Improved Predictions of Reactor Antineutrino Spectra

TL;DR

This work tackles the challenge of predicting reactor $ar{ u}_e$ spectra with high precision by combining two complementary strategies: an ab initio summation of beta-branches from fission-product decays and a mixed approach that leverages precise ILL electron spectra while calibrating the remaining contributions with nuclear-database information. The authors deliver a comprehensive ab initio calculation using ENSDF, pandemonium-corrected data, and JENDL, achieving predictions at the ~10% level and providing a detailed error budget, including off-equilibrium effects. They then refine the neutrino conversion by anchoring to ILL electron data and implementing branch-by-branch finite-size and weak-magnetism corrections, leading to new reference antineutrino spectra for $^{235}$U, $^{239}$Pu, and $^{241}$Pu with a consistent ~3% normalization shift. The results have practical impact on reactor neutrino analyses, suggesting a modest increase in detected flux and highlighting the need for improved nuclear data and reactor-specific off-equilibrium corrections to meet the precision goals of upcoming experiments.

Abstract

We report new calculations of reactor antineutrino spectra including the latest information from nuclear databases and a detailed error budget. The first part of this work is the so-called ab initio approach where the total antineutrino spectrum is built from the sum of all beta-branches of all fission products predicted by an evolution code. Systematic effects and missing information in nuclear databases lead to final relative uncertainties in the 10 to 20% range. A prediction of the antineutrino spectrum associated with the fission of 238U is given based on this ab initio method. For the dominant isotopes 235U and 239Pu, we developed a more accurate approach combining information from nuclear databases and reference electron spectra associated with the fission of 235U, 239Pu and 241Pu, measured at ILL in the 80's. We show how the anchor point of the measured total beta-spectra can be used to suppress the uncertainty in nuclear databases while taking advantage of all the information they contain. We provide new reference antineutrino spectra for 235U, 239Pu and 241Pu isotopes in the 2-8 MeV range. While the shapes of the spectra and their uncertainties are comparable to that of the previous analysis of the ILL data, the normalization is shifted by about +3% on average. In the perspective of the re-analysis of past experiments and direct use of these results by upcoming oscillation experiments, we discuss the various sources of errors and their correlations as well as the corrections induced by off equilibrium effects.

Figures (13)

• Figure 1: (Color online) Illustration of the detected antineutrino spectrum in the case of $^{235}$U fissions (blue solid curve). Units are arbitrary and oscillation effects are suppressed. The detected rate rises from the threshold value at about 1.8 MeV, reaches a maximum around 4 MeV and vanishes after 8 MeV. This shape is the result of folding the emitted spectrum (black dashed-dotted curve), parameterization taken from section \ref{['subsec:Param']} and beta-inverse cross section (red dashed curve).
• Figure 2: Comparison of $^{235}$U and $^{239}$Pu reference electron spectra from SchreckU5Pu9 with our predictions based on the ab initio approach. The predictions have no free parameters and the rates are normalized to one fission.
• Figure 3: Residues of the $^{235}$U electron spectra computed as the difference of our ab initio calculations minus reference data from SchreckU5Pu9 divided by reference data. Dashed-dotted curve: ENSDF data only; dashed curve: some ENSDF data replaced by pandemonium corrected data; solid curve: unmeasured $\beta$ emitters are added on top of previous curve, using the gross-theory calculations of the JENDL nuclear database and few remaining exotic nuclei described by our model (see text).
• Figure 4: ab initio calculation of the electron (dashed histogram) and antineutrino (solid histogram) spectra of $^{238}$U for a 450 days irradiation time. For comparison the predictions of Vogel81 for an infinite irradiation time are plotted as crosses for the electron spectrum and stars for the antineutrino spectrum.
• Figure 5: (Color online) The blue hatched area shows the contribution of our ab inito prediction (ENSDF + pandemonium corrected nuclei) relative to the ILL reference data. The missing contribution coming from unknown nuclei and remaining systematic effects of nuclear databases (red hatched area) is fitted using a set of 5 effective $\beta$-branches.