Integral experiment on bismuth slabs with D–T neutron irradiation for data validation

Abstract: Accurate nuclear data for neutron interactions with bismuth are crucial for applications in nuclear technology and radiation protection. This paper presents a comprehensive benchmark analysis that compares experimental data obtained at multiple angles and thicknesses with simulations based on four nuclear-data libraries: CENDL 3.2, ENDF/B-VIII.0, JENDL-5, and JEFF-3.3. The experiments involved neutron-leakage measurements of bis muth at three thicknesses and six angles, along with standard sample validations. Pulse time distributions were reconstructed using the maximum likelihood expectation maximization algorithm, and a silicon-carbide detector was employed to accurately distinguish between deuterium–tritium and deuterium–deuterium reaction products. Simulation models validated using polyethylene-sample results demonstrated calculated-to-experimental (C/E) values of 1 ± 0.03, thus confirming their reliability. The analysis revealed that CENDL-3.2 exhibited the best overall agreement in the elastic scattering region. In the discrete inelastic scattering region, JENDL-5 performed best at larger angles, while JEFF-3.3 was more accurate at smaller angles. In the continuous inelastic scatter ing region, JEFF-3.3 demonstrated the best overall performance, with CENDL-3.2 achieving good agreement at selected angles. In the (n,2n) reaction region, ENDF/B-VIII.0 provided C/E values closest to unity, while JENDL-5 ensured better consistency across the full energy spectrum. These findings highlight the importance of selecting appropriate nuclear-data libraries and emphasize the necessity for ongoing data refinement to im prove modeling accuracy.