Efficient numerical modeling of random rough surface effects in interconnect resistance extraction

Owing to the decreasing skin depth in high-speed analog and digital circuits, surface roughness is playing an increasingly important role in interconnect parasitic extraction. However, the random nature of surface roughness and the complicated electromagnetic behavior baffle satisfactory solutions to the extraction exercise. This paper utilizes a numerically formulated effective conductivity as an efficient measure of the rough surface effects in the resistance extraction to avoid the global discretization of interconnect surfaces. Two numerical methods based on different boundary element formulations are proposed to compute the mean effective conductivity for interconnects with random surface roughness. Numerical experiments compare these two methods in different rough surface patterns to evaluate their efficacy. Copyright (C) 2008 John Wiley & Sons, Ltd.