Repetition-rate dependence of gain compression in InGaAsP optical amplifiers using picosecond optical pulses

A two-level model of a semiconductor optical amplifier is used to calculate gain compression due to picosecond optical pulse injection. Gain compression (measured in decibels) is shown to depend linearly on output pulse energy, with a slope that is characterized by a saturation energy. At repetition rates such that the pulse period is comparable to the amplifier gain recovery time, the saturation energy is repetition-rate dependent. At low repetition rates, the saturation energy is a function only of device parameters. In the limit of very high repetition rate, gain compression is determined by the average power, rather than the pulse energy, and is characterized by a saturation power which is equal to the low-frequency saturation energy divided by the amplifier gain recovery time. The authors present measurements of gain compression in the three regimes of repetition rates. In all cases, the measurements confirm the gain compression characteristics predicted by the model. They describe the implications of the repetition-rate-dependent gain compression on the optimum choice of amplifier and its operating conditions. They show that a short gain recovery time results in an increased output power for a given acceptable level of gain compression, as well as an increased gain level for a given output power