Mechanism of High-Temperature CO2 Sorption on Lithium Zirconate

ZrO3 powders were prepared from lithium carbonate (Li2-. CO3) and zirconium oxide (ZrO2) by the solid-state method, and the CO2 sorption/desorption pr...
3 downloads 0 Views 168KB Size
Download PDF
Environ. Sci. Technol. 2003, 37, 1999-2004

Mechanism of High-Temperature CO2 Sorption on Lithium Zirconate JUN-ICHI IDA AND Y. S. LIN* Department of Chemical Engineering, University of Cincinnati, Mail Location 171, Cincinnati, Ohio 45221-0171

Lithium zirconate (Li2ZrO3) is one of the most promising materials for CO2 separation from flue gas at high temperature. This material is known to be able to absorb a large amount of CO2 at around 400-700 °C. However, the mechanism of the CO2 sorption/desorption process on Li2ZrO3 is not known yet. In this study, we examined the CO2 sorption/desorption mechanism on Li2ZrO3 by analyzing the phase and microstructure change of Li2ZrO3 during the CO2 sorption/desorption process with the help of thermogravimetric analysis (TGA), differential scanning calorimetry (DSC), and X-ray diffraction (XRD) analyses. Li2ZrO3 powders were prepared from lithium carbonate (Li2CO3) and zirconium oxide (ZrO2) by the solid-state method, and the CO2 sorption/desorption property was examined by TGA. It was shown that pure Li2ZrO3 absorbs a large amount of CO2 at high temperature with a slow sorption rate. Addition of potassium carbonate (K2CO3) and Li2CO3 in the Li2ZrO3 remarkably improves the CO2 sorption rate of the Li2ZrO3 materials. DSC analysis for the CO2 sorption process indicates that doped lithium/potassium carbonate is in the liquid state during the CO2 sorption process and plays an important role in improving the CO2 uptake rate. XRD analysis for phase and structure change during the sorption/ desorption process shows that the reaction between Li2ZrO3 and CO2 is reversible. Considering all data obtained in this study, we proposed a double-shell model to describe the mechanism of the CO2 sorption/desorption on both pure and modified Li2ZrO3.

Introduction At present, we rely greatly on fossil fuels to provide affordable electricity, which is the basis of the economy and our daily lives. Many energy forecasts predict that this situation will last for a while or possible even become worse in the future. Flue gas from power plants, especially from coal-burning power plants, is one of the major sources for the generation of CO2 as a greenhouse gas in the atmosphere. Therefore, development of CO2 reduction techniques is an urgent necessity (1). Since the flue gas from a coal-burner is hot (for example, around 400 °C after the economizer), it is highly desirable to separate CO2 from flue gas at a high temperature without cooling the flue gas to room temperature or even lower. Transformation of the captured warm CO2 into usable or valuable products (e.g., fuels or chemicals) (2), if it becomes possible, will improve further the efficiency and economics of overall processes for power generation and CO2 sequestration. The high-temperature CO2 separation process using an inorganic adsorbent or membrane is one of the more * Corresponding author e-mail: [email protected]; telephone: (513)556-2769; fax: (513)556-3473. 10.1021/es0259032 CCC: $25.00 Published on Web 04/02/2003

 2003 American Chemical Society

promising ways for CO2 capture, utilization, and sequestration. Various adsorbents have been studied for many years for use in adsorption processes for separation of CO2 from gas mixtures. However, most of the studies are for CO2 separation at ambient temperature. Carbon-based adsorbents (3), metal oxide sorbents (4), and hydrotalcite-like compounds (5) show CO2 sorption properties at relatively high temperature. However, CO2 sorption capacity is still low (e.g., 0.65 mol/kg at 400 °C for hydrotalcite) (5). A review covering these topics can be found in a paper by Yong et al. (6). Development of inorganic membranes for CO2 separation has also received increasing attention in the past few years. Zeolite membranes (especially of Y type) offer good permselectivity and high permeance for CO2 over N2 at low temperatures (350 °C. Li2ZrO3 is another material for CO2 separation that has received increasing attention in the past few years. Nakagawa and Ohashi (15, 16) examined a Li2ZrO3 adsorbent that absorbs/desorbs CO2 in the appropriate high-temperature range (around 400-800 °C). Since Li2ZrO3 has excellent CO2 sorption characteristics, such as a large CO2 sorption capacity (about 4.5 mol/kg) and a small volume change during CO2 sorption/desorption, it is considered as a material with potential for use in CO2 separation from flue gas at high temperature. It was also reported that the addition of lithium/ potassium carbonates to Li2ZrO3 increased the CO2 sorption rate when compared to the pure Li2ZrO3 (15, 16). Although they presumed the role of lithium/potassium carbonate, the detailed mechanism of CO2 sorption/desorption process on Li2ZrO3 is not understood yet. Understanding the CO2 sorption and desorption mechanism for Li2ZrO3 is critical to the design and development of an adsorbent or a membrane using Li2ZrO3 for CO2 separation at high temperature. In this study, we examined the CO2 sorption/desorption mechanism on Li2ZrO3 by analyzing phase and microstructure changes of Li2ZrO3 during the CO2 sorption/desorption process with the help of thermogravimetric analysis (TGA), differential scanning calorimetry (DSC), and X-ray diffraction (XRD) analyses.

Experimental Section Synthesis and Characterization of Li2ZrO3. Pure Li2ZrO3 powders were prepared by the solid-state method. Starting materials were reagent-grade Li2CO3 and ZrO2 (both from Alfa Aesar) in a 1:1 ratio. The materials were weighed, ground, and intimately mixed in an agate mortar with a suitable amount of acetone or ethanol. The mixtures were calcined in air at 850 and 1000 °C for 12 h. Powders of Li2ZrO3 with K2CO3 (from J. T. Baker) (referred to as the modified Li2ZrO3) were also prepared. Although the same preparation procedure was used as mentioned above, the calcination temperature was at 850 °C, and the composition in molar ratio of the starting materials (Li2CO3:ZrO2:K2CO3) was 1.1:1.0:0.2, respectively. All the obtained samples were analyzed by XRD (a Siemens D-50 XRD with Cu KR1 radiation). CO2 Sorption and Desorption on Pure and Modified Li2ZrO3. CO2 sorption/desorption properties of the obtained VOL. 37, NO. 9, 2003 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

1999

FIGURE 1. CO2 sorption on pure and modified Li2ZrO3 (at PCO2 ) 1 atm). pure and modified Li2ZrO3 were tested by TGA in a microelectronic recording balance system (CAHN C-1000). About 150 mg of pure or modified Li2ZrO3 powder was placed in the sample pan. The samples were first dried by passing dry air for 0.5-1 h at 400-500 °C. Then CO2 sorption was started by changing the purge gas from dry air to 100% CO2 or to dry air containing 50% CO2 at the same temperature. After that, CO2 desorption was performed by raising the temperature to 780 °C. At the last stage, the purge gas was changed from 100% CO2 or dry air containing 50% CO2 to dry air at 780 °C to increase the CO2 desorption rate. The gas flow rate was maintained at 150-200 mL/min by mass flow controllers. A tubular furnace was use to control the temperature. Characterization. To analyze the structural change of the modified Li2ZrO3 during the CO2 sorption/desorption process, the modified Li2ZrO3 was rapidly quenched to the room temperature after the CO2 sorption at 500 °C and also after the CO2 desorption at 780 °C. The quenched samples were analyzed by XRD. To examine the effect of lithium/potassium carbonate on the CO2 sorption of modified Li2ZrO3, DSCTGA (TA Instrument, SDT 2960) was carried out for both pure and modified Li2ZrO3 under the flow of CO2 (flow rate: 100 mL/min). The samples were first equilibrated at 50 °C, then heated to 110 °C at a ramping rate of 10 °C/min, held at 110 °C for 30 min, and finally heated to 1000 °C at a ramping rate of 10 °C/min.

Results CO2 Sorption. Figure 1 shows the weight uptake of pure Li2ZrO3 powder at 500 °C after the surrounding gas was switched from pure dry air to pure CO2 (PCO2 ) 1 atm). The figure shows a slow but clear increase in the sample weight. It is considered that this weight increase corresponds to CO2 sorption on Li2ZrO3 based on

Li2ZrO3 + CO2 ) Li2CO3 + ZrO2

(1)

It is clear that pure Li2ZrO3 can absorb CO2 up to around 20 wt % of its sample weight at 500 °C within 10 000 min. Although this material has very high CO2 sorption capacity, the sorption rate is low. Figure 1 also shows a weight uptake of the modified Li2ZrO3 powders under 100% CO2 flow at 400 °C. For this sample, the CO2 sorption rate is also slow at this temperature and close to that of the pure Li2ZrO3 at 500 °C. It should be noted that the theoretical maximum weight uptake for the pure and modified Li2ZrO3 is 29 and 23 wt %, respectively. It will take much longer for the actual weight uptake to reach the theoretical maximum in the CO2 sorption process. The CO2 sorption rate of the modified Li2ZrO3 is improved dramatically at 500 °C as shown in Figure 2. The sample 2000

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 37, NO. 9, 2003

FIGURE 2. CO2 sorption and regeneration on the modified Li2ZrO3. Sorption process: CO2 50% balanced by dry air at 500 °C. Desorption process: CO2 50% balanced by dry air at 780 °C f dry air at 780 °C. Gas flow rate: 150 mL/min. absorbs up to 20 wt % of CO2 of the sample weight within 250 min. This is about 40 times faster than the pure Li2ZrO3 sample even under 50% CO2 flow. About 80% of the absorbed CO2 is desorbed from the modified sample after the temperature is raised to 780 °C (within 30 min) in the same CO2containing gas. The remaining absorbed CO2 can be completely desorbed after the surrounding gas is switched to pure dry air. This completes the first sorption and desorption cycle. Figure 2 also shows the second sorption and desorption cycle, which is identical to the first one. It should be noted that, if the CO2 sorption amount is plotted against CO2 partial pressure, the sorption isotherm is characterized by a step function, which is quite different from the isotherm of the other adsorbents based on physical sorption. For the practical use of this sorbent in a coal-burning power plant, the heat required for sorbent regeneration could be obtained by bypassing warmer flue gas from the flue gas line before the economizer. Microstructure Change of Lithium Zirconate during CO2 Sorption/Desorption Process. The microstructure change of the modified Li2ZrO3 before and after sorption of CO2 was studied by XRD as illustrated in Figure 3. At the point (b) when the sample was quenched after CO2 sorption, most of the sample (about 90%) had already reacted with CO2. Also, at the point (c) when the sample was quenched after CO2 desorption, the CO2 was desorbed almost completely. Figure 4b,c shows the XRD patterns of the quenched samples. The XRD pattern of the modified Li2ZrO3 before CO2 sorption (original state) is also given in Figure 4a. Compared to Figure 4a, the XRD pattern in Figure 4b shows that the peaks of the Li2ZrO3 monoclinic structure completely disappeared and that the ZrO2 peaks are present instead. This result indicates that, after the CO2 sorption process, Li2ZrO3 reacted almost completely with CO2 to become ZrO2 and Li2CO3. However, the peaks of Li2CO3 were undetectable. This result may indicate that the Li2CO3 produced by the carbonation reaction is in the liquid state. The Li2CO3 that was produced becomes amorphous and undetectable by XRD after it is quenched. This will be discussed in detail with the DSC-TGA results. The XRD pattern in Figure 4c includes the peaks of only the monoclinic Li2ZrO3, without other peaks. This means that, after CO2 desorption at 780 °C, Li2CO3 and ZrO2 react again and return to Li2ZrO3 with a monoclinic structure by releasing the CO2. These results confirm that the reaction between Li2ZrO3 and CO2 is also reversible on the microstructural level during the CO2 sorption/desorption process.

FIGURE 3. Quenching points of Li2ZrO3 during CO2 sorption and desorption for microstructure analysis. Sorption process: CO2 50% balanced by dry air at 500 °C. Desorption process: CO2 50% balanced by dry air at 780 °C f dry air at 780 °C. Gas flow rate: 150 mL/min.

heat of Li/K mixture: approximately 170 J/g). This indicates that an excess amount of lithium and potassium carbonates exist as a mixture in the modified Li2ZrO3 and start to melt (this is referred to as a molten carbonate) at 500 °C before rapid CO2 sorption starts. The starting point of the rapid increase of the sample weight agrees well with the endothermic peak. A slightly broad exothermic peak corresponding to a weight decrease is also observed at around 800 °C, similar to that of the pure Li2ZrO3. This peak corresponds to the heat associated with the CO2 desorption reaction. As to be discussed later, these DSCTGA results are critical to the identification of the CO2 sorption/desorption mechanism on Li2ZrO3. CO2 Desorption Properties of Pure and Modified Lithium Zirconate. Figure 7 shows the comparison of the CO2 desorption process in the case of pure and modified Li2ZrO3. In both cases, about 80% of the absorbed CO2 is desorbed from the sample within 80 min for the pure Li2ZrO3 and within 50 min for the modified Li2ZrO3. The CO2 was desorbed after the temperature was raised to 780 °C in the same CO2containing surrounding gas as used in the CO2 sorption process. The remaining absorbed CO2 can be completely desorbed after the surrounding gas is switched to pure dry air. In contrast to the CO2 sorption process, however, CO2 desorption rates in both cases are similar. A small difference could come from the difference in the surrounding gas during the CO2 desorption process. The CO2 concentration in the surrounding gas for the pure Li2ZrO3 (100% CO2) is much higher than that for the modified case (50% CO2) at the beginning of the desorption process. This may result in a slightly faster CO2 desorption rate for the modified Li2ZrO3.

Discussion: CO2 Sorption Mechanism

FIGURE 4. Changes of XRD patterns of the Li2ZrO3 by CO2 sorption and desorption. Spectra a-c correspond to points a-c in Figure 3. DSC-TGA. Figure 5a,b shows the results of DSC-TGA under CO2 atmosphere for pure and modified Li2ZrO3, respectively. For pure Li2ZrO3, a broad exothermic peak is observed between 450 and 650 °C. This peak corresponds to the heat associated with the reaction between Li2ZrO3 and CO2 because the weight increase is observed at the same time and the reaction is exothermic. The sharp endothermic peak around 720 °C may be associated with both melting of the produced Li2CO3 and a CO2 desorption reaction. This is due to the fact that the sharp peak is at the same temperature as the melting point of Li2CO3 (723 °C) and also that the CO2 desorption process, indicated by the weight decrease, is endothermic. This result indicates that, in the case of pure Li2ZrO3, Li2CO3 produced during CO2 sorption reaction is in solid state. There is a sharp endothermic peak at around 500 °C for the modified Li2ZrO3 that is not present for the pure Li2ZrO3. It is known that the mixture of Li and K carbonate can form a eutectic mixture, as shown in Figure 6(17). Although the melting points of pure lithium and potassium carbonate are 723 and 891 °C, respectively, the melting point of their mixture is much lower with a minimum at 498 °C. This temperature agrees with the endothermic peak at around 500 °C (melting

Synthesis Process. To find out the CO2 sorption/desorption mechanism on Li2ZrO3, first, we should consider the preparation process of Li2ZrO3. During the preparation (calcination) process, Li2CO3 should be in the liquid state because its melting point (723 °C) is lower than the preparation temperatures (e.g., 850 °C). On the other hand, ZrO2 is in the solid state because its melting point is much higher (2700 °C). The reaction between a ZrO2 particle and the liquid Li2CO3 occurs on the surface of the ZrO2 particle, and it forms the Li2ZrO3. Immediately after the reaction between Li2CO3 and ZrO2 is complete, the size of the Li2ZrO3 produced should be slightly larger because of the volume change (134%) (15) but not so different from the original ZrO2. In the case of the modified Li2ZrO3, the mixture of lithium and potassium carbonates are also in the liquid state during the preparation process because the melting point of their mixture is much lower than the preparation temperature. In this case, an excess amount of lithium/potassium carbonates may remain on the surface of the Li2ZrO3 after the Li2ZrO3 formation reaction completes. After the preparation process, the pure and modified Li2ZrO3 are considered to be in the state depicted in Figure 8a,c. Experimental Evidences. The results obtained from the experiments that aid in the identification of the CO2 sorption mechanism are summarized below: (i) Pure Li2ZrO3 has a very slow sorption rate (at 500 °C). (ii) The CO2 sorption rate of modified Li2ZrO3 at 400 °C is similar to that of pure Li2ZrO3 but is dramatically improved at 500 °C (40 times faster). (iii) In the case of modified Li2ZrO3, the second sorption/ desorption cycle was identical to the first one. (iv) DSC-TGA results indicate that an excess amount of lithium/potassium carbonates exist as a mixture that has a low melting point (498 °C) in the modified Li2ZrO3. The mixture starts to melt at around 500 °C, and this agrees very well with the starting point of rapid CO2 sorption in the modified Li2ZrO3 case. (v) DSC-TGA results also show that, during the CO2 sorption reaction, the Li2CO3 that is produced is in the solid state in VOL. 37, NO. 9, 2003 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

2001

FIGURE 5. DSC-TGA results for pure and modified Li2ZrO3 (under CO2 purge gas).

FIGURE 6. Phase transition of carbonate mixture corresponding to the change of Li2CO3 mol % during CO2 sorption reaction.

FIGURE 7. CO2 desorption process for pure and modified Li2ZrO3. the case of pure Li2ZrO3 but mostly is in the liquid state in the case of the modified Li2ZrO3. (vi) CO2 desorption experiments show that there is little difference in CO2 desorption rate between pure and modified Li2ZrO3. 2002

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 37, NO. 9, 2003

FIGURE 8. Schematic illustration of carbonation mechanism on pure and modified Li2ZrO3. The above summary of the experimental results leads to the following conclusions. First, from results i and ii, it can be concluded that the reason for the sorption rate difference between pure and modified Li2ZrO3 should not be the particle size difference. This is because, even in the same modified Li2ZrO3 sample, the sorption rate changes dramatically when the temperature is changed from 400 to 500 °C. The consideration of the preparation process also supports this conclusion. Because the preparation process is essentially the same for both pure and modified Li2ZrO3 as mentioned before, different particle size cannot lead to such a difference in sorption rate (40 times). Result iii indicates that, after one cycle of the CO2 sorption/desorption process, the modified Li2ZrO3 returns to the original state. Results iv and v suggest that the melting of the lithium/ potassium carbonates mixture plays an important role in enhancing the CO2 sorption rate in the case of modified Li2ZrO3. This indicates that diffusion of CO2 in the solid Li2CO3 that is produced is the rate-limiting step in the pure Li2ZrO3 case. In the modified case, CO2 diffuses through the molten carbonate at a much faster rate. Results ii and vi also support this. The lithium/potassium carbonate mixture remains in

the solid state at 400 °C even in the modified case, and this will give a similar CO2 diffusion rate as that through the solid Li2CO3 in the pure Li2ZrO3 case. This can explain the similar sorption rate for the modified Li2ZrO3 at 400 °C and for the pure Li2ZrO3 at 500 °C. In the desorption process, the operational temperature (780 °C) is even higher than the melting point of pure Li2CO3 (723 °C). The produced Li2CO3 will be molten even in the pure Li2ZrO3 case. Since CO2 diffuses through liquid Li2CO3 or molten carbonate toward the outside, the desorption rate is similar in both pure and modified Li2ZrO3. This explains result vi. Proposed CO2 Sorption Mechanism. On the basis of the experimental evidences presented above and with consideration of the preparation process of Li2ZrO3, we propose a double-shell mechanism for the CO2 sorption on pure and modified Li2ZrO3 as shown in Figure 8. When pure Li2ZrO3 is exposed to CO2 at 500 °C (Figure 8a,b), CO2 molecules move to the surface of the Li2ZrO3 and react with Li2ZrO3 to form solid ZrO2 (monoclinic phase) and Li2CO3. The initial period of the relatively fast increase in the CO2 sorption uptake may correspond to the carbonation process before the formation of a dense Li2CO3 shell on the outer surface and a dense ZrO2 shell in the middle covering the unreacted Li2ZrO3 core. ZrO2 and Li2CO3 shells are both in the solid state because their melting points are higher than the temperature for CO2 sorption. After the formation of the two Li2CO3 and ZrO2 shells, the carbonation reaction can continue but at a much slower rate. As shown in Figure 8, CO2 may diffuse through the solid Li2CO3 shell and react with Li+ and O2- on the outer surface of the ZrO2 shell. Another reaction occurs on the surface of the unreacted Li2ZrO3 to generate Li+ and O2-. It is wellknown that ZrO2 contains a considerable number of defects in its crystals including oxygen vacancies (18, 19). Oxygen ions can jump through the oxygen vacancy sites, and Li+ can diffuse interstitially through the ZrO2 shell. The carbonation reaction is limited by the solid-state transport of CO2 in the Li2CO3 shell and by the transport of Li+ and O2- in the ZrO2 shell. Since the sizes of Li+ and O2- ions are much smaller than the molecule size of CO2, the diffusion of CO2 in the solid Li2CO3 layer is expected to be much slower than the diffusion of Li+ and O2- ions in the ZrO2 shell. Therefore, the diffusion of CO2 is more likely to be the rate-limiting step. In the case of modified Li2ZrO3 (Figure 8c,d), the lithium/ potassium carbonate layer covering Li2ZrO3 may melt and become a liquid layer (molten carbonate) because of its lower melting point. After that, CO2 diffuses through the molten carbonate layer and reacts with Li2ZrO3. In this case, Li2CO3 formed during the carbonation reaction is mixed with the molten carbonate to increase the volume of the molten carbonate layer. The Li:K ratio changes from 1:2 to 11:2 (Figure 6a,b) due to the addition of the Li2CO3 formed by the carbonation reaction. This lithium/potassium carbonate mixture is still present as a film covering the ZrO2 shell, as shown in Figure 8. The diffusion of CO2 in the molten carbonate, with a diffusivity of about 10-5 cm2/s at 500-600 °C (20-23), is much faster than that in the solid carbonate. This can well explain the higher CO2 sorption rate in the case of modified Li2ZrO3. We also propose a CO2 desorption mechanism as shown in Figure 9. In the desorption process, the temperature is raised to 780 °C. Since this temperature is higher than the melting point of Li2CO3, the produced Li2CO3 is in the liquid state even in the pure Li2ZrO3 case (Figure 9b). Because the lithium/potassium carbonate mixture in the modified Li2ZrO3 is also in the liquid state, the situation is considered the same for pure and modified Li2ZrO3 in the desorption process. Therefore, for both samples, CO2 desorption follows the same steps. At first, Li2CO3 reacts with ZrO2 on the surface to form Li2ZrO3 and CO2. The CO2 that is produced diffuses through

FIGURE 9. Schematic illustration of CO2 desorption process. the liquid Li2CO3 or liquid lithium/potassium carbonate mixture to the outside, and the Li2ZrO3 forms a dense shell in the middle covering the unreacted ZrO2. After formation of the Li2ZrO3 shell, the CO2 desorption process continues with Li+ and O2- diffusion through the solid Li2ZrO3 shell. This CO2 desorption process is considered essentially the same as the Li2ZrO3 preparation process. The temperatures of the CO2 desorption and Li2ZrO3 preparation processes are 780 and 850 °C, respectively. The atmosphere is air or nitrogen. Therefore, the regenerated ZrO2 will return to Li2ZrO3 with a similar size to the initial Li2ZrO3 after CO2 sorption. This can explain the fact that the second CO2 sorption/ desorption cycle is identical to the first one. Compared to other adsorbents (e.g., hydrotalcite), Li2ZrO3 offers excellent CO2 sorption characteristics in terms of large CO2 sorption capacity, infinite CO2/N2 selectivity, and good reversibility. However, the CO2 sorption rate, even for the modified Li2ZrO3, is still low. Further improvement in the CO2 sorption rate is very important for practical applications of this group of promising sorbent materials for CO2 separation. The sorption mechanism identified in this work provides guidelines for further research in this direction.

Acknowledgments This research was supported by Ohio Coal Development Office (OCDO, Subcontract OCRC3-00-1.C1.8) and Department of Energy (FG26-00NT40824). Charlie Cooper is acknowledged for the help in preparation of this manuscript.

Literature Cited (1) Draper, E. L.; Becker, R. A. Research and Development Needs for the Sequestration of Carbon Dioxide as Part of a Carbon Management Strategy; The National Coal Council: Washington, DC, 2000. (2) Ding, Y.; Alpay, E. Process Saf. Environ. Prot. 2001, 79, 45-51. (3) Song, H. K.; Lee, K. H. Sep. Sci. Technol. 1998, 33, 2039. (4) Yong, Z.; Mata, V.; Rodrigues A. E. J. Chem. Eng. Data 2000, 45, 1093. (5) Ding , Y.; Alpay, E. Chem. Eng. Sci. 2000, 55, 3461. (6) Yong, Z.; Mata, V.; Rodrigues A. E. Sep. Purif. Technol. 2002, 26, 195. VOL. 37, NO. 9, 2003 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

2003

(7) Morooka, S.; Kuroda, T.; Kusakabe, K. Proceedings of the 5th International Conference on Inorganic Membranes. The Membrane Society of Japan, New Energy and Industrial Technology Development Organization (NEDO) and Japan Fine Ceramics Center (JFCC), Najoya, Japan, 1998; p 136. (8) Lovallo, M. C.; Gouzinis, A.; Tsapatsis, M. AIChE J. 1998, 44, 1903. (9) Poshusta, J. C.; Tuan, V. A.; Falconer, J. L.; Noble, R. D. Ind. Eng. Chem. Res. 1998, 37, 3924. (10) de Vos, R. M.; Verweij, H. J. Membr. Sci. 1998, 143, 37. (11) Lu, Y. F.; Cao, G. Z.; Kale, R. P.; Prabakar, S.; Lopez, G. P.; Brinker, C. J. Chem. Mater. 1999, 11, 1223. (12) Keizer, K.; Burggraaf, A. J.; Vroon, Z. A. E. P.; Verweij, H. J. Membr. Sci. 1998, 147, 159. (13) Krishna, R.; Vlugt, T. J. H.; Smit, B. R. Chem. Eng. Sci. 1999, 54, 1751. (14) Van de Graaf, J. M.; Kapteijn, F.; Moulijn, J. A. AIChE J. 1999, 45, 497. (15) Nakagawa, K.; Ohashi, T. J. Electrochem. Soc. 1998, 145, 1344.

2004

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 37, NO. 9, 2003

(16) Nakagawa, K.; Ohashi, T. Electrochemistry 1999, 67, 618. (17) Mamantov, G.; Braunstein, J.; Mamantov, C. B. Advances in Molten Salt Chemistry, Vol. 4; Plenum Press: New York, London, 1981. (18) Stevens, R. Zirconia and Zirconia Ceramics; Magnesium Elektron Ltd.: Twickenham, U.K., 1986. (19) Lin, Y. S.; de Vries, K. J.; Brinkman, H. W.; Burggraaf, A. J. J. Membr. Sci. 1992, 66, 211-226. (20) Claes, P.; Thirion, B.; Glibert, J. Electrochim. Acta 1996, 41, 141146. (21) Spedding, P. L.; Millis, R. J. Electrochem. Soc. 1965, 113, 559. (22) Spedding, P. L.; Millis, R. J. Electrochem. Soc. 1965, 113, 594. (23) Spedding, P. L. J. Electrochem. Soc. 1970, 117, 177.

Received for review June 20, 2002. Revised manuscript received February 1, 2003. Accepted February 18, 2003. ES0259032