What is a MOSFET.

MOSFET stands for Metal-Oxide-Semiconductor Field-Effect Transistor. That mouthful is actually a description of what's inside — each word names a layer in the device. You're literally reading the recipe.

Every MOSFET has four terminals:

• Gate (G) — the control input. A voltage here decides whether the switch is on or off.
• Drain (D) — where current exits.
• Source (S) — where current enters.
• Body (also called bulk or substrate) — usually tied to source; sets the reference voltage of the underlying silicon.

Between the gate and the rest of the transistor sits a very thin layer of insulator — historically silicon dioxide (SiO₂), and in modern process nodes a high-k dielectric like hafnium dioxide (HfO₂). This insulator is critical. It means the gate never actually touches the current path. Instead, the gate controls current through an electric field reaching across the insulator. That's the "field-effect" part of the name.

How it actually switches

Take the most common type — an N-channel enhancement-mode MOSFET:

1. When the gate-to-source voltage (V_GS) is below a certain threshold voltage (V_th), the region between drain and source is empty of charge carriers. No current flows. Switch is OFF.
2. Raise V_GS above V_th, and the electric field from the gate pulls electrons up to the surface of the silicon underneath. Those electrons form a conductive inversion channel connecting drain to source. Current flows. Switch is ON.

The elegant part: in steady state, almost no current flows into the gate itself. You're just presenting a voltage across the insulator. The gate looks like a tiny capacitor, not a resistor. This is why chips can pack billions of MOSFETs onto a fingernail-sized die without melting from control losses.

Why they replaced everything

Before MOSFETs took over integrated circuits in the 1970s, most electronics used bipolar junction transistors (BJTs). BJTs work — but they're current-controlled. You need a continuous base current to keep them switched on, which wastes power on a scale that made billion-transistor chips impossible.

MOSFETs are voltage-controlled. Once the gate is charged to the right voltage, it stays there — no continuous current required. This makes them:

• Massively more efficient for logic circuits
• Cheap and dense to fabricate in massive arrays (the entire modern chip industry depends on this)
• Well-suited to CMOS — the pairing of N-channel and P-channel MOSFETs that consumes power only when switching, which is the foundation of every modern digital chip from a phone to an H100 GPU

The two families

Every schematic labels MOSFETs clearly by type:

• NMOS (N-channel) — conducts when gate voltage is high (above V_th). Fast, but leaks a bit more at small process nodes.
• PMOS (P-channel) — conducts when gate voltage is low (below source by |V_th|). Slower than NMOS due to hole mobility being lower than electron mobility, but complementary.

CMOS logic pairs them so that at any stable output state, one is always off. Static power draw is nearly zero — energy is only spent during the brief moment of switching. This is why your phone doesn't drain the battery when the screen is idle displaying a still image, even though billions of transistors are technically "holding state."

Where you'll actually meet them

• Inside every chip — a modern GPU has ~80 billion of them, most as CMOS logic gates and cache memory cells
• Power supplies — as high-current switches in buck converters, LDOs, and motor drivers
• Battery protection — as the "gate" that disconnects a lithium cell when it overcharges or gets too hot
• Audio amplifiers — as linear amplification stages (BJTs still compete in this specific corner)
• Every Arduino / microcontroller project that switches a motor, LED strip, or relay uses discrete MOSFETs on the output

The schematic symbol is a T-shape with a small gap where the gate meets the channel — the gap represents the oxide insulator. The direction of an arrow on the source terminal tells you N-channel vs P-channel.

The one detail that trips students up

MOSFETs have a threshold voltage, and it isn't zero. A common logic MOSFET might specify V_GS(th) of 1.8V — meaning the switch starts to conduct at 1.8V — but need 4-5V at the gate to fully turn on. If your microcontroller outputs 3.3V but you picked a MOSFET that needs 5V to fully saturate, the switch is only partially on. It dissipates heat, drops voltage across itself, and fails to properly drive its load.

The workaround is a logic-level MOSFET — one specifically designed to fully turn on at 3.3V or even 1.8V gate drive. The datasheet will list this explicitly (something like "V_GS(on) = 2.5V max at I_D = 10A"). Always check that number against whatever's driving the gate. Getting this wrong causes roughly 90% of "my MOSFET circuit doesn't work" problems on hobbyist forums.