Photons in LHC data @ 7 TeV Results from ATLAS and CMS Mathieu - PowerPoint PPT Presentation

Photons in LHC data @ 7 TeV Results from ATLAS and CMS Mathieu Aurousseau (LAPP) On behalf of the ATLAS and CMS Collaborations Higgs Hunting : Discussions on Tevatron and first LHC results 29-31 July 2010 Orsay, France Higgs Hunting 2010 1 Mathieu Aurousseau - LAPP Orsay (29-31 July 2010)

Outline I – Introduction • Physics with photons at the LHC • ATLAS and CMS : trackers and calorimeters II – Using photons in performance studies • Photons from 0 decays [CMS/ATLAS] • Conversions and Dalitz [CMS/ATLAS] III – Prompt photons results • Photon identification [CMS/ATLAS] • Photon conversions [CMS/ATLAS] • Beam-halo background [CMS] • Prompt isolated photon signal and purity [ATLAS] Higgs Hunting 2010 2 Mathieu Aurousseau - LAPP Orsay (29-31 July 2010)

I- Introduction Higgs Hunting 2010 3 Mathieu Aurousseau - LAPP Orsay (29-31 July 2010)

Perspective for H searches in CMS and ATLAS • Clean signature (2 high-E T isolated photons) • Huge background from QCD photons and jets For m H = 120 GeV => BR ~ 0.2% CMS What can we extract from the first 10-100 nb -1 ? Higgs Hunting 2010 4 Mathieu Aurousseau - LAPP Orsay (29-31 July 2010)

Photon production In today’s data q q gg • Among reconstructed qg gq di-jet photons :  Jets with high-pT 0 Compton Annihilation Pythia cross section @ 7 TeV misidentified as photons  Prompt photons (from ME and fragmentation) 0 separation -jet • / Born Box • Energy + angular resolution • Conversion recovery Fragmentation (Brem) • Prompt photon measurements :  Background to H , … H  Perturbative QCD tests • Isolation Higgs Hunting 2010 5 Mathieu Aurousseau - LAPP Orsay (29-31 July 2010)

ATLAS and CMS : Trackers ~1.2m ~2.1m Pixels Silicon strips Material in the Inner Detectors • Material => non-negligible fraction of photon conversions (up to 50% in Si)  Consequences on photon ID, energy resolution, etc.  Use of conversions as a tool for material mapping • ATLAS : add O (2X 0 ) before EM Calo (solenoid coil, cryostat) Higgs Hunting 2010 6 Mathieu Aurousseau - LAPP Orsay (29-31 July 2010)

ATLAS and CMS : Calorimetry • Pb + LAr sampling calorimeter • PbWO 4 scintillating crystals • 3 radial layers + pre-shower • Preshower in front of EE • Design energy resolution : • Design energy resolution : 10 % 300 MeV 2 . 9 % 125 MeV E E 0 . 7 % 0 . 3 % E E E E E E • Outside solenoid coil • Inside solenoid coil Higgs Hunting 2010 7 Mathieu Aurousseau - LAPP Orsay (29-31 July 2010)

II – Photons in performance studies 0 peaks • Reconstruction of • Material mapping with photon conversions and Dalitz in ATLAS Higgs Hunting 2010 8 Mathieu Aurousseau - LAPP Orsay (29-31 July 2010)

Photons from 0 Diphoton invariant mass spectrum L = 414.4 b -1 L = 123 nb -1 0 E T ( ) > 400 MeV p T ( ) > 900 MeV 0 σ = 14.9% E T ( ) > 300 MeV p T ( ) > 900 MeV ATLAS Preliminary Uniformity in • Channel-by-channel in-situ intercalibration precision : • Uniformity along : < 0.7% in EM Calorimeter 1.2% for | | < 0.8 ( 0 , symmetry) (| | < 2.37) • Agreement data/MC in energy scale ~ 1 – 3 % • Agreement data/MC in energy scale along ~ 1 – 2 % (Barrel – Endcap) Higgs Hunting 2010 9 Mathieu Aurousseau - LAPP Orsay (29-31 July 2010)

Converted photons in ATLAS and CMS Some discrepencies Pixel support shift ~ 1cm 3 pixel layers Corrected from overall shift between tracker and reference frame • Conversions :  Track-based conversion finding • Dalitz decays ( 0 e + e - )  Used to constrain material in the Beam pipe  Beam pipe = reference for material estimate Dalitz : Beam pipe 0 e + e - Higgs Hunting 2010 10 Mathieu Aurousseau - LAPP Orsay (29-31 July 2010)

III – Prompt photons results • Photon identification • Photon conversions • Non-collision backgrounds • Extraction of isolated photon signal and purity Higgs Hunting 2010 11 Mathieu Aurousseau - LAPP Orsay (29-31 July 2010)

Data samples • Integrated luminosity • (Super)Clusters  ATLAS : L = 15.8 nb -1  ATLAS : E Cl > 10 GeV  CMS : L = 74 nb -1  CMS : E SC > 20 GeV  | | < 2.37 / 2.5, out of crack (53.6 nb -1 for beam bkg study) In the  Remove problematic regions / • Event selection next plots anomalous signals  L1 trigger : E EM > 5 GeV  ATLAS : 2.3 10 5 photon cand.  ≥ 1 vertex with ≥ 3 tracks  CMS : 1.3 10 5 events EB Higgs Hunting 2010 12 Mathieu Aurousseau - LAPP Orsay (29-31 July 2010)

Identification of photons in ATLAS Variable Definition Loose Hadronic E T (Had)/E T selection leakage  Had. MIDDLE Leakage + R E 3x7 /E 7x7 Middle w 2 RMS of energy in in middle variables layer R E 3x3 /E 3x7 Tight w stot RMS of energy in in strips selection Asymetry between 1 st and 2 nd E ratio  Loose + max Strips 0 STRIPS variables Difference between 2 nd max and E min (between the 2 max) (different cuts for converted F side Fraction of energy in 7 cells and outside the core of 3 cells unconverted STRIPS photons) w s3 RMS of energy in in the core of 3 cells MIDDLE • Isolation  Treated separately (see after) Higgs Hunting 2010 13 Mathieu Aurousseau - LAPP Orsay (29-31 July 2010)

Shower shapes in the ATLAS EM Calorimeter Transverse | |<0.6 F side shape R Loose candidates R E 3x7 / E 7x7 E max1 E max2 Containment in strips E ratio (Fside) Loose candidates E out / E 7strips Strips E E max1 max 2 E ratio E E max1 max 2 Efficiencies computed Good agreement from H MC between Data (E T > 20 GeV) : and MC after Loose : ~ 95% loose ID cut Tight : ~ 88 % Higgs Hunting 2010 14 Mathieu Aurousseau - LAPP Orsay (29-31 July 2010)

Identification of photons in CMS Variable Definition R 9 E 3x3 /E SC jet Pixel seed Match with track in pixels Tracker Iso. Sum(p T ) in a ring between R = Separation of 0.04 and R = 0.4 conversions, choice of cluster ECAL Iso. Sum(E T (EM)) in a ring between R size = 0.06 and R = 0.4 HCAL Iso. Sum(E T (HCAL)) in a ring between R = 0.15 and R = 0.4 σ i -i Width in of the SC Hadronic H/E = E T (Had)/E T • Purity (from MC) ~ 50% fraction • Increases with E T Additional cut : E T > 30 GeV • Efficiencies computed on MC isolated photons :  EB : ≈ 90%  EE : ≈ 80% Higgs Hunting 2010 15 Mathieu Aurousseau - LAPP Orsay (29-31 July 2010)

Identification variables in the CMS ECAL Track Isol EE EB N – 1 σ i -i distributions (apply all cuts except the one plotted) Signal contribution EB • Good agreement between Data and MC R 9 • Signal contribution visible On selected photons Higgs Hunting 2010 16 Mathieu Aurousseau - LAPP Orsay (29-31 July 2010)

Photon conversions in CMS and ATLAS L = 62 nb -1 ATLAS Preliminary All normalized to 1 • Selected photons : • Selected photons :  No pixel seed  Tight selection  E T > 30 GeV  E T > 20 GeV  Track Iso < 2 GeV  Isolation < 3 GeV  ECAL Iso. < 4.2 GeV  HCAL Iso. < 2.2 GeV  H/E < 0.05  σ i -i < 0.01 (0.03) Distributions compatible with  | | < 0.2 isolated converted photons  | cot | < 0.3  P(vertex) > 5.10-4 Higgs Hunting 2010 17 Mathieu Aurousseau - LAPP Orsay (29-31 July 2010)

Non-collision background (I) Non-collision events = EMU/HE tags background to studies with photons or MET HALO • HALO :  No reco track  MET > 25 GeV  tagged in Had Endcap (HE) or BEAM Muon System Endcap (EMU) IP • PROMPT :  > 3 reco tracks  MET < 15 GeV  |t seed | < 3 ns HCAL  no tag ECAL Muon • CANDIDATE : System  > 3 reco tracks Time of seed in ECAL : t seed  MET > 25 GeV Angle of shower in - plane  photon with ET > 30 GeV w.r.t beam axis : A  |t seed | < 3ns  no tag Higgs Hunting 2010 18 Mathieu Aurousseau - LAPP Orsay (29-31 July 2010)

Non-collision background (II) t seed [ns] HALO HALO PROMPT • Data driven method to estimate candidate contamination • Estimate < 5.9 halo events in “candidate” sample (351 candidates) with MET > 25 GeV Higgs Hunting 2010 19 Mathieu Aurousseau - LAPP Orsay (29-31 July 2010)

Isolation • Energy in ring includes : Cone (0.4) Core (5x7)  Photon leakage out of the core EM + HAD EM • Depends on photon pT • Subtracted from the ring energy  Pile-up / Underlying Event effects • Ambient energy density from low E jets • Subtracted from the ring energy  Nearby hadronic activity Isolation energy This definition of isolation is closer to theoretical parton- Candidates level isolation already passing the loose ID cut Higgs Hunting 2010 20 Mathieu Aurousseau - LAPP Orsay (29-31 July 2010)

Signal extraction (I) Tight ID Control Control cut region region Two corrections : Fail cut • Correlation between ID and isolation cuts Control  Reverse only a subset (E ratio , E, F side , W s3 ) region of the strips shower shape variables to Pass cut minimize correlation • Signal contamination in control regions  Use fraction of signal in the control regions from MC (c 1 , c 2 , c 3 ) Isolated signal region Signal contamination Correlation Corrected formula : Data driven extraction of number of isolated signal Higgs Hunting 2010 21 Mathieu Aurousseau - LAPP Orsay (29-31 July 2010)