Material Dynamics Analysis with Deep Generative Model

This dataset contains time-resolved 3D XAFS-CT (X-ray Absorption Fine Structure Computed Tomography) imaging data of sulfidation dynamics in brass particles embedded in a rubber matrix. Ten representative particle clumps were selected from a larger set of approximately 4,956 particles and tracked over four aging intervals: 0, 3, 14, and 28 days. Each clump is stored in a separate folder indexed by an identifier (e.g., 119, 164, etc.), and contains four .npy files corresponding to the voxelized 3D structures at each time point (0d.npy, 3d.npy, 14d.npy, and 28d.npy). The voxel size is 0.65 µm, and each volume is formatted into a 32×32×32 grid. This dataset supports the study of microstructural and chemical evolution during sulfidation and is suitable for machine learning applications, including generative modeling and temporal analysis.See more details in: Matsui, H., Muramoto, Y., Niwa, R. et al. Machine learning-derived reaction statistics for 3D spectroimaging of copper sulfidation in heterogeneous rubber/brass composites. Commun Mater 4, 88 (2023). https://doi.org/10.1038/s43246-023-00413-z