Measurements and Simulations of
Charge Collection Efficiency of
p+/n Junction SiC Detectors
Francesco Moscatelli1,2, Andrea Scorzoni2,1, Antonella Poggi1, Mara
Bruzzi3, Stefano Lagomarsino3 , Silvio Sciortino3 , Mihai Lazar4 and
Roberta Nipoti1
1CNR-
IMM Sezione di Bologna, via Gobetti 101, 40129 Bologna, Italy
and INFN, Università di Perugia, via G. Duranti 93, 06125 Perugia, Italy
3Dipartimento di Fisica, Polo Scientifico di Sesto Fiorentino,Via Sansone 1 Firenze Italy
4CEGELY (UMR CNRS n°5005), INSA de Lyon, 20, Av. A. Einstein, 69621 Villeurbanne, France
2DIEI
This work was partially supported by the CERN RD50
Collaboration.and by the INFN SiCPOS project
Università degli Studi di Perugia
IMM Bologna
1
Outline
• Introduction
• Technological processes and I/V - C/V
measurements on p+/n diodes
• CCE setup and measurements
• Modeling of SiC detectors
– Motivations and simulation tool
– Results
• Conclusions
Università degli Studi di Perugia
IMM Bologna
2
Introduction
• Large Hadron Collider
(LHC) experiment (upgrade)
• Fast hadron fluences above
1016 cm-2 (after 10 years)
• Current silicon technology is
unable to cope with such an
environment
– Unreachable full depletion
voltage
– Very high leakage current
– Poor charge collection
efficiency
Università degli Studi di Perugia
IMM Bologna
3
Silicon Carbide
- large Eg (3-3.3 eV)
very low leakage current
- MIP (Minimum Ionizing Particle) generates 55 e/h pairs per m
- radiation hardness (?) (high atomic binding within the material)
- high quality crystals now available
- Schottky barrier detectors have been studied as -particle
detectors (100% of charge collection efficiency (CCE))*
- complex radiation detectors
an integrated electronic readout
on board of the detector chip. p/n junctions are needed
* F.
Nava, et al. , IEEE Transactions on Nuclear Science, Vol. 51, No. 1 (February, 2004).
Università degli Studi di Perugia
IMM Bologna
4
SiC Process: p+/n
n-
Epi (40 m) doping:
1 1015 cm-3
Al (350 nm) /
Ti (80 nm)
deposition
Ion implantation
Al+ @ 300°C
Annealing 1650°C 30 min
p+ doping (0.4 m)
= 41019 cm-3
p- doping (0.6 m)
= 51017 cm-3
Annealing 1000°C
in vacuum 2 min
n+
Università degli Studi di Perugia
IMM Bologna
5
I-V measurements
on p+/n diodes
In collaboration with INSACEGELY, Lyon France
3
0
1x10
1x10
n=1.5-1.6
-3
1x10
-5
1x10
n=2
-4
1x10
VBD
-6
1x10
-9
0
-7
1
2
3
4
V (V)
•
•
•
-3
1x10
-5
-7
1x10
-2
1x10
1x10
1x10
Ni
Reverse I-V
-1
1x10
-1
2
n
n+
Forward I-V
2
p+
1
J (A/cm )
1x10
J (A/cm )
p-
Ti-Al
1x10
5
1x10
-4000
-3000
-2000
V (V)
-1000
0
75% of diodes have good I-V curves
VBD is about 4 kV
Theoretical limit for this device: 5 kV
Università degli Studi di Perugia
IMM Bologna
6
CV measurements
1E16
Ti-Al
Ni
-3
n
n+
Ndop [cm ]
p+
1E15
5
10
15
20
25
30
Depth [m]
• Epi doping (1.11015 cm-3) confirmed
Università degli Studi di Perugia
IMM Bologna
7
CCE measurements setup
90Sr
0.1mCi
Amptek
Acquisition
system
S+PM trigger
Università degli Studi di Perugia
IMM Bologna
8
CCE measurements
32
Collection lenght
depleted region
p+/n
1600
28
L [m]
Number of collected charges
• Measurements on p+/n diodes: epi 1.1.1015 cm-3 40 µm,
• Max. applied voltage: 900V (30 µm depleted). Vdep (from theory) =1600V
1200
24
L
20
collected charge
55 pairs/ m
16
800
12
200
400
600
800
Reverse voltage [V]
200
400
600
800
Reverse voltage [V]
• 100% collection efficiency in the 30 µm deep depleted region using
measured lengths of depleted region
Università degli Studi di Perugia
IMM Bologna
9
Motivations for simulation
• Very high cost of SiC wafers
• Trade off between SiC wafer quality and available budget
• Suitability of device simulation for design optimization
• Introducing traps, we will be able to analyze which defects
are important to decrease the CCE
Simulation Tool
• DESSIS ISE-TCAD
– Discrete time and spatial solution to the
fundamental semiconductor equations
– 6H-SiC model available
Università degli Studi di Perugia
IMM Bologna
10
Grid and Heavy Ion crossing
Depleted region
Hole
density
Before crossing
After crossing
Heavy Ion crossing modeling available within DESSIS ISE-TCAD
Università degli Studi di Perugia
IMM Bologna
11
p+/n diode output signal
CC  0 I  dt
t
• Particle crossing
at 2.5 ns
Current [A]
• Collected Charge
3x10
-6
2x10
-6
1x10
-6
0
2x10
-9
3x10
-9
4x10
-9
Time [s]
epi doping (40 m) = 1015 cm-3
p+ doping (0.45 m) = 41019 cm-3
Università degli Studi di Perugia
IMM Bologna
12
Simulations of CC of Schottky diodes
50 m
Collection length (90Sr source, β particles )
Schottky contact:qB=1.6 eV
40
38 µm
35
L (m)
epi n
substrate
n+
30
25
simulations
measurements
20
15
0
20
40
60
80
100
Voltage [V]
* Measurements from:
F. Nava, et al. , IEEE Transactions on
Ohmic contact
Nuclear Science, Vol. 51, No. 1 (February, 2004).
Università degli Studi di Perugia
IMM Bologna
13
Simulations of CC of p+/n diodes
p+
measurements
simulations
40 µm
28
L [m]
ND = 1.11015 cm-3
32
ND = 71018 cm-3
60 µm
24
20
16
12
200
400
600
800
Reverse voltage [V]
50 µm
Università degli Studi di Perugia
CCE experimental results are
well reproduced
IMM Bologna
14
Conclusions
• p+/n junctions have been realized and
electrically characterized. Good forward and
reverse characteristics have been obtained
• First CCE experimental results on SiC pn
junctions: 100% collection efficiency in 30 m
using measured lengths of depleted region
• Development of a simulation model for SiC to
obtain good agreement with CC measurements
on Schottky and p+/n SiC diodes
Università degli Studi di Perugia
IMM Bologna
15
Future developments
• Radiation hardness will be verified.
• New SiC detectors will be realized taking into
account the simulation results.
• Using DLTS measurements and simulations,
we will be able to analyze which defects are
important to decrease the CCE
Università degli Studi di Perugia
IMM Bologna
16