N-polar GaN HEMTs Predicted for High-Power Terahertz Sources - Bioengineer.org

The N-polar GaN HEMT shift is a physics-level architecture change, not a process tweak. P-polar structures hit a hard ceiling at THz frequencies—gate leakage and breakdown voltage cap out. N-polar repositions the gate on the AlGaN barrier facing the substrate, fundamentally redistributing the electric field. This is a regime change, not an iteration. Upstream, MOCVD epitaxy must be re-qualified: N-polar c-plane orientation alters V-pit density and dislocation propagation, forcing IQE, Sumitomo, and Sanan into a 12-to-18-month revalidation cycle. Downstream, flip-chip packaging and on-chip antenna coupling—already the THz bottleneck—will see design-rule rewrites as device impedance shifts. Competitively, Wolfspeed and Qorvo have P-polar process IP deeply coupled to 5G mmWave production lines. Pivoting to N-polar for THz means either a parallel fab line or a foundry partnership—neither is capital-light. Chinese GaN players, operating under MOCVD export restrictions, may actually find N-polar a clean-sheet window where legacy P-polar IP moats are less entrenched. The 12-to-24-month tail: expect an IP filing wave in N-polar gate dielectrics, at least one European or Japanese IDM pilot-line announcement, and 3GPP Release 20 THz band power specs that will implicitly favor N-polar architectures. The real value is not the transistor itself—it is system-level integration and packaging IP. Whoever defines the packaging standard first locks in THz-era dominance.