System identification, a modelling approach derived from control theory, is applied to the identification of reaction mechanisms in SOFC electrodes. The method is illustrated with a simple reaction model which includes the adsorption of molecular oxygen on the electrode surface and the incorporation of oxygen ions into the electrolyte. The kinetics of the relevant reaction steps are identified using a distribution function of time constants which is calculated directly from electrochemical impedance spectra by a newly implemented deconvolution method. In contrast to the ubiquitous non-linear least squares curve fit of equivalent circuit models, no a priori circuit choice has to be made. Moreover, the distribution function is able to resolve several physical processes within one frequency decade, much more than could be revealed using equivalent circuit models. Physical processes correspond to peaks in the distribution function. Based on the reaction model, the influence of operating conditions on the peak parameters is simulated. Relationships between peak parameters and electrochemical rate constants of the reaction model are subsequently established providing a strong tool for the identification of reaction mechanisms and loss factors.
展开▼