Identification of crystal structures in atomistic simulation by predominant common neighborhood analysis

Structural transformation is prone to alteration under crystalline deformation for multiple structural symmetry groups. Such transformation usually includes a large-scale number representing atomic bonding and material deformation mechanisms in the vicinity of voids, dislocation cores and other defects for diverse crystal systems. Selected analytical methods should be able to distinguish prefect crystal structures of multiple crystalline systems while extracting distinct crystalline symmetry groups from nearby stacking faults, crack surfaces or other features. Some work has been done to optimize such identification in the form of Common Neighborhood Parameter CNP formulation. In this paper, an improved approach towards structural identification revolving around predominance of common nearest neighbors of atoms is introduced with two separate parameterizations for arbitrary crystal structures. The approach is called Predominant Common Neighborhood Parameter PCNP. The method is proposed to characterize cross-species interactions with more complex centrosymmetric space groups or crystals with no common first nearest neighbors for certain atomic species. We validate the method against conventional Centrosymmetry Parameter CSP and Common Neighbor Analysis CNA for multiple simulation conditions, where the introduced method illustrates higher sensitivity than CNP analysis for multiple geometries and defect characterization regardless of the atom's species. Further enhancement of the method has been introduced to include a complex, non-monoatomic centrosymmetric crystal with a P6/mmm space group, typically found in certain ceramic materials such as Zirconium Diboride.