Studies of bulk damage induced in
different silicon materials by 900 MeV
electron irradiation
S. Dittongo1,2, L. Bosisio1,2, D. Contarato3, G. D’Auria4,
E.Fretwurst5, J. Härkönen6, G. Lindström5, E. Tuovinen6
Dipartimento di Fisica, Universita’ di Trieste, Italy
2 INFN, Sezione di Trieste, Italy
3 DESY, Hamburg, Germany
4Sincrotrone Trieste S.C.p.A., Trieste, Italy
5 Institut für Experimentalphysik, Universität Hamburg, Germany
6 Helsinki Institute of Physics, Helsinki, Finland
1
Overview
• Introduction: why high-energy electrons?
• Irradiation:
– irradiated devices
– experimental conditions
• Experimental results:
– Effective dopant concentration
– Reverse leakage current, damage constant a
– Isothermal annealing effects
• Conclusions
Selenia Dittongo
IWORID 2004 – Glasgow, 25-29th July
2
Why high-energy electrons?
• In the last years, much effort devoted to study
– the radiation hardness of silicon detectors against different
particle types (charged hadrons, neutrons and g rays)
– the improvements achievable by using different silicon substrates
• By contrast, very few contributions devoted to damage induced by
high-energy (GeV) electrons
• In previous irradiations, we demonstrated the effectiveness of 900
MeV electrons in creating bulk damage
• In this work we extend these investigations to a wider sample of
silicon materials (standard and oxygenated float-zone, magnetic and
non-magnetic Czochralski, epitaxial silicon) and by reaching very high
fluence levels (6x1015 e/cm2)
Selenia Dittongo
IWORID 2004 – Glasgow, 25-29th July
3
Tested devices
p+/n-/n+ and n+/p-/p+ diodes fabricated on different silicon substrates (thickness
~300 mm), provided with a 100 mm wide guard ring, surronded by floating rings
material
substrate
processed by
resistivity
[O]
FZ & DOFZ
Wacker
(111) n-type
CiS
(Erfurt, Germany)
3-4 kW cm
1.2x1017 cm-3
(DOFZ)
FZ & DOFZ
Wacker
(111) n-type
Helsinki Institute
of Physics
1.2 kWcm
not measured
CZ
Sumitomo
(100) n-type
CiS
1.2 kW cm
8x1017 cm-3
Magnetic CZ
Okmetic
(100) n-type
Helsinki
1.0 kW cm
5-9x1017 cm-3
Magnetic CZ
Okmetic
(100) p-type
Helsinki
1.8 kW cm
5-9x1017 cm-3
ITME
CZ (111) n-type
CiS
50 Wcm
9x1016 cm-3
EPI
Selenia Dittongo
IWORID 2004 – Glasgow, 25-29th July
4
Experimental conditions
Irradiations
•
performed with the 900 MeV electron beam of the LINAC injector at Elettra
(Trieste, Italy)
•
fluence measured by a toroidal coil coaxial with beam
•
devices kept unbiased during irradiation, at room temperature (~25oC)
step
Fluence
(e/cm2)
Measurements
1
(1.06±0.01±0.04)x1015
2
(1.97±0.01±0.08)x1015
• devices electrically characterized by reverse I-V
and C-V measurements, performed ~1 day after
irradiation
3
(4.16±0.05±0.17)x1015
• C-V measurements @ 10 kHz
4
(4.97±0.05±0.20)x1015
5
(6.11±0.02±0.25)x1015
• currents normalized to 20oC
Selenia Dittongo
• isothermal annealing cycles up to 16 hours @ 80oC
IWORID 2004 – Glasgow, 25-29th July
5
Photocurrent vs. bias voltage
a quick and simple method to check for type inversion
• the diode is illuminated on the front or back side by the (dimmed) microscope light
• the I-V curve in the dark is subtracted from the I-V curve with light, to obtain the
photogenerated current vs. bias voltage
• comparing the voltage dependence of the photocurrent for the two situations (front and
back illumunation) helps in determining whether the substrate is inverted
15
DOFZ Helsinki -  = 4x1015 e/cm2
2
DOFZ Helsinki - = 2x10 e/cm
before type inversion
1.5E-06
Front_Illumination
VTD
5.0E-07
Back_Ilumination
0.0E+00
Photocurrent (A)
Photocurrent (A)
2.0E-06
1.0E-06
after type inversion
2.0E-06
1.5E-06
Front_Illumination
1.0E-06
Back_Ilumination
5.0E-07
VTD
0.0E+00
0
10
20
30
40
50
60
70
80
90 100
0
10
20
30
40
V_bias (V)
Selenia Dittongo
50
60
70
80
90
V_bias (V)
IWORID 2004 – Glasgow, 25-29th July
6
100
Effective dopant concentration: FZ & DOFZ
type inversion occurs at:
3x1015
e/cm2
~
for Helsinki devices
(higher initial doping)
5E+12
3
•
 ~ 5x1014 e/cm2 for CiS devices
(from previous irradiations)
Neff (1/cm )
•
4E+12
FZ - Helsinki
3E+12
DOFZ - Helsinki
2E+12
FZ - CiS
DOFZ - CiS
1E+12
0
-1E+12
-2E+12
post-inversion behaviour
-3E+12
-4E+12
•
lower Neff variations for DOFZ devices
•
Neff(Helsinki devices) ~ 70-80% Neff(CiS devices)
•
differences between Helsinki and CiS diodes likely due to different starting materials and oxygenation
procedures (CiS: oxygen diffusion performed in N2 environment for 72 hours @1150oC; Helsinki: 8 hour
oxidation in O2 @ 1050oC + 60 hour diffusion in N2 @ 1050oC)
Selenia Dittongo
0
2E+15
4E+15
6E+15
8E+15
2
fluence (e/cm )
IWORID 2004 – Glasgow, 25-29th July
7
Effective dopant concentration: CZ & MCZ
n-type CZ & MCZ devices:
n-type CZ & MCZ
6E+12
•
Neff decreases, reaches a minimum at ~4x1015
e/cm2 and then increases again
•
type inversion not observed even if the preirradiation dopant concentration is comparable
with FZ & DOFZ made in Helsinki
•
the oxygen concentration is higher than
for DOFZ (under investigation)
3
Neff (1/cm )
5E+12
CZ - CiS
4E+12
MCZ - Helsinki
3E+12
-
2E+12
1E+12
0
0
2E+15
4E+15
6E+15
8E+15
fluence (e/cm2)
p-type MCZ
5E+12
p-type MCZ devices:
•
substrate type inversion is not approched
very slow rate of decrease in Neff
(1.2x10-4 cm-1)
N eff (1/cm3 )
•
4E+12
3E+12
2E+12
1E+12
0
0
2E+15
4E+15
6E+15
8E+15
2
fluence (e/cm )
Selenia Dittongo
IWORID 2004 – Glasgow, 25-29th July
8
Effective dopant concentration: EPI
type inversion not observed:
the pre-irradiation dopant
concentration is even higher than
for CZ substrates...
•
the variations of Neff are small (<15%)
•
the rate of decrease in Neff is slow
(1.5x10-3 cm-1)
8E+13
7E+13
6E+13
N eff (1/cm3 )
•
5E+13
4E+13
3E+13
2E+13
1E+13
0
0
2E+15
4E+15
6E+15
8E+15
fluence (e/cm2 )
Selenia Dittongo
IWORID 2004 – Glasgow, 25-29th July
9
Leakage current & damage constant a
the increase in leakage current does not depend on substrate material
(as already observed after hadron irradiations)
after ~1 day @ room temperature
a =7.9x10-19 A/cm
6.E-03
hardness factor
kexp = a(900 MeV e-)/a(1 MeV n) = 2.0x10-2
Jleak (A/cm3)
5.E-03
4.E-03
at equal NIEL, high energy electrons are
~4.1 times less effective than 1 MeV neutrons
in degrading the carrier generation lifetime of
substrate material
3.E-03
2.E-03
EPI
FZ - CiS
MCZn
DOFZ Helsinki
1.E-03
0.E+00
0
2E+15
4E+15
fluence (e/cm2)
Selenia Dittongo
CZ
DOFZ - CiS
FZ - Helsinki
MCZp
6E+15
ktheo = NIEL(900 MeV e-)/NIEL(1 MeV n) = 8.1x10-2
ktheo/ kexp = 4.1
8E+15
the NIEL scaling hypothesis is not adequate
when comparing electrons with hadrons
IWORID 2004 – Glasgow, 25-29th July
10
Annealing of leakage current
•
The time evolution of the leakage current density for devices made on
different substrates and irradiated at different fluences follows a
common functional dependence on the annealing time
1.2
Jleak / Jleak,0
1.0
0.8
0.6
0.4
EPI - 6e15 e/cm2
CZ - 6e15 e/cm2
MCZn - 6e15 e/cm2
FZ Hels. - 4e15 e/cm2
DOFZ Hels - 4e15 e/cm2
0.2
0.0
0.1
1
CZ - 4e15 e/cm2
MCZn - 4e15 e/cm2
FZ Hels. - 2e14 e/cm2
DOFZ Hels - 2e15 e/cm2
10
100
1000
annealing time (min)
Selenia Dittongo
IWORID 2004 – Glasgow, 25-29th July
11
Annealing of Neff: FZ and DOFZ by Helsinki
•
before type inversion, a maximun in the effective acceptor concentration is reached after ~15
minutes, followed by a decrease
– in FZ device, substrate type inversion is observed after 4 hours @ 80oC
•
after type inversion, a minimum in the effective acceptor concentration is reached after
~15 (FZ)-60 (DOFZ) minutes (beneficial annealing), followed by an increase (reverse annealing)
•
slightly higher effect in FZ devices
 = 2x1015 e/cm2 - before type inversion
 = 4x10
100%
2
e/cm - after type inversion
80%
60%
D Neff / Neff 0
0%
D Neff / Neff0
15
-100%
-200%
FZ
DOFZ
type
inversion
-300%
FZ
DOFZ
40%
20%
0%
-20%
-40%
-400%
-60%
0.1
1
Selenia Dittongo
10
annealing time (min)
100
1000
0.1
1
10
annealing time (min)
IWORID 2004 – Glasgow, 25-29th July
100
1000
12
Annealing of Neff: CZ, MCZn and EPI
EPI devices (non-inverted):
•
15
e/cm
2
4.0%
after highest fluence (6x1015 e/cm2),
show an increase of effective donor
concentration with time
variations of the order of a few %
3.0%
D N eff / N eff,0
•
EPI devices -  = 6x10
2.0%
1.0%
0.0%
-1.0%
0.1
CZ & MCZn
100%
80%
10
annealing time [min]
100
1000
CZ - 4e15 e/cm2
CZ - 6e15 e/cm2
60%
D Neff / Neff,0
1
CZ and MCZn devices (non-inverted):
MCZn - 4e15 e/cm2
40%
MCZn - 6e15 e/cm2
•
20%
0%
-20%
•
-40%
-60%
0.1
1
10
annealing time (min)
Selenia Dittongo
100
show an oscillating behaviour, observed
also after hadron irradiation. Possible
reasons under investigation...
oscillation amplitude is larger at lower
fluences and for CZ devices
1000
IWORID 2004 – Glasgow, 25-29th July
13
Conclusions
• Leakage current: no difference is observed among different
materials, as already known after hadron irradiation
•
Neff:
• FZ & DOFZ: a slightly beneficial effect of oxygen diffusion is
observed; both show a saturation trend beyond 4x1015 e/cm2
• n-type CZ & MCZ: high fluence irradiations show that the
linear trend observed at lower fluences does not continue up
to type inversion; instead, Neff goes back up after reaching a
minimun at ~4x1015 e/cm2
• EPI & p-type MCZ: substrate type inversion is not even
approched up to 6x1015 e/cm2 ; very slow rate of decrease in
Neff (1.5x10-3 cm-1 in EPI, 1.2x10-4 cm-1 in p-MCZ)
Selenia Dittongo
IWORID 2004 – Glasgow, 25-29th July
14
a: comparison with other particles
•
Use asymptotic value of displacement cross section for 200 MeV electrons
[G.P.Summers et al., IEEE Trans.Nucl.Sci. 40(6) (1993), 1372-1378] (no higherenergy values available)
NIEL(900 MeV e)/NIEL(1 Mev n) = 8.106x10-2
particles
a(A/cm)
measured
1 MeV n equiv.
1.8 MeV e-
4.5x10-20
1.9x10-18
900 MeV e-
9.06x10-19
1.12x10-17
1 MeV n
4.0x10-17(*)
a (900 MeV e) / a (1 MeV n)
NIEL(900 MeV e) / NIEL(1 MeV n)
= 0.24 =
1
4.1
(*) M.Moll et al., NIM, vol. A426, pp.87-93, 1999
Selenia Dittongo
IWORID 2004 – Glasgow, 25-29th July
15