DIRECT LIQUID FUEL CELLS
4.11 FUEL CELLS USING ORGANIC LIQUIDS AS FUELS
In its chemical properties, ethanol is similar to methanol but has a considerable advantage over methanol of much lower toxicity. It must be pointed out that from an ecological viewpoint, ethanol is exceptional among all other types of fuel. In the oxidation of all organic fuels (e.g., natural gas, oil-derived products, coal), carbon dioxide is formed, which leads to the well-known global warming effect in the atmosphere of the Earth: a global temperature rise. This is equally true for chemical uses of these fuels in heat engines by ‘‘hot combustion’’ as for electrochemical uses in fuel cells by ‘‘cold combustion.’’
Ethanol can be obtained by the fermentation of various agricultural biomasses, which in turn are formed by photosynthesis involving CO2 and solar energy. This implies that combustion, hot or cold, of ethanol will not lead to the accumulation of excess CO2and will not upset the overall balance of this gas in the atmosphere. Using ethanol as an energy vector in essence is a 88 DIRECT LIQUID FUEL CELLS
practical way of using solar energy. Ethanol is the only chemical fuel in renewable supply. In Brasil, mass production of ethanol from biomass has already started, the corresponding infrastructure was created, and large part of automotive transport is transformed so as to work with ethanol rather than with gasoline. In the European Union, ethanol is treated as a desirable fuel additive. Attention should be drawn to current discussions of the ecological (and other) implications, including those of ethanol competing with foodstuff agriculture. The overall energy content of ethanol (about 8 kWh/kg) is rather close to that of gasoline (about 10 kWh/kg).
It is quite natural that all these considerations have led to enhanced interest in the uses of ethanol in fuel cells. A lot of research went into the development of direct ethanol fuel cells (DEFCs) during the last decade. Much of the research in this direction was performed in France in the group of Claude Lamy (see the review of Lamy et al., 2002).
The reactions that occur ideally in an ethanol–oxygen fuel cell, and the associated thermodynamic parameters, are as follows:
Anode : C2H5OHỵ3H2O!2CO2ỵ12Hỵỵ12e E0ẳ0:084 V ð4:8ị Cathode : 6O2ỵ12Hỵỵ12e!6H2O E0ẳ1:229 V ð4:9ị
Overall: C2H5OHỵ3O2!2CO2ỵ3H2O E0ẳ1:145 V ð4:10ị
DGẳ 1325 kJ=molẳ1:145 eV DH ẳ 1367 kJ=molẳ0:969 eV The structure of a fuel cell using ethanol (DEFC) is little different from that of a fuel cell using methanol (DMFC). The same proton-conducting mem- branes made of Nafion or its replacements are used. As the catalyst for the anodic process, here again the Pt–Ru system provides better electrical perfor- mance indicators than does pure platinum. An important difference between these two varieties of fuel cells is the higher working temperature required with ethanol. This is due to the fact that the electrochemical oxidation of ethanol occurs with greater difficulties than the analogous reaction of methanol and is markedly slower. For acceptable electrical parameters, therefore, higher temperatures, ranging up to 2001C, must be used. For this reason, the problems that arise when building DEFCs often are the same as those described in Section 3.7 for PEMFCs working at elevated temperatures.
In experimental DEFCs working at a temperature of 901C, values of power density of 33 mW/cm2were realized at a voltage of 0.55 V (Lamy et al., 2004), which is a rather modest result. Arico et al. in 1998 reported that in an ethanol fuel cell with a Nafion membrane containing added silica (SiO2), the power density attained 110 mW/cm2at a temperature of 1451C and a voltage of 0.35 V.
4.11 FUEL CELLS USING ORGANIC LIQUIDS AS FUELS 89
In 1995, when studying an ethanol fuel cell with a membrane made of PBI impregnated with concentrated phosphoric acid at a temperature of 1701C and a voltage of 0.30 V, Wang et al. obtained a power density of 75 mW/cm2. These data, although obained under different conditions and hardly comparable, show that when working at elevated temperatures, ethanol fuel cells may yield electrical performance indicators comparable to those of methanol fuel cells.
On the basis of these results, it is assumed by a number of workers that it will be possible to replace methanol with ethanol in membrane-type fuel cells working with liquid fuels. There is, however, a basic point that casts doubt on this conclusion.
In the electrochemical oxidation of methanol, carbon dioxide gas is the chief reaction product. The yields of other potential products of the oxidation reaction, such as formaldehyde, formic acid, and the like, are a few percent at most. Arico et al. (1998) concluded from a chromatographic analysis of the reaction products that the chief product of electrochemical oxidation of ethanol (with a yield of about 98%) is CO2, just as for methanol. This conclusion is inconsistent with the results obtained by other workers. Wang et al. (1995) studied the reaction products of ethanol and propanol oxidation by differential electrochemical mass spectrometry. They found that during the reaction, only 20 to 40% of the theoretical yield of CO2is produced, whereas acetaldehyde is formed to 60 to 80% (even traces of acetic acid are formed). Rousseau et al. (2006) used a high- performance liquid chromatograph for analysis of the products of ethanol oxidation. According to their data, about 50% aldehyde, 30% acetic acid, and only about 20% CO2are formed at a temperature of 901C at a platinum catalyst.
With Pt–Sn or Pt–Sn–Ru catalysts, somewhat different numbers were obtained:
15% aldehyde, 75% acid, and 10% CO2. It follows from these data that the composition of the reaction products depends heavily on the catalyst used for the electrode reaction. Conditions have not been found that would reliably provide CO2yields approaching 100%. This is of basic significance.
The formation of acetic acid by ethanol oxidation is a four-electron process:
C2H5OHỵH2O!CH3COOHỵ4Hỵỵ4e ð4:11ị while the formation of acetaldehyde is a two-electron process:
C2H5OH!CH3CHOỵ2Hỵỵ2e ð4:12ị For the sake of comparison, note that the electrochemical oxidation of ethanol to CO2according to reaction (4.8) is a 12-electron process.
The smaller number of electrons involved in the reaction leads to a con- siderable decrease in the de facto energy content of ethanol: from 8 kWh/kg for the 12-electron process to 2.6 kWh/kg for a four-electron process and to as little as 1.3 kWh/kg for a two-electron process.
90 DIRECT LIQUID FUEL CELLS
The drop in energy content is not the only negative consequence of these side reactions. In contrast to CO2gas, which is readily vented from the system into the atmosphere, a number of problems of removal and disposal arise when aldehyde and acetic acid are formed in the reaction.
It is surprising that in many publications on ethanol fuel cells, this aspect is not examined more closely. It is only in the conclusions of the review of Lamy et al. (2002) that the importance of this problem is stressed.*
The formation of acetaldehyde and acetic acid is due to the fact that rupture of the C–C bond in the original ethanol molecule would be required for CO2formation. In the chemical oxidation of ethanol at high temperatures (combustion), this bond is readily ruptured in the hot flame, and the only reaction product (in addition to water) is CO2. In electrochemical oxidation occurring at temperatures below 2001C, this bond is very difficult to rupture, and reactions involving the sole rupture of C–H bonds occur much more readily, thus leading to the side products noted.
The very idea of building ethanol fuel cells is very attractive and promising.
To this day, development of these fuel cells is still in its initial stages, a large amount of research in electrocatalysis being required to find ways of making fully adequate ethanol fuel cells. This work should lead to the development of basically new polyfunctional catalysts that would secure a high rate not only of low-temperature oxidation of organic substances through C–H bond rupture but simultaneously, reactions involving rupture of the C–C bonds. Only then will it be possible to think of a widespread application of ethanol fuel cells for future highly efficient, ecologically harmless electric vehicles.
In many publications cited in the literature, the possibility of using higher alcohols (e.g., 1-propanol, 2-propanol) and other, analogous organic com- pounds (e.g., ethylene glycol, acetaldehyde, dimethyl ether) in fuel cells has been studied. The electrochemical oxidation mechanism of these substances is more complex and associated with the formation of many more side products than in the case of ethanol. In all these papers the high theoretical specific energy content of these compounds is mentioned, and experimental results concerning the power density and a few other parameters are reported, but as a rule, nothing is said as to the depth of oxidation attained (the percentage of CO2among the reaction products).
4.11.2 Direct Formic Acid Fuel Cells
Formic acid (HCOOH) is a somewhat unusual type of ‘‘fuel’’ for fuel cells, and in many of its properties differs from other substances used as fuels. On the one hand, the theoretical energy content of formic acid (1.6 kWh/kg) is much lower than for all other reactants considered in this section. On the other hand, the equilibrium electrode potential for the oxidation of formic acid (0.171 V) is more negative than that for the other organic fuels; that is, in a thermodynamic
* This aspect was also mentioned (without discussion) in a review by Antolini (2007).
4.11 FUEL CELLS USING ORGANIC LIQUIDS AS FUELS 91
sense, formic acid is a very strong organic reducing agent. The thermodynamic EMF of a formic acid–oxygen cell is 1.45 V.
Formic acid has a number of properties that make its use in fuel cells very attractive. First, this substance is ecologically absolutely harmless (the U.S.
Food and Drug Administration allows it to be used as a food additive). Its only oxidation products are CO2and water. It is, essentially, not possible that side products or intermediates will be formed. Formic acid is a liquid. In aqueous solutions it dissociates, yielding HCOOions. This is of basic significance for its use in fuel cells with proton-conducting membranes. These membranes have a skeleton containing negatively charged ionic groups [in the case of Nafion, sulfonic acid groups (SO3)]. Because of the electrostatic repulsion, these groups hinder (or at least strongly retard) the penetration of formate ions into the membrane. Thus, in the case of formic acid, the effect that constituted the major difficulty for the development of methanol fuel cells, crossover of the anodic reactant from the anodic region through the membrane to the cathodic region, is practically inexistent.
The electrochemical oxidation of formic acid,
HCOOH!CO2ỵ2Hỵỵ2e ð4:13ị
occurs along one of two possible pathways:
1. By a step of catalytic decomposition (dehydrogenation):
HCOOH!CO2ỵH2 ðor 2MHadsị ð4:14aị followed by hydrogen ionization:
H2!2Hỵỵ2e ð4:14bị 2. By a chemisorption step, including dehydration:
HCOOH!MCOadsỵH2O ð4:15aị
followed by the steps of
H2OỵM!MHadsỵHỵỵe ð4:15bị and
MCOadsỵMOHads!CO2ỵHỵỵe ð4:15cị (here M is a metallic catalyst surface site).
92 DIRECT LIQUID FUEL CELLS
The overall current-producing reaction in the cell is
HCOOHỵ12O2 !CO2ỵH2O ð4:16ị The low theoretical energy content of this substance is due to the fact that only two electrons per formic acid molecule are involved in the reaction.
Yet the reaction is so simple that undesirable side products cannot be formed in it.
Formic acid is used in membrane-type fuel cells as an aqueous solution. A 20 M HCOOH solution contains about 75% formic acid. Owing to the low water content, the membrane is not sufficiently moistened in such a concen- trated solution, and its resistance increases (Rice et al., 2002). In solutions less concentrated than 5 M, the current densities that can be realized are low, owing to the slow reactant supply by diffusion to the catalyst surface. An optimum concentration for fuel cell operation is 10 to 15 M. In contrast to DMFCs, an increase in reactant concentration in DFAFCs does not produce complications related to reactant crossover.
At a temperature of 701C, a power density of about 50 mW/cm2was attained in 12 M HCOOH at a working voltage of 0.4 V. By comparison, the power density in a typical methanol fuel cell under the same conditions is about 30 mW/cm2.
With the experience gathered in the development of DMFCs, Pt–Ru catalysts were used for the anodic process in the early studies on DFAFCs.
Ha et al. (2006) showed that much better electrical characteristics can be obtained with palladium black as the catalyst. Importantly, with this catalyst one can work at much lower temperatures. In particular, at a temperature of 301C, power densities of 300 mW/cm2 were obtained with a voltage of 0.46 V, and about 120 mW/cm2 with a voltage of 0.7 V. The differences between the two catalysts probably are due to the fact that with Pt–Ru, formic acid oxidation follows the second of the mechanisms mentioned (chemisorption with dehydration), while palladium black is a highly effective catalyst for the dehydrogenation of formic acid, the first mechanism being followed on it. It must be pointed out that this effect is highly specific;
in methanol oxidation, the catalytic activity of palladium is lower than that of Pt–Ru.
Considering all these special features, it will be very convenient to use formic acid as a reactant in fuel cells of small size for a power supply in portable equipment ordinarily operated at ambient temperature. Such fuel cells are described in Chapter 14.
In a review by Yu and Pickup (2008), recent advances in DFAFCs are presented, focusing mainly on anodic catalysts for the electrooxidation of formic acid. The problem of formic acid crossover through Nafion membranes is also discussed.
4.11 FUEL CELLS USING ORGANIC LIQUIDS AS FUELS 93