Materials 4.0

Phase diagrams are fundamental to materials development. They provide the conditions for phase stabilities and transformations, and thereby a thorough thermodynamic understanding of materials. However, today’s phase diagrams are often based on scarce experimental input and rely on extrapolations. Ab initio approaches can provide useful complementary information, but they have been limited to zero-Kelvin or low-temperature approximations which are not representative for phase diagrams. Within the Materials 4.0 project, we utilize our unique expertise to develop and apply efficient statistical sampling techniques that are rooted in high-accuracy finite-temperature ab initio methods and are accelerated by machine learning potentials. These techniques facilitate an efficient and accurate determination of Gibbs energies and a wide range of other thermodynamic quantities. In particular the Gibbs energies constitute the fundamental input to the calculation of phase diagrams. With Materials 4.0, we thus provide an important contribution to the design of next-generation materials.