Simulation study for lifetime distribution of middle-of-line time-dependent dielectric breakdown affected by global and local spacing variations

Abstract

The impact of process fluctuations for the lifetime distribution of middle-of-line time-dependent dielectric breakdown was investigated by Monte-Carlo simulation. Global and local variations were simulated using a doubly truncated normal distribution. A goodness of fit to the generated data was determined statistically in terms of the Weibull distribution and clustering model. A change in the standard deviation of the global variation shows a large contribution to lifetime variation. However, it does not affect the average lifetime. A change in the standard deviation of the local variation contributes to both the average lifetime and the distribution variation. The upper limit of the process variation induces the convex upward shape of the lifetime distribution on the Weibull plot. The clustering model well explains the distribution shape.