The effect of anhydrite saturation on the fate of sulfur during fluid-present melting of subducting basaltic crust
Abstract
The apparent sulfur enrichment of sub-arc mantle is thought to derive from an oxidized downgoing slab, and it has been suggested that the slab-derived sulfate species is responsible for oxidizing the mantle wedge [1]. However, the conditions and extent of sulfur transfer from the subducting slab to the mantle wedge are poorly understood. In particular, the relative mobility of sulfur as a function of oxygen fugacity (fO2) is unconstrained at sub-arc depths. To add to our recent study on sulfur mobility during fluid-present melting of a sulfide-bearing basaltic crust [2], here we constrain the fate of sulfur during similar melting at relatively oxidizing conditions, i.e., at sulfate saturation. Experiments were performed using a piston cylinder device at P = 2-3 GPa, T = 950-1050 °C. A synthetic MORB + 6.8 wt.% H2O doped with 1 wt% S (added as pyrite) was contained in AuPd inner capsules and hematite-magnetite (HM: ~FMQ+3.9 to +4.6) mixture used as fO2 buffer was housed in Pt outer capsules, following the recently proposed design of ref. [3]. Sulfur concentration in quenched silicate glasses, the major element phase compositions, and fO2 of the experiments based on dissolved Fe contents in AuPd and added Pt sensor [4, 5], were determined using EPMA. All experiments contain silicate melt, cpx, garnet, anhydrite, rutile and/or Ti-magnetite, and are fluid saturated. The partial melt compositions are rhyolitic to rhyodacitic with increasing T and melting degree. Sulfur contents in the melt range from ~700 to 3000 ppm, and increase with increasing P and T, in agreement with published SCAS models [6, 7]. Mass balance calculations show that the proportion of sulfur dissolved in silicate melt can be >13% of the bulk sulfur at 1050 °C. However, at slab surface (<900 °C), the major part of the bulk sulfur present in the slab is dissolved in the aqueous fluid phase, the rest being stored as anhydrite crystals. Moreover, our results suggest that sulfur partition coefficient between aqueous fluid and silicate melt (DSf/m) decreases from ~200 to ~100 at 2 GPa and from ~100 to ~50 at 3 GPa, between 950 and 1050°C, respectively. Our study shows that at high T and fO2, DSf/m is lower than at low fO2 conditions of pyrrhotite saturation but, at slab surface Ts, tends to reach DSf/m values observed in reducing conditions. Although anhydrite-saturated slab partial melts may carry some S, slab-derived fluid is still the most effective agent of S transfer. Moreover, compared to pyrrhotite-saturated systems, the concentration of sulfur in fluid at anhydrite saturation is higher by no more than a factor of two, implying that metasomatism by ~0.5 wt.% oxidized slab-derived fluids, relative to the wedge, will be sufficient to enrich the arc source mantle, instead of ~1 wt.% fluid in reducing conditions [2]. Experiments at lower temperatures (800-900 °C) and fO2s intermediate between ~FMQ and FMQ+4, using NixPd1-x-NiO alloy buffers, are underway and will be presented. [1] Kelley and Cottrell (2009), Science 325, 605-607; [2] Jégo and Dasgupta (GCA, under review); [3] Jakobsson (2012), CMP, in press; [4] Médard et al. (2008), AmMin 93, 1838-1844; [5] Balta et al. (2011), AmMin 96, 1467-1474; [6] Li and Ripley (2009), EconGeol 104, 405-412; [7] Baker and Moretti (2011), RiMG 73, 167-213.
- Publication:
-
AGU Fall Meeting Abstracts
- Pub Date:
- December 2012
- Bibcode:
- 2012AGUFM.V33C2884J
- Keywords:
-
- 1030 GEOCHEMISTRY / Geochemical cycles;
- 1031 GEOCHEMISTRY / Subduction zone processes;
- 3630 MINERALOGY AND PETROLOGY / Experimental mineralogy and petrology;
- 8413 VOLCANOLOGY / Subduction zone processes