Nonlinear Integrated Optics

Abstract

In the early days of integrated optics the attractive features of nonlinear optical devices using a guided wave structure have been recognized very soon [1–4]. Due to the concentration of optical fields in a long waveguide structure with crosssection dimensions of the order of a wavelength, very efficient nonlinear interactions could be expected to generate new optical frequencies. Especially the possibility to achieve phase matching even in isotropic materials, utilizing the waveguide mode structure, stimulated the interest in integrated, nonlinear devices. Anderson and Mc Mullen [5] were the first who demonstrated experimentally phase matched difference frequency generation in a GaAs (isotropic!) waveguide. Suematsu [6] and Boyd [7] both predicted a very low threshold power for an integrated optical parametric oscillator in GaAs. Despite these promising aspects, no experimental efforts have been reported up to now, to obtain parametric oscillation in a GaAs waveguide structure. Probably the required high waveguide quality (low losses, extremely good homogeneity, no optical damage) and the need for a suitable index profile to yield a good overlap of the interacting fields prevented the development of efficient nonlinear devices in that material. Nevertheless, in a number of other material systems nonlinear effects, mainly second harmonic generation, were studied [8]. In particular, several methods to achieve phase matching in optical waveguides were demonstrated. Somekh and Yariv [9] and Tang et al. [10,11] used periodic structures; Burns, Andrews and Lee [12,13] developed a waveguide composition yielding “noncritical” phase matching; other authors used the waveguide mode dispersion or the material birefringence to get phase matching of the interacting modes.