ANATEXIS OF LUNAR CUMULATE MANTLE IN TIME AND SPACE: CLUES FROM TRACE-ELEMENT, STRONTIUM, AND NEODYMIUM ISOTOPIC CHEMISTRY OF PARENTAL APOLLO 12 BASALTS

Abstract

In an effort to elucidate the processes of lunar mantle melting, and the magma evolution of mare basalts within Oceanus Procellarum on the western lunar near-side, we have analyzed seven fine-grained to vitrophyric Apollo 12 basalts for trace-elements; five of these also have been analyzed for Nd and Sr isotopic compositions. These samples represent all three main groups identified among the Apollo 12 mare basalts and have been proposed as parental melts to their respective groups, i.e., olivine-, pigeonite-, and ilmenite-basalts. The sources for these low-Ti mare basalts are postulated to have formed from crystallization of a global magma ocean. Li-Be systematics, combined with REE data, indicate that the specific sources for the Apollo 12 low-Ti mare basalts were generated after 82-94% crystallization of this lunar magma ocean. In fact, it seems that all mare basalts analyzed from the Apollo collections were generated from cumulates precipitated in the last 20% of the magma ocean. Chemical compositions of fine-grained pigeonite and olivine basalts are consistent with 7-9% nonmodal (in proportions not defined by experimental petrology and phase equilibria) melting of a source consisting of 48% olivine, 30% calcic clinopyroxene, and 22% pigeonite (as per Neal et al., 1994b). Sm-Nd and Rb-Sr abundance data also suggest that the pigeonite- and olivine-basalt source contained from 0.3 to 0.5% trapped residual liquid from the magma ocean. The compositions of the two fine-grained ilmenite basalts are consistent with 5-7% partial melting of a source with subequal proportions of olivine (45.5%) and pigeonite (42.5%) and lesser amounts of clinopyroxene (11.5%) and entrained plagioclase (0.5%). Furthermore, the ilmenite source was nearly devoid of trapped liquid (<0.15%). A few of these samples do indicate minor post-extrusive fractionation, but most of the samples are considered to be unfractionated, primitive magmas that are parental to the other mare basalts.Isotopic systematics of the Apollo 12, fine-grained, parental basalts are consistent with their derivation from two distinct mantle source regions. Both of these sources were LREE depleted for extended periods of time: up to 600 million years for the ilmenite-basalt source and up to 900 million years for the pigeonite- and olivine-basalt source. Due in part to the relatively small proportion of low Sm/Nd, trapped, residual magma-ocean liquid in the source (<0.15%), the Nd isotopic compositions of the ilmenite basalts are among the most radiogenic ever analyzed from the Moon (εNd = +10.5 to +11.2, at 3.2 Ga). The mantle source for the olivine and pigeonite basalts contained a higher proportion of trapped, residual, magma-ocean liquid (0.3 to 0.5%), thus yielding less radiogenic Nd isotopic signatures (εNd = +4.3 to +4.7, at 3.2 Ga).By integrating this information on parental, low-Ti, Apollo 12 basalts with mare basalt and picritic glass data from other landing sites, as well as telescopic and remote-sensing data, we propose a model for melting of the lunar interior. The upper 400-500 km of the lunar mantle is a consequence of incipient melting of the Moon and formation of a global magma ocean. This magma ocean became progressively enriched in incompatible elements as it precipitated the cumulate upper mantle. This incompatible-element enriched liquid was also trapped in varying proportions in the differentiated cumulates. The earliest, extensive mare magmas (high-Ti mare basalts) were generated at shallow depths in the mantle from cumulate source-regions that had trapped relatively large proportions of this incompatible-element enriched, residual, magma ocean liquid. These source regions also contained the late-crystallizing phase ilmenite and, thus, generated high-Ti magmas. The trapped liquid component contained elevated abundances of heat-producing elements (K, U, and Th), increasing the fertility of associated source-regions. With time, melting moved progressively deeper in the mantle to source-regions with less of the residual, heat-producing, magma ocean liquid. Due to density contrasts, the ilmenite-bearing upper portions of the lunar mantle sank into the cumulate pile, possibly carrying more fertile material with it and allowing melting of more Mg-enriched source regions (to form the high-Ti picritic glass beads). Thus, the major controlling factors in the melting of the lunar interior could be the proportion of trapped, magma-ocean liquid in the cumulate source and the sinking of more fertile, ilmenite-bearing material into the lower mantle.