JP2008048405A - Photoelectric conversion apparatus, x-ray image pickup apparatus, and system having the apparatus - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide an apparatus which is excellent in read-out speed, high S/N, high tone level, and low cost for a photoelectric conversion apparatus. <P>SOLUTION: A photoelectric conversion apparatus has a photoelectric conversion circuit unit comprising a plurality of photoelectric conversion elements, switching elements, matrix signal wires, and gate drive wires arranged on the same substrate in order to output parallel signals, a driving circuit unit for applying a driving signal to the gate drive wire, and a read-out circuit unit for converting the parallel signals transferred through the matrix signal wires into serial signals to output them. The read-out circuit unit comprises at least one or more analog operational amplifiers connected with each of the matrix signal wires, transfer switches for transferring output signals from the respective matrix signal wires outputted through each amplifier, read-out capacitors, and read-out switches for successively reading out signals from the read-out capacitors as serial signals. <P>COPYRIGHT: (C)2008,JPO&INPIT

Description

  The present invention relates to a photoelectric conversion apparatus and a photoelectric conversion system having the apparatus, and more specifically, office equipment such as an X-ray detector, a digital copying machine, an electronic blackboard, and a facsimile machine for nondestructive inspection such as medical treatment and internal inspection. The present invention relates to a photoelectric conversion device that can be applied as an image input unit and the like, and a photoelectric conversion system having the device.

  Currently, in an X-ray imaging apparatus used for medical diagnosis, a so-called film that exposes X-rays to a human body, irradiates a fluorescent substance that converts X-rays transmitted through the human body into visible light, and exposes the fluorescence to a film. The method has become mainstream.

  However, not only in Japan, which is facing an aging society, but also worldwide, there is a strong demand for improved diagnostic efficiency in hospitals and more accurate medical devices. Under such circumstances, in a conventional film type X-ray imaging apparatus, it takes a long time to obtain a patient's X-ray image because there is a film development process on the way, and sometimes If the patient moves during X-ray imaging or if the exposure does not match, the imaging must be performed again. These are factors that hinder the improvement of the efficiency of medical treatment in hospitals, and the burden on patients is large, which becomes a major obstacle when aiming for a new medical society in the future.

  In recent years, the demand for “digitalization of X-ray image information” is increasing in the medical industry. If digitalization is achieved, the doctor can know the patient's X-ray image information at an optimal angle in real time, and the obtained X-ray image information is recorded and managed using a medium such as a magneto-optical disk. can do. If a facsimile or other communication method is used, the patient's X-ray image information can be sent to any hospital in the world in a short time.

  Further, even in non-destructive inspection represented by inspection inside an object such as a building frame, installation of various devices for X-ray imaging and imaging of necessary parts cannot be performed again and again. However, in the case of the film system, even in such an inspection, it is not known whether imaging of a necessary part is completed until development is completed. Further, since the expert's judgment is made after seeing the film developed, it is impossible to give an instruction for photographing or a treatment from another angle as needed.

  Accordingly, there is a high demand for X-ray image information of a desired part in real time even in such a field.

  Therefore, recently, an X-ray imaging apparatus using a CCD solid-state imaging element or an amorphous silicon photoelectric conversion element instead of a film has been proposed in order to meet the demand for “digitization of X-ray image information”.

  FIG. 1 is an equivalent circuit diagram of an example of a two-dimensional photoelectric conversion apparatus applicable to such an X-ray imaging apparatus. Although a 3 × 3 two-dimensional photoelectric conversion device is shown in FIG. 1 for the sake of simplicity of explanation, an actual photoelectric conversion device is composed of a larger number of bits depending on the purpose of the device.

In FIG. 1, T _1-1 , T _1-2 , T _1-3 , T _2-1 ,..., T _3-3 are switching elements, S _1-1 , S _1-2 , S _1-3 , respectively. S _2-1 ,..., S _3-3 are photoelectric conversion elements, SR 1 is a shift register, SR 2 is a shift register, G 1, G 2 and G 3 are gate drive wirings, M 1, M 2 and M are signal wirings, C 1 C2 and C3 are read capacitors, RES1, RES2 and RES3 are reset switches, CRES is a voltage pulse input unit for reset, OP is an operational amplifier, Ca is a storage capacitor (for example, an equivalent additional capacitor added to the wiring), U1, U2 and U3 are read switching elements, N1, N2 and N3 are gate drive wirings for the switching elements U1 to U3, 1 is a photoelectric conversion circuit section, and 2 is a read circuit section.

In Figure 1, the light hν which is incident on the photoelectric conversion element S1-1~S3-3 is photoelectrically converted by the photoelectric conversion elements S _1-1 to S _3-3, as a photoelectric conversion signal charges, each of the photoelectric conversion elements S _1-1 is accumulated in inter-electrode capacity to S _3-3. These photoelectric conversion signals pass through the transfer switches T1-1 to T3-3 and the signal wirings M1 to M3 and become parallel voltage outputs. Further, it is converted into a serial signal by the read switch circuit unit and taken out to the outside.

  In the configuration example of the photoelectric conversion device in FIG. 1, photoelectric conversion elements having a total number of pixels of 9 bits are grouped in 3 bits and divided into 3 rows. The above-described operations are sequentially performed in units of rows.

  FIG. 2 is a timing chart showing an example of the operation of the photoelectric conversion apparatus shown in FIG.

The optical information (hν) incident on the photoelectric conversion elements S _{1-1 to} S _3-3 in the first row is photoelectrically converted and used as signal charges between the electrodes in each of the photoelectric conversion elements S1-1 to S1-3. Accumulated in capacity. After a certain accumulation time has elapsed, a first voltage pulse for transfer is applied to the gate drive wiring G1 from the shift register SR1 for a time T1, and the transfer switch elements T1-1 to T1-3 are turned on. As a result, the signal charges stored in the interelectrode capacitances (S1-1 to S1-3) in the photoelectric conversion elements S1-1 to S1-3 pass through the signal wirings M1 to M3, respectively, and load capacitors C1. To C3, and the potentials V1 to V3 of the load capacitors C1 to C3 are increased by the amount of charge of the signal (transfer operation).

  Subsequently, voltage pulses are sequentially applied from the shift register SR2 to the gate driving wirings N1 to N3, and the readout switches U1 to U3 are sequentially turned on, thereby transferring the first row transferred to the load capacitors C1 to C3. The signal is converted into a serial signal, and after impedance conversion by the voltage follower type operational amplifier OP, a signal (Vout) for three pixels is output to the outside of the photoelectric conversion device during the time T3 (read operation).

  Thereafter, a reset voltage pulse CRES is applied to the reset switches RES1 to RES3 for a time T2, thereby resetting the load capacitors C1 to C3 and preparing for the read operation of the next row (reset operation).

  Thereafter, by sequentially driving the gate drive wirings G2 and G3 from the shift register SR2, the data of all the pixels of the photoelectric conversion elements S2-1 to S3-3 are output in time series.

  In general, in an area type (two-dimensional photosensor array) photoelectric conversion device, each operation of transfer, readout, and reset is sequentially performed for each row as described above. The image signal is output intermittently as indicated by Vout in FIG. That is, the time required to read one row is T1 + T3 + T2, and in order to read all the bits, the 3 × 3 two-dimensional photoelectric conversion device shown in FIG. 1 requires three times as long. For example, the size of the photoelectric conversion device of a medical X-ray imaging device is said to be about 40 cm × 40 cm when taking an example of an X-ray imaging device that images a lung portion. If it is formed at a pitch, the total number of pixels is 4000 × 4000, which is 16 million pixels, which is an enormous number of pixels. If the reading operation is simply performed with the configuration shown in FIG. 1, a time of 4000 × (T1 + T2 + T3) is required. Actually, since the time required for T3 becomes long, a plurality of readout circuit units (N) are provided, and N scanning is performed in parallel, so that all of the time is 4000 × {(T1 + T2 + T3) / N}. A configuration in which pixels are read is common.

  However, even with such a configuration, the time required to read one row of pixels (= 4000 / N) is a conventional photoelectric conversion device in which transfer, read, and reset operations are sequentially performed. Since the transfer time T1 and the reset time T2 are required every time when the pixels in each row are read, the scanning time of the photoelectric conversion device having a large number of pixels may be longer than expected. In particular, when the transfer switching elements (T1-1 to T3-3) are composed of amorphous silicon (hereinafter referred to as “a-Si”) TFT (Thin Film Transistor) having a high cost effect, the switching performance is simple. Since it is not sufficient as compared with a switching element made of crystalline silicon, it has a problem that can be improved in terms of further high-speed reading of the photoelectric conversion device.

  Although the load capacitors are shown as read capacitors C1 to C3 and capacitive elements in FIG. 1, in practice, no separate elements are required, and the gate electrodes of the switching elements T1-1 to T3-3 and the signal wiring M1 are not required. It is comprised by the capacity | capacitance (Cgs) between electrodes formed with the electrode of -M3 side. For example, the load capacitor (reading capacity) C1 has a capacitance of Cgs of the switching elements T1-1, T2-1 and T3-1 parasitic on the signal wiring M1 when the signal charge of S1-1 in the first row is transferred. Become sum. Similarly, for example, when the signal charge of S2-2 in the second row is transferred, the capacitance value of C2 is the sum of Cgs of the switching elements T1-2, T2-2 and T3-2 parasitic on the signal wiring M1. In other words, even if the signal charge of any photoelectric conversion element is transferred, the load capacity value (C1 to C3) is added with the capacity of three Cgs of the switching element.

  Similarly, when a two-dimensional photoelectric conversion device having 4000 × 4000 pixels is configured, the load capacity of each signal line in the matrix has a capacity of Cgs × 4000. On the other hand, the switching elements RES1 to RES3 in the readout circuit section effectively add to the input capacitance (Ca in FIG. 1) parasitic to the input of the analog operational amplifier (op-amp) OP when the signal charge of the load capacitance is converted into series. Will be transferred. When the transfer switching element is formed of a-Si, since the load capacity of Cgs × 4000 >> Ca, the signal potential of the load capacity is impedance-converted with almost no reduction.

Further, when a transfer operation is performed from the load capacitors (C1 to C3) to the operational amplifier OP via the switching elements (U1 to U3) controlled by the shift register SR2, it is generated due to thermal disturbance of carriers in the switching elements. There may be a problem that the S / N of the photoelectric conversion device may be lowered due to thermal noise. The effective value Vj of this thermal noise voltage is generally given by Vj = (4KTRB) ^1/2 (Vrms). Here, K is a Boltzmann constant 1.38 × 10 ⁻²³ (J / K), T is an absolute temperature (K), and B is a frequency bandwidth (Hz) of the system. R is the resistance value (Ω) of thermal noise generated by the resistor. In the case of this system, it may be considered as the on-resistance value (Ω) of the switching element.

Further, if the capacitance on the matrix side (Cgs × 4000) is CL and the input capacitance on the operational amplifier OP side is Ca, the frequency bandwidth B = 1 / (at thermal noise voltage Vj = (4KTRB) ^1/2 (Vrms). 4R (CL‖Ca)), and Vj = (4KTR / (4R (CL‖Ca))) ^1/2 = (KT / (CL‖Ca)) ^1/2 . Here, CL‖Ca is a series combined capacity of CL and Ca.

Incidentally, Qj = CV = (KT / (CL‖Ca)) ^1/2 (Vrms) in terms of the amount of charge. That is, the thermal noise voltage Vj generated in such a system is determined only by the Boltzmann constant K, the temperature T, and the capacitance C (= CL‖Ca), and is generally called KTC noise. Hereinafter, unless otherwise noted, the thermal noise voltage is referred to as “KTC noise”. This KTC noise is given by (KT / (CL‖Ca)) ^1/2 (Vrms) if simplified. Since CL >> Ca, the KTC noise is determined by approximately (kT / Ca) ^1/2 . In order to reduce this kind of noise, Ca may be increased, but there is a limit to increasing the capacitance formed in the integrated circuit (IC).

Similarly, when the load capacitance is reset to the reset potential by the reset switches RES <b> 1 to RES <b> 3, KTC noise is generated, and there is a problem that the S / N of the photoelectric conversion device is lowered. The KTC noise at the time of reset is given by (KT / CL) ^1/2 (V). KTC noise generated at the time of transfer and KTC noise generated at the time of reset appear as random noise of the photoelectric conversion device. In particular, a high-definition, high-gradation information such as a medical X-ray imaging apparatus requires a photoelectric conversion device having a higher S / N ratio than office machines such as copying machines and electronic blackboards. Yes, KTC noise can be a big problem.

  In the photoelectric conversion circuit section, when the capacitance value between the electrodes of one photoelectric conversion element is Cs, the load capacitance value in the matrix signal wiring is CL, and the accumulated signal charge total photoelectrically converted by the photoelectric conversion element is Q, switching for transfer The signal potential V of the load capacitance CL on the matrix signal wiring after being transferred by the element is given by V = Q / (CS + CL). Since one interelectrode capacitance CS is much smaller than the load capacitance CL formed by 4000 interelectrode capacitances Cgs, it is actually approximated by V = Q / CL. When the switching element having the interelectrode capacitance Cgs is formed of an a-Si semiconductor thin film, the film thickness in the manufacture of the thin film varies, resulting in an individual difference in the capacitance value of the load capacitance CL for each device, and a large output. There may be a problem that a device or a device with a small output is produced. In order to cope with this, when a system is configured, a measure such as adding a general-purpose operational amplifier and adjusting the gain is performed. However, in the above example, N general-purpose amplifiers are required, and the adjustment process is considered together. It also increases the cost of the device.

  In addition, N reading circuit units (ICs), such as medical devices, particularly in medical devices that require a high S / N ratio, do not extend the signal wiring from the viewpoint of noise resistance. It is not preferable, and it is desirable that a necessary circuit is mounted in the vicinity of the photoelectric conversion circuit unit. However, when a large number (N) of ICs are provided, the generated heat may increase the temperature of the photoelectric conversion circuit unit. In particular, when the switching element is an amorphous silicon TFT, it is said that the dark current at the time of OFF becomes large, and there is a case where the heat generated by the IC increases the fixed pattern noise as the photoelectric conversion device.

  For example, when the photoelectric conversion device part of a medical X-ray image pickup device is configured by a solid-state image pickup device, the amount of noise required for the entire device including the photoelectric conversion element is a signal if an image quality higher than the film method is to be obtained. It is said that it is 1 / 10,000 or less with respect to the dynamic range. That is, the performance of the A / D converter necessary to achieve “digitization of X-ray image information” is also required to have a resolution of 14 bits or more. Recently, a 16-bit A / D converter is also commercially available, but the conversion speed has been reduced with the increase in the number of bits, and an X-ray having a 4000 × 4000 pixel photoelectric conversion device as described above. There are currently no high-speed A / D converters of 14 bits or more that are practically and practically used for imaging devices.

  It is an object of the present invention to provide a photoelectric conversion device capable of shortening a reading scanning time and capable of reading at high speed, and a photoelectric conversion system having the device.

  It is another object of the present invention to provide a photoelectric conversion device that can read a high S / N signal with little thermal noise (KTC noise) and a photoelectric conversion system having the device.

  In addition, an object of the present invention is to provide a photoelectric conversion device capable of obtaining good image information with reduced fixed pattern noise and free of uneven density and unnecessary stripes, and a photoelectric conversion system having the device. .

  A further object of the present invention is to provide a photoelectric conversion device capable of obtaining image information with excellent gradation and a photoelectric conversion system having the device.

  In addition, the present invention has a photoelectric conversion device that can easily compensate for non-uniform characteristics due to variations in the fabrication of photoelectric conversion elements and the like and can promote further cost reduction, and the device. An object is to provide a photoelectric conversion system.

  In order to solve the above various problems, for example, a photoelectric conversion device according to the present invention includes a pixel including a photoelectric conversion element for converting light into an electric charge and a switch element for outputting an electric signal corresponding to the electric charge. Are arranged two-dimensionally, a photoelectric conversion circuit unit for outputting the electric signal from the plurality of pixels in a row unit as a parallel signal, and a drive circuit unit for controlling the conduction of the switch element in a row unit And a read circuit unit including a sample hold unit for sample-holding the parallel signal, and a read switch for sequentially reading the parallel signal from the sample hold unit as a serial signal. In the photoelectric conversion device, the driving circuit unit includes the switch in a row different from the predetermined row within a time when the readout circuit unit performs the reading of the predetermined row. Characterized in that for turning the switch element.

In the readout circuit section, the first stage analog operational amplifier connected to each wiring of the matrix signal wiring has a noise voltage density converted to Vn (V / (Hz) ^1/2 ) at this input terminal section. And an analog operational amplifier input terminal unit that has a frequency band B (Hz) that can sufficiently amplify a signal from the photoelectric conversion circuit unit, and that is generated when a switching element in the photoelectric conversion circuit unit is turned on. It is preferable to satisfy the relationship of Vn × B ^1/2 ≦ Tn with respect to the thermal noise effective voltage Tn (Vrms) of the switching element in FIG.

  Further, in the readout circuit unit, a capacitive element that allows only an AC component to pass is connected to the output wiring from the output terminal of the analog operational amplifier that is subordinately connected to each wiring of the matrix signal wiring. It is preferable to arrange a reset switch for reproduction.

  In the readout circuit unit, it is preferable that at least one analog operational amplifier connected to each of the matrix signal wirings has a function of varying the amplification factor by an external signal.

  In addition, in the readout circuit section, it is preferable that the analog operational amplifiers connected to the respective matrix signal wirings have a function of reducing current consumption by an external signal.

In addition, an A / D conversion circuit unit for converting the readout circuit unit 2 analog signal into a digital signal is connected, and this A / D conversion circuit unit includes an operational amplifier for amplifying the signal from the readout circuit unit. N (N: 2 or more) are arranged, N M-bit A / D converters are arranged, and the ratio of the amplification factors G1, G2,... GN of the N operational amplifiers is G1: G2: ...: GN = 2 ⁰ : 2 ¹ : ...: 2 ^N-1 is set, and the outputs of the N operational amplifiers are respectively input to the N A / D converters for reading. It is preferable that the output of one A / D converter is selected from the N A / D converters according to the output level of the analog signal from the circuit unit, and is output as an N + M−1 bit digital value. It is.

  In the readout circuit section, a separate analog operational amplifier (amplifier 2) is arranged in the vicinity of the first stage analog operational amplifier (amplifier 1) connected to each of the matrix signal wirings, and the amplifier 1 is amplified. The amplifier 2 is configured as a non-inverting amplifier with a gain of 1 or more, and the amplifier 2 is configured as a buffer amplifier with a gain of 1 and a reference potential serving as a reference for the operation of the amplifier 1 is supplied from the output terminal of the amplifier 2. It is preferable to do so.

  In the readout circuit section, a capacitive element that allows only an AC component to pass is connected to the output wiring from the output terminal of the analog operational amplifier that is subordinately connected to each wiring of the matrix signal wiring. It is preferable to dispose a reset switch for reproduction and interpose a resistance element between the capacitive element and the reset switch.

  In the readout circuit section, a capacitive element that allows only an AC component to pass is connected to the output wiring from the output terminal of the analog operational amplifier that is subordinately connected to each wiring of the matrix signal wiring. It is preferable that a reset switch for reproduction is disposed, a resistance element is interposed between the capacitor element and the reset switch, and a function of varying the open / close time of the reset switch by an external signal is provided.

  In the readout circuit section, a capacitive element that allows only an AC component to pass is connected to the output wiring from the output terminal of the analog operational amplifier that is subordinately connected to each wiring of the matrix signal wiring. It is preferable that a reset switch for reproduction is arranged and a low-pass filter circuit is subordinately connected to a terminal opposite to the terminal connected to the output of the analog operational amplifier among the terminals of the capacitive element.

  In the readout circuit unit, it is preferable that at least one analog operational amplifier connected to each of the matrix signal wirings has a function of varying a slew rate by an external signal.

  Furthermore, it is preferable that the photoelectric conversion element and the switching element in the photoelectric conversion circuit unit use an amorphous silicon semiconductor as a semiconductor material.

  The photoelectric conversion element has an amorphous silicon nitride insulating layer (a-SiNx) that blocks passage of the first metal thin film layer (first conductive layer), electron carriers, and hole carriers from the insulating substrate side as a lower electrode. A hydrogenated amorphous silicon photoelectric conversion layer (a-Si: H) as a semiconductor layer, an N-type injection blocking layer for blocking hole carrier injection, a transparent conductive layer as the upper electrode (second insulating layer) or the injection The switching element (thin film transistor) is composed of a first metal thin film layer as a lower gate electrode from the insulating substrate side, and a gate insulating layer of amorphous silicon nitride. (A-SiNx), hydrogenated amorphous silicon semiconductor layer (a-Si: H), N-type ohmic contact layer, source The drain electrode is composed of a transparent conductive layer or a second metal thin film layer, the photoelectric conversion element and the switching element are formed on the same insulating substrate, and in the refresh mode, hole carriers are second from the photoelectric conversion layer. An electric field is applied to the photoelectric conversion element in a direction leading to the metal thin film layer, and in the photoelectric conversion mode, carriers generated by light incident on the photoelectric conversion layer remain in the photoelectric conversion layer, and electron carriers are allowed to remain in the second metal thin film layer. An electric field is applied to the photoelectric conversion element in the direction leading to the light source, and hole carriers accumulated in the photoelectric conversion layer or electron carriers guided to the second metal thin film layer in the photoelectric conversion mode are detected as an optical signal. May be.

  Alternatively, the photoelectric conversion elements can be divided into a plurality of systems, the photoelectric conversion elements of each system can be independently set to the refresh mode, and the photoelectric conversion elements of each system can be independently set to the photoelectric conversion mode. Good. In addition, a wavelength converter such as a phosphor may be included. In addition, it is preferable to have a grid between the light source and the photoelectric conversion device.

  According to the present invention, in the readout scanning of each row, the row scanning can be performed with almost only the readout time compared to the operation time in a case where transfer, readout, and reset are combined into one set, and the readout of the photoelectric conversion device is significantly faster. Make it possible.

In the readout circuit section, the first stage analog operational amplifier connected to each wiring of the matrix signal wiring has a noise voltage density converted to Vn (V / (Hz) ^1/2 ) at this input terminal section. And an analog operational amplifier input terminal unit that has a frequency band B (Hz) that can sufficiently amplify a signal from the photoelectric conversion circuit unit, and that is generated when a switching element in the photoelectric conversion circuit unit is turned on. By satisfying the relationship of Vn × B ^1/2 ≦ Tn with respect to the thermal noise effective voltage Tn (Vrms) of the switching element in FIG. 5, it is possible to reduce the decrease in S / N due to KTC noise during transfer. it can.

  In the readout circuit section, a capacitive element that allows only an AC component to pass is connected to an output wiring from an output terminal of an analog operational amplifier that is subordinately connected to each wiring of the matrix signal wiring, and the capacitive element is DC-regenerated. By disposing a reset switch for this purpose, it is possible to reduce the S / N reduction due to KTC noise at the time of reset, which was a problem in the past. Then, by reducing the decrease in S / N due to the KTC noise, high-quality image information with less random noise can be obtained.

  Further, in the readout circuit section, the analog operational amplifiers connected to each of the matrix signal wirings have a function of reducing the current consumption by an external signal, so that switching for transfer due to heat generation of the IC is performed. The dark current when the element is OFF can be reduced, the fixed pattern noise is reduced, and an image having no shading or streaks in the image plane can be obtained.

  In addition, according to the present invention, the output of one A / D converter is selected from the N A / D converters according to the output level of the analog signal from the readout circuit unit, and [N + M -1] By outputting as a digital value of bits, photoelectric conversion signals can be converted into high-speed and high-speed A / D conversion, and a photoelectric conversion device with high gradation can be obtained. An X-ray imaging apparatus can also be provided.

  Furthermore, the photoelectric conversion element and the switching element in the photoelectric conversion circuit portion can be provided with a large-area photoelectric conversion device at low cost by forming an amorphous silicon semiconductor as a material. In addition, in the readout circuit section, at least one analog operational amplifier connected in cascade with each of the matrix signal wirings has a function of controlling the amplification factor by an external signal, so that the amorphous silicon semiconductor thin film can be manufactured. Gain variations caused by film thickness variations can be easily and inexpensively compensated.

  In the readout circuit section, a separate second analog operational amplifier (amplifier 2) is arranged in the vicinity of the first stage analog operational amplifier (amplifier 1) connected to each of the matrix signal wirings, and the amplifier 1 is configured as a non-inverting amplifier having an amplification factor of 1 or more, and the amplifier 2 is configured as a buffer amplifier having an amplification factor of 1. The reference potential serving as a reference for the operation of the amplifier 1 is output from the amplifier 2. By supplying from the terminal, the reference potential of the amplifier 1 is stabilized, an accurate photoelectric conversion signal can be obtained, and the S / N can be increased.

  In the readout circuit section, a capacitive element that allows only an AC component to pass is connected to the output wiring from the output terminal of the analog operational amplifier that is subordinately connected to each wiring of the matrix signal wiring. By arranging a reset switch for reproduction and interposing a resistance element between the capacitive element and the reset switch, a low-pass filter is formed at the time of DC reproduction, and the random noise of the analog operational amplifier can be reduced. N can be increased.

  In the readout circuit section, a capacitive element that allows only an AC component to pass is connected to the output wiring from the output terminal of the analog operational amplifier that is subordinately connected to each wiring of the matrix signal wiring. A reset switch for playback is arranged, a resistance element is interposed between the capacitive element and the reset switch, and a function for varying the open / close time of the reset switch by an external signal is provided. The S / N in the mode can be increased, and the frame rate can be increased in the moving image mode, thereby further improving the usability.

  In the readout circuit section, a capacitive element that allows only an AC component to pass is connected to the output wiring from the output terminal of the analog operational amplifier that is subordinately connected to each wiring of the matrix signal wiring. A reset switch for reproduction is arranged, and a low-pass filter circuit is subordinately connected to a terminal on the opposite side of the terminal connected to the output of the analog operational amplifier among the terminals of the capacitive element. Random noise can be reduced and S / N can be increased.

  Further, in the readout circuit section, at least one analog operational amplifier that is subordinately connected to each wiring of the matrix signal wiring has a function of changing the slew rate by an external signal, so that noise reduction is necessary. This is advantageous when reading a weak photoelectric conversion signal.

  The photoelectric conversion element includes a first metal thin film layer, an amorphous silicon nitride insulating layer (a-SiNx) that blocks passage of electron carriers and hole carriers, and a hydrogenated amorphous silicon photoelectric conversion as a lower electrode from the insulating substrate side. A layer (a-Si: H), an N-type injection blocking layer for blocking hole carrier injection, a transparent conductive layer as an upper electrode, or a second metal thin film layer disposed on a part of the injection blocking layer. The Further, the switching element includes a first metal thin film layer, an amorphous silicon nitride gate insulating layer (a-SiNx), and a hydrogenated amorphous silicon semiconductor layer (a-Si: H) as a lower gate electrode from the insulating substrate side. , An N-type ohmic contact layer, and a transparent conductive layer or a second metal thin film layer as source and drain electrodes. The photoelectric conversion element and the switching element are formed on the same insulating substrate, and in the refresh mode, an electric field is applied to the photoelectric conversion element in a direction in which hole carriers are guided from the photoelectric conversion layer to the second metal thin film layer. In the photoelectric conversion mode, carriers generated by light incident on the photoelectric conversion layer remain in the photoelectric conversion layer, and an electric field is applied to the photoelectric conversion element in a direction in which electron carriers are guided to the second metal thin film layer. By detecting hole carriers accumulated in the photoelectric conversion layer according to the mode or electron carriers introduced to the second metal thin film layer as an optical signal, it is possible to easily cope with the moving image mode.

  Further, the photoelectric conversion elements are divided into a plurality of systems, the photoelectric conversion elements of each system are independently set to the refresh mode, and the photoelectric conversion elements of each system are independently set to the photoelectric conversion mode. In the moving image mode, the frame can be substantially enlarged and a large number of continuous images can be obtained.

  More specifically, according to the present invention, in the readout circuit unit, the output signal amplified from the matrix signal line of each photoelectric conversion circuit unit is temporarily transferred to the readout capacitor by the transfer switch, and then sequentially switched by the readout switch. As a result, the readout scanning time for one row is significantly shortened compared to the conventional one. That is, according to the present invention, it is possible to achieve a photoelectric conversion device capable of high-speed reading and a photoelectric conversion system using the device.

  Further, according to the present invention, the low-noise performance is provided in the first-stage analog operational amplifier of the reading circuit unit, so that the output obtained is less affected by the KTC noise generated in the photoelectric conversion circuit unit. That is, according to the present invention, it is possible to provide a photoelectric conversion device having a high S / N and a photoelectric conversion system using the device.

  In addition, a capacitor element that allows only an AC component to pass through is connected in series to the output terminal of the analog operational amplifier, and a reset switch for reproducing the capacitor element by direct current is disposed, thereby generating KTC noise generated when the photoelectric conversion circuit unit is reset. S / N can be prevented from lowering due to a high-S / N photoelectric conversion device and a system thereof, and a high-quality image without roughness can be obtained.

  In addition, according to the present invention, in a state where a reading operation is not performed, that is, in a so-called standby state, the function of reducing the consumption current of the operational amplifier used in the reading circuit unit can suppress the heat generation of the IC, and the photoelectric conversion during the operation The dark current of the switching element in the circuit portion can be reduced. That is, according to the present invention, the fixed pattern noise of the photoelectric conversion device is reduced, and a good image without in-plane light and shade can be obtained.

According to the present invention, in the A / D conversion circuit unit, N operational amplifiers (N: 2 or more) for amplifying the signal from the readout circuit unit are arranged, and an M-bit A / D converter , And the ratio of the amplification factors G1, G2,..., GN of the N operational amplifiers is G1: G2: ...: GN = 2 ⁰ : 2 ¹ : ...: 2 ^{N- 1} and the outputs of the N operational amplifiers are respectively input to the N A / D converters, and the N A / Ds are output in accordance with the output level of the analog signal from the readout circuit unit. By selecting the output of one A / D converter from the converters and outputting it as a digital value of N + M−1 bits, the photoelectric conversion signal can be A / D converted with high resolution and high speed. That is, according to the present invention, image data with high gradation can be obtained. The photoelectric conversion apparatus of the present invention is also suitable for use in a photoelectric conversion system such as a high-performance medical X-ray imaging apparatus.

  Furthermore, by using an amorphous silicon semiconductor for the photoelectric conversion element and the switching element in the photoelectric conversion circuit section, a large-area photoelectric conversion device can be provided at a low cost with an easy process. In addition, by providing a function that can control the amplification factor of the analog operational amplifier of the matrix signal wiring in the readout circuit section by an external signal, it is easy to gain variations due to film thickness variations in the production of amorphous silicon semiconductor thin films. Therefore, cost reduction of the apparatus can be promoted.

  As described above, the photoelectric conversion device of the present invention is extremely excellent in terms of speed, S / N, gradation, and cost, and is strongly desired in the medical industry and industry in recent years. In response to the demands of “digitalization of information”, we are improving the diagnosis efficiency in hospitals all over the world, not only in Japan, which is facing an aging society, but also in nondestructive inspection of buildings and other various parts It is possible to improve inspection efficiency and subsequent handling efficiency.

  Hereinafter, the contents of the present invention will be described in detail as embodiments with reference to the drawings.

(Embodiment 1)
FIG. 3 is a circuit diagram of the photoelectric conversion device showing the first embodiment of the present invention. In order to simplify the explanation, the figure is composed of a total of 9 pixels of 3 × 3. Moreover, the same symbols are used for the same members as in FIG. S1-1 to S3-3 are photoelectric conversion elements for receiving visible light and converting them into electrical signals, and T1-1 to T3-3 are signals photoelectrically converted by the photoelectric conversion elements S1-1 to S3-3. This is a switch element for transferring charges to the matrix signal wirings M1 to M3. G1 to G3 are gate drive wirings of switches connected to the shift register (SR1) and connected to the switch elements T1-1 to T3-3. As described above, the matrix signal wiring M1 is added with three capacitances of the interelectrode capacitance (Cgs) of the switching element at the time of transfer, and is not represented as a capacitive element in FIG. The same applies to the other matrix signal wirings M2 and M3.

  Photoelectric conversion elements S1-1 to S3-3, switching elements T1-1 to T3-3, gate drive wirings G1 to G3, and matrix signal wirings M1 to M3 are displayed in the photoelectric conversion circuit unit 101 in the figure. Although not provided, each is disposed on one insulating substrate. Reference numeral 102 denotes a drive circuit unit including a shift register (SR1) for opening and closing the switch elements T1-1 to T3-3. A1 to A3 are operational amplifiers for amplifying the signal charges of the matrix signal wirings M1 to M3 and converting the impedance, and are illustrated only as buffer amplifiers constituting a voltage follower circuit in the drawing. Sn1 to Sn3 are transfer switches that read out the outputs of the operational amplifiers A1 to A3, that is, the outputs of the matrix signal wirings M1 to M3, and transfer them to the read capacitors CL1 to CL3.

  The read capacitors CL1 to CL3 are read by the read switches Sr1 to Sr3 via the buffer amplifiers B1 to B3 constituting the voltage follower circuit. Reference numeral 103 denotes a shift register (SR2) for switching the read switches Sr1 to Sr3. The parallel signals of CL1 to CL3 are serially converted by Sr1 to Sr3 and the shift register (SR2) 103, input to the operational amplifier 104 constituting the final voltage follower circuit, and further digitized by the A / D conversion circuit unit 105. The RES1 to RES3 are reset switches for resetting signal components stored in capacitors (three Cgs) added to the matrix signal wirings M1 to M3, and are reset to a certain reset potential by a pulse from the CRES terminal. (Reset to GND potential in the figure).

  Reference numeral 106 denotes a power source for applying a bias to the photoelectric conversion elements S1-1 to S3-3. The read circuit unit 107 includes buffer amplifiers A1 to A3, transfer switches Sn1 to Sn3, read capacitors CL1 to CL3, buffer amplifiers B1 to B3, read switches Sr1 to Sr3, a shift register SR2, a final stage operational amplifier 104, and a reset circuit. The switches RES1 to RES3 are configured.

  FIG. 4 is a timing chart showing the operation of the photoelectric conversion apparatus shown in FIG. Details of the operation will be described with reference to FIG. The signal charges photoelectrically converted by the photoelectric conversion elements S1-1 to S3-3 are accumulated for a certain period in the capacitance component formed in the photoelectric conversion element. The signal charges stored in the photoelectric conversion elements S1-1 to S1-3 in the first row are turned on only by the switching elements T1-1 to T1-3 by the gate pulse signal G1 of the shift register (SR1) 102 for t1 time. Then, it is transferred to a capacitance component (capacity corresponding to 3 Cgs of the switching elements T1-1 to T3-3) formed in each of the matrix signal wirings M1 to M3. In FIG. 4, M1 to M3 indicate the state of the transfer, and indicate the case where the signal amounts stored in the photoelectric conversion elements are different. That is, in the first row of photoelectric conversion elements (S1-1 to S1-3), the output level is S1-2> S1-1> S1-3. The signal outputs of the matrix signal wires M1 to M3 are impedance-converted by operational amplifiers A1 to A3, respectively.

  Thereafter, the switching elements Sn1 to Sn3 in the readout circuit section are turned “ON” for t2 time by the SMPL pulse shown in FIG. 4 and transferred to the readout capacitors CL1 to CL3, respectively. The signals of the read capacitors CL1 to CL3 are impedance-converted by buffer amplifiers B1 to B3, respectively. Thereafter, the read switches Sr1 to Sr3 are sequentially turned on by the shift pulses Sp1 to Sp3 from the shift register (SR2) 103, so that the parallel signal charges transferred to the read capacitors CL1 to CL3 are converted into a serial signal. And read. Assuming that the pulse widths of the shift pulses Sp1, Sp2, and Sp3 are Sp1 = Sp2 = Sp3 = t3, the time required for this serial conversion readout is t3 × 3. The serially converted signal is output from the operational amplifier 104 at the final stage and further digitized by the A / D conversion circuit unit 105.

  Vout shown in FIG. 4 indicates an analog signal before being input to the A / D conversion circuit unit. As shown in FIG. 4, the parallel signals of S1-1 to S1-3 in the first row, that is, the parallel signals of the signal potentials of the matrix signal wirings M1 to M3 are proportional to the Vth signal. Above, it is serially converted. Finally, the signal potentials of the matrix signal wirings M1 to M3 are reset to a constant reset potential (GND potential) via the reset switch elements RES1 to RES3 when the CRES pulse is “ON” for t4 time, and the next photoelectrical In preparation for the transfer of signal charges in the second row of the conversion elements S2-1 to S2-3. Similarly, the photoelectrically converted signals in the second and third rows are repeatedly read out.

  Here, as can be seen from FIG. 4, in the present invention, the matrix signal wirings M1 to M3 in the first row are within a time range of t3 × 3 necessary for the signal charge read operation of CL1 to CL3 in the first row. Two operations can be performed: a capacitance reset operation and a transfer operation by the gate pulse G2 of the photoelectric conversion elements S2-1 to S2-3 in the second row. That is, the time required for reading one row is t4 + t1 + t2, and this time can be made substantially equal to (t3 × 3) + t2. In the figure, the capacitance formed in one of the matrix signal wirings M1 to M3 is at most three of the interelectrode capacitance Cgs of the switching elements connected to the photoelectric conversion elements S2-1 to S2-3. .

  However, as described above, when an actual photoelectric conversion element is configured, the number of bits is several hundred to several thousand in one column, so that the capacitance value is very large as compared with the reading capacitor CL. As a result, the time t2 required for the transfer by the SMPL pulse is sufficient to be several times the time constant determined by the product of the capacitance value of the readout capacitor CL and the on-resistance value of the switch element Snx (x: 1 to 3). It becomes. If the reading circuit unit 107 is formed of an integrated circuit (IC) using normal crystalline silicon as a substrate material, the time t2 can be operated in a time sufficiently shorter than each time t1, t3, or t3 × 3. it can.

  That is, the time required for reading the signal charges of one row of photoelectric conversion elements can be set to approximately t4 + t1 = t3 × 3. In the case of the above-described example, this means that for reading one row, (time t1 required for transfer from the photoelectric conversion element to the matrix signal wiring) + (time t3 necessary for reading the signal of the matrix signal wiring) (* 3) + (time t4 required for resetting the capacitance component of the matrix signal wiring) is required, but in the present embodiment, (time required for reading the signal of the matrix signal wiring) One row can be read out at time t3 × 3), and the reading speed as a photoelectric conversion device is greatly reduced.

  FIG. 5A is a schematic top view of the photoelectric conversion circuit unit when the photoelectric conversion element and the switching element are configured using an amorphous silicon semiconductor thin film. FIG. 5B is a diagram in FIG. It is a schematic sectional block diagram in AB. The photoelectric conversion element 301 and the switching element 302 (amorphous silicon TFT, hereinafter simply referred to as TFT) are formed on the same substrate 303, and the lower electrode of the photoelectric conversion element 301 is the same as the lower electrode (gate electrode) of the TFT 302. The upper electrode of the photoelectric conversion element 301 is shared by the same second metal thin film layer 305 as the upper electrode (source electrode, drain electrode) of the TFT 302. The first and second metal thin film layers also share the gate driving wiring 306 and the matrix signal wiring 307 in the photoelectric conversion circuit portion. In FIG. 5A, a total of 4 pixels of 2 × 2 is described as the number of pixels. In FIG. 5A, the hatched portion is the light receiving surface of the photoelectric conversion element. Reference numeral 309 denotes a power supply line for applying a bias to the photoelectric conversion element. Reference numeral 310 denotes a contact hole for connecting the photoelectric conversion element 301 and the TFT 302.

Here, a method for forming the photoelectric conversion circuit portion in this embodiment will be described. First, a first metal thin film layer 304 of about 500 Å is deposited on the insulating substrate 303 by sputtering or resistance heating, and patterned by photolithography, and unnecessary areas are etched. The first metal thin film layer 304 serves as a lower electrode of the photoelectric conversion element 301 and a gate electrode of the switching element 302. Next, a-SiNx (311), a-Si: H (312), and an N ⁺ layer (313) are sequentially stacked by 3000, 5000, and 1000 angstroms, respectively, in the same vacuum by the CVD method. Each of these layers is an insulating layer / photoelectric conversion semiconductor layer / hole injection blocking layer of the photoelectric conversion element 301 and a gate insulating film / semiconductor layer / ohmic contact layer of the switching element 302 (TFT).

  Moreover, it is utilized also as an insulating layer of the cross | intersection (FIG. 5 (a) 314) of the 1st metal thin film layer 304 and the 2nd metal thin film layer 305. FIG. The film thickness of each layer is not limited to the above thickness, but is optimally designed according to the voltage used for the photoelectric conversion device, charge, the amount of incident light on the light receiving surface of the photoelectric conversion element, and the like. At least a-SiNx is preferably 500 angstroms or more, which cannot pass electrons and holes, and can sufficiently function as a gate insulating film of the TFT 302. After depositing each layer, an area to be a contact hole (see FIG. 5A 310) is dry-etched by RIE or CDE, and then aluminum (Al) is sputtered or resistance-heated as the second metal thin film layer 305. About 10,000 angstroms. Further, patterning is performed by photolithography, and unnecessary areas are etched.

The second metal thin film layer becomes the upper electrode of the photoelectric conversion element 301, the source and drain electrodes of the switching TFT 302, and other wirings. Simultaneously with the formation of the second metal thin film layer 305, the upper and lower metal thin film layers are connected through the contact hole portion 310. Further, in order to form a channel portion of the TFT 302, a part between the source electrode and the drain electrode is etched by the RIE method, and then unnecessary a-SiNx layer, a-Si: H layer, and N ⁺ layer are formed by RIE. Each element is separated by etching. Thus, the photoelectric conversion element 301, the switching TFT 302, other wirings (306, 307, 309), and the contact hole portion 310 are formed. In the schematic cross-sectional view of FIG. 5B, only two pixels are shown, but it goes without saying that a large number of pixels are simultaneously formed on the insulating substrate 303.

Finally, for the purpose of improving moisture resistance, each element and wiring are covered with a SiNx passivation film (protective film) 315. As described above, the common first metal thin film layer, a-SiNx, a-Si: H, N ⁺ layer, and second metal thin film layer in which the photoelectric conversion element, the switching TFT, and the wirings are simultaneously deposited. And only by etching each layer.

  By using the process mainly made of amorphous silicon semiconductor as described above, photoelectric conversion elements, switching elements, gate driving wirings, and matrix signal wirings can be fabricated on the same substrate at the same time. The conversion circuit unit can be provided easily and inexpensively.

  In general, an amorphous silicon TFT has a significantly high ON resistance because it has a lower electron mobility than a crystalline silicon switch element. For example, the ON resistance of a TFT having a channel size (W / L): 50 μm / 10 μm manufactured by the above process becomes a very large value of 8 megaohms when a 12 V bias (Vgs) is applied. When this photoelectric conversion circuit unit as shown in FIGS. 5A and 5B is configured with a pixel pitch of 100 μm using this TFT, the capacitance component formed in the photoelectric conversion element is 2 to 3 (pF). The time required for transfer from the photoelectric conversion element to the matrix signal wiring is approximately 20 (μsec) as the time constant τ. In order to perform the transfer sufficiently, a time several times the time constant is required. If a time of 4τ is provided, the pulse width of the TFT driving gate pulse is 80 (μsec).

One Cgs of the TFT is about 0.05 (pF), and when the number of pixels in one row is set to 4000, the capacitance component formed in one matrix signal wiring is 4000 × Cgs = 200 ( pF). The ON resistances of the reset switch elements (RES1 to RES3) in the readout circuit section shown in FIG. 3 can be easily manufactured from several hundred ohms to several kiloohms if the readout circuit section is an IC (crystalline silicon). The time constant τ _R necessary for the reset is less than 1 (μSEC) if the resistance component of the wiring can be ignored. However, the reset current passing path in the reset operation flows through the gate drive wiring (G1, G2, G3 in FIG. 3) via the Cgs of the TFT. If chromium is used as the material of the gate drive wiring and the photoelectric conversion circuit portion is configured as shown in FIGS. 5A and 5B, the resistance value due to the wiring is expected to increase. .

  If the wiring width is increased in order to reduce the resistance value, the occupation area of the light receiving surface of the photoelectric conversion element with respect to the pixel area of 100 (μm) × 100 (μm) is reduced, and the signal amount can be secured, and the wiring thickness is increased. If the thickness of the protective film 315 is increased, the coverage of the protective film 315 is lowered and there is a concern about reliability. Therefore, a wiring width of about 10 (μm) and a wiring film thickness of about 1000 (A) are reasonable in design.

When chromium is used for the gate drive wiring, the sheet resistance is about 2 ohm / □, and the wiring length is about 40 (cm) at 4000 (pixels) × 100 (μm) in the above example. The resistance is 80 (kiloohms). As a result, τ _R = 1 (μsec) is insufficient for resetting the capacitance of about 200 (pF) formed in the matrix signal wiring. In an actual reset operation, it becomes a two-dimensional distributed constant circuit and cannot be simply expressed by the CR time constant, but in order to sufficiently reset, 200 (pF) × 80 (kΩ) = 16 (μsec) Several times are required, and a time substantially equal to the driving gate pulse width of 80 (μsec) is required.

  Further, when the reading circuit unit (IC) to which the matrix signal wiring for 4000 pixels is connected is constituted by one IC, the size of the IC becomes very large and the yield of the IC itself is lowered. In addition, it takes too much time for one IC to read out data for one row = 4000 pixels in series. Therefore, the read circuit section is divided into an appropriate number: N, and N are operated simultaneously. For example, N is set so that it can be serially converted by the time (t1 + t4) of the sum of the transfer time (t1) from the photoelectric conversion element to the matrix signal wiring and the reset time (t4) of the matrix signal wiring. In the above example, transfer time t1: 80 (μsec) + reset time t4: 80 (μsec) = 160 (μsec), conversion rate for serial conversion (Sp pulse width of shift register 2: t3) is 1. If it is 6 (μsec), N = 20 readout circuit portions capable of inputting 100 pixels are prepared.

  In this example, the time required for reading one row is t1 + t4 + (t3 × 100) = 320 (μsec), whereas in the present invention, the time required for reading one row is t3 × 100 = 160 (μsec), which is substantially twice as fast.

  In addition, when the photoelectric conversion device does not require high speed, the transfer time t1 and the reset time t4 can be set longer even at the same reading speed as in the past, so that more sufficient transfer and reset are possible.

  In addition, t3 in the reading circuit portion can be easily made shorter than t3 = 1.6 (μsec) in the above example in the case of a normal IC of crystalline silicon. In this case, since the time required for reading one row is determined by ts + t4, there is no change in the reading speed, but the number (N) of ICs in the reading circuit portion can be reduced, so that the photoelectric conversion device can be manufactured at a lower cost. Can be provided.

  As described above, according to the present invention, in the readout scanning of each row, it is possible to perform the readout scanning of each row in a time required for only reading compared to the operation time of transfer + reading + reset which is indispensable in the prior art. Thus, the reading speed of the photoelectric conversion device can be greatly increased.

(Embodiment 2)
FIG. 6 is a circuit diagram of a photoelectric conversion device showing a second embodiment of the present invention, and is an example in which the photoelectric conversion circuit unit is configured by 3 × 3 = 9 pixels. The same components as those in FIG. 3 shown in the first embodiment are denoted by the same reference numerals, and description thereof is omitted. 6 differs from FIG. 3 in that the buffer amplifiers L1 to L3 connected to the matrix signal wirings in the reading circuit unit are changed to non-inverting amplifiers having an amplification factor G determined by the resistors R1 and R2. There is in point. Although not shown in FIG. 6, the operational amplifiers of the buffer amplifiers L1 to L3 are excellent in terms of very low noise performance as compared with other amplifiers. The amplification factor is 1+ (R2 / R1).

In general, an operational amplifier generates random voltage noise, which is dominated by noise generated in an internal transistor, particularly the first-stage transistor. For example, when the first stage portion is composed of a bipolar transistor, it is said that the thermal noise generated in the base resistance determines the noise amount of the operational amplifier. The amount of noise is generally expressed with respect to the unit bandwidth, and the unit is (Volt / (Hz) ^1/2 ). When the operational amplifier is used as a non-inverting amplifier as shown in FIG. 6, the amount of noise is also multiplied by 1+ (R2 / R1) according to the frequency band to be operated. Hereinafter, the noise generated in the operational amplifier is considered as a noise value before multiplying the amplification factor, that is, an input conversion noise voltage, and expressed as Vn (V / (Hz) ^1/2 ).

  In the present embodiment, Vn of the operational amplifiers (L1 to L3) shown in FIG. 6 is set to a certain fixed value or less. The certain value is a noise value due to so-called KTC noise generated during the transfer operation via the switching elements T1-1 to T3-3 in the photoelectric conversion circuit unit 101. That is, the amount of noise generated in the first operational amplifier unit (L1 to L3) of the reading circuit unit is set to be equal to or less than the amount of KTC noise generated in the photoelectric conversion circuit unit 101. Both noises are intrinsic noises that can occur in principle, and cannot be set to “zero” in design.

Next, taking the photoelectric conversion device having 4000 × 4000 = 16 million pixels described in the first embodiment as an example, each noise is estimated. When each layer of a-SiNx, a-Si, and N ⁺ is stacked at 3000, 5000, and 1000 angstroms and the pixel pitch is 100 μm, the capacitance (C1) in the photoelectric conversion elements S1-1 to S3-3 is about The readout capacitance (C2) of 3 pF and one of the matrix signal wirings M1 to M3 is 200 pF with Cgs × 4000. The KTC noise (Tn) generated when the transfer operation is performed by the switching elements (TFT) T1-1 to T3-3 is obtained as voltage noise on the capacitor C2 of the matrix signal wirings M1 to M3. × T × (C1‖C2)) ^1/2 / (C1 + C2). However, K: Boltzmann constant (1.38 × 10 ⁻²³ (J / K)), T is an absolute temperature, and C1‖C2 is a series combined capacity of C1 and C2.

  This noise Tn has a Gaussian distribution statistically and is represented by an effective noise voltage value (Vrms). When Tn at room temperature (300 K) is calculated, Tn = 0.55 (μVrms). On the other hand, noise generated in the operational amplifiers L1 to L3 differs depending on the frequency band B to be handled. As described in the first embodiment, if the transfer time is 80 (μs) + the reset time is 80 (μs), the operational amplifier has (1/160 (μs)) = 6.25 (kHz). A signal is input. Assuming that the operational amplifier is operated in the frequency band B of 25 (kHz), which is four times that, the transferred photoelectric conversion signal and the KTC noise generated in the photoelectric conversion circuit section are sufficiently amplified (G times). .

In addition, the effective noise An (= Vn × B ^1/2 ) of the operational amplifier input section generated in the operating frequency band is also amplified G times. The noise An generated in the operational amplifier and the KTC noise Tn in the photoelectric conversion circuit unit 101 are independent from each other, and the effective noise Jn in the amplifier input unit obtained by combining both noises is Jn = (An ² + Tn ² ) ^1/2. The effective total noise at the amplifier output terminal is Jn × G.

Here, when An >> Tn, Jn is determined by An, and the S / N as a photoelectric conversion device is disadvantageous. Therefore, An = Tn or An <Tn is desirable. When the frequency band B is 25 (kHz) in the above example, the input equivalent noise voltage Vn of the operational amplifier is 3.5 (nV / (Hz) from Tn ≧ (Vn × B ^1/2 ) (= An). ^1/2 ) or less is desirable. When Vn = 3.5 (nV / √Hz), the effective noise of the amplifier is equal to the KTC noise Tn, and the combined effective noise Jn at the amplifier input section is Jn = (Vn ² + Tn ² ) ^1/2 Therefore, it becomes √2 times Tn. That is, in the above example, Jn = 0.55 × 2 ^1/2 = 0.78 (μVrms).

  When the photoelectric conversion device in the present embodiment is used for an X-ray imaging device and an image comparable to the conventional film system is obtained, the S / N ratio required as the device is very high, and generally S / N = 10000 times or more. It is said that it is necessary. Here, an example in which a photoelectric conversion device in which photoelectric conversion elements are two-dimensionally arranged is used in an X-ray imaging device will be described first.

  FIG. 7 is a schematic cross-sectional view of a medical X-ray detection apparatus configured using a two-dimensional photoelectric conversion apparatus. The X-rays emitted from the X-ray source 1501 are irradiated on a specimen 1502 such as a human body (examined part of a patient's affected area or object), and are used for in-vivo information such as lung, bone, or lesion, or internal information such as structure or content space. Corresponding X-rays go to the grid plate 1503. The grid plate 1503 is arranged for the purpose of preventing the fluorescent material 1504 and the photoelectric conversion device 1506 from being irradiated with scattered X-rays in the specimen, such as an absorbing material 1507 that absorbs X-rays such as lead and aluminum. And a transparent material 1508 that transmits X-rays. The X-rays that have passed through the grid are irradiated to an X-ray visible conversion phosphor 1504 that is a wavelength converter, and are converted into a wavelength within a range having sensitivity of a photoelectric conversion element such as visible light. The fluorescence from the X-ray visible conversion phosphor 1504 is photoelectrically converted by the photoelectric conversion device 1506. In FIG. 7, reference numeral 1509 denotes a photoelectric conversion element, 1510 denotes a switching element, and 1511 denotes a protective film that protects the photoelectric conversion element 1509 and the switching element 1510. Reference numeral 1512 denotes an insulating substrate on which the photoelectric conversion element 1509 and the switching element 1510 are arranged.

  By the way, as shown in FIG. 7, when the X-ray visible conversion phosphor 1504 is brought into close contact with the photoelectric conversion device 1506, the illuminance obtained on the light receiving surface of the photoelectric conversion element 1509 is maximized, and the illuminance at that time is the fluorescence used. Although it depends on the dose of the body and the X-ray source, it is possible to secure an illuminance of about several (Lx). In the example of the above-described 4000 × 4000 photoelectric conversion device, the photoelectric current flowing to the photoelectric conversion element 1509 with 1 (Lx) light is about 5 (pA) with one photoelectric conversion element. If accumulated in the capacitor C1 of 3 (pF) in the photoelectric conversion element for a period of 500 (msec), the signal output S in the capacitor C2 of 200 (pF) of the matrix signal wiring after transfer through the switching TFT is 5 pF. × 500 msec / (3 pA + 200 pA) = 12.3 (mV). The noise value Jn in the matrix signal wiring 200 (pF) capacitor C2 is Jn = 0.78 (μVrms), the effective noise Jn is N, and the S / N ratio is 12.3 (mV) /0.78 ( μV) = 15800. That is, it becomes possible to sufficiently function as a photoelectric conversion unit of the X-ray imaging apparatus.

  The noise in the operational amplifier is not limited to Vn, and there is also thermal noise generated by resistors R1, R2, for example. It is easy to make it sufficiently smaller than the noise caused by Vn by reducing its resistance value. Further, the operational amplifier has a current noise component (In) in the input section. This can be sufficiently reduced against noise caused by Vn by using a field effect transistor as the first stage transistor of the operational amplifier. That is, the input equivalent noise voltage Vn of the operational amplifier greatly affects the S / N of the photoelectric conversion device, so that the Vn of the operational amplifier is defined in the present invention.

(Embodiment 3)
FIG. 8 is a circuit diagram of a photoelectric conversion device showing a third embodiment of the present invention, which is an example in which the photoelectric conversion circuit unit is configured by 3 × 3 = 9 pixels. The same components as those in FIG. 6 shown in the second embodiment are denoted by the same reference numerals, and description thereof is omitted. FIG. 8 differs from FIG. 6 in that in the reading circuit unit, capacitive elements CC1 to CC3 that allow only an AC component to pass are connected in the middle of the output wiring from the output unit terminals of the operational amplifiers L1 to L3. This is the point where reset switches D1 to D3 for reproduction are arranged. Also, impedance conversion buffer amplifiers A1 to A3 are connected to the capacitive elements CC1 to CC3.

  FIG. 9 is a timing chart showing the operation in FIG. 8. The operation relating to the capacitors CC1 to CC3 and the reset switches D1 to D3 is particularly described, and the other operations are the same as those in FIG. The operation of this embodiment will be described below with reference to FIGS.

CRES is a switch RES1 for resetting the capacitance CL (three Cgs of TFTT1-1 to T3-3, not shown in FIG. 8) formed in the matrix signal wirings M1 to M3 to the reset potential (GND). RES3 control signal. P1 shows how the potential of the node (for example, P1 in FIG. 8) of the matrix signal wiring changes. Originally, the node P1 must be reset to the reset potential GND by the CRES signal “Hi”. However, when the capacitor C2 of the matrix signal wiring is reset, thermal noise due to the on-resistance of the switch RES2 is stochastically generated as so-called KTC noise. The amount of noise Rn = (KT / C2) ^1/2 (Vrms). In the above-described example of C2 = 200 pF, Rn = 4.55 (μVrms), which exceeds the above-mentioned Jn = 0.78 (μVrms) and becomes the main cause of noise in the photoelectric conversion device.

  If reading is performed in a state where the noise amount Rn is superimposed on the capacitor C2 of the matrix signal wiring, the S / N as a photoelectric conversion device is naturally reduced. In the waveform P1 of FIG. 9, the amount of deviation from the reset potential GND after CRES “OFF” (indicated as “error” in FIG. 9) is due to the KTC noise generated at the time of reset. This noise is also multiplied by an amplification factor G = 1 + (R2 / R1) by the operational amplifiers L1 to L3. Although not shown in FIG. 8, the output of the operational amplifiers L1 to L3 always shows an output G times the waveform P1. P2 in FIG. 9 is a waveform of the counter electrode of the capacitive element connected in series to the output of the amplifier L2. That is, the waveform of the node P2 shown in FIG. 8 is represented.

  Here, a reset switch D2 is connected to the node P2, and is controlled by a control signal DRES. DRES turns “ON” almost simultaneously with CRES, and DRES turns “OFF” slightly after CRES “OFF”. During the period when DRES is “ON”, the reset potential GND is applied to the node P2. Even when DRES is “OFF” and the switch element D2 is in a high impedance state, the node P2 is held at the GND potential. In this state, for example, when the gate (G2) of the transfer TFT is “ON”, the signal charge stored in the capacitor of the photoelectric conversion element S2-2 is transferred to the capacitor C2 of the matrix signal wiring.

  The state is shown in the waveform P1 of FIG. 9, and the KTC noise Rn at the time of reset previously held after the completion of CRES is also superimposed during the transfer operation. However, the waveform of the node P2 during this transfer operation process is such that the DC component of Rn multiplied by G by the operational amplifiers L1 to L3 is blocked by the capacitive element CC2, so that the change in the potential of C2 due to the signal of the photoelectric conversion element. Appears G times. That is, the KTC noise at the time of reset is canceled. Thereafter, the output of P2 is transferred to the capacitive element CL2 by the SMPL pulse, and is serially converted by SR2 and output from the operational amplifier 104. The operation of this part is the same as that described in the first embodiment.

  As described above, in the present invention, in the readout circuit unit 107, the capacitive elements CC1 to CC3 that allow only an AC component to pass are connected to the output wiring from the output unit terminals of the operational amplifiers L1 to L3. By arranging reset switches 1 to D3 for direct current reproduction, KTC noise generated at the time of resetting the capacitance formed in each matrix signal wiring can be removed, so that the S / N ratio of the photoelectric conversion device is increased. A high-quality image can be obtained.

(Embodiment 4)
FIG. 10 is a circuit diagram of the photoelectric conversion device showing the fourth embodiment of the present invention, and is an example in which the photoelectric conversion circuit unit is configured by 3 × 3 = 9 pixels. The same components as those in FIG. 8 shown in the third embodiment are denoted by the same reference numerals, and description thereof is omitted. FIG. 10 differs from FIG. 8 in that operational amplifiers K1 to K3 having a function of variably controlling the amplification factor by an external signal are added to the reading circuit unit 107. In FIG. 10, four signal lines A1 to A4 for controlling the amplification factor from the outside are provided, and four amplification factors are selected. FIG. 11 shows a schematic circuit diagram of the inside of the operational amplifiers K1 to K3 having variable gain in FIG. The function will be briefly described below.

  Terminals A1, A2, A3, and A4 are terminals for inputting a signal for selecting an amplification factor from the outside, and only one of the four terminals is set to “Hi”. When a “Hi” signal is input to any one of the terminals A1, A2, A3, and A4, the switch elements S1, S2, S3, and S4 connected to the terminals A1, A2, A3, and A4 are turned on. . When any one of the switches is turned on, the operational amplifiers K1 to K3 are operated as non-inverting amplifiers. For example, when the resistance values of the resistors R3 to R7 are sufficiently larger than the on-resistance of each switch element and are all set to the same value R (Ω), the amplification factor becomes 1 + 1/4 = 1.25 by turning on S1. When S2 is turned on, the amplification factor is 1 + 2/3 = 1.66 times. When S3 is turned on, the amplification factor is 1 + 3/2 = 2.5 times, and when S4 is turned on, the amplification factor is 1 + 4/1 = 5 times. If the resistance values of the resistors R3 to R7 are appropriately selected, the other four desired amplification factors can be obtained.

In the present embodiment, an example in which the four amplification factors are switched by four control signals has been shown, but nothing is limited to four, and the amplification factors may be switched by a desired number of control signals. Further, if a multiplexer circuit is connected to the control terminal, 2 ^N types of switching can be performed by ^N external control signals.

  In response to the problem that individual differences in photoelectric conversion output occur due to manufacturing variations of the a-Si semiconductor thin film as described above, the photoelectric conversion device of the present invention amplifies the read circuit portion by an external signal. Since the rate can be controlled, output variations can be easily compensated, and the cost of the apparatus is reduced as a result.

(Embodiment 5)
FIG. 12 is an example of a circuit diagram of one operational amplifier configured in the reading circuit unit of the photoelectric conversion device according to the fifth embodiment of the present invention. In this figure, the feature of the present invention is that a switch element SWp controlled by a signal from the terminal PS is provided. An explanation of the operation related to the function of the switch element SWp will be described below.

  Terminals Vdd and Vss are power supply terminals of the operational amplifier, and a power supply of Vdd> Vss is turned on. Usually, if the GND of the system of the photoelectric conversion device is set to zero potential, + voltage is applied to Vdd and −voltage is applied to Vss, respectively. When the control signal from the PS terminal is not input to the switch element SWp, that is, when SWp is “OFF”, a current flows through the resistor R9, the diode D1, and the diode D2, and the base potential of the transistor Q7 is in the order of D1 and D2. Biased to a potential given by the direction threshold voltage. Then, the transistor Q7 is turned on, and the collector current I of the transistors Q6 and Q7 flows from the Vdd terminal to the resistor R8.

  Since Q6 and Q5, and Q6 and Q8 are in a current mirror configuration relationship, the collector currents of Q6 and Q6 are set to the collector current I of Q6 by making Q5, Q6, and Q8 transistors have the same performance. An equal current flows. Q5 is a constant current source for functioning as an operational amplifier. The bipolar transistors Q1 and Q2 are transistors in the input stage, and a current corresponding to the input differential voltage of the input terminals VIN (+) and VIN (−) flows into (or flows out from) the base of the transistor Q9, and the transistors Q8, Q9, Amplified at the output stage constituted by Q10 and output from the terminal Vout. The transistors Q3 and Q4 form a current mirror and function as an active load for the input stage transistors Q1 and Q2. In actual use, the input terminal VIN (−) terminal is used as a negative feedback circuit to which negative feedback is applied from the output terminal Vout, a non-inverting amplifier circuit, an impedance conversion circuit, a voltage follower circuit, or the like. Also, it is often used as an inverting amplifier circuit.

  Now, in general, when the operational amplifier is configured using a bipolar transistor as shown in FIG. 12, it depends on the resistance value used as the power supply current, but it also exceeds 100 μA in common sense. There are many cases. 3, 6, 8, and 10, when several operational amplifiers are connected to one matrix signal wiring, if a current consumption of 1 mA is required to read one pixel, 4000 × When a photoelectric conversion device having 4000 pixels is read, a power supply current of 1 mA × 4000 (column) = 4 (A) flows. If the power supply voltages of Vdd and Vss are +5 (V) and −5 (V), respectively, the power consumption of 40 (W) is required in the reading circuit unit. 3, 6, 8, and 10, the power is supplied to each operational amplifier even when SR1, SR2, or other switches are not operating, that is, when reading is not performed. , Will always be consumed. This is converted into heat by the reading circuit unit (IC), and the heat is radiated to the surroundings.

  The switch element SWp shown in FIG. 12 is intended to reduce the power consumption except during reading. The operation will be described below. At times other than reading, the switch element SWp is turned on by a control signal from the terminal PS so that no current flows through the diodes D1 and D2. Then, the transistors Q6 and Q7 are turned off, so that no current flows. At the same time, the collector currents of the transistors Q5 and Q8 are cut off. That is, the constant current source in the operational amplifier is cut off by the control signal from the terminal PS, and the current consumption can be greatly reduced. For example, if the switching element SWp is opened and closed by a voltage signal of 0 (V) / 5 (V), it may be configured by a MOS transistor.

  Thus, by providing the operational amplifier in the readout circuit section with a switch for reducing the current consumption as shown in FIG. 12, the photoelectric conversion circuit in which the heat generated in the readout circuit section (IC) is arranged around the switch is provided. Without increasing the temperature of the part, it is possible to reduce an increase in dark current when the TFT which is a switching element is “OFF”, and to reduce fixed pattern noise as a photoelectric conversion device. Needless to say, it is economical to reduce the power consumption of the reading circuit unit (IC) at times other than reading.

(Embodiment 6)
FIG. 13 is a schematic circuit diagram of an A / D conversion circuit unit of a photoelectric conversion device for explaining a sixth embodiment of the present invention. The A / D conversion circuit unit of this embodiment is mainly composed of three operational amplifiers, three A / D converters, two selector circuits, and one bit conversion circuit. The operation will be described below.

  The analog signal Va that has been serially converted by the readout circuit unit is input to three operational amplifiers in the A / D conversion circuit unit. The three operational amplifiers are referred to as amplifier 1, amplifier 2, and amplifier 3, and their amplification factors G1, G2, and G3 are set to a ratio of 1: 2: 4. Each amplification factor is determined by a resistance value connected to the operational amplifier. In the present embodiment, for the sake of explanation, the amplification factors G1, G2, and G3 of the amplifier 1, the amplifier 2, and the amplifier 3 are set to 1, 2, and 4 times, respectively. In addition, the signal Va from the reading circuit unit is output in the range of 0 (V) to 1 (V). That is, a signal exceeding 1 (V) or a negative voltage signal is not input to the A / D conversion circuit unit.

The signal Va from the reading circuit unit is amplified by the amplifier 1, the amplifier 2, and the amplifier 3, and the output of each amplifier is input to the A / D converters AD1, AD2, and AD3. Two reference voltages are input to the terminals REF + and REF− to the A / D converters AD1, AD2, and AD3, and an analog input signal is digitized with respect to a difference voltage between the reference terminals. In this embodiment, a 12-bit A / D converter is used. That is, it is digitized in 2 ¹² = 4096 steps. Two reference voltages of the A / D converter are set to 0 (V) and 4 (V).

  Further, since AD3 is connected to an operational amplifier having a gain of 4 times, A / D conversion is performed when Va is 0 (V) or more and 0.25 (V) or less. When the signal Va exceeds 0.25 (V), a logic signal “Hi” is output from the overflow terminal OF. Since AD2 is connected to an operational amplifier having a double gain, A / D conversion is performed when Va is 0 (V) or more and 0.5 (V) or less. When Va exceeds 0.5 (V), a logic signal “Hi” is output from the overflow terminal OF. Since AD1 is connected to an operational amplifier having a gain of 1 time, A / D conversion is performed when Va is 0 (V) or more and 1 (V) or less. When the signal Va exceeds 1 (V), a logic signal “Hi” is output from the overflow terminal OF terminal.

  Further, if the digital signal from AD3 and AD2 is input to the selector 1 and the OF terminal of AD3 is “Lo”, the digital signal from AD3 is output as it is, and the OF terminal of AD3 is “Hi”. Then, the digital signal from AD2 is output as it is. If the digital signal from selector 1 and AD1 is input to the selector 2 and the OF terminal of AD2 is “Lo”, the digital signal from selector 1 is output as it is, and the OF terminal of AD2 is “Hi”. If so, the digital signal from AD3 is directly output. That is, AD3 is output from the output terminal of selector 2 when Va is 0 (V) to 0.25 (V), and AD2 is output when Va is 0.25 (V) to 0.5 (V). When Va is 0.5 (V) to 1 (V), AD1 is output. The selector 1 and the selector 2 are identical in circuit, and FIG. 14 shows an example of a schematic circuit of the selector 1 in FIG.

  When the OF of each A / D converter is “Lo”, that is, when the signal Va is smaller than 0.25 (V), the digital output of AD1, AD2, and AD3 has an amplifier gain ratio G1: G2: G3 = 1. : 2: 4 ratio. That is, the AD2 digital output shifts the bit position of the AD1 digital output by 1 bit to the MSB side, and the AD3 digital output shifts the bit position of the AD2 digital output by 1 bit to the MSB side. There is a shift.

  For example, if the output of AD1 is {000100101101} from the MSB side, the output of AD2 is {001001011010} and the output of AD3 is {010010110100}.

  FIG. 15 is an example of a schematic circuit of the bit conversion circuit in FIG. In the bit conversion circuit, the input 12-bit digital signal of the selector 2, that is, the digital signal of the selected A / D converter is converted into 14 bits. At that time, a bit shift operation corresponding to the selected A / D converter is performed.

  For example, if the A / D converter AD1 is selected and its output is {10100100101101} from the MSB side, the 14-bit output of the bit conversion circuit is {1010010010110100}, and AD2 is selected and its output is MSB. If {100101001001} from the side, the 14-bit output of the bit conversion circuit is {01001010010010}, and if AD3 is selected and its output is {101010111010} from the MSB side, the 14-bit output of the bit conversion circuit Is {00101010111010}. A desired bit shift operation is performed according to the digital signal of the selected A / D converter by the input signals of the terminals SEL1, SEL2, and SEL3 of the bit conversion circuit unit. The signal can be created with a simple logic circuit by using a signal from the OF terminal of each A / D converter. In FIG. 15, if the terminal SEL1 is “Hi”, bit conversion is performed on the digital output of AD1, and if the terminal SEL2 is “Hi”, bit conversion is performed on the digital output of AD2, and the terminal SEL3 is “ If "Hi", bit conversion is performed on the digital output of AD3.

As a result, the A / D conversion circuit unit of the present embodiment is digitized in 2 ¹² = 4096 steps by AD3 when Va: 0 (V) to 0.25 (V), and Va: 0.25 (V). When .about.0.5 (V), it is digitized to 2 ¹¹ = 2048 steps by AD2, and when Va: 0.5 (V) to 1 (V), it is digitized to 2 ¹¹ = 2048 steps by AD1. That is, an analog signal from the reading circuit unit of Va: 0 (V) to 1 (V) can be divided into 4096 + 2048 + 2048 = 8192 stages and output as a 14-bit digital value. The 14-bit digital output is stored in, for example, a memory, and digital processing is performed using a computer.

In the present embodiment, Va is 0.25 (V) The following signals, the dynamic range: so that the quantization is performed in 1/2 ¹⁴ for 1 (V). That is, a low level signal of 1/4 or less is expressed with high resolution, and is particularly suitable for applications such as a medical X-ray imaging apparatus. Further, since offset components such as fixed pattern noise (FPN) in the dark state caused by the photoelectric conversion circuit section and FPN caused by the readout circuit are digitized with high resolution, when offset correction is performed. The accuracy of correction is improved.

In the present embodiment, the description has been given of the case where the number of operational amplifiers and the number of A / D converters are three in the A / D conversion circuit unit, but a plurality (N) of each may be provided. In addition, although the operational amplifier has been described with an amplification factor of 1, 2 and 4 times, it may not be G1: G2: G3 = 1: 2: 4, and other amplification ratios, for example, 2 times, 4 times, It may be 8 times. In that case, the reference voltage of the A / D converter may be set to 8 (V) according to the amplification factor. If there are N operational amplifiers, the gain ratio of each operational amplifier is 2 ⁰ : 2 ¹ : 2 ² :...: 2 ^N−1 , and ^N A / D converters may be used. In this embodiment, a 12-bit A / D converter is used, but any number of A / D converters may be used.

  As can be seen from the above description, if N operational amplifiers and N M-bit A / D converters are used in the A / D conversion circuit unit, a digital output of M + N−1 bits can be obtained, and a subsequent computer or memory can be obtained. Data can be processed as a digital value of M + N−1 bits in a data processing device using a circuit.

In addition, an analog signal of 1/2 ^N−1 or less with respect to the dynamic range can be digitally converted with substantially the same accuracy as when an M + N−1 bit A / D converter is used. This means that if M + N-1 bit A / D converters do not exist or if they cannot be used due to the conversion speed, N + M-bit A / D converters are used. It means that digital conversion equivalent to 1 bit can be achieved.

(Embodiment 7)
FIG. 16 is a circuit diagram of a photoelectric conversion device showing a seventh embodiment of the present invention. In FIG. 16, the number of pixels of the photoelectric conversion circuit unit 101 is not 3 × 3 pixels, but is described assuming a larger number of pixel configurations. Also, the capacitive elements CL1 to CL3, the switches Sn1 to Sn3, the amplifiers B1 to B3, and the switches Sr1 to Sr3 described in the readout circuit section of FIG. 6 are omitted in FIG. There are 128. Further, the shift register 103, the amplifier 104, and the A / D conversion circuit portion in FIG. 6 are also omitted in FIG.

  In FIG. 16, the number of inputs to the reading circuit unit 107 is described as 128. If the number of columns of the two-dimensional photoelectric conversion element circuit unit 101 is, for example, 2560 columns, 20 readout circuit units 107 (IC) are used. BND1 to BND128 indicate connection portions between the matrix signal wirings (M1 to M128) in the photoelectric conversion circuit portion and the readout circuit portion, and are connected by a wire bonding method or an anisotropic connection method.

  16 differs from FIG. 6 in that potentials (GND) serving as references for the first-stage operational amplifiers L1 to L128 for amplifying signals from the matrix signal wiring are supplied from the buffer amplifiers E1 to E128, respectively. Is a point. As described in the second embodiment, the operational amplifiers L1 to L128 are operational amplifiers that are intended to amplify signals from the photoelectric conversion circuit unit and have excellent characteristics in terms of noise performance. At the same time, when amplifying with the operational amplifier, in addition to the random noise generated with the operational amplifiers L1 to L128, there is also thermal noise generated with a resistor configured as a non-inverting amplifier. In particular, thermal noise (4KTRB) generated in the input resistors RA1 to RA128 inserted between the inverting terminal of the operational amplifier and GND results in amplification by the amplification factor of the non-inverting amplifier by the operational amplifiers L1 to L128. Therefore, it is required to reduce the input resistances of L1 to L128 in order to further reduce the thermal noise generated by the resistance.

  On the other hand, when signals from the photoelectric conversion circuit unit are input to the operational amplifiers L1 to L128, a current corresponding to the input voltage flows through the input resistors RA1 to RA128. For example, if the output voltage of the matrix signal wiring M1, that is, the input voltage of the operational amplifier L1, is V1, the current I1 flowing through the input resistor RA1 is I1 = V1 / RA1. That is, if the input resistance is reduced, the thermal noise is reduced and the current flowing through the input resistance is increased. The current flows through GND. If the impedance of GND is large, a voltage drop is caused by the current flowing through the input resistance. For example, if the number of GND supplied from the outside to the reading circuit unit 107 is one point, the GND is supplied from the point to the operational amplifiers L1 to L128.

  In the GND wiring, when the number of inputs is 128 as shown in FIG. 16, all the currents flowing through the 128 input resistors flow, and the operational amplifier existing far away from the GND supply point. The reference potential (GND) will fluctuate. In addition, the amount of fluctuation depends on the input signal of another signal wiring, and there is a possibility that a correct photoelectric conversion signal cannot be obtained. Although the amount of voltage drop of the reference potential can be reduced by increasing the line width of the GND wiring to be routed, this leads to an increase in the chip area and is not a desirable solution.

  Although the impedance can be lowered by supplying external GND to each of the operational amplifiers L1 to L128, it is not practical to provide the same number of lead-out pads as the number of inputs.

  In the present embodiment, separate buffer amplifiers E1 to E128 are provided for the operational amplifiers L1 to L128, respectively, and the reference potentials (GND) of the low noise amplifiers L1 to L128 are supplied from the outputs of the buffer amplifiers E1 to E128, respectively. . By doing so, even if the input resistances RA1 to RA128 of the operational amplifiers L1 to L128 are set small and as a result, the current flowing through the input resistance increases, the reference of the low noise operational amplifiers L1 to L128 always has a good GND potential. Since it can be supplied, an accurate photoelectric conversion signal can be obtained. Of course, since the input resistors RA1 to RA128 are small, the thermal noise generated by the input resistors is small and the S / N can be increased. The resistance value of the input resistor does not need to be unnecessarily small, and may be set in consideration of thermal noise generated by the resistor and noise generated by the buffer amplifiers E1 to E128.

  In this embodiment, the number of inputs to the reading circuit unit is described as 128. However, the number of inputs may be further increased.

(Embodiment 8)
FIG. 17 is a circuit diagram of a photoelectric conversion device for explaining an eighth embodiment of the present invention. The same members as those in FIG. 8 are denoted by the same reference numerals, and description thereof is omitted.

  17 differs from FIG. 8 in the following four points. First, the capacitive elements CC1 to CC3 are connected to the switch elements D1 to D3 via the resistance elements RB1 to RB3, respectively. Secondly, a signal for opening and closing the switching elements D1 to D3, that is, a DRES signal is created by the CRES signal and the delay circuit DL1 (or DL2), and is made selectable by an external control signal DSEL. Thirdly, a low-pass filter including buffer amplifiers F1 to F3, resistance elements RF1 to RF3, and capacitive elements CF1 to CF3 is provided between the capacitive elements CC1 to CC3 and the buffer amplifiers A1 to A3. The slew rate of the buffer amplifiers A1 to A3 that function when sampling the signals to the capacitive elements CL1 to CL3 is made variable by the control signal SR from the outside.

  Although the photoelectric conversion device of the present embodiment is described with 3 × 3 = 9 pixels, there is no difference in the main point even with a larger number of pixels. In addition, since there is no problem even if only one matrix signal wiring is described, this embodiment will be described with reference to the drawings only for the M1 matrix signal wiring.

  The operational amplifier L1 in the first stage of the reading circuit unit 107 for amplifying the signal from the photoelectric conversion circuit unit 101 is excellent in low noise performance as described in the second embodiment. The frequency band may be a band that can sufficiently amplify the photoelectric conversion signal transmitted by the transfer operation in the TFT in the photoelectric conversion circuit unit. However, when the operational amplifier L1 has a frequency band wider than necessary, the photoelectric conversion signal can be amplified, but the high-frequency component of random noise generated in L1 is also amplified. The noise of the high-frequency component appears at the output when the photoelectric conversion circuit unit is reset, and as a result, is terminated in the capacitive element CC1. This causes a loss of S / N.

  In addition, in the transfer of the photoelectric conversion signal by the TFT, the high frequency component of the noise of L1 is amplified and the S / N is also lowered. In other words, the performance required for the operational amplifier of L1 is a frequency band that can sufficiently amplify the photoelectric conversion signal transmitted by the transfer operation in the TFT in the photoelectric conversion circuit section, and a frequency band wider than necessary. Must not. However, when actually designing and manufacturing an operational amplifier, it is not easy to obtain an operational amplifier having a desired band with a simple circuit configuration. In particular, in the readout circuit of the present invention in which a plurality of operational amplifiers need to be formed on one chip, there is a possibility that it may vary several times although it does not vary up to one digit. Therefore, the operational amplifier L1 is designed in a wider frequency band in consideration of variation than the frequency band in which the photoelectric conversion signal transmitted by the transfer operation in the TFT in the photoelectric conversion circuit unit can be amplified.

  FIG. 18 shows a timing chart of the CRES signal and delay circuit DL1 output and DL2 output. The delay circuits DL1 and DL2 are circuits that cause a delay only at the falling edge of the CRES signal. The delay amount of DL2 is larger than that of DL1. A DRES signal having a different delay amount can be selected by an external control signal. When the CRES signal is ON, that is, when the readout capacitance of the matrix signal wiring in the photoelectric conversion circuit section is reset, the DRES signal is simultaneously ON. At that time, the noise generated in the operational amplifier L1 is limited by a primary low-pass filter (LPF) having a cutoff frequency fc = 1 / (2 · π · CC1 · RB1) determined by the capacitive element CC1 and the resistor RB1. Will be. Therefore, even if the frequency band of the operational amplifier L1 is set somewhat wider, the frequency band is effectively limited by the insertion of the resistor RB1, and noise on the high frequency side of L1 can be cut off.

  The buffer amplifier F1, the capacitive element CF1, and the resistive element RF1 function as a primary LPF when a photoelectric conversion signal is transferred via the TFT after the reset is completed. That is, the noise on the high frequency side of L1 can be blocked. The band of the buffer amplifier F1 is set to a slightly high band, and the constants of CF1 and RF1 are in a frequency band that can sufficiently amplify the photoelectric conversion signal transmitted by the transfer operation in the TFT in the photoelectric conversion circuit section, and more than necessary. A constant that does not result in a wide frequency band should be selected.

  The delay time of the fall of the DRES signal by the delay circuits DL1 and DL2 is controlled by the DSEL signal. Let Td be the delay time of the fall of the DRES signal. As described in the third embodiment, the capacitive element CC1 is AC-coupled and the timing operation shown in FIG. 6 is performed, so that the KTC noise generated when the readout capacitance (C2) of the matrix signal wiring is reset by CRES. Can be canceled. The amount of KTC noise VKTC to be clamped to CC1 by inserting the resistance element RB1 is given by VKTC = VT (1−EXP (−Td / CC1 · RB1), where the total amount of KTC noise is VT. By setting Td sufficiently long compared to CC1 and RB1, KTC noise can be stored sufficiently in CC1, and KTC noise can be canceled at the time of TFT transfer. This also means that the time required to read one row also increases, which means that the time to read photoelectric conversion signals for one frame becomes longer, for example, using the photoelectric conversion device of the present invention as an X-ray imaging device. Assuming that a high quality still image for one shot is obtained, Td may be set longer. Time increase video acquisition rate reading when that: leading to (frame rate number / sec) be reduced.

  In this embodiment, Td can be switched by the DSEL signal. As a result, the DRES signal with a long Td is used in the still image mode that requires a high image quality with a high S / N, and the Td is set short in the video mode with a high frame rate. The operation state can be easily switched by the DRES signal for each purpose such as using the DRES signal.

  Although the shift register 103 in the reading circuit is not shown in FIG. 17, when the circuit system is configured to output in synchronization with the basic clock, the photoelectric conversion signal (analog signal) serially converted by the 103 is the basic clock. Output in sync with. For example, when the basic clock is set to 10 MHz, the analog signal output current can be designed to 10 MHz. However, for example, due to restrictions on hardware such as the A / D conversion circuit unit 105, a memory connected in the subsequent stage, a system circuit including a CPU and other digital circuits, and the like in terms of software, an analog output current rate of 10 MHz is It may not be accepted. In such a case, for example, the basic clock may be operated at 5 MHz.

  In the present embodiment, the delay amount Td can be made constant at the operation timing for canceling the KTC noise. That is, the delay circuits DL1 and DL2 need only be created using a basic clock, and can be realized with a simple digital circuit.

  For example, DL1 may be delayed by 16 clocks, and DL2 may be delayed by 32 clocks. When DMH is set to “Hi” and DL2 is used at 10 MHz, Td = 3.2 (μsec) when DL2 is used. When DSEL is set to “Lo” and DL1 is used at 5 MHz, Td = 3.2 (μsec). That is, Td can be made constant by switching DSEL. The capacitance value of CC1 and the resistance value of RB1 may be appropriately set in consideration of the KTC noise, the band of the operational amplifier L1, and the reading speed, and the DRES signal can be switched according to the purpose.

  FIG. 19 shows an example of a specific circuit configuration example of the part forming the sample hold unit in FIG. 17, that is, the region of the buffer amplifier A1, the switch element Sn1, the capacitive element CL1, and the buffer amplifier B1. The value of the current flowing through the transistors Q16 and Q15 can be changed depending on whether the external control signal SR is on or off. This changes the amount of charge that can be charged to CL per unit time when the SMPL signal is on, that is, when the photoelectric conversion signal from the photoelectric conversion circuit section is sampled in the capacitor element CL. That is, SR changes the slew rate (V / μsec) of the buffer amplifier A1.

The noise generated in the buffer amplifier A1 is the inverse 1 / G (G: 1 + R2 / R1) of the gain of the amplifier L1 when converted at the input section of the reading circuit section. Like L1 and other amplifiers, it depends on the noise density (V / (Hz) ^1/2 ) and frequency band of the buffer amplifier A1. If the amplifier SR signal of the buffer amplifier A1 is set to “Lo”, the slew rate is lowered. In other words, it also means reducing the frequency band of the buffer amplifier A1. This means that, when the photoelectric conversion signal is sampled with the SMPL signal “Hi”, the noise of the buffer amplifier A1 accumulated in the capacitor element CL can be reduced, which is advantageous in terms of S / N. For example, when the signal from the photoelectric conversion circuit section is weak, the control signal SR is set to “Lo” to reduce the slew rate of A1, thereby reducing noise. For example, as described above, when the photoelectric conversion device is used as an X-ray imaging device, the amount of signal obtained is generally weak in the moving image mode because the X-ray dose is limited, and the switching function based on the control signal SR is effective. It becomes.

  In the present embodiment, the switching by the DSEL and SR control signals has been described only with “Hi” and “Lo” switching functions, but four, eight, sixteen,... Needless to say, it can easily be increased.

(Embodiment 9)
FIG. 20 is a schematic circuit diagram of a photoelectric conversion device showing a ninth embodiment of the present invention. In order to simplify the explanation, only 3 × 3 = 9 pixels are shown. The notation of photoelectric conversion elements S1-1 to S3-3 is different from FIG. Further, the power supply circuit unit for applying a bias to the photoelectric conversion element is different. The readout circuit portion is the same as that in FIG. 3, and the same reference numerals are used for the same constituent members. The method for creating the photoelectric conversion circuit portion is described in Embodiment 1. Therefore, a schematic top view and a schematic cross-sectional view of a photoelectric conversion element, a switching element (TFT), and the like are the same as FIGS. 5A and 5B, respectively.

As can be seen from FIGS. 5A and 5B, the photoelectric conversion element has the same layer configuration as the switching element, and is configured as a MIS type capacitor. However, the N ⁺ layer is used as the upper electrode of the photoelectric conversion element, which is different from a normal MIS capacitor for the purpose of entering light. The photoelectric conversion element is also a capacitive element, and the signal charge that has been subjected to photoelectric conversion is stored in an apparent capacity. In the present embodiment, an operation method such as accumulation of photoelectric conversion charge, transfer by TFT, and signal readout after resetting using a bias circuit provided with a photoelectric conversion element that is also a capacitor outside will be described. The reset operation of the above-described photoelectric conversion element is hereinafter referred to as “refresh”. Moreover, although the description of photoelectric conversion element S1-1-S3-3 is changed with FIG. 3, the 1st metal thin film layer said in FIG. 5 (a) and FIG.5 (b) is used for a photoelectric conversion element. The “G” electrode and the second metal thin film layer are referred to as “D” electrodes. However, the D electrode functions as an electrode including the N ⁺ layer as described above for the photoelectric conversion elements S1-1 to S3-3.

  First, device operation of a single photoelectric conversion element will be described. 22A to 22C are energy band diagrams for explaining the device operation.

FIG. 22A and FIG. 22B show the operation of the refresh mode and the photoelectric conversion mode of this embodiment, respectively, and the film of each layer of the photoelectric conversion element of FIG. 5A or FIG. The state in the thickness direction is shown. M1 is a lower electrode (G electrode) formed of the first metal thin film layer (Cr). The a-SiNx layer is an insulating layer that blocks the passage of both electrons and holes. The a-SiNx layer needs to have a thickness that does not cause a tunnel effect, and is set to 500 angstroms or more. a-Si is a photoelectric conversion semiconductor layer formed of an intrinsic semiconductor i layer. The N ⁺ layer is an N-type a-Si layer injection blocking layer formed to prevent hole injection into the a-Si layer. M2 is an upper electrode (D electrode) formed of the second metal thin film layer (A1).

In the present embodiment, D electrode is not completely cover the N ⁺ layer, the D electrode and the N ⁺ layer for between the the electron transfer is freely performed between the D electrode and the N ⁺ layer is always the same potential Yes, the following explanation assumes that.

  This photoelectric conversion element has two types of operation modes, a refresh mode and a photoelectric conversion mode, depending on how the voltages of the D electrode and G electrode are applied.

In FIG. 22A in the refresh mode, the D electrode is given a negative potential with respect to the G electrode, and the holes indicated by black circles in the i layer are guided to the D electrode by an electric field. At the same time, electrons indicated by white circles are injected into the i layer. At this time, some holes and electrons recombine and disappear in the N ⁺ layer and the i layer. If this state continues for a sufficiently long time, the holes in the i layer are swept out of the i layer.

In order to change from this state to FIG. 22B of the photoelectric conversion mode, a positive potential is applied to the D electrode with respect to the G electrode. Then, the electrons in the i layer are instantaneously guided to the D electrode. However, since the N ⁺ layer functions as an injection blocking layer, holes are not led to the i layer. When light is incident on the i layer in this state, the light is absorbed and an electron hole pair is generated. The electrons are guided to the D electrode by an electric field, and the holes move in the i layer and reach the interface between the i layer and the a-SiNx insulating layer. However, since it cannot move into the insulating layer, it remains in the i layer. At this time, electrons move to the D electrode and holes move to the interface of the insulating layer in the i layer, so that a current flows from the G electrode in order to maintain electrical neutrality in the photoelectric conversion element. This current is proportional to the incident light because it corresponds to the electron-hole pair generated by the light.

  After maintaining the photoelectric conversion mode in FIG. 22B for a certain period and then entering the refresh mode in FIG. 22A again, the holes remaining in the i layer are led to the D electrode as described above, and at the same time A current corresponding to this hole flows. The amount of holes corresponds to the layer amount of light incident during the photoelectric conversion mode period. At this time, a current corresponding to the amount of electrons injected into the i layer also flows, but since this amount is approximately constant, it may be detected by subtracting. That is, in this embodiment, the photoelectric conversion element outputs the amount of light incident in real time, and at the same time, can detect the total amount of light incident in a certain period.

  However, when the period of the photoelectric conversion mode becomes long for some reason or when the illuminance of incident light is strong, current may not flow even though light is incident. This is because a large number of holes remain in the i layer as shown in FIG. 22C, and the electric field in the i layer is reduced due to these holes, and the generated electrons are not guided and recombined with holes in the i layer. Because. If the light incident state changes in this state, the current may flow in an unstable manner. However, if the refresh mode is set again, the holes in the i layer are swept away and the current proportional to the light is again used in the next photoelectric conversion mode. Flows.

In the above description, when sweeping out the holes in the i layer in the refresh mode, it is ideal to sweep out all the holes, but it is effective only by sweeping out some holes, and the same current as above can be obtained. ,No problem. In other words, it is sufficient that the detection opportunity in the next photoelectric conversion mode is not in the state shown in FIG. 22C, the potential of the D electrode with respect to the G electrode in the refresh mode, the period of the refresh mode, and the N ⁺ layer injection blocking layer. What is necessary is just to decide the characteristic. Further, in the refresh mode, injection of electrons into the i layer is not a necessary condition, and the potential of the D electrode with respect to the G electrode is not limited to negative. This is because when many holes remain in the i layer, the electric field in the i layer is applied in the direction leading the holes to the D electrode even if the potential of the D electrode with respect to the G electrode is a positive potential. Similarly, the characteristics of the N ⁺ layer injection blocking layer are not necessarily required to be able to inject electrons into the i layer.

  Next, an example of operation of the photoelectric conversion device in FIG. 20 will be described with reference to a timing chart in FIG. The control signal VSC is for giving two types of bias to the bias line REF of the photoelectric conversion element, that is, the D electrode of the photoelectric conversion element. The D electrode becomes VREF (V) when VSC is “Hi”, and becomes VS (V) when “Lo”. Reference numerals 106A and 106B denote DC power supplies, which are a read power supply VS (V) and a refresh power supply VREF (V), respectively.

  First, the operation during the refresh period will be described. All the signals of the shift register 102 are set to “Hi”, and the CRES signal of the reading circuit portion is set to the “Hi” state. Then, all the switching TFTs (T1-1 to T3-3) are turned on, and the switch elements RES1 to RES3 in the reading circuit are also turned on, so that the G electrodes of all the photoelectric conversion elements become the GND potential. When the VSC signal becomes “Hi”, the D electrodes of all the photoelectric conversion elements are biased to the refresh power source VREF (negative potential). Then, all the photoelectric conversion elements S1-1 to S3-3 are in the refresh mode, and refresh is performed.

  Next, the photoelectric conversion period will be described. The VSC is switched to the “Lo” state, and the D electrodes of all the photoelectric conversion elements are biased to the reading power source VS (positive potential). Then, the photoelectric conversion element enters a photoelectric conversion mode. In this state, all the signals of the shift register 102 are set to “Lo”, and the CRES signal of the reading circuit portion is set to “Lo”. Then, all the switching TFTs (T1-1 to T3-3) are turned off, and the switch elements RES1 to RES3 in the readout circuit are also turned off, and the G electrodes of all the photoelectric conversion elements are in an open state in terms of DC. However, since the photoelectric conversion element is also a capacitor, the potential is maintained. However, at this time, since no light is incident on the photoelectric conversion element, no charge is generated. That is, no current flows.

When the light source is turned on in a pulse state in this state, light is irradiated to the D electrode (N ⁺ electrode) of each photoelectric conversion element, and so-called photocurrent flows. The light source is not particularly described in FIG. 20, but for example, in the case of a copying machine, it is a fluorescent lamp, an LED, a halogen lamp, or the like. An X-ray imaging apparatus is literally an X-ray source. In this case, a scintillator for X-ray visible conversion may be used. The photocurrents flown by these lights are accumulated in the respective photoelectric conversion elements as electric charges, and are retained even after the light source is turned off.

  Next, the reading period starts. The operation is exactly the same as that described in the first embodiment, and the description is omitted here.

  One image is obtained through the refresh period, the photoelectric conversion period, and the readout period. However, when a plurality of images such as a moving image are obtained, the above operation may be repeated. In this embodiment, the D electrodes of the photoelectric conversion elements are connected in common, and this common wiring is controlled to the potentials of the refresh power supply VREF and the read power supply VS by the VSC signal. The mode can be switched between the refresh mode and the photoelectric conversion mode. For this reason, a light output can be obtained with one TFT per pixel without complicated control.

(Embodiment 10)
FIG. 23 is a schematic circuit diagram of a photoelectric conversion device for explaining a tenth embodiment of the present invention. In FIG. 23, a set of photoelectric conversion elements and switching TFTs is simply represented by a rectangle for the sake of simplicity. The feature of this embodiment is that four power supply circuits for switching between the refresh power supply and the read power supply are provided, and in addition, four reset circuits for resetting the matrix signal wiring are provided. The power supply is switched by VSC1 to VSC4, and the reset switching is performed by CRES1 to CRES4. The portions corresponding to the operational amplifiers A1 to A3 in the readout circuit 107 in FIG. 20 are the operational amplifiers A1 to A3 in FIG. For simplification of the drawing, the circuit portion after the operational amplifier A1 is omitted, but may be considered to be exactly the same as FIG. Further, in FIG. 23, it is assumed that the total number of pixels is not 3 × 3 = 9 pixels but a larger number of pixels. Further, the number of columns in FIG. 23 is assumed to be 4 × N times (N: natural number).

  Next, an example of the operation of the present embodiment will be described with reference to the timing chart of FIG.

  When the G1 signal of the shift register (SR1) becomes “Hi” when the power supply switching VSC1 signal is in the “Hi” state and the reset switching CRES1 signal is also in the “Hi” state, the photoelectric conversion elements S1-1 and S1- 5, S1-9..., That is, the photoelectric conversion elements corresponding to the (4 × 1 + 1) th column (an integer of 1: 0 or more) in the first row are refreshed. Similarly, when G2, G3,... GN sequentially become “Hi” and the shift register operation is completed, the first column, the fifth column, the ninth column, that is, the (4 × 1 + 1) th column (1: 0 or more) All the photoelectric conversion elements corresponding to (integer) complete the refresh. During this period, the other rows of photoelectric conversion elements, ie, 2, 6, 10,..., 3, 7, 7, 11,..., 4, 8, 12,. VSC2, VSC3, and VSC4 corresponding to the column are in the “Lo” state (D electrode is biased to VS), and the signals of CRES2, CRES3, and CRES4 repeat the normal reset operation (the CRES signal in FIG. 4). Therefore, the reading operation by the reading circuit unit is being performed.

  Next, with the VSC1 transitioning from the “Hi” state to the “Lo” state and the CRES1 signal maintained in the “Hi” state, the shift register SR1 is operated from normal G1 to GN. When the shift register operation is completed, all the D electrodes of the photoelectric conversion elements corresponding to the first column, the fifth column, the ninth column, that is, the (4 × 1 + 1) th column (an integer of 1: 0 or more) are read power supply VS. Thus, the G electrode is in a state of holding the GND potential. That is, it has shifted to the photoelectric conversion mode. On the other hand, the photoelectric conversion elements in the other columns, that is, the photoelectric conversion elements of 2, 6, 10,..., 3, 7, 7, 11,. , VSC2, VSC3, and VSC4 are in the “Lo” state (D electrode is biased to VS), and the signals of CRES2, CRES3, and CRES4 repeat the normal reset operation (similar to the CRES signal of FIG. 4). In this state, the reading circuit unit is performing a reading operation.

  That is, in the process of refreshing the photoelectric conversion elements corresponding to the (4 × 1 + 1) -th column (an integer of 1: 0 or more) in a certain period and performing the operation of shifting to the photoelectric conversion mode in the next period, That is, the photoelectric conversion elements in the column are continuously performing the reading operation twice. As shown in FIG. 24, a series of these operations is performed by the (4 × 1 + 1) th column, (4 × 1 + 2) column, (4 × 1 + 3) column, (4 × 1 + 4) column (an integer greater than or equal to 1: 0) 4) is repeated while shifting the phase.

  If the light source is turned on / off at the timing as shown in FIG. 24, the three-line photoelectric conversion elements that perform the reading operation twice in succession will output dark state information in the first frame, Information on the state of light irradiation in the next frame is output. Although not shown in FIG. 23, dark state information (fixed pattern noise: FPN) can be corrected by subtracting them in the digital processing circuits after the A / D conversion circuit section in the subsequent stage. In addition, by interpolating the data of the photoelectric conversion elements in the column being refreshed with the data of the photoelectric conversion elements performing the reading operation on the adjacent columns, image data that is continuous in time series in the moving image mode Can be obtained.

  Performing the refresh operation separately for the four systems as described in the present embodiment is particularly effective when obtaining continuous moving image images, and can substantially increase the frame rate of moving images.

  The shift operation is performed so that the four stages are simultaneously turned on, such as turning on G1 to G4 simultaneously, then turning on G5 to G8, and then turning on G9 to G12 simultaneously. By doing so, average information of four adjacent photoelectric conversion elements in the row direction is obtained, and the scanning speed is reduced to ¼. Further, in the column direction, by performing averaging processing of pixel data of three systems (for three columns) that are not refreshed in the subsequent digital processing circuit, for example, this photoelectric conversion device is configured in 2000 rows × 2000 columns. Even in this case, the data can be compressed to 500 rows × 500 columns, and can be displayed on a commercially available inexpensive CRT. In this case, 4 × 4 = 16 pixels become one pixel in a frame that is convenient for refreshing, and the center of gravity of the pixel is shifted. However, since this is averaged in units of 4 frames, there is no particular problem.

  In this embodiment, the refresh has been described with four systems, but there is no particular need for four systems, and a plurality of systems may be set according to the purpose.

  As described above in detail, according to the present invention, it is possible to shorten the readout scanning time, and provide a photoelectric conversion device capable of high-speed reading and a photoelectric conversion system having the device.

  In addition, according to the present invention, it is possible to provide a photoelectric conversion apparatus that can read a high S / N signal with less generation of thermal noise (KTC noise) and a system including the apparatus.

  In addition, according to the present invention, it is possible to provide a photoelectric conversion device capable of obtaining good image information with reduced fixed pattern noise and free of uneven density and unnecessary stripes, and a photoelectric conversion system having the device. it can.

  Furthermore, according to the present invention, it is possible to provide a photoelectric conversion device capable of obtaining image information with excellent gradation and a photoelectric conversion system having the device.

  In addition, according to the present invention, it is possible to easily compensate for non-uniform characteristics due to variations in manufacturing of photoelectric conversion elements and the like, and to promote cost reduction and the photoelectric conversion device A system having the apparatus can be provided.

  It should be noted that the present invention is not limited to the various examples described in the above embodiment, and it goes without saying that the present invention can be appropriately modified and / or combined within the scope of the gist of the present invention.

It is a schematic circuit diagram for demonstrating an example of a photoelectric conversion apparatus. 6 is a timing chart for explaining an example of a driving method of the photoelectric conversion device. It is a schematic circuit diagram for demonstrating an example of a photoelectric conversion apparatus. 6 is a timing chart for explaining an example of a driving method of the photoelectric conversion device. It is a typical top view for explaining an example of a photoelectric conversion element. It is a schematic circuit diagram for demonstrating an example of a photoelectric conversion apparatus. It is a schematic block diagram for demonstrating an example of the apparatus which has a photoelectric conversion apparatus. It is a schematic circuit diagram for demonstrating an example of a photoelectric conversion apparatus. 6 is a timing chart for explaining an example of a driving method of the photoelectric conversion device. It is a schematic circuit diagram for demonstrating an example of a photoelectric conversion apparatus. It is a schematic circuit diagram for demonstrating an example of an operational amplifier. It is a schematic circuit diagram for demonstrating an example of an operational amplifier. It is a schematic circuit diagram for explaining an example of an A / D conversion circuit. It is a schematic circuit diagram for demonstrating an example of a selector circuit part. It is a schematic circuit diagram for demonstrating an example of a bit conversion circuit part. It is a schematic circuit diagram for demonstrating an example of a photoelectric conversion apparatus. It is a schematic circuit diagram for demonstrating an example of a photoelectric conversion apparatus. 18 is a timing chart for explaining an example of input and output to the delay circuit shown in FIG. It is a schematic circuit diagram for demonstrating an example of a sample hold circuit. It is a schematic circuit diagram for demonstrating an example of a photoelectric conversion apparatus. 6 is a timing chart for explaining an example of a driving method of the photoelectric conversion device. It is a schematic energy band diagram for demonstrating an example of the drive method of a photoelectric conversion apparatus. It is a schematic circuit diagram for demonstrating an example of a photoelectric conversion apparatus. 6 is a timing chart for explaining an example of a driving method of the photoelectric conversion device.

Explanation of symbols

S1-1 to S3-3 Photoelectric conversion elements T1-1 to T3-3 Switching element SR1 Shift register (for switching elements)
SR2 Shift register (for readout switch)
G1 to G3 Gate drive wiring M1 to M3 Matrix signal wiring 101 Photoelectric conversion circuit unit 104 Buffer amplifier 105 A / D conversion circuit unit 106 Photoelectric conversion element bias power supply 107 Read circuit unit R
ES1 to RES3 Switches A1 to A3 for resetting load capacitors formed in M1 to M3 Buffer amplifiers B1 to B3 Buffer amplifiers R1 to R10 Resistors CL1 to CL3 Read capacitors Sn1 to Sn3 Transfer switch Sr1 for transferring signals to the read capacitors ˜Sr3 Read switch 301 for sequentially reading the read capacitance signal Photoelectric conversion element 302 Switching element (TFT)
306 Gate drive wiring 307 Matrix signal wiring 310 Contact hole portion 314 Wiring cross portion 304 First metal thin film layer 305 Second metal thin film layer 311 a-SiN insulating thin film layer 312 a-Si semiconductor thin film layer 313 N + layer 303 Insulation Substrate 315 Protective films L1 to L3 Low noise amplifiers CC1 to CC3 AC coupling capacitors D1 to D3 DC regeneration switches K1 to K3 Variable gain operational amplifiers S1 to S3 Switches Q1 to Q10 for switching gains of K1 to K3 Bipolar transistors D1 and D2 Diode SWp Switch for controlling current consumption of operational amplifier 1501 X-ray source 1502 Human body 1503 Grid 1507 Material for absorbing X-ray 1508 Material for transmitting X-ray 1504 Phosphor 1511 for converting X-ray to visible light Protective film 1509 Photoelectric conversion element 15 0 switching element 1512 insulating substrate

Claims (11)

A plurality of pixels including a photoelectric conversion element for converting light into an electric charge and a switch element for outputting an electric signal corresponding to the electric charge are arranged two-dimensionally, and the electric signal is output from the plural pixels in a row unit. A photoelectric conversion circuit unit for outputting as a parallel signal;
A driving circuit unit for controlling conduction of the switch elements in units of rows;
A read circuit unit comprising: a sample hold unit for sample-holding the parallel signal; and a read switch for sequentially reading the parallel signal from the sample hold unit as a serial signal;
In a photoelectric conversion device having
The photoelectric conversion device, wherein the driving circuit unit conducts the switch elements in a row different from the predetermined row within a time when the reading circuit unit performs the reading of the predetermined row. The photoelectric conversion circuit unit outputs the parallel signal through a plurality of signal wirings in units of rows,
The sample and hold unit includes a capacitor provided for each signal wiring, and a transfer switch for transferring the parallel signal to the capacitor.
The time is included in a period of time from when the sample-and-hold circuit samples and holds a parallel signal in a predetermined row to when it samples and holds a parallel signal in a row different from the predetermined row. Photoelectric conversion device.   The switch further includes a reset switch for resetting the signal wiring to a constant potential, and the switching element in a row different from the predetermined row is turned on after the sample and hold circuit samples and holds the parallel signal in the predetermined row. The photoelectric conversion device according to claim 2, wherein the reset is performed during the operation.   The photoelectric conversion device according to claim 2, wherein the readout circuit unit further includes at least one analog operational amplifier between the signal wiring and the transfer switch.   The readout circuit unit includes a capacitive element connected in series between the analog operational amplifier and the transfer switch and to an output terminal of the analog operational amplifier, and a reset switch for resetting the capacitive element; The photoelectric conversion device according to claim 4, further comprising: a resistance element provided between the capacitance element and the reset switch.   The photoelectric conversion device according to claim 5, wherein the readout circuit unit further includes a low-pass filter circuit connected to a terminal opposite to a terminal connected to an output of the analog operational amplifier among terminals of the capacitive element.   The photoelectric conversion apparatus according to claim 4, wherein the analog operational amplifier has a function of changing a slew rate according to an external signal.   The photoelectric conversion device according to claim 1, wherein the switching element is a thin film transistor having an amorphous silicon semiconductor layer. The photoelectric conversion device according to any one of claims 1 to 8,
An X-ray imaging device comprising: a wavelength converter disposed on a light incident side of the photoelectric conversion device.   A system comprising: the X-ray imaging apparatus according to claim 9; and an X-ray source for irradiating the X-ray imaging apparatus with X-rays. A plurality of pixels including a photoelectric conversion element for converting light into an electric charge and a switch element for outputting an electric signal corresponding to the electric charge are arranged two-dimensionally, and the electric signal is output from the plural pixels in a row unit. A photoelectric conversion circuit unit for outputting as a parallel signal;
A driving circuit unit for controlling conduction of the switch elements in units of rows;
A read circuit unit comprising: a sample hold unit for sample-holding the parallel signal; and a read switch for sequentially reading the parallel signal from the sample hold unit as a serial signal;
In a photoelectric conversion device having
The photoelectric conversion device according to claim 1, wherein the readout of the predetermined row by the readout circuit unit and the conduction of the switch element in a row different from the predetermined row are performed in a time-overlapping manner.