Exoplanet bulk silicate composition as a function of host stellar elemental abundances, and its effects on long-term planetary evolution

One of the main goals of Exoplanetary Sciences is to reconcile our theoretical knowledge of terrestrial exoplanet systems with observations. To reach this goal, the interaction between it and the planetary interior needs to be studied, since the atmosphere is the only observable part of a terrestrial exoplanet. This atmosphere-interior interaction depends on properties of the interior, many of which are directly (density, viscosity) or indirectly (thermal evolution, layering) affected by the bulk composition of the planet. In order to better understand the variability in atmosphere-interior interaction between exoplanets, as well as the properties of the resulting atmosphere, we here constrain the range of terrestrial exoplanet bulk compositions.

To constrain this range, we approximate exoplanet composition by applying devolatilization to the bulk composition of the host star. We approximate exoplanet compositions by adjusting host-star compositions from a stellar catalogue according to the condensation temperature. We consider planetary differentiation by distributing elements between the core and mantle according to their tendency to stabilize oxide, thus obtaining a proxy for bulk silicate compositions. We include partitioning of light elements into the core. Lastly, we explore the effects of these compositions on the tendency to promote stable mantle stratification in the aftermath of magma-ocean freezing, using a thermodynamic model of crystallization, and on thermal evolution using a 1D parametrized convection model.

We find that mantle Mg/Si is an important control on mantle properties, since increased Mg/Si-ratios tend to decrease mantle viscosity by stabilizing soft minerals, such as olivine and ferropericlase, at the expense of pyroxene and stishovite (and corresponding high-pressure polymorphs). The Mg/Si of planets is shifted towards higher values by the slightly higher volatility of Si, and by the partitioning of Si into the core. We find that the Earth's mantle is below average in terms of bulk-silicate Mg/Si for planets in the galactic neighborhood. This result indicates that most terrestrial planets have a mantle viscosity lower than that of Earth. Earth is average in terms of bulk Fe/Si, and above average in terms of bulk Fe/Mg. We find that planets with relatively low Mg/Si and high Fe/Mg in their silicate envelopes cool slower because of high mantle viscosities, and because of their tendency to sustain double-layered convection in a stratified mantle.

Finally, we identify a number of end-member bulk planet compositions, which we recommend for use in modelling of terrestrial exoplanet interiors. These end-member compositions cover most of the variability in bulk terrestrial exoplanet compositions based on available stellar composition data. We also present mineralogical mantle profiles for these end-member compositions. In the future, we intend to explore the effects of these bulk-silicate planet compositions on surface tectonic style, and the related feedback on planetary cooling and volatile cycling.