1. CsI-TGEM vs. CsI-MWPC photodetector Some part of this work was performed in collaboration between CERN and Breskin group

2. There are two options for planar photodetectors: TGEMs/RETGEMs ( see P. Martinengo talk) MWPC (Currently used in ALICE and COMPASS RICH) or

3. Some considerations to be taken into account: I) Cherenkov light detection efficiency II) Discharges at low rates III) Discharges at high rate (additional gain drop withrate, cathode exciataion effect IV) Gases

4. I) Cherenkov light detection efficiency

5. Q TGEM =Q vac ε extr (E,λ) QE in gas QE in vacuum Fraction of photoelectrons extracted from PC (depends on E and gas) N pe =  Q TGEM (λ) I(λ)f pe f pe ~ exp(-A th /A 0 For MWPC:

6. Ε extr measurements Such curves were measures by many authors, see for example the lates publicationJ. Escada et al., JINST 4:P11025,2009J. Escada

7. Q TGEM =Q vac ε extr (E,λ) S eff ε coll (A) QE in gas QE in vacuum Active area Fraction of photoelectrons extracted from PC (depends on E and gas) Fraction of photoelectrons collected into the TGEM holes- depends on gas gain N pe =  Q TGEM (λ) I(λ)f pe f pe ~ exp(-A th /A 0 For a single TGEM/RETGEM

8. Q TGEM =Q vac ε extr (E,λ) S eff ε coll (A) ε transf1.. ε transf1 QE in gas QE in vacuum Active area Fraction of photoelectrons extracted from PC (depends on E and gas) Fraction of photoelectrons collected into the TGEM holes- depends on gas gain For cascaded TGEMs/RETGEMs =1 In avalanche mode

9. Electric filed on the top TGEM electrode Ε extr : estimations in the case of TGEM/RETGEM C. Azevedo, et al., 2010 JINST 5 P01002Azevedo

10. ε coll (A) measurements in Breskin group (single photoelectron mode) C. Azevedo, et al., JINST 5 P01002, 2010Azevedo

11. ε coll (A) measurements in our group at CERN (pulsed mode) ΔI sat ΔI back = ΔI sat ε coll A εcoll (A)= ΔI back / AΔI sat Pulsed D 2 lamp Drift mesh Back plane TGEM CsI Avalanche

12. ε coll (A): results of indirect measurements in laboratory Q TGEM ≈65%Q MWPC

13. Beam Cherenkov light 40mm 3mm 4.5mm Drift gap 10mm R/O pads 8x8 mm 2 Front end electronics (Gassiplex + ALICE HMPID R/O + DATE + AMORE) CsI layer Drift mesh Ne/CH 4 90/10 Indirect Q TGEM measurements on the beam(Summer2010) 4 mm CaF 2 window

14. ~20 o ~37 o

15. Time after coating [h] Normalized photocurrent CsI quality control HMPID level

16. Analysis of the beam test data shows that for the given geometrical layout the QE of TGEM (after geometrical corrections) is compatible to one of the HMPID (which confirms the scan data!) Monte Carlo simulations well reproduce the experimental data Q TGEM ≈70-80%Q MWPC Ne/CH4 Ne/CF 4

17. 2 3 4 5 1 Pad plain RETGEMs CsI Drift mesh Was manufactured New, exists Old, exist Was modified Old, exist Proto-4 (schematic side view) ~70 3 3 3 ~60 R c = ~135 ~30 11 30 Direct Q TGEM measurements (November 2011 ) C 6 F 14 radiator

18. 135 The top view of the frame #3 (from the electronics side) Feethroughts RETGEM supporting flame New holes Cherenkov ring TGEMs 4 3 1 2 5 6

19. Data are still under analysis, however qualitatively QE TGEM ≈ 50%Q MWPC

20. Raether limit: A max n 0 =Q max =10 6 -10 7 electrons, where n 0 is the number of primary electrons created in the drift region of the detector (Q max depends on the detector geometry and the gas composition/track density) II.1) Discharges in TGEMs/RETGEMs Low rate

21. General curve High rates

22. General limit for all micropattern gaseous detectors: ( a very similar curves were measured for MWPC as well, however the physics behind is different) See also : P. Fonte et al, NIM A419,1998, 405;Yu. Ivaniouchenkov et al, NIM A422,1999, 300; P. Fonte et al., Nuclear Science, IEEE Transactions 46, Issue 3, Part 1, June 1999 Page(s):321 - 325 P. Fonte et al ICFA Insrum Bull., Summer 1998 issue, SLAC-JOURNAL, ICFA-16 For physics behind this effect see V. Peskov et al., arXiv:0911.0463,2009arXiv:0911.0463

23. V. Peskov et al., JINST 5 P11004, 2010 Results obtained wit bare (not coated with CsI) TGEMS

24. Rate effect for CsI coated TGEMs Alice region

25. II.2 Discharges in MWPC The maximum gain is determined by a feedback loop

26. Discharges in thin wire detectors Primary avalanche Small gain

27. VUV photons Larger gain

28. Secondary photoelectrons

29. Secondary avalanches

30. Aγ=1 (Aγ ph =1 or Aγ + =1) Geiger mode in quenched gases Geiger discharge is not damaging. One can observed signals~1V directly on 1MΩ input of the scope (no amplifier is needed) c front =10 6 -10 7 cm/s

31. This discharge is not destructive because there is no any conductive bridge between the anode and the cathode

32. Space charge effect on gas amplification. In this figure taken from [A.H. Walenta, Physica Scripta 23, 1981, 354 ] G/G0 is the gas gain relative to zero counting rate, Q is a total charge in a single avalanche and n is particle rate/wire length. Rate effect in MWPC

33. V. Peskov et al., JINST 5 P11004, 2010

34. Discharges in thick wire detectors Anode wire (grounded via amplifier) Cylindrical cathode Avalanche -V

35. Transition to streamer occurs when An 0 ≥Q max =10 8 electros Self-quenched streamer Strimers give huge amplitudes but the are not harmful as well Streamers cannot propagate to the cathode because the electric field drops as 1/r Streamer

36. Signal’s amplitude in proportional and streamer modes Avalanche Streamer

37. …so in practice: in bare MWPC steamers (and streamers triggered discharges) may appear in “weak” regions (not well stretched wire, dust and cetera) or regions of dielectric surfaces

38. The maximum achievable gain is determined by a feedback loop: A m =1/γ, where γ is a probability of creation secondary electrons (as a results there are no sparks in presence of Ru or Fe source) Typical results obtained with CsI-MWPC: Voltage (V) Gain

39. III.3. Cathode excitation effect

40. Quantum efficiency vs. wavelength of metal (rhombus) and CsI (triangles) cathodes measured in as ingle-wire counter before a corona discharge (solid symbols) and immediately after the corona discharge was interrupted Changes in QE vs. time for Cu (rhombus) and CsI (triangles) photocathodes V. Peskov et al., arXiv:0911.0463, 2009arXiv:0911.0463 Cathode excitation effect in CsI single wire-counters CsI pc Metal pc CsI pc Metal pc

41. Conclusion: triple TGEM less suffering from the cathode excitation effect than MWPC Continuous discharge is possible due to the cathode excitation effect Recent measurements with CsI-MWPC and with CsI-TGEM

42. IV.Gases CsI-MWPC can efficiently operate only in CH4. This arise safety concerns In contrast TGEMs/RETGEMs can operate in many gases. This open possibility to use windowless design when GEM or TGEM/RETGEM operates in the same gas as uses for the Cherenkov radiator

43. Conclusions: ●CsI MWPC reached their operational limit in gain ( 5x10 4 ) and in QE (80-90% of Q vac ) ● Discharges are possible in MWPC due to the design features, defects and cathode excitation effect ● In contrast TGEMS/RETGEMs have several advantages and unexploited yet potentials: higher gain, possibility to increase A eff and thus QE, wider choice of gases, possibility to exploit windowless designs, less troubles from the cathode excitation effect… and more ● Thus it looks that TGEM/TRETGEMs is an attractive option for some gaseous RICH detectors and for this reason is still under consideration and studies for the ALICE RICH upgrade

44. Spare

45. TGEM+MWOC option (sugested by Hungarin team)

46. Feature #4 Cathode excitation effect (closely related to the rate effect physics-see P. Fonte et al., IEEE Nuc.Sci 46,1999,321)