There are two questions that are frequently asked about semiconductor surface research. First, in what ways do the bulk and surface properties of semiconductors depend on each other? Second, and particularly pertinent to an audience of chemists, how does the surface behavior of a semiconductor differ from a metal or an insulator? The first question becomes especially important when a semiconductor device is being considered. Channeling on transistors, high back currents on diodes, inversion layers-these are part of the catalogue of surface effects that device nianufacturers have worried about for years. The limited amount of research on semiconductor surfaces (wis-d-wis the bulk research effort) has explained how these effects occur and what one should do about them in principle. The reduction of these principles to practice has involved a great deal of chemistry and, on occasion, alchemy. But why the emphasis on devices when talking about semiconductor surfaces? It is because device technology has provided niost of the important experimental techniques and almost all of the impetus for surface studies. Pearson and Brattain,' in their "History of Semiconductor Research," have made this point quite clear.Most of the standard surface experiments performed on metal surfaces can be performed on semiconductors. Contact potential, adsorption, low energy electron diffraction, photoelectric emission, and field emission are becoming as commonly used in semiconductors as they are in metals surface work. Despite the fact that the same tools can be used on semiconductor surfaces, the important difference between the surface properties of semiconductor, metals, and insulators is precisely that which differentiates the bulk phenomena, namely, the band structure of the material. In a metal, if it is proper to speak of a space charge region at all, you would have one whose width was of the order of a fraction of the lattice spacing, owing to the enormous amount of free charge. In a semiconductor, the width of the space charge region varies as the inverse square root of the bulk carrier concentration. Xeedless to say, the range of the space charge thickness is many orders of magnitude. But perhaps the greatest difference between metal and semiconductor surface studies is that the lower carrier concentration allows us to carry out electrical measurements in the interior of the material which depend on what we do externally. Only thin evaporated metal films have this added versatility.2 For this reason semiconductor surface research is potentially the best way of examining charge transfer 830 For the moment let us turn to the second question.
