The COROT CCD test bench.
1 : Observatoire de Meudon, LESIA.
JT Buey 1, P. Bernardi 1, V. Lapeyrere 1, D. Tiphène 1 , B LeRuyet 1, R. Schmidt 1, J. Parisot 1
Meudon Observatory is involved in the development and making use of a CCD test bench dedicated to high precision photometry.
We are now testing the flight model, at least 10 chips, to :
* Check the measurements done by E2V (former Marconi, former EEV…).
* Realize specific calibrations : response versus the temperature ( and others).
* Establish criteria to choose the CCDs (4 over 10) for the flight.
CNES is responsible for the contract for the CCD flight model. E2V is the manufacturer of the CCD.
The Meudon observatory is responsible for the CCD test bench, the flight camera and electronics.
Current CCD 4280 in test with ZIF connector and
flex-ribond cable.
2k*2k of sensible area // 2k*2k of memory area
Main CCD characteristics :
• Thinned and backside illuminated : 90% of quantum
efficiency at 650nm..
• AIMO mode : dark current equals 0.5e/s at -40°C.
• Frame-Transfer : 0.225s for image transfer.
• Two outputs register : 10ms for a digitalized pixel.
• 30µm pic-to-pic flatness.
• Pixel size : 13.5*13.5µm. 2k*4k pixels.
• 3 sides buttable : 1.25mm between the sensitive area.
Main goals of the test bench :
•Technological test on the chip.
• Measurement of characteristics in function of
irradiation.
• Readout electronic and timing optimization.
• CCD Electrical Model characterization.
• CCD Flight Model characterization.
• Validation of high precision photometry.
The CCD test bench
*Temperature range : -55 to +40°C.
*Stabilization : < 0.05°C/hour.
*5 temperature probes : CCD and electronics.
* 3 axes motorized with 1 µm of precision.
The whole test bench : optics,
electronics, cryogenics... and the CCD!
The CCD 4280 inside the cryostat.
Temperature probe
CCD is read in windows (10) or all pixels.
Digitized pixel : 10µs.
Non digitized pixel : 1µs
Transfer line : 100µS
Read noise : < 10e-
Optical bandwith : 450-950nm
Spectral resolution : 10 or 1nm.
Optical source : pinhole or flat illumination.
We can simulate : different illuminations, temperature drift, jitter, defocusing...
The PRNU
The pixel response is a function of the wavelength, we show 3
typical images corresponding to 3 physical characteristics of the
CCD .
Quantum efficiency
In our application the Local PRNU on small surface is the most important
characteristic. It is calculated on 4096 windows of 32*32 pixels as the
standard-deviation (1s) of the pixels response of the different windows.
The PRNU is strongly dependant of the wavelength.
The quantum efficiency of the 10
flight models has been measured at
E2V, we plot here the mean value.
2,50
2,30
90
80
70
60
50
40
2,10
Local PRNU
100
Qe (%)
Local PRNU vs wavelength
Quantum efficiency
1,90
30
1,70
20
1,50
10
1,30
0
200
1,10
In the red (950nm)
---> Fringing
In the blue (420nm)
---> state surface and
AR coating.
CCD gain (µV/e-)
500
550
600
Voie droite
650
700
750
800
850
LinЋaire (Voie droite)
700
4500
3500
2500
500
The temperature dependence
of the quantum efficiency shows
great variation in the blue and red
part of the spectrum.
600
700
800
900
With typical target the variation of
the global response will be around
few 10-3 for 1°C variation.
10000
9000
8000
7000
6000
5000
4000
400
500
600
700
Wave le ngth (nm )
1000
y = -0,0046x + 4,0982
Parameters to choose the CCD
4,00
3,90
-52
-42
-32
-22
-12
-2
8
18
Comparison of 2 CCDs (flight models).
28
T (ЎC)
The Full Well Capacity is measured on the 10 CCDs
In different configurations :
•With flat illumination (comparison withE2V results).
•With spot illumination, at different positions and with
2 different PSF (exoplanet and asteroseismology)
Dark Current and irradiation
Aton
From the images acquired on the bench we use modelisation to determine the useful arae on the
CCD. We show here the process to calculate the jitter noise.
Flat fields
Convolution with the PSF
Dark
350-550 nm
sum
Calculation of the jitter noise
550-750 nm
Jitter
750-950 nm
At beginning of life the dark current is less than
0.5e-/s at -40°C.
After irradiation :
• Mean dark current is from 3 to 10e-/s at -40°C.
• Defects appear : 1 pixel over 10.000 (>100e-/s).
The effects of irradiation do not impact significantly
on the other characteristics of the CCD.
1000
11000
4,20
y = -0,0048x + 4,1152
900
Temperature Coefficient of the Quantum Efficiency
Wave le ngth (nm )
4,30
4,10
800
900
We also measured the variation of the
video signal with the different
polarization
The sensitivity is about 1e/1mv of bias
voltage.
500
400
LinЋaire (Voie gauche)
600
Wav elength (nm)
1500
Voie gauche
4,40
450
5500
FG/FG (ppm/Ў)
CCD Gain vs temperature
4,50
0,50
400
Temperature coefficient of the CCD Response
Global response
with temperature
500
0,70
CCD Response
Gain versus the temperature
Full Well Capacity
400
Wave le ngth (nm )
Qe/Qe (ppm/Ў)
In the green (700nm)
---> « center » of silicon
300
0,90
We obtain an image, each pixel
indicates the jitter noise at that position
We can choose the PSF for
seismo or exo field
Parameters priority
• For exoplanets :
• For seismology :
– Dark noise (will change!)
– Pixel capacity
– Jitter noise map
– Jitter noise map
– Quantum efficiency
– Quantum efficiency
– Pixel capacity
– Temperature coefficient
– Temperature coefficient
A way to show the different
parameters, the surface under the
curve increases with the quality of
the CCD.
All parameters
Best CCD!!
Benu
800
900
1000