SKM 2021 – scientific programme
Parts | Days | Selection | Search | Updates | Downloads | Help
MM: Fachverband Metall- und Materialphysik
MM 7: Topical Session Interface-Dominated Phenomena - Segregation and Embrittlement
MM 7.1: Invited Talk
Wednesday, September 29, 2021, 11:15–11:45, H8
Computational methods for grain boundary segregation in metallic alloys — •Lorenz Romaner1, Daniel Scheiber2, Vsevolod Razumovskiy2, Oleg Peil2, Christoph Dösinger1, and Alexander Reichmann1 — 1Department Materials Science, Montanuniversität Leoben, A-8700 Leoben — 2Materials Center Leoben Forschung GmbH, A 8700 Leoben
Modeling of grain boundary segregation phenomena is an important discipline of integrated computational materials design. Several computational methods, including in particular atomistic, thermokinetic or mechanical models are available to model grain boundary excess and to assess the associated material properties. Segregation energies plays a central role in this connection and large databases are being created to get a comprehensive overview over materials. With the availability of such databases, machine learning approaches can be used to learn the trends in the periodic table and get segregation energies even for alloys for which no data exist at present. We present an investigation on machine learning segregation energies obtained from density functional theory simulations. We will discuss the critical role of feature engineering and analyze different physical parameters including cohesive energies, solution energies, geometry of the segregation site and many more. Furthermore, we show results for a variety of metallic alloys focusing on the class of transition metals and on comparison with experiment. Finally, the challenges of machine learning of segregation energies and grain boundary engineering in general will be discussed.