1: \documentstyle[epsf,epsfig,rotate]{mn}
2:
3: \topmargin -1.5cm
4:
5: \newcommand{\be}{\begin{equation}}
6: \newcommand{\ee}{\end{equation}}
7: \newcommand{\gs}{\;\raisebox{-.8ex}{$\buildrel{\textstyle>}\over\sim$}\;}
8: \newcommand{\ls}{\; \raisebox{-.8ex}{$\buildrel{\textstyle<}\over\sim$}\;}
9:
10: %Journal abbreviations:
11:
12: \newcommand{\anrev}{{\it ARA\&A, }}
13: \newcommand{\apj}{{\it ApJ, }}
14: \newcommand{\apjl}{{\it ApJ(Letters), }}
15: \newcommand{\apjs}{{\it ApJS, }}
16: \newcommand{\aj}{{\it AJ, }}
17: \newcommand{\icar}{{\it Icarus, }}
18: \newcommand{\mnr}{{\it MNRAS, }}
19: \newcommand{\nat}{{\it Nat, }}
20: \newcommand{\sci}{{\it Sci, }}
21: \newcommand{\pasj}{{\it PASJ, }}
22: \newcommand{\pasp}{{\it PASP }}
23: \newcommand{\ana}{{\it A\&A, }}
24: \newcommand{\anas}{{\it A\&AS, }}
25: \newcommand{\anar}{{\it A\&AR, }}
26: \newcommand{\prl}{{\it Phys. Rev. Letters, }}
27: \newcommand{\qjl}{{\it QJRAS, }}
28:
29: \newcommand{\MJup}{M$_J$}
30:
31: \title
32: [Resonant planets in a disc]{Possible Commensurabilities Among
33: Pairs of Extrasolar Planets}
34: \author[R.P.Nelson \& J.C.B.Papaloizou]{Richard P. Nelson \&
35: John C.B. Papaloizou \\
36: Astronomy Unit, Queen Mary, University of London, Mile End Rd, London E1 4NS}
37:
38: \date{Received/Accepted}
39: % \pagerange{\pageref{firstpage}--\pageref{lastpage}}
40:
41: %\def\LaTeX{L\kern-.36em\raise.3ex\hbox{a}\kern-.15em
42: %T\kern-.1667em\lower.7ex\hbox{E}\kern-.125emX}
43:
44: \begin{document}
45: %\label{firstpage}
46:
47: \maketitle
48:
49: \begin{abstract}
50: We investigate the possible
51: commensurabilities to be expected
52: when two protoplanets in the Jovian
53: mass range, gravitationally interacting with each
54: other and an external
55: protoplanetary disc, are driven by disc induced
56: orbital migration of the outer protoplanet
57: into a commensurability which is then
58: maintained in subsequent evolution.
59: We find that for a variety
60: of protoplanet masses
61: and typical protoplanetary disc properties, as well as the
62: setting up of 2:1 commensurabilities of the type recently
63: observed in GJ876,
64: 3:1 commensurabilities are often formed, in addition to
65: 4:1, 5:1, and 5:2 commensurabilities which occur less frequently.
66: The higher order commensurabilites are favoured when
67: either one of the planets is massive, or the inner planet begins with
68: a significant orbital eccentricity.
69: Detection of such commensurabilities could yield
70: important information relating to the operation of protoplanet
71: disc interactions during and shortly post formation.
72:
73: \end{abstract}
74:
75: \begin{keywords} giant planet formation - extrasolar planets -
76: - orbital migration - resonance-protoplanetary discs - stars: individual:
77: GJ876
78: \end{keywords}
79:
80: \section{Introduction} \label{intro}
81: The recent discovery of a pair of extrasolar giant planets orbiting
82: in a 2:1 commensurability
83: around GJ876 (Marcy et al 2001)
84: has raised interesting questions about the post-formation
85: orbital evolution of this system. Assuming the system did not form
86: in its currently observed state, the existence of the commensurability
87: indicates that disc induced orbital migration is likely to have occurred
88: so as to gradually reduce the planetary orbital
89: separation until resonance was established.
90:
91: Simulations of two planets in the Jovian mass
92: range interacting with a disc have been performed
93: by Kley (2000) and Bryden et al (2000).
94: The latter authors found a tendency
95: for the two planets to open up gaps in their local vicinity, and for material
96: between the two planets
97: to be cleared, ending up interior to the inner
98: planet orbit or exterior to the outer planet orbit,
99: such that both planets orbit within an inner cavity.
100:
101: Snellgrove, Papaloizou \& Nelson (2001) (hereafter paper I)
102: performed a
103: simulation of a system consisting
104: of a primary star with two planets
105: moving under their mutual gravitational attraction and forces produced by
106: tidal interaction with an
107: externally orbiting gaseous disc.
108: Angular momentum exchange caused the outermost planet to migrate inwards
109: until a 2:1 commensurability with the inner planet was reached.
110: The subsequent dynamical interaction then resulted in the planets
111: migrating inwards together maintaining the commensurability.
112:
113: The simulation results in paper I could be well matched with
114: those of a simple $N$
115: body integration procedure used below, which incorporated
116: simple prescriptions for the migration and eccentricity damping of
117: the outer planet due to interaction with the disc.
118: Using this we investigate the range of commensurabilities
119: that might be expected in two planet systems as a consequence of protoplanet
120: disc interactions occurring in a standard protoplanetary disc model, where
121: we assume a fixed migration time and consider various eccentricity
122: damping rates.
123:
124: Depending on the nature of the interaction between the disc and the
125: outer planet, as well as the details of the resonance into which
126: the planets enter,
127: we distinguish between four types of orbital evolution
128: in the resonant migration phase which we denote as types A -- D. We
129: denote the orbital elements of the outer planet with a subscript `1'
130: and the inner planet with subscript `2'. Note that we find the evolution of
131: the system depends on the details of the resonance into which the planets
132: become locked. For higher order resonances such as 3:1 and 4:1,
133: commensurabilities may be maintained in which differing resonant angles
134: librate such that they do not
135: cover the full $(0 , 2 \pi ) $ domain
136: (see for example Murray \& Dermott 1999 for an extensive discussion).
137: For a $p:q$ commensurability,
138: we monitor the resonant angles defined by
139: \be \phi_{p,q,k}= p\lambda_1 -q\lambda_2 -p\varpi_1 +q\varpi_2
140: +k(\varpi_1 - \varpi_2).\ee
141: Here $\lambda_1, \lambda_2, \varpi_1$ and
142: $\varpi_2$ denote the mean longtitudes
143: and longtitudes of periapse for the planets $1$ and $2$
144: respectively. The positive integers $p$ and $q$
145: satisfy $p > q,$ and there are $p - q + 1$ possible values
146: of the positive integer $k$ such that $ q \le k \le p.$
147: Although there are $p - q + 1$ corresponding angles, no more than
148: two can be linearly independent.
149: This means that if libration occurs, either all
150: librate or only one librates. Both situations occur in our
151: integrations.
152: \\
153: {\bf Type A:} Eccentricities increase during migration until
154: they reach steady state equilibrium values, with possibly small oscillations
155: superposed. After this migration continues in a self-similar manner
156: and all the resonant angles defined above are in libration. \\
157: {\bf Type B:} Eccentricities
158: increase as the migration proceeds, until
159: $e_1 \ge 0.2$ at which point we estimate the outer planet
160: enters the outer disc.
161: Our simple model breaks down at this point.
162: Large values of $e_1$ arise when the eccentricity damping rate is small.
163: As in type A resonant migration all the resonant angles considered
164: go into libration. The difference is that $e_1$ exceeds $0.2$
165: before a steady state can be reached.
166: An experimental model of the evolution of a
167: two planet system corresponding to type B behaviour, and
168: based on our knowledge of the migration of an eccentric planet
169: interacting with a protoplanetary disc is presented in section~\ref{e2_0_0.1}.
170: \\
171: {\bf Type C:} For higher initial values of $e_2$,
172: we find a mode of evolution in which $e_1$ and $e_2$
173: rise continuously during the migration phase,
174: even for efficient damping of eccentricity. The issue
175: of the outer planet entering the disc again arises in type C migration. The
176: primary difference between type C evolution and types A and B is that
177: only one of the resonant angles defined above
178: is found to librate for the same apparent
179: commensurability, so that the evolution differs. \\
180: {\bf Type D:} This mode of resonant migration corresponds to small
181: values of $e_1 < 0.2 $ being maintained for the outer planet, whilst the
182: eccentricity of the inner planet grows to attain values close to unity.
183: In this respect
184: it differs from the other types of migration. This mode of
185: evolution is related to that described by Beust \& Morbidelli (1996)
186: when discussing the generation of star grazing comets in the $\beta$ Pictoris
187: system. All the resonant angles librate in type D
188: migration but with large amplitude.
189:
190: \section{Conservation of Energy and Angular Momentum}
191: We consider the consequences of a simple
192: application of the conservation
193: of energy and angular momentum.
194: In the case when a near steady state for the orbital
195: eccentricities is attained, a relationship between the orbital eccentricities
196: and the circularization and migration rates induced
197: by the disc is obtained for the case
198: of type A migration.
199:
200: We consider two planets with masses $m_1, m_2,$ osculating
201: semi-major axes $a_1, a_2,$ and eccentricities $e_1 ,e_2.$
202: These orbit a central mass $M_*.$
203: The total angular momentum is
204: \be J= J_1+J_2 = m_1\sqrt{GM_* a_1(1-e_1^2)} + m_2\sqrt{GM_* a_2(1-e_2^2)}\ee
205: and the energy $E$ is
206: \be E = -{GM_* m_1\over 2 a_1} -{GM_* m_2\over 2 a_2} \ee
207: We assume resonant self-similar migration in which $a_2/a_1,$ $e_1,$ and $e_2$
208: are constant. Then conservation of angular momentum gives
209: \be {dJ\over dt} =J_1{1\over 2 a_1}
210: {d a_1\over dt}\left( 1 +
211: {m_2\sqrt{a_2(1-e_2^2)}\over m_1\sqrt{a_1(1-e_1^2)}}\right)
212: =-T \label{CJ} \ee
213: and conservation of energy gives
214:
215: \be {dE\over dt} ={GM_* m_1\over 2 a_1^2} {d a_1\over dt}
216: \left(1+{m_2 a_1\over m_1 a_2}\right)
217: =-{n_1T\over \sqrt{1-e_1^2}} - D , \label{CE}\ee
218: where $n_1= \sqrt{GM_*/a_1^3}$ and
219: we suppose there is
220: a tidal torque $-T$ produced by the disc
221: which acts on $m_1.$
222: In addition we suppose there to be
223: an associated
224: tidally induced orbital energy loss rate
225: which is written as $n_1 T/\sqrt{1-e_1^2} +D,$
226: with
227: $D \equiv (GM_* m_1 e_1^2) /(a_1(1-e_1^2) t_c).$
228: Here $t_c$ is the circularization time of $m_1$
229: that would apply if the tidal torque and energy
230: loss rate acted on the orbit of $m_1$ with $m_2$ being absent.
231: In that case we would have $de_1/dt = -e_1/t_c.$
232: A migration time $t_{mig}$ can be defined through
233: $T= m_1\sqrt{GM_* a_1(1-e_1^2)}/(3 t_{mig}).$
234: This is the time for $n_1$ to increase by a factor of
235: $e$ if $m_2$ was absent and the
236: eccentricity $e_1$ was fixed. Note that $t_c$ and $t_{mig}$
237: are determined by the disc planet tidal interaction
238: and may depend on $e_1.$
239:
240: \noindent By eliminating ${d a_1\over dt}$ from (\ref{CE}) and (\ref{CJ})
241: we can obtain a relationship between $e_1, e_2, t_c,$ and $t_{mig}$ in the form
242: \be e_1^2 =
243: {t_c \left(1-e_1^2
244: -{(1-e_2^2)^{1/2}(1-e_1^2)^{1/2}\over a^{-3/2}_2 a_1^{3/2}}\right)
245: \over 3 t_{mig}\left( {m_1a_2\over m_2a_1} +
246: {a_2^{3/2}(1-e_2^2)^{1/2}\over a_1^{3/2}(1-e_1^2)^{1/2}}\right)}
247: \label{ejcons} \ee
248:
249: We comment that this is general in that it depends only on the conservation laws
250: and applies for any magnitude of eccentricity. However, it's derivation did
251: assume self-similar migration with equilibrium eccentricities
252: and so it does not apply in non equilibrium situations
253: for which the eccentricities grow continuously. Note too that
254: for a 2:1 commensurability for which $(a_1/a_2)^{3/2}=2,$ (\ref{ejcons}) reduces
255: in the case $e_i^2 \ll 1$
256: to the expression given in paper I which was obtained by analysis of the perturbation
257: equations directly:
258:
259: \be e_1^2 = {t_c m_2a_1\over 3 t_{mig}(2 m_1 a_2 + m_2a_1)} \label{ec1} \ee
260: The above determines the eccentricity of the outer planet
261: $e_1$ as a function of $t_c$ and $t_{mig}.$
262:
263: \section{Numerical Calculations}
264: \subsection{Model Assumptions and Physical Setup}
265: The basic assumptions of our model
266: are that the two planets orbit
267: within the inner cavity of a tidally truncated disc that lies
268: exterior to the outer planet.
269: Tidal interaction with this disc causes inwards migration of the outer planet
270: on a time scale of $t_{mig}$,
271: and also leads to eccentricity damping of the outer planet on a time scale of
272: $t_c$. We assume that the inner edge of the disc is such that an eccentricity
273: of $e_1 \; \gs \; 0.2$ will enable the outer planet to enter the disc, in basic agreement with the results of hydrodynamic simulations
274: (e.g. Nelson et al 2000).
275:
276: We have performed
277: three-body orbit integrations using a fifth-order Runge-Kutta scheme,
278: and the Burlisch-Stoer method as an independent check
279: (e.g. Press et al. 1993).
280: A torque was applied
281: to the outermost planet such that it migrated inwards on a time scale
282: of $t_{mig}= 10^4 $ local orbital periods
283: and a damping force proportional to the radial velocity
284: was applied in
285: the radial direction. This has a fixed
286: constant of proportionality such that for small $e_1$
287: and in the absence of disc torques,
288: the eccentricity damps on a time scale of
289: $t_c=600{\cal N} $ local orbital periods with values of
290: ${\cal N} = 1,10,100$ having been considered.
291: We comment that
292: we have adopted the simplification of neglecting
293: any possible dependence of
294: the damping force and
295: excepting the calculation
296: presented in section \ref{e2_0_0.1},
297: $t_{mig},$ on $e_1.$
298:
299: This procedure was shown to be capable of matching the results of
300: a detailed simulation in paper I, and
301: the value of $t_{mig}$ adopted is consistent with that found
302: from hydrodynamic simulations of protoplanetary discs
303: interacting with protoplanets in the Jovian mass range (e.g. Nelson et al 2000).
304: Consideration of the orbital parameters of GJ876 (see paper I)
305: suggests $ {\cal N} \sim 1.$ However, there is some uncertainty
306: in the functional dependence of the circularization rate of an orbiting
307: protoplanet nonlinearly interacting
308: with a tidally truncated protostellar disc as it would
309: depend on the distance to and form of the disc edge
310: (Goldreich \& Tremaine 1980),
311: and be sensitive to the presence of coorbital material
312: which can lead to eccentricity damping through coorbital Lindblad torques
313: (Artymowicz 1993).
314:
315: For length scale we adopt
316: a fiducial radius, $R$, which for simplicity is taken to be
317: 1 AU. However, scaling invariance allows this to be scaled to some
318: other value if required.
319: Runs were typically started with the outer planet
320: at radius $15.6R$ while the inner planet was started at radius $5R.$
321: A larger initial separation would be inconsistent with our model assumption
322: of there being two planets orbiting within a tidally cleared cavity. Simulations
323: by Bryden et al. (2000) suggest that the time scale to clear such a cavity
324: becomes comparable to the migration time if the ratio of planetary semimajor
325: axies becomes much larger than 3.
326: The outer planet was assumed, because of interaction with the disc,
327: to be in a circular orbit while
328: starting eccentricities $e_2 =0.0, 0.05, 0.1, 0.2, 0.3$ were
329: adopted for the inner planet.
330: For protoplanet masses we have considered
331: all permutations of $0.4,1$ and $4$
332: Jupiter masses
333: assuming the central star mass is 1 M$_{\odot}$.
334:
335: \subsection{Results}
336:
337: \begin{table}
338: \begin{center}
339: \begin{tabular}{|l|l|l|l|l|l|l|l|} \hline \hline
340: $m_1$ & $m_2$ & $\cal{N} $ & & Resonance & & & \\
341: \hline
342: & & &$e_2=0$& 0.05&0.1&0.2 & 0.3 \\
343: \hline
344: 4 & 4& 1 &3:1A &3:1A & 3:1A &4:1A &3:1A \\
345: 4 & 4& 10 &3:1B &4:1B & 4:1B &3:1B &4:1C\\
346: 4 & 4& 100&3:1B &3:1B & 3:1B &3:1B &5:1C\\
347: 4 & 1& 1 & 3:1A &3:1A & 4:1A &3:1A &2:1A\\
348: 4 & 1& 10 & 3:1B &3:1B & 3:1B &2:1A &4:1D\\
349: 4 & 1& 100 &3:1B & 3:1B & 3:1B &3:1C &2:1B\\
350: 4 & 0.4& 1 & 3:1A &3:1A& 3:1A &5:2A &4:1A\\
351: 4 & 0.4& 10 &3:1A & 3:1A& 3:1A &3:1A &2:1A\\
352: 4 & 0.4& 100 & 3:1B& 3:1B & 3:1B &2:1B &2:1B\\
353: \hline
354: 1 & 4& 1 & 2:1A& 3:1A& 3:1A &3:1A &3:1C\\
355: 1 & 4& 10 & 3:1B& 3:1B& 3:1B &5:1C &5:1C\\
356: 1 & 4& 100 & 3:1B& 3:1B& 4:1B &5:1C &3:1B\\
357: 1 & 1& 1 & 2:1A& 2:1A & 3:1A &4:1C &2:1A\\
358: 1 & 1& 10 & 2:1B& 2:1B & 2:1B &4:1B &5:1B\\
359: 1 & 1& 100 & 2:1B& 3:1B & 4:1B &3:1B &5:1B\\
360: 1 & 0.4& 1 &2:1A & 3:1A & 3:1A &2:1A &2:1A\\
361: 1 & 0.4& 10 &2:1B & 2:1B & 3:1B &3:1B &3:1B\\
362: 1 & 0.4& 100 & 2:1B& 3:1B & 3:1B &2:1B &3:1B\\
363: \hline
364: 0.4 & 4& 1 &2:1A & 3:1A & 3:1A &3:1C &3:1C\\
365: 0.4 & 4& 10 & 2:1B& 3:1B & 3:1B &5:1C &5:1C\\
366: 0.4 & 4& 100 & 3:1B& 3:1B & 4:1B &5:1C &5:2B\\
367: 0.4 & 1& 1 &2:1A & 2:1A & 3:1A &2:1A &4:1C\\
368: 0.4 & 1& 10 & 2:1B& 3:1B & 2:1B &2:1B &5:2C\\
369: 0.4 & 1& 100 &2:1B & 5:2B & 3:1B &5:2C &2:1B\\
370: 0.4 & 0.4& 1& 2:1A& 2:1A & 3:1A & 2:1A&2:1A\\
371: 0.4 & 0.4& 10 &2:1B & 2:1B & 2:1B &3:1B &3:2B\\
372: 0.4 & 0.4& 100 &2:1B & 3:1B & 2:1B &2:1B &2:1B\\
373:
374: \hline
375: \end{tabular}
376: \end{center}
377: \caption{This table indicates the outcome
378: and commensurability attained when a pair of masses $m_1, m_2$
379: measured in Jupiter masses
380: becomes resonantly coupled due to disc driven migration
381: of the outer planet. The initial eccentricity $e_2$ and $\cal{N}$
382: are indicated, $e_1$ is always initated as zero. The letters
383: A, B, C, and D in columns 4 -- 8 indicate whether types A, B, C, or D
384: evolution occurred.}
385: \label{tab1}
386: \end{table}
387:
388: \begin{figure}
389: \epsfig{file=fig1.ps,width=\columnwidth}
390: \caption{This figure shows the evolution of the planet
391: semimajor axes and
392: eccentricities for a pair
393: of protoplanets with $m_1=1$, $m_2=1$, initial $e_2=0$, and
394: ${\cal N}=1.$
395: This case underwent self-similar (type A) migration in 2:1
396: resonance after having attained
397: equilibrium eccentricities.
398: The outer planet is denoted by the upper line in the first panel
399: and the lower line in the second panel.
400: The unit of time is $10^5 (R/{\rm 1AU})^{3/2}$ yr.}
401: \label{fig1}
402: \end{figure}
403:
404: \begin{figure}
405: \epsfig{file=fig2.ps, angle =270, width=\columnwidth}
406: \caption
407: {This figure shows the evolution of the protoplanet
408: semimajor axes and
409: eccentricities for a pair
410: of protoplanets, each of 1 Jupiter mass with
411: ${\cal N}=100$ but with the migration rate of the outer planet
412: modified such that it reverses for $e_1>0.2.$
413: At small times, the upper two curves show $\log_{10} (a/R)$ while
414: the lower two curves show $e_1$, $e_2$. The upper curve of the two denoting
415: $\log_{10} (a/R)$ represents the outer planet, whereas at late times
416: the lower curve of the two representing $e_1$, $e_2$ represents the outer
417: planet.
418: The unit of time is $(R/{\rm 1AU})^{3/2}$ yr.}
419: \label{fig2}
420: \end{figure}
421:
422: \begin{figure}
423: \epsfig{file=fig3.ps, width=\columnwidth}
424: \caption{This figure shows the evolution of the protoplanet
425: semimajor axes and eccentricities for a pair of protoplanets,
426: with $m_1=1$, $m_2=4$, initial $e_2=0.3$, and ${\cal N}=1$. This case
427: provides an example of type C migration in 3:1 resonance. Here the
428: eccentricities of both planets grew continuously as the
429: system evolved, even with significant damping of $e_1$ by the disc.
430: Note that our simple model breaks down for $e_1 > 0.2$.
431: The outer planet is denoted by the upper line in the first panel and
432: the lower line in the second panel.
433: The unit of time is $10^5 (R/{\rm 1AU})^{3/2}$ yr.}
434:
435: \label{fig3}
436: \end{figure}
437:
438: \begin{figure}
439: \epsfig{file=fig4.ps, width=\columnwidth}
440: \caption
441: {This figure shows the evolution of the protoplanet
442: semimajor axes and
443: eccentricities for a pair
444: of protoplanets, with $m_1=4$, $m_2=1$, initial $e_2=0.3$, and
445: ${\cal N}=10$.
446: This run provides an example of type D migration in 4:1 resonance.
447: Here the eccentricity of the inner planet grows without limit during the
448: evolution, eventually reaching $e_2 \simeq 1$, whereas the eccentricity
449: of the outer planet remains relatively small. As $e_2$ approaches 1, a
450: dynamical instability occurs, causing the planets to be scattered.
451: The simplicity of our disc model does not allow this phase of evolution to
452: be modeled accurately. The outer planet is denoted by the upper line in the
453: first panel, and by the lower line in the second panel.
454: The unit of time is $10^5 (R/{\rm 1AU})^{3/2}$ yr.}
455: \label{fig4}
456: \end{figure}
457:
458: The results of our survey are shown in table~1.
459: In general for the shortest circularization rates with ${\cal N}=1$
460: and initial values of $e_2 \le 0.1$,
461: self-similar (type A) migration occurred.
462: For larger values of ${\cal N}$, type B migration was usually the preferred
463: outcome.
464:
465: For larger initial values of $e_2$ and ${\cal N}=1$, either type A or type C
466: migration occurred. For larger values of ${\cal N}$, types A, B, C, or D
467: were all found to occur, with a strong preference for types B and C.
468: Only one example of type D migration arose in our simulations.
469:
470: \subsection{Initial $e_2=0$}
471: For an inner planet on an initially circular orbit,
472: a 2:1 commensurability normally resulted,
473: except when one of the masses was $4$ Jupiter masses in which case
474: a 3:1 commensurability could occur.
475: Plots of the evolution of the semimajor axes and eccentricities
476: of a Jupiter mass pair with ${\cal N}=1$ which undergo self-similar
477: migration are illustrated in figure \ref{fig1}.
478: In this case steady equilibrium eccentricities
479: are obtained that are
480: consistent with equations
481: (\ref{ejcons}) and (\ref{ec1}).
482: We found that for the adopted migration rate, the transition
483: between a 2:1 and a 3:1 commensurability occured
484: for $m_1$ in the range 2--3 Jupiter masses when $m_2$ was
485: a Jupiter mass, with the transition mass being smaller
486: for larger migration rates.
487:
488: \subsection{ Initial $0 < e_2 \le 0.1$} \label{e2_0_0.1}
489: When the inner planet was started with an initial
490: eccentricity such that $0 < e_2 \le 0.1$,
491: trapping in a 3:1 commensurability was more common and could
492: occur even for the lowest mass pairs. Also a few cases
493: of trapping in a 4:1 commensurability and one 5:2 commensurability
494: were found. It was observed that the mode of migration
495: in these cases usually corresponded to type A for ${\cal N}=1$, and type B
496: for ${\cal N}=10$ or 100.
497:
498: In order to investigate the potential outcome of a type B
499: migration we considered a Jupiter mass pair with ${\cal N}=100$
500: in which case the eccentricity of the outer planet grows to exceed
501: $0.2.$ At this stage the assumed migration rate becomes inapplicable.
502: In fact migration
503: may reverse on account of penetration
504: of the disc by the outer planet. In this context we note that a simulation
505: of massive protoplanets by Papaloizou, Nelson \& Masset (2001)
506: indicated such a potential reversal for $e_1 > 0.2$
507: and linear torque calculations by Paploizou \& Larwood (2000)
508: suggested migration reversal for an embedded protoplanet with eccentricity
509: exceeding a few times the disc aspect ratio. As an experiment
510: we modified the disc induced migration rate of $m_1$ by a factor
511: $(1- (e_1/0.2)^2).$
512: The resulting evolution is shown in figure~\ref{fig2}
513: This adjustment can stall the migration rate
514: as $e_1$ approaches $0.2$ and lengthen the lifetime
515: of the system before a potential
516: orbital instability occurs due to the development of large
517: eccentricities in both planets. In fact in this case the eccentricity
518: of the inner planet reached $\sim 0.8$ after $\sim 2\times 10^6
519: (R/{\rm 1AU})^{3/2}$~y.
520: The system was found to be subsequently stable for at least a similar
521: time after a rapid disc removal.
522:
523: Thus we emphasise that the projected lifetimes of the resonantly migrating
524: systems discussed here are dependent on the nature of the disc
525: interaction, and at present are very uncertain. However, they could
526: approach the protoplanetary disc lifetime if the mode of evolution discussed
527: in the preceding paragraph was to occur.
528:
529: \subsection{Initial $0.2 \le e_2 \le 0.3$}
530: For initial values of $e_2$ in the range $0.2 \le e_2 \le 0.3$, it was observed
531: that type C evolution became an important consideration, with only a single
532: case of type D evolution being observed.
533: For type C evolution, as in type B, the eccentricities of both planets
534: were observed
535: to grow without reaching equilibrium, until $e_1 > 0.2$ at which point our
536: simple model breaks down. The long term evolution in this situation
537: may resemble that described in figure~\ref{fig2}. We illustrate the
538: behaviour of type C evolution in figure~\ref{fig3}.
539:
540: The single case of type D evolution obtained is illustrated in
541: figure~\ref{fig4}, which should be compared with figure~\ref{fig3}
542: since the only difference between these runs is the initial value of $e_2$
543: ($e_2=0.2$ in figure~\ref{fig3} and $e_2=0.3$ in figure~\ref{fig4}).
544: In this mode of evolution, the outer planet maintains
545: a modest eccentricity while the eccentricity of the inner planet
546: grows without limit. Although this mode of evolution appears to be
547: rather rare based on the results of our calculations, it is interesting to
548: note that it provides a method of generating very high eccentricities,
549: such as that observed in the system HD80606 where the planetary
550: eccentricity is $e \; \gs \; 0.9$ (Naef et al. 2001).
551:
552:
553: Even for initially high values of $e_2$, we still find a tendency for
554: lower mass pairs of planets to enter lower order resonances such as 2:1,
555: 3:1 and
556: even 3:2, whilst the higher mass pairs tend to occupy the higher order
557: commensurabilities, even resulting in capture into 5:1 on a significant number
558: of occassions.
559:
560: \subsection{The Case of GJ876}
561: We have run a number of calculations to examine the putative early evolution
562: of the observed system GJ876, which exists in a 2:1 commensurability.
563: The minimum masses of the two planets, in units where the central star
564: s of solar mass, are $m_1=6$ \MJup and $m_2=1.8$ \MJup. These parameters
565: suggest that trapping in higher order resonance may have been
566: expected rather than in 2:1. For initial conditions in which
567: $m_1=6$, $m_2=1.8$, initial $e_2=0$, capture into 3:1 resonance is favoured.
568: For initial $e_2=0.1$, 4:1 resonance results.
569: However, in a scenario in which the planetary masses were smaller
570: during the initial resonant capture, capture into 2:1 can occur.
571: Simulations were performed with $m_1=1$, $m_2=1.8$, initial $e_2=0$ and 0.1,
572: which all favour capture into 2:1. Assuming a mass for the outer
573: planet of $m_1=3$ produces capture in 3:1.
574:
575: Two possible conclusions to draw from this are that: \\
576: (1). resonant capture occurred in this system when the planets were
577: of smaller mass, and that the mass of the outer planet increased due to
578: accretion of gas from the disc.\\
579: (2). The planets formed sufficiently close together initially that
580: capture into resonances of higher order than 2:1 was not possible.
581:
582:
583: \section{Summary and Discussion} \label{blah}
584: In this paper we have
585: considered two protoplanets gravitationally interacting with each
586: other and a protoplanetary disc. The two planets orbit
587: interior to a tidally maintained disc cavity
588: while the disc interaction induces
589: inward migration.
590: We find that as well as 2:1 commensurabilities being formed,
591: 3:1 commensurabilities are obtained in addition to
592: 4:1, 5:1, and 5:2 commensurabilities which occur less frequently and usually
593: for larger initial eccentricities of the inner planet.
594: Possession of a small eccentricity $\sim 0.1$ before resonant locking takes
595: place would favour 3:1 commensurabilities if migration
596: over large separations has occurred.
597: Initially near circular orbits would favour 2:1 commensurabilities
598: except when one of the planets is significantly more massive than one Jupiter
599: mass.
600:
601: Up to now seven multiple planet systems have been detected
602: with two showing a 2:1 commensurability (GJ876, HD82943 -- see the website: \\
603: obswww.unige.ch/$\sim$udry/planet/new\_planet.html).
604: It is too early at present
605: to establish
606: the frequency of occurence of commensurabilities in multiple systems
607: but they may be fairly common.
608: If orbital migration has occured over a large scale
609: in extrasolar planetary systems the future detection
610: of additional commensurabilities of the type discussed here
611: is to be expected.
612: We also comment that in some cases a high eccentricity for the
613: inner planet orbit may be produced through resonant migration.
614: If, depending on the details
615: of disc dispersal and interaction,
616: a scattering followed by ejection of the outer planet
617: occurs, a single planet remaining on an eccentric orbit may result.
618: This will be a topic for future investigation.
619:
620:
621:
622:
623:
624: %\begin{acknowledgements}
625:
626: \centerline{\bf {Acknowledgements}}
627:
628: \noindent This work was supported by
629: PPARC grant number PPARC GR/L 39094.
630: %\end{acknowledgements}
631:
632: \begin{thebibliography}{}
633:
634: \bibitem[1993]{Arty}
635: {Artymowicz, P., 1993, \apj, 419, 155}
636:
637: \bibitem[1996]{Beust}
638: {Beust, H, Morbidelli, A., 1996, Icarus, 120, 358}
639:
640: \bibitem[2000]{Bryden2}
641: {Bryden, G., R{'o}zyczka, M., Lin, D. N. C., Bodenheimer, P, 2000,
642: \apj 540, 1091}
643:
644: \bibitem[1980]{GT80}
645: {Goldreich, P., Tremaine, S., 1980, \apj, 241, 425}
646:
647:
648: \bibitem[2000]{Kley3}
649: {Kley, W., 2000, \mnr 313, L47}
650:
651: %\bibitem[1998]{Marcy1}
652: %Marcy, G.W., Butler, R.P., 1998, \anrev 36
653:
654: %\bibitem[2000]{Marcy3}
655: %Marcy, G.W., Butler, R.P., 2000, \pasp, 112, 137
656:
657: \bibitem[2001]{Marcy4}
658: {Marcy, G.W., Butler, R.P., Fischer, D., Vogt, S.,
659: Lissauer, J., Rivera, E., 2001, \apj 556, 296}
660:
661: \bibitem[1999]{Murray}
662: {Murray, C.D., Dermott, S.F., 1999, Solar System Dynamics,
663: Cambridge University press.}
664:
665: \bibitem[2001]{Naef}
666: {Naef,D., Latham,D.W., Mayor,M., Mazeh,T., Beuzit,J.L., Drukier,G.A.,
667: Perrier-Bellet,C., Queloz,D., Sivan,J.P., Torres,G., Udry,S.,
668: Zucker,S., 2001, \ana 375, L27}
669:
670:
671:
672: \bibitem[2000]{Nelson}
673: {Nelson, R.P., Papaloizou, J.C.B., Masset, F.S., Kley, W., 2000,
674: \mnr 318, 18}
675:
676: \bibitem[2000]{Papaloizou5}
677: {Papaloizou, J.C.B., Larwood, J.D., 2000,
678: \mnr 315, 823}
679:
680:
681: \bibitem[2001]{Papaloizou4}
682: {Papaloizou, J.C.B., Nelson, R.P., Masset, F.S. 2001,
683: \ana 366, 263}
684:
685: \bibitem[1993]{Press}
686: {Press, W.H., Teukolsky, S.A., Vetterling, W.T., Flannery, B.P.,
687: 1993, Numerical Recipes in FORTRAN: The Art of Scientific
688: Computing, Cambridge University Press}
689:
690: \bibitem[2001]{Snel}
691: { Snellgrove, M., Papaloizou, J.C.B., Nelson, R.P., 2001,
692: \ana 374, 1092}
693:
694: \end{thebibliography}
695:
696: \end{document}
697: