"Incident light passes the first organic layer, which absorbs only the blue light component and converts to an electrical signal, and is transparent to the green and red components. The second organic layer absorbs only the green component, and the red component is detected by the CMOS image sensor. The organic layers are combined with transparent thin-film transistors, and the signals output from each of the layers can be combined to reproduce a color image.
This structure enables all color information of red, green and blue to be obtained within a single pixel, achieving a high-resolution image sensor that uses light more efficiently. We will continue to work reducing the pixel size and increasing the number of pixels, and accelerate R&D toward realizing a compact, high-resolution, single-chip camera."
Background on the Foveon X3 for people new to solid state image sensors (color separation based on absorption depth in silicon, not a color filter array). https://en.wikipedia.org/wiki/Foveon_X3_sensor
ReplyDeleteI think everyone in the image sensor community is very familiar with the pros and cons of the X3 and earlier (and later) approaches. But thanks for sharing the Wikipedia version.
DeleteI did not know about the Wiki page of the X3. I quickly checked it, and what I expected came true : the original "inventor" is not mentioned at all. It was dr. Wolffenbuttel of Delft University of Technology who demonstrated the technology for the very first time, just with a single pixel if I recall well. Foveon used the same concept and made a product out of it.
DeleteThanks for sharing Albert - I always thought that came from Merill and Mead
DeleteYes, Wolffenbuttel - 1994 https://doi.org/10.1016/0924-4247(94)80099-5 based on selective epitaxy, also US Patent 4,581,625 based on a-Si overlayers filed in 1983, and US 5,883,421 filed in '97 then Merrill's US 5,965,875 filed '98.
DeleteCertainly Merrill et al. truly refined and commercialized the "X3" technology. BTW, while at JPL (94-95), we transferred the JPL CMOS active pixel sensor technology to Merrill at National Semiconductor. He left a few years later to start Foveon. Good guy.
Can anyone comment on pixel size, 20um? It is fairly big comparing to current market trend. Is it because of 3-layer approach?
ReplyDeleteI think the large 20 µm pixel size is likely a consequence of two factors:
ReplyDelete1. Display-oriented TFT/OLED device design constraints.
2a. The supportable thickness of the blue and green channel organic films.
2b. The quantum efficiency achieved by the B & G films.