0712.1036/ms.tex
1: \documentclass[]{article}
2: \usepackage{emulateapj}
3: \usepackage{graphicx}
4: 
5: \advance \voffset by -.50cm\relax
6: 
7: 
8: 
9: \begin{document}
10: 
11: \title{The lowest-mass stellar black holes: catastrophic death of
12: neutron stars in gamma-ray bursts}
13: 
14: \author{K.\ Belczynski$^1$, R.\ O'Shaughnessy$^2$, V. Kalogera$^3$, 
15:         F. Rasio$^3$, R. E. Taam$^3$, T. Bulik$^4$}
16: \affil{$^1$ Los Alamos National Laboratory (Oppenheimer Fellow)}
17: \affil{$^2$ Penn State University}
18: \affil{$^3$ Northwestern University}
19: \affil{$^4$ Warsaw University}
20: 
21: 
22: 
23: \begin{abstract}
24: Mergers of double neutron stars are considered the most likely progenitors for 
25: short gamma-ray bursts. Indeed such a merger can produce a black hole with a 
26: transient accreting torus of nuclear matter (Lee \& Ramirez-Ruiz 2007, Oechslin 
27: \& Janka 2006), and the conversion of a fraction of the torus mass-energy to 
28: radiation can power a gamma-ray burst (Nakar 2006). Using available binary pulsar 
29: observations supported by our extensive evolutionary calculations of double 
30: neutron star formation, we demonstrate that the fraction of mergers that can form 
31: a black hole -- torus system depends very sensitively on the (largely unknown) 
32: maximum neutron star mass. We show that the available observations and models put 
33: a very stringent constraint on this maximum mass under the assumption that a 
34: black hole formation is required to produce a short gamma-ray burst in a double 
35: neutron star merger. Specifically, we find that the maximum neutron star mass 
36: must be within $2 - 2.5 M_\odot$. Moreover, a single unambiguous measurement of a 
37: neutron star mass above $2.5 M_\odot$ would exclude a black hole -- torus central 
38: engine model of short gamma-ray bursts in double neutron star mergers. Such an
39: observation would also indicate that if in fact short gamma-ray bursts are 
40: connected to neutron star mergers, the gamma-ray burst engine is best explained 
41: by the lesser known model invoking a highly magnetized massive neutron 
42: star (e.g., Usov 1992; Kluzniak \& Ruderman 1998; Dai et al. 2006; Metzger, 
43: Quataert \& Thompson 2007).
44: \end{abstract}
45: 
46: 
47: \section{Introduction}
48: 
49: Gamma-ray bursts (GRBs) have been separated into two classes: long-soft bursts, 
50: and short bursts (Nakar 2006,{Gehrels} {et~al.} 2007). 
51: The origin of long-soft bursts has been 
52: connected to the death of low-metallicity massive stars
53: (Piran 2005, {Gehrels} {et~al.} 2007).  
54: However, while observations support a binary merger origin for short bursts
55: (Nakar 2006,{Gehrels} {et~al.} 2007), the exact nature of the 
56: progenitor remains uncertain: 
57: they could be either double neutron stars (NS--NS) or  black hole -- neutron star
58: (BH--NS) binaries. The number of BH--NS binaries that both merge and produce GRBs 
59: is hard to estimate since (i) no such system has yet been observed; (ii) formation 
60: models are rather uncertain and predict very small BH--NS merger rates (likely too 
61: small to explain most of the short bursts); and (iii) theory suggests that the 
62: fraction of BH--NS mergers producing bursts depends sensitively on the black hole 
63: spin and spin--orbit orientation 
64: ({Belczynski} {et~al.} 2007b), but black hole birth
65: spins are not well constrained observationally or theoretically. On the other hand, 
66: NS--NS binaries are only observed in the Milky Way, but their properties and 
67: numbers are also in agreement with theoretical models, and their merger rate
68: is sufficient to explain the present-day short burst population (Nakar 2006,
69: Belczynski et al. 2007a).
70: 
71: We have performed an extensive theoretical study of high-mass binary stars 
72: (potential progenitors of NS--NS systems) using \texttt{StarTrack}, a population 
73: synthesis code incorporating the most up-to-date and detailed input physics for 
74: massive stars (Belczynski et al 2008). The code employs state-of-the-art predictions
75: for neutron star and black hole masses based on hydrodynamic core collapse 
76: simulations (Fryer \& Kalogera 2001)
77:  and detailed stellar structure and
78: evolution calculations for massive stars (Timmes, Woosley, \& Weaver 1996). Our models 
79: predict a Galactic NS--NS merger rate in the range $\sim 10-100 {\rm Myr}^{-1}$
80: (Belczynski et al. 2007a), in good agreement with the empirical estimate of $\sim 3-190 
81: {\rm Myr}^{-1}$ ({Kim}, {Kalogera}, \&
82:   {Lorimer}, 2006). The spread in our predicted rates originates 
83: from including the most significant model uncertainties associated with the 
84: treatment of dynamical mass transfer episodes (common envelope phases), which are 
85: involved in the formation of most double compact objects (Belczynski et al. 2007a).
86: 
87: 
88: \section{Results}
89: 
90: In Figure \ref{fig1} we compare short GRB rates with NS--NS merger rates in the 
91: present-day (redshift $0$) universe. Extrapolating the NS--NS merger rates to the 
92: local universe by assuming a star-forming density of $10^{-2}$ Milky 
93: Way-equivalents per ${\rm Mpc}^3$ \footnote{Introduction of different modes of star 
94: formation and redistribution of progenitors over old elliptical and younger spiral 
95: galaxies does not increase the predicted merger rates by more than a factor of 
96: $\sim 3$ O'Shaughnessy et al. 2007.}, we estimate the local universe 
97: NS--NS merger rate to be in the range $\sim 100-1000 {\rm Gpc}^{-3} 
98: {\rm yr}^{-1}$. By comparison, the estimated conservative lower limit on the 
99: short GRB rate is $\sim 10 {\rm Gpc}^{-3}{\rm yr}^{-1}$, based on the BATSE/SWIFT 
100: sample (Nakar 2006). This estimate relies on very conservative assumptions: (i) 
101: there is no collimation and (ii) there are no bursts dimmer than we have already 
102: observed, thus providing a true lower limit on the rate. Therefore, even adopting 
103: the most optimistic predictions for the NS--NS merger rate and the most pessimistic 
104: bound on the local short GRB event rate, the fraction of NS--NS mergers $f_{grb}$ 
105: that produce GRBs must be greater than at least $10^{-2}$ to explain the majority 
106: of known short bursts.
107: 
108: In this paper we start with the assumption  that all short GRBs are connected with
109: NS-NS system mergers that produce a black hole. We discuss the implications of  
110: relaxing this stringent assumption at the end of the paper.
111: 
112: From our models we also derive physical properties of double neutron stars, with 
113: individual masses of neutron stars being of particular interest. Figure 
114: \ref{fig2} shows the relation between progenitor (single star) mass and final 
115: remnant mass used in our evolutionary calculations. Mass transfer and other 
116: binary interactions change this simple picture, through both accretion and mass
117: loss, which can either increase or decrease an individual binary component mass.
118: However, (Belczynski et al. 2007b) argues that we do not expect significant mass 
119: accretion onto the components of NS--NS binaries. The population model we adopt 
120: for our discussion here produces NS mass distributions that appear consistent  
121: with the current observed NS-NS sample, at least in the extent of the mass 
122: ranges (Fig.\ 3). While mass transfer does influence the remnant masses (e.g., 
123: smearing the narrowly peaked mass distribution implied by Fig.\ 2), the 
124: qualitative structure is largely preserved, as one would expect from isolated 
125: stellar evolution combined with an initial mass function that falls steeply with 
126: increasing initial mass. The predicted neutron star mass distribution only 
127: very weakly depends on evolutionary model assumptions because the neutron
128: star formation mass is almost independent of the progenitor mass (Timmes et al.
129: 1996) and mass accretion in NS-NS progenitor binaries is rather small 
130: (Belczynski et al. 2007a). 
131: 
132: Depending on the masses in the progenitor binary and the highly uncertain 
133: nuclear equation of state, the final remnant of a NS--NS merger may or may not 
134: collapse to a black hole. We estimate the final gravitational mass of the compact remnant as
135: \begin{equation}
136: \label{eq:FinalRemnantMass}
137: M_{\rm rem} = 0.9 (M_{\rm ns,1}+M_{\rm ns,2} - 0.1 M_\odot),
138: \end{equation}
139: where the initial neutron star masses are denoted by $M_{\rm ns,1},\ 
140: M_{\rm ns,2}$ and we have assumed that the torus mass is sufficiently large to 
141: power a GRB (i.e., $\simeq 0.1 M_\odot$)
142: ({Setiawan}, {Ruffert}, \&
143:   {Janka} 2004, Lee \& Ramirez-Ruiz 2007)
144: and that 10\% of rest mass is lost in neutrinos (Lattimer \& Yahli 1989; Timmes et
145: al. 1996). Higher rest mass loss (e.g., Metzger, Thompson \& Quataert 2007) would only
146: strengthen our subsequent conclusions. Because stars more massive than $18 
147: M_\odot$ (progenitors of massive neutron stars with $M_{\rm ns} \simeq 1.8 M_\odot$) 
148: are much rarer than those forming lighter neutron stars ($M_{\rm ns} \simeq 1.35 
149: M_\odot$) we \emph{a priori} expect that most remnants from NS--NS mergers will 
150: have rather low mass $M_{\rm rem} \simeq 2.3 M_\odot$ (see eq.~1 for two $1.35 M_\odot$
151: neutron stars).
152: 
153: Neither observations nor nuclear theory have yet pinned down the maximum 
154: neutron star mass $M_{\rm ns,max}$ above which a black hole must form. Thus, the 
155: fraction of binary neutron stars which produce black holes and are able to 
156: power short GRBs is set by the fraction of mergers such that
157: \begin{equation}
158: M_{\rm rem} \geq M_{\rm ns,max}.
159: \end{equation}
160: We therefore calculate the fraction of our simulated NS--NS mergers that 
161: lead to black hole formation and a short GRB as a function of $M_{\rm ns,max}$; 
162: see Fig.~4. Observations of the highest mass neutron stars ($\lesssim 2 
163: M_\odot$;
164: ( {Barziv} {et~al.} 2001, {Ransom} {et~al.} 2005)
165:  and lowest mass 
166: black holes ($\gtrsim 3 M_\odot$; 
167: (Orosz 2003, Casares 2006) 
168: only weakly 
169: constrain this parameter. Remnant masses from NS--NS mergers ($M_{\rm rem}$) 
170: obtained both from our simulations and from observations all fall very 
171: close to the range $2.2-2.5 M_\odot$.
172: 
173: Comparing Figs 1 and 4 we immediately deduce that, since the fraction $f_{\rm grb}$
174: of NS--NS mergers that produce short GRBs must be greater than $10^{-2}$ 
175: (Fig.~1), the neutron star maximum mass $M_{\rm ns,max}$ must be less than $2.5 
176: M_\odot$ (Fig.~4). Because we lack a robust lower bound on the mass of the 
177: residual torus surrounding the black hole, we have adopted a conservative 
178: upper limit on $M_{\rm ns,max}$ obtained by assuming a negligible torus mass (i.e., 
179: replace $0.1 M_\odot$ with $0$ in eq.~1).
180: 
181: This result has been obtained with the assumption that all ($k=1.0$) short GRBs are 
182: connected with NS-NS mergers. It is however possible that only a fraction of short 
183: GRBs is produced in NS-NS mergers. How does our result depend on the fraction $k$ of 
184: short GRBs that are connected with NS-NS mergers?
185: The lower limit on $f_{\rm grb}$ is then decreased by $k$, see Figure~\ref{fig1}.
186: If $k \gtrsim 0.1$, then the limit lower limit becomes $f_{\rm grb}>10^{-3}$, and 
187: as is clearly seen from Figure~\ref{fig3}, the upper limit on the maximum 
188: mass of a neutron star remains unchanged. 
189: For the values of $k \lesssim 0.01$ the NS-NS mergers are not important for
190: overall short GRB population, as the mergers would consist of only $\lesssim
191: 1\%$ of the short GRBs. Therefore, in this case short GRBs do not provide
192: information about the merger product. 
193: 
194: 
195: \section{Discussion}
196: 
197: Our proposed limit on the maximum neutron star mass is still above the maximum 
198: masses allowed by almost all proposed models for the nuclear equation of state 
199: ({Lattimer} \& {Prakash} 2007). However, the proposed limit would remain unchanged 
200: even if a dramatic improvement in short GRB surveys led to a significantly larger 
201: lower bound on the local short GRB rate:  because of the sharp decrease in 
202: $f_{\rm grb}$ with $M_{\rm ns,max}$ shown in Fig.~4.
203: If, however, electromagnetic observations could constrain the 
204: least luminous short GRBs and thus provide an \emph{upper} bound on the short 
205: GRB rate, with gravitational wave observations at the same time accurately 
206: determining the NS--NS merger rate, then $f_{\rm grb}$ could also be constrained from 
207: \emph{above}. If only a fraction of NS--NS mergers produce short bursts, 
208: because $f_{\rm grb}$ depends so sensitively on $M_{\rm ns,max}$, the combination of upper 
209: and lower limits would constrain the maximum neutron star mass extremely 
210: tightly, even if the assumptions going into eq.~1 are relaxed. 
211: 
212: While our limit on the effective maximum neutron star mass is entirely 
213: empirical, detailed merger models including realistic relativistic dynamics, 
214: neutrino transport, magnetic fields, and potentially even energy extraction
215: from the final black hole remain under intense investigation (Janka \& Ruffert 
216: 1996; Oechslin \& Janka 2006). Many merger remnants are 
217: expected to be (temporarily) rotationally supported against collapse 
218: (Morrison, Baumgarte \& Shapiro 2004), with a ``hypermassive'' remnant neutron star 
219: eventually spinning down and collapsing to a black hole 
220: (Faber et al. 2006; Duez et al. 2007; Shibata \& Taniguchi 2006).
221: Our model only relies upon the current consensus on double neutron star 
222: mergers, as summarized by Oechslin \& Janka (2006): 
223: sufficiently massive binary mergers produce a black hole and only mergers 
224: that produce a black hole extract enough energy to power short GRBs.
225: 
226: Known Galactic black holes extend in mass up to $10-15 M_\odot$ (Casares 2006),
227: while two recently discovered black hole candidates in other galaxies 
228: (Orosz et al. 2007, Prestwich et al. 2007)  have even higher masses of 
229: $\simeq 16$ and $\gtrsim 24 M_\odot$. Clearly black holes can form with rather 
230: high masses in different types of environments. The lower mass limit is not 
231: well constrained observationally, as the highest-mass neutron stars barely 
232: reach $2 M_\odot$, while the lowest-mass black holes are above $3 M_\odot$. 
233: In order to explain the observed short GRBs with NS--NS mergers, under the 
234: assumption of black hole -- torus central engine model, we have shown that 
235: the maximum neutron star mass must be lower than $2.5 M_\odot$. 
236: However, pulsar surveys (Ransom et al. 2005) have discovered increasingly
237: more massive neutron stars. So far in our analysis we 
238: have included only NS--NS systems formed in the field.  There is one known 
239: relativistic double neutron star system in the Galactic globular cluster M15, 
240: that has probably formed through dynamical interactions.  This binary consists 
241: of two low-mass neutron stars ($1.36$ and $1.35 M_\odot$; Jacoby et al. 2006)
242: very similar to those in the Galactic field, so the results of our analysis 
243: are not changed by this isolated observation. Moreover, it was estimated that
244: no more than $10-30\%$ short GRBs can originate from mergers of double neutron 
245: stars formed in globular clusters ({Grindlay}, {Portegies Zwart}, \& {McMillan} 
246: 2006).
247: 
248: If any observation can be made that establishes unambiguously a pulsar mass 
249: (either in the field or a globular cluster) over $2.5 M_\odot$, this would 
250: exclude a black hole -- torus short GRB central engine model for double neutron 
251: star mergers. We note that a tentative mass measurement for a pulsar of 
252: $2.74 \pm 0.21 M_\odot$ was recently reported by (Freire et al. 2007). If this 
253: measurement is confirmed, the double neutron star mergers may still be possible 
254: progenitors for short GRBs. However, the central engine model will need to be 
255: reexamined.  In particular, it was proposed that a merger of two neutron stars 
256: may lead to the formation of a magnetar; a rapidly rotating highly 
257: magnetized and high mass neutron star (with or without a torus) that can lead 
258: to a short GRB (e.g., Usov 1992; Kluzniak \& Ruderman 1998; Dai et al. 2006; 
259: Metzger et al. 2007).  Gravitational wave observatories (LIGO, VIRGO) may 
260: provide the direct evidence that NS-NS merger can produce a short GRB if there 
261: is a coincidence of the burst and the inspiral gravitational wave signal. 
262: It may also be possible to distinguish a merger product (NS versus BH) 
263: from the shape of the merger and ringdown signal or from radio pulses if a 
264: magnetar was formed in a nearby gamma-ray burst event.  
265: 
266: \begin{acknowledgements}
267: We would like to thank Neil Gehrels, Duncan Lorimer, Scott Ransom, Ben Owen,
268: Jerome Orosz, Chris Stanek, John Beacom, Paulo Freire, Chunglee Kim, Todd
269: Thompson and an anonymous referee for very useful discussions. 
270: \end{acknowledgements}
271: 
272: 
273: \begin{references}
274: 
275: \reference{} Barziv, O., Kaper, L., Van Kerkwijk, M.~H., Telting, J.~H., \& Van
276:              Paradijs, J. 2001, \aap, 377, 925
277: \reference{} Belczynski, K., Kalogera, V., Rasio, F., Taam, R., Zezas, A.,
278:              Maccarone, T., \& Ivanova, N. 2008, ApJS, 174, 223
279: \reference{} Belczynski, K., Taam, R.~E., Kalogera, V., Rasio, F.~A., \& Bulik, T.
280:              2007, ApJ, 662, 504
281: \reference{} Belczynski, K., Taam, R.~E., Rantsiou, E., \& Sluys, M.~v.~d.
282:              2007, ApJ, submitted (astro-ph/0703131)
283: \reference{} Casares, J. 2006, IAU Symposium 238: "Black Holes: From
284:              Stars to Galaxies -Across the Range of Masses", in press 
285:              (astro-ph/0612312)
286: \reference{} Dai, Z., Wang, X., Wu, X., \&  Zhang, B. 2006, Science, 
287:              311, 1127
288: \reference{} Duez, M.~D., Liu, Y.~T., Shapiro, S.~L., Shibata, M., \& Stephens,
289:              B.~C. 2007, Proceedings of the Eleventh Marcel Grossmann
290:              Meeting, in press (gr-qc/0701145)
291: \reference{} Faber, J.~A., Baumgarte, T.~W., Shapiro, S.~L., \& Taniguchi, K. 2006,
292:               \apj, 641, L93
293: \reference{}  Freire, P.~C.~C., Ransom, S.~M., Begin, S., Stairs, I.~H., Hessels,
294:               J.~W.~T., Frey, L.~H., \& Camilo, F. 2007, To appear in the
295:               proceedings of "40 Years of Pulsars: Millisecond Pulsars, Magnetars, and
296:               More", in press (arXiv:0711.2028)
297: \reference{} Fryer, C.~L. \& Kalogera, V. 2001, \apj, 554, 548
298: \reference{} Gehrels, N., Cannizzo, J.~K., \& Norris, J.~P. 2007, New Journal of
299:              Physics, 9, 37
300: \reference{} Grindlay, J., Portegies Zwart, S., \& McMillan, S. 2006, Nature Physics,
301:              2, 116
302: \reference{} Janka, H.-T. \& Ruffert, M. 1996, \aap, 307, L33
303: \reference{} Jacoby, B., et al.\ 2006, ApJ, 644, L113
304: \reference{} Kim, C., Kalogera, V., \& Lorimer, D.~R. 2006, 
305:              Proceedings of "A life with stars", in press (astro-ph/0608280)
306: \reference{} Klu\'zniak, W., \& Ruderman, M.\ 1998, \apj, 505, L113 
307: \reference{} Lattimer, J.~M. \& Prakash, M. 2007, \physrep, 442, 109
308: \reference{} Lattimer, J.~M. \& Yahli, M. 1989, \apj, 340, 426
309: \reference{} Lee, W.~H. \& Ramirez-Ruiz, E. 2007, New Journal of Physics, 9, 17
310: \reference{} Lorimer, D.~R. et al., M. 2006, \apj, 640, 428
311: \reference{} Metzger, B., Quataert, E., \& Thompson, T. 2007, MNRAS,
312:              submitted (arXiv:0712.1233)
313: \reference{} Metzger, B.,  Thompson, T. \& Quataert, E.\ 2007, ApJ,
314:              659, 561 
315: \reference{} Morrison, I.~A., Baumgarte, T.~W., \& Shapiro, S.~L. 2004, \apj, 610, 941
316: \reference{} Nakar, E. 2007,  Physics Reports, 442, 166
317: \reference{} Oechslin, R. \& Janka, H.-T. 2006, \mnras, 368, 1489
318: \reference{} Orosz, J.~A. 2003, in IAU Symposium, Vol. 212, A Massive Star Odyssey: From
319:              Main Sequence to Supernova, ed. K.~van der Hucht, A.~Herrero, \&
320:              C.~Esteban, 365
321: \reference{} Orosz, J.~A. et al.\ 2007, \nat, 449, 872
322: \reference{} O'Shaughnessy, R., Kim, C., Kalogera, V., \& Belczynski, K. 2008, ApJ,
323:              672, 479
324: \reference{} Piran, T. 2005, Reviews of Modern Physics, 76, 1143
325: \reference{} Podsiadlowski, P., Langer, N., Poelarends, A.~J.~T., Rappaport, S.,
326:              Heger, A., \& Pfahl, E. 2004, \apj, 612, 1044
327: \reference{} Prestwich, A., et al.\ 2007, \apj, 669, L21
328: \reference{} Ransom, S., et al.\ 2005, Science, 307, 892
329: \reference{} Setiawan, S., Ruffert, M., \& Janka, H.-T. 2004, \mnras, 352, 753
330: \reference{} Shibata, M. \& Taniguchi, K. 2006, \prd, 73, 064027
331: \reference{} Timmes, F.~X., Woosley, S.~E., \& Weaver, T.~A. 1996, \apj, 457, 834
332: \reference{} Usov, V. 1992, Nature, 357, 472
333: 
334: \end{references}
335: 
336: 
337: \begin{figure}
338: \includegraphics[width=\columnwidth]{f1.ps}
339: \caption{\label{fig1} 
340: Comparison of the double neutron star merger rates and short GRB event rates.  
341: The solid black line and arrows indicate a firm lower bound on the short GRB  
342: event rate (Nakar 2006), based solely on the rate of detected bursts. Depending 
343: on the amount of beaming and the fraction of distant faint short GRBs that are 
344: missed, the true event rate is often estimated to be at least $10$ times larger 
345: (Nakar 2006). This lower limit is smaller than the double neutron star merger 
346: rate estimated for the Milky Way both from (i) observations of Galactic binary 
347: pulsars (filled blue region) and (ii) our population synthesis simulations 
348: (filled red region), when these two estimates are extrapolated to cosmological 
349: scales. Based on the maximum plausible double neutron star merger rate with the 
350: minimum plausible short GRB event rate, the fraction $f_{\rm grb}$ of binary mergers 
351: that lead to short GRBs should be greater than $10^{-2}$ if double neutron stars 
352: are the progenitors of short GRBs.
353: }
354: \end{figure}
355: 
356: 
357: \begin{figure}
358: \includegraphics[width=\columnwidth]{f2.ps}
359: \caption{\label{fig2}
360: Initial (Zero Age Main Sequence) mass to final compact object mass relation for 
361: single stars. This represents our current understanding of compact object 
362: formation. Stars below about $7.5 M_\odot$ form white dwarfs; stars in the narrow 
363: range around $8 M_\odot$ can potentially form very light neutron stars through 
364: electron capture supernovae (Podsiadlowski et al. 2004). More massive stars show a well defined 
365: bifurcation caused by different modes of energy transport in the stellar core: 
366: stars below $18 M_\odot$ form light neutron stars ($\simeq 1.35 M_\odot$), while 
367: stars above this mass form heavy neutron stars ($\simeq 1.8 M_\odot$). Above 
368: $\simeq 20 M_\odot$ stars experience partial fallback of material that can turn 
369: nascent neutron stars into black holes. Compact objects originating from stars 
370: of $\sim 20-22 M_\odot$ form either very heavy neutron stars or low-mass black 
371: holes depending on the unknown limiting mass between these two remnant types 
372: (expected to lie around $2-3 M_\odot$).
373: }
374: \end{figure}
375: 
376: 
377: \begin{figure}
378: \includegraphics[width=\columnwidth]{f3.ps}
379: \caption{\label{fig3}
380: Predicted mass distribution for neutron stars in merging double neutron star 
381: binaries. First born neutron stars are slightly heavier as they can accrete 
382: some matter from their unevolved binary companions. Population synthesis 
383: models (red and blue lines) are shown along with measured neutron star masses 
384: for the known double neutron star binaries. Although more observations are 
385: needed to constrain the shape of this distribution, the mass ranges of 
386: observed and predicted systems are in agreement. We use direct mass estimates 
387: for B1913$+$16, B1534$+$12, J0737$-$3039 and J1756$-$2251 (O'Shaughnessy et
388: al. 2008), 
389: while for J1906$+$0746 we assume that both neutron stars have masses of $1.3 
390: M_\odot$ (total system mass is $2.6 M_\odot$; Lorimer et al. 2006). The 
391: few compact objects found in our simulations with masses as high as $\simeq 
392: 2.5 M_\odot$ may well be low-mass black holes (see also Fig.~\ref{fig1}).
393: }
394: \end{figure}
395: 
396: 
397: 
398: 
399: \begin{figure}
400: \includegraphics[width=1.05\columnwidth]{f4.ps}
401: \caption{\label{fig4}
402: Gamma Ray Burst production efficiency as a function of the maximum neutron 
403: star mass in the framework of the double neutron star model for GRBs 
404: involving the formation of a black hole. 
405: The mass of the merger product is plotted 
406: here as the blue (observations) and red (theory) lines. There is a sharp drop 
407: in number of NS--NS systems that can form a merger with mass over $2.5 M_\odot$. 
408: For example, only 1 in $10^3$ NS--NS mergers can form a remnant with a mass 
409: of $2.5 M_\odot$ or higher. Therefore, if short GRBs are connected to NS--NS 
410: mergers, the maximum neutron star mass is required to be $M_{\rm ns,max} < 2.5 M_\odot$.
411: For comparison we show observed masses of the lowest-mass black holes (GRO 
412: J0422$-$32, GRS 1009$-$45; Casares 2006) and highest-mass neutron stars (Vela X-1, 
413: Terzan~5I, and Terzan~5J; Barziv et al. 2001, Ransom et al. 2005).
414: These observations, along with our findings, constrain the maximum mass of a neutron 
415: star to lie in the narrow range of $2-2.5 M_\odot$.
416: }
417: \end{figure}
418: 
419: 
420: 
421: \end{document}
422: 
423: 
424: 
425: 
426: