Berlin 2018 – wissenschaftliches Programm
Bereiche | Tage | Auswahl | Suche | Aktualisierungen | Downloads | Hilfe
MM: Fachverband Metall- und Materialphysik
MM 51: Topical Session (Symposium MM): Big Data in Materials Science - Managing and exploiting the raw material of the 21st century
MM 51.3: Vortrag
Donnerstag, 15. März 2018, 10:45–11:00, H 0107
Machine Learning Structural Descriptors on Nanocluster Catalysts — •Marc Jäger1, Eiaki Morooka1, and Adam Foster1,2 — 1Aalto University, Helsinki, Finland — 2Kanazawa University, Japan
Scientists have advanced significantly in producing nanoparticles with defined composition, size and morphology in the last decade. Due to this and because of their remarkable properties, nanoclusters have gained attention in heterogeneous catalysis. Nanoclusters differ from bulk metal behaviour, their catalytic properties are sensitive to changes in size and morphology. Nanoparticles like molybdenum disulfide are known to catalyze the hydrogen evolution reaction (HER). The combinatorial and structural space of nanoclusters is vast, so extensive modelling is difficult. Structural descriptors are used to describe the geometry of an adsorption site and to predict properties which indicate a high catalytic activity, in particular the hydrogen adsorption free energy. We analysed the performance of state-of-the-art structural descriptors (SOAP, MBTR and ACSF). Simulations can provide energetic and kinetic analysis of HER using DFT. The vast amount of possible nanoclusters, all potential candidates for catalysing the HER, requires reduction and interpolation of DFT calculations. This is tackled by merging the combinatorial space with the chemical compound space and applying machine learning on diverse datasets.