Basis set superposition error?

[ProkopHapala]

Most of time I spend optimizing geometry of hydrogen-bonded complexes, like DNA basepairs. B3LYP+D3 / cc-pVDZ is very good mathod for that as long as counterpointwise corrections of basis-set superposition error is applied. Without counterpointwise correction much large basis set (like aug-cc-pVTZ) is required to achieve comparable accuracy.

On this page can be easily compared different methods and basisets which shows how importaint this correction is for weakly bonded complexes. I also attached plot for hydrogen-bonded complexes from S22 dataset (CP is with correction, unCP is without correction).

• Is that any possibility to do some BSSE correton with QUICK?
• Idea would be apply this correction during geometry optimization so that optimized geometry of the dimer is also correct.
  • For example in Psi4 there is this function for that.
• if there is not way how to do BSSE correction with QUICK, is there some possibility to use at least aug-cc-pVDZ basiset?

[Madu86]

Is that any possibility to do some BSSE correton with QUICK? Idea would be apply this correction during geometry optimization so that optimized geometry of the dimer is also correct. For example in Psi4 there is this function for that.

Unfortunately, this is not implemented in QUICK at the moment.

if there is not way how to do BSSE correction with QUICK, is there some possibility to use at least aug-cc-pVDZ basiset?

I have added this basis set into QUICK. You can either pull merzlab/master once PR #292 is merged or apply the attached patch. aug-cc-pvdz_update.patch.tar.gz

[agoetz]

Larger basis sets are now included by default in QUICK since f functions are supported. I am leaving this issue open, adding label "enhancement" as some sort of automatic BSSE correction would be a new feature.