tt + production cross-section measurement at s=13TeV in the - - PowerPoint PPT Presentation

Titas Roy

Florida Institute of Technology

1

tt̅+𝜹 production cross-section measurement at √s=13TeV in the semileptonic channel

USLUA meeting, October, 2018

Introduction and Motivation

• The top quark being the heaviest quark, the way it interacts with a gauge boson makes an

interesting study and can be a window for new physics.

• This measurement can also be used for:
• a direct measurement of top quark charge: qt

2∝ σtt̅γ

• top quark electroweak coupling as qt and anomalous magnetic moment can be studied

through this process.

• tt̅ charge assymetry by selecting tt̅+γ events.
• tt̅+γ : important background in many BSM searches and in Higgs analyses.

2

jbfbf bfjb fbb bfbf ffknf jsf

W ¯ t t W+ q ¯ q

`

¯ ν

`

¯ b ¯ q0 q b γ W ¯ t t W+ g g

`

¯ ν

`

¯ b ¯ q0 q b γ

3

Event and Object Selection

• Top quark pair selection:
• Lepton Selection : exactly 1 high PT ,

isolated lepton

• Muon PT > 30 GeV, Ele PT > 35 GeV
• >=4 jets (PT > 30 GeV), at least one

btagged

• Photon Selection:
• >=1 isolated photon in barrel, PT > 20

GeV

Events 0.2 0.4 0.6 0.8 1

3

10 ×

Data Z+jets γ + t t γ W+ t t γ Z+ γ t+ WW/WZ/ZZ Single top ttV W+jets Multijet Uncertainty

(13 TeV)

• 1

35.8 fb

CMS

Preliminary 1 ≥

b

4, N ≥

j

e+jets, N

Photon Et (GeV) 20 40 60 80 100 120 140 Data/MC 0.5 1 1.5

Events 0.5 1 1.5 2 2.5 3

3

10 ×

Data Z+jets γ + t t γ W+ t t γ Z+ γ t+ WW/WZ/ZZ Single top ttV W+jets Multijet Uncertainty

(13 TeV)

• 1

35.8 fb

CMS

Preliminary 1 ≥

b

4, N ≥

j

e+jets, N

N Jets 4 5 6 7 8 9 10 Data/MC 0.5 1 1.5

Events 0.2 0.4 0.6 0.8 1 1.2

3

10 ×

Data Z+jets γ + t t γ W+ t t γ Z+ γ t+ WW/WZ/ZZ Single top ttV W+jets Multijet Uncertainty

(13 TeV)

• 1

35.8 fb

CMS

Preliminary 1 ≥

b

4, N ≥

j

e+jets, N

(GeV)

T

Electron p 40 60 80 100 120 140 160 180 200 Data/MC 0.5 1 1.5

Photon ET (GeV)

Prefit Prefit Prefit

Work in Progress Work in Progress Work in Progress

4

Analysis Strategy

tt힬 tt W힬, Z힬 Other Bkg

Non-Top Top Genuine 힬 Fake 힬

ChIso ChIso

Genuine and fake 힬 can be distinguished with Ch. Iso variable

M3

M3

Top and Non-top samples can be distinguished with M3 variable

Charged Hadron Isolation - sum of the transverse momentum of all charged hadrons in a cone centred around a photon.

〉 Events / GeV 〈

3

10

4

10

Data Isolated γ + t t Nonprompt γ + t t Isolated t t Nonprompt t t Isolated γ V+ Nonprompt γ V+ Other Isolated Other Nonprompt Uncertainty

(13 TeV)

• 1

35.8 fb

CMS

Preliminary e+jets

Photon Ch. Had. Isolation (GeV) 2 4 6 8 10 12 14 16 18 20 Data/MC 0.5 1 1.5 〉 Events / GeV 〈 5 10 15 20 25 30 35

Data Isolated γ + t t Nonprompt γ + t t Isolated t t Nonprompt t t Isolated γ V+ Nonprompt γ V+ Other Isolated Other Nonprompt Uncertainty

(13 TeV)

• 1

35.8 fb

CMS

Preliminary e+jets

M3 (GeV) 100 150 200 250 300 350 400 450 500 Data/MC 0.5 1 1.5

Isolated

M3 - the invariant mass of the three jets with the highest summed pT, expected to reconstruct the top quark mass.

Prefit Prefit

We want to distinguish tt̅+𝜹 from background

Work in Progress

Work in Progress

Photon Categories

Through generator matching we divide our photons into 4 categories based on where it is radiated from.

• Genuine : if the photon has a leptonic, bosonic or quark parents
• Photon is actually a Misidentified electron
• Hadronic Photon: radiated from a hadronic source, example: pion
• Hadronic Fake: when the reconstructed photon cannot be matched to a generator level

photon.

Isolated (signal) photons: Genuine photons and Misidentified electrons NonPrompt (background) photons: Hadronic photons & hadronic fakes: All MC samples are divided into Isolated and NonPrompt templates

5

Photon Category

Genuine Mis-id ele Hadronic Photon Hadronic Fake

Events 0.5 1 1.5 2 2.5 3 3.5 4 4.5

3

10 ×

Data γ + t t t t γ t+ Single top W+jets Z+jets γ W+ γ Z+ WW/WZ/ZZ ttV Multijet

(13 TeV)

• 1

35.8 fb

CMS

Preliminary 1 ≥

b

4, N ≥

j

e+jets, N Work in Progress

6

Data based template for Hadronic/Fake photons

• Data based templates used to avoid any mismodeling in NonPrompt

photons

• We extract the shape of these background photons from the σiηiη

sideband region of Charged Hadron Isolation which is dominated by nonprompt photons.

η η

σ Photon 5 10 15 20 25

3 −

10 × Events 0.5 1 1.5 2 2.5 3

3

10 ×

Data Genuine Photons MisID Electrons Hadronic Photons Hadronic Fakes

(13 TeV)

• 1

35.8 fb

CMS

Preliminary e+jets

Photon σiηiη

Photon charged-hadron isolation (GeV) 2 4 6 8 10 12 14 16 18 20 Normalized events 0.02 0.04 0.06 0.08 0.1

sideband

η η

σ Data MC Nonprompt photons

(13 TeV)

• 1

35.8 fb

CMS

Preliminary 1b ≥ 4j ≥ e+jets,

Data σiηiη sideband barrel barrel

sideband region signal region

Work in Progress

Work in Progress

7

Control Regions in the Fit

Events / GeV 500 1000 1500 2000 2500 3000 3500 4000

Data Isolated γ + t t Nonprompt γ + t t Isolated t t Nonprompt t t Isolated γ V+ Nonprompt γ V+ Other Isolated Other Nonprompt Uncertainty

(13 TeV)

• 1

35.8 fb

CMS

Preliminary e+jets

>=4jets,0btag 2 4 6 8 10 12 14 Data/MC 0.5 1 1.5

Events

>=4jets, 0 btag

Events / GeV 200 400 600 800 1000

3

10 ×

Data γ + t t t t γ V+ Other Uncertainty

(13 TeV)

• 1

35.8 fb

CMS

Preliminary e+jets, 0-photon

0-photon CR 100 200 300 400 500 600 Data/MC 0.5 1 1.5

>=4jets, 0 btag CR 0-photon CR

>=4jets, 0 btag

To get a better handle on W/Z + gamma process To get a better handle on ttbar process

Work in Progress Work in Progress

Fit Results

8

• Measuring the following:
• Fiducial σtt̅+𝜹 : Semileptonic top decay with a photon of PT>20 GeV, in barrel
• Inclusive σtt̅+𝜹 : inclusive in top decay and fiducial in photon cuts (PT>20 GeV, in barrel)
• NonPrompt photon correction factor extracted from the fit.
• Ratio (R) of tt̅+𝜹 to tt̅ cross-section also calculated

θ Δ )/ θ

• θ

(

2 − 1 − 1 2

r Δ

0.05 − 0.05

Pull Impact σ +1 Impact σ

• 1

0.181 ± = 2.68 r

θ Δ )/ θ

• θ

(

2 − 1 − 1 2

CMS Internal

r Δ

0.05 − 0.05

Pull Impact σ +1 Impact σ

• 1

0.182 ± = 2.93 r

Work in Progress Work in Progress

Fitting to extract the inclusive cross-section

• We get the following preliminary results in the lepton+jets channel:
• Inclusive cross-section = 2.63 ± 0.14(syst) ± 0.07(stat)pb 

• Ratio = (3.21 ± 0.18(syst) ± 0.06(stat)) × 10-3
• Fiducial cross-section being finalised
• Final results coming soon!

9

Summary

Backup

10

11

Trigger

• Mu: HLT_IsoMu24_v or HLT_IsoTkMu24_v
• Ele: HLT_Ele32_eta2p1_WPTight_Gsf_v

Muons

• Tight muon
• Tight Muon ID
• PT > 30 GeV, η < 2.4
• RelIso < 0.15
• Loose Muon:
• Loose Muon ID
• PT > 15 GeV, η < 2.4
• RelIso < 0.25

Electrons

• Tight electron
• Tight Electron ID
• PT > 35 GeV, η < 2.1
• Pass gap b/w barrel and endcap
• Impact parameter cuts :

d0< 0.05(barrel), d0< 0.1 (endcap) dz<0.1(barrel), dz<0.2(endcap)

• Veto Electrons:
• Pass Electron Veto ID
• PT > 15 GeV, η < 2.5
• Pass gap b/w barrel and endcap
• Impact parameter cuts as above

Jets

• PT > 30 GeV, η < 2.4
• Tight JetID
• Remove Jets close to a lepton if

dR(jet,lep)< 0.4

• Remove Jets close to a photon if

dR(jet,pho)< 0.1

• btagging : deep csv > 0.8484

Photon

• PT > 20 GeV, η < 2.5 (barrel);

1.4442 > η > 1.566 (endcap)

• Cut Based Medium Photon ID

Object Selection

12

• TTbar selection:
• Good PV
• Trigger -
• Muon: HLT_IsoMu24_v or HLT_IsoTkMu24_v
• Electron: HLT_Ele32_eta2p1_WPTight_Gsf_v
• Exactly 1 high PT, isolated lepton
• No loose leptons
• >= 4 Jets
• >= 1 Btagged Jets
• Overlap removal for the following samples: b/w ttgamma & ttbar, b/w VJets

& VGamma, b/w TGJets (t-channel) & ST (t-channel).

• Photon Selection
• Passes Preselection and has >= 1 photon

Event Selection

13

Overlap Removal

• To avoid double counting the same event in two different MC sample
• verlap removal is used.
• For example in tt̅+𝜹 and tt̅ sample, the same event maybe duplicated. To

avoid counting the same event twice, if an event in tt̅ is already present in tt̅+𝜹 kinematic phase space, then the event is removed from tt̅.

• Overlap removal is based on generator level kinematic cuts of the photon in

each sample.

• Overlap removal is performed between :
• tt̅+𝜹 & tt̅ (events removed from tt̅)
• V+𝜹 & V+jets (events removed from V+jets)
• ST+𝜹 & ST (events removed from ST)

14

Corrections to Simulation

• Correcting events with MisIdentified electrons
• Z->e+e- (e->𝜹): electron is mis-reconstructed as a photon.
• 3jets, 0btag CR: rich in background processes
• Fit to invariant mass of e+gamma

the uncertainty in the scale factor comes from:

• statistical uncertainty
• variation between different jet multiplicty

regions

• variation due to using a LO ZJets sample vs

a NLO ZJets sample (currently using LO ZJets)

Events/10 GeV 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8

3

10 ×

Data Sum MisIdentified Electron Background

(13 TeV)

• 1

35.8 fb

CMS

Preliminary e+jets

mass (GeV) γ e+ 40 60 80 100 120 140 160 Data/MC 0.5 1 1.5

Postfit

MisIDEle SF: 2.2 +/- 0.5

15

Fit parameters to extract tt̅+𝜹 and tt̅ cross-section

Sample Isolated Photon NonPrompt Photon t¯ t + γ 64.3% 0.39% t¯ t 4.6% 18.49% V + γ 4.6% 0.45% Other Bkg 4.4% 2.5%

• Three free floating parameters in this fit: σ(tt̅+𝜹),

NonPrompt photon correction factor, σ(ttbar)

• Backgrounds are floated with a gaussian constraint: V+γ

and Other bkg - 30%

16

σt¯

tγ,inlc(pT > 20GeV, barrel) = σt¯ tγ,inlc(pT > 13GeV, η < 2.5) × Aγ

σt¯

tγ,fid(pT > 20GeV, barrel) = σt¯ tγ,inlc(pT > 20GeV, barrel) × At¯ t

Aγ = Events with photon in the barrel and pT > 20 GeV Total events

At¯

t = Events with top quark pairs in signal phase space

Events with photon in the barrel and pT > 20 GeV

Cross-section calculation