Differentiation and decomposition of solid oxide fuel cell electrochemical impedance spectra

Abstract

Benefitting from high-temperature operating conditions, solid oxide fuel cells (SOFCs) exhibit high electricity efficiency and can be coupled with conventional engines such as gas turbines, to achieve cascaded energy utilization. However, high temperature inevitably accelerates material deterioration, and simultaneously complicates the on-line diagnosis of SOFCs. Electrochemical impedance spectroscopy (EIS) is a mature on-line testing technology that is been widely used in SOFC research. Using EIS analyses, researchers can clearly determine the ohmic resistance of pure ion/electron conduction and the magnitude of the polarization impedance of electrochemical processes or diffusion. However, the lack of decomposition of polarization impedance constitutes a limitation for the deeper understanding of SOFC operation. To better utilize information-rich EIS, this work used a typical Ni-Y2O3 stabilized zirconia (Ni-YSZ) anode-supported SOFC and designed a complete set of impedance tests by modifying the test temperature, anode operating atmosphere, and cathode operating atmosphere. In addition, this study combined two advanced impedance spectrum analysis methods: the analysis of differences in impedance spectra (ADIS) and distribution of relaxation time (DRT) methods to allow for a deeper comprehension of the impedance spectrum and obtain the characteristic frequency of each process in the SOFC. Based on the characteristic frequency obtained by the spectral decomposition, this study analyzed and explained the physical or (electro-) chemical origins of the impedance in each frequency band. In addition, the equivalent circuit method (ECM) was adopted to fit the SOFC impedance spectra by 6 (RQ) parallel circuits to summarize the impedance distribution. The ADIS method was more convenient than the DRT method and could be used to initially determine the range of characteristic frequencies. However, it failed to provide a clear decomposition for the high-frequency region of the impedance spectrum. The DRT method can yield a good decomposition of a single impedance spectrum, and the characteristic frequency of Ni-YSZ SOFC was separated into ~2, ~20, ~30, 1 × 103–1.5 × 103, 2 × 103–4 × 103, and ~4 × 104 Hz regions. The highest frequency (~4 × 104) region was dominated by charge transfer impedance at the YSZ/GDC (gadolinia doped ceria) or LSCF/GDC interface. Peaks at 2 × 103–4 × 103 and 1 × 103–1.5 × 103 Hz correspond to the H2 electrochemical reaction at the Ni/YSZ/H2 boundary and O2 reduction at the LSCF/O2 boundary, respectively. The impedance at lower frequencies can be explained by diffusion polarization impedance. In this study, ~20 Hz corresponds to the cathode diffusion impedance, and ~30 Hz corresponds to the anode diffusion resistance. At a frequency of ~2 Hz, the impedance magnitude is influenced by both cathode and anode diffusion. The results of this study combined three methods (ADIS, DRT, and ECM) for SOFC impedance investigations, and lays a foundation for future studies regarding SOFC performance degradation and degradation characteristics based on DRT analysis.