Experimental determination of the solubility of platinum in silicate melts

We report here new experimental results on the solubility of Pt in silicate melts at one atmosphere total pressure. Experiments were made with two different melt compositions: an anorthite-diopside eutectic and a CaO-free system (3SiO(2) . MgO . Al2O3, by weight). Oxygen fugacities ranged from pure oxygen to 10(-8) atm and temperatures from 1300 to 1560 degrees C. Glass samples were analysed after the experiments by instrumental neutron activation analysis (INAA). In experiments with An-Di melts, isotopically enriched Ca-40 (and correspondingly depleted Ca-46) was used to reduce the background from Sc-47 produced by neutron activation of Ca-46. In this way, the interference was reduced by a factor of 500, allowing the determination of Pt in glass samples by INAA down to a concentration of about 50 ppb. Over five orders of magnitude in oxygen fugacity, from pure oxygen to an f(o2) of 10(-5) atm, a systematic decrease of the Pt solubility with decreasing f(o2) Was observed. The slope of the log (Pt solubility) vs. log f(o2) correlation suggests a formal valence of +2 for Pt in the silicate melt, independent of temperature. The solubility in the CaO-free melt is about one-fourth of an order of magnitude lower than for the An-Di melt. At oxygen fugacities below 10(-5) atm, results show a large scatter. We suspect that formation of Pt nuggests leads to very inhomogeneous distribution of Pt in the melt. A literature survey shows large variations in Pt solubilities determined by Various authors. These discrepancies primarily result from the formation of Pt nuggets. A new mechanism of PGE nugget formation is suggested. The formation of PtFe alloys predicted to occur in terrestrial magmatic melts may be important for fractionating Pt from other noble metals. Copyright (C) 1997 Elsevier Science Ltd.