Constellation, le dépôt institutionnel de l'Université du Québec à Chicoutimi

Résumé

Over the past two decades the distribution of trace elements in base-metal sulfides (BMS) from magmatic sulfide deposits (i.e. Ni-Cu-PGE and PGE-dominated deposits) has been extensively investigated, and much information is now available. We have compiled the trace-element concentrations in BMS, obtained by LA-ICP-MS, from various magmatic sulfide deposits to understand whether they may record the ore forming processes. Among the samples, there are some of the most studied Ni-Cu-PGE (e.g. Aguablanca, Duluth, Jinchuan, Noril’sk-Talnakh, Sudbury, Voisey’s Bay and others), and PGE-dominated (e.g. Bushveld, Lac des Iles, Stillwater, Great Dyke and Penikat) deposits of the world. The results reveal that it is possible to separate BMS from Ni-Cu-PGE and PGE-dominated deposits based on the concentrations of highly chalcophile elements. For instance, pentlandite from Ni-Cu-PGE deposits has much lower Rh and Pd concentrations than those from PGE-dominated deposits. This is because of the higher R-factors in PGE-dominated deposits relative to Ni-Cu deposits, and consequently greater concentrations of these elements in the sulfide liquid. The BMS also record the fractional crystallization of the sulfide liquid. The concentrations of elements compatible with MSS and ISS are lower in BMS from more fractionated ores, whereas concentrations of incompatible elements increase in BMS from progressively more fractionated ores. Also, since both Se and Te behave as incompatible elements during crystallization of the sulfide liquid, but Te is more incompatible than Se, the Se/Te ratio in BMS decreases with progressive fractionation, and can also be used to track the differentiation of the sulfide liquid. Crustal assimilation by the parental silicate liquids can also be constrained using the concentration of slightly chalcophile elements in BMS. This is because high R-factors do not obscure the effect of crustal contamination for elements with low partition coefficients between sulfide and silicate liquids. We propose that a plot of As/Se vs Sb/Se in pentlandite may be appropriate for assessing crustal contamination. The pentlandite from ores that formed from more contaminated silicate liquids have higher As/Se and Sb/Se ratios. Finally, the composition of pyrite may be used to investigate the late- or post-magmatic alteration of BMS. Magmatic pyrite have higher Co/Se and Sb/As ratios relative to pyrite formed in other settings, which could be useful for the use of pyrite as an indicator mineral.