LHCb Muon System Technology Why has it changed
LHCb Muon System Technology Why has it changed ? Burkhard Schmidt, CERN • • • Outline: Overview of the Muon Detector Technologies Studies done with MRPCs Ageing studies of bakelite RPCs (talk PD 2 -2, G. Carboni) Ageing studies of MWPCs IEEE Symposium 2003, Portland Conclusions
The LHCb Muon Detector M 1 M 2 M 3 M 4 M 5 • 5 Muon stations, M 1 in front and M 2 -M 5 behind the calorimeters -> large variations in particle flux: 0. 2 k. Hz < Φ < 500 k. Hz • 435 m 2 of detector area have to be equipped Muon System is primarily used for L 0 muon triggering -> RPC-type technology seemed good choice for many regions -> Industrial production appealing However, particle fluxes are higher than in ATLAS and CMS -> Pushing the limits. IEEE. . . Symposium 2003, Portland
Muon System Technology Evolution TP: 02/1998 CPC 14% TDR: 05/2001 Addendum: 01/2003 MWPC 51% (up to 100 k. Hz/cm 2) MRPC: 86% RPC: 48% (up to 1 k. Hz/cm 2) up to 5 k. Hz/cm 2 ? : <1% MWPC 99. 9% GEM ? : 0. 13%
Comparison RPC - MRPC Multi-gap RPC: Single-gap RPC: V 0 d d’ ΔV g g’ g’ Veff g’ ΔV R= ρ·d/A V 0 d’ 0 V I=Φ·A·q g’ Veff = V 0 – 2·ρ·d·Φ·q ¾ V 0 ½ V 0 ΔV=0 ¼ V 0 0 V On intermediate plates: Q=const. -> ΔV = ρ·d·Φ·q Veff ΔV = ρ·d’·Φ·q ΔV = 0 Veff = V 0 – 2·ρ·d’·Φ·q Veff in single and multi-gap RPCs is the same (if d=d’) Multiple small avalanches in MRPCs have no advantage in terms of efficiency because the signal in the far gaps is reduced due to the weighting field. IEEE Symposium 2003, Portland
Comparison RPC - MRPC Rate Capability: SRPC MRPC 1. 8 k. Hz/cm 2 1. 9 k. Hz/cm 2 ρ ≈ 1010 cm ρ ≈ 1012 cm Rate capability is determined by the resistivity 11 cm We could not get thin melamine plates of ρ < 10 IEEE Symposium 2003, Portland
Experience with MRPC Problems encountered when operating MRPCs with dry gas: (resistive plates are of melamine-phenolic) • It changes the plate resistivity (from 6 x 1011 cm -> 6 x 1012 cm) • It causes large deformations: Melamine sheets (0. 8 mm) These problems could be reduced when flushing 50% of the gas mixture (C 2 H 2 F 4/ iso-C 4 H 10/SF 6 95/4/1) through water -> MRPC have been dropped quickly in LHCb, since they offered no advantages IEEE Symposium 2003, Portland
Aging Studies done with RPCs Test setup at Gamma Irradiation Facility (CERN) Measuring the bakelite resistivity Effects of aging on resistivity Flowing of humid gas Effects of aging on rate capability IEEE Symposium 2003, Portland
Test setup at GIF Source characteristics: 662 Ke. V from 740 GBq 137 Cs Counting rate: 1 3 k. Hz/cm 2 at 1 m from the source 2001: A in pos 1, B in pos 2002 -2003: both in pos 1 1: aging position 2: reference position 3: testbeam position The irradiation is measured by the accumulated charge per cm 2 (Qacc) we had previously shown that the resistivity of bakelite slabs is not affected by pure irradiation up to doses of 20 k. Gy Since the rate capability is inversely proportional to ρ, this has been the monitored quantity IEEE Symposium 2003, Portland
Measurement of resistivity A simple method has been developed to measure ρ continuously at GIF It requires the detector to be exposed to large flux of radiation The model is described in G. Carboni et al. NIM A 498 (2003) 135 It is based on the hypothesis that all the physical properties of an RPC must depend on Vgap = V 0 - RI S R depends exponentially on the temperature For bakelite we measured the value a = 0. 12 ± 0. 01 IEEE Symposium 2003, Portland d
Measurement of resistivity Current saturation with flux Current linearity with HV Prediction of the model: For fixed V 0 the current depends exponentially on the temperature through R IEEE Symposium 2003, Portland
Current vs time Current drawn at 10800 V corrected by temperature 8 months of irradiation 0. 4 C/cm 2 acc. charge (RPC A) In the same period RPC B was used as reference and accumulated only 0. 05 C/cm 2 I decrease by a factor 6 Qacc=0. 4 C/cm 2 jan 01 aug 01 ρ increase by the same factor IEEE Symposium 2003, Portland
ρ vs time 1999 -2001 RPC A RPC B Date Qacc (C/cm 2) ρ20 (1010 Ωcm) oct 99 0 <2 0 ~3 jan 01 0. 076 6. 6 ± 0. 5 - - mar 01 0. 11 8. 5 ± 0. 7 - - jul 01 0. 361 26 ± 2. 3 - - aug 01 0. 42 39 ± 4 0. 05 13 ± 1. 2 dec 01 0. 42 69 ± 6 - irradiating Remarks: Large increase for RPC A Evidence of increase not related to irradiation for both IEEE Symposium 2003, Portland
ρ vs time 2002 Test: both detectors now installed close to the source to measure ρ continuously Only ~0. 05 C/cm 2 accumulated charge (large ρ low current) Both detectors show a steady increase of ρ with time -> Drying up of bakelite (see also J. Va’vra’s talk yesterday DA 2 -1) IEEE Symposium 2003, Portland
ρ vs time 2003: humid gas no flow humid flow 1. 2% of vapor H 2 O added to the usual gas mixture Clear effect of humid gas, (see also ALICE 2002 results) but: on RPC B there is a sharp decrease of ρ on RPC A the effect is much reduced ρ rapidly resume to old values when dry gas flow is restored IEEE Symposium 2003, Portland
Effects on rate capability We define a RPC detector “capable” to stand a given rate if: efficiency > 95% (trigger requirement) at least 400 V plateau (safety requirement) HV < 11000 (streamer limitation) The detectors were tested at GIF by means of X 5 muon beam IEEE Symposium 2003, Portland
GIF test August 2001 RPC A Resistivity at 20 ºC ρA = 39 x 1010 Ωcm Rate capability at 20ºC ~ 640 Hz/cm 2 T = 25 °C IEEE Symposium 2003, Portland
GIF test July 2002 Resistivity at 20 ºC ρA = 110 x 1010 Ωcm Rate capability at 20ºC ~ 200 Hz/cm 2 T = 24. 5 °C IEEE Symposium 2003, Portland
Testbeam summary RPC A RPC B Date Temp (°C) ρ (1010 Ωcm) rate cap (Hz/cm 2) oct 99 23 <1 >3000 ~2 >3000 aug 01 25. 1 20 1150 - - jul 02 24. 5 65 350 45 380 Except for the 1999 data (affected by large uncertainties) the values are in agreement with the 1/ρ dependence of rate capability IEEE Symposium 2003, Portland
Conclusions on RPCs Aging effects on bakelite RPCs have been extensively studied for 3 years on two identical detectors with building resistivities around 1010Ωcm After 2 years of operation ρ had already increased to ~100 x 1010Ωcm reaching the value of ~200 x 1010Ωcm at the end of the third year Although irradiation contributes to the resistivity increase, we believe the effect is mainly related to dry gas flow Humid gas has been flowed with different response: RPC B shows a sharp decrease of resistivity, whereas in RPC A the effect is very much reduced Restoring dry gas flow has resulted again in fast resistivity increase Flow of humid gas doesn’t appear to be a practical method to recover detector performances Rate capability dropped from few k. Hz/cm 2 to about 200 Hz/cm 2 èLHCb dropped RPCs ! è This decision did not make the LHCb muon system more expensive ! IEEE Symposium 2003, Portland
Aging Studies done with MWPCs MWPC chamber characteristics Overview of aging tests done Results in terms of gain variations ΔG/G and Malter current Visible inspection of the chamber Next steps IEEE Symposium 2003, Portland
MWPC Characteristics HV HV Materials used for the chamber construction: • FR 4 (fire resistant fiber glass epoxy) Pitch: 2 mm Gap: 5 mm Wire: 30 µm HV: 2650 V Cathode field: 6. 2 k. V/cm Wire field: 262 k. V/cm Gain: 105 Total avalanche charge: 0. 74 p. C • Gold plated tungsten wire • Ni/Au plated PCBs • Araldite 2011 and Adekit A 145/50 • Natural rubber gasket for chamber closing • Epoxy glued kapton foil. Expected accumulated charges in 10 LHCb years: IEEE Symposium -> up to 1. 6 C/cm 2 on cathodes and 2003, 0. 3 Portland C/cm on wires
MWPC Aging studies Two long term aging tests have been performed: 1 st test at GIF in 2001, where charges of 0. 25 C/cm have been accumulated in 6 months. Open loop gas system has been used Gas mixture: Ar / CO 2 / CF 4 (40%, 50%, 10%) 2 nd test at ENEA Cassacia in June 2003, where charges of 0. 5 C/cm have been accumulated in 1 month (using the 25 k. Ci (!) Co source). One chamber has been in open loop, one in closed loop Gas mixture: Ar / CO 2 / CF 4 (40%, 20%) Parameters controlled : • Relative gas gain: In both tests one gap was used as reference gap to avoid complicated corrections for P and T variations. The reference gap was only for very short periods per day under HV and always flashed with the fresh gas. • Dark currents, including self-sustaining rest current following the beam off. The dark currents were measured with current monitors with a resolution of 1 n. A. IEEE Symposium 2003, Portland
Developments of currents with time B 1 and C 2 have been the reference gaps in the 2 chambers -> No visible deterioration of the gas gain IEEE Symposium 2003, Portland
Current ratio after/before irradiation Am Source tests: The ratios after/before of the source currents measured before and after irradiation averaged over the area ~360 cm 2 Gap S 1 S 2 A 1 Ratio(03/02) 1. 04 1. 08 1. 14 P, T corrected 1 1. 03 1. 05 Tests at GIF: Gap S 1 S 2 Current ratio 1. 18 1. 23 P, T corrected 1. 07 1. 11 Precision of the measurement: about 10% -> No aging in gaps S 1, S 2 and A 1. IEEE Symposium 2003, Portland
Rest current behaviour Rest currents in one of the chambers (in n. A): Date 11. 05. 03 28. 05. 03 HV (k. V) Gap S 1 Gap S 2 HV (k. V) Gap A 2 3. 1 9 6 2 17 3. 2 16 9 2. 1 21 3. 3 37 4 2. 2 30 33 2. 1 40 After GIF 08. 03 3. 3 6 Rest current: The currents were measured immediately after the source off has been switched off. IEEE Symposium 2003, Portland
Summary of Results from Cassacia Accumulated charges: Chamber Gap IC (C/cm) M 3 R 1 A, C, D 0. 43 S/A S 1 0. 52 S/A S 2 0. 42 S/A A 2 0. 38 No signs for aging from the measured parameters: Relative currents did not change. Malter currents were not observed. No variations in current ratios after/before Casaccia from measurements with GIF and with an Am-source. IEEE Symposium 2003, Portland
Analysis of Surfaces 1 st the wires. . . Strong FR 4 etching of open surfaces CF 4 is a phantastic terrible gas !! gas Bubbles under Kapton foil Detaching of the ground grid between the pads (due to FR 4 etching)
Analysis of Deposits on Cathodes Brownish deposit limited by the field wire Si Deposit reflecting wiring structure Presence of the deposit can provoke Malter current on the beam due to spikes of intensity. IEEE Symposium 2003, Portland
Effects in the Gas System Rotameter glass tube etching F Inner glass surface of a gas rotameter from closed loop is strongly eroded ( Flurine was found). O-rings (from NBR) used in valves of the gas system are damaged. IEEE Symposium 2003, Portland
Conclusions and next steps Conclusions: Although no drop in gas gain was visible after the Casaccia aging test, the materials exposed to CF 4 show strong surface etching. The effects are very similar for the chambers in the two gas system used in Casaccia (vented and re-circulation gas-system). The effect has been stronger the 2 nd aging test, where the gas mixture contained 20% CF 4, than in 1 st test, where 10% CF 4 have been used. Plans: Repeat the test at GIF in the coming months, where the test takes however 5 times longer than in Casaccia (source strength 25 k. Ci). New mixture with less/no CF 4. Both gas loops will be used. Control of O 2 and H 2 O and contamination is foreseen. IEEE Symposium 2003, Portland
- Slides: 30