Computational methods to systematically identify intercalation sites

Computational investigations of novel or existing cathode materials have the ability to speed-up the design of novel and improved cathode materials, either to replace the existing Li-ion cathodes, or to be used in non-Li-ion batteries.

Density functional theory is well suited for such explorations. For each material one needs to identify the lowest energy intercalation sites. Trial-and-error, ideally based on physical intuition, is often used to identify these intercalation sites. This approach is, however, not systematic and might miss the lowest energy intercalation sites. We therefore explore and compare three systematic and easily automatable approaches: one based on voronoi decompositions, one based on a bond-valence model, and one based on the electronic ground-state density, using vanadium pentoxide (V₂O₅) as test system.