A semiconductor device capable of amplification in addition to rectification. It is the basic unit in radio, television, and computer circuits, having almost completely replaced the thermionic valve. The point-contact transistor, which is now obsolete, was invented in 1948. It consists of a small germanium crystal with two rectifying point contacts attached to it; a third contact, called the base, makes a low-resistance nonrectifying (ohmic) connection with the crystal. Current flowing through the device between the point contacts is modulated by the signal fed to the base. This type of transistor was replaced by the junction transistor, which was developed in 1949–50. The field-effect transistor (FET) was a later invention. Bipolar transistors, such as the junction transistor, depend on the flow of both majority and minority carriers, whereas in unipolar transistors, such as the FET, the current is carried by majority carriers only.
In the bipolar junction transistor, two p-type semiconductor regions are separated by a thin n-type region, making a p−n−p structure. Alternatively, an n−p−n structure can also be used. In both cases the thin central region is called the base and one outer region of the sandwich is called the emitter, the other the collector. The emitter–base junction is forward-biased and the collector–base junction is reverse-biased. In the p−n−p transistor, the forward bias causes holes in the emitter region to flow across the junction into the base; as the base is thin, the majority of holes are swept right across it (helped by the reverse bias), into the collector. The minority of holes that do not flow from the base to the collector combine with electrons in the n-type base. This recombination is balanced by a small electron flow in the base circuit. The diagram illustrates the (conventional) current flow using the common-base type of connection. If the emitter, base, and collector currents are Ie, Ib, and Ic, respectively, then Ie=Ib+Ic and the current gain is Ic/Ib.
Field-effect transistors are of two kinds, the junction FET (JFET or JUGFET) and the insulated-gate FET (IGFET; also known as a MOSFET, i.e. metal-oxide-semiconductor FET). Both are unipolar devices and in both the current flows through a narrow channel between two electrodes (the gate) from one region, called the source, to another, called the drain. The modulating signal is applied to the gate. In the JFET, the channel consists of a semiconductor material of relatively low conductivity sandwiched between two regions of high conductivity of the opposite polarity. When the junctions between these regions are reverse-biased, depletion layers form, which narrow the channel. At high bias the depletion layers meet and pinch-off the channel completely. Thus the voltage applied to the two gates controls the thickness of the channel and thus its conductivity. JFETs are made with both n-type and p-type channels.
In the IGFET, a wafer of semiconductor material has two highly doped regions of opposite polarity diffused into it, to form the source and drain regions. An insulating layer of silicon dioxide is formed on the surface between these regions and a metal conductor is evaporated on to the top of this layer to form the gate. When a positive voltage is applied to the gate, electrons move along the surface of the p-type substrate below the gate, producing a thin surface of n-type material, which forms the channel between the source and drain. This surface layer is called an inversion layer, as it has opposite conductivity to that of the substrate. The number of induced electrons is directly proportional to the gate voltage, thus the conductivity of the channel increases with gate voltage. IGFETs are also made with both p-type and n-type channels. Because MOS devices cannot be formed on gallium arsenide (there are no stable native oxides of GaAs), metal semiconductor FETs (MESFET) devices are used. This makes use of Schottky barrier (see Schottky effect) as the gate electrode rather than a semiconductor junction.