I am trying to couple a 532nm laser with a 200 ns square pulse into an optical fiber (fused silica with damage threshold ~ 500 MW/cm^2). At lower intensities, < ~30 MW/cm^2, I see about what I would expect to see at the output. However, as I increase the energy, I see the development of a temporally-based attenuation in the pulse; this continues to get worse at increasing energy (pictures attached). The structure is very repeatable in time and seems the be insensitive to alignment into the fiber. I was curious if this is a behavior anyone is familiar with? And if there is a solution?


  • There is no visible damage to the fiber upon inspection after running it at these energies.
  • The transmission through the fiber is ~60%, regardless of the intensity (should be ~95%).
  • The spot quality at the output of the fiber looks good. That is, we are not propagating much through the cladding, and the spot is fairly uniform in intensity.
  • This is not a detector feature. If I move the detector closer to (or further away from) the fiber output I see the overall amplitude of the pulse change but the structure remains unaltered.
  • If I switch to a shorter fiber (5 ft as opposed to 30ft) I am able to transmit more energy before this structure begins to appear

enter image description

  • $\begingroup$ Various questions: i) How have you coupled the laser to the fiber? ii) Is the back reflection to the laser sufficiently attenuated? iii) Do you have an optical isolator? iv) What is the coupling loss? v) Have you observed the laser pulse directly before the fiber at the various pulse intensities? $\endgroup$ – Massimo Ortolano Jan 8 '17 at 21:39
  • $\begingroup$ Thank you for the comment. i) The laser is coupled to the fiber with an AR coated 500mm focal length lens. ii) The fiber is not AR coated, and we have not considered any effects from a back reflection. iii) The coupling loss seems to be on the order of 40% - based on placing an energy meter at the input and then the output. iv) I have observed the pulse entering the fiber with the same photodiode detector, and it exhibits the structure that I expect - a 200 ns top hat. $\endgroup$ – Ben Jan 8 '17 at 22:00
  • $\begingroup$ Is the fiber end cut at an angle with respect to the fiber's axis or the cut is orthogonal? $\endgroup$ – Massimo Ortolano Jan 8 '17 at 22:03
  • $\begingroup$ The fiber is cut and polished normal, at the input and at the output. And no we do not have an optical isolator (sorry I missed that question). $\endgroup$ – Ben Jan 8 '17 at 22:06
  • $\begingroup$ Ok, on one side there seems to be multiple reflections which seem to be (roughly) compatible with the fiber's length, but in that case they should have been present also at lower intensities; on the other side you might have a problem of back reflection into the laser. Beware that if you have a diode laser the back reflection can even damage it. $\endgroup$ – Massimo Ortolano Jan 8 '17 at 22:11

I'm not 100% sure this is the reason, but I figured I'd post it in hopes someone could comment on it.

This looks like it could be a result of stimulated Brillouin scattering. Such that there are either ion acoustic waves being excited or charge depletion sites causing back-scattering in the fiber. This behavior is very non-linear in intensity and fiber length, and that seems consistent with what we see. Additionally, Figure 5.5 in here shows something pretty similar to the temporal structure we observe with a single pulse.

I haven't found any references explicitly talking about 532nm light under these time scales in a multimode silica fiber, however. So I am not certain this is the culprit.

I also don't know if there's a way to fix it if this is indeed the case.


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.