Optimization of the Perturbation Amplitude for Impedance Measurements in a Commercial PEM Fuel Cell Using Total Harmonic Distortion

Abstract

Abstract:One of the most important measurement parameters in electrochemical impedance spectroscopy (EIS) is the perturbation amplitude. The optimum perturbation amplitude value corresponds with a balance between the signal-to-noise ratio improvement and the reduction of the harmonic generation due to nonlinear effects: Therefore, the optimum perturbation amplitude is the maximum amplitude that ensures a linear response of the system. Two strategies were considered in this work: a constant amplitude strategy and a frequency dependent amplitude strategy. On the one hand, for the constant amplitude strategy, the optimum perturbation amplitude for EIS measurements of an individual cell of a commercial PEM fuel cell stack was determined. In order to fulfill this aim, the impedance spectra (at different DC currents) of the individual cell of the commercial PEM fuel cell stack were measured using different perturbation amplitudes. The total harmonic distortion of the recorded voltage signal was determined in each case, applying a FFT based method. The optimum amplitude for each DC current corresponds to the amplitude that minimizes the critical total harmonic distortion value. On the other hand, for the frequency dependent amplitude strategy, the optimum amplitude at each frequency for EIS measurements of an individual cell of a commercial PEM fuel cell stack was determined.

Harmonic analysis based method for linearity assessment and noise quantification in electrochemical impedance spectroscopy measurements: Theoretical formulation and experimental validation for Tafelian systems

“…When the potentiostat was used to perform measurements in a different system (i.e. PEM fuel cell) exactly the same ℘ curves were obtained [56]: this indicates that these ℘ curves are characteristic of the measurement equipment, which is responsible for generating the perturbation, and applying it to the system. These curves may be affected by distortions in the grid signal that powers the measurement equipment, and by electric couplings between the measurement equipment and other devices connected to the same electrical grid.…”

“…Consequently, the quality of EIS measurements is given by the trade-off between the linearity condition fulfilment and the maximization of the signal-to-noise ratio [53]: the perturbation amplitude has to be low enough in order to guarantee the fulfilment of the linearity condition [54]; and it has to be high enough in order to achieve an acceptable signalto-noise ratio [55]. The optimum perturbation amplitude of a given system corresponds with the perturbation amplitude that maximizes the signal-to-noise ratio without violating the linearity condition [56].…”

CuO improved (Sn,Sb)O2 ceramic anodes for electrochemical advanced oxidation processes

“…In this research, a perturbation amplitude of 20 mV was used for all the EIS measurements. This amplitude was selected using the selection methodology presented in previous works …”

Improvement of the Electrochemical Behavior of (Sb, Sn, Cu)O Ceramic Electrodes as Electrochemical Advanced Oxidation Anodes

“…[44] For this reason, validation is a fundamental part of the preliminary analysis of experimental EIS spectra. [52,53] The uncompensated resistance is given by the high frequency intersect of the EIS spectrum with the real axis: [54] in this work, the uncompensated resistance for each overpotential was determined from the high frequency intersect of the experimentally measured impedance spectrum at that overpotential with the real axis. [47,48] In this work, the EIS measurements were performed at 100 frequencies logarithmically spaced between 10 kHz and 10 mHz.…”

“…This amplitude was selected using the selection methodology presented in previous works. [52,53] The uncompensated resistance is given by the high frequency intersect of the EIS spectrum with the real axis: [54] in this work, the uncompensated resistance for each overpotential was determined from the high frequency intersect of the experimentally measured impedance spectrum at that overpotential with the real axis. Both, the CVs and the pseudo-steady-state polarization curve, were IÀ R corrected using the procedure described by Hrbac and co-…”