1: %\documentclass[12pt,preprint]{aastex}
2:
3: %% manuscript produces a one-column, double-spaced document:
4:
5: %\documentclass[manuscript]{aastex}
6:
7: %% preprint2 produces a double-column, single-spaced document:
8:
9: \documentclass[preprint2]{aastex}
10:
11: %% If you want to create your own macros, you can do so
12: %% using \newcommand. Your macros should appear before
13: %% the \begin{document} command.
14: %%
15: %% If you are submitting to a journal that translates manuscripts
16: %% into SGML, you need to follow certain guidelines when preparing
17: %% your macros. See the AASTeX v5.x Author Guide
18: %% for information.
19:
20: \newcommand{\vdag}{(v)^\dagger}
21: \newcommand{\var}{\Omega}
22: \newcommand{\tha}{\theta}
23: \newcommand{\alp}{\alpha}
24: \newcommand{\beq}{\begin{equation}}
25: \newcommand{\eeq}{\end{equation}}
26: \newcommand{\bea}{\begin{array}}
27: \newcommand{\eea}{\end{array}}
28: \newcommand{\dw}{\Delta \Omega}
29:
30: %% You can insert a short comment on the title page using the command below.
31:
32: \slugcomment{Not to appear in Nonlearned J., 45.}
33:
34: %% If you wish, you may supply running head information, although
35: %% this information may be modified by the editorial offices.
36: %% The left head contains a list of authors,
37: %% usually a maximum of three (otherwise use et al.). The right
38: %% head is a modified title of up to roughly 44 characters. Running heads
39: %% will not print in the manuscript style.
40:
41: \shorttitle{Accretion of Heavy Elements } \shortauthors{Zhou et
42: al.}
43:
44: %% This is the end of the preamble. Indicate the beginning of the
45: %% paper itself with \begin{document}.
46:
47: \begin{document}
48:
49:
50: \title{Planetesimal Accretion onto Growing Proto-Gas-Giant Planets}
51:
52:
53: \author{Ji-Lin Zhou$^1$, Douglas N.C. Lin$^{2,3}$}
54: \affil{$^1$Department of Astronomy, Nanjing University, Nanjing
55: 210093, China; zhoujl@nju.edu.cn}
56:
57: \affil{$^2$UCO/Lick Observatory, University of California, Santa
58: Cruz, CA 95064, USA; lin@ucolick.org}
59: \affil{$^3$KIAA, Peking University,
60: Beijing 100871, China}
61:
62: \begin{abstract}
63:
64: The solar and extra solar gas giants appear to have diverse internal
65: structure and metallicities. We examine a potential cause for these
66: dispersions in the context of the conventional sequential accretion
67: formation scenario, which assumes the initial formation of cores
68: from protoplanetary embryos. In principle, gas accretion onto cores
69: with masses below several times that of the Earth is suppressed by
70: the energy released from the bombardment of residual planetesimals.
71: After the cores have attained their isolation masses, additional
72: mass gain through gas accretion enlarges their feeding zones and
73: brings a fresh supply of planetesimals. However, the relatively
74: low-mass cores have limited influence on exciting the eccentricities
75: of the newly embraced planetesimals. Due to their aerodynamical and
76: tidal interaction with the nascent gas disk, planetesimals on
77: eccentric orbits undergo slow orbital decay. We show that these
78: planetesimals generally cannot pass through the mean motion
79: resonances of the cores,
80: and the suppression of planetesimal bombardment
81: rate enables the cores to accrete gas with little interruption.
82: During
83: growth from the cores to protoplanets,
84: the resonance width of protoplanets increases with their masses.
85: When the resonances overlap with each other, the trapped planetesimals
86: become dynamically unstable and their eccentricity excitation is
87: strongly enhanced. Subsequent gas drag induces the planetesimals
88: to migrate to the proximity of the protoplanets and collide
89: with them. This process leads to the resumption and
90: a surge of planetesimal bombardment
91: during the advanced stage of the protoplanet growth. The most
92: massive intruders are the residual earth-mass protoplanetary
93: embryos that may survive the passage through the protoplanet
94: envelopes and increase their core masses. This mechanism may
95: account for the diversity of the core-envelope
96: structure between Jupiter, Saturn and the metallicity dispersion
97: inferred from the transiting extra solar planets. During the
98: final formation stage of the proto-gas-giants, their tidal torque
99: induces the formation of gaps in the gas disk. This perturbed
100: structure of gas disk also leads to the accumulation of planetesimals
101: outside the feeding zone of the protoplanets. The surface density
102: enhancement promotes the subsequent buildup of cores for
103: secondary gas giant planets outside the orbit of the first-born
104: protoplanets and the formation of eccentric multiple planet systems.
105: \end{abstract}
106:
107: \keywords{Giant planet; planetary formation; solar nebula; planet
108: dynamics; n-body simulation }
109:
110:
111: \section{Introduction}
112:
113: Models of the interior structure of giant planets in our Solar System
114: suggest that giant planets contain heavy elements with total masses up to
115: several tens Earth masses ($M_\oplus$) (e.g., Wuchterl et al. 2000).
116: Due to uncertainties in the equation of state,
117: the internal distribution of the heavy elements is poorly determined.
118: Guillot, Gautier \& Hubbard (1997) constructed models of Jupiter which
119: indicate the presence of a core with mass in the range $0-12
120: M_\oplus$ and a total mass of heavy elements reaching $11-45
121: M_\oplus$. With a slight modification in the equation of state used in
122: these models, Saumon \& Guillot (2004) deduced a smaller upper limit
123: for the core mass of Jupiter. They find a Jupiter model with core mass of
124: $0-11M_\oplus$ and total heavy elements of $8-39 M_\oplus$. Their models
125: suggest Saturn may have a core with mass $9-22M_{\oplus}$ and a total
126: heavy-element mass of $13-28M_{\oplus}$.
127:
128: Super-solar metallicity and massive core have also been inferred from
129: the transit observation of a compact Saturn-mass extra solar planet,
130: HD149026b (Sato et al. 2005). However, transit observations of
131: several other close-in extra solar planets indicate a large dispersion
132: in the planetary mass-size distribution. Since the contraction rate
133: during the evolution of gas giant planets is determined by the ratios of
134: their cores to envelope masses (Bodenheimer et al. 2001,
135: Burrows et al. 2000), a large spread in their mass-radius
136: relation is indicative of a wide dispersion in their internal
137: structure.
138:
139: The existence of massive cores in Saturn and HD149026b provides a
140: strong support for the conventional sequential accretion scenario
141: (Perri \& Cameron 1974, Mizuno 1980, Bodenheimer \& Pollack 1986,
142: Pollack et al. 1996), which is based on the assumption that gas
143: giant planets form through three major stages:
144:
145: \noindent S1) Embryo-growth stage: protoplanetary cores formed and grew mainly by
146: the bombardment of planetesimals onto them. Although low-mass cores (up
147: to a few $M_\oplus$) can also attract a small amount of gas,
148: the envelopes initially built up at a much slower rate than the
149: growth of the core. The cores attain isolation masses at the end
150: of this initial stage.
151:
152: \noindent S2) Quasi-hydrostatic sedimentation stage: the
153: accretion of planetesimals tapers as their supply in the feeding
154: zone is depleted by the cores. As the dissipation rate also
155: decreases along with the declining influx of planetesimal
156: bombardment, thermal energy continues to diffuse out of the
157: envelope. This loss of entropy allows a quasi-hydrostatic
158: sedimentation and the growth of the gaseous envelope.
159:
160: \noindent S3) Runaway gas-accretion stage: when the mass of the
161: gas becomes comparable to
162: that of the core, the rate
163: of gas sedimentation increases with the intensified flux of radiation
164: transfer. The characteristic growth time scale of the protoplanets decreases
165: with their masses. This runaway stage continues until the gas supply
166: is exhausted by either the formation of a tidally induced gap near the
167: protoplanet orbit or the depletion of the entire nascent disk.
168:
169: While this paradigm has been widely accepted, many uncertainties
170: remain. One major issue concerns the protracted transition through
171: the second stage (S2). In the early models (Pollack et al. 1996),
172: this stage persists over a time scale longer than the observationally
173: inferred depletion time scale of the disk ($\sim$ a few Myr), because any
174: increase in the protoplanet mass also leads to an expansion of its
175: feeding zone and a resurgence of planetesimal accretion, which tends to
176: slow down the gas accretion. A simple extrapolation of the early
177: models would imply a low probability of gas-giant formation, which is
178: incompatible with the observationally inferred ubiquity (with a probability $\sim
179: 0.1-0.15$) of gas giant planets around nearby solar-type stars (Marcy
180: 2005).
181:
182: Another unsolved issue is the compositional diversity. Regardless of
183: the large dispersion in the core mass, the average metallicity of
184: Jupiter is about twice solar while that of Saturn is an order of
185: magnitude larger than that of the Sun. There are several potential
186: mechanisms which may lead to these differences:
187:
188: \noindent
189: M1) The critical mass of the cores ($M_{cr}$) needed for the onset of
190: efficient gas accretion ({\i.e.} from quasi-hydrostatic sedimentation
191: (S2) to runaway gas-accretion stage (S3)) depends on their
192: planetesimal bombardment rate ($\dot M_p$) and the rate of radiation
193: transfer (Ikoma et al. 2000). Both processes
194: may be stochastic and dependent on the inventory of residual
195: planetesimals and the dust-to-gas ratio.
196:
197: \noindent
198: M2) During the runaway gas-accretion stage (S3), a fraction of the pre-existing
199: cores may be eroded and mixed into the envelope.
200:
201: \noindent
202: M3) The metallicity of the accreted gas in the runaway gas-accretion stage (S3)
203: may depend on the
204: epoch of protoplanet formation, since its value increases during the
205: transition from protostellar to debris disks.
206:
207: \noindent
208: M4) During and after the runaway gas-accretion stage (S3),
209: gaseous planets may also gain mass in
210: heavy elements through giant impacts by nearby residual planetesimals
211: and protoplanetary embryos. Most intruders are disrupted during their
212: passage through the envelope. However, colliding embryos with several
213: $M_\oplus$ may be able to reach the cores and increase their masses.
214:
215: Here, we explore how several relevant physical processes may act
216: together to overcome the growth challenge of gas giants and introduce
217: metallicity diversity. The main aim of this paper is to examine the
218: dynamical interaction of a growing proto-gas-giant planet with its
219: neighboring planetesimals and protoplanetary embryos in a gaseous
220: environment. This analysis is relevant in two contexts. Our first
221: objective is to evaluate whether process M1 can significantly reduce
222: the planetesimals accretion rate onto relatively low-mass (a few
223: $M_\oplus$) cores. A reduction of the energy dissipation associated
224: with planetesimal accretion would enable the gas to settle onto the
225: cores at a more rapid speed than the early models, thus shorten the
226: transition from stages of quasi-hydrostatic sedimentation
227: (S2) to runaway gas-accretion (S3). The reduced influx of the
228: planetesimals would also lower the replenishment of dust and the
229: contribution to the opacity in the envelope. Protoplanetary models
230: show that the suppression of opacity enhances both the heat flux
231: through the radiative region and the gas accretion rate (Ikoma,
232: et al. 2000, Hubickyj et al. 2005).
233:
234: This promising avenue to bypass the gas-accretion barrier also implies
235: a limited acquisition of heavy elements during the initial evolution
236: of proto planets when their mass $M_p$ is only a few Earth masses. In
237: order to account for the rich abundance of heavy elements in the gas
238: giants, especially in their envelopes, we need to consider a possible
239: mechanism for the resumption of planetesimal-accretion during the
240: runaway gas-accretion stage (S3) when
241: the protoplanets have acquired a major fraction of their present-day
242: gaseous envelopes (with $M_p \sim M_J$ where $M_J$ is Jupiter's mass).
243: Our second goal is to assess the efficiency of planetesimal accretion
244: under the influence of process M4. The first objective has
245: implications for the ubiquity of gas giants whereas the second is
246: linked to the structural diversity of gas giants.
247:
248:
249: In \S 2 of this paper, we first present a dynamical model for this
250: process and estimate the various time scales associated with both
251: aerodynamical and tidally-induced gas drag. In order to simplify the
252: problem, we assume a prescribed model for the evolution of protoplanet mass
253: which is based on the Bondi formula for idealized,
254: spherically symmetric, unimpeded accretion. We also utilize an {\it ad
255: hoc} uniform accretion prescription to illustrate the dominant
256: physical processes which determine the dynamical evolution of the
257: system. In \S3, we show that the combined effects of the
258: protoplanet's perturbation and the planetesimals' aerodynamical and
259: tidal interaction with their nascent disks lead to the formation of a
260: planetesimal gap when $M_p$ is a few $M_\oplus$. With a reduced rate
261: of planetesimal bombardment at the onset of the gas accretion, the
262: envelopes contract more rapidly and the gas accretion time scale can be
263: shortened from the early models. In \S4, we show that the planetesimal
264: accretion rate increases rapidly at the late stage of the
265: protoplanet's gas accretion. We also provide evidence to demonstrate
266: the accumulation of planetesimals outside the asymptotic feeding zone.
267: This effect can promote the formation of second-generation proto-gas-giant
268: planets. Finally, we summarize our results and discuss their
269: implications in \S 5.
270:
271: \section{Dynamical model}
272:
273: In this section, we study the growth of a protoplanet from the beginning
274: of quasi-hydrostatic sedimentation stage (S2),
275: i.e., after its core reaches an isolation mass.
276: The protoplanet is moving in a gaseous and planetesimal disk. The
277: orbits of the protoplanet and planetesimals are perturbed by the gas drag.
278: We first estimate the effects of gas drag in \S 2.1. Models of the protoplanet and
279: planetesimals are given in \S 2.2 and \S 2.3, respectively.
280:
281:
282: \subsection{Gravitational gas drag}
283: \label{sec:gasdrag}
284:
285: There are three physical processes that are acting
286: on the protoplanet and planetesimals during their evolution: aerodynamical gas drag,
287: gravitational tidal drag and dynamical friction.
288: Dynamical friction on the most massive embryos by the low-mass
289: planetesimals plays an important role during the stage of runaway
290: growth of protoplanetary embryos (e.g., Wetherill \& Stewart 1989,
291: Palmer et al. 1993, Kokubo \& Ida 1996,
292: Goldreich et al. 2004). During the subsequent oligarchic stage,
293: massive embryos emerge
294: to perturb the velocity dispersion ($\sigma$) of the residual
295: planetesimals (Ida \& Makino 1993, Kokubo \& Ida 1998).
296: Numerical simulations show that the collisions
297: generally lead to coagulation and the embryos attain most of the mass
298: in heavy elements (Kokubo \& Ida 2000, Leinhardt \& Richardson 2005).
299: In this limit, we assume that the effect of dynamical friction can be
300: incorporated into the gravitational drag due to embryo-gas
301: interaction.
302:
303: \subsubsection{Aerodynamical gas drag}
304:
305: The gas drag on small particles is in the form of aerodynamical drag
306: (e.g., Adachi et al. 1976, Tanaka \& Ida 1999). The
307: acceleration of a planetesimal with mass $m$ by the aerodynamical drag
308: has the form
309: \beq {\bf f}_{\rm aero}=-\frac{1}{2m} C_D \pi S^2 \rho_{g}
310: |{\bf U}|{\bf U},
311: \label{faero}
312: \eeq
313: where $C_D=0.5$ is the drag coefficient for objects with
314: large Reynold's number,
315: %(Landau and Lifthitz 1999),
316: $S$ is the radius of the planetesimal, $\rho_{\rm g}$ is the density
317: of gas, ${\bf U}={\bf V}_{\rm k}-{\bf V}_{\rm g}$ is the relative velocity,
318: ${\bf V}_{\rm k}$ and ${\bf V}_{\rm g}$ are the velocity vectors of
319: the planetesimal's Keplerian motion and gas motion respectively.
320:
321: The motion of the gas is subject to both the stellar gravity and
322: its own pressure gradient. In an unperturbed (by the protoplanet)
323: region of the disk, the balance of forces in the radial direction
324: gives
325: \beq
326: \frac{V_{\rm g}^2}{R}=\frac{V_{\rm c}^2}{R}+\frac{1}{\rho_{\rm g}}\frac{dP}{dR},
327: \label{vg}
328: \eeq
329: where $V_{\rm c} = \sqrt{GM_\ast/R}$ is the pressure-free circular velocity
330: at the radial distance $R$ to the host star, and $P$ is the gas
331: pressure. In a stable disk, $P=\rho_{\rm g} c_{\rm s}^2$, where $c_{\rm s}=(\frac{kT}
332: {\mu m_H})^{1/2}$ is the sound speed of an ideal gas in an isothermal
333: environment, with $k$ the Boltzmann constant, $m_H$ the proton mass, T the gas temperature
334: and $\mu$ the average molecular weight of the gas. We take $\mu=2.34$ for the gas with solar
335: composition, thus $c_{\rm s}=5.95 \times 10^3 $cm s$^{-1}
336: \sqrt{T/ {\rm K}}$. Suppose that
337: \beq
338: \begin{array}{l}
339: \rho_g=\rho_{g0} (\frac{R}{\rm 1AU})^{-s_{\rho}}, \\
340: c_{\rm s}^2=c_{s0}^2 (\frac{R}{\rm 1AU})^{-s_{\rm c}}, \\
341: T=T_0 (\frac{R}{\rm 1AU})^{-s_T}(\frac{M_*}{M_\odot})^\beta,
342: \end{array}
343: \eeq
344: where the subscript $_0$ denotes the value of the corresponding
345: quantity at 1AU (and $M_*=M_\odot$ in the equation of $T$).
346: From equation (\ref{vg}) we get
347: \beq
348: V_{\rm g}= V_c (1-2 \eta (R))^{1/2},
349: \label{vgeta}
350: \eeq
351: where
352: \beq
353: \eta= \frac{(s_{\rho}+s_{\rm c})}{2} (\frac{c_{\rm s}}{V_{\rm c}})^2,
354: \label{eta}
355: \eeq
356: and
357: \beq
358: \left( \frac{c_{\rm s}}{V_c} \right)^2=4 \times 10^{-6}
359: \left(\frac{T_0}{K} \right) \left(\frac{R}{\rm 1AU} \right)^{1-s_T}
360: \left(\frac{M_*}{M_\odot} \right)^{\beta-1}.
361: \label{csovk}
362: \eeq
363:
364: %where $V_{\rm k} = \Omega_{\rm k} R$ and $\Omega_{\rm k} = \sqrt{G M_\ast/R^3}$
365: %and the Keplerian velocity and angular frequency at the radial
366: %distance $R$ to the spin axis of disk,
367: %\beq
368: %\eta \equiv = - {1 \over 2 \rho_{gas} R \Omega_{\rm k}^2} {\partial p
369: %\over \partial R},
370: %\eeq
371:
372:
373: Throughout this paper, we use the minimum mass nebula model
374: (Hayashi et al. 1985) to evaluate fiducial model parameters, though our
375: basic algorithm can be applied to a more general disk. In this
376: model, the gas is heated to an equilibrium temperature by the central
377: star with temperature
378: \beq
379: T = 280\left( {R \over 1 {\rm AU}} \right)^{-1/2} \left(
380: {M_\ast \over M_\odot} \right).
381: \eeq
382: The surface density of the gas disk is given by
383: \beq
384: \Sigma_{\rm g}= f_g \Sigma_{g0} \left(\frac{R}{1{\rm
385: AU}} \right)^{-3/2},
386: \label{mingas}
387: \eeq
388: where $\Sigma_{g0}=1700 ~{\rm g ~cm^{-2}}$, $f_g$ is
389: a scaling factor so that $f_g=1$ corresponds to the minimum mass nebula.
390: The corresponding density of gas disk is
391: \beq \rho_{\rm g}= 1.4 \times 10^{-9} {\rm g ~cm^{-3}} f_g
392: \left(\frac{R}{1{\rm AU}} \right)^{-11/4}.
393: \label{rhogas}
394: \eeq
395: Since we have $s_\rho=11/4,s_c=1/2$, substituting them into equations
396: (\ref{eta}) and (\ref{csovk}),
397: $\eta$ has the form of
398: \beq
399: \eta(R)=0.0018 \left(\frac{R}{1AU} \right)^{1/2}.
400: \eeq
401: In more realistic models, the magnitude of $\eta$ can be modified by the surface irradiation,
402: internal viscous dissipation, and the radiation transfer through the
403: disk (Garaud \& Lin 2007).
404:
405:
406: Assuming the protoplanet or a planetesimal has a density $\rho$,
407: its radius can be expressed as
408: \beq
409: S=5.2 \times 10^{-3} {\rm AU} \left(\frac{M}{M_{\odot}}
410: \right)^{1/3} \left(\frac{\rm 1 ~g~cm^{-3}}{\rho} \right)^{1/3}.
411: \label{S}
412: \eeq
413: In terms of the above expression, equation (\ref{faero}) can be expressed as
414: \beq
415: \begin{array}{ll}
416: {\bf f}_{\rm aero}
417: = & - 10^{-7}C_D f_g~{\rm AU}^{-1}
418: \left(\frac{R}{1{\rm AU}} \right)^{-11/4} \\ &
419: \left(\frac{M}{M_\odot}
420: \right)^{-1/3} \left(\frac{\rho}{\rm 1~ g ~cm^{-3}}\right)^{-2/3}
421: |{\bf U}|{\bf U}. \end{array}
422: \label{faofin}
423: \eeq
424: %where $f_g = \Sigma_{g}/\Sigma_{g0}=\rho_{\rm g}/\rho_{\rm g0}$ is a scaling factor.
425:
426: The aerodynamical gas drag decreases the semi-major axis $a$, eccentricity
427: $e$, and inclination $i$ of the planetesimal orbit. The average
428: time scales for the evolution of these orbital elements are given by
429: Adachi et al. (1976)
430: \beq
431: \begin{array}{ll}
432: \frac{1}{\tau_{a,a}} & \equiv \frac{1}{a} \left(\frac{da}{dt}\right)_{\rm aero}
433: \\ & = -\frac{2}{\tau_{\rm aero}} \left(\frac58 e^2+\frac12 i^2+ \eta^2
434: \right)^{1/2} \left( \eta+\frac{17}{16}e^2+\frac18 i^2 \right)
435: \\
436: \frac{1}{\tau_{a,e}} & \equiv \frac{1}{e} \left(\frac{de}{dt} \right)_{\rm aero}
437: =\frac{2}{i} \left(\frac{di}{dt} \right)_{\rm aero}
438: \\ & =
439: -\frac{1}{\tau_{aero}} \left(\frac58 e^2 +\frac12 i^2+ \eta^2 \right)^{1/2},
440: \label{daaero}
441: \end{array}
442: \eeq
443: where $\tau_{\rm aero}$ is a time scale given as:
444: \beq \begin{array}{ll}
445: \tau_{\rm aero}& =\frac{2m}{\pi C_D s^2 \rho_{\rm g} V_{\rm k}(a)} \\
446: & \approx
447: \frac{3.5 \times 10^3 {\rm yr}}{f_{\rm g}} \left(\frac{M}{10^{17} {\rm g}} \right)^{1/3}
448: \left(\frac{\rho}{\rm 2g cm^{-3}} \right)^{2/3}
449: \left(\frac{a}{\rm 5AU} \right)^{13/4}.
450: \end{array}
451: \label{tgas}
452: \eeq
453: Note that $\tau_{a,a} \sim \tau_{a,e} / (\eta+ e^2+i^2) > > \tau_{a,e}$.
454:
455: \subsubsection{Tidal effect of gas disk}
456:
457: Tidal interaction between a gas disk and a
458: protoplanet leads to an effect similar to gas drag,
459: which is particularly important for the dynamical
460: evolution of protoplanets with large masses (Goldreich
461: \& Tremaine 1980, Ward 1986, 1989, 1997, Artymowicz 1993). We adopt
462: the form of acceleration from Kominami and Ida (2002):
463: \beq
464: \mathbf{f}_{\rm tidal}=-\frac{{\bf V}_k-{\bf V}_{\rm g}}{\tau_{\rm tidal}},
465: \label{tidal}
466: \eeq
467: where ${\bf V}_{\rm g}$ is the velocity of gas motion,
468: and we consider it in circular orbits, and $\tau_{\rm tidal}$ is the time scale defined as
469: (Ward 1989, Artymowicz 1993)
470: \beq
471: \begin{array}{ll}
472: \tau_{\rm tidal} & =\left(\frac{M}{M_\ast} \right)^{-1}
473: \left(\frac{\Sigma_{g}a^2}{M_\ast} \right)^{-1}
474: \left(\frac{c_{\rm s}}{V_{\rm c}} \right)^4 \Omega_{\rm k}^{-1} \\
475: & \approx {5 \times 10^{4} {\rm yr} \over f_g}
476: ~ \left(\frac{M}{10^{27} {\rm g}} \right)^{-1} \left(\frac{a}{5 {\rm AU}} \right)^2,
477: \eea
478: \label{ttidal}
479: \eeq
480: where $\Omega_k$ is the Keplerian frequency of circular motion.
481: Since dynamic friction of planetesimals
482: on protoplanets have similar expressions of acceleration and
483: time scale as those of the disk tide, i.e., equations
484: (\ref{tidal}) and (\ref{ttidal}) (see Appendix of Kominami \& Ida 2002), we
485: do not consider the effect of dynamical friction particularly in this work.
486:
487: Tidal drag decreases the eccentricities and
488: inclinations of the embryo orbits.
489: In principle, tidal interaction between embryos and the
490: disk also lead directly to the decay of embryo orbits (Ward 1997). The
491: rate of this ``type I migration'' is determined by an imbalance in
492: the torque from disk regions interior and exterior to the embryo
493: orbits. Linear analysis for this process is evaluated for
494: idealized background surface density (Tanaka
495: et al. 2002). The results of linear analysis imply that,
496: in a solar nebula environment, embryos more massive than the Earth
497: would migrate rapidly toward their host stars, thus gas giants
498: would rarely form (Ida \& Lin 2007). In general, both intrinsic (Rice
499: \& Armitage 2003, Laughlin et al. 2004, Nelson \&
500: Papaloizu 2003) and self-excited turbulence (Koller et al. 2003) may
501: lead to nonlinear evolution of disk structure, readjustment of the
502: torque distribution, and the suppression of type I migration.
503: However, these structural adjustments do not modify the
504: efficiency of eccentricity damping since the contribution from both sides
505: of the disk is cumulative rather than cancelling. Thus, we include
506: here the effect of tidally-induced eccentricity damping but neglect
507: that of type I migration. Nevertheless, there is an associated change
508: in the semi-major axes of the planetesimals and embryos. Within
509: $O(e^2,i^2)$, the average time scales for the evolution of these
510: orbital elements are given by,
511: \beq
512: \begin{array}{l}
513: \frac{1}{\tau_{t,a}} \equiv \frac{1}{a} \left(\frac{da}{dt} \right)_{\rm tidal}
514: =-\frac{1}{8\tau_{\rm tidal}}( 5e^2 + 2i^2 ) \\
515: \frac{1}{\tau_{t,e}} \equiv \frac{1}{e} \left(\frac{de}{dt} \right)_{\rm tidal}
516: =-\frac{1}{\tau_{\rm tidal}} \left(1-\frac{13}{32} e^2 -\frac18 i^2 \right) \\
517: \frac{1}{i} \left(\frac{di}{dt}\right)_{\rm tidal}=-\frac{1}{2\tau_{\rm tidal}}
518: \left(1+\frac{11}{16}e^2 +\frac{3}{16} i^2 \right).
519: \end{array}
520: \label{datide}
521: \eeq
522: See Appendix for a brief derivation.
523:
524: \begin{figure}
525: \vspace{4cm}
526: \special{psfile=f1.eps voffset=-80 hoffset=-20
527: vscale=30 hscale=30 angle=0}
528: %\plotone{f1.eps}
529: \caption{\small Damping time scales in the
530: semi-major axis $a$ and eccentricity $e$ due to the aerodynamical
531: (Eq. [13]) or tidal drag (Eq.[17]) for different mass of a
532: planetesimal (or a protoplanet) located at 5 AU. For aerodynamical
533: drag, a density of $\rho=2 {\rm ~g ~cm^{-3}}$ is assumed.
534: \label{fig1}}
535: \end{figure}
536:
537: The times scales for the orbital decay of planetesimals or protoplanets under aerodynamical
538: and tidal drag are shown in Fig.1 (See page 9).
539: During the disk evolution
540: and depletion (on a time scale $\sim 10^{6-7}$ yr), aerodynamical
541: drag is more effective for small planetesimals with mass $\le 10^{23}$g,
542: while tidal drag is more important for embryos with mass
543: $ \geq 10^{24}$g.
544: The surface density of the gas is globally depleted on a time scale
545: $\tau_{\rm dep} \sim 1-3 {\rm Myr}$, so the magnitude of all
546: time scales of the gas drag increases.
547:
548:
549:
550: \subsection{Protoplanet model}
551:
552: In this paper, we study the growth of an isolated protoplanet as the
553: progenitor of Jupiter. We analyze the dynamical evolution of its
554: nearby residual planetesimals and embryos subject to the
555: perturbation of the protoplanet and gas drag.
556: We start at the beginning of the quasi-hydrostatic
557: sedimentation stage (S2) of the protoplanet formation.
558: The following simplifications are adopted in our
559: simulations:
560:
561: \noindent
562: i) After its core has acquired an isolation mass, the protoplanet is
563: assumed to accrete gas with prescribed rates. The gas accretion model
564: is described in \S 2.2.1.
565:
566: \noindent
567: ii) The dynamical feedback of planetesimals to the protoplanet is
568: neglected because random orbital phases usually counteract each other.
569: The accumulative feed-back perturbations by its co-existing
570: nearby embryos on the protoplanet will be studied in
571: the future.
572:
573: \noindent
574: iii) The protoplanet does not have a significant radial excursions
575: during the evolution.
576:
577: The last approximation is consistent with
578: neglecting type I migration. Although close-in planets may have
579: attained their present-day orbits through extensive type II migration
580: (Lin et al. 1996), these events occur after the
581: proto-gas-giant planets have already acquired most of their masses and
582: on the viscous evolution time scale of the protostellar disks (Lin \&
583: Papalozou 1986). Here we focus our investigation primarily on the
584: acquisition of planetesimals during the formation stages of a protoplanet.
585:
586: Under the above simplifications, we place a protoplanet in an orbit
587: with elements $a_p=5$ AU, $e_p=0.01$, and $i_p=0.005$. According to
588: equation (\ref{datide}), the eccentricity and inclination of such an isolated
589: protoplanet embedded within the solar nebula would be damped by the
590: tidal drag soon, on a time scale shorter than its growth time scale prior to
591: gap formation but longer than the synodic period of nearby embryos and
592: planetesimals (see below). In dynamical equilibrium, the recoil
593: motion of the protoplanet is balanced by the tidal damping to yield
594: the assumed eccentricity and inclination, especially during the onset
595: of the quasi-hydrostatic sedimentation stage (S2) when the mass of
596: the protoplanet is modest.
597: During the runaway gas-accretion stage (S3), a gas gap forms near the protoplanet and
598: $\tau_{\rm tidal}$ becomes longer than its growth time through gas accretion.
599:
600:
601: \subsubsection{Inner and outer feeding zones}
602: \label{sec:feeding}
603:
604: Planetesimals grow into protoplanetary embryos (which significantly
605: perturb the motion of their neighboring planetesimals) and cores
606: (which accrete gas) through cohesive collisions (Safronov 1969). The
607: region from where a protoplanet has a non-zero probability to accrete
608: planetesimals during a single azimuthal passage is referred
609: to as its feeding zone. Neglecting the host stars' tidal perturbation
610: on their equation of motion (but including the Keplerian
611: shear), all planetesimals with semi-major axis $a$ and eccentricity $e
612: > \delta_a \equiv \vert a/a_{\rm p} -1 \vert$ are contained in the
613: feeding zone of the protoplanet centered on its semi-major axis $a_{\rm p}$.
614: However, many synodic periods ($\tau_{\rm syno} \sim 2 P_{\rm k} / (3\delta_a)$,
615: where $P_{\rm k}$ is the period of Kepler motion) may be needed before
616: close encounters can occur. The accretion rate onto the protoplanet
617: with a mass $M_{\rm p}$ and radius $R_c$ is (Safronov 1969)
618: \beq
619: \bea{ll}
620: \dot M_{\rm p} & \sim \pi R_c^2 \rho_{\rm d} \left( {2 G M_{\rm p} \over R_c \sigma^2} \right)
621: \sigma \\
622: & \sim \left( {2 \pi \over e^2 + 2.25 \delta_a^2} \right)
623: \left( {R_c \over a_{\rm p}} \right)
624: \left( {M_{\rm p} \over M_\ast} \right) \Sigma_{\rm d} a_{\rm p}^2 \Omega_{\rm k},
625: \eea
626: \label{eq:mdotp}
627: \eeq
628: where $\sigma = \sqrt{e^2 + 9 \delta_a^2/4 } \Omega_{\rm k} a_{\rm p}$, $H_{\rm p} \sim
629: \sigma/\Omega_{\rm k}$, $\rho_{\rm d} \sim \Sigma_{\rm d}/ 2 H_{\rm p}$, and $\Sigma_{\rm d}$ are the
630: velocity dispersion, scale height, spatial and surface density of the
631: planetesimals, respectively. The eccentricities of the planetesimals are excited by
632: the secular perturbation of the protoplanet, and the average excursions of
633: the eccentricities and inclinations of the planetesimals per synodic encounter can be expressed
634: as (Hasegawa \& Nakazawa 1990),
635: \beq
636: <\Delta e>\approx \left(\frac{1.9h}{\delta_a} \right)^3 \delta_a,
637: ~~ <\Delta i>\approx
638: \left(\frac{1.3h}{\delta_a} \right)^3i,
639: \label{de}
640: \eeq
641: where $h \equiv (M_{\rm p} / 3M_\ast)^{1/3}$ is the scaled
642: Hill's radius of the protoplanet. Note that secular perturbation does not significantly
643: change the semi-major axes of the planetesimals. Over $\tau_{\rm syno}$ the eccentricities
644: of planetesimals with mass larger than $10^{18}$ g are not
645: significantly damped by either aerodynamical or tidal gas drag. In
646: principle, nonlinear diffusion can lead to further eccentricity growth
647: in a gas-free environment. But it proceeds on time scales much longer
648: than $\tau_{\rm syno}$ and is effectively suppressed by the gas damping
649: effect (Zhou et al. 2007).
650:
651: Equation (\ref{de}) implies $e \sim \delta_a$ at $\delta_a \sim 1.5 h $.
652: We define this location to be the boundary between the inner and outer
653: feeding zone. In the inner feeding zone, the planetesimals undergo
654: radial excursions which cross the protoplanet orbit at each
655: azimuthal conjunction. The two-body formula is a reasonable
656: approximation to their encounters. In the outer feeding zone where
657: $\delta_a > 1.5 h$, two-body effects alone cannot lead to close
658: encounters. However, when the tidal perturbation of the host star is
659: also included in the equation of motion, the feeding zone expands to
660: include more distant planetesimals. A simple approach to approximate
661: the orbits of the planetesimals is to use a restricted three-body
662: approximation. Although the perturbation of the protoplanet induces
663: periodic (or chaotic if the planetesimal has relatively large energy)
664: variations on the planetesimals' star-centric orbital elements
665: $a,e,i$, the motions of the planetesimals are constrained by the
666: Jacobi integral, which can be expressed as
667: \beq
668: E_J= \frac{1}{2}(e^2+i^2)-\frac{3 }{8}\delta_a^2 +\frac{9}{2}h^2+O(h^3).
669: \label{EJ}
670: \eeq
671:
672: Planetesimals with positive Jacobi energy reside in the feeding zone
673: (Hayashi et al. 1977), i.e. they have finite close encounter
674: probability per $\tau_{\rm syno}$. In the $(a,e)$ and $(a,i)$ planes, the
675: boundary of the feeding zone $E_J=0$ is on a branch of hyperbolic
676: curves. The half width of the feeding zone is $\delta_a \sim 2\sqrt{3} h$
677: for planetesimals with $e < \delta_a$. In the outer feeding zone
678: where $\delta_a/h \sim 1.5-3.5$, $e < \delta_a $, the planetesimals can engage in
679: close encounters and possible physical collision with the protoplanet only within
680: certain ranges of the longitude of periapse.
681: Through nonlinear diffusion, the orbit-orientation angle undergoes random
682: walk (Zhou et al. 2007) and over many $\tau_{\rm syno}$'s
683: planetesimals may occasional venture into the
684: collision ``key holes'' of the protoplanet. The range of the phase for direct
685: collision decreases with $\delta_a$ and vanishes in the limit
686: $\delta_a > 2\sqrt{3} h$. Thus, the accretion rate of
687: planetesimals onto the protoplanet from the outer feeding zone with $\delta_a \sim
688: 1.5-3.5 h$ is significantly reduced from the value of $ \dot M_{\rm p}$ in
689: equation (\ref{eq:mdotp}).
690:
691: However, in a gaseous environment, the accretion of planetesimals
692: in the entire feeding zone can be enhanced by the gas drag effect,
693: especially in the outer feeding zone. Although the eccentricity
694: excitation of the planetesimals in this location is limited by equation
695: (\ref{de}), tidal damping of their eccentricities leads to orbital decay. In
696: \S\ref{sec:pgap} we show that planetesimals drift in from the
697: outer edge of the feeding zone to the proximity of the protoplanet
698: where the gravitational perturbation is intense and close
699: encounters occur more frequently. It is such induced orbital decay
700: rate rather than the protoplanet accretion rate ($ \dot M_{\rm p}$)
701: that determines the time scale for the clearing of the feeding zone.
702: Along the way, these
703: migrating planetesimals pass through the mean motion
704: resonances of the protoplanet and their eccentricities
705: are excited to large amplitudes. The
706: subsequent gas drag leads to orbital decay. With relative small
707: $\tau_{\rm aero}$ and $\tau_{\rm tidal}$, small planetesimals and
708: large embryos pass through the mean motion resonances. But some
709: intermediate-mass planetesimals have relatively long $\tau_{t,a}$
710: and $\tau_{a,a}$ and may be trapped by the mean motion
711: resonances (see \S\ref{sec:resonance}).
712:
713: \subsubsection{Isolation core mass and gas accretion model}
714: \label{sec:isolat}
715:
716: A planetary core temporarily halts its growth by accreting planetesimals
717: when it attains an isolation mass $M_{\rm iso}$ (Lissauer 1987). If
718: all the planetesimals in the feeding zone can be accreted by the
719: protoplanet, the magnitude of $M_{\rm iso}$ would be determined by
720: $\Sigma_{\rm d}$ and the width of the core feeding zone ($\Delta a$):
721: \beq
722: M_{\rm iso}=2\pi \Sigma_{\rm d} a_{\rm p}\Delta a.
723: \eeq
724: It is useful to scale the magnitude of $\Sigma_{\rm d}$ with that of the
725: fiducial minimum solar nebula (Hayashi 1981) outside the snow line,
726: \beq
727: \Sigma_{\rm d, min} = 30 (a/ 1{\rm AU})^{-3/2} ~{\rm g ~cm^{-2}},
728: \label{dust}
729: \eeq
730: by a multiplicative factor $f_{\rm d}$. Numerical simulations (Kokubo \& Ida
731: 2002) indicate that the feeding zone of a core has a width of $\Delta_a =
732: 10 a_{\rm p} h$ which is slightly larger than twice $2 \sqrt{3} a_{\rm p} h$ for
733: low-eccentricity planetesimals. This minor expansion is
734: due to the eccentricity excitation associated with nonlinear
735: diffusion (rather than linear secular perturbation) over the time
736: scale for reaching the feeding zone, $\sim 3 M_{\rm iso}/\dot M_{\rm p}
737: (M_{\rm iso})$. From these values, we obtain
738: \beq
739: M_{\rm iso}=2 M_{\oplus} f_{\rm d} ^{3/2}
740: \left(\frac{a_{\rm p}}{5{\rm AU}} \right)^{3/4}.
741: \eeq
742: We place our protoplanet at the present-day location of Jupiter,
743: i.e. at 5AU which is slightly outside the snow line. In a minimum
744: mass nebula, this radius is a preferred location for the onset of gas
745: giant formation because 1) the isolation mass of the embryos is a few
746: $M_\oplus$ and 2) the time scale for the buildup of the embryo with
747: isolation mass is comparable to or shorter than the disk depletion
748: time scale (Ida \& Lin 2004). We take $f_{\rm d}=2$ in this study, so
749: the protoplanet core in $5$ AU has a mass of $M_{p0}=5.67~M_{\oplus}$
750: before the onset of gas accretion.
751:
752: Cores with isolation masses attract nearby gas because their surface
753: escape speed is much larger than the sound speed of the disk
754: gas. Nevertheless, heat is released from the contraction of gas onto
755: the cores. When the protoplanet mass is low (a few $M_\oplus$),
756: inefficient heat transfer in its envelope leads to the buildup of a
757: high pressure gradient to balance its gravity and slow down the
758: accretion rate in the quasi-hydrostatic sedimentation stage (S2).
759: Early numerical models of protoplanetary
760: structure indicate that the main heat-diffusion bottleneck is in the
761: outer radiative region (Pollack et al. 1996). The gas accretion
762: rate would be greatly enhanced if the grain-dominated opacity is
763: suppressed (Ikoma et al. 2000, Hubickyj et al. 2005).
764: A possible
765: mechanism for major opacity reduction is dust coagulation and
766: sedimentation through the outer radiative region.
767: However, the grains may also be replenished by much larger colliding
768: planetesimals which disintegrate during their passages through the
769: envelope. The clearing of planetesimals from the feeding zones of the
770: accreting cores would greatly reduce the resupply rate.
771:
772: As a first approximation, we assume the heat transfer barrier can
773: be bypassed by the depletion of the feeding zone, and the process of
774: gas accretion is not impeded by the radiative feedback. At 5AU, when
775: the protoplanet mass is well below $M_J$, its Bondi radius $R_b =G M_{\rm p}
776: /c_{\rm s}^2$ is smaller than its Hill's radius $R_h = h a_{\rm p}$ and the disk scale
777: height $H/a_{\rm p}= c_{\rm s}/ V_{\rm k}= 0.07(a_{\rm p}/5~{\rm AU})^{1/4}$ in a minimum
778: mass nebula. For such a protoplanet, the disk gas in the background
779: is homogeneous and the tidal effect of its host star can be neglected.
780: For most of our calculations, we adopt the conventional Bondi formula
781: (Frank et al. 2002) for spherical accretion, in which the
782: accretion rate is given as
783: \beq \dot{M}_{BD} =
784: \frac{ \pi G^2 \rho_{\rm g}}{c_{\rm s}^3}\bar{\alpha} M^2 \equiv \xi
785: \bar{\alpha} M^2, \label{bd}
786: \eeq
787: where $\bar{\alpha}$ is a constant of order unity determined by the state equation
788: of the gas.
789: For a protoplanet at 5 AU, we find $ \xi \approx 1.2\times 10^{-32}f_{\rm g} {\rm g}^{-1}
790: {\rm yr}^{-1} =24 f_{\rm g} M_{\small \odot}^{-1} {\rm yr}^{-1}$. By integrating
791: equation (\ref{bd}), we obtain
792: $M=M_0 [1-(1-M_0/M_f)(t-t_0)/\tau_{\rm grow}]^{-1}$ for $t-t_0 \le \tau_{\rm grow}$,
793: where $t_0$ is the epoch when the accretion begins,
794: $\tau_{grow}$ is the time scale of the protoplanet growth to a mass of $M_{\rm f}$,
795: \beq
796: \tau_{\rm grow}\equiv \frac{1}{\bar{\alpha}
797: \xi} \left(\frac{1}{M_0}-\frac{1}{M_{\rm f}} \right)\approx \frac{1}{24
798: f_{\rm g} \bar{\alpha}\mu_0} {\rm yr}, \label{tgt}
799: \eeq
800: where $\mu_0=M_0/M_{\odot}$, $M_0$ is the mass of protoplanet at
801: the onset of gas accretion, and we assume that $M_{\rm f} >> M_0$ in the
802: approximation of equation (\ref{tgt}). In a minimum mass nebula, a
803: $5.67~M_{\oplus}$ protoplanet at 5 AU has $\tau_{\rm grow}
804: \sim 10^5$ yr. This time scale only refers to the growth time scale
805: associated with the gas accretion. The emergence of such a massive core
806: and the transition from stages of embryo-growth (S1) to quasi-hydrostatic sedimentation
807: (S2) may take longer time.
808:
809:
810: The validity of the assumed homogeneous, unimpeded, spherical accretion
811: flow onto the protoplanet is questionable when its mass becomes comparable
812: to that of Jupiter because $R_b \sim R_h \sim H$. The effect of
813: differential rotation in the disk and the tidal torque by the host
814: star channel the accretion flow through a protoplanetary disk. We are
815: primarily interested in the dynamical evolution of residual
816: planetesimals and embryos near the outer regions of the feeding zone
817: ($\delta_a > h$) which is not strongly perturbed by the distribution
818: of gas within $R_h$. As the protoplanet attains its asymptotic mass,
819: its tidal interaction with the gas disk leads to the formation of a gap
820: (see \S{\ref{sec:gasgap}). The open of gap reduce the accretion rate
821: from the unimpeded $\dot M \sim 10^{-3} M_J$ yr$^{-1}$ in
822: equation (\ref{bd}) by several orders of magnitude
823: (Dobbs-Dixon et al. 2007). However, this transition occurs rapidly and we can approximate
824: it with an abrupt termination of its growth. In order to take into
825: account the uncertainties in the boundary conditions, we consider a
826: range of values for $\bar{\alpha}$ and $f_{\rm g}$ so that
827: $\tau_{\rm grow}=10^{3}-10^6$ yr in the following calculations.
828:
829: In the Bondi model, the growth time scale decreases with the mass of the
830: protoplanet. As we show below, this growth pattern can lead
831: to the initial quenching of planetesimal bombardment and the
832: late-stage capture of residual embryos and planetesimals. In order to
833: highlight this behavior, we consider a second series of models
834: with an {\it ad hoc} prescription in which a constant gas accretion
835: rate,
836: \beq
837: \dot{M}_{LN}=\frac{1}{\tau_{\rm grow}}(M_{\rm f}-M_0),
838: \label{ln}
839: \eeq
840: onto the protoplanet is assumed. Figure 2 shows the evolution of
841: protoplanetary mass by accreting gas on a time scale of $\tau_{\rm
842: grow}=10^5$~yr according to these two models.
843: The initial mass of the isolated core is set as
844: $5.67 ~ M_{\oplus}$. Gas accretion is most effective around
845: $\tau_{\rm grow}=10^5$~yr in the Bondi model.
846:
847: \begin{figure}
848: %\plotone{f2.eps}
849: \vspace{5cm}
850: \special{psfile=f2.eps voffset=-80 hoffset=-20
851: vscale=30 hscale=30 angle=0}
852: \caption{\small Increase of protoplanet mass in two gas-accretion models:
853: Bondi accretion (Eq. [\ref{bd}])
854: and linear accretion (Eq. [\ref{ln}]) on a time scale $\tau_{\rm grow}=10^5$ yr.
855: The initial mass of the planet is $5.67~ M_{\oplus}$ and the final
856: mass is 1 Jupiter mass.
857: \label{fig2}}
858: \end{figure}
859:
860: \subsubsection{Gap opening in gaseous disk}
861: \label{sec:gasgap}
862:
863: When a protoplanet grows to a sufficiently large mass, a gap in the gas
864: disk forms around its orbit (Lin \& Papaloizou 1979). The formation
865: of the gap has two important effects. As we have already indicated
866: above, gap formation greatly reduces the accretion rate onto the
867: protoplanet and effectively terminates its growth. The clearing of
868: the gas also reduces the magnitude of the drag on the residual
869: planetesimals and embryos (see \S\ref{sec:gasdrag}). Here we briefly
870: describe our prescription for the emergence of tidally induced gaps in
871: the disk.
872:
873: The critical mass of a planet ($M_c$) over which it can open a gap is
874: determined (Lin \& Papaloizou 1993) by a viscous and a thermal
875: condition,
876: \begin{equation} \frac{M_{c, v}}{M_{\odot}}=\frac{40\nu}{a_{\rm p}^2\Omega_{\rm k}}= 40\alpha
877: \left(\frac{c_{\rm s}}{a_{\rm p}\Omega_{\rm k}} \right)^{2},
878: \end{equation}
879: \begin{equation} \frac{M_{c, t}}{M_{\odot}}=3
880: \left(\frac{H}{a_{\rm p}}\right)^3,
881: \end{equation}
882: where $\nu= \alpha c_{\rm s}^2/\Omega_{\rm k} $ is the kinematic viscosity, and
883: $\alpha$ is a parameter (Shakura \& Sunyaev 1973). At 5AU in a minimum
884: mass nebula, a protoplanet continues to grow until its mass
885: reaches $M_{c,t} \sim 300~ M_{\oplus}$ so that both the
886: thermal and viscous conditions are satisfied .
887:
888: The minimum width of the gap is the Hill's radius $R_h$
889: of the protoplanet.
890: Numerical simulation indicates that the gap extends to $2-3 R_h$. As
891: the mass of the protoplanet mass approaches that of Jupiter,
892: its unimpeded growth time scale reduces to less than $10^3$ yr,
893: which is shorter than the
894: viscous diffusion time scale across all regions of the disk wider than
895: $R_h$. Consequently, gas within a gap-feeding zone is rapidly depleted.
896: In the low-viscosity limit (where $\alpha < < 1$), we estimate the
897: half-width of the gap $\Delta$ on the assumption that all the gas in
898: the gap is accreted on to the protoplanet and contributes to its
899: asymptotic mass. Let us denote the mass increment as $\delta M$. In the absence of viscous
900: mass diffusion, $\delta M= 8 \pi\int_{a_{\rm p}}^{a_{\rm p}+\Delta}\rho_{\rm g} a H
901: da $, and the half-width of the gap is given by
902: \beq \frac{\Delta}{a_{\rm p}}=160 \left(\frac{\delta M }{M_{\odot}}
903: \right) \left(\frac{a_{\rm p}}{5AU} \right)^{-1/2}.
904: \label{gapwidth}
905: \eeq
906: From the above minimum-mass nebula parameters and the present-day
907: $a_{\rm p}=5AU$, $\delta M+M_{\rm p0} =10^{-3}M_{\odot}$, we find
908: $\Delta \approx 0.8 $ AU, which is approximately $2.3 R_h$ and
909: comparable to the value found in numerical simulations (cf Bryden
910: et al. 2000).
911:
912: %It may also be modified in the vicinity of the protoplanet by its
913: %tidal torque (see \S\ref{sec:gasgap}).
914: % When a sufficiently
915: % massive protoplanet satisfied both the viscous and thermal
916: % criteria, a gap is formed near its orbit to terminate its gas
917: % accretion (see \S\ref{sec:gasgap}).
918:
919: The gap-opening in the gas disk near a massive protoplanet may change
920: the kinematics of nearby planetesimals.
921: At the inner edge of the gas
922: gap, $\eta$ is enlarged due to the sharp pressure contrast, and the
923: orbital decay rate of planetesimals is enhanced. However, at the
924: outer edge of the gas gap, the direction of gas pressure is
925: inwards and the pressure gradient is
926: positive, resulting in $V_g> V_c$ according to equation (\ref{vg}), and
927: $\eta$ is negative by equation (\ref{vgeta}) (Lin \&
928: Papaloizou 1993, Bryden et al. 2000). Thus the tail wind of
929: the gas induces the planetesimals to migrate outward. The severe
930: depletion of gas in the gap region also reduces the drag at that
931: location. However, these feedback effects are negligible until
932: the protoplanet has opened a gap and attained most of its
933: asymptotic mass. After the gap formation,
934: the sign of $\dot a$
935: becomes positive (see Eq. [\ref{daaero}]) for low-eccentricity planetesimals (with $e^2 <
936: \vert \eta \vert$) and they move beyond the gap, whereas the
937: high-eccentricity planetesimals continue to migrate inward.
938:
939: The modification of $V_{\rm g}$ near the outer edge of the gas gap
940: also locally changes the sign of the tidal torque on the protoplanetary
941: embryos (see Eq. [\ref{tidal}]). In principle, this process
942: induces the embryos to undergo outward migration, analogous to the
943: type I inward migration. But similar to the aerodynamical drag, the
944: damping of the embryo eccentricities dominates the
945: evolution of their semi-major axes.
946:
947: As the depletion of gas in the gap reduces the local density
948: of the gas disk, we take ${\bf f}=0$ in
949: the corresponding aerodynamical and tidal drag formulas (Eqs.
950: [\ref{faero}] and [\ref{tidal}]) inside the gap in our simulations.
951: We assume that gas in the gap is depleted on a time scale $\tau=\Delta
952: r^2/\nu\sim H^2/\nu $, which gives $\tau \approx 2 \alpha^{-1}$
953: yr. Thus for $\alpha=10^{-3}$, we have $\tau\approx 2000$
954: yr. However, to simplify the problem, we neglect the migration effect
955: due to gap formation, which is very interesting and
956: will be studied elsewhere.
957:
958:
959: \subsection{Planetesimal model}
960:
961: In this paper, we study the evolution of planetesimals with initial
962: semi-major axes $a\in [3.2,8.5]$ AU, i.e., $a\in [a_{\rm p}-5a_{\rm p} h_{\rm f},
963: a_{\rm p}+10a_{\rm p} h_{\rm f}]$ with $a_{\rm p}=5$AU, $h_{\rm f}=(\mu_{\rm f}/3)^{1/3}$ and
964: $\mu_{\rm f}=10^{-3}$. We choose a larger region for the outer part, because
965: due to gas drag and eccentricity damping, planetesimals on both side
966: of the protoplanet suffer orbital decay. Exterior to the protoplanet,
967: planetesimals migrate toward its feeding zone.
968:
969: Numerical simulations indicate that during the oligarchic growth, the
970: feeding zone of the most massive embryos may indeed contain less
971: massive planetesimals and embryos (Kokubo \& Ida 2002). In order to
972: examine the mass-dependent collision probability of planetesimals with
973: a growing protoplanet, we carry out several simulations,
974: each with a population of uniform mass planetesimals. The individual
975: planetesimal mass under investigation ranges from $10^{17}
976: -10^{27}$g. The planetesimals are treated as test particles in the
977: simulation (i.e. they do not impose any gravitational
978: perturbation on each other or on the protoplanet) except when we
979: evaluate the magnitude of the gas drag. The low-mass range corresponds
980: to km-size objects, which can withstand the aerodynamical drag in a
981: minimum mass nebula with $\tau_{a, a} > \tau_{\rm dep}$ and be retained by
982: the disk. We did not simulate the interaction between a protoplanet
983: and embryos more massive than $10^{27}$g because it would not be
984: adequately approximated by the restricted three-body approach. A more
985: comprehensive treatment of the interaction between a population of
986: comparable mass embryos will be presented elsewhere (see Zhou et al. 2007).
987:
988: We choose three sets of models for detail analysis. In Models 1, 2,
989: and 3, we set the mass of each planetesimal to be $m=10^{19}$ g (Model 1),
990: $10^{24}$ g (Model 2), and $10^{27} $ g(Model 3). The planetesimals
991: in Model 2 correspond to the transitional objects as planetesimals
992: evolve into oligarchies, which perturb the velocity dispersion of their
993: neighbors. The embryos in Model 3 correspond to the
994: isolation mass in a minimum mass nebula interior to the snow line.
995: These three representative models
996: are also chosen to illustrate the relative importance of the gas drag
997: effects. The dominant physical process for eccentricity damping is
998: aerodynamical drag for the low-mass planetesimals in Model 1 and tidal
999: drag for the embryos in Model 3. The eccentricity damping time scale
1000: is the longest for the intermediate-mass oligarchies in Model 2.
1001:
1002: In order to build up sufficient samples for a statistical analysis, we
1003: use 1000 planetesimals for each simulation. We normalize these models
1004: with $\Sigma_{\rm d} = f_{\rm d} \Sigma_{\rm d, min}$, so the total mass of
1005: planetesimals in the region $[a_{\rm in},a_{\rm out}]$ is:
1006: \beq
1007: \bea{ll}M_{\rm tot} & =\int_{a_{\rm in}}^{a_{\rm out}} f_{\rm d} ~\Sigma_{\rm d, min}
1008: 2\pi a da \\
1009: & =14 f_{\rm d} M_{\oplus}
1010: \left[ \left(\frac{a_{\rm out}}{\rm ~1AU} \right)^{1/2}
1011: -\left(\frac{a_{\rm in}}{\rm 1~AU} \right)^{1/2} \right].
1012: \eea
1013: \eeq
1014: Thus the total mass of the planetesimals in $[3.2,8.5]$ AU with $f_{\rm d}
1015: =2$ is $31.5~ M_{\oplus}$.
1016:
1017: Due to their interaction with each other and the turbulent gas, the
1018: velocity dispersion of planetesimals, $\delta {\bf V}= {\bf V}-{\bf V}_{\rm k}$,
1019: is expected to have a Gaussian distribution in Cartesian coordinates.
1020: The corresponding eccentricities and
1021: inclinations of the test particles follow a Rayleigh distribution
1022: \cite{gl92}. The initial average eccentricities and inclinations of
1023: the planetesimals are taken as $0.0007$ and $0.00035$, respectively.
1024:
1025: \section{Formation of planetesimal gap during gas accretion}
1026:
1027: For our numerical simulations, we use the Regularized Mixed Variables
1028: Symplectic (RMVS3) method in the SWIFT packet (Levison \& Duncan 1994). %\cite{ld94}.
1029: This
1030: algorithm is well suited for the computation of close encounters
1031: of planetesimals with protoplanets (but not the inter-particle encounters).
1032: The time step of the
1033: integration is set as $0.05$ yr, i.e., $\sim 1/200$ of the period of
1034: the protoplanet orbit. All planetesimals which venture within the
1035: physical radius of the protoplanet are assumed be
1036: accreted by it. The density of planetesimals is taken as $\rho_{\rm
1037: tp}= {\rm 2 ~g ~cm^{-3}}$ in equation (\ref{faofin}). The radius of the
1038: protoplanet is given by equation (\ref{S}]) with two limiting values of $\rho_{\rm p}=
1039: {\rm 1 ~g ~cm^{-3}}$ and $0.125 ~{\rm g ~cm^{-3}}$. The quantity
1040: $\rho_{\rm p}= {\rm 1 ~g ~cm^{-3}}$ corresponds to that of the core with radius $R_c$.
1041: In reality, the effective capture cross section of a protoplanet is determined
1042: by the planetesimal masses, the relative speed as well as the protoplanet mass
1043: (which determines the density distribution in its gaseous
1044: envelope). During the initial gas accretion stage of a protoplanet,
1045: its envelope has a relatively small mass and does not significantly
1046: modify the capture cross section. After the onset of efficient gas
1047: accretion, the density in the outer regions of the envelope is also low as
1048: it undergoes dynamical collapse. Gas accumulates near the core as its
1049: inward motion is halted. The radius of this location is determined by
1050: the efficiency of radiation transfer but is a few times larger than
1051: $R_c$. After the protoplanet has attained its asymptotic mass and gas
1052: accretion onto it is quenched, the radius of a typical gas giant
1053: quickly reduces to twice that of Jupiter. In order to take these
1054: possibilities into account and still keep our investigation within
1055: relatively general and simple bounds, we consider a second set of
1056: capture criterion for intruding planetesimals by doubling the
1057: protoplanet's effective radius (which corresponds to assuming a
1058: homogeneous density $\rho_{\rm p}= {\rm 0.125 ~g ~cm^{-3}}$ for the
1059: protoplanet).
1060:
1061: There are also planetesimals scattered to solar distances larger than
1062: $100$ AU or smaller than $1$ AU. These particles are respectively
1063: classified as ejectors to the outer solar system or Sun crashers. The
1064: planet mergers, ejectors, and Sun crashers are registered and removed
1065: from the subsequent evolution.
1066:
1067: The eccentricity of the planetesimal orbits are excited by the
1068: protoplanet and damped by gas drag. Eccentricity damping also lead to
1069: a decline in the semi-major axes (see \S\ref{sec:gasdrag}). The energy of
1070: the planetesimal orbits may also be modified by resonant interaction
1071: and close encounters with the protoplanet. Planetesimals with $a<1$
1072: AU are not able to collide with the protoplanet in their subsequent
1073: evolutions, therefore an inner boundary of $1$ AU is adequate for the
1074: objectives of the present investigation. In all the simulations
1075: presented here, we assume the gas disk has an initial surface density
1076: of the minimum solar nebula (see Eq. [\ref{rhogas}]). After the formation
1077: of the protoplanet core, a uniform exponential depletion of gas on a
1078: time scale of $10^6$ yr is also assumed. In most but not all
1079: models, this time scale is larger than the magnitude of $\tau_{\rm
1080: grow}$. The solid disk has an initial surface density of twice the
1081: minimum solar nebula (see Eq. [\ref{dust}]).
1082:
1083: \subsection{Opening of planetesimal gap}
1084: \label{sec:pgap}
1085:
1086: An important phenomenon in the evolution of planetesimals in a gaseous
1087: environment is the opening of a planetesimal gap around the
1088: protoplanet. The opening of a planetesimal gap in a gas-free environment
1089: was discussed by Rafikov (2001). This situation is analogous to
1090: planetary rings being shepherded by a satellite. Due to the
1091: differential Keplerian motion, inter-particle encounters lead to
1092: angular momentum transfer and the diffusion of the ring. But in the
1093: proximity of the protoplanet, tidal perturbation from the protoplanet tends to expel
1094: the planetesimals away from it (Goldreich \& Tremaine 1978, 1980, Lin
1095: \& Papaloizou 1979). When the tidal torque exceeds the rate of
1096: inter-particle angular momentum exchange, a gap centered on the protoplanet
1097: forms.
1098:
1099: During the growth of the embryos and the formation of the core, the
1100: residual planetesimals attain a MRN size distribution, i.e.
1101: most of the planetesimals' surface area and mass are contributed by
1102: the small and large planetesimals respectively. During the stages of
1103: quasi-hydrostatic sedimentation (S2) and runaway gas-accretion (S3),
1104: the maximum embryo size has already increased, so that the
1105: collision frequency of the smaller residual planetesimals becomes
1106: small compared with that of their synodic encounters with the
1107: protoplanet. In this low-collision-frequency limit, the
1108: secular perturbation of protoplanet on the residual planetesimals
1109: needs to be taken
1110: into account (Franklin et al. 1980, see \S\ref{sec:feeding}).
1111:
1112: Although the energy dissipation rate associated with planetesimal
1113: collisions is reduced with the collision frequency, the excited
1114: planetesimals also experience eccentricity damping from the disk gas.
1115: Since $\tau_{a,e}$ and $\tau_{t, e}$ in equations (\ref{daaero})
1116: and (\ref{datide}) are much larger than
1117: $\tau_{\rm syno}$ (over which span the eccentricities of
1118: planetesimals are excited to $< \Delta e >$),
1119: the gas drag does not directly reduce the planetesimal eccentricities from
1120: that in equation (\ref{de}).
1121: However, this process is accompanied
1122: by a slow radially inward drift on time scales $\tau_{a,a} \sim e^{-2}
1123: \tau_{a,e}$ and $\tau_{t,a} \sim e^{-2} \tau_{t,e} $
1124: (see Eqs. [\ref{daaero}] and [\ref{datide}]) .
1125: This drift speed is faster for planetesimals
1126: inside than those outside the feeding zone of the protoplanet because they
1127: are excited to larger $< \Delta e >$.
1128:
1129: The inward drift causes planetesimals interior to the protoplanet
1130: orbit to drift away from the feeding zone of the protoplanet. But it also
1131: brings the planetesimals exterior to the protoplanet orbit into its
1132: feeding zone, where their orbital responses become chaotic, resulting
1133: in capture or close encounters. Planetesimals on both sides of the
1134: protoplanet orbit are evacuated and a gap eventually appears. Beyond
1135: the feeding zone, the planetesimal orbits evolve slowly because the
1136: protoplanet perturbation is much weaker so that the magnitude of $<
1137: \Delta e >$ is much smaller. Some external planetesimals are trapped
1138: onto mean motion resonances of the protoplanet, where their
1139: eccentricities are increased. Since the gas gap is confined to within
1140: the feeding zone, the residual disk gas damps the planetesimal
1141: eccentricities to modest equilibrium values.
1142:
1143:
1144: \begin{figure}
1145: \vspace{5.5cm}
1146: %\plotone{f3.eps}
1147: \special{psfile=f3.eps voffset=-90 hoffset=-40 vscale=35 hscale=35
1148: angle=0}
1149: \caption{\small Configurations of survived planetesimals with mass
1150: $m=10^{19}$ g in the $(a,e)$ plane during evolution.
1151: The protoplanet accretes gas according to the Bondi model on a time scale
1152: $\tau_{\rm grow}=10^5$ yr. The
1153: dotted and solid lines are the boundaries of the
1154: feeding zone ($E=0$) at time $t=10^4$ yr and $t=10^5$ yr,
1155: respectively.
1156: The gray dot at $a \approx 5$ AU shows the position of the
1157: protoplanet. Some locations of mean motion resonances between planetesimals and the protoplanet
1158: are shown in plot (d).
1159: \label{fig3}}
1160: \end{figure}
1161:
1162: Figure 3 illustrates the distribution of low-mass ($m=10^{19}$ g)
1163: planetesimals (Model 1) in the $(a,e)$ plane at different epochs. In
1164: this model, we adopt the Bondi gas-accretion prescription and set the
1165: time scale of $\tau_{\rm grow}=10^5$ yr. At $t=10^4$ yr, $M_{\rm p}
1166: \sim 6 M_\oplus$, $\tau_{\rm syno} \sim 10^3$ yr, and the eccentricities of
1167: the planetesimals within the current feeding zone are excited to
1168: values $< \Delta e > \sim 0.01-0.1$ (Fig.3a). After $t=10^5$ yr,
1169: the protoplanet acquires its full asymptotic mass. There are many
1170: planetesimals inside the final feeding zone of the protoplanet. A
1171: planetesimal gap begins to form, albeit at a slower rate than the
1172: expansion of the hypothetical feeding zone (Fig.3b). At the outer
1173: regions of the feeding zone, the eccentricities of many planetesimals
1174: are greatly excited. The V shape of the planetesimals' ($a,e$)
1175: distribution indicates that many are scattered analogous to the
1176: KBO's. After $t \ge 10^6$ yr, planetesimals in the feeding zone
1177: are completely cleared, as those interior to the protoplanet orbit
1178: drift inward while those exterior to it are scattered outward (Fig.3c,
1179: 3d). The analysis for the clearing time scale of the planetesimal
1180: gap is presented in \S\ref{sec:pgaptime}.
1181:
1182: Since the opening of the planetesimal gap is the result of gas drag and the
1183: protoplanet perturbation, the time scale for their eccentricity
1184: damping and therefore gap-clearing depends on their mass (Ida \& Lin
1185: 1996). Figure 4 displays the configurations of survived planetesimals
1186: in the $(a,e)$ plane for planetesimals with mass $m=10^{23}$ g (Model
1187: 2) and protoplanetary embryos with $m=10^{27}$ g (Model 3),
1188: respectively.
1189:
1190: \begin{figure}
1191: \vspace{5.5cm}
1192: \special{psfile=f4.eps voffset=-100 hoffset=-40 vscale=35 hscale=35
1193: angle=0}
1194: \caption{\small
1195: Configurations of survived planetesimals with mass
1196: $m=10^{23}$ g and $m=10^{27}$ g in the $(a,e)$ plane during evolution.
1197: The protoplanet accretes gas by the Bondi model on a time scale
1198: $\tau_{\rm grow}=10^5$ yr. The
1199: dotted lines and solid lines are the boundaries of the feeding zone ($E=0$)
1200: at time $t=10^4$ yr and $t=10^5$ yr,
1201: respectively. The cases of $t= 10^4 $ yr are similar to Fig.3.
1202: The gray dot at $a=5$ AU shows the position of the protoplanet.
1203: \label{fig4}}
1204: \end{figure}
1205:
1206:
1207: During the buildup of the protoplanet asymptotic mass (at $t=10^5 $
1208: yr in Figs. 4a and 4c), residual planetesimals are found inside the
1209: feeding zone in all models. A closer inspection indicates that the
1210: opening of the planetesimal gap is less efficient in Model 2 than
1211: in Models 1 and 3. This minor difference supports the conjecture that
1212: the clearing of the gap is due to the combined action of the protoplanet
1213: perturbation and gas drag, because the $e$-damping and $a$-decay rates
1214: of the intermediate-mass planetesimals (represented by Model 2) are
1215: smaller than those of the low-mass planetesimals and high-mass embryos
1216: (see Fig.1 for the drift time scale). On time scales much longer than
1217: $\tau_{\rm grow}$, the intermediate-mass continues to occupy the edge
1218: regions of the feeding zone. Many planetesimals are also trapped in
1219: the outer mean motion resonances with increased eccentricity
1220: (Fig.4b). In contrast, the drift speed of the embryos ($m=10^{27}$ g)
1221: is faster than those in Models 1 and 2, and all the embryos in the
1222: feeding zone are cleared out after $t=10^6$ yr (Fig.4d).
1223: Nevertheless, the width of the planetesimal gap is limited to $\sim 2
1224: \sqrt{3} h a_{\rm p}$.
1225:
1226: \begin{figure}
1227: \vspace{5cm}
1228: \special{psfile=f5.eps voffset=-100 hoffset=-40 vscale=35 hscale=35
1229: angle=0}
1230: %\plotone{f5.eps}
1231: \caption{\small
1232: Semi-major axes distribution of survived planetesimals
1233: at four epochs during the mass growth of the protoplanet.
1234: Each bar corresponds to $0.1$ AU. The mass of individual
1235: planetesimal is $10^{23}$ g. The protoplanet accretes gas by the Bondi
1236: model on a time scale
1237: $\tau_{\rm grow}=10^5$ yr. $N_0$ is the total number of
1238: planetesimals survived at that time. The solid curve corresponds to the
1239: initial density profile. The opening of a planetesimal gap leads to a
1240: slight enhancement of surface density near the boundary of the feeding
1241: zone.
1242: \label{fig5}}
1243: \end{figure}
1244:
1245: The opening of a planetesimal gap near the protoplanet is also shown in
1246: Fig. 5 with the intermediate-mass ($10^{23}$ g) planetesimals. At
1247: $t=\tau_{grow} = 10^5$ yr, the radial distribution of the
1248: planetesimals show a diffusion profile around the corotation radius of
1249: the protoplanet. Some survival planetesimals near the protoplanet are
1250: caught onto horseshoe or tadpole orbits. Figure 6 plots two examples in the tadpole
1251: orbits librating around the $L_4$ and $L_5$ points, respectively.
1252: In Fig. 5, there is a slight
1253: enhancement of surface density beyond the edge of the feeding zone due
1254: to the resonant trapping of inward-drifting planetesimals, which
1255: will be discussed in \S 4.3. This accumulation of
1256: planetesimals increases the local surface density and $M_{\rm iso}$,
1257: promotes the growth rate of protoplanetary cores and the
1258: emergence of secondary proto-gas-giant planets.
1259:
1260: \begin{figure}
1261: \vspace{5cm}
1262: \special{psfile=f6.eps voffset=-80 hoffset=-20 vscale=30 hscale=30 angle=0}
1263: %\plotone{f6.eps}
1264: \caption{\small Tadpole
1265: orbits of two surviving planetesimals in model Fig.5d. The
1266: coordinate is set as origin at the mass center of the sun and protoplanet (with mass growing),
1267: and corotating with the protoplanet at
1268: $\approx 5AU$. The planetesimals' asymptotic eccentricity is
1269: $e\approx 0.001 $. The dot around $(5,0)$ marks the
1270: orbit of the protoplanet. The planetesimals have semi-major axis
1271: $a=4.875$ and $a=5.038$ at $t=10^6$ yr, respectively.
1272: %$e=1.9\times 10^{-3}, i=0.90\times 10^{-3}0.00090$, and
1273: %$a=4.805$, $e=5.2\times 10^{-3}, i=1.6\times 10^{-3}0.00090$ at time $t=10^6$ yr.
1274: \label{fig6}}
1275: \end{figure}
1276:
1277: \subsection{Planetesimal gap clearing time scale}
1278: \label{sec:pgaptime}
1279:
1280: There are several relevant time scales to be considered. In
1281: \S\ref{sec:feeding}, we showed that, under the protoplanet
1282: perturbation, planetesimals with overlapping 2-body (neglecting the
1283: stellar tide) orbits extend to the boundary of the inner feeding zone
1284: where $\delta_a < 1.5 h$. The time scale of depletion of the total
1285: planetesimal population in this region, $M_{\rm p, f} =4 \pi \Sigma_{\rm d}
1286: \delta_a a_{\rm p}^2$, can be derived in view of equation (\ref{eq:mdotp}) with
1287: $e \sim <\delta e> $ from equation (\ref{de}),
1288: \beq
1289: \bea{ll}
1290: \tau_{\rm p, f} & \equiv {M_{\rm p, f} \over {\dot M}_{\rm p} }
1291: \simeq \left( {3 P_k \over 4 \pi} \right) \left( {\delta_a \over h}
1292: \right)^3 \cdot \\
1293: & \left( 1 + {4 \over 9} \left( { 1.9 h \over \delta_a} \right)^6
1294: \right) \left( {a_{\rm p} \over R_c} \right)
1295: \simeq \left( {a_{\rm p} \over R_c} \right) P_k
1296: \eea
1297: \eeq
1298: where $P_k$ is the Keplerian period. So $\tau_{\rm p, f} < 10^5$ yr.
1299:
1300: This rapid depletion time scale is only applicable to
1301: the planetesimals in the inner feeding zone where $\delta_a < 1.5 h$.
1302: Planetesimals in the outer feeding zone with $\delta_a/h \sim 1.5
1303: -3.5$ only occasionally venture into the Roche
1304: lobe of the protoplanet. Based on the discussions in the previous section, we now derive
1305: the time scale for planetesimals to migrate across the boundary
1306: between the inner and outer feeding zone. Since $\tau_{\rm p, f}$ is
1307: relatively small, the duration of the migration across the outer
1308: feeding zone essentially corresponds to the clearing
1309: time scale of the planetesimal gap.
1310:
1311: Suppose the protoplanet has a mass $\mu=M_{\rm p}/M_\ast$ at time $t$,
1312: so its instantaneous normalized Hill radius is $h=(\mu/3)^{1/3}$.
1313: In terms of the protoplanet's asymptotic (at $t\ge \tau_{\rm grow}$)
1314: normalized Hill's radius $h_{\rm f}$, the scaled distance of a planetesimal
1315: is defined as
1316: \beq
1317: b_{\rm f}\equiv \frac{\delta_a}{h_{\rm f}},
1318: \label{bf}
1319: \eeq
1320: where $\delta_a=|a/a_p-1|$. According to equation (\ref{daaero}),
1321: the speed of the inward drift for a
1322: planetesimal with mass $m\le 10^{23}$ g is given as
1323: \beq
1324: \dot{a}_{\rm aero}\approx - 1.7 \frac{a_{\rm p} }{\tau_{\rm
1325: aero}} e ^3,
1326: \label{dadt}
1327: \eeq
1328: where we suppose $e^2 \gg \eta$, because the eccentricity of
1329: planetesimals inside the feeding zone could be excited up to $\sim
1330: 0.1$ (e.g., Figs.3-4). In \S\ref{sec:feeding}, we evaluate the average
1331: excursions of $<\Delta e>$ per encounter. (As small initial inclination
1332: is expected, we neglect the modulation in $\Delta i$.) Since $\tau_{a,e}
1333: > > \tau_{\rm syno}$, we can replace $e$ in equation (\ref{dadt}) by $<\Delta e>$
1334: in equation (\ref{de}) to obtain
1335: \beq
1336: (\dot{b}_{\rm f})_{\rm aero}
1337: = \frac{550 }{\tau_{\rm aero}}\frac{1}
1338: { b_{\rm f}^6 }\frac{h^9}{h_{\rm f}^7},
1339: \eeq
1340: where the dependence on $h_{\rm f}$ is introduced for the purpose of
1341: normalization. With the growth of the protoplanet mass, $h$ also
1342: increases with time. However, in order to simplify the problem, we
1343: first assume $h$ as a constant $h_{\rm f}$ (an approximation to be justified
1344: {\it a posteri}) so that
1345: \beq
1346: \left(b_{\rm f} \right)_{\rm aero}=\beta_1 \left(\frac{h}{h_{\rm f}}
1347: \right) \left(\frac{h^2 }
1348: {\tau_{\rm aero}}t \right)^{1/7}, ~~(b_{\rm f} \le 2\sqrt{3} \left(\frac{h}{h_{\rm f}} \right) ),
1349: \label{dbaein}
1350: \eeq
1351: where $\beta_1=3.2$. The limiting value ($2 \sqrt{3}$) for $b_{\rm f}$
1352: corresponds to the width of the entire feeding zone. Planetesimals
1353: drift from the protoplanet orbit to the inner boundary of the
1354: asymptotic feeding zone and from the outer boundary of the asymptotic
1355: feeding zone to the protoplanet orbit on a time scale of
1356: \beq
1357: \left(T_{\rm open} \right)_{\rm aero}=
1358: (\frac{2\sqrt{3}}{\beta_1})^7\frac{\tau_{\rm aero}}{h^2}.
1359: \label{topenae}
1360: \eeq
1361: The above expression is obtained by equating the left hand size of equation
1362: (\ref{dbaein}) with $2\sqrt{3} h/h_{\rm f}$.
1363:
1364:
1365: According to the above
1366: equation, gap formation for the low-mass planetesimals proceeds on a
1367: time scale $\tau_{aero}/h^2$. Substituting $\tau_{\rm aero}$ from equation
1368: (\ref{tgas}) and assuming a constant protoplanet mass, we find
1369: $T_{\rm open} \sim 3 (M_J/M_{\rm p})^{2/3}$ Myr for Model 1. We carry out
1370: two comparisons, Models 1a and 3a, in which the mass of the protoplanet
1371: is fixed to that of Jupiter so that $h=h_{\rm f}$. We compare, in Figs. 7a
1372: and 7c, the results of the numerical simulations with that in
1373: equation (\ref{dbaein}). Numerically, we determine $b_{\rm f}$ from the
1374: distribution plots such as Fig. 5 at some typical epoch. The
1375: magnitude of $b_{\rm f}$ is defined to be the maximum half-width of the gap
1376: within which only planetesimals on horseshoe or tadpole orbits (around
1377: $5$AU) survived.
1378:
1379: The qualitative agreement between the functional dependence of $b_{\rm f}$ on $h$ in
1380: the numerical results and the expression in equation (\ref{dbaein}) supports
1381: the interpretation that the clearing of the gap is regulated by the
1382: orbital decay of the planetesimals in the feeding zone. (This
1383: agreement is particularly good during the expansion of the gap through
1384: the initial outer feeding zone where the migration scenario is most
1385: applicable.) However, in comparison with the expression in equation
1386: (\ref{dbaein}), a factor of $2-3$ for $\beta_1$
1387: is needed to fit the numerical simulations. This difference in the
1388: magnitude of $\beta_1$ is caused by an underestimate in the analytical
1389: expression for $<\Delta e>$ which did not take into account the
1390: cumulative eccentricity modulation during each secular cycle.
1391: From equation ({\ref{topenae}) , the expansion of the gap is
1392: stalled with an asymptotic width $\sim 2 \sqrt{3} h$ at $\sim
1393: \tau_{\rm aero}$ in the numerical simulations. A similar analysis also
1394: applies to the massive embryos for which the tidal drag is more
1395: appropriate.
1396:
1397: We now return to the more realistic models in which the planetesimal
1398: gap formation proceeds during the growth of the protoplanet. Despite
1399: the increases of the protoplanet mass, the above approximations
1400: would essentially be valid if $\tau_{\rm grow} > T_{\rm open}$. In
1401: this limit, planetesimal gaps with the instantaneous feeding zone
1402: width ($2\sqrt{3} h$) form and expand with the mass of the
1403: protoplanet. Based on the assumption that most planetesimals in the
1404: feeding zone can collide with the protoplanet core and are the main
1405: contributors to the initial growth of the protoplanet, Pollack et al.
1406: (1996) derived their bombardment rate onto the core from the
1407: expansion rate of its feeding zone. In that model, the
1408: suppression of the gas accretion rate due to the energy dissipation of
1409: planetesimal accretion has been taken into account.
1410:
1411: However, in the limit of modest $M_{\rm p}$ (a few $M_\oplus$), both
1412: the protoplanet's $h$ and the planetesimal eccentricities are very small so that
1413: $T_{\rm open} > \tau_{\rm grow}$ even for a protracted protoplanetary
1414: growth. In this case, the feeding zone expands faster than they can
1415: be cleared out (especially in Model 2 in Fig.4). Both the protoplanet
1416: mass and feeding zone width attain their asymptotic values on a time
1417: scale $\tau_{\rm grow}$, after which gap formation and clearing of the
1418: planetesimals in the feeding zone proceed on a time scale $T_{\rm
1419: open}$. This gradual mass ramp up significantly delays the clearing
1420: of the gap. During the advanced stage, the assumption of constant
1421: $M_{\rm p}$ is again satisfied and the values of $h$ can be replaced by
1422: $h_{\rm f}$ in the above equations. In the next section, we show that not
1423: all the planetesimals in the feeding zone collide with the
1424: protoplanet core and their collision rate may be substantially lower
1425: than that estimated by Pollack et al. (1996). This effect can
1426: reduce the barrier for the gas accretion rate onto the protoplanet
1427: envelope.
1428:
1429: Using the same approach, we deduce the width and time scale associated
1430: with the gap-opening of intermediate and high-mass ($>10^{23}$ g)
1431: embryos. In this case, the gravitational tidal drag provides the
1432: dominant eccentricity damping effect. From equation (\ref{datide}), we find
1433: \beq
1434: \dot{a}_{\rm tidal}\approx - \frac{5}{8}
1435: \frac{a_{\rm p} }{\tau_{\rm tidal }} e ^2,
1436: \label{datd}
1437: \eeq
1438: and
1439: \beq
1440: \left(b_{f} \right)_{\rm tidal}= \beta_2 \left(\frac{h}{h_{\rm f}} \right)
1441: \left(\frac{h }{\tau_{\rm tidal}}t \right)^{1/5}, ~~(b_{\rm f} \le 2\sqrt{3} \left(\frac{h}{h_{\rm f}} \right)),
1442: \label{dbtdin}
1443: \eeq where $\beta_2=2.7$. The time scale to open a gap is given as,
1444: \beq
1445: \left(T_{\rm open} \right)_{\rm tidal}=(\frac{2\sqrt{3}}{\beta_2})^5 \frac{\tau_{\rm tidal}}{h}.
1446: \eeq
1447: Figure 7 shows the evolution of the width of the planetesimal gap
1448: determined from our numerical simulations of Models 1 and 3. In
1449: Figs. 7c-7d, the protoplanet accretes gas according to the Bondi
1450: accretion prescription with $\tau_{\rm grow}=10^5$ yr, as in the
1451: cases of Figs. 3-5. The solid curves are theoretical predictions
1452: given by equations (\ref{dbaein}) and (\ref{dbtdin}) with different
1453: coefficients. The functional forms again are in general agreement,
1454: though there is a factor of 3 discrepancy in the coefficient of
1455: $\beta_2$.
1456:
1457: \begin{figure}
1458: \vspace{5cm}
1459: \special{psfile=f7.eps voffset=-100 hoffset=-30 vscale=35 hscale=35 angle=0}
1460: %\plotone{f7.eps}
1461: \caption{\small Evolution of the width of the planetesimal gap.
1462: The protoplanet has constant mass $\mu=10^{-3}$ in panels (a) and (b), and grows from
1463: $\mu=1.7\times 10^{-5}$ to $\mu=10^{-3}$ in panels (c) and (d) through gas-accretion
1464: with a Bondi gas-accretion model of $\tau_{\rm grow}=10^5$. The solid curves are theoretical
1465: predictions given by equations (\ref{dbaein}) and (\ref{dbtdin}) with
1466: coefficients: (a) $\beta_1=7.4$, (b) $\beta_2=8.0$, (c) $\beta_1=9.6$,
1467: (d) $\beta_2=8.0$.
1468: The squares and triangles denote the inner and outer boundaries, respectively.
1469: The dashed curves in panels (c) and (d) show the width of the
1470: feeding zone defined by $2\sqrt{3}|a-a_{\rm p}|/(a_{\rm p}h)$ at time $t$.
1471: \label{fig7}} \end{figure}
1472:
1473:
1474: Outside the boundary of the feeding zone, the rate of the eccentricity
1475: excitation by the protoplanet is small. According to equations
1476: (\ref{daaero}) and (\ref{datide}), the migration speed becomes
1477: correspondingly smaller by several orders of magnitude. Though the
1478: migration speed outside the gap can be obtained following similar
1479: lines of reasoning, such an analysis is not crucial for this study and
1480: we will not discuss it further.
1481:
1482: \subsection{Dependence on gas accretion model and time scale }
1483:
1484: In the above simulations, the time scale of gas accretion onto the
1485: protoplanet, $\tau_{\rm grow}$, is set to be $10^5$ yr. In order
1486: to determine the dependence of the collision efficiency of
1487: planetesimals on the growing time scale of the protoplanet, we adopt a
1488: range of $\tau_{\rm grow}$, from $10^3$ to $10^6$ yr, with both the
1489: Bondi (\ref{bd}) and linear (\ref{ln}) prescriptions for gas
1490: accretion. Figure 8 illustrates the collision ($P_{\rm col}$) and escape
1491: ($P_{\rm esc}$) probabilities of intermediate-mass ($10^{23}$ g)
1492: planetesimals (in Model 2) during the growth of the protoplanet. Since the
1493: evaluation of the protoplanet radius is based on the present
1494: density of gas giants, $P_{\rm col}$ should be regarded as a lower limit.
1495: Although protoplanets have extended envelopes, most of their mass
1496: resides in the core and the density in the envelope decreases rapidly
1497: with radius. Small and modest-size planetesimals may be captured by
1498: the protoplanet when they pass through its envelope. But the large
1499: embryos can only merge with the protoplanet through direct collisions
1500: with the core.
1501:
1502: \begin{figure}
1503: \vspace{5cm}
1504: \special{psfile=f8.eps voffset=-100 hoffset=-30 vscale=35 hscale=35 angle=0}
1505: \caption{\small Distributions of the planetesimal collision probability
1506: $P_{\rm col}$ (a-c) and escape probability $P_{\rm esc}$ (d-f) as
1507: a function of the evolution time. The time scales for
1508: gas accretion are $\tau_{\rm grow}=10^4,10^5,10^6$ yr, respectively.
1509: In the Bondi gas-accretion model, the planetesimals collide with
1510: the protoplanet mainly around the late stage of gas accretion, which
1511: is quite different from that in the linear gas-accretion
1512: model. \label{fig8}}
1513: \end{figure}
1514:
1515: Throughout the evolution, $T_{\rm open} > \tau_{\rm grow}$, so that the
1516: expansion of the feeding zone outpaces the clearing process (Figs. 3-4).
1517: Consequently, major epochs for planetesimal collisions with the
1518: protoplanet occur around $t=\tau_{\rm grow}$ in Bondi gas-accretion
1519: model (Fig.8). The duration of the epoch of intense bombardment is
1520: $\sim 10^{\log (\tau_{\rm grow}) \pm 0.25}$ for $\tau_{\rm grow} \ge
1521: 10^{5}$ yr. In contrast, major collision events would occur
1522: much earlier if the protoplanet mass increases according to the
1523: hypothetical linear gas-accretion prescription. Most of these
1524: collisions occur at $ t \le \tau_{\rm grow}$. The dichotomy between
1525: these two types of accretion arises because the Bondi
1526: prescription leads to a runaway process, in which most of
1527: the protoplanet mass is attained only at the very end when
1528: $t=\tau_{\rm grow}$. However, the hypothetical linear accretion
1529: introduces a much earlier ramp up in the
1530: protoplanets mass. Consequently its physical radius, gravitational
1531: perturbation, and width of feeding zone also grow quickly, inducing
1532: an earlier phase of intensified collisions.
1533:
1534: Also note that both $P_{\rm esc}$ and $P_{\rm col}$ are normalized to the
1535: initial planetesimal population in the entire computational domain
1536: which covers twice the width of the asymptotic feeding zone of the
1537: protoplanet. The total cumulative statistics suggest that,
1538: the fraction of the
1539: original planetesimal population in the feeding zone which collides
1540: with the protoplanet is comparable to that scattered to the outer disk
1541: regions. With both prescriptions, the ejections of planetesimals
1542: occur around $t \ge \tau_{\rm grow}$. For a given density, the
1543: radius and surface escape speed $V_{\rm esp}$ of the
1544: protoplanet are
1545: proportional to $M_{\rm p}^{1/3}$. During the initial growth stages of the
1546: protoplanet, its $V_{\rm esp}$ is small compared with the local
1547: Keplerian velocity. Scattering from grazing encounters excite the
1548: planetesimal eccentricities rather than eject them. At an advanced
1549: growth stage of the protoplanet(when its $M_{\rm p} \sim M_J$), however,
1550: scattering with impact parameter larger than a few planetary radii
1551: can lead to large-angle deflections and the escape of planetesimals
1552: (Lin \& Ida 1997). The ratio of $P_{\rm esc}/P_{\rm col}$ would increase with
1553: the semi-major axis ($a_{\rm p}$) and effective density ($\rho_{\rm p}$)
1554: of the protoplanet(Ida \& Lin 2004).
1555: The latter quantity is likely to increase after the gas accretion onto
1556: the protoplanet envelope is quenched by its tidal barrier
1557: (Dobbs-Dixon et al. 2007).
1558:
1559:
1560: \begin{figure}
1561: \vspace{5cm}
1562: \special{psfile=f9.eps voffset=-90 hoffset=-20 vscale=32 hscale=32 angle=0}
1563: \caption{\small Dependence of
1564: collided planetesimal mass on $\tau_{\rm grow}$ and gas accretion model.
1565: (a) Evolution of the planetesimal
1566: mass colliding with the protoplanet during its gas accretion.
1567: (b) Variation of collided solid mass with the
1568: time scale of gas accretion $\tau_{\rm grow}$ in the two different
1569: models. (c) A comparison of the survived planetesimals with mass
1570: $m=10^{23}$ g in Bondi and linear gas-accretion models during the
1571: evolution. (d) Evolution of solid and gaseous compositions of the
1572: protoplanet. In this paper, we suppose the solid and gaseous disks
1573: have a surface density two times and one time the minimum solar
1574: nebula, respectively.
1575: \label{fig9}}
1576: \end{figure}
1577:
1578: We now scale our model with the minimum mass nebula model. By adopting
1579: the same radial dependence but twice the magnitude in surface density
1580: as that in the minimum mass nebula, the total solid mass in the region
1581: $a\in [3.2,8.5]$ AU is $31.5 ~M_{\oplus}$ in
1582: our models(\S 2.3). We deduce the rate of accretion by multiplying this total
1583: solid mass with the planetesimal accretion probability. Figure 9a
1584: shows the solid mass ($M_{\rm col}$) that collides with the protoplanet
1585: as a function of its growth time scale ($\tau_{\rm grow}$). With
1586: Bondi accretion, the magnitude of $M_{\rm col}$ attains
1587: a maximum value with $ \tau \in [10^5, 10^6]$ yr (Fig. 9b), which is
1588: $6 \sim 7 ~M_{\oplus}$. (Once again, these values are
1589: applied to compact protoplanets and should be regarded as lower
1590: limits.) With the same $\tau_{\rm grow}$, the collided solid mass in
1591: the linear accretion prescription is $1\sim 2 ~M_{\oplus}$ less
1592: because the more rapid initial growth of the protoplanet leads to
1593: early excitation and evacuation of the feeding zone. This rapid
1594: initial ramp up of the protoplanet mass causes many planetesimals to
1595: drift inward due to gas drag or be ejected before the feeding zone
1596: attains its asymptotic width (Fig.9c). In the limit of large
1597: $\tau_{\rm grow}$, a protoplanet with Bondi gas-accretion model can
1598: accrete planetesimals more effectively.
1599:
1600: Figure 9d shows the evolution of solid (planetesimals) and gas
1601: (including dust grains) component accreted by the protoplanet. The
1602: protoplanet has more solid mass (including the initial core which is
1603: $\sim 6 M_\oplus$) in the beginning of Bondi accretion due to the
1604: inefficient gas accretion. But after $M_{\rm p}=M_{\rm sod} + M_{\rm gas} > 7
1605: M_\oplus$, the gas accretion rate far exceeds that of the planetesimal
1606: accretion rate because $T_{\rm open} > \tau_{\rm grow}$. While the accreted
1607: gas may carry small dust particles with it, the suppression of initial
1608: planetesimal accretion reduces the feedback effect which limits the
1609: gas accretion rate. This late addition
1610: of planetesimals can lead to the enrichment of protoplanet envelope
1611: to super-solar metallicity.
1612:
1613:
1614: \begin{figure}
1615: \vspace{5cm}
1616: \special{psfile=f10.eps voffset=-80 hoffset=-20 vscale=30 hscale=30 angle=0}
1617: \caption{\small Dependence of collision rate on planetesimal mass ($m$).
1618: (a) Evolution of the total solid mass collided onto
1619: the protoplanet with individual planetesimal mass
1620: ranging from $10^{17}-10^{27}$ g in the Bondi gas-accretion model of
1621: $\tau_{\rm grow}= 10^5$ yr. (b) Dependence of the collided
1622: planetesimals' mass on their individual masses.
1623: \label{fig10}}
1624: \end{figure}
1625:
1626: We now examine the dependence of planetesimal accretion efficiency on
1627: their mass $m$. This dependence arises because the eccentricity
1628: damping time scale is a function of $m$ (see Fig.1). Figure 10 shows
1629: the evolution of collided solid mass with individual planetesimal mass
1630: ranging from $10^{17}\sim 10^{27}$ g. The maximum collision rate occurs
1631: with individual planetesimal mass $m\in 10^{20\sim 26}$ g,
1632: which corresponds to planetesimals with radius of $50\sim 5000 $
1633: km. These planetesimals also have the largest $\tau_{\rm aero}$ and
1634: $\tau_{\rm tidal}$. The planet with smaller density ($\rho =0.125$ g cm$^{-3}$) has a bigger collision rate.
1635: In Fig. 11, we plot the
1636: probabilities of the planetesimals survived, collided, ejected or
1637: crashed after $2\times 10^6$ yr in a Bondi accretion model with
1638: $t_{\rm grow}=10^5 $ yr. We find that at least $~2/3$ of
1639: planetesimals have survived at an epoch $\sim 10 t_{\rm
1640: grow}$. In this model, the initial width of the planetesimal disk is
1641: $15 a_{\rm p} h_f$ out of which a planetesimal gap with a width $\sim 7 a_{\rm p} h_f$
1642: is evacuated. Although the final distribution of the planetesimal disk
1643: is more extended (see Fig.5), most of the surviving planetesimals $(>2/3)$
1644: are located within a range of nearly $ 8 a_{\rm p} h_f$ from the protoplanet.
1645: Thus, the density outside the feeding zone is slightly enhanced on
1646: average after the protoplanet obtains its asymptotic mass.
1647:
1648:
1649:
1650: \begin{figure}
1651: \vspace{5cm}
1652: \special{psfile=f11.eps voffset=-80 hoffset=-20 vscale=30 hscale=30 angle=0}
1653: \caption{\small
1654: Probability of survivers, colliders, ejectors($a>100$ AU) and
1655: Sun cashers ($a<1$ AU) as functions of the planetesimals' mass $m$ in the range
1656: $10^{17}-10^{27}$ g at $t=2\times 10^6 $yr.
1657: The gas-accretion model is Bondi accretion. The model parameters are
1658: $\tau_{\rm grow}= 10^5$ yr, protoplanet density $\rho=0.125$ g cm$^{-3}$.
1659: \label{fig11}}
1660: \end{figure}
1661:
1662: \subsection{Resonant trapping}
1663: \label{sec:resonance}
1664:
1665: In both Figures 3 and 4, the accumulation of planetesimals near the
1666: outer mean motion resonances of protoplanet is noticeable, particularly
1667: for the intermediate-mass Model 2. The first-order mean motion
1668: resonances are located both interior and exterior to $a_{\rm p}$ where the
1669: period ratio can be expressed either as $(p+1):p$ or $p:(p+1)$.
1670: The corresponding ratio of semi-major axes with $a_{\rm p}$ is $\alpha =
1671: (1+1/p)^{-2/3} $ or $(1+1/p)^{2/3}$ respectively. In each
1672: case, the protoplanet mass grows from 5.6$M_\oplus$ to 300 $M_\oplus$. With its
1673: asymptotic mass, the 2:1 and 3:2 resonances of the protoplanet are outside
1674: its feeding zone, whereas mean motion resonances up to 10:9 are beyond
1675: $2 \sqrt{3} h$ at the onset of the simulations.
1676:
1677: Using a simple pendulum model (\S8, Murray \& Dermott 1999), the width
1678: and libration time scale associated with these mean motion resonances
1679: can be derived as
1680: \beq
1681: \Delta_{\rm res} = \left( {16 R_r (\alpha) \mu
1682: e_{\rm res} \over 3} \right)^{1/2} a_{\rm p} ,
1683: \eeq
1684: \beq
1685: \tau_{\rm res} = \left( {3 p^2 e_{\rm res} R_r (\alpha) \mu }
1686: \right)^{-1/2} P_{\rm k},
1687: \eeq
1688: where $P_{\rm k}$ is the Keplerian period, and the magnitude of $R_r(\alpha)=\alpha
1689: [p+1+\frac12 \alpha D] b^{p+1}_{1/2}$ is
1690: an increasing function of $p$. The minimum value of eccentricity in the
1691: resonance is $< \Delta e>$ (see Eq. [\ref{de}]).
1692: During the resonant passage, an adiabatic invariant constrains
1693: the eccentricity change to be
1694: \beq
1695: \Delta e_{\rm res}^2 = {\Delta_{\rm res} \over p a_{\rm p}} .
1696: \eeq
1697: In the limit that $ \Delta e_{\rm res} $ due to resonant
1698: perturbation is larger than $ < \Delta e >$ due to secular
1699: perturbation, it can be substituted by $e_{\rm res}$ so that
1700: \beq
1701: e_{\rm res} \simeq \left( {16 R_r (\alpha) \mu \over 3 p^2}
1702: \right)^{1/3},
1703: \eeq
1704: \beq
1705: \Delta_{\rm res} = \left( {16 R_r (\alpha) \mu \over 3}
1706: \right)^{2/3} {a_{\rm p} \over p ^{1/3}},
1707: \label{eq:reswidth}
1708: \eeq
1709: \beq
1710: \tau_{\rm res} = 12^{-1/3} \left( p R_r (\alpha) \mu
1711: \right)^{-2/3} P_{\rm k}.
1712: \eeq
1713:
1714: Substituting $e_{\rm res}$ into equations (\ref{daaero}) and (\ref{datide}), we
1715: find that the damping of the resonantly excited eccentricity leads to
1716: a characteristic migration time across $\Delta_{\rm res}$, which is
1717: \beq
1718: \tau_x = {\Delta_{\rm res} \over {\dot a_{\rm aero} (e_{\rm res}) }}
1719: = \left( \frac{ 16 p R_r (\alpha) \mu}{3} \right)^{1/3}\frac{\tau_{aero}}{2}
1720: \eeq
1721: for small planetesimals. For the more massive embryos,
1722: \beq
1723: \tau_x = {\Delta_{\rm res} \over {\dot a_{\rm tide} (e_{\rm res}) }}
1724: = { 8\over 5 } p \tau_{\rm tide},
1725: \eeq
1726: which is independent of the protoplanet mass.
1727:
1728:
1729: Planetesimals are trapped in resonances when $\tau_x > \tau_{\rm res}$. As
1730: a planetesimal approaches a protoplanet, it encounters resonances with
1731: increasing $p$ and $R_r(\alpha)$, so the magnitude of $\tau_{\rm
1732: res}$ decrease with little change in the magnitude of $e_{\rm
1733: res}$. In principle, it should be easier for the strong resonance
1734: close to the protoplanet to capture the planetesimals because $\tau_x$
1735: becomes larger than $\tau_{\rm res}$ for sufficiently large
1736: $p$'s. However, for relatively large $p$'s, the normalized distance
1737: separating the $p:(p+1)$ and $(p-1):p$ mean motion resonances,
1738: \beq
1739: {\Delta_{p, p-1} \over a_{\rm p}} \simeq { 2 \over 3 p^{2} },
1740: \eeq
1741: is a decreasing function of $p^2$ but independent of the protoplanet mass.
1742: In contrast, equation (\ref{eq:reswidth}) indicates that the magnitude of
1743: $\Delta_{\rm res}/a_{\rm p}$ increases with $\mu = M_{\rm p}/M_\ast$
1744: and decreases with $p^{1/3}$.
1745: When the protoplanet attains
1746: \beq
1747: \mu > \mu_{p, c} = 2^{-5/2} 3^{-1/2} R_r(\alpha)^{-1} p^{-5/2},
1748: \label{eq:overlap}
1749: \eeq
1750: the width of its $p:(p+1)$ resonance (i.e. $\Delta_{\rm res}$) becomes
1751: larger than the separation between it and the $(p-1):p$ resonance
1752: (i.e. $\Delta_{p, p-1}$). Overlapping resonances generally lead
1753: to dynamical instabilities which excite the eccentricities of the
1754: trapped planetesimals and modify their semi-major axes (Murray \& Dermott
1755: 1999).
1756:
1757: The expression in equation (\ref{eq:overlap}) indicates that the critical mass
1758: for the overlapping resonances is a decreasing function of $p$. During
1759: the growth of the protoplanet, planetesimals located in the initial
1760: inner feeding zone become destabilized and collide with the protoplanet
1761: first. But the planetesimals captured onto the more distant low-order
1762: mean motion resonances may remain trapped during the growth of the protoplanet.
1763: For example, the resonant capture condition is most easily
1764: satisfied for the ``distant'' 3:2 and 2:1 mean motion resonances. As
1765: the protoplanet grows, the resonance
1766: strengthen increases and the libration time scale is reduced. Both
1767: the resonant probability and the characteristic eccentricity of the
1768: resonant planetesimals increase. In the standard model with
1769: $\tau_{\rm grow} = 10^5$ yr, the mass doubling time scale prior to the termination
1770: of growth is $\sim 10^3$ yr, which is comparable to the libration
1771: time scale in the mean motion resonances. Planetesimals that are
1772: loosely bound to the mean motion resonances have longer libration time
1773: scale than $\tau_{\rm res}$. They are shaken by the rapid evolution of
1774: the protoplanet's gravitational potential and cannot respond through
1775: adiabatic adjustments. This impulse leads to a late episode of
1776: planetesimal bombardment, which can introduce metallicity and structure
1777: diversity to the gas giant planets. The impact of this impulsive
1778: shake up is most pronounced in the rapid gas accretion models (see Figs.
1779: 8a and 8d), where the resonant capture becomes ineffective.
1780:
1781: The condition for resonant trapping is also more easily satisfied for
1782: the intermediate mass planetesimals, because their eccentricity damping
1783: and orbital migration time scales are relatively long. These
1784: tendencies are clearly evident in Figs. 3 and 4. During the
1785: epoch of gas depletion, the damping time scales $\tau_{\rm aero}$ and
1786: $\tau_{\rm tide}$ lengthen, which again provide favorable conditions for
1787: the capture of residual planetesimals into the mean motion resonances.
1788:
1789: Finally, the excess density in the mean motion resonances is determined
1790: by the migration rate outside the resonances. Through orbital decay,
1791: planetesimals from the regions outside the first-born protoplanet
1792: congregate near its mean motion resonance. The enhancement of the
1793: local surface density promotes the formation of secondary gas giants.
1794:
1795: \begin{figure}
1796: \vspace{5cm}
1797: \special{psfile=f12.eps voffset=-120 hoffset=-40 vscale=50 hscale=50 angle=0}
1798: %\vspace{-7cm}
1799: \caption{\small
1800: Configuration of tidal drag perturbation on a planetesimal (see Appendix).
1801: \label{fig12}}
1802: \end{figure}
1803:
1804:
1805: \section{Summary and discussion}
1806:
1807: In this paper, we investigate numerically the process and
1808: efficiency of planetesimal accretion onto a growing protoplanet. We
1809: consider the stage after the formation of the protoplanet core and
1810: during the accretion of its envelope. We use a physical Bondi formula
1811: and an {\it ad hoc} linear prescription to evaluate the gas accretion
1812: rate. The seed protoplanet is placed at $5$ AU with initial mass
1813: $5.67~M_{\oplus}$, accreting gas in a time scale $\tau_{\rm
1814: grow}$. The results of our numerical simulation and analysis have
1815: implications for three issues:
1816:
1817: \subsection{Suppression of planetesimal accretion during the onset of
1818: gas accretion.}
1819:
1820: We first examine the initial growth of the protoplanet through gas
1821: accretion (transition from stages of embryo-growth (S1) to quasi-hydrostatic sedimentation
1822: (S2)). The accretion rate of the
1823: gas is suppressed by the bombardment of planetesimals, which generates
1824: heat needed to be redistributed efficiently. In the vicinity of
1825: the core, there are two regions that supply the bullet planetesimals.
1826: Through its secular perturbation, the protoplanet induces
1827: planetesimals within an inner feeding zone to attain radial excursion,
1828: which crosses its orbit during each azimuthal conjunction.
1829: Planetesimals within this band ($\sim 1.5 h a_{\rm p}$) engage in repeated
1830: close encounters. At 5AU, the local Keplerian speed is comparable to the
1831: surface escape speed of the protoplanet, so only a fraction of the close
1832: encounters will lead to physical collisions and the buildup of the core.
1833:
1834: The protoplanet also has an outer feeding zone at $1.5-3.5 h
1835: a_{\rm p}$. Due to the tidal perturbation of the host star and the
1836: protoplanet, planetesimals in this region occasionally cross the orbit
1837: of the protoplanet. However, the frequency of such encounters
1838: decreases with $\delta_a$ and vanishes at the boundary of the feeding
1839: zone. In a gas free-environment, many synodic periods are needed for
1840: the planetesimals in the outer feeding zone to be captured by the
1841: protoplanet. The presence of gas can lead to eccentricity damping,
1842: orbital migration, and diffusion of planetesimals from the outer to
1843: the inner feeding zone. The migration time scale is determined by the
1844: gravitational perturbation of the protoplanet as well as the efficiency of
1845: gas damping. At the onset of the gas-accretion stage, the mass of the
1846: protoplanetary core is relatively low, so the excited
1847: eccentricities of the planetesimals are relatively small.
1848: Consequently the migration rate from the outer to the inner feeding zone is slow.
1849:
1850: The low replenishment rate as well as the modest collision probability imply
1851: that the planetesimal bombardment rate onto the core is much less
1852: frequent than that inferred from the efficient consumption of all
1853: planetesimals engulfed by the expanding feeding zone (as assumed in
1854: the early models of Pollack et al. 1996). The suppression of
1855: planetesimal collisions also lead to a decline in the energy
1856: dissipation rate and a cutoff in the replenishment of grains in the
1857: gaseous envelope of the protoplanet. The elimination of these bottlenecks
1858: for gas accretion shortens the growth time scale for proto-gas-giant
1859: planets.
1860:
1861:
1862:
1863: \subsection{Chemical and structural diversity through late-stage
1864: bombardment}
1865:
1866: A substantial fraction of the super-solar chemical composition in
1867: Jupiter resides in its envelope. We explore the possibility that the
1868: chemical and core-envelope structural diversity may be due to
1869: late-stage bombardment by residual planetesimals. Due to the secular
1870: perturbations of the protoplanet and gas drag, exterior planetesimals
1871: (with $a > a_{\rm p}$) in the outer feeding zone diffuse into the inner
1872: feeding zone and engage in close encounters with the protoplanet. The
1873: interior planetesimals (with $a < a_{\rm p}$) undergo orbital decay out of
1874: the feeding zone. Both effects lead to the clearing of the feeding
1875: zone and the formation of a planetesimal gap.
1876:
1877: In tenuous regions of the disk, protoplanet grows gradually and the
1878: orbits of both planetesimals and the massive embryos decay slowly.
1879: The migration of the external planetesimals and embryos may be stalled
1880: near the outer mean motion resonances of the protoplanet. But during
1881: the growth of the protoplanet, its feeding zone expands and the mean
1882: motion resonances (with high $p$'s) overlap. The
1883: planetesimals accumulated in the resonances become dynamically
1884: unstable. The colliding embryos may penetrate deeply into the
1885: protoplanet envelope and become part of its core.
1886:
1887: However, in the relatively dense inner regions of the disk, the
1888: gas-accretion rate of protoplanet and the orbital decay
1889: rates of the planetesimals are both relatively high.
1890: Embryos pass through the mean motion
1891: resonances without any significant perturbation. Only the
1892: intermediate-mass planetesimals can be captured onto the mean motion resonances
1893: of the protoplanet. When the protoplanet becomes sufficiently
1894: massive to destabilize the resonance-trapped planetesimals, the planetesimals
1895: will collide with the protoplanet. However, as they do not
1896: have adequate mass to survive the passage through its envelope,
1897: they mostly supply the metallicity in the
1898: protoplanet envelope.
1899:
1900: These two possible outcomes may account for the structural diversity
1901: between Jupiter and Saturn. Jupiter formed early in a relatively dense
1902: region just beyond the snow line. Intermediate-mass planetesimal
1903: bombardments occurred during the advanced stage of its progenitor's
1904: growth so that it has a relatively low core mass but a metal-rich
1905: envelope. In contrast, Saturn formed during the depletion epoch in
1906: the relatively tenuous outer regions of the disk. In this case,
1907: massive embryos may be trapped in the mean motion resonances
1908: of the protoplanet when
1909: its mass is a few $M_\oplus$. Their orbits become unstable
1910: when Saturn acquired most of its present-day mass. The late-stage
1911: bombardment by these massive embryos may have contributed to
1912: substantial core of Saturn.
1913:
1914: Generally the longer the time scale of gas accretion, the more
1915: efficient a protoplanet accretes planetesimals. But when the time
1916: scale is comparable to the age of the gaseous disk (millions of years), a
1917: runaway type of gas accretion model like Bondi accretion is preferred
1918: for the protoplanet to acquire more efficient and late-stage
1919: planetesimal accretions (Figs.8-9). In a solid disk with surface
1920: density twice the minimum solar nebula, the solid mass
1921: colliding with a compact protoplanet is $6\sim 7 ~M_{\oplus}$,
1922: and it reaches maximum when $ \tau \in [10^5, 10^6]$ yr in the Bondi
1923: gas-accretion model. This mass is comparable to the initial core mass
1924: and should be regarded as a lower limit.
1925: In our calculations, the radius is calculated
1926: according to equation (\ref{S}). After the protoplanet acquired an atmosphere, the
1927: collision rate could be enhanced (Inaba \& Ikoma 2003) and a higher
1928: solid accretion rate may be expected.
1929:
1930: The accretion rates of planetesimals with different masses are
1931: determined. The mass of individual planetesimal that could be
1932: effectively accreted to the protoplanet lies in the range
1933: $10^{20}-10^{26}$ g, which corresponds to an embryo with
1934: radius of $50-5000 $ km (Fig.10). The relatively low mass
1935: planetesimals disintegrate in the protoplanet envelope, whereas
1936: massive protoplanetary embryos may survive their passage through
1937: the envelope and become a part of the protoplanet core.
1938:
1939: \subsection{The enhanced formation of multiple gas giant planetary
1940: systems}
1941:
1942: We also show that due to the gas drag and protoplanet secular
1943: perturbations, the density of the planetesimal disk could be slightly
1944: increased outside the orbit of the protoplanet, which will enhance the
1945: subsequent formation of external protoplanet cores (Figs.3-5).
1946:
1947: During their orbital decay, all planetesimals, with the
1948: exception of the most massive embryos, migrate sufficiently slowly that
1949: they become locked onto the mean motion resonances of the protoplanet. When
1950: the growth of the protoplanet is stalled, the resonant planetesimals residing
1951: outside the asymptotic feeding zone remain well separated from other
1952: mean motion resonances and are stably trapped in these
1953: resonances. The enhancement of the local surface density reduces the
1954: growth time scale of the embryos. The formation of
1955: second generation proto-gas-giant planets is promoted.
1956:
1957: The emergence of multiple gas giants close to each other's mean motion
1958: resonances may also lead to dynamical instabilities (Zhou et al. 2007).
1959: The resulting dynamical interaction may lead to mergers,
1960: ejections, and breakup of the system. We suggest that this may be a
1961: promising avenue for the generation of the large eccentricity
1962: distribution among the extra solar planets.
1963:
1964:
1965:
1966:
1967: \acknowledgments
1968:
1969: We thank Drs. M. Nagasawa, S. Aarseth and S. Dong for constructive
1970: discussions, and Dr. S. Aarseth for improving the manuscript.
1971: This work is supported by NSFC(10233020, 10778603), NCET (04-0468),
1972: NASA (NAGS5-11779, NNG04G-191G, NNG06-GH45G), JPL (1270927), NSF(AST-0507424, PHY99-0794).
1973:
1974: \appendix
1975:
1976: \section{Perturbations under tidal drag}
1977:
1978: Suppose the perturbing acceleration of tidal drag to a planetesimal
1979: has the form of equation (\ref{tidal}):
1980: \beq
1981: f_{\rm tidal}=-\frac{{\bf V}_k-{\bf V}_{\rm g}}{\tau_{\rm tidal}}
1982: \equiv A({\bf V}_k-{\bf V}_{\rm g}),
1983: \label{tidal2}
1984: \eeq
1985: where $A= 1/\tau_{\rm tidal} $. The velocity of the planetesimal and gas can be expressed in terms of the
1986: radial, azimuthal and normal components with unit vectors $\hat{r},\hat{\psi}$ and $\hat{h}$,
1987: respectively (Fig. 12):
1988: \beq \begin{array}{l}
1989: {\bf V}_k=v_0 (\cos \alpha ~\hat{r} +\sin \alpha ~\hat{\psi} ), \\
1990: {\bf V}_{\rm g}=\chi (\cos \epsilon ~\hat{\psi} +\sin \epsilon ~\hat{h} ),
1991: \end{array}
1992: \eeq
1993: where $\alpha$ is the angle from the radial to the velocity direction in the planetesimal
1994: orbit plane, and $\chi=\sqrt{GM_{\odot}/R}(1-2\eta (R))^{1/2}$. Suppose
1995: the gas is in a circular orbit,
1996: $\eta=0$, we obtain,
1997: \beq \chi=n a (\frac{1+e\cos f}{1-e^2})^{1/2} \cos ^{-1/2} \delta.
1998: \eeq
1999: The definitions of the angles are also shown in Fig.12.
2000: From spherical geometry (Adachi et al. 1976),
2001: \beq
2002: \begin{array}{l}
2003: \sin \epsilon = \cos i /\cos \delta , \\
2004: \cos \epsilon = \sin i \cos (f+\omega) / \cos \delta, \\
2005: \cos \delta = [1-\sin ^2 i \sin ^2(f+\omega )]^{1/2},
2006: \end{array}
2007: \eeq
2008: where $f$ is the true anomaly of the planetesimal orbit.
2009: Thus the perturbing acceleration of tidal drag can be expressed as:
2010: \beq
2011: \begin{array}{rl}
2012: f_{\rm tidal}= & \frac{naA}{\sqrt{1-e^2}}\{ e\sin f \hat{r}+(1+e\sin f)^{1/2}[(1+e\sin f)^{1/2}
2013: -\cos i \cos ^{-3/2}\delta ] \hat{\psi} \\
2014: & +(1+e\sin f)^{1/2} \sin i \cos (f+\omega) \cos ^{-3/2}\delta \hat{h}\} \\
2015: \equiv & \bar{R} \hat{r} +\bar{T} \hat{\psi} +\bar{N} \hat{h}
2016: \end{array}
2017: \eeq
2018: The evolution equations of
2019: the osculating elements under tidal drag obey (Murray \& Dermott 1999):
2020: \begin{equation}
2021: \begin{array}{ll}
2022: \frac{da}{dt} & = \frac{2}{n\sqrt{1-e^2}} [ \bar{R} e\sin f +\bar{T} (1+e\sin f)] \\
2023: &= \frac{2aA}{1-e^2}\{ 1+e^2+2e\cos f-(1+e\sin f)^{3/2}\cos i
2024: [1- \sin^2 i \sin^2 (f+\omega)]^{-3/4} \}\\
2025: \frac{de}{dt} &= \frac{\sqrt{1-e^2}}{na} [\bar{R}\sin f + \bar{T} (\cos f+\frac{\cos f+e}{1+e\cos f})] \\
2026: \ & =2A(e+\cos f) +A(1+e\cos f)^{-3/2}(2\cos f + e\cos^2 f+e)
2027: \cos i [1- \sin^2 i \sin (f+\omega)]^{-3/4} \\
2028: \frac{di}{dt}& = \frac{\sqrt{1-e^2}}{na}\frac{\cos (f+ \omega)}{1+e\cos f}\bar{N} \\
2029: &= A\sin i (1+e\cos f)^{-1/2} [1- \sin^2 i \sin (f+\omega)]^{-3/4} \cos^2 (f+\Omega).
2030: \end{array}
2031: \eeq
2032: We average over time $t$ in one period of Keperian motion according to :
2033: \beq
2034: <F>=\frac{1}{T}\int_{0}^{T}F dt =\frac{1}{2\pi}\int_{0}^{2\pi}\frac{F(1-e^2)^{3/2}}{(1-e\cos f)^{2}} df.
2035: \eeq
2036: This gives:
2037: \beq
2038: \begin{array}{ll}
2039: \frac{1}{a}<\frac{da}{dt}> & = \frac{A}{8}[(5e^2+2 i^2) +o(e^2,i^2)], \\
2040: \frac{1}{e}<\frac{de}{dt}> & = A [(1-\frac{13}{32} e^2-\frac12 i^2+\frac34 i^2 \sin^2\omega)+o(e^2,i^2)], \\
2041: \frac{1}{i}<\frac{di}{dt}> & = \frac{A}{2}[ (1-\frac{13}{32} e^2+\frac{35}{16}e^2\sin^2\omega
2042: +\frac{3}{16}i^2)+o(e^2,i^2)]. \\
2043: \end{array}
2044: \eeq
2045: We further average over one period of $d\omega/dt$ to eliminate the dependence of $\omega$,
2046: and recall $A=-1/\tau_{\rm tidal}$, which finally yields
2047: \beq
2048: \begin{array}{ll}
2049: \frac{1}{a}<\frac{da}{dt}> & = -\frac{1}{8\tau_{\rm tidal}}[(5e^2+2 i^2)+o(e^2,i^2)], \\
2050: \frac{1}{e}<\frac{de}{dt}> & = -\frac{1}{\tau_{\rm tidal}} [(1-\frac{13}{32} e^2-\frac{1}{8}i^2)+o(e^2,i^2)], \\
2051: \frac{1}{i}<\frac{di}{dt}> & = -\frac{1}{2\tau_{\rm tidal}}[ (1+\frac{11}{16} e^2+\frac{3}{16}i^2)+o(e^2,i^2)] . \\
2052: \end{array}
2053: \eeq
2054:
2055:
2056:
2057: \begin{thebibliography}{}
2058: %\bibitem[Aarseth, Lin \& Palmer 1993]{ALP93} Aarseth, S.J., Lin D.N.C., \& Palmer, P.L. 1993, ApJ, 403,351
2059: \bibitem[Adachi, Hayashi \& Nakazawa 1976]{ahn76} Adachi, I., Hayashi, C., \& Nakazawa, K. 1976,
2060: Prog. Theo. Phys. 56, 1756
2061: \bibitem[Artymowicz(1993)]{art93}Artymowicz, P., 1993, \apj, 419,155
2062: %\bibitem[Bodenheimer 1974]{bod74}Bodenheimer, P. 1974, Icarus, 23, 319
2063: %\bibitem[Bodenheimer 1980]{bod80}Bodenheimer, P. 1980, Protostars and Planets II.
2064: % eds. D. C. Black \& M. S. Matthews (Tucson : Univ. Arizona Press), 873
2065: \bibitem[Bodenheimer, Lin, \& Mardling 2001]{blm01}
2066: Bodenheimer, P., Lin, D. N. C., \& Mardling, R. A. 2001,\apj, 548, 466
2067: \bibitem[Bodenheimer \& Pollack 1986]{bp86}Bodenheimer, P., \& Pollack, J. B. 1986, Icarus, 67, 391
2068: \bibitem[Burrows et al. 2000]{bur00}Burrows, A., et al. 2000, \apj, 534, L97
2069: % Burrows, A.; Guillot, T.; Hubbard, W. B.; Marley, M. S.; Saumon, D.; Lunine, J. I.; Sudarsky, D.
2070: \bibitem[Bryden et al. 2000]{bry00} Bryden, G., et al. 2000, \apj, 540, 1091
2071: %\bibitem[Bryden, Lin \& Ida 2000]{bli00}Bryden, G., Lin, D.N.C. ,\& Ida, S. 2000, ApJ, 544, 481
2072: %\bibitem[Cameron 1978]{cam78}Cameron, A.G.W. 1978, Moon Planets, 18, 5
2073: %\bibitem[DeCampli \& Cameron 1979]{dc79}DeCampli, W.M. and Cameron, A.G. 1979, Icarus 38, 367
2074: \bibitem[Dobbs-Dixon, Li \& Lin 2007]{dll07}Dobbs-Dixon, I., Li, S. L., \& Lin, D. N. C. 2007,
2075: Tidal Barrier and the Asymptotic Mass of Proto Gas-Giant Planets, arXiv:astro-ph/0701269
2076: \bibitem[Frank, King \& Raine 2002]{fkr02} Frank, J., King, A., \& Raine, D. 2002,
2077: Accretion power in Astrophysics. (Cambridge: Cambridge Univ. Press)
2078: \bibitem[Franklin {\it et al.} 1980]{fra80}
2079: Franklin, F. A., et al. 1980, Icarus, 42, 271 %Lecar, M., Lin, D. N. C., Papaloizou, J.
2080: \bibitem[Garaud \& Lin 2007]{gl07}Garaud, P., \& Lin, D. N. C. 2007, \apj, 654, 606
2081: \bibitem[Goldreich, Lithwick \& Sari 2004 ]{gls04}
2082: Goldreich, P., Lithwick, Y., \& Sari, R. 2004, \apj, 614, 497
2083: \bibitem[Goldreich\ & Tremaine 1978]{gt78}Goldreich, P., \& Tremaine, S. 1978,
2084: \apj, 222, 850
2085: \bibitem[Goldreich\ & Tremaine 1980]{gt80}Goldreich P., \& Tremaine, S. 1980,
2086: \apj, 241, 425
2087: %Goldreich, Peter; Lithwick, Yoram; Sari, Re'em
2088: \bibitem[Greenzweig \& Lissauer 1992]{gl92}Greenzweig, Y., \& Lissauer, J. J. 1992, Icarus, 100, 440
2089: \bibitem[Guillot, Gautier \& Hubbard 1997]{ggh97}Guillot, T., Gautier,D., \& Hubbard, W. B. 1997, Icarus, 130, 534
2090: \bibitem[Hasegawa \& Nakazawa 1990]{hn90}Hasegawa, M., \& Nakazawa, K. 1990, A\&A, 227, 619
2091: \bibitem[Hayashi, Nakazawa \& Nakagawa 1985]{hay85}Hayashi, C., Nakazawa, K. \& Nakagawa, Y. 1985,
2092: in {\rm Protoplanets
2093: and Planets II.} ed. D. C. Black \& M. S. Mathew (Tucson: Univ. Arizona Press), 1100
2094: \bibitem[Hayashi 1981]{hay81}Hayashi, C. 1981, Prog. Theo. Phys. Suppl. 70, 35
2095: \bibitem[Hayashi, Nakazawa \& Adachi 1977]{kna77} Hayashi, C., Nakazawa K. \& Adachi, I. 1977,
2096: Publ. Astron. Soc. Japan, 29, 163
2097: %\bibitem[Hayashi, Nakazawa \& Nakagawa 1985]{hnn85} Hayashi, C., Nakazawa K.,\& Nakagawa, Y. 1985,
2098: %Protostars and Planets II. eds. D. C. Black \& M. S. Matthews (Tucson : Univ. Arizona Press), 1100
2099: \bibitem[Hubickyj, Bodenheimer \& Lissauer, 2005]{hub05}Hubickyj, O., Bodenheimer, P., \& Lissauer, J. J. 2005,
2100: Icarus, 179, 415
2101: %RevMexAA, 22, 83-86
2102: \bibitem[Ida \& Lin 1996]{il96} Ida, S., \& Lin, D. N. C. 1996, AJ, 112, 1239
2103: \bibitem[Ida \& Lin 2004]{il04} Ida, S., \& Lin, D. N. C. 2004, \apj, 604, 388
2104: \bibitem[Ida \& Lin 2007]{il07} Ida, S., \& Lin, D. N. C. 2007, \apj, submitted
2105: %Toward a Deterministic
2106: %Model of Planetary Formation IV: Core's Migration,
2107: %Self-Regulated Clearing and the late Epoch of Gas Giant Formation,
2108: \bibitem[Ida \& Makino 1993]{im93}Ida, S., \& Makino, J. Icarus, 106, 210
2109: \bibitem[Ikoma, Nakazawa \& Kiyoshi 2000]{ink00}
2110: Ikoma, M., Nakazawa, K., \& Emori, H. 2000, \apj, 537, 1013
2111: %Ikoma, Masahiro; Nakazawa, Kiyoshi; Emori, Hiroyuki
2112: \bibitem[Inaba \& Ikoma 2003] {ii03}Inaba, S., \& Ikoma, M. 2003, A\&A, 410, 711
2113: \bibitem[Kokubo \& Ida 1996]{ki96} Kokubo, E., \& Ida, S. 1996, Icarus, 123, 180
2114: \bibitem[Kokubo \& Ida 1998]{ki98} Kokubo, E., \& Ida, S. 1998, Icarus, 131, 171
2115: \bibitem[Kokubo \& Ida 2000]{ki00} Kokubo, E., \& Ida, S. 2000, Icarus, 143, 15
2116: \bibitem[Kokubo \& Ida 2002]{ki02} Kokubo, E., \& Ida, S. 2002, \apj, 581, 666
2117: \bibitem[Koller, Li \& Lin 2003]{kll03}Koller, J., Li, H., \& Lin, D. N. C 2003, \apj, 596, L91
2118: \bibitem[Kominami \& Ida 2002]{koi02}Kominami, J., \& Ida, S. 2002, Icarus 157,43
2119: %\bibitem[Kuiper 1951]{kui51} Kuiper G. P., 1951, In {\em Astrophysics: A Topical Symposium }(J.A.Hynek, Ed.) pp.357-324.
2120: % McGraw-Hill, New York
2121: %\bibitem[Landau \& Lifthitz 1999]{l999}Landau,L.D., \& Lifthitz, E.M. 1999, Fluid Mechanics, Butterworth-Heinemann
2122: \bibitem[Laughlin, Steinacker \& Adams 2004]{lsa04}
2123: Laughlin, G., Steinacker A., \& Adams, R. 2004, ApJ, 608, 489
2124: \bibitem[Levison \& Duncan 1994]{ld94}Levison, H. F., \& Duncan, M. J. 1994, Icarus 108, 18
2125: \bibitem[Lin , Bodenheimer \& Richardson (1996)]{lbr96}
2126: Lin, D. N. C, Bodenheimer, P., \& Richardson, D. C. 1996, Nature, 380, 606
2127: \bibitem[Lin \& Ida 1997]{li97} Lin, D. N. C., \& Ida, S. 1997 \apj, 477, 781
2128: \bibitem[Lin \& Papaloizou 1979]{pl79}
2129: Lin, D. N. C, \& Papaloizou, J. C. B. 1979, MNRAS, 186, 799
2130: \bibitem[Lin \& Papaloizou 1986]{lp86}
2131: Lin, D. N. C, \& Papaloizou, J. C. B. 1986, \apj, 309, 846
2132: \bibitem[Lin \& Papaloizou 1993]{lp93}
2133: Lin, D. N. C, \& Papaloizou, J. C. B. 1993, in
2134: \rm{Protostars and Planets III}, eds. E. H. Levy \& J.I. Lunine,
2135: (Tucson: Unv. Arizona Press), p.\ 749
2136: \bibitem[Leinhardt \& Richardson 2005]{lr05}
2137: Leinhardt, Z., \& Richardson, D. C. 2005, \apj, 625, 427
2138: \bibitem[Lissauer 1987]{lis87}Lissauer, J. J. 1987, Icarus, 69, 249
2139: \bibitem[Marcy et al. 2005]{mar05}Marcy, G., et al. 2005,
2140: Prog. Theo. Phys. Suppl. 158, 24
2141: \bibitem[Mizuno 1980]{Miz80}Mizuno, H. 1980, Prog. Theo. Phys. 64, 544
2142: \bibitem[Murray \& Dermott 1999]{md99}Murray, C. D., \& Dermott, S. F. 1999, Solar System Dynamcis (Cambridge: Cambridge Univ. Press)
2143: \bibitem[Nelson \& Papaliozou]{np03} Nelson, R. P., \& Papaloizou, J. C. B. 2003, MNRAS, 339, 993
2144: \bibitem[Palmer, Aarseth \& Lin 1993]{pal93} Palmer, P. L., Lin, D. N. C., \& Aarseth, S. J. 1993,
2145: \apj, 403, 336
2146: \bibitem[Perri \&Cameron 1974]{PC74} Perri, F., \& Cameron, A. G. W. 1974, Icarus, 22, 416
2147: \bibitem[Pollack {\it et al.} 1996]{pol96} Pollack, J. B., et al. 1996, Icarus, 124, 62
2148: \bibitem[Rafikov 2001]{raf01}Rafikov, R. R. 2001, AJ, 122, 2713
2149: \bibitem[Rice \& Armitage 2003]{ra03} Rice, W. K. M., \& Armitage, P. J. 2003, \apj, 598, L55
2150: \bibitem[Safronov 1969]{saf69}Safronov, V.S. 1969, Evolution of the Protoplanetary Cloud and Formation of
2151: the Earth and Planets (Moscow: Nauka)
2152: \bibitem[Sato et al. 2005]{sato05}Sato, B., et al. 2005, \apj, 633, 465
2153: \bibitem[Saumon \& Guillot 2004]{sg04}Saumon, D., \& Guillot, T. 2004, ApJ, 609, 1170
2154: \bibitem[Shakura \& Sunyaev 1973]{ss73}Shakura, N. I., \& Sunyaev, R.A. 1973, AAp, 24, 337
2155: %\bibitem[Thommes, Duncan \& Levison 2003]{tdl03} Thommes, E.W., Duncan, M.J., \& Levison, H.F. 2003, Icarus 161, 431
2156: %\bibitem[V$\ddot{o}$lk et al. 1980 ]{volk80}V$\ddot{o}$lk, H.J., Jones, F.C., Morfill, G.E., \& R$\ddot{o}$ser, S. 1980,
2157: %A\&A, 85,316
2158: \bibitem[Tanaka \& Ida 1999]{ti99}Tanaka, H., \& Ida, S. 1999, Icarus, 139, 350
2159: \bibitem[Tanaka, Takeuchi \& Ward 2002]{ttw02}Tanaka, H., Takeuchi, T., \& Ward, W. R.
2160: 2002, \apj, 565, 1257 %Tanaka, Hidekazu; Takeuchi, Taku; Ward, William R.
2161: \bibitem[Ward 1986]{wd86}Ward, W. R. 1986, Icarus, 67, 164
2162: \bibitem[Ward 1989]{wd89}Ward, W. R. 1989, \apj, 336, 526
2163: \bibitem[Ward 1997]{wd97}Ward, W. R. 1997, Icarus, 126, 261
2164: %\bibitem[Weidenschilling 1984]{Wei84} Weidenschilling, S. 1984, Icarus, 60, 553
2165: %\bibitem[Weidenschilling, \& Cuzzi 1992]{WC92} Weidenschilling, S. \& Cuzzi, J.N. 1992, Protostars and Planets III. eds. E. Levy \& M. Matthews (Tucson : Univ. Arizona Press), 1031
2166: \bibitem[Wetherill \& Stewart 1989]{ws89}Wetherill, G. W., \& Stewart, G. R. 1989, Icarus, 77, 330
2167: \bibitem[Wuchterl, Guillot \& Lissauer, 2000]{wgl00} Wuchterl, G., Guillot, T., \& Lissauer,J. J. 2000,
2168: in Protostars and Planets IV eds. V. Mannings, A. P. Boss \& S. S. Russell (Tucson : Univ. Arizona Press), 1081
2169: \bibitem[Zhou, Lin \& Sun 2007]{zls07}Zhou, J. -L., Lin, D. N. C., \& Sun Y. -S. 2007,
2170: % Post-Oligarchic Evolution of Protoplanetary Embryos and the Stability of
2171: % Planetary Systems,
2172: \apj, accepted (astro-ph 0705.2164)
2173:
2174: \end{thebibliography}
2175:
2176:
2177:
2178: \end{document}
2179: