Development of Large Area and of Position Sensitive Timing RPCs A.Blanco, C.Finck, R. Ferreira Marques, P.Fonte, A.Gobbi, A.Policarpo and M.Rozas LIP, Coimbra, Portugal. GSI, Darmstadt, Germany Univ. de Coimbra, Portugal. Univ. de Santiago de Compostela, Spain. ISEC, Coimbra, Portugal. 21
Plan of the Presentation What are timing RPCs? Earlier development of timing RPCs for ALICE-TOF. Recent developments for medium and low multiplicity applications. Development of very high frequency front-end electronics. A large area timing RPC (.16 m 2 ). Position sensitive single-gap timing RPCs VCI 2
Why timing RPCs? What are timing RPCs? Resolution: 1-3 ps 12 m 2 16, channels (determined by occupancy) standard PM-based TOF clearly impossible timing RPCs But many other experiments also have or will have large TOF systems: STAR@RHIC, BNL: 6 channels FOPI@GSI: 6 channels HARP@CERN PHENIX... (published specifications)
Timing RPCs general features (many configurations are possible) What are timing RPCs? -HV R R Gaseous detectors made with flat, resistive or metallic+resistive electrodes. Several (~4) thin (~.3 mm) gas gaps Atmospheric pressure operation Both charge and time readout Offline correction for time-charge correlation. ADC TDC 1-2 ns 1 ps Time-charge correlation
Earlier work done for the ALICE-TOF project Development of glass timing RPCs A CERN, Coimbra, ITEP collaboration 4-gap glass-metal RPC cells Multichannel module (32 ch.)
Earlier work done for the ALICE-TOF project Main results on single chambers (taken with 7 Gev/c negative pions at CERN) Time resolution (with optimized electronics) σ = = 2 2 68 49 47 ps =47 ps Resolution of the reference counter 3σ ε = 99.5 % for MIPs Main results on the multichannel module [ NIM A 449 (2) 295 ] σ = 88 ± 9 ps ε = 97 ±.5 % Crosstalk < 1% [ CERN-EP/99-166 ]
Development of new front-end electronics We wanted: precise timing easily available inexpensive stable VCI 7
Very high frequency front-end electronics based on commercial chips Input signal (experimental) i=i e st s=8.9 GHz Development of new front end electronics Pre-amplifier based on the INA-5163 chip (HP/Agilent) 2.5 GHz bandwidth 2 db power gain 3 db noise figure Amplifier-discriminator-delay module 1 ps σ resolution above 1 fc Time resolution per channel (s/ 2) (ps) Realistic tests using chamber-generated signals 7 6 5 4 3 2 1 RPC at Threshold=12.5 fc RPC at threshold=25 fc RPC at threshold=5 fc Pulser at threshold=25 fc 1 1 1 Signal fast charge per channel (fc)
A large area timing RPC (.16 m 2 ) Why? Study the dependence of the performance on the pad area. May be useful (HARP experiment at CERN). Advantages when compared with scintillators? good time resolution insensitivity to the magnetic field compact mechanics VCI 9
Detector structure Large area timing RPC 4-gap device Ordinary 3 mm window glass Copper strips Side view 1,6 m t1 t2 Top view 5 cm 4 timing channels Active area = 1 cm 16 cm =.16 m 2 (4 cm 2 electronic channel)
Cross section and connection schematics Large area timing RPC HV 1 GΩ 1 GΩ 1 MΩ HV distribution network... 2.2 nf Detector 1.6 m Signal path Electrodes applied to the edge of the middle glasses to define the potential at HV/2 (no floating electrodes) Noiseless.3 mm glass fiber spacers The dark count rate was only 2 khz for the whole surface
Electronics assembly details Large area timing RPC
Typical distributions Charge distribution Events / 2 fc 1 3 3.5 pc 2 1 2 1 ε 98 % Large area timing RPC 1 1 2 4 6 Time distribution Events/2 ps 1 3 1 2 1 2 4 6 8 Fast charge (pc) 64 2-35 2 = 54 ps s= 63.4 ps Time resolution essentially independent from electrode size 6 4 2 ±1.5 σ fit 1 1-1 -5 5 1 3 ps tails 1-1 -5-3 3 5 1
Timing and efficiency across the strips Large area timing RPC 1% Glass edge Strip edge Strip B Strip A Strip edge Glass edge 9% 8% 7% Efficiency 6% 5% 4% 3% Size of the trigger region Strip A-Charge Strip B-Charge Strip A-Time Strip B-Time Strips A&B-Charge 2% 1% % -7-6 -5-4 -3-2 -1 1 2 3 4 5 6 7 Centerof the trigger region across the strips (cm) Resolution (ps s) 1 9 8 7 6 5 Strip A Strip A - all events Strips A+B Strip B Strip B - all events 4-7 -6-5 -4-3 -2-1 1 2 3 4 5 6 7 Centerof the trigger region across the strips (cm) Time resolution between 6 and 75 ps σ across the strips. But up to 9 % inter-strip crosstalk...
Timing and efficiency along the strips Large area timing RPC Time resolution (ps s) Time efficiency 1% 99% ε = 95 to 98 % 98% 97% 96% Strip A 95% Strip B 94% Strips A+B 93% -8-7 -6-5 -4-3 -2-1 1 2 3 4 5 6 7 8 1 9 σ = 5 to 75 ps 8 7 6 5 4-8 -7-6 -5-4 -3-2 -1 1 2 3 4 5 6 7 8 2.5% Center of the trigger region along the strips (cm) Center of the trigger region along the strips (cm) 3 ps tails 2.% 1.5% 1.% tails < 2 %.5%.% -8-7 -6-5 -4-3 -2-1 1 2 3 4 5 6 7 8 Center of the trigger region along the strips (cm)
Position resolution along the strips D t/2 (ns) Fit residuals (ns) 8 6 4 y=.79 x +.1 v=14.1 cm/ns 2-2 Strip A -4-6 Strip B -8-8 -7-6 -5-4 -3-2 -1 1 2 3 4 5 6 7 8.3 Very good linearity (± 1.4 cm) Large area timing RPC.2.1 -.1 -.2-8 -7-6 -5-4 -3-2 -1 1 2 3 4 5 6 7 8 Center of the trigger region along the strips (cm) Events/bin 4 35 3 25 2 15 1 5 5. cm σ X = 1.2 cm (.75% of detector length) -35-3 -25 TDC bins
Position sensitive single-gap timing RPCs Why? A concept for small ( 1 m 2 ) and accurate TOF counters in medium or high multiplicity applications. Aim for the best possible performance by correcting for position dependent effects (e.g. counter edges, mechanical defects) and by considering multilayer configurations. Study in detail the performance of single gaps and compare with 4-gap counters. It is generally useful for TOF measurements to have some position information from the TOF counter. Advantages when compared with scintillators? low cost very good timing accuracy (resolution + tails) insensitivity to the magnetic field compact mechanics VCI 17
Chamber construction Position sensitive single-gap counters Time only version Signal path lenght independent from avalanche position (avoid positiondependent time variations) HV Position sensitive version HV 4 cm Signal Time signal 2 mm thick black glass lapped to ~1µm flatness metal box (no crosstalk) 3 µm thick high ρ glass disk (corners) Well carved into the glass (avoid dark currents from the spacer) X right XY readout plane X left 1 strips for each coordinate at 4 mm pitch RC passive network out left Y-strips (on PCB) out right RC passive network X-strips (deposited on glass)
Typical distributions Charge distribution Events/bin 3 25 2 15 1 5 Position sensitive single-gap counters Inefficiency peak 1 4 ε = 75 % 1 2 1 5 1 2 4 6 8 1 Fast charge (a.u.) Time distribution 67 2-4 2 = 54 ps Time resolution of single-gap and 4-gap counters is similar! Events/1ps 1 3 1 2 1 1 Events=26186 σ = 66.6 ps (54 ps) 3σ Tails=1.9 % 3ps Tails=.36 % 3σ 15 1 5 ±1.5 σ fit -1-5-3 3 5 1 1-1 -5-3 3 5 1
Efficiency and time resolution Time resolution (ps s) 1 9 8 7 6 5 4 #2 (time only) #3 (time + XY) Position sensitive single-gap counters Efficiency Resolution 2.4 2.6 2.8 3 3.2 3.4 Applied Voltage (kv) 85 8 75 7 65 6 55 Efficiency (%) σ = 5 to 6 ps ε = 75 to 8 % No influence from XY readout
Position resolution Position sensitive single-gap counters Counter #4 Events/.5mm 15 1 5 Trigger edge (3 mm) Chamber edge edges 3 mm resolution 3 mm FWHM -25-2 -15-1 -5 5 1 15 2 25 Position along X (mm) Events/.5mm 15 1 5 4 mm strip pitch -25-2 -15-1 -5 5 1 15 2 25 Position along Y (mm)
Edge effects? Position sensitive single-gap counters Counter #1 All events Sigma=72.4 ps (6 ps) 3-Sigma Tails=3. % 3ps Tails=1.7 % Events/4ps 1 2 1 1 1-1 -5 5 1 Essentially no edge effects Center Sigma=74. ps (62 ps) 3-Sigma Tails=3. % 3ps Tails=1.5 % Edge & Corner Sigma=76.9 ps (66 ps) 3-Sigma Tails=2.9 % 3ps Tails=1.8 % Events/4ps 1 2 1 1 1-1 -5 5 1 Events/4ps 1 2 1 1 1-1 -5 5 1
Origin of timing tails? Position sensitive single-gap counters Counter #2 Y (mm) = events on 3σ tails Events/mm 2 No special incidence of tails on particular locations (edge for instance) X (mm) Events/.5mm 8 6 4 2-2 -1 1 2 Events/.5mm 6 4 2-2 -1 1 2 Tails/.5mm 15 1 5-2 -1 1 2 X (mm) Tails/.5mm 1 5-2 -1 1 2 Y (mm) right tails left tails 3σ band Events/1ps 1 3 Events=26186 σ = 66.6 ps (55 ps) 1 2 3ps Tails=.36 % 1 1 3σ Tails=1.9 % 3σ 1-1 -5-3 3 5 1
Correction of timing variations along the counter Position sensitive single-gap counters Counter #2 Y (mm) Variation ot the average measured time (ps) X (mm) σ = 67 ps (54 ps) 3σ Tails=1.9 % 3ps Tails=.36 % σ = 68 ps (55 ps) 3σ Tails=2. % 3ps Tails=.34 % 1 3 Events=26186 Events=766 Events/1ps 1 2 1 1 Events/1ps 1 2 1 1 1-1 -5-3 3 5 1 1-1 -5-3 3 5 1 There are no position dependent corrections to be made
Multilayer detector configurations (for small and accurate TOF systems) 4 layers of single-gap chambers Position sensitive single-gap counters Single layer of 4-gap chambers Layer ε = 75 %, σ = 6 ps + Events/1ps 4 1 1 3 1 2 1 1 3 σ Tails=1.4 % 3ps Tails=.% σ = 32 ps ε = 94.9 % Events/1ps 1 3 1 2 1 1 σ = 5 ps @ ε = 99 % 3 σ Tails=1.5 % 3ps Tails=.2% Events/1ps Events=51215 1-1 -5-3 3 5 1 1 5 π + 1% K @ 3 ps -3 3 5 π + 1% K @ 2 ps 1-1 -5 5 1 Events/1ps 4 2 tail-dominated -3 3 5 Events/1ps 1 5-3 3 5 Events/1ps 4 2-3 3 5
Multilayer detector configurations (for small and accurate TOF systems) 3 layers of double-gap chambers Position sensitive single-gap counters 4 layers of single-gap chambers Layer ε = 93 %, σ = 55 ps σ = 6 ps @ ε = 75 % 3 σ Tails=3.3 % 3ps Tails=.81% Events/1ps 1 3 1 2 1 1 3 σ Tails=1. % 3ps Tails=.% σ = 33 ps ε = 97.5 % Pure gaussian Events=455 1-1 -5 5 1 Events/1ps Events/1ps 1 3 1 2 1 1 Large tails Single layer 1-1 -5 5 1 1 4 1 3 1 2 1 1 3 σ Tails=1.6 % 3ps Tails=.% σ = 33 ps ε = 94.6 % Very small tails Events=5699 4 layers 1-1 -5 5 1
Conclusions After 2 years of development the timing RPC technology is now well established at resolutions between 5 and 8 ps σ in a variety of chamber configurations. Large electrode areas, up to 8 cm 2 per electronic channel, have been tested without significant performance degradation. Simultaneous time and position measurements are also possible without any performance degradation. Single-gap counters without measurable edge effects or position dependent time variations were demonstrated. Multilayer configurations can provide resolutions from 3 to 35 ps σ, essentially free of timing tails.