Semi-leptonic and Dileptonic Top-Quark Decays at ATLAS Raphael - - PowerPoint PPT Presentation

Semi-leptonic and Dileptonic Top-Quark Decays at ATLAS

Raphael Mameghani

IMPRS/GK Young Scientist Workshop at Ringberg

23rd July 2007

1/ 25

Outline

1

Top pair production at LHC & ATLAS

2

Ratio of semileptonic and fully leptonic decays

3

Semileptonic channel

4

Dileptonic channel

5

Summary & outlook

2/ 25

Large Hadron Collider

pp accelerator with 27 km circumference superconducting magnets √s = 14TeV design luminosity: 1034cm−2s−1 (start: 1033cm−2s−1) bunch crossings every 25 ns

3/ 25

ATLAS

width: 44m, height: 22m, weight: 7000t together with CMS the universal detectors of LHC toroid magnet containing the muon system solenoid magnet for the inner detector

4/ 25

Top Pair Production at the LHC

Mainly (87%) gg → t¯ t: But also (13%) q¯ q → t¯ t: Expectation: σ(pp → t¯ t) ≈ 830 pb (NLO) 8·106 events per year at

L dt = 10 fb−1

5/ 25

Top Pair Decays

t → W + b (≈ 100%) W → e, µ, τ + ν (each 1/9) W → q¯ q (2/3) combinatorics ⇒ 5%: signature 2 jets 4 2 thereof b-jets 2 2

• char. leptons

1 2 missing ET 1 30%:

6/ 25

Reasons for a Ratio Measurement

A cross section ratio might compensate Experimental Uncertainties luminosity (as N = σ ·

L dt)

(energy and momentum scale uncertainties which might affect counting efficiencies) Theoretical Uncertainties parton density functions unknown effects of higher order But: two channels

7/ 25

Branching Ratio Fully Leptonic / Semileptonic

Standard Model expectation Rℓℓ/ℓ = BR(t¯

t→2ℓ+2ν+2q) BR(t¯ t→ℓ+ν+4q) = Nℓℓ/Nℓ = 1/6

discrepancies might occur due to rare top decays, e.g. t → H+ + b H+ → τν,¯ cs ⇒ deficit of electrons and muons Previous Examination (ATLAS Design Report) statistical precision for 1 year with

L dt = 10 fb−1:

∆Rℓℓ/ℓ/Rℓℓ/ℓ ≈ 0.5% Smeared 4-vectors with the cuts pT(ℓ) > 20GeV, E /T> 20GeV,

• min. 2 b-jets with pT > 20GeV

Here with full detector simulation and no b-jet identification

8/ 25

Semileptonic Channel

Semileptonic Channel

9/ 25

Signal & Background

Semileptonic Signal (≈ 250pb) t¯ t → ℓν +4jets, with ℓ = e,µ (MC@NLO) Other No All Hadronic t¯ t Background (≈ 210pb) t¯ t → ℓνℓν +2jets,τ +X +2jets (MC@NLO) hadronic τ decays look quite similar to jets in the detector, in the following assumed that τ identification is not working W + jets Background (≈ 790pb) q¯ q → W → ℓν + QCD jets ALPGEN & HERWIG irreducible contribution (apart from kinematics)

10/ 25

QCD Background I

Only ATLFAST simulation available (Generators: Alpgen + Pythia) Cross sections: # Alpgen partons cross sections [pb] 3 4 766 000 4 549 000 5 64 000 6+ 30 000 Fake Leptons/Electrons A jet may be reconstructed as an electron Together with E /Tfrom limited energy resolution ⇒ Background for semileptonic channel Would need 100 million fully simulated events! Fake jets manually: P(jet → e) = 10−3 = const. assumed

11/ 25

QCD Background II

partons events fake electron events factor 3 225 834 402 076 1.78 4 200 000 508 064 2.54 5 430 106 1 466 860 3.41 6+ 446 900 2 249 220 5.03 σeff = σ ·P(j → e)·factor partons σeff [pb] 3 8483 4 1394 5 218 6+ 151 No energy dependence of P(j→e) considered here Correlations as each jet in event is used once as an electron

12/ 25

Cut Flow

1 2 3 4 5 6 7 8 9

3

10

4

10

5

10

6

10

7

10 1 2 3 4 5 6 7 8 9

3

10

4

10

5

10

6

10

7

10 n events

4j ν l → t t X τ 2j, ν l ν l → t t + 4j ν τ , ν l → W QCD fake l

red: t¯ t semileptonic brown:

• ther no all

hadronic t¯ t blue: W + jets yellow: QCD with fake electrons 1: 1 ℓ PT > 20 GeV, |η| < 2.5 2: 3 jets > 40 GeV + 1 jet > 20 GeV (all |η| < 2.5) 3: E /T > 20 GeV 4: in 3 jets with max. ∑ PT : |m2jets −mW | < 10 GeV 5: mtotal < 900 GeV 6:

• lead. 3 jets |cosθ ∗|’s < 0.7

9: without cuts 5 & 6 Cuts on missing transversal energy and number

• f

jets are the most important ones

13/ 25

Hadronic Top

Hadronic top mass:

/ GeV

top

M 50 100 150 200 250 300 350 400 450 500 Events 1000 2000 3000 4000 5000 6000 7000 / GeV

top

M 50 100 150 200 250 300 350 400 450 500 Events 1000 2000 3000 4000 5000 6000 7000 Mass of Hadronic Top after Preselection

4j ν l → t t X τ 2j, ν l ν l → t t + 4j ν τ , ν l → W QCD fake l

3 jets with highest vec- tor sum PT (t¯ t tend to recoil) But do not take the event if no pair within the 3 jets has an invari- ant mass less than 10 GeV away from the W mass!

14/ 25

Leptonic Top

With mass constraint two solutions for neutrino pz M2

W = (Eℓ +Eν,T )2 −(pℓ,x +pν,x)2 −(pℓ,y +pν,y)2 −(pℓ,z +pν,z)2

Take smaller solution

/ GeV

top

M 50 100 150 200 250 300 350 400 450 500 Events 2000 4000 6000 8000 10000 12000 14000 / GeV

top

M 50 100 150 200 250 300 350 400 450 500 Events 2000 4000 6000 8000 10000 12000 14000 Mass of Leptonic Top after Preselection

4j ν l → t t X τ 2j, ν l ν l → t t + 4j ν τ , ν l → W QCD fake l

Invariant mass from neutrino vector + lepton + remaining jet with highest PT

) / GeV

top,lept

• M

top,had

(M

• 200 -150 -100
• 50

50 100 150 200 Events 1000 2000 3000 4000 5000 6000 ) / GeV

top,lept

• M

top,had

(M

• 200 -150 -100
• 50

50 100 150 200 Events 1000 2000 3000 4000 5000 6000 Top Mass Difference after Preselection

4j ν l → t t X τ 2j, ν l ν l → t t + 4j ν τ , ν l → W QCD fake l

Difference between hadronic and lep- tonic top mass

15/ 25

Missing Transverse Energy before & after all Cuts

T

missing E 20 40 60 80 100 120 140 Events

4

10

5

10

6

10

T

missing E 20 40 60 80 100 120 140 Events

4

10

5

10

6

10 / GeV

T

Missing E

4j ν l → t t X τ 2j, ν l ν l → t t + 4j ν τ , ν l → W QCD fake l

T

missing E 20 40 60 80 100 120 140 Events 100 200 300 400 500 600

T

missing E 20 40 60 80 100 120 140 Events 100 200 300 400 500 600

(after cuts) / GeV

T

Missing E

4j ν l → t t X τ 2j, ν l ν l → t t + 4j ν τ , ν l → W QCD fake l

16/ 25

Top Masses after all Cuts

/ GeV

top

M 50 100 150 200 250 300 350 400 450 500 Events 500 1000 1500 2000 2500 / GeV

top

M 50 100 150 200 250 300 350 400 450 500 Events 500 1000 1500 2000 2500

Mass of Hadronic Top

4j ν l → t t X τ 2j, ν l ν l → t t + 4j ν τ , ν l → W QCD fake l

/ GeV

top

M 50 100 150 200 250 300 350 400 450 500 Events 500 1000 1500 2000 / GeV

top

M 50 100 150 200 250 300 350 400 450 500 Events 500 1000 1500 2000

Mass of Leptonic Top

4j ν l → t t X τ 2j, ν l ν l → t t + 4j ν τ , ν l → W QCD fake l

17/ 25

Without Cut on Missing Transverse Energy

/ GeV

top

M 50 100 150 200 250 300 350 400 450 500 Events 500 1000 1500 2000 2500 3000 3500 / GeV

top

M 50 100 150 200 250 300 350 400 450 500 Events 500 1000 1500 2000 2500 3000 3500

Mass of Hadronic Top

4j ν l → t t X τ 2j, ν l ν l → t t + 4j ν τ , ν l → W QCD fake l

T

missing E 20 40 60 80 100 120 140 Events 200 400 600 800 1000

T

missing E 20 40 60 80 100 120 140 Events 200 400 600 800 1000

(after cuts) / GeV

T

Missing E

4j ν l → t t X τ 2j, ν l ν l → t t + 4j ν τ , ν l → W QCD fake l

Hadronic top mass & missing ET without E /Tcut QCD background increases by ≈ 1 order of magnitude

18/ 25

Dileptonic Channel

Dileptonic Channel

19/ 25

MC Samples

Leptonic t¯ t Signal (≈ 40pb) t¯ t → ℓνℓν +2jets, with ℓ = e,µ Again no b quark identification info used in my analysis Other No All Hadronic t¯ t Background (≈ 420pb) t¯ t → ℓν +4jets,τ +X +2jets (5200 sample) Assume that τ identification is not available Z Background (≈ 5000pb) Z → ℓℓ + jets from hard interaction Diboson Background (≈ 35pb) ZZ,WZ,WW → leptons + jets from hard interaction

20/ 25

W + jets Background I

Cross sections: W decay cross sections [pb] eν 17 740 µν 17 740 τν 17 170 Fake Electrons A jet may be reconstructed as an electron E /T from neutrino ⇒ Background for di-leptonic channel Manually: P(jet → e) = 10−3 = const. assumed Random charge ±e assigned to fake electrons

21/ 25

W + jets Background II

W decay events fake electron events factor eν 347 500 342 285 0.985 µν 215 200 153 149 0.712 τν 98 007 85 976 0.877 σeff = σ ·P(j → e)·factor W decay σeff [pb] eν 17.2 µν 12.4 τν 15.1 No energy dependence of P(j→e) considered here Correlations as each jet in event is used once as an electron

22/ 25

Early cut flow - work in progress

1 2 3 4 5 6 7 8 9

2

10

3

10

4

10

5

10

6

10

7

10 1 2 3 4 5 6 7 8 9

2

10

3

10

4

10

5

10

6

10

7

10 n events

2j ν l ν l → t t X τ 4j, ν l → t t Y → ZZ, WZ, WW X + jets → Z X + jets → W

red di-leptonic signal brown

• ther t¯

t blue di-boson events green Z + jets yellow W + jets

cut explanation 1 missing ET > 35 GeV 2 exactly 2 recon. leptons 3

• pposite lepton charge

4

• min. 2 jets > 40 GeV

5 di-lepton mass cut

23/ 25

Di-lepton Mass

Invariant mass of two leading leptons:

/ GeV

ll

M 50 100 150 200 250 300 Events 200 400 600 800 1000 1200 1400 1600 1800 2000 2200 / GeV

ll

M 50 100 150 200 250 300 Events 200 400 600 800 1000 1200 1400 1600 1800 2000 2200 Mass of Dibosons/Z

2j ν l ν l → t t X τ 4j, ν l → t t Y → ZZ, WZ, WW X + jets → Z X + jets → W

veto for events with leptons of same kind: 85 GeV < Mℓℓ < 95 GeV Mℓℓ > 50 GeV

24/ 25

Summary & Outlook

Semileptonic channel: ATLFAST QCD samples give some impressions of the QCD background shape P(j → e) had to be assumed Cuts on E /T and the number of jets are the most important

• nes

Dileptonic channel: Model W + n jets as QCD in semileptonic channel For P(j→e) = 10−3 the contribution is still small but larger than the di-boson background Still working on cuts Together: Estimate achievable precision for the ratio Rℓℓ/ℓ in the early phase of LHC & ATLAS

25/ 25