Comment on “A possible definition of basis set superposition error”

In their recent Letter Davidson and Chakravorty proposed a framework for analysis of basis set effects which trouble calculations of potential energy surfaces for weakly interacting species. Their analysis suggests that the complete basis set interaction energy differs by some monomer- and dimer-type non-additive corrections from the conventional counterpoised interaction energy obtained in a finite basis set. In this Comment we point out that the dimer-type correction does not fit the definition of non-additivity applied for the monomer terms and that the sum of the non-additive corrections reduces to a component of the interaction energy. This component disappears if the finite basis set is flexible enough to describe all important terms of the interaction energy. Thus the discrepancy between the counterpoise-corrected and the complete basis set interaction energy is exclusively determined by the suitability of the basis set used to reproduce the interaction energy terms.