Experimentally determined melting and phase relations in a volatile bearing mica-pyroxenite system – implications for mantle metasomatism and alkaline volcanism.

Heterogeneous assemblages within the lithospheric mantle represent some of the most enriched domains within the earth for alkaline, rare-earth and volatile elements. Volatile-rich mica-bearing pyroxenites are among these assemblages and were crucial in the recognition of metasomatism as a mantle process. However much of the experimental work until recently has focused on four-phase peridotites which are largely devoid of volatiles (i.e. C, H, N, S) and moderately volatile elements like fluorine unless they have been metasomatised. In domains where peridotite and pyroxenites coexist the chemical fingerprint of metasomatism is challenging to untangle, principally due to a lack of foundational phase relation studies in complex systems. Using a complex synthetic mica-pyroxenite system we will present new experimental results of melting and phase relations, and melt chemistry. Our experiments span 900 – 1400°C at pressures of 1, 2.5 and 5 GPa in a system containing carbon, water, sulfur, and fluorine, as well as 29 trace elements. Across all pressures we see the generation of a fluid phase at low temperature that acts as the precursor to a silicate or carbonated melt. The lower pressure silicate melts border the foidite field in TAS and straddle the leucitite and shoshonite divide in K2O vs. SiO2 space, while maintaining a K2O/Na2O value ranging from 0.9–3 in most experiments. The higher pressure carbonated melts from 5 GPa initially show low silica (~37wt%) that decreases to a low of ~27 wt% SiO2 as magnesite is consumed before increasing to roughly 41 wt% once major melting occurs at higher temperature. K2O/Na2O for these experiments ranges between 7 – 9 for the higher SiO2 melts, and up to 17 for the lowest SiO2 melt. The melts we've generated from our mica-pyroxenite assemblage can contribute to the explanation of a range of alkaline magmas while also having significant metasomatic potential at the point of melt generation.