Investigation of evaluated nuclear data in the prediction of inherent neutron sources

TL;DR

The paper addresses accurate prediction of inherent neutron sources from spontaneous fission and $(\alpha,n)$ reactions in MOX and UO$_2$ fuels. It critically evaluates the SOURCES-4C data and modelling by comparing alpha-lines, stopping powers, and neutron spectra against modern evaluated data libraries (e.g., JENDL-5, ENDF/B-VIII.1, JEFF-3.3) and uses PHITS-based straggling assessments, complemented by an OpenMC implementation of spontaneous fission. Key findings include underestimation of $^{235}$U alpha lines in SOURCES, generally reliable stopping-power data with caveats for MOX-fuel Pu, and a simplistic $(\alpha,n)$ emission-spectrum model in SOURCES that breaks down at higher $E_{\alpha}$ where continuum channels dominate; updated spectra from modern libraries are recommended. The work advocates incorporating detailed spontaneous-fission spectra into evaluations and Monte Carlo tools, demonstrated by the new OpenMC SF implementation, to improve accuracy in safety analyses and fuel-design studies.

Abstract

Quantifying inherent neutron sources in matter, particularly $(α, n)$ reactions and spontaneous fission, is important in nuclear engineering and other fields. The SOURCES code is a common tool for calculating the yield and spectrum of such neutrons. This paper critically examines all modelling assumptions and nuclear data in SOURCES and proposes alternative approaches where applicable. For $(α, n)$ reactions, we show that the alpha emission lines for $^{235}\mathrm{U}$ should be updated. Furthermore, we compare four different stopping power data sets for alpha particles slowing down and propose measurements to constrain mixed oxide nuclear fuel data. We use the computer code PHITS to show that energy and angular straggling during the slowing down of alpha particles in the material of interest is unimportant. Then, we compare the cross section and emission spectrum of $(α, n)$ reactions in SOURCES to recently evaluated data libraries. Importantly, the modelling of SOURCES for the emission spectrum seems too simple and may need to be updated. Finally, we compare data on spontaneous fission and show that while the neutron yield from SOURCES is reliable, some discrepancy is found with the neutron spectrum of evaluated data libraries. Complementing this work is an implementation of spontaneous fission in the Monte Carlo code OpenMC.

Figures (8)

• Figure 1: Comparison of stopping power in O, Pb, U and Pu from four different datasets: SOURCES libraries, SRIM, ASTAR and ATIMA. The stopping power in these databases has negligible dependence on the isotope composition. It depends on the atomic density.
• Figure 2: Scatter plot of the relative mean error and relative standard deviation in stopping power between SOURCES and the compared data sets. For each element, only the largest relative mean error between SOURCES and the other three stopping power datasets is shown.
• Figure 3: Distribution of alpha particles that traverse a material of thickness $0.5\times r(E_0)$ and which have initial energy 0.5 MeV. The left panel shows distribution in energy, while the right panel shows distribution in position transverse to the beam axis. The different curves have varying values of the step size parameter deltc.
• Figure 4: Relative standard deviation in energy loss (left panel) and transverse position of alpha particles (right panel) traversing matter. The standard deviation is plotted as a function of the initial alpha particle energy, for varying thickness of the traversed material, $\delta r$ with $r$ the particles' range.
• Figure 5: The cross section of $(\alpha,n)$ reactions as a function of the alpha particle energy. The comparison is performed between Sources4C data sets, JENDL-5 libraries and TENDL-2019 libraries; results are shown for four nuclides.