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

Observation of light-by-light scattering 
 in lead-lead collisions in the ATLAS experiment Agnieszka Ogrodnik (AGH UST), 
 on behalf of the ATLAS Collaboration Interpreting the LHC Run 2 Data and Beyond, 27.05-31.05.2019

Introduction • 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 ( α em ) 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 ~ Z 4 ) • EM field is treated as a beam of quasi-real photons with 
 PRC 93 (2016) 044907 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 LbyL scattering in Pb+Pb collisions in the ATLAS experiment A. Ogrodnik � 2 ILHC-ICTP2019

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 LbyL scattering in Pb+Pb collisions in the ATLAS experiment A. Ogrodnik � 3 ILHC-ICTP2019

First ATLAS LbyL result Nature Physics 13 (2017) 852 • ATLAS detector is optimised for detection of high energy 12 Events / 3 GeV -1 ATLAS Data, 480 µ b particles γ γ → γ γ MC Pb+Pb s =5.02 TeV 10 - + e e MC γ γ → NN • Many ingredients ine ffi cient for low-E T region: 
 CEP γ γ MC 8 trigger, photon reconstruction, photon identification Signal selection with Aco < 0.01 6 • LbyL scattering events have a very simple signature: 
 4 the signal selection included: 2 • Two photons with E T > 3 GeV and | η | < 2.4, m γγ > 6 GeV 0 0 5 10 15 20 25 30 m [GeV] • Back-to-back topology: p γγ γ γ T < 2 GeV, diphoton 
 reduced acoplanarity, Aco = 1-| Δ ϕ |/ π < 0.01 • 13 events found in the signal region, 7.3 signal 
 events and 2.6 background events are expected • Excess corresponds to 4.4 σ statistical 
 significance over background only hypothesis • Considered as an evidence of the process LbyL scattering in Pb+Pb collisions in the ATLAS experiment A. Ogrodnik 4 � ILHC-ICTP2019

Analysis setup in 2018 - trigger 1.4 L1 _ TE5 efficiency ATLAS Preliminary • In the 2015 ATLAS LbyL analysis the trigger was a major 1.2 Pb+Pb, s =5.02 TeV NN source of ine ffi ciency 1 • Strong emphasis was put to improve the trigger strategy 0.8 N = 2 clusters in 2018 heavy-ion data-taking (especially at the Level-1) cluster 0.6 E > 1.5 GeV T Aco < 0.2 • Main improvements are: clusters 0.4 -1 Data, 480 b µ • Optimised calorimeter noise settings 0.2 Fit to data 0 • Lower requirement on total E T at Level-1 : 0 5 10 15 20 25 cluster1 cluster2 ATLAS-CONF-2016-111 E + E [GeV] T T • Finally two approaches used: requirement of two EM Level-1 trigger efficiency 1 ATLAS clusters with E T > 1 GeV OR one such EM cluster and total E T at Level-1 above 4 GeV 0.8 • Optimisation performed on Xe+Xe sample from 2017 0.6 • Some modifications at High Level Trigger (redefinition of Pb+Pb s =5.02 TeV NN 0.4 -1 Data 2018, 1.7 nb forward gap & relaxing the veto on activity in the tracker) Fit to data 0.2 Stat • E ffi ciency at Level-1 improved in 2018 wrt 2015 Stat syst ⊕ 0 4 6 8 10 12 14 arXiv:1904.03536 cluster1 cluster2 E + E [GeV] T T LbyL scattering in Pb+Pb collisions in the ATLAS experiment A. Ogrodnik 5 � ILHC-ICTP2019

Analysis setup in 2018 - photon ID arXiv:1904.03536 • The nominal ATLAS photon ID working points 1.2 Efficiency ATLAS Simulation optimised for high-E T photons 1 • Already in 2015 a dedicated photon ID was used 0.8 to increase the e ffi ciency in low-E T region 0.6 0.4 • The improvements introduced in 2018 analysis: Pb+Pb s =5.02 TeV NN Photon reconstruction 0.2 Low-E photon PID T High-E photon PID • Switch from cut-based selection to neural T 0 5 10 15 20 25 true photon E [GeV] network based T 1.05 Photon PID efficiency ATLAS • Use three additional shower shape variables 1 0.95 • Use η dependent ID 0.9 0.85 • Better background rejection 0.8 • ID e ffi ciency on signal MC maintained 95% , in 0.75 Pb+Pb s =5.02 TeV NN -1 Data 2018, 1.7 nb FSR photons 0.7 agreement with e ffi ciency measured with FSR MC γ γ → γ γ 0.65 photons 0 5 10 15 20 25 Photon E [GeV] T LbyL scattering in Pb+Pb collisions in the ATLAS experiment A. Ogrodnik � 6 ILHC-ICTP2019

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 di ff erent shape of acoplanarity distribution • Fakes (calo noise, cosmics) • Others , found negligible (exclusive di-meson production (e.g. π 0 π 0 ), γγ → ττ , γγ → qq, 
 γγ → ee γγ , γγ → η b → γγ , γ Pb → ϒ → 3 γ , 
 ion bremsstrahlung) LbyL scattering in Pb+Pb collisions in the ATLAS experiment A. Ogrodnik � 7 ILHC-ICTP2019

Signal selection arXiv:1904.03536 • Events are selected using following requirements: • Two photons • Identification: NN working point • E T > 3 GeV, | η | < 2.37 • LbyL scattering topology • m ɣɣ > 6 GeV • Veto extra particle activity : 
 to suppress e+e- background • Requiring no tracks (p T > 100 MeV) 
 and no pixel tracks (p T > 50 MeV, | Δη | < 0.5 photon-pixelTrk matching) • Selecting back-to-back topology : to suppress fakes and CEP background • p T ɣɣ < 1 GeV (2 GeV for m ɣɣ > 12 GeV) • Diphoton acoplanarity < 0.01 
 LbyL scattering in Pb+Pb collisions in the ATLAS experiment A. Ogrodnik � 8 ILHC-ICTP2019

Results from 2018 arXiv:1904.03536 • 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 8.2 σ (expected 6.2 σ ) • 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) • 50 18 Events / 0.2 GeV Events / 0.005 Events / GeV ATLAS 25 ATLAS ATLAS 45 16 Pb+Pb s = 5.02 TeV Pb+Pb s = 5.02 TeV Pb+Pb s = 5.02 TeV NN NN NN 40 14 20 35 -1 -1 -1 Data 2018, 1.7 nb Data 2018, 1.7 nb 12 Data 2018, 1.7 nb Signal ( ) Signal ( ) Signal ( ) γ γ → γ γ 30 γ γ → γ γ γ γ → γ γ 15 10 CEP gg CEP gg CEP gg → γ γ → γ γ → γ γ 25 ee ee γ γ → ee γ γ → γ γ → 8 Sys. unc. Sys. unc. Sys. unc. 20 10 6 15 4 10 5 2 5 0 0 0 0 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 5 10 15 20 25 30 0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2 A m [GeV] γ γ p [GeV] γ γ φ T LbyL scattering in Pb+Pb collisions in the ATLAS experiment A. Ogrodnik 9 � ILHC-ICTP2019

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 8.2 σ • 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 LbyL scattering in Pb+Pb collisions in the ATLAS experiment A. Ogrodnik � 10 ILHC-ICTP2019