LUNAR APATITES AND SILICATE LIQUID IMMISCIBILITY

In the recent years, we have seen a flourish of research data on OH in lunar apatite and its implications of water in lunar mantle. Meanwhile, the challenges in using apatite to infer water in the lunar mantle have continued to grow. Boyce et al. (2014) evaluated the uncertainties in unknown thermodynamic properties of volatile partitioning and the fluid chemistry. However, a less recognized challenge comes with lunar apatite and silicate liquid immiscibility (Taylor et al., 2014). Apatite is a late-stage mineral, after extensive crystallization of a basaltic melt. In lunar samples, apatite appears in the late-stage mesostasis that often displays signs of silicate liquid immiscibility (SLI). Thus, the process of SLI contributes to additional complexity in understanding the partitioning behaviors of volatiles, as exemplified by Pernet-Fisher et al. (2014a, b).

In order to address the relationship between SLI and lunar apatite, we have investigated the textural and compositional variations of apatite and other phosphates in mesostasis in lunar basalts (Pernet-Fisher et al., 2014a, b). We demonstrated that mesostasis pockets in a single lunar basalt represent isolated heterogeneous systems that have evolved independently from each other, each displaying distinct ranges in OH concentrations (a factor of 5 in a single thin-section, and up to a factor of >10 within an individual sample) and in D/H values by a factor of 2-3 in one section. Further, apatites in these mesostasis pockets appear in both the K-rich (felsic) and REEP-rich (Fe-basaltic) melts formed by SLI, which is a product of the Fenner-trend fractional crystallization of typically FeO-rich lunar magmas at low fO₂ values (e.g., < IW). Apatites crystallized in equilibrium with these two immiscible melts would require an additional set of OH-partitioning coefficients when back-calculating magmatic water contents.

These results demonstrate the need for critical considerations on: 1) OH phases in heterogeneous mesostasis; 2) phosphate phases formed by SLI; and 3) more than one OH-bearing phosphate – i.e., merrillite/whitlockite and apatite. The lack of incorporation of these factors into any OH modelling from phosphate minerals casts serious doubt on many previous-calculated OH contents of parental magmas and subsequent estimates of water in the lunar mantle.