Tracking Detector Material Issues
for the sLHC
Hartmut F.-W. Sadrozinski
SCIPP, UC Santa Cruz, CA 95064
Hartmut F.-W. Sadrozinski, US ATLAS Upgrade Meeting Nov 10, 2005
1
Outline of the talk
- Motivation for R&D in new Detector Materials
- Radiation Damage
- Initial Results with p-type Detectors
- Expected Performance
- R&D Plan
- Much of the data from RD50 http://rd50.web.cern.ch/rd50/
- In collaboration with Mara Bruzzi and Abe Seiden
- Presumably this is relevant for both strips and pixels
- Will not discuss 3-D detectors here
Announcement:
2nd Trento Workshop on Advanced Detector Design
(focus on 3-D and p-type SSD)
Feb 15. –16. 2006
Hartmut F.-W. Sadrozinski, US ATLAS Upgrade Meeting Nov 10, 2005
2
Motivation for R&D in New Detector Materials
- The search for a substitute for silicon detectors (SSD) has come up empty.
- Radiation damage in SSDs impacts the cost and operation of the tracker.
- What is wrong with using the p-on-n SSD a la SCT in the upgrade?
- Type inversion requires full depletion of the detector
- Anti-annealing of depletion voltage constrains thermal management
- Large depletion voltages require high voltage operation
- Slower collection of holes wrt to electrons increases trapping
- What is wrong with using the n-on-n SSD a la ATLAS pixels in the upgrade?
- Cost: double-sided processing about 2x more expensive
- Type inversion changes location of junction
(but permits under-depleted operation)
- Strip isolation challenging, interstrip capacitance higher?
-Potential solution: SSD on p-type wafers (“poor man’s n-on-n”)
- Single-sided processing, no change of junction
- Strip isolation problems still persist
- Need to change the wafer properties to reduce the large depletion voltages: MCz
Hartmut F.-W. Sadrozinski, US ATLAS Upgrade Meeting Nov 10, 2005
3
Charge collection efficiency CCE on n-side
G. Casse, 1st RD50 Workshop, 2-4 Oct. 2002
n-side read-out after irradiation.
1060nm laser CCE(V) for the highest
dose regions of an n-in-n (7.1014p/cm2)
and p-in-n (6.1014p/cm2) irradiated
LHC-b full-size prototype detector.
Hartmut F.-W. Sadrozinski, US ATLAS Upgrade Meeting Nov 10, 2005
4
Radiation Effects in Silicon Detectors
Basic effects are the same for n-type and p-type materials.
- Increase of the leakage current.
- Change in the effective doping concentration (increased depletion voltage),
- Shortening of the carrier lifetimes (trapping),
- Surface effects (interstrip capacitance and resistance).
The consequence for the detector properties seems to vary widely.
- An important effect in radiation damage is the annealing,
which can change the detector properties after the end of radiation.
- The times characterizing annealing effects depend exponentially on the temperature,
constraining the temperature of operating and maintaining the detectors.
- Fluence dependent effects normalized to equivqlent neutrons (“neq”),
We use mostly proton damage constants and increase the fluence by 1/0.62.
Hartmut F.-W. Sadrozinski, US ATLAS Upgrade Meeting Nov 10, 2005
5
Radiation Induced Microscopic Damage in Silicon
particle
Sis
Frenkel pair
Vacancy + Interstitial
EK > 25 eV
Point Defects (V-V, V-O .. )
V
I
EK > 5 keV
clusters
Influence of defects on the material and device properties
charged defects
 Neff , Vdep
Trapping (e and h)
 CCE
e.g. donors in upper
and acceptors in
lower half of band
gap
shallow defects do not
contribute at room
temperature due to fast
detrapping
generation
 leakage current
Levels close to
midgap
most effective
Hartmut F.-W. Sadrozinski, US ATLAS Upgrade Meeting Nov 10, 2005
6
Leakage Current
Hadron irradiation
Annealing
6
10-2
-3
10
-4
10
10-5
10-6 11
10
•
•
n-type FZ - 7 to 25 Kcm
n-type FZ - 7 Kcm
n-type FZ - 4 Kcm
n-type FZ - 3 Kcm
p-type EPI - 2 and 4 Kcm
80 min 60C
1012
(t) [10-17 A/cm]
I / V [A/cm3]
10-1
n-type FZ - 780 cm
n-type FZ - 410 cm
n-type FZ - 130 cm
n-type FZ - 110 cm
n-type CZ - 140 cm
p-type EPI - 380 cm
1013
eq [cm-2]
1014
80 min 60C
5
1015
5
4
4
3
3
2
2
.
17
-3
oxygen enriched silicon [O] = 2 10 cm
parameterisation for standard silicon
1
[M.Moll PhD Thesis]
Damage parameter  (slope)
6
1
[M.Moll PhD Thesis]
0
1
10
100
1000
o
10000
annealing time at 60 C [minutes]
M. Moll, Thesis, 1999
I
α
V   eq
 independent of eq and
impurities
 used for fluence calibration
(NIEL-Hypothesis)
•
•
Oxygen enriched and
standard silicon show
same annealing
Same curve after
proton and neutron
irradiation
Hartmut F.-W. Sadrozinski, US ATLAS Upgrade Meeting Nov 10, 2005
7
Vdep and Neff depend on storage time and temperature
Stable Damage
N eff  N C 0 (1  e c )  [ g c  g a e

t
 a (T )
 g y (1  e
Beneficial Annealing
ShallowDonor Removal
 Neff [1011cm-3]
4
10
T = 300K
Vdep [Volt]
3
t
 y (T )
)]
Reverse Annealing
10
80min at 60°C
8
10

2
10
NY, = gY eq
Na = ga eq
6
4
NC
gC eq
2
NC0
1
10
0
5 kcm
1 kcm
500 cm
0
10
11
12
13
14
15
1
10
100
1000
10000
o
annealing time at 60 C [min]
G.Lindstroem et al, NIMA 426 (1999)
Short term: “Beneficial annealing”
Long term: “Reverse annealing”
M. Bruzzi, Trans. Nucl. Sci. (2000)
time constant : ~ 500 years (-10°C)
~ 500 days ( 20°C)
after inversion and annealing saturation Neff  b  
~ 21 hours ( 60°C)
Hartmut F.-W. Sadrozinski, US ATLAS Upgrade Meeting Nov 10,30min
2005 (80°C)
8
10
10
10
10
-2
Fluence [cm ]
10
•
•
Charge Collection Efficiency
 Partial depletion
Limited by:
 Trapping at deep levels
 Type inversion (SCSI)
Collected Charge:
Q  Qo   dep   trap
 dep 
d
W
 trap  e
1/e,h = βe,h·eq[cm-2]
c
t
From TCT measurements within RD50:
t ~ 0.2*1016 / , t ~ 0.2 ns for   1016 cm-
2.00E+04
1.50E+04
Trapping T from
Krasel et al
1.00E+04
Casse et al: ptype
5.00E+03
Trapping T scaled
by 2.4
0.00E+00
1.0E+14

W: Detector thickness
d: Active thickness
c : Collection time
t : Trapping time
1.0E+15
1.0E+16
2
Luckily this is excludedby CCE measurements:
 t ~ 0.48*1016 / 
Fluence 
[neq/cm2]
Trapping time
[ns]
3·1014
16
5·1014
1·1015
3·1015
9.6
4.8
1.6
Hartmut F.-W. Sadrozinski, US ATLAS Upgrade Meeting Nov 10, 2005
9
Defect Engineering of Silicon
Influence the defect kinetics by incorporation of impurities or defects: Oxygen
Initial idea: Incorporate Oxygen to getter radiation-induced vacancies
 prevent formation of Di-vacancy (V2) related deep acceptor levels
•Higher oxygen content  less negative space charge
One possible mechanism: V2O is a deep acceptor
VO
(not harmful at RT)
V
VO
V2O (negative space charge)
V2 in
clusters
Ec
V2O(?)
Carbonated
600
500
6
Standard
400
300
4
200
Oxygenated
2
0
0
VO
EV
8
Carbon-enriched (P503)
Standard (P51)
O-diffusion 24 hours (P52)
O-diffusion 48 hours (P54)
O-diffusion 72 hours (P56)
100
1
2
3
4
24 GeV/c proton [10 cm ]
14
-2
5
DOFZ (Diffusion Oxygenated Float Zone Silicon) RD48 NIM A465 (2001) 60
Hartmut F.-W. Sadrozinski, US ATLAS Upgrade Meeting Nov 10, 2005
10
Vdep [V] (300 m)
O
|Neff| [1012cm-3]
10
Caveat with n-type DOFZ Silicon
Discrepancy between CCE and CV analysis observed in n-type
(diodes / SSD, ATLAS / CMS, DOFZ / Standard FZ)
Vrev 95% Charge Coll. [V]
Author
standard - oxygenated
500
Casse et al.
Robinson et al.
Buffini et al.
Robinson et al.
Casse et al.
Lindstroem et al.
200
100
0
0
100
200
300
400
Vdep CV analysis [V]
To maximise CCE it is necessary to
overdeplete the detector up to :
Exp.
material
●
Robinson et 3x1014
al., NIM A 24GeV
461 (2001) p/cm2
ATLAS Oxygen. +
standard
■
Casse et al., 3-4x1014
NIM A 466 24GeV
(2001)
p/cm2
ATLAS Oxygen. +
standard
400
300
radiation
♦
500
▲
Lindström
et al., NIM
A 466
(2001)
Buffini et
al., NIM A
(2001)
1.65x1014 ROSE
24GeV
p/cm2
Oxygen.
<100>
1.1x1014
1MeV
n/cm2
Standard
<111>
CMS
Vbias ~ 2 Vdep
Hartmut F.-W. Sadrozinski, US ATLAS Upgrade Meeting Nov 10, 2005
11
Caveat:
The beneficial effect of oxygen in proton irradiated
silicon
microstrip
almost
disappear
in
CCE
measurements
G.Casse et al. NIM A 466 (2001) 335-344
ATLAS microstrip CCE
analysis after irradiation
with 3x1014 p/cm2
Hartmut F.-W. Sadrozinski, US ATLAS Upgrade Meeting Nov 10, 2005
12
CCE n-in-p microstrip detectors
 Miniature
n-in-p
microstrip
detectors (280m thick) produced by
CNM-Barcelona using a mask-set
designed by the University of
Liverpool.
CCE ~ 60% after 3 1015 p cm-2 at
900V( standard p-type)
CCE ~ 30% after 7.5 1015 p cm-2
900V (oxygenated p-type)
 Detectors read-out with a SCT128A
LHC speed (40MHz) chip
 Material: standard p-type
oxygenated (DOFZ) p-type
and
 Irradiation: 24GeV protons up to 3
1015 p cm-2 (standard) and 7.5 1015 p
cm-2 (oxygenated)
G. Casse et al., Nucl. Inst Meth A 518 (2004) 340-342.
At the highest fluence Q~6500e at Vbias=900V. Corresponds to:
ccd~90µm, trapping times 2.4 x larger than previously measured.
Hartmut F.-W. Sadrozinski, US ATLAS Upgrade Meeting Nov 10, 2005
13
Recent n-in-p Results
Important to check that there are no unpleasant
surprises during annealing.
800
700
600
ADC
Minutes at 80oC converted to days at 20oC using acceleration factor of
7430 (M. Moll).
900
500
300 V
400
500 V
300
200
G. Casse et al., 6th RD50 Workshop, Helsinki June 2-4 2005
http://rd50.web.cern.ch/rd50/6th-workshop/.
800 V
100
0
0
6
100
200
4
3
2
1
0
0
200
400
600
800
1000
1200
Days @ 20 oC
Signal ke-
Signal ke-
5
Detector after 7.5× 1015 p/cm2 showing
pulse height distribution at 750V after
annealing. (Landau + Gaussian fit)
20
18
16
14
12
10
8
6
4
2
0
300
400
Minutes @ 80 oC
300 V
500 V
800 V
0
500
1000
1500
2000
Days @ 20 oC
Detector with 1.1× 1015 p/cm2
Hartmut F.-W. Sadrozinski, US ATLAS Upgrade Meeting Nov 10, 2005
14
Expected Performance for p-type SSD
Details in : “Operation of Short-Strip Silicon Detectors based on p-type Wafers in the ATLAS Upgrade ID
M. Bruzzi, H.F.-W. Sadrozinski, A. Seiden, SCIPP 05/09
Conservative Assumptions:
p = 2.5·10-17 A/cm (only partial anneal)
Ctotal = 2 pF/cm
Vdep = 160V + b* ( with 2.7* 10-13 V/cm2) (no anneal)
(= 600V @  = 1016 neq/ cm2)
s2Noise = (A + B·C)2 + (2·I·s)/q A = 500, B = 60
S/N for Short Strips for different bias voltages:
35.0
35.0
30.0
25.0
25.0
20.0
20.0
S/N
S/N
30.0
300um, -20deg, 400V
300um, -20deg, 600V
300um, -20deg, 800V
15.0
200um, -20deg, 400V
200um, -20deg, 600V
200um, -20deg, 800V
no need for thin detectors,
unless n-type:
depletion vs. trapping
600V seems to be sufficient
15.0
10.0
10.0
5.0
5.0
0.0
1.E+12 1.E+13 1.E+14 1.E+15 1.E+16
0.0
1.E+12 1.E+13 1.E+14 1.E+15 1.E+16
Fluence [neq/cm2]
Fluence [neq/cm2]
Hartmut F.-W. Sadrozinski, US ATLAS Upgrade Meeting Nov 10, 2005
15
Expected Performance for p-type SSD, cont.
Noise for SiGe Frontend
(see talk by Alex Grillo)
Leakage current important:
Trade shaping time against operating temperature
( 20 ns & -20 oC vs. 10 ns & -10 oC )
Temperature:
-10 deg C
Fluence:
2.2·1015 neq/cm2 (short strips) 2.2·1014 neq/cm2 (long strips)
The maximum bias voltage is 600 V
Noise vs. Shaping time
S/N vs. Temperature
c=6, f=0
1500
20.0
C = 6, 10 ns
C = 6, 15 ns
C = 6, 20 ns
C = 15, 10 ns
C = 15, 15 ns
C = 15, 20 ns
c=6, f=2e15
c=6, f=2e15, 20deg
C=15, f=0
1000
C=15, f=2e14
15.0
S/N
RMS Noise [e-]
c=6, f=2e14
10.0
500
5.0
0.0
0
5
10
15
20
Shaping Time  [ns]
25
-35
-30
-25
-20
-15
-10
-5
Temperature [oC]
Hartmut F.-W. Sadrozinski, US ATLAS Upgrade Meeting Nov 10, 2005
16
Expected Performance for p-type SSD, cont.
2
Heat Generation in 300 m SSD
Temperature [oC]
(T)/ (20)
Only from active volume
 neq
3E+14
5E+14
1E+15
1E+15
3E+15
3E+15
3E+15
20
1
T 
E 1 1
I (T )  I (T0 )  exp( b   )
2 K  T0 T0 
 T0 
0
0.197
-10
0.0797
-20
0.0300
-30
0.0104
I
 
Volume
Generated Heat Flux [W/cm2]
Vbias [V] w [m] T = 20°C T=-10°C T=-20°C T=-30°C
290
300
1.05E-01 6.75E-03 2.35E-03 7.54E-04
376
300
2.27E-01 1.46E-02 5.09E-03 1.63E-03
400
247
3.98E-01 2.55E-02 8.90E-03 2.85E-03
591
300
7.15E-01 4.59E-02 1.60E-02 5.13E-03
400
157
7.62E-01 4.89E-02 1.70E-02 5.46E-03
600
193 1.40E+00 8.99E-02 3.13E-02 1.00E-02
800
223 2.16E+00 1.38E-01 4.82E-02 1.55E-02
Hartmut F.-W. Sadrozinski, US ATLAS Upgrade Meeting Nov 10, 2005
17
An Italian network within RD50: INFN SMART
n-type and p-type detectors processed at IRST- Trento
Edge structures
Pad detector
Test2
Test1
Square MG-diodes
Microstrip
detectors
Inter strip Capacitance test
Round MG-diodes
Wafers Split in:
1. Materials:
(Fz,MCz,Cz,EPI)
2. Process:
Standard
Low T steps
T.D.K.
3. Isolation:
Low Dose p-spray
High Dose p-spray
Hartmut F.-W. Sadrozinski, US ATLAS Upgrade Meeting Nov 10, 2005
18
SMART News: Annealing behaviour of MCz Si n- and p-type
Vdep variation with fluence (protons) and
annealing time (C-V):
Beneficial annealing of the depletion voltage:
14 days at RT, 20 min at 60 oC. 3 min at 80 oC.
Reverse (“anti-”) annealing starts
in p-type MCz: at 10 min at 80 oC , 250 min (=4 hrs) at 60 oC,
>> 20,000 min (14 days) at RT,
in p-type FZ : at 20 min at 60 oC
in n-type FZ: at 120 min at 60 oC.
G. Segneri et al. Submitted to NIM A,
presented at PSD 7, Liverpool , Sept. 2005
A. Macchiolo et al. Submitted to NIM A,
presented at PSD 7, Liverpool , Sept. 2005
Hartmut F.-W. Sadrozinski, US ATLAS Upgrade Meeting Nov 10, 2005
19
SMART News: Annealing behaviour of n- type MCz Si
(is n-type MCz inverted?)
N-type
M. Scaringella et al. presented at Large Scale Applications
and Radiation Hardness Florence, Oct. 2005
A. Macchiolo et al. Submitted to NIM A,
Hartmut F.-W. Sadrozinski, US ATLAS Upgrade Meeting Nov 10, 2005
20
presented at PSD 7, Liverpool , Sept. 2005
Inter-strip Capacitance
One of the most important sensor parameters contributing to the S/N ratio
Depends on the width/pitch ratio of the strips
and on the strip isolation technique (p-stops, p-spray).
Observe large bias dependence on p-type detectors, due to accumulation layer.
Cint [F]
Interstrip Capacitance
2.0E-11
1.8E-11
1.6E-11
1.4E-11
1.2E-11
1.0E-11
8.0E-12
6.0E-12
4.0E-12
2.0E-12
0.0E+00
14-5 250krad
Pre-rad
SMART 14-5
p-type FZ
low-dose spray
w/p = 15/50
Vdep = 85 V
(I. pitch
Henderson,
100 m
J. Wray,
D. Larson,
SCIPP)
Cint = 1.5 pF/cm
0
100
200
300
400
100
m
pitch
Bias
Voltage
[V]
500
Irradiation with 60Co
reduces
the bias dependence,
as expected.
Hartmut F.-W. Sadrozinski, US ATLAS Upgrade Meeting Nov 10, 2005
21
Status

Radiation hard materials for tracker detectors at SuperLHC are under study
by the CERN RD50 collaboration. Fluence range to be covered with
optimised S/N is in the range 1014-1016cm-2 . At fluences up to 1015cm-2 (Mid
and Outer layers of a SLHC detector) the change of the depletion voltage
and the large area to be covered by detectors is the major problem.

High resistivity MCz n-type and p-type Si are most promising materials.

Quite encouragingly, at higher fluences results seem better than first
extrapolated from lower fluence:
longer trapping times ( p-FZ, p-DOFZ)
delayed and reduced reverse annealing ( MCz SMART)
sublinear growth of the Vdep with fluence ( p - MCz&FZ)
delayed/supressed type inversion ( p- MCZ&FZ, MCz n- protons)

The annealing behavior in both n- and p-type SSD needs to be verified with
CCE measurements.
Hartmut F.-W. Sadrozinski, US ATLAS Upgrade Meeting Nov 10, 2005
22
R&D Plan:
- Need to confirm findings of C-V measurements
- Fabricate SSD on MCz wafers, both p- and n-type.
- Optimize isolation on n-side.
- Measure charge collection efficiency (CCE) on SSD,
pre-rad, post-rad, during anneal.
- Measure noise on SSD pre-rad, post-rad, during anneal.
Un-irradiated SMART SSD
Hartmut F.-W. Sadrozinski, US ATLAS Upgrade Meeting Nov 10, 2005
23
R&D Plan
Submission of 6” fabrication run within RD50
Goals:
-a.
-b.
-c.
-d.
-e.
-f.
-g.
P-type isolation study
Geometry dependence
Charge collection studies
Noise studies
System studies: cooling, high bias voltage operation,
Different materials (MCz, FZ, DOFZ)
Thickness
Wafer
MCz
DOFZ
FZ
MCz
Fz
MCz
bulk
p
p
p
n
n
n
#
7
5
5
3
2
3
Thickness
[um]
SSD
300
n-on-p
300
n-on-p
300
n-on-p
300
p-on-n +n-on-n (no backside
300
p-on-n +n-on-n (no backside
200
p-on-n +n-on-n (no backside
Hartmut F.-W. Sadrozinski, US ATLAS Upgrade Meeting Nov 10, 2005
24