I want to optimize the azobenzene structure, but I'm not sure which functional/basis set to use. This molecule has 24 atoms, 12 carbons, 10 hydrogens, and two nitrogen atoms. Is B3LYP/6-31G* reliable and if yes, why?
Thank you!
Hi Mohammadreza Sharghi ,
depending on the program you are using (for example orca), you can use the composite methods like PBEh-3c, B97-3c or r2SCAN-3c. They are well balanced methods with a fixed combination of functional, basis sets and additional corrections.
If you can not use the methods, try to use larger basis set. Your molecule is small so you can go to triple-zeta. Maybe avoid the old fashioned Pople basis sets and switch to Ahlrichs (def2-SVP, def2-TZVP, def2-QZVPP).
As for the functional, B3LYP geometries should be fine, but maybe even BLYP or PBE geometries might ok, if you are using a triple-zeta basis set.
Good luck and regards
Conrad
Thank you Conrad Hübler, what I was thinking about was to first optimize the structure, and then do a potential energy surface scan, but I was wondering if I optimize with B3LYP/6-31G*, and then do PES scan, would the results be reliable? I'll scan with wB97XD/def2TZVP, after initial optimization!
Well you should first optimise using a cheaper method or smaller basis set and then reoptimise using a more accurate method and larger basis set. So put a final optimisation using wB97XD/def2TZVP on top of your optimised structure using B3LYP. If you have def2TZVP already there, you can keep it consistent and use def2SVP for the first optimisation.
If you can afford it, give a Double hybrid with Spin component scaling and Dispersion correction a try. DSD functionals might nowadays be the best functionals for potential energy calculation.
Mixing different methods for a PES scan might be unreliable in the end, most likely in case the PES is not parallel across different methods. Might be unlikely for rigid molecules, however.
- Badges
- Science topic