Solid-state image sensor and imaging device
The solid-state image sensor addresses the challenge of large pixel size in traditional sensors by incorporating a reduced pixel cell design with a memory capacitor and counting circuit, enabling high integration and accurate photon detection for low light imaging applications.
Patent Information
- Application Number
- DE112018002674
- Authority / Receiving Office
- DE · DE
- Patent Type
- Patents
- Current Assignee / Owner
- Priority Date
- 2017-05-25
- Filing Date
- 2018-04-17
- Publication Date
- 2024-11-28
- Estimated Expiration
- 2038-04-17
AI Technical Summary
Traditional photo-counting solid-state image sensors are not suitable for high integration due to their large pixel cell size, which limits the number of pixels and hinders high-speed imaging capabilities, especially in applications requiring low light detection.
A solid-state image sensor design with a pixel cell configuration that includes an avalanche photodiode, memory capacitor, and transistors, allowing for reduced pixel size and high integration, featuring a memory with a laminate structure and a capacitor with electrodes and insulating layers, and a counting circuit that minimizes error in photon counting.
The design enables a multi-pixel semiconductor image sensor with a powerful photo-counting function and reduced pixel size, facilitating high integration and accurate photon detection, suitable for real-time low light imaging.
Smart Images

Figure 00000000_0000_ABST
Abstract
Description
[0001] The present disclosure relates to solid-state image sensors and in particular to a solid-state image sensor that detects weak light.
[0002] Solid-state image sensors that detect low light and have a photocounting function are known in the relevant technology (see, for example, JP H07-67043A and JP S61-152176A and DUTTON, Neale AW [et al.]: 9.8μm SPAD-based Analogue Single Photon Counting Pixel with Bias Controlled Sensitivity. In: Paper presented at International Image Sensor Workshop, 2013, pp. 1-4).
[0003] US 2009 / 0 225 210 A1 relates to a solid-state image sensor which divides a saturated charge and collects it in a storage capacitor.
[0004] From US patent 2016 / 0344965A1, a single-pixel detector is known that uses an anti-blooming transistor. The transistor has three main objectives, one of which is to dissipate excess electrical signal generated in the photodiode.
[0005] In JP 2014-60 519 A, a solid-state image sensor device is disclosed which performs analog-to-digital conversion of signals during the exposure time by carrying out pixel charge accumulation exclusively using photodiodes during the exposure time.
[0006] DUTTON, Neale AW [et al.]: 9.8μm SPAD-based Analogue Single Photon Counting Pixel with Bias Controlled Sensitivity. In: Paper presented at International Image Sensor Workshop, 2013, pp. 1-4, describes an image sensor with a variable photodiode voltage. The effects of the applied voltage are correlated with the sensor's sensitivity and error rate. The device can be operated in two modes: CTA and SCS.
[0007] US 2008 / 0 258 045 A1 describes another solid-state image sensor.
[0008] Unfortunately, traditional solid-state image sensors with photo-counting functionality are not always suitable for high integration, as their pixel cells are relatively large.
[0009] One object of the present disclosure is the provision of a solid-state image sensor having a photo-counting function and being better suited for high integration than that in the corresponding art, and an imaging device.
[0010] The problem is solved by a solid-state image sensor according to the independent claim.
[0011] Preferably, the storage device has a capacity that is at least five times greater than the capacity of floating diffusion.
[0012] Preferably, the storage device is a capacitor with a laminate structure.
[0013] Preferably, the storage device is a capacitor with an electrode, a semiconductor layer and an insulating layer arranged between the electrode and the semiconductor layer.
[0014] Preferably, the storage device is a capacitor with a first electrode, a second electrode and an insulating layer arranged between the first electrode and the second electrode.
[0015] Preferably, the avalanche photodiode, the flow diffusion, the transfer transistor, the first reset transistor, the second reset transistor, and the gain transistor are arranged in a semiconductor substrate. The capacitor comprises the first and second electrodes facing each other and the insulating layer located between the first and second electrodes. The first and second electrodes are arranged in a wiring layer region that is vertically oriented above a flat main surface of the semiconductor substrate, and the first electrodes in the pixel cells are connected to each other without a continuity.
[0016] Preferably, the first electrode is grounded at a predetermined potential through a passage outside the pixel array.
[0017] Preferably, the first reset transistor has a threshold voltage that is equal to or lower than the threshold voltage of the transfer transistor.
[0018] Preferably, a potential barrier between a source and a drain of the first reset transistor when a first voltage is applied to a gate of the first reset transistor is lower than a potential barrier between a source and a drain of the transfer transistor when a second voltage is applied to a gate of the transfer transistor.
[0019] Preferably, at least some of the charges collected in the cathode of the avalanche photodiode, which have a potential higher than the potential barrier between the source and the drain of the first reset transistor, are transferred to the drain of the first reset transistor.
[0020] Preferably, in operation in the second mode, the potential barrier between source and drain is higher than the potential of the flow diffusion initialized by the second reset transistor when a third voltage is applied to the gate of the transmission transistor.
[0021] Preferably, in operation in the first mode, the potential barrier between source and drain is equal to or lower than the potential of the flow diffusion initialized by the second reset transistor when a fourth voltage is applied to the gate of the transmission transistor.
[0022] The imaging device according to the present disclosure comprises the solid-state image sensor and a signal processing circuit that sets a bias voltage applied to the avalanche photodiode at one of the first bias voltages and the second bias voltage based on a signal output of the solid-state image sensor.
[0023] With the above configuration of the solid-state image sensor, a solid-state image sensor with a photocounting function can be used with a pixel cell containing fewer elements than in the corresponding type. For this reason, the pixel cell can be smaller than those in conventional solid-state image sensors.
[0024] Accordingly, the above configuration of the solid-state image sensor can provide a solid-state image sensor that has a photo-counting function and is better suited for higher integration than that of the corresponding type.
[0025] The imaging device with the above configuration includes a solid-state image sensor with the above configuration.
[0026] Accordingly, the above configuration of the imaging device can provide an imaging device that is better suited for higher integration than that of the corresponding type. BRIEF DESCRIPTION OF THE DRAWINGS Fig. Figure 1 is a circuit diagram representing a pixel cell according to one embodiment. Fig. Figure 2 is a diagram that illustrates an example of switching an operating mode. Fig. Figure 3 is a sectional view illustrating a first exemplary layout of the pixel cell according to the embodiment. Fig. Figure 4 is a sectional view illustrating a second exemplary layout of the pixel cell according to the embodiment. Fig. Figure 5 is a sectional view illustrating a third exemplary layout of the pixel cell according to the embodiment. Fig. Figure 6 is a sectional view illustrating a fourth exemplary layout of the pixel cell according to the embodiment. Fig. Figure 7 is a potential diagram 1 that represents the pixel circuit according to the embodiment in a Geiger multiplication mode. Fig. Figure 8A is a potential diagram 2 that represents the pixel circuit according to the embodiment in Geiger multiplication mode. Fig. Figure 8B is a potential diagram 2 that represents the pixel circuit according to the embodiment in Geiger multiplication mode. Fig. Figure 8C is a potential diagram 3 that represents the pixel circuit according to the embodiment in Geiger multiplication mode. Fig. 8D is a potential diagram 4 that represents the pixel circuit according to the embodiment in Geiger multiplication mode. Fig. Figure 8E is a potential diagram 5 that represents the pixel circuit according to the embodiment in Geiger multiplication mode. Fig. 8F is a potential diagram 6 that represents the pixel circuit according to the embodiment in Geiger multiplication mode. Fig. 8G is a potential diagram 7 that represents the pixel circuit according to the embodiment in Geiger multiplication mode. Fig. 8H is a potential diagram 8 that represents the pixel circuit according to the embodiment in Geiger multiplication mode. Fig. 8I is a potential diagram 9 that represents the pixel circuit according to the embodiment in Geiger multiplication mode. Fig. 8J is a potential diagram 10 that represents the pixel circuit according to the embodiment in Geiger multiplication mode. Fig. 8K is a potential diagram 11 that represents the pixel circuit according to the embodiment in Geiger multiplication mode. Fig. 8L is a potential diagram 12 that represents the pixel circuit according to the embodiment in Geiger multiplication mode. Fig. 9H is a potential diagram 13 that represents the pixel circuit according to the embodiment in Geiger multiplication mode. Fig. 9I is a potential diagram 14 that represents the pixel circuit according to the embodiment in Geiger multiplication mode. Fig. 9J is a potential diagram 15 that represents the pixel circuit according to the embodiment in Geiger multiplication mode. Fig. 9K is a potential diagram 16 that represents the pixel circuit according to the embodiment in Geiger multiplication mode. Fig. 9L is a potential diagram 17 that represents the pixel circuit according to the embodiment in Geiger multiplication mode. Fig. 10A is a timing diagram 1 that represents a photo-counting drive sequence in Geiger multiplication mode. Fig. 10B is a timing diagram 2 that represents the photo-counting drive sequence in a Geiger multiplication mode. Fig. 11A is a characteristic map 1 that illustrates the photo-counting properties in Geiger multiplication mode. Fig. 11B is map 2, which illustrates the photo-counting properties in Geiger multiplication mode. Fig. 11C is a characteristic map 1 that illustrates the photo-counting characteristics according to the comparison example in Geiger multiplication mode. Fig. 11D is the characteristic map 2, which illustrates the photo-counting characteristics according to the comparison example in Geiger multiplication mode. Fig. Figure 12A is a potential diagram 1 that represents the pixel circuit according to the embodiment in a linear multiplication mode. Fig. Figure 12B is a potential diagram 2 that represents the pixel circuit according to the embodiment in linear multiplication mode. Fig. 12C is a potential diagram 3 that represents the pixel circuit according to the embodiment in linear multiplication mode. Fig. 12D is a potential diagram 4 that represents the pixel circuit according to the embodiment in linear multiplication mode. Fig. Figure 13 is a time diagram illustrating a drive sequence in linear amplification mode. Fig. Figure 14 is a block diagram representing a configuration of an imaging device according to the embodiment. Fig. Figure 15A is a conceptual diagram representing a solid-state image sensor of the corresponding type. Fig. Figure 15B is a timing diagram illustrating the operation of the solid-state image sensor in the corresponding manner. Fig. 15C is a block diagram showing a counter-circuit of the solid-state image sensor in the corresponding manner. Fig. Figure 16 is a conceptual diagram illustrating the solid-state image sensor in its associated manner. Fig. 17A is a circuit diagram showing a pixel circuit of the corresponding type. Fig. 17B is a timing diagram that shows the operation of the pixel circuit in the corresponding manner. Fig. 17C is a time diagram that describes a process for counting the number of photons of the corresponding type. DESCRIPTION OF THE EXAMPLE EXECUTION (How an aspect of the present revelation was achieved)
[0027] Recently, a wide range of medical and biological applications, as well as radiation measurement, have required low-light sensors for the precise measurement of faint light, corresponding to a single photon. Among the currently widely used low-light sensors are photomultiplier tubes (PMTs). Unfortunately, PMTs, which are vacuum tube devices, have a size of approximately 10 mm × 10 mm, which limits the pixel count. Furthermore, imaging with a PMT requires processing information about points on a target object by scanning the target object within an XY plane, for example, and then converting this information into an image. Such processing poses challenges for real-time photography.Under these circumstances, there is a need for low-light sensors in the form of a solid-state image sensor to realize a low-light sensor with an increased pixel count and simultaneous high-speed performance.
[0028] As one of the solid-state image sensors that detect weak light, solid-state image sensors have been proposed that each count photons that cause a photoelectric conversion in a photodiode and transmit the counting result as a signal with a digital value to the outside.
[0029] For example, in solid-state image sensors according to JP H07-67 043 A and JP S61-152 176 A, when one or more photons cause a photoelectric conversion between the resets to the one in Fig. The time shown in 15B is determined by a photodiode with a circuit according to Fig. 15A detects the presence of a photon and generates a pulse signal. This pulse signal then increments the count value of the [missing information - likely a specific device or component]. Fig. The digital counter circuit shown in 15C is increased by one.
[0030] Unfortunately, if the digital counter circuit is mounted as a counter on a pixel as described in JP H07-67043A and JP S61-152176A, even a single-bit counter circuit increases the circuit diagram scale, as shown in Fig. Figure 16 illustrates this. Increasing the number of bits to achieve the desired gradation leads to a proportional increase in the number of circuit elements and signal lines, corresponding to the number of bits. It is evident that this results in an increase in the size of the pixel cells, which prevents an increase in the pixel count.
[0031] As described in DUTTON, Neale AW [et al.]: 9.8μm SPAD-based Analogue Single Photon Counting Pixel with Bias Controlled Sensitivity. In: Paper presented at International Image Sensor Workshop, 2013, pp. 1-4, a new type of solid-state image sensor has recently emerged. This sensor features a structure comprising an array of avalanche photodiodes (so-called Geiger-mode APDs) to which a load resistor is connected and a high voltage, equal to or greater than the breakdown voltage, is applied. In this solid-state image sensor, an analog circuit is mounted as a counter on a pixel circuit to overcome the problems of conventional digital counting circuits as described above. The pixel circuit is in Fig. 17A shown.
[0032] In this pixel circuit, a discharge occurs every time a photon reaches the target. Fig. Figure 17B illustrates the moment a charge enters a light-receiving element, a detector, from the holder MC, which maintains an initial voltage. However, a variation in the amplitude VIN, caused by the intensity of the photon energy, and a variation in the parasitic capacitance CP of the control transistor M7 result in a variation in the amount of charge discharged from the holder. Therefore, ΔV of the voltage VC of the holder MC is not constant. As shown in Fig. As shown in Figure 17C, the number of photons is measured by comparing the restoring voltage with the voltage VC of the holder MC after subtracting the ΔV integration. A variation in ΔV leads to difficulties in accurately counting the number of photons.
[0033] In light of such problems, the inventor designed a solid-state image sensor with a configuration that includes the following counting circuit. Using a transistor and a storage element, the counting circuit compensates for the amplitude change of a light-receiving element when photons cause a photoelectric conversion. The counting circuit thus eliminates errors in the amount of analog integration signal per photon, thereby reducing errors in the detected number of photons and the count value. Furthermore, the counting circuit significantly reduces the number of elements compared to a conventional digital counting circuit and allows the count value to be read using only a single analog signal line. Such a configuration can implement a multi-pixel semiconductor image sensor with a powerful photocounting function and a reduced pixel size. FORM OF EXECUTION
[0034] The solid-state image sensor according to one embodiment will now be described with reference to the drawings. [1. Configuration]
[0035] The solid-state image sensor according to the embodiment comprises a pixel arrangement with pixel cells arranged in a matrix.
[0036] Fig. Figure 1 is a circuit diagram representing pixel cell 10 according to the present embodiment.
[0037] As shown in the diagram, pixel cell 10 comprises an avalanche photodiode 1, a flow diffusion 2 that collects charges, a transfer transistor 3 that connects a cathode of the avalanche photodiode 1 and the flow diffusion 2, a first reset transistor 4 for resetting the charges collected in the cathode of the avalanche photodiode 1, a second reset transistor 5 for resetting the charges collected in the flow diffusion 2, an amplifier transistor 6 that converts the charge amount of the charges collected in the flow diffusion 2 into a voltage, a memory 7 that collects charges, and a counter transistor 8 that connects the flow diffusion 2 to the memory 7. Pixel cell 10 may also include a selector transistor 9 that connects the amplifier transistor 6 and the vertical signal line Vsig.
[0038] Here, the avalanche photodiode 1 has a first mode in which an amount of charge approximately proportional to the number of photons that cause a photoelectric conversion is collected in a cathode when photons enter the avalanche photodiode 1 under a first bias voltage (for example, -25 V), and a second mode in which a saturated amount of charge is collected in the cathode when photons enter the avalanche photodiode 1 under a second bias voltage (for example, -27 V) with a larger potential difference to that of the first bias voltage, and a photon causes a photoelectric conversion.
[0039] Here, the first operating mode of the avalanche photodiode 1, for collecting the charges that are approximately proportional to the number of photons causing the photoelectric effect in the cathode, is also referred to as the linear multiplication mode. The second operating mode of the avalanche photodiode 1, for collecting the saturated charge in the cathode when a single photon causes the photoelectric effect, is also referred to as the Geiger multiplication mode.
[0040] In the following, the avalanche photodiode 1 is referred to as APD, flow diffusion 2 as FD, transfer transistor 3 as TRN-TR, first reset transistor 4 as ADPRST-TR, second reset transistor 5 as FDRST-TR, amplifier transistor 6 as SF-TR and in some cases as MCT-TR.
[0041] Although the first bias voltage is, for example, -25 V and the second bias voltage is, for example, -27 V in this description, these specific examples are merely illustrative. In other words, the first bias voltage can have any value at which the amount of charge that is approximately proportional to the number of photons that cause a photoelectric conversion can be accumulated in the cathode. The second bias voltage can have any value that exhibits a larger potential difference compared to that of the first bias voltage and allows the saturated amount of charge to be accumulated in the cathode when a photon causes a photoelectric conversion.
[0042] The signals to be connected to pixel cell 10 will now be described.
[0043] APDRST is a signal for controlling the first reset transistor 4 and is an APD reset signal for resetting the cathode charges of the avalanche photodiode 1.
[0044] TRN is a signal for controlling the transmission transistor 3 and a transfer signal for transferring the charges of the cathode of the avalanche photodiode 1 to the flow diffusion 2.
[0045] FDRST is a signal to control the second reset transistor 5 and is an FD reset signal to reset the charges of the flow diffusion 2.
[0046] MCT is a signal for controlling the counting transistor 8 and is a photo-counting signal for controlling the connection state between flow diffusion 2 and memory 7.
[0047] SEL is a signal for controlling the selection transistor 9 and a row selection signal.
[0048] APDBIAS is a power supply for applying voltage to the anode of the avalanche photodiode 1.
[0049] APDRSD is a reset drain power supply for the avalanche photodiode 1.
[0050] FDRSD is a reset drain power supply for flow diffusion 2.
[0051] PIXVDDD is a power supply for amplifier transistor 6.
[0052] Pixel cell 10 can change the multiplication rate of avalanche photodiode 1 by changing APDBIAS.
[0053] For example, in Fig. As shown in Figure 2, APDBIAS can be set to -25 V at an illuminance greater than 0.1 lx, and the operating mode of the avalanche photodiode 1 can be linear multiplication mode. Similar to a standard image sensor, in this linear multiplication mode, the number of charges to be collected in the cathode of the avalanche photodiode 1 increases approximately proportionally to the illuminance.
[0054] In the event that the target object has an illuminance of less than 0.1 Ix, APDBIAS can be set to -27 V and the operating mode of the avalanche photodiode 1 can be Geiger multiplication mode. In this Geiger multiplication mode, avalanche breakdown in the avalanche photodiode 1 occurs through photoelectric conversion by a photon, and the amount of charge accumulated in the cathode reaches a saturated level. A white image is produced as a result of the saturated output from pixel cell 10, where the photon causes the photoelectric effect, and a black image is produced in a pixel where the photon does not cause the photoelectric effect. Accordingly, the image of the target object is represented with a pixel density with binary values of black and white. To express the pixel density in the form of a graduated image, each pixel cell 10 comprises memory 7.
[0055] Memory 7 with a larger capacity can extend the dynamic range of the output voltage, since such a memory can accumulate a larger amount of charge and integrate a larger number of photons.
[0056] Fig. Figure 3 illustrates a cross-sectional view of a first exemplary layout of pixel cell 10.
[0057] As shown in the drawing, the memory 7 in the first exemplary layout is implemented with a capacitor having a laminate structure comprising an electrode, a semiconductor layer and an insulating layer between the electrode and the semiconductor layer.
[0058] Fig. Figure 4 illustrates a sectional view of a second exemplary layout of pixel cell 10.
[0059] As shown in the drawing, the memory 7 in the second exemplary layout is implemented with a laminated capacitor and formed in a wiring layer region arranged vertically above a flat main surface of a semiconductor substrate. The capacitor can have a configuration comprising a first electrode 72 consisting of a TiN element, a second electrode 73 also consisting of a TiN element, and an insulating layer 71, for example consisting of a SiN element, located between the first electrode 72 and the second electrode 73. Here, the first electrode 72 and the second electrode 73 are positioned opposite each other, with an insulating layer 71 between them.
[0060] In pixel cell 10 of the second example layout, memory 7 is located in the wiring layer. Such a configuration can provide a pixel with a reduced size compared to that of pixel cell 10 in the first example layout.
[0061] In the second exemplary layout, it is assumed that photons enter the avalanche photodiode 1 perpendicularly from above the flat main surface of the semiconductor substrate (namely, from above the front surface). Therefore, a portion of the conduction layer should have an opening through which photons can enter the avalanche photodiode 1.
[0062] Fig. Figure 5 illustrates a sectional view of a third exemplary layout of pixel cell 10.
[0063] In this third exemplary layout, it is assumed that photons enter the avalanche photodiode 1 from the back of the semiconductor substrate.
[0064] In the third exemplary layout, photons enter pixel cell 10 from the back of the semiconductor substrate, thus eliminating the need for an opening in the wiring area. Consequently, a pixel cell 10 in the third exemplary layout can have a memory 7 with a larger capacity than the pixel cell 10 in the second exemplary layout. In the pixel cell 10 of the third exemplary layout, the opening area of the avalanche photodiode 1 can be increased to approximately the size of one pixel cell 10. Such a configuration can increase the photoelectric conversion efficiency of the pixel cell 10 in the third exemplary layout compared to that of the pixel cell 10 in the second exemplary layout.
[0065] Fig. Figure 6 illustrates a sectional view of a fourth exemplary layout of pixel cell 10.
[0066] As in Fig. 3 to 6, the first electrode (hereinafter referred to as the "VSSA electrode"), which is a flat-plate electrode image storage device 7, is connected to a fixed voltage VSSA. As in Fig. 3 to 5, in the first to third example layouts, the VSSA electrode is connected via a junction to the fixed voltage VSSA within pixel cell 10. In contrast, as in Fig. As illustrated in Figure 6, in the fourth example layout, the VSSA electrode in pixel cell 10 is connected without a through-hole to VSSA electrodes in adjacent pixel cells 10 and via through-holes to fixed-voltage VSSAs outside pixel cell 10, i.e., outside the pixel array. Therefore, in the fourth example layout, the wiring for the fixed-voltage VSSA and the connection between the VSSA electrode and the fixed-voltage VSSA do not need to be located within pixel cell 10. Thus, in pixel cell 10 in the fourth example layout, the memory 7 can be enlarged to approximately one cell size of pixel cell 10. Consequently, the memory 7 in pixel cell 10 in the fourth example layout can have a larger capacity than the memory 7 in pixel cell 10 in the third example layout.
[0067] Fig. Figure 7 is a potential diagram of pixel cell 10 when the avalanche photodiode 1 is in Geiger multiplication mode.
[0068] As described above, in the avalanche photodiode 1, in Geiger multiplication mode, the photoelectric conversion by a photon leads to an avalanche discharge, such that the amount of charge accumulated in the cathode exceeds the maximum accumulation amount. When the charges exceeding the maximum accumulation amount enter the flow diffusion 2, the memory 7, and adjacent pixel cells 10, a phenomenon known as blooming occurs.
[0069] Blooming causes mixed colors between pixels, hindering image resolution and thus image recognition. To prevent this, as in Fig. As shown in Figure 7, the potential of the first reset transistor 4 (APDRST-TR) in the off state is set lower than that of the transfer transistor 3 (TRN-TR) in the off state. In other words, the potential barrier between the source and drain of the first reset transistor 4 (APDRST-TR) when a first voltage (the voltage to turn off the first reset transistor 4 (APDRST-TR)) is applied to its gate is lower than the potential barrier between the source and drain of the transfer transistor 3 (TRN-TR) when a second voltage (the voltage to turn off the transfer transistor 3 (TRN-TR)) is applied to its gate. The adjustment procedure involves setting the threshold value Vt of the first reset transistor 4 (APDRST-TR) so that it is lower than the threshold value Vt of the transfer transistor 3 (TRN-TR).As a further measure to be considered, the signal level (LOW level) of APDRST when the first reset transistor 4 (APDRST-TR) is off is set so that it is higher than the signal level (LOW level) of TRN when the transfer transistor 3 (TRN-TR) is off. By setting the potential of the first reset transistor 4 (APDRST-TR) when it is off to a lower level than that of the transfer transistor 3 (TRN-TR) when it is off, the first reset transistor 4 dissipates excess charges generated in the avalanche photodiode 1 beyond what is required to an APDRSD power supply. In other words, among the charges collected in the cathode of the avalanche photodiode 1, at least some of the charges with a higher potential than the potential barrier between the source and the drain of the first reset transistor 4 (APDRST-TR) are transferred to the drain of the first reset transistor 4 (APDRST-TR).By using the APDRSD power supply as a so-called overflow drain, blooming can be prevented.
[0070] [2. Operation] The operation of the solid-state image sensor according to the present embodiment with pixel cell 10 with the above configuration will now be described with reference to the drawings.
[0071] Fig. 8A to 8L and 9H to 9L are potential diagrams of a photocount drive sequence, expressed with a potential when the operating mode of avalanche photodiode 1 is the Geiger multiplication mode. Fig. 10A and Fig. Figure 10B are timing diagrams illustrating the photo-counting drive sequence expressed in terms of timing when the operating mode of avalanche photodiode 1 is the Geiger multiplication mode. Fig. 11A and Fig. Figure 11B are character diagrams illustrating the photo-counting properties when the operating mode of avalanche photodiode 1 is the Geiger multiplication mode.
[0072] When the photo-counting drive sequence is started, the first reset transistor 4 (APDRST-TR), the second reset transistor 5 (FDRST-TR) and the counting transistor 8 (MCT-TR) are first switched on to initialize (reset) the cathode of the avalanche photodiode 1, the flow diffusion 2 and the memory 7.
[0073] As in Fig. As shown in 8A, the potentials of the cathode of avalanche photodiode 1, flow diffusion 2 and memory 7 are set to 3.3 V by initialization.
[0074] In the next step, the first reset transistor 4 (APDRST-TR), the second reset transistor 5 (FDRST-TR), and the counter transistor 8 (MCT-TR) are switched off. At this point, the potential on the turn-off stage of the first reset transistor 4 (APDRST-TR) is held lower than on the turn-off stage of the transmission transistor 3 (TRN-TR).
[0075] As in Fig. As shown in Figure 8B, the potential level of the first reset transistor 4 (APDRST-TR) is set to 1.8 V. The adjustment procedure is as described above.
[0076] In this state (the state in which the potential level of the first reset transistor 4 (APDRST-TR) is set to 1.8 V) pixel cell 10 is exposed to light.
[0077] As in Fig. 10A and Fig. In pixel cell 10B, the exposure time required to count a photon is 754 µs. Pixel cell 10 can count up to 20 photons for 16 ms at 60 fps (frames per second). When photons cause the photoelectric effect in the avalanche photodiode 1 during this exposure time of 754 µs, the amount of charge accumulated in the cathode of the avalanche photodiode 1 reaches a saturated level. The bridging function described above causes the charges to overflow to the APDRSD power supply through the potential-free barrier of the first reset transistor 4 (APDRST-TR). The cathode of the avalanche photodiode 1 has a capacitance of 1.5 fF, which corresponds to 1.5 V as the difference between 1.8 V and a reference voltage of 3.3 V during the initialization of the cathode of the avalanche photodiode 1. For this reason, the cathode of the avalanche photodiode 1 contains, as shown in Fig. Figure 8B shows 14000 electrons. Through this process (operation to overflow the charges to the APDRSD power supply under the influence of the blooming prevention function) a uniform amount of charge can accumulate in the cathode of the avalanche photodiode 1 when a photon causes the photoelectric effect.
[0078] In the next step, as in Fig. As shown in Figure 8C, the transfer transistor 3 (TRN-TR) is switched on to connect the cathode of the avalanche photodiode 1 and the flow diffusion 2 (FD). The potential level when transfer transistor 3 (TRN-TR) is switched on corresponds to a level sufficient to create a potential barrier of 2.8 V, rather than fully connecting the cathode of the avalanche photodiode 1 and the flow diffusion 2 (FD). In other words, the potential barrier between the source and the drain in the state where a third voltage (the voltage to switch on transfer transistor 3 (TRN-TR)) is applied to the gate of transfer transistor 3 (TRN-TR) is increased by the second reset transistor 5 (FDRST-TR) compared to the potential of the flow diffusion 2 in the initialization state. This serves to prevent malfunctions during operation, the details of which will be described later. At this point, the transmission transistor 3 (TRN-TR) is switched on in a linear state.
[0079] Switching on the transmission transistor 3 (TRN-TR) redistributes the charges accumulated in the cathode of the avalanche photodiode 1 between the cathode of the avalanche photodiode 1 and the flow diffusion 2 (FD). The amounts of charge to be redistributed are proportional to the capacitance of the cathode of the avalanche photodiode 1 and that of the flow diffusion 2 (FD). The capacitance of the cathode of the avalanche photodiode 1 is 1.5 fF and the capacitance of the flow diffusion 2 is 2.0 fF. For this reason, as in Fig. Figure 8D illustrates that 8000 electrons accumulated in flow diffusion 2 (FD) even after the transfer transistor 3 (TRN-TR) was switched off in the subsequent step.
[0080] In flow diffusion 2 (FD), slightly leaky charges are generated. Accordingly, as in Fig. As shown in Figure 8E, the counter transistor 8 (MCT-TR) is immediately switched on to transfer the charges to memory 7. At this point, the charges accumulated in the flow diffusion 2 (FD) are redirected to the flow diffusion 2 (FD) and memory 7. The amounts of charge to be redistributed are proportional to the capacitance of the flow diffusion 2 (FD) and the capacitance of memory 7. The capacitance of memory 7 is 20 fF. For this reason, as shown in Figure 8E, the charge is transferred to memory 7. Fig. 8F is shown, 7270 electrons are stored in memory 7.
[0081] The in Fig. 8A to 8F corresponds to one cycle of the photo-counting operation.
[0082] Note that even if two or more photons cause the photoelectric effect in the exposure time of 754 µs, the photoelectric effect is still recognized as the photoelectric effect caused by a single photon.
[0083] To detect the second photon, the cathode of avalanche photodiode 1 must be initialized once, as shown in Fig. 8G is shown. For this reason, the first reset transistor 4 (APDRST-TR) is switched on. Whether or not the flow diffusion 2 (FD) is initialized at this time depends on the ratio of the capacitance of the flow diffusion 2 (FD) to the capacitance of the memory 7. It is preferred that the flow diffusion 2 (FD) be initialized if the capacitance of the memory 7 is sufficiently larger than the capacitance of the flow diffusion 2 (FD). Initialization can reduce the effects of leakage charges in the flow diffusion 2 (FD). Here, the flow diffusion 2 (FD) is not initialized because it cannot be said that the capacitance of the memory 7 is sufficiently larger than the capacitance of the flow diffusion 2 (FD). This is because the charges accumulated in the memory 7 are reduced when the charges for a photon accumulated in the memory 7 (7270 electrons, as in Fig. (8G shown) are redistributed between storage 7 and the initialized flow diffusion 2 (FD). Its effects are particularly significant when one photon causes the photoelectric effect and a photon in the next cycle does not. Such effects do not occur when the first photon causes the photoelectric effect, and the second photon causes it in the next cycle, since 14,000 electrons are again accumulated in the cathode of the avalanche photodiode 1.
[0084] The in Fig. 8H to 8L is a sequence that corresponds to the one in Fig. 8B to 8F. By repeating the sequence similar to the above sequence 20 times, the charges generated by the photoelectric effect, which are caused by the photon at most 20 times, can be integrated into memory 7.
[0085] The charges integrated in memory 7 are converted into a voltage in amplifier transistor 6 (SF-TR) according to the number of photons causing the photoelectric effect in the state in which memory 7 is connected to flow diffusion 2 (FD) via counting transistor 8 (MCT-TR) and the voltage is output. Fig. 10A and Fig. Figure 10B illustrates the initial timing. Fig. 10A is a timing diagram for the case where the conversion of the charge integrated in memory 7 into a voltage in amplifier transistor 6 (SF-TR) and its output is followed by the conversion of the initialized charge in flow diffusion 2 (FD) into a voltage in amplifier transistor 6 (SF-TR) and its output. Fig. Figure 10B is a timing diagram for the case where the conversion of the initialized charge in flow diffusion 2 (FD) into a voltage in amplifier transistor 6 (SF-TR) and its output is followed by the conversion of the charge integrated in memory 7 into a voltage in amplifier transistor 6 (SF-TR) and its output.
[0086] When the number of charges integrated and accumulated in memory 7 is increased by the photoelectric effect continuously caused by photons, the number of newly integrated charges in memory 7 is reduced (see Fig. 11A). For this reason, as in Fig. Figure 11B shows that with an increasing number of photon counts, the slope of the output voltage tends to decrease. A larger capacity of the memory 7 can counteract this tendency even more noticeably. For comparison, the character diagrams in Fig. 11C and Fig. Figure 11D illustrates the photo-counting characteristics when the capacity of memory 7 is 6 fF and the operating mode of avalanche photodiode 1 is Geiger multiplication mode. As shown in Fig. As shown in Figure 11D, if the capacity of memory 7 is equal to 6 fF, i.e., if the ratio of the flow diffusion capacity 2 (FD) to the capacity of memory 7 is 1:3, the slope of the output voltage around the photocounter 10 or higher becomes so small that photocounter detection is difficult. The inventor, who conducted repeated tests and investigations, found that the capacity of memory 7 is preferably at least five times the flow diffusion capacity 2 (FD) in order to detect 20 photons in the photocounter.
[0087] Fig. Figures 9H to 9L are potential diagrams of a photo-counting drive sequence, which are represented by a potential when the first photon causes the photoelectric effect, and after the first photon, the second photon does not cause the photoelectric effect in the next cycle.
[0088] As in Fig. As shown in 9H, the photon does not cause the photoelectric effect even after illuminating pixel cell 10. For this reason, for example, only 10 electrons are accumulated in the cathode of the avalanche photodiode 1 due to noise. For this reason, as in Fig. As shown in Figure 9, even with the transmission transistor 3 (TRN-TR) switched on, the charges accumulated in the cathode of the avalanche photodiode 1 and the flow diffusion 2 (FD) are not redistributed to the incomplete connection of the cathode of the avalanche photodiode 1 and the charges accumulated in flow diffusion 2 (FD).
[0089] The potential level at the switch-on of the transfer transistor 3 (TRN-TR) is a level to create a potential barrier of 2.8 V, and not the level to fully connect the cathode of the avalanche photodiode 1 and the flow diffusion 2 (FD). This serves to prevent the redistribution of the charges accumulated in the cathode of the avalanche photodiode 1 and the charges accumulated in the flow diffusion 2 (FD) when a photon does not cause the photoelectric effect. For this reason, as in Fig. As shown in 9I, the redistribution of charges by the transfer transistor 3 (TRN-TR) is not carried out, and the flow diffusion 2 (FD) continuously maintains the amount of accumulated charges (here 730 electrons).
[0090] For this reason, as in Fig. 9K is shown, the amount of charges stored in memory 7 is not changed by switching on the counter transistor 8 (MCT-TR).
[0091] Fig. Figures 12A to 12D are potential diagrams that represent a drive sequence expressed by a potential when the operating mode of avalanche photodiode 1 is the linear multiplication mode. Fig. Figure 13 is a timing diagram illustrating a timing-expressed drive sequence when the operating mode of avalanche photodiode 1 is the linear multiplication mode.
[0092] If the operating mode of the avalanche photodiode 1 is the linear multiplication mode, the solid-state image sensor according to the present embodiment, including pixel cell 10, operates essentially like a standard image sensor.
[0093] When the drive sequence is started, the first reset transistor 4 (APDRST-TR), the second reset transistor 5 (FDRST-TR) and the counter transistor 8 (MCT-TR) are first switched on to initialize (reset) the cathode of the avalanche photodiode 1, the flow diffusion 2 and the memory 7.
[0094] As in Fig. As shown in 12A, the cathode of the avalanche photodiode 1, the flow diffusion 2 and the memory 7 are set to a potential of 3.3 V by initialization.
[0095] In the next step, the first reset transistor 4 (APDRST-TR), the second reset transistor 5 (FDRST-TR), and the counter transistor 8 (MCT-TR) are switched off. At this point, the turn-off potential of the first reset transistor 4 (APDRST-TR) is kept lower (here 1.8 V than that of the turn-off level of the transfer transistor 3 (TRN-TR)). The blooming prevention function described above then operates when the amount of charge accumulated in the cathode of the avalanche photodiode 1 reaches a saturated level and the charge overflows through the off-potential barrier of the first reset transistor 4 (APDRST-TR) to the APDRSD power supply.
[0096] In this state (the state in which the potential level of the first reset transistor 4 (APDRST-TR) is set to 1.8 V), pixel cell 10 is exposed to light. As in Fig. As shown in Figure 13, the maximum exposure time is 16 ms at 60 fps.
[0097] The exposure time can be set using an electronic shutter. The electronic shutter can be implemented by switching on the first reset transistor 4 (APDRST-TR) and the second reset transistor 5 (FDRST-TR) during exposure.
[0098] In the next step, the second reset transistor 5 (FDRST-TR) is switched on to initialize the flow diffusion 2 (FD), as shown in Fig. 12C is shown. After initialization, as in Fig. As shown in Figure 12D, the potential level of the transfer transistor 3 (TRN-TR) is regulated to 3.3 V to fully connect the cathode of the avalanche photodiode 1 and the flow diffusion 2 (FD). At the point when a fourth voltage (the voltage to turn on the transfer transistor 3 (TRN-TR)) is applied to the gate of the transfer transistor 3 (TRN-TR), the potential barrier between the source and drain is equal to or less than the potential of the flow diffusion 2 initialized by the second reset transistor 5 (FDRST-TR). The charges accumulated in the flow diffusion 2 (FD) are converted into a voltage in the amplifier transistor 6 (SF-TR), and the voltage is output. Fig. Figure 13 is a time diagram illustrating the case in which the conversion of the initialized charges of the flow diffusion 2 (FD) into a voltage in the amplifier transistor 6 (SF-TR) and its output is followed by the conversion of the charges accumulated in the flow diffusion 2 (FD) into a voltage in the amplifier transistor 6 (SF-TR) and its output.
[0099] [3. Application example] The imaging device 200 with a solid-state image sensor with the above configuration will now be described.
[0100] Fig. Figure 14 is a block diagram illustrating a configuration of the imaging device 200.
[0101] As shown in the drawing, the imaging device 200 comprises a solid-state image sensor 100 according to the present embodiment with the above configuration and a signal processing circuit 110 which sets a bias voltage for the avalanche photodiode 1 at a first bias voltage (for example -25 V) and a second bias voltage (for example -27 V) based on a signal output of the solid-state image sensor 100.
[0102] The signal processing circuit 110 includes the signal processing LSI 120 and the analog front-end circuit 130.
[0103] The solid-state image sensor 100 outputs the pixel signal SENS-OUT to the analog front-end circuit 130. The analog front-end circuit 130 then detects a signal component of SENS-OUT, applies a gain, and outputs an AFE-OUT signal. The signal processor LSI 120 then compares the AFE-OUT signal with a preset threshold to determine whether the target object has low illuminance. If the signal processor LSI 120 determines that the target object has low illuminance, it controls the voltage switch 140 using a control signal VOL-CNTL to set a bias voltage applied to the avalanche photodiode 1 to a second bias voltage (for example, -27 V). This switches the avalanche photodiode 1 into Geiger multiplication mode.In contrast, if the signal processor LSI 120 determines that the target object does not have low illuminance, the signal processor LSI 120 controls the voltage switch 140 using the control signal VOL-CNTL to set the bias voltage applied to the avalanche photodiode 1 to a first bias voltage (for example, -25 V). This switches the avalanche photodiode 1 into linear multiplication mode.
[0104] As described above, the imaging device can capture 200 images by switching the operating mode of the avalanche photodiode 1 between the Geiger multiplication mode and the linear multiplication mode according to the illuminance of the target object. INDUSTRIAL APPLICABILITY
[0105] The solid-state image sensor and imaging device according to this disclosure may be widely used in devices / equipment that take photographs. REFERENCE MARKS IN THE DRAWINGS 1 avalanche photodiode 2 Flow diffusion 3 transmission transistor 4 first reset transistor 5 second reset transistor 6 Amplification transistor 7 storage 8 Counting transistor 9 Selection transistor 10 pixel cell 71 Insulation layer 72 first electrode 73 second electrode 100 solid-state image sensors 110 Signal processing circuit 200 Imaging device
Claims
[1] Solid-state image sensor comprising: a pixel array having pixel cells (10) arranged in a matrix, wherein each of the pixel cells (10) comprises: an avalanche photodiode (1) having a first mode in which an amount of charge approximately proportional to the number of photons causing photoelectric conversion is collected in a cathode of the avalanche photodiode (1) when photons enter the avalanche photodiode (1) under application of a first bias voltage, and having a second mode in which a saturated amount of charge is collected in the cathode when photons enter the avalanche photodiode (1) under application of a second bias voltage having a larger potential difference from a potential difference of the first bias voltage and to which a photon causes photoelectric conversion; a flow diffusion (2) that accumulates charges; a transfer transistor (3) connecting the cathode of the avalanche photodiode (1) and the flow diffusion (2); a second reset transistor (5) for resetting charges accumulated in the flow diffusion (2); an amplifier transistor (6) for converting a quantity of charges accumulated in the flow diffusion (2) into a voltage; a storage device (7) that accumulates charges; and a counting transistor (8) connecting the flow diffusion (2) and the memory (7), characterized by : a first reset transistor (4) for resetting charges collected in the cathode of the avalanche photodiode (1); wherein in the case where the avalanche photodiode (1) is in the second mode, a potential barrier between the source and drain of the transfer transistor (3) is higher than the potential of the cathode of the avalanche photodiode (1) when the transfer transistor (3) is turned on; and in the case where the avalanche photodiode (1) is in the first mode, the potential of the cathode of the avalanche photodiode (1) is equal to the potential of the flow diffusion (2) and a potential barrier between the source and drain of the transfer transistor (3) is equal to or lower than the potential of the cathode of the avalanche photodiode (1) when the transfer transistor (3) is turned on. [2] A solid-state image sensor according to claim 1, wherein the memory (7) has a capacity that is at least five times the capacity of the flow diffusion (2). [3] A solid-state image sensor according to claim 1 or 2, wherein the memory (7) is a capacitor having a laminate structure. [4] A solid-state image sensor according to claim 1 or 2, wherein the memory (7) is a capacitor comprising an electrode (72), a semiconductor layer and an insulating layer (71) disposed between the electrode (72) and the semiconductor layer. [5] A solid-state image sensor according to claim 1 or 2, wherein the memory (7) is a capacitor comprising a first electrode (72), a second electrode (73) and an insulating layer (71) disposed between the first electrode (72) and the second electrode (73). [6] A solid-state image sensor according to claim 5, wherein the avalanche photodiode (1), the flow diffusion (2), the transfer transistor (3), the first reset transistor (4), the second reset transistor (5) and the amplification transistor (6) are arranged in a semiconductor substrate; the capacitor has the first electrode (72) and the second electrode (73) facing each other, and the insulating layer (71) arranged between the first electrode (72) and the second electrode (73); the first electrode (72) and the second electrode (73) are arranged in a wiring layer region arranged vertically above a flat main surface of the semiconductor substrate; and first electrodes (72) in the pixel cells (10) are connected to each other without a passage. [7] A solid-state image sensor according to claim 6, wherein the first electrode (72) is grounded at a predetermined potential through a via outside the pixel array. [8] A solid-state image sensor according to any one of claims 1 to 7, wherein the first reset transistor (4) has a threshold voltage equal to or lower than a threshold voltage of the transfer transistor (3). [9] A solid-state image sensor according to any one of claims 1 to 7, wherein a potential barrier between a source and a drain of the first reset transistor (4) upon application of a first voltage to a gate of the first reset transistor (4) is lower than a potential barrier between a source and a drain of the transfer transistor (3) upon application of a second voltage to a gate of the transfer transistor (3). [10] A solid-state image sensor according to claim 8 or 9, wherein, among the charges collected in the cathode of the avalanche photodiode (1), at least a part of the charges having a potential higher than the potential barrier between the source and the drain of the first reset transistor (4) is transferred to the drain of the first reset transistor (4). [11] A solid-state image sensor according to any one of claims 1 to 7, wherein, in an operation in the second mode, the potential barrier between the source and the drain upon application of a third voltage to the gate of the transfer transistor (3) is higher than a potential of the flow diffusion (2) initialized by the second reset transistor (5). [12] Imaging device comprising: the solid-state image sensor according to any one of claims 1 to 11; and a signal processing circuit (110) that sets a bias voltage applied to the avalanche photodiode (1) at one of the first bias voltage and the second bias voltage based on a signal output of the solid-state image sensor.
Citation Information
Patent Citations
Solid-state image pickup device
JP1986152176A
Solid-state image pickup element
JP1995067043A
Solid-state image sensor, control method thereof and electronic apparatus
JP2014060519A
Solid-state imaging device, method of driving solid-state imaging device and imaging apparatus
US20080258045A1
Imaging device
US20090225210A1
