Measurements of the Top Quark Mass at CDF Run II

1
Measurements of the Top Quark Mass at CDF Run II

• Jean-Francois Arguin
• University of Toronto

• Particle Physics Seminar
• Brookhaven National Laboratory
• May 26th, 2005

2
Outline

• Introduction, motivation
• Run I world average
• Mtop measurement using the template method
• Improving the systematics W?jj calibration
• Current result
• Other methods

3
The Top Quark

• Discovered only recently (1995 at CDF, DØ)
• No real surprise existence is required for
  viability of the Standard Model (SM)
• Renormalizability
• Theory/experiment agreement for observables
  sensitive to isospin nature of b-quarks
  

Most striking characteristics huge mass!
How large? 40 times Mbottom Comparable
to gold nucleus
4
Top Production and Decay

• Half-life of top 10-25s
• Top decays before hadronizing!

• QCD Production (6pb) dominates at Tevatron
• 85
• 15
• single-top EWK production predicted but never
  observed
• Decay in SM
• Thus,

5
Observed Final State and Selections

• W bosons decay either hadronically or
  leptonically.
• W decays define channel
• Dilepton 5
• Leptonjets 30
• All-hadronic 44
• Tausanything 21
• Typical event selections
• Well-identified electron(s) or muon(s)
• Large missing ET
• Several reconstructed jets identified in
  calorimeters

6
Motivation for Mtop inside SM framework

• At tree level, EW theory
• depend on 3 quantities, e.g. can choose
• Fine structure constant a
• Fermi constant GF
• Z boson mass MZ
• However, radiative corrections
• must be included

• Example W boson mass
• Receives corrections dominated by Mtop
• quadratic dependence on Mtop!

7
Motivation for Mtop inside SM framework

• Many corrections to SM predictions of EWK
  measurements dominated by Mtop
• Therefore, high accuracy Mtop measurement is
  crucial for
• Constrain unknown model parameters ? MHiggs
• Consistency check of SM

8
Motivation for Mtop beyond SM

• Top quark mass so large, can it be a special
  particle?
• Beyond SM top quark mass important for many
  theories
• mSugra
• dynamical EWSB theories (topcolor)
• MSSM superparticles constraints (assuming Higgs
  discovered)

• Precise measurements of the top quark
• mass will help constrain beyond SM theories

9
Summary of Run I Measurements

• Mtop meas. in Run I (100pb-1) world ave.
• Higgs mass fit
• Run II goals per experiment

10
The Tevatron

• Proton-antiproton collisions at
• Run II c.m. upgrade (1.8 TeV to 1.96 TeV)
• ? 35 increase in

11
The CDF Detector

• Calorimeters
• Central, wall, plug calorimeter
• Coverage
• EM reso.
• HAD reso.

• CDF II general purpose
• solenoidal detector
• 7 layers of silicon tracking
• B-tagging ttbar eff. 55
• COT drift chamber
• coverage
• Resolution
• Muon chambers
• Scintillator, proportional chamber interspersed
  with absorber
• Provide muon ID up-to

12
Challenge I for Measuring Mtop

• Statistical limitations (for leptonjets
  channel)
• Small statistics 25 b-tag leptonjets ev. / 100
  pb-1 for CDF
• Complicated final state to reconstruct
• Observed ljets final state
• Complications
• 12 possible jet-parton assignments (if 4-jets)
• B-tagging helps a lot

13
Challenge II for Measuring Mtop

• Jet energy scale uncertainty
• Dominant systematics
• Comes from uncertainty in modeling the behavior
  of jets (particles response, fragmentation)
• Current world average uncertainty (4.3GeV/c2)
• 2.6 GeV/c2 from JES
• 2.7 GeV/c2 from stat.
• As more data is accumulated JES will become
  dominant
• Run II JES uncert. 3GeV/c2

14

• Traditional Template Analysis (LeptonJets)

(Luminosity 318pb-1)
15
Selecting Leptonjets Events

• Lepton selection
• One isolated e or µ ET (pT)gt20GeV with ID
• Electron EM cluster with matched track
• Muon track matched to hits in muon chambers, MIP
  ionizing energy in calorimeter

• Large missing ETgt20GeV
• Measured in calorimeters
• gt4 jets
• Reconstructed with cone algorithm from
  calorimeter towers

Four events category
16
Top Quark Mass Reconstruction

• Chi-square kinematic fit

• Try all jet-parton assignments
• Assign b-tag jets to b-quarks
• Keep the mass from assignment yielding lowest ?2

More correct combinations with b-tags!
17
Mtop Templates

• Get analytical functions of reconstructed mass as
  a function of true top mass
• This way obtain smooth templates versus top mass

Reco. Mass and p.d.f.s vs true Mtop
18
Background contamination

• Background sources
• gt1-tag events
• Wjets mistag
• QCD multijets
• Wheavy flavor
• Others (WW/WZ, single top)
• 0-tag events
• Wjets
• QCD multijets

• Use MC events to get background
• expected distributions

19
Extracting Mtop

• Unbinned likelihood fit
• Compare data reconstructed mass distributions
  with templates extracted from Monte Carlo
• Background constrained to expectations
• Combined fit multiply subsamples likelihood

• Likelihood fit sanity check

20
Systematic Uncertainties
CDF Run II Preliminary (318 pb-1)
JES dominates!
CDF Run II Preliminary (318 pb-1)
21
Results on Data

• Fit applied to data
• Final result

22
Template Analysis with In situ W?jj Calibration
Yield best measurement of Mtop!
23
Jet Energy Scale and W?jj

• JES, future limitation for Mtop
• Uncertainties from JES already dominate
  statistical uncertainty
• Knowledge of JES will improve, but has
  limitations
• Understanding of QCD
• LHC top measurements will be even more
  compromised by JES

• How does it work?
• Identify jets coming from W
• Reconstruct their invariant mass mjj
• mjj strongly dependent on JES (GeV)
• MW uncertainty is completely negligible (lt 50
  MeV)
• mjj mostly independent of Mtop

JES calibration using W?jj can provide an elegant
solution!
JES from W?jj is mostly statistical ? scale
with luminosity!
24
Reconstructing W?jj

• How to reconstruct Mjj?
• Which jet comes from W?
• No ambiguity when 2 b-tags
• Otherwise keep all possible mjj and consider
  them equally
• 1,3,6 mjj per event with 2,1,0 b-tag
• This method works well because
• Less combinatorics for mjj than Mtop
• 25 better uncert. than kinematic fit

25
JES Constraint from W?jj

• mjj templates
• mjj varies significantly as a function of JES
• mjj approximately independent of Mtop
• mjj determines JES with little uncertainties
  from Mtop!

26
Application to Mtop Measurement?
1) Can we use W?jj to calibrate b-jets?

• 2) How to take into account correlations
  Mtop-JES?
• W?jj display some dependence on Mtop
• Therefore, fitted JES is correlated to true top
  mass
• Solution simultaneous fit of Mtop and JES

CDF Run II Preliminary (318 pb-1)
? Most b-jets energy scale can be set using W?jj
27
Adding a Dimension to Mtop Measurement

• Show 2-D templates Mtop, JES

• Extend the machinery
• Determine simultaneously Mtop and JES
• mt and mjj templates
• Each template depend on Mtop and JES
• Unit of JES 1s as defined by CDF jet group
• A priori knowledge is used as a constraint in
  likelihood

28
Results on Data I
Reconstructed top mass (318pb-1)
29
Result on Data II

• Very good agreement data-MC JES
• Combined W?jj and prior JES yield 20 improvement

Reconstructed dijet mass
30
Results on Data III
31
Implications for Higgs

• New result is world best measurement of Mtop!

Yield relevant constraint on the SM Higgs mass
(only based on this measurement)
(Run I world ave. constraint
32
SUSY and Mtop

• Compute SUSY particles corrections to EM
  observables (Heinemeyer et al. hep-ph/0412214)
• Run I data already slightly favors MSSM over SM
  ?2/d.o.f.27.2/16 (SM)
• 16.4/12 (MSSM)
• New CDF result favors lower SUSY mass scales

33
Future of Analysis with W?jj

• Improvement to traditional calibrations of JES
  expected to be limited in the future
• Using W?jj JES uncertainty becomes mostly
  statistical
• Can reach JES uncert. below 1 GeV/c2 in
• Run II
• Total Mtop uncertainty can reach 2 GeV/c2

34
Other Methods

• (Results with luminosity
• lt 200pb-1)

35
Dynamical Likelihood Method

• Similar to Run I best measurement (DØ) (Nature
  429, 638 (2004))
• Calculates probability for each event to be
  signal with a given top mass
• Calculation based on full Standard Model
  matrix-element
• Very statistically powerful!

Using 162pb-1 (syst. 6.2 GeV/c2)
New (very competitive) result expected in few
weeks!
36
Dilepton Analyses

• Unconstrained problem 2 neutrinos, 1 missing ET
  observables
• General approach assume some kinematic
  quantities are known (ex ? or F or neutrinos,
  PZttbar))
• Integrate of assumed quantities calculate weight
  that the assumed quantity explains the observed
  event given a true top mass

Neutrino weighting algorithm
New much improved result expected in few weeks!
37
More to Come

• More work on matrix-element analyses
• Application to dilepton events
• Background matrix-element
• Multivariate technique
• Use more variables, include W?jj constraints
• All-hadronic channel
• Challenging more background and combinatorics
• B-jets Lxy based measurement
• Statistically limited at Tevatron, but
  interesting because of very different systematics
  than other analyses

38
Conclusion

• Top quark mass measurement is one of highest
  priority of Tevatron
• Related to Higgs mass through radiative
  corrections

• Top mass measurement is a complicated task
• Few events available
• Event topology is complicated
• Large uncertainty from jet energy scale
• We demonstrated
• Several methods available to measure Mtop
• W?jj calibration can provide crucial improvement
  in understanding of jet energy scale
• very important for the future of the
  measurement at the Tevatron and LHC