Simulation of nickel surfaces using ab-initio and empirical methods

Abstract

Nickel is the catalyst of choice in state-of-the-art solid oxide fuels cells (SOFC) and constitutes the key component in commonly used anodes. However, the anode microstructure is known to change significantly under SOFC operation, leading to a deterioration of the device’s electrochemical performance. In order to assess the relevant degradation mechanisms and make qualitative lifetime predictions, a simulative multiscale approach is developed within the publicly funded project KerSOLife100. In this work, we describe the determination of the relevant nickel surface properties, surface and adsorption energies, on the atomistic level using ab-initio methods based on density functional theory (DFT). For the first time, we comprehensively compare the applicability of three exchange-correlation functionals to computationally significantly less expensive methods based on a selection of empirical potentials. While we found all three exchange-correlation functionals to be in qualitative agreement and within the experimental scatter, PBEsol yields the most accurate nickel surface energies. Our analysis and determined values should be helpful for future research on nickel surfaces. Even though one of the empirical potentials gives weighted surface energies within 10% accuracy, the DFT surface anisotropies are not reproduced by any used potential. While nickel adsorption energies on nickel surfaces predicted by the empirical potential are in qualitative agreement with our PBEsol results, hydrogen adsorption energies show even qualitative disagreement. Our assessment of the considered methods opens the door for further atomistic simulations of the anode properties. The derived results are transferred to the next-scale models to simulate the evolution of the anode microstruture.