1. Patrizia Rossi for the RICH Collaboration Laboratori Nazionali di Frascati- INFN (Italy) Physics motivations Status of the project Future Plans CLAS12 2nd European Workshop - March 7-11, 2011- Paris, France

2. CLAS12 Physics Program Hadron PID to achieve flavor tagging Hadron PID to strongly constrain the models Hadron PID to access rare processes Study of the internal nucleon dynamics: TMD distribution and fragmentation functions & GPDs Quark hadronization in the nuclear medium Spectroscopy This program requires good identification of  and K over the full kinematical range accessible with CLAS12 Particle identification is an essential part of any experiment, and has contributed substantially to our present understanding of elementary particles and their interaction

3. The power of a good PID Need to distinguish B d   from other similar topology 2-body decays and to distinguish B from anti-B using K tag. LHCb ( MC prediction) NO RICH With RICH

4. CLAS12 PID RICH SIDIS kinematics full pion / kaon / proton separation over whole accessible momentum range of 2 – 8 GeV for SIDIS exp.  /K separation of 4-5  @ 8 GeV/c for a rejection factor ~1000 GeV/c 12345678910  /K  /p K/p e/  HTCC TOF HTCC EC/PCAL RICH LTCC RICH LTCC RICH LTCC RICH

5. Concept of a RICH for CLAS12 Projective geometry:  6 radial sectors  1.2 m gap  ~ 3 m radius 25 o FTOF wall DC3 TORUS 538 cm 124 cm B ~ 40 G

6. RADIATOR RIC 8 mrad 2 mrad Aerogel mandatory to separate hadrons in the 2-8 GeV/c momentum range  collection of visible Cherenkov light  use of PMTs Freon+UV-light detection does not provide enough discrimination power in the 2- 8 GeV/c momentum range RICH RICH for CLAS12

7. RIC  Large Detector area (several m 2 )  Operation in magnetic field & with high intensity e - beam Challenging project: Innovative new technologies  Radiator Material  Photo-detectors  Electronics RIC ~ 6 m 2 entrance window 1 m depth From a proximity focusing to an “hybrid” RICH

8. 5 8 5 8 5 8 5 8 5 8 CC CKCK   RICH for CLAS12: status and plans MC Simulation for basic parameters studies ✔ (stand-alone GEANT-3 based code) - Aerogel refractive index and thickness - Photon detector pixel size - Gap dimension Fix the pixel size of PMTs < 1cm 5 8 5 8 5 8 5 8 5 8

9. RICH for CLAS12: status and plans MC Simulation with RICH geometry included into the CLAS12 GEMC package (Geant4/C++ based code) - focussing mirrors option studies in progress - Development of the reconstruction tracking algorithm of charged particles in progress Front-end & Readout Electonics - Available readout system to be customized for CLAS12 - Test of the modified system in CLAS12 conditions - Production of the needed boards - Quality checks/characterization Preparation of a Conceptual Design Report by this summer MC Simulation for basic parameters studies ✔ (stand-alone GEANT-3 based code) - Aerogel refractive index and thickness - Photon detector pixel size - Gap dimension Validation of simulations and check performances: RICH prototype construction - Procurements of parts done ✔ - Tests of the radiators and the photo-detectors at Frascati – setup installation started - Prototype beam tests Committed by UTFSM (Chile)

10. Transparent Silica Aerogel with High n New production technique invented for high refractive index greater than 1.05 –Optical quality degraded if sample with n>1.05 is produced in a conventional method –“Pinhole drying (PD)” method artificially shrinks alcogel to obtain high index –Transparency doubled for n>1.05 aerogel with this new method Makoto Tabata, Ichiro Adachi et al. for Belle II aerogel RICH group Some new developments also in Novosibirsk First use of high refractive index aerogel (n=1.13) in particle physics experiment [A.Yu.Barnyakov et.al., Nucl.Instr. and Meth. A 598 (2009) 163]

11. GEMC Simulations GEANT4 toolkit: Toward a complete simulation: realistic geometry / detailed optic effects track multiplicity / background full Cherenkov ring simulation chain Ongoing activities: Improve simulation Reduce costs  mirrors

12. The focusing Mirror System Goals: instrument only forward region reduce active area (~1 m 2 /sect) minimize interference with TOF system allow larger aerogel thickness (focalization) Low material budget Direct & reflected photons

13. The focusing Mirror System Preliminary studies with mirrors (to reduce instrumented area): - focalization capabilities shown - ring patterns for positive and negative mesons at different angles and momenta reconstructed spherical (elliptical) mirror within gap volume for backward refl. plane mirror just beyond radiator for forward reflections Different scenarios (refractive index, radiator thickness, mirror geometry) are being explored TOF Reflecting inside direct & reflected Low material budget Minimize detector area (~1 m 2 /sector) interference with FTOF

14. The reconstruction algorithm: Direct Ray Tracing (DRT) For each track, t, and particle hypothesis, h, use direct ray tracing for a large number of generated photons to determine the hit probability for each PMT The measured hit pattern is compared to the hit probability densities for the different hypotheses by a likelihood function. Hypothesis that maximizes is assumed to be true is the probability of a hit given the kinematics of track t and hypothesis h is the hit pattern from data = 1 if the ith PMT is hit = 0 if the ith PMT is not hit is the probability of no hit is the total number of expected PMT hits is a background term ANL+INFN/FE

15. Direct ring example Hit prob > 3 10 -3 200 trials per event Aerogel: - n=1.06 - thickness increasing with radius: 2 cm up to 13 deg 4 cm 13-15 deg 6 cm 15-17 deg 8 cm 17-20 deg 10 cm > 20 deg Mirror: 14 o -25 o PMTs: UBA M. Contalbrigo INFN/FE P PMT (i)

16. Average N p.e. N p.e. > 5 for reflected rings N p.e.> 12 for direct rings ++ -- 5°

17. LH  -LH K,p ++ -- Contamination as expected from the GEANT3 simulation! Very promising results also for the reflected events Contamination as expected from the GEANT3 simulation! Very promising results also for the reflected events 5°

18. Average N p.e. Mirror 14 o -25 o Mirror 14-35 o Better for negative hadrons ++ -- ++ --

19. LH  -LH K,p Mirror 14-25 o Mirror 14-35 o Better for negative hadrons ++ -- ++ --

20. Average N p.e. Mirror up to 35 o : Viable configuration Mirror up to 35 o : Viable configuration ++ -- Mirror 14°-25°

21. Average N p.e. - Mirror 14 o -25 o n=1.06 Aer. thick 2-4-6-8-10 cm n=1.03 Aer. thick 3-6-9-12-15 cm n=1.03 gives less photons regardless the increase thickness due to same assumed transmission length n=1.03 gives less photons regardless the increase thickness due to same assumed transmission length ++ -- ++ --

22. LH p -LH K,p - Mirror 14 o -25 o n=1.03 in principle good due to the larger Cherenkov angle separation n=1.06 Aer. thick 2-4-6-8-10 cm n=1.03 Aer. thick 3-6-9-12-15 cm ++ -- ++ --

23. n=1.06 better for patter recognition in the presence of backgrouns n=1.06 better for patter recognition in the presence of backgrouns Average N p.e. o ++ -- n=1.06

24. Average N p.e. o ++ --

25. Photo-detectors  Multi-anode PMTs  SiPM visible lightcompactsingle photon Small pad size REQUESTS: MA-PMTDimentional outline (mm 3 ) Effective area (mm 2 )Pixel size (mm 2 ) Comment R760026x26x2818x184.5x4.5 (4x4)Optimized for single photon Recommended by Hamam Low packing factor H8500-C52x52x2849x495.8x5.8 (8x8)Excellent packing factor Not optimized for single photon Not recommended by H. H8500-C-03UV glass window R8900-00-M1625x25x2820x204.8x4.8 (4x4)Optimized for single photon High packing factor Sensitive to B R8900-100-M16Super bialkali R1126523x232.8x2.8 (8x8)Optimized for single photon High packing factor Insensitive to B Available only 8x8 - Preliminary tests results with H8500 and R7600 at Glasgow U. - R8900 will be tested soon

26. Front-end & Readout Electronics Independent channels (unique!) with selectable gain for non-uniformity compensation Smart (reconfigurable) self- triggering by active FPGA (trigger topology scheme) Up to 4096 channels in compact form factor Fast Readout Compact (high density of the front end) Front-end & readout board developed by INFN-Genova group (front- end chip MAROC from IN2P3-Orsay) A.G. Argentieri et al. NIM A 617 (2010) 348–350

27. Players in the Game INSTITUTIONSResearchers ARGONNE NL3 INFN Bari, Ferrara, Genova, Frascati, Roma/ISS 13 GLASGOW U.2 JLAB2 U. CONN3 UTFSM (Chile)3 NEW COLLABORATORS, CONTRIBUTIONS ($, €..), MANPOWER, ARE VERY WELCOME TO JOIN THIS EXCTING PROJECT

28. Conclusions Good hadron identification is required for studies of the internal nucleon dynamics RICH technique is the clear choice when hadron identification is required at high momenta Preliminary studies show that aerogel plus visible light detection with MA-PMT can match the requirements for a RICH for CLAS12. Work is in progress to: - Improve simulation and reconstruction algorithm - Define a CDR by this summer - validate simulations and check performances by testing components and building a prototype Initial R&D funding available from INFN and ANL

29. Error on  Error on the particle velocity  measured with a RICH Depends on :  Number of detected photons  Error on the Cherenkov angle of the emitted photons  Emission point uncertainty Chromatic error Dispersion : 1  = n( ) cos  Photon impact point resolution (~photon detector granularity)

30. RICH Simulations Main output parameter:  k  = mean error on Cherenkov angle reconstruction of k and   K-  = (  K +   )/2 CC CKCK    contamination 1 ‰  Detection efficiency = 80% 4  97% 5  ~100% 6 

31. Hybrid ring example Hit prob Hit prob > 3 10 -3 200 trials per point Aerogel: - n=1.06 - thick. increasing with radius: 2-4-6-8-10 cm Mirror: 14-25 o PMTs: UBA

32. Wrong ID example Hit prob Hit prob > 3 10 -3 200 trials per point Aerogel: - n=1.06 - thickness increasing with radius: 2-4-6-8-10 cm Mirror: 14 o -25 o PMTs: UBA

33. MA-PMTsAdvantages: Uses known photomultiplier technology High quantum efficiency photocathodesare now available – Super-bialkali (typ. QE ~35%) – Ultra-bialkali (typ. QE ~43%) Good effective area -- ~80-85% Good packing density -- square geometry Good high current operationDisadvantages: Cost per channel – 60-150$ per pixel – significantly more expensive per area than single anode tube Quite Large Pixel-to-Pixel Gain Variation Pixel-to-Pixel Crosstalk (~2-5%)