Quantum interference between doubly and singly resonant top quark - - PowerPoint PPT Presentation

Quantum interference between doubly and singly resonant top quark production

Christian Herwig University of Pennsylvania APS DPF 2017

The ttbar and tWb processes have the same final state and thus there exists a quantum interference effect The interference is largest when tWb "looks like" ttbar ATLAS separately generates ttbar and Wtb at NLO+PS with Powheg+Pythia

b~ 6 w- 4 t~ b 5 w+ t u 1 u~ 2 g

u t b g

W +

W −

¯ u ¯ t ¯ b

u 1 u~ 2 g b~ 4 b b 3 w+ 5 t w- 6

u ¯ u g b ¯ b t

W + W −

b

ttbar tWb

|AW W bb|2 ∼ |A(W tb)|2 + |A(tt)|2

+2R{A(W tb)A(tt)}

|AW W bb|2 ∼ |A(W tb)|2 + |A(tt)|2

+2R{A(W tb)A(tt)}

Interference effects are estimated by comparing two ad-hoc prescriptions: Diagram Removal (DR) and Diagram Subtraction (DS) Their difference is assigned as a systematic uncertainty

ATLAS separately generates ttbar and Wtb at NLO+PS with Powheg+Pythia The ttbar and tWb processes have the same final state and thus there exists a quantum interference effect The interference is largest when tWb "looks like" ttbar

The interference is largest when Wtb "looks like" ttbar

|AW W bb|2 ∼ |A(W tb)|2 + |A(tt)|2

+2R{A(W tb)A(tt)}

Interference effects are estimated by comparing two ad-hoc prescriptions: Diagram Removal (DR) and Diagram Subtraction (DS) Their difference is assigned as a systematic uncertainty

ATLAS separately generates ttbar and Wtb at NLO+PS with Powheg+Pythia6

Define Diagram Removal (DR) single top to take only the piece

A(W tb)

Add’l details: Frixione et al. arXiv:0805.3067

The ttbar and tWb processes have the same final state and thus there exists a quantum interference effect

The interference is largest when Wtb "looks like" ttbar

|AW W bb|2 ∼ |A(W tb)|2 + |A(tt)|2

+2R{A(W tb)A(tt)}

Interference effects are estimated by comparing two ad-hoc prescriptions: Diagram Removal (DR) and Diagram Subtraction (DS) Their difference is assigned as a systematic uncertainty

ATLAS separately generates ttbar and Wtb at NLO+PS with Powheg+Pythia6

Define Diagram Removal (DR) single top to take only the piece

A(W tb)

Define Diagram Subtraction (DS) single top as the entire expression, minus a gauge-invariant term that exactly cancels when A(t¯

t)

M 2

bW → m2 t

Add’l details: Frixione et al. arXiv:0805.3067

The ttbar and tWb processes have the same final state and thus there exists a quantum interference effect

arXiv:0805.3067

magenta = 2L2b inclusive selection

DS/DR disagreement large in extreme (search) phase space

6 Herwig (Pennsylvania)

arXiv:0805.3067 DS/DR disagreement large in extreme (search) phase space

7 Herwig (Pennsylvania)

"Start worrying"

magenta = 2L2b inclusive selection

Run 2 LHC has provided us with millions of Wt events Can we use this data to improve our understanding of tt-Wt interference?

8 Herwig (Pennsylvania)

p p e+ µ− b b

9 Herwig (Pennsylvania)

p p e+ µ− b b

b~ 6 w- 4 t~ b 5 w+ t u 1 u~ 2 g

u t b g

W +

W −

¯ u ¯ t ¯ b

10 Herwig (Pennsylvania)

p p e+ µ− b b

u 1 u~ 2 g b~ 4 b b 3 w+ 5 t w- 6

u ¯ u g b ¯ b t

W + W −

b

11 Herwig (Pennsylvania)

p p e+ µ− b b

12 Herwig (Pennsylvania)

p p e+ µ− b b

. mtop ? . mtop ?

13 Herwig (Pennsylvania)

흂 t t b1 1 흂 b2 2

14

Let mij = m(bi, `j), and define

Herwig (Pennsylvania)

min-max m(b, `)

) ≡ min {max(m11, m22), max(m12, m21)}

흂 t t b1 1 흂 b2 2

15

Let mij = m(bi, `j), and define

Important Properties: min-max mbl < mtop for ttbar events not necessarily for Wt events!

Herwig (Pennsylvania)

min-max m(b, `)

) ≡ min {max(m11, m22), max(m12, m21)}

wrong pairing: mbl usually large correct pairing: mbl bounded by mtop

흂 t t b1 1 흂 b2 2

16

[GeV]

bl

reco min-max m 50 100 150 200 250 300 350 400 450 500 Events / 20 GeV

1 −

10 1 10

2

10

3

10

4

10

5

10

Total SM t t Wt (DR)

Wt (DS)

ATLAS Work in progress simulation

• 1

= 13 TeV, 36.1 fb s

Wt (DR) purity 0.5 1 DS / DR 0.5 1

Analysis strategy: measure this spectrum!

Herwig (Pennsylvania)

흂 t t b1 1 흂 b2 2

17

[GeV]

bl

reco min-max m 50 100 150 200 250 300 350 400 450 500 Events / 20 GeV

1 −

10 1 10

2

10

3

10

4

10

5

10

Total SM t t Wt (DR)

Wt (DS)

ATLAS Work in progress simulation

• 1

= 13 TeV, 36.1 fb s

Wt (DR) purity 0.5 1 DS / DR 0.5 1

Analysis strategy: measure this spectrum!

Herwig (Pennsylvania)

All Wt here is NLO+PS Powheg+Pythia6 (DR and DS)

Fiducial region: exactly 2 leptons, exactly 2 b-tagged jets mll > 10 GeV and |mll-mZ| > 5 GeV

18 Herwig (Pennsylvania)

Fiducial region: exactly 2 leptons, exactly 2 b-tagged jets mll > 10 GeV and |mll-mZ| > 5 GeV Single lepton triggers lepton pT > 28 GeV b-jet pT > 20 GeV tag at 60% efficiency WP , veto at 85%

19 Herwig (Pennsylvania)

[GeV] 450 [GeV]

bl

reco min-max m 50 100 150 200 250 300 350 400 450 Events / 40 GeV

2

10

3

10

4

10

5

10

Total SM t t +hf t t Wt (DR) Z+jets +V t t Diboson

ATLAS Internal simulation

• 1

= 13 TeV, 36.1 fb s SR

(b) Logarithmic scale

ATLAS Work in Progress

20 Herwig (Pennsylvania)

dominant backgrounds estimated from data using dedicated control regions (CRs) ttbar and Wt are treated together as the signal process

Estimating tt with additional heavy flavor

dilep_mbl_3b 50 100 150 200 250 300 350 400 450 Events / 40 100 200 300 400 500 600

Data Total SM t t

1.8 ×

+hf t t

Wt (DR)

1.2 ×

Z+jets

+V t t

ATLAS Work in progress

• 1

= 13 TeV, 36.1 fb s

Data / SM 0.5 1 1.5

tt+b Control Region

3b mbl [GeV]

21 Herwig (Pennsylvania)

normalize tt+b in a dedicated 3 b-jet CR Solution: if the identified b-jets aren’t from top decays ttbar can pass the kinematic endpoint! Problem:

(with ttbar and Wt treated together)

[GeV]

bl

truth min-max m 50 100 150 200 250 300 350 400 450 [GeV]

bl

reco min-max m 50 100 150 200 250 300 350 400 450 a.u. 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9

Response Matrix

mbl spectrum is unfolded to particle-level

ATLAS Work in Progress

22 Herwig (Pennsylvania)

[1/GeV]

bl

d m σ d σ 1

5 −

10

4 −

10

3 −

10

2 −

10

1 −

10 1

pseudo-data stat unc. syst ⊕ stat PhPy6 DR PhPy6 DS

ATLAS Simulation Internal

Bayesian Unfolding with 3 iterations [GeV]

bl

minimax m

100 200 300 400

1/Data

0.5 1 1.5

Results (blinded) Total uncertainty in tail is dominated by data statistics Sensitive to DR/DS differences: shape+normalization Data compared to state-of-the-art generators

ATLAS Work in Progress

23 Herwig (Pennsylvania)

DR pseudodata

simulation

• 1

= 13 TeV, 36.1 fb s SR

Conclusions

We expect to reduce the systematic uncertainty associated with ATLAS’s treatment of the tt-Wt interference We present the first measurement of the WWbb final state in a region of maximal tt-Wt interference Measurement is sensitive to the large differences between state-of-the-art generators

24 Herwig (Pennsylvania)

Backup

minus one. Black solid, red dashed, blue dotted, and green dot-dashed lines correspond to p(veto)

T

= 10, 30, 50, and 70 GeV respectively. The magenta solid line with open boxes is

• btained without imposing any veto.