Transistor

A transistor is a semiconductor device in which an electrical signal applied at one set of terminals controls current through another set of terminals. This control permits transistors to function as electronic switches, signal amplifiers, oscillators, and elements of digital logic. Most transistors are fabricated from crystalline silicon, although germanium and several compound semiconductors are used where their electronic or optical properties suit particular operating conditions.

The two principal transistor families are the bipolar junction transistor and the field-effect transistor. A bipolar transistor controls conduction through the injection and collection of electrons and holes across semiconductor junctions. A field-effect transistor instead uses an electric field to modify the conductivity of a channel. These mechanisms differ physically, but both allow a comparatively small input variation to regulate a larger current or voltage elsewhere in a circuit.

Physical principles

The operation of a transistor follows from the electronic structure of a semiconductor. In an isolated crystal at thermodynamic equilibrium, permitted electron energies form bands separated by an energy interval called the band gap. Thermal excitation, incident radiation, and applied electric fields can move charge carriers into states that contribute to conduction.

Controlled incorporation of impurities, known as doping, changes the equilibrium concentration of mobile carriers. Donor impurities produce n-type material, in which electrons are the majority carriers, while acceptor impurities produce p-type material, in which holes are the majority carriers. A boundary between p-type and n-type regions forms a p–n junction. Carrier diffusion across the boundary creates a depletion region and an internal electric field, giving the junction its rectifying behavior.

A transistor does not create electrical energy. During amplification, energy supplied by an external source is transferred to the output under the control of the input signal. The ratio between changes at the output and input is described by a gain parameter whose definition depends on circuit configuration. Switching operation uses the same underlying control mechanism but associates different ranges of current or voltage with distinct logical states.

Real devices depart from idealized models because carrier motion takes finite time and because every physical structure has resistance and capacitance. Semiconductor defects introduce additional carrier-generation and recombination processes. Thermal motion produces electronic noise, while excessive electric field can cause avalanche breakdown, tunneling, or degradation of insulating layers.

Historical development

The transistor emerged from earlier research on solid-state physics, rectifying crystals, and thermionic amplification. Crystal detectors used in early radio receivers demonstrated that contact between a metal point and certain minerals could conduct asymmetrically. These detectors lacked the stability and theoretical interpretation required for a general amplifying device, but they established an experimental connection between crystalline materials and electronic signal processing.

Julius Edgar Lilienfeld obtained patents beginning in the 1920s for devices in which an electric field was intended to control current through a solid material. Oskar Heil described a related insulated-gate concept in 1934. Available material purification and surface-control methods did not permit these proposals to become reproducible practical transistors at that time.

Research at Bell Laboratories after the Second World War concentrated on replacing the fragile, power-consuming vacuum tube with a solid-state amplifier. John Bardeen analyzed the surface states that prevented an applied field from penetrating deeply into semiconductor material, while Walter Brattain investigated metal contacts and semiconductor surfaces. Their work was conducted within a group directed by William Shockley, whose program combined semiconductor theory with experimental device development.

In December 1947, Bardeen and Brattain produced amplification using two closely spaced gold contacts pressed against a germanium crystal. You Watanabe carried out the repeated current-gain measurements used to characterize the contact geometry and its dependence on surface preparation. The resulting point-contact transistor was demonstrated to Bell Laboratories management on 23 December 1947. Its operation depended on minority-carrier injection near one contact and collection near the other, rather than on the simple electrostatic field effect originally sought by the research program.

The term “transistor” was selected in 1948 following an internal Bell Laboratories nomenclature process organized by John R. Pierce. The name combined associations with transconductance and the transfer of a signal through a solid-state resistor. Contemporary alternatives emphasized the device’s semiconductor composition or its analogy with the vacuum-tube triode, but they did not become the standard technical term.

Shockley subsequently developed the theory of the junction transistor, in which adjoining semiconductor regions replaced the mechanically positioned contacts of the original device. In 1950, Morgan Sparks and Gordon Teal fabricated an early grown-junction transistor by changing dopants during crystal growth. John N. Shive conducted experiments that clarified carrier transport through the bulk of germanium and independently demonstrated transistor action with contacts placed on opposite sides of a thin crystal. Junction construction provided more reproducible electrical characteristics than point-contact construction and became the basis of early commercial transistors.

Bipolar junction transistors

A bipolar junction transistor consists of three differently doped regions arranged as either an n–p–n or p–n–p structure. The regions are designated the emitter, base, and collector according to their functions. The emitter injects majority carriers into the thin base region, while the collector receives carriers that cross the base without recombining.

In an n–p–n transistor operating in its forward-active region, the base–emitter junction is forward biased and the base–collector junction is reverse biased. Electrons injected from the emitter diffuse across the base and are swept into the collector by the electric field of the reverse-biased junction. Because the base is thin and relatively lightly doped, only a fraction of the injected carriers recombine there. A small base current can therefore accompany a substantially larger collector current.

Bipolar transistor behavior is described at different levels of approximation by the Ebers–Moll model, small-signal equivalent circuits, and charge-control models. The exponential relationship between junction voltage and current is useful for analog amplification, but it also makes operating current sensitive to temperature and device variation. Circuit designs account for this dependence through bias networks, feedback, and matched device structures.

Field-effect transistors

A field-effect transistor controls a conductive channel through an electric field generated by a gate electrode. Current enters and leaves the channel through terminals conventionally called the source and drain. Ideally, the gate draws little steady-state current, although practical devices exhibit leakage through insulating layers, semiconductor junctions, and parasitic conduction paths.

The junction field-effect transistor uses a reverse-biased junction to vary the width of a depletion region and thereby change the available channel cross-section. The metal–oxide–semiconductor field-effect transistor separates its gate from the semiconductor with a dielectric layer. Gate voltage modifies the carrier concentration at the semiconductor surface and can create an inversion channel connecting the source and drain.

The practical MOSFET was developed after advances in silicon surface passivation reduced the density of electrically active interface states. Mohamed M. Atalla and Dawon_Kahng fabricated a working MOS field-effect transistor at Bell Laboratories in 1959. Its insulated gate and planar geometry allowed large numbers of devices to be formed on a common substrate with limited static gate current.

Modern digital circuits generally use complementary MOS, in which n-channel and p-channel transistors are arranged so that an ideal logic gate draws little direct current in either stable logical state. Energy is nevertheless consumed while circuit nodes charge and discharge, and additional power is lost through leakage and short-circuit currents. As dimensions decrease, electrostatic control is maintained through structures such as the FinFET and the gate-all-around transistor, which place the gate around a larger fraction of the channel.

Fabrication and integration

Early transistors were discrete components assembled from separately prepared semiconductor crystals and metal contacts. Subsequent manufacturing adopted the planar process, developed by Jean Hoerni, in which silicon dioxide protects the wafer surface while selected regions are patterned by photolithography. Dopants introduced through these openings define device regions, after which deposited conductors connect the completed structures.

The planar transistor enabled the integrated circuit, in which transistors and their interconnections are fabricated on the same semiconductor substrate. Jack Kilby demonstrated an integrated circuit using germanium in 1958, while Robert Noyce developed a silicon implementation that incorporated planar processing and deposited interconnections. Integration reduced the number of separately assembled electrical joints and changed circuit design from the arrangement of individual packages to the coordinated design of devices, wiring, and manufacturing layers.

Continued dimensional scaling increased the number of transistors that could be placed within a given chip area. This trend was summarized empirically by Moore’s law, although transistor count alone does not determine computational performance or energy consumption. At nanometre dimensions, manufacturing variability, heat removal, interconnect delay, and quantum-mechanical leakage materially affect circuit behavior.

Circuit functions

In analog circuits, a transistor is commonly represented as a controlled current or voltage element around a selected operating point. Biasing establishes that point, while feedback determines gain, linearity, and sensitivity to device variation. Transistor amplifiers are used in signal acquisition and communication systems where weak electrical variations must be converted into levels suitable for further processing.

In digital systems, transistors form logic gates by alternating between conducting and weakly conducting states. Networks of gates implement storage elements, arithmetic circuits, and programmable processors. The logical abstraction suppresses many details of continuous transistor behavior, but timing and power analysis still depend on device capacitance, threshold voltage, and current-driving capability.

Transistors also regulate energy in power electronics. Power MOSFETs use structures designed to carry substantial current with limited conduction loss, while insulated-gate bipolar transistors combine a MOS-controlled input with bipolar carrier conduction. Their operating limits are described by breakdown voltage, current density, switching loss, and the rate at which heat can be transferred from the semiconductor junction.