Higher order SAPT with RI for heavier elements

The Psi4 manual for SAPT calculations states the following:

“For heavier elements ( i.e. , second-row and beyond), the RI auxiliary basis is unsound for this role (insufficiently flexible).”

Is there some paper, where this issue is further discussed, so that I could see how it influences the results of DF-SAPT2+3 compared to SAPT2+3?

I saw a paper regarding DF-MP2 with Dunning basis sets and in that case the RI was shown to introduce negligible errors compared to the errors due to the choice of basis set [1]. So I am wondering why in the case of SAPT RI is a problem.

[1] F. Weigend, A. Köhn, C. Hättig, Efficient use of the correlation consistent basis sets in resolution of the identity MP2 calculations, J. Chem. Phys. 116 (2002) 3175–3183. doi:10.1063/1.1445115.

I believe the “problem” is that by default the DF-SAPT programme in Psi4 uses the same auxiliary basis for density fitting both the Fock type terms AND for the correlated terms. In an ideal world you would use a separate JKFit auxiliary basis for the Fock terms and an RI (I prefer to use the name MP2Fit, but it’s RI in the Psi4 manual) auxiliary basis for the correlated terms.

Somewhat fortuitously, this doesn’t cause problems for lighter elements. But, it does cause issues for heavier elements. The manual details a workaround if you’re doing SAPT0, but there isn’t one for higher-order SAPT.

1 Like