1: %%

2: %% Created 2005 10 24

3: %%

4: %% This is a manuscript file creates using AASTeX v5.x LaTeX 2e macros

5: %%

6: 

7: \documentclass[12pt]{emulateapj}

8: %\usepackage{amsmath}

9: %\usepackage{natbib}

10: \usepackage{ulem}

11: \bibliographystyle{apj}

12: 

13: \slugcomment{To be submitted to the Astrophysical Journal}

14: 

15: \begin{document}

16: 

17: \title{Interacting Binaries with Eccentric Orbits.  Secular 

18: Orbital Evolution Due To Conservative Mass Transfer} 

19: 

20: \author{J. F. Sepinsky, B. Willems, V. Kalogera, F. A. Rasio}

21: \affil{Department of Physics and Astronomy, Northwestern University,

22: 2145 Sheridan Road, Evanston, IL 60208}

23: \slugcomment{j-sepinsky, b-willems, vicky@northwestern.edu, and

24: rasio@northwestern.edu}

25: \shortauthors{Sepinsky, Willems, Kalogera, \& Rasio}

26: \shorttitle{Interacting Binaries with Eccentric Orbits}

27: 

28: \begin{abstract}

29: 

30: We investigate the secular evolution of the orbital semi-major axis and

31: eccentricity due to mass transfer in eccentric binaries, assuming

32: conservation of total system mass and orbital angular momentum. Assuming

33: a delta function mass transfer rate centered at periastron, we find

34: rates of secular change of the orbital semi-major axis and eccentricity

35: which are linearly proportional to the magnitude of the mass transfer

36: rate at periastron. The rates can be positive as well as negative, so

37: that the semi-major axis and eccentricity can increase as well as

38: decrease in time. Adopting a delta-function mass-transfer rate of

39: $10^{-9} M_\sun\,{\rm yr}^{-1}$ at periastron yields orbital evolution

40: timescales ranging from a few Myr to a Hubble time or more, depending on

41: the binary mass ratio and orbital eccentricity. Comparison with orbital

42: evolution timescales due to dissipative tides furthermore shows that

43: tides cannot, in all cases, circularize the orbit rapidly enough to

44: justify the often adopted assumption of instantaneous circularization at

45: the onset of mass transfer.  The formalism presented can be incorporated

46: in binary evolution and population synthesis codes to create a

47: self-consistent treatment of mass transfer in eccentric binaries.

48: 

49: \end{abstract}

50: 

51: \keywords{Celestial mechanics, Stars: Binaries: Close, Stars: Mass Loss}

52: 

53: \section{Introduction}

54: \label{sec-intro}

55: Mass transfer between components of close binaries is a common

56: evolutionary phase for many astrophysically interesting binary

57: systems.  Indeed, mass ejection and/or accretion is responsible for

58: many of the most recognizable phenomena associated with close

59: binaries, such as persistent or transient X-ray emission, neutron star

60: spin-up, and orbital contraction or expansion.  Theoretical

61: considerations of these and other associated phenomena in the

62: literature probe these systems quite effectively, yet they often do

63: not consider the effects of any eccentricity associated with the

64: binary orbit.  This can be of particular importance for binaries

65: containing a neutron star or a black hole, where mass loss and natal

66: kicks occurring during compact object formation may induce a

67: significant eccentricity to the binary \citep[e.g.][]{H83, BP95,

68:   K96}. After the formation of the compact object, tides tend to

69: circularize the orbit on a timescale which strongly depends on the

70: ratio of the radius of the compact object's companion to the orbital

71: semi-major axis.  Because of this, orbits are usually assumed to

72: circularize instantaneously when a binary approaches or begins a mass

73: transfer phase. 

74: 

75: Despite our generally well developed understanding of tidal

76: interactions in close binaries, quantitative uncertainties in tidal

77: dissipation mechanisms propagate into the determination of

78: circularization timescales. For example, \citet{MM05} have shown

79: that current theories of tidal circularization cannot explain observed

80: degrees of circularization of solar-type binaries in open clusters. 

81: Circularization of high-mass binaries, on the other hand, is currently

82: thought to be driven predominantly by resonances between dynamic tides

83: and free oscillation modes, but initial conditions play an

84: important role and an extensive computational survey of relevant parts

85: of the initial parameter space has yet to be undertaken \citep{WS99,

86: WS01, WVS03}.

87: 

88: Furthermore, assumptions of instantaneous circularization immediately

89: before or at the onset of mass transfer are in clear contrast with

90: observations of eccentric mass transferring systems. In the most recent

91: catalog of eccentric binaries with known apsidal-motion rates compiled

92: by \citet{PO99}, 26 out of the 128 listed systems are semi-detached or

93: contact binaries. Among these mass-transferring systems, $9$ have

94: measured eccentricities greater than $0.1$. In addition, many

95: high-mass X-ray binaries are known to have considerable orbital

96: eccentricities \citep{Rag05}. While mass transfer in these systems is

97: generally thought to be driven by the stellar wind of a massive O- or

98: B-star, it has been suggested that some of them may also be subjected

99: to atmospheric Roche-lobe overflow at each periastron passage of the

100: massive donor \citep[e.g.][]{P78}. 

101: 

102: \citet{H56}, \citet{K64-2}, and \citet{P64} were the first to study

103: the effects of mass transfer on the orbital elements of eccentric

104: binaries. However, their treatment was restricted to perturbations of

105: the orbital motion caused by the variable component masses.

106: \citet{MW83,MW84} extended these early pioneering studies to include

107: the effects of linear momentum transport from one star to the other,

108: as well as any other possible perturbations caused by the mass

109: transfer stream in the system. However, these authors derived the

110: equations governing the motion of the binary components with respect

111: to a reference frame with origin at the mass center of the binary,

112: which is not an inertial frame. Their equations therefore do not

113: account for the accelerations of the binary mass center caused by the

114: mass transfer (see \S~\ref{sec-appb}). 

115: 

116: More recent work on mass transfer in eccentric binaries has mainly

117: focused on smoothed particle hydrodynamics calculations of the mass

118: transfer stream over the course of a few orbits, without any

119: consideration of the long-term evolution of the binary \citep{L98,

120: Rea05}.

121: 

122: Hence, there is ample observational and theoretical motivation to

123: revisit the study of eccentric mass-transferring binaries.  In this

124: paper, our aim is to derive the equations governing the evolution of

125: the orbital semi-major axis and eccentricity in eccentric

126: mass-transferring binaries, assuming conservation of total system mass

127: and orbital angular momentum. In a subsequent paper, we will

128: incorporate the effects of mass and orbital angular momentum losses

129: from the system.

130: 

131: Our analysis is based on the seminal work of \citet{H69} who was the

132: first to derive the equations of motion of the components of eccentric

133: mass-transferring binaries while properly accounting for the effects

134: of the variable component masses on the stars' mutual gravitational

135: attraction, the transport of linear momentum from one star to the

136: other, the accelerations of the binary mass center due to the

137: redistribution of mass in the system, and the perturbations of the

138: orbital motion caused by the mass-transfer stream.  While the

139: equations of motion derived by \citet{H69} are valid for orbits of

140: arbitrary eccentricity, the author restricted the derivation of the

141: equations governing the evolution of the semi-major axis and

142: eccentricity to orbits with small initial eccentricities.

143: 

144: The paper is organized as follows. In \S\,2 and \S\,3 we present the

145: basic assumptions relevant to the investigation and derive the equations

146: governing the motion of the components of an eccentric mass-transferring

147: binary under the assumption of conservative mass transfer.  The

148: associated equations governing the rates of change of the semi-major

149: axis and the orbital eccentricity are derived in \S\,4, while numerical

150: results for the timescales of orbital evolution due to mass transfer as

151: a function of the initial binary mass ratio and orbital eccentricity are

152: presented in \S\,5. For comparison, timescales of orbital evolution due

153: to dissipative tidal interactions between the binary components are

154: presented in \S\,6.  \S\,7 is devoted to a summary of our main results

155: and a discussion of future work. In the appendices, lastly, we derive

156: an equation for the position of the inner Lagrangian point in eccentric

157: binaries with non-synchronously rotating component stars

158: (Appendix~\ref{sec-appa}), and present an

159: alternative derivation for the equations governing the secular evolution

160: of the orbital semi-major axis and eccentricity assuming instantaneous

161: mass transfer between two point masses (Appendix~\ref{sec-appc}).

162: 

163: \section{Basic Assumptions}

164: 

165: We consider a binary system consisting of two stars in an eccentric

166: orbit with period $P_{\rm orb}$, semi-major axis $a$, and

167: eccentricity $e$. We let the component stars rotate with angular

168: velocities $\vec{\Omega}_1$ and $\vec{\Omega}_2$ parallel to the

169: orbital angular velocity $\vec{\Omega}_{\rm orb}$, and assume the

170: rotation rates to be uniform throughout the stars. We also note that

171: the magnitude of $\vec{\Omega}_{\rm orb}$ varies periodically in time

172: for eccentric binaries, but its direction remains fixed in

173: space. Because of this, the stars cannot be synchronized with the

174: orbital motion at all times.

175: 

176: At some time $t$, one of the stars is assumed to fill its Roche lobe

177: and begins transferring mass to its companion through the inner

178: Lagrangian point $L_1$. We assume this point to lie on the line

179: connecting the mass centers of the stars, even though non-synchronous

180: rotation may cause it to oscillate in the direction perpendicular to

181: the orbital plane with an amplitude proportional to the degree of

182: asynchronism \citep{MW83}. Since the donor's rotation axis is assumed

183: to be parallel to the orbital angular velocity, we can safely assume 

184: that the transferred mass remains confined to the orbital plane.

185: 

186: We furthermore assume that all mass lost from the donor is accreted by

187: its companion, and that any orbital angular momentum transported by

188: the transferred mass is immediately returned to the orbit. The mass

189: transfer thus conserves both the total system mass and the orbital

190: angular momentum.

191: 

192: We also neglect any perturbations to the orbital motion other than

193: those due to mass transfer. At the lowest order of approximation,

194: these additional perturbations (e.g., due to tides, magnetic breaking,

195: or gravitational radiation) are decoupled from those due to mass

196: transfer, and can thus simply be added to obtain the total rates of

197: secular change of the orbital elements.

198: 

199: \section{Equations of Motion}

200: \label{sec-eom}

201: 

202: \subsection{Absolute Motion of the Binary Components}

203: 

204: \label{sec-pert}

205: 

206: Following \citet{H69}, we derive the equations of motion of the

207: components of an eccentric mass-transferring binary with respect to a

208: right handed inertial frame of reference $OXYZ$ which has an arbitrary

209: position and orientation in space (see Fig.~\ref{fig-coords}). We let

210: $M_i$ be the mass of star $i$ at some time $t$ at which mass is

211: transferred from the donor to the accretor, and $M_i + \delta M_i$ the

212: mass of the same star at some time $t+\delta t$, where $\delta t > 0$ is

213: a small time interval. With these notations, $\delta M_i < 0$

214: corresponds to mass loss, and $\delta M_i > 0$ to mass accretion. We

215: furthermore denote the point on the stellar surface at which mass is

216: lost or accreted by $A_i$. For the donor star, $A_i$ corresponds to the

217: inner Lagrangian point $L_1$, while for the accretor, $A_i$ can be any

218: point on the star's equator. For the remainder of the paper, we let

219: $i=1$ correspond to the donor and $i=2$ to the accretor.

220: 

221: \begin{figure*}

222: \epsscale{.9}

223: \plotone{f1.eps}

224: \caption{

225: Schematic representation of the reference frames and position vectors

226: adopted in the derivation of the equations of motion of the 

227: components of an eccentric mass-transferring binary.  The $Z$- and

228: $Z_1$-axes of the $OXYZ$ and $O_1X_1Y_1Z_1$ frames are perpendicular

229: to the plane of the page and are therefore not drawn. The geometry of

230: the system at time $t$ is shown in black, while the geometry at time

231: $t + \delta t$ is shown in gray.  The solid line connecting the origin

232: of the $O_1X_1Y_1Z_1$ frame at time $t$ to the origin of the frame at

233: time $t+\delta t$ represents the path the donor would have taken had

234: no mass transfer occurred.  The dashed line, on the other hand,

235: represents the perturbed motion of the donor's mass center (small

236: solid circles) due to the mass transfer. A similar perturbation is

237: imparted to the motion of the accretor. For clarity, this perturbed

238: motion and the $O_2X_2Y_2Z_2$ reference frame connected to the

239: accretor (see text) are omitted from the figure. The dotted line 

240: illustrates a possible path of a mass element transferred from the

241: donor to the accretor.  The element leaves the donor at the inner

242: Lagrangian point $L_1$ (asterisk) and accretes onto the companion at

243: the point $A_2$ (open star).

244: }

245: \label{fig-coords}

246: \end{figure*}

247: 

248: Because of the mass loss/gain, the center of mass of star $i$ at time

249: $t + \delta t$ is shifted from where it would have been had no mass

250: transfer taken place.  To describe this perturbation, we introduce an

251: additional right-handed coordinate frame $O_i X_i Y_i Z_i$ with a

252: spatial velocity such that its origin follows the unperturbed orbit of

253: star $i$, i.e., the origin of $O_iX_iY_iZ_i$ follows the path the

254: center of mass of star $i$ would have taken had no mass transfer

255: occurred.  Thus, at time $t$, the center of mass of star $i$ lies at the

256: origin of $O_i$, while at time $t+\delta t$ it has a non-zero position

257: vector with respect to $O_i$.  We furthermore let the $Z_i$-axis of the

258: $O_i X_i Y_i Z_i$ frame point in the direction of the orbital angular

259: momentum vector, and let the frame rotate synchronously with the 

260: unperturbed orbital angular velocity of the binary in the absence of

261: mass transfer. The direction and orientation of the $X_i$-axes are

262: then chosen such that at time $t$ the $X_i$-axis points along the

263: direction from the mass center of star $i$ to the mass center of its

264: companion.

265: 

266: To describe the shift in the mass center of star $i$ due to the mass

267: loss/gain, we denote the position vector of $O_i$ at times $t$ and $t

268: + \delta t$ with respect to the inertial frame by $\vec{R}_{_i}$ and

269: $\vec{R}_i^\prime$, respectively.  The position vector of the center

270: of mass of star $i$ at times $t$ and $t + \delta t$ is then given by

271: $\vec{R}_i$ and $\vec{R}_i^\prime + \delta \vec{r}_i^{\,\prime}$, where 

272: $\delta \vec{r}_i^{\,\prime}$ is the position vector of the

273: center of mass of star $i$ at time $t+\delta t$ with respect to

274: $O_i$.  Moreover, we denote by $\vec{r}_{A_i}$ 

275: and $\vec{r}_{A_i}^{\,\prime}$

276: the vectors from $O_i$ to the position where the point $A_i$ on the

277: stellar surface would be at times $t$ and $t + \delta t$, respectively,

278: had no mass been lost/accreted. The various position vectors at 

279: time $t + \delta t$ are related by

280: \begin{equation}

281: \left( M_i+\delta M_i \right) \left( \vec{R}_i^\prime 

282:   + \delta \vec{r}_i^{\,\prime} \right) = M_i\, \vec{R}_i^\prime 

283:   + \delta M_i\,\left( \vec{R}_i^\prime + \vec{r}_{A_i}^{\,\prime} 

284:   \right),

285: \label{eq-cm}

286: \end{equation}

287: which, at the lowest order of approximation in $\delta M_i$ and

288: $\delta \vec{r}_i^{\,\prime}$, yields

289: \begin{equation}

290: \delta \vec{r}_i^{\,\prime} = \frac{\delta M_i}{M_i}\, 

291:   \vec{r}_{A_i}^{\,\prime}.

292: \label{eq-DelR}

293: \end{equation}

294: As expected, the displacement of the center of mass

295: of star $i$ due to the mass loss/gain is directed along the line

296: connecting the center of mass of the star and the mass

297: ejection/accretion point.  

298: 

299: We furthermore denote with $\vec{\rho}_{A_i}^{\, \prime}$ the vector

300: from the center of mass of star $i$ at time $t+\delta t$ to the

301: position where the point $A_i$ would be at time $t+\delta t$, had no

302: mass been transferred between the binary components, and with 

303: $\delta\vec{\rho}^{\,\prime}_{A_i}$ the perturbation of this vector 

304: caused by the mass transfer.  It then follows

305: that $\vec{\rho}_{A_i}^{\, \prime} = \vec{r}_{A_i}^{\,\prime} - \delta

306: \vec{r}_i^{\,\prime}$ and thus, by definition, $\delta

307: \vec{\rho}_{A_i}^{\, \prime} = \delta \vec{r}_i^{\,\prime}$.  At the

308: lowest order of approximation in $\delta M_i$ and $\delta

309: \vec{r}_i^{\,\prime}$, equation~(\ref{eq-DelR}) therefore also yields

310: \begin{equation}

311: \delta \vec{\rho}_{A_i}^{\, \prime} = \frac{\delta

312:   M_i}{M_i}\, \vec{r}_{A_i}^{\, \prime}.

313: \label{eq-Delrho}

314: \end{equation}

315: The definitions of and the relations between these various position

316: vectors are illustrated schematically in Fig.~\ref{fig-coords}.

317: 

318: Next, we denote the absolute velocity of the center of mass of star

319: $i$ with respect to the inertial frame of reference at times $t$ and

320: $t + \delta t$ by $\vec{V}_i$ and $\vec{V}_i^\prime$, respectively,

321: and the absolute velocity of the ejected/accreted mass element by 

322: $\vec{W}_{\delta M_i}$.  The linear momentum $\vec{Q}_1$ of star~$1$ at 

323: time $t$ is then given by

324: \begin{equation}

325: \vec{Q}_1 = M_1\,\vec{V}_1,

326: \label{eq-Q1-1}

327: \end{equation}

328: and the total linear momentum $\vec{Q}_1^\prime$ of star~1 and the

329: ejected mass element at time $t+ \delta t$ by

330: \begin{equation}

331: \vec{Q}_1^\prime = (M_1 + \delta M_1) \vec{V}_1^\prime

332:   - \delta M_1\,\vec{W}_{\delta M_1}.

333: \label{eq-Q1-2}

334: \end{equation}

335: Similarly, the total linear momentum $\vec{Q}_2$ of star~2 and 

336: the mass element to be accreted at time $t$ is given by

337: \begin{equation}

338: \label{eq-Q2-1}

339: \vec{Q}_2 = M_2\,\vec{V}_2 + \delta M_i\,\vec{W}_{\delta M_i},

340: \end{equation}

341: and the linear momentum $\vec{Q}_2^\prime$ of star~2 at time $t+\delta 

342: t$ by

343: \begin{equation}

344: \label{eq-Q2-2}

345: \vec{Q}_2^\prime = (M_2 + \delta M_2) \vec{V}_2^\prime.

346: \end{equation}

347: 

348: At time $t+\delta t$ the velocity of the center of mass of star $i$

349: can be written as

350: \begin{equation}

351: \vec{V}_i^\prime = \vec{V}_{O_i}^\prime

352:    + \left( \vec{\Omega}_{\rm orb}^\prime 

353:    + \delta \vec{\Omega}_{\rm orb}^\prime

354:    \right) \times \delta \vec{r}_i^{\,\prime},  \label{eq-Vtdt}

355: \end{equation}

356: where $\vec{V}_{O_i}^\prime$ is the absolute velocity of the origin of

357: $O_i X_i Y_i Z_i$ at time $t + \delta t$, $\vec{\Omega}_{\rm

358: orb}^\prime$ is the orbital angular velocity of the binary at time

359: $t+\delta t$ in the absence of mass transfer, and $\delta

360: \vec{\Omega}_{\rm orb}^\prime$ is the perturbation of the orbital

361: angular velocity at time $t + \delta t$ due to the mass loss/gain of

362: the binary components.

363: 

364: In the limit of small $\delta t$, taking the difference between the 

365: linear momenta $\vec{Q}_i^\prime$ and $\vec{Q}_i$, dividing the 

366: resulting equation by $\delta t$, and noting that the absolute velocity

367: $\vec{V}_{O_i}$ of the origin of $O_i X_i Y_i Z_i$ at time $t$ is

368: equal to $\vec{V}_i$, yields

369: \begin{equation}

370: M_i \frac{d\vec{V}_{O_i}}{dt} = \vec{F}_i +

371:   \dot{M}_i\, \vec{U}_{\delta M_i}. \label{eq-dVi}  

372: \end{equation}

373: Here, $\vec{F}_i = d\vec{Q}_i/dt$ is the sum of all external forces

374: acting on star $i$, $\dot{M_i} = dM_i/dt$ is the mass loss/accretion

375: rate of star $i$, and  

376: \begin{equation}

377: \vec{U}_{\delta M_i} = \vec{W}_{\delta M_i} - \vec{V}_{O_i} 

378:   - \vec{\Omega}_{\rm orb} \times \vec{r}_{A_i}

379: \label{eq-Ui}

380: \end{equation}

381: is the relative velocity of the ejected/accreted mass element with

382: respect to the ejection/accretion point $A_i$. In the derivation of 

383: equation~(\ref{eq-Ui}), we have made use of equations (\ref{eq-DelR}) and 

384: (\ref{eq-Vtdt}) and restricted ourselves to first-order terms in 

385: the small quantities $\delta M_i$ and $\delta \vec{\Omega}_{\rm 

386: orb}^\prime$.

387: 

388: The absolute acceleration of the center of mass of star $i$ with

389: respect to the inertial frame $OXYZ$ is given by 

390: \begin{equation}

391: \frac{d^2 \vec{R}_i}{dt^2} = \vec{\gamma}_{O_i}

392:   + \vec{\gamma}_{\rm rel,i} + \vec{\gamma}_{\rm cor,i},

393: \end{equation}

394: where $\vec{\gamma}_{O_i} = d\vec{V}_{O_i}/dt$ is the

395: acceleration of the origin of $O_i X_i Y_i Z_i$ with respect to

396: $OXYZ$, $\vec{\gamma}_{\rm rel,i} = (\ddot{M}_i/M_i)\, \vec{r}_{A_i}$

397: is the relative acceleration of the center of mass of star $i$ with

398: respect to $O_i$, and $\vec{\gamma}_{\rm cor,i} = 2\,

399: (\dot{M}_i/M_i)\, (\vec{\Omega}_{\rm orb} \times \vec{r}_{A_i})$ is

400: the Coriolis acceleration of the center of mass of star $i$ with

401: respect to $O_i$\footnote{The centrifugal acceleration does not play a

402: role since it is proportional to $\delta \vec{r}_i^{\,\prime}$ which

403: vanishes for small $\delta t$.}. The expressions for

404: $\vec{\gamma}_{\rm rel,i}$ and $\vec{\gamma}_{\rm cor,i}$ follow from

405: the observation that $d\vec{\rho}^{\,\prime}_{A_i}/dt = 

406: (\dot{M}_i/M_i)\,

407: \vec{r}_{A_i}$, which one obtains by dividing equation~(\ref{eq-Delrho}) by

408: $\delta t$ in the limiting case of small $\delta t$.  The equation of

409: motion for the mass center of star $i$ with respect to the inertial

410: frame $OXYZ$ then becomes

411: \begin{eqnarray}

412: M_i \frac{d^2 \vec{R}_i}{dt^2} &=& \vec{F}_i 

413:   + \dot{M}_i \left( \vec{U}_{\delta M_i} + 2\,

414:   \vec{\Omega}_{\rm orb} \times \vec{r}_{A_i} \right)

415:   \nonumber \\ 

416:  &+& \ddot{M}_i\, \vec{r}_{A_i}.

417: \label{eq-mtn}

418: \end{eqnarray}

419: 

420: \subsection{Relative Motion of the Binary Components}

421: \label{relmot}

422: 

423: We can now obtain the equation describing the relative motion of the

424: accretor (star~2) with respect to the donor (star~1) by taking the

425: difference of the equations of motion of the stars with respect to

426: the inertial frame of reference. For convenience, we first decompose

427: the sum of the external forces acting on each star as

428: \begin{equation}

429: \vec{F}_i = - \frac{G\,M_1\,M_2}{\left| \vec{r} \right|^2}\,

430:   \frac{\vec{R}_i}{| \vec{R}_i |} + \vec{f}_i,

431: \end{equation}

432: where $G$ is the Newtonian constant of gravitation, and $\vec{f}_i$

433: the total gravitational force exerted on star $i$ by the particles in

434: the mass-transfer stream. It follows that

435: \begin{eqnarray}

436: \label{eq-R21}

437: \frac{d^2\vec{r}}{dt^2} &=& -\frac{G \left( M_1+M_2

438:     \right)}{\left| \vec{r} \right|^3}\, \vec{r} 

439:     + \frac{\vec{f}_2}{M_2}

440:     - \frac{\vec{f}_1}{M_1}  \nonumber \\

441: &+& \frac{\dot{M}_2}{M_2} \left( \vec{v}_{\delta M_2} +

442:     \vec{\Omega}_{\rm orb} \times \vec{r}_{A_2} \right) \nonumber \\

443: &-& \frac{\dot{M}_1}{M_1} \left( \vec{v}_{\delta M_1} +

444:     \vec{\Omega}_{\rm orb} \times \vec{r_{A_1}} \right) \\

445: &+& \frac{\ddot{M_2}}{M_2}\, \vec{r}_{A_2} 

446:   - \frac{\ddot{M_1}}{M_1}\, \vec{r}_{A_1}, \nonumber 

447: \end{eqnarray}

448: where $\vec{r} = \vec{R}_2 - \vec{R}_1$ is the position vector of the

449: accretor with respect to the donor, and $\vec{v}_{\delta M_i} =

450: \vec{W}_{\delta M_i} - \vec{V}_{i}$ is the velocity of the

451: ejected/accreted mass element with respect to the mass center of the

452: mass losing/gaining star. 

453: 

454: Equation~(\ref{eq-R21}) can be written in the form of a perturbed

455: two-body problem as

456: \begin{equation}

457: \frac{d^2\vec{r}}{dt^2} = -\frac{G \left( M_1+M_2 \right)}

458:   {\left| \vec{r} \right|^3}\,

459:   \vec{r} + S\, \hat{x} + T\, \hat{y} + W\, \hat{z},

460: \end{equation}

461: where $\hat{x}$ is a unit vector in the direction of $\vec{r}$,

462: $\hat{y}$ is a unit vector in the orbital plane perpendicular to

463: $\vec{r}$ in the direction of the orbital motion, and $\hat{z}$ is a

464: unit vector perpendicular to the orbital plane parallel to and in the

465: same direction as $\vec{\Omega}_{\rm orb}$.  The functions $S$, $T$,

466: and $W$ are found by taking the dot product of the perturbing force 

467: arising from the mass transfer between the binary components and the 

468: unit vector in the $\hat{x}$, $\hat{y}$, and $\hat{z}$ directions, 

469: respectively.  These vector components are

470: \begin{eqnarray}

471: S &=& \frac{f_{2,x}}{M_2} - \frac{f_{1,x}}{M_1}  

472: + \frac{\dot{M}_2}{M_2} \left( v_{\delta M_2,x} - |\vec{\Omega}_{\rm orb}|

473: |\vec{r}_{A_2}| \sin{\phi} \right) \nonumber \\

474: &-& \frac{\dot{M}_1}{M_1} v_{\delta M_1,x} 

475: + \frac{\ddot{M}_2}{M_2} |\vec{r}_{A_2}| \cos{\phi}

476: - \frac{\ddot{M}_1}{M_1} |\vec{r}_{A_1}|, \label{eq-S} \\

477: T &=& \frac{f_{2,y}}{M_2} - \frac{f_{1,y}}{M_1}  

478: + \frac{\dot{M}_2}{M_2} \left( v_{\delta M_2,y} 

479: + |\vec{\Omega}_{\rm orb}| |\vec{r}_{A_2}| \cos{\phi} \right) \nonumber \\ 

480: &-&\frac{\dot{M}_1}{M_1} \left( v_{\delta M_1,y}

481: + |\vec{\Omega}_{\rm orb}||\vec{r}_{A_1}| \right) +

482: \frac{\ddot{M}_2}{M_2} |\vec{r}_{A_2}| \sin{\phi}, \label{eq-T} \\

483: W &=& \frac{f_{2,z}}{M_2} - \frac{f_{1,z}}{M_1}, \label{eq-W}

484: \end{eqnarray}

485: where $\phi$ is the angle between $\hat{x}$ and the vector from the 

486: center of mass of the accretor to the mass

487: accretion point $A_2$, and the subscripts $x$, $y$, and $z$ 

488: denote vector components in the $\hat{x}$, $\hat{y}$, and $\hat{z}$ 

489: directions, respectively. In working out the vector products

490: $\vec{\Omega}_{\rm orb} \times \vec{r}_{A_2}$, we assumed that $A_2$ is 

491: located on the equator of the accreting star\footnote{For brevity, we refer to the point $A_2$ as lying on the 

492: stellar surface. Though, in practice, it can lie at any point near the 

493: star where the 

494: transferred mass can be considered to be part of the accretor. 

495: For instance, if an accretion disk has formed around the 

496: accretor, it would be equally valid to write $A_2$ as the point 

497: where the transferred mass impacts the outer edge of the accretion 

498: disk.}.  The terms contributing to the perturbed orbital motion can be 

499: categorized as follows: (i) term proportional to $\vec{f}_i$ represent 

500: gravitational perturbation on the binary components caused by mass 

501: elements in the mass-transfer stream; (ii) terms proportional to 

502: $\dot{M}_i$ represent linear momentum exchange between the mass donor 

503: and accretor; and (iii) terms proportional to $\ddot{M}_i$ represent 

504: shifts in the position of the mass centers of the mass donor and 

505: accretor due to the non-spherical symmetry of the mass loss or gain.  In 

506: the limiting case where both stars are treated as point masses

507: ($|\vec{r}_{A_1}| \rightarrow 0$ and $|\vec{r}_{A_2}| \rightarrow 0$), 

508: the only non-zero terms in the perturbed equations of motion are those 

509: due to gravitational perturbations of the mass transfer stream and the 

510: transport of linear momentum.

511: 

512: \subsection{Comparison with Previous Work}

513: \label{sec-appb}

514: 

515: The most recent study on the orbital evolution of eccentric

516: mass-transferring binaries has been presented by \citet[][hereafter MW83

517: and MW84, respectively]{MW83, MW84}. These authors extended the work

518: of \citet{H56}, \citet{K64-2}, and \citet{P64} by accounting for the

519: effects of linear momentum transport between the binary components, as

520: well as possible perturbations to the orbital motion caused by the

521: mass transfer stream. However, they also derived the equations

522: describing the motion of the binary components with respect to a frame

523: of reference with origin at the mass center of the binary, which, for

524: mass transferring systems, is not an inertial frame of reference. The

525: equations therefore do not account for the accelerations of the binary

526: mass center caused by the mass transfer. Here, we

527: demonstrate that if the procedure adopted by Matese \& Whitmire is

528: developed with respect to an inertial frame of reference that is not

529: connected to the binary, the resulting equations are in agreement with

530: those derived in \S\,\ref{relmot}.

531: 

532: The core of Matese \& Whitmire's derivation is presented in Section II

533: of MW83. While the authors choose to adopt a reference frame with

534: origin at the binary mass center early on in the investigation, the

535: choice of the frame does not affect the derivation of the equations of

536: motion up to and including their equation~(24). In particular, equation~(13) in

537: MW83, which, in our notation, reads

538: \begin{equation}

539: \label{eq-apb-p}

540: \vec{p}_i = M_i \dot{\vec{R}}_i - \dot{M}_i\vec{r}_{A_i}, 

541: \end{equation}

542: is valid with respect to any inertial frame of reference with

543: arbitrary position and orientation in space. The same applies to equation (1)

544: in MW84:

545: \begin{equation}

546: \label{eq-apb-pdot}

547: \dot{\vec{p}}_i = -G M_1 M_2\,

548:   \frac{\vec{R}_i-\vec{R}_{3-i}}{|\vec{R}_i-\vec{R}_{3-i}|^3}

549:   + \vec{f}_i + \vec{\Psi}_i.

550: \end{equation}

551: In these equations, $\vec{p}_i$ is the linear momentum of star $i$,

552: and $\vec{\Psi}_i=\dot{M}_i (\dot{\vec{R}}_i+\vec{v}_{\delta M_i} )$

553: is the amount of linear momentum transported by the transferred mass

554: per unit time (see equation~(3) of MW84). Substitution of

555: equation~(\ref{eq-apb-p}) into equation~(\ref{eq-apb-pdot}) then yields

556: \begin{eqnarray}

557: M_i\, \ddot{\vec{R}}_i = &-& GM_{3-i}\,

558:   \frac{\vec{R}_i-\vec{R}_{3-i}}{|\vec{R}_i-\vec{R}_{3-i}|^3} \nonumber \\

559:   &+& \frac{\vec{f}_i}{M_i} + \frac{\dot{M}_i}{M_i} 

560:   \left( \vec{v}_{\delta M_i} +

561:   \dot{\vec{r}}_{A_i} \right) + \frac{\ddot{M}_i}{M_i}\vec{R}_{A_i},

562: \label{eq-apb-mrddot}

563: \end{eqnarray}

564: and thus

565: \begin{eqnarray}

566: \label{eq-apb-eqm}

567: \frac{d^2\vec{r}}{dt^2} = &-&\frac{G \left( M_1+M_2

568:   \right)}{\left| \vec{r} \right|^3}\, \vec{r}

569:   + \frac{\vec{f}_2}{M_2} - \frac{\vec{f}_1}{M_1} \nonumber \\

570:   &+& \frac{\dot{M}_2}{M_2} \left( \vec{v}_{\delta M_2} +

571:     \dot{\vec{r}}_{A_2} \right)

572:   - \frac{\dot{M}_1}{M_1} \left( \vec{v}_{\delta M_1} +

573:     \dot{\vec{r}}_{A_1} \right) \nonumber \\

574:   &+& \frac{\ddot{M_2}}{M_2}\, \vec{r}_{A_2} 

575:   - \frac{\ddot{M_1}}{M_1}\, \vec{r}_{A_1}, 

576: \end{eqnarray}

577: where $\vec{r} = \vec{R}_2 - \vec{R}_1$.  Setting $\dot{\vec{r}}_{A_i}

578: = \vec{\Omega}_{\rm orb} \times \vec{r}_{A_i}$, this equation is in

579: perfect agreement with equation~(\ref{eq-R21}) derived in

580: \S\,\ref{relmot}.

581: 

582: In MW83 and MW84, the authors incorrectly set $\vec{R}_1=-M_2\,

583: \vec{r}/(M_1+M_2)$ and $\vec{R}_2=M_1\, \vec{r}/(M_1+M_2)$ in

584: equation~(\ref{eq-apb-mrddot}), which is valid only when the origin of the

585: frame of reference coincide with the mass center of the binary. For a

586: mass-transferring binary, such a frame is, however, not an inertial

587: frame and can therefore not be used for the derivation of the

588: equations of motion of the binary components. Instead of

589: equation~(\ref{eq-apb-eqm}), Matese \& Whitmire therefore find

590: equations~(7)--(8) in MW84, which lack the terms associated with the

591: acceleration of the binary mass center due to the mass transfer.

592: 

593: \section{Orbital Evolution Equations}

594: \label{sec-sec}

595: 

596: \subsection{Secular Variation of the Orbital Elements}

597: 

598: In the classical framework of the theory of osculating elements, the

599: equations governing the rate of change of the orbital semi-major axis 

600: $a$ and eccentricity $e$ due to mass transfer are obtained from the 

601: perturbing functions $S$ and $T$ as \citep[see, e.g.,][]{S60, BC61, D62, 

602: F70}

603: \begin{equation}

604: \frac{da}{dt} = \frac{2}{n(1-e^2)^{1/2}} [ S e \sin{\nu} + T ( 1

605:   + e \cos{\nu} )],

606: \label{eq-dadt}

607: \end{equation}

608: \begin{eqnarray}

609: \lefteqn{\frac{de}{dt} = \frac{(1-e^2)^{1/2}}{na}} 

610:   \nonumber \\

611:  & & \times \left\{ S \sin{\nu} +

612:   T \left[ \frac{2\cos{\nu} +e \left(1+\cos^2{\nu}

613:   \right)}{1+e\cos{\nu}} \right] \right\},

614: \label{eq-dedt}

615: \end{eqnarray}

616: where $n=2\pi/P_{\rm orb}$ is the mean motion and $\nu$ the true

617: anomaly. These equations are independent of the perturbing function

618: $W$ which solely appears in the equations governing the rates of

619: change of the orbital inclination, the longitude of the ascending

620: node, and the longitude of the periastron. 

621: 

622: After substitution of equations~(\ref{eq-S}) and (\ref{eq-T}) for $S$ and

623: $T$ into equations~(\ref{eq-dadt}) and~(\ref{eq-dedt}), the equations

624: governing the rates of change of the semi-major axis and eccentricity

625: contain periodic as well as secular terms. Here we are mainly

626: interested in the long-term secular evolution of the orbit, and so we

627: remove the periodic terms by averaging the equations over one orbital

628: period:

629: \begin{equation}

630: \left< {\frac{da}{dt}} \right>_{\rm sec} \equiv 

631:   \frac{1}{P_{\rm orb}} \int_{-P_{\rm orb}/2}^{P_{\rm orb}/2}

632:   {\frac{da}{dt}}\, dt,

633: \label{eq-dadtsec}

634: \end{equation}

635: \begin{equation}

636: \left< {\frac{de}{dt}} \right>_{\rm sec} \equiv 

637:   \frac{1}{P_{\rm orb}} \int_{-P_{\rm orb}/2}^{P_{\rm orb}/2}

638:   {\frac{de}{dt}}\, dt.

639: \label{eq-dedtsec}

640: \end{equation}

641: The integrals in these definitions are most conveniently computed in

642: terms of the true  anomaly, $\nu$. We therefore make a change of 

643: variables using

644: \begin{equation}

645: dt=\frac{(1-e^2)^{3/2}}{n(1+e\cos{\nu})^2}\,d\nu.

646: \label{eq-dnudt}

647: \end{equation}

648: For binaries with eccentric orbits, the resulting integrals can be

649: calculated analytically only for very specific functional

650: prescriptions of the mass-transfer rate $\dot{M}_1$ (e.g., when

651: $\dot{M}_1$ is approximated by a Dirac delta function centered on the

652: periastron, see \S\,\ref{sec-delta}). In general, the

653: integrals must be computed numerically.

654: 

655: \subsection{Conservation of Orbital Angular Momentum}

656: 

657: \label{sec-angmom}

658: 

659: Since the perturbing functions $S$ and $T$ depend on the properties

660: of the mass transfer stream, calculation of the rates 

661: of secular

662: change of the orbital semi-major axis and eccentricity, in principle,

663: requires the calculation of the trajectories of the particles in the

664: stream \citep[cf.][]{H69b}. As long as no mass is lost from the

665: system, such a calculation automatically incorporates the conservation

666: of total angular momentum in the system. Special cases of angular

667: momentum conservation can, however, be used to bypass the

668: calculation of detailed particle trajectories. Here, we adopt such a

669: special case and assume that any orbital angular momentum carried by

670: the particles in the mass-transfer stream is always immediately

671: returned to the orbit, so that the orbital angular momentum of the

672: binary is conserved.

673: 

674: The orbital angular momentum of a binary with a semi-major axis $a$ and 

675: eccentricity $e$ is given by

676: \begin{equation}

677: J_{\rm orb} = M_1M_2 \left[ \frac{Ga(1-e^2)}{M_1+M_2} \right]^{1/2},

678: \label{eq-J}

679: \end{equation}

680: so that

681: \begin{equation}

682: \frac{\dot{J}_{\rm orb}}{J_{\rm orb}} = \frac{\dot{M_1}}{M_1} +

683: \frac{\dot{M_2}}{M_2}  

684: - \frac{1}{2}\frac{\dot{M_1}+\dot{M_2}}{M_1+M_2}

685: + \frac{1}{2}\frac{\dot{a}}{a} - \frac{e\,\dot{e}}{1-e^2},

686: \label{eq-Jdot}

687: \end{equation}

688: where a dot indicates the time derivative.

689: 

690: In the case of eccentric orbits, substitution of 

691: equations~(\ref{eq-dadt}) and (\ref{eq-dedt}) into equation 

692: (\ref{eq-Jdot}) leads to 

693: \begin{eqnarray}

694: \frac{\dot{J}_{\rm orb}}{J_{\rm orb}} &=& 

695:   \frac{\dot{M_1}}{M_1} + \frac{\dot{M_2}}{M_2} 

696:   - \frac{1}{2}\frac{\dot{M_1}+\dot{M_2}}{M_1+M_2} \nonumber \\ 

697: &+& \frac{(1-e^2)^{1/2}}{n\,a \left(1+e\cos{\nu} \right)}\,T.

698: \label{eq-JT}

699: \end{eqnarray}

700: As we shall see in the next section, by setting $\dot{M}_1 + \dot{M}_2 =

701: 0$ and $< \dot{J}_{\rm orb}/J_{\rm orb} >_{\rm sec} =0$ and substituting

702: equation~(\ref{eq-T}) for $T$, equation~(\ref{eq-JT}) allows us to

703: calculate the $\hat{y}$-component of the final velocities of the 

704: accreting

705: particles as a function of their initial velocities without resorting to

706: the computation of the ballistic trajectories of the mass transfer

707: stream.  

708: 

709: In the limiting case of a circular orbit, equation~(\ref{eq-Jdot}) is 

710: usually used to derive the rate of change of the semi-major axis of 

711: circular binaries under the assumption of conservation of both total 

712: mass ($\dot{M}_1=-\dot{M}_2$) and orbital angular momentum ($\dot{J}_{\rm 

713: orb}=0$):

714: \begin{equation}

715: \frac{da}{dt} = 2a(\frac{M_1}{M_2}-1)\frac{\dot{M}_1}{M_1}.

716: \label{circ-da}

717: \end{equation}

718: 

719: The assumption of orbital angular momentum conservation over secular

720: timescales ($< \dot{J}_{\rm orb}/J_{\rm orb} >_{\rm sec} =0$) is a

721: standard assumption in nearly all investigations of conservative mass

722: transfer in binary systems \citep[e.g., ][]{SPH97, P98}, which is valid

723: over long timescales provided there is no significant storage of angular

724: momentum in the spins of the components stars, the accretion flow,

725: and/or the accretion disk.  In future work, we will investigate the

726: consequences of both mass and orbital angular momentum losses from the

727: binary on the evolution of the orbital elements.

728: 

729: \section{Orbital Evolution Timescales}

730: \label{orbevtim}

731: 

732: In order to assess the timescales of orbital evolution due to mass

733: transfer in eccentric binaries, we observe that, for eccentric binaries, 

734: mass transfer is expected to occur first at the periastron of the 

735: relative orbit, where the component stars are closest to each other.  We 

736: therefore explore the order of magnitude of the timescales assuming a 

737: delta function mass transfer profile centered at the periastron of the 

738: binary orbit

739: \begin{equation}

740: \dot{M}_1 = \dot{M}_0\, \delta \left( \nu \right),

741: \label{eq-del}

742: \end{equation}

743: where $\dot{M}_0 < 0 $ is the instantaneous mass transfer rate, and 

744: $\delta(\nu)$ is the Dirac delta function.

745: 

746: We calculate the rates of secular change of the orbital semi-major axis

747: and eccentricity from equations~(\ref{eq-dadt})--(\ref{eq-dnudt}), and 

748: neglect any gravitational attractions exerted by the particles in the

749: mass-transfer stream on the component stars. Hence, we set

750: \begin{eqnarray}

751: f_{1,x} &=& f_{2,x} = 0, \label{eq-f12x} \\

752: f_{1,y} &=& f_{2,y} = 0. \label{eq-f12y}

753: \end{eqnarray}

754: 

755: \label{sec-delta}

756: 

757: Substituting equations~(\ref{eq-T}), (\ref{eq-dnudt}), and 

758: (\ref{eq-del}) -- (\ref{eq-f12y}) into equation~(\ref{eq-JT}) for $< 

759: \dot{J}_{\rm orb}/J_{\rm orb} >_{\rm sec}=0$ then yields a relationship 

760: between the initial and final $\hat{y}$-component of the velocities of 

761: the transferred mass and the initial and final positions of the 

762: transferred mass given by

763: \begin{eqnarray}

764: qv_{\delta M_2,y} &+& v_{\delta M_1,y} = 

765:  na(1-q)\left(\frac{1+e}{1-e}\right)^{1/2}

766: - |\vec{\Omega}_{{\rm orb}, P}||\vec{r}_{A_1,P}| \nonumber \\

767: &-& q|\vec{\Omega}_{{\rm orb}, P}||\vec{r}_{A_2}|\cos{\phi_P} 

768: \left( \left.1- \frac{d\phi}{d\nu}\right|_{\nu=0} \right),

769: \label{eq-vr}

770: \end{eqnarray}

771: where the subscript $P$ indicates quantities evaluated at the periastron

772: of the binary orbit, $q=M_1/M_2$ is the binary mass ration, and we have 

773: used the relation

774: $d\delta(\nu)/d\nu = -\delta(\nu)/\nu$.  Assuming the transferred mass

775: elements are ejected by star~1 at the $L_1$ point with a velocity

776: $\vec{v}_{\delta M_1}$ equal to the star's rotational velocity at $L_1$,

777: we write

778: \begin{eqnarray}

779: \label{eq-vdm1x}

780: v_{\delta M_1, x} = 0,\\

781: \label{eq-vdm1y}

782: v_{\delta M_1,y}=-|\vec{\Omega}_{{\rm orb},P}||\vec{r}_{A_1, P}|.

783: \end{eqnarray}

784: Moreover, under the assumption that each periastron passage of the

785: binary components give rise to an extremum of $\phi(\nu)$, the

786: derivative $d\phi/d\nu|_{\nu=0}$ is equal to zero in

787: equation~(\ref{eq-vr}), so that

788: \begin{equation}

789: v_{\delta M_2, y} = |\vec{\Omega}_{{\rm orb},P}|\left[ |\vec{r}_{A_1,P}| 

790: \frac{(1-q)}{q} + |\vec{r}_{A_2}|\cos{\phi_P} \right].

791: \label{eq-vrel}

792: \end{equation}

793: 

794: For a binary with orbital period $P_{\rm orb} = 1\,{\rm day}$,

795: eccentricity $e=0.2$, and component masses $M_1=2\,M_\sun$ and

796: $M_2=1.44\,M_\sun$, with $|\vec{r}_{A_1,P}|$ the distance from star~1 to

797: $L_1$ (See Appendix A), and $|\vec{r}_{A_2}|\cos{\phi_P}\approx

798: 2.9\times10^5\,{\rm km}$ (the circularization radius around a compact

799: object for these binary parameters; see \citet{FKR}), the accreting

800: matter has a $\hat{y}$-velocity component of the order of $\sim-45\,{\rm

801: km}\,{\rm s}^{-1}$.  

802: 

803: 

804: After substitution of equation~(\ref{eq-dnudt}) and 

805: equations~(\ref{eq-f12x})--(\ref{eq-vr}), the integrals in

806: equations~(\ref{eq-dadtsec}) and (\ref{eq-dedtsec}) for the rates of

807: secular change of the orbital semi-major axis and eccentricity can be

808: solved analytically to obtain

809: \begin{eqnarray}

810: \left< \frac{da}{dt} \right>_{\rm sec} &=& 

811: \frac{a}{\pi}\frac{\dot{M}_0}{M_1}\frac{1}{(1-e^2)^{1/2}}\left[

812: qe\frac{|\vec{r}_{A_2}|}{a}\cos{\phi_P} \right. \nonumber \\

813:  &+& \left. e\frac{|\vec{r}_{A_1,P}|}{a} +

814: (q-1)(1-e^2) \right],

815: \label{eq-deltaa} 

816: \end{eqnarray}

817: \begin{eqnarray}

818: \left< \frac{de}{dt} \right>_{\rm sec} &=& 

819: \frac{(1-e^2)^{1/2}}{2\pi} \frac{\dot{M}_0}{M_1}\left[ 

820: q\frac{|\vec{r}_{A_2}|}{a}\cos{\phi_P} \right. \nonumber \\

821: &+& \left. \frac{|\vec{r}_{A_1,P}|}{a} + 2(1-e)(q-1) \right].

822: \label{eq-deltae}

823: \end{eqnarray}

824: We note that for a delta-function mass transfer

825: rate given by equation~(\ref{eq-del}) the $\hat{x}$-component of the

826: velocities $\vec{v}_{\delta M_1}$ and $\vec{v}_{\delta M_2}$ does not

827: enter into the derivation of theses equations due to the $\sin{\nu}$ 

828: term in equations~(\ref{eq-dadt}) and (\ref{eq-dedt}).  Furthermore, in 

829: the limiting case of a circular orbit,

830: equation~(\ref{eq-deltaa}) reduces to equation~(\ref{circ-da}), provided that

831: $\dot{M_1}$ in that equation is interpreted as the secular mean mass

832: transfer rate $\left<\right.\!\dot{M_1}\!\left.\right>_{\rm

833: sec}=\dot{M}_0/(2\pi)$.  In Appendix~B, we present an alternative

834: derivations to equations~(\ref{eq-deltaa}) and (\ref{eq-deltae}) in the

835: limiting case where the stars are treated as point masses.

836: 

837: The rates of secular change of the semi-major axis and orbital

838: eccentricity are thus linearly proportional to the magnitude of the

839: mass transfer rate at periastron. Besides the obvious dependencies on

840: $a$, $e$, $q$, and $M_1$, the rates also depend on the ratio of the

841: donor's rotational angular velocity $\Omega_1$ to the orbital angular

842: velocity $\Omega_{{\rm orb},P}$ at periastron through the position of 

843: the $L_1$

844: point, $\vec{r}_{A_1}$. A fitting formula for the position of the

845: $L_1$ point accurate to better than $4\,\%$ over a wide range of $q$,

846: $e$, and $\Omega_1/\Omega_{{\rm orb},P}$ is given by equation~(\ref{XL1fit}) in

847: Appendix~\ref{sec-appa}. While the fitting formula can be used to

848: obtain fully analytical rates of secular change of the semi-major axis

849: and eccentricity, we here use the exact solutions for the

850: position of $L_1$ obtained by numerically solving equation~(\ref{XL}) in

851: Appendix~\ref{sec-appa}.  For a detailed discussion of the 

852: properties of the $L_1$ point in eccentric binaries, we refer the 

853: interested reader to \citet{SWK07}.

854: 

855: \begin{figure*}

856: \plottwo{f2a.eps}{f2b.eps}

857: \caption{Orbital evolution timescales for a delta function mass

858:   transfer profile centered at the periastron of the binary orbit with 

859:   an instantaneous mass transfer rate of $\dot{M}_0 = - 

860:   10^{-9}\,M_\sun\,{\rm yr}^{-1}$. The timescales are calculated under

861:   the assumption that the donor rotates synchronously with the orbital

862:   angular velocity at periastron, and that the accretor is a

863:   $1.44\,M_\odot$ neutron star.  Shown at left (right) are the

864:   timescales for the evolution of the semi-major axis (eccentricity) as

865:   a function of the mass ratio, $q$, for a range of eccentricities, $e$.

866:   Regimes where the timescale is negative correspond to a decrease of

867:   the semi-major axis (eccentricity), while regimes where the

868:   timescale is positive correspond to an increase of the semi-major

869:   axis (eccentricity).

870: }

871: \label{delta1}

872: \end{figure*}

873: 

874: To explore the effects of mass transfer on the orbital elements of

875: eccentric binaries, we calculate the rates of secular change of the

876: semi-major axis and eccentricity and determine the characteristic

877: timescales $\tau_a = a/\dot{a}$ and $\tau_e = e/\dot{e}$. While the

878: actual timescales are given by the absolute values of $\tau_a$

879: and $\tau_e$, we here allow the timescales to be negative as well as

880: positive in order to distinguish negative from positive rates of

881: secular change of the orbital elements. We also note that since

882: $|\vec{r}_{A_1, P}| \propto a$ (see Appendix A), the timescales do not

883: explicitly depend on the orbital semi-major axis $a$ except through 

884: the ratio $|\vec{r}_{A_2}|/a$ of the radius of the accretor to the 

885: semi-major axis.  For convenience, we therefore assume the accretor to 

886: be a compact object with radius $|\vec{r}_{A_2}| << a$.  The timescales 

887: are found to be insensitive to terms containing $|\vec{r}_{A_2}|/a$ in 

888: equation (\ref{eq-deltaa}) and (\ref{eq-deltae}).  Varying 

889: $|\vec{r}_{A_2}|$ from $0$ to $0.01a$ changes the timescales by less 

890: that 10\%.  In what follows, we therefore set $|\vec{r}_{A_2}|=0$.  An

891: implicit dependence on $a$ may then still occur through the amplitude 

892: $\dot{M}_0$ of the mass transfer rate at periastron. Since incorporating 

893: such a dependence in the analysis requires detailed modeling of the 

894: evolution of the donor star, which is beyond the scope of this 

895: investigation, we here restrict ourselves to exploring the timescales of 

896: orbital evolution for a constant $\dot{M}_0$. The linear dependence of 

897: $\left <\dot{a} \right >_{\rm sec}$ and $\left < \dot{e} \right >_{\rm 

898: sec}$ on $\dot{M}_0$ in any case allows for any easy rescaling of 

899: our results to different mass transfer rates.  

900: 

901: In Fig.~\ref{delta1}, we show the variations of $\tau_a$ and $\tau_e$ as

902: functions of $q$ for $\dot{M}_0 = -10^{-9}\,M_\sun\,{\rm yr^{-1}}$ and

903: $e=0.0,0.1, \ldots, 0.9$. In all cases, the donor is assumed to rotate

904: synchronously with the orbital angular velocity at the periastron, and

905: the accretor is assumed to be a neutron star of mass

906: $M_2=1.44\,M_\odot$.  The timescales of the secular evolution of the

907: semi-major axis show a strong dependence on $q$, and a milder dependence

908: on $e$, unless $e \ga 0.7$. The timescales for the secular evolution of

909: the orbital eccentricity always depend strongly on both $q$ and $e$. 

910: These timescales can furthermore be positive as well as negative, so

911: that the semi-major axis and eccentricity can increase as well as

912: decrease under the influence of mass transfer at the periastron of the 

913: binary orbit.

914: 

915: From Fig.~\ref{delta1}, as well as equations~(\ref{eq-deltaa}) and

916: (\ref{eq-deltae}), it can be seen that, for a given ratio of the donor's

917: rotational angular velocity to the orbital angular velocity at

918: periastron, the line dividing positive from negative rates of secular

919: change of the orbital elements is a function of $q$ and $e$. This is

920: illustrated further in Fig.~\ref{contours} where the timescales of

921: orbital evolution are displayed as contour plots in the $(q,e)$-plane.

922: The thick black line near the center of the plots marks the transition

923: values of $q$ and $e$ where the rates of secular change of $a$ and $e$

924: transition from being positive (to the left of the thick black line) to

925: negative (to the right of the thick black line). Varying

926: $\Omega_1/\Omega_{{\rm orb},P}$ between $0.5$ and $1.5$ changes the

927: position of the transition line by less than $10\%$ in comparison to the

928: $\Omega_1/\Omega_{{\rm orb},P}=1$ case displayed in Figures~\ref{delta1}

929: and \ref{contours}.

930: 

931: \begin{figure*}

932: \plottwo{f3a.eps}{f3b.eps}

933: \caption{Contour plot of the orbital evolution timescales for the 

934:   semi-major axis and eccentricity in the

935:   $(q,e)$-plane for the same set of assumptions as adopted in 

936:   figure~\ref{delta1}.  Timescales for the evolution of the 

937:   semi-major axis are shown

938:   on the left; timescales for the evolution of the orbital

939:   eccentricity on the right. The different shades of gray designate

940:   regions of the $(q,e)$ parameter space with timescales (in Gyr) in the 

941:   ranges labeled in the plots.  The thick black line near the center

942:   of each plot designates the transition point where the rate of

943:   change of the semi-major axis or orbital eccentricity changes from

944:   positive (to the left of the thick black line) to negative (to 

945:   the right of the thick black line).}

946: \label{contours}

947: \end{figure*}

948: 

949: In the limiting case of a circular orbit, the orbit expands when

950: $q<1$ and shrinks when $q>1$, in agreement with the classical result

951: obtained from equation~(\ref{circ-da}). For non-zero eccentricities, the

952: critical mass ratio separating positive from negative values of $\left

953: < \dot{a} \right >_{\rm sec}$ decreases with increasing orbital

954: eccentricities. This behavior can be understood by substituting the

955: fitting formula for the position of the $L_1$ point given by

956: equation~(\ref{XL1fit}) in Appendix~\ref{appL1} into equation~(\ref{eq-deltaa})

957: and setting $\left< \dot{a} \right>_{\rm sec} =0$.  However, we can fit 

958: the critical mass ratio separating expanding from shrinking orbits with 

959: a simpler formula given by

960: \begin{equation}

961: q_{\rm crit} \simeq 1 - 0.4e + 0.18e^2.

962: \label{eq-dadttran}

963: \end{equation}

964: 

965: The critical mass ratio separating positive from negative values of

966: $\left < \dot{e} \right >_{\rm sec}$ is largely independent of

967: $e$. Proceeding in a similar fashion as for the derivation of

968: equation~(\ref{eq-dadttran}), we derive the critical mass ratio separating

969: increasing from decreasing eccentricities to be approximately given by

970: \begin{equation}

971: q_{\rm crit} \simeq 0.76 + 0.012e.

972: \label{eq-dedttran}

973: \end{equation}

974: 

975: Last, we note that a more quantitative numerical comparison between

976: the above approximation formulae for $q_{\rm crit}$ and the exact

977: numerical solutions shows that equations~(\ref{eq-dadttran}) and

978: (\ref{eq-dedttran}) are accurate to better than 1\%.

979: 

980: \section{Tidal Evolution Timescales}

981: 

982: A crucial question for assessing the relevance of the work

983: presented here is how the derived orbital evolution timescales compare

984: to the corresponding timescales associated with other orbital

985: evolution mechanisms such as tides. In Fig.~\ref{fig-tid-lin}, we show

986: the secular evolution timescales of the semi-major axis and orbital

987: eccentricity of a mass-transferring binary due to tidal dissipation in

988: the donor star as a function of $q$, for different values of the 

989: eccentricity, $e$. The timescales are strong functions of 

990: $|\vec{r}_{A_1}|/a$ and are determined as in \citet{HTP02}\footnote{Note 

991: that there is a typo in equation~(42) of \citet{HTP02}. The correct equation 

992: for $k/T$ for stars with radiative envelopes is (J. Hurley, Private 

993: Communication)

994: \[

995: \left( k/T\right)_r =

996:   1.9782 \times 10^4\left( MR^2/a^5 \right)^{1/2}

997:   \left(1+q_2\right)^{5/6}E_2\,{\rm yr}^{-1}. \nonumber

998: \]}

999: \citep[see also][]{Z77,Z78,H81}.  The radius $|\vec{r}_{A_1}|$ is

1000: determined by assuming the donor is on the zero-age

1001: main sequence and that the orbital separation is then obtained by 

1002: equating the radius of the donor \citep[given by][]{TEA96} to the 

1003: volume-equivalent radius of its Roche lobe at the periastron of the 

1004: binary orbit \citep[see][]{SWK07}.  As before, we assume the donor

1005: rotates synchronously with the orbital motion at periastron and that the 

1006: accretor is a $1.44\,{\rm M}_\sun$ neutron star.

1007: 

1008: \begin{figure*}

1009: \plottwo{f4a.eps}{f4b.eps}

1010: \caption{Timescales of orbital evolution due to tidal dissipation in a

1011:   Roche-lobe filling component of a close binary under the assumption

1012:   that the donor is a zero-age main-sequence star rotating

1013:   synchronously with the orbital angular velocity at the periastron,

1014:   and the accretor is a $1.44\,M_\odot$ neutron star.  Shown at left

1015:   (right) are the timescales for the evolution of the semi-major axis

1016:   (eccentricity) as a function of the mass ratio, $q$, for 

1017:   different orbital eccentricities $e$.  Regimes

1018:   where the timescales are negative correspond to a decrease of the

1019:   semi-major axis (eccentricity), while regimes where the timescales

1020:   are positive correspond to an increase of the semi-major axis

1021:   (eccentricity). The discontinuity at $q \simeq 0.87$ corresponds to

1022:   the transition from donor stars with convective envelopes ($M_1 \la

1023:   1.25\,M_\odot$) to donor stars with radiative envelopes ($M_1 \ga

1024:   1.25\,M_\odot$).}

1025: \label{fig-tid-lin}

1026: \end{figure*}

1027: 

1028: The timescales of orbital evolution due to tides range from a few Myr to

1029: more than a Hubble time, depending on the binary mass ratio and the

1030: orbital eccentricity. The discontinuity in the timescales at $q \simeq

1031: 0.87$ corresponds to the transition from donor stars with convective 

1032: envelopes ($M_1 \la 1.25\,M_\odot$) to donor stars with radiative 

1033: envelopes ($M_1  \ga 1.25\,M_\odot$) which are subject to different 

1034: tidal dissipation mechanisms.  It follows that tides do not 

1035: necessarily lead to rapid circularization during the early stages of 

1036: mass transfer, especially for orbital eccentricities $e \ga 0.3$. 

1037: Furthermore, for the adopted system parameters, the orbital eccentricity 

1038: always decreases, while the orbital semi-major axis can either increase 

1039: or decrease. Hence, in some regions of the parameter space, the effects 

1040: of tides and mass transfer are additive, while in other regions they are

1041: competitive.  This is illustrated in more detail in 

1042: figure~\ref{fig-tot_timescale} where we show the orbital evolution 

1043: timescales due to the combined effect of tides and mass transfer.  In 

1044: the calculations of the timescales, we have assumed that, at the lowest 

1045: order of approximation, the effects of tides and mass transfer are 

1046: decoupled.  The total rate of change of the orbital elements is then

1047: given by the sum of the rate of change of the orbital elements due to 

1048: tides and mass transfer.

1049: 

1050: \begin{figure*}

1051: \plottwo{f5a.eps}{f5b.eps}

1052: \caption{Orbital evolution timescales due to the combined effects of 

1053: tidal dissipation in a Roche Lobe filling component of a close binary 

1054: system and a delta function mass transfer with an amplitude 

1055: $\dot{M}_0=-10^{-9}\,M_\sun\,{\rm yr}^{-2}$ centered at the 

1056: periastron of the orbit under the assumption that 

1057: the donor is rotating synchronously with 

1058: the orbital angular velocity at periastron, and the accretor is a 

1059: $1.44\,M_\sun$ neutron star.  The contribution to the orbital 

1060: evolution timescales due to tides is determined under the 

1061: assumption that the donor is zero-age main-sequence star.  Shown at left 

1062: (right) are the timescales 

1063: for the evolution of the semi-major axis (eccentricity) as a function 

1064: of the mass ratio, $q$, for a range of eccentricities, $e$.  Regimes 

1065: where 

1066: the timescale is negative correspond to a decrease in the semi-major 

1067: axis (eccentricity), while regimes where the timescales is positive 

1068: correspond to an increase in the semi-major axis (eccentricity).  The 

1069: discontinuity at $q\simeq0.87$ corresponds to a transition of the 

1070: dominant tidal dissipation mechanism which is different for donor 

1071: stars with convective envelopes ($M_1\lesssim1.25\,M_\sun$) than for 

1072: donor stars with radiative envelopes ($M_1 \gtrsim1.25\,M_\sun$).}

1073: \label{fig-tot_timescale}

1074: \end{figure*}

1075: 

1076: When $q \la 0.87$ and $e \ga 0.4$, the effects of tides and mass 

1077: transfer on the orbital semi-major axis are always opposed, with the 

1078: orbital expansion due to mass transfer dominating the orbital shrinkage

1079: due to tides.  In the case of the orbital eccentricity, the 

1080: increase of the eccentricity due to mass transfer dominates the 

1081: decrease due to tides for mass ratios smaller than some critical 

1082: mass ratio which depends strongly on the orbital eccentricity.  Since 

1083: the timescales of orbital evolution

1084: due to mass transfer are inversely proportional to the magnitude

1085: $\dot{M}_0$ of the mass-transfer rate at periastron, the parameter space

1086: where and the extent to which mass transfer dominates increases with the

1087: rate of mass transfer at periastron.

1088: 

1089: \begin{figure*}

1090: \plottwo{f6a.eps}{f6b.eps}

1091: \caption{Contour plots of the total orbital evolution timescales in the 

1092: $(q,e$)-plane due to the combined effects of tidal dissipation and 

1093: mass transfer for the same set of assumptions adopted in 

1094: figure~~\ref{fig-tot_timescale}.  Timescales for the evolution of the 

1095: semi-major axis are shown at left; timescales for the evolution of the 

1096: orbital eccentricity are shown at right.  The different shades of gray 

1097: designate regions of the ($q,e$) parameter space with timescales (in 

1098: Gyr) in the ranges labeled in the plots.  The thick black line designates 

1099: the transition point where the rate of change of the semi-major axis or 

1100: orbital eccentricity changes from positive (to the left of the thick 

1101: black line) to negative (to the right of the thick black line).  The 

1102: vertical white line at $q\simeq0.87$ corresponds to a transition of the 

1103: dominant tidal dissipation mechanism which is different for donor 

1104: stars with convective envelopes ($M_1\lesssim1.25\,M_\sun$) than for 

1105: donor stars with radiative envelopes ($M_1 \gtrsim1.25\,M_\sun$).}

1106: \label{fig-tot_contour}

1107: \end{figure*}

1108: 

1109: In figure~\ref{fig-tot_contour}, we show the total orbital evolution

1110: timescale due to the sum of tidal and mass transfer effects as a contour

1111: plot in the $(q,e)$-plane.  The thick black lines indicate the

1112: transitions from positive (left of the thick black line) to negative

1113: (right of the thick black line) rates of change of the semi-major axis

1114: and eccentricity.  The white dividing line near $q\approx 0.87$

1115: corresponds to the transition between tidal dissipation mechanisms in

1116: stars with convective envelopes ($M_1\lesssim1.25\,{\rm M}_\sun$) and

1117: stars with radiative envelopes ($M_1\gtrsim1.25\,{\rm M}_\sun$).  It

1118: follows that there are large regions of parameter space where the

1119: combined effects of mass transfer and tidal evolution do not rapidly

1120: circularize the orbit.  In particular, for $q > 0.87$ and $e \ga 0.75$

1121: orbital circularization always takes longer than 10\,Gyr, while for $q$

1122: to the left of the thick black line the orbital eccentricity grows

1123: rather than shrinks.  For a given $q$ left of the thick black line, the

1124: timescales for eccentricity growth increase with increasing $e$ though,

1125: so that there is no runaway eccentricity growth.  Hence, for small $q$,

1126: mass transfer at the periastron of eccentric orbits may provide a means

1127: for inducing non-negligible eccentricities in low-mass binary or

1128: planetary systems.  The orbital semi-major axis, on the other hand,

1129: always increases when $e \la 0.55$, but can increase as well as decrease

1130: when $e \ga 0.55$, depending on the binary mass ratio $q$.  We recall

1131: that both the tidal and mass transfer orbital evolution time scales

1132: depend on the ratio of the donor's rotational angular velocity

1133: $\Omega_1$ to the orbital angular velocity $\Omega_{{\rm orb},P}$ and that we have

1134: set $\Omega_1/\Omega_{{\rm orb},P}=1$ in all figures shown.

1135: 

1136: \section{Concluding Remarks}

1137: 

1138: We developed a formalism to calculate the evolution of the

1139: semi-major axis and orbital eccentricity due to mass transfer in

1140: eccentric binaries, assuming conservation of total system mass and

1141: orbital angular momentum. Adopting a delta-function mass-transfer

1142: profile centered at the periastron of the binary orbit yields rates of

1143: secular change of the orbital elements that are linearly proportional to

1144: the magnitude $\dot{M}_0$ of the mass-transfer rate at the periastron.

1145: For $\dot{M}_0 = 10^{-9}\,M_\sun\,{\rm yr}^{-1}$, this yields timescales

1146: of orbital evolution ranging from a few Myr to a Hubble time or longer.

1147: Depending on the initial binary mass ratio and orbital eccentricity, the

1148: rates of secular change of the orbital semi-major axis and eccentricity

1149: can be positive as well as negative, so these orbital elements can 

1150: increase as well as decrease with time. 

1151: 

1152: Comparison of the timescales of orbital evolution due to mass transfer

1153: with the timescales of orbital evolution due to tidal dissipation shows

1154: that the effects can either be additive or competitive, depending on the

1155: binary mass ratio, the orbital eccentricity, and the magnitude of the

1156: mass-transfer rate at the periastron. Contrary to what is often assumed

1157: in even the most state-of-the-art binary evolution and population

1158: synthesis codes, tides do not always lead to rapid circularization

1159: during the early stages of mass transfer. Thus, phases of episodic mass

1160: transfer may occur at successive periastron passages and may persist for

1161: long periods of time.  As a first approximation, the evolution of the

1162: orbital semi-major axis and eccentricity due to mass transfer in

1163: eccentric binaries can be incorporated into binary evolution and

1164: population synthesis codes by means of equations~(\ref{eq-deltaa}) and

1165: (\ref{eq-deltae}) in which the mass-transfer rate is approximated by a

1166: delta-function of amplitude $\dot{M}_0$ centered at the periastron of

1167: the binary orbit.

1168: 

1169: In future papers, we will relax the assumption of conservation of total

1170: system mass and orbital angular momentum, and examine the effects of

1171: non-conservative mass transfer on the orbital elements of eccentric

1172: binaries. We also intend to study the onset of mass transfer in

1173: eccentric binaries in more detail, adopting realistic mass-transfer

1174: rates appropriate for atmospheric Roche-lobe overflow in interacting

1175: binaries as discussed by \citet{R88}.  We will consider individual 

1176: binary systems that are known to be eccentric and transferring mass 

1177: during periastron passage, as well as populations of eccentric 

1178: mass-transferring binaries and their descendants.

1179: 

1180: 

1181: %--------------------------------------------------------------------

1182: %--------------------------- BEGIN APPENDIX -------------------------

1183: %--------------------------------------------------------------------

1184: 

1185: \appendix

1186: 

1187: \section{Equipotential Surfaces and the Inner Lagrangian point in

1188:   Eccentric Binaries}

1189: \label{sec-appa}

1190: \label{appL1}

1191: %\setcounter{figure}{0}

1192: 

1193: \begin{figure}

1194: \plotone{f7.eps}

1195: \caption{Schematic diagram showing the vectors pertinent to the

1196:   derivation of the equipotential surfaces for a non-synchronous,

1197:   eccentric binary.  The center of mass of the system is shown

1198:   as a cross near the center of the diagram and the center of mass of

1199:   each star is shown as a filled circle at the star's center.  The 

1200:   vectors $\vec{R}_1$ and $\vec{R_2}$ begin at the center of mass of 

1201:   the system.}

1202: \end{figure}

1203: 

1204: A crucial element in the description of mass transfer in any binary

1205: system is the location of the inner Lagrangian point ($L_1$) through

1206: which matter flows from the donor to the accretor.  While a solution

1207: for the location of $L_1$ is not analytic, the case of circular orbits

1208: with synchronized components can be approximated by the formula

1209: \begin{equation}

1210: X_{L_1} = 0.5 + 0.22 \log{q},  \label{eq-XL1cs}

1211: \end{equation}

1212: where $X_{L_1}$ is the distance of the $L_1$ point from the mass

1213: center of the donor star in units of the distance between the stars,

1214: and  $q=M_1/M_2$ is the mass ratio of the binary defined as the ratio

1215: of the donor mass $M_1$ to the accretor mass $M_2$. For mass ratios in

1216: the range $0.1 < q < 15$, this formula is valid to an accuracy of

1217: better than 2\% \citep[e.g.][]{DR74}.  The above

1218: formula has been generalized by \citet{PS76} to

1219: include the effect of a non-synchronously rotating donor star: 

1220: \begin{equation}

1221: X_{L_1} = \left[ 0.53 - 0.03 \left( \frac{\Omega_1}{\Omega_{\rm orb}}

1222:   \right)^2 \right] \left( 1 + \frac{4}{9}\log{q} \right). 

1223:   \label{PS}

1224: \end{equation}

1225: Here, $\Omega_1$ is the rotational angular velocity of the donor, and

1226: $\Omega_{\rm orb}$ the orbital angular velocity of the binary. For $0

1227: < \Omega_1/\Omega_{\rm orb} < 1$ and $0.2 < q < 10$, the formula is

1228: valid to an accuracy of better than 3\%. It is to be noted though that

1229: \citet{PS76} derived the position of the $L_1$ point under the

1230: assumption that the orbital and rotational periods of the binary and

1231: its component stars are much longer than the dynamical timescale of

1232: the donor. Despite the better than 3\% accuracy of the fit given by

1233: equation~(\ref{PS}), the formula may therefore still break down because the

1234: underlying formalism is no longer valid.

1235: 

1236: In this paper, it is necessary to generalize the formula for the

1237: position of the $L_1$ point even further to account for a non-zero

1238: eccentricity as well as non-synchronous rotation. For this purpose, we

1239: determine the equipotential surfaces describing the shape of the

1240: components of a non-synchronous, eccentric binary. Our procedure is a

1241: generalization of the steps outlined by \citet{L63} for the

1242: derivation of the equipotentials of a non-synchronous, circular binary.

1243: 

1244: Here, we consider an eccentric binary system where the stars are

1245: considered to be centrally condensed and spherically symmetric, and

1246: thus can be well described by Roche models of masses $M_1$ and $M_2$.

1247: Their orbit is assumed to be Keplerian with semi-major axis $a$ and

1248: eccentricity $e$. Star~1 is furthermore assumed to rotate uniformly

1249: with a constant angular velocity $\vec{\Omega}_1$ parallel to the

1250: orbital angular velocity $\vec{\Omega}_{\rm orb}$. We note that for an

1251: unperturbed Keplerian orbit, the magnitude of $\vec{\Omega}_{\rm orb}$

1252: is a function of time, while the direction and orientation remain

1253: fixed in space.

1254: 

1255: As is customary, we determine the equipotential surfaces with respect

1256: to a Cartesian coordinate frame $OXYZ$ with origin $O$ at the mass

1257: center of star 1, and with the $Z$-axis pointing along

1258: $\vec{\Omega}_1$. The $X$- and $Y$-axes are co-rotating with the star

1259: at angular velocity $\Omega_1$. Since non-synchronously rotating

1260: binary components are inevitably subjected to time-dependent tides

1261: invoked by their companion, the mass elements will oscillate with

1262: frequencies determined by the difference between the rotational and

1263: orbital frequencies. Here, we neglect these tidally induced

1264: oscillations as well as any other type of bulk motion of matter due

1265: to, e.g., convection. This approximation is valid as long as the

1266: characteristic timescale associated with the motion of the mass

1267: elements is sufficiently long compared to the star's dynamical

1268: timescale (more extended discussions on the validity of the

1269: approximation can be found in \citet{L63}, \citet{S78}, and 

1270: \citet{SWK07}).  Star 1 is then completely stationary with respect to 

1271: the co-rotating frame of reference. 

1272: 

1273: With respect to the $OXYZ$ frame, the equation of

1274: motion of a mass element at the surface of star 1 is given by

1275: \begin{equation}

1276: \ddot{\vec{r}}_1 = \ddot{\vec{r}}_0-\ddot{\vec{R}}_1 - \vec{\Omega}_1

1277: \times \left( \vec{\Omega}_1 \times \vec{r}_1 \right) - 2\,

1278: \vec{\Omega}_1 \times \dot{\vec{r}}_1,  

1279: \label{eq-r1}

1280: \end{equation}

1281: where $\vec{r}_0$ and $\vec{r}_1$ are the position vectors of the mass

1282: element with respect to the mass center of the binary and 

1283: star 1, respectively, and $\vec{R}_1$ is the position vector of the center

1284: of mass of star 1 with respect to the center of mass of the binary, and  

1285: $-\vec{\Omega}_1 \times ( \vec{\Omega}_1 \times \vec{r}_1 )$ and $-

1286: 2\, \vec{\Omega}_1 \times \dot{\vec{r}}_1$ are the centrifugal and

1287: Coriolis acceleration with respect to the rotating coordinate frame. 

1288: 

1289: Under the assumption that, in an inertial frame of reference, the only

1290: forces acting on the considered mass element are those resulting from

1291: the pressure gradients in star 1 and the gravitational attractions of

1292: star 1 and its companion, the acceleration of the mass element with

1293: respect to the binary's center of mass is given by

1294: \begin{equation}

1295: \ddot{\vec{r}}_0 = -\frac{1}{\rho}\, \vec{\nabla} P - \vec{\nabla} 

1296: \left( -G\, \frac{M_1}{|\vec{r}_1|} - G\, \frac{M_2}{|\vec{r}_2|} 

1297: \right),

1298: \label{eq-r0}

1299: \end{equation}

1300: where $G$ is the Newtonian constant of gravitation, $\rho$ the mass

1301: density, $P$ the pressure, and $\vec{r}_2$ the position vector of the

1302: mass element with respect to the mass center of star 2.  The gradient in 

1303: equation~(\ref{eq-r0}) and subsequent equations in this Appendix are 

1304: taken with respect to $X$, $Y$, and $Z$.

1305: 

1306: The acceleration of the center of mass of star 1 with respect to the

1307: binary mass center, on the other hand, is given by

1308: \begin{equation}

1309: \ddot{\vec{R}}_1 = -\frac{GM_2}{D^2} \frac{\vec{R}_1}{|\vec{R}_1|}, 

1310: \label{eq-R}

1311: \end{equation}

1312: where $D(t)$ (simplified to $D$ in what follows) is the time-dependent

1313: distance between star 1 and star 2.  

1314: 

1315: Substituting equations~(\ref{eq-r0}) and~(\ref{eq-R}) into

1316: equation~(\ref{eq-r1}), we obtain 

1317: \begin{equation}

1318: \ddot{\vec{r}}_1 = -\frac{1}{\rho} \vec{\nabla} P - \vec{\nabla} \left( 

1319: -G\,

1320: \frac{M_1}{|\vec{r}_1|} - G\, \frac{M_2}{|\vec{r}_2|} \right) +

1321: \frac{GM_2}{D^2} \frac{\vec{R}_1}{|\vec{R}_1|} - \vec{\Omega}_1

1322: \times ( \vec{\Omega}_1 \times \vec{r}_1 ) - 2\, \vec{\Omega}_1 \times

1323: \dot{\vec{r}}_1.

1324: \label{eq-r1.2}

1325: \end{equation}

1326: 

1327: When the $X$-axis coincides with the line connecting the mass centers

1328: of the two stars, the third and fourth term in the right-hand member

1329: of equation~(\ref{eq-r1.2}) can be written as the gradient of a potential

1330: function as  

1331: \begin{equation}

1332: \frac{GM_2}{D^2} \frac{\vec{R}_1}{|\vec{R}_1|} = 

1333:  - \vec{\nabla} \left( \frac{GM_2}{D^2} X \right),

1334: \end{equation}

1335: \begin{equation}

1336: \vec{\Omega}_1\times ( \vec{\Omega}_1 \times \vec{r}_1 ) = -\vec{\nabla}

1337: \left[ \frac{1}{2} |\vec{\Omega}_1|^2 \left( X^2 + Y^2 \right) \right],

1338: \end{equation}

1339: where $X$ and $Y$ are the Cartesian coordinates of the mass element

1340: under consideration. With these transformations, equation~(\ref{eq-r1.2})

1341: becomes 

1342: \begin{equation}

1343: \ddot{\vec{r}}_1 = -\frac{1}{\rho} \vec{\nabla} P - \vec{\nabla} V_1 - 

1344: 2\, \vec{\Omega}_1 \times \dot{\vec{r}}_1,  \label{a8}

1345: \end{equation}

1346: where

1347: \begin{equation}

1348: V_1 = -G \frac{M_1}{|\vec{r}_1|} - G \frac{M_2}{|\vec{r}_2|} -

1349: \frac{1}{2} |\vec{\Omega}_1|^2 (X^2 + Y^2) + \frac{GM_2}{D^2} X.

1350: \end{equation}

1351: 

1352: Since star 1 is assumed to be static in the rotating frame,

1353: $\dot{\vec{r}}_1 = \ddot{\vec{r}}_1 = 0$, so that equation~(\ref{a8})

1354: reduces to

1355: \begin{equation}

1356: \vec{\nabla} P = - \rho \vec{\nabla} V_1.

1357: \end{equation}

1358: This equation governs the instantaneous shape of the equipotential 

1359: surfaces at the instant when the $X$-axis coincides with

1360: the line connecting the mass centers of the two stars. However, since

1361: this instant is entirely arbitrary, the equation is generally valid

1362: and can be used to determine the shape of the star at any phase of the

1363: orbit, with appropriate re-definition of the $X$-axis at every

1364: instant. A similar equation was derived by \citet{A76} and

1365: \citet{W79}.

1366: 

1367: Next, we look for the critical equipotential surface for

1368: which matter starts to flow from star 1 to star 2 through the inner

1369: Lagrangian point $L_1$. For circular binaries with synchronously

1370: rotating component stars, the $L_1$ point is always located on the

1371: line connecting the mass centers of the two stars. \citet{MW83} have

1372: shown that non-synchronous rotation combined with a misalignment 

1373: between the spin axis of star 1 and the orbital plane may cause the 

1374: $L_1$ point to oscillate in the $Z$-direction.  As a first 

1375: approximation, we here neglect

1376: these oscillations and assume the $L_1$ point to be located on the

1377: line connecting the mass centers of the two stars. The position of the

1378: $L_1$ point is then obtained by setting $dV_1/dX=0$. The resulting

1379: equation for the position of $L_1$ is

1380: \begin{equation}

1381: G \frac{M_1}{X^2} - G \frac{M_2}{(X-D)^2} -

1382:   |\vec{\Omega}_1|^2 X + G \frac{M_2}{D^2} = 0,

1383: \end{equation}

1384: which can be rewritten in dimensionless form as

1385: \begin{equation}

1386: \frac{q}{X_{L_1}^2} - \frac{1}{(X_{L_1}-1)^2} 

1387:   - f^2 X_{L_1} (1+q)(1+e) + 1 = 0. 

1388:   \label{XL}

1389: \end{equation}

1390: Here, $X_{L_1}=X/D$ is the position of the $L_1$ point on the $X$-axis

1391: in units of the distance between the two stars, $q=M_1/M_2$ is the

1392: binary mass ratio, $f=\Omega_1/\Omega_{{\rm orb},P}$ is the ratio of the star 1's

1393: rotational angular velocity to the orbital angular velocity at

1394: periastron, and $\Omega_{{\rm orb},P}^2=G(M_1+M_2)(1+e)/D_p^3$. In the latter

1395: expression, $D_P=a(1-e)$ is the distance between the stars at

1396: periastron. Note that in the particular case of a circular orbit,

1397: equation~(\ref{XL}) reduces to the equation for the position of the $L_1$

1398: point derived by \citet{K63} and \citet{PS76}.

1399: 

1400: Equation~(\ref{XL}) must be solved numerically for the location of

1401: $L_1$.  Since we have made no explicit assumptions about the relative

1402: location of the two star in the orbit, and we have let the distance

1403: between the stars $D=a(1-e^2)/(1+e\cos{\nu})$, with $\nu$ the true

1404: anomaly, be an explicit function of time, this equation characterizes

1405: the position of the $L_1$ point at any position in the orbit.

1406: Figure~\ref{fig-l1} shows the relative change

1407: \begin{equation}

1408: \Delta_{X_{L_1}}= \frac{X_{L_1}(\nu)-X_{L_1}(\nu=0)}{X_{L_1}(\nu=0)}

1409: \end{equation}

1410: in the position of the

1411: $L_1$ point as a function of the true anomaly for $q=1$, $f=1$,

1412: and a variety of eccentricities.  The position of the $L_1$ point over

1413: the course of a single orbit varies by less than 15\% when $e \la

1414: 0.1$, and by 30--40\% when $e \ga 0.2$. At apastron, $\Delta_{X_{L1}}$

1415: increases with increasing eccentricity when $e \la 0.5$, and decreases

1416: with increasing eccentricity when $e \ga 0.5$. This turn-around at $e

1417: \simeq 0.5$ is caused by the increasing deviations from synchronous

1418: rotation at apastron with increasing orbital eccentricity.

1419: 

1420: \begin{figure}

1421: \plotone{f8.eps}

1422: \caption{The relative change in the position of the $L_1$ point as a

1423:   function of true anomaly $\nu$, for $q=1$, $f=1$, 

1424: and a range of

1425:   eccentricities.}

1426: \label{fig-l1}

1427: \end{figure}

1428: 

1429: We have also derived a fitting formula for the position of the $L_1$

1430: point at the periastron of the binary orbit as a function of $q$, $e$,

1431: and $f$:

1432: \begin{equation}

1433: X_{L_1} = 0.529+0.231\log{q} - f^2(0.031 + 0.025 e)(1+0.4\log{q}).

1434: \label{XL1fit}

1435: \end{equation}

1436: This fitting formula is accurate to better than $4\,\%$ for mass

1437: ratios $0.1 \le q \le 10$, eccentricities $0 \le e \le 0.9$, and

1438: ratios of stellar rotational angular velocities to orbital angular

1439: velocities at periastron $0 \le f \le 1.5$. In the limiting case of a

1440: circular, synchronized binary, equation~(\ref{XL1fit}) reduces to

1441: equation~(\ref{eq-XL1cs}), while for a circular, non-synchronized binary,

1442: equation~(\ref{XL1fit}) reduces to equation~(\ref{PS}).  A detailed discussion of 

1443: the properties of all five Lagrangian points and the volume of the 

1444: critical Roche Lobe in eccentric binaries is presented in \citet{SWK07}

1445: 

1446: %******************

1447: %*** Appendix C ***

1448: %******************

1449: 

1450: \section{Orbital Evolution Due to Instantaneous Mass Transfer Between 

1451: Two Point Masses}

1452: \setcounter{figure}{0}

1453: \label{sec-appc}

1454: 

1455: When the components of a mass-transferring binary are treated as point

1456: masses and mass transfer is assumed to be instantaneous, an alternative

1457: derivation of the equations governing the secular change of the orbital

1458: semi-major axis and eccentricity may be obtained from the equations for

1459: the specific orbital energy and angular momentum.  This method is

1460: commonly used to relate the post- to pre-supernova orbital parameters of

1461: binaries in which one of the components undergoes an instantaneous

1462: supernova explosion \citep[e.g.][]{H83, BP95, K96}. As in

1463: Section~\ref{orbevtim}, we assume conservation of total system mass and

1464: describe the mass transfer by a delta function mass-transfer rate

1465: centered at the periastron of the binary orbit [see equation~(\ref{eq-del})].

1466: 

1467: The relative orbital velocity, $V_{\rm rel}$, of two stars with masses

1468: $M_1$ (the donor star) and $M_2$ (the accreting star) in a binary with

1469: orbital semi-major axis $a$ and eccentricity $e$ is given by

1470: \begin{equation}

1471: V_{\rm rel}^2 = G(M_1+M_2) \left( \frac{2}{|\vec{r}|} - \frac{1}{a} 

1472: \right),

1473: \label{eq-C-V}

1474: \end{equation}

1475: where $G$ is the Newtonian constant of gravitation, and $\vec{r}$ is the

1476: relative position vector of the accretor with respect to the donor. 

1477: Under the assumption that mass elements are transferred instantaneously

1478: between the two stars, the distance $|\vec{r}|$ is the same right before

1479: and right after mass transfer, so that the rate of change of the orbital

1480: semi-major axis due to instantaneous mass transfer is given by

1481: \begin{equation}

1482: \frac{da}{dt} = \frac{2a^2V_{\rm rel}}{G(M_1+M_2)}\frac{dV_{\rm rel}}{dt}.

1483: \label{eq-C-dadt}

1484: \end{equation}

1485: 

1486: Next, the specific orbital angular momentum of the binary written in 

1487: terms of the binary component masses and orbital elements is given by

1488: \begin{equation}

1489: |\vec{r}\times\vec{V}_{\rm rel}|^2 = G(M_1+M_2)a(1-e^2).

1490: \label{eq-C-J}

1491: \end{equation}

1492: Noting that $|\vec{r}\times\vec{V}_{\rm rel}|=|\vec{r}||\vec{V}_{\rm

1493: rel}|$ at periastron, and proceeding in a similar way as for the

1494: derivation of equation~(\ref{eq-C-dadt}), we derive the rate of change

1495: of the orbital eccentricity due to instantaneous mass transfer to be

1496: \begin{equation}

1497: \frac{de}{dt} = \frac{2a(1-e)|\vec{V}_{\rm 

1498: rel}|}{G(M_1+M_2)}\frac{d|\vec{V}_{\rm rel}|}{dt}. 

1499: \label{eq-C-dedt}

1500: \end{equation}

1501: 

1502: Finally, we eliminate $V_{\rm rel}$ from equations~(\ref{eq-C-dadt})

1503: and~(\ref{eq-C-dedt}) to obtain equations for the rates of changes of

1504: the orbital semi-major axis and eccentricity in terms of the binary

1505: orbital elements and component masses. For this purpose, we use the

1506: conservation of linear momentum to write the rate of change of the

1507: relative orbital velocity due to mass-transfer as

1508: \begin{equation}

1509: \frac{d\vec{V}_{\rm rel}}{dt} = \frac{\dot{M}_2}{M_2}\, \vec{v}_{\delta 

1510: M_2} - \frac{\dot{M}_1}{M_1}\, \vec{v}_{\delta M_1},

1511: \label{eq-C-vrelvec}

1512: \end{equation}

1513: where $\vec{v}_{\delta M_1}$ and $\vec{v}_{\delta M_2}$ are the relative

1514: velocities of the transferred mass elements with respect to the mass

1515: center of the donor and accretor, respectively.  To obtain an equation 

1516: for the change in the magnitude of $\vec{V}_{\rm rel}$, we write the 

1517: vector components of $\vec{V}_{\rm rel}$, $\vec{v}_{\delta M_1}$, and 

1518: $\vec{v}_{\delta M_2}$ with respect to the $\hat{x}$, $\hat{y}$, and 

1519: $\hat{z}$ unit vectors introduced in Section~\ref{relmot}, and take the 

1520: dot product of equation~(\ref{eq-C-vrelvec}) with $\vec{V}_{\rm rel}$. 

1521: Since, $\vec{V}_{\rm rel} = V_{\rm rel}\, \hat{y}$ at the periastron of 

1522: the binary orbit, it follows that

1523: \begin{equation}

1524: \frac{dV_{\rm rel}}{dt} = -\frac{\dot{M}_1}{M_1} \left( qv_{\delta 

1525: M_{2,y}} + v_{\delta M_{1,y}} \right).

1526: \label{eq-C-vrel}

1527: \end{equation}

1528: Assuming conservation of orbital angular momentum it then follows from

1529: equations~(\ref{eq-Jdot}) and~(\ref{eq-C-dadt})--(\ref{eq-C-vrel}) that

1530: \begin{equation}

1531: qv_{\delta M_2,y} + v_{\delta M_1,y}= na(1-q)\left(\frac{1+e}{1-e} 

1532: \right)^{1/2},

1533: \label{eq-C-vdm2}

1534: \end{equation}

1535: where we have used Kepler's 3rd law and that $|\vec{V}_{\rm rel}| = 

1536: \left(G(M_1+M_2)/a\right)^{1/2}\left((1+e)/(1-e)\right)^{1/2}$ at 

1537: the periastron of the binary orbit.  This equation is identical to 

1538: equation~(\ref{eq-vr}) in the limiting

1539: case where $|\vec{r}_{A_1}| \rightarrow 0$ and $|\vec{r}_{A_2}| 

1540: \rightarrow 0$.  Substituting equations~(eq-C-vdm2) and (eq-C-vrel) into 

1541: equations~(\ref{eq-C-dadt}) and

1542: (\ref{eq-C-dedt}) and averaging over the orbital period, we find

1543: \begin{equation}

1544: \left< \frac{da}{dt} \right>_{\rm sec} = 

1545: \frac{a}{\pi}\frac{\dot{M}_0}{M_1}\left( 1-e^2 

1546: \right)^{1/2}(q-1)

1547: \label{eq-C-dadtsec}

1548: \end{equation}

1549: \begin{equation}

1550: \left< \frac{de}{dt} \right>_{\rm sec}= 

1551: \frac{1}{\pi}\frac{\dot{M}_0}{M_1}(1-e^2)^{1/2}(1-e)(q-1).

1552: \label{eq-C-dedtsec}

1553: \end{equation}

1554: These equations are identical to equations~(\ref{eq-dadtsec})

1555: and(\ref{eq-dedtsec}) in the limiting case where

1556: $|\vec{r}_{A_1}|\rightarrow0$ and $|\vec{r}_{A_2}|\rightarrow0$.

1557: 

1558: \acknowledgements

1559: 

1560: We are grateful to an anonymous referee for insightful comments which

1561: helped to improve the paper, as well as to Ronald Taam for useful

1562: discussions, and to Richard O'Shaughnessy for discussions regarding

1563: methods used to fit the position of the $L_1$ point in eccentric,

1564: non-synchronous binaries. This work is partially supported by a NASA

1565: Graduate Fellowship (NNG04GP04H/S1) to J.S., and a NSF CAREER Award

1566: (AST-0449558), a Packard Fellowship in Science and Engineering, and a

1567: NASA ATP Award (NAG5-13236) to V.K.

1568: 

1569: %\bibliography{MT}

1570: \begin{thebibliography}{37}

1571: \expandafter\ifx\csname natexlab\endcsname\relax\def\natexlab#1{#1}\fi

1572: 

1573: \bibitem[{{Avni}(1976)}]{A76}

1574: {Avni}, Y. 1976, \apj, 209, 574

1575: 

1576: \bibitem[{{Brandt} \& {Podsiadlowski}(1995)}]{BP95}

1577: {Brandt}, N., \& {Podsiadlowski}, P. 1995, \mnras, 274, 461

1578: 

1579: \bibitem[{{Brouwer} \& {Clemence}(1961)}]{BC61}

1580: {Brouwer}, D., \& {Clemence}, G.~M. 1961, {Methods of celestial mechanics} (New

1581:   York: Academic Press, 1961)

1582: 

1583: \bibitem[{{Danby}(1962)}]{D62}

1584: {Danby}, J. 1962, {Fundamentals of celestial mechanics} (New York: Macmillan,

1585:   1962)

1586: 

1587: \bibitem[{{Drobyshevski} \& {Reznikov}(1974)}]{DR74}

1588: {Drobyshevski}, E.~M., \& {Reznikov}, B.~I. 1974, Acta Astronomica, 24, 189

1589: 

1590: \bibitem[{{Eggleton}(1983)}]{Egg83}

1591: {Eggleton}, P.~P. 1983, \apj, 268, 368

1592: 

1593: \bibitem[{{Fitzpatrick}(1970)}]{F70}

1594: {Fitzpatrick}, P.~M. 1970, {Principles of celestial mechanics} (New York,

1595:   Academic Press [1970])

1596: 

1597: \bibitem[{{Frank, King, \& Raine}(2002)}]{FKR}

1598: {Frank}, J., {King}, A., \& {Raine}, D. 2002, {Accretion Power in 

1599: Astrophysics Third Edition} (Cambridge University 

1600: Press [2002])

1601: 

1602: \bibitem[{{Hadjidemetriou}(1969{\natexlab{a}})}]{H69b}

1603: {Hadjidemetriou}, J.~D. 1969{\natexlab{a}}, \apss, 3, 330

1604: 

1605: \bibitem[{{Hadjidemetriou}(1969{\natexlab{b}})}]{H69}

1606: ---. 1969{\natexlab{b}}, \apss, 3, 31

1607: 

1608: \bibitem[{{Hills}(1983)}]{H83}

1609: {Hills}, J.~G. 1983, \apj, 267, 322

1610: 

1611: \bibitem[{{Huang}(1956)}]{H56}

1612: {Huang}, S.~S. 1956, \aj, 61, 49

1613: 

1614: \bibitem[{{Hurley} {et~al.}(2002){Hurley}, {Tout}, \& {Pols}}]{HTP02}

1615: {Hurley}, J.~R., {Tout}, C.~A., \& {Pols}, O.~R. 2002, \mnras, 329, 897

1616: 

1617: \bibitem[{{Hut}(1981)}]{H81}

1618: {Hut}, P. 1981, \aap, 99, 126

1619: 

1620: \bibitem[{{Kalogera}(1996)}]{K96}

1621: {Kalogera}, V. 1996, \apj, 471, 352

1622: 

1623: \bibitem[{{Kruszewski}(1963)}]{K63}

1624: {Kruszewski}, A. 1963, Acta Astronomica, 13, 106

1625: 

1626: \bibitem[{{Kruszewski}(1964)}]{K64-2}

1627: ---. 1964, Acta Astronomica, 14, 241

1628: 

1629: \bibitem[{{Layton} {et~al.}(1998){Layton}, {Blondin}, {Owen}, \&

1630:   {Stevens}}]{L98}

1631: {Layton}, J.~T., {Blondin}, J.~M., {Owen}, M.~P., \& {Stevens}, I.~R. 1998, New

1632:   Astronomy, 3, 111

1633: 

1634: \bibitem[{{Limber}(1963)}]{L63}

1635: {Limber}, D.~N. 1963, \apj, 138, 1112

1636: 

1637: \bibitem[{{Matese} \& {Whitmire}(1983)}]{MW83}

1638: {Matese}, J.~J., \& {Whitmire}, D.~P. 1983, \apj, 266, 776

1639: 

1640: \bibitem[{{Matese} \& {Whitmire}(1984)}]{MW84}

1641: ---. 1984, \apj, 282, 522

1642: 

1643: \bibitem[{{Meibom} \& {Mathieu}(2005)}]{MM05}

1644: {Meibom}, S., \& {Mathieu}, R.~D. 2005, \apj, 620, 970

1645: 

1646: \bibitem[{{Petrova} \& {Orlov}(1999)}]{PO99}

1647: {Petrova}, A.~V., \& {Orlov}, V.~V. 1999, \aj, 117, 587

1648: 

1649: \bibitem[{{Petterson}(1978)}]{P78}

1650: {Petterson}, J.~A. 1978, \apj, 224, 625

1651: 

1652: \bibitem[{{Piotrowski}(1964)}]{P64}

1653: {Piotrowski}, S.~L. 1964, Acta Astronomica, 14, 251

1654: 

1655: \bibitem[{{Pratt} \& {Strittmatter}(1976)}]{PS76}

1656: {Pratt}, J.~P., \& {Strittmatter}, P.~A. 1976, \apjl, 204, L29

1657: 

1658: \bibitem[{{Pribulla}(1998)}]{P98}

1659: {Pribulla}, T. 1998, Contrib. Astron. Obs. Skalnat\'e Pleso, 28, 101

1660: 

1661: \bibitem[{{Raguzova} \& {Popov}(2005)}]{Rag05}

1662: {Raguzova}, N.~V., \& {Popov}, S.~B. 2005, ArXiv Astrophysics e-prints

1663: 

1664: \bibitem[{{Reg{\"o}s} {et~al.}(2005){Reg{\"o}s}, {Bailey}, \&

1665:   {Mardling}}]{Rea05}

1666: {Reg{\"o}s}, E., {Bailey}, V.~C., \& {Mardling}, R. 2005, \mnras, 358, 544

1667: 

1668: \bibitem[{{Ritter}(1988)}]{R88}

1669: {Ritter}, H. 1988, \aap, 202, 93

1670: 

1671: \bibitem[{{Savonije}(1978)}]{S78}

1672: {Savonije}, G.~J. 1978, \aap, 62, 317

1673: 

1674: \bibitem[{{Sepinsky} {et~al.}(2007){Sepinsky}, {Willems}, \& 

1675: {Kalogera}}]{SWK07}

1676: {Sepinsky}, J.~F., {Willems}, B., \& {Kalogera}, V. 2007, \apj, in press

1677: 

1678: \bibitem[{{Soberman} {et~al.}(1997){Soberman}, {Phinney}, \& {van den 

1679: Heuvel}}]{SPH97}

1680: {Soberman}, G.~E., {Phinney}, E.~S., \& {van den Heuvel}, E.~P.~J. 1997, 

1681: \aap, 327, 620

1682: 

1683: \bibitem[{{Sterne}(1960)}]{S60}

1684: {Sterne}, T.~E. 1960, {An introduction to celestial mechanics} (Interscience

1685:   Tracts on Physics and Astronomy, New York: Interscience Publication, 1960)

1686: 

1687: \bibitem[{{Tout} {et~al.}(1996){Tout}, {Pols}, {Eggleton}, \& {Han}}]{TEA96}

1688: {Tout}, C.~A., {Pols}, O.~R., {Eggleton}, P.~P., \& {Han}, Z. 1996, \mnras,

1689:   281, 257

1690: 

1691: \bibitem[{{Willems} {et~al.}(2003){Willems}, {van Hoolst}, \&

1692:   {Smeyers}}]{WVS03}

1693: {Willems}, B., {van Hoolst}, T., \& {Smeyers}, P. 2003, \aap, 397, 973

1694: 

1695: \bibitem[{{Wilson}(1979)}]{W79}

1696: {Wilson}, R.~E. 1979, \apj, 234, 1054

1697: 

1698: \bibitem[{{Witte} \& {Savonije}(1999)}]{WS99}

1699: {Witte}, M.~G., \& {Savonije}, G.~J. 1999, \aap, 350, 129

1700: 

1701: \bibitem[{{Witte} \& {Savonije}(2001)}]{WS01}

1702: ---. 2001, \aap, 366, 840

1703: 

1704: \bibitem[{{Zahn}(1977)}]{Z77}

1705: {Zahn}, J.-P. 1977, \aap, 57, 383

1706: 

1707: \bibitem[{{Zahn}(1978)}]{Z78}

1708: {Zahn}, J.-R. 1978, \aap, 67, 162

1709: 

1710: \end{thebibliography}

1711: 

1712: \end{document}

1713: