Measurement of solar neutrino interaction rate below 3.49 MeV in Super-Kamiokande-IV

TL;DR

This paper advances low-energy solar neutrino measurements by integrating a boosted decision tree-based event selection with the wideband intelligent trigger in Super-Kamiokande-IV to suppress radioactive backgrounds below 3.49 MeV. By analyzing data from 2.99–3.49 MeV, it observes a solar neutrino signal with a data-to-unoscillated-MC ratio of $0.307^{+0.091}_{-0.090} \,( ext{stat}) \,\pm\,0.066 \,( ext{syst})$, corresponding to a flux of $(1.61^{+0.48}_{-0.47} \,\pm\,0.35) \times 10^{6} \,\text{cm}^{-2}\text{s}^{-1}$ and a $2.76\sigma$ significance. While this single low-energy bin aligns with expectations from the MSW framework, the large systematic and statistical uncertainties mean it does not decisively confirm an upturn in the $ u_e$ survival probability $P_{ee}$; nonetheless, the result demonstrates the potential of advanced trigger and ML-based selection to extend solar-neutrino measurements into the transition region. The measured low-energy data have a modest impact on the global $P_{ee}$ fits, with the strongest constraints continuing to come from higher-energy solar neutrino data.

Abstract

Super-Kamiokande has observed $^{8}\text{B}$ solar neutrino elastic scattering at recoil electron kinetic energies ($E_{kin}$) as low as 3.49 MeV to study neutrino flavor conversion within the sun. At SK-observable energies, these conversions are dominated by the Mikheyev-Smirnov-Wolfenstein effect. An upturn in the electron neutrino survival probability in which vacuum neutrino oscillations become dominant is predicted to occur at lower energies, but radioactive background increases exponentially with decreasing energy. New machine learning approaches provide substantial background reduction below 3.49 MeV such that statistical extraction of solar neutrino interactions becomes feasible. This article presents an analysis of the solar neutrino interaction rate at $E_{kin}$ < 3.49 MeV with the full SK-IV period, using data from a wideband intelligent trigger when available and with a boosted decision tree for event selection. A solar neutrino signal is observed between 2.99 MeV < $E_{kin}$ < 3.49 MeV with $2.76σ$ significance and a data to unoscillated MC ratio of $0.307^{+0.112}_{-0.111}$. This additional low energy data has a negligible effect on the $1σ$ intervals of the fits to the solar neutrino energy spectrum but has a noticeable effect on the best fit when using the exponential parameterization.

Figures (12)

• Figure 1: The time variation of the SLE trigger efficiency estimated by the MC simulation of $\mathrm{^{8}B}$ solar neutrinos shown as the average of 30-day intervals. The green and red plots represent the lowest two energy bins of skiv, and the 2.99-3.49MeV energy bin (blue) is discussed in this article. After changing the trigger threshold (orange dashed line) from 34 hits to 31 hits, trigger efficiency improved significantly. The start of data-taking with the separate WIT system (violet dashed line) is shown for reference.
• Figure 2: The trigger efficiency of the WIT system as a function of reconstructed electron kinetic energy based on solar $\mathrm{^{8}B}$ MC. The blue, orange, and red plots show the cumulative efficiency following the pre-trigger, the cut on the number of hits coincident with potential vertices, and the cut on the final vertex position, respectively.
• Figure 3: Distributions of the most important reconstructed variable inputs to the BDT in the WIT dataset for signal (blue) and background (orange) following preselection cuts. The numbers refer to the indices in Table \ref{['tab:vars']}. Histograms are normalized to an area of 1. Variable importance is determined by the XGBoost "Total gain" metric xgboost.
• Figure 4: BDT background rejection calculated as the inverse of the background acceptance rate as a function of signal efficiency for the WIT dataset in 2.49MeV - 3.49MeV. The performance of the existing cuts used in skiv except those described in Section \ref{['sec:final_sel']} is shown as a red $\times$.
• Figure 5: Solar angle distributions for the WIT dataset in all MSG bins for data selection (red), calculated background shape (black), and calculated signal shape added to background shape (blue) all with $1\sigma$ statistical error bands.