Regensburg 2022 – wissenschaftliches Programm
Bereiche | Tage | Auswahl | Suche | Aktualisierungen | Downloads | Hilfe
MM: Fachverband Metall- und Materialphysik
MM 25: Data Driven Materials Science: Computational Frameworks / Chemical Complexity
MM 25.1: Topical Talk
Mittwoch, 7. September 2022, 15:45–16:15, H46
Automated atomistic calculation of thermodynamic and thermophysical data — •Jan Janssen1,2, Tilmann Hickel1,3, and Jörg Neugebauer1 — 1Max-Planck-Institut für Eisenforschung, Düsseldorf, Germany — 2Los Alamos National Laboratory, Los Alamos, USA — 3Bundesanstalt für Materialforschung und -prüfung (BAM), Berlin, Germany
A major challenge in predicting the properties of materials at realistic conditions is the accurate inclusion of finite temperature effects. Doing this on an ab initio level often requires complex simulation protocols. These complex protocols, which often couple several specialized codes, make a quantitative description of error propagation and uncertainty quantification a critical issue.
To handle this high level of complexity we have developed an integrated development environment (IDE) called pyiron - http://pyiron.org. pyiron has been specifically designed to scale simulation protocols from the interactive prototyping level up to the high throughput level, all within the same software framework.
We highlight two recent success stories towards automated calculation of phase diagrams: We first discuss with the automated convergence for all key parameters in DFT codes, followed by the calculation of melting points with a guaranteed precision of better than 1K. These fully automated high-precision tools allow us to study trends over the periodic table in an efficient and systematic way. Examples how such high-throughput screenings allow to develop new strategies in designing materials will be given.