No abstract

The current-interrupt technique and electrochemical impedance spectroscopy were employed in order to study the behavior of a polymer electrolyte fuel cell ͑PEFC͒ cathode containing 30 wt % Nafion and 70 wt % Pt/C. The steady-state polarization curves were also recorded. The experimental results were analyzed with help of the mathematical models developed in Part I of this paper. The effect of a varying oxygen pressure and humidity on the dynamic response of the cathode was investigated. The double-layer capacitance, Tafel slope, oxygen solubility, a group containing the effective O 2 diffusion coefficient and agglomerate size, and finally, the effective proton conductivity in the cathode were obtained. The parameter values were reasonable and attest the robustness of the agglomerate model for describing the PEFC cathode. At low humidity, a second, low-frequency loop was observed that was attributed to the membrane behavior.

Electrochemical impedance spectroscopy ͑EIS͒ and steady-state models have been developed for the porous hydrogen electrode with water concentration dependence and water transport in a polymer electrolyte fuel cell membrane. Because the hydrogen electrode performance is influenced by its water content, the hydrogen electrode model was coupled to the membrane model. The EIS model for the hydrogen electrode gave three to four loops in the complex plane plots. The high-frequency semicircle was attributed to the Volmer reaction and the medium-frequency semicircle to the hydrogen adsorption. The additional low-frequency loops were connected to changes in the hydrogen electrode performance with water concentration, due to changes in kinetics or proton conductivity. Those loops appear in a frequency range depending on the water transport in the membrane, changing with D/L m 2 , where D is the water diffusivity and L m is the membrane thickness. Modeling of the membrane alone showed that the membrane gives rise to a loop in EIS. The difference between the high-and low-frequency intercepts of the loop is idR/di, where the high-frequency intercept is equal to the membrane resistance. The loop appears in the same frequency range as the hydrogen electrode low-frequency loops and thus overlaps. A749) unless CC License in place (see abstract). ecsdl.org/site/terms_use address. Redistribution subject to ECS terms of use (see 155.33.16.124 Downloaded on 2014-11-05 to IP A755 Journal of The Electrochemical Society, 153 ͑4͒ A749-A758 ͑2006͒ A755 ) unless CC License in place (see abstract). ecsdl.org/site/terms_use address. Redistribution subject to ECS terms of use (see 155.33.16.124 Downloaded on 2014-11-05 to IP A756 Journal of The Electrochemical Society, 153 ͑4͒ A749-A758 ͑2006͒ A756 ) unless CC License in place (see abstract). ecsdl.org/site/terms_use address. Redistribution subject to ECS terms of use (see 155.33.16.124 Downloaded on 2014-11-05 to IP Figure 10. Modeled impedance spectra at 300 mA/cm 2 for a full cell at different membrane thicknesses, corresponding to ͑a͒ one, ͑b͒ two, and ͑c͒ three Nafion 1035 membranes, respectively. Insert: Frequency ͑ f͒ and angular velocity ͑͒ at the maximum imaginary part for different membrane thicknesses ͑constant diffusion coefficient D 0 = 7.2 ϫ 10 −10 m 2 /s͒ ͑d͒ and diffusion coefficients ͑constant membrane thickness, two Nafion 1035 mem-branes͒ ͑e͒.

Influence of water on membrane and anode performance was studied with steady-state and electrochemical impedance spectroscopy ͑EIS͒ measurements using a symmetrical cell with hydrogen on both sides. Both full-cell and half-cell measurements were performed. To obtain half-cell data a new reference electrode approach was demonstrated based on porous references in a four-electrode setup. A varying membrane resistance with current density was obtained using current interrupt and EIS measurements. The EIS measurements showed two semicircles at 10 4 Hz and 0.01-0.1 Hz, respectively. The first corresponds to hydrogen adsorption and the second to the water dependence of the electrode performance and membrane resistance. The low-frequency semicircle appears in a frequency range depending on the membrane thickness. The loop corresponding to the discharge of the double-layer capacitance through the Volmer reaction appears at frequencies too high to be experimentally measurable. The experimental data were in good agreement with the model developed in Part I of this paper. The model was also successfully fitted to experimental full cell data at different current densities and membrane thicknesses. The experiments confirmed that the low-frequency semicircle is attributed to the water dependence of both anode and membrane performance.

Electrochemical impedance spectroscopy (EIS) and steady-state models have been developed to investigate the influence of water transport on the membrane and electrode performance, with focus on the low-frequency impedance. Models for the membrane, hydrogen anode and oxygen cathode were connected in order to take the influence of water concentration on proton conductivity and hydrogen kinetics into account. At low frequencies, below 1 Hz, a pseudo-inductive loop was predicted, resulting from the overlap of the responses from anode and membrane. The anode response could be coupled to changes in the kinetics and polymer conductivity in the active layer, and the membrane response to changes in conductivity with changing water profile. The low frequency capacitive part was attributed to drying of the anode side of the membrane, while the inductive part was attributed to the rehydration of the membrane with water produced at the cathode. The loop appeared at a frequency proportional to 1/L 2 , where L is the membrane thickness. The model was successfully fitted to experimental data at different membrane thicknesses, relative humidities and current densities. The modeled data follow the same trends as experimental data, giving an increase in impedance at dry conditions and with thicker membranes.Losses in polymer electrolyte fuel cells (PEFCs) have been widely investigated. Most studies have focused on membrane resistance, as well as the cathode performance, while losses from the anode are usually assumed to be small or negligible. Common electrochemical techniques utilized to study the performance losses are steady-state polarization curves, current interrupt and electrochemical impedance spectroscopy (EIS). EIS is a powerful tool for elucidation of ratelimiting processes in PEFCs, since it can separate processes with different rate constants. However, it is sometimes challenging to distinguish between different processes taking place. When the cathode is studied using EIS, the impedance of the whole cell is usually measured, assuming that the anode impedance can be neglected and that the membrane is acting as a pure resistance.In part I of this paper, 1 low humidity effects in polymer electrolyte fuel cells (PEFCs) are studied using electrochemical impedance spectroscopy (EIS). The effect of membrane thickness, relative humidity and current density was studied experimentally with focus on a low frequency loop. The size of the loop increased with increasing membrane thickness and decreasing humidity. The response also shifted to a lower frequency with a thicker membrane. A similar loop was studied in an earlier study 2,3 in a cell with hydrogen on both sides. A full cell model was developed, including water transport in the membrane, and fitted to experimental data. It was seen that the change in water content in the membrane and anode resulted in a low frequency loop due to changes in the membrane conductivity and anode kinetics. To describe a fuel cell under real conditions, the model must also include an oxygen redu...