Recent Advances in Automating TEM-based Characterization of Nuclear Materials
Abstract: Recently, an interest has grown in the development of high-performance nuclear materials for advanced reactor concepts. To meet the deployment schedule for these advanced reactor deployments, high-throughput characterization tasks are needed to shorten the materials qualification timeline. This talk will highlight research directions focused on automating the quantitative workflows using transmission electron microscopy (TEM) for radiation resistance evaluations. Specifically, general concepts around machine learning (ML) tasks including object detection and quantification will be presented. The presentation will progress towards a discussion regarding challenges associated with deploying common ML architectures to TEM-based quantification and recent advances in solutions to resolve these issues. These solutions include our efforts in developing techniques for large-scale training database formation using synthetic data generation. The power of the automated approaches will be demonstrated on in-situ TEM ion irradiation experiments where direct links between defect nucleation, mobility, and growth based on alloy chemistry and microstructure can be extracted in mere seconds to minutes of data post processing. The presentation will conclude with a discussion on the status of developing augmented reality like TEM-based experiments using the presented ML approaches through innovative deployments of edge-computing devices in microscopy-based workflows.
Biography: Dr. Kevin Field is an Associate Professor in the Department of Nuclear Engineering and Radiological Sciences at the University of Michigan where his research specializes in alloy development and radiation effects. His active research interests include advanced electron microscopy and scattering-based characterization techniques, additive/advanced manufacturing for nuclear materials, and the application of machine/deep learning techniques for advanced innovation in characterization and development of material systems. Prof. Field moved to the University of Michigan in the Fall of 2019 after six years at Oak Ridge National Laboratory. Prof. Field is also Vice President and co-founder of Theia Scientific, LLC, a company focused on bringing machine learning approaches to microscopy through innovative edge-computing frameworks. Prof. Field has presented and published numerous manuscripts on radiation effects in various material systems relevant for nuclear power generation including irradiated concrete performance, deformation mechanisms in irradiated steels, and radiation tolerance of enhanced accident tolerant fuel forms. Dr. Field received his B.S. (2007) from Michigan Technological University in Materials Science & Engineering and his M.S. (2009) and Ph.D. (2012) from the University of Wisconsin – Madison in Materials Science with a focus on segregation phenomena in ion and neutron irradiated ferrous-based alloys. Dr. Field’s work has been recognized through several avenues including receiving the prestigious Alvin M. Weinberg Fellowship from ORNL in 2013 and being awarded the UT-Battelle Award for Early Career Researcher in Science and Technology in 2018 and Department of Energy Early Career Award in 2020.
