Intrinsic polarization superjunctions - Semiconductor Today

Intrinsic polarization superjunctions mark a paradigm shift, not an incremental upgrade. Replacing precision-alternating doping columns with fixed ferroelectric polarization charges sidesteps the two-decade GaN p-type doping bottleneck and decouples superjunction yield from dopant uniformity. Upstream, Veeco and AMEC see doping-precision specs de-prioritized; the new bottleneck shifts to polarization-layer control and AlGaN/InAlN epitaxial wafer quality, elevating material suppliers' (I-HQ in Taiwan, China; JX Nippon) bargaining power. Downstream, the RDS(on)-BV trade-off for 800V/1200V devices is redrawn, pressuring BOM costs in EV traction inverters and data-center power stages. Supply-chain risk migrates from equipment chokepoints to material chokepoints. Concentration in a handful of epitaxial wafer fabs makes long-term offtake contracts existential. Ion-implantation vendors (Axcelis, ASMPT) face structural order erosion as that process step shrinks. Competitively, Infineon and STMicro hold deep SiC superjunction patent moats. A validated polarization path hands Rohm, Navitas, and other GaN players an architectural leapfrog window. Expect dense cross-licensing negotiations and IEEE/IEC standard-definition battles within 12-24 months. The 24-month tail: if polarization-charge stability clears AEC-Q101 for 25-year automotive life, SiC's cost edge in superjunctions compresses sharply, and GaN power devices become credible in 800V traction by 2027. This is the industry's structural pivot from doping engineering to polarization engineering.