Application of machine learning in adsorption energy storage using metal organic frameworks: A review

Tackling the issues posed by climate change and the need to reduce greenhouse gas emissions has led to the development of novel adsorbent materials tailored for clean energy solutions. Advancements in machine learning (ML) have enabled significant progress in identifying, designing, and optimizing materials with enhanced efficiency and economic viability for clean energy applications. This review provides an overview of key ML techniques and their applications in the development of robust adsorbent materials, with particular emphasis on thermal adsorption energy storage. We examine recent progress in using ML models to predict the adsorption capacities of various gases, including HQ, CH4, COQ, and biogas, in metal-organic frameworks (MOFs). Quantitatively, ML models have achieved prediction accuracies with mean absolute errors as low as 0.1 mmol/g for hydrogen adsorption and 0.05 mmol/g for COQ adsorption in select MOFs. Pivotal case studies demonstrate how ML has expedited the performance enhancement, stability prediction, and material identification processes for MOFs, with a comparison drawn between the adsorption performance of MOFs and zeolites, where MOFs have shown up to 50 % higher gas uptake in some cases. Finally, we discuss challenges such as limited high-quality data and algorithmic complexity, while highlighting future opportunities for integrating ML with MOFs to improve adsorption energy storage. This review offers critical insights for advancing ML-assisted MOF research in clean adsorption energy applications.