Development of a Model for Predicting the Adsorption Performance of Zeolites and Designing New Zeolites

Climate change is currently one of the most serious environmental problems. The main cause of climate change is carbon dioxide, which accounts for approximately 80% of all anthropogenic greenhouse gases. The development of technology to separate, recover, store, and reuse carbon dioxide is required. In this study, we focused on carbon dioxide separation technology for flue gas through the physical adsorption method using zeolites. The amount of carbon dioxide adsorbed by zeolites varies depending on the structural parameters, such as the Si/Al ratio and the loaded cations. We used two adsorption isotherms, Langmuir and Freundlich, and set two adsorption parameters for each, and used machine learning to predict the logarithm of the adsorption parameters, q_max and K for the Langmuir equation and n and a for the Freundlich equation, from structural information on zeolite obtained from the literature. Then, using this model, we searched for the characteristics of zeolites with higher carbon dioxide adsorption capacity than zeolites obtained from the literature based on the structural information on zeolites not used in the model construction and were able to find zeolites with higher carbon dioxide adsorption capacity than existing zeolites.