-L9E-

NO2LLVAXL ‘3HL J,V S3ISAHd XOAV?d AAWIH

-89E-

aql103 paw suonw pur? suog~a~a aA!st-qau! spa~as ~a%?rJl 1a~a1 aaJq1 v ~uo!leay!map! uonm ap!AoJd 0.1 > 101 uo@aJ aql II! sJaqureqa $3~ ‘Jalau~uo~Ea aql aprslno 'sou -u$nau a!$a8Jaua palaalapun30 aauasaJd aq) alwpu! UED~X~M (+#KsJaua asJaasu~?J~ %~~ssfu~aq~aJnsEau~o~ pasnospaxe sJa)aur~op2:,aq~ 5a~epypueauo~~aa~apues1afaJns -earn PUE k3!u1ap! 01 pasn are days 'z'p N Ikl3 o sa!l!p!deJopnasd 01 punxa pw a@~ @QtlUI!ZE aIOqM aql JIaAO3 SJa]XII!.IO@3 (w/%0< N p”~3/(p0~3)D) a!UO.Ipaq pUE

(W/%LI - 2ua3/(“a3)~)3!~augeuro~13ara ‘p!ouafos aqlap!slng '(%g'o $ sd%r.o = ~~/(~~)o) uo!uqosaJ uwuauIotu 1ualla3xa UE ql!M ‘~auku @p!oualos 8u!wpuo3 -Jadns l-p'1 e u! pasJaunu! s! qD!qM (3,~3)Jaqur~qa 8ulyaeJ$ @~luaa aql u! paJnsEaru s! saIa!wd pa%qa30 UnwazuouI aqL .sLEaap yJvnb 3 pue q ~1013 say.m hrepuo -3as &uap~ 01 pasn s! puv (2u7/ LT N "PO) smaq aql 01 JvIna!puadrad uo!laa~!p aql u! uo~lm~~suoaa~ yae~l as!aaJd saprAoJd ‘adId tuvaq aql ap!s]no d[alE!pauuu~ lsn[paw -01 rr‘(XI\S)J013alap XalBA UOXI!S JaLE[ JUOJ pJEq UOW!pEJ ‘aS!OU-MO1 MaU V 'SJaq -urcqa uonw PUE sJalauIuo@a Lq papuno.uns Jalauronaads a!lauk?Eur E 30 SIS!SUOD I!

'4wa ,;aJaqMasla Ip2jap u! paquasap QaA!sualxa uaaq seq Jovalap da3 aqL wJrvaqaq~o~le~na!pua&aduop2aJ~aqlu!

wrlg~ -p~‘au~~un2aqaq~%u0~u1a og = *D 30 qlp!MvZkqAequo~nq!Jwpue!ssnEf) e q$!M uo!%aJ snoyuuq E %u!anpoJd sti S'E hraaa ssoJa saqaunq aq,L 'aam?n~a ureaq aql past?aJaap PUE aur!la3!1 ureaq aql paAoJduq suo!rrasu! asaqL .suo@aJ uo!vasJaw! @a pur? da3 aql IE ldaaxa ‘surEaq aql %+redas ,suo!uasuy agElsoJvaIa30 ]as e 01

SvJEq~‘SpouadNw-PEP s661-~661 PW ~661-2661 aql WJnp T-Sz-~a IFFY x z3o saysou"nI paqaEaJ d[au!lnoJ 1~ ‘r-sZ-tu3 OE~~ 30 1ClysouFnI Tads Jo3 pau%!saa

'ppo~ aql u!Jap!~lo:, BJaualsaq8g aql SF ‘AaL 8.1 30 BJaua sserujo-Jaluaa E ql!M ‘Japqlo3 uoJwAaL & qEIyua,I aqL

SIOOJ, @~uampadx~ 1-z

ms aw 30 Jolaas q aql30 dpnls aq101 suo!mnqguo3Jo[w apwu 0s~ aAEq‘Jtqna!rrt?du! da3 PUE ‘JapyIIoa uoJ$EAaLaaql‘p 'Jas u!uMoqs aqI[!~ sv .uonEAaL aq~~eparpnis yJEnb hEaq r([UOaqluaaqlou SEq~JRIbdolaql‘JaAaMOH '~361 ~!%I~uO!SS~O~~~!~XI!~J~~~~~~~ uonr?AaLqvIvad aql30 sp108aql3oauo uaaqseqyJvnbdolaql3ohraAoasyp aqL

.qJvnb q aql3oJauwd z/l+ = SIP 30 dlrssaaau aql ‘aJo3aJaql pue‘yrenb q aql30 ydsos! yeaM aqlJo3 Z/I-- = &130 luaus -@SST aqlparroddnssmauIaJnsEamaq]aJaqm S‘a~s-d~~prre LmL3dw qq t -a+a -

u! ad?ry hlaunudse pleMyDvq-pJEAuo3 aq] “0.13 palat2Jjxa ‘ymnb q aql30 ~uaura.rns~auI urdsos! yeaM aql dqpa3npoJd SEM yrenb doI aql30 a3uals!xa aq] ~03 aauap!Aa Jaqvn3 pu~'~a~qnopu~dsos~~aME30sJaqu1ama~y~enbdo~puv q aq~‘~ySaq~u!‘rl~JE~~!s

'TJenb do) aql uoyuyap 1(q s! Jaumd s!qL 's/z+30 a%qaE ql!MJauwdynnb q ~30 aauals!xa aqlpa!Idur! uoldal L aql30 aauals!xa aql PUE yrenb q aql~o3 a%qa E/I- aql‘aJo3aJaqJ ~(Lp1m3 E u! saZ.nzqD q.nmb aq] a= $J pue ‘sa%q:, uoldaI aql aJE ,~‘sy~vnb paJo[oa 30 Jaqwnu aq) s! "l\i aJaqM‘0 = &"K"N + @z)IInus! dlyn23uaA!t?e u! sa%Eq:,uo~a3papwq-yalaql IF 30 ums aql3~ Lruo pm 3! ,hoaql aaJ3 dpmorn? ue SF @is) Iapotq pJwws aqL

.yJEnb q paJaAoas!p @au aqlol E/I- aZhq3 e 30 iuamu%!sse aql paMoI@ saaueuosaJ MOJJEU asaqL 'c3/Aaf) S.6 lnoqE30 ssew e le sa3uwosaJ J aql30 aauals!xa aql pauuyuo:, ,Asga le %uu a8E -Jois -a+a smog aql %aAoas!p a! Ja33E uoos PUE ,‘qepnuas 1~ suo!s!1103 snalanu -uoloJd Aaf) @)p u! a3wuosaJ uonuup E SE paJaAoas!p SEM yJt?nb q aql ‘~~61 III

.salElsua2!a uo!laEJalu!yeaM aql PUE SSEW aql uaaMlaq %u!x~ awudoJdde UE pue ‘saFIpuE3 yJEnb aaJql30 aauals!xa aql qPnoJq1 ,uxawk &

-,,x aql u! ~0~~~10~~ da pahlasqo aq) pau?Idxa leql ,s]uaurn%E @a!$aJoayl uo paseq SEM uo!lEd!a!lue aql30 md ISI?aI IV .r([nue~ yJvnb,,palEd!a!wE ]sour,, aq$ uaaq sEq uo!] -eJaua% yrenb pJ!q~ aql‘gL61 u! ,uoldaI L a4130 hraAoas!p aql aau!s .uo!lEJaua8 yJEnb pJy]aq$30 sa~sdqdq~!MSnOwdUOUdS s!Jap~~~oauo~~Aa~a~]e w!sLqdJo~~?g I(AEaH

's 'aas UT uaA@ SF ‘Jolaas q aql u! sluaruaJnst?atu uo!w[o!A-d3 SE IIaM s~‘sa~sICqdqrenb dolJo3 svadsoJd aJn$y aql )e ~oopno3a!Jqe‘&Eurd 'uosazu "a aql30 uo!]ehlasqo aql ssnwp os@ I[!M a& .&C/I/~ t & ‘,,a I(EDap aql u! hauuudse luapuadap-am!1 aql30 wacuaJnseam sa~~~~]eis-~o~'~u~~~aJd~uos~sEqduraq~~~‘Jap~~~o~uo~eAa~aq~~e sl1nsa.I sa!sLqdg ~uaaaJo~pa~oAaps~puo!laag 'ss~u1y~~nbdo~aq~3ouo~]eu~a~aplaaJ!pa~EJaAE-ppoM aql pue uopsas ssoJ3 uo!lanpoJd do1 aql30 IuauraJnseaur aql ssnas!p aM ‘JEIna!wd UI 'E 'aas u!& pue da3 IE saysdqd ynnb doI 30 snlels aql azwuuuns aM ‘z 'aas II! Ma!hlaAo @a!~o~s!q~a!~qe Jal3v 'qEI&IrJa~ ~2 lap!1103 &uoJaAa~ aql le sa!sdqd JOAE~ dAvaq 30 Iaa[qns aql uo pau!Elqo sl[nsaJ IuaaaJ 30 Ma!AaJ e masaJd aM ‘Jaded s!ql q

top search. The detection efficiency for tt events is improved by the inclusion of trig- gers based on &.

A schematic view of the CDF detector is shown in Fig. l(a).

(a)

CDF Detector BBCkWBTd

Figure 1: Isometric views of (a) CDF and (b) D9 detectors.

The D9 detector and data collection system are also described elsewhere.” The D9 detector has a hermetic, compensating sampling calorimeter with fine longitudinal and transverse segmentation in pseudorapidity and azimuthal angle. The energy resolu- tions are slightly better than those measured at CDF (g(E,,)/E,, N 15%/G) and

(g(Ehad)/Ehad N 50%/G). Since there is no central field, charged particle tracks are reconstructed with a sign degeneracy using drift chambers located between the in- teraction region and the calorimeter. Electrons are identified by a transition radiation detector. Muons are detected by reconstructing tracks in proportional drift tubes before and behind a set of magnetized iron toroids located outside the calorimeter which pro- vide some momentum measurement with a resolution of ,(P,)/P, = 0.3%& @ 17.% for pseudorapidities in the range In/ < 3.0. The good calorimeter hermeticity provides a good missing transverse energy resolution. The CDF transverse energy resolution is approximately 20% worse than D9. A schematic view of the D9 detector is shown in Fig. l(b).

3 Top Quark at the Tevatron Collider

As mentioned previously, the top quark has been searched since the discovery of its partner, the b quark, in 1977. The first indication that the top quark was a heavy object came from the measurement of a large mixing parameter zd in the B”-9 mixing, first observed by the UA113 and ARGUS14 Collaborations.

Precise electroweak fits from LEP have constrained the top quark mass with ever increasing precision since the turn-on of LEP in 1990, as shown in Fig. 2.

01 1988 1990 1992 1994 1996 1998

Year

Figure 2: Historical evolution r5 of the indirect fits to the top quark mass from precise electroweak measurements at LEP (circles). The solid and dashed lines indicate the 95% C.L. on the lower Mtop bound from direct e+e- and pij searches. The last points are the direct measurements from CDF (triangles) and D9 (inverted triangles).

The first evidence for the top quark’s existence was published by CDF” in 1994, with a 2.8~ excess of events over the background expectations. Under the assumption of top production, CDF measured a top mass of M top = 174116 GeV/c’ and a production cross section ctt = 13.9’2:: pb. Both CDF17 and D9" announced the definitive top discovery in 1995, reporting mass values of:

Mop = 176 f 8 i 10 G&/c’ (CDF),

Mtop = 199i$ * 22 GeV/c’ (DQ),

-369-

log,

, [M

/(1

W/2

)]

. Dilepton channel (ee, p/1, or ep), corresponding to a branching ratio of N 5% of the total decay.

l Lepton + jets channel (e or p), corresponding to a branching ratio of N 30%.

l Hadronic channel, corresponding to the remaining N 45% of the total decay.

In addition, there are N 21% of the tf decays containing a r lepton in the final state.

3.1 Top Quark Production at the Tevatron Collider

3.1.1 Dilepton Channel

Given the process t? + W+W-bb + l+Z-vob& the final state for this channel is determined by two oppositely charged leptons with high transverse momenta, large &, and two b-jets. The dominant backgrounds are WW, Z -+ r+r-, and the Drell-Yan production. The dilepton channel is expected to have a very good signal-to-background ratio. However, this decay mode is still limited by statistics, and therefore, not ideal (yet) for a precise measurement of the top mass.

The CDF dilepton search starts with the identification of a lepton (e or p) with PT > 20 GeV/c and satisfying a set of isolation requirements in the cental region (1~1 < 1.0). The second lepton is required to have PT > 20 GeV/c and to satisfy a looser set of isolation requirements. The two leptons must be oppositely charged, and events with ee or 1-1~ candidates where the invariant mass is between 75 and 105GeV/cZ are rejected as being consistent with Z” candidates.

In order to reject the Drell-Yan events, CDF requires & > 25GeV. For events with & < 50 GeV, the azimuthal angle between the &r and the closest lepton or jet must be greater than 20” to reduce the background coming from Z --t T’T and Drell-Yan events, where a mismeasured jet produces an artificial &“r. Finally, all the events are required to have at least two jets with Ep > 10 GeV and InI < 2.0. Background contributions are estimated from a combinations of data and Monte Carlo simulations.

Figure 5(a) shows the distribution of the nine dilepton candidates (one ee, one ppL, and seven ep events) surviving these cuts in the total integrated luminosity of 110 pb-’

at CDF. The DQ analysis proceeds in a parallel way, the major differences being the cuts

on the lepton PT (PT > 15 (20) GeV/c for the ep, /1/1 (ee) channels) and a minimum requirement on HT (HT > 120 GeV for the “electron” channels, or HT > 100 GeV for the “muon” channel). HT is defined as the scalar sum of the transverse energies Egt of

(4 Rtm I di lqm data ! LO9 pb’), CUFpn4imin.q

NJ2

(b)

0 50 IW 150 2W 250 3W 35j ,4X i

lw

Figure 5: The azimuthal angle Ad between the J??r vector and the nearest lepton orjet vs J?r in CDF dilepton events. (a) The small points show the distribution expected for the tT signal, while the larger symbols represent the data. (b) HT distributions of dilepton events in the D @ analysis. The observed events are shown by the darker histograms.

-371-

c

a

Figure 6: Aplanarity vs HT for the ev analysis in D @ for data, tt Monte Carlo events, multijet background, and W-t 4 jets Monte Carlo background. The dashed lines indi- cate the cuts.

bb jets (- 30 events), 2 + II (- 26 events), and diboson production (- 15 events). Clearly, additional background rejection is required. D$9 and CDF solve the problem of the rejection of W+jets QCD backgrounds with three different approaches:

l Event Shape analysis24 (D@ ). This approach relies on the fact that, for heavy top, the overall event is different (more spherical, and with more energy) than nor- mal QCD W-t multijet events. The variables used by D(Z) to discriminate against background include the event aplanarity A and the already-mentioned HT. No

attempt is made at the identification of the original flavor of the jets in the event.

. b-quark tagging through the semileptonic decay (CDFz5 and Doz4). This ap- proach identifies the b-nature of the jets present in the event through the presence of a soft lepton embedded in the jets and originated by a semileptonic decay of the parent b quark (SLT, or Soft Lepton Tagging).

. b-quark tagging by mean of displaced vertices or displaced tracks (CDF25). This approach relies on the finite b-quark lifetime and the superb precision of the Sil- icon Vertex (SVX) detector to identify displaced vertices (SECVTX tagging) or displaced tracks (JPB tagging).

Table 2: Cuts used by CDF and D @ for the lepton + jets search.

CDF W All e(b) e + jets/b

pi- 20 GeV 20 GeV 20 GeV GT 20 GeV 25(20) GeV 20 GeV

Number of Jets L 3, I I< 2 T 2 4,Iq I< 2.5 2 4,I I< 2.5 7

@et

J 2

15 GeV 15 GeV 2 > 0.065

20 GeV > > 0.040

HT > 180 GeV > 110 GeV

-373-

-PLE-

5uo!]Epm!s 01~3 aluom snoyA pm ewp aql 103 salqym Jjy pm y u! sluaAa 30 uo!lnqys!p da aq) sMoqs 8 a.m%g

‘sluaaa slaf + uoldal

.uogmpoJd do1 01 I(lajaIduIo3 palnqyle am smq la[-p +M pm? -E +M aql u! SluaAa 30 ssa3xa aqA

.suo!lwadxa ql!M luauraa&k ~00% u! an2 slep pahlasqo aql ‘puno.6+?q dq palcqndod dlalaIduro3 aq 01 pal3adxa s! q3!qM ‘mq pa[ 1 + M aql q .kmafz33a %u@ihl pamwarn aql 8u!sn 01~3 aluom aqltuo~3 palvugsa am 3~ JO ‘33M ‘qqM 01 anp sluaaa la[!qnm +A 30 uog~13 aql apqM ‘sla[ aayym! UK pamseam aleI %1sg11 e uo pawq an? suogel3adxa punoJ%yzq aqJ ‘Aa3 II! swqlpo%[e %!%%?J LTs pm xh,y aql .ral3e sluaAa iaf&pu +M paMasq0 30 laquxnu aql shioqs L am8y

.suog!sodwo3 ymb-q puv ymnb-3 luaJaJJ!p ql!M elep 30 saldum IDalas 01 pasn aq 1~23 pm2 [dpApadSaJ ‘(~7s) O&W pw ‘(a& %6-

‘(x,~,yaas) %p~] luaJa33!p am sla[y.mb-3 ~03 s[oo] %u@?e] asaql3o sa!may&a aql ‘SD! -wuauyy hap pm atupa3!1 luaJa33p aql 01 anp ‘Lpual3 ‘(~71s) %L- puv ‘(a& %zz-

‘(x~~~s) %EZ- am stool %I!%%$ aaq$ aql~o3 la! ym+q lad sapua!Dy3a pm!dLl aqL

Table 3: Summary of the lepton + jets counting experiments.

Sample 1 DO / CDF i Event Shape Observed 19

Background 9.7 f 1.7 Expected (M,, N 170 GeV) 14.1 f 3.1

b + 1X Observed 11 Background 2.4 310.5 Expected (&fin N 170 GeV) 5.8 f 1.0

Displaced Vertex Observed -

Background -

Expected (M,, N 170 GeV) -

22 7.2 f 2.1

-

40 24.3 f 3.5

9.6 34 Events (42 Tags)

8.4 f 1.4 19.8 f 4.0

Table 4: Predicted signal and backgrounds in the CDF Hadronic Top search. -2% S+B 2 5 jets 2 6 jets

No tag lL5OO l/200 1 b - tag l/ lOO l/30 2 b - tag l/20 l/10

]

which includes requiring the CET 2 300 GeV and C > 0.75, where the centrality C is defined as C = HT/& and B is the invariant mass of the multijet system. After these kinematical cuts a requirement of one b-tagged jet produces the multiplicity distribution shown in Fig. 9(a) which shows a clear excess of events over the expected background from QCD contributions.

A similar analysis from D$?Jz7 uses a selection based on the discriminating power of a neural network fed with 14 variables which include kinematical quantities (like aplanarity and total scalar energy) as well as soft lepton tagging of b-quark jets. The output of the neural network is shown in Fig. 9(b).

The cross section values obtained by the two collaborations for the all-hadronic decay modes are:

0, = lO.l?;:; pb (CDF)

a,, = 7.1 + 3.2pb (DQ).

Figure 9: (a) CDF multiplicity distribution for all hadronic analysis with the tagged events and the background expectation. (b) Distribution of the final neural network out- put in the D$II all hadronic analysis, showing the results of a fit of the observed tagged distribution to the predictions for signal and background, with Mtop = 180 GeV/c2.

3.1.5 Summary of the utz Cross-Section Measurements

The summary of the a,, cross-section measurements performed at the Tevatron is shown in Fig. 10. The D @ and CDF measurements are in good agreement with the averageZ8 of a, = 6.7 f 1.3pb, which is slightly larger than, but still in good agreement with, the theoretical SM predictions of Fig. 4.

3.2 Top Mass Determination

A precise measurement of the top quark mass plays a central role in our understanding of the mechanism for the symmetry breaking in the SM. A precise direct measurement can provide a consistency check of the experimental data from different sources, and a combination of the top and W mass measurements provides information on the mass of the Higgs boson.

As for the cross-section measurement, the top mass determination can be performed in any one of the three different decay topologies of the tt event (dilepton, lepton +

jets, and all hadronic). At the time of the top discovery, the lepton + jets chan- nels were studied more extensively given the relative large signal-to-background ratio,

-375-

-9LE-

‘uInuI!U!W Sl! 0lPadsa.I qlrM pooq!Iay@oI aa@au aql u! a8ueq3 IFun-3Ieq e o) spuods -au03 Kuywa3un aql aJaqm ‘,3/~a3 8.p T 6.9~1 = *TM saJnstzaur da3 ‘sfql WOJ~

L’S E’S 8'P @FL

0'1 SO 1'1 w3 alum ‘dad L’I E'I E'O punoJ%ycw~

8'1 9'Z I’E uog~2pe~ uonIg

O’S VP 8-i J3 w D!UOJ~EH IIv s?aC + uoTda7 uoldapa amiog

‘SluawaJnseaw ssem doI da3 aql u! saguyJa3un zymals& 30 A.rvunun~ :s a[qg,

‘(‘I) I I ‘~!LI u! UMoqs s! saIdunzs luaJa3gp Jno3 aql30 uogeuFqtuo3 v .uogDun3 pooqyayg-SoI aql30 adeqs aqi sMoqs lasu! aql ‘uognqys!p qma 01 .(e) II %,I u! uwoqs saIduresqns luapuadapu! KI@~gsyis Jno3 aqi oiu! siuaha aq] %!p!A&)qns 30 sls!suoc~ aIdums aqi uog!gvd 01 KEM Iwgdo aql lsql paleJlsuourap sa!pnls OIJQ aluom .uognqgs!p sseru-palanrlsuoDaJ aql 30 ssauMo.ueu aql put2 ‘O~VJ punoJi?y3uq-ol-@?!s aql ‘s)uaAa pahlasqo 30 Jaqumu aw ~I!M aseaJDu! 01 pal3adxa s! luauIaJnsuauI ssrw ymnb do1 aql30 uo!s!3aJd aqL ‘palDa[aJ am 01 < ,X q]!M s]uaAs %uoyu!quroD @~oleu!quro3 aIq!ssod Kwur aql30 lno ,X Isa -MOI qlJM uogEJn8yuo3 aql SasooqD pun suomd-q 01 pauk?!sw aJE Sla[ pa%k?$-~~S PIE? xhs ieql SaJ!nbaJ da3 ‘aAoqr! paqp3sap auraq3s %uyly 32 aqi Bu!MoIIo~ wuaAa ~8 8u~uymo~ aIdu.res ssuw B sauyap uop3aIas aAoqe aqL ‘aaJq1ls.g aql se sluauxaJ!nbaJ k put? J3 aws aql K~s~J?s lsnw laf quno3 aql ‘]uasaJd s! %l qms ou 31 .sla[ aql I@ uo paMoIIe am s8el~~s aI!qM ‘(~a3 g < J3) sla[ 3wpeaI aaq aql uo paMoI@ LIUO are S%l XhS WIql!JO+ LTS JO XAS aql Kq pa88el S! sia[ %npvaI mo3 30 au0 pap!AOJd ‘p’z > IhI put2 Aag 8 < J3 aq 01 paxvIaJ s! la! qJ.nIo3 aql uo S$UaUIaJ!nbaJ aql ‘aZNII?l -daDz aql aseaJsu! 01 JapJo q ‘z > Jhl puv ~93 91 7 J3 pahIasq0 ue aAuq lsnur qD!qM 30 aaJq1 ‘1UaAa qxa II! paJ!nbaJ am Slaf JnO3 1seaI ]v ‘(Aas 0~ < J&) sluamaJ!nb -aJ L&pm (D/A33 0~ < fd) uoldaI @nsn aqlql!M S~JVIS uog3a~as IuaAa aql 6z‘da3

UI .Kwap do] e u! paAJasqo am sla[ .mo3 I@ 3~ KIuo aIqesn SF poqlaur aql ‘KI.reaIa

PJV = ‘Qq puv ‘&&7 = (yL-ZC)H ‘&$iy = (czf)M Buraq SlU!~JlSUO3 SSBUI aql) UO!lVA -Jasuo3 30 suogEnba 0~ pue ‘sagguunb (UMOT JO) pamseam p6 ‘sa1qy.mA zs a.m aJay aDu!s ly 32 e saA@ uoguhIasuo3 um]uauIour-KZJaua pm2 sassvm u~ouy 30 asn aqL

.P&-Z,c t qM t Zf .

pUE‘r,bZ~tqMtl~ 0

‘9urq+h? + “?I$ t dd . -

:sapnIDu! dpnls Japun s!saqlodKq aqL

.sluawpadxa OM] aql Kq pauvalap sswu a%JaAv aq3 sv IIaM su ‘da3 pw da Kq syJauraJnsuaw ssem do1 ayl aqpxap 11~~ suoyas %!Mo~~o~ aqL .(y& N “a.! ,~/~a3 F; -) uo!s!DaJd PO08 KJaA e ql!~ swzu yJtmb doI aql30 aspaImouy aql %!MoI@ am qcyM SIool %!doIaaap ‘sIauueq:, Kwap z+o.lpoy 11~ pue uo$daftp aql30 s!sK@ue aql pue sJoUa cpmalsds

30 %!pUElSJapUn J!aql paAOJdUI! SUO!Vz?JOqt?IIO3 qlOq Uaql aXI!S ',3/AaD ()Z N 30

K$uga3un uv ql!M u~ouy SEM ssem do1 aql gj6I u! ‘1InSaJ )! sv ysaqlodKq SSBUI do1 aql 01 lg pauyJisuo:, B %uyojlad 30 Kl!I!q!ssod aw pue ‘O!IEJ %u!qcwJq a&I aql

.suog3!paJd @3gaJoaql30 a%UJ aql s! UMoqs OsIv ‘da puu da3 Kq pa!pnis sIauuc?q:, aq) u! paJnssaw uoyas SSOJD uog3npoJd 2~ aqL :OI aJ&k$

qdr:‘g-L’t

qd g'f-w+roi qd g'f-f'ttzx qdri-i'ztz'g qd VE-vttzx

Figure 11: CDF top quark mass distributions for the Zepton + jets sample. The points are the data, the dark area is the top signal+background resulting from the fit, while the lightly shaded area is the background alone. The plots on (a) show the four independent samples, while (b) shows the combination of the four samples.

The CDF systematic uncertainties for the various tZ decay modes considered here are listed in Table 5. The final measurement is

Mop = 175.9 * 4.8(stat.) f 4.9(syst.) GeV/c’ (CDF - lepton + jets).

D @ performs a two-dimensional likelihood fit to extract the top mass value. One variable in the two-dimensional distribution is the best fit mass obtained by the 2C analysis of the data. The other variable is a top discriminant, which provides a dis- tinct separation between the top signal and the background, without biasing the mass analysis. DQ uses two discriminants 3o based on the following four variable:

l @T

a.4

. HTz/Clpzl, where HT2 is defined as the HT minus the ET of the leading jet. This variable measures the centrality of the event.

. (AR~)E~n/(E&+@T), where (AR?) is the minimum AR between all pairs of jets and ET mzn is the smaller jet ET from the minimum AR pair. This variable measures the extent to which the jets are clustered together.

These variables are combined in a Neural Network (NN) and a Low Bias (LB) discriminant to provide the kind of separation illustrated in Fig. 12(a). D(D then fits the two-dimensional distributions to templates determined from simulated ti events and background estimated using a combination of Monte Carlo and data. The experiment obtains a mass measurement of MtOP = 173.3&5.6(stat.) GeV/c* shown in Fig. 12(b). The systematic effects, coming mostly from the jet energy scale and the Monte Carlo modeling, sum to f5.5 GeV/c2. The final mass obtained by DQ in the Zepton + jets

channel is:

Mtop = 173.3 f 5.6(stat.) f 5.5(syst.) GeV/c* (Do - Zepton + jets).

6)

172 GeV top 1 ..mn DO..

..00~00.*.

..monom...

..lloca....

4 . . . . . . . . . . ..~CCll..

3

. ..m[Im....

. ..*m.....

. . . . . . . . . .

. . . . . . . . . . 0 00 120 160 zoo 240 280

(a) background

Fitted mass True mass K m 164 I.& . . . . . . I

Data

Fitted top quark mass (GeV/c’)

Figure 12: (a) Events per bin in the D NN, mfzt plane for the D@J neural network dis- criminant analysis, showing the expectation for top, background, and data. (b) Results from the D @ lepton + jets mass analysis with the ‘D)LB discriminant for events poor in top signal, rich in top signal, and the final log-likelihood distribution.

3.2.2 Dilepton Top Mass Measurement

Due to the presence of two neutrinos, dilepton events do not contain enough information for a constrained fit. Therefore, to determine the top mass, one must use some other

-377-

II

Norm

aliz

ed P

rob.

/(5

CM/c

’)

Even

ts/(

10 G

&/c’

) 2

mo

- N

w c

03

4 1

,,I

8 -

,I,,

r ,,,

,I ,,I

7

7 .I/

reduce this background, events with at least one identified SVX b-jet are required to pass strict kinematic criteria that favor tf production and decay. To determine the top quark mass, full kinematic reconstruction is applied to the sample of events with six or more jets. All combinations are tried, with the constraint that an SVX-tagged jet must be assigned to a b-parton. The data sample consists of 136 events, of which 108 f 9 events are expected to come from background. The reconstructed three-jet mass distribution is shown in Fig. 15. The inset shows the shape of the difference log- likelihood as a function of top mass. With the systematic uncertainty shown in Table 5, the CDF measurementzg is:

Mop = 186.0 f lO.O(stat.) * 5.7(syst.) GeV/c’ (CDF - all hadronic)

Figure 15: Reconstructed mass distribution for all hadronic events with at least one b-tag. Also shown are the background distribution (shaded) and the tf Monte Carlo events added to background (hollow). The inset shows the log-likelihood and the fit used to determine the top mass.

3.2.4 Top Mass Summary

Figure 16(a) shows the summary of the direct measurements of the top quark mass at the Tevatron collider. When the appropriate correlations are taken into account between the mass systematic errors in CDF and D@, the world average is determined to be

Mt, = 174.3 f 5.1 GeV/c* (Tevatron average)

Figure 16(b) shows the relative weight of the various mass determinations on the final Tevatron average. When the top and W boson masses are interpreted in the frame- work of the SM, Fig. 17 shows the correlation between the top and the W masses as a function of the Higgs mass. With the latest values of the W and top masses and the LEP measurements3’ the data seems to favor a light Higgs, with mH < 260 GeV/c2 at 95% C.L.

(b) Relative weight in top mass average

-175.93~7.2 GeV/c2 Lep+kt -167.4f11.4 G&c Dil

-186.Qt11.5 GcV/c Had

-172.1f7.1 GcV/c* Comb.

_ 173.3f7.8 GeVk L.ep+Jet

168.4f12.8 G&/c Dil

BCDFltjets KDFallhad ElCDSdilepton too tzo 140 M O 180 *c?o 220

Top Mass (GIN& EJW dilepton I W Itjets

Figure 16: Summary of all the top quark mass measurements from (a) CDF and DQ, and (b) relative weight of the various single measurements in the overall average.

3.3 Other Top Quark Measurements

3.3.1 Measurement of /b&l

In the previous discussions, II&\ has been assumed to be approximately equal to 1. From the knowledge of the tagging efficiencies, the number of dilepton, and lepton +

jets events with one, two, or no jets tagged as b quark, the following ratio can be derived:

-379-

events (Fig. 18) CDF35 finds

F. = 0.55 f 0.32(stat.) f O.l2(syst.)

4 b Quark at the Tevatron

The principal motivation for studying b-quark physics in the context of the SM arises from the possibility of gathering valuable information on the CKM matrix elements. In fact, a study of b decays allows access to five of the nine CKM elements (V&, I$, Vtdr V,,,

and I&), some of which (I&, V,,) would be very hard to study in decays of the top quark system.

Traditionally, b-quark physics has been the domain of e+e- machines. However, UAl already demonstrated the possibility of studying b physics at a hadron collider. CDF, with a superb mass resolution and vertex detection capabilities, has really ex- panded the b-physics program achievable at a hadron collider. The D @ experiment has also published several b-physics results, 36 but due to the lack of precision momentum measurement of charged particles and the absence of a precision microvertex detector, D @ is not ideally suited to study the b sector with the same broad coverage.

A hadron machine has several advantages (and some disadvantages) compared to an e+e- machine at the ‘I (45’). All species of B hadrons can be produced at the Tevatron Collider (B+, B”, Bf, B,, and A*), with a large production cross section (gb N 50 pb,

while c~(~s) N 1 nb and uZo+bb N 7 nb at LEP). This very large cross section results in about 5 x 10’ bb pairs produced during Run I at the Tevatron detectors. Unfortu- nately, the inelastic cross section is three orders of magnitude larger, which puts very specific requirements on the trigger system designed to recognize b-hadrons for further processing. Moreover, the b-quark production cross section drops almost exponentially with the transverse momentum of the produced b quark. This puts the trigger threshold for b-physics events on a collision course with the experiment’s data acquisition (DAQ) bandwidth.

All b physics triggers at CDF and D(i? are based on leptons, with the possibility of requiring both single leptons and dileptons events. As an example, CDF dilepton trigger consists of a dimuon trigger with PT > 2 GeV/c for both muon legs, and an el trigger with PF > 3 GeV/c and Ef > 5 GeV. The dimuon trigger is the source of the J/i sample, and both dilepton triggers are used for b mixing analysis. The thresholds for single-lepton triggers are higher with PT > 7.5 GeV/c for muons and ET > 8 GeV

Longitudinal W Fraction

35 (CDF Preliminary)

Lepton + Jet Channel

Combined F, = 0.55

Lepton Pt (GeVlc)

Figure 18: The CDF lepton PT spectra for the sum of longitudinal W boson decays, transverse background.

-381-

a

defined as

ct* = L,,xM(J/$d) WJ/N ’

where L,, is the distance between the reconstructed decay vertex and the average beam position in the transverse plane.

From Monte Carlo studies (where Mn, is set to 6.27 GeV/c’) most signal events are expected to have 4 5 M(J/$d) 5 6 GeV/c*. To select possible decays of long lived particles, CDF requires ct* > 60pm. This cut is removed later for the lifetime measurement.

Starting with a sample of 196,000 J/g reconstructed in the SVX and then rejecting candidates compatible with being B+ + J/$K+ decays, and events where the electron was identified as coming from a photon conversion, CDF finds a sample of 31 J/t/d

candidates (19 J/tie, and 12 J/$p). The main sources of background are expected to be due to real J/$ which form a good displaced vertex when paired with a hadron misidentified as a third lepton, and to bb events with one B hadron decaying to a J/$

and the other B hadron decaying semileptonically, with a topology compatible with having the J/+ and the lepton in a common vertex. Hadrons misidentified as the third lepton are found to be the main source of background. For muons, this is due to light hadrons (pions or kaons) which punch through the calorimeter and are then detected in the muon chambers, or decays in flight producing a muon with a kink small enough to be well-linked to the track of the hadron. For electrons, this happens when the shower of a hadron in the electromagnetic calorimeter is indistinguishable from that of an electron. The contribution of these sources is estimated from a J/$J + track sample obtained by releasing the lepton identification criteria on the third track. This sample is then weighted with the probability, estimated from real data as a function of PT, that a hadron is misidentified as a lepton. With this method, CDF also obtains the mass shape of the background. Real J/+l background from bb events is estimated from a Monte Carlo simulation.

A summary of all the background sources and the estimated signals in both channels is given in Table 6.

The number of B, mesons and the statistical significance of the excess is also esti- mated from a likelihood fit of the J/$ mass distribution, as shown in Fig. 21(a). The mass shapes for the signal and background are constrained to the results of signal sim- ulation and background measurement, respectively. The only free parameter returned by the fit is the number of B, mesons, N(B,) = 20.4$~. The null hypothesis (i.e., the

CDF Preliminary 14 ,,m,, ,I,,,,

(4 (b) CDF Preliminary

I Jhy+e and J/yf+p

-Data (B, Candidates)

Calculated Signal 1

Calculated Background

f 8 lo-'

i d & a 3 6 5 ; 4

1

* lo-'

4 5 6 7 8 ^ . . M(J/~+lepton) (Gel7

Figure 21: (a) Mass distribution of B, candidates. The result of the fit for the B, signal and the measured background are superimposed. (b) Pseudolifetime distribution for data (crosses) with the result of the fit for signal (shaded histogram) and background (dashed line).

0 -1 -0.8 -0.6 -0.4 -0.2 0 0.2 0.4 0.6 0.8 1 -

Figure 22: Indirect constraints44 on the unitary triangle.

-383-

p can be measured by comparing the relative decay rates of B” and @to the com- mon CP eigenstate mode J/1/K:. By exploiting the interference between the direct decay path (B” + J/$Ki) and the mixed decay path (B” + B” --t J/1/K:), /3 can be measured through the time-dependent asymmetry

S'(t)-Be(t) ACP = - Byt)+BO(t) = sin(2/3) sin(Am,&),

where B’(t) and p(t) are the number of decays to J/$K,‘& at the time t, assuming that the meson produced at t = 0 was a B” or a B”, respectively. The effect of the mixing between B” and BO appears through the mass difference Am,, while the CP-phase difference between the two decay amplitudes appears via the factor sin(2P). Indirect evidence shown in Fig. 22 implies 0.30 < sin(2p) < 0.88 at 95% C.L., while OPAL recently reported45 sin(2P) = 3.2’;:: f 0.5 using the same decay channel.

During Run I, CDF collected approximately 200 B”, p + J/iKg decays. Al- though this sample is not sufficient to allow a precise measurement of sin(2/3), it is the largest reconstructed sample of Do, p + J/$Ki decays in the world, and it can be studied4’j to determine the feasibility of this measurement in a hadron collider. As for the B, discovery, the selection of B”, B” + J/$Ki candidates starts from the J/ii, + /1+/1- reconstruction. The muon tracks are required to be measured in the SVX detector, thereby obtaining a precise determination of the J/+ vertex. The other pairs of oppositely charged tracks in the event are then searched for those consistent with the Kg + 7r+r- decay hypothesis, where the Kg decay point is significantly displaced from the J/$ vertex. Each Ki candidate is then combined with the J/4 can-

didate in a four-particle fit which requires the Ki to point back to the J/$ vertex, and the combined J/$Ki system to point back at the primary vertex. The mass calculated by the fit has a typical resolution of 0‘~ N 9 MeV/c2. The proper decay length has a typical resolution of N 50 pm. Figure 23(a) shows the distribution of positive-lifetime candidate events as a function of the normalized mass 1M,v = (Af~rr - Mo)/o~lr, where MO is the central value of the B” mass peak (5.277 GeVlc’). A maximum likeli- hood fit yields 198 & 17 mesons. Since the CP asymmetry varies in time as sin(AmJ), it reaches its maximum close to a proper decay length of N 1000 pm, which is a region in which the background is strongly suppressed, as shown in Fig. 23(b).

4.2.1 Same-Side Flavor Tagging

Once the sample of B’s is obtained, the next step is to determine (“tag”) whether they were Be’s or Brj when they were produced.

50 45 (b) ct ) 200,m 1 40 i 1 35 30 25 20 i

Figure 23: Normalized mass distribution for (a) B” t J/$Ki candidates with ct 2 0, and (b) ct 2 200 pm. The curve is the Gaussian signal plus background from the full likelihood fit.

(b)

~\ ’ Candidate track ?TR

Figure 24: (a) A simple picture of b quarks hadronizing into B mesons, and (b) defini- tion of tagging candidate based on p?‘.

-385-

3

shown in Fig. 25(a) would be a cosine curve of amplitude one. An amplitude smaller than one is an indication of a “dilution” of the measurement. The dilution, Do, is related to the mistag probability and to the observed asymmetry by:

Aobs. CP = VO sin(28) sin(AmJ),

v 0

_ NRS - Nws

NRS + Nws = 1 - 2~nListag,

where NRS are events with the correct-sign correlation, and NWS events with the wrong-sign correlation. For this same-side tagging method, Do N 20%. This dilu- tion determination, measured on the data, is necessary for the extraction of sin(28).

4.2.2 Tagging Asymmetry

The SST technique tags approximately 65% of the B”, B” + J/$Kg decays. Figure 26 shows the sideband-subtracted asymmetry in bins of the proper decay-time, where the asymmetry is calculated by counting the sideband-subtracted number of positive tags, N+ , and negative tags, N-, in each proper decay-time bin:

A(ct) = N-(ct) - N+(d) N-(d) + N+(d)’

The signal and background samples were defined according to the dashed region shown in Fig. 23. The events in the signal region generally prefer negative tags (i.e., a positive asymmetry), whereas events in the sideband regions favor positive tags (nega- tive asymmetry). As noted before, however, the signal purity is high at large ct, and the sideband subtraction is, correspondingly, a small effect.

Two fits are shown in Fig. 26. The dashed curve gives the results of a simple x2 fit of the function A0 sin(Amdt) to the binned asymmetries, where Amd has been fixed to its 1996 world-averaged4Q value of 0.474 ps-‘. The solid curve is the result of an unbinned maximum likelihood fit which incorporates both signal and background distributions in mass and proper decay-time. Sideband and negative-lifetime events are included to help constrain the background distributions. The likelihood function also incorporates resolution effects and corrections for systematic biases, such as the small inherent charge asymmetry favoring positive tracks resulting from the wire plane orientation in the main CDF drift chamber. Clearly, the result is dominated by the sample size and not by the particular fitting procedure applied to the data. Also shown

CDF: B:/B:

Figure 26: The sideband-subtracted tagging structed J/$Ki proper decay length (points). discussed in the text, while the inset shows the

Figure 27: 95% confidence belts for the sin(2p) the actual measurement; (a) intersection with interval, and (b) the world summary of sin(2/3)

-387-

5.1 Potentials for Top Physics in Run II

The reach of top physics at the Tevatron during Run II has been studied.52 Each detector will record N 160 dilepton and N 1200 lepton+jets events, with N 500 of them double- tagged. For the cross-section measurement, the limiting factor will probably be the error on the luminosity, which can be measured to N 5% using the W + Iv rate. Therefore, the error on the production cross section is expected to be around S-10%.

The error on the mass measurement is presently dominated by statistics, but the uncertainty on the jet energy scale is a close second. As it is usual in the study of sys- tematic errors, the understanding of the energy scale will likely improve with improving statistics (by studying, for instance, the 2 + multijet events). The total uncertainty on A4rop in Run II should be at the level of 2 GeV/c’ (i.e., at the 1% level).

The precision on the R measurement should improve to 2%, corresponding to a 95% C.L. of IV,,1 2 0.20, while limits of B(t + 4~) < 3 x 10m3 and B(t + q.Z’) < 0.02 should be achievable.

From single top production, which is directly proportional to lV,#, the Tevatron experiments will be able to determine I&l* to N 10%. Other fields of study may arise from the large top quark mass and the expected (or, rather, hopedfor) connection with the symmetry breaking mechanism of the Standard Model.

5.2 Potentials for B Physics in Run II

The goal of the B physics community in the next decade is the measurement of the unitary triangle parameters. Therefore, we will concentrate on the expectations for this kind of physics, even though it is expected that the Tevatron experiments will continue exploring the “standard” b physics avenues, like lifetimes, B” and B, mixing, and heavy states searches.

Studies of CP violation will concentrate on the measurement of sin(2p) and sin(2a) in B” + J/$K: and B” + rr+7~- decays. CDF has demonstrated that the Tevatron Collider environment allows these kind of measurements. Moreover, CDF plans to also look for CP violation in B, + D,K and B + DK, probing sin(2y).

As shown in the discussion of the sin(2/3) measurements, the important ingredients for the CP measurements are the efficiency E to tag the original b flavor and the dilu- tion D in the tagging. These ingredients are combined in a figure of merit (ED’) that compares different tagging methods, since the error on the determination of a given CP asymmetry is proportional to l/m, where N is the number of reconstructed

decays. The e2)* for the three tagging methods discussed in the CP asymmetry mea- surements at CDF are shown in Table 7.

Table 7: Summary of the dilepton counting experiments.

Based on the experience gathered in Run I, CDF expects to reconstruct -10,000 B” + J/$Ki decays in Run II. This takes into account the increased luminosity, the extended silicon vertex detector, and the upgraded muon detector. This expectation is conservative because it does not include the possibility of using the J/lc, + e+e-

decay channel, which will be possible in Run II with a contribution of N 5000 more reconstructed B” decays. The asymmetry and its error are defined as:

A = ZDsin2/3,

a(sin2/3) = 44 Q) F 63 (sin 2p)-,

where @ means addition in quadrature. Using the values determined in the Run I aual- ysis and extrapolating according to Table 7, CDF expects to measure sin(2/3) with an error of c7(sin 2p) = 0.071 @ 0.044 = 0.08.

The identification and reconstruction of B + K+K, for the determination of sin(2cY), is very challenging at hadronic colliders because of the very low branching ratio (- 10e5) and the huge combinatorial background from the low PT charged tracks produced in pp collisions. CDF made an effort to study the trigger and the background rejection. A fully dedicated trigger, based on the idea of selecting tracks with large impact parameter (do > 100pm) is thought to be sufficient to limit the event rate. As- suming 10,000 events in 2 fb’, tV2 = 5.4% and a signal-to-background ratio of one to four, CDF expects to measure the asymmetry with an error of 0~ = 0.12.

-389-

b