Microstructural modelling for prediction of effective properties in porous SOFC electrodes

A modelling framework for the prediction of effective properties in porous SOFC electrodes is presented. The model consists of: i) a packing algorithm to numerically reconstruct the microstructure, and ii) a Monte Carlo method to calculate the effective transport properties. This modelling technique improves the accuracy of the prediction of effective properties beyond percolation theory estimates. In addition, the numerical reconstruction does not rely on existing samples and complex instrumentations (for example, X-ray tomography, FIB-SEM analyses) required by other reconstruction methods. The packing algorithm enables to numerically generate a representative sample of the electrode microstructure with the desired particle size distribution, composition and porosity. Sintering phenomena are simulated by increasing the overlap among the particles, pore-former particles are accounted for during the packing generation [1]. The model is capable to simulate packings of spherical particles as well as of agglomerates of spheres. The reconstructed samples are then analyzed with a Monte Carlo method [2]. Random walk simulations are used to determine the transport properties in gas and solid phase, such as the effective diffusivity and the effective electric conductivity. Other geometric quantities can be evaluated, such as the pore size distribution, the surface area exposed to the gas phase, the three-phase boundary length. In this study, effective properties as a function of porosity and particle size for random packings of spherical particles are calculated. The results are compared with independent experimental data, revealing a good agreement for both gas and solid phase properties. Effective properties of agglomerates of particles are also presented and compared with the results obtained for spherical particles. The comparison highlights that particle agglomeration significantly increases the mean pore size while reducing the effective gas diffusivity and the specific surface area exposed to gas phase (Figure 1). The presented modelling technique can be used to improve SOFC electrode design and to support the interpretation of experimental data.