You are on page 1of 5

G Model MEMSCI-11189; No.

of Pages 5

Journal of Membrane Science xxx (2012) xxx–xxx

Contents lists available at SciVerse ScienceDirect

Journal of Membrane Science
journal homepage:

Silver-molten carbonate composite as a new high-flux membrane for electrochemical separation of CO2 from flue gas
Nansheng Xu a,b , Xue Li a , Maxwell A. Franks c , Hailei Zhao b , Kevin Huang a,∗
a b c

Department of Mechanical Engineering, University of South Carolina, Columbia, SC 29208, USA School of Materials Science and Engineering, University of Science and Technology Beijing, Beijing 100083, China South Carolina Governor’s School for Science and Mathematics, Hartsville, SC 29550, USA

a r t i c l e

i n f o

a b s t r a c t
One of the major concerns of consuming fossil fuels to produce useful form of energy is the emission of CO2 , a greenhouse gas that can cause climate change and ultimately threaten the survival of humanity. Controlling CO2 -emission is an urgent but only practical solution to stabilize CO2 concentration in the atmosphere. In this paper, we report our effort to capture CO2 from a simulated flue gas by utilizing a dualphase mixed carbonate-ion and electron conducting membrane, namely molten carbonate and silver. The obtained CO2 and O2 flux densities are the highest among the published metal-molten carbonate systems and relatively stable over 80-h period. The measured CO2 and O2 flux densities not only exhibit similar activation energy but also have the ratio of 2:1, favorably confirming the surface electrochemical reaction of CO2 + 1/2O2 + 2e = CO3 2− . These results demonstrate the Ag–MC dual-phase composite as a promising high-flux membrane for high-temperature electrochemical CO2 separation from flue gas with high selectivity. © 2012 Elsevier B.V. All rights reserved.

Article history: Received 22 November 2011 Received in revised form 31 January 2012 Accepted 1 February 2012 Available online xxx Keywords: Gas separation Membrane Mixed conductor Flux density Chemical potential

1. Introduction Fossil fuels are still the dominant energy source powering our modern society, largely due to their competitive cost and wide accessibility. However, one of the main byproducts from consuming fossil fuels is carbon dioxide (CO2 ), a greenhouse gas that can cause climate change and ultimately threaten the survival of humanity. An alarming statistic published in 2008 shows that a nearly 40% of 5.8 billion metric tons of the total US anthropogenic CO2 emissions was resulted from burning fossil fuels for power generation [1,2]. Therefore, to effectively mitigate CO2 emissions from the use of fossil fuels has drawn a great deal of attentions globally in recent years. It has risen to be not only a scientific topic but also a political agenda. There are four general strategies currently being considered for the abatement of CO2 emissions: demand-side conservation, supply-side efficiency improvement, potential increase in nuclear and renewable energy supplies and implementation of carbon capture and storage (CCS) technology on fossil-fueled power plants [3,4]. Of all the options, CCS is regarded as the sole practical solution to achieve considerable CO2 emission reduction from fossil-fueled power plants in the near term [4].

∗ Corresponding author. Tel.: +1 803 777 0204. E-mail addresses:, (K. Huang). 0376-7388/$ – see front matter © 2012 Elsevier B.V. All rights reserved. doi:10.1016/j.memsci.2012.02.001

To realize the ultimate CO2 storage/sequestration, separation and capture of CO2 from a CO2 -containing industrial process stream is the first step. The CO2 capture technologies currently being developed and demonstrated are targeting at three different stages/processes depending on the type of a power plant: post-combustion, oxyfuel combustion and pre-combustion [4–6]. The pre-combustion CO2 capture is more suitable for integrated gasification combined cycle (IGCC) power plants whereas the postcombustion and oxygen-combustion captures are more applicable to conventional pulverized coal-fired (PC) and natural gas-fired power plants [3]. However, the existing state-of-the-art CO2 capture technologies are primarily based on chemical/physical absorptions by a liquid solvent or a solid sorbent and are yet ready for commercialization at the larger scale necessary for power plant application. The parasitic loads (steam and power) required to support CO2 capture would decrease power generation capacity by nearly one-third and increase levelized cost-of-electricity (COE) by as much as 80%. Even if successfully scaled up, it would not be cost effective at their current level of process development [3,7]. In comparison to solvent- and sorbent-based CO2 capture processes, conventional membrane-based technology presents cost advantages when applied to pre- and post-combustion CO2 capture: no steam or chemical load, delivering high-pressure CO2 and driving CO shifting reaction to completion [8–10]. However, the major challenges are the limited selectivity, thus product purity, and poor compatibility with high-temperature process stream.

Please cite this article in press as: N. Xu, et al., Silver-molten carbonate composite as a new high-flux membrane for electrochemical separation of CO2 from flue gas, J. Membr. Sci. (2012), doi:10.1016/j.memsci.2012.02.001

G Model MEMSCI-11189; No. of Pages 5 2

N. Xu et al. / Journal of Membrane Science xxx (2012) xxx–xxx

downstream OTM or recycled back to the combustion chamber for oxy-combustion [6]. Since single-phase MOCC is presently nonexistent from a material perspective [12], the MOCC membrane in Fig. 1 is illustrated as a composite of two phases: one being carbonate-ion conducting and another being oxide-ion conducting. Compared to the carbonate-ion conductor, the oxide-ion conductor has much lower ionic conductivity in the temperature of 500–650 ◦ C, making it the limiting factor to the total CO2 flux. In fact, the highest permeability of CO2 of a LSCF + MC dual-phase membrane measured so far is only in the order of magnitude of 10−9 mol m−2 s−1 Pa−1 at 650 ◦ C in one study [13] and 10−12 mol m−1 s−1 Pa−1 for other ceramiccarbonates dual-phase materials (note: different unit is used for this study) at the same temperature in other studies by both experiment [14] and modeling [15]. Similarly, a MECC illustrated in Fig. 1 can also be comprised of two phases: one being a carbonate-ion conductor and another being an electronic conductor. Compared to the MOCC, the CO2 flux of a MECC can be higher, simply because of higher carbonate-ion and electron conductivities (for example, CO2− =

Fig. 1. Schematic of the working principles of MECC and MOCC based CO2 separation membranes: (a) MOCC; (b) MECC.

Therefore, developing an alternative high-temperature membrane technology to separate/capture CO2 more selectively and efficiently for existing power plants is in great need. One category of gas separation membranes that has the potential to meet these requirements is the mixed ionic and electronic conductors [11–14]. Different from those absorption and sizeexclusion counterparts, an electrochemical transport membrane only allows the electrochemically active species to transport through the membrane. Therefore, its selectivity is exclusive. Since this membrane normally operates at elevated temperatures, it also has an excellent compatibility with any high-temperature process stream. The best example of this category of materials is the oxygen transport membranes, or OTMs, which have been demonstrated in producing pure oxygen and synthetic gas at large scale. The key feature of these materials is the concomitant fast oxide-ion and electronic transport enabling high oxygen flux. Applying this principle to the CO2 separation membrane, two types of mixed conductors can be envisioned: (1) mixed carbonate-ion and oxide-ion conductor (MOCC); (2) mixed carbonate-ion and electron conductor (MECC). Fig. 1 shows the working principles of these two membranes along with CO2 -separation enabling surface reactions. Based on these fundamental reactions, it is evident that MOCC is more suitable for CO2 separation from a reducing stream such as water-gas shift gas (CO2 + H2 ) where the chemical gradients of CO2 and O2 exist in opposite direction across the membrane. On the other hand, MECC is more adequate for CO2 separation from an oxidizing stream such as flue gas (CO2 , O2 , N2 ) where the chemical gradients of CO2 and O2 exist in the same direction across the membrane. In this case, both CO2 and O2 will permeate through the membrane, the mixture of which can be further separated by a

0.3 S/cm and e = 6.3 × 105 S/cm at 650 ◦ C). The first reported MECC was made of a stainless steel and molten carbonate (MC) dual-phase composite (SS–MC) [11]. The reported permeability is almost one order of magnitude higher than that of MOCC (e.g., 2.5 × 10−8 mol m−2 s−1 Pa−1 at 650 ◦ C). However, the chemical interaction between SS and MC is a potential problem affecting long-term stability. In this work, we report for the first time a new dual-phase MECC membrane comprising of Ag and MC. The selection of Ag against SS is primarily based on the fact that Ag has an excellent chemical compatibility with MC. The high and stable electronic conductivity of Ag along with its operatability in the temperature range of 500–650 ◦ C can also favor a high and stable CO2 flux. Since the research in this field is scarce at present time, these promising results are anticipated to inspire more studies to this important subject in the very near future.

2. Experimental procedures 2.1. Preparation of Ag–MC MECC The molten carbonate (MC) phase used in this work is a binary eutectic mixture of Li2 CO3 and K2 CO3 (99%, Alfa Aesar) in a molar ratio of 62:38, prepared by solid state reaction at 600 ◦ C for 2 h. The obtained melt was then broken up and ball-milled into submicrometer powders. The powders were then intimately mixed with silver powders (1–3 m, 99.5%, Alfa Aesar) in a predetermined volumetric ratio of 50:50, followed by second-round ball milling. After drying, the obtained powders were pressed uniaxially into pellets of 19 mm in diameter with a static pressure of 150 MPa. The Ag–MC pellets were finally fired in air at 650 ◦ C for 2 h. 2.2. Permeation characterization The CO2 permeation characteristics of the Ag–MC were studied with a homemade permeation cell, the configuration of which is schematically shown in Fig. 2. The Ag–MC pellet was first sealed to a supporting alumina tube using a modified sealant containing a mixture of a ceramic cement (Ceramabond 552-VFG, Aremco Products) and a commercial glass powder (Schott glass GM31107). A second short alumina tube was then bonded to the top of the sample for the purpose of guiding the feed gas. The assembly is then subject to the curing condition required by the Ceramabond 552-VFG cement: room temperature 1–4 h and 93 ◦ C for a minimum of 4 h. The feed gas is a pre-mixed CO2 –O2 gas in a molar ratio of 1:1 flowing at

Please cite this article in press as: N. Xu, et al., Silver-molten carbonate composite as a new high-flux membrane for electrochemical separation of CO2 from flue gas, J. Membr. Sci. (2012), doi:10.1016/j.memsci.2012.02.001

G Model MEMSCI-11189; No. of Pages 5

N. Xu et al. / Journal of Membrane Science xxx (2012) xxx–xxx


Fig. 3. XRD patterns obtained from the polished surface of an Ag–MC sample before and after long-term flux measurement.

Fig. 2. Schematic of experimental setup for permeation measurement.

concentrations swept from a 650 ◦ C-reactor and analyzed by GC is 1.65% CO2 , 0.90% O2 and 0.01% N2 . This corresponds to a CO2 flux density of 0.919 mL min−1 cm−2 and an O2 flux density of 0.501 mL min−1 cm−2 . The leakage flux density for N2 is about 0.006 mL min−1 cm−2 , which implies a respective leakage flux density of 0.015 mL min−1 cm−2 for CO2 and O2 . This level of leakage accounts for insignificant 1.6% and 3.0% of the total flux densities for CO2 and O2 , respectively. 2.3. Other characterizations The phases in the Ag–MC were examined by X-ray diffraction (XRD) using a MiniFlexTM II bench-top XRD system diffractometer (Rigaku Corporation, Japan). The diffraction angles were scanned from 2Â = 20–90◦ at a step size of 2◦ /step. The microstructure of the sample was captured using a field emission scanning electron microscopy (FESEM, Zeiss Ultra). In order to better distinguish the Ag and MC phases, the back scattered electron (BSE) mode was particularly used. 3. Results and discussion 3.1. XRD analysis The XRD pattern of the synthesized Ag–MC was collected from the surface of a polished sample and is displayed in Fig. 3. Besides the Ag and the carbonate phases, it is evident that no other phases are detectable within the limit of X-ray diffractometer, suggesting no chemical reaction occurred between Ag and carbonate during synthesis process. The measured flux densities of CO2 and O2 at a thickness of 1.67 mm are shown in Fig. 4 as a function of reciprocal temperature. As expected, both fluxes increase with temperature and follow closely Arrhenius relationship up to 650 ◦ C. Above 650 ◦ C, they begin to decrease with temperature. This behavior has also been observed in the SS–MC membrane [11], which is likely related to the loss of MC as a result of Ag sintering suggested by the postmortem SEM analysis (see later in Fig. 7(b)). Overall, the level of CO2 and O2 flux densities of Ag–MC membrane is outstanding, achieving 0.82 mL cm−2 min−1 and 0.43 mL cm−2 min−1 for CO2 and O2 , respectively, at 650 ◦ C. The calculated apparent activation energies are 0.68 and 0.61 eV for CO2 and O2 flux densities, respectively. The close value for the two different active species seems to infer that the activation processes for CO2 and O2 are most likely involved in the same electrochemical reaction that enables the ultimate bulk transport of CO3 2− . This is exactly suggested by the

100 mL min−1 , resulting in 50 mL min−1 CO2 and 50 mL min−1 O2 . A tracer gas, N2 , was also added to the feed gas at a flow rate of 20 mL min−1 for leak detection. At the permeate side, high-purity helium (99.999%) at a flow rate of 50 mL min−1 was used as the sweeping gas, the composition of which was analyzed by an online micro-GC (Varian 490-GC, Varian). To ensure the accuracy, the GC was calibrated with three standard gas compositions for each gas of interest (CO2 , O2 , N2 ). The final flux density of each active gas was calculated from an averaged gas composition of a total of ten consecutive readings. For all gas flows, commercial mass flow controllers (Smart-Trak, 50 Series) specifically calibrated for the gas under use were employed to control the mass flow rates. The temperature was varied from 500 to 700 ◦ C in an interval of 10–20 ◦ C. The long-term stability test was only performed at 650 ◦ C. The flux densities of the CO2 (total), O2 (total) and N2 (leakage), were calculated using the following equations:

 cN Q  JN  2,leak = (1 − c − 2 − c ) × S  cO2  N2 CO2  c Q
CO ,total S (1 − cCO2 − cO2 − cN2 )  2   cO2 Q J  O2 ,total = ×

(1) (2) (3)





(1 − cCO2 − cO2 − cN2 )


where, cN2 , cCO2 and cO2 are the measured concentrations of N2 (leakage), CO2 (total) and O2 (total), respectively. Q is the flow rate of the helium sweeping gas, and S is effective reaction area of the sample. The concentration of N2 was found very low (<0.05%) for all measurements, indicating a well sealed cell. The final fluxes of the permeated CO2 and O2 were corrected JCO2 = JCO2 ,total − 2.5 × JN2,leak (4) by where the correction coef(5) JO2 = JO2 ,total − 2.5 × JN2,leak ficient of 2.5 is the ratio between CO2 flow rate (50 mL min−1 ) and N2 flow rate (25 mL min−1 ). The low N2 concentration in the permeate gas ensures a low leakage flux, and therefore better accuracy of the obtained results. As an example, a representative set of gas

Please cite this article in press as: N. Xu, et al., Silver-molten carbonate composite as a new high-flux membrane for electrochemical separation of CO2 from flue gas, J. Membr. Sci. (2012), doi:10.1016/j.memsci.2012.02.001

G Model MEMSCI-11189; No. of Pages 5 4

N. Xu et al. / Journal of Membrane Science xxx (2012) xxx–xxx

Fig. 4. Temperature dependence of CO2 and O2 flux densities.

Fig. 6. Long-term stability of CO2 and O2 flux densities measured at 650 ◦ C.

Fig. 5. Permeance of Ag–MC membrane as a functional of temperature.

reaction shown in Fig. 1(b). Furthermore, the flux density of CO2 is almost twice of that of O2 in the whole temperature range, also confirming the electrochemical reaction mechanisms described in Fig. 1(b). In order to compare with the literature data, the flux densities in Fig. 4 have been converted into permeance defined as JCO2 / PCO2 [11], and are shown in Fig. 5. The permeance of Ag–MC reaches 1.4 × 10−7 mol m−2 s−1 Pa−1 at 650 ◦ C, which is a factor of six higher than that reported in SS–MC membrane even though the CO2 /O2 chemical gradient in this study is only one-half of that used in ref. [11]. Compared to the SS–MC membrane, the high level of perme-

ation performance exhibited by the Ag–MC membrane is attributed to (i) better chemical compatibility of MC with Ag; (ii) better catalytic activity over CO2 and O2 activation processes, and (iii) better electronic conductivity of Ag in both oxidizing and reducing atmospheres. The long-term stability of CO2 and O2 flux densities in the Ag–MC membrane is shown in Fig. 6. It is interesting to note that the fluxes increase with time during the first 20 h and then slightly decrease in the next 60 h. To understand this phenomenon, SEM analysis was conducted on both prior- and post long-term stability test samples, the SEM images of which are shown in Fig. 7. The XRD pattern of the post-test sample shown in Fig. 3 did not reveal any noticeable difference in phase composition from the original sample. By comparing the morphologies of the two samples, however, it becomes evident that excessive grain growth of Ag has taken place during the measurement. The consequence of this grain growth is two-fold: reduction in sample size, both in diameter and thickness, and repelling of MC from the Ag network. The first effect can explain the initial increase of fluxes as the result of reduced thickness. The second effect is the loss of MC, which may account for the decrease in flux. In fact, some escaped MC from the sample was evident on the surface of alumina supporting tube after the test. To retain MC in the Ag network will be one of the focuses of our future study. The knowledge learned from conventional molten carbonate fuel cell technology will be considered to guide this research.

Fig. 7. A typical microstructure of the Ag–MC membrane in BSE mode: (a) prior test; (b) post long-term stability test. White phase: Ag; dark phase: MC.

Please cite this article in press as: N. Xu, et al., Silver-molten carbonate composite as a new high-flux membrane for electrochemical separation of CO2 from flue gas, J. Membr. Sci. (2012), doi:10.1016/j.memsci.2012.02.001

G Model MEMSCI-11189; No. of Pages 5

N. Xu et al. / Journal of Membrane Science xxx (2012) xxx–xxx


4. Conclusions An Ag–MC dual-phase composite has been demonstrated for the first time with high CO2 and O2 fluxes in the temperature range of 500–650 ◦ C. The permeated CO2 flux density is the highest among the reported similar metal–MC dual-phase membranes, reaching 0.82 mL cm−2 min−1 at 650 ◦ C. The close activation energy and nearly 2:1 ratio for CO2 and O2 flux densities compellingly confirm the CO2 transport mechanism predicted from electrochemical surface reactions. The peak flux density observed at 650 ◦ C is hypothesized to be related with the loss of MC from the Ag network. Long-term stability test shows an initial increase in flux density for the first 20 h, followed by a gradual decrease for the rest of 60 h. This transient behavior is explained based on the Ag sintering mechanism that was revealed by SEM analysis conducted on a prior- and post-test sample. Acknowledgement The authors would like to thank DARPA (W91CRB-10-1-0007) for financial support. References
[1] U.S. Energy Information Administration, Energy Perspective 1949-2010, October 19, 2011.

[2] U.S. Energy Information Administration, Annual Energy Outlook 2011, April 26, 2011. [3] DOE/NETL, Carbon dioxide capture and storage RD&D roadmap, December 2010. [4] DOE/NETL, Advanced carbon dioxide capture R&D program: technology update, September 2010. [5] R. Bredesen, K. Jordal, O. Bolland, High-temperature membranes in power generation with CO2 capture, Chem. Eng. Process. 43 (2004) 1129–1158. [6] B.J.P. Buhre, L.K. Elliott, C.D. Sheng, R.P. Cupta, T.F. Wall, Oxygen-fuel combustion technology for coal-fired power generation, Prog. Ener. Comb. Sci. 31 (2005) 283–307. [7] C.F. Hendricks, Carbon dioxide removal from coal-fired power plants, Energy Environ. 1 (1994) 19–49. [8] D. Shekhawal, D.R. Luebke, H.W. Pennline, A review of carbon dioxide selective membranes, DOE/NETL, 2003. [9] H. Kawamura, et al., Dual-ion conducting lithium zirconate-based membranes for high temperature CO2 separation, J. Chem. Eng. Jpn. 38 (5) (2005) 322–328. [10] E.J. Granite, T. O’Brien, Review of novel methods for carbon dioxide separation from flue and fuel gases, Fuel Processing Technol. 86 (2005) 1423–1434. [11] S.J. Chung, J.H. Park, D. Li, J.I. Ida, I. Kumakiri, Y.S. Lin, Dual-phase metalcarbonate membrane for high-temperature carbon dioxide separation, Ind. Eng. Chem. Res. 44 (2005) 7999–8006. [12] K.S. Lackner, A.C. West, J.L. Wade, Membrane for carbon dioxide separation at high temperatures, in: U. Columbia (Ed.), WO2006113674, United States, 2006. [13] M. Anderson, Y.S. Lin, Carbonate–ceramic dual-phase membrane for carbon dioxide separation, J. Membr. Sci. 357 (2010) 122–129. [14] J. Wade, C. Lee, A.C. West, K.S. Lackner, Composite electrolyte membranes for high temperature CO2 separation, J. Membr. Sci. 369 (2011) 20–29. [15] J. Wade, K.S. Lackner, A.C. West, Transport model for a high temperature, mixed conducting CO2 separation membrane, Solid State Ionics 178 (2007) 1530–1540.

Please cite this article in press as: N. Xu, et al., Silver-molten carbonate composite as a new high-flux membrane for electrochemical separation of CO2 from flue gas, J. Membr. Sci. (2012), doi:10.1016/j.memsci.2012.02.001