Mixing properties of phengitic micas and revised garnet‐phengite thermobarometers

Mixing properties for muscovite–celadonite–ferroceladonite solid solutions are derived from combining available experimental phase equilibrium data with tabulated thermodynamic data for mineral end‐members. When a partially ordered solution model is assumed, the enthalpy of mixing among the end‐members muscovite–celadonite–ferroceladonite is nearly ideal, although the Gibbs energies of muscovite–celadonite and muscovite–ferroceladonite solutions are asymmetric due to an asymmetry in the entropy of mixing. Thermodynamic consistency is achieved for data on phengite compositions inassemblages with (a) pyrope+kyanite+quartz/coesite (b) almandine+kyanite+quartz/coesite (c)talc+kyanite+quartz/coesite and (d) garnet–phengite pairs equilibrated both experimentally at high temperatures and natural pairs from low‐grade schists. The muscovite–paragonite solvus has been reanalysed using the asymmetric van Laar model, and the effects of the phengite substitution into muscovite have been quantitatively addressed in order to complete the simple thermodynamic mixing model for the solid solution among the mica end‐members. Results are applied to a natural pyrope–coesite–phengite–talc rock from the Western Alps, and to investigate the conditions under which biotite‐bearing mica schists transform to whiteschist‐like biotite‐absent assemblages for average pelite bulk compositions.