High-speed optical detection device

Abstract

The present application relates to a high-speed light detection device. A circuit is disclosed, comprising: a light detector, comprising a first readout terminal and a second readout terminal, the second readout terminal being different from the first readout terminal; a first readout circuit coupled to the first readout terminal and configured to output a first readout voltage; a second readout circuit coupled to the second readout terminal and configured to output a second readout voltage; and a common-mode analog-to-digital converter, comprising: a first input terminal coupled to a first voltage source; a second input terminal coupled to a common-mode generator, the common-mode generator being configured to receive the first readout voltage and the second readout voltage and generate a common-mode voltage, the common-mode voltage being between the first readout voltage and the second readout voltage; and a first output terminal configured to output a first output signal corresponding to a current amount generated by the light detector.

Description

[0001] This application is a divisional application of the application with national application number 201880035981.X (international application number is PCT / US2018 / 025949, international application date is April 3, 2018, and the name of the invention is “High-speed optical detection device”).

[0002] CROSS-REFERENCE TO RELATED APPLICATIONS

[0003] This application is a continuation-in-part of and claims the benefit of U.S. Patent Application No. 15 / 908,447, filed on February 28, 2018, which claims the benefit of U.S. Provisional Application No. 62 / 465,139, filed on February 28, 2017, U.S. Provisional Application No. 62 / 479,322, filed on March 31, 2017, U.S. Provisional Application No. 62 / 504,531, filed on May 10, 2017, U.S. Provisional Application No. 62 / 485,003, filed on April 13, 2017, U.S. Provisional Application No. 62 / 511,977, filed on May 27, 2017, U.S. Provisional Application No. 62 / 534,179, filed on July 18, 2017, U.S. Provisional Application No. 62 / 534,179, filed on September 21, 2017 Provisional Patent Application 62 / 561,266, U.S. Provisional Patent Application 62 / 613,054 filed on January 3, 2018, priority to U.S. Provisional Patent Application 62 / 617,317 filed on January 15, 2018; and is a continuation-in-part of and claims the benefit of U.S. Patent Application 15 / 338,660 filed on October 31, 2016, which claims priority to U.S. Provisional Patent Application 62 / 294,436 filed on February 12, 2016, U.S. Provisional Patent Application 62 / 271,386 filed on December 28, 2015, and U.S. Provisional Patent Application 62 / 251,691 filed on November 6, 2015, all of which are incorporated by reference in their entirety.

[0004] This application also claims priority to U.S. Provisional Application No. 62 / 481,131 filed on April 4, 2017, U.S. Provisional Application No. 62 / 511,977 filed on May 27, 2017, U.S. Provisional Application No. 62 / 542,329 filed on August 8, 2017, U.S. Provisional Application No. 62 / 561,256 filed on September 21, 2017, U.S. Provisional Application No. 62 / 581,720 filed on November 5, 2017, U.S. Provisional Application No. 62 / 581,777 filed on November 5, 2017, and U.S. Provisional Application No. 62 / 596,914 filed on November 11, 2017, all of which are incorporated by reference in their entirety. Technical Field

[0005] The present invention relates to detecting light using a photodetector. Background Art

[0006] The light propagates in free space or an optical medium coupled to a photodetector that converts the optical signal into an electrical signal for processing. Summary of the invention

[0007] According to one of the inventive aspects of the objectives described in the present invention, light reflected by a three-dimensional object can be detected by a photodetector of an imaging system. This photodetector can convert the detected light into multiple electric charges. Each photodetector may include two sets of switches for collecting charges. The charge collection process controlled by the two sets of switches can be switched in a time sequence so that the imaging system can confirm the phase information of the detected light. The imaging system can use the phase information to analyze characteristics associated with the three-dimensional object, including depth information or material composition. The imaging system can also use the phase information to analyze characteristics associated with the following: eye gesture recognition, gesture recognition, three-dimensional object scanning / video recording, motion tracking, and / or augmented / virtual reality applications.

[0008] Generally speaking, an innovative aspect of the subject matter described herein can be embodied as a circuit comprising: a photodetector comprising a first readout terminal and a second readout terminal, the second readout terminal being different from the first readout terminal; a first readout subcircuit comprising a first MOSFET transistor and a second MOSFET transistor, the first MOSFET transistor comprising a first gate terminal coupled to a first control voltage source, a first channel terminal, and a second channel terminal coupled to the first readout terminal of the photodetector; the second MOSFET transistor comprising a second gate terminal coupled to a second control voltage source, a first channel terminal coupled to a supply voltage source, and a second gate terminal coupled to a second control voltage source. a third channel terminal connected to the supply voltage node, and a fourth channel terminal coupled to the first channel terminal; and a second readout subcircuit comprising a third MOSFET transistor and a fourth MOSFET transistor, the third MOSFET transistor comprising a third gate terminal coupled to the first control voltage source, a fifth channel terminal, and a sixth channel terminal coupled to the second readout terminal of the photodetector; and the fourth MOSFET transistor comprising a fourth gate terminal coupled to the second control voltage source, a seventh channel terminal coupled to the supply voltage node, and an eighth channel terminal coupled to the fifth channel terminal. When operating the circuit, the first control voltage source generates a first control voltage, the first control voltage is configured to establish a first voltage difference and a second voltage difference, the first voltage difference being a difference between a supply voltage of the supply voltage node and a first voltage of the first readout terminal, and the second voltage difference being a difference between the supply voltage of the supply voltage node and a second voltage of the second readout terminal.

[0009] Implementations of the circuit may include one or more of the following features: For example, when operating the circuit, the first control voltage is configured to cause the first MOSFET transistor and the third MOSFET transistor to operate in a subcritical region or a saturation region, respectively.

[0010] In some embodiments, the first voltage difference and the second voltage difference are greater than or equal to ten percent of the supply voltage.

[0011] In some embodiments, when operating the circuit, the first control voltage source reduces a first dark current collected through the first readout terminal and a second dark current collected through the second readout terminal compared to a circuit without the first MOSFET transistor and the third MOSFET transistor.

[0012] In some embodiments, the photodetector further comprises a p-type doped body, the first readout terminal and the second readout terminal comprise a plurality of n-type doped regions, and the first MOSFET transistor and the third MOSFET transistor are both n-type MOSFET transistors.

[0013] In some embodiments, the photodetector further comprises an n-type doped body, the first readout terminal and the second readout terminal comprise a plurality of n-type doped regions, and the first MOSFET transistor and the third MOSFET transistor are both p-type MOSFET transistors.

[0014] In some embodiments, the light detector is a switching light detector and is configured for time-of-flight detection.

[0015] In some embodiments, the photodetector further comprises a light absorbing region, and the light absorbing region comprises germanium. The photodetector further comprises a first control terminal and a second control terminal. The photodetector may comprise a groove, and at least a portion of the light absorbing region is embedded in the groove.

[0016] Another innovative aspect of the subject matter of the present invention can be embodied in a method for operating a circuit, the circuit comprising a photodetector, the photodetector having a first readout terminal coupled to a first readout subcircuit and a second readout terminal coupled to a second readout subcircuit. The method comprises: generating a first control voltage by a first control voltage source coupled to the first readout subcircuit and the second readout subcircuit, the first control voltage being configured to operate a first MOSFET transistor of the first readout subcircuit and a third MOSFET transistor of the second readout circuit in a subthreshold region or a saturation region, respectively; and performing a photodetector readout step, comprising setting a first output terminal of the first readout subcircuit to a fifth voltage and setting a second output terminal of the second readout subcircuit to a sixth voltage; wherein the first control voltage source is controlled to establish a first voltage difference and a second voltage difference, the first voltage difference being the difference between a supply voltage of the first readout subcircuit and the second readout subcircuit and a first voltage of the first readout terminal, and the second voltage difference being the difference between the supply voltage and a second voltage of the second readout terminal.

[0017] Another innovative aspect of the subject matter described in the present invention can be embodied in a circuit comprising: a light emitting device comprising a cathode coupled to a first supply voltage node, and an anode; a MOSFET transistor comprising a gate terminal coupled to an input signal source, a first channel terminal coupled to the anode of the light emitting device, and a second channel terminal coupled to a second supply voltage node; a first inductor comprising a first terminal coupled to a third supply voltage node or a current source, and a second terminal coupled to the anode of the light emitting device; and a second inductor comprising a third terminal coupled to the gate terminal of the MOSFET transistor, and a fourth terminal; wherein a second inductance value of the second inductor is set so that an LC resonant frequency associated with the gate terminal of the MOSFET transistor corresponds to an input frequency of the input signal source.

[0018] Specific embodiments of the circuit may include one or more of the following features. For example, the circuit may further include a first capacitor disposed between the input signal source and the gate terminal of the MOSFET transistor, the first capacitor including a first terminal coupled to the gate terminal of the MOSFET transistor, and a second terminal coupled to the input signal source, wherein the fourth terminal of the second inductor may be coupled to a MOSFET bias source.

[0019] In some embodiments, when operating the circuit, the MOSFET bias source is controlled to adjust a duty cycle of light output by the light emitting device.

[0020] In some embodiments, the light emitting device comprises a light emitting diode array or a laser diode array.

[0021] Another innovative aspect of the subject matter described in the present invention can be embodied in a circuit, the circuit comprising: a photodetector comprising a first readout terminal and a second readout terminal, the second readout terminal being different from the first readout terminal; a first readout circuit coupled to the first readout terminal and configured to output a first readout voltage; a second readout circuit coupled to the second readout terminal and configured to output a second readout voltage; and a common-mode analog-to-digital converter comprising a first input terminal coupled to a first voltage source and a second input terminal coupled to a common-mode generator, the common-mode generator being configured to receive the first readout voltage and the second readout voltage and generate a common-mode voltage, the common-mode voltage being between the first readout voltage and the second readout voltage; and a first output terminal configured to output a first output signal corresponding to an amount of current generated by the photodetector.

[0022] Specific embodiments of the circuit may include one or more of the following features. For example, the circuit may further include a differential analog-to-digital converter; the differential analog-to-digital converter includes: a third input terminal coupled to the first readout circuit, the third input terminal is configured to receive the first readout voltage; a fourth input terminal coupled to the second readout circuit, the fourth input terminal is configured to receive the second readout voltage; and a second output terminal is configured to output a second output signal corresponding to a time difference ranging information generated by the photodetector; wherein the circuit is operated to synchronously generate the first output signal and the second output signal.

[0023] In some embodiments, the first readout circuit includes: a first capacitor coupled to the first readout terminal; and a first source follower circuit coupled to the first capacitor and configured to generate the first readout voltage. The second readout circuit includes: a second capacitor coupled to the second readout terminal; and a second source follower circuit coupled to the second capacitor and configured to generate the second readout voltage.

[0024] In some embodiments, the first readout circuit includes: a first MOSFET transistor including a first gate terminal, a first channel terminal and a second channel terminal, the first gate terminal being coupled to a first control voltage source, and the second channel terminal being coupled to the first readout terminal of the photodetector; a second MOSFET transistor including a second gate terminal, a third channel terminal and a fourth channel terminal, the second gate terminal being coupled to a second control voltage source, the third channel terminal being coupled to a supply voltage node, and the fourth channel terminal being coupled to the first channel terminal; a first capacitor being coupled to the first channel terminal of the first MOSFET transistor; and a first source follower circuit being coupled to the first capacitor and being configured to generate the first readout voltage. The second readout circuit includes: a third MOSFET transistor, including a third gate terminal, a fifth channel terminal and a sixth channel terminal, the third gate terminal is coupled to the first control voltage source, and the sixth channel terminal is coupled to the second readout terminal of the photodetector; a fourth MOSFET transistor, including a fourth gate terminal, a seventh channel terminal and an eighth channel terminal, the fourth gate terminal is coupled to the second control voltage source, the seventh channel terminal is coupled to the supply voltage node, and the eighth channel terminal is coupled to the fifth channel terminal; a second capacitor is coupled to the fifth channel terminal of the third MOSFET transistor; and a second source follower circuit is coupled to the second capacitor and is configured to generate the second readout voltage.

[0025] In some embodiments, the first voltage source includes a third source follower circuit.

[0026] Another innovative aspect of the subject matter of the present invention can be embodied in a method for operating a circuit, the method being suitable for measuring a performance characteristic of a time-of-flight detection device, the time-of-flight detection device comprising a photodetector, the photodetector having a first readout terminal and a second readout terminal, the first readout terminal being coupled to a first readout circuit and configured to output a first readout voltage, the second readout terminal being coupled to a second readout circuit and configured to output a second readout voltage. The method comprises: measuring a dark current of the photodetector by measuring a common-mode output signal between the first readout voltage and the second readout voltage in the absence of ambient light and a time-of-flight light signal; determining whether the dark current of the photodetector is greater than a first value; and determining that the time-of-flight detection device does not meet a performance specification when the dark current of the photodetector is greater than the first value.

[0027] Specific embodiments of the method may include one or more of the following features. For example, measuring the dark current of the photodetector includes: measuring the common-mode output signal between the first readout voltage and the second readout voltage one or more times by a 1-bit analog-to-digital converter or a multi-bit analog-to-digital converter in the absence of ambient light and a time-difference ranging light signal; and determining the dark current based on the one or more measured common-mode output signals.

[0028] In some embodiments, the one or more measurements are multiple measurements, wherein each of the multiple measurements corresponds to a different integration time or a different replica voltage input to the 1-bit analog-to-digital converter or the multi-bit converter.

[0029] In some embodiments, the method further includes: when there is a time difference ranging optical signal, measuring a differential mode output signal between the first readout signal and the second readout signal to measure a demodulation contrast of the time difference ranging detection device; determining whether the demodulation contrast of the time difference ranging detection device is lower than a second value; and when the demodulation contrast of the time difference ranging detection device is lower than the second value, determining that the time difference ranging device does not meet the performance specification.

[0030] This embodiment and other embodiments may each include one or more of the following features as appropriate. Germanium is an efficient absorber material for infrared wavelengths, which can reduce the problem of slower photocarrier generation at greater substrate depths when less efficient absorbers (e.g., silicon) are used. For photodetectors with n- and p-doped regions made at different depths, the transmission distance of photocarriers is limited by the depth of the absorber material rather than the width. Therefore, even if an efficient absorber material with a short absorption length is used, the distance between the p- and n-doped regions can be shortened, so that a small bias voltage can generate a strong field and increase the operating speed. For such a photodetector, two groups of switches can be inserted and configured to be staggered to collect photocarriers of different optical phases for a ToF system. The increase in operating speed allows the use of higher modulation frequencies in a ToF system, thereby obtaining a higher depth resolution. In a ToF system, the peak intensity of the light pulse increases when its duty cycle is reduced, which can improve the signal-to-noise ratio (and depth accuracy) while keeping the energy consumption of the ToF system unchanged. This can be achieved when the operating speed is increased, so that the duty cycle of the light pulse can be shortened without deforming the pulse shape. In addition, when germanium is used as the absorption region, light pulses with wavelengths greater than 1 μm can be used. NIR with longer wavelengths (such as 1.31 μm, 1.4 μm, 1.55 μm) are generally considered safer for the human eye, so light pulses of constant wavelength can be output at higher intensities while meeting eye safety requirements, thereby improving the signal-to-noise ratio (and even depth accuracy).

[0031] The details of one or more embodiments are described in the accompanying drawings and the detailed description below. Other potential features and advantages will become apparent from the detailed description, drawings, and claims. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] Figure 1A , 1B , 1C and 1D are examples of switching photodetectors.

[0033] Figure 2A , 2B , 2C and 2D are examples of switching photodetectors.

[0034] Figure 3A , 3B , 3C and 3D are examples of switching photodetectors.

[0035] Figure 4A , 4B , 4C, 4D and 4E are examples of switch photodetectors.

[0036] Figure 4F-4I is an example of a design in which an absorber layer is selectively formed on a substrate.

[0037] Figures 5A-5C is an example of a light detector.

[0038] Figure 5D-5K is an example of a switching photodetector.

[0039] Figure 6A-6B is an example of a switching photodetector.

[0040] Figure 7A-7B is a cross-sectional view of an example structure of integrating a lens into a photodetector.

[0041] Figures 8A-8C is an example of a switch that turns a light detector on and off.

[0042] Figures 9A-9E are examples of electrical terminals for switching photodetectors.

[0043] Figures 10A-10I is an example structure of a photodetector having an absorption region and a substrate.

[0044] Figures 11A-11F 2 are top and side views of an example of a switching photodetector.

[0045] Figures 12A-12H 2 are top and side views of an example of a switching photodetector.

[0046] Figures 13A-13G 2 are top and side views of an example of a switching photodetector.

[0047] Figures 14A-14B is a top view of an example of a switching light detector.

[0048] Figures 15A-15G is a side view of an example structure of sensor pixel isolation.

[0049] Figures 16A-16J is a cross-sectional view of an example structure of a photodetector.

[0050] Figures 17A-17E A cross-sectional view of an example structure showing surface modification in the absorption region.

[0051] Figures 18A-18G 2 are top and side views of an example of a switching photodetector.

[0052] Figures 19A-19H 2 are top and side views of an example of a switching photodetector.

[0053] Figures 20A-20L 2 are top and side views of an example of a switching photodetector.

[0054] Figures 21A-21F 2 are top and side views of an example of a switching photodetector.

[0055] Figures 22A-22D 2 are top and side views of an example of a switching photodetector.

[0056] Figures 23A-23B 2 are top and side views of an example of a switching photodetector.

[0057] Figures 24A-24G 2 are top and side views of an example of a switching photodetector.

[0058] Figures 25A-25H 2 are top and side views of an example of a switching photodetector.

[0059] Fig.26 is an example of a unit cell for a rectangular photodetector.

[0060] Fig. 27 is an example of a rectangular switched photodetector with phototransistor gain.

[0061] Fig.28A is a block diagram of an example of an imaging system.

[0062] Figures 28B-28C An example of a technique for determining characteristics of an object using an imaging system is shown.

[0063] Fig.29 An example of a flow chart for determining characteristics of an object using an imaging system is shown.

[0064] Fig.30 A block diagram showing an example of a receiver unit for Time-of-Flight (ToF) detection.

[0065] Figures 31A-31I Architectural and cross-sectional views showing examples of a ToF receiver cell with increased capacitance values.

[0066] Fig.32 A block diagram showing an example of a receiver unit for ToF detection.

[0067] Figures 33A-33E A cross-sectional view showing an example of a receiver unit for ToF detection.

[0068] Fig.34 A cross-sectional view showing an example of a bonding procedure of a receiver unit for ToF detection.

[0069] Fig.35 An architectural diagram showing the circuitry for operating a ToF pixel.

[0070] Fig.36A and Fig.36B A side view of an example of a test apparatus is shown.

[0071] Fig.37A An example of a circuit for digitally measuring a ToF pixel is shown.

[0072] Fig.37B and Fig.37C An architectural diagram showing an example of a pixel circuit.

[0073] Fig.37D An architectural diagram showing an example of a common mode detection circuit.

[0074] Fig.37E Display and Operation Fig.37A An example of a timing diagram for the associated circuit.

[0075] Fig.37F An example of a flow chart for characterizing the performance of a ToF detection device is shown.

[0076] Fig.38A and 38B An architectural diagram showing a circuit for operating a light emitting device.

[0077] In the various drawings, like reference numbers and names refer to like elements. It should also be understood that the various exemplary embodiments shown in the drawings are merely illustrative representations and are not necessarily drawn to scale. DETAILED DESCRIPTION

[0078] A photodetector can be used to detect an optical signal and convert the optical signal into an electrical signal that can be further processed by other circuits. In time-of-flight (ToF) applications, the depth information of a three-dimensional object can be determined by the phase difference between the transmitted optical pulse and the detected optical pulse. For example, a two-dimensional array of pixels can be used to reconstruct the three-dimensional image of a three-dimensional object, where each pixel can include one or more photodetectors for obtaining the phase information of the three-dimensional object. In some embodiments, ToF applications use light sources with wavelengths in the near-infrared (NIR) range. For example, the wavelength of a light-emitting diode (LED) can be 850 nanometers (nm), 940 nm, 1050 nm, or from 1.3 micrometers (μm) to 1.6 μm. Some photodetectors can use silicon as the absorption material, but silicon is not an efficient absorber of NIR wavelengths. Specifically, optical carriers can be generated deep in the silicon substrate (e.g., with a depth greater than 10 μm), and these optical carriers will slowly drift and / or diffuse to the junction of the photodetector, which results in a reduction in the operating speed of the device. Secondly, to minimize power consumption, a small voltage amplitude is usually used to control the operation of the photodetector. For a larger absorption region (e.g., with a diameter of 10 μm), a small voltage amplitude can only form a small transverse / longitudinal electric field across the entire larger absorption region, which will affect the drift speed of the optical carriers sweeping across the absorption region. Therefore, the operating speed of the device is further limited. For ToF applications using NIR wavelengths, a photodetector using germanium-silicon (molecular formula GeSi) as the absorption material solves the technical problems discussed above. In this application, the term "photodetector" can be used interchangeably with the term "optical sensor". In this application, the germanium-silicon alloy referred to by the term "GeSi" ranges from 99% germanium (i.e., 1% silicon) to 1% germanium (i.e., 99% silicon). In this application, the germanium-silicon layer can be formed by blanket epitaxy technology, selective epitaxy technology, or other applicable technologies. Secondly, the absorption layer containing the germanium-silicon layer can form a plane, a mesa top surface, or a trench bottom surface surrounded by an insulator (e.g., oxide, nitride), a semiconductor (e.g., silicon, germanium), or a combination thereof. In addition, a strained super lattice structure or a multiple quantum well structure including multiple layers such as alternating layers of germanium-silicon with different compositions can be used in the absorption layer. Also, a silicon layer or a germanium-silicon layer with a low germanium concentration (e.g., <10%) can be used to protect the surface of a germanium-silicon layer with a high germanium concentration (e.g., >50%), which can reduce a dark current or a leakage current on the surface of the germanium-silicon layer with a high germanium concentration.

[0079] Figure 1A 1 is an example of a switch photodetector 100, wherein the switch photodetector 100 is used to convert an optical signal into an electrical signal. The switch photodetector 100 includes an absorption layer 106 formed on a substrate 102. The substrate 102 can be any substrate suitable for forming semiconductor devices on it. For example, the substrate 102 can be a silicon substrate. The absorption layer 106 includes a first switch 108 and a second switch 110.

[0080] In general, the absorption layer 106 receives an optical signal 112 and converts the optical signal 112 into a plurality of electrical signals. The absorption layer 106 may be intrinsic, p-type, or n-type. In some embodiments, the absorption layer 106 may be formed of a p-type germanium silicon material. The absorption layer 106 is selected to have a high absorption coefficient in a desired wavelength range. For NIR wavelengths, the absorption layer 106 may be a germanium silicon platform, where the germanium silicon absorbs photons in the optical signal 112 and generates electron-hole pairs. The material composition of germanium and silicon in the germanium silicon platform may be selected for a specific technology or application. In some embodiments, the absorption layer 106 is designed to have a thickness t. For example, for a wavelength of 850nm or 940nm, in order to have a large quantum efficiency, the thickness of the germanium silicon platform may be about 1 μm. In some embodiments, the surface of the absorption layer 106 is designed to have a specific shape. For example, the germanium silicon platform may be circular, square, or rectangular, depending on the spatial profile of the optical signal 112 on the surface of the germanium silicon platform. In some embodiments, the absorption layer 106 is designed to have a lateral dimension d for receiving the optical signal 112. For example, the SiGe plane may be circular or rectangular, where d ranges from 1 μm to 50 μm.

[0081] A first switch 108 and a second switch 110 are fabricated in the absorption layer 106, and the first switch 108 is coupled to a first control signal 122 and a first readout circuit 124. The second switch 110 is coupled to a second control signal 132 and a second readout circuit 134. Generally speaking, the first readout circuit 124 or the second readout circuit 134 determines to collect electrons or holes according to the control of the first control signal 122 and the second control signal 132.

[0082] In some embodiments, the first switch 108 and the second switch 110 are fabricated to collect a plurality of electrons. In such a case, the first switch 108 includes a p-type doped region 128 and an n-type doped region 126. For example, the p-type doped region 128 may have a p+ doping, wherein the activated dopant concentration may be as high as the fabrication technology allows; for example, when the absorber layer 106 is germanium and doped with boron, about 5×1020 cm -3 In some embodiments, the doping concentration of the p-type doping region 128 may be less than 5×10 20 cm -3 , thereby simplifying the manufacturing complexity (although increasing the contact resistance). The n-type doped region 126 may have an n+ doping, wherein the activated dopant concentration may be as high as the manufacturing technology can achieve; for example, when the absorber layer 106 is germanium and doped with phosphorus, it is about 1×10 20 cm -3 In some embodiments, the doping concentration of the n-type doping region 126 may be less than 1×10 20 cm -3 , thereby simplifying the manufacturing complexity (although increasing the contact resistance). The distance between the p-type doping region 128 and the n-type doping region 126 depends on the process design rules. Generally speaking, the closer the distance between the p-type doping region 128 and the n-type doping region 126, the higher the switching efficiency of generating photocarriers. However, reducing the distance between the p-type doping region 128 and the n-type doping region 126 may increase a dark current associated with the PN junction between the p-type doping region 128 and the n-type doping region 126. Therefore, the distance can be set depending on the performance of the switch photodetector 100. The second switch 110 includes a p-type doping region 138 and an n-type doping region 136, the p-type doping region 138 is similar to the p-type doping region 128, and the n-type doping region 136 is similar to the n-type doping region 126.

[0083] In some embodiments, the p-type doped region 128 is coupled to the first control signal 122; for example, the p-type doped region 128 can be coupled to a voltage source, wherein the first control signal 122 can be an AC voltage signal from the voltage source. In some embodiments, the n-type doped region 126 is coupled to the readout circuit 124. The readout circuit 124 can be a three-transistor configuration consisting of a reset gate, a source-follower, and a selection gate, or any circuit suitable for handling charge. In some embodiments, the readout circuit 124 can be fabricated on the substrate 102. In some other embodiments, the readout circuit 124 can be fabricated on another substrate and integrated or co-packaged with the switch photodetector 100 by die / wafer bonding or stacking.

[0084] The p-type doped layer 138 is coupled to the second control signal 132. For example, the p-type doped region 138 can be coupled to a voltage source, wherein the second control signal 132 can be an AC voltage signal and its phase is opposite to the first control signal 122. In some embodiments, the n-type doped region 136 can be coupled to the readout circuit 134. The readout circuit 134 can be similar to the readout circuit 124.

[0085] The first control signal 122 and the second control signal 132 are used to control the collection process of multiple electrons generated by absorbing photons. For example, when the first control signal 122 is different from the second control signal 132, an electric field is formed between the p-type doping region 128 and the p-type doping region 138, and the free electrons drift to the p-type doping region 128 or the p-type doping region 138 according to the direction of the electric field. In some embodiments, the first control signal 122 can be fixed at a voltage value Vi, and the second control signal 132 can be changed between Vi±ΔV. The drift direction of the electrons depends on the bias value. Accordingly, when one of the switches (e.g., the first switch) is turned on (i.e., the electrons drift to the p-type doping region 128), the other switch (e.g., the second switch 110) is turned off (i.e., the electrons are blocked by the p-type doping region 138). In some embodiments, the first control signal 122 and the second control signal 132 can have different voltage values.

[0086] Generally, the difference between the Fermi level of the p-type doped region and the Fermi level of the n-type doped region forms an electric field between the two regions. In the first switch 108, the electric field is formed between the p-type doped region 128 and the n-type doped region 126. Similarly, in the second switch 110, the electric field is formed between the p-type doped region 138 and the n-type doped region 136. When the first switch 108 is turned on and the second switch 110 is turned off, the electrons drift to the p-type doped region 128, and the electric field between the p-type doped region 128 and the n-type doped region 126 further moves the electrons to the n-type doped region 126. The readout circuit 124 can then process the charges collected by the n-type doped region 126. Conversely, when the second switch 110 is turned on and the first switch 108 is turned off, the electrons will drift to the p-type doping region 138, and the electric field between the p-type doping region 138 and the n-type doping region 136 will further move the electrons to the n-type doping region 136. The readout circuit 134 can then process the charges collected by the n-type doping region 136.

[0087] In some embodiments, a voltage may be applied between a p-type doped region and an n-type doped region of a switch to operate the switch in an avalanche regime to increase the sensitivity of the dual switch photodetector 100. For example, in the case where the silicon germanium layer 106 comprises silicon germanium, when the distance between the p-type doped region 128 and the n-type doped region 126 is about 100 nm, a voltage less than 7 volts may be applied to establish an avalanche gain between the p-type doped region 128 and the n-type doped region 126.

[0088] In some embodiments, substrate 102 can be coupled to an external control 116; for example, substrate 102 can be coupled to electrical ground or a predetermined voltage that is lower than the voltage of n-type doped regions 126 and 136. In some other embodiments, substrate 102 can be floating or not coupled to any external control.

[0089] Figure 1B is an example of a switch optical detector 160, wherein the switch optical detector 160 is used to convert an optical signal into an electrical signal. The switch optical detector 160 is similar to Figure 1A 100, but the first switch 108 and the second switch 110 further include an n-type well region 152 and an n-type well region 154, respectively. In addition, the absorption region 106 can be a p-type doped region. In some embodiments, the doping amount of the n-type well regions 152 and 154 can range from 10 15 cm -3 Up to 10 17 cm -3 , the doping amount of the absorption region 106 can range from 10 14 cm -3 Up to 10 16 cm -3 .

[0090] The arrangement of the p-type doping region 128, the n-type well region 152, the p-type doped absorption region 106, the n-type well region 154, and the p-type doping region 138 forms a PNPNP junction structure. In general, the PNPNP junction structure selectively reduces a leakage current from the first control signal 122 to the second control signal 132, or a leakage current from the second control signal 132 to the first control signal 122. The arrangement of the n-type doping region 126, the p-type doped absorption region 106, and the n-type doping region 136 forms an NPN junction structure. In general, the NPN junction structure selectively reduces a charge coupling from the first readout circuit 124 to the second readout circuit 134, or a charge coupling from the second readout circuit 134 to the first readout circuit 124.

[0091] In some embodiments, the p-type doped region 128 is formed entirely within the n-type well region 152. In some embodiments, the p-type doped region 128 is partially formed in the n-type well region 152; for example, a portion of the p-type doped region 128 can be formed in the n-type well region 152 by implanting a p-type dopant, and another portion of the p-type doped region 128 can be formed in the absorber layer 106 by implanting a p-type dopant. Similarly, in some embodiments, the p-type doped region 128 is formed entirely within the n-type well region 154. In some other embodiments, the p-type doped region 138 is partially formed in the n-type well region 154. In some embodiments, the depth of the n-type well regions 152 and 54 is shallower than the depth of the p-type doped region.

[0092] Figure 1C is an example of a switch photodetector 170, wherein the photodetector 170 is used to convert an optical signal into an electrical signal. The switch photodetector 170 is similar to Figure 1A The switch photodetector 100 shown in FIG. 1 is shown in FIG. 1 , but the absorption layer 106 further includes an n-type well region 156. In addition, the absorption region 106 can be a p-type doped region. In some embodiments, the doping amount of the n-type well region 156 can range from 10 15 cm -3 Up to 10 17 cm -3 , the doping amount of the absorption layer 106 can be in the range of 10 14 cm 3 Up to 10 16 cm -3 .

[0093] The arrangement of the p-type doping region 128, the n-type well region 156, and the p-type doping region 138 forms a PNP junction structure. In general, the PNP junction structure selectively reduces a leakage current from the first control signal 122 to the second control signal 132, or a leakage current from the second control signal 132 to the first control signal 122. The arrangement of the n-type doping region 126, the p-type doped absorption layer 106, and the n-type doping region 136 forms an NPN junction structure. In general, the NPN junction structure selectively reduces a charge coupling from the first readout circuit 124 to the second readout circuit 134, or a charge coupling from the second readout circuit 134 to the first readout circuit 124. In some embodiments, when the n-type well region 156 is deep enough, the arrangement of the n-type doped region 126, the p-type doped region 106, the n-type well region 156, the p-type doped absorption region 106 and the n-type doped region 136 can form an NPNPN junction structure to further reduce a charge coupling from the first readout circuit 124 to the second readout circuit 134, or from the second readout circuit 134 to the first readout circuit 124.

[0094] In some embodiments, p-type doped regions 128 and 138 are completely formed in n-type well region 156. In some embodiments, p-type doped regions 128 and 138 are partially formed in n-type well region 156; for example, a portion of p-type doped region 128 can be formed in n-type well region 156 by implanting p-type dopants, and another portion of p-type doped region 128 can be formed in absorber layer 106 by implanting p-type dopants. In some embodiments, the depth of the n-type well region is shallower than the depth of p-type doped regions 128 and 138.

[0095] Figure 1D is an example of a switch photodetector 180, wherein the photodetector 180 is used to convert an optical signal into an electrical signal. The switch photodetector 180 is similar to Figure 1A The switch photodetector 100 is shown in FIG. 1 , but the switch photodetector 180 further includes a p-type well region 104, and n-type well regions 142 and 144. In some embodiments, the doping amount of the n-type well regions 142 and 144 may be in the range of 10 16 cm -3 Up to 10 20 cm -3 The doping amount of the p-type well region 104 can be in the range of 10 16 cm -3 Up to 10 20 cm -3 .

[0096] In some embodiments, the absorption layer 106 may not completely absorb photons in the incident light signal 112; for example, if the SiGe platform does not completely absorb photons in the incident NIR light signal 112, the NIR light signal 112 may penetrate deep into the silicon substrate 102, and the silicon substrate 102 may absorb the photons that penetrate deep therein and generate slow-recombining photocarriers deep in the silicon substrate 102. These slow-recombining photocarriers may negatively affect the operating speed of the switching light detector. Secondly, the photocarriers generated in the silicon substrate 102 may be collected by adjacent pixels, which may cause undesirable signal cross-talk between pixels. In addition, the photocarriers generated in the silicon substrate 102 may cause the substrate 102 to be charged, which may cause reliability issues for the switching light detector.

[0097] To remove slow-recombining photocarriers, the switching photodetector 180 may include a connection that shorts the n-well regions 142, 144 to the p-well region 104. For example, the connection may be formed by connecting the p-well region 104 and the n-well regions 142, 144 through a silicide process or by depositing metal pads. The short between the n-well regions 142, 144 and the p-well region 104 allows photocarriers generated in the substrate 102 to recombine at the shorting node, thereby improving the operating speed and / or reliability of the switching photodetector. In some embodiments, the p-well region 104 is used to passivate or reduce the electric field around the interface defects between the absorber layer 106 and the substrate 102 to reduce the dark current of the device.

[0098] although Figures 1A-1D Not shown, but in some embodiments, the optical signal can enter the switch optical detector from the back side of the substrate 102 of the switch optical detector. One or more optical elements (e.g., microlenses or light guides) can be fabricated on the back side of the substrate 102, and the optical signal is focused, straightened, defocused, filtered, or processed according to the lens design in the optical element.

[0099] although Figures 1A-1D Not shown, but in some other embodiments, the first switch 108 and the second switch 110 may be fabricated in an alternative manner to collect holes instead of electrons. In this case, the p-doped region 128 and the p-type doped region 138 would be replaced by n-type doped regions, the n-type doped region 126 and the n-type doped region 136 would be replaced by p-type doped regions, the n-type well regions 142, 144, 152, 154 and 156 would be replaced by p-type well regions, and the p-type well region 104 would be replaced by an n-type well region.

[0100] although Figures 1A-1D Not shown, but in some embodiments, the absorber layer 106 may be bonded to a substrate after forming the switch light detectors 100, 160, 170, and 180. The substrate may be any material that allows the photoelectric signal 112 to be transmitted to the switch light detector. For example, the substrate may be a polymer or glass. In some embodiments, one or more optical elements (e.g., microlenses or light guides) may be fabricated on the carrier substrate, and the optical signal may be focused, aligned, defocused, filtered, or otherwise processed according to the lens design.

[0101] although Figures 1A-1DNot shown, but in some embodiments, the switch photodetectors 100, 160, 170, and 180 can be bonded (e.g., by metal-metal bonding, oxide-oxide bonding, hybrid bonding) to a second substrate containing control signal circuitry and / or readout circuitry and / or phase lock loop (PLL) and / or analog-to-digital converter circuitry. A metal layer can be deposited on top of the switch photodetector to act as a reflector to reflect the light signal incident from the back side of the substrate 102. Adding a metal layer like a reflector can increase the absorption efficiency (quantum efficiency) of the absorption layer 106; for example, by adding a reflective metal layer, the absorption efficiency of the photodetector operating in the long NIR wavelength range of 1.0 μm to 1.6 μm can be greatly improved. An oxide layer can also be included between the metal layer and the absorption layer to increase reflectivity. The metal layer can also serve as a bonding layer during the wafer bonding process. In some embodiments, one or more switches similar to the first switch 108 and the second switch 110 can be added to link the control signal / readout circuitry.

[0102] although Figures 1A-1D Although not shown, in some embodiments, the absorption layer 106 may be partially or completely embedded or recessed in the substrate 102 to ease the surface topography and facilitate manufacturing. The aforementioned technology is disclosed in the early publication of U.S. Patent No. 20170040362A1, entitled “Germanium-Silicon Light Sensing Apparatus”, and is incorporated herein by reference.

[0103] Figure 2A 2 is an example of a switch photodetector 200, wherein the switch photodetector 200 is used to convert an optical signal into an electrical signal. The first switch 208 and the second switch 210 are fabricated on a substrate 202. The switch photodetector 200 includes an absorption layer 206 fabricated on the substrate 202. The substrate 202 can be any substrate suitable for disposing semiconductor devices thereon. For example, the substrate 202 can be a silicon substrate.

[0104] In general, the absorption layer 206 receives an optical signal 212 and converts the optical signal 212 into a plurality of electrical signals. The absorption layer 206 is similar to the absorption layer 106. The absorption layer 206 can be intrinsic, p-type, or n-type. In some embodiments, the absorption layer 206 can be formed of a p-type germanium silicon material. In some embodiments, the absorption layer 206 can include a p-type doped region 209. The p-type doped region 209 can repel photoelectrons transmitted from the absorption region 206 to the substrate 202, thereby increasing the operating speed; for example, the p-type doped region 209 can have a p+ dopant. The concentration of the p+ dopant can be as high as the manufacturing technology can achieve; for example, when the absorption layer 206 is germanium and is doped with boron, it is about 5×10 20 cm -3 In some embodiments, the doping concentration of the p-type doping region 209 may be less than 5×10 20 cm -3 , thereby simplifying the manufacturing complexity (although increasing the contact resistance). In some embodiments, the p-type doped region 209 may be a gradient p-type doped region.

[0105] A first switch 208 and a second switch 210 may be fabricated in the substrate 202. The first switch 208 is coupled to a first control signal 222 and a first readout circuit 224. The second switch 210 is coupled to a second control signal 232 and a second readout circuit 234. Generally speaking, the first readout circuit 224 or the second readout circuit 234 determines to collect electrons or holes according to the control of the first control signal 222 and the second control signal 232. The first control signal 222 is similar to the first control signal 122, and the second control signal 232 is similar to the second control signal 132; the first readout circuit 224 is similar to the first readout circuit 124, and the second readout circuit 234 is similar to the second readout circuit 134.

[0106] In some embodiments, the first switch 208 and the second switch 210 are fabricated to collect the electrons generated by the absorption region 206. In such a case, the first switch 208 includes a p-type doped region 228 and an n-type doped region 226. For example, the p-type doped region 228 may have a p+ doping, wherein the activated dopant concentration may be as high as the fabrication technology can achieve; for example, when the substrate 202 is silicon and doped with boron, about 2×10 20 cm -3 In some embodiments, the doping concentration of the p-type doping region 228 may be less than 2×10 20 cm -3 , thereby simplifying the manufacturing complexity (although increasing the contact resistance). The n-type doped region 226 may have an n+ doping, wherein the activated dopant concentration may be as high as the manufacturing technology can achieve; for example, when the substrate 202 is silicon and is doped with phosphorus, it is about 5×1020 cm -3 In some embodiments, the doping concentration of the n-type doping region 226 may be less than 5×10 20 cm -3 , thereby simplifying the manufacturing complexity (although increasing the contact resistance). The distance between the p-type doping region 228 and the n-type doping region 226 depends on the process design rules. Generally speaking, the closer the distance between the p-type doping region 228 and the n-type doping region 226, the higher the switching efficiency of generating photocarriers. The second switch 210 includes a p-type doping region 238 and an n-type doping region 236. The p-type doping region 238 is similar to the p-type doping region 228, and the n-type doping region 236 is similar to the n-type doping region 226.

[0107] In some embodiments, the p-type doped region 228 is coupled to the first control signal 222, and the n-type doped region 226 is coupled to the readout circuit 224. The p-type doped region 238 is coupled to the second control signal 232. The n-type doped region 236 is coupled to the readout circuit 234. The first control signal 222 and the second control signal 232 are used to control the collection process of multiple electrons generated by absorbing photons. For example, when the absorption layer 206 absorbs multiple photons in the light signal 212, electron-hole pairs are generated and drift or diffuse into the substrate 202. When a voltage is applied, if the first control signal 222 is different from the second control signal 232, an electric field is formed between the p-type doped region 228 and the p-type doped region 238, and free electrons drift from the absorption layer 206 to the p-type doped region 228 or the p-type doped region 238 according to the direction of the electric field. In some embodiments, the first control signal 222 can be fixed at a voltage value Vi, and the second control signal can be changed between Vi±ΔV. The drift direction of the electrons depends on the bias value. Accordingly, when one of the switches (e.g., the first switch 208) is turned on (i.e., the electrons drift to the p-type doped region 228), the other switch (e.g., the second switch 210) is turned off (i.e., the electrons are blocked by the p-type doped region 238). In some embodiments, the first control signal 222 and the second control signal 232 may have different voltage values.

[0108] In the first switch 208, an electric field is formed between the p-type doping region 228 and the n-type doping region 226. Similarly, in the second switch 210, an electric field is formed between the p-type doping region 238 and the n-type doping region 236. When the first switch 208 is turned on and the second switch 210 is turned off, electrons drift to the p-type doping region 228, and the electric field between the p-type doping region 228 and the n-type doping region 226 further moves the electrons to the n-type doping region 226. The readout circuit 224 can then process the charges collected by the n-type doping region 226. Conversely, when the second switch 210 is turned on and the first switch 208 is turned off, electrons drift to the p-type doping region 238, and the electric field between the p-type doping region 238 and the n-type doping region 236 further moves the electrons to the n-type doping region 236. The readout circuit 234 can then process the charges collected by the n-type doping region 236.

[0109] In some implementations, a voltage may be applied between a p-type doped region and an n-type doped region of a switch so that the switch operates in an avalanche mechanism to increase the sensitivity of the switch photodetector 200. For example, in the case where the substrate 202 comprises silicon germanium, when the distance between the p-type doped region 228 and the n-type doped region 226 is approximately 100 nm, a voltage of less than 7 volts may be applied to establish an avalanche gain between the p-type doped region 228 and the n-type doped region 226.

[0110] In some embodiments, the p-type doped region 209 can be coupled to an external control 214, for example, the p-type doped region 209 can be coupled to electrical ground. In some embodiments, the p-type doped region 209 can be floating or not coupled to any external control. In some embodiments, the substrate 202 can be coupled to an external control 216; for example, the substrate 202 can be coupled to electrical ground or a predetermined voltage that is lower than the voltage of the n-type doped regions 226 and 236. In some other embodiments, the substrate 202 can be floating or not coupled to any external control.

[0111] Figure 2B is an example of a switch optical detector 250, wherein the switch optical detector 250 is used to convert an optical signal into an electrical signal. The switch optical detector 250 is similar to Figure 2A 2. The switch photodetector 200 shown in FIG. 2 is shown in FIG. 2, but the first switch 208 and the second switch 210 further include an n-type well region 252 and an n-type well region 254, respectively. In addition, the absorption layer 206 can be a p-type doped region, and the substrate 202 can be a p-type doped substrate. In some embodiments, the doping amount of the n-type well regions 252 and 254 can range from 10 15 cm -3 Up to 10 17 cm -3 , the doping amount of the substrate 202 can range from 10 14 cm-3 Up to 10 16 cm -3 .

[0112] The arrangement of the p-type doping region 228, the n-type well region 252, the p-type doping substrate 202, the n-type well region 254, and the p-type doping region 238 forms a PNPNP junction structure. In general, the PNPNP junction structure selectively reduces a leakage current from the first control signal 222 to the second control signal 232, or a leakage current from the second control signal 232 to the first control signal 222. The arrangement of the n-type doping region 226, the p-type doping substrate 202, and the n-type doping region 236 forms an NPN junction structure. In general, the NPN junction structure selectively reduces a charge coupling from the first readout circuit 224 to the second readout circuit 234, or a charge coupling from the second readout circuit 234 to the first readout circuit 224.

[0113] In some embodiments, the p-type doped region 228 is formed entirely within the n-type well region 252. In some embodiments, the p-type doped region 228 is formed partially within the n-type well region 252; for example, a portion of the p-type doped region 228 can be formed in the n-type well region 252 by implanting a p-type dopant, and another portion of the p-type doped region 228 can be formed in the substrate 202 by implanting a p-type dopant. Similarly, in some embodiments, the p-type doped region 238 is formed entirely within the n-type well region 254. In some embodiments, the p-type doped region 238 is formed partially within the n-type well region 254. In some embodiments, the depth of the n-type well region 252 is shallower than the depths of the p-type doped regions 228 and 238.

[0114] Figure 2C is an example of a switch photodetector 260, wherein the photodetector 260 is used to convert an optical signal into an electrical signal. The switch photodetector 260 is similar to Figure 2A 2. The switch photodetector 200 shown in FIG. 2 is shown in FIG. 2, but the substrate 202 further includes an n-type well region 244. In addition, the absorption layer 206 can be a p-type doped region, and the substrate 202 can be a p-type doped substrate. In some embodiments, the doping amount of the n-type well region 244 can range from 10 15 cm -3 Up to 10 17 cm -3 , the doping amount of the absorption layer 206 and the substrate 202 can range from 10 14 cm -3 Up to 10 16 cm -3 .

[0115] The arrangement of the p-type doping region 228, the n-type well region 244, and the p-type doping region 238 forms a PNP junction structure. In general, the PNP junction structure selectively reduces a leakage current from the first control signal 222 to the second control signal 232, or a leakage current from the second control signal 232 to the first control signal 222. The arrangement of the n-type doping region 226, the p-type doped substrate 202, and the n-type doping region 236 forms an NPN junction structure. In general, the NPN junction structure selectively reduces a charge coupling from the first readout circuit 224 to the second readout circuit 234, or a charge coupling from the second readout circuit 234 to the first readout circuit 224. In some embodiments, when the depth of the n-type well region 244 is deep enough, the arrangement of the n-type doped region 226, the p-type doped substrate 202, the n-type well region 244, the p-type doped substrate 202, and the n-type doped region 236 can form an NPNPN junction structure to further reduce a charge coupling from the first readout circuit 224 to the second readout circuit 234, or a charge coupling from the second readout circuit 234 to the first readout circuit 224. In some embodiments, the n-type well region 244 also effectively reduces the potential energy barrier that electrons can experience when flowing from the absorption layer 206 to the substrate 202.

[0116] In some embodiments, the p-type doped regions 228 and 238 are completely formed in the n-type well region 244. In some other embodiments, the p-type doped regions 228 and 238 are partially formed in the n-type well region 244; for example, a portion of the p-type doped region 228 can be formed in the n-type well region 244 by implanting a p-type dopant, and another portion of the p-type doped region 228 can be formed in the substrate 202 by implanting a p-type dopant. In some embodiments, the depth of the n-type well region 244 is shallower than the depth of the p-type doped regions 228 and 238.

[0117] Figure 2D is an example of a switch photodetector 270, wherein the photodetector 270 is used to convert an optical signal into an electrical signal. The switch photodetector 270 is similar to Figure 2A 200, but further includes one or more p-type well regions 246 and one or more p-type well regions 248. In some embodiments, the one or more p-type well regions 246 and the one or more p-type well regions 246 can be part of a ring structure; the ring structure surrounds the first switch 208 and the second switch 210. In some embodiments, the doping amount of the one or more p-type well regions 246 and 248 can range from 10 15 cm -3 Up to 10 20 cm -3The one or more p-type well regions 246 and 248 described above isolate photoelectrons from adjacent pixels.

[0118] although Figures 2A-2D Not shown, but in some embodiments, the optical signal can enter the switch optical detector from the back side of the switch optical detector substrate 202. One or more optical elements (e.g., microlenses or light guides) can be fabricated on the back side of the substrate 202, where the optical signal is focused, straightened, defocused, filtered, or otherwise processed according to the lens design.

[0119] Despite Figures 2A-2D Not shown, but in some embodiments, the first switch 208 and the second switch 210 may be fabricated in an alternative manner to collect holes rather than electrons; in this case, the p-type doped region 228, the p-type doped region 238, and the p-type doped region 209 would be replaced by n-type doped regions, the n-type doped region 226 and the n-type doped region 236 would be replaced by p-type doped regions, the n-type well regions 252, 254, and 244 would be replaced by p-type well regions, and the p-type well regions 246 and 248 would be replaced by n-type well regions.

[0120] although Figures 2A-2D Not shown, but in some embodiments, the absorber layer 206 may be bonded to a substrate after forming the switch light detectors 200, 250, 260, and 270. The carrier substrate may be any material that allows the light signal 212 to be transmitted to the switch light detector; for example, the substrate may be a polymer or glass. In some embodiments, one or more optical elements (e.g., microlenses or light guides) may be fabricated on the carrier substrate, where the light signal is focused, aligned, defocused, filtered, or otherwise processed according to the lens design.

[0121] although Figures 2A-2D Not shown, but in some embodiments, the switch photodetectors 200, 250, 260, and 270 can be bonded (e.g., by metal-metal bonding, oxide-oxide bonding, hybrid bonding) to a second substrate containing control signal circuitry and / or readout circuitry and / or phase locked loop and / or analog-to-digital converter circuitry. A metal layer can be deposited on top of the switch photodetector to act as a reflector to reflect the light signal incident from the back side of the substrate 202. Adding a metal layer like a reflector can increase the absorption efficiency (quantum efficiency) of the absorption layer 206; for example, by adding a reflective metal layer, the absorption efficiency of the photodetector operating at long NIR wavelengths in the range of 1.0 to 1.6 μm can be greatly improved. An oxide layer can be included between the metal layer and the absorption layer to increase reflectivity. The metal layer can also serve as a bonding layer during the wafer bonding process. In some embodiments, one or more switches similar to the first switch 208 and the second switch 210 can be added to link the control signal / readout circuitry.

[0122] although Figures 2A-2D Not shown, but in some embodiments, the absorption layer 206 may be partially or completely embedded or recessed in the substrate 202 to mitigate the surface topography and facilitate manufacturing. The aforementioned technology is disclosed in the early publication of US Patent No. 20170040362A1.

[0123] Figure 3A is an example of a switch photodetector 300, wherein the switch photodetector 300 is used to convert an optical signal into an electrical signal. Figure 3A In the embodiment of the present invention, first switches 308a and 308b, and second switches 310a and 310b are fabricated on a substrate 302 and are in a vertical arrangement. A feature of the switch photodetector 100 or the switch photodetector 200 is that the larger the optical window size d, the longer the photoelectron transmission time for electrons to drift or diffuse from one switch to the other switch. This affects the operating speed of the switch photodetector. The switch photodetector 300 can further increase the operating speed by vertically arranging the p-type doping region and the n-type doping region. With such a vertical arrangement, the photoelectron transmission distance will be limited by the thickness t of the absorption layer (e.g., about 1 μm) rather than the optical window size d of the absorption layer (e.g., about 10 μm). The switch photodetector 300 includes an absorption layer 306 fabricated on the substrate 302. The substrate 302 can be any substrate suitable for arranging semiconductor devices thereon; for example, the substrate 302 can be a silicon substrate.

[0124] In general, the absorption layer 306 receives an optical signal 312 and converts the optical signal 312 into a plurality of electrical signals. The absorption layer 306 is similar to the absorption layer 206. The absorption layer 306 can be intrinsic, p-type, or n-type. In some embodiments, the absorption layer 206 can be formed of a p-type silicon germanium material. In some embodiments, the absorption layer 306 can include a p-type doped region 309; the p-type doped region 309 is similar to the p-type doped region 209.

[0125] The first switches 308a and 308b, and the second switches 310a and 310b are fabricated in the substrate 302. It should be noted that although Figure 3A Two first switches 308a and 308b and two second switches 310a and 310b are shown, but the number of first switches and second switches can be more or less. The first switches 308a and 308b are coupled to a first control signal 322 and a first readout circuit 324, and the second switches 310a and 310b are coupled to a second control signal 332 and a second readout circuit 334.

[0126] Generally speaking, the first readout circuit 324 or the second readout circuit 334 decides to collect electrons or holes according to the control of the first control signal 322 and the second control signal 332. The first control signal 322 is similar to the first control signal 122, the second control signal 332 is similar to the second control signal 132, the first readout circuit 324 is similar to the first readout circuit 124, and the second readout circuit 334 is similar to the second readout circuit 134. In some embodiments, the first switches 308 and 308b, and the second switches 310a and 310b are fabricated to collect a plurality of electrons generated by the absorption layer 306. In such a case, the first switches 308a and 308b include p-type doped regions 328a, 328b and n-type doped regions 326a, 326b, respectively. For example, the p-type doped regions 328a and 328b may have a p+ doping, wherein the activated dopant concentration may be as high as the fabrication technology can achieve; for example, when the substrate 302 is silicon and doped with boron, it is about 2×10 20 cm -3 In some embodiments, the doping concentration of the p-type doping regions 328a and 328b may be less than 2×10 20 cm -3 , thereby simplifying the manufacturing complexity (although increasing the contact resistance). The n-type doped regions 326a and 326b may have an n+ doping, wherein the activated dopant concentration may be as high as the manufacturing technology can achieve; for example, when the substrate 302 is silicon and is doped with phosphorus, about 5×10 20 cm -3 In some embodiments, the doping concentration of the n-type doping regions 326a and 326b may be less than 5×10 20 cm -3 , thereby simplifying manufacturing complexity (although increasing contact resistance). The distance between the p-type doping region 328a and the n-type doping region 326a depends on the process design rules; for example, the distance between the p-type doping region 328a and the n-type doping region 326a can be controlled based on the energy associated with the implanted dopant. Generally speaking, the closer the distance between the p-type doping region 328a / 328b and the n-type doping region 326a / 326b, the higher the switching efficiency of generating photocarriers. The second switches 310a and 310b respectively include p-type doping regions 338a and 338b, and n-type doping regions 336a and 336b. The p-type doping region 338a / 338b is similar to the p-type doping region 328a / 328b, and the n-type doping region 336a / 336b is similar to the n-type doping region 326a / 326b.

[0127] In some embodiments, p-type doped regions 328a and 328b are coupled to a first control signal 322. N-type doped regions 326a and 326b are coupled to a readout circuit 324. P-type doped regions 338a and 338b are coupled to a second control signal 332. N-type doped regions 336a and 336b are coupled to a readout circuit 332. The first control signal 322 and the second control signal 332 are used to control the collection process of a plurality of electrons generated by absorbing photons. For example, when the absorption layer 306 absorbs photons in the optical signal 312, electron-hole pairs are generated and drift or diffuse to the substrate 302. When a voltage is applied, if the first control signal 322 is different from the second control signal 332, multiple electric fields are established between the p-type doping region 309 and the p-type doping region 328a / 328b or the p-type doping region 338a / 338b, and free electrons drift from the absorption layer 306 to the p-type doping region 328a / 328b or the p-type doping region 338a / 338b according to the direction of the electric field. In some embodiments, the first control signal 322 can be fixed at a voltage value Vi, and the second control signal 332 can be changed between Vi±ΔV. The drift direction of the electrons depends on the bias value. Accordingly, when one set of switches (e.g., the first switches 308a and 308b) is turned on (i.e., the electrons drift to the p-type doping regions 328a and 328b), the other set of switches (e.g., the second switches 310a and 310b) is turned off (i.e., the electrons are blocked by the p-type doping regions 338a and 338b). In some implementations, the first control signal 322 and the second control signal 332 can have different voltage values.

[0128] In each first switch 308a / 308b, an electric field is established between the p-type doping region 328a / 328b and the n-type doping region 326a / 326b. Similarly, in each second switch 310a / 310b, an electric field is established between the p-type doping region 338a / 338b and the n-type doping region 336a / 336b. When the first switches 308a and 308b are turned on and the second switches 310a and 310b are turned off, electrons will drift to the p-type doping regions 328a and 328b, and the electric field between the p-type doping region 328a and the n-type doping region 326a will further move the electrons to the n-type doping region 326a. Similarly, the electric field between the p-type doping region 328b and the n-type doping region 326b will also move the electrons to the n-type doping region 326b. The readout circuit 324 can then process the charges collected by the n-type doping regions 326a and 326b. Conversely, when the second switches 310a and 310b are turned on and the first switches 308a and 308b are turned off, the electrons will drift to the p-type doping regions 338a and 338b, and the electric field between the p-type doping region 338a and the n-type doping region 336a will further move the electrons to the n-type doping region 336a. Similarly, the electric field between the p-type doping region 338b and the n-type doping region 336b also moves the electrons to the n-type doping region 336b. The readout circuit 334 can then process the charges collected by the n-type doping regions 336a and 336b.

[0129] In some implementations, a voltage may be applied between a p-type doped region and an n-type doped region of a switch so that the switch operates in an avalanche mechanism to increase the sensitivity of the switch photodetector 300. For example, in the case where the substrate 302 comprises silicon germanium, when the distance between the p-type doped region 328a and the n-type doped region 326a is approximately 100 nm, a voltage of less than 7 volts may be applied to establish an avalanche gain between the p-type doped region 328a and the n-type doped region 326a.

[0130] In some embodiments, the p-type doped region 309 can be coupled to an external control 314, for example, the p-type doped region 309 can be coupled to electrical ground. In some embodiments, the p-type doped region 309 can be floating or not coupled to any external control. In some embodiments, the substrate 302 can be coupled to an external control 316; for example, the substrate 302 can be coupled to electrical ground. In some embodiments, the substrate 302 can be floating or not coupled to any external control.

[0131] Figure 3B is an example of a switch optical detector 360, wherein the switch optical detector 360 is used to convert an optical signal into an electrical signal. The switch optical detector 360 is similar to Figure 3AThe illustrated switch photodetector 300, however, the switch photodetector 360 further includes an n-type well region 344. In addition, the absorption region 360 can be a p-type doped region, and the substrate can be a p-type doped substrate. In some embodiments, the doping amount of the n-type well region 344 can range from 10 15 cm -3 Up to 10 17 cm -3 , the doping amount of the substrate 302 can range from 10 14 cm 3 Up to 10 16 cm -3 .

[0132] The arrangement of the p-type doping region 328a, the n-type well region 344, and the p-type doping region 338a forms a PNP junction structure; similarly, the arrangement of the p-type doping region 328b, the n-type well region 344, and the p-type doping region 338b also forms a PNP junction structure. In general, the PNP junction structure selectively reduces a leakage current from the first control signal 322 to the second control signal 332, or a leakage current from the second control signal 332 to the first control signal 322. The arrangement of the n-type doping region 326a, the p-type doping substrate 302, and the n-type doping region 336a forms an NPN junction structure; similarly, the arrangement of the n-type doping region 326b, the p-type doping substrate 302, and the n-type doping region 336b also forms an NPN junction structure. In general, the NPN junction structure selectively reduces a charge coupling from the first readout circuit 324 to the second readout circuit 334, or a charge coupling from the second readout circuit 334 to the first readout circuit 324. In some embodiments, n-type well region 344 is also effective in reducing the potential barrier that electrons can experience when flowing from absorber layer 306 to substrate 302 .

[0133] In some embodiments, the p-type doping regions 328a, 338a, 328b, and 338b are completely formed in the n-type well region 344. In some other embodiments, the p-type doping regions 328a, 338a, 328b, and 338b are partially formed in the n-type well region 344. For example, a portion of the p-type doping region 328a can be formed in the n-type well region 344 by implanting a p-type dopant, and another portion of the p-type doping region 328a can be formed in the substrate 302 by implanting a p-type dopant. In some embodiments, the depth of the n-type well region 344 is shallower than the depth of the p-type doping regions 328a, 338a, 328b, and 338b.

[0134] Figure 3C is an example of a switch photodetector 370, wherein the photodetector 370 is used to convert an optical signal into an electrical signal. The switch photodetector 370 is similar to Figure 3A300, but further includes one or more p-type well regions 346 and one or more p-type well regions 348. In some embodiments, the one or more p-type well regions 346 and one or more p-type well regions 348 can be part of a ring structure; the ring structure surrounds the first switches 308a, 308b and the second switches 310a, 310b. In some embodiments, the doping amount of the one or more p-type well regions can range from 10 15 cm -3 Up to 10 20 cm -3 The one or more p-type well regions 346 and 348 described above isolate photoelectrons from adjacent pixels.

[0135] Figure 3D is a cross-sectional view of an example of a switch light detector 380. Figure 3D In the embodiment, the p-type doping regions 328 a and 328 b of the first switches 308 a and 308 b and the p-type doping regions 338 a and 338 b of the second switches 310 a and 310 b may be arranged on a first plane 362 of the substrate 302 in an interdigitated arrangement. Figure 3D It is also shown that the n-type doping regions 326 a and 326 b of the first switches 308 a and 308 b and the n-type doping regions 336 a and 336 b of the second switches 310 a and 310 b may be arranged in an interdigitated manner on a second plane 364 of the substrate 302 .

[0136] although Figures 3A-3D Not shown, but in some embodiments, the optical signal can enter the switch optical detector from the back side of the switch optical detector substrate 302. One or more optical elements (e.g., microlenses or light guides) can be fabricated on the back side of the substrate 302, and the optical signal is focused, straightened, defocused, filtered, or processed according to the lens design at the optical element.

[0137] although Figures 3A-3D Not shown, but in some embodiments, the first switches 308a and 308b, and the second switches 310a and 310b may be alternatively fabricated to collect holes rather than electrons; in this case, the p-type doped regions 328a and 328b, the p-type doped regions 338a and 338b, and the p-type doped region 309 would be replaced by n-type doped regions, the n-type doped regions 326a and 326b, and the n-type doped regions 336a and 336b would be replaced by p-type doped regions, the n-type well region 344 would be replaced by a p-type well region, and the p-type well regions 346 and 348 would be replaced by n-type well regions.

[0138] although Figures 3A-3DNot shown, but in some other embodiments, the absorber layer 306 may be bonded to a substrate after the switch photodetectors 300, 360, 370, and 380 are fabricated. The substrate may be any material that allows the light signal 312 to be transmitted to the switch photodetector; for example, the substrate may be a polymer or glass. In some embodiments, one or more optical elements may be fabricated on the carrier substrate, and the light signal may be focused, aligned, defocused, filtered, or otherwise processed according to the lens design.

[0139] although Figures 3A-3D Not shown, but in some embodiments, the switch photodetectors 300, 360, 370, and 380 can be bonded (e.g., by metal-metal bonding, oxide-oxide bonding, hybrid bonding) to a second substrate containing control signal circuitry and / or readout circuitry and / or phase locked loop and / or analog-to-digital converter circuitry. A metal layer can be deposited on top of the switch photodetector to act as a reflector to reflect the light signal incident from the back side of the substrate 302. Adding a metal layer like a reflector can increase the absorption efficiency (quantum efficiency) of the absorption layer 306; for example, by adding a reflective metal layer, the absorption efficiency of the photodetector operating at long NIR wavelengths in the range of 1.0 to 1.6 μm can be greatly improved. An oxide layer can be included between the metal layer and the absorption layer to increase reflectivity. The metal layer can also serve as a bonding layer during the wafer bonding process. In some embodiments, one or more switches similar to the first switch 308a (or 308b) and the second switch 310a (or 310b) can be added to combine the control signal / readout circuitry.

[0140] although Figures 3A-3D Not shown, but in some embodiments, the absorption layer 306 may be partially or completely embedded or recessed in the substrate 302 to mitigate the surface topography and facilitate manufacturing. The aforementioned technology is disclosed in the early publication of US Patent No. 20170040362A1.

[0141] Figure 4A 4 is an example of a switch photodetector 400, wherein the switch photodetector 400 is used to convert an optical signal into an electrical signal. The switch photodetector 400 includes an absorption layer 406 formed on a substrate 402. The substrate 402 can be any substrate suitable for disposing semiconductor devices thereon. For example, the substrate 402 can be a silicon substrate. The absorption layer 406 includes a first switch 408 and a second switch 410.

[0142] In general, the absorption layer 406 receives an optical signal 412 and converts the optical signal 412 into a plurality of electrical signals. The absorption layer 406 can be intrinsic, p-type, or n-type. In some embodiments, the absorption layer 406 can be formed of a p-type germanium silicon material. The absorption layer 406 is selected to have a high absorption coefficient in a desired wavelength range. For NIR wavelengths, the absorption layer 406 can be a germanium silicon platform, where the germanium silicon absorbs photons in the optical signal 412 and generates electron-hole pairs. The material composition of germanium and silicon in the germanium silicon platform can be selected for a specific technology or application. In some embodiments, the absorption layer 406 is designed to have a thickness t; for example, for a wavelength of 850nm or 940nm, in order to have a large quantum efficiency, the thickness of the germanium silicon platform can be about 1μm. In some embodiments, the surface of the absorption layer 406 is designed to have a specific shape; for example, the germanium silicon platform can be circular, square, or rectangular, depending on the spatial profile of the optical signal 412 on the surface of the germanium silicon platform. In some embodiments, the absorption layer 106 is designed to have a lateral dimension d for receiving the optical signal 412; for example, the SiGe platform can be circular or rectangular, where d ranges from 1 μm to 50 μm.

[0143] The first switch 408 and the second switch 410 are fabricated in the absorption layer 406 and the substrate 402. The first switch 408 is coupled to a first control signal 422 and a first readout circuit 424. The second switch 410 is coupled to a second control signal 432 and a second readout circuit 434. Generally speaking, the first readout circuit 424 or the second readout circuit 434 determines to collect electrons or holes according to the control of the first control signal 422 and the second control signal 432.

[0144] In some embodiments, the first switch 408 and the second switch 410 are fabricated to collect electrons. In the foregoing case, the first switch 408 includes a p-type doped region 428 implanted in the absorber layer 406, and an n-type doped region 426 implanted in the substrate 402. For example, the p-type doped region 428 may have a p+ doping, wherein the activated dopant concentration may be as high as the fabrication technology allows; for example, when the absorber layer 106 is germanium and doped with boron, about 5×10 20 cm -3 In some embodiments, the doping concentration of the p-type doping region 428 may be less than 5×10 20 cm -3 , thereby simplifying the manufacturing complexity (although increasing the contact resistance). The n-type doped region 426 may have an n+ doping, wherein the activated dopant concentration may be as high as the manufacturing technology can achieve; for example, when the substrate 402 is silicon and is doped with phosphorus, it is about 5×10 20 cm -3In some embodiments, the doping concentration of the n-type doping region 426 may be less than 5×10 20 cm -3 , thereby simplifying the manufacturing complexity (although increasing the contact resistance). The distance between the p-type doping region 428 and the n-type doping region 426 depends on the process design rules. Generally speaking, the closer the distance between the p-type doping region 428 and the n-type doping region 426, the higher the switching efficiency of the generated photocarriers. The second switch 410 includes a p-type doping region 438 and an n-type doping region 436, the p-type doping region 438 is similar to the p-type doping region 428, and the n-type doping region 436 is similar to the n-type doping region 426.

[0145] In some embodiments, the p-type doped region 428 is coupled to the first control signal 422; for example, the p-type doped region 448 can be coupled to a voltage source, wherein the first control signal 422 can be an AC voltage signal from the voltage source. In some embodiments, the n-type doped region 426 is coupled to a readout circuit 424. The readout circuit 424 is a three-transistor configuration consisting of a reset gate, a source follower, and a select gate, or any circuit suitable for handling charge. In some embodiments, the readout circuit 424 can be fabricated on the substrate 402. In some other embodiments, the readout circuit 424 can be fabricated on another substrate and integrated or co-packaged with the switch photodetector 400 using chip / wafer bonding or stacking technology.

[0146] The p-type doped region 438 is coupled to the second control signal 432; for example, the p-type doped region 438 can be coupled to a voltage source; wherein the second control signal 462 can be an AC voltage signal, and its phase is opposite to the first control signal 422. In some embodiments, the n-type doped region 436 is coupled to the readout circuit 434, which can be similar to the readout circuit 424.

[0147] The first control signal 422 and the second control signal 432 are used to control the collection process of multiple electrons generated by absorbing photons. For example, when the first control signal 422 is different from the second control signal 432, an electric field is formed between the p-type doping region 428 and the p-type doping region 438, and the free electrons drift to the p-type doping region 428 or the p-type doping region 438 according to the direction of the electric field. In some embodiments, the first control signal 422 can be fixed at a voltage value Vi, and the second control signal 432 can be changed between Vi±ΔV. The drift direction of the electron depends on the bias value. Accordingly, when a switch (for example, the first switch 408) is turned on (that is, the electrons drift to the p-type doping region 428), the other switch (for example, the second switch 410) is turned off (that is, the electrons are blocked by the p-type doping region 438). In some embodiments, the first control signal 422 and the second control signal 432 may have different voltage values.

[0148] In general, the difference between the Fermi level of the p-type doped region and the Fermi level of the n-type doped region (before equilibrium) forms an electric field between the two regions. In the first switch 408, the electric field is formed between the p-type doped region 428 and the n-type doped region 426. Similarly, in the second switch 410, the electric field is formed between the p-type doped region 438 and the n-type doped region 436. When the first switch 408 is turned on and the second switch 410 is turned off, electrons drift to the p-type doped region 428, and the electric field between the p-type doped region 428 and the n-type doped region 426 further transfers the electrons to the n-type doped region 426. The readout circuit 424 can then process the charges collected by the n-type doped region 426. Conversely, when the second switch 410 is turned on and the first switch 408 is turned off, the electrons will drift to the p-type doping region 438, and the electric field between the p-type doping region 438 and the n-type doping region 436 will further transfer the electrons to the n-type doping region 436. The readout circuit 434 can then process the charges collected by the n-type doping region 436.

[0149] In some embodiments, the substrate 402 is coupled to an external control 416. For example, the substrate 402 can be coupled to an electrical ground. In some embodiments, the substrate 402 can be floating or not coupled to any external control.

[0150] Figure 4B is an example of a switch optical detector 450, wherein the switch optical detector 450 is used to convert an optical signal into an electrical signal. The switch optical detector 450 is similar to Figure 4A 4. The switch photodetector 400 shown in FIG. 4 is shown in FIG. 4, but the first switch 408 and the second switch 410 further include an n-type well region 452 and an n-type well region 454, respectively. In addition, the absorption region 406 can be a p-type doped layer, and the substrate 402 can be a p-type doped substrate. In some embodiments, the doping amount of the n-type well region 452 can range from 10 15 cm -3 Up to 10 17 cm 3 , the doping amount of the substrate 402 can range from 10 14 cm -3 Up to 10 16 cm 3 .

[0151] The arrangement of the p-type doping region 428, the n-type well region 452, the absorption region 406, the n-type well region 454 and the p-type doping region 438 forms a PNPNP junction structure. Generally speaking, the PNPNP junction structure selectively reduces a leakage current from the first control signal 422 to the second control signal 432, or from the second control signal 432 to the first control signal 422.

[0152] The arrangement of the n-type doped region 426, the p-type doped substrate 402 and the n-type doped region 436 forms an NPN junction structure. Generally, the NPN junction structure selectively reduces a charge coupling from the first readout circuit 424 to the second readout circuit 434 or from the second readout circuit 434 to the first readout circuit 424.

[0153] In some embodiments, the p-type doped region 428 is formed entirely within the n-type well region 452. In some other embodiments, the p-type doped region 428 is partially formed in the n-type well region 452; for example, a portion of the p-type doped region 428 can be formed in the n-type well region 452 by implanting a p-type dopant, and another portion of the p-type doped region 428 can be formed in the absorber layer 406 by implanting a p-type dopant. Similarly, in some other embodiments, the p-type doped region 438 is formed entirely in the n-type well region 454. In some embodiments, the p-type doped region 438 is partially formed in the n-type well region 454. In some embodiments, the depths of the n-type well regions 452 and 454 are shallower than the depths of the p-type doped regions.

[0154] Figure 4C is an example of a switch photodetector 460, wherein the photodetector 460 is used to convert an optical signal into an electrical signal. The switch photodetector 460 is similar to Figure 4A 4. The switch photodetector 400 is shown in FIG. 4, but the absorption layer 406 further includes an n-type well region 456. In addition, the absorption region 406 can be a p-type doped region, and the substrate 402 can be a p-type doped substrate. In some embodiments, the doping amount of the n-type well region 456 can range from 10 15 cm -3 Up to 10 17 cm -3 The doping amount of the absorption layer 406 and the substrate 402 can be in the range of 10 14 cm 3 Up to 10 16 cm -3 .

[0155] The arrangement of the p-type doping region 428, the n-type well region 456 and the p-type doping region 438 forms a PNP junction structure. Generally speaking, the PNP junction structure selectively reduces a leakage current from the first control signal 422 to the second control signal 432 or from the second control signal 432 to the first control signal 422.

[0156] The arrangement of the n-type doped region 426, the p-type doped absorption layer 406 and the n-type doped region 436 forms an NPN junction structure. Generally speaking, the NPN junction structure selectively reduces a charge coupling from the first readout circuit 424 to the second readout circuit 434 or from the second readout circuit 434 to the first readout circuit 424.

[0157] In some embodiments, p-type doped regions 428 and 438 are formed completely within n-type well region 456. In some other embodiments, p-type doped regions 428 and 438 are formed partially in n-type well region 456; for example, a portion of p-type doped region 428 can be formed in n-type well region 456 by implanting p-type dopants, and another portion of p-type doped region 428 can be formed in absorber layer 406 by implanting p-type dopants. In some embodiments, the depth of n-type well region 456 is shallower than the depth of p-type doped regions 428 and 438.

[0158] Figure 4D is an example of a switch photodetector 470, wherein the photodetector 470 is used to convert an optical signal into an electrical signal. The switch photodetector 470 is similar to Figure 4C 4. The switching photodetector 460 is shown, but its n-well region 458 extends from the absorption region 406 to the substrate 402. In addition, the absorption region 406 can be a p-type doped region, and the substrate 402 can be a p-type doped substrate. In some embodiments, the doping amount of the n-type well region 456 can range from 10 15 cm -3 Up to 10 17 cm -3 The doping amount of the absorption layer 406 and the substrate 402 can range from 10 14 cm -3 Up to 10 16 cm -3 .

[0159] The arrangement of the p-type doping region 428, the n-type well region 458, and the p-type doping region 438 forms a PNP junction structure, which can further reduce a leakage current from the first control signal 422 to the second control signal 432, or from the second control signal 432 to the first control signal 422. The n-type doping region 426, the p-type doped substrate 402, the n-type well region 458, the p-type doped substrate 402, and the n-type doping region 436 form an NPNPN junction structure, thereby selectively reducing the charge coupling from the first readout circuit 424 to the second readout circuit 434, or from the second readout circuit 434 to the first readout circuit 424. In some embodiments, the n-type well region 458 can effectively reduce the potential energy barrier that electrons can perceive when flowing from the absorption layer 406 to the substrate 402.

[0160] Figure 4E is an example of a switch photodetector 480, wherein the photodetector 480 is used to convert an optical signal into an electrical signal. The switch photodetector 480 is similar to Figure 4A400, but the switch photodetector 480 further includes one or more p-type well regions 446 and one or more p-type well regions 448. In some embodiments, the one or more p-type well regions 446 and one or more p-type well regions 448 can be part of a ring structure; the ring structure surrounds the first switch 408 and the second switch 410. In some embodiments, the doping amount of the p-type well regions 446 and 448 can range from 10 15 cm -3 Up to 10 20 cm -3 The one or more p-type well regions 446 and 448 described above isolate photoelectrons from adjacent pixels.

[0161] although Figures 4A-4D Not shown, but in some embodiments, the optical signal may enter the switch optical detector from the back side of the switch optical detector substrate 402. One or more optical elements (e.g., microlenses or light guides) may be fabricated on the back side of the substrate 402, where the optical signal is focused, straightened, defocused, filtered, or otherwise processed according to the lens design.

[0162] although Figures 4A-4E Not shown, but in some embodiments, the first switch 408 and the second switch 410 may be fabricated in an alternative manner to collect holes rather than electrons; in this case, the p-type doped region 428 and the p-type doped region 438 would be replaced by n-type doped regions, the n-type doped region 426 and the n-type doped region 436 would be replaced by p-type doped regions, the n-type well regions 452, 454, 456, and 458 would be replaced by p-type well regions, and the p-type well regions 446 and 448 would be replaced by n-type well regions.

[0163] although Figures 4A-4E Not shown, but in some embodiments, the absorber layer 406 may be bonded to a substrate after forming the switch photodetectors 400, 450, 460, 470, and 480. The substrate may be any material that allows the photoelectric signal 412 to be transmitted to the switch photodetector; for example, the substrate may be a polymer or glass. In some embodiments, one or more optical elements (e.g., microlenses or light guides) may be fabricated on the carrier substrate, and the optical signal may be focused, aligned, defocused, filtered, or otherwise processed according to the lens design.

[0164] although Figures 4A-4ENot shown, but in some embodiments, the switch photodetectors 400, 450, 460, 470 and 480 can be bonded (e.g., by metal-metal bonding, oxide-oxide bonding, hybrid bonding) to a second substrate containing control signal circuitry and / or readout circuitry and / or phase-locked loop and / or analog-to-digital converter circuitry. A metal layer can be deposited on top of the switch photodetector to act as a reflector to reflect the light signal incident from the back side of the substrate 402. Adding a metal layer like a reflector can increase the absorption efficiency (quantum efficiency) of the absorption layer 406; for example, by adding a reflective metal layer, the absorption efficiency of the photodetector operating in the long NIR wavelength range of 1.0 to 1.6 μm can be greatly improved. An oxide layer can be included between the metal layer and the absorption layer to increase reflectivity. The metal layer can also serve as a bonding layer during the wafer bonding process. In some embodiments, one or more switches similar to the first switch 408 and the second switch 410 can be added to link the control signal / readout circuitry.

[0165] although Figures 4A-4E Not shown, but in some embodiments, the absorption layer 406 may be partially or completely embedded or recessed in the substrate 402 to mitigate the surface topography and facilitate manufacturing. The aforementioned technology is disclosed in the early publication of US Patent No. 20170040362A1.

[0166] Figure 4F-4I is an example of a design 490 for selectively forming an absorber layer on a substrate. Design 490 is, for example, an example of a design for manufacturing Figures 1A-4E The switching photodetector is shown. Figure 4F In the embodiment of the present invention, a groove 492 is formed on a substrate 402. The groove 492 can define a light detector area for a NIR pixel. The groove can be formed by lithography and then dry etching the substrate 402. The shape of the groove can correspond to the shape of the pixel and can be, for example, a square, a circle, or other suitable shape.

[0167] exist Figure 4G In the embodiment of the present invention, a dielectric layer may be deposited on top of the substrate, and a directional etch may be performed to form sidewall spacers 494. The directional etch may be an anisotropic dry etch. The aforementioned spacers 494 may be a dielectric material (e.g., various oxides and nitrides) and are used to separate the sidewalls of the absorption layer to be formed from the substrate 402. In some embodiments, the spacers 494 may be omitted, and the embedded portion of the absorption layer to be formed may directly contact the surface of the groove 492 formed in the substrate 402, such as the

[110] sidewall of the silicon substrate.

[0168] exist Figure 4HIn the embodiment, the germanium or germanium silicon absorption layer 496 can be selectively grown from the substrate 402. For example, the absorption layer 496 can be formed by epitaxial growth using a chemical vapor deposition (CVD) system. The formed absorption layer 496 is partially embedded in the groove 492 of the substrate 402. The absorption layer 496 can be, for example, Figures 1A-4E The absorbing layer of the switching photodetector described in .

[0169] exist Fig. 4I In the embodiment, the germanium or silicon germanium absorption layer 496 and the substrate 402 are planarized (or flattened) to form a fully embedded absorption layer 496. The germanium or silicon germanium absorption region 496 can be planarized by chemical-mechanical polishing (CMP) or any other suitable technique. In some embodiments, if the original surface morphology of the germanium or silicon germanium absorption layer 496 and the surface of the substrate 402 is acceptable to subsequent manufacturing process steps, the step of planarizing the germanium or silicon germanium absorption layer 496 relative to the surface of the substrate 402 can be omitted.

[0170] Figure 5A 5 is an example of a switch photodetector 500, wherein the switch photodetector 500 is used to convert an optical signal into an electrical signal. The switch photodetector 500 includes an absorption layer 506 formed on a substrate 502, and a first layer 508 formed above the absorption layer 506 and the substrate 502. The substrate 502 can be similar to the substrate 102 described above, and the absorption layer 506 can be similar to the absorption layer 106 described above, and can be formed of, for example, germanium or germanium silicon, and the concentration of the germanium component ranges from 1 to 99%. The background doping polarity of the germanium or germanium silicon absorption layer can be p-type and the doping amount can range from 10 14 cm -3 Up to 10 16 cm -3 The background doping level may be determined, for example, by explicit incorporation of doping or by material defects introduced when forming the absorption layer 506. The absorption layer 506 of the photodetector 500 has a platform structure supported by a substrate. Although the sidewalls are vertical in the illustrated example, the sidewall profile of the platform may vary depending on the growth and fabrication process of the absorption layer 506.

[0171] The first layer 508 covers an upper surface and multiple side surfaces of the absorption layer 506, and a portion of an upper surface of the substrate 502 on which the absorption layer 506 is formed. The first layer 508 can be formed of a material compatible with a CMOS process, such as amorphous silicon, polycrystalline silicon, epitaxial silicon, aluminum oxide family (e.g., Al 2 O 3), the silicon oxide family (e.g., SiO 2 ), the germanium oxide family (e.g., GeO 2 ), the Germanium-Silicon family (e.g., Ge 0.4 Si 0.6 ), the silicon nitride family (e.g., Si 3 N 4 ), high-k value materials (e.g., HfOx, ZnOx, LaOx, LaSiOx), and any combination thereof. The first layer 508 present on the surface of the absorption layer 506 may have different effects; for example, the first layer 508 may serve as a surface passivation layer of the absorption layer 506, thereby reducing dark current or leakage current generated by defects present on the surface of the absorption layer 506. In the case where the absorption layer 506 is a germanium absorption layer or a germanium silicon absorption layer, surface defects may be a significant source of dark current or leakage current, which may increase the noise intensity in the photocurrent generated by the photodetector 500. By forming the first layer 508 on the surface of the absorption layer 506, the dark current or leakage current may be reduced, thereby reducing the amount of noise in the photodetector 500. In other examples, the first layer 508 may adjust a Schottky barrier between a contact formed on the photodetector 500 and the absorption layer and / or the substrate 502. The aforementioned barrier adjustment effect will be described in detail as follows.

[0172] Figure 5B is an example of a switch optical detector 510, wherein the switch optical detector 500 is used to convert an optical signal into an electrical signal. The switch optical detector 510 is similar to Figure 5A 500, but the difference is that the absorption layer 506 is partially embedded in a trench formed on the substrate 502, and the photodetector 510 further includes a plurality of spacers 512. The plurality of spacers 512 can be a dielectric material, such as various oxides and nitrides that separate the sidewalls of the absorption layer 506 from the substrate 502. In some embodiments, the spacers 512 can be omitted, so that the embedded portion of the absorption layer 506 directly contacts a surface of the trench formed in the substrate 502, such as a

[110] sidewall of a silicon substrate. The above technology is disclosed in the early publication of U.S. Patent No. 20170040362A1.

[0173] Figure 5C is an example of a switch photodetector 520, wherein the switch photodetector 520 is used to convert an optical signal into an electrical signal. The switch photodetector 520 is similar to Figure 5B The switch photodetector 510 shown is different in that the absorption layer 506 is completely embedded in the groove formed on the substrate 502. The aforementioned technology is disclosed in the early publication of US Patent No. 20170040362A1.

[0174] Figure 5D is an example of a switch photodetector 530, wherein the switch photodetector 530 is used to convert an optical signal into an electrical signal. The switch photodetector 530 is similar to Figure 5B The switch photodetector 510 shown in FIG. 5 is different in that a first switch 532 and a second switch 542 are fabricated in the absorption layer 506 and the first layer 508. The first switch 532 is similar to Figure 1A The first switch 108 shown in FIG. 1 is similar to the first switch 108 shown in FIG. 1 , but further includes a first readout contact 535 coupled to a first n-type doped region 534, and a first control contact 538 coupled to a first p-type doped region 537. Similarly, the second switch 542 is similar to Figure 1A 1, but further includes a second readout contact 545 coupled to a second n-type doping region 544, and a second control contact 548 coupled to a second p-type doping region 547. The first p-type doping region 537 and the second p-type doping region 547 can be control regions, and the first n-type doping region 534 and the second n-type doping region 544 can be readout regions. The first readout contact 535 and the second readout contact 545 are respectively connected to corresponding readout circuits (similar to Figure 1A The first control contact 538 and the second control contact 548 are connected to corresponding control signals (similar to the readout circuits 124 and 134 shown). Figure 1A Control signals 122 and 132 are shown).

[0175] Contacts 535, 538, 545 and 548 provide electrical contacts for the corresponding doped regions and may be formed of different conductive materials. Exemplary contact materials include different metal silicides, Ta-TaN type Cu stacks, Ti-TiN type W stacks, aluminum and various combinations of these materials. In some embodiments, readout contacts 535 and 545 and control contacts 538 and 539 may be formed of different materials. Contacts 535, 538, 545 and 548 may have different physical structures. The diameter or width of the contact may be as small as tens of nanometers. Although only a single contact 535, 538, 545 or 548 is shown coupled to a doped region, similar to conventional semiconductor device manufacturing procedures, two or more contacts may be coupled to a doped region to reduce contact resistance or improve reliability.

[0176] Figure 5E is an example of a switch photodetector 550, wherein the switch photodetector 550 is used to convert an optical signal into an electrical signal. The switch photodetector 550 is similar to Figure 5DThe switch photodetector 530 is shown in FIG. 1 , except that the first switch 532 and the second switch 542 further include n-type well regions 539 and 549, and p-type well regions 536 and 546, respectively. The additional n-type well regions and p-type well regions can adjust the electronic and / or optical properties of the photodetector 550. In some embodiments, the doping amount of the n-type well regions 539 and 549, and the p-type well regions 536 and 546 can range from 10 15 cm -3 Up to 10 17 cm -3 .

[0177] The p-type well region 537, the n-type well region 539, a p-type absorption layer 506, the n-type well region 549, and the p-type doping region 547 form a PNPNP junction structure. In general, the PNPNP junction structure can selectively reduce a leakage current from the first control signal 122 to the second control signal 132, or a leakage current from the second control signal 132 to the first control signal 122. The n-type doping region 534, the p-type well region 536, the p-type absorption layer 506, the p-type well region 546, and the n-type doping region 544 form an NPN junction structure. In general, the NPN junction structure can selectively reduce a charge coupling from the first readout circuit 124 to the second readout circuit 134, or a charge coupling from the second readout circuit 134 to the first readout circuit 124.

[0178] In some embodiments, p-type doped region 537 is formed entirely within n-type well region 539. In some other embodiments, p-type doped region 537 is partially formed in n-type well region 539; for example, a portion of p-type doped region 537 can be formed in n-type well region 539 by implanting p-type dopants, and another portion of p-type doped region 128 can be formed in absorber layer 506 by implanting p-type dopants. Similarly, in some embodiments, p-type doped region 547 is formed entirely within n-type well region 549. In some other embodiments, n-type well regions 539 and 549 form a continuous n-type well region that includes at least a portion of p-type doped regions 537 and 547.

[0179] In some embodiments, the n-type doping region 534 is formed completely outside the p-type well region 536. In some other embodiments, the n-type doping region 534 is partially formed in the p-type well region 536; for example, a portion of the n-type doping region 534 can be formed in the p-type well region 536 by implanting an n-type dopant, and another portion of the n-type doping region 534 can be formed in the absorber layer 506 by implanting an n-type dopant. Similarly, in some embodiments, the n-type doping region 544 is formed completely outside the p-type well region 546. In some other embodiments, the n-type doping region 544 is partially formed in the p-type well region 546.

[0180] although Figure 5D and 5E A switching photodetector with a partially embedded absorption region 506 is shown, but the same structure can be used in a photodetector 500 with a non-embedded absorption layer 506, and a photodetector 520 with a fully embedded absorption layer to achieve similar effects.

[0181] For the convenience of drawing, n-type well regions 539 and 549, and p-type well regions 536 and 546 are drawn together; however, in actual implementation, these well regions can be implemented independently or in any combination.

[0182] Fig. 5F is an example of a switch photodetector 560, wherein the switch photodetector 560 is used to convert an optical signal into an electrical signal. The switch photodetector 560 is similar to Figure 5D The difference of the photodetector 530 shown in FIG. 5 is that the p-type doping regions 537 and 547 corresponding to the switches 537 and 547 are omitted; thus, the first control contact 538 and the second control contact 548 form a Schottky junction with the first layer 508. When the semiconductor is not intentionally doped or doped with a moderate dopant concentration, for example, less than 1×10 15 cm -3 A Schottky junction is an electrical junction formed between a metal and a semiconductor. A leakage current path is noted in a region 562 between the first control contact 538 and the second control contact 548 through the first layer 508 and the absorption layer 506. This leakage current path will be referred to as Figure 5G Provide detailed explanation.

[0183] Figure 5G 548. An example of an energy band diagram 570 is shown for a leakage current path formed between control contacts 538 and 548. Energy band diagram 570 illustrates different energy levels of charge carriers, such as an electron 573 and a hole 574, at different locations along a leakage current path. The vertical axis corresponds to an energy level E and the horizontal axis corresponds to a position x along the leakage current path, where the leakage current path is formed between control contacts 538 and 548. An exemplary scenario is shown where the potential of the first control contact 538 is higher than the potential of the second control contact 548 (e.g., the voltage of the first control signal 122 is lower than the voltage of the second control signal 132). The potential difference is represented by the slope of the entire energy band diagram decreasing from the first control contact 538 to the second control contact 548. Figure 5G The energy levels and positional relationships shown in are for illustration purposes only and are not actual numerical values.

[0184] An electron energy barrier 573 and a hole energy barrier 575 are examples of Schottky energy barriers. A Schottky junction is characterized by exhibiting a Schottky energy barrier, which is a potential energy barrier that electrons 572 and holes 574 need to overcome to pass through the Schottky junction. The values ​​of energy barriers 573 and 575 can vary depending on the work function of the materials of contacts 538 and 548. By selecting a suitable combination of materials for the contact and the first layer, the desired electron energy barrier 573 and hole energy barrier 575 can be set.

[0185] Electrons 572 must overcome an electron energy barrier 573 between the first control contact 538 and the first layer 508. By providing a sufficiently high electron energy barrier 573, the potential of the control signal provided to the first control contact is insufficient to overcome the energy barrier 573. Thus, the electron energy barrier 573 can block the electron 572 from being transferred to the absorption layer 506. If the electron 572 can pass through the absorption layer 506 and transfer to the first layer 508 adjacent to the second control contact 548 due to statistical fluctuations of the thermal energy of an electron 572 ("thermionic emission") or quantum tunneling, the electron 573 can overcome the electron energy barrier 573. Another electron energy barrier appears at a junction between the absorption layer 508 and the first layer 508, which further blocks the electron from being transferred to the second control contact 548, thereby reducing a leakage current generated by the electron transfer from the first control contact 538 to the second control contact 548.

[0186] Similarly, the hole 574 must overcome the energy barrier formed between the second control contact 548 and the first layer 508. By providing a sufficiently high hole energy barrier 575, the potential of the control signal 132 provided to the second control contact is insufficient to overcome the energy barrier 575. Thus, the hole energy barrier 575 can block the hole 574 from being transferred to the absorption layer 506. If the hole 574 can pass through the absorption layer 506 and transfer to the first layer 508 adjacent to the first control contact 538 due to statistical fluctuations in the thermal energy of a hole 574 ("thermionic emission") or quantum penetration response, the hole 574 can overcome the hole energy barrier 575. Another hole energy barrier appears at a junction between the absorption layer 508 and the first layer 508, which further blocks the hole 574 from being transferred to the first control contact 538, thereby reducing a leakage current generated by the hole transferring from the second control contact 548 to the first control contact 538.

[0187] When light strikes the absorber layer 506, a photon 576 of the light may be absorbed by an electron in a valence band of the absorber layer 506, thereby creating an electron-hole pair as indicated by the vertical arrow next to the photon 576. The photocurrent formed by the electron in the electron-hole pair is read by the readout contacts 535 and / or 545 corresponding to the readout circuit 124 and / or 134, and does not flow to the control contacts 538 and 548. In this case, the energy barrier formed by the junction between the first layer 508 and the absorber layer 506 can prevent this flow, thereby improving the photocurrent collection efficiency of the readout circuit.

[0188] When the first layer 508 (e.g., amorphous silicon, polysilicon, crystalline silicon, or silicon germanium) is inserted between the control contacts 538, 548 and the absorption layer 506 (e.g., a silicon germanium platform), the Schottky barrier of the metal semiconductor junction is changed, and as described above, the contacts 538 and 548 can partially shield the electrons or holes injected into the first layer 508. The energy consumption of a ToF pixel (such as the switching photodetector described herein) depends in part on the leakage current transmitted between the two control contacts 538 and 548 connected to the two control circuits. In this way, by partially shielding the injected electrons or holes through the contacts 537 and 548, the energy consumption of the ToF can be greatly reduced.

[0189] Figure 5H is an example of a switch photodetector 580, wherein the switch photodetector 560 is used to convert an optical signal into an electrical signal. The switch photodetector 580 is similar to Fig. 5F The switch photodetector 560 shown in FIG. 1 is different in that the switch photodetector 580 further includes n-type doped regions 539 and 549, and p-type well regions 536 and 546. The structures and effects of the n-type well regions 539 and 549, and the p-type well regions 536 and 546 have been described in detail. Figure 5E In addition, the n-type well regions 539 and 549 overlap a portion of the first layer 508 below the control contacts 538 and 548, which increases the voltage drop within the absorption layer 506. Increasing the voltage drop within the absorption layer 506 can increase the electric field strength established within the absorption layer 506, thereby improving the ability of the readout circuit 124 and / or 134 to capture photo-generated electrons through the corresponding readout contacts 535 and / or 545.

[0190] Fig.5I is an example of a switch photodetector 582, wherein the switch photodetector 560 is used to convert an optical signal into an electrical signal. The switch photodetector 582 is similar to Figure 5EThe switch photodetector 550 is shown, except that the first switch 532 is located on the substrate 502 and adjacent to the left side of the absorption region 506, and the second switch 542 is located on the substrate 502 and adjacent to the right side of the absorption region 506. The switch photodetector 582 is similar to the switch photodetector described above, but compared to the electrical contacts formed between the contacts (e.g., read contacts 535 and 545 or control contacts 538 and 548) and the germanium or germanium silicon absorption layer 506, the electrical contacts formed between the contacts and the silicon substrate 502 generally have a lower dark current or leakage current, which can be caused, for example, by the fact that the material defects of the substrate 502 are less than the material defects of the absorption layer 506; this makes the overall dark current or leakage current lower. Figure 5E The photodetector 550 is shown as being lower. In addition, due to the configuration of the switch on the substrate 502, photocarriers generated by absorption of light in the absorption region 506 can be transferred from the absorption region 506 to the substrate 502 before reaching the readout circuits 124 and 134. Depending on the specific geometry of the absorption region 506 and the spacer 512 and the materials thereof, the photocarriers can be conducted through the spacer 512, transferred around the spacer 512, or a combination thereof.

[0191] In some embodiments, p-type doped regions 537 and 547 may be similar to Fig. 5F Although the n-type well regions 539 and 549, and the p-type well regions 536 and 546 are shown together for the sake of drawing convenience, these wells may be omitted, implemented independently, or implemented in any combination.

[0192] Figure 5J is an example of a switch photodetector 586, wherein the switch photodetector 586 is used to convert an optical signal into an electrical signal. The switch photodetector 586 is similar to Fig.5I The switch photodetector 582 is shown, the difference is that the p-type doping regions 537 and 547 corresponding to the switches 532 and 542 are omitted; thus, the first control contacts 538 and 548 form a Schottky junction with the first layer 508. The effect of the Schottky junction has been Fig. 5F -H corresponding paragraphs are described in detail. Due to the changed geometry of the light detector 586 relative to the light detector 506, Figure 5G The energy band diagram 570 shown still applies to the region 562 in the photodetector 586 , but the energy barrier formed in the first layer 508 is now formed by the corresponding first layer 508 , the substrate 502 , and the spacer 512 .

[0193] Although the n-type well regions 539 and 549 and the p-type well regions 536 and 546 are depicted together for the sake of drawing convenience, these wells may be omitted, implemented independently, or implemented in any combination.

[0194] Figure 5Kis an example of a switch photodetector 588, wherein the switch photodetector 588 is used to convert an optical signal into an electrical signal. The switch photodetector 588 is similar to Fig.5I The switch photodetector 582 shown is different in that the first switch 532 further includes a second p-type doping region 537a, a third control contact 538a and a second n-type well region 539a, and the second switch 543 further includes a second p-type doping region 547a, a fourth control contact 548a and a second n-type well region 549a. The third control contact 538a is coupled to the second p-type doping region 537a, and the second n-type well region 539a contacts the second p-type doping region 537a; the fourth control contact 548a is coupled to the second p-type doping region 547a, and the second n-type well region 549a contacts the second p-type doping region 547a. The second p-type doping regions 537a and 537b are similar to the second p-type doping regions 537 and 547, respectively. The second n-type well regions 539a and 549a are similar to the second n-type well regions 539 and 549, respectively. The third control contact 538a is similar to the first control contact 538, and the fourth control contact 548a is similar to the second control contact 548. The third control contact 538a is connected to the first control signal 122, and the fourth control contact 548a is connected to the second control signal 132.

[0195] The first control contact 538 and the associated doped region do not directly contact the absorption region 506, so the electric field created by applying the first control signal 122 to the first control contact 538 within the absorption region 506 is similar to the electric field created by the first control signal 122. Figure 5E The photodetector 550 shown has a weaker electric field than the electric field created by the first control contact 538 directly contacting the absorption layer 506. By adding the third and fourth control contacts 538a and 548a, and the associated doped regions, the carrier collection control efficiency of the photodetector 586 can be improved. Fig.5I The light detector 582 is shown elevated and Figure 5E The photodetector 550 shown has similar carrier collection control efficiency; however, the advantage of reducing dark current or leakage current can still be partially retained by moving the contact to the substrate 502. In addition, the large electric field in the absorption region can increase the bandwidth of the photodetector, speed up the switching of the first switch 532 and the second switch 542, and the additional control contacts 538a and 548a can also increase the operating speed of the photodetector 584.

[0196] Although the third control contact 538a and the fourth control contact 548a are shown to share control signals 122 and 132 with the first control contact 538 and the second control contact 548, respectively, in some embodiments, contacts 538a and 548a may be coupled to control signals different from the first control signal 122 and the second control signal 132; for example, the control signal provided to the third control contact 538a may be less than the first control signal 122 provided to the first control contact 538; due to the proximity of the second p-type doped region 537a to the carriers generated by the absorption region 506, the control signal provided to the third control contact 538a can have a greater effect on the photogenerated carriers than the first control signal 122 provided to the first control contact 538; the same mechanism also applies to the control signal supplied to the fourth control contact 548a.

[0197] In some embodiments, the second p-type doping regions 537a and 547a can be omitted to form a Schottky junction, the effect of which can be seen in the description of FIG.5F-5H. For the convenience of drawing, the n-type well regions 539 and 549, and the p-type well regions 536 and 546 are drawn together; however, in actual implementation, these wells can be implemented independently or in any combination.

[0198] although Figure 5D-5K Various structures of switch photodetectors having partially embedded absorber layer 506 are described, but the structures described can also be applied to switch photodetectors having a fully protruding absorber layer 506 (e.g., Figure 5A ), and a switching photodetector having a structure completely embedded in the absorption layer 506 (eg, Figure 5C ), to achieve a similar effect.

[0199] exist Figure 5A-5K The photodetector described in the above can be incorporated into a front-side illumination (FSI) image sensor or a back-side illumination (BSI) image sensor. In the front-side illumination structure, light enters the photodetector from the top of the first layer 508. In the back-side illumination structure, light enters the photodetector from the bottom of the substrate 502.

[0200] The control region (e.g., p-type doped regions 537 and 547) and the readout region (e.g., n-type doped regions 534 and 544) may have different heights; for example, for photodetectors 530, 550, 560, and 580, and for any structure in which both the control region and the readout region are located in the absorption region 506, a portion of the absorption region 506 corresponding to the readout region or the control region may be etched, and the readout region or the control region may be formed on the etched portion, so that a vertical offset may be formed between the control region and the readout region. Similarly, for photodetectors 582, 586, and 588, and for any structure in which both the control region and the readout region are located in the substrate, a portion of the substrate 502 corresponding to the readout region or the control region may be etched, and the readout region or the control region may be formed on the etched portion, so that a vertical offset may be formed between the control region and the readout region.

[0201] In some embodiments, the lens may be disposed on an optical path of the incident light. The lens may be, for example, a micro ball lens or a Fresnel Zone Plate (FZP) lens. In other examples, when the substrate 502 is a silicon substrate, the lens may be directly formed on the substrate 502 by etching the substrate 502. The detailed structure of the lens will be described in detail in Figure 7A-7B Provide detailed explanation.

[0202] In some embodiments, the interface between the absorber layer 506 and the spacer 512 can be doped with n-type or p-type dopants to improve electrical isolation of holes and electrons. In some embodiments, the interface between the absorber layer 506 and the substrate 502 (e.g., the bottom interface) can be doped with n-type or p-type dopants to improve electrical isolation of holes and electrons.

[0203] Fig. 6A 6 is an example of a switch photodetector 600, wherein the switch photodetector 600 is used to convert an optical signal into an electrical signal. The switch photodetector 600 includes a substrate 502, an absorption region 506, a first switch 532, a second switch 542, and an anti-doping region 610. The anti-doping region 610 is disposed in the absorption region 506, and the first switch 532 and the second switch 542 are disposed on the absorption layer 506. The substrate 502, the absorption region 506, the first switch 532, and the second switch 542 have been Figure 5D Describe in the relevant paragraphs.

[0204] The anti-doped region 610 is a portion of the absorption region 506 that is doped with a dopant species to reduce a net carrier concentration in the absorption region 506. An undoped semiconductor material has a substantial concentration of charge carriers that can contribute to current conduction even in the absence of dopants, and the charge carrier concentration corresponds to the intrinsic carrier concentration of the semiconductor. The absorption region 506 is typically formed of a semiconductor material, such as silicon, germanium, or an alloy of the two, and has an associated intrinsic carrier concentration. The intrinsic carrier concentration may vary depending on various factors, such as material preparation methods and defect levels (defect concentrations). Material preparation methods include epitaxial growth, chemical vapor deposition (CVD), metal organic CVD (MOCVD), and physical vapor deposition (PVD), and different material preparation methods may produce different material defect levels. Generally speaking, the more material defects there are, the higher the intrinsic carrier concentration will be. For example, at room temperature, the intrinsic p-type like carrier concentration of bulk crystalline germanium is about 2×10 13 cm -3 , while epitaxial germanium can have a higher p-type intrinsic carrier concentration by an order of magnitude, about 5×10 14 cm -3 Depending on the material properties and the types of defects, semiconductor materials can be either p-type or n-type.

[0205] Reducing a leakage current of a switched photodetector (e.g., photodetector 600) to reduce power consumption is important for a ToF pixel. A portion of the leakage current of the switched photodetector comes from a leakage current conducted between control regions, such as the current conducted between p-type doped regions 537 and 547. One way to reduce the leakage current is to reduce a net carrier concentration in the absorption region 506 between the p-type doped regions 537 and 547. The net carrier concentration is the carrier concentration that can be obtained when conducting current, and can be determined by combining the contribution of the intrinsic carrier concentration and the extrinsic carrier concentration of impurities. By properly selecting the electrical type, type, and concentration of the impurities, the intrinsic carrier concentration can be compensated, or the semiconductor material can be "counter-doped" with dopants to have a lower net carrier concentration. In general, when the intrinsic carriers and the net carriers have the same polarity, such as both p-type or n-type, the leakage current between the control regions will be proportional to the net carrier concentration.

[0206] The type of dopant in the anti-doped region 610 can be selected based on different factors, such as the material forming the absorption region 506 or the type of dopant in the absorption region 506. For example, epitaxial germanium grown on a silicon substrate is generally a p-type material. In this case, an n-type dopant, such as phosphorus, arsenic, antimony, or fluorine can be used to dope the anti-doped region 610. Doping can be performed in different ways, including implantation, diffusion, and in-situ doping when growing the material. In some cases, dopants, such as fluorine, can protect defects. The protected defects will no longer serve as a source of charge carriers, and the net carrier concentration of the fluorine-doped absorption region 506 can be reduced and become closer to the intrinsic material.

[0207] The dopant concentration in the anti-doping region 610 can be selected based on the intrinsic carrier concentration of the absorption region 506. For example, the intrinsic carrier concentration is about 5×10 14 cm -3 An epitaxial germanium can have a thickness of about 5×10 14 cm -3 The anti-doping concentration can make the intrinsic carrier concentration of epitaxial germanium close to that of bulk crystalline germanium, which is about 2×10 13 cm -3 Generally speaking, the counter-doping concentration can be between 1×10 13 cm -3 Up to 1×10 16 cm -3In some embodiments, different regions of the counter-doped region 610 can have different dopant concentrations. For example, near a material interface (e.g., the bottom surface of the absorber 506) there may be a higher intrinsic carrier concentration (because of a higher number of defects), and a high counter-doping amount may provide better compensation. In some embodiments, the counter-dopant concentration may be higher than the intrinsic carrier concentration of the absorber region 506; in such a case, the polarity of the absorber region 506 may be changed from p-type to n-type, or from n-type to p-type.

[0208] Although the anti-doping region 610 is shown to completely cover the n-type doping regions 534 and 544, and the p-type doping regions 537 and 547, the anti-doping region 610 may also cover only the p-type doping regions 537 and 547, or the n-type doping regions 534 and 544. Second, although the anti-doping region 610 is shown as a continuous region, the anti-doping region 610 may also be two or more separate regions. In addition, although the anti-doping region 610 is shown as a portion of the absorption region 506, the anti-doping region 610 may also span the entire absorption region 506.

[0209] In some embodiments, the anti-doping region 610 can be used as a dopant diffusion suppressor and is used to provide a structure with a steep junction profile. The structure of the steep junction profile between the anti-doping region 610 and the p-type doping regions 537 and 547 can reduce leakage current, thereby reducing the energy consumption of the ToF pixel; for example, in the germanium absorption region 506, fluorine doping can suppress the diffusion of phosphorus dopants in the n-type doping region 534.

[0210] Generally, the counter-doped regions 610 may be implemented in different switching photodetectors to reduce leakage current between control regions.

[0211] In some embodiments, p-type doped regions 537 and 547 may be omitted, but this may result in the creation of a Schottky junction structure, the effect of which has been Figure 5F-5H Described in the relevant paragraphs.

[0212] Figure 6B is an example of a switch optical detector 620, wherein the switch optical detector 620 is used to convert an optical signal into an electrical signal. The switch optical detector 620 is similar to Fig. 6AThe switch photodetector 600 shown in FIG. 1 is different in that the first switch 532 and the second switch 542 further include n-type well regions 612 and 614, respectively. The additional n-type well regions can adjust the electrical and / or optical characteristics of the photodetector 620. In some embodiments, the doping amount of the n-type well regions 612 and 614 can range from 10 15 cm -3 Up to 10 17 cm -3 In some embodiments, the n-type well regions 612 and 614 can extend from the upper surface of the absorption region 506 to the lower surface of the anti-doped region 610 , or the interface between the absorption layer 506 and the substrate 502 .

[0213] The p-type doping region 537, the n-type well region 612, the anti-doping region 610, the n-type well region 614 and the p-type doping region 547 are arranged to form a PNINP junction structure. In general, the PNINP junction structure can selectively reduce a leakage current from the first control signal 122 to the second control signal 132, or a leakage current from the second control signal 132 to the first control signal 122.

[0214] In some embodiments, the p-type doped region 537 is formed entirely within the n-type well region 612. In some other embodiments, the p-type doped region is partially formed in the n-type well region 612; for example, a portion of the p-type doped region 537 can be formed in the n-type well region 612 by implanting a p-type dopant, and another portion of the p-type doped region 537 can be formed in the anti-doped region 610 by implanting a p-type dopant. Similarly, in some embodiments, the p-type doped region 547 is formed entirely within the n-type well region 614. In some other embodiments, the p-type doped region 614 is partially formed in the n-type well region 614. In some embodiments, the n-type well regions 612 and 614 form a continuous n-type well region that includes at least a portion of both the p-type doped regions 537 and 547.

[0215] The operating speed or bandwidth of a photodetector is an important performance parameter in applications and is beneficial for high-speed light detection, such as ToF detection. Of all the characteristics of a photodetector, the one that can significantly affect the bandwidth of a photodetector is the physical size of the photodetector, such as the area of ​​the photodetector that can receive light. For example, reducing the area of ​​a photodetector will increase the bandwidth of the photodetector, which will reduce the component capacitance, reduce the carrier transfer time, or a combination of the above. However, reducing the detection area of ​​a photodetector will also reduce the amount of light (i.e., the number of photons) measured by the photodetector; for example, for a predetermined light intensity per unit area, reducing the area of ​​the detector will reduce the light it can detect.

[0216] In applications that benefit from both high bandwidth and high detection efficiency, such as ToF detection, it is advantageous to add a microlens in front of the photodetector. The microlens can focus the incident light on the photodetector, allowing a small-area photodetector to detect incident light that is larger than its own area. For example, a design that combines the characteristics of a microlens and a spacer layer (SL) can make the distance between the microlens and the photodetector just an effective focal length of the microlens, allowing the incident light to be focused into a diffraction-limited spot that is on the order of the square of the wavelength of the incident light. This mechanism allows the photodetector area to be reduced while also reducing the potential disadvantages of reducing the photodetector area.

[0217] Fig. 7A A cross-sectional view of an example structure 700 for integrating a silicon lens into a photodetector is shown. The structure 700 includes a donor wafer 710 and a carrier wafer 730. The donor wafer 710 includes a plurality of pixels 720a-720c (collectively referred to as pixels 720), channels 714, metal pads 716, and a first bonding layer 712; the carrier wafer 730 includes a second bonding layer 732. The donor wafer 710 and the carrier wafer 730 are bonded to each other via the first bonding layer 712 and the second bonding layer 732. The substrate 710 may be similar to Figure 5A The substrate 502 shown, the absorption region 706 may be similar to Figures 5A-5K Absorption zone 506 is shown.

[0218] Pixels 720a-720c include absorption regions 706a-706c, and microlenses 722a-722c (collectively referred to as microlenses 722). Microlenses 722 are convex lenses and can be integrated in or on donor wafer 710. In applications that benefit from high light collection efficiency, such as ToF detection, additional microlenses 722 may be beneficial. The convex structure of microlenses 722 enables light incident on microlenses 722 to be focused on absorption regions 706, which can improve the light collection efficiency of pixels 720, thereby improving pixel performance. Pixels 720 having microlenses 722 configured on the back side of donor wafer 710 can be referred to as back-illuminated technology.

[0219] The characteristics of the microlens 722 affect its performance, including its geometric parameters and its constituent materials. The microlens 722 is generally implemented as a plano-convex structure; one surface of the microlens 722 facing the incident light is convex and has a radius of curvature, and the other surface bonded to the donor wafer 710 (the microlens 722 is bonded to the donor wafer 710 or integrated on the donor wafer 710) is a plane. The plano-convex structure of the microlens 722 can be manufactured using standard semiconductor process technology. The microlens 722 can have a height H L and a diameter D L, and may be separated from a lens-facing surface of the absorption region 706 by a height H O In some embodiments, H L The range can be from 1 to 4 μm, H O The range can be 8 to 12 μm, H A The range can be 1 to 1.5 μm, D L The range of H can be 5 to 15 μm. In some embodiments, for a spherical microlens 722, its radius of curvature can be set so that the focal length is approximately or equal to H O , so that the light can be optimally focused to the absorption region 706. The focal length and the radius of curvature can be determined by different simulation techniques, such as beam propagation method (BPM) and finite difference time domain (FDTD). In some embodiments, the microlens 722 is an aspherical lens.

[0220] The microlens 722 can be formed by different materials and processes. In general, a variety of materials that are transparent to the wavelength to be detected by the pixel 720 can be used to make the microlens 722. For example, the microlens 722 can be made of a medium to high refractive index material (e.g., a refractive index greater than 1.5), such as: crystalline silicon, polycrystalline silicon, amorphous silicon, silicon nitride, polymers or the like. At visible wavelengths, polymer materials are generally used to make microlenses. At NIR wavelengths, silicon is often used to make microlenses; this is because silicon is relatively transparent to NIR wavelengths and has a high refractive index (about 3.5 at a wavelength of 1000nm), so silicon is a suitable lens material for NIR wavelengths. In addition, silicon has high absorption for visible light (e.g., less than 800nm), and silicon microlenses can prevent a considerable amount of visible light from entering the absorption region 706, which is beneficial for applications of NIR wavelength detection (e.g., ToF detection). A crystalline silicon microlens 722 may be fabricated by patterning and etching the surface of a donor wafer 710, which is typically a crystalline silicon wafer. In other examples, polycrystalline silicon or amorphous silicon may be deposited on the surface of the donor wafer 710 and similarly fabricated by patterning and etching. Lenses formed by etching a crystalline silicon donor wafer 710, or etching polycrystalline silicon or amorphous silicon deposited on the donor wafer 710, are exemplary methods of integrating the microlens 722 on the donor wafer 710.

[0221] The patterning process of the microlens 722 can be performed, for example, using grayscale lithography. In grayscale lithography, the features to be patterned (e.g., microlenses) are exposed using local gradations of exposure dose, and the gradient thickness on the mask is pattern-transferred to facilitate development. For example, the mask can be first patterned to have a shape similar to the microlens 722, and then the microlens 722 is completed by transferring the patterned shape to the bottom of the underlying material (e.g., crystalline silicon donor wafer 710) using a semiconductor etching technique, such as a plasma-based directional etching technique. In some embodiments, the gradation mask of the local area of ​​exposure can be achieved, for example, by changing the fill-factor of the sub-wavelength features on the mask.

[0222] The absorption zone 706 may be similar to Figure 5A The absorber region 506 is shown, and the carrier wafer 730 may contain various electronic circuits coupled to the pixel 720. For example, the electronic circuits may be coupled via a structure such as a via 714. The via 714 may be coupled to a metal pad 716 and connected to external electronic devices via, for example, a conductive wire.

[0223] The carrier wafer 730 and the donor wafer 710 can be bonded or mechanically attached to each other by technology. For example, the first bonding layer 712 and the second bonding layer 732 can be oxide layers (e.g., silicon dioxide) and bonded to each other by oxide-oxide bonding technology. In other examples, the first bonding layer 712 and the second bonding layer 732 can be metals (e.g., copper) and bonded to each other by metal-metal bonding technology. In another example, the first bonding layer 712 and the second bonding layer 732 can be a combination of metal and oxide (e.g., silicon oxide and copper) and bonded to each other by hybrid bonding technology.

[0224] Figure 7B A cross-sectional view of an example structure 740 for integrating a silicon lens with a photodetector is shown. The structure 740 includes a microlens 742, an anti-reflection coating (ARC) layer 744, a spacer layer 746, a first layer 748, a second layer 750, a silicon layer 752, and a photodetector 754. The microlens 742 supports the ARC layer 744, and the spacer layer 746 supports the microlens layer 742. The silicon layer 752 can support the photodetector 754, or the photodetector 754 can be formed in the silicon layer 752. The first layer 748 and the second layer 750 can be intervening layers between the silicon layer 752 and the spacer layer 746.

[0225] The ARC layer 744 is used to reduce a reflectivity of light incident on the microlens 742. The ARC layer 744 may have a refractive index equivalent to the square root of the refractive index of the microlens 742, and a thickness equivalent to one quarter of the incident wavelength. In some embodiments, the ARC layer 744 may be formed of silicon dioxide. In some embodiments, the ARC layer 744 may be a multi-layer ARC formed of a multi-layer structure.

[0226] Structure 740 may be equivalent to an integrated lens in a back-illuminated (BSI) image sensor structure. For example, silicon layer 752 may be a silicon substrate, such as Fig. 7A The substrate 710 or Figure 5D The substrate 502 is shown; the light detector 754 may be, for example, Figure 5D The interface between the silicon layer 752 and the second layer 750 may correspond to Figure 5D The bottom surface of the substrate 502 opposite to the absorption region 506 is shown. In the BSI structure, the second layer 750 formed on the silicon layer 752 (e.g., the back side of the substrate 502) can include various structures and layers generally made on the BSI sensor wafer; the aforementioned various structures and layers include, for example, an ARC layer for reducing the reflection of light at the interface of the silicon layer 752, and a metal grid (metal grid), such as a tungsten grid, for blocking light that does not enter the portion of the silicon layer 752 for receiving light (e.g., the bottom of the microlens 742). The first layer 748 can be a thin layer to improve the adhesion between the spacer layer 746 and the second layer 750, and is used to increase the manufacturability and reliability of the structure 740. The material of the first layer 748 can be, for example, various dielectric materials (e.g., SiO 2 In some embodiments, the first layer 748 may be omitted depending on the interaction between the second layer 748 and the spacer layer 746 (eg, when the spacer layer 748 can achieve good bonding with the second layer 750).

[0227] Structure 740 can be made by providing a sensor wafer including a silicon layer 752, a photodetector 754, and a second layer 750, and sequentially depositing a first layer 748, a spacer layer 746, a microlens 742, and an ARC layer 744; then patterning and etching to expose a metal pad, which is similar to Fig. 7A The metal pad 716 is shown. The microlens 742 can be Fig. 7A The illustrated technique for forming microlenses 742 is patterned and etched. Although ARC layer 744 is shown here as being localized to the surface of microlenses 742, in general, ARC layer 744 may extend to other surfaces, such as the sides of microlenses 742 or the top surface of spacer layer 746.

[0228] Herein, various component characteristics of the structure 740 in a specific implementation are listed as examples for operation at a wavelength of 940 nm. The microlens 742 has a refractive index of 1.5316, a radius of curvature of 6 μm, a height of 4 μm, and a diameter D L The ARC layer 744 is made of SiO 2 740 is formed, and its refractive index at a wavelength of 940nm is 1.46, and has a thickness of 160.96nm. The spacer layer 746 has a refractive index of 1.5604 and a thickness of 10μm. The first layer 748 has a refractive index of 1.5507 and a thickness of 60nm. The second layer 750 has a tungsten grid and an ARC layer for the silicon layer 752. Although the above parameters have been provided, the characteristics of the structure 740 are still adjusted according to different operating wavelengths, materials, and sizes of the photodetector 754.

[0229] In some embodiments, the second layer 750 may be referred to as a "top layer" because the second layer 750 is formed on top of a back side of a silicon substrate of a BSI image sensor and may be tailored to improve the overall optical performance of the structure 740. As previously described, the second layer 750 includes a metal grid embedded in a dielectric layer, such as a SiO 2 When light enters the silicon layer 752 directly from air, SiO 2 The layer can be used as an ARC layer. However, since the refractive index of the additional microlenses 742, the spacer layer 746 and the first layer 748 is much greater than the refractive index of air (about 1.0), the SiO 2 Optical reflections at the interface of the silicon layer 752 and the first layer 748 and / or the spacer layer 746 of the stack may not be effectively reduced.

[0230] Table 1 shows simulation parameters and calculated penetrations corresponding to structure 740 , where layers and thicknesses are modified and / or approximated to allow structure 740 to achieve desired penetrations under different implementations.

[0231]

[0232] In Table 1, the second layer 750 corresponding to Case 1 includes a standard single layer SiO 2 The simulated transmittance is about 79%. For applications that require maximum light transmission, a loss of up to 21% of incident light is unacceptable. 2 Add a Si 3 N 4 layer, as SiO 2The problem of reduced transmittance can be alleviated by adding an intermediate layer between the Si layer 752 and the Si layer 753. 3 N 4 , the transmittance can be increased to about 97.6%. In this way, the interlayer can be classified as an ARC layer. Generally speaking, the refractive index is greater than that of SiO 2 Various optically transparent materials can be used to replace Si 3 N 4 ; For example, SiON, SiN, Al 2 O 3 , HfO 2 、ZrO 2 ,La 2 O 3 , and high-k materials (e.g., materials with a high dielectric constant). Suitable materials may have a refractive index higher than, for example, 1.6, 1.7, 1.8, 1.9, or 2.0. The thickness of the material should be adjusted to an odd multiple of a quarter of the wavelength of light entering the material.

[0233] Directly increase Si on the silicon layer 752 3 N 4 or high-k materials may result in an increase in the dark current of the photodetector 754, for example due to the increase in the dark current of the photodetector 754 compared to Si-SiO 2 Interface, Si-Si 3 N 4 In some embodiments, the second layer of SiO 2 Insert Si 3 N 4 The second SiO layer is inserted between the Si layer 752 to mitigate the increased dark current. 2 The thickness ranges from 10nm to 50nm, which will increase the corresponding transmittance to about 97.1% to 85%. 2 , which can reduce the increase of dark current while maintaining high optical transmittance.

[0234] As mentioned above, low leakage current flowing through the control region of a switched photodetector is an important performance parameter because low leakage current can reduce the power consumption of the device containing the photodetector. Another important performance parameter is the dark current flowing between a readout circuit and the control region of a switched photodetector. Dark current is a source of noise in the signal measured by the switched photodetector and will reduce the signal-to-noise ratio of a measured ToF signal.

[0235] Fig. 8Ais an example of a switch 800 for switching a light detector. The switch 800 can be used as a first or second switch in various switching light detectors. Figure 5A As described above, the switch 800 is formed in the absorption region 506 having the first layer 508. The switch 800 includes an n-type doped region 802, a readout contact 804, a low-doped n-type well region 806, a p-type doped region 812, a control contact 814, a low-doped p-type well region, and an n-type well region 818; the readout contact 804 is coupled to the n-type doped region 802, and the control contact 814 is coupled to the p-type doped region 812. The edges of the n-type doped region 802 and the p-type doped region 812 are separated by a distance S. The n-type doped region 802 and the p-type doped region 812 can be similar to Figure 5E The first n-type doped region 534 and the first p-type doped region 537 are shown. The readout contact 804 and the control contact 814 may be similar to Figure 5E The first readout contact 535 and the first control contact 538 in the embodiment of the present invention are shown in FIG. 8. The p-type doped region 812 may be a control region, and the n-type doped region 802 may be a readout region.

[0236] Sources of dark current formed in a lateral PIN diode include Shockley-Read-Hall (SRH) generation and band-to-band tunneling, wherein the lateral PIN diode includes a control region (p-type doped region 812), an absorption region 506 (undoped / intrinsic), and a readout region (n-type doped region 802). Surface defects present on the surface of the absorption region 506 may affect SRH generation. The additional first layer 508 reduces some of the surface defects, which can reduce the dark current generated by SRH. Increasing the distance S between the n-type doped region 802 and the p-type doped region 812 can also reduce the dark current, for example, because the electric field formed between the n-type doped region 802 and the p-type doped region is weakened, which also reduces the SRH generation rate between the two regions. For example, the distance S should be kept above 400 nm. However, increasing the distance S will reduce the bandwidth of the photodetector, for example, due to the increase in the carrier transfer time. Adding a low-doped n-type well region 806 , a low-doped p-type well region 816 , or a combination thereof can overcome the above trade-off problem.

[0237] The dopant concentrations of the low-doped regions 806 and 816 are lower than the dopant concentrations of the n-type doped region 802 and the p-type doped region 812, respectively. For example, the dopant concentrations of the low-doped regions 806 and 816 can be 1×10 17 cm -3 level, which is lower than that of the n-type doping region 802 and the p-type doping region 812 at 1×10 19 cm-3 Providing a low-doped region can reduce the problem of discontinuity of dopant concentration between the doping regions 802, 812 and the absorption region 506, and the low-doped region can have a dopant concentration of 1×10 15 cm -3 By providing a region with an intermediate dopant concentration, the electric field value at the edge of the doping regions 802 and 812 can be weakened. Weakening the electric field value can reduce the inter-band tunneling, thereby reducing the dark current between the two doping regions 802 and 812. In addition, the dark current generated by the SRH can be reduced. In general, the doping concentration of the low doping regions 806 and 816 can be adjusted according to different factors, such as the geometry of the switch, the doping concentration of the doping regions 802 and 812, and the doping concentration of the absorption region 506.

[0238] Figure 8B is an example of a switch 820 that switches the light detector on and off. The switch 820 is similar to Fig. 8A The switch 800 shown is different in that the low-doped regions 806 and 816 are replaced by a trench 822 formed in the absorption region 506, and the trench 822 is filled with a dielectric filler 824. The trench 822 filled with the dielectric filler 824 can reduce dark current.

[0239] The dielectric filler 824 is an electrically insulating material having a lower dielectric constant than the surrounding absorption region 506. The electric field is more able to penetrate regions with a low dielectric constant than regions with a high dielectric constant. By placing the filler trench 822 near the doped regions 802 and 812, some of the high electric field regions surrounding the doped regions 802, 812 and in the depletion region ("space charge region") around the doped regions 802, 812 are brought into the dielectric filler 824; thereby, SRH generation and / or band-to-band tunneling in the absorption region 506 is reduced. In addition, unlike the germanium absorption region 506, the dielectric filler 824 (e.g., SiO 2 ) is an insulator and does not generate SRH and / or inter-band tunneling. Therefore, dark current generated by SRH and / or inter-band tunneling (derived from high electric fields at the edges of doped regions 802 and 812) can be reduced.

[0240] The trench 822 may be formed by etching the absorption region by dry etching (e.g., plasma etching) or wet etching (e.g., liquid chemical bath). The trench 822 may be etched to a depth similar to the depth of the doped regions 802 and 812 (e.g., 10-200 nm). The trench 822 at least partially overlaps with the high electric field region around the n-type doped region 802 or the p-type doped region 812. In some embodiments, the trench 822 cuts into the doped regions 802 and 812 to remove a portion of the doped regions 802 and 812. After the trench 822 is formed, a first layer 508 may be deposited over the trench 822 to protect defects that appear on the surface of the trench 822. For a germanium absorption region 506, the first layer 508 may be, for example, amorphous silicon, polycrystalline silicon, germanium silicon, or a combination thereof. Thereafter, a dielectric filler 824 is filled in the trench 822, and the dielectric filler may be, for example, SiO 2 The dielectric filler 824 should be cleaned so that it does not have a significant concentration of impurities to avoid generating dark current.

[0241] In some embodiments, the depth of the trench can be deeper than the depth of the doped regions 802 and 812. For example, the depth of the doped regions 802 and 812 can be about 100 nm, and the depth of the trench can be up to 200 nm to reduce SRH generation and / or band-to-band tunneling. In some embodiments, up to 50% reduction in SRH generation and / or band-to-band tunneling around the doped regions 802 and 812 can be observed.

[0242] Figure 8C is an example of a switch 830 that switches the light detector on and off. The switch 830 is similar to Fig. 8A The switch 800 shown, but also includes Figure 8B Trench 822 and dielectric fill 824 are shown. Implementing low-doped regions 806 and 816, as well as trench 822, simultaneously can further reduce SRH generation and / or band-to-band tunneling compared to implementing low-doped regions 806, 816, or trench 822 alone.

[0243] In general, the dark current reduction achieved by using low doped regions 806 and 816, or trench 822, depends on the specific design of the switch and the overall design of the switch photodetector incorporating the switch. Figure 8C The switch shown includes both low-doped regions 806, 816 and trenches 822, but the key to implementing low-doped regions, trenches, or the like in combination still depends on the specific design of the switch in the switch photodetector. In addition, although only an example of a single trench is shown, in general, the trench 822 can be divided into two or more trenches.

[0244] although Figures 8A-8CIn the implementation of FIG. 5 , the first layer 508 and the n-type well region 818 are included, but in some embodiments, the first layer 508 and the n-type well region 818 can be omitted.

[0245] So far, description has been made with respect to different switch photodetectors and switches in the switch photodetectors. Now, description will be made with respect to different structures and elements of the switch photodetectors.

[0246] The switch photodetector is generally fabricated on a substrate, such as substrates 102, 202, 302, 402, and 502. The substrate is a carrier material on which the switch photodetector is fabricated. A semiconductor wafer is an embodiment of a substrate. The substrate can be part of the switch photodetector; however, in general, the substrate can simply provide a mechanical table for fabricating the switch photodetector. The substrate can be made of different materials, such as silicon, germanium, compound semiconductors (e.g., III-V, II-VI), silicon carbide, glass, and sapphire. The substrate can include different layers therein, such as a silicon on insulator (SOI) substrate including a base layer of silicon, an insulator layer (e.g., SiO2) on the base layer of silicon. 2 ), and a silicon device layer on an insulator layer. SOI may additionally include a device layer-insulator layer pair. For example, a dual-SOI wafer includes two device layer-insulator layer pairs.

[0247] The switch photodetector includes an absorption layer for absorbing incident light and converting the incident light into charge carriers. The absorption layers 106, 206, 306, 406, and the absorption regions 506, 706 are examples of absorption regions. The absorption region can be formed of various absorption materials that can absorb light at the operating wavelength of the switch photodetector. The material of the absorption region can include, for example, silicon, germanium, IV-IV semiconductor alloys (e.g., GeSn, GeSi), III-V compound semiconductors (e.g., GaAs, InGaAs, InP, InAlAs, InGaAlAs), and other materials in the third, fourth, and fifth groups of the periodic table. In some embodiments, the absorption region can be a region in a substrate; for example, a region in a silicon substrate can be used to absorb visible light.

[0248] In some embodiments, an absorption region can be defined in a light absorbing material by changing the material composition (e.g., different germanium silicon compositions), doping a region of the absorbing material (e.g., a counter-doped region), or forming an optical window for light to pass through (e.g., a tungsten grid opening in a BSI image sensor).

[0249] The absorber material may be deposited on a substrate. For example, the absorber material may be blanket-deposited on the substrate. In some embodiments, the absorber material may be deposited on an interposer formed on the substrate. In general, the interposer may be selected based on the absorber material, the substrate, or both. Such an interposer may improve the manufacturability of the device and / or improve the performance of the device. The material of the interposer may, for example, include silicon, graded germanium-silicon component materials, graded III-V materials, germanium, GaN, and SiC. A graded material is a material that changes its material composition along at least one direction. For example, in a graded germanium-silicon material, its germanium composition changes from 1% at one end to 99% at the other end. In general, the composition of the start and end can be set, for example, depending on the composition of the substrate and the composition of the absorber material.

[0250] In some embodiments, the material of the absorber layer can be epitaxially grown on the interposer in one or two steps. For example, the absorber layer (e.g., germanium, germanium silicon) can be deposited on a dielectric layer having an opening, and the dielectric layer opens to the substrate below it (e.g., a crystalline silicon substrate). When the absorber material is deposited on a substrate with mismatched lattice constants, the multi-step growth process can improve the material quality (e.g., reduce the number of material defects). The above technology is disclosed in U.S. Patent No. 9,786,715, entitled "High Efficiency Wide Spectrum Sensor", and is incorporated herein by reference.

[0251] Figures 9A-9E Shows an example of the electrical terminals used in a switch light detector. Fig. 9A , an electrical terminal 900 includes a region 902, a contact metal 904, and a doped region 906. Region 902 is a metal on which the electrical terminal 900 is formed, and may correspond to an absorption region (e.g., absorption region 506) or a substrate (e.g., substrate 502). Doped region 906 may be a p-type (acceptor) doped region or an n-type (donor) doped region depending on the type of dopant. Doped region 906 is typically doped to a high doping concentration (e.g., 1×10 19 Up to 5×10 20 cm -3 ) to allow ohmic contact to be formed between contact metal 904 and region 902. This amount of doping concentration may be referred to as "degenerate doping."

[0252] The contact metal 904 is a metal material and contacts the region 902 through the doped region 906. The contact metal can be selected from different metals or alloys according to the material of the region 902 and the dopant of the doped region 906; for example, it includes Al, Cu, W, Ti, Ta-TaN type Cu stack, Ti-TiN type W stack, and various metal silicides.

[0253] See also Fig. 9B ; Electrical terminal 910 is similar to Fig. 9A The electrical terminal 900 is shown, but the difference is that the doped region 906 is omitted. The contact metal 904 is directly disposed on the region 906 without passing through the doped region 906, which can form a Schottky contact, an ohmic contact, or an intermediate characteristic between the above two, depending on various factors, including the material of the region 902, the contact metal 904, and the amount of impurities or defects in the region 902.

[0254] See also Fig. 9C ; Electrical terminal 920 is similar to Fig. 9B The electrical terminal 910 is shown in FIG. 1 , but the difference is that a dielectric layer 922 is inserted between the contact metal 904 and the region 902. For example, for the crystalline germanium region 902, the dielectric layer 922 can be amorphous silicon, polycrystalline silicon, or silicon germanium. In other examples, for the crystalline silicon region, the dielectric layer 922 can be amorphous silicon, polycrystalline silicon, or silicon germanium. The insertion of the dielectric layer 922 can form a Schottky contact, an ohmic contact, or an intermediate characteristic between the above two combinations.

[0255] See also Fig.9D ; Electrical terminal 930 is similar to Fig. 9B The electrical terminal 910 is shown with the difference that an insulating layer 932 is inserted between the contact metal 904 and the region 902. The insulating layer 932 prevents direct current conduction from the contact metal 904 to the region 902, but after a voltage is applied to the contact metal 904, an electric field can be formed in the region 902. The formed electric field can attract or repel charge carriers into the region 902. The insulating layer 932 can be SiO 2 、Si 3 N 4 , or high-k value materials.

[0256] A switch to turn on or off the photodetector, e.g. Figure 5D The first switch 532 is shown to include a carrier control terminal and a carrier collection (readout) terminal. The carrier control terminal is a terminal that directs the photogenerated carriers in the region 902 in a certain direction, for example, toward the carrier collection terminal, by applying a control voltage, for example, by an external bias circuit. The operation of the carrier control terminal has been described in detail. Figure 1AThe description is made in the paragraphs related to control signals 122 and 132 in FIG. Different types of electrical terminals can be used to implement the carrier control terminal; for example, electrical terminals 900, 910, 920 and 930 can be used to implement the carrier control terminal.

[0257] The carrier collection terminal is a terminal for collecting photocarriers in region 902. The carrier collection terminal can be used to collect electrons (e.g., n-type doped regions) or holes (e.g., p-type doped regions). Figure 1A The description is made in the paragraphs related to the readout circuits 124 and 134 in FIG. Different types of electrical terminals can be used to implement the carrier collection terminals; for example, electrical terminals 900, 910, 920, and 930 can be used to implement the carrier collection terminals.

[0258] The number of carrier control terminals and carrier collection terminals can be adjusted, for example, according to the performance of the target device. For example, the switch photodetector can have the following exemplary structures: two carrier control terminals with two carrier collection terminals, two carrier control terminals with one carrier collection terminal, four carrier control terminals with two carrier collection terminals, and four carrier control terminals with four carrier collection terminals. Generally speaking, the number of carrier control terminals and carrier collection terminals of the switch photodetector is greater than one.

[0259] When the switch photodetector includes two or more control terminals, various combinations of the aforementioned electrical terminals may be used, for example, combining ohmic and Schottky / ohmic terminals (e.g., terminals 900 and 920), ohmic and insulating (e.g., terminals 900 and 930), insulating and Schottky / ohmic (e.g., 930 and 920), and ohmic and Schottky / ohmic and insulating terminals (e.g., terminals 900, 920, 930).

[0260] When the switching photodetector includes two or more carrier collection terminals, it can be implemented using Ohmic and Schottky / Ohmic terminals (e.g., terminals 900 and 920).

[0261] Electrical terminals may have different shapes depending on various considerations, such as manufacturability and device performance. Fig.9E A top view of examples of electrical terminals of different shapes is shown. The shape of the terminal 940 can be rectangular, triangular, circular, polygonal or a combination thereof. The corners of the terminal can be sharp or rounded. The shape can depend on doped regions, metal silicides, contact metals or a combination thereof.

[0262] The absorber and substrate can be configured in different structures, and the absorber can have different shapes based on different considerations, such as manufacturability and device performance. Figures 10A-10I, etc. show various structures of absorption regions and substrates. Fig. 10A In the embodiment, the structure 1000 includes a substrate 1002 and an absorption region 1004 protruding from the upper surface of the substrate 1002; the substrate 1002 may be similar to Figure 5D The substrate 502 shown, and the absorption region 1004 can be similar to Figure 5D Absorbing region 506 is shown. Structure 1000 may be fabricated by depositing an absorbing region 1004 on substrate 1002 and etching the absorbing region 1004 to form a protruding structure.

[0263] See also Fig. 10B ; Structure 1010 is similar to Fig. 10A The structure 1000 is shown, but further includes an interposer 1006 between the absorber layer 1004 and the substrate 1002. The interposer can act as a buffer layer to facilitate the growth of the absorber layer 1004 on the substrate 1002. The structure 1010 can be made by depositing the interposer 1006 on the substrate 1002, depositing the absorber layer 1004 on the interposer 1006, and then etching the absorber layer 1004 and the interposer 1006 to form a protruding structure.

[0264] See also Fig. 10C ; Structure 1020 is similar to Fig. 10A The structure 1000 is shown, but with the absorber layer 1004 partially embedded in the substrate 1002. The structure 1020 can be made by forming a recess in the substrate 1002 and selectively depositing the absorber layer 1004 in the recess. Optionally, the structure 1020 can be made by depositing a sacrificial layer over the substrate 1002, etching the sacrificial layer to form the recess in the substrate 1002, selectively depositing the absorber material, and removing the absorber material deposited outside the recess by a planarizing step, such as chemical-mechanical polishing (CMP), and finally removing the sacrificial layer by a selective etch, such as a wet chemical etch.

[0265] See 10D; structure 1030 is similar to Fig. 10C Structure 1020 is shown, but with absorber layer 1004 fully embedded in substrate 1002. Structure 1030 may be fabricated by forming a recess in substrate 1002, depositing a selective layer of absorber material over substrate 1002, and removing absorber material deposited outside the recess with a planarization step (e.g., a CMP step).

[0266] See also Fig.10E ; Structure 1040 is similar to Fig. 10D Structure 1030 is shown, but with an interposer 1006 interposed between absorber layer 1004 and substrate 1002 in recess 1004. Structure 1040 may be fabricated by forming a recess in substrate 1002, depositing a conformal layer of interposer 1006, depositing a blanket layer of absorber material over interposer 1006, and removing the absorber material and interposer deposited outside the recess with a planarization step (e.g., a CMP step).

[0267] See also Fig.10F ; Structure 1050 is similar to Fig.10E The structure 1040 is shown, but with a second interposer 1008 replacing the first interposer 1006 between a sidewall of the absorbing region 1004 and the sidewall of the recess in the substrate 1002. The structure 1050 can be fabricated by forming a recess in the substrate 1002, depositing a conformal layer of the second interposer 1008 and performing an anisotropic blanket etch to remove the second interposer 1008 along the vertical surfaces, depositing a conformal layer of the first interposer 1006, performing an anisotropic blanket etch to remove the first interposer 1006 along the non-vertical surfaces, depositing a selective layer of absorbing material, and removing the absorbing material and the first interposer deposited outside the recess in a planarization step (e.g., a CMP step). In an exemplary embodiment, the first interposer 1006 can be made of SiO 2 The second interposer 1008 may be formed of silicon germanium.

[0268] See also Figure 10G ; Structure 1060 is similar to Fig. 10A The structure 1040 is shown, but includes a tiered interposer 1062 with an absorbing region, and the absorbing region 1004 is embedded in the interposer 1062. The tiered interposer 1062 includes an opening 1064 to the substrate 1002, and a recess 1066 in which the absorbing region 1004 is embedded. The absorbing region 1004 contacts the substrate 1002 through the opening 1064. The structure 1060 can be fabricated by forming a deposited interposer on the substrate 1002, etching the opening 1064 through the entire thickness of the interposer, etching the recess 1066 in the interposer, depositing the absorbing region 1004 on the tiered interposer 1062, and removing the absorbing material deposited outside the recess by a planarization step (e.g., a CMP step).

[0269] See also Fig. 10H ; Structure 1070 is similar to Figure 10GThe structure 1060 is depicted, but includes a second interposer 1072 in which the recess 1066 is formed. The structure 1070 can be fabricated by depositing the first interposer 1062 on the substrate 1002, depositing the second interposer 1072, etching the first interposer 1062 and the second interposer 1072 to form the opening 1064, etching the recess 1066 in the second interposer 1072, depositing the absorber layer 1004, and removing the absorber material deposited outside the recess with a planarization step (e.g., a CMP step).

[0270] See also Fig.10I ; Structure 1080 is similar to Fig.10E The structure 1040 is depicted, but includes an opening 1084 formed in the interposer 1006. The absorbing region 1004 contacts the substrate 1002 through the opening 1084. The structure 1080 can be fabricated by forming a recess in the substrate 1002, depositing a conformal layer of the interposer 1006, etching the opening 1084, depositing a blanket layer of absorbing material over the interposer 1006, and removing the absorbing material and the interposer deposited outside of the recess with a planarization step (e.g., a CMP step).

[0271] The absorption region, the carrier control terminal and the carrier collection terminal may be arranged in different structures according to different considerations, such as manufacturability and device performance. Figures 11A-11B A top view and a side view of an exemplary switching photodetector 1100 are shown, wherein the carrier control terminal and the carrier collection terminal are disposed on a substrate, and a portion of the substrate is an absorption region. In this example, the switching photodetector 1100 includes a substrate 1102, an absorption region 1104, a plurality of carrier collection terminals 1106, and a plurality of carrier control terminals 1108, wherein the absorption region 1104 is a region in the substrate 1102. For example, for a silicon substrate 1102, the absorption region 1104 is formed in silicon, and the absorption region 1104 is used to absorb visible light. The absorption region can have different shapes; from a top view of the photodetector, the absorption region can be, for example, a square. The absorption region 1104 can extend from an upper surface of the substrate 1102 to a desired depth below the aforementioned upper surface. For example, the absorption region 1104 can extend below the upper surface of the substrate 1102 by 1 μm, 2 μm, 3 μm, 5 μm, or 10 μm. A pair of adjacent carrier collection terminals 1106 and carrier control terminals 1108 form a switch. The absorption region 1104 is disposed between the pair of adjacent carrier collection terminals 1106 and carrier control terminals 1108. In some embodiments, the pair of adjacent carrier collection terminals and carrier control terminals are symmetrically disposed around the absorption region 1104 (e.g., on opposite sides or four sides of the absorption region 1104). Such a symmetrical configuration can improve the matching of carrier control and collection performance of the two switches in a pair.

[0272] Figures 11C-11F Shows top and side views of a switched photodetector where the absorbing region is formed of a different material than the substrate. Figures 11C-11D The switching photodetector 1120 includes a substrate 1102 , an absorption region 1124 , a carrier collection terminal 1106 and a carrier control terminal 1108 . Fig. 11C A top view of the switch light detector 1120 is shown, and Fig.11D A side view of the switch light detector 1120 is shown. The switch light detector 1120 is similar to Figures 11A-11B 1100, but the difference is that the absorption region 1124 of the switching light detector 1120 is made of a material different from the substrate 1102. For example, the absorption region 1124 can be formed of germanium and the substrate 1102 can be a silicon substrate. The absorption region 1124 is completely embedded in a trench formed in the substrate 1102. Although the details of the embedded structure are not shown, the embedded absorption region 1124 can be, for example, Figures 10D-10F and Figure 5C This is achieved through a related narrative structure.

[0273] See also Fig.11E , the switch light detector 1130 is similar to Figures 11C-11D The switch photodetector 1120 is shown, but the absorption region 1124 is partially embedded in the substrate 1102. Although the details of the embedded structure are not shown, the partially embedded absorption region 1124 can be used, for example Fig. 10C and Figure 5B This is achieved through a related narrative structure.

[0274] See also Fig.11F , the switch light detector 1140 is similar to Figures 11C-11D The switch photodetector 1120 is shown, but the absorption region 1124 completely protrudes above the substrate 1102. Although the details of the fully protruding structure are not shown, the fully protruding absorption region 1124 can be used, for example Figures 10A-10B and Figure 5A This is achieved through a related narrative structure.

[0275] In some switch photodetector structures, the carrier collection terminal, the carrier control terminal, or both can be disposed in the absorption region. For the sake of brevity, the detailed implementation details of the substrate, the absorption region, the carrier control terminal, and the carrier collection terminal are not repeated here. Figures 12A-12BA top view and a side view of an example of a switch photodetector 1200 are shown, in which a carrier collection terminal is disposed on a substrate and a carrier control terminal is disposed on an absorption region. The switch photodetector 1200 includes a substrate 1202, an absorption region 1204, a light receiving region 1205, a plurality of carrier collection terminals 1206, and a plurality of carrier control terminals 1208. The light receiving region 1205 may indicate that input light is incident on a portion of the absorption region 1204 and may not be physically distinguished from the remaining portion of the absorption region 1204. For example, a combination of a light shield (e.g., a tungsten grid) and a microlens may block and focus the incident light on the light receiving region 1205. The carrier collection terminal 1206 is disposed on the substrate 1202, and the carrier control terminal 1208 is disposed at a position of the absorption region 1204 that does not overlap with the light receiving region 1205. For the switch photodetector 1200, the absorption region 1204 protrudes completely. For the switching light detector 1220, the absorption region 1204 can be as follows Fig. 12C The parts shown are embedded (in the substrate), or can be as Fig.12D Shown fully embedded (in substrate).

[0276] Figures 12E-12F The top view and side view of an example of a switch photodetector 1240 are shown, wherein the carrier collection terminal and the carrier control terminal are disposed on the absorption region. The switch photodetector 1240 is similar to Figures 12A-12B The switch photodetector 1200 is shown, but the difference is that the carrier collection region 1206 is now arranged in the absorption region 1204, rather than overlapping with the light receiving region 1205. For the switch photodetector 1240, the absorption region 1204 is completely protruding. Figure 12G For the switching light detector 1250, the absorption region 1204 can be partially embedded (in the substrate); Fig.12H For the switching photodetector 1260, the absorption region 1204 can be completely embedded (in the substrate).

[0277] Only Figures 12A-12H The light receiving region 1205 is shown as not overlapping with the carrier collecting terminal or the carrier control terminal, but in fact the light receiving region 1205 may overlap with at least a portion of the carrier control region, at least a portion of the carrier receiving region, and at least a portion of the n-type doping region or the p-type doping region. For example, such an overlap may occur in a pixel that applies the FSI and BSI structures.

[0278] In some embodiments of the switch photodetector, each switch may include more than one carrier collection terminal, more than one carrier control terminal, or more than one of the two. For the sake of brevity, the detailed implementation details of the substrate, the absorption region, the carrier control terminal, and the carrier collection terminal are not repeated here.Figures 13A-13G A top view of an example of a switched photodetector having a switch, wherein the switch comprises a plurality of carrier control terminals or a plurality of carrier collection terminals. Fig.13A In the embodiment, the switching photodetector 1300 includes a substrate 1302, an absorption region 1304, a light receiving region 1305, a plurality of substrate carrier collection terminals 1306, a plurality of substrate carrier control terminals 1308 and a plurality of absorber carrier control terminals 1309. The substrate carrier collection terminal 1306 is a carrier control terminal disposed on a substrate (e.g., substrate 1302). The substrate carrier control terminal 1308 is a carrier control terminal disposed on a substrate (e.g., substrate 1302). The absorber carrier control terminal 1309 is a carrier control terminal disposed in an absorption region (e.g., absorption region 1304). The effects and implementation details of the absorber carrier control terminal 1309 and the substrate carrier control terminal 1308 have been described in detail. Figure 5K In some embodiments, the substrate carrier collection terminal 1306, the substrate carrier control terminal 1308 and the absorber carrier control terminal 1309 will appear repeatedly in Fig. 13B in the second column.

[0279] exist Fig. 13B , the switch light detector 1310 is similar to Fig.13A The switch photodetector 1300 shown is different in that the substrate carrier control terminal 1308 is omitted and a pair of terminals 1306 and 1309 are added in the second row. The second pair of control and collection terminals adjacent to the first pair of control and collection terminals can operate independently of the first pair of control and collection terminals; of course, the second pair of control and collection terminals can also operate in conjunction with the first pair of control and collection terminals.

[0280] exist Fig. 13C , the switch light detector 1320 is similar to Fig. 13B The switch photodetector 1310 is shown with the difference that a substrate carrier collection terminal 1306 on the side of the light receiving region 1305 is removed. The combination of a pair of absorber carrier control terminals 1309 and corresponding substrate carrier collection terminals 1306 on the side of the light receiving region 1305 can act as a switch.

[0281] exist Fig.13D , the switch light detector 1330 is similar to Fig. 13B The switching photodetector 1310 is shown, with the difference that the substrate carrier collection terminal 1306 is moved to the absorption region 1304 to serve as the absorber carrier collection terminal 1307.

[0282] exist Fig.13E , the switch light detector 1340 is similar to Fig.13DThe switch photodetector 1330 is shown with the difference that an absorber carrier collection terminal 1307 on one side of the light receiving region 1305 is removed. The combination of a pair of absorber carrier control terminals 1309 and corresponding absorber carrier collection terminals 1307 on one side of the light receiving region 1305 can act as a switch.

[0283] exist Fig.13F , the switch light detector 1350 is similar to Fig.13D The switch photodetector 1330 is shown with the difference that an absorber carrier control terminal 1309 on the side of the light receiving region 1305 is removed. The combination of a pair of absorber carrier collection terminals 1307 and corresponding absorber carrier control terminals 1309 on the side of the light receiving region 1305 can act as a switch.

[0284] exist Figure 13G , the switch light detector 1360 is similar to Fig.13D The switch photodetector 1330 shown has the difference that four pairs of absorption carrier collection and control terminals 1307 and 1309 are symmetrically arranged around the light receiving area 1305. Any pair of terminals 1307 and 1309 can be used as a switch. Each switch can be operated alone or in cooperation with other switches. For example, the east and west switches can be controlled to act as the first switch, and the south and north switches can be controlled to act as the second switch. In other examples, the east and south switches can be controlled to act as the first switch, and the west and north switches can be controlled to act as the second switch.

[0285] Only Figures 13A-13G The light receiving region 1305 is shown as not overlapping with the carrier collecting terminal or the carrier control terminal, but generally speaking, the light receiving region 1305 may overlap with at least a portion of the carrier control region, at least a portion of the carrier receiving region, and at least a portion of the n-type doping region or the p-type doping region. For example, such an overlap may occur in a pixel using FSI and BSI structures.

[0286] For switches having two or more carrier control terminals, an independent control bias may be applied to each carrier control terminal, or a single bias may be applied to the carrier control terminals that are short-circuited to each other. Figures 14A-14B A top view of an example of a switched photodetector having a switch including a plurality of carrier control terminals is shown. Fig.14A , the switch light detector 1400 is similar to Fig.13AThe switch photodetector 1300 shown in FIG. The substrate carrier collection terminal 1306, the substrate carrier control terminal 1308 and the absorber carrier control terminal 1309 on the left side of the light receiving region 1305 form a first switch, and the substrate carrier collection terminal 1306, the substrate carrier control terminal 1308 and the absorber carrier control terminal 1309 on the right side of the light receiving region 1305 form a second switch 1420.

[0287] In switches 1410 and 1420, substrate carrier control terminal 1308 and absorber carrier control terminal 1309 may be short-circuited together and applied with a single bias, or may be individually applied with independent control biases. For example, a voltage V is applied to substrate carrier control terminal 1308 of first switch 1410. B1 , and apply a voltage V to the absorber carrier control terminal 1309 A1 Similarly, a voltage V is applied to the substrate carrier control terminal 1308 of the second switch 1420 B2 , and apply a voltage V to the absorber carrier control terminal 1309 A2 In some embodiments, control terminals near the light receiving region, such as absorber carrier control terminal 1309, can be individually applied with a control voltage V A1 and V A2 To guide the photogenerated carriers in the light receiving region 1305, so that the photogenerated carriers are directed to the voltage V shown in FIG. C1 and V C2 The substrate carrier collection terminal 1306 moves. At the same time, the substrate control terminal 1308 can be applied with a voltage V B1 and V B2 , thereby establishing a high electric field between the substrate carrier control terminal 1308 and the substrate carrier collection terminal 1306. When the electric field between the terminals 1308 and 1306 is high enough, a region of avalanche multiplication can be established between the terminals 1308 and 1306, thereby providing an avalanche gain to the photogenerated carriers guided to the substrate carrier collection terminal 1306 by the absorber carrier control terminal 1309. As a result, the photogenerated carriers can be multiplied by the avalanche gain, which increases the photocurrent signal generated by the switch photodetector 1400.

[0288] exist Fig. 14B , the switch light detector 1430 is similar to Fig.14AThe switching photodetector 1400 is shown, but the difference is that the substrate carrier collection terminal 1306 is relocated on the absorption region 1304 and used as the absorber carrier collection terminal 1407, and the substrate carrier control terminal 1308 is relocated on the absorption region 1304 and used as the absorber carrier control region 1409. The effect of different bias voltages on the terminals is similar to Fig.14A and its related paragraphs.

[0289] Only in Figures 14A-14B The light receiving area 1305 shown does not overlap with the carrier collection terminal or the carrier control terminal, but in fact the light receiving area 1305 can overlap with at least a portion of the carrier control region, at least a portion of the carrier receiving region, and at least a portion of the n-type doping region or the p-type doping region; for example, such overlap may occur in a pixel applying FSI and BSI structures.

[0290] In a typical implementation of an image sensor, a plurality of sensor pixels (e.g., photodetectors) may be arranged in an array to allow the image sensor to capture an image with a plurality of image pixels. In order to provide a high integration density, a plurality of sensor pixels disposed on a common substrate are arranged as close as possible. For a semiconductor substrate, such as a p-type doped silicon substrate, adjacent sensor pixels may cause electronic and / or optical crosstalk between the sensor pixels, which may, for example, reduce a signal-to-noise ratio of the sensor pixels. In such a case, the electrical isolation between the sensor pixels may be improved by introducing various isolation structures.

[0291] Figures 15A-15G is a side view of an example structure for sensor pixel isolation. Fig.15A In the embodiment of the present invention, an exemplary structure 1500 includes a substrate 1502, a plurality of sensor pixels 1510a and 1510b (collectively referred to as sensor pixels 1510), and an isolation structure 1506. The sensor pixels 1510a and 1510b include corresponding absorption regions 1504a and 1504b. Each imager pixel 1510 can be a switching photodetector, such as Figures 5A-5K A switching photodetector is shown. For clarity of illustration, sensor pixel 1510 is omitted.

[0292] Isolation structure 1506 may include electrical isolation between sensor pixels 1510a and 1510b. In structure 1500, the isolation structure extends from an upper surface of substrate 1502 to a predetermined depth within substrate 1502. In some embodiments, isolation structure 1506 is a doped region doped with p-type dopants or n-type dopants. The doping of isolation structure 1506 may establish a bandgap offset-induced potential energy barrier to hinder current from traversing isolation structure 1506 and improve electrical isolation between pixels 1510a and 1510b. In some embodiments, isolation structure 1506 is a trench filled with a semiconductor material, and the semiconductor material filled in the trench of isolation structure 1506 is different from the semiconductor material of substrate 1502. A potential barrier caused by bandgap offset may be established at the interface between the two different semiconductors of the substrate 1502 and the isolation structure 1506 to hinder the current from flowing through the isolation structure 1506 and improve the electrical isolation between the pixels 1510a and 1510b.

[0293] In some embodiments, isolation structure 1506 is a trench filled with a dielectric or an insulator. Isolation structure 1506 filled with a low conductivity dielectric or insulator can provide a region with high resistance between sensor pixels 1510a and 1510b, thereby hindering current flow through isolation structure 1506 and improving electrical isolation between pixels 1510a and 1510b.

[0294] Although only a single isolation structure 1506 is shown, in practice, multiple isolation structures 1506 may be provided in each adjacent pair of image sensing pairs 1510. For example, in a two-dimensional array of sensor pixels 1510, a single sensor pixel 1510 may be surrounded by four nearest sensor pixels 1510. In such a case, the isolation structure 1506 may be provided along the four nearest interfaces. In some embodiments, the isolation structure 1506 may be a continuous structure surrounding the sensor pixel 1510. The isolation structure 1506 may be shared at the interface between the pixels 1510.

[0295] Fig. 15B An example of a structure 1520 is shown, where the structure 1520 is similar to Fig.15A Structure 1500 is shown, but with the difference that both absorption regions 1504a and 1504b are completely embedded in substrate 1502.

[0296] Fig. 15C An example of a structure 1530 is shown, where the structure 1530 is similar to Fig.15AStructure 1500 is shown, but differs in that isolation structure 1506 extends from the upper surface of substrate 1502, through the entire depth of substrate 1520, to the lower surface of substrate 1502. Structure 1530 can remove the alternate conductive path between image sensor 1510 that diverts isolation structure 1506 and improve electrical isolation between sensor pixels 1510.

[0297] Fig.15D An example of a structure 1540 is shown, where the structure 1540 is similar to Fig. 15C Structure 1530 is shown, but with the difference that both absorption regions 1504a and 1504b are completely embedded in substrate 1502.

[0298] Fig.15E An example of a structure 1550 is shown. The structure 1550 includes a substrate 1502, sensor pixels 1510a, 1510b (collectively referred to as sensor pixels 1510), and isolation structures 1556a and 1556b (collectively referred to as isolation structures 1556). The isolation structures 1556a and 1556b are similar to Fig.15A The isolation structure 1556 is similar to the isolation structure 1506 described in the related paragraphs, but the difference is that the isolation structure 1556 is disposed on a portion of the substrate 1502 and just below the corresponding absorption layer 1504. The isolation structure 1556 disposed between the absorption region 1504 and the substrate 1502 can help confine the photocarriers to the absorption region 1504 and help reduce the photogenerated carriers leaking from the substrate 1502. For example, the sensor pixels 1510a and 1510b can be similar to Figure 5D The embodiment of the present invention is realized by means of a switching photodetector 530 in FIG. 1504 , which has all electrical terminals disposed on the absorption region 1504. In such a case, the electrical isolation provided by the isolation structure 1556 (e.g., a thin p-type doped layer) can improve the photocurrent collection efficiency and / or the bandwidth of the sensor pixel 1510.

[0299] Fig.15F is an example of structure 1560. Structure 1560 is similar to Fig.15E 1504a and 1504b are completely embedded in the substrate 1502, and the isolation structure 1556 partially or completely surrounds the absorption region 1504. For the isolation structure 1556 formed of an insulator or dielectric, the isolation structure 1556 may include an opening that is located below the absorber and partially surrounds the embedded absorption region 1504. For the isolation structure 1556 that is a doped region, the isolation structure 1556 may be a continuous structure that completely surrounds the embedded absorption region 1504 without an opening.

[0300] Although the isolation structure may be a doped region, a dielectric material, or an insulator as described above, in practice, the isolation structure may be a combination of these implementations. Figure 15G An example of structure 1570 is shown. Structure 1570 is similar to Fig.15A 1500, but the difference is that the isolation structure 1506 includes a first isolation structure 1576 and a second isolation structure 1577. The first isolation structure isolation structure 1576 can be a trench filled with a semiconductor material, or a trench filled with a dielectric or insulator, wherein the semiconductor material filled in the trench of the first isolation structure 1576 is different from the semiconductor material of the substrate 1502. The second isolation structure 1577 can be a doped region doped with a p-type dopant or an n-type dopant. Compared with the isolation structure implemented with only different materials or doped regions, the isolation structure 1504 implemented with different materials and doped regions can further improve the electrical isolation between the sensor pixels 1510. In some embodiments, a doped isolation can be used to form the second isolation structure 1577, and a material isolation through trench filling can be used to form the first isolation structure 1576, wherein the doped isolation is shallower than the material isolation.

[0301] The light detection efficiency of a photodetector (e.g., a switching photodetector) can be improved by adding various structures that adjust the characteristics of the photodetector. For example, reflectors, dielectric layers, and ARC layers can be added individually or simultaneously to achieve different effects, including increasing the light absorption rate through the absorption region, establishing an optical resonant cavity, and / or changing the spectral response of the photodetector. Figures 16A-16J is a cross-sectional view of an example structure of a photodetector. Fig.16A An example of a structure 1600 is shown. The structure 1600 includes a substrate 1602, an absorptive region 1604, and a metal reflector 1606, where the absorptive region 1604 forms a photodetector and the metal reflector 1606 reflects incident light.

[0302] As shown in the figure, the optical signal 1605 is incident from the top of the absorption region 1604, which can be regarded as an FSI structure. In such a structure, the optical signal 1605 may not be completely absorbed by the absorption region 1604, and a portion of the light may penetrate the absorption region 1604. The light that penetrates the absorption region 1604 but is not absorbed by the absorption region 1604 may reduce the light detection efficiency of the photodetector. By arranging a metal reflector 1606 on the lower surface of the substrate 1602, the optical signal 1605 that penetrates the absorption region 1604 is reflected, so that the optical signal 1605 that penetrates the absorption region 1604 can be reflected back to the absorption region 1604 and penetrate the absorption region 1604 for the second time, thereby improving the detection efficiency.

[0303] The portion of the optical signal 1605 absorbed by the absorption region 1604 may be a function of the optical absorption coefficient of the absorption region 1604 , the thickness of the optical absorption region 1604 along the light incident direction (along the vertical direction), and the wavelength of the optical signal 1605 .

[0304] The metal reflector 1606 may be formed of various optically reflective metals, such as copper, aluminum, gold, and platinum. In the light detector of structure 1600, the metal reflector 1606 may have a reflectivity greater than 50%, 60%, 70%, 80%, 90%, or 95%. The thickness of the metal reflector 1606 may be greater than a skin-depth of the metal. For example, the metal reflector 1606 may have a thickness ranging from 50 nm to 500 nm.

[0305] Fig. 16B An example of structure 1610 is shown. Structure 1610 is similar to Fig.16A The structure 1600 is similar to the structure 1600 shown in FIG. 1 , but the difference is that the structure 1610 further includes a dielectric layer 1608 disposed between the substrate 1602 and the metal reflector 1606. The dielectric layer 1608 can change an optical reflection spectrum of the metal reflector 1606. For example, the dielectric layer 1608 (e.g., SiO 2 ) due to a thin film interference caused by the metal reflector 1606 (e.g., an aluminum layer), the reflectivity of the metal reflector 1606 (e.g., an aluminum layer) for light of certain wavelengths incident thereon can be improved (e.g., the reflectivity can be improved from less than 90% to more than 97%), while the reflectivity of incident light of other wavelengths may be reduced.

[0306] Fig. 16C An example of structure 1620 is shown. Structure 1620 is similar to Fig.16A The structure 1600 is shown in FIG. 1 , but the difference is that the metal reflector 1606 of the structure 1600 is replaced by a dielectric reflector 1626. The dielectric reflector can be a single layer of dielectric film or a stack of various dielectric films. The dielectric reflector can be formed of various dielectric materials, such as SiO 2 、Si3N 4 , SiON and Si. At the operating wavelength of the photodetector of structure 16202, the dielectric mirror can have a reflectivity of up to 50%, 60%, 70%, 80%, 90% or 95%. The dielectric mirror 1626 can have a thickness in the range of 50nm to 4000nm.

[0307] Fig.16D An example of structure 1630 is shown. Structure 1630 is similar to Fig. 16CThe structure 1620 is shown, but the difference is that the dielectric reflector 1626 in the structure 1620 is replaced by a distributed Bragg (DBR) reflector 1632. The DBR reflector includes a plurality of first dielectric layers 1634 and a plurality of second dielectric layers 1636, which are alternately stacked on each other. The second dielectric layer 1636 has a refractive index different from that of the first dielectric layer 1634. The first dielectric layer 1634 and the second dielectric layer 1636 can have a thickness equivalent to one quarter of the operating wavelength in their respective dielectric materials, and their reflectivity and reflection bandwidth can depend on their respective thicknesses, refractive indices, and the number of first-second layer pairs.

[0308] Fig.16E 1640. The structure 1640 includes a substrate 1602, an absorption layer 1604, and an ARC layer 1648. The ARC layer 1648 can reduce the reflection of the optical signal 1605 when it is incident on the absorption region 1604. The ARC layer 1646 can be similar to Figure 7B ARC layer 744 is shown in FIG.

[0309] Fig.16F An example of structure 1650 is shown. Structure 1650 is similar to Fig.16A The structure 1600 shown in FIG. 1 is different in that the metal reflector 1606 is now disposed on the upper surface of the substrate 1602 on the side where the absorption region 1604 is disposed. The optical signal 1605 is now incident on the absorption region 1604 through the lower surface of the substrate 1602, which can be referred to as a BSI structure. The effect of the metal reflector 1606 is similar to that of the structure shown in FIG. Fig.16A The effect described in the corresponding paragraph.

[0310] Figure 16G An example of structure 1660 is shown. Structure 1660 is similar to Fig. 16B The structure 1610 shown in FIG. 1 is different in that the dielectric layer 1608 and the metal reflector 1606 are now disposed on the upper surface of the substrate 1602 on the side where the absorption region 1604 is disposed. The optical signal 1605 is now incident on the absorption region 1604 through the lower surface of the substrate 1602, which can be referred to as a BSI structure. The effect of the dielectric layer 1608 and the metal reflector 1606 is similar to that of the BSI structure. Fig. 16B The effect described in the corresponding paragraph.

[0311] Fig.16H An example of structure 1670 is shown. Structure 1670 is similar to Fig. 16CThe structure 1620 shown in FIG. 1 is different in that the dielectric reflector 1626 is now disposed on the upper surface of the substrate 1602 on the side where the absorption region 1604 is disposed. The optical signal 1605 is now incident on the absorption region 1604 through the lower surface of the substrate 1602, which can be referred to as a BSI structure. The effect of the dielectric reflector 1626 is similar to that of the structure 1620 in FIG. Fig. 16C The effect described in the corresponding paragraph.

[0312] Fig.16I An example of structure 1680 is shown. Structure 1680 is similar to Fig.16D The structure 1630 shown in FIG. 1 is different in that the DBR reflector 1632 is now disposed on the upper surface of the substrate 1602 on the side where the absorption region 1604 is disposed. The optical signal 1605 is now incident on the absorption region 1604 through the lower surface of the substrate 1602, which can be referred to as a BSI structure. The effect of the DBR reflector 1632 is similar to that of the structure shown in FIG. Fig.16D The effect described in the corresponding paragraph.

[0313] Fig.16J An example of structure 1690 is shown. Structure 1690 is similar to Fig.16E The structure 1640 shown in FIG. 1 is different in that the ARC layer 1648 is now disposed on the lower surface of the substrate 1602 (opposite to the absorption region 1604). The optical signal 1605 is now incident on the absorption region 1604 through the lower surface of the substrate 1602, which can be referred to as a BSI structure. The effect of the ARC layer 1648 is similar to that of the Fig.16E The effect described in the corresponding paragraph.

[0314] In general, the reflector structures, such as the metal reflector 1606, the dielectric 1608, the dielectric reflector 1626, and the DBR reflector 1632, can be fabricated using different methods. For example, the reflector structure can be directly deposited on the substrate 1602. Alternatively or additionally, the reflector structure can be fabricated on a separate substrate and bonded to the substrate 1602 using a wafer bonding technique.

[0315] Although the individual implementations have the metal reflector 1606, dielectric 1608, dielectric reflector 1626, and DBR reflector 1632 on the lower or upper surface of the substrate 1602, in practice, the structures described can be implemented on both sides of the substrate 1602. For example, the DBR reflector 1632 can be implemented on both sides of the substrate 1602, which can establish an optical resonant cavity around the absorption region 1604 and change a spectral response of the photodetector. In other examples, the ARC layer 1648 can be implemented on the upper surface of the substrate 1602, and the reflector structure (e.g., structures 1600, 1610, 1620, and 1630) can be implemented on the lower surface of the substrate to further improve the light detection efficiency of the photodetector. In general, the reflectors, such as the metal reflector 1606, the dielectric layer 1608, the dielectric reflector 1626, and the DBR reflector 1632 can be partially reflective and partially transmissive.

[0316] The surface of the absorption region can be modified in various ways to adjust various performance characteristics of the light detection. Example modifications of the surface of the absorption region also include: the addition of doped regions, the introduction of foreign elements, changes in material composition, the introduction of the morphology of the surface of the absorption region, and the deposition of dielectric or semiconductor materials. Example performance characteristics include, for example, light absorption efficiency, optical absorption spectrum, carrier absorption efficiency, dark current or leakage current, operating power of the photodetector, and bandwidth of the photodetector.

[0317] Figures 17A-17E A cross-sectional view of an example structure of an absorption zone surface modification. Fig.17A In the embodiment, a surface modified absorption region 1700 includes a germanium silicon-based absorption region 1704 and a surface modified layer 1706. The germanium silicon-based absorption region 1704 may be an absorption region of a switch light detector; the switch light detector may be, for example, Figure 5D A switching photodetector is shown.

[0318] The SiGe-based absorption region 1704 may have a Si x Ge 1-x Compound. For example, the composition (X) may vary from 0.01 to 0.99, and a composition (X) of 0.01 may allow the germanium-silicon-based absorption region 1704 to have properties close to those of germanium, and a composition (X) of 0.99 may allow the germanium-silicon-based absorption region 1704 to have properties close to those of silicon. The composition of the germanium-silicon-based absorption region may affect its light absorption efficiency at the absorption wavelength, and may also affect the overall optical absorption spectrum. For example, a high germanium concentration corresponding to a low (X) composition may absorb a large amount of near-infrared wavelengths (e.g., greater than 1 μm), compared to a high silicon composition corresponding to a high (X) composition.

[0319] The surface modification layer 1706 can change the optical and / or electrical properties of the germanium-based silicon absorption region 1704 and the photodetector having the absorption region 1704. The surface modification layer can be formed of various materials, such as amorphous silicon, polycrystalline silicon, epitaxial silicon, Si with variable composition (Y), and / or other materials. Y Ge 1-Y Compounds, Ge with variable composition (Z) Z Sn 1-Z Compounds, and any combination thereof.

[0320] In some embodiments, for Si x Ge 1-x For a germanium-silicon-based absorption region 1704 of a component, the surface modification layer 1706 may be a Si Y Ge 1-Y A layer in which the composition (Y) is different from the composition (X). For example, when the composition (Y) is higher than the composition (X), the surface modification layer 1706 has a high absorption coefficient for long wavelengths compared to the germanium silicon-based absorption region 1704. In this case, incident light of a long wavelength can be largely absorbed in the surface modification layer 706 without penetrating deep into the germanium silicon-based absorption region 1704. By absorbing incident light at the surface of the germanium silicon-based absorption region 1704, the bandwidth of the photodetector containing the absorption region 1704 can be improved, which is attributed to the reduction of the diffusion of photogenerated carriers in the light absorption region 1704. In some embodiments, for a pure germanium absorption region 1704 (i.e., X=0), the surface modification layer 1706 can be a Si Y Ge 1-Y In some embodiments, the composition of the surface modification region 1706 and the SiGe-based absorption region 1704 can be changed along a direction (e.g., vertical direction) to form a gradient SiGe absorption region 1704. The gradient of SiGe composition can further improve the bandwidth of the photodetector. In some embodiments, the surface modification layer 1706 can be a multi-layer structure. For example, a SiGe layer can be deposited on top of a SiGe-based absorption region 1704 to provide protection, and another silicon layer can be deposited on top of the SiGe layer for further protection.

[0321] In some embodiments, the surface modification layer 1706 may be a germanium-tin alloy Ge with a variable composition (Z). Z Sn 1-Z Adding tin to the surface modification layer 1706 can improve the optical absorption efficiency at long wavelengths, for example, beyond the germanium energy band (about 1.55 μm). Generally, the absorption rate of pure germanium will drop significantly beyond this energy band.

[0322] See also Fig. 17B; A surface modification layer 1710 includes a germanium silicon-based absorption region 1704 and a first doped region 1712. In some embodiments, the first doped region 1712 can be doped with p-type or n-type dopants. P-type or n-type dopants can change the electrical properties of the absorption region 1704. For example, due to the first doped region 1712, photogenerated electrons (or holes) can be repelled from the surface, thereby avoiding surface recombination, which provides high absorption efficiency when the first doped region has p-type or n-type dopants. In some embodiments, the first doped region 1712 can be doped with impurities, such as silicon or tin, to adjust the optical properties of the absorption region 1704.

[0323] See also Fig. 17C A surface modification layer 1720 is similar to the surface modification layer 1710, but differs in that it further includes a second doping region 1722. The second doping layer 1722 may be similar to the first doping region 1712 or have a different polarity, depth, or width, so that photogenerated carriers may be attracted by the second doping region 1722 and repelled by the first doping region 1712.

[0324] See also Fig.17D A surface modification layer 1730 includes a germanium-silicon-based absorption region 1704 and a plurality of dielectric wells 1732. The dielectric wells 1732 can be filled with different dielectrics, such as SiO 2 、Si 3 N 4 When the dielectric well is placed in a PN junction or sandwiched between surface electrical terminals, it can reduce dark current or leakage current, reduce the operating power of the photodetector and / or increase the bandwidth of the photodetector.

[0325] See also Fig.17E A switch light detector 1740 includes Fig. 17B A surface modified germanium absorber layer 1710 is shown. The switch photodetector 1740 is similar to Figure 1B The switch photodetector 160 is shown, but the difference is that it also includes a surface modification layer 1706, and Fig.11A Shown are carrier collection terminal 1106 and carrier control terminal 1108. Adding surface modification layer 1706 can improve various performance characteristics of switch photodetector 1740, such as light absorption efficiency, dark current or leakage current, operating power of the photodetector, and bandwidth of the photodetector.

[0326] Although a separate embodiment of surface modification of the absorption region is shown, in practice the surface modification can be implemented in different combinations to achieve the desired effect. For example, the surface modification layer 1706 can be implemented in combination with the first doped region 1712 and / or the second doped region 1706. In another example, the surface modification layer 1706 can be implemented in combination with the dielectric well 1732. In yet another example, the surface modification layer 1706 can be implemented in combination with the first doped region 1712 and / or the second doped region 1722, and the dielectric well 1732.

[0327] Various doped regions and wells, such as p-type doped regions and wells, and n-type doped regions and wells, can be placed at different locations in the absorption region, substrate, or interposer to change the performance characteristics of the device. Exemplary performance characteristics include: light absorption efficiency, light absorption spectrum, carrier collection efficiency, dark current or leakage current, operating power of the photodetector, and photodetector bandwidth.

[0328] The depth of the doped regions and wells can be determined based on various considerations, such as manufacturability and device performance. One or more doped wells and regions can be connected to a voltage or current source. One or more doped wells and regions can also be not connected to a voltage or current source (i.e., floating) and / or connected to each other (i.e., shorted).

[0329] Figures 18A-18B 1800 is a top view and a side view of an example of a switch light detector 1800. The switch light detector 1800 is similar to Figure 1B The switch photodetector 160 is shown, but also includes Fig.11A The carrier collection terminal 1106 and the carrier control terminal 1108 are shown. Figure 1B As described, the n-type well regions 152 and 154 can reduce a leakage current from the first control signal 122 to the second control signal 132 and can reduce charge coupling between the n-type doping regions 126 and 136. Reducing the leakage current can reduce the operating power of the switching photodetector 1800.

[0330] Figures 18C-18D A top view and a side view of an example of a switch light detector 1820 are shown. The switch light detector 1820 is similar to Figures 18A-18B The illustrated switching photodetector 1800, however, also includes a p-type well region 1822. The p-type doped region 1822 may be similar to Figure 2D P-type well regions 246 and 248 are shown. P-type well region 1822 can increase the absorption rate of photocarriers of the switching photodetector 1820 compared to the switching photodetector 1800.

[0331] In some cases, the n-type doped regions 126 and 136 cannot completely absorb the photogenerated carriers in the absorption region 106. In such cases, the photogenerated carriers may reach the material surface where material defects appear between the substrate 102 and the absorption region 106. The material defects can capture the photogenerated carriers and release these carriers after a period of time; the n-type doped regions 126 and 136 then collect the aforementioned carriers. This phenomenon of intercepting and releasing carriers by material defects at the interface and the subsequent collection of carriers by the n-type doped regions 126 and 136 may reduce the bandwidth of the switching photodetector 1800, which is due to the time delay effect of intercepting and releasing carriers. Therefore, the addition of the p-type well region 1822 can mitigate the reduction in bandwidth because it can prevent the carriers from being collected by the n-type doped regions 126 and 136 by not allowing the photocarriers to reach the interface between the absorption region 106 and the substrate 102.

[0332] Fig.18E A top view of an example of a switch light detector 1830 is shown. The switch light detector 1830 is similar to Figures 18C-18D The illustrated switching photodetector 1820 further includes a p-type well region 1832. The p-type well region 1832 is similar to the p-type well region 1822. The combination of the p-type well regions 1822 and 1832 surrounds the corresponding n-type doped regions 126 and 136, which can further prevent the photogenerated carriers from reaching the interface between the absorption region 106 and the substrate 102 and preventing the carriers from being absorbed by the n-type doped regions 126 and 136. Although separate p-type well regions 1822 and 1832 are shown here, the p-type well regions 1822 and 1832 can be combined into a C-shaped region surrounding the corresponding n-doped regions.

[0333] Figure 18F-18G 18 shows an exemplary top view and side view of the switch light detector 1830. The switch light detector 1840 is similar to Figures 18A-18B The switch photodetector 1800 is shown, but the difference is that the n-type well regions 152 and 154 are omitted, and further includes a p-type well region 1842. The p-type well region 1842 may be similar to Figure 2D 1842. The p-type well region 1842 surrounds the absorption region 106 embedded in the substrate 102. The p-type well region 1842 can block the photogenerated carriers of the absorption region 106 from reaching the substrate 102. Such blocking can increase the collection efficiency of the photocarriers of the switching photodetector 1840 compared to the switching photodetector 1800. The p-type doped region 1842 can be formed in the absorption region 106, the substrate 102, an intermediate layer between the absorption region 106 and the substrate 102, or a combination thereof.

[0334] Although separate implementations of n-type well regions 152 and 154 and p-type well regions 1822 , 1832 , and 1842 are shown, in practice the n-type well regions and p-type well regions may be implemented in combination to achieve the desired effect.

[0335] So far, several specific implementations of components of a switch photodetector and various arrangements of the components have been introduced. Next, the combination of the aforementioned components will be described. The combinations described here are not a complete list of all combinations.

[0336] Fig.19A -B shows a top view and a side view of an example of a switch light detector 1900. The switch light detector 1900 is similar to Figure 1A The switch photodetector 100 is shown in FIG. 1 , but the difference is that the absorption region of the switch photodetector 1900 is completely embedded in the substrate 102 and further includes Fig.11A The carrier collecting terminal 1106 and the carrier controlling terminal 1108 are shown. The light receiving area 1205 is Figures 12A-12B The presence of the p-type doped regions 128 and 138 results in an ohmic contact being formed at the interface between the carrier control terminal 1108 and the absorption region 106 .

[0337] Fig.19C -D shows a top view and a side view of an example of a switch light detector 1910. The switch light detector 1910 is similar to Figures 19A-19B The switching photodetector 1900 is shown, but the difference is that the p-type doped regions 128 and 138 are omitted. The omission of the p-type doped regions 128 and 138 causes the interface between the carrier control terminal 1108 and the absorption region 106 to form a Schottky junction.

[0338] Fig.19E -F shows a top view and a side view of an example of a switch light detector 1920. The switch light detector 1920 is similar to Fig.19A -B shows the switching photodetector 1900 , but the difference is that p-type doped regions 128 and 138 are added, and a carrier control terminal 1108 is added on each side of the light receiving region 1205 .

[0339] Figure 19G -H shows a top view and a side view of an example of a switch light detector 1930. The switch light detector 1930 is similar to Fig.19E -F shows a switching photodetector 1920, but the difference is that the p-type doping regions 128 and 138 are omitted. The omission of the p-type doping regions 128 and 138 causes the interface between the carrier control terminal 1108 and the absorption region 106 to form a Schottky junction.

[0340] Fig. 20A-B shows a top view and a side view of an example of a switch light detector 2000. The switch light detector 2000 is similar to Fig.19A -B shows a switch photodetector 1900, but the difference is that the Fig.10I The interposer 1006 is shown. Fig.10I As described in the corresponding paragraph, the intermediate layer 1006 has an opening deep into the substrate 102, the absorption region 106 fills the opening and contacts the substrate 102, and the opening is formed by the intermediate layer 1006. In some embodiments, the intermediate layer 1006 may be SiO 2 、SiN x 、AlO x , or any oxide or nitride based insulator.

[0341] Fig. 20C -D shows a top view and a side view of an example of a switch light detector 2010. The switch light detector 2010 is similar to Fig. 20A -B shows a switch photodetector 2000, but the difference is Fig. 20A -B interposer 1006 is replaced by another interposer 2012. The material of interposer 2012 is similar to that of interposer 1006, but the difference is that interposer 2012 is a uniform layer that spans an upper surface of substrate 102 and has an opening that reaches deep into the substrate. Absorption region 106 is embedded in the opening of interposer 2012. In some embodiments, interposer 2012 may be SiO 2 、SiN x 、AlO x , or any oxide or nitride based insulator.

[0342] Fig.20E -F shows a top view and a side view of an example of a switch light detector 2020. The switch light detector 2020 is similar to Fig. 20C -D shows a switching photodetector 2010, but the difference is that the p-type doping regions 128 and 138 are omitted. The omission of the p-type doping regions 128 and 138 causes the interface between the carrier control terminal 1108 and the absorption region 106 to form a Schottky junction.

[0343] Figure 20G -H shows a top view and a side view of an example of a switch light detector 2030. The switch light detector 2030 is similar to Fig. 20C -D shows the switch photodetector 2010, but the difference is Fig. 20C -D's interposer 2012 is replaced by another interposer 2032. Interposer 2032 is similar to Fig. 20C-D, but the difference is that the interposer 2032 has a first opening 2034 that reaches deep into the substrate 102 , and a second opening 2036 that is larger than the first opening 2034 , and the second opening 2036 opens toward the upper surface of the interposer 2032 .

[0344] Fig.20I -J shows a top view and a side view of an example of a switch light detector 2040. The switch light detector 2040 is similar to Figure 20G -H shows a switching photodetector 2030, but the difference is that the p-type doping regions 128 and 138 are omitted. The omission of the p-type doping regions 128 and 138 causes the interface between the carrier control terminal 1108 and the absorption region 106 to form a Schottky junction.

[0345] Figure 20K -L shows a top view and a side view of an example of a switch light detector 2050. The switch light detector 2050 is similar to Figure 20G -H shows a switching photodetector 2030, but the difference is that n-type well regions 152 and 154 are added. Figure 1B are described in the relevant paragraphs.

[0346] Fig.21A -B shows a top view and a side view of an example of a switch light detector 2100. The switch light detector 2100 is similar to Fig.19A -B shows a switching photodetector 1900, but the difference is that the n-type doping regions 126 and 136, the p-type doping regions 128 and 138, the carrier collection terminal 1106 and the carrier control terminal 1108 are moved from the absorption region 106 to the substrate 102. Such terminals 1106 and 1108 can be referred to as substrate carrier collection terminals and substrate carrier control terminals.

[0347] Fig. 21C -D shows a top view and a side view of an example of a switch light detector 2110. The switch light detector 2110 is similar to Fig.21A -B shows a switching photodetector 2100, but the difference is that the absorber p-type doped regions 2128 and 2138 and the absorber carrier control terminal 2108 are arranged in the absorption region 106. The substrate carrier absorption terminal 1106, the substrate carrier control terminal 1108 and the absorber carrier control terminal 2108 can be similar to those in Fig.14A The substrate carrier collecting terminal 1306, substrate carrier controlling terminal 1308 and absorber carrier controlling terminal 1309 are described in the relevant paragraphs and have similar effects.

[0348] Fig.21E-F shows a top view and a side view of an example of a switch light detector 2120. The switch light detector 2120 is similar to Fig. 21C -D shows a switching photodetector 2110, but the difference is that the absorber p-type doped regions 2128 and 2138 are omitted. The omission of the absorber p-type doped regions 2128 and 2138 causes the interface between the absorber carrier control terminal 2108 and the absorption region 106 to form a Schottky junction.

[0349] Fig.22A -B shows a top view and a side view of an example of a switch light detector 2200. The switch light detector 2200 is similar to Fig.18F -G shows a switch photodetector 1840, but the difference is that the Fig.18A - n-type well regions 152 and 154 shown in B.

[0350] Fig. 22C -D shows a top view and a side view of an example of a switch light detector 2210. The switch light detector 2210 is similar to Fig. 21C -D shows a switch photodetector 2210, but the difference is that Fig.18A - n-type well regions 152 and 154 in B.

[0351] Fig.23A A top view showing an example of a switch light detector 2300, and Fig. 23B FIG. 2 shows a side view of an example of a switch light detector 2300 connected along an AA line. The switch light detector 2300 is similar to Fig. 21C -D shows a switching photodetector 2110, but the difference is that a p-type well region 2302 is added to the interface between the absorption region 106 and the substrate 102. The p-type well region 2302 can help to ease the carrier collection and release at the junction between the absorption region 106 and the substrate 102, which has been Fig. 18C -D is described in the relevant paragraph.

[0352] Fig.24A -B shows a top view and a side view of an example of a switch light detector 2400. The switch light detector 2400 is similar to Fig. 18C -D shows a switching photodetector 1820, but the difference is that the n-type well regions 152 and 154 are omitted.

[0353] Fig.24C FIG. 24 is an exemplary top view of a switch light detector 2410. The switch light detector 2410 is similar to Fig.18E The difference between the switch photodetector 1830 shown is that Fig.18E The p-type well regions 1822 , 1832 shown in FIG. 1 are merged into a continuous p-type well region 2412 .

[0354] Fig.24D -E shows a top view and a side view of an example of a switch light detector 2420. The switch light detector 2420 is similar to Fig.24A -B shows a switching photodetector 2400, but the difference is that a dielectric well 2422 is added in the n-type doped regions 126 and 136. The dielectric well 2422 is similar to Fig.17D 1732 is shown in the figure. The dielectric well 2422 is disposed in a portion of the n-type doped region 126 between the carrier collection terminal 1106 and the carrier control terminal 1108. The dielectric well 2422 can reduce the dark current between the carrier collection terminal 1106 and the carrier control terminal 1108. The depth of the dielectric well 2422 can be less than, equal to, or greater than the depth of the n-type doped region 126.

[0355] Fig.24F -G depicts a top view and a side view of an example of a switch light detector 2430. The switch light detector 2430 is similar to Fig.24D -E, but the difference is that the dielectric well 2422 is moved from the n-type doping regions 126 and 136 to the p-type doping regions 128 and 138. The depth of the dielectric well 2422 can be less than, equal to, or greater than the depth of the p-type doping region 128. In general, the dielectric well 2422 can be disposed at any position between the n-type doping region 126 and the p-type doping region 128, and between the n-type doping region 136 and the p-type doping region 138.

[0356] Fig.25A -B shows a top view and a side view of an example of a switch light detector 2500. The switch light detector 2500 is similar to Fig.19A -B shows a switch photodetector 1900, but the difference is that the Fig.16F The metal reflector 1606 is shown, wherein the metal reflector is disposed on an upper surface of the absorption region 106 having the carrier collection terminal 1106 and the carrier control terminal 1108. The metal reflector 2502 can be disposed above the light receiving region 1205. In some embodiments, the metal reflector 2502 can be implemented by the first metal layer (M1) or the second metal layer (M2) in the CMOS process, or a combination thereof.

[0357] Fig.25C -D shows a top view and a side view of an example of a switch light detector 2510. The switch light detector 2510 is similar to Fig.25A -B shows a switching photodetector 2500, but the difference is that the p-type doping regions 128 and 138 are omitted. The omission of the p-type doping regions 128 and 138 causes the interface between the carrier control terminal 1108 and the absorption region 106 to form a Schottky junction.

[0358] Fig.25E -F shows a top view and a side view of an example of a switch light detector 2520. The switch light detector 2520 is similar to Figure 20K -L shows the switch photodetector 2050, but the difference is that the Fig.16F The metal reflector 1606 is shown as the metal reflector 2502, wherein the metal reflector 2502 is disposed on an upper surface of the absorption region 106 having the carrier collection terminal 1106 and the carrier control terminal 1108. The metal reflector 2502 may be disposed above the light receiving region 1205. In some embodiments, the metal reflector 2502 may be implemented by the first metal layer (M1) or the second metal layer (M2) in the CMOS process, or a combination thereof.

[0359] Figure 25G -H shows a top view and a side view of an example of a switch light detector 2530. The switch light detector 2530 is similar to Fig.18F -G shows a switch photodetector 1840, but the difference is that the Fig.16F The metal reflector 1606 is shown as the metal reflector 2502, wherein the metal reflector 2502 is disposed on an upper surface of the absorption region 106 having the carrier collection terminal 1106 and the carrier control terminal 1108. The metal reflector 2502 may be disposed above the light receiving region 1205. In some embodiments, the metal reflector 2502 may be implemented by the first metal layer (M1) or the second metal layer (M2) in the CMOS process, or a combination thereof.

[0360] In a typical implementation of an image sensor, multiple sensor pixels (e.g., switching photodetectors) are arranged in an array to allow the image sensor to capture images with multiple image pixels. Viewed from the top of the image sensor, square sensor pixels with the same size on both sides can be a simple two-dimensional array. However, for some applications, such as ToF, some sensing pixels may not be square, but rectangular. For example, in Figure 1B In the embodiment, the switching photodetector 160 has two carrier control terminals (e.g., p-type doped regions 128 and 138), and two carrier collection terminals (e.g., n-type doped regions 126 and 136). The four terminals are generally arranged along a line, which results in a rectangular sensor pixel having a longer shape along the line where the terminals are arranged (e.g., Fig.18A Switching light detector 1800 is shown).

[0361] Such rectangular sensor pixels may make it difficult to arrange the pixels efficiently, for example due to design rules associated with semiconductor fabrication in a foundry. Design rules may restrict various minimum spacings of features such as doped regions, doped wells, dielectric wells, and germanium absorption regions. One approach to improve compactness and symmetry is to create a unit cell of a photodetector that contains four rectangular photodetectors. Fig.26 An example of a unit cell of a rectangular photodetector is shown. The unit cell 2600 comprises Fig.18A Four switch photodetectors 1800 are shown, and four isolation structures 2602 are respectively surrounded by the switch photodetectors 1800. The isolation structures 2602 are Fig.15A The unit cell 2600 can improve the compactness and symmetry of the sensor pixels on the rectangular unit cell.

[0362] Fig. 27 A top view of an example of a rectangular switched photodetector 2700 with phototransistor gain is shown. The switched photodetector 2700 is similar to Fig.18A 100. The switched photodetector 1800 is shown as a switched photodetector, but with the addition of an electron emitter 2710 on the substrate 102. The electron emitter 2710 can be similar to the n-type doped regions 126 and 136. The rectangular shape of the switched photodetector 1800 allows a photocurrent integration capacitor (e.g., a floating diffusion capacitor) to be coupled to a bipolar junction transistor (BJT) 2720; the BJT 2720 is an NPN bipolar junction transistor formed by the n-type doped regions 126 and 136, the p-type doped regions 128 and 138, and the electron emitter 2710. When the BJT 2720 has an appropriate bias voltage, it can provide phototransistor gain in response to an incident light signal, which can improve the light-to-photocurrent conversion efficiency of the photodetector 2700. For example, the bipolar junction transistor 2720 can be biased under the following conditions: the bias voltage of the n-type doping regions 126 and 136 is between 1V and 3V, the bias voltage of the p-type doping regions 128 and 138 is between 0V and 1V, and the bias voltage of the electron emitter 2710 is lower than the bias voltage of the corresponding n-type doping regions 126 and 136.

[0363] Generally, an external voltage should be applied to the electron emitter 2710 and / or the n-type doped regions 126 and 136 or shorted to the p-type doped regions by a metal connection to allow electrons to be emitted from the electron emitter 2710 .

[0364] Although various embodiments of switching photodetectors having specific combinations and arrangements of n-type and p-type regions and wells have been described, in general the polarity of the doped regions and wells can be reversed and similar operation and functionality achieved. For example, all instances of p-type wells and p-type doped regions can be converted to n-type wells and n-type doped regions, respectively, and all n-type well regions and n-type doped regions can be converted to p-type wells and p-type doped regions, respectively.

[0365] Fig.28A An exemplary imaging system 2800 is shown for determining characteristics of a target object 2810. The target object 2810 may be a three-dimensional object. The imaging system 2800 may include a transmitter unit 2802, a receiver unit 2804, and a processing unit 2806. In general, the transmitter unit 2802 emits light 2812 toward the target object 2810; the transmitter unit 2802 may include one or more light sources, control circuits, and / or optical elements. For example, the transmitter unit 2802 may include one or more NIR light emitting diodes or lasers; wherein the emitted light 2812 may be collimated by a collimating lens to pass in free space.

[0366] In general, the receiver unit 2804 receives reflected light 2814 reflected by the target object 2810. The receiver unit 2804 may include one or more photodiodes, control circuits, and / or optical elements. For example, the receiver unit 2804 may include an image sensor; wherein the image sensor includes a plurality of pixels fabricated on a semiconductor substrate. Each pixel may include one or more switch light detectors for detecting the reflected light 2814, wherein the reflected light 2814 may be focused onto the switch light detector. Each switch light detector may be a switch light detector disclosed in the present patent application.

[0367] In general, the processing unit 2806 processes the photocarriers generated by the receiver unit 2804 and determines the characteristics of the target object 2810. The processing unit 2806 may include control circuitry, one or more processors, and / or a computer storage interface that may store instructions for determining the characteristics of the target object 2810. For example, the processing unit 2806 may include: a readout circuit and a processor, wherein the processor may process information related to the collected photocarriers to determine the characteristics of the target object 2810. In some embodiments, the characteristics of the target object 2810 may be depth information of the target object 2810. In some embodiments, the characteristics of the target object 2810 may be material composition of the target object 2810.

[0368] Fig.28BAn exemplary technique for determining the characteristics of a target object 2810 is shown. The transmitter unit 2802 may generate a light pulse 2812, which may be modulated at a frequency of fm and a duty cycle of 50% as an example. The receiver unit 2804 may receive a reflected light pulse 2814 with a phase shift of Φ. The switch photodetector is controlled so that the readout circuit 1 reads out the collected charge Q1 whose phase is synchronized with the emitted light pulse, and the readout circuit 2 reads out the collected charge Q2 whose phase is opposite to the emitted light pulse. In some embodiments, the distance D between the imaging system 2800 and the target object 2810 may be obtained by the following formula:

[0369]

[0370] where c is the speed of light.

[0371] Fig.28C Another exemplary technique for determining a characteristic of a target object 2810 is shown. The transmitter unit 2802 can emit a light pulse 2812, which is illustratively pulsed at a frequency fm and a duty cycle of less than 50%, but by reducing the duty cycle of the light pulse by a factor N and increasing the intensity of the light pulse 2812 by this factor N, the signal-to-noise ratio of the received reflected light pulse 2814 can be improved, while keeping the energy consumption of the imaging system 2800 substantially unchanged. This can be achieved when the component bandwidth is increased so that the duty cycle of the light pulse can be shortened without deforming the pulse shape. The receiver unit 2804 can receive the reflected light pulse 2814 with a phase shift of Φ. The multi-gate photodiode is controlled so that the readout circuit 1 reads out the collected charge Q1' whose phase is synchronized with the transmitted light pulse, and the readout circuit 2 reads out the collected charge Q2' whose phase is delayed with the transmitted light pulse. In some embodiments, the distance D between the imaging system 2800 and the target object 2810 can be obtained by the following formula:

[0372]

[0373] Fig.29 An example of a flowchart 2900 for determining characteristics of an object by an imaging system is shown. The flowchart 2900 may be performed by the imaging system 2800, for example.

[0374] The system receives reflected light (step 2902 ). For example, the transmitter unit 2802 may transmit a NIR light pulse 2812 toward the target object 2810 . The receiver unit 2804 may receive a reflected NIR light pulse 2814 reflected from the target object 2810 .

[0375] The system determines phase information (step 2904). For example, the receiver unit 2804 may include an image sensor, wherein the image sensor includes a plurality of pixels fabricated on a semiconductor substrate. Each pixel may include one or more switching photodetectors for detecting the reflected light pulse 2814. The switching photodetectors may be of the type disclosed in this patent, wherein the phase information may be obtained by referring to Fig.28B or Fig.28C The technique described is used to determine.

[0376] The system determines the object characteristics (step 2906). For example, the processing unit 2806 can determine the object characteristics based on the phase information by referring to Fig.28B or Fig.28C The described technology determines the depth information of the object 2810.

[0377] In some embodiments, an image sensor includes a plurality of pixels fabricated on a semiconductor substrate, wherein each pixel may include one or more switching photodetectors 100, 160, 170, 180, 200, 250, 260, 270, 300, 360, 370, 380, 400, 450, 460, 470, and 480 as shown in FIG. Fig.28A and Fig.28B The isolation between these pixels can be achieved by insulating isolation, such as an oxide layer or a nitride layer, or by an injection isolation, such as using a p-type or n-type region to block signal electrons or holes, or by an intrinsic built-in energy barrier (such as using a germanium-silicon heterojunction).

[0378] So far, different switching photodetectors have been described, and how they can be used in time-of-flight detection systems (e.g. Fig.28A ; now, the receiver unit 2804 in the imaging system 2800 will be described. Fig.30 An example of a receiver unit 3000 for ToF detection is shown. The receiver unit 3000 includes a pixel array 3010, an amplifier array 3020, and an analog-to-digital converter (ADC) array 3030. The pixel array 3010 is electrically coupled to the amplifier array 3020, and the amplifier array 3020 is electrically coupled to the ADC array 3030.

[0379] The pixel array 3010 includes a plurality of photodetectors (e.g., the switching photodetectors described above) and a plurality of capacitors for storing photogenerated carriers from the switching photodetectors. The pixel array 3010 is a two-dimensional array (i.e., an M×N array) including M columns and N rows of photodetectors and capacitors. The capacitors may be integrated with the photodetectors or implemented separately. Exemplary capacitors include floating-diffusion capacitors, metal-oxide-metal (MOM) capacitors, and metal-insulator-metal (MIM) capacitors. The pixel array 3010 includes pixel transistors for controlling the operation of the photodetectors, such as controlling the charge reading of the switching photodetectors. The pixel array 3010 may be part of an imaging sensor that includes various optical elements associated with light detection, such as reflectors, lenses, and anti-reflection film layers.

[0380] The amplifier array 3020 includes one or more amplifiers 3022. The amplifier 3022 is used to amplify the electrical signals generated by the individual pixels of the pixel array 3010. The amplifier 3022 may be a voltage gain amplifier, and is used to amplify the voltage established by integrating the photocurrent on the capacitor. The amplifier 3022 may be a charge-voltage amplifier, and is used to convert the charge stored in the capacitor into a voltage for output. The amplifier 3022 may be a gain variable amplifier, which can optimize the detection sensitivity within the magnitude range of the light signal received by the pixel array 3010. The amplifier 3022 may be a differential amplifier, and is used, for example, to amplify the difference between the two output voltages of the switching light detector. Such a differential detection mechanism can improve the detection sensitivity of ToF.

[0381] Different embodiments of the amplifier array 3020 may have different numbers of amplifiers 3022. In some embodiments, each pixel of the pixel array 3010 is coupled to a dedicated amplifier 3022. Such a configuration may allow all pixels to be read simultaneously to maximize the image data acquisition rate. In some embodiments, each row or column of the pixel array shares one amplifier 3022; for example, for an M×N pixel array 3010, there may be M or N amplifiers 3022. Such a structure of sharing amplifiers 3022 may improve the scalability of the receiver unit 3000 when there are a large number of pixels (e.g., megapixels). In some embodiments, each row or column may be further divided into multiple subsections that share multiple amplifiers 3022. In some embodiments, for a small pixel array 3010, all pixels in the array may share one amplifier 3022. Generally speaking, a group of pixels in multiple columns and rows can be grouped and share one amplifier 3022; for example, for an M×N pixel array 3010, it can have K×L amplifiers 3022, where K≤M and L≤N.

[0382] The analog-to-digital converter array 3030 includes one or more analog-to-digital converters. The analog-to-digital converter is used to convert the analog voltage or current signal output by the amplifier 3022 into a digital output 3040 having N bits. The digital output 3040 can be, for example, Figure 5A The processing unit 506 shown receives and performs ToF detection. The number of output bits N is used to determine the resolution of the analog-to-digital converter, which can be set according to the sensitivity and conversion speed considerations of a given application. Examples of various forms of analog-to-digital converters include flash analog-to-digital converters (flash ADC), successive-approximation-register analog-to-digital converters (successive-approximation-register ADC), and delta-sigma analog-to-digital converters (delta-sigma ADC). The analog-to-digital converter may be a differential analog-to-digital converter and, for example, is used to convert the difference in the amplified voltage output by the differential amplifier 3022. Similar to various embodiments of the amplifier array 3020, the analog-to-digital converter array 3030 may have a different number of analog-to-digital converters. Depending on various design considerations (e.g., the required conversion speed), the number of analog-to-digital converters may be equal to (i.e., one-to-one correspondence) or less than (i.e., multiple amplifiers share one analog-to-digital converter) the number of amplifiers 3022 in the amplifier array 3020.

[0383] An imaging system (e.g., ToF system 500) can operate over a wide range of light signal levels. For example, the light signal level can be affected by ambient light conditions, the reflectivity of an object, or the distance between the object and the imaging system 500; under different operating conditions, the signal level can vary by several orders of magnitude (e.g., 2 or more orders of magnitude, 10 or more orders of magnitude, 100 or more orders of magnitude). Changes in the light signal level will generally cause a linearly proportional change in the photocurrent generated by the photodetector of the pixel.

[0384] The operation of the pixel array 3010 is to integrate the photocurrent generated by each photodetector in the pixel on a respective capacitor over a period of time (e.g., a rated integration time) to generate an electrical signal proportional to the detected light signal. For example, at the beginning of an image acquisition cycle, a capacitor may be charged with a preset voltage (e.g., 1.8V); the charge stored in the capacitor is determined by the formula Q = C*V, where C is the capacitance of the capacitor and V is the voltage of the capacitor. Thereafter, the charged capacitor will be charged with the photocurrent I ph Discharge is performed; wherein, I ph =ΔQ / Δt, i.e., the change in charge Q within a given change time t. The photocurrent I generated by the photodetector ph The magnitude of directly affects the discharge rate of the associated capacitor. The maximum integration time t max =Q / I ph , which is the time required for the capacitor to fully discharge. When the maximum integration time is shorter than the rated integration time and allows the capacitor to fully discharge (and more generally, when the capacitor is discharged to a second predetermined voltage), the pixel is said to be "corrupted" or "bloomed", and the electrical output of the pixel at this point is no longer proportional to the received light input, resulting in distortion of the acquired image or incorrect ToF measurement. Therefore, if the light signal is large enough to corrupt one or more pixels within a predetermined integration period, the integration of the signal can be terminated before the pixels are damaged. Premature termination of the integration of the light signal may result in the generation of sub-frames that are the result of integration after the rated integration time has been exceeded. Multiple sub-frames can be acquired when the rated integration time has been exceeded, and multiple sub-frames can be pre-processed to produce a single image frame.

[0385] Once the collection of the light signal is terminated, the electrical output of the pixel is amplified by the amplifier 3022 and converted by the analog-to-digital converter array 3030 into a digital output 3040 to generate a frame or sub-frame. Once the conversion is complete, the capacitor will be recharged to the preset voltage and the acquisition cycle repeats. Since large light signals cause the maximum integration time of each acquisition cycle to decrease accordingly, large light input signals cause the rate of generation of sub-frames and digital outputs 3040 to increase accordingly. In some applications, such as due to increased power consumption or increased system complexity to support increased data production rate, the increase in data output may not be desirable. Therefore, increasing the maximum integration time of the pixel, reducing the output data rate of the analog-to-digital converter, reducing the output data rate of the ToF receiver unit 3000, or a combination thereof are solutions to the needs.

[0386] One way to increase the maximum integration time or reduce the output data rate is to increase the capacitance of the capacitor associated with each pixel of the pixel array 3010. Increasing the capacitance by a constant factor can increase the maximum integration time by approximately the same constant factor and can increase a dynamic range of the ToF receiver unit. However, capacitors are physical structures fabricated at the device level, while photodetectors are fabricated at the back-end interconnect level, and the capacitance is typically linearly proportional to their total area. Therefore, the capacitance of the capacitor integrally formed on the photodetector is limited by the available space on the sensor wafer on which the pixel array 3010 is fabricated. These issues can be addressed by fabricating additional capacitors on a second wafer and bonding it to the sensor wafer to further increase the capacitance of each pixel.

[0387] Fig.31A and 31B An architectural diagram and a cross-sectional view of a ToF receiver unit 3100 with increased capacitance are shown. The ToF receiver unit 3100 includes an integrated circuit (IC) wafer 3110, a sensor wafer 3130, and a plurality of interconnects 3170. The IC wafer 3110 includes a first capacitor 3112 and a plurality of pixel transistors 3120. The sensor wafer 3130 includes a second capacitor 3132 and a ToF pixel 3140. The IC wafer 3110 and the sensor wafer 3130 are bonded together, such as by a wafer bonding process, and the interconnect 3170 is electrically coupled to the ToF pixel 3140, the first capacitor 3112, the second capacitor 3132, and the pixel transistor 3120. By utilizing available space on the IC wafer 3110 and the sensor wafer 3130 to fabricate the capacitors 3112 and 3132, the total capacitance value integrated into the ToF pixel 3140 can be increased by a factor of two compared to a structure where both the first and second capacitors are fabricated on a single wafer. The increase in capacitance value can increase the maximum integration time of the ToF pixel 3140 and reduce the subframe generation rate and corresponding data throughput by the same factor.

[0388] The ToF pixel 3140 may be a switch photodetector, such as the switch photodetector 100. The ToF pixel 3140 includes a first switch 3150 and a second switch 3160, each having a corresponding readout (collection) terminal 3152 and 3162 (R), and a control (regulation) terminal 3154 and 3164 (C). The first switch 3150, for example, may be similar to Figure 1A The first switch 108 shown, carrying a readout terminal 3152, may be similar to the n-type doped region 126, and carrying a control terminal 3154, may be similar to the p-type doped region 128. Similarly, the second switch 3160, for example, may be similar to Figure 1A The second switch 110 shown carries a readout terminal 3162 which may be similar to the n-type doped region 136 and a control terminal 3164 which may be similar to the p-type doped region 138. The ToF pixel 3140 may be a backside illuminated pixel, and the light signal may enter the ToF pixel 3140 from the back side of the sensor wafer 3130 opposite to the side on which the ToF pixel 3140 is fabricated.

[0389] The pixel transistor 3120 is a transistor configured to control the operation of the ToF pixel 3140. The pixel transistor 3120 includes a first readout transistor 3122 and a second readout transistor 3124 for collecting carriers from the readout terminals 3152 and 3162. The pixel transistor 3120 may include a readout circuit having a 3T structure (i.e., a three-transistor structure having a reset, a source follower, and a column select transistor), or may include a circuit similar to Figure 1A Pixel transistor 3120 may include control transistors 3126 and 3128 to provide control signals to control terminals 3154 and 3164. The control signals provided by control transistors 3126 and 3128 may be similar to Figure 1A Control signals 122 and 132 are shown.

[0390] The first capacitor 3112 and the second capacitor 3132 can be implemented using standard semiconductor IC manufacturing technology. For example, the first capacitor 3112 and the second capacitor 3132 include metal-oxide-metal (MOM) capacitors and metal-insulator-metal (MIM) capacitors. In some embodiments, the oxide or insulator of the capacitor can be made of a high-k dielectric constant material, such as Al 2 O 3 , HfO 2 、ZrO 2 or La 2 O 3Although the capacitors are illustrated as parallel plate capacitors, various structures having capacitance values ​​may be used as capacitors 3112 and 3132, including floating diffusion capacitors and metal-oxide-semiconductor (MOS) capacitors. Generally, each of first capacitor 3112 and second capacitor 3132 may be implemented as a set of capacitors in parallel, depending on manufacturability or performance considerations.

[0391] exist Fig.31B , the first capacitor 3112 and the second capacitor 3132 are electrically coupled to the input terminals (e.g., source or drain) of the readout transistors 3124 and 3122, respectively. The first readout transistor 3122 collects carriers from the carrier readout terminal 3152 and provides the collected carriers to the second capacitor 3132 via its output terminal (e.g., source or drain). The second readout transistor 3124 collects carriers from the carrier readout terminal 3162 and provides the collected carriers to the first capacitor 3112 via its output terminal (e.g., source or drain). In such a structure, a voltage can be applied to the gates of the readout transistors 3122 and 3124 to control the conversion of carriers from the ToF pixel 3140 to the corresponding transistors 3132 and 3112.

[0392] Although a specific association of capacitors 3112 and 3132 to readout terminals 3152 and 3162 has been described, generally the associations between capacitors 3112 and 3132 and readout terminals 3152 and 3162 can be interchanged and operate in a similar manner. The respective terminals of the first capacitor 3112 and the second capacitor 3132 that are not connected to the readout terminals 3124 and 3122 can be, for example, grounded, floating, or connected to a power supply.

[0393] The IC wafer 3110 and the sensor wafer 3130 may be bonded in various ways. For example, bonding techniques include metal-metal bonding, oxide-oxide bonding, and hybrid bonding. The interconnect 3170 may include a plurality of bonding pads 3172, which are used to electrically couple the interconnects 3170 formed on the IC wafer 3110 and the sensor wafer 3130. The bonding pads 3172 may be copper pillars or pads and may provide mechanical coupling between the IC wafer 3110 and the sensor wafer 3130.

[0394] although Fig.31A and 31B A single ToF pixel 3140 is shown, but in general, the receiver unit 3100 may include an array of ToF pixels 3140 connected to an array of pixel phototransistors 3120 .

[0395] although Fig.31A and31B The ToF pixel 3140 is shown to have two switches 3150 and 3160 and is coupled to two capacitors 3132 and 3112, but in general, the ToF pixel 3140 may include three or more switches and be electrically coupled to three or more capacitors.

[0396] Fig.31C An example architecture diagram of a ToF receiver unit 3180 with increased capacitance values ​​is shown. The ToF receiver unit 3180 is similar to Fig.31A The ToF receiver unit 3100 is shown, but the difference is that the first capacitor 3112 and the second capacitor 3132 are now directly electrically coupled to the corresponding load-bearing readout terminals 3162 and 3152, without the presence of an intervening transistor.

[0397] Fig.31D An example architecture diagram showing a ToF receiver unit 3182 with increased capacitance values. The ToF receiver unit 3180 is similar to Fig.31A 3100, but the difference is that the first capacitor 3132 is divided into first capacitors 3133 and 3134, and the second capacitor 3112 is divided into second capacitors 3113 and 3114. The second capacitors 3113 and 3133 are located on the sensor wafer 3130, and the capacitors 3114 and 3134 are located on the IC wafer 3110. The first capacitors 3113 and 3114 are connected in parallel to achieve a similar Fig.31A The first capacitors 3113 and 3114 are electrically coupled to the readout transistor 3124 and are configured to charge or discharge according to the carriers collected by the readout terminal 3162. Similarly, the second capacitors 3133 and 3134 are connected in parallel to achieve a similar Fig.31A The second capacitors 3133 and 3134 are electrically coupled to the readout transistor 3122 and are configured to be charged or discharged according to the carriers collected by the readout terminal 3152.

[0398] Generally, the IC wafer 3110 and the sensor wafer 3130 are manufactured separately. For example, the two wafers 3110 and 3130 may be manufactured in different foundries and / or at different points in time using different process technologies, different process nodes, which may affect the capacitance value of the capacitors manufactured on the wafers 3110 or 3130 due to limited manufacturing technology tolerances and variability. By dividing the first capacitor connected to the readout terminal 3162 into a sub-capacitor 3113 located on the sensor wafer 3130 and a sub-capacitor 3114 located on the IC wafer, and similarly dividing the second capacitor associated with the readout terminal 3152 into sub-capacitors 3133 and 3134, any change in the capacitor of one wafer will have an equal impact on the overall first or second capacitance value, thereby helping to reduce or eliminate potential imbalances in the two capacitance values ​​caused by any variability or mismatch between the IC wafer 3110 and the sensor wafer 3130.

[0399] Fig.31E An example architecture diagram showing a ToF receiver unit 3184 with increased capacitance values. The ToF receiver unit 3184 is similar to Fig.31D A ToF receiver unit 3182 is shown, but the difference is that the first capacitors 3113 and 3114, and the second capacitors 3133 and 3134 are now directly electrically coupled to the corresponding carrier readout terminals 3162 and 3152, without the presence of intervening transistors.

[0400] Fig.31F An example of an architecture diagram showing a ToF receiver unit 3186 with increased capacitance values. The ToF receiver unit 3186 is similar to Fig.31A 3100, but the difference is that the pixel transistor 3120 is moved from the IC wafer 3110 to the sensor wafer 3130. In some cases, the pixel transistor 3120 can be moved to an unoccupied space in the sensor wafer 3130 to surround the ToF pixel 3140. Such a configuration of the pixel transistor 3120 can improve the performance of the receiver unit 3186, and / or free up space on the IC wafer 3110 for other components of the receiver unit 3186, such as additional capacitors, memory, amplifiers, or analog-to-digital converters.

[0401] Figure 31G An architectural diagram showing an example of a ToF receiver unit 3188 with increased capacitance values. The ToF receiver unit 3188 is similar to Fig.31C3180 is shown, but the difference is that the pixel transistor 3120 is moved from the IC wafer 3110 to the sensor wafer 3130. In some cases, the pixel transistor 3120 can be moved to an unoccupied space in the sensor wafer 3130 to surround the ToF pixel 3140. Such a configuration of the pixel transistor 3120 can improve the performance of the receiver unit 3188, and / or free up space on the IC wafer 3110 for other components of the receiver unit 3188, such as additional capacitors, memory, amplifiers, or analog-to-digital converters.

[0402] Fig.31H An example architecture diagram of a ToF receiver unit 3190 with increased capacitance values ​​is shown. The ToF receiver unit 3190 is similar to Fig.31D 3182 is shown, but the difference is that the pixel transistor 3120 is moved from the IC wafer 3110 to the sensor wafer 3130. In some cases, the pixel transistor 3120 can be moved to an unoccupied space in the sensor wafer 3130 to surround the ToF pixel 3140. Such a configuration of the pixel transistor 3120 can improve the performance of the receiver unit 3190, and / or free up space on the IC wafer 3110 for other components of the receiver unit 3190, such as additional capacitors, memory, amplifiers, or analog-to-digital converters.

[0403] Fig.31I An example architecture diagram of a ToF receiver unit 3192 with increased capacitance values ​​is shown. The ToF receiver unit 3192 is similar to Fig.31E 3184 is shown, but the difference is that the pixel transistor 3120 is moved from the IC wafer 3110 to the sensor wafer 3130. In some cases, the pixel transistor 3120 can be moved to an unoccupied space in the sensor wafer 3130 to surround the ToF pixel 3140. Such a configuration of the pixel transistor 3120 can improve the performance of the receiver unit 3192, and / or free up space on the IC wafer 3110 for other components of the receiver unit 3192, such as additional capacitors, memory, amplifiers, or analog-to-digital converters.

[0404] Generally speaking, Fig.31A and Fig.31B The pixel transistor 3120 of the ToF receiver unit 3100 in Fig.31D The pixel transistor 3120 of the ToF receiver unit 3182 in Fig.31F The pixel transistor 3120 of the ToF receiver unit 3186 in Fig.31HThe pixel transistor 3120 of the ToF receiver unit 3190 can be controlled to select an appropriate overall capacitance value for a target integration time while minimizing noise from the back-end amplifier array 3020 and the analog-to-digital converter 3030.

[0405] Fig.32 A block diagram showing an example of a receiver unit 3200 for ToF detection is shown. The ToF receiver unit 3200 is similar to Fig.30 The receiver unit 3000 is shown, but further includes a memory module 3210 and a digital signal processing (DSP) module 3220. The memory module 3210 is electrically coupled to the digital output 3040 of the analog-to-digital converter 3030 and the input of the DSP module 3220. The DSP module 3220 outputs digital processed data as a DSP output 3230.

[0406] The memory module 3210 is configured to store digital outputs 3040 of the analog-to-digital converter 3030 corresponding to amplified electrical signals from the pixel array 3010. The memory module 3210 can store a plurality of digital outputs corresponding to subframes generated by a large number of input optical signals, buffer the subframes before output, or further process the digital outputs 3040. For example, the receiver unit 3200 can generate data outputs at a rate that is higher than the data conversion rate of a system receiving the DSP outputs 3230. Such an increased data generation rate can be achieved, for example, due to a large amount of optical signals or a burst-mode transmission of ToF image frames. Under such conditions, the memory module 3210 can store additional data when the receiver unit 3200 transmits the DSP outputs 3230 to the data receiving system.

[0407] The DSP module 3220 is configured to digitally process the digital data stored in the memory module 3210. The DSP module 3220 can be configured to perform various arithmetic operations, Boolean operations, or special digital operations, such as Fast Fourier Transform (FFT), on the data received from the memory module 3210. For example, the DSP module 3220 can process multiple subframes stored in the memory module 3210 into a single complete frame, or a portion of a complete frame containing a region of interest to output to a data receiving system. By processing multiple subframes, generating a single complete frame or a portion of a complete frame containing a region of interest, and outputting the complete frame or a portion of a single complete frame containing a region of interest to the data receiving system, the overall external data production of the receiver unit 3200 can be reduced. In some embodiments, the overall external data production of the receiver unit 3200 can be reduced by a factor corresponding to the number of subframes. In another example, the DSP module 3220 can process the data stored in the memory module 3210 to determine and filter the depth information from the ToF measurement.

[0408] The memory module 3210 and the DSP module 3220 may be implemented separately or in combination, for example. Figures 31A-31I The described ToF receiver with increased capacitance value is implemented in combination. The combination of memory module 3210 and DSP module 3220 and the increased capacitance value can reduce the amount of external data generated and reduce the number of subframes generated. Reducing the number of subframes can reduce the storage capacity required by memory module 3210. In addition, reducing the number of subframes can reduce the number of processing operations performed by DSP module 3220. Such reduction in the number of subframes and the corresponding reduction in memory capacity and or program operations can reduce the energy consumption of the ToF receiver unit.

[0409] In some embodiments, the DSP module 3220 may include Figure 5A The processing unit 506 is shown. The DSP module 3220 can be implemented, for example, with a general-purpose processor or an application-specific integrated circuit.

[0410] Fig.33A A cross-sectional view of an example of a receiver unit 3300 for ToF detection is shown. The receiver unit 3300 includes an IC wafer 3110 and a sensor wafer 3130. The IC wafer 3110 includes a pixel transistor array 3320, an amplifier array 3020, and an analog-to-digital converter 3030. The sensor wafer 3130 includes a ToF pixel array 3010 and a memory module 3210. The pixel transistor array 3320 is Fig.31AArray of pixel transistors 3120 described in . IC wafer 3110 and sensor wafer 3130 are bonded by wafer bonding. Interconnects 3170 and bonding pads 3172 electrically couple different elements of receiver unit 3300.

[0411] The memory module 3210 may be distributed in the unoccupied space around the ToF pixel array 3010 in the sensor wafer 3130. For example, in the receiver unit 3300 of the BSI structure, the space on the amplifier array 3020 and the analog-to-digital converter 3030 in the sensor wafer 3130 is unoccupied. By configuring the memory module 3210 in the unoccupied space above the amplifier array 3020 and the analog-to-digital converter 3030, the receiver unit 3300 may improve performance without increasing the size of the receiver unit 3300.

[0412] The memory module 3210 can be implemented in different structures and using different memory technologies. Example memory technologies include static random access memory (SRAM), dynamic random access memory (DRAM), flash memory, resistive RAM (ReRAM), magnetic RAM (MRAM), phase change RAM (PRAM), and ferroelectric RAM (FeRAM). Different memory technologies generally share a common architecture of a one-bit storage element coupled to a read / write transistor. For example, a DRAM bit includes a capacitor for storing a charge associated with 1 bit, and a transistor for reading from or writing to the capacitor. In other examples, an SRAM bit includes a flip-flop for storing a charge associated with 1 bit, and two transistors for reading from or writing to the flip-flop. Similarly, ReRAM has a variable resistance storage element, MRAM has a magnetic storage element, PRAM has a phase change storage element, and FeRAM has a ferroelectric storage element for storing 1 bit. In some embodiments, two or more memory technologies can be combined and work in conjunction with each other. Fig.33A In the illustrated embodiment, memory module 3210 includes read / write transistors and associated bit storage elements. In some embodiments, additional processing circuits may be included to further expand the functionality of memory module 3210. For example, a digital adder may be included to further process the bits stored by the memory.

[0413] Fig.33BA cross-sectional view of an example of a receiver unit 3330 for ToF detection is shown. The receiver unit 3330 is similar to Fig.33A Receiver unit 3300 is shown, but the difference is that memory module 3210 is replaced by a distributed memory module 3340. Distributed memory module 3340 includes a storage element sub-region 3342 and a read / write transistor sub-region 3344. Storage sub-region 3342 is located on sensor wafer 3130, and transistor sub-region is located on IC wafer 3110. In some embodiments, additional processing circuits, such as digital adders, can further process the bits stored by the memory.

[0414] The storage element 3342 is typically manufactured using specialized techniques and / or materials. For example, capacitors for DRAM are typically formed in silicon wafers using specialized techniques, such as deep trench etching, which may be incompatible with the manufacturing techniques used to manufacture the IC wafer 3110. In addition, the manufacturing process used to manufacture the IC wafer 3110 can be better optimized for manufacturing transistors, such as the read / write transistor 3344. For example, the manufacturing process used for the IC wafer 3110 can have a more advanced process node with a smaller minimum feature size than the sensor wafer 3130. Therefore, allowing the performance of the two sub-regions 3342 and 3344 to be optimized separately, separating the manufacturing of the storage element 3342 and the read / write transistor 3344 can improve the performance of the memory module 3340 and reduce the manufacturing complexity of the sensor wafer 3130 and the IC wafer 3110.

[0415] Fig.33C A cross-sectional view of an example of a receiver unit 3350 for ToF detection is shown. The receiver unit 3350 is similar to Fig.33A 3300 is shown, but the difference is that the memory module 3210 is now located in the IC wafer 3110. For some memory technologies, such as SRAM or flash memory, for a given technology node, the complete memory module can be provided by the CMOS foundry or a third-party supplier as an intellectual property (IP) core. Incorporating these IP cores to implement the memory module 3210 of the IC wafer 3110 can reduce research and development work.

[0416] Fig.33D A cross-sectional view of an exemplary receiver unit 3360 is shown, which is used for ToF detection. The receiver unit 3360 is similar to Fig.33CThe receiver unit 3350 is shown in FIG. 31 , but the difference is that the pixel transistor array 3320 is now located on the sensor wafer 3130. Configuring the pixel transistor array 3320 on the sensor wafer 3130 can improve the space utilization of the IC wafer 3110. For example, since the amplifier array 3020 and the analog-to-digital converter 3030 are located below the pixel transistor array 3320, the memory module 3210 can be set in an unoccupied position below the ToF pixel array 3010 to simplify the electrical connection between these components.

[0417] Fig.33E A cross-sectional view of an exemplary receiver unit 3370 is shown, which is used for ToF detection. The receiver unit 3370 is similar to Fig.33D The receiver unit 3360 is shown, but the difference is that the memory module 3210 is now located on the sensor wafer 3130.

[0418] Although the ToF pixel array 3010, the pixel transistor array 3320, the memory module 3210, the analog-to-digital converter 3030, and the amplifier array 3020 are shown as exemplary blocks for purposes of illustration, generally portions of these elements may span across the respective wafers 3110 and 3130. For example, the ToF pixels 3140 of the ToF pixel array 3010 and the pixel transistors 3120 of the pixel transistor array 3320 may be distributed on the chips 3110 and / or 3130, and the memory bits of the memory module 3210 or 3340 may be distributed in the spaces on the chips 3110 and / or 3130 that are not occupied by the ToF pixel array 3010 or the pixel transistor array 3320.

[0419] In general, additional electrical and optical components can be added to Figures 33A-33E Receiver unit described. Examples of electrical components include resistors, inductors, data processing circuits (e.g., processors, FPGAs, ASICs), bias circuits (e.g., for providing bias to control terminals 3154 and 3164 of sensor wafer 3130 and / or ToF pixel 3140), and light source driver circuits (e.g., for providing electrical pulses to transmitter unit 502 to generate light pulses). Examples of optical components include anti-reflective coatings (ARCs), microlenses, bandpass filters, and reflectors. Examples of microlenses include microsphere lenses, Fresnel zone plates, and integrated silicon microlenses.

[0420] In general, there may be an intermediate layer between the IC wafer 3110 and the sensor wafer 3130. The intermediate layer may provide various benefits, such as improved electrical coupling between the two chips, improved bonding quality and yield of the two chips, and improved optical performance of the receiver unit. The intermediate layer may be formed of various materials, such as dielectrics, polymers, and optical refractive index matching materials.

[0421] Although already Figures 33A-33E Bidirectional bonding is described with respect to IC wafer 3110 and sensor wafer 3130. However, in general, a receiver unit may be formed by bonding three or more chips. For example, additional IC wafers may be bonded to integrate additional capacitors to further increase the capacitance associated with pixel array 3010. In another example, additional IC wafers may be bonded to integrate additional memory elements to further increase the storage capacity of memory module 3210.

[0422] Generally speaking, Figures 33A-33E The sensor wafer 3130 of the receiver unit described in can be a front-illuminated sensor wafer or a back-illuminated sensor wafer.

[0423] In general, the sensor wafer 3130 , the ToF pixels 3140 , or both may be formed of Group III-V compound semiconductor materials, Group IV alloy semiconductor materials, or a combination thereof.

[0424] Fig.34 A cross-sectional view of an exemplary receiver unit 3400 is shown, which is used for ToF detection. The sensor wafer 3130 includes a ToF pixel 3140 and a back-end layer 3136. The IC wafer 3110 includes a pixel transistor 3120 and a back-end layer 3116. The back-end layers 3116 and 3136 include interconnects 3170 and bonding pads 3172, which are formed on the front sides of the wafers 3110 and 3130, respectively. The surfaces of the back-end layers 3116 and 3136 include dielectric surfaces and metal surfaces of the bonding pads 3172. Prior to bonding, the sensor wafer is first flipped so that the bonding pads 3172 of the sensor wafer 3130 face the bonding pads 3172 of the IC wafer 3110. The two wafers are contacted in a controlled manner, which may involve controlling the strength of the force, the temperature, and the formation environment. The bonding of the dielectric surface and the metal surface allows for hybrid bonding of the wafers 3110 and 3130, resulting in electrical and mechanical coupling between the two wafers.

[0425] In wafer bonding, the inversion of the sensor wafer 3130 allows the sensor wafer 3130 to receive the optical signal 3410 through its back side. The sensor wafer 3130 may be a silicon wafer, which is transparent to infrared wavelengths (e.g., >1.1 μm). As a result, the infrared optical signal 3410 may reach the ToF pixel 3140 through the back side of the sensor wafer 3130. This configuration is referred to as a backside illuminated (BSI) sensor.

[0426] In some embodiments, the back-end layer 3136 of the sensor wafer 3130 may include a reflector 3420. The reflector 3420 is located below the light absorption region of the ToF pixel 3140. Therefore, any light that is not absorbed by the ToF pixel 3140 will be reflected by the reflector 3420 when passing through the pixel 3140, and reflected back to the ToF pixel 3140, and further absorbed by the ToF pixel 3140. The reflector 3420 can be, for example, a metal mirror, a dielectric mirror, or a distributed Bragg reflector. The reflector 3420 can be a combination of a dielectric layer (e.g., silicon oxide or silicon nitride) and a metal layer. In some embodiments, the reflector 3420 can be a concave mirror configured to reflect light toward a focal point located within the ToF pixel 3140.

[0427] In some embodiments, the sensor wafer 3130 may include a partial reflector 3422. The partial reflector 3422 is formed on the back side of the sensor wafer 3130 and allows a portion of the light to pass through the ToF pixel 3140. The partial reflector 3422 may generate destructive interference at the interface between it and the air, so that the light transmitted to the partial reflector 3422 after being reflected by the reflector 3420 is further reflected back to the ToF pixel 3140. When such conditions are met, the cooperation of the partial reflector 3422 and the reflector 3420 forms a resonant cavity to allow multiple reflected lights to pass back and forth between the partial reflector 3422 and the reflector 3420. The formed resonant cavity can improve the detection efficiency of the ToF pixel 3140 at the resonant wavelength of the resonant cavity. The partial reflector 3422 may be, for example, a dielectric reflector or a distributed Bragg reflector. The partial reflector 3422 may have a transmittance substantially equal to the round-trip attenuation of the light passing through the ToF pixel 3140 and reflected by the reflector 3420.

[0428] In some embodiments, the back-end layer 3116 of the IC wafer 3110 may include a reflector 3424. After the sensor wafer 3130 and the IC wafer 3110 are combined, the reflector 3424 is located below the light absorption area of ​​the ToF pixel 3140. Thus, any light that passes through the ToF pixel 3140 but is not absorbed by the ToF pixel 3040 will be reflected by the reflector 3424 and returned to the ToF pixel 3140, and further absorbed by the ToF pixel 3140. The reflector 3424 can be, for example, a metal reflector, a dielectric reflector, or a distributed Bragg reflector. The reflector 3424 can be a combination of a dielectric layer (e.g., silicon oxide or silicon nitride) and a metal layer. In some embodiments, the reflector 3424 can be a concave mirror configured to reflect light to a focal point within the ToF pixel 3140.

[0429] In the previous paragraphs, methods for increasing integration time by increasing capacitance have been described. An important consideration when determining the integration time is the dark current of the photodetector, which is the current that flows in the absence of a light signal and ambient light. In general, the signal-to-noise ratio (SNR) of optical measurements made with a photodetector, such as ToF measurements made with a switched photodetector, is negatively impacted by the presence of dark current. For example, the SNR of optical measurements made with a photodetector is linearly proportional to the integration time. In addition, the integration time for a given capacitance may be limited by the dark current because the dark current allows the capacitor to discharge continuously even in the absence of a light signal and ambient light.

[0430] The dark current of a photodetector is generally an exponential function of the reverse bias voltage established across the cathode and anode of the photodetector. Therefore, reducing the reverse bias voltage by a controlled method while maintaining the overall operation of the photodetector can improve the SNR performance of the photodetector.

[0431] Fig.35 A schematic diagram of a circuit 3500 for operating a ToF pixel is shown. The circuit 3500 includes a first readout subcircuit 3510 and a second readout subcircuit 3530 coupled to a switch photodetector 3550. The first readout subcircuit 3510 includes a first first MOSFET transistor 3512 and a second MOSFET transistor 3520. The second readout subcircuit 3530 includes a third MOSFET transistor 3532 and a fourth MOSFET transistor 3540. The first readout subcircuit 3510 is coupled to a first source follower circuit 3560, and the second readout subcircuit 3530 is coupled to a second source follower circuit 3570. The first readout subcircuit 3510 and the first source follower circuit 3560 may be referred to as a first readout circuit, and the second readout subcircuit 3530 and the second source follower circuit 3570 may be referred to as a second readout circuit.

[0432] The switch photodetector 3550 includes a body 3551, a first readout terminal 3552 and a second readout terminal 3554. The switch photodetector 3550 can be implemented as any of the aforementioned switch photodetectors, for example Figure 1A The switch photodetector 100 is shown. The body 3551 can be similar to the light absorbing layer 106 or the substrates 202, 302 and 402, and is doped with a p-type dopant. The first readout terminal 3552 and the second readout terminal 3554 can be n-doped regions, for example, similar to Figure 1A n-type doped regions 126 and 136. According to the control operation of the switch photodetector 3550, the photocurrent generated by the switch photodetector 3550 can be collected by the first readout terminal 3552 or the second readout terminal 3554.

[0433] Each MOSFET transistor 3512, 3520, 3532 and 3540 includes a source terminal, a drain terminal and a gate terminal. The source terminal and the drain terminal may be the same in the underlying structure, but the difference lies in the direction of flow of current through the transistor. For example, for an N-type MOSFET (NMOS) transistor having a P-type channel region, the current may flow from the drain terminal to the source terminal via the channel region; and for a P-type MOSFET (PMOS) transistor having an N-type channel region, the current may flow from the source terminal to the drain terminal via the channel region. Since the source and drain are named based on convention, and since the underlying structure may be similar or identical, the source terminal and the drain terminal may be referred to as the first channel terminal and the second channel terminal when describing the connectivity between the MOSFET transistor and other circuits.

[0434] The gate terminal controls the current flowing through the source and drain terminals. For example, a control voltage greater than the critical voltage Vth allows current to flow through the source and drain terminals. Depending on the voltage of the source and drain terminals relative to the gate terminal, the operating mode of the MOSFET transistor can operate in the saturation region or the triode region. In the saturation region, the current flowing through the source and drain terminals will not produce drastic changes due to the voltage difference between the source and the gate (i.e., the output impedance of the transistor is high). In the triode region, the current flowing through the source and drain terminals is almost linearly proportional to the voltage difference between the source and the drain (i.e., the operation of the transistor is similar to a resistor). A control voltage less than the critical voltage can reduce the flow of current through the source and drain terminals. For example, as the control voltage is reduced below the critical voltage, the current can be reduced exponentially. This operating mode of the MOSFET transistor can be referred to as operating in a subthreshold region.

[0435] For the purpose of illustration, the circuit 3500 is implemented with an N-type MOSFET transistor. In the first readout subcircuit 3510, the source terminal of the first MOSFET transistor 3512 is coupled to the first readout terminal 3552, the drain terminal of the first MOSFET transistor 3512 is coupled to the source terminal of the second MOSFET transistor 3520, and this coupling node can be referred to as the first output node 3515 of the first readout subcircuit 3510. A capacitor can be coupled to the first output node 3515, and the capacitor can be similar to Fig.31A and 31B 3112 and 3132 are shown. The drain terminal of the second MOSFET transistor 3520 is coupled to the first supply node 3508. Similarly, in the second readout subcircuit 3530, the source terminal of the third MOSFET transistor 3532 is coupled to the second readout terminal 3554. The drain terminal of the third MOSFET transistor 3532 is coupled to the source terminal of the fourth MOSFET 3540, and this coupling node may be referred to as the second output node 3535 of the second readout subcircuit 3530. A capacitor may be coupled to the second output node 3535, and the capacitor may be similar to Fig.31A and 31B Capacitors 3112 and 3132 are shown. The drain terminal of the fourth MOSFET transistor 3540 is connected to the first supply node 3508.

[0436] The first supply node 3508 provides a first supply voltage to the first and second readout sub-circuits 3510 and 3530. The second supply node 3502 provides a second supply voltage to the first and second source follower circuits 3560 and 3570. Depending on various factors, including the particular process node, circuit design, characteristics of the switching photodetector 3550, a reset voltage of the capacitor coupled to the first output node 3515, and a charge-to-voltage conversion gain, one or more supply voltage sources may provide suitable first and second supply voltages to the first and second supply nodes 3508 and 3502. The first supply node 3508 may be referred to as V U node, and V U The first supply voltage of the node may be, for example, a user-defined voltage generated by an on-chip integrated circuit. The second supply node 3502 may be referred to as V E node, and V E The second supply voltage of the node may be, for example, an externally defined voltage generated by an off-chip power supply.

[0437] During operation of the ToF pixel, the first output node 3515 and the second output node 3535 are charged to a preset voltage through the second and fourth MOSFET transistors 3520 and 3540. For example, by applying a second control voltage 3506 (Vc2), the second and fourth MOSFET transistors 3520 and 3540 are operated in a saturation region or a triode region, and current can flow from the first supply node 3508 to the corresponding output nodes 3515 and 3535 and charge the nodes to a preset voltage. A second control voltage source 3507 coupled to the gate terminals of the second and fourth transistors 3520 and 3540 can be used to apply the second control voltage 3506. The second control voltage 3506 can be controlled to change the preset voltage (e.g., set to a supply voltage or a portion of a supply voltage) to charge the output nodes 3515 and 3535. Once the output nodes 3515 and 3535 are charged, the second control voltage 3506 may be set (e.g., to 0V) to turn off the second and fourth MOSFET transistors 3520 and 3540, which decouple the output nodes 3515 and 3535 from the first supply node 3508. The charging operation described may be referred to as a reset operation of the switching photodetector 3550, and the second and fourth MOSFETs 3520 and 3540 may be referred to as reset transistors. The reset operation may be a step within the readout step of the ToF pixel.

[0438] Once charging is completed, the electrical signal generated by the switching photodetector 3550 can begin to integrate. The first control voltage 3504 (Vc1) coupled to the corresponding gate terminals by the first control voltage source 3505 coupled to the gate terminals of the MOSFETs 3512 and 3532 can be controlled to start or stop integration. For example, the first control voltage 3504 can be set by the voltage source 3505 to allow the first and third MOSFETs 3512 and 3532 to operate in the triode region. When operating in the triode region, the photocurrent generated by the switching photodetector 3550 can flow through the drain and source terminals of the MOSFETs 3512 and 3532 and through the readout terminals 3552 and 3554. By discharging the corresponding capacitors charged to the preset voltage during the reset operation, the photocurrent flows through the output terminals 3552 and 3554, and such flow can be integrated at the output nodes 3515 and 3535.

[0439] The operation of the first and third MOSFET transistors 3512 and 3532 in the triode region is similar to replacing the first and third MOSFET transistors 3512 and 3532 with corresponding resistors (equivalent resistors) to couple the output nodes 3515 and 3535 to the corresponding readout terminals 3552 and 3554. The resistance of these effective resistors is typically of moderate value (e.g., 10 ohms to 10,000 ohms) that does not experience a significant voltage drop in response to the current flowing through the photodetector. For example, the photodetector current, which can be a combination of photocurrent and dark current, is typically a small current ranging from pA to μA, and the voltage drop across the resistor is also relatively small (e.g., ranging from nV to mV). As a result, the voltage of the readout terminals 3552 and 3554 is similar to the voltage of the output nodes 3515 and 3535 with a small voltage drop. When the output nodes 3515 and 3535 are charged to a preset voltage close to the first supply voltage of the first supply node 3508, the readout terminals 3552 and 3554 may experience a similar voltage when the reverse bias voltage across the junction of the photodetector 3550 is greater than the minimum reverse bias voltage required for proper operation of the photodetector 3550. This excessive reverse bias causes an increase in dark current, which may reduce the SNR of the output produced by the circuit 3500.

[0440] Photodetectors of various designs and material compositions may benefit from reverse bias control. In the materials used to form the absorption region of the photodetector, germanium may be more susceptible to dark current than silicon due to its higher material defect density than silicon, which is usually associated with germanium absorption reg...

Claims

1. A circuit comprising: a photodetector comprising a first readout terminal and a second readout terminal, the second readout terminal being different from the first readout terminal; a first readout circuit coupled to the first readout terminal and configured to output a first readout voltage; a second readout circuit coupled to the second readout terminal and configured to output a second readout voltage; and A common mode analog-to-digital converter, comprising: a first input terminal coupled to a first voltage source; a second input terminal coupled to a common mode generator, the common mode generator being configured to receive the first read voltage and the second read voltage and generate a common mode voltage between the first read voltage and the second read voltage; and A first output terminal is configured to output a first output signal corresponding to a current amount generated by the photodetector.

2. The circuit of claim 1 further comprising a differential mode analog-to-digital converter, the differential mode analog-to-digital converter comprising: a third input terminal coupled to the first readout circuit and configured to receive the first readout voltage; a fourth input terminal coupled to the second readout circuit and configured to receive the second readout voltage; and a second output terminal configured to output a second output signal corresponding to a time difference ranging information generated by the light detector, in, The circuit is operated to synchronously generate the first output signal and the second output signal.

3. The circuit of claim 1 , wherein the first readout circuit comprises: a first capacitor coupled to the first readout terminal; and a first source follower circuit coupled to the first capacitor and configured to generate the first read voltage; and The second readout circuit comprises: a second capacitor coupled to the second readout terminal; and A second source follower circuit is coupled to the second capacitor and is configured to generate the second readout voltage.

4. The circuit of claim 1 , wherein the first readout circuit comprises: A first MOSFET transistor comprising: a first gate terminal coupled to a first control voltage source; a first channel terminal; and a second channel terminal coupled to the first readout terminal of the photodetector; A second MOSFET transistor comprising: a second gate terminal coupled to a second control voltage source; a third channel terminal coupled to a supply voltage node; and a fourth channel terminal coupled to the first channel terminal; a first capacitor coupled to the first channel terminal of the first MOSFET transistor; as well as a first source follower circuit coupled to the first capacitor and configured to generate the first read voltage, and The second readout circuit comprises: A third MOSFET transistor comprising: a third gate terminal coupled to the first control voltage source; a fifth channel terminal; and a sixth channel terminal coupled to the second readout terminal of the photodetector; a fourth MOSFET transistor, comprising: a fourth gate terminal coupled to the second control voltage source; a seventh channel terminal coupled to the supply voltage node; as well as an eighth channel terminal coupled to the fifth channel terminal; a second capacitor coupled to the fifth channel terminal of the third MOSFET transistor; as well as A second source follower circuit is coupled to the second capacitor and is configured to generate the second readout voltage.

5. The circuit of claim 1, wherein the first voltage source comprises a third source follower circuit.

6. A method for measuring a performance characteristic of a time difference ranging detection device, the time difference ranging detection device comprising a photodetector, the photodetector having a first readout terminal and a second readout terminal, the first readout terminal coupled to a first readout circuit and configured to output a first readout voltage, the second readout terminal coupled to a second readout circuit and configured to output a second readout voltage, the method comprising: In the absence of ambient light and a time-of-flight optical signal, measuring a dark current of the photodetector by measuring a common-mode output signal between the first readout voltage and the second readout voltage; Determining whether the dark current of the photodetector is greater than a first value; and When the dark current of the photodetector is greater than the first value, it is determined that the time-of-flight detection device does not meet a performance specification.

7. The method of claim 6, wherein measuring the dark current of the photodetector comprises: Measuring the common mode output signal between the first readout voltage and the second readout voltage one or more times by a 1-bit analog-to-digital converter or a multi-bit analog-to-digital converter in the absence of ambient light and the time-of-flight optical signal; and The dark current is determined based on the one or more measured common mode output signals.

8. The method of claim 7, wherein the one or more measurements are multiple measurements; and Each of the multiple measurements corresponds to a different integration time or a different replica voltage input to the 1-bit ADC or the multi-bit ADC.

9. The method of claim 6, further comprising: When the time difference ranging optical signal is present, measuring a differential mode output signal between the first readout voltage and the second readout voltage to measure a demodulation contrast of the time difference ranging detection device; Determine whether the demodulation contrast of the time difference ranging detection device is lower than a second value; as well as When the demodulation contrast of the time difference ranging detection device is lower than the second value, it is determined that the time difference ranging detection device does not meet the performance specification.

Images