Catalytic materials and chemistry development using a synergistic combination of machine learning and ab initio methods

Abstract

First principles-based molecular modelling plays a crucial role in the development of novel catalytic materials and in the investigation of catalytic chemical reactions. However, the computational cost and/or the accuracy of these models remains a bottleneck in carrying out these simulations for complex or large scale systems, as in the case of catalysis. Over the past two decades, machine learning (ML) has made an impact in the field of computational catalysis. Modern-day researchers have started using machine learning-based data-driven techniques to overcome the limitations of these molecular simulations. In this review, we summarize the recent progress in the utilization of ML algorithms to assist molecular simulations, followed by its applications in the field of catalysis. Furthermore, we provide our perspective on promising avenues for research in the future regarding the incorporation of ML in molecular simulations in catalysis.