Near-chondritic relative abundances of sulfur (S), selenium (Se), and tellurium (Te) observed in mantle peridotites have been used to support the hypothesis of a carbonaceous chondrite-like late veneer added to the proto-Earth. However, the extent to which the observed S, Se, and Te compositions represent the signature of the silicate Earth remains a topic of debate. The concentrations of Se and Te in mantle-derived melts, such as mid-ocean ridge basalts (MORBs), can help clarify this issue, provided we have a precise understanding of the behavior of Se and Te during magmatic differentiation and mantle partial melting. Here we conduct laboratory experiments to determine the sulfide liquid–silicate melt partition coefficients of Se and Te (DSe,TeSul/Sil) at crust–mantle conditions. Our results indicate that DSe,TeSul/Sil range from 180 to 2200 for Se and from 1000 to 25,400 for Te, exhibiting an inverted U-shaped dependence on the FeOtot content in silicate melt. We parameterize DSe,TeSul/Sil as a multi-function of the compositions of silicate melt and sulfide liquid, with pressure (0.5–2.5 GPa), temperature (1273–1973 K), and oxygen fugacity (FMQ-5 to FMQ+1.5; FMQ refers to the fayalite–magnetite–quartz buffer) having negligible effects on DSe,TeSul/Sil. Applying our parameterization to magmatic differentiation can effectively account for the Se, Te, and Cu systematics observed in both MORBs and oxidized arc magmas. More significantly, when applied to mantle partial melting, alongside high-precision Se and Te concentration data from MORBs, our parameterization yields superchondritic S/Se, S/Te, and Se/Te ratios in the depleted MORB mantle and primitive mantle, compared to carbonaceous chondrites. These findings suggest that the S, Se, and Te abundances in the silicate Earth were likely established during the main accretion phase, with the late veneer playing only a minor role.
Liu et al. (Wed,) studied this question.