How to choose solution models for phase equilibria modeling of migmatitic metapelites?

English Abstract:

Phase equilibria modelling is a widely used tool to investigate the evolution of metamorphic rocks. One approach consists in fixing the bulk chemical composition and calculating the mineral phase assemblages by minimizing the Gibb's free energy. Because mineral phases share chemical elements and compete with each other's, the choice of mineral solution models is of uttermost importance. Despite their wide use, few studies have tested the reliability of this technique by reproducing well-controlled cases, such as experimental studies and none have investigated whether a set of solution models performs better than others. In this study, we investigate these issues by calculating phase equilibria in the Na-Ca-K-Mn-Fe-Mg-Al-Si-H-Ti-O chemical system for three partial-melting experiments using different set of solutions models. These experiments span a range of pelitic bulk compositions and were conducted between 500 and 1000 MPa, as well as between 750 to 1050°C. The software Perple_X is used because it is the only one offering total flexibility for the choice of models. The stability field, compositions, and proportions of the various mineral phases for each calculation are compared with experimental results. We focus on biotite, garnet and plagioclase, for which the compositional isopleths are most commonly used for thermobarometry conducted with phase equilibria modeling. We show that it is possible to reproduce experimental results with relatively good accuracy (discrepancies generally varying by 20 to 50°C for a given pressure), notably for the biotite-in, the melt-in and Ti-oxides equilibria However, a specific set of solution models is required, and it is not necessarily composed of the models generally used.