Accelerated discovery of porous materials for carbon capture by machine learning: A review

In the past decades, greenhouse gases (e.g., anthropogenic CO 2 and CH 4 ) have raised significant concerns due to the foreseeable dire consequences in climate change. Capturing them via adsorption using porous materials has drawn much attention due to their low synthesis and regeneration cost and high capacity. Recently, the flourishing machine learning (ML) has been introduced to various fields of materials science, which also has shown great potential in accelerating the materials discovery for carbon capture. In this article, we first describe the general workflow of applying ML to tackle materials problems. Then we systematically summarize the recent research progress in the application of ML for development of porous carbon and metal–organic frameworks for carbon capture. Finally, we discuss the existing challenges, possible solutions, and research directions. This article will inspire exploration of new frontiers in the carbon capture by development of ML in porous materials research in the future. Graphical abstract