燃料电池低频电化学阻抗谱峰强度和膜电极金属杂离子浓度的关联关系
A Theoretical and Experimental Study onElectrochemical Impedance Spectra of Polymer Electrolyte Membrane Fuel Cellsfor Cation Content Estimation
Masao ShibataTakahisa SuzukiYu Morimoto
Abstract
This paper proposes a method to estimatethe content of metal cations in polymer electrolyte membrane fuel cells byelectrochemical impedance spectroscopy． The experimental results revealed that contamination by metal cations produces alow－frequency （ca． 0．03 Hz） peak in the imaginary part of the impedance whenhydrogen gas is supplied to both the electrodes． The peak was explained as cation transport in the membrane by atheoretical analysis． The results show an approximatelylinear relationship between the peak height and content of the metal cations，which enables the estimation of the content of metal cations in the membranefrom the peak height withoutdisassembling the cell． This method would allow revealing the contaminationprogress by metal cations in operating fuel cells．
Table I． List of materials used in thisstudy
多因素分析
Figure 1． A schematic diagram ofexperimental setup．
Figure 2． Schematic diagrams of equivalentcircuits． （a） The simplest equivalent circuit， （b） Equivalent circuit used inthis study
Figure 3． Results of the EIS measurementunder the H2－H2 condition． （a） Effect of Co2＋ content in the Nyquist plot． （b）Effect of Co2＋ content on the imaginary part of the impedance with respect tothe frequency． （c） Effect of metal cation species （Co2＋， K＋， Fe2＋， and Ce3＋）．（d） Effect of ac amplitude． （e） Effect of MEA components． RH of supplied gas is60％， coolant temperature is 80°C， ac amplitude is 5 mV， metal cation species isCo2＋， and MEA＃1 in Table I is used， unless otherwise noted． The measurementswere conducted twice for one MEA sample except for the samples shown in （e）．The solid lines represent the average of the two spectrums and the translucentarea represent the upper and lower values of the two spectrums．
Figure 4． Comparison between theoreticaland experimental results． （a） Comparison between experimental result and thethree types of theoretical model results （xc is 0．077）， （b） Effect of cationratio． （c） Effect of RH with cation ratio （xc） of 0．077． （d） Effect of celltemperature with cation ratio of 0．077． （e） Effect of PEM thickness （cationratio is 0．042 for MEA＃1 and 0．043 for MEA＃5）． （f） Effect of metal cationspecies． RH of supplied gas is 60％， coolant temperature is 80°C， metal cation species is Co2＋， and MEA＃1 inTable I is used， unless otherwise noted． Solid lines show the experimentalresults， dotted lines represent the results of the theoretical analysis， anddiamond shaped plots represent the analytical solution obtained from H＋－Mz＋model （Eq． 5）． The experimental measurements were conducted twice for one MEAsample except for the samples shown in （e）． The solid lines represent theaverage of the two spectrums and translucent area represent the upper and lower values of thetwo． The diamond shaped plots for 30％RH in （c） and Ce3＋ in （f） are not shownbecause their frequencies were calculated to be lower than 0．01 Hz．
Figure 5． Result of parameter sensitivitystudy． （a） The impedance spectrum when the parameters （Dw，ξp） are changed by 5％from their default values． （b） The shift of the imaginary part of the impedance（Zimag） at 2 Hz with respect to the rate of the change from the default values．Default values for the parameters are shown in Table II． Average RH is 60％，temperature is 80°C， cation ratio is 0．077 with Co2＋． MEA＃1 in Table I isassumed．
Dw ：Diffusioncoefficient of water （m2 s−1）
ξi：Electroosmotic dragcoefficient of ion species i
p： Proton
Figure 6． Comparison between ex situ cation contentmeasurement and proposed method． （a） Relationship between cation ratio measuredby ex situ measurement （xc，ex） and peak height of imaginary part of theimpedance caused in low frequency region （in 0．01–0．1 Hz） （Zpeak） measured bythe EIS measurement under the H2－H2 condition． （b） Comparison between the exsitu cation content measurement （xc，ex） and proposed method （xc，est）． RH ofsupplied gas is 60％， coolant temperature is 80 °C， and metal cation species isCo2＋．
Conclusions
In this paper， a non－destructive method to estimate thecontent of a metal cation in the membrane using the EIS measurement was proposed．The experimental results showed that metal cation contamination produced a lowfrequency peak in the imaginary part of the impedance （lower than 0．1 Hz withNR212） when the EIS measurement was carried out with the bias potential of 0．0V， supplying hydrogen to both the sides of the MEA． The height of the peak is linearly proportional to the content of themetal cation in the membrane whileit is independent of the catalyst loading， catalyst layer structure， and gasdiffusion layer thickness． The theoretical analysis revealed that theimpedance peak can be explained by the reaction and ionic resistance shift dueto the through－plane distribution of the metal cation． This method is expected to reveal the metal cation contaminationprogress over time in operating fuel cells．
