Two published 16N beta-delayed alpha spectra can be described with essentially the same strong continuum, but they do not fit a single common weak-current amplitude when analyzed together, according to a new modeling study. The result identifies a model-level tension, not proof that the physical weak amplitudes are intrinsically different.
The work is an arXiv preprint, version v1, dated 26 August 2026. It uses corrected spectra reported by Azuma and Tang, together with a shared set of elastic alpha-12C scattering measurements.
A shared nuclear backdrop, tested two ways
The analysis asked whether the Azuma and Tang measurements could be described with common strong and weak-current coefficients while allowing independent overall normalizations. In practice, that meant testing whether both spectra could fit one nuclear description even if their overall scales differed.
The calculation used cluster effective field theory, or cluster EFT. It linked the decay amplitudes to inverse dressed propagators constrained by elastic alpha-12C scattering. These propagators are the model's way of carrying information about the shared strong interaction into the decay calculation.
The corrected beta-delayed alpha data contained 91 points for Azuma and 93 for Tang. The shared elastic-scattering block contained 11,392 measurements, covering alpha-particle energies from 2.6 to 6.7 MeV at 32 laboratory angles between 24.0 and 165.9 degrees.
The beta-delayed alpha points were treated as independent Gaussian measurements with their quoted errors. Elastic residuals were also treated as Gaussian, using a diagonal covariance matrix, so the calculation did not represent correlations between different points.
The spectra improve when fitted separately
Each spectrum was fitted together with the elastic data in an independent analysis. Using a fit score called chi-squared per data point, the present model gave 1.732 for Azuma and 2.028 for Tang, compared with 4.06 and 3.56 in the earlier fixed-propagator comparison. The fits achieved those lower values without adding an extra effective 1-minus resonance or a smooth remainder.
Conditioning on either spectrum also retained the common elastic description without qualitative distortion. The elastic chi-squared per data point was 6.4194 for the Azuma-conditioned fit and 6.696 for the Tang-conditioned fit.
To search the parameter space, independent differential-evolution runs selected reproducible minima before affine-invariant Markov chain Monte Carlo sampling. The samples supplied parameter estimates, credible intervals and pointwise uncertainty bands.
One joint fit exposes the remaining mismatch
The difficulty appeared when both spectra were required to share one weak-current description. With the raw fitting objective, the chi-squared per point was 4.3194 for Azuma, 10.9011 for Tang and 6.3663 for the elastic data. A sector-balanced objective, designed to give the three data blocks more equal influence, improved the decay-sector numbers to 3.3082 and 8.9926, but the elastic value was 6.4662 and the combined result still fell short of the quality reached by the independent fits.
The authors interpret the deterioration in the joint fits as robust tension within the minimal common-current, response-free model. They stress that this result does not establish intrinsically different physical weak amplitudes.
The fitted Azuma-to-Tang normalization ratio changed little with the objective, from 3.7489 for the raw fit to 3.7425 for the sector-balanced fit. The shared strong-sector parameters were likewise stable. The common weak-current coefficients were more sensitive, including a sign change in one coefficient.
The analysis does not yet deliver a capture prediction
The study does not extract a new S_E1 value at 300 keV, the low-energy capture quantity considered in the analysis. Electromagnetic short-distance couplings were not fixed by elastic scattering or by the beta-delayed alpha data alone, so this is not a new radiative-capture result.
The reported intervals are conditional on the chosen parameter domains, fixed branching-ratio inputs, fixed nonshared elastic parameters, uncorrelated pointwise errors and objective weighting. They do not establish that the minimal model is adequate.
The simultaneous model also omitted experiment-specific energy calibration, resolution and response functions. That leaves open whether those effects explain the remaining Azuma-Tang line-shape difference, or whether a response-aware analysis would still require an extension of the weak amplitude.
A narrower conclusion for future fits
The practical conclusion is that the two spectra can share a strong continuum in separate descriptions, while weak-current parameters can depend strongly on how the data are balanced in a joint fit. The authors intend the work to guide a future unified analysis of elastic scattering, decay and radiative capture, with the relevant experimental and theory uncertainties carried through.
Paper data and sources
Original title: Joint cluster-EFT analysis of $^{16}$N $β$-delayed $α$ spectra and $α$-$^{12}$C scattering
Authors: Jubin Park, Myeong-Hwan Mun, Shung-Ichi Ando
Journal/Repository: arXiv
Status: Preprint, not yet peer-reviewed
First online: 2026-08-26
DOI: Not available
Original paper · Full text