The Fate of Sulfur during Mantle Melting - Implications for Sulfur Concentration of OIB versus MORB Sources

Abstract:

Sulfur is one of the key volatiles that are less affected by degassing and may capture deeper magmatic processes and mantle source characteristics. S content of OIBs is controlled by S solubility of partial melts at sulfide saturation which depends on P, T, fO₂ and melt composition (e.g., [1]), S content of the mantle, and fractional crystallization of primary basalts. In this study, we coupled mantle melting and basalt differentiation models with sulfur solubility models and compared our results with the S abundance data of OIBs. The goal was to compare and contrast the efficiency of S extraction from MORB and OIB sources by partial melting and put constraints on the S content of the mantle source.

S contents in basalts from Loihi, Samoa, and Galapagos islands range from 800 to 2500 ppm with MgO from 11 to 4 wt.%. Modeling of S solubility change during isobaric fractional crystallization for basalts from these three islands with MELTS and an SCSS parameterization [1] showed that SCSS are distinctly higher than S concentration in the basalts at MgO>8wt.% (e.g., [2-3]). Assuming that there could be ≤20% S⁶⁺/SS present in the OIBs derived from a potentially more oxidized source [4], sulfide-undersaturated primitive melts with 700-900 ppm S are required to match the S concentration in the primitive OIBs. Derivation of sulfide undersaturated low-degree melts (Loihi: F=2-4%; [5]; Samoa: F=1-6%; [6]) from a mantle source with potential temperature (T_P) of 1420 °C requires that the mantle source S content is ≤50 ppm. For Galapagos spreading center with F~10% [7], ≤100 ppm S is required in the mantle source. While low F constrains low S abundance in OIB mantle, more sulfide could be present in the MORB source (eg. 146±35 ppm S, [8]) owing to higher average extent of melting. The estimated S content for OIB sources above should be the lower bounds since T_P in OIB mantle could be hotter and increasing T_P increases SCSS significantly because of increased FeO* in the deeper partial melt.

 [1] Li and Ripley (2005), Miner. Deposita; [2] Jenner and O’Neill (2012) Geochem. Geophys. Geosyst.; [3] Dixon and Clague (2001) J. Petrol; [4] Jugo et al. (2010) GCA, 74, 5924-5938. [5] Norman et al. (2002) CG, 183, 143-168. [6] Jackson and Hart (2006) EPSL, 245, 260-277. [7] Geist et al. (2005) J. Petrol. 46, 2197-2224. [8] Saal et al. (2002) Nature, 419, 451-455.