1: \def\etal{{\it et al.\thinspace}}
2: \def\mearth{{\rm\,M_\oplus}}
3: \def\msun{{\rm\,M_\odot}}
4:
5: \topmargin -0.5in
6:
7: %\documentclass[12pt,preprint,psfig]{aastex}
8: %\documentstyle[emulateapj,psfig,apjfonts]{article}
9: \documentclass[apjl]{emulateapj}
10: %\documentclass{emulateapj}
11:
12:
13: %\received{}
14: \accepted{Sept 18, 2008}
15: %\journalid{}{}
16: %\articleid{}{}
17:
18: \lefthead{}
19: \righthead{}
20:
21: \begin{document}
22:
23: \shorttitle{Mean Motion Resonances from Planet-Planet Scattering}
24: \shortauthors{Raymond et al.}
25:
26: \title{Mean motion resonances from planet-planet scattering}
27:
28: \author{Sean N. Raymond\altaffilmark{1},
29: Rory Barnes\altaffilmark{2},
30: Philip J. Armitage\altaffilmark{3},
31: \& Noel Gorelick\altaffilmark{4}}
32:
33: \altaffiltext{1}{NASA Postdoctoral Program Fellow, Center for Astrophysics and
34: Space Astronomy, University of Colorado, Boulder CO 80309; raymond@lasp.colorado.edu}
35: \altaffiltext{2}{Lunar and Planetary Laboratory, University of Arizona,
36: Tucson, AZ; rory@lpl.arizona.edu }
37: \altaffiltext{3}{JILA, University of Colorado, Boulder CO 80309; pja@jilau1.colorado.edu}
38: \altaffiltext{4}{Google, Inc., 1600 Amphitheatre Parkway, Mountain View, CA
39: 94043; gorelick@google.com}
40: %%\email{}
41:
42: \begin{abstract}
43: Planet-planet scattering is the leading mechanism to explain the large
44: eccentricities of the observed exoplanet population. However, scattering has
45: not been considered important to the production of pairs of planets in mean
46: motion resonances (MMRs). We present results from a large number of numerical
47: simulations of dynamical instabilities in 3-planet systems. We show that MMRs
48: arise naturally in about five percent of cases. The most common resonances we
49: populate are the 2:1 and 3:1 MMRs, although a wide variety of MMRs can occur,
50: including high-order MMRs (up to eleventh order). MMRs are generated
51: preferentially in systems with uneven mass distributions: the smallest planet
52: is typically ejected after a series of close encounters, leaving the
53: remaining, more massive planets in resonance. The distribution of resonant
54: planets is consistent with the phase-space density of resonant orbits, meaning
55: that planets are randomly thrown into MMRs rather than being slowly pulled
56: into them. It may be possible to distinguish between MMRs created by
57: scattering vs. convergent migration in a gaseous disk by considering planetary
58: mass ratios: resonant pairs of planets beyond $\sim$ 1 AU with more massive
59: outer planets are likely to have formed by scattering. In addition,
60: scattering may be responsible for pairs of planets in high-order MMRs (3:1 and
61: higher) that are not easily populated by migration. The frequency of MMRs
62: from scattering is comparable to the expected survival rate of MMRs in
63: turbulent disks. Thus, planet-planet scattering is likely to be a major
64: contributor to the population of resonant planets.
65: \end{abstract}
66:
67: \keywords{ planetary systems: formation --- methods: n-body simulations}
68:
69: \section{Introduction}
70:
71: The current sample of exoplanets exhibits several interesting dynamical
72: features (Butler \etal 2006): here we focus on two of these. First, there is
73: a vast range of planetary eccentricities, from zero to $>0.9$, with a median
74: of 0.2 (0.27 for planets past 0.1 AU that have not been affected by tides;
75: Rasio \etal 1996; Jackson \etal 2008). Second, mean motion resonances (MMRs)
76: in multiple planet systems appear to be relatively common (e.g., Marcy
77: \etal 2001). There are 31 currently-known multiple planet systems comprising
78: 44 pairs of adjacent planets, of which ten (23\%) show some evidence of
79: resonances (Table 1). However, the evidence for resonances is tentative for
80: all but a few cases.
81:
82: Dynamical instabilities in systems of two or more planets can explain the wide
83: eccentricity distribution of exoplanets (Rasio \& Ford 1996; Weidenschilling
84: \& Marzari 1996; Lin \& Ida 1997). Instabilities arise on timescales that are
85: related to the planets' initial separation (Marzari \& Weidenschilling 2002;
86: Chatterjee \etal 2008), and lead to close encounters between planets and
87: subsequent ejections or mergers. In the aftermath of close encounters, the
88: surviving planets can statistically reproduce the observed eccentricity
89: distribution of exoplanets (Adams \& Laughlin 2003; Juric \& Tremaine 2008;
90: Chatterjee \etal 2008).
91:
92: MMRs are thought to arise primarily from convergent migration in gaseous
93: protoplanetary disks (Snellgrove \etal 2001; Lee \& Peale 2002). Indeed,
94: models show that capture in the 2:1 and 3:2 MMRs is a particularly common
95: occurrence (Thommes 2005; Pierens \& Nelson 2008; Lee \etal 2008). However,
96: MRI-derived turbulence can act to remove planets from resonance. Adams \etal
97: (2008) estimate that only 1\% of resonant systems should remain for a disk
98: lifetime of 1 Myr.
99:
100: In this paper we attempt to reconcile the planet-planet scattering scenario
101: with the population of resonant exoplanets. We numerically
102: investigate dynamical instabilities in systems of three planets located at
103: $\sim$ 2-10 AU with a variety of mass distributions. We find that MMRs are a
104: common occurrence, arising in 5-10 percent of unstable systems. Our
105: simulations populate a range of MMRs, including the 2:1, 3:2, 3:1, 4:1 and
106: extending up to much higher-order (Table 2). MMRs are populated by scattering
107: at random into stable regions; the density of resonant orbits (i.e., the
108: fraction of phase space that undergoes resonant oscillations) is consistent
109: with the scattered resonant systems. We propose several ways to discriminate
110: between scattering and convergent migration as the source of exoplanet MMRs.
111:
112: \section{Methods}
113:
114: Our simulations started with three planets randomly separated by 4-5 mutual
115: Hill radii ($R_{H,m} = 0.5 (a_1+a_2) ([M_1+M_2]/3M_\star)^{1/3}$, where $a$ is
116: the semimajor axis and $M$ the mass). This spacing was chosen to produce
117: instabilities on timescales of at least the $\sim 10^5$ year timescale of
118: runaway gas accretion\footnote{Indeed, instabilities occurred on timescales
119: from 100 years to 98 Myr with a median of a few $\times 10^5$ years. In
120: addition, about 1/4 of simulations were stable for 100 Myr which shows that we
121: started close to the stability boundary.} (Pollack \etal 1996; Marzari
122: \& Weidenschilling 2002). The outermost planet was placed two Hill radii
123: interior to 10 AU; cases with more massive planets and therefore larger Hill
124: radii therefore had the innermost planet closer to the star than for cases
125: with lower-mass planets (see below). Planets were given zero eccentricity and
126: mutual inclinations of less than 1 degree, and the stellar mass was 1 $\msun$.
127:
128: We considered a range of planetary mass distributions. For our two largest
129: sets (1000 simulations each) we randomly selected planet masses according to
130: the observed distribution of exoplanet masses: $dN/dM \propto M^{-1.1}$
131: (Butler \etal 2006). In the ``mixed1'' set we restricted the planet mass
132: $M_p$ to be between a Saturn mass $M_{Sat}$ and three Jupiter masses
133: $M_{Jup}$. For our ``mixed2'' set, the minimum planet mass was decreased to
134: 10 $\mearth$. We also performed four ``Mequal'' sets (500 simulations each)
135: with equal mass planets for $M_p = 30 \mearth$, $M_{Sat}$, $M_{Jup}$, and $3
136: M_{Jup}$. Finally, the ``Mgrad'' sets (250 simulations each) contained radial
137: gradients in $M_p$. For the JSN set, in order of increasing orbital distance,
138: $M_p$ = $M_{Jup}$, $M_{Sat}$, and $30 \mearth$. For the NSJ set, these masses
139: were reversed, i.e., the $M_{Jup}$ planet was the most distant. The 3JJS and
140: SJ3J sets had, in increasing radial distance, $M_p$ = $3 M_{Jup}$, $M_{Jup}$
141: and $M_{Sat}$, and $M_p$ = $M_{Sat}$, $M_{Jup}$ and $3 M_{Jup}$, respectively.
142:
143: \begin{figure}[t]
144: \centerline{\plotone{f1.eps}}
145: \caption{Cumulative eccentricity distribution of the known exoplanets
146: beyond 0.1 AU (thick grey), compared with our scattering simulations.}
147: \label{fig:edist}
148: \end{figure}
149:
150: Each simulation was integrated with the hybrid version of the {\tt Mercury}
151: integrator (Chambers 1999). All planets were assigned physical densities of
152: 1.3 $g \,cm^{-3}$ and collisions were treated as inelastic mergers. We used a
153: 20 day timestep which tests show introduces an error of less than 1 part in
154: 10$^5$ for perihelion distances larger than 0.5 AU. In almost all cases
155: energy was conserved to better than one part in 10$^4$ for the entire 100 Myr
156: simulation, which Barnes \& Quinn (2004) showed is adequate precision to test
157: stability. However, in some cases, energy was poorly conserved; those cases
158: were rerun with a 5 day timestep. After this step, simulations with poor
159: energy conservation were removed from the analysis.
160:
161: \section{Results}
162:
163: Figure~\ref{fig:edist} shows that four of our sets of simulations match the
164: exoplanet eccentricity distribution -- mixed1, Mequal:Jup, Mequal:Sat, and
165: Mequal:30 $\mearth$ -- with P values from K-S tests greater than 0.01.
166: However, given the increasing number of low-mass exoplanets, we believe that
167: our mixed1 and mixed2 simulations are the most realistic initial conditions.
168: If scattering is the source of exoplanet eccentricities, then
169: soon-to-be-discovered systems with lower-mass planets should indeed tend to
170: have lower eccentricities (Ford \& Rasio 2008).
171:
172:
173: \begin{figure}[t]
174: \centerline{\plotone{f2.eps}}
175: \caption{Evolution of a system that produced a pair of planets in the 3:1 MMR.
176: {\bf Top:} The three planets' semimajor axes $a$, perihelia $q$ and aphelia
177: $Q$. The inner (green), middle (black), and outer (red) planets are 43, 105,
178: and 16 $\mearth$, respectively. {\bf Bottom:} Evolution of the 3:1 resonant
179: argument $\theta_3 = 3 \lambda_2 - \lambda_1 - (\varpi_1+\varpi_2)$. Resonant
180: libration starts immediately after ejection of the outer planet.}
181: \label{fig:evol31}
182: \end{figure}
183:
184:
185: We found MMRs by examining resonant arguments for simulations which produced
186: pairs of planets with period ratios close to commensurate values. A pair of
187: planets is in resonance if any resonant argument $\theta_i$ librates rather
188: than circulates. For MMR p+q : p, arguments are of the form
189: \begin{equation}
190: \theta_{i} = (p+q) \lambda_1 - p \lambda_2 -q \varpi_{1,2} \\
191: \end{equation}
192: \noindent where $\lambda$ are mean longitudes, $\varpi$ are longitudes
193: of pericenter, and subscripts 1 and 2 refer to the inner and outer planet,
194: respectively (e.g., Murray \& Dermott 1999).
195:
196: Figure~\ref{fig:evol31} shows the evolution of a typical simulation that
197: created a resonant system. The instability started 55.8 Myr into the
198: simulation, causing a series of close encounters. Within a few hundred
199: thousand years the outer planet was ejected and the inner two planets swapped
200: places. The two remaining planets are on stable orbits in the 3:1 MMR, and
201: all three resonant arguments $\theta_{1,2,3}$ librate with amplitudes between
202: 120$^\circ$ and 160$^\circ$.
203:
204: A variety of MMRs is populated by scattering (see Table 2). Most common are
205: the 2:1 and 3:1, but higher-order MMRs exist up to eleventh order (13:2). The
206: resonant libration amplitudes tend to be large, with a median of 110$^\circ$
207: and several cases with amplitudes of $\sim 170^\circ$. This contrasts with
208: MMRs from migration which are created in a dissipative environment and should
209: be much smaller. MMRs occur preferentially in cases with mixed mass
210: distributions, especially those with a positive mass gradient such as
211: Mgrad:NSJ. MMRs are relatively rare for equal mass planets, and they tend to
212: arise more often after collisions rather than ejections, which contrasts with
213: the mixed and Mgrad cases. This may explain why MMRs have not been found in
214: previous studies (except for isolated cases in Adams \& Laughlin 2003 and
215: Chatterjee \etal 2008).
216: %\footnote{We found two cases of the 1:1 MMR but those simulations
217: %conserved energy poorly and were excluded.}
218:
219: MMRs appear to be populated at random: any stable region of parameter space
220: can be accessed by scattering. To test this hypothesis, we calculated the
221: phase space density of resonant orbits within 10\% of the 2:1 and 3:1 MMRs for
222: planetary mass ratio of 1/3, 1, and 3, with $M_{inner}+M_{outer}=400 \mearth$.
223: For each MMR we ran $\sim$ 22,000 3-body (star + two planets) simulations for
224: 1 Myr. The semimajor axis of the inner planet was fixed at 5 AU (2:1 MMR) or
225: 4 AU (3:1 MMR). We sampled four parameters: the orbital period ratio, the
226: inner and outer planets' eccentricities, and the relative apsidal orientation.
227: Inclinations (of $<1^\circ$) and mean longitudes were sampled at random.
228: Resonant orbits were found by libration of resonant arguments.
229: %: ($M_{inner},M_{outer}$) = (100 $\mearth$, 300 $\mearth$), (200 $\mearth$,
230: %200 $\mearth$), and (300 $\mearth$,100 $\mearth$).
231:
232: The density of 2:1 resonant orbits is higher for a more massive outer planet
233: (Figure~\ref{fig:res21}).\footnote{For a more detailed study of the 2:1 MMR,
234: see Marzari \etal (2006) and Michtchenko \etal (2008a, 2008b).} Almost all of
235: the 2:1 and 3:1 resonant orbits from scattering are found in areas of high
236: resonant density, and near-resonant ``false alarms'' lie in areas of low
237: density. Thus, scattering does indeed appear to populate MMRs at random.
238: This explains why the 2:1 MMRs from our mixed1 and mixed2 simulations, with no
239: initial mass gradients, have a median $M_{inner}/M_{outer}$ of 0.5. In
240: addition, the Mgrad:NSJ ($M_{inner}/M_{outer}\approx 1/3$) systems formed a
241: large number of 2:1 MMRs while the Mgrad:JSN ($M_{inner}/M_{outer} \approx 3$)
242: cases formed far fewer.
243:
244: For the parameter space we sampled, the integrated 3:1 resonant density is
245: $\sim 40\%$ less than the 2:1 density. However, the available parameter space
246: is not evenly populated by scattering. For example, scattering causes more
247: massive planets to be closer to the star (Chatterjee \etal 2008). Indeed, the
248: integrated 2:1 resonant density for $M_{inner}/M_{outer} = 1/3$ is 66\% higher
249: than the 3:1 density. This explains the increased number of 2:1 vs. 3:1 MMRs
250: in our sample -- 74 cases of the 2:1 MMR and 47 of the 3:1 (61\% more in 2:1).
251:
252: \begin{deluxetable}{c|c|c|c|c}[t]
253: \scriptsize
254: \tablewidth{0pc}
255: \tablecaption{Candidate Resonant Planetary systems\tablenotemark{1}}
256: \renewcommand{\arraystretch}{.6}
257: \tablehead{
258: \\
259: \colhead{System} &
260: \colhead{$a_1,a_2$} &
261: \colhead{$e_1,e_2$} &
262: \colhead{$M_1,M_2$} &
263: \colhead{MMR} \\
264: \colhead{(pair)} &
265: \colhead{(AU)} &
266: \colhead{ } &
267: \colhead{($M_{Jup}$)} }
268: %\colhead{ }}
269: \startdata
270: \\
271: GJ 876 c-b & 0.13, 0.2078 & 0.27, 0.025 & 0.56, 1.935 & 2:1\\
272:
273: HD 73526 b-c & 0.66, 1.05 & 0.19, 0.14 & 2.9, 2.5 & 2:1\\
274:
275: HD 82943 c-b & 0.746, 1.19 & 0.359, 0.219 & 2.01, 1.75 & 2:1\\
276:
277: HD 128311 b-c & 1.099, 1.76 & 0.25, 0.17 & 2.18, 3.21 & 2:1\\
278:
279: $\mu$ Arae d-b & 0.921, 1.497 & 0.067, 0.128 & 0.522, 1.676 & 2:1\\
280:
281: GJ 317 b-c & 0.95, 2.35 & 0.19, 0.42 & 1.2, 0.83 & 4:1\tablenotemark{2} \\
282:
283: HD 108874 b-c & 1.051, 2.68 & 0.07, 0.25 & 1.36, 1.018 & 4:1\\
284:
285: HD 17156 b-c & 0.159, 0.481 & 0.6717, 0.136 & 3.111, 0.063 & 5:1\\
286:
287: HD 202206 b-c & 0.83, 2.55 & 0.435, 0.267 & 17.4, 2.44 & 5:1\\
288:
289: HD 208487 b-c & 0.49, 1.8 & 0.32, 0.19 & 0.45, 0.46 & 7:1\\
290:
291: \enddata
292: \tablenotetext{1}{See http://www.lpl.arizona.edu/$\sim$rory/research/xsp/dynamics/.}
293: \tablenotetext{2}{Johnson \etal (2007) did not determine apsidal angles, but
294: Barnes \& Greenberg (2008) used a stability analysis to predict that the
295: system must be in the 4:1 MMR.}
296: \end{deluxetable}
297:
298:
299: The density of resonant orbits can explain other features of the population of
300: resonant planets. The resonant planets tend to have smaller eccentricities
301: than the non-resonant planets (Fig.~\ref{fig:edist}). Indeed, resonant
302: systems underwent an average of about five times fewer close encounters
303: between planets before the system stabilized ($\sim$ 30-50 vs. 100-$>$200
304: encounters), compared with the median outcome. The time between the first
305: instability and stabilization for resonant cases was $\sim$ 50,000 years, a
306: factor of about five shorter than for the non-resonant systems.
307:
308: \begin{figure}[t]
309: \centerline{\plotone{f3.eps}}
310: \caption{Resonant density for the 2:1 MMR. For these simulations, the inner
311: planet's semimajor axis was fixed at 5 AU, and its eccentricity was varied
312: between 0.05 and 0.25. The color corresponds to the fraction of orbits in
313: resonance for a given value of the outer planet's semimajor axis and
314: eccentricity (see color bar), averaged over 8 simulations with different
315: apsidal alignments.}
316: \label{fig:res21}
317: \end{figure}
318:
319: Low-order MMRs preferentially arise in systems containing a planet with a
320: large Safronov number $S$. $S$ is the ratio of the escape speed from a
321: planet's surface to the escape speed from the system, $S =
322: (M_p/M_\star)^{1/2}\, (a_p/R_p)^{1/2}$, where $M_\star$ is the stellar mass,
323: $a_p$ and $R_p$ are the planet's orbital distance and radius, respectively
324: (Safronov 1969). Planets with larger $S$ give stronger velocity kicks and
325: thereby reduce the number of encounters needed to eject a planet. Indeed, the
326: 2:1 and 3:1 MMRs correlate with systems with at least one planetary $S$ value
327: above 4. In the mixed2 set, which is the only set with a significant range in
328: $S$, the median $S_{max}$ for the 2:1 and 3:1 MMRs is 4.5, as compared with
329: 3.5 for all unstable mixed2 simulations. A K-S test shows that the two
330: samples are indeed different at the 99.9\% confidence level. This constrains
331: where the 2:1 or 3:1 MMRs can arise as a function of $M_p$, $M_\star$, $a_p$,
332: and $R_p$: only systems with relatively high-mass planets ($M_p \gtrsim
333: M_{Jup}$) can generate these MMRs close-in. In contrast, high-order (4:1 and
334: higher) MMRs from the mixed2 set match the sample of unstable cases and so are
335: not constrained.
336:
337: The MMRs we found are numerically robust. The median fractional integration
338: error dE/E for the 170 resonant systems was $2.6 \times 10^{-8}$, far below
339: the $\sim10^{-4}$ limit for determining stability (Barnes \& Quinn 2004), and
340: smaller than the median dE/E of $1.2 \times 10^{-7}$ for all unstable systems.
341: MMRs tend to arise in cases with short encounter times and relatively low
342: final eccentricities. Those situations yield smaller dE/E than for the more
343: common stronger encounters that lead to very eccentric planets. Thus, our
344: cutoff of dE/E $< 10^{-4}$ allows us to accurately sample both eccentricities
345: and MMRs.
346:
347: \section{Discussion and Conclusions}
348:
349: Planet-planet scattering creates MMRs. The typical path to a resonant system
350: involves several close encounters between one smaller and two larger planets.
351: After a relatively short time of instability, the smaller planet is destroyed,
352: usually via ejection (78\% of all cases) or collision (22\%), leaving behind a
353: pair of resonant planets. Relatively weak instabilities are probably very
354: common in planetary systems; one may even have occurred in our own Solar
355: System (Thommes \etal 1999). Nonetheless, only a fraction of unstable systems
356: produce resonant planets, typically 5-10\%. These systems have large
357: libration amplitudes and occupy a range of low-and high-order MMRs (Table 2).
358: Most of these resonances are indefinitely stable; we integrated the 170
359: resonant systems for an additional 1 Gyr and only 9 (5\%) left the resonance,
360: 4 cases leading to an additional system instability.
361:
362: \begin{deluxetable}{c|c|c|p{2cm}}[t]
363: %\vskip -0.5in
364: \scriptsize
365: \tablewidth{0pt}
366: \tablecaption{Resonances from scattering simulations}
367: \renewcommand{\arraystretch}{.6}
368: \tablehead{
369: \\
370: \colhead{Set} &
371: \colhead{Nsims ---} &
372: \colhead{N(\%) in} &
373: \colhead{MMRs}\\
374: \colhead{ } &
375: \colhead{unstable(\%)} &
376: \colhead{MMRs} &
377: \colhead{(\%)}}
378: \startdata
379: %\\
380: Mixed1 & 965--569 (59\%) & 27 (4.7\%) & 2:1 (1.6\%), 3:1 (1.6\%), 4:1
381: (0.7\%), 5:1, 6:1, 7:2, 9:2\\
382: %\\ \hline \\
383: Mixed2 & 982--744 (76\%) & 52 (7\%) & 3:2 (0.8\%), 2:1 (2.4\%), 3:1 (1.1\%),
384: 5:3, 4:1, 5:2, 5:1, 7:3, 6:1, 7:2, 8:3, 9:4, 10:3, 12:5, 11:2, 14:5\\
385: %\\ \hline \\
386: Mequal:3$M_J$ & 368--241 (65\%) & 1 (0.4\%) & 7:1\\ %\tablenotemark{1}
387: %\\ \hline \\
388: Mequal:$M_J$ & 452--232 (51\%) & 4 (1.7\%) & 2:1, 3:1, 4:1 (0.9\%)\\
389: %\\ \hline \\
390: Mequal:$M_{Sat}$ & 390--362 (93\%) & 14 (3.9\%) & 2:1 (1.9\%), 3:1 (0.6\%),
391: 5:2, 5:1 (0.6\%), 6:1, 7:1\\
392: %\\ \hline \\
393: Mequal:30$\mearth$ & 367--365 (99\%) & 10 (2.7\%) & 2:1 (1.9\%), 3:1, 11:6\\
394: %\\ \hline \\
395: Mgrad:JSN & 250--206 (82\%) & 13 (6.3\%) & 2:1 (1\%), 3:1 (1.9\%), 4:1,
396: 5:2, 5:1, 7:3, 8:3, 13:2\\
397: %\\ \hline \\
398: Mgrad:NSJ & 245--221 (90\%) & 30 (14.6\%) & 2:1 (9.2\%), 3:1 (2.3\%), 5:2,
399: 7:3, 7:2, 11:5\\
400: %\\ \hline \\
401: Mgrad:3JJS & 250--150 (60\%) & 4 (2.7\%) & 3:1, 4:1 (1.3\%), 11:3\\
402: %\\ \hline \\
403: Mgrad:SJ3J & 245--219 (89\%) & 16 (6.5\%) & 3:1 (5.7\%), 4:1
404: %\\ \\ \hline \\
405: \enddata
406: %\tablenotetext{1}{Note the closer spacing of planets in this case (3.5-4
407: %mutual Hill radii as opposed to 4-5).}
408: \end{deluxetable}
409:
410: It may be possible to tell apart resonant exoplanets created via scattering
411: from those created via convergent migration. In fact, only one resonant
412: system appears to be inconsistent with a scattering origin due to its very
413: low-amplitude libration (GJ 876; Marcy \etal 2001). If two planets are
414: trapped in the 2:1 or 3:2 MMR and the inner planet is the more massive, then
415: migration can be stopped or even reversed (Masset \& Snellgrove 2001; Crida \&
416: Morbidelli 2007). However, if the outer planet is the more massive then
417: inward migration continues. In contrast, scattering produces planets in a
418: variety of MMRs (including the 3:2 and 2:1) with a wide range in mass ratios
419: and a preference for the outer planet to be more massive. Thus, scattering is
420: likely to be responsible for systems past $\sim$ 1 AU with 2:1 or 3:2 resonant
421: planets and a more massive outer planet. The HD 128311 and $\mu$ Arae systems
422: are good candidates for creation via scattering (Table 1; see also S\'andor \&
423: Kley 2006).
424:
425: Several extra-solar systems show tentative evidence for high-order MMRs --
426: 4:1, 5:1 and even 7:1 (Table 1; Johnson \etal 2007; Correia \etal 2005;
427: Gregory 2007; Short \etal 2008). No study to date has shown that migration
428: could capture planets in MMRs of higher order than 2:1, although we encourage
429: expanded studies of this process.\footnote{Highly-damped bodies can undergo
430: resonant shepherding by the 6:1 or even 8:1 MMRs (Raymond
431: \etal 2006; Mandell \etal 2007). However, as bodies grow the damping
432: decreases, and shepherded planets do not survive in resonance.} Scattering
433: produces a wide range of high-order resonances (Table 2). Thus, if the
434: current candidate high-order MMR systems are confirmed (Table 1), then
435: scattering is likely to be the responsible mechanism.
436:
437: Turbulence in gaseous disks may destroy MMRs, leaving perhaps only $\sim$ 1\%
438: of planet pairs in resonance (Adams \etal 2008). This effect is stronger for
439: higher-order MMRs. However, the timescale for MMR destruction is sensitive to
440: the strength of MRI turbulence which is very uncertain. In particular, if the
441: MRI is fully or partially suppressed by the low ionization fraction in the
442: inner protoplanetary disk (Gammie 1996) then the survival prospects for
443: resonant planets would be improved. Nonetheless, if only a few percent of
444: systems remain in resonance, then scattering and migration may provide a
445: comparable number of MMRs.
446:
447: Our simulations do not account for any dissipation. However, instabilities
448: may occur while some gas remains in the disk (Moeckel \etal 2008). In that
449: case, MMRs could still result from scattering (Lee \etal 2008) and perhaps
450: have smaller libration amplitudes.
451:
452: In conclusion, we have identified a new mechanism for the creation of
453: exoplanet systems in MMRs. Unfortunately the current data do not allow a
454: conclusive determination of a resonance, let alone precise descriptions of the
455: resonant argument oscillation. Nonetheless, our scattering model has
456: important distinctions from the convergent migration model: high-order MMRs,
457: large-amplitude resonant libration, and low-order MMRs with
458: $M_{inner}/M_{outer} < 1$. As the orbital properties of exoplanets are better
459: determined, it should be possible to distinguish between these scenarios.
460:
461: \vskip .2in
462: We thank Google for access to their machines. We are grateful to Greg
463: Laughlin, Dimitri Veras, and an anonymous referee for helpful input.
464: S.N.R. was supported by the NASA Postdoctoral Program administered by Oak
465: Ridge Associated Universities through a contract with NASA. R.B. acknowledges
466: support from NASA's PG\&G grant NNG05GH65G and NASA Terrestrial Planet Finder
467: Foundation Science grant 811073.02.07.01.15.
468:
469: %\newpage
470: \begin{thebibliography}{999}
471:
472: \bibitem[Adams
473: \& Laughlin(2003)]{2003Icar..163..290A} Adams, F.~C., \& Laughlin, G.\ 2003, Icarus, 163, 290
474:
475:
476: \bibitem[Adams et al.(2008)]{2008ApJ...683.1117A} Adams, F.~C., Laughlin,
477: G., \& Bloch, A.~M.\ 2008, \apj, 683, 1117
478:
479:
480: \bibitem[Barnes
481: \& Greenberg(2008)]{2008IAUS..249..469B} Barnes, R., \& Greenberg, R.\ 2008, IAU Symposium, 249, 469
482:
483:
484: %\bibitem[Barnes
485: %\& Greenberg(2007)]{2007ApJ...665L..67B} Barnes, R., \& Greenberg, R.\ 2007,
486: %\apjl, 665, L67
487:
488:
489: \bibitem[Barnes
490: \& Quinn(2004)]{2004ApJ...611..494B} Barnes, R., \& Quinn, T.\ 2004, \apj, 611, 494
491:
492:
493: \bibitem[Butler et al.(2006)]{2006ApJ...646..505B} Butler, R.~P., et al.\
494: 2006, \apj, 646, 505
495:
496:
497: \bibitem[Chambers(1999)]{1999MNRAS.304..793C} Chambers, J.~E.\ 1999,
498: \mnras, 304, 793
499:
500:
501: \bibitem[Chatterjee et al.(2007)]{2007astro.ph..3166C} Chatterjee, S.,
502: Ford, E.~B., Matsumura, S.,
503: \& Rasio, F.~A.\ 2007, ArXiv Astrophysics e-prints, arXiv:astro-ph/0703166
504:
505:
506: \bibitem[Correia et
507: al.(2005)]{2005A&A...440..751C} Correia, A.~C.~M., Udry, S., Mayor, M., Laskar, J., Naef, D., Pepe, F., Queloz, D., \& Santos, N.~C.\ 2005, \aap, 440, 751
508:
509:
510: \bibitem[Crida
511: \& Morbidelli(2007)]{2007MNRAS.377.1324C} Crida, A., \& Morbidelli, A.\ 2007, \mnras, 377, 1324
512:
513:
514: \bibitem[Ford
515: \& Rasio(2007)]{2007astro.ph..3163F} Ford, E.~B., \& Rasio, F.~A.\ 2007, ArXiv Astrophysics e-prints, arXiv:astro-ph/0703163
516:
517:
518: \bibitem[Gammie(1996)]{1996ApJ...457..355G} Gammie, C.~F.\ 1996, \apj, 457,
519: 355
520:
521:
522: \bibitem[Gregory(2007)]{2007MNRAS.374.1321G} Gregory, P.~C.\ 2007, \mnras,
523: 374, 1321
524:
525:
526: \bibitem[Jackson et al.(2008)]{2008ApJ...678.1396J} Jackson, B., Greenberg,
527: R., \& Barnes, R.\ 2008, \apj, 678, 1396
528:
529:
530: \bibitem[Johnson et al.(2007)]{2007ApJ...670..833J} Johnson, J.~A., Butler,
531: R.~P., Marcy, G.~W., Fischer, D.~A., Vogt, S.~S., Wright, J.~T.,
532: \& Peek, K.~M.~G.\ 2007, \apj, 670, 833
533:
534:
535: \bibitem[Juric
536: \& Tremaine(2007)]{2007astro.ph..3160J} Juric, M., \& Tremaine, S.\ 2007, ArXiv Astrophysics e-prints, arXiv:astro-ph/0703160
537:
538:
539: %\bibitem[Laughlin et al.(2004)]{2004ApJ...608..489L} Laughlin, G.,
540: %Steinacker, A., \& Adams, F.~C.\ 2004, \apj, 608, 489
541:
542:
543: \bibitem[Lee et al.(2008)]{2008arXiv0801.1926L} Lee, A.~T., Thommes, E.~W.,
544: \& Rasio, F.~A.\ 2008, ArXiv e-prints, 801, arXiv:0801.1926
545:
546:
547: \bibitem[Lin
548: \& Ida(1997)]{1997ApJ...477..781L} Lin, D.~N.~C., \& Ida, S.\ 1997, \apj, 477, 781
549:
550:
551: \bibitem[Mandell et al.(2007)]{2007ApJ...660..823M} Mandell, A.~M.,
552: Raymond, S.~N., \& Sigurdsson, S.\ 2007, \apj, 660, 823
553:
554:
555: \bibitem[Marcy et al.(2001)]{2001ApJ...556..296M} Marcy, G.~W., Butler,
556: R.~P., Fischer, D., Vogt, S.~S., Lissauer, J.~J.,
557: \& Rivera, E.~J.\ 2001, \apj, 556, 296
558:
559:
560: \bibitem[Marzari
561: \& Weidenschilling(2002)]{2002Icar..156..570M} Marzari, F., \& Weidenschilling, S.~J.\ 2002, Icarus, 156, 570
562:
563:
564: \bibitem[Marzari et
565: al.(2006)]{2006A&A...453..341M} Marzari, F., Scholl, H., \& Tricarico, P.\ 2006, \aap, 453, 341
566:
567:
568: \bibitem[Masset
569: \& Snellgrove(2001)]{2001MNRAS.320L..55M} Masset, F., \& Snellgrove, M.\ 2001, \mnras, 320, L55
570:
571:
572: \bibitem[Michtchenko et al.(2008)]{2008MNRAS.387..747M} Michtchenko, T.~A.,
573: Beaug{\'e}, C., \& Ferraz-Mello, S.\ 2008a, \mnras, 387, 747
574:
575:
576: \bibitem[]{} Michtchenko, T.~A.,
577: Beaug{\'e}, C., \& Ferraz-Mello, S.\ 2008b, \mnras, in press.
578:
579:
580: \bibitem[Moeckel et al.(2008)]{2008arXiv0807.4186M} Moeckel, N., Raymond,
581: S.~N., \& Armitage, P.~J.\ 2008, ArXiv e-prints, 807, arXiv:0807.4186
582:
583:
584: %\bibitem[Papaloizou
585: %\& Nelson(2003)]{2003MNRAS.339..983P} Papaloizou, J.~C.~B., \& Nelson, R.~P.\ 2003, \mnras, 339, 983
586:
587:
588: \bibitem[Pierens
589: \& Nelson(2008)]{2008A&A...482..333P} Pierens, A., \& Nelson, R.~P.\ 2008, \aap, 482, 333
590:
591:
592: \bibitem[Pollack et al.(1996)]{1996Icar..124...62P} Pollack, J.~B.,
593: Hubickyj, O., Bodenheimer, P., Lissauer, J.~J., Podolak, M.,
594: \& Greenzweig, Y.\ 1996, Icarus, 124, 62
595:
596:
597: \bibitem[Rasio et al.(1996)]{1996ApJ...470.1187R} Rasio, F.~A., Tout,
598: C.~A., Lubow, S.~H., \& Livio, M.\ 1996, \apj, 470, 1187
599:
600:
601: \bibitem[Rasio
602: \& Ford(1996)]{1996Sci...274..954R} Rasio, F.~A., \& Ford, E.~B.\ 1996, Science, 274, 954
603:
604:
605: \bibitem[Raymond et al.(2006)]{2006Sci...313.1413R} Raymond, S.~N.,
606: Mandell, A.~M., \& Sigurdsson, S.\ 2006, Science, 313, 1413
607:
608:
609: \bibitem[Safronov(1969)]{1969QB981.S26......} Safronov, V.~S.\ 1969, 1969.,
610:
611: \bibitem[S{\'a}ndor \& Kley(2006)]{2006A&A...451L..31S} S{\'a}ndor, Z., \&
612: Kley, W.\ 2006, \aap, 451, L31
613:
614: %\bibitem[Shen
615: %\& Turner(2008)]{2008arXiv0806.0032S} Shen, Y., \& Turner, E.~L.\ 2008, ArXiv
616: %e-prints, 806, arXiv:0806.0032
617:
618:
619: \bibitem[Short et al.(2008)]{2008arXiv0803.2935S} Short, D., Welsh, W.~F.,
620: Orosz, J.~A., \& Windmiller, G.\ 2008, ArXiv e-prints, 803, arXiv:0803.2935
621:
622:
623: \bibitem[Snellgrove et
624: al.(2001)]{2001A&A...374.1092S} Snellgrove, M.~D., Papaloizou, J.~C.~B., \& Nelson, R.~P.\ 2001, \aap, 374, 1092
625:
626:
627: \bibitem[Thommes(2005)]{2005ApJ...626.1033T} Thommes, E.~W.\ 2005, \apj,
628: 626, 1033
629:
630:
631: \bibitem[Thommes et al.(1999)]{1999Natur.402..635T} Thommes, E.~W., Duncan,
632: M.~J., \& Levison, H.~F.\ 1999, \nat, 402, 635
633:
634:
635: \bibitem[Weidenschilling
636: \& Marzari(1996)]{1996Natur.384..619W} Weidenschilling, S.~J., \& Marzari, F.\ 1996, \nat, 384, 619
637:
638: \end{thebibliography}
639:
640:
641:
642:
643: \end{document}
644:
645: