An Anode Hole Injection Percolation Model for Oxide Breakdown - The 'Doom's Day' Scenario Revisited

This work explores stress dependent oxide breakdown using a comprehensive and unique numerical model that integrates, for the first time, the two leading theories of oxide breakdown: the theory of anode hole injection and the percolation theory. In sharp contrast to the traditional percolation (TP) model, this new nonuniform percolation (NP) theory (that includes the effects of surface roughness and localized feedback through the weak spots) successfully predicts experimentally observed stress dependence of the critical trap density (N sub (crit)). In this paper, we systematically (a) explore, interpret, and qualify the assumptions of TP model, (b) reexamine the predictions for ultimate gate oxide thickness based on reliability concerns, and (c) question the validity of voltage scaling used for accelerated testing for ultrathin oxides.