Muons in ATLAS Overview Muon physics Muon spectrometer

Muons in ATLAS Overview Muon physics Muon spectrometer Muon trigger Muon reconstruction Muon validation www. slac. stanford. edu/~cla/Leptons/Muons. html 5 March 07 Claus Horn (SLAC): Muon Overview 1

Muon Physics High p. T muons appear in many interesting physics signatures: q SM Higgs q MSSM Higgs, A->mm q Z’, W’ q b decays q t decays H->ZZ*->mmmm Needs: high trigger efficiency and good momentum resolution Different types of muons: q isolated (high p. T/low p. T) q in jets q cosmics q halos 5 March 07 3 -10 Ge. V muons are important for b-physics Claus Horn (SLAC): Muon Overview 2

Muon cross section @ Ecm=14 Te. V Muon rates @ 10 -34 cm-2 s-1: 30 k. Hz for p. T>8 Ge. V 1 k. Hz for p. T>20 Ge. V Event rate after trigger: ~100 Hz p. T<10 Ge. V: pion/Kaon decays dominate p. T>10 Ge. V: b, c decays dominate 5 March 07 Claus Horn (SLAC): Muon Overview 3

η=1. 4 η=2. 0 η=1. 05 Muon Spectrometer Barrel: |h|<1 Transition region: 1<|h|<1. 4 End-caps: |h|<2. 7 η=2. 4 η=3. 0 Three layer structure both for barrel and end-caps (The EE(=extra end-cap) ring was canceled) 22 m x 42 m 5 March 07 Claus Horn (SLAC): Muon Overview 4

Magnetic Field q Inner detector: solenoid field of Bz~2 T q Muon system: toroidal field to measure muon trajectories: q 8 superconducting coils in barrel and each end-cap q main field component: Bf~0. 5 T Magnetic field lines in x-y: Bending power: • Increases with eta • Complicated transition region 5 March 07 Claus Horn (SLAC): Muon Overview 5

Tracking detectors: Monitored Drift Tube chamber (MDT) § Three stations, resolution: 50 mm Cathode Strip Chamber (CSC) § replace MDTs in inner station for 2<h<2. 7 (better rate capacity and radiation harder) Trigger system: Resistive Plate Chamber (RPC) § At two outer stations of the barrel § Time resolution: <2 ns Thin Gap Chamber (TGC) § For end-cap triggers h<2. 4 Low p. T trigger: p. T> 6 Ge. V High p. T trigger: p. T>20 Ge. V Number of readout channels: 400 k for tracking, 900 k for trigger 5 March 07 Claus Horn (SLAC): Muon Overview 6

MDT chamber: q Two separated multilayer of drift tubes q Tube diameter: 3 cm Number of chambers: 1194 Track reconstruction: 1. Segment finding 2. Fitting min c 2 Chamber naming: B=barrel, E=end-cap I=inner, M=middle, O=outer S=small, L=large; F=feet, R=rail and: BOG+BOH=BOF p. T measurement from sagitta in magnetic field: 5 March 07 Claus Horn (SLAC): Muon Overview 7

Muon Triggering Reconstruction algorithms were adapted to run in HLT on ROIs. Full trigger chain: 5 March 07 Claus Horn (SLAC): Muon Overview 8

Low p. T trigger turn on at 6 Ge. V Event Filter output rates: Total rate for typical scenarios: mu(6): 3000 Hz mu(20): 25 Hz (by Stefania Spangolo) 5 March 07 Claus Horn (SLAC): Muon Overview 9

Muon Reconstruction Two containers in AOD: Staco. Muon. Collection Muid. Muon. Collection Using muon system only: Combined with Inner Detector: Low p. T algorithm: (muons don't reach outer stations) Muonboy MOORE STACO Mu. ID Mu. Tag Mu. IDLow. Pt New: Mu. Girl: started as independent low p. T algorithm, now extending to become a complete id tool Finally one should be chosen, maybe combination ? 5 March 07 Claus Horn (SLAC): Muon Overview 10

STACO / Muonboy / Mu. Tag Muonboy: 3 D pattern recognition, proceeding outwards Oldest algorithm, started in Fortran, now converted to C++ STACO (STAtistical COmbination) principle: Combined In. Det & Mu. Sp parameter fitting using covariance matices Pi: track parameter vectors Ci: their covariance matrices Mu. Tag: Instead of sagitta, extrapolate In. Det tracks to medium station and compare to hit. Hole closed now! missing 5 March 07 Claus Horn (SLAC): Muon Overview 11

MOORE / Mu. ID MOORE (Muon Object Oriented REconstruction) q Pattern recognition from regions of activity in MS q Parameters given at first measured point Mu. ID (Muon IDentification): q using Mu. Sp, In. Det and calorimeter q track matching with c 2< cut-off q parameters given at interaction point Performance: (by Steffaia Spangolo) 5 March 07 Claus Horn (SLAC): Muon Overview 12

More MOORE examples MOOR pull for 1/p. T: (by S. Willocw) Moore/Muid performance against eta for Z->mm: 5 March 07 Claus Horn (SLAC): Muon Overview 13

Mu. Girl Procedure: Start with ID -> extrapolate to MS -> find segments ANN trained against cavern background using variables: q # hits (MDT, TCG, RPC) q # segments q # hits/segment Number of MDT segments per track Performance: 5 March 07 Claus Horn (SLAC): Muon Overview 14

Muon Tracks 4 types of tracks: q in. Det. Track. Particle n q muon. Spectrometer. Track. Particle n n q From standalone muon finding (Mu. Sp only) Parameters at entrance to the muon system muon. Extrapolated. Track. Particle n q Inner detector track e. g. from low-p. T algorithm (In. Det only) Previous extrapolated to interaction point combined. Muon. Track. Particle n Combination of inner and standalone muon tracks (exist for both algorithm containers) Kinematics and quality cuts: q Number of hits q p. T > 3 Ge. V, |z 0|< 5 March 07 q c 2/dof q isolation Claus Horn (SLAC): Muon Overview 15

Momentum resolution From single muon simulation: Inner Detector Resolution in Mu. Sp only Muon Spectrometer combined (by Eric Lancon) (from TDR) Influence of track quality on p. T: 5 March 07 Claus Horn (SLAC): Muon Overview (by Benjamin Ruckert) 16

Calorimeter energy loss Standard approach uses parameterization in (f, h) Description worsens towards higher energies New approach: use measured energy in each event (only used in high energy tail) Performance: (by Eric Lancon) Procedure: 1. Start from track extrapolated to first calorimeter layer 2. Sum up energy deposited in cone 3. Do this for each layer Multiple scattering parameterized at each layer 5 March 07 Claus Horn (SLAC): Muon Overview (by K. Bachas) 17

Calorimeter based muon id Use pattern of energy deposition for m/p seperation. Use cases: Tagging muons (isolated / second muon in B jets) Cross check muon system Derive efficiency from data 9 different regions (in h and energy) with specific set of variables for likelihood, e. g. : Result for H->4 m sample: (by L. R. Flores-Castillo) 5 March 07 Claus Horn (SLAC): Muon Overview 18

Muon Validation Activities: q Check supostious effects event by event q Compare hits with extrapolation from previous station q Pile-up events for: tuning, evaluation, developments q Check for software bugs MC truth muons Reconstructed muons matching found, lost, fake Found Fakes Lost (by David Adams) 5 March 07 Claus Horn (SLAC): Muon Overview (by N Benekos) 19

Comparing muon algorithms for different physics samples Package by David Adams Classification into: Efficinecy q q Combined: Mu. Sp & In. Det tracks with good fit Split: combined with bad fit Standalone: Mu. Sp tracks propagated to ip. Inner: only In. Det track different processes Momentum resolution (by David Adams) 5 March 07 Both STACO and Mu. ID show good performance Efficiency 80 -95% Fakes around 0. 01/event Working less well in high-p. T jets (STACO 86%, Mu. ID 73%) Claus Horn (SLAC): Muon Overview 20

Muon efficiency and resolution from data Tag and probe method: Z->m+m. Cut on invariant mass and Df Tag muon requires track in In. Det and Mu. Sp (by Matthias Schott) 2% stat. precision with 30 pb-1! System. error: 0. 23%+0. 25% 5 March 07 Claus Horn (SLAC): Muon Overview 21

Physics Studies Processes under investigation: micro black holes decay isotropically to a lage number of high p. T muons 5 March 07 Claus Horn (SLAC): Muon Overview 22

Muon event Visualization with PERSINT readable output can be produced with athena job. Options flag from ROD, AOD and ESD. Can display: q Muon containers (standalone & combined) q Calorimeter objects (Calo. Cells, Egamma) q Inner detector tracks q MC truth Matching of ID and MS tracks: STACO track 5 March 07 Claus Horn (SLAC): Muon Overview 23

Work in progress • Fix BOG MDT shift digitization bug • Check for Moore efficiency drop near h=2. 0 • Material map need optimization for low p. T muons • Restructuring of reconstruction software • Use NN for calorimeter based muon id • Improve magnetic field description 5 March 07 Claus Horn (SLAC): Muon Overview 24

Conclusions q Muons are important for new discoveries q ATLAS has a powerful muon spectrometer q Many different reconstruction algorithms with good performance q Interesting new projects www. slac. stanford. edu/~cla/Leptons/Muons. html 5 March 07 Claus Horn (SLAC): Muon Overview 25

Additional Information 5 March 07 Claus Horn (SLAC): Muon Overview 26

Big Wheel RIBS 5 March 07 Claus Horn (SLAC): Muon Overview 27

5 March 07 Claus Horn (SLAC): Muon Overview 28

Energy loss after Bethe and Bloch: Material in front of muon system: 5 March 07 Claus Horn (SLAC): Muon Overview 29

Another PERSINT example Get detailed info by clicking on: Spectrometer tracks q ID tracks q Calorimeter deposits (Tile/LArg) q Hit positions q 5 March 07 Claus Horn (SLAC): Muon Overview 30

References Benjamin Ruckert (LMU Muenchen): Muon reconstruction and the search for LQ, 22 March 06 Eric Lancon (Saclay): Muon combined performance, 22 March 06 Matthias Schott (U. Munich): Tag and Probe, May 06 David Adams (BNL): ATLAS muon system performance, 29 th May 06 N. Benekos (MPL): Muon Sftware Validation Algorithms, Okt. 06 S. Willocq (UMass): Moore and Muid status, 30 th May 2006 L. R. Flores-Castillo (UWisconsin): Calorimeter based muon id, 30 th May 2006 5 March 07 Claus Horn (SLAC): Muon Overview 31