I'm afraid there's no hard evidence as of 100% fill factor. This is rather a target figure for all manufacturers and it only takes a bit of wishful thinking to jump to conclusions. The sad fact is that the active part of the pixels only gets around 30% of the light. Microlenses, traps for stray photoelectrons and other solutions to circumvent this only artificially increase the fill factor, adding noise as well. The overall result needs to be heavily processed in order to present the beautiful, uniformly lit, grainless surfaces that people take as a digital attribute.
Servus, Alin Ryan wrote: RKB> No, not really. The huge majority of area on a ccd is sensitive to RKB> incoming light, even moreso with modern cmos. Just because the diode RKB> junction is sometimes small doesn't mean it doesn't "see" everything RKB> thru the microlens. RKB> e.g. The DCS-14n chip uses the entire p-junction base (or back if you RKB> prefer) of the semiconductor to detect light, and therefore claims a RKB> 100% fill factor. I assume this is even without the microlenses, but RKB> even the crappiest interline chip gets 70% these days. (which would be a RKB> tiny chip anyway, with small sites, hence small distances) RKB> Here's someone else that agrees with me: RKB> http://www.binbooks.com/books/photo/i/l/541E6AF912&orig=1 RKB> The 1Ds also claims a 100% fill factor. And 100% of 8.8 microns doesn't RKB> leave a lot of "dead space". RKB> dpreview has a dated discussion of this, but any google search can show RKB> you the newer chips have 100% coverage: RKB> http://www.dpreview.com/learn/Glossary/Camera_System/Sensor_01.htm RKB> So if we agree that anything much less than 400nm (.4microns) is below RKB> visible, then a (big pixeled) 8 micron image sensor with 95% or more RKB> fill factor has a suboptical distance between it's 'eyes'. RKB> I'd argue that this gap is small enough not to matter since it's built RKB> to a far better tolerance than the camera lenses anyway. Someone care RKB> to compare this numerically to the boundaries between film grain of an RKB> E6 emulsion? RKB> This is getting like the misinformation about edge fall off because the RKB> wells are too deep that we had last year. Where's that argument now? I RKB> still stand behind my yield statement. RKB> "Oh no, the sky is falling, and the camera is making up information RKB> between pixels".
