Towards Quantum-Accurate Modeling of Polycrystalline Grain Boundary Environments

The properties of grain boundaries (e.g., diffusion, mobility, and solute segregation) are strongly dependent on the array of the local atomic environments present at the boundary. Therefore, to fully understand the structure-property relationships for grain boundaries, it is essential to study grain boundaries with atomic resolution. However, the most important general low-symmetry grain boundaries are inaccessible to quantum mechanical simulation methods. As a result, researchers tend to focus on high-symmetry coincident site lattice (CSL) grain boundaries in bicrystals, which are accessible to quantum methods, as opposed to the more complex polycrystalline grain boundaries. In this talk, using solute segregation for illustration, we will discuss the shortcomings of using CSL grain boundaries as a general model for grain boundary environments. In addition, we will discuss our recently developed algorithmic framework that can be used to study grain boundary local atomic environments in polycrystals with quantum accuracy. Using this framework, we are able to directly compute from quantum methods the full spectrum of solute segregation energies in polycrystals, and to build a comprehensive segregation database across the alloy space.