Ce_0.9Gd_0.1O_2−δ membranes coated with porous Ba_0.5Sr_0.5Co_0.8Fe_0.2O_3−δ for oxygen separation

Abstract

Robust oxygen ion conducting membranes based on doped ceria oxides can be used as oxygen permeation membranes with a short circuit to provide the required electronic conduction. Previous methods have coated both surfaces of the ion conducting electrolyte membrane with expensive noble metals as the electronic conducting phase to allow the electron shuttling required for oxygen reduction and oxidation to take place between the two membrane surfaces. During operation of the membrane, the atmosphere on the two sides of the membrane is different. The feed side is exposed to air, whereas the permeated side may be exposed to CO₂ or reducing gases such as CH₄ or H₂. At high operating temperatures, such exposure to different gases requires the use of different materials to prepare the membranes, giving opportunities for further optimisation and the reduction of costs. In this work, a novel Ce_0.9Gd_0.1O_2−δ membrane coated on the surface exposed to air with a cost-effective mixed conductive layer of Ba_0.5Sr_0.5Co_0.8Fe_0.2O_3−δ was developed to deliver a highly stable oxygen flux for use in clean energy applications or as a membrane reactor for chemical synthesis. The membrane coated with Ba_0.5Sr_0.5Co_0.8Fe_0.2O_3−δ improved the flux of oxygen compared with membranes coated with pure Ag. A triple phase boundary theory is put forward to explain the observed improvement in the oxygen flux.

---

1 Introduction

The annual global production of oxygen from air separation is of the order of hundreds of millions of tons and is the largest sector of the industrial gases market, with sales worth more than US$5 billion in 2014.¹ Oxygen is used in a variety of industries, including the chemical industry, in pharmaceuticals, petroleum, glass, cement and ceramics, and in pulp/paper and metal manufacturing. The power generation industry currently only uses 4% of the oxygen produced worldwide.² However, this energy market is expected to expanded in the near future with the deployment of clean energy technologies to decrease CO₂ emissions, such as the IGCC and Oxyfuel projects. These clean energy projects use oxygen as the feed gas. It is estimated that, by 2040, the energy industry will make up 60% of the oxygen market, using around 2 million tons of oxygen per day.²

Industrial oxygen is currently produced on an intermediate scale at a lower purity (<93%) by pressure swing adsorption, or on a larger scale (>1000 tons per day) at a higher purity (>93%) by cryogenic distillation; both methods are expensive. Cryogenic distillation, in which air is cooled down to −185 °C, is a mature technique that has been used since the early 1900s and there is little room for improvement. These conventional high-cost oxygen production techniques largely prevent the widespread use of clean energy technologies. To combat extreme climate change, we need to develop more cost-effective oxygen production technologies to make clean energy more economically feasible.

Among the emerging technologies, ceramic membranes have the greatest potential in oxygen separation techniques on the basis of energy efficiency and lower capital investment. Since 1997, Air Products and the US Department of Energy have invested over US$150 million in advancing ion transporting membrane (ITM) technology.³ Research began in the 1980s when an appreciable oxygen permeation flux through a dense perovskite membrane with the general formula La_1−xSr_xCo_1−yFe_yO_3−δ was reported by Teraoka et al.⁴ Since then, a variety of perovskite ceramics with oxygen permeation properties have been reported. These have a general formula ABO_3−δ, where the A and B sites are either individually or jointly occupied by La, Sr, Ba, Ca or Zr and Mg, Al, Ti, Cr, Mn, Fe, Co, Ni, Cu, Ga, Zr, Ta, W, Sc or Zn, respectively.^5–20 As a result of the structural flexibility of the perovskite structure, 90% of the metals in the periodic table have been investigated in the synthesis of perovskite structures.²¹ One common feature of these membranes is that they are able to simultaneously conduct both oxygen ions and electrons at increased temperatures and are thereby classified as mixed ionic and electronic conducting (MIEC) ceramic membranes. The working principles of these MIEC ceramic membranes are shown schematically in Fig. 1a. When exposed to a partial pressure gradient, oxygen can permeate through these perovskite membranes via surface transferring reactions and bulk diffusion without a need for the external electric loadings required by pure ionic conductors (Fig. 1b).²² To sustain a high oxygen flux, the membranes must not only possess sufficient ionic and electronic conductivities, but also require good surface reactions for the exchange of oxygen between its gaseous molecular state and solid lattice oxygen.

Fig. 1 Schematic diagram of different designs of oxygen permeation membranes. (a) Mixed ionic–electronic conducting oxide membranes; (b) pure oxygen ionic conducting membranes with an external power source; (c) ITM reactors for syngas production; and (d) pure oxygen ionic conducting membranes with short circuit coating.

Progress has been made during the last two decades in the optimisation of materials and the scale-up of engineering applications. An oxygen flux of up to 14.5 ml min⁻¹ cm⁻² at 950 °C was achieved on a thin Ba_0.5Sr_0.5Co_0.8Fe_0.2O_3−δ (BSCF) membrane supported on a hollow fibre geometry under the oxygen pressure gradient created by non-pressurized air and sweep gas.²³ Long-term operational stability over thousands of hours has been achieved on BSCF and La_0.6Sr_0.4Co_0.2Fe_0.8O_3−δ membranes for oxygen production under vacuum conditions or non-reacting sweep gas models.^5,24,25 The ITM project at Air Products has been advanced to Phase 5 testing and has an oxygen-producing capability of 2000 tons per day.²⁶ Inspired by the success of using membrane technology for the production of pure oxygen, an ambitious project has been initiated to develop ITM syngas to combine air separation and gas oxidation in a single unit (Fig. 1c). If successful, this novel technology would greatly advance both the petrochemical industry and the production of clean energy because it could save 20–30% of the capital costs of the overall production of synthesis gas and H₂.³ In particular, it would help to improve the viability of the Oxyfuel project by retrofitting existing power plants rather than developing new boiler systems.²⁶

These ambitious targets pose stricter requirements on the stability of the membrane material. As shown in Fig. 1c, the permeate side of the membrane is exposed to reducing gases such as CH₄ or acid gases such as CO₂ for high-temperature reactions. However, none of these membranes can survive long term in such harsh environments at high temperatures because the perovskite oxides are easily damaged by reaction with these gases.^27,28 Fortunately, robust membranes that can tolerate these gases can be fabricated from fluorite-type ion conducting ceramics such as yttria-stabilized zirconia (YSZ) and gadolinium- (or samaria-) doped ceria (GDC/SDC), as illustrated by their wide applications as high-temperature electrolytes for oxygen ion transport in solid oxide fuel cells.^29–31 Despite their proven robustness, their application as oxygen production membranes is fairly limited as a result of their poor electronic conductivity and more complex membrane design using an oxygen pump with external electrical loadings (Fig. 1b). To simplify the membrane design, noble metal powders such as Pt, Au, Pd and Ag have been mixed with YSZ, SDC or GDC to prepare a dual-phase membrane in which the metal and ion conductor phases individually transfer oxygen ions and electrons simultaneously. However, it is difficult to guarantee a continuous conducting path for each phase, which often leads to very low oxygen fluxes as a result of mismatch between the two phases.²²

A novel ion conducting ceramic membrane based on the configuration of solid oxide fuel cell with a short circuit has recently been reported.³² Noble metals were decorated as a thin coating layer on the ceramic surface instead of mixing with the ceramics. Together with an electronic conducting sealant (e.g. Ag paste) as the short circuit, the porous metal layer functions as a continuous electrically conducting phase for electrons shuttling between the two membrane surfaces during the oxygen-exchange reactions. Noble Ag or Pt were decorated on both membrane surfaces as the electronic conducting phase to verify this concept (Fig. 1d),^32,33 but the high cost of these precious metals is a disadvantage in terms of future large-scale applications. However, some MIEC perovskites are good electronic conductors with sufficient oxygen-exchange kinetics and high stability in air to be used as the coating material, replacing the Ag or Pt layer in the air (feed) side of the GDC, YSZ or SDC membranes during air separation and therefore reducing costs.

In the work reported here, we tested this membrane design using porous MIEC perovskite to coat the membrane surface facing the air side, but leaving the permeate side exposed to the CO₂-containing sweep gas coated with a precious metal (Ag). BSCF was used as the perovskite phase and the ion conducting membrane was GDC. To fully understand the effects of different electronic conducing phases from the ceramics or precious metals on the overall transport of oxygen, some membrane samples were coated on both sides with BSCF and their performances were compared with those of Ag-coated membranes.

2 Experimental

Ce_0.9Gd_0.1O_2−δ (GDC) and Ba_0.5Sr_0.5Co_0.8Fe_0.2O_3−δ (BSCF) ceramic oxides were synthesized by a combined EDTA–citrate complexing sol–gel process. The metal element precursors were from the nitrate salts Ce(NO₃)₂·xH₂O, Gd(NO₃)₂·xH₂O, Ba(NO₃)₂, Sr(NO₃)₂, Co(NO₃)₂·6H₂O and Fe(NO₃)₃·9H₂O, which were purchased from Aldrich and used as received. Stoichiometric quantities of the metal nitrates in aqueous solution were dissolved in distilled water. EDTA and citric acid were then added as complexing agents. NH₃·H₂O was added to the solution to control the pH at around 6–8. The molar ratios of the total metal ions, EDTA and citric acid were controlled at 1 : 1 : 2. The solution was heated at 90 °C to evaporate the water and to obtain a transparent gel. The gel was pre-fired at 250 °C and heated in air for 5 h at 700 (GDC) or 950 °C (BSCF) to obtain the ceramic powder with the desired structure.

To fabricate the membrane, 0.4 g of the ceramic powder was pressed into a disk-shaped membrane in a 15.0 mm diameter stainless-steel mould under a hydraulic pressure of approximately 1.5 × 10⁸ Pa. These green membranes (about 0.8 mm in thickness) were further sintered at 1350 °C for 10 h at a ramping/cooling rate of 2 °C min⁻¹. BSCF was dispersed inside an ink vehicle solution and was then coated onto the surfaces of some GDC membranes by brush-painting at least three times, followed by calcination at 1000 °C in air for 2 h. An Ag slurry was applied to the surface of other GDC membranes by a similar brush-painting method and the membranes were subsequently calcined at 600 °C for 2 h.

To investigate the phase composition of the membranes, XRD analysis was carried out using a Bruker D8 Advance X-ray diffractometer with Cu Kα radiation at 40 kV and 30 mA. Scanning electron microscopy (SEM) images were obtained using a Zeiss EVO 40XVP microscope at an accelerating voltage of 15 kV.

A Shimadzu 2014A gas chromatograph equipped with a 5 Å capillary column and a thermal conductivity detector was used to determine the oxygen concentration during the high-temperature oxygen permeation test. To set up the permeation cell, an Ag paste was applied as the sealant to fix the disk membrane onto a dense quartz tube. The effective membrane area was about 0.45 cm². The partial pressure of oxygen in the feed stream was 0.21 atm. Helium was applied as the sweep gas to transport the permeated oxygen to the gas chromatograph to determine the concentration. Assuming Knudsen diffusion for the leakage of nitrogen and oxygen through the pores or cracks, the fluxes of leaked N₂ and O₂ are related by:

The O₂ permeation flux was then calculated by subtracting the leaked oxygen using the following equation:

where C_O2 and C_N2 are the measured concentrations of oxygen and nitrogen in the gas on the permeate side, respectively, F is the flow-rate of the exit gas on the permeate side (ml min⁻¹) and S is the geometrical surface area of the membrane on the sweep side (cm²).

3 Results and discussion

XRD analysis was carried out to determine the phase components of the GDC membranes and the coating decorations. Fig. 2 shows the typical XRD patterns of the membranes used for the oxygen permeation test. As shown in Fig. 2a, the characteristic peaks of GDC are located at the 2θ angles of 28° (111), 33° (200), 47° (220), 52° (311), 58° (222), 76° (331) and 79° (420), which agrees with previously reported results.^34,35 Fig. 2b shows the XRD patterns of the GDC with a coating of BSCF on the surface. All the peaks could be clearly assigned to GDC or BSCF. No other undesired phase was observed, which rules out any reactions between these two phases and proves the stable coexistence of GDC and BSCF. Fig. 2c shows the existence of the Ag coating. Compared with the strong diffraction signal from Ag, the GDC peaks are relatively weak and not as well distinguished as in Fig. 2a and b.

Fig. 2 XRD patterns of (a) GDC, (b) GDC coated with BSCF and (c) GDC coated with Ag.

Fig. 3a and b show the SEM images of the surface and cross-sectional views of the pure GDC membrane with a grain size of 0.5–1.0 μm. The dense structure of the sintered GDC membrane is observed not only from the surface, but also from the cross-sectional view (Fig. 3b) as no apparent porosity can be detected in the SEM image. The BSCF was initially dispersed inside an ink vehicle solution and then coated on the GDC membrane, followed by sintering at a high temperature to increase the interfacial adhesion. Fig. 3c shows the SEM image of the porous coating structure. The pore size is in the range 2–3 μm and the porous BSCF coating layer has a thickness of about 5.0 μm, as marked by the horizontal line in Fig. 3d. In contrast, the thick supporting bulk layer is the dense GDC structure (inset, Fig. 3d). The GDC membrane coated with Ag paste has a porous structure with a thickness of about 6.0 μm (Fig. 3e and f). The porous structure, either BSCF or Ag, is used to promote the oxygen permeation flux.^9,32,36

Fig. 3 SEM images of: (a) surface of GDC; (b) cross-section of GDC; (c) surface of GDC membrane decorated with BSCF coating; (d) cross-section of GDC membrane decorated with BSCF coating; (e) surface of GDC membrane decorated with Ag coating; and (f) cross-section of GDC membrane decorated with Ag coating.

To test the effectiveness of the BSCF coating in promoting oxygen transport through GDC membrane, three GDC membranes with similar 0.8 mm thickness but different coatings (samples a, b and c) were used to display the flux differences. All the oxygen permeation tests were repeated three times. Sample a had Ag coatings on both sides of the membrane, whereas sample b had an Ag coating on one side of the membrane and the other side was coated with BSCF. Sample c was coated on both sides with BSCF. Fig. 4 shows that a coating of BSCF (sample b or c) on the membrane significantly improved the oxygen flux at similar operating temperatures and sweep gas flow-rates.

Fig. 4 Oxygen permeation of GDC membrane coated with (a) Ag on both the feed side and the permeate side, (b) Ag on the feed side and BSCF on the permeate side and (c) BSCF on both the feed side and the permeate side. The sweep gas flow-rate was 100 ml min⁻¹.

For example, at 650 °C, the oxygen flux of sample a with the Ag coating on both sides of the membrane was 0.017 ml min⁻¹ cm⁻² (Fig. 4a), whereas for sample b, with a BSCF coating on the permeate side and an Ag coating on the feed side, the flux was improved by 16% to 0.02 ml min⁻¹ cm⁻² (Fig. 4b). When both sides of the membrane were coated by BSCF (sample c), the flux was further increased by 27% (compared with sample a) to 0.023 ml min⁻¹ cm⁻² at an operating temperature of 650 °C. When the temperature was continuously increasing, the oxygen flux of all three samples increased compared with the flux at lower temperatures. This promoting effect of temperature on the flux can be interpreted as temperature activation of the oxygen-exchange surface reactions and the bulk diffusion of oxygen ions. However, careful inspection of the increase in flux with the BSCF coating at different temperatures shows that this improvement in flux is more distinguishable at higher than at lower temperatures. For instance, at 800 °C sample a had a flux of 0.043 ml min⁻¹ cm⁻², whereas the respective fluxes of samples b and c were 0.075 and 0.088 ml min⁻¹ cm⁻², an improvement of 42 and 51%, respectively. The relatively large improvement in flux with the BSCF coating at higher temperatures indicates that the percentage of surface reaction resistance in the overall resistance (surface reaction plus bulk diffusion), which determines the limiting step for oxygen transport, gradually increases with increasing operating temperatures. However, to explain theoretically the larger improvement in oxygen permeation obtained by coating with a mixed conductor such as BSCF rather than a pure electronic conductor like Ag, we need to consider the two membrane surface reactions:

where stands for lattice oxygen, for an oxygen vacancy and e for an electron. At the membrane surface facing the air, molecular oxygen is reduced to lattice (or ionized) oxygen with the acceptance of electrons. On the other membrane surface facing the sweep gas, molecular oxygen and electrons are released via the oxidation of lattice oxygen. Thus each surface reaction only takes place in the respective triple phase boundary (TPB) areas. This is shown schematically in Fig. 5a: when the membrane is coated with Ag, the TPB areas marked by pink squares are the points where Ag, GDC and oxygen (or the sweep gas He) can meet to make the surface reactions possible. This requires the Ag coating layer to be continuous, uniform and porous, with a strong adherence to the GDC surface. However, the Ag coating produced via brush-painting and a high-temperature treatment is agglomerated; in this instance the GDC areas covered by the Ag agglomerates (marked in Fig. 5a by the dotted rectangle) are sacrificed and do not contribute to the surface reactions. In another words, the large Ag particles from the coating layer block the ionic bulk diffusion as a result of the very limited TPB area. However, this restriction is removed if the coating layer is replaced by a mixed conductor such as BSCF (Fig. 5b). Here, the TPB area is expanded from the area marked by the pink squares to the entire particle surface of BSCF, thus improving the kinetics of the oxygen surface reaction. Of course, the BSCF particles should adhere strongly to the GDC surface to decrease the interfacial resistance to ionic or electron transport. This is why a mixed conductor-BSCF coating on GDC improves the oxygen permeation flux compared with an Ag coating. A similar observation was made by Wang et al.³⁷ and Imashuku et al.,³⁸ who used a pulse laser method to deposit La_0.8Sr_0.2CoO_3−δ oxide on an SDC membrane.

Fig. 5 Schematic diagram showing the effects of different coatings on the oxygen permeation process: (a) pure electronic conductor of Ag; and (b) mixed conductor of BSCF.

In addition to temperature activation, Fig. 6 shows that the oxygen flux can also be increased by increasing the sweep gas flow-rate, which lowers the oxygen partial pressure on the permeate side, thus increasing the partial pressure gradient of oxygen. For example, the oxygen flux through the membrane of sample c increased from 1.02 to 1.21 ml min⁻¹ cm⁻² as the sweep gas flow-rate was increased from 80 to 120 ml min⁻¹ at 850 °C. In this work, which used ambient air and He as the feed and sweep gases, respectively, the largest oxygen partial pressure gradient only reached 0.21/0.001. However, under the reaction conditions in industrial applications, the air used as the feed gas is pressurized to 50 atm and the permeated oxygen is consumed in the reaction so that the oxygen partial pressure gradient can be as large as 10/10⁻¹⁶ atm. Under such circumstances, the oxygen flux is very fast. Therefore the factor that restricts the performance of the reaction is not the oxygen flux, but the stability of the membrane materials in real-life applications.

Fig. 6 Effect of sweep gas flow-rate on oxygen permeation fluxes.

If the oxygen flux is completely controlled by bulk diffusion as the membrane surface coating improves the oxygen-exchange surface reaction kinetics to a minimum resistance, we can estimate the maximum theoretical oxygen flux based on the Wagner equation:

where T is the absolute temperature, R is the ideal gas constant, F is the Faraday constant, L is the thickness of the membrane, σ_i is the oxygen ion conductivity of GDC, and and are the oxygen partial pressures at the feed and permeate sides of the membrane.

A number of studies have been carried out to measure the oxygen ion conductivity of the GDC membrane.^39–43 There may be minor differences in the conductivity values when using different synthesis methods for GDC. Table 1 shows the oxygen ion conductivity at different temperatures from previously published data.³⁹ The maximum theoretical oxygen flux was then calculated using the results in Table 1. Compared with the theoretical oxygen flux, the measured fluxes in this work were 10–40% smaller, indicating that the surface reaction kinetics contributed a certain resistance in limiting the overall oxygen transport.

Table 1 Oxygen ion conductivity and theoretical oxygen permeation flux of GDC membrane coated on both sides with BSCF

Temp (°C)	σi (S cm^−1)^39	Theoretical JO2 (ml min^−1 cm^−2)	Measured JO2 (ml min^−1 cm^−2)
600	4.05 × 10^−3	0.018	0.016
650	8.44 × 10^−3	0.039	0.023
700	1.29 × 10^−2	0.063	0.038
750	2.03 × 10^−2	0.104	0.065

---

Two types of coated GDC membranes (samples c and d) were tested for long-term use by adding CO₂ to the sweep gas on the permeate side. Sample c has BSCF coatings on both sides of the membrane, but sample d consisted of GDC coated by Ag on one side of the membrane and BSCF on the other. Fig. 7 shows the stability performance tests of two typical membranes. During the test, the BSCF coating side of the membrane in sample d faced the feed air and the side with the Ag coating faced the sweep gas containing CO₂. As expected, the oxygen flux of sample c coated on both sides with BSCF showed a continuous decrease when the sweep gas was switched from pure He to a mixture with 10% CO₂ (Fig. 7a). At 800 °C and after 600 minutes, the flux for sample c decreased by more than 50%. After the sweep gas had been changed back to pure helium, the oxygen flux could not be recovered to its original value as a result of the formation of carbonate from the reactions between the alkaline earth metals (such as Ba/Sr) and CO₂.^27,44 In sharp contrast, sample d worked very well under these conditions. The GDC membrane with coatings of BSCF and Ag showed very stable oxygen fluxes in either pure He or mixtures containing CO₂. Fig. 7b shows that, after introducing CO₂ into the He sweep gas, the flux value was slightly lowered, but could be maintained at a stable value of 0.044 ml min⁻¹ cm⁻². The temporary decrease in flux with the introduction of CO₂ is a result of the stronger adsorption of CO₂ to the membrane surface relative to helium, not due to the reaction between CO₂ and GDC. The wide application of GDC as a solid electrolyte in solid oxide fuel cells verifies that it is sufficiently stable to withstand gases such as CO₂ and CH₄.⁴⁵ Therefore when the sweep gas was switched back to pure He, the oxygen flux was quickly recovered to the original value.

Fig. 7 Long-term oxygen permeation tests through GDC membranes coated with (a) BSCF on both sides and (b) BSCF on feed side and Ag on the sweep side. The sweep gas flow-rate is 100 ml min⁻¹.

4 Conclusions

Based on the working principles of solid oxide fuel cells, a short circuit robust ion conducting ceramic membrane configuration has been proposed to overcome the weak stability of ceramic membranes in atmospheres of CO₂ or reducing gases at high temperatures.³² However, this reported design had a high cost because it used expensive noble metals on both sides of the membrane. This study demonstrated the feasibility of using a cheaper mixed conducting ceramic coating for electronic conduction on robust ion conducting ceramic membranes for oxygen separation. To illustrate the concept, GDC membranes coated with BSCF or Ag paste were used as ion conducting membranes with a mixed or pure electronic conductor, respectively. The experimental results show that the replacement of the Ag coating on the GDC membrane by BSCF significantly improved the oxygen flux as a result of the increased TPB area, with improved oxygen-exchange surface reaction kinetics. The maximum oxygen flux was 0.13 ml min⁻¹ cm⁻² at 850 °C with a dense GDC thickness of 0.8 mm and with both sides of the membrane coated by BSCF. The GDC membrane with a BSCF coating facing the air and an Ag coating exposed to a CO₂ mixture was very stable during a long-term permeation test, making it suitable for clean energy applications such as the Oxyfuel project and in membrane reactors for the production of syngas.

Acknowledgements

The authors are grateful for the financial support provided by the Australian Research Council through the Future Fellow Program (FT12100178). Mr Chi Zhang acknowledges the PhD scholarship provided by Curtin University.

References

1. PRWeb, http://www.prweb.com/releases/2013/11/prweb11347927, (accessed November 2013).
2. World nuclear association, http://www.world-nuclear.org/info/Energy and-Environment/-Clean-Coal-Technologies/, (accessed May 2014).
3. T. Foster, Air products, Air Separation Technology-Ion Transport Membrane (ITM), Allentown, 2008.
4. Y. Teraoka, H. M. Zhang, S. Furukawa and N. Yamazoe, Chem. Lett., 1985, 14, 1743.
5. Z. P. Shao, W. S. Yang, Y. Cong, H. Dong, J. H. Tong and G. X. Xiong, J. Membr. Sci., 2000, 172, 177.
6. Z. B. Zhang, D. J. Chen, Y. Gao, G. M. Yang, F. F. Dong, C. Chen, F. Ciucci and Z. P. Shao, RSC Adv., 2014, 4, 25924.
7. D. Z. Han, J. H. Wu, Z. F. Yan, K. Zhang, J. Liu and S. M. Liu, RSC Adv., 2014, 4, 19999.
8. H. H. Wang, R. Wang, D. T. Liang and W. S. Yang, J. Membr. Sci., 2004, 243, 405.
9. S. Baumann, J. M. Serra, M. P. Lobera, S. Escolástico, F. Schulze Küppers and W. A. Meulenberg, J. Membr. Sci., 2011, 377, 198.
10. X. Y. Lin, T. Kerstiens and T. Markus, J. Membr. Sci., 2013, 438, 83.
11. F. Schulze-Küppers, S. Baumann, W. A. Meulenberg, D. Stöver and H.-P. Buchkremer, J. Membr. Sci., 2013, 433, 121.
12. M. Katsuki, S. R. Wang, M. Dokiya and T. Hashimoto, Solid State Ionics, 2003, 156, 453.
13. J. Luyten, A. Buekenhoudt, W. Adriansens, J. Cooymans, H. Weyten, F. Servaes and R. Leysen, Solid State Ionics, 2000, 135, 637.
14. T. Ramos and A. Atkinson, Solid State Ionics, 2004, 170, 275.
15. Y. Zou, W. Zhou, S. M. Liu and Z. P. Shao, J. Eur. Ceram. Soc., 2011, 31, 2931.
16. B. Meng, Z. G. Wang, Y. Y. Liu, X. Y. Tan, J. C. D. da Costa and S. M. Liu, Sep. Purif. Technol., 2011, 78, 175.
17. X. Z. Chen, L. Huang, Y. Y. Wei and H. H. Wang, J. Membr. Sci., 2011, 368, 159.
18. M. Shang, J. H. Tong and R. O'Hayre, RSC Adv., 2013, 3, 15769.
19. V. V. Kharton, S. B. Li, A. V. Kovalevsky and E. N. Naumovich, Solid State Ionics, 1997, 96, 141.
20. J. X. Yi, S. J. Feng, Y. B. Zuo, W. Liu and C. S. Chen, Chem. Mater., 2005, 17, 5856.
21. K. Zhang, J. Sunarso, Z. Shao, W. Zhou, C. Sun, S. Wang and S. Liu, RSC Adv., 2011, 1, 1661.
22. J. Sunarso, W. A. Meulenberg, S. Baumann, J. M. Serra, S. Liu, J. C. D. da Costa and Y. S. Lin, J. Membr. Sci., 2008, 320, 13.
23. A. Leo, S. Smart, S. Liu and J. C. D. da Costa, J. Membr. Sci., 2011, 368, 64.
24. D. Schlehuber, E. Wessel, L. Singheiser and T. Markus, J. Membr. Sci., 2010, 351, 16.
25. X. Tan, Z. Wang, B. Meng, X. Meng and K. Li, J. Membr. Sci., 2010, 352, 189.
26. J. Repasky, D. McCarthy, P. Armstrong and M. Carolan, Workshop on Technology Pathways Forward for Carbon Capture & Storage on Natural Gas Power Systems, Washington DC, 2014.
27. A. Waindich, A. Mobius and M. Müller, J. Membr. Sci., 2009, 337, 182.
28. J. Yi and M. Schroeder, J. Membr. Sci., 2011, 378, 163.
29. J. Kim and Y. S. Lin, J. Membr. Sci., 2000, 167, 123.
30. C. S. Chen, B. A. Boukamp, H. J. M. Bouwmeester, G. Z. Cao, H. Kruidhof, A. J. A. Winnubst and A. J. Burggraaf, Solid State Ionics, 1995, 76, 23.
31. T. J. Mazanec, T. L. Cable and J. G. Frye Jr, Solid State Ionics, 1992, 53, 111.
32. K. Zhang, Z. Shao, C. Li and S. Liu, Energy Environ. Sci., 2012, 5, 5257.
33. K. Zhang, L. Liu, Z. Shao, R. Xu, J. C. D. da Costa, S. Wang and S. Liu, J. Mater. Chem. A, 2013, 1, 9150.
34. J. Xue, Q. Liao, Y. Y. Wei, Z. Li and H. H. Wang, J. Membr. Sci., 2013, 443, 124.
35. J. Xue, Q. Zheng, Y. Y. Wei, K. J. Yuan, Z. Li and H. H. Wang, Ind. Eng. Chem. Res., 2012, 51, 4703.
36. Z. G. Wang, H. Liu, X. Y. Tan, Y. G. Jin and S. M. Liu, J. Membr. Sci., 2009, 345, 65.
37. L. Wang, S. Imashuku, A. Grimaud, D. Lee, K. Mezghani, M. A. Habib and Y. Shao-Horn, ECS Electrochem. Lett., 2013, 2, F77.
38. S. Imashuku, L. Wang, K. Mezghani, M. A. Habib and Y. Shao-Horn, J. Electrochem. Soc., 2013, 160, E148.
39. S. Lubke and H.-D. Wiemhofer, Ber. Bunsenges. Phys. Chem., 1998, 102, 642.
40. B. C. H. Steele, Curr. Opin. Solid State Mater. Sci., 1996, 1, 684.
41. R. V. Mangalaraja, S. Ananthakumar, M. Paulraj, K. Uma, M. López, C. P. Camurri and R. E. Avila, Process. Appl. Ceram., 2009, 3, 137.
42. H. Yahiro, Y. Eguchi, K. Eguchi and H. Arai, J. Appl. Electrochem., 1988, 18, 527.
43. J. Liu and W. Weppner, Ionics, 1999, 5, 115.
44. M. Arnold, H. H. Wang, J. Martynczuk and A. Feldhoff, J. Am. Ceram. Soc., 2007, 90, 3651.
45. W. Wang, C. Su, Y. Z. Wu, R. Ran and Z. P. Shao, Chem. Rev., 2013, 113, 8104.

---