Sony 3-Layer with Shield Patent

Just got this patent application which give also useful information about the technical details of the issues valid for all 3-layer sensors.

http://www.freepatentsonline.com/20130009263.pdf

As I considered before the foveon sensor have much less color separation which requires strong off-diogonal color correction for sRGB. That result in more noise because of taking the difference between the layer channels. Color inaccuracy to sRGB with nonstandard spectral light sources or mixed light. Layer channel interaction because the depletion zones extend to each other if a channel goes to saturation. And finally the spectral sensitivity depend on light incidence angle which mean apture dependence. All that need to processed or corrected before jpg in camera or in raw conversion.

The sony patent application 20130009263 only address with two additional shield layers the nonlinear interaction between the layers. So from engineering perpective there is a decision to have either a complicated 5-layer buried implant stacking or more numerical processing. All other issues remain and in addition the two shield layers loose photons if taken to VDD.

Some idiotic question could be: If there are 5-layer deep implant technology why not use this for improved 5x3 color matrix conversion instead of shielding.

To me it seems that sony proves all foveon based alternatives to have an edge to Canon/Nikon/Toshiba/Fujitsu/ST/Aptina.

Here some statements:

[0043]Similarly, since the charges generated by photoelectric conversion in the region between the photoelectric conversion regions 23G and 23R can be taken into and discharged through the discharge region 24, the solid-state imaging device 21 can prevent crosstalk from being generated between the G and R signals.

[0048] Thus, the solid-state imaging device 21 can further separate the colors from each other and enhance the color reproducibility than in the related art. Accordingly, the solid-state imaging device 21 has better color separation performance and can improve the color reproducibility of captured images.

Better separation on the cost of lost photons in the shielding. The basic low color separation remain.

​[0136]FIG. 14A shows the color arrangement of the color filter 72 formed on the solid-state imaging device 71. The color filter 72 has magenta filters 72MG transparent to the light of blue and red wavelength bands and green filters 72G transparent to the light of green wavelength band, arranged in a checkered pattern. Accordingly, the solid-state imaging device 71 includes magenta pixels for outputting signals corresponding to the charges photoelectrically converted from the light of blue and red wavelength bands and green pixels for outputting signals corresponding to the charges photoelectrically converted from the light of green wavelength band.