The Effects of Cationic Contamination on Pem Hydrogen Fuel Cells
Brian Kienitz1, Harihara Baskaran2, and Thomas Zawodzinski2. (1) Case Western Reserve University, A.W.Smith Building, 10900 Euclid Avenue, Cleveland, OH 44106, (2) Chemical Engineering, Case Western Reserve University, A.W.Smith Building, 10900 Euclid Avenue, Cleveland, OH 44106
When cationic impurities, such as Na+, Ca2+, or metal ions, are present in a polymer electrolyte membrane fuel cell (PEMFC) the performance of the PEMFC can be significantly reduced. These contaminants can enter the membrane and catalyst layers from other fuel cell components or from the external environment. Cations replace protons in the membrane in accord with normal ion exchange processes. Cations, in general, have a higher affinity than protons for the sulfonic acid group sites in the polymer electrolyte. This is a problem because replacement of protons results in reduced conductivity, increased electro-osmotic drag and reduced water levels at saturation. To better understand how cationic contaminants affect a fuel cell the effects of contamination were studied and modeled. Polarization curves were taken on fuel cells contaminated with cationic species. These curves were used to determine the effects of contamination on the fuel cell membrane and catalyst layer. From these tests four main effects of cationic contaminants on PEMFC's were determined. The first observation was that the maximum obtainable current in the cell decreases. Also, it was noted that steady state was not reached at high over-potentials. The exchange current density is decreased. Finally, little change in apparent working membrane conductivity was detected although it is known that the conductivity of a fully poisoned membrane can be an order of magnitude lower than that of a fully protonated membrane. These effects can be explained through a membrane model that accounts for both the diffusion and migration of contaminants and protons in the system and assumes the flux of cationic contaminants is small compared to the flux of protons. The fluxes of ions in the membrane were described by the Nernst-Plank equation and measured data reported in the literature for conductivity, diffusion coefficients, water saturation and electro-osmotic drag. The steady state concentration profile for each ionic species in the membrane at a given current and impurity level was calculated from the flux equations. In addition, steady state concentration profiles were compared at levels of constant current with varying cationic contaminant concentration and varying current under constant contaminant concentration. From this model it is evident that cationic contaminants will always be more concentrated at the cathode side of the fuel cell when any current is drawn. This profile arises due to the interplay between flux due to diffusion and migration in the membrane with the electrode blocking to one ion. Since the cationic flux is low or zero the two components to the flux must balance each other. In contrast, the proton flux must be constant through out the membrane. Near the anode migration dominates this flux. Near the cathode the proton flux is dominated by diffusion. At the point of maximum current the fraction of cationic impurities occupying acid sites is unity near the cathode/membrane interface. Conversely, the proton concentration at this point must be zero since we are using a fixed site ionomer membrane. The transport number of protons moving by migration must also be zero at this point since the flux due to migration is proportional to concentration. Therefore, the maximum current density in the cell occurs when proton flux at the cathode is due to diffusion only. This maximum current value is not a true limiting current in the classical sense since a finite value for Ohmic over-potential can be determined for the maximum current. Beyond this voltage drop no steady state concentration profile exists. Cationic contaminants pile up on the cathode end of the fuel cell and the current heads towards zero. Therefore it is hypothesized that the apparent change in limiting current is actually an effect of the membrane and not actually a limiting current effect. The onset current and potential of this maximum current effect was determined from the model. It was determined that this maximum current effect would not be noticeable in normal fuel cell operation unless cationic contaminants exceeded roughly 50% of the total acid sites. The dependence of this maximum current on membrane saturation, impurity valency and membrane thickness were also investigated. It was found that decreasing the saturation and increasing the valency of the contaminant resulted in a lowering of the maximum current for a particular contaminant concentration. The trend of dependence of maximum current on membrane thickness showed that thin membranes with the same percent contamination had higher maximum currents but when the total number of available sites for contamination was accounted for thicker membranes had higher maximum currents. The catalyst layer is also affected by the concentration profile of cations in the cell since it is partially made of polymer electrolyte. The reduced proton concentrations lower the reaction rate which in turn increases over-potentials by lowering the apparent exchange current density. This model gives insight into the main observed effects of fuel cells operating with cationic concentration. A reduction in maximum current, onset of an unsteady state regime, and little relative change in the working conductivity of a membrane under cationic contamination were accounted for by modeling ion transport in the membrane. This model can potentially be used to aid in the diagnostics of contaminated fuel cells and the design of future fuel cell components.