Sulfur concentration of mare basalts at sulfide saturation at high pressures and temperatures-Implications for S in the lunar mantle

Low estimate of S in the bulk silicate moon (BSM) [e.g., 1] suggests that sulfide in the lunar mantle is likely exhausted during melting. This agrees with estimates of HSE depletion in the BSM [2], but challenges the S-rich core proposed by previous studies [e.g., 3]. A key parameter to constrain the fate of sulfide during mantle melting is the sulfur carrying capacity of the mantle melts (SCSS). However, the SCSS of variably high-Ti lunar basalts at high P-Tare unknown. Basalt-sulfide melt equilibria experiments were run in graphite capsules using a piston cylinder at 1.0-2.5 GPa and 1400-1600 °C, on high-Ti (Apollo11, 11.1 wt.%; [4]) and intermediate-Ti (Luna16, 5 wt.%; [5]) mare basalts. At 1.5 GPa, SCSS of Apollo11 increases from 3940 ppm S to 5860 ppm, as temperature increases from 1400 °C to 1600 °C. And at 1500 °C, SCSS decreases from 5350 ppm S to 3830 ppm, as pressure increases from 1 to 2.5 GPa. SCSS of Luna16 shows a similar P-T dependence. Previous models [e.g., 6] tend to overestimate the SCSS values determined in our study, with the model overprediction increasing with increasing melt TiO2. Consequently, we derive a new SCSS parameterization for high-FeO* silicate melts of variable TiO2content. At multiple saturation points [e.g., 7], the SCSS of primary lunar melts is 3500-5500 ppm. With these values, 0.02-0.05 wt.% sulfide (70-200 ppm S) in the mantle can be consumed by 2-6% melting. In order to generate primary lunar basalts with S of 800-1000 ppm [1], sulfide in the mantle must be exhausted, and the mode of sulfide cannot exceed 0.025 wt.% (100 ppm S). This estimate corresponds with lower end values in the terrestrial mantle and further agrees with previous calculations of HSE depletion in the BSM [2]. [1] Hauri et al.,2015, EPSL; [2] Day et al.,2007, Science; [3] Jing et al., 2014, EPSL; [4] Synder et al.,1992, GCA; [5] Warren & Taylor, 2014, Treatise on Geochemistry; [6] Li & Ripley, 2009, Econ.Geol ; [7] Krawczynski & Grove, 2012, GCA.