Multiwfn forum / Dispersion correcting factor in DFT functionals

Re: Dispersion correcting factor in DFT functionals

dummy@example.com (saeed_E) — Fri, 12 Jan 2024 20:28:26 +0000

Thanks a lot.
Saeed

Re: Dispersion correcting factor in DFT functionals

dummy@example.com (sobereva) — Fri, 12 Jan 2024 20:03:33 +0000

saeed_E wrote:

I would like to know that the zero-damping function is requested using "em=gd3"

Yes

Re: Dispersion correcting factor in DFT functionals

dummy@example.com (saeed_E) — Fri, 12 Jan 2024 19:43:49 +0000

Dear Tian,
Too many thanks for your highly kind attention to reply my post with your valuable comments.
Indeed, I asked some questions to make myself quite ensure. Please let me know that all above-mentioned discussions are correct or not. I would like to know that the zero-damping function is requested using "em=gd3" keyword. Is it true?

Sincerely,
Saeed

Re: Dispersion correcting factor in DFT functionals

dummy@example.com (sobereva) — Fri, 12 Jan 2024 19:07:30 +0000

I never use D2, which is completely out-of-date.
For most functionals, always use em=gd3bj to employ DFT-D3(BJ). For B3LYP, it will be B3LYP-D3(BJ)
Only for a few functionals, including M06-2X, only zero-damping form of DFT-D3 can be used, then it can be written as M06-2X-D3 or more expicitly, M06-2X-D3(0).

Dispersion correcting factor in DFT functionals

dummy@example.com (saeed_E) — Thu, 11 Jan 2024 11:50:43 +0000

Dear Tian,

It is well established that the lack of dispersion inclusion within DFT computations can be resolved through an empirical dispersion term. Generally, two types of dispersion corrections could be included: 1- using the original damping function (known as zero-damping function); 2- using the newer damping function known as Beck-Janson (BJ) damping function. Consequently, for instance,
#B3lyp/6-31G(d) em=gd2 or gd3 requests the D2 or D3 zero damping function to include dispersion term in the B3LYP functional. On the other hand, B3lyp/6-31G(d) em=gd2(BJ) or gd3(BJ) requests the D2 or D3 beck-janson damping function to include dispersion term in the B3LYP functional. All these statement correct?

Meantime, for some DFT functionals already included some extent of dispersion through parametrization, using GD3(BJ) leads to an overestimation of dispersion effect. Thus, these functionals such as M06-2X should be included just using zero-damping function as #M062X/6-31G(d) em=GD3. From writing point of view in a manuscript, the #M062X/6-31G(d) em=GD3 is equal to M06-2X-GD3 or M06-2X-D3(0). Is it true?

Sincerely yours,
Saeed