Measurement of muon misidentification rates in Z events for the - PowerPoint PPT Presentation

Measurement of muon misidentification rates in Z → µµ events for the ATLAS detector Johannes Mellenthin IMPRS Young Scientists’ Workshop, Ringberg 2013 23.07.2013 Johannes Mellenthin (Ringberg 2013) Muon misidentification rates 23.07.2013 1 / 21

Outline Muon reconstruction with the ATLAS detector 1 Description of the measurement methods 2 Discussion of results 3 Conclusions & Outlook 4 Johannes Mellenthin (Ringberg 2013) Muon misidentification rates 23.07.2013 2 / 21

The ATLAS detector Multipurpose detector at the LHC at CERN (Geneva) Record proton proton collisions at √ s = 7 TeV (2011) and 8 TeV (2012) After 2015 up to 13 TeV High instantaneous luminosity 7 · 10 33 cm − 2 s − 1 , after upgrade 1 · 10 34 cm − 2 s − 1 3 main components inner detector calorimeters muon spectrometer p Tile calorimeters LAr hadronic end-cap and forward calorimeters LAr electromagnetic calorimeters Toroid magnets Pixel detector Calorimeters Transition radiation tracker Muon chambers Solenoid magnet Muon spectrometer Semiconductor tracker |η| < 2.7 Inner tracking detector |η| < 2.5 acceptance gap in |η|<0.1 Johannes Mellenthin (Ringberg 2013) Muon misidentification rates 23.07.2013 1 / 21

Muon reconstruction with ATLAS Use of all subdetectors Inner detector: tracking & momentum measurement with a solenoid magnetic field Calorimeter: isolation and energy loss Muon spectrometer: tracking & momentum measurement with a toroidal magnetic field & muon identification Performance goal: Momentum measurement with a 10 % accuracy for 1 TeV Muons p Tile calorimeters LAr hadronic end-cap and forward calorimeters LAr electromagnetic calorimeters Toroid magnets Pixel detector Calorimeters Transition radiation tracker Muon chambers Solenoid magnet Muon spectrometer Semiconductor tracker |η| < 2.7 Inner tracking detector |η| < 2.5 acceptance gap in |η|<0.1 Johannes Mellenthin (Ringberg 2013) Muon misidentification rates 23.07.2013 2 / 21

Muon reconstruction Ideal case: combination of two tracks in the inner detector and the muon spectrometer Combined muons no MS track (acceptance gap) 95 % of all cases minimum deposit calorimeter muon in all calo layers Highest resolution and purity MS track ID track no ID track (|η| > 2.5) All subdetectors need to be ID track standalone muon combined muon instrumented and operational MS track Acceptance loss in uninstrumented regions of the muon spectrometer Johannes Mellenthin (Ringberg 2013) Muon misidentification rates 23.07.2013 3 / 21

Muon reconstruction Recover efficiency in incompletely instrumented regions through additional algorithms → Combination of inner detector track with spectrometer hits that don’t form an independent track ( segment tagged muons ) → Muon spectrometer track with no associated inner detector track ( standalone muon ) → Combination of inner detector track with minimum energy deposit in the calorimeter ( calorimeter muon ) Efficiency gained at the price of reduced purity Johannes Mellenthin (Ringberg 2013) Muon misidentification rates 23.07.2013 4 / 21

Muon reconstruction Efficiency of combination of combined and segment-tagged muons Johannes Mellenthin (Ringberg 2013) Muon misidentification rates 23.07.2013 5 / 21

Muon background in physics analysis Real muons out of heavy flavour decays → b-lifetime: Muon displaced from primary vertex → Surrounding jet activity in the calorimeter Real muons out of pion / kaon decays → Characteristic kink in the track at pion / kaon decay point Fake muons from jets (punch-through) → Recognizable through large energy deposit in the calorimeter Cosmic muons → Tracks do not emerge from primary vertex Johannes Mellenthin (Ringberg 2013) Muon misidentification rates 23.07.2013 6 / 21

Example: H → ZZ ∗ → 4 µ High muon efficiency required for 4 µ final state → Combine all previously mentioned reconstruction methods to obtain maximum efficiency Main background ZZ diboson production → Low below m 4 µ ∼ 180 GeV But: Processes with non-prompt muons become important → t ¯ t and Z + jets → Understanding of these muons important for measuring the new Higgs properties Johannes Mellenthin (Ringberg 2013) Muon misidentification rates 23.07.2013 7 / 21

Analysis strategy Goal: Study appearance of non-prompt muon background → Validate the prediction of the detector simulation Observe background-like muons in a well controlled environment Use Z → µµ decays well known physics process → Know that any additional muons must be non-prompt High statistics at LHC Easy to select with high purity Comparable environment to Higgs search Johannes Mellenthin (Ringberg 2013) Muon misidentification rates 23.07.2013 8 / 21

Analysis strategy counts / 0.5 GeV 6 10 ATLAS Work in Progress Selection of Z → µµ samples 1 5 10 Require presence of two reconstructed Data Zjets 4 10 t t WZ combined muons WW ZZ 3 10 from the primary vertex no surrounding jet activity 2 10 Require opposite muon charges 10 Require invariant dimuon mass within 10 Data/MC 82 84 86 88 90 92 94 96 98 100 1.05 GeV of the Z mass 1 82 84 86 88 90 92 94 96 98 100 M [GeV] ll Johannes Mellenthin (Ringberg 2013) Muon misidentification rates 23.07.2013 9 / 21

Analysis strategy Selection of Z → µµ samples 1 Collect background candidates 2 In selected events look for presence of objects that could give rise to non-prompt muons For the contribution of pion / kaon decays use inner detector tracks any track with transverse momentum above 10 GeV For the contribution of secondary muons from heavy flavor decays and punch-through use jets any reconstructed jets above 20 GeV Use all events that have at least one such candidate for further study Johannes Mellenthin (Ringberg 2013) Muon misidentification rates 23.07.2013 10 / 21

Analysis strategy Selection of Z → µµ samples 1 Collect background candidates 2 Look for reconstructed muons matching the background candidates 3 Exclude muons from the Z → µµ candidates Study fraction as function of several observables transverse momentum, pseudorapidity Activity surrounding the muon Muon impact parameter Processes with additional prompt muons: estimate using MC simulation Plot the fake rates as the fraction of background candidates with a matching muon Johannes Mellenthin (Ringberg 2013) Muon misidentification rates 23.07.2013 11 / 21

Combined and segment tagged Fake rates Fake rates as a function of transverse momentum Order of magnitude below 1 % in both cases Excellent agreement between simulation and data Jets: increase with p T (probability of emitting a muon with sufficient momentum or punch-through) Tracks: Maximum at 40 GeV decrease to lower and higher momentum Johannes Mellenthin (Ringberg 2013) Muon misidentification rates 23.07.2013 12 / 21

Calorimeter tagged Fake rates Fake rates as a function of transverse momentum Fake rates for tracks much higher Jets: similar to combined muons up to 40 GeV, then decrease due to reconstruction level requirements → High energy deposit in calorimeter prevents identification as calorimeter muon Simulation underestimates the fake rates for jets Tracks: increase with p T Johannes Mellenthin (Ringberg 2013) Muon misidentification rates 23.07.2013 13 / 21

Possible reduction of Fake rates Muon isolation Require low activity in an angular cone around the muon Prompt muons: no surrounding activity, high efficiency Non-prompt muons: often part of jets, high rejection counts / 0.5 GeV 6 10 ATLAS Work in Progress Data 5 10 Prompt muons non-Prompt muons 4 10 3 10 2 10 10 0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5 cone30 E / p T T Johannes Mellenthin (Ringberg 2013) Muon misidentification rates 23.07.2013 14 / 21

Effect of Isolation on Combined and segment-tagged muons Fake rates as a function of transverse momentum -1 10 fake rates fake rates ATLAS Work in Progress 0.014 ATLAS Work in Progress Data Simulation -2 10 0.012 Data isolation Simulation isolation -3 Data Simulation 10 0.01 0.008 Data isolation Simulation isolation -4 10 0.006 -5 10 0.004 -6 10 0.002 -7 0 10 20 40 60 80 100 120 20 40 60 80 100 120 jet p [GeV] track p [GeV] T T Jets: very strong reduction (Factor 10 - 100) Jet includes extra activity by definition Tracks: Strong reduction (Factor 2 - 10) Tracks: increase with p T Some tracks may not be part of jets High p T contamination by prompt muons ( WZ → 3 µ ν ) Johannes Mellenthin (Ringberg 2013) Muon misidentification rates 23.07.2013 15 / 21

Effect of Isolation on Calorimeter tagged muons Fake rates as a function of transverse momentum -1 0.1 10 fake rates fake rates ATLAS Work in Progress Data Simulation Data Simulation 0.09 -2 10 0.08 Data isolation Simulation isolation Data isolation Simulation isolation 0.07 -3 ATLAS Work in Progress 10 0.06 -4 0.05 10 0.04 -5 10 0.03 0.02 -6 10 0.01 -7 0 10 20 40 60 80 100 120 20 40 60 80 100 120 jet p [GeV] track p [GeV] T T Reduction not as strong as for combined muons All calorimeter muons have to pass a loose isolation cut at reconstruction → Further reduction not as strong Still noticeable reduction (Jets: Factor 5 - 100, Tracks Factor 2) Johannes Mellenthin (Ringberg 2013) Muon misidentification rates 23.07.2013 16 / 21