Toevolution

With the touch of our fingers, we can capture the world using image sensors, semiconductor devices. The size of image sensors has likewise shrunk to the size of our fingers as a result of advancements in semiconductor technology. In 2022, there will likely be more than 1.7 trillion photographs taken. This demonstrates the pervasiveness of image sensors in our daily lives. These sensors can now record tiny and macroscopic information that the human eye is unable to record. The image sensor's main job is to collect light from the environment and transform it into electrical signals. Based on the architecture for converting light to electrical signals and the device structure, image sensors can be divided into Charge-Coupled Devices (CCD) and Complementary Metal Oxide Semiconductor (CMOS) image sensors. The latter are utilised for photographing and filming in all types of digital cameras.

This article will cover the development of image sensors from CCD to CMOS, its key features and benefits, applications, the distinction between front-side and backside illuminated CMOS image sensors, key market competitors, and leading image sensor patent assignees.

A Brief History of Image Sensors

MOS image sensors have been around for a while. When the vidicon tube was the standard technique for capturing images and video in the middle of the 1960s, they were initially designed and invented. Although the image quality was superb, there were a number of drawbacks, including its enormous size, sluggish video, and heavy weight. It additionally experienced significant fixed patterns and temporal noise.

In an effort to fix the MOS image sensor issue, CCD image sensors were created in 1969 and swiftly took over the market. In a CCD sensor, each pixel charge is converted to an analogue voltage, amplified, and digitalized via an output. The usage of MOS image sensors had been limited to a few specific applications throughout the 1970s, such as spectroscopy, while CCD dominated the market for image sensors because of its intrinsically lower fixed-pattern noise (FPN). The CCD market was monopolised by a few number of businesses, which made it difficult for universities and businesses to do CMOS image sensor (CIS) research.

CIS overcame the drawbacks of CCD image sensors, including their high energy consumption, sluggish serial data stream reading speed, low uniformity and image quality, poor link performance, and short battery life. On the CMOS sensor chip, the picture data is converted into digital data when each pixel in a CIS goes through a charge/voltage conversion. The benefit of parallel picture data read-out from a CMOS sensor allows for a higher frame rate at equivalent resolutions.

Faster frame rates are possible with CIS because it has fewer analogue to digital conversions (ADC) per pixel column than CCD. The most recent CMOS sensors have significantly improved over time, and they are now on par with or superior to CCD in terms of overall value, image quality, and imaging speed. The CMOS image sensors also enable scaling by enabling the integration of image acquisition and processing circuitry on the same chip using a conventional CMOS process.

After learning about the development of CIS and the history of image sensors, let's talk about the key elements that make this technology perform so well.

Critical Components of a CMOS Image Sensor

A CMOS image sensor contains four crucial parts:

• Interconnects with Microlens 
• Color Filter (CF) 
• Transistors
• Photodiode (PD) (PD)

A CIS consists of a collection of tiny lenses that focus light onto photodiodes. The colour filters divide the RGB (red, green, and blue) components of directed light. A photodiode (PD) is an apparatus that gathers photons or catches light and transforms the gathered photoelectrons into electrical signals. The signals that are conveyed inside the image sensor by connecting structures or metal wiring are controlled by the pixel transistors. Based on how the photodiode and connection layers are arranged in relation to one another, the CMOS image sensors can be divided into Front-Side Illuminated (FSI) and Backside Illuminated (BSI) types.

We shall go through the benefits of CMOS sensors over CCD image sensors in the next section.

Advantages of CMOS Image Sensors

Using CMOS image sensors instead of CCD image sensors has many benefits. The following discussion covers the four key benefits:

A CMOS image sensor has a much lower energy usage. It is essential because less heat dissipation results in a better camera with lower noise.

Due to the improved pixel architecture, it offers a better sensitivity, which is advantageous in low-light conditions. Short-wave infrared radiation and the near-infrared range are more sensitive.

Less dark noise will lead to higher image fidelity.

Because the read-out logic may be incorporated into the same chip, the camera size may be lower.

CMOS Image Sensors: Front-Side Illuminated (FSI) vs. Backside Illuminated (BSI)

The front-side illuminated (FSI) architecture used by standard CMOS image sensors starts with a microlens array at the top, followed by colour filters, an interconnect layer beneath the colour filters, and photodiodes at the bottom. Light reaches pixels in FSI image sensors through the front metal wiring and focuses on the photosensitive area (photodiode). The photosensitive area is essentially guaranteed because for larger pixels, the ratio of the height of the optical stack to the pixel size is small. As a result, the FSI image sensors perform better with bigger pixels.

There was a push to make pixels smaller as the industry worked to make CMOS image sensors smaller or pack more pixels into a given area to improve resolution. Unfortunately, the shrinking pixel size resulted in a decreased fill factor, a longer optical path, and a reduction in the light absorption by metal wiring, which hampered the performance of the FSI sensor.

The photodiode layer was positioned above the connection layer in the BSI (Backside Illuminated) architecture to overcome the issue. The absorption, reflection, and flare of the FSI metal wiring layer are eliminated because it separates electrical components from light and enables the optical route to be modified independently. Additionally, the optical stack of the BSI pixel is significantly smaller and produces a greater QE (quantum efficiency) than the FSI.

BSI for Today and Tomorrow

Either an increase in performance or a decrease in size of the image sensors is required.

The supporting substrate in traditional BSI sensors was replaced by a pixel layer that was built over a chip in the stacked sensor architectures that the BSI sensors eventually developed. The stacked sensor architectures made it possible to place complex circuits on a tiny chip. Since the pixel and circuit sections are created on separate chips, specialised manufacturing techniques are employed to create high-quality pixel and circuit sections, allowing for better resolution and a smaller overall size.

To connect the pixel and circuit die, these stacked BSI structures made use of through-silicon vias (TSV) in the pixel array's perimeter. Cu-Cu direct bonding between the stacked pixel and circuit dies was developed as a result, allowing for the decrease of the size of the sensors and the saving of space in the peripheral.

The emphasis on self-driving cars has stimulated extensive research and development in areas including time of flight (ToF), single-photon avalanche diode (SPAD) pixels, and the integration of artificial intelligence (AI) into image sensors. While AI provides quicker processing of the acquired image for pattern/object identification within the sensor, ToF and SPAD enable measurement of an object's distance from the detector (such as a car, person, road sign, etc.).

Smaller pixels in CMOS image sensors are advantageous due to higher resolution, minimal power consumption, and reduced costs. These advantages continue to be driving forces behind current and upcoming CIS technology improvements, and they are anticipated to accelerate industry expansion.

Due to their various advantages, CMOS image sensors are used in a wide range of products and sectors. Therefore, the focus of our next discussion will be on its practical applications.

Areas of Use for CMOS Image Sensors

Numerous industries, such as mobile, computing, security, healthcare, automotive, and defence, use CMOS image sensors. Image sensors have seen a dramatic transformation as a result of the inclusion of camera systems in mobile devices and the rising consumer demand for smartphones. In order to deliver high-quality photographs while minimising the size of the sensor module within smartphones, image sensors continue to be an area of constant development. In addition, image sensors are used in smartphones for a variety of security applications, such as facial recognition, in-screen fingerprint recognition, iris recognition, etc., that enable user authentication.

Based on research in the area of autonomous vehicles that employ a variety of image sensors for the development of Driver Monitoring Systems (DMS) and Advanced Driver Assistance Systems, CIS has lately found a substantial application in the automotive sector (ADAS).

Drones, robots, virtual reality, augmented reality, mixed reality, and consoles with gesture controls are some devices that can use image sensors.

Key CMOS Image Sensor Market Trends

• Digital single-lens reflex (DSLR) cameras, cellphones, and tablets all use CMOS image sensors in various ways. In the upcoming years, the market for CIS is anticipated to remain consistent as smartphone manufacturers concentrate on enhancing the camera quality.
• In addition, it is anticipated that the trend toward dual-camera smartphones across all major smartphone suppliers will support demand. Two sensors are necessary for the underlying technology of dual-camera smartphones. They work together to create a single image by combining the coloured and monochrome images produced by the many CMOS image sensors. Customers no longer consider two cameras to be sufficient, as Samsung's Galaxy S22 Ultra has five total cameras—four on the back and one on the front.
• The inclination of the customer for dual and 3D cameras will have a big impact on CMOS volumes. In addition, it is projected that emerging technologies like telescopic lens drones, virtual reality, robots, and augmented reality will fuel this expanding business.
• Consumption of consumer electronics has increased on a global scale. The retail trade revenue for consumer electronics and home appliances in China totaled CNY 96.31 billion in June 2019, according to the National Bureau of Statistics of China. The market for CMOS image sensors is anticipated to be driven by a variety of electronic products including cameras.
• In the next years, the market is expected to be driven by the Asia-Pacific region. Because of its expanding middle class and high demand for consumer electronics, China is expected to lead the increase. According to data from the National Bureau of Statistics China, the country's market for manufacturing computers, communication equipment, and electronic devices grew by 13.1 percent in the fiscal year 2018.
• The number of smart city projects in India is steadily expanding thanks to government initiatives. The requirement for electronics solutions, such as surveillance, monitoring, and maintenance, will become increasingly important, driving the demand for CMOS image sensors.
• The healthcare sector will also benefit from the use of CMOS image sensors as non-invasive testing tools for scanning. By 2023, it is expected that India's healthcare sector will be worth roughly USD 133 billion.
• By 2030, it is anticipated that the Asia-Pacific area will contribute more than 25% of all smart homes sold worldwide, with sales rising to USD 120 billion. Japan is currently in front of the field. However, the region's demand for CMOS image sensors will increase as smart homes become more popular.
• We will concentrate our discussion on the market's top players globally in order to gain a better understanding of the current and future situation of the CMOS image sensor market.