1: \documentclass[12pt,preprint]{aastex}
2: %%\documentclass{article}
3: %\usepackage{graphicx}
4: %\usepackage{subfigure}
5:
6: \newcommand{\ps}{\mbox{$P_{\rm s}$}}
7: \newcommand{\pb}{\mbox{$P_{\rm orb}$}}
8: \newcommand{\rn}{\mbox{$R_{\rm m}$}}
9: \newcommand{\rc}{\mbox{$R_{\rm c}$}}
10: \newcommand{\my}{\mbox{$M_{\odot}{\rm yr}^{-1}$}}
11: \newcommand\lsim{\lower0.5ex\hbox{$\; \buildrel < \over \sim \;$}}
12: \newcommand\gsim{\lower0.5ex\hbox{$\; \buildrel > \over \sim \;$}}
13:
14: \title
15: {EXPLORATION OF THE $\ps-\pb$ RELATION FOR WIND-FED X-RAY PULSARS
16: %SPIN EVOLUTION OF NEUTRON STARS IN MASSIVE BINARIES
17: }
18:
19: \author
20: {Hai-Lang Dai, Xi-Wei Liu and Xiang-Dong Li}
21:
22: \affil{Department of Astronomy, Nanjing University,
23: Nanjing 210093, P. R. China}
24:
25: \email{hldai@nju.edu.cn; lixd@nju.edu.cn; liuxw@nju.edu.cn}
26:
27:
28: \date{}
29:
30: %\address{Astronomy Department, University of California, Berkeley, CA 94720}
31: \begin{document}
32:
33: %\maketitle
34:
35: \begin{abstract}
36: We have investigated the relation between the orbital periods
37: ($\pb$) and the spin periods ($\ps$) of wind-fed X-ray pulsars in
38: high-mass X-ray binaries (HMXBs), based on population synthesis
39: calculations of the spin evolution of neutron stars during the
40: pre-HMXB stage. We show that most of the neutron stars either have
41: steady accretion or still reside in the radio pulsar phase when
42: the donor star starts evolving off the main sequence. In the
43: former case the values of $\ps$ can be decelerated to be $\sim
44: 10^2-10^3$ s depending on $\pb$.
45: %Our results are in rough agreement with the distributions of HMXBs
46: %in the $\ps-\pb$ diagram.
47: We briefly discuss the possible origin of the $\ps-\pb$
48: correlation in Be/X-ray binaries, and the existence of HMXBs with
49: main sequence donors. We have also investigated the evolution of
50: the inclination angle between the magnetic and spin axes of
51: neutron stars in a massive binary, suggesting secular alignment of
52: the magnetic and spin axes during their evolution.
53: \end{abstract}
54:
55: \keywords{stars: early-type --- X-rays: binaries --- stars: evolution --- stars: neutron}
56:
57: \section{INTRODUCTION}
58: High-mass X-ray binaries (HMXBs) can be roughly divided into two
59: types: a supergiant star and a Be star are contained respectively.
60: The X-ray source often is a pulsar and powered by accretion of the
61: material offered by the companion star. In the supergiant systems,
62: either Roche-lobe or stellar wind accretion occurs, while in the
63: Be systems commonly only the latter process takes place since the
64: Be star is well inside its Roche lobe \citep{tau05}.
65:
66: The relation between the spin periods $\ps$ and the orbital
67: periods $\pb$ of HMXBs was first studied by
68: \citet{cor84,cor85,cor86}, who pointed out that there may be a
69: $\ps\propto P_{\rm orb}^{4/7}$ correlation for neutron stars in
70: systems with radially expanding winds, while $\ps\propto P_{\rm
71: orb}^{2}$ for Be systems. \citet{van87} noticed that the spin and
72: orbital periods agree reasonably well with the former correlation
73: for the supergiant systems in which the X-ray source is powered by
74: stellar wind accretion. The $\ps-\pb$ relation has been studied by
75: \citet{ste86} in terms of the equilibrium spin period, where the
76: corotation radius equals to the magnetospheric radius defined by
77: equating the ram pressure of accreting matter with the magnetic
78: pressure of the dipole field of the neutron star
79: \citep[e.g.,][]{dav73,lam73}. However, in order to obtain
80: quantitative agreement between the predicted and observed spin
81: periods for given orbital periods, the assumption of the
82: equilibrium spin periods requires a mass accretion rate two orders
83: of magnitude lower than observed for supergiant systems
84: \citep{ste86}. \citet{wat89} presented a comprehensive explanation
85: for the $\ps-\pb$ relation for HMXBs. For supergiant systems they
86: suggested that the difference between the required and the
87: observed mass accretion rates can be explained by assuming that
88: the present-day periods are the equilibrium spin periods for the
89: stellar wind of the optical star when it was still on main
90: sequence (MS). The $\ps-\pb$ correlation for Be/X-ray binaries was
91: suggested to result from the equilibrium spin realized by
92: interaction between the neutron star magnetic field and the Be
93: star's equatorial wind - the wider orbits expose to the neutron
94: stars to a lower wind density on average and hence lower accretion
95: rate. The latter point was criticized by \citet{king91}, who
96: proposed that the observed $\ps-\pb$ relation was probably the
97: result of an earlier evolutionary stage, as for supergiant
98: binaries. By calculating the angular momentum carried by the Be
99: star's wind material, \citet{li96} showed that for Be/X-ray
100: binaries in relatively narrow orbits ($\pb\lsim$ 100 days), the
101: equilibrium spin could be reached by angular momentum transfer via
102: a disk, formed in the equatorial wind of the Be stars; in wider
103: binaries, the low angular momentum of the wind matter prevents the
104: existence of an accretion disk, and the neutron star's spin
105: remains close to its previous equilibrium value at an earlier
106: evolutionary stage. \citet{zha04} have calculated the detailed
107: spin evolution of a wind-fed neutron star in OB binaries prior to
108: the HMXB phase, by simulating the time-development of the mass
109: loss rate and radius expansion of a 20 $M_{\odot}$ donor star.
110:
111: The above-mentioned investigations are usually either
112: semi-analytic or numerical only for individual cases. To better
113: understand the distribution of HMXBs in the $\ps-\pb$ diagram, an
114: evolutionary population synthesis incorporated with neutron star
115: spin evolution is needed. In the present paper, we have conducted
116: a Monte Carlo population synthesis study of the spin evolution of
117: a neutron star in a massive binary. Due to the large theoretical
118: uncertainties on the formation and evolution of HMXBs, our work is
119: limited to the spin evolution of neutron stars before the HMXB
120: phase. The theoretical considerations are described in \S 2. The
121: calculated results are presented in \S 3 and their implications
122: for accretion processes in HMXBs are discussed in \S 4.
123:
124: \section{MODEL}
125: \subsection{Spin evolution}
126:
127: We consider a $1.4 M_{\odot}$ magnetized neutron star in binaries
128: with a massive MS companion star. We adopt a simplified version of
129: the theoretical model outlined by \citet{dav81} to calculate the
130: neutron star's spin evolution before steady accretion occurs, as
131: briefly described below.
132:
133: \noindent{\em Case a: the pulsar phase}
134:
135: The newborn neutron star first appears as a rapidly rotating,
136: radio pulsar after its birth in a supernova explosion, provided
137: that its radiation is strong enough to expel the wind material
138: coming from the companion star outside the radius of the light
139: cylinder $r_{\rm lc}=c\ps/2\pi$ or the Bondi accretion radius
140: $r_{\rm G}=2GM/v_{\infty}^2$ \citep{bon44}, where $G$ is the
141: gravitational constant, $M$ the mass of the neutron star and
142: $v_{\infty}=10^8v_8$\,cms$^{-1}$ the relative wind velocity at the
143: neutron star's orbit. The neutron star spins down due to magnetic
144: dipole radiation and/or particle emission,
145: \begin{equation}
146: I\dot{\Omega}_{\rm s}=-\frac{2}{3}\frac{\mu^2\Omega_s^3}{c^3},
147: \end{equation}
148: where $I$ is the moment of inertia, $\mu=10^{30}\mu_{30}$\,Gcm$^3$
149: the magnetic dipole moment, and $\Omega_{\rm s}$ the angular
150: velocity of the neutron star, respectively.
151:
152: The pulsar phase will break down in two distinct circumstances.
153:
154: (i) The pulsar phase will cease when the wind material penetrates
155: inside $r_{\rm lc}$, and the resulting $\ps$ is derived by
156: balancing radiation pressure from the pulsar with the stellar wind
157: ram pressure at $r_{\rm lc}$ \citep{dav81},
158: \begin{equation}
159: P_{\rm a} \simeq 0.8 \mu_{30}^{1/3} \dot{M}_{15}^{-1/6}
160: (M/M_{\odot})^{1/3}v_8^{-5/6}\,{\rm s},
161: \end{equation}
162: where $\dot{M}=10^{15}\dot{M}_{15}$ \,gs$^{-1}$ is the mass flow
163: rate onto the neutron star.
164:
165: (ii) The pulsar phase will end when the pressure gradient become
166: important as the outer boundary $R_{\rm a}$ of the envelope where
167: the density $\rho\simeq \rho_{\rm \infty}$ approaching to $r_{\rm
168: G}$, i.e., $R_{\rm a}\simeq r_{\rm G}$, and the corresponding
169: $\ps$ is \citep{dav81},
170: \begin{equation}
171: P_{\rm a}\simeq
172: 1.2\dot{M}_{15}^{-1/4}\mu_{30}^{1/2}v_8^{-1/2}\,{\rm s}.
173: \end{equation}
174:
175:
176: %The pulsar phase will cease when the wind material penetrates
177: %inside either $r_{\rm lc}$ or $r_{\rm G}$. The resulting $\ps$ is
178: %then \citet{dav81},
179: %\begin{equation}
180: %P_{\rm a} = 0.8 \mu_{30}^{1/3} \dot{M}_{15}^{-1/6}
181: %(M/M_{\odot})^{1/3}v_8^{-5/6}\,{\rm s},
182: %\end{equation}
183: %or
184: %\begin{equation}
185: %P_{\rm a}=1.2\dot{M}_{15}^{-1/4}\mu_{30}^{1/2}v_8^{-1/2}\,{\rm s},
186: %\end{equation}
187: %where $\dot{M}=10^{15}\dot{M}_{15}$ \,gs$^{-1}$ is the mass
188: %accretion rate.
189:
190: \noindent{\em Case b: the propeller phase}
191:
192: The propeller phase follows the pulsar phase. During this period
193: the magnetospheric radius $\rn=[\mu^4/(2GM\dot{M}^2)]^{1/7}$, is
194: larger than the corotation radius $\rc=(GM/\Omega_{\rm
195: s}^2)^{1/3}$. The plasma interacts with the neutron star
196: magnetosphere, but further accretion is inhibited by the
197: centrifugal barrier, and the infalling matter is accelerated
198: outward, taking away the angular momentum of the neutron star.
199: Though the propeller effect has be investigated extensively, there
200: exist large uncertainties in the efficiency of angular momentum
201: loss during the propeller regime
202: \citep[e.g.][]{pr72,is75,dav81,wan85,ik01}. Here we assume that
203: the infalling material is ejected with the corotating velocity at
204: $\rn$, and the spin-down torque is \citep{wan85,jia05}
205: \begin{equation}%\label{}
206: N=I\dot{\Omega}_{\rm s}=-\dot{M}R_{\rm m}^2\Omega_{\rm s}.
207: \end{equation}
208: The typical spin-down time-scale $\tau=|\Omega_{\rm
209: s}/\dot{\Omega}_{\rm s}|$ can be estimated to be
210: \begin{equation}
211: \tau\simeq 2.2\times
212: 10^4\mu_{30}^{-8/7}\dot{M}_{15}^{-3/7}(M/M_{\odot})^{2/7}I_{45}\,{\rm
213: yr}.
214: \end{equation}
215: The spin-down process ends until $\ps$ reaches the equilibrium
216: spin period
217: \begin{equation}
218: P_{\rm eq}\simeq
219: 17\mu_{30}^{6/7}\dot{M}_{15}^{-3/7}(M/M_{\odot})^{-5/7}\,{\rm s},
220: \end{equation}
221: and we assume that steady accretion takes place afterwards
222: \citep[see however,][]{al76,el76}.
223:
224: The spin period of the neutron star may be further changed after
225: $P>P_{\rm eq}$ in the following accretion phase. But we stop the
226: calculations when either $P_{\rm eq}$ is reached within the MS
227: lifetime or the companion star evolves off the MS (so that strong
228: wind accretion or Roche lobe overflow occurs) \citep[see
229: also][]{wat89}. In this work we do not consider narrow HMXBs with
230: Roche-lobe overflow (like SMC X$-$1, Cen X$-$3, and LMC X$-$4),
231: where the neutron stars are most likely to accrete from an
232: accretion disk. For wind-fed systems like Vela X$-$1, numerical
233: calculations \citep[e.g.][]{fry88,mat92,anz95,ruf99} have
234: suggested that there are no significant angular momentum transfer
235: onto the neutron star when radially-expanding wind material
236: accretes onto the neutron star. This may result in only small
237: deviation from the instantaneous (equilibrium if reached) spin
238: periods at the beginning of the accretion phase. {\em CGRO/BATSE}
239: observations have shown a random walk in their spin frequencies
240: with alternating spin-up and spin-down \citep{bil97}. The spin
241: evolution in Be/X-ray pulsars has a completely different picture,
242: and will be discussed separately.
243:
244: \subsection{Evolution of the inclination angle}
245: Besides the spin evolution, we have considered the evolution of
246: the inclination angle $\chi$ between the spin and magnetic axes in
247: wind-fed neutron stars. This part of work is partially motivated
248: by \citet{bul03}. By analyzing the light curves of 89
249: accretion-powered pulsars these authors obtained an upper limit on
250: the inclination angles $\chi<50\degr$, suggesting that the
251: magnetic axis tended to be aligned with the rotation axis. Similar
252: analyzes and conclusions have also been conducted by
253: \citet{wan81}, \citet{lea91}, and \citet{lea95}. More recently
254: \citet{bla05} presented the contemporaneous high-energy and radio
255: observations of the HMXB 4U 2206+54 conducted with
256: \textit{INTEGRAL} and VLA, which firmly indicate that 4U 2206+54
257: hosts a magnetic accreting neutron star. The absence of pulsations
258: are most likely to be due to geometrical effect that the
259: inclination angle of the neutron star is very small so that no
260: pulsation can be detected.
261:
262: \citet{wan82} show that the inclination angle in a binary X-ray
263: pulsar will vary as a result of its interaction with the
264: surrounding plasma, becoming smaller (larger) during spin-down
265: (spin-up). According to Eqs. (27), (36c), and (37c) in \citet
266: {wan82}, we can derive the changing rate of $\chi$ as follows,
267: \begin{eqnarray}
268: \dot{\chi}&=&\left\{
269: \begin{array} { ll}
270: N\cot\chi/(I\Omega_{\rm s}),&\ \chi\gtrsim\theta_{\rm c}\\
271: 3N\sin\chi/(I\Omega_{\rm s}\sin^2\theta_{\rm c}),&\
272: \chi\lesssim\theta_{\rm c}
273: \end{array}
274: \right.
275: \end{eqnarray}
276: where $\theta_c$ is the critical value of the meridional angle
277: measured from the north magnetic pole.
278:
279: \subsection{Evolution of the mass flow rate onto neutron star}
280:
281: We employed an evolutionary population synthesis method to
282: investigate the distribution of the orbital periods and companion
283: masses for the natal neutron star binaries (i.e. at the moment
284: when the neutron star was born). We started with a large set of
285: primordial binaries and generated the systems that evolved to
286: contain a neutron star and a MS companion. The initial mass
287: function of \citet{Kro93} was taken for the masses of the primary
288: star (the progenitor of the neutron star of mass $M_1$). For the
289: secondary star (of mass $M_2$), we assumed a uniform distribution
290: of the mass ratio for $0<q\equiv M_2/M_1\le 1$. A uniform
291: distribution of $\ln a$ was also taken for the binary separation
292: $a$. We adopted the binary population synthesis code developed by
293: \citet{hur00,hur02} to evolve the primordial binaries. This code
294: incorporates evolution of single stars with binary-star
295: interactions, such as mass transfer, mass accretion,
296: common-envelope (CE) evolution, collisions, supernova kicks, tidal
297: friction and angular momentum loss mechanisms. Most of our
298: adopted parameters are the same as the standard ones described in
299: \citet{hur02}. The star formation rate parameter is $S = 7.6085$
300: yr$^{-1}$, corresponding to a rate of $\sim 0.02$ yr$^{-1}$ for
301: core-collapse supernovae (SNe) in our Galaxy, assume all the stars
302: with masses $> 8M_{\odot}$ die through SNe. During the SN
303: explosions, we apply a Maxwellian distribution in the kick
304: velocities with a mean of 265 kms$^{-1}$ imparted on the newborn
305: neutron stars \citep{hobb05}. The treatment of Roche-lobe overflow
306: (RLOF) mass transfer in the primordial binary is presented
307: specifically in \citet{hur02}, and here we describe the stability
308: criterion of mass transfer briefly. Mass transfer via Roche-lobe
309: overflow takes place on either nuclear, thermal, or dynamical
310: time-scales, depending on whether the primary remains in thermal
311: equilibrium as it loses mass, and the radius of the primary
312: increases faster than the Roche-lobe. Stars with deep surface
313: convective zones, e.g. giants or naked helium giants, are
314: generally unstable to dynamical timescale mass loss and will enter
315: a CE evolution. The stable mass accretion rate of the secondary
316: star is limited by Eddington accretion rate. Actually, the
317: secondary may be spun up and become a Be star when it accretes
318: enough mass. This circumstance however, is not included in our
319: consideration since the origin of Be phenomena is still unclear
320: and it is difficult to model the mass transfer processes in
321: Be/X-ray binaries. The CE parameter $\alpha$ was set to be 1 as a
322: typical value and we varied it from 0.1 to 2 in the calculations
323: \citep{dew00, tau01}. Our product is a set of ``incipient" neutron
324: star in massive binaries with a particular distribution of $\pb$
325: and $M_{2}$ (shown in Fig.~1). Other binary parameters, such as
326: the radii, surface temperatures and luminosities of the companion
327: stars can also be obtained. These parameters were then used to
328: evaluate the mass loss rates from the companion stars and the mass
329: flow rates onto the neutron stars.
330:
331: For comparison with the observed properties of HMXBs, our
332: calculations are limited to systems with $\pb<1000$ days and
333: $10M_{\odot}\lsim M_2\lsim 30M_{\odot}$. The mass loss rate
334: $\dot{M}_2$ was estimated with the prescription described by
335: \citet{nie90},
336: \begin{equation}
337: -\dot{M}_2=9.6\times10^{-15}
338: R_2^{0.81}L_2^{1.24}M_2^{0.16}\,M_{\odot}{\rm yr}^{-1},
339: \end{equation}
340: where $R_2$ and $L_2$ are the radius and luminosity of the
341: companion star. All the quantities in Eq.~(8) are evaluated in
342: solar units. Assume that the stellar wind expands isotropically at
343: a speed of $v_{\rm w}$, the wind density $\rho_{\rm w}$ at the
344: orbit of the neutron star is
345: \begin{equation}
346: \rho_{\rm w}=-\dot{M}_2/(4\pi a^2v_{\rm w}),
347: \end{equation}
348: and the mass flow rate onto neutron star is roughly given by
349: \citep{bon44}
350: \begin{equation}
351: \dot{M}=\pi r_{\rm G}^2\rho_{\rm w}v_{\infty}.
352: \end{equation}
353:
354:
355: \section{RESULTS}
356:
357: We have done calculations of the spin and inclination evolutions
358: for $6\times10^5$ neutron star binary systems based on the
359: theoretical model presented in \S 2. For the initial neutron star
360: magnetic fields $B$ we assumed that $\log B$ is distributed
361: normally with a mean of 12.5 and a standard deviation of 0.3. No
362: field decay is considered. The initial distribution of the
363: inclination angle $\chi$ is randomly distributed in the range of
364: $(0,\pi/2)$. We set the initial spin periods to be distributed
365: uniformly between $10$ and $100$ ms, though they have little
366: influence on the final results. We stop our calculations when
367: either $\ps$ reaches $P_{\rm eq}$ or the companion star begins to
368: evolve off the MS with the arguments afore.
369:
370: The calculated results are presented in Figs.~2$-$5. Panels (a)
371: and (b) in Figs.~2$-$3 correspond to the relative wind velocity at
372: the neutron star's orbit $v_8=1$ and 2 respectively. Figure 2
373: shows the final distribution of all the binaries in the $\ps-\pb$
374: diagram. The relative numbers of binary systems are displayed
375: sketchily with the darkness of the shaded regions. According to
376: our calculation, we find that, when $v_8=1$, around $68\%$ neutron
377: stars can reach the equilibrium periods to allow wind accretion
378: within the MS lifetime $t_{\rm MS}$ , while others would still be
379: in either the pulsar ($\sim 31\%$) or the propeller ($\sim 1\%$)
380: phases. This can be clearly seen in Fig~2, where the former two
381: groups of neutron stars occupy the distinct upper and lower shaded
382: regions. The small number of stars in the propeller phase
383: originates from its much shorter duration \citep{dav81}. If $v_8$
384: is increased to be 2, the mass flow rates onto the neutron stars
385: will be lower by a factor of $\sim 16$ according to Eqs.~(9) and
386: (10), further extending the spin-down time in the pulsar phase. So
387: the corresponding numbers become $\sim 42\%$, 55\% and 3\%,
388: respectively. The larger wind velocity also induces longer
389: equilibrium periods. Figure 2 reveals that up to half of neutron
390: stars in binary systems cannot reach the equilibrium periods,
391: indicating that there might be hundreds of ``sleeping" neutron
392: star $+$ MS star binaries like PSR B1259$+$63 in the Galaxy (see
393: also Fig.~4 below). They could be observed as radio pulsars or in
394: X-rays depending on whether the interaction between the neutron
395: star and the companion's winds becoming active. Most of them have
396: $\ps\sim 1$ s and $\pb$ from tens of to hundreds of days.
397:
398: To compare the calculated results with observations of HMXBs, we
399: show the distribution of those neutron star binaries with $\ps\geq
400: P_{\rm eq}$ and of the observed HMXBs in Fig.~3. Asterisks and
401: diamonds mark the Be and supergiant wind-fed HMXBs respectively,
402: and crosses for Roche-lobe overflow systems \citep[data are taken
403: from][]{rag05}. We first discuss the persistent, supergiant
404: systems. Obviously the spin periods of neutron stars can approach
405: $P_{\rm eq}$ if the total spin-down timescale $t_{\rm spin}$ is
406: less than the MS lifetime $t_{\rm MS}$ of the companion star. As
407: we argued before, for this type of systems, the current spin
408: periods are likely to be around the equilibrium periods $P_{\rm
409: eq}$ (MS) attained during the MS stage because of the inefficient
410: angular momentum transfer in the accretion phase. For these
411: systems to be observed as HMXBs, the accretion rate must have been
412: increased to be high enough that the spin periods are larger than
413: the (current) equilibrium period $P_{\rm eq}$ (sg) when the
414: companion star becomes a supergiant \citep{ste86,wat89}, which is
415: plotted with the dotted line in Fig.~3 according to Eq.~(6). Here
416: we take $M_2=20M_{\odot}$, $\dot{M}_2=10^{-6}\,\my$, and
417: $\mu_{30}=3$ \citep{cob04} as typical values of the supergiant
418: systems. This period roughly serves as the lower limit for $\ps$
419: of the neutron stars in supergiant HMXBs. Moreover, since the mass
420: accretion rates generally decrease with $\pb$, only those in
421: narrow orbits ($\pb\lesssim 15$ d) can have X-ray luminosities
422: $\gtrsim 10^{36}$ ergs$^{-1}$ as observed.
423:
424: %The reason that we consider both $P_{\rm eq}$ and $P_{\rm br}$ is
425: %that the latter was derived under the assumption that there is a
426: %static quasi-spherical envelope around the neutron star
427: %\citep{dav81}, which may not be appropriate for accreting X-ray
428: %pulsars in supergiant HMXBs, which are usually surrounded by a
429: %small disk even in wind-accretion phase \citep{fry88}. Similar
430: %situations exist in Be/X-ray binaries. Actually we can see from
431: %Fig.~4 that $P_{\rm eq}$ in accretion phase may be more suitable
432: %as the lower limit of $\ps$ for both types of systems.
433:
434: Be/X-ray binaries generally have much longer orbital periods than
435: the supergiant systems, and are usually transient X-ray sources
436: observable during outbursts. The structure of the Be star's winds
437: is quite complicated, consisting of a relatively dense, slowly
438: expanding, disk-like, equatorial wind and a fast, isotropically
439: expanding, polar wind \citep{wat89}. \citet{wat89} suggested that
440: the $\ps-\pb$ correlation for Be/X-ray binaries could be explained
441: by setting $\ps=P_{\rm eq}$ obtained in the disk winds. However,
442: current understanding of the Be disk winds \citep{oka01} is very
443: different from the radial wind model used by \citet{wat89}. The
444: evolution of the Be star's disk and the high eccentricities of the
445: orbits also mean that the mass flow rate onto neutron stars in Be
446: systems are always highly variable on both long-term and
447: short-term timescales. Because of these complications we did not
448: calculated the spin evolution of neutron stars accreting from the
449: disk winds, but present constraints on their possible locations in
450: the $\ps-\pb$ diagram instead. We first note that the calculated
451: $P_{\rm eq}$ (MS), under the assumption of isotropic winds, can be
452: regarded as the upper limits for $\ps$ of Be/X-ray binaries, since
453: $\dot{M}$ is much lower from the polar winds than from the disk
454: winds (these values should be taken as the averaged ones since we
455: have ignored the eccentricity of the Be star binaries). Then,
456: similar as in the supergiant systems, we can set the lower limit
457: of $\ps$ for Be/X-ray binaries to be the equilibrium period
458: $P_{\rm eq}$ (disk) in outbursts when the the disk winds dominate
459: accretion. From the observed data of 36 Be/X-ray binaries compiled
460: by \citet{rag05}, we have derived a correlation between the
461: maximum luminosities and the orbital periods,
462: \begin{equation}
463: \log (L_{\rm X,max}/10^{35}\,{\rm ergs}^{-1})=4.53(\pm 0.66)
464: -1.50(\pm 0.33)\log(\pb/{\rm d}).
465: \end{equation}
466: Insert Eq.~(11) and $\mu_{30}=3$ into Eq.~(6) we obtain
467: \begin{equation}
468: \log P_{\rm eq}({\rm disk})\simeq -0.29+0.64\log(\pb/{\rm d}),
469: \end{equation}
470: %\begin{equation}
471: %\log P_{\rm br}({\rm e})=-0.29+1.1\log(\pb/{\rm d}).
472: %\end{equation}
473: which is plotted in the solid line in Fig.~3. We see that a large
474: fraction of Be/X-ray binaries lie between $P_{\rm eq}$ (MS) and
475: $P_{\rm eq}$ (d). The peculiar location of A 0535$-$669 may be due
476: to its relatively low magnetic field.
477:
478: The cumulative luminosity distributions of HMXBs based on our
479: population synthesis calculations are shown in Fig.~4 with
480: different values of the parameters $\alpha$, $v_8$, and $\sigma$.
481: The results show that change of these parameters do not
482: significantly influence the final outcome, which are roughly
483: compatible with the observed luminosity distributions of HMXBs in
484: our Galaxy \citep{gri02}.
485:
486: In Fig.~5 we plot the inclination angle distribution for X-ray
487: pulsars in HMXBs. It is easily seen that most of the inclination
488: angles are within 1 rad, in general agreement with the analysis of
489: the observation data \citep{bul03}. The evolution of the
490: inclination angle seems to be insensitive to the values of the
491: relative wind velocity.
492:
493: \section{DISCUSSION}
494: We have calculated the spin evolution of neutron stars in a
495: massive binary. The main ideas in this work are of two folds.
496: First, to appear as X-ray binaries, the spin periods of neutron
497: stars should be longer than the current (instantaneous)
498: equilibrium periods. Second, to satisfy this condition, the
499: neutron stars should be spun-down enough during the previous
500: propeller phase when the companion stars had relatively weak
501: winds. These two requirements set constraints on the spin period
502: distribution in wind-fed X-ray pulsars. Since $\dot{M}$ depends on
503: the orbital separations, this may result in the particular
504: distributions of the supergiant and Be HMXBs in the $\ps-\pb$
505: diagram. The numerical results presented in Figs.~2 and 3, roughly
506: accounting for the main features of the $\ps-\pb$ distribution of
507: HMXBs, lend support to the original suggestions by \citet{cor84},
508: \citet{ste86}, and \citet{wat89}. Our calculations of the
509: alignment during spin-down imply that most HMXBs may have
510: relatively small inclination angles.
511:
512: We note that our preliminary results are subject of lots of
513: uncertainties and simplified treatment adopted. In case of
514: isotropic wind accretion, a considerable fraction of neutron stars
515: can reach $P_{\rm eq}$ when the companion star is still on the MS.
516: The detail number is determined not only by the binary
517: evolutionary processes, but also critically by the propeller
518: mechanism. The spin-down torque (Eq.~[4]) adopted in this work is
519: among the most efficient ones \citep[][and references
520: therein]{jia05}, and the results should be taken as to be the most
521: optimistic. For less efficient spin-down torques ones, the number
522: of binaries in the propeller phase will obviously increase due to
523: the longer spin-down timescales. The magnitude of $P_{\rm eq}$
524: depends on the mass loss rate, the wind velocity, and the magnetic
525: field strength of the neutron star. With the mass loss rates given
526: by Eq.~(9) and typical magnetic fields of $\sim 3\times 10^{12}$
527: G, the neutron star's spin can be decelerated to be $\sim
528: 10^2-10^3$ s within $t_{\rm MS}$ of the secondary star. The very
529: long period $\ps=10^4$ s of 2S 0114$+$650 \citep{hal00}, however,
530: may be explained by a ultra-strong initial magnetic field ($B\gsim
531: 10^{14}$ G), in order to be spun-down efficiently by the propeller
532: effect \citep{li99}.
533:
534: The structure of the Be star's winds is much more complex and
535: variable than in supergiant systems, and the simple propeller
536: spin-down may not be applicable to this type of systems. It is
537: likely that the $\ps-\pb$ distribution and correlation of Be/X-ray
538: binaries result from the balance between the spin-up during
539: outbursts and spin-down during quiescence. Assume that the spin-up
540: and spin-down torques are $\dot{M}_{\rm o}(GMR_{\rm mo})^{1/2}$
541: and $-\dot{M}_{\rm q}R_{\rm mq}^2\Omega_{\rm s}$ respectively, we
542: can derive the following expression for the equilibrium period
543: \citep[see also][]{men99},
544: \begin{equation}
545: P_{\rm eq}=2\pi(\frac{GM}{R_{\rm mo}^3})^{-1/2}(\frac{\dot{M}_{\rm
546: q}R_{\rm mq}^2}{\dot{M}_{\rm o}R_{\rm mo}^2})(\frac{t_{\rm
547: q}}{t_{\rm o}})\propto \dot{M}_{\rm o}^{-3/7}(\frac{\dot{M}_{\rm
548: q}}{\dot{M}_{\rm o}})^{3/7}(\frac{t_{\rm q}}{t_{\rm o}}).
549: \end{equation}
550: Here the subscripts o and q denote quantities evaluated during
551: outbursts and quiescence, respectively. All the three terms on the
552: right hand side of Eq.~(13) are likely to increase with $\pb$,
553: which might account for the observed $\ps-\pb$ correlation of
554: Be/X-ray binaries.
555:
556: Our calculated results also indicate the existence of HMXBs with MS
557: donors in the Galaxy. These systems, as the progenitors of
558: supergiant HMXBs, are less luminous than the latter (but with
559: similar spin periods). Figure 4 suggests that there could be a few
560: hundred of these sources in the Galaxy with X-ray luminosities
561: ranging from $\sim 10^{33}$ to $10^{35}$ ergs$^{-1}$ (If the
562: propeller effect is not considered, the total number of X-ray
563: binaries, most of which are Be/X-ray binaries, can reach a few
564: thousand).
565: %however, the numbers of HMXBs with MS donors not only in high
566: %luminosity but in low luminosity are much larger than the numbers
567: %of HMXBs with supergiant donors according to the results of
568: %population synthesis. The results in Fig. 4. are consistent well
569: %with the observations \citep{gri02}, which indicate that there are
570: %about 1000 HMXBs with luminosity larger than $10^{33}$ ergs$^{-1}$
571: %when we take the typical values of $\alpha=2$, $v_8=2$, and
572: %$\sigma=190$ kms$^{-1}$ respectively and may be detected in
573: %sensitive X-ray surveys.
574: %In Fig.~6 we show the distributions of
575: %their X-ray luminosities and X-ray lifetimes. Most of them can
576: %shine in X-rays with luminosities of $\sim 10^{32}-10^{36}$
577: %ergs$^{-1}$ for a few million years.
578: The HMXB 4U2206$+$54 is likely to be the prototype of this kind of
579: sources \citep{ribo05}. A number of papers have also suggestted
580: that some neutron stars receive low kick speeds of $\leq50$
581: kms$^{-1}$ at birth \citep{pfahl02, pod04, dew05}. If all the
582: neutron stars are born with such small kicks, our calculations
583: show that there would be about $4-5$ times more X-ray sources
584: produced.
585: %In this way, the assumed natal kick distribution of our
586: %model doesn't have significant
587: %effect on our results. %\citet{wan02} have reported hundreds of
588: %X-ray sources with luminosities $\sim 10^{33}-10^{35}$ ergs$^{-1}$
589: %in the the Galaxy center region, which contains $\lsim 1\%$ of the
590: %total Galactic population.
591: %\citet{pfa02} attributed a significant
592: %fraction of these X-ray sources to neutron stars accreting from
593: %the winds of the MS donor stars. Our calculations, taking into
594: %account neutron star spin evolution, suggest that HMXBs with MS
595: %donors could contribute at most a few of these sources \citep[see
596: %also][]{liu06}.
597: % However, more
598: %detailed investigation by \citet{liu06} shows that the number of
599: %these sources falls short of the observed ones if the propeller
600: %effect is taken in account. It remains to see how much they could
601: %contribute to the Galactic ridge X-ray emission along with
602: %coronally active binaries and cataclysmic variables \citep{saz05}.
603:
604: \acknowledgments We are grateful to an anonymous referee for careful
605: reading the manuscript and constructive comments that significantly
606: improved the manuscript. This work was supported by Natural Science
607: Foundation of China under grants 10025314 and 10573010.
608:
609:
610: \begin{thebibliography}{}
611: \bibitem[Anzer \& B\"orner(1995)]{anz95}
612: Anzer, U., \& B\"orner, G. 1995, A\&A, 299, 62
613:
614: \bibitem[Arons \& Lea(1976)]{al76}
615: Arons, J., \& Lea, S. M. 1976, ApJ, 207, 914
616:
617: \bibitem[Blay et al.(2005)]{bla05}
618: Blay, P., Rib$\acute{o}$, M., Negueruela, I., Torrej$\acute{o}$n,
619: J. M., Reig, P. et al. 2005, 438, 963
620:
621: \bibitem[Bildsten et al.(1997)]{bil97}
622: Bildsten, L. et al. 1997, ApJS, 113, 367
623:
624: \bibitem[Bondi \& Hoyle(1944)]{bon44}
625: Bondi, H., \& Hoyle F. 1944, MNRAS, 104, 273
626:
627: \bibitem[Bulik et al.(2003)]{bul03}
628: Bulik, T., Gondek-Rosi$\acute{n}$ska, D., Santangelo, A., Mihara,
629: T., Finger, M. et al. 2003, A\&A, 404, 1023
630:
631: \bibitem[Coburn et al.(2004)]{cob04}
632: Coburn, W. et al. 2004, ApJ, 580, 394
633:
634: \bibitem[Corbet(1984)]{cor84}
635: Corbet, R. H. D. 1984, A\&A, 141, 91
636:
637: \bibitem[Corbet(1985)]{cor85}
638: Corbet, R. H. D. 1985, SCR, 40, 409
639:
640: \bibitem[Corbet(1986)]{cor86}
641: Corbet, R. H. D. 1986, MNRAS, 220, 1047
642:
643: \bibitem[Davidson \& Ostriker(1973)]{dav73}
644: Davidson, K., \& Ostriker, J. P. 1973. Astrophys. J., 179, 585
645:
646: \bibitem[Davies \& Pringle(1981)]{dav81}
647: Davies R. E., \& Pringle J. E. 1981, MNRAS, 151, 351
648:
649: \bibitem[Dewi, Podsiadlowski \& Pols(2005)]{dew05}
650: Dewi, J. D. M., Podsiadlowski, Ph., \& Pols, O. R. 2005, MNRAS,
651: 368, 1742
652:
653: \bibitem[Dewi \& Tauris (2000)]{dew00}
654: Dewi, J. D. M., \& Tauris, T. M. 2000, A\&A, 360, 1043
655:
656: \bibitem[Elsner \& Lamb(1976)]{el76}
657: Elsner, R. F., \& Lamb, F. K. 1976, Nat, 262, 356
658:
659: \bibitem[Fryxell \& Taam(1988)]{fry88}
660: Fryxell, B. A., \& Taam, R. E. 1988, ApJ, 335, 862
661:
662: %\bibitem[Galloway, Wang \& Morgan(2005)]{gal05}
663: %Galloway D. K., Wang Z. X., Morgan E. H., 2005, submitted to ApJ
664:
665: \bibitem[Grimm et al.(2002)]{gri02}
666: Grimm, H.-J., Gilfanov, M., \& Sunyaev, R. 2002, A\&A, 391, 923
667:
668: \bibitem[Hall et al.(2000)]{hal00}
669: Hall, T. A. et al. 2000, \apj, 536, 450
670:
671: \bibitem[Hobbs et al.(2005)]{hobb05}
672: Hobbs, G., Lorimer, D. R., Lyne, A. G., \& Kramer, M. 2005, MNRAS,
673: 360, 963
674:
675: \bibitem[Hurley et al.(2000)]{hur00}
676: Hurley, J. R., Pols, O. R., \& Tout, C. A. 2000, MNRAS, 315, 543
677:
678: \bibitem[Hurley et al.(2002)]{hur02}
679: Hurley, J. R., Tout, C. A., \& Pols, O. R. 2002, MNRAS, 329, 897
680:
681: \bibitem[Ikhsanov(2001)]{ik01}
682: Ikhsanov, N. R. 2001, A\&A, 368, L5
683:
684: \bibitem[Illarionov \& Sunyaev(1975)]{is75}
685: Illarionov, A. F., \& Sunyaev R. A. 1975, A\&A 39, 185
686:
687: \bibitem[Jiang \& Li(2005)]{jia05}
688: Jiang, Z.-B., \&, Li, X.-D. 2005, ChJA\&A, 5, 487
689:
690: \bibitem[King(1991)]{king91}
691: King, A. R. 1991, MNRAS, 250, 3
692:
693: \bibitem[Kroupa, Tout \& Gilmore(1993)]{Kro93}
694: Kroupa, P., Tout, C. A., \& Gilmore, G. 1993, MNRAS, 262, 545
695:
696: \bibitem[Lamb, Pethick \& Pines(1973)]{lam73}
697: Lamb, F. K., Pethick, C. J., \& Pines, D. 1973. ApJ, 184, 271
698:
699: \bibitem[Leahy(1991)]{lea91}
700: Leahy, D. A. 1991, MNRAS, 251, 203
701:
702: \bibitem[Leahy \& Li(1995)]{lea95}
703: Leahy, D. A., \& Li, L. 1995, MNRAS, 277, 1177
704:
705: \bibitem[Li \& van den Heuvel(1996)]{li96}
706: Li, X.-D., \& van den Heuvel, E. P. J. 1996, A\&A, 314, L13
707:
708: \bibitem[Li \& van den Heuvel(1999)]{li99}
709: Li, X.-D., \& van den Heuvel, E. P. J. 1999, \apj, 513, L45
710:
711: %\bibitem[Liu \& Li(2006)]{liu06}
712: %Liu, X.-W. \& Li, X.-D. 2006, A\&A, 449, 135
713:
714: \bibitem[Matsuda et al.(1992)]{mat92}
715: Matsuda, T. et al. 1992, MNRAS, 255, 183
716:
717: \bibitem[Menou et al.(1999)]{men99}
718: Menou, K. et al. 1999, ApJ, 520, 276
719:
720: \bibitem[Nieuwenhuijzen \& de Jager(1990)]{nie90}
721: Nieuwenhuijzen, H., \& de Jager, C. 1990, A\&A, 231, 134
722:
723: \bibitem[Okazaki \& Negueruela(2001)]{oka01}
724: Okazaki, A. T., \& Negueruela, I. 2001, A\&A, 377, 161
725:
726: \bibitem[Pfahl et al.(2002)]{pfahl02}
727: Pfahl, E., Rappaport, S., Podsiadlowski, Ph., \& Spruit, H. 2002,
728: ApJ, 574, 364
729:
730: \bibitem[Podsiadlowski et al.(2004)]{pod04}
731: Podsiadlowski, Ph., Langer, N., Poelarends, A. J.T., Rappaport,
732: S., Heger, A., \&Pfahl, E. D. 2004, ApJ, 612, 1044
733:
734: \bibitem[Pringle \& Rees(1972)]{pr72}
735: Pringle, J. E., \& Rees, M. J. 1972, A\&A, 21, 1
736:
737: %\bibitem[Pfahl, Rappaport, \& Podsiadlowski(2002)]{pfa02}
738: %Pfahl, E. D., Rappaport, S., \& Podsiadlowski, Ph. 2002, ApJ, 571,
739: %L37
740:
741: \bibitem[Raguzova \& Popov(2005)]{rag05}
742: Raguzova, N. V., \& Popov, S. B. 2005, astro-ph/0505275
743:
744: \bibitem[Ribo et al.(2005)]{ribo05}
745: Ribo, M., Negueruela, I., Blay, P., Torrej\'on, J. M., \& Reig, P.
746: 2006, 449, 687
747:
748: \bibitem[Ruffert(1999)]{ruf99}
749: Ruffert, M. 1999, A\&A, 346, 861
750:
751: %\bibitem[Sazonov et al.(2005)]{saz05}
752: %Sazonov, S., Revnivtsev, M., Gilfanov, M., Churazov, E., \&
753: %Sunyaev, R. 2005, A\&A, submitted (astro-ph/0510049)
754:
755: \bibitem[Stella et al.(1986)]{ste86}
756: Stellar, L., White, N. E., \& Rosner, R. 1986, ApJ, 308, 669
757:
758: \bibitem[Tauris \& Dewi(2001)]{tau01}
759: Tauris, T. M., \& Dewi, J. D. M. 2001, A\&A, 369,170
760:
761: \bibitem[Tauris \& van den Heuvel(2005)]{tau05}
762: Tauris, T. M., \& van den Heuvel, E. P. J. 2005, in Compact
763: stellar X-ray sources (astro-ph/0303456)
764:
765: \bibitem[van den Heuvel \& Rappaport(1987)]{van87}
766: van den Heuvel, E.P.J., \& Rappaport, S. 1987, in Proc. IAU
767: Symposium 98, Physics of Be Stars, Eds. A. Slettebak and T.P.
768: Snow, p.291
769:
770: %\bibitem[Wang, Gotthelf, \& LangWang(2002)]{wan02}
771: %Wang, Q. D., Gotthelf, E. V., \& Lang, C. C. 2002, Nat, 415, 148
772:
773: \bibitem[Wang \& Robertson(1985)]{wan85}
774: Wang, Y.-M., \& Robertson, J. A. 1985, A\&A, 151, 361
775:
776: \bibitem[Wang \& Robnik(1982)]{wan82}
777: Wang, Y.-M., \& Robnik, M. 1982, A\&A, 107, 222
778:
779: \bibitem[Wang \& Welter(1981)]{wan81}
780: Wang, Y.-M., \& Welter, G. L. 1981, A\&A, 102, 97
781:
782: \bibitem[Waters \& van Kerkwijk(1989)]{wat89}
783: Waters, L. B. F. M., \& van Kerkwijk, M. H. 1989, A\&A, 223, 196
784:
785: \bibitem[Webbink(1985)]{web85}
786: Webbink R. F. 1985, in Interacting Binary Stars, Eds. J. E.
787: Pringle, R. A. Wade, Cambridge Univ. Press, Cambridge, p. 39
788:
789: \bibitem[Zhang, Li \& Wang(2004)]{zha04}
790: Zhang, F., Li, X. D., \& Wang, Z. R. 2004, ChJA\&A, 4, 320
791:
792:
793: \end{thebibliography}
794:
795: \clearpage
796: \begin{figure}
797: \plottwo{f1a.eps}{f1b.eps}
798: %\includegraphics{fig1.eps}
799: \caption{The initial orbital period distribution of neutron star
800: $+$ massive MS star binaries (left) and the incipient mass
801: distribution of the MS companion stars (right).}
802: \end{figure}
803:
804: %\clearpage
805: %\begin{figure}
806: %\includegraphics{fig2.eps}
807: %\caption{The incipient mass distribution of the MS companion
808: %stars.}
809: %\end{figure}
810:
811: \clearpage
812: \begin{figure}
813: \plottwo{f2a.eps}{f2b.eps}
814: \caption{The distribution of neutron star binaries in the $\ps-\pb$ diagram
815: when $\ps=P_{\rm eq}$ or at the age of $t_{\rm MS}$. The color bars denote
816: the relative numbers in different regions. The left and right panels correspond
817: to $v_8=1$ and 2, respectively.}
818: \label{}
819: \end{figure}
820:
821: \clearpage
822: \begin{figure}
823: \plottwo{f3a.eps}{f3b.eps}
824: \caption{The $\ps-\pb$ distribution of wind-fed HMXBs. The dashed and
825: solid lines represent the lower limits of the spin periods for
826: supergiant and Be systems respectively. Asterisks and diamonds
827: mark the Be and supergiant wind-fed HMXBs, and
828: crosses for Roche-lobe overflow systems, respectively. The left and right
829: panels correspond to $v_8=1$ and 2, respectively.
830: }
831: \label{f}
832: \end{figure}
833:
834: \clearpage
835: \begin{figure}
836: \plottwo{f4a}{f4b}
837: \clearpage
838: \plottwo{f4c}{f4d}
839: \caption{The luminosity distribution of wind-fed HMXBs based on
840: population synthesis calculations. {\bf Top:} We take $v_8=1$, $\sigma=190$ kms$^{-1}$
841: ($\alpha=0.1$ and $1$ in the left and right panels
842: respectively).
843: {\bf Bottom:} We take $\alpha=2$, $v_8=2$ ($\sigma=190$ kms$^{-1}$ and $265$ kms$^{-1}$ in the left
844: and right panels respectively).
845: }
846: \label{}
847: \end{figure}
848:
849: \clearpage
850: \begin{figure}
851: \plottwo{f5a.eps}{f5b.eps}
852: \caption{The relative distribution of the inclination angles of HMXBs. The left and right
853: panels correspond to $v_8=1$ and 2, respectively.
854: }
855: \label{}
856: \end{figure}
857:
858: %\clearpage
859: %\begin{figure}
860: % \plottwo{fig61.eps}{fig62.eps}
861: % \caption{The luminosity and lifetime distribution of HMXBs with MS donors. The left and right
862: % panels correspond to $v_8=1$ and 2, respectively.
863: % }
864: % \label{}
865: %\end{figure}
866:
867: \end{document}
868: