Splice Loss Optimization of Single Mode Fiber and Erbium Doped Fiber—Experiments and Issues

In metro and long haul networking applications, Erbium Doped Fiber Amplifiers (EDFAs) are used to amplify weak optical signals. Manufacturing of EDFAs is primarily a fusion splicing process in which both Single Mode Fibers (SMFs) and Erbium Doped Fibers (EDFs) are utilized. One of the critical operations is the splicing of an SMF to an EDF, a dissimilar fiber splicing process. Splice losses between these fibers need to be optimized, and the process is highly reliant on the properties of the EDF. Mode Field Diameter (MFD), spectral attenuation at peak wavelength and concentration of erbium along its length vary from batch to batch. The splice loss is dependent on some of these properties and must be taken into consideration. With this background, research was conducted to study the properties of EDFs and its applicability in the splicing process. Having considered the characteristics of the EDF in different wavelength regions, experiments were designed to optimize the losses between an SMF and an EDF. In the C-band (1525–1565 nm), erbium atoms absorb most of the transmitted power (in absence of a 980/1480 nm laser pump). Splice losses measured in these regions are dependent upon the absorption properties and would not depict a true picture of the splice loss. Since the incident power is absorbed, an alternate approach would be to launch extremely low power (<−27 dBm). In this case, the absorption losses should be minimal. As C-band is highly absorptive, launching power in the range of 1310 nm would be another possible scenario. The ‘cutback’ method was also employed to determine the losses in the C-band region. Statistical methods such as the Design of Experiments (DOE) were used to study the properties of the EDF and its response to various splicing parameters and wavelengths. Splice loss trends at various power levels were also investigated. The primary intent of these experiments was to translate the results and their utility into the manufacturing of EDFAs, wherein a multitude factors creep into the splicing scenario. The best method would be the one that consistently yields a low splice loss, since these are critical to minimize the noise figure of the EDFA.