Solubility and speciation of sulfur in silicate melts under crustal conditions

Lauren R Gorojovsky , Bernard J. Wood

We have determined the solubility of sulfur as either sulfide (S2-) or sulfate (S6+) in a wide range of silicate melts at 1 atm pressure and temperatures of 1050° to 1250°C. The method involved suspension of the melt in either a mixture of CO2-CO-SO2 (sulfide solubility) or SO2 and air (sulfate solubility) for periods of up to 120 hours. Sulfur concentrations, measured by electron microprobe were converted into sulfide capacity CS2- and sulfate capacity CS6+ using (Fincham and Richardson; 1954):
logCS2-= log⁡[S2-] + 1/2 log(fS2/fO2)
logCS6+= log⁡[SO42-] - 1/2 logfS2 - 3/2 logfO2
[S2-] and [SO42-] refer to weight % sulfur dissolved in the melt as sulfide and sulfate respectively. Our new results demonstrate that extrapolation of earlier data to temperatures below 1200°C yields good agreement for sulfide capacity but overestimates sulfate capacity. This means that sulfide is appreciably more stable relative to sulfate in crustal magmas (temperatures <1200°C) than previously calculated.
A major consequence is that the crossover from S2- at low fO2 to S6+ at high fO2 shifts upwards by ~0.6 logfO2 units relative to the FMQ buffer as temperature declines from 1200° to 1050°C. The large temperature effect on sulfur speciation also means that there is electron exchange between Fe2+ and S6+ during magma cooling and ascent leading to high measured Fe3+/Fe2+ in quenched melts, high calculated fO2 and relatively reduced sulfur with high S2-/S6+even at +2-3 logfO2 units above the FMQ buffer. This self-oxidation mechanism at low temperatures is a major contribution to the observation that hydrous S-bearing arc magmas are more oxidised than MORB, which are generated at low fO2 and which erupt at higher temperatures. Furthermore, these oxidised hydrous melts are sulfide saturated and should precipitate an Fe-rich sulfide throughout their path of ascent and differentiation in lower, mid and upper -crustal levels. Given that such sulfides would scavenge Cu and other chalcophile metals we suggest that the occurrence of Cu-(±Au) porphyry deposits is governed less by the capacity of magmas to remain Cu-rich, and more by the architecture and evolution of the cumulate pile, the timing and depth of volatile saturation, and the efficiency of Cl-rich fluids in later mobilisation of metals.