A new angle on stacking faults: Breaking the edge-on limit in high-resolution defect analysis

This record contains datasets and code related to the publication: N. Karpstein et al., A new angle on stacking faults: Breaking the edge-on limit in high-resolution defect analysis (submitted to journal). A preprint has been uploaded to arXiv: doi:10.48550/arXiv.2506.15510 A readme file containing descriptions of datatypes can be found in the main folder. This record has been updated to provide a separate Readme file with instructions and further details for the code, additional comments within the code, as well as a demo dataset on which the code can be tested. Abstract: The nature of stacking faults—whether intrinsic or extrinsic—plays a pivotal role in defect-mediated processes in crystalline materials. Yet, current electron microscopy techniques for their reliable analysis remain limited to either conventional fringe-contrast imaging of inclined faults or atomic-resolution imaging of edge-on configurations. Here, we overcome this dichotomy by introducing a high-resolution scanning transmission electron microscopy (HRSTEM) method that enables full structural discrimination of inclined stacking faults, as demonstrated for various faults in fcc and L1₂ crystals. This approach eliminates a long-standing geometric constraint on high-resolution analysis, providing comprehensive access to stacking faults on all glide planes along the widely used [001] and [110] zone axes. We demonstrate the robustness of the method in a CoNi-based superalloy, achieving clear discrimination of fault types even in overlapping configurations and foil thicknesses exceeding 100 nm. Simulations reveal that fault-induced de-channeling is key to contrast formation and is strongly governed by the fault’s depth within the sample. Leveraging this effect, we further establish a route to artificially generate ultrathin TEM lamellae—bounded by the stacking fault itself—thereby enhancing contrast for atomic-scale studies of long-range ordering, compositional fluctuations, and nanoclustering. Code is released under the MIT license, everything else is under the CC BY 4.0 license.