Seminar Rajdeep Dasgupta

Abstract

Silicate differentiation of Earth is aided chiefly by shallow mantle partial melting that generates basaltic melts. However, mantle melting commences at deeper depths, aided by trace quantities of carbon. Although several lines of evidence exist, based on laboratory experiments and studies of natural samples, that carbon plays a central role in several types of silica-poor magmas, no straightforward tool existed for estimating the depth of generation of natural silica-deficient magmas. I will present our work on a melt composition-based thermobarometer that allows estimating the depth of final melt-mantle equilibration of mantle-derived, a wide range of silica-poor magmas from carbonatite, kimberlite, to alkali basalts (Sun and Dasgupta, 2020). I will present the application of this newly available tool to two different data sets: (1) kimberlites from continental settings (Sun and Dasgupta, 2020) and (2) intraplate ocean island basalts (OIBs; Sun and Dasgupta, 2023). I hope to convince that global data sets of kimberlite major element geochemistry through time likely suggest thinning of continental lithospheric mantle. I will further show that the averaged primary melt compositions of alkaline OIBs are more CO₂-rich than those of subalkaline OIBs, but both series yield final equilibration with the mantle at ~100–150 km and mantle potential temperatures of ~1430–1530 °C. Using this estimation, I will argue that alkaline and subalkaline OIBs are generated from carbon-rich and carbon-poor domains, respectively, in the mantle plume sources, rather than deriving from two different sources in the plume and lithospheric mantle or variable extents of melting of a homogeneous source. Despite the source heterogeneity, CO₂ in the primary melts of individual island groups displays strong positive correlations with Nb and Ba for the less-degassed deep, OIB source mantle. Using our analysis, we estimate that the deep mantle on average contains ~330–400 ppm carbon, significantly higher than those estimated for the MORB source globally (e.g., Dasgupta and Aubaud, 2025).