Dear Mustafa,

As far as the method is concerned my personal opinion is to go with b3lyp. Now about the basis it depends on the exact system and the accuracy that you need... You can have a look in the portal https://bse.pnl.gov/bse/portal and find the basis set that fits to your needs. Remember, if you want to compare your values for the different systems you need to use the same method + basis set for all your calculations!!!

I hope this is helpful.

B3LYP is a good choice, however it is a good idea to add atom-pairwise dispersion corrections to the energy (D3). Moreover in such systems You should use at least triple zeta basis set to get good results, e.g. 6-311G**.

Dear Mustafa.

I am utterly against B3LYP, especially when it is coupled with Grimme correction. There are some articles which show the failure of B3LYP and B3LYP-D3. Also, all the time I have problem with the convergency of Minesotta functionals, altought it has been shown that M062X-D3 is the best hybrid functional. I suggest double-hybrid functionals especially DSD-BLYP-D3 if not Gordon functioan wb97xd is a good choice. Generally, the quality of the results of DFT+Disp depends on the parametrization!. Do not think if you apply Dispersion corrections you get necessarily the reliable results!...also, I do not know your sources allow you to use large basis set or not, if yes I suggest you Dunning basis set augmanted TZ one

I want to use Gaussian09 for binding energy calculations of my system and I want to include DFT-D3 in the calcualtion. If any one knows how use the B3LYP-D3 in Gaussian09, please help me. Thanks

Dear Sadanandam,

I do not suggest you B3LYP, use another one, coupling of Grimme correction with B3LYP makes the results worse. If you stick to it, use this command line

# b3lyp/xxxx IOp(3/124=30) keywords.