Dear Geant4 developers and users,
In the context of the ANAIS-112 experiment, we have performed detailed comparison between neutron calibration data and Geant4 simulations using 252Cf fission sources in NaI(Tl) detectors. In the study, we have investigated the production of 128I through neutron capture on 127I and the subsequent decay signatures observed in the detectors which are easily identifiable over the backgrounds.
Our analysis indicates a systematic excess in the simulated contribution from 128I, evident both around the 31 keV region (corresponding to the Te K-shell de-excitation following electron capture decay) and above a few hundreds keV (corresponding to 200-1600 keV). These discrepancies are observed consistently across the nine ANAIS-112 detectors. The observed behavior suggests that the neutron capture cross section for (127I(n,gamma)128I) implemented in Geant4 is overestimated. You can find further information in Section 4.3.1.2 of https://arxiv.org/abs/2510.19469
To quantify the overestimation in the production of 128I, we performed a simultaneous fit to the ANAIS-112 single-hit spectra across all nine detectors, using the spectral shape of the 128I decay. A single normalization parameter was allowed to float to match the simulated and experimental spectra from 200 to 1600 keV. This procedure yielded the following overestimation factors for the capture cross section for the two different Geant versions studied:
- Geant4 v9.4.p01 → (37.0 ± 0.3)% overestimation
- Geant4 v11.1.1 → (20.9 ± 0.4)% overestimation
After applying these correction factors, the agreement between data and simulation significantly improved in both the high-energy continuum (dominated by the beta spectrum and where the fit was conducted) and the 31 keV peak. Recent experimental measurements of the (127I(n,gamma)128I) cross section (e.g., Gandhi et al., 2021) also points to an overestimation of this cross section, but only information at three energies is available. Notice that our results correspond to the convolution of the Geant4 cross-section with the neutron spectra of the 252Cf source after partial thermalization in the ANAIS-112 lead shielding, because the source was placed outside.
In summary, both Geant4 versions overestimate the effective neutron capture rate producing 128I in NaI(Tl). The discrepancy highlights the need for a review or update of the (127I(n,gamma)128I) cross section in Geant4’s neutron data libraries.
We would appreciate any insights on the origin of these cross-section differences between Geant4 versions and whether updates or validations for iodine neutron capture data are currently planned. From our side, we will continue our measurements planning using the Cf-source in different shielding configurations trying to further pinpoint the identifed overestimation. Finally, should it be of interest to the collaboration, we would be willing to share relevant data or quantitative studies to help further investigate the observed disagreement and to inform the choice of an appropriate correction.
Thank you in advance and best regards,
Tamara Pardo on behalf of the ANAIS research team
CAPA, Universidad de Zaragoza