Is CrystalNN suitable for amorphous systems？

[sek1ro-yuzzz]

I am using pymatgen to analyze amorphous solid-state electrolytes NaTaCl6. I tried using the RDF method to determine coordination numbers, but the cutoff distance (first minimum) is quite large (e.g., ~4.5 A for Na-Cl), which leads to an overestimated coordination number.

I have read the paper "Benchmarking Coordination Number Prediction Algorithms on Inorganic Crystal Structures" (https://pubs.acs.org/doi/10.1021/acs.inorgchem.0c02996), but it is still unclear to me if CrystalNN is applicable to amorphous systems.

Does CrystalNN work reliably for amorphous structures?

Thanks!

[JaGeo]

Hey @sek1ro-yuzzz ,
CrystalNN has been benchmarked on typical inorganic crystal structures. I am not aware of an application for an amorphous system. In any case, I would recommend to check and benchmark cutoffs and weights carefully. I am also not sure how fast it will be for an amorphous system.