1: %\documentclass{aastex}

2: \documentclass[12pt,preprint]{aastex}

3: %\documentclass[manuscript]{aastex}

4: %\usepackage{graphicx,apjfonts,emulateapj5}

5: 

6: \newcommand{\rd}{{\rm d}}

7: \newcommand{\bF}{{\bf F}}

8: \newcommand{\bV}{{\bf V}}

9: \newcommand{\bJ}{{\bf J}}

10: \newcommand{\bW}{{\bf \Omega}}

11: \newcommand{\br}{{\bf r}}

12: \newcommand{\be}{{\bf e}}

13: \newcommand{\bx}{{\bf x}}

14: \newcommand{\by}{{\bf y}}

15: \newcommand{\bz}{{\bf z}}

16: \newcommand{\he}{{\hat{\bf e}}}

17: \newcommand{\hx}{{\hat{\bf x}}}

18: \newcommand{\hy}{{\hat{\bf y}}}

19: \newcommand{\hz}{{\hat{\bf z}}}

20: \newcommand{\bT}{{\bf T}}

21: \newcommand{\kmps}{\rm km\,s^{-1}}

22: 

23: \newcommand{\ba}{\begin{eqnarray}}

24: \newcommand{\ea}{\end{eqnarray}}

25: 

26: \newcommand{\lp}{\left(}

27: \newcommand{\rp}{\right)}

28: \newcommand{\lb}{\left[}

29: \newcommand{\rb}{\right]}

30: \newcommand{\ls}{\left<}

31: \newcommand{\rs}{\right>}

32: 

33: 

34: \begin{document}

35: \title{Spin-Kick Correlation in Neutron Stars: Alignment Conditions and Implications}

36: \author{Chen Wang\altaffilmark{1,2}, 

37:   Dong Lai\altaffilmark{2,1}, 

38:   J. L. Han\altaffilmark{1}}

39: \altaffiltext{1}{National Astronomical Observatories, Chinese Academy of

40:   Sciences, Jia 20 Datun Road, Chaoyang District, Beijing, 100012, China;

41:   wangchen@bao.ac.cn, hjl@bao.ac.cn}

42: \altaffiltext{2}{Department of Astronomy, Cornell University, 

43: Ithaca, NY 14853; dong@astro.cornell.edu}

44: 

45: 

46: 

47: \begin{abstract}

48: Recent observations of pulsar wind nebulae and radio polarization

49: profiles revealed a tendency of the alignment between the spin and

50: velocity directions in neutron stars. We study the condition for

51: spin-kick alignment using a toy model, in which the kick consists of

52: many off-centered, randomly-oriented thrusts.  Both analytical

53: considerations and numerical simulations indicate that spin-kick

54: alignment cannot be easily achieved if the proto-neutron star does not

55: possess some initial angular momentum, contrary to some previous

56: claims.  To obtain the observed spin-kick misalignment angle

57: distribution, the initial spin period of the neutron star must be

58: smaller than the kick timescale. Typically, an initial period of a

59: hundred milliseconds or less is required.

60: 

61: \end{abstract}

62: 

63: \keywords{neutron star --- pulsar kick}

64: 

65: 

66: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

67: 

68: \section{Introduction}

69: 

70: It is well known that pulsars have much larger space velocities than

71: their progenitors, implying a kick at neutron star (NS) birth (e.g.,

72: Lorimer et al.~1997; Arzoumanian et al.~2002; Chatterjee et al.~2005;

73: Hobbs et al.~2005; Winkler \& Petre~2006). The physical mechanism for

74: the kick, however, remains unclear (e.g., Lai 2004; Janka et

75: al.~2005).  One of the reasons that it has been difficult to constrain

76: various kick mechanisms is the lack of correlations between kick

77: velocity and the other properties of NSs. This situation has been

78: changed due to the recent high-resolution Chandra X-ray observations

79: of pulsar wind nebulae (e.g., Pavlov et al. 2000; Helfand et al. 2001;

80: Ng \& Romani 2004), which provided the evidence for spin-kick

81: alignment for several young pulsars (e.g, Lai et al.~2001;

82: Romani~2004; Wang et al.~2006)\footnote{A recent re-analysis of the

83: proper motion of the Crab pulsar (Ng \& Romani~2006) indicates that

84: the spin-kick misalignment angle is $26\arcdeg\pm3\arcdeg$}.

85: 

86: 

87: Another well-known method to determine the spin axis of pulsars is by

88: the linear polarization profile of radio emission.  If the

89: polarization profile could be described by rotating vector model

90: (RVM), one can constrain the projected spin axis by the polarization

91: angle at the center of the pulse. Previous attempts using this method

92: have yielded ambiguous results (e.g., Morris et al.~1979; Anderson \&

93: Lyne 1983; Deshpande et al.~1999) mainly because for many pulsars the

94: polarization profiles are not well described by the RVM (Weisberg et

95: al.~1999). With Parkes surveys (e.g., Manchester et al.~1996, 2001),

96: many well-calibrated polarization profiles became available, and more

97: pulsar rotation measures to eliminate the Faraday rotation effect

98: (e.g. Han et al.~2006). Moreover, proper motions for more than 200

99: pulsars have been determined (Hobbs et al.~2005).  By selecting

100: pulsars with well calibrated polarization and proper motion

101: measurements, Wang et al.~(2006) have obtained spin-kick misalignment

102: angle for 24 pulsars, and the data revealed a strong tendency of

103: spin-kick alignment. Johnston et al.~(2005) independently obtained

104: similar results for 25 pulsars based on different sample.

105: 

106: 

107: On the other hand, one can constrain NS kicks using the orbital

108: properties of NS binary systems (e.g., Dewey \& Cordes~1987; Fryer \&

109: Kalogera~1997; Willem et al.~2004; Thorsett et al.~2005).  In Wang et

110: al.~(2006), we obtained constraints on the kick magnitudes and

111: directions for various NS binaries, including double NS systems,

112: binaries with massive main-sequence companions, and binaries with

113: massive white-dwarf companions. We found that the kick velocity is

114: misaligned with the NS spin axis in a number of systems, and the NS

115: spin period (when available) in these systems is generally longer than

116: several hundreds milliseconds.

117: 

118: 

119: What is the implications of the apparent spin-kick alignment for many

120: pulsars?  One possibility which is widely discussed (e.g., Johnston et

121: al.~2005) comes from Spruit \& Phinney (1998). They suggested that the

122: initial spin of NS may originate from off-centered kicks even when the

123: proto-NS has no angular momentum. They further suggested that if one

124: imagine that the kick is composed of many random thrusts, then with

125: multiple thrusts, alignment may be easily achieved.

126: 

127: In this paper, we systematically study the condition of spin-kick

128: alignment using a toy model similar to that of Spruit \& Phinney

129: (1998).  We consider both the cases of zero and finite initial

130: proto-NS spin.  In \S 2 we introduce our toy model for kicks and

131: describe our simulation procedure. In \S3 we derive approximate but

132: analytic conditions for spin-kick alignment.  In \S 4, we present our

133: simulation results (especially the distribution of spin-kick

134: misalignment angle) under different initial conditions.  We find that,

135: consistently with our analytical estimate, without initial spin of the

136: proto-NS, it is difficult to achieve spin-kick alignment.  This is

137: contrary to some previous claims (e.g., Spruit \& Phinney~1998;

138: Johnston et al. 2005).  However, with sufficiently short initial spin

139: period (less than the timescale for each kick thrust), spin-kick

140: alignment can be achieved. We discuss the implications of our results

141: in \S5.

142: 

143: 

144: %trusts, alignment can be achieved. The purpose of this paper is to

145: %explore this issue systematically, and to understand under what

146: %conditions will alignment be achieved

147: 

148: 

149: 

150: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

151: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

152: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

153: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

154: 

155: \section{A Toy Model for Kicks}

156: 

157: The basic equations governing the evolution of the center-of-mass

158: velocity $\bV$ and angular velocity $\bW$ of a proto-NS (mass $M$,

159: radius $R_{NS}$) are

160: \begin{eqnarray}

161: M\frac{\rd \bV}{{\rd} t} &=& \bF, \label{eq:Newton2}\\

162: \frac{\rd \bJ}{{\rd} t}  &=& \br \times \bF. \label{eq:am}

163: \end{eqnarray}

164: Here $\bF$ is the kick force, and $\br$ specifies the location where

165: the force is applied, $\bJ = I \bW$ is the angular momentum, with

166: $I=kMR_{\rm NS}^2$ the moment of inertia.  We adopt $M=1.4 M_\sun$,

167: $R_{\rm NS}=10\,{\rm km}$ and $k=0.4$.

168: 

169: We model the kick force on the NS as consisting of $n$ thrusts,

170: $\bF_i$, $i=1, 2, ...., n$, each has a duration $\tau_i$.  During each

171: thrust, we construct a ``temporary'' body frame (xyz) corotating with

172: the star so that $\bF_i$ is constant in this frame (see Fig.~1). We

173: specify the kick force $\bF_i$ by the magnitude $F_i$ ($=F$, the same

174: for all thrusts) and two angles $\alpha_i, \beta_i$.  We choose

175: ($\alpha_i$, $\beta_i$) randomly distributed in the range of of

176: $0\arcdeg<\alpha_i<30\arcdeg$ and $0\arcdeg<\beta_i<360\arcdeg$.  Each

177: thrust acts at the position $\br_i$, which is specified by the

178: spherical coordinates $(r_i, \theta_i, \phi_i)$ in the nonrotating

179: frame (XYZ) at the beginning of the thrust. Note that choosing a

180: different range of $\alpha_i$ is equivalent to choosing different

181: $r_i$, as long as $r_i\sin\alpha_i$ remains the same. During the course

182: of each thrust (duration $\tau_i$), the $\br_i$ changes in the XYZ

183: frame as the body rotates.  In our simulation, we assume $\tau_i=\tau$

184: is the same for all thrusts.  Thus the total kick time is $T_{\rm

185: kick} = \sum_i\tau_i = n\tau$. We set the total momentum ${\cal P}

186: \equiv FT_{\rm kick}$ at the fixed range ${\cal P}=M{\cal

187: V}=M(500-2000)\,{\rm km\,s^{-1}}$ in all simulations. We fix

188: $r_i=30\,{\rm km}$ for all thrusts. For $(\theta_i,

189: \phi_i)$, we consider two possibilities: (i) The kick position is

190: randomly distributed on a sphere in the XYZ frame; (ii) The kick

191: position is randomly distributed on a sphere in the body frame of the

192: NS. Our simulation results reported in \S 4 refer to the first

193: case. We have found that the results of the second case are similar.

194: 

195: \begin{figure}

196: \begin{center}

197: \includegraphics[angle=-90, scale=0.6]{f1.ps}

198: \caption

199:      {Geometric model used in our simulations. XYZ is a nonrotating

200:       ``fixed'' frame centered at the neutron star. A thrust force

201:       $\bF_i$ is applied at the position $\br_i$ (specified by the

202:       spherical coordinates $r_i$, $\theta_i$, $\phi_i$. The body

203:       frame xyz is constructed with the $z$-axis along $\br_i$, and

204:       the $x$-axis in the meridional direction.  The direction of

205:       $\bF_i$ is specified by the two polar angles

206:       $\alpha_i,\beta_i$.}

207: \label{fig1}

208: \end{center}

209: \end{figure}

210: 

211: For each thrust, $\br_i$ and $\bF_i$ are fixed in the the

212: ``temporary'' body frame, i.e., $\br_i=r_i\hz$, and

213: $\bF_i=F_i(\sin\alpha_i\cos\beta_i\hx+\sin\alpha_i\sin\beta_i\hy

214: +\cos\alpha_i\hz)$. To find the components of the force $\bF_i$ and

215: torque $\br_i\times\bF_i$ in the inertial frame (XYZ), we need to

216: solve for the time evolution of the body axes $\he$ ($=\hx$, $\hy$ or

217: $\hz$).  The (XYZ) frame and (xyz) frame are related by the rotation

218: matrix $\bT$:

219: \begin{eqnarray}

220:   \left( \begin{array}{c} X \\ Y \\ Z \end{array} \right) = 

221:   \bT \left( \begin{array}{c} x \\ y \\z \end{array} \right)

222:       = \left( \begin{array}{ccc} 

223:         T_{11} & T_{12} & T_{13} \\

224:         T_{21} & T_{22} & T_{23} \\

225:         T_{31} & T_{32} & T_{33}          

226:         \end{array} \right)

227:         \left( \begin{array}{c} x \\ y \\z \end{array} \right).

228: \end{eqnarray}

229: So we have the expressions of the body axes in the (XYZ) frame:

230: \begin{eqnarray}

231: \hx = \left( \begin{array}{c} T_{11} \\ T_{21} \\ T_{31} \end{array} \right), \quad

232: \hy = \left( \begin{array}{c} T_{12} \\ T_{22} \\ T_{32} \end{array} \right), \quad

233: \hz = \left( \begin{array}{c} T_{13} \\ T_{23} \\ T_{33} \end{array} \right). \label{eq:eexp}

234: \end{eqnarray}

235: At the beginning of each thrust

236: \begin{eqnarray}

237:   \bT = \left( \begin{array}{ccc} 

238:         \cos\theta_i\cos\phi_i & -\sin\phi_i & \sin\theta_i\cos\phi_i \\

239:         \cos\theta_i\sin\phi_i & \cos\phi_i& \sin\theta_i \sin\phi_i \\

240:         -\sin\theta_i      & 0         & \cos\theta_i         

241:         \end{array} \right).

242: \end{eqnarray}

243: The body axis $\he$ evolves according to 

244: \begin{eqnarray}

245: \frac{\rd \he}{\rd t}=\bW\times\he. \label{eq:eevo}

246: \end{eqnarray}

247: Substituting Eq.~(\ref{eq:eexp}) in Eq.~(\ref{eq:eevo}), we obtain the

248: evolution of each component of $\bT$, e.g., $\rd T_{11}/\rd t=\Omega_2

249: T_{31}-\Omega_3 T_{21}$, et al., where $\Omega_{1,2,3}$ are the three

250: components of $\bW$ in the XYZ frame.

251: 

252: 

253: Using Eqs.~(\ref{eq:Newton2})~--~(\ref{eq:eevo}), we can directly

254: simulate the movement and the rotation of the NS. Consider a star with

255: initial velocity $\bV_0=\bV_{\rm init}$ and angular velocity

256: $\bW_0=\bW_{\rm init}$.  Suppose it receives a thrust $\bF_1$ at a

257: random position $\br_1$ with duration $\tau_1$.  We use the 4th order

258: Runge-Kutta method to integrate equations~(\ref{eq:Newton2}),

259: (\ref{eq:am}) and (\ref{eq:eevo}) to obtain $\bV_1$ and $\bW_1$, the

260: velocity and rotation rate after the first thrust.  If the star

261: receives a new thrust, we just select a new body frame according to

262: the new position of the thrust, considering $\bV_1$ and $\bW_1$ as the

263: initial velocity and angular velocity, and repeat the calculation as

264: in the first thrust and so on.

265: 

266: 

267: 

268: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

269: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

270: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

271: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

272: \section{Analytic Consideration}

273: 

274: Consider a proto-NS without initial spin. The kick consists of $n$

275: thrusts, each with the same duration $\tau$. We assume $n\gg1$ in this

276: section. After the first thrust, the star receives an angular

277: velocity

278: \begin{equation}

279: \Delta\Omega=\Omega_1=\frac{F\tau r \sin \alpha}{I}

280:  =\frac{F\tau r \sin \alpha}{kMR_{NS}^2}, \label{eq:Omega}

281: \end{equation}

282: and a velocity $\Delta V$ not more than $F\tau/M$. Note that after the

283: first thrust, the star's spin ${\bf\Omega_1}$ is always perpendicular

284: to its velocity ${\bf V}_1$. If the duration of each thrust is larger

285: than the spin period caused by the first thrust, i.e.

286: \begin{equation}

287: \tau \gtrsim \frac{2\pi}{\Delta\Omega}, \label{eq:aligncondition}

288: \end{equation}

289: then the second thrust will be rotationally averaged such that the net

290: thrust will be along $\hat {\bW_1}$ (the unit vector along

291: $\bW_1$). similar argument applies to additional thrusts. The final

292: characteristic velocity and angular velocity are given by

293: \begin{equation}

294: \ls V_f^2 \rs=\ls\lp \sum_i \Delta \bV_i \cdot {\hat \bW_1}\rp^2\rs

295:   \sim \frac{1}{3} n \Delta V ^2,

296: \end{equation}

297: \begin{equation}

298: \ls \Omega_f^2 \rs=\ls\lp \sum_i \Delta \bW_i \cdot {\hat \bW_1}\rp^2\rs

299:   \sim \frac{1}{3} n \Delta \Omega ^2.

300: \end{equation}

301: The typical final velocity and angular velocity are aligned, with

302: \begin{equation}

303: \bV_f \sim \sqrt{\frac{n}{3}}\Delta V {\hat \bW_1}, \quad

304: \bW_f \sim \sqrt{\frac{n}{3}}\Delta \Omega {\hat \bW_1}.

305: \end{equation}

306: 

307: Equation~(\ref{eq:aligncondition}) is a sufficient condition for

308: spin-kick alignment but not a necessary one. It's convenient to define

309: a critical ratio

310: \begin{equation}

311: n_c\equiv \lp \frac{2\pi}{\Delta\Omega\tau}\rp^2. \label{eq:criticalratio}

312: \end{equation}

313: For $\tau \lesssim 2\pi/\Delta\Omega$ or $n_c\gtrsim1$, spin-kick

314: alignment may or may not be achieved. If $n_c\gtrsim n$ or $\tau

315: \lesssim 2\pi/(\sqrt{n}\Delta\Omega)$, different thrusts add up in 

316: a random walk fashion. The final spin and kick are of order

317: \begin{equation}

318: V_f \sim \sqrt{n}\Delta V, \quad

319: \Omega_f \sim \sqrt{n}\Delta \Omega,

320: \end{equation}

321: with random angle between the spin and kick.

322: 

323: 

324: For $1\ll n_c \lesssim n$, or $2\pi/(\sqrt{n}\Delta\Omega) \lesssim

325: \tau \ll 2\pi/\Delta\Omega$, the situation is more complicated.

326: For the first $n_c$ thrusts ($i=1$, 2, ..., $n_c$), the thrust

327: duration $\tau$ satisfies $\tau \lesssim 2\pi/(\sqrt{i}\Delta\Omega)$,

328: thus the characteristic velocity and rotation rate are

329: \begin{equation}

330: V_i \sim \sqrt{i}\Delta V, \quad

331: \Omega_i \sim \sqrt{i}\Delta \Omega, \qquad i=1, 2, \cdots, n_c

332: \end{equation}

333: with random directions between $\bV_i$ and $\bW_i$.  For the remaining

334: thrusts ($i=n_c+1$, $n_c+2$, ..., $n$), $\tau \gtrsim 2\pi/\Omega_i$,

335: so that rotationally averaging is effective.

336: The final velocity and angular velocity are

337: \begin{eqnarray}

338: \bV_f &\sim& \bV_{n_c}+\sqrt{\frac{n-n_c}{3}}\Delta V {\hat \bW_{n_c}},\\

339: \bW_f &\sim& \bW_{n_c}+\sqrt{\frac{n-n_c}{3}}\Delta \Omega {\hat \bW_{n_c}}  

340:         \sim \lp \sqrt{n_c}+\sqrt{\frac{n-n_c}{3}}\rp \Delta\Omega {\hat \bW_{n_c}}.

341: \end{eqnarray}

342: Here $\bV_{n_c}\sim\sqrt{n_c}\Delta V$ has random direction compared

343: to $ {\hat \bW_{n_c}}$. So spin and kick will be aligned when

344: $\sqrt{(n-n_c)/3} \gg \sqrt{n_c}$, which means

345: \begin{equation}

346: n\gg 4n_c.

347: \end{equation}

348: Otherwise, spin and kick will be misaligned.

349: 

350: To summarize, spin-kick alignment/misalignment depends on the critical

351: ratio $n_c$ (see Eq.~\ref{eq:criticalratio}). Let $F={\cal P}/T_{\rm

352: kick}=M{\cal V}/T_{\rm kick}$, $r=f_\Omega R_{NS}$, we find

353: \begin{equation}

354: n_c = \lp \frac{2\pi k R_{\rm NS} n}{f_{\Omega}\tau {\cal V}\sin\alpha}\rp^2

355:     \simeq \lp \frac{n^2}{40 f_{\Omega} {\cal V}_{1k}T_1\sin\alpha}\rp^2,

356: \end{equation}

357: where we have used $T_{\rm kick}=n\tau$, $T_1=T_{\rm kick}/(1~{\rm s})$,

358: and ${\cal V}_{1k}={\cal V}/(10^3~{\rm km~s}^{-1})$. 

359: For 

360: \begin{equation}

361: n_c\lesssim {\rm max}(n/4, 1), \label{eq:aligncond}

362: \end{equation}

363: spin and kick will be aligned, 

364: while for $n_c\gtrsim {\rm max}(n/4, 1)$, spin and kick will be misaligned.   

365: 

366: 

367: If the NS has initial spin $\Omega_{\rm init}$, the sufficient

368: alignment condition, equation~(\ref{eq:aligncondition}), should be modified

369: to

370: \begin{equation}

371: \tau \gtrsim \frac{2\pi}{{\rm max}(\Delta\Omega,\,\,\Omega_{\rm init})}.

372: \label{eq:alignwithinit}

373: \end{equation}

374: 

375: 

376: 

377: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

378: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

379: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

380: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

381: \section{Simulation Results}

382: 

383: In our model, the key parameters are $n$ and $\tau$ or $T_{\rm kick}$,

384: as well as the initial spin period $P_{\rm init}$.  Depending on the

385: kick mechanisms, the total kick duration $T_{\rm kick}$ ranges from

386: $0.1\,s$ to a few seconds (e.g., Lai et al. 2001; Socrates et

387: al. 2005; Scheck et al. 2006; Burrows et al. 2006a, b). Note that we

388: choose ${\cal P}=FT_{\rm kick}$ in the range of $M(500-2000){\rm

389: km\,s^{-1}}$ and other parameters such that the final distributions of

390: kick velocity and spin period of NSs qualitatively agree with

391: observations (see Hobbs~et al.~2005).

392: 

393: In Fig.~\ref{fig:N}~--~\ref{fig:tau}, we present simulations of 20000

394: pulsars without initial spin ($\Omega_{\rm init}=0$).  The final spins

395: of the NSs are all due to the off-centered thrusts.  In

396: Fig.~\ref{fig:N}, we fix the total kick duration to $T_{\rm

397: kick}=1$\,s, while changing the number of thrusts: $n=5$,~10,~20.  For

398: these cases, $n_c\sim \lb n/(7\sqrt{V_{1k}})\rb^4$ (see

399: Eq.~\ref{eq:aligncond}). We find that an aligned distribution is

400: produced for $n=5$, but not for $n=10$ or 20.

401: 

402: In Fig~\ref{fig:T}, we consider different values of total kick

403: duration $T_{\rm kick}=0.1$~s, 0.5~s, 1~s, while fixing the number of

404: thrusts to $n=5$. For these cases, $n_c\sim 0.2/(V_{1k}T_1)^2$. So we

405: find that for large $T_{\rm kick}$, an aligned $\gamma$ distribution

406: is produced.

407: 

408: 

409: Figure~\ref{fig:tau} shows the cases with the same thrust duration

410: $\tau=0.2$~s, while the number of thrusts are $n=5,~10,~20$. Here

411: $n_c\sim \lb n/(60V_{1k}\tau_1)\rb^2$ (where $\tau_1 = \tau /1~{\rm

412: s}$). Since either $n_c\lesssim 1$ or $n_c\lesssim n/4$ is

413: satisfied for these cases, the $\gamma$ distributions all show an

414: tendency of alignment.

415: 

416: Note that in the above three figures, the kick velocities are all

417: distributed at a few hundred kilometers per second and the final spin

418: periods are distributed from 10 to hundreds of milliseconds, in

419: agreement with observations. Although an aligned $\gamma$

420: distribution can be produced under certain conditions (see

421: Eq.~\ref{eq:aligncond}) without initial spin, the distributions are

422: significantly broader than what was observed (Johnston et al.~2005;

423: Wang et al.~2006; see Fig.~\ref{fig:obs}).

424: 

425: Figure~\ref{fig:P1} shows the simulations with different initial spin

426: period $P_{\rm init}=500$\,ms, 100\,ms and 50\,ms, all with $T_{\rm

427: kick}=1\,$s and $n=10$. Clearly, for $P_{\rm init}\lesssim\tau$, 

428: rotational averaging is effective, and spin-kick alignment is easily

429: achieved. Figure~\ref{fig:obs} compares our simulation results

430: with the observed spin-kick misalignment angles based on pulsar

431: polarization profiles (see Wang et al.~2006), taking into account of

432: the sky projection effect. We see that for the $P_{\rm

433: init}=50$\,ms simulation depicted in Fig.~\ref{fig:P1}, the simulated 

434: spin-kick distribution agrees with observational data. With 

435: $P_{\rm init}=100$\,ms (other parameters being the same), 

436: the simulated distribution is broader than the data. The key condition

437: for producing alignment is Eq.~(\ref{eq:alignwithinit}).

438: 

439: 

440: Figures~\ref{fig:vpg1} and \ref{fig:vpg2} show the distribution of the

441: misalignment angle $\gamma$ as a function of $V_f$ and $P_f$. We see

442: that when the $\gamma$ distribution is broad (Fig.~\ref{fig:vpg1}),

443: pulsars with different $V_f$'s have similar range of $\gamma$'s.  On

444: the other hand, based on our toy model simulations, for an aligned

445: $\gamma$ distribution (Fig.\ref{fig:vpg2}), high-$V_f$ pulsars have a

446: strong tendency for spin-kick alignment. We have attempted to test such

447: $V_f - \gamma$ correlation in the existing sample for 24 pulsars. The

448: current data does not show such correlation, probably because of the small 

449: sample or large error in various measurements.

450: 

451: 

452: 

453: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

454: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

455: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

456: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

457: \section{Conclusion}

458: 

459: In this paper, we have developed a toy model to study the conditions

460: for pulsar spin-velocity alignment in supernova kicks.  We have

461: focused on the idea (Spruit \& Phinney 1998) that multiple

462: off-centered thrusts to the proto-neutron star may result in spin-kick

463: alignment.  We found that without initial angular momentum, the

464: alignment cannot be easily produced. To obtain the observed spin-kick

465: alignment distribution based on radio pulsar polarization data

466: (Johnston et al.~2005; Wang et al.~2006), the proton-neutron stars

467: should have appreciable rotation rate, with period less than the

468: timescale of each kick thrust. The typical initial period required is

469: $\lesssim 100$~ms.

470: 

471: Currently, the most conservative (and promising) kick mechanisms are

472: ``hydrodynamically driven kicks''. In particular, large-scale

473: convections, instabilities or wave modes developed in the

474: neutrino-heated mantle behind the shock and in the proto-neutron star

475: may naturally lead to asymmetric explosion (e.g., Thompson 2000;

476: Scheck et al.~2004,2006; Blondin \& Mezzacappa~2006; Foglizzo et

477: al.~2005; Burrows et al.~2006a, b; Yamasaki \& Yamada~2006). The supernova

478: simulations cited above do not include initial angular momentum, and

479: the resulting kicks are randomly distributed. It is possible that with

480: even a small rotation, the hydrodynamical instabilities may

481: preferentially develop along the rotation axis. If the kick timescale

482: is long (as indicated by recent simulations), rotational averaging may

483: be effective and a preferentially aligned spin-kick distribution can

484: be produced.

485: 

486: Other kick mechanisms (such as those based on asymmetric neutrino

487: emissions in the proto-neutron star; e.g. Duncan \& Thompson 1992; Lai

488: \& Qian 1998; Arras \& Lai 1999a,b; Socrates et al.~2005) and the

489: ``electromagnetic rocket'' effect (Harrison \& Tademaru 1975; Lai et

490: al.~2001) can easily result in spin-kick alignment, but they require

491: more extreme conditions (such as superstrong magnetic field or very

492: rapid spin) for the proto-neutron stars (see Lai 2004; Wang et

493: al.~2006 and references therein).

494: 

495: \acknowledgments

496: 

497: This work is supported by National Natural Science Foundation of China

498: (10328305, 10473015 and 10521001). D.L. has also been supported in part

499: by NSF grant AST 0307252 and NASA grant NAG 5-12034.  D.L. thanks NAOC

500: (Beijing) for hospitality during the course of the work.

501: 

502: 

503: 

504: 

505: \begin{thebibliography}{}

506: 

507: \bibitem[]{} Anderson, B. \&  Lyne, A. G. 1983, Nature, 303, 597

508:   % On the origin of pulsar velocities

509: 

510: \bibitem[]{} Arras, P. \& Lai, D. 1999a, PhRvD, 60, 3001

511:   % Neutrino-nucleon interactions in magnetized neutron-star matter: The effects of parity violation

512: 

513: \bibitem[]{} Arras, P. \& Lai, D. 1999b, \apj, 519, 745

514:   % Can Parity Violation in Neutrino Transport Lead to Pulsar Kicks?

515: 

516: \bibitem[]{} Arzoumanian, Z., Chernoff, D. F., \& Cordes, J. M. 2002, \apj, 568, 289

517:   % The Velocity distribution of isolated pulsars

518: 

519: \bibitem[]{} Blondin, J. M. \& Mezzacappa, A. 2006, \apj, 642, 401

520:   % The Spherical Accretion Shock Instability in the Linear Regime

521: 

522: \bibitem[]{} Burrows, A. et al. 2006a, \apj, 640, 878

523:   % A New Mechanism for Core-Collapse Supernova Explosions

524: 

525: \bibitem[]{} Burrows, A. et al. 2006b, ApJ in press, astro-ph/0610175

526:   % Features of the Acoustic Mechanism of Core-Collapse Supernova Explosions

527: 

528: \bibitem[]{} Chatterjee, S. et al. \apj, 630, 61

529:   % Getting Its Kicks: A VLBA Parallax for the Hyperfast Pulsar B1508+55

530: 

531: \bibitem[]{} Deshpande, A. A., Ramachandran, R., \& Radhakrishnan, V. 1999, \aap, 351, 195

532:   % The observational evidence pertinent to possible kick mechanisms in neutron stars

533: 

534: \bibitem[]{} Dewey, R. J. \& Cordes, J. M. 1987, \apj, 321, 780

535:   % Monte Carlo simulations of radio pulsars and their progenitors

536: 

537: \bibitem[]{} Duncan, R. C. \& Thompson, C. 1992, \apj, 392, 9

538:   % Formation of very strongly magnetized neutron stars - Implications for gamma-ray bursts

539: 

540: \bibitem[]{} Foglizzo, T., Scheck, L., \& Janka, H. -Th, SF2A-2005: Semaine de l'Astrophysique Francaise, meeting held in Strasbourg, Edited by F. Casoli, T. Contini, J.M. Hameury and L. Pagani, published by EdP-Sciences, Conference Series, 2005, p. 483 (astro-ph/0507636)

541:   % Neutrino-driven convection versus advection in core collapse supernovae

542: 

543: \bibitem[]{} Fryer, C. L. \& Kalogera, V. 1997, \apj, 489, 244

544:   % Double neutron star systems and natal neutron star kicks

545: 

546: \bibitem[]{} Han, J. L. et al. 2006, \apj, 642, 868

547:   % Pulsar Rotation Measures and the Large-Scale Structure of the Galactic Magnetic Field

548: 

549: \bibitem[]{} Harrison, E. R. \& Tademaru, E. 1975, \apj, 201, 447

550:   % Acceleration of pulsars by asymmetric radiation

551: 

552: \bibitem[]{} Helfand, D. J., Gotthelf, E. V., \& Halpern, J. P. 2001, \apj, 556, 380

553:   % Vela Pulsar and Its Synchrotron Nebula

554: 

555: \bibitem[]{} Hobbs, G., Lorimer, D. R., Lyne, A. G., \& Kramer, M. 2005, \mnras, 475

556:   % A statistical study of 233 pulsar proper motions 

557: 

558: \bibitem[]{} Janka, H.-T., Scheck, L., \& Kifonidis, K. et al. 2005, ASPC, 332, 363

559:   % Supernova Asymmetries and Pulsar Kicks -- Views on Controversial Issues

560: 

561: \bibitem[]{} Johnston, S., Hobbs, G., Vigelang, S., Kramer, M., Weisberg, J. M., \& Lyne, A. G., 2005, \mnras, 364, 1397

562:   % Evidence of alignment of the rotation and velocity vectors in pulsars

563: 

564: \bibitem[]{} Lai, D. 2004, in Cosmic Explosions in 3D: Asymmetries in

565: Supernovae and Gamma-ray Bursts. eds. P. Hoflich et al.  (Cambridge

566: Univ. Press), p.276 (astro-ph/0312542)

567: % Neutron star kicks and supernova asymmetry

568: 

569: \bibitem[]{} Lai, D., Chernoff, D. F., \& Cordes, J. M. 2001, \apj, 549, 1111

570:   % Pulsar jets: Implications for neutron star kicks and initial spins

571: 

572: \bibitem[]{} Lorimer, D. R., Bailes, M., \& Harrison, P. A. 1997, \mnras, 289, 592

573:   % Pulsar statistics - IV. Pulsar velocities

574: 

575: \bibitem[]{} Lai, D. \& Qian, Y. Z. 1998, \apj, 505, 844

576:   % Neutrino Transport in Strongly Magnetized Proto-Neutron Stars and the Origin of Pulsar Kicks: The Effect of Asymmetric Magnetic Field Topology

577: 

578: \bibitem[]{} Manchester, R. N. et al. 2001, \mnras, 328, 17

579:   % The Parkes multi-beam pulsar survey - I. Observing and data analysis systems, discovery and timing of 100 pulsars

580: 

581: \bibitem[]{} Manchester, R. N. et al. 1996, \mnras, 279, 1235

582:   % The Parkes southern pulsar survey - I. Observing and data analysis systems and initial results.

583: 

584: \bibitem[]{} Morris, D., et al. 1979, \aap, 73, 46

585:   % Intrinsic position angles of polarization for 40 pulsars

586: 

587: \bibitem[]{} Ng, C.-Y. \& Romani, R. W. 2004, \apj, 601, 479

588:   % Fitting Pulsar Wind Tori

589: 

590: \bibitem[]{} Ng, C.-Y. \& Romani, R. W. 2006, \apj, 644, 445

591:   % Proper Motion of the Crab Pulsar Revisited

592: 

593: \bibitem[]{} Pavlov, G. G., Sanwal, D., \& Garmire, G. P. 2000, \aaps, 196, 3704

594:   % Observations of the Vela Pulsar and its Compact Nebula with the Chandra high Resolution Camera

595: 

596: \bibitem[]{} Romani, R. W. 2004, astro-ph/0404100

597:   % Pulsar Kicks: Spin and Kinematic Constraints

598: 

599: \bibitem[]{} Scheck, L., Kifonidis, K., Janka, H. -Th., \& Mueller, E. 2006, \aap, 457, 963

600:   % Multidimensional Supernova Simulations with Approximative Neutrino Transport I. Neutron Star Kicks and the Anisotropy of Neutrino-Driven Explosions in Two Spatial Dimensions

601: 

602: \bibitem[]{} Scheck, L., Plewa, T., Janka, H.-Th., Kifonidis, K., \& Müller, E. 2004, PhRvL, 92, 1103

603:   % Pulsar Recoil by Large-Scale Anisotropies in Supernova Explosions

604: 

605: \bibitem[]{} Spruit, H. C. \& Phinney, E. S. 1998, \nat, 393, 139

606:   % Birth kicks as the origin of pulsar rotation

607: 

608: \bibitem[]{} Socrates, A., Blaes, O., Hungerford, A., \& Fryer, C. L. 2005, \apj, 632, 531

609:   % The Neutrino Bubble Instability: A Mechanism for Generating Pulsar Kicks

610: 

611: \bibitem[]{} Thorsett, S. E., Dewey, R. J., \& Stairs, I. H. 2005, \apj, 619, 1036

612:   % Studies of the Relativistic binary pulsar PSR B1534+12, II, Origin and Evolution

613: 

614: \bibitem[]{} Thompson, C. 2000, \apj, 534, 915

615:   % Accretional Heating of Asymmetric Supernova Cores

616: 

617: \bibitem[]{} Wang, C., Lai, D., \& Han, J. L. 2006, \apj, 639, 1007

618:   % Neutron Star Kicks in Isolated and Binary Pulsars: Observational Constraints and Implications for Kick Mechanisms

619: 

620: \bibitem[]{} Weisberg, J. M. et al. 1999, \apjs, 121, 171

621:   % Arecibo 1418 MHZ Polarimetry of 98 Pulsars: Full Stokes Profiles and Morphological Classifications

622: 

623: 

624: \bibitem[]{} Willems, B., Kalogera, V., \& Henninger, M. 2004, \apj, 616, 414

625:   % Pulsar kicks and spin tilts in the close double neutron stars PSR J0737-3039, PSR B1534+12, PSR B1913+16

626: 

627: \bibitem[]{} Winkler, P. F. \& Petre, R. 2006, astro-ph/0608205

628:   % Direct Measurement of Neutron-Star Recoil in the Oxygen-Rich Supernova Remnant Puppis A

629: 

630: \bibitem[]{} Yamasaki, T. \& Yamada, S. 2006, \apj, 650, 291

631:   % Standing Accretion Shocks in the Supernova Core: Effects of Convection and Realistic Equations of State  

632: 

633: \end{thebibliography}

634: 

635: \begin{figure}

636: \centering

637: \includegraphics[height=15cm, angle=-90]{f2.ps}

638: \caption{ The distribution of the spin-kick misaligned angle $\gamma$

639: (upper panel), final velocity $V_f$ (lower-left panel) and final spin

640: period $P_f$ (lower-right panel) of NSs in simulations with $T_{\rm

641: kick}=1\,{\rm s}$ and different number of thrusts $n=$5, 10, 20 (from

642: thick lines to thin lines). The initial spin of the NS is $\Omega_{\rm

643: init}=0$.

644: \label{fig:N}

645: }

646: \end{figure}

647: 

648: \begin{figure}

649: \centering

650: \includegraphics[height=15cm, angle=-90]{f3.ps}

651: \caption{ Same as Fig.~\ref{fig:N}, except for different values of the

652: total kick duration $T_{\rm kick}=$0.1s, 0.5s, 1s (from thick lines to

653: thin lines).  The number of thrusts is fixed at $n=5$, and the initial

654: spin is $\Omega_{\rm init}=0$.

655: \label{fig:T}

656: }

657: \end{figure}

658: 

659: \begin{figure}

660: \centering

661: \includegraphics[height=15cm, angle=-90]{f4.ps}

662: \caption{ Same as Fig.~\ref{fig:N}, except that the 

663: duration of each thrust is fixed at $\tau=0.2$\,s, and 

664: the number of thursts are $n=5$, 10, 20

665: (from thick lines to thin lines). The other parameters keep

666: the same as Fig.~\ref{fig:N}.

667: \label{fig:tau}

668: }

669: \end{figure}

670: 

671: \begin{figure}

672: \centering

673: \includegraphics[height=15cm, angle=-90]{f5.ps}

674: \caption{ The spin-kick alignment distributions from observations and

675: simulaions. The points (with error bars) are based on radio pulsar

676: polarization and proper motion data (see Fig.~1 in Wang et al.~2006,

677: note that the spin-kick angle for the Crab pulsar has been changed

678: from $8\arcdeg\pm20\arcdeg$ to $26\arcdeg\pm3\arcdeg$, based on a

679: recent analysis, see Ng \& Romani~2006): each point represents a

680: pulsar with measured angle between the projected spin axis and the

681: proper motion direction. Note that because of the orthorgonal mode

682: phenomena, there is a $90^\circ$ degeneracy for the spin axis inferred

683: from the pulsar polarization profile.  Thus we have folded Fig.~1 of

684: Wang et al. around $45^\circ$ (e.g., a $60^\circ$ data point is

685: identified with $30^\circ$).  The histograms are based on our

686: simulations with $P_{\rm init}=50$~ms, $T_{\rm kick}=1$~s and $n=10$

687: (see Fig.~\ref{fig:P1}): The dashed line gives the actual

688: (unprojected) spin-kick angle distribution, the solid and dot-dashed

689: lines give the projected spin-kick angles in the plane of the sky when

690: the line of sight is inclined with respect to the initial spin axis by

691: $60\arcdeg$ and $30\arcdeg$, respectively. To compare with

692: observational data, the simuation results are also folded around

693: $45^\circ$.

694: \label{fig:obs}

695: }

696: \end{figure}

697: 

698: 

699: \begin{figure}

700: \centering

701: \includegraphics[height=15cm, angle=-90]{f6.ps}

702: \caption{ Same as Fig.~\ref{fig:N}, except for different values of the

703: initial spin period $P_{\rm init}=$50\,ms, 100\,ms, 500\,ms (from

704: thick lines to thin lines).  The other parameters are $T_{\rm

705: kick}=1\,{\rm s}$, $n=10$.

706: \label{fig:P1}

707: }

708: \end{figure}

709: 

710: 

711: \begin{figure}

712: \centering

713: \includegraphics[height=15cm, angle=-90]{f7.ps}

714: \caption{ The distribution of the spin-kick misalignment angle

715: $\gamma$ as a function of the final velocity $V_f$ (left panel) and

716: the period $P_f$ (right panel) for a simulation which produces a broad

717: $\gamma$ distribution. The simulation parameters are $T_{\rm

718: kick}=1\,{\rm s}$, $N=20$, $\Omega_{\rm init}=0$.

719: \label{fig:vpg1}

720: }

721: \end{figure}

722: 

723: \begin{figure}

724: \centering

725: \includegraphics[height=15cm, angle=-90]{f8.ps}

726: \caption{ The distribution of the spin-kick misalignment angle

727: $\gamma$ as a function of final velocity $V_f$ (left panel) and period

728: $P_f$ (right panel) for a simulation which produces an aligned $\gamma$

729: distribution. The simulation parameters are $T_{\rm kick}=1\,{\rm

730: s}$, $N=10$, $P_{\rm init}=50$\,ms.

731: \label{fig:vpg2}

732: }

733: \end{figure}

734: 

735: \end{document}

736: 

737: