Realizing the Data-Driven, Computational Discovery of Metal-Organic Framework Catalysts

TL;DR

This paper reviews recent advances and challenges in applying data science and machine learning to accelerate the discovery and design of metal-organic framework catalysts, highlighting future directions for high-throughput screening.

Contribution

It provides a comprehensive overview of current work and discusses challenges and solutions for integrating data-driven methods into MOF catalyst discovery.

Findings

Recent progress in computational screening of MOF catalysts

Identification of key challenges in data-centric MOF catalysis

Proposed solutions to accelerate adoption of machine learning in the field

Abstract

Metal-organic frameworks (MOFs) have been widely investigated for challenging catalytic transformations due to their well-defined structures and high degree of synthetic tunability. These features, at least in principle, make MOFs ideally suited for a computational approach towards catalyst design and discovery. Nonetheless, the widespread use of data science and machine learning to accelerate the discovery of MOF catalysts has yet to be substantially realized. In this review, we provide an overview of recent work that sets the stage for future high-throughput computational screening and machine learning studies involving MOF catalysts. This is followed by a discussion of several challenges currently facing the broad adoption of data-centric approaches in MOF computational catalysis, and we share possible solutions that can help propel the field forward.