1. Marco Musy INFN and University of Milano-Bicocca Pylos, June 2002 Aerogel as Cherenkov radiator for RICH detectors for RICH detectors

2. The LHCb experiment  Particle ID needed between 1-150 GeV/c Proton-proton Proton-proton interactions at √ s = 14 TeV at LHC  Two RICH systems with 3 Cherenkov radiators  Acceptance: 10-300 mrad (bending plane) 10-250 mrad (non-bending plane) RICH1 RICH2 Local Luminosity 2 x 10 32 cm -2 s -1

3. 3 Momentum LHCb RICHes detectors RICH2 RICH2 RICH1 RICH1 m

4. Aerogel as Cherenkov radiator Light, solid quartz-like structure SiO 2 Physical properties: low density,  = 0.15 g/cm³ n = 1.01 ÷ 1.10, (n = 1 + 0.20  ) A = 95.88 ± 0.04, C = (6.44 ± 0.01) 10ˉ³ μm ¯ 4 /cm A = 91.97 ± 0.05, C = (7.22 ± 0.01) 10ˉ³ μm ¯ 4 /cm A = 88.18 ± 0.06, C = (6.95 ± 0.01) 10ˉ³ μm ¯ 4 /cm Novosibirsk tile 10x10 cm² tile 7x8 cm² tile1 + tile2 Aerogel type : Novosibirsk, Boreskov Institute of Catalysis, Russia (hygroscopic) SP30 Matsushita Electric works Ldt, Japan (hydrophobic) T = A e -Cd/  4 A is the long  transmittance C is the clarity coefficient 8cm

5. Ageing tests with γ 60 Co (E  = 1.3 MeV, 1.7 MeV) Dose : 420 rad/min

6. 1 year LHCb operation Ageing tests with protons Source of radiation: Proton beam 24 GeV/c Flux : 9 10 9 p/cm 2 /s Spot size : 2 x 2 cm 2 Depletion in Transmittance of ~1% after 1 year run (w.r.t. non irradiated sample taken as a reference) 13

7. Humidity tests Expose hygroscopic aerogel tile to humid air (70%) Measure water absorption through weight Measure Transmittance in range 200-800 nm Loss of 30% at 300 nm Loss of 15% at 400 nm Loss of 8% at 500 nm

8. Test Test beam Set-up at CERN Beam from CERN-PS : πˉ and p/π in the range 6 – 10 GeV/c ( Δp/p = 1%)

9. Quantum Efficiency of the 4 photocathodes > 20% ( =280-380nm) Bialkali photocathode, K 2 CsSb Fountain shaped electric field, demagnification factor ≈ 2.3 Silicon pad sensor 2048 pixels (16 sectors x 128 pads 1x1 mm² 2.3x2.3 mm² granularity on ph.cathode ) Hybrid Photo Detectors Hybrid Photo Detectors AEROGEL test beam

10. Aerogel Glass filter Place holder  Rayleigh scattering  Refraction on the boundaries  Light absorption  Light detection on photocathode  Photocathode transparency... All relevant processes are considered in the simulation: Beam axis photons Geant4 Aerogel tile Mirror Photo Detectors Silicon layers Monte Carlo description

11. Sector #4 Sector #8 Ring region Out of ring sect 4 sect 8 Test beam results 9 Gev/c π ¯ beam 4 cm aerogel Novosibirsk noise/pad < 2%

12. Photoelectron yield Number of ph.electrons Novosibirsk 4 cm aerogel 8 cm aerogel On ring Integrate signal across the measured arcs and compare with Monte Carlo Evaluate nr. photoelectrons: - on ring, |R-R| < 3σ - out of Cherenkov ring

13. Novosibirsk No filterFilter D263 4 cm 8.3 ± 0.3 10.0 ± 1.1 5.6 ± 0.2 6.4 ± 0.7 8 cm 10.7 ± 0.4 12.9 ± 1.1 8.4 ± 0.3 8.9 ± 1.0 Photoelectron yield cont’d Contributions to total error: background subtraction (±1σ): ~ 5% inefficient or noisy pads : ~ 4% definition of ‘active region’ (±1mm): 2% separation of on-ring/off-ring (±2mm): 3% signal losses outside ADC thresholds (±1σ): 3% Data MC results are normalised to 2π acceptance on-ring off-ring Npe 8 cm 4 cm 8 cm 4 cm No filterD263 4 cm (off-ring) 1.00 ± 0.10 0.81 ± 0.08 0.57 ± 0.04 0.55 ± 0.04 8 cm (off-ring) 1.10 ± 0.10 1.06 ± 0.11 0.94 ± 0.07 0.84 ± 0.10 results are in units of 10 ¯ ² /cm ²

14. Ring reconstruction Thickness No filterFilter D263Glass θ c σ θ 4 cm 250.0 5.4 248.7 4.0 247.1 5.0 246.8 3.0 243.6 5.3 243.2 3.8 8 cm 246.8 5.8 245.0 3.9 245.4 4.8 243.7 3.0 246.0 5.3 -- 6 cm250.2 8.7250.9 5.8251.3 5.4 8 cm249.5 9.8250.3 6.2 -- Data MC Data Novosibirsk Matsushita rad Study resolution as a function of - filter type - aerogel thickness - aerogel type single photoelectron Results per single photoelectron are (mrad): -- Data -- Monte Carlo 4cm Novosibirsk (no filter) θcθc

15. Ring reconstruction cont’d Contribution to angular resolution is determined with the simulation : Resolution is expected to scale as A/√N + k (in the 3σ ring region) Sourceσ (mrad) Pixelling1.3 Chromaticity2.5 Point spread func. + Emission point 1.1 Beam divergence0.7 Allignement2.2 TOTAL3.8 -- Fit to data Resolutions differ by ~20-40% in MC with respect to the Data. Still under investigation.

16.  / p separation at PID performance 8 GeV 6 GeV 10 GeV rad SINGLE ph.e. 6.1  4.8  3.1  θ ~30,000 events

17. PID performance cont’d Evaluate separations N σ = Δθ/σ θ, scaling with the N pe and extrapolate to the total acceptance Energyθpθp θπθπ NσNσ 6 GeV194.0±7.8243.6±2.99.3 8 GeV216.4±4.1244.3±2.88.1 10 GeV224.8±3.0242.8±2.36.8 π-ring p-ring Clear π/p separation For 4 cm aerogel + filter:

18. Conclusion  The use of aerogel as Cherenkov radiator has become reliable in high energy particle physics  Test beam has shown a photon yield which agrees with the Monte Carlo expectations  Good PID ability in the momentum range 6 – 10 GeV/c  Further studies are on the way