Abstract Scope |
This talk first reviews a series of our studies to compute grain boundary (GB) counterparts to bulk phase diagrams. Here, a grand challenge is to construct complexion diagrams for general GBs as function of five macroscopic (crystallographic) degrees of freedom (DOFs). A most recent collaborative study uses genetic algorithm-guided deep learning to predict GB properties in a 7-D space (5 macroscopic DOFs plus temperature and composition) [Materials Today 2020]. Furthermore, several examples of complexions formed at general GBs are discussed. Complex interfacial superstructures with gradients in the chemical, structural, and bonding characters were characterized and modeled at both general GBs [Materials Horizons 2020] and phase boundaries [Science Advances 2021]. An electrochemically induced GB transition was discovered and modeled, which can induce abnormal grain growth [arXiv: 2012.15862]. Here, thermodynamic models, DFT calculations, hybrid Monte Carlo and molecular dynamics simulations, machine learning, and advanced microscopy are combined to understand these interfaces. |