Astronomical CCD's
There are a variety of CCD architectures that have been developed
over time with varying strengths
Full Frame or Slow Scan CCD's
Full Frame CCD
Most common type used in Astronomy
–
Good low noise readout
–
Cooled for low dark current
–
Must be mechanically shuttered to avoid streaking during
readout
–
QE is Compromised by gate structure
Incoming photons
●
p-type silicon
n-type silicon
625m
Silicon dioxide insulating layer
Polysilicon electrodes
Interline Transfer CCD
Very fast electronic shuttering
●
One clock pulse moves electrons to non-sensitive transfer
register
●
Light loss due to silicon used for transfer
●
Not commonly used in astronomy. Best in video applications
Interline Transfer CCD
QE Losses can be helped with addition of micro-lenses
Frame Transfer CCD
●
Fast electronic shuttering by use of storage array
–
Shuttering not as fast as interline, but close to full-frame
CCDs in performance
Frame Transfer CCD
●
Loss of focal plane area is disadvantage
●
Best for fast framing where mechanical shutters can't work
●
●
Eliminates most of photon integration dead-time. Readout can
be in parallel to integration
–
Great if needing to maximize S/N over a fixed period of tim
–
Full frame CCD's can lose > 1/2 time on sky due to readou
Generally smaller than full frame CCD's so not good for large
area imaging
CCD Pixels and Thinning
●
●
●
Pixels are formed by the potentials of the gates and P-N
structure of the silicon
Photoelectrons need to be generated in or near the depletion
region of the silicon to prevent recombination
Thick CCDs have problems with very red or very blue light from
absorption and Fresnel losses
Electric potential
The electric field structure in a CCD defines to a large degree its Quantum Efficiency (QE). Consider
first a thick frontside illuminated CCD, which has a poor QE.
Cross section through a thick frontside illuminated CCD
In this region the electric potential gradient
is fairly low i.e. the electric field is low.
Potential along this line
shown in graph above.
Any photo-electrons created in the region of low electric field stand a much higher chance of
recombination and loss. There is only a weak external field to sweep apart the photo-electron
and the hole it leaves behind.
CCD Pixels and Thinning
●
●
●
Pixels are formed by the potentials of the gates and P-N
structure of the silicon
Photoelectrons need to be generated in or near the depletion
region of the silicon to prevent recombination
Thick CCDs have problems with very red or very blue light
from absorption and Fresnel losses
–
–
●
2
R=∣( n1−n 2)/( n1 + n2 )∣
Glass n = 1.5, R = 4%
Silicon n=3.6, R=32%
What if we remove the bulk Silicon and bring the photons in
the other direction?
Electric potential
In a thinned CCD , the field free region is simply etched away.
Cross section through a thinned CCD
There is now a high electric field throughout the
full depth of the CCD.
This volume is
etched away
during manufacture
Problem : Thinned CCDs may have good blue
response but they become transparent
at longer wavelengths; the red response
suffers.
Red photons can now pass
right through the CCD.
Photo-electrons created anywhere throughout the depth of the device will now be detected. Thinning
is normally essential with backside illuminated CCDs if good blue response is required. Most blue
photo-electrons are created within a few nanometers of the surface and if this region is field free,
there will be no blue response.
Deep Depletion CCDs
Electric potential
Ideally we require all the benefits of a thinned CCD plus an improved red response. The solution is to use a
CCD with an intermediate thickness of about 40m constructed from Hi-Resistivity silicon. The increased
thickness makes the device opaque to red photons. The use of Hi-Resistivity silicon means that there are no field
free regions despite the greater thickness.
Cross section through a Deep Depletion CCD
Problem :
Hi resistivity silicon contains much lower
impurity levels than normal. Very few wafer
fabrication factories commonly use this
material and deep depletion CCDs have to
be designed and made to order.
Red photons are now absorbed in
the thicker bulk of the device.
There is now a high electric field throughout the full depth of the CCD. CCDs manufactured in this way
are known as Deep depletion CCDs. The name implies that the region of high electric field, also known as
the ‘depletion zone’ extends deeply into the device.
Another problem commonly encountered with thinned CCDs is ‘fringing’. This is greatly reduced
in deep depletion CCDs. Fringing is caused by multiple reflections inside the CCD. At longer
wavelengths, where thinned chips start to become transparent, light can penetrate through and be
reflected from the rear surface. It then interferes with light entering for the first time. This can give
rise to constructive and destructive interference and a series of fringes where there are minor
differences in the chip thickness.
The image below shows some fringes from an EEV42-80 thinned CCD
For spectroscopic applications, fringing can render some thinned CCDs unusable, even those
that have quite respectable QEs in the red. Thicker deep depletion CCDs , which have a much
lower degree of internal reflection and much lower fringing are preferred by astronomers
for spectroscopy.
CCD Pixels and Thinning
●
●
Bare silicon is now available for A-R coating further enhancin
QE performance, with values often above 90%
Coatings can be wavelength tuned for a particular application
CCD Readout
At the end of the Horizontal register is a readout node
Capacitance of node converts charge to voltage
20m
Output Drain (OD)
Gate of Output Transistor
Output Source (OS)
SW
R
RD
OD
Output Node
Reset
Transistor
Reset Drain (RD)
Summing
Well
R
Output
Node
Serial Register Electrodes
OS
Summing Well (SW)
Last few electrodes in Serial Register
Output
Transistor
Substrate
On-Chip Amplifier 1.
The on-chip amplifier measures each charge packet as it pops out the end of the serial register.
+5V
RD and OD are held at
constant voltages
SW
0V
-5V
SW
R
RD
OD
+10V
R
0V
Reset
Transistor
Summing
Well
--end of serial register
Output
Node
Vout
Output
Transistor
(The graphs above show the signal waveforms)
OS
Vout
The measurement process begins with a reset
of the ‘reset node’. This removes the charge
remaining from the previous pixel. The reset
node is in fact a tiny capacitance (< 0.1pF)
On-Chip Amplifier 2.
The charge is then transferred onto the Summing Well.
Vout is now at the ‘Reference level’
+5V
SW
0V
-5V
SW
R
RD
OD
+10V
R
0V
Reset
Transistor
Summing
Well
--end of serial register
Output
Node
Vout
Output
Transistor
OS
Vout
There is now a wait of up to a few tens of
microseconds while external circuitry measures
this ‘reference’ level.
On-Chip Amplifier 3.
The charge is then transferred onto the output node.
Vout now steps down to the ‘Signal level’
+5V
SW
0V
-5V
SW
R
RD
OD
+10V
R
0V
Reset
Transistor
Summing
Well
--end of serial register
Output
Node
Vout
Output
Transistor
This action is known as the ‘charge dump’
OS
Vout
The voltage step in
Vout is as much as
several V for each electron contained
in the charge packet.
On-Chip Amplifier 4.
Vout is now sampled by external circuitry for up to a few tens of microseconds.
+5V
SW
0V
-5V
SW
R
RD
OD
+10V
R
0V
Reset
Transistor
Summing
Well
--end of serial register
Output
Node
Vout
Output
Transistor
OS
Vout
The sample level - reference level will be
proportional to the size of the input charge
packet.
CCD Readout
In the real world there is noise!!!
●
Reset noise - CDS
●
FET Noise
●
Switching transients
CCD Readout
Dual-Slope Integrator
●
Reset CCD amp, connect to output capacitor and integrate up
●
Dump the charge, invert the signal and integrate down
Reset switch
Pre-Amplifier
.
Inverting Amplifier
-1
OS
OS
CCD
R
RC=
C
RL
Input Switch
Integrator
Polarity Switch

ADC
(1 sample
Per pixel)
Computer Bus
R RD OD
CCD Readout
Dual-Slope Integrator
●
Wait for signal to settle on both up and down
●
Integrator averages signal, reducing noise in the measurement
●
Averaging is proportion to time integrating
–
Many cameras have selectable pixel rates from 100K to 110M pixels per second
–
Low signal levels, slow pixel rates and better averaging
–
High signal levels, photon noise dominates and readouts
can be sped up without impacting data quality
CCD Readout Defects
●
Blooming – Wells fill up and spill over into other pixels. 1-d
effect due to channel stops between columns
Spillage
Spillage
CCD Readout Defects
●
Charge traps, hot pixel clusters, Cosmic rays
Bright
Column
Cluster of
Hot Spots
Cosmic rays
CCD Mosaics
Telescope time is often very limited, but large scale images are
sometimes needed
–
Mosaic images can be made with a single CCD by jogging
the telescope around the sky and combining the images in
post-processing
–
If the scale of the field needed is 10 times the size of the
CCD, 100x the integration time is required to build a mosa
–
REALLY big CCD's are not possible due to restrictions of
wafer processing techniques
CCD Mosaics
●
●
3 and 4 side butt-able CCDs are now available in large forma
Allows for packing of CCDs in the focal plane of a telescope
with gaps as small as 50 microns
ESO OmegaCam
For VLT Survey Telescope.
Camera ready for use.
256 Megapixel
Large Synoptic Survey Telescope
4° FOV  74 cm 
WFS
3GigaPixel
Will survey whole available sky
every 3 nights.
XXX
XX
X
X
XX
XXX
XXX
XX
X
X
XX
XXX
8.4m primary
200 x 4k x 4k Detectors
3.5 degree field of view
20-30 TByte per night
2 second mosaic readout.
May require hybrid detector
technology
Large Mosaic Projects
1E+10
LSST
SNAP (space)
Pan-STARRS
Number of pixels
1E+09
CFHT & SAO Megacam
SLAC VXD3
GAIA (space)
SDSS
1E+08
ESO omegacam
UH4K
lots of 8K mosaics!
1E+07
NOAO4K
INTWFC
1E+06
1990
1992
1994
1996
1998
2000
2002
Year
2004
2006
2008
2010
2012
2014
Electron Multiplying CCDs
●
Desire for fast readout and low light levels?
–
●
Rapid time series (occultations, fast variables etc)
This implies read-noise limited images, and fast readouts are
even worse than the typical 6 e- of slow readout
–
Ideal detectors count single photon events. Traditional photoelectron amplification e.g. PMT's allows for discrimination of
single photons
–
CCD's have to collect 30-40 e- per pixel to avoid read noise
limits
–
Still want to use a CCD since single pixel detectors are so
limiting in both QE and lack of spatial coverage.
Electron Multiplying CCDs
●
●
CCD's are carefully tailored so clock voltages and clock shapes
efficiently move charge from pixel to pixel
–
In addition to efficiency (CTE), levels are also set to avoid a
phenomenon called impact ionization
–
Electrons moving in a crystal can scatter off the crystal
structure and release kinetic energy into the lattice, with a
probability that a carrier pair is generated.
–
Usually this is bad, and akin to dark (leakage) current.
EEV corporation has developed CCDs with a special
amplification horizontal register designed to allow amplifications
of factors of 10's to 1000's of e- per photo-electron
Electron Multiplying CCDs
Avalanche multiplication takes
place in Multiplication Register,
using an HV clock (40-45Volts).
1000e- signal out
Multiplication register
Standard MOSFET
amplifier
Commercially available :
CCD60 128 x 128
Multiplication register
1e- in
Normal Serial register
Image Area
Multiplication register
Store Area
CCD87/97 512x512
CCD201 1K x 1K
Flexible Operation
E2V CCD201
Electron Multiplying
Amplifier
Conventional Amp.
3e noise
clock left
To Observe:
Absorption lines
Bright Emission Lines
Emission lines superimposed on bright continuum
clock right
<<1e noise
To Observe:
Faint Emission Lines
EMCCD becomes competitive
at lower exposure levels
With EM Gain: At low illumination, photons are resolved.
Also visible is the Clock Induced Charge (CIC) which
is the dominant noise source (typical value 0.03e-)
Without EM Gain: object only just visible above noise.
EMCCD range of operation
So at higher expsoure levels the EMCCD actually performs worse than a normal CCD
8
7
6
SNR
5
4
3
Conventional CCD 5e- noise
2
L3EMCCD
CCD with no CIC
1
photon noise limit
0
0
10
20
30
40
50
Photons per pixel per frame
EMCCD wins due to zero read noise
EMCCD loses due to multiplication noise
60
Early EMCCD Demonstrations
1) Crab Nebula pulsar at 180fps
2) Photon interference in the lab
Cryogenic CCD87 imaging a faint pinhole
Threshold
Raw input frames
Particles?
Thresholded and accumulated
Waves?
CMOS Detectors
●
●
●
CMOS detectors are closer to IR arrays than CCD's but work
the same silicon wavelength range
Commonly in use now in everything from smart phones to high
end SLR cameras, but only beginning to be adapted to
Astronomy
Each pixel has a readout
CCD Versus CMOS
CCD
Pros
Cons
CMOS
Relaxed design rules (1-2um).
Simple design.
Large format.
Hi QE.
Low noise.
High dynamic range.
100% fill factor.
Proven, mature technology.
Random Access pixels: flexible readout.
Hi Speed combined with low noise (at least
when multiple samples per pixel are
combined)
On-chip guiding possible using guiding
sub-window within science field.
No shutter needed :
continuous observations.
Integration of controller onto sensor:
‘camera-on-a-chip’
Low power.
Radiation tolerant.
Relatively slow, generally only 2-4
amplifiers per chip.
Blooming.
Requires shutter
No exposure during readout
(unless half of silicon is sacrificed as a
frame store).
Bad pixel can wipe out a full column.
Sub-micron design rules.
5-10nm dielectrics : rupture at 5V.
Small amplifiers difficult to optimise for low
noise.
Poor linearity (can be overcome with
multiple sampling)
Low QE, <100% fill factor.
Pixel cross talk issues.
CMOS Sensitivity
Thin depletion region means low red sensitivity.
If depletion depth could be increased, CMOS will
start to look very attractive.
Sony 12.84 Mpixel
CMOS sensor
Microlenses solve fill factor problem:
photodiode
photodiode
transistors
CMOS pixel structure
Hybrid CMOS
Solution is hybridisation : Silicon multiplexer bump bonded to Silicon absorber
CCD-like QE + all the advantages of CMOS
However, still very expensive, CCDs still have the
upper hand.
Rockwell HySiVi
4k x 4k detector
For ‘blank-cheque’
applications.
May be used for LSST
where 15 second exposures
must be read out in 1-2 seconds