Towards a more practical two-feldspar geothermometer

Abstract

The thermodynamic basis of several recent attempts to formulate a simple two-feldspar geothermometer is discussed, together with a review of earlier empirical geothermometers and ones based on experimental studies in the ternary feldspar system. It is shown that double-binary thermometers which involve the combination of regular solution mixing models for the binary alkali feldspar system with ideal mixing in plagioclases do not give a satisfactory representation of two-feldspar relations, especially for albite-rich compositions where a critical point exists. Thermometers based on mixing parameters for ordered alkali feldspar frameworks are even more unjustified both because low-plagioclases are certainly non-ideal, and because of uncertainty in knowing the degree of Al-Si order in the alkali feldspar when exchange equilibrium was achieved. A ‘thermodynamic’ thermometer requires knowledge of ternary activities which are at present unknown. Experimental determinations of relationships in the ternary feldspar system are reviewed and the correct general form of the thermometer constructed using mainly the experimental data of Seck (1971a) and Smith and Parsons (1974). Chemographic tests for equilibrium between feldspar pairs are suggested and petrographie features discussed. In an appendix new values are given of Margules parameters calculated for binary disordered alkali feldspars from recent solvus data up to 15 kbars, and their physico-chemical basis examined. We suggest that accurate representations of the mixing properties of disordered alkali feldspars using Margules parameters are at present premature.