Elizaveta Shabalina

1
Top Quark Physics at the Tevatron
From discovery to precision measurements

• Elizaveta Shabalina
• University of Illinois (Chicago)

2
Motivations for Studying Top

• Special place in the Standard Model (SM)
• Only known fermion with a mass at the natural
  electroweak scale (40 times larger than its
  isospin partner, b-quark)
• Large Yukawa coupling to Higgs boson (Gt1)
• Special role in precision electroweak physics
• Window into the problem of ElectoWeak Symmetry
  Breaking (EWSB)?
• Top quark mass sets severe constraints on SM
  extensions
• New physics may appear in production (e.g.
  topcolor) or in decay (e.g. charged Higgs).
• Can only be studied at Tevatron prior to LHC.

3
A Brief Top History
Top quark was expected in the SM of electroweak
interactions as a partner of b-quark in SU(2)
doublet of weak isospin for the third family of
quarks

• Observed by CDF and D0 in 1995 in first 70 pb-1
  of Run I data.
• Final Run I top analyses based on 110 pb-1.
• Production cross sections in many channels
• Mass 174.3 ? 5.1 GeV (CDF/DØ combined)
• Event kinematics
• W helicity measurement
• Limits on single top production, rare/non-SM
  decays.

Overall consistency with the Standard Model. But
all our knowledge is based on 100 analyzable top
events ? analyses statistics-limited
4
Tevatron collider in Run II

• The Tevatron is a proton-antiproton collider with
  980 GeV/beam
• 36 p and p bunches ?396 ns between bunch crossing
  
• Increased from 6x6 bunches with 3.5ms in Run I
• Increased instantaneous luminosity
• Run II goal 30 x 1031 cm2 s-1
• Current 45 x 1031 cm2 s-1

5
Tevatron Peak Luminosity
Typical recent stores 4-5?1031
7?1031
Run IIa goal 8?1031
6
Tevaton performance
Performance above design/base
Integrated Luminosity (fb-1)
-
Start of Fiscal Year
7
Integrated luminosity at D0
delivered
403 pb-1
on tape
314 pb-1
Data collection efficiency 90
Start of physics-quality data spring 2002

• Most of the results presented today are obtained
  with 120?200 pb-1
• Goal for 2004 additional 310-380 pb-1 delivered.

8
The upgraded detectors
D0
CDF

• New tracking silicon and fibers in 2 T magnetic
  field
• Upgraded muon system
• Upgraded DAQ/trigger (displaced track soon

• New bigger silicon, new faster drift chambers,
  TOF
• New scintillating tile and end-plug calorimeter
  and increased coverage of the muon system
• Upgraded DAQ/trigger,. displaced-track trigger

9
Top physics in Run II

• Run II
• with high precision we hope to answer
  questions such as
• Why is top so heavy ?
• Is it or the third generation special ?
• Is top involved with EWSB ?
• Is it connected to new physics ?

• New results
• Top quark pair production cross section
• Single top quark production
• new top mass world average

(FCNC)
10
Top Quark production at the Tevatron
B(t?Wb) 100

• Cross section ? basic number, starting point for
  all top physics
• In Run II (1.96 TeV) is expected to be 30
  higher than in Run I (1.8 TeV)

85
15

• Requires detailed understanding of detector,
  backgrounds and selection efficiencies

11
Production cross section

• Run I measurement
• CDF
• D0

• Test of QCD
• Latest calculations (NNLO CTEQ6M parton
  densities)
• Departures from QCD could indicate non SM
  physics
• decays of SUSY states
• Top-color objects

• Current uncertainty is statistics limited

RunII (2 fb-1)
12
Top quark decays
44.4
21.1
t

• W decay modes determine top quark final state
• Dilepton (ee, µµ, eµ)
• Both Ws decay leptonically
• BR 5
• Lepton (e or µ) jets
• One W decays leptonically, another one
  hadronically
• BR 30
• All-hadronic
• Both Ws decay hadronically
• BR 44
• thad X
• BR 23

All jets
mjet
ejet
14.8

14.8
1.2 1.2
2.5
Most favorable channels for top physics
More challenging backgrounds, but measurements
still possible
13
Top cross section dilepton channels

• Event selection
• 2 high PT isolated charged leptons (e,m)
• Neutrinos large missing ET
• At least 2 jets (b-jets)
• Large scalar sum of all measured objects ETs
  (leptons, jets)
• Backgrounds
• Physics determined from Monte Carlo
• WW/WZ/ZZ, Z?tt
• Instrumental determined from data
• fake leptons in Wjets and QCD
• Drell Yan (Z/g? ee,mm) with fake missing ET
  (eµ is not affected)

14
D0 ee, eµ, µµ final states
156 pb-1
143 pb-1
140 pb-1
Top contribution is calculated assuming
15
D0 ee, eµ, µµ final states
13.1?4.75.9 (stat)?1.71.7 (sys) ? 0.9 (lumi)
pb
19.1?9.613.0 (stat) ?2.63.1 (sys) ? 1.2(lumi)
pb
11.7?14.119.7 (stat)?8.24.1(sys) ? 0.8 (lumi)
pb
14.3?4.35.1 (stat)?2.0 1.9 (sys) ? 0.9 (lumi)
pb
16
D0 dilepton channels eµ candidate
Soft lepton tag
Isolated muon
Secondary vertex
17
CDF ee, eµ, µµ final states

• Events with 1 tight and 1 loose e or µ 1ee,
  3 µµ, 9eµ

?tt 8.73.9(stat) ? 1.4(sys) ? 0.5(lumi) pb

-2.6

• Events with 2 tight leptons
  1ee, 2 µµ, 4eµ

?tt 8.14.4(stat) ? 1.6(sys) ? 0.5(lumi) pb
-3.4
18
CDF leptontrack final states

• Require 1 isolated lepton and 1 isolated track
  (instead of isolated lepton)
• 20 higher acceptance but lower S/B
• different background composition

Measured cross-section for different jet ET and
track pT
?tt 6.92.7 (stat) ? 1.2(sys) ? 0.4(lumi) pb
?2.4
19
Top cross section leptonjets channels

• Event preselection
• 1 high PT isolated charge
• lepton (e,m)
• Neutrinos large missing ET
• Large jet multiplicity
• Backgrounds
• Wjets and multijet production (QCD) with fake
  isolated lepton or fake missing ET
• Features
• Larger yield but higher background
• Additional techniques
• make use of event topology
• tag b jets

jet
?
p
b
jet
jet
jet
t(?Wb) t(?Wb)
e,m
qq
20
Leptonjets channels overview
21
Leptonjets topological variables

• Topological variables to separate signal from
  background
• Minimize systematic uncertainty by reducing
  sensitivity to Jet Energy Scale
• Small correlations
• Angular quantities
• sphericity
• aplanarity
• Ratios of energy-dependent quantities
• HT2
• KTmin

Sphericity
Aplanarity
HT2/HZ
KTmin
22
D0 leptonjets topological cross section
23
D0 leptonjets topological cross section
8.8 4.1 (stat) 1.6 (sys) ? 0.57(lumi) pb
-2.1
-3.7
?3.7
?2.1
6.03.4 (stat) 1.6 (sys) ? 0.39(lumi) pb
?3.0
?1.6
7.22.6 (stat) 1.6 (sys) ? 0.47(lumi) pb
?1.7
?2.4
24
CDF leptonjets topological cross section

• Combine 7 kinematic variables in Neural Network
  to separate signal from background

Nj?4
Nj?3
25
CDF leptonjets with Soft Lepton Tag

• Soft Lepton Tag
• Exploits the b quarks semi-leptonic decays
• These leptons are soft and not isolated
• Probability to tag a event 15, fake rate
  (QCD jet) 3.6

• same preselection as topological analysis
• 3 jets

Cross section
?tt 4.14.0 (stat) ? 1.9(sys) pb
126 pb-1
?2.8

• D0 has presented this analysis with 92 pb-1 to
  be updated soon with 150 pb-1

26
Top cross section lifetime b-tagging

• jet is tagged as b jet
• If signed decay length significance Lxy/?(Lxy)gtcut

Secondary Vertex Tag (SVT)
Counting Signed Impact Parameter tag (CSIP D0)

• S IP/?(IP)
• Jet is positively tagged if it has
• at least two tracks with Sgt3 or
• at least three tracks with Sgt2

• Signature of a b decay is a displaced vertex
• Long lifetime of b-hadrons (c? 450µ) boost
• B hadrons travel Lxy 3mm before decay with
  large charge track multiplicity

• b-tagging at hadron machines is proved to be
• crucial for top discovery in Run I
• essential for Run II physics program

27
B-tagging analysis at D0
after tagging
before tagging
W njets
N bckg tag
lepton MET njets
from data
QCD
cross section is determined from the excess
of the observed number of tagged events w.r.t.
the predicted background for njets3
28
D0 signal and background summary

• Event tagging probability requires
• b-tagging efficiency
• c-tagging efficiency
• light jet tagging efficiency (mistag rate)

• Performance highlights
• tt tagging probability 56
• Single Top (s channel) 52
• Wbb 52
• Wj 0.3 (mistag rate)

ejets, 2 tag
ejets, 1 tag
Number of events
Number of events
excess
29
D0 signal and background summary
lepton jets
1 tag
2 tags
tt contribution is calculated assuming ? 7 pb
30
CDF cross section with b-tagging

• event b-tagging efficiency 55, fake tag
  rate (QCD jets) 0.5

162 pb-1
162 pb-1
?tt 5.61.2(stat) 1.0(sys) pb
?1.0
?0.7
31
CDF cross section with b-tagging and kinematic
fits

• Use event shape information in tagged events to
  further separate signal from background
• Examine various kinematic quantities leading jet
  ET, subleading jet ET, their sum

• Fit templates to extract signal and background
  fractions
• Signal templates from tt MC, background templates
  from data W3jets events without b-tag, QCD
  contribution from non-isolated lepton

Wbb
tt
?tt 6.01.5(stat) ? 0.8 (sys) pb
-1.8
?1.8
32
D0 all jets channel with SVT

• Derive SVT tag rate function in the same multijet
  events. Apply to untagged sample to predict
  background shape

• Final state 6 jets, 2 b-quark jets
• Overwhelmed by QCD multijet background
  impossible to extract signal without tagging
  b-jet(s)
• Three NNs combine various kinematic variables
  apply successive cuts on the outputs

33
All jets channel discriminating variables
Variables are designed to address different
aspects of the background

• Energy Scale HT, ?s
• Soft non-leading Jets H3jT, ET5,6,ltNjgt
• Event Shape sphericity, aplanarity

• Rapidity Centrality, lth2gt
• Top Properties Top and W Mass Likelihood, MWW,
  Mtt, min dijet masses

34
D0 all jets channel cross section
220 events pass all cuts Expect 186 5(stat)
7.5(sys)
0.75
35
CDF all jets channel cross section
36
Top pair production cross section summary
37
Electroweak production of the top quark

• Access to Wtb vertex
• Only vertex which can be sensitively probed at
  hadron colliders
• Measure Vtb directly, test CKM unitarity
• Test V?A structure of the Wtb vertex

• Not observed yet, despite the expected
• large rate (sst 40 stt )
• Existing Run I upper limits (_at_ 95 CL)
• CDF ss lt 18 pb, st lt 13 pb, sst lt 14 pb
• DØ ss lt 17 pb, st lt 22 pb

38
Search for single top quark production

• Experimental signature similar to tt?ljets
• Same signature as SM Higgs associated production
  
• W2 jets bin!
• Single top samples have less objects in the final
  state
• larger background

• Backgrounds Wjets, tt, multijets (misidentified
  leptons), diboson
• Experimental strategy
• Optimize event preselection to maximize signal
  acceptance while reducing backgrounds/misreconstru
  cted events as much as possible.
• b-tagging is extremely important.
• Make use of topological information to further
  discriminate signal from background

39
CDF single top quark search

• Event preselection
• 1 high PT isolated charge lepton (e,m)
• large missing ET
• two jets, at least one with a b-tag
• Topological selection
• 140 GeV lt Mlnb lt 210 GeV
• Leading jet pT lt 30 GeV (only t-channel search
• Good agreement between observed number of events
  and predicted sum of background and signal
• Consider discriminating variables
• Combined (st channels) search HT pT
  (lepton)MET? pT(jet)
• t-channel search Q(lepton)?? (untagged jet)

40
CDF single top quark search

• Maximum likelihood fit to data HT or Q??
  distributions using a sum of templates determined
  from MC single top (PYTHIA), tt (HERWIG),
  non-top Wbb (ALPGEN)

Fitted signal content compatible with zero
st lt 8.5 pb _at_ 95 CL
sst lt 13.7 pb _at_ 95 CL
L162 pb-1
41
D0 single top quark search

• Event preselection
• 1 high PT isolated charge lepton (e,m)
• large missing ET
• 2lt jetslt4, leading jet pT lt 25 GeV, other pT lt
  15 GeV
• At least 1 b-tag SVT or soft muon tag
• No topologocal selection applied yet large room
  for improvement
• 4 orthogonal analyses
• ejets SVT, soft muon tag
• µjets SVT, soft muon tag

• Background estimates
• Multijets (QCD) from data
• Wjets from data using measured inclusive tag
  rate functions on a multijets sample.
• tt from ALPGEN MC

mjets/SVT
42
D0 single top quark search
Expected cross-section limits _at_95 CL from
combination of ejets and µjets for SVT and SLT
analysis channels

• Sensitivity already better than Run I butthis
  is just half of the analysis!!
• Significantly increased statistical sensitivity
  expected from use of topological information

43
The Top Quark Mass

• Fundamental SM parameter
• needed to determine ttH coupling
• Enters as a parameter in the calculation of
  radiative corrections to other SM observbles
• Experimental handles
• B tagging reduce background combinatorial
• Data driven systematics scale with 1/?N (energy
  scale, gluon radiation)

CDF/D0 2 fb-1goal!
With 2 fb-1 data (Run II) ?Mt 3
constrain ?Mh/Mh to 35
44
D0 Run I top quark mass update

• Old approach
• Lowest ?2 solution from kinematic fit
• Topological discriminant is used to separate
  signal from background
• Mass estimate is made by 2D fit in fitted mass
  and descriminant and compared to MC template
• New approach
• Calculate event probability to be a tt event
• Consider all combinations, assign each a weight
  and sum their probabilities
• Increase purity by cutting on background
  probability 22 events (out of 71 4 jet events
  survive)

Mtop 180.1 3.6 (stat) 4.0 (syst) GeV
Improvement in statistocal error is equivalent to
2.4 times more data
New Run I D0 top mass average Mt 179.0 5.1
GeV
45
New top quark mass average

• Two updates
• New D0 Run I ljets mass measurement
• Latest CDF results (PRD63 032003, 2001)

New Tevatron CDFD0 result Mtop178 4.1
GeV Breakdown of errors 2.7 statistical, 3.3
systematical error
2.6 jet energy scale 1.65 signal modelling 0.12
Monte Carlo 0.88 background
0.83 U-noise and multiple interactions 0.35 fit
46
Consequences for SM Higgs
47
Conclusions and Outlook

• The top quark is back!
• First Run II results cover a variety of channels
  and topics and are improving rapidly. All
  measurements consistent with the Standard Model
  but still statistics limited
• The Tevatron is the unique top quark factory
  until LHC
• We still expect at least 50x more data compared
  to Run I!

It is the start of a program of precision top
physics ? and hopefully top surprises ? at the
Tevatron
48
Backup slides
49
New Run I Top Mass (D0)
Template Method
Optimized Method (ME)

• All the events are presented to the same
  template.
• The template corresponds to a probability
  distribution for the entire sample, using several
  variables calculated from MC simulations.
• The features of individual events are integrated
  (averaged) over the variables not considered in
  the template.

• Each event has its own probability distribution.
• The probability depends on all measured
  quantities (except for unclustered energy).
  
• The full information contained in each event is
  contributed to the probability well-measured
  events contribute more than poorly-measured
  events.

50
Matrix Element Method
W(y,x) is the probability that a parton level
set of variables y will be measured as a set of
variables x
dns is the differential cross section Matrix
element
? Mtop is to be estimated
Probability of measuring the event
f(q) is the probability distribution that a
parton will have a momentum q
Measured, reconstructed objects in the event

• Leading-Order ttbar-gtleptonjets matrix element,
  PDFs
• 12 jet permutations, all values of P(?)
• Phase space of 6-object final state
• Detector resolutions

• Only Wjets, 80
• VECBOS subroutines for Wjets
• Same detector resolutions as for signal
• All permutations, all values of P(?)
• Integration done over the jet energies

• Convolute probability to include all conditions
  for accepting or rejecting an event
• Form a Likelihood as a function of Top Mass, F0
  (longitudinal fraction of W bosons)