Abstract

Spatial-hole-burning (SHB) is known to play an important role in DFB lasers and has been studied extensively in the literature. To predict the dynamic behaviour in the presence of SHB, several DFB laser models have been developed. In most of the existing models, a linear gain profile with a constant differential gain is assumed^[1]. In reality, however, the material gain in MQW lasers is not a linear function of carrier density. As was pointed out recently^[2-3], the material gain will saturate rapidly at high carrier densities in MQW lasers. This is one of the characteristics of QW lasers compared with bulk semiconductor lasers, mainly due to the step-like density of states in the QW active layer^[2-3]. This is so-called linear gain saturation^[2]. This effect has been experimentally confirmed by studying the dependence of laser dynamics on the number of quantum wells for λ/4-shifted MQW DFB lasers^[2]. In the recent paper^[2], the authors included the linear gain saturation effect in the simple FP laser model and tried to explain the experimental results. There was, however, a clear discrepancy between the experimental results and the theoretical predictions. The experimental results showed that the relaxation oscilation frequency becomes maximum with 10 quantum-wells, while the simple FP model predicted that it will increase when the number of quantum-wells is increased. This discrepancy is understandable because that the model they used is only valid for FP lasers and not applicable to DFB lasers, especially for the quarter-wave shifted DFB lasers where the SHB is of great importance in predicting the dynamic behaviours of these lasers. In this paper, we will show that the above discrepancy can be resolved by the combined effects of linear gain saturation in MQW structures and SHB in DFB lasers based on our newly-developed MQW DFB laser simulator^[4-5].

© 1994 Optical Society of America