We provide new estimates for the abundance of heat-producing elements in the lunar mantle by using SIMS techniques to measure the concentrations of thorium and samarium in lunar pyroclastic glasses. Lunar pyroclastic glasses are utilized in this study because they represent quenched products of near-primary melts from the lunar mantle and as such, they provide compositional information about the mantle itself. Thorium and samarium were measured because: (1) Th is not significantly fractionated from Sm during partial melting of the pyroclastic glass source regions, which are dominated by olivine and pyroxene. Therefore, the Th/Sm ratios that we measure in the pyroclastic glasses reflect the Th/Sm ratio of the pyroclastic glass source regions. (2) Strong correlations between Th, U, and K on the Moon allow us to use measured Th concentrations to estimate the concentrations of U and K in the pyroclastic glasses. (3) Th, Sm, U, and K are radioactive elements and as such, their concentrations can be used to investigate heat production in the lunar mantle.

The results from this study show that the lunar mantle is heterogeneous with respect to heat-producing elements and that there is evidence for mixing of a KREEP component into the source regions of some of the pyroclastic glasses. Because the source regions for many of the glasses are deep (⩾400 km), we propose that a KREEP component was transported to the deep lunar mantle. KREEP enriched sources produce 138% more heat than sources that do not contain KREEP and therefore, could have provided a source of heat for extended periods of nearside basaltic magmatism. Data from this study, in conjunction with models for the fractional crystallization of a lunar magma ocean, are used to show that the average lunar mantle contains 0.15 ppm Th, 0.54 ppm Sm, 0.039 ppm U, and 212 ppm K. This is a greater enrichment in radiogenic elements than some earlier estimates, suggesting a more prolonged impact of radiogenic heat on nearside basaltic volcanism.