| |
| |
|
| |
Cascaded Raman Self-frequency Shifted Soliton Generation in an Er/Yb-doped Fiber Amplifier D. H. Kim, J. U. Kang and J. B. Khurgin |
| |
Motivation:
The polarization and power dependences of cascaded Raman self-frequency shifted solitons (CRSSS) generation in a polarization maintaining high power Er/Yb fiber amplifier have been experimentally investigated. The experimental results show that the number and the amount of cascaded Raman solitons generations and the frequency shifts, respectively, dependent on the input polarization state and the gain of the all-PM fiber amplifier. The numerical modeling was also conducted based on the experimental parameters and the experimental results are shown to be in good agreement with the theory.
Experiments:
The experimental setup consists of a passively mode-locked Er-doped fiber laser and a high power all-PM Er/Yb-doped fiber amplifier as shown in Fig. 1. The Er-doped fiber laser is wavelength tunable and operates at a repetition rate of 10 MHz with a maximum output power of ~0.6 mW and a pulse width of 0.7 ps. This laser was used to provide a seed pulse for the experiments. The all-PM commercial 23 dBm Er/Yb doped fiber amplifier was used for the generation of CRSSS and its overall length of the gain stage was approximately 5 m. In order to characterize the polarization dependence in the SSFS, a quarter- and half-waveplates were inserted in front of the polarized beam splitter (PBS) to convert the non-polarized light from the fiber laser to linear polarization. Additional half-waveplate after the PBS was used to rotate the state of polarization relative to the fast axis(TM) of the PM-EDFA. The measurement of the gain and power dependent soliton generation was done with a variable attenuator in front of the PM-EDFA and by varying the pump current of the PM-EDFA. The output from the PM-EDFA was monitored by three detection systems: optical spectrum analyzer, autocorrelator, and an optical power meter. The connections between the amplifier and all the other devices were done using the standard SMF-28 fiber patch cords. The output power level of the amplifier was varied from 3.6 mW to 76 mW for the experiments. Based on our preliminary result, we found that the zero dispersion wavelength of our entire experimental setup was approximately 1540 nm and the effect of the self-frequency shift was the largest near the zero dispersion wavelength.
Fig. 1. Schematic of the experimental setup for the generation of cascaded Raman solitons. The inset shows the angle (θ) between the polarization of the input seed laser pulse and fast-axis (TM) of the PM-EDFA.
Results
Increase in generation of cascaded Raman solitons can be seen as the optical pump power increases. Also the generation depends on polarization of input laser. (Fig.2)
 
 
Fig. 2. Spectra for the pulses exiting from Er/Yb-doped fiber amplifier with input pulse of five different polarization angles (θ) and different optical pump powers (a) 3.6 mW, (b) 28 mW, (c) 52 mW, (d) 76 mW. The numbers correspond to the polarization angle θ in units of degree.
Fig. 3 shows soliton shifts as a function of input power.
 
Fig. 3. (a) The experimental results for the power dependent spectral width are summarized (squares) along with the simulation results (triangles). The square-root fitting line is shown together. (b) Soliton self-frequency shifts can be seen in the simulated spectra with 5 different output power levels.
The polarization dependence experimental results and the corresponding simulation results are shown in Fig. 4a and 4b.
 
Fig. 4. (a) The polarization angle dependent spectral width of the CRSSS is shown for the experimental results of 28 mW and the simulation result of 27 mW. The lines are guide for the eye. (b) Spectra obtained from numerical simulation solving coupled NLSE for 5 different polarization angles.
In the Fig. 4a we plotted the amount of soliton self-frequency shift as the function of the input polarization angle. It is clear from the figure that the nonlinear effect is minimum when the input light is at 45 degrees where the power ratio between the slow and fast axes is unity. However, as the polarization angle approached both zero and 90 degrees, the nonlinear effect increasingly became larger. This, we believe, is because the ratio of cross-phase modulation to the self-phase modulation in an isotropic medium such as silica fiber is 2/3. Thus, in the absence of four-wave mixing term, when the light is launched at 45 degrees, the total nonlinear phase modulation becomes about 83% compared to the case when the light is launched along one of the principal axes of the fiber. However as the power ratio between the Slow/Fast or Fast/Slow modes becomes larger, so as the total nonlinear phase modulation.
Reference
“Cascaded Raman self-frequency shifted soliton generation in an Er/Yb-doped fiber amplifier”, D.H.Kim, J.U. Kang and J.B. Khurgin, Appl. Phys. Lett. 81 , 2695 (2002)
|
| |
| >> BACK |
|
| |
|
|
|
|
