I was trying to calculate difference between triplet and singlet for O2 molecule however obtained value is incorrect in comparison to literature (after that I would like to use this algorithm for different molecules). How to do it correctly? I was using DFT calculation, TD-SCF method B3LYP with 6-31G basis set...
Thank you in advance
Hi Michal,
Instead of using TD-SCF, why don't you calculate the singlet and the triplet in two separate calculations? In the case of the triplet, apart from the specification of the multiplicity in the line previous to the coordinates, which would be 3 in this case, you should also specify that you want to use an unrestricted (U) formalism within DFT. If you're using B3LYP, then you should write UB3LYP/6-31G.
In any case, given the size of the molecule, I'd recommend you to use a much larger basis set if you pretend to achieve good accuracy. For instance, try def2-TZVP (it is implemented in gaussian).
Hope it helps.
Sergi
Hello,
my couple cents: actually, it is possible to omit specifying the unrestricted formalism explicitly, even if you simply write B3LYP, the Gaussian software (G09, for instance) would use the unrestricted formalism for open shell systems by default. Also, for much more accurate calculations, I would recommend using such approach as CCSD(T) with basis sets like aug-cc-pVDZ or cc-pVDZ (implemented in Gaussian). Hope it is helpful, too.
Best regards,
Aleksey.
Hi Michael,
The poor results in comparison to literature data are a consequence of the small base set. If you use a larger base sets such as 6-311 + G (3df, 3pd), aug-cc-pV5Z, or Def2QZVPP (implemented in G09), you'll get a much better agreement with literature data. The method is also very important. I would recommend you try it with M062X and MP2, and then you may try CCSD(T) as well.
Hope this is helpful to you.
Regards,
Zoran
- Badges
- Science topic