &  $(M_{\odot})$ & $(Y/N)$ 
853:  \\ 
854: \tableline
855:  1 & 9.0 & 131072 & 1.3 & 6.51 & 1.16 & 104 & 0.03  & 96 & 0.54 & 2786 & Yes \\
856:  2 & 7.0 & 131072 & 1.3 & 5.67 & 5.98 &   5 & 0.60  & -  & -  &  -   & No  \\
857:  3 & 7.0 & 131072 & 0.5 & 6.47 & 2.71 &  37 & 0.08  & 3  & 2.55 &  258 & Yes \\ 
858:  4 & 5.0 & 131072 & 1.3 & 5.20 & 18.36 &  -   &  -  & -  & -  &  -   & No  \\
859:  5 & 3.0 & 131072 & 1.3 & 4.91 & 39.75 &  -   &  -  & -  & -  &  -   & No  \\
860: \end{tabular}
861: \tablecomments{1: The first column indicates the cluster model, followed by 
862: the dimensionless central potential $W_0$ in the 2nd column. The number of stars in the 
863: cluster and the half-mass radius are given in the 3rd and 4th columns, respectively. 
864: The 5th column shows 
865:  the logarithm of central density followed by the logarithm of the central
866:  relaxation time. The 7th column gives the total number of collisions that
867:  occur up to 3 Myrs, followed by the average time between collisions. The 9th
868:  and the 10th columns indicate the number of collisions leading to runaway
869:  mergers and the time when runaway merging starts. The mass of the runaway
870:  star produced at the  end of the runaway merging process is given in the 11th
871:  column. The last column shows whether runaway
872:  merging happens or not.}
873: 
874: \end{center}
875: \end{table}
876: 
877: 
878: \onecolumn
879: \begin{table}
880: \begin{center}
881: \caption{Properties of clusters with initial mass segregation introduced within a certain radius \label{tab2}}
882: \begin{tabular}{cccccccccccc}
883: \tableline\tableline
884:  ${}_1$ & ${}_2$ & ${}_3$ & ${}_4$  & ${}_5$  & ${}_6$  & ${}_7$  &
885:  ${}_8$ &
886:  ${}_9$ & ${}_{10}$ & ${}_{11}$ & ${}_{12}$ \\
887:  $Model$ & $W_0$ & $N_{Star}$ & $r_h$ & $M_{IMS}$  & $m_{min}$ & $log~\rho_c$ &
888: $T_{\rm rel,c}$ &  $Col$ & $\langle T_{\rm col} \rangle$ & $Col_{\rm rm}$ & $RM$ \\
889:        &       &         & $(pc)$ &$(r \le R_{005})$    & $(M_{\odot})$& $(M_{\odot}/pc^3)$ &
890: $(Myr)$ &    & $(Myr)$ & & $(Y/N)$ \\ 
891: \tableline
892: 
893:  6 & 7.0 & 124420 & 1.3 & 0.05 & 30.0 & 5.52 & 3.54 & 3 & 1.00 & - &  No  \\
894:  7 & 7.0 & 124305 & 1.3 & 0.05 & 50.0 & 5.52 & 3.78 & 7 & 0.43 & - &  No  \\
895:  8 & 7.0 & 124201 & 1.3 & 0.05 & 90.0 & 5.52 & 3.84 & 2 & 1.50 & - &  No  \\
896: \end{tabular}
897: \tablecomments{2: The first and second columns indicate the cluster model and 
898: the dimensionless central potential $W_0$. The 3rd column shows the number of stars in the 
899: cluster followed by the half-mass radius in the 4th column. The 5th column gives 
900: the fraction of total mass of cluster (which is contained within the 5 \% lagrangian 
901: radius) where the first scenario of IMS is applied. We choose some of these stars 
902: randomly and assign them with new masses which are larger than the minimum mass 
903: indicated in the 6th column. The logarithm of central density and the logarithm of the central
904:  relaxation time are given in the 7th and 8th columns. The 9th and 10th
905:  columns indicate the total number of collisions that occur up to 3 Myrs and
906:  the average time between collisions. The 11th columns gives the number of
907:  collisions leading to runaway mergers. Here we see that none of these
908:  collisions leads to a runaway merger process. The last column shows whether
909:  runaway merging happens or not.}
910: 
911: \end{center}
912: \end{table}
913: 
914: %\clearpage
915: 
916: \onecolumn
917: \begin{table}
918: \begin{center}
919: \caption{Properties of simulated clusters with initial mass segregation introduced below a certain energy \label{tab3}}
920: \begin{tabular}{cccccccccccc}
921: \tableline\tableline
922:  ${}_1$ & ${}_2$ & ${}_3$ & ${}_4$  & ${}_5$  & ${}_6$  & ${}_7$  &
923:  ${}_8$
924:  & ${}_9$ & ${}_{10}$ & ${}_{11}$ & ${}_{12}$ \\
925:  $Model$ & $W_0$ & $N_{Star}$ & $r_h$ & $M_{IMS}$ &$m_{min}$ & $log~\rho_c$ &
926: $T_{\rm rel,c}$ & $Col$ & $\langle T_{\rm col}\rangle$ & $Col_{\rm rm}$ & $RM$ \\
927:        &    &       &  $(pc)$  &$(\rm{lowest} E_{tot})$ & $(M_{\odot})$& $(M_{\odot}/pc^3)$ &
928: $(Myr)$     &   &  $(Myr)$ &  & $(Y/N)$ \\ 
929: \tableline
930:  9 & 7.0 & 124420 & 1.3 & 0.05 & 30.0 & 5.59 & 8.24  & 2 & 1.50 & - &  No \\
931:  10 & 7.0 & 124297 & 1.3 & 0.05 & 50.0 & 5.58 & 8.69  & 6 & 0.50 & - & No  \\
932:  11 & 7.0 & 118805 & 1.3 & 0.10 & 30.0 & 5.54 & 8.07  & 1 & 3.00 & - & No  \\
933:  12 & 7.0 & 106669 & 1.3 & 0.20 & 30.0 & 5.47 & 6.55  & 2 & 1.50 & - & No  \\
934:  13 & 7.0 & 106669 & 0.7 & 0.20 & 30.0 & 6.01 & 3.25  & 12 & 0.25 & - & No  \\
935:  14 & 7.0 & 106669 & 0.6 & 0.20 & 30.0 & 6.21 & 2.57  & 8 & 0.38 & - & No  \\
936:  15 & 7.0 & 106669 & 0.5 & 0.20 & 30.0 & 6.45 & 1.96  & 20 & 0.15 & - & No  \\
937: \end{tabular}
938: \tablecomments{Same as Table~2 except that the 5th column indicates the
939:   fraction of stars with lowest total energy which were replaced by 
940:   massive stars. The minimum masses $m_{min}$ of these stars are given in column 6.}
941: 
942: 
943: \end{center}
944: \end{table}
945: 
946: \end{document}
947: