For a given wavelength $\lambda$, is there a limit on the size of the sensor pixel, regardless of the optical equipment attached to the sensor or the nature of the censor?

  • $\begingroup$ Thanks to the quantum mechanical nature of light deep-sub-wavelength optical system can be built quite successfully. The quantum efficiency of individual pixels in modern detectors is high, and as pixels are getting smaller, each pixel has to accumulate or count ever smaller numbers of photons (the total photon count rate is set by the total power of the incident light) which can, at least theoretically, be used to increase the dynamic range of the detector. Ideally each pixel would also record the wavelength and timing of the photon, which would allow for multispectral high speed imaging. $\endgroup$ – CuriousOne May 12 '15 at 8:54
  • $\begingroup$ @CuriousOne: is there a limit on how deep sub-wavelength is theoretically attainable? Rayleigh limit probably doesn't apply because one theoretically can create a lens with arbitrary diameter. Still, can one squeeze a wave of length $\lambda$ into a pixel of size, say, $\lambda/100$? $\endgroup$ – Michael May 12 '15 at 13:57
  • $\begingroup$ One can take this to atomic resolution, below that there is no material to build optics from, of course. Look at it this way: single atoms can emit and absorb photons. That's the limit because nucleons do not respond to optical frequencies except under unrealizable magnetic fields (if you manage to make a local field of 1e10-1e11T, the game changes, again). So as long as you can devise techniques (like plasmon resonance) to make sever atoms or electrons on atomic distances in a metal to do something at the frequency of visible light (or IR), you have some form of "optics". $\endgroup$ – CuriousOne May 12 '15 at 20:12
  • $\begingroup$ Although in theory one might be able to go down to atomic size, in practice, the limit will be reached at a much larger size. This limit will be set by how small can a CMOS crystal be made and how "cleanly" can its small signal be amplified and separated from "noise." $\endgroup$ – Guill May 14 '15 at 7:44
  • 1
    $\begingroup$ @uhoh, I am one of those engineers you were "making miserable in the process". :-) BTW, I didn't say the industry was breaking the laws of Physics, only that the industry was breaking the barriers that were considered the laws of Physics just a few years ago. Things like printing deeply sub-wavelength features, you know, $\endgroup$ – Michael Sep 7 '16 at 14:56

For visible photons, I believe this is really a quantum mechanical question. The absorption of light into a detector is a quantum phenomena. It some way, it is the coupling of the frequency of light with a dipole in the atomic media. That dipole moves with the frequency of the light and this coupling can cause absorption and photoelectrons to be formed as the excited electron decays.

From this viewpoint, they limit of absorption is based upon the coupling efficiency as a function of the confinement of the optical field.

This is not restricted to quantum mechanics. One such classical dipole is your TV antenna. It is much less than $\lambda$ in size, yet it absorbs RF photons. Patch and coiled antennas have a lateral dimension even smaller than a wavelength.

Metamaterials have been made from split ring resonators which form capacitive and inductive features much smaller than $\lambda$.

  • $\begingroup$ A dipole antenna has a physical length which is slightly less than half a wavelength and acts as a resonant element. $\endgroup$ – Farcher Mar 17 '16 at 14:36

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.