How do CMOS image sensors work?
Article By : In-Chul Jeong
How do CMOS image sensors work and which process techniques matter in their development?
Innovations in CMOS image sensor (CIS) technology continue to enhance digital imaging landscape. While the demand has been driven by smartphone makers, leveraging enhanced photo-taking capabilities to differentiate their devices from the competition, there is also a growing market for applications in the automotive, security, medical, and manufacturing sectors.
The tiny CMOS image sensors, comparable in function to the human eye’s retina, now rival what could only be done before with large, expensive camera equipment. And, compared to smartphones, the new applications are putting an even greater emphasis on the need to advance CIS technology.
Therefore, instead of merely capturing an image for viewing by the human eye, CIS technology is now capturing data to power a host of new use cases from autonomous vehicles and virtual reality (VR) to next-generation medical imaging and high-tech surveillance systems.
Figure 1 The demand for advanced CIS technology comes from a range of applications. Source: SK Hynix
So, what needs to happen to ensure CIS that technology keeps up with this demand for more advanced applications? First, let’s take a quick look at how CIS technology works. Then, we’ll highlight the five categories of manufacturing process technology unique to CIS that will require continued advancements.
How CIS works
At its most fundamental level, CIS technology is tasked with converting light from the camera lens into digital data to create a picture of what’s in view. When light energy in the visible light wavelength range of 400 to 700 nm is condensed on the photodiode (PD) of the silicon substrate, the silicon surface of a CMOS image sensor receives the light energy to form an electron-hole pair.
The electron generated in this process is converted into voltage through floating diffusion (FD) and then into digital data through an analog-to-digital converter (ADC). The data is sent to a processor to create a digital description, usually an image, of what’s in view.
CIS manufacturing techniques
Producing such a sophisticated sensor requires specific manufacturing technologies that can be classified into five categories.
1. Deep PD formation process technology
The persistent consumer demand for enhanced image quality has led to a competition on how to increase pixel density and resolution in mobile CIS, which in turn, has further accelerated the development of CIS process technology. To get there, the pixel size needs to be reduced further to accommodate a greater number of pixels on the same-sized chip.
Figure 2 This schematic diagram shows photodiode structure change along with reduction in pixel size. Source: SK Hynix
A deep photodiode is also critical to avoid deterioration in image quality. To secure a sufficient full well capacity (FWC) in small pixels, it requires patterning and implementing technologies with difficulty levels that go beyond existing semiconductor memory. For that, it’s essential to secure a high aspect ratio of more than 15:1 to implant mask process technology that can block high-energy ion implantation, following the evolving industry trend of ever-higher ratios.
2. Pixel-to-pixel isolation process technology
The technology to isolate pixels from one another is crucial when it comes to high-definition CIS. Chipmakers utilize different isolation technologies. Using one that’s less developed could introduce image defects such as color mixing and color spreading.
Increasingly, as higher pixel density and resolution become common requirements, isolation becomes an important criterion for image quality in the CIS market. Beyond that, issues have also been known to occur during the isolation process and, for that reason, efforts are being made to select better equipment and develop new recipes for improving yield and product quality.
3. Color filter array (CFA) process technology
Color filter array (CFA), a process unique to CIS domain, is not common in the semiconductor memory processes. The CFA process generally consists of a color filter (CF) that filters the incident light into red, green, and blue for each wavelength range and a microlens (ML) that boosts condensing efficiency. To create robust image quality, it is important to evaluate R/G/B color materials and develop technologies that optimize parameters such as shape and thickness.
Recently, a series of high-quality and highly-functional CIS products have been released; they are based on technologies such as Quad Bayer, and they are complemented with the basic form of CFA.
Figure 3 A color filter array is comprised of a color filter and a microlens. Source: SK Hynix
4. Wafer stacking process technology
Wafer stacking—literally, attaching two wafers together—is an essential technology for producing high-pixel and high-definition CIS products. For high-pixel CIS products, pixel arrays and logic circuits are formed on individual wafers separately, which are then attached during the middle of the process by using a technique called wafer bonding.
Figure 4 Wafer stacking significantly improves CIS capabilities. Source: SK Hynix
Wafer stacking technology, which has been adopted by most CIS chipmakers, is continuously evolving in various dimensions.
5. Control technology for CIS yield and quality
One of the most fundamental requirements in the CIS product development and mass production process is the control of metallic contamination. Since CIS products are sensitive to contamination several times more than memory products and the contamination directly affects product yield and quality, various contamination control technologies are mandated.
Beyond that, plasma damage control is also an important factor. Since the deterioration of image properties such as hot pixels occurs due to the damage caused during the process, it’s necessary to manage key processes accurately.
CIS’s future outlook
It is no exaggeration to say that the effectiveness of CIS-driven applications will be determined by the process technology. The way these individual processes interact with each other will also play a major role. It’s not enough to optimize a single aspect of the manufacturing process; they have to be optimized to organically complement each other.
The payoff, though, is immense. Virtually every business segment from manufacturing to healthcare services to surveillance can take advantage of new CIS technology to improve operations. With a richer, more detailed view of the world, companies across all industries will be able to create more intelligent and sophisticated products and services that benefit their end customers and society as a whole.
This article was originally published on EDN.
In-Chul Jeong is CIS process team leader at SK Hynix.
Related articles:
- Future of Vision: Sensor Tech Is Shifting from Images to Information
- Digital camera design, part 5: Basic noise considerations for CMOS image sensors
- Image sensor defects can ruin an image
- Autonomous Vision Chip for IoT Applications
- The Slow But Steady Rise of the Event Camera
- 10 image sensors to consider for your next industrial, medical, or smartphone design
