Computational Materials Science最新文献_第2页

Mapping and characterization of surface-dependent electronic properties and morphological changes in the cubic phase of crystalline perovskite CsPbBr3 晶体钙钛矿CsPbBr3立方相表面依赖电子性质和形态变化的制图和表征

IF 3.3 3区材料科学 Q2 MATERIALS SCIENCE, MULTIDISCIPLINARY

Computational Materials Science

Pub Date : 2026-10-01 Epub Date: 2026-01-06 DOI: 10.1016/j.commatsci.2025.114477

Marcos de C. Leite , Gabriel X. Pereira , Lucas M. Farigliano , Gustavo M. Dalpian , Juan Andrés , Amanda F. Gouveia

Lead halide perovskites are attracting considerable interest across a wide range of applications, from gas sensors to energy conversion and utilization. Here, the cubic phase of crystalline perovskite CsPbBr

_{3}

is proposed as a probe to shed light on subtle structural and electronic changes that control surface-dependent electronic properties and morphology, using a computational approach based on density functional theory calculations. We carried out first-principles density functional theory calculations to obtain the surface-dependent properties (band structures, density of states, and surface energies) of low Miller-index (001), (110), and (111) surfaces with different terminations of CsPbBr

_{3}

. Additionally, the atomic arrangements and stability of these surfaces were characterized to provide a close match between experimental field-emission scanning electron microscopy images and computational simulations. We demonstrate a practical application of the Wulff construction by leveraging computed surface energies to determine a complete map of available morphologies that are consistent with experimental observations. Our findings reveal how the exposed surfaces on the morphology influence the electronic properties, elucidating the atomic-level synergy between surface-dependent electronic properties and morphological changes in CsPbBr

_{3}

, and providing a theoretical foundation and design principles for enhancing perovskite stability through surface engineering.

卤化铅钙钛矿在从气体传感器到能量转换和利用的广泛应用中引起了相当大的兴趣。本文采用基于密度泛函理论计算的计算方法，提出了晶体钙钛矿CsPbBr3的立方相作为探针，以揭示控制表面依赖电子性质和形貌的细微结构和电子变化。我们通过第一性原理密度泛函理论计算得到了具有不同末端的CsPbBr3的低米勒指数（001）、（110）和（111）表面的表面依赖性质（能带结构、态密度和表面能）。此外，对这些表面的原子排列和稳定性进行了表征，以提供实验场发射扫描电子显微镜图像和计算模拟之间的密切匹配。我们展示了Wulff结构的实际应用，利用计算表面能来确定与实验观察一致的可用形态的完整地图。本研究揭示了CsPbBr3表面暴露对其电子性能的影响，阐明了CsPbBr3表面依赖的电子性能与形态变化之间的原子水平协同作用，为通过表面工程增强钙钛矿稳定性提供了理论基础和设计原则。

{"title":"Mapping and characterization of surface-dependent electronic properties and morphological changes in the cubic phase of crystalline perovskite CsPbBr3","authors":"Marcos de C. Leite , Gabriel X. Pereira , Lucas M. Farigliano , Gustavo M. Dalpian , Juan Andrés , Amanda F. Gouveia","doi":"10.1016/j.commatsci.2025.114477","DOIUrl":"10.1016/j.commatsci.2025.114477","url":null,"abstract":"<div><div>Lead halide perovskites are attracting considerable interest across a wide range of applications, from gas sensors to energy conversion and utilization. Here, the cubic phase of crystalline perovskite CsPbBr<span><math><msub><mrow></mrow><mrow><mn>3</mn></mrow></msub></math></span> is proposed as a probe to shed light on subtle structural and electronic changes that control surface-dependent electronic properties and morphology, using a computational approach based on density functional theory calculations. We carried out first-principles density functional theory calculations to obtain the surface-dependent properties (band structures, density of states, and surface energies) of low Miller-index (001), (110), and (111) surfaces with different terminations of CsPbBr<span><math><msub><mrow></mrow><mrow><mn>3</mn></mrow></msub></math></span>. Additionally, the atomic arrangements and stability of these surfaces were characterized to provide a close match between experimental field-emission scanning electron microscopy images and computational simulations. We demonstrate a practical application of the Wulff construction by leveraging computed surface energies to determine a complete map of available morphologies that are consistent with experimental observations. Our findings reveal how the exposed surfaces on the morphology influence the electronic properties, elucidating the atomic-level synergy between surface-dependent electronic properties and morphological changes in CsPbBr<span><math><msub><mrow></mrow><mrow><mn>3</mn></mrow></msub></math></span>, and providing a theoretical foundation and design principles for enhancing perovskite stability through surface engineering.</div></div>","PeriodicalId":10650,"journal":{"name":"Computational Materials Science","volume":"264 ","pages":"Article 114477"},"PeriodicalIF":3.3,"publicationDate":"2026-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145923412","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"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

Molecular dynamic simulation study on 3MOAl2O3 3SiO2 [M = Ba, Sr, Ca, Mg, Zn and Mn] glasses 3MOAl2O3 3SiO2 [M = Ba, Sr, Ca, Mg， Zn和Mn]玻璃的分子动力学模拟研究

IF 3.3 3区材料科学 Q2 MATERIALS SCIENCE, MULTIDISCIPLINARY

Computational Materials Science

Pub Date : 2026-10-01 Epub Date: 2026-01-13 DOI: 10.1016/j.commatsci.2026.114486

Veeramohan Rao M

Aluminosilicate glasses incorporating alkaline earth metals, Zn, and Mn are of significant interest within the realms of material science and geoscience. The structural implications of the MO/Al₂O₃ ratio, specifically at a value of 3, in aluminosilicate glasses remain incompletely elucidated. In the present study, molecular dynamic simulations are employed to investigate the influence of cation field strength on the structural and elastic characteristics of aluminosilicate glasses. The cation field strength is observed to induce alterations in structural properties attributes such as bond length, coordination number, bond angle distribution, and the presence of oxygen species including FO, NBO, BO, and TBO. The glass transition temperature and elastic constants are determined through these simulations. It is found that an increase in cation field strength correlates with an elevation in elastic constants and a reduction in glass transition temperature. These findings provide atomic-scale insights into the effects of cation field strength on the properties of glasses.

含有碱土金属、锌和锰的铝硅酸盐玻璃在材料科学和地球科学领域具有重要意义。MO/Al2O3比在铝硅酸盐玻璃中的结构意义，特别是当其值为3时，仍未完全阐明。本文采用分子动力学模拟方法研究了阳离子场强度对铝硅酸盐玻璃结构和弹性特性的影响。观察到阳离子场强会引起结构属性的改变，如键长、配位数、键角分布以及氧（包括FO、NBO、BO和TBO）的存在。通过这些模拟确定了玻璃化转变温度和弹性常数。发现阳离子场强的增加与弹性常数的升高和玻璃化转变温度的降低有关。这些发现提供了原子尺度上对阳离子场强对玻璃性质影响的见解。

引用次数: 0

Development and validation of interatomic potential for Sc and Al–Sc alloys: Thermodynamics, solidification, and intermetallic ordering Sc和Al-Sc合金原子间势的发展和验证：热力学、凝固和金属间有序

IF 3.3 3区材料科学 Q2 MATERIALS SCIENCE, MULTIDISCIPLINARY

Computational Materials Science

Pub Date : 2026-10-01 Epub Date: 2026-01-13 DOI: 10.1016/j.commatsci.2025.114443

Avik Mahata

We present a second-nearest-neighbor Modified Embedded Atom Method (2NN–MEAM) potential for Scandium (Sc) and Aluminum-Scandium (Al–Sc) alloys that unifies cohesive, thermodynamic, and solidification behavior within a single transferable framework. The Sc component accurately reproduces cohesive energy, lattice constants, defect energetics, and the experimental melting point obtained from two-phase coexistence, demonstrating reliable description of both hcp and liquid phases. The Al–Sc binary interaction parameters were fitted using the L1

_{2}

–Al

_{3}

Sc reference and benchmarked against first-principles and calorimetric data. The potential reproduces the strong negative formation enthalpy of Al

_{3}

Sc (–0.45 eV atom⁻¹), correct relative stability of competing phases, and realistic elastic properties. Mixing enthalpies of the liquid alloy agree with ideal-associated-solution and CALPHAD models, confirming that the potential captures exothermic Al–Sc association in the melt. Molecular-dynamics simulations of solidification reveal the expected temperature and composition dependence of homogeneous nucleation. Pure Al crystallizes readily, while Al–1 at.% Sc exhibits a longer incubation and slower growth at the same absolute temperature due to reduced undercooling and solute drag. Within the alloy, ordered Al

_{3}

Sc-type L1

_{2}

embryos appear spontaneously, with Sc atoms occupying cube-corner (B) sites surrounded by twelve Al neighbors. Energy–volume trajectories confirm that the potential links thermodynamics to microstructural evolution. Overall, the developed 2NN–MEAM potential provides a quantitatively grounded basis for modeling melting, solidification, and intermetallic ordering in Sc and Al–Sc systems, enabling future multicomponent alloy design and large-scale nucleation studies.

我们提出了钪（Sc）和铝-钪（Al-Sc）合金的第二近邻修正嵌入原子法（2NN-MEAM）潜力，该方法在一个可转移的框架内统一了内聚、热力学和凝固行为。Sc组分精确地再现了从两相共存得到的内聚能、晶格常数、缺陷能量和实验熔点，证明了对hcp和液相的可靠描述。采用L12-Al3Sc标准拟合Al-Sc二元相互作用参数，并以第一性原理和量热数据为基准。该势再现了Al3Sc （-0.45 eV原子−1）的强负生成焓，正确的竞争相相对稳定性和真实的弹性性质。液态合金的混合焓与理想相关溶液和CALPHAD模型一致，证实了该势捕获了熔体中的放热Al-Sc结合。凝固的分子动力学模拟揭示了均匀形核的预期温度和成分依赖性。纯铝容易结晶，而Al - 1不易结晶。在相同的绝对温度下，由于过冷和溶质阻力的减少，Sc的培养时间更长，生长速度较慢。在合金内部，有序的al3sc型L12胚胎自发出现，Sc原子占据立方体角(B)的位置，周围环绕着12个Al邻居。能量-体积轨迹证实了热力学与微观结构演化之间的潜在联系。总的来说，开发的2NN-MEAM潜力为Sc和Al-Sc体系的熔化、凝固和金属间有序建模提供了定量基础，使未来的多组分合金设计和大规模成核研究成为可能。

{"title":"Development and validation of interatomic potential for Sc and Al–Sc alloys: Thermodynamics, solidification, and intermetallic ordering","authors":"Avik Mahata","doi":"10.1016/j.commatsci.2025.114443","DOIUrl":"10.1016/j.commatsci.2025.114443","url":null,"abstract":"<div><div>We present a second-nearest-neighbor Modified Embedded Atom Method (2NN–MEAM) potential for Scandium (Sc) and Aluminum-Scandium (Al–Sc) alloys that unifies cohesive, thermodynamic, and solidification behavior within a single transferable framework. The Sc component accurately reproduces cohesive energy, lattice constants, defect energetics, and the experimental melting point obtained from two-phase coexistence, demonstrating reliable description of both hcp and liquid phases. The Al–Sc binary interaction parameters were fitted using the L1<span><math><msub><mrow></mrow><mrow><mn>2</mn></mrow></msub></math></span>–Al<span><math><msub><mrow></mrow><mrow><mn>3</mn></mrow></msub></math></span>Sc reference and benchmarked against first-principles and calorimetric data. The potential reproduces the strong negative formation enthalpy of Al<span><math><msub><mrow></mrow><mrow><mn>3</mn></mrow></msub></math></span>Sc (–0.45 eV atom<sup>−1</sup>), correct relative stability of competing phases, and realistic elastic properties. Mixing enthalpies of the liquid alloy agree with ideal-associated-solution and CALPHAD models, confirming that the potential captures exothermic Al–Sc association in the melt. Molecular-dynamics simulations of solidification reveal the expected temperature and composition dependence of homogeneous nucleation. Pure Al crystallizes readily, while Al–1 at.% Sc exhibits a longer incubation and slower growth at the same absolute temperature due to reduced undercooling and solute drag. Within the alloy, ordered Al<span><math><msub><mrow></mrow><mrow><mn>3</mn></mrow></msub></math></span>Sc-type L1<span><math><msub><mrow></mrow><mrow><mn>2</mn></mrow></msub></math></span> embryos appear spontaneously, with Sc atoms occupying cube-corner (B) sites surrounded by twelve Al neighbors. Energy–volume trajectories confirm that the potential links thermodynamics to microstructural evolution. Overall, the developed 2NN–MEAM potential provides a quantitatively grounded basis for modeling melting, solidification, and intermetallic ordering in Sc and Al–Sc systems, enabling future multicomponent alloy design and large-scale nucleation studies.</div></div>","PeriodicalId":10650,"journal":{"name":"Computational Materials Science","volume":"264 ","pages":"Article 114443"},"PeriodicalIF":3.3,"publicationDate":"2026-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145974143","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"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

The effect of disorder and mechanical deformation on the vibrational density of states in 2D silica 无序和机械变形对二维二氧化硅中态振动密度的影响

IF 3.3 3区材料科学 Q2 MATERIALS SCIENCE, MULTIDISCIPLINARY

Computational Materials Science

Pub Date : 2026-10-01 Epub Date: 2026-01-13 DOI: 10.1016/j.commatsci.2025.114425

Tobias Focks, Bernd Markert, Franz Bamer

The vibrational density of states provides essential information about the characteristics of glassy materials. In particular, both mechanically and structurally induced changes in the vibrational density of states, as well as the emergence of quasilocalized modes, remain a topic of discussion. Thus, this paper aims to shed light on the vibrational density of states in model network glasses. The results show its dependence on the level of disorder and the evolution of the vibrational density of states under shear deformation. We have found that the tendency for eigenvalues to change depends on heterogeneity, deformation state, and the occurrence of rearrangement processes in the network.

态的振动密度提供了关于玻璃材料特性的基本信息。特别是，机械和结构引起的状态振动密度的变化，以及准局域模式的出现，仍然是一个讨论的话题。因此，本文旨在揭示模型网络玻璃中状态的振动密度。结果表明，剪切变形作用下，振动密度的变化与失序程度和状态振动密度的演化有关。我们发现特征值的变化趋势取决于网络的异质性、变形状态和重排过程的发生。

引用次数: 0

Deep learning detection of topological defects in confined two-dimensional nematics 受限二维向列图拓扑缺陷的深度学习检测

IF 3.3 3区材料科学 Q2 MATERIALS SCIENCE, MULTIDISCIPLINARY

Computational Materials Science

Pub Date : 2026-10-01 Epub Date: 2026-01-14 DOI: 10.1016/j.commatsci.2026.114508

Ignacio Palos-Reynoso , Humberto Híjar

We present a neural network-based algorithm for the identification and classification of topological defects in two-dimensional nematic liquid crystals confined within square geometries. The nematic configurations are generated through Nematic-Multiparticle Collision Dynamics, a mesoscopic simulation method that captures both hydrodynamic and orientational fluctuations. Our supervised learning framework is trained on synthetic images labeled with topological defects of positive and negative charge. In the inference stage, we employ local winding number estimations to propose candidate defect locations, which are then evaluated by the neural network to determine their authenticity and type. The algorithm achieves robust classification performance, with a macro-averaged

F_{1}

score of 0.92, indicating balanced precision and recall across all defect classes, with most misclassifications arising from director field fluctuations near the corners of the confinement domain. Beyond static identification, our method enables the temporal tracking of defect dynamics, including annihilation events. This work demonstrates the potential of deep learning tools to extract and quantify topological information in fluctuating soft matter systems.

提出了一种基于神经网络的二维向列液晶拓扑缺陷识别与分类算法。向列型结构是通过向列多粒子碰撞动力学生成的，这是一种介观模拟方法，可以捕获流体力学和方向波动。我们的监督学习框架是在带有正负电荷拓扑缺陷的合成图像上进行训练的。在推理阶段，我们使用局部圈数估计来提出候选缺陷位置，然后通过神经网络对候选缺陷位置进行评估以确定其真实性和类型。该算法具有鲁棒的分类性能，宏观平均F1分数为0.92，表明所有缺陷类别的精度和召回率平衡，大多数错误分类是由约束域角附近的方向场波动引起的。除了静态识别之外，我们的方法还支持缺陷动态的时间跟踪，包括湮灭事件。这项工作证明了深度学习工具在波动软物质系统中提取和量化拓扑信息的潜力。

{"title":"Deep learning detection of topological defects in confined two-dimensional nematics","authors":"Ignacio Palos-Reynoso , Humberto Híjar","doi":"10.1016/j.commatsci.2026.114508","DOIUrl":"10.1016/j.commatsci.2026.114508","url":null,"abstract":"<div><div>We present a neural network-based algorithm for the identification and classification of topological defects in two-dimensional nematic liquid crystals confined within square geometries. The nematic configurations are generated through Nematic-Multiparticle Collision Dynamics, a mesoscopic simulation method that captures both hydrodynamic and orientational fluctuations. Our supervised learning framework is trained on synthetic images labeled with topological defects of positive and negative charge. In the inference stage, we employ local winding number estimations to propose candidate defect locations, which are then evaluated by the neural network to determine their authenticity and type. The algorithm achieves robust classification performance, with a macro-averaged <span><math><msub><mrow><mi>F</mi></mrow><mrow><mn>1</mn></mrow></msub></math></span> score of 0.92, indicating balanced precision and recall across all defect classes, with most misclassifications arising from director field fluctuations near the corners of the confinement domain. Beyond static identification, our method enables the temporal tracking of defect dynamics, including annihilation events. This work demonstrates the potential of deep learning tools to extract and quantify topological information in fluctuating soft matter systems.</div></div>","PeriodicalId":10650,"journal":{"name":"Computational Materials Science","volume":"264 ","pages":"Article 114508"},"PeriodicalIF":3.3,"publicationDate":"2026-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145974222","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"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

A framework for automated extraction and validation of experimental arguments in zeolite synthesis 沸石合成中实验参数的自动提取和验证框架

IF 3.3 3区材料科学 Q2 MATERIALS SCIENCE, MULTIDISCIPLINARY

Computational Materials Science

Pub Date : 2026-10-01 Epub Date: 2026-01-12 DOI: 10.1016/j.commatsci.2026.114495

Ziming Yu, Song He, Wenli Du

Natural language processing (NLP), particularly information extraction techniques, has enabled the automatic extraction of structured experimental procedures from synthesized text, significantly accelerating the automation of materials research and development. However, two key challenges in zeolite synthesis hinder the advancement and application of these cutting-edge technologies: (1) Dense and lengthy experimental arguments pose challenges for models in comprehensively extracting all relevant information. (2) The reliance on manual verification for extracted experimental arguments poses a challenge to fully automating experimental workflows. Furthermore, while large language models demonstrate impressive general capabilities, their practical deployment in specialized domains is often constrained by limited task-specific accuracy, high training and inference costs, and latency. To address these challenges, we propose a novel framework that enhances argument extraction while significantly reducing verification time. We first develop a compact, domain-specific multi-event extraction model TabAMR that integrates abstract meaning representation (AMR) with slotted table to capture dense semantics and long-range dependencies. TabAMR attains an average F1 score of 90.39 for argument classification, demonstrating improvements of 1.43 and 0.96 over the SOTA model in extracting argument-dense samples and samples with long argument lengths, respectively. Furthermore, we design Experimental Argument Error Judgement (EAEJ) method for automatically judging the correctness of experimental arguments extracted by TabAMR. EAEJ achieves an average accuracy of 93.63% and requires only one-seventh of the time needed for traditional machine reading approach. The framework enhances machine reading capabilities in the field of zeolite synthesis and significantly reducing the need for human intervention, thereby advancing the automation of zeolite synthesis and facilitating the discovery of new zeolites.

自然语言处理（NLP），特别是信息提取技术，使得从合成文本中自动提取结构化实验过程成为可能，极大地加速了材料研究和开发的自动化。然而，沸石合成的两个关键挑战阻碍了这些前沿技术的进步和应用：(1)密集和冗长的实验参数给模型全面提取所有相关信息带来了挑战。(2)依赖人工验证提取的实验参数对实验工作流程的完全自动化提出了挑战。此外，虽然大型语言模型展示了令人印象深刻的一般功能，但它们在专门领域的实际部署通常受到特定于任务的有限准确性、高训练和推理成本以及延迟的限制。为了解决这些挑战，我们提出了一个新的框架，可以在显著减少验证时间的同时增强论点提取。我们首先开发了一个紧凑的，特定于领域的多事件提取模型TabAMR，它将抽象意义表示（AMR）与槽表集成在一起，以捕获密集语义和远程依赖关系。TabAMR在参数分类方面的平均F1得分为90.39，在提取参数密集样本和长参数长度样本方面分别比SOTA模型提高了1.43和0.96。在此基础上，设计了实验论证错误判断（EAEJ）方法，对TabAMR提取的实验论证的正确性进行自动判断。EAEJ的平均准确率为93.63%，所需时间仅为传统机器阅读方法的七分之一。该框架增强了沸石合成领域的机器读取能力，大大减少了人工干预的需要，从而推进了沸石合成的自动化，促进了新沸石的发现。

{"title":"A framework for automated extraction and validation of experimental arguments in zeolite synthesis","authors":"Ziming Yu, Song He, Wenli Du","doi":"10.1016/j.commatsci.2026.114495","DOIUrl":"10.1016/j.commatsci.2026.114495","url":null,"abstract":"<div><div>Natural language processing (NLP), particularly information extraction techniques, has enabled the automatic extraction of structured experimental procedures from synthesized text, significantly accelerating the automation of materials research and development. However, two key challenges in zeolite synthesis hinder the advancement and application of these cutting-edge technologies: (1) Dense and lengthy experimental arguments pose challenges for models in comprehensively extracting all relevant information. (2) The reliance on manual verification for extracted experimental arguments poses a challenge to fully automating experimental workflows. Furthermore, while large language models demonstrate impressive general capabilities, their practical deployment in specialized domains is often constrained by limited task-specific accuracy, high training and inference costs, and latency. To address these challenges, we propose a novel framework that enhances argument extraction while significantly reducing verification time. We first develop a compact, domain-specific multi-event extraction model TabAMR that integrates abstract meaning representation (AMR) with slotted table to capture dense semantics and long-range dependencies. TabAMR attains an average F1 score of 90.39 for argument classification, demonstrating improvements of 1.43 and 0.96 over the SOTA model in extracting argument-dense samples and samples with long argument lengths, respectively. Furthermore, we design Experimental Argument Error Judgement (EAEJ) method for automatically judging the correctness of experimental arguments extracted by TabAMR. EAEJ achieves an average accuracy of 93.63% and requires only one-seventh of the time needed for traditional machine reading approach. The framework enhances machine reading capabilities in the field of zeolite synthesis and significantly reducing the need for human intervention, thereby advancing the automation of zeolite synthesis and facilitating the discovery of new zeolites.</div></div>","PeriodicalId":10650,"journal":{"name":"Computational Materials Science","volume":"264 ","pages":"Article 114495"},"PeriodicalIF":3.3,"publicationDate":"2026-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145974223","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"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

Physically interpretable models for predicting lattice parameters, tunnel geometry, and symmetry in hollandite-type materials 预测荷兰石型材料中晶格参数、隧道几何形状和对称性的物理可解释模型

IF 3.3 3区材料科学 Q2 MATERIALS SCIENCE, MULTIDISCIPLINARY

Computational Materials Science

Pub Date : 2026-10-01 Epub Date: 2026-01-09 DOI: 10.1016/j.commatsci.2026.114494

Mingyang Zhao , Zhisen Feng , Bin Lei , Xiao Chen , Zhenfeng Tong

Hollandite-type materials are structurally versatile ceramics used in nuclear waste immobilization, ion transport, and energy storage, yet quantitative and physically interpretable composition-to-structure prediction remains limited. In this work, we develop a suite of interpretable multivariate regression models, benchmarked against machine learning approaches, to predict key crystallographic features of hollandite materials, including lattice constants a and c, tunnel bottleneck size (d-d), and crystallographic symmetry. Using physically motivated descriptors such as ionic-radius mismatch, oxidation states, electronegativity, and tunnel occupancy, we reveal a pronounced directional anisotropy in composition-structure relationships. Lattice constant a is accurately captured by a multivariate linear regression (MLR) model dominated by steric mismatch effects, whereas lattice constant c and d-d require compact multivariate polynomial regression (MPR) models to describe essential nonlinear and interaction effects arising from coupled steric and electronic contributions. A regression-based symmetry model further enables reliable classification of tetragonal and monoclinic phases using physically interpretable descriptors alone. Benchmarking against artificial neural networks (ANN) and support vector regression (SVG) confirms that the interpretable regression models achieve comparable accuracy while offering direct mechanistic insight. The validated regression models are subsequently applied to predict structural features of previously uncharacterized hollandite compositions, including actinide-substituted titanates and Mn- and Sn-based analogues. Overall, this work establishes a data-efficient, mechanism-aware framework for composition-driven structure prediction in tunnel-structured oxides, providing a transferable strategy for accelerated materials design.

荷兰石型材料是一种结构多样的陶瓷，用于核废料固定化、离子传输和能量储存，但定量和物理上可解释的成分-结构预测仍然有限。在这项工作中，我们开发了一套可解释的多元回归模型，以机器学习方法为基准，预测荷兰石材料的关键晶体特征，包括晶格常数a和c，隧道瓶颈尺寸（d-d）和晶体对称性。利用离子半径失配、氧化态、电负性和隧道占用等物理动机描述符，我们揭示了成分-结构关系中明显的方向各向异性。晶格常数a是由空间错配效应主导的多元线性回归（MLR）模型准确捕获的，而晶格常数c和d-d需要紧凑的多元多项式回归（MPR）模型来描述由耦合的空间和电子贡献引起的基本非线性和相互作用效应。基于回归的对称模型进一步实现了四方相和单斜相的可靠分类，仅使用物理可解释的描述符。针对人工神经网络（ANN）和支持向量回归（SVG）的基准测试证实，可解释的回归模型在提供直接机制洞察力的同时达到了相当的准确性。经过验证的回归模型随后被应用于预测以前未表征的荷兰酸盐成分的结构特征，包括锕系取代的钛酸盐和锰基和锡基类似物。总的来说，这项工作为隧道结构氧化物的成分驱动结构预测建立了一个数据高效、机制感知的框架，为加速材料设计提供了一种可转移的策略。

{"title":"Physically interpretable models for predicting lattice parameters, tunnel geometry, and symmetry in hollandite-type materials","authors":"Mingyang Zhao , Zhisen Feng , Bin Lei , Xiao Chen , Zhenfeng Tong","doi":"10.1016/j.commatsci.2026.114494","DOIUrl":"10.1016/j.commatsci.2026.114494","url":null,"abstract":"<div><div>Hollandite-type materials are structurally versatile ceramics used in nuclear waste immobilization, ion transport, and energy storage, yet quantitative and physically interpretable composition-to-structure prediction remains limited. In this work, we develop a suite of interpretable multivariate regression models, benchmarked against machine learning approaches, to predict key crystallographic features of hollandite materials, including lattice constants <em>a</em> and <em>c</em>, tunnel bottleneck size (d-d), and crystallographic symmetry. Using physically motivated descriptors such as ionic-radius mismatch, oxidation states, electronegativity, and tunnel occupancy, we reveal a pronounced directional anisotropy in composition-structure relationships. Lattice constant <em>a</em> is accurately captured by a multivariate linear regression (MLR) model dominated by steric mismatch effects, whereas lattice constant <em>c</em> and d-d require compact multivariate polynomial regression (MPR) models to describe essential nonlinear and interaction effects arising from coupled steric and electronic contributions. A regression-based symmetry model further enables reliable classification of tetragonal and monoclinic phases using physically interpretable descriptors alone. Benchmarking against artificial neural networks (ANN) and support vector regression (SVG) confirms that the interpretable regression models achieve comparable accuracy while offering direct mechanistic insight. The validated regression models are subsequently applied to predict structural features of previously uncharacterized hollandite compositions, including actinide-substituted titanates and Mn- and Sn-based analogues. Overall, this work establishes a data-efficient, mechanism-aware framework for composition-driven structure prediction in tunnel-structured oxides, providing a transferable strategy for accelerated materials design.</div></div>","PeriodicalId":10650,"journal":{"name":"Computational Materials Science","volume":"264 ","pages":"Article 114494"},"PeriodicalIF":3.3,"publicationDate":"2026-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145923414","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"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

A new interatomic potential for mixed Mg-Al-Ga-In spinels 混合Mg-Al-Ga-In尖晶石的新原子间势

IF 3.3 3区材料科学 Q2 MATERIALS SCIENCE, MULTIDISCIPLINARY

Computational Materials Science

Pub Date : 2026-10-01 Epub Date: 2026-01-15 DOI: 10.1016/j.commatsci.2026.114492

Michael J.D. Rushton , Michael W.D. Cooper , Ghanshyam Pilania , Blas P. Uberuaga

While density functional theory (DFT) has become the de facto approach for accurate simulation of materials at the atomic scale, there are many aspects of materials that are simply out of reach of DFT methods. In particular, finite temperature properties such as diffusivities, the structure and properties of grain boundaries and interfaces, and the study of defect properties in complex alloys are computationally challenging for DFT methods. Recently, a new class of spinels in which three cations order over two sublattices was discovered. In order to predict the properties of these types of structures, classical potentials are a must. In this work, we derive a new classical potential for Mg-bearing spinels in which the B cations are Al, Ga, and/or In. The potential does well in describing the DFT energetics of various spinel structures as a function of chemistry and inversion. In particular, it reproduces the thermodynamically favorable MgAlGaO₄ structure while correctly predicting that neither MgAlInO₄ nor MgGaInO₄ are stable. Further, it reproduces physical trends in elastic properties as compared against experiment.

虽然密度泛函理论（DFT）已经成为在原子尺度上精确模拟材料的事实上的方法，但材料的许多方面是DFT方法无法达到的。特别是，有限的温度性质，如扩散系数、晶界和界面的组织和性质，以及复杂合金缺陷性质的研究，对DFT方法的计算具有挑战性。最近，发现了一类新的尖晶石，其中三个阳离子在两个亚晶格上有序。为了预测这类结构的性质，经典势是必须的。在这项工作中，我们推导了含镁尖晶石的一个新的经典势，其中B阳离子是Al， Ga和/或In。势很好地描述了各种尖晶石结构的DFT能量学作为化学和反转的函数。特别是，它再现了热力学有利的MgAlInO4结构，同时正确地预测了MgAlInO4和MgGaInO4都不是稳定的。此外，与实验相比，它再现了弹性性能的物理趋势。

{"title":"A new interatomic potential for mixed Mg-Al-Ga-In spinels","authors":"Michael J.D. Rushton , Michael W.D. Cooper , Ghanshyam Pilania , Blas P. Uberuaga","doi":"10.1016/j.commatsci.2026.114492","DOIUrl":"10.1016/j.commatsci.2026.114492","url":null,"abstract":"<div><div>While density functional theory (DFT) has become the de facto approach for accurate simulation of materials at the atomic scale, there are many aspects of materials that are simply out of reach of DFT methods. In particular, finite temperature properties such as diffusivities, the structure and properties of grain boundaries and interfaces, and the study of defect properties in complex alloys are computationally challenging for DFT methods. Recently, a new class of spinels in which three cations order over two sublattices was discovered. In order to predict the properties of these types of structures, classical potentials are a must. In this work, we derive a new classical potential for Mg-bearing spinels in which the B cations are Al, Ga, and/or In. The potential does well in describing the DFT energetics of various spinel structures as a function of chemistry and inversion. In particular, it reproduces the thermodynamically favorable MgAlGaO<sub>4</sub> structure while correctly predicting that neither MgAlInO<sub>4</sub> nor MgGaInO<sub>4</sub> are stable. Further, it reproduces physical trends in elastic properties as compared against experiment.</div></div>","PeriodicalId":10650,"journal":{"name":"Computational Materials Science","volume":"264 ","pages":"Article 114492"},"PeriodicalIF":3.3,"publicationDate":"2026-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145974141","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"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

First-principles study on high-pressure behavior of terahertz spectrum in TATB crystal TATB晶体中太赫兹光谱高压行为的第一性原理研究

IF 3.3 3区材料科学 Q2 MATERIALS SCIENCE, MULTIDISCIPLINARY

Computational Materials Science

Pub Date : 2026-10-01 Epub Date: 2026-01-10 DOI: 10.1016/j.commatsci.2026.114489

Cheng He-ping , Chen Shi , Peng Hui , Liu Qiao , Chen Tu-nan

This study investigates the structural evolution and terahertz (THz) vibrational spectra of 1,3,5-triamino-2,4,6-trinitrobenzene (TATB) crystals under hydrostatic pressures up to 20 GPa using first-principles calculations with Tkatchenko–Scheffler (TS) dispersion correction. The calculations accurately reproduce the key structural parameters, including the lattice parameters, bulk modulus, and its pressure derivative. All terahertz absorption peaks show a systematic nonlinear blueshift with increasing pressure. By fitting the pressure–volume data to the Vinet and third-order Birch–Murnaghan equations of state, we obtain the mode Grüneisen parameters across the terahertz range (2.06–8.88 THz).These parameters display strong frequency dependence: they are markedly larger in the low-frequency region (2.06–4.47 THz, 7.75–13.62) and become smaller and nearly constant at higher frequencies (7.13–8.88 THz, 1.74–1.88). This result highlights a fundamental distinction in how pressure affects vibrations: low-frequency modes, governed by intermolecular forces, respond strongly, whereas high-frequency intramolecular vibrations remain much less sensitive. The work thus provides insights into the high-pressure vibrational behavior of TATB and offers a valuable theoretical reference for future terahertz-based studies of energetic materials under extreme conditions.

本研究利用Tkatchenko-Scheffler （TS）色散校正的第一原理计算，研究了1,3,5-三氨基-2,4,6-三硝基苯（TATB）晶体在高达20gpa静水压力下的结构演变和太赫兹（THz）振动谱。计算准确再现了关键结构参数，包括晶格参数、体积模量及其压力导数。所有太赫兹吸收峰均随压力的增加呈现系统的非线性蓝移。通过将压力-体积数据拟合到Vinet和三阶Birch-Murnaghan状态方程，我们获得了太赫兹范围内（2.06-8.88 THz）的模式grisen参数。这些参数表现出很强的频率依赖性：它们在低频区域（2.06-4.47 THz, 7.75-13.62）明显较大，在高频区域（7.13-8.88 THz, 1.74-1.88）变得较小且几乎恒定。这一结果强调了压力如何影响振动的一个基本区别：由分子间力控制的低频模式反应强烈，而高频分子内振动则不那么敏感。因此，这项工作提供了对TATB高压振动行为的见解，并为未来在极端条件下基于太赫兹的含能材料研究提供了有价值的理论参考。

{"title":"First-principles study on high-pressure behavior of terahertz spectrum in TATB crystal","authors":"Cheng He-ping , Chen Shi , Peng Hui , Liu Qiao , Chen Tu-nan","doi":"10.1016/j.commatsci.2026.114489","DOIUrl":"10.1016/j.commatsci.2026.114489","url":null,"abstract":"<div><div>This study investigates the structural evolution and terahertz (THz) vibrational spectra of 1,3,5-triamino-2,4,6-trinitrobenzene (TATB) crystals under hydrostatic pressures up to 20 GPa using first-principles calculations with Tkatchenko–Scheffler (TS) dispersion correction. The calculations accurately reproduce the key structural parameters, including the lattice parameters, bulk modulus, and its pressure derivative. All terahertz absorption peaks show a systematic nonlinear blueshift with increasing pressure. By fitting the pressure–volume data to the Vinet and third-order Birch–Murnaghan equations of state, we obtain the mode Grüneisen parameters across the terahertz range (2.06–8.88 THz).These parameters display strong frequency dependence: they are markedly larger in the low-frequency region (2.06–4.47 THz, 7.75–13.62) and become smaller and nearly constant at higher frequencies (7.13–8.88 THz, 1.74–1.88). This result highlights a fundamental distinction in how pressure affects vibrations: low-frequency modes, governed by intermolecular forces, respond strongly, whereas high-frequency intramolecular vibrations remain much less sensitive. The work thus provides insights into the high-pressure vibrational behavior of TATB and offers a valuable theoretical reference for future terahertz-based studies of energetic materials under extreme conditions.</div></div>","PeriodicalId":10650,"journal":{"name":"Computational Materials Science","volume":"264 ","pages":"Article 114489"},"PeriodicalIF":3.3,"publicationDate":"2026-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145923034","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"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

A general LLM-powered text mining framework: Applied to extract high entropy alloys 一个通用的基于llm的文本挖掘框架：用于提取高熵合金

IF 3.3 3区材料科学 Q2 MATERIALS SCIENCE, MULTIDISCIPLINARY

Computational Materials Science

Pub Date : 2026-10-01 Epub Date: 2026-01-09 DOI: 10.1016/j.commatsci.2025.114476

Haolun Yuan , Jun Zeng , Jie Zuo , Xin Wang , Dingguo Xu

In this paper, we present a general framework for automating the information extraction process from materials science literature. Our aim is to meet the increasing demand for large-scale databases in both research and engineering. The text mining part consists of three continuous stages: labeling, extraction, and post-processing, which are all powered by large language models (LLMs). Through these successive stages, the framework enables the extraction of material data from both text and tables. It supports the generation of high-quality databases with only a moderate level of prior knowledge about the extraction targets and minimal coding effort, thereby facilitating the rapid development of data-driven models from the ground up. The Framework was applied to high entropy alloys (HEAs) research papers and constructed a comprehensive database of 5393 records encompassing mechanical properties, phase information, and processing histories. Such a database provides a valuable foundation for investigating process–structure–property relationships in alloys, which may support both mechanistic understanding and data-driven design. To assess the quality of the database, we also trained machine learning models to accurately predict phase and yield strength. Our database of HEAs provides a rich resource for future data-driven design of new alloy materials.

在本文中，我们提出了一个自动化材料科学文献信息提取过程的通用框架。我们的目标是满足研究和工程领域对大规模数据库日益增长的需求。文本挖掘部分包括三个连续的阶段：标记、提取和后处理，这些阶段都由大型语言模型（llm）提供支持。通过这些连续的阶段，框架可以从文本和表中提取重要数据。它支持生成高质量的数据库，只需要对提取目标有一定程度的先验知识和最少的编码工作，从而促进从头开始的数据驱动模型的快速开发。该框架应用于高熵合金（HEAs）的研究论文，构建了包含力学性能、相信息和加工历史的5393条记录的综合数据库。这样的数据库为研究合金的工艺-结构-性能关系提供了有价值的基础，可以支持机理理解和数据驱动设计。为了评估数据库的质量，我们还训练了机器学习模型来准确预测相位和屈服强度。我们的HEAs数据库为未来新型合金材料的数据驱动设计提供了丰富的资源。

{"title":"A general LLM-powered text mining framework: Applied to extract high entropy alloys","authors":"Haolun Yuan , Jun Zeng , Jie Zuo , Xin Wang , Dingguo Xu","doi":"10.1016/j.commatsci.2025.114476","DOIUrl":"10.1016/j.commatsci.2025.114476","url":null,"abstract":"<div><div>In this paper, we present a general framework for automating the information extraction process from materials science literature. Our aim is to meet the increasing demand for large-scale databases in both research and engineering. The text mining part consists of three continuous stages: labeling, extraction, and post-processing, which are all powered by large language models (LLMs). Through these successive stages, the framework enables the extraction of material data from both text and tables. It supports the generation of high-quality databases with only a moderate level of prior knowledge about the extraction targets and minimal coding effort, thereby facilitating the rapid development of data-driven models from the ground up. The Framework was applied to high entropy alloys (HEAs) research papers and constructed a comprehensive database of 5393 records encompassing mechanical properties, phase information, and processing histories. Such a database provides a valuable foundation for investigating process–structure–property relationships in alloys, which may support both mechanistic understanding and data-driven design. To assess the quality of the database, we also trained machine learning models to accurately predict phase and yield strength. Our database of HEAs provides a rich resource for future data-driven design of new alloy materials.</div></div>","PeriodicalId":10650,"journal":{"name":"Computational Materials Science","volume":"264 ","pages":"Article 114476"},"PeriodicalIF":3.3,"publicationDate":"2026-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145923415","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"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