Simulated and Experiment-Mimicking Electron Diffraction Image Dataset for Single-Walled Carbon Nanotube Chirality Classification

This dataset contains simulated electron diffraction images of single-walled carbon nanotubes (SWCNTs). The dataset covers 266 distinct SWCNT chiralities with diameters ranging from 0.48 to 2.3 nm and chiral angles from 0° to 30°.Two types of diffraction images are provided. (1) Base simulated diffraction patterns represent the calculated diffraction-intensity distributions and include the source patterns used to generate both the training and validation datasets. For each chirality, these patterns cover CNT tilt angles from 0° to 20.5° in 0.5° increments, corresponding to 42 tilt angles and 11,172 chirality–tilt combinations. (2) Experiment-mimicking diffraction images were generated by applying image augmentation designed to reproduce characteristics encountered in experimental TEM diffraction measurements. These images cover tilt angles from 0.5° to 20.5° in 1° increments, corresponding to 21 tilt angles and 5,586 chirality–tilt combinations.The experiment-mimicking image generation incorporates experimentally relevant features and perturbations, including direct-beam and beam-stop features, blooming artifacts, random noise, blurring, contrast variation, cropping, and random erasing. Each image is associated with a known SWCNT chirality (n,m) and CNT tilt angle, providing structural labels for the development, validation, and benchmarking of chirality-classification and diffraction-analysis methods.The dataset is associated with the study “AI-Based Chirality Classification for Carbon Nanotubes from Electron Diffraction Patterns Using Vision Transformer,” Carbon 257 (2026) 121672.