Overview

Early bipolar junction transistors were too slow for practical applications in telecommunications. One approach to speed up the flow of the minority carriers from the emitter to the collector by incorporating an 'electric field' into the base region, the so-called ' drift transistor,' was proposed by Herbert Kroemer in 1953 [1]. The drift transistor used the concept of a doping-engineered electric field in the base to reduce the electron base transit time. An eight-fold increase in the theoretical frequency was predicted as compared to Shockley's 'diffusion' bipolar transistors. This could be achieved by using not a uniform doping in the base but one that decreased exponentially from the emitter end to the collector end. While working out the detail, Herbert Kroemer realised that

a drift field may also be generated through a variation of the energy gap itself by making the base region from a nonstoichiometric mixed crystal of different semiconductors with different energy gaps (for example, SiGe), with a composition that varies continuously through the base (translation from Reference 2.)

This was not yet the full general design principle, but it constituted the original conception (see Figure 6.1) of what has become known as the heterostructure bipolar transistor, and ultimately of the heterostructure device field in general [2]. Heterostructures, as it is known today, may be defined as heterogeneous semiconductor structures built from two or more different semiconductors, in such a way that the transition region or interface between the different materials plays an...