Apatite-Melt Volatile Partitioning Under Lunar Conditions

Introduction: The volatile component of apatite [Ca5(PO4)3(F,Cl,OH)] makes it a popular choice for studying the abundance and distribution of indigenous lunar volatiles [e.g. 1-5]. The thermodynamic behavior of volatile partitioning into apatite is, however, poorly constrained. Limited experimental studies of apatite/melt partitioning have focused on terrestrial systems [6], chlorine-brine systems [7], or Martian conditions [8]. Recent studies applicable to basaltic magmas have varied parameters in each experiment [9] making it difficult to extract individual effects of temperature, pressure, and melt composition. Modeling of apatite/melt partitioning of volatiles in crystallizing melts [10] demonstrated the importance of quantifying abundances of all of the main volatiles, as they are major structural components in apatite. At present, experimentally-derived partition coefficients remain undetermined for lunar compositions and conditions. In this study, fluorapatite, depleted in OH and Cl, was grown in equilibrium with silicate melt to provide a baseline for further experiments. These experiments were conducted with bulk compositions similar to that of late-stage mesostasis regions in Apollo lunar basalt samples [11] in which apatite is typically found. Once partitioning behavior in a nominally ‘dry’ fluorapatite system is constrained additional experiments will add in varying amounts of OH and Cl. The aim is to enable accurate and precise tracking of the effects of volatiles on partitioning. Once these effects are known robust back-calculations of volatile contents in lunar melts, from which apatite crystallized, could be attempted. Methods: Experimental methods. Experiments were conducted in an end-loaded piston cylinder at VU University Amsterdam using a talc-pyrex assembly. For each experiment a graphite bucket was filled with starting material and closed with a graphite lid. The bucket was inserted into a triple crimped and welded Pt capsule. The double-capsule was then placed in an alumina inner tube. A W97Re3-W75Re25 thermocouple (T/C) was used to measure and control T throughout the experiment. The sample was placed in the hotspot of the assembly, 2 mm from the T/C tip end, shown to be 10 C hotter than the T/C reading [12]. Runs were performed using a “hot piston-in” technique to achieve the desired P-T conditions. Each mixture was heated to 1550 C (super liquidus) at 100 C/minute, dwelled for 10 minutes, and subsequently cooled to final experimental temperatures at 50 C/hr. During the heating cycle P was kept low, ~300 psi, until the final T was reached and then increased to ~715 psi (equivalent to 1 GPa). Experiments were left to dwell for 21 hrs at final T. Runs were quenched by turning off the power. Experiments were conducted for 7, 18, 46, and 72 hrs to assess equilibrium.