On the accuracy of numerical Hartree–Fock energies

Abstract.

It is demonstrated that numerical Hartree–Fock (HF) energies reported in the literature in some cases have errors in the milliHartree range. The main cause of these errors is due to the use of too small a value for the ‘practical infinity’ parameter in the finite difference method for generating the results. By systematically investigating the convergence with respect to the computational parameters, HF energies accurate to at least 1 microHartree are generated for 42 diatomic systems containing first and second row elements, encompassing both cationic, neutral and anionic systems.