Sandra Leone, INFN PisaLHC School Lecce June 2006
Top Physics at CDF Top Physics at CDF ……(while waiting for the LHC)(while waiting for the LHC)
Sandra LeoneSandra LeoneINFN Pisa
Italo-Hellenic School of Physics 2006The Physics of LHC: Theoretical Tools and Experimental Challenges
Martignano, Lecce, June 12-18, 2006
Sandra Leone, INFN PisaLHC School Lecce June 2006
OutlineOutline
� Motivations for studying top � A brief history� Top production and decay� The Tevatron & CDF� Detector issues� Identification of final states� Cross section measurement� Single top production� Mass determination� Study of top properties� The lesson for LHC
bscdu
t
Sandra Leone, INFN PisaLHC School Lecce June 2006
Motivations for Studying TopMotivations for Studying Top
� Only known fermion with a mass at the natural electroweak scale.
� Similar mass to tungsten atomic # 74 35 times heavier than b quark.⇒Why is Top so heavy?⇒Is top involved in EWSB?
�(Does (2 √ 2 GF)-1/2 ≈ Mtop mean anything?)⇒Special role in precision electroweak physics?⇒Is top, or the third generation, special?
� New physics BSM may appear in production (e.g. topcolor) or in decay (e.g. Charged Higgs).
Sandra Leone, INFN PisaLHC School Lecce June 2006
A Brief History of TopA Brief History of Top� Top quark was expected in the
Standard Model (SM) of electroweak interactions as a partner of b-quark in SU(2) doublet of weak isospin for the third family of quarks
(weak isospin of b can be inferred from the forward-backward asymmetry in e+e- � bb)
� Anomaly free SM requires the sum of the family charges to be zero: given the b (and the tau lepton) there should be a 2/3 charge quark
• 1977-1994: increasing lower top mass limits
Sandra Leone, INFN PisaLHC School Lecce June 2006
A Brief History of TopA Brief History of Top� First top evidence in 1994 in CDF
data, 19 pb-1, 15 events on a background of 6, ⇒ 2.8 σ excess, not enough to claim
discovery
� Confirmed in 1995 by CDF and D0 in first ~70 pb-1 of run 1 data (4.8 σ ).
� Final Tevatron Run 1 top analyses based on ~110 pb-1.
⇒Production σσσσ in many channels.⇒Mass: 178.0 ± 4.3 GeV (CDF/DØ) ⇒Study of several aspects of
event kinematics.⇒Limits on single top production,
rare/non-SM decays.� Overall consistency with the SM � But only ~100 analyzable top events
⇒analyses statistics-limited.
Sandra Leone, INFN PisaLHC School Lecce June 2006
Why Why TevatronTevatron Top results in a school on LHC Top results in a school on LHC physics?physics?
� Tevatron is the only place until the LHC turns on, to study the top quark.
�LHC will be a top factory, producing ∼ 2800 ttbar events per hour at low luminosity (1033)
�Top quark will be one of the major subject of study at LHC
�Moreover will be a background to many other processes -> it will be important to have it under control
�It’s not easy: looking for relatively rare events -> accurate determination of background necessary
�The LHC environment will be a difficult one as well
Sandra Leone, INFN PisaLHC School Lecce June 2006
A more cultural perspective of TopA more cultural perspective of Top……
About 5 orders of magnitude range in quark masses!
�Top analyses as playground to learn how to optimize the cuts to reduce the bckgnd contamination, optimize the simulation, and realize that you’ll have to use real data in order to study all the instrumental effects
LHC?⊙
Sandra Leone, INFN PisaLHC School Lecce June 2006
Production and Decay BasicsProduction and Decay Basics
85%
15%
NB: qq, gg fractions reversed at LHC
t W+
b
tW
-
b
SM predicts: BR( t → Wb) ≈ 100%
σtheory ≈ 7 pb
Event topology
determined by the decay
modes of the 2 W’s (W+W-) in final state
b-jet: identify via secondary vertex or soft lepton tag
At the Tevatron top quarks are mainly pair produced via
strong interaction:
Sandra Leone, INFN PisaLHC School Lecce June 2006
Top Decay : add. Motivation for Studying TopTop Decay : add. Motivation for Studying Top
� In the SM, assuming V-A coupling with a CKM matrix parameter
|Vtb| = 1 for the t → bW decay vertex, one gets (LO):
Γ(t → bW ) ≈ 175 MeV (MT/MW)2 (MT,MW >> Mb)
⇒⇒⇒⇒ Γ(t → bW ) ≈ 1.5 GeV ⇒⇒⇒⇒ τ(top) ≈ 4 x 10-25 s� Non-perturbative QCD hadronization takes place in a time of order:
Λ-1QCD ∼ (100 MeV)-1 ∼ 10-23 s
⇒ top decays before hadronizing, as free quark (no top hadrons, no toponium spectroscopy)
⇒the top quark provides the first opportunity to study the decay characteristics of a “bare” quark.
� t → Ws, t → Wd allowed but suppressed by factors of ∼ 10-3 and ∼ 5 x 10-5 respectively
Sandra Leone, INFN PisaLHC School Lecce June 2006
tt--ttbarbar Final StatesFinal States� Dilepton (ee, µµ, eµ)
⇒BR = 5%⇒2 high-PT leptons + 2 b-jets + large missing-ET
� Lepton (e or µ) + jets⇒BR = 30%⇒single lepton + 4 jets (2 from b’s) + missing-ET
� All-hadronic⇒BR = 44%⇒six jets, no missing-ET
� τhad +X⇒BR = 21%
Most favorable channels for top
physics
More challenging backgrounds, but
measurements still possible
e-e (1/81)
mu-mu (1/81)
tau-tau (1/81)
e -mu (2/81)
e -tau (2/81)
mu-tau (2/81)
e+jets (12/81)
mu+jets (12/81)
tau+jets (12/81)
jets (36/81)Dilepton
L+Jets
Sandra Leone, INFN PisaLHC School Lecce June 2006
The The TevatronTevatronRun II: √s = 1.96 TeV
36 p and pbar bunches �396 ns between bunch crossing
Started in spring 2001
The Fermilab Tevatron has been the only place, and will be until the LHC turns on, to study the top quark.
Sandra Leone, INFN PisaLHC School Lecce June 2006
Tevatron Peak LuminosityTevatron Peak Luminosity
In may 2004 Typical stores: 5 – 6 x1031
1.6 x 1032
Peak Luminosities above 1 x 1032 cm-2 s-1 are common
Record: initial lum ~ 1.8 x 1032 cm-2 s-1 on 11/10/2005After a commissioning period, data “good for physics’’ since Feb. 2002
Sandra Leone, INFN PisaLHC School Lecce June 2006
A Brief History of CDFA Brief History of CDF
� 1985: First collisions with partial detector
� 1987: Core detector in place. Jet Physics
� 1988-89: “Run 0” 4x the expected data, seen lots of W/Z’s
� 1992-1995 : “Run I” -added silicon detector. Top quark discovered!
� 2001-present: Run II begins with essentially a new detector, higher collisions energy and more data.
� 2004: First Run II physics papers published
Sandra Leone, INFN PisaLHC School Lecce June 2006
The CDF Run II DetectorThe CDF Run II Detector– new front-end,DAQ
– trigger• new L1 track trigger
• new L2 secondary vertex tracker (SVT)
– new silicon tracker• 7 layers |η|<2
• 3-D reconstruction
– new central tracker• Nstereo=Naxial=48
� ∆∆∆∆pt/pt<0.001pt
– Time of flight
– new forward calorimeter
– µ,e id to |η|η|η|η|=2
Inherited from Run I:• Central Calorimeter• Muon System (some new)• Solenoid
Sandra Leone, INFN PisaLHC School Lecce June 2006
CDF Run II DetectorCDF Run II Detector
Central tracking
|η| < 1.0
Muon Chambers
|η| < 1.5
Central+Plug
Calorimetry
|η| < 3.6
Silicon tracking
|η| < 2.0
10m
Sandra Leone, INFN PisaLHC School Lecce June 2006
CDF Integrated LuminosityCDF Integrated Luminositydelivered ~1.6 fb-1
on tape ~1.3 fb-1
Taking data efficiency L(recorded)/L(delivered)
commonly > 85%
Running stably since Feb. 02 Physics on 0.9 – 1.2 fb-1
Start of physics-quality
data
Error on luminosity is ± 6%�1.6% due to CerenkovLuminosityCounter systematic error on rate�4.0% due to CLC acceptance sys�3.8% due to limited knowledge of pp inelastic cross section.
Sandra Leone, INFN PisaLHC School Lecce June 2006
Principles for Particle IdentificationPrinciples for Particle Identification
Beam direction
Sandra Leone, INFN PisaLHC School Lecce June 2006
Principles for Particle IdentificationPrinciples for Particle Identification
Corresponding energy deposition in calorimeter consistent with MIP
Extrapolated track matching signal in muon chambers
Corresponding energy deposition in EM calorimeter
Small Had/EM
shower shape consistent with that expected for an electron
A track in the Central Outer Chamber
A track in the Central Outer Chamber
Muons: (|η| < 1)Electrons: (|η| < 2.8)
Sandra Leone, INFN PisaLHC School Lecce June 2006
Principles for Particle IdentificationPrinciples for Particle Identification
No associated track
Projective tower geometry.
Fixed cone algorithm in η−φ, ∆∆∆∆R = 0.4
Missing Transverse energy:
ET = Σi ETi • ni
ET = - ET
Energy deposition in em calorimeter
Energy deposition in em and had calorimeter
No interaction seen in the detector
Photon:Jet: Neutrino:
Sandra Leone, INFN PisaLHC School Lecce June 2006
More details on JetsMore details on Jets
A jet is a complicated object:
�measured by calorimeter towers �defined by a clustering algorithm
�Doing analysis with jets requires the energy of the measured jet to be converted to the energy of the parent parton or particle jet.
�From measured jet energies to parton energies we need to correct for:
�Instrumental effects�Physics effects�Jet Algorithm effects
Sandra Leone, INFN PisaLHC School Lecce June 2006
“Physics at a Physics at a hadronhadron collidercollider……
…Is all about the trigger!” High pT leptonHigh ET jet, photonHigh Missing ET (MET)
Examine each pp collision Select few interesting events (<70 Hz)
Store for further offline analysis Keep 1 out of 25,000
0.0000015 Hz15 fbpp→WH→ℓνbb (if MH=120GeV)
0.0002 Hz2 pbpp→tt→WWbb→ℓνbbX
0.04 Hz0.5 nbpp→ZX→ℓℓX
0.4 Hz5 nbpp→WX→ℓνX
1 kHz10 µbpp→bb (b pT>6 GeV, |η|<1)
6 MHz60 mbInelastic pp
Event RateCross-sectionProcess
Assume L =100x1030 cm-2s-1, ℓ=electron or muon
Sandra Leone, INFN PisaLHC School Lecce June 2006
Comparison of Cross SectionsComparison of Cross Sections
� In 1 fb-1 of data, ~ 1.4 x 1014 total collisions, one in every 1010 producing a ttbarevent (c.f. one every 2.5 x 106 producing a W event)
Note: LHC will be a top factory, producing ∼ 2800 ttbar events per hour at low luminosity (1033)
with 1 fb-1
1.4 x 1014
1 x 1011
6 x 106
6 x 105
14,0005,000
1000 ~ 100100 ~ 10
σσσσ Tevatron @ NLO: 6.7 ± 1.2 pb
σσσσ LHC @ NLO: ~830 ± 100 pb
Sandra Leone, INFN PisaLHC School Lecce June 2006
Method to identify a top signalMethod to identify a top signal� Start from “counting events” passing cuts in all decay
channels
⇒Optimization of signal region with respect to SM background processes (control region)
� Background dominates the production of ttbar pairs by several orders of magnitude
� How to separate signal from background:
⇒Top events have very distinctive signatures
�Decay products (leptons, neut., jets) have large pT’s
�Event topology: central and spherical
�Heavy flavor content: always 2 b jets in the final state!
Sandra Leone, INFN PisaLHC School Lecce June 2006
DileptonDilepton ChannelChannel
ℓ
ℓ
jet
jet
b
b
νννν
νννν
p p
ET
+
-
jet
jet
b
b
νννν
νννν
pp p
ET
ET
Signature:� Two high pT Isolated leptons
opposite sign
� Veto Z, cosmic, conversion
� ∆Φ(ET,l/j)>20o, or ET>50 GeV
� ET > 25
� Two jets with ET>10 GeV
� Total ET > 200 GeV (HT = Scalar summed ET of jets, leptons, and ET )
Expect: S/B ~ 9
Dominant backgrounds: Drell-Yan, W+jets (“fakes”)
Relatively clean:
�Top and a small amount of SM bkgds
�Down side is small event samples
Sandra Leone, INFN PisaLHC School Lecce June 2006
DileptonDilepton Channel: Channel: eeee, , eeµµµµµµµµ, , µµ (750 µµ (750 µµ (750 µµ (750 µµ (750 µµ (750 µµ (750 µµ (750 pbpb−−11))))))))
64 events on 19 ± 4 background12 ee, 24 µµ, 28 eµ ttbar
signal bin
Jet MultiplicityPT
Sandra Leone, INFN PisaLHC School Lecce June 2006
DileptonDilepton KinematicsKinematicsHt:Scalar summed ET of jets,
leptons, and missing ET
Leptons transverse energy
With higher statistics in Run IIData follow SM expected
distribution of top + bkgd
RunI: had seen hints of discrepancy in kinematic distributions
Missing ET
Sandra Leone, INFN PisaLHC School Lecce June 2006
DileptonDilepton background overviewbackground overview�Instrumental backgrounds
�Drell-Yan (ee, µµ)
�False ET from mismeasured leptons, jets
�Fake leptons
�W+jets with jet misidentified as lepton
�Use data whenever possible
�Physics backgrounds
�Diboson (WW/WZ/ZZ) and Z � ττττττττ
�Real leptons, ET, jets
�Evaluate using MC
�Determine bkgd in 0j, 1j bins to give confidence in signal bin prediction
Sandra Leone, INFN PisaLHC School Lecce June 2006
DileptonDilepton: DY (: DY (eeee, , µµµµ) background) background• Large cross section but no intrinsic
ET
� False Et
� Detector coverage isn’t 4π� Reconstruction isn’t perfect
� Tails of ET resolution critical� Simulation doesn’t accurately model
this� Extra cuts in “Z window”� Estimate residual contamination:
� Use loose Z data to normalize (subtract expected non-Z’s)
� Use MC to distribute inside -outside Z window, across jet bins
ET > 25
ET > 25
Sandra Leone, INFN PisaLHC School Lecce June 2006
DileptonDilepton: Fake lepton backgrounds: Fake lepton backgrounds
114 +/- 3169j100
65 +/- 975j70
49 +/- 651j20
PredObsvDIL
� Determine lepton fake rate from jets in j50 sample
� Cross-check fake rates in other samples with jets
� Apply fake rates to jets in W+jets data sample
ET of observed (predicted) fake tracks in green (black)
# o
bserv
ed (
pre
dic
ted)
fakes
Jet20
Jet70
Sandra Leone, INFN PisaLHC School Lecce June 2006
Dilepton:ZDilepton:Z --> > ττττ, , dibosondiboson bkgdsbkgds� In both cases:
⇒Real missing energy⇒Jets from decays or
initial/final state radiation
� Estimates derived using PYTHIA, ALPGEN+HERWIG MC, normalized to theoretical xsecs
� Correct for underestimation of extra jets in MC⇒Determine jet bin reweighting
factors for Z � ττ from Z�ee, µµ data
⇒Reweight WW similarly
Sandra Leone, INFN PisaLHC School Lecce June 2006
Why Measure the ttbar Cross Section?Why Measure the ttbar Cross Section?� Basic engineering number, absolute measurement (→ very
difficult!) , prior step to any top property study.
� Requires detailed understanding of backgrounds and selection efficiencies.
� Test of SM
⇒Departures from QCD prediction could indicate nonstandard production mechanisms, i.e. production through decays of SUSY states.
∫××
−=
LdtA
NNtt backobs
εσ )(
_
Sandra Leone, INFN PisaLHC School Lecce June 2006
DileptonDilepton: : tttt acceptance and efficiencies acceptance and efficiencies
Acceptance:� Determine from PYTHIA MC
(mt=175 GeV)� Apply trigger efficiencies, lepton
ID MC correction factors, luminosity weights for different detector categories
22%
25%53%
Lepton efficiencies:� Determined in Z -> ℓ ℓ using second leg of Z’s
� Get efficiency for data and MC, estimate difference and derive scale factor ( 90%, 79%, 83% for ee, µµ, eµ respectively)
Final acceptance x eff = 0.732 %
Sandra Leone, INFN PisaLHC School Lecce June 2006
Dilepton:SignalDilepton:Signal acceptance acceptance systematicssystematics
2.4Monte Carlo Generators- compare acceptance of PYTHIA to HERWIG
4.4Initial- and Final-state radiation- ISR: difference from no-ISR sample
- FSR: parton-matching method, different PYTHIA tune
1Parton Distribution Functions (PDF’s)- default CTEQ5L vs MRST PDF’s, different αs samples
3.1Jet Energy Scale- vary jet corrections by ±1σ, for evts w/≥2 jets
4Lepton ID efficiency- Variation of data/MC SF with isolation
DIL(%)Systematic
Must take into account many effects potentially contributing to the acceptance uncertainty:
Sandra Leone, INFN PisaLHC School Lecce June 2006
DileptonDilepton: Systematic Uncertainty on Background : Systematic Uncertainty on Background EstimateEstimate
30-50Fake Estimate- J20, J50, J70, J100 x-check
42Drell-Yan Estimate- Absolute scale (data driven), Monte Carlo shape
~20WW, WZ, ZZ estimate- Compare WW+0p+(njet scaling) to WW+2p
13-30Jet Energy Scale – same procedure on bkgnd acc.
4Lepton efficiency – same as signal
DIL(%)Systematic
Sandra Leone, INFN PisaLHC School Lecce June 2006
DileptonDilepton: Cross section results : Cross section results
∫××
−=
LdtA
NNtt backobs
εσ )(
_
SM: σ = 6.7 pb (mt = 175 GeV/c2)
~750 pb-1
σσσσ(tt) = 8.3 ± 1.5 (stat) ± 1.0 (syst) + 0.5 (lumi) pb
Sandra Leone, INFN PisaLHC School Lecce June 2006
DileptonDilepton Channel: lepton + track (LTRK) Channel: lepton + track (LTRK)
Signature: looser requirements for second lepton:
� 1 lepton+1 isolated track opposite sign
�pT>20 GeV
�|ηηηη| < 1
� Et > 25 GeV
� ≥ 2 central jets (|ηηηη| < 2, ET>20)
�Sensitive also to ττττ lepton final states (~20% from ττττ)
46 events on 14.6 ± 3.6 background30 e+track, 16 µ+track ttbar
signal bin
Jet Multiplicity
σσσσ(Mtop=178) = 9.9 ± 2.1(stat) ± 1.3(syst) ± 0.6(lum) pb
Sandra Leone, INFN PisaLHC School Lecce June 2006
DileptonDilepton event displayevent display
jet
jet
ννννe1
e2
•2 electrons (ET1=73 GeV, PT2=63 GeV)
• Missing ET = 59 GeV
• 2 central jets + 1 forward jet
e1
e1e1
e2
jetjet
Sandra Leone, INFN PisaLHC School Lecce June 2006
ee--µµ eventeventCDF Run II Preliminary (360 pb-1)MET: 91 GeV, ϕ = 2.8
CEM electron: E
T= 61 GeV,
ϕ = 4.8
tl (= CMX muon): p
T= 102 GeV,
ϕ = 0.8
Jet 1: E
T= 43 GeV,
tagged
Jet 2: E
T= 37 GeV,
tagged
Jet 3: E
T= 36 GeV
Sandra Leone, INFN PisaLHC School Lecce June 2006
tttt DileptonDilepton decays with decays with ττhadhad�We look for anomalously large or small τ τ τ τ lepton rate in top decay
�One W decays into ττττ , the other into µµµµ or e
� ττττ decay mode: semileptonic, 1 charged hadron (~50%) or 3 charged hadrons (~15%)
�Signature: high PT e or µµµµ, large ET, ≥ 2 jets and a object passing ττττ identification cuts:
⇒ PTτ > 15 GeV/c
⇒Isolated Σ PT < 1 GeV/c
⇒electron and muon removal
�HT>205 GeV is requested to reduce the background.
�Use W-> τ ντ ντ ντ ν to understand τ τ τ τ ID
Sandra Leone, INFN PisaLHC School Lecce June 2006
tttt DileptonDilepton decays with decays with ττhadhad
32Obs. Data
0.92 ± 0.05stat1.32 ± 0.05stat ttbar
1.32 ± 0.301.43 ± 0.31Bckground
Muon+ τhadEle + τhadSample In ~360 pb -1 expect ~2.2
signal events. Background dominated by jets faking τ’s.
Results:
Sandra Leone, INFN PisaLHC School Lecce June 2006
Single lepton channel (Lepton + jets)Single lepton channel (Lepton + jets)Signature:
�One high pT Isolated lepton
�Veto Z, cosmic, conversion, dilepton
�ET > 20 GeV
�3 or more jets with ET>15 GeV|η| < 2.0
�S/B ~ 1/6
4028461064768183Secondary vertex sample (~700 pb -1)
W+>=4 jetsW + 3 jetsW + 2 jetsW + 1 jet
ℓ
ℓ
l
b
b
νννν
p p
E T
l
jet
jet
jet
Sandra Leone, INFN PisaLHC School Lecce June 2006
Lepton + jets Channel: HLepton + jets Channel: HTT cutcut� HT is the scalar sum of ET of jets, leptons, and ET
� HT is a powerful discriminator for Top signal
� HT > 200 GeV keep 96% of signal and reject 38% of background
An HT cut increases the systematics due to Top mass dependence, Energy scale and Heavy flavor fraction, but still small compared with other syst(SF, lum)
Sandra Leone, INFN PisaLHC School Lecce June 2006
Lepton +jets channel: counting eventsLepton +jets channel: counting events� Require ≥ 1 jet to be identified as a b-jet
(“b-tagged”)
� Motivation of using b-tags:⇒Reduce the backgrounds, especially W + light flavor jets
events while keeping good efficiency on ttbar signal
⇒B tag improves S/B from 1/6 �3/1
� Count ttbar candidate events
� Predict rates for SM non-top processes in tagged W+jets, excess in ≥3 jets is top
� Use data as much as possible to determine background contamination (non-W QCD, fake tags)
� Use MC when necessary (diboson, W+heavy flavor)
Sandra Leone, INFN PisaLHC School Lecce June 2006
Tagging Tools: Tagging Tools: VertexingVertexing and Soft and Soft MuonsMuons
B hadrons in top signal events:
are long-lived and massive may decay semileptonically
Top Event Tag EfficiencyFalse Tag Rate (QCD jets)
60%0.5%
15%~2%
Identify low-pt muon from decayVertex of displaced tracks
Sandra Leone, INFN PisaLHC School Lecce June 2006
Double bDouble b--tagged tagged Lep+TrkLep+Trk event at CDFevent at CDF
Sandra Leone, INFN PisaLHC School Lecce June 2006
Vertex Tagging AlgorithmVertex Tagging Algorithm
� Take advantage of the long lifetime of B hadrons: τ(b) ∼ 1 ps(cτ ≈ 450 µm) ⇒ B hadrons travel Lxy ∼ 3mm before decay
� Select good quality tracks with large impact parameter.� Try to reconstruct a vertex with ≥ 2 traks
⇒ first pass searches for at least 3 tracks with loose kin and Pt >2 GeV/c
⇒ second pass looks for 2 tracks vertices with tighter cuts on tracks qualityand Pt> 3 GeV/c
� A jet containing a vertex is considered b-tagged if has large (positive) decay length significance: Lxy/σLxy > 7.5 (typically σLxy ∼ 150 µm)
Sandra Leone, INFN PisaLHC School Lecce June 2006
Vertex bVertex b--Tagging EfficiencyTagging Efficiency� Use events with back-to-back jets,
non-isolated electron, require away jet to be tagged (enriched in heavy flavor)
� The efficiency of b-tagging determined by the ratio of the number of double tagged to single tagged events, in data and MC
� Define a Scale Factor between data and MC tagging efficiency (usually
εεεεDATA < εεεε MC) SF ≈ 0.91 ± 0.07⇒ SF < 1 due to # of good vertex tracks
higher in MC
� Check in generic jets the ET
dependence of the b-tagging efficiency
�Efficiency for tagging at least one jet in a ttbarevent (L+>=3 jets, including data-MC scaling):
ε ≈ 60 ± 4 %
Sandra Leone, INFN PisaLHC School Lecce June 2006
Vertex Tagging BackgroundVertex Tagging Background� Most from W+heavy flavor and W+mistags
� W+mistags: from generic jet data
� In generic jets heavy flavor pairs are produced by both direct production and gluon splitting. In W+jets by gluon splitting only.
⇒The fraction of Wbb, Wcc events is determined from MC and scaled to the observed number of W events in each jet multiplicity bin
� The (smaller) QCD (non-W) background is evaluated from data: lepton isolation vs Missing ET method (standard)
Sandra Leone, INFN PisaLHC School Lecce June 2006
W + W + mistagsmistags backgroundbackground
� “Fake tags” background:⇒The fake background is measured from generic jet data
using negative decay-length tags �Assume that positive mistag rates are well described by negative
(Lxy <0) tag rates. �Probability to have a negative Lxy (~ 0.5% per jet)
– it is parametrized as a function of the jet ET, jet silicon track multiplicity, η
– it is used to estimate the mistag rate:» Negative- non-physical tags = positive mistags
⇒This probability matrix is then applied to W+jets events to obtain the background estimate in our sample
�
Sandra Leone, INFN PisaLHC School Lecce June 2006
NonNon--W W ““QCDQCD”” Background: Background: ““IsoIso vsvs METMET””
The simplest and most discriminating difference between an “isolated” electron as the one coming from the W decay and a jet is the Isolation.There is NO correlation between Missing Transverse Energy and Isolation for dijet events faking a W->eνννν candidate.→”Isorel vs MET” method
Sandra Leone, INFN PisaLHC School Lecce June 2006
Lepton + jets with Lepton + jets with ≥≥ 1 vertex tag: results1 vertex tag: results
≥ 1 SecVerteX = 314 events on 70 ± 8 bkgnd
Mis tag from
Generic jet data
Method II:
based on MC
MET vs Isolation
method: data driven
top signal region
Number of jets per event:
Sandra Leone, INFN PisaLHC School Lecce June 2006
Tagged Jets PropertiesTagged Jets Properties
The tagged events contain b quarks, as
seen by the decay length Lxy distribution
Sandra Leone, INFN PisaLHC School Lecce June 2006
Tagged Jets PropertiesTagged Jets Properties
The tagged jets are kinematicallyconsistent with the expectation from top, as seen by their ET distribution
Sandra Leone, INFN PisaLHC School Lecce June 2006
Summary of secondary vertex countingSummary of secondary vertex counting
1581565141029Observed positive tags
463329-Observed double tags
1.9 ±0.57.2±1.324 ± 5-Double tags backg
53±6 17±2408 ±53921 ± 113Total background
4028461064768183Events before tagging
W + ≥ 4 jW + 3 jW + 2 jW + 1 j
Require HT > 200 GeV for 3, ≥ 4 jets bins only,
~700 pb-1
Sandra Leone, INFN PisaLHC School Lecce June 2006
Lepton +jets cross section Lepton +jets cross section
� High pt lepton plus missing ET
� N(jet) 3 � 1 jet SEC(ondary) VTX� HT>200 GeV
∫⋅⋅
−≥+=
LdtA
BKGNjetsWNtt
εσ
)()3()(
A: ttbar acceptance : b-jet eff. for ttbarε
≥
8.2 ± 0.6 (stat) ± 1.0 (syst) pb=)(ttσ
≥
Assuming a top mass of 178 GeV/c2
~700 pb-1
Sandra Leone, INFN PisaLHC School Lecce June 2006
Lepton +jets with Lepton +jets with ≥≥ 2 vertex tag2 vertex tagssLower statistics but low background: 79 events observed in top region, 9.1 ± 1.8 bkgd expected
b-tagging efficiency and its uncertainty enters twice here
2 jet bin:29 events observed, ~ 24±5 bkg.~ 7 ttbar expected
# of jets per event:
top signal region
=)(ttσfor a top mass of 178 GeV/c2
8.8+1.2-1.1(stat.)+2.0
-1.3 (sys.) pb
Sandra Leone, INFN PisaLHC School Lecce June 2006
Soft Lepton TaggingSoft Lepton Tagging� Leptons from semi-leptonic decays of B have a softer pT
spectrum than W/Z leptons and are less isolated
� Soft Muon Tagger based on a “Global χ2” to identify low-pt muons: “Global χ2” combines information from muon matching variables There is no calorimetry or isolation requirements
� b-tagging with SLTµ: Track with dR slt-jet<0.6,
|dZ slt-Zvtx|<5cm & pT>3 GeV/c
are considered by the SLTµ tagger An event is tagged if at least
one SLTµ tag is found
Sandra Leone, INFN PisaLHC School Lecce June 2006
Lepton + jets: Soft Lepton Tag countingLepton + jets: Soft Lepton Tag counting
Number of jets per event
MET vs isolation method
Predicted by data in/out
of Z-mass window.
From MC and theory
Alternative way to secondary vertex b-tag
(b→µν→µν→µν→µνX, b→→→→c →→→→ µνµνµνµνX)
top signal region
Predicted by fake rate
matrix made by photon
+ jets sample.
Sandra Leone, INFN PisaLHC School Lecce June 2006
Lepton + Jets: Soft Lepton Tag ResultsLepton + Jets: Soft Lepton Tag Results
Tagged Events: 139 48 13 7 20
Assuming a top mass of 175 GeV/c2
Sandra Leone, INFN PisaLHC School Lecce June 2006
Soft Lepton Tag PropertiesSoft Lepton Tag Properties
Jet ET distribution in W + ≥≥≥≥ 3 jets for SLtagged events and expectations from background and ttbar (scaled to the measured cross section of 5.2 pb).
Comparison of the SLTµ pT distribution in W + ≥≥≥≥ 3 jets for SLtagged events and for expectations from background and ttbar (scaled to the measured cross section of 5.2 pb).
Sandra Leone, INFN PisaLHC School Lecce June 2006
All All HadronicHadronic channelchannelVery difficult channel!!
Signature:
• 6 or more jets
• kinematical selection
• ≥≥≥≥ 1 vertex tag.
Background estimate based on data.
Observe 143 tags excess over 683 expected background in signal region.
pbsyststattt
3.5
4.3
7.4
3.2 8.7)()(5.28.7 +−
+− =±=σ
Assuming a top mass of 175 GeV/c2
7.5 ± 1.7 (stat) +3.3- 2.2 (syst) = 7.5 +3.7
-2.8 pb
Sandra Leone, INFN PisaLHC School Lecce June 2006
Top cross section summaryTop cross section summary
√√ss--Dependence:Dependence:
⇒ Main data driven systematics(jet energy scale, ISR, εbtag) scale with 1/√N :
⇒ RunII(2fb-1) δσδσδσδσtt/σσσσtt <10%
Sandra Leone, INFN PisaLHC School Lecce June 2006
22ndnd PartPart
Top Physics at CDFTop Physics at CDFand lessons for LHC and lessons for LHC
Sandra Leone, INFN PisaLHC School Lecce June 2006
OutlineOutline
� Motivations for studying top � A brief hystory� Top production and decay � The Tevatron & CDF� Detector issues� Identification of final states� Cross section measurement� Mass determination� Single top production� Study of top properties� The lesson for the LHC
Yesterday Introductory
Part
Sandra Leone, INFN PisaLHC School Lecce June 2006
You cannot trust MC to predict instrumental effects/backgrounds involving distribution tails
Is our detector/simulation particularly bad?
No! You always have at some level this kind of effects
We are afflicted by very “rare” instrumental
backgrounds
We are looking for very rare events
What can we do?
Optimize the cuts to reduce the bckgndcontamination, optimize with available data
the simulation, but in the end use real datawhenever possible
Sandra Leone, INFN PisaLHC School Lecce June 2006
Cannot rely (completely) on MC when instrumental backgrounds are concerned,
distribution tails are concerned
Are MC useless?Are MC useless?
NO! of course, they are necessary tools!
� Use MC to model the kinematical and geometrical acceptance of signal and physics backgrounds
� Trust ME calculations� Trust basic kinematical description of data� Trust basic description of detector features
� Check MC predictions on control samples
BUT
Sandra Leone, INFN PisaLHC School Lecce June 2006
Final states:Final states:
ℓ
ℓ
jet
jet
b
b
νννν
νννν
p p
ET
+
-
jet
jet
b
b
νννν
νννν
pp p
ET
ET
Dilepton:
b
b
νννν
p p
E T
jet
jet
jet
ℓ
ℓ
Lepton + jets
Sandra Leone, INFN PisaLHC School Lecce June 2006
Once reOnce re--established top established top existanceexistance, , letlet’’s measure its properties. s measure its properties. First property is its First property is its massmass
Sandra Leone, INFN PisaLHC School Lecce June 2006
Top mass is a fundamental SM parameter
• Related through radiativecorrections to other EW observables.
CDF&D0
RUNII
Why is Top Mass so special?Why is Top Mass so special?
• Very important for precision tests of SM.•Together with Mw and other electroweak precision
measurements, it constrains MHiggs
• By far the largest quark mass, largest mass of all known fundamental particles
Sandra Leone, INFN PisaLHC School Lecce June 2006
What is the top mass?What is the top mass?� Particle masses are not “directly measured”
⇒ one can only measure cross sections, decay rates� Measurement of Mtop: at Tevatron, LHC:
⇒kinematic reconstruction, fit to invariant mass distribution⇒Best measurement from lepton +jets
� Experimental accuracy of Mtop :⇒Measurement comparison data from Monte Carlo⇒ you measure the mass that is implemented in your MC
� measured mass is not strictly model independent
� Situation at the Tevatron:⇒δδδδMtop = 2.3 GeV (Tevatron; today)
� Projections at the LHC:⇒δδδδMtop < 1 GeV with 10 fb-1
����Will Tevatron get there first?
Sandra Leone, INFN PisaLHC School Lecce June 2006
Top Mass Measurement ChallengesTop Mass Measurement Challenges
Why so challenging? It is a difficult measurement Many combinations of leptons and jets:
Events are complicated! Measurements are not perfect!Experimental observations are not as pretty as Feynman diagrams!
�Missing neutrino�Confusion in ID assignment (add. Jets from ISR/FSR, b-tag: not 100% correct)
�Link observables to parton-level energies�Large syst uncert. from jet energy scale�Need accurate detector simulation
l
ν
W+
W-
t
t
b-jet
b-
jet
jet
jet
X Constraints
Method: reconstruct Mtop with 2 constraints: M(W+)=M(W-), M(t)=M(tbar)
ℓ
Sandra Leone, INFN PisaLHC School Lecce June 2006
Lepton + Lepton + ≥≥ 4 jets: template method4 jets: template method
Wbb MCData
tt MC
Datasets
Massfitter
Templates
χ2 mass fitter:�Finds top mass that fits event best� All event info into one number � 12 parton/jet matching assignments possible, 2 longit. neutrino possible, use b-tag to reduce permutations � test for consistency with top using kinematic constraints � choose combination with lowest χ2
Likelihoodfit ResultLikelihood fit:
� fit resulting mass distribution to MC background + top signal templates at different values of Mtop
� Best template to fit data gives mass� Constraint on background normalization
Sandra Leone, INFN PisaLHC School Lecce June 2006
The new CDF top mass measurement The new CDF top mass measurement in in
Lepton+Jets channel with 680 pbLepton+Jets channel with 680 pb--11
� Subdivision improves statistical uncertainty.⇒Pure and well reconstructed events contribute more to result.
⇒Adds 0-tag events.
� Subdivision does not improve systematic uncertainty.⇒Most systematics, including jet energy scale, are highly
correlated among the samples.
Improve stat. power of the method dividing the sample in 4 categories of events that have different backg. contamination and different sensitivity to the top mass
0.6:11:14:110:1S:B
ET>2115>ET>8ET>15ET>8j4
ET>21ET>15ET>15ET>15j1-j3
0-tag1-tag(L)1-tag(T)2-tagCateg.
Sandra Leone, INFN PisaLHC School Lecce June 2006
Top Mass reconstruction in Top Mass reconstruction in MontecarloMontecarlo eventsevents
Sandra Leone, INFN PisaLHC School Lecce June 2006
� W+heavy flavor jets(bb,cc,c)⇒Heavy flavor fraction from MC⇒Normalized to data
� W+jets(mistag)⇒Use measured mistag rate,
applied to the data� Multijet:non-W (jet->e, track->µµµµ)
⇒Estimated from data� Single top, diboson (WW,WZ)
⇒Estimated from MC
Mass analysis event selection & BackgroundsMass analysis event selection & Backgrounds
57
4.0±±±±1.3
2tag
75
31.±±±±7.
1tagL
no est22.±±±±5.bkgds
108120Data
1tagT 0tagIn 680 pb-1
360 candidate events used for top mass measurement
Sandra Leone, INFN PisaLHC School Lecce June 2006
3.4TOTAL
0.4MC statistics
0.2b-tagging
0.6b-jet energy scale
1.0Background shape
0.3Generators
0.4PDFs
0.4FSR
0.4ISR
3.1Jet Energy Scale
∆Mtop (GeV/c2)Systematic
Systematic Errors (based on 318pbSystematic Errors (based on 318pb--11))
Can we improve jet systematic?
Dominated by jet energy scale:
Sandra Leone, INFN PisaLHC School Lecce June 2006
3.4TOTAL
0.4MC statistics
0.2b-tagging
0.6b-jet energy scale
1.0Background shape
0.3Generators
0.4PDFs
0.4FSR
0.4ISR
3.1Jet Energy Scale
∆Mtop (GeV/c2)Systematic
0.6Total
0.4Semi-leptonicdecay
0.3Color flow
0.4Heavy quark fragmentation
∆Mtop
(GeV/c2)
B-jet energy
scale
Systematic ErrorsSystematic Errors
Though we find that 70% of JES uncertainty comes from b-jet, b-jet uncertainty is mainly due to generic jet corrections: only 0.6 GeV/c2
additional uncertainty on Mtop due to b-jet-specific systematics
Sandra Leone, INFN PisaLHC School Lecce June 2006
Jet Energy Jet Energy MismeasurementMismeasurement
Precision on the determination of jet energies is necessary for Mtop measurements.Jets may be mismeasured due to a variety of effects:
� Instrumental effects: calorimeter non-linearities non-instrumented regions non-compensating calorimeter
�Physics effectsContribution from underlaying eventmultiple ppbar interactionsthere are different types of jets
�Algorithm effects: might not capture all particles (out of cone)low energy jets might not be possible to define
Non-uniform
response
Diff. resp.
of π0/πi+-
Non-linearity
Shower, frag.
Sandra Leone, INFN PisaLHC School Lecce June 2006
Jet Energy CorrectionJet Energy Correction
Determine true “parton” E from measured jet E in a cone 0.4
The correction factor depends on jet ET and η and is meant to reproduce the average jet ET correctly, (not to reduce the jet fluctuations around this mean)
A set of corrections was developed for generic jets:⇒Absolute corrections (photon-jet balancing)⇒Relative corrections (central-forward calorimeters, dijet balancing)
Out-of-Cone: correction to partonUnderlying event”top-specific correction” to light quark jets and b-jets separately
Non-uniform
response
Diff. resp.
of π0/πi+-
Non-linearity
Shower, frag.
Sandra Leone, INFN PisaLHC School Lecce June 2006
Jet Energy Jet Energy SystematicsSystematics
A lot of work has been done to reduce the syst. from jet-energy scale (a factor of two improvement compared to start of RunII). The new Run II systematic uncertainties are now better than Run I.
RunII
RunII 2004
RunI
Frac. Syst. uncert. vs Pt
Central region
�What is jet energy scale “JES”?
�Measures how incorrect is our nominal jet energy measurement.�Units of σσσσ: correspond to one s.d. of jet energy uncertainty
⇒ Accounts for ET, η dependence.
Sandra Leone, INFN PisaLHC School Lecce June 2006
WorldWorld’’s Best Top Quark Mass:s Best Top Quark Mass:MMtoptop+ JES simultaneous fit+ JES simultaneous fit
� Measure JES in situ.� Perform simultaneous fit.
⇒ Extend 1-D template (only on reconstructed Mtop) to maximize sensitivity to Jet Energy Scale:⇒ Mtop and JES are simultaneously determined in likelihood fit using shape comparisons ofReconstructed Mtop and, Reconstructed Mjj
distributions, taking correlations between them
Sandra Leone, INFN PisaLHC School Lecce June 2006
l
n
W+
W-
t
t
b-jet
b-jet
jet
jet
JES and JES and MtopMtop measurementsmeasurements-- 2D Template Analysis 2D Template Analysis --
Simultaneous fit to JES and Mtop using top mass and W mass templates:
Mt ( true Mtop, JES), Mjj ( true Mtop, JES)
� Identify jets coming from W• All non-btagged jets pairs are taken
into account equally.• 1/3/6 Mjj per event with 2/1/0 b-tag
� Reconstruct their invariant mass Mjj
� Mjj strongly dependent on JES• Make Mjj templates by varying JES • Fit data with Wjj to measure JES!
� MW uncertainty is negligible (< 50 MeV)
� Mjj mostly independent of Mtop
� This scale is applied to b-jets and light-quark jets
JES from W→jj is mostly
statistical → luminosity scale !
Mjj(W)
Sandra Leone, INFN PisaLHC School Lecce June 2006
2D Top Mass Result with 680 pb-12D Top Mass Result with 680 pb2D Top Mass Result with 680 pb--11
2
top GeV/ (syst) 3.1(stat) 5.24.173 cM ±±=
mtreco in data w/ global best fit overlaid
± 1.7 (stat) ± 1.8 (JES)
Using 360 candidates in 680 pb-1 we measure:
Mtop = 173.4 ± 2.8 GeV/c2
Best single measurement in the world!
Better than TevatronSummer 2005 average!
Sandra Leone, INFN PisaLHC School Lecce June 2006
JES Results with 680 pb-1JES Results with 680 pbJES Results with 680 pb--11
• Combined W→jj and prior
JES yield 40% improvement
Reconstructed dijet (W) mass:
Black curve from global best fit.
∆∆∆∆JES = -0.3 ± 0.6 (stat. + Mtop) σσσσc
Sandra Leone, INFN PisaLHC School Lecce June 2006
Systematic Uncertainty in 680 pbSystematic Uncertainty in 680 pb--11
0.4Bkgd JES
0.7Residual JES
0.1B-tagging
0.2FSR
0.3Generators
0.5Bkgd Shape
0.5ISR
1.3TOTAL
0.3MC stats
0.3PDFs
0.6B-jet energy scale
∆∆∆∆Mtop
(GeV/c2)Systematic
Systematics are largely due to uncertainties in modeling.
From pT and η dependence of modeling uncertainties
Model constrained by Z+jets data
Sandra Leone, INFN PisaLHC School Lecce June 2006
Top mass: matrix element (ME) methodTop mass: matrix element (ME) method� Optimizes the use of kinematic and dynamic information� Calculate a probability per event to be signal or background as a
function of the top mass� Signal probability for a set of measured jets and lepton (x)
� Likelihood simultaneously determines Mtop, Jet Energy Scale, and signal fraction:
� All jet-parton assignments and neutrino solutions are considered, weighted.
� Select events with exactly 4 jets, well described by LO ME.
P(x; M top ,JES) =
1
σdq 1dq 2 f(q 1) f(q 2 ) d σ (y; M top )∫ W(x, y, JES)
L ( f top, M top , JES ) ∝ ∏i
Nevents
f top Ptop ,i ( M top , JES ) + (1 − f top ) Pbkgd ,i (JES )( )
Differential cross sectionDifferential cross sectionDifferential cross sectionDifferential cross section:
LO ME (qq->tt) only
Transfer function: Transfer function: Transfer function: Transfer function: probability
to measure x when parton-level
y was produced
Sandra Leone, INFN PisaLHC School Lecce June 2006
Matrix Element (L+J) ResultsMatrix Element (L+J) Results——680 pb680 pb--11
� Exactly 4 jets with ≥ 1 b tags
� JES here is a constant multiplicative factor.⇒Edata = EMC/JES
� JES = 1.02 ± 0.02.
⇒Consistent with template method
� Virtually identical sensitivity with fewer events!
2
top GeV/ (syst) 4.1(stat) 5.21.174(ME/LJ) cM ±±=Using 118 candidates in 680 pb-1
Sandra Leone, INFN PisaLHC School Lecce June 2006
Top Mass: Top Mass: DileptonDilepton channel, ME techniquechannel, ME technique
~1 27≥1 b tags
19.4 ± 3.464≥0 b tags
BackgroundEvents� Harder to reconstruct Mtop in dilepton events: two neutrinos make system underconstrained.
� Likelihood is similar to L+jets.
� No W reson. → no fit for JES
�Add ME for dominant bkgds:DY+jets, WW+jets, fakes
2
top GeV/ (syst) 1.3(stat) 5.45.164 cM ±±=
Restrict sample to b-tagged events:Mtop = 162.7 ± 4.6 ± 3.0 GeV/c2
750 pb-1
probability density as a function of top pole mass
Sandra Leone, INFN PisaLHC School Lecce June 2006
Top mass summaryTop mass summary
CDF April 2006: 172.4± 1.5 ± 2.2
Tevatron March 2006: 172.5± 1.3 ± 1.9
Sandra Leone, INFN PisaLHC School Lecce June 2006
Keep an eye onKeep an eye on……
� Discrepancy btw L+jets, Dilepton ch. measurements…?
� Is it statistical?� Is there a missing
systematic?� Is our assumption of
SM ttbar incorrect??
Stay tunedStay tuned……
Sandra Leone, INFN PisaLHC School Lecce June 2006
Consequences on the Higgs MassConsequences on the Higgs Mass
Preferred MHiggs
MHiggs = 89 +42-30 GeV
MHiggs < 175 GeV @ 95% CL
This limit increases to 207 GeVwhen including the LEP-2 direct search limit of 114 GeV shown in yellow
Sandra Leone, INFN PisaLHC School Lecce June 2006
The Future Prospects for top massThe Future Prospects for top massThe Future Prospects for top mass
� Using W→jj: JES uncertainty becomes mostly statistical
So we can reach JES uncert. < 1 GeV/c2 in Run II
� Reaching total δδδδMtop < 1.5 GeV/c2 can be possible with full Run II dataset
Now !
Achieved 1.3% precision of the Mtop measurement
Sandra Leone, INFN PisaLHC School Lecce June 2006
SingleSingle--Top Quark ProductionTop Quark Production
Theoretical cross section predictions at √s = 1.96 TeV
0.88 ±±±± 0.11 pb 1.98 ±±±± 0.25 pb
Top quark production via the weak interaction
B.W. Harris et al. Phys. Rev. D 66 054024 (2002)
Vtb Vtb
s-channel t-channel
Sandra Leone, INFN PisaLHC School Lecce June 2006
Single Top EWK ProductionSingle Top EWK Production
s –channel (W*)
btWqq * →→→→→→→→
t -channeltqqb '→→→→
tW associated production−−−−→→→→ tWbg
< 14 pb< 13 pb< 18 pbCDF
< 22 pb<17 pbD0
62+17 -4 pb247 ±±±± 25 pb10.6 ±±±± 1.1 pbLHC σσσσNLO
~ 0.1 pb1.98 ±±±± 0.25 pb0.88 ±±±± 0.11 pbTevatron σσσσNLO
Combined
(s+t)Associated tWt-channels-channel
B.W. Harris et al.:Phys.Rev.D66,054024 T.Tait: hep-ph/9909352
Z.Sullivan Phys.Rev.D70:114012 Belyaev,Boos: hep-ph/0003260
RunI
95% C.L.
(mtop=175 GeV/c2)
Sandra Leone, INFN PisaLHC School Lecce June 2006
Cross section smaller than top pair production, but ➡Observation of single top allows direct access to Vtb CKM matrix element: cross section ∝ |Vtb|
Test non-SM phenomena� 4th generation, FCNC couplings like t →→→→ Z/γγγγ c, anomalous Wtbcouplings
Potentially useful for Higgs searches� single top has same final state as Higgs+W (associated) production
Why Single Top? Why Single Top?
only indirectly
known
Direct measurement
only via single-top quark production
Sandra Leone, INFN PisaLHC School Lecce June 2006
Single Top Signature & BackgroundSingle Top Signature & Background
ss--channelchannel
tt--channelchannelb quark produced WITH the top
b quark from the top decay
lepton + Missing ET
b quark from the top decay
lepton + Missing ET
extra light quark
at NLO an additional b from gluon splitting (difficult to measure)
Backgrounds:⇒ W/Z + jets production
⇒ Top pair production
⇒ Multijet events
Sandra Leone, INFN PisaLHC School Lecce June 2006
CDF Single Top Search StrategyCDF Single Top Search Strategy
� W selection:
� - ET > 20 GeV/c (central and forward electrons)
� - pT > 20 GeV/c (central muons)
� - Missing ET > 20 GeV
� Jets selection:
� - Exactly 2 jets, at least one is b-tagged
� - Jet ET > 15 GeV, lηl < 2.8
696 pb-1
Separate Channel Search
2D Neural Network discriminant
+ likelihood fit
Combined Channel Search
1D Neural Network discriminant
+ likelihood fit (Bayesian approach))
Sandra Leone, INFN PisaLHC School Lecce June 2006
Separate Channel searchSeparate Channel search
(syst)pb(stat)0.6σ0.10.1
1.90.6cht
++++−−−−
++++−−−−−−−− ====
t-channel:
σσσσ < 3.1 pb @ 95% C.L.
s-channel:
σσσσ < 3.2 pb @ 95% C.L.
CDF applies a maximum likelihood fit to the network 2D output and the best fits for t- / s-channel are:
t-channel:
s-channel:
The resulting upper limits are:
pb (syst)(stat)0.3σ0.5
0.3-2.20.3chs
++++++++−−−−−−−− ====
Sandra Leone, INFN PisaLHC School Lecce June 2006
� Best fit value:
Combined channel searchCombined channel search
The resulting upper limit
on the cross section is:
σσσσsingle-top < 3.4 pb
at 95% C.L.
(syst)pb(stat)0.8σ0.20.3
1.30.8st
++++−−−−
++++−−−−++++ ====
Sandra Leone, INFN PisaLHC School Lecce June 2006
Measurement of Top Quark Production Measurement of Top Quark Production and Decay Propertiesand Decay Properties
� Let’s continue on the natural path of measuring top quark properties in order to confirm SM or find deviations from it
⇒ Is top quark adequately described by the Standard Model?
Top spin polarization
Production Cross Section
Resonance production ?
Production kinematics
Sandra Leone, INFN PisaLHC School Lecce June 2006
� Several theoretical scenarios:⇒Heavy 4th generation quark
(He et. Al. Hep-ph/0102144) � Consistent with EWK data
⇒ “beautiful mirrors”models (Wagner et. Al. Hep-ph/0109097)�Accomodates LEP b forward-backward asymmetry result
� CDF performs 2D fit: HT = ΣjetsET + ET,l + MET and Mreco from χ2 mass fit in Lepton + Jets channel
� Uses binned likelihood fit and Bayesian limit:⇒Rule out at 95% CL a t’ with
m(t’) < 258 GeV/c2
Search for Heavy Quark Search for Heavy Quark tt’’→→WqWq
Sandra Leone, INFN PisaLHC School Lecce June 2006
Searches for Searches for ttbarttbar ResonancesResonances
pp → X0 → tt
� Various exotic models predict the existence of particles decaying to ttbar: Topcolor-Assisted Technicolor
(Hill, Phys Lett. B345, 483 (1995); Hill and Parke PRD49, 4454 (1994))
� Extends technicolor models and attempts to explain EWSB by introducing a new strong interaction
� Predicts new massive bosons “topgluons” and a topcolor Z’
Sandra Leone, INFN PisaLHC School Lecce June 2006
History: Previous MeasurementsHistory: Previous Measurements
� Observed quite intriguing excess around 500 GeV� Had a similar although smaller excess in Run 1
CDF
Run 1
� Lepton + jets:CDF
Phys.Rev.Lett. 85, 2062 (2000)
Sandra Leone, INFN PisaLHC School Lecce June 2006
Latest CDF measurement (682pbLatest CDF measurement (682pb--11))
With about twice more data the excess disappeared!
95% CL upper limits for the cross section for resonance production
Mtt distribution from data vsexpected SM background in the >400 GeV search region
>725
Sandra Leone, INFN PisaLHC School Lecce June 2006
Top LifetimeTop Lifetime� The SM predicts a top lifetime of
~ 4 x 10-25 s ⇒A measured deviation from
“zero lifetime” could mean:�A much larger top lifetime�New Physics?o Anomalous top production by a long-
lived parent particle, oro A long-lived background to SM top.
•Use Lepton+Jets top pair sample with a b tagged jet.
•Measure signed lepton impact parameter (d0)
•Calibration:
•Used Drell-Yan events near Z resonance to understand the d0resolution.
•Confirmed technique correctly measures tau lifetime in Z → ττ
Sandra Leone, INFN PisaLHC School Lecce June 2006
Top Lifetime (2)Top Lifetime (2)� Backgrounds:
⇒Prompt: W+jets, Drell-Yan, Diboson
⇒Displaced lepton: W/Z decaying to τ, semileptonic b & c decays, photon conversions
� Results:⇒Using 157 events
�32 expected background
First Direct LimitFirst Direct Limit
On Top LifetimeOn Top Lifetime
Data compared to fit result
Sandra Leone, INFN PisaLHC School Lecce June 2006
W Helicity Measurement in top decayW Helicity Measurement in top decay� Top decays before it can hadronize, because width Γt ~ 1.5 GeV > ΛQCD.⇒Decay products preserve information about parent particles.
⇒Unique opportunity to study the weak interactions of a bare quark, with a mass at the natural electroweak scale!
� SM Prediction:⇒W helicity in top decays is fixed by Mtop, MW, and V-A structure
of the tWb vertex.
⇒W helicity reflected in kinematics of W decay products
Sandra Leone, INFN PisaLHC School Lecce June 2006
W Helicity Measurement, contd.W Helicity Measurement, contd.
-1 -0.8 -0.6 -0.4 -0.2 -0 0.2 0.4 0.6 0.8 10
0.5
1
1.5
2
2.5
3
3.5
4
-
φcos
)φw
(co
s
-1 -0.8 -0.6 -0.4 -0.2 -0 0.2 0.4 0.6 0.8 10
0.5
1
1.5
2
2.5
3
3.5
4
0
φcos
)φw
(co
s
-1 -0.8 -0.6 -0.4 -0.2 -0 0.2 0.4 0.6 0.8 10
0.5
1
1.5
2
2.5
3
3.5
4
+
φcos
)φw
(co
s
22
0
2 )cos1(8
3)cos1(
8
3)cos1(
8
3)(cos
blblblblFFFw −−−− +⋅+−⋅+−⋅= +− ϕϕϕϕ
The angular dependence of the semileptonic decay in the W rest frame is given by:
3.021
2≈
+=− ω
ωF 7.0
21
10 ≈
+=
ωF 0=+F
where ω= MW2/Mtop
2
parameter to measure
SM (V-A) predictions (for mb=0):
left rightlong.
[V+A: 70% long., 30% r.h.]
Sandra Leone, INFN PisaLHC School Lecce June 2006
W W HelicityHelicity Measurement: MMeasurement: Mlblb22
Use the invariant mass of the lepton and the b jet, assuming Use the invariant mass of the lepton and the b jet, assuming mmbb = 0,= 0,
Three data samples used:Three data samples used:
�� Single b tagged Single b tagged leplep + jets+ jets
�� Double b tagged Double b tagged leplep + jets+ jets
�� DileptonDilepton
F+ = 0.02 ± 0.07
F+ < 0.09 at 95% C.L.
Sandra Leone, INFN PisaLHC School Lecce June 2006
θ* is the angle between the P of the charged lepton in the W boson rest frame and the W P in the top quark rest frame. In order to determine the cos(θ*) distribution in the data, the kinematics of the ttbar events are fully reconstructed
Lepton+jets sample
W W HelicityHelicity measurement:cosmeasurement:cosθθ**
F0 = 0.85 +0.15-0.22 (stat.) ± 0.06 (syst.)
F+ = 0.05 +0.11-0.05 (stat.) ± 0.03 (syst.)
F+ < 0.26 at the 95% C.L.
Sandra Leone, INFN PisaLHC School Lecce June 2006
Top Branching fractions R=Top Branching fractions R=B(tB(t��Wb)/B(tWb)/B(t��WqWq))
� Cross section measurements assume t→Wbalways, never t→Wq.
� b-tag rates in tt events test that assumption, depend on B(t→Wb) = R, tag efficiency = ε:
� Measure from the ratio of tag rates the product R εεεε: assume ε and measure R or vice-versa.
(Rε) Rε) Rε)
Rε
Sandra Leone, INFN PisaLHC School Lecce June 2006
Measuring Measuring R R
assuming 3 generations
•Integrated luminosity of 162 pb-1
Sandra Leone, INFN PisaLHC School Lecce June 2006
Outlook from the Outlook from the TevatronTevatron
� Run II measurements of cross section & mass are improving rapidly.
� Other analyses (W helicity, single top…) are making excellent progress.
� A program of precision top physics—and hopefully top surprises— is well established.
� Tevatron performing well: experiments very busy in staying after the data and using all of them!
� We still expect 5x more data 5x more data in hand before LHC starts!
DØ / CDF
Run 2a
Goal
Run II with first 2 fb-1 will provide a chance to:
�Measure Mw to < ±40 MeV/c2
�Measure Mtop to < ±2.0 GeV/c2
Sandra Leone, INFN PisaLHC School Lecce June 2006
Extended Extended TeVTeV goalsgoals� Luminosity plan 2002� Goals 2002 - 05 accomplished� Goals 2006 ~ 500 - 890 pb-1
⇒Tevatron is performing very well⇒Up to now accomplished
expectations0.080.08FY02
8.564.41Total
2.421.16FY09
2.371.14FY08
1.530.63FY07
0.890.50FY06
0.670.39FY05
0.380.31FY04
0.220.20FY03
Design plan
Luminosity/yr (fb-1)
Base plan luminosity/yr (fb-1)
Year
0
1
2
3
4
5
6
7
8
9
0 1 2 3 4 5 6 7 8 9 10
End FY
fb-1
Base plan
Design plan
Base plan ~ no recycler
LLLLmax = 2-3 x 1032
Sandra Leone, INFN PisaLHC School Lecce June 2006
Top Top ExpectedExpected statisticsstatistics @ LHC@ LHC
≈ 80 000 000100 days : 1001034High luminosity
(> 2010?)
≈ 8 000 000100 days : 101033Low luminosity
(2008)
≈ 80 00010 days : 0.11032Very beginning
(~ 2007?)
Number of inclusivet t events
Integrated luminosity
(fb-1)
Luminosity
(cm-2s-1)
Data taking
Sandra Leone, INFN PisaLHC School Lecce June 2006
Top @ LHC: Top @ LHC: Some Thoughts (from a nonSome Thoughts (from a non--LHCLHC’’erer))
� With the first data: ⇒ understand the detector and trigger
�Many detector-level checks (tracking, calorimetry)
⇒Minimize MC dependency & check detector simulation
⇒Exploit available ‘candle’ physics signals:
�Presence and mass of the W±, Z0, top-quark
⇒Calibrate:�Jet energy scale�b-jets & tagging (require good alignment: cannot
come immediately)�Balance in transverse plane (missing energy)
“Only after full understanding of these the road to discovery starts…”
M. Cobal
Sandra Leone, INFN PisaLHC School Lecce June 2006
Top @ LHC: Top @ LHC: Some Thoughts (from a nonSome Thoughts (from a non--LHCLHC’’erer))
The 3 golden suggestions:
� Measure everything ( ….as much as possible….) in data
⇉ But try to have soon a well-tuned simulation available..
� Understand b-tagging:
⇉ learning how to deal with high luminosity and multiple interactions is crucial : efficiency decreases and mis-tag rates rise
� Make sure the calibration triggers are in:
⇉ Low-pT leptons
⇉ Jets at various thresholds
� Likely have to prescale a lot
� Don’t let them get turned off at high luminosity!
Sandra Leone, INFN PisaLHC School Lecce June 2006
l
νννν
W+
W-
t
t
b-jet
b-jet
jet
jet
Top @ LHC: Top @ LHC: Some Thoughts (from a nonSome Thoughts (from a non--LHCLHC’’erer))
Yesterday’s discovery could be tomorrow’s calibration
⇉Top events as “tools” to calibrate the “basics”:
�Clean source of b jets: Measure b-tag efficiency (or data/MC scale factor) in top events?
�Invariant mass of jets should add up to W mass: jet energy scale calibration
�Missing energy calibration: 4-momentum of single neutrino in each event can be constrained from event kinematics
Sandra Leone, INFN PisaLHC School Lecce June 2006
The Road Ahead the The Road Ahead the FemptobarnFemptobarn Era!Era!
� The Tevatron program in the pre-LHC era is very exciting & rich
� The standard model top is pretty healthy⇒ We are beginning to have sensitivity to the unexpected
� More data makes us smarter !
⇒ It is not just the luminosity factor
⇒ We become more daring and more creative
⇒New channels open up� For calibrations
� For cross-checks
� For understanding
� Let me advertise the many opportunities to do great physics at the Tevatron!!
� Many opportunities to develop expertise which will be crucial for the LHC physics program