Experimental investigation of the transport of sulfur from the subducting crust to the mantle wedge

Subduction zone magmas are considered to be important carriers of sulfur from the sub-arc mantle wedge to the arc crust - through deposition of sulfide ores, and to the atmosphere - through volcanic degassing. Slab-derived sulfur is also proposed to be linked to the oxidation state of the mantle wedge [1]. However, the origin of the sulfur enrichment of most of arc magmas and the transport of sulfur from the subducting slab to the mantle wedge are poorly understood. Here we report experimental measurements of the sulfur content at sulfide saturation (SCSS) of slab-derived hydrous partial melts at a pressure of 2.0 GPa and a temperature range of 800-1050 °C. A synthetic MORB + 6.8 wt.% H₂O doped with 1 wt% S (added as pyrite) was used as starting material. The experiments were conducted in a piston-cylinder device with samples contained in Au inner capsules and Ni-NiO (fO₂ = FMQ+0.5) or Co-CoO mixtures in Au-Pd outer capsules. Sulfur concentrations in quenched silicate glass and the major element composition of the experimental phases were determined by EPMA. All the experiments contain garnet, cpx, rutile, pyrrhotite, and fluid with amphibole, quartz, and silicate melt present at 800, 800-950, and 850-1050 °C, respectively. The partial melt composition ranges from rhyolitic to dacitic with increasing temperature and melting degree (up to ~30 wt% partial melt). At 1000-1050 °C, the pyrrhotite crystals are almost completely consumed. At all the temperatures investigated, sulfur concentrations in melt are very low, from 60 to >300 ppm S, but consistent with previous experiments at lower pressures [2, 3]. Sulfur contents of the melts appear to be controlled by sulfur fugacity fS₂ (calculated from the composition of pyrrhotite crystals) and temperature. Bulk mass balance calculations show that the proportion of sulfur dissolved in the silicate melt is always very low, i.e., less than 1 wt% of the amount of sulfur initially added to the system is transferred to the partial melt. Our preliminary results suggest that more than 99 wt% of the bulk sulfur present in the slab are partitioned between the pyrrhotite crystals and the aqueous fluid phase, and sulfur solubility in the fluid phase appears to increase with increasing temperature. This suggests that the slab-derived fluid phase must be considered as a more efficient agent for the transport of sulfur from the slab to the mantle wedge as compared to slab melts, so that particular magmas whose petrogenesis is thought to involve slab partial melting may be relatively sulfur-poor. However, in the absence of excess hydrous fluid and for low-T conditions (<850 °C) relevant for downgoing slab at sub-arc depths, most of crustal sulfide is expected to remain stable in the subducting crust. [1] Kelley and Cottrell (2009), Science 325, 605-607. [2] Botcharnikov et al. (2004), Chem. Geol. 213, 207-225. [3] Liu et al. (2007), Geochim. Cosmochim. Acta 71, 1783-1799.