ALBERT

Description: The refined structure (Vieillard & Tardy, 1988) and polyhedral (Chermak & Rimstidt, 1989, 1990; Holland, 1989; Robinson & Haas, 1983) models offer the most realistic approaches to predicting unknown thermodynamic properties (ΔHf, ΔGf, S°, Cp) of various minerals. However, routine application of the former is currently limited by the lack of model parameters for most silicate minerals. On the other hand, the polyhedral approach is potentially more powerful than alternative ΔG°f and ΔH°f estimation algorithms inasmuch as Pauling's rules for silicate minerals underlie this concept. Nonetheless, free energies of formation estimated with current polyhedral models do not yet permit accurate prediction of phase equilibria. The repeated clustering of minerals into different distortion fields (Robinson et al, 1971; Fleet, 1976; this study) as well as the separation of apparently homologous group of polyhedra into different distortion-polarization energy fields (this study) suggest assumptions that nearest neighbour ionic interactions can account for the overall features of mineral structures are too simplistic. Inclusion of polyhedral distortions and measures of covalent contributions improve estimates of thermodynamic properties based on the polyhedral approach. However, further development of the proposed estimation algorithm is limited by gaps in the thermochemical database and thus emphasizes the need for continued experimental determinations of thermodynamic properties for structurally-coherent mineral groups.