R

1
RD for a T2K TPC with GEM Readout

• Thorsten Lux
• IFAE/ UAB
• for the GEM Europe Group

2
Outline

• Who are we?
• Assembling of the GEM module
• First measurements
• Future RD plans
• Summary

3
Our Team
Geneva Bari Barcelona
Raphael Schroeter Nicolas Abgrall Anselmo
Cervera Alain Blondel Didier Ferrere Erik
Perrin Pierre Béné etc.
Federico Nova Ana Rodriguez Gabriel Jover Sofia
Andringa T.L. Federico Sánchez Alex Carles
Emilio Radicioni
For future measurements group from Valencia will
join!
4
Production of the GEM Module

• 12 unframed GEMs were delivered by CERN (LHCb)
• Quality check 500 V were applied to each GEM
  with current limit of 5 nA (2 GEMs were rejected)
• GEM foils were glued to frame
• Protection resistors were soldered
• HV supply cables were mounted
• best 6 GEMs were assembled in 2 GEM towers on the
  pad plane
• Mounting of the GEM module on the end-plate

5
Glueing of the GEMs
1
4
2
3
6
Glueing of the GEMs
1
3
12 hours of drying
2
7
Protection Resistors and Cableling
8
Assembling of the GEM Module
9
Assembling of the GEM Module
1
4
3
2
10
Basics of a GEM Tower

• drift field 60-110 V/cm (P5) and 140 V/cm
  (ArCO2)
• transfer/ induction field set to 2000 V/cm
• GEM voltage 310-320 V (P5) and 320 (-330 V)
  (ArCO2)

cathode side
anode side
11
The Setup

• 2 x 3 GEM structures
• 12 ALICE boards
• approx. 1500 readout channels (48 rows, 31
  columns)
• staggered pads
• pad size 8 x 8.3 mm2
• Gases P5 and magnetic field (up-to 0.4 T) and
  ArCO2 (9010)
• Cosmic muons

12
Measurements with P5

• measurements beginning of November
• simpler gas mixture (ArCH4 955)
• idea by varying the magnetic field to simulate
  different pad sizes
• but

• Huge field distortions ( of the order of cm)
• 1-2 sparks per hour (even without voltage across
  GEMs)
• only inclined tracks
• Electronic problems (in some runs)

13
Field Distortions
Simple 2D Maxwell simulation shows reason Guard
ring much to narrow!
14
A few Results for P5
Gain? 3?106/(2 ?102) ? 1.5?104
vd ? 3.9 ?s/cm vdth ? 4.0 ?s/cm
15
Improvement of the Setup
2D Maxwell simulation (potential)
improved design
shield
16
Measurements with ArCO2

• 2 measurement seasons 12/2005 and 01/2006
• no magnetic field
• due to slow gas, the measurements were performed
  in chamber slices
• improved setup -) no field
  distortions -) perpendicular
  alignment
• But

17
Loss of one Sector

• shortly after the first tracks a spark occurred
  during ramping up the GEM module
• consequence short between both sides of GEM1 of
  the top GEM tower ( few Ohms)
• during the spark there was no voltage
  difference across GEM1 but only between pad plane
  and GEM1
• operation of bottom GEM tower was still possible
  (and there was no spark at all afterwards)
• ? data taking with only one GEM tower

18
Broken GEM

• after 10 days of data taking with 1 sector the
  detector was dismantled
• the GEM module was brought back to the clean
  room and step by step dismantled
• special emphasis on broken GEM and pad plane
  below

19
Broken GEM

• some small spark remains seen on the pad plane
• some spots due to etching process on GEM surface
  (LHCb test sample)
• but no obvious reason for the short was detected
  
• after inspecting the short disappeared and the
  GEM could be ramped up again to 500 V (but was
  replaced)
• ? Most probably tiny particle between GEM and pad
  plane

Quality test procedure has to be reviewed
20
After the Re-assembling

• wonderful straight tracks over both sectors
• whole procedure was done within 2 days by 2
  persons
• and the best no sparks at all afterwards!
• data will be analysed soon (data taking ended
  last Friday !)

front view
side view
21
Pulse Shape and Drift Velocity
Pulse shapes for a 2 hit cluster
vd ? 0.92 ?s/cm vdth ? 0.9 ?s/cm
85 ?m delay
75 ?m delay
22
Attachment
2 runs with and without delay directly taken one
after the other
12.12.2006
15.12.2006
2 runs with 3 days and 2 gas bottle changes in
between
23
Gas Gain Dependence
320 V

• runs with different GEM voltage 320 V and 330 V
• mean cluster charge increases from 900 adc to
  1800 adc
• GEMs were stable during both runs

330 V
24
Point Resolution Method
Not using the likelihood method
Points reconstructed in the pad centre
25
Point Resolution Method

• parameterisation of the PRF as function of the
  ratio (charge width/ pad width)
  ( from ILC TPC group Hamburg)
• input from MAGBOLTZ diffusion coefficient and
  defocussing term (diffusion within GEM tower)
• taking z position into account correction to
  cluster position

26
Overall Residual Distributions
27
Measured PRF
28
Track Distributions
Track angles
?2
29
Point Resolution

• Cuts
• Only 2 and 3 hit clusters
• ?track lt 0.1 rad
• Clusters included in the fit
• Refitting of the track

Preliminary
RMS
?420 ?m
30
Residual Distributions
31
Future RD Plans

• try to fully exploit advantage of GEMs with
  respect to resolution ? ratio charge cloud with
  to pad width important
• might have nice side effects -) closer
  to drift velocity peak ? not so sensitive to
  temperature effects
  -) less sensitive to oxygen
  impurities
• -) etc.
• measurements with another pad plane ( pad size,
  geometry)

32
Summary

• with a great teamwork we gained a lot of
  experience!
• a GEM module was succesfully constructed,
  assembled and operated
• one GEM was (part time) damaged most probably
  tiny particle was overseen between GEM and
  padplane ? the quality test procedure has to be
  improved
• afterwards the GEM module was absolutely stable
  during operation (no sparks at all)
• the best point resolution was measured for ArCO2
  to 420 ?m
• ArCO2 is a quite delicate gas mixture and we
  want to investigate and test alternative gas
  mixture for the T2K TPC
• data analysis ongoing
• measurments to exploit the full advantage of
  GEMs due to the defocussing in the GEM tower are
  planned for the future (with magnetic field)