Multiwfn forum / Definition of “compatibility” with diffuse functions

Re: Definition of “compatibility” with diffuse functions

dummy@example.com (sobereva) — Wed, 29 Oct 2025 01:15:09 +0000

1 ADCH charge more negative than -0.6 or -0.7. ADCH charge can be calculated by Multiwfn using 7-11-1.
2 It is reasonable
3 Should be
4 This fully depends on which wavefunction analysis method will be used. Only less than 30% analysis methods in Multiwfn is incompatible with diffuse functions. For example, if calculating Mayer bond order and Mulliken charge, then def2-TZVPD should no be used, unless the atoms with diffuse functions are very far from the region of interest.

Re: Definition of “compatibility” with diffuse functions

dummy@example.com (wham09) — Tue, 28 Oct 2025 04:54:48 +0000

1. What is the bar for significantly negatively charged atom? mulliken or nbo charge? For example, the atom with the most negative nbo charge in my molecule has nbo charge of -0.59.
2. Would it work to use def2-TZVPD for only the significantly negatively charged atom and def2-TZVP for the rest?
3. If the molecule is overall neutral, but contains an atom that is significantly negatively charged, should diffuse functions be used?
4. If I use def2-TZVPD for only one atom in a molecule with total 14 atoms, would the output be incompatible with wavefunction analysis?

Re: Definition of “compatibility” with diffuse functions

dummy@example.com (sobereva) — Tue, 28 Oct 2025 04:26:46 +0000

For anions (more specifically, for significantly negatively charged atoms), def2-TZVP is inadequate, at least for evaluating energy and properties related to electronic structure.

Re: Definition of “compatibility” with diffuse functions

dummy@example.com (wham09) — Tue, 28 Oct 2025 04:23:38 +0000

Would you also agree with “it retains the merit of diffuse basis sets for calculation of, for example, anions”?

Re: Definition of “compatibility” with diffuse functions

dummy@example.com (sobereva) — Tue, 28 Oct 2025 04:14:11 +0000

According to my experience, def2-TZVP is safe for any wavefunction analysis.

Re: Definition of “compatibility” with diffuse functions

dummy@example.com (wham09) — Tue, 28 Oct 2025 04:00:47 +0000

I see. Then, would it be correct to state that def2-TZVP can be used for Multiwfn functions that are incompatible with diffuse functions, but it retains the merit of diffuse basis sets for calculation of, for example, anions?

Re: Definition of “compatibility” with diffuse functions

dummy@example.com (sobereva) — Tue, 28 Oct 2025 03:00:49 +0000

Usually, a GTF with exponent less than 0.05 is regarded as a diffuse function.
Although def2-TZVP doesn't contain diffuse functtions, it is well known that it is more diffuse than 6-311G series. Note that diffuseness of 6-311G is inadequate, making its result frequently unsatisfactory (see J. Chem. Phys., 91, 7305 (1989)).

Definition of “compatibility” with diffuse functions

dummy@example.com (wham09) — Tue, 28 Oct 2025 02:19:10 +0000

Dear Prof. Lu,

I know that many functions of Multiwfn are incompatible with diffuse functions. But I’m confused with the definition of diffuse functions.
What I attached here is the basis set information for the same atom (boron) in the same molecule, left one using 6-311+g** and right one using def2-TZVP.
Although 6-311+g** is said to have diffuse functions (and def2-TZVP does not), the def2-TXVP setting also contains a S-type function with an exponent smaller than 0.1 (the last S-type in the right figure, exponent=0.06056). Also, the def2-TZVP setting contains an additional D-type function (exponent=0.199) that is more “diffuse” than the sole D-type function in the 6-311+g** setting (exponent=0.401).
So, what is the exact standard for calling a basis set diffuse or not?