Wavelength Conversion Using SOAs: Simulations and Experiment
Paul R. Prucnal,Bing C Wang,Ivan Glesk
DOI: https://doi.org/10.1364/ipr.2002.ithh5
2002-01-01
Abstract:As optical networks advance toward a dynamic multiwavelength all-optical structure, all optical wavelength conversion technologies become vital in the efficient utility of available fiber bandwidth. Experimental demonstrations of optical wavelength conversion usually involves a nonlinear element to transfer the information from the signal wavelength to that of a local laser source. In the case of Non-return-to-zero (NRZ) signals, the local laser source is usually a CW narrow linewidth semiconductor laser. For Return to Zero (RZ) signals, the local laser sources commonly employ a modelocked fiber laser or a modelocked semiconductor laser or a gain switched DFB to generate low jitter optical clock. Various materials and devices can be used as the nonlinear element for wavelength conversion, with the two most common being the optical fiber and semiconductor optical amplifier, though Lithium Niobate has also attracted attention in recent years as a switching element. The advantage of using optical fiber for wavelength conversion is that it is relatively cheap and has low levels of noise. However, the switching energy required to generate the nonlinearity in optical fiber is high compared to that of the SOA [1]. Wavelength conversion using SOA typically uses either four-wave mixing or crosses gain/phase modulation, with the latter method often used in conjunction with an interferometric topology [2]. The advantages of cross phase modulation of SOA in interferometers are the relatively low switching energy required, the fast switching speed, and the ability to integrate the whole structure on a single substrate.