Pub Date : 2026-03-25DOI: 10.1038/s41524-026-02048-3
Rhyan Barrett, Johannes C. B. Dietschreit, Julia Westermayr
Simulating long-range interactions remains a significant challenge for molecular machine learning (ML) potentials due to the need to accurately capture interactions over large spatial regions. In this work, we integrate the multipole expansion into equivariant ML potentials to model long-range interactions present in QM/MM simulations more accurately. By incorporating the multipole expansion, we are able to effectively capture environmental long-range effects in both ground and excited states. Benchmark evaluations demonstrate the superior performance of including higher-order features from atoms in the environment. To showcase the efficacy of our model, we accurately predict properties such as energies and forces for nickel complex systems and simulate the nonadiabatic excited-state dynamics of a ring-opening reaction in solution. Furthermore, we show that transfer learning from foundational models trained without any explicit environment enhances data efficiency, reducing the need to generate large QM/MM datasets before training. These examples demonstrate the versatility of our approach, paving the way for efficient, accurate, and scalable simulations of complex molecular systems and materials across electronic states.
{"title":"Incorporating long-range interactions via the multipole expansion into ground and excited-state molecular simulations","authors":"Rhyan Barrett, Johannes C. B. Dietschreit, Julia Westermayr","doi":"10.1038/s41524-026-02048-3","DOIUrl":"https://doi.org/10.1038/s41524-026-02048-3","url":null,"abstract":"Simulating long-range interactions remains a significant challenge for molecular machine learning (ML) potentials due to the need to accurately capture interactions over large spatial regions. In this work, we integrate the multipole expansion into equivariant ML potentials to model long-range interactions present in QM/MM simulations more accurately. By incorporating the multipole expansion, we are able to effectively capture environmental long-range effects in both ground and excited states. Benchmark evaluations demonstrate the superior performance of including higher-order features from atoms in the environment. To showcase the efficacy of our model, we accurately predict properties such as energies and forces for nickel complex systems and simulate the nonadiabatic excited-state dynamics of a ring-opening reaction in solution. Furthermore, we show that transfer learning from foundational models trained without any explicit environment enhances data efficiency, reducing the need to generate large QM/MM datasets before training. These examples demonstrate the versatility of our approach, paving the way for efficient, accurate, and scalable simulations of complex molecular systems and materials across electronic states.","PeriodicalId":19342,"journal":{"name":"npj Computational Materials","volume":"14 1","pages":""},"PeriodicalIF":9.7,"publicationDate":"2026-03-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147506163","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 : 2026-03-24DOI: 10.1038/s41524-026-02039-4
Evan R. Antoniuk, Peggy Li, Nathan Keilbart, Stephen Weitzner, Bhavya Kailkhura, Anna M. Hiszpanski
{"title":"Publisher Correction: Active learning enables generation of molecules that advance the known Pareto front","authors":"Evan R. Antoniuk, Peggy Li, Nathan Keilbart, Stephen Weitzner, Bhavya Kailkhura, Anna M. Hiszpanski","doi":"10.1038/s41524-026-02039-4","DOIUrl":"https://doi.org/10.1038/s41524-026-02039-4","url":null,"abstract":"","PeriodicalId":19342,"journal":{"name":"npj Computational Materials","volume":"235 1","pages":""},"PeriodicalIF":9.7,"publicationDate":"2026-03-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147506548","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 : 2026-03-24DOI: 10.1038/s41524-026-02044-7
Aleksandr Avdoshin, Nikita A. Matsokin, Thanh-Nam Huynh, Dmitry I. Sharapa, Karin Fink, Felix Studt, Wolfgang Wenzel, Mariana Kozlowska
Single-atom catalysts (SACs) provide isolated, well-defined metal sites that are suited for mechanistic modeling in porous materials such as metal-organic frameworks (MOFs). However, the influence of framework topology and mass transport on catalytic outcomes remains poorly understood. Here we develop a multiscale kinetic model for ethylene oligomerization in Ni-grafted NU-1000 that combines density functional theory (DFT)-derived free-energy barriers with adsorption and diffusion descriptors. The framework predicts product distributions under realistic reaction conditions. The simulations show that flow-mode operation favors selective C4H8 formation across a temperature range. This selectivity window progressively narrows with increasing effective diffusion length and catalytic-site density, as longer residence times enhance chain growth beyond dimerization. In contrast, batch-mode operation shifts the product distribution toward heavier olefins. These trends provide practical guidance for tuning operating conditions and material properties to achieve desired selective Ni-MOF catalysts.
{"title":"Multiscale kinetic model of ethylene oligomerization in Ni-NU-1000 metal-organic framework","authors":"Aleksandr Avdoshin, Nikita A. Matsokin, Thanh-Nam Huynh, Dmitry I. Sharapa, Karin Fink, Felix Studt, Wolfgang Wenzel, Mariana Kozlowska","doi":"10.1038/s41524-026-02044-7","DOIUrl":"https://doi.org/10.1038/s41524-026-02044-7","url":null,"abstract":"Single-atom catalysts (SACs) provide isolated, well-defined metal sites that are suited for mechanistic modeling in porous materials such as metal-organic frameworks (MOFs). However, the influence of framework topology and mass transport on catalytic outcomes remains poorly understood. Here we develop a multiscale kinetic model for ethylene oligomerization in Ni-grafted NU-1000 that combines density functional theory (DFT)-derived free-energy barriers with adsorption and diffusion descriptors. The framework predicts product distributions under realistic reaction conditions. The simulations show that flow-mode operation favors selective C4H8 formation across a temperature range. This selectivity window progressively narrows with increasing effective diffusion length and catalytic-site density, as longer residence times enhance chain growth beyond dimerization. In contrast, batch-mode operation shifts the product distribution toward heavier olefins. These trends provide practical guidance for tuning operating conditions and material properties to achieve desired selective Ni-MOF catalysts.","PeriodicalId":19342,"journal":{"name":"npj Computational Materials","volume":"9 1","pages":""},"PeriodicalIF":9.7,"publicationDate":"2026-03-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147506158","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 : 2026-03-23DOI: 10.1038/s41524-026-02034-9
Hendrik Schrautzer, Moritz A. Goerzen, Bjarne Beyer, Soumyajyoti Haldar, Pavel F. Bessarab, Stefan Heinze
Reliable control of skyrmion lifetime is essential for realizing spintronic devices, yet the role of higher-order exchange—which can lead to skyrmion stabilization—remains largely unexplored. Here we calculate lifetimes of isolated skyrmions and antiskyrmions at transition-metal interfaces based on an atomistic spin model that includes all fourth-order exchange terms. Within harmonic transition-state theory, we evaluate both energetic and entropic contributions and find substantially enhanced lifetimes when higher-order exchange is included. The four-spin four-site interaction raises the energy barrier and lowers the curvature of the energy landscape at the collapse saddle point, increasing the pre-exponential factor. We show that skyrmions and antiskyrmions can remain thermally stable even without Dzyaloshinskii-Moriya interaction (DMI), and that tuning the four-spin term by a small amount modulates the prefactor over orders of magnitude. Our results identify higher-order exchange as a promising route to stabilize topological magnetic textures—in particular in systems lacking DMI—and to engineer their thermally activated decay.
{"title":"Impact of higher-order exchange on the lifetime of skyrmions and antiskyrmions","authors":"Hendrik Schrautzer, Moritz A. Goerzen, Bjarne Beyer, Soumyajyoti Haldar, Pavel F. Bessarab, Stefan Heinze","doi":"10.1038/s41524-026-02034-9","DOIUrl":"https://doi.org/10.1038/s41524-026-02034-9","url":null,"abstract":"Reliable control of skyrmion lifetime is essential for realizing spintronic devices, yet the role of higher-order exchange—which can lead to skyrmion stabilization—remains largely unexplored. Here we calculate lifetimes of isolated skyrmions and antiskyrmions at transition-metal interfaces based on an atomistic spin model that includes all fourth-order exchange terms. Within harmonic transition-state theory, we evaluate both energetic and entropic contributions and find substantially enhanced lifetimes when higher-order exchange is included. The four-spin four-site interaction raises the energy barrier and lowers the curvature of the energy landscape at the collapse saddle point, increasing the pre-exponential factor. We show that skyrmions and antiskyrmions can remain thermally stable even without Dzyaloshinskii-Moriya interaction (DMI), and that tuning the four-spin term by a small amount modulates the prefactor over orders of magnitude. Our results identify higher-order exchange as a promising route to stabilize topological magnetic textures—in particular in systems lacking DMI—and to engineer their thermally activated decay.","PeriodicalId":19342,"journal":{"name":"npj Computational Materials","volume":"2 1","pages":""},"PeriodicalIF":9.7,"publicationDate":"2026-03-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147496684","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 : 2026-03-20DOI: 10.1038/s41524-026-02041-w
Xiao Fu, Jing Xu, Qifan Yang, Xuhe Gong, Jingchen Lian, Liqi Wang, Zibin Wang, Ruijuan Xiao, Hong Li
The rapid development of computational materials science powered by machine learning (ML) is gradually leading to solutions to several previously intractable scientific problems. One of the most prominent is machine learning interatomic potentials (MLIPs), which expedites the study of dynamical methods for large-scale systems. However, as a promising field, high-entropy (HE) solid-state electrolytes (SEs) remain constrained by trial-and-error paradigms, lacking systematic computational strategies to address their huge and high-dimensional composition space. In this work, we establish a dual-stage ML framework that combines fine-tuned MLIPs with interpretable feature-property mapping to accelerate the high-entropy SEs discovery. Using Li3Zr2Si2PO12 (LZSP) as a prototype, the fine-tuned CHGNet-based relaxation provides atomic structure for each configuration, the structure features - mean squared displacement (SF-MSD) model predicts the ionic transport properties and identifies critical descriptors. The theoretical studies indicate that the framework can satisfy the multiple requirements including computational efficiency, generalization reliability and prediction accuracy. One of the most promising element combinations in the quinary HE-LZSP space containing 4575 compositions is identified with a high ionic conductivity of 4.53 mS/cm as an application example. The framework contains generalizability and extensibility to other SE families.
{"title":"High-entropy solid electrolytes discovery: a dual-stage machine learning framework bridging atomic configurations and ionic transport properties","authors":"Xiao Fu, Jing Xu, Qifan Yang, Xuhe Gong, Jingchen Lian, Liqi Wang, Zibin Wang, Ruijuan Xiao, Hong Li","doi":"10.1038/s41524-026-02041-w","DOIUrl":"https://doi.org/10.1038/s41524-026-02041-w","url":null,"abstract":"The rapid development of computational materials science powered by machine learning (ML) is gradually leading to solutions to several previously intractable scientific problems. One of the most prominent is machine learning interatomic potentials (MLIPs), which expedites the study of dynamical methods for large-scale systems. However, as a promising field, high-entropy (HE) solid-state electrolytes (SEs) remain constrained by trial-and-error paradigms, lacking systematic computational strategies to address their huge and high-dimensional composition space. In this work, we establish a dual-stage ML framework that combines fine-tuned MLIPs with interpretable feature-property mapping to accelerate the high-entropy SEs discovery. Using Li3Zr2Si2PO12 (LZSP) as a prototype, the fine-tuned CHGNet-based relaxation provides atomic structure for each configuration, the structure features - mean squared displacement (SF-MSD) model predicts the ionic transport properties and identifies critical descriptors. The theoretical studies indicate that the framework can satisfy the multiple requirements including computational efficiency, generalization reliability and prediction accuracy. One of the most promising element combinations in the quinary HE-LZSP space containing 4575 compositions is identified with a high ionic conductivity of 4.53 mS/cm as an application example. The framework contains generalizability and extensibility to other SE families.","PeriodicalId":19342,"journal":{"name":"npj Computational Materials","volume":"14 1","pages":""},"PeriodicalIF":9.7,"publicationDate":"2026-03-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147496685","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 : 2026-03-19DOI: 10.1038/s41524-026-02037-6
Nathan Linton, Parampreet Singh, Dilpuneet S. Aidhy
The compositional complexity and chemical randomness of high entropy alloys (HEAs) make conventional atomic-scale calculations, such as density functional theory (DFT), prohibitively expensive for property prediction. One key property of interest is the vacancy formation energy (({E}_{v}^{f})), which plays a crucial role in diffusion and microstructure evolution. In this work, we present a machine learning (ML) framework that eliminates the need for DFT calculations by predicting ({E}_{v}^{f}{rm{s}}) in HEAs using models trained on binary and ternary alloys. Our approach first relaxes face-centered cubic (FCC) structures using a fine-tuned CHGNet model and then uses the resulting configurations as input into a crystal graph convolutional neural network (CGCNN) to predict both Bader charges and ({E}_{v}^{f}{rm{s}}). Incorporating Bader charges as descriptors introduces DFT-informed electronic structure information into the model, significantly improving prediction accuracy compared to using elemental features alone. Furthermore, we demonstrate that the model’s performance generalizes well to other alloy systems with minimal fine-tuning, offering a robust and efficient path toward high-throughput defect property prediction in complex alloys.
{"title":"Framework to completely bypass expensive DFT calculations via graph neural networks for vacancy formation energy predictions in FCC high entropy alloys","authors":"Nathan Linton, Parampreet Singh, Dilpuneet S. Aidhy","doi":"10.1038/s41524-026-02037-6","DOIUrl":"https://doi.org/10.1038/s41524-026-02037-6","url":null,"abstract":"The compositional complexity and chemical randomness of high entropy alloys (HEAs) make conventional atomic-scale calculations, such as density functional theory (DFT), prohibitively expensive for property prediction. One key property of interest is the vacancy formation energy (({E}_{v}^{f})), which plays a crucial role in diffusion and microstructure evolution. In this work, we present a machine learning (ML) framework that eliminates the need for DFT calculations by predicting ({E}_{v}^{f}{rm{s}}) in HEAs using models trained on binary and ternary alloys. Our approach first relaxes face-centered cubic (FCC) structures using a fine-tuned CHGNet model and then uses the resulting configurations as input into a crystal graph convolutional neural network (CGCNN) to predict both Bader charges and ({E}_{v}^{f}{rm{s}}). Incorporating Bader charges as descriptors introduces DFT-informed electronic structure information into the model, significantly improving prediction accuracy compared to using elemental features alone. Furthermore, we demonstrate that the model’s performance generalizes well to other alloy systems with minimal fine-tuning, offering a robust and efficient path toward high-throughput defect property prediction in complex alloys.","PeriodicalId":19342,"journal":{"name":"npj Computational Materials","volume":"27 3 1","pages":""},"PeriodicalIF":9.7,"publicationDate":"2026-03-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147496686","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 : 2026-03-19DOI: 10.1038/s41524-026-02029-6
Khaled Badawy, Lianxi Zheng, Nirpendra Singh
Robust two-dimensional magnets are essential for next-generation spintronics. Using first-principles calculations, we demonstrate that only the antiferromagnetic 1H- and 1T-CrSe2 exhibit stable magnon dispersions. The preferred stability of 1T phase originates from spin-ordering polarization of correlated Cr-d states among three low-lying crystal-field levels. These levels are localized with distinct orbital character in the 1T phase, but delocalized in the 1H phase. The full occupation of low-lying levels leads to antiferromagnetic exchange, yielding Néel temperatures of 310 K (1T) and 274 K (1H). By introducing 25% Se line defects in CrSe2 monolayer, we predict a novel monolayer Cr2Se3 in H and T phases (analogous to their parent 1H/1T-CrSe2). Both Cr2Se3 phases are stable and are half-metallic, with spin (↓) band gaps of 1.39 eV (H) and 2.28 eV (T). Cr2Se3/h-BN heterostructures preserve the electronic properties, indicating feasible growth on h-BN substrates. In both phases, the partial occupation of the low-lying crystal-field levels enhances ferromagnetic exchange through hopping between occupied and unoccupied orbitals. Remarkably, Curie temperatures based on the Heisenberg Hamiltonian reach 547 K (H) and 606 K (T). The H phase satisfies the Stoner criterion, while the Heisenberg-like T phase shifts toward the Stoner regime under 2–4% biaxial tensile strain. These results position Cr2Se3 as a promising half-metallic 2D magnet.
坚固的二维磁体对下一代自旋电子学至关重要。利用第一性原理计算，我们证明了只有反铁磁性的1H-和1T-CrSe2表现出稳定的磁振子色散。1T相的优先稳定性源于三个低洼晶体场能级间相关Cr-d态的自旋有序极化。这些能级在1T相具有明显的轨道特征，但在1H相不具有局域性。完全占据低洼能级导致反铁磁交换，产生310 K （1T）和274 K （1H）的nsamel温度。通过在CrSe2单分子层中引入25%的Se线缺陷，我们预测在H和T相中形成新的Cr2Se3单分子层（类似于它们的母体1H/1T-CrSe2）。两个Cr2Se3相都是稳定的半金属相，自旋（↓）带隙分别为1.39 eV (H)和2.28 eV (T)。Cr2Se3/h-BN异质结构保留了电子性能，表明在h-BN衬底上生长是可行的。在这两个相中，低能级晶体场的部分占据通过已占据轨道和未占据轨道之间的跳跃增强了铁磁交换。值得注意的是，基于海森堡哈密顿量的居里温度达到了547 K (H)和606 K (T)。在2-4%的双轴拉伸应变下，H相满足Stoner准则，而类海森堡T相则向Stoner模式转移。这些结果表明Cr2Se3是一种很有前途的半金属二维磁铁。
{"title":"Discovery of a novel half metallic 2D Cr2Se3 monolayer with high Curie temperature from correlated antiferromagnetic 2D CrSe2","authors":"Khaled Badawy, Lianxi Zheng, Nirpendra Singh","doi":"10.1038/s41524-026-02029-6","DOIUrl":"https://doi.org/10.1038/s41524-026-02029-6","url":null,"abstract":"Robust two-dimensional magnets are essential for next-generation spintronics. Using first-principles calculations, we demonstrate that only the antiferromagnetic 1H- and 1T-CrSe2 exhibit stable magnon dispersions. The preferred stability of 1T phase originates from spin-ordering polarization of correlated Cr-d states among three low-lying crystal-field levels. These levels are localized with distinct orbital character in the 1T phase, but delocalized in the 1H phase. The full occupation of low-lying levels leads to antiferromagnetic exchange, yielding Néel temperatures of 310 K (1T) and 274 K (1H). By introducing 25% Se line defects in CrSe2 monolayer, we predict a novel monolayer Cr2Se3 in H and T phases (analogous to their parent 1H/1T-CrSe2). Both Cr2Se3 phases are stable and are half-metallic, with spin (↓) band gaps of 1.39 eV (H) and 2.28 eV (T). Cr2Se3/h-BN heterostructures preserve the electronic properties, indicating feasible growth on h-BN substrates. In both phases, the partial occupation of the low-lying crystal-field levels enhances ferromagnetic exchange through hopping between occupied and unoccupied orbitals. Remarkably, Curie temperatures based on the Heisenberg Hamiltonian reach 547 K (H) and 606 K (T). The H phase satisfies the Stoner criterion, while the Heisenberg-like T phase shifts toward the Stoner regime under 2–4% biaxial tensile strain. These results position Cr2Se3 as a promising half-metallic 2D magnet.","PeriodicalId":19342,"journal":{"name":"npj Computational Materials","volume":"17 1","pages":""},"PeriodicalIF":9.7,"publicationDate":"2026-03-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147496688","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}
Machine learning interatomic potentials (MLIP) are powerful tools for using large-scale molecular dynamics (MD) to evaluate material properties, including the performance of solid-state electrolytes (SSEs). While there are many efforts for constructing universal big MLIP models, their accuracies and speeds of inference still need to be improved for many practical applications. Another approach is to develop a system-specific MLIP model relying on active learning strategy. Although much cheaper than training a big model, using the conventional procedure, it still requires large numbers of active learning loops and the corresponding DFT calculations to ensure convergency. Here, we propose a single-shot workflow that significantly accelerates small MLIP model development by leveraging the capabilities of the big model (using MACE as one example) and requiring only a few hundred additional DFT calculations. Our workflow comprises two stages, first the MACE model itself is fine-tuned to make it more accurate for the given system, second a smaller MLIP model (using NEP as one example) is distilled from the fine-tuned MACE model. We employed a MACE-driven sampling strategy, carried out additional DFT calculations without relying on active learning iterations. We show that fine-tuned MACE model can inherit the stability of the pretrained model, and fine-tuning the pretrained MACE model is much more DFT data efficient comparing to training a start-from-scratch NEP model. In the second stage, the fine-tuned MACE model provides the dataset to train the NEP model, allows the final NEP model to carry out large scale MD simulations with competitive accuracy. This integrated workflow establishes a systematic pathway for rapid MLIP development via small additional DFT dataset, with potential applications to many material systems.
{"title":"Constructing machine learning interatomic potentials with minimum amount of ab initio data","authors":"Wentao Zhang, Xingxing Wu, Chen Wang, Siyu Hu, Yueyang Liu, Lin-Wang Wang","doi":"10.1038/s41524-026-02023-y","DOIUrl":"https://doi.org/10.1038/s41524-026-02023-y","url":null,"abstract":"Machine learning interatomic potentials (MLIP) are powerful tools for using large-scale molecular dynamics (MD) to evaluate material properties, including the performance of solid-state electrolytes (SSEs). While there are many efforts for constructing universal big MLIP models, their accuracies and speeds of inference still need to be improved for many practical applications. Another approach is to develop a system-specific MLIP model relying on active learning strategy. Although much cheaper than training a big model, using the conventional procedure, it still requires large numbers of active learning loops and the corresponding DFT calculations to ensure convergency. Here, we propose a single-shot workflow that significantly accelerates small MLIP model development by leveraging the capabilities of the big model (using MACE as one example) and requiring only a few hundred additional DFT calculations. Our workflow comprises two stages, first the MACE model itself is fine-tuned to make it more accurate for the given system, second a smaller MLIP model (using NEP as one example) is distilled from the fine-tuned MACE model. We employed a MACE-driven sampling strategy, carried out additional DFT calculations without relying on active learning iterations. We show that fine-tuned MACE model can inherit the stability of the pretrained model, and fine-tuning the pretrained MACE model is much more DFT data efficient comparing to training a start-from-scratch NEP model. In the second stage, the fine-tuned MACE model provides the dataset to train the NEP model, allows the final NEP model to carry out large scale MD simulations with competitive accuracy. This integrated workflow establishes a systematic pathway for rapid MLIP development via small additional DFT dataset, with potential applications to many material systems.","PeriodicalId":19342,"journal":{"name":"npj Computational Materials","volume":"4 1","pages":""},"PeriodicalIF":9.7,"publicationDate":"2026-03-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147464952","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}
The rapid development of two-dimensional van der Waals heterostructures has sparked notable interest in optoelectronic applications. However, issues such as lattice mismatch or a misalignment of the constituent layers can drastically suppress charge transfer for these interlayer transitions. Here, we construct a new type-II MoS2/Ti2CO2 heterojunction using density functional theory and non-adiabatic molecular dynamics simulations, revealing the optimal band alignments across various stacking configurations. The optimized heterointerface exhibits ultrafast charge separation, with electron and hole transfer completing within 4.6 fs and 228.8 fs, respectively, and a prolonged carrier lifetime of 1.53 ns. Compared to pristine monolayers, the heterointerface displays broader light absorption from the visible to the UV spectrum. This optoelectronic performance is further enhanced by biaxial strain, which effectively tunes the photoresponse, resulting in a high theoretical power conversion efficiency of 12.89%. These findings offer valuable guidance for designing high-performance MoS2-based heterostructures for next-generation optoelectronic and energy conversion devices.
{"title":"Photoinduced ultrafast charge transfer and enhanced optoelectronics in MoS2/Ti2CO2 van der Waals heterojunction","authors":"Xianke Yue, Zhong Zhou, Xiaodong Wang, Qi An, Kolan Madhav Reddy","doi":"10.1038/s41524-026-02035-8","DOIUrl":"https://doi.org/10.1038/s41524-026-02035-8","url":null,"abstract":"The rapid development of two-dimensional van der Waals heterostructures has sparked notable interest in optoelectronic applications. However, issues such as lattice mismatch or a misalignment of the constituent layers can drastically suppress charge transfer for these interlayer transitions. Here, we construct a new type-II MoS2/Ti2CO2 heterojunction using density functional theory and non-adiabatic molecular dynamics simulations, revealing the optimal band alignments across various stacking configurations. The optimized heterointerface exhibits ultrafast charge separation, with electron and hole transfer completing within 4.6 fs and 228.8 fs, respectively, and a prolonged carrier lifetime of 1.53 ns. Compared to pristine monolayers, the heterointerface displays broader light absorption from the visible to the UV spectrum. This optoelectronic performance is further enhanced by biaxial strain, which effectively tunes the photoresponse, resulting in a high theoretical power conversion efficiency of 12.89%. These findings offer valuable guidance for designing high-performance MoS2-based heterostructures for next-generation optoelectronic and energy conversion devices.","PeriodicalId":19342,"journal":{"name":"npj Computational Materials","volume":"270 1","pages":""},"PeriodicalIF":9.7,"publicationDate":"2026-03-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147464954","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 : 2026-03-13DOI: 10.1038/s41524-026-02036-7
Yuting Wang, Weikang Wu, Jianzhou Zhao
The interplay between strong electronic correlations, unconventional superconductivity, and symmetry-protected topology provides a fertile ground for discovering exotic quantum states. In this work, we investigate the correlated electronic structure and topological properties of the heavy fermion material CeCoGe3 using density functional theory combined with dynamical mean-field theory calculations. Our results reveal a crossover from high temperature incoherent states to low temperature coherent heavy quasiparticles, accompanied by a mass enhancement of m*/mDFT ~ 52.6 at T = 25 K. The interplay between electronic correlation, spin-orbit coupling and the noncentrosymmetric I4mm crystal symmetry stabilizes flat topological nodal lines within 10 meV of the Fermi level, which could contribute a significant density of states. The proximity of topological nodal lines to the Fermi surface suggests a potential role in mediating pressure induced unconventional superconductivity. Our work establishes CeCoGe3 as a prototype topological nodal line Kondo semimetal. The coexistence of strong correlation, non-trivial band topology and superconductivity indicates CeCoGe3 as a potential candidate for realizing topological superconductivity.
{"title":"Flat topological nodal lines in heavy-fermion compound CeCoGe3","authors":"Yuting Wang, Weikang Wu, Jianzhou Zhao","doi":"10.1038/s41524-026-02036-7","DOIUrl":"https://doi.org/10.1038/s41524-026-02036-7","url":null,"abstract":"The interplay between strong electronic correlations, unconventional superconductivity, and symmetry-protected topology provides a fertile ground for discovering exotic quantum states. In this work, we investigate the correlated electronic structure and topological properties of the heavy fermion material CeCoGe3 using density functional theory combined with dynamical mean-field theory calculations. Our results reveal a crossover from high temperature incoherent states to low temperature coherent heavy quasiparticles, accompanied by a mass enhancement of m*/mDFT ~ 52.6 at T = 25 K. The interplay between electronic correlation, spin-orbit coupling and the noncentrosymmetric I4mm crystal symmetry stabilizes flat topological nodal lines within 10 meV of the Fermi level, which could contribute a significant density of states. The proximity of topological nodal lines to the Fermi surface suggests a potential role in mediating pressure induced unconventional superconductivity. Our work establishes CeCoGe3 as a prototype topological nodal line Kondo semimetal. The coexistence of strong correlation, non-trivial band topology and superconductivity indicates CeCoGe3 as a potential candidate for realizing topological superconductivity.","PeriodicalId":19342,"journal":{"name":"npj Computational Materials","volume":"10 4 1","pages":""},"PeriodicalIF":9.7,"publicationDate":"2026-03-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147454740","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：