Journal of Membrane Science 407–408 (2012) 164–175 Contents lists available at SciVerse ScienceDirect Journal of Membrane Science journal homepage: www.elsevier.com/locate/memsci CF4 plasma surface modification of asymmetric hydrophilic polyethersulfone membranes for direct contact membrane distillation Xing Wei a,b , Baolong Zhao a , Xue-Mei Li a,∗ , Zhouwei Wang a , Ben-Qiao He c , Tao He a,b,∗ , Biao Jiang a a b c Shanghai Advanced Research Institute, CAS, Shanghai 201203, China Nanjing University of Technology, Nanjing 210009, China Tianjin University of Technology, Tianjin 300160, China a r t i c l e i n f o Article history: Received 22 December 2011 Received in revised form 5 March 2012 Accepted 12 March 2012 Available online 20 March 2012 Keywords: Polyethersulfone Membrane distillation CF4 plasma Surface modification a b s t r a c t This paper describes the use of CF4 plasma modification of a hydrophilic membrane into a hydrophobic one for membrane distillation. Plasma surface modification conditions were optimized with respect to plasma glow discharge power and treatment duration using a flat sheet PES membrane. The modified membranes were characterized by X-ray photoelectron spectroscopy (XPS), SEM, contact angle measurements, pore size distribution, liquid entry pressure and atomic force microscopy. Results revealed that the plasma modification converted hydrophilic membranes of a contact angle 0◦ into hydrophobic ones with water contact angle above 120◦ . Fluorination was ascribed to the wettability change of the membrane from hydrophilic to hydrophobic via insertion and possibly deposition. Direct contact membrane distillation of the hollow fibers using 4% NaCl solution yielded a water flux of 45.4 kg/m2 h at a feed temperature of 63.3 ◦ C. A rather high evaporation efficiency of the membrane distillation process was estimated in comparison with literature results. Direct contact membrane distillation (DCMD) stability test showed a water flux of 42.1 kg/m2 h using 4 wt% NaCl as feed (at the temperature of 60.5 ± 0.2 ◦ C). No leakage was observed for 54 h indicating a stable membrane performance. The high evaporation efficiency and water flux were ascribed most probably to the high porosity of the base membrane. © 2012 Elsevier B.V. All rights reserved. 1. Introduction Membrane distillation (MD) is a thermally driven process that depends on the difference of the partial water vapor pressure across a non-wetting, hydrophobic, porous membrane [1,2]. Emerged nearly 50 years ago [3–5], no large scale MD plants have been implemented yet for desalination. There are several scientific and technological challenges that hamper its industrial applications [6,7]. The major barriers include MD membrane and module design, membrane pore wetting, low permeate flow rate, and flux decay as well as uncertain energy and economic costs. These challenges have attracted scientists and engineers striving for the best membrane performance, module and process design [8–11], among which the selection of membrane materials was the most important. A key requirement for distillation membrane is that the membrane should not be wetted by water. Therefore, in early works, commercial hydrophobic membranes were used because of their intrinsically hydrophobic characteristics that resist the pore wetting of the membranes [12]. But these hydrophobic membranes ∗ Corresponding authors at: Shanghai Advanced Research Institute, Chinese Academy of Sciences, China. Tel.: +86 21 20325162; fax: +86 21 20325034. E-mail address: [email protected] (T. He). 0376-7388/$ – see front matter © 2012 Elsevier B.V. All rights reserved. doi:10.1016/j.memsci.2012.03.031 usually give low permeability in MD process. It is believed that hydrophobic materials may cause severe temperature polarization and thereby lower the evaporation efficiency in the membrane distillation process due to their good thermal conductivity. On the other hand hydrophilic membrane may suffer less temperature polarization and demonstrate higher evaporation efficiency because of their high thermal resistance. In recent years, a series of works have reported the development of dual layer membrane consists of a hydrophobic layer and a hydrophilic support. Khayet et al. [13–15] prepared a series of hydrophobic/hydrophilic flat-sheet membrane for direct contact membrane distillation (DCMD) by phase inversion. These membranes had a composite structure with a thin hydrophobic layer (thus low resistance to water vapor diffusion) and a thick hydrophilic sublayer (a low conductive heat loss). Although the concept was theoretically attractive, results did not reflect significant advantages. Qtaishat et al. [16] continued their work on the double layer membrane by mixing fluorinated macromolecules (SMMs) with polyetherimide (PEI). The membranes showed a contact angle of 100◦ at the top surface with a LEPw value of 4.7 bars. Unfortunately, the water flux in DCMD was fairly low, about 18 L/m2 h at a feed temperature of around 65 ◦ C and a distillate temperature of 15 ◦ C. A separation factor above 99% was observed. The reasons may lie in that the X. Wei et al. / Journal of Membrane Science 407–408 (2012) 164–175 structure of the hydrophobic layer was not optimized and thereby may impart higher mass transfer resistance. Surface modification by plasma polymerization for the formation of a hydrophobic layer on a hydrophilic base membrane was conducted as well. For example, Kong et al. [17,18] modified a hydrophilic microporous cellulose nitrate membrane surface via plasma polymerization of octafluorocyclobutane (OFCB) and vinyltrimethylsilicon/carbon tetrafluoride (VTMS/CF4 ). The membrane was tested in a DCMD system using a 0.3–0.5 M NaCl solution as feed. A water flux of 32.0 kg/m2 h was observed at a feed and cold distillate temperature of 70/25 ◦ C respectively. Unfortunately, the membrane showed a salt rejection of only 92.1%, indicating the salt transfer across the membrane. They studied the plasma glow discharge time on the membrane performance and found that longer polymerization time yielded membrane with higher salt rejection up to 99% but the flux decreased significantly possibly due to that longer reaction time led to thicker coating layer and consequently a higher mass transport resistance and eventually a higher salt rejection and a lower flux. The researchers from Sirkar’s group [8,19–21] have reported in a number of publications on hydrophobic/hydrophilic hollow fiber membranes having a thin layer of microporous coating of silicon fluoropolymer plasma polymerized on the fiber outer surface. The modified hydrophobic PP membranes showed significantly high water flux 79 kg/(m2 h) at 90 ◦ C in a cross flow module [21]. It should be noted that all these surface modification has been mainly focused on the one side of membrane surface, not in the membrane matrix and the other side of the membrane. Plasma surface modification has shown advantages in changing the surface wettability of the materials in the nanometer scale, without affecting the bulk properties, and has been widely used in membrane surface modification [22–25]. However, so far, plasma treatment in membranes has been mostly focused on improving the hydrophilicity of the membrane for better fouling resistance [26,27]. Plasma polymerization has been widely used for the surface modification of membranes as mentioned above. CF4 plasma has been used to improve the membrane hydrophobicity and fluorinated membrane has been tested for gas permeation [28–32] and for blood compatibility [33], but using CF4 plasma without other monomers for membrane surface modification has not yet been widely applied for membrane distillation. CF4 plasma treatment showed a moderate etching and a strong fluorination effect which introduced fluorine functional groups in the material. Therefore CF4 plasma surface modification can be used to reduce the surface energy, enhance the material surface roughness and make the material surface more hydrophobic. We have recently conducted CF4 plasma enhanced chemical vapor deposition (CVD) to make superhydrophobic composite resins [34]. It was found that the process depends heavily on the process parameters. At an optimal condition, a suitable etching and fluorination yields a superhydrophobic surface. It is believed that fluorination and deposition of fluorocarbon materials was the main reason for the wettability change of the surface but was not elucidated clearly. Based on our previous work on surface modification using CF4 plasma for polyester resins, we are going to investigate the use of CF4 plasma modification of hydrophilic base membranes for membrane distillation. The surface modification process was investigated in order to optimize the treatment conditions by changing the glow discharge power and treatment time. The surface treatment effects were characterized by contact angle measurements, liquid entry pressure and X-ray photo electron spectroscopy. The membrane performance was evaluated in a direct contact membrane distillation of 4% NaCl solutions. The evaporation efficiency of the membrane distillation process was estimated and compared with literature reports in order to assess the performance of the membrane and to shed light on the new directions for 165 the preparation of high performance of membrane distillation membrane. 2. Experimental 2.1. Chemicals Sodium chloride (NaCl, AG grade) was supplied by Nanjing Ningshi Chemical Reagent Co., Ltd. All chemicals were used as received. Deionized water was used in the direct contact membrane distillation (DCMD) processes. 2.2. Membrane Permanently hydrophilic polyethersulfone (PES) flat sheet and hollow fiber membranes were provided by Nanjing Altrateck Co. Ltd. The thickness of the PES flat membrane was measured with a micrometer and the average of 8 points is reported. The inner and outer diameters of PES hollow fiber membrane were determined via a zoom biological microscope (Nikon YS1000) equipped with a CCD camera monitored by a computer. The accuracy was within 0.02 mm. 2.3. CF4 plasma surface modification Plasma treatment was performed on a M4L plasma system provided by PVA TePla Co. Ltd. The membranes, either flat-sheet or hollow fibers, were placed on the plate of the plasma chamber. The chamber was evacuated to 50 mTorr at a speed of 40 L/min. Then argon gas was injected at a rate of 100 standard cubic centimeters (SCCM) with a glow discharge at 45 W for 30 s. The system was evacuated to 100 mTorr. Then CF4 was injected at a speed of 18 SCCM till the pressure of 200 mTorr. Radio frequency (RF) glow discharge power was set in the range of 50–400 W and the treatment time was 5–40 min. Thereafter, the system was evacuated again to a pressure below 100 mTorr and N2 was introduced into the chamber to atmosphere pressure for 15 min. The membranes were taken out and preserved for further investigation. 2.4. Instrumentation X-ray photoelectron spectroscopy (K-alpha X-ray photoelectron spectroscopy, Thermo Fisher). Surface survey data was collected followed by high resolution scans over C1s (278–298 eV), O1s (525–545 eV), S2s (222–242 eV), N1s (392–402 eV) and F1s (675–695 eV). Peak areas were calculated using Gaussian fit program. Relative peak area ratios were calculated according to published results. Liquid entry pressure (LEPw) membrane porosity was measured by weight difference at wet and dry state. After the membrane sample wetted by ethanol, samples were immersed in water to replace the ethanol within the membrane pores. Wet weight was measured by wiping off surface water on the membrane and the dry weight was measured after membrane was completely dried in an oven at 65 ◦ C. The membrane porosity was obtained by: ε= mwet − mdry H2 O Vmem × 100 (1) in which ε is the membrane porosity, mwet and mdry are the wet and dry weight of the sample, H2 O is the density of water and Vmem is the total membrane volume. Water contact angles (CA) were evaluated by means of a contact angle goniometer (Maist DropMeter A-100P) at room temperature by deposition of a 5 L water droplet on the membrane surface. Each reported value was calculated by averaging six measurements 166 X. Wei et al. / Journal of Membrane Science 407–408 (2012) 164–175 at different points on the sample. The CA of hollow fiber membrane inner surface were measured according to literature [35]. Part of the hollow fiber was encased in a clear cylinder, and high vacuum grease was used to seal the clearance between the fiber and the cylinder. The fiber was slowly submerged into the water and the height difference (h) between the top of the fiber and water surface was recorded upon water protrusion from the top of the fiber. The experiments were repeated four times. The contact angles are determined from: hgd i = cos−1 − 4l (2) where is water density, h is the height between the water protrusion and water surface in the beaker and g is acceleration due to gravity, di is the inner diameter of the fiber, and l is the surface tension of water. Scanning electron microscopy samples were prepared by cryogenic breaking. The samples were allowed to dry under vacuum at 30 ◦ C overnight and then coated with a thin layer of gold. For low magnifications, HITACH TM100 was used for analysis. For photos at high magnifications, field emission scanning electron microscopy (FESEM HITACHI JapanS-4800) was utilized for analysis. Pore size analysis was conducted with capillary flow porometry (Model CFP-1100-A, supplied by Porous Material Inc.). The membrane was pre-wetted with commercial low surface tension liquid Porewick (surface tension of 16 dyn/cm based on the supplier’s datasheet). After mounting the sample onto the test cell, the measurement was managed with a program consisting of wet-run and dry-run. The wet-run was realized by replacing the wetting liquid within certain pore size by compressed air at certain pressure till the membrane was dried out (wet-run). Then the air flow rate of the membrane was tested by decreasing the air pressure (dry-run). Membrane pore size was determined by commercial software from PMI based on Young–Laplace equation: p = 2 cos r (3) where p, , , r are the pressure difference, surface tension, contact angle and membrane pore size, respectively. In case that the membrane pore size was out of the measurement range of PMI, bovine serum albumin (BSA) rejection measurement was used for pore size characterization [36]. For BSA rejection measurement, a cross-flow setup was employed. The concentration of BSA was monitored by UV photospectrometry (UV-2802, UNICO (Shanghai) Instruments Co., Ltd.) and determined according to the measured standard curve. Fig. 1. Schematic of direct contact membrane distillation process: (1) test cell; (2) digital thermometer; (3) flowmeter; (4) circulation pump; (5) distillate tank with thermostatic jacket; (6) feed tank with thermostatic jacket; (7) overflow collector. The salt rejection rate R was determined according to the following equation: R= Cf − Cp × 100 Cf (5) where Cf , Cp representing the NaCl concentration of the feed and the permeate, respectively. Mettler-Toledo FE30 conductivity meter was used for monitoring the conductivity of the feed and distillate. A conductivity and salt concentration standard curve was established by plotting the conductivities of a series of solutions against their salt concentrations. The concentrations of the experimental solutions were then determined according to the plot. For the flat-sheet membrane, both feed and distillate flow rates were kept constant at 0.36 m/s. For the hollow fiber membrane module, feed flows inside of the fiber at a rate in the range of 0.5–3.0 m/s, and the cold water circulates outside of the hollow fiber within the range of 0.17–1.36 m/s. The inlet temperature at the cold side was kept at 20.0 ± 3 ◦ C and the hot water temperature was varied within 40.0–75.0 ◦ C. To keep the feed concentration constant, additional deionized water was added to the feed side according to the amount of water that has been transferred across the membrane. Stability of the hollow fiber membranes performance was tested for 6 h/day and for 9 days continuously. 2.5. DCMD experiments Fig. 1 schematically illustrates the direct contact membrane distillation process. The membrane was placed in the membrane module as shown in Fig. 2. The feed solution and distillate (deionized water) were circulated concurrently. The inlet and outlet temperatures of the feed and distillate were read out with digital thermometer (calibrated with a standard mercury thermometer). Water transferred across the membrane from feed to distillate was collected as an overflow. Water flux was calculated according to FMD = V A·t (4) With FMD , V, A, and t representing the water flux, L/m2 h; amount of permeate (L or kg), effective membrane surface area (m2 ), and the time duration (h), respectively. Fig. 2. Schemes of the flat and hollow fiber membrane modules with the dimensions of the test cell. X. Wei et al. / Journal of Membrane Science 407–408 (2012) 164–175 100 100 80 80 CA(o) 120 CA(o) 120 60 Top Bottom 40 167 60 Top Bottom 40 20 20 0 0 0 50 100 150 200 250 300 350 400 450 0 CF4 plasma power(W) Fig. 3. Water contact angles of top and bottom surface of modified PES flatsheet membrane as a function of the CF4 treatment power. The pretreatment was using argon plasma at a glow discharge power of 45 W with a duration of 30 s (Ar 45 W 30 s), and the CF4 treatment time was 10 min. Between each operation, the membrane module was kept in the setup. 3. Results and discussion A flat sheet polyethersulfone (PES) membrane was used for optimization of the plasma treatment process. The PES membrane was hydrophilic with a static water contact angle of 60.0 ± 2.0◦ and 0◦ at the top and bottom surface, respectively. It was mechanically robust and its water flux was 1070 LMHBar. The membrane showed no water transfer when applied directly in a DCMD process. In order to use this membrane in MD, surface modification with CF4 plasma treatment was carried out. 3.1. Plasma treatment Plasma treatment effect is strongly dependent on the glow discharge power and treatment duration. In order to optimize the treatment condition, a series of experiments were carried out with respect to these two issues, respectively. The pretreatment was realized with argon plasma. The purpose of pretreatment with argon was to make the membrane free of dust particles and ready for further treatment. The condition was optimized by varying the treatment time and glow discharge power. It was found that at 45 W at 30 s the pretreatment was sufficient. Other plasma gases, for example O2 , should not be used because oxygen atom can be inserted in the surface, which is not favorable for the creation of hydrophobic surface. The surface after pretreatment was then exposed to CF4 plasma and the influence of the glow discharge power of CF4 plasma on the membrane water contact angle was investigated with treatment duration set for 10 min. As shown in Fig. 3, the original membrane showed a water contact angle of 60.0 ± 2.0/0◦ at the top/bottom surfaces. After treated with the CF4 plasma, at 25 W for 10 min, the top surface contact angle increased slightly to 71.0 ± 2.0◦ while the CA of bottom surface remained unchanged. At 50 W, the CA of both surfaces increased significantly to 115.0 ± 2.0◦ . At 100 W, the CA of the surfaces increased slightly to 117.0 ± 2.0◦ /119.0 ± 2.0◦ for top/bottom surfaces. Further increment of the treatment power resulted in very slight change in the contact angle indicating the surface modification has reached saturation. We chose 200 W for further surface modification. The influence of the CF4 plasma treatment time at 200 W on the membrane contact angle was shown in Fig. 4. Similar to the trend in glow discharge power effect, the water contact angle change 10 20 30 40 CF4 plasma treat time(min) Fig. 4. Water contact angles of top and bottom surfaces of modified PES flatsheet membrane as a function of CF4 treatment. The pretreatment was argon 45 W 30 s, and the CF4 treatment glow discharge power was 200 W. showed initially significant increase within short time and thereafter a level-off. At 5 min of treatment, the top surface contact angle increased from 62.0◦ to 113.0◦ and the bottom from 0◦ to 118.0◦ . It appears that from 5 min to 40 min, the contact angle at the top surface increased slightly from 113.0◦ to 125.0◦ , and the contact angle of the bottom surface increased from 118.0◦ to 124.0◦ . Water contact angle is a surface property related to the surface composition, roughness and surface porosity [37]. In general, a rough and highly porous surface normally shows a high water contact angle. Scanning electron microscopy (SEM) images of the PES membrane structure are shown in Fig. 5. The membrane is asymmetric in the cross section (Fig. 5a) with a dense skin layer (Fig. 5b) and a porous support (Fig. 5c). Small pores of 20–30 nm are observable from the top surface (Fig. 5d) while the bottom surface shows much larger pores in the range of 1.0–4.0 m, in agreement with the cross section observation. The asymmetric structure might be the reason for the difference in the water contact angles between the top and bottom surface both before and after plasma modification [37]. To guarantee sufficient surface modification, we have chosen 200 W 30 min for CF4 glow discharge and 45 W 30 s for argon pretreatment. The characteristics of PES membranes before and after plasma treatment are listed in Table 1. The membrane thickness, porosity, gas permeability remained unchanged. Pore size measurement by PMI failed to give exact bubble point due to the low limit in the set pressure, however, indicating that the bubble pore size is below 70 nm, which agrees to the SEM observations with the presence of pores of 20–30 nm (Fig. 5d). The modification is more pronounced in the change of liquid entry pressure of water (LEPw). The LEPw of the PES before and after treatment were approximately 0.1 bar and 3.7 bar, respectively. LEPw is an indication of the ability of a hydrophobic membrane against wetting in the MD process. If the LEPw is low, water Table 1 Characteristics of PES membrane before and after surface treatment. Properties Units Before After Membrane thickness Porosity Average pore size Surface pore size CA (t/b) LEPw Elongation at break Mechanical strength PWP m % nm nm deg bar % N kg/m2 h bar 201 ± 14 79.2 ± 1.8 <70 40 58.0/0 0.1 19.7 ± 3.1 3.0 ± 0.8 1070 201 ± 14 79.2 ± 1.8 <70 40 124.0/125.0 3.7 ± 0.1 13.6 ± 2.3 2.8 ± 0.2 – 168 X. Wei et al. / Journal of Membrane Science 407–408 (2012) 164–175 Fig. 5. SEM images of PES flatsheet membranes. (a) Cross-section at 500×; (b) cross-section of the top layer at 5000×; (c) cross-section of the bottom layer at 5000×; (d) top surface at 30K×; (e) bottom surface at 30K×. could be pressed easily inside the pore of the membrane leading to pore wetting and possibly solute leakage. Therefore, a membrane with a higher LEPw is expected to perform better than that with a lower one. Based on the LEPw and contact angle changes of the PES membrane, one may conclude that the hydrophilic membrane has indeed been transformed into a hydrophobic one. It is noted that the mechanical property of the membrane changed. In literature, when surface modification was carried out on membrane, the mechanical properties for example, mechanical strength is normally strengthened [17,18]. However, the mechanical strength decreased in our case. In order to elucidate the surface modification mechanism, we have carried out X-ray photoelectron spectroscopy (XPS). 3.1.1. Plasma modification mechanism Plasma modification is a rather complicated process including etching, atomic insertion, deposition and polymerization. From our former research [34] and published literatures [25,38], the main effects of CF4 plasma treatment are etching and F atom insertion, and possibly deposition depending on the operation conditions. XPS is always used to assess the composition of a surface and moreover can be used to analyze the bonding status of an atom. XPS survey scans (Fig. 6a and b) revealed that before surface modification, C is the major component and after surface modification, F becomes dominant. The atomic concentrations of different elements are listed in Table 2. After surface modification, the carbon atomic concentration decreased drastically from 74.0% to 43.4% and F becomes dominant with 50.8% abundance. The F/C ratio increased from 0 to 1.2 indicating that CF4 plasma treatment has brought a fluorinated layer to the surface, which is responsible for the wettability change of the membrane Carbon XPS spectra (with simulation) before and after modification are shown in Fig. 6c and d, respectively. It can be seen that the before modification, C atoms were mainly in the form of C–C, C–H status. After surface modification, carbon atoms were present as CF2 –CF2 , C–F, CF3 , among which CF2 –CF2 accounts for the highest percentage. Based on this analysis we envisioned the following scenario for the process of the CF4 plasma modification as depicted in Fig. 7. A plasma gas is overall neutral but is in fact a mixture of cations, anions and radicals. These particle species are highly active and energetic and thus react drastically when collide into any substances on their way. As such, fluorine elements are introduced to the membrane surface leading to a hydrophobic membrane. Furthermore, the gaseous species penetrates into the membrane pores or channels inside the porous structure of the membrane (as depicted in Fig. 7), and thereby modifies the interior of the membrane, which helps to avoid pore wetting during membrane distillation. In some cases, when CF3 species is inserted in the Table 2 PES membrane top surface composition before and after modification (at.%). Sample C1s O1s S2p N1s Cl2p F1s O/C F/C Original Modified 74.0 43.4 17.5 4.8 7.0 1.0 1 0.0 0.9 0.00 0.0 50.8 0.2 0.1 0.0 1.2 X. Wei et al. / Journal of Membrane Science 407–408 (2012) 164–175 169 1.0M C1s (a) 250.0k (b) F1s 800.0k Intensity (a.u.) Intensity (a.u.) 200.0k O1s 150.0k 100.0k S2p 50.0k 600.0k 400.0k N1s 200.0k C1s N1s O1s S2p 0.0 0 200 400 600 800 1000 1200 0.0 1400 0 200 Bonding Engergy/eV 40k (c) C-C/C-H 400 600 800 18.0k C1s Intensity (a.u.) Intensity (a.u.) 1200 1400 C1s (d) CF2-CF2 15.0k 30k 20k 10k 1000 Bonding Engergy/eV C-N 12.0k C-C/C-H CH2-CF2 9.0k 6.0k 3.0k C-F CF3 N-C=O pi-pi* 0 0.0 280 282 284 286 288 290 292 294 296 298 Bonding energy/eV 280 282 284 286 288 290 292 294 296 298 Bonding energy/eV Fig. 6. XPS survey scans of the PES flatsheet membranes (a, b) and carbon spectra before and after surface modification: (a) original surface; (b) surface modified membrane surface; (c) carbon spectra before surface modification; (d) carbon spectra after surface modification (with simulations). polymer chain, the C–C bond is broken and the polymer chain becomes shorter, which might explain the lower mechanical strength after surface modification. We used atomic force microscopy (AFM) to characterize the PES membrane before and after surface modification as shown in Fig. 8. Tapping mode AFM height images showed granular particulate structures for both before and after surface modification. The membrane pores were assigned to the cavities among the granular structure. The average grain size before surface modification was 51 ± 5 nm, and 65 ± 5 nm after surface modification. The growth of the grain size was ascribed to the deposition of the fluorinated layer. By the grain size analysis, we estimated that a fluorinated layer of less than 10 nm was deposited on the PES membrane, which helps to the clear the surface modification mechanism. 3.1.2. DCMD properties of hydrophobic flat sheet PES membrane Using a 4 wt% NaCl water solution as feed, the water flux and separation performance of CF4 plasma treated PES flat sheet membranes were studied. Fig. 9 shows the water flux as a function of feed water temperature from 47.4 to 74.5 ◦ C. At a feed inlet temperature of 47.4 ◦ C, the water flux was 12.6 kg/m2 h, when the feed temperature increased to 74.5 ◦ C, the water flux increased to 40.9 kg/m2 h. Since the membrane thickness was about 200 m, it was expected that if the membrane thickness was 100 m, the membrane flux would be much higher. A salt rejection of 99.97% was obtained. It has been reported, for example, the surface plasma polymerization modified cellulose nitrate membrane showed a water flux of 32.0 kg/m2 h at a feed and cold distillate temperature of 70/25 ◦ C, respectively with a salt rejection of 92.1% [17,18]. At similar operating condition, our membrane has shown a higher flux with much higher salt rejection. Probably our membrane is more homogeneous in surface modification and thus less prone to pore wetting and thereby higher salt rejection. 3.2. PES hollow fiber membrane With the successful surface modification of flat sheet membrane, we continued surface modification of the PES hollow fiber membranes due to advantages of higher packing density and wider availability. Fig. 7. Schematic illustration of CF4 plasma modification of the porous membranes. 3.2.1. Characteristics of PES hollow fiber membrane Similar to the flat-sheet PES membrane, the hollow fiber PES membrane is asymmetric in the cross section with a relatively dense inner skin and a sponge-like porous support (Fig. 10a–c). At 5000× magnification, the inner surface (Fig. 10d) appeared relatively smooth and the outer surface (Fig. 10e) showed open 170 X. Wei et al. / Journal of Membrane Science 407–408 (2012) 164–175 Fig. 8. Tapping mode AFM height images of the PES membrane before (a) and after surface modification. Image size: 1 m × 1 m, Z range: 250 nm. methods reported in literature [35] with CA of 120◦ /96◦ for outer/inner surfaces. 2 Flux (kg/m h) 45 30 15 45 60 75 o Tf ( C) Fig. 9. Effect of the feed inlet temperature (Tf ) on the water flux in DCMD for surface modified PES flat membrane. Feed was 4 wt% NaCl water solution, and distillate was deionized water (18.3 ± 1.9 ◦ C). Flow rate at both feed and distillate was 0.36 m/s. pores. The inner surface of the hollow fiber membranes shows much smaller pore size than its flat sheet membrane counterparts. Table 3 lists the characteristics of the hollow fiber membranes. The BSA rejection of the membrane was about 40% indicating that the membrane was a UF membrane with relatively large pores. The inner and outer diameters of the hollow fiber were measured to be 0.78/1.28 mm, respectively, in agreement with the standard measure of ultrafiltration hollow fiber membranes. Water permeability of the membrane was 1436 ± 260 kg/m2 h bar, higher than the flat sheet membrane. After CF4 plasma treatment, the membrane was changed from hydrophilic to hydrophobic. The water contact angle of the hollow fiber membrane was monitored according to Table 3 Characteristics of PES hollow fiber membranes. Properties Units Original Surface modified ID/OD PWP BSA rejection LEPw Inner/outer CA m kg/m2 h bar % bar degree 780/1280 1436 ± 260 40 NA 0/0 NA NA 3.1 ± 0.3 96 ± 2/120 ±2 NA: not applicable. 3.2.2. Surface composition analysis Fig. 11 shows the atomic composition before and after plasma modification of the hollow fiber membranes. Similar to the flat sheet membrane, the hollow fiber membrane does not contain any fluorine atoms before surface modification as expected. At the inner surface, after surface modification the F/C ratio increased slightly to 0.07 however with a contact angle of 96.4◦ . For the outer surface, the F/C ratio was 1.09, similar to that in the flat PES membrane and the CA was 120◦ (Table 4). From this comparison, it can be seen that the degree of fluorination inside the hollow fiber was less compared to outer surface possibly due to that the plasma had to diffuse through the membrane wall to reach the inner surface leading to asymmetric modification of the hollow fiber membrane. It was nevertheless very interesting to test the membrane distillation performance of such a membrane. 3.2.3. DCMD properties of hydrophobic PES membrane The MD performance of the PES hollow fiber was tested using 4% NaCl as a feed and the feed temperature varied from 45.3 to 73.8 ◦ C. The water flux plotted against the feed temperature showed an exponential increase from 20.4 to 66.7 kg/m2 h against the feed temperature (Fig. 12). Throughout the experiment, the permeate conductivities were all below 10 s/cm, and the salt rejections were as high as 99.97%, indicating a nearly complete rejection of NaCl. This result was very interesting in that it showed that the asymmetrically surface modified PES hollow fiber has shown a good MD performance. Moreover, compared to flat sheet membrane, the water flux of the hollow fiber is about 50% higher, which might be due to less mass transfer resistance of the base membrane and possibly improved hydrodynamic conditions. Fig. 13 shows the effect of the feed NaCl concentration at 63.3 ◦ C on the water flux. Following the increase in the NaCl concentration from 4% to 16%, the water flux declined from 45.4 Table 4 Surface elements analysis of PES hollow fiber membranes. Sample C1s N1s O1s F1s S2p O/C F/C Original-inner Original-outer Modified-inner Modified-outer 80.67 75.94 72.31 44.37 1.64 2.27 2.16 0.00 12.49 16.51 14.95 5.94 0.00 0.00 4.91 48.46 5.19 5.28 5.68 1.23 0.15 0.22 0.21 0.13 0.00 0.00 0.07 1.09 X. Wei et al. / Journal of Membrane Science 407–408 (2012) 164–175 171 Fig. 10. SEM photos of PES hollow fiber membranes. (a) Cross-section at 500×; (b) inner skin layer at 5000×; (c) outer skin layer at 5000×; (d) inner surface at 5000×; (e) outer surface at 5000×. to 30.1 kg/m2 h. The reduction of water vapor pressure at high solute concentrations was probably the main cause for the decrease in permeation flux. However, with the increase of feed concentration, the concentration polarization might also become more announced and thus impart the decrease of water flux in the MD process. 3.2.4. Effect of the flow rates on the DCMD flux Membrane distillation process combines the mass transfer along with heat transfer. Both on the feed side and the distillate side, there exists the temperature polarization and the concentration polarization. The flow rate may alter both thereby affecting the water flux of the process. Fig. 14 shows the effects of the feed flow rate in the bore of the membrane on the permeate flux. The cold distillate was circulating at the shell side. The water flux increases gradually with the increase of the flow rate of hot NaCl solution. At higher flow rate temperature polarization could be significantly suppressed because of the improvement in hydrodynamic conditions and the trans-membrane temperature difference is thus increased. This result agrees well with literature [39]. Fig. 15 shows the effect of the flow rate of the distillate at the shell side of the membrane on the permeate flux. The flux increased slightly when the flow rate of the cold distillate increased from 0.17 to 0.68 m/s. Further increase in the flow rate of the cold distillate did not give significant improvement in water flux. Estimation of the Reynold number at the shell side of the membrane indicated that at the flow rate of 0.34 m/s, the Reynolds was around 1900. At higher flow rate, the turbulent flow may occur, thus further increase in the hydrodynamics does not add much to the improvement in the water flux. 3.2.5. Performance stability In membrane distillation, it is well-known that the long term stability of the membrane is a key issue for the practical application of MD. In order to evaluate the performance stability of our PES hollow fiber membrane, we carried out an experiment using a NaCl solution as feed at 60.5 ◦ C. The membrane module was tested in total 54 h for 9 days, 6 h per day. During the test interval, the membrane module was in contact with both feed and distillate. Fig. 16 shows the DCMD water flux against time. The flux increased slightly from 38.8 to 42.1 kg/m2 h in the first two days, and then declined slowly to 36.4 kg/m2 h. The first increase was due to the removal of air bubbles adsorbed onto the membrane walls. Later on, the flux stayed stable around 39 kg/m2 h and after 40 h, a slight drop was observed to 36.4 kg/m2 h. As for salt rejection, the cold distillate water’s conductivity decreased slightly from 7.81 to 7.45 s/cm throughout the experiments showing a nearly 100% salt rejection. Pore wetting has been ascribed as the main reason for salt leakage and low flux. Thus no salt leakage was observed indicating there was no membrane pore wetting throughout the whole test indicating that our membrane holds potentials to compete with other hydrophobic membranes. 172 X. Wei et al. / Journal of Membrane Science 407–408 (2012) 164–175 400.0k 400.0k (a) Intensity (a.u.) Intensity (a.u.) C1s 300.0k 250.0k 200.0k O1s 150.0k 0.0 250.0k O1s 200.0k 150.0k S2p 0 C1s 300.0k N1s 100.0k N1s 100.0k 50.0k (b) 350.0k 350.0k S2p 50.0k 200 400 600 800 1000 1200 0.0 1400 0 200 400 600 800 1000 1200 1400 1200 1400 Bonding Engergy/eV Bonding Engergy/eV 1.2M 160.0k (c) 120.0k O1s 100.0k 80.0k F1s 60.0k N1s 40.0k S2p 800.0k 600.0k 400.0k 200.0k C1s 20.0k 0.0 F1s C1s Intensity (a.u.) Intensity (a.u.) (d) 1.0M 140.0k S2p 0.0 0 200 400 600 800 1000 1200 1400 0 200 N1s 400 O1s 600 800 1000 Bonding Engergy/eV Bonding Engergy/eV Fig. 11. XPS survey scans of PES hollow fiber membranes before and after CF4 plasma treatment. (a) Original inner surface; (b) original outer surface; (c) modified inner surface; (d) modified outer surface. Modification parameters: Ar 45 W 30 s, CF4 200 W 30 min. 3.3. Evaporation efficiency of the hollow fiber membranes Evaporation efficiency (EE) of the MD process is defined as the ratio between the heat transfer contributing to water flux (latent heat) and the total feed heat-loss in the module at different temperatures, which is an indication of the efficiency of energy used to produce condensate [2]. It can be calculated according to [40] EE = FMD A Hv (6) mf C̄P (tf1 − tf2 ) where FMD , A, Hv , C̄p , mf , tf1 , and tf2 representing the MD flux, membrane area, latent heat of vaporization, average heat capacity of water, feed mass flow rate, and feed inlet and outlet temperatures, respectively. As shown in Fig. 17, EE increased from 56% to 79% along with the increase of the feed temperature from 47.4 ◦ C to 66.7 ◦ C in a nearly exponential relationship. From 60 ◦ C on, further increase in the feed temperature did not lead to increase in the EE. The initial increase in EE with temperature was probably due to the exponential relationship between the partial vapor pressure and the temperature. The 60 70 55 50 Flux (kg/m .h ) 45 2 2 Flux (kg/m .h ) 60 50 40 30 40 35 30 25 20 20 45 50 55 60 65 70 75 o Tf ( C) Fig. 12. The DCMD flux of the CF4 plasma modified PES hollow fiber membrane as an effect of the feed temperature. Salt solution: flow rate 2.0 m/s. Cold distillate water: 17.5 ◦ C, flow rate 0.68 m/s. 0 4 8 12 16 NaCl concentration(wt%) Fig. 13. The DCMD flux of the CF4 plasma modified PES hollow fiber membrane as an effect of the feed NaCl concentration. Feed solution. 2.0 m/s, 63.3 ± 0.2 ◦ C. Cold distillate water: 0.68 m/s, 20 ± 0.5 ◦ C. X. Wei et al. / Journal of Membrane Science 407–408 (2012) 164–175 173 90 Evaporation Efficiency(%) 45 2 Flux (kg/m .h ) 40 35 30 25 80 70 60 50 20 0.5 1.0 1.5 2.0 2.5 3.0 40 -1 Flow rate m.s 60 55 Flux ( kg/m2h ) 50 45 40 35 30 25 20 0.4 0.6 0.8 1.0 1.2 1.4 -1 Flow rate m.s Fig. 15. The DCMD flux of the CF4 plasma modified PES hollow fiber membrane as an effect of the flow rate of the distillate, 20 ± 0.5 ◦ C at the shell side of the membrane on the permeate flux. Feed solution NaCl 4%: 63.3 ± 0.2 ◦ C, 2.0 m/s. 50 45 40 Flux (kg/m2h) 50 55 60 65 o 70 75 Feed inlet temperature( C) Fig. 14. The DCMD flux of the CF4 plasma modified PES hollow fiber membrane as an effect of the flow rate of the feed solution. Feed NaCl 4% with a temperature of 63.3 ± 0.2 ◦ C at the bore side of the membrane and cold distillate at the shell side with a temperature of 20 ± 0.5 ◦ C, 0.68 m/s. 0.2 45 35 30 25 20 Fig. 17. Evaporation efficiency of the hollow fiber PES membranes at different feed inlet temperatures. permeation flux is proportional to the vapor pressure difference, or to the vapor pressure of the feed temperature since the distillate vapor pressure is significantly lower than the feed vapor pressure. Therefore, the permeation flux shows an exponential relation with the feed temperature, thus EE against temperature as well. It was observed that when the feed temperature reached 74.5 ◦ C, the EE value declined slightly to 76%. At higher feed temperature, the heat loss at the feed side to environment may become more significant resulting in lower EE. Zhang et al. [41] prepared a non-woven supported PTFE flat microfiltration membrane with a highest EE of 50% at the feed inlet temperature of 70 ◦ C. Bonyadi et al. [40] prepared PVDF hollow fiber membrane with a EE of 58% at the feed inlet temperature of 72 ◦ C. Wang et al. [42] prepared mixed matrix PVDF membrane by addition of organophilic clay into the membrane dope solution with the highest EE of 50% at the feed inlet temperature of 80 ◦ C. In contrast, our membrane showed much higher EE than literature reports. The difference might be related to that the base membrane is hydrophilic instead of hydrophobic. It is well known that hydrophilic materials have lower thermal conductivity than their hydrophobic counterparts. When used in membrane distillation, the low thermal conductivity of the base membrane may show higher evaporation efficiency due to less temperature polarization. However, the exact causes need further investigation and confirmation. Recently, very high thermal efficiency or evaporation efficiency values in the range between 70 and 85% were reported by Sirkar’s group [21] using a cascade of cross-flow hollow fiber membrane distillation device integrated with a heat exchanger. The evaporation efficiency values reported in this paper are quite close to what has been observed in literature. Thus, application of the membranes in this work might achieve at least comparable thermal efficiency or evaporation efficiency by adopt of similar cross-flow distillation device with a heat exchanger. 15 3.4. Performance comparison with literature results 10 5 0 0 5 10 15 20 25 30 35 40 45 50 55 Time(hr) Fig. 16. Performance stability of the surface modified PES hollow fiber membrane. Hot NaCl solution: 4 wt% NaCl, 60.5 ± 0.2 ◦ C, 2.0 m/s. Cold distillate water: 20 ± 0.5 ◦ C, 0.68 m/s. Hydrophilic asymmetric PES flat and hollow fiber membranes were surface modified by CF4 plasma treatment. Although the present PES membranes, both flat and hollow fiber membranes, had pore size significantly smaller than those of conventional MD membranes, our membranes showed a quite high water flux up to 66.7 kg/m2 h at a feed temperature of 73.8 ◦ C as seen in Fig. 12. Table 5 lists the performance of 9 different hollow fiber membranes used in DCMD for various separation purposes. Scientifically, it 174 X. Wei et al. / Journal of Membrane Science 407–408 (2012) 164–175 Table 5 Comparison of the performance in DCMD of the hollow fiber membranes in the literatures and the present work. Membrane PP Accurel® S6/2* Modified PP MXFR 3* PVDF PVDF-2 PVDF/PAN dual-layer PVDF HF with nanoscale pores Clay particle PVDF hollow fiber PVDF* Surface modified PES hollow fibera ID/OD (mm) 1.8/2.6 0.63/0.33 0.9/1.2 0.97/1.4 0.52/1.2 1.0/1.38 0.86/1.1 0.92/1.47 0.78/1.2 Pore size (m) 0.22 0.2 0.25 NA 0.41 NA 0.44 0.19–0.42 <0.070 Porosity Feed 0.73 0.65 0.75 – 0.8 0.90 0.80 0.80–0.85 0.79 Distillate Salt conc. tf (◦ C) Shell /Lumen tp (◦ C) Tap water (∼650 s/cm) 1 wt%NaCl 5 mg/L fluoride 1000 ppm NaCl 3.5 wt% NaCl 3.5 wt% NaCl 3.5 wt% NaCl 3.5 wt% NaCl 4 wt% NaCl 90 90 80 85 90 81.3 86 70 73.8 Shell Lumen Shell Shell Lumen Lumen Lumen Lumen Shell 20 15–17 20 20 16.5 17.5 20.5 25 20 Flux (kg/m2 h) Ref 33.3 78.8 35.6 37.3 55.2 79.2 ± 1.2 70.1 67 66.7 ± 4.9 [44] [8] [45] [46] [47] [42] [40] [48] This study Note: NA, not available. a Maximal pore size, others are mean pore size. is difficult to make a solid conclusion due to variation in membrane dimension, feed temperature/feed concentration, flow rate in the bore side and the shell side and most notably the membrane module design. Nevertheless, the present plasma modified PES membrane showed relatively high water flux. It should be note worthy that most of the existing surface modified membranes focused on mainly the top surface or surfaces. In our case, the whole membranes were attempted to be modified, which made it less prone to pore wetting and may be more fouling, scaling resistant. Moreover, according to Li and Sirkar [8], the surface modified PP membrane showed a flux of 78.8 kg/m2 h at a feed temperature of 90 ◦ C in a cross flow configuration, which helped to reduce the temperature polarization. Our membranes were operated in a concurrent flow fashion. At optimized conditions, our membrane should give higher water flux, which is currently under investigation. In addition, both our membrane and the PP membranes are based on commercial products. The systematic research work on the PP membranes at pilot scale indicates that the present membranes are promising in future desalination applications. Most of the PVDF membranes did not show high flux. Yet, very high water flux was obtained by blending PVDF with clay and utilization of delamination technology to form a relatively thin-wall hollow fiber membrane [42]. The water flux of our PES membrane is comparable to the PVDF membrane at a same feed temperature. The water vapor flux in DCMD depends on membrane porosity, and water vapor pressure at the feed-side pore mouth which is strongly affected by temperature polarization in the feed side thermal boundary layer, membrane thickness, pore connectivity and tortuosity, thermal conductivity of the membrane material and the distillate side water vapor pressure. Under similar working conditions, high porosity, pore connectivity and low thermal conductivity may contribute most to the good performance. The membrane thermal conductivity km is generally used to evaluate the membrane conductivity and is given by [43]. km = (1 − ε)ks + εkg (7) where ε is the membrane void fraction (in this case using porosity), and ks and kg are the thermal conductivities of the solid membrane material and of the vapor/air within the pores, respectively. For instance, the thermal conductivity of Millipore PVDF (GVHP) membrane was 0.0858 W m−1 K−1 . Wang et al. [42] reported that the thermal conductivity of a mixed matrix hollow fiber PVDF membrane with a porosity as high as 0.9 was about 0.04–0.05 W m−1 K−1 . Calculation of thermal conductivity of the PES membrane is 0.052 W m−1 K−1 (the porosity = 0.79; thermal conductivity of PES, Ks (PES) = 0.16 W m−1 K−1 ). Surprisingly, the estimated thermal conductivity of present PES membrane was in line with that of the earlier reported PP or PVDF membranes. This comparison indicates that the thermal conductivity of PES membrane may not be the main cause for the high water flux in MD process. The membrane porosity of the PES membrane is relatively higher than some of the membranes, however, is still at the same level as most of the membranes (as listed in Table 5). Other parameters, such as pore connectivity, may contribute significantly to the performance of the PES membranes. However, at this stage, we have no solid scientific proof to confirm this. A more quantitative measure is worthy of investigation in future research. 4. Conclusions Hydrophilic asymmetric PES flat and hollow fiber membranes were surface modified by CF4 plasma treatment. Results showed that after plasma modification the membranes become hydrophobic. The plasma modification mechanism was explored and fluorination was ascribed to the main cause for the converting the hydrophilic membranes into the hydrophobic ones. DCMD tests of the membrane demonstrated that both flat and hollow fiber membranes were good membranes materials in membrane distillation with high water flux and salt rejection. The high salt rejection was ascribed to the homogeneous surface modification by plasma treatment and thus was less prone to pore wetting and eventually less salt leakage in the MD processes. The evaporation efficiency of present MD processes were calculated and compared with literature results. The use of hydrophilic membrane was ascribed to the high evaporation efficiency in the MD process. Finally a long-term stability evaluation of the hollow fiber membrane showed a quite stable water flux and 100% salt rejection. Overall, we have demonstrated a novel approach for converting a hydrophilic ultrafiltration membrane to a hydrophobic one that has shown satisfactory performance in membrane distillation. Further work is expected to investigate and the applicability of this process to other materials and its effect on membrane performance and the scale up processes for large scale applications. Acknowledgements The authors would like to thank the partial financial support from National Natural Science Fund China (Project nos. 20976083, 21176119), the National Key Basic Research Program of China (973 Program) (Project nos. 2012CB932800(TH), 2012CB720903(XML), 2009CB623402). China-Israel Joint Research Program from MOST. X. Wei et al. / Journal of Membrane Science 407–408 (2012) 164–175 List of symbols A Cp Cf EE FMD Hv mdry mf mwet p r t tf1 tf2 V C̄p ε H2 O effective membrane surface area (m2 ) salt concentration in distillate (g/L) salt concentration in the feed (g/L) evaporation efficiency (%) water flux (L/m2 h) latent heat of vaporization (kJ/kg) dry weight (kg) feed mass flow rate (kg/h) wet weight (kg) pressure difference (Pa) membrane pore size (m) time of the experiment (h) feed inlet temperatures (K) feed outlet temperatures (K) amount of water permeate through the membrane during experiment (L or kg) average heat capacity of water (J/kg K) membrane porosity density of water (kg/m3 ) References [1] K. Smolders, A.C.M. Franken, Terminology for membrane distillation, Desalination 72 (1989) 249–262. [2] K.W. Lawson, D.R. Lloyd, Membrane distillation, J. Membr. Sci. 124 (1997) 1–25. [3] B.F. Dodge, Review of distillation processes for the recovery of fresh water from saline waters, Adv. Chem. Ser., 1963. [4] M.E. Findley, Vaporization through porous membranes I & CE process, Des. Dev. 6 (1967) 226–230. [5] M.E. Findley, V.V. Tanna, Y.B. Rao, C.L. Yeh, Mass and heat transfer relations in evaporation through porous membranes, AIChE J. 15 (1969) 483–489. [6] E. Curcio, E. Drioli, Membrane distillation and related operations—a review, Sep. Purif. Rev. 34 (2005) 35–86. [7] M. Khayet, Membrane distillation, in: N.N. Li, A.G. Fane, W.S.W. Ho, T. Matsuura (Eds.), Advanced Membrane Technology and Applications, John Wiley & Sons, Inc., Hoboken, NJ, USA, 2008. [8] B. Li, K.K. Sirkar, Novel membrane and device for direct contact membrane distillation-based desalination process, Ind. Eng. Chem. Res. 43 (2004) 5300–5309. [9] J. Gilron, L. Song, K.K. Sirkar, Design for cascade of crossflow direct contact membrane distillation, Ind. Eng. Chem. Res. 46 (2007) 2324–2334. [10] M. Khayet, T. Matsuura, Preparation, Characterization of polyvinylidene fluoride membranes for membrane distillation, Ind. Eng. Chem. Res. 40 (2001) 5710–5718. [11] D.E. Suk, G. Pleizier, Y. Deslandes, T. Matsuura, Effects of surface modifying macromolecule (SMM) on the properties of polyethersulfone membranes, Desalination (2002) 303–307. [12] E. Drioli, Y. Wu, Membrane distillation: an experimental study, Desalination 53 (1985) 339–346. [13] M. Khayet, T. Matsuura, Application of surface modifying macromolecules for the preparation of membranes for membrane distillation, Desalination 158 (2003) 51–56. [14] M. Khayet, J.I. Mengual, T. Matsuura, Porous hydrophobic/hydrophilic composite membranes: application in desalination using direct contact membrane distillation, J. Membr. Sci. 252 (2005) 101–113. [15] M. Khayet, T. Matsuura, J.I. Mengual, Porous hydrophobic/hydrophilic composite membranes: estimation of the hydrophobic-layer thickness, J. Membr. Sci. 266 (2005) 68–79. [16] M. Qtaishat, M. Khayet, T. Matsuura, Novel porous composite hydrophobic/hydrophilic polysulfone membranes for desalination by direct contact membrane distillation, J. Membr. Sci. 341 (2009) 139–148. [17] Y. Kong, X. Lin, Y. Wu, J. Chen, J. Xu, Plasma polymerization of octafluorocyclobutane and hydrophobic microporous composite membranes for membrane distillation, J. Appl. Polym. Sci. 46 (1992) 191–199. [18] Y. Wu, Y. Kong, X. Lin, W. Liu, J. Xu, Surface-modified hydrophilic membranes in membrane distillation J, Membr. Sci. 72 (1992) 189–196. [19] B. Li, K.K. Sirkar, Novel membrane and device for vacuum membrane distillation-based desalination process, J. Membr. Sci. 257 (2005) 60–75. 175 [20] L.M. Song, B. Li, K.K. Sirkar, J.L. Gilron, Direct contact membrane distillationbased desalination: novel membranes, devices, larger-scale studies, and a model, Ind. Eng. Chem. Res. 46 (2007) 2307–2323. [21] H. Lee, F. He, L. Song, J. Gilron, K.K. Sirkar, Desalination with a sacade of cross flow hollow fiber membrane distillation devices integrated with a heat exchanger, AIChE J. 57 (2011) 1780–1794. [22] H. Yasuda, M. Gazicki, Biomedical applications of plasma polymerization and plasma treatment of polymer surfaces, Biomaterials 3 (1982) 68–77. [23] E.-G. Schlosser, Plasma polymerization, Angew. Chem. 99 (1985), 284–284. [24] F. Poncin-Epaillard, B. Chevet, J.-C. Brosse, Functionalization of polypropylene by a microwave (433 MHz) cold plasma of carbon dioxide. Surface modification or surface degradation? Eur. Polym. J. 26 (1990) 333–339. [25] M.B. Olde Riekerink, J.G.A. Terlingen, G.H.M. Engbers, J. Feijen, Selective etching of semicrystalline polymers: CF4 gas plasma treatment of poly(ethylene), Langmuir 15 (1999) 4847–4856. [26] K.R. Kull, M.L. Steen, E.R. Fisher, Surface modification with nitrogen-containing plasmas to produce hydrophilic, low-fouling membranes, J. Membr. Sci. 246 (2005) 203–215. [27] H.-Y. Yu, Y.-J. Xie, M.-X. Hu, J.-L. Wang, S.-Y. Wang, Z.-K. Xu, Surface modification of polypropylene microporous membrane to improve its antifouling property in MBR: CO2 plasma treatment, J. Membr. Sci. 254 (2005) 219–227. [28] N. Inagaki, N. Kobayashi, M. Matsushima, Gas separation membranes made by plasma polymerization of perfluorobenzene/CF4 and pentafluorobenzene/cf4 mixtures, J. Membr. Sci. 38 (1988) 85–95. [29] S. Borisov, V.S. Khotimsky, A.I. Rebrov, S.V. Rykov, D.I. Slovetsky, Y.M. Pashunin, Plasma fluorination of organosilicon polymeric films for gas separation applications, J. Membr. Sci. 125 (1997) 319–329. [30] K.S. Houston, D.H. Weinkauf, F.F. Stewart, Gas transport characteristics of plasma treated poly(dimethylsiloxane) and polyphosphazene membrane materials, J. Membr. Sci. 205 (2002) 103–112. [31] C.L. Chapman, D. Bhattacharyya, R.C. Eberhart, R.B. Timmons, C.-J. Chuong, Plasma polymer thin film depositions to regulate gas permeability through nanoporous track etched membranes, J. Membr. Sci. 318 (2008) 137–144. [32] G. Obuskovic, K.K. Sirkar, Liquid membrane-based CO2 reduction in a breathing apparatus, J. Membr. Sci. 389 (2012) 424–434. [33] G. Clarotti, F. Schue, J. Sledz, K.E. Geckeler, W. Göpel, A. Orsetti, Plasma deposition of thin fluorocarbon films for increased membrane hemocompatibility, J. Membr. Sci. 61 (1991) 289–301. [34] G. Li, X. Wei, W. Wang, T. He, X. Li, Modification of unsaturated polyester resins (UP) and reinforced UP resins via plasma treatment, Appl. Surf. Sci. 257 (2010) 290–295. [35] J.H. Zhang, J.D. Li, M. Duke, Z.L. Xie, S. Gray, Performance of asymmetric hollow fibre membranes in membrane distillation under various configurations and vacuum enhancement, J. Membr. Sci. 362 (2010) 517–528. [36] Y. Liu, G.H. Koops, H. Strathmann, Characterization of morphology controlled polyethersulfone hollow fiber membranes by the addition of polyethylene glycol to the dope and bore liquid solution, J. Membr. Sci. 223 (2003) 187–199. [37] X.M. Li, D. Reinhoudt, M. Crego-Calama, What do we need for a superhydrophobic surface? A review on the recent progress in the preparation of superhydrophobic surfaces, Chem. Soc. Rev. 36 (2007), 1529-1529. [38] B.A. Cruden, M. Rao, S.P. Sharma, M. Meyyappan, Fourier-transform infrared and optical emission spectroscopy of CF4 /O-2/Ar mixtures in an inductively coupled plasma, J. Appl. Phys. 93 (2003) 5053–5062. [39] L. Martinez-Diez, F.J. Florido-Diaz, M.I. Vazquez-Gonzalez, Study of evaporation efficiency in membrane distillation, Desalination 126 (1999) 193–198. [40] S. Bonyadi, T.S. Chung, R. Rajagopalan, A novel approach to fabricate macrovoidfree and highly permeable PVDF hollow fiber membranes for membrane distillation, AIChE J. 55 (2009) 828–833. [41] J.H. Zhang, N. Dow, M. Duke, E. Ostarcevic, J.D. Li, S. Gray, Identification of material and physical features of membrane distillation membranes for high performance desalination, J. Membr. Sci. 349 (2010) 295–303. [42] K.Y. Wang, S.W. Foo, T.S. Chung, Mixed matrix PVDF hollow fiber membranes with nanoscale pores for desalination through direct contact membrane distillation, Ind. Eng. Chem. Res. 48 (2009) 4474–4483. [43] J. Phattaranawik, R. Jiraratananon, A.G. Fane, Heat transport and membrane distillation coefficients in direct contact membrane distillation, J. Membr. Sci. 212 (2003) 177–193. [44] M. Gryta, Influence of polypropylene membrane surface porosity on the performance of membrane distillation process, J. Membr. Sci. 287 (2007) 67–78. [45] D.Y. Hou, J. Wang, B.Q. Wang, Z.K. Luan, X.C. Sun, X.J. Ren, Fluoride removal from brackish groundwater by direct contact membrane distillation, Water Sci. Technol. 61 (2010) 3178–3187. [46] M. Gryta, M. Barancewicz, Influence of morphology of PVDF capillary membranes on the performance of direct contact membrane distillation, J. Membr. Sci. 358 (2010) 158–167. [47] S. Bonyadi, T.S. Chung, Flux enhancement in membrane distillation by fabrication of dual layer hydrophilic–hydrophobic hollow fiber membranes, J. Membr. Sci. 306 (2007) 134–146. [48] X. Yang, R. Wang, L. Shi, A.G. Fane, M. Debowski, Performance improvement of PVDF hollow fiber-based membrane distillation process, J. Membr. Sci. 369 (2011) 437–447.
© Copyright 2026 Paperzz