| About this Abstract |
| Meeting |
2025 TMS Annual Meeting & Exhibition
|
| Symposium
|
Characterization of Materials through High Resolution Coherent Imaging
|
| Presentation Title |
Physics-Informed Self-Supervised Learning of Structural Morphology Imaged by Scanning X-Ray Diffraction Microscopy |
| Author(s) |
Aileen Luo, Tao Zhou, Ming Du, Martin V. Holt, Andrej Singer, Mathew J. Cherukara |
| On-Site Speaker (Planned) |
Aileen Luo |
| Abstract Scope |
Scanning X-ray nanodiffraction microscopy is a powerful technique for spatially resolving nanoscale structural morphologies by diffraction contrast. One of the critical challenges in experimental nanodiffraction data analysis is in disentangling the crystalline lattice information from the effects of the zone plate optics. The convergence angle of nanoscale focusing optics creates simultaneous dependency of the far-field scattering data on three independent components of the local strain tensor, which are resolvable through a spatially mapped sample tilt series. Yet, traditional data analysis is computationally expensive and prone to artifacts. Here, we present NanobeamNN2.0, a convolutional neural network that learns lattice strain and rotation angles from simulated diffraction of a focused X-ray nanobeam by an epitaxial thin film. NanobeamNN2.0 has a built-in physics model, eliminating the need for labeled data during training. We demonstrate that a case study of this approach on experimental data and discuss the potential advantages in enabling real-time analysis. |
| Proceedings Inclusion? |
Planned: |
| Keywords |
Machine Learning, Thin Films and Interfaces, Modeling and Simulation |