After amplification in optic fiber amplifier pulse autocorrelation (red line in the attached picture) width decrease as compared to initial pulse autocorrelation (black line). Fiber amplifier dispertion is normal, pulse chirp is positive. Pulse amplified spectrum width about 12 nm, output averaged power 2 W. What possible reason of pulse width shrinkage ?
What is the duty cycle or pulse rate and wavelength? What is the spectral width of the input waveform? What is peak power, gain coefficient and fibre effective area within the amplifier?
With 2 W average power, the peak power could be sufficient for substantial chirp from self phase modulation. The amplifier dispersion is normal, but depending on the gain coefficient and active fibre length, only in the last fraction of a metre of fibre will the optical power and SPM-induced chirp be sufficiently high for appreciable pulse broadening.
What is the dispersion of fibre within the amplifier after the gain section, and of any patch cords between amplifier and autocorrelator? With output pulses chirped by 12 nm, a few metres of standard single mode fibre with anomalous dispersion could be sufficient to cause the pulse compression you observe.
Hi Alan,
I attached pulse spectrums before amplifier (black line) and amplified (red line). Pulse repetition rate 23 MHz, input power 10 mW, output 2 W. In set-up wasn't included any anomalous dispersion elements, so pulse compression by wavelength dependent delay is excluded. So pulse width only increasing is supposed as pulse amplified in all-normal dispersion scheme.
My only guess for such effect of compression is a some pulse profile distortion as it amplified. I simulated that case but always amplified pulse autocorrelation function is broader as initial. So pulse always uniformly (in correspondence of every time point in it profile) amplified. I assumed parabolic gain profile with width 20 nm.
Arkadiy,
There seems to be quite a lot happening. The output spectrum is broadened, presumably due to Kerr effect non-linearity in the fibre - self phase modulation (SPM) etc. There is also a significant shift of energy from short to longer wavelengths.
The input spectrum is fairly regular, with narrow peaks at 1034 and 1037.5 nm, and a weaker peak near 1036.6 nm. The output spectrum is more complex, with satellite peaks roughly +/-1 nm from the original narrow peaks, and a new strong and broader peak at 1040 nm. The spacing is suggestive of a process such as four wave mixing (FWM) creating new spectral components.
Where is the amplifier gain spectrum centred and what is its spectral profile? Unlikely to be the cause of the frequency shift, but worth checking if you have not done so already.
A more likely cause is stimulated Raman scattering (SRS) in the fibre. This couples energy from short to longer wavelengths, transferring some of the photon energy to lattice vibrations. The strength of the Raman coupling increases with frequency separation up to around 13 THz or 50 nm.
A FWM seeded pulse at 1040 nm would be narrower than the input pulse, and would grow at the expense of the 1034 nm line via SRS. However, the Raman process conserves photons, so is not a complete explanation of the pulse narrowing. You may need to consider how the optical phase of the new spectral components evolves. There is significant energy transferred beyond 1040 nm, and this may have a significant effect on the evolution of the pulse shape.
Biao Sun "Is there a trend of the pulse width with the increased power?"
There is a potentially complex interaction between non-linear phase shift and chromatic dispersion.
In the absence of chromatic dispersion or wavelength-selective attenuation, the pulse shape is unchanged by phase modulation.
Kerr effect non-linearity typically broadens the spectrum of an input pulse, so that the pulse width is more sensitive to chromatic dispersion. One approach to create narrow pulses uses self-phase modulation (SPM) to induce chirp in a smooth un-chirped pulse, and combines this with anomalous dispersion to reduce chirp and compress the pulse.
SPM combined with anomalous dispersion (i.e. phase velocity slower for low optical frequencies than for high frequencies) is responsible for modulation instability. MI selectively amplifies some envelope perturbations at moderate powers, and leads to chaotic instability and break-up of pulses at high signal powers.
Single channel MI gain is not usually a concern in fibre with normal dispersion. Spectral broadening can still occur and increase the sensitivity to linear chromatic dispersion, so that pulse broadening or narrowing is more pronounced.
- Badges
- Science topic
- Similar topics
- Papers