A transistor is a three-terminal semiconductor device in which a small voltage or current applied to one terminal controls a much larger current flowing between the other two. In hardware, it serves two fundamental roles: an electronic switch that represents the ones and zeros of digital logic, and an amplifier that enlarges weak analog signals. Every processor, memory chip, power supply and radio receiver is built from transistors.
Within semiconductors, the transistor is the basic building block of integrated circuits. A modern processor or system on a chip contains tens of billions of transistors etched into a single piece of silicon, and the steady shrinking of these devices since the first working transistor of 1947 explains most of the growth in computing performance over the following decades.
How a transistor controls current
A transistor works by changing the conductivity of a semiconductor region with an electrical signal. Pure silicon conducts poorly, but doping it with impurities creates n-type regions rich in free electrons and p-type regions rich in holes. Junctions between these regions form barriers that a control signal can lower or raise, which turns the path between the main terminals on, off or anywhere in between.
The two dominant families use different control mechanisms. In a bipolar junction transistor (BJT), a small base current injects charge carriers that allow a current up to a few hundred times larger to flow from collector to emitter. In a field-effect transistor (FET), a gate voltage creates an electric field that forms or removes a conductive channel between source and drain, with almost no steady gate current. Because a FET is voltage-controlled, it draws power mainly when it switches, which makes it the natural choice for dense digital logic.
The table below maps the terminals of the two main families and the role of each.
| Function | BJT terminal | FET terminal | What it does |
|---|---|---|---|
| Control input | Base | Gate | Receives the signal that sets the current |
| Current source side | Emitter | Source | Supplies the charge carriers |
| Current output side | Collector | Drain | Collects carriers and connects to the load |
| Control quantity | Base current | Gate-source voltage | Determines how strongly the device conducts |
Main transistor types and their uses
Transistor types differ in control mechanism, voltage rating, speed and efficiency, and each dominates a different part of electronics. The MOSFET (metal-oxide-semiconductor field-effect transistor), described on its own page, accounts for the overwhelming majority of transistors ever manufactured because it scales well and pairs naturally into CMOS logic. Other types remain essential in analog, radio-frequency and power circuits.
| Type | Control | Typical strengths | Typical applications |
|---|---|---|---|
| BJT | Base current | High gain, low noise, predictable analog behavior | Audio amplifiers, sensor interfaces, discrete switching |
| JFET | Gate voltage on a junction | Very high input impedance, low noise | Instrument preamplifiers, analog front ends |
| Planar MOSFET | Gate voltage through an oxide | Simple, cheap, easy to integrate | Older logic nodes, power switching, analog chips |
| FinFET | Gate wrapped on three sides of a fin | Better channel control at small sizes | Processors from about the 22 nm generation onward |
| Gate-all-around (GAA) | Gate surrounding stacked nanosheets | Best electrostatic control, lower leakage | Leading-edge logic at the 3 nm and 2 nm classes |
| IGBT | MOS gate driving a bipolar stage | Handles hundreds to thousands of volts | Motor drives, inverters, induction heating |
| GaN and SiC transistors | Gate voltage, wide-bandgap material | Fast switching, high voltage, low losses | Compact chargers, electric vehicle inverters |
The move from planar transistors to FinFET and then to gate-all-around structures was driven by one problem: as channels became shorter, the gate lost control over them and leakage current rose sharply. Wrapping the gate around more of the channel restores that control and lets designers keep the supply voltage near 0.7 to 0.9 V.
Key parameters engineers check
A transistor is specified by a handful of electrical limits and performance figures, and exceeding any absolute maximum rating can destroy it instantly. The values below are typical ranges for small-signal and logic devices; power transistors extend them by orders of magnitude.
| Parameter | Typical range | What it affects |
|---|---|---|
| Current gain (hFE, beta) | About 50 to 300 for small-signal BJTs | Base current needed to drive a load |
| Base-emitter voltage | About 0.6 to 0.7 V for silicon BJTs | Minimum drive voltage to turn the device on |
| Threshold voltage | About 0.2 to 0.5 V in modern logic; 1 to 4 V in discrete MOSFETs | Gate voltage at which a channel forms |
| On-resistance (RDS(on)) | Milliohms in power MOSFETs to ohms in small devices | Conduction loss and heating |
| Collector-emitter saturation voltage | About 0.1 to 0.3 V | Voltage lost across a fully on BJT switch |
| Leakage current | Picoamperes to nanoamperes per device | Standby power of large chips |
In digital chips, the decisive figure is switching energy. Dynamic power follows the relation P = α × C × V² × f, where α is the activity factor, C the switched capacitance, V the supply voltage and f the clock frequency. Because voltage enters squared, lowering it from 1.0 V to 0.8 V cuts dynamic power to 64 percent, a 36 percent saving at the same frequency, which is why each new transistor generation fights for lower operating voltage.
Transistor scaling and its physical limits
Transistor scaling is the process of making devices smaller so that more fit on a chip at lower cost per function. For decades, transistor counts doubled roughly every two years, the trend known as Moore’s law, and Dennard scaling let each smaller transistor also run at lower voltage with constant power density. Dennard scaling broke down in the mid-2000s, when leakage prevented further voltage reduction, which is why clock speeds stalled near a few gigahertz and the industry turned to multiple cores.
Node names such as 5 nm or 3 nm are marketing labels, not physical dimensions: no feature of a 3 nm transistor measures exactly 3 nm. Real progress is better judged by transistor density per square millimeter and by performance per watt. Scaling is also constrained by photolithography cost, by quantum tunneling through ultrathin gate oxides and by the physical subthreshold swing limit of about 60 mV per decade of current at room temperature, which restricts how sharply a conventional transistor can switch off. Approaches such as chiplets and 3D stacking now supplement shrinking.
Common mistakes and misconceptions
Most transistor failures in practice come from wrong drive conditions or ignored thermal limits, not from defective parts. The following points come up repeatedly in design reviews.
- Driving a BJT switch without enough base current leaves it in the linear region, where it heats up instead of saturating.
- Using a standard MOSFET with a 3.3 V logic signal often fails because its threshold and specified RDS(on) assume a 10 V gate drive; a logic-level MOSFET is required.
- Omitting a flyback diode across a relay coil or motor lets the inductive voltage spike destroy the transistor at turn-off.
- Assuming that a smaller node is always faster ignores design, voltage and thermal limits, which often matter more than the label.
- Treating the transistor as an ideal switch overlooks its leakage, capacitance and switching time, which dominate behavior at high frequency.
In simple terms
A transistor is a tiny electrically controlled valve. Think of a water tap: a small turn of the handle controls a large flow of water. In a transistor, a small signal plays the role of the handle, and it decides how much electric current passes through.
When the valve is fully open or fully closed, the transistor acts as a switch, and billions of such switches working together let a computer calculate. When it is partly open, it copies a weak signal in a stronger form, which is how a microphone signal becomes loud enough to drive a speaker.
A practical example
An engineer designs a controller in which a 3.3 V microcontroller pin must switch a 12 V relay coil that draws 60 mA. The pin cannot supply that current or voltage, so an NPN transistor acts as a low-side switch between the coil and ground. With a worst-case current gain of 100, the transistor needs at least 0.6 mA of base current.
To guarantee full saturation, the engineer supplies about four times that minimum. A 1 kΩ base resistor gives (3.3 V minus 0.7 V) divided by 1,000 Ω, or 2.6 mA of base current, well within the pin’s capability. A flyback diode across the coil absorbs the voltage spike at turn-off. In saturation, the transistor drops about 0.2 V, so it dissipates only about 12 mW (0.2 V × 60 mA) and runs cold, and the relay switches reliably for years.