Impedance spectroscopy is a powerful technique for detailed study of the electrochemical and transport processes that take place in fuel cells and electrolysis cells, including solid oxide cells (SOCs). Meaningful analysis of impedance measurements is nontrivial, however, because a large number of modeling parameters are fit to the many processes which often overlap in the same frequency ranges. Also, commonly used equivalent circuit (EC) models only provide zero-dimensional (0-D) approximations of the processes of the two electrodes, electrolyte and gas transport. Employing improved analytical techniques to provide good guesses for the modeling parameters, like transforming the impedance data to the distribution of relaxation times (DRT), together with experimental parameter sensitivity studies, is the state-of-the-art approach to achieve good EC model fits. Here we present new impedance modeling methods which advantageously minimize the number of modeling parameters and the parameters used have direct physicochemical meaning. This is accomplished by (i) employing an improved cell model where the representative 0-D resistive-capacitive type EC elements are replaced by analytical 1-D porous electrode and 2-D gas transport models which have fewer unknown parameters for the same number of processes, (ii) use of a new model fitting algorithm, “multi-fitting”, in which multiple impedance spectra are fit simultaneously with parameters linked based on the variation of measurement conditions, (iii) constraining the parameter values during fitting to ranges of physically reasonable values. Using these methods, the number of fitting parameters for four impedance spectra measured with isolated changes to the fuel and oxidant gas compositions, has been reduced from 80 to 21-34 depending on the model. The obtained results include structural parameters like porosity and tortuosity; or if those characteristics are known, use of even fewer fitting parameters is possible. The methods have been implemented in a software package written by one of the authors, which also implements many previously used impedance analysis methods and integrates the analysis process in a modular workflow – data validation (Kramers-Kronig), clean-up, visualization (DRT and others), modeling (nonlinear least-squares fitting), and final plotting for publication.
Proceedings of 11th European Sofc and Soe Forum 2014, 2014
Main Research Area:
European fuel cell 2014 - 11th European SOFC and SOE Forum 2014