Two new symmetry-adapted perturbation theories for the calculation of intermolecular interaction energies

Abstract

We outline two new symmetry-adapted perturbation theories (SAPTs) and present some results obtained with them. The first is superior to the symmetrized Rayleigh–Schrodinger (SRS) theory in that it corrects a fundamental defect of that theory, namely, that carried to infinite order the SRS theory cannot predict even the ground-state energy for most interacting atoms and molecules. The new theory includes correction terms which have a large cumulative effect, but which, order by order, make only small contributions. When applied to interacting closed-shell systems and truncated after first order in the wave function, it is equal in accuracy to the SRS theory. Thus, it provides both an understanding of why the SRS theory gives results of useful accuracy and justification for its continued use when truncated to low order. The second new SAPT also corrects the SRS theory's flaw, but achieves significantly greater accuracy than the SRS theory when truncated after first order in the wave function. Applied to the interaction between the open-shell Li and H atoms, a critical test case, the second theory gives the ground-state dissociation energy with an error of 1%, whereas the SRS theory is in error by 38%. The LiH molecule is a critical test case because its physical ground-state energy, like that of nearly all systems, lies in a continuum of accessible states which violate the Pauli exclusion principle.