Electron charge identification and study of the electroweak WW jj … · 2017-09-12 · Data20.3fb1 LHC pp p s= 13TeV Data0.08 36.1fb1 Standard Model Production Cross Section Measurements

Electron charge identification and study of the electroweakW±W±jj production with the ATLAS detector at 13 TeV

Giulia GonellaAlbert-Ludwigs-Universitat Freiburg

The Final HiggsTools Meeting

Durham, UK - September 12th, 2017

Overview

Vector boson scattering

The role in the EWK simmetry breaking investigation

Features

The WW channel

Background composition

Electron charge mis-identification: efficiencies measurement

Analysis overview

Electron charge mis-identification: background in the W±W±jj analysis

Summary

Giulia Gonella - W±W± jj production 12.09.2017 1 / 24

Why keeping working after the Higgs boson discovery?

• Is this the only Higgs boson in Nature?

• Is it doing the Higgs boson’s job?

Different paths to follow to answer (one or more of) these questions


Why keeping working after the Higgs boson discovery?

• Is this the only Higgs boson in Nature?

• Is it doing the Higgs boson’s job?

Different paths to follow to answer (one or more of) these questions


Many tools for the higgs investigation

• Measuring with higher precision its properties and couplings

• Investigating the high mass region

• Measuring the electroweak (EWK) vector boson interactions

The Standard Model (SM)

lagrangian predicts triple (TGC)

and quartic (QGC) gauge bosons

vertices

Lgauge = −1

4BµνBµν −

1

4W aµνWµν

a

= LGC + LTGC + LQGC

! Pure EWK verteces are suppressed by αEWK → rare processes

Ultimately: aiming to study quartic couplings


Many tools for the higgs investigation

• Measuring with higher precision its properties and couplings

• Investigating the high mass region

• Measuring the electroweak (EWK) vector boson interactions

The Standard Model (SM)

lagrangian predicts triple (TGC)

and quartic (QGC) gauge bosons

vertices

Lgauge = −1

4BµνBµν −

1

4W aµνWµν

a

= LGC + LTGC + LQGC

! Pure EWK verteces are suppressed by αEWK → rare processes

Ultimately: aiming to study quartic couplings


Vector boson scattering

The role of Vector Bosons Scattering

• CERN Large Hadron Collider is providing promising statistics to access these

processes

• quartic couplings can be accessed in Vector Boson Scattering signatures

At hadron colliders VBS can be idealized as VVjj

at leading order (LO):

• two vector bosons (i.e. their respective decay products)

• two outgoing jets

BUT

VBS diagrams are not separately gauge invariant and must be studied in conjunction

with additional Feynman diagrams leading to the same VVjj final state.


VVjj final state diagrams

Theoretically there are two classes of physical processes.

• Electroweak production: Only Weak interaction

• O(α6EWK )

• VBS signal in it

• It contains also:

purely EWK process which give the same final state

processes with 3 decaying vector boson (only 1 decaying hadronically)


VVjj final state diagrams

Theoretically there are two classes of physical processes.

• Electroweak production: Only Weak interaction

• Strong production: Both strong and EWK interaction

• O(α4EWKα

2S )


The WW channel

WW scattering

Without a light SM Higgs boson the VBS amplitude of longitudinally polarized W

bosons increases with√

s and violates unitarity at energies around 1 TeV.

↓

The SM Higgs boson should avoid this problem

WW scattering is a key process to probe EWKSBWe can establish if the Higgs boson can preserve unitarity of the VBS at all energies

• Test of the Higgs boson nature

• The discovered Higgs boson

contribute fully to the EWKSB

↓

WW interaction remain weak at high

energies

• Model independent research ofalternative theory

• The discovered Higgs boson is

partially responsible for the EWKSB

↓

WW interaction get strong at high

energy

arXiv:1412.8367


https://arxiv.org/abs/1412.8367

Choosing the right process: W±W±jj channel

C. Gumpert, PhD Thesis, CERN-THESIS-2014-290

W±W±jj production:

• lower background from QCD production: σEWK/σQCD '11 ( σEWK/σQCD '0.6

for opposite sign)

The same-charge selection restricts the number of diagrams involved


https://cds.cern.ch/record/2003240/files/CERN-THESIS-2014-290-rev.pdf

Choosing the right process: W±W±jj channel

C. Gumpert, PhD Thesis, CERN-THESIS-2014-290

W±W±jj production:

• lower background from QCD production: σEWK/σQCD '11 ( σEWK/σQCD '0.6

for opposite sign)

The same-charge selection restricts the number of diagrams involved


https://cds.cern.ch/record/2003240/files/CERN-THESIS-2014-290-rev.pdf



SM processes can mimic the signature ``′ + E missT +2 jets

Type Sources Reduction

1 SC leptons

WZ+jets

veto on third leptonZZ+jets

ttV

2 OC charge mis-ID

tt → `ν`νbb b-jet veto +SC req.

W±W∓ + jets

Z/γ∗+jets→ `±`∓+jets large E missT +mZ peak excl.

3 Jets, γ mis-reco

W +jets

tt → `νjjbbb-jet veto+E miss

Tsingle top

Wγ+jets tight isolation+veto on third lepton


It’s important to keep in mind that the other SM processes have muchlarger cross section compared to our signal!

pp

total (x2)

inelastic

JetsR=0.4

dijets

incl .

γ

fid.

pT > 125 GeV

pT > 25 GeV

nj ≥ 1

nj ≥ 2

nj ≥ 3

pT > 100 GeV

W

fid.

nj ≥ 0

nj ≥ 1

nj ≥ 2

nj ≥ 3

nj ≥ 4

nj ≥ 5

nj ≥ 6

nj ≥ 7

Z

fid.

nj ≥ 1

nj ≥ 2

nj ≥ 3

nj ≥ 4

nj ≥ 5

nj ≥ 6

nj ≥ 7

nj ≥ 0

nj ≥ 1

nj ≥ 2

nj ≥ 3

nj ≥ 4

nj ≥ 5

nj ≥ 6

nj ≥ 7

ttfid.

total

nj ≥ 4

nj ≥ 5

nj ≥ 6

nj ≥ 7

nj ≥ 8

t

tot.

Zt

s-chan

t-chan

Wt

VVtot.

ZZ

WZ

WW

ZZ

WZ

WW

ZZ

WZ

WW

γγ

fid.

H

fid.

H→γγ

VBFH→WW

ggFH→WW

H→ZZ→4ℓ

H→ττ

total

WV

fid.

Vγ

fid.

Zγ

W γ

ttW

tot.

ttZ

tot.

ttγ

fid.

WjjEWK

fid.

ZjjEWK

fid.

WWExcl.

tot.

Zγγ

fid.

Wγγ

fid.

WWγ

fid.

ZγjjEWKfid.

VVjjEWKfid.

W ±W ±

WZ

σ[p

b]

10−3

10−2

10−1

1

101

102

103

104

105

106

1011 Theory

LHC pp√s = 7 TeV

Data 4.5 − 4.9 fb−1

LHC pp√s = 8 TeV

Data 20.3 fb−1

LHC pp√s = 13 TeV

Data 0.08 − 36.1 fb−1

Standard Model Production Cross Section Measurements Status: July 2017

ATLAS Preliminary

Run 1,2√s = 7, 8, 13 TeV


The challenges

Many challenges enter this analysis:

• very low cross-section compared to other SM processes

• many SM backgrounds can be reduced with dedicated cuts, but still some of

them need to be estimated

• big impact of background from non-prompt leptons (fakes)

• big impact of background from charge mis-reconstruction

↓

always better to estimate these backgrounds with data-driven methods (“Data model

data better than Monte Carlo”)

In the following focus on charge mis-identification background

In order to estimate the charge mis-ID background it’s important to precisely know

the probability of an electron to have its charge wrongly reconstructed:

• measurement of electron charge mis-identification efficiencies

• data-driven estimation of background from charge mis-identification


Electron charge mis-identification:efficiencies measurement

The source of charge mis-ID

Charge (mis-)reconstruction

• Conversion of bremsstrahlung photon

e-primary electron

conversion electron

γbrem

e+

e-primary electron

reconstructed electronwith wrong charge

γbrem

e+

• Wrong track reconstruction

primary electron


• very rare effects

• very hard to properly model in detector simulation

• the effect have to be measured on data

In ATLAS objects measurements are performed inside Combined Performance groups. This one in particular has

been carried out in the Electron-photon CP group





e-primary electron

conversion electron

γbrem

e+

e-primary electron


γbrem

e+


primary electron











e-primary electron

conversion electron

γbrem

e+

e-primary electron


γbrem

e+


primary electron








Measuring the efficiencies

GeVllM60 70 80 90 100 110 120

even

ts/G

eV

0

500

1000

1500

2000

2500

3000

310×

OS events

SS events

ATLAS Work in progress

• Double differential measurement in

4×6 bins in pT × |η|

• measured in Z → e+e− events

• εi probability electron charge mis-reconstructed in

bin i . Assume εi independent

Nexpsc = np = n

[(1− εi ) εj +

(1− εj

)εi

]Building likelihood function:

• binomial counting: probability of nsc SC events:( nnsc

)pnsc (1− p)n−nsc

• approximated Poisson distribution

P(nsc|εi , εj

)=

(Nexpsc )nsc e−Nexp

sc

nsc!≡ L(εi , εj )

• L =∏

i,j L(εi , εj )

Efficiencies obtained by minimizing −ln(L)

|η|0 0.5 1 1.5 2 2.5

mis

-ID

∈

3−10

2−10

1−10

[25,60] GeV∈ T

p

[60,90] GeV∈ T

p

[90,130] GeV∈ T

p

[130,1000] GeV∈ T

p

ATLAS Internal, 2016 data-133.9 fb

Tp

40 60 80 100 120 140

mis

-ID

∈3−10

2−10

1−10


[0.0,0.6]∈| η| [0.6,1.1]∈| η| [1.1,1.52]∈| η| [1.52, 1.7]∈| η| [1.7, 2.3]∈| η| [2.3, 2.5]∈| η|

arxiv:1612.01456Giulia Gonella - W±W± jj production 12.09.2017 13 / 24


Measuring the efficiencies

GeVllM60 70 80 90 100 110 120

even

ts/G

eV

0

500

1000

1500

2000

2500

3000

310×

OS events

SS events

ATLAS Work in progress

• Double differential measurement in

4×6 bins in pT × |η|

• measured in Z → e+e− events

• εi probability electron charge mis-reconstructed in

bin i . Assume εi independent

Nexpsc = np = n

[(1− εi ) εj +

(1− εj

)εi

]Building likelihood function:

• binomial counting: probability of nsc SC events:( nnsc

)pnsc (1− p)n−nsc

• approximated Poisson distribution

P(nsc|εi , εj

)=

(Nexpsc )nsc e−Nexp

sc

nsc!≡ L(εi , εj )

• L =∏

i,j L(εi , εj )

Efficiencies obtained by minimizing −ln(L)

|η|0 0.5 1 1.5 2 2.5

mis

-ID

∈

3−10

2−10

1−10

[25,60] GeV∈ T

p

[60,90] GeV∈ T

p

[90,130] GeV∈ T

p

[130,1000] GeV∈ T

p


|η|0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5

] 0R

adia

tion

leng

th [X

0

0.5

1

1.5

2

2.5 ServicesTRTSCTPixelBeam-pipe

ATLASSimulation

arxiv:1612.01456Giulia Gonella - W±W± jj production 12.09.2017 13 / 24


Scale factors

Tp

40 60 80 100 120 140

rate

s

3−10

2−10

1−10

datamcATLAS Internal

[1.52, 1.7]∈| η|

Tp

40 60 80 100 120 140

SF

SS

0.6

0.8

1

1.2

1.4

|η|0 0.5 1 1.5 2 2.5

rate

s

3−10

2−10

1−10 datamcATLAS Internal

[90, 130] GeV∈ T

p

|η|0 0.5 1 1.5 2 2.5

SF

SS

0.6

0.8

1

1.2

1.4

• efficiencies highly dependent on process and kinematic

selection

• scale factors are produce where these effects cancel

• they can also be used to study the effeiciencies of other

processes

• scale factors are derived for right charge and wrong

charge electrons to correct MC

SFwrong =εdata

εMCSFright =

1− εdata

1− εMC

0.8506 1.0662 0.8508 1.0948

0.7524 0.6884 0.7634 1.1763

0.6716 0.7525 0.9382 0.9647

0.6803 0.6679 0.7773 0.9437

0.9007 1.0002 0.9707 1.0326

0.9095 1.0091 1.1332 0.9740

GeVT

p40 60 80 100 120 140

|η|

0

0.5

1

1.5

2

2.5

0.7

0.8

0.9

1

1.1

1.0000 1.0000 1.0003 0.9997

1.0002 1.0007 1.0010 0.9989

1.0007 1.0012 1.0006 1.0006

1.0021 1.0047 1.0052 1.0023

1.0010 1.0000 1.0011 0.9979

1.0024 0.9995 0.9894 1.0030

GeVT

p40 60 80 100 120 140

|η|

0

0.5

1

1.5

2

2.5

0.7

0.8

0.9

1

1.1

• corrections for wrong charge electrons up to ∼30%

• corrections for right charge electrons never bigger than 0.5%Giulia Gonella - W±W± jj production 12.09.2017 14 / 24

Analysis overview

Prelude: W±W±jj analysis topology

• 2 high-pT jets in the forward regions

(2 jets with highest pT )

• No color exchange in the hard

scattering process → rapidity gap in

the central part of the detector

• large mjj and mWWjj

jj y∆

0 1 2 3 4 5 6 7 8 9

arbi

trar

y un

its

0

0.01

0.02

0.03

0.04

0.05

0.06

0.07

0.08

0.09

0.1jj, electroweak±W± W

jj, strong±W± WATLAS Internal

Plot: "CutNJets/DYjj"

[GeV]jjm

0 100 200 300 400 500 600 700 800 900 1000

arbi

trar

y un

its

3−10

2−10

jj, electroweak±W± W


Plot: "CutNJets/Mjj"

Two analysis regions could be defined, using the EWK specific topologies as

discriminant cuts:

• Inclusive region: signal = EWK+QCD

• VBS region: signal = EWK → additional cut on |∆yjj | > 2.4Giulia Gonella - W±W± jj production 12.09.2017 15 / 24

Prelude: W±W±jj analysis topology

• 2 high-pT jets in the forward regions

(2 jets with highest pT )

• No color exchange in the hard

scattering process → rapidity gap in

the central part of the detector

• large mjj and mWWjj

jj y∆

0 1 2 3 4 5 6 7 8 9

arbi

trar

y un

its

0

0.01

0.02

0.03

0.04

0.05

0.06

0.07

0.08

0.09

0.1jj, electroweak±W± W


Plot: "CutNJets/DYjj"

[GeV]jjm

0 100 200 300 400 500 600 700 800 900 1000

arbi

trar

y un

its

3−10

2−10

jj, electroweak±W± W


Plot: "CutNJets/Mjj"

Two analysis regions could be defined, using the EWK specific topologies as

discriminant cuts:

• Inclusive region: signal = EWK+QCD

• VBS region: signal = EWK → additional cut on |∆yjj | > 2.4Giulia Gonella - W±W± jj production 12.09.2017 15 / 24

Studies of EWK-QCD separation and interference

Interference between O(α4

QCD

)strong and O

(α6

EWK

)electroweak W±W±jj

production to be studied:

∣∣∣MWWjjSM

∣∣∣2 =∣∣∣MWWjj

QCD +MWWjjEWK

∣∣∣2 =∣∣∣MWWjj

QCD

∣∣∣2 +∣∣∣MWWjj

EWK

∣∣∣2 +∣∣∣MWWjj

INT

∣∣∣2

• interference varies from few percent to few ten-percent

• negative in some regions

• will be included as systematic uncertainty on the EWK

signal


Contributions at final selection

• analysis performed in four channels: eµ, µe, µµ and ee

• different background composition between channels

• main backgrounds: WZ , fakes and charge mis-ID (not µµ channel)

As anticipated the main challenges in this analysis are related to the fake background

and the charge mis-identification one


Electron charge mis-identification:background in the W±W±jj analysis

Charge mis-ID background

The technique

• charge mis-ID background estimated from data

• same-charge events are estimated from opposite-charge events

• opposite charge data are scaled with

w =ε1 + ε2 − 2ε1ε2

1− (ε1 + ε2 − 2ε1ε2)

[GeV]l0T

p20 40 60 80 100 120 140 160 180 200 220

effic

ienc

y

0.005

0.01

0.015

0.02

0.025

0.03

0.035

0.04bkg/ee/Zjetsbkg/ee/top/Nom/ttbar

ATLAS Work in progress-1= 13 TeV, 36.1 fbs

• ε1,2: charge mis-ID rates for e1,2

• rates appear to be process

dependent


The data-driven estimate

The final aim

We want a data-driven estimation of the charge mis-identification background

remaining in the SR

opposite-charge data

same-charge estimation

where

ε1,2: charge misID rates. P for `1,2 charge to be mis-identified

→ rates from data are needed to build the weights


From SF to rates

0.8506 1.0662 0.8508 1.0948

0.7524 0.6884 0.7634 1.1763

0.6716 0.7525 0.9382 0.9647

0.6803 0.6679 0.7773 0.9437

0.9007 1.0002 0.9707 1.0326

0.9095 1.0091 1.1332 0.9740

GeVT

p40 60 80 100 120 140

|η|

0

0.5

1

1.5

2

2.5

0.7

0.8

0.9

1

1.1

1.0000 1.0000 1.0003 0.9997

1.0002 1.0007 1.0010 0.9989

1.0007 1.0012 1.0006 1.0006

1.0021 1.0047 1.0052 1.0023

1.0010 1.0000 1.0011 0.9979

1.0024 0.9995 0.9894 1.0030

GeVT

p40 60 80 100 120 140

|η|

0

0.5

1

1.5

2

2.5

0.7

0.8

0.9

1

1.1

↓

• apply the SF to MC so that it matches the probability of mis-identification of data

• get rates from MC using the truth information

ε =wrong charge ele

all ele

0.00023 0.00018 0.00021 0.00016 0.00022 0.00026 0.00057 0.00166

0.00036 0.00037 0.00027 0.00033 0.00016 0.00031 0.00083 0.00214

0.00057 0.00086 0.00050 0.00052 0.00051 0.00050 0.00148 0.00287

0.00138 0.00121 0.00135 0.00098 0.00083 0.00133 0.00266 0.00565

0.00214 0.00221 0.00212 0.00148 0.00141 0.00196 0.00377 0.01195

0.00396 0.00268 0.00273 0.00100 0.00277 0.00527 0.00511 0.01435

0.00434 0.00597 0.00594 0.00423 0.00334 0.00610 0.00932 0.01826

0.00987 0.00984 0.01071 0.00859 0.00738 0.01039 0.01937 0.03169

0.00777 0.00666 0.00830 0.00711 0.00457 0.00930 0.01841 0.03699

0.00847 0.00758 0.00797 0.00781 0.00727 0.00888 0.02190 0.03853

0.00711 0.00792 0.00806 0.00940 0.00741 0.01330 0.02139 0.03957

0.01196 0.00948 0.01290 0.01050 0.01097 0.01368 0.02636 0.044380.01029 0.01171 0.01532 0.01407 0.01311 0.02164 0.03569 0.053570.02454 0.02373 0.02548 0.02640 0.02288 0.03324 0.05156 0.094270.01930 0.02662 0.02645 0.03144 0.03297 0.03627 0.05422 0.088030.01754 0.02799 0.02593 0.03324 0.03276 0.04560 0.05427 0.09273

GeVl0T

p30 40 50 60 70 80 90

l0 |η|

0

0.5

1

1.5

2

2.5

0.01

0.02

0.03

0.04

0.05

0.06

0.07

0.08

0.09


Energy corrections needed

Up to this point of the workflow a simple scaling is applied to OC data:

• the integral is modified

• the distribution of charge flip estimation is identical to OC data one

• SC events have lower energy wrt OC one, because of the energy leakage due to

Bremsstrahlung emission

arbi

trar

y un

its

0.01

0.02

0.03

0.04

0.05

0.06

0.07

0.08

0.09

0.1 data, same-charge

data, opposite-chargeATLAS Work in progress-1 = 13 TeV, 36.1 fbs

[GeV]llm

75 80 85 90 95 100 105

ratio

00.5

11.5

2








arbi

trar

y un

its

0.01

0.02

0.03

0.04

0.05

0.06

0.07

0.08

0.09


DY MC, opposite-charge

DY MC, same-charge

data, opposite-charge

ATLAS Work in progress-1 = 13 TeV, 36.1 fbs

[GeV]llm

75 80 85 90 95 100 105

ratio

00.5

11.5

2








arbi

trar

y un

its

0.01

0.02

0.03

0.04

0.05

0.06

0.07

0.08

0.09


charge flip estimate

DY MC, opposite-charge

DY MC, same-charge

data, opposite-charge

ATLAS Work in progress-1 = 13 TeV, 36.1 fbs

[GeV]llm

75 80 85 90 95 100 105

ratio

00.5

11.5

2


The scaling and the smearing

The four-momentum of OC data used for the estimation is modified

• pT is shifted towards smaller values

• pT is smeared to match the worse resolution

pcorrectedT = pscaled

T + dE

The scaling is based on the energy response

response =preco

T

ptruthT

− 1

which is different for correctly reconstructed and wrongly reconstructed electrons

- 1l0,truth

T/pl0,reco

Tp

0.8− 0.6− 0.4− 0.2− 0 0.2 0.4 0.6 0.8

arbi

trar

y un

its

0

0.01

0.02

0.03

0.04

0.05

0.06

0.07

0.08 stat. unc. Drell-YanATLAS Internal

-1 = 13 TeV, 36.1 fbs

channelν±eν±e

- 1l0,truth

T/pl0,reco

Tp

0.8− 0.6− 0.4− 0.2− 0 0.2 0.4 0.6 0.8

l0 η

2.5−

2−

1.5−

1−

0.5−

0

0.5

1

1.5

2

2.5

0

500

1000

1500

2000

2500

3000

3500

4000ATLAS Internal

-1 = 13 TeV, 36.1 fbs

channelν±eν±e


The scaling and the smearing

The four-momentum of OC data used for the estimation is modified

• pT is shifted towards smaller values

• pT is smeared to match the worse resolution

pcorrectedT = pscaled

T + dE

The scaling is based on the energy response

response =preco

T

ptruthT

− 1

which is different for correctly reconstructed and wrongly reconstructed electrons

- 1l0,truth

T/pl0,reco

Tp

0.8− 0.6− 0.4− 0.2− 0 0.2 0.4 0.6 0.8

arbi

trar

y un

its

0

0.01

0.02

0.03

0.04

0.05

0.06 stat. unc. Drell-YanATLAS Internal

-1 = 13 TeV, 36.1 fbs

channelν±eν±e

- 1l0,truth

T/pl0,reco

Tp

0.8− 0.6− 0.4− 0.2− 0 0.2 0.4 0.6 0.8

l0 η

2.5−

2−

1.5−

1−

0.5−

0

0.5

1

1.5

2

2.5

0

5

10

15

20

25

30

35

40ATLAS Internal-1 = 13 TeV, 36.1 fbs

channelν±eν±e


The estimation in the ee channel

Eve

nts

/ 1 G

eV

0

200

400

600

800

1000

1200

1400 Data stat. unc.±W± W Top

VV + VVV γ W+

γ Z+ charge-flip

fakes

ATLAS Internal-1 = 13 TeV, 36.1 fbs

channelν±eν±e

[GeV]llm

70 75 80 85 90 95 100 105 110

Dat

a / M

C

00.20.40.60.8

11.21.41.61.8

2E

vent

s / 5

GeV

0

500

1000

1500

2000

2500

3000 Data stat. unc.

±W± W Top VV + VVV γ W+

γ Z+ charge-flip

fakes


channelν±eν±e

[GeV]l0

Tp

20 40 60 80 100 120 140 160 180 200 220

Dat

a / M

C

00.20.40.60.8

11.21.41.61.8

2

Eve

nts

/ 0.2

0

200

400

600

800

1000

1200 Data stat. unc.±W± W Top

VV + VVV γ W+

γ Z+ charge-flip

fakes


channelν±eν±e

l0η

2− 1− 0 1 2

Dat

a / M

C

00.20.40.60.8

11.21.41.61.8

2

• the ee channel is dominated by charge mis-ID background

• the estimation seems to match quite nicely with data

• studies and optimizations ongoing


Summary

Summary

Two main working areas have been shown...

→ electron charge mis-ID SF measurement

• performed within the ATLAS e/γ combined performance group

• official recommendations provided to physics analyses inside the collaboration

• the work will be documented in the e/γ paper

→ estimation of the charge mis-ID background in the W±W±jj analysis

• process of estrapolation of the rates for the analysis has been shown

• the impact of this background is huge in the SR, much work needed to get a

precise estimation

• the technique for this has been studied and implemented inside the analysis

framework

• further corrections have been performed

WW scattering is a key process for SM and EWSB mechanism

• same charge final state is a clean signature of electroweak production

• on the other side the low cross-section and the large background from charge

mis-identification make the W±W±jj analysis a difficult path

Very challenging time yet to come to put Standard Model to stringent test


Thank you!

Backup

Phase space for fiducial cross sections (slide 8)

• Sherpa

• defined at parton level

• plepT ≥ 25 GeV

• |ηlep | ≤ 2.5

• ∆R lep ≥ 0.3

• minimum m`` of all charged lepton pair combinations ≥ 20 GeV

• at least 2 jets pjetT ≥30 GeV

• |ηjet | ≤ 4.5

• ∆R lep,jet ≥ 0.3

• mjj ≥ 500 GeV

• |∆η| ≥ 2.4

ATLAS Run 1 results in the W±W±jj channel

First evidence for W±W±jj production and EWK-only W±W±jj production @ 8 TeV

σfidincl = 2.1 ± 0.5 (stat) ± 0.3 (syst) fb

• Significance: 4.5σ

• σexp = 1.52 ± 0.11 fb

σfidVBS = 1.3 ± 0.4 (stat) ± 0.2 (syst) fb

• Significance: 3.6σ

• σexp = 0.95 ± 0.06 fb

Goal for Run 2

Measurement of the EWK production cross section → 5σ observation reachable within

LHC operation plans

Yields Run 1 analysis

Inclusive region VBS region

e±e± e±µ± µ±µ± e±e± e±µ± µ±µ±

W±W±jj QCD 0.89 ± 0.15 2.5 ± 0.4 1.42 ± 0.23 0.25 ± 0.06 0.71 ± 0.14 0.38 ± 0.08

W±W±jj EWK 3.07 ± 0.30 9.0 ± 0.8 4.9 ± 0.5 2.55 ± 0.25 7.3 ± 0.6 4.0 ± 0.4

Total background 6.8 ±1.2 10.3 ± 2.0 3.0 ± 0.6 5.0 ± 0.9 8.3 ± 1.6 2.6 ± 0.5

Total predicted 10.7 ± 1.4 21.7 ± 2.6 9.3 ± 1.0 7.6 ± 1.0 15.6 ± 2.0 6.6 ± 0.8

Data 12 26 12 6 18 10

50 signal candidates with 20 background expectation for ISR

34 signal candidates with 16 background expectation for VBS

Fiducial regions Run 1 analysis

Two fiducial regions are defined to follow the selections applied in the analysis

• Inclusive region:• 2 same sign leptons with

• pT > 25 GeV

• |η| < 2.5

• m`` > 20 GeV

• ∆R`` =√

(∆Φ)2 + (∆η)2 > 0.3

• At least 2 anti-kt jets with

• R = 0.4

• pT > 30 GeV

• |η| < 4.5

• ∆R`j > 0.3

• mjj (highest pT jets) > 500 GeV

• E missT > 40 GeV

• VBS region: cuts above + rapidity cut

|∆yjj | > 2.4

Run 1 event selection

Events selection used for the analysis:

• event cleaning

• exactly two selected leptons with m`` > 20 GeV

• veto events with additional veto leptons

• q`1× q`2

> 0

• pT > 25 GeV

• |m`` −mZ | > 10 GeV in the ee channel

• E missT ≥ 40 GeV

• at least two jets with pT > 30 GeV and |η| < 4.5

• b-jet veto

• mjj > 500 GeV

• |∆yjj | > 2.4 - VBS analysis region

Cross section changes from 8 TeV to 13 TeV

Signal samples generated with Sherpa 2.1.1, LO up to 3 additional partons:

QCD EWK

8 TeV 13 TeV 13 TeV/8 TeV 8 TeV 13 TeV 13 TeV/8 TeV

σprod [fb] 10.1 25.9 2.6 16.4 43.0 2.6

Rates dependency on material

• Rate of wrongly reconstructed charge strongly depends on material distribution:

Z and R for the first detector interaction using Z MC truth:

wrong sign interact primarily with beam pipe and pixel

Current selection

• p`1,`2T >27 GeV → under optimization studies

• m`` > 20 GeV → under optimization studies

• q`1× q`2

> 0

• 3rd lepton veto

• pjT > 30 GeV → under optimization studies

• |η|j < 4.5

• E missT ≥ 30 GeV

• mjj > 500 GeV → under optimization studies

• |∆yjj | > 2.4 → under optimization studies

Energy correction derivation

• response is derived

response =preco

T

ptruthT

− 1

for right charge and wrong charge electrons, differentially in η

• defining α as

α =1 + responseright

1 + responsewrong

• pT is corrected as:

pscaledT =

poldT

α

• smearing is based on dE , random number in Gauss distribution:

dE = Gauss

(0 ,sqrt

(err2

wrong

(1 + responsewrong )2−

err2right

(1 + responseright )2

))pscaled

T

Acceptances and efficiencies for W±W±jj

• Calculated on Sherpa 2.2.1

• σtotEWK = 37.4fb and σtot

QCD = 23.5fb

• acceptance on fiducial selection is a few percent

→ σfidEWK = 1.97fb and σtot

QCD = 0.30fb

CMS analysis - Selection and yields

ATLAS CMS

dataset 2015+2016 2016

Ldx 36.1 fb−1 35.9 fb−1

p`T 26 GeV 25/20 GeV

|η| 2.5 2.4/2.5

pjT /E j

T 25/30 GeV 30 GeV

m`` 20 GeV 20 GeV

E missT 30 GeV 40 GeV

Z -veto (ee) 15 GeV 15 GeV

mjj 500 GeV 500 GeV

∆yjj/∆ηjj 2.4 2.5

max(z∗` ) - 0.75

veto 3` < 10 GeV <10 GeV

veto τhad - >18 GeV

veto b-jet 3 3

With respect to ATLAS analysis:

• less WZ background

• higher non-prompt

Zeppenfeld variable z∗ =

∣∣∣η`−(ηj1|ηj2

)/2∣∣∣∣∣∣∆ηjj

∣∣∣

CMS analysis - Some details

General information

• charge flip rate: 0.01% in the barrel, 0.3% in the endcap (ATLAS: 0.07% barrel,

3.5% endcap)

• charge flip contribution estimated from MC with data scale factors

• fake-factor method yields 30% uncertainty

• third lepton CR (with Z mass window) yields 20%-40% uncertainty

• significance extracted from 2-dim fit in m`` and mjj

Signal:

• purely EWK6, contributions from EWK4 are subtracted

• theoretical uncertainties 12% (αs ), 5% (PDF) and 4.5% (EWK6-EWK4

interference)

• LO cross-section from Madgraph (also signal sample): σtheofid = 4.25± 0.21 fb−1

Cross section result:

σfid (W±W±jj) = 3.83± 0.66(stat.)± 0.35(syst.) fb

CMS analysis - BSM interpretation

Anomalous Quartic Couplings

• EFT lagrangian of nine C- and

P-conserving dim-8 operators

modifying quartic couplings

Doubly charged Higgs boson

• Doubly charged Higgs bosons

predicted by models with Higgs triplet

field

• couplings depend on m(H±) and

sin2 θH

• Georgi-Machacek model of Higgs

triplet considered

• limits presented for

σVBF (H±±)× B(H±± →W±W±)

and sH ≡ sin θH

Scale factor systematics

Four sources:

• m`` variation (nominal: 15GeV)

• trigger matchig requirement on sub-leading candidate

• background subtraction on/off

• truth matching comparison

These are combined and propagated in the derivation of the data-driven background

→ variation up/down (7-15%)

→ energy correction on/off (2-7%)

Documents

Electron charge identification and study of the electroweak WW jj … · 2017-09-12 · Data20.3fb1 LHC pp p s= 13TeV Data0.08 36.1fb1 Standard Model Production Cross Section Measurements