R. Becker-Szendy q, R.M. Bionta °, C.B . Bratton h, D . Casper s, R. Claus q, B. Cortez ', S.T. Dye d , S . Errede m, G.W. Foster n, W. Gajewski a, K. Ganezer t , M. Goldhaber c, T.J . Haines k, P.G . Halverson a, E. Hazen d, T.W. Jones f, D. Kielczewska a,g, W.R. Kropp a, J.G. Learned e, J.M. LoSecco ', S . Matsuno e, J . Matthews b, G. McGrath e, C. McGrew a, R. Miller ', M.S . Mudan f, H.S. Park °, L . Price a, F. Reines a, J . Schultz a, S . Seidel D. Sinclair b, H.W. Sobel a, J.L . Stone d , L.R. Sulak d , R. Svoboda i, G . Thornton `, J.C. Van der Velde b and C. Wuest °

The IMB-3 experiment was a large water Cherenkov ring imaging detector with a fiducial mass of 3.3 kton. During a 7.6-kton-year exposure (similar to 4.6 x 10(33) nucleon yr) 935 contained events were observed. The observed rate and characteristics are consistent with the expected backgrounds from atmospheric neutrinos. Lower limits on the nucleon lifetime are set for a wide variety of proposed decay modes. [S0556-2821(98)00323-3].

A 2.1 ktonyr exposure of data from the Irvine-Michigan-Brookhaven detector has yielded 72 atmospheric neutrino events with a vertex contained inside the fiducial volume and at least 0.95 GeV of visible Cerenkov energy. The ratio of these two ratios (muonlike/total)(Data)/(muonlike/total)(MC) was found to be 1.1(-0.12)(+0.07)(stat) +/- 0.11(syst). The zenith angle dependence of this ratio of ratios is consistent with being flat. The region of sin(2)(2 theta) > 0.5 and delta m(2) > 9.8 x 10(-3) eV(2) has been excluded to the 90% confidence level for nu(mu) --> nu(e) oscillations while the region of sin(2)(2 theta) > 0.7 and delta m(2) > 1.5 X 10(-2) eV(2) has been excluded to the 90% confidence level for nu(mu) --> nu tau oscillations.

Received 5 August 1997DOI:https://doi.org/10.1103/PhysRevLett.79.2754.3©1997 American Physical Society

The measured fraction of muon-like, single-track, atmospheric neutrino events over a 7.7 kton-yr exposure of IMB is 0.36+/-0.02(stat)+/-0.02(syst) as compared to an expectation of 0.51+/-0.01(stat)+/-0.05(syst). No significant dependence of this fraction on zenith angle or momentum is seen. In addition, upward-going muons from higher energy atmospheric neutrino interactions are used to search for nu(mu) oscillations into nu(tau) by comparing the measured rate with the expected rate from a conservative flux calculation. The ratio of upward-going muons which stop in the detector to those which exit is also used to search for deviations from the expected energy spectrum. No evidence for oscillations is found in either analysis. Finally, a search is made for an astrophysical component to the detected neutrino flux from both energetic point sources and gamma-ray bursts and also for a possible seasonal variation of the high energy neutrino flux due to atmospheric density changes. No evidence for any of these effects is found.

The TMB detector (named after its founding institutions: University of California, Irvine, the University of Michigan and Brookhaven National Laboratory) collected data on a wide range of phenomena for over eight years. It was the first and the largest of the ring imaging water Cherenkov detectors. The detector consisted of 8000 metric tons of ultra-pure water instrumented with 2048 photomultiplier tubes (PMTs). The PMTs were placed on the roof, floor, and walls of the detector in a lattice of approximately 1 m spacing. It made measurements of contained events that ranged in energy from 15 MeV up to 1.5 GeV. This paper describes the calibration of the IMB detector. This procedure was accurate and stable over a wide range of physical variables. It was used with little change throughout the entire eight-year lifetime of the experiment. The IMB calibration is a model for future large-scale detectors that employ the water Cherenkov technique.

A test has been conducted at KEK, Japan using beams of electrons and muons in a 1 kiloton water Cherenkov detector instrumented with IMB3 phototubes and electronics to evaluate IMB`s algorithms for identifying electrons and muons. This identification is important because the IMB3 detector`s results on the atmospheric neutrino anomaly depend on the proper identification of the electrons and muons produced in neutrino charged-current interactions. Preliminary results are presented.

A temporal correlation analysis between moderate- (60 Mev less than or equal to E(nu) less than or equal to 2500 MeV) and high-energy (E(nu), greater than or equal to 2000 MeV) neutrino interactions within the IMB-3 detector and a catalog of gamma-ray bursts (GRBs) has been performed. The neutrino interactions consist of two types: the moderate-energy interactions that are contained within the volume of IMB-3 and the upward-going muons produced by high-energy nu(mu) interactions in the rock around the detector. No evidence is found for moderate- or high-energy neutrino emission from GRBs nor for any neutrino/neutrino correlation. The nonobservation of nu/GRB correlations allows upper limits to be placed on the neutrino flux associated with GRBs.

The temporal structure of low-energy (20-60 MeV) neutrino interactions within the IMB-3 detector during 863 days of livetime between 1986 May and 1991 March has been analyzed. The neutrino data are consistent with expected cosmic-ray-induced neutrino interactions with no bursts evident (excluding SN 1987A). When combined with the 327 effective livedays of IMB-1 data, we place a 90% C.L. upper limit of less-than-or-equal-to 0.71 Galactic supernovae yr-1. We have also performed a temporal correlation analysis with gamma-ray bursts (GRBs) using a subset of the low-energy neutrino data. No significant correlations were observed for coincidence windows of 1 minute, 1 hour, or 1 day.

If Gamma Ray Bursts (GRBs) are associated with a stellar collapse‐like phenomenon then it is resonable to expect neutrino production to occur at the source. We have performed a temporal correlation analysis with GRBs using the IMB low‐energy neutrino dataset during 809 days of livetime between 1986 and 1990. No correlations were observed placing a 90%. C.L. limit of 0.046 ν interactions per GRB. The dependence of the GRB distances to neutrino yield using volume and shell distribution models is discussed. Lower limits are derived which exclude galactic stellar collapse‐like models.

An improved limit on the flux of magnetic monopoles in the vicinity of the solar system is obtained, assuming that monopoles strongly catalyze nucleon decay (the Rubakov-Callan effect). Flux limits are presented for monopole velocities from 10(-5)c to 10(-1)c and for monopole-nucleon cross sections between 10(-27) cm2 and 10(-21) CM2. For a representative velocity beta almost-equal-to 10(-3) , and cross section sigma almost-equal-to 10(-24) CM2, We obtain a limit F(m) < 2.7 x 10(-15) Cm-2 sr-1 sec-1 and for sigma almost-equal-to 10(-25) cM2, F(m) < 1.0 X 10(-15) CM-2 sr-1 sec-1 at 90% C.L.

The IMB experiment, a large water Cherenkov detector which began data collection in September 1982, has undergone several upgrades to improve light collection, on-line processing power, data throughput and buffering, calibration, and operating efficiency. The current device, known as IMB-3, enjoys a factor of four light collection advantage over its precursor. Since May 1986, it has been used to search for such diverse phenomena as nucleon decay, dark matter, neutrino oscillation, and magnetic monopoles, and to study stellar collapse and cosmic rays. Due to its large size and long exposure time IMB presents unique challenges. The design and operation of the IMB-3 detector are described in detail.

Neutrino interactions from a 7.7 kton yr exposure of the IMB-3 detector are analyzed. A total of 935 contained events radiating over approximately 50 MeV of Cerenkov-equivalent energy and consistent with atmospheric neutrino interactions are identified. Of these, 610 have a single Cerenkov ring. Single-ring interactions are efficiently separated into those containing a showering particle (produced mainly by nu(e)) and those containing a nonshowering particle (produced mainly by nu(mu)). In the momentum range 100 < p(e) < 1500 MeV/c and 300 < p(mu) < 1500 MeV/c, the fraction of nonshowering events is 0.36+/-0.02(stat)+/-0.02(syst). Based on detailed models of neutrino production and interaction, a fraction of 0.51+/-0.01(stat)+/-0.05(syst) is expected. This deficit of nonshowering, or excess of showering, events relative to the total is supported by an independent analysis of muon decay signals. In the same sample 33+/-2(stat)% of events are accompanied by one or more muon decays, while 43+/-1(stat)% are expected. Further studies that could reduce systematic errors and discover the cause of these discrepancies are suggested.

Muon neutrinos produced as a result of cosmic-rav interactions with the atmosphere are used to search for nu(mu) oscillations into nu(tau) by comparing the measured rate of upward-going muons in the Irvine-Michigan-Brookhaven detector with the expected rate. In addition, the ratio of upward-going muons which stop in the detector to those which exit is used to search for deviations from the expected spectrum. This latter technique is free of flux and cross-section normalization uncertainties. No evidence for oscillations is found. 90% C.L. limits on delta-m2 are derived in the range (1-2) x 10(-4) eV2 for sin(2)2-theta > 0.5.