Back-thinned TDI-CCD / Technical information

Technical information
Back-thinned TDI-CCD
Back-thinned TDI (time delay integration)-CCDs allow
acquiring high S/N images even under low-light conditions
during high-speed imaging and the like. TDI operation
yields dramatically enhanced sensitivity by integrating the
exposure of a moving object. The back-thinned structure
ensures high quantum efficiency over a wide spectral
range from the ultraviolet to the near infrared region (200
to 1100 nm).
[Figure 1-1] TDI operation illustration
TDI-CCD
TDI-CCD
Camera
Lens
Signal
intensity
Direction of object movement
TDI operation
In CCD operation, a signal charge is transferred to the
output section while being held in potential wells so as
not to mix with other individual charges. TDI operation
makes good use of this CCD charge transfer principle,
and it is an effective technique for imaging a moving
object or a still object while scanning it with a CCD
sensor that is itself being moved.
Normally, an image focused on the CCD sensor is output
as a signal corresponding to the focused position. This
means that the image focused within the integration
time must stay in the same position on the CCD sensor.
If, for some reason, the focused position is shifted,
then the image S/N will deteriorate. When an object is
moving, the focused position will shift, causing the image
to blur or, in some cases, no image to appear.
The TDI operation, in contrast, is a unique operation that
captures images of a moving object. In FFT-CCD, signal
charges in each column are vertically transferred during
charge readout. TDI operation synchronizes this vertical
transfer timing with the movement of the object, so
signal charges are integrated by a number of times equal
to the number of vertical stages of the CCD pixels.
In TDI operation, the signal charges must be transferred
1
in the same direction at the same speed as those of the
object to be imaged. These speeds are expressed by
equation (1).
v = f × d ……… (1)
v : object speed, charge transfer speed
f : vertical CCD transfer frequency
d : pixel size (transfer direction)
In Figure 1-2, when the charge accumulated in the first
stage is transferred to the second stage, another charge
produced by photoelectric conversion is simultaneously
accumulated in the second stage. Repeating this
operation continuously until reaching the last stage
M (number of vertical stages) results in a charge
accumulation M times greater than the initial charge.
This shows that the TDI operation enhances sensitivity
up to M times higher than ordinary linear image sensors.
(If the number of vertical stages is 128, the sensitivity will
be 128 times higher than ordinary linear image sensors.)
Since the accumulated signal charges are output for each
column from the CCD horizontal shift register, a twodimensional continuous image can be obtained. TDI
operation also improves sensitivity variations compared
to two-dimensional operation mode.
[Figure 1-2] Schematic of integrated exposure in TDI
operation
Charge transfer,
object movement
What is a back-thinned TDI-CCD?
Time1
Time2
Time3
First stage
Last stage M
Charge
1
KMPDC0139EA
[Figure 1-3] Imaging examples in TDI operation
[Figure 1-5] Imaging in TDI operation (continuous
image during drum rotation)
(a) Imaging of fast moving object
TDI-CCD
(2048 × 128 pixels)
Direction of charge transfer
Direction of object
movement
KMPDC0266EB
(b) Imaging of fast rotating object
TDI-CCD
(2048 × 128 pixels)
2
Direction of charge transfer
Features
High sensitivity (UV to near IR)
Hamamatsu TDI-CCDs employ back-illuminated
structure and ensure high sensitivity in the UV to near
infrared region (200 to 1100 nm).
Direction of
object rotation
[Figure 2-1] Spectral response (without window)
(Typ. Ta=25 °C)
7000
KMPDC0267EA
In Figure 1-3 (b), when the CCD is put in two-dimensional
operation and the drum is imaged while in idle, a clear
image with no blurring is obtained as shown in Figure
1-4 (a). However, when the drum is rotating, the image
is blurred as shown in Figure 1-4 (b). Shortening the
shutter time captures an unblurred image, but the image
becomes dark as shown in Figure 1-4 (c). Using a TDICCD acquires clear, continuous images with no blurring
as shown in Figure 1-5 since charge transfer is performed
in the same direction at the same speed as those of the
rotating drum.
Photosensitivity [V/(μJ · cm2)]
Drum
6000
5000
S10200-02-01
S10201-04-01
S10202-08-01
S10202-16-01
4000
3000
2000
1000
0
200
300 400
500 600
700 800 900 1000 1100
Wavelength (nm)
KMPDB0268EB
[Figure 1-4] Imaging in two-dimensional operation
(a) When drum is in idle
(b) When drum is rotating
(c) When drum is rotating (with shutter time shortened)
2
[Figure 2-2] Quantum efficiency vs. wavelength
(without window)
(b) Bidirectional transfer
Object
Scan
(Typ. Ta=25 °C)
100
Back-thinned TDI-CCD
Movement
90
Quantum efficiency (%)
80
70
60
50
(c) Camera with bidirectional transfer TDI-CCD
40
30
TDI camera
20
10
Front-illuminated CCD
0
200
300
400
500
600
700
800
Object
900 1000 1100
Wavelength (nm)
KMPDB0269EC
KMPDC0503EA
High-speed line rate using multiple ports
To achieve continuous imaging of high-speed moving
samples, multiple amplifiers are arranged in the TDICCD, and images are read out in parallel. This results in
high-speed line rate. The pixel rate is 30 MHz/port, and
the line rate is 50 kHz on the S10200-02-01, S10201-0401, and S10202-08-01 and 100 kHz on the S10202-16-01.
[Figure 2-3] Sensor structure [typical example: S10201-04-01,
2048 (H) × 128 (V) pixels, 4 ports per side × 2
(bidirectional transfer)]
128 pixels
Port
Port
Port
512 pixels
Port
512 pixels
Port
Port
512 pixels
Port
512 pixels
Bidirectional
transfer
Port
Bidirectional vertical transfer is possible.
KMPDC0268EA
When an object is scanned multiple times, the bidirectional
transfer function of the TDI-CCD eliminates the need to
return the camera as shown in Figure 2-4 (a), and thus the
inspection throughput can be improved.
[Figure 2-4] Camera scan direction
Blooming (overflow) is a phenomenon that occurs when
high-intensity light enters the photosensitive area and
the resulting signal charge exceeds a specific level. This
excess charge then overflows into adjacent pixels and
transfer region. A technique to prevent this is called
anti-blooming which provides a drain to carry away the
excess charge [Figure 2-5].
Anti-blooming structures for CCDs are roughly divided
into a lateral type and a vertical type, and our CCDs
use the lateral type. The lateral type structure has an
overflow drain formed along the pixels or charge transfer
channels. This structure has the drawback that the fill
factor is reduced when used for front-illuminated CCDs.
However, this problem can be avoided when used for
back-thinned CCDs [Figure 2-6].
When controlling the anti-blooming function by means
of the overflow drain voltage (VOFD) and overflow gate
voltage (VOFG), these applied voltages may cause charge
to flow from the drain to the pixel or decrease the
saturation charge. The applied voltages must be set to
appropriate values [Figure 2-7, 2-8].
[Figure 2-5] Imaging examples
(a) Without anti-blooming (b) With anti-blooming
(a) Unidirectional transfer
Object
Scan
Movement
3
Anti-blooming
[Figure 2-8] Voltage setting and anti-blooming
(schematic)
High
[Figure 2-6] Anti-blooming structure (lateral type) and potential
(structure in which overflow drain is provided for two pixels)
A
Overflow gate voltage
Area in which charge flows
into pixels
A’
Area in which anti-blooming
takes effect
Area in which blooming
occurs
Low
[Top view of vertical
pixel area]
OFG
OFD
N
+
N
N
P-EPI
P+
Poly-Si
SiO2
N- channel
3
Barrier
(Clock voltage: low)
Storage
(Clock voltage: high)
[A-A’ cross section]
KMPDC0286EB
[Figure 2-7] Schematic diagram of anti-blooming
(lateral type)
VOFD
VOFG
Comparison between previous products and new products
[Figure 3-1] Characteristic comparison between
previous products and new products
Charge drift
Potential
VPXV
Previous product
New product
S10200-02
S10201-04
S10202-08
S10202-16
S10200-02-01
S10201-04-01
S10202-08-01
S10202-16-01
Typ.
30
30
Max.
35
40
CCD node efficiency
3.5
9.5
µV/e-
Readout noise
(30 MHz)
100
35
e- rms
Dynamic range
1000
2857
-
See datasheet.
See datasheet.
-
300
150
Ω
Parameter
Output
signal
frequency
Operating voltage
Output impedance
N+
High
KMPDC0496EA
N
P
Overflow drain voltage
Low
Channel stop
1 pixel
Unit
MHz
+
P
N-
[Figure 3-1] Output waveforms (fc=30 MHz)
P
(a) Previous product
(b) New product
Potential at which blooming occurs
1 pixel
Vertical low level
Appropriate potential
1 pixel
Potential level of the overflow drain
Potential at which charge flows into pixels
Potential
Vertical high level
KMPDC0285EA
10 ns/div.
10 ns/div.
KMPDB0405EA
Thanks to the increased CCD node efficiency and
optimized amplifier design, the new products produce
output waveforms that are closer to the ideal waveform,
with greater output amplitude and improved bandwidth
than the previous product [Figure 3-1].
4
4
increases, the effects of dark shot noise may increase,
in which case an appropriate heat dissipation measures
need to be taken [Figure 4-3].
If the dark offset of each column needs to be corrected,
use the output from the effective pixels that is generated
when there is no incident light (dark state). Note that the
blank pixel output does not include signals generated by
vertical pixels.
How to use
Reducing spurious signals
When the back-thinned CCD is viewed from the light
input side, the horizontal shift register is covered by the
thick area of the silicon (insensitive area) [Figure 4-1], but
long-wavelength light may pass through the insensitive
area. If this light is received by the horizontal shift register,
it can cause spurious signals.
Spurious signals are mixed into the actual signal. If the
horizontal transfer time period is longer than the total
of the integration times of TDI operation, the effect of
spurious signals increases.
If the effect of spurious signals is large, measures need to
be taken such as adjusting the light irradiation position or
shielding the horizontal shift register.
[Figure 4-2] Dark output vs. element temperature
(typical example)
Dark output (e-/pixel)
10000
(Line rate: 50 kHz, number of vertical stages: 128)
1000
100
10
Reducing effects of dark output
1
20
Dark output is an output current that flows when no
light is input. For CCDs for measurement applications,
the dark output is typically expressed as the number of
electrons generated per pixel per second (unit: electron/
pixel/s). In TDI operation, since the dark current that
is generated by the pixels of each column is integrated
over the number of vertical stages, the dark current is
expressed as the number of electrons generated per
column (unit: electron/pixel), and its magnitude varies
depending on the line rate, the number of stages, and the
like. As such, at high-speed line rates, the dark current is
extremely small.
Dark output nearly doubles for every 5 to 7 °C increase in
temperature [Figure 4-2]. When the element temperature
30
40
50
60
Element temperature (°C)
KMPDB0407EA
[Figure 4-1] Device structure (typical example: S10202-08-01, schematic of CCD chip as viewed from top of dimensional outline)
OSb3
OSb6
OSb7
OSb8
OSa3
OSa6
OSa7
OSa8
OSa1
Thinning
128 stages
OSb2
OSa2
OSb1
Thinning
8 blank pixels
512 pixels
V=128
H=512 × 8 (ports)
KMPDC0252EC
5
[Figure 4-3] Noise vs. element temperature
(typical example)
Clocks and output waveforms during high-speed operation
(Line rate: 50 kHz, number of vertical stages: 128)
100
For the clock waveforms of the horizontal shift register,
we recommend that ringing be reduced as much as
possible and that the waveforms cross at 50% ± 10% of
the clock amplitude [Figure 4-5].
If the drive conditions are not appropriate, saturation
charge, CCD transfer efficiency, readout noise, and the
like may not meet the characteristic values listed in the
datasheet. Furthermore, adjust the waveform applied to
the reset gate so that flat regions are created in the OS
output waveform’s DC level (reset level) and signal level
[Figure 4-6]. The driver circuit requires a mechanism for
fine-tuning these clock timings.
Total noise
Noise (e- rms)
Readout noise
10
Dark shot noise
1
20
30
40
50
60
Element temperature (°C)
KMPDB0406EA
[Figure 4-5] Timing chart
(horizontal shift register, reset gate)
Tpwh
Heat generation from sensor
Tprh
Tpfh
P1H
[Figure 4-4] Element temperature vs. operation time
(S10201-04-01, our evaluation circuit,
typical example)
P2H
Tprr
Tpwr
Tpfr
RG
KMPDC0497EA
[Figure 4-6] OS output waveform example
(fc=30 MHz)
Reset field through
DC level
(reset level)
Signal
The TDI-CCD performs high-speed readout. Because
of its multiport structure, the sensor may become
hot. Since the dark current increases as the element
temperature increases, appropriate heat dissipation
measures may be necessary depending on the situation.
For the heat dissipation methods, see “Image sensors”
under “Precautions.”
The power consumption during charge transfer is
proportional to the square of the operating voltage
amplitude and readout frequency. In this case, the power
consumption by the horizontal shift register whose
readout frequency is large is dominant. Therefore, in the
horizontal shift register on the side that is not reading
out, to reduce heat generation, the drive voltage is set to
DC voltage so that unneeded charge is discarded (see the
timing chart on the datasheet).
Figure 4-4 is an example showing the relationship
between the element temperature and operation time
when our evaluation circuit is used (the circuit system is
sealed and without any heat dissipation measures).
Signal level
The circuit must be optimized to obtain an ideal
waveform as shown above.
KMPDB0409EA
70
High-speed signal processing circuit
Element temperature (°C)
60
Without cooling fan
50
For a CCD signal processing circuit that requires highspeed readout at several megahertz or faster, it is difficult
for a circuit constructed only of discrete components to
achieve high-speed clamp operation and fast capacitor
charging/discharging characteristics.
A high-speed signal processing circuit can be constructed
by using an analog front-end IC (a single IC chip consisting
of CDS, gain, and offset circuits, A/D converter, etc.)
optimized for CCD signal processing [Figure 4-7].
40
With cooling fan
30
20
10
0
0
5
10
15
20
25
30
35
Operation time (min)
KMPDB0408EA
6
[Figure 4-7] High-speed signal processing circuit example (using analog front-end IC)
To ClockDriver
+DRV
0.1 μF
+
4.7 μF
0.1 μF
+DRV
+VOD or +VRD
10 μF
0.1 μF
4.7 μF
0.1 μF
24
23
22
21
20
19
18
17
16
15
14
13
0.1 μF
+1.8 VLDOOUT
OSxx
4.7 μF +
0.1 μF
2.2 kΩ
0.1 μF
0.1 μF
0.1 μF
0.1 μF
RG
IOVDD
LDOOUT
CLI
AVSS
AVDD
CCDINP
CCDINM
AVSS
AVDD
REFT
REFB
D11
D10
D9
D8
D7
DRVDD
DRVSS
D6
D5
D4
D3
D2
12
11
10
9
8
7
6
5
4
3
2
1
To FPGA
LDOEN
SL
SDI
SCK
GPO1
GPO2
VD
HD
DVSS
DVDD
D0(LSB)
D1
25
26
27
28
29
30
31
32
33
34
35
36
0.1 μF
0.1 μF
+3 V
37
38
39
40
41
42
43
44
45
46
47
48
+3 V
+1.8 VLDOOUT
+
RGVDD
HL
RGVSS
H4
H3
HVDD
HVSS
H2
H1
NC
D13(MSB)
D12
AD9979
+1.8 VLDOOUT
To FPGA
0.1 μF
KMPDC0500EA
Readout noise and output signal frequency
In general, lowering the output signal frequency reduces
the CCD readout noise [Figure 4-8]. Note that when
the output signal frequency is lowered, the line rate is
also lowered. This causes an increase in the dark output
component during charge transfer, and its shot noise may
affect the total noise.
The readout noise varies depending on various factors
including the readout circuit.
[Figure 4-8] Readout noise vs. output signal frequency
(typical example)
50
Readout noise (e- rms)
40
30
20
10
0
10
20
30
40
Output signal frequency (MHz)
KMPDB0410EA
7
Exposure adjustment
In TDI operation, the exposure can be varied by changing
the line rate. Also, adding a filter to the optical system to
adjust the light level is another effective method. Note
that our standard products do not have a function for
adjusting the exposure by switching the number of
vertical stages.
Two-dimensional operation
In addition to TDI operation, Hamamatsu TDI-CCDs can
perform two-dimensional operation. This is sometimes
used to verify an optical system or for initial evaluation. A
two-dimensional operation image when light is incident
is shown in Figure 4-9.
A fixed zigzag pattern may appear when the image
contrast is enhanced, and the output difference is in the
order of a few percent. This is because of the sensitivity
difference in each pixel, and the effect varies depending
on the wavelength of the incident light. In each column,
the composition of pixels of different sensitivities is the
same. Therefore, in TDI operation, the average sensitivity
of the pixels in each column is the same.
The clock timing chart for two-dimensional operation is
shown in Figure 4-10.
[Figure 4-9] Image when uniform light is incident
during two-dimensional operation
(1 port: 512 × 128 pixels)
5
Output circuit structure
FDA (floating diffusion amplifier) is the most popular
method for detecting the signal charge of a CCD. The
FDA consists of a node for detecting charges and a
MOSFET (MOS1) for reset and MOSFETs (MOS2 to 6)
for charge-to-voltage conversion connected to the node
[Figure 5-1]. The charge transferred to the detection node
is converted into a voltage by MOSFETs for conversion
via the relation Q = C V. The detection node is reset by the
MOSFET for reset to the reference level (voltage on RD)
in order to read the next signal.
Noise accompanying the charge detection by FDA is
determined by the capacitance of the node but can be
almost entirely eliminated by CDS (correlated double
sampling) invented by White.
The signal charge output timing is synchronized with the
timing at which the summing gate (SG) goes from high
level to low level, which is the last clock gate for the shift
register.
The output voltage undergoes an impedance conversion
(gain < 1) through the three-stage source follower circuit
and transmitted as OSA and OSB. Note that the external
load resistor (2.2 kΩ) in Figure 5-1 is not included in
the back-thinned TDI-CCD, so it must be connected
externally.
Detail
Correcting output variations
Variations occur in the output of each port due to
differences in the characteristics of readout amplifiers,
differences in the circuit wiring lengths, and so on.
Moreover, variations in the output may also occur between
columns depending on the operating conditions. As such,
we recommend adding a correction function if necessary.
[Figure 4-10] Timing chart of two-dimensional operation
(a) Port A readout
Integration time
(shutter open)
OSb
RGb
H
L
P2Hb, SGb
H
L
P1Hb
H
L
Readout time (shutter closed)
Tprv, Tpwv, Tpfv
Tovrv
TGb
P1V
P3V
H
L
H
L
TGa
H
L
P2V
Tovrv
H
L
H
L
1
2
3
128
P1Ha
P2Ha, SGa
RGa
OSa
Detail
Tovr
TGa
H
L
P1Ha
H
L
P2Ha, SGa
H
L
RGa
H
L
Tprh, Tpwh, Tpfh
Tprs, Tpws, Tpfs
Tprr, Tpwr, Tpfr
OSa
D1
D2
D3...D8
S1
S2
S509
S253
S510
S254
S511
S255
S512: S10200-02-01, S10201-04-01, S10202-08-01
S256: S10202-16-01
8
(b) Port B readout
Detail
OSb
D1
D2
D3...D8
S1
S2
S509
S253
S510
S254
S511
S255
S512: S10200-02-01, S10201-04-01, S10202-08-01
S256: S10202-16-01
Tprr, Tpwr, Tpfr
RGb
H
L
P2Hb, SGb
H
L
P1Hb
H
L
TGb
H
L
Tovr
Tprh, Tpwh, Tpfh
Integration time
(shutter open)
Tprs, Tpws, Tpfs
Readout time (shutter closed)
OSb
RGb
H
L
P2Hb, SGb
H
L
P1Hb
H
L
Tprv, Tpwv, Tpfv
Tovrv
TGb
H
L
P2V
H
L
H
L
P3V
H
L
TGa
H
L
P1Ha
H
L
P2Ha, SGa
H
L
RGa
H
L
P1V
1
2
Tovrv
3
128
OSa
KMPDC0498EB
[Figure 5-1] CCD output section using FDA
C10000 series TDI camera (related product)
RD
Hamamatsu offers C10000 series TDI cameras with builtin S10201-04-01 back-thinned TDI-CCD and driver circuit.
MOS1
RG
P1
SG
OD
OG
MOS2
MOS4
MOS6
MOS3
Signal charge
Cfd
MOS5
OSA
OSB
C10000-801 (built-in S10201-04-01)
External load
resistor 2.2 kΩ
Product information
Charge transfer
www.hamamatsu.com/all/en/C10000-801.html
KMPDC0502EA
9
Information described in this material is current as of June, 2015.
Product specifications are subject to change without prior notice due to improvements or other reasons. This document has been carefully prepared and
the information contained is believed to be accurate. In rare cases, however, there may be inaccuracies such as text errors. Before using these products,
always contact us for the delivery specification sheet to check the latest specifications.
Type numbers of products listed in the delivery specification sheets or supplied as samples may have a suffix "(X)" which means preliminary specifications or a
suffix "(Z)" which means developmental specifications.
The product warranty is valid for one year after delivery and is limited to product repair or replacement for defects discovered and reported to us within
that one year period. However, even if within the warranty period we accept absolutely no liability for any loss caused by natural disasters or improper
product use.
Copying or reprinting the contents described in this material in whole or in part is prohibited without our prior permission.
www.hamamatsu.com
HAMAMATSU PHOTONICS K.K., Solid State Division
1126-1 Ichino-cho, Higashi-ku, Hamamatsu City, 435-8558 Japan, Telephone: (81) 53-434-3311, Fax: (81) 53-434-5184
U.S.A.: Hamamatsu Corporation: 360 Foothill Road, Bridgewater, N.J. 08807, U.S.A., Telephone: (1) 908-231-0960, Fax: (1) 908-231-1218
Germany: Hamamatsu Photonics Deutschland GmbH: Arzbergerstr. 10, D-82211 Herrsching am Ammersee, Germany, Telephone: (49) 8152-375-0, Fax: (49) 8152-265-8
France: Hamamatsu Photonics France S.A.R.L.: 19, Rue du Saule Trapu, Parc du Moulin de Massy, 91882 Massy Cedex, France, Telephone: 33-(1) 69 53 71 00, Fax: 33-(1) 69 53 71 10
United Kingdom: Hamamatsu Photonics UK Limited: 2 Howard Court, 10 Tewin Road, Welwyn Garden City, Hertfordshire AL7 1BW, United Kingdom, Telephone: (44) 1707-294888, Fax: (44) 1707-325777
North Europe: Hamamatsu Photonics Norden AB: Torshamnsgatan 35 16440 Kista, Sweden, Telephone: (46) 8-509-031-00, Fax: (46) 8-509-031-01
Italy: Hamamatsu Photonics Italia S.r.l.: Strada della Moia, 1 int. 6, 20020 Arese (Milano), Italy, Telephone: (39) 02-93581733, Fax: (39) 02-93581741
China: Hamamatsu Photonics (China) Co., Ltd.: B1201, Jiaming Center, No.27 Dongsanhuan Beilu, Chaoyang District, Beijing 100020, China, Telephone: (86) 10-6586-6006, Fax: (86) 10-6586-2866
Cat. No. KMPD9004E03 Jun. 2015 DN
10