Multiwfn forum

Marlen

Member

Registered: 2026-03-02

Posts: 4

NTOs from "plain text input"

Dear Professor Lu,

I have probably found a weird behaviour in MultiWFN. I'm trying to generate NTOs from an ORCA output (after converting .gbw into .molden with orca_2mkl .. -molden), and it works well, however, once I reformat the ORCA output into the "plain text input" as described in the section 3.21.A.3 of the most recent manual (p. 237), the resulting NTOs (in .molden format) are different. What could be the reason of such discrepancy? Do I format the plain text input in a wrong way? An example is attached. Thank you!

nitrochlorobenz.out
nitrochlorobenz.molden.input.zip
nitrochloro_NTO_1-2_from_out.molden.zip

nitrochlorobenz.txt
nitrochloro_NTO_1-2_from_txt.molden.zip

Best regards,
Roman

sobereva

Tian Lu (Multiwfn developer)

From: Athens, Greece

Registered: 2017-09-11

Posts: 2,217

Website

Re: NTOs from "plain text input"

Dear Roman,

It is not a bug of Multiwfn. You forgot to fix the MO numbering convention when converting information from ORCA output to plain text file. In ORCA, MO index starts from 0, while in Multiwfn, the index starts from 1. So, all MO indices in the plain text file provided to Multiwfn must also start from 1, that means you should add all MO indices by 1 in the current plain text file.

I will modify the manual to make ORCA users aware of this point.

Best regards,

Tian

Marlen

Member

Registered: 2026-03-02

Posts: 4

Re: NTOs from "plain text input"

Dear Professor Lu,

thank you very much for the clarification! That's indeed something really easy to miss

Best regards,
Roman

Marlen

Member

Registered: 2026-03-02

Posts: 4

Re: NTOs from "plain text input"

Dear Professor Lu,

I've got a small follow-up question. Now with the numbering fixed, the output NTO has AO coefficients exactly like the NTO generated from
"native" ORCA output, but for some reason, the eigenvalue of the NTO is exactly double. I realize it is not the real eigenvalue/energy but just the orbital contribution instead; I just would like to ask for a better understanding.

nitrochloro_fixed_numbering_NTO_1-2_from_txt.zip
nitrochlorobenz_fixed_numbering.txt

Best regards,
Roman

sobereva

Tian Lu (Multiwfn developer)

From: Athens, Greece

Registered: 2017-09-11

Posts: 2,217

Website

Re: NTOs from "plain text input"

Dear Roman,

This is because you didn't modify configuration coefficients to follow the convention of Gaussian, which is also the convention of the plain text file of Multiwfn. In the case of TDDFT of Gaussian, when reference state is closed-shell, the coefficients are normalized to 0.5 (because alpha and beta part are the same, only one of them is printed); however, in the plain text file you provided they normalize to 1.0. (In contrast, when Multiwfn directly loads configuration coefficients from ORCA output file, the convention is automatically converted).

Best regards,

Tian

Marlen

Member

Registered: 2026-03-02

Posts: 4

Re: NTOs from "plain text input"

Dear Professor Lu,

thank you very much for your answer again!

Best regards,
Roman