Transcript Solid State Detectors

Solid State
Detectors- 3
T. Bowcock
Schedule
1
2
3
4
5
6
Position Sensors
Principles of Operation of Solid State
Detectors
Techniques for High Performance
Operation
Environmental Design
Measurement of time
New Detector Technologies
2
Techniques for High
Performance Operation
• Strip Detectors
– Design and Fabrication Issues
• What to avoid!
3
Review...
• In the p-strip in n-bulk (“p-in-n”) Al
-V
detectors
Si
-+
• Vdep=100V
• Energy to create electron hole pair is
– 3.6eV ( not 1.1eV-why? )
• Average energy lost/mm
– 39keV (108eh/ mm)
4
1.2
1
Drift
0.8
250V Irrad
0.6
250v Unirrad
60V Unirrad
0.4
0.2
97
89
81
73
65
57
49
41
33
25
9
17
1
0
• Electric field in Depleted
region linear
– 300mm detector
– at 100V E=3.0keV/cm
• Diffusion/Drift by multiple
collisions
• Takes 7ns for e’s, 20ns for
holes
 x2 
dN
dx
 exp  
N
4
Dt


D
kT
m
q
Higher diffusion at low
temps!
5
Ballistic Deficit
Charge lost is known as the ballistic deficit
1.2
1
0.8
250V Irrad
0.6
250v Unirrad
60V Unirrad
0.4
0.2
Collection time
97
89
81
73
65
57
49
41
33
25
17
9
1
0
6
Strip Pitch and Readout
Pitch and resolution
• Select it:
d
Single strip has d/12
d/10
7
Choosing the Pitch
• Why not make it
infinitely small
– transverse diffusion
• 10-20 microns
– construction
– readout electronics!
• Readout pitch
– not necessarily the
same as diode
pitch (cost$$$)
75mm readout
(25mm diode)
8
Intermediate Strips
• Work by capacitive coupling
– induced current/charge is that seen by the
electrons and holes (not a post-facto charge
sharing!)
• Why no broader strips ?
– Interstrip capacitance <1pF
Need field map!
9
Intermediate Strips?
• Loose signal
• An option if
– limited by resources
– little noise in electronics (slow e’s)
• Optimal choice is
– readout each strip
• pitch and width evaluated by FEA
– pitch between 20 microns and 100 microns
10
Performance
Resolution
12
microns
10
50 mm with intermediate strip
8
Series1
6
Series2
4
25mm readout
2
0
0
10
20
30
Signal/noise
11
Resolution
• Test your resolution
– series of particles
of known position
• testbeam telescope
• cosmic telescope
• longwavelength
laser
12
Checking Resolution
• Tests
– laser
Optical fiber
Focus to 5 mm
• problems?
• transparancy
– cosmics
• slower
– testbeam
• expensive
• labour intensive
1064nm
Si transparent
13
Two Track Resolution
• Reconstruction position as a function
of proximity of one track to another
14
Occupancy
• Best to reduce
occupancy
– 1% considered the
benchmark
• 10% too high
• Reduce the length of
strips
– usually about 6cm
– reduce to 1cm for
example
15
AC Coupling Revisited
• e=0.34pF/cm
• 200nm oxide
– 10pF/cm
• Greater than
Interstrip
capacitance
• Electronics at
ground!
16
Double Sided
-V
--
++
--
++
• Needs AC
coupling!
• Correlation
of signals
• Strips can
run opposite
directions
0V
– 2D style r/o
17
Double Sided Detector
• Would like
electronics at one
end
• Can get correlated
measurement (E)
giving x/y
measurement
• Reduces fakes
• Punchthrough
18
Double Metal
• Add another
routing layer
• more processing
via
•
•
Expense can double
Built in stresses in SiO2
can warp Si wafer badly
19
Double Metal
Can also use to
route on single
sided detectors
20
Strips
21
Example of Double Metal
Detectors
• LHCb prototypes
22
Bond Pads
• Structure you will often see
Typically 80 by 200 microns
23
n-strip detectors
• We can make single sided n-strip
detectors (note depletion!)
24
Field Plates
• MOS structures
25
p-stops
• Individual p-stops
26
Operating Voltage
• High (overvoltage is desirable)
– 250V
– reduced ballistic deficit
• BUT
– introduces very high field regions?
• Avalanche will set in if field exceeds 30V/m
27
Analysis of structure
28
Electric Field
Sample field map
29
Guard Rings
• Reduce
fields at
edge
30
Micro-Discharges
• Discharges may be seen as in
increase in the noise with voltage
31
Si Choices
• Resistivity
• n-type
– p-strips
– n-strips
– double sided
• p-type
• Crystal orientation
32
Benchmark measures
• Charge Collection Efficiency
• Partial Depletion
• Ballistic Deficit
33
Fabrication
• Control of all steps critical
• Of special interest
– resistor values
– implantation
– oxide quality for breakdown
– quality of lithography
34
Quality Assurance
• Job of the physicist is to measure all
the key parameters of the detectors
– IV and CV
– interstrip capacitance
– resistor values
– lightspot response
35
Readout Chain
36
F/E Electronics
• Binary vs Analog
• Amplifier Characteristics
– rise time and falltime
– undershoot
• Digital Performance
– pipeline & logic
• Noise
37
Hybrid Design
38
Noise
• Hybrid is often a source of noise
– bad grounding for electronics
– bad grounding for supplies to detector
– sensor,analog and digital all connected
• The detector, f/e electronics and the
hybrid should be regarded as one unit
or MODULE
39
Module and Mounting
40
Material Budget
• Ideally should be as low as possible
– avoid high mass materials
• gold
• Good detector about 1% of a radiation
length
41
Example: DELPHI barrel
42
Offline Analysis
• Can give
improvement in
resolution
w
L
R
w

PR  A  x 
2

w

PL  A  x 
2

P d  PL d Adx d
x'  R

 x
PR  PL
Aw w
Only true if charge uniform and if
the width of the cluster matches
the strip width
d
x
In general we have a Gaussian
distribution of width determined by
the diffusion coefficient (for normal
incidence)
43
Offline
• Corrections for the angle of the track
and the known (measured) charge
sharing can give great improvement
– 20 to 30% in the case of 25 microns
pitch detectors
• Good software must accompany good
hardware
• Removal of deltas
44
7 things to avoid
•
•
•
•
•
•
•
Picking the wrong technology
Picking the wrong manufacturer($)
Not enough Quality Control
Bad design limiting operation
Noise in system
Treating sensor and hybrid separately
Bad analysis
45
Summary
• We have all the elements now to think
about real detectors in real
environments
– design issues
– noise problems
• See how we design a detector for
LHCb
46