Fibers and Polymers 2016, Vol.17, No.10, 1558-1568 DOI 10.1007/s12221-016-6544-2

ISSN 1229-9197 (print version) ISSN 1875-0052 (electronic version)

Preparation and Characterization of Inner-pressure Hollow Fiber Composite Membrane via Two-way Coating Technique for CO2/CH4 Separation Hongbin Li1*, Wenying Shi1, Yuheng Su1, Qiyun Du2, Hongying Zhu1, and Xiaohong Qin1,3 1

School of Textiles Engineering, Henan Engineering Laboratory of New Textiles Development, Henan Institute of Engineering, Zhengzhou 450007, P. R. China 2 State Key Laboratory of Separation Membranes and Membrane Processes, Tianjin Polytechnic University, Tianjin 300387, P. R. China 3 School of Textiles Science, Donghua University, Shanghai 201620, P. R. China (Received May 18, 2016; Revised July 29, 2016; Accepted August 8, 2016) Abstract: High-selectivity inner-pressure hollow fiber composite (HFC) membrane for CO /CH separation was prepared through the Two-way coating (TWC) technique. The blends of poly(vinylamine) (PVAm)/polyvinyl alcohol (PVA) were coated onto porous hollow fiber polysulfone (PSF) ultrafiltration (UF) membrane with an effective membrane area of 0.4 m . The effects of fabrication parameters on the permselectivity of the resultant HFC membrane were investigated and the optimum preparation conditions were obtained as follows: coating time for 30 min and air blowing time for 30 min after the coating. The prepared HFC membrane showed the typical characteristic of fixed carrier membrane with a high selectivity of CO and CH : the separation factor of CO /CH (40 vol% CO at 25 C and 0.2 MPa) was 36.6 and the CO permeability was 56.3 GPU. Field emission scanning electron microscopy (FESEM) images indicated that the HFC membrane prepared by TWC technique had a uniform coating layer along the whole hollow fiber. Membrane permselectivity showed almost no difference between different membrane sections. The HFC membrane showed a good stability during the continuous testing process of 540 h. And the HFC membrane preserved at 30 C and 40 % humidity exhibited a good durability with a basically unchanged separation factor after 30 days. 2

4

2

o

2

4

2

4

2

2

o

Keywords: Poly(vinylamine), Hollow fiber, Composite membrane, Two-way coating, CO /CH separation 2

4

[12,13]. Among them, PVAm has a high density of primary amino groups for CO2 facilitated transport carriers where the reversible reactions of CO2 with amino carriers could well facilitate the CO2 transport and thus result in a high CO2 permselectivity. Besides, the coating liquid was PVAm aqueous solution which was environment-friendly. Many researchers have successfully employed PVAm as the separation layer for the removal of CO2 [5,9,12]. However, the PVAm active layer has high crystallinity which results in its brittle properties especially at high feed pressure and the reduction of the membrane’s permselectivity [5]. Blending is believed to be an effective method to decrease the PVAm crystallinity. Polyvinyl alcohol (PVA) was often selected as the modifier due to its excellent film-forming properties, flexibility and tensile strength [13,14]. The incorporation of PVA could enhance the PVAm polymeric network with good membrane-forming properties and mechanical stability through the entanglement of the PVA chains with the PVAm chains. The previous study has reported that the PVAm/PVA blend composite membrane could act as a CO2 facilitated transport membrane in the presence of water and show excellent permeation performance of CO2/CH4 mixture [15]. However, most of the researches are centered on the improvement of separation performance through the studies of PVAm concentration, post-treatment (including drying process and silicone rubber cast to avoid possible surface defects) [8], operation conditions such as feed pressure, temperature and humidity and support properties improvement [5,15].

Introduction Removal of CO2 from energy gas (e.g. syngas and natural gas) and flue gas (e.g. landfill gas) are becoming more important for clean energy supply and environmental remediation [1,2]. Compared with conventional gas separation methods, membrane separation technology as a new type process for gas separation has attracted more and more attention due to its large interfacial area, cost saving, and superior durability [3,4]. Most of the research is focused on the preparation of composite membranes. Recently, many researchers have studied new active materials to prepare membranes with both high permeability and selectivity of CO2 such as polyvinylamine/polyethlene glycol (PVAm) [5], the hydrolysate of polyvinylpyrrolidone (PVP) [4], poly (N,N-dimethylaminoethyl methacrylate)-poly (ethylene glycol methyl ethermethacrylate) (PDMAEMA-PEGMEMA) [6], polyaniline (PANI) [7] and pentaerythrityl tetraethylenediamine (PETEDA) [8], etc. Fixed carrier membranes containing amino groups show high permselectivity as well as good stability. The strong affinity between amino groups and CO2 molecules leads to the high CO2 permeability and CO2/gas selectivity of membranes. Novel polymeric membrane materials containing amino groups for the preparation of fixed carrier membranes has been explored such as polyvinylamine (PVAm) [5,9], polyallylamine [10], polyethylenimine (PEI) [11], polyamide *Corresponding author: [email protected] 1558

Hollow Fiber Gas Separation Membrane

Few studies about coating technique are reported and generally flat sheet substrates were used as support membranes. Hollow fiber with high-packing density and self-support is appropriate to be used as the substrates in the preparation of composite membrane. The coating method is an import factor to prepare high-performance composite membrane especially for the hollow fiber membrane with long fiber length. Just as in practical application, the concentration polarization and membrane fouling along the longer hollow fiber is more serious. The conventional coating process is immersion coating which is complex and needs many man-made operations. During module fabrication process of outer-skinned HFC membranes, external functional layers are vulnerable to stick and rub against each other especially for the coating of water soluble substance such as PVAm, PVA, etc. The coating layers would be ruptured and even stripped from the outer surface of the substrate. Therefore, after the completion of coating, another coating layer of silicone rubber is often required which would further reduce membrane gas permeation rate. Moreover, performance differences between different sections in a long fiber filament are larger. Thus, the studies on the preparation of outer-skinned HFC membranes usually used fiber membrane with very small effective membrane area. In addition, a uniform liquid film after coating is difficult to obtain followed by the drying process which results in an even thickness of the active layer. In our previous report [16], a novel coating method-twoway coating (TWC), has been successfully explored and used in the interfacial polymerization process of highperformance hollow fiber polyamide composite nanofiltration (NF) membrane with the polyamide layer formed on the fiber lumen side. The inlet and outlet valves can be switched periodically so that the flow directions of coating solution are regularly changed. And thus a homogeneous and uniform active layer is formed on fiber inner surface. Inner-pressurized hollow fiber composite membrane with different fiber length and membrane area can be obtained via TWC technique. In this study, HFC membrane used for CO2/CH4 separation was prepared by TWC technique with the active layer of the blends of commercial poly(vinylamine) (PVAm)/polyvinyl alcohol (PVA) on the fiber lumen side. The effects of fabrication conditions on HFC membrane separation performance of CO2/CH4 were systematically investigated. The structure of active layer was observed through field emission scanning electron microscopy (FESEM). The performance stability during the continuous testing and the durability of HFC membrane preserved at 30 oC and 40 % humidity were also studied.

Fibers and Polymers 2016, Vol.17, No.10

vacuum oven at 80 oC for 12 h before use. Polyvinylamine (PVAm, MW 340,000) was purchased from BASF. The raw PVAm was re-precipitated in acetone/ethanol mixtures and then dissolved in distilled water. PVAm aqueous solution with high purity was obtained by ion exchange through the immersion of strongly basic anion exchange resin. Polyvinyl alcohol (PVA, MW 72,000), dimethylacetamide (DMAc) and polyethylene glycol (PEG) (Mw 400) were all supplied from Tianjin Kemiou reagent Co., Ltd. (China). The mixed gas (CO2: CH4=40: 60 vol%) used for permeation test was purchased from Henan Yuanzheng Hi Tech Co. (China). Preparation of Hollow Fiber PSF Support Membrane Hollow fiber PSF support membrane was prepared by drywet phase inversion method. The spinning dope was prepared by dissolving PSf in the DMAc at 80 oC for about 8 h to form homogeneous solution. After degassing under vacuum at 30 oC for 24 h, the solution was extruded through a tube-in-orifice spinneret with inside/outside diameters of 0.8/1.3 mm. Meanwhile, the core liquid with different content of DMAc aqueous solution was extruded from the inner tube of the spinneret. Table 1 listed the spinning parameters in this study. The as-spun hollow fiber membranes were immersed in de-ionized water and washed for several times to remove the residual solvent. Fabrication of Hollow Fiber PVAm/PVA Composite Membrane The coating solution of PVAm/PVA blends was obtained by adding PVA aqueous solution into a PVAm aqueous solution of the same concentration. After ultrasonic vibration for 5 min and stirring the mixture overnight, the coating solution was obtained for use. The concentration of PVAm/ PVA blend solution was fixed at 2 wt% with the blend ratio of 1:1. The main components of TWC equipment were illustrated in Figure 1 and the specific procedure for the preparation of HFC membranes was as follows. Table 1. Spinning conditions of hollow fiber PSF support membrane Spinning condition Composition of spinning dope (wt%) (PSF/DMAc/PEG400) Spinning pot temperature ( C) Metering pump temperature ( C) Metering pump rotation frequency (Hz) Air gap (cm) Bore liquid temperature ( C) Bore liquid flow rate (ml·min ) Coagulation temperature ( C) Fiber take-up frequency (Hz) Fiber take-up temperature ( C) o

o

o

-1

Experimental Materials Polysulfone (PSF) (Solvay, Udel-3500) was dried in a

1559

o

o

Value 19/76/5 30 30 22 3 30 10 30 18 30

1560

Fibers and Polymers 2016, Vol.17, No.10

Hongbin Li et al.

Hitachi S-4800, Japan) images with an average value of five points. The pure water flux of five membrane modules was tested in a cross-flow filtration cell at 25 oC. Each membrane module was composed of 150 fibers with a length of 125 cm and its effective membrane area was 0.4 m2. Membranes were initially pre-compacted at 0.20 MPa to get a steady flux. Then, water flux was measured at 0.10 MPa and calculated by the following equation. V F = ----At

Figure 1. The preparation flow diagram of HFC membrane via TWC technique.

Membrane Module Fabrication The effective membrane area for each module was about 0.4 m2. Each module was composed of 150 fibers with a length of 125 cm and tied into one bundle. Both ends of the bundle were sealed with epoxy resin. HFC Membrane Preparation Hollow fiber membrane modules were installed in TWC system and the excessive liquid in fiber lumen was blown out with compressed air. Valves 1 and 4 were switched on and meantime valves 2 and 3 were closed. The coating solution was circulated for a certain time. Afterwards, solution flow direction was changed by opening valves 2, 3 and closing valves 1, 4 to. Subsequently, turn off the pump and unscrew valves 7, 8. Excess coating solution could be drained from the fiber lumen by gravity. Then, compressed air blew through the fiber inner surface via turning on valves 5 and 8 for a certain time, followed by changing the blowing direction through opening the valves 6 and 7. In this context, the coating of PSF hollow fiber is completed from two directions. Similarly, the blowing procedure is also carried out from two directions. Post-treatment Membrane modules were then preserved at 30 oC and 40 % humidity in a climatic chamber (H/HWHS -800L, Shanghai Husheng Laboratory Instrument Factory, China) at least 24 h before performance testing. The prepared HFC membranes were divided into five sections with an average length of 200 mm. These five sections were designated as S1, S2, S3, S4 and S5 from the top to the bottom of the filament, successively. Characterization of PSF Hollow Fiber Substrate Membrane Membrane inner/outer diameters were measured through the Field emission scanning electron microscopy (FESEM,

(1)

F-water permeate flux (l·m-2·h-1), V-water permeation volume (L), A-the effective membrane area (m2), t-filtration time (h). The PEG-20,000 concentrations in the feed and the permeate solutions were measured by a UV-vis spectrophotometer (TU-1901, Purkinje General Instrument Co. Ltd., China) at a wavelength of 510 nm, respectively. The rejection (R) is calculated as: C R = 1 – ------p Cf

(2)

where Cp and Cf were PEG concentrations in the permeate and feed solutions, respectively. The porosity of PSF hollow fiber substrate membrane (ε) was evaluated via the ratio of pore volume to membrane geometrical volume as described in equation (3) [17,18]. 10 cm length of PSF hollow fiber samples was immersed in pure water for at least 12 h. Membrane samples were taken out and wiped with filter papers in order to drain off the excess water absorbed on fiber surface. Then, weigh the wet membranes weight (Ww). Afterwards, membrane samples were dried until a constant mass (Wd) was reached. ( Ww – W d ) - × 100% ε = ----------------------Alρ

(3)

where Ww and Wd were the wet and dry membranes weights (g), respectively. A, l, and ρ were the effective area (cm2), average thickness (cm) and pure water density at atmosphere temperature (g·cm-3), successively. The tensile strength and elongation-at-break of hollow fiber membranes were examined using a Hounsfield tensile tester (LLY-06, Laizhou Electron Instrument Co., Ltd., China) at a strain rate of 2 mm·min-1 at room temperature. The tested membrane samples were 10 cm length. At least five values were measured to get a reliable value. Membrane Morphology Field emission scanning electron microscopy (FESEM, Hitachi S-4800, Japan) was used to observed cross-sectional morphology of hollow fiber membranes. Membrane specimens were frozen in liquid nitrogen, fractured and then sputtered with gold.

Hollow Fiber Gas Separation Membrane

Fibers and Polymers 2016, Vol.17, No.10

permeance ratio as equation (2). JCO αCO2/CH4 = ---------2 JCH4

1561

(5)

Results and Discussion Characterization of PSF Hollow Fiber Substrate The Structural Parameter of PSF Hollow Fiber Membrane The prepared PSF hollow fiber substrate membrane was dried at 30 oC and 40 % humidity in a climatic chamber for 48 h. The permselectivity of an ideal composite membrane is mainly dominated by the top active layer on the porous support layer [19]. Additionally, the properties of substrate membrane had an important effect on the permselectivity of the composite membrane [20]. The perfect substrate membranes have appropriate pore size and pore shape with short straight flow passage which can reduce the diffusion resistance of the fast gas as well offer the good mechanical support for the coating layer. Table 2 listed the properties of the resultant PSF hollow fiber substrate membrane. It could be seen that PSF substrate membrane prepared from the drywet spinning exhibited an excellent properties which had a high porosity and tensile strength. The Morphologies of PSF Hollow Fiber Membrane Figure 3 showed the cross section morphologies of PSF hollow fiber membrane. It could be clearly seen that PSF substrate exhibited a typical asymmetric structure which was composed of a support layer with double finger-like pores and a denser skin layer with sponge-like pores. Previous reports have verified membranes with this structure would have a large bursting pressure, a high water flux and tensile strength (as listed in Table 2) which are particularly appropriate to be used as the substrates of HFC membranes [21,22]. The Permselectivity of PSF Hollow Fiber Membrane Figure 4 showed the permselectivity of PSF hollow fiber substrate membrane at different feed pressure. It could be seen from Figure 4 that the permeance of CO2 and CH4 increased as the feed pressure varied from 0.1 to 0.6 MPa. The selectivity of CO2/CH4 kept at around 0.68 which suggested that Knudsen flow (with a selectivity of CO2/CH4 about 0.60) dominates the gas transport process in the porous PSF substrate prepared in this study [23]. The mean effective pore size of porous membrane was obtained according to the traditional solute transport approach by ignoring the influence of the steric and hydrodynamic interaction between solute and membrane pores [24]. The mean pore size of prepared PSF substrate membrane was

Figure 2. Schematic diagram of the experimental setup used for measuring gas permeation performance of membranes. (1) feed gas; (2), (6), (8), (11), (15) valves; (3), (12) rotameter; (4), (13) pressure gauge; (5) HFC membrane module and thermostatic container; (7) soap-film flow meter; (9) gas chromatograph; (10) sweep gas (14) humidifier.

Gas Permeation Experiment HFC membrane permselectivity was measured using CO2/CH4 (40/60, vol%) mixed gas as the feed gas. The schematic diagram of the membrane permselectivity measurement was illustrated in Figure 2. H2 was used as the sweep gas on the permeate side under the atmospheric pressure. The gas permeation of all membranes was tested at the same feed pressure of 0.2 MPa and feed temperature of 30 oC except for otherwise indicated. The sweep gas flow was measured via a soap bubble meter and the composition of permeate gas was analyzed through a gas chromatograph (7820A, Agilent, USA). The permeances (Ji) of CO2 and CH4 were obtained as equation (1). The permeate flux (Ni) of CO2 and CH4 was calculated via the measurement of flow rate and composition of outlet sweep gas. N Ji = -------iΔPi

(4)

where i, Ni and ΔPi were the gas species, permeate flux and partial pressure difference between the membrane upstream and downstream sides. The permeance (Ji) unit was described using GPU. It equaled to 10-6 cm3 (STP)/(cm2 . . s cmHg). The CO2/CH4 selectivity was designated as separation factor (αCO2/CH4) and calculated through their Table 2. Properties of the resultant PSF hollow fiber substrate membrane Inner/outer diameters (mm/mm) PSF substrate membrane 0.76/1.32 Properties

PEG-20,000 Rejection (%) 98.2

Pure water flux (l·m ·h ) 162 -2

-1

Porosity (%) 80.2

Tensile strength Elongation-at-break (MPa) (%) 7 150

1562

Fibers and Polymers 2016, Vol.17, No.10

Hongbin Li et al.

Figure 3. Cross section morphologies of PSF membrane (a1) ×65 and (a2) ×250.

Figure 5. Variations of the permselectivity of the HFC membrane with different coating time.

Figure 4. Permselectivity of the dry PSF hollow fiber substrate membrane at different feed pressure.

calculated to be about 8.3 nm which was in the range of Knudsen diffusion through pores (1-100 nm). Effect of Coating Conditions on the Permselectivity of HFC Membranes Coating Time The effects of coating times on the permselectivity of the resultant HFC membranes were investigated and the results were shown in Figure 5. It could be seen from Figure 5 that

the permeance of the HFC membranes after being coated decreased significantly compared with the PSF substrate membrane. With the coating time increasing from 0 to 30 min, the separation factor (αCO2/CH4) increased obviously. Afterwards, the value of αCO2/CH4 kept almost unchanged. On the contrary, the permeances (JCO2 and JCH4) showed the opposite variation trends. During the TWC process, the coating solution would penetrate into the pores in the PSF substrate and the thin selective layer would formed and gradually thicken on the PSF inner surface. Consequently, the selectivity of HFC membrane gradually enhanced and the permeances of both JCO2 and JCH4 decreased when the coating time was lower than 30 min. With the extension of coating time to 50 min, the thickness increasing of coating layer resulted in the continuous decline of the permeances and the stability of the selectivity. Considering that the separation factor reached the maximum, the coating time of 30 min was selected. Air Blowing Time in Fiber Lumen after PVAm/PVA Solution Coating The effects of air blowing time after PVAm/PVA solution coating on membrane separation performane were shown in

Hollow Fiber Gas Separation Membrane

Fibers and Polymers 2016, Vol.17, No.10

1563

Figure 6. It should be noted that the blowing time here is a total value in two coating directions. It could be seen from Figure 6(a) that with the increasing of air blowing time, the separation factor of CO2/CH4 initially increased and afterwards kept almost constant. The variations of the permeance of CO2 and CH4 exhibited a similar trend as shown in Figure 6(b). When the air blowing time was shorter than 30 min, the fiber lumen saturated with the viscous coating solution could not be completely opened. Consequently, a thick and uneven coating layer was formed on the inner surface of hollow fiber which resulted in a poor permselectivity. The further experiment indicated that HFC membrane prepared through a short air blowing time (less than 30 min) after PVAm/PVA solution coating had a extremely bad separation performance under higher testing pressure (>0.5 MPa) (the results were not listed here). This was ascribed to the formation of an uneven coating layer even if the active later was relatively thick. As shown in Figure 6, when the air blowing time increased from 10 to 30 min, the increasing rate of CO2 permeance was higher than that of CH4 which indicated that

more CO2 molecules penetrated through the gradually thinning selective layer. The thinner selective layer would lead to the more sensitive CO2 transport competition in fixed carrier membranes and thus the membranes would exhibit a higher CO2 permeance [25]. Deng [15] pointed that the facilitated effect will be sensitively decreased with increasing membrane thickness and the facilitated transport effect for CO2 is more obvious for the thinner membranes. This meant that more CO2 molecules would permeate through the thinner selective layer due to the enhanced facilitated effect. When the air blowing time reached 30 min, hollow fiber lumen gradually opened up with a constant thickness of the coating layer. With the expansion of air blowing time over 30 min, the air flow over the coating layer surface only played a role in the blowing away the water molecules of the PVAm/PVA aqueous solution. Too long blowing time can cause the rapid drying of coating layer. This resulted in a high crystallinity of the PVAm/PVA blend polymers and caused the permselectivity deterioration of HFC membranes [25]. Effect of Pressure Difference Figure 7(a) and (b) showed the separation performance of the HFC membrane at different feed pressure. The HFC

Figure 6. The variations of HFC membrane separation performance with different air blowing time after PVAm/PVA solution coating.

Figure 7. Permselectivity of HFC membrane at different feed pressure.

1564

Fibers and Polymers 2016, Vol.17, No.10

Figure 8. The mechanism of facilitated transport process through a fixed carrier HFC membrane.

Hongbin Li et al.

membrane used here and in the following text was prepared in conditions: 30 min of coating time, 30 min of air blowing time in fiber lumen after PVAm/PVA solution coating. The feed gas was saturated with water vapor in the humidifier before entering into the HFC membrane module. As can be seen in Figure 7(a), the CO2/CH4 selectivity of the HFC membrane decreased with the feed pressure increasing from 0.15 to 0.5 MPa. Afterwards, the decline trend of the separation factor gradually slowed down. It could be obtained from previous reports [5,25] that the permeation mechanism of CH4 through the membrane is solutiondiffusion, while CO2 permeates the membrane mainly by a facilitated transport mechanism. Generally, the characteristic permeance of facilitatedtransport membrane would decrease initially with the increasing of feed partial pressure due to the saturation of carriers. When the carriers are saturated, the permeance will

Figure 9. Cross section morphologies of different fiber sections of HFC membranes (b1-f1) ×10.0k and (b2-f2) ×50.0k.

Hollow Fiber Gas Separation Membrane

Fibers and Polymers 2016, Vol.17, No.10

1565

Figure 9. Continued.

stabilize and the solution-diffusion mechanism will take over. Figure 8 illustrated the mechanism of CO2 transport through the fixed carrier membrane. CO2 molecules transport through the water-swollen fixed carrier membrane in the form of HCO3- and diffuse as ions in liquid. It can be seen in Figure 7(b) that at low feed pressure (0.15-0.5 MPa), the CO2 permeance decreased rapidly with increasing feed pressure whereas the CH4 permeance was nearly constant. When the feed pressure was between 0.5 and 1.5 MPa, CO2 permeance kept nearly constant whereas CH4 permeance had an obvious increase. This was due to the high pressure broke the increase of gas diffusion resistance derived from the packed and less water-swollen membrane. Consequently, the CH4 permeance increased when the feed pressure was high. The obvious decrease of CO2 permeance in the beginning and the apparent increase of CH4 permeance at higher feed pressure resulted in the continuous decline of separation factor of CH4/CO2 as shown in Figure 7(a). Morphologies of Different Composite Membrane Sections In order to investigate the effect of TWC process on fiber membrane structure and permeation performance, HFC membranes prepared in this study were divided into five sections with an average length of 22 cm. Then, the five sections were made into five membrane module according to the method on Section Membrane module fabrication. Five membrane modules were designated as S-1, S-2, S-3, S-4 and S-5 from the upper end to the bottom of HFC membrane, respectively.

Figure 9 showed the cross section morphologies of different fiber sections (S1-S5) It could be obviously seen that a thin coating layer with a thickness of about 160180 nm was formed on the inner surface of PSF fiber. The layer was smooth and tightly attached on the sponge-like pores of PSF support membrane surface. Coating layer thickness of different fiber sections were characterized by FESEM analysis and the specific values of S1, S2, S3, S4 and S5 were 189.3, 163.4, 154.7, 158.0 and 173.0 nm, respectively. The data indicated that a relatively uniform active layer from the upper end to the bottom of the hollow fiber prepared by TWC technique was formed and this contributed to the excellent permselectivity of HFC membrane. Comparison of Separation Performance of Different Composite Membrane Sections Table 3 showed the separation performance of different membrane sections. It could be seen that the permeance of CO2 and separation factor of CH4/CO2 showed almost no difference between different membrane sections with only a small fluctuation of ±1.6 % and ±1 %, respectively. The Table 3. Comparison of separation performance of different membrane sections J

Section (GPU) α

CO2

CO2/CH4

S-1 54.3 35.7

S-2 54. 7 36.3

S-3 55.1 36.6

S-4 54.6 36

S-5 54 35.5

1566

Fibers and Polymers 2016, Vol.17, No.10

Hongbin Li et al.

similar separation performance of different membrane sections was mainly due to their even coating layers as shown in FESEM images. This result indicated that TWC was an effective method to prepare HFC membrane with a uniform separation performance along the filament.

The Durability of HFC Membrane The prepared HFC membrane was placed at 30 oC and 40 % humidity in a climatic chamber. Data of separator performance were obtained every 24 h (i.e. 1 day) to characterize the membrane durability. To make the comparison

more straightforward, data of membrane permeance and selectivity after the preservation at constant temperature and humidity in this study were normalized. Figure 11 showed the variations of HFC membrane permselectivity with preservation time. It could be seen from Figure 11(a) that the separator factor of CO2/CH4 showed slight decrease with only 3.6 % reduction after 30 days' preservation. And the permeance as shown in Figure 11(b) exhibited a slow decline for JCO2 and a slight increase for JCH4. These results suggested that the HFC membrane preserved at 30 oC and 40 % humidity would exhibit a good durability with a basically unchanged separation performance after 30 days. The slight deterioration of membrane separation performance was due to the gradual crystallization of PVAm during the long time preservation [25]. The increase of crystallinity can induce the reduction of gas permeance which has been clearly elaborated in the previous literature [5,25]. Crystalline region can act as an impermeable barrier to gas molecules and prolong the penetrating path in crystalline-amorphous interface than in the amorphous region, thus decreasing the diffusion coefficient [26]. Moreover, previous literature has pointed that membranes made from poly(vinylamine) (PVAm) have high crystallinity tendency due to the strong

Figure 10. Stability test of HFC membrane (a) CH /CO separation factor; (b) CH and CO permeace.

Figure 11. The variations of HFC membrane permselectivity with preservation time.

Performance Stability of HFC Membrane The separation performance was continuously tested to characterize the performance stability of HFC membrane. Figure 10 showed the variations of separation performance of HFC membrane with testing time. It could be seen that with the testing time increasing over 540 min, the CH4/CO2 separation factor, CH4 and CO2 permeace kept almost constant. This result suggested that the separation performance of HFC membrane could be well maintained after continuous testing which was due to the stable morphology of PVAm/ PVA coating layer.

4

4

2

2

Hollow Fiber Gas Separation Membrane

Fibers and Polymers 2016, Vol.17, No.10

1567

Table 4. Permselectivity of other PVAm gas membrane for CO /CH separation 2

Code a b c d e f g h i

Membrane material

J

CO2

PVAm/PEG/PES CNT/PVAm/PVA/PSF PVAm/PVA/PSF PVAm/PVA/PSF Crosslinked PVAm/PSF PVAm/PSF PVAm/PSF PVAm/PVA/PSF PVAm/PVA/PSF

(GPU)

5.8 12.8 201 201 2.3 4 15.9 35.6 55

α

4

CO /CH (vol%) Pure gas 10/90 10/90 35/65 Pure gas Pure gas 50/50 10/90 40/60 2

CO2/CH4

62 45 45 40 26.9 206 31.4 17 36

4

Testing pressure (bar) 12.7 2 2 2 2 1.3 1.3 2 2

Module type 1

TFC TFC TFC TFC TFC HFC HFC HFC HFC

Ref. [5] [27] [25] [28] [29] [30] [30] [13,22] This work

permselectivity of HFC membrane prepared in this work shows a better comprehensive performance with a moderate CO2 permeation and CO2/CH4 separator factor. It should be pointed that the HFC membranes prepared through TWC technique in this study has much larger effective membrane area and uniform performance than other HFC membranes reported in previous literature in Table 4.

Conclusion

Figure 12. Comparison permselectivity.

of

different

PVAm

membrane

intermolecular interaction from the hydrogen atoms [3,7]. The high crystallinity not only deteriorates membrane’s permselectivity but also makes the membranes brittle. Indeed, we cast the PVAm aqueous solution onto clean glass plate. And put it under the atmosphere circumstance for 30 days, it could be found that the thin PVAm film was brittle and some cracks emerged locally on the surface of thin PVAm film. This is the reflection of gradual crystallization of PVAm. Table 4 compared the permselectivity of the HFC membrane prepared in this work and other PVAm gas membranes for CO2/CH4 separation reported in previous literatures [5,13,22,25,27-30]. It could be seen more researches focused on the study of thin-film composite (TFC) PVAm membrane. Thus, the research on HFC membranes for CO2/ CH4 separation is particularly important. However, as listed in Table 4, HFC membranes exhibited poorer permselectivity than that of TFC membranes which is an urgent problem to be solved. In addition, Comparison of different PVAm membrane permselectivity was shown in Figure 12. The

High-selectivity hollow fiber composite (HFC) membrane was prepared by the Two-way coating (TWC) of poly (vinylamine) (PVAm)/polyvinyl alcohol (PVA) onto porous hollow fiber polysulfone (PSF) ultrafiltration (UF) membrane. It was found that the coating time and air blowing time after the aqueous coating had an important effect on the permselectivity of the HFC membrane. The prepared HFC membrane showed the typical characteristic of fixed carrier membrane. The comprehensive performance of HFC membrane was obtained when the feed pressure was 0.2 MPa. Membrane permselectivity showed almost no difference between different membrane sections. In addition, the continuous testing process of 540 h and the long time preservation at 30 oC and 40 % humidity indicated that HFC membrane had a good stability and durability. These results were mainly due to the smooth and even coating layer formed on the inner surface of hollow fiber through the TWC method. Based on the above results, TWC technique was proven to be a promising method for preparing high-selectivity innerpressure hollow fiber composite membrane for the CO2/CH4 separation.

Acknowledgment The authors gratefully acknowledge the funding for the Project supported by the National Natural Science Foundation of China (No. 51403052), the Science and Technology Development Project of Zhengzhou (No. 153PKJGG133), the Doctoral Fund of Henan Institute of Engineering (No.

1568

Fibers and Polymers 2016, Vol.17, No.10

D2015027), the open fund of Henan Engineering Laboratory of New Textiles Development (No. GCSYS201603) and the Program for Innovative Research Team (in Science and Technology) in University of Henan Province (No. 15IRTSTHN011).

References 1. S. C. Li, Z. Wang, C. X. Zhang, M. M. Wang, F. Yuan, J. X. Wang, and S. C. Wang, J. Membr. Sci., 436, 121 (2013). 2. Y. W. Gong, Z. Wang, and S. C. Wang, Chem. Eng. Process., 45, 652 (2006). 3. X. W. Yu, Z. Wang, Z. H. Wei, S. J. Yuan, J. Zhao, J. X. Wang, and S. C. Wang, J. Membr. Sci., 362, 265 (2010). 4. Y. Zhang, Z. Wang, and S. C. Wang, Desalination, 145, 385 (2002). 5. C. H. Yi, Z. Wang, M. Li, J. X. Wang, and S. C. Wang, Desalination, 193, 90 (2006). 6. P. F. Ji, Y. M. Cao, X. M. Jie, and Q. Yuan, J. Nat. Gas Chem. 19, 560 (2010). 7. J. Zhao, Z. Wang, J. X. Wang, and S. C. Wang, J. Membr. Sci., 403-404, 203 (2012). 8. Z. Wang, M. Li, Y. Cai, J. X. Wang, and S. C. Wang, J. Membr. Sci., 290, 250 (2007). 9. M. Sandru, T. J. Kim, and M. B. Hägg, Desalination, 240, 298 (2009). 10. Y. Cai, Z. Wang, C. H. Yi, Y. H. Bai, J. X. Wang, and S. C. Wang, J. Membr. Sci., 310, 184 (2008). 11. H. Matsuyama, A. Terada, T. Nakagawara, Y. Kitamura, and M. Teramoto, J. Membr. Sci., 163, 221 (1999). 12. T. J. Kim, H. Vrålstad, M. Sandru, and M. B. Hägg, J. Membr. Sci., 428, 218 (2013). 13. M. W. Uddin and M. B. Hägg, J. Membr. Sci., 423-424, 150 (2012).

Hongbin Li et al.

14. M. W. Uddin and M. B. Hägg, J. Membr. Sci., 423-424, 143 (2012). 15. L. Y. Deng, T. J. Kim, and M. B. Hägg, J. Membr. Sci., 340, 154 (2009). 16. H. B. Li, W. Wang, and Y. F. Zhang, J. Appl. Polym. Sci., 131, 12636 (2014). 17. W. T. Yan, Z. Wang, J. H. Wu, S. Zhao, J. X. Wang, and S. C. Wang, J. Membr. Sci., 498, 227 (2016). 18. B. K.Thakur and S. De, Sep. Purif. Technol., 93, 6 (2012). 19. U. Beuscher and C. H. Gooding, J. Membr. Sci., 132, 213 (1997). 20. L. Zhao, Y. X. Chen, B. Wang, C. H. Sun, S. Chakraborty, K. Ramasubramanian, P. Dutta, and W. S. Winston Ho, J. Membr. Sci., 498, 1 (2016). 21. R. H. Du, X. S. Feng, and A. Chakma, J. Membr. Sci., 279, 76 (2006). 22. L. L. Zhang, G. H. He, W. Zhao, F. Nie, X. C. Li, and M. Tan, J. Membr. Sci., 371, 141 (2011). 23. J. C. Poshusta, R. D. Noble, and J. L. Falconer, J. Membr. Sci., 160, 115 (1999). 24. X. Li and T. S. Chung, Appl. Energ., 114, 600 (2014). 25. L. Y. Deng and M. B. Hägg, J. Membr. Sci., 363, 295 (2010). 26. V. Compan, L. F. D. Castillo, S. I. Heranandez, M. Marlopez-Gonzalez, and E. Riande, J. Polym. Sci. B: Polym. Phys., 48, 634 (2010). 27. L. Y. Deng and M. B. Hägg, Int. J. Greenh. Gas. Con., 26, 127 (2014). 28. L. Deng and M. B. Hägg, Int. J. Greenh. Gas. Con., 4, 638 (2010). 30. C. M. Dong, Z. Wang, C. H. Yi, and S. C. Wang, J. Appl. Polym. Sci., 101, 1885 (2006). 29. T. J. Kim, B. A. Li, and M. B. Hägg, J. Polym. Sci. Pt. BPolym. Phys., 42, 4326 (2004).