Observation of light-by-light scattering in lead-lead collisions in - - PowerPoint PPT Presentation

SLIDE 1

Observation of light-by-light scattering 
 in lead-lead collisions in the ATLAS experiment

Interpreting the LHC Run 2 Data and Beyond, 27.05-31.05.2019 Agnieszka Ogrodnik (AGH UST), 


• n behalf of the ATLAS Collaboration

SLIDE 2

LbyL scattering in Pb+Pb collisions in the ATLAS experiment ILHC-ICTP2019

• A. Ogrodnik
• Light-by-light (LbyL) scattering is a rare QED process, which is forbidden in classical

electrodynamics

• In QED, the γγ → γγ reaction proceeds at lowest order in the 


fine-structure constant (αem) via virtual one-loop box diagrams 
 involving fermions

• Photon-photon interactions can be observed in heavy-ion 


collisions due to large EM fields associated with relativistic 
 ions (cross-sections scale with ~Z4)

• EM field is treated as a beam of quasi-real photons with 


small virtuality (equivalent photon approximation)

• Two datasets from Pb+Pb collisions at 5.02 TeV were collected 


by ATLAS detector and resulted in LbyL measurements

• 0.48 nb-1 in 2015: Nature Physics 13 (2017) 852
• 1.73 nb-1 in 2018: arXiv:1904.03536 NEW
• Also CMS experiment reported evidence of LbyL 


based on 2015 Pb+Pb data: arXiv:1810.04602

Introduction

• 2

PRC 93 (2016) 044907

SLIDE 3

LbyL scattering in Pb+Pb collisions in the ATLAS experiment ILHC-ICTP2019

• A. Ogrodnik

Possible interpretations

• Measurement of LbyL is sensitive to new physics
• Possible interpretations include establishing new limits on specific BSM models
• Axion-like particles searches, ATL-PHYS-PUB-2018-018
• Born-Infeld extension of QED, PRL 118 (2017) 261802
• First EFT constraints on nonlinear Lorentz-violating operators in QED, 

• Phys. Rev. D 99 (2019) 056016
• 3

SLIDE 4

LbyL scattering in Pb+Pb collisions in the ATLAS experiment ILHC-ICTP2019

• A. Ogrodnik

First ATLAS LbyL result

• ATLAS detector is optimised for detection of high energy

particles

• Many ingredients inefficient for low-ET region: 


trigger, photon reconstruction, photon identification

• LbyL scattering events have a very simple signature: 


the signal selection included:

• Two photons with ET > 3 GeV and |η| < 2.4, mγγ > 6 GeV
• Back-to-back topology: pγγ

T < 2 GeV, diphoton 


reduced acoplanarity, Aco = 1-|Δϕ|/π < 0.01

• 13 events found in the signal region, 7.3 signal 


events and 2.6 background events are expected

• Excess corresponds to 4.4σ statistical 


significance over background only hypothesis

• Considered as an evidence of the process
• 4

Nature Physics 13 (2017) 852

[GeV]

γ γ

m 5 10 15 20 25 30 Events / 3 GeV 2 4 6 8 10 12 ATLAS =5.02 TeV

NN

s Pb+Pb Signal selection with Aco < 0.01

• 1

b µ Data, 480 MC γ γ → γ γ MC

• e

+

e → γ γ MC γ γ CEP

SLIDE 5

LbyL scattering in Pb+Pb collisions in the ATLAS experiment ILHC-ICTP2019

• A. Ogrodnik

Analysis setup in 2018 - trigger

• In the 2015 ATLAS LbyL analysis the trigger was a major

source of inefficiency

• Strong emphasis was put to improve the trigger strategy

in 2018 heavy-ion data-taking (especially at the Level-1)

• Main improvements are:
• Optimised calorimeter noise settings
• Lower requirement on total ET at Level-1:
• Finally two approaches used: requirement of two EM

clusters with ET > 1 GeV OR one such EM cluster and total ET at Level-1 above 4 GeV

• Optimisation performed on Xe+Xe sample from 2017
• Some modifications at High Level Trigger (redefinition of

forward gap & relaxing the veto on activity in the tracker)

• Efficiency at Level-1 improved in 2018 wrt 2015
• 5

[GeV]

cluster2 T

+ E

cluster1 T

E 5 10 15 20 25 L1_TE5 efficiency 0.2 0.4 0.6 0.8 1 1.2 1.4 ATLAS Preliminary =5.02 TeV

NN

s Pb+Pb, > 1.5 GeV

cluster T

E = 2

clusters

N < 0.2

clusters

Aco

• 1

b µ Data, 480 Fit to data

[GeV]

cluster2 T

E +

cluster1 T

E

4 6 8 10 12 14

Level-1 trigger efficiency

0.2 0.4 0.6 0.8 1 ATLAS

=5.02 TeV

NN

s Pb+Pb

• 1

Data 2018, 1.7 nb Fit to data Stat syst ⊕ Stat

ATLAS-CONF-2016-111 arXiv:1904.03536

SLIDE 6

LbyL scattering in Pb+Pb collisions in the ATLAS experiment ILHC-ICTP2019

• A. Ogrodnik

Analysis setup in 2018 - photon ID

• The nominal ATLAS photon ID working points
• ptimised for high-ET photons
• Already in 2015 a dedicated photon ID was used

to increase the efficiency in low-ET region

• The improvements introduced in 2018 analysis:
• Switch from cut-based selection to neural

network based

• Use three additional shower shape variables
• Use η dependent ID
• Better background rejection
• ID efficiency on signal MC maintained 95%, in

agreement with efficiency measured with FSR photons

• 6

[GeV]

T

true photon E 5 10 15 20 25 Efficiency 0.2 0.4 0.6 0.8 1 1.2

ATLAS Simulation

=5.02 TeV

NN

s Pb+Pb Photon reconstruction photon PID

T

Low-E photon PID

T

High-E

[GeV]

T

Photon E

5 10 15 20 25

Photon PID efficiency

0.65 0.7 0.75 0.8 0.85 0.9 0.95 1 1.05

ATLAS =5.02 TeV

NN

s Pb+Pb FSR photons

• 1

Data 2018, 1.7 nb MC γ γ → γ γ

arXiv:1904.03536

SLIDE 7

LbyL scattering in Pb+Pb collisions in the ATLAS experiment ILHC-ICTP2019

• A. Ogrodnik

Backgrounds

• Various background sources considered:
• Exclusive dielectron production γγ → e+e-:
• Contribute to the signal region when tracks

not reconstructed, or bremsstrahlung photons emitted

• Central Exclusive Production (CEP) gg → γγ:
• Event kinematics very similar to signal, but

different shape of acoplanarity distribution

• Fakes (calo noise, cosmics)
• Others, found negligible (exclusive di-meson

production (e.g. π0π0), γγ → ττ, γγ → qq, 
 γγ → eeγγ, γγ → ηb → γγ, γPb → ϒ → 3γ, 
 ion bremsstrahlung)

• 7

SLIDE 8

LbyL scattering in Pb+Pb collisions in the ATLAS experiment ILHC-ICTP2019

• A. Ogrodnik

Signal selection

• Events are selected using following requirements:
• Two photons
• Identification: NN working point
• ET > 3 GeV, |η| < 2.37
• LbyL scattering topology
• mɣɣ > 6 GeV
• Veto extra particle activity: 


to suppress e+e- background

• Requiring no tracks (pT > 100 MeV)


and no pixel tracks (pT > 50 MeV, |Δη| < 0.5 photon-pixelTrk matching)

• Selecting back-to-back topology: to suppress fakes and CEP background
• pTɣɣ < 1 GeV (2 GeV for mɣɣ > 12 GeV)
• Diphoton acoplanarity < 0.01 

• 8

arXiv:1904.03536

SLIDE 9

LbyL scattering in Pb+Pb collisions in the ATLAS experiment ILHC-ICTP2019

• A. Ogrodnik

Results from 2018

• In total 59 events passing the signal selection are observed, with 


a background expectation of 12 ± 3 events

• Signal significance measured with events having Aco < 0.005, where 42 events pass signal

selection and 6 ± 2 background events are expected

• Observed signal significance over the background only hypothesis 


is of 8.2σ (expected 6.2σ)

• The corresponding fiducial cross section is 78 ± 13 (stat.) ± 7 (syst.) ± 3 (lumi.) nb
• SM predictions: 51 ± 5 nb (Szczurek et al.) and 50 ± 5 nb (SuperChic3)
• 9

0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08

φ

A 5 10 15 20 25 30 35 40 45 50 Events / 0.005 ATLAS = 5.02 TeV

NN

s Pb+Pb

• 1

Data 2018, 1.7 nb ) γ γ → γ γ Signal ( γ γ → CEP gg ee → γ γ

• Sys. unc.

5 10 15 20 25 30 [GeV]

γ γ

m 2 4 6 8 10 12 14 16 18 Events / GeV ATLAS = 5.02 TeV

NN

s Pb+Pb

• 1

Data 2018, 1.7 nb ) γ γ → γ γ Signal ( γ γ → CEP gg ee → γ γ

• Sys. unc.

0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2 [GeV]

γ γ T

p 5 10 15 20 25 Events / 0.2 GeV ATLAS = 5.02 TeV

NN

s Pb+Pb

• 1

Data 2018, 1.7 nb ) γ γ → γ γ Signal ( γ γ → CEP gg ee → γ γ

• Sys. unc.

arXiv:1904.03536

SLIDE 10

LbyL scattering in Pb+Pb collisions in the ATLAS experiment ILHC-ICTP2019

• A. Ogrodnik

Summary

• Light-by-light scattering was observed with the ATLAS detector

using data from Pb+Pb collisions at 5.02 TeV from 2018

• Precision of the new measurement greatly improved thanks to the

larger dataset and improved analysis techniques:

• The signal significance gives an observation with 8.2σ
• Measured fiducial cross-section: 


78 ± 13 (stat.) ± 7 (syst.) ± 3 (lumi.) nb

• Ratio of the measured cross-section to the SM predictions, 


51 ± 5 nb (Szczurek et al.) and 50 ± 5 nb (SuperChic3), is 
 1.53 ± 0.33 and 1.56 ± 0.33, respectively

• 10

SLIDE 11

LbyL scattering in Pb+Pb collisions in the ATLAS experiment ILHC-ICTP2019

• A. Ogrodnik

Backup

• Performance of

exclusive e+e- pairs - they are used for many cross-checks in the analysis

• MC simulation

is normalized to integrated luminosity

• 11

10 20 30 40 50 60 70 80 [GeV]

ee

m 1 10

2

10

3

10

4

10

5

10

6

10 Events / 2 GeV ATLAS = 5.02 TeV

NN

s Pb+Pb ee selection

• 1

Data 2018, 1.7 nb ee MC → γ γ

• Sys. unc.

10 20 30 40 50 60 70 80 [GeV]

ee

m 0.6 0.8 1 1.2 1.4 Data / MC 3 − 2 − 1 − 1 2 3

ee

y 500 1000 1500 2000 2500 Events / 0.1 ATLAS = 5.02 TeV

NN

s Pb+Pb ee selection

• 1

Data 2018, 1.7 nb ee MC → γ γ

• Sys. unc.

3 − 2 − 1 − 1 2 3

ee

y 0.6 0.8 1 1.2 1.4 Data / MC

0.5 1 1.5 2 2.5 [GeV]

ee T

p 500 1000 1500 2000 2500 3000 3500 4000 4500 5000 Events / 0.1 GeV ATLAS = 5.02 TeV

NN

s Pb+Pb ee selection

• 1

Data 2018, 1.7 nb ee MC → γ γ

• Sys. unc.

0.5 1 1.5 2 2.5 [GeV]

ee T

p 0.6 0.8 1 1.2 1.4 Data / MC 2 4 6 8 10 12 14 16 18 20 22 24 [GeV]

e T

p 5000 10000 15000 20000 25000 Events / GeV ATLAS = 5.02 TeV

NN

s Pb+Pb ee selection

• 1

Data 2018, 1.7 nb ee MC → γ γ

• Sys. unc.

2 4 6 8 10 12 14 16 18 20 22 24 [GeV]

e T

p 0.6 0.8 1 1.2 1.4 Data / MC

arXiv:1904.03536

SLIDE 12

LbyL scattering in Pb+Pb collisions in the ATLAS experiment ILHC-ICTP2019

• A. Ogrodnik

Backup

• Additional

kinematic distributions for the signal region

• 12

2 4 6 8 10 12 14 [GeV]

T

Leading photon E 5 10 15 20 25 30 35 40 45 50 Photons / GeV ATLAS = 5.02 TeV

NN

s Pb+Pb

• 1

Data 2018, 1.7 nb ) γ γ → γ γ Signal ( γ γ → CEP gg ee → γ γ

• Sys. unc.

3 − 2 − 1 − 1 2 3

γ γ

y 5 10 15 20 25 Events / 0.6 ATLAS = 5.02 TeV

NN

s Pb+Pb

• 1

Data 2018, 1.7 nb ) γ γ → γ γ Signal ( γ γ → CEP gg ee → γ γ

• Sys. unc.

0.5 1 1.5 2 2.5 3 3.5 4

γ γ

η Δ 5 10 15 20 25 30 35 40 Events / 0.4 ATLAS = 5.02 TeV

NN

s Pb+Pb

• 1

Data 2018, 1.7 nb ) γ γ → γ γ Signal ( γ γ → CEP gg ee → γ γ

• Sys. unc.

3 − 2 − 1 − 1 2 3 η photon 5 10 15 20 25 30 35 40 45 Photons / 0.6 ATLAS = 5.02 TeV

NN

s Pb+Pb

• 1

Data 2018, 1.7 nb ) γ γ → γ γ Signal ( γ γ → CEP gg ee → γ γ

• Sys. unc.

arXiv:1904.03536