1: \documentclass[12pt,preprint]{aastex}
2: \newcommand{\bm}[1]{\mbox{\boldmath{$#1$}}}
3:
4: \shorttitle{Infall of planetesimals onto growing giant planets}
5: \shortauthors{Shiraishi and Ida}
6:
7: \begin{document}
8: \title{Infall of planetesimals onto growing giant planets: onset of runaway
9: gas accretion and metallicity of their gas envelopes}
10: \author{Masakazu Shiraishi and Shigeru Ida}
11: \affil{Department of Earth and Planetary Sciences, Tokyo Institute of Technology,
12: Ookayama, Meguro-ku, Tokyo 152-8551, Japan}
13: \email{ida@geo.titech.ac.jp}
14:
15: \begin{abstract}
16: We have investigated the planetesimal accretion rate
17: onto giant planets that are growing through gas accretion, using
18: numerical simulations and analytical arguments.
19: We derived the condition for gap opening
20: in the planetesimal disk, which is determined by
21: a competition between the expansion of the planet's Hill radius
22: due to the planet growth
23: and the damping of planetesimal eccentricity due to gas drag.
24: We also derived the semi-analytical
25: formula for the planetesimal accretion rate
26: as a function of ratios of the rates of
27: the Hill radius expansion, the damping,
28: and planetesimal scattering by the planet.
29: The predicted low planetesimal accretion rate due to gap opening
30: in early gas accretion stages quantitatively shows that "phase 2,"
31: which is a long slow gas accretion phase before onset of
32: runaway gas accretion, is not likely to occur.
33: In late stages, rapid Hill radius expansion
34: fills the gap, resulting in significant planetesimal accretion,
35: which is as large as several $M_{\oplus}$ for Jupiter and Saturn.
36: The efficient onset of runaway gas accretion
37: and the late pollution may reconcile
38: the ubiquity of extrasolar giant planets with
39: metal-rich envelopes of Jupiter and Saturn inferred from
40: interior structure models.
41: These formulae will give deep insights into formation of extrasolar gas giants
42: and the diversity in metallicity of transiting gas giants.
43: \end{abstract}
44: \keywords{planetary systems: formation -- solar system: formation}
45:
46: %*+*+*+*+*+*+*+*+*+*+*+*+*+*+*+*+*+*+*+*+*+*+*+*+*+*+*+*+*+*+*+*+*+*+*+*+*+*+*+*+
47: \section{Introduction}
48:
49: Models of the interior structure of Jovian planets in our solar system
50: suggest that Jupiter and Saturn would contain much more amount
51: of heavy elements in their envelopes than that assuming
52: the solar metallicity \citep{SG04}.
53: This may imply that significant amount of planetesimals was accreted
54: onto the planets together with gas accretion from the protoplanetary disk.
55: However, the orbital calculations \citep[e.g.,][]{TI97,ZL07} showed that
56: the coupled effect of excitation of planetesimals' eccentricities
57: due to gravitational scattering by the planet
58: and their damping by aerodynamical and/or dynamical drag
59: tends to open up a gap in the planetesimal disk,
60: resulting in truncation of planetesimal infall
61: onto the planet's gas envelope.
62:
63: \citet{PO96} {\it a priori} assumed the maximally efficient
64: planetesimal accretion
65: during gas accretion phase. As a result of increase in the planet's mass,
66: the width of its feeding zone, which is proportional to
67: cubic root of the mass, expands.
68: They assumed that planetesimals in the expanded zones are accreted
69: with the fastest rate for circular orbits of planetesimals.
70: Their assumption can be consistent with the anticipated metal-rich envelopes
71: of Jupiter and Saturn, but is inconsistent with the eccentricity excitation
72: and gap formation shown by the above orbital integrations.
73:
74: Furthermore, the assumption of the maximal planetesimal accretion
75: results in long "phase 2" that is very inefficient gas accretion
76: phase before onset of runaway gas accretion.
77: As explained in \S 2, envelope contraction starts when core's mass
78: ($M_{\rm c}$) becomes larger than a critical core mass ($M_{\rm c,hydro}$).
79: For $M_{\rm c} > M_{\rm c,hydro}$,
80: pressure gradient no more supports envelope
81: gas hydrodynamically against the increased core's gravity
82: \citep{Mizuno80, IK00}.
83: \citet{PO96} showed that
84: heat generation due to the assumed planetesimal
85: accretion associated with gas accretion
86: supports the envelope quasi-hydrodynamically (in other words,
87: it increases the critical core mass; eq.~[\ref{eq:crit_core_mass}]),
88: after the onset of envelope contraction.
89: The quasi-hydrodynamical state is called ``phase 2''
90: and it may last for more than Myrs.
91: However, the inefficient gas accretion would be inconsistent with the
92: ubiquity of extrasolar giant planets \citep{IL08}.
93:
94: Recently, likelihood of phase 2 is re-addressed.
95: \citet{Fortier07} showed that even if the gap formation is neglected,
96: more realistic planetesimal accretion rate based on oligarchic growth
97: \citep{KI98, KI02, Thommes03} significantly suppresses
98: the duration of phase 2.
99: \citet{ZL07} showed that planetesimals around a protoplanet
100: with $M_{\rm c} \sim M_{\rm c,hydro}$
101: are gravitationally shepherded and cannot be accreted.
102: It suggests non-existence of phase 2.
103: They also showed that the anticipated metal-rich envelopes
104: of Jupiter and Saturn is not inconsistent with it, because
105: such shepherding occurs only in early stages.
106: The planet starts accreting planetesimals
107: when its mass becomes comparable to
108: that of gas giants
109: because the planetesimals trapped in mean-motion
110: resonances are released by resonance overlapping due to
111: the planet's mass increase.
112: The planetesimal accretion no more halts gas accretion
113: onto such a massive planet.
114: Through numerical simulations with two different simple gas accretion
115: prescriptions, they estimated that total accreted mass can be
116: as large as several earth masses.
117:
118: The idea by \citet{ZL07} reconciled the efficient formation of
119: gas giants with the anticipated metal-rich envelope.
120: However, since they showed only numerical results with limited
121: prescriptions for gas accretion, it is not clear
122: that the accreted planetesimal mass for more realistic gas accretion rate
123: is as much as that they obtained.
124: Furthermore, they discussed suppression of phase 2 only
125: qualitatively.
126: As shown below, the planetesimal accretion rate does depend on gas accretion
127: speed as well as the planet's mass.
128: Hence, for evaluation of amount of planetesimal infall for
129: more realistic gas accretion models and quantitative
130: discussion on the suppression of "phase 2," general formulae
131: for gas accretion rate as a function of the planet's mass ($M$)
132: and its increase rate ($\dot{M}$) is needed.
133:
134: In the present paper, through orbital integrations,
135: we clarify the physical mechanism to determine
136: the planetesimal accretion rate and
137: derive detailed semi-analytical formulae for
138: the accretion rate as a function of $M$ and $\dot{M}$.
139: We find that the total infall mass of planetesimals can be
140: as much as several earth masses even for more realistic
141: gas accretion models
142: and quantitatively show that phase 2 is unlikely to occur.
143:
144: The outline of this paper is as follows.
145: We summarize gas accretion processes onto planets in \S 2.
146: The method of our calculation and initial setup is described in \S 3.
147: With artificial simple gas accretion models,
148: we clarify intrinsic physics that determines
149: the planetesimals accretion rate and
150: derive semi-analytical formulae for the accretion rate (\S 4.1 to 4.3).
151: Applying the formulae to realistic gas accretion models,
152: we discuss the metallicity of Jupiter and Saturn
153: envelope (\S 4.4).
154: We also discuss ``phase 2'' and find
155: that phase 2 is not likely to occur (\S 4.5).
156: The conclusion is in \S 5.
157:
158: %*+*+*+*+*+*+*+*+*+*+*+*+*+*+*+*+*+*+*+*+*+*+*+*+*+*+*+*+*+*+*+*+*+*+*+*+*+*+*+*+
159:
160: \section{Gas Accretion Onto a Core}
161:
162: As mentioned in \S 1,
163: when the core mass becomes larger than a critical core mass,
164: pressure gradient no more supports envelope
165: gas hydrodynamically against the core's gravity and
166: hydrostatic atmosphere does not exist.
167: After that, heat generation due to gas envelope
168: contraction itself supports the envelope against
169: dynamical collapse, thus the envelope undergoes
170: quasi-static contraction.
171: The contraction allows gas inflow from the disk
172: into Bondi radius of the planet, so that the contraction
173: rate is almost equal to gas accretion rate of the planet.
174:
175: Here, we briefly summarize the
176: prescriptions for this process for later use.
177: The critical core mass depends on planetesimal
178: accretion rate onto the core ($\dot{M}_{\rm c}$)
179: and the grain opacity ($\kappa_{\rm gr}$) associated
180: with the disk gas. Based on a series of numerical models,
181: \citet{IK00} found that the critical core mass
182: for break-down of hydrostatic atmosphere is
183: \begin{equation}
184: M_{\rm c,hydro} \simeq
185: 10 \left( \frac{\dot{M}_{\rm c}}{10^{-6}M_{\oplus} {\rm yr}^{-1}}\right)^{0.2-0.3}
186: \left( \frac{\kappa_{\rm gr}}{\kappa_{\rm gr}^{\rm P}}\right)^{0.2-0.3}M_{\oplus},
187: \label{eq:crit_core_mass}
188: \end{equation}
189: where $\kappa_{\rm gr}^{\rm P}$ $(\sim 1{\rm cm^2 g^{-1}})$
190: is the grain opacity given by \citet{PO85},
191: who assumed dust grains with interstellar
192: abundance and size distributions.
193: Faster accretion and higher opacity (relatively
194: large $\dot{M}_{\rm c}$ and $\kappa_{\rm gr}$)
195: result in a warmer planetary atmosphere and
196: an enhanced pressure gradient, so $M_{\rm c,hydro}$ is larger
197: \citep{Stevenson82,IK00}.
198:
199: \citet{PO96} assumed the most efficient planetesimal accretion
200: induced by expansion of feeding zone due to increase of the
201: planet mass,
202: the rate of which is $\sim 10^{-6}M_{\oplus} {\rm yr}^{-1}$,
203: for $M_{\rm c} \sim 10 M_{\oplus}$.
204: When the planet with $M_{\rm c} \sim 10 M_{\oplus}$ becomes
205: isolated consuming planetesimals in its feeding zone,
206: gas envelope starts contraction and the induced
207: planetesimal accretion from the expanded region of the
208: feeding zone increases $M_{\rm c,hydro}$
209: up to $\sim M_{\rm c}$ (eq.~[\ref{eq:crit_core_mass}])
210: and stall gas accretion. This self-regulated process works
211: on more than Mrys until
212: $M_{\rm c}$ exceeds $\sim 20 M_{\oplus}$.
213: This is called ``phase 2.''
214: However, as we show in \S 4.5,
215: the rate of the planetesimal accretion
216: induced by gas accretion is not generally large enough to
217: maintain phase 2.
218: Then, gas accretion dominant phase starts.
219:
220: For $M_{\rm c} \sim M_{\rm c,hydro}$,
221: heat generation due to planetesimal accretion
222: marginally equilibrates with the core's gravity.
223: In the quasi-static contraction stage, heat generation
224: due to {\it envelope contraction}
225: marginally equilibrates with the gravity of the planet
226: with total mass $M$ (including envelope mass).
227: The Kelvin-Helmholtz contraction timescale
228: is equivalent to planet mass increase timescale
229: $\tau_{\rm g,acc} = M/\dot{M}$.
230: Replacing $M_{\rm c}$ and $\dot{M}_{\rm c}$ by
231: $M$ and $M/\tau_{\rm g,acc}$ in eq.~(\ref{eq:crit_core_mass}),
232: $\tau_{\rm g,acc}$ is given by
233: \begin{equation}
234: \tau_{\rm g,acc} \simeq
235: 10^7 \left( \frac{M}{10M_{\oplus}}\right)^{-({\rm 2.3-4})}
236: \left( \frac{\kappa_{\rm gr}}{\kappa_{\rm gr}^{\rm P}}\right) \; {\rm yrs}.
237: \label{eq:tau_g,acc}
238: \end{equation}
239: Detailed numerical simulations of quasi-static evolution of
240: the gaseous envelope \citep{IK00,IG06} show consistent results
241: at the onset of runaway gas accretion in which the envelope and
242: core masses are nearly equal.
243: Although \citet{PO03} suggested $\kappa_{\rm gr}
244: \sim 0.01\kappa_{\rm gr}^{\rm P}$ through the numerical
245: simulations of coagulation and sedimentation of dust grains in the atmosphere,
246: the amount and size distribution of dust grains in the atmosphere are
247: highly uncertain.
248: Here, we adopt the results by \citet{IG06}
249: with $\kappa_{\rm gr}=\kappa_{\rm gr}^{\rm P}$,
250: \begin{equation}
251: \tau_{\rm g,acc} =
252: 10^{6.5} \left( \frac{M}{10M_{\oplus}}\right)^{-3.5}
253: \textrm{yrs} \, ,
254: \label{eq:ikoma_model}
255: \end{equation}
256: as a fiducial ``realistic'' gas accretion model.
257:
258: When eq.~(\ref{eq:ikoma_model}) is extrapolated to
259: large $M$ ($\ga$ 100 M$_\oplus$), it may give
260: unrealistically fast supply of gas from the disk.
261: Hence, we limit the gas accretion rate as bellow.
262: \citet{TW02} showed through two-dimensional local hydrodynamic simulations,
263: the mass infall to the circumplanetary subdisk
264: from the protoplanetary disk is limited by
265: \begin{equation}
266: \frac{\dot{M}}{M} \simeq 6 \times 10^{-4} f_g
267: \left( \frac{a}{5\textrm{AU}} \right)^{-1.5}
268: \left( \frac{M}{M_{\oplus}} \right)^{0.3} \textrm{yr}^{-1} \, ,
269: \label{eq:tanigawa_model}
270: \end{equation}
271: where $f_g$ is a scaling factor for disk gas surface density
272: defined by eq.~(\ref{eq:surface_gas_density}).
273: We use this limit with $f_g = 0.7$.
274:
275: Another limit is Bondi gas accretion, the rate of which
276: is given by $\dot{M} = \pi r^2_B \rho_{gas} c_s$,
277: where $\rho_{gas} = \Sigma_{g}/(2H)$ is the spatial
278: density of gas disk and $H$ is the disk scale hight,
279: $r_B = 2GM/c_s^2$ is the Bondi radius,
280: $c_s = H\Omega$ is the sound speed and
281: $\Omega = \sqrt{GM_{\odot}/a^3}$
282: is the Keplerian angular velocity.
283: Adopting the temperature distribution in the limit of
284: an optically thin disk (Hayashi 1981),
285: $T = 2.8\times 10^2 (r/\textrm{1AU})^{-1/2} \, \textrm{K}$,
286: the Bondi gas accretion limit is
287: \begin{equation}
288: \frac{\dot{M}}{M} = 0.7 \times 10^{-3}
289: \left( \frac{a}{\textrm{5AU}} \right)^{-2}
290: \left( \frac{M}{M_{\oplus}} \right) \, \textrm{yr}^{-1} \, .
291: \label{eq:bondi_model3}
292: \end{equation}
293: As figure \ref{fig:Increase_timescale} shows,
294: the timescale ($=M/\dot{M}$) in eq.~(\ref{eq:tanigawa_model})
295: is generally longer than the Bondi accretion timescale,
296: so an actual lower limit for gas accretion timescale
297: is given by eq.~(\ref{eq:tanigawa_model}).
298:
299:
300: \section{Calculation Setup}
301: \subsection{Orbital Integration}
302: %
303: We numerically calculate the orbital evolution of
304: a swarm of planetesimals in the vicinity of a protoplanet's orbit
305: embedded in a gaseous disk.
306: The protoplanet grows accreting gas with a given rate.
307: The planetesimals are treated as massless test particles
308: and neglect their interactions.
309: We assume that the protoplanet has a fixed circular orbit.
310:
311: The planetesimals' orbits are affected by the gravitational force
312: of the growing protoplanet and drag force from disk gas,
313: \begin{equation}
314: \bm{f}_{\rm gas} = - \frac{\bm{v} - \bm{v}_{\rm gas}}{\tau_{\rm damp}}\, ,
315: \label{eq:gas_drag}
316: \end{equation}
317: where $\bm{v}$ and $\bm{v}_{\rm gas}$ are the velocity of a planetesimal
318: and disk gas.
319: The gas motion is a circular Keplerian motion.
320: In some runs, we adopted slightly slower rotation speed of
321: the disk gas due to radial pressure gradient in disk gas
322: \citep[e.g.,][]{AD76}, which induces inward migration of
323: planetesimal orbits.
324: However, we found that the inward migration hardly changed
325: the results of planetesimal accretion rate onto the protoplanet.
326: We here show the results without the inward migration.
327: We set a damping timescale of the gas drag,
328: $\tau_{\rm damp}$ $(= e/\dot{e})$,
329: as a constant parameter for all the planetesimals throughout a run
330: in order to make clear the effect of the damping force.
331:
332:
333: We follow the prescription of gas surface density distribution
334: by \citet{IL04},
335: \begin{equation}
336: \Sigma_g = 210 f_g \left( \frac{a}{\textrm{5AU}} \right)^{-3/2} \ \textrm{g$\,$cm$^{-2}$} \, ,
337: \label{eq:surface_gas_density}
338: \end{equation}
339: where $f_g$ is the scaling parameter and
340: $f_g = 0.7$ corresponds to the gas surface density
341: of the minimum mass solar nebular model, $\Sigma_{g,{\rm MMSN}}$ \citep{HA81}.
342: For simplicity, we neglect a gap in the gas disk, which may
343: be opened up by the perturbations from a massive
344: protoplanet \citep[e.g.,][]{LP93}
345: and assume the above unperturbed $\Sigma_g$ everywhere.
346: With this $\Sigma_g$, a given value of $\tau_{\rm damp}$ corresponds to
347: individual planetesimal mass \citep{AD76,TI99},
348: \begin{equation}
349: m = 3 \times 10^{17} f_g^3 \left( \frac{e}{0.1} \right)^3
350: \left( \frac{\tau_{\rm damp}}{10^5\textrm{yr}} \right)^3
351: \left( \frac{\rho_{\rm pl}}{1\textrm{g$\,$cm$^{-3}$}} \right)^{-2}
352: \left( \frac{a}{\textrm{5AU}} \right)^{-39/4} \ \textrm{g} \, ,
353: \label{eq:m_pl_gas}
354: \end{equation}
355: where $a$, $e$, and $\rho_{\rm pl}$ are semi-major axis,
356: eccentricity, and material density of planetesimals, respectively.
357: If gravitational drag \citep[e.g.,][]{TW04} is considered
358: in stead of aerodynamical gas drag,
359: \begin{equation}
360: m = 4.5 \times 10^{26} f_g^{-1}
361: \left( \frac{\tau_{\rm damp}}{10^5\textrm{yr}} \right)^{-1}
362: \left( \frac{a}{\textrm{5AU}} \right)^{2} \ \textrm{g} \, ,
363: \label{eq:m_pl_gas2}
364: \end{equation}
365: although in this case, interactions among the planetesimals
366: could be important.
367:
368: The orbits of planetesimals are numerically integrated by using the fourth-order Hermite scheme \citep{MA92} with the hierarchical timestep \citep{MA91}.
369: The equation of motion of particle $k$ is given by
370: \begin{equation}
371: \frac{d^2\bm{r}_k}{dt^2} = -GM_\odot \frac{\bm{r}_k}{|\bm{r}_k|^3}
372: - \frac{GM (\bm{r}_k - \bm{r})}{|\bm{r}_k - \bm{r}|^3}
373: - \frac{GM \bm{r}}{|\bm{r}|^3} + \bm{f}_{\rm gas}\, ,
374: \label{eq:planetesimals}
375: \end{equation}
376: where $M$ and $\bm{r}$ is the mass and position of the protoplanet.
377: The first term to the last one represent the gravity from the central star,
378: the gravitational perturbation from the protoplanet,
379: the indirect term and the gas drag force, respectively.
380: We set $M_* = M_\odot$.
381:
382: When a planetesimal contacts the surface of the protoplanet,
383: the planetesimal is removed after recording the collision.
384: The planet mass is unchanged.
385: The physical radius of a protoplanet is
386: determined by its mass and internal density $\rho$ as
387: \begin{equation}
388: R = \left( \frac{3M}{4\pi \rho} \right)^{1/3},
389: \label{eq:physical_radius}
390: \end{equation}
391: We set $\rho = 1$gcm$^{-3}$ in all simulations.
392: The dependence of the planetesimal accretion rates
393: on $\rho$ will be discussed in \S 4.4.
394:
395: Although we neglect gravitational forces of planetesimals,
396: mass of planetesimals is specified in order to calculate
397: the amount of mass accretion onto the protoplanet
398: (regarding ``effective'' mass for gas drag force, see below).
399: Assuming equal-mass planetesimals,
400: they are initially distributed in the range
401: $a_{\rm in} < a < a_{\rm out}$ to satisfy the surface mass density
402: \begin{equation}
403: \Sigma_d = 3.8 f_d \left( \frac{a}{\textrm{5AU}} \right)^{-3/2} \textrm{g}\,\textrm{cm}^{-2}\, ,
404: \label{eq:surface_mass_density}
405: \end{equation}
406: where $f_d$ is a scaling factor. As is the case for $f_g$,
407: $f_d = 0.7$ corresponds to MMSN.
408: The inner and outer boundaries are
409: $a_{\rm in} = a_p(1 - 5 h_{\rm f})$ and
410: $a_{\rm out} = a_p(1 + 10 h_{\rm f})$,
411: where $a_p$ is the semi-major axis of the protoplanet and $h_{\rm f}$
412: is the reduced Hill radius for final mass of the planet ($M_{\rm f}$).
413: The reduced Hill radius of a protoplanet $h$ is
414: Hill radius $r_{\rm H}$ scaled by $a_p$,
415: \begin{equation}
416: h = r_{\rm H}/a_p = \left( \frac{M}{3M_*} \right)^{1/3} \, .
417: \label{eq:reduced_Hill_radius}
418: \end{equation}
419:
420: In all numerical simulations, we adopt
421: $a$ = 5AU, $M_{\rm f} = M_J$ (Jupiter mass), and
422: $f_d = f_g$.
423: Accordingly, $h_{\rm f} = 6.8 \times 10^{-2}$,
424: $a_{\rm in} = 3.3$ AU, and $a_{\rm out} = 8.4$AU.
425: We will derive the dependences on $a$, $M_{\rm f}$, and $f_d (= f_g)$
426: by analytical arguments and discuss the results with
427: different parameter values.
428: Total mass of planetesimals within the region
429: $a_{\rm in} < a < a_{\rm out}$ is
430: $\sim 20 f_d M_\oplus$.
431: The number of planetesimals in most runs is $N = 20000$.
432: With an assumption that planetesimals have an equal mass, their
433: individual mass is
434: $m_{pl} \simeq 1.1 \times 10^{-3} f_d M_\oplus
435: = 6.6 \times 10^{24} f_d$ g.
436: In our simulations, we specify
437: $\tau_{\rm damp} = 10^6, 10^5, 10^4$ yrs and $\infty$ (gas-free case),
438: independent of the values of $m_{pl}$.
439: Except for the gas-free case,
440: the above values of $m_{pl}$ is much larger than the
441: values in eq.~(\ref{eq:m_pl_gas}) for the given $\tau_{\rm damp}$,
442: so planetesimals that we use correspond to ``super particles''
443: representing many smaller planetesimals.
444: Since we neglect interactions among planetesimals,
445: such ``super particles'' treatment is not inconsistent.
446: Initial eccentricity and inclination of planetesimals are taken as $e_0$ = $i_0$ = 0.001 for all simulations.
447: Since $e$ and $i$ are quickly pumped up by perturbations from the protoplanet, the choice of $e_0$ and $i_0$ does not affect results.
448:
449: \subsection{Growth of a Protoplanet}
450:
451: Since we consider the phase after isolation of protoplanets,
452: we assume that the growth of the protoplanets
453: is dominated by accretion of surrounding disk gas
454: but not by planetesimals.
455: As we will show later,
456: this assumption is valid, because amount of
457: accreted gas is much larger than the anticipated
458: amount of accreted planetesimals.
459:
460: In \S 2, we described the prescription for gas accretion.
461: In the numerical simulations,
462: we use simple artificial gas accretion models
463: in order to make clear what conditions regulate the planetesimal
464: accretion rate. From the results with the artificial models,
465: we derive semi-analytical formulae for the planetesimal
466: accretion in general forms (\S 4.3).
467: Applying the formulae to the more realistic gas accretion
468: rate in \S 2, we calculate the total mass of planetesimal infall
469: into the envelope of Jupiter and Saturn in \S 4.4.
470:
471: The simple artificial gas accretion models are expressed by
472: \begin{equation}
473: \frac{dM}{dt} \equiv \alpha M^p \, ,
474: \label{eq:arti_model}
475: \end{equation}
476: where $\alpha$ is the integration constant determined by
477: the boundary condition.
478: We set the condition as $M = M_0$ for $t = 0$ and
479: $M = M_{\rm f}$ for $t = t_{\rm f}$.
480: Following \citet{ZL07}, we set the protoplanet at $5$AU
481: with its initial mass $M_0 = 5.67 M_\oplus$ and
482: final mass $M_{\rm f} = M_J$ (Jupiter mass).
483: We adopt $t_{\rm f} = 10^5 $yr for numerical simulation,
484: following their nominal cases.
485: The growth with $p = 2$ and 0 correspond to the Bondi and
486: linear models
487: in \citet{ZL07}.
488: Figure \ref{fig:Artificial_massup} shows the evolution of
489: the mass of a protoplanet by accreting gas
490: for $p = 2, 1, 0, -2$.
491: Here we assume that $M_0 > M_{\rm c,hydro}$.
492: The consistency of the assumption is checked in \S 4.5.
493:
494: \section{Results of Orbital Calculation}
495: \label{sec:result}
496:
497: \subsection{Overall Evolution}
498:
499: Figure \ref{fig:Orbit_snapshot} shows the snapshots of
500: the distributions of planetesimals on the $b$-$e/h$ plane,
501: where $h$ is defined by eq.~(\ref{eq:reduced_Hill_radius}).
502: The scaled orbital separation $b$ is defined by
503: \begin{equation}
504: b = \frac{a - a_p}{ha_p} \, ,
505: \label{eq:barameter}
506: \end{equation}
507: where $a_p$ is semimajor axis of the protoplanet.
508: The protoplanet is fixed at the origin (i.e., $b = 0$ and $e = 0$).
509: The growth rate of the protoplanet $\dot{M}$ is $\propto M^2$.
510: To avoid busy plots, we show only 1000 planetesimals
511: in this figure.
512: We integrate the evolution of planetesimals
513: for $3 \times 10^5$yrs.
514: Since $t_{\rm f} = 10^5$yrs,
515: we set $\dot{M}=0$ for $t=1$--$3 \times 10^5$yrs, which
516: corresponds to termination of gas accretion due to gap formation
517: in the gas disk, although we neglect the effect of
518: gas density depletion on drag force.
519: The damping timescale is $\tau_{\rm damp} = 10^4$yr.
520:
521: In figure \ref{fig:Orbit_snapshot},
522: we also drew the Jacobi energy $E_J$,
523: \begin{equation}
524: E_J = \frac{1}{2}((e/h)^2+(i/h)^2) - \frac{3}{8}b^2 + \frac{9}{2} + O(h).
525: \label{eq:jacobi_energy}
526: \end{equation}
527: In the figure, we also include higher order terms of $h$ in $E_J$.
528: In the circular restricted three-body problem, $E_J$ is conserved
529: between before and after scattering by the protoplanet (on average,
530: both ($e^2+i^2$) and $b^2$ increase).
531: Since only planetesimals with $E_J \geq 0 $
532: can enter the Hill sphere of the protoplanet \citep[e.g.,][]{HA77},
533: we regard the region $E_J > 0$ as the feeding zone of the protoplanet.
534: When $e/h, i/h \la 1$, the width of feeding zone is
535: $b \simeq 2\sqrt{3}$.
536:
537: In the top panel ($t=1000$yrs), the planetesimals in the vicinity of
538: the protoplanet are scattered and their $e$ and $b$ increase
539: along a constant $E_J$ curve.
540: The planetesimal eccentricities $e$
541: are damped in the panel of $t = 3\times 10^4$yrs
542: because of $\tau_{\rm damp} = 10^4$yrs.
543: Since the gas drag damps $e$ keeping $b$ almost constant,
544: all the planetesimals except for those trapped in horseshoe orbits
545: go out of the feeding zone.
546: As the protoplanet grows up, its feeding zone expands.
547: Since $b \propto M^{-1/3}$,
548: $b$ of planetesimals decreases with time, but
549: scattering opposes it.
550: Since $\dot{M} \propto M^2$, the expansion accelerates with time.
551: Eventually, the expansion overwhelms the gap opening due to a coupling
552: effect of scattering and gas drag damping, so that
553: planetesimals go into the feeding zone in the
554: panel of $t = 1\times 10^5$yrs.
555: After $t = 1\times 10^5$yrs, the increase of $M$ stopped,
556: so a gap in the planetesimal disk is again produced (the bottom panel).
557: Planetesimals that have sufficiently large $b$ are
558: captured in proper mean motion resonances.
559: Although we neglect inward migration due to slightly
560: slower rotation of gas than Keplerian rotation due to pressure
561: gradient, damping of $e$ results in small decrease in $b$
562: due to angular momentum conservation.
563: Such inward migration causes resonance trapping.
564: The evolution of the gap width is consistent with the
565: result by \citet{ZL07}.
566:
567: Thus, after initial relaxation, planetesimals are shepherded
568: and planetesimal accretion rate is very low, until
569: efficient planetesimal accretion re-starts in late stage.
570: \citet{ZL07} showed through orbital simulation of planetesimals
571: that planetesimal accretion occurs only in late stage of gas accretion.
572: They suggested that
573: in late stage, planet's mass becomes large and
574: the mean motion resonances overlap to release planetesimals
575: captured in the resonances.
576: So, they concluded that mass of the protoplanet controls
577: the accretion rate of planetesimals onto the protoplanet.
578: This effect indeed determines supply of planetesimals to
579: the regions near the feeding zone.
580: However, whether the planetesimals near the feeding zone
581: are shepherded or not is determined by gap opening
582: that is a result of competing processes of the feeding zone
583: expansion and scattering/damping, so
584: the values of $\dot{M}$ play an important role as well as $M$.
585:
586: \subsection{Dependence of Planetesimal Accretion Rate on Planet's Mass}
587: \label{subsec:evolution_of_planetesimals_accretion_rate}
588:
589: The evolution of the planetesimal accretion rate
590: for $\tau_{\rm damp} = 10^6, 10^5, 10^4$ yrs and gas-free case
591: is plotted as a function of $M$ in Figure \ref{fig:mlmp7_mass}.
592: The four lines in each panel represent various
593: gas accretion models ($p = 2, 1, 0, -2$).
594: Initial mass of the protoplanet $M_0$ is set as 5.67 M$_\oplus$.
595: Starting with 20,000 planetesimals
596: (i.e., planetesimal mass $\sim 1.1 \times 10^{-3} f_dM_{\oplus}$),
597: we calculated for $10^5$yrs
598: (the growth timescale $t_{\rm f} = 10^5$yrs).
599:
600: To see the dependence on $M$ more clearly,
601: we plot the scaled planetesimal accretion rate $\dot{M}/R^2$,
602: where $R$ is the physical radius of the protoplanet.
603: Through the numerical simulations, we found that planetesimals are
604: likely to experience 2-D accretion rather than 3-D.
605: The 2-dimensional accretion rate is
606: \begin{equation}
607: \frac{dM}{dt} \sim 2R\Sigma_d \left( \frac{v_{\rm esc}}{v_{\rm rel}} \right) v_{\rm rel} = \sqrt{\frac{32\pi \textrm{G}\rho}{3}}\Sigma_d R^2,
608: \label{eq:size-independent_accretion_rate}
609: \end{equation}
610: where $\Sigma_d$, $v_{\rm esc} = \sqrt{2\textrm{G}M/R}$ and $v_{\rm rel}$ are
611: the surface density of the planetesimals, escape velocity from the
612: protoplanet's surface and relative velocity between the protoplanet
613: and planetesimals, respectively.
614: The scaled accretion rate $(\dot{M}/R^2)$ is
615: determined by effective $\Sigma_d$ in the feeding zone
616: for fixed internal density of the protoplanet ($\rho$).
617: In our simulations, the total mass of the planetesimals is not
618: significantly decreased, so the effective $\Sigma_d$ is
619: determined by scattering by the planet, gas drag, and
620: Hill radius expansion due to the planet growth.
621:
622: Figure \ref{fig:mlmp7_mass} shows that the
623: planetesimal accretion rates for $f_d = 0.7$.
624: The planetesimal accretion rates as a function of $M$
625: depends on the parameter $p$.
626: For $p = -2$ and 0, the scaled planetesimal accretion rate
627: decreases with $M$, which suggests that a gap in the
628: planetesimal disk is formed when $M$ becomes large.
629: On the other hand,
630: for $p=2$, the protoplanet may grow so fast in the late stage
631: that the feeding zone expansion overwhelms the gap formation,
632: as shown in figure \ref{fig:Orbit_snapshot}.
633: The dependence on $p$ implies that
634: the accretion rate is not a function solely of $M$,
635: but depends on $\dot{M}$ as well as $M$,
636: because $p$ determines the $\dot{M}$--$M$ relation.
637:
638: \subsection{Dependence of Planetesimal Accretion Rate on
639: Gap Formation Parameters}
640:
641: Here we show that competition among
642: the feeding zone expansion, scattering, and eccentricity
643: damping regulates flux of planetesimals across the boundaries of
644: the feeding zone ($E_J = 0$), that is, the planetesimal accretion rate.
645: We consider change rates of $b^2$ and $(e/h)^2$ of planetesimals
646: (we neglect the contribution from $i$ because $i$ is usually correlated
647: to $e$ and $i < e$).
648:
649: Evolution of planetesimals on the $b^2$--$(e/h)^2$ space
650: due to gravitational scattering by the protoplanet, damping of
651: eccentricity by gas drag, and expansion of Hill radius by
652: mass increase of the protoplanet is expressed by
653: the change rates, $v_{\rm scat}$, $v_{\rm damp}$,
654: and $v_{\rm H}$, on the plane.
655: Since $b \propto h^{-1} \propto M^{-1/3}$,
656: \begin{equation}
657: v_{\rm H} \equiv \frac{d b^2}{dt}{\rm (growth)}
658: = (bh)^2 \frac{d h^{-2}}{dt}
659: = -\frac{2}{3} b^2 \frac{\dot{M}}{M}
660: \simeq - \frac{8}{\tau_{\rm g,acc}},
661: \label{eq:hill_velocity}
662: \end{equation}
663: where $\tau_{\rm g,acc} = M/\dot{M}$ is the timescale
664: of planet mass increase.
665: In the last equation, we used $b \simeq 2\sqrt{3}$, which is
666: the location of the feeding zone for $e/h \la 1$, for simplicity.
667: With $\tau_{\rm damp} = e/\dot{e}$,
668: \begin{equation}
669: v_{\rm damp} \equiv \frac{1}{2} \frac{d (e/h)^2}{dt}{\rm (damping)}
670: = - \frac{(e/h)^2}{\tau_{\rm damp}}.
671: \end{equation}
672: The factor $(1/2)$ in the definition is added for
673: the more simple form of the final expression and
674: better fit with numerical results.
675: Evolution due to the scattering is increase of $b^2$ and
676: $(e/h)^2$ on average, along a constant $E_J$ curve ($E_J \sim 0$).
677: The corresponding change rate is
678: \begin{equation}
679: v_{\rm scat} \equiv \frac{d b^2}{dt}({\rm scattering})
680: =\frac{4}{3}\frac{d (e/h)^2}{dt}({\rm scattering}).
681: \end{equation}
682: Assuming long-range gravitational interaction with $(e/h) \la 1$,
683: linear calculation \citep{GT82,HN90} showed that
684: $b$ of a planetesimal is increases by
685: $\delta b \simeq 30 b^{-5}$ during each encounter.
686: Numerical calculation showed that for $b \sim 3$--4,
687: $\delta b$ is overestimated by a factor $\sim 10$ \citep{ID90}.
688: Since the scattering occurs at every synodic time
689: $T_{\rm syn} \simeq 2\pi a_p/(\frac{3}{2} b r_{\rm H} \Omega_K)$,
690: \begin{equation}
691: v_{\rm scat} = 2b\frac{d b}{dt}({\rm scattering})
692: \simeq 2 b \frac{0.1\delta b}{T_{\rm syn}}
693: \simeq \frac{6}{b^4} \frac{(3/2)bh}{T_{\rm K}}
694: \simeq 0.22 \frac{h}{T_{\rm K}},
695: \label{eq:scattering_velocity}
696: \end{equation}
697: where $T_{\rm K}=2\pi/\Omega_{\rm K}$ is Keplerian period and
698: $b \simeq 2\sqrt{3}$ is again used.
699:
700: Since the feeding zone is determined by the values of
701: $E_J$ and the scattering does not change the values,
702: the condition of gap opening would be $v_{\rm damp} \ga v_{\rm H}$.
703: If inward migrations of planetesimals due to gas drag or
704: type I migration of the protoplanet is considered
705: but growth of a protoplanet is neglected,
706: the gap formation condition is similarly derived,
707: replacing $v_{\rm H}$ by $db^2/dt$ due to
708: gas drag \citep{TI97} or type I migration \citep{TI99}.
709: These effects can suppress growth of the protoplanets
710: before they attain their isolation masses \citep{TI97,TI99}.
711:
712: When $v_{\rm damp} \la v_{\rm H}$, the gap is not created
713: and planetesimals are engulfed by the expanding feeding zone.
714: The engulfment rate would be determined by
715: $v_{\rm H}/v_{\rm scat}$, because $v_{\rm H}$ and
716: $/v_{\rm scat}$ have opposite
717: directions to each other in the $b^2$ components.
718: Thus, it is expected that for $v_{\rm damp} < v_{\rm H}$,
719: the accretion rate would be regulated by
720: \begin{equation}
721: \xi \equiv | \frac{v_{\rm H}}{v_{\rm scat}} |
722: \simeq 37 h^{-1} \frac{T_{\rm K}}{\tau_{\rm g,acc}}
723: \simeq 4.1
724: \left(\frac{a_p}{5\textrm{AU}} \right)^{3/2}
725: \left( \frac{M}{M_{\oplus}} \right)^{-1/3}
726: \left( \frac{\tau_{\rm g,acc}}{10^{4} \textrm{yrs}} \right)^{-1} \, ,
727: \label{eq:xi}
728: \end{equation}
729: while for $v_{\rm damp} > v_{\rm H}$,
730: the accretion rate would be regulated by
731: \begin{equation}
732: \eta \equiv | \frac{v_{\rm H}}{v_{\rm damp}} |
733: \simeq \frac{8}{(e/h)^2} \frac{\tau_{\rm damp}}{\tau_{\rm g,acc}}
734: \simeq 0.8
735: \left( \frac{\tau_{\rm damp}}{10^{4} \textrm{yrs}} \right)^{1/2}
736: \left( \frac{\tau_{\rm g,acc}}{10^{4} \textrm{yrs}} \right)^{-1}
737: \left( \frac{M}{M_\oplus}\right)^{-1/6}
738: \left(\frac{a_p}{5\textrm{AU}} \right)^{3/4},
739: \label{eq:eta}
740: \end{equation}
741: where we used eq.~(\ref{eq:equili_e}) in Appendix for $(e/h)^2$.
742: Since $\tau_{\rm g,acc}=M/\dot{M}$ and
743: $\tau_{\rm damp}$ and $h$ are functions of $M$,
744: the planetesimal accretion rate would depend on
745: $\dot{M}$ as well as $M$.
746: We show that the numerical results agree with the above
747: argument and derive formulae for
748: the planetesimal accretion rate as a function of $\xi$ and $\eta$.
749:
750: Figure \ref{fig:mlmp7_xsi} shows the evolution of
751: the scaled planetesimal accretion rate as
752: a function of $\xi = v_{\rm H}/v_{\rm scat}$
753: for $p=2,1,0$ and $-2$ in the cases of
754: $\tau_{\rm damp} = 10^6, 10^5, 10^4$ yrs and gas-free case.
755: As suggested in the above discussion,
756: figure \ref{fig:mlmp7_xsi} shows that
757: in the ranges of $\eta > 1$
758: [equivalently, $\xi > 5(\tau_{\rm damp}/10^5\textrm{yr})^{-1/2}
759: (M/M_{\oplus})^{-1/6}(a_p/5\textrm{AU})^{-3/4}$],
760: the scaled planetesimal accretion rate is independent
761: of planet gas accretion
762: models with different $p$ (different $\dot{M}$--$M$ relations) and
763: different $\tau_{\rm damp}$.
764: This confirms that the accretion rate is
765: determined by $\xi$ for $\eta>1$ (non-gap cases).
766: The planetesimal accretion rate in this case is
767: given by
768: \begin{equation}
769: \frac{dM_{\rm solid}}{dt} =
770: 10^{\beta}
771: \left( \frac{R}{R_\oplus}\right)^{2}
772: f_d \left( \frac{v_{\rm H}}{v_{\rm scat}} \right)^{\alpha}
773: \, \textrm{M}_\oplus \, \textrm{yr}^{-1} \, ,
774: \label{eq:fitting_line1}
775: \end{equation}
776: with $\alpha \simeq 0.8$ and $\beta \simeq -6$
777: that are obtained from our numerical results
778: by the least square fitting.
779: The fitting line, eq.~(\ref{eq:fitting_line1}),
780: is expressed by thick solid lines in the plots.
781: For $\eta < 1$, the accretion rate declines, which
782: corresponds to gap opening in the planetesimal disk.
783:
784: Figure \ref{fig:mlmp11_eta} shows the evolution of
785: the scaled planetesimal accretion rate as
786: a function of $\eta = v_{\rm H}/v_{\rm damp}$.
787: In the range of $\eta < 1$, the scaled planetesimal
788: accretion rate is independent of planet accretion
789: models with different $\dot{M}$--$M$ relations and
790: different $\tau_{\rm damp}$.
791: This confirms that the accretion rate is
792: determined by $\eta$ for $\eta<1$.
793: From our numerical results,
794: the planetesimal accretion rate in this case is given by
795: \begin{equation}
796: \frac{dM_{\rm solid}}{dt} =
797: 10^{\beta}
798: \left( \frac{R}{R_\oplus}\right)^{2}
799: f_d \left( \frac{v_{\rm H}}{v_{\rm damp}} \right)^{\alpha}
800: \, \textrm{M}_\oplus \, \textrm{yr}^{-1} \, ,
801: \label{eq:fitting_line2}
802: \end{equation}
803: with $\alpha \simeq 1.4$ and $\beta \simeq -6$.
804:
805: When the planet's mass has grown to $M$,
806: the total mass of planetesimals that infall in the envelope
807: ($M_{\rm solid}(M)$) is
808: obtained by integrating
809: \begin{equation}
810: \frac{dM_{\rm solid}}{dM} = \frac{dM_{\rm solid}}{dt} \frac{\tau_{\rm g,acc}}{M}
811: \label{eq:M_solid}
812: \end{equation}
813: from 0 to $M$.
814: In figure~\ref{fig:Arti_Compare_Amass},
815: $M_{\rm solid}(M)$ evaluated by the above semi-analytical formulae
816: is compared with that obtained by orbital calculations
817: for individual gas accretion models in cases of
818: $\tau_{\rm damp} = 10^6$ yrs, $10^5$ yrs and $10^4$ yrs.
819: The semi-analytical formulae well reproduce the results of
820: orbital calculations except for early stages
821: in which $M_{\rm solid}$ is so small that statistical fluctuation is large.
822: The formulae also reproduce numerical results
823: in their figure~9a in \citet{ZL07}.
824:
825: \subsection{Application to Jupiter and Saturn}
826:
827: In the preceding subsection, we investigated planetesimal accretion
828: onto growing protoplanets with artificial gas accretion models
829: and obtained semi-empirical formulae of the planetesimal accretion rate.
830: Applying this formulae to the more realistic gas accretion models
831: in \S 2, we discuss the metallicity of envelopes of Jupiter and Saturn.
832:
833: Integrating eq.~(\ref{eq:M_solid}) with eq.~(\ref{eq:ikoma_model})
834: to $M_{\rm f}$, we estimate total mass of the accreted planetesimals
835: in the cases of Jupiter ($M_{\rm f} = 318M_{\oplus}$,
836: $a_p = 5.2$ AU) and Saturn ($M_{\rm f} = 95M_{\oplus}$, $a = 9.55$ AU).
837: The evolution of $M_{\rm solid}$ is plotted in figure \ref{fig:Amass_DAMP}.
838: The three curves show
839: the results with $\tau_{\rm damp}=10^4, 10^5$ and $10^6$ yrs.
840: It is likely that gas giants were inflated during
841: gas accretion phase.
842: For a fixed $M$, $dM/dt \propto f_d \sqrt{\rho} R^2 \propto f_d \sqrt{R}$
843: (eq.~[\ref{eq:size-independent_accretion_rate}]).
844: In the figure, we plot the accreted planetesimal mass $M_{\rm solid}^*$
845: for $R=2R_1$ and $f_d=2$, where $R_1$ is
846: the physical radius for mass $M$ and $\rho = 1$gcm$^{-3}$.
847: For other $R$ and $f_d$, the accreted mass is
848: $M_{\rm solid} = (R/2R_1)^{1/2} (f_d/2) M_{\rm solid}^*$.
849:
850: All the results show similar qualitative features of
851: evolution of $M_{\rm solid}$.
852: Planetesimal accretion is inhibited in early stages
853: by gap formation, but
854: rapid planetary growth due to gas accretion in later stages allows
855: planetesimal accretion.
856: With $\tau_{\rm damp}=10^6$ yrs,
857: $M_{\rm solid} \simeq 6 (R/2R_1)^{1/2} (f_d/2) M_{\oplus}$
858: both for Jupiter and Saturn.
859: For shorter $\tau_{\rm damp}$, $M_{\rm solid}$ is smaller
860: due to easier gap formation.
861: We also did calculations starting from different core masses.
862: The resultant $M_{\rm solid}$ hardly changed, because
863: $dM_{\rm solid}/dM$ is negligibly small when
864: $M$ is small and gap is opened.
865: The amount of predicted $M_{\rm solid}$ can be
866: as large as that inferred from the internal structure model \citet{SG04},
867: if the planets are inflated and/or relatively large $f_d$ is
868: considered.
869:
870: For the same $M$, $\rho$ and $f_d$,
871: $M_{\rm solid}$ is larger for larger $a_p$.
872: Although Saturnian mass is $1/3$ of Jovian mass,
873: our model predicts that the mass of planetesimals
874: falling into Saturnian envelope is comparable to
875: that into Jovian envelope.
876: More detailed internal structure models will
877: test our prediction.
878:
879: \subsection{Phase 2}
880:
881: So far, we have assumed that gas accretion immediately
882: starts when $M_{\rm c}$ exceeds $M_{\rm c,hydro}$
883: without undergoing ``phase 2.''
884: In the previous subsection, we predicted
885: the planetesimal accretion rate as a function of planetary mass
886: based on the realistic gas accretion model.
887: With this accretion rate, we show that ``phase 2'' is not
888: likely to occur.
889:
890: In the nominal model (J1 model) in \citet{PO96},
891: $f_d \simeq 2.5$, $a_p = 5.2$AU and $M_{\rm c} \simeq 10 M_{\oplus}$.
892: Then, they found that $\dot{M}_{\rm c} \simeq 10^{-6} M_{\oplus}/$yr is
893: maintained during ``phase 2'' with their maximally
894: efficient planetesimal accretion model.
895: As shown in eq.~(\ref{eq:crit_core_mass}),
896: this $\dot{M}_{\rm c}$ can marginally support gas envelope
897: around a $10 M_{\oplus}$ core.
898:
899: First, we derive the condition for gap opening with
900: a realistic $\tau_{\rm g,acc}$ given by eq.~(\ref{eq:ikoma_model}).
901: Substituting eq.~(\ref{eq:ikoma_model}) into eq.~(\ref{eq:eta}),
902: \begin{equation}
903: \eta \simeq 0.8 \times
904: 10^{-6} \left(\frac{\tau_{\rm damp}}{10^{4} \textrm{yrs}}\right)^{1/2}
905: \left(\frac{M}{M_{\oplus}}\right)^{3.3}
906: \left(\frac{a_p}{5\textrm{AU}}\right)^{3/4}.
907: \label{eq:eta2}
908: \end{equation}
909: With $a_p = 5.2$AU, $M \sim M_{\rm c} \sim 10 M_{\oplus}$ and
910: $\tau_{\rm damp} = 10^6$ yrs, we obtain $\eta \simeq 2 \times 10^{-2} \ll 1$.
911: Then, the gap should be opened up.
912: Our formula for $\eta < 1$ gives
913: \begin{equation}
914: \dot{M}_{\rm solid} \simeq 2.2 \times 10^{-6} f_d
915: \left( \frac{\rho_{\rm p}}{\rm 1gcm^{-3}} \right)^{-1/6}
916: \left( \frac{\tau_{\rm damp}}{10^4{\rm yrs}} \right)^{7/10}
917: \left( \frac{\tau_{\rm g,acc}}{10^4{\rm yrs}} \right)^{-7/5}
918: \left( \frac{M}{M_{\oplus}} \right)^{13/30}
919: \left( \frac{a_p}{5{\rm AU}} \right)^{21/20} \ \rm M_{\oplus}/yr.
920: \label{mdot_gap}
921: \end{equation}
922: Substituting eq.~(\ref{eq:ikoma_model}) into this equation,
923: \begin{equation}
924: \dot{M}_{\rm solid} \simeq 0.9 \times 10^{-14} f_d
925: \left( \frac{\rho_{\rm p}}{\rm 1gcm^{-3}} \right)^{-1/6}
926: \left( \frac{\tau_{\rm damp}}{10^4{\rm yrs}} \right)^{7/10}
927: \left( \frac{M}{M_{\oplus}} \right)^{16/3}
928: \left( \frac{a_p}{5{\rm AU}} \right)^{21/20} \ \rm M_{\oplus}/yr.
929: \label{mdot_gap2}
930: \end{equation}
931: For $\tau_{\rm damp} = 10^6$ yrs, $f_d \simeq 2.5$, $a_p = 5.2$AU
932: and $M \simeq 10 M_{\oplus}$,
933: $\dot{M}_c \simeq 1.1 \times 10^{-7} M_{\oplus}/$yr,
934: which is one order smaller
935: than the planetesimal accretion rate that \citet{PO96} assumed.
936:
937: We examine the possibility of phase 2 for other $f_d $ and $a_p$.
938: For phase 2 to occur, $\dot{M}_{\rm c}$ must be maintained
939: to be as large as $\dot{M}$ for $M_{\rm c} \sim M_{\rm c,hydro}$.
940: Core mass can be approximately identified by core isolation mass
941: beyond the ice line \citep{KI98,KI02,IL04},
942: \begin{equation}
943: M_{\rm c,iso} \simeq 4.6 f_d^{3/2} \left(\frac{a_p}{5{\rm AU}}\right)^{3/4} M_{\oplus}.
944: \end{equation}
945: From eq.~(\ref{eq:crit_core_mass}) with the exponent derived by
946: assuming eq.~(\ref{eq:ikoma_model}), the accretion rate required by
947: occurrence of phase 2 is
948: \begin{equation}
949: \dot{M}_{\rm solid,2}
950: \simeq 10^{-6} \left(\frac{M_{\rm c}}{10M_{\oplus}}\right)^{4.5}
951: M_{\oplus}/{\rm yr}
952: \sim 3 \times 10^{-8} f_d^{6.75}
953: \left(\frac{a_p}{5{\rm AU}}\right)^{3.4} M_{\oplus}/{\rm yr}.
954: \label{eq:mdot_phase_2}
955: \end{equation}
956: Substituting $M_{\rm c,iso}$ into $M$ in eq.~(\ref{eq:eta2}),
957: \begin{equation}
958: \eta \simeq 1.3 \times 10^{-4} f_d^5
959: \left(\frac{\tau_{\rm damp}}{10^{4} \textrm{yrs}}\right)^{1/2}
960: \left(\frac{a_p}{5\textrm{AU}}\right)^{13/4}.
961: \end{equation}
962: So, $\eta < 1$ is equivalent to
963: \begin{equation}
964: f_d < 3.8 \left(\frac{\tau_{\rm damp}}{10^{6} \textrm{yrs}}\right)^{-1/10}
965: \left(\frac{a_p}{5\textrm{AU}}\right)^{-13/20}.
966: \end{equation}
967: For this range of $f_d$ and $a_p > 3$AU (the ice line),
968: eq.~(\ref{mdot_gap}) with
969: $M$ replaced by $M_{\rm c,iso}$ is always smaller than
970: $\dot{M}_{\rm solid,2}$ given by eq.~(\ref{eq:mdot_phase_2})
971: (see figure \ref{fig:Phase2}).
972: For $\eta > 1$, on the other hand,
973: \begin{equation}
974: \dot{M}_{\rm solid} \simeq 1.5 \times 10^{-5} f_d
975: \left( \frac{\rho_{\rm p}}{\rm 1gcm^{-3}} \right)^{-1/6}
976: \left( \frac{\tau_{\rm g,acc}}{10^4{\rm yrs}} \right)^{-4/5}
977: \left( \frac{M}{M_{\oplus}} \right)^{2/5}
978: \left( \frac{a_p}{5{\rm AU}} \right)^{6/5} \ \rm M_{\oplus}/yr.
979: \label{mdot_nogap}
980: \end{equation}
981: In the range of $f_d$ and $a_p$ that satisfy $\eta > 1$,
982: eq.~(\ref{mdot_nogap}) can reach $\dot{M}_{\rm solid,2}$
983: only at $a_p > 15$AU and $f_d \sim 1$,
984: in which gas giant formation is unlikely
985: \citep{IL04}.
986: Thus, the predicted $\dot{M}_c$ never reaches the values
987: required for phase 2.
988: We conclude that phase 2 is not likely to occur for formation of
989: giant planets.
990: This conclusion is consistent with the ubiquity of
991: extrasolar gas giant planets.
992:
993:
994: \section{Conclusion}
995:
996: We have investigated the planetesimal accretion rate onto growing giant planets
997: through numerical simulations and analytical arguments.
998: The planet mass ($M$) is increased with assumed gas accretion rate onto the planet,
999: and orbits of planetesimals in the vicinity of the planet's orbit are
1000: integrated with the effect of gas drag, but without self-gravity of the
1001: planetesimals.
1002:
1003: We first performed simulations with several different
1004: artificial gas accretion rates to clarify intrinsic physics determining
1005: the planetesimal accretion rate.
1006: A gap in the planetesimal disk is opened by a coupling effect of
1007: gravitational scattering by the planet and gas drag damping.
1008: Here, the gap formation means that most planetesimals are get out of the
1009: feeding zone of the planet.
1010: The scattering increases both $e$ and
1011: $b$ keeping Jacobi energy constant,
1012: where $e$ is orbital eccentricity and $b$ is difference in semimajor axis
1013: between the planet and the planetesimals.
1014: Changes in $e$ and $bh$ are of the same order, where $h$ is
1015: reduced Hill radius defined by $(M/3M_\ast)^{1/3}$.
1016: Since the gas drag predominantly damps $e$ after the scattering,
1017: the gap is formed.
1018: On the other hand, the width of the feeding zone is
1019: proportional to $h$.
1020: Thus, the planet growth inhibits
1021: gap formation and competes with the scattering/damping process.
1022:
1023: We derived the condition for the gap formation by
1024: comparison between the eccentricity damping rate ($v_{\rm damp}$)
1025: and the rate of expansion of the feeding zone due to the planet growth
1026: ($v_{\rm H}$).
1027: When $v_{\rm H}/v_{\rm damp} > 1$, the gap is not formed.
1028: Then, the planetesimal accretion rate ($dM_{\rm solid}/dt$) is
1029: scaled by the ratio of the scattering rate
1030: $v_{\rm scat}$ to $v_{\rm H}$.
1031: The numerical results are fitted as
1032: \begin{equation}
1033: \frac{dM_{\rm solid}}{dt} =
1034: 10^{-6}
1035: \left( \frac{R}{R_\oplus}\right)^{2}
1036: f_d \left( \frac{v_{\rm H}}{v_{\rm scat}} \right)^{0.8}
1037: \, \textrm{M}_\oplus \, \textrm{yr}^{-1} \, ,
1038: \end{equation}
1039: where $R$ is physical radius of the planet and $f_d$ is
1040: a scaling factor for surface density of the planetesimals
1041: (eq.~[\ref{eq:surface_mass_density}]).
1042: When the gap is formed ($v_{\rm H}/v_{\rm damp} < 1$),
1043: the accretion rate is significantly depleted.
1044: We found that the accretion rate is scaled by
1045: $v_{\rm H}/v_{\rm damp}$ as
1046: \begin{equation}
1047: \frac{dM_{\rm solid}}{dt} =
1048: 10^{-6}
1049: \left( \frac{R}{R_\oplus}\right)^{2}
1050: f_d \left( \frac{v_{\rm H}}{v_{\rm damp}} \right)^{1.4}
1051: \, \textrm{M}_\oplus \, \textrm{yr}^{-1} \, .
1052: \end{equation}
1053:
1054: Applying these formulae to the more realistic gas accretion models
1055: described in \S 2, we found the followings:
1056: \begin{enumerate}
1057: \item In early stages when $M \sim O(10)M_{\oplus}$, a gap is opened
1058: in the planetesimal disk.
1059: The planetesimal accretion rate is smaller than that required
1060: for phase 2 to occur.
1061: This ensures efficient formation of gas giants, which may be consistent
1062: with the ubiquity of extrasolar giant planets.
1063: \item In later stages ($M \ga O(100)M_{\oplus}$),
1064: the expansion of the feeding zone overwhelms the gap opening process,
1065: so the gap is filled.
1066: Then, the planetesimal accretion becomes efficient.
1067: \item The amount of infalling planetesimals into
1068: the envelopes of Jupiter and Saturn in the late stages can be as large as
1069: several $M_{\oplus}$, which may be consistent with
1070: interior models for these planets.
1071: \end{enumerate}
1072: In this "realistic" model, we assumed that planetesimals
1073: are infinitely supplied.
1074: However, if the accreted mass is significant, planetesimals distributed
1075: in the regions inside isolated strong mean motion resonances can be consumed.
1076: In that case, release of planetesimals from the resonance capture
1077: by resonance overlapping due to planet mass increase may also become
1078: a important factor \citep{ZL07}.
1079:
1080: \citet{Guillot06} pointed out the correlation that
1081: the amount of solid components of extrasolar transiting gas giants increases
1082: with metallicity of their host stars that is proportional to $f_d$.
1083: This trend is consistent with our formulae, because
1084: $dM_{\rm solid}/dt \propto f_d$.
1085: As this example shows, the analysis here
1086: will give deep insights into formation of
1087: extrasolar gas giants and their diversity.
1088:
1089:
1090:
1091:
1092: \acknowledgments %
1093: This work is supported by JSPS.
1094:
1095:
1096: \section*{Appendix}
1097:
1098: The magnitude of $(e/h^2)$ in \S 4.3 is determined by
1099: a balance between damping due to the gas drag and
1100: excitation due to the planet's perturbations.
1101: Since in the non-gap case, planetesimals are engulfed by the
1102: feeding zone mainly through the parameter range of $(e/h) \la 1$,
1103: we use eq.~(\ref{eq:scattering_velocity})
1104: for definition of the parameter $\xi$.
1105: However, gap opening is caused by
1106: damping of relatively high orbital eccentricity, so
1107: we use the formula of excitation of
1108: planetesimal eccentricity due to the protoplanet's
1109: perturbations for $(e/h) \ga 1$, in evaluating $(e/h)^2$.
1110: Then the scattering timescale is given approximately by
1111: Chandrasekahr's two-body scattering formula
1112: \citep[e.g.,][]{Stewart_Ida00,Ohtsuki02},
1113: \begin{equation}
1114: \tau_{e,{\rm scat}} \simeq
1115: \frac{1}{n_p \pi (GM/(e v_{\rm K})^2)^2 e v_{\rm K} \ln \Lambda},
1116: \end{equation}
1117: where $\ln \Lambda \sim 3$ and
1118: $n_p$ is spatial density of the protoplanet, which is
1119: given by inverse of volume of the planetesimal disk
1120: in the feeding zone,
1121: $1/(2 \pi a_p \times 4\sqrt{3} h a_p)(e v_{\rm K}/\Omega_{\rm K})$.
1122: Then,
1123: \begin{equation}
1124: \tau_{e,{\rm scat}}
1125: \simeq \frac{8\sqrt{3}\pi (e/h)^4}{27\pi} h^{-1} \frac{T_{\rm K}}{2\pi}
1126: \simeq 1 \times 10^{2} (e/h)^4
1127: \left( \frac{M}{M_\oplus}\right)^{-1/3}
1128: \left(\frac{a_p}{5\textrm{AU}} \right)^{3/2} \textrm{yrs.}
1129: \end{equation}
1130: From $\tau_{e,{\rm scat}} = \tau_{\rm damp}$, we obtain
1131: \begin{equation}
1132: (e/h)^2 \simeq 10 \left( \frac{\tau_{\rm damp}}{10^4 \textrm{yrs}} \right)^{1/2}
1133: \left( \frac{M}{M_\oplus}\right)^{1/6}
1134: \left(\frac{a_p}{5\textrm{AU}} \right)^{-3/4} \textrm{.}
1135: \label{eq:equili_e}
1136: \end{equation}
1137:
1138:
1139: \begin{thebibliography}{} %
1140: \bibitem[Adachi et al.(1976)]{AD76} Adachi, I., Hayashi, C., \& Nakazawa, K. 1976, Prog. Theor. Phys., 56, 1756
1141: \bibitem[Fortier et al.(2007)]{Fortier07}
1142: Fortier, A., Benvenuto, O. G. \& Brunini, A. 2007,
1143: A\&A, 473, 311.
1144: \bibitem[Guillot et al.(2006)]{Guillot06}
1145: Guillot, T., Santos, N. C., Pont, F., Iro, N., Melo, C. \& Ribas, I. 2006,
1146: A\&A, 473, L21.
1147: \bibitem[Goldreich \& Tremaine(1982)]{GT82} Goldreich, P., \& Tremaine, S. 1982, Ann. Rev. Astron. Astrophys. 20, 249
1148: \bibitem[Hasegawa \& Nakazawa(1990)]{HN90} Hasegawa, M. \& Nakazawa, K. 1990, Astro. Astrophys, 227, 619
1149: \bibitem[Hayashi et al.(1977)]{HA77} Hayashi, C., Nakazawa, K., \& Adachi, I. 1977, Publ. Astron. Soc. Jpn., 2, 163
1150: \bibitem[Hayashi(1981)]{HA81} Hayashi, C. 1981, Prog. Theor. Phys. Suppl., 70, 35
1151: \bibitem[Ida(1990)]{ID90} Ida, S. 1990, Icarus, 88, 129
1152: \bibitem[Ida \& Lin(2004)]{IL04} Ida, S., \& Lin, D. N. C. 2004, ApJ, 604, 388
1153: \bibitem[Ida \& Lin(2008)]{IL08} Ida, S., \& Lin, D. N. C. 2008, ApJ, 673, 487
1154: \bibitem[Ikoma \& Genda (2006)]{IG06} Ikoma, M., \& Genda, H. 2006, ApJ, 650, 1150
1155: \bibitem[Ikoma et al.(2000)]{IK00} Ikoma, M., Nakazawa, K., \& Emori, H. 2000, ApJ, 537, 1013
1156: \bibitem[Kokubo \& Ida(1998)]{KI98} Kokubo, E., \& Ida, S. 1998, Icarus, 131, 171
1157: \bibitem[Kokubo \& Ida(2002)]{KI02} Kokubo, E., \& Ida, S. 2002, ApJ, 581, 666
1158: \bibitem[Lin \& Papaloizou(1993)]{LP93}
1159: Lin, D. N. C., \& Papaloizou, J. C. B. 1993,
1160: in Protostars and
1161: Planets III, ed. E. H. Levy and J. I. Lunine (Tucson:Univ. of Arizona
1162: Press), 749
1163: \bibitem[Makino(1991)]{MA91} Makino, J. 1991, Publ. Astron. Soc. Jpn., 4, 859
1164: \bibitem[Makino \& Aarseth(1992)]{MA92} Makino, J., \& Aarsetth, S.J. 1992, Publ. Astron. Soc. Jpn., 44, 141
1165: \bibitem[Mizuno(1980)]{Mizuno80}
1166: Mizuno, H. 1980, Prog. Theor. Phys. Suppl., 64, 54
1167: \bibitem[Nakazawa \& Ida(1988)]{NI88} Nakazawa, K., Ida, S. 1988, Prog. Theor. Phys. Suppl., 96, 167
1168: \bibitem[Ohtsuki et al.(2002)]{Ohtsuki02}
1169: Ohtuski, K., Stewart G. R. \& Ida, S. 2002,
1170: Icarus, 155, 436
1171: \bibitem[Podolak(2003)]{PO03} Podolak, M. 2003, Icarus, 165, 428
1172: \bibitem[Pollack et al.(1985)]{PO85} Pollack, J. B., McKay, C. P., \& Christofferson, B. M. 1985, Icarus, 64, 471
1173: \bibitem[Pollack et al.(1996)]{PO96} Pollack, J. B., Hubickyj, O., Bodenheimer, P., Lissauer, J.J., Podolak, M., \& Greenzweig, Y. 1996, Icarus, 124, 62
1174: \bibitem[Saumon \& Guillot(2004)]{SG04} Saumon, D., Guillot, T. 2004, ApJ, 609, 1170
1175: \bibitem[Stevenson(1982)]{Stevenson82}
1176: Stevenson, D. ~J. 1982, P\&SS, 30, 755
1177: \bibitem[Stewart \& Ida(2000)]{Stewart_Ida00}
1178: Stewart, G. R. \& Ida, S. 2000, Icarus, 143, 28
1179: \bibitem[Tanaka \& Ida(1997)]{TI97} Tanaka, H., \& Ida, S. 1997, Icarus, 125, 302
1180: \bibitem[Tanaka \& Ida(1999)]{TI99} Tanaka, H., \& Ida, S. 1999, Icarus, 139, 350
1181: \bibitem[Tanaka \& Ward(2004)]{TW04}
1182: Tanaka, H. \& Ward, W. R., 2004, \apj, 602, 388
1183:
1184: \bibitem[Tanigawa \& Watanabe(2002)]{TW02} Tanigawa, T., \& Watanabe, S. 2002, ApJ, 580, 506
1185: \bibitem[Thommes et al.(2003)]{Thommes03}
1186: Thommes, E. W., Duncan, M. J. \& Levison, H. F. 2003,
1187: Icarus, 161, 431.
1188: \bibitem[Zhou \& Lin(2007)]{ZL07} Zhou, J.-L, Lin, D. N. C. 2007, ApJ, 666, 447
1189: \end{thebibliography}
1190:
1191: \clearpage
1192:
1193: \begin{figure}[btp]
1194: \epsscale{0.6} %
1195: \plotone{f1.eps} %
1196: \caption{Gas accretion timescales for a planet with mass $M$
1197: at $a = 5 \textrm{AU}$.
1198: Solid line is an extrapolation of the model by \citet{IG06}.
1199: Dashed and dotted-dashed lines represent limits by \citet{TW02} and
1200: Bondi accretion.
1201: }
1202: \label{fig:Increase_timescale}
1203: \end{figure}
1204: \clearpage
1205:
1206: \begin{figure}[tbp]
1207: \epsscale{0.6} %filescale
1208: \plotone{f2.eps} %filename
1209: \caption{Evolution of the protoplanet mass
1210: according to the simple power-law gas accretion models
1211: ($\dot{M} \propto M^{p}$).
1212: Initial and final masses are $M_0 = 5.67 M_{\oplus}$ and
1213: $M_{\rm f} = M_J$, where $M_J$ ($= 10^{-3}M_{\odot}$) is a Jupiter mass.
1214: Growth timescale $t_{\rm f} = 10^5$ yr.
1215: After $t > t_{\rm f}$, we set $M = M_{\rm f} = const$.}
1216: \label{fig:Artificial_massup}
1217: \end{figure}
1218: \clearpage
1219:
1220: \begin{figure}[htbp]
1221: \epsscale{.60} %
1222: \plotone{f3.eps}
1223: \caption{Orbital evolution of
1224: a swarm of a planetesimals on the $b$-$(e/h)$ plane.
1225: We adopt $p = 2$, $\tau_{\rm damp} = 10^4$yrs
1226: and $t_{\rm f} = 10^5$ yrs.
1227: The planet is fixed at $e/h = b = 0$ ($a_p = $5 AU).
1228: The horizontal axis $b$ expresses
1229: $(a-a_p)/h$ where $a$ is the semimajor axis of planetesimals.
1230: Solid and dotted lines represent
1231: the boundaries of the feeding zone (i.e. Jacobi energy $E_J = 0$)
1232: and those at $t = 0$, respectively.
1233: The time evolution of the latter is caused by
1234: increase in $h$.
1235: The selected number of planetesimals is 1000 in 3.3 AU $\leq a \leq$ 8.4 AU
1236: at $t=0$.
1237: The numbers of planetesimals are 998($10^3$yr), 992($3\times 10^4$yr), 904($10^5$yr), and 878($3\times 10^5$yr).}
1238: \label{fig:Orbit_snapshot}
1239: \end{figure}
1240: \clearpage
1241:
1242: \begin{figure}[tbp]
1243: \epsscale{0.8} %
1244: \plotone{f4.eps}
1245: \caption{Evolution of the planetesimal accretion rate onto the
1246: growing planet as a function of the protoplanet mass ($M$).
1247: (a) The results in the gas-free case,
1248: the cases of (b) $\tau_{\rm damp} = 10^6$ yrs, (c) $10^5$ yrs
1249: and (d) $10^4$ yrs.
1250: The four lines in the each panel represent the results
1251: with various gas accretion models ($p = 2, 1, 0, -2$).
1252: Initial mass of the protoplanet $M_0$ is set as 5.67$M_\oplus$.
1253: The systems initially consist of 20,000 planetesimals,
1254: so the individual planetesimal masses
1255: correspond to $\simeq 7.7 \times 10^{-4}M_{\oplus}$.}
1256: \label{fig:mlmp7_mass}
1257: \end{figure}
1258: \clearpage
1259:
1260: \begin{figure}[tbp]
1261: \epsscale{0.8} %
1262: \plotone{f5.eps}
1263: \caption{Evolution of the planetesimal accretion rate
1264: as a function of $\xi = v_H / v_{\rm scat}$ for $p = 2,1,0$ and $-2$.
1265: (a) The results in the gas-free case,
1266: the cases of (b) $\tau_{\rm damp} = 10^6$ yrs, (c) $10^5$ yrs
1267: and (d) $10^4$ yrs.
1268: The fitting formula, eq.~(\ref{eq:fitting_line1}),
1269: is expressed by thick solid lines in the plots.}
1270: \label{fig:mlmp7_xsi}
1271: \end{figure}
1272: \clearpage
1273:
1274: \begin{figure}[tbp]
1275: \epsscale{1.00} %
1276: \plottwo{f6a.eps}{f6b.eps}
1277: \caption{Evolution of the scaled planetesimal accretion rate
1278: as a function of $\eta = v_H / v_{\rm damp}$ for $p = 2,0$ and $-2$
1279: in the case of $\tau_{\rm damp} = 10^4$ yrs (left panel) and for
1280: $\tau_{\rm damp} = 10^6,10^5$ and $10^4$ yrs in the case of $p = -2$ (right panel).
1281: The fitting formula, eq.~(\ref{eq:fitting_line2}),
1282: is expressed by thick solid lines.}
1283: \label{fig:mlmp11_eta}
1284: \end{figure}
1285: \clearpage
1286:
1287: %comparison with numerical and semi-analytical result
1288: \begin{figure}[tbp]
1289: \epsscale{1.00} %
1290: \plottwo{f7a.eps}{f7b.eps}
1291: \caption{Comparison between the numerical simulations and the semi-analytical results.
1292: The left and right panels show the results
1293: for gas accretion with $p = 2$ and $p = 0$, respectively.
1294: The thin and thick curves represent
1295: the numerical and semi-analytical results.}
1296: \label{fig:Arti_Compare_Amass}
1297: \end{figure}
1298: \clearpage
1299: %insert picture
1300:
1301: %accretion mass with realistic mass-up model
1302: \begin{figure}[tbp]
1303: \epsscale{1.0} %
1304: \plottwo{f8a.eps}{f8b.eps} %
1305: \caption{The evolution of cumulative mass of accreted planetesimals
1306: as a function of $M$ in the case with $M_0 = 10M_\oplus$.
1307: The left and right panels show the results
1308: for $a = 5.2$ AU and $a = 9.55$ AU, which correspond to
1309: Jupiter and Saturn.
1310: Here, $R=2R_1$ and $f_d=2$ are assumed, where $R_1$ is
1311: the physical radius for mass $M$ and $\rho = 1$gcm$^{-3}$.
1312: For other $R$ and $f_d$, the accreted mass is
1313: multiplied by $(R/2R_1)^{1/2} (f_d/2)$.
1314: }
1315: \label{fig:Amass_DAMP}
1316: \end{figure}
1317: \clearpage
1318: %insert picture
1319:
1320: %Phase2 region
1321: \begin{figure}[tbp]
1322: \epsscale{0.6}
1323: \plotone{f9.eps}
1324: \caption{The parameter range in which
1325: phase 2 can occur, which is expressed by the shaded regions.
1326: The region above the solid line represents $\eta > 1$.
1327: Phase 2 can occur in the regions above the dashed line for $\eta < 1$ and
1328: in the regions below the dot-dashed line for $\eta > 1$.
1329: }
1330: \label{fig:Phase2}
1331: \end{figure}
1332: \clearpage
1333: %insert picture
1334: \end{document}
1335: