Lunar evolution: a combined numerical modelling and HPT experimental study

Abstract

Recent studies, some of them using data from the Apollo seismic network from the 1960's and 1970's, others using newer data, have shown that part of the lunar core may still be fluid today. Furthermore, a possible partial melt zone has been detected in the deep mantle, just above the core-mantle boundary. These melt occurrences are hard to explain, since a small body like the Moon, which is approximately 40 times smaller than Earth, generally cools very fast. An internal heat source is required to cause present-day high temperatures in the deep lunar mantle and core. This combined experimental and numerical modelling research shows that an enrichment in radioactive elements in a dense mineral layer may provide this heat source. It is generally agreed upon that the lunar surface was covered with a magma ocean during its early history. When this magma ocean cooled, the crystallisation of the magma resulted in a compositionally layered mantle due to subsequent crystallisation of different minerals with decreasing solidus temperatures. Most minerals are denser than the melt they crystallise from and will therefore sink to the bottom of the magma ocean. An exception is plagioclase: this mineral has a lower density that the melt and has therefore floated to the surface to form the early lunar crust. The final layer to crystallise, at shallow depth below the crust, contains a significant amount of the high-density, iron- and titanium-rich mineral ilmenite (nominally (Fe,Mg)TiO3). Furthermore, it is strongly enriched in radioactive elements due to the general incompatibility of these elements in the crystal structures of the lunar minerals. This high-density cumulate layer on top of lower-density mineral cumulates causes a gravitationally unstable mineral stratification, resulting in an overturn of the lunar mantle. The influence of the high-density, radioactive element enriched, ilmenite-rich layer during and after mantle overturn has been the focus of this study. Experimental determination of the incompatibility of the radioactive elements Th, U and K in the crustal plagioclase, combined with literature values for the incompatibility in other lunar minerals, has been used to estimate the concentrations of these elements in the different mineral cumulates for use as numerical model input parameters. Furthermore, an experimental study, determining the thermal equation of state of the iron end-member of orthopyroxene, orthoferrosilite, has been performed. Equations of state for the lunar minerals, as calculated from semi-empirical thermodynamic models, have been used in our final numerical convection models to improve the description of the relative density contrasts between the minerals as a function of pressure (depth) and temperature in the lunar mantle. Results of these models show that the high concentrations of radioactive elements, transported to the deep mantle with the ilmenite-rich cumulate, can provide the heat source required to cause outer core and deep mantle melting.