To mitigate the climate crisis, it is essential to actively reduce the carbon dioxide that has already been emitted. For this purpose, technology that directly captures carbon dioxide from the air (Direct Air Capture, DAC) is gaining attention. Domestic researchers successfully identified promising carbon capture candidate materials among metal-organic frameworks (MOFs), which are the core materials of DAC technology, using artificial intelligence (AI)-based machine learning technology.
Professor Kim Ji-han and his research team at the Korea Advanced Institute of Science and Technology (KAIST) announced on the 29th that they developed a machine learning-based simulation technique capable of quickly and accurately selecting MOFs suitable for capturing carbon dioxide from the atmosphere through collaborative research with researchers from Imperial College London.
The research team developed a machine learning-based simulation that can precisely predict the interactions between MOFs, carbon dioxide, and water, enabling the calculation of the adsorption properties of MOF materials at a speed significantly faster than before while maintaining quantum-level prediction accuracy.
Using the developed system, they conducted a large-scale screening of over 8,000 experimentally synthesized MOF structures, discovering more than 100 promising carbon capture candidate materials. In particular, they proposed new candidate materials that had not been identified by existing classical simulations and also suggested seven key chemical features useful for the design of DAC materials by analyzing the correlation between the chemical structure of MOFs and their adsorption performance.
The research team explained, "The simulation technique proposed in this study improves both precision and computational efficiency compared to existing methods, significantly reducing the time and resources required for discovering high-performance DAC materials," and noted, "This can be expanded not only to MOFs but also to various porous materials and adsorption systems."
The results of this study were published in the international journal 'Matter' on June 12.
References
Matter(2025), DOI: https://doi.org/10.1016/j.matt.2025.102203