Optical pixel with optical concentrator and full-depth deep isolation trench for improved low-light performance
The pixel architecture with a large aperture and deep trench isolation, coupled with light collection structures, addresses high dark current and noise in digital night vision, achieving superior low-light imaging performance.
Patent Information
- Application Number
- JP2024068176
- Authority / Receiving Office
- JP · JP
- Patent Type
- Applications
- Current Assignee / Owner
- Priority Date
- 2023-04-26
- Filing Date
- 2024-04-19
- Publication Date
- 2024-11-08
Smart Images

Figure 2024159599000001_ABST
Abstract
Description
[Technical field]
[0001] Field The disclosed technology relates generally to imaging applications and image sensor technology, and more particularly to pixel architectures that utilize light concentrators to reduce dark current. [Background technology]
[0002] background Light concentrators can be used in a variety of applications, including solar power generation, camera lens design, etc., to allow for efficient harvesting of available light. In the case of camera lens design, light concentrators can help reduce the exposure time needed to produce a sharper image by directing light to the camera sensor.
[0003] For night vision, analog image intensifier (I2) tube devices have become the mainstream technology. In the latest (Gen III) I2 tube devices, incident photons strike a GaAs photocathode, generating electrons that are accelerated towards a microchannel plate (MCP), which amplifies the electrons with a gain of up to 10,000 times. The amplified electrons are further accelerated towards a phosphor screen to form a low-light image scene. The excellent low-light image quality of I2 tubes can be achieved by the extremely low dark current due to the large band gap of the GaAs photocathode, 1.42 eV (compared to the band gap of silicon, 1.12 eV). The extremely low read noise of I2 tubes is due to the high gain provided by the MCP. I2 tubes can also have a high frame rate equivalent to 1000 Hz. This is associated with a P43 phosphor decay time of 1 ms. The use of GaAs material as the photocathode in I2 tubes can result in very low dark shot noise. However, the dominant noise is the photon shot noise. I2 tubes tend to be bulky, can require special manufacturing, and can have limited mean time to failure (MTTF).
[0004] Typically, the signal-to-noise ratio (SNR) of an image is used to quantify image quality in low light. A simplified formula for SNR is given by Equation 1 below:
number
[0005] In formula (1), S is the electronic signal, S dark is the dark signal of electrons, F is the additive noise factor, n read is the input-referred read noise floor of the electronic readout circuitry and G is the gain of the pixel / sensor.
[0006] In recent years, several digital night vision solutions have been developed to take advantage of standard silicon microelectronic processing techniques. Examples include electron multiplying CCDs (EMCCDs), single photon avalanche diodes (SPADs), and CMOS sensors. The native read noise of EMCCDs is typically as high as between 50e- and 100e- input-referred. EMCCDs can achieve very low equivalent read noise with high gain (G>>1), but suffer from high dark current and have an excess read noise factor F of about 1.4. To reduce the effects of dark current, the operating temperature of a typical EMCCD is set at -40 degrees Celsius or lower, which may be impractical for many applications.
[0007] Although SPAD sensors can achieve true read noise-free operation (G is infinite), their performance is typically limited by a higher dark count rate (DCR). State-of-the-art CMOS pixels can have read noise below 1e- and low dark current that is acceptable for many applications. For CMOS pixels, F is 1 and G is 1 in equation (1). However, in extreme low light conditions (such as moonless night sky, overcast and cloudy night sky), the illumination level can be below 0.1 mlux. Under such conditions, only I2 tubes can provide acceptable performance.
[0008] Various combinations of analog and digital binning may be used to improve low light sensitivity, but these methods may also increase dark current and read noise. In analog binning, for example, signals from two or more adjacent pixels may be combined in the charge domain before being sensed by the sense node. Due to the circuitry involved in supporting the binning operation, the capacitance of the effective sensing node (i.e., floating diffusion FD) may increase, which may cause an increase in read noise. In digital binning, signals from two or more adjacent pixels may be combined in the digital domain, which may increase the read noise according to the square root of the binned pixel count. In both analog and digital binning, the dark current typically increases proportionally to the binned pixel count, increasing the contribution of dark shot noise.
[0009] For sensors with large pixel sensing areas, the corresponding dark current may be high due to the large silicon device area and the interface with shallow trench isolation (STI). In addition, for typical 4T CMOS pixels, the large pixel sensing area tends to have image retention issues due to the long travel distance of the charge collected in the photodiode. Image retention typically appears as fixed pattern noise (FPN) on the image and can significantly degrade image quality for night vision applications. Despite recent advances in low light sensors, there is still a need for improved digital devices with low light performance that can match or exceed that of I2 tubes. [Prior art documents] [Non-patent literature]
[0010] [Non-Patent Document 1] S. -G. Wuu, H. -L. Chen, H. -C. Chien, P. Enquist, RM Guidash and J. McCarten, “A Review of 3-Dimensional Wafer Level Stacked Backside Illuminated CMOS Image Sensor Process Technologies,” in IEEE Transactions on Electron Devices, vol. 69, no. 6, pp. 2766-2778, June 2022. [Non-Patent Document 2] Y. Oike, "Evolution of Image Sensor Architectures With Stacked Device Technologies," in IEEE Transactions on Electron Devices, vol. 69, no. 6, pp. 2757-2765, June 2022. [Non-Patent Document 3] J. Gambino, et al, “CMOS image sensor with high refractive index lightpipe”, IISW 2009. Summary of the Invention
[0011] The disclosed technology includes a pixel architecture for imaging devices with reduced dark current and improved signal-to-noise ratio. The pixel architecture includes a light-sensing pixel characterized by a light-receiving aperture having a first dimension D defined by a unit pixel pitch, a sensing region having a second dimension d smaller than the first dimension D of the unit pixel pitch, the sensing region being defined within the boundaries of a first full depth deep trench isolation (FDTI), and a light-collecting structure configured to receive light incident on the light-receiving aperture and to concentrate and direct the received light to the sensing region.
[0012] Certain exemplary embodiments of the disclosed technology include a night vision device with reduced dark current and improved signal-to-noise ratio. The night vision device includes an array of light-sensing pixels, each light-sensing pixel of the array characterized by a light-receiving aperture having a first dimension D defined by a unit pixel pitch, a sensing region having a second dimension d smaller than the first dimension D of the unit pixel pitch, the sensing region being defined within the boundaries of a first full depth deep trench isolation (FDTI), and a light-collecting structure configured to receive light incident on the light-receiving aperture and to concentrate and direct the received light to the sensing region.
[0013] A method for manufacturing an imaging device for reducing dark current and improving signal-to-noise ratio is disclosed. The method includes forming a pixel array, where each pixel of the pixel array is fabricated by forming a sensing region having a dimension d on a wafer, forming a full depth deep trench isolation (FDTI) to border the sensing region, forming a light collection structure, and forming a gapless microlens array on the pixel array, where each gapless microlens of the gapless microlens array defines a light-receiving aperture characterized by a dimension D greater than d and is configured to receive light incident on the light-receiving aperture and focus and direct the received light to the sensing region.
[0014] More specifically, the present invention provides: [1] a light receiving aperture having a first dimension D defined by a unit pixel pitch; a sensing area having a second dimension d smaller than the first dimension D of the unit pixel pitch, the sensing area being defined within a boundary of a first full depth deep trench isolation (FDTI); a light collecting structure configured to receive light incident on the light receiving aperture and to concentrate and direct the received light onto the sensing area; A light-sensing pixel, characterized by a pixel architecture for an imaging device having reduced dark current and improved signal-to-noise ratio, [2] The pixel architecture of [1], wherein the light-sensing pixel is a back-side illuminated (BSI) CMOS pixel or a front-side illuminated (FSI) CMOS pixel; [3] The pixel architecture of [1] , wherein the light collecting structure comprises a light pipe waveguide; [4] The pixel architecture of [1] , wherein the light focusing structure includes one or more of a binary optical lens and a diffraction grating-based lens; [5] The pixel architecture of [1] , wherein the light-collecting structures include one or more of a gapless microlens and an internal microlens; [6] The pixel architecture of [1] , wherein the sensing region comprises a photodiode, a photoconductor, a single-photon avalanche diode (SPAD), or a non-silicon-based detector made from one or more of InGaAs, InP, and Germanium; [7] The pixel architecture of [1], further comprising a second FDTI having a dimension substantially equal to D and surrounding the first FDTI; [8] The pixel architecture of [7], in which the second FDTI is filled to reduce or eliminate pixel crosstalk; [9] The pixel architecture of [1], further comprising an embedded texture on a silicon surface or embedded within said sensing area;
[10] The pixel architecture of [1] further comprising a metallic reflector structure;
[11] The pixel architecture of [1] further comprising a color filter;
[12] The pixel architecture of [1], where the ratio D / d is greater than or equal to 1.5;
[13] The pixel architecture of [1], where the ratio D / d is greater than or equal to 2.0;
[14] The pixel architecture of [1], where the ratio D / d is 5.0 or greater;
[15] An array of light-sensing pixels, each light-sensing pixel of the array comprising: a light receiving aperture having a first dimension D defined by a unit pixel pitch; a sensing area having a second dimension d smaller than the first dimension D of the unit pixel pitch, the sensing area being defined within a boundary of a first full depth deep trench isolation (FDTI); a light collecting structure configured to receive light incident on the light receiving aperture and to concentrate and direct the received light onto the sensing area; The array of light-sensing pixels, a night vision device having reduced dark current and improved signal-to-noise ratio,
[16]
[15] a night vision device according to
[15] , wherein each light sensing pixel of the array is a backside illuminated (BSI) CMOS pixel or a frontside illuminated (FSI) CMOS pixel;
[17] The light collecting structure is Light pipe waveguide, Gapless microlens, Internal microlenses, and Binary Optical Lens
[15] a night vision device, including one or more of the following:
[18]
[15] , wherein each light-sensing pixel of the array includes a second FDTI having a dimension substantially equal to D and surrounding the first FDTI, the second FDTI being metal filled to reduce or eliminate pixel crosstalk;
[19] Each light-sensing pixel of the array is an embedded texture on the silicon surface of said sensing area; A metal reflector structure, and Color Filters
[15] a night vision device, including one or more of the following:
[20]
[15] A night vision device having a ratio D / d of 1.5 or more; [twenty one] forming a pixel array, each pixel of the pixel array comprising: forming a sensing area having a dimension d on a wafer; forming a full depth deep trench isolation (FDTI) to border the sensing area; forming a light collecting structure; forming a gapless microlens array over the pixel array, each gapless microlens of the gapless microlens array defining a light-receiving aperture characterized by a dimension D greater than d and configured to receive light incident on the light-receiving aperture and to focus and direct the received light onto the sensing area; forming the pixel array, A method for manufacturing an imaging device having reduced dark current and improved signal to noise ratio, comprising: [twenty two]
[21] , wherein forming a gapless microlens array on the pixel array comprises forming a microlens array structure using photolithography and further applying a material reflow or material etching to the microlens array structure; [twenty three] forming the light collecting structure, Light pipe waveguide, Gapless microlens, Internal microlenses, and Binary Optical Lens
[21] the method of
[21] comprising forming one or more of: [twenty four] forming the sensing region an embedded texture on the silicon surface of the sensing area; and metal reflector structure
[21] The method of
[21] , comprising forming one or more of:
[0015] Other embodiments, features, and aspects of the disclosed technology are described in detail herein and are considered a part of the claimed disclosure, and can be understood with reference to the following detailed description, the accompanying drawings, and the claims. [Brief description of the drawings]
[0016] Reference is now made to the accompanying drawings and flow diagrams, which are not necessarily drawn to scale.
[0017] [Figure 1] FIG. 2 shows a concentrator that utilizes a Fresnel lens to focus light received from an aperture having a dimension D onto a smaller target having a dimension d; <Dである。 [Diagram 2] We demonstrate a theoretical upper limit on the number of photons per pixel capturing scene illumination of 0.1 mlux across visible and near-infrared wavelengths (400 nm - 1100 nm) captured at 30 frames per second using an F#1.4 lens, which corresponds to an ideal blackbody light source of 2856 K. [Diagram 3] Figure 3A shows an example cross-sectional view of a monochrome back-illuminated (BSI) CMOS pixel with a light pipe concentrator having a significantly reduced light sensing area according to an exemplary embodiment of the disclosed technology. Figure 3B shows an alternative wafer stacking configuration that may be utilized with the BSI CMOS pixel (such as the top of the BSI CMOS pixel shown in Figure 3A), which may include a logic carrier wafer bonded with a sensor wafer using wafer stacking. [Figure 4]1 illustrates an example of a top view of a BSI CMOS pixel having an optical pixel size D and a first full depth deep trench isolation (FDTI) surrounding a sensing area having a smaller dimension d. The optical pixel size D may correspond to the pixel pitch of an array of pixels according to an exemplary embodiment of the disclosed technology. [Diagram 5] 5 shows an example of a top view of a BSI CMOS pixel with similar features to that shown in FIG. 4 with the addition of a second FDTI (of dimension approximately D) bordering the optical pixel, according to an exemplary embodiment of the disclosed technique. [Figure 6] 1 shows an example cross-sectional view of a color backside illuminated (BSI) CMOS pixel with a light pipe concentrator having a significantly reduced light sensing area, in accordance with an exemplary embodiment of the disclosed technology. [Figure 7] 1 illustrates an example cross-sectional view of a color front-side illuminated (FSI) CMOS pixel with a light pipe concentrator having a significantly reduced light sensing area, in accordance with an exemplary embodiment of the disclosed technology. [Figure 8] 1 illustrates an example cross-sectional view of a monochrome back-illuminated (BSI) single-photon avalanche diode (SPAD) pixel with a light pipe concentrator having a significantly reduced light sensing area, in accordance with an exemplary embodiment of the disclosed technology. [Figure 9] 1 illustrates an example cross-sectional view of a color back-side illuminated (BSI) single-photon avalanche diode (SPAD) pixel with a light pipe concentrator having a significantly reduced light sensing area, in accordance with an exemplary embodiment of the disclosed technology. [Figure 10] 1 shows an example cross-sectional view of a color back-side illuminated (BSI) single-photon avalanche diode (SPAD) pixel with a significantly reduced light sensing area bordered by a metal-filled FDTI, a light pipe, and a smaller FDTI to reduce crosstalk between pixels, in accordance with an exemplary embodiment of the disclosed technology. [Figure 11] 1 illustrates an example cross-sectional view of a color back-side illuminated (BSI) single-photon avalanche diode (SPAD) pixel with internal microlenses for light focusing, in accordance with an exemplary embodiment of the disclosed technology. [Figure 12] An example of a cross-sectional view of a color backside-illuminated (BSI) single-photon avalanche diode (SPAD) pixel having a diffraction grating structure or a binary optical lens for directing incident light into a light pipe, according to an exemplary embodiment of the disclosed technology, is shown. [Figure 13] An example of a cross-sectional view of a color backside-illuminated (BSI) single-photon avalanche diode (SPAD) pixel having a significantly reduced light sensing area, with metal-filled FDTI, a light pipe, and a half-pitch gapless microlens for reducing crosstalk between pixels and for directing incident light into the light pipe for further light concentration, according to an exemplary embodiment of the disclosed technology, is shown. [Figure 14] A flow diagram of a method, according to an exemplary embodiment of the disclosed technology, is shown. **DETAILED DESCRIPTION**
[0018] Detailed Description The disclosed technology includes a novel pixel architecture that can reduce dark current and improve signal-to-noise ratio, particularly for low-light-level sensing applications. Some exemplary embodiments of the disclosed technology utilize a light sensing pixel having a large light receiving aperture characterized by a dimension D approximately equal to the pixel pitch (i.e., the pixel center-to-center spacing), a light concentrating structure, and a pixel sensing area characterized by a small sensing area having a dimension d < D (i.e., smaller than the light receiving aperture D), thereby enabling collection of more photons for dark current and / or read noise reduction. According to some exemplary embodiments, the ratio D / d can be configured as needed. In some exemplary embodiments, the pixel sensing area can be defined within a deep trench isolation boundary having a dimension of approximately d. Some exemplary embodiments of the disclosed technology may also include an additional deep trench isolation boundary having a dimension of approximately D for reducing crosstalk between pixels. In some exemplary embodiments, the deep trench isolation boundary can be metal-filled.
[0019] According to certain exemplary embodiments of the disclosed technology, analog and / or digital binning may be used to further improve the associated low light sensitivity of the pixel. Additionally, noise-free "optical binning" may be achieved by utilizing light concentration in the new pixel architecture. Certain exemplary embodiments of the disclosed technology may also enable the fabrication of the new pixel architecture using standard microelectronics foundry manufacturing processes utilizing silicon substrates, which may provide manufacturing, reliability, and cost reduction advantages over previous devices such as analog image intensifier (I2) tubes.
[0020] In photon sensing and imaging devices, the signal-to-noise ratio (SNR) of an image is limited by the total number of photons per pixel per frame time. To achieve higher SNR under extreme low illumination, exemplary embodiments of the disclosed technology utilize a large "effective" physical pixel size (defining the incident light receiving aperture) and a focusing structure that concentrates the collected incident light onto a smaller actual sensing pixel device.
[0021] 1 illustrates the concept of concentrating light. For example, a concentrator 102 may receive incident light 104 over its aperture dimension D and concentrate the incident light 104 to an area 106 having a smaller dimension d, thereby reducing the incident illuminance (lumens / m 2 ) to increase
[0022] Figure 2 shows the theoretical upper limit of the number of photons per pixel capturing scene illumination of 0.1 mlux across visible and near infrared wavelengths (400 nm - 1100 nm) captured at 30 frames per second using an F#1.4 lens, which corresponds to a blackbody of 2856 K. For a moonless night sky, the corresponding illumination power spectrum nearly matches an ideal 2856 K blackbody that can be used to calculate the photon count. To achieve an SNR of 1, one photon is required based on equation (1). At 0.1 mlux, the minimum pixel size required is 5.0 μm.
[0023] FIG. 3A illustrates an example cross-sectional view of a monochrome back-illuminated (BSI) CMOS pixel 300 in accordance with an exemplary embodiment of the disclosed technology.
[0024] 3B illustrates an alternative wafer stacking configuration that may be utilized with the BSI CMOS pixel 300 (such as the top of the BSI CMOS pixel shown in FIG. 3A), which may utilize, for example, a logic carrier wafer 320 bonded to the sensor wafer using wafer stacking to provide additional circuitry or functionality. Some embodiments of wafer stacking may utilize techniques, materials, and the like, as described in S.-G. Wuu, H.-L. Chen, H.-C. Chien, P. Enquist, RM Guidash, and J. McCarten, “A Review of 3-Dimensional Wafer Level Stacked Backside Illuminated CMOS Image Sensor Process Technologies,” in IEEE Transactions on Electron Devices, vol. 69, no. 6, pp. 2766-2778, June 2022, which is incorporated by reference herein as if fully set forth. Certain embodiments of wafer stacking may utilize techniques, materials, etc. as described in Y. Oike, “Evolution of Image Sensor Architectures With Stacked Device Technologies,” in IEEE Transactions on Electron Devices, vol. 69, no. 6, pp. 2757-2765, June 2022, which is incorporated by reference herein as if set forth in full.
[0025] Returning to Figure 3A, the exemplary pixel 300 includes a gapless microlens 302 that can receive incident light over an effective receiving aperture having a dimension D, and a light pipe 304 concentrator that further concentrates the incident light to a sensing area having a smaller dimension d. In this exemplary embodiment, a photodiode 312 may be utilized to detect the concentrated light. In other implementations described below, other detectors such as photoconductors, single photon avalanche diodes (SPADs), etc. may be utilized.
[0026] The light pipe 304 collector may include an outer region characterized by a first refractive index N1 and a central region characterized by a second refractive index N2>N1 such that light entering the top of the light pipe 304 collector is contained within the higher refractive index region (N2) by total internal reflection, similar to the waveguiding properties of an optical fiber. The light pipe 304 collector material may be selected to have very little / minimal absorption for light having wavelengths between 300 nm and 1200 nm. According to certain exemplary embodiments of the disclosed technology, various profiles, shapes, materials, and fabrication techniques of the light pipe 304 collector may be implemented without departing from the scope of the disclosed technology, as described in J. Gambino, et al, “CMOS image sensor with high refractive index lightpipe”, IISW 2009, which is incorporated herein by reference. According to certain exemplary embodiments of the disclosed technology, the dimensions of the actual pixel sensing area, the height of the light pipe, the sidewall profile of the light pipe, the curvature of the gapless microlens, and / or associated material properties may be optimized based on optical simulations, such as by finite-difference time-domain (FDTD) techniques.
[0027] 3A, certain embodiments of the disclosed technology utilize a first full depth deep trench isolation (FDTI) 310 to define, for example, a smaller device area (having a dimension d), thereby achieving a reduction in dark current. In certain exemplary embodiments, the FDTI 310 may extend completely through the silicon epilayer of the pixel 300 and may be characterized by a vertical boundary around the detector 310 having an approximate (horizontal inner extent) area characterized by a dimension d.
[0028] According to certain exemplary embodiments of the disclosed technology, the exemplary pixel 300 may include one or more of an optional back metal shield 306, a textured surface 308, a metal reflector 314, a metal routing layer 316, and a carrier wafer 318. Generally, the optional back metal shield 306 in combination with the texture 308 may increase the light trapped within the pixel area FDTI 310 and improve the absorption of NIR light. In certain exemplary embodiments, the texture 308 may be on the surface as shown in FIG. 3A or embedded elsewhere within the sensing area. For example, the texture 308 may be located on and near the metal reflector 314. In certain exemplary embodiments, the texture 308 may be located along one or more walls of the first FDTI 310. In certain exemplary embodiments, the carrier wafer 318 may be a dummy bulk wafer (shown at the bottom of FIG. 3A) or a logic carrier wafer 320 (shown in FIG. 3B) having logic and / or memory circuits, for example, to provide additional functionality to improve the performance of the image sensor. According to certain exemplary embodiments of the disclosed technology, the logic carrier wafer 320 may be bonded to the pixel wafer by wafer stacking techniques such as microbumps, through silicon vias (TSVs), direct Cu-Cu hybrid bonding, etc. The above-referenced components shown in FIG. 3A and / or FIG. 3B may be the same as or similar to components shown in other exemplary embodiments described below with reference to FIG. 6-FIG. 13. One or more of these features or components may be included to improve near infrared (NIR) light sensitivity, which may be important for many low light imaging applications. For example, for imaging applications in a moonless night sky, there are more photons in the NIR wavelength range than in the visible wavelength range. According to certain exemplary embodiments of the disclosed technology, the textured surface 308 may be utilized to improve the quantum efficiency (QE) of the near infrared light (NIR). In certain exemplary embodiments, a metallic reflector 314 may be utilized to further increase the NIR QE.
[0029] For CMOS pixels, dark current consists of three components: generation current in the depletion region, diffusion current, and surface generation, each of which depends on the pixel dimensions. State-of-the-art CMOS pixels can already achieve very low dark current. However, to further reduce the dark current for large pixels, shrinking the sensing area of the actual device (except for camera cooling, which is not practical for most low-light applications) may be the most effective way to reduce the dark current.
[0030] In addition to the direct benefit of reduced dark current for smaller pixel sensing areas (d), the reduced size of the sensing area may also provide the benefit of lower read noise. For example, in a CMOS pixel, the input-referred read noise is typically determined by the Float Diffusion (FD) conversion gain in units of "μV / e-", which is the inverse of the FD capacitance. Due to the miniaturization of the charge transfer gate (TX) and reduced coupling, the FD capacitance can be significantly smaller for smaller pixel sizes than for larger pixels, resulting in a significantly higher FD conversion gain. By further combining the disclosed technology with other FD technologies, such as distal FD, the FD capacitance can be further reduced. Additionally, for typical 4T CMOS pixels, the large pixel sensing area tends to have image retention issues due to the long travel distance of the charge collected in the photodiode. Image retention typically manifests as fixed pattern noise (FPN) on the image and can significantly reduce image quality for night vision applications. By using smaller pixels, the charge transfer can be significantly improved and therefore the FPN noise can be significantly reduced.
[0031] FIG. 4 illustrates an example of a top view of a pixel 400 having a sensing device 404 characterized by an “effective” optical pixel size D402 and a device pixel size d406 that is smaller than the optical pixel size D402. As described above, a lens and / or light pipe may “collect” incident light over the optical pixel size D402 and focus the light onto an area of size d406. According to an exemplary embodiment, the ratio D / d may be configured as needed. For example, the ratio D / d may be 1.5 or greater. In an exemplary embodiment, the ratio D / d may be 2 or greater. In an exemplary embodiment, the ratio D / d may be 5 or greater. In an exemplary embodiment, the pixel 400 includes a full depth deep trench isolation (FDTI) 408 that borders the sensing device 404. In an exemplary embodiment, the trench may be filled with a polysilicon dielectric. Other dielectric materials may be utilized to fill the FDTI 408 without departing from the scope of the disclosed technology.
[0032] In an exemplary implementation, the optical pixel size D402 may correspond to the pixel pitch (i.e., center-to-center spacing) of the array of pixels. According to an exemplary embodiment, the inner and / or outer regions 410 of the FDTI 408 may be silicon-based. In other exemplary embodiments, the inner and / or outer regions 410 of the FDTI 408 may be non-silicon-based, for example, made from one or more of InGaAs, InP, germanium, etc. The sensing device 404 may include a photodiode, a photoconductor, a single-photon avalanche diode (SPAD), or any combination thereof. In an exemplary embodiment, the pixel 400 may be a back-side illuminated (BSI) device. In other embodiments, the pixel 400 may be a front-side illuminated (FSI) device.
[0033] 5 shows an example of a top view of a pixel 500 having similar features as those described above with reference to FIG. 4, with the addition of a second FDTI 502 bordering the pixel 500. In an example implementation, the second FDTI 502 may be metal filled. An advantage of the disclosed technology is that it may allow for the reduction or complete elimination of optical and / or electrical crosstalk. For example, due to the smaller pixel device area 404 (defined in the first FDTI 408), the distance between adjacent pixel device areas may be increased.
[0034] In an exemplary embodiment, the first FDTI 408 and / or the second FDTI 502 may be filled with oxide or air. However, such an embodiment may not completely block optical crosstalk between pixels, since light may still penetrate such isolation trenches. However, in an embodiment, the first FDTI 408 and / or the second FDTI 502 may be filled with metal (Al, tungsten, Cu, etc.) that may completely block light. However, if the metal filling is near the sensing device 404 region, typically the metal filling will adversely affect the dark current. Thus, in an embodiment, the first FDTI 408 may be filled with air, oxide, polysilicon, etc., and the second FDTI 502 may include a metal filling. In this regard, the second FDTI 502 may be located far away from the small sensing device 404 region, for example, to avoid adversely affecting the dark current while eliminating crosstalk between pixels. In an example implementation, the area between the first FDTI 408 and the second FDTI 502 may be used to create other circuitry to support / enhance the performance or functionality of the image pixel / sensor. For example, one-time programmable memory (OTPM) cells may be placed in the area between the first FDTI 408 and the second FDTI 502 to store defective pixel location information, per-pixel offsets to reduce dark signal non-uniformity (DSNU), and / or photon response non-uniformity (PRNU).
[0035] FIG. 6 illustrates an example cross-sectional view of a color back-illuminated (BSI) CMOS pixel 600 (similar to pixel device 300 described above with reference to FIG. 3A) with an additional color filter 602 according to an exemplary embodiment of the disclosed technology. In an exemplary embodiment, the color filter 602 may be part of a color filter array (CFA). This exemplary embodiment illustrates how the disclosed technology can be used for different wavelength filtering applications, including but not limited to pixels designed for monochrome, color, or hyperspectral applications. In an exemplary embodiment, the color filter 602 may include dye-based and / or pigment-based materials. In an exemplary embodiment, the color filter 602 may include a diffraction grating-based filter and / or a nanostructure-based filter. In an exemplary embodiment, the color filter 602 may include a thin film multilayer structure. In an exemplary embodiment, the color filter 602 may include a Fabry-Perot-based optical filter.
[0036] According to certain exemplary embodiments of the disclosed technology, the curvature, height, material, and / or other properties of the associated lens 302 may be individually optimized for each color pixel to achieve the best light collection for different color pixels (such as green, red, blue, or NIR wavelengths). In certain exemplary embodiments, the fill material, height, and / or other properties of the light pipe 304 may be individually optimized for each color pixel to achieve the best light collection results.
[0037] 7 illustrates an example cross-sectional view of a color front-illuminated (FSI) CMOS pixel 700, according to an exemplary embodiment of the disclosed technology. As described above with respect to pixel 600 illustrated in FIG. 6 (except for the BSI configuration of pixel 600), FSI pixel 700 may include a gapless microlens 702 (which may be part of a lens array), a light pipe 724 concentrator, and / or a color filter 704 (which may be part of a color filter array). Microlens 702 and / or light pipe 724 concentrator may enable collection of light onto a receiving aperture D and focusing the light onto a sensing area d, according to an exemplary embodiment of the disclosed technology.
[0038] Because the exemplary pixel 700 shown in FIG. 7 is designed to be front-illuminated (as opposed to back-illuminated designs in other embodiments described herein), the FSI CMOS pixel 700 may include one or more metal layers 706, 708, 710. In an exemplary implementation, the metal layers 706, 708, 710 may be connected by one or more vias 728. In an exemplary implementation, the one or more metal layers 706, 708, 710 and the one or more vias 728 may provide access to the transistor 714, for example, access from the photodiode to a sensing floating diffusion (FD) node, reset, and / or transfer of charge. In an exemplary implementation, the sensing region d may include a P+ region 716 and an N well 718, which may form a pinned photodiode for sensing concentrated incident light.
[0039] According to certain exemplary embodiments of the disclosed technology, the exemplary pixel 700 may include deep trench isolation 720, for example, in a silicon epi layer 732. Certain embodiments may include a P++ substrate 722, for example, on the backside of the pixel 700. Certain exemplary embodiments of the pixel 700 may include a metal aperture layer 712. Certain exemplary embodiments of the pixel 700 may include an anti-reflective coating 730.
[0040] FIG. 8 illustrates an example cross-sectional view of another monochrome back-illuminated (BSI) single-photon avalanche diode (SPAD) pixel 800 with a light pipe concentrator having a significantly reduced light sensing area, according to an exemplary embodiment of the disclosed technology. The BSI SPAD pixel 800 may include similar design features as those described above with reference to the BSI CMOS pixel 300 shown in FIG. 3A, with the main difference being that the sensing area of the BSI SPAD pixel 800 may include a SPAD 802. The exemplary pixel 800 may include one or more of a back metal shield 306, a textured surface 308, a metal reflector 314, a metal routing layer 316, and a carrier wafer 318. As described above with respect to FIGS. 3A and 3B, the carrier wafer 318 may be a dummy bulk wafer (shown at the bottom of FIG. 3A) or a logic carrier wafer 320 (shown in FIG. 3B) having logic and / or memory circuits, for example, to provide additional functionality to improve the performance of the image sensor. According to certain exemplary embodiments of the disclosed technology, the carrier wafer 318 and / or the logic carrier wafer 320 may be bonded to the pixel wafer by wafer lamination techniques such as microbumps, through silicon vias (TSVs), direct Cu-Cu hybrid bonding, and the like.
[0041] As previously mentioned, the BSI SPAD pixel 800 may include a light collection structure comprised of a gapless microlens 302 and a light pipe 304. Certain exemplary implementations of the disclosed technology of this SPAD pixel 800 may include an embedded texture structure 308 to improve NIR QE.
[0042] Generally, SPAD pixels may be characterized by a read noise of 0e-. However, a major drawback of most SPAD pixels is their high dark count rate (DCR), which is roughly comparable to the dark current of CMOS pixels. In SPADs, the DCR is mainly due to the electric field strength of the avalanche region, the volume of the avalanche region, and / or the excessive bias voltage. Other dark current factors (such as diffusion current and surface generation) may also have an impact and can be reduced by shrinking the pixel device area, as in CMOS pixels. If the SPAD pixel device area is smaller, the avalanche region 802 can be much smaller and the excessive bias voltage to reach the avalanche state can be significantly reduced. According to certain exemplary embodiments of the disclosed technology, such factors can be utilized to contribute to making the DCR much smaller based on the disclosed technology. However, unlike CMOS pixels, the lowest DCR of a SPAD may not correspond to a minimum pixel size. Based on optical and electrical simulations of the device, the lowest DCR SPAD pixel can be achieved for moderate pixel sizes, for example, between about 3 μm and about 6 μm.
[0043] FIG. 9 illustrates an example cross-sectional view of a color back-illuminated (BSI) single-photon avalanche diode (SPAD) pixel 900 with a light pipe concentrator having a significantly reduced light sensing area according to an exemplary embodiment of the disclosed technology. The color BSI SPAD pixel 900 may include similar design features as those described above with reference to the monochrome BSI CMOS pixel 800 described above in FIG. 8, with the main difference being that the color BSI SPAD pixel 900 may include a color filter 902, which may be part of the array in some embodiments. Although a color BSI SPAD pixel 900 is shown in FIG. 9, some embodiments of the disclosed technology may also be used to create an FSI SPAD pixel with significantly reduced DCR, similar to the FSI CMOS pixel 700 described above with reference to FIG. 7. In this case, the sensing area d may utilize a SPAD.
[0044] 10 shows an example cross-sectional view of a color back-side illuminated (BSI) single-photon avalanche diode (SPAD) pixel 1000 having a metal-filled (second) FDTI 1004 to reduce or completely eliminate crosstalk between pixels, according to an exemplary embodiment of the disclosed technology. As with other examples described above, the color BSI SPAD pixel 1000 can include a metal-filled (second) FDTI 1004, a light pipe, and a significantly reduced light-sensing area SPAD 802 bordered by a smaller first FDTI 310.
[0045] FIG. 11 illustrates an example cross-sectional view of a color back-side illuminated (BSI) single-photon avalanche diode (SPAD) pixel 1100 having an internal microlens 1102 for light focusing, in accordance with an exemplary embodiment of the disclosed technique.
[0046] 12 shows an example cross-sectional view of a color back-side illuminated (BSI) single-photon avalanche diode (SPAD) pixel 1200 having a grating / metalens / nanostructure 1202 for directing incoming light into a light pipe, according to an exemplary embodiment of the disclosed technology. In this example, the lens 1202 may be fabricated using a 1D grating, a 2D grating, a nanocolumn / pillar, a metalens, or other structure. In an exemplary embodiment, the lens 1202 may be fine-tuned to a desired wavelength range for each color pixel. In an exemplary embodiment, this same type of (non-conventional) lens structure 1202 may be utilized in any of the embodiments described herein.
[0047] FIG. 13 shows an exemplary cross-sectional view of a color back-illuminated (BSI) single-photon avalanche diode (SPAD) pixel 1300 having a half-pitch gapless microlens 1302 for directing incident light to a light pipe for further collection according to an exemplary embodiment of the disclosed technology. As in other examples described above, the color BSI SPAD pixel 1300 may include a metal-filled (second) FDTI 1004, a light pipe, and a significantly reduced light-sensing area SPAD 802 bordered with a smaller first FDTI 310. According to an exemplary embodiment of the disclosed technology, reducing the size of the microlens 1302 to half the unit pixel size may improve the collection efficiency. This color BSI SPAD pixel 1300 is also shown as having a significantly reduced light-sensing area bordered with a metal-filled FDTI, a light pipe, and a smaller FDTI 310 for reducing crosstalk between pixels.
[0048] Although a SPAD-based BSI pixel is shown in Figures 8-13, the disclosed techniques may also be used for FSI SPAD pixels with reduced DCR. For example, rather than using a photodiode-based detector in sensing area d of Figure 7, a SPAD structure (similar to SPAD 802 shown in Figure 8) may be used.
[0049] It should be appreciated that any of the embodiments disclosed herein may utilize a second metal-filled FDTI (such as the FDTI 502 described with reference to FIG. 5, or the FTDI 1004 described with reference to FIG. 10), for example, to reduce or completely eliminate pixel crosstalk.
[0050] It should be appreciated that any of the embodiments disclosed herein may utilize a top gapless microlens and a buried internal microlens (such as microlens 1102 described above with reference to FIG. 11).
[0051] It should be further appreciated that any of the embodiments disclosed herein may utilize non-conventional lenses (such as lens 1202 described above with reference to FIG. 12), which may be fabricated as 1D gratings, 2D gratings, nanocolumns, nanopillars, metalenses, binary optical lenses, Fresnel lenses, and / or other structures. According to certain exemplary implementations of the disclosed technology, the non-conventional lens structures may be designed for specific wavelength ranges for each color pixel.
[0052] It should be further appreciated that any of the embodiments disclosed herein may utilize half (or smaller) pitch gapless microlenses and light pipes. By reducing the size of the microlenses to half the unit pixel size, focusing efficiency may be improved.
[0053] It should be further appreciated that any of the embodiments disclosed herein may be applied to CMOS or SPAD pixels, or charge-coupled device (CCD) pixels made by wafer stacking techniques (two wafers, three wafers, or more). Thus, the disclosed techniques may be applicable to FSI and / or BSI applications that utilize wafer stacking. For example, wafer bonding and stacking techniques may be utilized to bond wafers to add additional functionality.
[0054] Certain embodiments of the disclosed technology may be applied to other pixel designs for use with different wavelength ranges or other non-silicon based materials, and may be particularly beneficial in pixelated devices where the dark current of the device is highly dependent on pixel dimensions. As described herein, the performance of the device may be improved by using light collection (i.e., optical binning) to concentrate the collected photons onto a greatly reduced device area.
[0055] It should be appreciated that other pixel materials may be utilized without departing from the scope of the disclosed technology. For example, certain embodiments of the disclosed technology may employ pixels having materials such as germanium, microbolometers, III-V materials (GaAs, InGaAs, etc.) for SWIR, MWIR, LWIR, and / or VLWIR applications. Light collection structures may be compatible with such different materials and may include microlenses, diffraction gratings, nanostructures, metalenses, light pipes, internally embedded microlenses, etc.
[0056] Technical advantages of the disclosed technology may include one or more of: reduced pixel dark current, reduced pixel crosstalk, reduced CMOS pixel read noise, reduced pixel delay, reduced SPAD pixel excess bias voltage, reduced SPAD pixel DCR, and / or increased SNR for low light images.
[0057] FIG. 14 is a flow diagram of a method 1400 for manufacturing an imaging device with reduced dark current and improved signal-to-noise ratio by forming a pixel array, where one or more of each pixel of the pixel array may be manufactured by the method 1400 according to an exemplary embodiment of the disclosed technology. In block 1402, the method 1400 includes forming a sensing region having a dimension d on a wafer. In block 1404, the method 1400 includes forming a full depth deep trench isolation (FDTI) to border the sensing region. In block 1406, the method 1400 includes forming a light-collecting structure. In block 1408, the method 1400 includes forming a gapless microlens array on the pixel array, where each gapless microlens of the gapless microlens array defines a light-receiving aperture characterized by a dimension D greater than d and is configured to receive light incident on the light-receiving aperture and focus and direct the received light to the sensing region.
[0058] According to certain example implementations of the disclosed technology, forming a gapless microlens array on a pixel array may include forming a microlens array structure using photolithography and further applying a material reflow or material etch to the microlens array structure.
[0059] In certain example embodiments, forming the light collection structures may include forming light pipe waveguides, gapless microlenses, internal microlenses, and / or binary optical lenses.
[0060] According to certain exemplary embodiments of the disclosed technology, forming the sensing region may include forming an embedded texture and / or a metallic reflector structure on a silicon surface of the sensing region.
[0061] As described herein, the disclosed technology can utilize built-in textures on the Si surface to improve near-infrared (NIR) quantum efficiency (QE) and metal reflector structures to further increase the NIR QE.
[0062] Certain exemplary embodiments of the disclosed technology may be utilized to fabricate pixels with reduced read noise. Because input-referred read noise may be a function of floating diffusion (FD) conversion gain, which is inversely proportional to FD capacitance, input-referred read noise may be reduced by combining FD technology (such as distal FD) with the disclosed technology to enable smaller pixel sizes.
[0063] As disclosed herein, the ratio D / d may be configured, for example, by adjusting design parameters. For example, the ratio D / d may be 1.5 or greater. In an exemplary implementation, the ratio D / d may be 2 or greater. In an exemplary embodiment, the ratio D / d may be 5 or greater. According to an embodiment, the ratio may be configured as needed by specifying one or more of: (a) the dimension d of the sensing area; (b) the light collection structure (including geometry and material); and / or (c) the design of the microlenses.
[0064] In one exemplary embodiment, improved signal-to-noise ratio can be achieved by reducing the read noise resulting from smaller charge transfer gate (TX) size and reduced coupling, as well as further reduction in float diffusion (FD) capacitance through FD technologies such as distal FD.
[0065] In certain example implementations, improved signal-to-noise ratios may be achieved through a reduction in pixel crosstalk, both optical and electrical, and / or through an increase in the distance between adjacent pixel device areas.
[0066] Certain exemplary embodiments of the disclosed technology can be utilized to fabricate BSI and / or FSI pixels with reduced pixel crosstalk, both optical and electrical, due to the increased distance between adjacent pixel device areas and by virtue of the smaller pixel device sensing area defined by its border FDTI.
[0067] Certain aspects of the disclosed technology may provide digital images that are comparable to or better than images (in low light conditions) that can be produced only by image intensifier systems. As described herein, certain aspects of the disclosed technology provide imaging arrays configured to convert detected photons into a digital image.
[0068] As mentioned above, the systems and methods described herein may, in some aspects, provide for processing of devices at the wafer level. For example, such a wafer may include a plurality of pixel devices as described herein. Compared to conventional image intensifier manufacturing techniques that fabricate a single image intensifier at a time, many pixel devices according to various aspects of the disclosed technology may be fabricated on a single wafer, thereby increasing throughput and / or reducing cost per device through parallel processing.
[0069] In various aspects, the wafer may include a "sensing" array subcomponent that includes multiple photodiodes, SPADs, etc., each with its own light collector. In one exemplary embodiment, the disclosed techniques may include aligning an array of microlenses with multiple photodiodes / SPADs.
[0070] Numerous features and advantages have been set forth in the foregoing description together with structural and functional details. While the disclosed technology has been disclosed in several forms, it will be apparent to those skilled in the art that numerous modifications, additions, and deletions may be made, particularly with respect to the shape, size, and arrangement of parts, without departing from the spirit and scope of the disclosed technology as set forth in the following claims and its equivalents. These encompass various alternatives, modifications, and equivalents, as will be appreciated by those skilled in the art. [Explanation of symbols]
[0071] 102 Concentrator 104 Incident light 106 areas 300 pixels 302 Micro Lens 304 Light Pipe 306 Back metal shield 308 Textured Surfaces 310 FDT 312 Photodiode 314 Metal reflector 316 Metal Routing Layer 318 Carrier Wafer 320 Logic Carrier Wafer 400 pixels 402 Optical pixel size D 404 Sensing Devices 406 Device pixel size d 408 FDTI 410 areas 500 pixels 502 Second FDTI 600 pixels 602 Color Filter 700 pixels 702 Micro Lens 704 Color Filter 706 Metal layer 708 Metal layer 710 Metal layer 712 Metal aperture layer 714 Transistor 716 P+ area 718 N-well 720 Deep Trench Isolation 722 P++ board 724 Light Pipe 728 Via 730 Anti-Reflection Coating 732 Silicon epitaxial layer 800 pixels 802 SPAD 900 Color BSI SPAD Pixels 902 Color Filter 1000 Color BSI SPAD Pixels 1004 FDTI 1100 Color BSI SPAD Pixels 1102 Microlens 1200 Color BSI SPAD Pixels 1202 Diffraction Gratings / Metalens / Nanostructures 1300 Color BSI SPAD Pixels 1302 Micro Lens
Claims
1. a light receiving aperture having a first dimension D defined by a unit pixel pitch; a sensing area having a second dimension d smaller than the first dimension D of the unit pixel pitch, the sensing area being defined within a first full depth deep trench isolation (FDTI) boundary; a light collecting structure configured to receive light incident on the light receiving aperture and to concentrate and direct the received light onto the sensing area; A light-sensing pixel, characterized by A pixel architecture for an imaging device with reduced dark current and improved signal-to-noise ratio, comprising:
2. The pixel architecture of claim 1 , wherein the light-sensing pixels are back-side illuminated (BSI) CMOS pixels or front-side illuminated (FSI) CMOS pixels.
3. The pixel architecture of claim 1 , wherein the light collection structure comprises a light pipe waveguide.
4. 10. The pixel architecture of claim 1, wherein the light collection structure comprises one or more of a binary optical lens and a grating-based lens.
5. 10. The pixel architecture of claim 1, wherein the light collection structures include one or more of a gapless microlens and an internal microlens.
6. 10. The pixel architecture of claim 1, wherein the sensing region comprises a photodiode, a photoconductor, a single-photon avalanche diode (SPAD), or a non-silicon based detector made from one or more of InGaAs, InP, and germanium.
7. The pixel architecture of claim 1 , further comprising a second FDTI having a dimension substantially equal to D and surrounding the first FDTI.
8. The pixel architecture of claim 7 , wherein the second FDTI is filled to reduce or eliminate pixel crosstalk.
9. The pixel architecture of claim 1 , further comprising an embedded texture on a silicon surface or embedded within the sensing area.
10. The pixel architecture of claim 1 , further comprising a metallic reflector structure.
11. The pixel architecture of claim 1 further comprising a color filter.
12. The pixel architecture of claim 1 , wherein the ratio D / d is greater than or equal to 1.
5.
13. The pixel architecture of claim 1 , wherein the ratio D / d is greater than or equal to 2.
0.
14. The pixel architecture of claim 1 , wherein the ratio D / d is greater than or equal to 5.
0.
15. An array of light-sensing pixels, each light-sensing pixel of the array comprising: a light receiving aperture having a first dimension D defined by a unit pixel pitch; a sensing area having a second dimension d smaller than the first dimension D of the unit pixel pitch, the sensing area being defined within a first full depth deep trench isolation (FDTI) boundary; a light collecting structure configured to receive light incident on the light receiving aperture and to concentrate and direct the received light onto the sensing area; The array of light-sensing pixels, 1. A night vision device with reduced dark current and improved signal-to-noise ratio comprising:
16. 16. The night vision device of claim 15, wherein each light-sensing pixel of the array is a back-side illuminated (BSI) CMOS pixel or a front-side illuminated (FSI) CMOS pixel.
17. The light collecting structure is Light pipe waveguide, Gapless microlens, Internal microlenses, and Binary Optical Lens 16. The night vision device of claim 15, comprising one or more of:
18. 16. The night vision device of claim 15, wherein each light-sensing pixel of the array includes a second FDTI having a dimension substantially equal to D and surrounding the first FDTI, the second FDTI being metal filled to reduce or eliminate pixel crosstalk.
19. Each light-sensing pixel of the array is an embedded texture on the silicon surface of said sensing area; A metal reflector structure, and Color Filters 16. The night vision device of claim 15, comprising one or more of:
20. 16. A night vision device according to claim 15, wherein the ratio D / d is 1.5 or greater.
21. forming a pixel array, each pixel of the pixel array comprising: forming a sensing area having a dimension d on a wafer; forming a full depth deep trench isolation (FDTI) to border the sensing area; forming a light collecting structure; forming a gapless microlens array over the pixel array, each gapless microlens of the gapless microlens array defining a light-receiving aperture characterized by a dimension D greater than d and configured to receive light incident on the light-receiving aperture and to focus and direct the received light onto the sensing area; forming the pixel array, 2. A method for manufacturing an imaging device having reduced dark current and improved signal-to-noise ratio, comprising:
22. 22. The method of claim 21 , wherein forming a gapless microlens array on the pixel array comprises forming a microlens array structure using photolithography and further applying a material reflow or material etch to the microlens array structure.
23. forming the light collecting structure, Light pipe waveguide, Gapless microlens, Internal microlenses, and Binary Optical Lens 22. The method of claim 21, comprising forming one or more of:
24. forming the sensing region an embedded texture on the silicon surface of the sensing area; and metal reflector structure 22. The method of claim 21, comprising forming one or more of:
