Following its invention and demonstration in the late 1940s by Bardeen, Brattain, and Shockley at Bell Laboratories, the bipolar junction transistor, or BJT, became the first commercially successful three-terminal solid-state device. Its commercial success was based on its structure. In the structure of the BJT, the active base region of the transistor is below the surface of the semiconductor material, making it much less dependent on surface properties and cleanliness. Thus, it was initially easier to manufacture BJTs than MOS transistors. Commercial bipolar transistors were available in the late 1950s. The first integrated circuits, resistor-transistor logic gates, and operational amplifiers consisting of a few transistors and resistors appeared in the early 1960s.

While the FET has become the dominant device technology in modern integrated circuits, bipolar transistors are still widely used in both discrete and integrated circuit design. In particular, the BJT is still the preferred device in many applications that require high speed and/or high precision. Typical of these application areas are circuits for the growing families of wireless computing and communication products.

The bipolar transistor is composed of a sandwich of three doped semiconductor regions and comes in two forms: the npn transistor and the pnp transistor. Performance of the bipolar transistor is dominated by minority-carrier transport via diffusion and drift in the central region of the transistor. Because carrier mobility and diffusivity are higher for electrons than holes, the npn transistor is an inherently higher-performance device than the pnp transistor. In Part III of this book, we will learn that the bipolar transistor typically offers a much higher voltage gain capability than the FET. On the other hand, the BJT input resistance is much lower, and a dc-bias current must be supplied to the control electrode.

Our study of the BJT begins with a discussion of the npn transistor, followed by a discussion of the pnp device. The transport model, a simplified version of the Gummel-Poon model, is developed and used as our mathematical model for the behavior of the BJT. The mathematical model is also recast in the classical Ebers-Moll model of the BJT. Four regions of operation of the BJT are defined and simplified models developed for each region. Examples of circuits that can be used to bias the bipolar transistor are presented. The chapter closes with a discussion of the worst-case and Monte Carlo analyses of the effects of tolerances on bias circuits.