Migration of planets into and out of mean motion resonances in protoplanetary discs: analytical theory of second-order resonances

Abstract

Recent observations of Kepler multi-planet systems have revealed a number of systems with planets very close to second-order mean motion resonances (MMRs, with period ratio $1:3$, $3:5$, etc.) We present an analytic study of resonance capture and its stability for planets migrating in gaseous disks. Resonance capture requires slow convergent migration of the planets, with sufficiently large eccentricity damping timescale $T_e$ and small pre-resonance eccentricities. We quantify these requirements and find that they can be satisfied for super-Earths under protoplanetary disk conditions. For planets captured into resonance, an equilibrium state can be reached, in which eccentricity excitation due to resonant planet-planet interaction balances eccentricity damping due to planet-disk interaction. We show that this "captured" equilibrium can be overstable, leading to partial or permanent escape of the planets from the resonance. In general, the stability of the captured state depends on the inner to outer planet mass ratio $q=m_1/m_2$ and the ratio of the eccentricity damping times. The overstability growth time is of order $T_e$, but can be much larger for systems close to the stability threshold. For low-mass planets undergoing type I (non-gap opening) migration, convergent migration requires $q \lesssim 1$, while the stability of the capture requires $q\gtrsim 1$. These results suggest that planet pairs stably captured into second-order MMRs have comparable masses. This is in contrast to first-order MMRs, where a larger parameter space exists for stable resonance capture. We confirm and extend our analytical results with $N$-body simulations, and show that for overstable capture, the escape time from the MMR can be comparable to the time the planets spend migrating between resonances.