Dear Geant4 Community,
We are writing to report on significant discrepancies observed in the neutron physics Modelling between Geant4 v9.4.p01 and v11.1.1 during in situ neutron calibration studies for the ANAIS-112 experiment.
Our simulations involved a direct comparison of results obtained with v9.4.p01 (utilizing G4NDL3.14) and v11.1.1 (based on G4NDL4.7). We detail below the differences encountered, specifically concerning elastic and inelastic scattering processes in iodine:
- The simulation using Geant4 v11.1.1 exhibits a prominent de-excitation gamma-ray line at 295 keV. This line corresponds to an excited state of 127I. According to evaluated nuclear data, this state is not expected to be populated via neutron inelastic scattering which is confirmed by our measurement. This line is not present in the v9.4.p01 simulation.
- Version v11.1.1 predicts a substantially lower rate for the measured inelastic peak of 57 keV compared to v9.4.p01.
- The newer v11.1.1 predicts a higher elastic scattering signal.
We find a much better agreement with data for version 9.4, despite v11.1.1 is a more recent release.. You can find further information in Section 4.3.1 of the thesis: https://arxiv.org/abs/2510.19469
We appreciate the continuous effort devoted to the maintenance and development of the Geant4 toolkit. Nevertheless, given the strong dependence of dark matter experiments such as ANAIS-112 on an accurate modelling of neutron interactions, we would be grateful if members of the Geant4 Collaboration could provide some insight into the origin of the discrepancies we observe between the G4NDL 3.14 and G4NDL 4.7 data libraries. In particular, we would like to know whether these discrepancies were identified or investigated internally during the transition between these G4NDL releases. Our results suggest that the observed differences may be especially relevant for iodine, while other isotopes might be better described by the newer libraries. In this context, any information on the underlying causes of these differences, as well as guidance on which version of the elastic-scattering data is considered the most reliable for specific isotopes, would be highly appreciated. 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 experimental spectral shape for this transition.
Thank you in advance and best regards,
Tamara Pardo on behalf of the ANAIS research team
CAPA, Universidad de Zaragoza