Accelerated Percolation Path Identification in Twisted Bilayer Graphene

In this work, we present a novel computational workflow for accelerating the identification of percolation pathways in twisted bilayer graphene of nearly point-charge ions. The method uses the charge density from a single ab initio calculation using Density Functional Theory and requires that the percolating ion only weakly influences the charge density of the host material. The method is composed of three steps. First the intercalation sites in the bilayer are identified, then a graph describing the possible migrations between those is generated, and last a path-finding algorithm is used to discover the lowest-cost percolation paths. We have applied this workflow to Li-diffusion in 21 different twist-angle structures of twisted bilayer graphene, which could be imagined as a potential anode material in Li-ion batteries. We found that it yields physically plausible pathways in all examined cases and observed a significant relationship between the twist angle and the ease-of-percolation, highlighting the value of computational studies in mapping percolation paths. Our method is general and much faster than that conventionally used to determine percolation paths. Therefore, the method enables the efficient investigation of percolation pathways in diverse materials, including other 2D heterostructures and even 3D crystalline materials with trivial alterations.