Hierarchical Bayesian Thermonuclear Rate for the $^7$Be(n,p)$^7$Li Big Bang Nucleosynthesis Reaction

International audienceBig Bang nucleosynthesis provides the earliest probe of standard model physics, at a time when the universe was less than 1000 seconds old. It determines the abundances of the lightest nuclides, which give rise to the subsequent history of the visible matter in the universe. This work derives new 7Be(n,p)7Li thermonuclear reaction rates based on all available experimental information. This reaction sensitively impacts the primordial abundances of 7Be and 7Li during big bang nucleosynthesis. We critically evaluate all available data and disregard experimental results that are questionable. For the nuclear model, we adopt an incoherent sum of single-level, two-channel, R-matrix approximation expressions, which are implemented into a hierarchical Bayesian model, to analyze the remaining six data sets we deem most reliable. In the fitting of the data, we consistently model all known sources of uncertainty, including discrepant absolute normalizations of different data sets, and also take the variation of the neutron and proton channel radii into account, hence providing less biased estimates of the 7Be(n,p)7Li thermonuclear rates. From the resulting posteriors, we extract R-matrix parameters (E r , , ) and derive excitation energies and partial and total widths. Our fit is sensitive to the contributions of the first three levels above the neutron threshold. Reaction rates were computed by integrating 10,000 samples of the reduced cross section. Our 7Be(n,p)7Li thermonuclear rates have uncertainties between 1.5% and 2.0% at temperatures of ≤1 GK. We compare our rates to previous results and find that the 7Be(n,p)7Li rates most commonly used in big bang simulations have uncertainties that are too optimistic