Introduction

Date	Time	Title	Speaker
Dec. 6th	15:00 - 15:30	Machine Learning approach in designing catalytic alloy nanoparticle	Prof. YongJoo Kim

Abstract

Searching for an optimal component and composition of multi-metallic alloy catalysts, comprising two or more elements, is one of the key issues in catalysis research. Due to the exhaustive data requirement of conventional machine learning (ML) models and the high cost of experimental trials, current approaches rely mainly on the combination of density functional theory (DFT) and ML techniques. In this study, a significant step was taken toward overcoming limitations by the interplay of experiment and active learning that can effectively search for an optimal component and composition of multi-metallic alloy catalysts. The active learning model suggests process parameters that are likely to yield high catalytic performance, and the model is iteratively updated using fabricated solid-solution nanoparticles by the carbothermal shock (CTS) method by examining their electro-catalytic performance for the hydrogen evolution reaction (HER). An optimal metal precursor composition of Pt0.65Ru0.30Ni0.05 is located, which exhibits an HER overpotential of 54.2 mV; this HER potential is superior to that of a pure Pt catalyst (68.2 mV). This result indicates the success of the constructed model by only utilizing the precursor mixture composition as input data, thereby improving the overpotential by searching for an optimal component and composition inside the system. This method appears to be widely applicable since we were able to use it to determine an optimal component and composition of electrocatalyst for HER with no obvious restriction to the types of catalysts to which it can be applied.