The CMS Particle Flow Algorithm

Similar documents
ATLAS jet and missing energy reconstruction, calibration and performance in LHC Run-2

Future prospects for the measurement of direct photons at the LHC

CMS Physics Analysis Summary

The ATLAS Run 2 Trigger: Design, Menu, Performance and Operational Aspects

ATLAS-CONF October 15, 2010

Performance of muon and tau identification at ATLAS

Measurement of Jet Energy Scale and Resolution at ATLAS and CMS at s = 8 TeV

Jet Reconstruction and Energy Scale Determination in ATLAS

Physics potential of ATLAS upgrades at HL-LHC

Physics object reconstruction in the ATLAS experiment

PERFORMANCE OF THE ATLAS MUON TRIGGER IN RUN 2

Application of the Tau Identification Capability of CMS in the Detection of Associated Production of MSSM Heavy Neutral Higgs Bosons Souvik Das

Di muons and the detection of J/psi, Upsilon and Z 0 Jets and the phenomenon of jet quenching

Physics with Tau Lepton Final States in ATLAS. Felix Friedrich on behalf of the ATLAS Collaboration

Search for high mass diphoton resonances at CMS

Conference Report Mailing address: CMS CERN, CH-1211 GENEVA 23, Switzerland

Matt Nguyen Rencontres QGP-France September 13 th, 2013

Muon reconstruction performance in ATLAS at Run-2

Top quarks objects definition and performance at ATLAS

Particle Flow Algorithms

arxiv: v1 [hep-ex] 2 Nov 2010

The CMS Experiment: Status and Highlights

How to Measure Top Quark Mass with CMS Detector??? Ijaz Ahmed Comsats Institute of Information Technology, Islamabad

PoS(EPS-HEP 2013)508. CMS Detector: Performance Results. Speaker. I. Redondo * CIEMAT

First V+jets results with CMS. Vitaliano Ciulli (Univ. & INFN Firenze) V+jets workshop, 8-10 Sep 2010, Durham

CALICE scintillator HCAL

ATLAS Calorimetry (Geant)

The rejection of background to the H γγ process using isolation criteria based on information from the electromagnetic calorimeter and tracker.

Search for heavy BSM particles coupling to third generation quarks at CMS

CMS Note Mailing address: CMS CERN, CH-1211 GENEVA 23, Switzerland

Upgrade of ATLAS and CMS for High Luminosity LHC: Detector performance and Physics potential

Atlas Status and Perspectives

Electroweak Physics at the Tevatron

Z 0 /γ +Jet via electron decay mode at s = 7TeV in

Physics at Hadron Colliders

PoS(CORFU2016)060. First Results on Higgs to WW at s=13 TeV with CMS detector

Validation of Geant4 Physics Models Using Collision Data from the LHC

Hà γγ in the VBF production mode and trigger photon studies using the ATLAS detector at the LHC

Reconstruction and identification of hadronic τ decays with ATLAS

Studies on the e + e - spectrum with the first data of the CMS experiment at the Large Hadron Collider

Measurement of the Inclusive Isolated Prompt Photon Cross Section at CDF

Risultati dell esperimento ATLAS dopo il run 1 di LHC. C. Gemme (INFN Genova), F. Parodi (INFN/University Genova) Genova, 28 Maggio 2013

PoS(DIS 2010)190. Diboson production at CMS

Measurement of charged particle spectra in pp collisions at CMS

High Pt Top Quark Mass Reconstruction in CMS

Improving the Jet Reconstruction with the Particle Flow Method; an Introduction

PoS(ICHEP2012)311. Identification of b-quark jets in the CMS experiment. Sudhir Malik 1

The ATLAS muon and tau triggers

Lepton and Photon reconstruction and identification performance in CMS and ATLAS

Boosted top quarks in the ttbar dilepton channel: optimization of the lepton selection

Higgs-related SM Measurements at ATLAS

W and Z boson production in p p collisions from Run II of the Tevatron Collider

Missing Transverse Energy. Performance with CMS

Measurement of Quenched Energy Flow for Dijets in PbPb collisions with CMS

Tau Lepton Reconstruction in ATLAS. Tau 2016 Conference, Beijing, 21st September 2016

THE ATLAS TRIGGER SYSTEM UPGRADE AND PERFORMANCE IN RUN 2

Discovery Physics at the Large Hadron Collider

How to find a Higgs boson. Jonathan Hays QMUL 12 th October 2012

HL-LHC Physics with CMS Paolo Giacomelli (INFN Bologna) Plenary ECFA meeting Friday, November 23rd, 2012

Reconstruction in Collider Experiments (Part IX)

Discovery of the W and Z 0 Bosons

Performance of the CMS missing transverse momentum reconstruction in pp data at s = 8 TeV

Precise measurements of the W mass at the Tevatron and indirect constraints on the Higgs mass. Rencontres de Moriond QCD and High Energy Interactions

arxiv: v3 [hep-ex] 21 Mar 2014

Preparations for the ATLAS Heavy Ion Physics Program at the LHC. Deepak Kar IKTP, TU Dresden On behalf of the ATLAS Collaboration

Jet quenching in PbPb collisions in CMS

Jet Energy Calibration. Beate Heinemann University of Liverpool

Measurement of the associated production of direct photons and jets with the Atlas experiment at LHC. Michele Cascella

The search for standard model Higgs boson in ZH l + l b b channel at DØ

Identification and rejection of pile-up jets at high pseudorapidity with the ATLAS detector

Measurement of the Higgs Couplings by Means of an Exclusive Analysis of its Diphoton decay

Measurement of jet production in association with a Z boson at the LHC & Jet energy correction & calibration at HLT in CMS

Upgrade of the CMS Forward Calorimetry

Early physics with Atlas at LHC

Boosted Top Resonance Searches at CMS

Data-driven approaches to pile-up treatment at the LHC

PoS(EPS-HEP2011)250. Search for Higgs to WW (lνlν, lνqq) with the ATLAS Detector. Jonas Strandberg

Search for high mass resonances in the diphoton and Zγ channels at LHC

Observation of a New Particle with a Mass of 125 GeV

Top Quark Mass Reconstruction from High Pt Jets at LHC

Higgs and Z τ + τ in CMS

Thesis. Wei Tang. 1 Abstract 3. 3 Experiment Background The Large Hadron Collider The ATLAS Detector... 4

Status of the LHCb experiment and minimum bias physics

Differential photon-jet cross-section measurement with the CMS detector at 7 TeV

Tau Reconstruction at CMS. Evan Friis

ATLAS NOTE. September 26, 2016

La ricerca dell Higgs Standard Model a CDF

VBF SM Higgs boson searches with ATLAS

CMS Higgs Results Adi Bornheim Caltech

ATLAS: Status and First Results

Study of Z+Jet in CMS

Dilepton Forward-Backward Asymmetry and electroweak mixing angle at ATLAS and CMS

Upgrade of ATLAS Electron and Photon Triggers and Performance for LHC Run2

Higgs couplings and mass measurements with ATLAS. Krisztian Peters CERN On behalf of the ATLAS Collaboration

Higgs cross-sections

Design of the new ATLAS Inner Tracker for the High Luminosity LHC era

Measurement of the Z ττ cross-section in the semileptonic channel in pp collisions at s = 7 TeV with the ATLAS detector

ATLAS EXPERIMENT : HOW THE DATA FLOWS. (Trigger, Computing, and Data Analysis)

Electroweak results. Luca Lista. INFN - Napoli. LHC Physics

Jet energy measurement in the ATLAS detector

Transcription:

PROCEEDINGS of CHEF 13 he CMS Particle Flow Algorithm Laboratoire Leprince-Ringuet, INP3-CNRS E-mail: Florian.Beaudette@cern.ch A particle flow event-reconstruction algorithm has been successfully deployed in the CMS experiment and is nowadays used by most of the analyses. It aims at identifying and reconstructing individually each particle arising from the LHC proton-proton collision, by combining the information from all the subdetectors. he resulting particle-flow event reconstruction leads to an improved performance for the reconstruction of jets and ME, and for the identification of electrons, muons, and taus. he algorithm and its performance will be described. he commissioning phase, during which it was demonstrated that the algorithm was performing as expected from the simulation up to a high level of precision, will be presented. Finally, a selection of recent improvements in the CMS analyses obtained thanks to the particle-flow algorithm will be discussed. Calorimetry for High Energy Frontiers - CHEF 13, April -5, 13 Paris, France Speaker.

he CMS Particle Flow Algorithm 1. he Particle Flow Algorithm he Particle Flow (PF) algorithm aims at identifying and reconstructing all the particles from the collision by combining optimally the information of the different subdetectors. he ingredients for an efficient PF algorithm arise from these principles: to maximise the separation between charged and neutral hadrons, a large field integral and a good calorimeter granularity are of primary importance. An efficient tracking is a key item as well as a material budget in front of the calorimeters as small as possible. he CMS detector [1] fulfill several of these conditions with a field integral more than twice larger than in other past or existing experiments; an electromagnetic calorimeter with an excellent resolution and granularity, and a tracker system fully exploited by the iterative tracking algorithm. he CMS PF algorithm [] relies on a efficient and pure track reconstruction, on a clustering algorithm able to disentangle overlapping showers, and on an efficient link procedure to connect together the deposits of each particle in the sub-detectors. Simplifying, the algorithm can be described as follows; a detailed description can be found in []. he tracks are extrapolated through the calorimeters, if they fall within the boundaries of one or several clusters, the clusters are associated to the track. he set of track and cluster(s) constitute a charged hadron and the building bricks are not considered anymore in the rest of the algorithm. he muons are identified beforehand so that their track does not give rise to a charged hadron. he electrons are more difficult to deal with. Indeed, due to the frequent Bremsstrahlung photon emission, a specific track reconstruction [3] is needed as well as a dedicated treatment to properly attach the photon clusters to the electron and avoid energy double counting. Once all the tracks are treated, the remaining clusters result in photons in case of the electromagnetic calorimeter (ECAL) and neutral hadrons in the hadron calorimeter (HCAL). Once all the deposits of a particle are associated, its nature can be assessed, and the information of the sub-detectors combined to determine optimally its four-momentum. In case the calibrated calorimeter energy of the clusters, which is simply a linear combination of the ECAL and HCAL energy deposits, associated to a track is found to be in excess with respect to the track momentum at more than one sigma, the excess is attributed to an overlapping neutral particle (photon or hadron), carrying an energy corresponding to the difference of the two measurements. he resulting list of particles, namely charged hadrons, photons, neutral hadrons, electrons and muons, is then used to reconstruct the jets, the missing transverse energy (E miss ), to reconstruct and identify the taus from their decays products and to measure the isolation of the particles. he performance of the algorithm have been first been studied with simulated events []. About 9% of the jet energy is carried by the charged hadrons and the photons the energies of which are measured with a high precision by the tracker and the ECAL respectively, while the remaining 1%, carried by the neutral hadrons, is measured with the hadron calorimeter (HCAL) with a 1%/ p E resolution. As visible in Fig. 1, where the jets reconstructed with the sole calorimeters and the jets reconstructed with the PF candidates are compared for di-jets events in the barrel, between 95% and 97% of the jet energy, depending on the jet p, is reconstructed in the case of the PF compared to 6%-8% for the calorimeter jets (left plot). In addition, the gain in resolution can be up to a factor of 3 (right plot). Furthermore, the angular resolution is improved by a factor -3.

he CMS Particle Flow Algorithm CMS Preliminary Jet Response. -. -.4 -.6 -.8 Particle-Flow Jets Calo-Jets < η < 1.5 CMS Preliminary Jet-Energy Resolution.45.4.35.3.5..15.1.5 Corrected Calo-Jets Particle-Flow Jets < η < 1.5 1 3 4 5 6 p (GeV/c) 1 p [GeV/c] Figure 1: (left) Jet response for di-jets events in the barrel reconstructed with the particle flow (red triangles) and the calorimeters (blue open squares). (right) Jet resolution for di-jets events in the barrel reconstructed with the particle flow (red triangles) and the calorimeters (blue open squares). he hadronic decays of the t-leptons result in narrow jets made of a small number of particles, and do not contain neutral hadrons. Consequently, the reconstruction of the t-leptons is highly improved by the PF reconstruction. Firstly, the measurement of the t visible energy as shown in Fig. is spectacularly enhanced; secondly the identification of the particles provides access to the decay channel. CMS Preliminary 3 5 15 PFlowaus Caloaus η < 1.4 1 5-3 - 1 3 E [GeV] Figure : Resolution in E for hadronic t jets from Z decays, in the barrel, reconstructed using the particle flow (filled), and calorimeters (hollow).. Commissioning with the first data A large fraction of the commissioning work was carried out with the first minimum-bias events resulting from the proton-proton collisions delivered by the LHC at a center-of-mass ( p s) energy of 9 GeV[4]. It continued with the first collisions at p s=.36ev. he number of high p and isolated leptons, coming from e.g. W ± or Z, and high p jets, was however very limited until the 3

he CMS Particle Flow Algorithm LHC switched to p s=7 ev with an increasing luminosity [5, 6]. An integrated luminosity of a few nb 1 was sufficient to check the calibration of the calorimeters. For the ECAL calibrations, photons from p decays have been used. All the photons with h < 1 and at least 4MeV of energy were paired. In addition the energy of the pair was required to be larger than 1.5 GeV. he resulting invariant-mass distribution of the photon pairs in the barrel can be seen in Fig. 3 where it can observed that the agreement of the average reconstructed, as obtained from a Gaussian fit from the signal is in agreement with the world average within 1%. A similar level of agreement with the Geant-based simulation is obtained. he calibration of the hadrons is assumed to be identical for charged and neutral hadrons. hus, Number of photon pairs 15 1 fit m π = 135. ±.1 MeV/ c σ mπ = 13. ±.1 MeV/ c 7-eV Data,.1 nb CMS Preliminary 1 Number of photon pairs 15 1 fit m π = 136.9 ±. MeV/ c σ mπ = 1.8 ±. MeV/ c Simulation CMS Preliminary 1 5 5.5.1.15..5.3.35.4.45.5 Mass (GeV/c ).5.1.15..5.3.35.4.45.5 Mass (GeV/c ) Figure 3: Photon-pair invariant-mass distribution in the barrel ( h < 1.) for the data (left) and simulation (right) charged hadrons could be used to check the calibration. A quite pure sample of charged hadrons has been obtained by requiring high quality tracks, with p > 1GeV/c, p > 3GeV/c. In addition, the energy of the HCAL cluster had to exceed 1 GeV. Finally a loose isolation criteria was applied. he energy measured by the calorimeters could then be compared with the track momentum. he distribution thus obtained is presented in Fig. 4 and it can be observed that the slope of the calibrated energy (in red) is close to 1. For illustration, the raw energy is also presented in the same plot. he small deviation at low momentum, also present in the simulation (see [5] ) is a bias of the check, and is attributed to neutral particles overlapping with the charged hadron clusters. Once the linking procedure and the particle calibration were proven to work in the data as expected from the simulation, it was time to compare the jet distributions. he jet multiplicity, p and angular distributions, which can be found in [4, 5] exhibit a very satisfactory data-simulation level of agreement. he access to the jet composition, i.e. the fraction of the jet energy carried by the photons, charged and neutral hadrons in the tracker acceptance or hadronic and electromagnetic energy deposit in the forward region, was studied as a function of the jet h (Fig. 5) and p, showing again, the goodness of the simulation. Such studies have been continuously carried out during the data taking, as illustrated in Fig. 6 extracted from [7] where the jet composition is shown as a function of the jet p in the barrel region. he simulation reproduces well the data even with a high pile-up multiplicity. he reconstruction was further commissioned, first by studying the E miss, a quantity which is known to be sensitive to detector effects. In the case of the particle-flow event reconstruction, the 4

he CMS Particle Flow Algorithm Calibrated Calo Energy (GeV) 1 9 8 7 6 5 4 3 CMS Preliminary 1 (Barrel) 1 7-eV Data, 7.5 nb 7-eV Data, 7.5 nb (Calibrated) (Uncalibrated) 1 3 4 5 6 7 8 9 1 rack momentum (GeV/c) Figure 4: Average raw (black) and calibrated (red) calorimeter response as a function of the track momentum for charged hadrons selected in the data in the barrel. Mean Fraction of Jet Energy 1. 1.8.6 CMS Preliminary 1 s = 7 ev, DAA (6. nb ) Anti-k R=.5 corr p > 5 GeV/c Mean Fraction of Jet Energy 1. 1.8.6 CMS Preliminary 1 s = 7 ev, MC Anti-k R=.5 corr p > 5 GeV/c.4.4.. -4-4 PFJet -4-4 PFJet η Figure 5: Reconstructed jet energy fractions as a function of pseudo-rapidity in the data (left) and in the simulation (right). From bottom to top in the central region: charged hadrons, photons, electrons, and neutral hadrons. In the forward regions: hadronic deposits, electromagnetic deposits. is simply the modulus of the vector sum of the reconstructed particle transverse momenta. As visible in Fig. 7, the distribution obtained in the data is reproduced by the simulation with an impressive agreement over several order of magnitudes [5]. Even more challenging, the scalar sum of the particle momenta, ÂE, for which no noise cancellation can occur, is also faithfully reproduced by the simulation, proving the refined knowledge of the detector included in the simulation. With the large statistics available at the end of the data-taking, boosted Z and g+jets events can E miss be used to study the performance of the E miss reconstruction. As visible in Fig. 8, an impressive data-simulation agreement on the resolution is obtained [8]. he reconstruction of the electrons and the muons, at high p and also at low p was, of course, also commissioned, with J/y! `` events (Fig. 9) as well as W ±! ` ± n events as described in [4, 5]. As mentioned earlier, the PF is well suited to quantify the isolation of the leptons and 5

he CMS Particle Flow Algorithm PF energy fraction.9.8.7.6.5.4.3..1 CMS preliminary, L = 1.6 fb s = 8 ev MC: histo, Data: marker Charged pile-up Charged hadrons Photons Neutral hadrons Electrons Muons η <1.3 ag-and-probe method 3 1 1 (GeV) Figure 6: Jet composition as a function of the jet p for jets with h < 1.3 for data (histogram) and simulation (markers). p Figure 7: Distribution of the E miss (histogram). (left) and SE (right) for di-jet events from data (dots) and simulation of the photons. he precision needed by some analyses, e.g. [9], requires clean event selections to be applied. As an example, Z! µµg events allowed the photon isolation around photons to be carefully monitored [1] as shown in Fig. 1 where a data-simulation agreement over more than two orders of magnitude is observed. 3. Impact on the analyses he gain in terms of higher efficiency, lower fake-rate, better resolution for the jet, E miss and t-lepton reconstruction, together with the reduced systematics had, obviously, a major impact on the majority of the analyses carried out in CMS since the advent of the PF. he benefits of the PF for the leptons (electrons and muons) and for the isolated photons appeared a bit later, in the period when a cutting-edge lepton and photon identification as well as the mitigation of the pile-up were crucial in the context of the Higgs boson search. raditionally, the lepton and photon isolation are computed as the sum of the transverse momenta and energies of the tracks and deposits in the ECAL and the HCAL in a cone around the particle considered. In the PF approach, the tracks and 6

he CMS Particle Flow Algorithm ) [GeV] Y σ(e 5 4 3 Z/γ q Z µµ Z ee γ+jets > 1 GeV PF E CMS Preliminary 1 1. fb at s = 8 ev 1 Data/MC 1. 1.8.5 1 1.5.5 E [ev] Figure 8: Resolution of the PF E miss projection along the y-axis as a function of SE for events with Z and photon. Results are shown for Zµµ events (full blue circles), Zee events (open red circles), and photon events (full green squares). Events/.1 GeV/c 14 CMS Preliminary 1 7-eV data, L = 57.5 nb 1 1 8 6 N sig = 4 ± 19 M = 3.7 ±.1 GeV/c = 93 ± 8 MeV/c 4 1.5.5 3 3.5 4 4.5 m ee [GeV/c ] Figure 9: Invariant-mass distribution of the J/y candidates decaying into an electron pairs in the first data. energy deposits are simply replaced by the reconstructed particles. he gain brought by the PF isolation is illustrated in Fig. 11-left for electrons with p < 1GeV/c in the endcaps. he signal and background efficiencies are relative to electrons already identified, mostly with shower-shape and track-cluster matching criteria. For the 87% efficiency working point used in the analysis, a factor of two in background rejection is achieved using the PF isolation with respect to the subdetectorbased isolation. A similar background reduction is obtained for 1 < p < GeV/c electrons. he PF isolation not only gives superior performance, it also provides handles to mitigate the effect of the pile-up events. Since the charged hadrons have, by construction, a corresponding track, it is possible to impose that the charged hadrons considered for the isolation computation come from the primary vertex. As a result, the average energy contained in the isolation carried by the charged hadron exhibits little dependency on the pile-up event multiplicity, measured by the number of reconstructed vertices, as it can observed on the blue curve in Fig. 11-right. he neutrals from pile-up events falling in the isolation cannot be identified and their contribution is firstly 7

he CMS Particle Flow Algorithm CMS preliminary, s = 8 ev, L = 19.6 fb Events/. GeV 3 1 1 Z µµγ MC 8eV Data 1 1 3 4 5 6 7 8 9 1 Neutral EM isolation (GeV) Figure 1: he electromagnetic isolation sum for pre-selected photons in simulated of the Z! µµg events compared to data for the ECAL barrel. background efficiency CMS Preliminary 1.9 det. iso (11 def.).8.7 PF iso (1 def.).6.5 s = 8 ev, L = 5.1 fb <E flow> [GeV] 15 CMS Preliminary, s = 8 ev, L dt = 19.6 fb 5 ρ charged particles neutral particles.4.3..1.8.85.9.95 1 signal efficiency 1 5 η < 1, p > GeV 5 1 15 5 3 # vertices Figure 11: (left) For low p electrons in the endcaps, background efficiency as a function of the signal efficiency with detector isolation (red circles) and PF isolation (blue squares).(right) For electrons in the central region with p > GeV/c: evolution of the average event energy density (open circles) and energy within the isolation cone carried by the charged particles (blue circles) and neutral particles (red circles). estimated before being subtracted. Several methods exist to do so, and one of them is illustrated in Fig. 11-right where it can be seen that both the average energy density (r) and the contribution from neutrals scale linearly with the number of vertices. As a result, the average energy neutral energy caused by the pile-up events in the isolation cone is r multiplied by a constant. he isolation of neutrals thus corrected has almost no dependency on the pile-up and so does the global PF-isolation as can be checked in Fig. 1 where the efficiency of an electron isolation cut is plotted as a function of the number of vertices. he efficiency is measured with a tag-and-probe procedure in a sample of Z! ee events. he efficiency is completely stable in the barrel and decreases by about 1% in the endcaps. Once again, a spectacular data-simulation agreement is obtained. 8

he CMS Particle Flow Algorithm Figure 1: Efficiency of the isolation requirement for identified electrons as obtained with a tag-and-probe method with Z! ee events in the barrel (left) and in the endcaps (right). 4. Conclusion he CMS detector is well suited for an efficient PF algorithm. he PF has been deployed in CMS and is currently used in most of the analyses. Even though the concept is simple, the implementation required some work. he performance of jets, E miss reconstruction, lepton reconstruction and identification have been significantly improved. Most of the key ingredients of the algorithm have been commissioned with the very first data, in a short timescale, demonstrating the accuracy of the CMS simulation from the beginning. he Particle Flow has not only proven to work well with high pile-up conditions, it appears as the only way to preserve and even improve the performance of CMS for the next rounds of data-taking. Acknowledgments I would like to thank the organizers for the very pleasant and interesting conference. References [1] S. Chatrchyan et al. he CMS experiment at the CERN LHC. JINS, 3:S84, 8. [] CMS Collaboration. Particle-Flow Event Reconstruction in CMS and Performance for Jets, aus and E miss. (CMS-PAS-PF-9), 9. [3] Adam, Wolfgang and Frühwirth, R. and Strandlie, Are and odor,. Reconstruction of Electrons with the Gaussian-Sum Filter in the CMS racker at the LHC. 5. [4] CMS Collaboration. Commissioning of the Particle-flow Event Reconstruction with the first LHC collisions recorded in the CMS detector. (CMS-PAS-PF), 1. [5] CMS Collaboration. Commissioning of the Particle-Flow reconstruction in Minimum-Bias and Jet Events from pp Collisions at 7 ev. (CMS-PAS-PF-), 1. 9

he CMS Particle Flow Algorithm [6] CMS Collaboration. Particle-flow commissioning with muons and electrons from J/Psi and W events at 7 ev. (CMS-PAS-PF-3), 1. [7] CMS Collaboration. Jet Energy Corrections and Uncertainties. Detector Performance Plots for 1. (CMS-DP), 1. [8] CMS Collaboration. ME performance in 8 ev data. (CMS-PAS-JME-), 13. [9] S. Chatrchyan et al. Observation of a new boson at a mass of 15 GeV with the CMS experiment at the LHC. Phys.Lett., B716:3 61, 1. [1] Photon ID performance with 19.6 fb 1 of data collected at p s = 8 ev with the CMS detector. (CMS-DP3), 13. 1