Discovery of novel reticular materials for carbon dioxide capture using GFlowNets

• Published in: Digital discovery Vol. 3; no. 3; pp. 449 - 455
• Main Authors: Cipcigan, Flaviu, Booth, Jonathan, Barros Ferreira, Rodrigo Neumann, Ribeiro dos Santos, Carine, Steiner, Mathias
• Format: Journal Article
• Language: English
• Published: 13.03.2024
• ISSN: 2635-098X; 2635-098X
• DOI: 10.1039/d4dd00020j

Artificial intelligence holds promise to improve materials discovery. GFlowNets are an emerging deep learning algorithm with many applications in AI-assisted discovery. Using GFlowNets, we generate porous reticular materials, such as Metal Organic Frameworks and Covalent Organic Frameworks, for applications in carbon dioxide capture. We introduce a new Python package (matgfn) to train and sample GFlowNets. We use matgfn to generate the matgfn-rm dataset of novel and diverse reticular materials with gravimetric surface area above 5000 m 2 g −1 . We calculate single- and two-component gas adsorption isotherms for the top-100 candidates in matgfn-rm. These candidates are novel compared to the state-of-art ARC-MOF dataset and rank in the 90th percentile in terms of working capacity compared to the CoRE2019 dataset. We identify 13 materials with CO 2 working capacity outperforming all materials in CoRE2019. After further analysis and structural relaxation, two outperforming materials remain. GFlowNets discover reticular materials with simulated CO 2 working capacity outperforming all materials in CoRE2019.