Berlin 2024 – scientific programme

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

MM: Fachverband Metall- und Materialphysik

MM 63: Functional Materials: Performance, Reliability and Degradation

MM 63.1: Talk

Thursday, March 21, 2024, 15:30–15:45, C 230

Combining DFTB and Structure Mapping for the Prediction of Transition Paths in the Deactivation of ZnO@Cu Catalysts — •Artem Samtsevych, Chiara Panosetti, Karsten Reuter, and Christoph Scheurer — Fritz-Haber-Institut der MPG, Berlin

Solid-solid transformations are common in the aging of functional materials like catalysts. Understanding these transformations at the atomistic level is thus crucial for a resilient design. In practice, this requires identifying minimum energy pathways between basins on a complex free energy surface. While chain-of-state methods help obtain corresponding pathways, they are generally challenged by the exponential growth of the number of possible transition pathways with system size and the computational cost of the underlying first-principles, typically density-functional theory (DFT), energy evaluations.

Here we address both challenges by employing geometry- and topology-based mapping techniques for the efficient generation of suitable initial transition pathways and a machine learning-based optimization of density-functional tight binding (DFTB). The former techniques map the atomic structures and unit cells or the graphs of interatomic bonds of the connected basins. The latter optimization of the DFTB repulsive potential [1] establishes this technique as a computationally efficient surrogate for the DFT energetics. We illustrate the combined general workflow by studying the aging process in ZnO@Cu catalysts, which involves the transformation of the ZnO overlayer from a graphitic-like to wurtzitic structure.

[1] C. Panosetti et al., J. Chem. Theory Comput. 16, 21818 (2020).

Keywords: Solid-solid transformations; Density functional tight binding; Machine learning; Catalysis