Skip to content
Documentation
Advanced Transistor Tech
FinFET

FinFET

For about forty years, every transistor in every chip was “planar” — built flat, like a pancake, on the surface of the silicon wafer. Then, in the mid-2000s to early 2010s, the industry did something radical: it stood the transistor’s channel up on its edge, turning it into a thin vertical fin, and wrapped the gate around three of its four sides. That change — the FinFET — is one of the single biggest structural shifts in transistor history, and it’s the reason chips kept getting smaller and more efficient well past the point where flat transistors had run out of room.

The problem FinFETs were built to solve

A MOSFET works by using a gate to control whether current can flow between a source and a drain, through a channel underneath the gate. In a traditional planar transistor, the gate sits directly on top of a flat channel, controlling it from exactly one side. As transistors shrank generation after generation, that single point of control became a serious liability: the channel got so short that the gate increasingly struggled to shut off current flow completely, even in the “off” state. Current would leak through the channel anyway — a problem called short-channel leakage — wasting power and generating unwanted heat across billions of transistors.

🧩 Think of it like… trying to pinch a garden hose shut. If you press down on the hose from just one side (planar gate, one-sided control), a determined trickle of water can still squeeze past underneath, especially as the hose gets thinner and floppier. But if you wrap your entire hand around the hose from three sides, you get a much firmer, more complete shutoff. A FinFET's gate wraps around the channel from three directions instead of one, giving it dramatically better grip on when current flows and when it doesn't.

What actually changed: the channel becomes a fin

The core idea of a FinFET is simple to state, even though it took years of engineering to manufacture reliably: instead of laying the channel flat on the wafer surface, you etch it into a thin vertical ridge — the “fin” — standing up out of the silicon. The gate material is then deposited over the top of the fin and down both of its sides, wrapping around three of the fin’s four faces (the bottom stays attached to the substrate). This gives the gate far more surface area of contact with the channel, and control from three directions instead of one.

Planar MOSFET(cut along current flow: source → drain)SubstrateSourceDrainChannelGatecontrol from 1 side (top)

FinFET(cut across the fin, through the channel)

SubstrateFin (channel)Gatewraps 3 sides: top + 2 sidewalls

Note that the two cross-sections above are cut in different directions, which is standard practice when illustrating FinFETs: the planar view is cut along the direction current flows (source to drain), while the FinFET view is cut across the fin, at the channel, to show how the gate wraps around it. The fin’s source and drain sit further along its length, outside this particular slice.

Why three-sided control matters so much

The technical term for how completely a gate controls its channel is electrostatic control. More contact area between gate and channel means the gate’s electric field dominates the channel more thoroughly, which produces two big benefits: the transistor switches off far more completely when it should be off (less leakage, less wasted power), and it can be made shorter without losing that control — which is exactly what’s needed to keep shrinking transistors generation after generation. FinFETs became standard across the industry starting around the 22nm/14nm process generations (see What Does “7nm” Actually Mean? for why those node names don’t map directly to a physical measurement) and remained the dominant transistor architecture for roughly a decade.

FinFETs aren’t just theoretically better — they were a manufacturing triumph. Etching thousands of perfectly uniform, nanometer-precision vertical fins across an entire wafer, repeatably, was a major achievement that required advances across lithography, etching, and deposition (see Etching and Deposition).

The next step: going all the way around

FinFETs wrap the gate around three sides, but the bottom of the fin still connects to the substrate, leaving one side uncontrolled. The natural next step — controlling all four sides — is exactly what Gate-All-Around transistors do, and it’s where the industry has been transitioning as FinFETs themselves start to reach their own scaling limits.

Key takeaways

  • Planar transistors control their channel from only one side (the top), which becomes a leakage problem as channels get shorter.
  • FinFETs stand the channel up as a vertical “fin” and wrap the gate around three sides — top and both sidewalls — giving much stronger electrostatic control.
  • Better control means less current leakage when the transistor is off, and the ability to keep shrinking transistors without losing that control.
  • FinFETs became the industry-standard transistor structure starting around the 22nm/14nm generations.
  • The FinFET’s remaining limitation — an uncontrolled fourth side at the fin’s base — is what led to the development of gate-all-around transistors.
Gate-All-Around