Transcript Document

Epitaxial Silicon Detectors for Particle Tracking
Overview on Radiation Tolerance at Extreme Hadron Fluence
G. Lindström(a), E. Fretwurst(a), F. Hönniger(a), G. Kramberger(b),
M. Moll(c), E. Nossarzewska(d), I. Pintilie(e), R. Röder(e)
(a) Institute for Experimental Physics, University of Hamburg
(b) Jozef Stefan Institute, University of Ljubljana
(c) CERN
(d) ITME Institute of Electronic Materials Technology, Warsaw
(e) National Institute for Materials Physics, Bucharest
(f) CiS Institut für Mikrosensorik gGmbH, Erfurt
Joint Project: Bucharest-Hamburg-Ljubljana
with ITME (EPI), CiS (diodes), CERN (PS-p)
 Motivation
 Material parameters
 Results for proton and neutron irradiated devices
 S-LHC scenario: simulation and experimental data
 Summary
1
G. Lindström, University of Hamburg, 10th European Symposium on Semiconductor Detectors, Wildbad Kreuth,12-16 June 2005
Motivation
 LHC upgrade to Super-LHC
Luminostity LHC: L ~ 1034 cm-2s-1  S-LHC: L ~ 1035 cm-2s-1
Expected radiation at inner pixel layer (4 cm) under S-LHC conditions:
Fluence for fast hadrons: 1.61016 cm-2 Dose: 420 Mrad
 Proposed Solution: thin EPI-SI for small size pixels
Si best candidate: low cost, large availability, proven technology
thickness doesn’t matter! CCE limited at 1016 cm-2 by e  20 m, h  10 m
small pixel size needed, allows thin devices with acceptable capacitance for S/N
high Neff,0 (low ρ) provides large donor reservoir, delays type inversion
 Needed data for prediction of S-LHC operational scenario:
 Reliable projection of damage data to S-LHC operation
 Full annealing cycles at elevated temperatures
 Charge collection efficiency at very high fluences
2
G. Lindström, University of Hamburg, 10th European Symposium on Semiconductor Detectors, Wildbad Kreuth,12-16 June 2005
Material Parameters
75 mu
50 mu
25 mu
102
18
10
5
SIMS 25 m
SIMS 50 m
SIMS 75 m
simulation 25 m
simulation 50 m
simulation 75m
1017
5
16
10
5
0
10 20 30 40 50 60 70 80 90 100
Depth [m]
 Oxygen depth profiles
SIMS-measurements after diode processing
O diffusion from substrate into epi-layer
Resistivity [ cm]
O-concentration [1/cm3]
5
101
100
-1
10
25 m, C-V method
50 m, C-V method
50 m, spreading resistance
75 mum, C-V method
10-2
0
20
40
60
Depth [m]
80
 Resistivity profiles
SR  before diode process, C-V on diodes
interstial Oi + dimers O2i
SR coincides well with C-V method
[O] 25 µm > [O] 50 µm
Excellent homogeneity in epi-layers
process simulation yields reliable [O]
(SIMS-measurements: A. Barcz/ITME, Simulations: L. Long/CiS)
3
100
(SR-measurements: E. Nossarzewska, ITME)
G. Lindström, University of Hamburg, 10th European Symposium on Semiconductor Detectors, Wildbad Kreuth,12-16 June 2005
Typical Annealing Curves
140
1000
23 GeV protons
120
23 GeV protons
Ta=80oC
o
Vfd [V]
80
50 m EPI, 9.7.1014 cm-2
60
50 m FZ, 1.1.1015 cm-2
Vfd [V]
Ta=80 C
100
100
40
9.1015 cm-2
6.1015 cm-2
2.1015 cm-2
20
1.1015 cm-2
3.1014 cm-2
0 0
10
10
101
102
103
Annealing time [min]
104
105
100
101
102
103
Annealing time [min]
104
105
Comparison EPI- with FZ-device:
Typical annealing behavior of EPI-devices:
Vfd development:
Vfd development:
short term: EPI increasing, FZ decreasing
long term: EPI decreasing, FZ increasing
Inversion only(!) during annealing ()
 EPI not inverted, FZ inverted, already
 EPI never inverted at RT (t < 500d),
even for 1016
immediately after irradiation
4
(100 min @ 80C ≈ 500 days @ RT)
G. Lindström, University of Hamburg, 10th European Symposium on Semiconductor Detectors, Wildbad Kreuth,12-16 June 2005
Parameterization of Annealing Results
Change of effective “doping“ concentration: Neff = Neff,0 – Neff (,t(T))
Standard parameterization: Neff = NA(,t(T)) + NC() + NY(,t(T))
Annealing components:
5.0.1013
Neff [ cm-3]
 Short term annealing  NA(,t(T))
Ta=80oC
4.0.1013
 Stable damage  NC()
3.0.1013
NC = NC0(1-exp(-cΦeq) + gCΦeq
gC negative for EPI (effective positive
space charge generation!)
NY,2
2.0.1013
NA
.
NY,1
13
1.0 10
 Long term (reverse) annealing:
NC
0.0 0
10
5
1
10
2
3
10
10
Annealing time [min]
4
10
5
10
Two components:
 NY,1(,t(T)), first order process
 NY,2(,t(T)), second order process
NY1, NY2 ~ Φeq, NY1+NY2 similar to FZ
G. Lindström, University of Hamburg, 10th European Symposium on Semiconductor Detectors, Wildbad Kreuth,12-16 June 2005
Stable Damage Component
2.0.1014
25 m, 80oC
50 m, 80oC
400
50 m, 60 C
o
300
1.0.1014
200
100
0.0
0
2.1015
4.1015 6.1015
eq [cm-2]
8.1015
1016
0
150
100
5.1013
50 m
25 m
0
0
2.1015 4.1015 6.1015 8.1015
eq [cm-2]
1016
Neff(t0): Value taken at annealing time t0 at which Vfd maximum
 No space charge sign inversion after proton and neutron irradiation
Introduction of shallow donors overcompensates creation of acceptors
 Protons: Stronger increase for 25 µm compared to 50 µm
 higher [O] and possibly [O2] in 25 µm (see SIMS profiles)
 Neutrons: Similar effect but not nearly as pronounced
most probably due to less generation of shallow donors
and as strong influence of acceptors
6
50
G. Lindström, University of Hamburg, 10th European Symposium on Semiconductor Detectors, Wildbad Kreuth,12-16 June 2005
0
Vfd (t0)[V] normalized to 50 m
Neff(t0) [cm-3]
500
Reactor neutrons
Ta = 80oC
Neff (t0) [cm-3]
600
23 GeV protons
3.0.1014
Vfd(t0) [V] normalized to 50 m
1014
Shallow Donors, the real issue for EPI
-Comparison of 25, 50 and 75 µm Diodes-
SIMS 25 m
SIMS 50 m
SIMS 75 m
simulation 25 m
simulation 50 m
simulation 75m
1017
5
1016
5
0
10 20 30 40 50 60 70 80 90 100
Depth [m]
SIMS profiling:
[O](25µm) > [O](50µm) > [O](75µm)
Stable Damage:
Neff(25µm) > Neff(50µm) > Neff(75µm)
TSC Defect Spectroscopy:
[BD](25µm) > [BD](50µm) >
[BD](75µm)
2.1014
Neff(t0) [cm-3]
1018
5
23 GeV protons
Stable Damage after
PS p-irradiation
25 m, 80 oC
50 m, 80 oC
75 m, 80 oC
Generation of
positive space charge
most pronounced in
25 µm diodes
1014
0
0
TSC (pA/
signal(pA/
NormalisedNormalised
m) m)
TSC signal
75 mu
50 mu
SIMS profiling of [O]
25 mu
O-concentration [1/cm3]
5
2.1015
4.1015 6.1015
-2
75 m eq [cm ]
0.4
50 m
25 m
1.2
8.1015
1016
CiOi
-/0
V2
CiOi
+?
0/++
BD
1.0
-/0
V2 +?
0.2
0.8
0.6
0.0
0.4
80
100
0/++
BD
0.2
120
140
160
75 m
Temperature (K)
50 m
25 m
0.0
80
100
120
140
Temperature (K)
160
Defect spectroscopy
after PS p-irradiation
Generation of recently
found shallow donors
BD (Ec-0.23 eV)
180strongly related to [O]
Possibly caused by
O-dimers, outdiffused
from Cz with larger
180
diffusion constant
dimers monitored
by IO2 complex
Strong correlation between [O]-[BD]-gC
generation of O (dimer?)-related BD reason for
superior radiation tolerance of EPI Si detectors
7
G. Lindström, University of Hamburg, 10th European Symposium on Semiconductor Detectors, Wildbad Kreuth,12-16 June 2005
Annealing Curves for 25 µm and 50 µm
4.0.1014
.
23 GeV protons
Ta=80oC
14
Neff [cm-3]
3.0 10
2.0.1014
1.0.1014
50 m
25 m
0.0
-1.0.1014
-2.0.1014 0
10
101
102
103
Annealing time [min]
104
105
 Nearly identical time dependence, constant difference between both curves
 Shift due to higher introduction of shallow donors in 25 µm epi-Si
 Concentration of donors not affected by long term annealing
 Radiation induced donors are stable at 80°C (verified by TSC)
8
G. Lindström, University of Hamburg, 10th European Symposium on Semiconductor Detectors, Wildbad Kreuth,12-16 June 2005
Examples of parameter evaluation from annealing:
Fluence Dependence of NY1
23 GeV protons
.
2.0.1014
25 m, 80oC
1.0.1014
0.0
0
2.1015
4.1015 6.1015
eq [cm-2]
Reactor neutrons
Ta = 80oC
1.0.1014
50 m, 80oC
25 m
50 m, 60oC
50m
8.1015
1016
 Introduction rate for protons:
Independent of epi-layer thickness and
annealing temperature
gY1 = 3.110-2 cm-1
9
14
2.0 10
NY1 [cm-3]
NY1 [cm-3]
3.0.1014
0.0
0
2.1015
4.1015 6.1015
eq [cm-2]
8.1015
1016
 Introduction rate for neutrons:
Value for 25 µm slightly lower compared
to 50 µm, average value:
gY1 = 1.710-2 cm-1
G. Lindström, University of Hamburg, 10th European Symposium on Semiconductor Detectors, Wildbad Kreuth,12-16 June 2005
Reverse Annealing Time Constants
104
105
23 GeV protons
Reactor neutrons
Ta = 80oC
o
60 C
104
10
80oC
102
101
0
2.1015
106
4.1015 6.1015
eq [cm-2]
8.1015
1016
104
60oC
80oC
50 m, 60oC
50 m, 80oC
25 m, 80oC
103
1014
25 m
25 m
50 m
50 m
2.1015
4.1015 6.1015
eq [cm-2]
8.1015
1016
 Protons:
5
10
0
Y,2
Y,1
101
23 GeV protons
Y,2 [min]
103
102
50 m
50 m
25 m
10
Y [min]
Y,1 [min]
3
above 1015 cm-2 Y1 and Y2  constant
Ratio Y1(60°C)/ Y1(80°C) = 11
Ratio Y2(60°C)/ Y2(80°C) = 10
below 1015 cm-2 Y2  1/eq
 Neutrons:
above 21015 cm-2 Y1 and Y2  constant
1015
eq [cm-2]
1016
Y1(50 µm) = 1.3102 min, Y1(25 µm) = 3.3102 min
Y2(50 µm) = 2.9103 min, Y2(25 µm) = 5.3103 min
G. Lindström, University of Hamburg, 10th European Symposium on Semiconductor Detectors, Wildbad Kreuth,12-16 June 2005
Temperature Dependence of Y1 and Y2
23 GeV protons
100 oC 80 oC
106
5
60 oC
40 oC
20 oC
Ea = 1.3 eV
Y [min]
10
104
103
Y : FZ-Si
102
Y1: EPI-Si
Y2: EPI-Si
101
100
2.6
2.8
3
3.2
-3 -1
1/T [10 K ]
 Arrhenius relation: 1/ Y = k0exp(-Ea/kBT)
3.4
High resistivity FZ Si:
M. Moll, PhD thesis
Y1 component: Ea = 1.3 eV, k0 = 1.5  1015 s-1
Y2 component: Ea = 1.3 eV, k0 = 5.4  1013 s-1
11
G. Lindström, University of Hamburg, 10th European Symposium on Semiconductor Detectors, Wildbad Kreuth,12-16 June 2005
Reliability check: Annealing at 20°C
Example: EPI 50 µm, Φp = 1.01·1016 cm-2
300
250
parameter
20°C data fit
80°C data fit
-8.9·1013 cm-3
-8.6 ·1013 cm-3
150
Stable dam.
NC
3 ·1013 cm-3
5 ·1013 cm-3
100
Benef. Ann.
Na
Rev. Ann.
NY1
2.2 ·1014 cm-3
1.8 ·1014 cm-3
experimental data
fit: Hamburg model
VFD [V]
200
50
0
0
50
100
150
200
250
annealing time at 20oC [days]
 20°C annealing results can be fitted with Hamburg model
 RT Analysis agrees reasonably well with results from 60°, 80°C
 Projection from high T annealing to behavior at RT reliable
12
G. Lindström, University of Hamburg, 10th European Symposium on Semiconductor Detectors, Wildbad Kreuth,12-16 June 2005
Charge collection efficiency
Charge Collection Efficiency
23 GeV protons
1.0
25 m, 80oC
0.9
50 m, 80oC
50 m, 60oC
 CCE degradation linear with fluence if the
devices are fully depleted
CCE = 1 –  ,  = 2.710-17 cm2
0.8
0.7
0.6
0
 Charge collection efficiency measured with
244Cm -particles (5.8 MeV, R30 µm)
Integration time window 20 ns
 CCE(1016 cm-2) = 70 %
2.1015
4.1015 6.1015
eq [cm-2]
8.1015
1016
 Charge collection efficiency measured with
90Sr electrons(mip’s), shaping time 25 ns
 CCE no degradation at low temperatures !
degradation linear with fluence if the devices
CCE measured after n- and p-irradiation
 CCE(Φp=1016 cm-2) = 2400 e (mp-value)
Extracted trapping parameters coincide with
those measured in FZ diodes for small Φ, for
large Φ less trapping than expected !
13
G. Lindström, University of Hamburg, 10th European Symposium on Semiconductor Detectors, Wildbad Kreuth,12-16 June 2005
Current Generation
0.5
Ta = 80oC
ta = 8 min
I/V [Acm-3]
0.4
0.3
50 m
25 m
0.2
I =  x eq x Vol
0.1
0.0
0
2.1015
4.1015 6.1015
eq [cm-2]
8.1015
1016
Result almost identical to FZ silicon:
Current related damage rate α = 4.1·10-17 Acm-1
(Small deviations in short term annealing)
14
G. Lindström, University of Hamburg, 10th European Symposium on Semiconductor Detectors, Wildbad Kreuth,12-16 June 2005
S-LHC CERN scenario experiment
Proposal: Storage of EPIdetectors during beam off
periods at RT (in contrast to
required cold storage for FZ)
250
50 m after 50 min@80C annealing
200
Vfd [V]
Stable donor generation at
high Φ would lead to larger
Vfd, but acceptor generation
during RT anneal could
compensate this.
150
50 m simulation
100
Check by
dedicated experiment:
 Experimental parameter:
Irradiation:
fluence steps  2.21015 cm-2
irradiation temperature  25°C
After each irradiation step
annealing at 80°C for 50 min,
corresponding 265 days at 20°C
15
25 m after 50 min@80C annealing
50
0
0
25 m simulation
2.1015
4.1015
6.1015 8.1015
eq [cm-2]
1016
 Simulation:
reproducing the experimental scenario
with damage parameters from analysis
Excellent agreement between
experimental data and simulated results
 Simulation + parameters reliable!
G. Lindström, University of Hamburg, 10th European Symposium on Semiconductor Detectors, Wildbad Kreuth,12-16 June 2005
S-LHC operational scenario
simulation results
600
Vfd [V]
S-LHC scenario
500
50 m cold
50 m warm
400
25 m cold
S-LHC: L=1035cm-2s-1
Most inner pixel layer
operational period per year:
100 d, -7°C, Φ = 3.48·1015cm-2
25 m warm
300
200
beam off period per year
265 d, +20°C (lower curves)
-7°C (upper curves)
100
0
0
365
730
1095
time [days]
1460
1825
 RT storage during beam off periods extremely beneficial
 Damage during operation at -7°C compensated by 100 d RT annealing
 Effect more pronounced for 50 µm: less donor creation, same acceptor component
 Depletion voltage for full SLHC period less than 300 V
16
G. Lindström, University of Hamburg, 10th European Symposium on Semiconductor Detectors, Wildbad Kreuth,12-16 June 2005
Summary
Thin low resistivity EPI diodes (grown on Cz) are extremely radiation tolerant
No type inversion observed up to Φeq = 1016 cm-2 for protons and neutrons,
space charge sign remains positive even during long time annealing
Annealing experiments at elevated temperatures reveal stable donor generation
(in contrast to FZ), overcompensating the damage induced acceptor concentration
Elevated temperature annealing results verified at 20°C
Dedicated CERN scenario experiment shows benefits of RT storage in contrast to
required cooling for FZ silicon, experimental results in excellent agreement with
simulations using the Hamburg model parameters
Simulation of real SLHC operational scenario with RT annealing during beam off
periods show that Vfd kept at <300V (50μm), <150V (25μm) for full 5 y SLHC
Donors identified: BD point defects (EC-0.23 eV) most likely related to O-dimers
strong correlation found between [O](SIMS)-[BD](TSC)-NC(annealing exp.)
Trapping almost as expected from FZ, 2400 e measured for mips after 1016 cm-2
Current related damage rate 4·10-17 Acm-1 (as known for FZ)
17
G. Lindström, University of Hamburg, 10th European Symposium on Semiconductor Detectors, Wildbad Kreuth,12-16 June 2005