Abstract

Compression of a chirped (frequency modulated) optical pulse is an attractive method to reach very short pulse duration [1]. Theoretical analysis of a compression stage in a typical linear dispersive system, as a pair of gratings or a prism, is straightforward. In the recently demonstrated [2] compact device for generation of high energy femtosecond pulses, however, broad-band linearly chirped pulses obtained from a fast-tuned diode laser are compressed in an optical fiber. Fiber nonlinearity may significantly influence the pulse shape and has, therefore, to be taken into account while modelling the compression stage. Inclusion of a nonlinear term into the propagation equations implies that no analytical solution can be obtained and numerical marching-on schemes have to be applied in order to simulate pulse shape evolution along the fiber. The problem with applying the standard numerical schemes for a compression of a strongly chirped pulse is that the pulse duration may decrease by several order of magnitudes and hence a large computational window and a very dense computational grid is required to resolve the pulse shape both and the beginning and the end of the fiber. For methods involving Fourier transformation a very dense sampling is needed even at the beginning of the compression stage because of broad spectral width of the chirped pulse. This makes standard numerical methods very inefficient. To circumvent the problem we propose to simulate the pulse evolution using the time frame that is compressed along the fiber.

© 1995 Optical Society of America