AI-Guided Design and Property Prediction for Zeolites and Nanoporous Materials

4 Generating, Managing, and Mining Big Data in Zeolite Simulations

Daniel Schwalbe-Koda and Rafael Gómez-Bombarelli

Department of Materials Science and Engineering, Massachusetts Institute of Technology, Massachusetts, USA

4.1 Introduction

Accelerated materials innovation is imperative to address global challenges in energy, the environment, or healthcare [1,2]. Mitigating climate change, for example, requires a fast decarbonization of the economy, including hard-to-decarbonize sectors such as industry or transportation [3]. Enabling this transition requires discovering catalysts for sustainable chemical transformations [4]. However, catalyst discovery has long relied on trial-and-error. Finding stable and affordable materials that can catalyze given chemical reactions with high activity/selectivity requires exploring a combinatorial number of structures and compositions, often aggravated by the high-dimensional problem of synthesizing the materials themselves. As such, strategies to accelerate the design of novel catalysts are urgently required.

Zeolites are among the catalysts of interest for sustainable applications. Their mature use in the petrochemical industry makes them interesting candidates for a variety of reaction pathways involving thermocatalysis [5,6], especially because these materials are known to have a large topological diversity [7,8]. Unfortunately, zeolites are no exception to the trial-and-error discovery efforts typically seen in the field of catalysis. ...