Pulse compression using fiber gratings as highly dispersive nonlinear elements

Pulse compressors rely on two separate sections. The first section is for bandwidth generation through self-phase modulation and chirp linearization through normal dispersion. In conventional compressors this first section consists of a normal dispersion fiber of appropriate length. The second section is for compensating this linear chirp through anomalous dispersion, typically a prism pair or grating pair. In this way a transform-limited input pulse is compressed into an almost-transform-limited pulse. This scheme is quite different from chirped fiber gratings that are used in reflection to compensate existing chirp: no extra bandwidth is generated and nonlinear effects are not necessary. We propose a scheme for optical pulse compression utilizing an apodized fiber grating in transmission as the nonlinear dispersive element for the first section of the compressor. Near the band edge, on the long-wavelength side of the stop band of the grating, the normal quadratic dispersion is orders of magnitude greater than in a standard optical fiber. Therefore the first section of the compressor may be scaled down in length and the constraints placed on these systems may be relaxed. In this paper we discuss the limitations and the design of such fiber-grating compressors. Analysis and numerical simulation show efficient pulse compression. Further numerical simulation reveals that sufficiently far from the band edge the fiber grating can be modeled as an effective homogeneous medium obeying the nonlinear Schrodinger equation. (C) 1998 Optical Society of America.