Comprehensive overview of machine learning applications in MOFs: from modeling processes to latest applications and design classifications

Abstract

As an emerging class of nanoporous materials, metal–organic frameworks (MOFs) have the advantages of designability and structural and functional tunability, compared with traditional porous materials, which are widely used in various fields. The structural adjustability of MOFs provides the possibility of infinite material generation and a huge material space. At present, tens of thousands of MOFs have been synthesized and the number continues to grow at an alarming rate, which makes it difficult to explore the application prospects of all materials only by traditional experimental methods. Therefore, more efficient alternative methods are urgently needed to identify and screen MOFs. As a powerful data analysis tool, machine learning (ML) has shown great potential in the materials field, which can intuitively and quickly analyze the structure–property relationship and guide the rational design and preparation of reticular materials such as MOFs. This review systematically presents the complete workflow and cutting-edge developments in ML applications in the field of MOF research covering data preparation, algorithm selection, model evaluation, model optimization and application status. Further, rational design methods and future challenges are discussed. This review aims to provide the new paradigm of the combination of ML and MOFs and promote ML applied in MOF research efficiently.