Materials property predictions have improved from advances in machine learning algorithms, delivering materials discoveries and novel insights through data-driven models of structure-property relationships. Nearly all available models rely on featurization of materials composition, however, whether the exclusive use of structural knowledge in such models has the capacity to make comparable predictions remains unknown. Here we employ a deep neural network model to decode structure-property relationships in crystalline materials without explicitly considering chemical compositions. The focus is on classification of crystal systems, mechanical elasticity, electronic band gap, and phase stability. Our model utilizes a three-dimensional (3D) momentum space representation of structure from elastic x-ray scattering theory that exhibits rotation and permutation invariance. We perform novel ablation studies to help interpret the model performance by perturbing the physically meaningful input features (i.e., the diffraction patterns) instead of tuning the architecture of the learning model as in conventional ablation methods. We find that the spatial symmetry of the 3D diffraction patterns, which reflects crystalline symmetry operations, is more important than the diffraction intensities contained within for the model to make a successful classification. Our work showcases the potential of using statistical learning models to help understand materials physics, rather than performing predictive and generative tasks as in most materials informatics research. We also argue that learning the crystal structure genome in a chemistry-agnostic manner demonstrates that some crystal structures inherently host high propensities for optimal materials properties, which enables the decoupling of structure and composition for future codesign of multifunctionality.
Property-Guided Diffusion for Inverse Design of Crystalline Materials
Harish-Chandra Research Institute(哈里什·钱德拉研究所);National Institute of Science Education and Research (NISER)(国家科学教育研究中心);Homi Bhabha National Institute(霍米·巴巴国立研究所)
Anisotropic magnetic property of single crystals $R$V$_6$Sn$_6$ ($R$ = Y, Gd - Tm, Lu)
采用Sn助熔剂法合成了$R$ V ${_6}$ Sn ${_6}$($R$ = Y,Gd-Tm,Lu)单晶,并通过磁化强度、电阻率和比热测量表征了它们的物理性能。所有样品的粉末X射线衍射图都可以很好地用六方晶系HfFe $_6$ Ge $_6$型结构进行索引,其中稀土原子形成六方晶系层,钒原子形成可果美层。在高温下,磁矩轴承稀土($R$ = Gd-Tm)的磁化率测量遵循居里-外斯行为。由磁化率曲线的多晶平均值估算的有效磁矩与游离$R^{3+}$离子的值一致。除了GdV $_6$ Sn $_6$外,其它稀土磁矩均表现出强的磁各向异性。当$R$ = Tb - Ho时,易磁化方向为$c$轴;当$R$ = Er,Tm时,易磁化方向为$ab$面. $R$ V ${_6}$ Sn ${_6}$中的钒离子不具有磁矩。$R$ = Y和Lu的化合物表现出顺磁性金属的典型特征。在低温下,磁有序性由磁化强度、比热和电阻率确定:GdV $_6$ Sn $_6$的$T_{N} = 4.9$ ~K最高,HoV $_6$ Sn $_6$的$T_{N} = 2.3$ ~K最低。当$R$ = Er和Tm时,在1.8~K范围内没有观察到磁有序.磁有序温度与de Gennes标度的轻微偏差表明金属$R$ V ${_6}$ Sn ${_6}$化合物中稀土磁矩之间的主要Ruderman-Kittel-Kasuya-Yosi…
Molecular Dynamics Studies of Dislocations in CdTe Crystals from a New Bond Order Potential
Cd 1-xZnxTe(CZT)晶体是辐射探测的主要半导体材料,但探测器级材料的高成本限制了其应用。高晶体成本主要是由导致低制造产率的性质不均匀性引起的。虽然过去已经做出了巨大的努力来减少CZT中的Te夹杂物/沉淀物,但这并没有导致材料性能均匀性的预期改善。除Te粒子外,位错胞也会引起电场扰动和相关的性质不均匀性。因此,材料的进一步改进需要理解和控制CZT晶体中的位错。在这里,我们使用最近开发的CZT键级势进行代表性的分子动力学模拟,以研究CdTe中29种不同类型的可能位错的构型、能量和迁移率(即,x = 1)晶体。本文将探索一种有效的方法来推导热激活位错运动的激活自由能和激活体积。我们的重点是深入了解材料中的重要位错,并为改善CZT晶体中位错网络结构的实验工作提供指导。
Discovery and recovery of crystalline materials with property-conditioned transformers
University College London(伦敦大学学院);Earth Rover Program(地球漫游者计划);Queen Mary University of London(伦敦大学玛丽女王学院);NOMATEN Centre of Excellence, National Centre for Nuclear Research(国家核研究中心NOMATEN卓越中心)
Beyond Structure: Invariant Crystal Property Prediction with Pseudo-Particle Ray Diffraction
Guangzhou Municipal Key Laboratory of Materials Informatics, Advanced Materials Thrust, The Hong Kong University of Science and Technology (Guangzhou)(广州材料信息重点实验室,先进材料方向,香港科技大学(广州));The Chinese University of Hong Kong(香港中文大学);Data Science and Analytics Thrust, The Hong Kong University of Science and Technology (Guangzhou)(数据科学与分析方向,香港科技大学(广州));Department of physics, The City University of Hong Kong(香港城市大学物理系);Materials Genome Institute, Shanghai University(材料基因组研究所,上海大学)
Adaptable Method for Crystal Design across Diverse Constraints and Objectives with Pretrained Property Predictors
The University of Tokyo(东京大学);OMRON SINIC X Corp.(OMRON SINIC X公司);National Institute for Materials Science(国家材料科学研究所);National Institute of Advanced Industrial Science and Technology(国家先进工业科学和技术研究院)
PDDFormer: Pairwise Distance Distribution Graph Transformer for Crystal Material Property Prediction
Software Engineering Institute, East China Normal University(东华大学软件工程院);School of Chemistry and Molecular Engineering, East China Normal University(东华大学化学与分子工程学院);School of Geospatial Information, Information Engineering University(信息工程大学测绘信息学院);School of Communication and Electronic Engineering, East China Normal University(东华大学通信与电子工程学院);School of Artificial Intelligence, University of Chinese Academy of Sciences(中国科学院大学人工智能学院)
Material Property Prediction with Element Attribute Knowledge Graphs and Multimodal Representation Learning
Institute of Computing Technology, Chinese Academy of Science(中国科学院计算技术研究所);Ningbo Institute of Information Technology Application, Chinese Academy of Sciences (CAS)(中国科学院宁波信息技术应用研究院)
DenseGNN: universal and scalable deeper graph neural networks for high-performance property prediction in crystals and molecules
School of Materials Science and Engineering, Shanghai Jiao Tong University(上海交通大学材料科学与工程学院);Zhangjiang Institute for Advanced Study, Shanghai Jiao Tong University(上海交通大学张江高等研究院);Materials Genome Initiative Center, Shanghai Jiao Tong University(上海交通大学材料基因组计划中心)
LLM4Mat-Bench: Benchmarking Large Language Models for Materials Property Prediction
Princeton University(普林斯顿大学);Vertaix;Acceleration Consortium, University of Toronto(多伦多大学加速联盟);Vector Institute for Artificial Intelligence(向量人工智能研究所);Schwartz Reisman Institute for Technology and Society(施瓦茨·赖斯曼技术与社会研究所);University of Toronto(多伦多大学)