Multiwfn forum

Heliumqt

Member

Registered: 2024-01-26

Posts: 3

Convergence Criterion Gaussian 16 C01 question

Dear Dr. Tien,

I am trying to perform a single point calculation on a large system using the M062x functional and def2-tzvpd basis set. However, I have tried every other convergence method (qc,xqc,yqc) etc and it did not work. However, I lowered the criterion to conver=6, and that seem to work. My question is would it be better if i set the criterion to conver=7 or conver=6 or conver=4? I am not as familiar with the default setting for single point in gaussian, and also not quite sure if increasing the number (hypothetically lets say convergence criterion in gaussian is conver=4) to conver=5 or conver=6, which of those two is less accurate or "relaxed" as compared to the default.

sobereva

Tian Lu (Multiwfn developer)

From: Beijing

Registered: 2017-09-11

Posts: 2,014

Website

Re: Convergence Criterion Gaussian 16 C01 question

If you just need to obtain electronic energy at a given structure, scf=conver=6 is acceptable for DFT calculation. Because when convergence threshold is reached in this case, usually energy has converged to <1E-7, which is fine enough.

PS: My given name is Tian