Top Catal DOI 10.1007/s11244-015-0486-6

ORIGINAL PAPER

Sulfur Tolerance of Au–Mo–Ni/GDC SOFC Anodes Under Various CH4 Internal Steam Reforming Conditions Ch. Neofytidis1,2 • M. Athanasiou1,2 • S. G. Neophytides1 • D. K. Niakolas1

Ó Springer Science+Business Media New York 2015

Abstract The present work refers to the short communication of a first series of results on how Au and/or Mo addition can affect the stability of modified Ni/GDC anodes for the reaction of internal CH4 steam reforming, in the presence of H2S. Specifically, it is shown that Ni/GDC is stable in the presence of 10 ppm H2S, but only in the case where 100 vol% of H2 is the anode feed. In the case where CH4 and H2O (diluted in helium carrier gas) comprise the anode feed, then at ratios equal to S/C = 2 or S/C = 0.13 the performance of Ni/GDC shows severe degradation, while the Au–Mo–Ni/GDC anode has the best and most stable electrocatalytic behavior. Finally, there is a first attempt to investigate the effect of the electrocatalyst’s loading on sulfur tolerance. Keywords SOFCs  Ni/GDC anodes  Au and/or Mo doping  Degradation  Sulfur poisoning and tolerance  CH4 internal steam reforming (ISR)

1 Introduction Solid oxide fuel cells (SOFCs) are a viable high temperature fuel cell technology, which converts chemical energy directly into electricity with high efficiencies [1, 2]. The main advantages of SOFCs can be classified to: (i) the

& D. K. Niakolas [email protected] 1

Foundation for Research and Technology, Institute of Chemical Engineering Sciences (FORTH/ICE-HT), Stadiou str. Platani, Rion, 26504 Patras, Greece

2

Department of Chemical Engineering, University of Patras, Caratheodory 1 St, 26504 Patras, Greece

enhanced reaction kinetics due to the wide range of intermediate and high operating temperature (500–1000 °C) (ii) the higher outlet temperature (ease of waste heat utilization) and (iii) their fuel feed flexibility. Furthermore, SOFCs have the ability to operate directly with various hydrocarbon (H/C) fuels, without the need of an external reformer. This type of operation, known as internal reforming (I.R.), is based on the catalytic activity of the fuel exposed anode. It improves and simplifies the system integration, whereas it provides the additional benefit that part of the heat generated in the cell, by electrochemical reactions and ohmic heating, is directly used for the endothermic reforming reactions [3, 4]. Specifically, the latter property is one of the key competitive advantages for the commercialization of this technology, in both stationary and mobile applications, since it provides the ability to decrease the total operation cost of the SOFC systems, by the direct use of economically derived fuels and/or by avoiding the installation of additional high-grade purifying apparatus [4–7]. The fuel flexibility of SOFCs in combination with the growing need to reach real life applications and the commercialization standards have urged the research community to deal with more realistic SOFC operating conditions and especially to use real fuels and dynamic operation. This approach suggests that several fuels can be considered as potential realistic candidates for future SOFCs. Some typical examples are: (i) Natural Gas (N.G.) (ii) biogas, a mixture of mainly CH4 and CO2, or (iii) bio syngas that contains H2 and CO from the gasification of biomass [8]. The drawback in all of the above possible fuels is that, apart from various carbon components, they also comprise sulfur impurities in a wide range of concentrations, which depends on the main fuel source (e.g. Table 1) and the extent of processing/desulfurization. Therefore, even

123

Top Catal Table 1 Typical composition and H2S concentration of some fuel sources [7, 17] Fuel type

Typical composition

H2S concentration

Coal syngas

H2, CO, CO2, H2O, N2

100–300 ppm

Biogas

H2, CO, CO2, CH4 H2O, N2

50–200 ppm

Sour (natural) gas

H2, CO2, N2, CH4, C2H6

[1 %

though the I.R. process of H/C-based fuels appears more attractive compared to the mainstream technology, which is based on using hydrogen as the fuel, there still remain several critical issues to be solved. One of the most important is the anode‘s degradation, which is caused by unfavorable reactions and poisoning of the current State of the Art (SoA) Ni-based anode materials. More precisely, nickel-based electrodes suffer from several degradation factors. For example: (i) they are prone to carbon deposition and this is because Ni catalyzes the formation of carbon deposits from H/Cs under reducing conditions [9], (ii) the deposited carbon further induces Ni corrosion (dusting), leading to loss of conductivity and to direct structural damage of the anode [10, 11], (iii) they present poor redox stability, (iv) nickel has the tendency to form agglomerates after prolonged fuel cell operation due to either thermal and/or overpotential sintering [12], and (v) they exhibit low tolerance to poisoning from sulfur species [13]. The deactivation from sulfur is one of the most detrimental reasons for the SOFCs failure. Sulfur usually appears in the form of H2S and though there are desulfurization processes of the fuels prior the feeding in the cell systems [14], the remaining H2S concentrations typically vary in the range between 0.1 and 15 ppm [8, 15, 16], which have been proven critical for the cell‘s performance and lifetime. Moreover, in regards to the current technological level and the available real fuels it cannot be decided yet, whether it is more cost efficient to remove completely sulfur from the fuel or to develop tolerant anodes for a specific application region that is generally considered up to 15 ppm. Therefore, worldwide ongoing research effort focuses primarily on the development of efficient and tolerant anode materials against carbon deposition and/or sulfur poisoning in case of a desulfurization system fault [17–19]. The improvement of the electrodes‘ performance and tolerance against sulfur poisoning has been mainly approached by several modifications of the electrocatalyst structure/composition and there has been significant progress on the development of SOFC anode materials, summarized in several published reviews [4, 7, 20–22]. In brief, some general requirements that SOFC anode materials should possess are: (i) excellent catalytic activity towards electro-oxidation of the possible fuels, (ii) high electronic conductivity, (iii) adequate ionic conductivity so

123

that the active sites can be extended beyond the electrolyte–anode (TPB) interfaces, (iv) suitable porosity that allows the fuel molecules to easily flow towards and the reaction products to be removed from the active sites, (v) robust thermal and mechanical stability, (vi) effectual durability and compatibility with other SOFC components during cell fabrication and operation, (vii) low cost and facile manufacturing [7, 21]. Ni/YSZ for example is one of the most examined cermets and the general observation for it is that short term exposure towards H2S in the ppm range always results in acute voltage drop, which can be reversible or not. This degradation is primarily ascribed to the dissociative adsorption of H2S towards the formation of nickel sulfide, which is catalytically inactive for the H2 electrochemical oxidation reaction (Eq. 1) [17, 22, 23]. H2ðgÞ þ O2 ! H2ðgÞ OðgÞ þ 2e

ð1Þ

Furthermore, studies using Ni/GDC [24] have shown that this composition is another good candidate for operating in the presence of H2S. It has been observed that in H2S-containing H2 fuels the degradation in performance for the H2 electro-oxidation is substantially smaller compared to that on Ni/YSZ anodes. Recent studies on NiO/GDC-based materials have yielded [25, 26] ternary Ni–Au–Mo/GDC cermets, which are under investigation regarding their carbon and sulfur behavior/tolerance, both as catalysts and as anode electrodes, as well as their fabrication feasibility and commercialization in the SOFCs field. The present work refers to the short communication of a first series of obtained results on how Au and/or Mo addition can affect the stability of modified Ni/GDC anodes for the reaction of internal CH4 steam reforming, in the presence of H2S. In particular, the experiments comprise electrochemical stability measurements in the case where 100 vol% of H2 is the anode feed and for the ISR of CH4 at ratios equal to S/C = 2 or S/C = 0.13, all in the presence of 10 ppm H2S. Finally, there is a first attempt to investigate the effect of the electrocatalyst’s loading on the sulfur tolerance.

2 Experimental 2.1 Preparation of Binary Au–NiO/GDC, Mo–NiO/ GDC and Ternary Au–Mo–NiO/GDC Anode Cermets Binary Au–NiO/GDC and Mo–NiO/GDC anode powders with nominal loading 3 wt% Au or Mo, respectively, were prepared via the deposition–precipitation (D.P.) method. The ternary Au–Mo–NiO/GDC anode powder was

Top Catal

prepared by applying the deposition–coprecipitation (D.CP.) method, while the nominal loading of Au and Mo remained at 3 wt%, respectively. After filtering, each precipitate was dried at 110 °C for 24 h. All dried powders were calcined at 600 °C for 90 min. Further details regarding the preparation of the modified powders can be found in already published studies [25, 27]. 2.2 Preparation of Solid Oxide Electrolyte Supported Cells The solid oxide electrode assemblies (SOEAs) comprised circular shaped planar electrolyte-supported SOFC membranes manufactured by Kerafol with a diameter of 25 mm. The mechanical support consisted of *300 lm thick 8YSZ electrolyte. There are two series of cells that have been examined. In the first one, the anode was applied by means of stepwise addition of droplets of a slurry with a precision micropipette. In the second series of cells the anode functional layer (AFL) was deposited by using the screen printing technique. For the former method the anode slurry contained an amount of NiO/GDC anode powder (Marion Technologies) (modified with Au, Mo and Au–Mo), terpineol (Sigma-Aldrich) as the dispersant, PVB (polyvinylbutyral, Sigma-Aldrich) as binder and EMSURE isopropanol (MERCK) as solvent. For the screen printing method the prepared paste comprised an amount of commercial NiO/GDC (Marion Technologies) anode powder (modified with Au, Mo and Au–Mo), terpineol (SigmaAldrich) as the dispersant and PVB (polyvinyl butyral, Sigma-Aldrich) as binder. After the deposition of the slurry (with the micropipette method) the anode assembly was dried and finally sintered at 1250 °C for 5 h with a heating and cooling ramp rate of 2 °C/min. Alternatively, the cells where the anode was deposited by screen printing, were sintered at 1150 °C for 2 h with a heating and cooling ramp rate of 2 °C/min. The loading of the anodes, for the cells prepared with the micropipette method, lied in the range between 15 and 25 mg/cm2 ‘‘high-AFL’’ with 1.7 cm2 geometric surface area, while in those cells prepared with screen printing the anode loading was 5 mg/cm2 ‘‘lowAFL’’ with 1.7 cm2 geometric surface area. The cathode side comprised porous layers of YSZ/LSM and LSM, commercially available from Fuel Cell Materials (FCM), which were applied by means of screen printing and calcined at 1150 °C for 2 h. Finally, the planar fuel cell system was attached on a YSZ tube and sealed airtight by using glass sealing material (KeraGlas), purchased from Kerafol. 2.3 Electrochemical Testing The electrochemical experiments were carried out on galvanostatic mode at 850 °C under various H2O/CH4 (S/C)

ratios and the addition of 10 ppm H2S, with a flow rate of 10 cm3/min H2S/He. In particular, the anode was exposed to three different feed conditions: S/C = 2 (H2O = 6 cm3/ min, CH4 = 3 cm3/min), S/C = 0.13 (H2O = 6 cm3/min, CH4 = 46 cm3/min) and pure H2. The total fuel supply in all cases was, Ftotal = 100 cm3/min and in the case of internal CH4 steam reforming the reactants were diluted in Helium carrier gas. Further details are reported in the corresponding caption figures. H2O was introduced in the SOFC reactor in the form of steam. Prior its evaporation liquid water was circulated in the system by means of a pressurized liquid H2O vessel, which is connected with a liquid water mass flow controller. The cathode compartment was fed with pure Oxygen. Both cathodic and anodic flow rates were adjusted at 100 cm3/min and all lines and valves of the experimental setup were heated at 160 °C to prevent water condensation. Pt-meshes (99.9 % purity) were used as current collectors on the anode and cathode side, which were evenly pressed on the electrodes through spring force so as to achieve better current collection. The short term stability tests and electrochemical impedance spectra (EIS) EIS recording were carried out by using an AUTOLAB potentiostat/galvanonstat model 84693. In particular, the electrochemical effect of H2S on the ohmic (Rohm) and polarization resistance (Rpol) was measured by recording the EIS on a galvanostatic mode at 28 mA/cm2 with an amplitude of 15 mA in the frequency range between 100 kHz and 20 mHz.

3 Results and Discussion 3.1 Characterization Figure 1 shows SEM images of a binary 3Au–NiO/GDC anode layer, prepared with the micropipette method and of a ternary 3Au–3Mo–NiO/GDC AFL, prepared by screen printing. The images were collected in a HR-SEM (Zeiss SUPRA 35VP). The micrograph of the top side in Fig. 1a shows a porous ‘‘high-AFL’’ electrode with the size of particles in the order of 1 lm. The cross section micrograph (Fig. 1b) shows that the thickness of the ‘‘high-AFL’’ electrode is in the order of 80 lm. The specific cell was not broken properly during the presented cross section SEM analysis, resulting to the loss of the electrolyte. Therefore, the electrolyte layer is not depicted in Fig. 1b and the analysis is focused on the electrode. The observed thin layer of *15 lm, is in first contact with the electrolyte and it is an adhesion layer of the same material with the electrode, which was deposited first for achieving a good adhesion and connectivity between the electrolyte and the electrode. The micrograph of the top side in Fig. 1c shows a porous ‘‘low-AFL’’ Au–Mo–NiO/GDC electrode with the

123

Top Catal

Fig. 1 a SEM of the top side of a 3Au–NiO/GDC ‘‘high-AFL’’ anode layer [44]. b SEM cross section perpendicular to the anode/YSZ electrolyte of a ‘‘high-AFL’’ [44]. c SEM of the top side of a 3Au–

3Mo–NiO/GDC ‘‘low-AFL’’ anode layer. d SEM cross section perpendicular to the anode/YSZ electrolyte of a ‘‘low-AFL’’ 3Au– 3Mo–NiO/GDC

size of particles not higher than 500 nm. The latter difference in the particle size, compared to the ‘‘high-AFL’’ electrode, is mainly attributed to the lower sintering temperature (1150 °C) and most probably to the different type of modified cermet. The cross section micrograph (Fig. 1d) of the same electrode shows a thickness in the order of 20 lm. Finally, it has to be mentioned that for the cells where the AFL was deposited by means of screen printing, much better results have been achieved regarding the adhesion between the anode and the electrolyte, despite the fact that all of them have been calcined at lower temperature. This is mainly attributed to the properties of the components in the prepared paste in combination with its viscosity.

series refers to the ‘‘high-AFL’’ loading cells, prepared with the micropipette method. Another, detail is that the cells were examined sequentially under the various mentioned fuel feeds. Specifically, each cell was first subjected to the reaction feed of 100 vol% H2 plus 10 ppm H2S and the following step comprised an overnight regeneration period at 850 °C, by using a mixture of 30 vol% H2/He. After that step the same cell was subjected to the next reaction feed (H2O/CH4 = 2 plus 10 ppm H2S), followed by a second overnight regeneration period. The final reaction mixture was the one with ratio H2O/CH4 = 0.13 plus 10 ppm H2S, followed by the final reactivation step. All cells were examined by applying the same experimental protocol and the same reaction feeds sequence. The recovery ability together with other details regarding the electrochemical measurements are described below.

3.2 Electrochemical Performance and Stability of Au–Mo–Ni/GDC Anodes in the Presence of 10 ppm H2S In terms of the electrochemical stability tests in the presence of 10 ppm H2S, the experiments comprise two series of measurements. The first deals with the ‘‘low-AFL’’ loading cells, prepared with screen printing. The second

123

3.2.1 Electrochemical Performance and Stability of Au– Mo–Ni/GDC Anodes in H2S/H2 Figure 2 shows the performance of the ‘‘low-AFL’’ cells in the presence of 10 ppm H2S/H2 at 850 °C. Specifically, those that comprised Ni/GDC and 3Au–Ni/GDC as anode

Top Catal

Fig. 2 Stability diagrams of blank and Au–Mo modified ‘‘low-AFL’’ Ni/GDC anodes under 10 ppm H2S/H2. The experiments were performed at T = 850 °C, j = 28 mA/cm2 and Ftotal = 100 cc/min

exhibited lower performance compared to the 3Au–3Mo– Ni/GDC, whereas they also presented a slight degradation step of the operational voltage after some time of exposure. Particularly, for Ni/GDC the degradation step appeared after 60 min of operation, while after that step the performance remained rather stable. The situation was similar for 3Au–Ni/GDC, but the step was less intense compared to that of Ni/GDC and appeared after 150 min of operation. On the other hand, the cell with the ternary anode had a quite stable performance for the examined time period. Figures 3 and 4 show the comparison of the I–V plots and the impedance spectra of each cell, respectively. The

Fig. 3 I–V curves of blank and Au–Mo modified ‘‘low-AFL’’ Ni/ GDC anodes before (t0) and after (tEND) the stability tests under 10 ppm H2S/H2. The experiments were performed at T = 850 °C and Ftotal = 100 cc/min

Fig. 4 EIS measurements of blank and Au–Mo modified ‘‘low-AFL’’ Ni/GDC anodes before (t0) and after (tEND) the stability tests under 10 ppm H2S/H2 at T = 850 °C. All of them have been recorded on galvanostatic mode at 28 mA/cm2

presented measurements have been recorded at two time periods. One at the beginning of the stability (t0) and a second at the end of the stability (tEND). It is also worthy to be mentioned that the performance of the cells under H2, before the introduction of 10 ppm H2S in the reaction feed (not shown here), was similar to those at the beginning of the stability (to). Specifically, the I–V curves (Fig. 3) are linear for all cells indicating that high ohmic resistance, rather related to the electronic resistance of the films, is the prevailing cause of the overpotential and the low performance. The 3Au–Ni/GDC anode had the worst performance at the beginning of the stability, Ni/GDC showed a medium performance, while 3Au–3Mo–Ni/GDC was the best of all yielding the higher current density. The trend on sulfur tolerance did not change by the completion of the stability tests and the ternary anode seems to be the least affected from sulfur. On the other hand, the correlation between Ni/GDC and 3Au–Ni/GDC is inverted, since the former anode seems to be more poisoned than the Aumodified one. It is also noteworthy that the I–V plots of Au–Ni/GDC, after the stability test, and the obtained current densities from the non-modified Ni/GDC are significantly lower as compared to those of the ternary anode, which is an indication for the intrinsic electrocatalytic activity of each material under the specific experimental conditions. The EIS measurements (Fig. 4) have been used to clarify the H2S effect on the ohmic and polarization resistance of each anode functional layer. The ohmic resistance (Rohm) is obtained from the high frequency intercept on the real (Z’) axis and the total resistance of the cell (Rt) is obtained from the low frequency intersect on the

123

Top Catal

real axis. Rohm is the sum of ohmic resistances comprising the contribution from (i) the electrolyte, (ii) the electrode ohmic resistance, (iii) the current collector of the electrode and (iv) the ohmic contribution from the connecting wires. Rt is ascribed to the sum of ohmic and polarization resistances (Rt = Rohm ? Rpol) and from this equation someone can derive the polarization resistance (Rpol). The specific values from the spectra in Fig. 4 are presented in the following Table 2. At this point it has to be specified that in the case of 3Au–Ni/GDC, where the EIS do not cross the (Z’) axis at the high frequencies, the Rohm values have been measured by extrapolating the impedance curve so as to intersect the X-axis. The first remarks from the EIS are deduced from the spectra at the beginning of the stability tests (t0). In particular, the ternary anode exhibits the lowest Rohm and Rpol, compared to the other two cells (see also the values in Table 2), which can be primarily ascribed to the higher electron conductivity and improved structural properties of the electrochemical interface between the Ni–Au–Mo ternary metal phase and GDC of this electrode. On the other hand, the binary 3Au–Ni/GDC cell shows quite high ohmic resistance, which is an indication that the specific anode functional layer is not a good electron conductor and actually worse that the non-modified Ni/GDC. Moreover, the value of Rpol for the Ni/GDC anode is slightly lower than that of 3Au–Ni/GDC. The impedance spectra after the stability tests (tEND) intensify the superiority of 3Au–3Mo–Ni/GDC. Both Rohm and Rpol have been slightly increased compared to those of Ni/GDC and 3Au–Ni/GDC, showing that the poisoning effect of sulfur on the ternary electrode is significantly less. On the other hand, the comparison between the non-modified and the binary cell suggests that the effect of sulfur on Ni/GDC is more intense on the polarization resistance. The fact that the Rpol of the 3Au–Ni/GDC cell did not increase in the same extent as that of Ni/GDC, suggests that despite the fact that Au-modification seems to affect negatively the electron conductivity of the cermet, there is a positive impact on the sulfur tolerance properties of the electrochemical interface between the electrolyte and the anode layer. The increase in Rpol can also be correlated with the

degradation steps in the stability measurements of Fig. 2. These observations can be realized as the outset of degradation on the electrode‘s performance and can be ascribed to the dissociative adsorption of H2S at the nickel sites, which leads to the gradual poisoning of the electrode‘s activity for the electrochemical oxidation of H2 [reaction (1)]. The next set of the herein presented experiments comprises similar stability measurements on the ‘‘high-AFL’’ series of cells and the exhibited performance provides some first results on the effect of the electrode‘s loading on the cells‘ tolerance against sulfur poisoning. Figure 5 depicts the stability diagrams of blank and of Au–Momodified Ni/GDC, in the presence of 10 ppm H2S/H2. In particular, the worst cell was the one that comprised 3 wt% Mo–Ni/GDC as anode, showing the lowest electrocatalytic activity in terms of operating voltage. Unfortunately, due to technical reasons the EIS recording on these ‘‘high-AFL’’ cells was not possible. Specifically, the cermet was very easily detached from the electrolyte and the cells presented quite high values of Rpol. Therefore, the evaluation of sulfur poisoning on their electrochemical properties was not reliable. Nevertheless, all of the ‘‘high-AFL’’ cells presented a more stable behaviour, compared to the ‘‘lowAFL’’ samples, and the main conclusion is that by increasing the loading of the electrocatalyst there is enhancement of the stability against 10 ppm of H2S. This is most probably attributed to the increase of the active sites, which are responsible for the electrochemical oxidation of H2 (Eq. 1). Up today, several studies have been reported in literature regarding sulfur—tolerant anodes and the mechanisms of the induced poisoning in SOFCs. The majority of them examines the effect of a sulfur-containing fuel on the

Table 2 Values of ohmic (Rohm) and polarization (Rpol) resistance of blank and Au–Mo modified ‘‘low-AFL’’ Ni/GDC anodes at the beginning (t0) and after the completion (tEND) of each stability measurement under 10 ppm H2S/H2 at T = 850 °C Sample

Ni/GDC

Time period

t0

tEND

3Au–Ni/GDC

3Au–3Mo–Ni/GDC

t0

tEND

to

tEND

Rohm (ohm cm2)

6.4

8.0

15.0

17.5

4.7

5.5

Rpol (ohm cm2)

5.7

12.3

6.9

9.3

2.4

2.6

123

Fig. 5 Stability diagrams of blank and Au–Mo modified ‘‘high-AFL’’ Ni/GDC anodes under 10 ppm H2S/H2. The experiments were performed at T = 850 °C, j = 23 mA/cm2 and Ftotal = 100 cc/min

Top Catal

performance and durability of SOFC anodes by varying several factors, such as the: (i) operating temperature, (ii) applied current/voltage load, (iii) time exposure, (iv) H2S concentration, (v) the fuel composition and (vi) the examined anodes [7]. The outcome comprises various results on anode deactivation and materials application, summarized in several studies and reviews [7, 17, 20–22, 24], which can be classified into two broad categories mainly depending on the examined type of anode and the applied fuel feed composition. Overall, the research interest has been focused on nickel-based and Ni-free anode materials, which have been examined either under various H2S/H2 feeds or in the presence of H/C-based fuels [7]. In regards to the currently SoA Ni-based electrodes, they have been repeatedly reported to exhibit severe degradation in fuels containing only a few ppm of H2S and this is mainly due to the high vulnerability of nickel to sulfur poisoning. Some of the commonly published remarks, in the case where H2 is the main anode fuel, are that: (i) there is an initial cell voltage drop, which is followed by loss of the SOFC performance, (ii) sulfur poisoning especially on systems operating at intermediate temperatures (typically 700–850 °C) [6] can be irreversible and (iii) degradation at high temperature—SOFCs ([900 °C) can be reversible [28, 29], but not thoroughly especially if the cells have been poisoned by concentrations of H2S higher than 20 ppm at 900 °C [7, 30]. In the case where H2 is the main fuel and H2S is present as an impurity, like in the measurements presented in Figs. 2, 3, 4, and 5, the basic knowledge for possible sulfur poisoning mechanisms on Ni-based anodes has been reported from research groups in which Sasaki et al. [5, 28] and Cheng et al. [21, 31–33] participated. The generally accepted conclusion is that even small amounts of sulfurcontaining compounds result in performance loss and degradation on Ni-based anodes. This phenomenon has been attributed to: (i) strong physical absorption/chemisorption of H2S [(reaction (2)] at surface active sites that lead to reduction of surface area for electrochemical reactions (e.g. H2 oxidation) and/or (ii) sulfidation of the anode materials (particularly of Ni particles), resulting in loss of catalytic activity, conductivity and stability [reactions (3) and (4)] [7]. H2 Sg $ HSads þ Hg=ads $ Sads þ H2g=ads

ð2Þ

Ni þ H2 S $ NiS þ H2

ð3Þ

3Ni þ xH2 S $ Ni3 Sx þ xH2

ð4Þ

In conclusion, sulfur is always expected to adsorb on the nickel surface and along the triple phase boundary (TPB), thus blocking the active sites for fuel oxidation. Moreover, the rapid adsorption of sulfur onto the nickel surface and their interaction are the key steps to understand the initial

sharp degradation of the SOFC performance, which cannot be associated with the formation of nickel sulfides. On the other hand, the successive long-term degradation that leads most of the times to the complete deactivation of the cell is generally more related to the formation of the various nickel sulfide compounds, but the occurring processes are still under investigation [7]. The above conclusions have been mainly derived from studies on Ni/YSZ anodes under H2S/H2 conditions and provide useful general data for the interpretation of some of the H2S poisoning and reactivation effects on Ni-based anodes. However Ni/GDC, which is the main examined material in the presented study, shows to me more tolerant in H2S/H2 fuels and this behavior is primarily ascribed to the properties of ceria. In particular, sulfur poisoning of CeO2 is usually related to the formation of ceria oxysulfide (Ce2O2S) compound, which is considered as an intermediate state under the fuel cell operational reducing conditions. One general remark from the relative literature on Ni/Ceria based anodes, examined in the presence of H2S/ H2 reaction feeds, is that Ce2O2S could be the predominant form of ceria in high ([800 ppm) H2S concentrations, whereas this compound marks the beginning for the failure of the electrode and subsequently of the cell. On the other hand, Ce(SO4)2 could be the main form under lower sulfur amounts that allows ceria to act as a H2S absorbent, without being fully poisoned. However, Luo et al. [34] in their investigation of sulfur poisoning on Pd/Ceria catalysts reported that the various Ce–O–S compounds can easily convert from one form to another, depending on the reaction conditions, suggesting that the relative stability of the sulfide [Ce2O2S] and sulfate [Ce(SO4)2] compounds is related to the PO2 . In particular, the latter species under reducing conditions, similar to those in SOFCs operation, seem to interconvert fairly easily [7, 34]. The above suggestions highlight the complexity in the chemistry of the Ce–O–S system and the necessity for more investigation [7]. The herein examined concentration of 10 ppm H2S, in the case where the majority of the reaction feed is H2, cannot be considered as the main cause for the detrimental poisoning of Ceria (in GDC) and subsequently the degradation in the performance of Ni/GDC. This is because the specific concentration is low and the studied Ni/GDC anodes, either modified or not, are capable to withstand the degradation effects under the applied operating conditions of temperature, applied current and exposure time. Consequently, the observed differentiations can be primarily attributed to the sulfur poisoning of nickel and the induced effects that result from the modification of the cermets. Overall, the stability tests alongside with the I–V and EIS measurements coincide and show that the cell comprising 3Au–3Mo–Ni/GDC as AFL has the best

123

Top Catal

performance at the beginning of the stability and at the same time is less sulfur-poisoned at the end of the test. Despite the fact that the cells were not investigated on how long they would manage to operate until their complete degradation, there is a clear distinction between the three examined cells suggesting that, Au–Mo–Ni/GDC has the potential to be the most tolerant. Another noteworthy remark is that the studied electrolyte supported cells were not prepared with an optimized configuration. The purpose was to investigate solely the modified cermets, as a single anode functional layer, without the addition and interference of other layers (e.g. adhesion and/or current collection layer), which could decrease the Rohm and Rpol values, causing thus an increase of the cells‘ performance. Consequently, the cells‘ configuration justifies up to a certain extent the high values in the resistances, while the differences are mainly attributed to the intrinsic variations in the properties of the examined anode functional layers. Finally, it is obvious that the modification with Au and Mo is better compared to that of Au or Mo solely and this is an extra hint that there is a positive synergy of the doping elements with Ni, which enhances the sulfur tolerance of the electrode under the specific experimental conditions. In regards to the suggestion for the synergy between Au–Mo– Ni, this has been resulted in already published studies from our group [25, 26], which focus on the characterization of these cermets. 3.2.2 Electrochemical Performance and Stability of Au– Mo–Ni/GDC Anodes for the Internal CH4 Steam Reforming Plus 10 ppm H2S In the case where the cells were examined under internal CH4 steam reforming conditions with the ratio S/C = 2 plus 10 ppm H2S (Fig. 6) the results become more clear in terms of the most tolerant anode. Specifically, the cells with Ni/GDC and 3Au–Ni/GDC degraded instantly, while the 3Au–3Mo–Ni/GDC electrode lasted for almost 2.5 h. A noteworthy remark is that the examined conditions are quite different compared to those where the main fuel is H2. In the presence of CH4 and H2O (and generally with H/C-based fuels) the anodes, apart from the already mentioned properties, should also exhibit appreciable electrochemical and catalytic/reforming activity. Taking into consideration that the increased partial pressure of H2 is advantageous for the tolerance against sulfur contaminants, one of the major issues under the ISR operation is the low partial pressure of H2, which is progressively produced along the cell and is readily consumed electrochemically. If hydrogen is the only reactant fuel, then its electrochemical oxidation (Eq. 1) is the main sulfur affected process. In the case of the ISR process the degradation is even faster due to the poisoning of the reforming reaction too [35, 36].

123

Fig. 6 Stability diagrams of blank and Au–Mo modified ‘‘low-AFL’’ Ni/GDC anodes under CH4 steam reforming conditions (S/C = 2) plus 10 ppm H2S. The experiments were performed at T = 850 °C, j = 28 mA/cm2 and Ftotal = 100 cc/min. The reaction feed comprised, 6 vol% H2O, 3 vol% CH4 and 10 ppm H2S in helium

Thus, regarding the internal reforming of H/Cs, all of the discussed reactions; (Eq. 1) and (Eq. 5)–(Eq. 9) should be considered [8] as prone to be hindered [7]. Furthermore, it should be mentioned that small amounts of sulfur can also inhibit carbon formation on Ni-based catalysts. This is ascribed to the preferential binding of S on the surface step sites of nickel, which are active for the reforming and the carbon formation activities, and blocks its nucleation sites [35, 37]. CO oxidation: CO þ 1/2O2 $ CO2

ð5Þ

Water Gas Shift: CO þ H2 O $ CO2 þ H2

ð6Þ

CH4 Dry reforming: CH4 þ CO2 $ 2CO þ 2H2

ð7Þ

CH4 Steam reforming: CH4 þ H2 O $ CO þ 3H2 ð8Þ CH4 Partial oxidation: CH4 þ 1=2O2 $ CO þ 2H2 ð9Þ Therefore, some of the possible reasons that explain the performance of the examined ‘‘low-AFL’’ series in Fig. 6, could be that: (i) the cells that comprise Ni/GDC and 3Au– Ni/GDC can be catalytically active for the CH4 steam reforming reaction but the poisoning effect of H2S is direct and detrimental for the cells‘ catalytic performance, leading to H2 depletion in the reactants feed, (ii) apart from H2S poisoning, the effect of H2O should also be taken into account because its presence induces the anodes‘ re-oxidation and enhances further the deactivation of the cells. The above factors highlight the superiority of the ternary 3Au–3Mo–Ni/GDC anode, which managed to operate for

Top Catal

substantially more time under the same operating conditions. Nevertheless, the poisonous effect of H2S was once again detrimental, leading inevitably to the complete degradation of the cell after 2.5 h of exposure. Another consequence resulting from the complete degradation of the cells under S/C = 2 plus 10 ppm H2S, is that the ‘‘low-AFL’’ Ni/GDC and 3Au–Ni/GDC did not recover during the overnight re-generation step. As a result these cells could no longer operate for the next measurement in sequence under S/C = 0.13 plus 10 ppm pf H2S. Interestingly, the cell comprising 3Au–3Mo–Ni/GDC recovered and managed to operate for approximately 55 min (Fig. 7). In respect to the effect of higher anode loading, Fig. 8 depicts the performance of the same series of cells, under CH4 (ISR) with S/C = 2 and the addition of 10 ppm of H2S. It is clearly observed that 3Mo–Ni/GDC presented the worst performance both in terms of durability and electrocatalytic activity. Ni/GDC showed a bit better performance, but still degraded very fast. 3Au–Ni/GDC showed higher electrocatalytic performance, but the measurement was interrupted due to technical failure reasons. Once again the cell that comprised 3Au–3Mo–Ni/GDC as anode, had the best performance. The latter was quite stable for almost 130 min and remained more active compared to the other cells. The electrocatalytic activity and tolerance of the studied ‘‘high-AFL’’ cells under S/C = 0.13 plus 10 ppm H2S (Fig. 9) is better, compared to that of the ‘‘low-AFL’’ (Fig. 7), showing that the higher-loading electrodes operated for a longer period. Specifically, Ni/GDC remained the worst anode with the lower operating potential and the

Fig. 8 Stability diagrams of blank and Au–Mo modified ‘‘high-AFL’’ Ni/GDC anodes under CH4 steam reforming conditions (S/C = 2) plus 10 ppm H2S at T = 850 °C. The experiments were performed at j = 23 mA/cm2 and Ftotal = 100 cc/min. The reaction feed comprised, 5 vol% H2O, 2.5 vol% CH4 and 10 ppm H2S in helium

Fig. 9 Stability diagrams of blank and Au–Mo modified ‘‘high-AFL’’ Ni/GDC anodes under CH4 steam reforming conditions (S/C = 0.13) plus 10 ppm H2S at T = 850 °C. The experiments were performed at j = 23 mA/cm2 and Ftotal = 100 cc/min. The reaction feed comprised, 5 vol% H2O, 38 vol% CH4 and 10 ppm H2S in helium

Fig. 7 Stability diagram of the ‘‘low-AFL’’ 3Au–3Mo–Ni/GDC anode under CH4 steam reforming conditions (S/C = 0.13) plus 10 ppm H2S at T = 850 °C. The experiment was performed at j = 28 mA/cm2 and Ftotal = 100 cc/min. The reaction feed comprised, 6 vol% H2O, 46 vol% CH4 and 10 ppm H2S in helium

shorter life time. On the other hand, Au–Mo–Ni/GDC lasted for almost 200 min. Au and Mo-modified anodes performed similarly, though Au–Ni/GDC was slightly better, but both of them were not as stable as the ternary anode. The results and trend coming from the ‘‘high-AFL’’ samples are in agreement with those in the ‘‘low-AFL’’ series, proving that the cell with Au–Mo–Ni/GDC anode has the best and most stable performance. Moreover, it is obvious that the durability of the ‘‘high-AFL’’ cells improved significantly compared to the ‘‘low-AFL’’ series, under the same examined conditions, despite the fact that

123

Top Catal

the former cells exhibited worse characteristics (e.g. adhesion of the electrode on the electrolyte). The extended durability is mainly attributed to the effect of the higher electrodes‘ loading. Specifically, there is an increase of the catalytically active sites, responsible for the reaction of the internal CH4 steam reforming thus leading to the production of H2 in the reaction feed, which enables the cells to operate electrochemically for a longer period. After the investigation of the cells under the various reaction conditions it was decided that it would be useful to perform one more stability measurement with a cell that comprised a ‘‘low-AFL’’ ternary 3Au–3Mo–Ni/GDC anode, similarly prepared with screen printing and 5 mg/ cm2 loading. The difference in this test is that only one reaction mixture was applied and specifically the S/C = 0.13 plus 10 ppm of H2S. Finally, in this measurement the reactants and products were analyzed by online gas chromatography using a Varian CP-3800 gas chromatograph with a thermal conductivity detector. A Porapak Q column (80–100 mesh, 1.8 m 9 1/8in 9 2 mm), in parallel with a Carbosieve S-11 column (80–100 mesh, 2 m 9 1/8in 9 2 mm) were used for the analysis of CH4, H2, CO, CO2 and H2O at 130 °C. The scope of this test was to examine the performance and durability of the most promising electrode by applying a reaction mixture, in which the detrimental effects from sulfur poisoning and carbon deposition coexist. Figure 10 depicts the stability diagram, where there is an initial slight degradation during the first 25 min. After that first period the performance remained stable for approximately 140 min and finally it started to gradually degrade up to the point where the

Fig. 10 Stability diagram of 3Au–3Mo–Ni/GDC ‘‘low-AFL’’ Ni/ GDC anodes under CH4 steam reforming conditions (S/C = 0.13) plus 10 ppm H2S at T = 850 °C. The experiment was performed at j = 28 mA/cm2 and Ftotal = 100 cc/min. The reaction feed comprised, 6 vol% H2O, 46 vol% CH4 and 10 ppm H2S in helium

123

Fig. 11 I–V curves of a 3Au–3Mo–Ni/GDC ‘‘low-AFL’’ anode before (t0) and after (tEND) a stability test under CH4 steam reforming conditions (S/C = 0.13) plus 10 ppm H2S at T = 850 °C and Ftotal = 100 cc/min. The reaction feed comprised, 6 vol% H2O, 46 vol% CH4 and 10 ppm H2S in helium

voltage dropped to zero. Similarly the I–V (Fig. 11) and EIS (Fig. 12) measurements showed that the electrode performed well at the beginning of the stability. On the other hand, at the end of the test the I–V plots and the impedance spectra provide a clear indication for the deterioration of the cell. The latter is also shown by comparing the production rates of H2 and CO before and after the stability test. Indicatively the rates before the stability

Fig. 12 EIS measurements on a 3Au–3Mo–Ni/GDC ‘‘low-AFL’’ anode before (t0) and after (tEND) a stability test under CH4 steam reforming conditions (S/C = 0.13) plus 10 ppm H2S at T = 850 °C. Both of them have been recorded on galvanostatic mode at j = 28 mA/cm2 and Ftotal = 100 cc/min. The reaction feed comprised, 6 vol% H2O, 46 vol% CH4 and 10 ppm H2S in helium

Top Catal

(t = 0) were: rH2t¼0 ¼ 5:9 lmol=s; rCOt¼0 ¼ 1:8 lmol=s. Right after the completion of the stability (t = END) the corresponding values for the same rates were rH2t¼END ¼ 1:0 lmol=s and rCOt¼END ¼ 0:3 lmol=s, proving the catalytic deactivation of the electrocatalyst. Finally, the specific cell was subjected to a re-activation period under 30 vol% H2/He at 850 °C for 48 h and interestingly the reaction rates of the catalytically produced H2 and CO increased and were exactly the same like those before the stability test. On the downside, the electrochemical properties of the cell did not fully recover. Therefore, the above stability test verified that sulfur poisons the catalytic activity for the internal reforming reaction and the electrochemical interface between the electrode and the electrolyte. Furthermore it is suggested that, at least for the applied experimental conditions, the poisoning of the catalytic activity is reversible, while the poisoning of the electrochemical interface is not. However, these results are currently under further investigation, so as to be better clarified by the collection and comparison of more results under various experimental conditions. As it concerns the literature information in the SOFCs field there are some publications dealing with similar investigations, but the number is relatively limited. Specifically there are reports with various fuel compositions, and not H2 as the main component of the feed, including H2–CO, CH4, partially reformed CH4 and simulated biogas [5, 7, 28, 29, 38, 39]. In summary, the basic research approach is to study the most commonly used anode materials (e.g. Ni/YSZ) in the presence H/C-based fuels with the addition of various H2S concentration and by applying several operating conditions (e.g. at open circuit voltage, under different current loads, various reaction temperatures and fuel compositions) [7]. In the case of H/C-based fuels it is made clear that equally to the efforts that focus on the optimization of SOFC operational parameters, the composition of the cermets is another important aspect for the stable cell operation. One critical point that needs to be considered is that apart from nickel the electrode material should also contain an oxide, which alongside with the necessary electronic and ionic properties should also contribute to the catalytic/ reforming activity and in total to the cell‘s stability. In this respect there have been published several investigations on Ni-based compositions and some of the results tend to support Ni/Ceria based anodes and that particularly Ni/ GDC shows an increased sulfur tolerance in the presence of H/C-based fuels. This promising behavior has been mainly ascribed to the specific properties of GDC and particularly of CeO2, which plays a double operational role for both H/Cs reforming and H2S tolerance. The latter oxide is not severely poisoned in H/C fuel mixtures that contain H2S up

to 100 ppm and retains its catalytic/reforming activity, thus enabling the production of H2 that is further electrochemically oxidized for the power generation [7]. The approach that is followed in the presented study, particularly aiming to H/C-based fuel mixtures with sulfur concentrations varying in the region of 1–100 ppm, focuses on the modification of NiO/GDC. Currently, it is accepted that the targeted modification of SoA anode cermets, through the formation of Ni-alloys (or solid solutions), can mitigate carbon- and sulfur-induced deactivation, while maintaining the inherent material structure and SOFC performance. On this basis, the applied preparation methods of D.P. and D.CP. may allow insertion of the modifiers, without blocking the pores of the cermet and by maintaining the necessary high electronic conductivity and electrochemical/electrocatalytic activity [7] of the material. Furthermore, the effectiveness of metal sulfides that have already been examined as hydrodesulfurization (HDS) catalysts, leads to the suggestion that a balance between electrochemical performance and catalytic activity could be achieved through the proper materials selection and anode modification, by applying examples from the HDS research field. Indicatively, the transition metal sulfides (TMS) catalysts used in refineries, are generally molybdenum (Mo) or tungsten (W) sulfides supported on alumina and promoted by group VIIIA elements (e.g. Co or Ni). Studies by Chianelli et al. [40] and Kibsgaard et al. [41] have established the fundamental properties of TMS and concluded that their catalytic activity and selectivity arises from the electronic and structural properties of the sulfides themselves. Moreover, noble metals have also been suggested [42] as a promising option for improving the HDS properties and could also be considered as another approach for the preparation of Ni-based electrodes less prone to carbon and sulfur poisoning [7]. Consequently, it has been decided to modify the NiO/ GDC cermet with Au and/or Mo so as to be investigated as a possible sulfur tolerant material, having also in mind the already exhibited promising behavior of these modified cermets in respect to carbon deposition resistance [25, 27, 43, 44]. The proposed strategy resulted in the ternary Au– Mo–Ni/GDC cermet and shows to have a positive impact on the tolerance against 10 ppm of H2S, both in the case where H2 is the main fuel and in the case of CH4 internal steam reforming. Furthermore, it seems that there is a versatile positive effect from the applied modification, which induces (i) better electronic conductivity properties, (ii) better electrochemical interface between the electrode and the electrolyte and (iii) enhancement of the electrocatalyst’s sulfur tolerance. There are already published studies from our group, where it has been presented that the Au–Mo modification

123

Top Catal

of NiO/GDC, via the deposition coprecipitation method, seems to affect the structure of nickel towards the formation of a ternary Ni–Au–Mo solid solution [25]. Therefore, it can be suggested that the latter structural modification protects, up to a certain extent, nickel against sulfur adsorption and retards the rates of reactions (2)–(4), thus allowing the reforming of CH4 and preventing the depletion of hydrogen in the gas feed. Another speculation is that sulfur may adsorb and react preferentially with H2 on Au–Mo, thus affecting the equilibrium of reaction (2) during the period where the cell shows stable performance, and the nickel sites are able to act catalytically. In this respect Au–Mo–Ni alloyed sites may act as a HDS catalyst. The prerequisite for a stable performance is to achieve higher HDS rate than the detrimental H2S dissociative adsorption on the Ni (steam reforming) active sites. In addition, the intrinsic properties of ceria provide an extra stability factor against the 10 ppm of H2S, especially in the cases where the partial pressure of H2 is high. Finally, it should also be taken into account that carbon deposition is also hindered, which can be also beneficial. Overall, the combination of the above effects enables the cell to operate for a longer period, but it does not seem enough to fully protect the electrode from complete poisoning and degradation.

cases of CH4 internal steam reforming. Specifically, the induced positive effect is versatile providing (i) better electron conductivity properties, (ii) better electrochemical interface between the electrode and the electrolyte and (iii) enhancement of the electrocatalyst‘s sulfur tolerance. On the other hand the binary anodes, comprising Au–Ni/GDC and Mo–Ni/GDC, did not exhibit any improvement. The observed positive synergy on the ternary electrode can be attributed to the structural modification of nickel through the formation of a ternary Au–Mo–Ni solid solution [25]. It can be further suggested that the latter structural modification protects, up to a certain extent, nickel against sulfur adsorption. Therefore, the electrode maintains its catalytic activity for a longer period and the depletion of H2 in the reactants feed is delayed. The latter observations are currently under further investigation, so as to clarify the reasons for the superior performance of the Au–Mo–Ni/GDC electrode and to better understand the effect from the structural modification on nickel on the possible electrocatalytic reactions that take place. Acknowledgments The authors thank Dr. V. Dracopoulos at FORTH/ICEHT for the SEM images and our reviewers for their useful comments. The research leading to this review was funded by the European Union‘s Seventh Framework Programme (FP7/20072013) for the Fuel Cells and Hydrogen Joint Technology Initiative under the Projects ROBANODE and T CELL with grand agreement Numbers: 245355 and 298300, respectively.

4 Conclusions It has been shown that in the cases where the majority of the reaction feed is H2 then the concentration of 10 ppm H2S, under the investigated operating conditions of temperature, applied current and exposure time, cannot be considered as the main cause for the detrimental poisoning and subsequent degradation of Ni/GDC based anodes. This is because the specific concentration is low and the studied Ni/GDC electrodes, either modified or not, are fairly tolerant due to the intrinsic properties of ceria. Moreover, it is a fact that sulfur poisons the interface between the anode and the electrolyte and affects the electrochemical processes, but at the same time or even faster sulfur adsorbs and preferentially blocks the catalytically active sites that are responsible for the steam reforming of CH4. The result is depletion of H2 and finally the degradation of the cell. Therefore, by increasing the electrocatalyst‘s loading there is concomitant increase of the catalytically active sites, leading to generally more tolerant and stable behavior. The applied deposition coprecipitation of Au and Mo on NiO/GDC resulted in the ternary Au–Mo–Ni/GDC cermet and seems to be a good step towards the preparation of tolerant SOFC electrodes in the presence of 10 ppm of H2S, both in the case where H2 is the main fuel and in the

123

References 1. Escudero MJ, Irvine JTS, Daza L (2009) Development of anode material based on La-substituted SrTiO3 perovskites doped with manganese and/or gallium for SOFC. J Power Sources 192(1):43–50. doi:10.1016/j.jpowsour.2008.11.132 2. Singh A, Hill JM (2012) Carbon tolerance, electrochemical performance and stability of solid oxide fuel cells with Ni/yttria stabilized zirconia anodes impregnated with Sn and operated with methane. J Power Sources 214:185–194. doi:10.1016/j.jpowsour. 2012.04.062 3. Peters R, Riensche E, Cremer P (2000) Pre-reforming of natural gas in solid oxide fuel-cell systems. J Power Sources 86(1–2):432–441. doi:10.1016/s0378-7753(99)00440-1 4. Mogensen D, Grunwaldt JD, Hendriksen PV, Dam-Johansen K, Nielsen JU (2011) Internal steam reforming in solid oxide fuel cells: status and opportunities of kinetic studies and their impact on modelling. J Power Sources 196(1):25–38. doi:10.1016/j. jpowsour.2010.06.091 5. Sasaki K, Haga K, Yoshizumi T, Minematsu D, Yuki E, Liu R, Uryu C, Oshima T, Ogura T, Shiratori Y, Ito K, Koyama M, Yokomoto K (2011) Chemical durability of solid oxide fuel cells: influence of impurities on long-term performance. J Power Sources 196(22):9130–9140 6. Appleby AJ, Foulkes FR (2004) Fuel cell handbook, 7th edn. EG&G Technical Services, Inc., Morgantown 7. Niakolas DK (2014) Sulfur poisoning of Ni-based anodes for solid oxide fuel cells in H/C-based fuels. Appl Catal A 486:123–142. doi:10.1016/j.apcata.2014.08.015

Top Catal 8. Hagen A (2013) Sulfur poisoning of the water gas shift reaction on anode supported solid oxide fuel cells. J Electrochem Soc 160(2):F111–F118. doi:10.1149/2.060302jes 9. Prasad DH, Ji HI, Kim HR, Son JW, Kim BK, Lee HW, Lee JH (2011) Effect of nickel nano-particle sintering on methane reforming activity of Ni-CGO cermet anodes for internal steam reforming SOFCs. Appl Catal B 101(3–4):531–539 10. Chen T, Wang WG, Miao H, Li T, Xu C (2011) Evaluation of carbon deposition behavior on the nickel/yttrium-stabilized zirconia anode-supported fuel cell fueled with simulated syngas. J Power Sources 196(5):2461–2468 11. Chun CM, Bhargava G, Ramanarayanan TA (2007) Metal dusting corrosion of nickel-based alloys. J Electrochem Soc 154(5):C231–C240. doi:10.1149/1.2710215 12. Faes A, Hessler-Wyser A, Presvytes D, Vayenas CG, Van herle J (2009) Nickel-zirconia anode degradation and triple phase boundary quantification from microstructural analysis. Fuel Cells 9(6):841–851. doi:10.1002/fuce.200800147 13. Lussier A, Sofie S, Dvorak J, Idzerda YU (2008) Mechanism for SOFC anode degradation from hydrogen sulfide exposure. Int J Hydrog Energy 33(14):3945–3951. doi:10.1016/j.ijhydene.2007. 11.033 14. Halinen M, Saarinen J, Noponen M, Vinke IC, Kiviaho J (2010) Experimental analysis on performance and durability of SOFC demonstration unit. Fuel Cells 10(3):440–452. doi:10.1002/fuce. 200900152 15. Brightman E, Ivey DG, Brett DJL, Brandon NP (2011) The effect of current density on H2S-poisoning of nickel-based solid oxide fuel cell anodes. J Power Sources 196(17):7182–7187. doi:10. 1016/j.jpowsour.2010.09.089 16. Weber A, Dierickx S, Kromp A, Ivers-Tiffe´e E (2013) Sulfur poisoning of anode-supported SOFCs under reformate operation. Fuel Cells 13(4):487–493. doi:10.1002/fuce.201200180 17. Gong M, Liu X, Trembly J, Johnson C (2007) Sulfur-tolerant anode materials for solid oxide fuel cell application. J Power Sources 168(2):289–298. doi:10.1016/j.jpowsour.2007.03.026 18. Atkinson A, Barnett S, Gorte RJ, Irvine JTS, McEvoy AJ, Mogensen M, Singhal SC, Vohs J (2004) Advanced anodes for high-temperature fuel cells. Nat Mater 3(1):17–27. doi:10.1038/ nmat1040 19. Trembly JP, Bayless DJ, Gemmen RS (2006) In: Symposium of international Pittsburgh coal conference 20. Sun C, Stimming U (2007) Recent anode advances in solid oxide fuel cells. J Power Sources 171(2):247–260. doi:10.1016/j.jpow sour.2007.06.086 21. Cheng Z, Wang J-H, Choi Y, Yang L, Lin MC, Liu M (2011) From Ni-YSZ to sulfur-tolerant anode materials for SOFCs: electrochemical behavior, in situ characterization, modeling, and future perspectives. Energy Environ Sci 4(11):4380–4409. doi:10.1039/c1ee01758f 22. Wang W, Su C, Wu Y, Ran R, Shao Z (2013) Progress in solid oxide fuel cells with nickel-based anodes operating on methane and related fuels. Chem Rev 113(10):8104–8151. doi:10.1021/ cr300491e 23. Dong J, Cheng Z, Zha S, Liu M (2006) Identification of nickel sulfides on Ni–YSZ cermet exposed to H2 fuel containing H2S using Raman spectroscopy. J Power Sources 156(2):461–465. doi:10.1016/j.jpowsour.2005.06.016 24. Zhang L, Jiang SP, He HQ, Chen X, Ma J, Song XC (2010) A comparative study of H2S poisoning on electrode behavior of Ni/ YSZ and Ni/GDC anodes of solid oxide fuel cells. Int J Hydrog Energy 35(22):12359–12368 25. Niakolas DK, Athanasiou M, Dracopoulos V, Tsiaoussis I, Bebelis S, Neophytides SG (2013) Study of the synergistic interaction between nickel, gold and molybdenum in novel modified NiO/GDC cermets, possible anode materials for CH4

26.

27.

28.

29.

30.

31.

32.

33.

34.

35.

36.

37.

38.

39.

40.

41.

42.

fueled SOFCs. Appl Catal A 456:223–232. doi:10.1016/j.apcata. 2013.02.024 Niakolas DK, Athanasiou M, Neophytides SG, Bebelis S (2011) Characterization and carbon tolerance of new Au–Mo–Ni/GDC cermet powders for use as anode materials in methane fuelled SOFCs. ECS Trans 35(1):1329–1336 Niakolas DK, Ouweltjes JP, Rietveld G, Dracopoulos V, Neophytides SG (2010) Au-doped Ni/GDC as a new anode for SOFCs operating under rich CH4 internal steam reforming. Int J Hydrog Energy 35(15):7898–7904 Sasaki K, Susuki K, Iyoshi A, Uchimura M, Imamura N, Kusaba H, Teraoka Y, Fuchino H, Tsujimoto K, Uchida Y, Jingo N (2006) H2S poisoning of solid oxide fuel cells. J Electrochem Soc 153(11):A2023–A2029. doi:10.1149/1.2336075 Matsuzaki Y, Yasuda I (2000) The poisoning effect of sulfurcontaining impurity gas on a SOFC anode: part I. Dependence on temperature, time, and impurity concentration. Solid State Ion 132(3–4):261–269. doi:10.1016/s0167-2738(00)00653-6 Li TS, Wang WG, Chen T, Miao H, Xu C (2010) Hydrogen sulfide poisoning in solid oxide fuel cells under accelerated testing conditions. J Power Sources 195(20):7025–7032. doi:10. 1016/j.jpowsour.2010.05.009 Cheng Z, Zha S, Liu M (2007) Influence of cell voltage and current on sulfur poisoning behavior of solid oxide fuel cells. J Power Sources 172(2):688–693. doi:10.1016/j.jpowsour.2007. 07.052 Cheng Z, Zha SW, Aguilar L, Wang D, Winnick J, Liu ML (2006) A solid oxide fuel cell running on H2S/CH4 fuel mixtures. Electrochem Solid State Lett 9(1):A31–A33. doi:10.1149/1. 2137467 Cheng Z, Liu M (2007) Characterization of sulfur poisoning of Ni–YSZ anodes for solid oxide fuel cells using in situ Raman microspectroscopy. Solid State Ion 178(13–14):925–935. doi:10. 1016/j.ssi.2007.04.004 Luo T, Vohs JM, Gorte RJ (2002) An examination of sulfur poisoning on Pd/Ceria catalysts. J Catal 210(2):397–404. doi:10. 1006/jcat.2002.3689 Rostrup-Nielsen JR, Sehested J, Nørskov JK (2002) Hydrogen and synthesis gas by steam- and CO2 reforming. Advances in catalysis, vol 47. Academic Press, New York, pp 65–139. doi:10. 1016/S0360-0564(02)47006-X Rasmussen JFB, Hagen A (2009) The effect of H2S on the performance of Ni-YSZ anodes in solid oxide fuel cells. J Power Sources 191(2):534–541. doi:10.1016/j.jpowsour.2009.02.001 Abild-Pedersen F, Lytken O, Engbæk J, Nielsen G, Chorkendorff I, Nørskov JK (2005) Methane activation on Ni(111): effects of poisons and step defects. Surf Sci 590(2–3):127–137. doi:10. 1016/j.susc.2005.05.057 Haga K, Adachi S, Shiratori Y, Itoh K, Sasaki K (2008) Poisoning of SOFC anodes by various fuel impurities. Solid State Ion 179(27–32):1427–1431. doi:10.1016/j.ssi.2008.02.062 Hauch A, Hagen A, Hjelm J, Ramos T (2014) Sulfur poisoning of SOFC anodes: effect of overpotential on long-term degradation. J Electrochem Soc 161(6):F734–F743. doi:10.1149/2.080406jes Chianelli RR, Berhault G, Torres B (2009) Unsupported transition metal sulfide catalysts: 100 years of science and application. Catal Tod 147(3–4):275–286. doi:10.1016/j.cattod.2008.09.041 Kibsgaard J, Tuxen A, Knudsen KG, Brorson M, Topsøe H, Lægsgaard E, Lauritsen JV, Besenbacher F (2010) Comparative atomic-scale analysis of promotional effects by late 3d-transition metals in MoS2 hydrotreating catalysts. J Catal 272(2):195–203. doi:10.1016/j.jcat.2010.03.018 Babich IV, Moulijn JA (2003) Science and technology of novel processes for deep desulfurization of oil refinery streams: a review. Fuel 82(6):607–631. doi:10.1016/s0016-2361(02)003241

123

Top Catal 43. Papaefthimiou V, Shishkin M, Niakolas DK, Athanasiou M, Law YT, Arrigo R, Teschner D, Ha¨vecker M, Knop-Gericke A, Schlo¨gl R, Ziegler T, Neophytides SG, Zafeiratos S (2013) On the active surface state of nickel-ceria solid oxide fuel cell anodes during methane electrooxidation. Adv Energy Mater 3(6):762–769. doi:10.1002/aenm.201200727

123

44. Souentie S, Athanasiou M, Niakolas DK, Katsaounis A, Neophytides SG, Vayenas CG (2013) Mathematical modeling of Ni/ GDC and Au–Ni/GDC SOFC anodes performance under internal methane steam reforming conditions. J Catal 306:116–128. doi:10.1016/j.jcat.2013.06.015