- Email: [email protected]
Effect of sulfur contaminants on MCFC performance
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y x x x ( 2 0 1 4 ) 1 e9
Available online at www.sciencedirect.com
ScienceDirect journal homepage: www.elsevier.com/locate/he
Effect of sulfur contaminants on MCFC performance I. Rexed*, C. Lagergren, G. Lindbergh School of Chemical Science and Engineering, Applied Electrochemistry, KTH Royal Institute of Technology, SE-100 44 Stockholm, Sweden
article info
abstract
Article history:
Molten carbonate fuel cells (MCFC) used as carbon dioxide separation units in integrated
Received 12 January 2014
fuel cell and conventional power generation can potentially reduce carbon emission from
Received in revised form
fossil fuel power production. The MCFC can utilize CO2 in combustion flue gas at the
7 March 2014
cathode as oxidant and concentrate it at the anode through the cell reaction and thereby
Accepted 12 March 2014
simplifying capture and storage. However, combustion flue gas often contains sulfur di-
Available online xxx
oxide which, if entering the cathode, causes performance degradation by corrosion and by poisoning of the fuel cell. The effect of contaminating an MCFC with low concentrations of
Keywords:
both SO2 at the cathode and H2S at the anode was studied. The poisoning mechanism of
Molten carbonate fuel cell (MCFC)
SO2 is believed to be that of sulfur transfer through the electrolyte and formation of H2S at
Performance degradation
the anode. By using a small button cell setup in which the anode and cathode behavior can
SO2
be studied separately, the anodic poisoning from SO2 in oxidant gas can be directly
Electrochemical impedance spec-
compared to that of H2S in fuel gas. Measurements were performed with SO2 added to
troscopy (EIS)
oxidant gas in concentrations up to 24 ppm, both for short-term (90 min) and for long-term (100 h) contaminant exposure. The poisoning effect of H2S was studied for gas compositions with high- and low concentration of H2 in fuel gas. The H2S was added to the fuel gas stream in concentrations of 1, 2 and 4 ppm. Results show that the effect of SO2 in oxidant gas was significant after 100 h exposure with 8 ppm, and for short-term exposure above 12 ppm. The effect of SO2 was also seen on the anode side, supporting the theory of a sulfur transfer mechanism and H2S poisoning. The effect on anode polarization of H2S in fuel gas was equivalent to that of SO2 in oxidant gas. Copyright ª 2014, Hydrogen Energy Publications, LLC. Published by Elsevier Ltd. All rights reserved.
Introduction The fuel cell technology is an important part of the energy equation to break our dependency on fossil fuels and to decrease the carbon footprint of power production. Suitable for distributed power generation with high efficiency conversion of energy, fuel cells can simplify the transition from fossil to renewable fuels on the market. The molten carbonate
fuel cell (MCFC) is a high temperature fuel cell with non-noble catalysts and a molten carbonate electrolyte. Oxygen is reduced with CO2 at the cathode, while hydrogen is oxidized at the anode to produce water and CO2 (reactions 1 and 2). In the electrolyte, carbonate ions provide the means of ionic transfer between the cathode and the anode. Unlike other fuel cells, CO2 needs to be supplied to the cathode as well as oxygen, which is in practice solved by recirculating a part of the anode exhaust to be mixed with air at the cathode inlet [1].
* Corresponding author. Tel.: þ46 8 7906615. E-mail address: [email protected] (I. Rexed). http://dx.doi.org/10.1016/j.ijhydene.2014.03.068 0360-3199/Copyright ª 2014, Hydrogen Energy Publications, LLC. Published by Elsevier Ltd. All rights reserved.
Please cite this article in press as: Rexed I, et al., Effect of sulfur contaminants on MCFC performance, International Journal of Hydrogen Energy (2014), http://dx.doi.org/10.1016/j.ijhydene.2014.03.068
2
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y x x x ( 2 0 1 4 ) 1 e9
The operating temperature of 650 C is sufficiently high for steam reforming of natural gas, i.e. methane, to hydrogen, and also allows for CO, which acts as a poison in low temperature fuel cells, to be utilized as fuel or to produce hydrogen in the water-gas shift reaction (reaction 3) [2]. The flexibility towards the fuel composition makes the MCFC ideal to use for power production fueled with biogas produced from renewables or from anaerobic digestion of organic residues or waste sludge, which may vary greatly in their composition [3e5]. The anode and cathode are made of nickel, typically alloyed with Al or Cr to prevent creeping of the anode [6], and in-situ lithiated nickel oxide, respectively. The separator consists of a matrix of LiAlO2, impregnated with carbonate electrolyte. The high temperature makes the conductivity high, which allows the electrodes and separator to be thick (0.5e0.7 mm) compared to low temperature fuel cells. The anode and cathode side reactions are as follows: Anode : Cathode :
H2 þ CO2 3 /H2 O þ CO2 þ 2e
(1)
1 2O2 þ CO2 þ 2e /CO2 3
(2)
The water-gas shift reaction: H2 þ CO2 )/H2 O þ CO
(3)
It is well established that the high quality waste heat of the MCFC can be utilized for combined heat and power (CHP) operation, such as steam- [7] or gas-turbine [8] combined cycles to increase the overall efficiency. A novel concept is the use of the MCFC in CCS (Carbon Capture and Sequestration) applications in combination with conventional combustion or gas turbine power plants to reduce the carbon dioxide emissions from energy production. The cell reaction of the MCFC includes the migration of carbonate ions from the cathode to the anode, which corresponds to a depletion of CO2 in the oxidant feed gas and a subsequent enrichment of CO2 in the anode exhaust gas. By feeding the cathode with combustion flue gas the MCFC can operate as a fuel cell, giving an addition of power output to the power plant, while concentrating CO2 to the anode exhaust where it can be separated at high concentration for removal and storage. This makes it an attractive alternative to conventional CCS technologies. Ideally, only CO2 and water is released from the anode. However, thermodynamics limit both the utilization of fuel and oxidant, which set a minimum concentration of CO2 at the cathode, and leaves unreacted fuel at the anode exhaust. Case studies give removal rates of CO2 for combined MCFC and combustion plants of about 60e80% [9e11]. While some conventional CCS technologies (e.g. absorption by amines, oxy-fuel combustion, pre- or post- combustion capture) can reach as high as 90%, the advantage of the MCFC is that the overall penalty on plant efficiency for a comparable reduction in CO2 emissions is decreased by the power generation of the MCFC [12]. The system is also benefited by the redundancy of anode exhaust gas recirculation to the cathode inlet, since CO2 is supplied in the flue gas. The effect on fuel cell performance and degradation by using flue gas is a concern for the use of MCFC as a CCS application. Combustion flue gas differs in composition from the oxidant gas used in standalone operation (air þ CO2) by having significantly lower
concentrations of both oxygen and carbon dioxide, which affect the fuel cell performance negatively. More alarmingly, flue gas also often contains high concentrations of SO2 which has a detrimental effect on fuel cell performance and lifetime [3]. Traditionally, contaminant degradation has mainly been a concern for the anode side due to the fact that reformate gas, depending on its source, contains many trace components which are directly harmful to MCFC performance and lifetime. In biogas, commonly found fuel cell contaminants are e.g. hydrogen sulfide, halides and siloxanes, of which especially H2S poisoning of the anode is known to decrease performance and lifetime of MCFC [13]. Although a high sulfur content in natural gas and biogas is a problem also in conventional power generation, due to the corrosiveness of sulfur at high temperatures, the severe penalty on performance combined with demands for improvements in fuel cell lifetime incur a more strict tolerance limit for MCFC applications, around 1 ppm sulfur in fuel gas [13e16]. As trace amounts of sulfur are difficult and expensive to clean [17], finding tolerance criteria and quantifying poisoning effects of trace amounts of H2S in fuel gas has therefore been a prioritized issue for MCFC research. The use of flue gas to feed the MCFC cathode now highlights the need to study the degradation phenomena related to SO2 at the cathode. In this paper the effect on fuel cell performance and degradation by exposure to SO2 in oxidant gas is studied. In focus is the effect on the anode side, as a result of sulfur being transferred from the cathode to the anode side and released as H2S. Also studied is the effect of H2S in the fuel gas with different compositions, as the partial pressure of hydrogen and carbon dioxide affect the poisoning mechanism of the anode.
Effect of sulfur contaminant on MCFC Sulfur poisoning by SO2 of MCFC has been considered as a result of anode exhaust gas recirculation to the anode. Any H2S present in the exhaust gas will be oxidized to SO2 (reaction 4) together with residual fuel in a catalytic burner before entering the cathode. According to Weaver and Winnick [14] SO2 is converted to sulfate by the reaction (5) after which the sulfate ions migrate to the anode under load where they accumulate in the electrolyte and release H2S according to the reverse reaction (6). The poisoning effect would then, in addition to corrosion of metallic parts, be divided into a primary and a secondary poisoning mechanism. The primary poisoning is the loss of electrolyte by reaction (5), and the secondary poisoning would be that of poisoning from H2S released from the electrolyte at the anode. Indeed, in one of few experimental studies with SO2 in the oxidant stream a similar poisoning effect as from H2S was observed [3]. However, little experimental work has been made to quantify this poisoning effect of SO2 in oxidant gas as compared to the effect of H2S in fuel gas. H2 S þ 3=2O2 /H2 O þ SO2 2 CO2 3 þ SO2 þ 1 2 O2 /CO2 þ SO4
(4) (5)
Please cite this article in press as: Rexed I, et al., Effect of sulfur contaminants on MCFC performance, International Journal of Hydrogen Energy (2014), http://dx.doi.org/10.1016/j.ijhydene.2014.03.068
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y x x x ( 2 0 1 4 ) 1 e9
Hydrogen sulfide is a known poison for the nickel catalyst and the negative effect on fuel cell performance is well documented [3,14,18e21]. However, different theories for the interactions between H2S and the Ni electrode and the electrolyte are supported in literature. The interactions between H2S and the molten carbonate cause a loss of electrolyte by replacement of the carbonate at the anode by sulfate or sulfide ions (reactions 6 and 7) [14]. 2 H2 S þ CO2 3 þ 3H2 O/SO4 þ CO2 þ 4H2
(6)
2 H2 S þ CO2 3 /H2 O þ CO2 þ S
(7)
H2S is known to cause poisoning of the Ni-anode surface by formation of NiS which kinetically hinders the oxidation of hydrogen [20e22]. It may react chemically to form NiS (reaction 8), or electrochemically as sulfide in the electrolyte (reaction 9), with the nickel anode to form nickel sulfide. The standard potential for NiS and Ni3S2 is 0.756 V and 0.829 V vs. SOE, respectively. Nickel sulfide blocks the catalytic sites on the anode, decreasing the active surface area and affecting the performance negatively. As the anodic potential at OCV is more negative than that for nickel sulfide formation, no oxidation of H2S will occur, however, under current load the anodic overpotential will make the anode potential more positive and allow for nickel sulfide to be formed. It has been found that at low concentrations of H2S and at OCV, the main poisoning is adsorption or chemisorption of H2S dissolved into the electrolyte at the nickel surface [16,20], which inhibit the water-gas shift reaction (reaction 3) and reduce the fuel value [3]. xNi þ yH2 S/Nix Sy þ yH2
(8)
xNi þ y S2 /Nix Sy þ 2xe
(9)
Under load, H2S can also be oxidized at the anode to produce sulfate or sulfide ions if the anode potential increases above the standard potential of respective reaction, 1.073 V and 0.986 V vs. SOE for sulfide and sulfate, respectively (reactions 10 and 11) [20]. 2 þ 5CO2 þ 5H2 O þ 8e 5CO2 3 þ 4H2 þ H2 S/S
(10)
2 5CO2 3 þ H2 S/SO4 þ 5CO2 þ H2 O þ 8e
(11)
Operating conditions are the subject of the design of a fuel cell, and affect the poisoning mechanism of the anode. A hydrogen rich fuel increases the H2/H2S ratio, which affects the adsorption of H2S to the nickel surface, and promotes the oxidation of H2 instead of the competing reaction (9). A high water content in the anodic gas has been shown to make the regeneration of the fuel cell after H2S poisoning faster, presumably by promoting desulfurization of the electrolyte by the reversed reaction (7) [20]. The applied current changes both the anodic potential and the fuel utilization, which may change the partial pressure of reactants locally. Furthermore, nickel sulfide formation on the nickel surface also affects the performance by changing the morphology of the porous structure by blocking catalytic sites [16]. It also changes the wetting properties of the electrolyte [18], which can change the electrolyte distribution and cause a decrease of the
3
performance. Finally, the operating pressure and temperature are important factors which affect the poisoning.
Experimental Measurements were performed on circular button cells with a geometrical area of 3 cm2, operated at atmospheric pressure and at a temperature of 650 C. The setup and assembly has been described in detail in e.g. Ref. [23]. The NieCr anode and in-situ oxidized NiO cathode was cut from sheets of electrode material supplied by Ansaldo Fuel Cells. The matrix was LiAlO2 and the electrolyte was (Li0.62/K0.38)2 CO3. The reference electrodes were gold wires immersed in the same electrolyte as in the cell in an atmosphere of 33% O2 and 67% CO2, in this paper denoted the standard oxygen electrode (SOE). Humidification of anode gas was controlled with a bubbler at a set temperature. Electrochemical performance was evaluated with polarization measurements (IeV curves) and electrochemical impedance spectroscopy (EIS), using a Solartron 1285 potentiostat and 1255 frequency response analyzer. EIS measurements are performed from 10 kHz to 10 mHz with a voltage perturbation of 10 mV, if no other information is given. The small scale of the button cell and open cell configuration with impinging gas flow allow for homogeneous conditions throughout the fuel cell. Inlet and outlet conditions for gas composition can be applied to the entire cell and the Nernst equation for inlet gas composition and temperature can be assumed to be valid throughout the entire fuel cell. The gas flow for both anode and cathode are sufficiently high to ensure a stoichiometric surplus by order of magnitude. This is equivalent to a low utilization factor (Uf) for both fuel and oxidant. Post-test analysis of anode and cathode material was performed with scanning electron microscopy with energy dispersive X-ray spectroscopy (SEM-EDS) to detect residual sulfur on the anode and cathode surface. In button cell testing studying the effect of H2S on the anode side, an exposure time of 90 min was found to be sufficient to detect poisoning effects [24]. Similar to this, shortterm testing of H2S at the anode and SO2 at the cathode was performed by adding the contaminant to either anode of cathode gas for 90 min, after which the performance was measured with IeV curves and EIS measurements. The measurements were performed in the order of cell, cathode and anode, which resulted in a longer exposure time of contaminant before measurements over the separate electrodes than the stipulated 90 min for the cell measurements for shortterm contamination tests. During long-term tests the cell performance was measured continuously before, during, and after the test period. After contaminant testing, the cell was regenerated by changing to clean gas and applying a constant load of 100 mA cm2, with the purpose of removing any reversible poisoning effects due to interactions between residual sulfur compounds and electrode or electrolyte. Regeneration was performed overnight for approximately 20 h or until stable performance was reached. Contaminants were added to a gas stream of pre-mixed oxidant or fuel gas (AGA gas) stream in N2 carrier gas after humidification to avoid scrubbing of sulfur contaminant. To maintain a constant flow rate to the cell the flow was balanced
Please cite this article in press as: Rexed I, et al., Effect of sulfur contaminants on MCFC performance, International Journal of Hydrogen Energy (2014), http://dx.doi.org/10.1016/j.ijhydene.2014.03.068
4
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y x x x ( 2 0 1 4 ) 1 e9
with a second addition of N2 to compensate for the change in contaminant þ carrier gas flow when the contaminant level is changed or when clean gas was used. The SO2 poisoning effect was measured with oxidant gas consisting of 15% O2, 30% CO2 and 55% N2 diluted with 90 ppm SO2 in N2 carrier gas and the anodic gas composition consisted of 29% H2, 5% CO2 and 66% N2 humidified at 40 C. The resulting compositions of oxidant gas was 14% O2, 27% CO2 and 59% N2 for long-term test with 8 ppm of SO2. The flow rates of dry gases for the anode and cathode was 150 and 110 mL/min, respectively. The composition of oxidant gas for the short-term test with up to 24 ppm of SO2 was 11% O2, 23% CO2 and 66%, with a flow rate of dry gases for both the anode and cathode of 200 mL/min. The effect of poisoning from H2S at concentrations in the range of 1e4 ppm was tested by adding 350 ppm H2S in N2 carrier gas to the anode gas for two different anode gas compositions. The first anodic gas composition consisted of 75% H2, 19% CO2 and 6% N2 and was humidified at 60 C. The second anodic gas composition consisted of 29% H2, 5% CO2 and 66% N2, and was humidified at 60 C. The oxidant gas consisted of 15% O2, 30% CO2 and 55% N2. The total flow rate of dry gases for both anode and cathode was 150 mL/min.
Results e discussion In a first set of experiments, the cell was exposed to 8 ppm of SO2 in the oxidant gas for 100 h. The cell was kept at a load of 100 mA cm2 during the test period. In order to accelerate any effect on the anode side the fuel gas composition used was low in H2 (29%) and also the humidification temperature was low (40 C). The performance was measured continuously
Fig. 1 e a) Cell voltage, b) anode voltage vs. SOE and c) cathode voltage vs. SOE at different current density with 8 ppm SO2 in oxidant gas.
with IeV curves before, during, and after the addition of SO2. Fig. 1 shows the cell voltage together with anode and cathode measured separately; at OCV and at 100, 200 and 300 mA cm2. The degradation in cell voltage (Fig. 1(a)) can be seen to progress during the extent of the time of exposure to SO2, with a stronger effect at higher current density. When the oxidant gas is switched to clean gas the cell voltage is only regenerated to a small extent, despite operating with in such conditions for more than 100 h. The regeneration of the cell voltage is smaller than expected from studies of low concentrations of H2S in fuel gas, however, low hydrogen and low water content has been shown to effectively reduce the voltage regeneration [20]. In addition, the anode and cathode potentials vs. SOE were measured in parallel with the cell voltage. The effect of SO2 in oxidant gas is clearly seen on the anode (Fig. 1(b)), and on the cathode (Fig. 1(c)). The anodic potential shows a larger effect of regeneration than what can be seen for the cell, which is explained by the cathode potential which shows a continuously decreasing trend at all current densities even after changing to clean gas. In Fig. 2 it can be seen that the cell polarization during the 100 h of exposure to 8 ppm SO2 increases with exposure time to SO2 and is more pronounced at higher current density. The cathodic contribution to the voltage degradation is increasing linearly with exposure time, while the anodic contribution is accelerated with high current density and longer exposure time. The effect on the anodic potential clearly indicates a poisoning effect, and a transfer of sulfur from cathode to anode. This is assumed to progress through dissolution of SO2 and formation of sulfate ions in the electrolyte (reaction 5), which migrates to the anode side where they are converted to H2S. The increase in cathodic polarization is indicative of a progressive increase in the ohmic resistance, as will be discussed in the following section. Two possible causes for the increase in resistance are electrolyte loss by replacement of carbonate with sulfate ions in the electrolyte, which may lower conductivity, and the formation of a corrosion layer at
Fig. 2 e Polarization curves showing overvoltage (DE) of cell, anode and cathode before contamination test and after 20 h, 45 h, 75, and 100 h exposure to 8 ppm SO2.
Please cite this article in press as: Rexed I, et al., Effect of sulfur contaminants on MCFC performance, International Journal of Hydrogen Energy (2014), http://dx.doi.org/10.1016/j.ijhydene.2014.03.068
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y x x x ( 2 0 1 4 ) 1 e9
5
Fig. 3 e Nyquist diagram of cell impedance at a) OCV and b) 100 mA cmL2, after different exposure time to 8 ppm SO2 in oxidant gas. (B 2.5 kHz, D 250 Hz, , 25 Hz >2.5 Hz).
the electrode-current collector interface, which would increase the contact resistance. Although this does not satisfyingly explain the continued decrease in cathodic voltage after the oxidant has been changed to clean gas, the difference in pore structure and the faster kinetics of hydrogen oxidation compared to the oxygen reduction reaction makes the cathode more sensitive to changes in the electrolyte distribution due to changed wetting properties or electrolyte loss, and it
can be expected that the cathode performance is more strongly affected than the anode performance if electrode degradation is contributing to the increase in resistance. A continuous accelerated electrolyte loss could also prove to be a major cause of degradation for long-term exposure to SO2. The result from impedance measurements of the cell is presented in the Nyquist diagram in Fig. 3. The impedance curves are shifted to the right with longer exposure times of
Fig. 4 e Nyquist diagram of a) anode, and b) cathode impedance at OCV, after different exposure time to 8 ppm SO2 in oxidant gas. (B 2.5 kHz, D 250 Hz, , 25 Hz >2.5 Hz). Please cite this article in press as: Rexed I, et al., Effect of sulfur contaminants on MCFC performance, International Journal of Hydrogen Energy (2014), http://dx.doi.org/10.1016/j.ijhydene.2014.03.068
6
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y x x x ( 2 0 1 4 ) 1 e9
SO2, which indicate an increased cell resistance; although at low frequency the scattering of impedance data complicates the analysis of the cell impedance. The ohmic resistance can be identified by the intercept with the real axis at high frequency in the Nyquist diagram. During 100 h of exposure to SO2 in oxidant gas, the ohmic resistance increases from 0.134 U to 0.161 U after 45 h, and to 0.173 U after 100 h of exposure. The poisoning effect of the anode by reaction (9) is assumed to be better studied under load, as the formation of NixSy has a standard potential more positive than the anodic OCV. Furthermore, poisoning effects on the anode are more pronounced under load, as seen in Fig. 2. However, no significant difference in terms of poisoning effects can be seen in the cell impedance at OCV (Fig. 3(a)) compared to 100 mA cm2 (Fig. 3(b)). The cell impedance at 100 mA cm2 shows a decreased mass transfer arc compared to the impedance at OCV, which is due to the water production at the anode under current load. In the corresponding Nyquist plot for the anode (Fig. 4(a)) there is no visible increase in ohmic resistance during the measurement period. The anodic impedance data show the same scattering at low frequency as for the cell impedance, why it is difficult to analyze any change in mass transfer properties of the anode. An increase in this region was expected as it has been attributed to a competing reaction of H2S and H2 for hydrogen oxidation sites at the nickel surface when testing the effect of H2S in low concentrations in the fuel gas [24]. The cathode impedance curve (Fig. 4(b)) is shifted to the right, which indicates a higher resistance, although there is no high frequency intercept for an accurate reading of the ohmic resistance. However, the shift in the cathodic impedance curve is most likely corresponding to the increase in cell resistance and the cause of the cathodic voltage degradation during the long-term contamination testing (Fig. 1(c)). The high frequency loop in the cathode Nyquist diagram is an artifact stemming from the reference electrode and is not of relevance for the fuel cell performance.
Fig. 5 e Polarization curves showing overvoltage (DE) of cell, anode and cathode with different concentration of SO2 added in oxidant gas for 90 min. Regeneration with clean gas 20 h.
The short-term exposure to contaminant tests were performed with increased levels of SO2 (2, 4, 6, 8, 12, 18 and 24 ppm). Polarization curves are shown in Fig. 5. The effect on cell performance of SO2 in concentration lower than 12 ppm for as short time as 90 min was negligible, and thus these data are omitted. When the SO2 concentration was increased to 18 ppm and 24 ppm at a current density of 230 mA cm2 the voltage drop compared to clean gas was DEcell ¼ 15 mV and 30 mV, respectively (Fig. 5). When comparing the IeV curves for the separate electrodes, it must first be mentioned that the reference electrode measurements contributes to a small difference between the summation of the anodic and cathodic voltage and the measured cell voltage. A possible explanation to this is the different exposure time for cell and electrode measurements, as explained in the experimental procedure. However, it can be concluded that the increase in anode polarization at 230 mA cm2 with 18 ppm and 24 ppm of SO2 compared to clean gas is larger than the increase in cathode polarization; DEan ¼ þ15 mV and þ30 mV compared to DEcat ¼ 7 mV and 15 mV (Fig. 5). The effect on the anodic polarization is increasingly nonlinear above 100 mA cm2 with 24 ppm of SO2 in oxidant gas, which can be interpreted as the result of the concentration- and voltage dependent formation of NixSy by reaction (9). After regeneration with clean gas for 20 h after exposure to 24 ppm of SO2 for 90 min no remaining poisoning effect can be seen for the cell (Fig. 5). Fig. 6 shows the cell voltage as a function of time during the exposure to 12, 18 and 24 ppm of SO2 and the following regeneration period. It can be seen that the rate of voltage decay is increasing with higher concentration of SO2, but also that the voltage decay continues after changing to clean gas for approximately the same time as the contamination period, before the voltage is recuperated to a steady state value. This is interpreted to be an effect of accumulated sulfate in the electrolyte which continues to poison the anode by conversion to H2S according to the reversed reaction (6). The poisoning effect on the anode is eventually reduced, as the concentration of sulfate in the electrolyte decreases and no
Fig. 6 e Cell voltage at constant load 100 mA cmL2 during exposure to 12, 18 and 24 ppm of SO2 in oxidant gas for 90 min, and the following regeneration with clean gas.
Please cite this article in press as: Rexed I, et al., Effect of sulfur contaminants on MCFC performance, International Journal of Hydrogen Energy (2014), http://dx.doi.org/10.1016/j.ijhydene.2014.03.068
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y x x x ( 2 0 1 4 ) 1 e9
7
Fig. 7 e Nyquist diagram of cell impedance at a) OCV, and b) 100 mA cmL2, for different concentration of SO2 added in oxidant gas for 90 min. (B 2.5 kHz, D 250 Hz, , 25 Hz >2.5 Hz).
more H2S is produced. Up to 12 ppm SO2 the rate of sulfate formation is not sufficient to raise the concentration of H2S at the anode above the level where a poisoning effect is visible, and thus no such effect is seen for the following regeneration with clean gas. In order to verify the concentration of H2S at the anode, gas analysis of anode exhaust is necessary, however, that was not available for this work. The Nyquist diagram of the cell impedance with 12e24 ppm of SO2 in oxidant gas (Fig. 7) shows a slight increase in the cell resistance by a shift in the high frequency intercept. The low frequency region, especially under load, is difficult to analyze due to scattering of the impedance data in this region. No significant difference can be seen between cell impedance at OCV and under load. The anode impedance (Fig. 8(a)) shows no increase of ohmic resistance, similar to results of the long-term exposure test. However, the cathode impedance (Fig. 8(b)) shows a slight shift in the impedance curve, although less than what was found after the long-term exposure test. The result of the short-term contamination tests differs from that of long term exposure tests both in the concentration at which the addition of SO2 in oxidant gas has a clearly negative effect on cell performance, and in terms of the effect of cell regeneration. In the first perspective, it is expected that extended exposure time results in more severe poisoning, as sulfate dissolved into the electrolyte would accumulate in the electrolyte and create an increasingly more harmful atmosphere at the anode. The anode polarization was effectively regenerated with clean gas, both after long-term and shortterm poisoning. This is in agreement with what has been found in studies of poisoning of MCFC anodes by H2S in biogas
simulated fuel gas [13,24]. However, in contrast to long-term contamination tests, no effect was seen on the cathode after short-term contamination and regeneration with clean gas. The effect of adding 1e4 ppm of H2S to the anode gas was tested with different anode gas compositions with the purpose of comparing the effect of SO2 at the cathode and H2S at the anode. The poisoning effect was measured for one composition with high concentration of H2 (75% H2, 19% CO2 and 6% N2), and one composition which was diluted with nitrogen (29% H2, 5% CO2 and 66% N2). The anode gas was humidified at 60 C. The H2 concentration in the anode gas affects the poisoning of the anode by changing the H2/H2S partial pressure ratio, which is important for the competing reaction (8). A low H2 concentration will simplify the formation of nickel sulfide and accelerate the degradation of the anode. The polarization curves are shown in Fig. 9. In contamination tests with the high H2 composition of anode gas, no significant effect can be seen on the polarization of the cell in the concentrations tested. In order to accelerate the poisoning effect at low concentration of H2S, the contaminant gas was added in a diluted anodic gas to the fuel cell. When using the anode gas with low concentration of H2 there is an increase in cell polarization for 2 and 4 ppm of H2S. The voltage degradation is further highlighted in the anodic polarization curves. For concentrations of H2S higher than 1 ppm, anode polarization is significant, and the poisoning effect is increased with higher concentration of H2S and higher current density. The cathodic polarization curves show no effect on polarization. The increased effect of H2S at the anode under load compared to OCV is discussed in Ref. [20]; the poisoning at
Please cite this article in press as: Rexed I, et al., Effect of sulfur contaminants on MCFC performance, International Journal of Hydrogen Energy (2014), http://dx.doi.org/10.1016/j.ijhydene.2014.03.068
8
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y x x x ( 2 0 1 4 ) 1 e9
Fig. 8 e Nyquist diagram of a) anode, and b) cathode impedance at OCV, for different concentration of SO2 added in oxidant gas for 90 min. (B 2.5 kHz, D 250 Hz, , 25 Hz >2.5 Hz).
OCV or low polarization is caused only by adsorption and chemisorption on the nickel surface, but at high current density nickel sulfide may be produced in the electrode reaction, which can block the anode surface for hydrogen oxidation. The effect on the anodic polarization of H2S poisoning is similar to that of short-term tests with SO2 at the cathode. This supports the theory of sulfur transfer from the cathode to
the anode in the electrolyte and subsequent release of H2S at the anode as being the main cause of anode poisoning. The results from post-test analysis by SEM-EDS did not show any residual sulfur at the anode or the cathode surface, although a small poisoning effect was detected by electrochemical analysis of the fuel cell performance. In a study with 95 ppm of H2S in fuel gas [20], sulfur compounds was identified with SEM-EDS analysis, however, the concentration of H2S at the anode in this study was most likely too low to leave detectable amounts at the anode surface after regeneration.
Conclusions
Fig. 9 e Polarization curves showing overvoltage (DE) of cell, anode and cathode with clean gas and after 90 min exposure to 1e4 ppm H2S in anode gas. Regeneration is performed with clean gas for 20 h. Anode gas composition: High H2 (75% H2, 19% CO2, 6% N2), low H2 (29% H2, 5% CO2, 66% N2) humidified at 60 C.
SO2 added to the cathode gives an immediate and significant effect on the anode, which indicates a transfer mechanism of sulfur from anode to cathode. This is assumed to occur by SO2 dissolving into the electrolyte as sulfate ions which migrate to the cathode to be released as H2S, and by poisoning of the anode catalyst by interactions between nickel and H2S. Longterm exposure to 8 ppm SO2 at the cathode resulted in increased polarization of both anode and cathode, with a stronger effect at high current density. The cathode polarization increased linearly while the anode polarization was accelerating with high current density and longer exposure time. This indicates anode side poisoning by H2S. The effect of the short-term contamination tests on anode polarization is visible for concentrations higher than 12 ppm SO2. The short exposure time shows that the transfer mechanism of sulfur from the cathode to the anode is fast, although no detailed data on transfer kinetics can be derived from this study. The poisoning effect is increasing with exposure time
Please cite this article in press as: Rexed I, et al., Effect of sulfur contaminants on MCFC performance, International Journal of Hydrogen Energy (2014), http://dx.doi.org/10.1016/j.ijhydene.2014.03.068
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y x x x ( 2 0 1 4 ) 1 e9
and a delayed poisoning effect is seen after clean gas was introduced at the cathode side for regeneration. This indicates an accumulation of sulfate in the electrolyte, which results in an increase of H2S concentration at the anode with time. The anode poisoning effects were largely reversible by regeneration with clean gas. The cathode, however, experienced irreversible polarization effects after 100 h exposure to 8 ppm SO2. The increase in cell resistance after long-term exposure was attributed to the cathode. The cause may be either carbonate being replaced by sulfate by reaction with SO2 dissolved in the electrolyte, or the build-up of a corrosion layer at the electrode-current collector interface. The result shows that SO2 in oxidant gas is of great concern for the use of flue gas to feed the cathode of MCFCs. Poisoning of the anode was further supported by comparing the effect on the anode polarization from adding SO2 at the cathode with the effect of adding H2S at the anode, which was similar in the two experiments. However, impedance measurements did not reveal detailed information about the predominant poisoning mechanism, partly due to a reoccurring disturbance in the low frequency region in which mass transfer is studied.
references
[1] Farooque M, Maru H. Carbonate fuel cells: milliwatts to megawatts. J Power Sources 2006;160:827e34. [2] Clarke S, Dicks A, Pointon K, Smith T, Swann A. Catalytic aspects of the steam reforming of hydrocarbons in internal reforming fuel cells. Catal Today 1997;38:411e23. [3] Watanabe T, Izaki Y, Mugikura Y, Morita H, Yoshikawa M, Kawase M, et al. Applicability of molten carbonate fuel cells to various fuels. J Power Sources 2006;160:868e71. [4] Sanchez D, Monje B, Chacartegui R, Campanari S. Potential of molten carbonate fuel cells to enhance the performance of CHP plants in sewage treatment facilities. Int J Hydrogen Energy 2013;38:394e405. [5] Cigolotti V, McPhail S, Moreno A. Nonconventional fuels for high-temperature fuel cells: status and issues. J Fuel Cell Sci Technol 2009;6:021311. [6] Selman JR. Research, development, and demonstration of molten carbonate fuel cell Systems. In: Blomen LJMJ, Mugerwa MN, editors. Fuel cell systems. New York: Plenum Press; 1993. pp. 345e463. [7] Varbanov P, Klemes J, Shah R, Shihn H. Power cycle integration and efficiency increase of molten carbonate fuel cell systems. J Fuel Cell Sci Technol 2006;3:375e83.
9
[8] Massardo A, Bosio B. Assessment of molten carbonate fuel cell models and integration with gas and steam cycles. J Eng Gas Turbines Power-Trans ASME 2002;124:103e9. [9] Desideri U, Proietti S, Sdringola P, Cinti G, Curbis F. MCFCbased CO2 capture system for small scale CHP plants. Int J Hydrogen Energy 2012;37:19295e303. [10] Chiesa P, Campanari S, Manzolini G. CO2 cryogenic separation from combined cycles integrated with molten carbonate fuel cells. Int J Hydrogen Energy 2011;36:10355e65. [11] Campanari S, Chiesa P, Manzolini G. CO2 capture from combined cycles integrated with molten carbonate fuel cells. Int J Greenh Gas Control 2010;4:441e51. [12] Discepoli G, Cinti G, Desideri U, Penchini D, Proietti S. Carbon capture with molten carbonate fuel cells: experimental tests and fuel cell performance assessment. Int J Greenh Gas Control 2012;9:372e84. [13] Ciccoli R, Cigolotti V, Lo Presti R, Massi E, McPhail SJ, Monteleone G, et al. Molten carbonate fuel cells fed with biogas: combating H2S. Waste Manag; 2010:1018e24. [14] Weaver D, Winnick J. Sulfation of the molten carbonate fuelcell anode. J Electrochem Soc 1989;136:1679e86. [15] Di Giulio N, Bosio B, Cigolotti V, Nam SW. Experimental and theoretical analysis of H2S effects on MCFCs. Int J Hydrogen Energy 2012;37:19329e36. [16] Cigolotti V, McPhail S, Moreno A, Yoon S, Han J, Nam S, et al. MCFC fed with biogas: experimental investigation of sulphur poisoning using impedance spectroscopy. Int J Hydrogen Energy 2011;36:10311e8. [17] Hernandez S, Scarpa F, Fino D, Conti R. Biogas purification for MCFC application. Int J Hydrogen Energy 2011;36:8112e8. [18] Kawase M, Mugikura Y, Watanabe T. The effects of H2S on electrolyte distribution and cell performance in the molten carbonate fuel cell. J Electrochem Soc 2000;147:1240e4. [19] Uchida I, Ohuchi S, Nishina T. Kinetic-studies of the effects of H2S impurity on hydrogen oxidation in molten (Li þ K)CO3. J Electroanal Chem 1994;369:161e8. [20] Zaza F, Paoletti C, LoPresti R, Simonetti E, Pasquali M. Studies on sulfur poisoning and development of advanced anodic materials for waste-to-energy fuel cells applications. J Power Sources 2010;195:4043e50. [21] Vogel WM, Smith SW. The effect of sulfur on the anodic H2 (Ni) electrode in fused Li2CO3eK2CO3 at 650 C. J Electrochem Soc 1982;129:1441e5. [22] Devianto H, Yoon SP, Nam SW, Han J, Lim TH. The effect of a ceria coating on the H2S tolerance of a molten carbonate fuel cell. J Power Sources 2006;159:1147e52. [23] Lagergren C. Electrochemical performance of porous MCFC cathodes, Ph.D. thesis. Stockholm: KTH; 1997. [24] Devianto H, Simonetti E, McPhail S, Zaza F, Cigolotti V, Paoletti C, et al. Electrochemical impedance study of the poisoning behaviour of Ni-based anodes at low concentrations of H2S in an MCFC. Int J Hydrogen Energy 2012;37:19312e8.
Please cite this article in press as: Rexed I, et al., Effect of sulfur contaminants on MCFC performance, International Journal of Hydrogen Energy (2014), http://dx.doi.org/10.1016/j.ijhydene.2014.03.068
Available online at www.sciencedirect.com
ScienceDirect journal homepage: www.elsevier.com/locate/he
Effect of sulfur contaminants on MCFC performance I. Rexed*, C. Lagergren, G. Lindbergh School of Chemical Science and Engineering, Applied Electrochemistry, KTH Royal Institute of Technology, SE-100 44 Stockholm, Sweden
article info
abstract
Article history:
Molten carbonate fuel cells (MCFC) used as carbon dioxide separation units in integrated
Received 12 January 2014
fuel cell and conventional power generation can potentially reduce carbon emission from
Received in revised form
fossil fuel power production. The MCFC can utilize CO2 in combustion flue gas at the
7 March 2014
cathode as oxidant and concentrate it at the anode through the cell reaction and thereby
Accepted 12 March 2014
simplifying capture and storage. However, combustion flue gas often contains sulfur di-
Available online xxx
oxide which, if entering the cathode, causes performance degradation by corrosion and by poisoning of the fuel cell. The effect of contaminating an MCFC with low concentrations of
Keywords:
both SO2 at the cathode and H2S at the anode was studied. The poisoning mechanism of
Molten carbonate fuel cell (MCFC)
SO2 is believed to be that of sulfur transfer through the electrolyte and formation of H2S at
Performance degradation
the anode. By using a small button cell setup in which the anode and cathode behavior can
SO2
be studied separately, the anodic poisoning from SO2 in oxidant gas can be directly
Electrochemical impedance spec-
compared to that of H2S in fuel gas. Measurements were performed with SO2 added to
troscopy (EIS)
oxidant gas in concentrations up to 24 ppm, both for short-term (90 min) and for long-term (100 h) contaminant exposure. The poisoning effect of H2S was studied for gas compositions with high- and low concentration of H2 in fuel gas. The H2S was added to the fuel gas stream in concentrations of 1, 2 and 4 ppm. Results show that the effect of SO2 in oxidant gas was significant after 100 h exposure with 8 ppm, and for short-term exposure above 12 ppm. The effect of SO2 was also seen on the anode side, supporting the theory of a sulfur transfer mechanism and H2S poisoning. The effect on anode polarization of H2S in fuel gas was equivalent to that of SO2 in oxidant gas. Copyright ª 2014, Hydrogen Energy Publications, LLC. Published by Elsevier Ltd. All rights reserved.
Introduction The fuel cell technology is an important part of the energy equation to break our dependency on fossil fuels and to decrease the carbon footprint of power production. Suitable for distributed power generation with high efficiency conversion of energy, fuel cells can simplify the transition from fossil to renewable fuels on the market. The molten carbonate
fuel cell (MCFC) is a high temperature fuel cell with non-noble catalysts and a molten carbonate electrolyte. Oxygen is reduced with CO2 at the cathode, while hydrogen is oxidized at the anode to produce water and CO2 (reactions 1 and 2). In the electrolyte, carbonate ions provide the means of ionic transfer between the cathode and the anode. Unlike other fuel cells, CO2 needs to be supplied to the cathode as well as oxygen, which is in practice solved by recirculating a part of the anode exhaust to be mixed with air at the cathode inlet [1].
* Corresponding author. Tel.: þ46 8 7906615. E-mail address: [email protected] (I. Rexed). http://dx.doi.org/10.1016/j.ijhydene.2014.03.068 0360-3199/Copyright ª 2014, Hydrogen Energy Publications, LLC. Published by Elsevier Ltd. All rights reserved.
Please cite this article in press as: Rexed I, et al., Effect of sulfur contaminants on MCFC performance, International Journal of Hydrogen Energy (2014), http://dx.doi.org/10.1016/j.ijhydene.2014.03.068
2
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y x x x ( 2 0 1 4 ) 1 e9
The operating temperature of 650 C is sufficiently high for steam reforming of natural gas, i.e. methane, to hydrogen, and also allows for CO, which acts as a poison in low temperature fuel cells, to be utilized as fuel or to produce hydrogen in the water-gas shift reaction (reaction 3) [2]. The flexibility towards the fuel composition makes the MCFC ideal to use for power production fueled with biogas produced from renewables or from anaerobic digestion of organic residues or waste sludge, which may vary greatly in their composition [3e5]. The anode and cathode are made of nickel, typically alloyed with Al or Cr to prevent creeping of the anode [6], and in-situ lithiated nickel oxide, respectively. The separator consists of a matrix of LiAlO2, impregnated with carbonate electrolyte. The high temperature makes the conductivity high, which allows the electrodes and separator to be thick (0.5e0.7 mm) compared to low temperature fuel cells. The anode and cathode side reactions are as follows: Anode : Cathode :
H2 þ CO2 3 /H2 O þ CO2 þ 2e
(1)
1 2O2 þ CO2 þ 2e /CO2 3
(2)
The water-gas shift reaction: H2 þ CO2 )/H2 O þ CO
(3)
It is well established that the high quality waste heat of the MCFC can be utilized for combined heat and power (CHP) operation, such as steam- [7] or gas-turbine [8] combined cycles to increase the overall efficiency. A novel concept is the use of the MCFC in CCS (Carbon Capture and Sequestration) applications in combination with conventional combustion or gas turbine power plants to reduce the carbon dioxide emissions from energy production. The cell reaction of the MCFC includes the migration of carbonate ions from the cathode to the anode, which corresponds to a depletion of CO2 in the oxidant feed gas and a subsequent enrichment of CO2 in the anode exhaust gas. By feeding the cathode with combustion flue gas the MCFC can operate as a fuel cell, giving an addition of power output to the power plant, while concentrating CO2 to the anode exhaust where it can be separated at high concentration for removal and storage. This makes it an attractive alternative to conventional CCS technologies. Ideally, only CO2 and water is released from the anode. However, thermodynamics limit both the utilization of fuel and oxidant, which set a minimum concentration of CO2 at the cathode, and leaves unreacted fuel at the anode exhaust. Case studies give removal rates of CO2 for combined MCFC and combustion plants of about 60e80% [9e11]. While some conventional CCS technologies (e.g. absorption by amines, oxy-fuel combustion, pre- or post- combustion capture) can reach as high as 90%, the advantage of the MCFC is that the overall penalty on plant efficiency for a comparable reduction in CO2 emissions is decreased by the power generation of the MCFC [12]. The system is also benefited by the redundancy of anode exhaust gas recirculation to the cathode inlet, since CO2 is supplied in the flue gas. The effect on fuel cell performance and degradation by using flue gas is a concern for the use of MCFC as a CCS application. Combustion flue gas differs in composition from the oxidant gas used in standalone operation (air þ CO2) by having significantly lower
concentrations of both oxygen and carbon dioxide, which affect the fuel cell performance negatively. More alarmingly, flue gas also often contains high concentrations of SO2 which has a detrimental effect on fuel cell performance and lifetime [3]. Traditionally, contaminant degradation has mainly been a concern for the anode side due to the fact that reformate gas, depending on its source, contains many trace components which are directly harmful to MCFC performance and lifetime. In biogas, commonly found fuel cell contaminants are e.g. hydrogen sulfide, halides and siloxanes, of which especially H2S poisoning of the anode is known to decrease performance and lifetime of MCFC [13]. Although a high sulfur content in natural gas and biogas is a problem also in conventional power generation, due to the corrosiveness of sulfur at high temperatures, the severe penalty on performance combined with demands for improvements in fuel cell lifetime incur a more strict tolerance limit for MCFC applications, around 1 ppm sulfur in fuel gas [13e16]. As trace amounts of sulfur are difficult and expensive to clean [17], finding tolerance criteria and quantifying poisoning effects of trace amounts of H2S in fuel gas has therefore been a prioritized issue for MCFC research. The use of flue gas to feed the MCFC cathode now highlights the need to study the degradation phenomena related to SO2 at the cathode. In this paper the effect on fuel cell performance and degradation by exposure to SO2 in oxidant gas is studied. In focus is the effect on the anode side, as a result of sulfur being transferred from the cathode to the anode side and released as H2S. Also studied is the effect of H2S in the fuel gas with different compositions, as the partial pressure of hydrogen and carbon dioxide affect the poisoning mechanism of the anode.
Effect of sulfur contaminant on MCFC Sulfur poisoning by SO2 of MCFC has been considered as a result of anode exhaust gas recirculation to the anode. Any H2S present in the exhaust gas will be oxidized to SO2 (reaction 4) together with residual fuel in a catalytic burner before entering the cathode. According to Weaver and Winnick [14] SO2 is converted to sulfate by the reaction (5) after which the sulfate ions migrate to the anode under load where they accumulate in the electrolyte and release H2S according to the reverse reaction (6). The poisoning effect would then, in addition to corrosion of metallic parts, be divided into a primary and a secondary poisoning mechanism. The primary poisoning is the loss of electrolyte by reaction (5), and the secondary poisoning would be that of poisoning from H2S released from the electrolyte at the anode. Indeed, in one of few experimental studies with SO2 in the oxidant stream a similar poisoning effect as from H2S was observed [3]. However, little experimental work has been made to quantify this poisoning effect of SO2 in oxidant gas as compared to the effect of H2S in fuel gas. H2 S þ 3=2O2 /H2 O þ SO2 2 CO2 3 þ SO2 þ 1 2 O2 /CO2 þ SO4
(4) (5)
Please cite this article in press as: Rexed I, et al., Effect of sulfur contaminants on MCFC performance, International Journal of Hydrogen Energy (2014), http://dx.doi.org/10.1016/j.ijhydene.2014.03.068
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y x x x ( 2 0 1 4 ) 1 e9
Hydrogen sulfide is a known poison for the nickel catalyst and the negative effect on fuel cell performance is well documented [3,14,18e21]. However, different theories for the interactions between H2S and the Ni electrode and the electrolyte are supported in literature. The interactions between H2S and the molten carbonate cause a loss of electrolyte by replacement of the carbonate at the anode by sulfate or sulfide ions (reactions 6 and 7) [14]. 2 H2 S þ CO2 3 þ 3H2 O/SO4 þ CO2 þ 4H2
(6)
2 H2 S þ CO2 3 /H2 O þ CO2 þ S
(7)
H2S is known to cause poisoning of the Ni-anode surface by formation of NiS which kinetically hinders the oxidation of hydrogen [20e22]. It may react chemically to form NiS (reaction 8), or electrochemically as sulfide in the electrolyte (reaction 9), with the nickel anode to form nickel sulfide. The standard potential for NiS and Ni3S2 is 0.756 V and 0.829 V vs. SOE, respectively. Nickel sulfide blocks the catalytic sites on the anode, decreasing the active surface area and affecting the performance negatively. As the anodic potential at OCV is more negative than that for nickel sulfide formation, no oxidation of H2S will occur, however, under current load the anodic overpotential will make the anode potential more positive and allow for nickel sulfide to be formed. It has been found that at low concentrations of H2S and at OCV, the main poisoning is adsorption or chemisorption of H2S dissolved into the electrolyte at the nickel surface [16,20], which inhibit the water-gas shift reaction (reaction 3) and reduce the fuel value [3]. xNi þ yH2 S/Nix Sy þ yH2
(8)
xNi þ y S2 /Nix Sy þ 2xe
(9)
Under load, H2S can also be oxidized at the anode to produce sulfate or sulfide ions if the anode potential increases above the standard potential of respective reaction, 1.073 V and 0.986 V vs. SOE for sulfide and sulfate, respectively (reactions 10 and 11) [20]. 2 þ 5CO2 þ 5H2 O þ 8e 5CO2 3 þ 4H2 þ H2 S/S
(10)
2 5CO2 3 þ H2 S/SO4 þ 5CO2 þ H2 O þ 8e
(11)
Operating conditions are the subject of the design of a fuel cell, and affect the poisoning mechanism of the anode. A hydrogen rich fuel increases the H2/H2S ratio, which affects the adsorption of H2S to the nickel surface, and promotes the oxidation of H2 instead of the competing reaction (9). A high water content in the anodic gas has been shown to make the regeneration of the fuel cell after H2S poisoning faster, presumably by promoting desulfurization of the electrolyte by the reversed reaction (7) [20]. The applied current changes both the anodic potential and the fuel utilization, which may change the partial pressure of reactants locally. Furthermore, nickel sulfide formation on the nickel surface also affects the performance by changing the morphology of the porous structure by blocking catalytic sites [16]. It also changes the wetting properties of the electrolyte [18], which can change the electrolyte distribution and cause a decrease of the
3
performance. Finally, the operating pressure and temperature are important factors which affect the poisoning.
Experimental Measurements were performed on circular button cells with a geometrical area of 3 cm2, operated at atmospheric pressure and at a temperature of 650 C. The setup and assembly has been described in detail in e.g. Ref. [23]. The NieCr anode and in-situ oxidized NiO cathode was cut from sheets of electrode material supplied by Ansaldo Fuel Cells. The matrix was LiAlO2 and the electrolyte was (Li0.62/K0.38)2 CO3. The reference electrodes were gold wires immersed in the same electrolyte as in the cell in an atmosphere of 33% O2 and 67% CO2, in this paper denoted the standard oxygen electrode (SOE). Humidification of anode gas was controlled with a bubbler at a set temperature. Electrochemical performance was evaluated with polarization measurements (IeV curves) and electrochemical impedance spectroscopy (EIS), using a Solartron 1285 potentiostat and 1255 frequency response analyzer. EIS measurements are performed from 10 kHz to 10 mHz with a voltage perturbation of 10 mV, if no other information is given. The small scale of the button cell and open cell configuration with impinging gas flow allow for homogeneous conditions throughout the fuel cell. Inlet and outlet conditions for gas composition can be applied to the entire cell and the Nernst equation for inlet gas composition and temperature can be assumed to be valid throughout the entire fuel cell. The gas flow for both anode and cathode are sufficiently high to ensure a stoichiometric surplus by order of magnitude. This is equivalent to a low utilization factor (Uf) for both fuel and oxidant. Post-test analysis of anode and cathode material was performed with scanning electron microscopy with energy dispersive X-ray spectroscopy (SEM-EDS) to detect residual sulfur on the anode and cathode surface. In button cell testing studying the effect of H2S on the anode side, an exposure time of 90 min was found to be sufficient to detect poisoning effects [24]. Similar to this, shortterm testing of H2S at the anode and SO2 at the cathode was performed by adding the contaminant to either anode of cathode gas for 90 min, after which the performance was measured with IeV curves and EIS measurements. The measurements were performed in the order of cell, cathode and anode, which resulted in a longer exposure time of contaminant before measurements over the separate electrodes than the stipulated 90 min for the cell measurements for shortterm contamination tests. During long-term tests the cell performance was measured continuously before, during, and after the test period. After contaminant testing, the cell was regenerated by changing to clean gas and applying a constant load of 100 mA cm2, with the purpose of removing any reversible poisoning effects due to interactions between residual sulfur compounds and electrode or electrolyte. Regeneration was performed overnight for approximately 20 h or until stable performance was reached. Contaminants were added to a gas stream of pre-mixed oxidant or fuel gas (AGA gas) stream in N2 carrier gas after humidification to avoid scrubbing of sulfur contaminant. To maintain a constant flow rate to the cell the flow was balanced
Please cite this article in press as: Rexed I, et al., Effect of sulfur contaminants on MCFC performance, International Journal of Hydrogen Energy (2014), http://dx.doi.org/10.1016/j.ijhydene.2014.03.068
4
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y x x x ( 2 0 1 4 ) 1 e9
with a second addition of N2 to compensate for the change in contaminant þ carrier gas flow when the contaminant level is changed or when clean gas was used. The SO2 poisoning effect was measured with oxidant gas consisting of 15% O2, 30% CO2 and 55% N2 diluted with 90 ppm SO2 in N2 carrier gas and the anodic gas composition consisted of 29% H2, 5% CO2 and 66% N2 humidified at 40 C. The resulting compositions of oxidant gas was 14% O2, 27% CO2 and 59% N2 for long-term test with 8 ppm of SO2. The flow rates of dry gases for the anode and cathode was 150 and 110 mL/min, respectively. The composition of oxidant gas for the short-term test with up to 24 ppm of SO2 was 11% O2, 23% CO2 and 66%, with a flow rate of dry gases for both the anode and cathode of 200 mL/min. The effect of poisoning from H2S at concentrations in the range of 1e4 ppm was tested by adding 350 ppm H2S in N2 carrier gas to the anode gas for two different anode gas compositions. The first anodic gas composition consisted of 75% H2, 19% CO2 and 6% N2 and was humidified at 60 C. The second anodic gas composition consisted of 29% H2, 5% CO2 and 66% N2, and was humidified at 60 C. The oxidant gas consisted of 15% O2, 30% CO2 and 55% N2. The total flow rate of dry gases for both anode and cathode was 150 mL/min.
Results e discussion In a first set of experiments, the cell was exposed to 8 ppm of SO2 in the oxidant gas for 100 h. The cell was kept at a load of 100 mA cm2 during the test period. In order to accelerate any effect on the anode side the fuel gas composition used was low in H2 (29%) and also the humidification temperature was low (40 C). The performance was measured continuously
Fig. 1 e a) Cell voltage, b) anode voltage vs. SOE and c) cathode voltage vs. SOE at different current density with 8 ppm SO2 in oxidant gas.
with IeV curves before, during, and after the addition of SO2. Fig. 1 shows the cell voltage together with anode and cathode measured separately; at OCV and at 100, 200 and 300 mA cm2. The degradation in cell voltage (Fig. 1(a)) can be seen to progress during the extent of the time of exposure to SO2, with a stronger effect at higher current density. When the oxidant gas is switched to clean gas the cell voltage is only regenerated to a small extent, despite operating with in such conditions for more than 100 h. The regeneration of the cell voltage is smaller than expected from studies of low concentrations of H2S in fuel gas, however, low hydrogen and low water content has been shown to effectively reduce the voltage regeneration [20]. In addition, the anode and cathode potentials vs. SOE were measured in parallel with the cell voltage. The effect of SO2 in oxidant gas is clearly seen on the anode (Fig. 1(b)), and on the cathode (Fig. 1(c)). The anodic potential shows a larger effect of regeneration than what can be seen for the cell, which is explained by the cathode potential which shows a continuously decreasing trend at all current densities even after changing to clean gas. In Fig. 2 it can be seen that the cell polarization during the 100 h of exposure to 8 ppm SO2 increases with exposure time to SO2 and is more pronounced at higher current density. The cathodic contribution to the voltage degradation is increasing linearly with exposure time, while the anodic contribution is accelerated with high current density and longer exposure time. The effect on the anodic potential clearly indicates a poisoning effect, and a transfer of sulfur from cathode to anode. This is assumed to progress through dissolution of SO2 and formation of sulfate ions in the electrolyte (reaction 5), which migrates to the anode side where they are converted to H2S. The increase in cathodic polarization is indicative of a progressive increase in the ohmic resistance, as will be discussed in the following section. Two possible causes for the increase in resistance are electrolyte loss by replacement of carbonate with sulfate ions in the electrolyte, which may lower conductivity, and the formation of a corrosion layer at
Fig. 2 e Polarization curves showing overvoltage (DE) of cell, anode and cathode before contamination test and after 20 h, 45 h, 75, and 100 h exposure to 8 ppm SO2.
Please cite this article in press as: Rexed I, et al., Effect of sulfur contaminants on MCFC performance, International Journal of Hydrogen Energy (2014), http://dx.doi.org/10.1016/j.ijhydene.2014.03.068
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y x x x ( 2 0 1 4 ) 1 e9
5
Fig. 3 e Nyquist diagram of cell impedance at a) OCV and b) 100 mA cmL2, after different exposure time to 8 ppm SO2 in oxidant gas. (B 2.5 kHz, D 250 Hz, , 25 Hz >2.5 Hz).
the electrode-current collector interface, which would increase the contact resistance. Although this does not satisfyingly explain the continued decrease in cathodic voltage after the oxidant has been changed to clean gas, the difference in pore structure and the faster kinetics of hydrogen oxidation compared to the oxygen reduction reaction makes the cathode more sensitive to changes in the electrolyte distribution due to changed wetting properties or electrolyte loss, and it
can be expected that the cathode performance is more strongly affected than the anode performance if electrode degradation is contributing to the increase in resistance. A continuous accelerated electrolyte loss could also prove to be a major cause of degradation for long-term exposure to SO2. The result from impedance measurements of the cell is presented in the Nyquist diagram in Fig. 3. The impedance curves are shifted to the right with longer exposure times of
Fig. 4 e Nyquist diagram of a) anode, and b) cathode impedance at OCV, after different exposure time to 8 ppm SO2 in oxidant gas. (B 2.5 kHz, D 250 Hz, , 25 Hz >2.5 Hz). Please cite this article in press as: Rexed I, et al., Effect of sulfur contaminants on MCFC performance, International Journal of Hydrogen Energy (2014), http://dx.doi.org/10.1016/j.ijhydene.2014.03.068
6
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y x x x ( 2 0 1 4 ) 1 e9
SO2, which indicate an increased cell resistance; although at low frequency the scattering of impedance data complicates the analysis of the cell impedance. The ohmic resistance can be identified by the intercept with the real axis at high frequency in the Nyquist diagram. During 100 h of exposure to SO2 in oxidant gas, the ohmic resistance increases from 0.134 U to 0.161 U after 45 h, and to 0.173 U after 100 h of exposure. The poisoning effect of the anode by reaction (9) is assumed to be better studied under load, as the formation of NixSy has a standard potential more positive than the anodic OCV. Furthermore, poisoning effects on the anode are more pronounced under load, as seen in Fig. 2. However, no significant difference in terms of poisoning effects can be seen in the cell impedance at OCV (Fig. 3(a)) compared to 100 mA cm2 (Fig. 3(b)). The cell impedance at 100 mA cm2 shows a decreased mass transfer arc compared to the impedance at OCV, which is due to the water production at the anode under current load. In the corresponding Nyquist plot for the anode (Fig. 4(a)) there is no visible increase in ohmic resistance during the measurement period. The anodic impedance data show the same scattering at low frequency as for the cell impedance, why it is difficult to analyze any change in mass transfer properties of the anode. An increase in this region was expected as it has been attributed to a competing reaction of H2S and H2 for hydrogen oxidation sites at the nickel surface when testing the effect of H2S in low concentrations in the fuel gas [24]. The cathode impedance curve (Fig. 4(b)) is shifted to the right, which indicates a higher resistance, although there is no high frequency intercept for an accurate reading of the ohmic resistance. However, the shift in the cathodic impedance curve is most likely corresponding to the increase in cell resistance and the cause of the cathodic voltage degradation during the long-term contamination testing (Fig. 1(c)). The high frequency loop in the cathode Nyquist diagram is an artifact stemming from the reference electrode and is not of relevance for the fuel cell performance.
Fig. 5 e Polarization curves showing overvoltage (DE) of cell, anode and cathode with different concentration of SO2 added in oxidant gas for 90 min. Regeneration with clean gas 20 h.
The short-term exposure to contaminant tests were performed with increased levels of SO2 (2, 4, 6, 8, 12, 18 and 24 ppm). Polarization curves are shown in Fig. 5. The effect on cell performance of SO2 in concentration lower than 12 ppm for as short time as 90 min was negligible, and thus these data are omitted. When the SO2 concentration was increased to 18 ppm and 24 ppm at a current density of 230 mA cm2 the voltage drop compared to clean gas was DEcell ¼ 15 mV and 30 mV, respectively (Fig. 5). When comparing the IeV curves for the separate electrodes, it must first be mentioned that the reference electrode measurements contributes to a small difference between the summation of the anodic and cathodic voltage and the measured cell voltage. A possible explanation to this is the different exposure time for cell and electrode measurements, as explained in the experimental procedure. However, it can be concluded that the increase in anode polarization at 230 mA cm2 with 18 ppm and 24 ppm of SO2 compared to clean gas is larger than the increase in cathode polarization; DEan ¼ þ15 mV and þ30 mV compared to DEcat ¼ 7 mV and 15 mV (Fig. 5). The effect on the anodic polarization is increasingly nonlinear above 100 mA cm2 with 24 ppm of SO2 in oxidant gas, which can be interpreted as the result of the concentration- and voltage dependent formation of NixSy by reaction (9). After regeneration with clean gas for 20 h after exposure to 24 ppm of SO2 for 90 min no remaining poisoning effect can be seen for the cell (Fig. 5). Fig. 6 shows the cell voltage as a function of time during the exposure to 12, 18 and 24 ppm of SO2 and the following regeneration period. It can be seen that the rate of voltage decay is increasing with higher concentration of SO2, but also that the voltage decay continues after changing to clean gas for approximately the same time as the contamination period, before the voltage is recuperated to a steady state value. This is interpreted to be an effect of accumulated sulfate in the electrolyte which continues to poison the anode by conversion to H2S according to the reversed reaction (6). The poisoning effect on the anode is eventually reduced, as the concentration of sulfate in the electrolyte decreases and no
Fig. 6 e Cell voltage at constant load 100 mA cmL2 during exposure to 12, 18 and 24 ppm of SO2 in oxidant gas for 90 min, and the following regeneration with clean gas.
Please cite this article in press as: Rexed I, et al., Effect of sulfur contaminants on MCFC performance, International Journal of Hydrogen Energy (2014), http://dx.doi.org/10.1016/j.ijhydene.2014.03.068
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y x x x ( 2 0 1 4 ) 1 e9
7
Fig. 7 e Nyquist diagram of cell impedance at a) OCV, and b) 100 mA cmL2, for different concentration of SO2 added in oxidant gas for 90 min. (B 2.5 kHz, D 250 Hz, , 25 Hz >2.5 Hz).
more H2S is produced. Up to 12 ppm SO2 the rate of sulfate formation is not sufficient to raise the concentration of H2S at the anode above the level where a poisoning effect is visible, and thus no such effect is seen for the following regeneration with clean gas. In order to verify the concentration of H2S at the anode, gas analysis of anode exhaust is necessary, however, that was not available for this work. The Nyquist diagram of the cell impedance with 12e24 ppm of SO2 in oxidant gas (Fig. 7) shows a slight increase in the cell resistance by a shift in the high frequency intercept. The low frequency region, especially under load, is difficult to analyze due to scattering of the impedance data in this region. No significant difference can be seen between cell impedance at OCV and under load. The anode impedance (Fig. 8(a)) shows no increase of ohmic resistance, similar to results of the long-term exposure test. However, the cathode impedance (Fig. 8(b)) shows a slight shift in the impedance curve, although less than what was found after the long-term exposure test. The result of the short-term contamination tests differs from that of long term exposure tests both in the concentration at which the addition of SO2 in oxidant gas has a clearly negative effect on cell performance, and in terms of the effect of cell regeneration. In the first perspective, it is expected that extended exposure time results in more severe poisoning, as sulfate dissolved into the electrolyte would accumulate in the electrolyte and create an increasingly more harmful atmosphere at the anode. The anode polarization was effectively regenerated with clean gas, both after long-term and shortterm poisoning. This is in agreement with what has been found in studies of poisoning of MCFC anodes by H2S in biogas
simulated fuel gas [13,24]. However, in contrast to long-term contamination tests, no effect was seen on the cathode after short-term contamination and regeneration with clean gas. The effect of adding 1e4 ppm of H2S to the anode gas was tested with different anode gas compositions with the purpose of comparing the effect of SO2 at the cathode and H2S at the anode. The poisoning effect was measured for one composition with high concentration of H2 (75% H2, 19% CO2 and 6% N2), and one composition which was diluted with nitrogen (29% H2, 5% CO2 and 66% N2). The anode gas was humidified at 60 C. The H2 concentration in the anode gas affects the poisoning of the anode by changing the H2/H2S partial pressure ratio, which is important for the competing reaction (8). A low H2 concentration will simplify the formation of nickel sulfide and accelerate the degradation of the anode. The polarization curves are shown in Fig. 9. In contamination tests with the high H2 composition of anode gas, no significant effect can be seen on the polarization of the cell in the concentrations tested. In order to accelerate the poisoning effect at low concentration of H2S, the contaminant gas was added in a diluted anodic gas to the fuel cell. When using the anode gas with low concentration of H2 there is an increase in cell polarization for 2 and 4 ppm of H2S. The voltage degradation is further highlighted in the anodic polarization curves. For concentrations of H2S higher than 1 ppm, anode polarization is significant, and the poisoning effect is increased with higher concentration of H2S and higher current density. The cathodic polarization curves show no effect on polarization. The increased effect of H2S at the anode under load compared to OCV is discussed in Ref. [20]; the poisoning at
Please cite this article in press as: Rexed I, et al., Effect of sulfur contaminants on MCFC performance, International Journal of Hydrogen Energy (2014), http://dx.doi.org/10.1016/j.ijhydene.2014.03.068
8
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y x x x ( 2 0 1 4 ) 1 e9
Fig. 8 e Nyquist diagram of a) anode, and b) cathode impedance at OCV, for different concentration of SO2 added in oxidant gas for 90 min. (B 2.5 kHz, D 250 Hz, , 25 Hz >2.5 Hz).
OCV or low polarization is caused only by adsorption and chemisorption on the nickel surface, but at high current density nickel sulfide may be produced in the electrode reaction, which can block the anode surface for hydrogen oxidation. The effect on the anodic polarization of H2S poisoning is similar to that of short-term tests with SO2 at the cathode. This supports the theory of sulfur transfer from the cathode to
the anode in the electrolyte and subsequent release of H2S at the anode as being the main cause of anode poisoning. The results from post-test analysis by SEM-EDS did not show any residual sulfur at the anode or the cathode surface, although a small poisoning effect was detected by electrochemical analysis of the fuel cell performance. In a study with 95 ppm of H2S in fuel gas [20], sulfur compounds was identified with SEM-EDS analysis, however, the concentration of H2S at the anode in this study was most likely too low to leave detectable amounts at the anode surface after regeneration.
Conclusions
Fig. 9 e Polarization curves showing overvoltage (DE) of cell, anode and cathode with clean gas and after 90 min exposure to 1e4 ppm H2S in anode gas. Regeneration is performed with clean gas for 20 h. Anode gas composition: High H2 (75% H2, 19% CO2, 6% N2), low H2 (29% H2, 5% CO2, 66% N2) humidified at 60 C.
SO2 added to the cathode gives an immediate and significant effect on the anode, which indicates a transfer mechanism of sulfur from anode to cathode. This is assumed to occur by SO2 dissolving into the electrolyte as sulfate ions which migrate to the cathode to be released as H2S, and by poisoning of the anode catalyst by interactions between nickel and H2S. Longterm exposure to 8 ppm SO2 at the cathode resulted in increased polarization of both anode and cathode, with a stronger effect at high current density. The cathode polarization increased linearly while the anode polarization was accelerating with high current density and longer exposure time. This indicates anode side poisoning by H2S. The effect of the short-term contamination tests on anode polarization is visible for concentrations higher than 12 ppm SO2. The short exposure time shows that the transfer mechanism of sulfur from the cathode to the anode is fast, although no detailed data on transfer kinetics can be derived from this study. The poisoning effect is increasing with exposure time
Please cite this article in press as: Rexed I, et al., Effect of sulfur contaminants on MCFC performance, International Journal of Hydrogen Energy (2014), http://dx.doi.org/10.1016/j.ijhydene.2014.03.068
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y x x x ( 2 0 1 4 ) 1 e9
and a delayed poisoning effect is seen after clean gas was introduced at the cathode side for regeneration. This indicates an accumulation of sulfate in the electrolyte, which results in an increase of H2S concentration at the anode with time. The anode poisoning effects were largely reversible by regeneration with clean gas. The cathode, however, experienced irreversible polarization effects after 100 h exposure to 8 ppm SO2. The increase in cell resistance after long-term exposure was attributed to the cathode. The cause may be either carbonate being replaced by sulfate by reaction with SO2 dissolved in the electrolyte, or the build-up of a corrosion layer at the electrode-current collector interface. The result shows that SO2 in oxidant gas is of great concern for the use of flue gas to feed the cathode of MCFCs. Poisoning of the anode was further supported by comparing the effect on the anode polarization from adding SO2 at the cathode with the effect of adding H2S at the anode, which was similar in the two experiments. However, impedance measurements did not reveal detailed information about the predominant poisoning mechanism, partly due to a reoccurring disturbance in the low frequency region in which mass transfer is studied.
references
[1] Farooque M, Maru H. Carbonate fuel cells: milliwatts to megawatts. J Power Sources 2006;160:827e34. [2] Clarke S, Dicks A, Pointon K, Smith T, Swann A. Catalytic aspects of the steam reforming of hydrocarbons in internal reforming fuel cells. Catal Today 1997;38:411e23. [3] Watanabe T, Izaki Y, Mugikura Y, Morita H, Yoshikawa M, Kawase M, et al. Applicability of molten carbonate fuel cells to various fuels. J Power Sources 2006;160:868e71. [4] Sanchez D, Monje B, Chacartegui R, Campanari S. Potential of molten carbonate fuel cells to enhance the performance of CHP plants in sewage treatment facilities. Int J Hydrogen Energy 2013;38:394e405. [5] Cigolotti V, McPhail S, Moreno A. Nonconventional fuels for high-temperature fuel cells: status and issues. J Fuel Cell Sci Technol 2009;6:021311. [6] Selman JR. Research, development, and demonstration of molten carbonate fuel cell Systems. In: Blomen LJMJ, Mugerwa MN, editors. Fuel cell systems. New York: Plenum Press; 1993. pp. 345e463. [7] Varbanov P, Klemes J, Shah R, Shihn H. Power cycle integration and efficiency increase of molten carbonate fuel cell systems. J Fuel Cell Sci Technol 2006;3:375e83.
9
[8] Massardo A, Bosio B. Assessment of molten carbonate fuel cell models and integration with gas and steam cycles. J Eng Gas Turbines Power-Trans ASME 2002;124:103e9. [9] Desideri U, Proietti S, Sdringola P, Cinti G, Curbis F. MCFCbased CO2 capture system for small scale CHP plants. Int J Hydrogen Energy 2012;37:19295e303. [10] Chiesa P, Campanari S, Manzolini G. CO2 cryogenic separation from combined cycles integrated with molten carbonate fuel cells. Int J Hydrogen Energy 2011;36:10355e65. [11] Campanari S, Chiesa P, Manzolini G. CO2 capture from combined cycles integrated with molten carbonate fuel cells. Int J Greenh Gas Control 2010;4:441e51. [12] Discepoli G, Cinti G, Desideri U, Penchini D, Proietti S. Carbon capture with molten carbonate fuel cells: experimental tests and fuel cell performance assessment. Int J Greenh Gas Control 2012;9:372e84. [13] Ciccoli R, Cigolotti V, Lo Presti R, Massi E, McPhail SJ, Monteleone G, et al. Molten carbonate fuel cells fed with biogas: combating H2S. Waste Manag; 2010:1018e24. [14] Weaver D, Winnick J. Sulfation of the molten carbonate fuelcell anode. J Electrochem Soc 1989;136:1679e86. [15] Di Giulio N, Bosio B, Cigolotti V, Nam SW. Experimental and theoretical analysis of H2S effects on MCFCs. Int J Hydrogen Energy 2012;37:19329e36. [16] Cigolotti V, McPhail S, Moreno A, Yoon S, Han J, Nam S, et al. MCFC fed with biogas: experimental investigation of sulphur poisoning using impedance spectroscopy. Int J Hydrogen Energy 2011;36:10311e8. [17] Hernandez S, Scarpa F, Fino D, Conti R. Biogas purification for MCFC application. Int J Hydrogen Energy 2011;36:8112e8. [18] Kawase M, Mugikura Y, Watanabe T. The effects of H2S on electrolyte distribution and cell performance in the molten carbonate fuel cell. J Electrochem Soc 2000;147:1240e4. [19] Uchida I, Ohuchi S, Nishina T. Kinetic-studies of the effects of H2S impurity on hydrogen oxidation in molten (Li þ K)CO3. J Electroanal Chem 1994;369:161e8. [20] Zaza F, Paoletti C, LoPresti R, Simonetti E, Pasquali M. Studies on sulfur poisoning and development of advanced anodic materials for waste-to-energy fuel cells applications. J Power Sources 2010;195:4043e50. [21] Vogel WM, Smith SW. The effect of sulfur on the anodic H2 (Ni) electrode in fused Li2CO3eK2CO3 at 650 C. J Electrochem Soc 1982;129:1441e5. [22] Devianto H, Yoon SP, Nam SW, Han J, Lim TH. The effect of a ceria coating on the H2S tolerance of a molten carbonate fuel cell. J Power Sources 2006;159:1147e52. [23] Lagergren C. Electrochemical performance of porous MCFC cathodes, Ph.D. thesis. Stockholm: KTH; 1997. [24] Devianto H, Simonetti E, McPhail S, Zaza F, Cigolotti V, Paoletti C, et al. Electrochemical impedance study of the poisoning behaviour of Ni-based anodes at low concentrations of H2S in an MCFC. Int J Hydrogen Energy 2012;37:19312e8.
Please cite this article in press as: Rexed I, et al., Effect of sulfur contaminants on MCFC performance, International Journal of Hydrogen Energy (2014), http://dx.doi.org/10.1016/j.ijhydene.2014.03.068