Microsoft Word-World first mass productive 0.8ã”ł pixel size image sensor with new optical isolation technology to minimize optical loss for high sensitivity.docx

More and more pixels are needed to implement CMOS image sensors with high resolution while reducing the size of the chips. As the beam size of the incoming light nears the pixel size, the light loss from the conventional metallic grid increases. Development of an architecture that minimizes optical losses is key technology to secure pixel performance of sub-micro sized pixel. In this paper, we analyzed the light loss according to the pixel size and the optical structure. Based on this, we are going to introduce a new isolation structure that minimizes optical losses applied to the world's first 0.8um pixel size product. Introduction One of the most important features of the smartphone market in recent years is the hiring of high-resolution multiple cameras. Competition to decrease size of pixels to implement high resolution is accelerating even while area of image sensors is being controlled as much as possible. Pixel size has declined significantly over the years, image sensor products with sub-micrometer sized pixel coming to the market.[1] As pixel size is decreasing continuously, it is expected that pixel size will decrease as well as visible light wave level.[2] It is very difficult to maintain sensitivity of small sized pixel due to reduction of photons that goes through micro-lens and limitations of light diffraction. For these reasons, the technology to maximize the quantum efficiency of pixels in order to maintain picture quality in image sensors is important. We made a conclusion from the light loss analysis in image sensors that the metallic grid structure commonly used is not suitable for small sized pixels. In this paper, a new structure with reduced light loss was proposed. Analysis and result Light loss in common image sensors is affected by the light collection efficiency of the micro-lens, the color filter transmittance, the structure of the metallic grid structure, and the performance of the anti-reflective layer of silicon. We analyzed the major light loss components through 3D finite-difference time-domain (FDTD) simulations, color filter properties, and device spectrometry. Table1 is a summary of the analysis results, and at 0.8um pixels it was shown that the major light loss component was caused by the metallic grid structure. Micro lens Color filter Metallic grid Anti reflective layer normalized light loss ratio 3.8% 21.4% 71.4% 3.3% Table 1 Summary of major light loss components at 0.8um pixels We prepared three structures, such as Table2, which experimentally demonstrates the light loss rate from metallic grids. Structure 1 is a typical image sensor, and metallic grids were removed in Structure 2. By comparing these two structures, the R04