A petrologic study on the effect of mantle overturn: Implications for evolution of the lunar interior

Abstract Lunar mantle overturn caused by gravitational instability of the Fe-Ti rich KREEP layer (formed as the last 5% of a crystallizing magma ocean, and emplaced between the overlying anorthitic crust and the underlying lunar mantle) is a process that would introduce Fe-Ti enriched bodies deep inside the lunar interior. These chemical heterogeneities in the lunar mantle may be the source of the very Fe-Ti enriched near-primary Apollo basalts. Also, the Fe-Ti and KREEP enriched layer deep inside the Moon may be responsible for the 5–30% partial melt seismically detected close to the core-mantle boundary (CMB). This is assuming that the partial melt is neutrally buoyant at P-T conditions of the CMB. Here, we experimentally investigate the phase equilibria of the overturned Fe-Ti rich layer mixed with the mantle, at P-T conditions deep inside the lunar interior, focusing on the partial melt compositions formed. Our aim is to test (a) whether potential partial melt compositions formed near the CMB are neutrally buoyant with respect to the surrounding mantle, hence, stable; (b) if the partial melts formed within the lunar interior are positively buoyant and ascend, whether they can reproduce chemical characteristics of Apollo basalts. The densities calculated for the Fe-Ti rich partial melts from this study, using the physical parameters from previous studies, range from lower to higher values compared to that of the lunar mantle. This provides a basis for future investigations to experimentally constrain better the densities of these partial melts. Depending on the buoyancy of the partial melts, the following two scenarios are likely to happen. Firstly, if the partial melts are neutrally buoyant at the CMB, 5–30% partial melt would constrain the CMB temperatures between 1330(±1) and 1470(±19) °C. This can be used by future studies to derive the selenotherm better. Secondly, if the partial melts are positively buoyant, they should ascend and react with the mantle along their path. Upon reaching shallow depths below the crust, they may likely assimilate any Fe-Ti rich layer that was left over from the gravitational overturn, as well as undergo olivine fractionation upon pooling in a shallow magma chamber. We modeled assimilation-fractional crystallization of the partial melts using the Gibb’s-free minimization algorithm alphaMELTS. Our results show that reactive ascent of Fe-Ti rich partial melts through the lunar mantle and subsequent olivine fractionation in a shallow magma chamber is a promising way to evolve the melt compositions to converge with the lunar basalts better. Shallow level assimilation of Fe-Ti rich lithology post reactive-ascent through the mantle is also feasible, but only for low degrees of assimilation.