Observation of the east-west anisotropy of the atmospheric neutrino flux (Jan. 12, 1999) The Super-Kamiokande Collaboration T.Futagamiu, Y.Fukudaa, T.Hayakawaa, K.Inouea, K.Ishiharaa, H.Ishinoa, Y.Itowa, T.Kajitaa, J.Kamedaa, S.Kasugaa, K.Kobayashia, Y.Kobayashia, Y.Koshioa, M.Miuraa, M.Nakahataa, S.Nakayamaa, Y.Obayashia, A.Okadaa, K.Okumuraa, N.Sakuraia, 9 M.Shiozawaa, Y.Suzukia, H.Takeuchia, Y.Takeuchia, Y.Totsukaa, S.Yamadaa, M.Earlb, 9 A.Habigb, E.Kearnsb, M.D.Messierb, K.Scholbergb, J.L.Stoneb, L.R.Sulakb, C.W.Walterb, 9 1 M.Goldhaberc, T.Barszczakd, D.Casperd, W.Gajewskid, W.R.Kroppd, S.Mined, L.R.Priced, n M.Smyd, H.W.Sobeld, M.R.Vaginsd, K.S.Ganezere, W.E.Keige, R.W.Ellsworthf, S.Tasakag, a A.Kibayashih, J.G.Learnedh, S.Matsunoh, V.J.Stengerh, D.Takemorih, T.Ishiii, J.Kanzakii, J T.Kobayashii, K.Nakamurai, K.Nishikawai, Y.Oyamai, A.Sakaii, M.Sakudai, O.Sasakii, 2 1 S.Echigoj, M.Kohamaj, A.T.Suzukij, T.J.Hainesk,d, E.Blaufussl, B.K.Kiml, R.Sanfordl, R.Svobodal, M.L.Chenm,J.A.Goodmanm, G.W.Sullivanm, J.Hilln, C.K.Jungn, K.Martensn, 1 v C.Maugern, C.McGrewn, E.Sharkeyn, B.Virenn, C.Yanagisawan, W.Dokio, M.Kirisawao, 9 S.Inabao, K.Miyanoo, H.Okazawao, C.Sajio, M.Takahashio, M.Takahatao, K.Higuchip, 3 1 Y.Nagashimap, M.Takitap, T.Yamaguchip, M.Yoshidap, S.B.Kimq, M.Etohr, A.Hasegawar, 1 T.Hasegawar, S.Hatakeyamar, T.Iwamotor, M.Kogar, T.Maruyamar, H.Ogawar, J.Shirair, 0 9 A.Suzukir, F.Tsushimar, M.Koshibas, Y.Hatakeyamat, M.Koiket, M.Nemotot, 9 K.Nishijimat, H.Fujiyasuu, Y.Hayatou, Y.Kanayau, K.Kaneyukiu, Y.Watanabeu, / h D.Kielczewskav,d, J.S.Georgew,∗, A.L.Stachyraw, L.L.Waiw,§, R.J.Wilkesw, K.K.Youngw,† p - aInstitute for Cosmic Ray Research, University of Tokyo, Tanashi, Tokyo 188-8502, Japan o bDepartment of Physics, Boston University, Boston, MA 02215, USA r st cPhysics Department, Brookhaven National Laboratory, Upton, NY 11973, USA a dDepartment of Physics and Astronomy, University of California, Irvine, Irvine, CA 92697-4575, USA : v eDepartment of Physics, California State University, Dominguez Hills, Carson, CA 90747, USA i X fDepartment of Physics, George Mason University, Fairfax, VA 22030, USA r gDepartment of Physics, Gifu University, Gifu, Gifu 501-1193, Japan a hDepartment of Physics and Astronomy, University of Hawaii, Honolulu, HI 96822, USA iInstitute of Particle and Nuclear Studies, High Energy Accelerator Research Organization (KEK), Tsukuba, Ibaraki 305-0801, Japan jDepartment of Physics, Kobe University, Kobe, Hyogo 657-8501, Japan kPhysics Division, P-23, Los Alamos National Laboratory, Los Alamos, NM 87544, USA. lDepartment of Physics and Astronomy, Louisiana State University, Baton Rouge, LA 70803, USA mDepartment of Physics, University of Maryland, College Park, MD 20742, USA nDepartment of Physics and Astronomy, State University of New York, Stony Brook, NY 11794-3800, USA oDepartment of Physics, Niigata University, Niigata, Niigata 950-2181, Japan pDepartment of Physics, Osaka University, Toyonaka, Osaka 560-0043, Japan qDepartment of Physics, Seoul National University, Seoul 151-742, Korea rDepartment of Physics, Tohoku University, Sendai, Miyagi 980-8578, Japan sThe University of Tokyo, Tokyo 113-0033, Japan tDepartment of Physics, Tokai University, Hiratsuka, Kanagawa 259-1292, Japan 1 uDepartment of Physics, Tokyo Institute of Technology, Meguro, Tokyo 152-8551, Japan vInstitute of Experimental Physics, Warsaw University, 00-681 Warsaw, Poland wDepartment of Physics, University of Washington, Seattle, WA 98195-1560, USA Abstract The east-west anisotropy, caused by the deflection of primary cosmic rays in the Earth’s magnetic field, is observed for the first time in the flux of at- mospheric neutrinos. Using a 45 ktyear exposure of the Super-Kamiokande · detector, 552e-likeand633µ-likehorizontally-going events areselected inthe momentum range between 400 and 3000 MeV/c. The azimuthal distribution of e-like and µ-like events agrees with the expectation from atmospheric neu- trino flux calculations that account for the geomagnetic field, verifying that the geomagnetic field effects in the production of atmospheric neutrinos in the GeV energy range are well understood. 96.40.Tv,14.60.Pq,95.85.Ry Typeset using REVTEX 2 The primary cosmic ray flux approaching the Earth is known to be almost isotropic [1]. However, since the Earth has a magnetic field and the primary cosmic rays are positively charged, an angular anisotropy is produced for those primaries which reach and interact in the atmosphere. This azimuthal anisotropy, called the east-west effect, was discovered in the 1930’s as a deficit of secondary cosmic rays (muons) arriving from the easterly direction compared to a westerly direction [2]. In fact, this effect was used to infer that the primary cosmicraysarepositivelycharged. Theazimuthalanisotropyvariesaccordingtotheposition on the Earth and the momentum and charge of the primary cosmic ray. The anisotropy is largest for the lowest momenta particles. The anisotropy can be characterized by a cut-off momentum, the maximum momentum at which the incoming particle cannot reach the atmosphere. At the Super-Kamiokande experiment, the cut-off momentum for protons averages about 10 GeV/c. Since Super-Kamiokande is located close to the geomagnetic equator (25.8◦ N geomagnetic latitude), the cut-off momentum for horizontally arriving protons from the east is 50 GeV/c, considerably higher than the average over directions. ∼ Figure 1 illustrates typical allowed and forbidden tracks of cosmic rays. This results in the depletion of primary cosmic ray interactions east of Kamioka. Neutrinos produced in the atmosphere by these cosmic rays should likewise be depleted in the easterly direction. Inthispaperwereportthefirstobservationoftheeast-westanisotropyintheatmospheric neutrinoflux. Weobserveaneutrinoeast-west effectwhichisinagreement withexpectations based on detailed flux calculations and a Monte Carlo simulation of neutrino interactions. This confirms that the angular dependance of the atmospheric neutrino flux calculations in the GeV energy range are reasonable. The Super-Kamiokande detector is a 50 kt water Cherenkov detector located at the Kamioka Observatory, ICRR, Univ. of Tokyo. The center of the detector is at 36◦25′33′′ N, 137◦18′37′′ E, 371.8 m above the sea level. To reduce cosmic ray muons, it is placed at 2700 meters-water-equivalent below the peak of Mt. Ikenoyama. It is cylindrically shaped with the height of 41.4 meters and the diameter of 39.3 meters, and divided into the inner and outer detectors by a stainless steel support structure with a pair of opaque sheets. The inner detector is viewed by 11,146 photo-multiplier tubes (PMTs) of 50-cm diameter, facing inward to detect the neutrino events. The outer detector, a layer of water of 2.6 to 2.75 m thick, has 1885 PMTs of 20-cm diameter used to veto entering events and to tag exiting events. The energy and direction of a neutrino are measured from the momentum of the final state charged lepton which is produced in a charged current interaction of the neutrino on nucleus, ν + N l + X. The flavor of the final state lepton is used for tagging the → flavor of the neutrino. With the 45 kt yr exposure of the Super-Kamiokande, a total of · 4077 fully-contained (FC) single-ring atmospheric neutrino events were observed with the following requirements: (1)totalnumber of hitsinouter detector less than25,andno spatial cluster with more than 10 hits, (2) total charge collected in the inner detector more than 200 p.e.s, (3) the ratio (maximum p.e. in any single PMT)/(total p.e.s) less than 0.5, (4) the time interval from preceding event larger than 100 µs, (5) the vertex position should be inside the fiducial volume, 2 meters inside from the inner detector wall, and (6)single-ring events with the momentum higher than 100 MeV/c and 200 MeV/c for e-like and µ-like events, respectively. To determine the direction of neutrinos from the observed leptons, only single Cherenkov ring events were used. Then they were separated into e-like and µ-like 3 events according to their showering or non-showering signature. The particle identification efficiency was estimated to be better than 99% [5]. For analysis of the east-west anisotropy, the following additional cuts were applied: (7) momentum was required to be between 400 MeV/c and 3000 MeV/c for both e-like and µ-like events and (8) cosine of the zenith angle of the leptons was required to be between -0.5 and 0.5 . Criterion (7) selects the momentum region where the east-west anisotropy is expected to be most significant. The east-west anisotropy of the flux is expected to be larger for lower energy neutrinos because of the cutoff momentum. On the other hand, the poor determination of the neutrino direction at low energy tends to wash out the flux anisotropy. The estimated mean scattering angle in the momentum region of criterion (7) is 36 degrees for both νe,νµ charged current events. The contamination of neutral current events in these e-like and µ-like events were estimated to be 9.0% and 1.3%, respectively. Criterion (8) is to select neutrino events from near the horizon where the effect of the geomagnetic cutoff is maximum. After these criteria, 552 e-like and 633 µ-like events remained. These data were com- pared with Monte Carlo simulations with two independent flux calculations [3,4]. These calculations represented the geomagnetic field by a multipole expansion of the spherical harmonic function. Cutoffs are calculated for all zenith and azimuthal angles at the Super- Kamiokande detector site by backtracing antiprotons through the three dimensional map of the geomagnetic field. Figure 2 shows the azimuthal distributions for e-like and µ-like events. The number of Monte Carlo events was normalized to the number of data events to compare the azimuthal shape without regard to the 20% uncertainty in normalization and the deficit of µ-like events most likely due to neutrino oscillations [5–7]. The Monte Carlo events we used here assumed no neutrino oscillations. The lengths of the flight paths of the neutrinos vary with zenith angle but do not vary with the azimuthal angle. Therefore, neutrino flavor oscillations have little effect on east-west anisotropy. The estimated effect is far below the statistical accuracy of the data. The azimuthal angles are divided into 8 bins, 45 degrees each. Agreement between the 2 data and the Monte Carlo is good: the χ values were 5.1/7DOF and 2.6/7DOF for e-like and µ-like events, respectively. A quantitative comparison between the data and Monte Carlo was performed using a Kuiper test [8,9]. This test calculates a binning and starting-point free probability that observed data are the result of an assumed distribution. The Kuiper statistic V is defined as: V = max [S (φ) P(φ)]+ max [P(φ) S (φ)], N N 0<φ<2π − 0<φ<2π − where φ is the azimuthal angle, S (φ) is a cumulative probability function from data and N P(φ) is the one from Monte Carlo. The significance is obtained from the statistic V∗ = V(√n+0.155+0.24/√n), and defined as ∞ Prob = 2X(4j2V∗2 1)exp( 2j2V∗2), − − j=1 where n is number of events. From this test, the probability that the azimuthal distribution of the data originated from a flat parent distribution was 0.0008% (20%) for e-like (µ-like) events. The azimuthal 4 distribution of e-like events is inconsistent with a flat distribution at more than 99% CL. Also, the probabilities that the data match the Monte Carlo in shape with the flux of Ref. [3] were 42% for e-like events and 92% for µ-like events. For a Monte Carlo with neutrino oscillationswith(∆m2,sin22θ)=(2.2 10−3 eV2,1.0)[7],theprobabilitywasthesamewithin × 1% for both e-like and µ-like events. As expected, the probability did not change much with the addition of neutrino oscillations. With current statistics, the deficit of the westward-going events is apparent in the e- like but not in the µ-like data sample. The smaller anisotropy observed in µ-like data is consistent with a statistical fluctuation. The expected anisotropy is similar for ν and ν e µ [3,4]; the calculated anisotropy for ν is only about 1.1 times larger than that for ν in the e µ relevant energy region. The study of the east-west anisotropy as a function of lepton momentum is important for our understanding of the neutrino-lepton angular correlation and the geomagnetic field effect on the production of atmospheric neutrinos. Figure 3 shows the east-west asymmetry (N N )/(N +N ) as a function of lepton momentum, where N (N ) represents the E W E W E W − number of eastward(westward)-going events. Here the azimuthal distribution was divided into two bins, and the same zenith angle cut (8) was used. The data and the Monte Carlo 2 2 agreed well; the χ values were 5.9/6DOF (5.8/5DOF) for e-like (µ-like) events. The χ values of a comparison of the data and a straight line at (N N )/(N +N ) = 0 were E W E W − 26.5/6DOF (10.5/5DOF) for e-like (µ-like) events. As expected from the large neutrino- lepton scattering angle, the measured east-west asymmetry was small below 400MeV/c. Above3GeV/c,theneutrinosoriginatefromprimarycosmicraysthatareminimallyaffected by the geomagnetic field, so the flux is symmetric from the east and west. Insummary,wehaveobservedforthefirsttimeclearevidencefortheeast-westanisotropy in the atmospheric neutrinos, based on the data from 45 kt-yr of exposure of the Super- Kamiokande detector. A deficit of westward-going e-like events was seen at more than 99% confidence level, while for µ-like events, it was statistically less significant. However, the azimuthal distributions of these events and the momentum dependence of the east-west anisotropy for both e-like and µ-like agreed well with the prediction based on detailed neu- trino flux calculations. This observation strongly suggests that the geomagnetic field effects in the production of atmospheric neutrinos in the GeV energy range are well understood. We gratefully acknowledge the cooperation of the Kamioka Mining and Smelting Com- pany. The Superk-Kamiokande experiment was built from, and has been operated with, funding by the Japanese Ministry of Education, Science, Sports andCulture, andthe United States Department of Energy. 5 REFERENCES † Deceased ∗ Present address: Space Radiation Laboratory, California Inst. of Technology, Pasadena, CA 91125 USA § Present address: Department of Physics, Stanford University, CA 94305, USA [1] See, for sample, K.Nagashima, K.Fujimoto and R.M.Jacklyn, J. Geophys. Res. 103, 17429 (1998); K.Munakata et al., Phys. Rev. D 56, 23 (1997). [2] T. H. Johnson, Phys. Rev. 43, 307, 381 (1933); T. H. Johnson, Phys. Rev. 48, 287 (1935). ; see also B. Rossi, ’Cosmic Rays’, McGraw Hill (1964) and references therein. [3] M. Honda et al., Phys. Rev. D 52, 4985 (1995). [4] P. Lipari, T. K. Gaisser and T. Stanev, Phys. Rev. D 58, 073003 (1998). [5] Y. Fukuda et al., Phys. Lett. B433, 9 (1998). [6] Y. Fukuda et al., Phys. Lett. B 436, 33 (1998). [7] Y. Fukuda et al., Phys. Rev. Lett. 81, 1562 (1998). [8] M. A. Stephens, J. Royal Statist. Soc. B32, 115 (1970). [9] W. H. Press, S. A. Teukolsky, W. T. Vetterling, and B. P. Flannery, Numerical Recipes in FORTRAN, Cambridge University Press, second edition (1992). [10] S. Barr et al., Phys. Lett. B214, 147 (1988). 6 FIGURES the Earth forbidden track (low energy) SK north pole fi n B allowed track (high energy) FIG. 1. Typicalallowedandforbiddentracksoftheincidentcosmicrays. Thecurvedlineshows the allowed trajectory of the cosmic ray which has momentum larger than cut-off momentum and the dotted curved line shows forbidden trajectory with momentum below cut-off. ’SK’ represents the Super-Kamiokande detector. 100 100 s80 s80 nt nt e e v v e e of 60 of 60 r r e e b b m40 m40 u u n n 20 20 e-like m -like 0 0 0 p /2 p 3p /2 2p 0 p /2 p 3p /2 2p f f FIG. 2. Azimuthal angle distributions of e-like and µ-like events. The crosses represent the data points, the histogram drawn by solid line (dashed-line) shows the prediction of the Monte Carlo based on the flux of Ref. [3] ( [4]). Data are shown with statistical errors. The Monte Carlo has 10 times more statistic than data. The Monte Carlo histogram is normalized to the total number of the real data. φ represents the azimuthal angle. φ = 0, π/2, π and 3π/2 shows particles going to north, west, south, and east, respectively. 7 0.4 e-like 0.3 ) W N 0.2 + E 0.1 N ( 0 )/ W -0.1 N -E-0.2 N (-0.3 -0.4 2 3 4 10 10 10 0.4 m -like 0.3 ) W N 0.2 + E 0.1 N ( 0 )/ W -0.1 N -E-0.2 N (-0.3 -0.4 2 3 4 10 10 10 Lepton Momentum (MeV/c) FIG. 3. East-west asymmetry (N N )/(N +N ) as a function of lepton momentum for E W E W − e-like and µ-like events. The crosses with error bars represent the data and the hatched region represents the prediction based on the flux of Ref. [3] ; error bars are statistical. 8