Prediction, determination and validation of phase diagrams via the global study of energy landscapes

Abstract

Abstract Traditionally, the determination of phase diagrams has followed the inductive paradigm, where experimental observations provide the phase boundaries in more or less detail and phenomenological and semi-phenomenological models are employed to interpolate between the experimental data points, and by extrapolation to predict the shape of the phase boundaries in experimentally inaccessible regions. Over the past fifteen years, a new methodology has been developing, the aim of which is the prediction, determination and validation of phase diagrams in chemical systems without any recourse to experimental information. The founding stone of this deductive approach is the global study of the energy landscape of the chemical system. In this review, we present an introduction to the concept of energy landscapes in the context of phase diagram calculations, an overview over the large variety of methods employed to study energy landscapes and compute free energies on the ab initio level, and a number of typical examples of first-principle predictions, determinations and validations of phase diagrams.