Ab initio Standard-Model prediction of $\eta\to 3\pi$ decay and the correction $\Delta Q$ to the chiral low-energy theorem for the quark-mass ratio $Q$

Weinberg's 1975 paper "The U(1) problem" left four issues open: the dynamical origin of a proposed dipole (ghost-Goldstone) mechanism for the $\eta$; the domain of validity of quark-mass perturbation theory; the factor-of-three gap between the predicted $\eta\to\pi^+\pi^-\pi^0$ width ($\simeq 65$ eV) and experiment ($\simeq 200$ eV at the time); and the possible existence of nonperturbative anomaly effects. The first, second, and fourth issues are resolved or superseded: instantons break $U(1)_A$ nonperturbatively ('t Hooft 1976), removing the need for the dipole construction and establishing nonperturbative anomaly effects, later quantified by Witten and Veneziano; and chiral perturbation theory turned quark-mass expansions into a controlled power counting. The rate gap was closed phenomenologically by higher-order chiral corrections plus all-orders $\pi\pi$ final-state rescattering in dispersive analyses. What remains open is a first-principles Standard-Model prediction of the isospin-breaking $\eta\to 3\pi$ amplitudes (partial widths, neutral-to-charged ratio, Dalitz-plot parameters), with the normalization not anchored to chiral input or $\eta$-decay data, together with a decisive test of whether the correction $\Delta Q$ to the low-energy theorem relating the quark-mass ratio $Q$ to meson masses is second-order small, as chiral perturbation theory predicts, or numerically enhanced, as the $\sim 1.5\sigma$ offset between lattice and phenomenological determinations of $Q$ suggests.

Background

Quantum chromodynamics (QCD) with three light quark flavors has an approximate chiral symmetry $SU(3)_L\times SU(3)_R$ of the massless limit, spontaneously broken to flavor $SU(3)$; the corresponding pseudo-Goldstone bosons (massless particles arising from spontaneous symmetry breaking, here made slightly massive by the small $u$, $d$, $s$ quark masses) are the pions, kaons, and the octet partner of the $\eta$. The singlet axial symmetry $U(1)_A$ resists this treatment: if its current were conserved and spontaneously broken, a ninth nearly massless Goldstone boson would have to exist. In 1975 Weinberg analyzed this "U(1) problem" under the assumption that the singlet axial current is effectively conserved, and showed by current algebra — the pre-effective-field-theory technology of symmetry-current commutation relations — that the resulting Goldstone boson would obey $m_{\eta_L}^2 \le 3 m_\pi^2 \approx (240\ \mathrm{MeV})^2$, flagrantly contradicted by the observed $m_\eta \approx 548$ MeV.

To reconcile a conserved $U(1)_A$ current with the heavy $\eta$, Weinberg proposed a "dipole" alternative: $U(1)_A$ is spontaneously broken by two Goldstone fields of opposite metric signature whose couplings to gauge-invariant operators are equal, so that the unphysical pole cancels from physical matrix elements while the doublet carries the $\eta$ quantum numbers. The paper acknowledged the equal-coupling property as an assumption without dynamical derivation. Within this framework it also estimated the isospin-violating decay $\eta\to\pi^+\pi^-\pi^0$, finding $\Gamma\simeq 65$ eV — a factor of three below the then-current experimental value — and it flagged as uncontrolled both the treatment of quark masses as small perturbations and the possible role of nonperturbative anomaly effects beyond the perturbative triangle anomaly (the quantum breaking of a classical axial-current conservation law by loop diagrams). These are the four open items of the source paper.

The premise of the dipole construction failed within a year: 't Hooft showed that instantons — nonperturbative gauge-field configurations invisible in perturbation theory — break $U(1)_A$ nonperturbatively, so the singlet current is not conserved and no ninth Goldstone boson is required ('t Hooft 1976, both papers). Witten and Veneziano subsequently quantified this mechanism in the large-$N_c$ limit ($N_c$ is the number of quark colors), relating the heavy singlet masses to the topological susceptibility $\chi_t$ of pure Yang–Mills theory (the vacuum fluctuations of the gluonic topological charge density) through $m_{\eta'}^2 + m_\eta^2 - 2 m_K^2 \simeq 2 N_f \chi_t/F_\pi^2$; lattice simulations now reproduce the $\eta$–$\eta'$ spectrum from correlation functions of the gluonic topological charge density (Su et al. 2025). The quark-mass-expansion question was superseded rather than answered: chiral perturbation theory (χPT), the systematic low-energy effective field theory of QCD expanding simultaneously in momenta and quark masses (with $m_q = O(p^2)$ in the chiral counting), replaced ad hoc quark-mass perturbation theory with a controlled power counting.

The decay $\eta\to 3\pi$ violates isospin (the $SU(2)$ flavor symmetry interchanging $u$ and $d$), and — because electromagnetic contributions to its amplitude are suppressed by a classic current-algebra result (Sutherland 1966) — it is dominated by the quark-mass difference $m_d - m_u$. Its rate therefore measures the quark-mass double ratio $Q^2 \equiv \frac{m_s^2 - \hat m^2}{m_d^2 - m_u^2}, \qquad \hat m = \tfrac{1}{2}(m_u + m_d),$ numerically $Q \approx 22$. The leading current-algebra prediction for $\Gamma(\eta\to\pi^+\pi^-\pi^0)$ is $\approx 65$–$66$ eV; one-loop χPT corrections raise it to $\approx 160(50)$ eV (Gasser–Leutwyler 1985); two-loop corrections give the slowly converging chiral series $\{1, 1.72, 2.22\}\times$ leading order (Bijnens–Ghorbani 2007); and the observed $\sim 300$ eV is reached only when $\pi\pi$ final-state rescattering is resummed to all orders in Khuri–Treiman-type dispersive analyses (three-body dispersive frameworks whose subtraction constants are fixed by χPT and $\pi\pi$-scattering data), the resummation being dominated by the broad scalar $\pi\pi$ resonance near 700 MeV (historically called $\epsilon(700\ \mathrm{MeV})$, now $f_0(500)$) — an effect Weinberg's 1975 paper itself listed among the candidate improvements. The most complete such analysis (Colangelo–Lanz–Leutwyler–Passemar 2018) fixes all free parameters, reproduces the measured Dalitz distributions (the decay distribution in the two-body invariant masses), predicts the neutral-to-charged width ratio $B = 1.44(4)$ (Particle Data Group, the particle-physics data compilation: $1.426(26)$ and $1.48(5)$), and resolves the sign puzzle of the neutral Dalitz slope ($\alpha = -0.0303(12)$ versus the measured $-0.0318(15)$). The 1975 rate gap is thus understood phenomenologically.

Two gaps remain, and they are linked by the normalization of the amplitude. First, the dispersive prediction is not a first-principles one: its overall normalization is anchored to next-to-leading-order χPT (the normalization constant $H_0 = 1.176(53)$), not derived from the Standard Model's parameters, and lattice QCD — in principle an ab initio evaluation of QCD by simulation on a discretized space-time grid — "is not yet able to describe the complex decay dynamics of $\eta$ and $\eta'$ mesons comprehensively" (Gan et al. 2022). Second, the chiral low-energy theorem (a relation exact up to higher chiral orders) $Q^2(1 + \Delta Q) = \frac{\hat M_K^2\,(\hat M_K^2 - \hat M_{\pi^+}^2)}{\hat M_{\pi^+}^2\,(\hat M_{K^0}^2 - \hat M_{K^+}^2)},$ where hatted meson masses are evaluated in QCD with the electromagnetic coupling turned off and $\hat M_K$ is the isospin-averaged kaon mass, predicts the correction $\Delta Q$ to be of second order in the chiral counting (of order $M^2$ in the light-quark-mass scale $M$) — negligible relative to the first-order corrections $\Delta S$, $\Delta R$, each of order 5% (strictly, the reparametrization-invariant variant $\Delta\tilde Q = O(M^2)$ vanishes through next-to-leading order). Yet lattice values of $Q$ average higher than the $\eta\to 3\pi$ determination by about $1.5\sigma$ (FLAG 2024, the Flavour Lattice Averaging Group review: $Q = 22.5(5)$ for $N_f = 2+1+1$ and $23.3(5)$ for $N_f = 2+1$, versus the phenomenological $22.1(0.7)$), a pattern that would indicate an anomalously enhanced $\Delta Q$; the direct evaluation gives $\Delta Q = -0.011(13)(43)$ (Leutwyler 2024), consistent with zero but too imprecise to decide. Whether $\Delta Q$ is second-order small or numerically enhanced — and whether the Standard Model alone predicts the $\eta\to 3\pi$ amplitude from its parameters — remains open.

Problem Statement

Predict the isospin-violating decay $\eta\to 3\pi$ from first principles in the Standard Model — with the overall normalization of the amplitude determined by the theory's parameters (quark-mass ratios, strong coupling, electromagnetic coupling) rather than anchored to chiral low-energy-theorem input or fitted to $\eta$-decay data — and thereby settle the status of the chiral low-energy theorem for that normalization. The two deliverables are parts of one objective, a first-principles Standard-Model account of the isospin breaking measured by this decay, since the same low-energy theorem fixes the amplitude normalization that deliverable 1 must reproduce independently.

Concretely, determine from the underlying QCD + QED dynamics at physical quark masses, with isospin breaking, electromagnetic effects, and the three-body final state treated with controlled uncertainties:

1. the decay amplitudes $T(s,t,u)$ for $\eta\to\pi^+\pi^-\pi^0$ and $\eta\to 3\pi^0$ as functions of the two-body invariant masses $(s,t,u)$, and from them the partial widths $\Gamma(\eta\to\pi^+\pi^-\pi^0)$ and $\Gamma(\eta\to 3\pi^0)$, the neutral-to-charged ratio $B = \Gamma(\eta\to 3\pi^0)/\Gamma(\eta\to\pi^+\pi^-\pi^0)$, and the leading Dalitz-plot slope parameters, each with a quantified statistical and systematic uncertainty budget; and

2. the implied value of the quark-mass ratio $Q$ — equivalently the correction $\Delta Q$ in the low-energy theorem $Q^2(1+\Delta Q) = \hat M_K^2(\hat M_K^2 - \hat M_{\pi^+}^2)/[\hat M_{\pi^+}^2(\hat M_{K^0}^2 - \hat M_{K^+}^2)]$, with hatted meson masses evaluated in QCD with the electromagnetic coupling turned off and $\hat M_K$ the isospin-averaged kaon mass — deciding, at a stated precision, whether $\Delta Q$ is second-order small in the chiral power counting as chiral perturbation theory predicts (negligible at the achieved precision) or numerically enhanced at the $\mathcal{O}(5\%)$ level of the first-order corrections $\Delta S$, $\Delta R$, as the persistent $\sim 1.5\sigma$ offset between lattice and $\eta\to 3\pi$ determinations of $Q$ would indicate.

A determination of $Q$ from a different observable, such as the $\eta'\to\eta\pi\pi$ Dalitz distributions, is not a first-principles $\eta\to 3\pi$ normalization.

References

1. Steven Weinberg, "The U(1) problem", Physical Review D 11, 3583–3593 (1975). DOI: 10.1103/PhysRevD.11.3583, https://doi.org/10.1103/PhysRevD.11.3583
2. G. 't Hooft, "Symmetry Breaking Through Bell–Jackiw Anomalies", Physical Review Letters 37, 8–11 (1976). DOI: 10.1103/PhysRevLett.37.8, https://doi.org/10.1103/PhysRevLett.37.8
3. G. 't Hooft, "Computation of the Quantum Effects Due to a Four-Dimensional Pseudoparticle", Physical Review D 14, 3432–3450 (1976). DOI: 10.1103/PhysRevD.14.3432, https://doi.org/10.1103/PhysRevD.14.3432
4. Edward Witten, "Current Algebra Theorems for the U(1) Goldstone Boson", Nuclear Physics B 156, 269–283 (1979). DOI: 10.1016/0550-3213(79)90031-2, https://doi.org/10.1016/0550-3213(79)90031-2
5. G. Veneziano, "U(1) Without Instantons", Nuclear Physics B 159, 213–224 (1979). DOI: 10.1016/0550-3213(79)90332-8, https://doi.org/10.1016/0550-3213(79)90332-8
6. J. Gasser and H. Leutwyler, "η → 3π to One Loop", Nuclear Physics B 250, 539–560 (1985). DOI: 10.1016/0550-3213(85)90494-8, https://doi.org/10.1016/0550-3213(85)90494-8
7. Johan Bijnens and Karim Ghorbani, "η → 3π at Two Loops in Chiral Perturbation Theory", Journal of High Energy Physics 0711, 030 (2007). DOI: 10.1088/1126-6708/2007/11/030, arXiv:0709.0230, https://arxiv.org/abs/0709.0230
8. M. T. Hansen, F. Romero-López and S. R. Sharpe, "Decay amplitudes to three hadrons from finite-volume matrix elements", Journal of High Energy Physics 04, 113 (2021). DOI: 10.1007/JHEP04(2021)113, arXiv:2101.10246, https://arxiv.org/abs/2101.10246
9. A. V. Anisovich and H. Leutwyler, "Dispersive analysis of the decay η → 3π", Physics Letters B 375, 335–342 (1996). DOI: 10.1016/0370-2693(96)00192-X, arXiv:hep-ph/9601237, https://arxiv.org/abs/hep-ph/9601237
10. Gilberto Colangelo, Stefan Lanz, Heinrich Leutwyler and Emilie Passemar, "Dispersive analysis of η → 3π", European Physical Journal C 78, 947 (2018). DOI: 10.1140/epjc/s10052-018-6377-9, arXiv:1807.11937, https://arxiv.org/abs/1807.11937
11. L. Gan, B. Kubis, E. Passemar and S. Tulin, "Precision tests of fundamental physics with η and η′ mesons", Physics Reports 945, 1–105 (2022). DOI: 10.1016/j.physrep.2021.11.001, arXiv:2007.00664, https://arxiv.org/abs/2007.00664
12. H. Leutwyler, "Testing χPT with the masses of the Nambu–Goldstone bosons", International Journal of Modern Physics A 39, 2441006 (2024). DOI: 10.1142/S0217751X24410069, arXiv:2306.12227, https://arxiv.org/abs/2306.12227
13. Y. Aoki et al. [Flavour Lattice Averaging Group (FLAG)], "FLAG review 2024", Physical Review D 113, 014508 (2026). DOI: 10.1103/nfzp-p5dn, arXiv:2411.04268, https://doi.org/10.1103/nfzp-p5dn
14. N. Yamanaka, "Sketch of the resolution of the axial U(1) problem without chiral anomaly", arXiv:2411.02792 (2024), https://arxiv.org/abs/2411.02792
15. Yue Su, Nan Wang, Long-Cheng Gui, Jun Hua, Jian Liang and Jun Shi, "η and η′ mesons from N_f = 2+1 lattice QCD at the physical point using topological charge operators", Physical Review D 112, 114516 (2025). DOI: 10.1103/cmzr-h8kx, arXiv:2505.19775, https://arxiv.org/abs/2505.19775
16. Adolfo Guevara, Feng-Kun Guo and Hao-Jie Jing, "Novel method for determining the light quark mass ratio using η′→ηππ decays", Physical Review D 114, 014021 (2026). DOI: 10.1103/j982-c64r, arXiv:2502.02837, https://arxiv.org/abs/2502.02837
17. A. Browman, J. DeWire, B. Gittleman, K. M. Hanson, E. Loh and R. Lewis, "Radiative Width of the η Meson", Physical Review Letters 32, 1067–1070 (1974). DOI: 10.1103/PhysRevLett.32.1067, https://doi.org/10.1103/PhysRevLett.32.1067
18. V. Chaloupka, C. Bricman, A. Barbaro-Galtieri, D. M. Chew, R. L. Kelly, T. A. Lasinski, A. Rittenberg, A. H. Rosenfeld, T. G. Trippe, F. Uchiyama, N. Barash-Schmidt, P. Soding and M. Roos [Particle Data Group], "Review of Particle Properties", Physics Letters B 50, 1–198 (1974). https://inspirehep.net/literature/91816
19. D. G. Sutherland, "Current algebra and the decay η → 3π", Physics Letters 23, 384–385 (1966). DOI: 10.1016/0031-9163(66)90477-X, https://doi.org/10.1016/0031-9163(66)90477-X
20. D. Diakonov and M. I. Eides, "On the ghost pole solution of the U(1) problem", LENINGRAD-81-639 preprint (1981). INSPIRE: https://inspirehep.net/literature/165492