1. The RICH prototype V. Lucherini, M. Mirazita, S. Pereira, D. Orecchini, A. Orlandi, A. Viticchie’ - INFN Frascati L. Barion, M. Contalbrigo, R. Malaguti, A. Movsisyan, L.L. Pappalardo - INFN Ferrara L. Lagamba, R. Perrino - INFN Bari P. Musico - INFN Genova E. Cisbani – ISS/INFN Roma R. Montgomery, J. Phillips - Glasgow University M. Hoek - Institut fur Kerniphysik, Mainz V. Kubarovsky, P. Rossi – JLab F. Benmokhtar – Duquesne University, USA C. Colvin, A. Losada – CNU University, USA M. Turisini - UTFSM, Chile 1

2. The CLAS12 RICH beam pipe particle’s trajectory target  DC1 DC2 DC3 The RICH performances are a combination of:  Geometric effects photon path length pixel size photodetector coverage  can be easily simulated; example: photon emission point wrt photon path length  Optical effects aerogel characteristics mirror reflectivity  may be very complicated: need validation Prototype construction Limits set by: overall size, few m 3 availability of components, e.g.: 28 MAPMTs, commercial glass mirrors in the protoype 2  Photon detection direct photons: - 1m path length - thin (2cm) aerogel radiator reflected photons: - spherical and planar mirrors - ~3m path length - thick (6cm) aerogel radiator - double pass through thin aerogel - focalization

3. Prototype construction: direct light Radiator H8500 beam H8500 radiator 1 meter Same geometry as in the CLAS12 RICH -same gap length -same aerogel thickness Different PMT disposition - smaller ring coverage 3 extrapolation to the CLAS12 RICH is straightforward

4. planar mirrors aerogel absorber H8500 radiator mirror Radiator H8500 curved mirror beam mirrors + aerogel Prototype construction: reflected light Same concept of CLAS12 RICH -spherical and planar mirrors -same aerogel thicknesses Different geometry -smaller gap length -no mirror focalization -smaller ring coverage 4 extrapolation to the CLAS12 RICH needs MC simulation

5. The prototype in the T9 experimental Hall at CERN RICH GEMs beam RICH GEM scintillators (trigger) threshold gas Cerenkov  /K separation) beam pions few % kaons 5

6. Track reconstruction and PID GEM for track reconstruction RICH GEM scintillators (trigger) threshold gas Cerenkov beam pions few % kaons Upstream GEM Downstream GEM Threshold Cerenkov counter for PID pions kaons 6

7. The MAPMT electronics Maroc3 front end electronics 2 control board with 4 back planes up to 16 front end cards per back plane 64 channels per card, 4096 total channels preamplifier, adjustable from 1/8 to 4 ADC control board back planes MAROC front end cards Linux DAQ program (MAROC+GEM+CC) Event transfer to disk in single or multi event mode 7

8. MAPMT signal reconstruction On-line event display 8

9. Event reconstruction 9 o Ring center fixed from GEM track o Ring radius fitted with MAPMT hits o Hits far from the ring: background o Cerenkov angle computed using geometrical optics background hit pions p=8 GeV/c aerogeln=1.05 t=2 cm about 1 bkg hit per event Rayleigh - scintillation - external light - etc

10. direct light measurements 10

11. Pion resolution pions p=8 GeV/c aerogeln=1.05 t=2 cm 11 Number of p.e. dependence single photon angular resolution  C = 307.6 mrad   = 1.402±0.008 mrad  1pe = 4.55 ±0.17 mrad  0  0 Non-Cherenkov term: tracking, alignments, etc

12. 3  pion band pbar band kaons Efficiencies Pion/Kaon ID p=6 GeV/c p=7 GeV/c p=8 GeV/c pions kaons(x60) 12 Number of  separation Rejection factor

13. Single photon resolution normal glass MAPMTs produce 1 photon less but 30% better resolution than UV glass MAPMTs UV photons H8500C H8500C-03 14+14 MAPMT with normal/UV glass in alternated positions Compare ring reconstruction 13 residuals (mm)

14. reflected light measurements 14

15. The curved mirrors JLab mirror: prototype of HTCC mirrors - elliptical: a1=1850 mm a2=1613 mm Glass mirror - spherical: f=900 mm 15 Radiator H8500 curved mirror beam mirrors + aerogel Compare different technology mirrors

16. The planar mirrors eight planar mirrors aerogel tiles 5 new 3 old 16 Jun12 Dec12 Radiator H8500 curved mirror beam mirrors + aerogel Jun12 Dec12

17. Comparison without/with absorbers =13.1 =5.3 Fitted ring radius distribution (mm) 17 Yield reduction: 60% significant amount of light survives beam: P=6 GeV/c glass spherical mirror aerogel radiator:n=1.05t=6cm aerogel absorbers:n=1.05t=2 cm o In/Out without absorbers with absorbers

18. Cerenkov angle resolution smaller Npe Pion resolution 18  R = 2.7 mm  R = 3.8 mm absorber  1pe [mm]  0 [mm] NO8.81±0.250.22±0.04 YES9.10±0.210.35±0.09 non-zero constant term misalignments Relevant in the prototype only  may be subtracted P=6 GeV/c  1pe [mrad]   [mrad] direct4.55±0.171.41±0.02 refl. NO6.12±0.171.71±0.05 refl. YES6.27±0.142.41±0.06 Ring Radius

19. Pion/kaon separation Number of  separation 19 Without absorbers pions kaons(x60) With absorbers pions kaons(x60) Without Absorbers:n  =3.6 With Absorbers:n  =2.8

20. Mirror comparison JLab: N=13.1 Marcon: N=13.1 JLab: N=5.3 Marcon: N=5.8 20 Glass spherical vs JLab elliptical No GEM tracking used for this analysis Ring radius and center fitted from MAPMT hits  worse resolutions with JLab mirror scaling factors to compensate for the elliptical shape without absorbers with absorbers

21. From the prototype to the CLAS12 RICH 21  External (e.g. misalignments) effects may be neglected  Better ring coverage (close to 89% MAPMT pack. fraction) no more than 70% coverage in the prototype better resolution, smaller event loss  Non-UV glass window MAPMTs: +15% resolution H12700  Aerogel quality improvements 5 of the 8 absorbers with lower transparency more photon yield smaller yield loss for reflected photons  Larger path length for reflected photons decrease the effect of the emission point uncertainty MAPMT in the spherical mirror focal plane

22. Simulation of the prototype 22 Validation of the MC simulations using prototype data -detailed description of the setup -real tracks Study various contributions to the measured resolutions -digitalization of the MAPMT response -mirror reflectivity -aerogel characteristics: ref. index and dispersion, Rayleigh, surface effects, etc

23. aerogel filter n( ) n=1 air n=1.47 measured radius Cerenkov radius full wavelength spectrum selected wavelength Aerogel dispersion Study of Cherenkov ring radius as a function of Direct light setup with 23 fit MC nominal refractive index 1.05 prototype results - direct light setup Novosibirsk quotation: n 2 (400nm) = 1+0.438  → n(400nm)=1.0492

24. Comparison data/MC 24 direct light setup reflected light setup without absorbers with absorbers

25. Resolution decomposition 25 MAPMTs Contributions to the measured resolutions with the prototype Aerogel DATA  MISAL contribution not relevant for the CLAS12 RICH  Larger contributions from aerogel  Some of the contributions will be smaller in the CLAS12 RICH (EMIS, ASURF)  Use lucite to suppress UV and Rayleigh  Time coincidence may reduce the backgrounds  Background effects (Rayleigh, part of the DIGI) will be better managed by the CLAS12 RICH event reconstruction Residual

26. Conclusion 26 We had a successful campaign of test beam to study the various features of the foreseen RICH detector  A large-scale prototype has been build to test the various features of the CLAS12 RICH  direct light measurements indicate that the CLAS12 RICH will meet the PID requirements for small angle particles  reflected light measurements indicate that the necessary PID requirements may be reached also for large angle particles  MC simulation are able to reproduce the prototype results  CLAS12 RICH performances computed by simulations  M. Contalbrigo talk

27. support slides 27

28. Prototyping activity July-August 2011exploratory test beam at CERN (hadrons) - small scale prototype - encouraging results July 2012test beam at Frascati BTF (electrons) - setup of DAQ and electronics July-August 2012 Nov-Dec 2012test beam of a large scale prototype (hadrons) - test of the direct light configuration - proof-of-principle of the reflected light configuration - exploratory test of SiPM July-August 2013test beam at Frascati BTF (electrons) - digital readout of MAROC 28

29. Ring fit (x i, y i ) hits of the event, i=1,N(X C, Y C )ring center Rring radius Minimization of 3par fit: X C, Y C and R are fit using the MAPMT data 1par fit: X C, Y C are fixed from GEM tracking and R is fit using the MAPMT data 29

30. GEM alignment 3par fit  =1.62mrad 1par fit before  =2.55 mrad 1par fit after  =1.40 mrad Alignment constants for the GEM position (X 0, Y 0 ) (X 1, Y 1 ) dermined by minimization of 30

31. Background cut  = res. of integrated distribution No cut results Ring radius resolution vs cut to remove bkg hits 31

32. Systematic studies: aerogel production pions p=8 GeV/c aerogeln=1.05 t=2 cm same “nominal” characteristics different production date BLUE: latest productionRED: older production With newest production techniques: -more photon yield -better chromatic performances 32

33. Aerogel quality – reflected light 33 3par fit: no GEM tracking information o no absorbers o 8old absorbers o 5new+3old absorbers (the best old tiles) Number of p.e. Ring radius (mm)  comparing old/new absorbers: same yield, better resolution  comparing without/with(5new+3old) absorber - 40% photons surviving - some degradation in the single photon resolution With old aerogel production tiles degradation due to double pass increases by few%

34. fit MC nominal refractive index 1.06 aerogel filter n( ) n=1 air n=1.47 measured radius Cerenkov radius full wavelength spectrum selected wavelength Aerogel dispersion Study of Cherenkov ring radius as a function of Direct light setup with 34 prototype results - direct light setup

35. Residuals – MC comparison 35 res = d(Hit, Ring) red: data black: MC PMT 0 PMT 27

36. SiPM measurements 36

37. SiPMT setup RICH GEM scintillators (trigger) SiPM threshold gas Cerenkov beam pions few % kaons one commercial 8x8 matrix with 3x3mm 2 pixel size two custom 4x8 matrix with 3x3mm 2 pixel size with preamp stage water cooling system with a Peltier cell from -25 o to +25 o TDC spectra measured in 3ns window 37

38. SiPMT test RICH GEM scintillators (trigger) SiPM threshold gas Cerenkov beam pions few % kaons one commercial 8x8 matrix with 3x3mm 2 pixel size two custom 4x8 matrix with 3x3mm 2 pixel size with preamp stage water cooling system with a Peltier cell from -25 o to +25 o TDC spectra measured in 3ns window 38

39. Measurements at +25 o commercial matrix custom matrix + preamp Best pixel: -strong variations of signal with the bias 4% occupancy  about 24 p.e. in full ring -preamp important -high level of bkg, reduction with high threshold 39

40. Measurements at -25 o commercial matrix custom matrix + preamp Best pixel: -more stable behaviour with bias -reduction of bkg even at lower thresholds 10 -4 bkg comparable with H8500 40