Unraveling the size fluctuation and shrinkage of nanovoids during in situ radiation of Cu by automatic pattern recognition and phase field simulation

Void formation is an important aspect of irradiation response of metals. In situ transmission electron microscopy observation for void evolution during irradiation is an effective technique for studying void evolution. However, the amount of data collected during in situ studies drastically overwhelm the current capability for manual data analyses. Here, we used a data-driven approach where a convolutional neural network combined with greedy matching to detect and track nanovoid evolutions and migrations. This approach was able to discover the surprising phenomena of void size fluctuation and shrinkage during irradiation of Cu with pre-existing nanovoids. Phase-field simulations revealed the fundamental mecha-nism behind this in situ observed phenomenon of void size fluctuation.(c) 2022 Elsevier B.V. All rights reserved.

Automated summary

Void formation is an important aspect in the irradiation response of metals. In situ transmission electron microscopy was used to observe the evolution of voids during irradiation. By combining a convolutional neural network with greedy matching, a data-driven approach was utilized to detect and track nanovoid evolutions and migrations, revealing surprising phenomena of void size fluctuation and shrinkage in irradiated Cu with pre-existing nanovoids. Phase-field simulations provided insights into the fundamental mechanism behind the observed void size fluctuation phenomenon.