View
217
Download
0
Embed Size (px)
Citation preview
1
Viktor Veszprémi(Purdue University, CDF Collaboration)
SUSY 2005, Durham
Search for the SM Higgs Bosonat the CDF Experiment
Search for the SM Higgs Bosonat the CDF Experiment
2
Gluon fusion:
• Large cross section• But no search potential below Mh=~135 GeV
• Indistinctive signature of a pair of b-jets• Signal swamps in QCD multi-jet processes
“Higgsstrahlung” (W/Z associated production):
• Smaller cross section• Good search potential both at lower and higher Higgs mass• Easier to trigger on the decay products of the W/Z bosons
Viktor Veszprémi, SUSY 2005, Durham
Standard Model Higgs ProductionStandard Model Higgs ProductionStandard Model Higgs ProductionStandard Model Higgs Production
3
mH<114.4 GeV Excluded at 95 % C.L.
EWK fits prefer light SM Higgs < ~260 GeV
Exclu
ded
Exclu
ded
Mh (GeV)
Exluced by LEP
H->bb dominated region:2 analyses with results1 analysis in progress
H->WW dominated region:2 analysis result
Viktor Veszprémi, SUSY 2005, Durham
Standard Model Higgs Decay Standard Model Higgs Decay ModesModes
Standard Model Higgs Decay Standard Model Higgs Decay ModesModes
4
• Charged lepton final states:
• Low Higgs mass scenario
• Requires good b-jet selection, lepton identification, Missing Et reconstruction, di-jet invariant mass reconstruction
• High Higgs mass scenario
• Requires efficient lepton identification
• Final state with no charged leptons and low Higgs mass:
• Requires good b-jet selection, Missing Et reconstruction, di-jet invariant mass reconstructionViktor Veszprémi, SUSY 2005, Durham, England
CDF SM Higgs AnalysesCDF SM Higgs Analyses
5
Tagging B-jetsTagging B-jets
Tagging jets with the SecVtx algorithm is a powerful way of selecting b-jets
Jet is tagged:• decay length Lxy/Lxy > 7• Lxy>0 identifies real heavy flavor jet
Tagged jet
Prim. vertex
2nd vertexLxy>0
• Currently using the tight tagger
• Development of a “forward b-jet tagging” is in progress to increase acceptance
• Mistag rate is ~1%
Viktor Veszprémi, SUSY 2005, Durham, England
6
• Reconstuction of Z decaying to b-jets– Measure jet energy scale and
resolution– Provides a tool for
investigating b-jet specific jet energy corrections
• Looking for Z in double tagged events with – no additional jets above 10
GeV– Jets are back-to-back
topology
• 3394±515 Zbb events were found in a sample of 85,784 double-tagged events. (333pb-1)
• We can observe the Z peak
Reconstruction of Zbb decays in CDFReconstruction of Zbb decays in CDF
Viktor Veszprémi, SUSY 2005, Durham , England
7
WWHgg
W+ e+
W- e-
Event Selection:
• exactly two identified leptons (veto WZ, ZZ)
• no jets with Et > 15 GeV and || < 2.5 (veto top-pairs)
• > 25 GeV
• If < 50 GeV: (veto Z/)
• not 76 < Mee/uu < 106 unless Missing ET significance larger than 3 (veto WZ,ZZ,Z/)
• leptons having opposite charge signs (veto WZ)
• di-lepton invariant mass Mll < half the Higgs mass. (discriminates from WW, WZ, ZZ)
2~ h
ll
MM
Spin of Higgs is 0
W spins are anti-parallel
Leptons from W’s arealigned
Di-lepton mass isconstrained
Viktor Veszprémi, SUSY 2005, Durham , England
20),( leptonclosestETTE
TE
8
WWHgg
(SM) sourceExpectation
(Mh=180 GeV)
WW 6.49 +/- 0.76WZ 0.18 +/- 0.02ZZ 0.06 +/- 0.01
Drell-Yan ee 0.87 +/- 0.44DY mumu 0.43 +/- 0.19DY tautau 0.03 +/- 0.01
ttbar 0.02 +/- 0.01fakes 0.81 +/- 0.25
total bg 8.90 +/- 0.98HWW 0.17 +/- 0.02data 8
Small di-lepton invariant mass
Leptons originating from the signal process tend to be back-to-back in the transverse plain
Viktor Veszprémi, SUSY 2005, Durham , England
9
WWHgg
tlltC EMpM 22
Clustered mass:
The 95% C.L. effective cross-section limit of the Higgs boson
Viktor Veszprémi, SUSY 2005, Durham , England
10
• Like-sign di-lepton selection (“1”=more, “2”=less energetic)
• pT,1>20 GeV, pT,2>6 GeV• reject conversions, cosmics, Zleptons • • Signal region: pT,2>16 GeV, for mH<160 GeV• Harden pT,2 cut to 18 GeV for mH>160 GeV
Category Events in 193.5 pb-1
Conversions 0.61 0.61
Fake Leptons 0.12 0.01
Other sources* 0.22 0.10
Total background 0.95 0.64
Data 0
*Other backgrounds: Di-boson, top, WqqSM WH signal: 0.03 events (mH=160 GeV)
Viktor Veszprémi, SUSY 2005, Durham , England
11
Results
Viktor Veszprémi, SUSY 2005, Durham, England
Acceptance includes W branching fractions
12
bbHW Signature:
• isolated lepton with ET > 20 GeV or pt > 20 GeV in ||<1.0
• Missing ET > 20 GeV
• 2 taggable jets (require 1 or 2 tags)
Below is a double-tagged WH event
Applied selection:
• Events with an additional lepton with sign opposite to the primary is removed (Z and conversion removal)
• 2-jet bin contains the signal
• inclusive single / exclusive double tag is required -> single tag search is more sensitive
Viktor Veszprémi, SUSY 2005, Durham , England
13
bbHW
Background Source Events in 319 pb-
1
W+light flavor 39.3 3.1
54.0 18.4
19.5 6.6
16.8 4.3
Di-boson+Z+- 5.0 1.1
Non-W 16.5 3.2
14.1 2.5
9.6 2.0
Total Background 174.7 26.3
Observed Data 187
bbW ccW cW
tt
bt
Viktor Veszprémi, SUSY 2005, Durham, England
Number of background events in the 2-jet bin:
W+h.f. normalization
Signal Top background normalization
14
Dijet invariant mass in the inclusive single tag sample
The 95 % C.L. limit set on the effective cross-section of the Higgs boson
Viktor Veszprémi, SUSY 2005, Durham , England
bbHW
15
Z-> ν ν, H-> b bZ-> ν ν, H-> b b
• Signal has a distinctive topology
• Large missing transverse energy
• two jets (one is b-tagged)
Missing ETb-jet
b-jet
y
x
• This signature proved to be the most sensitivein the Run I and recent analyses at CDF
Viktor Veszprémi, SUSY 2005, Durham , England
Basic Selection cuts:
• Jet requirements
• 1st Jet ET > 40 GeV
• 2nd Jet ET > 20 GeV
• No other jet with ET> 15 GeV
• ET > 70 GeV
• Minimum 1 b-tag
16
Analysis OutlineAnalysis Outline
We performed a blind analysis
Extended Signal Region
• Veto events with leptons
• Missing ET and 2nd leading jet are not parallel
• Cut optimization is performed in this region based on Monte Carlo simulation before looking at the real data
Viktor Veszprémi, SUSY 2005, Durham , England
Control Region 1 – QCD h.f.
• Veto events with identified leptons
• Missing ET and 2nd leading jet are parallel
Control Region 2 – EWK
• Require min. 1 lepton
• Missing ET and 2nd leading jets are not parallel
• Optimized cuts are tested in this region before looking at the real data in the Signal Region
17
QCD heavy flavor normalizationQCD heavy flavor normalization
Control Region 1: • QCD multi-jet is the most dominant process because of its large cross-section
• A good description of this background allows us to efficiently reduce it with optimized cuts
Viktor Veszprémi, SUSY 2005, Durham , England
2nd jet
Fake Missing ET
1st jet
~180o
A di-jet QCD event:
QCD events also have a particular topology:
• jets are back-to-back
• “Missing ET“ points along the leading jet• Events are simulated by a b-filtered
Pythia Monte Carlo sample which is normalized to data
• Mistags (tagged light flavor jet events) are estimated from Data before tagging
18
Cut OptimizationCut Optimization
Viktor Veszprémi, SUSY 2005, Durham , England
The following optimized cuts were developed on a benchmark signal MC sample (Mh=120 GeV)
Selection cut ZH 120
288.9 pb-1
Acceptance (%)
Basic Cuts 0.2050.0035 5.92 0.1 0.026
0.205 0.0035 5.920.1 0.027
significance 0.1830.0033 5.230.095 0.031
0.1610.0031 4.680.089 0.037
Di-jet mass cut 0.1260.016 3.640.079 0.062
8.0,1 Tst EJet
THGeVJetET
st 601
• The last selection is on the di-jet mass. A 20 GeV mass window is applied to the data and Monte Carlo around the reconstructed Higgs mass peak
• Performed a counting experiment
BS /
19
ResultResult
Viktor Veszprémi, SUSY 2005, Durham , England
The top plot : di-jet invariant mass in Control Region 2 after optimization, before the mass-window cut
The bottom plot is the same in the Signal RegionFor the 120 GeV Higgs mass we apply a 80-120 GeV window cut:
SM background prediction: 4.36 1.02 events• QCD : 11.4%• Top : 20.5%• EWK : 18.2%• Light flavor mistag: 50%
Observed: 6 events.With a 95 % C.L. we expect the limit to be for the
Observed limit:
pbbbBRZH 4.16.3)()(
pbbbBRZH 5.4)()(
20
Signal Event CandidateSignal Event Candidate
Viktor Veszprémi, SUSY 2005, Durham , England
Data events from Signal Region• passing all selection cuts• candidate in the 80-120 GeV mass window
Double-tagged event
Missing ET
144.8 GeV
Di-jet invariant mass = 82 GeV
Leading Jet ET = 100.3 GeV
Second Jet ET = 54.7 GeV
21
Summary of Limits set at CDFSummary of Limits set at CDF
Viktor Veszprémi, SUSY 2005, Durham , England
• First analysis with 200-300 pb-1 data completed• Nearly 1 fb-1 on CDF tape• Estimate sensitivity:
• 2 fb-1: exclude Higgs up to mass of 120 GeV• 8 fb-1: exclude Higgs up to mass of 135 GeV
22
Backup Slides
23
FY02FY03
FY04
FY05
Increasing Tevatron LuminosityIncreasing Tevatron Luminosity
Viktor Veszprémi, SUSY 2005, Durham , England
Tevatron8 fb-1
Tevatron2 fb-1LE
Pexcl
ud
ed
Future expectations based on the Higgs sensitivity study in 2003.
Below is the instantaneous and integrated luminosities recorded at CDF so far
24
Missing ET ReconstructionMissing ET Reconstruction
2nd jet
Fake Missing ET
1st jet
180o
A di-jet QCD event:
QCD events also have a particular topology:
• jets are back-to-back
• Fake missing ET points along the leading jet
Real missing ET – produced by undetected particles
• Neutrinos
Fake missing ET
• Beam/detector effects• A set of clean-up cuts are applied to eliminate beam effects
• Unbalanced, mis-measured jets
• The primary source of ET in QCD events
• Since the cross-section of QCD is large, most of the triggered events are like this
• Makes QCD the main background :
Viktor Veszprémi, SUSY05, Durham, England