Statistics Reveal Hidden Coulomb Physics Inside the Light Meson Spectrum

For more than half a century, physicists have organized the bewildering zoo of meson resonances along elegant, straight lines known as Regge trajectories, in which a hadron’s squared mass grows linearly with both its orbital angular momentum and its radial excitation number. The picture, rooted in a semiclassical image of a rotating relativistic string with quarks at its endpoints, has survived decades of experimental scrutiny. Yet a nagging question has persisted: how exactly linear is this law when the data are examined with modern statistical rigor? A new open-access study in The European Physical Journal C by S. S. Afonin of Saint Petersburg State University and the NRC Kurchatov Institute now answers that question with unusual precision, and the answer carries a surprising twist rooted in relativistic quantum mechanics.

The stakes are higher than they might first appear. Light non-strange mesons, built from up and down valence quarks, are the cleanest laboratory available for studying the nonperturbative dynamics of quantum chromodynamics. Because the Higgs-generated current masses of the up and down quarks are nearly negligible, at roughly two to five MeV, almost all of the observable mass of these particles is generated by spontaneous chiral symmetry breaking and dynamic gluon dressing. Unlike heavy quarkonia, where large bare masses rigidify the quantum fluctuations, the light non-strange sector exposes confinement, dynamic mass generation, and short-range relativistic interactions in their purest form. Mapping the excitation spectrum of these mesons therefore offers direct insight into the structural dynamics of nucleons and atomic nuclei, which are built from the very same valence quarks.

Fortunately, the experimental dataset is unusually rich. Decades of high-energy hadronic production experiments have established long, continuous excitation bands of non-strange mesons extending up to total angular momentum J equal to six, providing the statistical leverage needed for genuine multi-parameter optimization. Afonin assembled a compilation of 85 light meson states, of which 27 are especially well-established benchmarks with robust quantum number assignments and minimal experimental errors. The pion was deliberately excluded, since its inclusion in Regge or string frameworks remains an open problem, and states with significant hidden strangeness were also set aside because nonlinearity is already appreciable there.

The analysis compared three competing models. Model 1 is the traditional linear trajectory with independent radial and orbital slopes. Model 2 enforces a universal slope, collapsing the spectrum into the simple form in which squared mass is proportional to the sum of the radial and orbital quantum numbers, a degeneracy reminiscent of the Coulomb problem in atomic physics. Model 3 keeps the universal slope but adds a nonlinear correction proportional to one over the quantity l plus one, motivated by the leading-order relativistic correction to the principal quantum number for a Dirac particle bound in a Coulomb field. This Dirac-Coulomb term is expected to matter most for S-wave states, where the centrifugal barrier vanishes and the quark-antiquark wave function probes very short distances dominated by one-gluon exchange rather than linear confinement.

The statistical machinery was as important as the physics. Because experimental uncertainties on meson masses vary by two orders of magnitude, a naive chi-squared fit would be dominated by a handful of states. Afonin instead introduced a universal theoretical model uncertainty, combined with the experimental errors through first-order error propagation from masses to squared masses, and determined that theoretical uncertainty self-consistently by enforcing the condition that the reduced chi-squared equals exactly one. The global optimization ran on a two-tier numerical routine combining the CERN MINUIT algorithm with SciPy root-finding solvers, and parameter errors were extracted from the Hessian curvature matrix of the log-likelihood function. Model selection then relied on the Akaike and Bayesian information criteria, which penalize unnecessary parameters, supplemented by Wald tests and a systematic scan of the likelihood across the full physical range of the theoretical error scale.

The results are striking. On the 27 benchmark states, both linear models fail badly, requiring intrinsic model uncertainties of 87 and 91 MeV respectively, and splitting the slopes barely helps, indicating that independent radial and orbital slopes are a statistical over-parameterization. Adding the Dirac-Coulomb term, however, cuts the required intrinsic error by more than half, down to 42 MeV, and the extracted correction of 0.410 plus or minus 0.097 GeV squared is significant at about 4.2 standard deviations. When the analysis is scaled to the full dataset of 85 states, the universal slope remains remarkably stable, shifting only from about 1.153 to 1.189 GeV squared, and the significance of the nonlinear term climbs to roughly 6.0 standard deviations, crossing the five-sigma discovery threshold conventionally demanded in high-energy physics. Model 3 wins decisively under both information criteria, with a Bayesian information advantage of 10.8, which on Jeffreys’ scale constitutes decisive evidence.

The most physically illuminating result comes from the truncation tests. When all S-wave states are removed, the significance of the Dirac-Coulomb term collapses, falling from 6 sigma to about 3 sigma on the global sample and to an insignificant 1.7 sigma on the benchmark subset, while the Bayesian criterion reverses its preference in favor of the simpler linear model. If the nonlinear term were a mere overfitting artifact, its significance would have survived the deletion of a single row of the spectrum. Instead, the distortions are demonstrably localized in the l equals zero sector, exactly where quantum mechanics predicts that a short-range relativistic interaction should dominate. For states with orbital angular momentum of one or higher, the centrifugal barrier pushes the wave function outward and restores the classical, linear string-like behavior. In other words, the light non-strange meson spectrum exhibits not only a Coulomb-like degeneracy but also a fine splitting analogous to the first relativistic correction in the Coulomb problem.

Robustness checks reinforce the conclusion. A parametric scan treating the theoretical uncertainty as a free variable between 0.01 and 0.10 GeV reveals that the chi-squared decays monotonically as the error scale grows, but the Akaike and Bayesian criteria trace a U-shaped profile whose minimum for Model 3 sits near 0.042 GeV, well within the typical phenomenological benchmark of 0.05 GeV. Throughout the entire scanned range, the nonlinear trajectory maintains a Bayesian advantage greater than ten units over the linear alternatives. A correlation analysis of the covariance matrix further shows that the slope and the Coulomb magnitude are only moderately correlated, at 0.647, proving that the nonlinear term models distinct low-angular-momentum physics rather than mimicking a global shift of the trajectory slope.

Beyond its theoretical elegance, the framework offers a practical tool for hunting exotic particles. One of the long-standing puzzles of hadron spectroscopy is whether the lightest scalar glueball, a bound state of pure gluons predicted by lattice QCD, hides among the scalar isoscalar resonances f0(1370), f0(1500) and f0(1710). The new analysis shows that the masses of these three candidates lie on the radial trajectories of ordinary quark-antiquark states: f0(1370) and f0(1710) fit neatly into the established clusters, while f0(1500) matches the expected position of a state with a dominant hidden-strangeness component between neighboring clusters. A glueball admixture is therefore not needed to explain their masses, a conclusion consistent with recent coupled-channel classifications of scalar isoscalar mesons.

The study demonstrates how modern statistical inference, from self-consistent error propagation to information criteria and likelihood scans, can extract crisp physical statements from a phenomenological spectrum long thought to be understood. The universal slope of the light non-strange Regge trajectories is real, the deviations from linearity are real too, but they belong to a specific and identifiable mechanism: short-range relativistic gluon exchange that distorts the S-wave states. With the methodology now validated, the natural next step is to extend the same statistical treatment to heavy-light systems and heavy quarkonia, where short-range interactions play an even more dominant role in shaping the bound-state structure, and to continue using trajectory statistics as an automated sieve for separating ordinary quark-antiquark resonances from anything more exotic.

Subject of Research: Statistical analysis of linear and nonlinear Regge trajectories in the light non-strange meson spectrum

Article Title: Advanced statistical analysis of linear and nonlinear Regge trajectories for light non-strange mesons

Article References: Afonin, S. S. (2026). Advanced statistical analysis of linear and nonlinear Regge trajectories for light non-strange mesons. The European Physical Journal C, 86(9), Article 1076. https://doi.org/10.1140/epjc/s10052-026-16378-5

Image Credits: AI Generated

DOI: 10.1140/epjc/s10052-026-16378-5

Keywords: Regge trajectories, light mesons, QCD, hadron spectroscopy, chi-squared minimization, Akaike information criterion, Bayesian information criterion, Dirac-Coulomb correction, glueball, quark-antiquark states, confinement, nonperturbative dynamics