Lunar central peak mineralogy and iron content using the Kaguya Multiband Imager: Reassessment of the compositional structure of the lunar crust

Ryder and Wood (1977) suggested that the lunar crust becomes more mafic with depth because the impact melts associated with the large Imbrium and Serenitatis basins are more mafic than the surface composition of the Moon. In this study, we reexamine the hypothesis that the crust becomes more mafic with depth; we analyze the composition of crater central peaks by using recent remote sensing data and combining the best practices of previous studies. We compute the mineralogy for 34 central peaks using (1) nine-band visible and near-infrared data from the Kaguya Multiband Imager, (2) an improved version of Hapke's radiative transfer model validated with spectra of lunar soils with well-known modal mineralogy, and (3) new crustal thickness models from the Gravity Recovery and Interior Laboratory data to examine the variation in composition with depth. We find that there is no increase in mafic mineral abundances with proximity to the crust/mantle boundary or with depth from the current lunar surface and, therefore, that the crust does not become more mafic with depth. We find that anorthosite with very low mafic abundance (“purest anorthosite” or PAN) is a minority constituent in these peaks, and there is no clear evidence of a distinct PAN-rich layer in the middle crust as previously proposed. The composition of most of the central peaks we analyze is more mafic than classically defined anorthosites with an average noritic anorthosite composition similar to that of the lunar surface.