Printer Friendly

Mitigation of fuel cell degradation through MEA design.


Hydrogen PEM fuel cells offer a viable way of providing the electrical energy to power a vehicle. With rapid reflueling and acceptable range on a tank of fuel, they provide a complementary technology to battery-powered EVs that will ultimately offer a route to decarbonising transport.

However, the demands on an automotive fuel cell stack provide a challenging

environment for key components within the membrane electrode assembly (MEA). Frequent start-up shut-down events and freeze starts can create conditions that cause corrosion within the fuel cell. The risk of fuel starvation to individual cells within a large stack of several hundred cells is also a potential cause of corrosion which will inevitably lead to degradation and reduced stack lifetime. While there are system-level mitigation strategies that can be employed in terms of design, controls and interventions, the most efficient solution is to improve the intrinsic stability of the key cell components themselves.

Many of the mechanisms that lead to degradation can be addressed through the design of the MEA. Potential excursions that drive corrosion of cathode catalysts layers have been found, perhaps surprisingly, to be reduced by design changes in the anode catalyst. Anodes can be protected from corrosion caused by cell reversal through the incorporation of additional oxygen evolution reaction functionality built into the catalyst layer. We have demonstrated significant improvements to MEA stability and durability through such design changes. In particular, significant improvements in cell reversal tolerance of MEAs with novel anodes are described.

CITATION: Petch, M., Burton, S., Hodgkinson, A., O'Malley, R. et al., "Mitigation of Fuel Cell Degradation Through MEA Design," SAE Int. J. Passeng. Cars - Mech. Syst. 9(1):2016.


PEM fuel cells can be deployed in a range of applications from small scale consumer electronic devices requiring a few watts through to stationary power plants delivering hundreds of kilowatts. The maximum power output of a fuel cell stack will be dependent on the total surface area of the membrane electrode assemblies of which it is built up. Typical automotive fuel cell stacks are required to deliver in the order of 80 to 100 kilowatts of power and to achieve this would comprise a few hundred MEAs each with an area of a few hundred square centimetres.

The demands on an automotive fuel cell stack are usually much greater than those on a stationary system. An automotive fuel cell is expected to start rapidly from sub-zero temperatures with minimal energy input; function at a wide range of ambient temperatures and humidity and deliver load following the power demands of the vehicle which could be anything from a few watts to the maximum power of around 100 kilowatts. These requirements, together with the durability expectations of over 100,000 miles and tens of thousands of start-ups put a high demand on the components of the MEA.

There are many degradation mechanisms that can affect the life expectancy of an MEA which in-turn can be accelerated by the specific demands of an automotive system. Membrane degradation through chemical corrosion and mechanical stress limited the life expectancy of MEAs in the early years though recent improvements in membrane technology has gone a long way to addressing this issue. However, the drive to use thinner membranes in order to improve power density may require further improvements.

Further, there are a number of requirements of automotive fuel cell stacks that can significantly affect the stability of the catalytic electrodes. For example, there are specific conditions that can occur during start-up and shut-down that can cause corrosion of the cathode electrode. When a stack is shut down and isolated from the fuel supply hydrogen in the anode will slowly defuse out or be oxidised by oxygen that defuses through the membrane. Thus, over time the anode will contain some oxygen instead of hydrogen; this is then displaced by hydrogen at start up. However, during the short time that the anode contains both hydrogen and oxygen it is possible to set up an internal cell that in turn creates a high potential on the opposing cathode. Under these conditions severe corrosion of the carbon of the cathode catalyst can occur. To mitigate this degradation mechanism, there is a drive to find catalyst supports that are resistant to the corrosion mechanism. There are other ways to help mitigate against this degradation mechanism which are shown later in this paper.

Cell reversal is another mechanism that can cause severe corrosion damage to catalyst layers. This effect is caused when there is insufficient fuel available at the anode of some of the cells of a stack to support the current that is being drawn. When this occurs the potential of the fuel starved anodes will increase until they reach a level at which a reaction can occur that will provide the protons to support the current being drawn. With conventional anodes the reaction will corrode the carbon in the electrode as shown below:

C + 2[H.sub.2]O [right arrow] C[O.sub.2] + [4H.sup.+] + [4e.sup.-]

With the standard types of carbon used in anode catalyst layers the carbon corrosion reaction will begin to take place below 1.4 V and becomes fast above that potential. By using a more corrosion resistant carbon the corrosion reaction can be delayed until a higher potential but without another mitigation strategy the higher potentials are reached and the carbon still corrodes.

To reduce or prevent the carbon corrosion reaction requires the facilitation of an alternative reaction that can generate the protons that are needed at a potential at or below the potential for carbon corrosion. Water electrolysis (or oxygen evolution) is such a reaction:

2[H.sub.2]O [right arrow] [O.sub.2] + [4H.sup.+] + [4e.sup.-]

For this reaction to occur in preference to carbon corrosion the anode layer must contain an oxygen evolution reaction (OER) catalyst; the platinum catalysts used for hydrogen oxidation are not particularly active for this reaction. Ruthenium Iridium oxides are known to be active for water electrolysis [1] and such catalysts have been proven to impart cell reversal tolerance to anode electrodes. However, it is also known that ruthenium can dissolve at elevated potentials and if this occurs in the anode layer, not only will the anode lose its reversal tolerance but the ruthenium can migrate to the cathode where it can decrease the activity of the oxygen reduction catalyst [2, 3] and thus lower the performance of the fuel cell stack.

In this paper we will show the advantage of improved cell reversal tolerant anodes using new test methods that are designed to show the relative differences in stability between various anode designs.


The amount of time an anode electrode will spend under reversal over the lifetime of an automotive stack is not known, but various OEMs have made estimates which can equate to tens of hours. From such estimates various targets have been proposed and from these specific tests applied. Often these tests involve forcing a fixed current through a fuel starved anode for a specific period of time which can be up to the target of many hours. Whilst this method can demonstrate the stability of a particular design it doesn't necessarily represent what will happen in a real system; reversal events are usually of short time duration and many such short events will not have the same effect as one long reversal event.

We have developed a series of protocols that allow us to monitor the change in performance after a number of short reversal events and measure the loss in carbon due to corrosion during the test.

Tests were carried out in a single cell hardware with an active area of 217 c[m.sup.2] connected to a Hydrogenics test stand with a load bank capable of driving a current. The anode and cathode exhausts were equipped with Vaisala CARBOCAP IR-based C[O.sub.2] probes to monitor the products of carbon corrosion. Before each test, a new MEA would be conditioned by operating at 500 mA/c[m.sup.2] at 100% relative humidity (RH). MEAs all comprised JMFC proprietary membranes and cathodes were all a standard JMFC design.

6 Second Reversal Events

A polar curve is measured under specific conditions to determine beginning of life performance. The anode is then purged with nitrogen to remove all the hydrogen fuel. A current of 200 mA/c[m.sup.2] is then forced through the cell for 6 seconds; during this process the cell voltage, resistance and anode and cathode C[O.sub.2] levels are recorded. After the reversal event we continue to flow nitrogen across the anode whilst recording the C[O.sub.2] levels to ensure we capture all the corrosion products created during the reversal event. The hydrogen is then restored to the anode and a polarisation curve recorded. The reversal cycles are then repeated until a significant performance loss is determined.

5 Minute Reversal Events

The above procedure is repeated using a fresh MEA except this time each reversal event is maintained for five minutes. The procedure is repeated until significant performance loss is determined.

45 Minute Reversal Events

The above procedure was repeated using a fresh MEA except this time each reversal event lasted 45 minutes.


The MEA is initially conditioned using standard current density-hold methods and polarisation data obtained. The load is then removed such that the cell moves to OCV. The anode gas feed is switched from [H.sub.2] to synthetic air (79:21 [N.sub.2]:[O.sub.2]) and so any C[O.sub.2] subsequently monitored is solely from carbon corrosion without the background level of 'standard air'; (the cathode is also switched to synthetic air). The anode air flow is maintained for 4 minutes before returning the MEA to normal [H.sub.2]/air conditions with the immediate re-introduction of load. The cell performance is monitored at medium and a high current density points and the cycle repeated. After each set of 10 SU/SD cycles, the MEA is reconditioned under full RH, 1 A/c[m.sup.2] conditions to ensure adequate hydration of the MEA.


The 6 second cell reversal tolerance test was first carried out on a standard MEA with no specific inbuilt cell reversal tolerance. The anode catalyst comprises platinum supported on a Ketjen EC300J carbon.

Fig. 1 shows the cell voltage for the first two reversal events. At each reversal event the cell voltage falls to below -1.0 V (i.e. an anode voltage of 1.8 V assuming a cathode potential of 0.8 V at 200 mA/c[m.sup.2]) and then continues to drop further through the 6 second duration of the reversal event. The voltage profiles of subsequent reversal events match those of the first two.

Fig. 2 shows the concentration of C[O.sub.2] in the anode exhaust during and after the reversal event. Despite the time lag, it is clear that a significant quantity of C[O.sub.2] is produced during each of the reversal events. Interestingly, we can see that the amount of C[O.sub.2] produced increases over the first five reversal events; it remained at that level for the next five.

Fig. 3 shows the polar curve performance data for the MEA at beginning of life and after each of the six second reversal events. It is clear that after each reversal event performance is lost such that after 10 six second reversal events the MEA has lost over 60 mV at 1.2 A/c[m.sup.2].

When a similar MEA was subjected to the 5 minute reversal event test the anode was destroyed after less than 3 minutes so the test could not be completed. This shows how quickly an MEA can become inoperable if starved of fuel during operation.

An MEA designed to be cell reversal tolerant was put through the same tests. This MEA comprised a more corrosion resistant hydrogen oxidation catalyst support and a new oxygen evolution reaction catalyst that does not contain ruthenium.

Fig.4 shows the voltage across this MEA during two of the 6 second reversal events. Here the cell voltage falls but only as far as -0.9 V where it remains for the duration of the reversal event.

Fig. 5 shows the C[O.sub.2] level in the anode exhaust during the first five reversal events is significantly lower than that seen for the standard MEA indicating that the majority of the current is supported by the water electrolysis reaction rather than the carbon corrosion reaction. The maximum C[O.sub.2] concentration observed was 30 ppm compared to 2000 ppm for the standard unprotected MEA.

Fig. 6 shows the change in performance after each reversal event is significantly lower than for the standard MEA with a total loss of only 10 mV after 10 six second reversal events.

The cycle of reversal events was continued; Fig. 7 shows a summary of the performance decay in terms of voltage at 1.2 A/c[m.sup.2] verses number of reversal events. After 90 six second reversal events the performance had decreased by 50 mV at 1.2 A/c[m.sup.2]. Further diagnostics showed that not all loses were as a result of anode degradation; after some reconditioning 10 mV was recoverable.

This cell reversal tolerant MEA design was subjected to the 5 minute reversal event protocol.

Fig. 8 shows that the cell voltage over the first 5 minute reversal event. It can be seen that the voltage initially falls to around -0.9 V as it did in the 6 second cycles, but over the course of the 5 minutes it does drop a little further suggesting that the electrolysis reactions becomes a little less favorable.

This is supported by the anode exhaust C[O.sub.2] data shown in Fig. 9. During the first event, the C[O.sub.2] concentration continues to increase over the 5 minute event. Though it is not certain why the relative rate of the carbon corrosion increases compared to the rate of water electrolysis, one explanation could be that the electrolysis reaction becomes mass transport limited. Both reactions require water but as the water is consumed around the particles of OER catalyst then the electrolysis reaction becomes limited by the transport of water to those sites. The carbon of the hydrogen oxidation catalyst has significantly more surface area than the OER catalyst and is therefore less likely to be mass transport limited. After the first few cycles the peak level of C[O.sub.2] decreases, presumably as the most corrodible carbon is lost and the water electrolysis becomes comparatively more favorable. However, the fact that during the time scale of the event the OER can become mass transport limited means that longer reversal events are more damaging than shorter events.

The cumulative damage to the MEA from 20 five minute cycles can be seen in fig.10: after the 20 reversal events the performance at 1.2 A/c[m.sup.2] has decreased by only 50 mV. An MEA without this CRT function was inoperable after one 5 minute cycle.

It has been shown that the relative rate of the two reactions, carbon corrosion and water electrolysis, will determine the amount of damage a reversal event will cause to an anode catalyst layer. By catalysing the water electrolysis the damage to the anode layer can be significantly reduced. By making an anode layer with an OER catalyst and no corrodible carbon it should therefore be possible have even more cell reversal tolerance. An MEA comprising such an anode layer was subjected to the five minute reversal event test.

Fig. 11 shows the cumulative performance loss for this MEA over 62 five minute reversal event cycles is 32 mV at 1.2 A/c[m.sup.2].

The same MEA design was subjected a similar test with 45 minute reversal events. After 31 of these 45 minute reversal events, (just under 24 hours of reversal at 200 mA/c[m.sup.2]), the cumulative performance loss for this MEA was 81 mV at 1.2 A/c[m.sup.2].

These results show that, by the incorporation of a stable OER catalyst and decreasing the corrodible nature of the HOR catalyst, it is possible to make an MEA that is extremely cell reversal tolerant.

Start-Up Shut Down Tolerance

The reactions that cause corrosion of the anode electrode during cell reversal events are the same as those that result in the corrosion of the cathode electrode during the start-up shut-down. Therefore, it should therefore be possible to use a similar methodology to protect a cathode and this is indeed the case.

Fig. 12 shows the cumulative degradation of two different MEAs over a series of aggressive start-up shutdown cycles. MEA 1 has a standard Pt/C cathode catalyst layer; MEA 2 has the same cathode except for the addition of the OER catalyst. We can see from the chart that the addition of the OER catalyst increases the stability of the MEA.

There are other MEA design changes that can further improve the stability of an MEA to the effects of this start-up mechanism and these have been presented elsewhere.4


Cell reversal events can cause significant damage to an anode. Though the damage from events is cumulative, longer reversal events are more damaging than shorter ones. The effects of reversal can be reduced by the addition of an OER catalyst and modification to the HOR catalyst support. A similar approach can be taken to improving the durability of an MEA to the corrosion mechanisms that can occur at start-up.


[1.] Wilkinson D. P., Gascoyne J.M., Taylor J. L., Knights S. D., Campbell S. A., Ralph T. R.; US Patent 6936370

[2.] Gancs L., et al., Electrochem. and Solid State Lett., 10 (9), 2004, B150-B154

[3.] Piela et al., J. Electrochem. Soc., 151 (2), 2004, A2053-A2059

[4.] O'Malley R. et al., 226th Meeting of the Electrochemical Society (PEFC 14), Cancun, Oct 2014, Session F3: A3-2: Abstract and Presentation #

Mike Petch, S. Burton, A. Hodgkinson, R. O'Malley, and W. Turner

Johnson Matthey Fuel Cells
COPYRIGHT 2016 SAE International
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 2016 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Author:Petch, Mike; Burton, S.; Hodgkinson, A.; O'Malley, R.; Turner, W.
Publication:SAE International Journal of Passenger Cars - Mechanical Systems
Article Type:Report
Date:Apr 1, 2016
Next Article:Investigation of characteristics of gear meshing noise under three-axis driving planetary gear set.

Terms of use | Privacy policy | Copyright © 2019 Farlex, Inc. | Feedback | For webmasters