Fast neutron leakage spectra for assemblies containing thorium, D{u2082}O and natural uranium

Integral measurements with a ³He spectrometer have been made to test the validity of the differential data of the ENOF/B-II files for thorium, natural uranium and heavy water. The experimental measurements of fast neutron spectra have been in two lattices of thorium metal and D₂O, a cylindrical array of close-packed natural uranium metal slugs in air, a D₂O sphere and cylinder, and a cylinder of thorium nitrate. A ²³⁵U fission disc, irradiated by a beam of thermal neutrons from the V.P.I. & S.U. reactor. provided a fast neutron source. Neutron leakage spectra were taken over the energy range between 100 keV and l.65 MeV. The spectrometer was calibrated using neutrons from the ³H(p,n)³He reaction in a Van de Graaff accelerator.

Experimental results were compared with leakage spectra calculated. with the multigroup transport code, ANISN. For the thorium metal and D₂O lattices and for the D₂O sphere and cylinder, agreement between calculated and measured spectra was attained confirming the validity of the ENDF/B-II cross-sections and pointing to the usefulness of the ³He spectrometer for fast neutron leakage spectra determinations.

For the natural uranium assembly, the calculated leakage spectrum between about 0.743 MeV and 1.65 MeV was significantly below the measured spectrum. This discrepancy supports a recent suggestion by H. Alter of Atomics International that the ENOF/B-II data files contain significant errors in the transport cross-sections for ²³⁸U and that the suspect data are in need of closer examination and revision.

As a further test of the applicability of the calculational procedures and ²³⁸U cross-sections, transport analysis using the ANISN code was applied to predict the central spectrum of the Japanese Atomic Research Institute (JAERI) I-5 fast assembly. These calculations were compared with the results of published ³He spectrometer measurements of the fast neutron spectrum for this 20% enriched uranium assembly. The results of the comparison reveal a measured spectrum slightly harder than the spectrum determined by the calculations using ENDF/BII cross-section data, and gives further evidence to support the conclusions regarding the suspect ²³⁸U data files.