Transcript A Study of Micro Pattern Gas Detector Readout Options for

ILC TPC resolution studies with charge dispersion
in MPGDs with a resistive anode
Madhu Dixit
Carleton University & TRIUMF
IPNS KEK 14/05/2005
The International Linear Collider (ILC)
• 2007 - LHC ready to explore new physics and to search for Higgs,
supersymmetry, hidden new dimensions, etc.
• ILC critical to understanding LHC discoveries
– Detailed study of Higgs and SUSY particles
– Precision measurements
• Higgs e+ e- -> Z H -> l l X
• ∆Mtop≈ 100 MeV, ∆top≈ 2%
• ∆MZ & ∆MW ≈ 5 MeV (from 30 MeV)
• ∆(sin2) ≈ 10-5 (from 2·10-4)
• Global Design Effort (GDE) timeline:
–
–
–
–
2006 accelerator CDR
2008 accelerator TDR, experimental collaborations, detector CDRs
2009 - Detector TDRs, construction
2015 - Physics at ILC
• ILC tracker resolution ∆(1/pT) ~ 5 x10-5 (GeV/c)-1 (10 times better than at
LEP!)
M. Dixit
IPNS KEK 14/10/05
2
ILC tracker requirements
• Small cross sections  100 fb, low rates, no fast trigger.
• Higgs measurements & SUSY searches require:
– High granularity continuous tracking for good pattern recognition.
– Good energy flow measurement in tight high multiplicity jets.
– Excellent primary and secondary b, c,  decay vertex reconstruction.
• TPC is an ideal tracker for ILC.
– Momentum resolution goal (1/pT) ~ 5.10-5 (GeV-1) achievable with vertex
+ Si inner tracker + TPC with ∆(1/pT) ~ 2 x 10-4 (GeV-1)
• ILC TPC tracker goals:
– 200 track points with (r, ) = 100 m, (r, z) = 500 m
– 2 track resolution < 2mm in (r, ) and < 5 mm in (r, z)
– dE/dx resolution < 5%
M. Dixit
IPNS KEK 14/10/05
3
ILC detector concepts
Silicon tracker (B=5T)
TPC (B=4T)
TPC (B=3T)
SiW ECAL
SiW ECAL (medium)
W/Scint ECAL (large)
“SiD”
“LDC”
“GLD”
M. Dixit
IPNS KEK 14/10/05
4
ILC TPC (TESLA design)
cm

B=4T

E
M. Dixit
IPNS KEK 14/10/05
5
Conventional TPCs never achieve their potential!
Example:Systematic effects in Aleph TPC at LEP
TPC wire/pad readout
100 µm
•Average Aleph resolution ~ 150 µm
•About 100 µm best for all drift distances
• Limit from diffusion  (10 cm drift) ~ 20 µm;  (2 m drift) ~ 90 µm
M. Dixit
IPNS KEK 14/10/05
6
Micro Pattern Gas Detector (MPGD) Readout for ILC TPC
•Transverse diffusion sets the ultimate limit on TPC
resolution.
•ILC TPC resolution goals close to the diffusion limit.
•Wire/pad TPC resolution inherently limited by ExB &
track angle systematic effects.
•A TPC read out with a MPGD endcap could meet the
ILC resolution challenge if the precision of pad charge
centroid determination could be improved.
•What is the best achievable resolution with
conventional techniques?
M. Dixit
IPNS KEK 14/10/05
7
Micro Pattern Gas Detectors (MPGD)
Unlike wires, MPGDs have no preferred direction - negligible ExB effect
The Gas Electron Multiplier (GEM)
Micromegas
Drift region
Drift region
~ 50 m
~ mm
MPGDs achieve excellent  40 µm resolution with 200 µm wide pads.
Conventional wire readout TPCs use cathode pads of width ~ a few mm.
Proposed ILC TPC channel count ~ 1.5x106 with 2 mm wide pads.
Narrower pads would lead to increased detector cost & complexity.
M. Dixit
IPNS KEK 14/10/05
8
Worldwide R&D effort for ILC TPC
Ron Settles main coordinator
Europe
RWTH Aachen
DESY
U Hamburg
U Karlsruhe
UMM Krakow
MPI-Munich
NIKHEF
BINP Novosibirsk
LAL Orsay
IPN Orsay
U Rostock
CEA Saclay
PNPI St. Petersburg
America
Carleton U
Cornell/Purdue
LBNL
MIT
U Montreal
U Victoria
Asian ILC gaseoustracking groups
Chiba U
Hiroshima U
Minadamo SU-IIT
Kinki U
U Osaka
Saga U
Tokyo UAT
U Tokyo
NRICP Tokyo
Kogakuin U Tokyo
KEK Tsukuba
U Tsukuba
Other USA
MIT (LCRD)
Temple/Wayne
State (UCLC)
Yale
Large task list
Resolution studies.
Ion feedback studies.
Gas studies for better
resolution and low neutron
background.
Low mass field cage and
endcap.
High density low power
electronics.
Analysis and simulation
software.
ILC TPC R&D plans

1) Demonstration phase
•Continue work for ~1 year with small prototypes on mapping out parameter space,
understanding resolution, etc, to prove feasibility of an MPGD TPC. For Si-based
ideas this will include a basic proof-of-principle.

2) Consolidation phase
•Build and operate “large” prototype (Ø ≥ 70cm, drift ≥ 50cm) which allows any
MPGD technology, to test manufacturing techniques for MPGD endplates, fieldcage
and electronics. Design work would start in ~1/2 year, building and testing another ~
2 years.

3) Design phase
•After phase 2, the decision as to which endplate technology to use for the LC TPC
would be taken and final design started.
mm^2, B = 1T
Many groups are working on MPGD TPC
R&D. Prototype results from “Some
Results -Summer 2005” (Ron Settles)
Point resolution,
Micromegas
Saclay/Orsay/Berkeley
--Diffusion measurements
 (r,) < 100 m possible
B = 1T
1x10mm^2
pads
--At moment only achieved
for short drift (intrinsic )
for gain~5000 (350V mesh),
noise~1000 e
--ongoing effort…
Prototype results contd.
Point resolution
GEM
DESY group. (r,)
measured for GEMs with
2x6mm2 pads
B=4T
Victoria group has
achieved ~ 100 µm
resolution for short drift
distances with narrower
1.2x7mm2 pads.
Gas:P5
30cm
Prototype Results - cont.
Point resolution, GEM
--Example of (r,) measured at Aachen GEMs with 2x6mm2 pads by
comparing track position with a Si hodoscope.
--In general (also for Micromegas) the resolution is not as good as
expected from diffusion.
Ideas to improve the MPGD TPC resolution
•Narrower pads leading to increased complexity & a
larger number of readout channels.
•Disperse track charge after gas gain over a larger area
to improve pad centroid with wide pads.
•For the GEM, large transverse diffusion in the high Efield field in transfer and induction gaps provides a
natural mechanism to disperse the cluster charge.
•Measurements with prototype GEM-TPCs have shown
that this increased diffusion is insufficient in a high B
field for the ILC-TPC to achieve the resolution target
with ~ 2 mm pads.
•Explore other concepts to disperse the charge
Carleton setup for MPGD resolution studies with x rays
•Point source ~ 50 µm collimated 4.5 keV x rays.
•Aleph TPC preamps.  Rise= 40 ns,  Fall = 2 s.
•DAQ - 500 MHz Tektronix digital scope.
M. Dixit
IPNS KEK 14/10/05
15
An idea - Measure the induced signals in a GEM
GEM
Anode pads
Proportional wire
Cathode pads
• Short ~ 200 ns signal
We measure x  y  70 µm
•But this technique requires
expensive high frequency pulse
shape sampling electronics
M. Dixit
IPNS KEK 14/10/05
16
Another idea- Position sensing from
charge dispersion in MPGDs with a resistive anode
Analogy:
Charge division is used to measure the avalanche position
on a proportional wire.
Deposit point charge at t=0
Telegraph equation (1-D):
L  2Q Q
1  2Q


2
R t
t
RC x 2
Solution for charge density
on the wire(L ~ 0)
RC x RC
4t
Q(x,t) 
e
4 t
2
Generalize the concept of 1-D charge division to 2-D

M. Dixit
IPNS KEK 14/10/05
17
Charge dispersion in a GEM with a resistive anode
M. Dixit
IPNS KEK 14/10/05
18
Equivalent circuit for currents in a GEM with an
intermediate resistive anode
Current generators
Resistive anode foil
Signal pickup pads
Pad amplifier
M. Dixit
IPNS KEK 14/10/05
19
A photon event in the resistive anode GEM test cell
Charge cluster size ~ 1 mm ; signal detected by ~7 anodes (2 mm width)
M. Dixit
IPNS KEK 14/10/05
20
Improving resolution with charge dispersion in a MPGD
with a resistive anode M.S.Dixit et.al., Nucl. Instrum. Methods A518 (2004) 721.
•Modified GEM 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.
Equation for surface charge
density function on the 2-dim.
continuous RC network:
(r)
 1 2 1  
t

Q
 2 

RC r
r r 
 (r,t) 
M. Dixit
RC
2t e
(r,t) integral
over pads
r 2 RC
4t
mm
IPNS KEK 14/10/05
r / mm
ns
21
Charge dispersion signal for a GEM
Simulation versus measurement
(2 mm x 6 mm pads) Collimated ~ 50 m 4.5 keV x-ray spot on pad centre.
Detailed simulation includes effects of, longitudinal & transverse diffusion,
gas gain, detector pulse formation, charge dispersion & preamplifier rise and
fall time effects. For tracks, include effects of unequal primary clusters.
Difference = induced signal (not included in
simulation) studied previously:
MPGD '99 (Orsay), LCWS '00
Primary signal: Fast large
amplitude main pulse on charge
collecting pad. Simulated primary
pulse is normalized to the data.
M. Dixit
Secondary signal: The dispersion pulse on
the neighboring pad is slower & smaller.
Simulated secondary pulse normalization
is the same as for the primary.
IPNS KEK 14/10/05
22
Resistive anode Micromegas
530 k/ Carbon loaded Kapton resistive anode was used with GEM. This was
replaced with more uniform higher resistivity 1 M/ Cermet for Micromegas.
Drift Gap
Al-Si
Cermet on
mylar
MESH
Amplification Gap
50 m pillars
M. Dixit
IPNS KEK 14/10/05
23
Charge dispersion signals in Micromegas
Single event (2 mm wide pads)
2nd neighbor (note different scale)
2 x 4 channel Tektronix
X-ray spot centred on pad 2
Two 1st neighbors
Ar/CO2 90/10, Gain ~ 3000
1st neighbor peak ~ 100 ns
after the primary pulse peak
Slow rising 2nd neighbor pulse
~ 25 MHz digitization could
Primary signal
replace pulse shape sampling
M. Dixit
IPNS KEK 14/10/05
24
GEM pad response function for collimated x rays
Simulation versus measurement
Ionization from 50 m
collimated x-rays.
2x6 mm2 pads
(Solid line)
Scan across width
Pad
22
Pad
23
Pad
24
Measured PRF deviates
from simulation due to
anode RC nonuniformities.
M. Dixit
IPNS KEK 14/10/05
25
Resistive anode double-GEM spatial resolution
Collimated ~ 50 m x-ray spot
2x6 mm2 pads
•GEM resolution ~ 70 m.
•Similar resolution measured for a Micromegas with
a resistive anode readout using 2 mm x 6 mm pads
M. Dixit
IPNS KEK 14/10/05
26
A fringe benefit: High Micromegas gain with a resistive anode
Argon/Isobutane 90/10
Resistive anode suppresses sparking stabilizing Micromegas
Extremely high gains without breakdown are possible
M. Dixit
IPNS KEK 14/10/05
27
Carleton cosmic ray test MPGD-TPC
•15 cm drift length with GEM or
Micromegas readout
•B=0 (so far)
•Ar:CO2/90:10 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.
M. Dixit
The resolution was next measured
with a charge dispersion resistive
anode readout with a double-GEM
& with a Micromegas endcap.
IPNS KEK 14/10/05
28
Simulation - GEM TPC cosmic event with charge dispersion
(track Z drift distance ~ 67 mm, Ar/CO2 90/10 gas)
Detailed model
simulation including
longitudinal &
transverse diffusion, gas
gain, detector pulse
formation, charge
dispersion & preamp rise
& fall time effects.
2x6 mm2 pads
Simulation
Data
Centre pad amplitude used for normalization - no other free parameters.
M. Dixit
IPNS KEK 14/10/05
29
The pad response function (PRF)
• The PRF is a measure of signal size as a function of track
position relative to the pad.
• For charge dispersion non charge collecting pads have signals
in contrast to conventional direct charge readout.
• Unusual highly variable charge dispersion pulse shape; both
the rise time & pulse amplitude depend on track position.
• We use pulse shape information to optimize the PRF.
• The PRF can, in principle, be determined from simulation.
• However, system RC nonuniformities & 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
which was used for calibration. The remaining data was used
for resolution studies.
M. Dixit
IPNS KEK 14/10/05
30
GEM & Micromegas PRFs for TPC track
Ar:CO2 (90:10) 2x6 mm2 pads
The pad response function maximum for longer drift
distances is lower due to Z dependent normalization.
GEM PRFs
Micromegas PRFs
Micromegas PRF is narrower due to the use of higher
resistivity anode & smaller diffusion after avalanche gain
M. Dixit
IPNS KEK 14/10/05
31
PRFs with the GEM & the 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 its FWHM (z) & base
width (z).
• PRFs determined from the data have been fitted to a
functional form consisting of a ratio of two symmetric
4th order polynomials.
(1 a2 x 2  a4 x 4 )
PRF[x,(z),,a,b] 
2
4
(1 b2 x  b4 x )
a2 a4 b2 & b4 can be written down in terms of 
and  & two scale parameters a & b.
M. Dixit
IPNS KEK 14/10/05
32
Track fit using the the PRF
Track at: xtrack= x0+ tan()yrow
(Ai  PRF i ) Ai 2
2 =   []
rows i=pads
6 mm
2 mm
M. Dixit
Determine x0 &  by minimizing
2for
the entire event

One parameter fit for xrow (track
position for a given row) using 
Bias = Mean of residuals (xrow-xtrack)
as a function of xtrack
Resolution =  of track residuals for
tracks with || < 5
IPNS KEK 14/10/05
33
Bias corrections with GEM & with Micromegas
Initial bias
Initial bias
Remaining bias after correction
M. Dixit
Remaining bias after correction
2 mm pads
2 mm pads
GEM
Micromegas
IPNS KEK 14/10/05
34

What is the diffusion limit of resolution for a gaseous TPC?
Resolution depends on electron statistics.
Electron number N fluctuates from event to event.
.
2
0 includes noise & systematic effects.
C
2
2
d z
x  0 
Cd = diffusion constant; z = drift distance
N eff
Neff  <N> the average number of electrons
= 1/<1/N> the inverse of average of 1/N
Gain fluctuations also affect Neff
M. Dixit
IPNS KEK 14/10/05
35
Simulation for the effective number of electrons for resolution
2 mm x 6 mm pads - Ar/CO2 90/10
Cosmic ray
momentum spectrum
Mostly
muons at
dE/dx in Argon
Measured pad pulse
height distribution
Total
sea level
Ionization
•Statistics of primary ionization & cluster size distribution.
•dE/dx dependence on momentum.
. •Account for track angle & detector acceptance effects.
•Use simulation to scale measured pulse heights to electron number.
•Neff = 1/<1/N> determined from pulse height distribution.
•Neff ≈ 38.9  10% (Naverage= 57)
M. Dixit
IPNS KEK 14/10/05
36
Measured TPC transverse resolution for Ar:CO2 (90:10)
R.K.Carnegie et.al.,
NIM A538 (2005) 372
R.K.Carnegie et.al.,
to be published
CD 2
0 
z
Ne
2
Unpublished
[Neff = 38.9]
Compared to conventional readout, resistive readout gives better resolution for the
GEM and the Micromegas readout. The z dependence follows the expectations from
transverse diffusion
& electron statistics.

M. Dixit
IPNS KEK 14/10/05
37
Summary
• TPC with MPGD readout is a very well suited technology for
the ILC.
• Traditional readout has difficulty achieving TPC resolution
goal, unless narrower pads are used.
• With charge dispersion, the cluster charge can be dispersed in a
controlled way such that relatively wide pads can be used
without sacrificing resolution. With such a readout system, it
may be feasible to achieve ILC TPC resolution goal with
relatively wide pads both for the GEM and the Micromegas
readout.
• With R&D, ILC TPC resolution goals appear within reach.
• Beam test at KEK (next week) important step in developing
the MPGD readout for the ILC TPC.
M. Dixit
IPNS KEK 14/10/05
38