SOI enables back-illuminated CMOS and CCD image sensors with higher performance at lower cost.
The consumer still and cell phone camera market is caught in a technology bind brought about by component pricing pressures. While consumers increasingly demand higher resolution (higher pixel density) cameras, the market is also very sensitive to manufacturing cost.
The core of every digital imaging system consists of either a CMOS or a CCD image sensor. These image sensors are employed in two possible configurations, front-illuminated or back-illuminated. Currently, lower performance, front illuminated sensors dominate the consumer imaging market due to the reduced cost. However, in more demanding industrial and scientific markets, back illuminated CCD imagers are preferred, as the premium for higher performance is justified.
Manufacturers have struggled with the fact that higher pixel density leads to larger chip size and potentially lower yield, ultimately resulting in higher manufacturing costs. To maintain the balance between cost and demand, image sensors need to have both higher pixel density and smaller pixel size.
However, shrinking pixel size limits the performance of image sensors. The ratio of light sensing area to signal processing area, known as fill factor, is reduced as pixel size becomes smaller. As more functionality is squeezed into each pixel, the fill factor reduces even more rapidly. Traditionally, micro-lenses have been used as a solution to achieve higher fill factor. Small pixels require more advanced micro-lens manufacturing techniques though, adding to the complexity of fabrication and increased final camera cost.
On the other hand, back-illuminated imagers offer 100% fill factor with substantially higher sensitivity. They also enable flexible metal routing for more effective circuit interconnection. Hence, they are ideally suited for smaller pixel arrays. Still, many challenges exist before mass adoption of backilluminated technology can become a reality. High volume manufacturability, cost, and pixel-to-pixel cross-talk issues all must be overcome.
Leveraging over 25 years of experience fabricating back-illuminated imagers, Sarnoff has been working to perfect a proprietary fabrication technique to finally break down these barriers. This innovative breakthrough is achieved through the use of customized SOI wafers and standard IC processing equipment. Imagers fabricated using this technology have a wide spectral response, from deep UV to near IR, with improved quantum efficiency. It also enables a compact optical module, an important aspect for today’s ultra-thin camera modules.
To make back-illuminated CMOS sensors more cost-effective and better suited for high-volume applications, Sarnoff has been adopting SOI technology.
A key step to achieve back-illuminated images involves thinning the silicon substrate in the range of 5 to 10 µm (sometimes even thinner) with extreme thickness uniformity. It is highly challenging to achieve such uniformity over the entire 200mm wafer and often results in lower manufacturing yield. Because the thinning is performed as the last step in the process sequence, any yield loss has a significant cost impact. As image sensor manufacturers migrate to 300mm wafers, this problem is expected to get even worse. Also, to obtain high and stable sensitivity, anti-reflective coatings need to be deposited on the thinned surface.
However, starting with an SOI substrate, one can use the buried oxide layer (BOX) as a natural etchstop. This simplifies manufacturing significantly and brings the cost down substantially. Also, this method allows easier scalability from 200 to 300 mm wafers. Careful choice of the buried oxide thickness can provide the desired anti-reflective coating. Additionally, use of SOI allows room for innovative design, which can offer many other benefits including lower power, and better parasitics.
This lower cost approach to back thinning is now available to sensor manufacturers and designers. The table below summarizes key challenges associated with smaller pixel imaging systems and how SOI and back-illumination helps in resolving them.