[PPT] - Wild Card ATLAS Haichen Wang University of California, Berkeley PowerPoint Presentation

SLIDE 1

Wild Card ATLAS

Haichen Wang

University of California, Berkeley Lawrence Berkeley National Laboratory

On behalf of the ATLAS Collaboration

54th Rencontres de Moriond Electroweak Interactions and Unified Theories 17 March 2019

SLIDE 2

Measurement of ttH in Diphoton Final State using 139 fb-1 data collected by the ATLAS experiment during the LHC Run-2

Haichen Wang

University of California, Berkeley Lawrence Berkeley National Laboratory

On behalf of the ATLAS Collaboration

54th Rencontres de Moriond Electroweak Interactions and Unified Theories 17 March 2019

SLIDE 3

Overview

The associated production of the Higgs boson with top

quarks provides a direct access to the top-Higgs coupling

The diphoton channel is one of the leading channels in the

ttH measurements ○ The diphoton resonant decay is an unambiguous signature for the Higgs boson ○ Robust background estimation from diphoton mass sidebands

3

SLIDE 4

Photon performance in ATLAS

4

Photons are selected using a cut-based multivariate

discriminant based on shower shape variables in the EM calorimeter

10 20 30 40 50 60 70 80 90 100 0.5 0.6 0.7 0.8 0.9 1 1.1

ID

ε

< 40 µ Data, γ ll → Z < 40 µ MC, γ ll → Z > 40 µ Data, γ ll → Z > 40 µ MC, γ ll → Z

γ <2.37, unconverted ⏐ η ⏐ <1.37, 1.52< ⏐ η ⏐ FixedCutLoose isolation

ATLAS Preliminary

1

=13 TeV, 44 fb s 10 20 30 40 50 60 70 80 90 100 [GeV]

T

E 0.96 0.98 1 1.02 1.04 1.06

MC

ε /

Data

ε

A “tight” identification, typically used in analysis, has an efficiency

> 90% for high pT photons, and a rejection at 103 - 104

SLIDE 5

B-tagging performance in ATLAS

5

B-jets are tagged using a multivariate discriminant combining

tracking, secondary vertex and decay chain information

The b-tagging is calibrated with a ttbar control sample
In the ttH analysis, a b-tagger with a 77% efficiency is used,

corresponding to a rejection of light jet at the level of a few hundreds

30 40

2

10

2

10 × 2 [GeV]

T

Jet p 0.5 0.6 0.7 0.8 0.9 b-jet tagging efficiency

ATLAS Preliminary

1

= 13 TeV, 80.5 fb s = 70% Single Cut OP

b

∈ MV2,

Data (stat. unc.) Data (total unc.) MC t t

SLIDE 6

The diphoton sample in ATLAS

Select two energetic and well isolated photons PTγ1/mγγ > 0.4, PTγ2/mγγ > 0.35 |η| < 1.37 or 1.52 < |η| < 2.37 Quality requirement - Isolation and identification criteria ~ 1.5 million events with 105 GeV < mγγ < 160 GeV at 139 fb-1 Assume the theoretical prediction, at 139 fb-1, the LHC should have produced

~ 7,000,000 Higgs bosons
~ 70,000 via ttH production
~ 160 in the ttHγγ channel

for the ATLAS experiment

6

10000 20000 30000 Events / 1.0 GeV

Data Signal + background Continuum background Preliminary ATLAS

1 −

= 13 TeV, 79.8 fb s = 125.09 GeV

H

m Diphoton Fiducial

110 120 130 140 150 160 [GeV]

γ γ

m 500 − 500 Data - Cont. Bkg

SLIDE 7

Strategy

Use photons to tag the Higgs Boson Use jets (b-jets), leptons, and ET

miss to capture the

characteristics of top quarks

7

γ

Directly use properties of the objects in the event to train a multivariate discriminant

SLIDE 8

Multivariate Training

Training variables

4-momenta of photons, jets, leptons
Whether or not a jet is b-tagged
Missing transverse energy and its φ direction

This discriminant is trained with

Signal - Powheg Monte Carlo that models signal events
Background - data control sample where the photon

quality (isolation and/or identification) requirement is reversed ○ Mostly γγ + jets events, our main background before selection ○ See Jennet Dickinson’s talk for our understanding of the background composition

8

SLIDE 9

Training variables

Some example training variables

Photons Jets MET Leptons

The training algorithm is a Boosted Decision Tree (BDT)

9

SLIDE 10

Multivariate ttH discriminant

The BDT is trained for events with a lepton (leptonic) and events without a lepton (hadronic), separately

Events with low BDT scores are removed The remaining events are classified into multiple categories with different signal-to-background-ratios (S/Bs) based on the BDT scores, to maximize the sensitivity

10

BDT Output 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 Fraction of Events 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1

0.9 0.91 0.92 0.93 0.94 0.95 0.96 0.97 0.98 0.99 1 0.05 0.1 0.15 0.2 0.25 0.3 0.35

Cont. Bkg.

NTI Control Region H t t H Higgs t Non-t

ATLAS Preliminary

1

= 13 TeV, 139 fb s Had region BDT Output 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 Fraction of Events 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1

0.7 0.75 0.8 0.85 0.9 0.95 1 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8

Cont. Bkg.

NTI Control Region H t t H Higgs t Non-t

ATLAS Preliminary

1

= 13 TeV, 139 fb s Lep region

SLIDE 11

Performance of Categorization

Event yields in a narrow mass window around 125 GeV

The S/B goes beyond 1 in the best “hadronic” and “leptonic”

Diphoton mass distributions (hadronic)

12

110 120 130 140 150 160 [GeV]

γ γ

m 2 4 6 8 10 12 14 16 18 Events / 1.375 GeV

Data Continuum Background Total Background Signal + Background

Preliminary ATLAS

1

= 13 TeV, 139 fb s

= 125.09 GeV

H

m Had 1 category

110 120 130 140 150 160 [GeV]

γ γ

m 5 10 15 20 25 30 Events / 1.375 GeV

Data Continuum Background Total Background Signal + Background

Preliminary ATLAS

1

= 13 TeV, 139 fb s

= 125.09 GeV

H

m Had 2 category

110 120 130 140 150 160 [GeV]

γ γ

m 20 40 60 80 100 Events / 1.375 GeV

Data Continuum Background Total Background Signal + Background

Preliminary ATLAS

1

= 13 TeV, 139 fb s

= 125.09 GeV

H

m Had 4 category

110 120 130 140 150 160 [GeV]

γ γ

m 20 40 60 80 Events / 1.375 GeV

Data Continuum Background Total Background Signal + Background

Preliminary ATLAS

1

= 13 TeV, 139 fb s

= 125.09 GeV

H

m Had 3 category

SLIDE 13

Diphoton mass distributions (hadronic)

13

110 120 130 140 150 160 [GeV]

γ γ

m 2 4 6 8 10 12 14 16 18 Events / 1.375 GeV

Data Continuum Background Total Background Signal + Background

Preliminary ATLAS

1

= 13 TeV, 139 fb s

= 125.09 GeV

H

m Had 1 category

110 120 130 140 150 160 [GeV]

γ γ

m 5 10 15 20 25 30 Events / 1.375 GeV

Data Continuum Background Total Background Signal + Background

Preliminary ATLAS

1

= 13 TeV, 139 fb s

= 125.09 GeV

H

m Had 2 category

110 120 130 140 150 160 [GeV]

γ γ

m 20 40 60 80 100 Events / 1.375 GeV

Data Continuum Background Total Background Signal + Background

Preliminary ATLAS

1

= 13 TeV, 139 fb s

= 125.09 GeV

H

m Had 4 category

110 120 130 140 150 160 [GeV]

γ γ

m 20 40 60 80 Events / 1.375 GeV

Data Continuum Background Total Background Signal + Background

Preliminary ATLAS

1

= 13 TeV, 139 fb s

= 125.09 GeV

H

m Had 3 category

110 120 130 140 150 160 [GeV]

γ γ

m 2 4 6 8 10 12 14 16 18 Events / 1.375 GeV

Data Continuum Background Total Background Signal + Background

Preliminary ATLAS

1

= 13 TeV, 139 fb s

= 125.09 GeV

H

m Had 1 category

SLIDE 14

Diphoton mass distributions (hadronic)

14

110 120 130 140 150 160 [GeV]

γ γ

m 2 4 6 8 10 12 14 16 18 Events / 1.375 GeV

Data Continuum Background Total Background Signal + Background

Preliminary ATLAS

1

= 13 TeV, 139 fb s

= 125.09 GeV

H

m Had 1 category

110 120 130 140 150 160 [GeV]

γ γ

m 5 10 15 20 25 30 Events / 1.375 GeV

Data Continuum Background Total Background Signal + Background

Preliminary ATLAS

1

= 13 TeV, 139 fb s

= 125.09 GeV

H

m Had 2 category

110 120 130 140 150 160 [GeV]

γ γ

m 20 40 60 80 100 Events / 1.375 GeV

Data Continuum Background Total Background Signal + Background

Preliminary ATLAS

1

= 13 TeV, 139 fb s

= 125.09 GeV

H

m Had 4 category

110 120 130 140 150 160 [GeV]

γ γ

m 20 40 60 80 Events / 1.375 GeV

Data Continuum Background Total Background Signal + Background

Preliminary ATLAS

1

= 13 TeV, 139 fb s

= 125.09 GeV

H

m Had 3 category

110 120 130 140 150 160 [GeV]

γ γ

m 2 4 6 8 10 12 14 16 18 Events / 1.375 GeV

Data Continuum Background Total Background Signal + Background

Preliminary ATLAS

1

= 13 TeV, 139 fb s

= 125.09 GeV

H

m Had 1 category

115 120 125 130 135 140 [GeV]

γ γ

m 0.02 0.04 0.06 0.08 0.1 0.12 0.14 0.16 / 0.5 GeV

γ γ

m 1/N dN/d Simulation Preliminary ATLAS

1

= 13 TeV, 139 fb s = 125 GeV

H

, m γ γ → H

Had 1 MC Signal Model Lep 3 MC Signal Model

SLIDE 15

Diphoton mass distributions (leptonic)

15

110 120 130 140 150 160 [GeV]

γ γ

m 2 4 6 8 10 12 Events / 1.375 GeV

Data Continuum Background Total Background Signal + Background

Preliminary ATLAS

1

= 13 TeV, 139 fb s

= 125.09 GeV

H

m Lep 2 category

110 120 130 140 150 160 [GeV]

γ γ

m 2 4 6 8 10 12 Events / 1.375 GeV

Data Continuum Background Total Background Signal + Background

Preliminary ATLAS

1

= 13 TeV, 139 fb s

= 125.09 GeV

H

m Lep 3 category

110 120 130 140 150 160 [GeV]

γ γ

m 2 4 6 8 10 12 14 Events / 1.375 GeV

Data Continuum Background Total Background Signal + Background

Preliminary ATLAS

1

= 13 TeV, 139 fb s

= 125.09 GeV

H

m Lep 1 category

SLIDE 16

Weighted diphoton mass distribution

16

A nice way to visualize the power of categorization is to draw the S/B weighted mass distribution

110 120 130 140 150 160 [GeV]

γ γ

m 5 10 15 20 25 30 Sum of Weights / 1.375 GeV

Data Continuum Background Total Background Signal + Background

Preliminary ATLAS

1

= 13 TeV, 139 fb s

= 125.09 GeV

H

m All categories ln(1+S/B) weighted sum

SLIDE 17

Summary of expected and observed event yields

17

20 40 60 80 100 120 140 160 180

Events

Data =1.4) µ H ( t t H Higgs t Non-t

Cont. Bkg.

Had categories Lep categories

Preliminary ATLAS

1

=13 TeV, 139 fb s

Had 4 Had 3 Had 2 Had 1 Lep 3 Lep 2 Lep 1

20

Data - Bkg.

=1.4) µ H ( t t

All numbers calculated in a mass window containing 90% of the ttH signal events

SLIDE 18

Observed significance and rate

18

110 120 130 140 150 160 [GeV]

γ γ

m 5 10 15 20 25 30 Sum of Weights / 1.375 GeV

Data Continuum Background Total Background Signal + Background

Preliminary ATLAS

1

= 13 TeV, 139 fb s

= 125.09 GeV

H

m All categories ln(1+S/B) weighted sum

The measured ttH cross section times H→γγ branching ratio is while the SM expectation is

The expected significance of the ttH process is 4.2 σ, the

bserved is 4.9 σ

SLIDE 19

Observed significance and rate

19

The expected significance of the ttH process is 4.2 σ, the

bserved is 4.9 σ

The measured ttH cross section times H→γγ branching ratio is The signal strength (obs/SM) is measured to be

110 120 130 140 150 160 [GeV]

γ γ

m 5 10 15 20 25 30 Sum of Weights / 1.375 GeV

Data Continuum Background Total Background Signal + Background

Preliminary ATLAS

1

= 13 TeV, 139 fb s

= 125.09 GeV

H

m All categories ln(1+S/B) weighted sum

SLIDE 20

Breakdown of systematic uncertainties

20

SLIDE 21

Summary

21

ATLAS measurement of the ttH → γγ process from the full

LHC Run-2 data set was reported

The ttH process is observed in the diphoton decay

mode with a significance of 4.9 σ

The ttH cross section times H → γγ branching ratio is

measured to be

Details are available in ATLAS-CONF-2019-004

SLIDE 22

Back up

22

SLIDE 23

Unweighted diphoton mass distribution

23

110 120 130 140 150 160 [GeV]

γ γ

m 20 40 60 80 100 120 140 Events / 1.375 GeV

Data Continuum Background Total Background Signal + Background

Preliminary ATLAS

1

= 13 TeV, 139 fb s

= 125.09 GeV

H

m All categories

SLIDE 24

Weighted diphoton mass distribution

24

110 120 130 140 150 160 [GeV]

γ γ

m 5 10 15 20 25 30 Sum of Weights / 1.375 GeV

Data Continuum Background Total Background Signal + Background

Preliminary ATLAS

1

= 13 TeV, 139 fb s

= 125.09 GeV

H

m All categories ln(1+S/B) weighted sum

SLIDE 25

Unweighted diphoton mass distribution

25