Optical fibre-grating pulse compression

Abstract

Numerical simulation is used to consider non-linear pulse propagation in fibres and subsequent pulse compression in a dispersive delay line. It is shown that for small initial pulse powers the conventional non-linear Schrodinger equation (NSE) is quite accurate to describe the process of pulse propagation in fibres. In this case initially symmetrical pulses undergo squaring and spectral broadening in fibres, and frequency chirp is linearized over most of the pulse, while shapes of the pulse, spectrum and frequency chirp remain symmetrical at the output of the fibre. There is a certain optimum fibre lengthZopt which is determined by the initial pulse parameters and fibre characteristics for pulse compression in the dispersive delay line. When the fibre lengthZ>Zopt the optical wave breaking effect distorts the linearity of the frequency chirp and thus deteriorates the quality of the compressed pulse. The region of NSE approximation accuracy is determined. It is demonstrated that at increase of the initial pulse power (initial pulse width makes no difference) the NSE approximation becomes inaccurate. So the pulse dynamics in the fibre were described by the modified NSE derived in the higher-order approximation of the method of slowly varying amplitudes from Maxwell's equations. In this case the shock wave appears at the trailing edge of the pulse, which accelerates the wave breaking process. This results in a decrease of the optimum fibre length and deterioration of compressed pulse parameters, compared with the NSE case. Spectral windowing of the extreme Stokes components of the pulse spectrum permits significant improvement in the quality of the compressed pulse. The main features of the compression of pulses with asymmetrical initial shape are also considered.