Robust solid oxide cells for alternate power generation and carbon conversion

Abstract

Solid oxide cells (SOCs) are potentially useful for versatile energy conversion and storage applications. This work demonstrates the feasibility of SOCs that are operated in alternate power generation and carbon conversion modes. SOCs with lanthanum strontium vanadate (LSV) hydrogen electrodes have been proved to be competent candidates for such applications. SOCs with LSV-based hydrogen electrodes exhibit salient catalytic activity in hydrogen and various simulated feedstocks, e.g. syngas, biogas, town gas, and coal gas. LSV electrodes seem to perform better in the electrolyser mode than in the fuel cell mode. LSV electrodes are unlikely to be coked by the deposited carbon when exposed to carbon-forming gases. Most interestingly, LSV undergoes continuous activation during the course of operation, instead of being poisoned, when exposed to gases containing 50 ppm H₂S. LSV-based reversible SOCs under alternating electrolyser and fuel cell operations have been demonstrated with negligible performance degradation over 500 h.

---

1. Introduction

International commitment on a sustainable society and low-carbon future demands green energy conversion and storage technologies.¹ In an ideal scenario, primary energy sources are converted to secondary energy sources, mostly in electrical energy, with high energy conversion efficiency and a low carbon footprint. On the other hand, temporarily surplus electrical energies should be efficiently stored in various forms such as flywheels, dammed water, heat accumulators, batteries, and fuels. Currently available technologies, e.g. wind,² nuclear,³ solar,⁴metal hydrides,⁵ and batteries,⁶ largely fail to meet at least one of the key targets such as cost, reliability, flexibility, efficiency, and large scale deployment. A technology that generally fulfils these requirements is a solid oxide cell (SOC). SOCs can be operated both in fuel cell and electrolyser modes. The fuel cell mode, i.e. solid oxide fuel cell (SOFC), has been extensively studied for three decades.^7,8 The electrolyser mode, i.e. solid oxide electrolysis cell (SOEC), was explored in 1980s for hydrogen production from steam electrolysis.⁹ Those developments were not continued because of the low energy price in 1990s. Recently, SOECs have been revived by employing the as-developed SOFC technology,^10,11 possibly due to the soaring energy cost and worldwide consensus on green and low-carbon economy. The advantages of SOCs are the high energy conversion and storage efficiency and the fuel flexibility. The electrical efficiencies of SOFCs ranging from 40% (simple and small systems) to 50% (hybrid systems) had been demonstrated, and the theoretical efficiency is projected up to 60%.¹² The overall efficiencies of SOECs can reach 87–93% for megawatt scale SOEC stacks.¹³

In 2007, centralized power plants represented 43% of the CO₂ emission.¹⁴ The most critical task towards a low-carbon economy is to reduce CO₂ emission or the carbon density. Current carbon capture and sequestration (CCS) technologies, however, are highly energy-intensive and inefficient.¹⁵ Additionally, captured CO₂ is mostly stored in inert absorbents, thus interrupting the carbon cycle.¹⁶ To overcome this problem, a “carbon-neutral” carbon capture and conversion method using SOCs is proposed and illustrated in Fig. 1. The SOC stacks are coupled to a centralized power station, which emits greenhouse gases such as CO₂ and H₂O, and renewable/nuclear power generation facilities. This demonstrates an energy mixed concept whereby the energy/power is generated from different possible sources. Most SOCs will operate in fuel cell mode, i.e. SOFCs, in the case of heavy electrical load demand. They are fed with hydrocarbon fuels such as natural gas, coal gas, town gas, and biogas. The waste thermal energy is recuperated and used to preheat the inlet fuels. The effluent gases are condensed (to remove water), compressed, and stored as feedstock reservoir for carbon conversion (Fig. 1). Alternatively, off-peak electricity from nuclear or renewable energy sources is used to co-electrolyze CO₂ and H₂O and convert them into H₂/CO-rich gases. The H₂/CO-rich fuels are essentially synthetic gas, which is the building block for manufacturing different types of hydrocarbon fuels (Fig. 1). The co-electrolysis of CO₂ and H₂O through SOEC produces O₂ as the by-product, which can be channelled to the cathodes of SOFCs as oxidant. In comparison with air as the oxidant, O₂ is supposed to improve the system efficiency significantly. The O₂-rich air can also be channelled to the fossil fuel power plant to improve the combustion efficiency of the heat engines and boilers. The merit of this design is that the CO₂ generated from power plants is not released into the atmosphere but is converted into chemical fuels. The integrated system tends to reduce CO₂ emissions and favour a carbon-neutral economy. If incorporated into power grid, the SOCs may mitigate the power instability due to demand fluctuations and intermittent supply of power from renewable energy sources (e.g. wind, solar, and tidal).

Fig. 1 Schematic illustration of a synergic power generation and carbon conversion route.

As indicated in Fig. 1, the hydrogen electrodes† of such kinds of SOCs must have considerable fuel flexibility, sulfur tolerance, and carbon resistance. SOFCs employing robust hydrogen electrodes can be fed with H₂, CO, syngas, and gasified reformates from natural gas, coal, and heavier hydrocarbons. Direct hydrocarbon SOFCs have also been demonstrated.¹⁷ In contrast, the feedstock of SOECs can be a mixture of CO₂ and H₂O, added with certain amount of H₂ to prevent the electrocatalysts from oxidation. The state-of-the-art Ni-cermet electrodes, nevertheless, are prone to carbon filament precipitation,¹⁸ dusting, and dry corrosion when exposed to hydrocarbon fuels. Ni-cermets are also known to be poisoned by ppm levels of H₂S, though they can be regenerated when exposed to low concentration and temporarily high concentration of H₂S.^19,20Ni-cermet based SOECs are easily poisoned even by ppb levels of H₂S in CO₂ electrolysis²¹ and CO₂–H₂O co-electrolysis.^22,23 In the literature, various mitigation approaches on Ni–cermet SOFC anodes had been proposed, including the addition of water^24,25 and low melting point metals²⁶ into the fuels, and decoration of NiO particles with ceria^27–29 or niobia.³⁰sulfide^31–33 and oxide^34–41 anodes had also been explored extensively. The sulfur tolerance of oxide SOEC negative electrodes, however, has hardly been reported. It is worth noting that many practical SOC feedstocks, e.g. natural gas, coal gas, syngas, town gas, biogas, and reformed military fuels, contain several to thousands ppm level of H₂S.

Lanthanum strontium vanadate may be the candidate hydrogen electrode, which generally fulfils the stringent criteria. Historically, lanthanum strontium vanadate La_1-xSr_xVO₃ (0 ≤ x ≤ 1) are of interest because of their metal–insulator transition,⁴² high electrical conductivity,^42,43 and hence, are plausible candidates for high-T_c superconductors.⁴⁴ Hui and Petric reported that “Strontium doped LaVO₃ is stable under SOFC fuel outlet conditions..., however… is unlikely to be useful as anode materials” a decade ago.⁴⁵ However, studies showed that La_1-xSr_xVO₃ was active in electrochemical H₂S oxidation.^35,46,47,48SOFC with La_0.7Sr_0.3VO₃ anode is among the best of all reported results, with excellent sulfur tolerance up to percentage level.⁴⁶ Nevertheless, it seems that La_0.7Sr_0.3VO₃⁴⁹ and La_0.7Sr_0.3VO₃–YSZ composite anodes³⁵ have poor catalytic activity for H₂ and CH₄ oxidation. The electrode performance was not improved by replacing the YSZ with Gadolinia-doped ceria (GDC) in La_0.7Sr_0.3VO₃-based anodes.⁴¹ Recently, cerium strontium vanadate^39,50 and molybdenum-doped strontium vanadate⁵¹ have also been extensively studied. Vanadate has been proved to be of excellent sulfur tolerance, but low catalytic activity to fuel oxidation undermines its candidature for SOFC anodes. Our recent work demonstrated that a careful sintering process and in situreduction route can improve the anode performance of La_1-xSr_xVO₃ (x = 0.2, 0.3, 0.4, 0.5) by an order of magnitude.⁵² Thorough experimental and theoretical studies depicted the dynamic behaviour, double layer capacitance, three phase boundaries, and electrochemically active zones of highly active LSV/YSZ interfaces.^53–55

La_1-xSr_xVO₃electrodes under SOEC mode, and further SOCs with alternative operations between SOFC and SOEC, have hitherto not been reported. In this work, we demonstrate that SOCs with La_1−xSr_xVO₃hydrogen electrodes exhibit salient catalytic activity, long-term stability, sulfur tolerance, and carbon resistance when exposed to various feedstocks out of the H₂–H₂O–CO–CO₂–CH₄–H₂S system. The La_1-xSr_xVO₃-based SOC has been proven to be robust for power generation and carbon conversion in the SOFC and SOEC, respectively.

2. Experimental

2.1 Powder synthesis

40 at% strontium-doped lanthanum vanadate (La_0.6Sr_0.4VO_3–δ, hereafter simplified as LSV) powders were synthesized via a citric route with La(NO₃)₃·6H₂O (Fluka, 99.0%), Sr(NO₃)₂ (Fluka, 99.0%), NH₄VO₃ (Sigma-Aldrich, 99.5%), and C₆H₈O₇·H₂O (Fluka, 99.5%) as starting materials. The molar ratio of total metal ions over citric acid was 2 : 3. Stoichiometric amounts of the raw materials were dissolved in deionised water. The solution was heated at 70 °C and under magnetic stirring. A dark blue gel was obtained after dissipating excess water. It was dehydrated overnight at 100 °C and then grounded in an agate mortar. The precursor was baked at 600 °C for 2 h. The LSV powers were further baked at 600 °C for 2 h and at 800 °C for 6 h. 8 mol% yttria-stabilized zirconia (YSZ, Tosoh Japan) was used as-received.

2.2 Electrode fabrication

Right amounts of LSV and YSZ powders were roll-milled overnight in isopropanol with YSZ balls. The slurry was subsequently dried at 80 °C in ambient air. The obtained LSV–YSZ mixture was added with ink vehicles (Fuel Cell Materials, the USA) and ground in the agate mortar to form the electrode paste. Green electrodes were prepared by screen-printing the ink onto 1 mm-thick YSZ electrolyte substrates. The geometric area of electrodes was 0.46 cm². The catalytic layer, LSV (50 wt%) –YSZ, were sintered at 1215 °C for 2 h. The gas diffusion layer, LSV (70 wt%) –YSZ, was screen-printed onto the catalytic layer and sintered at 1205 °C for another 2 h. The Pt paste was also applied on the opposite side of YSZ substrate to represent the counter and the reference electrodes. All the Pt layers were sintered at 900 °C for 30 min.

2.3 Half cell testing

The half cell configuration is given in Fig. 2. Both the counter and reference electrodes were exposed to the ambient air. After purging with N₂, the test rig was heated up in H₂ with a ramping rate of 2 °C/min. The operating temperature was 900 °C. At the hydrogen electrode side, the H₂S-free feedstock was modulated from the corresponding pure gases via Brooks® 4-channel mass flow controller (MFC). H₂S was introduced to the feedstock by using a certified gas blend of H₂S and H₂ (1000 ppm H₂S balanced with H₂). To obtain a H₂O-containing feedstock, the gas blend was channelled through a humidifier with a temperature control of ± 0.1 °C. The total flow rate of feedstock was maintained at 100 sccm. Impedance responses were recorded via a 1255B frequency response analyzer coupled to a 1470E electrochemical interface (Solartron Analytical). The frequency range was 100 000–0.05 Hz. The sampling rate was 10 points/decade of frequency. Long-term stability tests were performed viachronopotentiometry, i.e. monitoring the electric potentials under given current densities. The sampling rate of chronopotentiometry was 0.2 points/s.

Fig. 2 Schematic diagram of the flow channel, test rig, and half cell configuration.

2.4 Characterization and calculation

The X-ray diffraction pattern was recorded on a Shimadzu 6000 with Cu Kα radiation. The step size was 0.02°. The scan rate was 2°/min. Low vacuum scanning electron microscopy (SEM, JEOL JSM–6360) and field-emission SEM (FESEM, JEOL 7600F) was used for low and high magnification surface morphology observations, respectively. Qualitative element detection in the electrodes was carried out via the energy-dispersive X-ray spectroscopy (EDX, Oxford Instruments X-Max), which were coupled to the JEOL JSM–6360. The chemical states of sulfur were characterized by X-ray photoelectron spectroscopy (XPS, Kratos XSAM-800) with Mg Kα radiation. The binding energy of C1s of adventitious carbon at 282.2 eV was taken as an internal standard. Thermochemical calculations were implemented viaHSC Chemistry 5 (Outokumpu Research Oy, Finland). The vanadium–oxygen–sulfur tertiary predominance diagrams were generated on the basis of minimum Gibbs energy.

3. Results and discussion

3.1 Structure

As-synthesized La_1-xSr_xVO₃ is known to be a mixture of LaVO₄ and Sr₂V₂O₇. La_1-xSr_xVO₃ is of perovskite structure but always coexists with ineradicable impurity phases even after prolonged exposure to reducing atmosphere.⁵² The impurity phases were eliminated in the reduced La_1-xSr_xVO₃ that was prepared via a citric route instead of solid state reactions.⁵⁴Fig. 3 shows the XRD pattern of single phase LSV powders that were reduced in pure H₂ at 900 °C for 5 h. This result is consistent with our previous work on La_0.8Sr_0.2VO₃.⁵⁴ The XRD pattern can be indexed as a cubic structure, with a = 3.8815(6) Å, V = 58.48 ų, and space group Pm-3m. As stated in ref. 51, however, it could be orthorhombic because the differences among the three lattice parameters are negligibly small. It is thus better to classify LSV as pseudo-cubic perovskite structure.

Fig. 3 X-ray diffraction pattern of La_0.6Sr_0.4VO_x (LSV) as reduced in pure H₂ at 900 °C for 5 h.

3.2 Electrode behaviour in clean feedstocks

Impedance spectroscopy shows good electrode performance of the reversible SOCs. The polarization resistances (R_p) of the LSV electrode in H₂ are 0.54, 0.43, 0.44 Ω cm² under the potential biases (E_bias) of −100, 0, and 100 mV, respectively (Fig. 4a). The high R_p under E_bias = −100 mV is due to the severe starvation of H₂O. With the H₂O content of the feedstock increased to 9 vol%, R_p are 0.11, 0.22, 0.44, 0.56 Ω cm² under E_bias of −300, −100, 0, 100 mV, respectively (Fig. 4b). These results indicate that LSV seems to be of higher catalytic activity in electrolyser than in fuel cell. Under the open circuit voltage (OCV) condition, the La_0.75Sr_0.25Cr_0.5Mn_0.5O₃ (LSCM) and La₄Sr₈Ti₁₁Mn_0.5Ga_0.5O_37.5SOFC anode showed R_p of 0.47⁵⁶ and 0.20⁵⁷ Ω cm² in 97H₂–3H₂O at 900 °C, respectively. LSV is thought to be comparable to these more well-established SOFC oxide anode materials. In contrast, only a few half-cell results of oxide SOEC negative electrodes were reported. Based on rough comparisons with the single cell results of LSCM^58,59 and symmetric cell results of Sr₂Fe_1.5Mo_0.5O_6-δ,⁶⁰LSV could be regarded as one of the best oxides for SOEC negative electrodes.

Fig. 4 Impedance responses of the LSV electrodes in pure H₂ (a) and 91H₂–9H₂O (b), under either positive (fuel cell) or negative (electrolysis) potential biases. The electrolyte contribution has been subtracted from the overall impedance.

Fig. 5 shows the chronopotentiometry curves of the LSV electrode in various simulated feedstocks (Table 1), where the half cell is polarized under alternating current densities of ± 0.2 A cm⁻². The fresh SOFC electrode shows the overpotential (η) of 89 mV in H₂ and a degradation rate (κ) of 0.023%/h. η of the LSV electrode is −67 mV when exposed to 91H₂–9H₂O and operated as electrolyser. It is worth noting that the η of SOEC is 22 mV smaller than that of SOFC, even with only 9% of H₂O. This is consistent with the impedance spectroscopy results that LSV is more active in the electrolyser than in fuel cell. The electrode potential oscillates for 2–5 h and then levels off when significant amounts of CO₂ are incorporated into the feedstock, i.e. 36H₂–9H₂O–55CO₂ and 73H₂–9H₂O–18CO₂. This phenomenon might be due to the surface reconstruction of LSV under these oxidising environments. A typical composition of desulfurized CH₄-rich biogas, 10H₂–35CO₂–55CH₄ (Table 1), is then introduced to test the carbon resistance of LSV. κ is as high as 1.7%/h for the SOFC operated under this carbon-forming feedstock. It implies that the reaction sites have possibly been blocked by the carbon deposition due to CH₄ pyrolysis. After 16 h exposure in 10H₂–35CO₂–55CH₄, the choked electrode is then electrolysed in 80H₂–20CO₂ for 8 h with the purpose of surface reduction and carbon removal via dry reforming. The regenerated electrode exhibits a slight performance improvement when the feedstock is shifted to 55H₂–15CO₂–30CO, a typical composition of desulfurized coal gas, though the degradation rate is significant in both electrolyser and fuel cell modes. These preliminary results show that LSV is active and robust towards H₂, clean biogas, and coal gas in reversible SOCs.

Fig. 5 Long-term chronopotentiometry of the LSV electrodes exposed to H₂S-free feedstocks, under open circuit voltage, 0.2 A cm⁻² (fuel cell mode), and −0.2 A cm⁻² (electrolyser mode).

Table 1 Composition, open circuit potential, tendency of carbon formation of gases that were utilized in chronological order. The operating temperature is 900 °C. The carbon formation is predicted by assuming the thermodynamic equilibrium and minimising the Gibbs energy of the gas mixture

Feedstock (100 sccm in total)						OCV (V, vs.Pt/air) /Place occurred	Carbon formation	Remarks
H2	H2O	CO	CO2	CH4	H2S
100	—	—	—	—	—	−1.145 (Fig. 5);	—	—
						−1.112 (Fig. 6);
						−1.120 (Fig. 8)
91	9	—	—	—	—	−1.037 (Fig. 5)	—	—
36	9	—	55	—	—	Fig. 5	No	Humidified H2-rich biogas
73	9	—	18	—	—	Fig. 5	No	Humidified H2-rich biogas (with low CO2)
10	—	—	35	55	—	Fig. 5	Yes	CH4-rich biogas
80	—	—	20	—	—	Fig. 5	No	H2-rich biogas (with low CO2)
55	—	30	15	—	—	−0.988 V (Fig. 5)	No	Coal gas by hydrogasification
100	—	—	—	—	50 ppm	−1.128 V (Fig. 8)	—	—
						−1.121 V (Fig. 8)
70	—	30	—	—	—	Fig. 8	Yes	Syngas/desulfurized town gas
70	—	30	—	—	50 ppm	−1.121 V (Fig. 8)	Yes	Town gas
65	—	30	5	—	50 ppm	−1.003 V (Fig. 8)	No	Coal gas
75	—	—	25	—	50 ppm	Fig. 8	No	H2-rich biogas (with low CO2)
60	—	—	20	20	50 ppm	−0.943 (Fig. 8)	Yes	Mixture of H2-rich biogas and CH4-rich biogas
97	3	—	—	—	50 ppm	−1.060 (Fig. 8)	—	—

---

3.3 Electrode behaviour in carbon-free and H₂S-tainted feedstocks

50 ppm H₂S was added to H₂ in order to evaluate the sulfur tolerance of LSV. Fig. 6 shows the potential responses of LSV electrodes subjected to fuel cell polarization in clean and H₂S-tainted H₂. η suddenly jumps from 91 mV to 102 mV when the H₂fuel is tainted by 50 ppm H₂S. This fast electrode degradation is most probably related to the H₂S adsorption onto LSV that blocks the reaction sites in a simultaneous way. The degradation lasts 3 h and results in the η change of 1 mV, after what the adsorption/desorption equilibrium of H₂S is supposed to be established. The LSV electrode undergoes a performance improvement of 0.012%/h (i.e. κ = −0.012%/h), rather than degradation, in the following 112 h. η finally reaches 97 mV after the 130-h fuel cell operation, which is only 6 mV higher than that in pure H₂. The salient long-term stability is also evident in the microstructural analysis. As shown in Fig. 7a and 7b, the electrode microstructure has been little changed after 130 h fuel cell operation. It is of great interest to note that LSV is catalytically active in H₂–50 ppm H₂S and has being continuously improved at least up to a few hundred hours. To further demonstrate the suitability of LSV towards practical gases, the LSV electrode is subjected to a consecutive testing on simulated town gas, biogas, and coal gas with 50 ppm H₂S (Table 1). The LSV electrode exhibits non-negligible performance degradation, κ = 0.090%/h, under the fuel cell polarization in 70H₂–30CO (Fig. 8a), a simulated syngas. Interestingly, the LSV electrode enjoys performance improvement, κ = −0.003%/h, merely by the addition of 50 ppm H₂S into the feedstock. After the 52 h fuel cell operation, R_p in this syngas is only 0.07 Ω cm² larger than that of the fresh electrode in H₂ (Fig. 9). A coal gas composition is simulated by replacing 5% of H₂ with CO₂ in the previous feedstock, i.e. 65H₂–30CO–5CO₂–50 ppm H₂S. The η of the LSV electrode exposed to this coal gas is −68 mV. It decreases to −61 mV after the electrolysis for 10 h (Fig. 8a). After the standby exposure to H₂–50 ppm H₂S for 3 h, the LSV electrode is subjected to fuel cell polarization in H₂–50 ppm H₂S. The η in H₂–50 ppm H₂S is almost unchanged as compared to the fresh LSV electrode exposed to H₂ (cf.Fig. 8a and 8b). 75H₂–25CO₂–50 ppm H₂S, a typical H₂–rich biogas with low CO₂ concentration, is then introduced. The LSV electrode exhibits slight activation with κ of −0.087%/h in H₂–50 ppm H₂S and −0.069%/h in 75H₂–25CO₂–50 ppm H₂S (Fig. 8b). These promising results indicate that LSV, as a SOChydrogen electrode, performs satisfactorily in practical gases where carbon deposition hardly occurs.

Fig. 6 Long-term chronopotentiometry of the LSV electrodes under 0.2 A cm⁻² (fuel cell mode), with response to the gas shift from pure H₂ to H₂-50 ppm H₂S.

Fig. 7 Typical low-magnification SEM images of LSV electrodes: freshly reduced (a), operated in fuel cell mode and in H₂–50 ppm H₂S for 127 h (b), and operated under alternating SOFC/SOEC operations in various H₂S-tainted and carbon-forming feedstocks for 500 h (c).

Fig. 8 Long-term chronopotentiometry of the LSV electrodes exposed to H₂S-tainted feedstocks. The whole figure is divided into (a) to (d) merely for the clear illustration purpose.

Fig. 9 Impedance spectra of the LSV electrode as shown in Fig. 8a: (a) the freshly prepared electrode in pure H₂ and (b) the electrode after 70-hour fuel cell polarization and in the 70H₂–30CO–50 ppm H₂S atmosphere.

3.4 Electrode behaviour in carbon-forming and H₂S-tainted feedstocks

Once the feedstock is shifted to 60H₂–20CO₂–20CH₄–50 ppm H₂S, a simulated gas blend of H₂-rich and CH₄-rich biogases that is prone to carbon formation, the electrode potential is suddenly increased from −0.868 V to −0.692 V. This is mainly due to the OCV reduction after the gas shifting (see OCVs in Table 1). Nevertheless, a large k of 0.377%/h has been observed during the 12 h fuel cell polarization (Fig. 8b). Two degradation mechanisms might exist: (i) carbon coking of the electrochemical reaction sites and (ii) surface oxidation of LSV. To eliminate the ambiguity the LSV electrode is subsequently subjected to electrolysis for 30 h. It is believed that the carbon removal via dry reforming of CH₄ is somewhat suppressed because CO₂ is reduced to CO during the electrolysis. The CH₄ pyrolysis thus dominates and the carbon deposition is more remarkable in electrolyser mode than in fuel cell mode. If the degradation is controlled by the carbon coking mechanism, the electrode degradation after 30 h electrolysis should be more serious than that before the electrolysis.

In contrast to this prediction, the electrode potential has been hardly changed. As shown in Fig. 8b and 8c, the electrode potential is −0.664 V and −0.639 V for the LSV electrode before and after 30 h electrolysis, respectively. More interestingly, continuous electrode activation with k = −0.05%/h has been observed in the 300 h fuel cell operation (Fig. 8c). These results obviously rule out the carbon coking mechanism. The most probably mechanism is that the “oxidized” LSV surface has been reduced and reconstructed during the 30 h electrolysis, which is prior to the 300 h fuel cell operation. η is −51 mV and κ is only 0.028%/h in the following electrolysis operation in 97H₂–3H₂O–50 ppm H₂S (Fig. 8d). It is worth highlighting that this LSV electrode, even after the alternating SOFC/SOEC operations for 500 h, shows better performance than those LSV electrodes exposed to H₂–9H₂O (Fig. 5) and 65H₂–30CO–5CO₂–50 ppm H₂S (Fig. 8a)

Improved microstructure could be a plausible reason for the continuous electrode activation. It is, however, difficult to explain why both electrodes are activated but their microstructures are distinctly different for the two LSV electrodes that are addressed in Fig. 6 and 8. Note that the “fine” microstructure of the LSV electrode exposed to carbon-forming gases, as confirmed by the EDX measurements, is ascribed to the effect of deposited carbon (Fig. 7c). The exact mechanism of the electrode activation is currently unknown. These promising results, nevertheless, demonstrate that LSV performs satisfactorily at least up to a few hundred hours even with the presence of carbon deposition and sulfur.

3.5 Sulfur in LSV electrodes

While LSV is of excellent carbon resistance, it is still unclear why it is activated, instead of poisoned, by the H₂S-containing feedstocks. SEM observations showed negligible contributions of H₂S on modifying the electrode microstructure (cf.Fig. 7a and 7b). This is conceivable because the H₂S concentration is only 50 ppm and most sulfur-containing species, e.g. S, CS₂, are volatile at 900 °C. One of the possible mechanisms is the formation of metal sulfides when the LSV electrode is exposed to H₂S-contaning feedstocks. These sulfides might be catalytically active for fuel cell and electrolysis reactions. Fig. 10 shows high-resolution surface morphology images of the three LSV electrodes that are (i) freshly prepared, (ii) exposed to H₂–50 ppm H₂S as corresponding to Fig. 6, and (iii) exposed to various feedstocks as corresponding to Fig. 8, respectively. No fine structures are observed in the freshly prepared electrode (Fig. 10a). Modest amounts of nanoparticles are observed in the LSV electrode with 120 h exposure to H₂S (Fig. 10b). A large quantity of relatively uniform-sized nanoparticles, with diameters of 5.5 ± 0.9 nm (Fig. 11), exist on the electrode surfaces that have been exposed to 50 ppm H₂S for 510 h (Fig. 10c).

Fig. 10 Field-emission SEM images of LSV electrode surfaces: (a) freshly reduced, (b) after the fuel cell testing in H₂–50 ppm H₂S for 120 h (Fig. 6), and (c) after the long term testing as given in Fig. 8.

Fig. 11 High magnification surface morphology images of LSV surfaces, as shown in Fig. 10c. For the deposited particles the mean size is 5.5 nm and the standard deviation is 0.9 nm.

The nanoparticles could be either precipitated electrocatalysts after prolonged annealing or certain metal sulfides resulting from the reaction with H₂S. It is tentatively assumed that LSV is the matrix of these nanoparticles, as vanadium-based oxides have good affinity to sulfur while YSZ is essentially inert towards H₂S. This assumption is substantiated by relevant thermochemical calculations. Fig. 12a shows the V–O–S predominance diagram at 900 °C. V₂S₃ is found to be the most stable phase under the SOC operating conditions. The concentration of these nanoparticles in LSV electrodes, however, is still below the detection limit of EDX and XRD. In contrast, Fig. 13 shows very weak S2p photoelectron lines of the LSV electrode that was exposed to H₂–50 ppm H₂S for 120 h (Fig. 6). The background is so strong that the doublet peaks of the S2p spectrum cannot be resolved. The binding energy (BE) of S2p spectral line of the unknown sulfur-containing species is as high as 166.6 ± 0.1 eV. This BE value is higher than the reported values on V₂S₃ and other possible sulfides for 3 to 4 eV, as listed in Table 2. Surprisingly, it is very close to the S2p3/2 spectral lines of several alkali and transition metal sulfites and sulfates (Table 2). It is quite possible that the metal sulfides have decomposed during the cooling down of the half-cells. Alternatively, they are thermodynamically unstable in air and at room temperatures. As shown in the V–O–S predominance diagram at 25 °C (Fig. 12b), it is clearly evident that VOSO₄ is the only dominant phase at 25 °C and in wide range of oxygen partial pressures. Thus, the high BE S2p lines observed in this study is quite possibly the S2p spectrum of VOSO₄, which results from the oxidation of V₂S₃ at ambient conditions.

Fig. 12 The V–O–S predominance diagrams at 900 °C (a) and at 25 °C (b).

Fig. 13 XPS spectra of S2p for the LSV electrode exposed to H₂–50 ppm H₂S for 120 h.

Table 2 Binding energy (eV) of S2p spectral lines in sulfur, sulfides, sulfites, and sulfates

Formula	Binding energy (eV)			Reference
	S2p	S2p1/2	S2p3/2
Unknown	166.6 ± 0.1	—	—	This work
VS1.08	—	163.0 ± 0.2	162.0 ± 0.2	61
VS	—	—	162.3 ± 0.2	62
SrS	—	—	160.8 ± 0.2
VS	162.9	—	—	63
V3S4	162.2	—	—
V2S3	163.0	—	—
V5S8	162.8	—	—
VS	161.6	—	—	64
V2S3	161.9	—	—	65
SrS	—	—	160.2	66
S	—	—	164.8	67
NaSO3	—	—	166.4	68
	—	—	166.5	69
	—	—	166.6	70
	167.0	—	—	71
	—	—	167.2	72
CaSO3	166.9	—	—	71
K2SO3	—	—	167.5	72
Na2SO4	—	—	167.5	73
CoSO4	—	—	168.00	70
NiSO4	—	—	168.05	70

---

In literature, various vanadium-based sulfides had been reported with high electrical conductivity⁷⁴ and catalytic activity towards H₂S oxidation.⁷⁵ The in situ generated nanoparticles on LSV surfaces, as observed in this study, may lead to enlarged reaction areas and enhanced reaction kinetics, e.g.adsorption/desorption, dissociation, and spillover of intermediate surface species. To some extent, the nanoparticle decoration on the LSV surface is very similar to the impregnation of nano-sized ionic conductors into the electrodes.⁷⁶ The ultimate goal of impregnation is to decorate the electrocatalyst surfaces with catalytically active nanoparticles. One of the fatal shortcomings of impregnated nanoparticles is their thermal instability. In contrast, the in situ precipitated metal sulfide nanoparticles on LSV surfaces seem to be very stable for more than 500 h. These nanoparticles are generated continuously by the reactions between LSV and H₂S during the course of high temperature SOC operations. This is distinctive to the impregnated nanoparticles, which are largely eliminated at elevated temperatures because of the particle sintering. This “in situnanoparticles” approach represents a novel route in tailoring nanostructured electrodes in high temperature energy and conversion devices.

4. Conclusions

The feasibility of reversible SOCs for alternate power generation and carbon conversion had been demonstrated. SOCs with LSV-based hydrogen electrodes survived and showed promising catalytic activity in various simulated gases, e.g. H₂, syngas, biogas, town gas, and coal gas. The LSV was a pseudo-cubic perovskite structure. The LSV electrodes exhibited better performance for SOCs operated in the electrolyser mode than in the fuel cell mode. The polarization resistances of LSV electrodes, operated in various atmospheres, polarization modes, and potential biases, were in the range 0.1 to 0.5 Ω cm². Steady state polarization results showed that the overpotential of LSV electrodes was mostly not higher than 100 mV, under the alternating current density of ±0.2 A cm⁻². The LSV electrodes were not choked by deposited carbon when exposed to carbon-forming gases. Most interestingly, LSV electrodes underwent continuous activation, rather than being poisoned, when exposed to various gases containing 50 ppm H₂S. LSV-based reversible SOCs, operated under fuel cell and electrolyser modes for more than 500 h, had demonstrated negligible performance degradation in carbon-forming and H₂S-containing gases. Nanostructured fine structures were observed in the LSV electrodes after prolonged exposure to 50 ppm H₂S. XPS results and thermochemical calculations showed that, quite probably, these nanoparticles were V₂S₃ but they were oxidized to VOSO₄ during the cooling down of the LSV electrodes.

References

1. International Energy Agency, Energy Technology Perspectives 2010-Scenarios&Strategies to 2050, 2010.
2. J. Li, Renewable Sustainable Energy Rev., 2010, 14, 1154–1171.
3. N. Z. Muradov and T. N. Veziroglu, Int. J. Hydrogen Energy, 2008, 33, 6804–6839.
4. M. Pagliaro, A. G. Konstandopoulos, R. Ciriminna and G. Palmisano, Energy Environ. Sci., 2010, 3, 279–287.
5. K.-F. Aguey-Zinsou and J.-R. Ares-Fernandez, Energy Environ. Sci., 2010, 3, 526–543.
6. J. Baker, Energy Policy, 2008, 36, 4368–4373.
7. N. Q. Minh, J. Am. Ceram. Soc., 1993, 76, 563–588.
8. A. J. Jacobson, Chem. Mater., 2010, 22, 660–674.
9. W. Doenitz, R. Schmidberger, E. Steinheil and R. Streicher, Int. J. Hydrogen Energy, 1980, 5, 55–63.
10. A. Hauch, S. D. Ebbesen, S. H. Jensen and M. Mogensen, J. Electrochem. Soc., 2008, 155, B1184–B1193.
11. D. M. Bierschenk, J. R. Wilson and S. A. Barnett, Energy Environ. Sci., 2011, 4, 944–951.
12. US Department of Energy, Fuel Cell Handbook, 7th edn, Morgantown, 2004..
13. Q. Fu, C. Mabilat, M. Zahid, A. Brisse and L. Gautier, Energy Environ. Sci., 2010, 3, 1382–1397.
14. International Energy Agency, How the Energy Sector can Deliver on a Climate Agreement in Copenhagen, 2009.
15. B. W. Terwel, F. Harinck, N. Ellemers and D. D. L. Daamen, Int. J. Greenhouse Gas Control, 2011, 5, 181–188.
16. G. Ferey, C. Serre, T. Devic, G. Maurin, H. Jobic, P. L. Llewellyn, G. De Weireld, A. Vimont, M. Daturi and J. S. Chang, Chem. Soc. Rev., 2011, 40, 550–562.
17. S. McIntosh and R. J. Gorte, Chem. Rev., 2004, 104, 4845–4865.
18. M. L. Toebes, J. H. Bitter, A. J. van Dillen and K. P. de Jong, Catal. Today, 2002, 76, 33–42.
19. Y. Matsuzaki and I. Yasuda, Solid State Ionics, 2000, 132, 261–269.
20. S. W. Zha, Z. Cheng and M. L. Liu, J. Electrochem. Soc., 2007, 154, B201–B206.
21. S. D. Ebbesen and M. Mogensen, J. Power Sources, 2009, 193, 349–358.
22. S. D. Ebbesen and M. Mogensen, Electrochem. Solid-State Lett., 2010, 13, D106–D108.
23. S. D. Ebbesen, C. Graves, A. Hauch, S. H. Jensen and M. Mogensen, J. Electrochem. Soc., 2010, 157, B1419–B1429.
24. N. Lakshminarayanan and U. S. Ozkan, Appl. Catal., A, 2011, 393, 138–145.
25. L. Yang, Z. Cheng, M. L. Liu and L. Wilson, Energy Environ. Sci., 2010, 3, 1804–1809.
26. O. A. Marina, C. A. Coyle, M. H. Engelhard and L. R. Pederson, J. Electrochem. Soc., 2011, 158, B424–B429.
27. H. Kurokawa, T. Z. Sholklapper, C. P. Jacobson, L. C. De Jonghe and S. J. Visco, Electrochem. Solid-State Lett., 2007, 10, B135–B138.
28. J. W. Yun, S. P. Yoon, J. Han, S. Park, H. S. Kim and S. W. Nam, J. Electrochem. Soc., 2010, 157, B1825–B1830.
29. R. Lohsoontorn, D. J. L. Brett and N. P. Brandon, J. Power Sources, 2008, 183, 232–239.
30. S. H. Choi, J. H. Wang, Z. Cheng and M. Liu, J. Electrochem. Soc., 2008, 155, B449–B454.
31. N. U. Pujare, K. W. Semkow and A. F. Sammells, J. Electrochem. Soc., 1987, 134, 2639–2640.
32. C. Yates and J. Winnick, J. Electrochem. Soc., 1999, 146, 2841–2844.
33. M. Liu, G. L. Wei, J. L. Luo, A. R. Sanger and K. T. Chuang, J. Electrochem. Soc., 2003, 150, A1025–A1029.
34. R. Mukundan, E. L. Brosha and F. H. Garzon, Electrochem. Solid-State Lett., 2004, 7, A5–A7.
35. L. Aguilar, S. W. Zha, Z. Cheng, J. Winnick and M. L. Liu, J. Power Sources, 2004, 135, 17–24.
36. S. W. Zha, Z. Cheng and M. L. Liu, Electrochem. Solid-State Lett., 2005, 8, A406–A408.
37. Z. R. Xu, J. L. Luo, K. T. Chuang and A. R. Sanger, J. Phys. Chem. C, 2007, 111, 16679–16685.
38. X. C. Lu and J. H. Zhu, J. Electrochem. Soc., 2008, 155, B1053–B1057.
39. N. Danilovic, J. L. Luo, K. T. Chuang and A. R. Sanger, J. Power Sources, 2009, 192, 247–257.
40. S. W. Zha, P. Tsang, Z. Cheng and M. L. Liu, J. Solid State Chem., 2005, 178, 1844–1850.
41. M. Cooper, K. Channa, R. De Silva and D. J. Bayless, J. Electrochem. Soc., 2010, 157, B1713–B1718.
42. M. Sayer, R. Chen, R. Fletcher and A. Mansingh, J. Phys. C: Solid State Phys., 1975, 8, 2059–2071.
43. J. B. Webb and M. Sayer, J. Phys. C: Solid State Phys., 1976, 9, 4151–4164.
44. A. V. Mahajan, D. C. Johnston, D. R. Torgeson and F. Borsa, Phys. Rev. B: Condens. Matter, 1992, 46, 10973–10985.
45. S. Q. Hui and A. Petric, Solid State Ionics, 2001, 143, 275–283.
46. L. Aguilar, S. W. Zha, S. W. Li, J. Winnick and M. Liu, Electrochem. Solid-State Lett., 2004, 7, A324–A326.
47. Z. Cheng, S. W. Zha, L. Aguilar and M. L. Liu, Solid State Ionics, 2005, 176, 1921–1928.
48. Z. Cheng, S. W. Zha, L. Aguilar, D. Wang, J. Winnick and M. L. Liu, Electrochem. Solid-State Lett., 2006, 9, A31–A33.
49. C. Peng, J. L. Luo, A. R. Sanger and K. T. Chuang, Chem. Mater., 2010, 22, 1032–1037.
50. C. T. G. Petit, R. Lan, P. I. Cowin, J. T. S. Irvine and S. W. Tao, J. Mater. Chem., 2011, 21, 525–531.
51. A. Aguadero, C. de la Calle, D. Perez-Coll and J. A. Alonso, Fuel Cells, 2011, 11, 44–50.
52. X. M. Ge and S. H. Chan, J. Electrochem. Soc., 2009, 156, B386–B391.
53. X. M. Ge, C. F. Fu and S. H. Chan, Electrochem. Commun., 2011, 13, 792–795.
54. X. M. Ge, C. F. Fu and S. H. Chan, Phys. Chem. Chem. Phys., 2011, 13, 15134.
55. X. M. Ge, C. F. Fu and S. H. Chan, Electrochim. Acta, 2011, 56, 5947–5953.
56. S. W. Tao and J. T. S. Irvine, Nat. Mater., 2003, 2, 320–323.
57. J. C. Ruiz-Morales, J. Canales-Vazquez, C. Savaniu, D. Marrero-Lopez, W. Z. Zhou and J. T. S. Irvine, Nature, 2006, 439, 568–571.
58. X. Yang and J. T. S. Irvine, J. Mater. Chem., 2008, 18, 2349–2354.
59. C. Jin, C. H. Yang, F. Zhao, D. Cui and F. L. Chen, Int. J. Hydrogen Energy, 2011, 36, 3340–3346.
60. Q. A. Liu, C. H. Yang, X. H. Dong and F. L. Chen, Int. J. Hydrogen Energy, 2010, 35, 10039–10044.
61. H. F. Franzen and G. A. Sawatzky, J. Solid State Chem., 1975, 15, 229–232.
62. H. F. Franzen, M. X. Umaña, J. R. McCreary and R. J. Thorn, J. Solid State Chem., 1976, 18, 363–368.
63. W. Bensch and R. Schlogl, J. Solid State Chem., 1993, 107, 43–57.
64. J. Gopalakrishnan, T. Murugesan, M. S. Hegde and C. N. R. Rao, J. Phys. C: Solid State Phys., 1979, 12, 5255.
65. D. Lichtman, J. H. Craig, V. Sailer and M. Drinkwine, Appl. Surf. Sci., 1981, 7, 325–331.
66. R. P. Vasquez, J. Electron Spectrosc. Relat. Phenom., 1991, 56, 217–240.
67. J. M. Thomas, R. H. Williams, M. Barber and I. Adams, J. Chem. Soc., Faraday Trans. 2, 1972, 68, 755–764.
68. K. M. Abraham and S. M. Chaudhri, J. Electrochem. Soc., 1986, 133, 1307–1311.
69. C. D. Wagner and J. A. Taylor, J. Electron Spectrosc. Relat. Phenom., 1982, 28, 211–217.
70. X. R. Yu, F. Liu, Z. Y. Wang and Y. Chen, J. Electron Spectrosc. Relat. Phenom., 1990, 50, 159–166.
71. R. V. Siriwardane and J. M. Cook, J. Colloid Interface Sci., 1986, 114, 525–535.
72. N. H. Turner, J. S. Murday and D. E. Ramaker, Anal. Chem., 1980, 52, 84–92.
73. W. E. Swartz, K. J. Wynne and D. M. Hercules, Anal. Chem., 1971, 43, 1884–1887.
74. N. Kijima, S. Ikeda, I. Shimono, T. Matsumoto, S. Tsuji, K.-i. Kumagai and S. Nagata, J. Solid State Chem., 1996, 126, 189–194.
75. K. T. Li, M. Y. Huang and W. D. Cheng, Ind. Eng. Chem. Res., 1996, 35, 621–626.
76. S. P. Jiang, Mater. Sci. Eng., A, 2006, 418, 199–210.

---

Footnote

† When hydrogen electrode is referred in SOCs, it is the anode for SOFCs and the negative electrode for SOECs.

---