The performance of the ATLAS Initial detector layout for B-physics channels B. Epp, V.M. Ghete, D. Kuhn, Y.J. Zhang Institute for Experimental Physics University of Innsbruck, Austria ¨ OPG FAKT, Weyer, September 2004
Introduction Purpose of this studies: • Inner Detector(ID) performance w.r.t. quantities relevant to B 0 s → D s π analysis ( D s → φπ , φ → K + K − ) • Validation of software in the ATLAS framework ATHENA • Validation of generation-simulation-reconstruction-analysis chain for: – Initial layout – Complete layout Various comparisons were performed: • one-step comparisons: change one step in the software chain and perform – comparison of different layouts – comparison of different versions of the reconstruction software • multi-step comparisons: more steps were different ¨ V . B . Epp ATLAS initial detector layout performance OPG FAKT, Sept 2004
ATLAS Full Detector Inner Detector Electromagnetic and Hadronic Calorimeters Muon Detectors ¨ V . B . Epp ATLAS initial detector layout performance OPG FAKT, Sept 2004
ATLAS Inner Detector ID Inner Detector (ID): • cylinder with barrel and end- cap parts • discrete semiconductor pixel and strip detectors • transition radiation tracker: straw-tubes interspersed with a radiator = e/π ⇒ separation • inside solenoid: 2 T magn. field Staged detector: = ⇒ Initial layout versus Complete Layout Detector part barrel end-cap Initial / Complete cylinders disks on each side Pixel Detectors 2 / 3 2 / 3 Semi-Conductor Tracker (SCT) 4 / 4 9 / 9 Transition Radiation Tracker (TRT) axial tubes C-Wheels: no/yes ¨ V . B . Epp ATLAS initial detector layout performance OPG FAKT, Sept 2004
Software chain Software step Software release Event generator Pythia 6.205, with B-physics tuning 50k events Athena framework, software release 5.0.0 Detector simulation full detector: initial layout complete layout atlsim framework, software release 6.0.2 (via AFS) realistic, non-uniform magnetic field Pileup simulation atlsim framework, software release 7.0.2 ∼ L= 2 × 10 33 cm − 2 s − 1 min bias events with muon safety factor = 1 Reconstruction package xKalman++ with non-uniform magnetic field atrecon framework, software release 6.6.0 Athena, software release release 7.0.0 Analysis Combined Ntuples CTVMFT vertex fit package (uniform magnetic field) ¨ V . B . Epp ATLAS initial detector layout performance OPG FAKT, Sept 2004
Definitions of quantities • φ , D s , B 0 s mass resolutions and peak positions: = ⇒ Fit core with a Gaussian function • B 0 s d xy transverse decay length: distance between interaction point and reconstructed B 0 s decay vertex projected onto the transverse plane = ⇒ Fit sum of two Gaussian functions, same mean value • B 0 s g -factor fractional resolution: ( g − g 0 ) /g 0 c · p B 0 B 0 � � � � g = M B 0 s / , g 0 = M B 0 s / c · p T s s T true = ⇒ Fit with a Gaussian function • B 0 s proper time: t = d xy g proper time resolution t − t 0 = ⇒ Fit sum of two Gaussian functions, same mean value Fits with sum of two Gaussian functions: For comparison, force for second sample the ratio from the first sample. ¨ V . B . Epp ATLAS initial detector layout performance OPG FAKT, Sept 2004
Initial vs Complete layout: Atrecon rec., no pile-up Particle Characteristic Initial layout Complete layout Nominal mass (MeV) 1019.4 1019.4 φ Peak position (MeV) 1019 . 6 ± 0 . 0 1019 . 6 ± 0 . 0 Resolution (MeV) 3 . 8 ± 0 . 0 3 . 8 ± 0 . 0 Nominal mass (MeV) 1968.5 1968.5 D s Peak position (MeV) 1969 . 9 ± 0 . 1 1969 . 3 ± 0 . 1 Resolution (MeV) 15 . 4 ± 0 . 1 15 . 6 ± 0 . 1 Nominal mass (MeV) 5369.3 5369.3 B 0 Peak position (MeV) 5369 . 3 ± 0 . 7 5367 . 4 ± 0 . 7 s Resolution (MeV) 46 . 0 ± 0 . 6 46 . 4 ± 0 . 6 400 Entries / 0.01 GeV Initial Complete No deterioration of the mass resolutions for 300 the Initial layout w.r.t. Complete layout is seen. 200 100 0 5.2 5.4 5.6 M KK ππ [ GeV ] ¨ V . B . Epp ATLAS initial detector layout performance OPG FAKT, Sept 2004
Initial vs Complete layout: Atrecon rec., no pile-up Distribution Characteristic Initial layout Complete layout Peak position ( µ m ) − 0 . 8 ± 1 . 5 1 . 5 ± 1 . 4 Fraction G 1 (%) 72 . 4 ± 2 . 2 idem d xy − d 0 Fraction G 2 (%) 27 . 6 ± 2 . 2 idem xy Resolution G 1 ( µ m ) 84 . 3 ± 2 . 4 86 . 4 ± 1 . 5 Resolution G 2 ( µ m ) 212 . 7 ± 9 . 1 195 . 9 ± 4 . 1 Peak position (fs) − 1 . 0 ± 1 . 3 0 . 4 ± 1 . 3 Fraction G 1 (%) 63 . 1 ± 3 . 7 idem t − t 0 Fraction G 2 (%) 36 . 9 ± 3 . 7 idem Resolution G 1 (fs) 69 . 8 ± 3 . 2 73 . 2 ± 1 . 4 Resolution G 2 (fs) 157 . 5 ± 9 . 4 153 . 1 ± 2 . 9 400 Entries / 0.02 ps Initial Complete A slight tendency toward broader cores for 300 Complete (more dead material) and larger tails for Initial (missing pixel layer) 200 layout can be seen. 100 0 -0.4 -0.2 0 0.2 0.4 t-t 0 [ ps ] ¨ V . B . Epp ATLAS initial detector layout performance OPG FAKT, Sept 2004
Initial vs Complete layout: Atrecon rec., no pile-up B 0 s → D s π Initial layout Complete layout Initial/Complete Simulated events 50000 50000 Reconstructed events σ = 46 . 0 MeV σ = 46 . 4 MeV 3 σ 5901 6269 2 σ 5542 5865 1 σ 4016 4270 Reconstructed events / 10 fb − 1 3 σ 3324 3547 93.7% 2 σ 3122 3304 94.5% 1 σ 2262 2405 94.1% 6% of events are lost for the Initial layout w.r.t. the Complete layout. Differences are due to the track reconstruction. Overall: no significant deterioration of quantities for Initial layout w.r.t Complete layout. ¨ V . B . Epp ATLAS initial detector layout performance OPG FAKT, Sept 2004
Initial layout: Atrecon versus Athena rec., no pile-up B-physics conditions require a non-default tracking strategy for xKalman pattern recognition: • typical B-decay tracks at low p T • missing second pixel layer in Initial layout = ⇒ pattern recognition fails or finds track with wrong hit in b-layer Parameter settings for alternate xKalman tracking strategy: • silicon clusters per track ≥ 7 • no cut on hits in pixel layers (including b-layer) • minimum number of no-shared hits per track = 4 • selectivity of reconstruction = 4 • minimum number of TRT clusters per track = 9 • maximum number of holes = 22 ¨ V . B . Epp ATLAS initial detector layout performance OPG FAKT, Sept 2004
Initial layout: Atrecon versus Athena rec., no pile-up Distribution Characteristic Atrecon Athena Peak position ( µ m ) − 0 . 8 ± 1 . 5 − 2 . 1 ± 1 . 6 Fraction G 1 (%) 72 . 4 ± 2 . 2 idem d xy − d 0 Fraction G 2 (%) 27 . 6 ± 2 . 2 idem xy Resolution G 1 ( µ m ) 84 . 3 ± 2 . 4 92 . 4 ± 1 . 6 Resolution G 2 ( µ m ) 212 . 7 ± 9 . 1 228 . 6 ± 4 . 7 ( g − g 0 ) /g 0 Peak pos. (%) 0 . 007 ± 0 . 011 − 0 . 013 ± 0 . 010 Resolution (%) 0 . 80 ± 0 . 01 0 . 81 ± 0 . 01 Peak position (fs) − 1 . 0 ± 1 . 3 − 1 . 3 ± 1 . 4 Fraction G 1 (%) 63 . 1 ± 3 . 7 idem t − t 0 Fraction G 2 (%) 36 . 9 ± 3 . 7 idem Resolution G 1 (fs) 69 . 8 ± 3 . 2 76 . 6 ± 1 . 4 Resolution G 2 (fs) 157 . 5 ± 9 . 4 174 . 4 ± 3 . 3 Relatively large difference in d xy and proper time t resolutions: = ⇒ Athena worse by ∼ 10% than Atrecon. Explanation after cross check using Athena software inside Atrecon: predominantly due to different way of computing errors associated to pixel cluster position (simplified error calculation used in Athena) ¨ V . B . Epp ATLAS initial detector layout performance OPG FAKT, Sept 2004
Initial layout: Atrecon versus Athena rec., no pile-up B 0 s → D s π Atrecon Athena Atrecon/Athena Simulated events 50000 50000 Reconstructed events σ = 46 . 1 MeV σ = 46 . 4 MeV 3 σ 5901 6459 2 σ 5542 6038 1 σ 4016 4386 Reconstructed events / 10 fb − 1 3 σ 3324 3639 91.3% 2 σ 3122 3401 91.8% 1 σ 2262 2471 91.5% 400 Entries / 0.02 ps Atrecon rec. Athena rec. 9% less events in Atrecon, due to different way 300 of computing errors on the track parameters in Athena and Atrecon. 200 100 0 -0.4 -0.2 0 0.2 0.4 t-t 0 [ ps ] ¨ V . B . Epp ATLAS initial detector layout performance OPG FAKT, Sept 2004
Initial layout - pile-up versus no pile-up All samples were reconstructed with Athena only. An influence of pile-up on individual track reconstruction is seen for the Λ 0 b → J/ψ ( µ + µ − )Λ 0 ( pπ − ) channel: the fraction of reconstructed tracks with an incorrect hit in b-layer is increasing (especially in the low p T region). Λ 0 b : σ single Gauss fit no pile-up pile-up pile-up/no pile-up mass (MeV) 26 . 9 ± 0 . 6 28 . 4 ± 0 . 8 105 . 6% proper-time (fs) 108 . 1 ± 2 . 2 115 . 2 ± 2 . 6 106 . 6% N pile rec / N nopile 94 . 0% rec Broader resolutions and degradation of efficiency of ∼ 6% for pile-up events. ( Λ fails to match J/ψ vertex) Similar effects are not seen on B 0 s mass, d xy or t -resolutions or number of reconstructed events for B 0 s pile-up events (possible explanation: secondary particles decay before b-layer). ¨ V . B . Epp ATLAS initial detector layout performance OPG FAKT, Sept 2004
Recommend
More recommend
