Abstract

Recent advances in thermodynamic modelling now allow phase equilibrium studies of partial melting in mafic and ultramafic sources under upper mantle conditions. This potentially offers several advantages over experimental investigations, including the fact that capsule-sample interactions are avoided, there is no uncertainty over the volatile content of the charge, the process is much faster, and open system behaviour can be investigated with relative ease. Two different thermodynamic datasets are in common use for phase equilibrium investigations of mantle rocks, compiled alongside a range of constantly evolving phase solution models. This study compares the results of published experimental studies on partial melting in the upper mantle with phase equilibrium modelling results from the popular pMELTS software (Ghiorso et al., 2002), which uses an embedded dataset from Berman (1988), and with results from the Rcrust software (Mayne et al., 2016), which uses the dataset of Holland & Powell (2011). The aim of this comparative study is to investigate the suitability of each combination of software specific calculation routines, thermodynamic datasets and phase solution models to provide a viable alternative to conventional experimental techniques when studying magma genesis from the upper mantle. The main points of interest are how the phase assemblages, proportions, and compositions predicted by the modelling compare with those produced in the experiments. The set of experimental results used for comparison includes both peridotitic and eclogitic compositions, investigated over a pressure range of 1.0–3.0 GPa and a temperature range of 1165–1950 °C. The results indicate that using the Holland & Powell (2011) dataset and appropriately selected phase solution models provides the best match with experimental results in terms of melt volume, melt composition and the assemblage of coexisting phases produced during partial melting under upper mantle conditions.

Full Text
Published version (Free)

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call