Answers are grounded in this paper and cite their sources.
We extend FROST (Fock radial orbital shooting with Taylor series) from atoms to diatomic molecules and obtain the Hartree–Fock (HF) energy of H2 at the internuclear distance R = 1.4 bohr to 103 significant figures, that of HeH+ to 101 and that of LiH to 26, each certified by an exact upper bound, where about 11, 9 and 12 figures were known before. In prolate spheroidal coordinates both nuclear cusps are absorbed and the nuclear attraction becomes linear. Expanding the orbitals and their pair potentials in Legendre polynomials of the angular coordinate η then turns the HF equations into ordinary differential equations in ξ that all have the form (ξ2−1)u′′+2ξu′ = g. This is the molecular counterpart of the radial equations of the earlier papers, with a regular singular point on the internuclear axis instead of the nucleus. The expansion of the orbitals is the only truncation besides the wall, and it is variational, while that of the potentials is exact, because Poisson’s equation is diagonal in the Legendre index. The equations are solved by the Taylor-series multiple shooting and homotopy of the atomic code, with no basis-set matrix and no self-consistentfield iteration. What is new is step control that resolves the rapidly growing Legendre modes of the high-order potentials, and a homotopy and wall continuation on a few channels only, after which a single stage on all channels, whose Jacobian is factorised in double precision, reaches the target accuracy. The certificate evaluates the energy of the piecewise-polynomial orbitals exactly: the oneelectron integrals are rational, and Neumann’s expansion of 1/r12 gives the Coulomb and exchange integrals as rationals plus rational multiples of logarithms of the knots, which are rounded outwards. The rate at which the energy converges in the number of Legendre channels is set by the complex singularities of the HF orbitals near each nucleus, found for helium in the first paper. Mapped onto the η interval, they predict the decay of the Legendre channels of H2, HeH+ and LiH to within 6%, with no adjustable parameter, and they show why LiH, and molecules of heavier atoms generally, need many more channels.
No takes yet. Share an insight, caveat, or question.
Adam L. Baskerville (2026) studied this question.
Synapse has enriched 8 closely related papers on similar clinical questions. Consider them for comparative context: