Identification problems of muon and electron events in the Super-Kamiokande detector

TL;DR

This paper reevaluates the muon-electron event identification accuracy in the Super-Kamiokande detector, revealing significantly higher misidentification rates than previously estimated due to stochastic fluctuations.

Contribution

It introduces comprehensive Monte Carlo simulations that account for fluctuations, providing more accurate misidentification rates for muon and electron events.

Findings

Misidentification rate for sub-GeV electrons and muons is at least 20%.

Multi-GeV misidentification rate is several percent.

Most misidentified events originate from muons.

Abstract

In the measurement of atmospheric nu_e and nu_mu fluxes, the calculations of the Super Kamiokande group for the distinction between muon-like and electronlike events observed in the water Cerenkov detector have initially assumed a misidentification probability of less than 1 % and later 2 % for the sub-GeV range. In the multi-GeV range, they compared only the observed behaviors of ring patterns of muon and electron events, and claimed a 3 % mis-identification. However, the expressions and the calculation method do not include the fluctuation properties due to the stochastic nature of the processes which determine the expected number of photoelectrons (p.e.) produced by muons and electrons. Our full Monte Carlo (MC) simulations including the fluctuations of photoelectron production show that the total mis-identification rate for electrons and muons should be larger than or equal to 20 % for sub-GeV region. Even in the multi-GeV region we expect a mis-identification rate of several % based on our MC simulations taking into account the ring patterns. The mis-identified events are mostly of muonic origin.