Any finite system is delimited by interfaces. In this trivial sense interfaces are ubiquitous. However, modern epitaxial techniques seek to modify the properties of materials by the creation of interfaces. “Band gap engineering,” the attempt to tailor the electronic properties of semiconductors by interleaving many dissimilar layers, is an example of this approach. Interfaces between lattice-matched, isostructural, crystalline systems, (differing only in composition) are technologically the most advanced and thus most widely used. These “chemical” interfaces are formed by the introduction of dopant impurities, or by the stacking of dissimilar materials. Chemical interfaces formed by the epitaxial growth of dissimilar materials are the subject of this article.

A fundamental tenet of modern epitaxy is the tailoring of properties through control of structure. We will therefore outline how the “structure” of a chemical interface may be defined and experimentally determined. Determining the interfacial structure is a major experimental challenge, inextricably bound with our understanding of how structure affects other properties. The article will thus briefly discuss this vital link. We will conclude with an outline of the way chemical interfaces, in reality systems far from equilibrium, can relax, and how their relaxation mechanisms shed light on the fundamental properties of solids. Because the GaAs/AlGaAs system is the most technologically advanced, and to be concrete, we will illustrate the discussion by reference primarily to this system. Many of the concepts and experimental results, however, are more generally valid.