decrease more slowly when the temperature of the oxygen humidifier was raised. This indicates that less KOH is lost by the cell as the humidity of the oxygen in contact with the cathode is raised. But the weeping is never completely stopped. This observation is in agreement with the interpretation of the weeping effect in terms of a vapor-pressure gradient. The more humid oxygen would allow water to be condensed at the cathode, thus diluting the KOH at the electrode-gas surface and so reducing the difference in concentration across the thickness of the porous layer of the electrode. We have also observed that the weeping worsens with aging of the electrode. This could be due to a deterioration of the hydrophobic Teflon layer that is intended to prevent leakage from the cathode. It is known that PTFE is attacked by concentrated KOH solutions and we have made our own observations of the modifications of electrode surface layers. Scanning electron microscopy with energy dispersive x-ray analysis (EDAX) was used to measure the fluorine concentration in surface layers of three electrodes: one unused electrode, one that had been soaked in 7 M KOH for 500 hours and a third that was soaked in 14 M KOH for the same time. The treated electrodes were found to have lower surface concentrations of fluorine than the untreated one and the fluorine loss was greater at the higher KOH concentration. The fluorine is probably replaced by OH groups: this would result in the FIFE becoming less hydrophobic and could facilitate leakage through this layer. Other more quantitative studies of cathode weeping are now under way and should give more precise information about its nature. Conclusion The present study of water management in an H2/O2 fuel cell has shown that the excess hydrogen flow can be optimized, in terms both of flow rate and of humidity, and that this optimal condition corresponds to the case where the rate of water production is equal to its rate of elimination by the hydrogen. A numerical model of this process has been developed that gives a realistic description of the behavior of the system. Observations concerning the cathode weeping effect have also been made. These suggest that the loss of KOH can be reduced by humidifying the oxygen flowing through the cell. If this technique were to be adopted, the added water introduced into the cell would have to be included in our numerical model. In any case, this model should be of considerable value in the interpretation of future observations concerning the fuel cell operation.
RkJQdWJsaXNoZXIy MTU5NjU0Mg==