No, the potential energy curve for a diatomic is the electronic ground state energy as a function of nuclear separation and should be a smooth curve. Gaussian and other quantum chemistry programs get it smooth. The discontinuity may be a basis inconsistency between what was used on the two atoms, but when using completely consistent (identical) basis sets for both atom, and switching to a more properly bound one like Ar F minus the result is smooth, but the BSSE effect (overbinding in smaller basis) is very apparent.