Effects of polymer binders on separation performance of the perovskite-type 4- bore hollow fiber membranes. Separation and Purification Technology

Transcription

1 Accepted Manuscript Effects of polymer binders on separation performance of the perovskite-type 4- bore hollow fiber membranes Zhe Song, Zhicheng Zhang, Guangru Zhang, Zhengkun Liu, Jiawei Zhu, Wanqin Jin PII: S (17) DOI: Reference: SEPPUR To appear in: Separation and Purification Technology Received Date: 18 January 2017 Revised Date: 26 June 2017 Accepted Date: 26 June 2017 Please cite this article as: Z. Song, Z. Zhang, G. Zhang, Z. Liu, J. Zhu, W. Jin, Effects of polymer binders on separation performance of the perovskite-type 4-bore hollow fiber membranes, Separation and Purification Technology (2017), doi: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

2 Effects of polymer binders on separation performance of the perovskite-type 4-bore hollow fiber membranes Zhe Song, Zhicheng Zhang, Guangru Zhang*, Zhengkun Liu, Jiawei Zhu, Wanqin Jin* State Key Laboratory of Materials-Oriented Chemical Engineering, Jiangsu National Synergetic Innovation Center for Advanced Materials, College of Chemical Engineering, Nanjing Tech University, 5 Xinmofan Road, Nanjing , P. R. China Zhe Song and Zhicheng Zhang contributed equally. *To whom all correspondence should be addressed: wqjin@njutech.edu.cn, guangru.zhang@njtech.edu.cn. Contact information: Prof. Wanqin Jin State Key Laboratory of Materials-Oriented Chemical Engineering, Nanjing Tech University, 5 Xinmofan Road, Nanjing , P. R. China Tel.: ; Fax: ; wqjin@njtech.edu.cn, guangru.zhang@njtech.edu.cn. 1

3 Abstract Most perovskite oxides are commonly recognized as less tolerance to sulfur and sulfur-containing species. In a hollow fiber fabrication, polymer binders (sulfur-containing polyether sulphone (PES) and sulfur-free polyetherimide (PEI)) are expected to have different impacts on the membrane performance in terms of both permeability and stability. In this study, BaCo 0.7 Fe 0.22 Nb 0.08 O 3-δ (BCFN) 4-bore hollow fibers using PES and PEI as binders were prepared by a phase inversion and sintering technique. The sintering behavior, mechanical strength, crystal structure, morphology, composition, oxygen permeation and long-term stability were investigated systematically. Formation of sulfate was identified on the surface and in the bulk of the membrane fabricated via the PES route. The sulfur poisoning was confirmed taking place during not only the sintering stage but also the operation stage. The sulfur-containing PES contributed to a lower oxygen permeation flux than the sulfur-free PEI and also a severe degradation of permeation flux that there was more than 65% decrease over the original value within the first 100 hours at 650 o C. Membrane fabricated via the PEI route can operate for more than 300 h without obvious degradation. Key words: Perovskite oxide, Hollow fiber membrane, Polymer binder, Oxygen permeation, Sulfur poisoning 2

4 1. Introduction In recent years, perovskite-type oxides have raised considerable attention due to their versatile properties and wide applications, such as in solid oxide fuel cells [1, 2], oxygen permeable membranes for oxygen separation [3-5] and partial oxidation of hydrocarbons [6-8]. Among these applications, tremendous efforts have been focused on the perovskite-type oxygen permeable membranes because of their theoretically 100% selectivity for oxygen and their potential application in the pure oxygen production at elevated temperature. It can significantly reduce the oxygen production cost compared with the traditional oxygen producing technology (e.g. cryogenic process) and is much beneficial for energy and environment [5]. Currently, the most common perovskite-type oxides for the oxygen permeable membrane are barium- or strontium-containing oxides, such as Ba 0.5 Sr 0.5 Co 0.8 Fe 0.2 O 3-δ [9], La 1-x Sr x Co 1-y Fe y O 3-δ (0 x 1, 0 y 1) [10], BaBi 0.05 Sc 0.1 Co 0.85 O 3-δ [11], BaCo 0.7 Fe 0.2 Sn 0.1 O 3-δ [12], BaCo 0.7 Fe 0.22 Nb 0.08 O 3-δ [13], SrCo 0.9 Nb 0.1 O 3-δ F 0.1 [14]. These materials show a promising oxygen permeability. However, the critical problem of these materials is their poor chemical stability in sulfur dioxide and hydrogen sulfide [15-19], as these gases thermodynamically prefer to react with barium and strontium, and produce sulfate or sulfite which is considered oxygen non-permeable. These reactions will consequently cause a non-stoichiometry of cations of the bulk materials, and sulfur can also occupy the oxygen vacancy and block the oxygen transport path [17, 20]. These negative factors caused by gaseous sulfur-containing species would consequently decrease the membrane permeability and its stability and most of the time this degradation is irreversible. Although the sulfur poisoning during permeation operation from gaseous sulfur-containing species has been already widely realized and studied, very few work has looked into a membrane fabrication process where there is a potential sulfur contamination from polymer 3

5 binders. Up to now, there are two main kinds of polymer binders used for fabrication of hollow fiber membranes: one kind is sulfur-containing polymer, e.g. polysulphone (PS) [21, 22], polyether sulphone (PES) [22, 23]; and another kind is sulfur-free polymer, e.g. polyetherimide (PEI) [23, 24]. Sulfur-containing polymer has good spinning and membrane forming properties, for example a good rheological condition. Therefore, sulfur-containing binders, especially PES, are most common accepted and being widely used for hollow fiber fabrication currently [25, 26]. However, there were some reactions between the sulfur-containing binders and the solid perovskite oxides. Liu et al. [27] reported a small amount of BaSO 4 was formed in hollow fiber due to the fact that sulfur-containing polymer binder (PES) would react with Ba 0.5 Sr 0.5 Co 0.8 Fe 0.2 O 3-δ hollow fiber membrane, and it is the first paper noticed the poisoning effect of sulfur-containing binder on the membrane performance. Diniz da Costa et al. [28] further pointed that the formation of non-ionic conducting domains of sulfate would result in a hindering oxygen ionic diffusion. They [28] also found the sulfur-free polymer binders, such as PEI, could overcome this problem and increase the oxygen permeation flux as compared to the hollow fibers prepared using sulfur-containing binders. The problems caused by the sulfur-containing polymer binder have been realized, however, some fundamental functions of different binders on the perovskite-type membranes have not been well studied yet, such as their effects on membrane sintering, thermal expansion behavior, mechanical strength, phase structure, morphology and long-term stability. To study these fundamental problems, in this work we chose BaCo 0.7 Fe 0.22 Nb 0.08 O 3-δ (BCFN) with barium fully occupied A-site as an example material. It was manifested a high oxygen permeation performance and stability at both high and intermediate temperature (e.g. 650 o C) in a disk shape membrane [13]. 4

6 4-bore hollow fiber configuration was selected due to its approved advantages such as high mechanical strength, enhanced oxygen permeability and flexibility in control of morphology and microstructure. By using the sulfur-containing PES and sulfur-free PEI binders, two kinds of BCFN 4-bore hollow fiber membranes were prepared via a combined phase inversion and sintering technique. The sintering and thermal expansion behavior, mechanical strength, crystal structure, morphology, composition and oxygen permeation performance of the two kinds of membranes were systematically investigated. The main purposes of this work are to provide a deeper understanding of sulfur poisoning by polymer binders in the process of membrane fabrication and help to optimize the performance of perovskite-type hollow fiber membranes. 2. Experimental 2.1 Preparation of BCFN powder BaCo 0.7 Fe 0.22 Nb 0.08 O 3-δ (BCFN) powder was synthesized by a solid-state reaction method. Stoichiometric amounts of BaCO 3, Co 2 O 3, Fe 2 O 3 and Nb 2 O 5 (all chemicals with a purity of 99.9% and purchased from Sinopharm, China) were mixed by ball-milling in ethanol for 24 h, and the green powder was obtained after drying in air at 70 o C for 24 h. The BCFN oxide powder was further calcined at 900 o C for 5 h with a heating and cooling rate of 2 o C min -1. Finally, calcined powder was grinded and sieved (300 meshes) and ready for fabricating BCFN 4-bore hollow fiber precursor. 2.2 Fabrication of 4-bore hollow fiber membranes 4-bore hollow fiber membranes were fabricated by a combined phase inversion and sintering technique that was reported in our previous work [29, 30]. The spinning slurry with a composition of 5

7 67.0 wt% BCFN powder, 26.5 wt% 1-methyl-2-pyrrolidinone (NMP), 0.5 wt% polyvinylpyrrolidone (PVP), and 6 wt% polyether sulphone (PES) or polyetherimide (PEI) was ball-milled for 24 h. BCFN with PEI as binder (BCFN-PEI) and BCFN with PES as binder (BCFN-PES) 4-bore hollow fiber precursors were obtained by a custom-made 4-bore spinneret. Deionized water was used as internal and external coagulants. The extruded hollow fiber was immersed in the water for 1 h, and then transferred in ethanol for 24 h to complete solidification. After dried at room temperature, hollow fiber precursors were then sintered at 1120 o C for 10 h with a heating and cooling rate of 2 o C min -1, and the two kinds of as-prepared BCFN-PEI and BCFN-PES 4-bore hollow fiber membranes were obtained. 2.3 Characterizations The sintering behavior of hollow fiber precursors was measured in stagnant air using the apparatus of dilatometer (Netzsch, model DIL 402C, Germany). The length of hollow fiber precursor was 4 mm. The shrinkage measurement was tested in the range from 25 o C to 1150 o C with a heating rate of 5 o C min -1. The thermal expansion behavior of the as-prepared hollow fiber membranes (sintered at 1120 o C for 10 h) was also measured by the dilatometer from 25 o C to 950 o C with a heating rate of 5 o C min -1. The mechanical strength of the as-prepared BCFN-PEI and BCFN-PES 4-bore hollow fiber membranes was measured by a three-point bending test using a tensile (Model CMT6203, USA). The hollow fiber membrane samples were fixed on a sample holder with a spec gauge length of 40 mm. The crosshead speed was set at 0.01 cm min -1. The load (F m ) was recorded for each sample when the hollow fiber was fractured. 6

8 The crystal structure of the as-prepared BCFN powder, fresh and used BCFN-PEI and BCFN-PES 4-bore hollow fiber membranes were analyzed by X-ray diffraction (XRD, Bruker, model D8 Advance) using a Cu Kα radiation. The patterns of experimental diffraction were in the range of 20º 2θ 80º with an increment of 0.02º. The generator voltage was 40 kv and current was 20 ma. Morphology and microstructure of membranes were examined via a scanning electron microscope (SEM, Hitachi S-4800, Japan). Energy dispersive X-ray spectroscopy (EDX) was employed by the S-4800 scanning electron microscope (Hitachi, Japan) at an excitation of 20 kv to study elemental distributions on the membrane surface. X-ray photoelectron spectrometer (XPS Thermo ESCALAB 250, USA) equipped with an Al Ka X-ray source ( ev) was utilized to analyze the surface chemical composition and surface element valence state of the membranes. 2.4 Oxygen permeation measurement The measurement of oxygen permeation flux through BCFN-PEI and BCFN-PES 4-bore hollow fiber membranes was tested by a custom-made high-temperature permeation device depicted in our previous work [31, 32]. A 30 mm long dense BCFN 4-bore hollow fiber membrane was sealed between two dense quartz tubes (i.d.=2.6 mm, o.d.=4 mm) by a silver sealant (The effective length of membrane for oxygen permeation is about 24 mm). The sealing part is inside the furnace and the sealing method in our experiment is high-temperature sealing. The membrane and two dense quartz tube were surrounded by a larger quartz tube (i.d.=10 mm, o.d.=14 mm). High-temperature permeation device was placed in a tube furnace. The operating temperature in this work was considered as the center temperature of the furnace. The oxygen permeation flux of the BCFN-PEI and BCFN-PES 4-bore hollow fiber membranes was measured in the temperature range between 7

9 650 o C and 900 o C. Synthetic air (21% O 2 and 79% N 2 ) was fed into the shell side and helium (He) was fed into the lumen side as a sweep gas. The gas flow rates were controlled by mass flow controllers (model D07-19B), and the furnace temperature was controlled by a programmable temperature controller (Model Al-708PA, Xiamen Yudian automation technology Co.). At each operating temperature, the oxygen permeation flux was continuously monitored to a steady state for the sake of obtaining accurate oxygen permeation flux. An on-line gas chromatograph (GC, Shimadzu, model GC-8A, Japan) detected the composition of sweep gas which was used to calculate the oxygen permeation flux. Due to the complex 4-bore and asymmetric structure, the membrane area is simply defined as the outer surface area of the 4-bore hollow fiber membrane [33]. The leakage of the oxygen due to an imperfect sealing at high temperatures was less than 0.5% during all experiments. Then the oxygen permeation flux was calculated as follows: Where and are the volume concentration of oxygen and nitrogen, respectively, in the lumen side outlet obtained from the GC measurements. Q (ml (STP) min -1 ) is the flow rate of the sweep gas, and S (cm 2 ) is the effective membranes area, which defined as the outer surface area. STP refers to the standard temperature and pressure. 3. Result and Discussion 3.1 Sintering behavior Sintering behavior is important to determine the microstructure and performance of the membranes. The sintering behaviors of the BCFN hollow fiber precursor via the PEI and PES routes were studied. The linear shrinkage ΔL/L 0 (which gives the length change of sample as a function of 8

10 temperature (T) and L 0 is the initial length of sample) was given in Fig. 1a. A similar sintering curve in the temperature range of 25 o C-1150 o C was observed in the membranes fabricated via PEI and PES routes, however, BCFN-PES hollow fiber membrane had a larger shrinkage than that of BCFN-PEI hollow fiber membrane, and the shrinkage at 1150 o C for BCFN-PEI and BCFN-PES membrane was 40% and 50%, respectively. Sintering shrinking rate was obtained from the derivative of ΔL with respect to time (t) at a certain temperature. As shown in Fig. 1b, several additional peaks (arrow marked) were identified on BCFN-PES membrane at 225 o C, 450 o C and 910 o C. This might be owing to reactions of sulfur (or other sulfur-containing compounds) with the BCFN bulk phase. It is worthy to note that a sharp transition point of sintering rate was found around 1140 o C for both membranes. This temperature was the melting point of the membrane. Therefore, the sintering temperature should be slightly lower than 1140 o C. We found that these two kinds of membranes sintered at 1120 o C for 10 h were gas-tight and dense. The sintering temperate of 1120 o C was selected for both the membranes. The shrinkage at 1120 o C for BCFN-PEI and BCFN-PES membrane was 33% and 37%, respectively. 9

11 Fig. 1. Temperature dependence of (a) sintering shrinkage (ΔL/L 0 ) and (b) sintering shrinkage rate (dl/dt) of the BCFN-PEI and BCFN-PES 4-bore hollow fiber precursors. 3.2 Crystal structure XRD was used to exam the crystal and phase structure of BCFN powder and 4-bore hollow fiber membrane fabricated via the PEI and PES routes. Fig. 2 shows XRD diffraction patterns of BCFN powder, the as-prepared BCFN-PEI and BCFN-PES hollow fiber membranes. Typical perovskite diffraction peaks of BCFN powder were observed and consistent with the reported value [13]. There were no additional peaks found on the BCFN-PEI membrane suggesting sulfur-free 10

12 polyether PEI binder had few effects on the crystal and phase structure of BCFN after the membrane sintering. However, for the membrane fabricated and sintered via the PES route, there were some additional peaks (at 25.84º, 26.82º, 28.72º and 43.68º) observed in the XRD patterns. These diffraction peaks were well agreed with the characteristic peaks of BaSO 4. The formed BaSO 4 (might or might not be stoichiometric) was mainly generated from the interactions of BCFN powder and sulfur-containing polymer PES binder during high temperature sintering process. The additional peaks in the sintering rate curves observed in Fig. 1b might support these reactions. Fig. 2. XRD patterns of the BCFN powder, as-prepared BCFN-PEI and BCFN-PES 4-bore hollow fiber membranes. The formed sulfate impurities were also confirmed by XPS (Fig. 3) where peaks were identified at the bonding energy of ev and ev (assigned to the S 2P 1/2 and S 2P 3/2 peaks, respectively), indicating the existence of sulfate. Unlike the gaseous sulfur dioxide or hydrogen sulfide poisoning during operation that sulfate or sulfite second phase tended to form on the outermost surface of membrane, these second phase might locate not only on the outermost surface but also in the bulk of the membrane fabricated via the PES route. The EDX analysis (Fig. 4) shows a clear distribution of sulfur in the bulk of the BCFN-PES membrane. Fig. 2b shows the peak 110 in 11

13 Fig. 2a. Compared with the BCFN powder, it is interesting that the perovskite lattice structure of BCFN-PEI shrunk slightly (peak 110 shifts towards the higher angle) which might be owing to A- or B-site deficiency [34]. However, a shift towards lower angle of the 110 peaks of the BCFN-PES membrane suggests a lattice expansion. This might be owing to a partial substitution of oxygen or occupation of oxygen vacancy by sulfur which has a larger atomic radius than oxygen [35]. Fig. 3. X-ray photoelectron spectra of sulfur element analysis of the as-prepared BCFN-PES 4-bore hollow fiber membrane. 12

14 Fig. 4. SEM-EDX analysis of the as-prepared BCFN-PES 4-bore hollow fiber membrane. 3.3 Morphology and microstructure and mechanical strength As shown in Fig. 5, the geometry, microstructure and surface morphology of the hollow fiber membranes fabricated via the PEI and PES routes were compared and studied by a scanning electron microscope. Fig. 5a and Fig. 5b show the cross-sectional view of the as-prepared BCFN-PEI and BCFN-PES 4-bore hollow fiber membranes, respectively. The two kinds of membranes exhibited well-formed 4-bore structure. An asymmetric wall structure of the two kinds of membranes was also observed in Fig. 5c and Fig. 5d, a finger-like structure had formed near both the inner and outer walls, and a sponge-like layer appeared at the center of the membrane. The formation of the wall structure was attributed to the interactions among the solvent, non-solvent and binder during the phase inversion process. That the formation of the finger-like pore at both the inner and outer walls was related to a rapid precipitation and the sponge-like layer at the center wall was attributed to a 13

15 slow precipitation. There was no significant influence on the geometry and microstructure of hollow fiber membranes so far obtained from the two routes. The surface morphology of the as-prepared BCFN-PEI and BCFN-PES 4-bore hollow fiber membranes were showed in Fig. 5e-h. The surfaces (at both the outer and inner side) of the fresh BCFN-PEI and BCFN-PES membranes were dense and had clear grain boundaries. However, the grain size of the BCFN-PEI membrane (~10 μm) was obviously larger than that of the BCFN-PES membrane (~2 μm). One of the reasons might be the second-phase particle inhibited the grain growth of bulk material [36]. In addition, the second phase might also contribute to a low thermal expansion [37]. As can be seen in Fig. 6, BCFN-PES membrane showed a lower thermal expansion than that of BCFN-PEI membrane in temperature range of 300 o C o C. The hollow fiber membranes with the same materials shaped by different binders exhibited different mechanical strength. The three-point bending strength test was performed on BCFN-PES and BCFN-PEI membranes which were sintered at 1120 o C for 10 h. The breaking load of BCFN-PEI membrane was approximately 11.5 N, while the breaking load of BCFN-PES membrane was approximately 9.0 N, suggesting a lower mechanical strength of membrane fabricated via the PES route. When using PES as binder, some sulfate was formed around BCFN particle or on the surface BCFN particle. Formed sulfate inhabited BCFN grain growth, probably causing that the mechanical strength of BCFN-PES hollow fiber membrane is lower than that of BCFN-PEI membrane. 14

16 Fig. 5. SEM images of the as-prepared BCFN-PEI and BCFN-PES 4-bore hollow fiber membranes: (a) cross-section of BCFN-PEI membrane, (b) cross-section of BCFN-PES membrane, (c) wall of BCFN-PEI membrane, (d) wall of BCFN-PES membrane, (e) outer surface of fresh BCFN-PEI membrane, (f) outer surface of fresh BCFN-PES membrane, (g) inner surface of fresh BCFN-PEI membrane, (h) inner surface of fresh BCFN-PES membrane. 15

17 Fig. 6. Thermal expansion coefficient of the as-prepared BCFN-PEI and BCFN-PES 4-bore hollow fiber membranes in air atmosphere. 3.4 Oxygen permeation flux The effects from different polymer binders on the oxygen permeability of 4-bore hollow fiber membranes were also investigated by an oxygen permeation test. Fig. 7a shows the temperature dependence of oxygen permeation flux through the BCFN-PEI and BCFN-PES membranes. (Air and He flow rates were 120 ml(stp) min -1 and 80 ml(stp) min -1, respectively). An uptrend of oxygen permeation flux with temperature increase was observed on both membranes, this was because oxygen surface exchange reaction rate and oxygen bulk diffusion rate were improved with elevated temperature [37]. If we compared each temperature, clearly, BCFN-PEI membrane had a higher oxygen permeation flux than the BCFN-PES membrane. For example, for the BCFN-PEI 4-bore hollow fiber membrane, oxygen permeation fluxes were 1.3 and 8.0 ml min -1 cm -2 at 650 o C and 900 o C, respectively. For the BCFN-PES hollow fiber membrane, the oxygen permeation fluxes were 0.8 and 5.1mL min -1 cm -2 at 650 o C and 900 o C, respectively. In addition, the activation energies (Ea) for oxygen permeation through BCFN-PEI membrane were obviously lower than the BCFN-PES membrane (Fig. 7b). B.C.H. Steele [38] pointed out that oxygen permeation flux of mixed-conducting membrane should be more than 1 ml min -1 cm -2 for industrial application in 16

18 1992. However, Peter Vang Hendriksen et al. [39] pointed out that current target for economic feasibility of mixed-conducting membrane oxygen permeation flux is 10 ml min -1 cm -2. Hence, BCFN-PEI 4-bore hollow fiber membrane (oxygen permeation flux of 8 ml min -1 cm -2 at 900 o C) is a potential candidate for oxygen separation for industrial application in the future. Obviously, sulfur-containing polymer PES binder plays a negative effect on the permeation. The formation non-ionic conducting phase BaSO 4 on the surface and in the bulk might be the main reason for the lower permeability. However, complexity needed to be realized in a permeation process. Factors related to microstructure, including surface morphology and grain boundary diffusion may also influence overall oxygen transportation [40, 41]. For example, as discussed above, different polymer binder would result in differences in grain size and amount of grain boundaries, and these would contribute to the oxygen permeation flux in either positive or negative way. 17

19 Fig. 7. (a) Temperature dependence of oxygen permeation flux through the as-prepared BCFN-PEI and BCFN-PES 4-bore hollow fiber membranes. (b) Arrhenius plots of oxygen permeation flux through the BCFN-PEI and BCFN-PES membranes. 3.5 Long-term stability For a practical application, key challenges are lowering the operation temperature and achieving a long-term stability. Fig. 8 shows the long-term oxygen permeable performance of BCFN-PEI and BCFN-PES 4-bore hollow fiber membranes at 650 o C. The oxygen permeation flux of membranes was measured over 300 h (Air and He flow rates were fixed on 120 ml(stp) min -1 18

20 and 80 ml(stp) min -1, respectively). A sharp reduction in oxygen permeation flux of BCFN-PES membrane (more than 65% decrease over the original value) within the first 100 hours was observed and less than 0.3 ml(stp) cm -2 min -1 permeation flux was obtained over 300 h operation. However, a much more stable permeation was observed on BCFN-PEI hollow fiber membrane and there was only less than 10% decrease in the oxygen permeation flux throughout the test period and stabled over 1 ml(stp) cm -2 min -1, showing a promising long-term stability. In the PES route, second phases, i.e. BaSO 4, produced from the reactions between BCFN powder and PES during the sintering of membrane would mainly result in the reduction of oxygen permeation flux compared to that of membrane fabrication via the PIE route. Without certain knowledge of the reaction kinetics, it would be unlikely to exclude the possibility of the sulfur poisoning reaction during permeation operation. Fig. 8. Long-term stability of BCFN-PEI and BCFN-PES 4-bore hollow fiber membranes at 650 o C for 300 h. To gain insight into the reasons of magnificent degradation of the oxygen permeation flux in BCFN-PES membrane, the surface morphology, crystal structure and bulk composition were studied on the membranes after 300 h operation at 650 o C. SEM images of the used BCFN-PEI and 19

21 BCFN-PES hollow fiber membranes are showed in Fig. 9. The inner and outer surfaces of the used BCFN-PEI (Fig. 9a and Fig. 9b) and BCFN-PES (Fig. 9c and Fig. 9d) membranes were dense and crack free. However, the grain boundaries were not visually observed after a 300 h operation. Fig. 10 shows XRD patterns of used BCFN-PEI and BCFN-PES hollow fiber membranes after 300 h operation at 650 o C. The crystalline structure of the BCFN-PEI membrane was unchanged and kept consistent with fresh BCFN-PEI membrane, suggesting excellent chemical stability in a long-term operation. BaSO 4 diffraction peaks were observed on the BCFN-PES membrane but the relative intensity of BaSO 4 diffraction peaks of used BCFN-PES membrane became stronger than that of fresh BCFN-PES membrane. This phenomenon suggested more BaSO 4 was formed during operation. As shown in Table 1, similar element distributions were observed on BCFN-PEI membrane before and after experiment indicating a relative stable phase structure. However, the content of sulfur on BCFN-PES membrane inner surface before and after used was increased from 2.08 atom% to 5.2 atom%, respectively. And the content of barium was changed from atom% to atom%, respectively, supporting more BaSO 4 formed during the experiment. This result implied that the decrease of oxygen permeation flux of BCFN-PES membrane was caused by the fact that sulfur species in membrane which had incorporated barium species formed BaSO 4 could migrate from bulk to membrane surface and more BaSO 4 enriched on the surfaces during oxygen permeation process at intermediate temperature. 20

22 Fig. 9. SEM images of the used BCFN-PEI and BCFN-PES 4-bore hollow fiber membranes: (a) inner surface of used BCFN-PEI membrane, (b) outer surface of used BCFN-PEI membrane, (c) inner surface of used BCFN-PES membrane, and (d) outer surface of used BCFN-PES membrane. Fig. 10. XRD patterns of the BCFN-PEI and BCFN-PES 4-bore hollow fiber membranes before and used on-stream at 650 o C after 300 h. 21

23 Table 1. EDX analysis of fresh and used BCFN-PEI and BCFN-PES 4-bore hollow fiber membranes Samples Fresh BCFN-PEI Used BCFN-PEI Fresh BCFN-PES Used BCFN-PES Not detected Elemental content (atom %) Ba Co Fe Nb S O Conclusions The effects of different binders on separation performance of BCFN 4-bore hollow fiber membrane were systematically investigated in this study. Sulfur poisoning (formation of sulfate) was observed on the surface and in the bulk of BCFN membrane during the sintering process when using a sulfur-containing polymer polyether sulphone (PES) as binder. The formed sulfate results in lower mechanical strength, smaller particle size and lower oxygen permeability of membranes. In addition, more sulfate formed and gradually migrated from bulk to inner surface during the operation stage, which causes oxygen permeation performance lost. However, BCFN membrane with sulfur-free polymer polyetherimide (PEI) as binder could effectively avoid sulfur poisoning. By comparison, the oxygen permeation flux of BCFN-PEI hollow fiber membrane was ml min -1 cm -2 in the testing temperature range of 650 o C-900 o C, which is higher than the BCFN-PES membrane ( ml min -1 cm -2 ). Meanwhile, the long-term stability of BCFN-PEI hollow fiber membrane was superior to the BCFN-PES membrane. The oxygen permeation flux of BCFN-PEI membrane has almost no degradation during 300 h operation at 650 o C. 22

24 5. Acknowledgements This work was supported by the Innovative Research Team Program by the Ministry of Education of China (Grant No. IRT13070) and the Topnotch Academic Programs Project of Jiangsu Higher Education Institutions (TAPP). 23

25 References [1] Z.P. Shao, S.M. Haile, A high-performance cathode for the next generation of solid-oxide fuel cells, Nature. 431 (2004) [2] W. Jung, HL. Tuller, Investigation of surface Sr segregation in model thin film solid oxide fuel cell perovskite electrodes, Energy Environ. Sci. 5 (2012) [3] S.M. Hashim, A.R. Mohamed, S. Bhatia, Current status of ceramic-based membranes for oxygen separation from air, Adv. Colloid. Interface. Sci. 160 (2010) [4] Z.B. Zhang, D.J. Chen, F.F. Dong, Z.P. Shao, Efficient and CO 2 -tolerant oxygen transport membranes prepared from high-valence B-site substituted cobalt-free SrFeO 3 δ, J. Membr. Sci., 495 (2015) [5] J. Sunarso, S. Baumann, J.M. Serra, W.A. Meulenberg, S. Liu, Y.S. Lin, J.C.D. da Costa, Mixed ionic-electronic conducting (MIEC) ceramic-based membranes for oxygen separation, J. Membr. Sci. 320 (2008) [6] X.Y. Tan, L.L. Shi, G.Z. Hao, B. Meng, S.M. Liu, La 0.7 Sr 0.3 FeO 3-alpha perovskite hollow fiber membranes for oxygen permeation and methane conversion, Sep. Purif. Technol., 96 (2012) [7] H. Jiang, Z. Cao, S. Schirrmeister, T. Schiestel, J. Caro, A Coupling Strategy to Produce Hydrogen and Ethylene in a Membrane Reactor, Angew. Chem. Int. Ed. 49 (2010) [8] J. Kniep, Y.S. Lin, Partial Oxidation of Methane and Oxygen Permeation in SrCoFeOx Membrane Reactor with Different Catalysts, Ind. Eng. Chem. Res. 50 (2011) [9] Z.P. Shao, W.S. Yang, Y. Cong, H. Dong, J.H. Tong, G.X. Xiong, Investigation of the permeation behavior and stability of a Ba 0.5 Sr 0.5 Co 0.8 Fe 0.2 O 3-delta oxygen membrane, J. Membr. Sci. 172 (2000) [10] Y. Teraoka, H.M. Zhang, S. Furukawa, N. Yamazoe, OXYGEN PERMEATION THROUGH 24

26 PEROVSKITE-TYPE OXIDES, Chem. Lett. (1985) [11] J. Sunarso, S. Liu, Y.S. Lin, J.C. Diniz da Costa, High performance BaBiScCo hollow fibre membranes for oxygen transport, Energy Environ. Sci. 4 (2011) [12] Z.B. Zhang, Y.B. Chen, M.O. Tade, Y. Hao, S.M. Liu, Z.P. Shao, Tin-doped perovskite mixed conducting membrane for efficient air separation, J. Mater. Chem. A., 2 (2014) [13] Y.F. Cheng, H.L. Zhao, D.Q. Teng, F.S. Li, X.G. Lu, W.Z. Ding, Investigation of Ba fully occupied A-site BaCo 0.7 Fe 0.3 x Nb x O 3 δ perovskite stabilized by low concentration of Nb for oxygen permeation membrane, J. Membr. Sci. 322 (2008) [14] J.W. Zhu, G.P. Liu, Z.K. Liu, Z.Y. Chu, W.Q. Jin, N.P. Xu, Unprecedented Perovskite Oxyfluoride Membranes with High-Efficiency Oxygen Ion Transport Paths for Low-Temperature Oxygen Permeation, Adv. Mater., 28 (2016) [15] S. Engels, T. Markus, M. Modigell, L. Singheiser, Oxygen permeation and stability investigations on MIEC membrane materials under operating conditions for power plant processes, J. Membr. Sci. 370 (2011) [16] J. Gao, L.P. Li, Z. Yin, J.C. Zhang, S.M. Lu, X.Y. Tan, Poisoning effect of SO 2 on the oxygen permeation behavior of La 0.6 S r0.4 Co 0.2 Fe 0.8 O 3 δ perovskite hollow fiber membranes, J. Membr. Sci. 455 (2014) [17] Y. Alqaheem, A. Thursfield, G. Zhang, I.S. Metcalfe, The impact of sulfur contamination on the performance of La 0.6 Sr 0.4 Co 0.2 Fe 0.8 O 3 δ oxygen transport membranes, Solid State Ionics. 262 (2014) [18] E. Walker, S.C. Ammal, S. Suthirakun, F. Chen, G.A. Terejanu, A. Heyden, Mechanism of Sulfur Poisoning of Sr 2 Fe 1.5 Mo 0.5 O 6-δ Perovskite Anode under Solid Oxide Fuel Cell Conditions, J. Phy. Chem. 25

27 C. 118 (2014) [19] S.G. Li, W.Q. Jin, P. Huang, N.P. Xu, J. Shi, Y.S. Lin, Tubular lanthanum cobaltite perovskite type membrane for oxygen permeation, J. Membr. Sci. 166 (2000) [20] J. Xie, Y.-W. Ju, T. Ishihara, Influence of sulfur impurities on the stability of La 0.6 Sr 0.4 Co 0.2 Fe 0.8 O 3 cathode for solid oxide fuel cells, Solid State Ionics (2013) [21] H.H. Wang, C. Tablet, J. Caro, Oxygen production at low temperature using dense perovskite hollow fiber membranes, J. Membr. Sci. 322 (2008) [22] B. Meng, Z.G. Wang, Y.Y. Liu, X.Y. Tan, J.C.D. da Costa, S.M. Liu, Preparation and oxygen permeation properties of SrCo 0.9 Nb 0.1 O 3 δ hollow fibre membranes, Sep. Purif. Technol. 78 (2011) [23] A. Leo, S. Liu, J.C. Diniz da Costa, Production of pure oxygen from BSCF hollow fiber membranes using steam sweep, Sep. Purif. Technol. 78 (2011) [24] J.W. Zhu, G.R. Zhang, G.P. Liu, Z.K. Liu, W.Q. Jin, N.P. Xu, Perovskite Hollow Fibers with Precisely Controlled Cation Stoichiometry via One-Step Thermal Processing, Adv. Mater., 29 (2017). [25] S.J. Zhuang, Y. Li, M. Zuo, X.Y. Tan, B. Meng, N.T. Yang, S.M. Liu, Dense composite electrolyte hollow fibre membranes for high temperature CO2 separation, Sep. Purif. Technol., 132 (2014) [26] X.Y. Tan, Y.T. Liu, K. Li, Mixed conducting ceramic hollow-fiber membranes for air separation, AIChE J. 51 (2005) [27] S.M. Liu, G.R. Gavalas, Preparation of oxygen ion conducting ceramic hollow-fiber membranes, Ind. Eng. Chem. Res., 44 (2005) [28] A. Leo, S. Smart, S.M. Liu, J.C. Diniz da Costa, High performance perovskite hollow fibres for oxygen separation, J. Membr. Sci. 368 (2011)

28 [29] J.W. Zhu, S.B. Guo, Z.Y. Chu, W.Q. Jin, CO 2 -tolerant oxygen-permeable perovskite-type membranes with high permeability, J. Mater. Chem. A. 3 (2015) [30] S.B. Guo, J.W. Zhu, Z.K. Liu, X. Jiang, Z.C. Zhang, W.Q. Jin, Enhanced High Oxygen Permeation of Mixed-Conducting Multichannel Hollow Fiber Membrane via Surface Modified Porous Layer, Ind. Eng. Chem. Res. 54 (2015) [31] Z.K. Liu, G.R. Zhang, X.L. Dong, W. Jiang, W.Q. Jin, N.P. Xu, Fabrication of asymmetric tubular mixed-conducting dense membranes by a combined spin-spraying and co-sintering process, J. Membr. Sci. 415 (2012) [32] C. Zhang, X.F. Chang, X.L. Dong, W.Q. Jin, N.P. Xu, The oxidative stream reforming of methane to syngas in a thin tubular mixed-conducting membrane reactor, J. Membr. Sci. 320 (2008) [33] J.W. Zhu, Z.Y. Dong, Z.K. Liu, K. Zhang, G.R. Zhang, W.Q. Jin, Multichannel mixed-conducting hollow fiber membranes for oxygen separation, AIChE J. 60 (2014) [34] L. Ge, W. Zhou, R. Ran, S.M. Liu, Z.P. Shao, W.Q. Jin, N.P. Xu, Properties and performance of A-site deficient (Ba 0.5 Sr 0.5 ) 1 x Co 0.8 Fe 0.2 O 3 δ for oxygen permeating membrane, J. Membr. Sci. 306 (2007) [35] T. Yang, F. Li, D.G. Xia, Au@BICUVOX10 composite cathode for novel structure low-temperature solid-oxide fuel cells, J. Power Sources. 195 (2010) [36] G.R. Zhang, Z.K. Liu, N. Zhu, W. Jiang, X.L. Dong, W.Q. Jin, A novel Nb 2 O 5 -doped SrCo 0.8 Fe 0.2 O 3-delta oxide with high permeability and stability for oxygen separation, J. Membr. Sci. 405 (2012) [37] B.A. Vanhassel, J.E. Tenelshof, H.J.M. OXYGEN PERMEATION FLUX THROUGH LA1-YSRYFEO3 LIMITED BY CARBON-MONOXIDE OXIDATION RATE, Appl. Catal. A: Gen

29 (1994) [38] B.C.H. Steele, OXYGEN ION CONDUCTORS AND THEIR TECHNOLOGICAL APPLICATIONS, Mat. Sci. Eng. B-Solid. 13 (1992) [39] P. V. Hendriksen, P. H. Larsen, M. Mogensen, F. W. Poulsen, K. Wiik, Prospects and problems of dense oxygen permeable membranes, Catal. Today, 56 (2000) [40] Y. Liu, X.F. Zhu, M.R. Li, R.P. O'Hayre, W.S. Yang, Nanoparticles at Grain Boundaries Inhibit the Phase Transformation of Perovskite Membrane, Nano Lett. 15 (2015) [41] H.H. Wang, C. Tablet, A. Feldhoff, H. Caro, Investigation of phase structure, sintering, and permeability of perovskite-type Ba 0.5 Sr 0.5 Co 0.8 Fe 0.2 O 3-delta membranes, J. Membr. Sci. 262 (2005)

30 29

31 Highlights 1. The effects of different polymer binders on membrane properties were investigated. 2. Sulfur poisoning was confirmed taking place during sintering and operation stage. 3. Sulfate impurities were distributed on the surfaces and in the bulk of membrane. 4. Sulfate has a negative effect on oxygen permeation flux and long-term stability. 30