Neural Network of Defect Kinetics for Multi-Principal Element Alloys

The emergent multi-principal element alloys (MPEAs), commonly known as high entropy alloys, provide a vast compositional space to search for radiation-resistant materials for advanced nuclear reactor applications. However, the vast composition space of MPEAs makes identifying the desired composition and defect mechanisms a complex task. This presentation will introduce a machine learning strategy, specifically neural networks, to overcome this challenge. The machine learning-driven models have proven effective and efficient in predicting defect migration energy barriers across the complete compositional spectrum of MPEAs. The successful implementation of neural network kinetics model holds significant promise for harnessing defect kinetics in the huge compositional space. It could significantly accelerate the alloy selection process, paving the way for engineering new alloy compositions with enhanced radiation performance.