How a Transistor Works: From Silicon to the MOSFET

Q: Why is silicon used for transistors instead of another material?

Silicon is abundant and cheap, made from ordinary sand (SiO2). Its natural oxide, SiO2, is an excellent insulator that is essential for the MOSFET's gate. Its band gap of 1.12 eV is high enough to stay stable at room temperature yet low enough to be switched by small voltages, and seventy-plus years of process engineering have made silicon fabrication extremely mature and cheap to scale.

Q: Why is CMOS logic so energy-efficient?

The SiO2 gate oxide is an insulator, so essentially no current flows from the gate into the channel — the gate draws almost zero static current. CMOS logic pairs N-channel and P-channel MOSFETs so that in any steady logic state, one of the pair is always off, blocking the path from supply to ground. Power is only dissipated during the brief moment a gate switches state, not while it holds a value, which is what makes CMOS chips so efficient at billions of switches per second.

A three-terminal switch that changed everything

A transistor is a three-terminal semiconductor device that can amplify or switch an electrical signal. In digital electronics it behaves as a binary switch — on (conducting) or off (blocking) — which maps directly onto the 1s and 0s of computation. It was invented in December 1947 at Bell Labs by William Shockley, John Bardeen and Walter Brattain, work that earned the trio the 1956 Nobel Prize in Physics and launched the digital age. Modern MOSFET transistors have three connections: Source (current enters), Drain (current leaves) and Gate (the control input that switches it on or off) — in a bipolar transistor these are called Emitter, Collector and Base instead.

Why silicon, and what doping does

Silicon is a semiconductor: its conductivity sits between a metal and an insulator. Pure crystalline silicon has four valence electrons, each shared in a covalent bond with a neighbouring atom, leaving very few free charge carriers — so it conducts poorly on its own. Doping — deliberately adding trace impurities — changes that. Adding phosphorus (five valence electrons) leaves a spare electron free to move, creating N-type silicon that conducts via electrons. Adding boron (three valence electrons) creates a "hole" that behaves as a mobile positive charge, giving P-type silicon that conducts via holes. Silicon wins as the industry's material of choice because it is abundant and cheap (made from ordinary sand), because its natural oxide SiO₂ is an excellent insulator, and because seventy-plus years of process refinement have made it extraordinarily well understood.

live demo · I-V curves, load line and Q-point● LIVE

The p-n junction and the MOSFET switch

Where P-type and N-type silicon meet, a p-n junction forms: electrons diffuse across and recombine with holes, leaving a depletion region whose built-in electric field (about 0.6–0.7 V for silicon) resists further diffusion. That junction is the basis of every diode, transistor, LED and solar cell. The MOSFET (Metal-Oxide-Semiconductor Field-Effect Transistor), first built in 1960 and now the type behind almost every modern chip, puts a thin SiO₂ gate oxide over the channel between an N-type source and drain sitting in a P-type substrate. With no gate voltage the channel has no free electrons and blocks current — off. Once the gate voltage passes the threshold voltage (typically 0.3–0.7 V in modern chips), its field pulls electrons into the channel, forming a conductive inversion layer from source to drain — on. Because SiO₂ is an insulator, almost no current leaks from gate to channel, so a MOSFET draws essentially zero static current — the reason CMOS logic is so energy-efficient, dissipating power only at the moment it switches.

From one switch to a billion-transistor CPU

CMOS logic pairs an N-channel and a P-channel MOSFET to build gates: a NOT gate needs just two transistors, NAND and NOR each use four, an XOR uses eight, and a 1-bit full adder needs around 28. Stack enough of these and you get an 8-bit adder (~224 transistors), a 64-bit floating-point unit (millions), and a full modern CPU (tens of billions). In 1965 Intel co-founder Gordon Moore observed that transistor counts on a chip doubled roughly every 18–24 months — Moore's Law — a trend that held for fifty years, taking chips from 2,300 transistors (the 1971 Intel 4004) to roughly 20 billion in 2026's leading-edge processors, where a single transistor gate is only about ten silicon atoms wide and quantum tunnelling starts to set physical limits. Beyond silicon, the field is exploring 3D transistor stacking, wide-bandgap materials like GaN and SiC for power electronics, carbon nanotube transistors, spintronics and fundamentally different paradigms like quantum computing.

Frequently asked questions

Why is silicon used for transistors instead of another material?

Silicon is abundant and cheap, made from ordinary sand (SiO₂). Its natural oxide, SiO₂, is an excellent insulator essential for the MOSFET's gate. Its band gap of 1.12 eV is high enough to stay stable at room temperature yet low enough to be switched by small voltages, and seventy-plus years of process engineering have made silicon fabrication extremely mature and cheap to scale.

How does a MOSFET switch on and off?

An N-channel MOSFET has a P-type silicon substrate with two N-type regions, the source and drain, separated by a channel under a thin SiO₂ gate oxide. With no voltage on the gate, the channel has no free electrons and current cannot flow — the transistor is off. Once the gate voltage exceeds the threshold voltage, its electric field attracts electrons into the channel, forming a conductive inversion layer between source and drain, switching the transistor on.

Why is CMOS logic so energy-efficient?

The SiO₂ gate oxide is an insulator, so essentially no current flows from the gate into the channel — the gate draws almost zero static current. CMOS logic pairs N-channel and P-channel MOSFETs so that in any steady logic state one of the pair is always off. Power is only dissipated during the brief moment a gate switches state, not while it holds a value, which is what makes CMOS chips so efficient at billions of switches per second.

Try it live

Everything above runs in your browser — open Transistor BJT / MOSFET and drag the load line to watch a device move between cutoff, active/triode and saturation as you change base current, gate voltage and supply resistance. Nothing is installed, nothing is uploaded.

▶ Open Transistor BJT / MOSFET simulation