Density functional theory(DFT)has helped propel the advance of electrocatalysis in the past two decades.In view of its massive use,it is worth asking how reliable DFT is for the prediction of adsorption energies,which...Density functional theory(DFT)has helped propel the advance of electrocatalysis in the past two decades.In view of its massive use,it is worth asking how reliable DFT is for the prediction of adsorption energies,which are paramount in computational electrocatalysis models.Here,we provide an experimental-computational approach to break down overall adsorption-energy errors into separate gas-phase and adsorbed-phase contributions.The method is evaluated using experimental data and various exchange-correlation functionals and materials for C-and O-containing species.Our main conclusion is that no functional is simultaneously accurate for adsorbates and molecules,as adsorbed-phase errors are visibly different from gas-phase errors.Importantly,total,gas-phase,and adsorbed-phase errors are correlated,revealing intrinsic DFT limitations and enabling the elaboration of swift correction routines.To illustrate the benefits of our approach,we deconvolute and correct all errors in CO_(2)electroreduction to CO and find an agreement with experiments close to chemical accuracy for numerous transition-metal electrodes and all scrutinized functionals.展开更多
Most AI-for-Materials research to date has focused on ideal crystals,whereas real-world materials inevitably contain defects that play a critical role in modern functional technologies.The defects break geometric symm...Most AI-for-Materials research to date has focused on ideal crystals,whereas real-world materials inevitably contain defects that play a critical role in modern functional technologies.The defects break geometric symmetry and increase interaction complexity,posing particular challenges for traditional ML models.Here,we introduce Defect-Informed Equivariant Graph Neural Network(DefiNet),a model specifically designed to accurately capture defect-related interactions and geometric configurations in point-defect structures.DefiNet achieves near-DFT-level structural predictions in milliseconds using a single GPU.To validate its accuracy,we perform DFT relaxations using DefiNet-predicted structures as initial configurations and measure the residual ionic steps.For most defect structures,regardless of defect complexity or system size,only 3 ionic steps are required to reach the DFT-level ground state.Finally,comparisons with scanning transmission electron microscopy(STEM)images confirm DefiNet’s scalability and extrapolation beyond point defects,positioning it as a valuable tool for defect-focused materials research.展开更多
The development of high-performance materials for microelectronics,energy storage,and extreme environments depends on our ability to describe and direct property-defining microstructural order.Our present understandin...The development of high-performance materials for microelectronics,energy storage,and extreme environments depends on our ability to describe and direct property-defining microstructural order.Our present understanding is typically derived from laborious manual analysis of imaging and spectroscopy data,which is difficult to scale,challenging to reproduce,and lacks the ability to reveal latent associations needed for mechanistic models.Here,we demonstrate a multi-modal machine learning(ML)approach to describe order from electron microscopy analysis of the complex oxide La_(1−x)Sr_(x)FeO_(3).We construct a hybrid pipeline based on fully and semi-supervised classification,allowing us to evaluate both the characteristics of each data modality and the value each modality adds to the ensemble.We observe distinct differences in the performance of uni-and multi-modal models,from which we draw general lessons in describing crystal order using computer vision.展开更多
基金financial support from MICIU/AEI/10.13039/501100011033by the European Union,and grant MOE-T2EP10222-0007 from the Ministry of Education,Singapore。
文摘Density functional theory(DFT)has helped propel the advance of electrocatalysis in the past two decades.In view of its massive use,it is worth asking how reliable DFT is for the prediction of adsorption energies,which are paramount in computational electrocatalysis models.Here,we provide an experimental-computational approach to break down overall adsorption-energy errors into separate gas-phase and adsorbed-phase contributions.The method is evaluated using experimental data and various exchange-correlation functionals and materials for C-and O-containing species.Our main conclusion is that no functional is simultaneously accurate for adsorbates and molecules,as adsorbed-phase errors are visibly different from gas-phase errors.Importantly,total,gas-phase,and adsorbed-phase errors are correlated,revealing intrinsic DFT limitations and enabling the elaboration of swift correction routines.To illustrate the benefits of our approach,we deconvolute and correct all errors in CO_(2)electroreduction to CO and find an agreement with experiments close to chemical accuracy for numerous transition-metal electrodes and all scrutinized functionals.
基金supported by the Natural Science Foundation of Guangdong Province (Grant No. 2025A1515011487)Ministry of Education, Singapore, Tier 1 (Grant No. A-8001194-00-00)+1 种基金Tier 2 (Grant No. A-8001872-00-00)lunder its Research Center of Excellence award to the Institute for Functional Intelligent Materials (I-FIM, project No. EDUNC-33-18-279-V12)K.S.N. is grateful to the Royal Society (UK, grant number RSRP\R\190000) for support.
文摘Most AI-for-Materials research to date has focused on ideal crystals,whereas real-world materials inevitably contain defects that play a critical role in modern functional technologies.The defects break geometric symmetry and increase interaction complexity,posing particular challenges for traditional ML models.Here,we introduce Defect-Informed Equivariant Graph Neural Network(DefiNet),a model specifically designed to accurately capture defect-related interactions and geometric configurations in point-defect structures.DefiNet achieves near-DFT-level structural predictions in milliseconds using a single GPU.To validate its accuracy,we perform DFT relaxations using DefiNet-predicted structures as initial configurations and measure the residual ionic steps.For most defect structures,regardless of defect complexity or system size,only 3 ionic steps are required to reach the DFT-level ground state.Finally,comparisons with scanning transmission electron microscopy(STEM)images confirm DefiNet’s scalability and extrapolation beyond point defects,positioning it as a valuable tool for defect-focused materials research.
基金supported by the Laboratory Directed Research and Development (LDRD) program at Pacific Northwest National Laboratory (PNNL). PNNL is a multiprogram national laboratory operated for the U.S. Department of Energy (DOE) by Battelle Memorial Institute under Contract No. DE-AC05-76RL0-1830Some sample preparation was performed at the Environmental Molecular Sciences Laboratory (EMSL), a national scientific user facility sponsored by the Department of Energy's Office of Biological and Environmental Research and located at PNNL. Ion irradiation work was performed, in part, at the Center for Integrated Nanotechnologies, an Office of Science User Facility operated for the U.S. Department of Energy (DOE) Office of Science by Los Alamos National Laboratory (Contract 89233218CNA000001) and Sandia National Laboratories (Contract DE-NA-0003525)+1 种基金This work was authored in part by the National Renewable Energy Laboratory (NREL) for the U.S. Department of Energy (DOE) under Contract No. DE-AC36-08GO28308. The views expressed in the presentation do not necessarily represent the views of the DOE or the U.S. Government. The U.S. Government retains and the publisher, by accepting the article for publication, acknowledges that the U.S. Government retains a nonexclusive, paid-up, irrevocable, worldwide license to publish or reproduce the published form of this work, or allow others to do so, for U.S. Government purposes. R.P. and R.B.C. gratefully acknowledge funding support for film synthesis from the National Science Foundation under award DMR-1809847R.B.C. also acknowledges funding support for data science and machine learning efforts from the National Science Foundation under award DMR-2045993.
文摘The development of high-performance materials for microelectronics,energy storage,and extreme environments depends on our ability to describe and direct property-defining microstructural order.Our present understanding is typically derived from laborious manual analysis of imaging and spectroscopy data,which is difficult to scale,challenging to reproduce,and lacks the ability to reveal latent associations needed for mechanistic models.Here,we demonstrate a multi-modal machine learning(ML)approach to describe order from electron microscopy analysis of the complex oxide La_(1−x)Sr_(x)FeO_(3).We construct a hybrid pipeline based on fully and semi-supervised classification,allowing us to evaluate both the characteristics of each data modality and the value each modality adds to the ensemble.We observe distinct differences in the performance of uni-and multi-modal models,from which we draw general lessons in describing crystal order using computer vision.
