JPH11220657A - Image sensor and its correction method - Google Patents

Abstract

(57) [Problem] To provide an image sensor capable of detecting weak light with high sensitivity and a correction method thereof. An image sensor according to the present invention includes light receiving units PD1 to PDn for converting an optical signal into an electric signal, and a plurality of CMOS inverters I1 connected in series to the light receiving units.
1, I21-I1n, I2n, the inputs of the first-stage CMOS inverters I11-I1n and the odd-numbered CMOs.
Switch transistors SW1 to SWn each having a source and a drain connected to the output of the S inverter.
The signal is amplified by the OS inverter. Therefore,
Weak light can be detected with high sensitivity.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an image sensor and a method of correcting the same, and more particularly, to an image sensor capable of detecting weak light with high sensitivity and a method of correcting the same.

[0002]

2. Description of the Related Art A conventional image sensor has a light receiving portion (photodiode) for converting a light signal into an electric signal, and detects the light signal by reading out the electric signal output from the light receiving portion. Further, it has been difficult for conventional image sensors to detect weak light with high sensitivity.

The reason why it was difficult to detect weak light with high sensitivity will be described below. Even if the weak light strikes the image sensor, the amount of photocharge generated per unit area of the light receiving unit due to the weak light is small. Therefore, as a method of increasing the amount of generated photocharge even slightly, it is conceivable to increase the area of the light receiving section. However, when the area of the light receiving section is increased, the diffusion capacitance formed by the photodiodes constituting the light receiving section also increases. Therefore, even if a large photocharge Q is generated by increasing the area of the light receiving section, the diffusion capacitance C also increases, so that the voltage V detected by the following equation (1) cannot be increased. V = Q / C (1)

Further, when the area of the light receiving section is large, the capacity of the light receiving section becomes large, so that there is a problem that the time for reading an optical signal from the light receiving section becomes long. FIG. 5 is a circuit diagram illustrating a reading method using a conventional image sensor. When the electric charge is transferred from C1 (light receiving unit capacitance) to C2, if V1 (voltage) and V2 are almost the same, the voltage between the ON resistance r _{ON of} Tr3 becomes small, so the electric charge transfer time Becomes longer. Furthermore, since the light receiving section must be reset (Tr2 is turned on) to a certain potential in order to store the next signal light, a corresponding reset time is required.

[0005]

As described above, the conventional image sensor has a problem that weak light cannot be detected with high sensitivity. In addition, there is a problem that the reading time becomes long.

The present invention has been made in view of the above circumstances, and an object of the present invention is to provide an image sensor capable of detecting weak light with high sensitivity and a method of correcting the image sensor. It is another object of the present invention to provide an image sensor capable of reading an optical signal from a light receiving section at high speed.

[0007]

In order to solve the above problems, an image sensor according to the present invention comprises a light receiving section for converting an optical signal into an electric signal, a CMOS inverter connected to the light receiving section, And a switch transistor for connecting / cutting off an input and an output, wherein the electrical signal output from the light receiving unit is amplified by the CMOS inverter.

In the above image sensor, since the electric signal output from the light receiving section is amplified by the CMOS inverter, the weak light can be detected with high sensitivity. Also,
By arranging a switch transistor in the image sensor and connecting the switch transistor to the input and output of the CMOS inverter, the performance of the CMOS inverter as the amplifying element can be optimized.

Further, the image sensor according to the present invention includes a light receiving portion for converting an optical signal into an electric signal, a plurality of CMOS inverters connected in series to the light receiving portion,
A switch transistor for connecting / disconnecting an input of the MOS inverter and an output of the odd-numbered stage CMOS inverter;
Wherein the electric signal output from the light receiving section is amplified by the CMOS inverters of the plurality of stages.

In the above image sensor, since a plurality of stages of CMOS inverters are connected in series to the light receiving section, the CM
The amplification factor can be increased as compared with the case where the OS inverter is connected in one stage. Further, by connecting the input of the first-stage CMOS inverter and the output of one of the odd-numbered CMOS inverters with a switch transistor, even if the overall gain of the multiple-stage CMOS inverter becomes extremely high, Oscillation caused by feedback through the transistor can be prevented.

A CMO connecting switch transistors
The output of the S inverter is preferably the output of the CMOS inverter of the first stage or an odd-numbered stage close to the first stage.
The reason is as follows. Assuming that the gain of one stage of the CMOS inverter is, for example, 100 times, the gain of five stages is 10 ¹⁰ times. Therefore, for example, the first-stage CMO
When connecting the output of the input and 5-stage CMOS inverter S inverter switching transistors, it takes 10 ¹⁰ times the feedback may cause oscillation by small voltage changes in the noise. On the other hand, for example, when the input and output of the first-stage CMOS inverter are connected by a switch transistor, the feedback becomes 100 times, so that no oscillation occurs.

Further, the CMOS having a plurality of the light receiving portions is provided.
Preferably, a transistor is connected to an output of the inverter, an output of the shift register is connected to a gate of the transistor, and an electric signal output from the light receiving unit is read by driving the shift register.

Further, in the image sensor correcting method according to the present invention, a plurality of light receiving sections for converting an optical signal into an electric signal, a CMOS inverter connected to each of the light receiving sections, and an input and an output of the CMOS inverter are connected. And a switch transistor for shutting off the light, wherein the CMOS inverter corrects the electric signal output from the light receiving unit by the CMOS inverter. The amplification factor of each electric signal amplified by each CMOS inverter when light is irradiated to each light receiving unit is measured in advance, and correction for equalizing each amplification factor is performed. In addition, the CM
It is preferable that a plurality of OS inverters are connected in series.

According to the above-described image sensor correction method, even if the total gain of the CMOS inverters connected to the respective light receiving units varies, the light having a certain intensity distribution is irradiated to the light receiving units. It is possible to know the exact intensity distribution of the irradiation light.

[0015]

DESCRIPTION OF THE PREFERRED EMBODIMENTS As can be seen from the above-mentioned equation (1), in the conventional image sensor, if the detection light is weak, even if the area of the light receiving section is increased, the photocharge Q eventually becomes small, Since the voltage V is also small, it is difficult to detect a weak optical signal. Therefore, the present inventor has considered that a weak signal voltage V output from the light receiving unit is amplified using a CMOS inverter. The CMOS inverter is used as the amplifying means because the CMOS inverter has a simple structure, the area occupied by the chip on which the image sensor is formed can be reduced, and the CMOS inverter is suitable for high integration.

FIG. 2 is a graph showing the input / output characteristics of the CMOS inverter. As can be seen from this graph, if the operating point for biasing the light receiving unit is not in the linear region shown in FIG. 2, even if the signal voltage slightly changes, the output of the CMOS inverter saturates immediately, and the weak light is sensitive. Detection cannot be performed well.

It is ideal that the intensity of the weak light to be detected is proportional to the output voltage amplified by the CMOS inverter. For that purpose, it is desirable that the operating point is set to a portion having the best linearity in the linear region of FIG. 2, that is, a linear portion and a linear portion is set to a long portion.

From the above, it is desirable that the operating point of the CMOS inverter, which is the bias point of the light receiving section, be near the center of the linear region in FIG. Furthermore, as can be seen from FIG. 2, when the operating point is near the center of the linear region, the gain (gain) G of the inverter is the largest, so that the performance of the inverter as the amplifying element is also the best. The gain G when the operating point is near the center of the linear region is shown in FIG.
(Vout / Vin).

The present inventor paid attention to the above, and as a method of setting the bias point of the light receiving section, as shown in FIG.
A method was adopted in which the input and output of the MOS inverter were short-circuited by the switch transistor Tr1. According to this method, the operating point (bias point) of the CMOS inverter is
As shown in FIG. 4, it is the intersection of the input / output characteristic curve of the CMOS inverter and the straight line of Vin = Vout.

The input / output characteristic curve of the CMOS inverter shown in FIG. 2 moves right and left in the figure depending on the values of the threshold voltage Vthp of the PMOS and the threshold voltage Vthn of the NMOS. However, if Vthp = Vthn, the operating point of the CMOS inverter is set near the center of the linear region of the input / output characteristics of the CMOS inverter as shown in FIG.

In the above description, the case where only one CMOS inverter is connected has been described. However, it is also possible to connect a plurality of CMOS inverters in series in order to increase the amplification factor. In this case, CMOS
Assuming that the amplification factor for one stage of the inverter is A, the total amplification factor when N stages are connected in series is A ^N , and a very large gain can be obtained.

Hereinafter, an embodiment of the present invention will be described with reference to the drawings. FIG. 1 is a configuration diagram showing an image sensor according to an embodiment of the present invention.
An example in which OS inverters are connected in two stages in series is shown.

This image sensor has n photodiodes PD1 to PDn as light receiving units. Note that n is a natural number. Each photodiode PD1 to PDn is 1
The outputs of the first-stage CMOS inverters are connected to the inputs of the second-stage CMOS inverters I21 to I2n. Each of the second-stage CMOS inverters I21 to I21
The output of I2n is connected to each of the switch transistors TR1 to TRn.
, And the gates of the switch transistors TR1 to TRn are connected to the outputs SR1 to SRn of the shift register 10. A / D
The converter is connected to a computer, and the computer is connected to a memory.

The input and output of each first-stage CMOS inverter are connected to the source and drain of each of the switch transistors SW1 to SWn, and the gate of each of the switch transistors SW1 to SWn is connected to an external terminal G1.
It is connected to the.

CMOS used for the image sensor of FIG.
The inverter has input / output characteristics shown in FIG. FIG. 4 shows that the input / output characteristic curve of the CMOS inverter and the straight line Vin =
The intersection of Vout is the operating point of the CMOS inverter of FIG. 1, and this operating point is Vin = 2.5V. CM at this time
The gain of the OS inverter is A = 23 times. In the image sensor shown in FIG. 1, two stages of CMOS inverters of the same size are connected in series, so that the total gain is A ² = about 530 times.

Next, the operation of the image sensor of FIG. 1 will be described. When resetting the image sensor, the switch transistors SW1 to SWn are turned on.

That is, the input to the external terminal G1 is VL
(= 0V), each of the switch transistors SW1 to SW
Since Wn is turned on, each of the CMOS inverters I11 to I11
The operating points of I1n and I21 to I2n become Vin = 2.5V, and the light receiving sections (photodiodes) PD1 to PDn are biased to 2.5V. At this time, the photocurrent (or dark current) generated in each light receiving unit is based on each switch transistor SW1 to SWn and each first stage CMOS inverter I
It is discharged to GND via 11-I1n. At this time, the output potentials of the second-stage CMOS inverters I21 to I2n remain at 2.5V.

Next, when reading the amplified voltage, the switch transistors SW1 to SWn are turned off. Then, photocharges accumulate in the light receiving units PD1 to PDn.

That is, the input to the external terminal G1 is VH
(= 5V), and each of the switch transistors SW1 to SW
When n is set to OFF, the photocharge generated in each light receiving unit has no place to be discharged, and is stored in the diffusion capacitance of each light receiving unit. Photocharge (signal component) generated at this time
Is Q and the diffusion capacitance of the light receiving unit is C, the potential of the light receiving unit is Q / C higher than when biased at 2.5V.
(V). This changed voltage Q /
C (V) is further converted to A by a two-stage CMOS inverter.
It is amplified ^twice. The amplified voltage can be sequentially read as a time-series signal by the switch transistors TR1 to TRn and the shift register 10.

According to the above embodiment, each light receiving section PD1
To the output signals of the CMOS inverters I11 and I11.
Since the light is amplified by 21 to I1n and I2n, weak light can be detected with high sensitivity. In addition, even when the capacity of the light receiving unit is large, reading can be performed at high speed. Further, since it is suitable for use of modern semiconductor integrated circuit technology, the sensor chip and peripheral circuits can be reduced in size, which contributes to cost reduction.

Further, by amplifying the output signal from the light receiving section by the inverter, the image sensor shown in FIG. 1 can be used as a sensor for ultraviolet rays having low photodiode sensitivity.

The time corresponding to the read time when reading an optical signal from the light receiving section is only the reset time of the light receiving section. Since the potential itself of the light receiving section has already appeared in the output Vout, there is no need for the one corresponding to the transfer time as described in the prior art. Further, when resetting, the reset time can be shortened. It is assumed that Vin has increased from the original reset voltage Vr by dv. Assume that the gain of the CMOS inverter is A (for example, 50). In the case of the present invention, when Vin = Vr, Vout = Vr (6), and when Vin = Vr + dv, Vout = Vr-A.dv (7). As can be seen from the above equation (7), the voltage Vx applied to the ON resistance of Tr1 (shown in FIG. 3) at the time of resetting
Is the potential difference between Vin and Vout, so that Vx = (A-1) .dv (8) Therefore, according to the present invention, the voltage applied to the resistor can be amplified, so that the potential difference can be amplified even when Vin and Vout are almost the same, and the reset time can be shortened.

In the above embodiment, two stages of CMOS inverters as amplifying elements are connected in series to each light receiving portion. However, three or more stages of CMOS inverters can be connected to further increase the amplification factor. It is. In this case, each switch transistor SW1 to determine the operating point
The source and the drain of SWn need to be connected so as to short-circuit the input of the first-stage CMOS inverter and the output of any one of the odd-numbered CMOS inverters. The reason is as follows.

In this connection, the first-stage CMO
When the input potential of the S inverter increases, the odd-numbered CM
Since the output potential of the OS inverter becomes low, when the switch transistors SW1 to SWn are turned on, the first stage C
It is automatically adjusted so that the input potential of the MOS inverter is equal to the output potential of the odd-numbered CMOS inverter. However, the switch transistors SW1 to SWn are connected to the input of the first-stage CMOS inverter and the CMO of the even-numbered stage.
When the output of the S-inverter is short-circuited and the switch transistors SW1 to SWn are turned on and the input potential of the first-stage CMOS inverter increases, the output potential of the even-numbered CMOS inverter also increases. Therefore, the automatic adjustment operation as described above is not exhibited.

It should be noted here that when a CMOS inverter short-circuited by a switch transistor is composed of multiple stages and the overall gain is extremely high, oscillation is caused by feedback via the switch transistor. There is a point. As a countermeasure, when the CMOS inverter is composed of N stages (N is 3 or more), the switch transistor is connected to the first stage C
There is a method of short-circuiting between the input and output of a MOS inverter. In this case, even if the gain of the entire N-stage CMOS inverter is A ^N times, the gain at the time of feedback is only A times, so that oscillation can be prevented.

Further, in the above embodiment, the shift register 10 is used to read out the output signals from the n light receiving units from one output line, but the n output signals from each light receiving unit are used. Is read using n output lines, there is no need to use a shift register.

The present invention can be applied to both line sensors and area sensors.

Next, a method of using the image sensor (sensor chip) of FIG. 1 for detecting an alignment (position adjustment) signal will be described.

The alignment is performed by obtaining an intensity distribution of light which changes depending on a difference in position. To detect a weak alignment signal, it is necessary to increase the gain of a CMOS inverter (amplifier). However, if an inverter having a large gain is manufactured on the same chip, it is difficult to keep the gain of each inverter constant. Therefore, the gain of each inverter may vary.

As a countermeasure, variations are obtained by once irradiating each light receiving portion (all pixels) with light of uniform intensity and reading the amplified output signal, and obtaining the gain of the inverter of each light receiving portion in advance. There is a method of correcting.

Hereinafter, a method of correcting the variation will be specifically described. The total gain of the two-stage CMOS inverter connected to each of the light receiving units PD1 to PDn is An (n = n
Natural number).

First, a uniform light having a wavelength λ and an intensity I is applied to each light receiving portion PDn (n = natural number). At this time, the output voltage OUTn (n = natural number) of each light receiving unit is measured. Assuming that a uniform charge Q is generated in each light receiving portion by the uniform light, 2
Output voltage A amplified by the CMOS inverter
OUTn (n = natural number) can be expressed as in the following equation (2). AOUTn = Q / C · An (C: diffusion capacity of light receiving unit) (2) Here, since the light irradiated to each light receiving unit is uniform light,
Is constant, and C can be regarded as constant if the shape of each light receiving unit is the same.

When the total gain A1 of the light receiving unit PD1 is used as a reference, the coefficients B2 to Bn for obtaining the variation of the total gain A2 to An of the other light receiving units PD2 to PDn with respect to this are
It can be expressed as in the following equation (3). Bn = AOUTn / AOUT1 (n is a natural number of 2 or more) (3)

From the equations (2) and (3), the following equation (4) is obtained.
Holds. Bn = An / A1 (n is a natural number of 2 or more) (4)

Therefore, the output voltage AOU amplified by each CMOS inverter when each light receiving section is irradiated with uniform light.
Measure Tn (n = natural number) and calculate coefficient Bn (n = natural number)
, The total gain A of each CMOS inverter
n variation can be known. Then, even if the total gain of the CMOS inverter varies, light having a certain intensity distribution is applied to the light receiving unit using the A / D converter, computer and memory of FIG. 1 using this coefficient Bn. The output voltage OUTn ′ (n = natural number) amplified by each CMOS inverter at the time of the correction is corrected as in the following equation (5). Thereby, the accurate intensity distribution Cn of the emitted light can be known. Therefore, variation in signal detection due to variation in sensitivity of each pixel can be suppressed. Cn = OUTn '/ Bn (5)

Depending on the wavelength of light to be detected, it may be difficult to irradiate uniform light over the entire light receiving area. In such a case, light of another wavelength capable of easily producing a uniform light is irradiated and the output voltage AOU described above is applied.
Tn (n = natural number) may be measured. In this case, assuming that the photoelectric conversion coefficient of the light receiving unit for the light of wavelength λ is K,
= K · I (photocharge Q, intensity I). Therefore, the photocharge Q in the above equation (1) differs depending on the wavelength λ. However, when the entire light receiving area is irradiated with uniform light, the photocharge Q generated in each light receiving section remains constant. Since the required An is constant irrespective of the wavelength, it is only necessary to irradiate light having a wavelength at which uniform light can be easily obtained to determine the variation of An.

Further, the wavelength of the uniform light may have a width. For example, a halogen lamp may be used as a light source.

[0048]

As described above, according to the present invention, the electric signal output from the light receiving section is amplified by the CMOS inverter. Therefore, it is possible to provide an image sensor capable of detecting weak light with high sensitivity and a correction method thereof. Further, according to the present invention, it is possible to provide an image sensor capable of reading out a signal from a light receiving section at a high speed.

[Brief description of the drawings]

FIG. 1 is a configuration diagram illustrating an image sensor according to an embodiment of the present invention.

FIG. 2 is a graph showing input / output characteristics of a CMOS inverter.

FIG. 3 is a circuit diagram illustrating a method for setting a bias point of a light receiving unit.

FIG. 4 is a graph illustrating an operating point of the CMOS inverter shown in FIG. 3;

[Explanation of symbols]

Reference Signs List 10 shift register Tr1 switch transistor Vin input voltage Vout output voltage PD1 to PDn photodiode I11 to I1n first-stage CMOS inverter I21 to I2n second-stage CMOS inverter TR1 to TRn switch transistor SR1 SRn: Shift register output SW1 to SWn: Switch transistor G1: External terminal

────────────────────────────────────────────────── ───

[Procedure amendment]

[Submission date] March 20, 1998

[Procedure amendment 1]

[Document name to be amended] Statement

[Correction target item name] Brief description of drawings

[Correction method] Change

[Correction contents]

[Brief description of the drawings]

FIG. 1 is a configuration diagram illustrating an image sensor according to an embodiment of the present invention.

FIG. 2 is a graph showing input / output characteristics of a CMOS inverter.

FIG. 3 is a circuit diagram illustrating a method for setting a bias point of a light receiving unit.

FIG. 4 is a graph illustrating an operating point of the CMOS inverter shown in FIG. 3;

FIG. 5 is a circuit diagram illustrating a reading method using a conventional image sensor.

[Description of Symbols] 10 shift register Tr1 switch transistor Vin input voltage Vout output voltage PD1 to PDn photodiode I11 to I1n first-stage CMOS inverter I21 to I2n second-stage CMOS inverter TR1 to TRn ... switch transistors SR1 to SRn ... shift register outputs SW1 to SWn ... switch transistors G1 ... external terminals

Claims (5)

[Claims] 1. A light receiving unit for converting an optical signal into an electric signal, a CMOS inverter connected to the light receiving unit, and a switch transistor for connecting / cutting off an input and an output of the CMOS inverter, The electric signal output from the CMO
An image sensor characterized by being amplified by an S inverter. 2. A light receiving unit for converting an optical signal into an electric signal, a plurality of CMOS inverters connected in series to the light receiving unit, an input of a first CMOS inverter and an odd-numbered CMOS
A switch transistor for connecting / disconnecting an output of the S inverter; and an amplifier for amplifying an electric signal output from the light receiving unit by the plurality of CMOS inverters. 3. A plurality of the light receiving sections, a transistor is connected to an output of the CMOS inverter, an output of a shift register is connected to a gate of the transistor, and an output from the light receiving section is generated by driving the shift register. The image sensor according to claim 1, wherein the read electric signal is read. 4. A photodetector comprising: a plurality of light receiving units for converting an optical signal into an electric signal; a CMOS inverter connected to each of the light receiving units; and a switch transistor for connecting / disconnecting an input and an output of the CMOS inverter. A correction method for correcting a difference in amplification factor of each light receiving unit in an image sensor that amplifies an electric signal output from the light receiving unit by the CMOS inverter; a method of irradiating each light receiving unit with light of uniform intensity. A method of correcting an image sensor, comprising: measuring an amplification factor of each electric signal amplified by a CMOS inverter in advance, and performing a correction for equalizing each amplification factor. 5. The method according to claim 4, wherein said plurality of CMOS inverters are connected in series.