The dissolution mechanism of sulphur in hydrous silicate melts. II: Solubility and speciation of sulphur in hydrous silicate melts as a function of fO2

Raman and X-ray absorption spectroscopy (XANES) measurements on a series of experimentally synthesised, sulphur (S)-bearing, hydrous silicate glasses were used to determine the S-speciation and S-oxidation state as a function of glass composition and oxygen fugacity (fO2) and to decipher the dissolution mechanism of S in silicate melts. Synthesised glasses include soda-lime (SLG), K2Si4O9 (KSG), albite and trondhjemite (TROND) compositions. A series of SLG and KSG glasses, doped with small quantities of Fe, was also studied in order to determine the effect of Fe/S on the S solubility. The experiments were performed in internally heated (IHPV) and cold seal (CSPV) pressure vessels at 200 MPa, 1000 and 850 °C and a range of fO2 from log fO2 = QFM − 2.35 to QFM + 4 (QFM is quartz–fayalite–magnetite oxygen buffer). stematic correlation of features in Raman and XANES spectra allows the identification of at least four different S-species in the glasses depending on fO2 and Fe/S of the system. In XANES spectra of Fe-free glasses SH−, H2S and SO42 − are visible as peaks at 2466, 2471.8 and 2482 eV, respectively. In Raman spectra peaks at 2574 and 990 cm− 1 indicate the presence of HS bonds and SO42 −, respectively, but SH− and H2S can not be distinguished using a Raman spectroscopy. In Fe-bearing glasses FeS bonding is identified at 2469 eV in the case of XANES and at 298, 372 and 420 cm− 1 in the case of Raman spectra. The intensities of peaks related to SH bonding systematically decrease and the intensities of peaks related to FeS bonding systematically increase with increasing Fe/S in both the XANES and the Raman spectra indicating that in the presence of Fe, FeS bonding is preferred over SH bonding. The total S solubility at sulphur saturation in the Fe-free melts is a function of the degree of melt polymerisation and it increases with increasing NBO/T (from 0.03 to 1.91 wt.% S). The S2 − species are more soluble than the S6 + species in contrast to previously studied Fe-bearing “natural” compositions. ange from S2 − to S6 + is observed at log fO2 = QFM − 1 to QFM + 1 which is ~ 1.5 log unit lower than the range of fO2 previously reported for Fe-rich compositions indicating that Fe influences not only the speciation but also the oxidation state of S in silicate melts at given redox conditions. The natural implications are that S6 + in Fe-poor magmas can be stable at lower fO2 than previously predicted and, hence, S6 + may act as an oxidising agent in the mantle wedge by successively oxidising Fe2 + to Fe3 + via the reaction H2SO4 + 9FeO = FeS + 4Fe2O3 + H2O. For the silicate melt generated in the mantle wedge and containing about 10 wt.% total FeO, the change in the Fe3 +/ΣFe ratio from 0.1 to 0.2 will correspond to an increase in the log fO2 from QFM − 0.5 to QFM + 1.5 and will require only 1000–3000 ppm S extracted from subducted slab.