SCF convergence for large hydrocarbon

[Researcher Joe]

Dear All,

I am trying to calculate single point energy of a large hydrocarbon molecule using B3LYP dft with 6-31G basis. The structure was optimized before with same functional and 3-21G basis. The SCF calculation seemingly does not converge (I attach input and output). Any thoughts on how to fix it are highly appreciated.

Thank you.

[Edoardo Aprà]

Here is a modified input that converges with the 6-31g basis.

I did symmetrize the geometry since it was fairly close to c3v (this modification does not necessarily help for the convergence issue, but it speeds up the calculation).

I have modified some of the convergence options and tightened some of the parameters (Coulomb screening and grid)

[Researcher Joe]

Thanks Edoardo, it works! I am curious what the line "convergence fast" does? Is it somewhere in the manual?

[Edoardo Aprà]

No, this option is not mentioned in the manual yet (thanks for the feedback).

convergence fast turns on a series of parameters that most often speed-up convergence, but not in 100% of the cases.

Here is an input snippet that would give you the same result as convergence fast

...

dft

convergence lshift 0. ncydp 0 dampon 1d99 dampoff 1d-4 damp 40

end

set quickguess t

task dft

[Researcher Joe]

Great! Thank you for the explanation.