Dresden 2020 – scientific programme

The DPG Spring Meeting in Dresden had to be cancelled! Read more ...

Parts | Days | Selection | Search | Updates | Downloads | Help

O: Fachverband Oberflächenphysik

O 124: Development of Novel Methods II

O 124.6: Talk

Friday, March 20, 2020, 11:45–12:00, WIL C107

Developing descriptors for the prediction of adsorption energies on metal oxides — •Wenbin Xu, Mie Andersen, and Karsten Reuter — Chair for Theoretical Chemistry and Catalysis Research Center, Technical University of Munich, Garching, Germany

Linear scaling relations of adsorption energy arise at transition metal (TM) catalysts as a consequence of the adsorbate valency, together with the properties of the d electrons of the surface. Already, this linearity is only approximate, with outlier adsorption energies easily deviating from the trend by up to 1 eV. At other materials classes, deviations can be even more significant [1], calling for improved approaches that yield reliable adsorption energies at still comparably low computational cost.

To this end, we have recently demonstrated the usefulness of a compressed sensing approach for TM and TM alloy catalysts [2]. The corresponding SISSO (sure independence screening and sparsifying operator) approach [3] allows the prediction of adsorption energies from descriptors that are expressed as nonlinear functions of intrinsic properties of the clean catalyst surface (so-called primary features), e.g. coordination numbers, d-band moments, and work function. Here, we extend this approach to TM oxide catalysts and oxygen evolution catalysis. Next to establishing a systematic first-principles database for the SISSO training, a key aspect of our work is to identify primary features for this class of materials. [1] X. Hong, et al., ACS Catal., 6, 4428 (2016). [2] M. Andersen, et al., ACS Catal., 9, 2752 (2019). [3] R. Ouyang, et al., Phys. Rev. Mater., 2, 083802 (2018).