An efficient stochastic integral equation method for modeling the influence of conductor surface roughness on interconnect ohmic loss

Abstract

Due to the shrinking skin depth in high-frequency analog and digital circuits, surface roughness now has a significant impact on interconnect ohmic loss, which tends to be ignored in conventional EDA tools. An effective conductivity (EC) has been proposed as the metric of this impact, but the large amount of computation resulted from the Monte-Carlo (MC) simulation prohibits practical use. In this paper, we propose a computationally efficient stochastic integral equation (SIE) method to statistically compute the EC as an alternative to the timeconsuming MC simulation. An effective second-order correction utilizing model order reduction (MOR) techniques is developed to substantially improve the accuracy at an acceptable cost. Numerical experiments then verify the efficacy of the proposed method. ©2007 IEEE.