About this Abstract |
| Meeting |
2027 TMS Annual Meeting & Exhibition
|
| Symposium
|
Four-Dimensional Scanning Transmission Electron Microscopy (4D STEM) for Quantitative Materials Characterization
|
| Presentation Title |
Dynamic Scan Patterns and Automated Data Processing for In Situ 4D-STEM |
| Author(s) |
Carter Shaw Francis, Scott Dahl, Nicholas Hagopian, Benjamin Bammes, Paul Voyles |
| On-Site Speaker (Planned) |
Carter Shaw Francis |
| Abstract Scope |
In situ 4D STEM provides unrivaled characterization of structural evolution at high spatial resolution, offering new insights into short-range structural changes induced by external stimuli such as strain, heating, electrical biasing, or irradiation. These experiments face several interrelated challenges, including limited temporal resolution, dose accumulation, limited signal, and large, complex datasets. Optimizing scan patterns and automating data collection and analysis can substantially reduce this data burden without sacrificing the information of interest. In this work, we present dynamic, custom scan patterns that correct for drift and restrict acquisition to an automatically defined area of interest, paired with automated data processing combining general image-processing techniques with neural networks. Each dataset is reduced to a representative set of diffraction vectors, used to measure structural evolution across spatial and temporal dimensions. |
| Proceedings Inclusion? |
Planned: |
| Keywords |
Characterization, Machine Learning, Solidification |