1:
2: \documentclass[12pt,preprint]{aastex}
3:
4: %\documentclass[manuscript]{aastex}
5: %\usepackage{amssymb}
6: %\documentclass[preprint2]{aastex}
7: % \documentclass[preprint2,longabstract]{aastex}
8: % \addtolength{\hoffset}{-1.2cm}
9: % \addtolength{\textwidth}{1.8cm}
10: % \addtolength{\voffset}{-2.5cm}
11: % \addtolength{\textheight}{6.0cm}
12: %\usepackage{txfonts}
13:
14: \shorttitle{Planetesimal Accretion in Binary Systems: the Effects of
15: Gas dissipation} \shortauthors{Xie & Zhou}
16: \begin{document}
17:
18: \title{PLANETESIMAL ACCRETION IN BINARY SYSTEMS: THE EFFECTS OF GAS DISSIPATION }
19: \author{Ji-Wei Xie and Ji-Lin Zhou}
20: \affil{Department of Astronomy, Nanjing University, Nanjing 210093,
21: China} \email{xjw0809@163.com}
22:
23:
24: \begin{abstract}
25: Currently, one of major problems concerning planet formation theory
26: in close binary systems is, the strong perturbation from the
27: companion star can increase relative velocities ($\triangle V$) of
28: planetesimals around the primary and thus hinder their growth.
29: According to previous studies, while gas drag can reduce the
30: $\triangle V$ between bodies of the same sizes by forcing orbital
31: alignment to planetesimals, it increases the $\triangle V$ among
32: bodies of different sizes. In this paper, focusing on the $\gamma$
33: Cephei binary system, we propose a mechanism that can overcome this
34: difficulty. We show that in a dissipating gas disk (with a typical
35: dissipating timescale of $\sim 10^5-10^6$ years), all the
36: planetesimals eventually converge towards the same forced orbits
37: regardless of their sizes, leading to much lower impact velocities
38: among them. These $\triangle V$ decrease processes progressively
39: increase net mass accretion and even trigger runaway growth for
40: large bodies (radius $>15$ km). The effect of size distribution of
41: planetesimals is discussed, and found to be one of the dominant
42: factors that determine the outcome of collisional evolution. Anyway,
43: it can be concluded that by including the gas dissipation in the
44: early stage of disk evolution, the conditions for planetesimal
45: accretion become much better, and the process from planetesimal to
46: planet-embryo can be carried out in close binary systems like
47: $\gamma$ Cephei.
48: \end{abstract}
49:
50: \keywords{methods: numerical --- planetary systems: formation}
51:
52: \section{INTRODUCTION}
53: With the increasing number of discovered planets in binary systems
54: and the belief that a majority of solar-type stars reside in binary
55: or multiple systems, problem of planetary formation in binary
56: systems becomes a crucial one. Most of discovered planet-bearing
57: binary systems are S-type systems (e.g. $\gamma$ Cephei system, see
58: Hatzes et al. 2003) in which planets orbit the primary star with a
59: companion star surrounding them on an outer orbit. According to the
60: classical planetary formation scenario, planets form in a
61: protoplanetary disk of gas and dust orbiting a protostar. The
62: formation process is usually treated in three stages (Lissauer 1993;
63: Papaloizou \& Terquem 2006; Armitage 2007): [S1.] formation of
64: kilometer-size plantesimals ($10^{18}-10^{22}$ g) from sticking
65: collisions of dust (Weidenschilling \& Cuzzi 1993) or from
66: gravitational fragmentation of a dense particle sub-disk near the
67: midplane of the protoplanetary disk (Goldreich \& Ward 1973) on
68: timescales of the order of $ 10^4$ years, [S2.] accretion of
69: plantesimals into planetary embryos ($10^{26}-10^{27}$ g, Mercury-
70: to Mars-size) through a phase of ``runaway" and ``oligarchic" growth
71: on a timescale of the order of $10^4 - 10^5$ years, depending on
72: initial planetesimal sizes, duration of the runaway growth period,
73: possible transition to oligarchic mode (Greenberg et al. 1978;
74: Wetherill \& Stewart 1989; Barge \& Pellat 1993; Kokubo \& Ida 1996,
75: 1998, 2000; Rafikov 2003, 2004). [S3.] giant impacts between
76: embryos, producing full-size ($10^{27}$ to $10^{28}$ g) terrestrial
77: planets in about $10^7-10^8$ years (Chambers \& Wetherill 1998;
78: Kokubo, Kominami \& Ida 2006; Levison \& Agnor 2003). Here we focus
79: on the stage II to see the influence of the companion on the
80: planetesimal accretion.
81:
82: The companion star, especially when it is on a close orbit with a
83: high eccentricity, may prevent planetary formation through reducing
84: the size of the accretion disk (Artymowicz \& Lubow 1994), and
85: exciting high relative velocities between colliding planetesimals
86: (Heppenheimer 1978; Whitmire et al. 1998). The relative velocity
87: ($\triangle V$) is a critical parameter, which determines whether
88: accretion or erosion dominates. Due to the perturbation by the
89: companion, $\triangle V$ may exceed the planetesimal escape velocity
90: ($V_{esc}\sim100\times{(R_{p}/100km)}$ m s$^{-1}$), and thus inhibit
91: runaway growth. Furthermore, $\triangle V$ can even exceed the
92: threshold velocity ($V_{ero}$) for which erosion dominates
93: accretion. Here $V_{ero}$ is a few times larger than $V_{esc}$,
94: depending on the prescription on collision.
95:
96: Since planetesimals orbit the star in a sub-Keplerian gas disk
97: (Adachi et al. 1976), the presence of gas drag does not only damp
98: the companion's secular perturbation, but it also forces a strong
99: periastron alignment of planetesimal orbits. This alignment
100: significantly reduces $\triangle V$ between equal-sized bodies, favoring the accretion process
101: (Marzari \& Scholl 2000). Nevertheless, the alignment forced by the
102: gas drag induces another problem. As the alignment is
103: size-dependent, it can only reduce $\triangle V$ between
104: planetesimals of the same sizes, and at the same time it increases
105: $\triangle V$ between planetesimals of different sizes. Thebault et
106: al. (2006) find that this differential orbital alignment is very
107: efficient, leading to a significant $\triangle V$ increase for any
108: departure from the exact equal-size condition ($R_{1}$ = $R_{2}$,
109: where $R_{1}$ and $R_{2}$ are the radiuses of the two colliding
110: bodies).
111:
112: Pervious studies adopted a steady gas disk in which dissipating
113: process was neglected and the local gas density was constant. This
114: assumption, which is valid only when the planetesimal accretion time
115: scale (of the order of $10^{4}$ to $10^{5}$ years) is much shorter than the dissipating
116: time scale of local gas density, is violated under the following
117: conditions. 1)When the disk viscosity is high or photoevaporation
118: from external star exists (Hollenbach et al. 1994, 2000; Matsuyama
119: et al. 2003), disks can dissipate very fast and have short lifetimes
120: within a few $10^{5}$ years. 2)It is suggested that the assumption
121: of a single time scale for disk dissipation is not correct, and there could be a wide spread of disk lifetimes, with a large fraction of short-lived disks (Bouwman et al. 2006). As calculated by
122: Matsuyama et al.(2003) and Alexander et al. (2006b), even for a disk
123: with a lifetime of the order of $10^{6}$ years, local density can
124: decrease by as many as two orders of magnitude within the first few
125: $10^{5}$ years. For these two considerations, therefore, a model
126: that includes gas dissipation is essential for studying planetesimal
127: accretion.
128:
129: In this paper, we consider a model in which gas density
130: progressively decreases, to see how the conditions of planetesimal
131: accretion are affected by the gas dissipating process. As expected,
132: the planetesimal growth conditions change to being
133: accretion-friendly due to an dissipation induced orbital
134: convergence, which reduces $\triangle V$ between bodies of different
135: sizes. We describe our numerical model and methods in section 2. In
136: section 3, first, we simply review the planetesimal dynamics under
137: the coupled influence of secular perturbation and gas drag, and then
138: present the results. Some related and crucial issues, such as the
139: radial drift, impact rate, erosion conditions and remanent gas, are
140: discussed in section 4. Finally, in section 5, we summarize this
141: paper.
142:
143: \section{NUMERICAL MODEL AND METHODS}
144: \subsection{Gas Disk Model}
145: We made the gas model similar to that of Thebault et al. (2004).
146: Following Weidenschilling and Davis (1985), the gas drag
147: can be expressed as:
148: \begin{equation}
149: \mathbf{F}={-Kv\mathbf{v}} ,
150: \end{equation}
151: where $\mathbf{F}$ is the force per unit mass, $\mathbf{v}$ the
152: relative velocity between the planetesimal and gas, $v$ the velocity
153: modulus, and $K$ is the drag parameter defined as:
154: \begin{equation}
155: K={ {3\rho_{g}C_{d}} \over {8\rho_{p} R_{p}} },
156: \end{equation}
157:
158: \begin{equation}
159: \rho_{g}=\rho_{g0}T^{-n}, T={t\over T_{s}+1},
160: \end{equation}
161: where $\rho_{g}$ is the local gas density with an initial value of
162: $\rho_{g0}$ , $\rho_{p}$ and $R_{p}$ the planetesimal density and
163: radius, respectively. $C_{d}$ is a dimensionless coefficient related
164: to the shape of the body ($\simeq 0.4$ for spherical bodies). The
165: $T^{-n}$ function, in which time $T$ is scaled by $T_{s}$, is used
166: to include the gas dissipation, and it is based on the analytic
167: similar solutions given by Lynden-Bell and Pringle (1974). Taking
168: typical parameters from Hartmann et al. (1998), where $n=3/2$,
169: $T_{s}=10^{5}$ years, we plot figure 1 to show the gas disk density
170: evolution vs. time. The gas disk is scaled by the Minimum Mass solar
171: Nebula (hereafter MMN for short) and has the same profile to the
172: MMN(Hayashi 1981). The initial gas density is 10 MMN, and the
173: corresponding disk mass is about 100 Jupiter mass. As shown in
174: figure 1, gas density rapidly decreases from 10 MMN to 0.5 MMN
175: within the first few $10^{5}$ years, and then it experiences a slow
176: damping process lasting for a few million years. This dissipation
177: model is consistent with current theoretic calculations (Matsuyama
178: et al. 2003; Alexander et al. 2006a, 2006b) and observations (Strom
179: et al. 1993; Haisch et al. 2001; Chen \& Kamp 2004), which suggest a
180: typical disk age of 1 million years with a large scatter from $0.1$
181: to $10$ million years. Notice that the effects of binarity on the
182: dissipation of gas disk are not taken into account because details
183: of these issues are poorly known at present.
184:
185: Our model implicitly assumes an axisymmetric gas disk with constant
186: circular streamlines and follows a classical Hayashi (1981) power
187: law distribution. We are aware that this is a crude simplification
188: for modeling gas disk in close binary systems. In reality, the gas
189: disk around the primary also ``feels" the companion's perturbation,
190: under which disk structure would vary from the simplified gas model.
191: For example, the companion's perturbation can induce spiral
192: structures within the disk(Artymowicz and Lubow, 1994). To fully
193: model the behavior of planetesimals in these complex gas disks, one
194: would probably have to rely on hydro-code modeling of the gas in
195: addition to N-body type models for planetesimals. Such an
196: all-encompassing gas plus planetesimals modeling goes beyond the
197: scope of our study in this paper, and it is certainly the direction
198: of further binary disk studies. Therefore, taking a first step here,
199: we just prefer a simplified approach where gas drag force is given
200: by equation (1). As discussed by some previous studies (Scholl et
201: al. 2007, Thebault et al. 2006), this kind of simplification, on the
202: average, is reasonable at least for the dynamical evolution of
203: kilometer-size planetesimals.
204:
205: \subsection{Initial Conditions}
206: We focus on the $\gamma$ Cephei system, which is a close S-type
207: binary planetary system, hence being a good example to test the
208: influence of the companion on planetesimal accretion. Most
209: parameters adopted in this paper are listed in table 1. The initial
210: gas disk has the same profile to MMN, but is denser by 10 times. We
211: concentrate on planetesimals of four radiuses ($R_{p}=2.5, 5, 15,
212: 50$ km). As stated by Thebault et al. (2006), for impacts between
213: small bodies(1 $< R_{p} <$ 10 km), the delivered kinetic energy
214: peaks at roughly $R_{1}\simeq1/2R_{2}$, where $R_{1}$ and $R_{2}$
215: are the radiuses of the two colliding bodies. For the bigger ones,
216: this $R_{1}/ R_{2}$ ratio is somewhat smaller. Hence, the relative
217: velocity $\triangle V(2.5,5)$ between bodies of $R_{p}=2.5$ km and
218: $R_{p}=5$ km can be typical example values for small planetesimals,
219: and $\triangle V(15,50)$ for large ones. All the planetesimals
220: initially have very small inclinations based on the work of Hale
221: (1994), which suggests that approximate coplanarity between the
222: equatorial and orbital planes exist for solar-type binary systems
223: with separations less than 30-40 AU. Since it is unrealistic that
224: all planetesimals form synchronously, some earlier formed
225: planetesimals may have been pumped up to eccentric orbits while some
226: others have just formed. For this reason, the initial planetesimal
227: orbits should have random eccentricities within the range from 0 to
228: $e_{max}$, where $e_{max}$ is the maximum eccentricity that pumped
229: up by the companion. In the $\gamma$ Cephei system, $e_{max}$ is
230: about $0.1$ at $2$ AU from the primary.
231:
232: One implicit initial condition in this paper is that, of course,
233: kilometer-size plametesimals have already formed when the disk
234: begins dissipating. At present, with the poor knowledge on
235: planetesimal formation in binary systems, whether this assumption is
236: valid or not is not for sure at all. According to current limited
237: knowledge on planetesimal formation around a single star,
238: kilometer-size planetesimal can form within $10^{3} - 10^{5}$ years
239: through sticking collision or by gravitational instability after
240: dust having settled down on the midplane (Lissauer 1993;
241: Weidenschilling 1997; Goldreich \& Ward 1973; Youdin \& Shu 2002).
242: In such case , the timescale of planetesimal formation can be much
243: shorter than that of gas disk dissipation (about $10^6$ years is
244: considered in this paper), and thus it is reasonable to assume that
245: the gas dissipation starts when a population of kilometer-size
246: planetesimals exist in the system.
247:
248: \subsection{Numerical Methords}
249: We performed two kinds of runs. First, we numerically integrated the
250: equations of motion for $1000$ independent planetesimals with
251: semi-major axes from $1$ to $4$ AU. The focus is put on the
252: time-evolution of orbital eccentricities and of orbital periastrons.
253: Second, we concentrate on the time-evolution of $\triangle V$ at a
254: specific region near $2$ AU from the primary star where a planet is
255: detected. This is the configuration of the $\gamma$ Cephei system
256: that we specifically consider here. Plantesimals are initially
257: distributed in a ring near $2$ AU. Since the planetesimal sizes
258: (order of km) are very small comparing to the system typical scale
259: (order of AU), it is very difficult to track all ``real" physical
260: impacts among these planetesimals (Brahic 1977, Charnoz et al.
261: 2001a, 2001b; Lithwick \& Chiang 2007, etc). In such case, we have
262: to resort to the classical ``inflated radius" assumption, which
263: assumes an artificially increased radius to each particle (e.g.
264: Brahic 1977; Thebault \& Brahic 1998; Marzari \& Scholl 2000). For
265: planetesimals considered here, an artificially increased radius
266: (about $10^{-5}-10^{-4}$ AU) of $100$ times larger than the ``real"
267: radius is adopted for each planetesimal.
268:
269:
270: In all the runs, we used the fourth order Hermite integrator (Kokubo
271: et al. 1998), including the gas drag force and the perturbation of
272: companion. As gas drag also forces inward drift of planetesimals, we
273: adopt following boundary conditions: bodies whose semi-major axes
274: are less than $R_{in}$(greater than $R_{out}$), will be reset to
275: $R_{out}$($R_{in}$), where $R_{in}$ and $R_{out}$ are the inner and
276: outer boundaries of planetesimal belt, respectively. In these
277: resetting processes, only the semi-major axes of those bodies are
278: changed, while other orbital elements are preserved.
279:
280:
281:
282: \section{RESULTS}
283: \subsection{Planetesimal Dynamics: the Secular Approximation}
284: Before presenting the results, let's review the planetesimal
285: dynamics in a perturbed system. Heppenheimer (1978) developed a
286: simplified theory for the evolution of planetesimal eccentricity
287: with time in binary systems. First, he defined two variables $h$ and
288: $k$ as
289: \begin{equation}
290: h=e_{p} sin(\varpi), k=e_{p} cos(\varpi) ,
291: \end{equation}
292: where $e_{p}$ is the planetesimal eccentricity and $\varpi $ its
293: periastron longitude defined with respect to that of the companion
294: star ($\varpi=\varpi_{p}-\varpi_{B}$, where $\varpi_{p}$ and
295: $\varpi_{B}$ are the periastron longitudes of the companion and the
296: planetesimal, respectively. Then, introducing in the Langrange
297: planetary equations, he obtained the following equations for $h$ and
298: $k$:
299: \begin{equation}
300: {{dh}\over{dt}} =Ak - B ,
301: \end{equation}
302:
303: \begin{equation}
304: {{dk}\over{dt}} =- Ah ,
305: \end{equation}
306: where the constants $A$, $B$ are
307: \begin{equation}
308: A = {3\over4} {M_{A} \over {n(1-e_{B}^{3/2})}},
309: B = {15\over16} {ae_{B}\over{n(1-e_{B}^{5/2})}},
310: \end{equation}
311: with $e_{B}$ the eccentricity of the binary system and $M_{A}$ the
312: mass of the primary star. $a$ and $n$ are the semi-major axis and
313: mean motion of the planetesimal, respectively. The units of mass,
314: distance, and time are normalized in such a way that the
315: gravitational constant $G$ and the sum of the masses of the two
316: stars are set equal to $1$. The semimajor axis of the binary $a_{B}$
317: is chosen as the units of length, so that the time is expressed in
318: units of $(1/2\pi)T_{B}$, where $T_{B}$ is the orbital period of the
319: binary system.
320:
321: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
322: In $h-k$ plane, there is an equilibrium point (where $dh/dt=0,
323: dk/dt=0$) for equations (5) and (6), which is referred to as E0 in
324: this paper. At E0, $e_{p}=e_{f}$ and
325: $\varpi_{p}=\varpi_{f}=\varpi_{B}$, where $e_{f}=B/A$ and
326: $\varpi_{f}=\varpi_{B}$ are the forced eccentricity and periastron
327: of the planetesimal respectively. If a planetesimal reaches the
328: equilibrium point E0, its eccentricity and periastron will fix on
329: $B/A$ and $\varpi_{B}$ forever.
330:
331: To compute the effect of gas drag on the variables $h$ and $k$,
332: Marzari and Scholl (2000) modified equations (5) and (6) as
333: following:
334: \begin{equation}
335: {{dh}\over{dt}} =Ak - B - Dh(h^2 + k^2)^{1/2},
336: \end{equation}
337:
338: \begin{equation}
339: {{dk}\over{dt}} =- Ah - Dk(h^2 + k^2)^{1/2},
340: \end{equation}
341: where $D$ is a coefficient to measure the gas drag force. According
342: to these equations, for a specific $D$, the planetesimal orbit will
343: quickly or slowly (depending on the $D$ value, a larger value $D$
344: leads to a faster speed) reaches another equilibrium point(different
345: with E0), with an equilibrium eccentricity below $B/A$. Furthermore, AND THIS IS THE CRUCIAL POINT OF THIS STUDY, if $D$ damps slowly(caused by gas dissipation),
346: planetesimals will shift their orbits from this equilibrium point
347: eventually toward E0.
348:
349: In figure 2, we illustrate these processes. In no gas case, the
350: motion in the $h-k$ plane is circulating around the equilibrium
351: point $E0$ that derived from equations (5) and (6). For the case
352: with gas drag in our gas disk model, motions are divided into the
353: following two phases: a) ``no dissipation phase" in the first few
354: $10^3$ years, in which gas disk dose not significantly dissipate and
355: planetesimals of different sizes quickly reach different equilibrium
356: points depending on their sizes(point $E1$ for bodies of 50 km, $E2$
357: for 20 km, see figure 2), b) ``dissipating phase", in which gas disk
358: gradually dissipates, at the same time all the motions shift along
359: the line $E4-E0$, and eventually fix on the same equilibrium point
360: $E0$ regardless of their sizes. We also analyze the effects of
361: initial orbits on the dynamical behavior. As shown in figure 2,
362: bodies with the same sizes(5 km) but different initial orbits (one
363: is at $I1$, the other is at $I2$) go through different paths($I1-E4,
364: I2-E4$) to reach the same equilibrium point($E4$). After that, they
365: both experience the same ``dissipating phase" from $E4$ to $E0$.
366: From this point, we can see that how to choose the initial
367: planetesimal orbits do not affect the final results which are based
368: primarily on the latter ``dissipating phase".
369:
370: The appearance of the dissipating phase and the dynamical behavior
371: of planetesimal orbits during this phase are very important because
372: they provide channels to reduce the differential phasing effect
373: induced by the size-dependence of gas drag. Based on the above
374: theoretical analysis, we can expect a relative velocity($\triangle
375: V$) decrease from the convergence of all the planetesimal orbits. In
376: the next two subsections, we will numerically simulate this
377: $\triangle V$ decrease process.
378:
379:
380: \subsection{Time-evolution of Eccentricity and Periastron}
381: We first performed a simulation in which 1000 planetesimals (4
382: equal-number groups: $R_{p}=2.5, 5, 15, 50$ km, mutual interactions
383: were neglected) were initially distributed between 1 AU and 4 AU
384: from the primary. Figure 3 shows the distributions of planetesimal
385: eccentricities and periastrons vs. semi-major axes at different
386: epoches. Beyond 3 AU, the distributions of planetesimal
387: eccentricities and periastrons are random because the shorter period
388: perturbation and mean motion resonances are dominant there. Thus,
389: hereafter only planetesimals within 3 AU are discussed. In figure
390: 3a(or b), every eccentricity (or periastron) reaches an equilibrium
391: value at 5,000 years. These equilibrium values, as discussed in the
392: above subsection and also pointed out by previous studies(Thebault
393: et al. 2006), depend on the balance between the perturbation by the
394: companion and the gas drag force. Due to the size-dependence of gas
395: drag force, bodies of different sizes reach different equilibrium
396: eccentricities (or periastrons). The four lines in each panel are
397: corresponding to bodies of four kinds of sizes ($R_{p}=2.5, 5, 15,
398: 50$ km). As the gas dissipates gradually, the equilibrium
399: eccentricities (or periastrons) move to larger values, but at the
400: same time the differences among them become smaller (see Fig.~3c(or
401: d)). After a long time (5,000,000 years, see Fig.~3e(or f)), almost
402: all eccentricities (or periastrons) converge towards $e_{f}$
403: ($\varpi_{f}$).
404:
405: \subsection{Time-evolution of Relative Velocity}
406: We perform another simulation to investigate the time-evolution of
407: $\triangle V$ in a specific place(at 2 AU from the primary). In this
408: calculation, 1000 Planetesimals were initially distributed with
409: major-axes between 1.5 and 3 AU. This planetesimal ring is wide
410: enough that to trace most of collisions at 2AU.
411:
412: The results are plotted in figure 4. Figure 4b and figure 4c show
413: the average eccentricity and periastron of bodies at 2 AU as the
414: functions of time, respectively. As disk gradually dissipates, all
415: the planetesimals converge towards the same forced orbits where
416: $e_{p}=e_{f},\varpi_{p}=\varpi_{B}$(also see E0 in figure 2). Figure
417: 4a plots the $\triangle V(R_{1},R_{2}$) as the function of time. It
418: is evident that the larger differences in orbital elements, the
419: larger value of $\triangle V$. From figure 4a, it appears that the
420: $\triangle V$ between bodies of equal-size are always small because
421: of the orbital alignment. However, the $\triangle V$ between bodies
422: of different sizes first increase quickly to high values (e.g.
423: $300\sim800$ m s$^{-1}$ ), then each of them experiences a
424: relatively slow decrease. This $\triangle V$ decrease is most
425: efficient for large bodies. For 15 km-size and 50 km-size bodies,
426: the relative velocity $\triangle V(15,50)\sim300$ m s$^{-1}$ is
427: much larger than their escape velocities $V_{esc}\sim$ 50 m s$^{-1}$
428: at the beginning. After about $3\times10^{5}$ years, $\triangle
429: V(15,50)$ get lower(about 40 m s$^{-1}$) than the escape velocities
430: of the large planetesimals, so that runaway growth can occur.
431:
432: To compare with the dissipating gas drag case showed in figure 4, we
433: perform one more case with constant gas drag. It shows, in figure
434: 5, that without gas dissipation every $\triangle V$ is forced on a
435: relatively high value determined by the equilibrium between the gas
436: drag force and secular perturbation. The main difference with the
437: dissipating gas case is that there is no late stage with
438: size-independent orbital phasing and thus no $\triangle V$ decrease.
439:
440:
441: \section{DISCUSSIONS}
442:
443: \subsection{Impact Rate}
444: As impacts of different types (between the same sizes or different
445: sizes) have totally different $\triangle V$ and thus different
446: outcomes (erosion, incomplete accretion, complete accretion and
447: runaway growth), the condition that which type of collision
448: dominates becomes crucial for planetesimal growth. Figure 6 plots
449: the distributions of impact rates for two cases: a)standard case,
450: b)random case. In both cases, we compute 1000 planetesomals whose
451: radius distribution is assumed as a gaussian function centered at 8
452: km with a dispersion $\triangle R=7$ km. The only difference between
453: them is the companion and gas drag are not included in the random
454: case. As shown in figure 6, for the random case, the distribution of
455: impact rates depends only on the initial size distribution: impacts
456: occur more often in the places where more planetesimals are
457: distributed for impacts between equal-sized bodies close to the center of the Gaussian. On the other hand, in the standard case, the
458: distribution is obviously size-dependent: impacts mainly occur
459: between bodies of different sizes. By comparing these two cases, it
460: is clear: under the coupled effect between gas drag and the
461: companion's perturbation, impacts between bodies of different sizes
462: are favored, while impacts between bodies of the same(or similar)
463: sizes are hindered. This result can be understood in this way: for
464: bodies of the same sizes, as they have the same forced orbits and
465: radial drifts, one can only collide with another when their
466: semimajors are very close; for bodies of different sizes, in
467: contrast, as they have different forced orbits and radial drifts,
468: one can cross many more planetesimal orbits on a much larger region.
469:
470: \subsection{Accretion or Erosion}
471: The key result of this paper presented in section 3 is: as gas
472: dissipates, all planetesimals eventually converge towards the same
473: forced orbits regardless of their sizes, leading to much lower
474: $\triangle V$ than in the constant-gas density case. To further see
475: the effects of these $\triangle V$ decreasing processes on
476: planetesimal collisional evolution(accretion or erosion), we then
477: perform a quantitative study.
478:
479: Following Kortenkamp \& Wetherill (2000), we adopt the disruption
480: limit given by Love and Ahrens (1996), and compute the net mass
481: accretion ratio (see Appendix for details) for every impact. Figure
482: 7 shows the time-evolution of net mass accretion ratios($A_{r}$) for
483: impacts between different size groups. For impacts between bodies of
484: the same sizes, net mass accretion ratios are not plotted, since the
485: $\triangle V$ are always low enough for runaway growth in such
486: cases. As shown in figure 7, it can be summarized as following: 1)
487: for small bodies ($R_{p}<5$ km), collisions always lead to erosion
488: during the first $7\times 10^5$ years, after which accretion occurs
489: with a progressively increasing $A_{r}$, 2) for intermediate
490: bodies($5<R_{p}<15$ km), $A_{r}$ is initially modest(75\%-80\%) and
491: will progressively increase (to 90\%-95\%) as the gas dissipates, 3)
492: for large bodies ($R_{p}>15$ km ), $A_{r}$ is always very
493: high($\geq95\%$), 4) for impact between a large($R_{p}>15$) km and a
494: small($R_{p}<5$ km) planetesimal, while the $\triangle V$ is high
495: and decreases slowly(see figure 4), $A_{r}$ is always high ($\geq
496: 95\%$). Therefore, to fully know the details of collisions among a
497: swarm of planetesimals will have to require an entire information of
498: the initial planetesimal size distribution, which is, however, not
499: clear at all with current knowledge.
500:
501: Here, for simplicity, we just perform four simplified tests assuming
502: for the planetesimal size distribution a gaussian and three power-law
503: functions, respectively. For the three power law cases, planetesimals have
504: distributions given by $N\propto m^{-1.7}$ (Makino et al. 1998) with
505: three radius ranges, namely 1 - 50 km, 2.5 - 50 km and 5 - 50 km. For the gaussian case, the radius distribution is assumed as the gaussian function centered at
506: 8 km with a dispersion $\triangle R=7$ km. Figure 8 plots the
507: time-evolution of the average $\triangle V$ and $A_{r}$ for these
508: four cases.
509: It shows, at the first few $10^3$ years(no dissipation
510: phase), the conditions for accretion or erosion totally depend on
511: the initial size distribution of planetesimals. In this phase, the
512: average $\triangle V$ is pumped up by the size-dependence of orbital
513: alignment, and thus the accretion is inefficient($A_{r}\sim75\%$)
514: for one power law case(5 - 50 km), dangerous ($A_{r}\sim30\%$) for
515: the gaussian case and another power law case(2.5 - 50 km), and even completely suppressed for the power law case(1 - 50 km) . However, after a few $10^5$ years(gas dissipation phase), all the $\triangle V$ get low enough and accretion is
516: efficient($A_{r}\geq 95\%$) for all the cases, regardless of the
517: initial size distribution of planetesimals. Notice that the smaller bodies we consider initially, the more time the system needs to become accretion-friendly. For the power law case with minimum size of 1 km, it indeed takes about $6\times10^5$ years before accretion is efficient. As discussed in the next subsection, this long timespan can worsen the radial drift problem.
518:
519: \subsection{Radial Drift}
520: Moving in the gas disk, planetesimals undergo a headwind by which
521: they are forced to progressively migrate inwards (Adachi et al.
522: 1976). In the above runs, we adopt a boundary condition described in
523: section 2.3 to keep all the planetesimals staying in our computing
524: zone ($1.5 - 3$ AU). This is reasonable only if the planetesimal
525: disk is extended enough so that planetesimals can flow into the
526: computing zone from the outer disk. However, theoretical
527: calculations of binary-disk interactions predict that companions
528: might truncate circumstellar disks at an out radius of $0.2 - 0.5$
529: times the binary semi-major axes (Artymowicz \& Lubow 1994). For
530: $\gamma$ Cephei system, $a_{B}=18.5$ AU, then the truncated disk
531: size is about $3.7-9.3$ AU. Therefore, there may be not enough
532: material supplied from the outer disk, and it means there should be
533: enough planetesimals staying in the computing zone for at least a
534: few $10^{5}$ years to form planets. For this reason, we performed a
535: simulation without any boundary condition to compare the results in
536: figure 4. We find most large bodies with $R_{p}=15$ km and
537: $R_{p}=50$ km stay in the computing zone, having $\triangle V$
538: curves similar to those in figure 4, while almost all the small
539: bodies with sizes of $R_{p}=2.5$ km and $R_{p}=5$ km are removed by
540: gas drag. This problem of ``too fast migration" will be even worse
541: when smaller bodies are considered, such as bodies with radiuses of
542: $1-10$ m. As shown in figure 8, for the power law case(1 - 50 km), there is $6\times10^5$ years timespan, during which erosion dominates and thus planetesimals are transformed into small fragments which are quickly removed by inward drift.
543:
544: Actually, fast inward drift induced by gas drag is a general problem in the classical
545: planet formation model(Lissauer 1993; Papaloizou \& Terquem 2006;
546: Armitage 2007), and several ways have been proposed to address this
547: issue. It is possible that large planetesimals ($R_{p}>10$ km, which is big enough to overcome the inward dirft) form directly via
548: gravitational instability in a few $10^3$ years (Goldreich \& Ward
549: 1973; Youdin \& Shu 2002). In addition, radial drift may allow small
550: bodies to pileup within the inner disk to form larger planetesimals(
551: Youdin \& Chiang 2004), and the present of turbulence in gas disk
552: can also reduce the radial drift(Durisen et al. 2005; Haghighipour
553: \& Boss 2003; Rice et al. 2004).
554:
555:
556: \subsection{Remanent Gas for gaseous Planet Formation}
557: In addition, there should be enough remanent gas to form a massive
558: gaseous planet, as required to fit the minimum mass ($\sim2$ jupiter
559: masses) of the planet detected in the $\gamma$ Cephei system. In
560: this paper, for a initial gas disk of 10 MMN(about 100 Jupiter
561: mass), after $5 \times 10^{5}$ years when most $\triangle V$ have
562: already decreased to low enough values, the remanent gas, according
563: to figure 1, is about 7 Jupiter masses. On the other hand, Kley and
564: Nelson (2007) suggest that the gas accretion onto a planet will be
565: highly efficient in the $\gamma$ Cephei system due to the large
566: induced planet orbital eccentricity. Their simulations indicate that
567: it needs a gas disk with only $\sim 3$ Jupiter masses to form a
568: gaseous planet of $\sim 2$ Jupiter masses. Therefore, it is possible
569: to form a massive gaseous planet in our dissipating gas model.
570:
571: \section{SUMMARY}
572: In this paper, focusing on the $\gamma$ Cephei system and
573: concentrating on planetesimal impact velocities($\triangle V$), we
574: numerically investigate the conditions for planetesimal accretion in
575: binary systems. We extend the studies of Thebault et al. (2004,
576: 2006) by including the effect of a dissipating gas disk. We confirm
577: some of their results that in a gas disk without dissipation,
578: differential orbital alignment is very efficient and increase
579: $\triangle V$ between bodies of different sizes to high values that
580: significantly inhabit planetesimal growth. Furthermore, we find that
581: by including gas dissipation, the differential phasing effect
582: induced by the size-dependence of gas drag can be reduced. In such
583: case, as gas density decreases, all planetesimals converge their
584: orbits towards the same forced orbits, regardless of their sizes.
585: This orbital convergence induced by gas dissipation is most
586: efficient for large bodies(15 - 50 km). Within $3\times10^5$ years,
587: $\triangle V(15,50)$ decrease to low enough values(about 40 m
588: s$^{-1}$ below the escape velocities of large bodies) for which
589: runaway growth is able to occur.
590:
591: In order to get more information of the collisional evolution, we
592: first discuss the impact rate distribution. We find, for binary
593: systems including gas drag, collisions between bodies of different
594: sizes are dominant due to the differential orbital alignment and the
595: size-dependence of the radial drift.
596: Considering this result, our mechanism which can reduce the
597: $\triangle V$ between bodies of different sizes, therefore, becomes
598: much more essential for planetesimal growth.
599:
600: By defining the net mass accretion ratio($A_{r}$), we then discuss
601: the conditions of accretion or erosion for a swam of planetesimals
602: with different size distributions. We find the size distribution is
603: a very crucial factor that influences the collisional evolution. For
604: the constant gas density case, it totally dominates the growth of
605: planetesimals, and accretions are only efficient between equal-sized bodies in such
606: case. On the other hand, for the dissipating gas density case,
607: effect of size distribution is dominant only at the beginning, and
608: after a few $10^5$ years, accretion( or even runaway growth) is
609: always favored, regardless of the initial size distribution of
610: planetesimals.
611:
612: Due to the companion's perturbation in a binary system, disk is
613: truncated to a smaller one and the planetesimals undergo a much
614: faster inward drift. These effects may induce a problem that whether
615: enough planetesiamls can remain in the planet-formation zone against
616: the inward migration. We perform some computations for this
617: consideration, and find most small bodies($R_{p}<$ 10 km) are
618: removed within a few $10^5$ years, while no significant influences
619: on large bodies($R_{p}>$ 15 km). Furthermore, the inward drift problem will be much more acute when the initial planetesimal population is composed mainly of small bodies($R_{p}<$ 2.5 km). In such case, erosion dominates for the first few $10^5$ years, and planetesimals are transformed into small fragments which are quickly removed by inward drift.
620:
621: Finally, we estimate the remanent gas for forming a gaseous planet.
622: In our dissipating gas disk model, after $5 \times 10^{5}$ when
623: $\triangle V$ among most of planetesimals have already decreased to low
624: enough values, the disk mass is about 7 Jupiter mass which is enough
625: to form a massive gaseous planet.
626:
627:
628: \acknowledgments
629: We thank the anonymous referee for valuable suggestions, and
630: W. Kley for useful discussions. This work is supported by
631: NSFC(10778603), National Basic Research Program of China(2007CB4800).
632:
633: \appendix
634: \section{APPENDIX}
635: \subsection{Net Mass Accretion Ratio} For the sake of simplicity,
636: colliding planetesimals, both the target and the projectile are
637: normally considered as nearly homogeneous and spherical bodies, and
638: all the collisions are treated as central impacts. Having these
639: assumptions, to describe a specific collision needs only three input
640: parameters: mass of target($M_{t}$), mass of projectile($M_{P}$),
641: and impact velocity ($V_{imp}$, namely the $\triangle V$ derived
642: from our simulations).
643:
644: Given a target and projectile of mass and radius $M_{t}$,$R_{t}$ and
645: $M_{p}$, $R_{p}$ respectively, the surface escape velocity of the
646: pair is
647: \begin{equation}
648: V_{esc}^2={2G(M_{t}+M_{p})\over(R_{t}+R_{p})},
649: \end{equation}
650: Where $G$ is the constant of gravity. The center of mass impact
651: energy available for fragmentation is given by
652: \begin{equation}
653: Q_{f}={k_{1}\over2}V_{imp}^2M_{t}M_{p}/(M_{t}+M_{p}),
654: \end{equation}
655: where the impact efficiency $k_{1}=0.5$ is the fraction of the
656: impact energy not lost to heating. Assuming the crushing strength
657: scaled by Love and Ahren (1996)
658: \begin{equation}
659: Q_{c}=24.2[R_{t}(cm)]^{1.13},
660: \end{equation}
661: where $R_{t}$ is the radius of the target in cm, then the mass of
662: material fragmented by the impact is
663: \begin{equation}
664: M_{f}=Q_{f}/Q_{c}.
665: \end{equation}
666:
667: As some fragments fall back on the target by the gravity, the mass
668: of material to escape is only a fraction of $M_{f}$ and given by
669: \begin{equation}
670: M_{e}=k_{2}M_{f}V_{esc}^{-2.25}
671: \end{equation}
672: (Greenberg et al. 1978), where $k_{2}=3\times10^6$ (cm
673: s$^{-1})^{2.25}$.
674: Here in this paper, we define a ratio as being
675: \begin{equation}
676: A_{r}=1-M_{e}/M_{p},
677: \end{equation}
678: to measure the fraction of mass accreted on
679: the target. If the derived $M_{e}\ge M_{p}$ there is no growth of the target, and $A_{r}=0$ is forced in such cases.
680: Figure 9 maps the $A_{r}$ in the $R_{1}$-$R_{2}$ plane with four
681: typical impact velocities: 100m/s, 300m/s, 600m/s, 1000m/s. As shown
682: in figure 9, bodies with radius below 5 km hardly accrete each
683: other, on the other hand, once one of the two colliding bodies has
684: radius larger than 15 km, accretion is always efficient.
685:
686:
687: \begin{thebibliography}{99}
688: \bibitem[\protect\citeauthoryear{Adachi}{1976}]{b1}
689: Adachi, I., Hayashi, C., \& Nakazawa, K. 1976, Prog. Theor. Phys.,
690: 56, 1756
691:
692: \bibitem[\protect\citeauthoryear{Alexander}{2006a}]{b2}
693: Alexander, R. D., Clarke, C. J., Pringle, J. E. 2006a, MNRAS, 369,
694: 216
695:
696: \bibitem[\protect\citeauthoryear{Alexander}{2006b}]{b3}
697: Alexander, R. D., Clarke, C. J., Pringle, J. E. 2006b, MNRAS, 369,
698: 229
699:
700: \bibitem[\protect\citeauthoryear{Armitage}{2007}]{b4}
701: Armitage, P. J. 2007, Lecture note on the formation and evolution of
702: planetary systems, Preprint (astro-ph/0701485)
703:
704: \bibitem[\protect\citeauthoryear{Artymowicz}{1994}]{b5}
705: Artymowicz, P. \& Lubow, S.H. 1994, AJ 421, 651
706:
707:
708: \bibitem[\protect\citeauthoryear{Barge}{1993}]{b6}
709: Barge, P., \& Pellat, R. 1993, Icarus, 104, 79
710:
711: \bibitem[\protect\citeauthoryear{Bouwman}{2006}]{b7}
712: Bouwman, J., Lawson, W. A., Dominik, C., Feigelson, E. D., Henning,
713: Th., Tielens, A. G. G. M., Waters,L. B. F. M. 2006, ApJ, 653, L57
714:
715: \bibitem[\protect\citeauthoryear{Brahic}{1977}]{b8}
716: Brahic, A. 1977, A\&A, 54, 895
717:
718: \bibitem[\protect\citeauthoryear{Chambers}{1998}]{b9}
719: Chambers, J. E., \& Wetherill, G. W. 1998, Icarus, 136, 304
720:
721: \bibitem[\protect\citeauthoryear{Charnoz}{2001a}]{b10}
722: Charnoz, S., \& Brahic, A. 2001b, A\&A, 375, L31
723:
724: \bibitem[\protect\citeauthoryear{Charnoz}{2001b}]{b11}
725: Charnoz, S., Thebault, P., \& Brahic, A. 2001a, A\&A, 373, 683
726:
727: \bibitem[\protect\citeauthoryear{Chen}{2004}]{b12}
728: Chen, C. H., \& Kamp, I. 2004, ApJ 602, 985
729:
730: \bibitem[\protect\citeauthoryear{Durisen}{2005}]{b13}
731: Durisen, R. H., Cai, K., Mejia, A. C., \& Pickett, M. K. 2005,
732: Icarus, 173, 417
733:
734: \bibitem[\protect\citeauthoryear{Goldreich}{1973}]{b14}
735: Goldreich, P., \& Ward, W. R. 1973, ApJ, 183, 1051
736:
737: \bibitem[\protect\citeauthoryear{Greenberg}{1978}]{b15}
738: Greenberg, R., Hartmann, W. K., Chapman, C. R., Wacker, J. F. 1978,
739: Icarus, 35, 1
740:
741: \bibitem[\protect\citeauthoryear{Haghighipour}{2003}]{b16}
742: Haghighipour, N., \& Boss, A. P. 2003, ApJ, 598, 1301
743:
744: \bibitem[\protect\citeauthoryear{Haisch}{2001}]{b17}
745: Haisch, K. E., Lada, E. A., \& Lada, C. J. 2001, ApJ 553, L153
746:
747: \bibitem[\protect\citeauthoryear{Hale}{1994}]{b18}
748: Hale, A. 1994, AJ, 107, 306
749:
750: \bibitem[\protect\citeauthoryear{Hartmann}{1998}]{b19}
751: Hartmann, L., Calvet, N., Gullbring, E., D'Alessio, P. 1998, ApJ,
752: 495, 385
753:
754: \bibitem[\protect\citeauthoryear{Hatzes}{2003}]{b20}
755: Hatzes, A.P., Cochran, W.D., Endl, M., McArthur, B., Paulson, D.,
756: Walker, G.A.H., Campbell, B., \& Yang, S. 2003, ApJ, 599, 1383
757:
758: \bibitem[\protect\citeauthoryear{Hayashi}{2003}]{b21}
759: Hayashi, C. 1981, PthPS, 70, 35
760:
761: \bibitem[\protect\citeauthoryear{Heppenheimer}{1978}]{b22}
762: Heppenheimer, T. A. 1978, A\&A, 65, 421
763:
764: \bibitem[\protect\citeauthoryear{Hollenbach}{1994}]{b23}
765: Hollenbach, D., Johnstone, D., Lizano, S., Shu, F. 1994, ApJ, 428,
766: 654
767:
768: \bibitem[\protect\citeauthoryear{Hollenbach}{2000}]{b24}
769: Hollenbach, D., Yorke, H.W., \& Johnstone, D. 2000, in Protostars
770: and Planets IV, ed. V. Mannings, A. P. Boss, \& S. S. Russell
771: (Tucson: Univ. Arizona Press), 401
772:
773: \bibitem[\protect\citeauthoryear{Kley}{2007}]{b25}
774: Kley, W., \& Nelson, R. 2007, On the Formation and Dynamical
775: Evolution of Planets in Binaries, preprint (astro-ph/07053421)
776:
777: \bibitem[\protect\citeauthoryear{Kokubo}{1996}]{b26}
778: Kokubo, E. \& Ida, S. 1996, Icarus, 123, 180
779:
780: \bibitem[\protect\citeauthoryear{Kokubo}{1998}]{b27}
781: Kokubo, E. \& Ida, S. 1998, Icarus, 131, 171
782:
783: \bibitem[\protect\citeauthoryear{Kokubo}{2000}]{b28}
784: Kokubo, E. \& Ida, S. 2000, Icarus, 143, 15
785:
786: \bibitem[\protect\citeauthoryear{Kokubo}{2006}]{b29}
787: Kokubo, E., Kominami, J., Ida, S. 2006, ApJ, 642, 1131
788:
789: \bibitem[\protect\citeauthoryear{Kokubo et al.}{1998}]{b30}
790: Kokubo, E., Yoshinaga K., \& Makino, J. 1998, MNRAS, 297, 1067
791:
792: \bibitem[\protect\citeauthoryear{Kortenkamp}{2000}]{b31}
793: Kortenkamp, S. J., Wetherill, G. W. 2000, Icarus, 143, 60
794:
795: \bibitem[\protect\citeauthoryear{Levison}{2003}]{b32}
796: Levison, H. F., \& Agnor, C. 2003, AJ, 125, 2692
797:
798: \bibitem[\protect\citeauthoryear{Lissauer}{1993}]{b33}
799: Lissauer, J. J. 1993, ARA\&A, 31, 129
800:
801: \bibitem[\protect\citeauthoryear{Lithwick}{2007}]{b34}
802: Lithwick, Y., Chiang, E. 2007, ApJ, 656, 524
803:
804: \bibitem[\protect\citeauthoryear{Love}{1996}]{b35}
805: Love, S. G., \& Ahrens, T. J. 1996, Icarus, 124, 141
806:
807: \bibitem[\protect\citeauthoryear{Lynden-Bell}{1974}]{b36}
808: Lynden-Bell, D., \& Pringle, J. E. 1974, MNRAS, 168, 603
809:
810: \bibitem[\protect\citeauthoryear{Makino}{1998}]{b37}
811: Makino, J., Fukushige, T., Funato, Y., Kokubo, E. 1998, New
812: Astronomy, 3, 411
813:
814: \bibitem[\protect\citeauthoryear{Marzari}{2000}]{b38}
815: Marzari, F. \& Scholl, H. 2000, ApJ, 543, 328
816:
817: \bibitem[\protect\citeauthoryear{Matsuyama}{2003}]{b39}
818: Matsuyama, I., Johnstone, D., \& Hartmann, L. 2003, ApJ, 582, 893
819:
820: \bibitem[\protect\citeauthoryear{Papaloizou}{2006}]{b40}
821: Papaloizou, J. C. B., Terquem, C. 2006, Rep. Prog. Phys., 69, 119
822:
823: \bibitem[\protect\citeauthoryear{Rafikov}{2003}]{b41}
824: Rafikov, R. R. 2003, AJ, 125, 942
825:
826: \bibitem[\protect\citeauthoryear{Rafikov}{2004}]{b42}
827: Rafikov, R. R. 2004, AJ, 128, 1348
828:
829: \bibitem[\protect\citeauthoryear{Rice}{2004}]{b43}
830: Rice, W. K. M., Lodato, G., Pringle, J. E., Armitage, P. J., \&
831: Bonnell, I. A. 2004, MNRAS, 355, 543
832:
833:
834: \bibitem[\protect\citeauthoryear{Marzari}{2000}]{b44}
835: Scholl, H., Marzari, F., Th¨¦bault, P. 2007, MNRAS, 380, 1119
836:
837: \bibitem[\protect\citeauthoryear{strom}{1993}]{b45}
838: Strom, S. E., Edwards, S., \& Skrutskie, M. F. 1993, in Protostars
839: and Planets III, ed. E. H. Levy \& J. I. Lunine (Tucson: Univ.
840: Arizona Press), 837
841:
842: \bibitem[\protect\citeauthoryear{Thebault}{1998}]{b46}
843: Thebault, P., Brahic, A. 1998, P\&SS, 47, 233
844:
845: \bibitem[\protect\citeauthoryear{Thebault}{2004}]{b47}
846: Thebault, P., Marzari, F., Scholl, H., Turrini,D., \& Barbieri,M.
847: 2004, A\&A, 427, 1097
848:
849: \bibitem[\protect\citeauthoryear{Thebault}{2006}]{b48}
850: Thebault, P., Marzari, F., \& Scholl, H. 2006, Icarus, 183, 193
851:
852: \bibitem[\protect\citeauthoryear{Weidenschilling}{1993}]{b49}
853: Weidenschilling, S. J., \& Cuzzi, J. N. 1993, in Protostars and
854: Planets III, ed. E. H. Levy \& J. I. Lunine (Tucson: Univ. Arizona
855: Press), 1031
856:
857: \bibitem[\protect\citeauthoryear{Weidenschilling}{1997}]{b50}
858: Weidenschilling, S. J. 1997, ASPC, 122, 281
859:
860: \bibitem[\protect\citeauthoryear{Weidenschilling}{1985}]{b51}
861: Weidenschilling, S. J. \& Davis, D. R. 1985, Icarus, 62, 16
862:
863: \bibitem[\protect\citeauthoryear{Wetherill}{1989}]{b52}
864: Wetherill, G. W., Stewart, G. R. 1989, Icarus, 77, 330
865:
866: \bibitem[\protect\citeauthoryear{Whitmire}{1998}]{b53}
867: Whitmire, D. P., Matese, J. J. \& Criswell L. 1998, Icarus, 132, 196
868:
869: \bibitem[\protect\citeauthoryear{Youdin \& Chiang}{2004}]{b54}
870: Youdin, A. N., \& Chiang, E. 2004, ApJ, 601, 1109
871:
872: \bibitem[\protect\citeauthoryear{Youdin \& Shu}{2002}]{b55}
873: Youdin, A. N., \& Shu, F. 2002, ApJ, 580, 494
874:
875: \end{thebibliography}
876:
877: \clearpage
878: \input{efigscaptions}
879: \clearpage
880: \input{tab1}
881:
882: \end{document}
883: