Revealing in-plane grain boundary composition features through machine learning from atom probe tomography data

Gibbs Adsorption and Zener Pinning Enable Mechanically Robust High‐Performance Bi2Te3‐Based Thermoelectric Devices

“…To better resolve the spatial distribution of Mg and possibly B, atom probe tomography (APT) characterizations were carried out. [ 56 ] We analyzed the APT data by the method developed by Zhou et al., [ 57 ] which can reveal the in‐plane chemical features and the Gibbsian interfacial excess that could not be identified by standard compositional analyses. Figure a shows the GB mesh created by recognizing GB traces using a convolutional neural network.…”

Analytical Methods for Superresolution Dislocation Identification in Dark-Field X-ray Microscopy

“…Over the past decade, materials science experiments have built momentum in using data-science methods to analyze microscopy results [25]. This has enabled key opportunities for experiments to quantify information about grain-boundaries [26], dislocations [27], and other defects in statistically significant populations. Similar approaches in subsurface X-ray microscopy have struggled due to limitations in the optical resolution, and much more difficult interpretation due to the very high strain-resolution, which can make it difficult to interpret overlapping features.…”

On Strong-Scaling and Open-Source Tools for High-Throughput Quantification of Material Point Cloud Data: Composition Gradients, Microstructural Object Reconstruction, and Spatial Correlations

“…triangles without connectivity information, representing a complex (set) of isosurfaces. Although MC has frequently been applied in the atom probe literature, mostly via its implementation in commercial software, few atom probers have discussed that the implementation of the topological rule set can differ between MC implementations [64]. These differences can result in eventually significant effects on the local topology and closure of the iso-surface, in particular when there is a strong sensitivity on the threshold value ϕ.…”

“…Felfer et al implemented computational geometry methods for this task [42,43,46]. Their so-called DCOM algorithm has influenced several authors [47,64,124]. Implemented in practice, these tools are semi-automated and may or not need manual mesh processing which is most conveniently performed with GUI-based tools like Blender.…”

“…Suggestions were made by individual researchers how data from the GUI of commercial software can be extracted into spreadsheets [59,60]. As far as tasks like composition, interface-based analyses, and microstructural-object-centric analyses are considered, we found only few cases where tools other than the commercial ones were used [61][62][63] (apart from the case of interfacial excess quantification [46,64]. )…”