Abstract
It is shown that orthogonal atomic orbitals, as introduced by Wannier for solid‐state problems and by Löwdin for molecular problems, do not lead to a simple way of eliminating the non‐orthogonality problem in the Heitler‐London method. Thus, if a calculation of the H2 molecule is made using orthogonal atomic orbitals in place of ordinary atomic orbitals, proceeding by the simple Heitler‐London method, the resulting diagonal energy of the lowest state shows no energy minimum. The orthogonal atomic orbitals are suitable for use, however, if we make a complete perturbation calculation, and for this purpose they are as convenient as LCAO's, but not appreciably superior.