Robinson, Katharine - Varied H concentration and isotopic composition in the Lunar Interior

Abstract: 
Water from the lunar interior has been measured in olivine-hosted melt inclusions, ferroan anorthosite plagioclase, residual glass in a KREEP basalt, and apatite grains from all other major lunar rock types (mare basalts, alkali suite, Mg-suite).The common goal of these measurements is two-fold: to constrain the water content of the bulk Moon, and to determine the source(s) of the Moon’s water via hydrogen isotopic ratios. Estimating the pre-eruptive water content of the parental magma from glasses and melt inclusions is fairly straightforward, and it was initially thought that similar estimates could be made using OH abundances in apatite. It has recently been shown that volatile partitioning into apatite is more complex than previously thought, invalidating estimates of parental melt water content from apatite. However, apatite data is still a useful recorder of D/H ratios and relative water contents might be discernible among different rock types.

Analysis of the picritic glass beads and melt inclusions showed that they originated from a magma with ~1000 ppm H2O.  The source region for that magma would have contained 100 ppm, assuming ~10% partial melting to form the pyroclastic magmas.  In contrast, our measurements of residual quenched glass in KREEP basalt fragments in 15358 contain 58-95 ppm H2O.  Based on the modal abundances of the glass and accounting for H loss, the initial melt would have contained ~100 ppm H2O.  The KREEP source would have thus contained ~10 ppm, an order of magnitude less than the picritic magmas.  These calculations are somewhat rough, but indicate that there are at least two possible water reservoirs in the lunar interior.

Water data from lunar apatite also indicates multiple reservoirs in the lunar interior.  Water content of apatite in the major rock suites varies by 10-50x and seems to be related to rock type.  KREEP-rich samples have the driest apatite, while mare basalt apatite is more water-rich.  Additionally, the delta-D values of apatite vary widely.  Elevated delta-D in mare basalts are almost certainly caused by H loss from lava flows, but some evolved, intrusive rocks also appear enriched in D (+200 to +500 permil).  However, there are some samples that fall in the range of the terrestrial upper mantle (-140 to +60 permil D).  Our measurements of apatite in quartz monzogabbro 15404, 55 have the lowest delta-D (-500) values reported from the Moon so far, indicating a third, low D source inside the Moon (Robinson et al. 2014 LPSC).

Varied water concentrations and delta-D in the lunar interior probably reflect a combination of processes involved in lunar formation, primary differentiation in the magma ocean, secondary magmatism, and addition of material to the Moon after accretion. Which process(es) dominated is far from clear. The first step in increasing our understanding of water distribution in the Moon is to pin down its variability in all rock types and relate that to other important parameters, such as the abundances of highly volatile elements.