Hi all,
I've noticed some functionals differ by as much as 2 kcal/mol in hydrogen bond disassociation when the relative total energy is corrected using the BSSE. Unfortunately, the literature I have accessed has all been gas phase and further more, Gaussian will not allow me to mix counter poise with a PCM.
Any idea how I can go about calculating the BSSE of a dimer system in an implicit solvent model?
Many thanks
Anthony
Hello Anthony,
One option for reduce the BSSE error is increase the basis set. For example, with 6-31G++(d) you will have an important BSSE but with i.e. 6-311++G(2df,2p) should reduce the error. Other possible basis set, that I think are much better in general, are the Dunning ones, one possibility is aug-cc-pVTZ . The idea should be increase the basis all that you can, reducing BSSE error. Of course, is depending on your resources and molecule size, but you can increase the accuracy.
Hi Joaquim, thanks for your reply.
I am using a basis set of 6-311++g(d,p) and I am a little reluctant to make it any bigger because the molecule is 65 atoms big.
Then, I think that the BSSE error should be the same in PCM model that in gas-phase. Maybe, doing the correction and seen the energetic effect with Counterpoise in gas-phase you could add a correction factor to energy in pcm model.
Hww do I go about getting that correction term? I know how to do it from the start with an unoptimised geometry in gas phase. But I am a little confused with the idea of used a correction term of a gas phase structure on its counterpart solvent phase structure. Would you suggest I take a single point energy calculation using the optimised geometry of the PCM+molecule system (obviously turn off SCRF) with counterpoise? Or are you suggesting I take the optimised structure and reoptimise it with no solvent and with counterpoise?
Thanks :)
I like more taking the PCM optimized structure (even so, the structure do not differ so much, maybe some distances could change a little bit). I don't know if there are any paper doing these kind of corrections. I've never tried something like that, was an idea only. The question you have made is very interesting and I hope some researcher could explain if there is a known solution.
I've taken SCF (kcal/mol) values of:
optimised with solvent: -721724.070855038
single point of the optimised but with solvent off: 721724.108380113
single point (BSSE) of the optimised but with solvent off: -721708.651864652
There is quite a bit of a difference between the BSSE corrected gas SP and the non-BSSE gas SP. However, given that I will be using these calculations to work out the disassociation energy of a hydrogen bond of water and the molecule, I don't think the error will be noticeable when I calculate E_hydbond = E_complex - (E_H2O + E_molecule). I am going to see how much of a difference the 6-311++G(2df,2p) basis set makes.
- Badges
- Science topic