1. LHCb Particle Identification and Performance
Paul Szczypka (on behalf of the LHCb Collaboration)

2. Introduction
LHCb is a forward arm spectrometer at the LHC and focuses on precise measurements of CP violation and rare decays of B-mesons.

3. Particle identification systems
The Rich Detectors, the Calorimeter Systems and Muon System all contribute to particle identification

4. Why do we need particle identification?
Typical measurement – invariant mass of B → ππ:
Signal decay
Signal and background topologically similar → require pion / kaon ID
Background
Invariant mass distribution of B → ππ decays including background

5. Other requirements for PID
RICH
Identify events of interest in the HLT
Reduce combinatronics in many-body events
CALO
Photon identification to reconstruct π0
MUON
Identify high pT muons for L0 trigger decision

6. Why two RICHes? aerogel RICH1 C4F10 RICH2 CF4
Chrenkov angle vs. particle momentum
Coverage of all particles from B → ππ decays
RICH1
RICH2
aerogel
C4F10
CF4
n = 1.03
n =
n =
Two RICHes and 3 radiators cover the particle phase space almost completely

7. RICH 1 Spherical Mirrors Secondary Flat Mirrors Magnetic shielding
Focus Cherenkov radiation
Mirror tilt removes photon detectors from detector acceptance
Secondary Flat Mirrors
Minimizes detector length
Magnetic shielding
Maintain bending power upstream of first (Trigger) Tracking station
Shield HPDs from fringe B-field
C4F10 and Aerogel Radiators
Covers 1 to 60 GeV/c
Acceptance of 25 – 300 mrad
RICH 1

8. RICH 2 CF4 Radiator Acceptance of 15 to 120 mrad
Spherical Mirror
Flat Mirror
Photon funnel+Shielding
Central Tube
Mirror Support Panel
Support Structure
CF4 Radiator
Acceptance of 15 to 120 mrad
Covers high momentum tracks from 20 to 100 GeV/c
Uses same photon detectors as RICH 1
HPDs sensitive to a single photon in the range λ = 200 to 600 nm
80mm
120mm

9. Hadron Identification with RICH
Kaon identification efficiency ~ 88%
Pion misidentification ~ 3%
Given a set of track ID hypotheses, the probability distribution for finding photons in each pixel of the detector is determined
PID likelihood formed from comparison with observed hit distribution
Particle ID hypotheses adjusted to maximize likelihood

10. Efficiency of π0 identification vs. pT (compared to MC truth)
Reconstruction of π0
Normal cluster
Efficiency of π0 identification vs. pT (compared to MC truth)
Merged π0
Resolved π0
Merged cluster
Photon candidates paired to reconstruct the resolved π0
Algorithm disentangles the photon pair from a merged cluster and reconstructs π0
Efficiency ~ 53%

11. Calorimeter – π0 reconstruction
Decay channel:
produces π0 with <pT> ≈ 3 GeV/c
Resolved π0 mass distribution
Merged π0 mass distribution
One γ reaches CALO
Both γ‘s reach CALO
B mass resolution: ~ 60 MeV/c2
Mass Resolution: ~10 MeV/c2
Mass Resolution: ~15 MeV/c2
π0 ID gives us a B0 mass of ~ 5.26 GeV/c2 with a resolution of ~ 60 MeV/c2

12. Resolution ~ 9 MeV/c2 (12 MeV/c2 downstream)
Muon Identification
Muons selected by searching for muon stations hits compatible with reconstructed track extrapolations
Compare track slopes and distance of muon station hits from track extrapolation to remove background
Global PID can be used to reduce the misidentification rate e.g. π ID from the RICHes
Mass dist for B0 → J/ψ(μμ)K0S
For p>3GeV/c
eff = 96.7  0.2 %, misid = 2.50  0.04 %
Resolution ~ 9 MeV/c2 (12 MeV/c2 downstream)

13. Particle ID is vital for the LHCb Physics Program!
On target for 2007