The physical curve has its bonding minimum near , with ionic endpoints at . Use electronic energies excluding the fixed . Define , , and . The one-electron contribution is
The Coulomb energy of the doubly occupied orbital is
so . For localized orthogonal orbitals, a transparent illustrative approximation neglects integrals involving the product and retains the on-site repulsion and intersite direct repulsion :
Since normally , spreading the occupied orbital over both atoms reduces the repulsion. At equilibrium the bonding one-electron term favours , so the minimum represents the Restricted Hartree-Fock method approximation to . At or , both electrons occupy one atom, giving the same ionic energy . At the orbital is antibonding and the one-electron contribution is raised.
The diagram below uses an illustrative set of integrals to show these landmarks; it is not a numerical calculation for the molecule. Zero orbital overlap alone does not force every exchange or three-index Coulomb integral to vanish, so the simplified expression is an explicit sketching approximation, not a consequence of orthogonality.
Figure 1.
Physical and noninteracting electronic energies as the occupied orbital changes
. Details.
No: the lowest dissociation channel is , which a doubly occupied helium-centred orbital can describe. In the heteronuclear molecular orbital diagram the helium site lies below the hydrogen site. As separation grows, the bonding orbital tends to helium , the antibonding orbital tends to hydrogen , and the two ground-state electrons remain in the lower orbital.
There is therefore no symmetry requirement to keep both sites equally occupied and no compulsory half-weight mixture of wrong ionic configurations at dissociation. The physical asymptote is already a closed-shell atom plus a bare proton. The alternative limit lies higher. A Restricted Hartree-Fock method curve approaches the Hartree–Fock helium energy rather than the exact correlated helium energy, so ordinary dynamic electron correlation is still missing, but this is not the homonuclear static-correlation catastrophe. In the preceding two-panel sketch, the helium hydride curves have the same qualitative dissociation channel with a small residual correlation offset, whereas the hydrogen restricted curve reaches a qualitatively wrong high limit.