Efficient and universal characterization of atomic structures through a topological graph order parameter

A new graph-based order parameter is introduced for the characterization of atomistic structures. The universal to any material/chemical system, and transferable all structural geometries. Three sets data are used validate both generalizability accuracy algorithm: (1) liquid lithium configurations spanning up 300 GPa, (2) condensed phases carbon along with nanotubes buckyballs at ambient high temperature, (3) a diverse set aluminum including surfaces, compressed expanded lattices, point defects, grain boundaries, liquids, nanoparticles, non-zero temperatures. also compared existing methods speed accuracy. Our uniquely classifies every configuration outperforms crystalline parameters studied here, opening door its use in multitude complex application spaces that can require fine configurational materials.