Electron Microscope Analysis of Low Temperature Damage in PEM Fuel Cells

By Eric L. Thompson

Department of Chemical Engineering, University of Rochester, Rochester, NY




PEM fuel cells electrochemically combine hydrogen and oxygen to produce electricity and water. Because of their environmental friendliness and use of alternative fuel, they are being investigated by the automotive industry as a replacement for the gasoline engine. In order to meet automotive targets, PEM fuel cells must be able to start, operate, and survive reliably from sub-freezing temperatures. Considering the facts that water is generated within the fuel cell, and is often provided as external humidification, these pose a significant concern for low temperature operation. In this study, several low temperature failure modes are investigated in PEM fuel cells that underwent multiple freeze start and operation events from -20 C. Samples were removed from the fuel cells and analyzed with electron microscopy techniques.



PEM (polymer exchange membrane) fuel cells are electrochemical devices in which hydrogen and oxygen (from air) are combined to produce water. Reactant gases enter the fuel cell through a set of flow fields. Often, the gases are humidified before delivery to the fuel cell. The purpose of the flow field is to evenly distribute the reactant gases as well as carry away the unused gases and product water. Once gases enter the fuel cell, they diffuse through a porous “backing layer”, commonly known as diffusion media. Gas can freely diffuse through this backing layer toward the electrode, where reaction occurs. Product water can also diffuse or flow away from the electrode, particularly on the cathode side. Often a micro-porous layer (known as an MPL) is included in the fuel cell structure, between each backing layer and electrode. The purpose of the MPL is to provide a transition between the backing layer and electrode and assist in product water transport. The electrodes are also a micro-porous structure of carbon spheres coated with a finely dispersed catalyst, which speeds the kinetics of the electrochemical reactions. At the anode, hydrogen is oxidized:


2H2 ↔ 4H+ + 4e-


While at the cathode oxygen is reduced to water according to the reaction:


O2 + 4H+ + 4e- ↔ 2H20


The hydrogen (anode) and oxygen (cathode) remain separated by a thin membrane made of a unique polymer material. This polymer (known as an ionomer) has the ability to conduct protons but insulates electrons, and gives the fuel cell the ability to function. On the anode side, hydrogen dissociates into protons (H+) and electrons (e-). The protons travel through the membrane to the cathode, but the electrons cannot. They are forced through an external circuit, where they can do useful work, such as powering an electric drive motor. The humidification of the membrane helps determine its ability to conduct protons, and is a primary reason for pre-humidification of the reactants. Often, the electrodes are permanently applied or hot-pressed to the membrane to form what is known as a membrane electrode assembly (MEA). The following sketch was taken from Fuel Cells-Green Power; an informational booklet provided by Los Alamos National Labs which is available online at www.education.lanl.gov/resources/fuelcells. This sketch nicely illustrates the structures and process occurring within the fuel cell.


Figure 1. Sketch of PEM fuel cell showing internal components


One can imagine that due to the formation of product water within the fuel cell, the presence of water-filled pores or flow field channels could lead to problems when retained water is allowed to freeze and undergo 10% volume expansion. Often a purge is performed at shutdown, prior to freezing, to remove liquid water (U.S Patents #6,479,177 B1, #5,798,186). This purge is ineffective at removing all water from within the cell, so some ice formation is inevitable. One aspect of this study is to determine the effect that freezing of residual water has on the structures within the fuel cell. The examined mechanisms of hypothesized damage are as follows:

·        Water is present, in either liquid or vapor phase, in all operating PEM fuel cells. Following operation, if the fuel cell is exposed to sub-freezing temperatures, this water may freeze and damage the internal structures. One way to distinguish between different flow field designs is their ability to remove and minimize the retention of bulk water and large droplets in flow field channels during operation. For the purpose of this study, a flow field design that is effective at removing bulk water and droplets is referred to as a water-clearing fuel cell. On the other hand, some flow field designs have been shown to trap or hold bulk liquid water and droplets in specific locations of the flow field. This type of flow field design is referred to as a water-trapping fuel cell. In the first part of this study, a comparison of damage to the backing layer and electrode structures from two different fuel cells; one with a water-trapping flow field, the other with water-clearing flow field is made. Ice damage to the backing layer and electrode is expected in the water-trapping fuel cell. Each fuel cell underwent multiple freeze start and sub-freezing operation events, in which product water could freeze and cause damage to the internal components. Microscope analysis includes low magnification images to illustrate large scale damage as well as high magnification images to determine if damage is occurring to fine structures that would inhibit the electrochemical process or diffusion of reactants.

·        Since the outlet of the cathode side is the exit for all product water generated over the entire electrode, it is presumably the wettest region within the fuel cell, and potentially is the most probable location for ice damage to occur. This is true regardless of whether the flow field design is water-trapping or water-clearing. Based on the knowledge that a water-clearing fuel cell does not retain much bulk water in the flow field, a question still remains whether the outlet region would undergo any damage after freezing, especially to the fine pore structures of the electrodes and MPL. Presumably these fine pore structures could contain significant water near the cathode out, even if the adjacent flow field does not have bulk water present. Samples of the backing layer, MPL and electrode structures were collected from the inlet and outlet regions of a water-clearing PEM fuel cell and compared to new samples of each material. Both high and low magnification SEM images are collected. Since less large-scale damage to backing layers is expected in water-clearing fuel cells, the main goal is to determine if any small-scale structural damage in the electrode and MPL is present.

·        Finally, a common technique for assisting cold starting of PEM fuel cells is to provide a mixture of reactant gases to a single electrode (International Patent WO 00/54356). The gas mixture should be below the lower explosion limit of 4% hydrogen in air. For example, 2.5% hydrogen is mixed with the air and delivered to the cathode electrode. This mixture reacts exothermically on the electrode catalyst, and provides additional heat to warm the cell up. Catalyst sintering (Oswalt ripening) is known to occur in dispersed catalysts at elevated temperatures. Since not much is known about the local electrode temperatures during this process, an investigation of the catalyst particle sizes was conducted with a TEM. Electrode sections were taken from the leading edge of a fuel cell electrode and compared to a new sample to investigate changes in particle size.





Observed Damage to Backing Layer and Membrane Electrode Assembly (MEA) in Water-Trapping Fuel Cell Compared to Water-Clearing Fuel Cell




Figure 2. New backing layer sample compared to damaged sample from bulk water location of a water-trapping fuel cell.





Figure 3. New backing layer sample compared to sample taken near cathode outlet region of water-clearing fuel cell.


Figures 2 and 3 illustrate the difference in backing-layer damage observed in two fuel cell types. Images of new samples are provided for reference and to illustrate the overall magnitude of the damage. It is apparent from Figure 2 that locations which trap or hold bulk water during operation undergo massive damage when exposed to freezing temperatures. The water-clearing design does not appear to suffer from this problem, as only minor damage was found near the cathode outlet. This was the worst example of backing-layer damage found in the water-clearing design, which otherwise appeared undamaged.




Figure 4. SEM (left) and Light Micrographs (right) showing damaged MEA samples having loss of electrode from freezing of water-filled channel of water-trapping fuel cell.




Figure 5. Damaged MEA sample showing broken backing-layer fiber lifting a section of electrode from membrane near water-filled channel of water-trapping fuel cell.





Figure 6. New MEA (electrode) sample compared to a sample taken from electrode near cathode inlet of water-clearing fuel cell at low magnification. Note that some of micro-porous layer remains adhered to electrode in post frozen sample.





Figure 7. High magnification images of new MEA (electrode) sample compared to a samples taken from damaged electrode of the water-trapping fuel cell and cathode inlet of a water-clearing fuel cell.



Figures 4 and 5 illustrate the massive damage observed at an MEA taken from a location of bulk water-filled channel in a water-trapping fuel cell design. Freezing of water has de-laminated the electrode off the membrane over the entire channel region. Also, evidence of the mechanism of how this occurs is given by the broken backing-layer fiber lifting the electrode structure away from the membrane. Figure 6 compares a new electrode to one removed from a water-clearing fuel cell. Note that although there appears to be more electrode cracking in the used sample, none of the electrode de-lamination has occurred as observed in the water-trapping design. Finally, Figure 7 examines damage at a smaller scale. This figure compares a new sample to those obtained from the post frozen fuel cells at high magnification. From this figure, it does not appear that any damage is occurring to the electrode structures on a very small scale.




New Backing-Layer, Micro-porous Layer (MPL), and MEA Electrodes Compared to Samples Obtained From Inlet and Outlet Regions of a Water-Clearing Fuel Cell



Figure 8. New backing-layer sample compared to samples taken from the cathode inlet and cathode outlet regions of a water-clearing fuel cell.





Figure 9. New micro-porous layer (MPL) sample compared to samples taken from the cathode inlet and cathode outlet regions of a water-clearing fuel cell at low magnification.





Figure 10. New micro-porous layer (MPL) sample compared to samples taken from the cathode inlet and cathode outlet regions of a water-clearing fuel cell at high magnification.





Figure 11. New MEA electrode sample compared to samples taken from the cathode inlet and cathode outlet regions of a water-clearing fuel cell at high magnification.


Figure 8 shows that some minor damage to the backing layer fibers and binder is occurring near the wet, cathode out region of a water-clearing fuel cell. The inlet region appears to look undamaged, as compared to the new sample. Figures 9 and 10 show the micro-porous layer. In the low magnification image, cracking is present in the new sample, as well as those obtained from the inlet and outlet regions of a water-clearing fuel cell. In the high magnification image, Figure 10, no major damage is observed in the micro structure of the MPL, even in the wet outlet region. Similarly, at high magnification, the electrode does not appear to show signs of damage following the freeze events, as shown in Figure 11.



New Electrode Catalyst Compared to Samples Obtained From Inlet and Outlet Regions of a Water-Clearing Fuel Cell



Figure 12. New electrode microtome section compared to sample taken from the cathode inlet region of a water-clearing fuel cell that underwent catalytic heating. (500,000x)





Figure 13. New electrode microtome section compared to sample taken from the cathode inlet region of a water-clearing fuel cell that underwent catalytic heating. (300,000x)



Figures 12 and 13 show some high magnification TEM images of 100 nm microtome sections of the electrode. The dark spots correspond to the dispersed catalyst. It is apparent from these figures that several large particles appear in the inlet region samples; however the vast majority of the catalyst appears to be dispersed with no significant ripening.




         Using light microscopy, sputter coating, SE, In-lens SE, microtomy, and TEM, internal fuel cells structures were examined.

         Massive damage to backing layer and electrode was observed in water-trapping fuel cell at flow-field channel locations that accumulate bulk liquid water. Freezing of this water damages internal fuel cell components.

       Crushed or broken backing layer fibers

       Electrode cracking and de-lamination from membrane

         Minor damage to backing layer fibers in wet outlet region of water-clearing fuel cell was observed.

         Large scale cracking was observed on the MEA electrode of water-clearing fuel cell following repeated cold start events.

       No damage observed in fine pore structure of electrode or micro-porous layers in even the wettest region of water-clearing fuel cell.

         No excessive sintering or ripening of catalyst observed in fuel cell that underwent catalytic heating to assist cold starting.




I would like to thank Brian McIntyre for his assistance throughout this project. I also wish to thank General Motors Fuel Cell Activities for providing me the opportunity to perform this study.

Please enter any comments, criticisms, questions, etc. below.

Your name:

Email address: