Re: Using different model chemistry for opt and NBO calculations

dummy@example.com (sobereva) — Sun, 04 Jul 2021 01:05:20 +0000

The calculation level of geometry optimization and NBO analysis can be different.

6-31G is too poor for geometry optimization purpose (especially for heteroatoms such as nitrogen, oxygen, sulfur...), you should at least use 6-31G*.

Using different model chemistry for opt and NBO calculations

dummy@example.com (msn_dm) — Sat, 03 Jul 2021 18:02:27 +0000

I am intended to optimize a system including more than 100 atoms. The standard model chemistry that I am going to use for the optimization process is pbepbe/6-31G. Since the structural properties do not depend on absolute energy differences (but relative energy differences), I am of the opinion that such a model chemistry is sufficient.

However, to account for long-range and dispersion effects to the total energy, I need a larger basis set and a functional belong to Rung four in Jacob's ladder (something like wB97XD/6-311G(2d) level of theory). I use the optimized structure obtained from the former calculation as input geometry for the NBO study.

Some people claim that both optimization and NBO model chemistry should be the same. The question is "why"?
I cannot understand the reason behind this idea. I am not intended to compare opt and NBO energies; so why should both calculations perform by the same method?