| About this Abstract |
| Meeting |
2024 TMS Annual Meeting & Exhibition
|
| Symposium
|
Irradiation Testing: Facilities, Capabilities, and Experimental Designs
|
| Presentation Title |
Accelerating the Pace of Radiation Damage Experiments through Novel Sample Geometries, Beam Line Architecture, and Machine Learning Analysis |
| Author(s) |
Kevin Field, Charles Hirst, Aäron Penders, Hangyu Li, Robert Renfrow, Alexander Flick, Kai Sun, Zhijie Jiao, Gary Was |
| On-Site Speaker (Planned) |
Kevin Field |
| Abstract Scope |
Radiation damage investigations have historically been limited due to the serial design of experiments and the manual quantification of defects. In this talk, the Michigan Ion Beam Laboratory’s recent advances in multi-dimensional experiments and real-time, automated, feature analysis will be discussed. Specifically, the implementation of novel sample geometries and beam gradients that enable a single irradiation experiment to explore multiple stress, dose, dose rate, or helium-to-damage ratios will be presented. In addition, the implementation of machine learning (ML) for automated feature detection during in-situ ion irradiation experiments in the transmission electron microscope will be highlighted. Our recent results show an ~80× acceleration in experimental throughput using ML-enabled quantification and 2,500+ discrete irradiation conditions can be generated on a single sample using a novel beam line shutter architecture coupled with microbeam analysis. The generation of multi-dimensional datasets through these advances will accelerate the exploration of irradiation parameter-space. |
| Proceedings Inclusion? |
Planned: |
| Keywords |
Characterization, Nuclear Materials, Machine Learning |