Diffusion Processes for Integrated Waveguide Fabrication in Glasses: a Solid-state Electrochemical Approach

-The process of solid-state film ion exchange (Ag +-Na +) in glass substrates for the fabrication of integrated, optical waveguides has been investigated as a solid-state electrochemical process. The ion exchange process is initiated by establishing a space-charge distribution in the more mobile charge carrier. The space charge establishes an intense electric field that drives oxygen to the surface to fuel dopant oxidation and subsequent diffusion. A one-dimensional model is presented in space-charge effects in glass having charge-blocking and permeable electrodes. The model results show that the rate-limiting step for initiating Ag + diffusion in glass is the motion of oxygen ions towards the silver anode to form a layer of silver oxide. The predictions of oxygen diffusion time compare favorably with the "dead times" in the current vs time curves observed during the ion exchange process. The dead time corresponds to the period of negligible diffusion of silver into glass and formation of a stable oxide layer. This was experimentally confirmed by Rutherford backscattering analysis. The variation of maximum diffusion current as a function of applied voltage and temperature was also studied and a theoretical electrochemical treatment is presented. 1. I N T R O D U C T I O N All optical communication systems rely on waveguides, most commonly optical fibers, for confining and transmitting optical signals over large distances. In confined regions such as computers or telecommunications switches these waveguides will have to be fabricated on or within monolithic substrates to save space and increase reliability. These integrated, optical waveguides (IOWs) in glass and crystal substrates are believed to be as critical to the future development of photonic systems as integrated circuits were to electronic systems. One important method for fabricating these devices is ion exchange in glasses or crystals. The ion exchange process has its roots planted hundreds of years in the past in the process of silver staining to produce stained glass windows. Its recent direct ancestor is the process for chemical strengthening of glasses (Nordberg et al., 1964). Investigators soon realized that the ion exchange process used for chemical strengthening also modified the refractive index of the glass, hence a waveguide could be formed. Ion exchange has since been used to fabricate opto-electronic circuits in crystals such as lithium niobate (Syms, 1988) and to fabricate passive splitting and distribution networks in glasses (Ramaswamy and Srivastava, 1988). Although several of these devices are commercially available, a detailed understanding of the ion exchange process is lacking and so prototyping or fine tuning of the manufacturing process has been an inefficient trial-and-error procedure. There are two major ion exchange processes in widespread use. The "ion exchange" technique uses t Corresponding author. a molten salt as the dopant source and the diffusion can be assisted by applying an external electric field (Ramaswamy and Srivastava, 1988). The solid-state, film diffusion technique uses a solid metal film as the dopant source (Najafi et al., 1986; Forrest et al., 1986; Viljanen and Leppihalme, 1980; Kaneko, 1990). Since pure thermal diffusion is so slow in the solid-state process, an electric field is normally applied to enhance the penetration of dopant ions. Both molten salt and solid-state processes are shown schematically in Fig. 1. The ion exchange techniques exploit the fact that any alteration of the physical or electronic structure of a glass or crystal yields a refractive index change. The introduction of a dopant into the substrate changes the local environment, hence forming the waveguide. Alkali ion exchange is generally used to produce light-guiding structures since the alkali ions are relatively mobile in glasses and crystals and can be exchanged quite easily for ions of higher polarizability or different size. The exchange ofAg + for Na + ions in glass has been the most widely explored process (Ramaswamy et al., 1988; Doremus, 1964; Albert and Lit, 1990; Honkanen et al., 1987), but many other combinations have been investigated, especially in crystals (Abou-el-Leil et al., 1991; Becker et al., 1992; Brinkman et al., 1991; Bierlein and Vanherzeele, 1989). Many investigators have attempted to model the ion exchange and solid-state diffusion processes as one-dimensional, field-assisted, transient diffusion problems (Li and Johnson, 1992; Abou-el-Leil and Cooper, 1979; Tervonen et al., 1991). Their goal was to predict the dopant profile inside the substrate. The