Hey all!
I am very much confused in choosing basis sets for C and O atoms in Gaussian package. In some papers diffusion and polarization are used for these atoms while not in others. So I am not able to decide whether I should take diffusion and polarization while studying adsorption of O2 and CO in simple metal clusters. Any suggestion would be appreciated.
Thank you.
Hey!
The use of polarization functions is (almost) mandatory nowadays to achieve accuracy in DFT calculations. In your system you have two complex small molecules respect to the electronic structure (O2 and CO) and polarization functions would help in the good description of the M-O2 or M-CO interaction. Regarding diffuse functions, I would recommend to use them if your system is not too big and/or if you have charged species along the reaction profile (especially negatively charged species).
Hope it helps you!
Ignacio
Dear Megha
If you are doing calculations for metals complexes, then for C and O, only polarization functions are good enough, (use basis functions like 6-31G(d,p)). Diffuse functions are more important for calculations with anions or very weak interactions like hydogen bond, halogen bond, chalcogen bond etc.
Thank you
Thank you Ignacio and Dipankar.
Dear Ignacio, my system contains 6-atoms only. According to you, accuracy would be more if I take diffusion also into picture and in such a small system taking diffusion would not take much time. So, I think it would be great if I take 6311++G(d,p) basis set for my calculations.
Thank you again. Your suggestions means a lot.
Hi again!
In that case (only 6 atoms), you can even try bigger basis set to find the basis set limit using a correlation consistent basis set (Dunning basis set for example). In any case, I would use at least triple-zeta basis set with polarization and diffuse functions!
Best,
Ignacio
Hello!
In that case (triple zeta), I should do some testing using D95V (double-zeta) and AUG-cc-pVTZ type basis sets.
Thank you again for the help.
What level of theory are you looking at? The Pople sets are somewhat antiquated. Use Dunning sets (cc-pVXZ series) for post-Hartree-Fock, and Karlsruhe (def2 series) or Jensen's polarization consistent basis sets (pc-n or pcseg-n series) for Hartree-Fock and density functional theory studies.
Hello Susi!
I am looking for adsorption of CO in Group-4 clusters using DFT based methods.
Thank you.
Then you probably want to give a closer look to the Karlsruhe sets, since they employ effective core potentials for relativistic effects and are available across pretty much the whole of the periodic table. You should use the same basis set for all the atoms in your system, i.e. not just C and O but also the heavy atoms.
The Karlsruhe sets appear to be available in Gaussian: "The newer redefinitions of these basis sets in [Weigend05, Weigend06] are requested with the keywords Def2SV, Def2SVP, Def2SVPP, Def2TZV, Def2TZVP, Def2TZVPP, Def2QZV, Def2QZVP, Def2QZVPP, and QZVP. Note that Def2SVPP corresponds to the “def2-SV(P)” basis set in [Weigend05]; all other names follow those in the paper with the hyphen removed."
Thank You Susi!
But as far as I understand we should not use effective core potentials (ECP) for small atoms. Say If I am using LanL2DZ for clusters, then should I use the same for C and O? Because I have seen in literature also results are much better when we use ECPs for heavy atoms like ( Au, Ag) and basis sets like TZVP, QZVP, 6-311G etc. for smaller atoms like C, O, N etc.
Best,
Megha
The Karlsruhe sets don't use ECPs for small atoms, only heavy ones. LANL2DZ is a bad basis - it's only double zeta and lacks polarization functions altogether!!! That's why I said use Karlsruhe sets for all your atoms.
dear Megha Sharma Hi,
please can I obtain the lecture about how basis set related with heavy or light atoms, I mean factors which selection of basis set depend on.
thank you
- Badges
- Science topic