In 1969, Philip W. Anderson introduced a self-consistent pseudopotential method for localized orbitals and used it to compute the Hückel resonance integral of benzene from first principles, obtaining β ≈ −2.43 eV — already in agreement with semiempirical values fitted to electronic spectra. On this basis he argued that the large correlation-induced renormalizations of π-electron one-electron parameters estimated earlier by Linderberg and Öhrn from single-particle propagator calculations cannot be correct, while conceding that his one-electron calculation could not by itself quantify the role of electron correlation: his screening potential itself contains a Wigner correlation hole, and his empirical comparison target is a configuration-interaction-fitted effective parameter. The open problem is to determine quantitatively how the effective one-electron description of π electrons in conjugated hydrocarbons decomposes into one-electron and correlation contributions — for the effective β parameter and for the π-electron excitation spectrum — and thereby to adjudicate between the two estimates. No published work performs the required decomposition, and the two research traditions behind these estimates have not intersected on this question.

Background

Hückel theory (1931) describes the π electrons of planar conjugated hydrocarbons — the electrons occupying the carbon 2p atomic orbitals perpendicular to the molecular plane — with a one-electron tight-binding model: each carbon atom contributes one orbital, neighboring orbitals couple through a single empirical resonance integral β, and the secular (eigenvalue) equation is written as if the atomic orbitals were orthonormal, even though the overlap S between neighboring 2pπ orbitals is far from negligible (S ≈ 0.25 in benzene). Fitted to electronic spectra, |β| ≈ 2.3–2.6 eV, and the theory reproduces bond orders, charge distributions, and level patterns remarkably well. This posed the justification problem Anderson's 1969 paper addressed: why does an overlap-free, nearest-neighbor one-electron form built from almost-atomic orbitals work?

Anderson's method constructs localized orbitals — wave functions centered on single atoms, analogous to the Wannier functions of solid-state physics (localized basis functions built from the eigenstates of an energy band) — as eigenfunctions of a one-electron self-consistent pseudopotential Hamiltonian. A pseudopotential is an effective potential that replaces the strong attraction of an atomic core by a much weaker net potential by adding a repulsive term that cancels the attractive part in the core region; Anderson's version makes the potential on every non-home atom depend self-consistently on that atom's localized orbital. The resulting orbitals are nonorthogonal but obey an exact secular equation of Hückel form with no overlap terms, and they turn out to be extremely close to unperturbed atomic orbitals: for benzene, the first-order distortion changes the neighbor overlap by δS ≈ 0.003 on S ≈ 0.25 and β by less than 0.01 eV. Using Slater-type 2pπ orbitals (exponent 1.625) and a core Coulomb potential screened à la Hartree and corrected by a Wigner exchange-correlation hole — the device of screening the electron–electron interaction by removing one full electron of charge in the neighborhood of the interacting electron, an early local correlation correction — Anderson obtained β = −0.177 Ry ≈ −2.43 eV at the experimental C–C distance of 1.40 Å.

The open correlation question is set by two opposing 1960s estimates. Linderberg and Öhrn (1965) applied single-particle propagator methods — Green's-function techniques that compute ionization and excitation energies including the self-energy corrections generated by the other electrons — to alternant hydrocarbons (conjugated systems whose π network is bipartite, so the molecular orbitals pair into mirror-related bonding and antibonding levels). Their estimates implied large correlation-induced renormalizations of the π-electron one-electron parameters, in the solid-state language of mass renormalization — the reduction of a quasiparticle's effective hopping or band width by electron–electron interactions, familiar from transition-metal d bands, where renormalization factors of order 1.5 occur. Anderson's counterargument was indirect: the semiempirical β is not a bare one-electron quantity but a parameter fitted within a configuration-interaction (CI) treatment of the excited states — CI expands the correlated wave function as a combination of determinants built from the one-electron orbitals — which he judged already to include correlation at roughly the level of the random-phase approximation (RPA, the standard summation of ring diagrams describing a dynamically screened electron–electron interaction); and his one-electron value, computed with a potential that itself contains a correlation hole, already matched it. He concluded that the large Linderberg–Öhrn effects cannot be correct, while explicitly leaving the actual correlation problem — an RPA or CI treatment built on the orthogonalized (Wannier-type) functions — for future work, together with applications to heteropolar molecules (the one-electron α, or Coulomb-integral, parameters), spin densities, structural effects, and σ-bond frameworks. The Pariser–Parr–Pople (PPP) model — the standard π-electron model that augments Hückel theory with explicit electron–electron Coulomb interactions on the molecular framework — provides the natural theater in which a correlated treatment can be compared against a fixed-parameter one-electron theory.

The unresolved target is therefore: how large, quantitatively, is the electron-correlation contribution to the effective Hückel parameters and to the π-electron excitation spectra of conjugated hydrocarbons — small enough that the one-electron localized-orbital calculation already captures the effective parameter, as Anderson claimed, or large enough to constitute a genuine renormalization, as Linderberg and Öhrn estimated?

Problem Statement

Determine quantitatively the role of electron correlation in the effective one-electron (Hückel) description of π electrons in conjugated hydrocarbons, and thereby adjudicate the Anderson–Linderberg–Öhrn discrepancy.

Specifically, under a controlled correlated treatment of the localized-orbital (orthogonalized-function) type or an equivalent effective-Hamiltonian scheme, with the scheme and its conventions declared explicitly:

  1. Define and justify a decomposition of the effective π-electron one-electron parameters — in the first instance the benzene resonance integral, whose one-electron localized-orbital value is β ≈ −2.43 eV at a C–C distance of 1.40 Å — into a one-electron part and a correlation correction, making explicit (a) how the correlation already embedded in the screened one-electron potential (the Wigner exchange-correlation hole of one full electron) is allocated between the two parts, and (b) which computed object is identified with the fitted semiempirical target, given that the latter is a configuration-interaction-derived effective parameter (approximately the centroid of the singly excited configurations) rather than a bare one-electron matrix element.

  2. Determine: (a) the decomposition computed quantitatively for at least the canonical reference system, benzene, under one stated convention and to a stated numerical tolerance for the correlation contribution to the benzene resonance integral; (b) an estimate of the sensitivity of that result to the orbital, potential, and scheme conventions; and (c) the corresponding correlation corrections to the π-electron excitation spectrum, compared with experiment.

  3. State and justify a verdict on whether the large correlation-induced renormalizations of the π-electron one-electron parameters estimated by Linderberg and Öhrn (1965) are real, identifying the origin of any discrepancy between their propagator estimates and the one-electron result, and stating the verdict relative to the decomposition convention adopted in item 1 — showing it to be independent of the orbital, potential, and scheme conventions considered in item 2, or else reporting explicitly how it depends on them. The verdict must be given against an explicit decision rule, declared before the decomposition is computed, that fixes what counts as a small versus a large correlation correction — a stated threshold on the correlation contribution to the resonance integral, or an equivalent criterion — and the rule itself must be justified.

The question is not whether correlated wave-function methods can reproduce benzene's excitation spectrum — they can — but whether the decomposition of the empirically successful one-electron parameterization into one-electron and many-body parts is small or large, and why.

References

  1. Philip W. Anderson, "Localized Orbitals for Molecular Quantum Theory. I. The Hückel Theory", Physical Review 181, 25–32 (1969), DOI: 10.1103/PhysRev.181.25, https://doi.org/10.1103/PhysRev.181.25
  2. Erich Hückel, "Quantentheoretische Beiträge zum Benzolproblem. I.", Zeitschrift für Physik 70, 204–286 (1931), DOI: 10.1007/BF01339530, https://doi.org/10.1007/BF01339530
  3. Rudolph Pariser and Robert G. Parr, "A Semi-Empirical Theory of the Electronic Spectra and Electronic Structure of Complex Unsaturated Molecules. I.", The Journal of Chemical Physics 21, 466–471 (1953), DOI: 10.1063/1.1698929, https://doi.org/10.1063/1.1698929
  4. J. A. Pople, "Electron interaction in unsaturated hydrocarbons", Transactions of the Faraday Society 49, 1375–1385 (1953), DOI: 10.1039/TF9534901375, https://doi.org/10.1039/TF9534901375
  5. Jan Linderberg and Yngve Öhrn, "Improved single-particle propagators in the theory of conjugated systems", Proceedings of the Royal Society of London. Series A. Mathematical and Physical Sciences 285, 445–456 (1965), DOI: 10.1098/rspa.1965.0116, https://doi.org/10.1098/rspa.1965.0116
  6. Yngve Öhrn and Jan Linderberg, "Propagators for Alternant Hydrocarbon Molecules", Physical Review 139, A1063–A1068 (1965), DOI: 10.1103/PhysRev.139.A1063, https://doi.org/10.1103/PhysRev.139.A1063
  7. Jan Linderberg and Yngve Öhrn, "Derivation and Analysis of the Pariser–Parr–Pople Model", The Journal of Chemical Physics 49, 716–727 (1968), DOI: 10.1063/1.1670129, https://doi.org/10.1063/1.1670129
  8. Osamu Tanimoto and Katsumi Shimada, "Calculation of orbital energies in π-electron systems with the Green's function method", Molecular Physics 23, 745–763 (1972), DOI: 10.1080/00268977200100741, https://doi.org/10.1080/00268977200100741
  9. Osamu Tanimoto and Kaoru Toyoda, "Calculation of lower excited electronic levels of benzene with the Green's function method", Molecular Physics 24, 1257–1264 (1972), DOI: 10.1080/00268977200102351, https://doi.org/10.1080/00268977200102351
  10. J. Čížek, J. Paldus, and I. Hubač, "Correlation effects in the low-lying excited states of the PPP models of alternant hydrocarbons. I. Qualitative rules for the effect of limited configuration interaction", International Journal of Quantum Chemistry 8, 951–970 (1974), DOI: 10.1002/qua.560080610, https://doi.org/10.1002/qua.560080610
  11. Marcello Baldo and Renato Pucci, "Self-consistent many-body theory for π electron systems. I. The ethylene molecule", The Journal of Chemical Physics 67, 4747–4755 (1977), DOI: 10.1063/1.434601, https://doi.org/10.1063/1.434601
  12. J. D. Weeks, P. W. Anderson, and A. G. H. Davidson, "Non-Hermitian representations in localized orbital theories", The Journal of Chemical Physics 58, 1388–1395 (1973), DOI: 10.1063/1.1679371, https://doi.org/10.1063/1.1679371
  13. Herbert Schlosser, "Pseudopotential Approaches to Localized Orbitals for Polyatomic Systems", The Journal of Chemical Physics 53, 4035–4039 (1970), DOI: 10.1063/1.1673875, https://doi.org/10.1063/1.1673875
  14. W. M. C. Foulkes and D. M. Edwards, "Perfect localized basis functions for solids: chemical pseudopotentials and the Kronig-Penney model", Journal of Physics: Condensed Matter 5, 7987–8004 (1993), DOI: 10.1088/0953-8984/5/43/013, https://doi.org/10.1088/0953-8984/5/43/013
  15. R. G. Woolley, "The density functional theorem and Hückel theories", Philosophical Magazine B 69, 745–753 (1994), DOI: 10.1080/01418639408240143, https://doi.org/10.1080/01418639408240143
  16. Johan Lorentzon, Per-Åke Malmqvist, Markus Fülscher, and Björn O. Roos, "A CASPT2 study of the valence and lowest Rydberg electronic states of benzene and phenol", Theoretica Chimica Acta 91, 91–108 (1995), DOI: 10.1007/s002140050090, https://doi.org/10.1007/s002140050090
  17. Alexander V. Mironenko, Lanie Leung, and Jiqing Zhuang, "Self-consistent equations for nonempirical tight-binding theory", The Journal of Chemical Physics 163, 164119 (2025), DOI: 10.1063/5.0276043, https://doi.org/10.1063/5.0276043