Title: Trakcing%20systems%20with%20Silicon%20with%20special%20reference%20to%20ATLAS-SCT
1Trakcing systems with Siliconwith special
reference to ATLAS-SCT
- Some generalities about tracking
- Special requirements in LHC environments
- About silicon
- About the ATLAS-SCT
2Some general considerations
Tracking measures particle 3-momenta
Particle track
Sagitta, s
Lever arm, L
r
Interaction point
Precision of sagitta measurement
(N position measurements)
3Requirements for good resolution
- Lever arm as long as possible (large p?large
detector) - Measurement of sagitta as precise as possible
- Magnetic field as large as possible
4A typical event Granularity is
essential
5Requirements to LHC tracker
- FAST (40MHz)
- Excludes standard Drift Chambers due to large
drift times - Spacial resolution a few tens of microns
- High granularity
- Radiation hard
- Must minimize material
- Drift Chambers would be optimal from this point
of view - Silicon is a good choice
6Does very precise tracking give very precise
momentum estimates?
- Not necessarily due to Multiple (coulomb)
scattering - Direction change due to a concentrated scatterer
x/X0 is the amount of material traversed in units
of radiation lengths
7Example Atlas SCT, 3 X0 /layer, and pixel
measurements inside (probably also about 3
X0per layer).
8Displacement of tracks in different tracker
layers as function of momentum due to MSC.
Excellent resolution never harms, but is
sometimes useless.
9The silicon strips of the ATLAS-SCT has a pitch
of 80 µm
- . So position resolution is 23 µm per hit.
- Standard deviation of a flat distribution
width/(v12)
10Basics of Silicon detectors
- Simple p-n junctions
- Reverse biased
- Junctions can be segmented into strips
300um thickness
From L.G. Johansen, Thesis (Bergen)
11A model for the diode Constant charge density in
the depletion region of the n-bulk, and heavily
doped p-side
Nd is the donor concentration
12Setting a reverse potential across the diode
depletes it to a depth given by (about)
13How to measure the depletion depth?
- Charge stored in bulk (charge density x volume
(Al)) - Capacitance
14The formula works!
20 detectors from Hamamatsu (L.G. Johansen,
thesis)
15Charge collection
- Typical detector thickness is 300 µm.
- Bethe-Bloch equation Landau fluctuations gives
a most probable energy loss of about 80keV - To create a free electron-hole pair you need
3.6eV - ?Signal is about 22000 electrons ( 3.5 fC)
16Energy loss distributions for 2 GeV
electrons,pions and protons, broader than Landau.
(Bak et al , NPB288681,1987)
17The ATLAS SCT barrel detector
- Pitch 80 µm (resolution 23 µm)
- 768 strips per detector
- Thickness 285 µm
- Size 6.36x6.40 cm2
- Should biassed to 350V
- p strips on n material
18Read-out electronics
- Must have low noise
- Noise scales with capacitance ? Size limitations
- microstrip detectors inter-strip capacitances
dominate. - Must be compact ?ASICs
- For ATLAS Must operate at 25 Mhz
- For ATLAS Must be radiation tolerant
19Signal from electrons from a Ru-106 source
(mostly minimum ionising particles). Fluctuations
are dominated by Landau fluctuations in the
deposited energy. Spectrum collected with fast
analog electronics Chip SCT128A (from
B.Pommersche, thesis, Bergen)
N/bin
Signal/noise
20Charge collection vs bias, 25 ns collection time
Difference in signal could be explained
by differences in detector thickness
From B. Pommeresche, Cand. Scient thesis, (U of
Bergen) We must over-deplete the detectors to
collect all the charge in time
21What do we look for to assess the quality of a
detector?
- Depletion voltage
- Inter-strip capacitances
- Radiation hardness
- ?will require biassing to high volts
- Leakage currents must under control at high volts
22The leakage current nightmare
- Current through the bulk No problem
- The Problem Currents on detector surface, around
corners and who knows from where... - High currents into the readout destroys the
electronics, to avoid it we take the following
measures - Capacitively (AC) coupled aluminium readout
strips - Guard ring structures around active detector area
(connect to ground to suck out current) -
23The nightmare (part II)
- Large currents result in high power dissipation
and heating of the detector system. - Must be able to control the current, if not fully
understood, it should at least be stable with
time! - Must test all detectors and detector modules for
leakage current.
24Careful detector design is required!
25A detail of the detector for ATLAS-SCT
(picture from L.G. Johansen, thesis)
26Leakage currents for some SCT detectors
(From L.G.Johansen, thesis (Bergen))
27The detector modules must be radiation hard
- All components tested in a proton beam to a
fluence of 3 x 1014 protons/cm2 - This is 50 more than expected for ten years of
LHC operation
28A few words on radiation damage
- The two most important effects are
- 1 Crystal defects are created in such a way that
the effective doping gets more p-like with
fluence (dose). - Vdep decreases
- Type inversion
- Vdep increases
- 2 Leakage current increases
- ? increase in noise
- 3 Depletion from below
- n doping of back side
- preserves diode junction
29Development of leakage current with time
L.G. Johansen, thesis
30Signal/noise of an irradiated detector
Plot from L.G. Johansen (of course.)
31Leakage current doubles for a temperature
increase of 7 degrees
- ATLAS-SCT will be operated at about -10 degrees
C - Detector modules to be in thermal contact with
cooling agent.
32ATLAS-SCT readout electronics
- Digital readout (hit/no hit)
- Pros and cons.of digital electronics
- Rad. Hard.
- 128 readout channels per chip.
33Schematic of the ABCD3T chip
34The ALTAS SCT module
35ATLAS-SCT barrel module
- 4 detectors
- 1 baseboard (Patented TPG solution)
- Must be thermally conductive
- Hybrid with 12 chips, wraps around the
sensor-baseboard. - Strips are bonded together in pairs, to form 12
cm long strips. - About 3000 wire bonds per module
36A drawing of the module
37Production of about 2000 modules at 4 university
clusters around the world
- Necessary for efficient use of small resources at
each university. - Production clusters
- Japan
- US
- UK
- Scandinavia
38Equipment needed or developed
- Cleanrooms
- Tools for precision mounting (motorized jigs etc)
- Microscopes
- Metrology equipment (smart microscope)
- Bonding machines
- Setups for electric tests
39Module production
- Detector testing
- Glueing of detectors to baseboard (5 um
precision) - Testing (IV), metrology
- Hybrid testing and glueing
- Bonding
- Testing
40To mount to 5 micron precision is not trivial!
41Noise occupancy must be under control!
42Some module IV curves (nightmare, part III)
43But, in the end, the project seems to have been
successful
- Yield factor OK (above 85 )
- Module mounting on barrels in Oxford
- Transfer to CERN OK
- Cosmic tests OK
- Now, the barrel is in the ATLAS pit to be cabled
and tested further
44(No Transcript)
45(No Transcript)
46Conclusions
- Silicon tracking is very attractive in HEP
- But not at all trivial to make.
- Very cost and manpower intensive