Journal of Membrane Science 302 (2007) 109–114
Catalytic perovskite hollow fibre membrane reactors for methane oxidative coupling Xiaoyao Tan a,∗ , Zhaobao Pang a , Zi Gu b , Shaomin Liu b,∗∗ a
School of Chemical Engineering, Shandong University of Technology, Zibo 255049, PR China The ARC Centre for Functional Nanomaterials, School of Engineering, The University of Queensland, Brisbane OLD 407, Australia Received 28 April 2007; received in revised form 8 June 2007; accepted 16 June 2007 Available online 22 June 2007
Abstract Using the prepared La0.6 Sr0.4 Co0.2 Fe0.8 O3 (LSCF) hollow fibre membranes via the phase inversion and sintering technique, the catalytic perovskite hollow fibre membrane reactors (HFMRs) without or with additional OCM catalyst inside the fibre lumen were designed and tested for the methane oxidative coupling reaction. LSCF powder possesses almost zero C2 selectivity, but the blank LSCF hollow fibre membrane reactor shows a higher C2 selectivity up to 71.9%. Application of additional SrTi0.9 Li0.1 O3 catalyst in the fibre lumen is able to increase methane conversion and oxygen permeation rate remarkably but decrease the C2 selectivity. There is an optimum temperature to obtain the maximum C2 selectivity for either the packed or the blank membrane reactor. The membrane reactor exhibits appreciable advantages over the traditional packed bed reactor. A maximum C2 yield close to 21% can be achieved in the packed hollow fibre membrane reactor. © 2007 Elsevier B.V. All rights reserved. Keywords: Oxidative coupling of methane; Hollow fibre; Perovskite membrane; Catalytic membrane reactor
1. Introduction As petroleum reserves are dwindling around the world, conversion of natural gas into more useful chemicals and fuels is recognized as the next step to sustain economic growth and maintain fuel supplies [1,2]. Considerable world-wide researches have been conducted to develop various commercially viable processes for methane conversion via the direct or indirect routes. Compared to the indirect conversion technologies that need the energy-intensive step of syngas formation, the direct route that converts methane into higher hydrocarbons in one step by the oxidative coupling reactions (OCM) is more economically attractive and consequently has been intensively studied [3–8]. In the past two decades, considerable efforts have been placed on the development of OCM catalysts in order to make the product yields commercially feasible [9–13]. Although the C2 yield has gradually reached above 20% [14,15], it is still less than the 30% threshold for commercial consideration. In addition, vari∗
Corresponding author. Tel.: +86 533 2313676; fax: +86 533 2313676. Corresponding author. E-mail addresses: [email protected]
(X. Tan), [email protected]
(S. Liu). ∗∗
0376-7388/$ – see front matter © 2007 Elsevier B.V. All rights reserved. doi:10.1016/j.memsci.2007.06.033
ous new types of reactors have also been applied to improve the C2 yield such as countercurrent moving-bed chromatographic reactors , fluidized-bed reactors, riser simulator reactors based on “oxygen multiple injection”, fixed bed reactors with different distributed oxygen feed modes , gas recycle electrocatalytic or catalytic reactor separator and polytropic fixed-bed reactors . Although some of the C2 yields achieved using these reactor configurations are well above 20%, most of the processes proposed are too complicated to be practical. As an alternative, membrane reactors combining the reaction and separation in one unit possess many advantages for OCM including enhanced selectivity/yield and safety in operation [20–25]. Since oxygen is introduced discretely through the membrane along the length direction of the reactor into the reaction side, local oxygen concentrations in the reaction zone can be easily controlled to prevent the deep oxidation of products, thus resulting in higher selectivity. The membranes applied for oxygen addition can be either porous or dense [7,8,14,15,21–30]. Comparatively, the porous membranes such as alumina, zirconia and vycor glass possess high stability but low oxygen selectivity, while these dense mixed oxygen ionic and electronic conducting membranes are oxygen semipermeable and thus air can be used as the oxygen source without contaminating the products with
X. Tan et al. / Journal of Membrane Science 302 (2007) 109–114
nitrogen and nitrogen oxides, leading to remarkably reduced operation costs . However, the reported performance of these mixed conducting membranes for OCM is not ideal. In addition to the poor catalytic properties of membrane reactors, undesired reactor configuration is perhaps another drawback [21–23,28–30]. Hollow fibre membranes in this context appears to be a promising alternative as the hollow fibres can provide much larger area per unit volume than the disc or tubular membranes. In our previous work, La0.6 Sr0.4 Co0.2 Fe0.8 O3-a (LSCF) perovskite hollow fibre membranes were prepared and applied for OCM . A maximum C2 yield of 15.3% was achieved by introducing the mixture of methane and argon into the blank hollow fibre membranes. However, noticeable gaseous oxygen was found in the product stream with the concentration of 4–6%. Obviously, the total amount of oxygen provided by the blank membrane reactor was larger than the amount consumed by OCM reaction. It is hopeful that, by using additional OCM catalysts to expedite reaction rate between methane and oxygen, a higher C2 yield may be achieved. In the present work, the membrane reactor performances for OCM with or without OCM catalysts packed inside the fibre lumens have been evaluated and compared with the conventional pack-bed reactor.
Fig. 1. Schematic diagram of the LSCF hollow fibre membrane reactor.
particles were collected in an electrostatic precipitator. Detailed characterization of the STL catalysts were described elsewhere . In order to confirm the catalytic activity of the catalyst, 0.5 g STL powders were loaded inside a fixed bed reactor composed of quartz tube with 18 mm in diameter and 250 mm in length. A K-type thermocouple located at the catalyst bed was used to monitor the reaction temperature. The mixture of methane, oxygen, and the dilution gas (helium) were co-fed into the reactor tube. Gas compositions of the products were measured online using a gas chromatograph (Agilent 6890N) fitted with a carbon sieve S-II column with length of 3 m and TCD as the detector. Highly pure helium was used as the carrier gas and the flow rate was 30 cm3 /min.
2. Experimental 2.3. Hollow ﬁbre membrane reactors for methane coupling 2.1. Preparation of LSCF hollow ﬁbre membranes LSCF hollow fibre membranes were prepared by the phase inversion/sintering technique. The detailed procedures were described elsewhere [32–34]. In this work, the starting spinning mixture was composed of 70.49 wt% LSCF powders (surface area of 9.53 m2 /g and d50 = 0.6 m, purity >99.9%) purchased from Praxair Surface Technologies Corporation, 5.81 wt% polyethersulfone (PESf) [(Radel A-300), Ameco Performance, USA], 23.26 wt% 1-methyl-2-pyrrolidinone (NMP) [Synthesis Grade, Merck] and 0.44% polyvinyl pyrrolidone (PVP, K30) [from Fluka, Mw = 40,000]. A spinneret with the orifice diameter/inner diameter of 3.0/1.2 mm was applied to form the hollow fibre precursors. Deionized water and tap water were used as the internal and external coagulants, respectively. Sintering was carried out at 1300 ◦ C for 5 h, allowing for the formation of the gas-tight membranes. 2.2. Preparation of the OCM catalyst OCM catalyst, SrTi0.9 Li0.1 O3 (STL) was produced by a spray pyrolysis method briefly described below . The starting solution with a concentration of metal ions of 0.2 mol/L was prepared by dissolving stoichiometric quantities of Sr(NO3 )2 , C16 H36 O4 Ti and LiOH·H2 O [all from Merck, Synthesis Grade] in distilled water. The solution was ultrasonically pulverized under a resonant frequency of 1.6 MHz to form aerosol with an air stream. The aerosol was then introduced into a tubular furnace consisting of a heating zone of 300 ◦ C and a sintering zone of 900 ◦ C. The flow rate of the air carrier was 55.6 cm3 /s and the residence time in furnace tube was about 4 s. The formed
The catalytic hollow fibre membrane reactors (HFMRs) for methane coupling experiments were assembled with two gas-tight LSCF hollow fibre membranes with the i.d./o.d. of 1.2/1.7 mm and length of 250 mm, as shown in Fig. 1. In order to pack the above-prepared STL catalysts inside the hollow fibre membrane lumen, the STL paste prepared by ball-milling with anhydrous alcohol was injected into the lumen followed by subsequent drying for 24 h at room temperature. The loaded catalyst was approximately 0.15 g per hollow fibre with the packing length of 16 cm. A quartz tube (14 mm of inner diameter and 32 cm in length) was used to house the hollow fibres, which were placed in a pair of glass tubes respectively connecting to flexible silicone tubes. The sealing was achieved by using an organic sealant. A K-type thermocouple was positioned close to the centre of the hollow fibres to measure the temperature during operation. For comparison purposes, another HFMR with the same configuration containing two blank LSCF hollow fibres without packing catalysts was also assembled. The experimental setup for OCM using the HFMR is shown in Fig. 2. The membrane reactor was placed inside a tubular furnace where the uniform heating length is 5 cm. The temperature at the centre of the reactor was monitored and referred as the operation temperature. The methane and helium mixture was fed into the lumen side through the catalyst bed while the air supply was introduced cocurrently into the shell side of the reactor. The gas flow rates were controlled using mass flow controllers (D088B/ZM), which were calibrated by a soap bubble flow meter. The effluent flow rates were also measured by the soap bubble flow meter. All gas flows are quoted at standard temperature and pressure (STP). Gas compositions were measured online using
X. Tan et al. / Journal of Membrane Science 302 (2007) 109–114
Fig. 2. Schematic flowchart of the LSCF hollow fibre membrane reactor for OCM.
the gas chromatograph described above. Three analyses were performed for each experimental condition. The methane conversion and the species selectivity were calculated, respectively, by: Fout xCH4 × 100% (1) XCH4 = 1 − Fin yf Si =
Fout νi xi × 100% Fin yf − Fout xCH4
where Fin and Fout are the flow rates of methane feed (methanehelium mixture) and product stream respectively; yf and x are the methane feed mole fraction and the component mole fractions in the product stream respectively. vi is the number of carbon atoms in the carbon-containing products. By multiplying the methane conversion and C2 selectivity, C2 yield can be obtained. The oxygen permeation rate was determined by the change of oxygen contents in the air stream: FO2 =
0.79FAir xO s 1 − xO 2
is where Fair is the air feed flow rate, cm3 (STP)/min, and xO 2 the outlet molar fraction of oxygen in the air stream. Calculations of the carbon balance for each experiment showed that the errors were within 5%. Therefore, coking deposition on the catalysts and membranes may be negligible. Also, nitrogen balance indicated that the gas leakage through the joints or sealant was less than 1%, indicating its contribution to oxygen permeation can be neglected.
higher temperatures, methane conversion is gradually improved when the temperature rises. For C2 selectivity, it increases noticeably with increasing temperature from 998 to 1100 K and then decreases with temperature in the range from 1100 to 1198 K. C2 yield first increases with increasing temperature and then slightly decreases from 1100 to 1198 after reaching a maximum value of 11.9% at a temperature of 1098 k with a corresponding C2 selectivity of 42.7%. It should be mentioned that the catalytic properties of other catalysts for OCM like 10 wt%La2 O3 -20 wt%SrO/CaO and La0.6 Sr0.4 Co0.2 Fe0.8 O3-a (LSCF) were also investigated. We found that La–Sr–Ca catalyst exhibits much lower C2 selectivity and yield than STL; LSCF powder possesses almost zero selectivity to C2 product, but very high selectivity for methane combustion. In terms of C2 yield, STL is perhaps not the perfect OCM catalyst, but in this work, we still packed it inside the hollow fibre lumen to evaluate the performance of the catalytic hollow fibre membrane reactors. 3.2. Performance of the LSCF hollow ﬁbre membrane reactors for OCM The results of OCM in the LSCF hollow fibre membrane reactors at different temperatures are shown in Fig. 4, where the
3. Results and discussion 3.1. Performance of the SrTi0.9 Li0.1 O3 packed bed reactor for OCM Fig. 3 shows the OCM results over the SrTi0.9 Li0.1 O3 catalyst in a packed bed reactor at different temperatures, where the co-feed flow rates of CH4 , O2 and He were 19.7, 6.8 and 11.6 cm3 /min, respectively. Because of higher reaction rates at
Fig. 3. OCM over STL catalyst in the packed bed reactor (feed flow rates, FCH4 = 19.7 cm3 /min, FO2 = 6.8 cm3 /min, FHe = 11.6 cm3 /min).
X. Tan et al. / Journal of Membrane Science 302 (2007) 109–114
Fig. 4. OCM results (a, b, c) and the corresponding oxygen permeation rates (d) in the hollow fibre membrane reactors at different temperatures (feed flow rate of the mixture of 32.5% CH4 –He = 16 cm3 /min, FAir = 60 cm3 /min).
feed flow rates of the 32.5% CH4 –He mixture and the air were 16 and 60 cm3 (STP)/min, respectively. As expected in Fig. 4a, methane conversion increases with temperature as a result of the enhancement of reaction rate. At the same temperature, the methane conversion in catalyst-packed membrane reactor was remarkably higher than that in blank membrane reactor because of the improved reaction activity provided by additional STL catalyst. For example, at the temperature of 1198 K, methane conversion rose from 8 to 50% when STL catalyst was added. Fig. 4b describes the effects of temperature on C2 selectivity. For both membrane reactors, the C2 selectivity increases, and after reaching a maximum, then decreases with increasing temperature. Obviously, too high temperatures favor the deep oxidation of C2 products more than their formation. Noteworthy is that although the LSCF powder itself does not show any C2 selectivity, but the blank LSCF hollow fibres exhibit much higher selectivity for C2 product. The detailed mechanism for this phenomenon so far is still not clear. The highest C2 selectivity for the blank and the packed membrane reactors are 71.9 and 44.7%, respectively. The optimum temperature corresponding to the highest C2 selectivity in the absence of catalyst is 1198 K, but it is shifted to a lower temperature of 1148 K after the STL catalyst was packed. Due to the fact that STL catalyst is less selective than the blank LSCF membrane for OCM to C2 products, the application of additional catalyst in membrane reactor reduces the C2 selectivity. The C2 yields for the two kinds of membrane reactors are plotted as function of temperature in Fig. 4c. The C2 yields of both reactors always increase with increasing temperature within the investigated temperature range. The packed membrane reactor, though with a lower selectivity, shows a much higher C2 yield due to its higher methane conversion. The highest C2 yields achieved in this work are 21% and 14% for the packed and the blank membrane reactors, respectively, at the operation temperature of 1248 K. Based on the same membrane material, ten Elshof et al. studied the performance of LSCF disk-shaped membrane reactors in OCM and very low C2 yields (0.25–3%) were obtained . Comparing to this, the development of hollow fibre membrane reactors is a
big advancement because of their better performance in terms of product yield. However, the achieved C2 yield is still less than the 30% threshold for commercial consideration. The main reason is the mismatched oxygen species flux and catalytic properties provided by the reactors. OCM via such catalytic solid oxide membrane reactors is a complex process and involves both surface and gas phase reactions. The overall C2 yield or production rate is determined by the combined effects of surface catalytic activation of methane and oxygen species permeation rate. The maximum C2 yield can be achieved only under the conditions that the oxygen permeation flux and catalytic methane reaction rate match each other. The key to achieving the higher C2 yields in such hollow fibre membrane reactors is the integration of more appropriate OCM catalyst which possesses high C2 selectivity at lower oxygen partial pressures. Fig. 4d depicts the oxygen permeation rates against operating temperature in the packed and non-packed hollow fibre membrane reactors. As temperature increases, the mass transfer processes of oxygen permeation through the LSCF membranes, i.e., the bulk ionic diffusion and the surface exchange reaction, will be enhanced. On the other hand, the driving force (oxygen concentration gradient across the membrane) for oxygen permeation will also be improved by the rising temperature because the enhanced reaction with methane will consume more oxygen thus lowering the oxygen partial pressure in the methane side. Therefore, the oxygen permeation rates in both membrane reactors increase with increasing temperature as a result of the joint effects of the temperature on defect diffusion, surface reaction rate and thermal driving force. Further inspection of the oxygen flux profiles described in Fig. 4d indicates that, at the same temperature, the oxygen permeation rate through the packed membrane reactor is higher than that through the blank membrane reactor. This can be explained that larger oxygen partial pressure gradient is created by applying catalyst since more oxygen will react with methane arising from the application of the additional catalyst. Similar rules that the oxygen permeation rate is enhanced due to reactions were also observed by many other researchers [27,28].
X. Tan et al. / Journal of Membrane Science 302 (2007) 109–114
mum C2 selectivity for either the packed or the blank membrane reactor. The membrane reactor exhibits appreciable advantages over the packed bed reactor. A maximum C2 yield around 21% can be achieved using the packed hollow fibre membrane reactor although it is still lower than the 30% threshold for commercial consideration. In order to further improve the reactor performance in oxidative methane coupling reactions, hollow fibre membranes should be integrated with more proper catalysts with high C2 selectivity. Acknowledgements The authors gratefully acknowledge the research funding provided by the National Natural Science Foundation of China (NNSFC, No.20076025, No.20676073) and the financial support of the UQ Early Career Research Grants Scheme 2006001832. References
Fig. 5. Comparison of hollow fibre membrane reactor with the packed bed reactor.
3.3. Comparison of the LSCF hollow ﬁbre membrane reactor with traditional packed bed reactor A comparison of the reactor performance in terms of C2 selectivity and C2 yield against methane conversion between the hollow fibre membrane reactor and the packed bed reactor has been illustrated in Fig. 5. The methane conversion, C2 selectivity and C2 yield data were obtained at different temperatures while the feed flow rates were kept constant as described above. It can be seen that the C2 selectivity and C2 yield in the hollow fibre membrane reactor are higher than that in the packed bed reactor with an exception of the data recorded at 1098K, at which the methane conversions is around 30% (27.9 and 30.7, respectively). Note that the air feed flow rate in the membrane reactor was 60 cm3 /min, or the equivalent oxygen feed flow rate was around 12.2 cm3 /min, which is larger than that for the packed bed reactor, 6.8 cm3 /min. Moreover, the packed bed reactor needs more catalyst than that in the hollow fibre membrane reactor, and the product is contaminated with nitrogen which requires subsequent separation. Therefore, the membrane reactor exhibits appreciable advantages over the packed bed reactor. As mentioned above, a maximum C2 yield around 21% can be achieved using the packed hollow fibre membrane reactor although it is still lower than the 30% threshold for commercial feasibility. 4. Conclusion Dense LSCF hollow fibre membranes have been prepared by the phase inversion spinning/sintering method. Application of catalyst in the fibre lumens is able to increase methane conversion and oxygen permeation rate remarkably but to decrease C2 selectivity. There is an optimum temperature to obtain the maxi-
 J.H. Lunsford, Catalytic conversion of methane to more useful chemicals and fuel: a challenge for the 21st century, Catal. Today 63 (2000) 165.  H.J.M. Bouwmeester, Dense ceramic membranes for methane conversion, Catal. Today 82 (2003) 141.  G. Veser, J. Frauhammer, U. Friedle, Syngas formation by direct oxidation of methane reaction mechanisms and new reactor concepts, Catal. Today 61 (2000) 55.  L. Paturzo, F. Gallucci, A. Basile, G. Vitulli, P. Pertici, An Ru-based catalytic membrane reactor for dry reforming of methane-its catalytic performance compared with tubular packed bed reactors, Catal. Today 82 (2003) 57.  J.R. Frade, V.V. Kharton, A. Yaremchenko, E. Naumovich, Methane to syngas conversion. Part I. Equilibrium conditions and stability requirements of membrane materials, J. Power Sources 130 (2004) 77.  Y.S. Lin, Y. Zeng, Catalytic properties of oxygen semipermeable perovskite-type ceramic membrane materials for oxidative coupling of methane, J. Catal. 164 (1996) 220.  Y. Lu, A. Dixon, W.R. Mose, Y.H. Ma, Oxidative coupling of methane in a modified-alumina membrane reactor, Chem. Eng. Sci. 55 (2000) 4901.  F.T. Akin, Y.S. Lin, Oxidative coupling of methane in dense ceramic membrane reactor with high yields, AIChE J. 48 (2002) 2298.  G.E. Keller, M.M. Bhasin, Synthesis of ethylene via oxidative coupling of methane|: determination of active catalysts, J. Catal. 73 (1982) 9.  K. Otsuka, Q. Liu, M. Hatano, A. Morikawa, The catalysts active and selective in oxidative coupling of methane, alkali-doped samarium oxides, Chem. Lett. (1987) 1835.  G.J. Hutchings, M.S. Scurrell, J.R. Woodhouse, Partial oxidation of methane over samarium and lanthanum oxides: a study of the reaction mechanism, Catal. Today 4 (1989) 371.  T. Lin, Z.J. Ling, L. Lin, Roles of oxygen and carbon dioxide on methane oxidative coupling over CaO and Sm2 O3 catalysts, Appl. Catal. 115 (1994) 243.  R. Burch, G.D. Squire, S.C. Tsang, Comparative study of catalysts for the oxidative coupling of methane, Appl. Catal. 43 (1988) 105.  W. Wang, Y.S. Lin, Analysis of oxidative coupling of methane in dense oxide membrane reactors, J. Membr. Sci. 103 (1995) 219.  Y. Zeng, Y.S. Lin, Catalytic properties of yttria doped bismuth oxide ceramics for oxidative coupling of methane, Appl. Catal. A 159 (1997) 101.  A.L. Tonkovich, W.C. Robert, A. Rutherford, Enhanced C2 yields form methane oxidative coupling by means of a separative chemical reactor, Science 262 (1993) 221.  A. Pekediz, H.I. De Lasa, Methane oxidative coupling in a novel riser simulator reactor, Chem. Eng. Sci. 49 (1994) 4759.
X. Tan et al. / Journal of Membrane Science 302 (2007) 109–114
 V.R. Choudhary, S.T. Chaudari, A.M. Rajput, V.H. Rane, Beneficial effect of oxygen distribution on methane conversion and C2 selectivity in oxidative coupling of methane to C2 hydrocarbons over Lanthanum-promoted magnesium oxide, J. Chem. Soc., Chemmun. 20 (1989) 1526.  Y. Jiang, I.V. Yentekatis, I.V. Vayenas, Methane to ethylene with 85 percent yield in a gas recycle electrocatalytic reactor separator, Science 264 (1994) 1563.  S. Liu, X. Tan, K. Li, R. Hughes, Methane coupling using catalytic membrane reactors, Catal. Rev. 43 (2001) 147.  J.E. ten Elshof, H.J.M. Bouwmeester, H. Verweij, Oxidative coupling of methane in a mixed-conducting perovskite membrane reactor, Appl. Catal. A 130 (1995) 195.  Y. Lu, A.G. Dixon, W.R. Mose, Y.H. Ma, U. Balachandran, Oxygenpermeable dense membrane reactor for the oxidative coupling of methane, J. Membr. Sci. 170 (2000) 27.  F.T. Akin, Y.S. Lin, Oxygen permeation through oxygen ionic or mixedconducting ceramic membranes with chemical reactions, J. Membr. Sci. 231 (2004) 133.  A.M. Ramachandra, Y. Lu, Y.H. Ma, W.R. Moser, A.G. Dixon, Oxidative coupling of methane in porous vycor membrane reactors, J. Membr. Sci. 116 (1996) 253.  A.L.Y. Tonkovich, D.M. Jimenez, J.L. Zilka, G.L. Roberts, Inorganic membrane reactors for the oxidative coupling of methane, Chem. Eng. Sci. 51 (1996) 3051.  Y.K. Kao, L. Lei, Y.S. Lin, Optimum operation of oxidative coupling of methane in porous ceramic membrane reactors, Catal. Today 82 (2003) 255.
 X. Tan, K. Li, Oxidative coupling of methane in a perovskite hollow fiber membrane reactor, Ind. Eng. Chem. Res. 45 (2006) 142.  Z. Shao, G. Xiong, H. Dong, W. Yang, L. Lin, Synthesis, oxygen permeation study and membrane performance of a Ba0.5 Sr0.5 Co0.8 Fe0.2 O3-delta oxygenpermeable dense ceramic reactor for partial oxidation of methane to syngas, Sep. Purif. Technol. 25 (2001) 97.  H. Wang, Y. Cong, W. Yang, Oxidative coupling of methane in Ba0.5 Sr0.5 Co0.8 Fe0.2 O3-delta tubular membrane reactors, Catal. Today 104 (2005) 160.  Y.P. Lu, A.G. Dixon, W.R. Moser, Y.H. Ma, U. Balachandran, Oxygenpermeable dense membrane reactor for the oxidative coupling of methane, J. Membr. Sci. 170 (2000) 27.  E.A. Hazbun, Ceramic membrane and use thereof for hydrocarbon conversion, US Patent 4827071 (1989).  S. Liu, X. Tan, K. Li, R. Hughes, Preparation and characterization of SrCe0.95 Yb0.05 O2.975 hollow fiber membranes, J. Membr. Sci. 193 (2001) 249.  X. Tan, Y. Liu, K. Li, Preparation of La0.6 Sr0.4 Co0.2 Fe0.8 O3-␣ hollow fiber membranes for oxygen production by a phase-inversion/sintering technique, Ind. Eng. Chem. Res. 44 (2005) 61.  S. Liu, G.R. Gavalas, Oxygen selective ceramic hollow fiber membranes, J. Membr. Sci. 246 (2005) 103.  S.C. Tsai, Y.L. Song, C.C. Yang, C.S. Tsai, Ultrasonic spray pyrolysis for nanoparticles synthesis, J. Mater. Sci. 39 (2004) 3647.  Z. Pang, X. Tan, R. Ding, Z. Gu, S. Liu, Preparation, characterization and catalytic performance of SrTi0.9 Li0.1 O3 ultra-fine powders, Ceram. Int., in press.