J.K. Andersen

Polarization-insensitive nonlinear optical loop mirror demultiplexer with twisted fiber

We experimentally demonstrate reduction of the polarization sensitivity of a nonlinear optical loop mirror (NOLM) from 5 to 0.5??dB by use of 550??m of twisted dispersion-shifted fiber with a twist rate of 8 turns/m (24??turns/beat length). The twisting of the fiber induces circular birefringence and equates the parallel- and the orthogonal-polarization nonlinear phase-shift terms. Experimental results show that the polarization sensitivity monotonically decreases from 5??dB for nontwisted fiber to 0.5??dB for fiber that is twisted at a rate of 8??turns/m, and the twist rate should be more than 4??turns/m (>10 turns/beat length) for emulation of circularly polarized fiber. The minimum polarization sensitivity occurs when the control-pulse polarization is aligned with one of the eigenmodes of the twisted fiber. With the fiber twisted at a rate of 8??turns/m in the NOLM, the nonlinear transmission is 23% at a switching energy of 4??pJ/pulse. Simulations confirm the observed behavior and show that the remaining polarization sensitivity results from energy transfer between orthogonal modes of the signal pulse.

Polarization insensitive demultiplexing of 100-Gb/s words using a twisted fiber nonlinear optical loop mirror

We demultiplex 100-Gb/s words using a two-wavelength nonlinear optical loop mirror (NOLM) with polarization sensitivity <0.5 dB obtained by using fiber twisted at eight turns/m. This polarization insensitive demultiplexer is more suitable than conventional nontwisted fiber NOLMs for use in high-speed all-optical networks where the input states of polarization are arbitrary. The demultiplexer consists of synchronized erbium-doped fiber lasers, a 100-Gb/s fixed word encoder, and a NOLM using 450 m of twisted fiber with /spl lambda//sub 0/=1518 mn. For inputs of arbitrary polarization, we measure a/spl sim/5.5-ps timing window and contrast ratio >15 dB. To understand the improvement in performance of the system due to the polarization insensitivity, we apply a statistical method of measuring the Q for the twisted and nontwisted fiber demultiplexers. The Q of /spl sim/15 for the twisted fiber NOLM is improved over that of the NOLM with nontwisted fiber (Q/spl sim/10), and is consistent with an increased minimum nonlinear transmission.

Demultiplexing of arbitrarily polarized 100 Gb/s words using a twisted fiber nonlinear optical loop mirror

Summary form only given. We demonstrate a 100 Gb/s demultiplexer for data with arbitrary input states of polarization using a polarization insensitive, two-wavelength, nonlinear optical loop mirror (2/spl lambda/ NOLM). The device uses 450 m of twisted fiber to achieve polarization sensitivity 15 dB, and Q-parameter >15. Our demultiplexer combines the 2/spl lambda/ NOLM with a local laser and synchronization circuit based on a phase-locked-loop and an acousto-optic modulator and grating in the laser.

Propagation of 2-ps solitons through 100 km of standard fiber with temporal and spectral compensation by NOLMs and optical filters

We demonstrate propagation of 2-ps solitons through 100 km of standard fiber with erbium-doped fiber amplifiers (EDFAs) positioned every 25 km, recovering the input pulse width and spectral characteristics. We find that periodically placed optical bandpass filters and nonlinear optical loop mirrors (NOLMs) compensate for soliton self-frequency shift (SSFS), reject the accumulated dispersive wave, and self-time-gate the solitons.

Simultaneous measurements of fiber dispersion and nonlinearities using pulse phase measurements

The upgrading of existing communication systems to wavelength-division multiplexing and high-speed time-division multiplexing systems has focused attention on the effects of fiber nonlinearities. We show that both linear and nonlinear fiber parameters can be measured by observing phase changes in optical pulses propagating through the test fiber. In particular, the method is illustrated on standard Corning SMF-28 and the measured values are within 100% of nominal values with a standard deviation of 4%. The method consists of characterizing the intensity and phase of a reference pulse both at the input and the output of the fiber. By comparing the measured phases with the calculated phases the fiber parameters are determined.