the hermes recoil silicon detectorvertex2002/vprivate/stewart.pdf · 2002. 11. 8. · âuse an...
TRANSCRIPT
-
J Stewart Vertex 2002 1
The Hermes Recoil Silicon Detector
ââ Introduction Introductionââ Detector design considerations Detector design considerationsââ Silicon detector overview Silicon detector overviewââ TIGRE TIGRE microstrip microstrip sensors sensorsââ Readout electronics Readout electronicsââ Test beam resultsTest beam results
J. StewartDESY Zeuthen
DESY, Erlangen, and U. Glasgow
Vertex 2002
-
J Stewart Vertex 2002 2
The HERMES Spectrometer
â Forward Spectrometer and the 27 GeV polarized e+/e- Hera beam.â Large solid angle acceptance: |θx| < 170 mrad, 40 < |θy| < 140 mrad.â Momentum resolution for charged particles ~1% from 1 to 27 GeV.â Calorimeter resolution on the order of 5%.
Position ofRecoil Detector
-
J Stewart Vertex 2002 3
The Development of the Theory of GPDsHas Created a Great Deal of Interest in the
Study of Exclusive Processes.
ââ GPDs go beyond the probability ofGPDs go beyond the probability offinding a parton with momentumfinding a parton with momentumfraction x.fraction x.
ââ Provide a unified description of awide variety of physics processes.
ââ Measurements of exclusive photonMeasurements of exclusive photonproduction is the best way to studyproduction is the best way to studythe GPDs.the GPDs.
Deeply virtual Compton scatteringDVCS
-
J Stewart Vertex 2002 4
Limitations of the Existing Hermes Data
ââ missing mass resolution formissing mass resolution forDVCS candidate events atDVCS candidate events atHERMES: notHERMES: notsufficient to identifysufficient to identifyexclusive eventsexclusive eventsindividuallyindividually
ââ at present: limited energyat present: limited energyand position resolution atand position resolution atHERMESHERMES
ââ exclusivity only for a dataexclusivity only for a datasample and not forsample and not forindividual eventsindividual eventsMx resolution leads to negative values
2
e + p Y e + γ + p
-
J Stewart Vertex 2002 5
A Major Improvement in the HermesSpectrometer Is Needed!
The recoil protons need to bedetected!
-
J Stewart Vertex 2002 6
The Recoil Protons
â Lower momentum cutoff determined bymaterial between target and detectorÈ Place detector in vacuum!È Silicon detectors
â 0.135 < p < 1.4 GeV/c RealisticÈ Energy deposited in 300 µm silicon:
4.5 MeV to 86 keVÈ Dynamic range of 50+!
â Cover as much of 2π in phi as possible
50 < p < 1400 MeV/c0.1 < θ < 1.35 rad(10 to 80 degrees!)2π in φ
Want good t resolution Y Low t behavior is important!
-
J Stewart Vertex 2002 7
Background Suppression
â The transverse momenta of all particles are of comparable size.• Cuts based on the transverse momentum.• Transverse momentum resolution needs to be better than 10%.• Angular resolution needs to be better than 0.1 rad.
â Also cut on coplanarity.
Can achieve factor 5 in background suppression!
-
J Stewart Vertex 2002 8
â Use an existing silicon microstrip detector design.• Financial and time considerations preclude a custom design.
â Use the largest available double sided detector with less than1 mm pitch.• Adjust the target length to match the size of the silicon sensor.
â Minimize the material between the target gas and the siliconsensor.
â Measure the particle momentum using the relation between theenergy deposited in the silicon and the momentum.
â Cover as much of the 2π as possible.• HERMES has no coverage below 40 mrad in θy.
Recoil Detector Sensor Design Criteria
-
â 16 double sided sensors• each 99 mm × 99 mm
â 2 layer square tube orientation
â 76% Φ acceptanceâ 150 mm long target cellâ θ Coverage
• 0.4 < θ < 1.35 rad
Silicon Detector
Choose the TIGRE detector from MICRON semiconductor.
Readout Hybrid
TIGRESensor
Cooling
Target
-
J Stewart Vertex 2002 10
The Detector Coverageâ Deeply virtual Compton Scattering
(DVCS)â Combined Bethe Heitler and DVCSâ Protons from exclusive ρ
productionâ Protons from delta excitation
â 135 < P[MeV/c] < 400â 400 MeV/c proton Y3 MIPâ Minimum momentumâ Target + Flexfoil + 1st Siâ 135 < p < 250 MeV/c PID
Silicon Coverage
Need additional detectorsfor full coverage!
-
J Stewart Vertex 2002 11
3D Model of the Recoil Detector
-
J Stewart Vertex 2002 12
TIGRE Silicon Sensors
Parameter ValueSensor Size 99 mm x 99 mmActive Area 97.3 mm x 97.3 mmSilicon Thickness 300 mmStrip Pitch 758 mmStrip Separation 56 mmCoupling Capacitance 1 nFTotal Strip Capacitance 25 pFPolysilicon bias resistor 50 MWDepletion voltage 50 V MaxP-Stop TechnologySingle Guard Ring
Micron has produced detectors with7 MΩ polysilicon for HERMES
Manufacturer: Micron SemiconductorManufacturer: Micron Semiconductor
-
J Stewart Vertex 2002 13
Measurements of Sample TIGREs
Parameter Value
Single strip leakage current 40 nADepletion voltage 40 VCoupling capacitor 1.2 nF
Total strip capacitance 30 pFInterstrip capacitance 8 pF
Polysilicon bias resistor >60 MΩ
New sensor 7 MΩ
-
J Stewart Vertex 2002 14
Readout Electronics
â Used for the HERMES VC.â Can select between high and
low gain.â Neither pattern generator
nor ADC available.
â Used for the HERMESLambda Wheels
â Dynamic range a problem.â Both the pattern generator
and ADC are available.
â A pipeline chip is needed to accommodate the HERMES trigger.â A dynamic range of 70 ( 280 fC).â 10 MHz readout to match the HERA frequency.â Prefer a chip already used in HERMES.
Chip Selection Criteria:
APC HELIX
-
J Stewart Vertex 2002 15
Helix 128 - 3.0
â 0.8 µm CMOS process.â 10 MHz sampling frequency.â128 input channels.âAnalog pipeline 141 cells deep.
âPreamp-Shaper good noise char. âRadiation tolerant 220 krad. âDynamic range
+/- 40 fC or +/- 10 MIP
Designed by the ASIC lab. at the University of Heidelberg.
Preamp Shaper Buffer
Pipeline Pipeamp MUX
-
J Stewart Vertex 2002 16
First Tests Using Charge Injection
â Use an old Zeus MVD Hybridâ Connect the Helix via Zeus pitch
adapter to a general purpose pitchadapter.
â Capacitively couple an input pulse.â Readout with the Zeus laser test
stand.
-
J Stewart Vertex 2002 17
Helix Response
1 “MIP” =24,000 e-/h+/- “MIP”=Pos/Neg charge
on preamp
Conclusions:
âLinear response over+/-10 “MIP”
âSaturated at ~15 “MIP”
As Expected
-
J Stewart Vertex 2002 18
Readout Conceptual Design
âCapacitive coupling readout.âDoubles number of channels.
•32 Y64 Helix 3.0.âCan adjust dynamic range.
70 MIP Possible!
Proposed by W. Lange
Charge Division!
-
J Stewart Vertex 2002 19
First Prototype Constructed
ZEUS MVD hybrid Sensor
-
J Stewart Vertex 2002 20
DESY2 Test Beam
high gainchannel
low gainchannel
â First silicon prototype module has been tested in 1 GeV electron beam.• 50% charge collection efficiency due to large sensor capacitance (30 pF).• Signal to noise ratio of > 6.5 for a 1 MIP particle.• Indicates we can measure particles depositing energy up to 140 MIP.
Landau Fit Landau + Gaussian Fit
-
J Stewart Vertex 2002 21
Results of Charge Division Studies
ââ CCeffeff can be calculated from the difference in slopes for high and low gain can be calculated from the difference in slopes for high and low gainusing charge injection. Cusing charge injection. Ceffeff = 25 pF = 25 pF•• (agrees fairly well with published numbers C(agrees fairly well with published numbers Ceffeff = 31 pF) = 31 pF)
Charge Injection Test BeamCc (pF) Qhigh Qlow Qhigh Qlow22 67% 33% 72% 28%10 78% 22% 82% 18%4.7 86% 14% - -
-
J Stewart Vertex 2002 22
Present Status of Mechanical Design
SciFi Connector Holding Structure
Scattering Chamber
TargetCell
HERABeamline
TIGRESensors
Hybrid
CoolingCollimator
-
J Stewart Vertex 2002 23
Next Steps
â Set up a test stand using the HERMES pattern generatorand ADC.
â Test the components to be used for the first readouthybrid.
â Test the needed Helix 3.0 chips.â Design the layout and manufacture the first hybrid.â Assemble the first real module.â Vacuum Testing.â TEST BEAM
â ⋅ ⋅ ⋅⋅ ⋅ ⋅Hope to be ready for installation in spring 2004!
-
J Stewart Vertex 2002 24
SummaryâResponse of both the HELIX 3.0 and APC readout chips
to large pulses has been measured.â First prototypes have been constructed and tested in test
beam.âReadout using charge division has been shown to work.
• 50% charge collection due to large sensor capacitance.• S/N for 1 MIP is 6.5.• With a 5 pF coupling capacitor, it may be possible to measure
particles depositing 140 times the energy of a 1 MIP.
âHELIX 3.0 chosen for readout.â Work has started on our first readout hybrid.âA test stand is under construction at DESY Zeuthen.