About this Abstract |
Meeting |
2025 TMS Annual Meeting & Exhibition
|
Symposium
|
Characterization of Materials through High Resolution Coherent Imaging
|
Presentation Title |
ML-Guided Non-Destructive 3D Metrology of Functioning Devices With an X-Ray Laser |
Author(s) |
Oliver Hoidn, Aashwin Mishra, Matthew Seaberg, Apurva Mehta |
On-Site Speaker (Planned) |
Apurva Mehta |
Abstract Scope |
Watching the smallest cogs move in a complex machine shows how it functions and why it fails. The cogs driving many critical technologies, from microelectronics to batteries, are shrinking to a few nanometers. Nanoscale structures, e.g., crack tips, also control the functioning of larger components. We must see these cogs in operating devices with nanometer resolution but a field of view (FOV) of microns to see how they mesh with other cogs. Optical probes don’t have the resolution. Electron probes don’t have the needed FOV and are often destructive. X-rays have resolution and sufficient penetration to view large 3D volumes noninvasively. However, poor quality of X-ray optics limits resolutions. Coherent lensless imaging gains resolution but at the expense of reconstruction speed. I will discuss Physics-Informed Machine Learning approaches that speed up reconstruction by 1000 times and, by extracting information buried in noise, can accelerate data collection by another ten times. |
Proceedings Inclusion? |
Planned: |
Keywords |
Machine Learning, Mechanical Properties, Energy Conversion and Storage |