1. The Cylindrical GEM detector for the KLOE-2 Inner Tracker G. Bencivenni, LNF-INFN, Italy on behalf of the KLOE-IT group 1 LABORATORI NAZIONALI DI FRASCATI www.lnf.infn.it 52 nd International Winter Meeting on Nuclear Physics, Bormio 27-31 January 2014

2. 2 G. Bencivenni (LNF-INFN, Italy)  Introduction  GEM detector: principle of operation  The Cylindrical-GEM:  the idea  the construction  performance  Detector integration & preliminary results  Conclusions

3. 3 6 m 7m Lead/scintillating fiber 98% coverage of solid angle 88 modules (barrel + end-caps) 4880 PMTs (two side read-out) Lead/scintillating fiber 98% coverage of solid angle 88 modules (barrel + end-caps) 4880 PMTs (two side read-out) 4 m diameter × 3.3 m length 90% helium, 10% isobutane 12582/52140 sense/tot wires All-stereo geometry 4 m diameter × 3.3 m length 90% helium, 10% isobutane 12582/52140 sense/tot wires All-stereo geometry Electromagnetic Calorimeter Drift Chamber σ r φ = 150 µm, σ z = 2 mm σ V = 3 mm σ p /p = 0.4 % K, π momentum range of the order 100-200 MeV σ r φ = 150 µm, σ z = 2 mm σ V = 3 mm σ p /p = 0.4 % K, π momentum range of the order 100-200 MeV KS = 0.6 cm KL = 340 cm K  = 95 cm σ E /E =5.4%/ σ t = 54 ps/ σ E /E =5.4%/ σ t = 54 ps/  100 ps(calib) B = 0.52 T

4. 4 K s, η η To improve vertex reconstruction of K s, η and η’ decay K s - K L interference measurements, an upgrade of the tracking system has been proposed: and K s - K L interference measurements, an upgrade of the tracking system has been proposed: 1.  rφ  200 µm and  z  500µm low material budget: < 2% X 0 (to minimize MS of low momentum tracks, and low energy photon absorption) 2. low material budget: < 2% X 0 (to minimize MS of low momentum tracks, and low energy photon absorption) 3. 5-10 kHz/cm 2 rate capability K S → ππ vertex resolution will improve of about a factor 3 wrt only-DC config. (<1mm)  4 CGEM layers with radii from 130 to 205 mm from IP and before DC Inner Wall  700 mm active length  XV strips-pads readout (25 o ÷30 o stereo angle)  2% X 0 total radiation length in the active region 3 mm 2 mm Cathode GEM 1 GEM 2 GEM 3 Anode Read- out Cylindrical Triple GEM Cylindrical GEM (CGEM) detector is the adopted solution G. Bencivenni, LNF-INFN, Italy

5. The GEM (Gas Electron Multiplier) [F.Sauli, NIM A386 (1997) 531] is a thin (50 μm) metal coated kapton foil, perforated by a high density of holes (70 μm diameter, pitch of 140 μm) by using standard photo-lithographic technology. By applying 400-500 V between the two copper sides, an electric field as high as ~100 kV/cm is produced into the holes which act as multiplication channels for electrons produced in the gas by an ionizing particle. Gains up to 1000 can be easily reached with a single GEM foil. Higher gains (up to 10 4 ) and/or safer working conditions are usually obtained by cascading two or three GEM foils. 5 G. Bencivenni, LNF-INFN, Italy

6. 6 G. Bencivenni (LNF-INFN, Italy) Electron transparency (single-GEM) Electrons Diffusion Losses I - out = I - in. G. T (gain x transparency) Ions Ion collection CathodeAnode DriftField InductionField ~50% ~50% signal mainly due to electron motion

7. 7 G. Bencivenni (LNF-INFN, Italy) flexible geometry, arbitrary detector shapeflexible geometry, arbitrary detector shape: rectangular/square, annular, cylindrical ultra-light structure, very low material budget:ultra-light structure, very low material budget: ~3‰ X 0 /detector gas multiplication separated from readout stage, arbitrary readout pattern:gas multiplication separated from readout stage, arbitrary readout pattern: pad, strips (XY, UV), mixed … high rate capability:high rate capability: > 50 MHz/cm 2 Safe high gains:Safe high gains: > 10 4 high reliability:high reliability: discharge free; P d < 10 -12 per incoming particle rad. hard.:rad. hard.: up to 2.2 C/cm 2 integrated over the whole active area without permanent damages (corresponding to 10 years of operation at LHCb) good spatial resolution:good spatial resolution: down to 70 µm (with analog readout) (COMPASS) good time resolution:good time resolution: down to 3 ns (with CF 4 ) (LHCb)

8. COMPASS: COMPASS: 22 triple-GEM chambers, 310x310 mm 2 active area; 2-D charge readout (XY strips with 400 µm pitch), are the backbone of the inner tracking system of the Muon Spectrometer at the CERN SPS. GEM applications in HEP (C.Altunbas et al, Nucl.Instr.and Meth., A490(2002)177) 70 μ m r.m.s. APV25 128 chs analog output

9. HV filters CARIOCA-GEM FEE LV & LVDS card G~4000 G~17000 GEM applications in HEP 2.9ns r.m.s. Time resolution of two chambers in OR Ar/CO 2 /CF 4 =45/15/40 Drift = 3.5 kV/cm Transfer = 3.5 kV/cm Induction = 3.5 kV/cm LHCb: pad readout LHCb: 24 triple-GEM chambers, 20x24 cm 2 active area; pad readout (10x25 mm 2 ); for L0 muon triggering (*); extensively tested for ageing/etching (**). Recipe for geometry and gas mixture for FAST operation of the detector. Introduction of the GEM stretching idea. (*) G. Bencivenni et al., NIM A 494 (2002) 156 (**)P. de Simone et al., IEEE Trans. Nucl. Sci. 52 (2005) 2872

10. GEM applications in HEP TOTEM: half-moon shaped TOTEM: 40 triple-GEM half-moon shaped, inner radius 40 mm, outer 150 mm; mixed readout  radial pad rows (3x3  7x7 mm 2 ) and radial strips(400 µm pitch) VFAT readout 128 chs chip with digital output K. Kurvinen, 10 th Pisa Meeting on Advanced Detectors (Elba2006)

11. 11 Full sensitive (no spacers between electrodes) and very light detector To build a cylindrical electrode we exploit: the remarkable flexibility of polyimide the remarkable flexibility of polyimide the vacuum bag technique the vacuum bag technique The foil (GEM, anode, cathode) is wrapped on a fine machined PTFE cylindrical mould, then enveloped in a vacuum bag. After epoxy polymerization the cylindrical electrode is easily extracted. five cylindrical structures(“self supporting”: NO stretching is needed). The final detector is obtained inserting one into the other the five cylindrical structures (“self supporting”: NO stretching is needed). 11 G. Bencivenni (LNF-INFN, Italy) The CGEM is composed by five concentric cylindrical electrodes: the anode, 3 GEMs and a cathode.

12. The IT GEM foil 12 G. Bencivenni (LNF-INFN, Italy) new manufacturing procedureThe dimensions of the electrodes for the IT required a new manufacturing procedure 350 x 700 mm 2 (first time for an experiment)The CERN TE-MPE-EM workshop (Rui de Oliveira) produced large area GEM foils (up to 350 x 700 mm 2 ) using the single-mask technique (first time for an experiment) Every GEM foil is divided in 40 HV sectors (1.5 x 70 cm 2 ) on the top side and 4 HV sectors on the bottom side in order to reduce the energy released through the hole in case of discharge. Each sector is supplied separately: in case of short it can be disconnected (0.8% of a layer)

13. The readout of the IT The readout of the IT is a flexible kapton/copper circuit. The 2-D view is given by the X- strips (parallel to the axis of the CGEM) and V pads connected by vias to a common backplane. 13 G. Bencivenni (LNF-INFN, Italy)

14. GASTONE: the FEE for the IT 14 G. Bencivenni (LNF-INFN, Italy) Mixed analog-digital circuit Low input equivalent noise, low power consumption and high integrated chip 4 blocks: charge sensitive amplifier shaper leading-edge discriminator (programmable threshold) monostable (stretch digital signal for trigger) Sensitivity (pF)20 mV/fC Z IN 400 Ω (low frequency) C DET 1-50 pF Peaking time90-200 ns (1-50 pF) Noise (rms)800 e - + 40 e - /pF Channels/chip64 ReadoutLVDS/Serial Developed by INFN Bari and LNF

15. Quality check 15 G. Bencivenni (LNF-INFN, Italy) The GEM foils are tested in a N 2 flushed box for humidity reduction (RH below 10%) 600 V Each sector of the foil is supplied with up to 600 V O(1) h-1(<1nA) Discharge rate (O(1) h-1) and leakage current (<1nA) are monitored. < 2 Ω HV connections are checked to have R < 2 Ω ~ 4 hours. A complete test lasts ~ 4 hours. The GEM foil yield was around 85%.

16. The CGEM construction 16 G. Bencivenni (LNF-INFN, Italy) Three foils are spliced together along the kapton frame The large electrode foil is then rolled on the mould, glued and polymerized under the vacuum bag Fiberglass rings acting as spacers are glued at the two ends of the GEM electrode The cylindrical GEM electrode is ready to be extracted from the mould: the very low friction of the Teflon avoid mechanical damages on the foil.

17. Assembly and CR test 17 G. Bencivenni (LNF-INFN, Italy) The Vertical Insertion System provides the insertion of the five cylindrical electrodes, one into the other, ensuring a distance between the axis less than 100 μm over 1 m length. 90 Sr source test to check the functioning along the φ angle Cosmic-ray test with scintillators trigger and three PGEMs as external trackers

18. 18 G. Bencivenni (LNF-INFN, Italy) Ar/CO 2 =70/30 and B=0.5 T, average Lorentz angle α L ~ 8° displacement (dx, w spread The effect of the magnetic field (B ┴ E) is twofold: a displacement (dx, wrt the track impact direction) and a spread of the ionization charge over the readout plane (effect visible only on the “bending plane”). FWHM= p8 - p1 = 639.1 µm Mean = (p8+p1)/2 = 989.4 µm Garfield simulation

19. 19 G. Bencivenni (LNF-INFN, Italy) σ Z =370 µm σ r-φ =200 µm

20. 20 G. Bencivenni (LNF-INFN, Italy) increase of the magnetic field, requiring for higher gain to efficiently operate the detector. The increase of the magnetic field, increasing the spread of the charge over the readout strips (less charge is collected by each single pre-amp channel) results in an efficiency drop, thus requiring for higher gain to efficiently operate the detector.

21. IT assembly 21 G. Bencivenni (LNF-INFN, Italy) Final assembly of the KLOE-2 Inner Tracker, with the insertion of all the triple-CGEMs one into the other.

22. IT integration: insertion on the BP 22 G. Bencivenni (LNF-INFN, Italy)

23. IT integration 23 G. Bencivenni (LNF-INFN, Italy) IR sliding to the final position IR at the final position Shell extraction

24. IT gas system 24 G. Bencivenni (LNF-INFN, Italy) The IT is flushed at 24 l/h (100 cc/min per layer) Filters present along the inlets of each layer Isobutane sniffer close to the IT to detect gas leakage We can set the flux for each layer and monitor the overpressure of the line Gas parameters read on a dedicated PC Data transmission to slow control for the implementation of the alarm

25. HV and temperature slow control HV and temperature slow control 25 Final HV slow control is being set up A semi-graphical and a HTML interface are foreseen 6 probes on the innermost surface of the IT 4 probes on the FEE 4 probes on the water cooling circuit for the GASTONE boards

26. Very preliminary results Very preliminary results 26 G. Bencivenni (LNF-INFN, Italy) Efficiency Layer 185% Layer 289% Layer 390% Layer 457% Failure of one of the HV supply module Cosmic-ray muon sample selected with EMC Operational parameters still to be optimized Number of hit strips (strip multiplicity) per event for all layers and both views r-φ view Z view

27. 27 G. Bencivenni (LNF-INFN, Italy) Cosmic events Cosmic events

28. 28 G. Bencivenni (LNF-INFN, Italy) Cosmic events Cosmic events

29. 29 G. Bencivenni (LNF-INFN, Italy) Conclusions great robustness, long-term stability of operation, remarkable flexibility harsh environments.  Among Micro-Pattern Gas Detectors, the GEM technology has demonstrated great robustness, long-term stability of operation, remarkable flexibility and capability to accomplish different tasks in harsh environments. first CGEM based detector innovative concept of full sensitive low-mass vertex detector.  The first CGEM based detector has been successfully completed and installed, validating the innovative concept of full sensitive low-mass vertex detector. Big effort is still to be done to complete the commissioning of the detector Big effort is still to be done to complete the commissioning of the detector

30. G. Bencivenni (LNF-INFN, Italy) 30

31. 31

32. Operational parameters 32 G. Bencivenni (LNF-INFN, Italy) Gas mixture: Ar:i-C 4 H 10 90:10 (wrt CO 2 better quenching & smaller Lorentz angle) Ar:i-C 4 H 10 90:10 (wrt CO 2 better quenching & smaller Lorentz angle) Fields: 1/3/3/5 kV/cm Fields: 1/3/3/5 kV/cm GEM voltages: 295/285/280 V Gain: O(10 4 ) average Lorentz angle α L < 6°

33. G. Bencivenni (LNF-INFN, Italy) 33

34. 34 G. Bencivenni (LNF-INFN, Italy) The KLOE apparatus, consisting of a huge Drift Chamber and an Electromagnetic Calorimeter working in a 0.5 T axial magnetic field, has been upgraded with new sub-detectors, including an Inner Tracker, for a new data taking campaign. The required inner tracker performances are : 200 μm spatial resolution on the transverse plane and 500 μm along the beam line Material budget less than 2% X 0 5 kHz/cm 2 rate capability The Inner Tracker is composed of 4 coaxial cylindrical triple-GEMs with 700 mm active length Radii between 130 and 205 mm X-V stereo readout Very low mass detector

35. G. Bencivenni (LNF-INFN, Italy) 35 KLOE experiment

36. G.Bencivenni, LNF-INFN, PANIC08, November 11th 2008 36 RMS ~ 13 ns T max ~ 80 ns RMS ~ 200 ns T max ~ 750 ns standard zone gluing zone CGEM time spectra simulation Ar/CO 2 = 30/70 Time spectra with Ar/CO 2 = 30/70 gas mixture, obtained with CARIOCA- GEM. Ionization electrons, generated above a gluing region by a track, drift along the distorted field lines and then are efficiently driven and focused in the multiplication holes of the GEM. Ar/CO2 (70/30): 7 cm/μs from @ 3 kV/cm 7 cm/μs from @ 3 kV/cm 10 clusters in 3 mm 10 clusters in 3 mm

37. GEM Zoology Fig 1.a: top Fig 1.b: bottom Fig 2: small missing hole Fig 3: continuous discharges effect Rejected foil Typical discharge spotvTypical discharge spots

38. Fig 4a Fig 4b Fig 5a Over-etching top Rejected foil Over-etching bottom Missing holes topMissing holes bottom

39. Fig 7a Fig 7b Fig 7c Typical heavy “over-etching”, leading to continuous discharges (rejected foils)

40. 40 G. Bencivenni (LNF-INFN, Italy) GARFIELD Simulations

41. 41 G. Bencivenni (LNF-INFN, Italy) GARFIELD Simulations

42. Test results from 90 Sr source 42 G. Bencivenni (LNF-INFN, Italy) The profile of the source in 6 different positions is reconstructed by triggering the DAQ with a clock signal This fast test allows to check the cabling ξ ρ

43. Test results from cosmic rays events 43 G. Bencivenni (LNF-INFN, Italy) ρ ξ Events selected using External Tracking provided by 3 Planar Triple- GEM

44. 44 G. Bencivenni (LNF-INFN, Italy) Pre-insertion test After the final cabling each layer was tested peaks related to cosmic rays noise correlated to the V strips length X view V view Noise run with Layer 3 & 4

45. IT integration 45 G. Bencivenni (LNF-INFN, Italy) Faraday cage completed with a 18 μm shield connected to the PCB end caps Scintillators for cosmic-ray trigger mounted on a cylindrical rotating support for acquisition on different sectors of the Inner Tracker

46. IT integration 46 G. Bencivenni (LNF-INFN, Italy) Protection shell installation Interaction region movement Kapton protection on the HV boardsIR positioning