The degenerate case of four-wave mixing (also termed four-photon mixing) is a widely applied fundamental mechanism in which two pump photons combine in a nonlinear optical medium to form two new photons: one is down-shifted and the other is up-shifted in frequency from the pump, seen as the growth of two spectral sidebands. The parametric process can also be used for wavelength conversion; for example, a 1550 nm pump could be combined with signal pulses at 1300 nm inside an optical fiber where a copy of the signal pulses would then be generated at around 1.9 µm. These wavelengths satisfy energy conservation, 2/(1550 nm) = 1/(1300 nm) + 1/(1919 nm), but phase-matching between the interacting waves is also required to keep the process efficient over a sufficient fiber length. Suitable phase-matching can be achieved by dispersion-engineering the optical fiber: its core size and the refractive index relative to the surrounding cladding can be tuned to obtain an optimum dispersion profile. For example, generating photons with a large frequency shift from the pump requires that the group-velocity dispersion is close to zero at the pump wavelength.
The optical fiber should be highly nonlinear to increase the efficiency of the parametric process, meaning that the propagating light must be tightly confined to a small area (~10 µm2). For standard step-index fibers, this is achieved by having a small core with relatively large refractive index contrast to the cladding. In this regime, one can also significantly modify the dispersion profile through very small changes to the core size and refractive index step size. However, this large sensitivity also means that even small unintended geometric fluctuations can result in highly varying phase-matching conditions along the fiber. This has been found to significantly limit the efficiency of the four-wave mixing process and broaden the sidebands.
A series of recent papers in Optics Express by Bill P.-P. Kuo, Stojan Radic and collaborators, have provided an elegant solution to this problem. In issue 7 (March 27th 2012) they showed theoretically that for a conventional step-index nonlinear fiber to be used for a narrow band parametric amplifier, the core radius would have to be controlled at a scale of under 196 pm. Considering that 196 pm is close to the Si-O bond length of the silica comprising the core, it became clear that such fabrication tolerances could probably never be achieved in practice. The Authors suggested that a way around this problem would be to add an additional refractive index step to the core. It turned out that this allows tight confinement of the light within the core, as for conventional nonlinear fibers, but the sensitivity of dispersion to core radius fluctuations becomes two orders of magnitude smaller.
In issue 16 (July 30th 2012), the Authors demonstrated the fabrication and characterization of a fiber following the new design principle. They showed that the fiber can be stretched to effectively suppress Brillouin scattering, something which can also be done in conventional nonlinear fibers. The deformation of the fiber geometry due to the stretching would lead to significant change of the dispersion profile in a conventional nonlinear fiber, but the new fiber showed no measurable dispersion change.
Most recently, in issue 17 (August 13th 2012) the new fiber design was used for constructing a fiber optical parametric oscillator (FOPO) to generate light at 1.8 µm by pumping at 1535 nm. It was found that if the FOPO was based on a standard dispersion shifted fiber, the threshold pump power needed to be 2.5 times higher than with the newly designed fiber. This improvement was a direct consequence of having similar dispersion stability in both fibers, but the new fiber design allowed higher field confinement than the standard dispersion shifted fiber.
The work presented in this series of papers shows clearly that the additional refractive index step of the fiber core provides a very useful new design parameter: the field is still highly confined to a small core, but the dispersion is much more stable against geometrical fluctuations.
You must log in to add comments.