Advancing materials design by high-accuracy finite-temperature first principles calculations accelerated by machine learning potentials

Projektbeschreibung

Maschinelles Lernen für die Entwicklung fortgeschrittener Werkstoffe

Phasendiagramme zeigen Phasenbeziehungen im thermodynamischen Gleichgewichtszustand der Materie. Die meisten Phasendiagramme basieren allerdings auf knappen experimentellen Daten und beruhen häufig auf gewagten Hochrechnungen. Im Rahmen des EU-finanzierten Projekts Materials 4.0 wird an einer Lösung gearbeitet. Es wird eine hochpräzise thermodynamische Datenbank der ersten Prinzipien bereitstellen, um die Gestaltung von Werkstoffen auf die nächste Stufe zu heben. Da Ansätze der ersten Prinzipien, alias ab initio, keine experimentellen Daten erfordern, können sie dort eingesetzt werden, wo kein Experiment möglich ist. Sie wurden jedoch auf Annäherungen an niedrige Temperaturen beschränkt, die für Phasendiagramme nicht ideal sind. Das Projektteam wird Werkstoffe in Betracht ziehen, die für aktuelle wissenschaftliche Entwicklungen und das technologische Interesse relevant sind, wie beispielsweise Hybrid- und Superlegierungen.

Ziel

Phase diagrams have revolutionized materials development by providing the conditions for phase stabilities and transformations, and thereby a thorough thermodynamic understanding of materials design. However, the majority of today’s phase diagrams are based on scarce experimental input and often rely on daring extrapolations. Every multicomponent phase diagram relies on a fragile set of phase stabilities as very recent studies show.

Materials 4.0 will change this. It will raise materials design to the next level by providing a highly accurate first principles thermodynamic database. First principles, alias ab initio, approaches do not require any experimental input and can operate where no experiment is able to reach. However, they have been limited to zero Kelvin or low temperature approximations which are not representative of phase diagrams.

Materials 4.0 reaches far beyond this by utilizing my unique expertise in high-accuracy finite-temperature ab initio simulations. We will develop novel methods accelerated by machine learning potentials that facilitate a highly efficient determination of Gibbs free energies and migration barriers including all relevant finite-temperature excitation mechanisms. The methodology will be implemented in an easy-to-use open-source integrated development environment and made accessible to the community.

Materials 4.0 will consider materials relevant to current scientific developments and of technological interest, such as hydrides, lightweight alloys, superalloys, MAX phases, and high entropy alloys. A large ab initio thermodynamic database will be computed for elements across the periodic table. The main focus will be on phase stabilities of various phases, including dynamically unstable ones, and importantly liquids as well; all fully from ab initio. The phase stabilities will be put into practice by re-parametrizing binary phase diagrams and studying the implications on multicomponent phase diagrams.