Measurement of the triple gauge boson coupling αWφ from W+W− production in e+e− collisions at s=161 GeV

27 March 1997

PHYSICS ELSEMEiR

LETTERS B

Physics Letters B 397 (1997) 147- 157

Measurement of the triple gauge boson coupling crw+ from WfW- production in e+e- collisions at fi = 161 GeV OPAL Collaboration K. Ackerstaffh, G. Alexander “, J. Allisonr, N. Altekampe, K. AmeteweeX, K.J. Anderson i, S. Andersone, S. Arcelli b, S. Asai w, D. Axen ab, G. Azuelos’?‘, A.H. Ball 9, E. Barberio h, R.J. Barlow P, R. Bartoldus ‘, J.R. Batley e, J. Bechtluft n, C. BeestonP, T. Behnke h, A.N. Bella, K.W. Bell t, G. Bella”, S. Bentvelsen h, P. Berlichj, S. Bethke”, 0. Biebel n, S.D. Birdr, A. Biguzzie, V. Blobelz, I.J. Bloodwortha, J.E. Bloomera, M. Bobinskij, P. Bock k, H.M. Bosch k, M. Boutemeur ag, B.T. Bouwens ‘, S. Braibante, R.M. Brown t, H.J. Burckhart h, C. Burgard h, R. Btirginj, P. Capiluppi b, R.K. Carnegie f, A.A. Carter m, J.R. Carter e, C.Y. Chang 4, D.G. Charlton aa2,D. Chrisman d, P.E.L. Clarke O, I. Cohen “, J.E. Conboy O, O.C. Cooke P, M. Cuffiani b, S. Dado ‘, C. Dallapiccola 9, G.M. Dallavalle b, S. De Jonge, L.A. de1 Pozo d, K. Desch ‘, M.S. Dixitg, E. do Couto e Silval, M. Doucet r, E. Duchovni Y, G. Duckeck as, I.P. Duerdothr, D. Eatoughr, J.E.G. Edwards P, PG. Estabrooks f, H.G. Evans i, M. Evans”, F. Fabbri b, M. Fanti b, P. Fath k, A.A. Faust ac, F. Fiedler ‘, M. Fierro b, H.M. Fischer ‘, R. Folman Y, D.G. Fong q, M. Foucher 9, A. Ftirtjes h, P. Gagnon s, J.W. Gary d, J. Gasconr, S .M. Gascon-S hotkin 4, N.I. Geddes t, C. Geich-Gimbel’, T. Geralis t, G. Giacomelli b, P. Giacomelli d, R. Giacomelli b, V. Gibsone, W.R. Gibson”, D.M. Gingrichac,‘, D. Glenzinski i, J. Goldberg”, M.J. Goodrick e, W. Gorn d, C. Grandi b, E. Gross Y, J. Grunhaus “, M. Gruwe h, C. Hajdu Be, G.G. Hanson e, M. Hansroul h, M. Hapke m, C.K. Hargrove g, PA. Hart i, C. Hartmann ‘, M. Hauschild h, C.M. Hawkese, R. Hawkings’, R.J. Hemingway’, M. Herndonq, G. Hertenj, R.D. Heuerh, M.D. Hildrethh, J.C. Hille, S.J. Hilliera, T. Hilsej, P.R. HobsonX, R.J. Homer a, A.K. Honma =*‘, D. Horvath ae*3,R. Howard ab, R.E. Hughes-Jones P, D.E. Hutchcroft e, P. Igo-Kemenes k, D.C. Imrie ‘, M.R. Ingramr, K. Ishii w, A. Jawahery ‘J, P.W. Jeffreyst, H. Jeremie’, M. Jimack a, A. Joly r, C.R. Jones e, G. Jones r, M. Jones f, R.W.L. Jones h, U. Jost k, P. Jovanovic a, T.R. Junk h, D. Karlen f, K. Kawagoe w, T. Kawamoto w, R.K. Keeler aa, R.G. Kellogg 9, B.W. Kennedy t, J. Kirk ab, S. Kluth h, T. Kobayashi w, M. Kobelj, D.S. Koetke f, T.P. KokottC, M. Kolrepj, S. Komamiya w, T. Kress k, P. Krieger f, J. von Krogh k, P. Kyberd m, G.D. Lafferty P, R. Lahmann 9, W.P. Lai ‘, D. Lanske n, J. Lauber O, S.R. Lautenschlager ad, J.G. Layter d, D. Lazic u, 0370-2693/97/$17.00 0 1997 Published PI1 SO370-2693(97)00162-7

by Elsevier Science B.V. All rights reserved.

148

OPAL Colluboration/Physics

Letters B 397 (1997) 147-157

A.M. Leead, E. Lefebvre r, D. Lellouch Y, J. Letts e, L. Levinson Y, C. Lewis O, S.L. Lloyd m, F.K. Loebinger P, G.D. Long aa, M.J. Losty s,J. Ludwigj, A. Macchiolo b, A. MacphersonaC, M. Mannelli h, S. Marcellini b, C. Markus ‘, A.J. Martin m, J.P. Martin r, G. Martinez 9, T. Mashimo w, W. Matthews ‘, P. MattigC, W.J. McDonaldaC, J. McKennaab, E.A. Mckigney O, T.J. McMahon a, A.I. McNab m, R.A. McPherson h, F. Meijers h, S. Menke ‘, ES. Merritt i, H. Mes s, J. Meyer ‘, A. Michelini b, G. Mikenberg y, D.J. Miller O, R. Mir Y,W. Mohrj, A. Montanari b, T. Mori w, M. Morii w, U. Mtiller c, K. Nagai Y, I. Nakamura”, H.A. Neal h, B. Nellen ‘, B. Nijjharp, R. Nisius h, S.W. O’Neale”, F.G. Oakhams, F. Odorici b, H.O. Ogren ‘, N.J. Oldershaw P, T. Omori w, M.J. Oreglia i, S. Orito w, J. Palinkas afp4,G. Pasztor ae, J.R. Paterr, G.N. Patrick t, J. Pattj, M.J. Pearce a, S. Petzold”, P. Pfeifenschneider n, J.E. Pilcher’, J. PinfoldaC, D.E. Plane h, P. Poffenberger”, B. Poli b, A. Posthaus ‘, H. PrzysiezniakaC, D.L. Rees a, D. Rigby a, S. Robertson aa, S .A. Robins m, N. Rodning ac, J.M. Roney aa, A. Rooke O, E. Ros h, A.M. Rossi b, M. Rosvick=, P. Routenburg ac, Y. Rozen ‘, K. Rungei, 0. Runolfsson h, U. Ruppel n, D.R. Rust e, R. Rylko”, K. Sachsj, E.K.G. Sarkisyan”, M. Sasaki”, C. Sbarraab, A.D. Schaile as, 0. Schaileas, F. Scharfc, P. Scharff-Hansenh, P. Schenk”s, B. Schmitth, S. Schmittk, M. Schriiderh, H.C. Schultz-Coulonj, M. Schulz h, M. Schumacher ‘, h, P. SchtitzC, W.G. Scottt, T.G. Shears P, B.C. Shen d, C.H. Shepherd-Themistocleous P. Sherwood’, G.P. Siroli b, A. Sittler ‘, A. Skillman’, A. Skujaq, A.M. Smithh, T.J. Smithaa, G.A. Snowq, R. Sobieaa, S. Siildner-Remboldj, R.W. Springer ac, M. Sproston t, A. Stahlc, M. Steiertk, K. Stephensr, J. Steuerer ‘, B. StockhausenC, D. StromS, P. Szymanski t, R. Tafirout r, S.D. Talbot a, S. Tanaka w, P. Tar-as r, S. Tarem “, M. Thiergenj, M.A. Thomson h, E. von T6rneC, S. Towersf, I. Trigger r, T. Tsukamoto w, E. Tsur “, A.S. Turcot i, M.F. Turner-Watson h, P. Utzat k, R. Van Kooten”, M. Verzocchij, P. Vikas r, M. Vincter aa, E.H. Vokurkar, F. Wackerlej, A. Wagner ‘, C.P. Warde, D.R. Ward e, J.J. Ward O, P.M. Watkins a, A.T. Watson a, N.K. Watson a, P.S. Wells h, N. Wermes ‘, J.S. White aa, B. Wilkens j, G.W. Wilson ‘, J.A. Wilson a, G. Wolf Y, S. Wotton e, T.R. Wyatt P, S. Yamashita w, G. Yekutieli Y, V. Zacek r, D. Zer-Zion h a School of Physics and Space Research, University of Birmingham, Birmingham B1.5 271: UK b Dipartimento di Fisica dell’ Universita di Bologna and INFN, I-40126 Bologna, Italy c Physikalisches Institut, Universitiit Bonn, D-53115 Bonn, Germany e Department of Physics, University of California, Riverside CA 92.521, USA e Cavendish Laboratory, Cambridge CB3 OHE, UK f Ottawa-Carleton Institute for Physics, Department of Physics, Carleton University, Ottawa, Ontario KIS 5B6, Canada g Centre for Research in Particle Physics, Carleton University, Ottawa, Ontario KIS 5B6, Canada h CERN, European Organisation for Particle Physics, CH-1211 Geneva 23, Switzerland i Enrico Fermi Institute and Department of Physics, University of Chicago, Chicago IL 60637, USA j Fakultiitfiir Physik, Albert Ludwigs Universitiit, D-79104 Freiburg, Germany k Physikalisches Institut, Universitiit Heidelberg, D-69120 Heidelberg, Germany e Indiana University, Department of Physics, Swain Hall West I 17, Bloomington IN 47405, USA m Queen Mary and Westfield College, University of London, London EI 4NS, UK n Technische Hochschule Aachen, III Physikalisches Institut, Sommet$eldstrasse 26-28, D-52056 Aachen, Germany O University College London, London WCIE 6BT UK v Department of Physics, Schuster Laboratory, The University, Manchester Ml3 9PL, UK 9 Department of Physics, University of Maryland, College Park, MD 20742, USA

OPAL Collaboration/Physics

Letters B 397 (1997) 147-157

149

’ Laboratoire de Physique Nucleaire, Vniversite’ de Montreal, Montreal, Quebec H3C 3J7, Canada ’ University of Oregon, Department of Physics, Eugene OR 97403, USA t Rutheflord Appleton Laboratory, Chilton, Didcot, Oeordshire OX11 OQX, UK ’ Department of Physics, Technion-Israel Institute of Technology, Hatfa 32000, Israel ’ Department of Physics and Astronomy, Tel Aviv University, Tel Aviv 69978, Israel w International Centre for Elementary Particle Physics and Department of Physics, University of Tokyo, Tokyo 113, Japan and Kobe University, Kobe 657, Japan ’ Brunei University, Vxbridge, Middlesex VB8 3PH, UK Y Particle Physics Department, Weizmann Institute of Science, Rehovot 76100, Israel ’ Vniversitiit Hamburg/DESI: II Institut fir Experimental Physik, Notkestrasse 85, D-22607 Hamburg, Germany aa University of Victoria, Department of Physics, P.O. Box 3055, Victoria BC V8W 3P6, Canada a’ University of British Columbia, Department of Physics, Vancouver BC V6T 121, Canada ac University of Alberta, Department of Physics, Edmonton AB T6G 2J1, Canada ‘e Duke University, Department of Physics, Durham, NC 27708-0305, USA ae Research Institute for Particle and Nuclear Physics, H-1525 Budapest, PO. Box 49, Hungary af Institute of Nuclear Research, H-4001 Debrecen, P.O. Box 51, Hungary as Ludwigs-Maximilians-Vniversitiit Milnchen, Sektion Physik, Am Coulombwall 1, D-85748 Garching, Germany

Received 10 January 1997 Communicated by K. Winter

Abstract This letter describes a measurement of one of the anomalous triple gauge boson couplings using the first data recorded by the OPAL detector at LEP2. A total of 28 W-pan candidates have been selected for an integrated luminosity of 9.89+0.06 pbb’ recorded at a centre-of-mass energy of 161 GeV. We use these data to place constraints upon the coupling parameter c~wg. We analyse the predicted variation of the total cross-section for all observed channels and the distribution of kinematic variables in the semileptonic decay channels. We measure crw$ to be -0.61~$od.76:ZII0.35, which is consistent with the Standard Model expectation of zero. @ 1997 Published by Elsevier Science B.V.

1. Introduction In the initial phase of operation of LEP2, centre-ofmass energies of fi = 161 GeV have been attained, allowing the production of W+W- boson pairs for the first time in efe- collisions. The W+W- production process involves the triple gauge boson vertices between the W+W- and the Z” or photon. The measurement of these triple gauge boson couplings (TGCS) and the search for possible anomalous values is one of the principal physics goals at LEP2. The observable effects of such deviations have been studied extensively [ 11. Anomalous TGCS can affect both the total production cross-section and the shape of the differ-

1 And ’ And 3 And 4 And versity,

at TRIUMF, Vancouver, Canada V6T 2A3. Royal Society University Research Fellow. Institute of Nuclear Research, Debrecen, Hungary. Department of Experimental Physics, Lajos Kossuth UniDebrecen, Hungary.

ential cross-section as a function of the W- production angle. The relative contributions of each helicity state of the W bosons are also changed, which in turn affects the distributions of their decay products. Anomalous TGCS are associated with contributions to W*W- production that rise with fi and which would lead to violation of unitarity. Measurements at fi = 161 GeV are therefore somewhat less sensitive to TGCS than the measurements which will be possible at higher energies. A further effect at the W+Wthreshold is that the s-channel triple gauge boson diagrams are suppressed at tree level relative to the dominant t-channel neutrino exchange diagram by a factor of p2 = 1 - 4Mw2/s. However due to the width of the W the sensitivity at threshold is enhanced relative to this simple expectation. The first reason for this is that the W bosons can be off-shell and their average p is increased. This alone leads to a significant enhancement of the s-channel diagrams. Secondly the increase

150

OPAL Colhhorution/Physics

in ,6 results in an increase in the average W momentum and the measurement of this quantity is sensitive t0 ~OIIEhUS TGCS. In this letter we use the first sample of efe+ W+W- pair threshold production data accumulated by the OPAL detector. We use these data to place constraints upon one of the models adopted during the LEP2 workshop [ 11, the so-called W4 model. We present an analysis of the total production crosssection measured using all observed events and of the W momentum spectrum and angular distributions measured using the semileptonic decay channels.

1.1. The W4 model

The most general Lorentz invariant Lagrangian [ l] which describes the triple gauge boson interaction has up to fourteen independent terms, seven describing the WWy vertex and seven describing the WWZ vertex. This parameter space is very large, and it is not currently possible to measure all parameters independently. Assuming electromagnetic gauge invariance and C and P conservation this parameter set reduces to five, which we can write as &, K~, K~, A, and h, following the notation used in [ 1,2]. In the Standard Model & = K, = ICY = 1 and & = A, = 0. Triple gauge boson couplings contribute through virtual corrections to many observables measured at lower energies and the precise measurements at LEPl allow constraints to be placed upon the parameter space of these five quantities [ 31. As a result three specific linear combinations of these couplings have been proposed [ 1,2,4] which are not tightly constrained by LEPl data. These are: aa4 =

liK,

-

A&cos2 8,

cuw+ E A&cos2 8, cxw E A, with the constraints that AK~ = -Atc,tan2 8, + h& and h, = h,. The A indicates the deviation of the respective quantity from unity and 0, is the weak mixing angle. In this analysis we measure the aw+ coupling assuming and and cuw to be zero. This means that we allow only AK,, AK~ and A& to vary from zero under

Letters B 397 (1997) 147-157

the constraints that AgTcos2 13~= AK, and that AK, is as given above.

2. Data selection and simulated

event samples

The data were recorded during 1996 at fi = 161.3 f 0.2 GeV [ 51 using the OPAL detector, which is fully described elsewhere [ 6,7]. A total integrated luminosity of 9.89 pb-’ was recorded at this energy. The selection and reconstruction of W+Wevents produced by the reaction e+e- --f W+W- has been described in [ 81, which includes a description of lepton identification and jet finding. Events are selected based upon five different W pair decay combinations. -These are: qqqq events where both W bosons decay to a two-jet final state, semileptonic q+‘%t events where one W decays to a two-jet final state, and the other W decays to an electron, muon or tau, plus a neutrino and C-‘iieC’+~et events where both W bosons decay to a lepton plus a neutrino. A total of 28 events were selected with an expected background of 5.0 * 0.6 events. For the analysis of the kinematic distributions in the qq@t channels we have augmented the selections of [ 81 in order to reduce further the contamination from single Z” --+ qq or y -+ qq events. In the selection of [ 81 a single high momentum lepton and exactly two hadronic jets are required. The jets are formed from the tracks and calorimeter clusters of the hadronic system using the kl (“Durham”) [ 93 algorithm, and the total energy and momentum of each of the jets are corrected for double counting of energy [ lo]. The missing energy and momentum of the event are assigned to the neutrino. In this analysis we add the following re-quirements. In the qqeFe and qqpv, channels we apply a two dimensional cut in the plane of the invariant masses of the q?i and fie systems, Ml,, and Mqs. The selection requires that MS4 < 30 GeV + 1.65 Ml, and -Mqq > 135 GeV - Ml,. In the qqrv7 channel the cut on the cosine of the angle between the missing momentum vector and the beam direction is tightened to 1cos 6’miss 1 < 0.8 and two further cuts are applied. The invariant mass of the hadronic system, M,q is required to satisfy 40 GeV < Mqq -c 90 GeVand the cosine of the angle between the r jet direction and the missing momentum vector is required to be less than zero.

OPAL Collaboration/Physics Table 1 The selected number of events in the lepton channels after the additional cuts described in the text. The Standard Model expectation for the numbers of signal and background events are also shown Channel

q@K SW/l SWIG

Efficiency (%I

Selected events

70f3 73f3 29f3

3 2 6

Predicted signal

background

3.8 f 0.4 4.0 zt 0.5 1.6 zt 0.2

0.2 f 0.3 0.3 f 0.2 0.5 f 0.2

Details of the resulting signal, efficiencies, and background5 contaminations are shown in Tables 1 and 2. These are obtained using the same procedures as detailed in [8]. In comparison to the selections of [8] the Za/y -+ qq background is reduced from 5.7% to 1.0% and 1.8% to 0.8% in thee and p channels, respectively. The number of events selected from -the data does not change. In the qqrv, channel the background contamination from Z”/r -+ qq events is reduced from 22% to 10% and the number of events selected is reduced from seven to six. We use Monte Carlo samples to provide predictions of quantities as a function of “~4. The principal Standard Model sample is obtained using PYTHIA [ 111 and is generated using the current world average W boson mass of A4w = 80.33 f 0.15 GeV [ 121. The W+Wcross-section for this sample is within 2% of that predicted by the GENTLE program [ 131. To obtain samples corresponding to different anomalous values of aw4 we use EXCALIBUR [ 141 generating W pair diagrams only. Estimates of the different background processes are based primarily on the PYTHIA, EXCALIBUR and grc4f [ 151 generators. KORALW [ 161 is used in the estimation of systematic uncertainties. All events are passed through the full OPAL simulation program [ 171 and then subjected to the same reconstruction procedures as applied to the data.

5 In Table 2 the cross contamination between q?@f channels is shown. In particular there is a 16% contamination from qqeFe and q@FP events to the q@Fr channel. However it should be noted that no attempt has been made to try to reject these in the selection procedure. This is because due to our method of r reconstruction the e and fi contamination to r’s is effectively considered as signal.

Letters B 397 (1997) 147-157

151

-3. Kinematic variables for the qq&e event sample The qqfie events are the most straightforward to reconstruct since there is no ambiguity in assigning decay fermion pairs to each W nor in determining the -charges of each W. In contrast the qqqq channel has both of these problems and a higher background, leading to a much reduced sensitivity to TGCS. Therefore in this analysis we analyse the kinematic distributions only in the qi@l channels. For each q@e event we measure four variables: (1) cos0w, the production angle of the W- with respect to the e- beam direction, (2) pw, the momentum of the hadronically decaying W, cos t9J, the polar decay angle of the charged lepton with respect to the W flight direction measured in the W rest frame and (4) +i, the azimuthal decay angle of the charged lepton with respect to a plane defined by the W and the beam axis. The exact definition of the decay angles is described in [ 1,2,18]. We do not distinguish the charges of the quarks from the hadronically decaying W. As a result the decay angles of the quarks can only be determined with a two fold ambiguity which substantially reduces their sensitivity to TGCS. We therefore do not use these angles in this analysis. (3)

3.1. qqeve and q?lpijp events --In the case of qqev, and qqpvp events we use variables resulting from a kinematic fit. This fit demands energy and momentum conservation and a zero mass for the missing neutrino, resulting in one constraint. As input to the fit we use the electron momentum constructed from the direction measured by the tracking detectors and the energy measured in the electromagnetic calorimeters (except for an overlap region between the barrel and endcap in which case the measured track momentum is used instead). In the case of muons we use the momentum measured using the tracking detectors. We neglect initial state radiation when applying energy and momentum conservation in the reconstruction. However in the analysis we compare the data with Monte Carlo distributions which are subject to exactly the same reconstruction procedure and which include ISR in the simulation. Therefore

152

OPAL Colluboration/

Table 2 Details of the background

contaminations

Background

in the qqe&

contamination

%

ZY

w+w---t

q$$

1.0 0.8 10.0

3.1 4.5 16a

a The cross contamination considered as background.

from other qqfie

Physics Letters B 397 (1997) 147-157

q-QCN and q+irr

samples.

Other W’W-

zz.

Wev

Zee

Total

0.1 0.1 0.4

0.4 1.0 0.9

0.3 0.0 1.5

0.4 0.0 3.5

5.3 6.4 32

decays to the qTj7’li7channel is shown for information,

but as explained

in the text this is not

OPAL simulation mO.18 .Z SO.16 .* 0.14 n 2 0.12

20.08

0.1

0.06

z

0.08 0.06 0.04 0.02 0

0 -1

-0.5

0

0

0.5 6 ( cos ew :

6Pw)

KeVI

Y ‘3 0.25

.SO 16 T * io.14

b a 0.2 2

5012 t * 0.1 0.08 0.06 0.04 0.02 0

-1

-0.5

0

0.5 1 G(eos0;)

0

Fig. 1. The resolution of the kinematic variables from the q?@t events. All distributions show the difference -generated quantities. The solid line is for qije& and qwvP events and the dashed line is for q@rFr events.

the method compares like with like and should introduce no bias. This has been explicitly checked and at the present level of statistical accuracy any residual bias has been found to be negligible. We obtain cos 8w and pw by adding together the fitted four-momenta of the two jets. The charges of the W bosons are determined by the sign of the charged

between

reconstructed

and

lepton. The lepton decay angles are obtained by boosting the fitted lepton four-vectors to the parent W rest frame, and using the definitions given in the references above. The predicted resolution of each of these quantities obtained from Monte Carlo events is shown in Fig. 1.

OPAL Colluboration/Physics

2

Letters B 397 (1997) 147-157

153

5

. v1

2:!

4

W %

. 3 i? 2 1 0

-1

-0.5

0

0.5

1

1 0

0

20

40

60

pw

1 -1

-0.5

0

0.5 1 cos

0

e;

Fig. 2. The distributions of cos Ow, pw, cos 0: and 4; obtained from the eleven q?jfit events selected from the data. The hatched histogram shows the non qqfit background. These are compared with the distribution expected in the Standard Model using fully simulated Monte Carlo events. The predicted distributions for ow$ = +2(-2) are also shown as dotted (dashed) lines. Note: in the case of W+ + iv events the value of d* is shifted by rr in order to overlay distributions in the same plot.

3.2. ~@i7~ events -The qqrvr selection procedure results in three jet events, where one jet has been identified as a tau jet containing one or three tracks. From Monte Carlo studies this is found to make the correct selection of the tau in more than 97% of cases. We obtain cos Bw and pw by adding together the measured four-momenta of the two jets. The charges of the W bosons are determined by the sign of the sum of the charges of the tracks in the r jet. This yields the correct charge in 98% of the events where the Q-has been correctly identified. In order to reconstruct the decay angles, the flight direction of the 7 is approximated by the direction of its jet. The four unknown quantities can then be calculated using energy and momentum conservation. These are the energy of the r and the three-momentum

of the r neutrino originating directly from the W decay. The decay angles are then obtained as for the other qqfi[ events. The predicted resolutions for these quantities do not differ significantly from the resolutions obtained in the other q@e channels and are shown in Fig. 1. The reason for this is that reconstruction of all variables depends mainly upon the hadronic system. In Fig. 2 we show the distributions of all quantities -obtained from the 11 qqeFe, qqpuv, and qqrFT events added together. They are compared with the distributions expected in the Standard Model using fully simulated Monte Carlo events.

4. LYW~analysis

In this section we analyse the W+W- event sample in order to place limits upon crw4. We do this in two parts. In the first part we calculate the log likeli-

154

OPAL Collaborution/

Physics Letters B 397 (1997) 147-157

hood (log C) for observing the 28 ‘events in our data sample as a function of aw+. In the second part we calculate the IogC for the 11 qq&t events to have their measured distribution of kinematic variables, as a function of aw4. These likelihoods are independent and we add them to give the overall log C distribution from which we derive the results. 4.1. Analysis of the total number of observed events In this part of the analysis all channels are combined. The WIWproduction cross-section varies with cow+, having a minimum near the Standard Model value of zero and increasing parabolically for LYW~# 0. For each assumed value of “~4 we calculate the likelihood for the expected mean number of events to have resulted in our observed number of 28 events. The expected mean number of signal events as a function of aw+ is obtained by multiplying the expected cross-section by the recorded integrated luminosity value of 9.89f0.06 pb-‘. To calculate the expected cross-section we take advantage of the fact that this is a quadratic function of any TGC parameter. Five EXCALIBUR Monte Carlo samples with ~yw+ = -2, -1, 0, I, 2 were passed through the same selection requirements as the real data and the accepted cross-section found for all channels combined. All five points were found to lie on a parabola and this was parameterised and used to predict the crosssection for all other values of “~4. We use the W mass value of 80.33%0.15 GeV [ 121 which excludes the recent LEP results. The LEP beam energy is taken to be 80.65f0.10 GeV [ 51. The EXCALIBUR crosssections are multiplied by a factor of 1.026f0.026 to account for the slight difference [ 81 between EXCALIBUR and GENTLE, as the latter is more complete [ 191. The probability of finding the observed number of events was calculated assuming a Poisson distribution for the signal and background. The negative log L: distribution obtained from this procedure is shown in Fig. 3. 4.2. Analysis

of the differential

distributions

In the second part of this analysis a binned likelihood method is used to analyse the cos 0w, pw, cos 0; and 4; distributions. As discussed in the introduc-

Fig. 3. Likelihood distributions obtained from the cross-section (dotted) and differential distributions (dashed). The solid line is the distribution obtained by adding these together. In all cases the minimum value of the negative log likelihood has been subtracted.

tion each of these variables depends upon cow+. This is illustrated in Fig. 2, which shows the distributions of each, integrated over the other three variables, for our data sample and three Monte Carlo samples with cYw4 = -2,0,2. In this analysis pw is divided into three bins in the range [ 0,451 GeV, cos 8w and cos 0; are divided into five bins in the range [ - 1,l ] and 4; is divided into five bins in the range [--Z-, n] , giving a total of 375 bins which we label i. The log.fZ distribution is obtained in several steps. In the first step we parameterise in each bin the expected cross-section, before detector and acceptance effects, as a function of aw4. We denote this as e(aw6). It is obtained using large samples of Monte Carlo events without detector simulation. In each bin @-“(a~&) is parameterised using the quadratic dependence upon “~4, In the second step we calculate correction factors to include the effects of acceptance, resolution and the contribution of other W+W- decay channels to the sample. We use fully simulated Monte Carlo events generated with “~4 = 0 to obtain the factors, cki, which allow for events generated in true bin i being reconstructed in bin k. All WiWdecays are included

OPAL Collaboration/Physics

in the sample and all events which are reconstructed as qqfit events in bin k are counted in the correction factors cki. For a given bin, corresponding to a limited phase space region, these effects do not depend strongly on cuw4. The cki factors are therefore assumed to be independent of “~4. It has been explicitly checked that no significant bias is introduced by this approximation. Combining these terms results in the expected observed cross-section for each bin k, due to all W+Wchannels:

Letters B 397 (1997) 147-157

155

We have also checked these results using two independent analyses. In the first method an unbinned likelihood analysis was developed to analyse only dg/dcos Bw. In the second method the entire analysis was formulated using optimal observables [ 1,211. The measured optimal observable corresponding to ~yw+ was compared with the predicted value as a function of “~4 and the likelihood calculated. In all cases the log_C distributions were found to be consistent with the binned method applied to the corresponding observables.

4.3. Combined analysis In the next step the cross-section for non-W+Wbackground sources, u?, is estimated from the corresponding Monte Carlo sam les. The variation with % ~uw4 of the contribution to ffkkg from Wev events is considered in the systematic checks. The background cross section is added to give ,fr (~yw+ ) :

This expression now contains all of the information on the shape of the expected distribution in the fourdimensional space, as a function of “~4. We multiply this by a scale factor to give the prediction for the number of events in each bin, nk, where the scale factor is chosen such that the predicted total number of events is equal to the actual number of events observed in the data. This is to ensure that we do not incorporate any overall cross-section information in this part of the analysis. Finally the probability for observing the number of events seen in each bin for an expectation of nk is calculated using Poisson statistics. The statistical fluctuusing the ations in cki and ukbkg are taken into account method of Ref. [ 201. The negative log L distribution for the data set is shown in Fig. 3. In order to check the possible bias in the method, it was tested using different samples of Monte Carlo events taken as data events. These samples with values of cuwb ranging between -2 and 2 were passed through the full detector simulation. Each sample had 1500 events after selection cuts. The resulting fitted values for cuw4 were always consistent with the input values.

The individual 1ogfZ distributions shown in Fig. point. 3 are compatible with the same minimum These are independent measurements and therefore have been added together to give the overall 1ogL distribution shown as the solid line in the figure. We can express this as a measurement with the one standard deviation limits given by the values of LXW~ where A 1ogfZ = 0.5. This yields the result cuw4 = -0.61?:,$, where only the statistical error is included. The expected error for a sample of this size was evaluated by repeating the analysis using Monte Carlo events divided into sub-samples equivalent to the actual data sample. We find an expected statistical error of f0.68. To express this as a 95% confidence level (CL.) limit we take the points of intersection of the curve with the A 1ogL = 1.92 line. This yields the result: - 1.74 < CYW~< 0.94 at 95% C.L., where again only the statistical error is included in this range. It should be noted that the 1ogC distribution is not parabolic and therefore different C.L. ranges cannot be obtained simply from the one standard deviation values, but have to be determined directly from the distribution.

5. Systematic studies The Monte Carlo simulation of the measured quantities depends mainly upon the simulation of the jets from the W hadronic decay. Jet reconstruction has been studied and tuned extensively at LEPl, showing good agreement between distributions measured from data and Monte Carlo samples. We therefore expect

156

OPAL Collaboration/Physics

this to be adequate for the small number of events in this sample. As a direct test of cos 0w reconstruction we have used radiative Z events taken from the fi = 91 GeV data. By selecting events containing observed radiated photons with energies up to 20 GeV we obtain a sample of jet pairs exhibiting a similar acollinearity distribution to W decays. Assuming that no other photons have been radiated in the event the true direction of the Z” -+ qq system is opposite to that of the photon. The difference between the value of cos0 measured from the photon and that measured from the hadronic system is therefore strongly related to the true resolution of cos Bw. The results obtained from both data and Monte Carlo events agree well showing no significant differences in shape or width of the distributions. In order to quantify reasonable variations of jet reconstruction parameters we have used a sample of radiative Z/y events selected from the fi = 161 GeV data by requiring acollinear two jet events. The invariant mass distribution of the two jets is compared between data and Monte Carlo and a x2 is formed for the match. The jet energy scale and resolution for the Monte Carlo events are then artificially varied in order to minimise the x2. The best fit variations were then used to vary the Monte Carlo jet reconstruction in the aw+ analysis and the resulting changes caused to &w+ were added in quadrature and taken as a systematic error (0.16). We have evaluated the systematic errors due to Monte Carlo statistics (0.04), the uncertainty in the measured W mass and the LEP beam energy (0.24) and the uncertainties of the overall selection efficiency, background and luminosity measurements (0.06). We have varied the Monte Carlo generator to use both EXCALIBUR and KORALW to generate the Standard Model sample, resulting in a change of 0.10 to ~yw+. Since the shape of the distributions is important, regions where the efficiency changes rapidly were excluded and the analysis repeated. This procedure removed at most two events and caused a change of 0.15 to c~w4. We have studied the effect of both doubling the background and setting it to zero (0.02). We have also considered the variation with cuw4 of the background due to Wev events. The accepted cross-section for this process increases by a factor of approximately six for (a~4 / N 2. Allowing for this in the analysis caused a negligible change to

Letters B 397 (1997) 147-157

Table 3 Contributions to the error on a~+ Source

Error

Jet reconstruction MC statistics W mass and LEP beam energy Overall normalisation MC generator Acceptance variation Background

0.16 0.04 0.24 0.06 0.10 0.15 0.02

Total systematics Data statistics

0.35 +0.73 -0.61

aw4. The error due to all sources is listed in Table 3. To obtain the final 95% C.L. limits we have performed a similar procedure. The shift in the 95% upper and lower limits were obtained for each of the variations described above. These shifts were added in quadrature to obtain overall errors for both limits. The upper limit error was added linearly and the lower limit error was subtracted linearly to obtain the final limits.

6. Result and conclusion Adding the total systematic obtain the result a!w4 = -0.61?;,;7,:

we

f 0.35

and the corresponding -2.1

errors in quadrature

95% C.L. is

< crw4 < 1.6

Other experiments at the Tevatron, CDF [ 221 and DO [ 231, have previously reported limits on anomalous TGCS involving the WWZ coupling. However the constraints used in those publications are different to those of this analysis and the results are therefore not directly comparable. To allow an approximate comparison we have repeated our analysis using their sets of constraints but without detector simulation. We find statistical errors from this analysis which are approximately a factor of two greater than the CDF and DO results. To give an indication for the implication of this limit we use the unitarity relationship given in [ 1] /cuw4/ -

OPAL Collaboration/Physics

1~5.5(Mw/R~)~ where Au is the scale at which unitarity would be violated. Assuming this approximate relation our 95% C.L. upper limit corresponds to Au 2 220 GeV from this measurement alone.

Acknowledgements We particularly wish to thank the SL Division for the efficient operation of the LEP accelerator at the new energy of @ = 161 GeV and for their continuing close cooperation with our experimental group. We thank our colleagues from CEA, DAPNIA/SPP, CE-Saclay for their efforts over the years on the timeof-flight and trigger systems which we continue to use. In addition to the support staff at our own institutions we are pleased to acknowledge the Department of Energy, USA, National Science Foundation, USA, Particle Physics and Astronomy Research Council, UK Natural Sciences and Engineering Research Council, Canada, Israel Science Foundation, administered by the Israel Academy of Science and Humanities, Minerva Gesellschaft, Japanese Ministry of Education, Science and Culture (the Monbusho) and a grant under the Monbusho International Science Research Program, German Israeli Bi-national Science Foundation (GIF), Bundesministerium fiir Bildung, Wissenschaft, Forschung und Technologie, Germany,. National Research Council of Canada, Hungarian Foundation for Scientific Research, OTKA T-016660, and OTKA F-015089.

References [II Physics at LEPi!, Edited by G. Altarelli, T. Sjostrand and E Zwimer, Report on the LEP2 workshop 1995, CERN 96-01 (1996) Vol. 1, p. 525. [21 M. Bilenky, J.L. Kneur, EM. Renard and D. Schildknecht, Nucl. Phys. B 409 (1993) 22. 131 A. De Rujula, M.B. Gavela, F? Hemandez and E. Masso, Nucl. Phys. B 384 (1992) 3, K. Hagiwara, S. Ishihara. R. Szalapski and D. Zeppenfeld,

Letters B 397 (1997) 147-157

[4]

[5] [6]

[7] [ 81 [91

[lOI 1111 t121

t131 [I41

[I51 [I61 1171 [I81 [I91

[201 [211 r221 [231

15-l

Phys. Lett. B 283 (1992) 353, Phys. Rev. D 48 (1993) 2182. K. Hagiwara, R.D. Peccei, D. Zeppenfeld, K. Hikasa, Nucl. Phys. B 282 (1987) 253, M. Bile&y, J.L. Kneur, EM. Renard and D. Schildknecht, Nucl. Phys. B 419 (1994) 240. LEP Energy Working Group, private communication. The OPAL Collaboration, K. Ahmet et al., Nucl. Ins& Meth. A 305 (1991) 275, PP. Allport et al., Nucl. Instr. Meth. A 324 (1993) 34; Nucl. Instr. Meth. A 346 (1994) 476. B.E. Anderson et al., IEEE Transactions on Nuclear Science, 41 (1994) 845. The OPAL Collaboration, K. Ackerstaff et al., 16th October .1996, Phys. Lett. B 389 (1996) 416. N. Brown and W.J. Stirling, Phys. Lett. B 252 ( 1990) 657; S. Bethke, 2. Kunszt, D. Soper and W.J. Stirling, Nucl. Phys. B 370 (1992) 310; S. Catani et al., Phys. Lett. B 269 (1991) 432; N. Brown and W.J. Stirling, Z.&t. Phys. C 53 (1992) 629. The OPAL Collaboration, M.Z. Akrawy et al., Phys. Lett. B 253 ( 1990) 511. T. Sjastcand, Comp. Phys. Comm. 82 (1994) 74. The Particle Data Group, R.M. Bamett et al., Phys. Rev. D 54 (1996) 1; CDF Collaboration, E Abe et al., Phys. Rev. Lett. 75 (1995) 11, Phys. Rev. D 52 (1995) 4784; Phys. Rev. Lett. 65 (1990) 2243; Phys. Rev. D 43 (1991) 2070; UA2 Collaboration, J. Alitti et al., Phys. Lett. B 276 (1992) 354; UAl Collaboration, C. Albajar et al., Zeit. Phys. C 44 (1989) 15. D. Bardin et al., Nucl. Phys. B, Proc. Suppl. B 37 (1994) 148-157. EA. Berends, R. Pittau and R. Kleiss, Comp. Phys. Comm. 85 (1995) 437, EA. Berends and A.I. van Sighem, Nucl. Phys. B 454 ( 1995) 467. J. Fujimoto et al., KEK-CP-046, hep_ph/9603394, to be published in Comp. Phys. Comm. M. Skrzypek et al., Comp. Phys. Comm. 94 (1996) 216, M. Skrzypek et al., Phys. Lett. B 372 (1996) 289. J. Allison et al., Nucl. Ins& Meth. A 317 (1992) 47. R.L. Sekulin, Phys. Lett. B 338 (1994) 369. Physics at LEP2, Edited by G. Altarelli, T. Sjostrand and F. Zwirner, Report on the LEP2 workshop 1995, CERN 96-01 (1996) Vol. 2, p. 3. R. Barlow and C. Beeston, Comp. Phys. Comm. 82 (1994) 74. M. Diehl, 0. Nachtmann, Zeit. Phys. C 62 (1994) 397. The CDF Collaboration, F. Abe et al., Phys. Rev. Lett. 75 (1995) 1017. The DO Collaboration, S. Abachi et al., Phys. Rev. L&t. 75 (1995) 1023, S. Abachi et al., Phys. Rev. Lett. 77 (1996) 3303.