What Is a Field-effect Transistor (FET)?

A field-effect transistor (FET) is a three-terminal semiconductor device in which a voltage applied to a gate terminal modulates the conductance of a channel between source and drain terminals by means of an electric field, without injecting current into the gate itself. The gate voltage controls the density of mobile charge carriers in the channel, either accumulating them to create a conductive path or depleting them to suppress current flow. This mechanism makes the FET a voltage-controlled current source with very high input impedance at the gate, a property that distinguishes it from bipolar junction transistors and makes it the preferred device for most digital logic and for many analog applications where loading the driving circuit must be minimized.

The theoretical foundation for the field effect was described by Julius Lilienfeld in patents filed in 1926, though working devices based on his proposals were not demonstrated at the time. Practical junction FETs emerged in the early 1950s, and the metal-oxide-semiconductor FET (MOSFET) followed in 1960 following work at Bell Laboratories by Dawon Kahng and Martin Atalla. The MOSFET became the transistor of modern computing; the number of MOSFETs produced annually is estimated to exceed ten quintillion, making it by far the most manufactured object in human history.

Device Structure and Operating Modes

A MOSFET consists of a semiconductor substrate, typically silicon, into which two heavily doped regions, the source and drain, are implanted. A thin oxide layer, historically silicon dioxide grown by thermal oxidation, separates the gate metal (or in modern devices, polysilicon or a metal gate) from the substrate surface between the source and drain. The gate voltage controls the inversion layer, or channel, that forms at the oxide-semiconductor interface. Enhancement-mode devices require a threshold gate voltage to form the channel and are normally off; depletion-mode devices have a pre-existing doped channel and are normally on, requiring the gate to pinch off conduction. N-channel MOSFETs conduct through electrons; p-channel MOSFETs conduct through holes. Pairing both types in complementary MOS (CMOS) logic minimizes static power dissipation, which is why CMOS has dominated digital integrated circuit design since the 1980s. Device physics at this level of detail is treated comprehensively in IEEE Xplore publications on MOS device fundamentals.

Scaling and Short-Channel Effects

The history of the FET in integrated circuits is largely the history of scaling: shrinking the gate length to pack more transistors onto a chip and increasing their switching speed. As gate lengths crossed into the deep-submicron range in the 1990s and then into the nanometer range after 2000, short-channel effects became the dominant engineering challenge. When the gate length is comparable to the depletion regions of the source and drain, the gate loses some electrostatic control over the channel, increasing leakage current in the off state. Solutions have included thinner gate dielectrics, then the replacement of silicon dioxide with high-permittivity (high-k) dielectrics such as hafnium oxide to maintain gate capacitance without thinning the physical layer further. FinFET structures, in which the channel wraps around a thin vertical fin of silicon, improved three-dimensional electrostatic control below the 22 nm node. Gate-all-around nanosheet FETs, introduced at the 3 nm node, surround the channel on all sides, continuing the trend. The International Roadmap for Devices and Systems maintained by the IEEE IRDS initiative documents these successive scaling approaches.

Compound Semiconductor FETs

Beyond silicon, FETs are built on compound semiconductors such as gallium arsenide (GaAs), indium phosphide (InP), and gallium nitride (GaN) for applications where silicon's electron mobility is insufficient. GaAs metal-semiconductor FETs (MESFETs) use a Schottky gate directly on the semiconductor surface rather than an oxide and were among the first transistors to achieve useful gain at microwave frequencies in the early 1970s. High-electron-mobility transistors (HEMTs), which confine carriers in a two-dimensional electron gas at a heterojunction interface, deliver the highest electron mobility of any field-effect device and are used in low-noise amplifiers and power amplifiers from X-band through millimeter-wave frequencies. GaN HEMTs are particularly important in high-power RF applications because GaN supports ten times the breakdown electric field of GaAs. Comprehensive treatment of compound semiconductor FETs appears in Wiley's Encyclopedia of RF and Microwave Engineering.

Applications

The field-effect transistor has applications in a wide range of disciplines, including:

• Digital logic and memory in microprocessors and DRAM
• Radiofrequency amplification and switching in wireless communications
• Power conversion in motor drives, inverters, and voltage regulators
• Low-noise amplification in satellite receivers and radio telescopes
• Biosensing through functionalized gate surfaces for medical diagnostics