This review is devoted to the description of recent experimental results concerning the nonlinear optical properties of semiconductor-doped glasses SDGs with particular emphasis on the regime in which the energy of the incident photon is smaller than the energy gap. A considerable theoretical and experimental effort has been devoted in the last 10 years to the fundamental aspects of quantum-confined structures, which have properties somewhat intermediate between the bulk crystals and atoms or molecules. From this point of view, SDGs represent an easily available test system, and optical techniques have been a major diagnostic tool. Luminescence and absorption spectroscopy were extensively used to characterize the electronic states. The experiments aimed at the measurement of the real and imaginary parts of the third-order optical susceptibility of SDGs below the bandgap are described in some detail, and the results obtained with different techniques are compared. Besides the intrinsic fast nonlinearity due to bound electrons, SDGs may present a larger but much slower nonlinearity due to the free carriers generated by two-photon absorption. This implies that experiments have to be properly designed for separation of the two effects. In this article we stress the importance of a detailed structural characterization of the samples. Knowledge of the volume fraction occupied by the nanocrystals is necessary in order to derive from the experimental data the intrinsic nonlinearity and to compare it with the bulk nonlinearity. We discuss recent experiments in which the dependence of the intrinsic nonlinearity on the crystal size is derived by performing, on the samples, measurements of the real part and imaginary part of the nonlinear optical susceptibility and measurements of crystal size and volume fraction. Structural characterization is of interest also for a better understanding of the physical processes underlying the growth of crystallites in SDGs. The average size of nanocrystals can be tailored by controlling the temperature or time of the treatment. The major problem is the size dispersion of the crystallites, which is intrinsic to the diffusion process. At present, this is the major source of the undesired inhomogeneous broadening of the optical transition lines of the SDGs. Efforts are at present being made to fabricate materials, SDGs included, which embed nanocrystals with a reduced spread of sizes. The interest in the nonlinear optical properties is due not only to fundamental reasons but also to possible applications for optical devices. Generally speaking, resonant nonlinearities are much larger than non-resonant nonlinearities, but they are not necessarily the most interesting for applications because materials at resonance absorb the incident radiation and also present long response times. The studies below the bandgap seem to indicate that the values of the intrinsic nonlinearities of nanocrystals in the structures which are at present available are similar to those of the bulk. New and better controlled structures are now under development and have to be tested from the view-point of optical nonlinearities. In several situations SDGs cannot be modelled as an ensemble of freely standing nanocrystals, with the glass matrix playing the role of an inert support. Phenomena such as trapping and darkening, which are very probably connected with electronic states at the glass-semiconductor interface, may play a role in determining the optical response. They might give rise to an extrinsic optical nonlinearity which can be even larger than the intrinsic nonlinearity. The physical processes which are involved in these extrinsic nonlinearities are poorly understood and at present being investigated.
