Fermi National Accelerator Laboratory

FERMILAR-Pub-95/035-E

CDF

Measurement of the W Boson Mass

F. Abe et al. The CDF Collaboration

Fermi National Accelerator Laboratory P.O. Bm 500, Batavia, Illinois 60510

March 1995

Submitted to Physical Review Letters

$ Operated by Universities Research Association Inc. under Contract No. DE-AC02-76CH030W with the United States Department of Energy

This report UJDS prepared as an account of work sponsored by an agency of the United States Government. Neither the United States Government nor any agency thereof nor any of their employees, makes any warranty, express or implied, or assumes any legal liability or responsibility for the accuracy, completeness, or usefulness of any information, appamtus, product, or process disclosed, or represents that its use would not infringe privately owned rights. Reference herein to any specific commercial product, process, or service by trade name, trademark, manufacturer, or otherwise, does not necessarily constitute or imply its endorsement, recommendation, or favoring by the United States Government or any agency thereof, The views and opinions of authors expressed herein do not necessarily state or reflect those of the United States Government or any agency thereof.

CDF/PUB/ELECTROWEAK/PUBLIC, FERHILAB-PUB-95/035-1995

Measurement of the W Boson Mass

F. Abe,‘s M. G. Albrow, D. Amidei,‘s J. Antes,** C. Anway-Wiese,* G. Ap~llinari,~e H. A&i,’ M. Ataq7 P. Auchincloss,2” F. Azfar,s’ P. Azzi,20

N. Bacchetta,” W. Badgett,16 M. W. Bailey,‘* J. Bao,s’ P. de Barbaro,ss A. Barbaro-Galtieri,14 V. E. Barnes,s4 B. A. Barnett,‘2 P. Bartalini,s3 G. Ba~er,‘~

T. Baumarqg F. Bedeschi,s3 S. Behrends,s S. Belforte,*s G. Bellettini,ss J. Bellinger,3s D. Benjamin,ss J. Benlloch,l” J. Bensinger: D. Benton,sl

A. Beretvas,’ J. P. Berge,7 S. Bertolucci,s A. Bhatti,s* K. Biery,” M. Binkley,’ F. Bird,*g D. Bisello,” R. E. Blair,’ C. Blocker,sg A. Bodek,*s W. Bokhari,‘s

V. Bolognesi,23 D. Bortoletto,*4 C. Boswell,‘2 T. Boulos,i4 G. Brandenburg,g C. Bromberg,17 E. Buckley-Geer,’ H. S. Budd,ss K. Burkett,‘s G. Busetto,*’

A. Byon-Wagner,’ K. L. Byrum,’ J. Cammerata,‘* C. Campagnari,’ M. CampbelL’s A. Caner,’ W. Ctitherst4 D. Carlsmith, A. Castro,*’ Y. Ceq2’ F. Cervelli,*s

H. Y. Chao,ss J. Chapman,‘s M.-T. Cheng,zs G. Chiarelli,s T. Chikamatsu,sl C. N. Chiou,*s S. Cihangir,7 A. G. Clark,*s M. Cobal,*s M. Contreras,s J. Conway,s7 J. Cooper, 7 M. Cordelli,’ D. Crane,i J. D. Cunningham,s T. Daniels,‘s F. DeJongh,’

S. Delchampq7 S. Dell’Agnello,s3 M. Dell’Orso,ss L. Demortier,*s B. Denby,13 M. Deninno,* P. F. Derwent,ls T. Devlin,2’ M. Dickson,ss S. Donati,s3

R. B. Drucker,” A. Dunn,‘s K. Einsweiler,‘4 J. E. Elias,’ R. Ely,14 E. Engels, Jr.,*’ S. Eno,s D. Errede,” S. Errede,‘O Q. Fan,*s B. Farhat,‘s I. Fiori,* B. Flaugher,7

G. W. Foster,7 M. Frankliqg M. Frautschi,‘s J. Freemaq7 J. Friedmaq’s H. Frisch,’ A. Fry:’ T. A. Fuess,’ Y. Fuk’.‘&r3 S. Funaki,3’ G. Gru~liardi,*~ S. Galeottit3 M. Gallinaro,20 A. F. Ga&kel,24 S. Geer,’ D. W. Gerdes,” P. Giannetti,23

N. Giokaris,s6 P. Giromini,s L. Gladney,” D. Glenzinski,‘* M. Gold,” J. Gonzalez,” A. Gordon,g A. T. Goshaw,’ K. Gouliano~,*~ H. Grassmann, A. Grewal,*i

G. Grieco,23 L. Groer, s7 C. Grosso-Pilcher,s C. Haber,14 S. R. Hahn,’ R. Hamilton,g R. Handler,= R. M. HaqM K. Hara,31 B. Harral,*r R. M. Harris,’ S. A. Hauger,’

J. Hauser, C. Hawk,” J. Heinrich,sl D. Cronin-Hennessy,6 R. Hollebeek,*l L. Holloway, lo A. Hiilscher, l1 S. Hong, l6 G. Houk,*r P. Hu,*s B. T. H~tfman,~*

R. Hughes,2s P. Hurst,g J. Huston,l’ J. Huth,g J. Hylen,’ M. Incagli,s3 J. Incandela, H. 1~0,~~ H. Jensen,’ C. P. Jessop,g U. Joshi,’ R. W. Kadel,14

E. Kajfasr,“’ T. Kamoqm T. Kaneko,31 D. A. Kardelis,” H. Kasha,s4 Y. Kato,ig L. Keeble,30 R. D. Kennedy,*’ R. Kephart,’ P. Kesten,14 D. Kestenbaum,

R. M. Keup,‘O H. Keuteliaq7 F. Keyvaq4 D. H. Kim,’ H. S. Kim,” S. B. Kim,” S. H. Kim,31 Y. K. Kim,14 L. Kirsch,s P. Koehn,*s K. Kondo,3* J. Konigsberg,’ S. Kop~,~ K. Kordas, l1 W. Koska,’ E. Kovacs,“’ W. Kowald,6 M. Krasberg,”

J. KroR7 M. Kruse,s4 S. E. Kuhlmann,’ E. Kuns,*’ A. T. Laasanen,s4 N. Labanca,s3 S. Lammel, J. I. Lamouretq3 T. LeCompte,” S. Leone,*s J. D. Lewis,’ P. Limon,’

M. Lindgreq4 T. M. Liss,” N. Lockyer,*’ C. Loomis,*’ 0. Long,” M. Loreti,*’ Submitted to Physical Review Letters March 3, 1995.

1

E. H. Low,l’ J. Lu,~ D. Lucchesi,23 C. B. Lu~hini,‘~ P. Lukens,’ P. Maas,= K. Mae&ma,’ A. Maghakiaq’s P. Maksimovic,15 M. Mangano,23 J. Mansour,” M. Mariotti,*s .I. P. Marriner,’ A. Martin,“’ J. A. J. Matthews,‘* R. Mattingly,‘s

P. McIntyre,m P. Mele~e,*~ A. Menzione,ss E. Meschi,ss G. MichaiLg S. Mikamo,13 M. Miier,s R. Miller,” T. Mimashi,31 S. Miscetti,’ M. Mishina,ls H. Mitsushio,sl

S. Miyashita,31 Y. Morita,i3 S. Moulding,26 J. Mueller,27 A. Mukherjee,’ T. Muller,4 P. Musgrave,” L. F. Nakae,” I. Nakan~,~’ C. Nelson,’ D. Neuberger,*

C. Newman-Holmes,’ L. Nodulman,’ S. Ogawa,31 S. H. Oh,s K. E. Ohl,s4 R. Oishi,s’ T. Okusawa, *’ C. Pagliarone, s3 R. Paoletti,13 V. Papadimitriou,’

S. Park,’ J. Patrick,’ G. Pauletta,23 M. Paulini,14 L. Pescara,” M. D. Peters,‘* T. J. Phillips,s G. Piacentino,’ M. Pillai,2s R. Plunkett,’ L. Pondrom,ss

N. Prod&,” J. Proudfoot,’ F. Ptohos,’ G. Punzi,s3 K. Ragan,” F. Rimondi,2 L. Ristori,13 M. Roach-Belhno,s2 W. J. Robertsoqs T. Rodrigo,’ J. Romzno,s L. Rosensoq’s W. K. Sakumoto, *s D. Saltzberg,5 A. Sansoni,s V. Scarpine,s”

A. Schindler,‘* P. Scblabach,g E. E. Schmidt,’ M. P. Schmidt,3* 0. Schneider,“’ G. F. Sciacca,23 A. Scribano,23 S. Segler,’ S. Seidel,” Y. Seiya,31 G. Sganos,”

A. Sgolacchia,’ M. Shapiro,14 N. M. Shaw,24 Q. Shen,24 P. F. Shepard,” M. Shlmojima,s’ M. Shochet,s J. Siegrist,2Q A. Sii,“’ P. Sinervo,” P. Singh,22

J. Skarha,i2 K. Sliwa,32 D. A. Smith,23 F. D. Snider,12 L. Song,’ T. Song,‘” J. Spalding,’ L. Spiegel,’ P. Sphicas,” A. Spies,12 L. Stanco,2’ J. Steele,ss

A. Stefanini,23 K. St&l,” J. Strait,’ D. Stuart,’ G. Sulllvaqs K. Sumorok,15 R. L. Swartz, Jr.,” T. Takahashi,” K. Takikawa,31 F. Tartarelli,23 W. Taylor,”

P. K. Teng,28 Y. Teramoto,lg S. Tether,‘s D. Theriot,’ J. Thomas,lg T. L. Thomas,18 R. Thuq’s M. Timko,32 P. Tipton,*s A. Titov,s6 S. Tkaczyk,’

K. Tollefsoq2s A. Tollestrup,’ J. Tom&on, 24 J. F. de T~oconiz,~ J. Tseng,i2 M. Turcotte,2Q N. Turi~&~ N. Uemura, 31 F. Ukegawa, 21 G. Unal,21

S. van den Brink,12 S. Vejcik, 111,‘s R. Vidal,’ M. Vondracek,” R. G. Wagner,l R. L. Wagner,’ N. Wainer,’ R. C. Wz.lke~,~s C. H. Wang,28 G. Wang,23 J. Wang,s

M. J. Wang,2’ Q. F. Wang,2’ A. Warburton,” G. Watts,2” T. Watts,2’ R. Webb,sc C. Wendt,ss H. Wenzel,14 W. C. Wester, III,” T. We8thusing,10 A. B. Wicklund,’

E. Wicklund,’ R. Wiinson,21 H. H. Wiiam~,~’ P. Wiion,5 B. L. Winer,25 J. Wolinski,m D. Y. Wq’s X. Wu,s3 J. Wyss,” A. Yagil,’ W. Yao,i4 K. Yasu~ka,~~

Y. Ye,” G. P. Yeh,’ P. Yeh,sa M. Yin,’ J. Yoh,’ T. Yoshida,” D. Yovanovitch,’ I. Yu,~~ J. C. Yun,’ A. Zanetti,s3 F. Zetti,s3 L. Zhang,m S. Zhang,‘* W. Zhang,*l

and S. Zucchelli2

(CDF Collaboration)

' Aqmtnr Ndiond t&wator)r. Argomra, Illin& SO&99

2 111ildo Nuionde di Fish Nrclarr, Uniwernit~ of Bologna, I-JOI#S Bobgvu, Italy 3 Brand& Uni~cr~ily. W&horn, Marroehr~rtt~ O%?Sl

4 Umi*eraity of Cdifomia a* Lo, AnpIer, Lo, An&*, Coliforrh 5%%?4

2

’ Usiwerrity of Chicago,, CA&v, Rlinoia b0bS7

6 D& Univrrsilv, Durham, Nor& Car&no L7708

’ Fermi NalionJ Aurlenlor Lsbnlo~, B~fati~, Rlinoir 60510

a Lsbnlori Ntiodi di &sari, Ialilrlo Nasion& di Fiaiu Nuclure, LOOOld Fmscoti, Raly

’ Haned Univerrilv, Cambridge. Mormcharrttr OS138

lo Uni*rrril~ of Ill&*. Urbsna, Rlino* 81801 11 Imstiirlc of Particle PApica, McGill Umiwerait~, Afontmzl HSA 8T8, end Univerril~ of Tomnfo,

Tomdo MS 1.47, Cm.d.

l2 The JoAnr Hopbinr Usiverrit~, Baltimore, Ma4snd llll8

I3 Notions .Ea&mlo~ for RigA Ensrpl Phpica (KEK), TnArbs, Ibwabi 305. Jq-m

” Lawena Bwbrlc~ Lakmloq, BcrArlev, Colifonrio 947SO

I5 Masmdurlta halit~lc of Trdnologl, Cambridge, Mnrmchr~cll~ 06139

” Umi~er~il~ of Afichigar, Rm Allor, MicAigon VI09

” AficAigm Slate Umirrr~ilr, Rut LawinS, MirAigan @SW

‘~3 U&.r& of New Aferico, Al)rq.rrqs~ New Mmico 87131

I9 Ode City hi~cr~it~, O&a SSS, Jem

” Umieartita di Padova, Ietitrto Nasio~& di Fisk Nrc*orr, Sckonr di Podova, l-35151 Pado. It&

2’ Umi~rrail~ of Penrylvamiq PhiiadelpAia, Penrr~hosia 19101

22 Umiwrsit~ of Pitbbw&, Pittsbw& Pmm+ania IdtbO 23 Istifrfo Noliode di Fiaics N&ewe, Umiwrreil~ and Scrod. Norm& Srpwiorr of Piq 1-58100 Piss, Ital”

24 Prvdme Umiversil~, We,: Lojoycllr, Indimm d7907

25 Umiwcrd# of RocAuler, RocAaalrr, New Ywb t&S17

26 Rcc&cfellcr u nwaril~, New Yort, NW YorL 10011

27 Rrtgerr hivwrity, Piaularq, Ner Jsrrr~ 08.9Sd

2s Acsdemti Simiu, Takes 11569, Republic of CA&u

2Q Srpwcdrclimg Super Collider Lsbodo~, D&u, Trraa 75137

30 Tau ABY hieerritv, CoRqc Station, Tcru 77.4/S 31 lJti.erd, of Tdda, Tdnba, IbsrsLi 505, JV.WI

32TU/U”. mrdt,, hf&d, M~,~cAucl~ OS155

33 Uniwrd~ of Wiacouim, Mdi*a, Wieconrim 55706

34 Yale u *wewits, New Ihrm, Conncdicrl 06511

3

Abstract

We present a measurement of the mass of the W boson using data collected with the CDF detector during the 1992-93 collider run at the Fermilab Tevatron. A fit to the transverse mass spectrum of a sample of 3268 W -t pv events recorded in an integrated luminosity of 19.7 pb-’ gives a mass df& = 80.310 f 0.205 (stat.) f 0.130 (syst.) GeV/c2. A fit to the transverse mass spectrum of a sample of 5716 W + CY events recorded in 18.2 pb-’ gives a mass M& = 80.490f0.145 (stat.)&0.175 (syst.) GeV/c’. Combining the electron and muon results, accounting for correlated uncertainties, yields a mass MW = 80.410 f 0.180 GeV/c’.

PACS numbers: 13.38R, 12.15C, 14.80E

The relations between gauge boson mssses and the couplings of gauge bosons sdlow incisive tests of the standard model of the electroweak interactions [l]. The relationships are precisely specified at Born level; higher-order radiative corrections, which are sensitive to the top quark mass, Afttop, and the Riggs boson mass, Mnim, have also been calculated [2]. Measurements of the properties of the 2 boson, as well as measurements in atomic transitions, muon decay, and deep-inelastic scattering, tightly constrain the relationship between allowed v&es of the W mass, Mw, and A&, [3]. Precise measurements of A& and INtop, if inconsistent with the allowed range of predictions, could indicate the existence of new phenomena at or above the electroweak scale. Alternatively, within the confines of the standard model, such measurement8 predict M&w* The measurement of the W mass is unique among electroweak measurements in its sensitivity to charged currents at large momentum transfer.

This paper summarizes [4] a measurement of the W mass using W -+ pv and W + Ed decays observed in antiproton-proton (tip) collisions produced at the Fermilab Tewtron with a center-of-mass energy of 1800 GeV. The results are from a data sample with an integrated luminosity of 19.7 pb-‘, collected by the Collider Detector at Fermilab (CDF) during the period from August 1992 to May 1993 [5].

The CDF [6] is an azimuthally and forward-backward symmetric magnetic de- tector designed to study j$~ collisions at the Tevatron. We briefly describe here those aspects of the detector relevant to this analysis. The magnetic spectrometer consists of tracking devices inside a 3-m diameter, 5-m long superconducting solenoidal mag- net which operates at 1.4 T. A four-layer silicon microstrip vertex detector (SVX) [7], located directly outside the beryllium beampipe, is used to provide a precision mea- surement of the beam axis. Outside the SVX is a set of vertex time projection cham- bers (VTX), which provides ~-t [S] tracking, used to find the z position of the pp interaction (event vertex). Outside the VTX is the central tracking chamber (CTC), a 3.2-m long drift chamber used to measure the momentum of muons and electrons with up to 84 position measurements per track. The calorimeter is divided into

4

a central barrel (171 < l.l), end-plugs (1.1 < ]u] < 2.4), which form the pole pieces for the solenoidal magnet, and forward/backward module8 (2.4 < 171 < 4.2). The calorimeters are constructed as projective electromagnetic and hadronic towers [6]. The towers subtend approximately 0.1 in TJ by 15” in 4 (central) or 5” in 4 (plug and forward). The energies of electrons me measured in the central electromagnetic calorimeter (CEM). Muon8 are identified with the central muon chambers (CMU), situated outside the calorimeters in the region ]q] < 0.6.

This analysis uses the two-body decays W -+ pv and W ---) ev. Since the apparatus cannot detect the neutrino or measure the longitudinal component of the W momentum, there is insufficient information to reconstruct the invariant msss of the W. However, we can infer one additional kinematic quantity, the transverse com- ponent of the neutrino momentum, from a measurement of the transverse momentum imbalance in the calorimeters. For each event we have enough information to con- struct the transverse mass, MT = ((Et + E$)’ - (E$ + Eg)‘)‘i2, where E$ is the transverse energy [S] of the charged lepton (electron or muon), and EG is the trans- verse energy of the neutrino. The measurement of Mw is obtained from a detailed analysis of the Jacobian lineshape of the transverse mass distribution.

The transverse energy of the neutrino is calculated using the charged lepton energy or momentum and the net transverse energy of all other particles (the “recoil”), EF = -(EC + u). The recoil u is calculated as

ll= c ,+we* (~j.. q i, (1) towers

where the sum is over both electromagnetic and hadronic calorimeter towers, Etower is the energy measured in the tower, ti is the unit vector pointing in the direction of the center of the tower from the event vertex, and i is the unit vector in the radial direction [S]. The sum is carried out for towers in the region 1~1 < 3.6. Tower8 in proximity to the charged lepton are excluded from this sum; 30 MeV per excluded tower is added back in to account for average energy flow unrelated to the lepton [4].

The event selection is intended to produce a sample of W bosons with low background and well-understood lepton and neutrino kinematics. Electrons are re- quired to be within a restricted fiducial region of the CEM and have Es > 25 GeV [4]. Muons are required to be within the fiducial region of the CMU and to have fT > 25 GeV/c. Neutrinos are required to have Es > 25 GeV. In addition we require 1111 < 20 GeV, no jet [9] with & > 30 GeV, and no tracks with pi > 10 GeV/c other than that of the charged lepton. Events consistent with cosmic rays or 2 + ff are removed. The lepton track is required to come from an event vertex located within 60 cm of the detector center along the I axis. The W + gv sample consists of 3268 events with transverse masses in the range 65 < MT < 100 GeV/c2; the W + ev sample consists of 5718 events in the same MT range.

We estimate the background from the process W + TV + lvvv to be 0.8% of the W --t Ed and W + /LY samples. Events from Z -+ L! where one lepton is lost

5

Fl.esonance Measured Mass World-Average Mass ( MeV/cz) (MeV/cr)

TW) -+ PP 9460f2&6 9460.4 f 0.2 WS) + Plr 10029f5f6 10023.3 f 0.3 T(3S) + PP 10334f8f6 10355.3 f 0.5

z-t PB 91020 f 210 f 55 91187 f 7 z-+ ee 90880 f 185 31200 91187 f 7

Table 1: Measured masses (after scale calibrations) of the T + /+, Z -+ pp, and Z + ee resonances, compared to published values [lo]. The first uncertainty is sta- tistical and the second is systematic.

make up 0.1% of the W + Ed sample, and (3.6 f 0.5)% of the W + JL’Y sample. Backgrounds from W -+ TY + h + X, where h is a single charged hadron, Z + rr, WW and tZ production, and cosmic rays are estimated to be small [4].

The momentum scale and resolution of the tracking system and the energy scale and resolution of the CEM are measured from the data. The CTC is aligned by requiring that the ratio of calorimeter energy to track momentum, E/p, be charge- independent for high-m electrons. The momentum scale is determined from a sample of - 60,000 J/4 decays (see Figure la), which are also used to limit systematic effects on the scale such as non&c&ties and geometric variations. We find that the nominal scale. should be corrected down by a factor of 0.99984 f 0.00058 [4] for the J/+ mass to agree with the world average value, MJ/+ = 3096.88 f 0.04 MeV/c* [lo]. The scale is verified by measuring the Z and T masses (see Table 1). The uncertainty in the momentum scale, including the extrapolation from MJ/+, to A&, contributes an uncertainty of 50 MeV/c* on Mw ( see Table 2). The momentum resolution is determined from the width of the mass peak in a sample of 330 2 + pp events (see Figure lb) to be am/p; = 0.000810 f O.O00085(stat.) f O.OOOOlO(syst.) (GeV/c)-‘, and contributes 60 MeV/c* to the uncertainty on I&.

The CEM tower responses are equaliaed using E/p for electrons in a sample of - 140,000 events with ET > 9 GeV. The absolute CEM energy scale is transferred from the CTC momentum scale using E/p for electrons in the W + Ed sample (see Figure lc). This procedure contributes an additional 110 MeV/cs scale uncertainty on J~JV for the W -+ Ed channel, of which (see Table 2) 65 MeV/c* is statistical,

and 90 MeV/c* is systematic [4]. Th e energy resolution is (T)’ = (I*)* + ((1.0 f l.O)%)*, where the first term is measured with an electron testbeam [ll], and the second term is determined from a sample of 259 Z + ee decays (see Figure Id). The uncertainty in the energy resolution contributes 80 MeV/c* to the uncertainty on Mw. The width of the E/p distribution for the W -+ Ed sample is described well

6

by the measured resolutions in E and p. The reconstructed mass in Z + ee decays is also used as a check on the CEM energy scale, as shown in Table 1.

The detector response to the recoil ]u] is directly calibrated using Z --t ee decays, for which there is a good measurement of the true & from the measured electron energies. The Z + ee event sample is used as a table from which one can look up the measured response ]u] for a given pg. We assume that the response to the recoil from a W of a given pr is the same ss that to the recoil from a Z of the same pT.

Lineshapes in transverse mass corresponding to different W masses are simu- lated with a leading-order (i.e. pF=O) W Monte Carlo using the MR.9 D’_ parton dis- tribution functions [12]. The lineshapes include contributions from backgrounds [13]. To model the lineshape accurately, we need to incorporate a p? spectrum in the simulation. The similarity of the pr spectra of W and Z bosons observed in direct measurements [14] and in theoretical predictions [15] leads us to use the observed Z + ee pr spectrum, corrected for electron energy resolutions, M an initial guess for the pT” spectrum. We modify the shape of this spectrum in order to match the observed us distribution for the W events, where us is the component of the recoil perpendicular to the direction of the charged lepton. We find that the simplest mod- ification, scaling pr in the p$ distribution by a constant factor, gives good agreement for both electron and muon us distributions. We consider other modifications to the shape in estimating systematic errors; the uncertainty on A& due to the modelling of the pF spectrum is 45 MeV/c* [4].

Transverse mass spectra are generated for a range of W masses, at 100 MeV/cr intervals for W + ev, and 150 MeV/c* intervals for W + PY [4]. The value of the W width used is Dw=2.064 GeV [lS]. At each mass point, an unbinned log-likelihood is calculated for the hypothesis that the data are consistent with that mass. The log- likelihood values fit well to a parabola. The transverse mass spectra and the Monte Carlo lineshapes corresponding to the best fit mass are shown in Figure 2. We add 168 f 20 MeV/c* and 65 k 20 MeV/c* to the fitted masses in the muon and electron channels, respectively [4], to account for the effects of radiative W decay [17].

The value of the W mass extracted from the W + PY data is Itfk = 80.310f 0.205 (stat.) f 0.130 (syst.) GeV/c*. The mass from the W + Ed data is Af& = 80.490 f 0.145 (stat.) f 0.175 (syst.) GeV/c*. Accounting for correlations in the uncertainties, the combined data yield Afw = 80.410 f 0.180 GeV/c*. Fits with the W width unconstrained yield consistent results [4].

The measurement uncertainties are summarized in Table 2. The largest sys- tematic uncertainties, beyond those of the momentum and energy scales described above, are due to the limitations on determining the electron energy and muon mo- mentum resolutions, the W transverse and longitudinal production distributions, and the detector response to the recoil. Varying the parton distribution functions of the proton varies the distribution of the W longitudinal momentum, and, through accep- tance effects, the lineshape of the transverse mass spectrum, leading to an uncertainty

7

Uncertainty

I. Statistical

II. Energy Scale 1. Scale from J/$J 2. CTC Alignment 3. Calorimeter

a. Stat. on E/p b. Syst. on E/p

(&$*) (~~/k*) (iti;:;

145 205

120 50 50 50 50 50 15 15 15

110 - -

65 90

III. Other Systematics 1. e or p resolution 2. Input py 3. Recoil modeling 4. Parton distribution functions 5. e or p ID and removal 6. Trigger bias 7. Radiative corrections 8. W width 9. Higher-order corrections 10. Backgrounds 11. Fitting

130 80 45 60 50 25

0 20 20 20 10 10

120 90 60 45 25 60 60 50 50 10 5 25 - 20 20 20 20 20 20 25 - 10 -

TOTAL UNCERTAINTY 230 240 100

Table 2: Summary of uncertainties in the W mass measurement.

8

on Mw. We uze a measurement of the forward-backward charge asymmetry in W decays [18] to set a limit on this effect. Other sources of systematic uncertainty, each contributing 25 MeV/c* or less, are lepton identification (ID) and separation of the lepton energy deposit from the recoil energy sum (see Equation l), trigger bias, radiative corrections, the W width, higher-order QCD corrections to W production, backgrounds, and the fitting procedure. For the purpose of combining the W --t ev and W -+ PY measurements, we also list in Table 2 those components of the uncer- tainties common to the two channels. Details on the methods used to determine the systematics, and the checks on the methods used, are given in Reference 4.

This measurement of the W boson mass has an uncertsinty half that of the best previously published measurements [19, 201. Figure 3 shows the sensitivity in the Afw-Mtop plane of this result, M w = 80.410f0.180 GeV/c”, when combined with the value Aft+, = 176 f 13 GeV/c* [21], compared to theoretical predictions based on electroweak radiative corrections [22].

We thank the Fermilab staff and the technical staffa of the participating in- stitutions for their contributions. This work was supported by the U.S. Department of Energy and National Science Foundation; the Italian Istituto Nazionale di Fisica Nucleare; the Ministry of Science, Culture, and Education of Japan; the Natural Sci- ences and Engineering Research Council of Canada; the A. P. Sloan Foundation; and the Grainger Foundation.

References

[l] S. L. Glashow, Nucl. Phys., 22, 579 (1961); S. Weinberg, Phys. Rev. L&t., lQ, 1264 (1967); A. Salam, Proceedings of the Eighth Nobel Symposium, edited by N. Svarthohn (Ahnqvist and Wiksells, Stockholm, WiIey, New York, 1968), p. 367.

[2] For recent reviews, see W. HoIlik and W. Marciano, in “Precision Tests of the Standard Electroweak Model”, ed. by P. Langacker (World Scientific, Singapore, 1994), and B. A. Kniehl, KEK preprint KEK-TH-412, Sept. 1994, submitted to Int. J. Mod. Phys. A.

[3] J. L. Rosner, Rev. Mod. Phys., 64,115l (1992); erratum, ibid., 66, 1117 (1994).

[4] F. Abe et al., Fermilab preprint FERMILAB-PUB-95/033-E (1995), submitted to Phys. Rev. D. See also R. M. Keup, Ph. D. dissertation, University of Illinois, 1995, and D. P. Saltzberg, Ph. D. dissertation, University of Chicago, 1994.

[5] The W + ey sample is from an 18.2 pb-’ subset.

[6] F. Abe et al., Nucl. Instrum. Methods, A271, 387 (1988).

[7] D. Amidei et al., Nucl. Instrum. Methods, A350, 73 (1994).

9

[S] The CDF coordinate system has the .s (longitudinal) axis along the proton beam axis; r is the transverse coordinate in cyIindrical coordinates, measured from the beam axis. Pseudorapidity (7) is defined as w = -ln(tan(6‘/2)), where 6 is the polar angk relative to the proton-beam direction. “Transverse energy” (ET) denotes the transverse component of the vector whose magnitude is the energy of the particle and direction is parallel to the momentum of the particle.

[9] F. Abe et al., Phys. Rev., D45, 1488 (1992).

[lo] L. Montanet et al. (PDG), Review of Particle Properties, Phys. Rev., D50, (1994).

[ll] L. Ballra et of., Nuci. Instrum. Methods, A267, 272 (1988).

[12] A.D. Martin, R.G. Roberts and W.J. Stirling, RAL-92-099 (1993).

[13] Except smail bachgrounds in the muon channel from Z -+ TT and jet or heavy- flavor events faking W decays (see Reference 4).

[14] F. Abe et ol., Phys. Rev. Lett., 66, 2951 (1991); F. Abe et al., Phys. Rev. Lett., 67, 2937 (1991).

[15] M. H. Rena, University of Iowa preprint UIOWA-94-01, 1994.

[16] F. Abe et al., Phys. Rev. Lett., 73, 220 (1994).

[17] F. A. Berends and R. Kleiss, Z. Phys., C27, 365 (1985). The calculations are implemented in a Monte Carlo program described in R. G. Wagner, Comput. Phys. Commun., 70, 15 (1992).

[18] F. Abe et al., Phys. Rev. Lett., 74, 850 (1995).

(191 F. Abe et al., Phys. Rev. Lett., 62, 1005 (1989).

[20] J. AIitti et al. (UA2 Collaboration), Phys. Lett., B241, 150 (1990).

[21] F. Abe et d, FermiIab preprint FERMILAB-PUB-95/022-E, submitted to Phys. Rev. Ledt.

[22] The curves are. calculated using a FORTRAN program from F. Halzen and B.A. Kniehi (private communication), described in Nucl. Phys. B353,567 (1990). See Ref. 4 for details.

10

Gim

70 a, k c&9 ‘w

110 120

Figure 1: a) The dimuon mass spectrum near the J/+ mass peals, used to normalize the momentum scale. b) The dimuon mass spectrum near the Z rn88.s peak, used to determine the momentum resolution. c) The E/p spectrum for electrons from the W + w sample, used to determine the energy scale. d) The dielectron mass spectrum near the Z mass peah, used to determine the energy resolution. The solid line in (a) and the hi&grams in (b), (c), and (d) are Monte Carlo simulations, including radiative effects.

11

I I

CT E 1 (a> w + P

I I I I I

‘> ,A I M, = mT.31 +n 34 GPV

p 708 pw (?j 500 c! 400

> 300 c 200 ? w 100

~ O40

M, = 80.49f0.23 GeV

MT (GeV/c’)

Figure 2: Transverse msas spectra for a) W + pv decays and b) W + ev decays. The histograms arc from the simulation using the respective best-fit mass. The sxrows delimit the region used in the mass fit.

12

I I I I I

81.2 - W Mass (this measurement): M, = 80.41 f 0.18 GeV/c2

81 - CDF Measured Top Quark Mass:

80.8

N- 5 80.6

l? - 80.4 r’

80.2

80

MM=1 76 f 13 GeV/c2

140 160 180 200 220 240

M, (GeV/c’)

Figure 3: The data point represents this measurement of MW and the CDF mea- surement of the top quark mass of Mt,,=176f13 GeV/cs [21]. The curves are from a calculation [22] of the dependence of Mw on A& in the minimal standard model using several Higgs mawes. The bands are the uncertainties obtained by folding in quadrature uncertainties on a(@), Mz, and a,(@?).

13