Complementary Metal Oxide Semiconductor Image Research Articles

Metasurface-enabled on-chip multiplexed diffractive neural networks in the visible

Replacing electrons with photons is a compelling route toward high-speed, massively parallel, and low-power artificial intelligence computing. Recently, diffractive networks composed of phase surfaces were trained to perform machine learning tasks through linear optical transformations. However, the existing architectures often comprise bulky components and, most critically, they cannot mimic the human brain for multitasking. Here, we demonstrate a multi-skilled diffractive neural network based on a metasurface device, which can perform on-chip multi-channel sensing and multitasking in the visible. The polarization multiplexing scheme of the subwavelength nanostructures is applied to construct a multi-channel classifier framework for simultaneous recognition of digital and fashionable items. The areal density of the artificial neurons can reach up to 6.25 × 106 mm−2 multiplied by the number of channels. The metasurface is integrated with the mature complementary metal-oxide semiconductor imaging sensor, providing a chip-scale architecture to process information directly at physical layers for energy-efficient and ultra-fast image processing in machine vision, autonomous driving, and precision medicine.

An Alternate Approach to Backside Thinning: A Doping Selective Silicon Wet Etch

Backside illuminated technologies have been used in complementary metal–oxide–semiconductor image sensors for years and are finding a renewed interest and are being investigated for imagers and detectors in light detection and ranging applications. Coupled with the improved performance facilitated through 3D integration, the thinning of bonded wafers has become a crucial processing step. In this study, an alternate and potentially more cost-effective method has been explored using an HNA (hydrofluoric, nitric, acetic)wet etch for backside thinning of bonded epitaxial P+/P- silicon wafers. By utilizing a dopant concentration gradient, P+ silicon is selectively etched and P- silicon can be effectively used as an etch stop. Several HNA concentrations were explored to determine ratios that yield high P+ silicon etch rates while maintaining high P+/P- selectivity. Furthermore, bath life, spiking scheme, and repeatability were examined using a large wafer sample size with the goal of developing a full process flow.

Robust Dynamic Indoor Visible Light Positioning Method Based on CMOS Image Sensor

A real-time imaging recognition and positioning method based on visible light communication flat light source is proposed. This method images the visible light communication flat light source through the rolling shutter effect of the complementary metal-oxide semiconductor imaging sensor and obtains the rectangular area outline of the light source. The light and dark stripe information of image with the digital image processing method realizes light source matching recognition by defining the concept, the autocorrelation sequence, which can be used to obtain the identity of the light source, and the rectangular vertex coordinate information of flat light source achieves high-precision vision positioning on the basis of inertial measurement unit attitude sensor-assisted imaging. Simultaneously, the corresponding positioning module is developed for positioning testing. The test results indicate that the plane positioning error is less than 4.5 cm, and the positioning frequency is greater than 10 Hz, which provides a high-precision visual positioning solution for indoor positioning.

Characterization of low light performance of a complementary metal-oxide semiconductor sensor for ultraviolet astronomical applications

Complementary metal-oxide semiconductor (CMOS) detectors offer many advantages over charge-coupled devices (CCDs) for optical and ultraviolet (UV) astronomical applications, especially in space where high radiation tolerance is required. However, astronomical instruments are most often designed for low light-level observations demanding low dark current and read noise, good linearity, and high dynamic range, characteristics that have not been widely demonstrated for CMOS imagers. We report the performance, over temperatures from 140 to 240 K, of a radiation hardened SRI 4k × 2k back-side illuminated CMOS image sensor with surface treatments that make it highly sensitive in blue and UV bands. After suppressing emission from glow sites resulting from defects in the engineering grade device examined, a 0.077 me − / s dark current floor is reached at 160 K, rising to 1 me − / s at 184 K, rivaling that of the best CCDs. We examine the trade-off between readout speed and read noise, finding that 1.43 e − median read noise is achieved using line-wise digital correlated double sampling at 700 kpix / s / ch corresponding to a 1.5 s readout time. The 15 ke − well capacity in high gain mode extends to 120 ke − in dual gain mode. Continued collection of photogenerated charge during readout enables a further dynamic range extension beyond 106 e − effective well capacity with only 1% loss of exposure efficiency by combining short and long exposures. A quadratic fit to correct for non-linearity reduces gain correction residuals from 1.5% to 0.2% in low gain mode and to 0.4% in high gain mode. Cross-talk to adjacent pixels is only 0.4% vertically, 0.6% horizontally, and 0.1% diagonally. These characteristics plus the relatively large (10 μm) pixel size, quasi 4-side buttability, electronic shutter, and sub-array readout make this sensor an excellent choice for wide field astronomical imaging in space, even at far-UV wavelengths where sky background is very low.

Quantum Efficiency Simulation and Analysis of Irradiated Complementary Metal-Oxide Semiconductor Image Sensors

A quantum efficiency model of complementary metal-oxide semiconductor image sensors based on Shockley–Read–Hall and Auger recombination is developed using the technology computer-aided design tool, and the quantum efficiency degradation after irradiation is analyzed. By simulating the surface recombination velocity and depletion region width of the photodiode, the decrease in the quantum efficiency of complementary metal-oxide semiconductor image sensors under short and long incident light wavelengths is found to be caused by the increase in the surface recombination velocity and capture of optical carriers by radiation-induced defects in the epitaxial layer, respectively. In addition, a method to reduce the quantum efficiency degradation behavior of an irradiated pixel is discussed.

Visible‐Blind ZnMgO Colloidal Quantum Dot Downconverters Expand Silicon CMOS Sensors Spectral Coverage into Ultraviolet and Enable UV‐Band Discrimination

Selective spectral detection of ultraviolet (UV) radiation is highly important across numerous fields from health and safety to industrial and environmental monitoring applications. Herein, a nontoxic, visible-blind, quantum dot (QD)-based sensing scheme that expands the spectral coverage of silicon complementary metal-oxide-semiconductor (CMOS) sensors into the UV, enabling efficient UV detection without affecting the sensor performance in the visible and UV-band discrimination, is reported. This scheme uses zinc magnesium oxide (ZnMgO) QDs with compositionally tunable absorption across UV and high photoluminescence quantum yield in the visible. The efficient luminescence and large Stokes shift of these QDs are exploited herein to act as an efficient downconverting material that enhances the UV sensitivity of Si-photodetectors (Si-PDs). A Si-PD integrated with the QDs results in a ninefold improvement in photoresponsivity from 0.83 to 7.5mA W-1 at 260nm. Leveraging the tunability of these QDs, a simple UV-band identification scheme is further reported, which uses two distinct-bandgap ZnMgO QDs stacked in a tandem architecture whose spectral emission color depends on the UV-band excitation light. The downconverting stack enables facile discrimination of UV light using a standard CMOS image sensor (camera) or by the naked eye and avoids the use of complex optics.

Design of a CMOS image sensor pixel with embedded polysilicon nano-grating for near-infrared imaging enhancement.

Complementary metal-oxide semiconductor (CMOS) image sensor sensitivity in the near-infrared spectrum is limited by the absorption length in silicon. To deal with that limitation, we evaluate the implementation of a polysilicon nano-grating inside a pixel, at the transistor gate level of a 90 nm standard CMOS process, through opto-electrical simulations. The studied pixel structure involves a polysilicon nano-grating, designed with the fabrication layer of the transistor gate, which does not require any modifications in the process flow. The diffraction effect of the nano-grating increases the length of the light path in the photosensitive area and thus increases the photoelectric conversion efficiency. The nano-grating is integrated in combination with deep trench isolations to reduce cross talk between pixels. Coupled optical and electrical simulations report 33% external quantum efficiency improvement and 7% cross talk reduction at 850 nm.

Single event transient effect of frontside and backside illumination image sensors under proton irradiation

Complementary metal oxide semiconductor (CMOS) image sensor is susceptible to proton single event effect when being applied to space environment. Proton irradiation experiments with different energy values are carried out on a commercial FSI and BSI CMOS image sensors. The proton single event effect is analyzed by on-line testing, the maximum proton energy is 200 MeV, and the total fluence is 10<sup>10 </sup>particle/cm<sup>2</sup>. The single-event transient bright spots of different shapes are observed in the pixel array. By extracting the deposition energy and size, the effects of different energy protons on transient bright spot characteristics are compared, and the transient bright spot characteristics between FSI and BSI are also compared. Finally, the energy deposition distribution of the transient bright spot generated by proton in CMOS image sensor pixel unit is predicted by comparing the simulation method with the experimental results. The simulation results verify that the decrease of PPD depletion region thickness and the thinning of epitaxial layer are the main factors leading the proton energy deposition distribution to shift leftward in BSI image sensor.

Total Ionizing Dose Effects of the Color Complementary Metal Oxide Semiconductor (CMOS) Image Sensor at Different Bias

The experiments of total ionizing dose radiation effects on color CMOS image sensors at different biases were presented. Two bias conditions operated during 60Co γ irradiation, biased and unbiased. In the data processing, the data of four channels were extracted according to the sequence of Bayer array respectively to study the difference between each channel. The full well capacity and dark signal non-uniformity versus the cumulative doses were investigated. After γ-ray irradiation, the full well capacity of the samples decreased along with dark signal non-uniformity increased. Meanwhile, the descending amplitude of full well capacity was different in each channel, but the variation of dark signal non-uniformity was almost consistent. Moreover, the degradations of full well capacity and dark signal non-uniformity were more seriously under-biased.

Silicon Wafer Gettering Design for Advanced CMOS Image Sensors Using Hydrocarbon Molecular Ion Implantation: A Review

We have developed silicon epitaxial wafers with high gettering capability using hydrocarbon molecular ion implantation for advanced Complementary Metal-Oxide-Semiconductor (CMOS) image sensors. These wafers have three unique silicon wafer characteristics for improvement of CMOS device electrical parameter such as high metallic impurity gettering, oxygen out-diffusion barrier effects from Czochralski silicon (CZ) substrate and hydrogen passivation effect for interface state defect at Si/SiO <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sub> . We demonstrate that double epitaxial growth silicon wafers have an extremely high gettering capability during CMOS device fabrication process. We also found that gettering capability strongly dependence on oxygen impurity amount in hydrocarbon molecular ion implantation projection range. We believe that this novel silicon wafer can drastically contribute to the improvement of CMOS image sensor device performance such as white spot defect and dark current.

Effective interference mitigation scheme for multi-LED-based mobile optical camera communication.

Recently, optical camera communication (OCC) based on light-emitting diodes (LEDs) and complementary metal oxide semiconductor (CMOS) image sensors has received extensive attention due to its low cost and high flexibility. However, considering that the location of the user with a CMOS image sensor changes frequently, multi-LEDs are required to achieve seamless communication. In the paper, to mitigate random interchannel interference in overlapping areas of LED radiation, an effective identification recognition and sorting-based column selection are proposed. Meanwhile, to alleviate the distortion of bright and dark stripes caused by overlapping exposure time and user movement, an improved sampling scheme based on length estimation is proposed and experimentally demonstrated in a multi-LED-based mobile OCC system. Moreover, by varying the overlap rate of the adjacent LEDs' radiation area and the moving speed of the user, respectively, the performance of the system is investigated. The experimental results show that, the bit error rate (BER) performance of the system decreases slightly with the increasing of moving speed, but it is lower than the hard-decision forward error code rate limit. In addition, at the overlap rate of the LED radiation area of 4.3% and the moving speed of 80 cm/s, the BER is close to 10-6 using the proposed scheme.

All-in-one two-dimensional retinomorphic hardware device for motion detection and recognition.

With the advent of the Internet of Things era, the detection and recognition of moving objects is becoming increasingly important1. The current motion detection and recognition (MDR) technology based on the complementary metal oxide semiconductor (CMOS) image sensors (CIS) platform contains redundant sensing, transmission conversion, processing and memory modules, rendering the existing systems bulky and inefficient in comparison to the human retina. Until now, non-memory capable vision sensors have only been used for static targets, rather than MDR. Here, we present a retina-inspired two-dimensional (2D) heterostructure based retinomorphic hardware device with all-in-one perception, memory and computing capabilities for the detection and recognition of moving trolleys. The proposed 2D retinomorphic device senses an optical stimulus to generate progressively tuneable positive/negative photoresponses and memorizes it, combined with interframe differencing computations, to achieve 100% separation detection of moving trichromatic trolleys without ghosting. The detected motion images are fed into a conductance mapped neural network to achieve fast trolley recognition in as few as four training epochs at 10% noise level, outperforming previous results from similar customized datasets. The prototype demonstration of a 2D retinomorphic device with integrated perceptual memory and computation provides the possibility of building compact, efficient MDR hardware.

Hydrogen diffusion behavior in CH2P-molecular-ion-implanted silicon wafers for CMOS image sensors

Three-dimensional (3D) stacked complementary metal oxide semiconductor (CMOS) image sensors require reduction in leakage current due to interface state defects at the SiO2/Si interface. Hydrocarbon molecular ion implanted silicon wafers have identified the hydrogen termination effect of such wafer, contributing to high performance of CMOS image sensors. However, recently, low-temperature heat treatment during a device process has been shown to reduce the concentration of hydrogen diffusing from the hydrocarbon-molecular-ion-implanted region. Thus, a new method of molecular ion implantation with phosphorus added has been developed. In this paper, we present the results of our analysis of the diffusion behavior of hydrogen in a hydrocarbon molecule with phosphorus (CH2P) by reaction kinetic analysis and TCAD simulation. The results of reaction kinetic analysis show binding energies of 0.76 eV (C–H2 binding state) and 0.45 eV (P–H binding state). TCAD simulation results show that the 0.76 eV binding energy indicates that hydrogen is adsorbed in a carbon and silicon self-interstitial cluster (C/I cluster). On the other hand, the binding energy of 0.45 eV indicates that hydrogen is trapped in phosphorus complexes. Hydrogen in the CH2P-implanted region diffuses owing to the difference in between these binding states. Thus, a reduction in the interface state density of Si/SiO2 can be expected.

Reduction of Fluorine Diffusion and Improvement of Dark Current Using Carbon Implantation in CMOS Image Sensor

Recently, the demand of a high resolution complementary metal-oxide semiconductor (CMOS) image sensor is dramatically increasing. As the pixel size reduces to submicron, however, the quality of the sensor image decreases. In particular, the dark current can act as a large noise source resulting in reduction of the quality of the sensor image. Fluorine ion implantation was commonly used to improve the dark current by reducing the trap state density. However, the implanted fluorine diffused to the outside of the silicon surface and disappeared after annealing process. In this paper, we analyzed the effects of carbon implantation on the fluorine diffusion and the dark current characteristics of the CMOS image sensor. As the carbon was implanted with dose of 5.0 × 1014 and 1 × 1015 ions/cm2 in N+ area of FD region, the retained dose of fluorine was improved by more than 131% and 242%, respectively than no carbon implantation indicating that the higher concentration of the carbon implantation, the higher the retained dose of fluorine after annealing. As the retained fluorine concentration increased, the minority carriers of electrons or holes decreased by more Si-F bond formation, resulting in increasing the sheet resistance. When carbon was implanted with 1.0 × 1015 ions/cm2, the defective pixel, dark current, transient noise, and flicker were much improved by 25%, 9.4%, 1%, and 28%, respectively compared to no carbon implantation. Therefore, the diffusion of fluorine after annealing could be improved by the carbon implantation leading to improvement of the dark current characteristics.

Density Functional Theory Study on Dependence of Stability of Fe, Cu, and Ni Atoms on Surface Orientation of Si Crystal

The deep levels related to metal impurities degrade the yield of the complementary metal oxide semiconductor (CMOS) image sensors due to the increase in white pixels. Regarding the gettering technology, the stability of major contaminated metals, such as Fe, Ni, and Cu atoms near Si (100) wafer surface was clarified. However, recent high-performance devices including the latest CMOS image sensors involve complex surface structures including (110) and (111) surfaces. In this study, density functional theory calculations were performed to investigate the dependence of the stability of Fe, Ni, and Cu atoms on Si surface orientation. Our key findings are as follows. (1) The stability of Fe, Ni, and Cu atoms is in the order of (110) > (111) > (100) Si surface. (2) The stability of each metal is in the order of Cu > Fe > Ni regardless of surface orientation. (3) Fe, Ni, and Cu are unstable in the ultra-thin SiO2 film formed on the surface. These results suggest that controlling metal contamination, e.g., by using a chemical solution that removes the oxide film on the (110) and (111) surface, is very important in the manufacturing process for CMOS image sensors. In addition, thermal oxidation should be performed immediately when the Si surface is exposed.

Smartphone imaging technology and its applications

AbstractThanks to their portability, connectivity, and their image performance – which is constantly improving – smartphone cameras (SPCs) have been people’s loyal companions for quite a while now. In the past few years, multicamera systems have become well and truly established, alongside 3D acquisition systems such as time-of-flight (ToF) sensors. This article looks at the evolution and status of SPC imaging technology. After a brief assessment of the SPC market and supply chain, the camera system and optical image formation is described in more detail. Subsequently, the basic requirements and physical limitations of smartphone imaging are examined, and the optical design of state-of-the-art multicameras is reviewed alongside their optical technology and manufacturing process. The evolution of complementary metal oxide semiconductor (CMOS) image sensors and basic image processing is then briefly summarized. Advanced functions such as a zoom, shallow depth-of-field portrait mode, high dynamic range (HDR), and fast focusing are enabled by computational imaging. Optical image stabilization has greatly improved image performance, enabled as it is by built-in sensors such as a gyroscope and accelerometer. Finally, SPCs’ connection interface with telescopes, microscopes, and other auxiliary optical systems is reviewed.

Biologically Inspired Ultrathin Contact Imager for High‐Resolution Imaging of Epidermal Ridges on Human Finger

AbstractCompound eyes exhibit compelling vision schemes due to their unique visual ultrastructures. Here, a biologically inspired ultrathin contact imager (BUCI) for high‐resolution contact fingerprint imaging is reported. The BUCI contains ultrathin arrayed camera, light‐emitting diode arrays, and light guide plate (LGP). The ultrathin arrayed camera is fully packaged with microlens arrays, variable micropinhole arrays (MLA‐VMA), and a single complementary metal oxide semiconductor image sensor. The ultrathin arrayed camera shows a large field of view of 77.3° and a constant MTF50 of 89.4 cycles mm–1, relatively higher than a comparable single lens camera at the distance less than 36.8 mm. The channel images of a human fingertip are directly obtained by using frustrated total internal reflection and stitched to a single and high‐resolution full‐field image for a fingertip of 5.7 mm × 4.9 mm. The BUCI provides a new platform for high‐resolution contact imaging in biometric or healthcare applications.

A 2/3-in. 9.5-Mpixel CMOS Imager With High Frame Rate and HDR Capabilities

<italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">This study explains the architecture and readout schemes used to achieve a 2/3-in. 9.5-Mpixel complementary metal oxide semiconductor (CMOS) imager, with 3.6 Gpixels/s enabling high frame-rate operation in native UHD-1/4K ultrahigh-definition (UHD) resolution and full high-dynamic range (HDR) support. The presented imager demonstrates an architecture that supports high spatial and temporal resolutions and a high output data rate. To fulfill the transfer time requirements, an innovative row addressing scheme has been developed to toggle between the different rows</i> .

Compact computational spectrometer using a solid wedged low finesse etalon.

A novel, to the best of our knowledge, two-layer hybrid solid wedged etalon was fabricated and combined with a traditional imager to make a compact computational spectrometer. The hybrid wedge, comprised of ${{\rm Nb}_2}{{\rm O}_5}$ and Infrasil 302, was designed to operate from 0.4-2.4 µm. Initial demonstrations, however, used a complementary metal-oxide semiconductor (CMOS) imager and demonstrated operation from 0.4-0.9 µm with spectral resolutions ${\lt}\;{30}\;{{\rm cm}^{- 1}}$ from single snapshots. The computational spectrometer itself operates similarly to a spatial Fourier transform spectrometer (FTIR), but rather than use conventional Fourier-based methods or assumptions, the spectral reconstruction used a non-negative least-squares fitting algorithm based on analytically computed wavelength response vectors determined from extracted physical thicknesses across the entire two-dimensional wedge. This new computational technique resulted in performance and spectral resolutions exceeding those that could be achieved from Fourier processing techniques applied to this wedge etalon. With an additional imaging lens and translational scanning, the system can be converted into a hyperspectral imager.

Highly durable organic photodetector for complementary metal oxide semiconductor image sensors

Utilizing organic electronics compatible with conventional semiconductor fabrication processes is extremely difficult because of their low chemical resistivity and poor environmental durability. To preserve the intrinsic functionality of organic materials, only a few fabrication processes can be used. Moreover, it is essential to achieve process expandability and silicon-process compatibility to develop high-resolution electronics suitable for mass production. Therefore, we developed wet-process-compatible organic photodetectors by replacing the conventional shadow-mask process with photolithography. This suppresses particle deposition during the serial fabrication processes, providing high operational stability. The fabricated green organic photodiodes exhibit a low dark current (1.0 × 10−11 A/cm2) with high photon–electron conversion efficiency (EQE = 65%). The charge collection and charge separation efficiencies are stable (ηcc = 84.6% and ηcs = 97.7%, respectively). Moreover, the organic semiconductors are compatible with conventional wet- and dry-etching processes owing to thin-film encapsulation layers. Finally, the novel organic image sensor can withstand 500 h under 85 °C/85% relative humidity and 1000 thermal cycles (−55–125 °C). Because of its robustness and strong barrier properties, the novel process architecture reported herein can be extended to any organic electronic devices, including widely commercialized organic light-emitting diodes and organic photovoltaic devices.