Experimental Study of the Behaviour of Gold in Calc-alkaline Arc Magmas : Demonstration of the Effect of Sulphur and Geological Implications on the Formation of Gold Deposits
Résumé
AB: Copper-gold-molybdenum deposits represent major metal resources known to be spatially and temporally associated with intrusive arc magmatism (Hedenquist and Lowenstern, 1994). Although the bulk of the ore metals seem to originate from the magmas, there is as yet no consensus on processes responsible for their concentration. The fact that primary ore minerals are predominantly sulfides has led to the suggestion that sulfur may play an important role in metal enrichment processes at the magmatic stage (Sillitoe, 1997; Mungall, 2002). Previous experimental studies have shown that sulfides can largely fractionate gold from the melt, and therefore concluded that oxidized magmas are more likely to be Au-enriched. Here, we conducted high-pressure temperature experiments (4 Kbar, 1000°C) on three dacitic and two rhyolitic natural magmas from the Philippines (both adakitic and typical calc-alkaline compositions were used) over a range of fO2 corresponding to reducing (~NNO-1), moderately oxidizing (~NNO+1.5) and strongly oxidizing (~NNO+3) conditions as measured by solid Ni-Pd-O sensors. The experiments were carried out in gold capsules, the latter also serving as the source of gold, in the presence of variable amounts of H2O. The water content of experimental glasses was determined by Karl-Fisher Titration. Both S-free and S-bearing (~1 wt% S added) experiments were performed. Sulfur concentrations in experimental glasses were measured by electron microprobe, and thermodynamic calculations were used to compute the fS2 of the experimental charges. Gold concentrations in glasses were determined by LA-ICP-MS. Charges consist of dominant silicate melt quenched to glass plus minor silicate phases (mostly Cpx, rarely Opx), together with discrete magnetites in the most oxidized charges. All S-bearing charges are saturated with either sulfides or sulfates, together with discrete Au-Fe-S alloys, depending on the fO2. Our data show that gold solubility in S-free charges is low (30-240 ppb) but globally increases with fO2, suggesting the implication of oxidized species in the dissolution of gold. Gold solubility in S-bearing charges appears much higher, ranging from ~500 to 5200 ppb. Gold content is higher in reduced than in moderately oxidizing conditions, and is not affected by melt composition. In very oxidizing conditions, Au solubility is low and nearly constant, and does not seem affected by the presence of S ; it only shows a dependence on fO2 by following the same trend than S-free charges. At ~NNO+1.5, gold content of silicate melt appears to be the result of a complex competition between fO2 and fS2, whereas under more reduced conditions, the effect of fS2 appears the strongest, leading to a direct dependence of gold solubility on sulfur content. This study shows conclusively that Au transport and concentration in silicic melts can be dramatically enhanced by the presence of sulphur. A major consequence is that sulfide crystallization is not likely to scavenge all Au present in the magma, but may allow gold-enriched residual magmas, especially those that are moderately reduced (NNO-1 to NNO+1.5), to be emplaced in the upper crust. Gold enrichment in porphyry- and epithermal- type deposits can be directly linked with sulfur incorporation and abundances in magmas.