Quote from the page: "The heart of the detector is made out of a thin film of aluminum on a sapphire substrate. The aluminum is etched by standard photolithographic processes to form a meandering strip. When cooled to near absolute zero (less than one kelvin), the aluminum becomes superconducting. Like the vibrations of a tuning fork, current in the aluminum strip oscillates at a resonant frequency."
Seems like even the technology matures, we won't be seeing it implemented on DSLRs...unless scientists can find another way to tune CCD pixels to particular wavelengths at room temperature.
Anyway, since each pixel can only be tuned to 1 wavelength, Bayer mosaic is still required for single layer sensors. The only thing that can be eliminated is the RGB filter array that is used in conventional CCDs. The colour for each pixel still has to be interpolated from the surrounding pixels.
Yup... as mentioned in my first post.. the catch is 0K degrees.
On the net is not the complete article... what it mention is that it can tune to a single wavelength of a single photon... it implied "automatic" tuning for subsequent photons received. But the problem is that our currently pixel size... it can receive hundreds or thousands of photons at the same time so need some kind of averaging algorithm here.
Also, no interpolation is needed if the frequency data is embedded in the pixel output.. cuz the frequency IS the actual (reflected) colour.
This brings me to my thread implied intention... it means that other then Fevron 3 layer CCDs... there is another (currently available) technologically feasible way to determine colours per pixel location... theorically it's possible to build an "antenna" on the CCD (let's say by sacrificing half of every alternating pixles) this not only gives you added dynamic range (similar to fuji's CCD technology) but gives you the resolving power of Fevon types chip... all this without the complication of Fevon chips.