To cite this version : Belloir, Jean-Marc and Goiffon, Vincent and Virmontois, Cédric and Paillet, Philippe and Raine, Mélanie and Gilard, Olivier and Magnan, Pierre Dark Current Spectroscopy on Alpha Irradiated Pinned Photodiode CMOS Image Sensors. (2015) In: Proceedings of RADECS 2015, 14 September 2015 - 18
September 2015 (Moscou, Russian Federation).
Open Archive TOULOUSE Archive Ouverte (OATAO)
OATAO is an open access repository that collects the work of Toulouse researchers and makes it freely available over the web where possible.
This is an author-deposited version published in : http://oatao.univ-toulouse.fr/
Dark Current Spectroscopy on Alpha
Irradiated CMOS Image Sensors
J.-M. Belloir, V. Goiffon, P.Magnan, ISAE-SUPAERO, Toulouse, France, C. Virmontois, O. Gilard, CNES, Toulouse, France,
Context and goal
Space and nuclear environments contain particles which can displace atoms and create stable defects in the pixels of image sensors
Context and goal
Space and nuclear environments contain particles which can displace atoms and create stable defects in the pixels of image sensors
Some of these defects are located in the depleted volume of the pixels and can generate dark current
RADECS, Moscow, September 17, 2015
Context and goal
Space and nuclear environments contain particles which can displace atoms and create stable defects in the pixels of image sensors
Some of these defects are located in the depleted volume of the pixels and can generate dark current
The highest (most problematic) dark current increases, produced by
nuclear interactions, could be the superimposition of dark current contributions from many single defects
Context and goal
Space and nuclear environments contain particles which can displace atoms and create stable defects in the pixels of image sensors
Some of these defects are located in the depleted volume of the pixels and can generate dark current
The highest (most problematic) dark current increases, produced by
nuclear interactions, could be the superimposition of dark current contributions from many single defects
A better knowledge of the nature of the single defects is needed to develop more accurate radiation-induced dark current increase models
RADECS, Moscow, September 17, 2015
Context and goal
Space and nuclear environments contain particles which can displace atoms and create stable defects in the pixels of image sensors
Some of these defects are located in the depleted volume of the pixels and can generate dark current
The highest (most problematic) dark current increases, produced by
nuclear interactions, could be the superimposition of dark current contributions from many single defects
A better knowledge of the nature of the single defects is needed to develop more accurate radiation-induced dark current increase models
Dark Current Spectroscopy (DCS)
First proposed by R.D. McGrath in 1987 to study dark current in CCDs
RADECS, Moscow, September 17, 2015
Dark Current Spectroscopy (DCS)
First proposed by R.D. McGrath in 1987 to study dark current in CCDs Principle : a stable defect with energy level 𝐸𝑡 and an emission cross
section σ has:
only one possible dark current generation rate U:
U (e-/s)
∼
σvthni 2cosh |Et−Ei|kT only one possible temperature dependence of this rate:
Dark Current Spectroscopy (DCS)
First proposed by R.D. McGrath in 1987 to study dark current in CCDs Principle : a stable defect with energy level 𝐸𝑡 and an emission cross
section σ has:
only one possible dark current generation rate U:
U (e-/s)
∼
σvthni 2cosh |Et−Ei|kT only one possible temperature dependence of this rate:
Activation energy (eV) ∼ 0.63 + |Et − Ei|
If each pixel contains only one defect:
The dark current of individual defects can be measured
A similar dark current in different pixels is most likely due to an
identical defect
The most common defects are detected as peaks in the dark current
distribution
RADECS, Moscow, September 17, 2015
Which particle to test the DCS?
Requirement : maximum of one defect per pixelWhich particle to test the DCS?
Requirement : maximum of one defect per pixelThe particle must create only one defect
No nuclear reactions (high damage): no neutrons or high-energy ions
RADECS, Moscow, September 17, 2015
Which particle to test the DCS?
Requirement : maximum of one defect per pixelThe particle must create only one defect
No nuclear reactions (high damage): no neutrons or high-energy ions
Only Coulombic interactions: low energy ions (< 10 MeV)
But the Coulombic interaction probability increases with the atomic number
Which particle to test the DCS?
Requirement : maximum of one defect per pixelThe particle must create only one defect
No nuclear reactions (high damage): no neutrons or high-energy ions
Only Coulombic interactions: low energy ions (< 10 MeV)
But the Coulombic interaction probability increases with the atomic number
⇒ Choose a light ion (protons or alphas)
The TID induced dark current needs to be minimized:
Better DDD to TID ratio (NIEL to LET ratio) for the alpha particle, especially at low energies
RADECS, Moscow, September 17, 2015
Which particle to test the DCS?
Requirement : maximum of one defect per pixelThe particle must create only one defect
No nuclear reactions (high damage): no neutrons or high-energy ions
Only Coulombic interactions: low energy ions (< 10 MeV)
But the Coulombic interaction probability increases with the atomic number
Irradiated 4T-PPD CIS details
Custom CIS fabricated in a commercially available 0.18 µm technology
⇒ Fully parametrable sensor operation and image acquisition settings
4T-PPD CIS with a small pixel pitch:
⇒ Very low pre-irradiation dark current (∼ 3 e-/s at 22°C)
4
Technology
0.18 µm
Array size
256 x 256 pixels
Pitch
4.5 µm (4T-pixel)
Photodiode
Pinned photodiode
RADECS, Moscow, September 17, 2015 Jean-Marc Belloir
4 MeV alpha irradiation
Dark current (D.C.) measured at T=22°C
4 MeV alpha irradiation
RADECS, Moscow, September 17, 2015
Jean-Marc Belloir 5
Dark current (D.C.) measured at T=22°C Two DCS peaks:
P1 : +50 e-/s (delta
D.C. from the cold pixels)
P2 : +100 e-/s ⇒ P2 = 2P1?
Same peaks detected with end-of-range alphas P1 P2 Fluence (a/cm²) Cold pixels DCS peaks
P1 pixel (+ 50 e-/s) P2 pixel (+ 100 e-/s)
Very hot pixel (nuclear event) P2 pixel (+ 100 e-/s)
Dark signal during irradiation
(4 MeV alphas)
P1: dark signal step of 50 e-/s ⇒ single defect
RADECS, Moscow, September 17, 2015
Jean-Marc Belloir 6
P1 pixel (+ 50 e-/s) P2 pixel (+ 100 e-/s)
Very hot pixel (nuclear event) P2 pixel (+ 100 e-/s)
Dark signal during irradiation
(4 MeV alphas)
P1: dark signal step of 50 e-/s ⇒ single defect
P2:
Two 50 e-/s steps: P2 =
2P1
One 100 e-/s step: two P1 defects
forming at the same time?
P1 pixel (+ 50 e-/s) P2 pixel (+ 100 e-/s)
Very hot pixel (nuclear event) P2 pixel (+ 100 e-/s)
Dark signal during irradiation
(4 MeV alphas)
P1: dark signal step of 50 e-/s ⇒ single defect
P2:
Two 50 e-/s steps: P2 =
2P1
One 100 e-/s step: two P1 defects
forming at the same time?
another defect?
Nuclear event annealing: many 50 e-/s, 100 e-/s and 200 e-/s steps
Low fluence:
Many cold pixels at 3 e-/s (no defects)
Activation energy after irradiation
(end-of-range alphas)
Jean-Marc Belloir 7
Fluence = 2.108 a/cm²
RADECS, Moscow, September 17, 2015 Cold pixels
Low fluence:
Many cold pixels at 3 e-/s (no defects)
Two pixel clusters at +50 e-/s (P1): 0.70 and 0.75 eV ⇒ corresponds to two
different defects
One cluster at +100 e-/s (P2): ∽0.70 eV ⇒ same activation energy than P1
⇒ same defect than P1
Activation energy after irradiation
(end-of-range alphas)
Fluence = 2.108 a/cm² P1 P2 Two different defects Cold pixelsLow fluence:
Many cold pixels at 3 e-/s (no defects)
Two pixel clusters at +50 e-/s (P1): 0.70 and 0.75 eV ⇒ corresponds to two
different defects
One cluster at +100 e-/s (P2): ∽0.70 eV ⇒ same activation energy than P1
⇒ same defect than P1
High fluence:
No more cold pixels at 3 e-/s (they all contain defects)
Successive clusters every +50 e-/s
All at 0.70 eV: superposition of the same defect
Activation energy after irradiation
(end-of-range alphas)
Jean-Marc Belloir 7 Fluence = 2.108 a/cm² Fluence = 1.109 a/cm² P1 P2 P1 P2 Two different defects Cold pixelsIsochronal annealing (4 MeV alphas)
P1 and P2 ↗ after 200°C but ↘ after 260°C
same annealing behavior
⇒ same defect (P2 = 2P1)
the defect starts to anneal
at 260°C ⇒ divacancy V2 ? P1 P2 P3
Isochronal annealing (4 MeV alphas)
RADECS, Moscow, September 17, 2015
Jean-Marc Belloir 8
P1 and P2 ↗ after 200°C but ↘ after 260°C
same annealing behavior
⇒ same defect (P2 = 2P1)
the defect starts to anneal
at 260°C
⇒ divacancy V2 ?
A new peak P3 appears at
700 e-/s (defect stable at 260°C) P1 P2 P3
Activation energy after 260°C annealing
(4 MeV alphas)
Cold pixels P2 (+100 e-/s) P1 (+50 e-/s)Activation energy after 260°C annealing
(4 MeV alphas)
Cold pixels P2 (+100 e-/s) P1 (+50 e-/s)New defect P4 (+13 e-/s)
P1: the defect at 0.70 eV has annealed (only the defect at 0.75 eV remains) ⇒ Could correspond to the vacancy-phosphorus (only stable up to 150°C)
New defect P4: 0.80 eV, +13 e-/s
9 RADECS, Moscow, September 17, 2015
Activation energy after 260°C annealing
(4 MeV alphas)
Cold pixels P2 (+100 e-/s) P1 (+50 e-/s)New defect P4 (+13 e-/s)
P1
P2 New defect P3 (+700 e-/s)
10 Cold pixels P1 P4 2P4 P1+P4 P5 P4 and 2P4 peaks visible P1+P4 visible New defect P5: 0.70 eV, 70 e-/s
Activation energy after 280°C annealing
(end-of-range alphas)
RADECS, Moscow, September 17, 2015 Jean-Marc Belloir
Identification of the defects
Major deep-level radiation-induced point defects in silicon in the litterature:
Defect Dark current at T=22°C (e-/s) from EDistance
C (eV) Annealing temperature Divacancy V2 40 0.39 – 0.44 260 Vacancy-phosphorus VP 40 0.44 – 0.47 150 Divacancy-oxygen V2O ? ∼0.50 >330 Divacancy-oxygen V2O ? ∼0.43 >330 Divacancy-hydrogen V2H ? ∼0.44 ?
Identification of the defects
Peak Dark current at T=22°C (e-/s) Activation energy (eV) Distance from EC (eV) Annealing temperature (°C) Identity P1 (and P2) 50 0.70 0.46 < 260 VP P1 (and P2) 50 0.75 0.41 260-280 V2 P3 700 0.60 0.56 > 260 V2O? P4 13 0.80 0.36 > 280 V2O,V2H? P5 70 0.70 0.46 > 280 V2O,V2H?
Major deep-level radiation-induced point defects in silicon in the litterature:
Identification of our defects:
11 Defect Dark current at T=22°C (e-/s) from EDistance
C (eV) Annealing temperature Divacancy V2 40 0.39 – 0.44 260 Vacancy-phosphorus VP 40 0.44 – 0.47 150 Divacancy-oxygen V2O ? ∼0.50 >330 Divacancy-oxygen V2O ? ∼0.43 >330 Divacancy-hydrogen V2H ? ∼0.44 ? Jean-Marc Belloir
Summary
Two
custom CIS
have been
irradiated
and
annealed
with
4 MeV
and
end-of-range
alphas
Five
different radiation-induced defects
have been
detected by Dark Current Spectroscopy
Two of these defects have been identified from their
dark current
,
activation energy
and
annealing behavior
:
The
divacancy
V2
Summary
Two
custom CIS
have been
irradiated
and
annealed
with
4 MeV
and
end-of-range
alphas
Five
different radiation-induced defects
have been
detected by Dark Current Spectroscopy
Two of these defects have been identified from their
dark current
,
activation energy
and
annealing behavior
:
The
divacancy
V2
The
vacancy-phosphorus
VP
Jean-Marc Belloir RADECS, Moscow, September 17, 2015 12