The name bipolar is used because in this transistor both types of carriers are critical, as opposed to a field-effect transistor which is considered unipolar. The bipolar transistor is often called the junction transistor because its structure has two p-n junctions back-to-back. Not only was it the first transistor ever realized in practice, more important, its invention sparked rapid advancement of semiconductor fundamentals that revolutionized the electronics industry worldwide.
In December 1947, in the course of searching for a field-effect transistor, the point-contact transistor was discovered by Bardeen and Brattain of, then, Bell Telephone Laboratories. The point-contact transistor had two metal contacts (cat’s whiskers) pressed onto a piece of Ge. It was observed that the reverse current of one contact to the substrate was dependent on the forward current of the other contact that was close by. Within two months of that historical event, the junction transistor using p-n junctions was conceived by Shockley of the same group in early 1948. Demonstration of the junction transistor was made in 1950, using junctions grown from molten Ge. Bipolar transistors produced in the 1950s were typically made with alloyed junctions. The planar technology developed around 1960 started to favor silicon as the semiconductor material. Today, bipolar transistors enjoy a large market share, but they have been challenged by MOSFETs because of factors such as cost, yield, power, etc. Nevertheless, the bipolar transistor maintains a place in high-performance circuits because of its high transconductance.
The historical events surrounding the invention of the bipolar transistor are described by Shockley. In 1953 Bardeen, Brattain and Shockley were awarded the Nobel Prize in physics for their seminal work.
A bipolar transistor can be an n-p-n or p-n-p structure. It is critical that the middle layer (base) be thin, 1 . for high current gain. (The name “base” was obtained from the point-contact transistor where Ge was equivalent to the middle layer and served as the mechanical base.) Most of the bipolar transistors have vertical current flow although a lateral structure can be realized. A more advanced bipolar structure includes a self-aligned base contact for reduced base resistance and a poly-emitter contact for improved current gain.
A typical doping profile for an n-p-n transistor is needed. The emitter doping is generally higher than the base doping for high injection efficiency, and the collector doping is lower than the base doping so that the neutral base xB is a weak function of the collector-base voltage. A common approach to form the emitter and base regions is by a double-diffused technique in which the base is diffused first, followed by a shallower, heavier-doping diffusion of the emitter. The integrated doping in the base (neutral region excluding depletion) is defined as the Gummel number and it usually lies between 1012 and 1013 cm-2 for Si. It will be shown that a small Gummel number is critical for current gain, but the lower limit is set by punch-through between the collector and the emitter. Most commercial bipolar transistors are made from Si because it has a more mature technology. For microwave applications the n-p-n structure is preferred for its higher electron mobility. Although GaAs offers potential benefits in performance, its technology is more difficult for manufacturing.
A bipolar transistor can be operated in different regimes that are determined by the biases of the two junctions. The most important regime is the active (normal) mode where the emitter-base junction is forward biased and the collector-base junction is reverse biased. It is in this mode that current gain is realized.
The storage time severely limits the switching speed in digital circuits, and it is caused by excess charge injected from the collector to the base under forward bias. One way to reduce this minority-carrier injection is to add a Schottky barrier clamp in parallel to the collector-base junction. This Schottky diode absorbs the forward current between the base and the collector. Being a majority-carrier device, the Schottky diode has negligible minority-carrier injection.
Looking back to the first point-contact transistor, an ideal Schottky diode as the emitter would have provided negligible minority-carrier injection for transistor action. The observation of transistor action must mean that the point-contact Schottky emitter had non-ideal Schottky behavior such that minority-carrier injection was increased. This could arise in the presence of a thin dielectric layer between the metal wire and the semiconductor. A Schottky diode under large forward bias is also known to increase the minority-carrier injection. Most likely, these contacts were created by a forming process (by a current pulse) in which diffusion of impurities occurred and local p-n junctions were formed.