An empirical model for plagioclase equilibrium in hydrous melts

Abstract

Heat capacity models for anorthite (An) and albite (Ab) crystal and supercooled liquid, together with the assumption of ideal mixing of these components were used to derive average values of enthalpy and volume of melting from phase equilibrium data that are in significant disagreement with some recently published thermodynamic data. In an effort to find a means of calculating both liquidus temperature and crystal composition of plagioclase for a given hydrous melt composition, the activity models for feldspar components in hydrous melts and solubility calculations suggested by Burnham (1975) and the enthalpy data above were tested by comparing predicted and observed liquidi. These assumptions lead to satisfactory agreement in the systems Ab-An-H 2O and Ab-Si 4O 8-H 2O but liquidi in the system An-Si 4O 8-H 2O and complex systems differ radically from those calculated. For hydrous complex melts an empirical model using the above solubility and activity assumptions was fit to experimental data on coexisting melt and plagioclase compositions. Despite the demonstrable theoretical limitations of the assumptions involved, this empirical model apparently balances inaccuracies and reproduces the original data with absolute mean errors for 66 experiments of 17°C and 5 mol% An. It is sufficiently precise for use in kinetic crystallization models and may be useful as a geothermometer in some applications; it is probably not sufficiently accurate to be used as a geobarometer.