Creating the Radius Gap without Mass Loss
The observed exoplanet population features a gap in the radius distribution that separates the smaller super-Earths ($\lesssim$1.7 Earth radii) from the larger sub-Neptunes ($\sim$1.7--4 Earth radii). While mass loss theories can explain many of the observed features of this radius valley, it is difficult to reconcile them with potentially rising population of terrestrials beyond orbital periods of $\sim$30 days. We investigate the ability of gas accretion during the gas-poor phase of disk evolution to reproduce both the location of the observed radius gap and the existence of long-period terrestrial planets. Updating the analytic scaling relations of gas accretion rate accounting for the shrinking of the bound radius by hydrodynamic effects and deriving a more realistic disk temperature profile, we find that the late-stage gas accretion alone is able to carve out the observed radius gap, with slopes $R_{\rm gap} \propto P^{-0.096}$ and $R_{\rm gap} \propto M_\star^{0.15}$ for top-heavy; and $R_{\rm gap} \propto P^{-0.089}$ and $R_{\rm gap} \propto M_\star^{0.22}$ for bottom-heavy core mass distributions, in good agreement with observations. The general morphology of the primordial radius gap is stable against a range of disk gas density and disk accretion rate with the latter affecting mostly the population of large planets ($\gtrsim$3--4$R_\oplus$). The peaks and valleys in the radius distribution were likely set in place primordially while post-formation mass loss further tunes the exoplanetary population. We provide potential observational tests that may be possible with TESS, PLATO and Roman Space Telescope.