- Email: [email protected]
Hydrocarbon activation in solid state electrochemical cells
SOUD STATE IONICS
Solid State Ionics 57 (1992) 259-264 North-Holland
Hydrocarbon activation in solid state electrochemical cells I.S. Metc, alfe a, P.H. M i d d l e t o n m,P. Petrolekas a a n d B.C.H. Steele Department of Material~ and °Department of Chemical Engineering and Chemical Technology, Imperial Collegeof Science, Technologyand Medicine, London SW7, UK Received 21 October 1991;acceptedfor publication 30 November 1991
Hydrocarbonactivation in solid state electrochemicalcells is of interest both from a fuel cell perspective and for the development of ceramic electrochemicalreactorscapableof producingusefulchemicals. For both of these areas of research it is important to investigate and understand the behaviour of new electrode materials. Three different reactants, hydrogen,methane and propone, were studied over two different electrode materials, niobia-ceria and lithium ferrite (LiFeO2). Both cyclicvoltammetric work and steady-state electrode polarisation at 817°C showed that the reactivity of the gases decreased in the order, hydrogen> propene> methane, over the niobia-ceria electrode. Low temperature cyclicvoltammetric work at 550°C on the lithium ferrite showedthe presenceof an oxygenstate whichdisappeared as the temperature was increasedto 600°C.
1. Introduction Solid oxide fuel cells have been developed to operate on syngas fuels using nickel-zirconia anodes [1,2]. Overall efficiencies for energy production could be improved if it were possible to feed natural gas directly to the fuel cell without the need for an external reforming stage. It may be possible to use a premixed feed stream of natural gas and steam for the fuel cell with the reforming reaction taking place internally; this mode of oporation is the focus of considerable attention [3,4]. The highly endothermic nature of the reforming reaction means that there is a problem with high temperature gradients because of the cooling effect in the vicinity of the fuel inlet. Alternatively dry natural gas (or at least natural gas with a highly reduced steam content) could be used as the fuel cell feed. However, nickel-zirconia anodes are unsuitable for such a mode of operation because of high rates of carbon deposition on the nickel electrode resulting in a rapid loss of performance. At Imperial College a research programme has been initiated to investigate the direct electrochemical ox~" Paper presented at SSI-8, Lake Louise, Canada, October 2026, 1991. To whomcorrespondenceshouldbe addressed.
idation of hydrocarbons over alternative anode materials. Although methane is the primary constituent of natural gas there are also significant quantities of other hydrocarbons such as ethane and propane. These higher hydrocarbons can be of particular importance as regards carbon deposition. In addition to fuel cell applications the research involves developing the use of SOFC technology for the production of useful chemicals (this may occur simultaneously with the production of energy). In such a mode of operation, known as a ceramic electrochemical reactor, it is possible to control the supply of oxygen to the anode, where the partial oxidation process of a hydrocarbon may occur with an improved reaction selectivity. For both fuel ceU and ceramic electrochemical reactors it is therefore important to investigate the behaviour of a number of potential feed streams. In this study, feeds of hydrogen, methane and propene were selected. Hydrogen was studied because of its present importance as a fuel cell feed, methane because of interest in its direct oxidation in fuel cells and propene because it can undergo a partial oxidation resulting in the formation of acrolein (an important industrial process). Both areas of application also require the development and increased understanding of new anode
0167-2738/92/$ 05.00 © 1992ElsevierSciencePublishers B.V. All fightsreserved.
260
I.S. Metcalfe et al. / Hydrocarbon activation in solid state cells
materials. Recent work has shown ceria to be a very promising electrode material for methane oxidation [ 5 ]. However, during oxidation/reduction cycles the lattice parameter of ceria changes causing contraction and expansion leading to mechanical failure of the electrode. To overcome this problem doped ceria electrodes have been investigated. Results with niobia-ceria will be presented in this paper. Such ceria-based electrodes, whilst being of interest for the total oxidation of hydrocarbons, are not particularly suitable for partial oxidation processes and so data from a LiFeO2 electrode will also be presented. The technique of cyclic voltammetry was used to investigate the electrochemical behaviour of the electrodes in the different fuel atmospheres. Such a technique yields information regarding the polarisation behaviour of the cell and can be used to develop a mechanistic explanation of the electrode processes. Outlet gas compositions were not analysed, so it is impossible to report the selectivity to acrolein in the case of the partial oxidation of propene.
precursor powders obtained from Aldrich Chemicals Ltd. After mixing the appropriate proportions and calcining, the resultant catalyst powders were added to a tape casting slurry, in order to make the film electrodes. These were made by spreading the slurry over a pre-sintered zirconia electrolyte sheet and sintering at the appropriate temperature. The cell was completed by adding the platinum counter and reference electrodes. The cell housing was made of a machinable ceramic, and bonded together with a glass sealing compound. The total cell design constitutes a divided electrochemical cell, with fuel on one side and air on the other. Cyclic voltammetry and current interruption measurements were performed with a combination of standard instrumentation and in-house electronics; the data were recorded on xy and yt recorders. The design of the cell (see fig. 1 ) was influenced by the requirement to use prefabricated zirconia foil electrolyte sheets, and the need for rapid product analysis. The use of machinable heat resistant ceramic for the cell housing was preferred, even though this created a thermal mismatch between the material and the zirconia electrolyte. The alternative use of machinable glass (Macor) was less favoured because of its instability under reducing atmospheres
2. Experimental The oxide electrode materials were made by the solid state diffusion process, using high purity oxide
Anode Reference |
Glass s e a l / ~ " - , ~ / ~ / ~
SZ plate Cathode Reference 2
erio
"\
~,.
~7"ff"ff~[~-4-Reference I
V///////AI
*%
°e!"o
r//~. . . . . .
--
5cm
Fig. l. Schematic of the electrochemical cell.
--
j
261
LS. Metcalfe et al. ~Hydrocarbon activation in solid state cells
20-
hydr°~propene
16< E
12-
-4-8-12'
-~oo
i
-~oo'
-;oo'
-5oo' ' - ~ o o '
Potential(nnv)
-;oo
~oo' '
3oo
Fig. 2. Cyclic voltammogram of niobia--ceria electrode at 8170C. Sweep rate --30 mV/s. Hydrogen ( 10% in inert) flowrate= 155 ec/ rain. Propene (4% in inert) flowrate=270 cc/min. Methane ( 10% in inert) = 240 cc/min.
and also the problem of long term creep at temperaturesabove 800°C. The use of molten glass in the seals compensated for the thermal mismatch, the choice of glass being determined by the operating temperature range. The cavities within the cell blocks were machined to a depth of 1 mm, giving a total volume of 2 cm 3. This was considered adequate for rapid transfer of the products to the analysis section. The fuel electrode had a projected area of 9 cm 2. The gas supplies to both compartments of the cell were through narrow bore alumina tubes of I mm internal diameter. The flow rates were regulated using electronic mass flow controllers. The total inlet and outlet flows were monitored to check for leaks. The overall integrity of the seals was checked by analysis of residual air whilst purging the cell with helium. In this paper we report on the preliminary electrochemical studies from this cell design. In later publications we intend to also report on the product analysis by gas chromatography. Cyclic voltammetry was used as a convenient and rapid method for initial electrochemical investigations and, because it is a non-steady-state technique, it can reveal transient phenomena and be used to distinguish between" diffusion controlled processes and adsorption processes. In the work presented here we have used this method to give a qualitative description of system activity.
The current interruption method has been used to provide accurate IR free data for the purpose of obtaining the effective exchange current densities of the reactions.
3. Results and discussion 3.1. Ceria
Fig. 2 shows the results of cyclic voltammetric work with the niobia--eeria electrode at 817 °C with three fuels, hydrogen, propene and methane. As can be seen from the figure, higher currents are obtained in the presence of a hydrogen-containing atmosphere than in a propene-containing atmosphere, In turn, higher currents are observed in a propene-containing atmosphere than in one containing methane. These differences can be explained in terms of the reactivities of the individual molecules, hydrogen being very reactive and methane being a much more stable hydrocarbon than propene because of its symmetry. The observed hysteresis is ascribed to changes in the composition of the catalyst which, being nonstoichiometric, is dependent on the surface oxygen partial pressure and hence the electrode potential. Fig. 3 shows current--overpotential relationships for the niobia-ceria electrode exposed to the three
262
I.S. Metcalfe et al. / Hydrocarbon activation in solid state cells
6
~gen/~/~pr°penex x
4 <
E 'E
~x
x methane
2-
0-
(D I,_
:3
-2x
X
-4_
X
D
-S!
-500
!
-300
)
i
i
-- 100
100
Overpotential (my)
|
i
300
Fig. 3. Current-overpotential relationships for niobia--ceria electrode at 817"C. Hydrogen ( 10% in inert) flowrate-- 155 ee/min. Propene (4% in inert) flowrate= 270 ec/min. Methane ( 10% in inert) = 240 co/rain.
atmospheres. The overpotential values were obtained using the current interrupt method as de= scribed in the experimental section. The electrode effective exchange currents were obtained from the current--overpotential relationships by fitting the data to the Butler-Volmer equation (the solid lines in fig.
3), •
.
JTF
As expected, the value in the hydrogen atmosphere (2.3 mA) was found to be the largest, followed by that of propene (1.7 mA). From the data obtained it was impossible to determine the exchange current in the case of methane as a process other than the charge transfer process appeared to be rate determining.
3.2. LiFe02 Fig. 4 shows the low temperature behaviour of the catalyst at 550"C in an atmosphere of helium. At high positive potentials the current rises due to oxygen evolution, whereas at high negative potentials the cathodic currents are due to the reduction of the oxide catalyst. The hysteresis reflects the non-equilibrium changes in the composition of the catalyst as the electrode potential is varied over a compara-
tively short time scale (typically 2 min per scan). The most significant observation is the cathodic peak at + 20 inV. This was observed after scanning first in the anodic direction, then reversing at + 400 mV and scanning back in the cathodic direction. The nature of this peak was examined by carrying out diagnostic tests. First the variation in the peak current with scan rate was investigated. Fig. 5 shows a plot of peak current versus the square root of scan rate, the linear trend being indicative of a diffusionally controlled process (we have previously verified that this test can be used to probe diffusionally controlled processes by investigating the redox behaviour of nickel oxide [ 6 ] ). The alternative explanation of a surface adsorption process can be discounted since this would require a linear dependence of peak current on scan rate. The magnitude of the currents appear to be too small for this phenomena to be due to bulk oxygen, and moreover the peak was found to disappear completely when the temperature was raised to 600°C. Thus it is suggested that we are observing the existence of an area of local reduction probably close to the surface of the oxide. The nature of the oxygen species is uncertain but it could have a significant effect on the catalytic activity of this material. The reduction of bulk lattice oxygen can explain the monotonic increase in the cathodic current at more negative potentials. It should be em-
L$. Metcalfe et al. ~Hydrocarbon activation in solid state cells
263
16
12
q)
4
k. S--
--! C.)
J
0
/
-4
-8
|
-600
-400
i
!
-200
i
200
i
i
400
600
Potential (my) Fil~ 4. Cyclic voltammogram for LiFeO2 at 560 °C. Sweep rate = 30 m V / s . Helium flowrate = 30 cc/min.
phasised that the currents are non-steady-state and therefore do not reflect the equilibrium situation. Fig. 6 shows the cyclic voltammogram at 620°C for methane and, for comparison, helium. The opencircuit potential was observed to be - 300 mV, which is considerably less than that predicted by the Nernst equation ( - 8 0 0 mV), and is due to a combination of a mixed potential and the relative inactivity of methane at this temperature. The mixed potential is
caused by the competing reactions of the surface oxygen driving the potential in the positive direction and the methane molecules reacting to drive the potential in the negative direction.
4. Condusions The behaviour of three different reactants, hydro-
60
50
o-
k. t-
30
u o q) Q.
2o
lo
o 0
2. . . . .4. . . . . 6
8
;o
'
;2
'
;4
'
;,
'
18
square root of scan rate Fig. 5. Peak current versus sweep rate for niobia--ceria electrode at 560°C. Helium flowrate = 30 cc/min.
264
I.S. Metcalfe et al.
~Hydrocarbonactivation in solid state cells
!
1086-
E
4.2-
k. ::3
fD
0-2-
~4-
I
-400
I
t
-200
0
200
4.OO
Potential (my) Fig. 6. Cyclic voltammngram for LiFeO2 at 620°C. Sweep rate= 30 mY/s. Methane flowrate = 8.1 cc/min. Helium flowrate= 69 cc/min.
gen, methane and propene, was studied over two different electrode materials, niobia-ceria and lithium ferrite. Both cyclic voltammetric work and steadystate electrode polarisation showed that the reactivity of the gases decreased in the order, hydrogen> propene> methane. Low temperature cyclic voltammetric work at 550°C on the lithium ferrite showed the presence of an oxygen state which disappeared as the temperature was increased to 600°C. The use of cyclic voltammetry to study catalytic processes in conjunction with conventional catalytic techniques will be the subject of future work. Acknowledgement PHM acknowledges the support of the Ceramic
Electrochemical Reactors Club at Imperial College. PP acknowledges the support of the State Scholarships Foundation of Greece. References [ 1 ] D.C. Fee and J.P. Ackerman, Abstract of 1983 Fuel Cell Seminar (Orlando, FL. Nov. 1983) pp. 11-14. [ 2 ] A.O. Isenberg. Abstracts of 1982 Fuel Cell Seminar (Newport Beach, CA, Nov. 1982) pp. 154-156. [ 3 ] A.L. Lee, Internal Reforming for Solid Oxide Fuel Cells, Final Report DOE/MC/22045-2364 (Febr. 1987). [4] A.L. Lee, R.F. Zabransky and W.J. Huber, Ind. Eng. Chem. Res. 29 (1990) 766. [ 5 ] B.C.H. Steele, I. Kelly, H. Middleton and R. Rudkin, Solid State Ionics 28-30 (1988) 1547. [6 ] M.E. Seiersten and P.H. Middleton, Proc. 2nd Intern. Symp. on Solid Oxide Fuel Cells, ed. P. Zegers (Athens, Greece, July 1991 ).
Solid State Ionics 57 (1992) 259-264 North-Holland
Hydrocarbon activation in solid state electrochemical cells I.S. Metc, alfe a, P.H. M i d d l e t o n m,P. Petrolekas a a n d B.C.H. Steele Department of Material~ and °Department of Chemical Engineering and Chemical Technology, Imperial Collegeof Science, Technologyand Medicine, London SW7, UK Received 21 October 1991;acceptedfor publication 30 November 1991
Hydrocarbonactivation in solid state electrochemicalcells is of interest both from a fuel cell perspective and for the development of ceramic electrochemicalreactorscapableof producingusefulchemicals. For both of these areas of research it is important to investigate and understand the behaviour of new electrode materials. Three different reactants, hydrogen,methane and propone, were studied over two different electrode materials, niobia-ceria and lithium ferrite (LiFeO2). Both cyclicvoltammetric work and steady-state electrode polarisation at 817°C showed that the reactivity of the gases decreased in the order, hydrogen> propene> methane, over the niobia-ceria electrode. Low temperature cyclicvoltammetric work at 550°C on the lithium ferrite showedthe presenceof an oxygenstate whichdisappeared as the temperature was increasedto 600°C.
1. Introduction Solid oxide fuel cells have been developed to operate on syngas fuels using nickel-zirconia anodes [1,2]. Overall efficiencies for energy production could be improved if it were possible to feed natural gas directly to the fuel cell without the need for an external reforming stage. It may be possible to use a premixed feed stream of natural gas and steam for the fuel cell with the reforming reaction taking place internally; this mode of oporation is the focus of considerable attention [3,4]. The highly endothermic nature of the reforming reaction means that there is a problem with high temperature gradients because of the cooling effect in the vicinity of the fuel inlet. Alternatively dry natural gas (or at least natural gas with a highly reduced steam content) could be used as the fuel cell feed. However, nickel-zirconia anodes are unsuitable for such a mode of operation because of high rates of carbon deposition on the nickel electrode resulting in a rapid loss of performance. At Imperial College a research programme has been initiated to investigate the direct electrochemical ox~" Paper presented at SSI-8, Lake Louise, Canada, October 2026, 1991. To whomcorrespondenceshouldbe addressed.
idation of hydrocarbons over alternative anode materials. Although methane is the primary constituent of natural gas there are also significant quantities of other hydrocarbons such as ethane and propane. These higher hydrocarbons can be of particular importance as regards carbon deposition. In addition to fuel cell applications the research involves developing the use of SOFC technology for the production of useful chemicals (this may occur simultaneously with the production of energy). In such a mode of operation, known as a ceramic electrochemical reactor, it is possible to control the supply of oxygen to the anode, where the partial oxidation process of a hydrocarbon may occur with an improved reaction selectivity. For both fuel ceU and ceramic electrochemical reactors it is therefore important to investigate the behaviour of a number of potential feed streams. In this study, feeds of hydrogen, methane and propene were selected. Hydrogen was studied because of its present importance as a fuel cell feed, methane because of interest in its direct oxidation in fuel cells and propene because it can undergo a partial oxidation resulting in the formation of acrolein (an important industrial process). Both areas of application also require the development and increased understanding of new anode
0167-2738/92/$ 05.00 © 1992ElsevierSciencePublishers B.V. All fightsreserved.
260
I.S. Metcalfe et al. / Hydrocarbon activation in solid state cells
materials. Recent work has shown ceria to be a very promising electrode material for methane oxidation [ 5 ]. However, during oxidation/reduction cycles the lattice parameter of ceria changes causing contraction and expansion leading to mechanical failure of the electrode. To overcome this problem doped ceria electrodes have been investigated. Results with niobia-ceria will be presented in this paper. Such ceria-based electrodes, whilst being of interest for the total oxidation of hydrocarbons, are not particularly suitable for partial oxidation processes and so data from a LiFeO2 electrode will also be presented. The technique of cyclic voltammetry was used to investigate the electrochemical behaviour of the electrodes in the different fuel atmospheres. Such a technique yields information regarding the polarisation behaviour of the cell and can be used to develop a mechanistic explanation of the electrode processes. Outlet gas compositions were not analysed, so it is impossible to report the selectivity to acrolein in the case of the partial oxidation of propene.
precursor powders obtained from Aldrich Chemicals Ltd. After mixing the appropriate proportions and calcining, the resultant catalyst powders were added to a tape casting slurry, in order to make the film electrodes. These were made by spreading the slurry over a pre-sintered zirconia electrolyte sheet and sintering at the appropriate temperature. The cell was completed by adding the platinum counter and reference electrodes. The cell housing was made of a machinable ceramic, and bonded together with a glass sealing compound. The total cell design constitutes a divided electrochemical cell, with fuel on one side and air on the other. Cyclic voltammetry and current interruption measurements were performed with a combination of standard instrumentation and in-house electronics; the data were recorded on xy and yt recorders. The design of the cell (see fig. 1 ) was influenced by the requirement to use prefabricated zirconia foil electrolyte sheets, and the need for rapid product analysis. The use of machinable heat resistant ceramic for the cell housing was preferred, even though this created a thermal mismatch between the material and the zirconia electrolyte. The alternative use of machinable glass (Macor) was less favoured because of its instability under reducing atmospheres
2. Experimental The oxide electrode materials were made by the solid state diffusion process, using high purity oxide
Anode Reference |
Glass s e a l / ~ " - , ~ / ~ / ~
SZ plate Cathode Reference 2
erio
"\
~,.
~7"ff"ff~[~-4-Reference I
V///////AI
*%
°e!"o
r//~. . . . . .
--
5cm
Fig. l. Schematic of the electrochemical cell.
--
j
261
LS. Metcalfe et al. ~Hydrocarbon activation in solid state cells
20-
hydr°~propene
16< E
12-
-4-8-12'
-~oo
i
-~oo'
-;oo'
-5oo' ' - ~ o o '
Potential(nnv)
-;oo
~oo' '
3oo
Fig. 2. Cyclic voltammogram of niobia--ceria electrode at 8170C. Sweep rate --30 mV/s. Hydrogen ( 10% in inert) flowrate= 155 ec/ rain. Propene (4% in inert) flowrate=270 cc/min. Methane ( 10% in inert) = 240 cc/min.
and also the problem of long term creep at temperaturesabove 800°C. The use of molten glass in the seals compensated for the thermal mismatch, the choice of glass being determined by the operating temperature range. The cavities within the cell blocks were machined to a depth of 1 mm, giving a total volume of 2 cm 3. This was considered adequate for rapid transfer of the products to the analysis section. The fuel electrode had a projected area of 9 cm 2. The gas supplies to both compartments of the cell were through narrow bore alumina tubes of I mm internal diameter. The flow rates were regulated using electronic mass flow controllers. The total inlet and outlet flows were monitored to check for leaks. The overall integrity of the seals was checked by analysis of residual air whilst purging the cell with helium. In this paper we report on the preliminary electrochemical studies from this cell design. In later publications we intend to also report on the product analysis by gas chromatography. Cyclic voltammetry was used as a convenient and rapid method for initial electrochemical investigations and, because it is a non-steady-state technique, it can reveal transient phenomena and be used to distinguish between" diffusion controlled processes and adsorption processes. In the work presented here we have used this method to give a qualitative description of system activity.
The current interruption method has been used to provide accurate IR free data for the purpose of obtaining the effective exchange current densities of the reactions.
3. Results and discussion 3.1. Ceria
Fig. 2 shows the results of cyclic voltammetric work with the niobia--eeria electrode at 817 °C with three fuels, hydrogen, propene and methane. As can be seen from the figure, higher currents are obtained in the presence of a hydrogen-containing atmosphere than in a propene-containing atmosphere, In turn, higher currents are observed in a propene-containing atmosphere than in one containing methane. These differences can be explained in terms of the reactivities of the individual molecules, hydrogen being very reactive and methane being a much more stable hydrocarbon than propene because of its symmetry. The observed hysteresis is ascribed to changes in the composition of the catalyst which, being nonstoichiometric, is dependent on the surface oxygen partial pressure and hence the electrode potential. Fig. 3 shows current--overpotential relationships for the niobia-ceria electrode exposed to the three
262
I.S. Metcalfe et al. / Hydrocarbon activation in solid state cells
6
~gen/~/~pr°penex x
4 <
E 'E
~x
x methane
2-
0-
(D I,_
:3
-2x
X
-4_
X
D
-S!
-500
!
-300
)
i
i
-- 100
100
Overpotential (my)
|
i
300
Fig. 3. Current-overpotential relationships for niobia--ceria electrode at 817"C. Hydrogen ( 10% in inert) flowrate-- 155 ee/min. Propene (4% in inert) flowrate= 270 ec/min. Methane ( 10% in inert) = 240 co/rain.
atmospheres. The overpotential values were obtained using the current interrupt method as de= scribed in the experimental section. The electrode effective exchange currents were obtained from the current--overpotential relationships by fitting the data to the Butler-Volmer equation (the solid lines in fig.
3), •
.
JTF
As expected, the value in the hydrogen atmosphere (2.3 mA) was found to be the largest, followed by that of propene (1.7 mA). From the data obtained it was impossible to determine the exchange current in the case of methane as a process other than the charge transfer process appeared to be rate determining.
3.2. LiFe02 Fig. 4 shows the low temperature behaviour of the catalyst at 550"C in an atmosphere of helium. At high positive potentials the current rises due to oxygen evolution, whereas at high negative potentials the cathodic currents are due to the reduction of the oxide catalyst. The hysteresis reflects the non-equilibrium changes in the composition of the catalyst as the electrode potential is varied over a compara-
tively short time scale (typically 2 min per scan). The most significant observation is the cathodic peak at + 20 inV. This was observed after scanning first in the anodic direction, then reversing at + 400 mV and scanning back in the cathodic direction. The nature of this peak was examined by carrying out diagnostic tests. First the variation in the peak current with scan rate was investigated. Fig. 5 shows a plot of peak current versus the square root of scan rate, the linear trend being indicative of a diffusionally controlled process (we have previously verified that this test can be used to probe diffusionally controlled processes by investigating the redox behaviour of nickel oxide [ 6 ] ). The alternative explanation of a surface adsorption process can be discounted since this would require a linear dependence of peak current on scan rate. The magnitude of the currents appear to be too small for this phenomena to be due to bulk oxygen, and moreover the peak was found to disappear completely when the temperature was raised to 600°C. Thus it is suggested that we are observing the existence of an area of local reduction probably close to the surface of the oxide. The nature of the oxygen species is uncertain but it could have a significant effect on the catalytic activity of this material. The reduction of bulk lattice oxygen can explain the monotonic increase in the cathodic current at more negative potentials. It should be em-
L$. Metcalfe et al. ~Hydrocarbon activation in solid state cells
263
16
12
q)
4
k. S--
--! C.)
J
0
/
-4
-8
|
-600
-400
i
!
-200
i
200
i
i
400
600
Potential (my) Fil~ 4. Cyclic voltammogram for LiFeO2 at 560 °C. Sweep rate = 30 m V / s . Helium flowrate = 30 cc/min.
phasised that the currents are non-steady-state and therefore do not reflect the equilibrium situation. Fig. 6 shows the cyclic voltammogram at 620°C for methane and, for comparison, helium. The opencircuit potential was observed to be - 300 mV, which is considerably less than that predicted by the Nernst equation ( - 8 0 0 mV), and is due to a combination of a mixed potential and the relative inactivity of methane at this temperature. The mixed potential is
caused by the competing reactions of the surface oxygen driving the potential in the positive direction and the methane molecules reacting to drive the potential in the negative direction.
4. Condusions The behaviour of three different reactants, hydro-
60
50
o-
k. t-
30
u o q) Q.
2o
lo
o 0
2. . . . .4. . . . . 6
8
;o
'
;2
'
;4
'
;,
'
18
square root of scan rate Fig. 5. Peak current versus sweep rate for niobia--ceria electrode at 560°C. Helium flowrate = 30 cc/min.
264
I.S. Metcalfe et al.
~Hydrocarbonactivation in solid state cells
!
1086-
E
4.2-
k. ::3
fD
0-2-
~4-
I
-400
I
t
-200
0
200
4.OO
Potential (my) Fig. 6. Cyclic voltammngram for LiFeO2 at 620°C. Sweep rate= 30 mY/s. Methane flowrate = 8.1 cc/min. Helium flowrate= 69 cc/min.
gen, methane and propene, was studied over two different electrode materials, niobia-ceria and lithium ferrite. Both cyclic voltammetric work and steadystate electrode polarisation showed that the reactivity of the gases decreased in the order, hydrogen> propene> methane. Low temperature cyclic voltammetric work at 550°C on the lithium ferrite showed the presence of an oxygen state which disappeared as the temperature was increased to 600°C. The use of cyclic voltammetry to study catalytic processes in conjunction with conventional catalytic techniques will be the subject of future work. Acknowledgement PHM acknowledges the support of the Ceramic
Electrochemical Reactors Club at Imperial College. PP acknowledges the support of the State Scholarships Foundation of Greece. References [ 1 ] D.C. Fee and J.P. Ackerman, Abstract of 1983 Fuel Cell Seminar (Orlando, FL. Nov. 1983) pp. 11-14. [ 2 ] A.O. Isenberg. Abstracts of 1982 Fuel Cell Seminar (Newport Beach, CA, Nov. 1982) pp. 154-156. [ 3 ] A.L. Lee, Internal Reforming for Solid Oxide Fuel Cells, Final Report DOE/MC/22045-2364 (Febr. 1987). [4] A.L. Lee, R.F. Zabransky and W.J. Huber, Ind. Eng. Chem. Res. 29 (1990) 766. [ 5 ] B.C.H. Steele, I. Kelly, H. Middleton and R. Rudkin, Solid State Ionics 28-30 (1988) 1547. [6 ] M.E. Seiersten and P.H. Middleton, Proc. 2nd Intern. Symp. on Solid Oxide Fuel Cells, ed. P. Zegers (Athens, Greece, July 1991 ).