Study on the Accuracy of Several Beryllium Evaluations and Comparison of Measured and Calculated Data on Reaction Rates and Tritium Production Distributions

The accuracy of beryllium data, particularly the ⁹Be(n,2n) cross-section and the associated secondary neutrons energy distribution has direct impact on the prediction accuracy of important neutronics parameters such as in-blanket neutron spectra, tritium production and other reaction rates distributions. The measured values for these parameters obtained within the U.S./JAERI Collaborative Program on Fusion Integral Experiments were compared to the calculated ones with three evaluations for beryllium data. The ENDF/B-V, LANL, and ENDF/B-VI evaluations for beryllium (denoted Be5, (LANL)Be, and Be6, respectively) were processed, along with ENDF/B-V data of other materials, into 80-group neutron structure of the MATSX6 library. The processed library was used to calculate tritium production rate, neutron spectrum, and several foil activation reaction rates along the central axis of a Li₂O test assembly in which a 5 cm-thick beryllium zone precedes the Li₂O zone and acts as a neutron multiplier. The test assembly itself is located at one end of a Li₂CO₃ enclosure where the D-T neutron point source is located inside the inner cavity. Examination of the Be data indicates that the total Be(n,2n) cross-section of B6 is lower than that of B5 and (LANL)Be by ∼ 10–15% at high incident energies and that this cross-section is very similar in the case of B5 and (LANL)Be evaluation. However, the secondary energy distribution (SED) from this reaction is noticeably different among the three evaluations; the SED in Be6 evaluation does not extend down below ∼ IKeV while the (LANL)Be evaluation extends down below ∼ 1 KeV, but the cross-section in this energy range is negligible. On the other hand, in the Be5 evaluation, the SED extends down to very low energies. The impact of these differences on tritium production from ⁶Li(T₆) and ⁷Li(T₇) is that the calculated-to-measured values (C/E) for T₆ have improved with the Be6 by ∼ 3–5% (C/E = 1.05) as compared to those obtained by the Be5 and (LANL)Be evaluations in the bulk of the Li₂O zone. The C/E values for the high threshold reactions [e.g., ⁵⁸Ni(n,2n)] have also improved by 3–7%. By comparing the calculated spectra to the measured ones at various locations, it was noticed that the 14 MeV peak is better predicted by the Be6 evaluation.