燃料电池电解质膜化学降解的物理模型：气压、相对湿度、电压对化学降解的影响、实验和模型结果的对比

Physical modeling of chemical membranedegradation in polymer electrolyte membrane fuel cells： Influence of pressure，relative humidity and cell voltage

Georg A． Futter

Arnulf Latz

Thomas Jahnke

Abstract

Chemical membrane degradation causesdeterioration of critical membrane properties such as gas separation whichfinally causes failure of polymer electrolyte membrane fuel cells （PEMFCs）． Inorder to identify the underlying physical processes， a physics－based model ofchemical membrane degradation is implemented into the novel numerical frameworkNEOPARD－X． The existing 2D PEMFC model is extended to incorporate themechanisms of hydrogen peroxide formation and reduction， a redox cycle of ironcontaminants in the ionomer phase， radical formation due to Fenton＇s chemistryand radical attack on the polymer structure． Unzipping of the polymer backboneand scission of the side chains are considered as degradation mechanism． Thedegradation model is validated against experimental data obtained inaccelerated stress tests （ASTs）． From theoretical considerations， the influenceof chemical membrane degradation on the cell performance is revealed． The influence of pressure， relativehumidity and cell voltage on the chemical degradation is rationalized． Theoperating conditions strongly influence the kinetics and spacial distributionof the membrane degradation． Degradationis found to be most pronounced at elevated pressure， high relative humidity andhigh cell voltage close to the interface of anode catalyst layer and PEM．

最后一句话不是说阳极催化层和膜界面高电位，作者说的是电池电压最高的时候，位置在阳极催化层和膜界面处。

Fig． 1． Mechanisms represented in the  chemical membrane degradation model．

这篇文章的结论限于该模型。

Fig． 2． Comparison of simulated and  experimental polarization curves： a） linear current density scale． b）  logarithmic current density scale．

It should be noted that the membrane  electrode assemblies （MEAs） used in this study are non－automotive． Therefore，  the performance is not state－of－the－art． Since chemical membrane degradation  is expected to be most severe under dry conditions and at elevated  temperature， the

polarization curves were measured at a  temperature of 368．15 K（95度）， a pressure of ×  1．5＊10＾5 Pa and 30％ and 50％ relative humidity．

Table 1 Model parameters for validation．

Table 2 Model parameters for H2O2  －formation and －reduction

Table 3

The permeability values used in this  study for different layers of the electrodes．

Table 4 Set of chemical reactions  considered in the model

Fig． 4． a） Chemical structure of Nafion and  its coarse－grained simplification． b） Side－chain scission and unzipping along  the side－chain． c）Unzipping along the backbone．

Table 5 Set of degradation reactions  considered in the model．

Table 6 Initial conditions for the  degradation model

Table 7 Operating conditions for the ASTs

Fig． 5． a） Comparison between experimental  and simulated FER during the ASTs． b） Detailed analysis of the simulated FER  for case 2．

five membrane species backbone （bb），  activated backbone （bba）， trunk （t）， activated trunk （ta） and head group （h）

湿度越高，FER越多？和我的认知不一样。

Fig． 6． a） Simulated membrane thinning． Symbols  denote experimental observations． b） Simulated decrease of OCV．

OCV的差异并不大。

Table 8 Experimental and simulated  decrease of OCV／μV h −1

这个实验和仿真差异有些大。但是实验结果湿度较高，实验中的确OCV下降较快。

在2．5bar 70％RH时 case 3，OCV下降速率从Fig 6的数据计算（0．9675－0．955）V／130hr＝96uV／hr，这个表里面的数据似乎算错了。

Fig． 7． Simulated distributions of Fe 2＋  （left） and the volumetric FER （center） and the head group concentration  （right） in the PEM for cell voltages of 0．95 （top）， 0．85（center） and 0．75 V  （bottom）． The iron ion concentration and FER are displayed for the beginning  of the AST （case 3）， the head group concentration corresponds to 100 h of  chemical degradation．

At 0．95 V， the concentration of Fe 2＋ is  highest at the anode CL／PEM interface close to the anode gas inlet which also  corresponds to the cathode gas outlet in counter－flow． The overpotential for the  reduction of Fe 3＋ to Fe 2＋ Eq． （16） is approximately −0．53 Vat the anode，  while it is 0．03 Vat the cathode． Thus， Fe 2＋ is produced at the anode， while  it is consumed at the cathode． Since the gradient of ionic potential across the  PEM is small close to OCV （−8 V m −1 ）， Fe 3＋ is able to reach the anode  side．

In the plot for the volumetric FER， the  effect of the PTFE－reinforcement layer is visible． As sources and sinks for  all species are multiplied with the ionomer volume fraction ϕ ion ， the FER  is reduced in this part of the membrane． Though the maximum iron ion  concentration is close to the anode inlet， the FER increases along the  channel since hydrogen peroxide accumulates along the flow direction． The maximum  FER is located at anode CL／PEM interface close to the anode outlet in this  case．

最大Fe2＋和最大FER的位置竟然不一样。

After 100h of chemical degradation， the  concentration of sulfonic acid groups， i．e． head groups， in the anode－ and  cathode side of the PEM are reduced by 17％ and 13％ respectively．

Fig． 8． Simulated dependency of the FER  on the cell voltage．

Conclusions

In this study， models for gas transport  through the PEM， electrochemical hydrogen peroxide formation， transport and  an electrochemical reaction of iron ions， radical formation， polymer  structure and radical attack on the polymer were combined and implemented  into a comprehensive 2D， along－the－channel PEMFC model． The simulated cell  performance is in excellent agreement with the experiment． This was achieved  by reasonable adjustment of only a few parameters， underlining the predictive  capabilities of the performance model． Non－ideal relations for the  thermodynamics of hydrogen peroxide in a two phase－system are considered，  motivating a reconsideration of the reaction kinetics hydrogen peroxide formation．  An extended set of chemical reactions， leading to hydroxyl radical formation，  together with the corresponding temperature dependencies is considered． Where  data on the temperature dependency of the reaction kinetics is missing， a sensitivity  study was performed in order to highlight the importance of further research  in this area． For the kinetics of the unzipping－ and side－ chain－scission－mechanism，  literature values were employed where available． A coupling between chemical  membrane degradation and fuel cell performance is achieved through simulation  of membrane thinning and gas cross－over． The model is validated against  experimental data for the FER．图5中的测试结果误差／波动比较大。

Experimental and simulated evolution of OCV  during the ASTs are compared． Reasonable agreement between simulated and  experimental membrane thinning is achieved． The key findings of this study  are：

1． In order to depict chemical membrane  degradation in a PEMFC， the high affinity of hydrogen peroxide to the liquid  phase needs to be taken into account．

2． Chemical  degradation is most pronounced at the anode side of the PEM and highly  voltage dependent． To explain these observations， the presence of iron  ions， their （electro－）chemical reactions and their transport has to be  considered．

3． To simulate in operando chemical degradation，  the temperature dependencies of the radical formation reactions need to be  taken into account． The influence of the degradation reactions on the steady state  concentration of hydroxyl radicals must not be neglected．

4．  The rate constants of the degradation reactions in the confined spaces inside  the membrane may differ significantly from the values obtained from ex situ  measurements of model compounds in aqueous solution．

5． Chemical membrane degradation causes  thinning of the membrane which promotes hydrogen cross－over from the anode to  the cathode． This causes an increase of the internal short circuit current， a  reduction of OCV and therefore cell performance．

6． OCV reduction during the ASTs cannot  be explained by increasing hydrogen crossover due to membrane thinning alone．  Further investigations on the phenomenon are required．

7． Degradation increases with the operating  pressure due to an increase in oxygen cross－over and subsequent hydrogen  peroxide formation．

8．  Humidification increases chemical membrane degradation since it promotes gas  cross－over and reduces potential gradients by improving the ionic  conductivity．

目前不能理解作者的解释。

9． At OCV， chemical membrane degradation  is most pronounced as ionic potential gradients are small and iron ions are  able to move freely． With decreasing  cell voltage， iron ions get dragged to the cathode side and chemical membrane  degradation ceases．

仿真少看（看了也看不懂），实验多读算是读完这篇文章的感受。有能力的同仁多读读。