Evaluation of the reproducibility and crystal tracking precision of TEM goniometers in tomography experiments.

Nanocrystalline materials are the basis of many novel engineered systems, including batteries, nanocomposites, and glass ceramics. Three-dimensional electron diffraction (3D ED) has become a key technique for structural analysis of such materials, offering clear advantages over conventional X-ray diffraction. Commercial routine 3D ED acquisition allowing for measurements of crystals down to ∼750 nm is now standard, but pushing the measurable size towards a few tens of nanometers introduces new challenges, requiring robust crystal-tracking methods. At this scale, TEM automation, reliable object detection, and high mechanical precision of the goniometer are essential. PyFast-ADT is introduced as a modular automation framework for 3D ED data collection, extending the measurable size range through improved crystal tracking routines. Its Python architecture enhances shareability and promotes facility automation within the 3D ED and Cryo-EM communities. The PatchworkCC algorithm combines Cross-Correlation with Kalman Filtering to achieve fully automatic crystal tracking with improved accuracy and minimal user supervision. Characterization of goniometer reproducibility revealed a rapid decrease behaviour degrading precision, addressed by the HiPerGonio procedure, which stabilizes performance and supports optimal TEM/sample holder choices. Together, these developments enable fully automated 3D ED data collection on 25 nm nanocrystals embedded in a glass-ceramic matrix, increasing throughput up to sixfold and advancing reproducible, high-throughput structure determination at the nanometer scale.