Machine learning-assisted high-throughput computational screening of MOFs and advances in gas separation research

Metal organic frameworks (MOFs) have garnered extensive research interest in fields such as gas storage, adsorption separation, and catalysis due to their high surface area, large pore volume, and tunable structures. In recent years, the surge in the number of MOFs has posed unprecedented challenges in finding the ideal MOF for specific applications. In this scenario, high-throughput computational screening (HTCS) has become the most effective research method for screening high-performance target MOFs from a vast array of materials. HTCS will generate a large amount of multidimensional data, which can be further used for machine learning (ML) training. Recently, applying ML to HTCS of MOFs has become a new hotspot, which can not only reveal the potential structure-performance relationships of materials but also provide insights into their performance trends in different applications, especially in gas storage and separation. In this review, we highlight the latest advances in ML-assisted HTCS in the field of MOFs gas separation, systematically analyze the internal mechanism of ML and HTCS collaboration to improve screening efficiency in the search for high-performance MOFs, and explore the opportunities and challenges presented in this new field. © 2025 Materials China. All rights reserved.