What Is a Power MOSFET? Definition and Use

Key points

• Power MOSFETs switch by gate voltage, so drive power is Qg × VGS × frequency rather than a continuous base current
• Conduction loss is I² × RDS(on), and RDS(on) rises 40–70% from 25 °C to 125 °C
• Silicon MOSFETs dominate below 200 V; above 600 V silicon carbide wins on total loss

The structure, in one paragraph

A power MOSFET is built as a vertical structure: current flows from the drain on the die’s underside, up through a lightly doped drift region, across a channel formed in the body region under the gate, and out through the source on the top surface. The drift region sets the blocking voltage and, unfortunately, also sets the on-resistance. Thicker drift region means higher voltage rating and higher RDS(on) — this single trade-off explains most of the MOSFET market.

Two structural families matter for selection. Planar devices put the channel on the surface; they are rugged, cheap and stable, and dominate above 400 V. Trench devices dig the gate into a trench to increase channel width per unit area; they dominate below 200 V. A refinement of trench, the shielded gate trench (SGT), adds a source-connected shield under the gate that cuts gate charge — the reason modern 30–100 V synchronous rectifiers switch so cleanly.

Where the loss goes

Every MOSFET dissipates in three places, and a design that ignores any one of them will run hot.

Loss type	Formula	What controls it
Conduction	I²RMS × RDS(on) at temperature	Die area, voltage class, gate drive voltage
Switching	½ × VDS × ID × (tr + tf) × f	Gate charge, gate driver current, loop inductance
Drive	Qg × VGS × f	Gate charge, switching frequency
Output capacitance	Eoss × f	Coss non-linearity, bus voltage

At 50 kHz and 100 V, conduction usually dominates and a bigger die is the answer. At 500 kHz, switching and drive loss dominate and a smaller Qg matters more than a lower RDS(on). This is why “best MOSFET” has no universal answer: the ranking depends on your frequency, your current waveform and your cooling.

Choosing a voltage class

The rule that survives contact with real designs is: pick the class that gives 20–30% margin over the peak drain voltage including ringing, then stop.

Application	Bus voltage	Typical class	Why
12 V synchronous buck	12 V	30 V	Ringing is small, margin is cheap
48 V DC/DC, PoE	48–57 V	80–100 V	Load dump and hot-plug transients
48 V battery, power tools	54 V peak	60–80 V	Surge from motor inductance
Universal AC adapter	375 V DC	650 V	Mains plus flyback leakage ringing
PFC boost, 400 V output	400 V	650 V	Boost diode recovery spike
Solar boost, 600 V string	600 V	800 V or 900 V	Cold-weather open-circuit voltage rise

Jumping a class to “be safe” costs RDS(on) at the same die size — roughly double for one step at high voltage — which turns into heat and price. The correct move is to fix the clamp and layout first, then choose the class.

What the gate actually needs

A power MOSFET is not a logic gate input. At 10 A drain current, the gate is a capacitor of a few nanofarads that must be charged through the driver’s output impedance, and the switching time that results is what produces the loss.

Two practical numbers matter. First, Qg — if the part has 20 nC of gate charge and you switch at 200 kHz with 10 V drive, the driver dissipates 40 mW per device before it does any work. Second, the plateau voltage — during switching the gate sits at the Miller plateau while the drain voltage swings, and a weak driver keeps it there longer, which is where shoot-through and slow-turn-on loss come from.

Driving from a microcontroller pin directly is possible below about 100 kHz and a few amps, but the pin’s 20–50 mA source current turns a switching edge into a multi-microsecond ramp. That is a linear-mode operation, not a switching one, and the device will fail even though every rating in the datasheet looked comfortable.

Where MOSFETs stop being the answer

Three boundaries are worth stating plainly, because they are where designs fail rather than where they are merely suboptimal.

Above roughly 900 V silicon MOSFETs exist but their drift resistance makes them expensive and hot. Silicon carbide at 650 V and 1200 V carries the same current with lower total loss at high frequency, and IGBTs handle very high currents at low frequency.

In continuous body-diode conduction, a MOSFET is a poor diode. The body diode drop is around 1 V in silicon and 4–5 V in SiC, so any topology that parks current in the body diode for a significant fraction of the cycle is a thermal design error.

In linear mode, power MOSFETs have a narrow safe operating area. A device rated 100 A continuous can be destroyed by 5 A at 50 V for a few milliseconds if the die heats faster than the package can spread it. If your application includes hot-plug inrush or electronic load behaviour, check the SOA curve rather than the current rating.

What to look for in a supplier

For a switching design, the datasheet is only as useful as the conditions attached to it. A supplier that publishes RDS(on) at 25 °C only, Trr without a temperature, and no avalanche test conditions is telling you the numbers were measured for a catalogue page rather than for a design review.

The documents that actually decide a second source are: a full datasheet with characterised curves at your gate voltage, a lot-specific test record, and reliability data for the specific part number. Our MOSFET product pages state those per family, and the glossary defines each parameter with the practical consequence rather than the textbook definition.

This article describes general manufacturing practice for machined and fabricated components. It is not a substitute for a drawing-specific engineering review — tolerances, material selection and process capability depend on the individual part.