npj Computational Materials最新文献_第8页

Multiple Majorana zero modes realization based on superconducting topological crystalline metal ZrRuAs 基于超导拓扑晶体金属ZrRuAs的多Majorana零模实现

IF 9.7 1区材料科学

npj Computational Materials Pub Date : 2026-08-25 DOI: 10.1038/s41524-026-02292-7

Xiaoxu Wang, Jinyu Zou, Gang Xu

{"title":"Multiple Majorana zero modes realization based on superconducting topological crystalline metal ZrRuAs","authors":"Xiaoxu Wang, Jinyu Zou, Gang Xu","doi":"10.1038/s41524-026-02292-7","DOIUrl":"https://doi.org/10.1038/s41524-026-02292-7","url":null,"abstract":"The symmetry-protected multiple Majorana zero modes (MZMs) can be manipulated under external fields and have emerged as a promising pathway toward realizing topological quantum computing. While the suitable materials hosting multiple MZMs are still scarce, we propose a feasible candidate platform named superconducting topological crystalline metals (STCMs) that simultaneously possess symmetry-protected topological bands and intrinsic superconductivity. Model analyses demonstrate that the interplay among s-wave superconductivity, mirror symmetry-protected multiple surface Dirac cones, and the introduced spin splitting leads to high BdG Chern numbers of ({mathcal{N}}=pm {C}_{M}), where CM is the mirror Chern number of the STCM. First-principles calculations identify the experimentally synthesized superconductor ZrRuAs as a promising candidate with CM = 2, hosting two symmetry-protected surface Dirac cones. When integrated into a heterostructure with the ferromagnetic insulator (FMI) such as GdI2, a topological superconducting phase with ({mathcal{N}}=-2) can be realized, giving rise to two branches of MZMs. This new scheme offers advantages of structural simplicity and tunability, making the FMI/STCM heterostructure an ideal platform for investigating multipole MZMs and novel topological qubits.","PeriodicalId":19342,"journal":{"name":"npj Computational Materials","volume":"12 1","pages":""},"PeriodicalIF":9.7,"publicationDate":"2026-08-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148853496","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}

引用次数: 0

Quality-controlled active learning via Gaussian processes for robust structure–property learning in autonomous microscopy 基于高斯过程的质量控制主动学习在自主显微镜中的鲁棒结构-性质学习

IF 9.7 1区材料科学

npj Computational Materials Pub Date : 2026-08-21 DOI: 10.1038/s41524-026-02248-x

Jawad Chowdhury, Ganesh Narasimha, Jan-Chi Yang, Hiroshi Funakubo, Yoshitaka Ehara, Yongtao Liu, Rama Vasudevan

{"title":"Quality-controlled active learning via Gaussian processes for robust structure–property learning in autonomous microscopy","authors":"Jawad Chowdhury, Ganesh Narasimha, Jan-Chi Yang, Hiroshi Funakubo, Yoshitaka Ehara, Yongtao Liu, Rama Vasudevan","doi":"10.1038/s41524-026-02248-x","DOIUrl":"https://doi.org/10.1038/s41524-026-02248-x","url":null,"abstract":"Autonomous experimental systems are increasingly used in materials research to accelerate scientific discovery, but their performance is often limited by low-quality, noisy data. This issue is especially problematic in data-intensive structure–property learning tasks such as Image-to-Spectrum (Im2Spec) and Spectrum-to-Image (Spec2Im) translations, where standard active learning strategies can mistakenly prioritize poor quality measurements. We introduce a gated active learning framework that combines curiosity driven sampling with a physics-informed quality control filter based on Simple Harmonic Oscillator model fits, allowing the system to automatically exclude low fidelity data during acquisition. Evaluations on a pre-acquired dataset of band-excitation piezoresponse spectroscopy (BEPS) data from PbTiO3 thin films with spatially localized noise show that the proposed method outperforms random sampling, standard active learning, and multitask learning strategies. We further deployed the framework in real-time experiments on a separate PbTiO3 thin-film sample with heterogeneous domain structures, demonstrating its effectiveness in autonomous microscopy experiments. Overall, this work supports hybrid autonomy in self-driving labs, where physics-informed quality assessment and active decision-making work together for more reliable discovery.","PeriodicalId":19342,"journal":{"name":"npj Computational Materials","volume":"187 1","pages":""},"PeriodicalIF":9.7,"publicationDate":"2026-08-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148769236","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}

引用次数: 0

Self-supervised restoration of weak signal from noisy scattering and diffraction data in reciprocal space 互易空间中噪声散射和衍射数据弱信号的自监督恢复

IF 9.7 1区材料科学

npj Computational Materials Pub Date : 2026-08-21 DOI: 10.1038/s41524-026-02291-8

Haishuo Zhong, Xiuguo Chen, Jiahao Zhang, Dingxuan Deng, Shuo Liu, Shiyuan Liu

{"title":"Self-supervised restoration of weak signal from noisy scattering and diffraction data in reciprocal space","authors":"Haishuo Zhong, Xiuguo Chen, Jiahao Zhang, Dingxuan Deng, Shuo Liu, Shiyuan Liu","doi":"10.1038/s41524-026-02291-8","DOIUrl":"https://doi.org/10.1038/s41524-026-02291-8","url":null,"abstract":"Reciprocal-space imaging and measurement underpin a wide range of modern characterization techniques in scientific research. Nevertheless, radiation damage and limited photon budgets constrain the high-quality and efficient acquisition of noise-vulnerable weak signals in high-dynamic-range (HDR) reciprocal-space. To overcome the inherent unavailability of noise-free ground truth in most scientific measurement scenarios, here we introduce a self-supervised denoising network, SeDeNet. By leveraging a self-supervised denoising framework incorporating nonlinear correction, SeDeNet enables high-fidelity restoration of weak signal features in HDR scenarios using only noisy data. Extensive experiments on X-ray scattering and diffraction data from diverse samples validate the superior denoising performance and generalization of SeDeNet. Moreover, we demonstrate that SeDeNet achieves approximately 20-fold reduction in acquisition time while preserving ≥ 95% accuracy in one-dimensional grating sample’s parameter extraction. These results highlight a promising approach for the efficient acquisition of low-noise HDR scattering data, while maintaining physical consistency in structural parameter analysis.","PeriodicalId":19342,"journal":{"name":"npj Computational Materials","volume":"9 1","pages":""},"PeriodicalIF":9.7,"publicationDate":"2026-08-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148769238","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}

引用次数: 0

Inverse process design for microstructure engineering in FePt-X nanocomposite thin films: phase-field simulations and machine learning FePt-X纳米复合薄膜微结构工程的逆向工艺设计：相场模拟和机器学习

IF 9.7 1区材料科学

npj Computational Materials Pub Date : 2026-08-21 DOI: 10.1038/s41524-026-02288-3

Xiangwei Guo, Zebin Li, Mojue Zhang, Jingheng Chai, Wei-Heng Hsu, Timothy Klemmer, Olle Heinonen, Yingguo Peng, Ganping Ju, Jia-Mian Hu

{"title":"Inverse process design for microstructure engineering in FePt-X nanocomposite thin films: phase-field simulations and machine learning","authors":"Xiangwei Guo, Zebin Li, Mojue Zhang, Jingheng Chai, Wei-Heng Hsu, Timothy Klemmer, Olle Heinonen, Yingguo Peng, Ganping Ju, Jia-Mian Hu","doi":"10.1038/s41524-026-02288-3","DOIUrl":"https://doi.org/10.1038/s41524-026-02288-3","url":null,"abstract":"FePt-X nanocomposite thin film (X is a non-magnetic segregant) is a critical enabling material for ultrahigh-capacity magnetic recording, but a time- and cost-efficient strategy to identify process parameters that yield desirable microstructure features has remained scarce. Here, we present a computational framework, which employs high-throughput phase-field simulations for the forward prediction of process-microstructure relationship and integrates automated microstructure quantification and active learning for the inverse identification of process parameters. As a proof of concept, the framework is used to inversely identify the gradient energy coefficients that lead to microstructure features matching experimental observation in FePt-Al2O3 films. The identified parameters result in a closer agreement with experiments than expert best guesses. Notably, the framework yields a tenfold reduction in computational cost compared with brute-force high-throughput phase-field simulations. Our results demonstrate the potential of integrating phase-field simulations with active learning to significantly accelerate inverse process design for microstructure engineering, compared with using phase-field simulations alone.","PeriodicalId":19342,"journal":{"name":"npj Computational Materials","volume":"5 1","pages":""},"PeriodicalIF":9.7,"publicationDate":"2026-08-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148769237","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}

引用次数: 0

Chemical short-range order regulates hydrogen energetics and hydrogen-dislocation interactions in CoNiV 化学短程顺序调节CoNiV中氢能量和氢位错相互作用

IF 9.7 1区材料科学

npj Computational Materials Pub Date : 2026-08-20 DOI: 10.1038/s41524-026-02274-9

Beihan Chen, Dalia M. Sayed Ahmed, Yang Yang, Miaomiao Jin

{"title":"Chemical short-range order regulates hydrogen energetics and hydrogen-dislocation interactions in CoNiV","authors":"Beihan Chen, Dalia M. Sayed Ahmed, Yang Yang, Miaomiao Jin","doi":"10.1038/s41524-026-02274-9","DOIUrl":"https://doi.org/10.1038/s41524-026-02274-9","url":null,"abstract":"Chemical short-range order (CSRO) has emerged as a critical structural feature in concentrated alloys, yet its coupling with hydrogen remains an active discussion. Here, we develop a machine-learning interatomic potential for the Co-Ni-V-H system and investigate how CSRO regulates hydrogen energetics and dislocation behavior in CoNiV, an alloy with reported strong resistance to hydrogen embrittlement. We identify strong V-centered ordering that suppresses V–V clustering and remains robust up to 1000 K, with negligible changes upon the addition of 1.0 at.% hydrogen. Compared to a chemically random alloy, the ordered state exhibits higher average hydrogen solution energies and a reduced population of strongly binding sites, suggesting lower bulk hydrogen uptake. At partial dislocations, hydrogen preferentially segregates to tensile core regions, where the dislocation core acts as a shallow, reversible trap, with a much weaker effect compared to chemical trapping states. These results show that robust local chemical order plays a critical role in affecting H uptake, partitioning, and dislocation trapping in concentrated CoNiV alloys.","PeriodicalId":19342,"journal":{"name":"npj Computational Materials","volume":"159 1","pages":""},"PeriodicalIF":9.7,"publicationDate":"2026-08-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148769239","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}

引用次数: 0

Efficient quantum implementation of dynamical mean field theory for correlated materials 相关材料动力学平均场理论的高效量子实现

IF 9.7 1区材料科学

npj Computational Materials Pub Date : 2026-08-17 DOI: 10.1038/s41524-026-02289-2

Norman Hogan, Efekan Kökcü, Thomas Steckmann, Liam P. Doak, Carlos Mejuto-Zaera, Daan Camps, Roel Van Beeumen, Wibe A. de Jong, A. F. Kemper

{"title":"Efficient quantum implementation of dynamical mean field theory for correlated materials","authors":"Norman Hogan, Efekan Kökcü, Thomas Steckmann, Liam P. Doak, Carlos Mejuto-Zaera, Daan Camps, Roel Van Beeumen, Wibe A. de Jong, A. F. Kemper","doi":"10.1038/s41524-026-02289-2","DOIUrl":"https://doi.org/10.1038/s41524-026-02289-2","url":null,"abstract":"The accurate theoretical description of materials with strongly correlated electrons is a formidable challenge in condensed matter physics and computational chemistry. Dynamical Mean Field Theory (DMFT) is a successful approach that predicts behaviors of such systems by incorporating some of the correlated behavior using an impurity model, but it is limited by the need to calculate the impurity Green’s function. This work proposes a framework for DMFT calculations on quantum computers, focusing on near-term applications. It leverages the structure of the impurity problem, combining a low-rank Gaussian subspace representation of the ground state and a compressed, short-depth quantum circuit that joins state preparation with time evolution to compute Green’s functions. We demonstrate the convergence of the DMFT algorithm using the Gaussian subspace in a noise-free setting, and show the hardware viability of circuit compression by extracting the impurity Green’s function on IBM quantum processors for a single impurity coupled to three bath orbitals (8 qubits, 1 ancilla). We discuss potential paths toward realizing this quantum computing use case in materials science.","PeriodicalId":19342,"journal":{"name":"npj Computational Materials","volume":"42 1","pages":""},"PeriodicalIF":9.7,"publicationDate":"2026-08-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148769241","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}

引用次数: 0

Multi-fidelity computational screening of high-entropy MBenes for electrochemical CO2 reduction 电化学CO2还原高熵MBenes的多保真度计算筛选

IF 9.7 1区材料科学

npj Computational Materials Pub Date : 2026-08-17 DOI: 10.1038/s41524-026-02287-4

Sree Harsha Bharadwaj H, Raghavan Ranganathan

{"title":"Multi-fidelity computational screening of high-entropy MBenes for electrochemical CO2 reduction","authors":"Sree Harsha Bharadwaj H, Raghavan Ranganathan","doi":"10.1038/s41524-026-02287-4","DOIUrl":"https://doi.org/10.1038/s41524-026-02287-4","url":null,"abstract":"High-entropy MBenes (HE-MBenes) are two-dimensional transition metal borides whose metal sublattice is occupied by five or more principal elements, forming a largely unexplored class of electrocatalysts for CO2 electroreduction. These materials combine the lone-pair-driven CO2 activation intrinsic to MBenes with the configurational stabilization and chemical diversity of high-entropy systems, circumventing the linear scaling relationships (LSRs) that constrain monometallic surfaces. We report a multi-fidelity computational screening of all 56 equiatomic quinary HE-MBene compositions from a {Ti, V, Cr, Mo, Nb, Ta, Zr, Hf} elemental pool. A sequential funnel integrating DFT relaxation, formation-energy filtering, PDOS-guided active-site identification, and computational hydrogen electrode (CHE) free-energy profiling, accelerated by a MACE machine-learning interatomic potential across 1375 surface environments, narrows the space from 56 to 18 viable catalysts. Three compositions, CrNbZrMoTiB5, MoZrHfNbCrB5, and MoZrHfTaCrB5, exhibit spontaneously downhill CO2-to-CO free-energy profiles at zero applied potential (UL = 0.00 V vs. RHE), surpassing all previously reported CHE-computed limiting potentials for MBene and high-entropy-alloy catalysts. Bader charge analysis reveals that Hf/Zr-mediated electron donation to Cr-centred active sites stabilizes the rate-determining COOH* intermediate, decoupling LSRs. These findings establish HE-MBenes as a generalizable catalyst platform and provide a transferable multi-fidelity framework for multicomponent 2D materials discovery.","PeriodicalId":19342,"journal":{"name":"npj Computational Materials","volume":"93 1","pages":""},"PeriodicalIF":9.7,"publicationDate":"2026-08-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148769240","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}

引用次数: 0

Vibrational power spectra as a tool to benchmark universal machine-learning interatomic potentials for molecular systems: the OMOL-1k-MD data set 作为分子系统通用机器学习原子间势基准测试工具的振动功率谱：omol - k- md数据集

IF 9.7 1区材料科学

npj Computational Materials Pub Date : 2026-08-15 DOI: 10.1038/s41524-026-02286-5

Maximilian A. Bechtel, Julien Steffen

{"title":"Vibrational power spectra as a tool to benchmark universal machine-learning interatomic potentials for molecular systems: the OMOL-1k-MD data set","authors":"Maximilian A. Bechtel, Julien Steffen","doi":"10.1038/s41524-026-02286-5","DOIUrl":"https://doi.org/10.1038/s41524-026-02286-5","url":null,"abstract":"We present OMOL-1k-MD, a new dataset for the benchmark and training of universal machine-learning interatomic potentials (uMLIPs). It contains three independent ab initio molecular dynamics (AIMD) trajectories of 10 ps each for 1000 arbitrarily chosen neutral closed shell molecular systems from the OMOL25 dataset, calculated at the PBE level of theory at 300 K. In contrast to most other MLIP training or benchmark data sets, which are focused on local minimum structures, OMOL-1k-MD contains a total of 30 million structures sampled at room temperature. We benchmark several popular uMLIPs regarding their ability to reasonably reproduce the DFT dynamics. To this end, we compare the vibrational power spectra of the OMOL-1k-MD trajectories with uMLIP trajectories for the same set of structures. The comparison considers either the relative peak positions of different vibrations or the integrated vibrational free energies. This procedure enables a thorough benchmark and ranking of uMLIP performances apart from local minima. We found that the MACE MP-0 uMLIP performed best, followed by SevenNet and Orb OMat. The GFN2-xTB semiempirical method, however, performed better than all tested uMLIPs, proving the need for more off-equilibrium data in the parametrization of uMLIPs, where OMOL-1k-MD can be of high use.","PeriodicalId":19342,"journal":{"name":"npj Computational Materials","volume":"1 1","pages":""},"PeriodicalIF":9.7,"publicationDate":"2026-08-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148729034","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}

引用次数: 0

Universal alignment of defect thermodynamics in hematite 赤铁矿缺陷热力学的普遍对准

IF 9.7 1区材料科学

npj Computational Materials Pub Date : 2026-08-15 DOI: 10.1038/s41524-026-02281-w

Hao Chen, Baptiste Bienvenu, Mira Todorova, Christoph Freysoldt, Jörg Neugebauer

{"title":"Universal alignment of defect thermodynamics in hematite","authors":"Hao Chen, Baptiste Bienvenu, Mira Todorova, Christoph Freysoldt, Jörg Neugebauer","doi":"10.1038/s41524-026-02281-w","DOIUrl":"https://doi.org/10.1038/s41524-026-02281-w","url":null,"abstract":"First-principles defect calculations in hematite (α-Fe2O3) suffer from a >2 eV spread in reported formation energies, arising from (i) metastable or delocalized charge states and (ii) inconsistent error cancellation with chemical-potential boundaries across studies. We introduce a charge-localization protocol based on the Fe-3d occupation matrix that enforces physically meaningful small-polaron configurations and eliminates the spurious loss of intermediate charge states; the resulting thermodynamic charge-transition levels are essentially independent of computational parameters. To establish a common energy scale, we use the crossing point of the fully ionized Frenkel pair (({V}_{{rm{Fe}}}^{3-}) and ({{rm{Fe}}}_{i}^{3+})) as an intrinsic anchor and align the chemical potentials to parameter-independent experimental formation enthalpies. This reduces the inter-study deviation by ~80%, and reveals that the remaining absolute shifts are dominated by VBM displacement amplified by the large nominal defect charges. The framework provides a reproducible benchmark for hematite defect energetics and is directly transferable to other correlated oxides.","PeriodicalId":19342,"journal":{"name":"npj Computational Materials","volume":"42 1","pages":""},"PeriodicalIF":9.7,"publicationDate":"2026-08-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148729035","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}

引用次数: 0

MatUQ: a benchmark for uncertainty-aware out-of-distribution materials property prediction with graph neural networks MatUQ：一种基于图神经网络的不确定性感知非分布材料属性预测基准

IF 9.7 1区材料科学

npj Computational Materials Pub Date : 2026-08-15 DOI: 10.1038/s41524-026-02272-x

Liqin Tan, Xiean Wang, Yuexin Zou, Pin Chen, Qingsong Zou

{"title":"MatUQ: a benchmark for uncertainty-aware out-of-distribution materials property prediction with graph neural networks","authors":"Liqin Tan, Xiean Wang, Yuexin Zou, Pin Chen, Qingsong Zou","doi":"10.1038/s41524-026-02272-x","DOIUrl":"https://doi.org/10.1038/s41524-026-02272-x","url":null,"abstract":"Reliable uncertainty quantification (UQ) for graph neural networks (GNNs) under out-of-distribution (OOD) shifts remains insufficiently characterized in materials discovery. Existing benchmarks based on random splits can overestimate model reliability by underrepresenting structural extrapolation challenges. Here we introduce MatUQ, a benchmark built on structure-aware Smooth Overlap of Atomic Positions Leave-One-Cluster-Out (SOAP-LOCO) splitting, together with a training protocol that combines Deep Evidential Regression (DER) with dropout regularization, for evaluating GNN reliability under structural distribution shifts. Through systematic experiments spanning six materials datasets, twelve GNN architectures, and eight UQ strategies, we find that predictive accuracy and uncertainty quality are distinct capabilities that tend to decouple under OOD evaluation, and that uncertainty-metric leadership is largely non-transferable across datasets and target properties. We further find that as training data become scarce, the optimal strategy shifts from evidential-containing hybrids toward pure ensemble variance across all evaluated architectures, with the architecture holding the distributional optimum shifting correspondingly. Standalone evidential regression rarely attains per-model optima and benefits from hybrid pairing only under specific combinations of data density, inductive bias, and target metric. Monte Carlo dropout is less effective as a standalone uncertainty estimator under the tested configurations but can contribute within hybrid schemes. Overall, MatUQ provides a more rigorous benchmark for assessing uncertainty-aware GNNs in OOD materials discovery.","PeriodicalId":19342,"journal":{"name":"npj Computational Materials","volume":"1 1","pages":""},"PeriodicalIF":9.7,"publicationDate":"2026-08-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148728867","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}

引用次数: 0