In order to understand the influence of ordering behaviors on the thermodynamic and mechanical properties of multi-principal element alloys(MPEAs),the temperature-dependent thermodynamic properties and mechanical prop...In order to understand the influence of ordering behaviors on the thermodynamic and mechanical properties of multi-principal element alloys(MPEAs),the temperature-dependent thermodynamic properties and mechanical properties of FCC_CoNiV MPEAs were comparatively predicted,where the alloys were modeled as the ordered configurations based on our previously predicted site occupying fractions(SOFs),as well as disordered configuration based on traditional special quasi-random structure(SQS).The ordering behavior not only improves the thermodynamic stability of the structure,but also increases the elastic properties and Vickers hardness.For example,at 973 K,the predicted bulk modulus(B),shear modulus(G),Young’s modulus(E),and Vickers hardness(HV)of FCC_CoNiV MPEA based on SOFs configuration are 187.82,79.03,207.93,and 7.58 GPa,respectively,while the corresponded data are 172.58,57.45,155.14,and 4.64 GPa for the SQS configuration,respectively.The Vickers hardness predicted based on SOFs agrees considerably well with the available experimental data,while it is underestimated obviously based on SQS.展开更多
Computational materials science increasingly benefits from data management,automation,and algorithm-based decision-making for the simulation of material properties and behavior.Experimental materials science also chan...Computational materials science increasingly benefits from data management,automation,and algorithm-based decision-making for the simulation of material properties and behavior.Experimental materials science also changes rapidly by incorporation of‘machine learning’in materials discovery campaigns.The benefits including automation,reproducibility,data provenance,and reusability of managed data,however,are not widely available in the experimental domain.We present an implementation of an Active Learning loop with an interface to an experimental measurement device in pyiron as a demonstrator how to combine experimental and simulated data in one framework.Apart from the acceleration provided through active learning,additional acceleration of the experimental characterization is achieved by using prior knowledge from density functional theory simulations as well as predictions based on text mining using correlations in word embeddings.With data from all domains in the same framework,an untapped potential for the acceleration of materials characterization and materials discovery campaigns becomes available.展开更多
Recent advances in machine learning have demonstrated an enormous utility of deep learning approaches,particularly Graph Neural Networks(GNNs)for materials science.These methods have emerged as powerful tools for high...Recent advances in machine learning have demonstrated an enormous utility of deep learning approaches,particularly Graph Neural Networks(GNNs)for materials science.These methods have emerged as powerful tools for high-throughput prediction of material properties,offering a compelling enhancement and alternative to traditional first-principles calculations.While the community has predominantly focused on developing increasingly complex and universal models to enhance predictive accuracy,such approaches often lack physical interpretability and insights into materials behavior.Here,we introduce a novel computational paradigm—Self-Adaptable Graph Attention Networks integrated with Symbolic Regression(SA-GAT-SR)—that synergistically combines the predictive capability of GNNs with the interpretative power of symbolic regression.Our framework employs a self-adaptable encoding algorithm that automatically identifies and adjust attention weights so as to screen critical features from an expansive 180-dimensional feature space while maintainingO(n)computational scaling.The integratedSRmodule subsequently distills these features into compact analytical expressions that explicitly reveal quantum-mechanically meaningful relationships,achieving 23×acceleration compared to conventional SR implementations that heavily rely on first-principle calculations-derived features as input.This work suggests a new framework in computational materials science,bridging the gap between predictive accuracy and physical interpretability,offering valuable physical insights into material behavior.展开更多
Pretrained universal machine-learning interatomic potentials(MLIPs)have revolutionized computational materials science by enabling rapid atomistic simulations as efficient alternatives to ab initio methods.Fine-tuning...Pretrained universal machine-learning interatomic potentials(MLIPs)have revolutionized computational materials science by enabling rapid atomistic simulations as efficient alternatives to ab initio methods.Fine-tuning pretrained MLIPs offers a practical approach to improving accuracy for materials and properties where predictive performance is insufficient.However,this approach often induces catastrophic forgetting,undermining the generalizability that is a key advantage of pretrained MLIPs.Herein,we propose reEWC,an advanced fine-tuning strategy that integrates Experience Replay and Elastic Weight Consolidation(EWC)to effectively balance forgetting prevention with fine-tuning efficiency.Using Li_(6)PS_(5)Cl(LPSC),a sulfide-based Li solid-state electrolyte,as a fine-tuning target,we show that reEWC significantly improves the accuracy of a pretrained MLIP,resolving well-known issues of potential energy surface softening and overestimated Li diffusivities.Moreover,reEWC preserves the generalizability of the pretrained MLIP and enables knowledge transfer to chemically distinct systems,including other sulfide,oxide,nitride,and halide electrolytes.Compared to Experience Replay and EWC used individually,reEWC delivers clear synergistic benefits,mitigating their respective limitations while maintaining computational efficiency.These results establish reEWC as a robust and effective solution for continual learning in MLIPs,enabling universal models that can advance materials research through large-scale,high-throughput simulations across diverse chemistries.展开更多
Spin Hall effect(SHE)provides a promising solution to the realization of advantageous functionalities for spin-based recording and information processing.In this work,we conduct high-throughput calculations on the spi...Spin Hall effect(SHE)provides a promising solution to the realization of advantageous functionalities for spin-based recording and information processing.In this work,we conduct high-throughput calculations on the spin Hall conductivity(SHC)of antiperovskite compounds with the composition ZXM3,where Z is a nonmetal,X is a metal,and M is a platinum group metal.From an initial database over 4500 structures,we screen 295 structurally stable compounds and identify 24 compounds with intrinsic SHC exceeding 500(ℏ/e)(Ω^(⁻1)cm^(⁻1)).We reveal a strong dependence of SHC on spin-orbit coupling-induced energy splitting near the Fermi level.In addition,SHCs can be regulated through proper doping of electrons or holes.The present work establishes high-throughput database of SHC in antiperovskites which is crucial for designing future electric and spintronic devices.展开更多
It is urgent to establish a series of reasonable and general approaches to qualitatively and graphically characterize the four core effects of multi-principal element alloys(MPEAs)based on the inherent site preference...It is urgent to establish a series of reasonable and general approaches to qualitatively and graphically characterize the four core effects of multi-principal element alloys(MPEAs)based on the inherent site preference.In this contribution,a qualitatively and graphically characterizing approach to the diffusion behavior of interstitial nonmetallic atoms diffusing along the neighboring octahedra in MPEAs was explored intensively.For this purpose,the C atom diffusing along the neighboring octahedra in FCC_CoNiV MPEA with(V1.0000)1a(Co0.4445Ni0.4444V0.1111)3c,a constant ordered occupying configuration predicted in our previous paper,was demonstrated in detail.Six distinct diffusion paths along[110],[101],and[011]directions on XY,XZ,and YZ planes of FCC_CoNiV MPEA with forward and backward diffusion directions were explored one by one,respectively.The diffusion energy barrier,diffusion coefficient,diffusion constant,and activation energy were derived by employing first-principles calculations based on density functional theory alongside the Climbing Image Nudged Elastic Band method.Unlike diffusing behavior in pure metallic elements,the non-periodic diffusion energy barrier waves are revealed for the real FCC_CoNiV MPEA structure.The significant variations in the diffusion energy barriers are influenced by the atomic environment,particularly the interaction between V and C atoms,which enhances the localization of electrons and increases the overall diffusion energy barrier.The energy barriers show similar trends along six paths,but significant variations occur across different octahedral sites.展开更多
基金financially supported by the State Administration for Market Regulation,China(No.2021MK050)the National Natural Science Foundation of China(Nos.50971043,51171046,21973012)+3 种基金the Key Research and Development Program of China(Nos.2022YFB3807200,CISRI-21T62450ZD)the Natural Science Foundation of Fujian Province,China(Nos.2021J01590,2020J01351,2018J01754,2020J01474)the Student Research and Training Program(SRTP) of Fuzhou University,China(No.29320)Fujian Provincial Department of Science & Technology,China(No.2021H6011)。
文摘In order to understand the influence of ordering behaviors on the thermodynamic and mechanical properties of multi-principal element alloys(MPEAs),the temperature-dependent thermodynamic properties and mechanical properties of FCC_CoNiV MPEAs were comparatively predicted,where the alloys were modeled as the ordered configurations based on our previously predicted site occupying fractions(SOFs),as well as disordered configuration based on traditional special quasi-random structure(SQS).The ordering behavior not only improves the thermodynamic stability of the structure,but also increases the elastic properties and Vickers hardness.For example,at 973 K,the predicted bulk modulus(B),shear modulus(G),Young’s modulus(E),and Vickers hardness(HV)of FCC_CoNiV MPEA based on SOFs configuration are 187.82,79.03,207.93,and 7.58 GPa,respectively,while the corresponded data are 172.58,57.45,155.14,and 4.64 GPa for the SQS configuration,respectively.The Vickers hardness predicted based on SOFs agrees considerably well with the available experimental data,while it is underestimated obviously based on SQS.
基金funding from Deutsche Forschungsgemeinschaft(DFG)through project LU1175/26-1LZ and MS gratefully acknowledge the financial support provided by the China Scholarship Council(CSC number:202208360048)MS,LB,JN,and AL gratefully acknowledge funding by Deutsche Forschungsgemeinschaft(DFG)for CRC1625,project number 506711657,subprojects A01,A04,A05,INF.
文摘Computational materials science increasingly benefits from data management,automation,and algorithm-based decision-making for the simulation of material properties and behavior.Experimental materials science also changes rapidly by incorporation of‘machine learning’in materials discovery campaigns.The benefits including automation,reproducibility,data provenance,and reusability of managed data,however,are not widely available in the experimental domain.We present an implementation of an Active Learning loop with an interface to an experimental measurement device in pyiron as a demonstrator how to combine experimental and simulated data in one framework.Apart from the acceleration provided through active learning,additional acceleration of the experimental characterization is achieved by using prior knowledge from density functional theory simulations as well as predictions based on text mining using correlations in word embeddings.With data from all domains in the same framework,an untapped potential for the acceleration of materials characterization and materials discovery campaigns becomes available.
基金supported by National Natural Science Foundation of China (No.12374057)Fundamental Research Funds for the Central Universities. The work (S.T.) at Los Alamos National Laboratory (LANL) was performed at the Center for Integrated Nanotechnologies (CINT), a U.S. Department of Energy, Office of Science user facility at LANL.
文摘Recent advances in machine learning have demonstrated an enormous utility of deep learning approaches,particularly Graph Neural Networks(GNNs)for materials science.These methods have emerged as powerful tools for high-throughput prediction of material properties,offering a compelling enhancement and alternative to traditional first-principles calculations.While the community has predominantly focused on developing increasingly complex and universal models to enhance predictive accuracy,such approaches often lack physical interpretability and insights into materials behavior.Here,we introduce a novel computational paradigm—Self-Adaptable Graph Attention Networks integrated with Symbolic Regression(SA-GAT-SR)—that synergistically combines the predictive capability of GNNs with the interpretative power of symbolic regression.Our framework employs a self-adaptable encoding algorithm that automatically identifies and adjust attention weights so as to screen critical features from an expansive 180-dimensional feature space while maintainingO(n)computational scaling.The integratedSRmodule subsequently distills these features into compact analytical expressions that explicitly reveal quantum-mechanically meaningful relationships,achieving 23×acceleration compared to conventional SR implementations that heavily rely on first-principle calculations-derived features as input.This work suggests a new framework in computational materials science,bridging the gap between predictive accuracy and physical interpretability,offering valuable physical insights into material behavior.
基金supported by the Nano & Material Technology Development Programs through the National Research Foundation of Korea (NRF) funded by Ministry of Science and ICT (No. RS-2024-00407995 and No. RS-2024-00450102). The computations were carried out at Korea Institute of Science and Technology Information (KISTI) National Supercomputing Center (KSC-2025-CRE-0110) and at the Center for Advanced Computations (CAC) at Korea Institute for Advanced Study (KIAS).
文摘Pretrained universal machine-learning interatomic potentials(MLIPs)have revolutionized computational materials science by enabling rapid atomistic simulations as efficient alternatives to ab initio methods.Fine-tuning pretrained MLIPs offers a practical approach to improving accuracy for materials and properties where predictive performance is insufficient.However,this approach often induces catastrophic forgetting,undermining the generalizability that is a key advantage of pretrained MLIPs.Herein,we propose reEWC,an advanced fine-tuning strategy that integrates Experience Replay and Elastic Weight Consolidation(EWC)to effectively balance forgetting prevention with fine-tuning efficiency.Using Li_(6)PS_(5)Cl(LPSC),a sulfide-based Li solid-state electrolyte,as a fine-tuning target,we show that reEWC significantly improves the accuracy of a pretrained MLIP,resolving well-known issues of potential energy surface softening and overestimated Li diffusivities.Moreover,reEWC preserves the generalizability of the pretrained MLIP and enables knowledge transfer to chemically distinct systems,including other sulfide,oxide,nitride,and halide electrolytes.Compared to Experience Replay and EWC used individually,reEWC delivers clear synergistic benefits,mitigating their respective limitations while maintaining computational efficiency.These results establish reEWC as a robust and effective solution for continual learning in MLIPs,enabling universal models that can advance materials research through large-scale,high-throughput simulations across diverse chemistries.
基金supported by the National Natural Science Foundation of China(Grants Nos.12174450 and 11874429)the National Talents Program of China,the Science and Technology Innovation Program of Hunan Province(Grant No.2024RC1013)+3 种基金the Key Project of Hunan Provincial Natural Science Foundation(Grant No.2024JJ3029)the Hunan Provincial Key Research and Development Program(Grant No.2022WK2002)the Distinguished Youth Foundation of Hunan Province(Grant No.2020JJ2039),the Project of High-Level Talents Accumulation of Hunan Province(Grant No.2018RS3021)Program of Hundreds of Talents of Hunan Province,the State Key Laboratory of Powder Metallurgy,Start-up Funding and Innovation-Driven Plan(Grant No.2019CX023)of Central South University,Postgraduate Scientific Research Innovation Project of Hunan Province(Grants No.CX20230104)。
文摘Spin Hall effect(SHE)provides a promising solution to the realization of advantageous functionalities for spin-based recording and information processing.In this work,we conduct high-throughput calculations on the spin Hall conductivity(SHC)of antiperovskite compounds with the composition ZXM3,where Z is a nonmetal,X is a metal,and M is a platinum group metal.From an initial database over 4500 structures,we screen 295 structurally stable compounds and identify 24 compounds with intrinsic SHC exceeding 500(ℏ/e)(Ω^(⁻1)cm^(⁻1)).We reveal a strong dependence of SHC on spin-orbit coupling-induced energy splitting near the Fermi level.In addition,SHCs can be regulated through proper doping of electrons or holes.The present work establishes high-throughput database of SHC in antiperovskites which is crucial for designing future electric and spintronic devices.
基金supported by the National Natural Science Foundation of China(50971043 and 51171046)the Key Research and Development Program of China(CISRI-21T62450ZD)+1 种基金the Natural Science Foundation of Fujian Province(2014J01176,2018J01754,and 2021J01590)the Student Research and Training Program(SRTP)of Fuzhou University(27297).
文摘It is urgent to establish a series of reasonable and general approaches to qualitatively and graphically characterize the four core effects of multi-principal element alloys(MPEAs)based on the inherent site preference.In this contribution,a qualitatively and graphically characterizing approach to the diffusion behavior of interstitial nonmetallic atoms diffusing along the neighboring octahedra in MPEAs was explored intensively.For this purpose,the C atom diffusing along the neighboring octahedra in FCC_CoNiV MPEA with(V1.0000)1a(Co0.4445Ni0.4444V0.1111)3c,a constant ordered occupying configuration predicted in our previous paper,was demonstrated in detail.Six distinct diffusion paths along[110],[101],and[011]directions on XY,XZ,and YZ planes of FCC_CoNiV MPEA with forward and backward diffusion directions were explored one by one,respectively.The diffusion energy barrier,diffusion coefficient,diffusion constant,and activation energy were derived by employing first-principles calculations based on density functional theory alongside the Climbing Image Nudged Elastic Band method.Unlike diffusing behavior in pure metallic elements,the non-periodic diffusion energy barrier waves are revealed for the real FCC_CoNiV MPEA structure.The significant variations in the diffusion energy barriers are influenced by the atomic environment,particularly the interaction between V and C atoms,which enhances the localization of electrons and increases the overall diffusion energy barrier.The energy barriers show similar trends along six paths,but significant variations occur across different octahedral sites.
