Transformative Techniques in Ion Irradiation and Characterization for Breakthroughs in Materials for Extremes

In this work, we show our recent strides in accelerating the pace and enhancing the fidelity of radiation damage experiments through advancements in irradiation testing and characterization methods. By developing innovative sample geometries, refining experiment architectures, and integrating machine learning-assisted analysis and data visualization, we aim to address the grand challenge of understanding materials behavior under extreme conditions ¹, particularly where radiation at elevated temperatures and stress is pivotal to material performance. Ion irradiation experiments are indispensable for studying radiation effects, compensating for the scarcity of certain neutron sources (fast fission and fusion neutrons) and the current limitations of computational simulation tools in fully capturing the spatial and temporal phase space. The accelerated dose rates achievable with ion irradiation significantly speed up radiation effects testing, although traditional methods suffer from low fidelity and throughput due to static sample geometries, ion beam control, and human-based analysis.

Our approach seeks to improve ion beam irradiation testing by employing advanced sample geometries and dynamic beam applications, allowing for the exploration of multiple stress, dose, dose rate, temperature, and helium-to-damage ratios within a single irradiation experiment. This results in over three to ten times or more, the number of experimental conditions per irradiation, compared to conventional methods. In this work, we will highlight the use of tapered tensile creep geometries paired with non-contact extensometry to extract creep strain rates in static stress ranges of 80 to 230 MPa within a single simultaneous irradiation-creep sample, revealing the corresponding unique dislocation loop microstructures across the stress gradient. As part of this talk, we will also unveil a newly developed test station for performing stress- and temperature-gradient studies using microtensile testing machines. Additionally, our dynamic beam shuttering technique permits the generation of distinct single, dual, and triple beam irradiation conditions on a single sample or multiple samples. We will highlight recent results for helium and hydrogen gas implantation gradient studies to facilitate the rapid exploration of swelling responses and cavity stabilization behaviors in fusion-relevant steels. We also accelerate post-irradiation examination by deploying machine learning (ML) algorithms^2–6 for characterization. We will highlight recent advances in using ML to perform image stabilization during in-situ irradiations, which are devoid of key point references, providing new means to better understand defect mobility and growth in complex and dynamical responses of materials under irradiations. These ML-enhanced techniques, when coupled with our novel irradiation configurations, set a new standard for the examination and discovery of material responses under extreme conditions.

1. Wadsworth, J. et al. Basic Research Needs for Materials Under Extreme Environments. Report of the Basic Energy Sciences Workshop on Materials Under Extreme Environments, June 11-13, 2007. (2008) doi:10.2172/935440.
2. Li, W., Field, K. G. & Morgan, D. Automated defect analysis in electron microscopic images. NPJ Comput Mater 4, 36 (2018).
3. Jacobs, R. et al. Materials swelling revealed through automated semantic segmentation of cavities in electron microscopy images. Scientific Reports 2023 13:1 13, 1–13 (2023).
4. Shen, M. et al. A deep learning based automatic defect analysis framework for In-situ TEM ion irradiations. Comput Mater Sci 197, 110560 (2021).
5. Jacobs, R. et al. Performance, Successes and Limitations of Deep Learning Semantic Segmentation of Multiple Defects in Transmission Electron Micrographs. Cell Rep Phys Sci (2021).
6. Shen, M. et al. Multi defect detection and analysis of electron microscopy images with deep learning. Comput Mater Sci 199, 110576 (2021).