A simple regional analysis of transit times in bipolar transistors

Abstract

In the quasi-static approach the cut-off frequency or unity-gain bandwidth f T of a one-dimensional bipolar transistor can readily be calculated from the doping profile by means of a numerical device simulation program. The physical interpretation of such results in terms of the usual regional model is greatly facilitated by a new partition of the small-signal charge distribution. Only two region boundaries remain to be determined. The criterion employed for this purpose is unambiguous. The total transit time from emitter contact to collector contact (inversely proportional to f T ) is subdivided into five physically meaningful components. The method can readily be implemented in device simulation programs, making these easier to use as a design tool. It provides more detailed insight into the influence of the emitter doping gradient on f T . Examples illustrate the three effects that can cause the f T fall-off at high collector current densities: high injection in the base, base widening and quasi-saturation. In all cases the emitter transit time turns out to dominate the total transit time at very high current densities, albeit beyond the limit of practical interest.