Orbital Migration through Atmospheric Mass Loss

A central finding in the exoplanet field over the past decade is that the distribution of planets’ physical radii is bimodal. The two observed peaks are interpreted as separating planets that have retained significant atmospheres from bare cores whose atmospheres have been stripped through either photoevaporation or core-powered mass loss. Given the significant mass loss invoked by atmospheric escape models, we investigate the observational consequences on the planetary orbits. Hydrodynamic simulations of this atmospheric mass loss show that the morphology of the escaping gas depends on the strength of the incoming stellar wind. Under strong stellar winds that might be expected around young stars, the escaping atmosphere is funneled into a tail behind the planet. In this case, the trailing gas tail pulls on the evaporating planet gravitationally, causing it to lose energy and move toward a shorter orbital period. Full hydrodynamic simulations are computationally expensive, and therefore model only a few dozen planetary orbits. By contrast, atmospheric mass loss is expected to take place on timescales of ~100 million years that are computationally inaccessible. However, because the gas flow quickly becomes supersonic, streamlines follow ballistic trajectories that can be modeled through much cheaper N-body integrations. We present N-body integrations of this process, together with an approximate analytical model that shows good agreement over the relevant parameter space. We will show that over some plausible parameter ranges, planetary orbital periods can migrate by of order 1%, which we argue could help explain the observed pileups of planet pairs wide of mean motion resonances (integer period ratios). This would provide a new, independent observational constraint on atmospheric mass loss models.