The liquid water permeability of the anode and cathode gas diffusion layers are tailored for each cell according to its position within the fuel cell stack, so as to promote movement of water toward water transport plates and away from catalysts, especially cathode catalysts, taking into account that water moves toward the cooler part of the stack during the cooling (and possibly freezing) process. By controlling the water movement of each cell during the cooling process, the cold start performance of the stack can be improved.
It has been previously suggested that the startup procedure for a fuel cell stack at subfreezing temperature is hampered by the presence of ice in the porous catalyst layers of the electrodes. The ice prevents the reactant gases from reaching certain parts or even all of the electrodes' catalyst layer surfaces. To avoid such a situation, many proposals have been made for removing all of the water and water vapor from the stack when the stack is being shut down so that there is no possibility of ice being present upon re-establishing operation. Such systems are expensive, awkward, and quite time-consuming, and are certainly not at this time well suited for fuel cell power plants used in vehicles. The dry out of the cell stack assembly which is necessary for good cold start performance, can result in severe membrane stress, leading to untimely membrane failure.
Other approaches to the catalyst/ice problem include all sorts of heating methodologies, which are also expensive, cumbersome and require too much time, and are not well suited for vehicular applications.
Recognition of the fact that water in a fuel cell stack will tend to migrate toward the freezing front (toward the lower temperature along a temperature gradient), the liquid water permeability (water permeance) of gas diffusion layers (GDLs) is made lower than normal where a catalyst layer will be at a lower temperature than its corresponding water transport plate (WTP), and greater than normal where a catalyst layer will be at a higher temperature than its corresponding water transport plate. This gradation in GDL water permeance tailors the capability of the fuel cells to conduct water away from catalyst layers toward water transport plates, at either end of the stack, thus minimizing startup problems due to ice blockage of gas transport to the cells' catalyst layers.
Herein, the “anode end of the stack” and “anode end” are defined as the end of the stack at which the anode of the fuel cell closest to that end is closer to that end than the cathode of the closest fuel cell.
Specifically, at the anode end of the stack, each cells' anode water transport plate is closer to the stack end plate and therefore each WTP will be cooler than its associated anode catalyst layer, as the stack cools upon shutdown. As a result, during a shutdown procedure, water inventory normally tends to migrate through the anode gas diffusion layer (GDL) toward the water transport plate. Since this water migration is beneficial to fuel cell restart capability from a frozen condition, the GDL adjacent to each anode catalyst layer, at the anode end of the stack, has a greater than normal liquid permeability in order to promote water migration away from the anode catalyst layer.
On the other hand, at the anode end of the stack, the cathode catalyst layer is closer to the anode end plate and therefore colder than its associated cathode water transport plate. As a result, during a shutdown procedure, the fuel cell water inventory will normally migrate from the water transport plate (where it is abundant) toward the cathode catalyst layer. In order to impede this water flow, the cathode GDL is provided with lower than normal water liquid permeability.
When the stack temperature is below freezing, at the anode end of the stack, and freezing occurs in the small pores of the anode WTP, a decrease in the liquid pressure occurs drawing water out of the anode catalyst layer (toward the anode water transport plate) so that the anode catalyst layer dries out. On the other hand, as the water freezes in the small pores of the cathode catalyst layer, water is drawn out of the cathode water transport plate, through the cathode GDL and into the cathode catalyst layer. As the water is drawn into the cathode catalyst layer, the ice pressure increases, forcing small hydrophobic pores of the cathode catalyst layer, which are normally empty, to fill with ice. Once the pores of the cathode catalyst layer are filled, they are very difficult to empty. This cathode condition results in the performance loss seen after a boot strap start from freezing temperatures. While this phenomenon also works to fill the anode catalyst layer at the cathode end of the stack, the fuel cell is more tolerant of anode catalyst layer flooding due to rapid hydrogen/oxygen kinetics and hydrogen diffusion capability. Also, anode catalyst layer flooding is more easily recovered during normal fuel cell operation due to electro-osmotic drag of water from the anode electrode toward the cathode.
This water movement problem also exists in fuel cell power plants not utilizing water transport plates since there are small pores in the catalyst layers and water can move within the membrane electrode assembly itself. However, there is much less water inventory available to move within the cell (there is some liquid water in the GDLs and in the gas channels), so the problem is less severe.
The opposite situation occurs at the other end of the stack.
At the cathode end of the stack, the anode catalyst layer is closer to the cathode stack end plate and therefore cooler than its associated anode water transport plate as the stack cools upon shutdown. As a result, during a shutdown procedure, the fuel cell water inventory migrates from the water transport plate toward the anode catalyst layer. In order to impede this flow, the anode GDL at the cathode end of the stack is provided with lower than normal water permeability.
At the cathode end of the stack, the cathode water transport plate is closer to the cathode stack end plate, and therefore there is migration of water from the cathode catalyst towards the cathode water transport plate. To enhance this flow, the cathode GDL at the cathode end of the stack is provided with higher than normal permeability.
The arrangement herein may be utilized in several cells at each end of the stack, or up to one-half of the stack at each end of the stack if desired, but generally need not be utilized in every cell in the stack. For instance, applying the principles herein to 8 or 10 cells at either end of a stack will typically be sufficient to avoid ice blockage of reactant gases in the end cells. The arrangement may be used in fuel cell stacks with solid polymer electrolytes or liquid electrolytes. The arrangement may be used in power plants with external, internal, or some combination of water management systems, including evaporative cooling.
A second embodiment achieves a significant reduction in performance problems related to flooding electrode catalyst layers by taking advantage of the tolerance to flooding at the cell anodes referred to hereinbefore. In the second embodiment, the GDLs of cathodes and anodes at the anode end of the stack have lower than normal water permeability, while the GDLs of the cathodes and anodes at the cathode end of the stack have higher than normal water permeability.
A third embodiment also achieves a significant reduction in performance problems related to flooding of electrode catalyst layers by taking advantage of the tolerance to flooding at the cell anodes referred to hereinbefore. In the third embodiment, the GDLs of cathodes and anodes at the anode end of the stack have low water permeability, while at the cathode end of the stack, the GDLs of the cathodes have high water permeability and the GDLs of the anodes have low water permeability.
Other variations will become apparent in the light of the following detailed description of exemplary embodiments, as illustrated in the accompanying drawings.
Referring to
Similarly, air is provided through oxidant reactant gas flow field channels 27 which are depicted herein as being orthogonal to the fuel channels 20. The air channels 27 are formed on one surface of the cathode water transport plates 28 which have characteristics similar to those of water transport plates 21.
The catalysts are conventional PEM-supported noble metal coatings typically mixed with a perfluorinated polymer, such as that sold under the tradename NAFION® which may or may not also contain teflon. The PEM 10 consists of a proton conductive material, typically perfluorinated polymer, such as that sold under the tradename NAFION®. Water is transferred from the water channels 24 through the porous, hydrophilic WTPs 21 and the anode GDL 16, to moisturize the PEM. At the catalyst layer, a reaction takes place in which two hydrogen diatomic molecules are converted catalytically to four positive hydrogen ions (protons) and four electrons. The protons migrate through the PEM to the cathode catalyst. The electrons flow through the fuel cell stack out of the electrical connections and through an external load, doing useful work. The electrons arriving at the cathode combine with two oxygen atoms and the four hydrogen ions to form two molecules of water. The reaction at the anode requires the infusion of water to the anode catalyst, while the reaction at the cathode requires the removal of product water which results from the electrochemical process as well as water dragged through the PEM from the anode by moving protons (and osmosis).
The cathode catalyst layer 14 is similarly porous and the GDL 17 is porous to permit air from the channels 27 to reach the cathode catalyst and to allow product and proton drag water to migrate to the cathode WTP, where the water will eventually reach the water channels 24. In a power plant having an external water management system, the water will exit the stack for possible cooling, storage and return to the stack as needed.
Referring to
Along the bottom of
Variations in liquid water permeability may be achieved by adjusting the characteristics of the paper of which the GDL is formed, which is typically a mixture of fiber and particulate carbon, such as one of the readily available TORAY® papers, having suitable porosity and pore size for proper passage of reactant gas. The degree of hydrophobicity is then adjusted by adding an appropriate thin coating of a suitable polymer, such as PTFE. On the other hand, the paper can be produced with a desired hydrophobicity by including a suitable thermoplastic resin in the paper making process.
In the embodiment of
The embodiment of
As used herein, the gas diffusion layer is defined as being one or more layers interposed between an electrode and a water transport plate. It is sometimes called a support layer. Sometimes a support layer is referred to as having a substrate which is adjacent to the water transport plate as well as a microporous layer that is adjacent to the catalyst. Typically, the substrate will be relatively hydrophilic whereas the adjacent microporous layer will be relatively hydrophobic. Thus, a support comprising a substrate and a microporous layer will be referred to herein as a gas diffusion layer (GDL). On the other hand, a gas diffusion layer may only comprise what is essentially the same as a substrate layer of a two-layer gas diffusion layer. In this arrangement, the gas diffusion layer can be a single layer or it can be a dual layer or even have more than two layers.
The thickness, or porosity or wettability of the support layer may be adjusted in any combination to provide a greater or lesser impediment to the migration of water. However, the control of water permeability may also be imparted by the characteristics, particularly pore size and hydrophobicity, of the microporous diffusion layer, rather than the support.
The adjustments between high liquid water permeability GDLs and low liquid water permeability GDLs may, in some cases, be made on a relative basis, that is to say, having the anode end, cathode GDLs and the cathode end, anode GDLs with a water permeability which is some percentage of the water permeability of the anode end, anode GDL and the cathode end, cathode GDL. But generally, the absolute liquid water permeability of each GDL (or groups of GDLs) will be selected without regard to the liquid water permeability of other GDLs of the stack subject to other, different operational characteristics. Low liquid water permeability may range from near zero up to about 3×10−4 g/(Pa s m) and high liquid water permeability may exceed normal, which is about 3×10−4 g/(Pa s m).
Herein, the anode water transport plate 21 is illustrated as being separated from the cathode water transport plate 28, meeting at a seam which together form water passageways 24. However, it is possible that the water transport plates 21, 28 may be combined in some fashion without altering the advantage of the present arrangement.
Filing Document | Filing Date | Country | Kind | 371c Date |
---|---|---|---|---|
PCT/US2008/013601 | 12/11/2008 | WO | 00 | 5/13/2010 |