| Mode-locked fiber lasers have become indispensable tools in many fields as their use is no longer relegated to the optics community. In the future, their size will decrease and their applications will become far more prevalent than they are today. At present, the field is undergoing a cardinal shift as these devices have become commercially available in the last decade. This has put an emphasis on long-term performance and reliability as these devices are beginning to be integrated into complex systems in areas as diverse as medical optics, micro-machining, forensics, and tracking as well as their obvious use as laboratory tools or sources in telecommunications. This is also resulting in a transition from research to engineering.; Since the field of mode-locked lasers has been extensively studied for over forty years, one may expect that little has been overlooked. However, since the mode-locking phenomena is governed by nonlinear partial differential equations, a rich degree of effects exist and the field has not yet been exhausted. During the past two decades, the main emphasis has been on short-pulse generation; however, the main thrust of research is likely to change to producing high-power devices, which will result in limiting effects and thermal issues that are currently ignored for low-power sources. Finally, detailed studies have generally been performed numerically as analytic solutions only exist in limiting cases.; In this thesis, mode-locked fiber lasers are studied experimentally, numerically, and theoretically. The experimental work focuses on high-repetition rate, mode-locked cavities, which are then modeled numerically. A semi-analytic tool, which goes beyond the prior theories and includes all of the effects experienced by steady-state, mode-locked pulses as they propagate in a laser cavity, is also derived. The only caveats to this approach are an assumption of the pulse shape and the requirement that it not change during propagation through the laser cavity. Despite these limitations, it is found that the parameters predicted by our method deviate from those found through rigorous numerical simulations by 10% or less. |
