by Sunny Bains
University of Michigan researchers have shown that a simple fiber-based device can be used to convert a signal from one wavelength to another without the usual signal degradation caused by dispersion and other problems. The new system exploits cross-phase modulation, but uses a fiber Bragg grating (FBG) to shape and spectrally filter the converted pulse. In simulations, near-perfect conversion was possible. The design has yet to be proven in a fabricated system but, should the research prove successful, it could become an important step forward for truly flexible WDM networking.
Cross-phase modulation is a phenomenon caused by nonlinearities in optical devices, fibers, and semiconductor amplifiers among others. Essentially, as a high-power pulsed signal propagates through a fiber, the local intensity changes the local index of refraction. If a second laser beam is propagating through the fiber at the same time, this time as continuous-wave, then its phase will be modulated by the refractive index change as it propagates, as well as picking up a frequency shift. The phase change can be transformed into an intensity change by using interferometry. If the two beams are sufficiently close in wavelength then the intensity profile of the second beam should eventually resemble the first.
The converted pulse from cross-phase modulation would normally be very distorted (solid line) but, after filtering with a fiber Bragg grating, the desired pulse shape can still be extracted (dashed line). Below: Pulsed signal (λs) and continuous-wave (λc + δλs) beams interact inside a high-nonlinearity dispersion-shifted fiber (HNL-DSF). The resulting wavelength-converted (λc) signal is filtered using a fiber Bragg grating before being coupled out.
The complication is that to perform wavelength conversion, the two beams cannot be exactly the same. They will have slightly different dispersions through the fiber, thus causing the pulse shape of the newly modulated beam to be distorted. Though the quality might possibly be sufficient for a one-off conversion, it could not be used again to produce a reasonable-looking pulse.
The Michigan team modeled the whole process to find the ideal parameters—in particular, the length of fiber and signal needed to perform the necessary conversion.1 They then used their model to show how distorted the converted pulse shape was likely to be in a given set of circumstances. Finally, they showed that, by understanding this distortion, they could exactly compensate for it using an FBG (see figure). Because the FBG is wavelength-selective, researchers say they should be able to produce a universal converter by incorporating all of the possible compensation gratings into a single device.
For more information contact Herbert Winful at email@example.com.
1. V. E. Perlin and H. G. Winful, IEEE Photon. Technol. Lett. 14 (2) (February 2002).