Toward a TPC for the ILD
Dan Peterson
Laboratory for Elementary-Particle Physics, Cornell University
What is the ILD
What is a TPC
What is so special about the ILD TPC
Cornell contribution: endplate
20110714-LEPP-Peterson
1
Cornell is part of an international collaboration, LCTPC,
which has the goal of developing a TPC for a detector at ILC , or CLIC
most active
Cornell has responsibility for developing the mechanical structure for the endplate
for the ILD TPC and the collaborative Large Prototypes LP1, LP2.
R&D has been ongoing for about 5 years.
20110714-LEPP-Peterson
2
International Linear Collider (ILC) is
an e+e- machine, allowing
precision measurements by
directly accelerating the parton,
0.5 TeV CoM,
31 km in length,
2 detectors (“push-pull”): SiD, ILD .
The latest schedule:
TDR due early 2013
Site decision 2016
Oddly, construction is not on the map;
that depends on the “decision to proceed”.
20110714-LEPP-Peterson
3
The ILD (International Large Detector)
has a TPC for the central tracker
with
outer radius 1808 mm ,
inner radius 329 mm ,
half length 2350 mm .
The other detector,
SiD (Silicon Detector)
has, as its main feature,
no gaseous tracking.
20110714-LEPP-Peterson
4
What is a TPC ?
Time Projection Chamber
The Z projection of the track is measured in time
.
cathode open field cage anode
magnetic field parallel to electric field
track goes through
leaves a trail of ionization in the gas
electrons drift to anode
magnetic field reduces diffusion,
shape of trail of the track remains intact
gas amplification of drifted electrons
signals measured on pads
MWPC and MPGD gas amplification
20110714-LEPP-Peterson
5
ALICE, at LHC, has a TPC for the central tracker.
The physics goals require measurement of
thousands of tracks.
Efficiency goal, in very dense events, is ~90% .
As one can see on this cover figure,
many tracks are missed.
The Alice TPC has MWPC gas amplification.
There are 560k pads
with 2 endplates, 250 cm radius;
pad size ~ 0.7 cm2 .
This is typical for a MWPC TPC.
This is sufficient for the resolution goal,
σ(p)/p=.025 at 4GeV,
translated: σ(1/p) = 6 x 10-3 /Gev.
20110714-LEPP-Peterson
STAR
6
The MWPC gas-amplification TPC ,
like that implemented in ALICE, and STAR at RHIC,
is well-matched for studies of heavy ion collisions.
But, precision studies at the ILC require unprecedented resolution and efficiency.
The resolution and efficiency demands can be met with a
Micro Pattern Gas Detector (MPGD) readout, GEM or Micromegas.
Maintaining the resolution places unprecedented demands on the
mechanical endplate.
(explained on next few slides)
20110714-LEPP-Peterson
7
The precision and stability goal is ultimately based on
maintaining the Higgs mass measurement.
Ref. ILD LOI section 3.2.1.2
ILD momentum resolution goal (system) , (1/pT) = 2 x 10-5/GeV,
is achieved with a TPC-only resolution,
σ(1/pT) =10-4/GeV (1/60 of ALICE) .
Ref. ILD LOI section 3.3.1
The Higgs mass measurement in e+e- -> ZH, H -> μ+μis affected by the momentum resolution, beam energy spectrum,
and backgrounds.
The width (at generator level) is 730 MeV,
while the width (after track reconstruction) is 870 MeV;
thus, the contribution from tracking errors is 473 MeV.
σ(1/pT) is at the significance threshold.
Ref. ILD LOI section 4.3.2 (and references 86 and 87 of the ILD LOI)
Maintaining the goal σ(1/pT) requires that uncertainties in TPC
measurement locations, due to magnetic field uncertainties,
be limited to ~50μm.
Achieving the magnetic field certainty requires decoupling the
track-based magnetic field calibration from the precision
mechanical calibration. And this requires a precision survey and
stability of the endplate components to a similar accuracy: 50 μm.
20110714-LEPP-Peterson
8
Jet energy resolution and PFA:
from Klaus Mönig,
Vienna, Nov 2005
dE/E = 60% / E1/2
dE/E = 30% / E1/2
There are processes where WW and ZZ must be separated without beam constraints
(example e+e-  nnWW, nnZZ )
The requires a jet energy resolution of about dE/E = 30% / E1/2
Particle Flow Analysis (PFA) achieves this resolution goal.
Within a jet, charged energy is measured by the tracking;
remaining neutral energy is measured in the calorimeter.
Efficiency: Calorimeter hits associated with charged tracks must be matched and identified.
Material: Conversion and interactions in the endplate must be minimized.
20110714-LEPP-Peterson
9
Goals of the ILD TPC endplate
Micro Pattern Gas Detector (MPGD) readout module design requirement driven by the momentum resolution goal
MPGD gas-amplification provides a more localized signal;
the pad size is 0.1 cm2 , 1/7 of ALICE.
Modules must provide near-full coverage of the endplate.
Modules must be replaceable without removing the endplate.
Rigid – requirement is driven by the momentum resolution goal
limit is set to facilitate the de-coupled alignment of
magnetic field and
module positions.
Precision and stability of x,y positions < 50μm .
Low material - limit is set by ILD endcap calorimetry and PFA, total 25% X0
readout plane, front-end-electronics, gate 5%
cooling
2%
power cables
10%
mechanical structure
8%.
Thin – limit is again set by ILD endcap calorimetry and PFA
ILD will give us 100mm of longitudinal space
between the gas volume and the endcap calorimeter.
20110714-LEPP-Peterson
10
In 2008, Cornell constructed two endplates for the LCTPC Large Prototype (LP1).
Inside the chamber
Outside the chamber
The endplate construction was developed to provide the precision required for ILD,
precision features are accurate to ~30 μm,
but not to meet the material limit specified for the ILD TPC;
the bare endplate has mass 18.87 kg over an area of 4657 cm2,
(mass/area) / (aluminum radiation length (24.0 g/cm2) ) = 16.9% X0,
2x the goal.
Accuracy is achieved with a 5-step machining/stress-relief process developed at Cornell.
20110714-LEPP-Peterson
11
These were shipped August 2008 and February 2010,
and are currently in use in the LP1 TPC at DESY.
20110714-LEPP-Peterson
12
Developed over the past year,
the ILD endplate design is a space-frame
and shown here as the solid model used for
the Finite-Element-Analysis (FEA).
(inside view)
This is the “equivalent-plate” design space-frame; the separating members are thin plates.
This design has rigidity and material equivalent to a strut design, which has also been studied.
(More on that later.)
This model has a full thickness of 100mm, radius 1.8m, and a mass of 136 kg.
The material thickness is then 1.34g/cm2, 6% X0. This meets an ILD goal.
20110714-LEPP-Peterson
13
This ILD endplate design results from several studies
described in the following slides:
ILD endplate model: FEA of the model
LP1 current endplate: measurements and FEA of models
LP2 space-frame designs: FEA of models,
small test beams:
FEA and measurements.
Future studies will include
construction and measurements
of the new LP2 endplate.
“strut” space-frame design
LP1 2008 endplate
20110714-LEPP-Peterson
“equivalent plate”
space-frame design
small test beams
that represent
one diameter of the
LP2 endplate
14
FEA calculations of deflection and stress (stress is not shown)
Endplate Support:
outer and inner field cages
deflection=0.00991 mm/100N
100N is the force on LP1 due to 2.1 millibar overpressure
(area of ILD)/(area of LP1) =21.9
deflection for 2.1 millibar overpressure
on the ILD TPC endplate (2200N)
= 0.22 mm (220 μm) .
This is deflection; we are ultimately
interested in lateral precision and stability.
rigidity vs thickness
It was expected that the strength could be
improved by asking ILD for more longitudinal space.
However, the improvement with thickness
diminishes (deviates from power law) above 100mm.
Note small buckling in inner layer.
Rigidity is improved with modest increase
in the back-plate thickness.
20110714-LEPP-Peterson
15
In the first validation of the FEA,
deflection of the current LP1 endplate was measured and
compared to the FEA.
The load of 100N (22lbs) was placed
“uniformly” in the center module location.
Deflection, measured across 2 lines, agrees on average.
location number
7
6
5
20110714-LEPP-Peterson
4
3
2
1
0
-1
“deflection 1”
-2
-3
-4
Stress is <1% yield.
-5
-6
-7
“deflection 2”
16
Various ideas were considered for the ILD endplate.
Models of LP1-like endplates were tested with the FEA
mass
kg
LP1
18.87
16.9
33
1.5
8.93
8.0
68
3.2
Al 7.35
C 1.29
7.2
(68-168)
(3.2-4.8)
Al 7.35
6.5
168
4.8
8.38
7.5
23
Lightened
Al-C hybrid
(channeled plus fiber)
Channeled
material deflection stress
%X0 microns
Mpa
(yield: 241)
Space-Frame
4.2
(strut or equivalent plate)
Study of the solid models predicts
a favorable rigidity/material for the space-frame design.
20110714-LEPP-Peterson
17
Validation of the FEA with small test beams
The concern is that the FEA may not properly treat the joints.
2008
The small test beams represent sections of the LP2 endplate
across the diameter of the LP1 endplate (slide 14).
For each small test beam, there is a
solid model that was used for the FEA
and an assembly model for construction.
Al-C Hybrid
Deflection of the small beams was compared to the FEA.
100 mm
The “equivalent plate” space-frame was constructed in carbon-fiber.
20110714-LEPP-Peterson
18
Comparison of deflection for small test beams: FEA vs. measurements
20100408
mass
LP1
0.63 kg
20101111
center load
FEA
mm/100N
0.755
20101104
20110202
20110316
center load
center load
center loaded
MEASURED MEASURED MEASURED
mm/100N
mm/100N
mm/100N
0.88
Al-C hybrid ( 1.0 part fiber : 1 epoxy )
Al-C hybrid ( 0.25 part fiber : 1 epoxy )
LP1 (channeled)
0.37 kg
0.75
0.78
1.71
2.12
2.224
2.40
(channeled for the hybrid, but without adding carbon fiber)
space-frame (“strut”)
space-frame (“equivalent-plate”)
0.76 kg
0.111
0.111
0.11
0.12
0.14
The standard LP1 beam agrees with the FEA, except for some problems
in the first measurement, probably sloppy centering the load.
The “strut” space-frame agrees with the FEA. The model accurately predicts the
strength at the joints and the design is useful for ILD.
The “equivalent-plate” space-frame was made with commercial carbon-fiber
plates that were claimed to have the modulus of aluminum; the modulus is ~20% low.
The cast-in-place carbon-fiber does not come close to having the modulus of aluminum.
(Possibly, with a a heat-cured epoxy, the result would have been better.)
20110714-LEPP-Peterson
19
What have we learned
It is possible to simply reduce the material in the LP1 endplate
from 18.85kg to 7.35kg or 16.9% X0 to 6.5 X0,
with a deflection increase to from 33 μm to 168 μm.
But, when the lightened endplate design (or even the LP1 original 2008 design)
is scaled up to the ILD endplate the deflection is unacceptable.
Only a space-frame design provides the rigidity and material limit for the ILD.
The space-frame design for LP2, with 100mm total thickness
can be made with 7.5 X0 material
and 23 μm deflection.
When scaled up to the ILD endplate, the deflection is 220 μm .
Note that I am typically measure and compare longitudinal rigidity(deflection),
when we are ultimately concerned about lateral rigidity and stability.
Longitudinal rigidity is readily measured in the prototypes;
for now, it gives an indication of the relative lateral rigidity and stability.
Lateral rigidity and stability will be monitored in longer term studies of a new LP2 endplate.
20110714-LEPP-Peterson
20
The next phase of prototyping/validation in the ILD endplate study:
construction and measurement of a fully functional LP2 endplate in a space-frame design.
The solid model has been converted to an assembly model,
compatible with LP1 modules, field cage, field cage termination, alignment devices.
The “strut” space-frame is chosen because the construction is simpler.
The “equivalent plate” design was developed because the FEA failed for ILD in the “strut” design.
The “equivalent plate” while providing a simple model for the FEA, is more difficult and messy to construct.
How does one glue in the plates beyond #3 ?
There has been further lightening of the mullions.
Modification of the outer scallops (from that of the original solid model) allows 2-axis, vertical tool machining.
Mass: 6.56 kg in main plate, 0.81kg in back plate, 1.72kg in struts, =9.2 kg total
(LP1 2008 = 18.9 kg)
(Some mass is added due to the reality of the mounting fixtures and machining tools.)
20110714-LEPP-Peterson
21
The assembly model
is used as input to the
machining.
2011-06-23: space-frame endplate main plate after first machining step.
It is currently at the vendor, 3rd / final machining step is complete.
Some corrections are required. (Some of the holes are 0.0001 inch small on one end.)
20110714-LEPP-Peterson
22
New “module backframes”
will have ~50% of the original mass,
slightly more clearance.
New “mounting brackets”
are required to clear the strut mounts.
Different versions are required for
new and old module backframes.
The entire assembly will be
available for testing in 2012.
20110714-LEPP-Peterson
23
Next ~3 months:
The space-frame
LP2endplate
will be assembled in PSB 398.
Studies will include deflection under load,
precision and stability for in the x-y plane.
The new LP2 endplate (another prototype section of the ILD endplate) will be used
to further validate the FEA of the ILD endplate,
and understand complexities of the design.
20110714-LEPP-Peterson
24
FUTURE:
We are considering jumping back into the gating studies,
expanding on our earlier experience.
Cornell will build a new TPC for studying
wire gate designs adapted to the curved TPC module.
Accepts standard LP1 modules.
20110714-LEPP-Peterson
Other components fit in the rack:
cathode,
field shaping,
field cage termination,
ion detector.
25
Summary
There has been modeling and FEA at several scales of ILD development:
small beams, LP1, LP2, ILD.
The space-frame design is expected to provide the required rigidity
and is a viable construction.
A “strut” space-frame version of the LP1 endplate will be constructed
this summer for further study of this possible ILD design.
lateral rigidity and stability: much more work is required
The new space-frame version of the LP2 endplate will be used in this study.
The preliminary ILD spaceframe design can provide
0.22 mm deflection (2.1 millibar overpressure)
with a contribution of 6% X0 material (bare endplate)
and 2% X0 from the module back-frames.
20110714-LEPP-Peterson
26