Metal–Oxide–Semiconductor Field-Effect Transistor

A metal–oxide–semiconductor field-effect transistor (MOSFET) is a transistor in which the voltage applied to an insulated gate controls electric current through a semiconductor channel between source and drain terminals. The gate produces an electric field that changes the concentration of mobile charge carriers without requiring substantial steady gate current. MOSFETs serve as switches and amplifiers, both as individual components and within integrated circuits. Their complementary arrangement, complementary metal–oxide–semiconductor (CMOS), underlies much digital electronics. (infineon.com)

Structure and operating principle

A conventional bulk MOSFET has four terminals: gate, source, drain, and body, also called the bulk or substrate. The gate is separated from the semiconductor by a thin dielectric layer. In a basic planar n-channel device, heavily doped n-type source and drain regions lie in a p-type body. Their electrical properties are established through semiconductor doping. The gate lies over the region connecting them. (ocw.mit.edu)

A sufficiently positive gate voltage attracts electrons toward the semiconductor surface and repels holes. Above the threshold voltage, the surface develops an inversion layer: its dominant mobile carriers differ from those in the underlying body. This conducting channel links source and drain. A drain-to-source voltage then drives carriers along it, while gate voltage controls its carrier concentration. A p-channel device operates with holes and reversed voltage polarities. (ocw.mit.edu)

The gate ideally draws no direct current, but its capacitance must be charged and discharged whenever its voltage changes. Consequently, switching requires transient gate current. Body voltage also affects channel formation: in conventional bulk devices, changing source-to-body bias shifts threshold voltage, a phenomenon called the body effect. (infineon.com)

Device types and operating regions

MOSFETs are classified as n-channel or p-channel and as enhancement-mode or depletion-mode. Enhancement-mode devices are normally off at zero gate-to-source voltage; an appropriate gate bias creates a conducting channel. Depletion-mode devices already conduct at zero gate bias, and a bias of the opposite polarity suppresses conduction. (infineon.com)

For an ideal long-channel enhancement-mode n-channel MOSFET, let $V_{GS}$ denote gate-to-source voltage, $V_{DS}$ drain-to-source voltage, and $V_T$ threshold voltage. Three operating regions are conventionally distinguished:

• Cutoff: $V_{GS}<V_T$; current is neglected in the simplest model.
• Triode or linear region: $V_{GS}>V_T$ and $0\leq V_{DS}<V_{GS}-V_T$. The channel extends between source and drain, and the device behaves approximately as a voltage-controlled resistor at small $V_{DS}$.
• Saturation: $V_{GS}>V_T$ and $V_{DS}\geq V_{GS}-V_T$. The channel becomes pinched off near the drain, and current depends mainly on gate overdrive rather than drain voltage. Pinch-off does not mean that current stops. (ocw.mit.edu)

In the ideal saturation model,

where $\mu_n$ is electron mobility, $C_{\mathrm{ox}}$ gate-oxide capacitance per unit area, and $W/L$ the channel width-to-length ratio. This square-law expression is an approximation, not a universal description of MOSFET behavior. (ocw.mit.edu)

MOSFET saturation is useful for amplification; it should not be confused with the low-voltage saturation state of a bipolar transistor. A low-loss MOSFET switch normally operates in the low-resistance portion of the triode region when on. (ocw.mit.edu)

Circuit applications and power devices

CMOS circuits combine n-channel and p-channel MOSFETs. In a basic static logic gate, complementary networks connect the output to one supply rail while blocking a steady conducting path to the other. This produces low static power consumption, although leakage and switching still consume energy. CMOS supports processors, memory, and mixed digital–analog circuits. MOSFETs also provide controlled conduction for analog amplification. (infineon.com)

Power MOSFETs are optimized for switching larger currents and blocking higher voltages. Many use a vertical current path and connect body and source internally, leaving three external terminals. This arrangement includes an intrinsic p–n junction body diode, which can conduct reverse current independently of gate control. Its reverse-recovery behavior affects switching losses and transients. (infineon.com)

Important power-device parameters include on-state resistance $R_{DS(on)}$, breakdown voltage, gate charge, thermal resistance, and safe operating area. Conduction losses are approximately $I_D^2R_{DS(on)}$, while gate charging and overlapping voltage and current during transitions contribute additional losses. Threshold voltage indicates the onset of conduction under specified test conditions, not necessarily the gate voltage needed for low-resistance operation. (infineon.com)

Historical development and scaling

Mohamed Atalla and Dawon Kahng fabricated a successful insulated-gate field-effect transistor at Bell Labs in 1959 and demonstrated MOS operation in 1960. Thermally grown silicon dioxide reduced troublesome semiconductor surface states, allowing the gate field to control surface conduction effectively. (computerhistory.org)

The historical name does not require every later device to retain the original materials. Doped polysilicon gates became widespread, while advanced processes subsequently introduced metal gates and high-permittivity dielectrics. These dielectrics permit a physically thicker insulating layer while preserving strong gate coupling, reducing leakage caused by quantum tunneling. (intel.com)

Scaling also changed channel geometry. FinFETs use raised, fin-shaped channels with gate control on multiple sides. Gate-all-around transistors surround channels such as stacked nanosheets, strengthening electrostatic control at small dimensions. IBM researchers demonstrated stacked nanosheet gate-all-around devices in 2017, including metal-gate integration and multiple threshold voltages. These architectures extend the insulated-gate operating principle beyond the original planar structure. (research.ibm.com)