# Can a fibre preserve the shape of the light being transported through it?

Can some fibres retain the shape of the light being transported through them?

If I emit radiation towards a sample (in the shape of an "S") and this is reflected back into a fibre, will the same image come out of the other side of the fibre? Will I obtain an "S" shaped beam on the other side?

• Yes, it can work provided that the reflection incorporates a phase conjugation and the fiber is a multi-mode fibre. Oct 11, 2022 at 12:38

A conventional fiber cannot do this because the modes inside the fiber determine the distribution of energy of transported light. Thus the output distribution is entirely independent of the input distribution (single mode fiber) or scrambled (multimode fiber).

For imaging, so called "coherent fiber bundles" are used. These are an array of tightly packed multimode fibers. Each fiber acts as a pixel and relays the intensity arriving on it to the backside. If you look at the back face with a camera you will see a honeycomb pattern (the thin gaps between cores) superimposed over whatever the fiber face sees:

Note that the number of fibers is typically limited (1000-10,000s), so do not expect an HD image.

• Why do multimode fibers scramble the input? Don’t they preserve the amplitude in each mode? I could see phase varying between modes but that’s ok as long as you don’t need coherent imaging I guess. Oct 11, 2022 at 13:46
• @Jagerber48 Randomized phase is basically a diffuser, so each mode having different phase means the amplitude cannot recombine to form the input waveform. Oct 11, 2022 at 19:10
• Imaging can be done with incoherent light. If the input light is incoherent then phase shouldn't matter for reconstruction. How can I image a lightbulb? Oct 11, 2022 at 23:07
• I'm imagining using a lens to focus an image onto the input facet of a multimode fiber, then a second lens or microscope to re-image the output facet of the multimode fiber onto a sensor or retina Oct 11, 2022 at 23:09
• @Jagerber48 Incoherent light and a system with random phase are unrelated concepts. An example of an incoherent light source is the sun. An example of an object with random phase is fog. You can take a picture with sunlight but it's still hard to see through fog. You can try to image through a multimode fiber but you won't see much (at least without a lot of processing for phase retrieval). Oct 11, 2022 at 23:13

No, consider that since fiber optics are flexible and operate via total internal reflection that the number of reflections undergone by an extreme part of the image is variable with the variable cable geometry, therefore a single fiber optic cable doesn't have a definite focal length, and cannot maintain shape.

• An optical fiber does not work with total internal reflection. That is an old idea based on ray tracing. The better more correct way is to consider the modes that propagate in the fiber. Oct 11, 2022 at 12:36
• @flippiefanus, but the modes don't propagate with the same speed so the odds of reproducing the same beam profile at the far end of the five are astronomically low. Oct 11, 2022 at 14:30
• Yes, but that does not have anything to do with total internal reflection. Oct 12, 2022 at 3:03
• It's not applicable, we're talking about transmission of an image, ergo, you analyze it with a principal ray, we're not trying to model loss, junctions, or anything in the time domain. I'm sure there's an argument that total internal reflection isn't the whole story with regard to dispersion either, but it's sufficient to argue from an understandable point of view that a single strand of fiber optic cable of any sort isn't suited for transmitting an image in normal curvy operation. Oct 12, 2022 at 3:16

Yes it can, an it is used in endoscopes, look it up in wiki https://en.wikipedia.org/wiki/Endoscope

• Endoscopes do retain the "shape", but an endoscope does not use a fiber. It uses either a series of glass lenses, or a fiber bundle. The OP asks specifically about a single fiber (in which the answer to the OP's question is "no"). Oct 11, 2022 at 11:22
• Oh, but it would be possible using a fibre bundle. Thank you very much for both of your answers. Oct 11, 2022 at 13:12
• Except for narrow diameter endoscopes, most now put the sensor in the tip of the endoscope. Reason for this is that 2k or even 4k sensors can be only a few milliliters in diameter, but a 2k fiber bundle would be 8 or 10 mm in diameter and relatively inflexible. A 4k fiber bundle would be nearly an inch, plus power an accessory channels. Not something you want in a colonoscopy if you can avoid it! Oct 11, 2022 at 23:07