SiC MOSFET Body Diode Reliability

# SiC MOSFET Body Diode Degradation

Every power MOSFET has an intrinsic body diode built into its structure, and in a silicon device that diode is simply a known, stable electrical element — a forward drop to account for, nothing more. In a SiC MOSFET, forcing current through that same intrinsic body diode can trigger a slow, irreversible degradation of the device itself, because SiC's crystal structure carries a defect silicon never has to worry about: basal plane dislocations left over from substrate growth, which body-diode current recombination can actually cause to expand.

Basal plane dislocations are a growth-inherited crystal defect, and the energy released when carriers recombine near one is what drives it to grow. SiC substrate material, grown by a slow sublimation process, carries some population of basal plane dislocations — line defects lying along specific crystal planes, inherited from the growth process and present at some low density in essentially every commercial SiC wafer. Forward current through the body diode means electron-hole recombination happening throughout the device's drift region, and when that recombination occurs near a basal plane dislocation, the released energy can drive the dislocation to expand into a larger, triangular stacking fault — a genuinely new crystal defect that didn't exist before the body diode ever conducted.

The electrical consequence is a slow, cumulative rise in on-resistance and forward voltage drop, precisely because the growing stacking fault is itself a higher-resistivity region current has to route around. An expanded stacking fault isn't a catastrophic short or an open — it's a locally degraded region of the crystal with worse conductivity than the surrounding material, so current redistributes around it, and the device's overall on-resistance creeps upward measurably over repeated body-diode conduction cycles. Left unmanaged across a product's service life, that creeping Ron degradation can eventually push a device out of its own datasheet specification, a slow-motion reliability failure distinct from any of silicon power MOSFET's own aging mechanisms, since silicon's crystal structure simply doesn't carry an equivalent defect population to begin with.

Circuit-level avoidance, not device redesign, is the dominant real-world mitigation, because the simplest fix is just not relying on the body diode to conduct at all. Many SiC power circuits — synchronous rectification topologies especially — add an external, parallel Schottky diode specifically to carry freewheeling current instead of the MOSFET's own intrinsic body diode, sidestepping the degradation mechanism entirely by never forcing significant forward current through the vulnerable path in the first place. Where body-diode conduction can't be avoided, device manufacturers also screen substrate material for basal-plane-dislocation density and work to convert dislocations into less-harmful defect types during epitaxial growth, attacking the problem at the crystal-growth stage rather than leaving it entirely to circuit-level workarounds.

GaN dynamic on-resistance (611) and SiC body diode degradation both show up as the same electrical symptom — Ron creeping higher than the datasheet's best-case number — but through completely different physical mechanisms: trap states in an amorphous-enough buffer layer for GaN, a genuine crystal defect expanding inside the device for SiC. Wide-bandgap power semiconductors each inherited their own distinct reliability physics from the specific material system replacing silicon, not a single shared failure mode.