Resistive gate type CCD linear image sensors with electronic shutter / Technical information

Technical information
Resistive gate type
CCD linear image sensor with electronic shutter function
1. Features
 Long photosensitive area, high-speed transfer of charges from the photosensitive area, small image
lag
In ordinary CCD image sensors, one pixel contains multiple electrodes and a signal charge is
transferred by applying different clock pulses to those electrodes [Figure 1]. In photodiode type CCD
linear image sensors, because there is no potential slope in the active area, very large image lag occurs
for rectangular pixels [Figure 2]. In resistive gate structures, a single high-resistance electrode is
formed in the active area, and a signal charge is transferred by means of a potential slope that is
created by applying different voltages across the electrode [Figure 3]. Compared to a CCD area image
sensor which is used as a linear sensor by line binning, a resistive gate type CCD linear image sensor
offers higher speed transfer of charges in the photosensitive area. In addition, compared to photodiode
types, resistive gate types can transfer charges with small image lag in a short period of time even for
long photosensitive areas.
[Figure 1] Schematic diagram and potential of ordinary 2-phase drive CCD area image sensor
N-
P2V
N
N-
P1V
N
N-
P2V
N
N-
N
P
Potential
P1V
KMPDC0320EB
1
[Figure 2] Schematic diagram and potential of photodiode type CCD linear image sensor
STG
P+
N
TG
N-
N
Potential
P
KMPDC0373EB
[Figure 3] Schematic diagram and potential of resistive gate type CCD linear image sensor
REGL
REGH STG
TG
Resistive gate
P+
N
N-
N
Potential
slope
Potential
P
KMPDC0321EC
• Electronic shutter
The electronic shutter function can be used to accumulate a signal charge in as short a time as a few
microseconds. The shutter timing can be synchronized with an external trigger (see “4. Electronic
shutter and anti-blooming”).
• Anti-blooming function
Setting the all reset gate voltage enables controlling of not only the electronic shutter but also the
anti-blooming function (see “4. Electronic shutter and anti-blooming”).
2
• Capable of both non-MPP operation and MPP operation
When performing repeated measurements over a short integration time (e.g., 1 ms or less) in which
dark current can be ignored, use the sensor in the state where a potential slope is always formed at all
times (non-MPP operation). On the other hand, when accumulating over a long integration time, dark
current may be a problem. In this situation, apply a specific voltage to the resistive gate during
integration. This will set the portions under the resistive gate to the inverted state, thus reducing the
dark current (MPP operation, see “5. Non-MPP operation and MPP operation”).
• High sensitivity over a wide spectral range, nearly flat spectral response [Figure 4]
[Figure 4] Spectral response (without window)
(Typ. Ta=25 °C)
100
Quantum efficiency (%)
80
60
40
20
0
200
400
600
800
1000
1200
Wavelength (nm)
KMPDC0316EA
• Low etaloning
With back-thinned CCDs, when the incident light has a long wavelength, etaloning may occur due to
interference. This product uses a back-thinned CCD, but etaloning is reduced by adopting our unique
structure that is less prone to interference [Figure 5].
3
[Figure 5] Etaloning characteristics (typical example)
(Ta=25 °C)
110
100
Etaloning-improved type
Relative sensitivity (%)
90
80
70
60
Previous type
50
40
30
20
10
0
900 910 920 930 940 950 960 970 980 990 1000
Wavelength (nm)
KMPDC0284EB
4
[Table 1] Comparison of CCDs with similar pixel sizes (Typ.)
Product name
Resistive gate type CCD linear
image sensor with electronic
shutter function
CCD area image sensor
S11155/S11156-2048
Type no.
Features
Application example
CCD type
Photosensitive area
structure
Pixel pitch
Photosensitive area
length
CCE
Full well capacity
Dark current
(MPP operation)
S10420-1106-01




Low dark current
High sensitivity
No image lag
Low noise
S11071-1106




-01
(Previous type)
-02
(Improved type)
Spectrophotometry
(low-level light)
Back-thinned type
Low dark current
High sensitivity
No image lag
High-speed
readout
Spectrophotometry
(low-level light)
Back-thinned type
CCD
CCD
Resistive gate type CCD
14 µm
14 μm
896 μm
896 μm
6.5 μV/e 300 ke -
8 μV/e 200 ke -
3.5 ke -/pixel/s
3.5 ke -/pixel/s
14 μm
S11155 series: 500 μm
S11156 series: 1000 μm
8 μV/e 10 μV/e 200 ke
S11155 series:
S11155 series:
4 ke -/pixel/s
10 ke - /pixel/s
S11156 series:
S11156 series:
8 ke -/pixel/s
15 ke - /pixel/s
S11155 series:
S11155 series:
50 ke - /pixel/s
100 ke - /pixel/s
S11156 series:
S11156 series:
100 ke - /pixel/s
200 ke - /pixel/s
30 e - rms
6670
High-speed amp
5 MHz typ.




High-speed readout
Long photosensitive area
Small image lag
Electronic shutter function
Spectrophotometry
(low to high-level light)
Back-thinned type
(full line binning)
(full line binning)
-
-
6 e - rms
50000
Low-noise amp
0.25 MHz typ.
0.1 kHz typ.
30 e - rms
6670
High-speed amp
5 MHz typ.
0.65 kHz typ.
Integration time* 1
5 ms to 40 s
1.3 ms to 30 s
Image lag* 2
Electronic shutter
Readout time of
charges in the
photosensitive
area* 3
Anti-blooming
function
Horizontal register
light shield
None
None
None
None
-
-
Yes
(horizontal register)
Yes
(horizontal register)
Yes (storage gate)
-
-
Light shielding property on -02 is
better than -01.* 4
Dark current
(non-MPP
operation)
Readout noise
Dynamic range
Output circuit
Pixel rate
Line rate
(full line binning)
(full line binning)
2 kHz typ.
2 μs to 1 ms (Non-MPP operation)
1 ms to 7 s (MPP operation)
0.1% typ., 1% max.
Yes
S11155 series:
S11155 series:
80 μs
5 μs
S11156 series:
S11156 series:
300 μs
20 μs
*1: Varies depending on the operating conditions
*2: Varies depending on the light incident conditions and sensor operating conditions (see “6. Unread
charges”)
*3: Time when image lag in the photosensitive area becomes 0.1% typical or less
*4: See “(3) Light shielding of the horizontal shift register in 12. Other information”
5
2. Device structure
[Figure 6] Device structure (S11155/S11156-2048-01)
(a) Entire device drawing
Effective pixels
Thinning
Effective pixels
22
21
20
19
S2045
S2046
S2047
S2048
Thinning
23
Resistive gate D5 D6 D7 D8 D9 D10 S1 S2 S3 S4
18
17
16
D11 D12 D13 D14 D15 D16
Storage section
24
Horizontal shift register
D1 D2 D3 D4
15
D17 D18 D19 D20
14
13
1
Horizontal
shift register
2
3
4
5
6
7
8
9
10
11
12
KMPDC0339ED
(b) Enlarged view [dotted line section in (a) Entire device drawing]
14 µm 14 µm
Channel stop
REGL
Signal charge flow
Resistive gate
(photosensitive area)
REGH
Storage gate
(storage area)
STG
A
A'
ARD
ARG
TG
Transfer gate
Horizontal shift register
All reset
(Electronic shutter/
anti-blooming function)
P2H
P1H
KMPDC0374EB
6
[Figure 7] Device structure (S11155/S11156-2048-02)
(a) Entire device drawing
Effective pixels
Thinning
Horizontal
shift register
22 19 18 17 16
Effective
pixels
Horizontal CCD shift register
23
D77 D78 D79 D80
21
20
・
・
24
7
D1D2 ・
・
2
Resistive gate area D65 D66 D67 D68 D69 D70 S1 S2 S3 S4
・・
S2045
S2046
S2047
S2048
Thinning
・
・ D63 D64
8
D71 D72 D73 D74 D75 D76
・
・ Storage area
14
13
1
・
・ D63 D64
5
3
4
6
Horizontal CCD shift register
9
10
D77 D78 D79 D80
15
11
12
Horizontal
shift register
KMPDC0543EB
7
(b) Enlarged view [dotted line section in (a) Entire device drawing]
All reset
(Electronic shutter/
anti-blooming function)
P1H
P2H
Transfer gate
Storage gate
(storage area)
TG
ARG
ARD
STG
REGH
Signal charge flow
Resistive gate
(photosensitive area)
Channel stop
REGL
14 µm 14 µm
Area in which the light is incident
on the bottom side of the
photosensitive area
Signal charge flow
REGH
Storage gate
(storage area)
STG
ARD
ARG
TG
Transfer gate
Horizontal shift register
All reset
(Electronic shutter/
anti-blooming function)
P2H
P1H
KMPDC0585EA
In the case of the improved type (S11155/S11156-2048-02), signal charges generated by photoelectric
conversion at each pixel of the photosensitive area are directed upward or downward based on the
boundary line at the center of the photosensitive area and transferred. Then, the signals are combined
through the horizontal shift register and read out by a single amplifier. With the improved type, the
charge transfer distance can be reduced even with the same photosensitive area height. This has
enabled the reduction of the readout time (charge transfer time).
Note that since the improved type has a structure that combines signals with horizontal shift register
and reads out the results, there are more horizontal blank pixels than the previous type
(S11155/S11156-2048-01) (number of blank pixels: previous type: 4, improved type: 64). The
improved type has more number of horizontal readout pixels than the previous type, but the line rate is
the same (2 kHz typ.).
To accumulate as many signal charges as possible, the light incident position and distribution must be
made as horizontally symmetrical as possible at the center of the photosensitive area. As shown in
Figure 7 (b), if the light incident position and distribution are greatly offset upward or downward,
signal charges may exceed the saturation charge of the storage gate. The saturation charge of the top
and bottom storage gates is approximately 150 ke - each, and any charges exceeding this value are
8
discarded in the adjacent all reset drain (ARD). As a result, as shown in the graph of Figure 8, the
linearity of the “sum of the signals of each side” changes during the integration time in which the
“signal of the bottom side” has reached saturation. In addition, when the “signal of the bottom side”
saturates, the difference in the output of odd and even pixels becomes apparent. This phenomenon
occurs because the full well capacity of the storage gate is different between odd and even pixels (due
to the ARD position offset).
If the bias in the light cannot be avoided, the applied ARG or ARD voltage must be changed in order to
increase the saturation charge of the storage gate on one side (in this case, anti-blooming does not
work).
Example: ARG’s low voltage: +1 V → 0 V, ARD: +14 V → +13 V
[Figure 8] Output linearity when the light is incident on the bottom side of the photosensitive area
[typical example, see Figure 7 (b)]
250
Sum of both signals
200
Output (ke-)
Signal of the bottom side
150
100
Signal of the top side
50
0
Integration time
KMPDB0458EA
3. Signal detection flow
• Resistive gate (photosensitive area)
To perform charge transfer at high speeds, a resistive gate structure is used in the photosensitive area.
The back of the resistive gate is thinned, which enables the CCD to achieve high quantum efficiency
over a wide spectral range in the same manner as ordinary back-thinned CCDs. The signal charge that
is generated through photoelectric conversion at the resistive gate is transferred to the storage gate by
the potential slope that is formed by the difference in the voltages that are applied across the resistive
gate (applied to REGH and REGL).
Furthermore, the photosensitive area has a structure that suppresses the etaloning (an interference
phenomenon that is characteristic of back-thinned CCDs) that occurs in near infrared wavelengths.
Light incidence is suppressed in the areas of the chip other than the photosensitive area by a thick
layer of silicon substrate.
9
• Storage gate
The signal charge that is transferred from the resistive gate is accumulated in the storage gate. A
lateral type anti-blooming structure is next to the storage gate. This provides both an anti-blooming
function and an electronic shutter function (see “4. Electronic shutter and anti-blooming”).
• Transfer gate
The transfer gate is located between the storage gate and the horizontal shift register. When the
transfer gate is set to high level, the signal charge that is accumulated in the storage gate is
transferred to the transfer gate. When the transfer gate is then set to low level, the signal charge is
transferred to the horizontal shift register.
• Horizontal shift register
The signal charge that has been transferred to the horizontal shift register is transferred to the output
stage when clock pulses are applied to the horizontal shift register.
• Output stage
A MOSFET for charge-to-voltage conversion that is known as an FDA (floating diffusion amplifier) is
embedded in the output stage [Figure 9]. The signal charge is transferred to capacitance Cfd, and the
charge-to-voltage conversion occurs [Equation (1)].
Vfd = Signal charge / Cfd ・・・・・・・・・・・・ (1)
Vfd: output voltage
This voltage is impedance-converted through a two-stage MOSFET source follower circuit (gain < 1)
and output from the OS terminal. Because the CCD does not include an external load resistance (2.2
kΩ), it must be provided through an external circuit.
10
[Figure 9] CCD output section using FDA
RD
MOS1
RG
P1
SG
OD
OG
MOS2
MOS4
Signal charge
Cfd
0.4 mA typ.
MOS3
Vret
OS
External load resistance
2.2 kΩ
Charge transfer
KMPDC0383EB
4. Electronic shutter and anti-blooming
A lateral type overflow drain structure (one for 2 pixels) is provided next to the storage gates of the
resistive gate type CCD linear image sensor to achieve the electronic shutter function.
When the all reset gate (ARG) is set to low level, the signal charge is accumulated in the storage gate
(STG). This is the state when the electronic shutter is open [Figure 10 (a)]. Any signal charge that
exceeds the full well capacity is carried away into the all reset drain (ARD). This prevents blooming.
When the all reset gate is set to high level, the signal charge is carried away from the storage gate into
the all reset drain. Therefore, the signal charge is not accumulated. This is the state when the
electronic shutter is closed [Figure 10 (b)].
The full well capacity of the storage gate can be controlled by the all reset gate voltage. If the
saturation charge of the storage gate is increased, blooming may occur at the storage gate. If the
saturation charge of the storage gate is increased even further, blooming will occur in the latter-stage
horizontal shift register and the like.
11
[Figure 10] Schematic diagram and potential of all reset section (cross-section structure of the area
denoted by broken line A-A' in [Figure 6 (b)])
(a) ARG: low level, Electronic shutter: open [integration state (anti-blooming state)]
ARD
P+
N
N-
N+
ARG
STG
N-
N
P+
Potential
KMPDC0375EA
(b) ARG: high level, Electronic shutter: closed (reset state)
ARD
P+
N
N-
N+
ARG
STG
N-
N
P+
Potential
KMPDC0376EA
 Integration time
When an electronic shutter is used, the integration time is defined as the time from when the all reset
gate is set to low level to when the accumulated signal charge is transferred to the horizontal shift
register (the length of time until the transfer gate becomes low level [Figure 11]). Any signal charge
that is generated outside of this period is carried away into the all reset drain during the period that the
all reset gate is set to high level and is not read out as signals.
If the electronic shutter function is not used, the all reset gate is set to low level at all times. In this
situation, the integration time is the same as the readout cycle.
12
[Figure 11] Timing chart (when electronic shutter is in use)
Integration time
Readout cycle
All reset gate
Transfer gate
Horizontal shift register*
Readout period
KMPDC0287EC
5. Non-MPP operation and MPP operation
MPP (multi-pinned phase) operation is performed by setting the portions under all MOS structure gates,
which constitute the CCD electrodes, to the inverted state (except for storage gate). In MPP operation,
the oxide film interface is inverted by holes supplied from the channel stop region. The generation of
thermally excited electrons on the oxide film interface is drastically suppressed, thus resulting in a
state with low dark current. This state is known as pinning. By precisely applying the pinning voltage
(the gate voltage during pinning), the dark current can be greatly reduced. This operation is suitable
for integrations over a long period of time. Consider the amount of dark current generated and the
length of time needed to read out the signals (shutter time), and select non-MPP operation or MPP
operation according to the operating conditions.
(1) Non-MPP operation
In non-MPP operation, a different fixed voltage is applied to the resistive gate (REGH, REGL) to form a
potential slope at all times. In this situation, if a voltage lower than the pinning voltage is applied, the
potential slope does not form in the CCD channels, and the image lag increases. In non-MPP operation,
because the portions under the resistive gate are not in the pinning state, the dark current becomes
large. However, this operation is suitable for integrations over a short period of time where the effect
of dark current is small. Non-MPP operation is suitable when detecting intense light by using an
electronic shutter to decrease the integration time and reduce the incident light level.
The potentials at various times [Figure 12] in non-MPP operation are shown in Figure 13.
 T1
The signal charge that is generated through photoelectric conversion at the resistive gate (REG) is
transferred to the storage gate by the potential slope.
 T2
The all reset gate is set to high level. The signal charge is carried away into the all reset drain. Any
signal charge that was generated by incident light before this period is not read out (electronic shutter:
closed).
 T3
The all reset gate is set to low level, and signal charge is integrated (electronic shutter: open).
13
 T4
The transfer gate is set to high level. The signal charge that was accumulated in the storage gate is
transferred to the transfer gate. Note that a portion of the signal charge that is generated during this
period is also read out.
 T5
P1H is set to high level. Note that a portion of the signal charge that is generated during this period is
also read out.
 T6
The transfer gate is set to low level. The signal charge is transferred from the transfer gate to the
horizontal shift register (P1H). After this point, when clock pulses are applied to P1H and P2H, the
signal charge is transferred to the output stage via the horizontal shift register.
[Figure 12] Timing chart (S11156-2048-02, non-MPP operation)
Output period of one line
Tinteg
(electronic shutter: open)
Tpwar
(electronic shutter: closed)
ARG
REGH, REGL (REGH=+1 V, REGL=-7 V)
Tpwv Tovr
TG
Tpwh, Tpws
2
1
P1H
3..2127
2128
2129
2130...
N*1
P2H
SG
Tpwr
RG
*2
OS
D1
D2
D79
D80
D3..D70, S1...S2048, D77, D78
Normal readout period
T1
T2
Dummy readout period
T3
T4 T5 T6
*1: Set the total number N of clock pulses according to the integration time.
Apply clock pulses to appropriate terminals during dummy readout period.
*2: To prevent noise from being superimposed on the OS at the falling edge of the ARG clock pulse,
set the ARG rise/fall time to 200 ns or more.
KMPDC0377EC
14
[Figure 13] Potentials at various times in non-MPP operation
REGL
P+
REGH STG
N
TG
N-
P1H
N
N-
P2H
N
N-
N
P
T1
T2
Reset
T3
T4
T5
T6
KMPDC0379EB
(2) MPP operation
In MPP operation, the same pinning voltages are applied to both ends of the resistive gate to reduce
dark current during integration. MPP operation is suitable for detecting low-level light over a long time
period. However, because the image lag in the photosensitive area becomes large in this state, when
transferring the charge, to form a potential slope, a voltage must be applied to the resistive gate at the
same level as the voltage applied during non-MPP operation, and the readout time must be set to ①
or ② shown below. To apply anti-blooming at intense level, ① is recommended.
① Change the REGH and REGL voltage levels to high before the TG level becomes high to make the
high level time longer [Figure 14].
15
② Make the high level voltage period of REGH and REGL longer by synchronizing to the TG high level
period (Tpwv).
The potentials at various times [Figure 14] in MPP operation are shown in Figure 15.
 T1
The signal charge that is generated by light entering the resistive gate (REG) is accumulated in the
storage gate.
 T2
The all reset gate is set to high level. The signal charge is carried away into the all reset drain. Any
signal charge that was generated by incident light before this period is not read out (electronic shutter:
closed).
 T3
The all reset gate is set to low level, and signal charge is integrated (electronic shutter: open).
The voltage that is applied across the resistive gate is set to a voltage that is less than the pinning
voltage. This is the state in which the dark current generated in the resistive gate is reduced (the
potential slope does not form).
 T4
Different voltages are applied across the resistive gate. The potential slope forms, and the signal
charge that was generated by light entering the resistive gate is transferred to the storage gate. Note
that a portion of the signal charge that is generated during this period is also read out.
 T5
The transfer gate is set to high level. The signal charge that was accumulated in the storage gate is
transferred to the transfer gate. Note that a portion of the signal charge that is generated during this
period is also read out.
 T6
P1H is set to high level. Note that a portion of the signal charge that is generated during this period is
also read out.
 T7
The transfer gate is set to low level. The signal charge is transferred from the transfer gate to the
horizontal shift register (P1H). After this point, when clock pulses are applied to P1H and P2H, the
signal charge is transferred to the output stage via the horizontal shift register.
16
[Figure 14] Timing chart (S11155/56-2048-02, MPP operation)
Output period of one line
Tinteg
(electronic shutter: open)
Tpwar
(electronic shutter: closed)
ARG
REGH, REGL
Tpwreg
(REGH, REGL=-9.5 V)
(REGH=+1 V, REGL=-7 V)
Tregtr
Tpwv Tovr
TG
Tpwh, Tpws
2
1
P1H
3..2127
2128
2130...
2129
N*
P2H
SG
Tpwr
RG
OS
D1
D2
D79
D80
D3..D70, S1...S2048, D77, D78
Normal readout period
T1
Dummy readout period
T2
T3
T4
T5 T6 T7
* Set the total number N of clock pulses according to the integration time.
KMPDC0378EC
17
[Figure 15] Potentials at various times in MPP operation
REGL
P+
REGH STG
N
TG
N-
P1H
N
N-
P2H
N
N-
N
P
T1
T2
Reset
T3
T4
T5
T6
T7
KMPDC0380EB
6. Unread charges
Unread charges (image lag) are defined as a percentage of the signal level left unread of the incident
light level and are expressed by equation (2).
18
L = (Slag/S) × 100 [%] ・・・・・・・・・(2)
L: image lag
Slag: unread signal level
S: input signal level
Image lag occurs in the resistive gate and storage gate and is evaluated using different methods.
(1) Measurement method and characteristics of image lag in the resistive gate
Image lag in the resistive gate is measured in the following manner at the timing shown in Figure 16.
① Use a pulse-driven LED (peak emission wavelength: 660 nm), and adjust its output so that a signal
about half the full well capacity is applied to the photosensitive area. During this adjustment, set
ARG to low level when the LED is on. The input signal level at this point is S [Equation 2].
② While the LED is on, set the ARG to high level (electronic shutter: closed). The charges transferred
to the storage gate are carried away into the ARD (T1 in Figure 16).
③ When a given period (α) elapses after the LED is turned off, set the ARG to low level (electronic
shutter: open). Unread charges in the resistive gate (Slag) are accumulated in the storage gate
(T2).
④ Charges accumulated in the storage gate (see ③) are output (T3).
The longer the resistive gate, the longer the charge transfer time and the larger the image lag. With
the improved type, the charge transfer speed was increased by optimizing the resistive gate structure
and forming a steep potential slope between the resistive gates. Given the same charge transfer time,
the image lag is shorter on the improved type than the previous type [Figure 17].
[Figure 16] Timing chart example for measuring image lag in the resistive gate
1500 µs
500 µs
300 µs
(670 - α) µs
α
30 µs
High
LED
REG
ARG
TG
Low
High
Low
2 µs
High
Low
High
Low
Output
Unread signal (Slag)
T1
T2
T3
KMPDC0577EA
19
[Figure 17] Image lag in the resistive gate vs. time (α) (typical example)
(Ta=25 °C)
100
S11156-2048-01
Image lag (%)
10
S11155-2048-01
1
0.1
S11156-2048-02
0.01
S11155-2048-02
0.001
0
10
20
30
40
50
60
α (µs)
KMPDB0450EA
 Relationship between the image lag in the resistive gate and input light pulse width
Figure 19 shows the image lag when the input signal level is constant and the LED’s light emission
pulse width is changed at the timing shown in Figure 18. The larger the pulse width, the smaller the
image lag. As the signal charge that is converted from the incident light is transferred to the storage
gate by the potential slope of the resistive gate, the image lag will be smaller when a low level light is
applied over a long time period.
[Figure 18] Timing chart example for measuring image lag in the resistive gate
1500 µs
30 µs
TG
500 µs
Light emission
pulse width* α
µs
LED
2 µs
ARG
* 2 µs, 5 µs, 10 µs, 100 µs, 300 µs
KMPDC0578EA
20
[Figure 19] Image lag in the resistive gate vs. incident light pulse width (typical example)
(a) S11155-2048-01/-02 (α=5 μs)
(Average of all pixels)
100
S11155-2048-01
S11155-2048-02
Image lag (%)
10
1
0.1
0.01
2
5
10
100
300
Incident light pulse width (µs)
KMPDB0451EA
(b) S11156-2048-01/-02 (α=20 μs)
(Average of all pixels)
100
S11156-2048-01
S11156-2048-02
Image lag (%)
10
1
0.1
0.01
2
5
10
100
300
Incident light pulse width (µs)
KMPDB0452EA
(2) Measurement method and characteristics of image lag in the storage gate
As shown in Figure 20, a pulse-driven LED (peak emission wavelength: 660 nm) is used at about half
the saturation output. After sufficient time (approx. 500 s) elapses for the charge generated in the
resistive gate to be transferred to the storage, the TG voltage is set to high level to transfer the charge
to the horizontal shift register. Most of the input signal level (S) of equation (2) is read out during the
first output period, and the remaining signal (Slag) is output during the second output period.
It is possible to reduce the image lag in the storage gate as the TG high level period (Tpwv) is
21
increased. With the improved type, the potential structure of the storage gate is optimized to reduce
the image lag in the storage gate as compared to the previous type [Figure 21].
[Figure 20] Timing chart example for measuring image lag in the storage gate (Non-MPP operation)
300 µs
500 µs
High
LED
Low
High
Low
REG
First output period Second output period
Tpwv
High
ARG
Low
High
TG
Low
Output
Unread signal level (Slag)
Input signal level (S)
KMPDC0579EA
[Figure 21] Image lag in the storage gate vs. TG_high period (typical example)
(Ta=25 °C)
Image lag (%)
1
0.1
S11156-2048-01
0.01
S11156-2048-02
0.001
0.1
1
10
100
TG_high period (µs)
KMPDB0453EA
7. Dark current temperature dependence
Dark current has temperature dependence. For example, in the case of the S11156-2048-02, the
temperature dependence during MPP operation is as shown in Figure 22.
22
[Figure 22] Dark current temperature dependence (S11156-2048-02, MPP operation, typical example)
105
Dark current (ke-/pixel/s)
104
Double at 5 °C
103
102
Double at 7 °C
101
100
10-1
0
10
20
30
40
50
60
70
80
90
100
Temperature (°C)
KMPDB0454EA
8. Dark shot noise and dynamic range
As the integration time is increased, the dark shot noise is also increased. This causes the dynamic
range to be reduced. Figure 23 shows an example.
[Figure 23] Noise (readout noise + dark shot noise), dynamic range vs. integration time
(Typ. Td=25 °C)
Noise (e- rms)
104
Non-MPP (noise)
MPP (noise)
Non-MPP (dynamic range)
MPP (dynamic range)
104
103
103
102
102
101
101
10-3
10-2
10-1
100
101
102
103
Dynamic range
105
100
104
Integration time (ms)
KMPDB0455EA
9. Linearity
Figure 24 shows the linearity error (S11156-2048-02) when the incident light level is made constant,
and the signal charge level is varied by changing the integration time with the electronic shutter.
23
 Measurement conditions
• Light source output: constant, on at all times
• Tpwv [period in which TG is at high level]=20 s
• Average the data that has been acquired consecutively 64 times and then average all the channels
to derive the signal.
Linearity error = {1 – (Sm/Tm)/(S/Tinteg)} × 100 [%] ・・・・・・・・・(3)
Sm: signal level at one-half the saturation charge
Tm: integration time at one-half the saturation charge
S: signal
Tinteg: integration time
If the integration time (Tinteg) is short, the image lag increases, and the linearity error shifts in the
negative direction.
[Figure 24] Linearity error vs. signal charge (integration time varied with electronic shutter:
Tinteg=20 s to 5000 s)
(Typ. Tinteg: 20 µs to 5000 µs)
Linearity error (%)
5
0
-5
-10
102
103
104
105
106
Signal charge (e-)
KMPDB0456EA
10. Operation method
(1) Operating voltage
To reduce image lag, the operating voltage on the improved type was changed [Table 2].
24
[Table 2] Voltage applied to each terminal (comparison between the previous type and improved type)
S11155/S11156-2048-01
S11155/S11156-2048-02
Terminal
(Previous type)
(Improved type)
name
Min.
Typ.
Max.
Min.
Typ.
Max.
VOD
12
15
18
12
15
18
VRD
14
15
16
13
14
15
VARD
14
15
16
13
14
15
VARGH
7
8
9
7
8
9
VARGL
-2
-1.5
-1
0.5
1
2
VOG
2.5
3
3.5
2.5
3.5
4.5
VSTG
-
0
-
2.5
3.5
4.5
VSS
-
0
-
-
0
-
VREGHH
-4.5
-4
-3.5
0.5
1
1.5
VREGHL
-9
-8
-7
-10.5
-9.5
-8.5
VREGLH
-
VREGHH - 2.5
-
-
VREGHH - 8
-
VREGLL
-9
-8
-7
-10.5
-9.5
-8.5
Vret
-
1
2
-
1
2
VISH
-
VRD
-
-
VRD
-
VIGH
-9
-8
-
-10.5
-9.5
-
VPHH
4
5
6
5
6
8
VPHL
-8
-7
-6
-6
-5
-4
VSGH
4
5
6
5
6
8
VSGL
-8
-7
-6
-6
-5
-4
VRGH
7
8
9
7
8
9
VRGL
-6
-5
-4
-6
-5
-4
VTGH
8.5
9
9.5
9.5
10.5
11.5
VTGL
-7.5
-7
-6.5
-6
-5
-4
(2)Current flowing through terminals
Table 3 indicates an example of current that flows through each terminal (S11156-2048-02). Use this
as a reference to design your circuit.
25
[Table 3] Approximate current that flows through each terminal
S11155/ S11156-2048-01
Terminal name
S11155/S11156-2048-02
Supply voltage
Current
Supply voltage
Current
Typ.
(mA)
Typ.
(mA)
OS
-
-
-
-
OD
+15
+6
+15
+6
OG
+3
+0.1
+3.5
+0.1
SG
+5/-7
+0.1/-0.1
+6/-5
+0.1/-0.1
Vret
+1
-0.4
+1
-0.4
RD
+15
<<0.1
+14
<<0.1
REGL
-6.5/-8
-1/-0.2
-7/-9.5
-10/-0.2
REGH
-4/-8
+1/-0.2
+1/-9.5
+10/-0.2
P2H
+5/-7
+0.1/-0.1
+6/-5
+0.1/-0.1
P1H
+5/-7
+0.1/-0.1
+6/-5
+0.1/-0.1
IG2H
-8
-0.1
-9.5
-0.1
IG1H
-8
-0.1
-9.5
-0.1
ARG
+8/-1.5
+0.1/-0.1
+8/+1
+0.1/+0.1
ARD
+15
<+0.1
+14
<+0.1
ISH
+15(=VRD)
<<+0.1
+14(=VRD)
<<+0.1
STG
+3
+0.1
+3.5
+0.1
NC(STG)
-
-
-
-
TG
+9/-7
+0.1/-0.1
+10.5/-5
+0.1/-0.1
RG
+8/-5
+0.1/-0.1
+8/-5
+0.1/-0.1
The Vret and REGL power supplies must be of a sink type, and REGH of a source type. The power
supplies of other terminals must be of a source type if the applied voltage is positive and sink type if
negative. Inrush current and the like are different from the values in Table 3 (they depend on the
capacitance and resistance of the sensor and the driver circuit).
11. Use examples
(1) Controlling the incident light level with electronic shutter
If the incident light level is large, any signal charge that exceeds the saturation charge is carried away
into the all reset drain. At this point, the signal output saturates and becomes constant, making the
verification of the output difference between each pixel impossible. In such situations, the output
difference between each pixel can be verified by controlling the signal level by varying the integration
time with an electronic shutter [Figure 26]. Multiple signals with different integration times are
acquired, and the data is processed using the signals of unsaturated pixels and their integration times
to virtually increase the dynamic range. Note that the integration time can be reduced to 2 s in the
case of a resistive gate type CCD linear image sensor with an electronic shutter function.
26
[Figure 25] Timing example when electronic shutter is in use
(a) Long integration time
LED
Integration time
Integration time
ARG
TG
KMPDC0583EA
(b) Short integration time
LED
Integration time
Integration time
ARG
TG
KMPDC0584EA
[Figure 26] Output example when electronic shutter is in use
Long integration time
Output
Short integration time
Horizontal pixels
KMPDB457EA
27
(2) Integrating the signal obtained through multiple electronic shutter operations
If the product is operated at the timing shown in Figure 27, excitation operation is repeated due to the
spark light, and the resulting fluorescence signal charge is accumulated temporarily in the transfer
gate. Then, the horizontal shift register can be used to read out the signal of each channel. This
readout method is effective in increasing the signal level when the signal light obtained by each
excitation operation is small.
[Figure 27] Timing chart (when accumulating and reading out the signal obtained through multiple
electronic shutter operations)
Output period of one line
Tinteg1
Tpwar
Tinteg2
Tinteg
Spark light
Fluorescence
REG
Electronic shutter: open
Electronic shutter: closed
ARG
Tpwh
Tpws
TG
P1H
1
2
3.. 2067
2068
2069
2070
N
P2H
SG
RG
OS
D1
D2
D19 D20
D3..D10. S1...S2048. D11..D18
Normal readout period
Dummy readout period
KMPDC0581EA
12. Other information
(1) Measures to eliminate the effect of light from the output circuit
If the operating conditions are not appropriate, light may be generated from the output circuit. If this
generated light is received by the resistive gate, storage gate, or horizontal shift registers, a
phenomenon may occur in which the output is large for the first pixels that are read [Figure 28].
28
[Figure 28] Effect of the light generated by the output circuit (horizontal profile in a dark state, typical
example)
3000
Output (digital number)
2500
2000
1500
1000
500
0
0
50
100
150
200
250
300
Horizontal pixel no.
KMPDB0328EA
To reduce the effect of this light emission, the following measure is effective.
① Apply a positive voltage to the Vret terminal (the optimal voltage varies depending on the product).
② Make the horizontal shift register clock pulses (P1H, P2H, SG) cross each other at 50% ± 10% of
their amplitudes [Figure 29].
[Figure 29] Example of horizontal shift register clock pulse waveform
P1H
50%
P2H
KMPDC0381EA
③ After all pixels are read out, perform dummy readout in the horizontal direction until right before
TG goes to high level.
When performing a comparatively long integration, to carry away the charge that is accumulated in the
horizontal shift registers, even after all pixels are read out, perform dummy readout of the horizontal
shift register until right before the transferring of the signal charge to the transfer gate begins (see *1
of Figures 12 and 14).
(2) All reset gate (ARG) pulse timing and effect on OS output
If accumulation is started (the ARG pulse falls) at any time during the signal readout period, the ARG’s
clock feedthrough is mixed into the OS signal and becomes noise (see *2 of Figure 12). To reduce this
29
effect, in the datasheet (S11155/S11156-2048-01/-02), the minimum value of the ARG’s fall time
(Tpfar) is set to 200 ns.
(3) Light-shielding of horizontal shift register
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), and short-wavelength light hardly reaches the
horizontal shift register. However, long-wavelength light may pass through the insensitive area of the
silicon and may be received by the horizontal shift register [Figure 31].
If a mechanical shutter or the like is not used, light will enter the horizontal shift register during charge
integration and transfer, and the spurious signal will be superimposed on the actual signal. For example,
if a time invariant light enters the horizontal shift register, the output signal of each channel will
increase by the same proportion. These effects will be smaller with shorter horizontal transfer time
periods (higher output signal frequency).
The horizontal shift register is shielded from light in the following ways.
①
②
③
Apply light only during the integration period, read out the unneeded charges accumulated in the
horizontal shift register during this period, and then read out the actual signal.
Adjust the position where the light is directed to prevent light from entering the horizontal shift
register.
Attach a light-shielding mask to prevent light from entering the horizontal shift register.
With the improved type, the horizontal shift register has been moved away from the photosensitive
area to reduce the effects of the light transmitted by light path 1. [Figure 31].
In addition, the
light-shielding metal attached to the front side of the CCD chip reduces components of light
transmitted through the photosensitive area, reflected by the metal film of the package (light path 2),
and entering the horizontal shift register. As a result, the light input to the horizontal shift register was
reduced to approximately 1/40 to 1/50 (LED peak wavelength: 880 nm) as compared with the previous
product.
30
[Figure 30] Device structure (improved type)
Effective pixels
Thinning
Horizontal
shift register
22 19 18 17 16
Effective
pixels
・・
23
D77 D78 D79 D80
21
20
・・
24
7
D1D2 ・・
2
Resistive gate area D65 D66 D67 D68 D69 D70 S1 S2 S3 S4
S2045
S2046
S2047
S2048
Thinning
Horizontal CCD shift register
D63 D64
・・
8
D71 D72 D73 D74 D75 D76
・・ Storage area
14
13
1
・・
5
3
4
Horizontal CCD shift register
D63 D64
6
9
10
D77 D78 D79 D80
15
11
12
Horizontal
shift register
KMPDC0543EB
[Figure 31] Sensor cross section (cross section at the red dotted line in Figure 30), incident light path
to the horizontal shift register
Light path 1
CCD chip
Light path 2
Resistive gate electrode
Solder bump
Package
Metal film for light shielding
Horizontal shift register
(improved type)
Metal film
Horizontal shift register
(previous type)
KMPDC0582EA
Note that on the improved type, a charge transfer register (blank pixels of the horizontal shift register)
is arranged also on the side with the readout amplifier (area indicated by blue dotted line in Figure 30).
As this area may also detect long-wavelength light, it must be shielded.
• Light-shielding mask for horizontal shift register
To further increase the light-shielding property of the horizontal shift register, Hamamatsu
light-shielding mask can be mounted on the CCD chip [Figure 32]. The vertical aperture size of the
light-shielding mask is 400 m on the S11155-2048-02 and 900 m on the S11156-2048-02. Note that
the light-shielding effect varies depending on the light incident angle and the like.
31
[Figure 32] Sensor configuration (light-shielding mask mounting: only available for the -02 type)
650 µm max.
CCD chip
Mask
120 µm
Horizontal shift register
Effective area
Resistive gate
(photosensitive area)
50 µm
S11155: 400 µm
S11156: 900 µm
50 µm
Storage gate
Storage gate
Mask aperture area
Anti-reflection film
Horizontal shift register
Metal light-shielding film
KMPDC0580EA
(4) Power consumption of resistive gate
The resistive gate is composed of a high-resistance electrode. Applying a voltage across the resistive
gate transfers the signal charge. In this situation, the resistive gate consumes the power that depends
on the supply voltage and its electrode resistance [Table 4].
[Table 4] Power consumption and resistance of the resistive gate
Parameter
Type no.
Symbol
Min.
Typ.
Max.
1.4
2.5
12.5
50
100
160
0.7
1.3
6.3
S11156-2048-02
30
60
90
S11155-2048-01
0.5
2.5
4.5
0.4
0.7
1.4
1
5
9
0.7
1.1
2.2
S11155-2048-01
Power
consumption
Resistance
S11155-2048-02
S11156-2048-01
S11155-2048-02
S11156-2048-01
S11156-2048-02
PREG
R REG
Unit
mW
kΩ
(5) Output waveform example
The readout amplifier on the improved type is improved to support wide bandwidth.
32
Signal
DC level
(reset level)
Reset feedthrough
[Figure 33] OS output waveform example (operating conditions: Typ., unless otherwise noted)
Signal level
(a) S11155-2048-01 (fc=5 MHz typ.)
(b) S11155-2048-02 (fc=10 MHz max.)
(6) Sensor temperature
Figure 34 is a measurement example showing the relation between the sensor temperature and the
operating time when the S11156-2048-01 is operated with our evaluation circuit (circuitry is placed a
sealed case with no heat dissipation measures). The sensor temperature increases markedly when
operated at high speed. Since the rise in the sensor temperature increases the dark current and
accelerates device degradation, we recommend taking heat dissipation measures, such as attaching a
heatsink to the sensor or using a cooling fan.
If you want to achieve stable sensor temperature that is lower than room temperature, using a similar
product, S13255/6 series, with a one-stage TE-cooler is recommended.
[Figure 34] Sensor temperature vs. operating time (using a Hamamatsu evaluation circuit)
(Typ. Ta=25 °C)
50
Sensor temperature (°C)
45
fc=10 MHz
40
fc=5 MHz
35
30
25
20
0
20
40
60
80
100
120
Operation time (min)
KMPDB0331EA
33
13. Driver circuit
(1) Transfer clock pulse generator
As stated above, clock pulses with high and low levels of voltage amplitude are required to operate a
CCD. These clock pulses must drive the vertical shift register and horizontal shift register at high
speeds, which have an input capacitance of several hundred picofarads to several nanofarads. For this
purpose, MOS driver IC is commonly used to drive a CCD since it is capable of driving a capacitive load
at high speeds.
Normally, the timing signal generator circuit uses a PLD or FPGA that can generate TTL or CMOS logic
level output. The output voltage is +3.3 V or +5.0 V, so a level converter circuit must be connected to
the MOS driver IC.
In 2-phase CCD operation, the clock pulses for driving the vertical and horizontal shift registers must
overlap with each other. For this reason, a resistor Rd with an appropriate value (damping resistor: a
few ohms to several kiloohms) should be placed between the MOS driver IC and the CCD in order to
adjust the rise time and fall time of the clock pulses.
To minimize noise intrusion to the CCD from digital circuits, it is recommended that the analog ground
and digital ground be set to the same potential by the transfer clock pulse generator.
[Figure 35] Example of transfer clock pulse generator
VDD
10 µF
P1V
P2V
5
6
Rd
+Vcc
4
100 p
EL7212
2.2 k
2
7
20
2.2 k
Rd
18
17
16
15
14
13
12
11
Y1
Y2
Y3
Y4
Y5
Y6
Y7
Y8
3
G1
G2
100 p
2.2 k
10 µF
A1
A2
A3
A4
A5
A6
A7
A8
74HCT540
0.1 µF
2
3
4
5
6
7
8
9
P1V_in
P2V_in
1
19
10
2.2 k
VEE
VDD: high level voltage of clock pulse
VEE : low level voltage of clock pulse
Rd : damping resistor (a few to several dozen ohms)
KMPDC0052EB
(2) Bias generator circuit
The bias voltage is mainly applied to the peripheral section of the CCD output amplifier, so use a stable
power supply with low noise. It is also important to note the voltage accuracy, voltage fluctuation,
ripple, output current, and the like.
Figure 36 shows an example of a bias voltage generator circuit for the OD terminal. The reference
voltage is generated from the reference power supply IC and is set to a specified voltage value by the
amplifier making up the low-pass filter. This allows the user to obtain a highly stable and accurate
voltage with low noise. Note that there are cases where low-noise linear regulator is used as a bias
voltage generator circuit.
34
[Figure 36] Example of bias voltage generator circuit
51 k
51 k
0.1 μF
+24 V
7
VOD=+20 V
6
VOD
10
+
100
+24 V
1 μF
2
2 10 μF
Vref=+10 V 6
5
3
4
1 μF
+VI
VO
TRIM N.R.
GND
8
1 μF
4
KMPDC0303EA
• Signal processing circuit
Major sources of noise from a CCD are the well-known kT/C noise and 1/f noise. The kT/C noise is
generated by a discharge (reset operation) in the FDA. This noise is inversely proportional to the
square root of the node capacitance (Cfd) of the FDA and makes up a large percentage of the total
noise of a CCD. The 1/f noise is generated by the MOSFET constituting the FDA and is inversely
proportional to the frequency.
These noises degrade the S/N in the CCD system and therefore should be reduced as much as possible
in the signal processing circuit. A typical circuit for this purpose is a CDS circuit.
The operating principle of the CDS circuit is described below. Figure 37 shows an output waveform
from a CCD. As stated above, kT/C noise occurs during a reset period in the FDA. At the point where
the reset period has ended, the voltage level varies due to kT/C noise. Therefore, if data is acquired at
time T2, the S/N deteriorates by an amount equal to the kT/C noise variation. In contrast, acquiring
data at times T1 and T2 on the output waveform and then obtaining the difference between them will
extract only a signal component ∆V with the kT/C noise removed. DC components such as the offset
voltage component and reset feed-through are removed at the same time.
[Figure 37] CCD output waveform
Reset period
Signal output
period
Reset level
kT/C noise
DV
Signal output level
T1
T2
KMPDC0304EA
There are two types of CDS circuits: “Type 1” that uses a clamp circuit in combination with a
sample-and-hold amplifier (SHA), and “Type 2” that uses a SHA in combination with a differential
amplifier. Type 1 has a very simple circuit configuration [Figure 38]. But if the ON resistance of the
switch used in the clamp circuit is large, the amount of noise that can be removed will be small or a DC
voltage error will occur. Ideally, the ON resistance should be 0 Ω.
35
[Figure 38] CDS circuit block diagram (using clamp circuit and SHA)
Preamp (LPF)
Sample signal (T2)
Buffer amp
CCD output
Video output
Clamp signal (T1)
Sample-and-hold amp (SHA)
Clamp circuit
CCD output
Clamp signal (T1)
Sample signal (T2)
0V
Video output
KMPDC0305EA
Type 2 uses a larger number of components but removes noise more effectively than Type 1. However,
since Type 2 makes an analog calculation of the SHA output, the noise of the SHA itself may be added,
resulting in increased noise in some cases. The SHA noise should be small enough so that the kT/C
noise can be ignored.
[Figure 39] CDS circuit block diagram (using SHA and differential amplifier)
Preamp (LPF)
Sample-and-hold
amp (SHA)
Differential amp
CCD output
+
-
Sample signal 1 (T1)
Video output
Sample signal 2 (T2)
KMPDC0306EA
A circuit example of Type 1 is shown in Figure 40.
The preamp gain should be set high in order to sufficiently amplify the CCD output signal. Since the
CCD output signal contains DC voltage components, a capacitor is used for AC coupling. Note that this
capacitor can cause a DC voltage error if the preamp bias current is large. Therefore, a preamp with a
small bias current must be selected. A JFET or CMOS input amplifier is generally used. It is also
necessary to select a low-noise amplifier with a bandwidth wide enough to amplify the CCD output
waveform.
The clamp circuit is made up of capacitors and an analog switch. As the analog switch, we recommend
using a high-speed type having low ON resistance and small charge injection amount.
As with the preamp, the last-stage amplifier is AC-coupled via a capacitor, so a JFET or CMOS input
amplifier should be selected. In addition, a non-inverted amplifier must be configured to allow high
input impedance.
Incidentally the CCD provides a negative-going output while the last-stage amplifier gives a
positive-going output to facilitate analog-to-digital conversion. For this reason, an inverted amplifier is
36
connected after the preamp.
[Figure 40] CDS circuit example (using clamp circuit and SHA)
33 p
1k
+20 V
24
23
22
21
20
19
18
17
16
15
14
13
IG1V
IG2V
ISV
NC
SS
NC
NC
NC
TG
P1V
P2V
ISH
RG
RD
OS
OD
OG
SG
Th1
Th2
P2H
P1H
IG2H
IG1H
S9970-1007
1
2
3
4
5
6
7
8
9
10
11
12
0.1 μF
22 k
100 k
1.5 k
-15 V
1 μF 1 5
4
2
6
3 +
7
1 μF
+15 V
47 p
47 p
1k
1k
-15 V
100
1k
1 μF 8
2 4
3 +
7
1 μF
+15 V
1 μF
UI
2
100
1
4
3 +
1
6
1k
-15 V
1000 p
7
8
6
5
Video Out
1 μF
+15 V
-15 V
1 μF
1
2
3
IN
V+
COM NO
GND
V-
6
5
4
Clamp
1 μF
+15 V
KMPDC0053EC
37
Cat. No. KACC9005E03 Mar. 2016