Assessment of PEMFC Durability by Effluent Analysis
Doctoral thesis
View/ Open
Date
2011Metadata
Show full item recordCollections
- Institutt for kjemi [1403]
Abstract
A review of literature relevant for the understanding of chemical degradation in Polymer Electrolyte Membrane (PEM) systems was conducted. The proposed pathways to the formation of chemical species suspected of causing degradation were summarized. The structures of the most common PerFluorinated Sulfonic Acid (PFSA) ionomers were presented. The proposed mechanisms of chemical degradation were listed with reference to their point of attack on the ionomer structure. The role of Fluoride Emission Rate (FER) as a metric for chemical degradation rate was discussed. Several papers present results where fluorocarbon fragments are present in the effluent water. The quantification of these fragments was scarce, but results indicate that their amount may depend on fuel cell test hardware as well as the test operating conditions.
A novel methodology for simultaneous analysis of fluorine and fluoride were presented. By alkaline fusion, the fluorocarbons were digested and fluorine completely recovered as fluoride. Fluoride concentrations were then estimated for both digested and untreated samples by means of potentiometric analysis. By application of a fluoride selective electrode in acidic media, improved sensitivity was obtained. The proposed methodology was evaluated statistically. Using PolyTetraFluoroEthylene (PTFE) as a model compound, excellent recovery of fluorine was documented.
A review of papers presenting FER data was performed and the results were summarized with reference to the operating conditions testing were conducted. FER data span several decades in magnitude. A non-accelerated durability experiment was conducted to evaluate both FER as well as total Fluorine Emission Rates (FtER) throughout the experiment. Cathode and anode effluents were scrubbed in sodium hydroxide gas bubblers. In addition, a fluoride selective electrode was mounted in an anode gas bubbler in order to continuously monitor the accumulated fluoride concentration. The results show that the FtER to FER ratio is significantly higher than one. For the cathode FtER was initially high and decreasing with time. Cathodic FER was increasing although less pronounced than the FtER decrease. As a result adecreasing FtER to FER ration was found. At the anode, a correlation between onlineFER and the FER found from batch analysis was found; both increasing with time.Cathodic FtER and FER were always higher than their anode equivalent. Nocorrelation with fuel cell performance data was found. The results showed that for thenon-accelerated test protocol applied, significant amounts of fluorocarbons leaves thefuel cell. Further, the ratio of FtER to FER also changed with time, suggesting that theFER metric does not represent a constant fraction of the ionomer loss throughout theexperiment.
An accelerated durability experiment was conducted in order to evaluate FER as a function of gas flow rates at Open Circuit Voltage. It was found that a five-fold increase in gas flow rates resulted in FER that was more than five times higher. It was speculated that the increase in FER was caused by increased gas cross-over. For both gas flow levels, FER was sharply decreasing with time. Performance losses were higher for high gas flow rates, although a direct correlation between performance data and FER could not be obtained.
Experimental design was applied in order to evaluate several operational parameters and their impact on fuel cell performance and durability. An ON/OFF accelerated protocol was applied. Gas humidification level, clamping pressure and pressurized operation were used as input parameters. It was found that 100 % humidified reactant gases and pressurized operation gave superior performance and durability. FER was lower for high humidification, but higher for pressurized operation. In other words, the lowest FER was not found for the experiment showing superior durability. Interestingly, similar FER was recorded for repeated experiments where carbon flow fields were replaced with stainless steel ones. SER was found to be much higher than FER. As the ionomer structure suggests SER to be significantly lower than FER, it was believe that the ionomer was not the only source of sulfur. Sulfur levels were initially high, and highest for experiments with 100 % humidified gases. After an initial decay, SER nevertheless stabilized at a higher level than FER. It was not understood why the assumed contamination of sulfur did not wash out with time. An attempt to evaluate the background levels of fluoride and sulfate in the fuel cell test rig was attempted. It was found that significant levels of both were present at the test startup. Purging with inert gas did however reduce the fluoride and sulfate levels to acceptable levels after a few hours. It was also shown that Gas Diffusion Layers (GDL) was a possible source of fluoride and sulfur in the system