Modeling the Impact of Process Variation on Critical Charge Distribution

Abstract

In this paper, we investigate the impact of process variation on soft error vulnerability with Monte Carlo analysis. Our simulation results show that Qcritical variation (3sigma/mean) of four types of storage circuits caused by process variation can be as large as 13.6%. We also propose an empirical model to estimate the Qcritical variation caused by gate length and threshold voltage variations. Simulation results show that this simple model is very accurate. Based on this model, the dependence of Qcritical variation on gate length variation, threshold voltage variation, and correlation between gate lengths is studied, using 70 nm SRAM as benchmark circuit.