Perspectives on Shallow Mantle Melting from Thermodynamic Calculations

Understanding the relationship between mantle processes and formation of oceanic crust requires a method for modeling the composition and amount of melt produced as functions of temperature, pressure, and mantle composition. Empirical parameterizations constructed for this purpose have greatly enhanced our understanding of mantle melting, but they do not incorporate self-consistent relationships between phase equilibria and melting energetics. Here we describe a different approach, in which the melting process is calculated from thermodynamic models for minerals and silicate melts. These calculations represent the first attempt to use thermodynamic predictions to explore the connection between melting reactions and the energetic quantities (entropies, volumes, and heat capacities) that control mantle melting. This method holds particular promise for understanding aspects of mantle melting that cannot easily be inferred from empirical parameterization of phase equilibrium experiments (e.g. melt compositions at very low melt fraction, melt production during fractional fusion, and adiabatic paths for partially molten peridotite, with and without melt segregation.) Thermodynamic calculations of peridotite partial fusion have been performed using MELTS, which calculates silicate liquid-mineral equilibria by thermodynamic potential minimiTation (Ghiorso and Sack, 1994; Ghiorso et al., 1994). MELTS was developed to simulate low pressure crystall/zation of silicate liquids, but has been adapted to model peridotite melting at