Volatile-rich evolution of molten super-Earth L 98-59 d

Abstract

Small low-density exoplanets are sculpted by strong stellar irradiation, but their primordial compositions and subsequent evolution are still unknown. Two often-considered scenarios hold that they formed with rocky interiors and H$_2$-He atmospheres ('gas-dwarfs'), or alternatively with bulk compositions dominated by H$_2$O phases ('water-worlds'). Here, we constrain the possible range of evolutionary histories linking the birth conditions of low-density super-Earth L 98-59 d to recent observations using a coupled atmosphere-interior evolutionary model. We find that the observations can be explained by in-situ photochemical production of SO$_2$ in an H$_2$ background, indicative of a chemically-reducing mantle and substantial (1.8 mass pct.) early sulfur and hydrogen content, inconsistent with both the gas-dwarf and water-world scenarios. L 98-59 d's interior comprises a permanent magma ocean, allowing long-term retention of volatiles within its mantle over billions of years, consistent with California-Kepler Survey trends. Our analysis reveals an evolutionary pathway in which planets host volatile-rich atmospheres sustained by long-term magma ocean degassing, shaped by secular cooling, atmospheric erosion and photochemistry. Internal and environmental processes contribute to the observed diversity of super-Earth and sub-Neptune exoplanets.

Figures (4)

• Figure 1: Modelled planetary bulk-density over time (panel a). The initial hydrogen inventory of each bulk-density evolution track is shown by the line colour. Blue/green shaded regions are reference densities for a planet of this mass Zeng2019. Black dashed/dotted lines demarcate edges of the radius valley (at this planet mass) for low mass stars ho_shallower_2024. The region $\pm1\sigma$ compatible with the estimated bulk-density of this planet is indicated by the blue errorbar rajpaul_doppler_2024demangeon_warm_2021. Panel b: evolution of surface temperature (coloured) and stellar XUV energy flux (black). For visual clarity, this figure only shows cases with S/H=8.
• Figure 2: Projection of planetary bulk-density $\rho_p$ against several variables. Scatter points represent end points of simulations from our grid, projecting the calculated bulk density against other variables: H inventory from formation (panel a), mantle oxygen fugacity (panel b), S/H ratio from formation (panel c), total planet mass (panel d), atmospheric MMW (panel e), and mantle melt fraction (panel f). Point sizes represent the age of the planet at simulation end-points; largest scatter point size corresponds to the present day. Vertical blue line and shaded region ($\pm1\sigma$) highlights the observationally estimated $\rho_p$. Horizontal purple line show estimates on MMW (with $\pm1\sigma$ shaded) and H2S from free chemistry retrievals. Blue points are consistent with the observed density and MMW, orange points are only consistent with the observed density.
• Figure 3: Volatile loss and atmosphere contraction over time, through two stages of evolution. Bar heights in panel (a) highlight the total loss of volatiles between planet birth and observation, as percentages relative to total planet mass. Lighter and darker bar opacities indicate partitioning between the atmosphere and interior, respectively. Atmosphere elemental mass ratios relative to H are annotated. Panel (b) visualises the evolving atmospheric temperature profile, with an initial stage of rapid contraction due to cooling, followed by a later-stage of slower contraction due to mass loss. Dotted markers indicate convective regions. Profile line colours correspond to time, relative to model initialisation (colourbar). The colour bar is mapped to the x-axis of the inset, which plots radius $R_p$ as a function of time.
• Figure 4: Atmospheric composition and temperature profiles. Solid lines plot the volume mixing ratios for a selection of gases calculated with vulcan's SNCHO photochemical kinetic network. Dashed lines plot mixing ratios calculated without photoreactions. Scatter points show the median-estimates for photospheric chemical abundances (blue, lime) and temperatures (red) retrieved by banerjee_atmospheric_2024 and gressier_hints_2024. Scatter point error bars represent $\pm1\sigma$ ranges on the JWST retrieval posteriors. We assume a modest eddy diffusion coefficient $K_{zz}$ of 1e5cms and use the radiative-convective temperature solution obtained by agni (thick red line).