In-situ geochemistry of sulfides in highly metasomatized mantle xenoliths from Kerguelen, southern Indian Ocean
Résumé
Mantle xenoliths from the Kerguelen Archipelago record a complex multistage history involving a high degree (15 to 25%) partial melting that created a harzburgitic mantle completely stripped of Base Metal Sulfides (BMS), followed by pervasive melt-rock reaction with alkaline melts above the Kerguelen mantle plume. Subsequent reaction of the highly refractory protolith with small volumes of carbonate-rich silicate melts led to a re-enrichment in BMS (up to 0.05 wt.%). Two BMS precipitation mechanisms are suggested: immiscibility from the silicate-carbonate melt and sulfidation reactions from a CO2-rich supercritical fluid. In-situ analyses of chalcophile and siderophile elements (major and trace levels) in the BMS shed new light on their origin. The BMS phases that precipitated via immiscibility are metal-rich sulfide melts which progressively evolved toward Ni and Cu-rich end-members by cumulate fractionation of monosulfide solid solution (mss) during percolation inside the peridotites. Some cumulate mss have elevated and fractionated IPGE contents (200-900x C1-Chondrite abundances), indicating random digestion of preexisting Os, Ir, Ru-rich PGM by the percolating sulfide melt. The BMS that precipitated by sulfidation reactions from a CO2-rich vapour phase are subsolidus exsolution products from Cu-bearing but Ni-poorer mss. They have the highest concentrations of PGEs and show selective enrichment in S, Pd, Pt and Os over Cu, S, Ir, Ru and Rh. Their PGE compositions confirm experimental data, which demonstrate that S, Pd, Pt and Os can be efficiently transported in a CO2-rich supercritical fluid. Superchondritic (S/Se), (Os/Ir) and (Pd/Pt) in both bulk-rocks and individual sulfides are inferred to be the geochemical fingerprints of sulfide crystallisation from a CO2-rich vapour exsolved from a highly evolved carbonate-rich metasomatic melt.