basis set error

I had a problem trying to run a calculation on SnCl3 anion.  I used the same input as found in the following test input file:

MINP_SnCl2_def2-SVP_def2-ECP-28_DF-HF_geomopt

I just changed the structure to match SnCl3 and made the charge=-1.  I am getting the following error: 
 Basis set cc-pvtz for atom 1 (Sn) is not found!
 
 Fatal error in exec integ.
 Program will stop.

This is perplexing, as the MINP uses basis=def2-SVP.

For completeness, here is my MINP file:

# DF-HF geometry optimization for SnCl2 with the def2-SVP basis set and the def2-ECP-28 pseudopotential for Sn
basis=def2-SVP
calc=df-hf
mem=3000MB
gopt=full
charge=-1
orbloco=ibo

geom=xyz
4
model.mopin
Sn        0.000000000000        0.000000000000        0.000000000000
Cl        2.390000000000        0.000000000000        0.000000000000
Cl       -0.415020000000        0.000000000000       -2.353690000000
Cl       -0.415020000000       -2.038360000000        1.176850000000

(Note, I was having a similar problem trying to carry out a calculation using a Mo atom.  there must be something wrong with how I am setting the basis, but I have not been able to solve it yet.)

Thanks for any assistance with this!


 

The cc-pVTZ basis set is needed for the IBO localization since the minimal basis required for IBOs is constructed from this basis set. If you remove orbloco=ibo, it will run smoothly.

Does this mean I cannot construct IBOs for heavy elements such as Sn? On the basis set exchange, I do not see cc-pVTZ available for Sn, although cc-pVTZ-PP is available there. Thanks!!

I have not yet used IBOs for heavy element. Please check the literature what the recommendation is concerning heavy elements, and you can add the required basis sets by hand. Probably cc-pVTZ-PP is fine.

Thanks. How would I tell mrcc to look for cc-pVTZ-PP instead of cc-pVTZ when running orbloco=ibo? I don't see a keyword for that.

I have also a theory question: when running calc=LNO-CCSD(T) and orbloco=ibo or orbloco=pm, are the resulting localized orbitals in the MOLDEN file constructed with inclusion of electron correlation? I understand that mrcc may be generating the orbitals in the following way: density matrix -> cholesky guess -> ibo or pm. Is this correct?

Thanks very much.

For the moment, there no simple way to tell the program to use the PP basis sets. The corresponding code that generates the minimal basis set should be rewritten.
The IBOs are localized HF orbitals, correlation is not included. Yes, Cholesky localized orbitals are used as initial guess, but the HF density is used for that purpose.