The Forward Liquid Argon Calorimeter of the ATLAS Detector Geant4 Workshop' September. Triumf, Vancouver Patricia Méndez Lorenzo. CERN EP/SFT 1. Introduction and Generalities 2. FCAL at TestBeam 3. Simulation of FCAL with Geant4 4. Results of the Simulation 5. Summary 1
Introduction and Generalities Physics to be done at the LHC 1. Origin of the mass at the electroweak scale; search of Higgs boson ► H γγ for 90<M H <150 GeV ► H ZZ 4l for 130<M H <M Z ► WW, ZZ fusion with H ZZ, WW l,νν, 2jets for M H up 1 TeV ► pp WH, ZH, ttbar H, H bbbar (Excluded by LEP) for 80<M H <100 GeV Calorimetry plays a central role on subdetectors design at LHC 2. Deep study of the top and botton quarks 3. Search of new physics: SUSY, technicolours, extra-dimensions.... 2
General ATLAS Calorimeter System Goal: High precision measurements of energy and position of electrons, photons, jets and missing E T Requirements: ◘ Rapidity coverage ◘ Good electron reconstruction ◘ Excellent energy resolution ◘ Accurate measurement of the shower position Main Elements: ◙ One central cryostat barrel and 2T super-conducting solenoid ◙ Two endcaps, in each one: one electromagnetic, two hadronic wheels and one forward calorimeter Hadronic Tile Calorimeters EM Accordion Calorimeters Hadronic lAr EndCap Calorimeters Forward lAr Calorimeters 3
The ATLAS Forward lAr Calorimeter (FCAL) Goal of the FCAL Full coverage of the system for a good determination of the missing energy (undetected particles as neutrinos) Design ► 1 Electromagnetic copper/liquid argon module (FCAL1) ► 2 Hadronic modules (FCAL2, FCAL3) with tungsten as absorber Construction ► Several prototypes for FCAL1 constructed and tested successfully at Brookhaven and CERN ► Full depth pre-production prototypes for FCAL1 and FCAL2 (Module 0) designed and built in Arizona and Canada in 1998 Barrel EndCap EM Accordion Calorimeter Had EndCap Calorimeter FCAL Calorimeter 4
The FCAL at the TestBeam FCAL1 and FCAL2 prototypes (Module 0) have been tested at CERN in 1998 during a testbeam program with electrons, pions and muons ♦ Detectors and elements included in the TestBeam Setup included in the Simulation ♦ Description of FCAL module as close as possible to real detectors FCAL Module 0 test beam in Geant4 Particles enter the setup from the lower left corner FCAL1 Module 0 FCAL2 Module 0 Cryostat 5
Simulation of FCAL Module 0 with Geant4 Important parts: Geometry Cryostat EM Module and Had Modules TestBeam SetUp Physics Physical Processes Input Particles: electrons Visualization and histogramming Visualization packages Anaphe and Aida Simulation based on TestBeam'98 data included as a Geant4 advanced example called lAr_calorimeter 6
Simulation of FCAL Module 0 with Geant4 Detector Geometry Part int main(int argc, char** argv){ G4RunManager* runManager = new G4RunManager; FCALTestbeamSetup* detector = new FCALTestbeamSetUp; runManager ->SetUserInitialization(detector); FCALCryostatVolumes FCALTestbeamSetUpSD FCALEMModule (including EMParameters.input) FCALHadModule (including HadParameters.input) FCALEMModuleSD FCALHadModuleSD 7
Simulation of FCAL Module 0 with Geant4 Standard Advanced Examples Design for Definition of Particles and Physical Processes Classes ► FCALPhysicsList : Standard Particles and Physical Processes defined ► FCALPrimaryGeneratorAction: Definition of X, Y, Z, cos_X, cos_Y, cos_Z for electrons inside tracks_20GeV.dat, tracks_40GeV.dat, tracks_80GeV.dat and tracks_200GeV.dat in files of 1000 events each. particleRun ->SetParticlePosition particleRun ->SetParticleMomentumDirection Defined following parameter definitions inside those files Standard advanced examples design for the rest of classes 8
Simulation of FCAL Module 0 with Geant4 Visualization Development with OpenGL All visualization packages included in FCALVisManager class Initialization of the drawing via: (interactively) idle> /control/execute vis.mac Front View of the System 9
Simulation of FCAL Module 0 with Geant Anaphe Implementation of the Aida 3.0 Histograms MakeFile as simple as possible ifdef G4ANALYSIS_USE CPPFLAGS += 'aida-config --include' LDFLAGS += 'aida-config --lib' Sourcing some startup scripts at CERN setenv G4ANALYSYS_USE 1 setenv PATH ${PATH}: /afs/cern.ch/sw/lhcxx/share/LHCXX/latest/scripts source /afs/cern.ch/sw/lhcxx/share/LHCXX/latest/scripts ln -s /afs/cern.ch/sw/lhcxx/share/LHCXX/latest/scripts/* ~/bin/. (just once after the source) The example has been tested using the gcc compiler inside CERN RedHat 6.1 and
Results of the Simulation Histograms and Ntuples The example produces 6 histograms and ntuples saved in “fcal.his ” Group of histos and ntuples to show the AIDA mechanismus: (included in last release) Histo1 and Ntuple1 —— > Number of Tracks out of World Histo2 and Ntuple2 —— > Number of Secondary Particles Histo3 and Ntuple3 —— > Deposited Energy in FCAL1 Histo4 and Ntuple4 ——> Deposited Energy in FCAL2 Group of histos and ntuples to compare with data of 1998 test beam: (to include in next release) Histo5 and Ntuple5 ——> Reconstructed Energy in FCAL1 Histo6 and Ntuple6 —— > Reconstructed Energy in FCAL2 11
Results of the Simulation Comparison between Geant4 and TestBeam Data'98 Characteristics of the simulation: ♣ Variable to study: Energy deposited in FCAL1 and FCAL2 in the liquid argon of the tube electrode summed into tiles ♣ TestBeam data: ntuples of 5000 events each for 20, 40, 60 and 80 GeV ♣ Initial Simulation Conditions: 1000 electrons are jetted at beam energies of 20, 40, 60 and 80 GeV ♣ Range cut for secondary particles production: 1 mm ♣ Important effects to take into account: noise and digitization has to be parameterized inside the simulation 12
Results of the Simulation Definition of the variable inside the Simulation E tile rec = c MC x E tile vis + E noise (defined by event) E rec (per event) = Σ E tile rec c MC = Monte Carlo electron calibration constant: 1/C MC = / Cuts 0.5mm 1.0mm 2.0mm c MC [%] E tile vis = row electron signal recollected in each tile of the module E dep = Total energy deposited inside the module: E dep = E beam – E loss E loss = Energy losses outside the module E noise = correction factor calculated by gaussian smearing centered at 0 and a width of 8 GeV 13
Results of the Simulation 14 Statistical Test (Thanks to statistics testing team) Code design to compare Data with Simulation by a p-value test: Probability that both histos are compatible Test performed for FCAL1 Reconstructed Energy to 20, 40 and 60 GeV
Results of the Simulation 15
Results of the Simulation 16 Chi2 ≈ 300 n.d.f< 90 p-value quite small p-value very small due to an understimation of the statistical error from test beam data Quite important for these tests a good error estimation Further factors to take into accout: 1. Low energy electrons not considered into simulation 2. Constant smearing factor for all events into simulation
Summary 17 ♠ A simulation for the module 0 of the lAr calorimeter for the ATLAS detector has been successfully performed with the Geant4 toolkit ♠ Test beam'98 setup included in the simulation ♠ Results included in a Geant4 advanced example ♣ The statistical test results here presented are preliminary, still to be tested ♣ Only FCAL1 reconstructed energy has been checked in the comparison, further data variables have to be included in the whole analysis ♣ 2003 summer data must be tested in the simulation