Experimental differentiation of kimberlitic to carbonatitic melts: crystallization sequence and liquid line of descent

Abstract Kimberlites are volatile-rich, silica-undersaturated, mantle-derived magmas with an important role in studying mantle geochemistry, yet their compositional evolution during crystallization remains poorly constrained. This study defines the crystallization sequence and liquid line of descent of a reconstructed kimberlitic melt, using high-pressure experiments at 1–3 GPa and 1100–1400 °C. Early crystallization is dominated by olivine ± chromite, consistent with petrographic observations of natural samples, followed by clinopyroxene, ilmenite, perovskite, apatite, and phlogopite. The absence of clinopyroxene in natural kimberlitessuggests that kimberlitic melts remain above 1150 °C at 1 GPa, conditions at which clinopyroxene is not observed in our experiments. Substantial cooling of kimberlitic melt and related abundant crystallization likely occur in the crust, possibly linked to extensive degassing at shallow pressures. The experimental melts evolve continuously through decreasing SiO 2 and MgO, and enrichment in CaO, alkalis (Na, K), and volatiles (CO 2 , H 2 O), ultimately transitioning at ≤ 1150 °C to carbonatitic melts with 7–10 wt% SiO 2 , 6.5–7.5 wt% Na 2 O + K 2 O, and up to 35 wt% volatiles. Olivine (± clinopyroxene) fractionation drives Si depletion at almost constant MgO + FeO + CaO and moderate alkali-enrichment, such that the carbonate-silicate miscibility gap is bypassed. This evolution is in sharp contrast to mafic alkaline silicate melts where olivine + clinopyroxene crystallization causes Si enrichment hence promoting melt evolution towards the carbonate-silicate miscibility gap. Overall, the experimental results demonstrate a petrogenetic continuum between kimberlitic and carbonatitic melts and provide constraints on the crystallization conditions of kimberlitic melts.