Search for exotics with top at ATLAS
Diedi Hu
Columbia University
On behalf of the ATLAS collaboration
EPSHEP2013, Stockholm
July 18, 2013
Motivation
Motivation
Due to its large mass, top quark offers a unique window to search for new physics:
mt ∼ ΛEWSB (246GeV): top may play a special role in the EWSB Many extensions of the SM explain the large top mass by allowing the top to participate in new dynamics.
Approaches of new physics search Measure top quark properties: check with SM prediction.
Directly search for new particles coupling to top
tt resonances searches Search for anomalous single top production Vector-like quarks searches
(Covered in another talk ”Searches for fourth generation vector-like quarks with the ATLAS detector”)
- Search for exotics with tt and single top
- D.Hu, Columbia Univ.
- 2 / 16
−1
- tt resonance search with 14.3 fb
- @ 8TeV
tt resonance: benchmark models
2 specific models used as examples of sensitivity
Leptophobic topcolor Z0 [2]
Explains the top quark mass and EWSB through top quark condensate Z0 couples strongly only to the first and third generation of quarks
Narrow resonance: Γ/m ∼ 1.2%
Kaluza-Klein gluons [3]
Predicted by the Bulk Randall-Sundrum models with warped extra dimension Strongly coupled to the top quark
Broad resonance: Γ/m ∼ 15%
- Search for exotics with tt and single top
- D.Hu, Columbia Univ.
- 3 / 16
−1
- tt resonance search with 14.3 fb
- @ 8TeV
tt resonance: lepton+jets(I) [4]
Backgrounds SM tt, Z+jets, single top, diboson and W+jets shape: directly from MC W+jets normalization, multijet: estimated from data
Event selection
The boosted selection
One top decays hadronically and the other semileptonically.
Boosted selection helps improve efficiency in high signal mass region
1 high pT R=1.0 jet with
√
mjet > 100GeV, d12 > 40GeV;
1 R=0.4 jet close to lepton;
20
ATLAS Preliminary Simulation
s=8 TeV
18
≥ 1 R=0.4 b-jet(could overlap
16 µ + jets, combined
with the former 2 jets.)
µ + jets, boosted e + jets, combined e + jets, boosted
14 12 10 8
- 1 isolated lepton; ETmiss
- .
Resolved selection
6
3 or 4 R=0.4 jets with ≥ 1 b-jet
4
- 1 isolated lepton; ETmiss
- .
20
- 0
- 0.5
- 1
- 1.5
- 2
- 2.5
- 3
- 3.5
(Same as boosted channel)
mtt [TeV]
- Search for exotics with tt and single top
- D.Hu, Columbia Univ.
- 4 / 16
−1
- tt resonance search with 14.3 fb
- @ 8TeV
tt resonance: lepton+jets(II)
tt invariant mass reconstruction
pz (ν): from lepton+ETmiss system with mW constraint
Resolved: χ2 algorithm using mt and
Boosted: take R=1 jet as
- hadronically decaying top; build the
- mW , top pair pT balance as constraint
other top from ν, lepton and R=0.4 jet
0.3
ATLAS Preliminary
Simulation, s=8TeV
ATLAS Preliminary
Simulation, s=8TeV
0.25
0.25
Boosted
Resolved
0.2
Z’ (1.0 TeV)
Z’ (0.5 TeV)
0.2
Z’ (1.5 TeV)
Z’ (1.0 TeV)
Z’ (2.0 TeV)
Z’ (1.5 TeV) Z’ (2.0 TeV)
0.15
0.15
Z’ (3.0 TeV)
0.1
0.1
0.05
0
0.05
0
- 0
- 0.5
- 1
- 1.5
- 2
- 2.5
- 3
- 3.5
[TeV]
- 0
- 0.5
- 1
- 1.5
- 2
- 2.5
- 3
- 3.5
[TeV]
reco tt
reco tt
m
m
The tt mass spectra are used for statistical analysis
- Search for exotics with tt and single top
- D.Hu, Columbia Univ.
- 5 / 16
−1
- tt resonance search with 14.3 fb
- @ 8TeV
tt resonance: lepton+jets(III)
108 107 106 105 104 103 102 10
Dominant systematics on yields
5 × Z’ (1.5 TeV) 5 × (2.0 TeV)
ATLAS Preliminary
Data g
-1
L dt = 14.2 fb
∫
tt
KK
W+jets
Systematics Resolved Boosted
Multi-jets
Note: R=1. jet systematics
Other Backgrounds
tt
s = 8 TeV
normalization
- 8.0%
- 9.0%
also affect the resolved channel because only events failing the boosted selection will be examined with the resolved selection criteria.
JES of R=0.4 jets JES+JMS of R=1. jets
- 6.0%
- 0.70%
0.30%
4.0% 2.9%
17% 3.4% 6.0%
1
1.5 0
- 0.5
- 1
- 1.5
- 2
- 2.5
- 3
- 3.5
b-tag efficiency
PDF
1
0.5
- 0
- 0.5
- 1
- 1.5
- 2
- 2.5
- 3
- 3.5
reco
Exclusion ranges @ 95% C.L.
- m
- [TeV]
tt
14.3 fb−1 @ 8TeV [4]
s = 8 TeV
Obs. 95% CL upper limit Exp. 95% CL upper limit Exp. 1σ uncertainty
s = 8 TeV dt = 14.3 fb-1
Obs. 95% CL upper limit Exp. 95% CL upper limit Exp. 1 σ uncertainty
103 102 10
103 102 10
L dt = 14.3 fb-1
L
∫
∫
- Exclusion ranges
- Observed
Exp. 2 σ uncertainty
Exp. 2 σ uncertainty
Z0 0.5-1.8TeV
Kaluza-Klein gluon (LO)
Leptophobic Z’ (LO x 1.3)
ATLAS Preliminary
ATLAS Preliminary
KK gluon 0.5-2.0TeV
1
4.7 fb−1 @ 7TeV [5]
Exclusion ranges
1
10-1 10-2
10-1 10-2
Observed
Z0 0.50-1.74TeV
- 0.5
- 1
- 1.5
- 2
- 2.5
mass [TeV]
- 0.5
- 1
- 1.5
- 2
- 2.5
- 3
KK gluon 0.70-2.07TeV
g
Z’ mass [TeV]
KK
- Search for exotics with tt and single top
- D.Hu, Columbia Univ.
- 6 / 16
−1
- tt resonance search with 14.3 fb
- @ 8TeV
tt resonance: full hadronic decays [6]
102 10
Two complementary
Obs. 95% CL upper limit
Obs. 95% CL upper limit Exp. 95% CL upper limit Exp. 1σ uncertainty Exp. 2 σ uncertainty KK gluon (LO)
102
Exp. 95% CL upper limit Exp. 1σ uncertainty
techniques relying on jet
Exp. 2 σ uncertainty
10
1
Leptophobic Z' (LOx1.3)
substructure are used to identify top:
ATLAS
HEPTopTagger
ATLAS
HEPTopTagger
1
HEPTopTagger
10-1
s = 7 TeV
s = 7 TeV
Top Template Tagger
L dt = 4.7 fb-1
L dt = 4.7 fb-1
10-1
∫
∫
Exclusion ranges @ 95% C.L.
- 0.6
- 0.8
- 1
- 1.2
- 1.4
- 1.6
- 1.8
- 2
- 0.8
- 1
- 1.2
- 1.4
- 1.6
- 1.8
- 2
with 4.7 fb−1 @ 7TeV
gKK Mass [TeV]
Z' Boson Mass [TeV]
102 10
10
Use HepTopTagger
Obs. 95% CL upper limit Exp. 95% CL upper limit Exp. 1σ uncertainty Exp. 2σ uncertainty KK gluon (LO)
Obs. 95% CL upper limit Exp. 95% CL upper limit Exp. 1σ uncertainty Exp. 2σ uncertainty Leptophobic Z' (LOx1.3)
- Exclusion ranges
- Observed
Z0 0.70-1.00TeV
1
1.28-1.32TeV
1
KK gluon 0.70-1.48TeV
10-1
s = 7 TeV L dt = 4.7 fb-1
Use Top Template Tagger
s = 7 TeV L dt = 4.7 fb-1
10-1
∫
∫
- Exclusion ranges
- Observed
10-2
Top Template Tagger
Top Template Tagger
ATLAS
ATLAS
KK gluon 1.02-1.62TeV
- 1
- 1.2
- 1.4
- 1.6
- 1.8
- 2
- 1
- 1.2
- 1.4
- 1.6
- 1.8
- 2
gKK Mass [TeV]
Z' Boson Mass [TeV]
- Search for exotics with tt and single top
- D.Hu, Columbia Univ.
- 7 / 16
0
−1
@ 8TeV
W
→ tb search with 14.3 fb
W 0 → tb: theory [7]
W 0 carries new charged interaction as a result of enlarged symmetry group. Example of phenomenological models
Extra dimensions: Kaluza-Klein excitations of SM W Little Higgs: predicts new particles including W 0
Model-independent search
Use an effective model describing the coupling of the W 0 to the fermions.
Vi0j
0
fi γµg (1 + γ5) + gL0 (1 − γ5)W 0µfj + h.c. ¯
√
L =
Rij
ij
- 2
- 2
Search for W 0 in t+b decay channel
Explore models potentially inaccessible to leptonic search (W 0 → lν). WR0 can not decay to a lepton and νR if mν > mW
0
R
Leptophobic W 0 W 0 may couple more strongly to the third generation quarks
Can exploit the distinct signature of top quark.
- Search for exotics with tt and single top
- D.Hu, Columbia Univ.
- 8 / 16
0
−1
@ 8TeV
W
→ tb search with 14.3 fb
W 0 → tb: analysis strategy
Background
tt, Z+jets, single top, diboson, W+jets shape: from
MC
W+jets normalization, multijet: estimated from data
Event selection
Preselection: 1 lepton, ETmiss , 2 or 3 jets with ≥ 1 b-jet
Signal region: ≥ 2 b-jets, MW 0 ≥ 270GeV Control region:
≥ 2 b-jets, mW 0 < 270GeV: Derive W+jets normalization; verify the variable modeling =1 b-jet: check the modeling of kinematic variables
tb invariant mass reconstruction
pz (ν): from lepton+ETmiss system with mW constraint b from top: the b-jet/jet giving the mWb closest to mt b from W 0: the other b-jet or highest pT jet
- Search for exotics with tt and single top
- D.Hu, Columbia Univ.
- 9 / 16
0
−1
@ 8TeV
W
→ tb search with 14.3 fb
W 0 → tb: TMVA
Use Boosted Decision Tree discriminators (BDT)
1 BDT for 2-jet events:
14 variables used for training
Most discriminating variables: mtb, pT (t), ∆φ(l,top−jet)
1 BDT for 3-jet events:
13 variables used for training
Most discriminating variables: mtb, pT (t), sphericity
- Search for exotics with tt and single top
- D.Hu, Columbia Univ.
- 10 / 16
0
−1
@ 8TeV
W
→ tb search with 14.3 fb
W 0 → tb: result
Dominant systematics
tt MC generator
10%
b-tag efficiency
15% − 25% for events with ≥ 2b
The BDT output distributions are used for statistical analysis
Exclusion ranges @ 95% C.L. with 14.3 fb−1 @ 8TeV
Exclusion
- ranges
- Expected
- Observed
WL0 0.50 − 1.56TeV 0.50 − 1.74TeV WR0 0.50 − 1.72TeV 0.50 − 1.84TeV
- Search for exotics with tt and single top
- D.Hu, Columbia Univ.
- 11 / 16
∗
−1
@ 7TeV
b
→ Wt search with 4.7 fb
b∗ → Wt: theory [8]
Phenomenological models
Randall-Sundrum models; Composite Higgs models.
Consider b∗ produced singly through its coupling to a b and gluon
gs
2Λ
L = Gµνbσµν(kLbPL + kRb PR)b∗ + h.c.
¯
Exploit the weak coupling of excited-quarks to the third generation quarks
gs
+
µ
√
¯
L = 2 Wµ tγ (gLPL + gRPR)b∗ + h.c.
b∗ can also decay to bZ and bH cross-section and branching ratio of b∗ → Wt depend on the couplings and b∗ mass
general search for Wt resonances under 3 b∗ coupling scenarios
t
Left-handed: kRb = gR = 0
b∗
g
Right-handed: kLb = gL = 0
- b
- W
Vector-like: kLb = gL, kRb = gR
- Search for exotics with tt and single top
- D.Hu, Columbia Univ.
- 12 / 16
∗
−1
@ 7TeV
b
→ Wt search with 4.7 fb
b∗ → Wt: dilepton channel
Background
tt, Wt, diboson, Drell-Yan shape: from MC Drell-Yan normalization, and fake dileptons: estimated from data
Event selection
2 leptons with opposite charges, ETmiss 1 jet ee, µµ channels:
ATLAS
Data b* (800 GeV) Wt
10
-1
,
L dt = 4.7 fb
∫
tt Diboson Z(e+e-/µ+µ-)+jets Z(τ+τ-)+jets Fake dileptons
s = 7 TeV
1
10-1 10-2 10-3
m(ll) < 81GeV or m(ll) > 101GeV
Dilepton channel
- Veto to suppress Zττ
- :
∆φ(l
) + ∆φ(l
> 2.5
miss Tmiss T
- ,E
- ,E
)
- 1
- 2
Discriminating variable:
- 0
- 500
- 1000
- 1500
HT [GeV]
HT the scalar sum of ET of leptons, jet and
ETmiss
- Search for exotics with tt and single top
- D.Hu, Columbia Univ.
- 13 / 16