The Exoplanet Radius Valley from Gas-driven Planet Migration and Breaking of Resonant Chains

Abstract: The size frequency distribution of exoplanet radii between 1 and 4$R_{\oplus}$ is bimodal with peaks at $\sim$1.4 $R_{\oplus}$ and $\sim$2.4 $R_{\oplus}$, and a valley at $\sim$1.8$R_{\oplus}$. This radius valley separates two classes of planets -- usually referred to as "super-Earths" and "mini-Neptunes" -- and its origin remains debated. One model proposes that super-Earths are the outcome of photo-evaporation or core-powered mass-loss stripping the primordial atmospheres of the mini-Neptunes. A contrasting model interprets the radius valley as a dichotomy in the bulk compositions, where super-Earths are rocky planets and mini-Neptunes are water-ice rich worlds. In this work, we test whether the migration model is consistent with the radius valley and how it distinguishes these views. In the migration model, planets migrate towards the disk inner edge forming a chain of planets locked in resonant configurations. After the gas disk dispersal, orbital instabilities "break the chains" and promote late collisions. This model broadly matches the period-ratio and planet-multiplicity distributions of Kepler planets, and accounts for resonant chains such as TRAPPIST-1, Kepler-223, and TOI-178. Here, by combining the outcome of planet formation simulations with compositional mass-radius relationships, and assuming complete loss of primordial H-rich atmospheres in late giant-impacts, we show that the migration model accounts for the exoplanet radius valley and the intra-system uniformity ("peas-in-a-pod") of Kepler planets. Our results suggest that planets with sizes of $\sim$1.4 $R_{\oplus}$ are mostly rocky, whereas those with sizes of $\sim$2.4 $R_{\oplus}$ are mostly water-ice rich worlds. Our results do not support an exclusively rocky composition for the cores of mini-Neptunes.