Ok Yong-sik’s Group Developed the World’s First AI-Based Model for Predicting Carbon Dioxide Adsorption on Biomass and Waste-Derived Porous Carbons

The results were published in the supplementary cover article of Environmental Science and Technology, the world’s most authoritative journal in the field.
           

 

Professor Ok Yong-sik’s group in the Division of Environmental Science & Ecological Engineering, College of Life Sciences and Biotechnology, published an article entitled ‘Applied Machine Learning for Prediction of CO2 Adsorption on Biomass Waste-Derived Porous Carbons’ in the supplementary cover article of the world’s most authoritative journal in the field, Environmental Science and Technology (ES&T) published by the American Chemical Society. 

ES&T, one of the journals included in the Nature Index, is the most authoritative of all environmental journals covering environmental sciences, engineering and policy. 

Recently, in the field of carbon neutrality more and more attention is being paid to the value-added materials acquired by the upcycling of biomass and waste (e.g., biochar, porous carbons). In particular, using biomass and waste-based porous carbons for capturing carbon dioxide, the main culprit of global warming, is a central carbon capture and storage (CSS) technology. The technology is spotlighted as a sustainable technology that can address issues involving biomass, including byproducts from the agricultural and forestry industries, such as livestock excretions, as well as the environmental pollution caused by the inappropriate management of various types of waste, including waste plastics. However, since both biomass and waste-based porous carbons have different structural characteristics and diverse functional groups, their CO2 adsorption, particularly as affected by temperature and pressure, is difficult to accurately predict. 

Professor Ok’s group collaborated with an AI research team, including Professor Xiaonan Wang from National University of Singapore, to develop a machine learning-based algorithm by extracting 632 data sets from 76 articles published globally over the last decade. The developed machine learning-based algorithm showed excellent predictive power, and its capabilities were verified through an additional verification process based on independent data. 

Professor Ok said, “Using the algorithm, we are looking forward to predicting more accurately the CO2 adsorption by the porous carbon materials derived from various types of waste, and additionally providing the adsorption mechanisms.” He also explained, “Carbon-negative biomass-based materials, such as biochar, may be of great help in accomplishing the UN’s Sustainable Development Goals.” Professor Ok added, “In pursuit of commercialization, we will conduct further studies applying our model to the porous carbons generated by a real-world mass production process.”