a search for νμ→ντ oscillation

2 April 1998

Ž .Physics Letters B 424 1998 202–212

A search for n ™n oscillationm t

CHORUS Collaboration

E. Eskut a, A. Kayis a, G. Onengut a, R. van Dantzig b, M. de Jong b, J. Konijn b,Ö. Melzer b, R.G.C. Oldeman b, C.A.F.J. van der Poel b, J.W.E. Uiterwijk b,

J.L. Visschers b, A.S. Ayan c, E. Pesen c, M. Serin-Zeyrek c, R. Sever c, P. Tolun c,M.T. Zeyrek c, N. Armenise d, F. Cassol d, M.G. Catanesi d, M.T. Muciaccia d,

E. Radicioni d, S. Simone d, L. Vivolo d, A. Bulte e, P. Lendermann e,Ä. Meyer-Sievers e, T. Patzak e,1, K. Winter e, P. Annis f,2, M. Gruwe f,4,Ć. Mommaert f,6, M. Vander Donckt f,4, B. Van de Vyver f,7, P. Vilain f,8,

G. Wilquet f,8, P. Righini g, B. Saitta g, E. Di Capua h, C. Luppi h, P. Zucchelli h,Y. Ishii i, T. Kawamura i, M. Kazuno i, S. Ogawa i, H. Shibuya i, R. Beyer j,3,

J. Brunner j, D. Cussans j, J.P. Fabre j, R. Ferreira j,9, W. Flegel j, R. Gurin j,10,M. Litmaath j, L. Ludovici j,11, D. Macina j,12, R. Meijer Drees j,13, H. Meinhard j,

P. Migliozzi j, E. Niu j, H. Øveras j, J. Panman j, I.M. Papadopoulos j, F. Riccardi j,˙S. Ricciardi j, A. Rozanov j,14, D. Saltzberg j,15, R. Santacesaria j,11, G. Stefanini j,R. Tsenov j,16, Ch. Weinheimer j,17, H. Wong j,18, J. Goldberg k, K. Hoepfner k,3,M. Chikawa l, E. Arik m, I. Birol m, A.A. Mailov m, C.H. Hahn n,19, H.I. Jang n,

1 Now at College de France, Paris, France.2 Supported by Regione autonoma della Sardegna, Italy.3 Now at DESY, Hamburg, Germany.4 Fonds pour la Formation a la Recherche dans l’Industrie et dans l’Agriculture.5 Now at CERN, Geneva, Switzerland.6 Interuniversitair Instituut voor Kernwetenschappen.7 Nationaal Fonds voor Wetenschappelijk Onderzoek.8 Fonds National de la Recherche Scientifique.9 On leave of absence from LIP, Lisbon.

10 CASPUR, Rome, Italy.11 On leave of absence from INFN Sezione di Roma, Rome, Italy.12 Now at Universite de Geneve, Geneva, Switzerland.´ `13 Now at University of Washington, Seattle, USA.14 Now at CPPM CNRS-IN2P3, Marseille, France.15 Now at U.C.L.A., Los Angeles, USA.16 On leave of absence from Sofia University, Bulgaria, with support from the Bogazici University, Centre for Turkish-Balkan Physics

Research and Applications.17 Now at University of Mainz, Mainz, Germany.18 Now at Academia Sinica, Taipei, Taiwan.19 Now at the Changwon Nat. Univ., Changwon, Korea.

0370-2693r98r$19.00 q 1998 Elsevier Science B.V. All rights reserved.Ž .PII S0370-2693 98 00081-1

( )E. Eskut et al.rPhysics Letters B 424 1998 202–212 203

I.G. Park n, M.S. Park n, J.S. Song n, C.S. Yoon n, K. Kodama o, N. Ushida o,S. Aoki p, T. Hara p, G. Brooijmans q,20, D. Favart q, G. Gregoire q, J. Herin q,20,´ ´

ˆ q,3 r r r rV. Lemaitre , A. Artamonov , P. Gorbunov , V. Khovansky , V. Shamanov ,V. Smirnitsky r, D. Bonekamper s, D. Frekers s, D. Rondeshagen s, T. Wolff s,21,¨

K. Hoshino t, M. Kobayashi t, Y. Kotaka t, T. Kozaki t, M. Miyanishi t,M. Nakamura t, K. Nakazawa t,22, T. Nakano t, K. Niu t, K. Niwa t, Y. Obayashi t,

O. Sato t, T. Toshito t, S. Buontempo u, A. Cocco u, N. D’Ambrosio u,G. De Lellis u, A. Ereditato u, G. Fiorillo u, F. Garufi u, F. Marchetti-Stasi u,M. Messina u, V. Palladino u, S. Sorrentino u, P. Strolin u, K. Nakamura v,T. Okusawa v, T. Yoshida v, A. Capone w, D. De Pedis w, S. Di Liberto w,

U. Dore w, P.F. Loverre w, A. Maslennikov w,10, M.A. Mazzoni w, F. Meddi w,G. Piredda w, A. di Bartolomeo x, C. Bozza x, G. Grella x, G. Iovane x,

G. Romano x, G. Rosa x,23, Y. Sato y, I. Tezuka y

a CukuroÕa UniÕersity, Adana, Turkeyb NIKHEF, Amsterdam, The Netherlands

c METU, Ankara, Turkeyd UniÕersita di Bari and INFN, Bari, Italy`

e Humboldt UniÕersitat, Berlin, Germany 24¨f ( )Inter-UniÕersity Institute for High Energies ULB-VUB , Brussels, Belgium

g UniÕersita di Cagliari and INFN, Cagliari, Italy`h UniÕersita di Ferrara and INFN, Ferrara, Italy`

i Toho UniÕersity, Funabashi, Japanj CERN, GeneÕa, Switzerland

k Technion, Haifa, Israell Kinki UniÕersity, Higashiosaka, Japanm Bogazici UniÕersity, Istanbul, Turkey

n Gyeongsang National UniÕersity, Jinju, South Korea 25

o Aichi UniÕersity of Education, Kariya, Japanp Kobe UniÕersity, Kobe, Japan

q UniÕersite Catholique de LouÕain, LouÕain-la-NeuÕe, Belgium´r Institute for Theoretical and Experimental Physics, Moscow, Russian Federation

s Westfalische Wilhelms - UniÕersitat, Munster, Germany 24¨ ¨ ¨t Nagoya UniÕersity, Nagoya, Japan

u UniÕersita Federico II and INFN, Naples, Italy`v Osaka City UniÕersity, Osaka, Japan

w UniÕersita La Sapienza and INFN, Rome, Italy`x UniÕersita di Salerno and INFN, Salerno, Italy`

y Utsunomiya UniÕersity, Utsunomiya, Japan

Received 8 December 1997; revised 5 January 1998Editor: L. Montanet

20 Institut Interuniversitaire des Science Nucleaires.´21 Supported by a grant from Deutsche Forschungs Gemeinschaft.22 Now at Gifu University, Gifu, Japan.23 Now at Universita La Sapienza and INFN, Rome, Italy.`24 Supported by German Bundesministerium fur Bildung und Forschung under contract numbers 056BU11P and 057MS12P.¨25 Ž .Supported by the Korea Scince and Engineering Foundation, and the Ministry of Education through Research Fund BSRI-97-2407 ,

Republic of Korea.

( )E. Eskut et al.rPhysics Letters B 424 1998 202–212204

Abstract

CHORUS is an experiment searching for n ™n oscillation in the CERN wide band neutrino beam with a hybrid setupm t

consisting of a nuclear emulsion target followed by electronic detectors. The experiment has been taking data from 1994through 1997. A subset of the neutrino interactions collected in 1994 and 1995 have been analyzed, looking for n chargedt

current interactions where the t lepton decays to mn n . In a sample of 31,423 n charged current interactions, no nm t m t

candidates were found. For large Dm2 values, a limit on the mixing angle of sin22u -3.5=10y3 at 90% C.L. can bemt mt

set, thus improving the previous best result. q 1998 Elsevier Science B.V.

1. Introduction

CHORUS is an experiment designed to search forn ™ n oscillation through the observation ofm t

charged current interactions n N™tyX, followedt

by the decay of the t lepton. The experiment canprobe neutrino masses difference above few eVŽ 2 2 .Dm R1 eV . Massive neutrinos in this rangemt

have been proposed as candidates for the hot compo-w xnent of dark matter in the universe 1 , and the

n ™n oscillation search is further motivated if onem t

assumes a hierarchical pattern of neutrino masses.The experiment is performed in the CERN Wide-

Band Neutrino Beam, which contains mainly nm

with a contamination from n well below the level oft

sensitivity that can be reached in this experiment.Neutrino interactions occur in a target of nuclear

Žemulsion, whose exceptional spatial resolution be-. Žlow one micrometer and hit density 300 grainsrmm

.along the track allow a three dimensional ‘‘visual’’reconstruction of the trajectories of the t lepton andits decay products.

The experiment is sensitive to most of the decaychannels of the t; in this paper however we reporton the results obtained with a subsample where we

y yconfined ourselves to the t ™m n n decay search.t m

The detection of the decay topology in the nuclearemulsion, together with the reconstruction of theevent kinematics in the electronic detectors, makeCHORUS an essentially background free n appear-t

ance experiment.

2. The experimental setup

2.1. The neutrino beam

Ž .The CERN wide band neutrino beam WBBcontains dominantly muon neutrinos from pq and

qK decay, with a n contamination of 5% , a n ,nem e

contamination at a level of 1%. The estimated nt

w x y6background 2 is of the order of 3.3=10 nt

charged current interactions per n charged currentm

interaction, hence negligible. The n component ofm

the beam has an average energy of 27 GeV.

2.2. The apparatus

ŽThe short decay length of the t 1.5 mm onaverage, assuming that the n have the same energyt

.spectrum as the n beam and the large statisticsm

w xneeded to improve the best existing result 3 haveguided the design of the apparatus, which is de-

w xscribed in Ref. 4 . The hybrid setup, shown in Fig.1, is composed of an emulsion target, a scintillatingfiber tracker system, trigger hodoscopes, a magneticspectrometer, a lead-scintillator calorimeter and amuon spectrometer.

The emulsion target has a mass of 770 kg and asurface area of 1.42=1.44 m2. It consists of fourstacks of 36 plates each. Each plate has two 350 mmthick layers of nuclear emulsion on both sides of a90 mm thick plastic base.

Downstream of each emulsion stack there arethree sets of interface emulsion sheets of the samelateral dimensions as the emulsion target. Each inter-face emulsion sheet consists of a 800 mm acrylicbase coated, on both sides, with a 100 mm thickemulsion layer. They are placed between each emul-sion stack and the following fiber tracker module asshown in Fig. 2.

Scintillating fiber trackers locate the trajectoriesof charged particles produced in the neutrino interac-tion. The trajectories of these particles are extrapo-lated to the downstream face of the emulsion stackwith the help of the interface emulsion sheets. The


Fig. 1. General layout of the detector.

fiber tracker resolution on lateral position and direc-tion is ss150 mm and ss2 mrad, respectively.

Downstream of the target region, a magnetic spec-trometer allows the reconstruction of the charge andmomentum of charged particles. The spectrometerconsists of an air-core magnet of hexagonal shapeand three scintillating fiber trackers, one in front andtwo behind the air-core magnet, which record thecharged particles’ trajectory and deflection.

A lead scintillating fiber calorimeter following thespectrometer measures the energy and direction ofthe hadronic showers and allows neutral particlesdetection.

The calorimeter is followed by a muon spectrome-ter which identifies muons and measures their chargeand momentum. It is composed of magnetized irondisks and tracking devices. A momentum resolutionof 19% is achieved above 7 GeVrc by magnetic

Fig. 2. Layout of an emulsion stack and associated fiber trackers.


field deflection. By range measurement, a 6% resolu-tion is obtained below 7 GeVrc.

3. The data collection

The detector has been exposed to the WBB fromŽ .1994 to 1997. After a first run 1994–95 , the target

emulsion was replaced and the exposed emulsiondeveloped. During this period CHORUS collectedapproximately 969,000 triggers, corresponding to2.01=1019 protons on target. On the basis of neu-trino flux estimate, trigger efficiency, cross section,dead-time correction and target mass, about 320,000charged current n interactions are expected to havem

occurred in the emulsion target. Of these 250,932events have been reconstructed in the electronic de-tectors with an identified negative muon and aninteraction vertex in the emulsion.

I I4. The t ™m n n search analysist m

4.1. Principles

y yThe search for t ™m n n decays starts fromt m

the reconstructed events recorded in the electronicdetectors. The event selection, described later, pro-vides us with a set of events with a reconstructedmuon track and a prediction for the trajectory of themuon through the interface emulsion sheets. Themuon track is located in the interface emulsion sheetsusing automatic techniques and followed into theemulsion target in order to locate the plate where theinteraction had occurred. Automatic criteria are then

applied to select candidate events with the one-prongdecay topology we are looking for.

If the automatic decay search does not reject theevent, computer assisted eye scan of the vertex plateŽ .and downstream plates if necessary is performed,

y yin order to assess the presence of a t ™m n nt m

decay topology. If a candidate is found, all tracks aremeasured in the emulsion and connected to thosefound in the electronic detectors; in addition, theevent kinematics can be reconstructed.

4.2. EÕent selection

Only events with one negative muon, of momen-tum P -30 GeVrc, have been considered. The mo-m

mentum cut reduces the number of events to befurther analyzed to 177,827. Since muons from t

decay have a lower momentum, on average, thanthose produced in charged current interactions, thisselection would reject only a small fraction of gen-

Žuine n interactions 15% in the case that n have thet t

.same energy spectrum as the n beam , while reduc-m

ing the n charged current interactions that wouldm

have to be scanned to 71% of the total.So far, only 79,500 events of the 177,827 have

been scanned. Since the scanning procedure requiresthe application of fiducial cuts on the angle andposition of the muon track, the sample is furtherreduced to 67,813 events. The number of events ateach stage of the analysis are summarized in Table 1.

4.3. The Õertex location

The emulsion scanning procedure is fully auto-mated using computer controlled microscopes

Table 1y yCurrent status of the CHORUS t ™m n n searcht m

Year 1994 199519 19Protons on target 0.81=10 1.20=10

Emulsion triggers 422,000 547,000Charged current interactions expected 120,000 200,000Events with one identified muon and vertex predicted in emulsions 95,374 155,558Events with P -30 GeV 66,911 110,916m

Events scanned so far 41,931 37,569Events within the fiducial cuts 35,767 32,046Events with a vertex found in the emulsion 16,837 14,586


equipped with CCD cameras and fast processors: thisis the first time that fully automatic scanning isachieved. The processor, which is called track-selec-

w xtor 5 , is capable of identifying the tracks inside theemulsion and measuring their parameters on-line. Inthe scanning procedure, these systems are used tolocate the negative muon track in the interface emul-sion sheets, and to follow it inside the target, plateby plate, until it disappears in two consecutive plates.The first plate where the track disappears is calledthe Õertex plate. The vertex location procedure, calledscan-back, has been successful for 31,423 chargedcurrent events.

Once the vertex plate has been located, the decaysearch, described in more detail in the followingsection, is performed.

4.4. The decay search procedure

An event with a t lepton is identified by theŽ . ypresence of a change of direction kink of the t

y ytrack due to the t ™m n n semileptonic decay,t m

while a muon pointing to the vertex is the leadingmuon of a n charged current event. We consider asm

t candidates events satisfying the following criteria:Ž . yØ a decay signature kink along the m track;

Ø no other charged leptons at the primary vertex;

Fig. 3. Transverse momentum distribution in long decay pathŽ .events of data and of n N™ t ™m X MC simulation.t

Ø the muon transverse momentum P , with respectt

to the t candidate direction, is larger than250 MeVrc.

Ø the kink plate is located within 5 plates down-Žstream of the vertex plate corresponding to a

.maximum t decay path length of 3.95 mm .y yThe cut on P eliminates the K ™m n decays,t m

keeps the loss of efficiency for t events small and atthe same time discards the large majority of my

coming from n interactions at the level of automaticm

measurements. Fig. 3 shows the P distribution fort

the collected data sample compared with the MonteCarlo distribution expected for n events. Due to thet

ongoing development of the automatic scanning de-vices and procedures, different kink search algo-rithms have been applied to the analysis of 1994 and1995 data in order to optimize speed and efficiencyof the decay search procedure.

4.4.1. The 1994 data analysisThe decay search is performed following two

different procedures, which are applied concurrentlyon each event located.

)i short decay path search. This search is de-signed to detect decays in the vertex plate. Twodifferent approaches are followed according to thenumber of tracks reconstructed by the track-selectorat the upstream face of the plate following the vertexplate. If there is at least one track matched with the

Ž . ytracker predictions Fig. 4, topology a , the m

impact parameter is evaluated as the minimum dis-tance between the muon track and any other track of

Ž .the event Fig. 5 . Only events which show a largeimpact parameter, i.e. bigger than 2–8 mm accordingto the longitudinal position inside the plate, arevisually inspected and re-measured with semi-auto-matic techniques. For the events with no matchedhadrons, digital images of the microscope fields overthe whole vertex plate depth are recorded, and the

Žsearch for a kink continues off-line Õideo image.analysis .

)ii long decay path search. This procedure isdesigned to detect decays downstream of the vertexplate. The scan-back procedure follows the muontrack from one plate of the target to the next one,implicitly assuming the track straightness. A ‘long’


Fig. 4. Search for kink topologies, as described in the text.

decay path can thus be detected in two ways, accord-ing to its kink angle:Ø If the decay angle is larger than the scan-back

Ž .angular tolerance large angle kinks , the scan-back procedure stops and the kink plate is as-

Ž .sumed to be the vertex plate Fig. 4, topology b .Since no other emulsion tracks can be matched tothe fiber trackers predictions to apply the shortdecay path search, the video image analysis isundertaken.


Fig. 5. Impact parameter distributions in the search of short decayŽ .path decays of data and of n N™ t ™m X MC simulation.t

Ø If the kink angle is smaller than the tolerances ofŽ .the scan-back Fig. 4, topology c , the vertex

plate is indeed the interaction plate and decayscan be detected by measuring a track direction inthe plate immediately downstream of the vertexplate which is not compatible with measurements

Žin the other detectors emulsion plates, interface.emulsions, fiber tracker . If the measured trans-

verse momentum P fDuPP is larger thant m

250 MeVrc, the complete event is scanned byeye in five plates downstream of the vertex plate.

4.4.2. The 1995 data analysisCurrently only the long decay path search has

been performed on 1995 data, with the recentlydeveloped ‘‘parent track’’ search technique appliedfor large angle kink detection. As explained in the1994 long decay path search, whenever a large angledecay is found, the kink plate is assumed to be the

Ž .vertex plate Fig. 4, topology b . In the ‘parenttrack’ search technique the upstream part of thevertex plate is scanned in order to check the possibil-ity of associating a track with the muon. If theminimal distance between a track and the muon isless than 15 mm, this track will be considered its‘parent’. Checks are made to ensure that the parenttrack is not a passing-through track accidentally as-sociated to the muon. If a parent track is found, theevent is scanned by eye.

5. Analysis and result

A total of 31,423 muonic events have been ana-lyzed as described above, and no tau decays have

y ybeen found. By evaluating the t ™m n n detec-t m

tion efficiency this negatiÕe result is used to excludea significant region of the oscillation parameter space.

5.1. Oscillation sensitiÕity

In the two flavour mixing scheme, the result canbe expressed as an exclusion plot in the parameter

Ž 2 2 .space sin 2u ,Dm . The oscillation probabilitymt mt

depends on the number of observed n and n events.t m

The total number of expected charged current nm

interactions is given by

N sSP F Ps PA PdEHm n n nm m m

and the total number of observed n interactions witht

a muonic t decay is

N sSPsin2 2u PBRt mt

P F PP 2 Ps PA Pe PdEH n Dm n n km mt t t

sSPsin2 2u PBRP Imt

where

P 2 s C E, LŽ .HDmmt

2 w 2 x w x1.27PDm eV PL kmmt2Psin dLž /w xE GeV

andØ E is the incident neutrino energy;Ø L is the neutrino path-length to the emulsion

target;Ø u is the n yn mixing angle;mt m t

Ø Dm2 sm2 ym2 ;mt n nt m

ŽØ S is a constant factor identical for n and nm t

.interactions that takes into account the mass ofthe detector and the integrated neutrino flux;

Ø F is the n energy distribution;n mm

Ž .Ø C E, L is the n path distribution at a givenm

neutrino energy value;Ø s and s are the charged current n and nn n m tm t

differential cross sections;


Ø A and A are the energy dependent acceptancen nm t

and reconstruction efficiencies;Ø e is the energy dependent kink detection effi-k

ciency;y y ŽØ BR is the t ™m n n branching ratio 17.35"t m

w x.0.10%, 6 .Since we have observed zero candidate events, we

can express the 90% C.L. upper limit on sin2 2umt

according to:

2.37P F Ps PA dEH n n nm m m2sin 2u F 1Ž .mt N PBRP Im

In the above formula the numerical factor takes intoŽ .account the total systematic error 16% following

w xthe prescriptions given in 12 .For large Dm2 values, P 2 s1r2 and so themt Dmmt

previous formula can be simplified into

2P2.37Pr Prs A2 y3sin 2u F s3.5P10 2Ž .mt ² :N P e PBRm k

where² Ž .: ² Ž .:Ø r s s n r s n is the neutrino energys m t

w xweighted cross section ratio 13 . A value ofr s1.89"0.13 has been used; it takes into ac-s

count quasi elastic interactions, resonance produc-tion and deep inelastic reactions.

² Ž .: ² Ž .:Ø r s A n r A n is the cross sectionA m t

weighted acceptance ratio of n and n interac-m t

tions, its value has been evaluated to be 0.95"

0.07.² :Ø e is the average kink detection efficiency fork

the accepted events. Because of the different pro-cedures, its value is 0.54"0.06 for the 1994 datasample and 0.34"0.04 in 1995 where only thelong decay path search has been performed.A graphical representation of the oscillation pa-

rameter region excluded with the current data isshown in Fig. 6. Maximum mixing between n andm

n is excluded at 90% C.L. if Dm2 )1.5 eV 2.t mt

5.2. The t identification efficiency

The computation of the exclusion plot in theoscillation parameter space requires the knowledgeof the acceptance and reconstruction efficiencies Ant

and A and of the kink detection efficiency. Thesenm

w xFig. 6. Present result and previous limits from CCFR 8 , CDHSw x w x w x9 , CHARM II 10 and E531 3 .

efficiencies have been estimated by Monte Carlosimulation of the detector and of the scanning proce-dure. The scanning procedure and hence A andnt

A have evolved in the course of the analysis.nm

Several approaches have been used in computing theabove quantities so we can estimate a systematicerror of 7%.

The kink detection efficiency e has been evalu-k

ated by Monte Carlo simulation and has been aver-aged over the different kink search procedures de-scribed previously. The relative systematic error hasbeen evaluated to be 10%.

Since we expect Dq production in n inducedm

w xinteractions 7

n N™myDqXm

we can check experimentally the estimated kinkfinding efficiency on one-prong muonic charm de-cays

Dq™mqX 0nm

These events have two muons of opposite charge inŽ .the final state dimuons and – except for the parti-

cle charges – the topology of the muonic charmdecay is the same as that of the t decay. A sample ofpositive muons from dimuon events has been scannedto estimate the kink detection capability. A first


sample consisting of 10 dimuon events with a decaykink was selected by eye and the automatic t kinksearch applied. The kink finding efficiency of theautomatic procedure is consistent with our estimatesby Monte Carlo simulation. In an enlarged sample of17 events, including three prong decays, the distribu-tion of flight path length is in good agreement withthat obtained from the Monte Carlo simulation of theautomatic scanning procedure. Although the numberof events is too low to draw quantitative conclusions,we can take these results as a qualitative check of thesimulation of the automatic scanning procedure.

5.3. Backgrounds estimate

The main sources of potential background arecharm production and the n contamination of thet

beam.We expect less than 0.05 charm events in the

current sample from the anti-neutrino components ofthe beam:

q q yn n N™m e D XŽ .Ž .m e

followed by

Dy™myX 0

q Ž q.in which the m e escapes detection or is mis-identified.

Taking into account cross sections, branching ra-tio and kink detection efficiency, we expect thecurrent sample to contain less than 0.01 events due

w xto direct n contamination of the beam 2 .t

6. Conclusions

We have presented limits on n ™n oscillationm t

parameters based on the muonic decay channel ofthe t using a fraction of the data collected by theCHORUS experiment. The limit on sin2 2u formt

large Dm2 values is sin2 2u -3.5=10y3 at 90%mt mt

C.L., and improves the best previous result. A searchfor hadronic t decays using the sample of eventswithout a muon is in progress and results from

w xpartial statistics have been reported in 11 . By im-proving the efficiencies, by increasing the number oft decay channels searched for and by enlarging then search to the complete set of data we expect tot

reach the design sensitivity of the experiment. If nocandidates are found this corresponds to a limit ofsin2 2u -2=10y4 for large Dm2 values.mt mt

Acknowledgements

We gratefully acknowledge the help and supportof our numerous technical collaborators who con-tributed to the detector construction, operation, emul-sion pouring, development and scanning. We thankthe neutrino beam staff for the competent assistanceensuring the excellent performance of the facility.The accumulation of a large data sample in thisexperiment has been made possible also thanks tothe efforts of the crew operating the CERN PS andSPS. The general technical support from PPE, ECPand CN divisions is warmly acknowledged.

The experiment has been made possible by grantsfrom our funding agencies: the Institut Interuniversi-taire des Sciences Nucleaires and the Interuniversi-´

Ž .tair Instituut voor Kernwetenschappen Belgium ,Ž .The Israel Science foundation grant 328r94 and

the Technion Vice President Fund for the PromotionŽ . Ž .of Research Israel , CERN Geneva, Switzerland ,

the German Bundesministerium fur Bildung und¨Ž .Forschung Germany , the Institute of Theoretical

Ž .and Experimental Physics Moscow, Russia , theŽ .Istituto Nazionale di Fisica Nucleare Italy , the Japan

Private School Promotion Foundation and Japan So-Ž .ciety for the Promotion of Science Japan , the Korea

Science and Engineering Foundation and the Min-Ž .istry of Education Republic of Korea , the Founda-

tion for Fundamental Research on Matter FOM andthe National Scientific Research Organization NWOŽ .The Netherlands and the Scientific and Technical

Ž .Research Council of Turkey Turkey .We gratefully acknowledge their support.

References

w x1 Ya.B. Zel’dovic, I.D. Novikov, Relativistic Astrophysics,Ž .Nauka, Moscow, 1967; H. Harari, Phys. Lett. B 216 1989

413; J. Ellis, J.L. Lopez, D.V. Nanopoulos, Phys. Lett. B 292Ž .1992 189; H. Fritzsch, D. Holtmannspotter, Phys. Lett. B¨

Ž .338 1994 290.w x Ž .2 B. Van der Vyver, Nucl. Instr. and Meth. A 385 1997 91;

M.C. Gonzalez-Garcia, J.J. Gomez-Cadenas, Phys. Rev. DŽ .55 1997 1297.


w x Ž .3 N. Ushida et al., Phys. Rev. Lett. 57 1986 2897.w x4 E. Eskut et al., CHORUS Collaboration, CERN-PPEr97–

033, Nucl. Instr. and Meth., in print.w x Ž .5 S. Aoki et al., Nucl. Instr. and Meth. B 51 1990 466.w x Ž .6 Particle Data Group, Phys. Rev. D 54 1996 1.w x7 H. Abramowicz et al., CDHS Collaboration, Z Phys. C 15

Ž .1982 19; S.A. Rabinowitz et al., CCFR Collaboration,Ž .Phys. Rev. Lett. 70 1993 134; P. Vilain et al., CHARM II

Collaboration, Z. Phys. C, to be published.w x Ž .8 K.S. McFarland et al., Phys. Rev. Lett. 75 1995 3993.

w x9 F. Dydak et al. CDHS Collaboration, Phys. Lett. B 134Ž .1984 103.

w x10 M. Gruwe et al., CHARMII Collaboration, Phys. Lett. B 309´Ž .1993 463.

w x11 CHORUS Collaboration, in: Proceedings of the XVIII Inter-national Symposium on Lepton and Photon Interactions,Hamburg, 1997, contributed paper LP–189.

w x12 R.D. Cousins, V.L. Highland, Nucl. Instr. and Meth. A 320Ž .1992 331.

w x Ž .13 C. Jarlskog C.H. Albright, Nucl. Phys. B 84 1975 467.

a search for νμ→ντ oscillation

Documents