High precision solar neutrino spectroscopy with Borexino and JUNO

This PhD thesis focuses on the study of solar neutrinos with Borexino and JUNO. The first simultaneous analysis of all solar neutrino fluxes has been performed using Borexino Phase-II data. For this purpose, an innovative multivariate analysis has been designed and implemented. A soft- ware framework is created to perform such an analysis. Thanks to parallel computing techniques and graphics processing units (GPU), this software framework reduces the computing time by a factor of 200 compared with the existing tools. New analytical models of the detector response have been developed and benchmarked against simulations. A comprehensive study of the sensitivity and systematic uncer- tainties has also been performed. The results obtained from this analysis assuming the MSW-LMA model are: • ν(pp): (6.1 ± 0.5 +0.3 −0.5) × 1010 [cm−2s−1] • ν(7Be): (4.99 ± 0.11 +0.06 −0.08) × 109 [cm−2s−1] • ν(pep): (1.27 ± 0.19 +0.08 −0.12) × 108[cm−2s−1] (Assuming HZ ν(CNO)), and (1.39 ± 0.19 +0.08 −0.13) × 108[cm−2s−1] (Assuming LZ ν(CNO)) • ν(CNO): < 7.9 × 108(95% C.L.) [cm−2s−1] The estimation of the fluxes has been combined with an independent analysis of the 8B solar neutrino flux based on Borexino data in order to perform phenomenological studies on the solar models and neutrino oscillation models. It is found that, assuming the MSW-LMA model, the Low Metallicity model (B16-LZ) is rejected with a confidence level of 96.6% (2.2 σ); it is also found that, assuming the High Metallicity model (B16-HZ), the vacuum oscillation model is rejected with a confidence level of 98.2% (2.4 σ). The sensitivity of Borexino to CNO solar neutrinos has been studied assuming a constraint on the its major background 210Bi. Different methods to measure the 210Bi decay rate with its daughter 210Po have been considered and characterized. In order to reduce the systematic uncertainties, we have created a special dataset whose events are reconstructed using only the PMTs that have been operating stably over the considered data-taking period. This dataset is proved to be be useful for various studies, including testing fundamental assumptions made in the method to measure the 210Bi decay rate and improving the light yield monitoring in the Monte Carlo simulation. In the context of the JUNO experiment, fundamental aspects of the analysis and simulation framework needed for studying the solar neutrinos have been developed, including the event generator used in simulation, the algorithms used to suppress event pile-up effect and contribution of dark noise of PMTs, and energy/vertex reconstruction softwares optimized for sub-MeV events. The signal and backgrounds rates have been evaluated, as well as the efficiencies of various event selection for detecting the 8B solar neutrinos with JUNO. It is demonstrated that JUNO can achieve around 2 to 1 signal-to-background ratio in the observed energy range from 2 to 3 MeV, making it possible to test new physics models. A comprehensive study of the sensitivities of JUNO to different solar neutrino components has also been performed. According to our estimation, the evaluated sensitivities are • ν(7Be): 7% statistical + systematic uncertainty (assuming 0.5% light yield accuracy) • ν(CNO): 13% statistical + systematic uncertainty (assuming 0.1% light yield accuracy).