Experimental study of sulfur solubility in silico-carbonatitic melts at pressures of 6-10 GPa
Final Report Abstract
In this project, we set out to develop of an experimental method for investigating the equilibrium compositions of sulfide and silico-carbonatite melts at mantle pressures and temperatures with special emphasis on the need for efficient separation of carbonatitic and sulfide melts in order to avoid analytical artifacts via the “nugget effect”. Our scientific goal was to experimentally determine the sulfur concentration in sulfide-saturated carbonatitic and silico-carbonatitic melts at pressures of 6-10 GPa, temperatures of 1200-1500°C, at relatively oxidized (with Re-ReO2) and reducing (graphite saturation) conditions. Under reducing conditions, our approach proved effective in keeping sulfide melt isolated, but at the same time remaining in chemical “communication” with the silicocarbonatite melt. The silico-carbonatite melt migrated through thin fractures in an olivine container, but the sulfide droplets could not. The melt is essentially a Ca–Mg–Fe carbonatite, with no particular pressure or temperature dependence on bulk composition. The silicocarbonatite melt occurs in three textural situations: i) a large melt pool adjacent the sulfide within the olivine container, ii) a large pool between the olivine capsule and an outer Pt capsule, and iii) small pockets of melt trapped within the olivine container. In a given experiment, the majorelement composition of all three textural types is essentially identical. Sulfur contents are higher and more variable adjacent to the sulfide globule (generally 0.1–0.3 wt % S) compared to the other two types of melt pools. Our observations indicate that S solubility at sulfide saturation (SCSS) is a more a function of FeO content than other parameters, including temperature and pressure. Considering the S concentrations from textural type (i) and those of types (ii) and (iii) as opposing half-brackets constraining the equilibrium value, we find 0.02–0.10 wt % S can be dissolved in silico-carbonatite melts at graphite saturation at 7.5–10 GPa and 1400– 1600°C, depending on FeO content. This amount is about two times lower than the solubility of S in basaltic liquids (e.g., MORB). Thus, the extraction of silico-carbonatitic melts under reducing conditions will have a negligible influence on the sulfide budget of the residual upper mantle. In experiments under oxidising conditions of the Re-ReO2 buffer, sulfide was generally absent, with calcium-rich sulfate-carbonate melt globules being the S-bearing phase. The solubility of S is 1–3 wt %, corresponding to 2.5–7.5 wt % SO3. Here, saturation is indicated by the appearance of an immiscible Ca-carbonate-sulfate melt. However, since the amount of carbonatitic partial melting is limited by overall low C contents in the upper mantle, the vast majority of S (~90 %) will remain in the peridotitic residuum. On the other hand, under oxidizing conditions, melts derived from other sources (i.e. subducting slab or sediments) can transport significant S, but saturation will be reached very quickly when a reducing environment is encountered. This could be a mechanism for simultaneous formation of diamond and sulfide, which could explain the frequent occurrence of sulfide inclusions in diamond.
Publications
- (2019) Experimental study of sulfur solubility in silicate–carbonate melts at 5–10.5 Gpa. Chemical Geology 505 12–22
Woodland, A. B.; Girnis, A. V.; Bulatov, V. K.; Brey, G. P.; Höfer, H. E.
(See online at https://doi.org/10.1016/j.chemgeo.2018.12.008) - (2018) Sulfur Solubility in Silico-Carbonate Melts at high pressures . 16th Experimental Mineralogy, Petrology and Geochemistry (EMPG) Meeting, Clermont-Ferrand, 17-21 June 2018
Woodland AB, Girnis AV, Bulatov VK, Brey GP, Höfer HE
