Bereiche | Tage | Auswahl | Suche | Aktualisierungen | Downloads | Hilfe

MM: Fachverband Metall- und Materialphysik

MM 20: Data Driven Material Science: Big Data and Workflows III

MM 20.7: Vortrag

Dienstag, 19. März 2024, 12:00–12:15, C 243

Exploring high-entropy alloy transport properties through the lens of machine learning — •Ruiwen Xie¹, Ye Wei², Bo Peng³, Jiamu Liu³, Liuliu Han⁴, and Hongbin Zhang¹ — ¹Group of Theory of Magnetic Materials, Technical University of Darmstadt, Darmstadt, Germany — ²École Polytechnique Fédérale de Lausanne (EPFL), Switzerland — ³State Key Laboratory of Tribology in Advanced Equipment, Tsinghua University, Beijing, China — ⁴Department Microstructure Physics and Alloy Design, Max-Planck-Institut für Eisenforschung GmbH

The high-entropy alloys (HEAs), or the compositionally complexed alloys (CCAs), have attracted much attention due to their multifunctional properties with a vast chemical space to explore. For instance, the five-component HEAs contain approximately 4.6 million compositional combinations with a 1 at.% interval. Therefore, efficient sampling methods to navigate the chemical space for optimized properties are needed. As a showcase, we establish a workflow by combining the multi-objective Bayesian optimisation (MOBO) and active learning (AL), in order to explore the Ta-Nb-Hf-Zr-Ti system for compositions with optimal spin Hall conductivities and spin Hall angles. Additionally, a Monte Carlo beam search based AL algorithm is used to explore FeCoNi-based HEAs targeting for high saturation magnetization, high anomalous Hall conductivity and low electrical conductivity simultaneously.