700次NEDC工况下石墨双极板燃料电池和激光切割预涂层金属双极板燃料电池耐久性对比：金属污染物的影响

Trace－metal contamination in protonexchange membrane fuel cells caused by laser－cutting stains on carbon－coatedmetallic bipolar plates

Timon Novalin

Björn Eriksson

Sebastian Proch

Ulf Bexell

Claire Moffatt

Jörgen Westlinder

Carina Lagergren

Göran Lindbergh

Rakel Wreland Lindström

Abstract

Trace－metal contamination poses a threat toperformance and stability of proton exchange membrane fuel cells （PEMFCs）． Inthis study the source of origin and degree of metal dissolution fromcarbon－coated 316L bipolar plates （BPPs） are evaluated after a long－term PEMFCtest run under conditions resembling a real－life automotive application．Despite intact carbon－coating， metaldissolution occurs from uncoated oxycarbide stains on the plates’ surface．Which correlates with post－mortem detection of chromium， iron and nickel in themembrane electrode assembly． The rate ofcell voltage decrease throughout the high current operations is found to betwice as high in the presence of metal ions． Metal dissolution can be correlated with transients in cell voltage duringdynamic current load cycling as a result of temporary global fuel starvation．The observed difference in metaldissolution on the anode and cathode BPP indicates weak galvanic couplingbetween the bipolar plate（s） and the electrode layer（s）．

Three possible origins of trace－metalcontamination have been considered in this study as follows．

1Manufacture of membrane electrode assembly（MEA） and gas diffusion layer （GDL）

2Humidification water and balance－of－plant（BoP）

3Carbon－coated stainless－steel bipolarplates （BPPs）

Table 1 e Elemental composition of 316L  stainless steel， the substrate for the bipolar plates used in this study． Given  are the maximum allowed values while Fe is obtained from the balance．  Elements in low concentration， e．g．， Si， C， N， etc． have been neglected．

Fig． 1 e Experimental configurations of  the single－cell setup． A． Configuration for experiments with pure graphite  current collector （GCC）． B． Configuration for experiments with carbon－coated  stainless－steel bipolar plates （BPPs）．

Fig． 2 e Manufacturing route to the  bipolar plate with spiral flow field used in the single－cell setup to  investigate the trace－metal contamination from carbon－coated stainless－steel  bipolar plates with a gas outlet hole  in the middle． Such a plate is used at both anode and cathode side （Fig．  1 B）． 316L strip in 0．1 mm thickness is coated with graphite－like carbon in a  high－throughput physical vapor deposition （PVD） coating line （pre－coating）．  Laser cutting from the back side （5） leads to splashing and metal oxycarbide  stains on top of the carbon coating as found by SEM／EDS and AES．

这里使用激光切割的位置是气体的出口，不是气体的入口。即使是气体的入口，也只有加湿的气体通过，对耐久性结果有影响也令人不解。

Table 2 e Trace－metal content of relevant  ions determined by ICP－SFMS in the commercial， as－receivedMEA and GDL at BoL （entry  1 and 2） and after load cycling with two GCCs （entry 3） or with two BPPs  （entry 4） at EoL （see Fig． 1）． Note that for the EoL samples MEA and GDL were  inseparable and， hence， a combined value has been obtained．

耐久性后Fe含量90ppm

Fig． 3 e Load （current density） cycling  according to the NEDC． The green， yellow， and red dots mark positions of low，  medium， and high load A． Current density profile （right y－axis） and the  corresponding cell voltages after 1， 200， 400， 600， and 700 cycles，  respectively， for test configuration with GCCs． B． Cell voltage degradation  in the marked load regions with progressive cycling for test configuration  with GCCs． C． Current density profile （right y－axis） and the corresponding  cell voltages after 1， 200， 400， 600， and 700 cycles， respectively， for test  configuration with BPPs． D． Cell voltage degradation in the marked load  regions with progressive cycling for test configuration with BPPs．

这里提到NEDC，EU harmonized test protocols for PEMFC MEA testing in single cell configuration  for automotive applications． Petten （The Netherlands）： EU； 2015．

网上也有一个版本：European Commission， Joint Research Centre， De Marco， G．， Malkow，  T．， Tsotridis， G．， et al．， EU harmonised test protocols for PEMFC MEA testing  in single cell configuration for automotive applications， Publications  Office， 2016， https：／／data．europa．eu／doi／10．2790／54653

Fuel Cells Bulletin Volume 2016， Issue 4，  April 2016， Pages 9－10 https：／／doi．org／10．1016／S1464－2859（16）30088－8

还对这件事情进行了专门报道。

对比图a和图c，图c的电压存在超调，不知道是设备不同还是什么原因。

作者的解释：

It is suspected that the probes to  measure the interfacial contact resistance between the anode current collector／bipolar  plate and the anode GDL （blue in Fig． 1 A and B） led to hydrogen leakage． This  ultimately caused the slight transients observed in Fig．3 A（GCCs） and the more  heavy ones in Fig． 3C and Fig． S9 （BPPs） due to global hydrogen starvation and，  therefore， the dissolution of laser cutting stains．

每个工况1180s，700次循环230hr左右。

工况看的眼熟是吧，文献10kW大面积金属双极板燃料电池短堆600小时动态工况衰减分析［失效分析其八］中使用的也是NEDC工况，是1400s版本的。

Fig． 4 e Schematic surface models for  stained areas and their immediate carbon－coated surroundings． BPP as－cut  before single－cell tests （Fig． 2）； BPP from anode and cathode side after load  cycling （NEDC）， see Figs． 1 B and ． 3C and D． These surface models have been  derived by detailed investigation by SEM／EDS and AES， see supporting  information． It is apparent that as－cut and cathode plate are rather similar  while significant dissolution takes  place from the anode BPP．

The potential of the BPPs is decoupled （weak  galvanic coupling） from the catalyst layers （electrodes） and remains at OCP  even during transient behavior in the electrodes．

Fig． 5 e Temporal evolution of the ICR of  the anode and cathode side for test configuration with BPP （Fig． 1 A）

ICR能稳定或者增加比较好理解，ICR下降是第一次看到。和电池中的材料变化有关。

The observed initial decrease is  explained by swelling of the MEA resulting in a higher clamping pressure．

Conclusions

The aim of the study was to investigate  the contamination potential of carbon－coated stainless－steel bipolar plates  in operando fuel cells． This could not be done due to metal oxycarbide stains  that formed on top of the carbon coating during laser cutting． ICP－SFMS and  SEM／EDS investigations showed that metal ion dissolution takes place from  those stains at the anode bipolar plate during load cycling． High contamination of Fe （90 ppm） and Ni  （20 ppm） is found in the GDL／MEA at the end of life． This is attributed  to global fuel starvation due to rapid load cycling and changes in anode bipolar  plate potential as described by Hinds et al． The transients are speculated to  be originated in the insertion of metal probes to measure the anode contact  resistance and concurrent hydrogen leakage． We regard those stains as  valuable probes into the quality of the fuel cell operation， i．e．， presence／absence  of fuel starvation （transients）．

Investigation of the contact resistance  of the bipolar plates during load cycling showed low and stable values and  the absence of degradation exemplifying the benefits of such a coating  especially in the light of decoupled bipolar plates as described by Hinds et  al．

As next steps we will debug our  experimental setup and investigate the “stained” bipolar plates in the  absence of global fuel starvation． Moreover， the stains can easily be avoided by flipping the plates before laser  cutting and drill from the front instead of the back side． Another  interesting experiment will be the investigation of uncoated stainless－steel  plates and the evolution of contact resistance during load cycling． The AES  depth profiling method can then be used post mortem to investigate passive  （oxide） film formation on the steels．

这篇文章的有些证据被放在支撑数据里面。

Figure S1．SEM images recorded in secondaryelectron mode in increasing magnification from the center hole on the front（active） side of the “as－cut” bipolar plate used in the setup in Figure 1B（main text）． The image in the lower right shows “stains” on top of the carboncoating in blue and a rather unaffected area （carbon coated） in orange．

FigureS5．SEM images of the anodeside of the bipolar plate after load cycling． The blue square in A ismagnified in B to visualize the morphology of the stain． Dots indicatethe areas where the Auger survey spectra in C are acquired．

FigureS7．SEM images of the cathodeside of the bipolar plate after load cycling． The blue square in A ismagnified in B to visualize the morphology of the stain． Dots indicatethe areas where the Auger survey spectra in C are acquired．

这篇文章的证据缺少一环：

如果这些SEM下可见的金属杂质可以通过清洗等方式去除，是否可以让金属在膜电极中的沉积量大幅下降，不会影响耐久性？

如果不使用激光方法，而是用模具的方法切割孔，是否可以让金属在膜电极中的沉积量大幅下降，不会影响耐久性？毕竟预涂层容易让人诟病的是会有无涂层覆盖的切口。

如果不使用预涂层，而采用后涂层方案重新做该实验（作者中有sandvik的技术人员），可以让整个过程更加清晰，更有指导意义。