Zhenyu Wang
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Zhenyu Wang has a total of 125 co-authors. Here are the top co-authors by number of shared publications:
- Yanchao Wang
Key Laboratory of Material Simulation Methods and Software of Ministry of Education, College of Physics, Jilin University, Changchun, China
8 shared publications - Jian Lv
Jilin University
7 shared publications - Xiaoshan Luo
Key Laboratory of Material Simulation Methods and Software of Ministry of Education, College of Physics, Jilin University, Changchun, PR China
7 shared publications - Han Wang
Institute of Applied Physics and Computational Mathematics
3 shared publications - Yanming Ma
Jilin University
3 shared publications Cheng Ma
Key Laboratory of Material Simulation Methods and Software of Ministry of Education, College of Physics, Jilin University, Changchun, China
2 shared publications
Latest Publications
- Open Access
- 25 September 2026
- Published by American Association for the Advancement of Science (AAAS) in Science Advances
- Vol. 12 (39)
Boron-assisted synthesis of compositionally complex amorphous oxides via short-range-order-constrained generative design
Engineering short-range atomic order offers a promising route to design amorphous solids. Here, we establish a boron-assisted amorphization strategy using ApolloX, a theory-guided, short-range-order-constrained generative framework for identifying low-energy configurations in compositionally complex multielement systems. Using FeCoNiMoBO x as a model platform, ApolloX generates candidate amorphous structures across varied boron contents. Ab initio molecular dynamics simulations reveal that increasing the boron content suppresses atomic diffusion and crystallization, accompanied by the stabilization of BO 3 -centered local motifs that favor amorphous structure formation. Guided by these predictions, we synthesize three FeCoNiMoBO x compositions with distinct boron contents and use synchrotron scattering and electron microscopy to confirm their compositional fidelity, structural homogeneity, and targeted amorphous characteristics, thereby validating the predicted boron-dependent structural evolution. We further extend this strategy to a broader library of multimetal BO x compositions spanning diverse metal combinations and boron loadings. These results identify boron incorporation as an effective and generalizable means of enhancing amorphization and establish a theory-guided framework for the discovery of compositionally complex amorphous materials.
- 25 September 2026
- Published by American Association for the Advancement of Science (AAAS) in Science Advances
- Vol. 12 (39)
Open AccessBoron-assisted synthesis of compositionally complex amorphous oxides via short-range-order-constrained generative design
Engineering short-range atomic order offers a promising route to design amorphous solids. Here, we establish a boron-assisted amorphization strategy using ApolloX, a theory-guided, short-range-order-constrained generative framework for identifying low-energy configurations in compositionally complex multielement systems. Using FeCoNiMoBO x as a model platform, ApolloX generates candidate amorphous structures across varied boron contents. Ab initio molecular dynamics simulations reveal that increasing the boron content suppresses atomic diffusion and crystallization, accompanied by the stabilization of BO 3 -centered local motifs that favor amorphous structure formation. Guided by these predictions, we synthesize three FeCoNiMoBO x compositions with distinct boron contents and use synchrotron scattering and electron microscopy to confirm their compositional fidelity, structural homogeneity, and targeted amorphous characteristics, thereby validating the predicted boron-dependent structural evolution. We further extend this strategy to a broader library of multimetal BO x compositions spanning diverse metal combinations and boron loadings. These results identify boron incorporation as an effective and generalizable means of enhancing amorphization and establish a theory-guided framework for the discovery of compositionally complex amorphous materials.
- 20 March 2026
- Published by American Chemical Society (ACS) in The Journal of Physical Chemistry Letters
- Vol. 17 (13) , 3759-3765
Pressure-Induced Discovery of a Topological Phase of Bi 4 Br 4 with AA′′-Stacking
Bismuth halides Bi4X4 (X = Br, I) consist of quasi-one-dimensional van der Waals chains whose interchain stacking can be reshaped by pressure, potentially reprogramming topology. Using CALYPSO structure searches with first-principles enthalpy ranking, we uncover a previously unreported C2/m phase of Bi4Br4 featuring a novel AA″ stacking that becomes stable above ∼4 GPa and persists to at least 15 GPa. We map the transformation from known β-Bi4Br4 (AA stacking): under compression β develops an imaginary phonon mode, appears as a first-order saddle point in the crystal-solution landscape, and relaxes into the C2/m–AA″ structure with an estimated barrier of ∼40 meV/atom, explaining pressure-driven stacking rearrangement. With spin–orbit coupling, a band gap opens at symmetry crossings and a nontrivial inversion yields Z2 = (001; 0), identifying the AA″ phase as a weak topological insulator with gapless (100)/(010) surface states.
- 20 March 2026
- Published by American Chemical Society (ACS) in The Journal of Physical Chemistry Letters
- Vol. 17 (13) , 3759-3765
Pressure-Induced Discovery of a Topological Phase of Bi 4 Br 4 with AA′′-Stacking
Bismuth halides Bi4X4 (X = Br, I) consist of quasi-one-dimensional van der Waals chains whose interchain stacking can be reshaped by pressure, potentially reprogramming topology. Using CALYPSO structure searches with first-principles enthalpy ranking, we uncover a previously unreported C2/m phase of Bi4Br4 featuring a novel AA″ stacking that becomes stable above ∼4 GPa and persists to at least 15 GPa. We map the transformation from known β-Bi4Br4 (AA stacking): under compression β develops an imaginary phonon mode, appears as a first-order saddle point in the crystal-solution landscape, and relaxes into the C2/m–AA″ structure with an estimated barrier of ∼40 meV/atom, explaining pressure-driven stacking rearrangement. With spin–orbit coupling, a band gap opens at symmetry crossings and a nontrivial inversion yields Z2 = (001; 0), identifying the AA″ phase as a weak topological insulator with gapless (100)/(010) surface states.
- Top Cited PapersOpen Access
- 20 October 2025
- Published by Springer Nature in Nature Communications
- Vol. 16 (1) , 1-12
CrystalFlow: a flow-based generative model for crystalline materials
Deep learning-based generative models hold significant promise for exploring the configuration space of crystalline materials, though their application remains in its early stages. In this study, we present CrystalFlow, a flow-based generative model designed to address the unique challenges of this domain. By combining Continuous Normalizing Flows and Conditional Flow Matching with a graph-based equivariant neural network and symmetry-aware data representations, CrystalFlow efficiently models lattice parameters, atomic coordinates, and atom types. This architecture enables data-efficient learning and the generation of high-quality crystal structures. Our results indicate that CrystalFlow achieves performance comparable to state-of-the-art models on established benchmarks while exhibiting versatile conditional generation capabilities (e.g., predicting structures under specific pressures or material properties), and is approximately an order of magnitude more efficient than diffusion-based models in terms of integration steps.
- 20 October 2025
- Published by Springer Nature in Nature Communications
- Vol. 16 (1) , 1-12
Top Cited PapersOpen AccessCrystalFlow: a flow-based generative model for crystalline materials
Deep learning-based generative models hold significant promise for exploring the configuration space of crystalline materials, though their application remains in its early stages. In this study, we present CrystalFlow, a flow-based generative model designed to address the unique challenges of this domain. By combining Continuous Normalizing Flows and Conditional Flow Matching with a graph-based equivariant neural network and symmetry-aware data representations, CrystalFlow efficiently models lattice parameters, atomic coordinates, and atom types. This architecture enables data-efficient learning and the generation of high-quality crystal structures. Our results indicate that CrystalFlow achieves performance comparable to state-of-the-art models on established benchmarks while exhibiting versatile conditional generation capabilities (e.g., predicting structures under specific pressures or material properties), and is approximately an order of magnitude more efficient than diffusion-based models in terms of integration steps.
- Top Cited Papers
- 2 May 2025
- Published by American Chemical Society (ACS) in Journal of Chemical Theory and Computation
- Vol. 21 (9) , 4375-4385
DeePMD-kit v3: A Multiple-Backend Framework for Machine Learning Potentials
In recent years, machine learning potentials (MLPs) have become indispensable tools in physics, chemistry, and materials science, driving the development of software packages for molecular dynamics (MD) simulations and related applications. These packages, typically built on specific machine learning frameworks, such as TensorFlow, PyTorch, or JAX, face integration challenges when advanced applications demand communication across different frameworks. The previous TensorFlow-based implementation of the DeePMD-kit exemplified these limitations. In this work, we introduce DeePMD-kit version 3, a significant update featuring a multibackend framework that supports TensorFlow, PyTorch, JAX, and PaddlePaddle backends, and demonstrate the versatility of this architecture through the integration of other MLP packages and of differentiable molecular force fields. This architecture allows seamless back-end switching with minimal modifications, enabling users and developers to integrate DeePMD-kit with other packages using different machine learning frameworks. This innovation facilitates the development of more complex and interoperable workflows, paving the way for broader applications of MLPs in scientific research.
- 2 May 2025
- Published by American Chemical Society (ACS) in Journal of Chemical Theory and Computation
- Vol. 21 (9) , 4375-4385
Top Cited PapersDeePMD-kit v3: A Multiple-Backend Framework for Machine Learning Potentials
In recent years, machine learning potentials (MLPs) have become indispensable tools in physics, chemistry, and materials science, driving the development of software packages for molecular dynamics (MD) simulations and related applications. These packages, typically built on specific machine learning frameworks, such as TensorFlow, PyTorch, or JAX, face integration challenges when advanced applications demand communication across different frameworks. The previous TensorFlow-based implementation of the DeePMD-kit exemplified these limitations. In this work, we introduce DeePMD-kit version 3, a significant update featuring a multibackend framework that supports TensorFlow, PyTorch, JAX, and PaddlePaddle backends, and demonstrate the versatility of this architecture through the integration of other MLP packages and of differentiable molecular force fields. This architecture allows seamless back-end switching with minimal modifications, enabling users and developers to integrate DeePMD-kit with other packages using different machine learning frameworks. This innovation facilitates the development of more complex and interoperable workflows, paving the way for broader applications of MLPs in scientific research.
- Open Access
- 28 April 2025
- Published by Springer Nature in Research Square
Conditional Generative Modeling for Amorphous Multi-Element Materials
Amorphous multi-element materials offer unprecedented tunability in composition and properties, yet their rational design remains challenging due to the lack of predictive structure-property relationships and vast configurational space. Traditional modeling struggles to capture the intricate short-range order that dictates stability and functionality. We develop ApolloX, a pioneering predictive framework for amorphous multi-element materials, integrating physics-informed generative modeling with particle swarm optimization, using chemical short-range order as an explicit constraint. By systematically navigating the disordered energy landscape, ApolloX enables targeted design of thermodynamically stable amorphous configurations. It accurately predicts atomic-scale arrangements, including composition-driven metal clustering and amorphization trends, well-validated by experiments, while also guiding synthesis by leveraging sluggish diffusion to control elemental distribution. The resulting structural evolution, governed by composition, directly impacts catalytic performance, enhancing activity and stability with increasing amorphization. This predictive-experimental synergy transforms amorphous materials discovery, unlocking new frontiers in catalysis, energy storage, and functional disordered systems.
- 28 April 2025
- Published by Springer Nature in Research Square
Open AccessConditional Generative Modeling for Amorphous Multi-Element Materials
Amorphous multi-element materials offer unprecedented tunability in composition and properties, yet their rational design remains challenging due to the lack of predictive structure-property relationships and vast configurational space. Traditional modeling struggles to capture the intricate short-range order that dictates stability and functionality. We develop ApolloX, a pioneering predictive framework for amorphous multi-element materials, integrating physics-informed generative modeling with particle swarm optimization, using chemical short-range order as an explicit constraint. By systematically navigating the disordered energy landscape, ApolloX enables targeted design of thermodynamically stable amorphous configurations. It accurately predicts atomic-scale arrangements, including composition-driven metal clustering and amorphization trends, well-validated by experiments, while also guiding synthesis by leveraging sluggish diffusion to control elemental distribution. The resulting structural evolution, governed by composition, directly impacts catalytic performance, enhancing activity and stability with increasing amorphization. This predictive-experimental synergy transforms amorphous materials discovery, unlocking new frontiers in catalysis, energy storage, and functional disordered systems.
- Open Access
- 14 March 2025
- Published by AIP Publishing in Matter and Radiation at Extremes
- Vol. 10 (3)
Advances in high-pressure materials discovery enabled by machine learning
Crystal structure prediction (CSP) is a foundational computational technique for determining the atomic arrangements of crystalline materials, especially under high-pressure conditions. While CSP plays a critical role in materials science, traditional approaches often encounter significant challenges related to computational efficiency and scalability, particularly when applied to complex systems. Recent advances in machine learning (ML) have shown tremendous promise in addressing these limitations, enabling the rapid and accurate prediction of crystal structures across a wide range of chemical compositions and external conditions. This review provides a concise overview of recent progress in ML-assisted CSP methodologies, with a particular focus on machine learning potentials and generative models. By critically analyzing these advances, we highlight the transformative impact of ML in accelerating materials discovery, enhancing computational efficiency, and broadening the applicability of CSP. Additionally, we discuss emerging opportunities and challenges in this rapidly evolving field.
- 14 March 2025
- Published by AIP Publishing in Matter and Radiation at Extremes
- Vol. 10 (3)
Open AccessAdvances in high-pressure materials discovery enabled by machine learning
Crystal structure prediction (CSP) is a foundational computational technique for determining the atomic arrangements of crystalline materials, especially under high-pressure conditions. While CSP plays a critical role in materials science, traditional approaches often encounter significant challenges related to computational efficiency and scalability, particularly when applied to complex systems. Recent advances in machine learning (ML) have shown tremendous promise in addressing these limitations, enabling the rapid and accurate prediction of crystal structures across a wide range of chemical compositions and external conditions. This review provides a concise overview of recent progress in ML-assisted CSP methodologies, with a particular focus on machine learning potentials and generative models. By critically analyzing these advances, we highlight the transformative impact of ML in accelerating materials discovery, enhancing computational efficiency, and broadening the applicability of CSP. Additionally, we discuss emerging opportunities and challenges in this rapidly evolving field.
- Top Cited PapersOpen Access
- 12 November 2024
- Published by Springer Nature in npj Computational Materials
- Vol. 10 (1) , 1-10
Deep learning generative model for crystal structure prediction
- 12 November 2024
- Published by Springer Nature in npj Computational Materials
- Vol. 10 (1) , 1-10
Top Cited PapersOpen AccessDeep learning generative model for crystal structure prediction
- 1 July 2024
- Published by IOP Publishing in Reports on Progress in Physics
- Vol. 87 (7) , 078001
Quantitative regulation of electron–phonon coupling
Electron–phonon (e–p) coupling plays a crucial role in various physical phenomena, and regulation of e–p coupling is vital for the exploration and design of high-performance materials. However, the current research on this topic lacks accurate quantification, hindering further understanding of the underlying physical processes and its applications. In this work, we demonstrate quantitative regulation of e–p coupling, by pressure engineering and in-situ spectroscopy. We successfully observe both a distinct vibrational mode and a strong Stokes shift in layered CrBr3, which are clear signatures of e–p coupling. This allows us to achieve precise quantification of the Huang–Rhys factor S at the actual sample temperature, thus accurately determining the e–p coupling strength. We further reveal that pressure efficiently regulates the e–p coupling in CrBr3, evidenced by a remarkable 40% increase in S value. Our results offer an approach for quantifying and modulating e–p coupling, which can be leveraged for exploring and designing functional materials with targeted e–p coupling strengths.
- 1 July 2024
- Published by IOP Publishing in Reports on Progress in Physics
- Vol. 87 (7) , 078001
Quantitative regulation of electron–phonon coupling
Electron–phonon (e–p) coupling plays a crucial role in various physical phenomena, and regulation of e–p coupling is vital for the exploration and design of high-performance materials. However, the current research on this topic lacks accurate quantification, hindering further understanding of the underlying physical processes and its applications. In this work, we demonstrate quantitative regulation of e–p coupling, by pressure engineering and in-situ spectroscopy. We successfully observe both a distinct vibrational mode and a strong Stokes shift in layered CrBr3, which are clear signatures of e–p coupling. This allows us to achieve precise quantification of the Huang–Rhys factor S at the actual sample temperature, thus accurately determining the e–p coupling strength. We further reveal that pressure efficiently regulates the e–p coupling in CrBr3, evidenced by a remarkable 40% increase in S value. Our results offer an approach for quantifying and modulating e–p coupling, which can be leveraged for exploring and designing functional materials with targeted e–p coupling strengths.
- Open Access
- 23 April 2024
- Published by Springer Nature in Research Square
DPA-2: Towards a universal large atomic model for molecular and materials simulation
The rapid development of artificial intelligence (AI) is driving significant changes in the field of atomic modeling, simulation, and design. AI-based potential energy models have been successfully used to perform large-scale and long-time simulations with the accuracy of ab initio electronic structure methods. However, the model generation process still hinders applications at scale. We envision that the next stage would be a model-centric ecosystem, in which a large atomic model (LAM), pre-trained with as many atomic datasets as possible and can be efficiently fine-tuned and distilled to downstream tasks, would serve the new infrastructure of the field of molecular modeling. We show that DPA-2 can accurately represent a diverse range of chemical systems and materials, enabling high-quality simulations and predictions with significantly reduced efforts compared to traditional methods. Our approach paves the way for a universal large atomic model that can be widely applied in molecular and material simulation research, opening new opportunities for scientific discoveries and industrial applications.
- 23 April 2024
- Published by Springer Nature in Research Square
Open AccessDPA-2: Towards a universal large atomic model for molecular and materials simulation
The rapid development of artificial intelligence (AI) is driving significant changes in the field of atomic modeling, simulation, and design. AI-based potential energy models have been successfully used to perform large-scale and long-time simulations with the accuracy of ab initio electronic structure methods. However, the model generation process still hinders applications at scale. We envision that the next stage would be a model-centric ecosystem, in which a large atomic model (LAM), pre-trained with as many atomic datasets as possible and can be efficiently fine-tuned and distilled to downstream tasks, would serve the new infrastructure of the field of molecular modeling. We show that DPA-2 can accurately represent a diverse range of chemical systems and materials, enabling high-quality simulations and predictions with significantly reduced efforts compared to traditional methods. Our approach paves the way for a universal large atomic model that can be widely applied in molecular and material simulation research, opening new opportunities for scientific discoveries and industrial applications.
- 22 March 2024
- Published by American Physical Society (APS) in Physical Review B
- Vol. 109 (9) , 094117
Concurrent learning scheme for crystal structure prediction
Crystal structure prediction (CSP) and machine learning potential (MLP) are two fundamental methods for modern computational material discovery. While the former aims at efficient sampling of the potential energy surface (PES) for discovering new materials, the latter focuses on reproducing the PES to accelerate various atomic simulation tasks. In this work, we combine the two methods within a concurrent learning framework in an effort to generate efficient MLP models for accelerating CSP. The proposed scheme explores the PES through the swarm-intelligence calypso method, labels the most representative structures with quantum mechanical calculations, and learns the PES through a deep potential (DP) model. The process proceeds in an iterative, computationally efficient, and automated manner, leading to the collection of a most compact reference training set from which the resulting DP model is proven particularly suitable for accelerating calypso structure prediction. The scheme has been systematically benchmarked on binary magnesium-aluminium (Mg-Al) alloys and ternary lithium-lanthanum-hydrogen (Li-La-H) superhydrides, demonstrating its efficiency and reliability in DP model construction and calypso structure prediction. The proposed scheme represents a promising routine to perform the structure prediction of large or multicomponent systems.
- 22 March 2024
- Published by American Physical Society (APS) in Physical Review B
- Vol. 109 (9) , 094117
Concurrent learning scheme for crystal structure prediction
Crystal structure prediction (CSP) and machine learning potential (MLP) are two fundamental methods for modern computational material discovery. While the former aims at efficient sampling of the potential energy surface (PES) for discovering new materials, the latter focuses on reproducing the PES to accelerate various atomic simulation tasks. In this work, we combine the two methods within a concurrent learning framework in an effort to generate efficient MLP models for accelerating CSP. The proposed scheme explores the PES through the swarm-intelligence calypso method, labels the most representative structures with quantum mechanical calculations, and learns the PES through a deep potential (DP) model. The process proceeds in an iterative, computationally efficient, and automated manner, leading to the collection of a most compact reference training set from which the resulting DP model is proven particularly suitable for accelerating calypso structure prediction. The scheme has been systematically benchmarked on binary magnesium-aluminium (Mg-Al) alloys and ternary lithium-lanthanum-hydrogen (Li-La-H) superhydrides, demonstrating its efficiency and reliability in DP model construction and calypso structure prediction. The proposed scheme represents a promising routine to perform the structure prediction of large or multicomponent systems.