Artificial-Intelligence-Driven Characterization of Crystallographic Interfaces from Electron Microscopy

Characterizing crystallographic interfaces in synthetic nanomaterials is an important step for the design of novel materials. Trained materials scientists can assign interface structures of materials by looking at high-resolution imaging and diffraction data obtained by aberration-corrected scanning transmission electron microscopy (STEM). However, the high-acquisition rates in STEM pose a challenge to a purely human-based identification of interfaces or defects. As of today, STEM datasets are being massively accumulated, but they cannot be fully exploited due to the lack of automatic analysis tools. Here, we present AI-STEM, a newly developed AI tool, based on a a Bayesian neural network, for accurately extracting the key features of (poly)crystalline materials, i.e., crystal-structure prototype, lattice constant, and (relative) orientation from atomic-resolution STEM images. AI-STEM operates on both high-angle annular dark-field (HAADF) and convergent beam electron diffraction (CBED) images. It is trained on 25,080 simulated STEM images, and achieves excellent predictive performance for identifying crystal structure and lattice misorientations on experimental images.