Page 309 - High Temperature Solid Oxide Fuel Cells Fundamentals, Design and Applications
P. 309

Testing of  EZectrodes, Cells and Short Stacks  285

























                       0          20         40         60          80
                                           %H,O  in feed

           Figure 10.1 6  Observable deviation (OCV - Emf in mV) translated into percent loss of feed hydrogen as a
                    function ofwaterconcentration in the fuelgas, H2/H20mixturesnt  850°C[46].


           estimated from the OCV-Emf  deviation.  Reading the curve in another way, a
            given loss of H2 (e.g. 5% of the feed) causes a loss of driving potential of 60 mV if
            the feed contains only 3% water, whereas for a feed with  50% water the loss
           is only 5 mV.
             One way to exploit this knowledge is to assess the leak-rate of  a test system
           while running water-lean  (sensitive) gas mixtures, and transfer the observed
           leak to tests carried out in more water-rich feed gases. A practical approach to
           minimising  the  seal-leak  problem  in  testing  could  include  establishing  a
           reducing sweep atmosphere, e.g. 3% H2 in N2 outside the test compartment as
           illustrated in Fig. 10.5a. This would reduce the pH2 gradient to well below one
           order of  magnitude, and prevent incoming leakage of  02. At the same time the
           potential of  the cathode gas is far less sensitive to reaction of  O2 with H2 and
           production of H20, for example, a 5% injection of H2 in air reduces the potential
           by only 7 mV at 850°C.
             An electronic leak through the electrolyte is in principle another explanation
           for the differences in Emf  and OCV. If the electronic conductivity in such a leak
           decreases with cell voltage, a very flat i-V  curve may be obtained.
             Great care should always be taken when the OCV is far from the theoretical
           Emf. When the measured OCV lies close (-10  mV) to the calculated Emf, it is
           justified to calculate resistances based on OCV rather than Emf when the purpose
           is a detailed celI performance evaluation, thereby avoiding lumping problems of
           gas-tightness of the electrolyte or seals into the cell resistance. (If the purpose of
           the  calculation  is  economic  assessment  or  system  studies,  the  calculations
           should obviously be based on Emf.) However, if  large differences between Emf
   304   305   306   307   308   309   310   311   312   313   314