How camera sensors are made

Half the colored filters over a camera sensor are green, to match how human eyes work.

stations: 6
things upstream: 36
raw sources: 15
steps from the ground: 7

The short version

A camera sensor is a silicon chip made on a wafer, like a processor. Each pixel gets a photodiode and transistors, and a checkerboard of red, green and blue filters is laid on top. Wafers are sawed into dies, and each die is mounted in a package under protective glass.

Build the pixel array

Millions of pixels are patterned on a silicon wafer. Each has a pinned photodiode that collects light as charge and a few transistors to amplify it.

Goes in
Silicon wafer
Wire the readout

Metal layers connect every pixel to row and column circuits, so the whole array can be read out line by line.

Goes in
Copper
Lay the color mosaic

A Bayer filter is laid over the pixels. It is half green, one quarter red and one quarter blue, so each pixel records just one color.
Probe and saw the wafer

Dies are electrically tested, then separated by sawing.
Package under glass

Each die is wire-bonded into a package and sealed with a clear glass window.

Goes in
Glass
Check image quality

Sensors take test pictures so defective pixels can be mapped and corrected in software.

End of the lineCamera sensor Next it goes into smartphone, camera drone, satellite and 1 more.

Worth repeating

Bryce Bayer's 1976 patent used twice as many green elements as red or blue to mimic the human eye.
The pinned photodiode was invented at NEC in 1980 and is used in nearly all CCD and CMOS sensors.
Each pixel records only one color, and the other two are filled in by interpolation.