F1020 Journal of The Electrochemical Society, 162 (9) F1020-F1028 (2015)

0013-4651/2015/162(9)/F1020/9/$33.00 © The Electrochemical Society

Electrode Kinetics of the Ni Porous Electrode for Hydrogen

Production in a Molten Carbonate Electrolysis Cell (MCEC)
Lan Hu,z Göran Lindbergh, and Carina Lagergren∗
Applied Electrochemistry, Department of Chemical Engineering and Technology, KTH Royal Institute of Technology,
SE-100 44 Stockholm, Sweden

The purpose of this study was to elucidate the kinetics of a porous nickel electrode for hydrogen production in a molten carbonate
electrolysis cell. Stationary polarization data for the Ni electrode were recorded under varying gas compositions and temperatures.
The slopes of these iR-corrected polarization curves were analyzed at low overpotential, under the assumption that the porous
electrode was under kinetic control with mass-transfer limitations thus neglected. The exchange current densities were calculated
numerically by using a simplified porous electrode model. Within the temperature range of 600–650◦ C, the reaction order of hydrogen
is not constant; the value was found to be 0.49–0.44 at lower H2 concentration, while increasing to 0.79–0.94 when containing
25–50% H2 . The dependence on CO2 partial pressure increased from 0.62 to 0.86 with temperature. The reaction order of water
showed two cases as did hydrogen. For lower H2 O content (10–30%), the value was in the range of 0.47–0.67 at 600–650◦ C, while
increasing to 0.83–1.07 with 30–50% H2 O. The experimentally obtained partial pressure dependencies were high, and therefore not
in agreement with any of the mechanisms suggested for hydrogen production in molten carbonate salts in this study.
© 2015 The Electrochemical Society. [DOI: 10.1149/2.0491509jes] All rights reserved.

Manuscript submitted April 21, 2015; revised manuscript received June 8, 2015. Published July 1, 2015. This was Paper 1530
presented at the Chicago, Illinois, Meeting of the Society, May 24–28, 2015.

Electrolysis in molten carbonate salts at high temperatures is a see Table I. Ang and Sammells (AS)7 , and Jewulski and Suski (JS)8
promising method for hydrogen and/or syngas (H2 +CO) produc- suggested that the intermediate species of the charge-transfer step was
tion, especially when combining it with renewable electricity re- the hydroxyl ion (OH– ). Both these two mechanisms show identical
sources such as solar energy and wind power. Due to the favorable reaction orders (0.25) for hydrogen, carbon dioxide and water. Lu and
thermodynamic and kinetic conditions, high-temperature electrolysis Selman9 agreed with the AS mechanism, and they also noted that the
will attain higher overall efficiency and require lower applied volt- hydrolysis equilibrium reaction played an important role in carbonate
age when compared to low-temperature electrolysis.1 Electrolysis in melts. Nishina et al.10 proposed a modification of the JS mechanism,
molten carbonates has been evidenced by some authors, mainly by including a split up of the rate-determining step. The intermediate
converting CO2 into CO.2–5 Peelen et al.2 studied the electrochemical species M-HCO3 – was mentioned as plausibly rate-determining. But
reduction of CO2 into CO on a gold flag electrode in 62/38 mol% Weewer et al.12 showed that the rate-determining step was the first
Li/K carbonate mixture in the temperature range 575–700◦ C. Kaplan step of producing absorbed hydrogen and hydroxyl ion (OH– ) for Au
et al.3–4 observed the conversion of CO2 to CO by using a cell with a and was the final chemical reaction of OH– and CO2 for Ni. This gave
molten electrolyte consisting of lithium carbonate and lithium oxide, a higher dependence on the hydrogen partial pressure. The Ni porous
in which the electrodes were graphite (anode) and titanium (cathode), electrode was also experimentally studied for hydrogen oxidation,11,13
and the working temperature was 850–900◦ C. Chery et al.5 presented but the reaction mechanism is still not well defined.
thermodynamic calculations and experimental measurements on the So far, the kinetics and reaction mechanism on a porous Ni elec-
reduction of CO2 into CO on a gold flag or planar disk electrode trode as cathode for hydrogen and/or carbon monoxide production in a
with Li/K and Li/Na carbonate eutectics at temperatures from 575 molten carbonate electrolysis cell are not known. The overall electro-
to 650◦ C. However, most experiments were carried out on flag elec- chemical reaction on a Ni porous electrode for hydrogen production
trodes, which are different from porous electrodes when considering in MCEC mode is
electrode surface and mass-transfer limitations. In our previous study6
a cell operating at 650◦ C, with conventional molten carbonate fuel cell C O2 + H2 O + 2e− 
 H2 + C O3
2−
[1]
(MCFC) materials (Ni-based porous electrode and molten carbonate
electrolyte), was investigated also for electrolysis. The cell was found Although the kinetics of CO2 electrolysis generating CO is probably
to give lower polarization losses in MCEC mode than in MCFC mode, slower than electrolysis of H2 O producing hydrogen on a Ni-based
mainly due to the NiO electrode performing much better as anode in electrode,14 there is still the possibility of CO production through the
the electrolysis cell. Reversing the electrochemical cell will provide following reaction
new opportunities. The possibility of operating the molten carbon-
ate cell alternatingly as a fuel cell to produce electricity and as an 2C O2 + 2e− 
 C O + C O3
2−
[2]
electrolyzer to generate fuel gas may increase the economic benefits.
However, in the previous work,6 slightly higher polarization losses In this paper Reaction 1 is assumed to be the main electrochemical
on the Ni electrode were found for water electrolysis compared to cathodic reaction and Reaction 2 will therefore not be included when
hydrogen oxidation when using a fuel gas consisting of 64/16/20% analyzing the experimental data. The general procedure for elucidating
H2 /CO2 /H2 O. When the CO2 content was increased to 40%, the po- the mechanism of Reaction 1 is to determine the dependencies of the
larization resistance gave a similar or even a little lower value in the exchange current densities on the hydrogen, carbon dioxide and water
electrolysis cell than in fuel cell mode. To learn more about the use partial pressures.
of the cell as an electrolyzer it is of vital importance to elucidate the The aim of this paper is to determine the reaction orders of hy-
kinetics and reaction mechanism on the Ni porous electrode for hy- drogen, carbon dioxide and water. Using these as starting point the
drogen production, as well as to further study the disparity between kinetics and reaction mechanism on a Ni porous electrode for pro-
hydrogen oxidation and water electrolysis (hydrogen production) un- ducing hydrogen in MCEC mode may be proposed. For this purpose
der molten carbonate conditions. Several studies7–13 have been made a large amount of polarization data was investigated, concerning the
on the kinetics of hydrogen oxidation in a molten carbonate fuel cell, effect of gas composition and temperature on steady-state electrode
performance. From the slopes of the iR-corrected polarization curves
at low overpotential, the corresponding exchange current densities
∗
Electrochemical Society Active Member. were evaluated. Based on that, the partial pressure dependencies of
z
E-mail: [email protected] hydrogen, carbon dioxide and water could be determined.

Downloaded on 2016-04-08 to IP 130.237.74.37 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).
 Journal of The Electrochemical Society, 162 (9) F1020-F1028 (2015) F1021

Table I. The kinetics and reaction mechanisms for hydrogen oxidation in MCFC reported in the literature.7–13 i 0 = i 00 ( p H2 )γ1 ( pC O2 )γ2 ( p H2 O )γ3 ,
γ1 , γ2 and γ3 denote the electrochemical reaction orders of hydrogen, carbon dioxide and water. Experimental techniques are CV = cyclic
voltammetry, EIS = electrochemical impedance spectroscopy, CC = chronocoulometry, CA = chronoamperometry.

References Year Mechanisms and reaction orders Electrode Techniques

Ang and Sammells (AS)7 1980 H2 + 2M 
 2M − H Ni, Co wire Polarization
M − H + C O32−  O H − + C O2 + M + e− (r ds)
M − H + O H−   H2 O + M + e−
γ1 = 0.25, γ2 = 0.25, γ3 = 0.25
Jewulski and Suski (JS)8 1984 H2 + 2M 
 2M − H Modeling
2{M − H + C O32−  O H − + C O2 + M + e− } (r ds)
2O H − + C O2 
 H2 O + C O3
2−

γ1 = 0.25, γ2 = 0.25, γ3 = 0.25

Lu and Selman9 1984 H2 + 2M 
 2M − H Cu Polarization
M − H + C O32−  O H − + C O2 + M + e− (r ds)
M − H + O H−   H2 O + M + e−
H2 O + C O32− 
 2O H − + C O2
γ1 = 0.33, γ2 = 0.14, γ3 = 0.33
Nishina, Takahashi and Uchida10 1990 H2 + 2M 
 2M − H Ni, Pt flag CV, EIS, polarization, CC
M − H + C O32− 
 M − H C O3
2−
−
M − H C O32−  M − H C O3 + e− (r ds)
M − H C O3−  O H − + C O2 + M
2O H − + C O2 
 H2 O + C O3
2−

γ1 = 0.25, γ2 = 0.25, γ3 = 0.25

Yuh and Selman11 1991 γ1 = 0.84, γ2 = 0.36, γ3 = −0.42 Ni porous EIS
Weewer, Hemmes, and Wit12 1995 H2 + C O32− 
 Hads + O H − + C O2 + e− (rds) Ni, Au flag CA, EIS
Hads + C O32− 
 O H − + C O2 + e −
2O H + C O2 
−
 H2 O + C O3
2−

γ1 = 0.7 − 1.0, γ2 = 0.3 − 0.6, γ3 = −0.25 − 0

Lindbergh, Olivry and Sparr13 2000 γ1 = 0.5, γ2 = 0.5, γ3 = 0.7 Ni porous Polarization

Experimental Fuel Gases and Their Equilibrium Conditions

2
Experimental data were obtained from a 3 cm laboratory cell The inlet fuel gases were varied separately within three different
unit, where the cell components were provided by Ansaldo Fuel groups, see Table II. The concentration increment was 5% when in-
Cells, Italy. The cell has a porous Ni-Cr alloy as hydrogen elec- creasing the amount of hydrogen (5–50%), carbon dioxide (10–50%)
trode (cathode), and a porous Ni, oxidized and lithiated in situ, as and water (10–50%), respectively.
oxygen electrode (anode). The electrodes are separated by a porous Due to different reactions possibly taking place at the Ni electrode,
LiAlO2 matrix, which also supports the electrolyte, a eutectic mixture an essential problem is to know the actual gas composition inside the
of 62/38 mol% Li2 CO3 /K2 CO3 . The current collectors on the cath- cell when determining the partial pressure dependencies. One such
ode and anode sides are made of nickel and stainless steel SS316, reaction is the reversed water-gas shift reaction
respectively. Each current collector has two probes, one for measur-
ing the current and one for the potential measurement. The reference H2 + C O2 
 H2 O + C O [3]
electrode consists of gold wires in equilibrium with a gas mixture
containing 33/67% O2 /CO2 . Two identical reference electrodes are At higher temperatures (above 500◦ C), and in the presence of a nickel
placed in separate chambers filled with the same electrolyte as in the catalyst, the shift reaction will very rapidly establish equilibrium of
cell. These chambers are connected to the cell through a capillary with these gases.15 Therefore Reaction 3 will affect the gas compositions
a gold plug. The laboratory cell set-up used in this study is the same inside the cell and then probably influence the performance of the Ni
as that used in the previous study, and a schematic drawing is found electrode. The equilibrium constant
in Ref. 6. PH2 O · PC O
The cell was operated at three different temperatures, 600, 625 and Kp = [4]
650◦ C. A gas consisting of 15/30/55% O2 /CO2 /N2 was used for the PH2 · PC O2
NiO oxygen electrode. Different fuel gas mixtures of H2 /CO2 /H2 O/N2 can be found and calculated in the literature for 600◦ C (K p = 0.373)
for the Ni electrode and their equilibrium conditions are described in and 650◦ C (K p = 0.510),9,11 while being interpolated for the value
the following section. The concentration of H2 O in the fuel gases at 625◦ C (K p = 0.442). The fuel gas compositions calculated at shift
is controlled by the temperature of the water in the humidifier. In equilibrium at 600–650◦ C are shown in Table III.
order to reduce the effect of mass-transport limitations in the gaseous
phase,13 the flow rate of inlet gases was held high, at approximately
250 ml · min−1 for the Ni electrode. The gas flow rate for the NiO
electrode was also about 250 ml · min−1 , while the reference gas had Table II. Inlet fuel gas compositions using an increment of
a flow rate of 20 ml · min−1 . All the gases used in the experiments are 5% when changing the concentrations of H2 , CO2 , and H2 O,
certified gas mixtures from AGA Gas AB, Sweden. respectively.
The polarization curves for the Ni electrode were measured using
a Solartron Interface SI1287 supported by CorrWare software. The Gas Group No. H2 /% CO2 /% H2 O/% N2 /%
current-interrupt method was used to correct for the ohmic potential 1 5–50 25 25 Balance
drop. The negative current density refers to the electrolysis mode, 2 25 10–50 25 Balance
shown in the figures below. 3 25 25 10–50 Balance

Downloaded on 2016-04-08 to IP 130.237.74.37 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).
 F1022 Journal of The Electrochemical Society, 162 (9) F1020-F1028 (2015)

Table III. Gas compositions calculated at shift equilibrium at 600–650◦ C, in percent (%).

600◦ C 625◦ C 650◦ C

No. H2 CO2 H2 O CO H2 CO2 H2 O CO H2 CO2 H2 O CO N2

Group 1: An increase of H2 (5–50%) in the inlet fuel gases
1.1 3.74 23.74 26.26 1.26 3.59 23.59 26.41 1.41 3.45 23.45 26.55 1.55 45
1.2 7.64 22.64 27.36 2.36 7.37 22.37 27.63 2.63 7.12 22.12 27.88 2.88 40
1.3 11.67 21.67 28.33 3.33 11.30 21.30 28.70 3.70 10.96 20.96 29.04 4.04 35
1.4 15.80 20.80 29.20 4.20 15.35 20.35 29.65 4.65 14.94 19.94 30.06 5.06 30
1.5∗ 20.02 20.02 29.98 4.98 19.49 19.49 30.51 5.51 19.03 19.03 30.97 5.97 25
1.6 24.30 19.30 30.70 5.70 23.72 18.72 31.28 6.28 23.22 18.22 31.78 6.78 20
1.7 28.65 18.65 31.35 6.35 28.02 18.02 31.98 6.98 27.47 17.47 32.53 7.53 15
1.8 33.04 18.04 31.96 6.96 32.38 17.38 32.62 7.62 31.80 16.80 33.20 8.20 10
1.9 37.48 17.48 32.52 7.52 36.78 16.78 33.22 8.22 36.18 16.18 33.82 8.82 5
1.10 41.96 16.96 33.04 8.04 41.24 16.24 33.76 8.76 40.60 15.60 34.40 9.40 0
Group 2: An increase of CO2 (10–50%) in the inlet fuel gases
2.1 22.64 7.64 27.36 2.36 22.37 7.37 27.63 2.63 22.12 7.12 27.88 2.88 40
2.2 21.67 11.67 28.33 3.33 21.30 11.30 28.70 3.70 20.96 10.96 29.04 4.04 35
2.3 20.80 15.80 29.20 4.20 20.35 15.35 29.65 4.65 19.94 14.94 30.06 5.06 30
2.4∗ 20.02 20.02 29.98 4.98 19.49 19.49 30.51 5.51 19.03 19.03 30.97 5.97 25
2.5 19.30 24.30 30.70 5.70 18.72 23.72 31.28 6.28 18.22 23.22 31.78 6.78 20
2.6 18.65 28.65 31.35 6.35 18.02 28.02 31.98 6.98 17.47 27.47 32.53 7.53 15
2.7 18.04 33.04 31.96 6.96 17.38 32.38 32.62 7.62 16.80 31.80 33.20 8.20 10
2.8 17.48 37.48 32.52 7.52 16.78 36.78 33.22 8.22 16.18 36.18 33.82 8.82 5
2.9 16.96 41.96 33.04 8.04 16.24 41.24 33.76 8.76 15.60 40.60 34.40 9.40 0
Croup 3: An increase of H2 O (10–50%) in the inlet fuel gases
3.1 17.95 17.95 17.05 7.05 17.40 17.40 17.60 7.60 16.92 16.92 18.08 8.08 40
3.2 18.79 18.79 21.21 6.21 18.24 18.24 21.76 6.76 17.76 17.76 22.24 7.24 35
3.3 19.47 19.47 25.53 5.53 18.93 18.93 26.07 6.07 18.46 18.46 26.54 6.54 30
3.4∗ 20.02 20.02 29.98 4.98 19.49 19.49 30.51 5.51 19.03 19.03 30.97 5.97 25
3.5 20.47 20.47 34.53 4.53 19.97 19.97 35.03 5.03 19.52 19.52 35.48 5.48 20
3.6 20.86 20.86 39.14 4.14 20.37 20.37 39.63 4.63 19.94 19.94 40.06 5.06 15
3.7 21.18 21.18 43.82 3.82 20.72 20.72 44.28 4.28 20.30 20.30 44.70 4.70 10
3.8 21.46 21.46 48.54 3.54 21.02 21.02 48.98 3.98 20.61 20.61 49.39 4.39 5
3.9 21.70 21.70 53.30 3.30 21.28 21.28 53.72 3.72 20.89 20.89 54.11 4.11 0

∗ The standard inlet gas, 25/25/25/25% H2 /CO2 /H2 O/N2 .

Theory ductivities can be calculated for 600 and 625◦ C, using the Arrhenius
equation,18 see Table IV. It is assumed that the electrode reaction
When analyzing kinetic data for a porous electrode, it is important
occurs mainly on the exterior surface area of the agglomerates.13 By
to consider the current distribution along the depth of the electrode.16
assuming that the agglomerates are spherical and have a radius of
The current distribution is determined by the combination of the elec-
approximately 4.5 μm, the exterior agglomerate surface area, S, is
trode intrinsic activity, its specific surface area, the effective conduc-
estimated to be 2.7 · 105 m−1 .19 Furthermore, the electrochemically
tivities of the electrolyte and electrode, and the effective diffusion
active area is assumed to be constant for all gas compositions, and
path lengths of electroactive species. The exchange current density
will therefore only affect the value of the exchange current density
i 0 , varying with the concentration of the electroactive species and
and not the partial pressure dependencies.13 All parameter values
temperature, is determined by the reaction mechanism. In this study,
of the electrode and electrolyte used for solving Eq. 5 are listed in
exchange current densities are obtained by analyzing iR-corrected
Table IV.
polarization curves at low overpotential. In this region the relation be-
The dependencies of the exchange current density on hydrogen,
tween overpotential and current density is linear and it is assumed that
carbon dioxide and water partial pressures can be calculated from the
the electrode is mainly under kinetic control and that mass-transfer
following equation
limitations may be neglected.
Lagergren and Simonsson16,17 derived the expression for the total  a  b  c
polarization resistance for a porous electrode when taking the current i 0 = i 00 p H2 pC O2 p H2 O [6]
distribution into account. For the porous Ni electrode, the effective
The above equation can also be written in logarithmic form
conductivity of the solid material is high (about 105 S · m−1 ) and
does not contribute to the polarization resistance, leaving a simplified        
log (i 0 ) = log i 00 + a log p H2 + b log pC O2 + c log p H2 O
expression (see Ref. 13): [7]

dη RT
= [5]
di n F Si 0 κ Table IV. The parameter values of the electrode and electrolyte
used for Eq. 11.
where κ is the effective conductivity in the electrolyte phase. S is the
exterior surface area of the agglomerates, where the dominating part
600◦ C 625◦ C 650◦ C
of the electrode reaction takes place.
The porosity of the electrode is determined to be 53.7%, and S 2.7 · 105 m−1
the effective conductivity in the pore electrolyte is assumed to be n 2
5.0 S · m−1 at 650◦ C.13 Based on that, the effective electrolyte con- κ 4.1 S · m−1 4.6 S · m−1 5.0 S · m−1

Downloaded on 2016-04-08 to IP 130.237.74.37 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).
 Journal of The Electrochemical Society, 162 (9) F1020-F1028 (2015) F1023

Figure 1. iR-corrected polarization curves for the Ni electrode in electrolysis cell at 650◦ C with varied gas compositions. (a) H2 , (b) CO2 , and (c) H2 O.

where i 00 is the standard exchange current density, and a, b and c increases strongly from 600 to 625◦ C, but much less improvement is
denote the reaction orders of hydrogen, carbon dioxide and water, found when going from 625 to 650◦ C.
respectively. From Figs. 1 and 2, it shows that all these iR-corrected curves are
approximately linear in the low overpotential range, <40 mV. It is
assumed that the electrode in this range is mainly under kinetic con-
Results and Discussion trol, where mass-transfer limitations are negligible. The polarization
In order to obtain the kinetic data for the hydrogen-production re- resistances, dη
di
, are obtained from the slopes of these curves by linear
action, the electrochemical performance of the Ni electrode in MCEC
mode is investigated. Fig. 1 shows the iR-corrected polarization curves
for different gas compositions at 650◦ C. The blue curves in all the
figures correspond to the performance of the Ni electrode with the
standard inlet gas containing 25/25/25/25% H2 /CO2 /H2 O/N2 . The red
and green symbols refer to the lower and higher contents, respec-
tively, compared to the standard concentration of 25% when changing
H2 , CO2 and H2 O partial pressures separately. When increasing the
concentration of H2 in the inlet fuel gases, the polarization losses of
the Ni electrode decrease gradually. When containing close to 50%
H2 , no improvement or only a small improvement of the electrode
performance is shown, see Fig. 1a. At lower H2 content the down-
ward bending tendency of the polarization curves can be observed
at high current density, corresponding to mass-transfer limitations.
The bending seems to take place more obviously at lower CO2 and
H2 O concentrations, Figs. 1b and 1c (red symbols), respectively. This
makes sense since carbon dioxide and water are reactants in the elec-
trolysis cell. An increase of CO2 content improves the Ni electrode for
hydrogen production in the electrolyzer operation. Similarly, the elec-
trode polarization decreases when the water partial pressure is raised.
The effect of cell temperature on the polarization of the Ni electrode Figure 2. iR-corrected polarization curves for the Ni electrode in electrolysis
in the electrolysis cell is also investigated, showing an example for cell with the gas of 25/25/25/25% H2 /CO2 /H2 O/N2 at operating temperatures
the standard inlet gas in Fig. 2. The performance of the Ni electrode from 600 to 650◦ C.

Downloaded on 2016-04-08 to IP 130.237.74.37 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).
 F1024 Journal of The Electrochemical Society, 162 (9) F1020-F1028 (2015)

Figure 3. Exchange current densities as function of temperature with varied gas compositions. (a) H2 , (b) CO2 , and (c) H2 O.

regression analysis. The effect of gas composition on the electrolyte On the basis of exchange current densities obtained above, the
distribution in the electrode is not considered in the evaluation of these partial pressure dependencies of hydrogen, carbon dioxide, and wa-
experimental data, even though the electrolyte wetting properties of ter could be determined. Sparr et al.21 evidenced that the reversed
the Ni electrode strongly depend on gas composition.20 water-gas shift reaction is not in equilibrium for small cells in molten
The exchange current densities i 0 can be determined by solving carbonate salts, in spite of the fact that the reaction is assumed to
Eq. 5, using the parameter values listed in Table IV. The exchange establish equilibrium rapidly.15 Due to the uncertainty of the gas
current densities as function of temperature for different inlet fuel compositions reaching equilibrium or not in this study, two cases
gases are presented in Fig. 3. The exchange current densities for the with different gas compositions are investigated: (a) no shift re-
standard feeding gas, 25/25/25/25% H2 /CO2 /H2 O/N2 , are calculated action occurs (using the inlet gas compositions in Table II) and
to be in the ranges of 80–85, 115–125, and 130–150 A · m−2 at 600, (b) complete equilibrium is established due to the shift reaction (gas
625 and 650◦ C, respectively. The exchange current densities obtained compositions calculated in Table III). Both these cases are presented
in electrolysis operation are higher than the values reported for the in Figs. 4–6. The reaction orders can be determined from the slopes
porous Ni electrode operating in fuel cell mode.13 Due to the various by fitting the parameters a, b, and c in Eq. 7. And the results from
measuring techniques, calculating methods and operating conditions, the fitting are presented in Tables V and VI for the cases of the
it is difficult to directly compare the kinetic data reported by different inlet gas compositions and complete equilibrium gas compositions,
researchers. respectively.
When the temperature rises from 600 to 625◦ C, the exchange In Fig. 4a, the effect of hydrogen partial pressure on the exchange
current densities increase significantly for these three different groups current density is shown, and it can be seen that the result is not a
of gas compositions. But when going from 625 to 650◦ C the increase straight line. In this case the hydrogen partial pressure dependency
is much less. At 600◦ C the exchange current densities are in the range could be evaluated in two ways. At first, the obtained data points
of 40–140 A · m−2 under all measured gas compositions. In Fig. 3a are fitted to straight lines by means of linear regression (solid lines),
and 3b the respective increases of the H2 and CO2 concentration give resulting in values of 0.57–0.59 for hydrogen at 600–650◦ C. Secondly,
exchange current densities approximately in the interval of 60–220 the data points are fitted to curves in a more accurate way, resulting in
A · m−2 at 625◦ C, while these values lie in a slightly higher range, different dependencies for different hydrogen concentration intervals
80–245 A · m−2 , when increasing water content from 10 to 50%, seen (dashed lines). In Tables V and VI, the superscripts marked 1 and
in Fig. 3c. A similar tendency is seen at the operating temperature of 2 refer to the low and high concentration, respectively. The reaction
650◦ C. The ranges of exchange current densities are 65–220, 65–265, order of hydrogen is in the range of 0.49–0.44 (a 1 ) when containing
and 85–300 A · m−2 for increasing H2 , CO2 , and H2 O concentrations 5–25% H2 in the inlet gases. Then the partial pressure dependency
in the inlet gas compositions, respectively. increases to the interval of 0.79–0.94 (a 2 ) within the H2 content range

Downloaded on 2016-04-08 to IP 130.237.74.37 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).
 Journal of The Electrochemical Society, 162 (9) F1020-F1028 (2015) F1025

Figure 4. Effect of hydrogen partial pressure on the exchange current density at 600–650◦ C. (a) p(H2 ) from the inlet gas compositions, (b) p(H2 ) from the gas
compositions calculated at shift equilibrium. The vertical dotted line marks the point for change of slopes.

Figure 5. Effect of carbon dioxide partial pressure on the exchange current density at 600–650◦ C. (a) p(CO2 ) from the inlet gas compositions, (b) p(CO2 ) from
the gas compositions calculated at shift equilibrium.

Figure 6. Effect of water partial pressure on the exchange current density at 600–650◦ C. (a) p(H2 O) from the inlet gas compositions, (b) p(H2 O) from the gas
compositions calculated at shift equilibrium. The vertical dotted line marks the point for change of slopes.

Downloaded on 2016-04-08 to IP 130.237.74.37 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).
 F1026 Journal of The Electrochemical Society, 162 (9) F1020-F1028 (2015)

Table V. The dependency of exchange current densities on the partial pressure of the inlet fuel gas compositions i 0 = i 00 ( p H2 )a ( pC O2 )b ( p H2 O )c .
The superscripts 1 and 2 refer to the low and high concentration.

a c

Temperature/◦ C a a 1 (5–25% H2 ) a 2 (25–50% H2 ) b c a 1 (10–30% H2 O) a 2 (30–50% H2 O)

600 0.57 0.49 0.79 0.62 0.59 0.47 0.83
625 0.59 0.44 0.94 0.79 0.66 0.51 1.07
650 0.57 0.44 0.94 0.86 0.78 0.67 1.07

of 25–50%. Regardless of the way of evaluating data, the reaction for hydrogen production than does carbon dioxide and hydrogen. The
order of hydrogen does not strongly depend on the temperature in any reaction mechanisms for hydrogen production in molten carbonate
of the concentration intervals. In Fig. 4b the water-gas shift reaction is salts, following the reverse steps of hydrogen oxidation in the fuel
taken into account when calculating the gas compositions. However, cell, are suggested and discussed below.
similar dependencies of hydrogen partial pressure as in Fig. 4a are When considering the hydrogen oxidation mechanism proposed
found (listed in Table VI). Thus, the reversed water-gas shift reaction by Ang and Sammells7 and Jewulski and Suski,8 the processes
seems to have small effect, if any, on the hydrogen partial pressure of hydrogen production in electrolysis operation are expressed as
dependency. Reactions 8–10 and Reactions 11, 12 and 10, respectively. The method
Unlike the case of hydrogen, the reaction order of carbon dioxide for derivation of the theoretical reaction orders is given in Refs. 7 and
is dependent on the temperature, see Fig. 5a. The values are 0.62, 22. The symmetry factor, as generally found, is assumed to be β = 0.5
0.79 and 0.86 at 600, 625 and 650◦ C, respectively. But, much like the for the rate-determining step. The theoretical partial pressure depen-
case of hydrogen, a minor impact of the shift reaction on the carbon dencies are all derived as 0.25 for hydrogen, carbon dioxide and water,
dioxide partial pressure dependency is found in Fig. 5b. For the gases if the rate-determining step is Reaction 9 or Reaction 12, respectively,
at complete shift equilibrium, the reaction order increases from 0.59 and a low coverage of M H is assumed (M denotes metal). The last
to 0.80 with temperature. step for both reaction mechanisms is the chemical hydrogen evolution
In Fig. 6a the dependence of the exchange current density on the reaction. The theoretical values based on these two mechanisms are
partial pressure of water is analyzed. The evaluation is done in two not consistent with our experimental data.
ways as discussed for hydrogen above. For the first case (solid lines),
H2 O + M + e− 
 MH + OH
−
[8]
the reaction order of water increases from 0.59 to 0.78 at 600–650◦ C.
In the second case (dashed lines), the dependence values are in the
range of 0.47–0.67 (c1 ) at the H2 O concentration of 10–30%, while O H − + C O2 + M + e − 
 M H + C O3
2−
[9]
the reaction order increases to the interval of 0.83–1.07 (c2 ) at 30–
50% H2 O at these operating temperatures. In contrast to hydrogen and 2M H 
 H2 + 2M [10]
carbon dioxide, the shift reaction has an important influence on the
reaction order of water in electrolysis operation. The exchange current H2 O + C O32−  −
 2O H + C O2 [11]
densities as function of water partial pressure after equilibrium, in
Fig. 6b, more or less follow a straight line for each temperature.  
2 O H − + C O2 + M + e − 
 M H + C O3
2−
[12]
The values of reaction order for water are given by the slopes and
are 0.82, 0.94 and 1.13 at 600, 625 and 650◦ C, respectively. These
data approach the water partial pressure dependency with higher H2 O 2M H 
 H2 +2M [10]
concentration (30–50%) in the inlet fuel gases. Lu and Selman9 confirmed the AS mechanism and indicated the im-
If the reaction pathways of hydrogen production are assumed to portance of the hydrolysis equilibrium. The hydrolysis constant of
take place exactly as the reverse process of hydrogen oxidation in Reaction 11 is reported to be Kh = 1.79 · 10−4 for 650◦ C.9 On the
molten carbonate salts, the theoretical partial pressure dependencies basis of that, the hydrogen production reaction can be written as
of hydrogen, carbon dioxide and water would be the same in both follows:
operating modes. When compared to reaction orders for hydrogen
oxidation mentioned in Table I, the experimentally obtained data for H2 O + C O32−  −
 2O H + C O2 [11]
water electrolysis are still higher. But the partial pressure dependen-
cies evaluated in this study are rather close to the result of Lindbergh H2 O + M + e− 
 MH + OH
−
[8]
et al.,13 0.5 for hydrogen and carbon dioxide, and 0.7 for water in fuel
cell operation. Since similar porous electrode materials were used and O H − + C O2 + M + e − 
 M H + C O3
2−
[9]
analyzed by polarization data in both studies, it implies that hydrogen
production in MCEC mode may be the reversed reaction of hydrogen 2M H 
 H2 + 2M [10]
oxidation in fuel cell operation. When taking the water-gas shift re-
−
action into account, it is found that water shows higher reaction order If Reaction 8 occurs rather slowly, the OH ions provided for
Reaction 9 will be from the hydrolysis Reaction 11. In that case
the electroreduction of water with hydrogen adsorption 8 becomes the
Table VI. The dependency of exchange current densities on rate-determining step, a high reaction order of water is obtained, 2−β2
,
the partial pressure of the gas compositions calculated at shift or about 0.75, and the values of 0.25 and −0.25 are theoretically ob-
equilibrium i 0 = i 00 ( p H2 )a ( pC O2 )b ( p H2 O )c . The superscripts 1 and tained for hydrogen and carbon dioxide, respectively. The theoretical
2 refer to the low and high concentration. water partial pressure dependency seems close to the experimental
data (0.82–1.13). But the derived negative dependence (−0.25) for
a carbon dioxide does not show agreement with the positive one ob-
tained from the experimental results. If assuming Reaction 9 as the
Temperature/◦ C a a1 (3.5–20% H2 ) a 2 (20–42% H2 ) b c rate-determining step, all the reaction orders are β2 or about 0.25. Un-
600 0.55 0.47 0.74 0.59 0.82 der the assumption that Reactions 8 and 9 take place simultaneously
625 0.56 0.42 0.87 0.74 0.94 rather than as consecutive steps of a simple reaction, the possible
650 0.53 0.41 0.86 0.80 1.13 conclusion is that the reaction order for hydrogen is still 0.25, for

Downloaded on 2016-04-08 to IP 130.237.74.37 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).
 Journal of The Electrochemical Society, 162 (9) F1020-F1028 (2015) F1027

carbon dioxide somewhere between −0.25 and 0.25, and for water activation energy does not strongly depend on the hydrogen concen-
between 0.75 and 0.25. But these value ranges are all lower than the tration. When increasing the carbon dioxide and water from 10 to
data determined experimentally. 50% individually, the activation energy shows an increase, now rang-
Next, the mechanism of hydrogen production is based on the re- ing from 55 to 100 kJ · mol−1 for these two groups of gas compositions.
versed reaction of hydrogen oxidation suggested by Weewer et al.,12 The activation energy could be a clue to determining the rate-limiting
expressed as Reactions 11, 9 and 13. process in the fuel cell,24 and may also be used when evaluating the
Ni electrode as cathode in the molten carbonate electrolysis cell. The
H2 O + C O32−  −
 2O H + C O2 [11]
activation energy obtained in the present study indicates that the Ni
O H + C O2 + M + e 
−
 MH +
−
C O32− [9] electrode is under kinetic or mixed control. This differs from the rate-
limiting process in fuel cell mode, in which the porous Ni electrode
M H + O H − + C O2 + e −  anode is usually under mass-transfer limitation.11
 H2 + C O3
2−
[13]
12
Weewer et al. discussed that the Reaction 11 might be the rate- Conclusions
determining step for hydrogen oxidation on a Ni flag electrode, due
to the high reaction order of 1.5 found for H2 O. If Reaction 11 is also The polarization data for the Ni porous electrode in a molten car-
the rate-determining step for hydrogen production on the Ni porous bonate electrolysis cell were investigated under different gas compo-
electrode, the theoretical reaction orders will be zero for hydrogen sitions at three different temperatures. The exchange current densities
and carbon dioxide, and 1 for water. A possibility is that a chemical were evaluated numerically from the slopes of the polarization curves
reaction or the combination of a chemical reaction and electrochemical at low overpotential, assuming that the electrode was under kinetic
reactions is the rate-determining step in the porous electrode, due control.
to the slow rate of the chemical reaction. But in this case it does The partial pressure dependencies for hydrogen, carbon dioxide
not make sense concerning the dependency values of hydrogen and and water were determined in two cases, no water-gas shift reaction
carbon dioxide, i.e. no dependence occurs in the hydrogen production and complete shift equilibrium. The reaction order of hydrogen was
process. If Reaction 9 is assumed to be the rate-determining step, found in the range of 0.49–0.44 at low H2 concentration from 600 to
as discussed above, identical reaction orders of 0.25 apply for H2 , 650◦ C, changing to the interval 0.79–0.94 when containing 25–50%
CO2 and H2 O. Assuming Reaction 13 to be rate-determining, a high H2 in the inlet gas. The partial pressure dependency of carbon dioxide
reaction order of 0.75 is found for hydrogen, and of 0.25 for both increased from 0.62 to 0.86 with the increase of temperature. The
carbon dioxide and water. reversed water-gas shift reaction has a minor or almost no impact on
From the experimentally obtained reaction orders for hydrogen, the electrochemical reaction order of hydrogen and carbon dioxide.
carbon dioxide and water, regardless of shift equilibrium or not, they The reaction order of water was analyzed in two ways, just as for
do not reasonably satisfy any of the four mechanisms suggested above hydrogen. At lower H2 O concentration, 10–30% in the inlet gases,
for hydrogen production on a Ni porous electrode. Based on the re- the reaction order was in the range of 0.47–0.67 while it increased to
verse pathways of hydrogen oxidation, these mechanisms concern 0.83–1.07 for 30–50% H2 O at 600–650◦ C. The shift reaction plays a
flag electrodes. Thus different characteristics of the surface area on a major role in the water partial pressure dependency. The reaction order
porous electrode may affect the electrode kinetics and reaction mech- was found to be within the range of 0.82–1.13 at shift equilibrium in
anism. Additionally, in this study the polarization analysis is based on the temperature interval 600–650◦ C, approaching the results at higher
the assumption that the Ni electrode is mainly under kinetic control water content.
at low overpotential, where mass-transfer limitations are neglected. There were four reaction mechanisms suggested and discussed
However, the performance of the Ni electrode in the lab-scale cell, at concerning hydrogen production in a molten carbonate electrolysis
least in fuel cell mode, could be affected by the gas phase mass transfer cell, based on the reversed process of hydrogen oxidation in fuel
even at low overpotential, due to possibly not negligible concentration cell operation. However, the experimentally obtained reaction orders
gradients in the current collector.21 This may also be the case at the were high and therefore not consistent with the mechanisms suggested
Ni electrode in electrolysis cell mode. Another possible factor is that above. The interpretation of polarization data and the evaluation of ex-
the effect of gas composition on the electrolyte wetting properties change current densities were dependent on assumptions made about
is not taken into consideration when analyzing the polarization data. the surface area and the effective conductivities. The reaction orders
The wetting angle will affect the active surface area, and therefore could be affected if these two factors were dependent on gas compo-
it probably changes the available active surface area for the faradaic sitions and incorrect assumptions were made. It is also necessary to
reaction.23 The partial pressure dependencies will be changed, if the take into consideration the effect of mass-transfer limitations on the
active surface area or/and the effective conductivity is/are dependent performance of the Ni electrode, even though it is assumed to be neg-
on the gas compositions.13 ligible at low overpotential. Another explanation for this divergence
The number of transferring electrons has no impact on the reaction between experimental data and theoretical calculation might be the
orders, although changing n will change the value of the exchange possibility of co-electrolysis of water and carbon dioxide rather than
current density. In the MCEC mode, water electrolysis may not be the purely water electrolysis.
only reaction process in the cell. The electrolysis of carbon dioxide The activation energy of the Ni porous electrode for hydrogen
is not considered in this analysis of kinetic data. But due to the high production lies between 55 and 100 kJ · mol−1 for all the different gas
reaction order of CO2 (0.59–0.86), it is still another possible reaction compositions, indicating that the Ni electrode was under kinetic or
occurring in the cell. Therefore, future studies will be performed to mixed control.
determine the occurrence of electrolysis of carbon dioxide on the Ni
porous electrode. To study the reaction mechanism in more detail Acknowledgment
and more accurately, other electrochemical techniques could also be
employed to measure electrode performance and make qualitative The financial support of the China Scholarship Council (CSC) is
comparisons. appreciated. The cell components were provided by Ansaldo Fuel
According to the Arrhenius equation, the activation energy (E a ) Cells in Italy.
can be calculated from the following relationship:
List of Symbols
d (log i 0 )
E a = −2.303R [14] F Faraday’s number, C · mol−1
d (1/T ) i current density, A · m−2
For gas compositions in group 1, containing 5–50% hydrogen, the i0 exchange current density, A · m−2
obtained activation energy is in the range of 65–85 kJ · mol−1 . The i 00 standard exchange current density

Downloaded on 2016-04-08 to IP 130.237.74.37 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).
 F1028 Journal of The Electrochemical Society, 162 (9) F1020-F1028 (2015)

n number of electrons 3. V. Kaplan, E. Wachtel, K. Gartsman, Y. Feldman, and I. Lubomirsky, J. Electrochem.

p partial pressure, atm Soc., 157, B552 (2010).
R gas constant, J · mol−1 · K−1 4. V. Kaplan, E. Wachtel, and I. Lubomirsky, J. Electrochem. Soc., 159, E159 (2012).
5. D. Chery, V. Albin, V. Lair, and M. Cassir, Int. J. Hydrogen Energy, 39, 12330 (2014).
S specific surface area, m−1 6. L. Hu, I. Rexed, G. Lindbergh, and C. Lagergren, Int. J. Hydrogen Energy, 39, 12323
T temperature, K (2014).
a partial pressure dependency of hydrogen in MCEC 7. P. G. P. Ang and A. F. Sammells, J. Electrochem. Soc., 127, 1287 (1980).
8. J. Jewulski and L. Suski, J. Appl. Electrochem., 14, 135 (1984).
b partial pressure dependency of carbon dioxide in MCEC 9. S. H. Lu and J. R. Selman, J. Electrochem. Soc., 131, 2827 (1984).
c partial pressure dependency of water in MCEC 10. T. Nishina, M. Takahashi, and I. Uchida, J. Electrochem. Soc., 137, 1112 (1990).
11. C. Y. Yuh and J. R. Selman, J. Electrochem. Soc., 138, 3649 (1991).
12. R. Weewer, K. Hemmes, and J. H. W. de Wit, J. Electrochem. Soc., 142, 389 (1995).
13. G. Lindbergh, M. Olivry, and M. Sparr, J. Electrochem. Soc., 148, A411 (2001).
Greek 14. S. D. Ebbesen, R. Knibbe, and M. Mogensen, J. Electrochem. Soc., 159, F482 (2012).
15. S. H. Clarke, A. L. Dicks, K. Pointon, T. A. Smith, and A. Swann, Catal. Today, 38,
αA anodic transfer coefficient 411 (1997).
αC cathodic transfer coefficient 16. C. Lagergren and D. Simonsson, J. Electrochem. Soc., 144, 3813 (1997).
β symmetry factor 17. C. Lagergren, Electrochemical performance of porous MCFC cathodes, Ph.D. Thesis,
KTH Royal Institute of Technology, Sweden (1997).
γ1 partial pressure dependency of hydrogen in MCFC 18. S. Tanase, Y. Miyazaki, M. Yanagida, K. Tanimoto, and T. Kodama, Prog. Batt. Sol.
γ2 partial pressure dependency of carbon dioxide in MCFC Cells, 6, 195 (1987).
γ3 partial pressure dependency of water in MCFC 19. M. Sparr, Modelling and experimental investigation of the porous nickel anode in
κ effective conductivity of the electrolyte, S · m−1 the molten carbonate fuel cell, Ph.D. Thesis, KTH Royal Institute of Technology,
Sweden (2005).
η overpotential, V 20. M. Kawase, Y. Mugikura, and T. Watanabe, J. Electrochem. Soc., 147, 854 (2000).
21. M. Sparr, A. Bodén, and G. Lindbergh, J. Electrochem. Soc., 153, A1525 (2000).
22. R. Weewer, Study of electrochemcal processes and wetting phenomena at the molten
References carbonate fuel cell anode, Ph.D. Thesis, Delft University of Technology, Netherlands
(1990).
1. Y. Tao, H. Nishino, S. Ashidate, H. Kokubo, M. Watanabe, and H. Uchida, Elec- 23. M. Yoshikawa, A. Bodén, M. Sparr, and G. Lindbergh, J. Power Sources, 158, 94
trochim. Acta, 54, 3309 (2009). (2006).
2. W. H. A. Peelen, K. Hemmes, and J. H. W. de Wit, Electrochim. Acta, 43, 763 (1997). 24. C. Y. Yuh and J. R. Selman, J. Electrochem. Soc., 138, 3642 (1991).

Downloaded on 2016-04-08 to IP 130.237.74.37 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).