The Case for a Completely Solid Martian Mantle—No Basal Magma Layer Extant

Abstract

Recent results from the Mars InSight mission suggested the existence of a molten silicate layer atop the core‐mantle boundary. Geophysical modeling of this layer suggested that it must be denser than the overlying mantle but less dense than the core, and have either a viscosity similar to that of the overlying solid mantle or a strong chemical gradient. However, studies of silicate liquid viscosities at the appropriate pressure range for this layer (17–21 GPa) yielded values <1 Pa s, over 20 orders of magnitude smaller than those suggested by the geophysical models. Additionally, temperatures across the layer must be >2500 K to be completely molten with nearly isothermal temperature profiles, indicating a thoroughly mixed liquid, inconsistent with a chemical gradient. We constructed a set of models to test this liquid layer hypothesis using silicate liquid equations of state (EoS), 1D heat flow calculations, and convective scaling analysis to determine the properties and crystallization times of a possible liquid layer. We calculated the density and P‐wave velocities for a variety of liquid compositions, including bulk silicates and mineral endmembers. We also calculated crystallization times for various temperatures and viscosities, using the output parameters from the EoS calculations. Based on the EoS calculations, the layer must contain at least 20–25 wt.% FeO. Using the appropriate compositions and liquid viscosities, crystallization timescales fall between 3 and 4 Myr. From stability arguments and crystallization timescales, such a layer would be relatively short‐lived; thus, the putative seismically inferred molten silicate layer would have to be formed geologically recently. Viable formation mechanisms have yet to be found.