JP2006196742A - Imaging apparatus and drive method thereof - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide an imaging apparatus which can enlarge a dynamic range of light input while reducing influence of manufacturing dispersion of a photoelectric converter and a transfer transistor, and to provide its control method. <P>SOLUTION: The solid state imaging apparatus is constituted by disposing a unit picture element consisting of a photoelectric converter 401 for accumulating signal charge generated by incident light, a transfer 405 for transferring signal charge, a floating diffuser 403 whereto signal charge is transferred and an amplifier 404 for outputting amplified signal based on signal charge transferred to the floating diffuser in a plurality of lines. Each picture element separately has a charge retainer 402 which accumulates all the charge from immediately after exposure start to a certain time T in one exposure period. <P>COPYRIGHT: (C)2006,JPO&NCIPI

Description

  The present invention relates to an imaging apparatus that captures an image based on charges obtained by photoelectric conversion of incident light, and a driving method of the imaging apparatus, and in particular, an imaging apparatus that captures an image with a wide input dynamic range and the imaging The present invention relates to a method for driving the apparatus.

  Conventionally, as a technique for taking an image with a wide input dynamic range using a solid-state imaging device, for example, Patent Documents 1 and 2 can be cited.

  In Patent Document 1, as shown in FIG. 1, a charge overflowing from a photodiode 101 to a floating diffusion 103 is transferred via a transfer transistor 102, and the overflowed charge is read before reading a signal from the photodiode. Thus, a photoelectric conversion result at high luminance is obtained without saturation of the pixel signal. A graph of the photoelectric conversion result obtained is shown in FIG. 2, where 201 is a photoelectric conversion characteristic of a normal photodiode, and since the photodiode is saturated after the light amount of 202, the photoelectric conversion characteristic is also clipped to a constant value. . However, the electric charge overflowing at that time reaches the floating diffusion, and the characteristics of photoelectric conversion start to rise as indicated by 203. By appropriately converting the curve 203, a combined photoelectric conversion characteristic such as 204 having a large input dynamic range is obtained together with the straight line 201.

In Patent Document 2, a variable capacity of 201 is provided in part of the capacity of the floating diffusion, the capacity of the floating diffusion is changed in accordance with the brightness, saturation in the floating diffusion is prevented, and dynamic range expansion is realized. Yes.
US Pat. No. 6,307,195 JP 2000-1657545 A

  However, there are various problems in the technology for preventing saturation on the high luminance side, as typified by the above two technologies.

  First, regarding Patent Document 1, three problems will be taken up here. First, the saturation charge amount should ideally be determined uniquely for a photodiode, but the saturation charge amount varies in a certain range due to statistical variations in the manufacturing process. The influence will be described with reference to FIG. When the saturation charge amount varies, the overflow point is different, and the start point of 203 varies. Therefore, the combined photoelectric conversion characteristics vary as 301 in each pixel.

  Second, when a transfer transistor is provided between the photoelectric conversion unit and the floating diffusion, the threshold value of the transfer transistor varies statistically in the manufacturing process. If the threshold value varies, that is, the channel potential of the path through which the overflowing charge passes differs, and this further deteriorates the degree of variation 301 in the combined photoelectric conversion characteristics.

  Third, when the amplifying unit is provided in the pixel, the photoelectric conversion unit is connected to the input unit of the amplifying unit in the pixel, so that it is impossible to make the floating diffusion unit completely embedded. is there. Here, the buried structure is a structure in which an impurity diffusion region of a reverse conductivity type is formed on the surface of the impurity diffusion region for forming the floating diffusion portion. According to such a configuration, the floating diffusion portion is Generation of dark current on the surface of the impurity region for formation is reduced. Therefore, the reverse leakage current of the pn diode in the diffusion portion that forms the floating diffusion region is larger than that in the photoelectric conversion portion, and is therefore not suitable as a node for holding data. Since the charges overflowing in the floating diffusion portion are lost with time, a high S / N of the signal in the high luminance portion cannot be expected.

  Next, Patent Document 2 will be described. In this method, the number of saturated charges in the floating diffusion portion increases, but the number of saturated charges in the photoelectric conversion portion is not different from the conventional one. Therefore, it is effective in terms of expanding the input dynamic range after the floating diffusion portion, but further improvement is expected in terms of expanding the dynamic range of the optical input.

  In view of the above problems, an object of the present invention is to provide an image pickup device that can reduce the influence of manufacturing variations of photodiodes and transfer transistors and expand the dynamic range of light input, and a control method thereof.

  In view of the above problems, an imaging apparatus according to the present invention includes a photoelectric conversion unit that accumulates charges generated by incident light, a transfer unit that transfers the charges, and a floating diffusion unit that transfers the charges. , An imaging device configured by arranging unit pixels having an amplification unit that amplifies the charge transferred to the floating diffusion unit in a matrix, wherein at least one of the pixels is in one exposure period, In addition to the floating diffusion part, the first holding part for storing the electric charge photoelectrically converted from the start of exposure to a predetermined time T is provided.

The solid-state imaging device control method according to the present invention includes a photoelectric conversion unit that accumulates charges generated by incident light, a transfer unit that transfers the charges, and a floating diffusion unit that transfers the charges. A control method for an imaging apparatus configured by arranging unit pixels having an amplifying unit for amplifying the charge transferred to the floating diffusion unit in a matrix,
During one exposure period, the first photoelectric conversion result from the start of exposure to a predetermined time T is held in the charge holding unit, and then the second photoelectric conversion result from the predetermined time T to the end of exposure And outputting the first photoelectric conversion result and the second photoelectric conversion result.

  According to the present invention, it is possible to realize a solid-state imaging device in which the influence of manufacturing variations of the photoelectric conversion unit and the transfer transistor is reduced and the dynamic range of light input is expanded.

(Example 1)
A first embodiment of the present invention will be described. FIG. 4 is a diagram showing the constituent requirements of the pixel of the present invention. 401 is a photodiode serving as a photoelectric conversion unit that converts incident light into electric charge, 402 is a charge holding unit that accumulates electric charge from the start of exposure to a predetermined time T, 403 is a floating diffusion unit, and 404 is a voltage of the floating diffusion unit. Amplifiers 405, 406, and 407 for amplifying are switches that control the flow of charges. Reference numeral 408 denotes a switch for resetting the floating diffusion portion, the photodiode portion, and the charge holding portion. The unit pixels are arranged in a matrix, and pixels in the same column are connected to a common vertical output line 410 via a selection switch 409.

  Next, the operation of this apparatus will be described with reference to FIG. FIG. 5 shows the operation of each switch and the change in the photoelectric conversion result when focusing on the accumulation time of a certain pixel. Before time 501, all the switches 405, 406, and 408 are short-circuited (on), and the photodiode, the charge holding unit, and the floating diffusion are set to the reset state. Here, in the reset state, the pixel is designed so that the charge of the photodiode and the charge holding portion is completely discharged. Such a design is possible if the impurity concentration in the semiconductor substrate is set to such a value that the channel potential of the photodiode and the charge holding portion is higher than that of the floating diffusion at the time of reset.

  After that, by closing (turning off) the switches 405 and 406 at time 502, the charge of the photoelectric conversion result is held by the photodiode. Thereafter, after time T has elapsed, at time 503, the switch 407 is opened (on), and all charges in the photodiode are transferred to the charge holding portion. This transfer is possible if the channel potential of the photodiode is designed to be higher than that of the floating diffusion.

  After the switch 407 is closed, there is no charge inflow to the charge holding unit and no charge outflow from the holding unit, and accumulation is continued (started again) in the photodiode unit. After that, at time 504, first, the switch 406 is opened to transfer the charge in the charge holding portion to the floating diffusion, and after the charge-to-voltage conversion, the voltage value is read as a charge signal, and then the floating diffusion portion is turned on by the reset switch. After resetting, the charge of the photodiode portion is read out as a second signal at time 505. Here, it is assumed that “charge” is a negative charge. Needless to say, for positive charges, the magnitude relationship of channel potentials may be changed as appropriate. Here, the storage period is defined as 505 until the charge of the photodiode is read. This accumulation period may be separately defined by a mechanical shutter or the like.

  Next, there are two types of cases where the amount of light incident on the photodiode is small and the photodiode does not saturate during one accumulation period, and when the amount of light is large and the photodiode saturates during one accumulation period. Will be described as 506 and 507 in FIG.

  Reference numeral 506 denotes an example where the amount of incident light is small. As indicated at 508, at time 503, the charge of the photodiode is transferred to the charge holding portion, but the value 508 is insignificant. Thereafter, the photodiode continues normal accumulation. At time 504, the charge in the charge holding portion is converted into a voltage by floating diffusion, but the value is insignificant and does not have a high S / N. On the other hand, the photodiode continues to accumulate without saturating until time 505, and the value is converted again to the voltage 509 by the floating diffusion. In this case, it becomes possible to obtain a photoelectric conversion result having a high S / N by preferentially adopting 509 as a photoelectric conversion result in a later photographing system.

  On the other hand, reference numeral 507 denotes a case where the light quantity is such that the photodiode is saturated. Since the amount of light is large as in 511, a sufficiently large amount of charge can be obtained even in a short time T. On the other hand, since the amount of light is large, the photodiode is saturated at time 512, and the information on the photoelectric conversion result in the subsequent photodiodes is significantly degraded. As in the case where the amount of incident light 506 is small, the charge of the charge holding unit and the charge of the photodiode are read as voltages 513 and 514, respectively, but the signal output from the photodiode is saturated and the value is not reliable. It has become. On the other hand, the voltage of the charge holding unit is not saturated. In this case, the signal 513 is preferentially adopted as the photoelectric conversion result in the later photographing system, and the ratio of the total accumulation period (502 to 505) and the accumulation time (502 to 503) between the time T is used. By converting the signal 514, it is possible to obtain a photoelectric conversion result under a high luminance condition that saturates the photodiode.

  The photoelectric conversion characteristics obtained by such an operation are shown in FIG. The major difference is that the signal 515 on the high luminance side has linearity from the time when the amount of light is zero to the high luminance side where the photodiode is saturated unlike the conventional example. In addition, when all charges are transferred to the charge holding portion, this inclination is determined only by the inclination when the photoelectric conversion characteristic of only the photodiode is not saturated and the ratio of the time T during the accumulation time. It does not depend on manufacturing variations of various devices such as saturation charge. Therefore, even in an actual device in which various manufacturing variations can occur, the synthesized characteristic 516 can be obtained with good stability.

  Here, it is assumed that the photoelectric conversion results are obtained directly as 508, 509, 512, and 513, but the reset is fixed after the floating diffusion is reset as used in, for example, a conventional CMOS sensor or CCD. Pattern noise, thermal noise, etc. may be stored, a pixel signal may be read so as to be superimposed on the value, and two correlated signals may be subtracted to perform correlated double sampling. By doing so, the S / N of the signal can be further increased. However, it goes without saying that the effect of extending the input dynamic range of the present invention can be obtained without doing this.

  In addition, the charge holding portion is preferably composed of a capacitor such as an oxide film or a PN diode. At that time, when a PN diode is used, it is desirable to have a buried structure that can eliminate the influence of surface defects. Since the charge holding portion also needs to hold the charge for a long time, the embedded structure can reduce the leakage current when the PN junction is reverse biased, and can hold the charge with high S / N. realizable.

  Here, the switch 407 may be driven such that it is always turned on until time T and then turned off after time T, as indicated by reference numeral 601 in FIG. In this case, the design of the switch such that the relationship of the channel potential of the photodiode <the channel potential of the path of the switch 407 <the channel potential of the charge holding portion is satisfied, so that the charge of the photodiode is quickly transferred to the charge holding portion, And driving is desirable.

  Next, FIG. 7 shows an example of a layout for realizing the pixel shown in FIG. Numbers corresponding to those in FIG. 4 are assigned. Here, it is assumed that each device is made in a P-type semiconductor substrate, the photodiode is a buried photodiode, the charge holding portion is also a buried diode, the floating diffusion is an n + diffusion layer, and the switches 405, 406, 407, 408, and 409 are The MOS transistor / amplifier 404 is configured as a source follower composed of one MOS, and only the upper part of the photodiode is opened in the shape of 701 and the others are shielded from light. Needless to say, this is an example of a pixel for obtaining the effect of the present invention, and the present invention is not limited to this. For example, it may be configured in an N-type substrate so that minority carriers accumulated in the photodiode become holes.

(Example 2)
Next, a second embodiment will be described with reference to FIG. FIG. 8 is a circuit diagram example different from the first embodiment. Here, a charge holding unit is provided between the photodiode and the floating diffusion via a switch. Switches 801 and 802 are used to control the connection between the photodiode, the floating diffusion, and the charge holding unit. FIG. 9 shows an example of the layout. Parts having the same functions and roles as those in FIGS. 4, 7, and 8 are given the same numbers.

  FIG. 10 is a waveform diagram showing the operation timing of each switch and the state of the photodiode, the charge holding unit, and the floating diffusion in this embodiment. Points equivalent to those in FIG. 5 are given the same numbers. The operation of this pixel will be described with reference to FIG. When saturated or not saturated, there is no difference between the first embodiment and the description. What changes is the way the switch is driven. By turning on only 802 while keeping 801 off, only the charge in the charge holding portion can be read out while keeping the charge in the photodiode as it is. After all the charges in the charge holding portion are swept out, the floating diffusion is reset, and then both 801 and 802 are turned on to read out the charge from the photodiode to the floating diffusion. At this time, since the charge in the charge holding portion is already negligibly small, the second read result is composed of only the components of the photodiode. The new effect of the present embodiment is that the charge holding portion is provided between the photodiode and the floating diffusion, so that in the first embodiment, three photodiodes required to connect the photodiode, the floating diffusion, and the charge holding portion. This is because the number of switches is reduced to two, and the pixel size can be reduced.

(Example 3)
A third embodiment will be described. FIG. 11 shows (1) a photodiode, (2) a transistor between the photodiode and the charge holding unit, and (3) a charge holding unit on the line AA ′ of FIG. 9 in the configuration of the second embodiment. (4) Charge holding part-floating diffusion, (5) An example showing a channel potential of a series of floating diffusions. Here, charge is considered as electrons. As shown in FIG. 11A, as described in the first embodiment, in order to transfer all charges up to time T to the charge holding portion, the channel potential of the first holding portion is set to be equal to that of the photodiode. At the same time, it is necessary to set the channel potential higher than the channel potential. However, in order to improve the S / N, when the charge is read out to the floating diffusion, the channel potential of the charge holding portion is floating as shown in FIG. It is preferable to perform the read operation so that all charges are read out to the floating diffusion by lowering the channel potential of the diffusion.

  However, the channel potential of the charge holding portion that can satisfy the two conditions is usually limited, and the number of charges that can be held in the charge holding portion is naturally limited. Therefore, if the charge as shown in FIG. 11B flows in until the charge holding portion is saturated until time T, the condition that the charge holding portion reads all charges collapses. / N worsens. Further, as shown in FIG. 11C, even if all the charges are read out to the floating diffusion, it becomes difficult to store the charges in the charge holding portion to some extent.

  According to the present invention, as shown in FIG. 12, the channel potential of the charge holding portion is changed from when the charge flows in as shown in FIG. 12A and from the charge holding portion to the floating diffusion as shown in FIG. Is dynamically changed according to the time when data is read out, and even when there is a large amount of charge during time T, the saturation charge amount of the charge holding unit is secured and the read conditions for improving the S / N are also realized. Is possible.

  FIG. 13 is an example of a pixel layout in which the channel potential of the charge holding portion can be changed. FIG. 13A is a top view, which is different from the layout shown in FIG. 9 in that a gate 1301 is further provided on the charge holding portion. FIG. 13B is a cross-sectional view taken along line A-A ′ shown in FIG. 13A, and a new gate 1301 is formed using, for example, a second layer of polysilicon. Of course, a technique for realizing an equivalent function using only the polysilicon of the same layer may be used while devising the impurity concentration of the semiconductor layer, which is partially used in the conventional CCD. The effect of the present invention can be obtained by dynamically changing the channel potential, and various layer structures can be selected.

(Example 4)
A fourth embodiment will be described. FIG. 14 is a diagram illustrating a readout circuit of the present invention. Each column has a memory 1401 for storing pixel signals, and the analog memory is sequentially operated by the shift register 1402 and simultaneously outputted to the output terminal 1403 while simultaneously holding the signal output of each row. At that time, each column has a comparator 1404 for determining a signal level and a 1-bit digital memory 1405 for holding magnitude discrimination data as a result of the comparison. A pixel signal is input to one of the comparators, and a signal output Vsat obtained when the photodiode is saturated is supplied to the other terminal 1406. The comparison is controlled by a comparison instruction signal 1407. A terminal 1408 is an output terminal of the digital memory. Reference numeral 1409 denotes a sample hold switch for controlling the sample hold to the memory, and 1410 denotes an instruction signal for controlling the sample hold.

  Next, the operation of this circuit will be described using the timing chart of FIG. 15 and an output waveform example. FIG. 15A shows an example when the signal from the photodiode is not saturated. 1501 is an example of a signal output from a pixel appearing in a column. Since the photodiode is not saturated, it is better to read the signal from the photodiode in this case. The operation in such a case is as follows.

  First, at timing 1502, a signal from the charge holding portion appears in the column. At timing 1503, the sample hold signal 1410 is set in the hold state, and the signal from the charge holding unit is held in the memory. Next, at timing 1504, a signal from the photodiode appears in the column. The comparator compares the value with Vsat. Here, the signal of the photodiode is not saturated, and the comparison result is Hi here. In addition, when it is saturated, it is designed to output Lo. Sample and hold are performed at timings 1505 and 1506, respectively, and the value of the memory is replaced with the value of the photodiode at this point. In parallel with this, the Hi level of the comparison result is taken into the digital memory provided. With this operation, the memory stores the signal from the photodiode, and the digital memory stores the flag information indicating that the signal from the photodiode has been captured. Thereafter, these data are scanned and output to the outside.

  On the other hand, FIG. 15B shows an example where the signal from the photodiode is saturated. The same parts as those in FIG. 15A are given the same numbers. The major difference is the operation of the comparator after holding the charge retention signal. The comparator outputs a Lo level signal that the photodiode is saturated. Therefore, the operation of replacing the signal of the memory with the signal of the photodiode does not occur, and the flag information that the signal from the charge holding unit is finally taken into the memory and the signal from the charge holding unit is taken into the digital memory. Is stored. Thereafter, these data are scanned and output to the outside.

  The effects of the present embodiment are as follows. Conventionally, a signal at a high luminance and a signal at a low luminance have to be read twice. At that time, it took a long time to perform a scanning operation for reading data from the line memory to the outside. In the present invention, readout from the pixel is required twice, but scanning is only required once, so that it is possible to shorten the time for reading out one frame as compared with the conventional case.

  Here, detailed explanations of the photodiode signal and the charge holding portion signal are omitted. For example, they may be a value obtained by directly converting the photoelectric charge into a voltage, or may handle a signal with an increased S / N, excluding reset noise and fixed pattern noise. It goes without saying that the present invention can achieve an effect regardless of the characteristics of these signals.

(Application example to imaging system)
FIG. 16 illustrates an example of a circuit block in the case where the solid-state imaging device according to the first to fourth embodiments is applied to a camera. A shutter 1001 is provided in front of the taking lens 1002 and controls exposure. The amount of light is controlled by the diaphragm 1003 as necessary, and an image is formed on the solid-state imaging device 1004. A signal output from the solid-state imaging device 1004 is processed by a signal processing circuit 1005 and converted from an analog signal to a digital signal by an A / D converter 1006. The output digital signal is further processed by a signal processing unit 1007. The processed digital signal is stored in the memory 1010 or sent to an external device through the external I / F 1013. The solid-state imaging device 1004, the imaging signal processing circuit 1005, the A / D converter 1006, and the signal processing unit 1007 are controlled by a timing generation unit 1008, and the entire system is controlled by an overall control unit / arithmetic unit 1009. In order to record an image on the recording medium 1012, the output digital signal is recorded through a recording medium control I / F unit 1011 controlled by the overall control unit / arithmetic unit.

(A) It is drawing for demonstrating the example of the conventional pixel structure. (B) It is a figure for demonstrating the relationship between incident light quantity and a tax output. It is drawing for demonstrating the equivalent circuit schematic of the conventional pixel. It is drawing for demonstrating the subject of the prior art example. It is drawing for demonstrating the equivalent circuit of the apparatus of a 1st Example. 5 is a diagram for explaining an operation example of the circuit shown in FIG. 4. 6 is a diagram for explaining another operation example of the example shown in FIG. 5. It is drawing for demonstrating the top view of the pixel of a 1st Example. It is drawing for demonstrating the equivalent circuit of the pixel of a 2nd Example. It is drawing for demonstrating the top view of the pixel of a 2nd Example. It is drawing for demonstrating the operation example of a 2nd Example. (A)-(c) It is drawing for demonstrating the difference with the prior art of a 3rd Example. (A) It is drawing for demonstrating the potential state (at the time of charge transfer to a charge holding part) of a 3rd Example. (B) It is drawing for demonstrating the potential state (at the time of charge accumulation) of the 3rd Example. (A) It is drawing for demonstrating the top view of the pixel of a 3rd Example. (B) It is drawing for demonstrating sectional drawing of the pixel of a 3rd Example. It is drawing for demonstrating the column circuit of a 4th Example. It is drawing for demonstrating the operation example of a 4th Example. It is drawing for demonstrating the application example to an imaging system.

Claims (9)

A photoelectric conversion unit that accumulates charges generated by incident light, a transfer unit for transferring the charges, a floating diffusion unit to which the charges are transferred, and an amplification that amplifies the charges transferred to the floating diffusion unit An image pickup apparatus configured by arranging unit pixels having a portion in a matrix,
At least one of the pixels has an electric charge holding unit that accumulates electric charge photoelectrically converted from the start of exposure until a predetermined time T during one exposure period, separately from the floating diffusion unit.   The imaging apparatus according to claim 1, wherein the charge holding unit is disposed between the photoelectric conversion unit and the floating diffusion unit.   The imaging apparatus according to claim 1, wherein the charge holding unit is provided on a semiconductor substrate and includes a control unit that changes a channel potential of the charge holding unit.   2. Each column includes an analog memory for holding a pixel signal, a comparator for comparing the pixel signal and a first reference voltage for each column, and a digital memory for holding a comparison result. 4. The imaging device according to any one of items 1 to 3. A photoelectric conversion unit that accumulates charges generated by incident light, a transfer unit for transferring the charges, a floating diffusion unit to which the charges are transferred, and an amplification that amplifies the charges transferred to the floating diffusion unit A control method of an imaging device configured by arranging unit pixels having a portion in a matrix,
During one exposure period, the first photoelectric conversion result from the start of exposure to a predetermined time T is held in the charge holding unit, and then the second photoelectric conversion result from the predetermined time T to the end of exposure And outputting the first photoelectric conversion result and the second photoelectric conversion result.   The method for controlling an imaging apparatus according to claim 5, wherein the second photoelectric conversion result is read after reading the first photoelectric conversion result.   6. The first photoelectric conversion result is stored in a charge holding portion provided in a semiconductor substrate, and the channel potential of the charge holding portion is controlled to a binary value or more during one exposure period. 6. A method for controlling the imaging apparatus according to 6.   Each column has an analog memory for holding a signal from a pixel, and after holding the first photoelectric conversion result in the analog memory, the second photoelectric conversion result is read again as a signal from the pixel, and its value The method of controlling an imaging apparatus according to claim 5, wherein the value of the analog memory is rewritten according to 5. The image pickup apparatus according to claim 1, an optical system that forms an image of light on the image pickup apparatus, and a signal processing circuit that processes an output signal from the image pickup apparatus. An imaging system.

Images