MOLTEN CARBONATE FUEL CELLS
7.1 SPECIAL FEATURES OF HIGH-TEMPERATURE FUEL CELLS
The possibility of efficiently using the reaction heat for generating additional electrical energy
A high rate of the electrode reactions and relatively little electrode polarization, hence no need to use platinum catalysts
The possibility of using technical hydrogen with a large concentration of carbon monoxide and other impurities
The possibility of using carbon monoxide, natural gas, and a number of petroleum products directly through internal conversion of these fuels to hydrogen within the fuel cell itself
Yet changing over to higher temperatures implies a certain diminution of thermodynamic indices of the fuel cells. The Gibbs free energyDGof hydrogen oxidation by oxygen decreases with increasing temperature. It amounts to 1.23 eV at 251C but decreases to 1.06 eV at 6001C and to 0.85 eV at 10001C.
Fuel Cells: Problems and Solutions,By Vladimir S. Bagotsky Copyrightr2009 John Wiley & Sons, Inc.
125
(Identical figures in volts are the numerical values for the thermodynamic EMF of hydrogen–oxygen fuel cells at these temperatures.) On the other hand, the reaction enthalpy,DH, which at 251C has a value of 1.48 eV, will not change significantly with increasing temperature. For this reason the thermodynamic efficiency of the reactionZthermod=DG/DH also decreases with increasing temperature. At temperatures of 25, 600, and 10001C, it has values of 0.83, 0.72, and 0.57, respectively. In fuel cells operated at higher temperatures, numerous problems associated with the limits of chemical and mechanical stability of various materials that are used in them also become important.
7.2 STRUCTURE OF HYDROGEN–OXYGEN MCFCs
The working temperature of MCFCs is around 600 to 6501C. As an electrolyte, mixed melts containing 62 to 70 mol% of Li2CO3and 30 to 38 mol% of K2CO3, with compositions close to the eutectic point, are used in MCFCs. Sometimes, Na2CO3 and other salts are added to these melts. These liquid melts are immobilized in the pores of a ceramic matrix with fine pores made of sintered MgO or LiAlO2powders.
Porous metallic gas-diffusion electrodes are used. The anode consists of a nickel alloy with 2 to 10% chromium. The chromium that is added prevents a recrystallization and sintering of the porous nickel while it works as an electrode. This action is based on chromium forming a thin layer of chromium oxide at the nickel grain boundaries, interfering with surface diffusion of the nickel atoms.
The cathode consists of lithiated nickel oxide. Nickel oxide (NiO) is ap-type semiconductor that has rather low conductivity. When doped with lithium oxide, its conductivity increases tens of times, owing to a partial change of Ni2+ to Ni3+ ions. The lithiation is accomplished by treating the porous nickel plates with hot LiOH solution in the presence of air oxygen. The compound produced has a composition given as LiþxNi21þxNi3xþO. This lithiation of nickel oxide was first applied in 1960 by Bacon in his alkaline fuel cell (Section 2.2).
With its fine pores filled with the carbonate melt the matrix is a reliable protection against gases bubbling through and getting into the ‘‘wrong’’
electrode compartment. Therefore, there is no need to provide gas-diffusion electrodes with a special gas barrier layer.
The MCFC components have thicknesses as follows: the anode, 0.8 to 1.5 mm;
the cathode, 0.4 to 1.5 mm; and the matrix, 0.5 to 1 mm. In a battery of the filter- press type, the individual cells are separated by bipolar plates made of nickel- plated stainless steel contacting the anode with their nickel side, and the cathode with their steel side. All structural parts are made of nickel or nickel-plated steel.
In a working battery, the temperature of the outer part of the matrix electrolyte is lower than that of the inner part, so that in the outer part the electrolyte is solidified. This provides tight sealing around the periphery of the individual battery cells.
126 MOLTEN CARBONATE FUEL CELLS
In a hydrogen–oxygen MCFC, the following reactions take place:
Anode : H2ỵCO23 !H2OỵCO2ỵ2e E0ẳ0 V ð7:1ị
Cathode : 12O2ỵCO2ỵ2e!CO23 E0 ẳ1:06 V ð7:2ị
Overall : H2ỵ12O2!H2O E0ẳ1:06 V ð7:3ị
It is a special feature of the electrode reactions in MCFCs that unlike most other versions of fuel cells, the cathodic reaction consumes not only oxygen (or air) but also carbon dioxide (CO2). In the anodic reaction, CO2is evolved at the anode. It is imperative, therefore, in MCFC design to provide for the possibility of CO2evolved at the anode to return to the cathode (Figure 7.1).
The values of electrode potential given in reactions (7.1) and (7.2) refer to the condition where all gases involved in the reactions (H2, O2, CO2, and water vapor) are in their standard states (i.e., have partial pressures of 1 atm). During the reaction, in fact, their amounts and hence also their partial pressures change constantly. This implies that the equilibrium potentials of the electrodes also change. According to the Nernst equation, the potential of the hydrogen anode is
Hydrogen in
Electrical current
H2
Oxygen in O2
O2
CO2
CO2
CO2 CO2
e−
e− e−
e−
e− e−
e− e− e−
e−
H2O Water and
heat out
Anode
Electrolyte
Cathode Carbon dioxide in CO3−2
FIGURE 7.1 Schematic of reactant flow in an MCFC. (From Wikipedia, the free online encyclopedia.)
7.2 STRUCTURE OF HYDROGEN–OXYGEN MCFCs 127
given by
EaẳEa0ỵRT
2F lnpH2OpCO2 pH2
ð7:4ị
the potential of the oxygen electrode is given by
EcẳEc0ỵRT
2F lnpO1=22 pCO2 ð7:5ị If the reactant gases were enclosed, the potential of the hydrogen electrode would shift in the positive direction with the progress of H2consumption, while the potential of the oxygen electrode would shift in the negative direction with the progress in O2 and CO2 consumption. As a result, the overall value of thermodynamic EMF, EẳEcEa, would fall below the standard value of 1.06 V. The losses of overall efficiency of the energy conversion process that are associated with these changes have been termedNernst losses. In an operating fuel cell, however, fresh reactant gases are supplied continuously to the gas compartments, so that these losses are mitigated.
Depending on the partial pressures of the reactant gases, the open-circuit voltage of hydrogen–oxygen MCFCs has values of 1.00 to 1.06 V. A practically linear relation between cell voltageUand current densityiis a special feature of these fuel cells. This implies a practically constant value of the apparent internal resistance Rapp in equation (1.11). This leads to a more important voltage decrease with increasing cell discharge current than in other fuel cell types. At a pressure of the reactant gases of 1 bar and a current density of 100 mA/cm2, the voltage is about 0.85 V; at a current density of 150 mA/cm2 it is only about 0.6 V. Current densities above 150 mA/cm2are practically not used in MCFCs.
The linear relation between voltage and current density is not a result of the internal resistance of the cell being a purely ohmic resistance. Apart from the ohmic voltage drop across the electrolyte kept in the pores of the matrix, a marked contribution to the voltage decrease comes from polarization of the oxygen electrode. At a current density of 100 mA/cm2, this amounts to 0.5 V.
The polarization of the hydrogen electrode is considerably lower.
7.3 MCFCs WITH INTERNAL FUEL REFORMING
From the very outset of MCFC development, research workers were attracted by the fact that not only hydrogen but also carbon monoxide (CO) could be used as a reactant fuel (reducing agent). Carbon monoxide (as water gas, a mixture of CO and H2) is obtained readily by the steam gasification of coal:
CỵH2O!COỵH2 ð7:6ị
128 MOLTEN CARBONATE FUEL CELLS
This opens up possibilities for an indirect electrochemical utilization of huge coal reserves.
The possibility of a direct electrochemical oxidation of carbon monoxide was soon questioned. It was suggested that hydrogen, rather than carbon monoxide, is involved in the electrochemical reaction after being formed from CO by the Boudouard reaction:
COỵH2O!CO2ỵH2 ð7:7ị which occurs readily under the operating conditions of the carbonate fuel cell.
Sparr et al. (2006) could show that pure CO can actually be oxidized electrochemically according to the reaction
COỵCO23 !2CO2ỵ2e ð7:8ị although in fact the rate of this reaction is about 20 times lower than the rate of hydrogen oxidation [reaction (7.1)]. The exchange current density (as an index of reaction rates) is about 0.7 mA/cm2for the oxidation of water gas (a mixture of 56% H2+ 8% CO + 28% H2O); that for the oxidation of pure hydrogen is about 1 mA/cm2. This implies that the use (‘‘combustion’’) of CO in MCFCs occurs primarily, but not exclusively, via the intermediate formation of hydrogen. The possibilities of using products of gasification of biomass or residential waste that have a large CO content but include a variety of contaminants have been examined in a paper by Watanabe et al. (2006).
The rates of direct electrochemical oxidation of hydrocarbons (particularly of methane) in MCFCs are negligibly small. Therefore, natural types of fuel can be used in the fuel cells only after prior conversion (reforming) to hydrogen.
Since the reforming process is highly endothermic and requires a large heat supply, the idea sprang up that this process should be carried out within the fuel cell itself, where the heat of reaction evolved in the fuel cell could be utilized for the reforming (by analogy with the example of CO conversion to hydrogen reported above).
Two types of internal-reforming fuel cells (IRFCs) are distinguished: the direct internal-reforming fuel cells (DIRFCs), in which the reforming process takes place within the fuel cell at its anode catalyst, and the indirect internal- reforming fuel cells (IIRFCs), in which plates with special reforming catalysts are included within the battery stacks.
The most active company working in this field is FuelCell Energy (FCE) of Danbury, Connecticut, which, starting in the late 1970s, developed and delivered a large series of power plants designated as Direct Fuel Cells. These plants include a combined IRFC system. In the battery, special plates for prior fuel reforming are placed between groups of 8 to 10 DIRFCs. With this system one can achieve a
7.3 MCFCs WITH INTERNAL FUEL REFORMING 129
more uniform temperature distribution within the battery; heat is evolved in the fuel cells and is consumed at the reforming plates (Doyon et al., 2003).
A somewhat different version of reforming units combined with a battery of fuel cells was presented by the German company MTU under the designation Hot Modules. All the components of the MCFC system (the horizontally arranged fuel cell batteries, the catalytic burner of the anode tail gases, and a cathode recycle loop), as well as the fuel processing system, are placed into a common, thermally insulated vessel. In this version, special manifolds for the cathodic gases are not needed (Bischoff and Huppman, 2002).