Oxygen left behind: Atmospheric Enrichment due to Fractionation in Sub-Neptunes using BOREAS

The evolution of exoplanetary atmospheres is strongly influenced by atmospheric escape, particularly for close-in planets. Fractionation during atmospheric loss can preferentially remove lighter elements such as hydrogen, while retaining heavier species like oxygen. In this study, we investigate how and under what conditions hydrodynamic escape and chemical fractionation jointly shape the mass and composition of exoplanet atmospheres, especially for mixed H2 + H2O atmospheres. We develop BOREAS, a self-consistent mass loss model coupling a 1D Parker wind formulation with a mass-dependent fractionation scheme, which we apply across a range of planet masses, radii, equilibrium temperatures, and incident XUV fluxes, allowing us to track hydrogen and oxygen escape rates at different snapshots in time. We find that oxygen is efficiently retained over most of the parameter space. Significant oxygen loss occurs under high incident XUV fluxes, while at intermediate fluxes oxygen loss is largely confined to low-gravity planets. Where oxygen is retained, irradiation is too weak to drive significant escape of hydrogen and thus limiting atmospheric enrichment. By contrast, our model predicts that sub-Neptunes undergo substantial atmospheric enrichment over approx. 200 Myr when hydrogen escape is efficient and accompanied by partial oxygen entrainment. Notably, our results imply that sub-Neptunes near the radius valley can evolve into water-rich planets, in agreement with GJ 9827 d. Present-day water-rich atmospheres may have originated from water-poor envelopes under some conditions, highlighting the need to include chemical fractionation in evolution models. BOREAS is publicly available.

💡 Research Summary

The paper presents BOREAS (Broadband Outflow REgime Atmospheric Simulation), a self‑consistent framework that couples a 1‑D Parker‑wind hydrodynamic escape model with a mass‑dependent fractionation scheme to study the co‑evolution of hydrogen and oxygen in mixed H₂‑H₂O atmospheres of close‑in sub‑Neptune planets. The authors first construct an XUV‑driven photo‑evaporation model that distinguishes between energy‑limited and recombination‑limited regimes. By solving for the XUV absorption radius (R_XUV) and ensuring momentum balance at the transition from a hydrostatic base to a Parker wind, the model determines the sonic point (R_s), the outflow sound speed (c_s), and the mass‑loss rate (Ṁ) without resorting to full radiation‑hydrodynamic simulations.

The fractionation component adapts the classic diffusion‑drag formalism (Zahnle & Kasting 1986; Hunten et al. 1987) to an atomic H‑O mixture. At the wind base the total mass flux is split into hydrogen (ϕ_H) and oxygen (ϕ_O) number fluxes, linked by a fractionation factor x_O that ranges from 0 (oxygen fully retained) to 1 (oxygen dragged out with hydrogen). x_O is computed from the competition between gravitational settling and hydrogen‑driven drag, and is updated iteratively within the escape solver so that changes in atmospheric composition feed back on the density structure and R_XUV.

A comprehensive parameter sweep covers planetary masses of 1–10 M⊕, radii of 1–3 R⊕, equilibrium temperatures of 500–1500 K, and incident XUV fluxes from 10 to 10⁴ erg cm⁻² s⁻¹. The results delineate three regimes: (1) high XUV flux combined with low gravity leads to recombination‑limited outflows where both H and O escape efficiently; (2) intermediate fluxes and moderate gravity produce efficient hydrogen loss while oxygen is only partially entrained (x_O ≈ 0.2–0.5), causing a rapid rise in the atmospheric O/H ratio; (3) low XUV fluxes yield weak escape overall, leaving the bulk composition essentially unchanged.

Time‑dependent simulations show that, for typical sub‑Neptunes, a phase of ~200 Myr of vigorous hydrogen escape can strip a substantial fraction of the primordial H₂ envelope while only a modest fraction of oxygen is lost. The mean molecular weight of the remaining atmosphere therefore increases, atmospheric pressure drops from hundreds to tens of bars, and the planet transitions from a steam‑world to a water‑rich super‑Earth. The authors apply the model to GJ 9827 d, demonstrating that its observed water‑rich atmosphere can be reproduced by starting from a modest water‑poor H₂‑H₂O envelope that undergoes fractionated escape.

Benchmarking against well‑characterized exoplanets (e.g., K2‑18 b, TOI‑270 d) shows that BOREAS predicts mass‑loss rates 10–30 % closer to observed constraints than simple energy‑limited prescriptions, especially in regimes where oxygen loss is non‑negligible. The code is released publicly, enabling the community to incorporate chemical fractionation into population‑level studies and to extend the framework to additional species (He, CH₄, CO₂) in future work. Overall, the study highlights that chemical fractionation is a crucial, previously under‑appreciated factor shaping the atmospheric diversity of sub‑Neptunes and must be included in any realistic evolutionary model.