Geochemistry of the Yumen picrites-basalts from the Emeishan large igneous province: Implications for their mantle source, PGE behaviors, and petrogenesis

Abstract

Olivine compositions are often used to infer the source lithology of basaltic rocks, but the effect of source melting conditions on olivine chemistry remains poorly studied. Here we present mineral chemistry, whole-rock major and trace elements, SrNd isotopes, and platinum-group element (PGE) concentrations for high-Ti picritic-basaltic rocks from Yumen area. These data provide new constraints on the source lithology and petrogenesis of these high-Ti rocks in the Emeishan large igneous province (ELIP). The picrites have primitive compositional features (e.g., high Mg# values, depleted SrNd isotopic compositions, no PGE depletions and/or fractionations). Some basaltic rocks have negative NbTa anomalies, enriched SrNd isotopic compositions, and depleted and/or fractionated PGE contents, indicative of fractional crystallisation, variable degrees of crustal contamination, and magmatic sulfide segregation. Combined with published data for the ELIP, the compositions of the primitive magmas and 10000Zn/Fe and FC3MS (FeO/CaO − 3 × MgO/SiO2, all in wt%) values of the ELIP high-Ti picrites are similar to those of garnet peridotite-derived experimental melts. Based on the primitive magma compositions, the ELIP high-Ti melts could have been produced by mantle melting at 4.0–5.4 GPa and 1620–1660 °C, and started to crystallise olivine at 1410–1550 °C, based on the Al-in-olivine thermometer, which are higher temperatures than for mid-ocean ridge basalts and ELIP low-Ti magmas. The high-Ni and low-Mn olivine phenocrysts in the ELIP high-Ti samples crystallised from peridotite-derived, MgO-rich melts at high temperatures. Variations in the physicochemical source melting conditions was enough to explain the differences in olivine Ni and Mn contents in the ELIP high- and low-Ti picrites-basalts. Additionally, the almost constant Pd, Pt, and Rh contents with decreasing Mg# values indicate intermediate partition behaviors in the ELIP magmas. No correlations exist between whole-rock 10,000 × Pd/Yb and Mg#, (Th/Nb)N, and εNd(t) values, implying that selective assimilation of crustal sulfur resulted in magmatic sulfide saturation. We propose that the ELIP high-Ti magmas were derived from a deep-sourced garnet peridotite, and experienced various igneous processes during ascent through the continental lithosphere to form different types of ELIP high-Ti basaltic rocks.