Abstract

The partitioning of platinum-group elements (PGE; Os, Ir, Ru, Rh, Pt, and Pd) and Au between sulfide melt and silicate melt (i.e., DPGEsul) exerts a critical control on the PGE composition of the Earth’s crust and mantle, but previous estimates have been plagued by experimental uncertainties and vary through several orders of magnitude. Here we present direct experimental measurements of DPGEsul, based on in situ microanalysis of the sulfide and silicate melt, with values ranging from ∼4×105 (Ru) to ∼2–3×106 (Ir, Pt). Our measurements of DPGEsul are >100 times larger than previous results but smaller than anticipated based on comparison of alloy solubilities in sulfide melts and S-free silicate melts. The presence of S in the silicate melt greatly increases alloy solubility. We use our new set of partition coefficients to develop a fully constrained model of PGE behavior during melting which accurately predicts the abundances of PGE in mantle-derived magmas and their restites, including mid-ocean ridge basalts, continental picrites, and the parental magmas of the Bushveld Complex of South Africa. Our model constrains mid-ocean ridge basalt (MORB) to be the products of pooled low and high degree fractional melts. Within-plate picrites are pooled products of larger degrees of fractional melting in columnar melting regimes. A significant control on PGE fractionation in mantle-derived magmas is exerted by residual alloy or platinum group minerals in their source. At low pressures (e.g., MORB genesis) the mantle residual to partial melting retains primitive mantle inter-element ratios and abundances of PGE until sulfide has been completely dissolved but then evolves to extremely high Pt/Pd and low Pd/Ir because Pt and Ir alloys form in the restite. During melting at high pressure to form picrites or komatiites Ir alloy appears as a restite phase but Pt alloy is not stable due to the large effect of pressure on fS2, and of temperature on fO2 along an internal oxygen buffer, which causes large increases in alloy solubility. The magmas parental to the Bushveld Complex of South Africa appear, at least in part, to be partial melts of mantle that has previously been melted to the point of total sulfide exhaustion at low pressure, closely resembling mantle xenoliths of the Kaapvaal craton. Using the new extremely large DPGEsul the world-class Merensky Reef and UG2 Pt deposits of the Bushveld Complex can readily be modeled as the result of sulfide saturation due to mixing of magmas with unremarkable PGE contents, obviating the need to postulate anomalously PGE-rich parent magmas or hydrothermal inputs to the deposits.

Full Text
Published version (Free)

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call