0807.1761/ms.tex
1: \documentclass[12pt,preprint]{aastex}
2: 
3: \shorttitle{planet formation in turbulent disks} 
4: 
5: \shortauthors{Ida et al.}
6: 
7: \begin{document}
8: 
9: \title{Accretion and destruction of planetesimals in turbulent disks}
10: 
11: \author{Shigeru Ida}
12: \affil{Tokyo Institute of Technology,
13: Ookayama, Meguro-ku, Tokyo 152-8551, Japan}
14: \email{ida@geo.titech.ac.jp}
15: 
16: \and 
17: 
18: \author{Tristan Guillot, Alessandro Morbidelli}
19: \affil{Observatoire de la C\^{o}te d'Azur,
20: CNRS UMR 6202, BP 4229, 06304 Nice Cedex 4, France}
21: \email{guillot@oca.eu, morby@oca.eu}
22: 
23: 
24: \begin{abstract}
25: We study the conditions for collisions between planetesimals to be
26: accretional or disruptive in turbulent disks, through analytical
27: arguments based on fluid dynamical simulations and orbital
28: integrations.  In turbulent disks, the velocity dispersion of
29: planetesimals is pumped up by random gravitational perturbations from
30: density fluctuations of the disk gas.  When the velocity dispersion is
31: larger than the planetesimals' surface escape velocity, collisions
32: between planetesimals do not result in accretion, and may even lead to
33: their destruction. In disks with a surface density equal to that of
34: the ``minimum mass solar nebula'' and with nominal MRI turbulence, we
35: find that accretion proceeds only for planetesimals with sizes above
36: $\sim 300$\,km at 1AU and $\sim 1000$\,km at 5AU.  We find that
37: accretion is facilitated in disks with smaller masses. However, at
38: 5AU and for nominal turbulence strength, 
39: km-sized planetesimals are in a highly erosive regime
40: even for a disk mass as small as a 
41: fraction of the mass of Jupiter. The existence of
42: giant planets implies that either turbulence
43: was weaker than calculated by standard MRI models 
44: or some mechanism was capable of producing Ceres-mass planetesimals 
45: in very short timescales. 
46: In any case, our results show that in the presence of
47: turbulence planetesimal accretion is most difficult in massive
48: disks and at large orbital distances.
49: \end{abstract}
50: 
51: \keywords{solar system: formation --- planets and satellites: formation 
52:     --- accretion, accretion disks --- turbulence}
53: 
54: \section{Introduction}
55: 
56: It is often considered that the evolution of protoplanetary disks and
57: the consequent accretion of gas by the central protostar are driven by
58: turbulent viscosity due to a Magneto-Rotational-Instability (MRI)
59: \citep[e.g.,][]{bal91}.  \citet{lau04} and \citet{nel04} carried out
60: fluid dynamical simulations of MRI and found that the random torques
61: due to the turbulent density fluctuations give rise to a random walk
62: in semimajor axes of planetesimals.  \citet{ric03} pointed out through
63: model calculations that the random walk expands the effective feeding
64: zone of protoplanets, and may lead to rapid formation of large cores
65: for gas giants.  Through a Fokker-Planck treatment, \citet{johnson06}
66: also pointed out the importance of the random walk in planet accretion.  
67: Adopting the semi-analytical formula for the random torque 
68: derived by \citet{lau04}, \citet{ogi07} performed
69: N-body simulations for the late stages of terrestrial planet
70: accretion with a disk significantly depleted in gas, starting from
71: Mars-mass protoplanets.
72: They found that the MRI
73: turbulence indeed helps to reduce the number of
74: accreted terrestrial planets which is otherwise too large compared to
75: our Solar System.
76: 
77: However, \citet{nel05} found through direct integrations of
78: the orbits of protoplanets in a MRI turbulent disk that orbital
79: eccentricities are also excited.
80: \citet{britsch08} also found a similar feature 
81: in a self-gravitating disk.
82: While the random walk itself is favorable to
83: the growth of protoplanets by avoiding isolation, the
84: excitation of their eccentricities, which had been neglected in
85: \citet{ric03} and \citet{johnson06}, 
86: is a threat for planetesimal accretion processes
87: because of increased collision velocity. 
88: Unfortunately, \citet{nel05}'s orbital integrations were 
89: limited to 100--150 Keplerian times
90: and neglected collision processes, so that it is not possible to
91: conclude from that work whether planetesimals should grow or be eroded
92: in the presence of turbulence. 
93: 
94: In the present article, we explore by which paths planetesimals may
95: have grown to planet-sized bodies in turbulent disks. 
96: Because the level of density fluctuations due to the MRI turbulence is
97: not well determined, we choose to study the qualitative effects
98: of the turbulence on the accretion of planetesimals and 
99: their dependence on the key parameters of the problem, in particular
100: the progressive removal of the gas disk. 
101: In \S 2, we summarize the conditions for the accretion and destruction
102: of planetesimals in terms of their orbital eccentricities. 
103: In \S 3, we analytically derive the equilibrium eccentricity
104: for which the excitation due to turbulence is balanced
105: by damping due to tidal interactions with the disk gas,
106: aerodynamic gas drag, and collisions.
107: Comparing the equilibrium eccentricities with critical 
108: eccentricities for accretion and destruction,
109: we derive critical physical radii and masses of
110: planetesimals for accretion or destruction.
111: The results are applied to viscously evolving disks (\S 4).
112: We then discuss possible solutions for the problem of the formation of
113: planetesimals and planets (\S 5).
114: 
115: 
116: \section{Accretion and destruction conditions}
117: 
118: We summarize the accretion and destruction conditions below.  From
119: energy conservation, the collision velocity ($v_{\rm coll}$) between
120: two planetesimals (labeled 1 and 2) satisfies
121: \begin{equation}
122: E = \frac{1}{2}v_{\rm coll}^2 - \frac{G(M_1 + M_2)}{R_1 + R_2}
123: = \frac{1}{2}v_{\rm rel}^2,
124: \end{equation}
125: where $M_j$ and $R_j$ are the mass and physical radius of a
126: planetesimal $j$ ($j=1,2$), and $v_{\rm rel}$ is their relative
127: velocity when they are apart from each other.  When the velocity
128: dispersion of planetesimals $v_{\rm disp}$ is larger than the Hill
129: velocity that is given by $(M/3M_{\ast})^{1/3} v_{\rm K}$, where
130: $M_{\ast}$ is the mass of the host star and $v_{\rm K}$ is
131: the Keplerian velocity, $v_{\rm rel}$ is
132: approximated by $v_{\rm disp}$ \citep[e.g.,][]{IN89,I90}.  The total
133: energy then becomes:
134: \begin{equation}
135: 2E = v_{\rm coll}^2 - v_{\rm esc}^2 \simeq v_{\rm disp}^2,
136: \end{equation}
137: where $v_{\rm esc}$ is the (two-body) surface escape velocity
138: defined by
139: \begin{equation}
140: v_{\rm esc} = \sqrt{ \frac{2G(M_1 + M_2)}{R_1 + R_2} }.
141: \end{equation}
142: 
143: Collisional dissipation decreases the energy by some fraction of
144: $v_{\rm coll}^2/2$.
145: If $v_{\rm disp} \ll v_{\rm esc}$, the 
146: collisional dissipation results in $E < 0$ after collision.
147: On the other hand, $E$ is likely to be still positive 
148: after a collision with $v_{\rm disp} \gg v_{\rm esc}$.
149: Thus, for moderate dissipation,
150: the condition for an accretional collision is
151: $v_{\rm disp} < v_{\rm esc}$ \citep[e.g.,][]{ohtsuki93}.
152: Since the orbital eccentricity $e \simeq v_{\rm disp}/v_{\rm K}$, 
153: a collision should result in accretion for
154: $e < e_{\rm acc}$, where
155: \begin{equation}
156: \begin{array}{lll}
157: e_{\rm acc} & \simeq \frac{v_{\rm esc}}{v_{\rm K}}
158:  & \simeq 0.28 \left(\frac{M}{M_{\oplus}}\right)^{1/3}
159:   \left(\frac{\rho_{\rm p}}{3{\rm gcm}^{-3}}\right)^{1/6}
160:   \left(\frac{r}{1{\rm AU}}\right)^{1/2} \\
161:  & & 
162: \simeq 0.036 
163:   \left(\frac{R}{10^3{\rm km}}\right)
164:   \left(\frac{\rho_{\rm p}}{3{\rm gcm}^{-3}}\right)^{1/2}
165:   \left(\frac{r}{1{\rm AU}}\right)^{1/2}.
166: \end{array}
167:   \label{eq:e_acc}
168: \end{equation}
169: In the above relation, $\rho_{\rm p}$ is the bulk density of
170: the planetesimals and
171: $M \sim M_j$ (for simplicity, $M_1 \sim M_2$ is assumed).
172: The physical radius $R$ is given by
173: \begin{equation}
174: R=7.8 \times 10^8 (M/M_{\oplus})^{1/3} (\rho_{\rm p}/3{\rm gcm}^{-3})^{-1/3}
175: {\rm cm}.
176: \end{equation} 
177: 
178: A collision results in destruction if the collision velocity
179: is such that the
180: specific kinetic energy of a collision ($v_{\rm col}^2/2$) exceeds
181: \begin{equation}
182: Q_{\rm D} \simeq 
183:   \left[ Q_0 
184:   \left(\frac{R}{1{\rm cm}}\right)^{a} + 
185:   3 B \left(\frac{\rho_{\rm p}}{3{\rm gcm}^{-3}}\right) 
186:   \left(\frac{R}{1{\rm cm}}\right)^{b} \right] {\rm erg/g},
187: \label{eq:Q_D}
188: \end{equation}
189: where $Q_0$ is the material strength, 
190: $B \simeq 0.3$--2.1, $a \simeq -0.4$, and $b \simeq 1.3$ 
191: \citep{Benz_Asphaug99}.
192: For basalt rocks or water ice, 
193: $Q_0 \simeq 10^7$--$10^8$ \citep{Benz_Asphaug99}, but it can take
194: a significantly smaller value for loose aggregates.
195: \footnote{For porous materials, $Q_0$ is rather higher (W. Benz,
196: private communication).}
197: 
198: We adopt $Q_0 = 3 \times 10^7$ as a nominal value.
199: Self-gravity (the second term in the r.h.s.) dominates the material
200: strength when $R \ga 100$ m. 
201: In this regime, adopting $B \sim 1$, 
202: a collision results in destruction for
203: $e > e_{\rm dis}$, where 
204: \begin{equation}
205: \begin{array}{ll}
206: e_{\rm dis} \simeq \frac{\sqrt{2 Q_{\rm D}}}{v_{\rm K}}
207:  & \simeq 0.50 \left(\frac{M}{M_{\oplus}}\right)^{0.22}
208:   \left(\frac{\rho_{\rm p}}{3{\rm gcm}^{-3}}\right)^{0.28}
209:   \left(\frac{r}{1{\rm AU}}\right)^{1/2} \\
210:  & \simeq 0.13 \left(\frac{R}{10^3{\rm km}}\right)^{0.65}
211:   \left(\frac{\rho_{\rm p}}{3{\rm gcm}^{-3}}\right)^{0.5}
212:   \left(\frac{r}{1{\rm AU}}\right)^{1/2}. 
213: \end{array}
214: \label{eq:e_D}
215: \end{equation}
216: 
217: \section{Equilibrium eccentricities}
218: 
219: We first derive the equilibrium eccentricities of
220: planetesimals at which the excitation by the MRI turbulence is
221: balanced by damping due to drag and/or collisions.  
222: Comparing the
223: estimated eccentricities with $e_{\rm acc}$ and $e_{\rm dis}$, 
224: we then evaluate the outcome of
225: collisions between planetesimals as a function of planetesimal size,
226: turbulent strength, and surface density of disk gas.
227: 
228: For an easy interpretation, we provide in this section analytical
229: relations based on a model in which the gas and
230: solid components of disk surface density are scaled with the
231: multiplicative factors $f_g$ and $f_d$:
232: \begin{equation}
233: \Sigma_{g} = 2400 f_g \left(\frac{r}{1\mathrm{AU}}\right)^{-3/2} \,\mathrm{g\, cm}^{-2},
234: \label{eq:Sigma_g}
235: \end{equation}
236: and 
237: \begin{equation}
238: \Sigma_{d} = 10 f_d \eta_{\rm ice} \left(\frac{r}{1\mathrm{AU}}\right)^{-3/2} \,\mathrm{g\, cm}^{-2},
239: \label{eq:Sigma_d}
240: \end{equation}
241: where $\eta_{\rm ice} \simeq 3$--4
242: is an enhancement factor of $\Sigma_d$ due to ice condensation.
243: If $f_g=f_d=1$, $\Sigma_{g}$ and $\Sigma_d$ are 
244: 1.4 times those of the minimum mass solar nebula model \citep{hayashi81}.
245: 
246: In this section, we also use the disk temperature distribution 
247: obtained in the optically thin limit \citep{hayashi81},
248: \begin{equation}
249: T \simeq 280
250: \left(\frac{r}{1\, {\rm AU}}\right)^{-1/2}
251: \left(\frac{L_*}{L_{\odot}} \right)^{1/4}
252: \; {\rm K},
253: \label{eq:T_opt_thin_disk}
254: \end{equation}
255: where $L_*$ and $L_{\odot}$ are the stellar and solar luminosities,
256: respectively. The corresponding sound velocity is
257: \begin{equation}
258: c_s = 1.1 \times 10^5 \left( \frac{r}{1\,{\rm AU}} \right)^{-1/4}
259: \left(\frac{L_*}{L_{\odot}} \right)^{1/8}
260: \; {\rm cm/s}.
261: \label{eq:sound_velocity_hayashi}
262: \end{equation}
263: Since the disk scale height is given by $h = \sqrt{2}c_s/\Omega_{\rm K}$
264: (assuming that $T$ is vertically uniform in the disk),
265: eqs.~(\ref{eq:Sigma_g}) and (\ref{eq:sound_velocity_hayashi})
266: yield the disk gas density at the midplane as
267: \begin{equation}
268: \rho_g = \frac{\Sigma_g}{\sqrt{\pi}h}
269:        = 2 \times 10^{-9} f_g
270:          \left( \frac{r}{1\, {\rm AU}} \right)^{-11/4}
271:          \; {\rm gcm}^{-3}.
272: \label{eq:rho_midplane}
273: \end{equation}
274: 
275: 
276: \subsection{Excitation}
277: 
278: The orbital eccentricities of planetesimals are pumped up both by the
279: random gravitational perturbations from density fluctuations of disk
280: gas, as well as by mutual gravitational scattering among
281: planetesimals.  Assuming planetesimals have equal masses, their
282: orbital eccentricities should be excited to at most $\sim e_{\rm acc}$
283: by the mutual scattering \citep[e.g.,][]{Safronov69}.  
284: As will be shown below, the value of
285: this excentricity is smaller than that due to the turbulent
286: excitation, except for very large planetesimals ($\sim 10^3$\,km or
287: larger), and/or in the case of significantly depleted gas disks.
288: For simplicity, in this work we choose to neglect the possibility that mutual
289: scattering dominates over turbulent excitation. It is therefore important to
290: note that our results may be slightly optimistic when concerning the
291: possibility of accretion of massive planetesimals. 
292: 
293: The orbital eccentricities that result from the turbulent density
294: fluctuations in the disk are provided by \citet{ogi07} on the basis of
295: orbital integrations with empirical formula by \citet{lau04}, as
296: \begin{equation}
297: e \sim 0.1 \gamma 
298:   \left(\frac{\Sigma_g}{\Sigma_{g,1}}\right)
299:   \left(\frac{r}{1{\rm AU}}\right)^{2}
300:   \left(\frac{t}{T_{\rm K}}\right)^{1/2}
301:   =
302:   0.1 f_g \gamma 
303:   \left(\frac{r}{1{\rm AU}}\right)^{-1/4}
304:   \left(\frac{t}{1{\rm year}}\right)^{1/2},
305: \label{eq:e_random}
306: \end{equation}
307: where $\Sigma_{g,1}$ is $\Sigma_{g}$ at 1AU with $f_g = 1$
308: (eq.~[\ref{eq:Sigma_g}]) and
309: $\gamma$ is a non-dimensional parameter to express
310: the disk turbulence.
311: \footnote{Although \citet{ogi07} suggested that eq.~(\ref{eq:e_random})
312:   may be enhanced by a factor 10 by the inclusion of $m=1$ modes, the $m=1$
313:   modes actually enhance only the amplitude of random walk in
314:   semimajor axis ($\Delta a$) but not the eccentricity.  
315:   Since higher $m$ modes fluctuate over shorter timescales, 
316:   they tend to cancel out on the orbital period of a planetesimal.
317:   For these modes, $\Delta a/a$, which is due to time variation 
318:   of the potential, is much smaller than $\Delta e$, because the 
319:   latter is also excited by the non-axisymmetric structure.
320:   The inclusion of slowly varying $m=1$ modes enhances 
321:   $\Delta a/a$ up to the order of $\sim e$.
322:   On the other hand, the definition of $\Gamma$ in eqs.~(5) and
323:   (34) of \citet{ogi07} should be multiplied by $\pi$.  We use
324:   eq.~(\ref{eq:e_random}) for the eccentricity excitation, which is
325:   consistent with an orbital calculation including $m=1$ modes
326:   (figure~\ref{fig:e_evol}).}
327: Although \citet{ogi07} showed the results only at $\sim 1$AU,
328: we here added a dependence on $r$ using scaling arguments
329: (see the Appendix).
330: Orbital integration for other $r$ show a consistent dependence.
331: Note that $\Delta a/a \sim e$, where
332: $\Delta a$ is the amplitude of random walk in semimajor axis.  Since
333: $e \ll 1$, the radial distance $r$ and semimajor axis $a$ are
334: identified here.
335: 
336: From the simulation results by \citet{lau04}, the value of $\gamma$ may
337: be $\sim (1/3) (\delta \rho/\rho) \sim 10^{-3}$--$10^{-2}$ for MRI
338: turbulence.  
339: In this paper, we use $\gamma = 10^{-3}$ as a fiducial value.
340: Interestingly, with a quite
341: different approach, \citet{johnson06} derived a similar formula
342: for $\Delta a/a$ with the same dependences on $r$, $\Sigma_g$ and $t$.
343: If $e \simeq \Delta a/a$, their formula is consistent with ours.
344: They suggested that $\gamma \sim \alpha$ 
345: or $\alpha^{1/2} h/a$ where $h$ is disk scale height
346: and $\alpha$ is the parameter for the alpha prescription
347: for turbulent viscosity \citep{alpha}.
348: For $\alpha = 10^{-3}$--$10^{-2}$, 
349: their estimate is also similar to our fiducial value.
350: 
351: The top panels in fig.~\ref{fig:e_evol}
352: show the results of an orbital integration 
353: with 4-th order Hermite scheme for
354: the evolution of $e$ and $\Delta a/a$ 
355: with turbulent perturbations but without any damping.
356: Five independent runs with different random number seeds 
357: for the generation of turbulent density fluctuations 
358: \citep{ogi07} are plotted in each panel. 
359: The initial $e$ and $i$ are $10^{-6}$.
360: For $f_g = 1$, $\gamma = 0.01$, and $r=1$AU, as used 
361: in fig.~\ref{fig:e_evol}, 
362: eq.~(\ref{eq:e_random}) is reduced to
363: $e \sim 10^{-3}(t/1{\rm year})^{1/2}$.
364: To highlight the effect of turbulence, we used
365: a larger value of $\gamma$ than the fiducial value.
366: The evolution of the root mean squares of the five runs 
367: in fig.~\ref{fig:e_evol} agrees with eq.~(\ref{eq:e_random})
368: within a factor of $\sim 2$.
369: From eq.~(\ref{eq:e_random}), the excitation timescale is
370: \begin{equation}
371: \tau_{\rm exc} = \frac{e}{de/dt} 
372:   \simeq 2 \times 10^{2} 
373:   \gamma^{-2} e^2
374:   \left(\frac{\Sigma_g}{\Sigma_{g,1}}\right)^{-2}
375:   \left(\frac{r}{1{\rm AU}}\right)^{-4} T_{\rm K} 
376:   = 2 \times 10^{2} f_g^{-2} \gamma^{-2} e^2
377:   \left(\frac{r}{1{\rm AU}}\right)^{1/2} \; {\rm years}.
378: \label{eq:t_exc}
379: \end{equation}
380: 
381: 
382: 
383: \subsection{Damping}
384: 
385: The eccentricity damping processes are
386: i) tidal interaction with disk gas, ii) aerodynamical gas drag,
387: and iii) inelastic collisions.
388: The tidal damping timescale (i) is derived by \citet{tan04} as
389: \begin{equation}
390: \tau_{\rm tidal} \simeq
391:             1.3
392:             \left(\frac{M}{M_{\odot}}\right)^{-1}
393:             \left(\frac{\Sigma_{g} r^2}{M_{\odot}}\right)^{-1}
394:             \left(\frac{c_s}{v_{\rm K}}\right)^{4}
395:             \Omega_{\rm K}^{-1}
396:             \simeq 
397:  3 \times 10^{2} f_g^{-1} 
398:   \left(\frac{M}{M_{\oplus}}\right)^{-1}
399:   \left(\frac{r}{1{\rm AU}}\right)^2 \; {\rm years}.
400: \label{eq:t_tidal}
401: \end{equation}
402: The gas drag damping timescale (ii) is derived by \citet{Adachi76} as
403: \begin{equation}
404: \tau_{\rm drag} \simeq
405:  \frac{M v_{\rm disp}}{\pi R^2 \rho_g v_{\rm disp}^2}
406: \simeq
407:  2 \times 10^{4} f_g^{-1} e^{-1} 
408:   \left(\frac{M}{M_{\oplus}}\right)^{1/3}
409:   \left(\frac{\rho_{\rm p}}{3{\rm gcm}^{-3}}\right)^{2/3}
410:   \left(\frac{r}{1{\rm AU}}\right)^{13/4} \; {\rm years}.
411: \label{eq:t_drag}
412: \end{equation}
413: 
414: For simplicity, we evaluate the damping timescale due to inelastic
415: collision as the mean collision time of planetesimals, assuming that
416: all the planetesimals have the same mass $M$.  Since in the size
417: distribution caused by collision cascade, collisions with
418: comparable-sized bodies and those with smaller ones 
419: contribute similarly, the neglection of the size
420: distribution may not be too problematic.
421: Since we look for the conditions in
422: which collisions are non-accretional, we consider the case with
423: $v_{\rm disp} > v_{\rm esc}$.  Assuming that the gravitational
424: focusing factor $[1 + (v_{\rm esc}/v_{\rm disp})^2] \sim 1$, the
425: collision damping timescale is
426: \begin{equation}
427: \tau_{\rm coll} \simeq
428:  \frac{1}{n \pi R^2 v_{\rm disp}}
429: \simeq
430:  2 \times 10^{7} f_d^{-1} \eta_{\rm ice}^{-1} 
431:   \left(\frac{M}{M_{\oplus}}\right)^{1/3}
432:   \left(\frac{\rho_{\rm p}}{3{\rm gcm}^{-3}}\right)^{2/3}
433:   \left(\frac{r}{1{\rm AU}}\right)^{3} {\rm years},
434: \label{eq:t_coll}
435: \end{equation}
436: where $n$ is the spatial number density of planetesimals.
437: Note that $n \sim (\Sigma_d/M)/(v_{\rm disp}/ \Omega_{\rm K})$.
438: 
439: 
440: \subsection{Equilibrium eccentricity}
441: 
442: We now equate eq.~(\ref{eq:t_exc}) with 
443: eqs.~(\ref{eq:t_tidal}), (\ref{eq:t_drag}), and (\ref{eq:t_coll}),
444: respectively,
445: to obtain an equilibrium eccentricity for each damping process.
446: For simplicity and to a good approximation, the actual equilibrium
447: eccentricity can be approximated as the minimum of the three
448: equilibrium eccentricities. 
449: From eqs.~(\ref{eq:t_exc}) and (\ref{eq:t_tidal}), 
450: \begin{equation}
451: \begin{array}{ll}
452: e_{\rm tidal} & \simeq
453: 1.2 f_g^{1/2} \gamma  
454:   \left(\frac{M}{M_{\oplus}}\right)^{-1/2}
455:   \left(\frac{r}{1{\rm AU}}\right)^{3/4} \\
456:  & \simeq
457: 24 f_g^{1/2} \gamma  
458:   \left(\frac{R}{10^3{\rm km}}\right)^{-3/2}
459:   \left(\frac{\rho_{\rm p}}{3{\rm gcm}^{-3}}\right)^{-1/2}
460:   \left(\frac{r}{1{\rm AU}}\right)^{3/4}.
461: \end{array}
462: \label{eq:e_tidal}
463: \end{equation}
464: With eq.~(\ref{eq:t_drag}), 
465: \begin{equation}
466: \begin{array}{ll}
467: e_{\rm drag} & \simeq
468: 4.6 f_g^{1/3} \gamma^{2/3}  
469:   \left(\frac{M}{M_{\oplus}}\right)^{1/9}
470:   \left(\frac{\rho_{\rm p}}{3{\rm gcm}^{-3}}\right)^{2/9}
471:   \left(\frac{r}{1{\rm AU}}\right)^{11/12} \\
472:  & \simeq
473: 0.23 f_g^{1/3} \gamma^{2/3}  
474:   \left(\frac{R}{1{\rm km}}\right)^{1/3}
475:   \left(\frac{\rho_{\rm p}}{3{\rm gcm}^{-3}}\right)^{1/3}
476:   \left(\frac{r}{1{\rm AU}}\right)^{11/12}.
477: \end{array}
478: \label{eq:e_drag}
479: \end{equation}
480: For $f_g = 1$, $\gamma = 0.01$, and $r=1$AU,
481: eq.~(\ref{eq:e_drag}) predicts that
482: $e_{\rm drag} \simeq 0.045$ for $M/M_{\oplus}=10^{-6}$
483: and $e_{\rm drag} \simeq 0.01$ for $M/M_{\oplus}=10^{-12}$.
484: An orbital integration in the middle and bottom panels in 
485: fig.~\ref{fig:e_evol} shows that the results agree with the 
486: analytical estimate within a factor $\sim 1.5$.  
487: With eq.~(\ref{eq:t_coll}), 
488: \begin{equation}
489: \begin{array}{ll}
490: e_{\rm coll} & \simeq
491: 3.2 \times 10^2 f_g (f_d \eta_{\rm ice})^{-1/2} \gamma
492:   \left(\frac{M}{M_{\oplus}}\right)^{1/6}
493:   \left(\frac{\rho_{\rm p}}{3{\rm gcm}^{-3}}\right)^{1/3}
494:   \left(\frac{r}{1{\rm AU}}\right)^{5/4} \\
495:  & \simeq
496: 3.6 f_g (f_d \eta_{\rm ice})^{-1/2} \gamma
497:   \left(\frac{R}{1{\rm km}}\right)^{1/2}
498:   \left(\frac{\rho_{\rm p}}{3{\rm gcm}^{-3}}\right)^{5/6}
499:   \left(\frac{r}{1{\rm AU}}\right)^{5/4}.
500: \end{array}
501: \label{eq:e_coll}
502: \end{equation}
503: 
504: In figure \ref{fig:e_eq}, the equilibrium eccentricity, $e_{\rm eq} =
505: {\rm min}(e_{\rm tidal}, e_{\rm drag}, e_{\rm coll})$, is plotted 
506: with solid lines as a function of the planetesimal radius $R$,
507: the corresponding planetesimal mass being $M = 2.1 \times 10^{-3}
508: (R/10^3{\rm km})^3 (\rho_{\rm p}/3 {\rm gcm}^{-3}) M_{\oplus}$,
509: Note again that the effect of mutual planetesimal
510: scattering is neglected.  For bodies with more than Lunar to Mars
511: masses, tidal damping is dominant.  This yields a
512: decrease in the equilibrium eccentricity with
513: increasing planetesimal radius for $R \ga 100$km.
514: For smaller mass bodies, gas drag damping dominates tidal damping
515: and the equilibrium eccentricity increases with increasing $R$.
516: For the smallest planetesimal sizes (the regions with the slightly steeper
517: positive gradient), collision damping is dominant, but with a
518: significant contribution of gas drag damping.
519: 
520: The limiting mass and radius at which $e_{\rm drag}$
521: (eq.~[\ref{eq:e_drag}]) and $e_{\rm acc}$ (eq.~[\ref{eq:e_acc}]) cross
522: are
523: \begin{equation}
524: \begin{array}{l}
525: M_{\rm acc} \simeq 
526: 3.0 \times 10^{-4} f_g^{3/2} 
527:   \left(\frac{\gamma}{10^{-3}}\right)^{3}
528:   \left(\frac{\rho_{\rm p}}{3{\rm gcm}^{-3}}\right)^{-1/4}
529:   \left(\frac{r}{1{\rm AU}}\right)^{9/8} M_{\oplus}, 
530: \label{eq:m_acc}
531: \\
532: R_{\rm acc} \simeq 
533: 5.2 \times 10^2 f_g^{1/2} 
534:   \left(\frac{\gamma}{10^{-3}}\right)
535:   \left(\frac{\rho_{\rm p}}{3{\rm gcm}^{-3}}\right)^{-5/12}
536:   \left(\frac{r}{1{\rm AU}}\right)^{3/8} {\rm km}.
537: \label{eq:r_acc}
538: \end{array}
539: \end{equation}
540: The accretion of planetesimals is possible for 
541: $M > M_{\rm acc}$ ($R > R_{\rm acc}$).
542: In the top panel of fig.~\ref{fig:e_eq}
543: ($\gamma = 10^{-3}$ and $f_g = 1$),
544: planetesimal accretion proceeds in a range of
545: $R$'s in which the solid line ($e_{\rm eq}$) is 
546: located below the dashed line ($e_{\rm acc}$), that is,
547: only if a body is larger than Ceres.
548: Such large planetesimals can be formed 
549: by a different mechanism than pairwise accretion
550: such as self-gravitational instability 
551: in turbulent eddies \citep[e.g.,][]{Johansen07}.
552: When the disk gas is removed, accretion becomes
553: possible for smaller planetesimals 
554: (the 2nd panel of fig.~\ref{fig:e_eq}).
555: On the other hand, if turbulence is stronger
556: ($\gamma \sim 10^{-2}$), planetesimal accretion requires
557: more than 1000 km-sized bodies.
558: This appears to be an insurmountable
559: barrier to accretion, even for depleted gaseous disks ($f_g \sim
560: 0.1$), as shown in the the 3rd panel of fig.~\ref{fig:e_eq}.
561: Finally, at large orbital radii, planetesimal accretion is 
562: even more difficult (the bottom panel).
563: 
564: Another critical mass (radius) is the point
565: at which $e_{\rm drag}$ (eq.~[\ref{eq:e_drag}]) and 
566: $e_{\rm dis}$ in the gravity regime (eq.~[\ref{eq:e_D}]) cross,
567: \begin{equation}
568: \begin{array}{l}
569: M_{\rm dis} \simeq 
570:   6 \times 10^{-11} f_g^{3.3} 
571:   \left(\frac{\gamma}{10^{-3}}\right)^{6.7}
572:   \left(\frac{\rho_{\rm p}}{3{\rm gcm}^{-3}}\right)^{-0.6}
573:   \left(\frac{r}{1{\rm AU}}\right)^{4.2} M_{\oplus},
574: \label{eq:m_D}
575: \\
576: R_{\rm dis} \simeq 
577: 3 f_g^{1.1} 
578:   \left(\frac{\gamma}{10^{-3}}\right)^{2.2}
579:   \left(\frac{\rho_{\rm p}}{3{\rm gcm}^{-3}}\right)^{-0.53}
580:   \left(\frac{r}{1{\rm AU}}\right)^{1.4} {\rm km}.
581: \label{eq:r_D}
582: \end{array}
583: \end{equation}
584: Planetesimals with $M < M_{\rm dis}$ ($R < R_{\rm dis}$) 
585: are disrupted by collisions
586: down to the sizes for which material strength is dominant
587: (see below).
588: For $\gamma \sim 10^{-3}$ and $f_g \sim 1$
589: (the top panel of fig.~\ref{fig:e_eq}),
590: planetesimals with sizes larger than several km radius survive
591: but without growing, while smaller planetesimals 
592: are disrupted.
593: 
594: When a planetesimal is smaller than $\sim 100$ m in size,
595: it is bounded by material strength rather than 
596: self-gravity.  
597: In the regime of material strength, 
598: $Q_{\rm D} \sim Q_0 (R/1{\rm cm})^{-0.4}$.
599: The body is not disrupted if
600: $\sqrt{2Q_{\rm D}}/v_{\rm K} > e_{\rm drag}$, which
601: is equivalent to
602: \begin{equation}
603: \begin{array}{l}
604: M \la M_{\rm mat} \simeq 0.8 \times 10^{-17} f_g^{-1.9} 
605:   \left(\frac{\gamma}{10^{-3}}\right)^{-3.7}
606:   \left(\frac{\rho_{\rm p}}{3{\rm gcm}^{-3}}\right)^{-0.9}
607:   \left(\frac{r}{1{\rm AU}}\right)^{2.8} M_{\oplus},
608: \label{eq:m_D2} \\
609: R \la R_{\rm mat} \simeq 16 f_g^{-0.62} 
610:   \left(\frac{\gamma}{10^{-3}}\right)^{-1.25}
611:   \left(\frac{\rho_{\rm p}}{3{\rm gcm}^{-3}}\right)^{-0.62}
612:   \left(\frac{r}{1{\rm AU}}\right)^{2.8} {\rm m}.
613: \label{eq:r_D2}
614: \end{array}
615: \end{equation}
616: Since in this regime, collision damping is slightly stronger than
617: gas drag, actual values of $M_{\rm mat}$ and $R_{\rm mat}$ are determined by
618: $\sqrt{2Q_{\rm D}}/v_{\rm K} > e_{\rm coll}$, so
619: they are slightly larger than the above estimate (see fig.~\ref{fig:e_eq}).
620: When $\tau_{\rm drag} \Omega_{\rm K} \la 1$, 
621: the planetesimals' motions are coupled
622: to that of the gas.  The collision velocity then cannot be expressed
623: in terms of orbital eccentricity. This limiting size is however
624: much smaller than $R_{\rm mat}$.
625: The collision cascade would hence stop at 
626: $M \sim M_{\rm mat}$ ($R \sim R_{\rm mat})$.
627: Regions for which the dotted lines ($e_{\rm dis}$) in
628: fig.~\ref{fig:e_eq} have negative gradients correspond to the material strength regime.
629: In the depleted disk case, $e_{\rm dis}$ is always larger than
630: $e_{\rm eq}$, so that the disruptive regions do not exist
631: (see the 2nd panel of fig.~\ref{fig:e_eq}).
632: Note that $R_{\rm mat} \propto Q_0^{0.94}$.
633: If the planetesimals are loose aggregates so that
634: $Q_0 < 3 \times 10^7$ (the value
635: for basalt rocks or water ice), the limiting size $R_{\rm mat}$ is smaller.
636: 
637: 
638: \section{Accretion/destruction of planetesimals in an evolving disk}
639: 
640: We now put these various critical physical radii in the context of the
641: evolution of the protoplanetary disk.  In order to investigate the
642: effect of departures from power-law relations of the surface density
643: and temperature profiles in real disks, we also present in this section
644: results obtained from a 1D disk model that includes an
645: $\alpha$-viscosity and photoevaporation \citep[see][]{Guillot06,Hueso05}. 
646: The parameters used in the model presented
647: here are a turbulent viscosity $\alpha=0.01$ and an evaporation
648: parameter $T_{\rm atm}=100\,$K (the temperature of the
649: evaporation part of the outer disk).  
650: Another choice of the parameters would affect
651: the results only marginally.
652: 
653: In the numerical calculation in this section, 
654: we evaluate the equilibrium eccentricities $e_{\rm eq}$ as a
655: function of planetesimal radius by solving the following relation:
656: \begin{equation}
657: \tau_{\rm exc}^{-1}=\tau_{\rm tidal}^{-1}
658:   +\tau_{\rm drag}^{-1}+\tau_{\rm coll}^{-1},
659: \end{equation}
660: where the different timescales are given by eqs.~(\ref{eq:t_exc}) to
661: (\ref{eq:t_coll}). 
662: The survival physical radius for accretion $R_{\rm acc}$ is then found, 
663: for each orbital radius in the protoplanetary
664: disk and for each timestep, by solving the equation,
665: \begin{equation}
666: e_{\rm eq}(R_{\rm acc})=e_{\rm acc}(R_{\rm acc}),
667: \end{equation}
668: where $e_{\rm acc}$ is given by eq.~(\ref{eq:e_acc}).
669: When the mean kinetic energy
670: is larger than the strength of a planetesimal, the 
671: a collision is highly erosive.
672: We then obtain the range $R < R_{\rm dis}$ corresponding to
673: the highly erosive collisions by solving the equation,
674: \begin{equation}
675: e_{\rm eq}(R_{\rm dis}) = {\sqrt{2Q_{\rm D}(R_{\rm dis})}\over v_{\rm K}},
676: \end{equation}
677: where $Q_{\rm D}$ is given by eq.~(\ref{eq:Q_D}).
678: 
679: Figures~\ref{fig:accrete_gam1d-2} to \ref{fig:accrete_gam1d-4} show
680: our results for three values of the turbulent excitation parameter,
681: $\gamma=10^{-2}$, $10^{-3}$ (our fiducial value), and $10^{-4}$. Each
682: figure shows, for three orbital distances, 1, 5 and 30\,AU, the
683: planetesimal physical radii corresponding to 
684: accretive and erosive regions are
685: plotted as a function of the total mass remaining in the disk. 
686: Since the disk mass decreases with time as a result of viscous 
687: evolution and photoevaporation, a decrease in disk mass
688: corresponds to evolution in time.
689: As shown in the previous section, planetesimal accretion
690: becomes easier as the disk becomes less massive simply because the
691: turbulent excitation, directly proportional to the
692: local surface density of the gas, becomes weaker. 
693: However, after some point, the disk becomes too light to provide a
694: sufficient amount of gas to form Jupiter-mass gas giants.
695: 
696: The figures also show as thin black lines the values obtained for a
697: disk that follows the slope in surface density versus orbital distance
698: defined for the MMSN (eq.~\ref{eq:Sigma_g})
699: as a function of a disk mass.
700: The disk mass in this model is given by 
701: $3.4 \times 10^{-2} f_g (r_{\rm edge}/100{\rm AU})^{1/2} M_{\odot}$,
702: where $r_{\rm edge}$ is the outer edge radius of the disk.
703: Although the original MMSN model by \citet{hayashi81} used 
704: $r_{\rm edge} = 35$ AU,
705: we here adopt $r_{\rm edge} = 1000$ AU
706: (for comparison, our fiducial alpha disk model with $T_{\rm atm}=100$\,K
707: extends up to a maximum of 350\,AU). 
708: 
709: These are found to be in excellent agreement with the analytical 
710: expressions
711: derived in the previous section, with small differences arising from
712: the simplifications inherent to the analytical approach. 
713: Larger differences are found between the power-law disk and the
714: $\alpha$-disk models mostly because of the difference in slopes
715: ($d\ln\Sigma/d\ln r=-3/2$ for the former, $\sim -1$ for the
716: $\alpha$-disk) which implies that a given disk mass does not
717: correspond to the same surface density with two models, 
718: the difference being larger
719: at smaller orbital distances. 
720: However, the qualitative features of accretive and erosive
721: regions are similar to each other.
722: It should be noted that the models also
723: differ in their temperature profiles, but this is found to be less
724: important.  
725: 
726: 
727: In the highly turbulent regime presented in
728: fig.~\ref{fig:accrete_gam1d-2}, a self-sustained regime of accretion
729: becomes possible only when planetesimals have become very
730: large/massive, with sizes generally well over 100km. This case also
731: yields a sustained area of high erosion where
732: the average kinetic energies of
733: planetesimals are above their internal energies. It is
734: difficult to imagine how planetary cores can form in this context
735: especially if they have to grow large enough to form giant planets. 
736: 
737: With smaller perturbations from the turbulent disk
738: (fig.~\ref{fig:accrete}), planetesimals have more possibilities to
739: accrete: with time, as the disk mass decreases, 
740: the inner disk rapidly moves out of the highly
741: erosive regime, while erosion still remains important at large orbital
742: distances. 
743: With a turbulence strength parameter $\gamma=10^{-4}$, corresponding to
744: a very weak turbulence (fig.~\ref{fig:accrete_gam1d-4}), the presence
745: of a highly erosive regime centered around $\sim 300$m planetesimals
746: is limited only to the outer regions ($\ga 30$ AU), 
747: and the zone rapidly shrinks
748: as the circumstellar gaseous disk disappears. 
749: 
750: In order to put these findings into context, we also show the mass of
751: the disk when Jupiter is believed to have started accreting its
752: gaseous envelope. These values are calculated by assuming that the
753: planet growth has been limited mostly by viscous diffusion in the disk,
754: with the protoplanet capturing between 10\% and 70\% of the mass flux
755: at its orbital distance in the disk evolution model 
756: by \citet{Hueso05} \citep[for details, see][]{Guillot06}.
757: If giant planets have to form, 
758: at some time 
759: corresponding to the disk mass interval defined by the hashed areas in
760: figs.~\ref{fig:accrete_gam1d-2} to \ref{fig:accrete_gam1d-4},
761: protoplanetary cores must be already large enough
762: to start accreting the
763: surrounding hydrogen and helium gas.
764: 
765: We can now define three important disk masses, and their corresponding
766: disk ages (with the caution that ages are inherently model-dependent
767: and are provided here for illustrative purposes only, on the basis of
768: our particular model of $\alpha$-disk evolution with
769: photoevaporation):
770: \begin{enumerate}
771: \item The maximum mass of the disk, following the collapse of the
772:   molecular cloud. This mass can vary quite significantly from one disk
773:   formation/evolution model to another. 
774:   For the particular model shown here, it is
775:   of the order of $0.25\,\rm M_\odot$, for an age of 0.6\,Myrs. 
776: \item The disk mass necessary for Jupiter to grow to its present mass
777:   if it captures 10\% of the mass flux at its orbital distance. For
778:   realistic disk models, this depends weakly on parameters such as
779:   $\alpha$ and the disk evaporation rate. In our case, it corresponds
780:   to $M_{\rm disk}=0.035\,\rm M_\odot$ and an age of 1.95\,Myrs.
781: \item The disk mass necessary for Jupiter to grow to its present mass
782:   if it captures 70\% of the mass flux at its orbital distance. For
783:   our model, $M_{\rm disk}=0.0054\,\rm M_\odot$ (about 5 times the
784:   mass of Jupiter) and an age of 2.85\,Myrs.
785: \end{enumerate}
786: 
787: 
788: Table~\ref{tab:radii} provides the values of the physical
789: radii that define the
790: accretive regime and the highly erosive (disruptive) regime of
791: figs.~\ref{fig:accrete_gam1d-2} to \ref{fig:accrete_gam1d-4}, namely
792: $R_{\rm acc}$ and $R_{\rm dis}$.  In our
793: Solar System, the existence of Jupiter implies that either turbulence
794: was low, the planet grew from a protoplanetary core formed in the
795: inner solar system, or a mechanism was able to lead to
796: the rapid formation of embryos larger than 240 km in radius 
797: at 5 AU (17 km in
798: the low-turbulence case, 1080 km in the high-turbulence case) 
799: by the time when the
800: disk mass had decreased to $5\times 10^{-3}\,\rm M_\odot$. In the last
801: case, it appears that a mechanism such as the standard gravitational
802: instability \citep[e.g.,][]{Safronov69,GW73} would not work
803: because of the turbulence, but formation of relatively large
804: protoplanets in eddies or 
805: vortexes \citep[e.g.,][]{Johansen07,Barge95} is a promising possibility.
806: 
807: 
808: \section{Conclusion and discussion}
809: 
810: We have investigated the critical physical radii for collisions between
811: planetesimals to be accretional ($R > R_{\rm acc}$)
812: or disruptive ($R < R_{\rm dis}$) in turbulent disks,
813: as functions of turbulent strength 
814: ($\gamma \sim O(\Delta \rho/\rho)$),
815: disk gas surface density, and orbital radius. 
816: The results presented here highlight 
817: the fact that MRI turbulence poses a great
818: problem for the growth of planetesimals: generally, only those with
819: sizes larger than a few hundred km are in a clearly
820: accretive regime for a nominal value of $\gamma \sim 10^{-3}$. 
821: The others generally collide with velocities greater than their own surface
822: escape velocities. 
823: For some of them, more severely in the
824: kilometer-size regime, collisions are likely to be disruptive. 
825: The problem is greater when the disk is still massive and at
826: large orbital distances.
827: Also, if turbulence is stronger than $\gamma \sim 10^{-2}$,
828: planetesimal accretion becomes extremely difficult.
829: 
830: However, the rate of occurrence of
831: extrasolar giant planets around solar-type stars is inferred to be as
832: large as $\sim 20$\% \citep{Cumming08}, and depends steeply on
833: the metallicity of the host star \citep{Fischer05,santos04}. This
834: strongly suggests that the majority of extrasolar giant planets were
835: formed by core accretion followed by gas accretion onto the cores
836: \citep{IL05}.  
837: Thus, planetesimals should commonly grow to planetary masses
838: before the disappearance of gas in protoplanetary disks.
839: 
840: The possibilities to overcome the barrier are in
841: principle as follows (their likelihood is commented below):
842: \begin{enumerate}
843: \item Large $M$: 
844: Large planetesimals with sizes of 100 to 1000\,km are
845: formed directly in turbulent environment by a mechanism
846: other than collisional coagulation, jumping over the erosive
847: regime for physical radii.
848: 
849: \item Small $\Sigma_g$: Planetesimals start their accretion to
850:   planet-size only after the disk surface density of gas has declined
851:   to sufficiently small values.
852: 
853: \item Small $\gamma$: Planetesimals form in MRI-inactive regions
854:   (``dead zones'') of protoplanetary disks.
855: \end{enumerate}
856: 
857: Concerning point 1, 
858: the first-born planetesimals with sizes larger than $R_{\rm acc}$ may be formed
859: rapidly by an efficient capture of $\sim$meter-size boulders in
860: vortexes \citep{Johansen07}.  Such large planetesimals may be
861: consistent with the size distribution of asteroids (Morbidelli et
862: al. 2008).
863: Even if the first-born planetesimals are not as large, a small
864: fraction of them could continue to grow larger than 
865: $R_{\rm acc}$ by accreting smaller bodies, because 
866: accretion is not completely cut off as soon as
867: $v_{\rm disp} \ga v_{\rm esc}$
868: (there is always a small possibility for accretion) and the large
869: planetesimals would not be disrupted by smaller ones.  This
870: possibility, however, must be examined by a more detailed growth model
871: taking into account the effect of fragmentation and the size distribution of
872: planetesimals, which we neglected in this paper.
873: 
874: Concerning point 2, 
875: we have shown that planetesimals are most fragile at early times, in
876: massive disks, and at large orbital distances. We therefore suggest
877: that the growth towards planet sizes may be delayed due to MRI
878: turbulence, and then proceed from inside out: planetesimals should
879: start accretion first close to the star, then progressively at larger
880: orbital distances, as the gas surface density declines.  The
881: possibility to delay planet formation while keeping non-migrating
882: km-size planetesimals is noteworthy because it would help planetary
883: systems resisting to type-I migration: they would grow in a gas disk
884: that is less dense, and for which migration timescales may be
885: considerably increased.  \citet{IL08}, \citet{Alibert05}, and
886: \citet{Kominami06} showed that type-I migration must be 
887: lowered by one to two orders of magnitude from
888: the linear calculation \citep{Tanaka02} to provide an explanation
889: for the existence of a population of giant planets in agreement with
890: observations.  This ``late formation'' scenario is 
891: consistent with the noble gas enrichment in Jupiter
892: \citep{Guillot06}. 
893: However, in order to form gas giants,
894: core accretion and gas
895: accretion onto the cores must proceed fast enough to capture
896: Jupiter-mass amount of gas from the decaying gas disk.
897: Once the size of the largest planetesimals exceeds $\sim 1000$km,
898: their eccentricities are damped by tidal drag and dynamical friction 
899: from small bodies.  
900: Most of the other small bodies
901: may be ground into sizes smaller than 1\,km and their eccentricities
902: could be kept very small by gas drag and collision damping.  This
903: could facilitate the runaway accretion of cores to become large enough
904: ($\ga 10 M_{\oplus}$) for the onset of runaway gas accretion.  This
905: issue also has to be addressed by a detailed planetesimal growth model
906: taking into account a size distribution.
907: The likelihood of relatively rapid gas accretion without long ``phase 2''
908: is discussed by \citet{Shiraishi}.
909: If planetesimal sizes are relatively small, 
910: gas drag damping opens up a gap in the planetesimal disk 
911: around the orbit of a core and truncates planetesimal accretion 
912: onto the core.  The truncation of heating due to planetesimal
913: bombardment enables the core to efficiently accrete disk gas.
914: 
915: Concerning point 3, 
916: the MRI inactive region (``dead zones'') may exist in inner disk
917: regions in which the surface density is large enough to prevent cosmic and
918: X rays from penetrating the disk \citep{gammie96,sano00}.  The
919: preservation of a dead zone can also contribute to stall type-I
920: migration by converting it to type-II migration \citep{Matsumura07}
921: or by creating a local region with a positive radial gradient of disk
922: pressure near the ice line \citep{IL08b}.  However, dead zones can be
923: eliminated by turbulent mixing/overshoot
924: \citep{Varniere06,Turner07,Ilgner08}, a self-sustaining mechanism
925: \citep{Inutsuka05}, and dust growth \citep{sano00}.  The last effect
926: comes form the fact that small dust grains are the most efficient
927: agents for charge recombination.  
928: According to grain growth, the ionization of the disk and 
929: its coupling with the magnetic field become stronger to activate
930: MRI turbulence.
931: We remark that if MRI turbulence is
932: activated, collisions are disruptive and they re-produces small grains 
933: to decrease the ionization degree.
934: This self-regulation process might maintain a marginally dead state
935: and keep producing small dust grains.  This might be related with
936: relative chronological age difference ($\sim 2$Myr) between chondrules
937: and CAIs \citep[e.g.,][]{Kita05}.  Whether dead zones exist or not is
938: one of the biggest issues in evolution of protoplanetary disks and
939: planet formation.  A more detailed analysis of planetesimal accretion in
940: turbulent disks could impose a constraint on this issue.
941:  
942: At large orbital distances (10's of AU), the existence of a highly
943: erosive regime that lasts until late in the evolution of the
944: protoplanetary disk is an important feature of this scenario.
945: It shows that the entire mass of solids is highly reprocessed by
946: collisions, in qualitative agreement with the paucity of presolar
947: grains (intact remnants from the molecular cloud core) found in
948: meteorites.
949: It also prevents the growth of large
950: planetesimals and helps to maintain a large population of small grains
951: in the disks.  This is in qualitative agreement with observations that do
952: not indicate a significant depletion of micron-sized grains with time,
953: contrary to what would be predicted in the absence of turbulence 
954: \citep{Dullemond05,Tanaka05}.
955: 
956: In conclusion, the existence of MRI turbulence may be a threat to
957: planetesimal accretion.  Given the uncertainties related to these explanations,
958: we cannot provide a definitive scenario for the formation of
959: protoplanetary cores. However, it offers several promising hints to
960: explain important features of planet formation as constrained by
961: today's observations of protoplanetary disks, exoplanets and
962: meteoritic samples in the Solar System.
963: 
964: 
965: \acknowledgments
966: 
967: This research was supported by the Sakura program
968: between Japan and France, and by the CNRS
969: interdisciplinary program {\it ``Origine des plan\`etes et de la
970:   Vie''} through a grant to T.G. and A.M.
971: 
972: 
973: 
974: \section*{Appendix}
975: 
976: Here we derive the $r$-dependence in eq.~(\ref{eq:e_random}).
977: If the equation of motion is scaled by a reference radius $r_1$
978: and $T_{\rm K1}$ where $T_{\rm K1}$ is a Keplerian period at $r_1$,
979: the only remaining non-dimensional parameter in the equations is 
980: \begin{equation}
981: \gamma \Gamma(r_1) = \gamma \frac{64 \Sigma_g r^2}{\pi M_{\odot}}|_{r_1}
982: \end{equation} 
983: (see eqs.~[4], [5], [6] in \citet{ogi07}).
984: Consider the equation of motion scaled by $r_1$ and $T_{\rm K1}$
985: and that scaled by $r_2$ and $T_{\rm K2}$.  
986: If $\gamma$ is the same and $\Gamma(r_1)=\Gamma(r_2)$, 
987: these two scaled-equations of motion are identical and
988: evolution of eccentricity, which is a non-dimensional quantity, 
989: must be identical in terms of the scaled time.
990: Note that the magnitude of excited $e$ should 
991: be proportional to $\gamma \Gamma$.     
992: Since 
993: \begin{equation}
994: \Gamma(r) = \frac{\Sigma_g(r)}{\Sigma_g(r_1)} 
995:             \left(\frac{r}{r_1}\right)^2 \Gamma(r_1),
996: \label{eq:Gamma_r}
997: \end{equation} 
998: and Ogihara et al.'s eq.~(34) derived for
999: $r = 1$AU is proportional to $\gamma \Gamma(1{\rm AU})$,
1000: the formula for arbitrary $r$ is given by
1001: replacing a year by $T_{\rm K}(r)$ and 
1002: $\gamma$ by $\gamma (\Gamma(r)/\Gamma(1{\rm AU}))$ in their
1003: equation.  As a result,
1004: \begin{equation}
1005: e \sim 0.1 \gamma 
1006:   \left(\frac{\Sigma_g}{\Sigma_{g,1}}\right)
1007:   \left(\frac{r}{1{\rm AU}}\right)^{2}
1008:   \left(\frac{t}{T_{\rm K}}\right)^{1/2},
1009: \end{equation}
1010: where $\Sigma_{g,1}$ is $\Sigma_{g}$ at 1AU with $f_g = 1$
1011: (eq.~[\ref{eq:Sigma_g}]) and the numerical factor was
1012: corrected as explained in the footnote in \S 3.1.
1013: Assuming the simple power-law model defined by eq.~(7),
1014: \begin{equation}
1015: e \sim 0.1 f_g \gamma 
1016:   \left(\frac{r}{1{\rm AU}}\right)^{-1/4}
1017:   \left(\frac{t}{1{\rm year}}\right)^{1/2}.
1018: \end{equation}
1019: 
1020: \clearpage
1021: 
1022: \begin{thebibliography}{}
1023: 
1024: \bibitem[Adachi et al.(1976)]{Adachi76}
1025: Adachi, I., Nakazawa, K., \& Hayashi, C. 1976, PASJ, 29, 163
1026: 
1027: \bibitem[Alibert et al.(2005)]{Alibert05}
1028: Alibert, Y., Mousis, O., Mordasini, C. \& Benz, W.
1029: 2005, \apjl, 626, 57
1030: 
1031: \bibitem[Balbus \& Hawley(1991)]{bal91} 
1032: Balbus, S. A., \& Hawley, J. F. 1991,
1033: \apj, 376, 214
1034: 
1035: \bibitem[Barge \& Someria(1995)]{Barge95} 
1036: Barge \& Someria, 1995, \aa, 295, 1
1037: 
1038: \bibitem[Benz \& Asphaug(1999)]{Benz_Asphaug99}
1039: Benz, W. \& Asphaug, E. 1999, Icarus, 142, 5
1040: 
1041: \bibitem[Britsch et al.(2008)]{britsch08}
1042: Britsch, M., Clarke, C. J. \& Lodato, G. 2008, MNRAS, 385, 1067
1043: 
1044: \bibitem[Cumming et al.(2008)]{Cumming08}
1045: Cumming, A. Marcy, G.W. Butler, R.P. Fischer, D.A. Vogt, S.S. \& Wright,
1046: J.T.  2008, in preparation.
1047: 
1048: \bibitem[Daisaka et al.(2006)]{Kominami06}
1049: Daisaka, K. J., Tanaka, H. \& Ida, S. 2006, Icarus, 185, 492 
1050: 
1051: \bibitem[Dullemond \& Dominik(2005)]{Dullemond05}
1052: Dullemond, C. P. \& Dominik, C. 2005,
1053: A\&A, 434, 971
1054: 
1055: \bibitem[Fischer \& Valenti(2005)]{Fischer05}
1056: Fischer, D. A. \& Valenti, J. A. 2005.
1057: \apj, 622, 1102
1058: 
1059: \bibitem[Gammie(1996)]{gammie96} 
1060: Gammie, C. F., 1996,
1061: Layered accretion in T Tauri disks, \apj, 457, 355
1062: 
1063: \bibitem[Goldreich \& Ward(1973)]{GW73}
1064: Goldreich, P. \& Ward, W. 1973,
1065: \apj, 
1066: 
1067: \bibitem[Guillot \& Hueso(2006)]{Guillot06}
1068: Guillot, T. \& Hueso, R. 2006,
1069: MNRAS, 367, L47
1070: 
1071: \bibitem[Hayashi(1981)]{hayashi81} 
1072: Hayashi, C., 1981,
1073: Prog. Theor. Phys. Suppl, 70, 35
1074: 
1075: \bibitem[Hueso \& Guillot(2005)]{Hueso05}
1076: Hueso, R. \& Guillot, T. 2005,
1077: A\&A, 442, 703
1078: 
1079: \bibitem[Ida(1990)]{I90}
1080: Ida, S. 1990, Icarus, 88, 129
1081: 
1082: \bibitem[Ida \& Lin(2005)]{IL05}
1083: Ida, S. \& Lin, D. N. C. 2005,
1084: \apj, 626, 1045
1085: 
1086: \bibitem[Ida \& Lin(2008a)]{IL08}
1087: ------. 2008,
1088: \apj, 673, 487
1089: 
1090: \bibitem[Ida \& Lin(2008b)]{IL08b}
1091: ------. 2008, \apj, in press
1092: 
1093: \bibitem[Ida \& Makino(1993)]{IM93}
1094: Ida, S. \& Makino, J. 1993,
1095: Icarus, 106, 210
1096: 
1097: \bibitem[Ida \& Nakazawa(1989)]{IN89}
1098: Ida, S. \& Nakazawa, K. 1989,
1099: \aa, 224, 303
1100: 
1101: \bibitem[Ilgner \& Nelson(2008)]{Ilgner08}
1102: Ilgner, M., Nelson, R.~P., 2008, 
1103: astro-ph:0802.4409
1104: 
1105: \bibitem[Inaba \& Ikoma(2003)]{Inaba03}
1106: Inaba, S. \& Ikoma, M.
1107: \aa, 410, 711
1108: 
1109: \bibitem[Inutsuk \& Sann(2005)]{Inutsuka05}
1110: Inutsuka, S. \& Sano, T.
1111: \aa, 410, 711
1112: 
1113: \bibitem[Johansen et al.(2007)]{Johansen07}
1114: Johansen, A., Oishi, J. S., Low, M. M., Klahr, H.,
1115: Thomas, H. \& Youdin, A. 2007.
1116: Nature, 448, 1022
1117: 
1118: \bibitem[Kita et al.(2005)]{Kita05}
1119: Kita, N. T. et al. 2005,
1120: in Chondrites \& the protoplanetary disk,  
1121: ASP conf. series 341, pp. 558-587
1122: 
1123: \bibitem[Johnson et al.(2006)]{johnson06} 
1124: Johnson, E. T., Goodman, J. \& Menou, K. 2006,
1125: \apj, 647, 1413
1126: 
1127: \bibitem[Laughlin et al.(2004)]{lau04}
1128: Laughlin, G., Steinacker, A., \& Adams, F. C. 2004, 
1129: \apj, 608, 489
1130: 
1131: \bibitem[Matsumura et al.(2007)]{Matsumura07}
1132: Matsumura, S., Pudritz, R. E. \& Thommes, E. W. 2007,
1133: \apj, 660, 1609
1134: 
1135: \bibitem[Nelson(2005)]{nel05} Nelson, R. P., 2005,
1136: \aap, 443, 1067
1137: 
1138: \bibitem[Nelson \& Papaloizou(2004)]{nel04} 
1139: Nelson, R. P., \& Papaloizou, J. C. B., 2004,
1140: \mnras, 350, 849
1141: 
1142: \bibitem[Ohtsuki (1993)]{ohtsuki93} 
1143: Ohtsuki, K. 1993, Icarus, 106, 228 
1144: 
1145: \bibitem[Ogihara et al.(2007)]{ogi07} 
1146: Ogihara, M., Ida, S. \& Morbidelli, A. 2007,
1147: Icarus, 188, 522
1148: 
1149: \bibitem[Rice \& Armitage(2003)]{ric03} Rice, W. K. M., Armitage, P J., 2003,
1150: \apj, 598, 55
1151: 
1152: \bibitem[Safronov(1969)]{Safronov69}
1153: Safronov, V. 1969,
1154: Evolution of thr Protoplanetary Cloud and Formation of
1155: the Earth and Planets (Moscow: Nauka Press)
1156: 
1157: \bibitem[Santos et al. (2004)]{santos04} Santos, N., Israelian, G., \&
1158:   Mayor, M., 2004, \aa, 415, 1153
1159: 
1160: \bibitem[Sano et al.(2000)]{sano00} Sano, T., Miyama, S. M., Umebayashi, T., \& Nakano, T., 2000,
1161: \apj, 543, 486
1162: 
1163: \bibitem[Shakura \& Sunyaev(1973)]{alpha}
1164: Shakura, N. I. \& Sunyaev, R. A. 1973, A\&A, 24, 337
1165: 
1166: \bibitem[Shiraishi \& Ida(2008)]{Shiraishi} 
1167: Shiraishi, M. \& Ida, S. 2008,
1168: \apj, in press.
1169: 
1170: \bibitem[Tanaka et al.(2002)]{Tanaka02} 
1171: Tanaka, H., Takeuchi, T. \& Ward, W. R., 2002,
1172: \apj, 565, 1257
1173: 
1174: \bibitem[Tanaka et al.(2005)]{Tanaka05}
1175: Tanaka, H., Himeno, Y.  \& Ida, S. 2005, \apj, 625, 414
1176: 
1177: \bibitem[Tanaka \& Ward(2004)]{tan04} Tanaka, H., \& Ward, W. R., 2004,
1178: \apj, 602, 388
1179: 
1180: \bibitem[Turner et al.(2007)]{Turner07} 
1181: Turner, N. J., Sano, T. and Dziourkevitch, N., 2007,
1182: \apj, 659, 729
1183: 
1184: \bibitem[Varni{\`e}re \& Tagger(2006)]{Varniere06} 
1185: Varni{\`e}re, P. \& Tagger, M., 2007,
1186: \apj, 446, L13
1187: 
1188: 
1189: 
1190: \end{thebibliography}
1191: 
1192: \clearpage
1193: %Radii Accretion/destruction (in km)
1194: \begin{table}
1195: \caption{
1196: Limits of the accretive regime ($R > R_{\rm acc}$) 
1197: and the highly erosive (disruptive) 
1198: regime $R < R_{\rm dis}$, as a function of orbital distance and disk mass.}
1199: \begin{tabular}{r|cc|cc|cc} \hline
1200:  
1201:   & \multicolumn{2}{c|}{1 AU}  &  \multicolumn{2}{c|}{5 AU}  & \multicolumn{2}{c}{30 AU}  \\
1202: \cline{2-3} \cline{4-5} \cline{6-7}
1203: $M_{\rm disk}$ & $R_{\rm acc}$ & $R_{\rm dis}$  & $R_{\rm acc}$ & $R_{\rm dis}$  & $R_{\rm acc}$  & $R_{\rm dis}$ \\
1204: $\rm [M_\odot]$  & [km] & [km] & [km] & [km] & [km] & [km] \\
1205: \hline
1206: \multicolumn{7}{l}{Fiducial case: $\gamma= 10^{-3}$; $\rho_{\rm p}=3\,{\rm g\,cm^{-3}}$; $Q_0=3\times 10^7$; $B=1.0$}\\ \hline
1207:       0.25 &   280. &  $[1.14-0.061]$ &  1280. &  $[82.-0.00047]$ &  1680. &  $[1000.-0.00047]$ \\
1208:      0.035 &   86. &      $--$       &   440. &  $[5.3-0.026]$ &    850.  &  $[200.-0.0024]$ \\
1209:     0.0054 &   46. &      $--$       &   240. &  $[1.05-0.057]$ &    590.  &  $[37.-0.0069]$ \\
1210: \hline
1211: \multicolumn{7}{l}{Low turbulence case: $\gamma= 10^{-4}$; $\rho_{\rm p}=3\,{\rm g\,cm^{-3}}$; $Q_0=3\times 10^7$; $B=1.0$}\\ \hline
1212:      0.25 &    16. &      $--$       &   150. &  $[0.30-0.0099]$ &   590.  &  $[22.-0.0099]$ \\
1213:     0.035 &    3.9 &      $--$       &   36. &     $--$         &   220.  &  $[1.07-0.053]$ \\
1214:    0.0054 &    1.7 &      $--$       &   17. &     $--$         &   103.  &      $--$ \\
1215: \hline
1216: \multicolumn{7}{l}{High turbulence case: $\gamma= 10^{-2}$; $\rho_{\rm p}=3\,{\rm g\,cm^{-3}}$; $Q_0=3\times 10^7$; $B=1.0$}\\ \hline
1217:      0.25 &   2540. &  $[300.-0.0026]$ & 3920. &  $[2770.-2\times 10^{-5}]$ &  4250. & $[3220.-2\times 10^{-5}]$\\
1218:     0.035 &   870.  &  $[26.-0.0090]$ & 1590. &  $[710.-0.0012]$ &            2180. & $[1460.-0.00012]$ \\
1219:    0.0054 &   510.  &  $[7.4-0.018]$ &  1080. &  $[230.-0.0025]$ &            1580. & $[970.-0.00033]$\\
1220: \hline
1221: \multicolumn{7}{l}{High material resistance: $\gamma= 10^{-3}$; $\rho_{\rm p}=3\,{\rm g\,cm^{-3}}$; $Q_0=10^8$; $B=2.0$}\\ \hline
1222:      0.25 &   280.  &  $[0.26-0.15]$  &  1280. &  $[27.-0.0013]$  &  1680. &   $[700.-0.0013]$ \\
1223:     0.035 &   86.  &     $--$        &   440. &  $[1.6-0.068]$ &    850. &   $[66.-0.0068]$ \\
1224:    0.0054 &   46.  &     $--$        &   240. &  $[0.27-0.15]$ &     590. &   $[12.-0.019]$\\
1225: \hline 
1226: \multicolumn{7}{l}{Low material resistance:$\gamma= 10^{-3}$; $\rho_{\rm p}=3\,{\rm g\,cm^{-3}}$; $Q_0=10^7$; $B=0.3$}\\ \hline
1227:      0.25 &   280.  &  $[11.-0.026]$ &   1280. &  $[550.-0.00018]$ & 1680. &   $[1440.-0.00018]$ \\
1228:     0.035 &   86.  &  $[0.56-0.087]$ &   440. &  $[41.-0.011]$    &  850. &   $[590.-0.00097]$ \\
1229:    0.0054 &   46.  &     $--$        &   240. &  $[9.2-0.024]$  &    590. &   $[240.-0.0028]$
1230: \end{tabular}
1231: \label{tab:radii}
1232: \end{table}
1233: 
1234: \clearpage
1235: 
1236: \begin{figure}
1237: \epsscale{0.8}
1238: \plotone{f1}
1239: \caption{Evolution of the eccentricity (left panels) and of the
1240:   semimajor axis (right panels) as a function of time, 
1241:   for different planetesimal masses 
1242:   (top to bottom). 
1243:   A single planetesimal is integrated in a turbulent disk and
1244:   5 independent runs 
1245:   with different random number seeds 
1246:   for the generation of turbulent density fluctuations are shown in each panel.
1247:   {\it Top panels\/}:
1248:   We consider no tidal and gas drag. 
1249:   {\it Middle panels\/}: Results
1250:   including gas drag for planetesimals of
1251:   $M=10^{-6}M_{\oplus}$. 
1252:   {\it Bottom panels\/}: Results
1253:   for planetesimals of $M=10^{-12}M_{\oplus}$. In all cases, we
1254:   assume an initial orbital distance of 1AU, $f_g=1$ and
1255:   $\gamma=0.01$.}
1256: \label{fig:e_evol}
1257: \end{figure}
1258: 
1259: \clearpage
1260: 
1261: \begin{figure}
1262: \epsscale{0.8}
1263: \plotone{f2}
1264: \caption{Equilibrium eccentricities ($e_{\rm eq}$)
1265: as a function of the physical radius $R$ of planetesimals
1266: (solid lines).
1267: The values of $e_{\rm eq}$ is determined by the minimum of
1268: individual equilibrium eccentricities, $e_{\rm tidal}, e_{\rm drag}$, 
1269: and $e_{\rm coll}$.
1270: The critical values for accretion and destruction,
1271: $e_{\rm acc}$ and $e_{\rm dis}$, are also plotted by
1272: dashed and dotted lines, respectively.
1273: At $r=1$AU, the bulk density $\rho_{\rm p} = 3 {\rm gcm}^{-3}$ is assumed,
1274: while $\rho_{\rm p} = 1 {\rm gcm}^{-3}$ at 5 AU.
1275: The mass of the planetesimals is given by
1276: $M = 2.1 \times 10^{-3} (R/10^3{\rm km})^3 
1277: (\rho_{\rm p}/3 {\rm gcm}^{-3}) M_{\oplus}$.
1278: }
1279: \label{fig:e_eq}
1280: \end{figure}
1281: 
1282: \clearpage
1283: 
1284: \begin{figure}
1285: \epsscale{0.7}
1286: \plotone{f3}
1287: \caption{
1288:   Evolution of characteristic physical radii of planetesimals 
1289:   as a function of disk mass,
1290:   at several orbital distances in the disk: 1AU (top), 5AU (middle)
1291:   and 30AU (bottom). 
1292:   The solid and dot-dashed curves correspond to the boundary radius 
1293:   for accretion regime ($R_{\rm acc}$) and to that of the highly erosive 
1294:   regime ($R_{\rm dis}$).
1295:   Two disk models have been used: a
1296:   simple power-law model with $d\ln \Sigma_g/d\ln r = -3/2$
1297:   (eq.~[\ref{eq:Sigma_g}])
1298:   with an outer cut-off radius of 1000AU
1299:   (thin lines), and an alpha-disk
1300:   model with $\alpha=0.01$ and $T_{\rm atm}=100\,$K 
1301:   \citep[see][]{Guillot06} (thick lines). The hashed region corresponds to 
1302:   the range of disk mass (equivalently, the range of time if
1303:   disk evolution is given)
1304:   during which Jupiter must start accreting hydrogen/helium gas 
1305:   (assuming it grabs between 10\% and 70\% of the disk mass 
1306:   flux at its orbital
1307:   distance). In the simulations, MRI turbulence is supposed to be
1308:   high, with $\gamma=10^{-2}$. We also choose $Q_0=3\times 10^7$
1309:   and $B=1$ (see eq.~[\ref{eq:Q_D}]). }
1310: \label{fig:accrete_gam1d-2}
1311: \end{figure}
1312: 
1313: \clearpage
1314: 
1315: \begin{figure}
1316: \epsscale{0.7}
1317: \plotone{f4}
1318: \caption{Evolution of characteristic physical radii 
1319:   $R_{\rm acc}$ and $R_{\rm dis}$ 
1320:   of planetesimals as a function of disk
1321:   mass. The parameters and labels are the same as 
1322:   those in fig.~\ref{fig:accrete_gam1d-2} but for a medium
1323:   turbulence case ($\gamma=10^{-3}$) (our fiducial case).}
1324: \label{fig:accrete}
1325: \end{figure}
1326: 
1327: \begin{figure}
1328: \epsscale{0.7}
1329: \plotone{f5}
1330: \caption{Evolution of characteristic physical radii 
1331:   $R_{\rm acc}$ and $R_{\rm dis}$ 
1332:   of planetesimals as a function of disk
1333:   mass. The parameters and labels are the same as
1334:   those in fig.~\ref{fig:accrete_gam1d-2} but for a weak
1335:   turbulence case ($\gamma=10^{-4}$).}
1336: \label{fig:accrete_gam1d-4}
1337: \end{figure}
1338: 
1339: \end{document}
1340: 
1341: 
1342: 
1343: 
1344: 
1345: