Neural Network Kinetics Method Predicts Atomic Diffusion in Complex Materials
IRG-1 researchers, Assistant Professor Penghui Cao of Mechanical and Aerospace Engineering and his student Bin Xing, have developed a groundbreaking Neural Network Kinetics (NNK) method to model and predict atomic diffusion in compositionally complex materials. This innovative framework leverages artificial neural networks to simulate the chemical and structural evolution of materials, addressing the challenges posed by chemical complexity. By studying NbMoTa alloys, the team discovered a critical temperature at which chemical ordering reaches a peak, and diffusion heterogeneity is highest.
The NNK approach provides a scalable solution for exploring the vast potential of diffusion-related properties in advanced materials, potentially unlocking extraordinary new material properties.
This work has been published in Nature Communications. The authors also include IRG-1 faculty Timothy J. Rupert and Xiaoqing Pan.