Title: The importance of silicate vapor in determining the structure, radii, and envelope mass fractions of sub-Neptunes

Abstract: Substantial silicate vapor is expected to be in chemical equilibrium at temperature conditions typical of the silicate-atmosphere interface of sub-Neptune planets, which can exceed 5000 K. Previous models of the atmospheric structure and evolution of these exoplanets, which have been used to constrain their atmospheric mass fractions, have neglected this compositional coupling. In this work, we show that silicate vapor in a hydrogen-dominated atmosphere acts as a condensable species, decreasing in abundance with altitude. The resultant mean molecular weight gradient inhibits convection at temperatures above $\sim 4000$ K, inducing a near-surface radiative layer. This radiative layer decreases the planet's total radius compared to a planet with the same base temperature and a convective, pure H/He atmosphere. Therefore, we expect silicate vapor to have major effects on the inferred envelope mass fraction and thermal evolution of sub-Neptune planets. We demonstrate that differences in radii, and hence in inferred atmospheric masses, are largest for planets which have larger masses, equilibrium temperatures, and atmospheric mass fractions. The effects are largest for younger planets, but differences can persist on gigayear time-scales for some sub-Neptunes. For a $10 M_\oplus$ planet with $T_\mathrm{eq}=1000$ K and an age of $\sim 300$ Myr, an observed radius consistent with an atmospheric mass fraction of 10% when accounting for silicate vapor would be misinterpreted as indicating an atmospheric mass fraction of 2% if a H/He-only atmosphere were assumed. The presence of silicate vapor in the atmosphere is also expected to have important implications for the accretion and loss of primordial hydrogen atmospheres.