Multiwfn forum

wham09

Member

Registered: 2025-09-19

Posts: 39

Clarifications on diffuse functions for multiwfn

Dear Prof. Lu,

I’d be really grateful if I could get additional clarifications on dealing with diffuse functions (and some additional questions on ZPE correction)

  1. If I use the complete set of ma-def2 series that you provided, how should I cite it when I publish?

  2. If the net charge of my molecule is overall -1, but each atom is not significantly negatively charged (ADCH charge > -0.5), diffuse functions are not required. Correct?

  3. In my reaction A + B -> [A-B], some atoms in the product [A-B] are significantly negatively charged, but not in the reactant forms (A and B). For the sake of reaction energy calculation, should these atoms be calculated using diffuse functions for both reactant (A and B) and product ([A-B]) calculations?

  4. A molecule is overall neutral, and there is no atom with a significant negative charge. But it is a tight ion-pair, in which the cation and anion counterparts are electrostatically bound and deform each other's structure. The binding is strong, but it is still non-covalent. Should diffuse functions be widely used?

  5. When using Shermo for a solute, the sophisticated way of input for Shermo setting is:
E = E(gas, high-level basis set) + [E(SMD, M05-2X/6-31G*) - E(gas, M05-2X/6-31G*)] + 1.89 kcal/mol.
  Am I correct?

  6. When I do opt-freq calculations using SCRF(IEFPCM), the ZPE values are, not surprisingly, different from the values from gas-phase optimization. This will be same for the molecules in F38/10 database used to obtain the scale factor. Then, shouldn't the ZPE scale factor become different?

  7. For wavefunction analysis of a solute, it is recommended to calculate with high-level basis set & SMD. But at the same time, for the calculation of the solute's solvation energy, the same structure should be calculated with M05-2X/6-31G*. Am I correct?

Last edited by wham09 (2025-10-30 03:01:18)

sobereva

Tian Lu (Multiwfn developer)

From: Athens, Greece

Registered: 2017-09-11

Posts: 2,144

Website

Re: Clarifications on diffuse functions for multiwfn

1 Theor Chem Acc (2011) 128:295–305

2 Correct

3 Should

4 If the ion pair is not of interest, you can ignore diffuse functions; if it is important, then diffuse functions should be added to heavily negatively charged atoms in the ion pair (at least when calculating single point)

5 Yes

6 AFAIK, there is no published frequency correction factor for solvent environment, usually  it is assumed that the optimal scale factor in solvent environment does not differ notably with respect to that in gas phase.

7 Correct. But it is noted that most wavefunction analysis methods have relatively low requirement on basis set (the level for geometry optimization is already adequate for most wavefunction analyses).

wham09

Member

Registered: 2025-09-19

Posts: 39

Re: Clarifications on diffuse functions for multiwfn

Thank you very much for the clarification. I have a follow-up question for just #4:
  My situation is as following:
   1) The ion-pairing is important in my molecule. I want to study the interaction between the cationic and anionic counterparts.
   2) One atom (out of total 70~80 atoms) is significantly negatively charged (calc. level: PBE0-D3(BJ) / ma-TZVP or def2-TZVP)
  In the multiwfn manual on ETS-NOCV, it says “If there is no special reason, using diffuse functions is deprecated.” Does my situation fit into the “special reason”? Is ETS-NOCV absolutely meaningless when ma-TZVP is used?

Additional question: What if the ion-pairing is important, but there is no atom that is significantly negatively charged?

Last edited by wham09 (2025-11-05 03:01:18)

sobereva

Tian Lu (Multiwfn developer)

From: Athens, Greece

Registered: 2017-09-11

Posts: 2,144

Website

Re: Clarifications on diffuse functions for multiwfn

You can add diffuse functions for these specific atoms.

I don't understand your additional question.

wham09

Member

Registered: 2025-09-19

Posts: 39

Re: Clarifications on diffuse functions for multiwfn

About the additional question, sorry about the confusion. What I meant is as below:
My molecule is an ion-pair complex, consisting of a cation and an anion close together (total charge=0). The inter-fragment interaction between the cationic and the anionic counterparts is the primary interest. At the same time, the cationic and anionic charges in the two fragments are delocalized enough so that none of the individual atoms (in the structure of the ion-pair complex) are significantly negatively charged (ADCH charges of all atoms more positive than -0.5).

sobereva

Tian Lu (Multiwfn developer)

From: Athens, Greece

Registered: 2017-09-11

Posts: 2,144

Website

Re: Clarifications on diffuse functions for multiwfn

To fully guarantee the accuracy and avoid comments from reviewers, adding diffuse functions to atoms in the anionic part is the safest way, though simply using def2-TZVP for all atoms should also work fine.