← Catalogue · The r₁₂ frontier
P31 Measured frontier 3/5 · frontière déclarée Frontière r₁₂

Portée dérivée — frontière r₁₂ constitutive

La portée dérivée bat Hartree-Fock à bas Z mais la frontière r₁₂ explicite est déclarée constitutive — structurelle, pas paramétrique — pour He au-delà de 2 %.

◫ Simulation figure(s)

Figure P31

JSON results — converted to tables

Machine-readable artefacts frozen by SHA-256. Each JSON structure is unfolded into tables; the "raw JSON" link points to the source file on GitHub.

p31_portee.json 17 table(s) · 71 rows Raw JSON on GitHub
ChampValeur
chantierP31
titrePortee derivee du facteur a deux corps — frontiere r12 ?

Explanation — context & formalism

P31 — The r₁₂ Frontier Declared Constitutive

Field: Atomic / Quantum Chemistry / Foundations Verdict: ⚠️ Frontier Declared Constitutive (3/5 formal, 0/3 rules)

Problem

Final test announced in P30: three entirely derived range constraints, zero parameter, tested on the kinematic factor u(r₁₂) = ½ r₁₂ e^(−βr₁₂) (Kato cusp imposed, factorization imposed at large distance).

Three rules tested

Rule (0 parameter)βE_min (Ha)vs split-ζ −2.8347
R1 — orbital scale: β = (a+b)/21.65−2.790worse by 0.045
R2 — orthogonality: ⟨u′ cos θ₁₂⟩ = 02.53−2.821worse by 0.013
R3 — density: ⟨r₁²⟩_ΨJ = ⟨r₁²⟩_Φ0.71−2.742worse by 0.092
β free (diagnostic, grid edge)≥ 4−2.842does not even reach
Split-ζ alone (same integrator)−2.835reference
Exact (Pekeris)−2.9037

Reading

The three rules are physically distinct and all honest — and orthogonality nearly touches the truth. R2 (canceling the factor's leakage into the one-body sector) selects β = 2.53, close to the free edge, and loses only 0.013 Ha against split-ζ: it is the best, and its logic is that of the method. R3 (preserving density) imposes too soft a factor, R1 (orbital scale) too wide. The free diagnostic decides: energy wants the factor extinguished (β → ∞, J → 1). In this family, energy always prefers radial correlation alone.

The failure is structural, not parametric

P30 had localized the problem in range; P31 closes it: even free, range does not save the family. The reason is visible in the local identity: the gain of 1/r₁₂ at contact is paid by the kinetic term u′² and by the crossing u′ cos θ₁₂ (a+b); at one term, these three contributions cannot simultaneously be good — the angular flexibility of a Hylleraas expansion (Σ c_nlm s^n t^l u^m) is needed, whose coefficients are free parameters. The method refuses this price.

Consequence, published as bound not defeat: the calculation's map has an exact frontier — everything discrete, kinematic and one-body is derivable (24 successes); the correlated continuous two-body response is not without form freedom. The failures P22 (Q_ββ magnitudes), P23 (quenching), P24 (gaps), P26/P29 (skins, shell) and P30–P31 all converge on this now-measured wall.

Verdict

CriterionResult
Contact derivable (cusp selected by energy, P30)
Range not derivable✅ (published)
No single-term factor adds angular to radial✅ (published)
Three derived range rules tested
Bound not crossed❌ (documented as constitutive frontier)

Document chain

Document source

Reports (PDF verdict notes)

</> Simulation — Python scripts

p31_portee.py

⌗ Cross-references

§ Related glossary entries

Frontière r₁₂ — Limite constitutive de la machine : la réponse corrélée continue à deux corps (r₁₂ explicite) n'est pas dérivable sans liberté de forme. Déclarée en P31, loi en P32, renforcée en P33.
Hartree–Fock — Référence champ moyen à une particule que la portée dérivée doit battre pour déclarer un succès de corrélation.

See the full glossary →