Is there a semi-molten layer at the base of the lunar mantle?

Abstract

Interpretation of the data obtained throughout the period of more than 40 years by Lunar Laser Ranging yields an interesting observation: the tidal quality factor of the Moon, which determines the magnitude of ongoing energy dissipation, follows a different frequency dependence than is measured for rocks in laboratory conditions. When the self-gravity of the lunar body is taken into account, the detected frequency dependence can be interpreted as a signal coming from strong dissipation at the lunar base, indicating a deep-seated layer with low viscosity and possibly containing partial melt. Such a layer would be consistent with the non-detection of deep farside moonquakes by nearside seismic stations and is often associated with the ilmenite-bearing cumulates, that are thought to have descended onto the core-mantle boundary during the lunar magma ocean solidification. Alternative models of a melt-free lunar mantle have also been proposed. These models fit the tidal quality factor Q at the monthly frequency but were unable to explain its observed frequency dependence. In this study, we propose a melt-free model, in which the frequency dependence of lunar Q emerges due to elastically accommodated grain-boundary sliding (GBS) in the lunar mantle. We discuss the implications of such a model and compare it with the traditional approach, which assumes a highly dissipative basal layer. For both alternatives, we perform a Bayesian inversion of the measured tidal parameters (tidal quality factor Q at the monthly and the annual frequency, degree-2 Love numbers h2, k2, and degree-3 Love number k3) and predict either the conditions at the base of the lunar mantle or the relaxation time of elastically accommodated GBS. Since the two alternative models prove to be indistinguishable from each other by tidal measurements, we conclude with an outlook on future observations.