Optimization of the microstructure of porous electrodes plays an important role in the enhancement of the performance of solid oxide fuel cells. For this, microstructural models based on percolation theory have proven useful for the estimation of the effective material properties of the electrode material, assumed to consist of a binary mixture of spherical electron and ion conducting particles. In this work, we propose an extension of prior approaches for calculating the effective size of the three-phase boundary, which we judge to be physically more sound and, in particular, well suited for characterizing mixtures of particles of different sizes. This approach is then employed in a one-dimensional cell level model encompassing the entire set of processes of gas transport, electronic and ionic conduction as well as the electrochemical reactions. The impact of the electron and ion conducting particle sizes, their volume fraction and their size ratio on the performance of the fuel cell are investigated in a parametric study. Under certain conditions, cathode microstructures having electronic conducting particles of size different from that of the ionic conducting particles become preferable and yield a higher maximum power density when compared to the best possible configuration of monodisperse particles. (C) 2012 Elsevier B.V. All rights reserved.