Title:From CO$_2$- to H$_2$O-dominated atmospheres and back -- How mixed outgassing changes the volatile distribution in magma oceans around M dwarf stars

Abstract:We investigate the impact of CO$_2$ on TRAPPIST-1 e, f and g during the magma ocean stage. These potentially habitable rocky planets are currently the most accessible for astronomical observations. A constraint on the volatile budget during the magma ocean stage is a link to planet formation and also needed to judge their habitability. We perform simulations with 1-100 terrestrial oceans (TO) of H$_2$O with and without CO$_2$ and for albedos 0 and 0.75. The CO$_2$ mass is scaled with initial H$_2$O by a constant factor between 0.1 and 1.
The magma ocean state of rocky planets begins with a CO$_2$-dominated atmosphere but can evolve into a H$_2$O dominated state, depending on initial conditions. For less than 10 TO initial H$_2$O, the atmosphere tends to desiccate and the evolution may end with a CO$_2$ dominated atmosphere. Otherwise, the final state is a thick (>1000 bar) H$_2$O-CO$_2$ atmosphere. Complete atmosphere desiccation with less than 10 TO initial H$_2$O can be significantly delayed for TRAPPIST-1e and f, when H$_2$O has to diffuse through a CO$_2$ atmosphere to reach the upper atmosphere, where XUV photolysis occurs. As a consequence of CO$_2$ diffusion-limited water loss, the time of mantle solidification for TRAPPIST-1 e, f, and g can be significantly extended compared to a pure H$_2$O evolution by up to 40 Myrs for albedo 0.75 and by up to 200 Mrys for albedo 0. The addition of CO$_2$ further results in a higher water content in the melt during the magma ocean stage. Our compositional model adjusted for the measured metallicity of TRAPPIST-1 yields for the dry inner planets (b, c, d) an iron fraction of 27 wt-%. For TRAPPIST-1 e, this iron fraction would be compatible with a (partly) desiccated evolution scenario and a CO$_2$ atmosphere with surface pressures of a few 100 bar. A comparative study between TRAPPIST-1 e and the inner planets may yield the most insights about formation and evolution scenarios.