The Sub-Neptune to Rocky Planet Transition Mediated by Atmospheric Escape and Magma Ocean Evolution

Abstract

A great mystery in exoplanet science is the nature of planets between the sizes of Earth and Neptune: what are their compositions and how do they evolve? Though absent in our solar system, they are the most common type of planet in the galaxy. What can we learn about our solar system by studying these exoplanets? Three strongly related observations provide crucial clues: the bimodal, period-dependent exoplanet radius distribution separating super-Earths and sub-Neptunes at 1.7 - 2 Earth radii, known as the radius valley; evidence for sustained hydrodynamic escape of planetary atmospheres, both inside and outside of the solar system; and magma oceans that mediate this escape. Here I describe a two-pronged approach that combines predictive models and observations to connect the radius valley, atmospheric escape, and atmosphere-interior exchange to work toward a general model of small planet evolution. First, I describe the discovery of the radius valley sub-Neptune TOI-1695 b, with a mass-radius profile consistent with a water-rich composition. Along with eight other planets, TOI-1695 b is a “keystone planet,” for which competing formation/evolution mechanisms predict diverging compositions. Thermally-driven escape scenarios predict keystone planets to be airless, rocky planets that have lost their primordial atmospheres. Conversely, gas-depleted formation proposes that rocky planets form after the bulk of the H/He-rich disk dissipates, and predicts keystone planets to be enveloped. My statistical analysis reveals that thermally-driven mass loss is not dominating the sculpting of the radius valley for planets around low-mass stars, unlike for those around Sun-like stars. However, several keystone planets are also inconsistent with gas-depleted formation, pointing to more diverse, stochastic evolution pathways. Second, I describe a model I developed from first principles to capture previously overlooked physical processes that are necessary for understanding small planet evolution: metal-enrichment/planetary oxidation resulting from escape-driven mass fractionation, and volatile exchange with magma oceans. My model, IsoFATE, models mass fractionation in exoplanet atmospheres through an atomic diffusion framework first developed to understand enhanced D/H ratios and noble gas abundances in the solar system. Improving on previous numerical models, IsoFATE smoothly captures the super-Earth to sub-Neptune transition and robustly predicts strong He and D fractionation throughout the radius valley. Third, I coupled IsoFATE to a magma ocean/equilibrium chemistry model, Atmodeller, to include volatile exchange with molten interiors. The coupled model predicts a novel exoplanet demographic feature, the oxidation gradient, marking increasing planetary oxidation with decreasing planet radius across the radius valley. Population-level studies with IsoFATE make testable predictions of largely unexplored sub-Neptune classes with helium- and O$_2$-dominated atmospheres. Simulated planets with thin atmospheres below the radius valley also demonstrate the importance of magma oceans for atmospheric retention. Fourth, I designed and led two surveys to test the prediction of one such candidate “helium world:” LHS~1140b, a super-Earth orbiting a nearby, inactive mid-M dwarf with an equilibrium temperature of 230 K. I detected escaping metastable helium in the upper atmosphere of LHS~1140b with the high-resolution WINERED spectrograph. The best fit Parker wind/radiative transfer model suggests an H$_2$-depleted, helium-dominated outflow, with a mass-loss rate below the critical level to drag heavier species, consistent with the IsoFATE prediction. This marks the first detection of an atmosphere on an Earth-like exoplanet and places constraints on the cosmic shoreline. The discovery also highlights the importance of atmospheric fractionation for the chemical evolution of exoplanets, opening a new window to observe ongoing fractionation, a process which likely contributed to enhanced D/H on Earth, Mars, and Venus. Finally, I present preliminary results from an ongoing survey to measure the X-ray flux of LHS~1140 over time to investigate whether the variable helium escape from LHS~1140b is driven by variable stellar XUV flux. At the time of this writing, two 64 ks observations are planned for January and July, 2026 with the XMM-Newton space telescope. In a preliminary analysis of the first dataset, I find that the X-ray flux is consistent with a 2018 X-ray measurement, suggesting that XUV variability may not fully explain the variable atmospheric escape. I present stellar activity models that fit the current data. Future observations will help elucidate the enigmatic nature LHS~1140b, which may represent the first of an exciting, newly discovered class of planets: helium worlds.