Machine Learning‐Driven Prediction and Design of High‐Performance Metal‐Organic Frameworks for CO₂ Capture

Metal‐organic frameworks (MOFs) has been extensively employed in CO 2 capture because of high Brunauer–Emmett–Teller (BET) surface area, large pore volume, modifiability, and designability. The high diversity of metal clusters and organic ligands offers both hope and challenges for the design of novel MOFs with excellent CO 2 adsorption performance. Here we propose an automated nanoporous materials discovery method, powered by machine learning and genetic algorithm, for the generative design of MOFs. We demonstrate the automated design process of MOFs, with the goal of efficiently capturing CO 2 . The method provides a theoretical study for the design of MOFs with the ability to selectively capture CO 2 .