Preparation and characterization of polyamide thin-film composite (TFC) membranes on plasma-modified polyvinylidene fluoride (PVDF)

Preparation and characterization of polyamide thin-film composite (TFC) membranes on plasma-modified polyvinylidene fluoride (PVDF)

Journal of Membrane Science 344 (2009) 71–81 Contents lists available at ScienceDirect Journal of Membrane Science journal homepage:

2MB Sizes 3 Downloads 85 Views

Journal of Membrane Science 344 (2009) 71–81

Contents lists available at ScienceDirect

Journal of Membrane Science journal homepage:

Preparation and characterization of polyamide thin-film composite (TFC) membranes on plasma-modified polyvinylidene fluoride (PVDF) Eun-Sik Kim a , Young Jo Kim b , Qingsong Yu b , Baolin Deng a,∗ a b

Department of Civil & Environmental Engineering, University of Missouri, Columbia, MO 65211, USA Center for Surface Science and Plasma Technology, Department of Mechanical & Aerospace Engineering, University of Missouri, Columbia, MO 65211, USA

a r t i c l e

i n f o

Article history: Received 24 February 2009 Received in revised form 19 June 2009 Accepted 23 July 2009 Available online 6 August 2009 Keywords: Low-temperature plasmas Surface modification Polyvinylidene fluoride Thin-film composite membrane Support layer Water permeability

a b s t r a c t The use of polyvinylidene fluoride (PVDF) membrane as a support substrate layer for making a polyamide thin-film composite (TFC) has been hindered by the hydrophobic nature of PVDF. In this study, low temperature plasmas were used to treat a commercial PVDF membrane in order to improve its hydrophilicity and make it suitable for fabrication of a polyamide–polyvinylidene fluoride thin-film composite (PAPVDF) membrane. The plasma treatment used oxygen, methane, or their 1:1 mixture gas to produce glow discharge for treatment and surface modification of the PVDF membrane. The membrane surfaces were characterized using water contact angle measurements, X-ray photoelectron spectroscopy (XPS), and field emission scanning electron microscopy (FE SEM). Pure water permeability and salt rejection tests were performed to evaluate the performance of the resulting PA-PVDF. Our experimental results indicated that plasma treatment of PVDF membrane could significantly reduce its water contact angle and thus produce a hydrophilic surface that could be then used as the support substrate layer for the fabrication of TFC membrane through interfacial polymerization. Our experimental results also demonstrated that PA-PVDF membrane supports not only demonstrated higher pure water permeability, but also higher salt rejections than those of TFC membrane with polysulfone as the support. © 2009 Elsevier B.V. All rights reserved.

1. Introduction Membranes have been widely used for solid/liquid separation and removal of soluble pollutants [1]. The use of membranes has particularly increased in water and wastewater treatment fields because of more stringent government regulations. To obtain optimal operation of a filtration system, a membrane must possess a high water flux and have good rejection of the targeted materials, minimum fouling, and high physical and chemical stabilities [2,3]. Good nanofiltration (NF) and reverse osmosis (RO) membranes can largely meet these criteria and are now recognized as the best technologies for water management [4]. These systems are not only important for traditional desalination but also for ultrapure water production and water contamination controls [5]. NF/RO membranes are mostly composed of thin-film composites (TFC) with an extremely thin surface active layer (thin-layer) on top of a much thicker, porous support substrate layer (support layer). The thin-layer performs the essential function of NF/RO process through control of solubilization and diffusion of solutes and water. Several polymers are good candidates for the thin-layer, including polyamide, polyester amide, polyethylene and polysilane

∗ Corresponding author. Tel.: +1 573 882 0075; fax: +1 573 882 4784. E-mail addresses: [email protected], [email protected] (B. Deng). 0376-7388/$ – see front matter © 2009 Elsevier B.V. All rights reserved. doi:10.1016/j.memsci.2009.07.036

[6–9]. Among these, polyamide (PA) is commonly used because of its good separation and permeation characteristics. Candidates for the TFC support layer include polysulfone (PSU)-based ultrafiltration (UF) or microfiltration (MF) membranes, which act as porous substrates to permeate the feeding solutions and provide mechanical strength. Polysulfone (PSU) has good thermal and decent chemical stabilities (tolerant to a wide range of pH values) and PSU-based TFC can be easily fabricated because of its relatively high hydrophilicity, which is suitable for the interfacial polymerization of PA in aqueous solutions. Recently, extensive studies have been reported on NF/RO membrane performances [10,11], structures and separation mechanisms [12,13], conditions of membrane fouling [14,15], polymer types of active layers [16,17], pore size distributions and transport parameters [18], and solute–solution interactions and enhancements of performance or other properties [19]. In these studies, almost all the researchers used PSU as the support layer of composites [8,12,20,21]. The support substrate layer based on PSU, however, has relatively low mechanical strength and poor resistance to chemicals such as aromatic hydrocarbons, ketones, ethers, and esters [1]. The pressure limit of PSU is below 100 psi when it is in the flat sheet form and 25 psi as hollow fiber membranes. Moreover, PSU raw materials are quite flexible and thus, it is difficult to hold them in a specific apparatus. From these weaknesses, there is clearly a need to explore potential replacements that are superior to PSU


E.-S. Kim et al. / Journal of Membrane Science 344 (2009) 71–81

Fig. 1. Schematic diagram of experimental process of plasma treatment and interfacial polymerization.

microfiltration membranes as support layers for TFC membrane fabrication. Fluorinated polymers are used widely in various fields because of their chemical, thermal, and mechanical stabilities, as well as toughness and resistance to corrosion [22]. For example, polyvinylidene fluoride (PVDF) is extensively used in building components, metal finishing binders, and other membrane applications. For these reasons, it is considered an attractive candidate to replace PSU as support layers of TFC membranes. One of the problems of using PVDF as support layers is its high surface hydrophobicity, which makes it difficult to adhere with other materials on the surface, thus improper for TFC membrane fabrication without modification through the aqueous interfacial polymerization process. The objective of this study was to investigate whether the PVDFbased microfiltration membrane could be modified to make it suitable for polyamide TFC membrane preparation. Common techniques for surface modification include chemical etching, plasma treatment and polymerization, and physical transition methods such as ultraviolet (UV), gamma (␥) ray or electron beam irradiations [23]. Among these methods, plasma treatment has numerous advantages over the other methods including rapidity, simplicity, and low investment expenses [24,25]. Additionally, plasma treatment does not affect the bulk properties of materials. In this study, plasmas of oxygen, methane, or their 1:1 mixture gases were used to treat PVDF surfaces for different time periods; and conventional interfacial polymerization processes were applied to form thin film on support layers. The experimental schematic is shown in Fig. 1. Surfaces were characterized using surface contact angle measurements and water uptake experiments. TFC membranes were then fabricated using unmodified and plasma-modified PVDF supports, and subsequently evaluated using pure water permeability and salt rejection tests. Field emission scanning electron microscope (FE SEM) and X-ray photoelectron spectroscopy (XPS) were also

used to characterize the chemical and physical properties of the membranes. 2. Experimental 2.1. Materials To fabricate and to compare the performances of TFC membranes, two types of commercial microporous (MF) membranes, including flat sheet PVDF membranes (pore size: 0.45 ␮m, GE Osmonics) and PSU flat sheet membranes (30,000 molecular weight cut-off (MWCO), Sterlitech Corp.), were used in this study. Both PSU and PVDF membranes were cut into a circular shape with a diameter of 47 mm for plasma treatment, surface characterization, TFC membrane preparation, and membrane performance tests. 1,3,5-Benzenetricarbonyl trichloride (TMC, >98%, Aldrich) and mphenylene diamine (MPD, >99%, Fluka) were used for interfacial polymerization to fabricate TFC membranes. Deionized (DI) water produced by Millipore DI system (Synergy 185, 18.2 M) was used in the experiments and all the chemicals used in this study were ACS reagents grade. 2.2. Plasma modification The PVDF and PSU membranes were cleaned with ethanol (Fisher) in a sonicator for 5 min, then washed with DI water and dried in desiccators for 24 h. A schematic illustration of the plasma reactor is shown in Fig. 2. Plasma modification was performed in a Pyrex bell-jar plasma reactor (volume; 80 L), in which two internal electrodes were arranged with 105 mm apart. An aluminum plate positioned in the middle of the two electrodes was used as the sample holder, on which the membranes were attached. Mechanical and booster pumps were used to evacuate the reactor chamber to a

E.-S. Kim et al. / Journal of Membrane Science 344 (2009) 71–81


Fig. 2. Experimental set-up of plasma treatment system for membrane modifications.

base pressure of less than 3 mTorr. Reactive gases were then introduced through the mass flow controller (MKS Instruments, Inc.) into the plasma reactor to a preset pressure. A mass flow controller and gas channel control module were used to calibrate the gas flow rates and ratio of gas mixtures. Plasmas were then initiated and sustained with a radio frequency (RF) power generator (13.56 MHz, RFX-600, Advanced Energy Industries, Inc.). Plasma treatment of the membranes was performed for a predetermined time interval (irradiating time was 10–180 s). A 10 W of plasma power was chosen to prevent ion etching effects in the discharges [26]. 2.3. PA-PVDF membrane preparation Prior to performing interfacial polymerization, the untreated and plasma-treated membrane substrates were immersed in an aqueous casting solution containing 2.0 wt.% of MPD monomers for 3 min, then a foam roller was used to remove excess solution on the surface. The membrane substrates absorbed with MPD were then immediately immersed in an organic solution containing 0.2 wt.% TMC in n-hexane (>98.5%, Sigma–Aldrich) for interfacial polymerization. TMC solution was spread over the amine-impregnated substrate for 3 min before the extra hexane solution was drained off. The material was washed with pure hexane again, rinsed with DI water, and finally dried at 110 ◦ C in an oven for 3 min. The prepared PA-PVDF membranes thus prepared were further dried at room temperature in ambient air for 24 h before membrane evaluation tests. 2.4. Characterization of the membranes 2.4.1. Contact angle measurements and uptake tests The contact angles of all membrane samples were measured at room temperature (23 ± 1 ◦ C). Contact angle measurements were performed by the sessile drop method [27] using DI water and a micro syringe. Images of the water droplets on membrane surfaces were captured using a computer-controlled video capturing system (Video Contact Angle System, ASC products 2500 XE). The membrane samples were placed on a vertically and horizontally adjustable sample stage. Directly above the sample stage was the tip of a syringe which was set to automatically dispense a predetermined amount of water by the computer. After the water droplet

was placed on the sample surface, a snapshot of the image was taken immediately. After placing markers around the perimeter of the water droplet, the VCA-2500 Dynamic/Windows software calculated right and left contact angles as well as droplet volume. At least 10 measurements at different sites of each membrane sample were measured with the measurement error range below ±2◦ . The uptake amount of the diamine solution is an important factor for the first step of interfacial polymerization (react with amine group in aqueous phase). Uptake tests were performed by two different methods to minimize experimental errors. First, the sample membranes were immersed in DI water or MPD DI water solution at room temperature for 2 h to reach saturation. Wet membranes were then measured for their wet weight by a digital balance and dried at 60 ◦ C in an oven for 24 h for dry weight measurements. Another method was to start with measuring the dry weight of the membrane samples. The measured dry membranes were then soaked in DI water or amine solutions for 2 h. After excess aqueous solution on the surfaces was removed by a foam roller, the membrane samples were weighed again, and the water or solution uptake was calculated by Eq. (1): U=

W − W  w d Wd

× 100


where U is the water uptakes ratio and Ww and Wd are the weights of wet and dry membranes, respectively. 2.4.2. Performances evaluation Pure water permeability was determined by direct volume measurements of the collected permeate at different time intervals. It was standardized using two types of filtration systems: one is suitable for measuring of MF and UF membranes (low-pressure system), and another is appropriate for NF and RO membrane standards (high-pressure system). Because MF and UF membranes were permeated under relatively lower pressure than NF and RO membranes, MF permeability in a high-pressure system was detected with a wider error range from our pre-experiments. In this study, different types of membranes (with different structures before and after the interfacial polymerization process) were configured and evaluated by both systems. A low-pressure system was used for support layers before interfacial polymerization and a highpressure system was applied to TFC membranes. Experimental


E.-S. Kim et al. / Journal of Membrane Science 344 (2009) 71–81

Fig. 3. Experimental filtration set-ups for membrane testing. (a) MF and UF membrane set-up, low-pressure system (stable under 5–60 psi) and (b) NF and RO filtration set-up, high-pressure system (stable under 100–800 psi).

filtration set-ups are described in Fig. 3. The set-up for the lowpressure system was simple and contained a dead-end membrane holder. It was equipped with a gas cylinder (nitrogen, >99%) to support the pressure of the 4-L dispensing pressure vessel (Millipore Corp.) which provided feed water pressure constantly. The membrane holder (Stirred cell 8200, Millipore Corp.) had an effective membrane area of 28.7 cm2 and 200 ml water solution capacity. The high-pressure system was composed of a dead-end filtration module with an 8-L feed solution tank and a circular high pressure filter holder (Model: XX4504700, effective membrane area, 9.6 cm2 , Stainless steel, Millipore Corp.) with dimensions of 8.6 cm in diameter, 4.4 cm in height, and a maximum inlet pressure of 10,000 psi. In this high-pressure system, a Hydra-cell® pump (M-03 series, Wanner Engineering, Inc.) with a digital speed controller (PowerFlex 4 AC Drive, Rockwell Automation), a pulsation dampener (Model: 110-260, Stainless steel construction, Wanner Engineering, Inc.), and a pressure gauge (Swagelok) were also included. Prior to the permeability measurements, pure water from the feed tank was pressurized through the membranes for at least 1 h in order to stabilize the membranes. Membrane samples were placed in the filter holder with the thin-layer facing the incoming feed solution. Feed

pressure was mainly controlled through a pump speed controller and the detailed input pressure was also calibrated using a bypass valve (C46 types, Wanner Engineering, Inc.). Permeate weights were monitored continuously by an electrical balance (Scout Pro, Ohaus) interfaced with PC (LabView 8.2, National Instruments). Compressed air was supplied to maintain constant system pressure by a pulsation dampener. At least three tests were performed for each membrane sample under different pressure ranges. The average value was calculated and linear regression method was applied to calculate flow rates. Applied pressure was 10–50 psi for the lowpressure system and 100–400 psi for the high-pressure system at room temperature. The following equation was used to calculate water permeability in terms of liter per square meter per hour per unit pressure: LP =

Vp AtP


where LP is the hydraulic permeability, Vp is the permeate volume, A is the effective membrane area, t is the time, and P is the applied pressure. The salt rejection tests were performed by measuring the electrical conductance of the permeate and feed solutions. The salt

E.-S. Kim et al. / Journal of Membrane Science 344 (2009) 71–81


rejection tests for TFC membranes were conducted at 300 psi using a NaCl (Fisher) solution with initial concentration of 2000 ppm at room temperature in a feed solution tank. Prior to the initiation of a salt rejection test, conductivities of the solutions were measured. The membrane sample placed in a membrane holder was first washed with DI water at 150 psi for 1 h to remove surface impurities and compaction of membrane. A conductivity/TDS meter (HACH Company) was used to measure the conductance in feed and permeate NaCl solution, and their differences were used to calculate the rejection ratio with the following equation: R=


Cp Cf

× 100


where R is the rejection ratio, Cp and Cf are the conductances of permeate and feed, respectively. 2.4.3. Field emission scanning electron microscopy (FE SEM) and X-ray photoelectron spectroscopy (XPS) The physical morphology changes of the membrane surface upon plasma treatment were examined by field emission scanning electron microscope (FE SEM) using a Hitachi S-4700 FE SEM. All samples were cut into pieces less than 50 mm2 and attached to aluminum holders using carbon tape, then coated with platinum prior to FE SEM examination. Chemical compositions of the PVDF membranes were investigated using X-ray photoelectron spectroscopy (XPS, Kratos Axis 165 Photoelectron Spectrometer). The XPS detector used was an unmonochromated Mg source and spectra were obtained in hybrid lens mode with pass energy 80 eV for survey scans and 20 eV for window scans. CasaXPS (computer aided surface analysis for XPS, Ver 2.3.13) software was used for high resolution peak analysis and atomic proportions as well as curve-fitted modeling for core-level spectra. 3. Results and discussion 3.1. Membrane modification and characterization The surface contact angle changes via different plasma treatment times are shown in Fig. 4. Water contact angles were used to assess the surface hydrophilic/hydrophobic properties of membranes. Plasma modifications were performed with different types of plasma sources as a function of the plasma exposure time. Three types of plasma gases, including oxygen (O2 ), methane (CH4 ), and their 1:1 ratio mixture, were used for membrane surface modification. From Fig. 4, it can be seen that, initially, within 60 s plasma exposure, rapid reduction of water contact angles was observed with all of the membrane samples; after 60 s exposure, the water contact angle started to level off. The initial water contact angle of the untreated PVDF was 119◦ , while plasma-modified PVDF was reduced to 52◦ when oxygen and methane mixture was used. In the case of PSU, plasma treatment reduced its water contact angle from 80◦ to 29◦ . The maximum water contact angle decrease obtained for PVDF was 43.8%, and a maximum decrease of 35.5% was obtained for PSU. The lowest contact angle value obtained for PVDF upon plasma treatment was 52◦ , which was much lower, i.e., much more hydrophilic than that of unmodified PSU membrane, which had a water contact angle of 80◦ . The decreases in contact angles can be explained by the formation of hydrophilic moieties on membrane surfaces after the plasma treatment. It is known that the main difficulty in fabricating TFC membranes with PVDF as support layers is the hydrophobic properties of PVDF. The unmodified PVDF samples have a water contact angle of nearly 120◦ . Decreases in the water contact angle or increases in surface hydrophilicity are primarily controlled by the grafting of hydrophilic polar moieties on the surface through plasma modifications. Usually, oxygen plasma is very effective for enhancing

Fig. 4. Effects of plasma modifications on surface contact angles for PVDF membranes (solid lines) and PSU membranes (dashed lines). Plasma conditions: (1) oxygen, (2) methane, and (3) oxygen-methane mixture plasma, 50 mTorr and predetermined time intervals: 10, 30, 60, 120 and 180 s.

the hydrophilicity of many polymer surfaces due to the formation of oxygen-containing functional group and oxidations. However, when excessive oxygen plasma species exist on the surface, oxygen plasma-induced etching of the surface polymers can occur. Plasma etching reactions could chop off the polymer chain and thus degrade the polymer materials into oligomers to form weak boundary layers on the surface, which could result in surface contamination and adversely affect adhesion properties. In the case of plasma treatment using the oxygen and methane gas mixture, a significant improvement in surface hydrophilicity was obtained on PVDF membranes. At the same time, it was expected that using the oxygen/methane mixture in plasma treatment could minimize the undesired plasma etching effects that were often observed with oxygen plasmas. Water and amine-solution uptakes of plasma-modified membranes were defined in weight percentages of the dried membrane samples and weight loss of wetted membrane samples. Fig. 5 shows the percentages of water and amine-solution uptakes of the plasma-modified membrane samples with varying plasma exposure durations. Unmodified PVDF membranes had very low uptake percentages for both water and amine-solution, but uptake percentages were dramatically increased after plasma modification with pure oxygen and the oxygen/methane gas mixture. Methane-modified membranes still had low water uptake and slightly increased percentages of amine-solution uptake compared to water. Through the results, it can be noted that uptake percentage of the plasma-treated PVDF membrane samples initially showed very sharp increases with plasma treatment times up to 60 s before leveling off with longer plasma treatment times. This result is consistent with the change in water contact angle with plasma treatment time as shown in Fig. 4. The maximum uptakes of water and amine-solution in PVDF membrane samples were 61 and 62 wt.%, respectively, with 180 s (3 min) plasma treatment time using oxygen/methane gas mixture. Such an increase in water uptake in the plasma-treated PVDF membrane samples demonstrated the possibility of using PVDF membrane as support layers for TFC membrane fabrication. The uptake percentages for PSU membrane samples were higher than those of PVDF membrane samples. Because PSU is a natu-


E.-S. Kim et al. / Journal of Membrane Science 344 (2009) 71–81

Fig. 5. Uptakes of water (solid line with filled symbols) and amine-solution (dashed line with open symbols) by (a) PVDF and (b) PSU membranes under different plasma treatment times and gases types.

rally hydrophilic material, its uptake levels for aqueous solution were better than those of hydrophobic PVDF. As shown in Fig. 5, the water uptakes of PSU decreased after oxygen and methane plasma treatments. This behavior may be explained by the possible structure changes on PSU surfaces. Oxygen plasma can change surface morphology, making defects and porous structures by the etching process, so surface structure changes as well as uptake ability. Plasma treatment on PSU membranes usually results in the extension of pore sizes of the membranes [28]. Even if their hydrophilic properties are enhanced by plasma modification, the uptakes of solution are affected more by the structural changes. With plasma-treated membranes using oxygen/methane mixture plasmas, maximum uptakes of water and amine-solution of 76% and 78% were obtained after plasma modifications, respectively. These results indicate a 34% increase in uptake for the plasmatreated PSU membrane samples compared to the unmodified ones. XPS survey spectra for unmodified and plasma-modified PVDF are shown in Fig. 6. The fluorine (F 1s), oxygen (O 1s) and carbon (C 1s) envelopes were chosen as major peaks because these chemical compositions were directly correlated with PVDF polymer structure and plasma sources applied in this experiment. Untreated PVDF showed high F 1s and C 1s peaks, but they contained small amounts of O 1s (less than 1 mol%), which was based on the origin of PVDF composition. Plasma treatments by pure oxygen and the oxygen/methane mixture showed significant increases in the O 1s peak, indicating the incorporation of oxygen-containing moieties, while the F 1s and C 1s peak intensities decreased. This means that an increase in oxygen concentration on the PVDF membrane surface corresponds to a decrease in fluoride concentration. Treatment with pure oxygen plasma increased surface oxygen functionality, but the relative amount was lower when the oxygen/methane

Fig. 6. XPS survey spectra of unmodified and plasma-modified PVDF membrane. All samples were plasma treated by identical experimental condition: 50 mTorr and 180 s treatments.

gas mixture was used. Treatment with pure methane plasma did not provide oxygen moieties but significantly increased the relative carbon contents and reduced fluorine contents on the surface. This could be due to plasma polymer deposition on the membrane surface in pure methane plasmas. Table 1 summarizes the peak positions and atomic percentages that were calculated for unmodified and plasma-modified PVDF membranes. The highest O/C ratio was obtained when the oxygen/methane mixture plasma was used. The chemical components that corresponded to different plasmas in C 1s and O 1s peaks were analyzed by core-level spectra. Fig. 7(a)–(d) shows the core-level spectra of C 1s and O 1s in unmodified and variously plasma-modified PVDF membranes. Pristine PVDF consisted of two main peak components in C 1s because the chemical components of PVDF were mainly composed with 286.4 eV for CH2 species and 290.9 eV for CF2 species with a 1:1 ratio [29]. The C 1s peak of oxygen plasma-treated PVDF was curvefitted with pristine CH2 , CF2 species and C–O and C O at 286.7 and 289.1 eV, respectively [29]. C–O and C O components were not detected in untreated PVDF, thus it might be considered that assigned from oxygen plasmas. Moreover, oxygen plasma changed the ratio of CH2 and CF2 species compared to untreated PVDF. Carbon from CH2 was decreased with treatment that oxygen plasma because it can be etched selectively CH2 components on the PVDF surface. Oxygen plasma was employed by O 1s envelope on the PVDF surface which analyzed two different functionalities, corresponded to O–C at 534.7 and O C at 532.1, respectively [29,30]. The two components of equal intensity were also assigned in mixture plasma-treated PVDF, but their relative mol% decreased with untreated PVDF. Three different functionalities, C–O, C O and C–H, were employed from mixture plasma; these were good indications

E.-S. Kim et al. / Journal of Membrane Science 344 (2009) 71–81


Table 1 Surface chemical composition by XPS analysis of plasma-modified PVDF membrane samples. Membrane types

Unmodified PVDF Oxygen PVDF Mixture gas PVDF Methane PVDF

F 1s

O 1s

C 1s


Position binding energy (eV)

Atomic conc. (%)

Position binding energy (eV)

Atomic conc. (%)

Position binding energy (eV)

Atomic conc. (%)

686.5 686.5 686.5 686.0

44.34 41.48 40.22 26.21

531.0 531.5 531.5 532.5

0.52 4.03 6.47 .27

286.0 286.5 286.5 286.0

55.15 54.49 53.31 72.51

0.009 0.074 0.121 0.018

Fig. 7. XPS core-level spectra of unmodified and plasma-modified PVDF membrane. Left side figures presented C 1s envelopes and right side figures were indicated O 1s envelopes. (a) Untreated, (b) oxygen plasma, (c) mixture gas plasma, and (d) methane plasma treated.


E.-S. Kim et al. / Journal of Membrane Science 344 (2009) 71–81

Fig. 8. Pure water flux of PVDF (a) and PSU (b) membranes with and without plasma modifications.

that mixture gas plasma-treated PVDF was highly modified and these functionalities were observed as grafting sites with oxygen. These results were proven by the relative mol% of O 1s envelope in survey peak, also it can be shown by the components in corelevel peak. Mixture plasma employed 62% of C O group and 38% of C–O groups by curve-fitted with components. Large amounts of carbonyl groups resulted in more hydrophilic surfaces then react with hydrophilic moiety. In these results, they have lower contact angles and performed well in interfacial polymerization processes. Methane plasma treatment caused CF2 :CH2 ratio changes where CH2 intensity was two times that of CF2 peaks. This indicated that carbon functionalities were deposited on PVDF surfaces by methane plasma. A very small oxygen peak was detected for the methane plasma treatment, but it was not a major factor to change the chemical composition on hydrophobic PVDF. 3.2. Membrane permeability and salts rejections Prior to fabricating composite membranes, the plasma-modified and unmodified membrane support layers were first characterized with pure water permeability measurements. Fig. 8 shows the water permeability data obtained with plasma-modified and unmodified PSU and PVDF membranes. Based on the results of contact angle measurements and water and amine-solution uptakes tests, plasma-modified membrane samples with 3 min of plasma exposure were selected for permeability measurements. PSU samples were measured under pressures from 10 to 50 psi and PVDF samples were measured under pressures from 10 to 30 psi. Because these membranes have much larger pores than usual UF membranes, a low-pressure system with a maximum pressure of 50 psi can be used for their permeability measurements. All the unmodified membrane samples showed linear dependence of water

permeability on the applied pressures. In comparison, all the plasma-modified PVDF samples showed a slightly enhanced permeability for pure water. The plasma-modified PSU samples had similar performances to those of the unmodified control samples. The slight increase in water permeability of modified PVDF membranes could be due to the surface morphology changes induced by plasma treatment [31]. Both plasma-modified and unmodified microporous membranes of PVDF and PSU were used as support substrate layers individually for TFC membrane fabrication through interfacial polymerization. The membranes were then evaluated by permeability measurements on a high-pressure system. Pure water permeability and salt rejections for the prepared TFC membranes are depicted in Fig. 9. Please note that the data points with a plasma treatment time of 0 s are for the TFC membranes prepared with the unmodified PSU or PVDF membrane as support layers. The results showed that TFC membranes prepared on unmodified PVDF membranes had a very high permeability (∼2.4 L/(m2 h psi)) and low salt rejection as tested in the high-pressure system. It appeared that the unmodified PVDF membranes were too hydrophobic to react with diamine-aqueous solution and thus the subsequent interfacial polymerization could not occur on the unmodified PVDF surface. It is clear from Fig. 9 that all of the plasma-modified TFC samples showed an increase in salt rejections with increasing plasma treatment time for the support layers. Permeability of pure water, however, decreased with increased plasma treatment time of the support layers. These inverse trends between water permeability and salt rejections were similarly observed on typical RO membranes and are consistent with plasma-modified membrane performances in general. In the case of PSU as support layers, plasma modification definitely improved membrane performances as compared to the TFC membrane on unmodified PSU. The best results were achieved with plasma treatment using the oxygen/methane (1:1) gas mixture with 180 s plasma exposure for both the PVDF and PSU membranes as support layers. It is worth pointing out that the TFC membranes prepared on PVDF support layers showed higher salt rejection and water permeability than those prepared using PSU support layers. This result indicates that plasma treatment using oxygen/methane (1:1) gas mixture provided PVDF support layers with a suitable surface for diamine absorption and subsequent interfacial polymerization to form the TFC. Moreover, oxygen-containing moieties induced by plasma treatment not only resulted in improved surface hydrophilicity but also provided the reactive sites on PVDF surfaces to chemically bond with diamine, which joins the interfacial polymerization to form TFC with excellent adhesion the PVDF support layers [31]. The TFC membrane prepared on the oxygen plasma-treated PVDF support layers also had better permeability in accordance with the change in hydrophilicity. In comparison, TFC membranes fabricated on methane plasma-treated PVDF showed a higher permeability but lower salt rejection. This is likely due to the poor interactions between the support layer and diamine-aqueous solution, and consequently, a lesser formation of thin film on the surface during interfacial polymerization. This result is also consistent with the amine-solution uptake test results. Fig. 10 shows the SEM images of the unmodified and plasma-modified PVDF membrane surfaces (a and b) using the oxygen/methane (1:1) gas mixture with 3 min of plasma treatment and TFC coatings (c and d) prepared on the unmodified and plasmamodified PVDF membranes, respectively. The mean pore size and surface morphology of the plasma-modified PVDF membrane surface were similar to those of the unmodified controls. However, many of the polymer fibers across the pore found on unmodified PVDF membranes, as shown in Fig. 10(a), broke down and disappeared upon plasma treatment. Plasma treatment apparently caused surface changes on the polymer membranes [32]. Fig. 10(c

E.-S. Kim et al. / Journal of Membrane Science 344 (2009) 71–81


Fig. 9. Pure water permeability (left) and salts rejection (right) as a function of plasma treatment time for PSU (open circles) and PVDF (filled circles). (a) Oxygen plasma, (b) mixture gas plasma, and (c) methane plasma.

and d) are the SEM images of PA thin-layers formed on the PVDF supported surface. With the unmodified PA-PVDF (Fig. 10(c)), pore sizes were reduced after the TFC coating, but the membrane still showed open-pore structures. This explains the high permeability and low salt rejection of the membrane, because the hydrophobic nature of the PVDF is unsuitable for diamine interfacial polymerization. In contrast, TFC formed on the plasma-modified PVDF support layer showed significantly more coating materials (Fig. 10(d)). All the pores on the PVDF support surfaces were covered with a well-established thin-layer of PA, which functioned in salt rejection the same as RO membranes. Fig. 11 shows the cross-sectional SEM images for the progression of interfacial poly-

merization of PA-PVDF membranes. Unmodified PVDF (Fig. 11(a)) showed a bulk property of polymer membrane, and their porous structures were disordered in the matrix. Plasma-modified PVDF (Fig. 11(b)) exhibited a similar structure to unmodified membrane. Well-established PA-PVDF (Fig. 11(c)) showed the structure that polyamide thin-film layered on the plasma-modified PVDF membrane. The thickness of thin film was approximately 2.8–3.0 ␮m, and the whole membrane thickness was 138.0–142.0 ␮m. According to the results, mixture gas plasma-modified PVDF could be a support material for TFC membrane and the well-established PA-PVDF could improve salt rejection rates and enhance water permeability.


E.-S. Kim et al. / Journal of Membrane Science 344 (2009) 71–81

Fig. 10. Plane SEM micrographs of membranes. (a) Unmodified PVDF, (b) plasma-modified PVDF (3 min, mixture gas), (c) TFC using unmodified PVDF, and (d) TFC using plasma-modified PVDF (b). All scales were represented in 1 ␮m.

Fig. 11. Cross-sectional SEM micrographs of plasma-modified interfacial polymerization process. (a) Unmodified PVDF, (b) plasma-modified PVDF, and (c) PA-PVDF.

4. Conclusions To date, PVDF membrane has not been used as a support layer for TFC membrane because of its hydrophobic nature. This study has demonstrated that the surface hydrophilic/hydrophobic properties of PVDF membrane can be tailored with the appropriate plasma modification and thus can be used for fabricating TFC membranes by the conventional interfacial polymerization method. XPS and SEM analyses of the unmodified and plasma-modified PVDF membranes indicate that plasma treatment can induce both chemical and physical changes on the membrane surfaces. In general, plasma modification of PVDF membranes enhances membrane water uptakes and hydrophilic properties on the surface, but the oxygen/methane (1:1) gas mixture is most effective. With the plasma-modified PVDF support, a well-established PA thin-layer forms which contributes to the elevated salt rejection rates while maintaining high water permeability. The results suggest that plasma treatment is an effective tool for membrane surface mod-

ification in terms of tailoring their surface chemical and physical properties including surface morphology, hydrophilicity, and suitability as support layers for TFC membrane fabrication through interfacial polymerization.

Acknowledgements We greatly appreciate the analytical support from Mr. Louis Ross Jr. at the University of Missouri (SEM) and Mr. Jeff Wight at Missouri University of Science and Technology (XPS). Financial support for this research was provided by the U.S. National Science Foundation (No. BES-0296109).

References [1] R.W. Baker (Ed.), Membrane Technology and Applications, 2nd ed., John Wiley & Sons Ltd., West Sussex, England, 2004.

E.-S. Kim et al. / Journal of Membrane Science 344 (2009) 71–81 [2] N. Park, S. Lee, S.R. Yoon, Y.H. Kim, J. Cho, Foulants analyses for NF membranes with different feed waters: coagulation/sedimentation and sand filtration treated waters, Desalination 1–3 (2007) 231. [3] L.-F. Liu, S.-C. Yu, Y. Zhou, C.-J. Gao, Study on a novel polyamide–urea reverse osmosis composite membrane (ICIC-MPD). I. Preparation and characterization of ICIC-MPD membrane, J. Membr. Sci. 1–2 (2006) 88. [4] S. Lee, J. Cho, M. Elimelech, Influence of colloidal fouling and feed water recovery on salt rejection of RO and NF membranes, Desalination 1 (2004) 1. [5] Y. Zhou, S. Yu, M. Liu, C. Gao, Preparation and characterization of polyamide–urethane thin-film composite membranes, Desalination 1–3 (2005) 189. [6] B. Mottershead, S.J. Eichhorn, Deformation micromechanics of model regenerated cellulose fibre–epoxy/polyester composites, Compos. Sci. Technol. 10 (2007) 2150. [7] Y. Zhiyong, E. Mielczarski, J. Mielczarski, D. Laub, P. Buffat, U. Klehm, P. Albers, K. Lee, A. Kulik, L. Kiwi-Minsker, A. Renken, J. Kiwi, Preparation, stabilization and characterization of TiO2 on thin polyethylene films (LDPE). Photocatalytic applications, Water Res. 4 (2007) 862. [8] U. Razdan, S.S. Kulkarni, Nanofiltration thin-film-composite polyesteramide membranes based on bulky diols, Desalination 1 (2004) 25. [9] M. Ando, Y. Ohsawa, H. Naito, Y. Kanemitsu, Photoconduction of (silicon nanocrystals)–(organic polysilane) composites, J. Organomet. Chem. 1–2 (2003) 243. [10] Z. Yong, Y. Sanchuan, L. Meihong, G. Congjie, Polyamide thin film composite membrane prepared from m-phenylenediamine and m-phenylenediamine-5sulfonic acid, J. Membr. Sci. 1–2 (2006) 162. [11] I.J. Roh, J.-J. Kim, S.Y. Park, Mechanical properties and reverse osmosis performance of interfacially polymerized polyamide thin films, J. Membr. Sci. 1–2 (2002) 199. [12] Y. Song, P. Sun, L.L. Henry, B. Sun, Mechanisms of structure and performance controlled thin film composite membrane formation via interfacial polymerization process, J. Membr. Sci. 1–2 (2005) 67. [13] P.S. Singh, S.V. Joshi, J.J. Trivedi, C.V. Devmurari, A.P. Rao, P.K. Ghosh, Probing the structural variations of thin film composite RO membranes obtained by coating polyamide over polysulfone membranes of different pore dimensions, J. Membr. Sci. 1–2 (2006) 19. [14] A. Asatekin, A. Menniti, S. Kang, M. Elimelech, E. Morgenroth, A.M. Mayes, Antifouling nanofiltration membranes for membrane bioreactors from selfassembling graft copolymers, J. Membr. Sci. 1–2 (2006) 81. [15] S.H. Kim, S.-Y. Kwak, B.-H. Sohn, T.H. Park, Design of TiO2 nanoparticle selfassembled aromatic polyamide thin-film-composite (TFC) membrane as an approach to solve biofouling problem, J. Membr. Sci. 1 (2003) 157. [16] M.M. Jayarani, S.S. Kulkarni, Thin-film composite poly(esteramide)-based membranes, Desalination 1 (2000) 17.


[17] A.L. Ahmad, B.S. Ooi, J.P. Choudhury, Preparation and characterization of co-polyamide thin film composite membrane from piperazine and 3,5diaminobenzoic acid, Desalination 1–3 (2003) 101. [18] R.R. Sharma, R. Agrawal, S. Chellam, Temperature effects on sieving characteristics of thin-film composite nanofiltration membranes: pore size distributions and transport parameters, J. Membr. Sci. 1–2 (2003) 69. [19] P. Xu, J.E. Drewes, T.-U. Kim, C. Bellona, G. Amy, Effect of membrane fouling on transport of organic contaminants in NF/RO membrane applications, J. Membr. Sci. 1–2 (2006) 165. [20] L. Liu, A. Chakma, X. Feng, Preparation of hollow fiber poly(ether block amide)/polysulfone composite membranes for separation of carbon dioxide from nitrogen, Chem. Eng. J. 1–2 (2004) 43. [21] A. Wahab Mohammad, N. Hilal, M. Nizam Abu Seman, A study on producing composite nanofiltration membranes with optimized properties, Desalination 1–3 (2003) 73. [22] A. Kelly, S.T. Mileiko, Fabrication of Composites, Elsevier Science Pub. Co., New York, 1983. [23] C. Elsner, M. Lenk, L. Prager, R. Mehnert, Windowless argon excimer source for surface modification, Appl. Surf. Sci. 10 (2006) 3616. [24] T.D. Tran, S. Mori, M. Suzuki, Plasma modification of polyacrylonitrile ultrafiltration membrane, Thin Solid Films 9 (2007) 4148. [25] D.E. Weibel, C. Vilani, A.C. Habert, C.A. Achete, Surface modification of polyurethane membranes using acrylic acid vapour plasma and its effects on the pervaporation processes, J. Membr. Sci. 1–2 (2007) 124. [26] S.A. Voronin, M.R. Alexander, J.W. Bradley, Time-resolved mass and energy spectral investigation of a pulsed polymerising plasma struck in acrylic acid, Surf. Coat. Technol. 3–4 (2006) 768. [27] H. Yasuda, Luminous Chemical Vapor Deposition and Interface Engineering, Marcel Dekker, New York, 2005. [28] M.L. Steen, L. Hymas, E.D. Havey, N.E. Capps, D.G. Castner, E.R. Fisher, Low temperature plasma treatment of asymmetric polysulfone membranes for permanent hydrophilic surface modification, J. Membr. Sci. 1 (2001) 97. [29] G. Beamson, D. Briggs, High Resolution XPS of Organic Polymers: The Scienta ESCA 300 Database, John Wiley & Sons, Chichester, 1992. [30] P. Bebin, E. Robert, Prud’homme X-ray photoelectron spectroscopy of miscible poly(methyl methacrylate)/poly(styrene-co-acrylonitrile) and immiscible poly(methyl methacrylate)/polyacrylonitrile polymer surfaces metallized by nickel, J. Polym. Sci. Part B: Polym. Phys. 8 (2004) 1408. [31] H.I. Kim, S.S. Kim, Plasma treatment of polypropylene and polysulfone supports for thin film composite reverse osmosis membrane, J. Membr. Sci. 1–2 (2006) 193. [32] J. Hopkins, J.P.S. Badyal, Plasma modification of poly(ether sulfone), Macromolecules 19 (1994) 5498.