Next-generation fission and fusion reactors impose unprecedented demands on structural materials, requiring simultaneous resistance to high temperatures, high-dose irradiation, and aggressive corrosion. Designing materials that harness the intrinsic properties of multiple elements and their synergistic interactions has emerged as a key strategy to achieve such integrated performance. To guide this design paradigm, a mechanistic understanding of chemically and structurally complex systems is essential. However, such understanding is currently constrained by the lack of high-fidelity interatomic potentials (IAPs) that enable predictive, large-scale atomistic simulations. Here, we employ, for the first time, a multi-task, physics-informed pretraining strategy with the large atomic model (LAM) to systematically evaluate the construction and predictive capability of IAPs for complex nuclear alloy systems. Using Ta-Nb-W-Mo-V as a representative case, the resulting DPA2-5E model—trained solely on the quinary dataset—significantly outperforms conventional machine learning IAPs, demonstrates superior transferability to lower-order subsystems, and accurately reproduces cascade damage and stress-strain behavior. Furthermore, this approach extends to nuclear-relevant structures and corrosive/oxide environments, enabling high-fidelity IAPs and large-scale simulations at reactor extremes.
{"title":"Scaling reliable interatomic potentials to complex nuclear alloys via pretrained atomic models","authors":"Mingxuan Jiang, Biao Xu, Yixin Deng, Shihua Ma, Ji-Jung Kai, Fei Gao, Huiqiu Deng","doi":"10.1038/s41524-025-01950-6","DOIUrl":"https://doi.org/10.1038/s41524-025-01950-6","url":null,"abstract":"Next-generation fission and fusion reactors impose unprecedented demands on structural materials, requiring simultaneous resistance to high temperatures, high-dose irradiation, and aggressive corrosion. Designing materials that harness the intrinsic properties of multiple elements and their synergistic interactions has emerged as a key strategy to achieve such integrated performance. To guide this design paradigm, a mechanistic understanding of chemically and structurally complex systems is essential. However, such understanding is currently constrained by the lack of high-fidelity interatomic potentials (IAPs) that enable predictive, large-scale atomistic simulations. Here, we employ, for the first time, a multi-task, physics-informed pretraining strategy with the large atomic model (LAM) to systematically evaluate the construction and predictive capability of IAPs for complex nuclear alloy systems. Using Ta-Nb-W-Mo-V as a representative case, the resulting DPA2-5E model—trained solely on the quinary dataset—significantly outperforms conventional machine learning IAPs, demonstrates superior transferability to lower-order subsystems, and accurately reproduces cascade damage and stress-strain behavior. Furthermore, this approach extends to nuclear-relevant structures and corrosive/oxide environments, enabling high-fidelity IAPs and large-scale simulations at reactor extremes.","PeriodicalId":19342,"journal":{"name":"npj Computational Materials","volume":"85 1","pages":""},"PeriodicalIF":9.7,"publicationDate":"2026-01-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145938277","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"材料科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2026-01-09DOI: 10.1038/s41524-025-01863-4
Rebecca K. Lindsey, Awwal D. Oladipupo, Sorin Bastea, Bradley A. Steele, I-Feng W. Kuo, Nir Goldman
{"title":"Hierarchical transfer learning: an agile and equitable strategy for machine-learning interatomic models","authors":"Rebecca K. Lindsey, Awwal D. Oladipupo, Sorin Bastea, Bradley A. Steele, I-Feng W. Kuo, Nir Goldman","doi":"10.1038/s41524-025-01863-4","DOIUrl":"https://doi.org/10.1038/s41524-025-01863-4","url":null,"abstract":"","PeriodicalId":19342,"journal":{"name":"npj Computational Materials","volume":"82 1","pages":""},"PeriodicalIF":9.7,"publicationDate":"2026-01-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145938275","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"材料科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2026-01-09DOI: 10.1038/s41524-025-01944-4
Jiaxuan Li, Nikita Rybin, Taowei Wang, Alexander Shapeev, Xiaotong Chen, Bing Liu
{"title":"Revealing the diffusion mechanism of Cs in amorphous and polycrystalline SiC by actively trained moment tensor potentials","authors":"Jiaxuan Li, Nikita Rybin, Taowei Wang, Alexander Shapeev, Xiaotong Chen, Bing Liu","doi":"10.1038/s41524-025-01944-4","DOIUrl":"https://doi.org/10.1038/s41524-025-01944-4","url":null,"abstract":"","PeriodicalId":19342,"journal":{"name":"npj Computational Materials","volume":"252 1","pages":""},"PeriodicalIF":9.7,"publicationDate":"2026-01-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145938274","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"材料科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2026-01-09DOI: 10.1038/s41524-025-01868-z
Shenda Jiang, Israel Greenfeld, Lin Yang, Weilong Yin, Xiaodong He, H. Daniel Wagner
Two-dimensional materials (2DMs), possessing atomic-scale thickness, are prone to brittle fracture under loading conditions, which can lead to catastrophic failure. As their structural dimensions approach the nanoscale, conventional linear elastic fracture mechanics (LEFM) based on continuum assumptions is deficient in capturing the underlying failure mechanisms and accurately predicting potential crack instability. This limitation emphasizes the critical need for a new theoretical approach suited to the fracture behavior of 2DM systems. We propose a unified fracture mechanics (UFM) criterion that systematically incorporates two key physical mechanisms governing brittle fracture in 2DMs at the nanoscale, namely nonlinear elasticity and atomic-scale discreteness. By introducing two corrective parameters, for nonlinearity and quantization, the UFM model successfully resolves the limitations of LEFM in predicting failure. This is particularly important in the short crack regime, as small defects are frequent in 2DMs. The theoretical predictions show excellent agreement with molecular dynamics simulations of five different types of 2DMs and accurately capture the fracture strength of both cracked and defect-free structures. In addition, we present an empirical method that allows the fracture behavior of 2DMs to be estimated directly from their intrinsic structural and elastic properties. The unified theoretical framework is applicable not only to the materials simulated in this study but may also be applied to a broader class of atomically thin brittle systems.
{"title":"Unified fracture criterion for brittle 2D materials","authors":"Shenda Jiang, Israel Greenfeld, Lin Yang, Weilong Yin, Xiaodong He, H. Daniel Wagner","doi":"10.1038/s41524-025-01868-z","DOIUrl":"https://doi.org/10.1038/s41524-025-01868-z","url":null,"abstract":"Two-dimensional materials (2DMs), possessing atomic-scale thickness, are prone to brittle fracture under loading conditions, which can lead to catastrophic failure. As their structural dimensions approach the nanoscale, conventional linear elastic fracture mechanics (LEFM) based on continuum assumptions is deficient in capturing the underlying failure mechanisms and accurately predicting potential crack instability. This limitation emphasizes the critical need for a new theoretical approach suited to the fracture behavior of 2DM systems. We propose a unified fracture mechanics (UFM) criterion that systematically incorporates two key physical mechanisms governing brittle fracture in 2DMs at the nanoscale, namely nonlinear elasticity and atomic-scale discreteness. By introducing two corrective parameters, for nonlinearity and quantization, the UFM model successfully resolves the limitations of LEFM in predicting failure. This is particularly important in the short crack regime, as small defects are frequent in 2DMs. The theoretical predictions show excellent agreement with molecular dynamics simulations of five different types of 2DMs and accurately capture the fracture strength of both cracked and defect-free structures. In addition, we present an empirical method that allows the fracture behavior of 2DMs to be estimated directly from their intrinsic structural and elastic properties. The unified theoretical framework is applicable not only to the materials simulated in this study but may also be applied to a broader class of atomically thin brittle systems.","PeriodicalId":19342,"journal":{"name":"npj Computational Materials","volume":"57 1","pages":""},"PeriodicalIF":9.7,"publicationDate":"2026-01-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145938276","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"材料科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2026-01-08DOI: 10.1038/s41524-025-01915-9
Youheng Chen, Jiajia Han, Chen Yang, Cuiping Wang, Xingjun Liu
The stacking fault energy (SFE) in concentrated alloys is highly sensitive to the local atomic environment. Understanding this relationship necessitates extensive sampling of configurations, which significantly increases computational demands. To address this challenge, we propose a graph neural network (GNN)-based framework utilizing unrelaxed bulk and stacking fault structures as inputs to directly predict SFE. We first investigate two key extrapolation capabilities for bulk formation energy prediction: scale extrapolation (predicting formation energy in larger supercells) and compositional extrapolation (predicting formation energy for compositions beyond the training set). Leveraging the validated model, we concurrently predict formation energies of both bulk and stacking fault configurations to compute the SFE. The structural similarity between these configurations enables efficient parameter sharing, accelerating model convergence. The framework demonstrates excellent interpretability and robust compositional extrapolation capabilities in predicting SFEs. Furthermore, leveraging its exceptional compositional extrapolation, we integrate the model with Monte Carlo simulations to successfully predict ordering behavior and solute segregation at stacking faults. Finally, we introduce a hierarchical training strategy that further reduces data requirements. Collectively, our work establishes a unified and efficient framework for robust prediction of planar fault energies in complex concentrated alloys.
{"title":"A crystal graph convolutional neural network framework for predicting stacking fault energy in concentrated alloys","authors":"Youheng Chen, Jiajia Han, Chen Yang, Cuiping Wang, Xingjun Liu","doi":"10.1038/s41524-025-01915-9","DOIUrl":"https://doi.org/10.1038/s41524-025-01915-9","url":null,"abstract":"The stacking fault energy (SFE) in concentrated alloys is highly sensitive to the local atomic environment. Understanding this relationship necessitates extensive sampling of configurations, which significantly increases computational demands. To address this challenge, we propose a graph neural network (GNN)-based framework utilizing unrelaxed bulk and stacking fault structures as inputs to directly predict SFE. We first investigate two key extrapolation capabilities for bulk formation energy prediction: scale extrapolation (predicting formation energy in larger supercells) and compositional extrapolation (predicting formation energy for compositions beyond the training set). Leveraging the validated model, we concurrently predict formation energies of both bulk and stacking fault configurations to compute the SFE. The structural similarity between these configurations enables efficient parameter sharing, accelerating model convergence. The framework demonstrates excellent interpretability and robust compositional extrapolation capabilities in predicting SFEs. Furthermore, leveraging its exceptional compositional extrapolation, we integrate the model with Monte Carlo simulations to successfully predict ordering behavior and solute segregation at stacking faults. Finally, we introduce a hierarchical training strategy that further reduces data requirements. Collectively, our work establishes a unified and efficient framework for robust prediction of planar fault energies in complex concentrated alloys.","PeriodicalId":19342,"journal":{"name":"npj Computational Materials","volume":"3 1","pages":""},"PeriodicalIF":9.7,"publicationDate":"2026-01-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145919889","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"材料科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2026-01-07DOI: 10.1038/s41524-025-01927-5
Anvesh Nathani, Arthur RC McCray, Yingtao Liu, Hanping Ding, Pejman Kazempoor, Shuozhi Xu, Colin Ophus, Iman Ghamarian
Electron diffraction(ED) often used to solve for unknown structures or refine existing ones. Existing methods for automated ED analysis often struggle with challenges such as computational expense and experimental noise. This study introduces a deep learning framework to accelerate and improve crystal structure determination from diffraction patterns. The methodology treats each diffraction pattern as a relational graph of Bragg spots. Spot features are encoded using a 1D convolutional network, from which a relational attention aggregator constructs an orientation-agnostic graph. This graph is processed by a Graphormer encoder enhanced with Mixture-of-Experts layers, allowing the model to learn complex crystallographic relationships efficiently. Trained and tested on a large dataset of simulated diffraction patterns, the model achieved a crystal system classification accuracy of 89.2% and a space group accuracy of 70.2% from single patterns, significantly outperforming a state-of-the-art random forest baseline (74.2% and 57.8%, respectively). By aggregating predictions across multiple zone axes, these accuracies improved to 96.5% and 79.5%. The model also demonstrated robust performance on experimental data of gold nanoparticles, producing plausible classifications consistent with known orientation degeneracies. By unifying relational graph reasoning with specialized expert networks, this work presents a robust and automated framework for high-throughput materials characterization.
{"title":"Accelerating electron diffraction analysis using graph neural networks and attention mechanisms","authors":"Anvesh Nathani, Arthur RC McCray, Yingtao Liu, Hanping Ding, Pejman Kazempoor, Shuozhi Xu, Colin Ophus, Iman Ghamarian","doi":"10.1038/s41524-025-01927-5","DOIUrl":"https://doi.org/10.1038/s41524-025-01927-5","url":null,"abstract":"Electron diffraction(ED) often used to solve for unknown structures or refine existing ones. Existing methods for automated ED analysis often struggle with challenges such as computational expense and experimental noise. This study introduces a deep learning framework to accelerate and improve crystal structure determination from diffraction patterns. The methodology treats each diffraction pattern as a relational graph of Bragg spots. Spot features are encoded using a 1D convolutional network, from which a relational attention aggregator constructs an orientation-agnostic graph. This graph is processed by a Graphormer encoder enhanced with Mixture-of-Experts layers, allowing the model to learn complex crystallographic relationships efficiently. Trained and tested on a large dataset of simulated diffraction patterns, the model achieved a crystal system classification accuracy of 89.2% and a space group accuracy of 70.2% from single patterns, significantly outperforming a state-of-the-art random forest baseline (74.2% and 57.8%, respectively). By aggregating predictions across multiple zone axes, these accuracies improved to 96.5% and 79.5%. The model also demonstrated robust performance on experimental data of gold nanoparticles, producing plausible classifications consistent with known orientation degeneracies. By unifying relational graph reasoning with specialized expert networks, this work presents a robust and automated framework for high-throughput materials characterization.","PeriodicalId":19342,"journal":{"name":"npj Computational Materials","volume":"391 1","pages":""},"PeriodicalIF":9.7,"publicationDate":"2026-01-07","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145908188","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"材料科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2026-01-07DOI: 10.1038/s41524-025-01923-9
Daye Zheng, Xingliang Peng, Yike Huang, Yinan Wang, Duo Zhang, Zhengtao Huang, Zefeng Cai, Linfeng Zhang, Mohan Chen, Ben Xu, Weiqing Zhou
We propose a non-collinear spin-constrained method that generates training data for deep-learning-based magnetic model, which provides a powerful tool for studying complex magnetic phenomena that requires large-scale simulations at the atomic level. First, we propose a basis-independent projection method for calculating atomic magnetic moments by applying a radial truncation to numerical atomic orbitals. A double-loop Lagrange multiplier method is utilized to ensure the satisfaction of constraint conditions while achieving accurate magnetic torque. The method is implemented in ABACUS with both plane wave basis and numerical atomic orbital basis. We benchmark the iron (Fe) systems and analyze differences from calculations with the plane wave basis and numerical atomic orbitals basis in describing magnetic energy barriers. Based on an automated workflow composed of first-principles calculations, magnetic model, active learning, and dynamics simulation, more than 30,000 first-principles data with the information of magnetic torque are generated to train a deep-learning-based magnetic model DeePSPIN for the Fe system. By utilizing the model in large-scale molecular dynamics simulations, we successfully predict Curie temperatures of α-Fe close to experimental values.
{"title":"Integrating deep-learning-based magnetic model and non-collinear spin-constrained method: methodology, implementation and application","authors":"Daye Zheng, Xingliang Peng, Yike Huang, Yinan Wang, Duo Zhang, Zhengtao Huang, Zefeng Cai, Linfeng Zhang, Mohan Chen, Ben Xu, Weiqing Zhou","doi":"10.1038/s41524-025-01923-9","DOIUrl":"https://doi.org/10.1038/s41524-025-01923-9","url":null,"abstract":"We propose a non-collinear spin-constrained method that generates training data for deep-learning-based magnetic model, which provides a powerful tool for studying complex magnetic phenomena that requires large-scale simulations at the atomic level. First, we propose a basis-independent projection method for calculating atomic magnetic moments by applying a radial truncation to numerical atomic orbitals. A double-loop Lagrange multiplier method is utilized to ensure the satisfaction of constraint conditions while achieving accurate magnetic torque. The method is implemented in ABACUS with both plane wave basis and numerical atomic orbital basis. We benchmark the iron (Fe) systems and analyze differences from calculations with the plane wave basis and numerical atomic orbitals basis in describing magnetic energy barriers. Based on an automated workflow composed of first-principles calculations, magnetic model, active learning, and dynamics simulation, more than 30,000 first-principles data with the information of magnetic torque are generated to train a deep-learning-based magnetic model DeePSPIN for the Fe system. By utilizing the model in large-scale molecular dynamics simulations, we successfully predict Curie temperatures of α-Fe close to experimental values.","PeriodicalId":19342,"journal":{"name":"npj Computational Materials","volume":"28 1","pages":""},"PeriodicalIF":9.7,"publicationDate":"2026-01-07","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145908460","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"材料科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Suppressing critical current density ( Jc ) fluctuations in Josephson junctions is essential for improving the reproducibility and scalability of superconducting quantum processors. Despite many elucidations of microscopic mechanisms, the physical modulation of Jc by atomic-scale disorder at the metal-insulator interface remains elusive. Here, we reveal that interfacial bonding topology distortions are the dominant source that regulates Jc uniformity. We identify a new disorder metric, Interface Bonding Topology Factor (IBTF), that captures bond-angle fluctuations and oxygen-coordination heterogeneity within Jc variations. Through multivariate analysis, Jc is exponentially correlated with interface disorder and barrier thickness ( d ) by Jc ∝ e −IBTF⋅ d , explaining 91.88% of the observed Jc inhomogeneity. We establish IBTF as a tunable physical degree of freedom whose suppression efficacy enhances significantly with increasing d , and demonstrate its active modulation by twin boundary engineering in electrodes. This work provides a device-oriented strategy and a tunable physical metric beyond single-feature control for scalable high-performance quantum processors.
{"title":"Revealing the role of interface disorder in modulating critical current density of Josephson junctions","authors":"Chuanbing Han, Huihui Sun, Yonglong Shen, Junling Qiu, Peng Xu, Fudong Liu, Bo Zhao, Xiaohan Yu, Weilong Wang, Shuya Wang, Qing Mu, Benzheng Yuan, Lixin Wang, Chaofeng Hou, Zheng Shan","doi":"10.1038/s41524-025-01941-7","DOIUrl":"https://doi.org/10.1038/s41524-025-01941-7","url":null,"abstract":"Suppressing critical current density ( <jats:italic>J</jats:italic> <jats:sub>c</jats:sub> ) fluctuations in Josephson junctions is essential for improving the reproducibility and scalability of superconducting quantum processors. Despite many elucidations of microscopic mechanisms, the physical modulation of <jats:italic>J</jats:italic> <jats:sub>c</jats:sub> by atomic-scale disorder at the metal-insulator interface remains elusive. Here, we reveal that interfacial bonding topology distortions are the dominant source that regulates <jats:italic>J</jats:italic> <jats:sub>c</jats:sub> uniformity. We identify a new disorder metric, Interface Bonding Topology Factor (IBTF), that captures bond-angle fluctuations and oxygen-coordination heterogeneity within <jats:italic>J</jats:italic> <jats:sub>c</jats:sub> variations. Through multivariate analysis, <jats:italic>J</jats:italic> <jats:sub>c</jats:sub> is exponentially correlated with interface disorder and barrier thickness ( <jats:italic>d</jats:italic> ) by <jats:italic>J</jats:italic> <jats:sub>c</jats:sub> ∝ <jats:italic>e</jats:italic> <jats:sup> −IBTF⋅ <jats:italic>d</jats:italic> </jats:sup> , explaining 91.88% of the observed <jats:italic>J</jats:italic> <jats:sub>c</jats:sub> inhomogeneity. We establish IBTF as a tunable physical degree of freedom whose suppression efficacy enhances significantly with increasing <jats:italic>d</jats:italic> , and demonstrate its active modulation by twin boundary engineering in electrodes. This work provides a device-oriented strategy and a tunable physical metric beyond single-feature control for scalable high-performance quantum processors.","PeriodicalId":19342,"journal":{"name":"npj Computational Materials","volume":"31 1","pages":""},"PeriodicalIF":9.7,"publicationDate":"2026-01-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145903107","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"材料科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：