8.1 Introduction

With the aid of band theory we have succeeded in classifying solids into metals, insulators, and semiconductors. We are now going to consider semiconductors (technologically the newest class) in more detail. Metals and insulators have been used for at least as long as we have been civilized, but semiconductors have found application only in the last century, and their more widespread application dates from the 1950s. During this period the electronics industry has been (to use a hackneyed word justifiably) revolutionized, first by the transistor, then by microelectronic circuitry. Each of these in succession, by making circuitry much cheaper and more compact, has led to the wider use of electronic aids, such as computers, in a way that is revolutionary in the social sense too.

Perhaps the key reason for this sudden change has been the preparation of the extremely pure semiconductors, and hence the possibility of controlling impurity; this was a development of the 1940s and 1950s. By crystal pulling, zone-refining, and epitaxial methods it is possible to prepare silicon and germanium with an impurity of only 1 part in 10 ¹⁰. Compare this with long-established engineering materials, such as steel, brass, or copper where impurities of a few parts per million are still virtually unattainable (and for most purposes, it must be admitted, not required). Probably the only other material that has ever been prepared with purity comparable...