Transcript SEE, TID and DD results on Onsemi HAS2 APS

SEE, TID and DD results on
ON Semiconductor HAS2 APS
ESA CNES Final Presentation Days
5th - 6th June 2013, ESA-ESTEC The Netherlands
D. Hervé, M. Beaumel (Sodern)
D. Van Aken (ON Semiconductor)
M. Sauvagnac, P. Pourrouquet (TRAD)
M. Poizat (ESA)
SEE, TID and DD results on ON Semiconductor HAS2 APS
PRESENTATION OVERVIEW

Introduction

Summary of available data on HAS2

HAS2 SEE test results
–
–
–
–

heavy ion
proton
electron
detection of secondary photon
HAS2 TID and DD test results
– TID induced by gamma and electron
– DD induced by proton, electron and neutrons

5th-6th June 2013
Conclusions
ESA CNES Presentation Days, ESA ESTEC
P. 2
SEE, TID and DD results on ON Semiconductor HAS2 APS
INTRODUCTION
– OBJECTIVES:
• To improve the HAS2 radiation dataset for its use in high radiation environment
– ESA ACTIVITIES
• “Evaluation of STR performance in high radiation
environments” – 2010-2011
• “Radiation Characterization of Laplace RH
Optocouplers, Sensors and Detectors” – 2011-2012
– SUPPORT TO JUICE MISSION
• Formerly ESA’s contribution to LAPLACE mission with
Jupiter Ganymede Orbiter (JGO) reformulated to
JUICE ESA standalone mission in March 2011
(includes 2 fly-bys of Europa)
• Exposure to highly energetic trapped-Jovian electrons
5th-6th June 2013
ESA CNES Presentation Days, ESA ESTEC
P. 3
INTRODUCTION (cont’d)
– SUPPORT TO SOLAR ORBITER MISSION
• Distance from the sun as close as 0.25 AU
• Exposure to high peak flux of solar flare proton
– INPUTS FOR MISSIONS with HIGH PARTICLE FLUX & FLUENCE
• Investigate particle-induced SET by direct ionization in sensors pixel array
(considered to be one of the key issues of STR robustness)
• Collect radiation data at high TID levels, high DD levels on APS technology
• Compare TID and DD electron data wrt gamma irradiation TID data
• Investigate annealing behavior after radiation and its bias dependence
• Compare DD neutron results to DD proton results
• Apply those investigation on the HAS2 sensor
5th - 6th June 2013
ESA CNES Presentation Days, ESA ESTEC
P. 4
HAS2 CMOS image Sensor
– HAS2: High Accuracy star tracker CMOS
active pixel Sensor version 2
– Manufacturer: ON Semiconductor
– Technology: CMOS 0.35 µm Plessey (UK)
– Array of 1024x1024 pixels of 18x18 µm²
– Photodiode-type pixel with 3T readout
circuit → rolling shutter operation
– On-chip functions:
• Windowing addressing circuitry
• PGA, 12 bits ADC, T° sensor, analogue MUX for
external inputs digitization
– Readout modes
• Destructive (on chip FPN cancellation)
• Non Destructive (true CDS and off-chip
FPN cancellation) – NDR mode
– Hardened Design against SEE, TID & DD
effects, developed under ESA contract and
tailored for STR application
5th - 6th June 2013
HYDRA STR
ESA CNES Presentation Days, ESA ESTEC
P. 5
SEE, TID and DD results on ON Semiconductor HAS2 APS
SUMMARY OF AVAILABLE DATA (prior to Laplace study)
5th-6th June 2013
ESA CNES Presentation Days, ESA ESTEC
P. 6
SEE, TID and DD results on ON Semiconductor HAS2 APS
SUMMARY OF AVAILABLE DATA (this study)
5th-6th June 2013
ESA CNES Presentation Days, ESA ESTEC
P. 7
SEE, TID and DD results on ON Semiconductor HAS2 APS
HEAVY ION SINGLE EVENT EFFECT TESTING
HEAVY IONS TESTING
 Heavy ion Facility HIF
UCL(B) cocktail n°2
 Beam angle: 0°, 45°
and 60° tilt
 Purpose: investigate
low LET range and
Non Destructive Readout
mode operation

83Kr25+
ref image partially
corrupted (2 upsets)
image
partially corrupted
(1 upset)
ref image partially
at 4095 LSB
1.0E-03
TEST PRINCIPLE
 Images are grabbed under beam to
check for abnormal operation of the
sensor (address or integration time
errors, offsets shifts, image/line
corruptions...) for various operating
conditions
 Image error count: cross section vs LET
5th-6th June 2013
83Kr25+Video
83Kr25+
Address registers upset cross section [cm²]

1.0E-04
Sample A DR mode
Sample A NDR mode
Sample B DR mode
NO UPSET
Sample B NDR mode
1.0E-05
Sample C DR mode
Sample C NDR mode
Worst-case DR mode Weibull fit
Worst-case NDR mode Weibull fit
1.0E-06
0
10
20
30
40
50
60
70
Effective LET [MeV.cm²/mg]
ESA CNES Presentation Days, ESA ESTEC
P. 8
SEE, TID and DD results on ON Semiconductor HAS2 APS
DIRECT-IONIZATION SINGLE EVENT TRANSIENTS

PROTON TESTING
 Paul Scherrer Institute (PSI – CH)
 Energies from 100 MeV to 230 MeV
 Dosimetry accuracy of 5%
 Beam angle: 0° and 60° tilt

TEST PRINCIPLE
 Proton-induced SET signal above sensor noise
 Images grabbed under beam at short integration
time (2 ms) to collect isolated events at moderate
particle flux (1.0e+06 p+/cm²/s)
Corrected from dark signal
5th-6th June 2013
ESA CNES Presentation Days, ESA ESTEC
P. 9
DIRECT-IONIZATION SINGLE EVENT TRANSIENTS

PROTON TESTING
 Data of interest are the distributions of proton
induced signal size. As signal is spread over
several pixels (crosstalk effect), pixel clustering is
performed to obtain total signal of the spots
8.2 µm
 Calculation with MULASSIS: charge collection
layer thickness is tuned to fit the experimental
histograms with the calculated charge deposition
distribution (good agreement with the whole
histograms: distribution width and skewed shape)
100 MeV protons
DR mode 2 ms
 NDR mode: No upset detected in HAS internal
registers and multiplexer up to 230 MeV
8.2 µm
230 MeV protons
DR mode 2 ms
NDR Reference image
5th - 6th June 2013
NDR Final image
200 ms integration time
ESA CNES Presentation Days, ESA ESTEC
P. 10
SEE, TID and DD results on ON Semiconductor HAS2 APS
DIRECT-IONIZATION SINGLE EVENT TRANSIENTS


ELECTRON TESTING
 CURIE Institute CLINAC 2300 C/D
 Dose rate 1 to 10 Gy/min
 Beam energy 12 and 22 MeV
 Beam rotation: 0° and 28.2° tilt
 Pulsed flux: 16 -160 Hz
 Dosimetry accuracy ±10%
 Image grabber operating under beam flux
 Shielded with 10 cm of lead
TEST PRINCIPLE
 Rolling shutter operation combined w/
beam pulses  horizontal stripes, size
tuned by integration time
 Electron single SET signal (~ 350 e-)
close to detector noise (~100 e-)
 Facility adapted to cumulate 8-9 electrons
hit per pixel  electron-induced signal is
reliably detected
5th-6th June 2013
Pulsed Beam
Full frame image
1 Gy/min 12 MeV
ESA CNES Presentation Days, ESA ESTEC
P. 11
DIRECT-IONIZATION SINGLE EVENT TRANSIENTS
 ELECTRON TESTING
 Charge collection layer thickness obtained by fitting the
theoretical electron charge deposition distribution with
the experimental pulses using different models
 Poisson statistics + average LET: not accurate enough to
model the signal of accumulated hits (skewed part)
 Refined models: Straggling (particle to particle energy
deposition variation) and Monte-Carlo models in good
agreement with a large part of the histograms
Poisson distribution
7 µm
5th - 6th June 2013
Straggling distribution
10.2 µm
Monte-Carlo distribution
9.5 µm
ESA CNES Presentation Days, ESA ESTEC
P. 12
ELECTRON INDUCED SINGLE EVENT TRANSIENTS

SECONDARY PHOTON DETECTION
– Čerenkov effect: shockwave of photons emitted when
electrons exceed the speed of light in the medium, with
increased production of photon at short wavelength
(appears as a blue glow)
– Luminescence effect: excitation of atom electrons to higher energy levels (or orbits) by
ionizing radiation, atom relaxes with the emission of photons in the visible wavelengths
– Issue for JUICE: generation of parasitic light in instruments optics and lenses
– Tests performed with a specific half-masked window and 9.5 mm of silica
– Signal = clear zone – masked zone
– Measured signal: ~ 450 LSB (factor 2 below expected signal  discrepancy likely due to
the rough photometric rules applied to estimate the photon flux reaching the detector)
5th - 6th June 2013
ESA CNES Presentation Days, ESA ESTEC
P. 13
SEE, TID and DD results on ON Semiconductor HAS2 APS
TOTAL DOSE & DISPLACEMENT DAMAGE RESULTS

MAIN OBSERVATIONS AT HIGH IRRADIATION LEVEL
 All parts operational after irradiation (144,5 krad(Si) TID and 2.4109 MeV/g DD)
 Standby current increase during Cobalt-60 irradiation (x3 operational current)
 Loss of sensitivity during neutron / proton irradiations (33% at max level)
 Dark current increase
exhibits a non linear
dependence for parts
biased ON: rate of increase
starts to saturate around
~100 krad(Si)
 Parts biased OFF behave
almost linearly
5th-6th June 2013
ESA CNES Presentation Days, ESA ESTEC
P. 14
COMPARISON BETWEEN ELECTRON AND GAMMA
IRRADIATION TEST RESULTS
DARK CURRENT in e-/s/krad[Si]
Mean DCNU
Co-60 @ +25°C (this work)
192
15.8
Co-60 @ +25°C (previous study)
166
20.2
Co-60 @ −20°C (previous study)
4.4
1.5
10 MeV e- @ +19°C (this work)
121
21.3
10 MeV e- @ −20°C (previous
study)
3.0
1.5
Dark signal after 34 krad(Si)
• 9.71011 e-(11 MeV)/cm² (red histogram)
• Co60 source (green curve)
5th - 6th June 2013
ESA CNES Presentation Days, ESA ESTEC
P. 15
DARK CURRENT ANNEALING BEHAVIOR

BIAS DEPENDENCE OF ANNEALING EFFECTS




Specific annealing sequence has been applied on 3 parts after 34 krad at room temperature
ON bias: 5 adjacent windows were dynamically addressed in hard to soft reset NDR mode
OFF bias: 0 volt was applied on all pins
Reverse annealing behavior was observed on part biased OFF, but could recover quickly
when part was turned ON (dark current defect can be electrically activated or deactivated)
 Annealing enhanced on addressed lines (reset of photodiodes) wrt unaddressed (no reset)
 Those trends were not observed on DCNU (very stable at room temperature)
Annealing: applied in
NDR mode on 5 windows
Test : Full Frame image
acquired in DR mode
5th - 6th June 2013
ESA CNES Presentation Days, ESA ESTEC
P. 16
COMPARISON BETWEEN PROTON, ELECTRON AND
NEUTRON IRRADIATION TEST RESULTS

DARK CURRENT SPIKES
– Proton / neutron irradiation  DC
histogram exhibits a tail made of a
small number of pixels with high dark
signal (spikes / hot pixels)
– DC spikes intensity is higher for proton
compared to neutrons, while less DC
spikes are observed
3.04E+10 protons (100 MeV)/cm²
32.5°C 200 ms integration time
7.89E+07 MeV/g
Slope: 1450 LSB/s/dec
Density: 3.7%
– Total signal of DC spikes can be
integrated:


0
0
DC spike   x  ydiff ( x)  dx   x 
kc  ln( 10)
 dx
kxln(10)
e
Fit
Extrapolation
RADECS 2012:
M. Beaumel et al. “New Radiation Test
Results on HAS2 CMOS Image Sensor”
– And plotted against NIEL for each
irradiation species 
5th - 6th June 2013
ESA CNES Presentation Days, ESA ESTEC
P. 17
PROTON IRRADIATION TEST RESULTS
DD levels obtained from NIEL application to proton fluence
 Negligible spread between devices (same for neutrons)
 Good correlation between the various energies
 No effect of bias on spike generation (2 parts biased OFF
and 2 parts biased ON per energy)
5th - 6th June 2013
ESA CNES Presentation Days, ESA ESTEC
P. 18
SEE, TID and DD results on ON Semiconductor HAS2 APS
CONCLUSIONS
 Large set of radiation test campaign performed on HAS2 ready to
be used for sensor degradation and transient effects modeling in
harsh radiation environment
 Experimental results confirm HAS2 radiation hardness properties
and robustness when submitted to: heavy ion, proton, electron,
gamma, neutrons
 HAS2 flown on SPOT6 and PROBA2 and many other to come
5th-6th June 2013
ESA CNES Presentation Days, ESA ESTEC
P. 19
SEE, TID and DD results on ON Semiconductor HAS2 APS
ACKNOWLEDGMENTS
 M. Beaumel, L. Majewski, E. Anciant (SODERN)
 M. Sauvagnac, P. Pourrouquet, Y. Padie (TRAD)
 D. Van Aken, T. Vermeiren, D. D'Onofrio (ON Semiconductor)
 S. Ph. Airey, M. Poizat, Ch. Erd, Ch. Poivey (ESA)
 N. Fournier-Bidoz (Paris Curie Institute),
B. Nickson (ESTEC),
B. Van Houdt (SCK CEN),
G. Berger (UCL),
W. Hajdas, I. Britvitch (PSI)
5th-6th June 2013
ESA CNES Presentation Days, ESA ESTEC
P. 20