First‐Principles Density Functional Theory and Machine Learning Technique for the Prediction of Water Adsorption Site on PtPd‐Based High‐Entropy‐Alloy Catalysts

Abstract

AbstractThe water‐gas shift reaction (WGSR) is employed in industry to obtain high‐purity H2 from syngas, where H2O adsorption is an important step that controls H2O dissociation in WGSR. Therefore, exploring catalysts exhibiting strong H2O adsorption energy (Eads) is crucial. Also, high‐entropy alloys (HEA) are promising materials utilized as catalysts, including in WGSR. The PtPd‐based HEA catalysts are explored via density functional theory (DFT) and Gaussian process regression. The input features are based on the microstructure data and electronic properties: d‐band center (εd) and Bader net atomic charge (δ). The DFT calculation reveals that the εd and δ of each active site of all HEA surfaces are broadly scattered, indicating that the electronic properties of each atom on HEA are non‐uniform and influenced by neighboring atoms. The strong H2O‐active‐site interaction determined by a highly negative Eads is used as a criterion to explore good PtPd‐based WGSR catalyst candidates. As a result, the potential candidates are found to have Co, Ru, and Fe as an H2O adsorption site with Ag as a neighboring atom, that is, PtPdRhAgCo, PtPdRuAgCo, PtPdRhAgFe, and PtPdRuAgFe.