Dynamical Origin of Extrasolar Planet Eccentricity Distribution

We explore the possibility that the observed eccentricity distribution ofextrasolar planets arose through planet-planet interactions, after the initialstage of planet formation was complete. Our results are based on ~3250numerical integrations of ensembles of randomly constructed planetary systems,each lasting 100 Myr. We find that for a remarkably wide range of initialconditions the eccentricity distributions of dynamically active planetarysystems relax towards a common final equilibrium distribution, well describedby the fitting formula dn ~ e exp[-1/2 (e/0.3)^2] de. This distribution agreeswell with the observed eccentricity distribution for e > 0.2, but predicts toofew planets at lower eccentricities, even when we exclude planets subject totidal circularization. These findings suggest that a period of large-scaledynamical instability has occurred in a significant fraction of newly formedplanetary systems, lasting 1--2 orders of magnitude longer than the ~1 Myrinterval in which gas-giant planets are assembled. This mechanism predicts no(or weak) correlations between semimajor axis, eccentricity, inclination, andmass in dynamically relaxed planetary systems. An additional observationalconsequence of dynamical relaxation is a significant population of planets(>10%) that are highly inclined (>25deg) with respect to the initial symmetryplane of the protoplanetary disk; this population may be detectable intransiting planets through the Rossiter-McLaughlin effect.Comment: Accepted to ApJ, conclusions updated to reflect the current observational constraint