Deconvolution of Gas Diffusion Polarization in Ni/Gadolinium-Doped Ceria Fuel Electrodes

Abstract

Electrochemical impedance spectroscopy and the subsequent impedance data analysis by the distribution of relaxation times (DRT) represent a powerful method to unfold electrochemical processes in solid oxide cells. This approach has been successfully demonstrated for electrolyte [1, 2] as well as anode-supported [3] solid oxide cells exhibiting a nickel / yttria-stabilized zirconia (Ni/YSZ) fuel electrode. Applying this methodology, all physicochemical loss processes occurring in the cells could be identified by their relaxation frequencies and subsequently quantified by means of CNLS-fitting.However, in the case of cells exhibiting a Ni/Gadolinium-doped ceria (Ni/CGO) fuel electrode, an unambiguous process assignment in the DRT has not been feasible up to now, since the charge transfer and gas diffusion overlap in the spectrum [4, 5]. The latter is caused by the oxygen non-stoichiometry of CGO, which results in a large chemical capacity, shifting the relaxation frequency of the charge transfer process towards lower relaxation frequencies. The resulting overlap of charge transfer and gas diffusion process in the spectrum impedes a direct deconvolution of these two processes [6].Facing this challenge, we here present a solution of isolating and quantifying the gas diffusion process at the fuel electrode with the help of ternary fuel mixtures exhibiting different gas diffusion coefficients. The gas diffusion coefficients of hydrogen and steam were altered by changing the inert gas component between nitrogen and helium respectively. As the two gas mixtures exhibit identical reactant and reaction product concentrations, the charge transfer polarization is not affected. The differences in polarization resistance R pol, which become visible in figure 1, are solely caused by the differences in gas diffusion polarization [2].In this contribution the detailed procedure to quantify the gas diffusion polarization and an effective gas transport parameter predicting the gas diffusion resistance for arbitrary fuel mixtures are introduced. Furthermore, the applicability of this approach for the parameterization of a zero-dimensional cell model [3] for a solid oxide cell with Ni/CGO fuel electrode will be presented.