Theoretical basis for the statistics of dielectric breakdown

Abstract

The engineers' approach to dielectric breakdown has traditionally been one of testing leading to life prediction and material comparison. On the other hand, the scientist has attempted to understand the mechanism of specific breakdown processes. It is shown here that the former approach is prone to ambiguity and that a need exists to understand not just the physics of breakdown processes but also their resulting statistics. A number of statistical models are developed and discussed with the aid of experimental data. Finally it is shown how the general trends obtained can be used to aid the interpretation of the engineers' data.