Abstract

Pulsed nonlinear-optical devices suffer from the well-known problem of pulse shaping when the response time of the nonlinearity is much shorter than the pulse duration. This situation will prevail in silica Tiber devices for pulses longer than 100 fs and has been observed in nonlinear couplers¹ and in Sagnac interferometers.² Solitons were first proposed as the solution to this generic problem³ in the context of Mach-Zehnder interferometers but have subsequently been applied to Sagnac interferometers⁴ and nonlinear couplers.⁵ All these devices are interferometric in the sense that two paths are available for the light propagation. The principle of soliton operation can be understood in terms of the effect of dispersion in these devices. When dispersive effects are unimportant, the device responds only to the instantaneous intensity of the pulse leading to the pulse being reshaped by the nonlinear device response. When dispersion is important, every part of the pulse is connected to every other part, leading to a collective behavior in which the phase shift induced at the peak of the pulse can be communicated to the wings. This is exactly the circumstance under which it is possible to form solitons. Therefore, perhaps it is not surprising that they turn out to be central to the successful operation and understanding of these nonlinear-optical devices.

© 1990 Optical Society of America