Re-examination of the lunar magma ocean cumulate overturn hypothesis: Melting or mixing is required

There is a long-standing hypothesis that the last fraction of the lunar magma ocean crystallized into a layer of dense, Ti-rich cumulate minerals at shallow depths ( ~ 100 km) early in the moon's history. Many questions remain about the stability of these high-Ti cumulates. It has been suggested that the cumulates subsequently sank deep into the moon because of gravitational instability, but high-Ti material is required at shallower depths by 3.5 Ga to create the high-Ti mare basalts and picritic glasses. The high-Ti material may have re-erupted from depth, or some or all of it may have remained at shallow depths throughout lunar history. Data on phase stabilities, bulk compositions, densities, and temperatures of melting and crystallizing in addition to results from numerical modeling suggest that the high-Ti cumulates would sink only under highly specific conditions. Five scenarios for sinking high-Ti cumulate materials are examined, and only two are found plausible. In particular, it is found that simple sinking of solidified high-Ti cumulates is unlikely because the temperature at which the cumulates solidify is low, and viscosity under these conditions is very high. It is, however, possible that high-Ti cumulates mixed with a substantial fraction of olivine would have viscosity low enough to allow them to sink as solids. Further, because clinopyroxene and ilmenite melt in a ratio of 2:1, remelted high-Ti cumulates would be negatively buoyant and sink as liquids, percolating downward through the underlying mantle and beginning to recrystallize ilmenite at 200 km depth, making a hybrid, heterogeneous mantle.