It was noted that standard simple reaction–diffusion models assumed an infinite oxide thickness and a zero initial interface trap density, even though this was true of real MOS devices. The reaction–diffusion model was solved numerically here by taking account of the finite oxide thickness and of an initial trap density. The results showed that trap generation/passivation as a function of stress/recovery time was strongly affected by the condition of the gate-oxide/poly-Si boundary. When an absorbent boundary was considered, the reaction–diffusion model was more consistent with measured interface-trap data from CMOS devices under a bias temperature stress. The results also showed that non-negligible initial traps should affect the power index, n, when a power-law dependence of the trap generation upon stress time, tn, was observed in the diffusion-limited region of the reaction–diffusion model.

Boundary Condition and Initial Value Effects in the Reaction–Diffusion Model of Interface Trap Generation/Recovery. L.Yong, H.Daming, L.Wenjun, L.Mingfu: Journal of Semiconductors, 2009, 30[7], 074008