1: %\documentclass[12pt,preprint]{aastex}
2: \documentclass{emulateapj}
3: \usepackage{apjfonts}
4: \def\msun{{\rm ~M}_{\odot}}
5: \def\rsun{{\rm ~R}_{\odot}}
6: \def\myr{{\rm ~Myr}}
7: \def\mdot{\dot M}
8: \def\mpy{{\rm ~M}_{\odot} {\rm ~yr}^{-1}}
9: \def\medd{{\rm M_{Edd}}}
10: \def\ledd{{\rm L_{Edd}}}
11: \def\racc{{R_{\rm acc}}}
12: \def\rcap{{R_{\rm cap}}}
13: \def\vrel{{v_{\rm rel}}}
14: \def\cs{{c_{\rm s}}}
15: \def\nasr#1{{\sl New \ Astr. Rev.,} { #1}}
16: \def\lapprox{\;\rlap{\lower 2.5pt % approximately smaller
17: \hbox{$\sim$}}\raise 1.5pt\hbox{$<$}\;}
18: %\usepackage{graphicx}
19: %\usepackage{natbib}
20: %\usepackage[usenames,dvips]{color}
21: \begin{document}
22:
23: \title{The Interaction of Stellar Objects within a Common Envelope}
24:
25: \author{Paul M. Ricker\altaffilmark{1}
26: and Ronald E. Taam\altaffilmark{2}}
27:
28: \altaffiltext{1}{Department of Astronomy, University of Illinois, 1002 West Green Street, Urbana, IL 61801; pmricker@uiuc.edu.}
29: \altaffiltext{2}{Department of Physics and Astronomy, Northwestern University, 2131 Tech Drive, Evanston, IL 60208; r-taam@northwestern.edu.}
30:
31: \begin{abstract}
32: We use high-resolution, three-dimensional hydrodynamic simulations to study
33: the hydrodynamic and gravitational interaction between stellar companions
34: embedded within a differentially rotating common envelope. We evaluate the
35: contributions of the nonaxisymmetric gravitational tides and ram pressure forces
36: to the drag force and, hence, to the dissipation rate and the mass accumulated
37: onto the stellar companion. We find that the gravitational drag dominates the
38: hydrodynamic drag during the inspiral phase, implying that a
39: simple prescription based on a gravitational capture radius
40: significantly underestimates the dissipation rate and overestimates the inspiral
41: decay timescale. Although the mass accretion rate fluctuates
42: significantly, we observe a secular trend leading to an effective rate
43: that is significantly less than the rate based on a gravitational capture
44: radius. We discuss the implications of these results within the context of
45: accretion by compact objects in the common-envelope phase.
46: \end{abstract}
47:
48: \keywords{binaries: close -- hydrodynamics -- stars: evolution}
49:
50: \section{Introduction}
51: To understand the evolution of binary star systems, it is essential to analyze
52: the interactions between their stellar components. Examples of
53: such influences include the spin-orbit tidal interaction and mass transfer,
54: as well as interactions that result in the loss of mass and angular momentum.
55: Equally important are the interactions of stars orbiting about
56: their common center of mass within a differentially rotating common envelope.
57: It is generally accepted that such an evolutionary stage is essential for the
58: formation of short-period binary systems containing compact objects (see, e.g.,
59: Iben \& Livio 1993 and Taam \& Sandquist 2000). In this case, the interaction
60: determines the orbital evolution of the system and the conditions under which
61: the common envelope is ejected, leading to the survival of a remnant binary system
62: or to a merger that forms a rapidly rotating single star. The amount of mass and
63: angular momentum accreted by the inspiralling components during this
64: phase also has direct implications for the properties of
65: the compact object population in binary systems.
66:
67: Lacking multidimensional hydrodynamical simulations of
68: the common-envelope phase, the initial numerical and semi-analytical
69: studies of the problem used simple prescriptions for the stellar interactions
70: based on the pioneering work by Hoyle \& Lyttleton (1939) and Bondi \& Hoyle (1944), as
71: generalized by Bondi (1952). These seminal studies focused on the idealized
72: problem of the capture of matter by a gravitating point object moving supersonically
73: with respect to a uniform medium. In this framework, a gravitational capture radius,
74: $\rcap$, plays an important role in determining the rates of mass accretion and
75: energy dissipation. $\rcap$ is given by
76: \begin{equation}
77: \label{Eqn:rcap} \rcap = {2 G M \over \vrel^2 + \cs^2}\ ,
78: \end{equation}
79: where $M$ is the mass of the gravitating object, $\vrel$ is the velocity of the
80: object with respect to the medium, and $\cs$ is
81: the local speed of sound. When a density gradient with scale height $H$ is
82: present, the effective accretion radius $\racc$ is (Dodd \& McCrea 1952)
83: \begin{equation}
84: \label{Eqn:racc} \racc = {\rcap \over 1 + (\rcap/2H)^2}\ .
85: \end{equation}
86: The energy dissipation rate is then $L_{\rm d} \approx \pi
87: \racc^2 \rho \vrel^3$, where $\rho$ is the upstream density. To improve on
88: these estimates,
89: hydrodynamic effects were approximated analytically by Ruderman \& Spiegel
90: (1971), Wolfson (1977), and Bisnovatyi-Kogan et al.\ (1979) as well as
91: numerically by Hunt (1971, 1979), Shara \& Shaviv (1980), and
92: Shima et al.\ (1985). These early multidimensional simulations considered
93: axisymmetric flow, and their results have been used to calibrate the energy
94: loss rate. In particular, the drag coefficients obtained from such simulations
95: (see, e.g., Shima et al.\ 1985) have been used to estimate the rate of energy
96: dissipation in the common envelope.
97:
98: However, many of the simplifying assumptions underlying these studies are
99: inadequate for direct application to common-envelope interactions.
100: The flow is nonaxisymmetric and distinctly nonuniform, reflecting the
101: existence of velocity or density gradients (the density may span several scale
102: heights within $\racc$). The effect of relaxing these assumptions
103: has been studied in two dimensions by Fryxell \& Taam (1988) and Taam \&
104: Fryxell (1989) and in three dimensions by Sawada et al.\ (1989) and Ruffert
105: (1999). These studies could not encompass the full complexity of
106: common-envelope interactions, since the envelope's self-gravity was ignored.
107: Furthermore, because the companions move in elliptical
108: orbits, their cores interact with matter that has already been affected
109: in previous orbital phases. Thus, the state of the gas and its environment
110: in these calculations must be regarded as highly idealized.
111:
112: Within the past decade, three-dimensional numerical studies of the common-envelope
113: phase that have relaxed earlier geometrical assumptions have been carried
114: out by Sandquist et al.\ (1998, 2000), DeMarco et al.\ (2003a,b), and Taam \&
115: Ricker (2006). Recently, we have carried out high-resolution adaptive mesh
116: refinement (AMR) simulations of common-envelope evolution with effective
117: resolutions of $2048^3$ (R. E. Taam \& P. M. Ricker, in preparation), allowing the interaction of the
118: stars within the common envelope to be examined and quantified.
119:
120: In this Letter we report on some results of our numerical studies. We focus on
121: analyzing a single high-resolution simulation to determine the hydrodynamic and
122: gravitational contributions to the drag forces affecting the orbital motion of
123: the stellar components during the early inspiral phase. We also analyze the
124: accumulation of mass by the stellar components within the common envelope to
125: compare their magnitudes to estimates based on an accretion radius
126: formalism. In \S~2, we briefly describe the numerical method and our assumed
127: model for the binary system. Descriptions of the method of analysis and the
128: numerical results are presented in \S~3. Finally, we summarize our results and
129: comment on their possible implications for applications involving compact
130: objects in short-period binary systems.
131:
132: \section{Numerical Method}
133: For our simulations we use FLASH, a parallel AMR code that
134: includes both grid- and particle-based numerical methods (Fryxell et al.\ 2000).
135: The Euler equations describing the evolution of the common-envelope gas are solved
136: using the piecewise parabolic method (Colella \& Woodward 1984). We use AMR
137: to refine the mesh on the basis of the second derivative of gas density, for shock capturing,
138: and on the basis of the positions of the stellar components, for high force resolution in their
139: vicinities. This method significantly improves on stationary nested-grid techniques
140: because the refined region can change to track the motion of both stars, thus
141: allowing investigations of systems with components of similar mass.
142:
143: However, even with AMR the core of an evolved red giant star (the progenitor
144: of a white dwarf) is too compact compared to the size of the computational domain
145: (a factor of 40,000 smaller) given the large number of timesteps we require. Since
146: the core has a much higher mean density than does the common envelope, its strongest
147: interaction with the gas is gravitational. Therefore, we model the red giant core
148: and the companion object using spherical particle clouds containing $2\times10^5$
149: particles each, obtaining forces on the clouds using cloud-in-cell (CIC) interpolation.
150: (We use clouds rather than single particles to avoid problems related to CIC force
151: anisotropy.) The cloud radii (ie., $R \sim R_\odot$) are taken to be 3 times the
152: smallest zone spacing. The particles in each cloud move rigidly together with the
153: cloud's center of mass. More details are given by Taam
154: \& Ricker (2006).
155:
156: We considered the common-envelope evolution of a binary system consisting of a $1.05
157: \msun$ red giant having a $0.36 \msun$ core and a $0.6 \msun$ companion with an orbital
158: period of 44.2 days. The initial red giant model was taken from a one-dimensional stellar
159: evolution code developed by Eggleton (1971, 1972). We interpolated this model onto a
160: three-dimensional Cartesian grid with nine levels of refinement. Since each block contained $8^3$ zones, and since the coarsest level contained one block, this corresponded to a minimum zone
161: spacing of $2\times10^{10}$~cm (a factor of 2048 smaller than the domain size). We
162: artificially damped transient motions in the red giant envelope, evolving this model for
163: a dynamical timescale. We then turned off the damping and added the particle cloud representing
164: the companion star. At this point, the red giant core and envelope, as well as the companion,
165: were given circular orbital velocities and a spin angular velocity equal to 95\% of
166: the synchronous value.
167:
168: \section{Numerical Results}
169: After an initial phase of 12~days, the orbital separation decreased monotonically
170: with time, with an increasing rate of orbital shrinkage. During the later phase
171: ($\sim$ 43~days), the orbital separation decreased by a factor of 3 from $4.3 \times
172: 10^{12}$ to $1.4 \times 10^{12}$~cm, as a result of the strong drag forces
173: on the cores. In the physical system, the drag has both hydrodynamic and
174: gravitational components, and, in our simulation, the gravitational component dominates.
175:
176: This can be seen by defining a measured hydrodynamic drag force $F_{\rm d}$
177: on each core via
178: \begin{equation}
179: \label{Eqn:flux}{\bf F_{\rm d}} = \int \rho {\bf v_{rel}} ({\bf v_{rel}} \cdot {\bf n}) dA\ ,
180: \end{equation}
181: where the integration is taken over a closed spherical surface centered on the core and
182: {\bf n} is the unit vector perpendicular to its surface. While the particle clouds
183: have definite radii, these are determined by numerical constraints and do not correspond to
184: the physical radii of the objects that the clouds represent. Moreover, the physical radii
185: can be much smaller than the finest
186: zone spacing, depending on the type of companion star. Hence, we studied the effect of
187: choosing different (resolvable) radii for the spherical surfaces used to measure
188: the drag, ranging from $3.5\times10^{10}$ to $1.0 \times10^{12}$~cm.
189: We found that the hydrodynamic drag on the companion reaches a characteristic
190: early inspiral value (at 34 days) of $\sim 9\times 10^{32}$~dynes at a radial scale
191: of $3\times 10^{11}$~cm from its center, only $\sim 10$\% of the gravitational drag
192: on this object. At this radial scale, the hydrodynamic drag on the red giant core
193: is at most $4\times10^{32}$~dynes, about 1/4 of the gravitational drag on it.
194: In general, the total
195: gravitational drag is one to two orders of magnitude larger than the hydrodynamical
196: drag. The drag force causes the orbital energy to dissipate at a rate given by
197: the line integral of the drag force along the orbital path; in the case presented here,
198: the gravitational drag corresponded to a dissipation rate $L_{\rm d} \sim 8 \times
199: 10^{39}$ ergs s$^{-1}$.
200:
201: \begin{figure}
202: \plotone{f1.eps}
203: \caption{\label{Fig:flow}
204: Flow field near the companion
205: at $39.27$~days.
206: Colors denote the logarithm of gas density in the $x-y$ plane, in units of g~cm$^{-3}$.
207: Arrows indicate the gas velocity in every fourth zone. The arrow in the lower
208: right-hand corner corresponds to 106~km~s$^{-1}$. The blue arrow at the center
209: indicates the companion's velocity. The red giant core lies just
210: off the bottom of the plot. The inset shows the region immediately surrounding the
211: companion, enlarged 3$\times$ per dimension. The vector scale is the same as in
212: the main image.}
213: \end{figure}
214:
215: The flow near the inspiraling companion leads to the accretion of both mass
216: and angular momentum. The flow pattern and density distribution in this region are
217: illustrated in Figure~\ref{Fig:flow}. The density increases as one approaches the embedded object,
218: revealing a high density contrast, with scale length corresponding to $\sim 10^{11}$ cm,
219: which is the direct result of the companion's gravitational influence in the common envelope. In
220: the immediate region surrounding the embedded companion,
221: radiation pressure is comparable to gas pressure. Matter is shown to be significantly
222: deflected by the companion. The velocity drops off more gradually than the density,
223: with a scale length $\sim 5$ times larger. In this region, typical
224: velocities are of the order of 150 km s$^{-1}$, whereas the velocity of the companion is of the
225: order of
226: $\sim 50$ km s$^{-1}$. The gas flows are subsonic, and there is no evidence for the presence
227: of shocks.
228:
229: We note that our calculation does not include an explicit model for the embedded boundary
230: condition imposed by the surface of the companion star. A realistic inner boundary
231: condition would fall somewhere between two extreme types: a fully absorbing boundary,
232: corresponding to accretion onto a black hole, and a fully reflecting boundary,
233: perhaps corresponding to a star with a surface that retains very little accreted
234: material. In the latter case we would expect to see a clear accretion shock. Because
235: we allow gas to accumulate in the companion's vicinity,
236: it heats up due to compression and creates a back pressure that resists further accretion.
237: Thus, our calculation is intermediate between the two cases, and the accretion rate
238: that we measure should be regarded as only indicative.
239:
240: Since the particle cloud does not provide a boundary for the gas on the
241: grid, we computed the mass flux (using the respective equivalents to
242: Eq.~\ref{Eqn:flux}) for a range of assumed radii. (The accreted angular momentum
243: rates could not be reliably calculated close to the companion due to insufficient
244: numerical resolution. The length scale over which the pressure gradients would
245: be calculated is comparable to the size of the control region.) We found that
246: the mass accretion rate fluctuates significantly on timescales of less than a day.
247: Thus, we integrated this rate to determine the accumulation
248: of mass with time. In Figure~\ref{Fig:macc}, the accumulated mass is illustrated as a function of
249: time for different control radii, measured from the companion's position, ranging from
250: $3.5 \times 10^{10}$ to $2.1 \times 10^{11}$ cm. A secular trend is evident,
251: although variations are found to be larger at greater distances from the companion. The
252: numerical results show that the magnitude of the accreted mass is a function of the radius
253: of the control surface, with lower accreted mass associated with smaller control
254: surfaces. In particular, only about 1/3 of the matter at the outermost surface
255: reached the particle cloud. It appears that the differences in the mass accumulation
256: decrease with time at smaller control surfaces, whereas they increase with time at larger
257: radii, suggesting that estimates for the mass accretion rate should be based on
258: evaluations at small radii. After an evolution of 39 days,
259: the companion object would have accreted $\sim 5 \times 10^{-4}
260: \msun$, leading to an effective mass
261: accretion rate of about $0.005 \mpy$ (assuming that the angular momentum does not
262: significantly impede the accretion process). During the first 35 days of evolution,
263: instantaneous mass accretion rates rarely
264: exceed $0.05 \mpy$, and then they quickly rise to $\sim 0.1 \mpy$.
265:
266: Equations (\ref{Eqn:rcap}) and (\ref{Eqn:racc}) can be used to infer an expected
267: accretion rate $\dot{M}_{\rm BH} = \pi \racc^2 \rho \vrel$
268: on the basis of the gravitational capture radius formalism.
269: Using the companion mass for $M$ and taking $\rho$, $\vrel$, $\cs$, and $H$ as
270: observed in the simulation, we find $\racc \approx 8.0 \times 10^{10}$~cm and
271: $\dot{M}_{\rm BH} \approx 3.2 \mpy$. This estimate is significantly
272: larger than the instantaneous or average mass accretion rate determined from the
273: simulation. Although our lack of an embedded boundary condition implies that our
274: measured accretion rates are only indicative, we note that the rate expected in the
275: gravitational capture radius formalism would allow for the accretion of the
276: entire red giant envelope onto the companion within less than 80 days. Thus,
277: this formalism must significantly overestimate the real
278: accretion rate. The physical reason for this discrepancy is evident in
279: Figure~\ref{Fig:flow}. Whereas the capture radius formalism assumes a supersonic
280: wind flowing directly past the accretor, the common-envelope system develops a
281: relatively slow-moving region near the companion, and matter accretes subsonically.
282: Most of the gas is accelerated by the companion's potential and flung out to
283: large radius, where it becomes unavailable for later accretion due to its
284: angular momentum. In response, the companion spirals inward toward the red giant.
285:
286: \begin{figure}
287: \epsscale{0.9}
288: \plotone{f2.eps}
289: \caption{\label{Fig:macc} Variation of mass, in units of solar mass, accumulated within a range of radii for the
290: control surface centered on the embedded companion as a function of time in days.}
291: \end{figure}
292:
293: \section{Discussion}
294:
295: We have quantitatively described the interaction of stars within a common envelope
296: on the basis of the analysis of the early inspiral stage of a $1.05 \msun$ red giant
297: and a $0.6 \msun$ binary companion. The orbital decay is dominated by the nonaxisymmetric
298: gravitational drag associated with the self-gravitating matter in the common envelope.
299: On the basis of high-resolution three-dimensional hydrodynamic simulations, this drag is 1-2
300: orders of magnitude greater than the hydrodynamic drag. As a consequence, the
301: orbital decay timescale is much shorter than that derived from analyses based on the
302: Hoyle-Lyttleton-Bondi picture of accretion from a uniform medium by a gravitating point
303: mass moving at supersonic speeds. In this latter description the gravitational drag
304: associated with an accretion wake is not significantly larger than the hydrodynamic
305: drag. The effect of long-range gravitational interactions is critical for reliable estimates
306: of the orbital decay timescale and energy dissipation rate. In this picture the drag and the
307: mass accretion rate are not as directly related as they are in the Hoyle-Lyttleton-Bondi-type
308: description, because the dominant drag term is gravitational, rather than hydrodynamic, in
309: origin. This decoupling is reflected in the much larger ratio $L_{\rm d} / \dot{M} \sim
310: 10^{15}-10^{16}$~cm$^2$~s$^{-2}$ measured in the simulation than expected from the
311: gravitational capture radius formalism ($\sim 10^{14}$~cm$^2$~s$^{-2}$).
312:
313: Although the mass accretion rates estimated from our simulation should be regarded
314: as only indicative due to the lack of a detailed inner boundary treatment for the
315: companion, the structure of the flow (dominated as it is by tidal effects) strongly
316: suggests that the true mass accretion rate should be much smaller than the rate
317: expected in the gravitational capture radius formalism. The effective capture radius,
318: on the basis of the observed ambient density and relative velocity, is almost an order of
319: magnitude smaller than the expected value. In any case, the expected value would
320: lead to an unrealistic level of accretion over the common-envelope
321: period. Furthermore, the inspiral time suggested by the early evolution
322: of our three-dimensional simulation is much shorter than that found in earlier
323: one- and two-dimensional calculations that are based on the Bondi-Hoyle-Lyttleton picture.
324: Assuming that the discrepancy in mass
325: accretion rate continues into the deep inspiral phase, the total accumulation of
326: matter onto the companion should be much smaller than previously expected.
327:
328: These results would have little effect on the mass of an embedded main-sequence star
329: because of its tendency to expand as a result of the high entropy within the common
330: envelope (see Hjellming \& Taam 1991); however, it can significantly affect the outcome
331: for neutron stars within a common envelope. In particular, the
332: estimated accretion rates exceed $10^{-3} \mpy$, for which steady state accretion flows
333: with neutrino losses are possible (Chevalier
334: 1989; Houck \& Chevalier 1991). At these hypercritical mass accretion rates, photons
335: are trapped in the flow and the Eddington limit is not applicable. On the basis of this
336: hypercritical accretion flow regime, Chevalier (1993) and Brown (1995) suggest that
337: neutron stars embedded in the common envelope would accrete sufficient mass to form
338: low-mass black holes (although see Chevalier 1996), and,
339: hence, the formation of binary radio pulsars would require an
340: evolutionary scenario involving progenitor stars of nearly equal mass (Brown 1995).
341: However, the population synthesis of binary black holes and neutron stars by Belczynski
342: et al.\ (2002), including hypercritical accretion, resulted in an average accretion of $0.4
343: \msun$. Such a high rate of mass accretion is inconsistent with the observed masses of binary
344: radio pulsars ($\sim 1.35 \msun$; Thorsett \& Chakrabarty 1999) and indicates the need for
345: a reduction in accreted matter during the common-envelope phase (see also Belczynski et al.\ 2007). Our calculations show that the necessary reduction may arise naturally as a
346: result of a more realistic treatment of the common-envelope phase. Consequently, this
347: reduction could also lead to a reduction in the number of low-mass
348: black holes, depending on the maximum mass of neutron stars, resulting from the accretion
349: induced collapse of massive accreting neutron stars in the common-envelope phase. Similarly,
350: the mass accretion, which was found to be as large as several solar masses for black hole
351: accretors, would also be reduced, thereby affecting the masses and spins of double black holes
352: emerging from the common-envelope phase.
353:
354: Further investigations are planned to examine the generality of these results regarding
355: mass accretion and to quantify the importance of these processes for determining the
356: properties (mass and spin) and ultimate fate of the compact components in short-period binary
357: system populations. Such studies are not only important for determining the masses of binary
358: neutron star and black hole systems resulting from the common-envelope phase (Belczynski et al.\ 2007), but also their orbital periods, which directly influence the expected merger rates of
359: such binary populations as sources for gravitational wave detection in the advanced LIGO
360: experiment.
361:
362: \acknowledgments
363: P.M.R. acknowledges helpful conversations with Charles Gammie and Stu Shapiro.
364: This work was partially supported by the National Center for Supercomputing
365: Applications under allocation AST040024.
366: Partial support has also been provided by the NSF through grants
367: AST-0200876 and AST-0703950. FLASH was developed largely by the
368: DOE-supported ASC/Alliances Center for Astrophysical Thermonuclear Flashes at
369: the University of Chicago.
370:
371: \begin{thebibliography}{}
372: \bibitem[]{bkb02} Belczynski, K., Kalogera, V., \& Bulik, T.\ 2002, \apj, 572, 407
373: \bibitem[]{btkrb07} Belczynski, K., Taam, R. E., Kalogera, V., Rasio, F. A., \& Bulik, T. 2007,
374: \apj, 662, 504
375: \bibitem[]{bkkkl79} Bisnovatky-Kogan, G. S., Kazhdan, Ya., M., Klypin, A. A., Lutskii, A. E.,
376: \& Shakura, N. I. 1979, Soviet Astr., 23, 201
377: \bibitem[]{b52} Bondi, H. 1952, \mnras, 112, 195
378: \bibitem[]{bh44} Bondi, H. \& Hoyle, F. 1944, \mnras, 104, 273
379: \bibitem[]{b95} Brown, G. E. 1995, \apj, 440, 270
380: \bibitem[]{c89} Chevalier, R. 1989, \apj, 346, 847
381: \bibitem[]{c93} Chevalier, R. 1993, \apj, 411, L33
382: \bibitem[]{c96} Chevalier, R. 1996, \apj, 459, 322
383: \bibitem[]{cw84} Colella, P., \& Woodward, P. R. 1984, {\sl J. Comp. Phys.}, 54, 174
384: \bibitem[]{dsmht03a} De Marco, O., Sandquist, E. L., Mac Low, M., Herwig, F., \& Taam, R. E.
385: 2003a, RevMexAA, 15, 34-37
386: \bibitem[]{dsmht03b} De Marco, O., Sandquist, E. L., Mac Low, M., Herwig, F., \& Taam, R. E.
387: 2003b, RevMexAA, 18, 24-30
388: \bibitem[]{dm52} Dodd, K. N., \& McCrea, W. J. 1952, \mnras, 112, 205
389: \bibitem[]{e71} Eggleton, P. P. 1971, \mnras, 151, 351
390: \bibitem[]{e72} Eggleton, P. P. 1972, \mnras, 156, 361
391: \bibitem[]{f00} Fryxell, B., Olson, K., Ricker, P., et al.\ 2000, \apjs, 131, 273
392: \bibitem[]{f88} Fryxell, B. A., \& Taam, R. E. 1988, \apj, 335, 862
393: \bibitem[]{hc91} Houck, J. C., \& Chevalier, R. 1991, \apj, 376, 234
394: \bibitem[]{hl39} Hoyle, F. \& Lyttleton, R. A. 1939, Proc. Cambridge Phil. Soc., 35, 405
395: \bibitem[]{h71} Hunt, R. 1971, \mnras, 154, 141
396: \bibitem[]{h79} Hunt, R. 1979, \mnras, 188, 83
397: \bibitem[]{ht91} Hjellming, M. S., \& Taam, R. E. 1991, \apj, 370, 709
398: \bibitem[]{il93} Iben, I., \& Livio, M. 1993, \pasp, 105, 1373
399: \bibitem[]{rs71} Ruderman, M. A., \& Spiegel, E. A. 1971, \apj, 165, 1
400: \bibitem[]{r99} Ruffert, M.\ 1999, \aap, 346, 861
401: \bibitem[]{stb00} Sandquist, E., Taam, R. E., \& Burkert, A. 2000, \apj, 533, 984
402: \bibitem[]{stcb98} Sandquist, E., Taam, R. E., Chen, X., Burkert, A., \& Bodenheimer, P.
403: 1998, \apj, 500, 909
404: \bibitem[]{smabl89} Sawada, K., Matsuda, T., Anzer, U., B\"orner, G., \& Livio, M. 1989,
405: \aap, 221, 263
406: \bibitem[]{ss80} Shara, M. M., \& Shaviv, G. 1980, \apss, 67, 427
407: \bibitem[]{smts85} Shima, E., Matsuda, T., Takeda, H., \& Sawada, K. 1985,
408: \mnras, 217, 367
409: \bibitem[]{tf89} Taam, R. E., \& Fryxell, B. A., 1989, \apj, 339, 297
410: \bibitem[]{tr07} Taam, R. E., \& Ricker, P. M. 2006, \nasr in press (astro-ph/0611043)
411: \bibitem[]{ts00} Taam, R. E., \& Sandquist, E. L. 2000, \araa, 38, 113
412: \bibitem[]{tc99} Thorsett, S.E., \& Chakrabarty, D. 1999, \apj, 512, 288
413: \bibitem[]{w77} Wolfson, R. 1977, \apj, 213, 200
414: \end{thebibliography}
415:
416: \end{document}
417: