power mosfet igbt cross section

# One Added Layer: How the IGBT Grew Out of the Power MOSFET

same vertical structure, one swapped layer, a very different tradeoff curve

THE ONE LAYER THAT CHANGES EVERYTHING Power MOSFET (VDMOS) source / gate (top) p-body / channel n⁻ drift region n⁺ drain (backside) electrons only — one carrier type resistance set purely by drift-region doping IGBT — identical, but one layer swapped gate / emitter (top) p-body / channel n⁻ drift region p⁺ collector — ADDED LAYER holes injected from below too conductivity modulation: drift region floods with carriers, Ron collapses cost: those holes linger at turn-off → tail current bipolar's base-drive wall never had a voltage-controlled gate at all — this MOSFET gate is what both devices share B. Jayant Baliga, General Electric, patent filed ~1982 latch-up risk: that added p⁺ layer forms a parasitic PNPN thyristor — fixed by a buffer layer and cell geometry

The power MOSFET already solved bipolar's base-drive and second-breakdown problems with a voltage-controlled gate and a positive temperature coefficient.

The IGBT keeps that gate, but trades the MOSFET's single-carrier simplicity for a second carrier type above the voltage where resistance would otherwise dominate.

WHY THE SWAP ONLY PAYS OFF ABOVE ~200-600V Rₒₙ blocking voltage → MOSFET: Rₒₙ ∝ V^2.4–2.6 IGBT: modulation holds Rₒₙ down crossover ≈ 200–600V MOSFET territory fast, no tail current IGBT territory lower Rₒₙ, accepts tail current below the crossover, the MOSFET's single-carrier simplicity wins outright — no reason to pay the tail-current cost applications: motor drives, industrial inverters, EV traction, solar/wind inverters

GaN and SiC wide-bandgap devices pick up this same tradeoff curve one generation further out — named, not retold, in this series' closing step

Go deeper with CFSGPT

Get AI-powered deep-dives, save terms, and run advanced simulations — free account.

Create Free Account