Direct prediction of gas adsorption via spatial atom interaction learning
Abstract Physisorption relying on crystalline porous materials offers prospective avenues for sustainable separation processes, greenhouse gas capture, and energy storage. However, the lack of end-to-end deep learning model for adsorption prediction confines the rapid and precise screen of crystalli...
| Main Authors: | , , , , , , , , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Nature Portfolio
2023-11-01
|
| Series: | Nature Communications |
| Online Access: | https://doi.org/10.1038/s41467-023-42863-6 |
| _version_ | 1797636808382611456 |
|---|---|
| author | Jiyu Cui Fang Wu Wen Zhang Lifeng Yang Jianbo Hu Yin Fang Peng Ye Qiang Zhang Xian Suo Yiming Mo Xili Cui Huajun Chen Huabin Xing |
| author_facet | Jiyu Cui Fang Wu Wen Zhang Lifeng Yang Jianbo Hu Yin Fang Peng Ye Qiang Zhang Xian Suo Yiming Mo Xili Cui Huajun Chen Huabin Xing |
| author_sort | Jiyu Cui |
| collection | DOAJ |
| description | Abstract Physisorption relying on crystalline porous materials offers prospective avenues for sustainable separation processes, greenhouse gas capture, and energy storage. However, the lack of end-to-end deep learning model for adsorption prediction confines the rapid and precise screen of crystalline porous materials. Here, we present DeepSorption, a spatial atom interaction learning network that realizes accurate, fast, and direct structure-adsorption prediction with only information of atomic coordinate and chemical element types. The breakthrough in prediction is attributed to the awareness of global structure and local spatial atom interactions endowed by the developed Matformer, which provides the intuitive visualization of atomic-level thinking and executing trajectory in crystalline porous materials prediction. Complete adsorption curves prediction could be performed using DeepSorption with a higher accuracy than Grand canonical Monte Carlo simulation and other machine learning models, a 20-35% decline in the mean absolute error compared to graph neural network CGCNN and machine learning models based on descriptors. Since the established direct associations between raw structure and target functions are based on the understanding of the fundamental chemistry of interatomic interactions, the deep learning network is rationally universal in predicting the different physicochemical properties of various crystalline materials. |
| first_indexed | 2024-03-11T12:39:28Z |
| format | Article |
| id | doaj.art-3c549d11d97c40929d86c58c8bc4a33b |
| institution | Directory Open Access Journal |
| issn | 2041-1723 |
| language | English |
| last_indexed | 2024-03-11T12:39:28Z |
| publishDate | 2023-11-01 |
| publisher | Nature Portfolio |
| record_format | Article |
| series | Nature Communications |
| spelling | doaj.art-3c549d11d97c40929d86c58c8bc4a33b2023-11-05T12:22:47ZengNature PortfolioNature Communications2041-17232023-11-011411910.1038/s41467-023-42863-6Direct prediction of gas adsorption via spatial atom interaction learningJiyu Cui0Fang Wu1Wen Zhang2Lifeng Yang3Jianbo Hu4Yin Fang5Peng Ye6Qiang Zhang7Xian Suo8Yiming Mo9Xili Cui10Huajun Chen11Huabin Xing12Key Laboratory of Biomass Chemical Engineering of Ministry of Education, College of Chemical and Biological Engineering, Zhejiang UniversityCollege of Computer Science and Technology, Zhejiang UniversityCollege of Computer Science and Technology, Zhejiang UniversityKey Laboratory of Biomass Chemical Engineering of Ministry of Education, College of Chemical and Biological Engineering, Zhejiang UniversityKey Laboratory of Biomass Chemical Engineering of Ministry of Education, College of Chemical and Biological Engineering, Zhejiang UniversityCollege of Computer Science and Technology, Zhejiang UniversityCollege of Computer Science and Technology, Zhejiang UniversityCollege of Computer Science and Technology, Zhejiang UniversityKey Laboratory of Biomass Chemical Engineering of Ministry of Education, College of Chemical and Biological Engineering, Zhejiang UniversityKey Laboratory of Biomass Chemical Engineering of Ministry of Education, College of Chemical and Biological Engineering, Zhejiang UniversityKey Laboratory of Biomass Chemical Engineering of Ministry of Education, College of Chemical and Biological Engineering, Zhejiang UniversityCollege of Computer Science and Technology, Zhejiang UniversityKey Laboratory of Biomass Chemical Engineering of Ministry of Education, College of Chemical and Biological Engineering, Zhejiang UniversityAbstract Physisorption relying on crystalline porous materials offers prospective avenues for sustainable separation processes, greenhouse gas capture, and energy storage. However, the lack of end-to-end deep learning model for adsorption prediction confines the rapid and precise screen of crystalline porous materials. Here, we present DeepSorption, a spatial atom interaction learning network that realizes accurate, fast, and direct structure-adsorption prediction with only information of atomic coordinate and chemical element types. The breakthrough in prediction is attributed to the awareness of global structure and local spatial atom interactions endowed by the developed Matformer, which provides the intuitive visualization of atomic-level thinking and executing trajectory in crystalline porous materials prediction. Complete adsorption curves prediction could be performed using DeepSorption with a higher accuracy than Grand canonical Monte Carlo simulation and other machine learning models, a 20-35% decline in the mean absolute error compared to graph neural network CGCNN and machine learning models based on descriptors. Since the established direct associations between raw structure and target functions are based on the understanding of the fundamental chemistry of interatomic interactions, the deep learning network is rationally universal in predicting the different physicochemical properties of various crystalline materials.https://doi.org/10.1038/s41467-023-42863-6 |
| spellingShingle | Jiyu Cui Fang Wu Wen Zhang Lifeng Yang Jianbo Hu Yin Fang Peng Ye Qiang Zhang Xian Suo Yiming Mo Xili Cui Huajun Chen Huabin Xing Direct prediction of gas adsorption via spatial atom interaction learning Nature Communications |
| title | Direct prediction of gas adsorption via spatial atom interaction learning |
| title_full | Direct prediction of gas adsorption via spatial atom interaction learning |
| title_fullStr | Direct prediction of gas adsorption via spatial atom interaction learning |
| title_full_unstemmed | Direct prediction of gas adsorption via spatial atom interaction learning |
| title_short | Direct prediction of gas adsorption via spatial atom interaction learning |
| title_sort | direct prediction of gas adsorption via spatial atom interaction learning |
| url | https://doi.org/10.1038/s41467-023-42863-6 |
| work_keys_str_mv | AT jiyucui directpredictionofgasadsorptionviaspatialatominteractionlearning AT fangwu directpredictionofgasadsorptionviaspatialatominteractionlearning AT wenzhang directpredictionofgasadsorptionviaspatialatominteractionlearning AT lifengyang directpredictionofgasadsorptionviaspatialatominteractionlearning AT jianbohu directpredictionofgasadsorptionviaspatialatominteractionlearning AT yinfang directpredictionofgasadsorptionviaspatialatominteractionlearning AT pengye directpredictionofgasadsorptionviaspatialatominteractionlearning AT qiangzhang directpredictionofgasadsorptionviaspatialatominteractionlearning AT xiansuo directpredictionofgasadsorptionviaspatialatominteractionlearning AT yimingmo directpredictionofgasadsorptionviaspatialatominteractionlearning AT xilicui directpredictionofgasadsorptionviaspatialatominteractionlearning AT huajunchen directpredictionofgasadsorptionviaspatialatominteractionlearning AT huabinxing directpredictionofgasadsorptionviaspatialatominteractionlearning |