Energy refinement at different basis set

zako

Hello everyone,
I have studied a chemical organic reaction catalysed by the Lewis acid "SnCl4" using DFT methods. I have used mixed basis sets (6-311G(d,p) for the atoms C, H, O and Cl; and SDD for Sn atom). Is it possible to make energy refinement by performing a single point calculations at the advanced basis set Def2-TZVPP? If yes, what's the best way, Def2-TZVPP for all the atoms, or keep 6-311G(d,p) for the atoms C, H, O and Cl; and Def2-TZVPP for only Sn atom?




Many thanks in advance

dnlx

Other than the Sn atom, you should use the better basis set at least for any other atoms involved in bond formation or breaking in your reaction. Therefore, "Def2-TZVPP for all the atoms" is better than "Def2-TZVPP for only Sn atom".

sobereva

If you can easily afford the cost, simply using def2-TZVPP for all atoms is certainly the best choice.

zako

Thank you all very much for your valuable answers, and i would like to inquire about :
1- Isn't there any issue or contradictions performing single-point energy calculation using "Karlsruhe basis set" (Def2-TZVPP), despite i performed the optimization calculations using "pople basis set (6-311G(d,p) and ECP basis set (SDD))"  ?
2- Concerning the thermochemical values (Enthalpy, Entropy and Free Gibbs Energy), Can i use the electronic energy obtained using (Def2-TZVPP) to calculate a new the thermochemical values ? i mean add the new electronic energy to the old thermal corrections to get new values of thermochemical values. Is it a safe and correct procedure erroneous ?

dnlx

zako 发表于 2025-2-3 19:08
Thank you all very much for your valuable answers, and i would like to inquire about :
1- Isn't the ...

1. No issues or contradictions.
2. Yes you can and you should. This is a standard practice. See the second note on this page
https://www.faccts.de/docs/orca/5.0/tutorials/prop/thermo.html

This G-E(el) only depends on the geometry and frequencies, and thus can be computed in a lower level method, such as DFT, and later combined with the Eel
of a higher level method, such as DLPNO-CCSD(T) to achieve a very good prediction of the G0

zako

dnlx 发表于 2025-2-4 04:26
1. No issues or contradictions.
2. Yes you can and you should. This is a standard practice. See t ...

Dear dnlx, Thank you very much
Since you mentioned the DLPNO-CCSD(T) method for single point energy calculations, which is only can be used in ORCA program, i would like to how to switch from Gaussian calculations, since i performed the optimisation and vibration calculations using Gaussian package to Orca DLPNO-CCSD(T) calculation ?