OSCILLATION INVESTIGATIONS IN SOUDAN 2: ATMOSPHERIC νμ → ντ AND ${\rm n} \to \bar{n}$ IN IRON

Abstract

The 5.90 fiducial kiloton-year exposure of the Soudan-2 iron tracking calorimeter is analyzed for effects of neutrino oscillations. Using fully contained single track, single shower and multiprong events, we determine the atmospheric νμ/νe ratio-of-ratios in the sub-GeV Eν regime to be R=0.68±0.12±0.06. Assuming this anomalously low R-value to be the result of νμ flavor disappearance via νμ → ντ oscillation, we distribute fully contained and partially contained events into bins of log (L/Eν) and fit to oscillation hypotheses using extended maximum likelihood. The region in the ( sin 22θ, Δm2) plane allowed at 90% CL is obtained using the Feldman-Cousins procedure: 0.52< sin 2 2θ≤1.0 and 3×10-4<Δm2<2×10-2 eV 2. A different oscillation process which could occur in our contained event samples is baryon instability in the form of neutron-antineutron oscillations. Our search requires candidate [Formula: see text] occurrences to have ≥ 4 prongs (tracks and showers) and to have kinematics compatible with [Formula: see text] annihilation within a nucleus. We observe five candidate events, with estimated atmospheric neutrino plus cosmic ray backgrounds totalling 4.5 ± 1.2 events. Previous experiments with smaller exposures observed no candidates, with estimated background rates similar to this experiment. We set a lifetime lower limit at 90% CL for the [Formula: see text] oscillation time in iron: TA(Fe)>7.2×1031 years. The corresponding lower limit for oscillation of free neutrons is [Formula: see text] seconds.