Post on 19-Jan-2016
description
transcript
Beijing, 5 February 2007
Micromegas TPC prototype results & Electronics Developments
Madhu DixitCarleton University & TRIUMF
On behalf of LC TPC collaboration
ILC tracking review - Beijing 5 February, 2007
Beijing, 5 February 2007 Madhu Dixit 2
•ILC tracker goal: r ≤ 100 m including stiff 90° 2 m drift tracks •Anode wire/cathode pad TPC resolution limited by ExB effects•Negligible ExB effects for Micro Pattern Gas Detectors (MPGD)•TESLA TPC TDR : 2 mm x 6 mm pads (1,500,000 channels) with GEMs or Micromegas •LC TPC R&D: 2 mm pads too wide with conventional readout
•For the GEM ~ 1 mm wide pads (~3,000,000 channels) •Even narrower pads would be needed for the Micromegas
•The new MPGD readout concept of charge dispersion can achieve good resolution with ~ 2 mm x 6 mm pads.•R&D summary - mainly on the Micromegas TPC readout option
Micro-Pattern Gas Detector development for the ILC TPC
Beijing, 5 February 2007 Madhu Dixit 3
Micromegas - a parallel plate gas avalanche detectorMicromesh supported by ~ 50 m pillars above anode
Beijing, 5 February 2007 Madhu Dixit 4
50 cm diameter Micromegas, 50 cm max. drift distance 1024 read out pads, Star TPC 20 MHz 10 bit digitizers
2 T superconducting magnet
Berkeley Orsay Saclay cosmic ray TPC tests
1 mm x 10 mm pads
Beijing, 5 February 2007 Madhu Dixit 5
Beijing, 5 February 2007 Madhu Dixit 6
Beijing, 5 February 2007 Madhu Dixit 7
4 Gev/c beam tests at KEK with standard readout
(2.3 mm x 6.3 mm pads)
Beijing, 5 February 2007 Madhu Dixit 8
MP TPC with Micromegas - standard readoutKEK PS 4 Gev/c hadron test beam (2.3 mm x 6.3 mm pads)
Resolution limited by pad width at high magnetic fields, worst at short drift distances
Beijing, 5 February 2007 Madhu Dixit 9
•Disperse track charge after gas gain to improve centroid determination with wide pads.•For the GEM, large transverse diffusion in the transfer & induction gaps provides a natural mechanism to disperse the charge.•No such mechanism for Micromegas•The GEM readout will still need ~ 1 mm wide pads to achieve ~ 100 m ILC resolution goal
Charge dispersion on a resistive anode - a mechanism to disperse the MPGD avalanche charge. It makes position sensing insensitive to pad width.The technique works for both the GEM and the Micromegas
Improving MPGD resolution without using narrower pads
Beijing, 5 February 2007 Madhu Dixit 10
•Modified MPGD anode with a high resistivity film bonded to a readout plane with an insulating spacer.•2-dimensional continuous RC network defined by material properties & geometry.•Point charge at r = 0 & t = 0 disperses with time.•Time dependent anode charge density sampled by readout pads.
€
∂ρ∂t
=1
RC
∂2ρ
∂r2+
1
r
∂ρ
∂r
⎡
⎣ ⎢
⎤
⎦ ⎥
€
⇒ ρ(r, t) =RC
2t
−r2RC
4 te
ρ(r,t) integral over pads
ρ(r) Q
mm ns
Charge dispersion in a MPGD with a resistive anode
Equation for surface charge density function on the 2-dim. continuous RC network:
M.S.Dixit et.al., Nucl. Instrum. Methods A518 (2004) 721.
Beijing, 5 February 2007 Madhu Dixit 11
Micromegas resistive anode readout structure
Drift Gap
MESHAmplification Gap
25 m Al/Si coated mylar
Surface resistivity ~1 M/
50 m pillars
Beijing, 5 February 2007 Madhu Dixit 12
•15 cm drift length with GEM or Micromegas readout •B=0•Ar+10% CO2 chosen to simulate low transverse diffusion in a magnetic field.•Aleph charge preamps. Rise= 40 ns, Fall = 2 s. •200 MHz FADCs rebinned to digitization effectively at 25 MHz.•60 tracking pads (2 x 6 mm2) + 2 trigger pads (24 x 6 mm2).
The GEM-TPC resolution was first measured with conventional direct charge TPC readout.
The resolution was next measured with a charge dispersion resistive anode readout with a double-GEM & with a Micromegas.
Cosmic ray TPC tests with MPGD charge dispersion readout
Beijing, 5 February 2007 Madhu Dixit 13
Charge dispersion pulses & pad response function (PRF)
•Non-standard variable pulse shape; both the rise time & pulse amplitude depend on track position. •The PRF is a measure of signal size as a function of track position relative to the pad.•We use pulse shape information to optimize the PRF.•The PRF can, in principle, be determined from simulation. •However, system RC non-uniformities & geometrical effects introduce bias in absolute position determination.•The position bias can be corrected by calibration. •PRF and bias determined empirically using a subset of data used for calibration. Remaining data used for resolution studies.
Beijing, 5 February 2007 Madhu Dixit 14
GEM & Micromegas track Pad Response Functions Ar+10%CO2 2x6 mm2 pads
GEM PRFs Micromegas PRFs
Micromegas PRF narrower due to higher resistivity anode & smaller diffusion than in GEM after avalanche gain.
The pad response function (PRF) amplitude for longer drift distances is lower due to Z dependent normalization.
Beijing, 5 February 2007 Madhu Dixit 15
PRF[x,Γ(z),Δ(z),a,b] =(1+ a2x
2 + a4x4 )
(1+b2x2 +b4x
4 )
a2 a4 b2 & b4 can be written down in terms of Γ and Δ & two scale parameters a & b.
Track PRFs with GEM & Micromegas readout
The PRFs are not Gaussian.The PRF depends on track position relative to the pad.
PRF = PRF(x,z)PRF can be characterized by FWHM Γ(z) & base width Δ(z).PRFs determined from the data parameterized by a ratio of two symmetric 4th order polynomials.
Beijing, 5 February 2007 Madhu Dixit 16
Track fit using the the PRF
Determine x0 & by minimizing 2
for the entire event
2 mm
6 mm
Track at: xtrack= x0+ tan() yrow
16
2
2 ∑∑=
⎟⎟⎠
⎞⎜⎜⎝
⎛
∂−
=rows padsi i
ii
APRFA
Definitions:
- residual: xrow-xtrack
- bias: mean of xrow-xtrack = f(xtrack)
- resolution: standard deviation of residuals
Beijing, 5 February 2007 Madhu Dixit 17
Bias corrections for the GEM & for Micromegas
GEM Micromegas
2x6 mm2 pads
Initial bias Initial bias
Remaining bias after correction
2x6 mm2 pads
Remaining bias after correction
Beijing, 5 February 2007 Madhu Dixit 18
Transverse resolution (B=0) - Cosmic Rays Ar+10%CO2
R.K.Carnegie et.al., NIM A538 (2005) 372
K. Boudjemline et.al., NIM A - in press
€
02 +
CD2
Nez
A. Bellerive et al, LCWS 2005, Stanford
Compared to conventional readout, charge dispersion gives better resolution for the GEM and the Micromegas.
Beijing, 5 February 2007 Madhu Dixit 19
Track display - Ar+5%iC4H10 KEK 4 GeV/c hadrons Micromegas 2 mm x 6 mm pads B = 1 T
Zdrift = 15.3 cmmain pulse
Beijing, 5 February 2007 Madhu Dixit 20
Pad Response Function / Ar+5%iC4H10
Micromegas+Carleton TPC 2 x 6 mm2 pads, B = 1 T
xtrack – xpad / mm
norm
aliz
ed a
mpl
itude
0 < z < 0.5 cm 0 .5 < z < 1 cm 1 < z < 1.5 cm 1.5 < z < 2 cm 2 < z < 2.5 cm 2.5 < z < 3 cm
3 < z < 3.5 cm 3.5 < z < 4 cm 4 < z < 4.5 cm 4.5 < z < 5 cm 5 < z < 5.5 cm 5.5 < z < 6 cm
6 < z < 6.5 cm 6.5 < z < 7 cm 7 < z < 7.5 cm 7.5 < z < 8 cm 8 < z < 8.5 cm 8.5 < z < 9 cm
4 pads / ±4 mm
30 z regions / 0.5 cm step
Beijing, 5 February 2007 Madhu Dixit 21
Pad Response Function Ar+5%iC4H10
9 < z < 9.5 cm 9.5 < z < 10 cm 10 < z < 10.5 cm 10.5 < z < 11 cm 11 < z < 11.5 cm 11.5 < z < 12 cm
12 < z < 12.5 cm 12.5 < z < 13 cm 13 < z < 13.5 cm 13.5 < z < 14 cm 14 < z < 14.5 cm 14.5 < z < 15 cm
xtrack – xpad / mm
norm
aliz
ed a
mpl
itude
4 pads / ±4 mm
PRF parameters•a = b = 0
Δ = base width = 7.3 mm
Γ = FWHM = f(z)The parameters depend on TPC gas & operational details
Beijing, 5 February 2007 Madhu Dixit 22
Bias for central rows / Ar+5%iC4H10 B = 1 T
xtrack / mm (± 14 mm)
Res
idua
l / m
m (
± 0.
15 m
m)
bias before bias after
± 20 m
correction
row 4
row 5
row 6
Beijing, 5 February 2007 Madhu Dixit 23
Extrapolate to B = 4T Use DTr = 25 µm/cm Resolution (2x6 mm2 pads) Tr 100 m (2.5 m drift)
KEK beam test - Transverse resolution Ar+5%iC4H10
E=70V/cm DTr = 125 µm/cm (Magboltz) @ B= 1T
€
x = σ 02 +
Cd2 ⋅z
Neff
4 GeV/c + beam ~ 0°, ~ 0°
0= (52±1) mm
Neff = 220 (stat.)
Micromegas TPC 2 x 6 mm2 pads
•Transverse diffusion strongly suppressed at high B fields.Examples (4 T):DTr~ 25 m/cm (Ar/CH4 91/9) Aleph TPC gas ~ 20 m/cm (Ar/CF4 97/3)
Beijing, 5 February 2007 Madhu Dixit 24
5 T cosmic tests of charge dispersion at DESY COSMo (Carleton, Orsay, Saclay, Montreal) Micromegas
TPC
~ 50 m av. resolution over 15 cm (diffusion negligible)100 m over 2 meters appears feasible
Preliminary
Preliminary
Nov-Dec, 2006
DTr= 19 m/cm, 2 x 6 mm2 pads
Beijing, 5 February 2007 Madhu Dixit 25
Gain ~ 4700Gain ~ 4700 Gain ~ 2300Gain ~ 2300
Sample cosmic ray tracks - data taken at high & at low gains (B = Sample cosmic ray tracks - data taken at high & at low gains (B = 0.5 T)0.5 T)
The effect of lower gain on resolution
1
2
3
4
Beijing, 5 February 2007 Madhu Dixit 26
Gain ~ 4700 B=0.5 T
The resolution and 0 still good at low gain
Gain ~ 2300 B=0.5 T
COSMo TPC transverse resolution - Cosmic rays DESY magnet
Beijing, 5 February 2007 Madhu Dixit 27
Gain dependence on B field for Ar+5%C4H10
Micromegas gain constant to within ~ 0.5% up to 5 Tesla
Beijing, 5 February 2007 Madhu Dixit 28
Micromegas gain with a resistive anode Micromegas gain with a resistive anode Argon/Isobutane 90/10
Resistive anode suppresses sparking & improves Micromegas HV stability
Beijing, 5 February 2007 Madhu Dixit 29
Simulating the charge dispersion phenomenon
•The charge dispersion equation describe the time evolution of a point like charge deposited on the MPGD resistive anode at t = 0. •No standard pulse shape. For improved understanding & to compare to experiment, one must include the effects of:
•Longitudinal & transverse diffusion in the gas.
•Intrinsic rise time Trise of the detector charge pulse.
•The effect of preamplifier rise and fall times tr & tf.
•And for particle tracks, the effects of primary ionization clustering.
M.S.Dixit and A. Rankin, Nucl. Instrum. Methods A566 (2006) 281.
Beijing, 5 February 2007 Madhu Dixit 30
Charge dispersion signals for the GEM readoutSimulation vs. measurement for Ar+10%CO2 (2 x 6 mm2 pads) Collimated ~ 50 m 4.5 keV x-ray spot on pad centre.
Simulated primary pulse is normalized to the data.
Difference = induced signals (MPGD '99, Orsay & LCWS 2000) were not included in simulation).
Primary pulse normalization used for the simulated secondary pulse
Beijing, 5 February 2007 Madhu Dixit 31
GEM TPC charge dispersion simulation (B=0) Cosmic ray track, Z = 67 mm Ar+10%CO2
Centre pulse used for normalization - no other free parameters.
2x6 mm2 pads
Simulation
Data
Beijing, 5 February 2007 Madhu Dixit 32
Micromegas gas gain measurements at Saclay
100
1000
10000
100000
50 55 60 65 70 75 80 85 90 95 100
Mesh bias (kV/cm/atm)
Gain
Iso : 1%
Iso : 2%
Iso : 3%
Iso : 4%
Iso : 5%
CF4 : 3%, Iso : 1%
CF4 : 3%, Iso : 2%
CF4 : 3%, Iso : 3%
CH4 : 6.5%
CH4 : 8%
CH4 : 9%
CH4 : 10%
CH4 : 8%, CF4 : 3%
CH4 : 8%, CF4 : 5%
CH4 : 8%, CF4: 10%
CH4 : 10%, CF4 : 3%
CH4 : 5%, CO2 : 3%
CH4 : 10%, CO2 : 10%
CO2 : 10%
CO2 : 20%
CO2 : 30%
CO2 : 10%, Iso 2%
CO2 : 10%, Iso 5%
CO2 : 10%, Iso 10%
CF4 : 3%, CO2 : 1%
CF4 : 3%, CO2 : 3%
CF4 : 3%, CO2 : 5%
Iso : 2%, CH4 : 10%
Iso : 5%, CH4 : 10%
Iso : 10%, CH4 : 10%
Ethane 10%
Ethane 5%
Ethane 3.5%
Ethane 2%
Ethane 3.5% - CO210%Ethane 3.5% - CF4 3%
Ethane 3.5% - CF410%Ethane 3.5% - Iso 2%
Mixtures of gases containing argon
iC4H10
Gaz froids: CO2, CH4
C2H6
Mesh : 50 m gap of 10x10 size
(David Attie, EUDET Meeting, Munich 18 October, 2006)
Beijing, 5 February 2007 Madhu Dixit 33
Bulk Micromegas obtained by lamination of a woven grid on an anode with two photo-imageable films. The pillars hold the mesh on the whole surface: no frame needed.
New development: Bulk Micromegas (2004)I. Giomataris et al., CERN-Saclay collaboration, NIM A 560 (2006).
Beijing, 5 February 2007 Madhu Dixit 34
Advantages of Bulk Micromegas•Large surfaces at low cost (1000 € for a 34cm x 36cm detector)
•Almost no dead area•Very robust (robust mesh held everywhere)•Insensitive to dust (the mesh is dust-tight down to 30 µ)
•Good/excellent uniformity of the gap•Can be segmented and repaired
New bulk development: excellent resolution and stability
(KEK test Jan.07, 55Fe with no collimation)
pillar
Beijing, 5 February 2007 Madhu Dixit 35
Bulk Micromegas is cut with a 2 mm border
Bulk Micromegas development for T2K
Fully engineered projectMinimum dead space between panels ~ 8 mm
Beijing, 5 February 2007 Madhu Dixit 36
•T2K will have 3 TPCs•72 Micromegas modules •Total area ~ 9 m2
•124416 readout channels
Large panel Micromegas for T2K TPCs
Beijing, 5 February 2007 Madhu Dixit 37
Electronics
•Development for LP TPC based on ALICE TPC ALTRO digitizing electronics.•TPC requirements: highest flexibility in terms of pad geometry and shape of pad panels. •Design for 1 x 4 mm2 pads •Proposal: 32 channels modules, where each channel corresponds to an area of around 4 mm2.•2000 channels of 40 MHz ALTRO chips available. •Plans to acquire more channels - up to 10,000 for LP TPC tests.
3815 November 2006 Luciano Musa / CERN
anode wire
pad plane
drift region88s
PASA ADC DigitalCircuit
RAM
8 CHIPS (16 CH / CHIP) 8 CHIPS (16 CH / CHIP)
CUSTOM IC(CMOS 0.35m) CUSTOM IC (CMOS 0.25m )
DETECTOR Front End Card (128 CHANNELS)
557 568 PADS (3200 CH / RCU)
gat
ing
gri
d ALTRO
RCU
Front End and Readout Electronics
Captoncable
CustomBackplane
power consumption < 40 mW / channel
power consumption < 40 mW / channel
L1: 6.5s 1 KHz
L2: < 100 s 200 Hz
1 MIP = 4.8 fC
S/N = 30 : 1
DYNAMIC = 30 MIP
CSA SEMI-GAUSS. SHAPER
GAIN = 12 mV / fCFWHM = 190 ns
10 BIT
10 MHz
• LINEARIZATION
• BASELINE CORR.
• TAIL CANCELL.
• ZERO SUPPR.
MULTI-EVENT
MEMORY
Alice TPC
Beijing, 5 February 2007 Madhu Dixit 39
PASA designed for wire chamber pulses with long ion tails
Semi-Gaussian output pulse with ~200 ns integration produced for the digitizer. (base width ~ 450 ns)
Tail cancellation and shaping
Time (s)
Wire TPC charge pulse
PASA
0 2 4 6 8 10
Beijing, 5 February 2007 Madhu Dixit 40
PASA preamplifier-shaper not suitable for MPGD-TPC readout
(ns)0 200 400 600 800 1000
With 2 s charge preamp decay time
Charge pulse rise times will be much longer for ILC TPC tracks (dominated by charge collection). •Up to ~ 500 ns to collect the charge due to longitudinal diffusion and track angles.•ILC TPC resolution near statistical limit of diffusion.•Must collect over 90% of electrons for best resolution.•No optimum shaping time to achieve both good single hit and 2-track resolution•Better to digitize charge pulse directly without shaping.
GEM charge pulse - point x ray source
Beijing, 5 February 2007 Madhu Dixit 41
New preamp design to replace PASALuciano Musa
Beijing, 5 February 2007 Madhu Dixit 42
Table1. Specifications of the Programmable Charge Amplifier (PCA)
Conversion Gain Programmable in the range 10-30 mV/fC (see fig. 2)
Dynamics (Max Signal / noise) ~ 2000
Peaking Time Programmable in the range 30-100ns (see fig. 3)
Signal Polarity Programmable for positive and negative pulses
Power consumption < 8mW / channel (<1mW in standby mode)
Nr. Channels 32
Area 0.2 mm2/channel
Technology CMOS 130nm
Programmable gain Programmable peaking & decay times
Luciano Musa
Beijing, 5 February 2007 Madhu Dixit 43
Leif Jonsson Eudet meeting - Munich Nov. 2006
Beijing, 5 February 2007 Madhu Dixit 44
Summary•MPGD-TPC has difficulty achieving good resolution with wide pads •With charge dispersion, the charge can be dispersed in a controlled way. Wide pads can be used without sacrificing resolution. Charge dispersion works both for the GEM and the Micromegas.•At 5 T, an average ~ 50 m resolution has been demonstrated with 2 x 6 mm2 readout pads for drift distances up to 15 cm. •Electronics development on track•The ILC-TPC resolution goal, ~100 m for all tracks, appears feasible.