You just need to do calculations in the aug-cc-pVTZ and aug-cc-pVQZ (or, if this is too much, aug-cc-pVDZ) and then add the extrapolated correlation energy to the HF energy from the larger of the two basis sets. The necessary formula for 2-basis extrapolation for energy can be easily derived from fromula 2 in the paper of Halkier et al.:Chemical Physics Letters 286 1998. 243–252
Good luck! :)
I strongly recommend a three-point extrapolation (double, triple, quadruple zeta). We have used this approach rather successfully (see Eq. 1 in the attached reprint).
Certainly, the three point extrapolation is likely to be more accurate. I would also add, that for the aug-cc-pVnZ basis sets, the D,T extrapolation often gives a very different result than the T,Q or Q,5 ones, which is probably due to rather poor performance of the smallest (aug-DZ) basi set. Therefore I would say that while it is surely best to use the T,Q,5 extrapolation scheme, in cases for which the aug-cc-pV5Z calculations are impossible (for mere technical reasons) it is probably better to use the T,Q two-point extrapolation than the D,T,Q one )as the latter contains the uncertain results coming from the aug-cc-pVDZ calculations).
Best regards!
Marcin
Kindly tell me about the keywords which are used to get CBS extrapolated correlation energy while using Aug-cc-PVTZ set and plz attach the input or output file.
The kind of CBS extrapolation Marcin and I have been discussing is not based on a keyword. One performs two or three separate calculations on the same geometry with different basis sets (see Marcin's posts) and uses an appropriate equation to get the CBS energy estimate. Gaussian has a keyword "CBSExtrapolate" but I have not used it and cannot comment.
@B.R. ramachandran
which software did u used for carrying out these extrapolation calculations and for parameter optimizations.
@ Marcin Andrzejak
How to calculate extrapolated correlation energy. PLz explain.
Definition of correlation energy is Ecorr = Ex - E_HF, where Ex is the energy calculated at some post-HF level. If you need to get Ecorr after extrapolating to CBS limit, you should extrapolate the HF energy to CBS limit as well. As explained in the reprint, no software is used for extrapolation. You solve a system of 3 eqs. for 3 unknowns.
- Badges
- Science topic