1: \documentclass{emulateapj} %astro-ph
2: %\documentclass[12pt,preprint]{aastex} %ApJ
3: \usepackage{graphicx}
4: \usepackage{color,calc}
5: %\usepackage{natbib}
6: \usepackage{amsmath}
7: \usepackage{amssymb}
8: \usepackage{amstext}
9:
10: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
11: \newcommand{\note}[3]{{\color{red}\bf [#1: #2 -- #3]}}
12: \newcommand{\Fewbody}{{\em Fewbody\/}}
13:
14: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
15: \begin{document}
16:
17: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
18: \title{X-Ray Binaries and the Current Dynamical States of Galactic Globular Clusters}
19: \shorttitle{X-Ray Binaries and the Current Dynamical States of Galactic Globular Clusters}
20: \submitted{ApJL, accepted} %astro-ph
21: \author{John M. Fregeau\altaffilmark{1,2}}
22: \shortauthors{FREGEAU}
23: \affil{Department of Physics and Astronomy, Northwestern University, Evanston, IL 60208}
24: \altaffiltext{1}{fregeau@northwestern.edu}
25: \altaffiltext{2}{Chandra Fellow}
26:
27: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
28: \begin{abstract}
29: It has been known for over 30 years that Galactic globular clusters (GCs) are overabundant
30: by orders of magnitude in bright X-ray sources per unit mass relative to the disk
31: population. Recently a quantitative understanding of this phenomenon has developed,
32: with a clear correlation between the number of X-ray sources in a cluster, $N_X$, and the cluster's
33: encounter frequency, $\Gamma$, becoming apparent. We derive a refined version
34: of $\Gamma$ that incorporates the finite lifetime of X-ray sources
35: and the dynamical evolution of clusters. With it we find we are able
36: to explain the few clusters that lie off the $N_X$--$\Gamma$ correlation,
37: and resolve the discrepancy between observed GC core radii
38: and the values predicted by theory.
39: Our results suggest that most GCs are still in the
40: process of core contraction and have not yet reached the thermal equilibrium
41: phase driven by binary scattering interactions.
42: \end{abstract}
43:
44: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
45: \keywords{globular clusters: general --- methods: numerical --- stellar dynamics}
46:
47: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
48: \section{X-Ray Sources and Cluster Dynamics}\label{sec:dyn}
49: It was realized more than 30 years ago that Galactic globular clusters (GCs) are overabundant
50: by orders of magnitude in bright X-ray sources per unit mass relative to the
51: disk population \citep{1975ApJ...199L.143C,1975Natur.253..698K}.
52: It was quickly understood that strong dynamical scattering interactions of
53: binaries in the dense cluster cores should be responsible for this overabundance
54: \citep{1987IAUS..125..187V}. With the advances in X-ray astronomy
55: made possible by observatories such as {\em Chandra}, the relationship
56: between X-ray sources and core cluster dynamics has recently
57: been quantitatively studied. \citet{2003ApJ...591L.131P} performed {\em Chandra}
58: observations of many Galactic GCs down to a limiting luminosity
59: of $4 \times 10^{30}\,{\rm erg}/{\rm s}$ in the 0.5--6 keV range (which includes
60: low-mass X-ray binaries [LMXBs] in outburst and quiescence, cataclysmic variables [CVs], millisecond
61: pulsars [MSPs], and magnetically active main sequence binaries [ABs]), and looked
62: for correlations between the number of X-ray sources in each cluster and
63: properties of the cluster itself. They found the strongest correlation with
64: the ``encounter frequency'' $\Gamma$, a rough estimate of the current
65: dynamical encounter rate in the cluster. More recently, \citet{2003ApJ...598..501H}
66: and \citet{2006ApJ...646L.143P} have isolated the quiescent LMXBs (qLMXBs)
67: and CVs, respectively, from the X-ray source populations, and have shown that their numbers are
68: indeed consistent with dynamical formation.
69:
70: These results represent quantitative,
71: empirical evidence that dynamical encounters are responsible for the formation
72: of X-ray sources in clusters. However, they suffer from at least a few drawbacks.
73: First, the correlation between the number of X-ray sources, $N_X$, and the
74: encounter frequency appears to be sub-linear, with $N_X \propto \Gamma^{0.74 \pm 0.36}$,
75: although for LMXBs the exponent is $0.97 \pm 0.5$ \citep{2003ApJ...591L.131P}.
76: Second, there are three clusters for which $N_X$
77: is significantly larger than predicted by $\Gamma$. In the original
78: work of \citet{2003ApJ...591L.131P} it was already clear that
79: NGC 6397 has an $N_X$ that is $\sim 5$ times larger than predicted by the $N_X$--$\Gamma$ correlation.
80: Recent observations show that $N_X$ is factor of $\sim 2$
81: times that predicted by $\Gamma$ for NGC 7099 \citep{2007ApJ...657..286L},
82: and a factor of $\sim 20$ for Ter 1 \citep{2006MNRAS.369..407C}.
83: The common thread among these three clusters is that they are observationally
84: ``core-collapsed,'' while all others in the \citet{2003ApJ...591L.131P} sample are not
85: \citep{2006AdSpR..38.2923G}.
86: (A possible exception is NGC 6752, whose collapsed core status is debated
87: \citep{2003ApJ...595..179F,1995ApJ...439..191L}.) A cluster is observationally
88: termed core-collapsed if its surface brightness profile is consistent with a cusp
89: at the limit of resolution, making it more difficult to measure
90: the core radius. As described below, the collapsed core status of a cluster
91: is linked to its dynamical state, implying that cluster evolution
92: complicates the $N_X$--$\Gamma$ correlation.
93:
94: \begin{figure}
95: \begin{center}
96: \includegraphics[width=0.9\columnwidth]{f1.eps}
97: \caption{Number of observed cluster X-ray sources with
98: $L_X\gtrsim 4 \times 10^{30}\,{\rm erg}/{\rm s}$ for several Galactic
99: GCs versus the encounter rate $\Gamma$.
100: The power-law fit and data points are from \citet{2003ApJ...591L.131P} with the exception
101: of NGC 7099 \citep{2007ApJ...657..286L} and Ter 1 \citep{2006MNRAS.369..407C}.
102: The $N_X$ error bars for Ter 1, NGC 6397, and NGC 7099 represent source counting
103: noise and background source uncertainty, but for the remaining
104: clusters represent only background uncertainty.
105: \label{fig:gamma}}
106: \end{center}
107: \end{figure}
108:
109: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
110: \section{Understanding Cluster Core Radii}\label{sec:rcrh}
111:
112: The evolution of a GC, being a bound self-gravitating system, is very similar
113: to the evolution of a star, and comprises three main phases.
114: In the ``core-contraction phase,'' the first phase of evolution,
115: the cluster's core contracts on a
116: relaxation timescale, much like a pre-main sequence star. Once the core density becomes
117: large enough for binary stars to begin strongly interacting
118: dynamically, and thus generating energy via super-elastic encounters, the cluster settles into the
119: ``binary-burning phase,'' analogous to the main sequence in stars. In this phase the dynamical
120: properties of the cluster core remain roughly constant
121: \citep[e.g.,][]{1991ApJ...370..567G,2003gmbp.book.....H,2007ApJ...658.1047F}.
122: Once the binary
123: population is exhausted in the cluster core, it will collapse\footnote{Note that due to
124: clashing naming conventions between observation and theory, this core collapse does not
125: necessarily correspond to the observational definition, as discussed in the text. We
126: therefore suggest that some non-dynamical term be used in place of the observational
127: ``core-collapsed'' designation, or at least that the distinction between the observational
128: and theoretical uses of the term be noted when it is used.}
129: via the gravothermal instability,
130: leading to extremely high central densities. Deep in collapse, an energy producing event,
131: such as an interaction of a dynamically-formed binary, will reverse the collapse, causing the core to
132: rebound and enter the ``gravothermal oscillation phase,'' in which the core continues to collapse
133: and rebound \citep{2003gmbp.book.....H}. For a graphical representation of the three main phases of
134: cluster evolution, see Figure 1 of \citet{1991ApJ...370..567G}, Figure 5 of \citet{2003ApJ...593..772F},
135: or Figure 29.1 of \citet{2003gmbp.book.....H}.
136:
137: Since there is strong observational evidence that Galactic GCs were born with
138: significant binary fractions \citep{1992PASP..104..981H},
139: and since the binary-burning phase is the longest-lived phase of cluster evolution
140: \citep[perhaps tens of Hubble times;][]{1991ApJ...370..567G,2003gmbp.book.....H,2007ApJ...658.1047F,2007MNRAS.374..344T},
141: it is widely
142: believed that most clusters observed today should be in this phase. Early approximate
143: calculations suggested that
144: the ratio of core to half-mass radius, $r_c/r_h$, in the binary burning phase, is broadly
145: consistent with observations of the $\sim 80$\% of Galactic GCs that are
146: {\em not} observationally core-collapsed \citep{1991ApJ...370..567G,2003ApJ...593..772F}.
147: Recently, however, more accurate simulations have shown that the early calculations overestimate
148: $r_c/r_h$ by a factor of 10 or more \citep{2006MNRAS.368..677H,2007ApJ...658.1047F}.
149: These latest results are quite difficult to ignore, since they represent the concordance
150: of two completely independent cluster evolution codes---one direct $N$-body, with minimal
151: approximations and a natural inclusion of binary interactions; the other the approximate
152: H\'enon Monte Carlo method, with direct few-body
153: integration of binary interactions. The values of $r_c/r_h$
154: in the binary-burning phase agree quite well between the two codes, and furthermore agree quite well with
155: semi-analytical theory \citep{1994ApJ...431..231V}. The result is that now only
156: the core-collapsed clusters agree with the predicted values of $r_c/r_h$ in the
157: binary-burning phase, implying that if most clusters are in this phase
158: some other energy generation mechanism is responsible for the measured core sizes.
159: Several suggestions have been put forth for the energy
160: source, including, most notably, central intermediate-mass black holes (IMBHs) \citep{2006astro.ph.12040T}.
161:
162: One key feature of the evolution of GCs that has escaped careful
163: attention, though, is the timescale of the initial phase of core contraction.
164: Many numerical simulations have shown that it is of order
165: $\sim 10$ {\em initial} half-mass relaxation times
166: \citep{1991ApJ...370..567G,2003MNRAS.343..781G,2006MNRAS.368..677H,2007ApJ...658.1047F}.
167: Since the
168: current half-mass relaxation time for most clusters is $\sim 1$ Gyr and was likely much longer
169: in the past, the core contraction phase may easily last
170: longer than a Hubble time \citep[e.g.,][]{2007MNRAS.379...93H}.
171: In fact, recent $N$-body simulations have shown that
172: for a range of initial binary fractions, the evolution of $r_c/r_h$ over a Hubble time
173: is a simple core contraction with no evidence of a binary-burning phase being reached
174: \citep{2007MNRAS.379...93H}. Thus the solution to the discrepancy between theory and
175: observations in $r_c/r_h$ could be the most mundane one, namely that core-collapsed
176: clusters are in the binary-burning phase while the rest are still undergoing core contraction.
177:
178: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
179: \section{A Refined $\Gamma$}\label{sec:gamma}
180:
181: First introduced by \citet{1987IAUS..125..187V}, the encounter frequency
182: $\Gamma$ is an estimate of the {\em current} dynamical interaction rate
183: in the cluster, which is assumed to be proportional to the current
184: number of observable X-ray sources. We refine this predictive quantity by including, among
185: other factors, the recent history of the evolution of the core properties.
186: We start with the general form of the interaction
187: rate, then specialize to the \citet{1987IAUS..125..187V} $\Gamma$,
188: and our new version.
189:
190: The interaction rate between two species of objects can be written generally as
191: \begin{equation}\label{eq:main}
192: \Gamma \equiv \frac{dN_{\rm int}}{dt} = \int\!\!\!\int n_1 n_2 \sigma_{12} \left| {\bf v}_{12} \right|
193: f({\bf v}_{12}) d^3{\bf v}_{12}\,d^3{\bf r} \, ,
194: \end{equation}
195: where $n_i$ is the number density of species $i$, $\sigma_{12}$
196: is the interaction cross section between the two species, ${\bf v}_{12}$ is their relative
197: speed with $f$ its distribution function,
198: and the integral is carried out over relative velocity
199: space and volume. For the dynamical creation of LMXBs, species 1 represents stellar binaries,
200: while species 2 represents neutron stars. The dynamical formation of CVs and other
201: low-luminosity X-ray sources is rather more complicated, so species 1 and 2 generally represent single and
202: binary star systems \citep[see, e.g., Figure 7 of][]{2006MNRAS.372.1043I}.
203: Typically, the integral in eq.~(\ref{eq:main}) is
204: approximated as
205: \begin{equation}\label{eq:standardgamma}
206: \frac{dN_{\rm int}}{dt} \propto \rho_c^2 r_c^3 / v_\sigma \, ,
207: \end{equation}
208: where $\rho_c$ is the core mass density, $v_\sigma$ is the core velocity dispersion,
209: the integral has been approximated by the core value,
210: the gravitational-focusing dominated cross section has been used,
211: and $\rho_{\rm c,1}/\rho_{\rm c,2}$ is assumed to be constant for all clusters.
212: Additionally, it should be pointed out that only the proportionality
213: in eq.~(\ref{eq:main}) has been preserved since several factors (some of which are not
214: constant among clusters) have been dropped. More accurate approximations of eq.~(\ref{eq:main})
215: have been used, including numerical integrals over cluster models \citep{2003ApJ...591L.131P},
216: but all are estimates of the {\em current} interaction rate.
217:
218: Dynamically formed X-ray binaries (XRBs) are known to have finite detectable lifetimes. For LMXBs,
219: this lifetime varies from $\sim 10^5$--$10^7$ yr for red giant donors, to $\sim 1$ Gyr for main-sequence
220: companions, to a few Gyr for ultracompacts \citep{2007arXiv0706.4096I}. For CVs, this
221: lifetime is $\sim 1\,{\rm Gyr}$ (N. Ivanova, priv.\ comm.). An additional complication
222: is that dynamically formed XRBs do not necessarily turn on as X-ray sources immediately following a
223: strong interaction. In fact, the interaction that places a binary on the path to becoming
224: an observable X-ray source typically occurs several Gyr before it becomes detectable
225: \citep{2006MNRAS.372.1043I,2007arXiv0706.4096I}.
226:
227: Since the XRBs we see now formed a few Gyr
228: or more ago, and since the recent dynamical history of clusters may have been quite variable,
229: it is clear that the current number of observable sources should be proportional to the
230: interaction rate integrated over time. We thus write
231: \begin{equation}
232: \Gamma \equiv N_{\rm int} = \int\!\!\!\int\!\!\!\int n_1 n_2 \sigma_{12} \left| {\bf v}_{12} \right|
233: f({\bf v}_{12}) d^3{\bf v}_{12}\,d^3{\bf r}\,dt \, .
234: \end{equation}
235: We perform the time integration over an interval $t_x$, the typical detectable
236: lifetime of an XRB, to a time $t_\ell$, the typical timescale between
237: strong interaction and observability, in the past.
238: We leave $t_x$ and $t_\ell$ as parameters, but take $t_x=1\,{\rm Gyr}$ and
239: $t_\ell=3\,{\rm Gyr}$ as canonical values.
240: We further simplify the integral by writing
241: \begin{equation}
242: N_{\rm int} = \int_{t_0-t_\ell-t_x}^{t_0-t_\ell} f_b f_{\rm co} n_c^2 \sigma_{\rm strong} \sqrt{2} v_\sigma
243: \frac{4\pi}{3} r_c^3 \,dt \, ,
244: \end{equation}
245: with $f_b$ the core binary fraction, $f_{\rm co}$ the core compact object fraction,
246: $n_c$ the core number density, $r_c$ the core radius, $v_\sigma$
247: the 1-D core velocity dispersion, and $\sigma_{\rm strong}$ the cross section for a strong
248: interaction between a binary and a single star. The factor of $\sqrt{2}$
249: is from taking the difference of two Maxwellian velocity distributions.
250: The cross section is
251: $\sigma_{\rm strong} \approx \pi a 2 G M / (\sqrt{2} v_\sigma)^2$, where $a$ is the
252: typical semi-major axis of a binary, $M=3m$ is the total mass of the binary--single
253: system, and $m$ is the typical stellar mass. Taking $f_b$ and $f_{\rm co}$ to be
254: constant over the time integral and substituting the
255: definition of the core radius, $r_c^2=9v_\sigma^2/4\pi G m n_c$
256: \citep{2003gmbp.book.....H}, yields
257: \begin{equation}\label{eq:81}
258: N_{\rm int} = f_b f_{\rm co} \frac{81 \sqrt{2} a}{4 G m}
259: \int_{t_0-t_\ell-t_x}^{t_0-t_\ell} v_\sigma^3 r_c^{-1} \,dt \, .
260: \end{equation}
261: This expression is similar to eq.~(\ref{eq:standardgamma}), but with the core
262: properties integrated over time.
263: We have kept all numerical factors for the sake of completeness---only
264: the proportionality represented by this equation is needed for what follows.
265: By substituting in for the time evolution of the core quantities in the integrand,
266: one can use this expression to differentiate among the three different phases of cluster
267: evolution.
268: We exclude the gravothermal oscillation phase, since in this phase a cluster
269: should have no more than a few binaries in its core \citep{2003gmbp.book.....H}, and all clusters
270: considered here with measured core binary fractions show evidence for many more
271: binaries than this
272: \citep[note that a recent estimate puts the core binary fraction of NGC 6397 at $15\pm 1$\%;][]{davisposter}.
273:
274: In the binary-burning phase the core radius and velocity dispersion are constant, so
275: the integral is easy to evaluate. For the core contraction phase, we adopt the time
276: evolution of the core radius shown in the $N$-body models of \citet{2007MNRAS.379...93H},
277: approximated as
278: $r_c(t) = r_{c,0} (10-9t/t_0)$, with $r_{c,0}$ the current core
279: radius and $t_0$ the current cluster age. Since there is essentially
280: no core energy support in the core contraction phase, for the relationship between core
281: radius and central velocity dispersion we adopt the self-similar collapse model,
282: with $v_\sigma^2 \propto r_c^{-0.21}$ \citep{1987gady.book.....B}. The ratio of the number
283: of interactions for a cluster in the binary-burning phase to the same cluster in core
284: contraction is then
285: \begin{equation}
286: \frac{N_{\rm int,bb}}{N_{\rm int,cc}} = \frac{t_x}{t_0}
287: \frac{2.835}{\left(\frac{t_0}{t_0+9t_\ell}\right)^{0.315}-
288: \left(\frac{t_0}{t_0+9t_\ell+9t_x}\right)^{0.315}} \, ,
289: \end{equation}
290: which has a minimum of $2.0$ and a maximum of $17.8$ in the range
291: $t_x=10^{-4}$--$3\,{\rm Gyr}$,
292: $t_\ell=1$--$10\,{\rm Gyr}$, for $t_0=13\,{\rm Gyr}$.
293: For the canonical values of $t_x=1\,{\rm Gyr}$
294: and $t_\ell=3\,{\rm Gyr}$ with $t_0=13\,{\rm Gyr}$, the value is $5.0$. Since the
295: number of X-ray sources should scale roughly linearly with the number of interactions,
296: this suggests that if a cluster is in the binary-burning phase (and has been for a time
297: $t_\ell+t_x$ to the present), it should have $\sim 5$ times
298: as many X-ray sources than it would if it were in the core contraction phase.
299: (If the \citet{2003ApJ...591L.131P} exponent of $0.74$ is adopted
300: this factor is $3.3$.)
301: Interestingly, the three clusters in Figure~\ref{fig:gamma} that
302: are observed to be core collapsed are
303: the three that have a significantly larger $N_X$
304: than predicted by the standard $N_X$--$\Gamma$ correlation, by a factor of
305: $\sim 2$ to $\sim 20$. NGC 6397 has an overabundance factor of
306: $\sim 2$, NGC 7099 a factor of $\sim 5$, and Ter 1 a factor of $\sim 20$.
307: However, we note that the estimate of $\Gamma$ for Ter 1 is more uncertain
308: than the rest since it comes directly from eq.~(\ref{eq:standardgamma}) and not
309: an integral over the cluster profile, and since the velocity dispersion
310: is only estimated as it has not been measured observationally.
311: The similarity between the observed overabundances and those predicted by our
312: simple revision of $\Gamma$
313: suggests that the observationally core-collapsed clusters are indeed in the
314: binary-burning phase, while the rest are still in the process of core contraction.
315:
316: A natural objection to this conclusion is that cluster metallicity may explain
317: the X-ray source overabundance. Studies of bright LMXBs associated with star
318: clusters in external galaxies have shown a strong correlation between cluster
319: metallicity and LMXB incidence, with metal-rich clusters being much more likely
320: to harbor an LMXB \citep[by a factor of 3 or more;][]{2004ApJ...613..279J,2007ApJ...660.1246S}.
321: However, this result appears to hold only for bright LMXBs. When looking at
322: the low-luminosity LMXB and CV populations in our Galaxy (which comprise the majority
323: of sources in Figure~\ref{fig:gamma}), a correlation between
324: source incidence and cluster metallicity is not clearly apparent \citep{2006ApJ...651.1098H}.
325: In any case, the metallicities of the three overabundant clusters are not significantly
326: larger than those of the other clusters in Figure~\ref{fig:gamma}, with [Fe/H]
327: values of -1.30, -1.95, -2.12 for Ter 1, NGC 6397, and NGC 7099, respectively,
328: with the rest of the clusters ranging in value from -0.34 (NGC 6440) to
329: -1.75 (NGC 6093) \citep{1996AJ....112.1487H}.
330: %Most notably, NGC 6366
331: %and NGC 6121 are {\em more} metal-rich than the three overabundant clusters,
332: %with [Fe/H] values of -0.82 and -1.20, respectively \citep{1996AJ....112.1487H}.
333:
334: Another possibility is that the XRBs in the low-$\Gamma$ clusters
335: may be mainly primordial, in which case their number should scale with
336: cluster mass. However, were this the case, a look at cluster
337: absolute magnitudes shows that NGC 6121 would have
338: roughly as many sources as NGC 7099 and more than NGC 6397 and Ter 1, and
339: NGC 6366 would have more than Ter 1 \citep{1996AJ....112.1487H}.
340: Note that the 2--5 sources detected in the low-$\Gamma$ cluster NGC 288 (not included in this
341: study), with $\Gamma\approx 0.53$ in the \citet{2003ApJ...591L.131P} normalization,
342: are likely primordial \citep{2006ApJ...647.1065K}.
343:
344: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
345: \section{Discussion}\label{sec:disc}
346:
347: This {\em letter} presents the confluence of three suggestive observational and theoretical
348: results into a self-consistent picture. The first is that of the clusters
349: that have been observed sufficiently to determine their XRB population, the three
350: that are core-collapsed are the same three that have a significant X-ray source overabundance
351: (of a factor of $\sim 2$ to $\sim 20$).
352: The second is the semi-analytical result derived in this {\em letter} that a cluster in the binary
353: burning phase for the last few Gyr should have $\sim 5$ times more dynamically formed X-ray sources than if it
354: were in the core contraction phase for the same time. The third is the recently confirmed discrepancy
355: between observations and theory for the core radii of Galactic GCs,
356: which suggests that only the observationally core-collapsed clusters are in the binary-burning phase.
357: In light of these facts, the conclusion that seems strongly suggested
358: is that most Galactic GCs are currently still in the core contraction phase,
359: while only the $\sim 20\%$ of clusters that are core-collapsed are in the binary
360: burning phase. This goes counter to the widely held belief that most clusters
361: are currently in the binary burning phase, and complicates the many existing studies that have assumed
362: cluster core properties that are constant with time.
363:
364: The implications of this result are manifold. There are many studies of the dynamical production of
365: interesting source populations in clusters which assume core properties that are constant with time. These
366: include predictions of the formation of blue stragglers \citep[e.g.,][]{2004ApJ...605L..29M}, the evolution of
367: the core binary fraction \citep{2005MNRAS.358..572I}, and tidal-capture binaries
368: \citep[e.g.,][]{1994ApJ...423..274D}, among others.
369: Revising the results may be as simple as scaling predicted source numbers,
370: but may not be so simple for other quantities.
371:
372: Studies of GC evolution have shown that clusters starting from
373: very different initial conditions evolve toward a common range of values in many
374: observable structural parameters in the binary-burning phase, including the
375: concentration and ratio of core to half-mass radius
376: \citep{2003ApJ...593..772F,2006MNRAS.368..677H,2007ApJ...658.1047F}.
377: Since most clusters may
378: not be in the binary-burning phase after all, their observed properties are likely
379: to be more strongly correlated with their initial conditions. This makes modeling
380: of clusters a bit more complicated, but on the other hand allows one to more readily deduce
381: something about the initial properties of clusters.
382:
383: Perhaps anticlimactically, our results suggest that the alternative energy sources
384: recently proposed for supporting GC cores are not
385: required. These include the suggestion of IMBHs in {\em many}
386: Galactic GCs \citep{2006astro.ph.12040T},
387: enhanced stellar mass loss from stellar evolution of
388: physical collision products \citep{chatterjeeposter},
389: mass segregation of compact remnants in young clusters \citep{2004ApJ...608L..25M},
390: or evaporation of the stellar-mass black hole subsystem in young clusters
391: \citep{2007MNRAS.379L..40M}.
392:
393: Although the picture painted in this {\em letter} is a suggestive one, there
394: are still several caveats and limitations to our analysis. In the
395: derivation of our refined $\Gamma$ we assume that the core binary fraction
396: and abundance of compact objects are constant over the time of integration.
397: Neither is strictly true, although we expect they will not vary enough
398: to significantly change the overabundance value we derive.
399: We have used only one possible expression for the evolution of the core radius
400: in core contraction. While we expect the general behavior to be very similar to
401: what we have assumed here, more work needs to be done to determine if it is
402: universal. Additionally, our analysis ignores the effect of Galactic tidal stripping on
403: cluster mass, which would make a cluster appear overabundant in X-ray
404: sources, and may be relevant for NGC 6397 \citep{2003ApJ...591L.131P}.
405: On the observational side, there are some uncertainties in
406: evaluating $\Gamma$, which is dependent on quantities that are somewhat
407: difficult to measure for core-collapsed clusters.
408:
409: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
410: \acknowledgements
411:
412: The author thanks T. Fragkos, C. Heinke, N. Ivanova, V. Kalogera, D. Pooley, F. Rasio,
413: and J. Sepinsky for helpful discussions, D. Pooley in particular
414: for data and for pointing out the importance of primordial X-ray populations
415: in low-$\Gamma$ clusters, and the referee P. Edmonds for comments and
416: suggestions that improved this work.
417: JMF acknowledges support from Chandra theory grant TM6-7007X and Chandra
418: Postdoctoral Fellowship Award PF7-80047.
419:
420: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
421: \bibliographystyle{apj}
422: \bibliography{apj-jour,main}
423:
424: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
425: \clearpage
426:
427: \end{document}
428: