Noise characteristics of nonlinear semiconductor optical amplifiersin the Gaussian limit

This paper addresses noise properties of nonlinear semiconductor optical amplifiers. From a basic point of view, noise properties of nonlinear optical amplifiers are sufficiently different from those of linear amplifiers to warrant detailed modeling which has not been formulated previously. From a practical point of view, nonlinear semiconductor optical amplifiers are important for future all-optical signal-processing applications which may involve the operation of these devices in a saturated regime. Nonlinear amplifiers are also common in systems operating near 1300 nm and in integrated booster amplifiers. Under nonlinear operating conditions, amplifier noise contains a narrow-band contribution that comes about due to the nonlinear coupling of noise and gain. The more conventional broadband spontaneous noise also changes as the inversion factor becomes power-dependent and varies along the amplifier axis. We analyze noise in nonlinear amplifiers in the Gaussian limit (meaning, for fields consisting of large photon numbers) for CW or NRZ modulated signals and separately for short pulses. We consider the case of a single input as well as configurations of multi-input signals interacting via four-wave mixing. Using a specific detection system for the calculations of electronic signal-to-noise ratios, we demonstrate a reduction in the narrow-band electronic noise due to saturation in the single input case. We also demonstrate a vast advantage of using short pulses in four-wave-mixing applications     

Experimental study of the statistical properties of nonlinearly amplified signals in semiconductor optical amplifiers

We examine the effect of nonlinear amplification on the noise properties of an amplified signal. We show that the output noise statistics can be approximated as Gaussian over a wide range of practical parameters. In addition, we show that nonlinear amplification of noise carrying signals may result in an enhancement of the signal-to-noise ratio (SNR). Basic principles dictate that this can not improve the bit error rate performance of a communication link, as long as an optimal detection scheme is used. Hence, we conclude that for communication systems that exploit nonlinear amplification in semiconductor optical amplifiers, evaluation of performance on the basis of the SNR can be misleading and should be made with great care.     

Improved bit-error-rate performance in frequency conversion of high-bit-rate data based on cross gain compression in semiconductor optical amplifiers

We describe a detailed calculation of the bit-error-rate performance of an optical frequency converter based on cross gain saturation in a semiconductor optical amplifier. Performance limits related to the slow-gain dynamics of the amplifier, as well as to noise, are presented and a modification to the detection system based on an adaptive decision threshold level that corrects for the performance deterioration due to dynamical effects is proposed.     

Calculation of bit error rates in all-optical signal processingapplications exploiting nondegenerate few-wave mixing in semiconductoroptical amplifiers

We present a theoretical analysis of the noise properties in a nonlinear semiconductor optical amplifier. We concentrate on amplifiers operating as nonlinear media mediating four-wave mixing processes. We calculate error probabilities for different configurations that use four-wave mixing in all-optical signal manipulation such as frequency conversion and time domain demultiplexing. We identify a data induced nonlinear interference process as having a major adverse impact on the performance of nonlinear semiconductor optical amplifiers and propose the use of a variable decision threshold to overcome it. We compare the performance of the variable threshold to that using an optimal fixed threshold and show significant improvement in performance. We treat the cases of NRZ modulated data as well as the case of short optical pulses and demonstrate a vast advantage for the short pulses in terms of the required input field intensities and the achievable bit error rates     

The statistics of polarization-dependent loss in opticalcommunication systems

In this letter, we study the statistics of polarization-dependent loss (PDL) in optical systems and evaluate its evolution with system length. We show rigorously that within the range of conceivable parameters in optical links, the distribution of PDL is Maxwellian when it is expressed in decibels. The accumulation of the mean-square PDL with system length stays linear in most systems, but may assume an exponential growth in very long systems with large PDL. The analytical results are compared with numerical simulations and an excellent agreement is observed     

Four-wave mixing among short optical pulses in semiconductoroptical amplifiers

We describe the dependence of four-wave mixing conversion efficiency among short optical pulses in a semiconductor optical amplifier. We show a strong dependence on pulse shape and pulse overlap, predicted theoretically and confirmed in experiments     

Cross-phase modulation in an L-band EDFA

We measure the cross talk in a wavelength-division-multiplexed (WDM) system arising from cross-phase modulation (XPM) in an L-band erbium-doped fiber amplifier (EDFA). This effect has the potential of becoming the dominant nonlinear crosstalk mechanism in L-band WDM systems using standard fiber. We discuss how appropriate amplifier design can reduce the XPM in the EDFA     

Cancellation of timing and amplitude jitter in symmetric linksusing highly dispersed pulses

We study the properties of symmetric dispersion compensation in optical links using highly dispersed pulse transmission. We show analytically that by splitting the dispersion compensation equally between the input and output of the link, complete cancellation of the timing and amplitude jitter can be obtained in systems where the power profile is symmetric about the center. We explain the dynamics of this cancellation and show, theoretically and experimentaily, that with practical system parameters, symmetric compensation may lead to a considerable improvement in performance     

The effect of strong inline filtering on the amplitude jitter in long optical systems

The accumulation of amplitude jitter in ultralong-haul (ULH) systems deploying strong optical filters is studied. It is shown that in such systems that operate with nonsoliton waveforms, the selective inclusion of filters may lead to the enhancement or suppression of amplitude noise accumulation, depending on the filter's positioning. This paper explains the improvement in performance that was demonstrated experimentally in a 60-span 12.5-Gb/s system with 12 periodically positioned bandlimited dynamic gain equalizers (DGE) (Optical Fiber Communication Conf., 2005). It is also shown that the same noise suppression as in (Optical Fiber Communication Conf., 2005) achieved by such a system can be achieved with as few as two strong filters if their positioning is properly optimized.     

PMD penalties in long nonsoliton systems and the effect of inline filtering

We consider the polarization-mode dispersion (PMD) tolerance of long nonsoliton systems. We find that the effect of PMD in such systems is to derail the nonlinear evolution of the waveform from its PMD-free trajectory at an early stage of transmission, leading to extremely high penalties at the receiver end. It is shown that PMD affects the propagating bandwidth in a way that is strongly correlated with the received penalty. Use of inline optical filters is shown to be extremely efficient in reducing the penalties caused by PMD.