It seems to me that as the wavelength of the light gets longer than the object one is trying to view, the image should start blurring out, however i'm not sure if quantum mechanics makes this impossible, but would taking say 100 low quality separate images of the same blurry molecule be reconstructable into 1 HQ image of the molecule without the noise factor having too much of an influence?
Dear Tristan Sachsenweger Ballantyne!
When you try to obtain sub-wave length resolution, the captured image is distorted by diffraction.
Counterexample:
https://en.wikipedia.org/wiki/Diffraction_formalism#/media/File:Diffraction2vs5.jpg
This is diffraction on slits. As you can see, there are no difference between this picture and picture of imaged on screen laser-spots' array. So, there are no bijection: two different objects gives identical images. As corollary, there are no image restoration algorithm, which can restore "molecule image".
However your idea is very interesting. It sound similar to this method
https://en.wikipedia.org/wiki/X-ray_crystallography#/media/File:X_ray_diffraction.png
and, probably, you can achieve success by using image restoration algorithms to X-ray range.
Sincerely,
Daniil D. Stupin
Dear Daniil D. Stupin,
Thank you for clarifying that there is no bijection between the image and a molecule for longer wavelengths. Combining image restoration algorithms and X-ray diffraction seems to be the way to go.
Cheers,
Tristan Sachsenweger Ballantyne
Dear Mikkel Schou Nielsen, According to the nature paper, they currently achieved a half-pitch resolution of 154nm using a wavelength of 435nm light source using FPM, which so far seems to be mainly limited by the tech required to manufacture accurate and compact led arrays. With proteins being in the size range of ~10nm It would be interesting to see if it could resolve individual proteins using ~30nm UV LEDs.
Dear Colleagues!
I agree with Dr. Mikkel Schou Nielsen message about possibility of obtaining resolution in times better than wave-length. My counterexample just demonstrate, that it is impossible to obtain arbitrary high resolution (in conventional optical microscopy resolution is limited).
About suggestion of the usage 30-nm wave-length.
Unfortunately, there are no LEDs on such wave-length, because it corresponds to 41 electron-Volt photon energy. No one semiconductor has such energy gap. However, Free Electron laser could be used to obtain 30-nm radiation.
But even in this case we will face new difficulties. Namely, there are no lens for so short-wave length light, and, for my best knowledge, only diffraction mirrors could be used for creating optical image. Finally, so high energy photons can damage protein molecule.
To my mind, for study protein molecule with visible light range it is better to use Near-field scanning optical microscope
https://en.wikipedia.org/wiki/Near-field_scanning_optical_microscope
in which is possible to overcome diffraction limit. It should be noted, that this method is different to optical conventional microscopy, because Near-field scanning optical microscope scan sample, and image is collected line-by-line, contrary to optical microscopy, where image is captured simultaneously full.
Sincerely,
Daniil D. Stupin
- Badges
- Science topic
- Similar topics
- Computer Science
- Images