- Email: [email protected]
Surface modification of a polyamide reverse osmosis membrane for chlorine resistance improvement
Journal of Membrane Science 415–416 (2012) 192–198
Contents lists available at SciVerse ScienceDirect
Journal of Membrane Science journal homepage: www.elsevier.com/locate/memsci
Surface modification of a polyamide reverse osmosis membrane for chlorine resistance improvement Young-Nam Kwon a, Sungpyo Hong b, Hyoungwoo Choi b, Taemoon Tak b,n a b
School of Urban and Environmental Engineering, Ulsan National Institute of Science and Technology (UNIST), 100 Banyeon-ri, Ulsan 689-798, South Korea Department of Biosystems and Biomaterials Science and Engineering, Seoul National University, 599 Gwanangno, Gwanak-gu, Seoul 151-921, South Korea
a r t i c l e i n f o
a b s t r a c t
Article history: Received 28 December 2011 Received in revised form 25 April 2012 Accepted 29 April 2012 Available online 22 May 2012
A surface modified polyamide (PA) thin film composite (TFC) membrane was prepared using in-situ polymerization of sorbitol polyglycidyl ether (SPGE) on the membrane surface immediately after interfacial polymerization of TFC membranes. This modification was conducted to protect the chlorine-sensitive sites of the PA membrane using a chlorine-resistant hydrophilic SPGE polymer. The optimum preparation condition of the modified PA TFC reverse osmosis membranes was investigated, and then the successful modification of the membrane was confirmed using various analytical tools including Fourier transform infrared spectroscopy, a zeta potential analyzer, a contact angle analyzer, X-ray photoelectron microscopy, scanning electron microscopy, and atomic force microscopy. The modification converted the surface of the membrane to a more neutral, hydrophilic, and smooth surface. With increasing SPGE concentration in the coating solution, molecular overlapping of the coating polymer led to a denser coating layer, which resulted in declined flux but increased salt rejection. Chlorination tests showed that the modification of the membrane using SPGE ring-opening polymerization improved chlorine stability. & 2012 Elsevier B.V. All rights reserved.
Keywords: Polyamide Surface modification Chlorine resistance Reverse osmosis
1. Introduction Reverse osmosis (RO) is a membrane-based desalination process in which relatively pure water is transported across the membrane by the pressure gradient between the feed and permeate water, while dissolved organics and inorganics are rejected due to the chemical nature of the membrane material and the physical structure of the membrane [1]. Since the development of composite polyamide (PA) membranes using interfacial condensation method by Cadotte [2], drastic progress in RO membrane technology has been made. However, widespread use of the membrane is limited due to the adsorption of retained organic and/or inorganic materials and the subsequent decline in performance [3]. Fouling, which is defined as the (ir)reversible deposition of retained materials, can be mitigated through pretreatment of feed water before the membrane filtration and minimized by a physical or chemical cleaning process after filtration. Among the techniques controlling membrane fouling, the addition of chlorine to the feed water is one of most convenient method and has been adopted in many RO applications for deactivating microorganisms causing membrane biofouling. Since the PA membrane is reported to be very sensitive to chlorine [4–8], reducing agents need to be used in the dechlorination process. Partial failure of the dechlorination
n
Corresponding author. Tel.: þ82 2 880 4621. E-mail addresses: [email protected] (Y.-N. Kwon), [email protected] (T. Tak).
0376-7388/$ - see front matter & 2012 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.memsci.2012.04.056
process can result in the deterioration of PA membranes, making the whole membrane useless. Various approaches have been devised to develop chlorineresistant membrane by eliminating chlorine-sensitive sites of the PA membrane or protecting the sensitive sites using chlorineresistant coating materials. Amide nitrogen and aromatic rings of the PA membrane are sites that can be easily attacked by chlorine. Kawaguch and Tamura [9] chlorinated secondary and tertiary amide to investigate chlorinating behavior in the presence or absence of amidic hydrogen and demonstrated that the amide nitrogen is a key chlorination reaction site. Uemura and Kurihara [10] observed that a cross-linked N-substituted PA TFC membrane is more chlorine tolerant compared to a typical cross-linked PA TFC membrane. Glater and Zachariah [11] investigated the effect of chlorine on a PA membrane and concluded that chlorine attacks on aromatic rings change the hydrogen-bonding forces. Chlorination of the aromatic ring in the PA membrane can be blocked by the substitution of a deactivating functional group such as –NO2 or –SO3H at the ring position [12]. Son and Jegal [13] developed TFC RO membranes using several different amines and trimesoyl chloride (TMC) and showed that the m-phenylene diamine (MPD) derivative with functional groups such as –CH3 and –OCH3 has better chlorine tolerance. Surface modification such as grafting or physical adsorption is a potential technique to improve chlorine resistance of commercially available PA membranes. Du and Zhao [14] prepared poly (N,N-dimethylaminoethyl methacrylate) (PDMAEMA) positively charged nanofiltration membranes using an interfacial crosslinking reaction and demonstrated that the membrane was stable
Y.-N. Kwon et al. / Journal of Membrane Science 415–416 (2012) 192–198
during 28-days storage in 5 ppm NaOCl preservatives. Based on the results of Du and Zhao, Kang et al. [15] coated PA RO membranes with the PDMAEMA and demonstrated that the coated membrane shows better chlorine resistance. Polyvinyl alcohol (PVA) is a hydrophilic and electric neutral polymer that has been applied to the RO membrane to improve hydrophilicity and, thus, antifouling properties. Furthermore, the PVA coating was reported to provide enhanced chlorine resistance to the membrane [16]. Considering long-term stability, the cross-linked structure of these coating materials is desirable. However, modification using a grafting or physical adsorption method requires another cross-linking step, which might be economically undesirable. Therefore, it is economically favorable to form the cross-linked coating using only one step. The objective of this study was to develop a novel method that is suitable for surface modification of an aromatic PA RO membrane to improve its chlorine tolerance. Cross-linked sorbitol polyglycidyl ether (SPGE) was polymerized on the PA RO membrane surface via in situ ring-opening polymerization. N,Ndimethylaminopropylamine (DMAP) and glycerol were used as a ring-opening agent and a humectant. The physicochemical properties of the membrane surface were analyzed using various analytical tools and compared before and after the coating process. Permeability and selectivity of the virgin and modified membranes were evaluated by filtration of a 2000 ppm NaCl aqueous solution at 25 1C under the operating pressure of 1.5 MPa. Exposures of the membranes to 100 ppm free chlorine solutions during various exposure times were carried out to evaluate membrane chlorine resistances.
2. Experimental materials and methods
193
organic solvent, Isol-C (SK Chemical Inc., Korea), was selected to dissolve the acid chloride and prepare the organic phase solution for the interfacial polymerization. The reaction time for the interfacial polymerization was 90 s. N,N-dimethylamino propylamine (DMAP) (Acros Organics, USA) and sorbitol polyglycidyl ether (SPGE) (JSI Co., Ltd., Korea) were used to modify the PA membrane surface. 2.2. Modification of PA membranes The top layer of the PA TFC membrane has residual carboxylic acids (–COOH) [17]. The formation of the carboxylic acid is due to the hydrolysis of acyl chloride groups (–COCl) that were remained after the reaction with amine. Therefore, further modification of nascent TFC membrane is possible through a chemical reaction between the unreacted acyl chloride groups and other aminecontaining chemicals when the further reaction is conducted before hydrolysis of the acyl chloride is complete [18]. After the interfacial polymerization reaction between TMC (0.06 wt%) and MPD (2.25 wt%), excess hydrocarbon solution was removed and the membrane was subsequently dipped into a solution of DMAP in DI water for 5 min. The membranes were rinsed with DI water to remove the residual DMAP solution and then the membrane was dipped into a solution containing SPGE and glycerol. The glycerol worked as a humectant, while the DMAP, a tertiary amine, was used as both an anchor to hold the SPGE to the membrane surface and a ring-opening agent since tertiary amine can initiate ring opening of the epoxy moiety of the SPGE polymer [19]. The ring-opening polymerization reaction of SPGE was carried out in a convection oven at 60 1C for 5 min, and the glycerol prevented the membrane from drying out in the oven (Scheme 1). All of the membranes were thoroughly washed with and stored in DI water before test.
2.1. Chemicals and reagents Polysulfone layer on a nonwoven fabric was prepared using polysulfone beads with a number average molecular weight of 35,000 Da (Solvay Advanced Polymers, L.L.C., USA) and N,N-dimethyl formamide (DMF) (Acros Organics, USA). The chemicals used in the PA skin layer formation include 2.25 wt% m-phenylene diamine (MPD) (Sigma-Aldrich) dissolved in deionized (DI) water and 0.06 wt% trimesoyl chloride (TMC) (Sigma-Aldrich). A proprietary
2.3. Membrane surface characterization and performance evaluation The chemical composition changes after modification of the PA membrane that confirm the successful coating of the DMAP and SPGE on the PA TFC membrane were investigated using Nicolet Magna 550 Fourier transform infrared spectroscopy (FTIR) (Midac, USA) and SIGMA PROBE X-ray photoelectron spectroscopy (XPS) (Thermo VG
Scheme 1. Schematic drawing of the surface modification process: (a) polyamide (PA) thin film composite (TFC) membrane, (b) DMAP-treated TFC membrane and (c) SPGE-treated TFC membrane.
194
Y.-N. Kwon et al. / Journal of Membrane Science 415–416 (2012) 192–198
Scientific Ltd., UK). The atomic percent of carbon, oxygen, and nitrogen of each sample was monitored and the ratios between atoms were calculated from the atomic percent. The streaming potential of the membrane was measured in the range of pH 3–10 using a BI-EKA streaming potential analyzer (Brookhaven, USA) with a 1 mM KCl electrolyte solution. Contact angle measurements were performed using an FTA˚ ˚ 200 contact angle analyzer (First Ten Angstroms, USA) using the sessile drop technique. At least five measurements were taken at different locations on the membrane surface to determine the average contact angle value. The surface features of the membranes were investigated using an FEI XL30 Sirion Scanning Electron Microscope (SEM). Since the PA samples are not electrically grounded to the sample holder, all of the samples were sputter-coated with a thin layer of gold to minimize sample charging. Atomic force microscopy (AFM) (PSIA, Korea) was used to quantify the roughness of RO membranes in a tapping mode using a nitride cantilever. The filtration performances of both virgin and surface-modified membranes were evaluated using a cross-flow membrane filtration unit. The effective area of the RO membrane was 28 cm2 (with 7 cm length 4 cm width size). Before the filtration test, RO membranes were compacted using DI water for at least 1 h under 2.0 MPa, and the temperature was adjusted to 2571 1C. Initial performance of the membrane was measured at 1.5 MPa using 0.2 wt% NaCl solution. Flow rate was maintained at 4 L/min, while reservoir volume was 20 L. The conductivity of the feed and permeate was measured to calculate salt rejection of the membranes using Orion model 115. All membranes were tested using a total of six membranes for RO performance, results of which have been averaged. Chlorination of the membranes was conducted in a soaking solution of 100 ppm free chlorine at various exposure times: 540, 1620, and 3780 ppm hr Cl2 at pH 7.
3. Results and discussion 3.1. Optimum preparation condition for membrane modification PA TFC membranes have carboxylic acids on their surfaces due to the hydrolysis of acyl chloride groups that remained after the interfacial polymerization. Modification of the membrane using the reactivity of residual acyl chloride immediately after the interfacial polymerization was conducted to improve the chlorine resistance. The virgin polyamide membrane prepared before the modification showed a salt rejection of 98.9% and a water flux of about 28.6 gfd for a feed solution containing 0.2 wt% NaCl at 1.5 Mpa. To find an optimum preparation condition of modified PA TFC RO membranes, the concentration effects of DMAP (ring-opening promoter), SPGE (coating material), and glycerol (humectant) during the modification on the permeation properties were systematically investigated. The concentration effect of DMAP on membrane performance was investigated under the following membrane preparation condition: 0.1 wt% SPGE and 2.0 wt% glycerol in DI water (Fig. 1a). With increasing concentrations of the DMAP solution, water flux monotonically increased but salt rejection gradually decreased. TFT-30 type commercial polyamide membranes are reported to have 9% carboxylic acid in the membrane [20], indicating that the amount of acyl chloride groups on the nascent PA membrane is 9%. The primary amine group in DMAP can form chemical bonds with an unreacted acyl chloride group of nascent PA skin layer and residual acyl chloride monomer (or oligomer) sitting on the membrane, which could react with the remaining –NH2 group in the membrane structure right after the interfacial polymerization. Therefore, it is thought that the reaction of the acyl chloride with DMAP decreased the degree of cross-linking of the PA skin layer. Tertiary amine DMAP worked as both an anchor to hold the SPGE to the membrane surface and an initiator to open the epoxy rings of the SPGE polymer. Degree of polymerization is inversely proportional to the concentration of initiators since the initiators terminate the polymer reactions, producing short polymers on the membrane. The rinse step conducted after the modification could wash out the short polymers from the membranes. When the membrane was treated with DMAP, the membrane turns to be more neutral at pH 7 due to the removal of carboxylic acid forming group, acyl chloride. The reduced electrostatic repulsion interactions inside polymer chains decreased salt rejection due to the decreased Donnan effect. The decreased degree of crosslinking and washing-out
Fig. 1. Effect of (a) DMAP, (b) SPGE and (c) glycerol concentration on the performance of surface-modified RO membrane measured in 0.2 wt% NaCl aqueous solution at 1.5 Mpa, 25 1C, and pH 7; (a) at SPGE¼ 0.1 wt%, glycerol¼ 2 wt%, (b) at DMAP¼ 0.1 wt%, glycerol ¼2 wt%, and (c) at DMAP¼ 0.1 wt%, SPGE¼ 0.1 wt%. DMAP, N,N-dimethylaminopropylamine; SPGE, sorbitol polyglycidyl ether; RO, reverse osmosis.
effect along with charge neutralization of surface charge after DMAP treatment might have effected on the enhanced passage of water and decline of salt rejection. Fig. 1b shows the membrane performance as a function of SPGE concentration. The concentrations of DMAP and glycerol were fixed at 0.1 and 2.0 wt%, respectively. The water flux continuously decreased. On the other hand, the salt rejection of the modified membranes slightly decreased at SPGE concentrations o0.05 wt% and then continuously increased. The observed performance change with increasing SPGE concentration may be explained in terms of the charge neutralization like at the case of DMAP and hydraulic resistance that resulted from the surface coating layer (electrically neutral SPGE coating on the membrane surface). When SPGE is polymerized via the ring-opening mechanism on the surface of the RO membrane, the surface charge density is reduced. As a result, the salt passage increases due to the decreased Donnan effect (o0.05 wt%). In contrast, at higher concentrations, molecular overlapping and aggregation of the polymer on the membrane surface occur and a denser layer of polymer forms on the membrane surface. The decreased electrical repulsion effect can be compensated by the increased resistance arising from the coating layer. Subsequently, the salt rejection of the modified membranes
Y.-N. Kwon et al. / Journal of Membrane Science 415–416 (2012) 192–198 increased as the coating solution concentration increases to 40.05 wt%. Water flux and salt rejection of modified membrane increased with increasing concentration of glycerol from 0.5% to 2.0%, and then there was no further change of flux and rejection (Fig. 1c). Glycerol was not dominantly involved in the chemical reaction between epoxy rings, but glycerol as a humectant prevented deformation and damage of the RO membrane during the thermal ring-opening reaction in an oven. The optimum preparation condition of the modified PA membrane was obtained by changing the concentrations of DMAP, SPGE, and glycerol: DMAP¼ 0.1 wt%, SPGE¼ 0.1 wt%, and glycerol¼ 2 wt%. The membranes prepared under the optimum condition were used to investigate the change of surface properties after modification and to investigate chlorine stability.
195
bonds and incorporation of tertiary amine. The membrane showed an increased atomic percent of oxygen after SPGE coating, indicating that the SPGE coating layer was successfully formed on the membrane surface. Peak fitting of high resolution XPS scan for oxygen 1s confirms the successful formation of SPGE on the membrane (Fig. 3). For the virgin, DMAP-treated, and SPGE-coated membranes, the deconvoluted peaks were normalized with respect to the peak with the lowest binding energy (corrected to 285 eV). The virgin and DMAP-coated membranes had only two peaks: a major peak that account for amide oxygen
3.2. Change of surface properties due to the modification 3.2.1. Spectroscopic analysis Ring-opening polymerization of SPGE on the membrane surface was confirmed by comparing FTIR spectroscopy findings of virgin and surface-modified membranes. Since the active layer is very thin and the penetration depth of the FTIR beam is greater than the skin layer’s thickness, the FTIR spectra of the membrane samples shows information about both the active layer and the polysulfone support layer. The spectra of virgin and modified RO membranes presented characteristic peaks of polysulfone such as 1584 cm 1 and 1243 cm 1. The major characteristic peaks of aromatic PA are shown in Fig. 2 at 1660 cm 1 (amide I, C–O stretching mode) and 1544 cm 1 (amide II, N-H bending mode). There were no remarkable differences in the FTIR spectra between the virgin and DMAP-treated membranes. However, spectroscopy of the SPGE-coated membranes shows that new peaks corresponding to –CH2– stretching motion at around 3000 cm 1 appeared and the absorption intensity of the peak (3300 cm 1) attributing –OH stretching motion increased. It can be explained as the introduction of hydroxyl groups after the ring-opening reaction of epoxy in SPGE. Further investigation was carried out using XPS. The atomic percentages of nitrogen, carbon, and oxygen on the surface of virgin and modified membranes are shown in Table 1. After the virgin membrane was treated with DMAP, the nitrogen content of the membrane slightly increased due to the formation of new amide
Fig. 3. High-resolution XPS scans for oxygen 1s peaks.
Fig. 2. Attenuated total reflectance-Fourier transform infrared transmittance spectra of the surface of the active layer of (a) polyamide membrane, (b) N,Ndimethylaminopropylamine-treated membrane and (c) sorbitol polyglycidyl ether-treated membrane.
Table 1 Atomic percent of elements composing the active layer of a polyamide membrane, an N,N-dimethylaminopropylamine (DMAP)-treated membrane and a sorbitol polyglycidyl ether (SPGE)-treated membrane. Sample
Polyamide DMAP SPGE
XPS surface elemental analysis
Relative ratio
C (%)
O (%)
N (%)
O/C
N/C
O/N
75.6 7 0.4 75.9 7 1.3 72.4 7 0.5
14.4 70.1 13.6 70.7 18.9 71.2
10.07 0.4 10.57 0.6 8.77 0.7
0.19 0.18 0.26
0.13 0.14 0.12
1.44 1.30 2.17
XPS, X-ray photoelectron spectroscopy.
Fig. 4. Zeta potential of the surface of the active layer of (a) polyamide membrane, (b) N,N-dimethylaminopropylamine-treated membrane and (c) sorbitol polyglycidyl ether (SPGE)-treated membrane.
196
Y.-N. Kwon et al. / Journal of Membrane Science 415–416 (2012) 192–198 and a minor peak that corresponds to carboxylic oxygen. After coating with SPGE, the new peak for C–O–C was observed at 533.0 eV.
Fig. 5. Sessile drop contact angle of the surface of the active layer of (a) polyamide membrane, (b) N,N-dimethylaminopropylamine (DMAP)-treated membrane and (c) sorbitol polyglycidyl ether (SPGE)-treated membrane.
3.2.2. Zeta potential and contact angle measurements The zeta potential of the RO membranes is presented in Fig. 4 as a function of pH. The virgin membrane exhibited negative zeta potentials in the range of pH 5–10, and the zeta potential became less negative with decreasing pH. The DMAPtreated membrane, unlike conventional PA membranes, showed positive zeta potential in the range of pH 3–6. The observed decrease of negative zeta potential was caused by introduction of tertiary amine group of DMAP and protonation of the amine under acidic conditions (82). After the membrane was coated with SPGE, it exhibited a negative zeta potential over the entire pH range. The result indicates that the formation of a neutral SPGE coating layer covers the charge of the membrane surface. Hydrophilicity of membranes before and after membrane modification was investigated using sessile drop contact angle measurements. A drop of pure water was delivered onto the membrane surface and the angle between membrane– water interface and water–air interface was measured as a measure of hydrophilicity. As shown in Fig. 5, the water contact angle decreased in the order of: virgin PA membrane (62 72.7), DMAP-treated membrane (427 2.6), and SPGEcoated membrane (297 2.0). This is mainly due to the surface coverage with hydrophilic amine or hydroxyl groups. The increases of surface hydrophilicity of
Fig. 6. Scanning electron microscopy and atomic force microscopy images (including morphological statistics) of (a, d) polyamide membrane, (b, e) N,N-dimethylaminopropylamine (DMAP)-treated membrane, and (c, f) sorbitol polyglycidyl ether (SPGE)-treated membrane.
Fig. 7. Effect of (a) N,N-dimethylaminopropylamine (DMAP); (b) sorbitol polyglycidyl ether (SPGE); (c) glycerol concentration on the chlorine resistance of surfacemodified reverse osmosis membrane. Coating conditions: (a) SPGE ¼0.1 wt%, glycerol ¼2 wt%; (b) DMAP ¼0.1 wt%, glycerol ¼2 wt%; (c) DMAP¼ 0.1 wt%, SPGE¼ 0.1 wt%. Chlorination condition: 3780 ppm hr Cl2 at pH 7. Filtration conditions: 0.2 wt% NaCl aqueous solution, 1.5 MPa operating pressure, 25 1C, and pH 7.
Y.-N. Kwon et al. / Journal of Membrane Science 415–416 (2012) 192–198 the modified membranes compared with that of the virgin membrane indicates successful formation of SPGE on the RO surface, a finding that is consistent with the FTIR analysis. 3.2.3. Surface morphology of the modified RO membrane Surface morphology of the virgin and modified membranes was investigated using SEM (Fig. 5a, b, and c) and AFM (Fig. 5d, e, and f). A ridge and valley
197
structure on the membrane surface was observed in SEM images, causing irregularities and local variations of surface properties such as contact angle. There was no noticeable change observed between before and after modification, but the surface-modified membranes showed relatively smoother surfaces. AFM image analysis was carried out for further investigation of surface morphology. Three-dimensional 5 mm scan images for the unmodified and modified membranes were taken, and the root mean square roughness (Rms) and peak-to-valley (Rp-v) distance of the membrane surface was calculated and listed at the bottom table in Fig. 6. After the membrane was coated with SPGE, the Rms of the modified membrane decreased from 47 nm to 27 nm and the Rpv of the modified membrane decreased from 374 nm to 182 nm. AFM analysis results also revealed that the modification of the membrane by coating makes the membrane surface smoother. The decreased surface roughness subsequently may improve the antifouling properties along with enhancing the hydrophilicity [21–23].
3.3. Effect of modification on chlorine resistance
Fig. 8. Salt passage changes of virgin and modified polyamide membranes after chlorination under the condition of 540, 1620, 3780 ppm hr (Cl2) and pH 7. Coating conditions of the modified membrane were 0.1 wt% N,N-dimethylaminopropylamine, 0.1 wt% sorbitol polyglycidyl ether, and 2.0 wt% glycerol.
Table 2 Atomic percent of elements on virgin and chlorinated membranes. Sample type
XPS surface elemental analysis C(%)
Polyamide Modified polyamide
Virgin 77.1 Chlorinated 72.8 Virgin 74.6 Chlorinated 70.8
XPS, X-ray photoelectron spectroscopy.
Relative ratio
O(%) N(%) Cl(%)
Cl/N
12.0 13.8 17.0 18.2
0.47
10.9 9.1 8.4 8.9
0 4.3 0 2.1
0.24
The chlorine resistance of the surface-modified RO membrane was investigated in terms of salt passage. The amount of salt passing through the membrane per unit area per unit time at a feed concentration of 2000 ppm NaCl was defined as salt flux (Fs), while the difference in salt flux between the chlorinated membrane and the virgin membrane over the salt flux of the virgin membrane was designated as RDFs/100. The value of RDFs, which indicated the percentage of salt flux change after chlorination compared with that of the virgin membrane, was used to indicate how tolerant the modified membrane had become. Fs ¼ 3154 FluxðgfdÞ ð12Rejection=100Þ RDFs ¼ ðFsðchlorinatedÞ 2FsðvirginÞ Þ=FsðvirginÞ 100
ðmg=m2 hrÞ ð%Þ
A higher RDFs value indicates that the membrane is more susceptible to the chlorine attack and, thus, more salt passed through the membrane. The concentration effect of DMAP, SPGE, and glycerol on membrane chlorine resistance is shown in Fig. 7. RDFs increased gradually with increasing DMAP concentration. The increment of salt passage change, that is, lower chlorine resistance, can be explained as the lower cross-linking density of DMAP-treated membranes with increasing concentration, a finding that is consistent with that shown in Fig. 1a. On the other hand, the salt passage after chlorination decreased as the SPGE and glycerol concentrations increased. This phenomenon is attributed to the fact that the SPGE-coating layers prevented direct contact between chlorine and PA membrane surface and the glycerol which was rinsed away after fabrication prevented damage of the RO membrane during thermal ring-opening
Fig. 9. Scanning electron microscopy photographs of (A) virgin polyamide, (B) surface-modified polyamide and (a) chlorinated polyamide, (b) chlorinated surface-modified polyamide (10,000 ppm hr Cl2 at pH 7).
198
Y.-N. Kwon et al. / Journal of Membrane Science 415–416 (2012) 192–198
reaction in an oven. Chlorine resistance of the membranes modified under the optimum preparation condition were investigated compared with those of the virgin membrane under three different soaking conditions: 540, 1,620, and 3780 ppm hr Cl2 at pH 7. After exposure of the membranes to the chlorine solutions, the membranes were removed from the chlorine solution and thoroughly rinsed with DI water to remove any unreacted chlorine. Surface-modified membranes allowed less salt passage compared to the virgin membrane after the chlorine soaking test under all three conditions (Fig. 8). This result shows that SPGE coating layer on the PA membrane surface efficiently prevented chlorine attack. To evaluate the degree of chlorination quantitatively, XPS analysis was conducted. The surface chemical compositions of both virgin and surface modified membranes before and after chlorination are listed in Table 2. When the membranes were soaked in chlorine solutions, chlorine was detected that had not been seen before the chlorination. However, the atomic ratio of Cl/N decreased by half after the PA membrane surface was modified with SPGE. All of these results show drastically enhanced chlorine resistance of the surface-modified RO membrane. Fig. 9 shows surface SEM images of virgin membrane and membrane chlorinated at 10,000 ppm hr Cl2 and pH 7. Both the virgin and modified membranes showed a typical ridge and valley surface structure before chlorination. However, an obvious change in surface morphology was observed after chlorination. It is clearly shown that the initial morphology of the virgin membrane was ruptured but the SPGE-coated membrane kept its initial morphology after chlorination.
4. Summary In this study, the preparation of chlorine-resistant RO membranes using the surface coating method was studied and the chlorine resistances of modified membranes were evaluated. A chlorine-resistant coating layer was formed using a multi-functional epoxy compound, SPGE, via ring-opening polymerization. The effects of DMAP, SPGE, and glycerol on membrane permeation properties were investigated. With increasing DMAP solution concentration, water flux increased and salt rejection decreased due to the reduction of further cross-linkable sites. The introduction of neutral SPGE led to decreased salt rejection at lower concentrations. In contrast, at higher SPGE concentrations, molecular overlapping of the polymer led to the formation of a denser coating layer, which resulted in increased salt rejection. Glycerol used as a humectant prevented the membrane from drying out during the ring-opening reaction in an oven and increased membrane performance. The modified membrane was characterized using FTIR, XPS, zeta potential, and contact angle. Each result supported the successful incorporation of an SPGE coating layer on the PA membrane and showed the modification resulted in a membrane surface that was more neutral, hydrophilic, and smooth. The chlorine exposure test using a NaOCl solution showed that the modified surface membranes were much less susceptible to chlorine attack and had enhanced chlorine stability.
Acknowledgments This research was supported by Basic Science Research Program through the National Research Foundation of Korea (NRF)
funded by the Ministry of Education, Science and Technology (2011–0002457). The help and assistance of the Woongjin Chemical Co., Ltd. are greatly acknowledged. References [1] N.N. Li, A.G. Fane, W.S.W. Ho, M. T, Advanced membrane tehcnology and applications, Wiley, 2008. [2] J.E. Cadotte, Interfacially synthesized reverse osmosis membrane U.S. Patent 4,277,344, 1981. [3] J. Kucera, Reverse osmosis, Wiley, 2010. [4] Y.N. Kwon, R. Joksimovic, I.C. Kim, J.O. Leckie, Effect of bromide on the chlorination of a polyamide membrane, Desalination 280 (2011) 80–86. [5] Y.N. Kwon, J.O. Leckie, Hypochlorite degradation of crosslinked polyamide membranes. Changes in hydrogen bonding behavior and performance, J. Membr. Sci. 282 (2006) 456–464. [6] Y.N. Kwon, J.O. Leckie, Hypochlorite degradation of crosslinked polyamide membranes. Changes in chemical/morphological properties, J. Membr. Sci. 283 (2006) 21–26. [7] Y.N. Kwon, C.Y. Tang, J.O. Leckie, Change of membrane performance due to chlorination of crosslinked polyamide membranes, J. Appl. Polym. Sci. 102 (2006) 5895–5902. [8] Y.N. Kwon, C.Y. Tang, J.O. Leckie, Change of chemical composition and hydrogen bonding behavior due to chlorination of crosslinked polyamide membranes, J. Appl. Polym. Sci. 108 (2008) 2061–2066. [9] T. Kawaguchi, H. Tamura, Chlorine-resistant membrane for reverse osmosis. Preparation of chlorine-resistant polyamide composite membranes, J. Appl. Polym. Sci. 29 (1984) 3369–3379. [10] T. Uemura, M. Kurihara, Chlorine resistance of reverse osmosis membranes and changes in membrane structure and performance caused by chlorination degradation, Bull. Soc. Sea Water Sci., Japan 57 (2003) 498–507. [11] J. Glater, M.R. Zachariah, Mechanistic study of halogen interaction with polyamide reverse-osmosis membranes, ACS Symp. Ser. (1985) 345–358. [12] J. Glater, S.K. Hong, M. Elimelech, Search for a chlorine-resistant reverse osmosis membrane, Desalination 95 (1994) 325–345. [13] S.H. Son, J. Jegal, Preparation and characterization of polyamide reverseosmosis membranes with good chlorine tolerance, J. Appl. Polym. Sci. 120 (2011) 1245–1252. [14] R.H. Du, J.S. Zhao, Properties of poly (n,n-dimethylaminoethyl methacrylate)/ polysulfone positively charged composite nanofiltration membrane, J. Mem. Sci. 239 (2004) 183–188. [15] G.D. Kang, C.J. Gao, W.D. Chen, X.M. Jie, Y.M. Cao, Q. Yuan, Study on hypochlorite degradation of aromatic polyamide reverse osmosis membrane, J. Mem. Sci. 300 (2007) 165–171. [16] N. Li, Z.Z. Liu, S.G. Xu, Dynamically formed poly (vinyl alcohol) ultrafiltration membranes with good anti-fouling characteristics, J. Mem. Sci. 169 (2000) 17–28. [17] R.J. Petersen, Composite reverse-osmosis and nanofiltration membranes, J. Mem. Sci. 83 (1993) 81–150. [18] G.D. Kang, M. Liu, B. Lin, Y.M. Cao, Q. Yuan, A novel method of surface modification on thin-film composite reverse osmosis membrane by grafting poly(ethylene glycol), Polymer 48 (2007) 1165–1170. [19] D. Ratna, Handbook of thermoset resins, iSmithers Rapra Publishing, 2009. [20] J.Y. Koo, R.J. Petersen, J.E. Cadotte, Esca characterization of chlorine-damaged polyamide reverse osmosis membrane, Polym. Prepr. Div. Polym. Chem. Am. Chem. Soc. 27 (1986) 391–392. [21] M. Elimelech, X.H. Zhu, A.E. Childress, S.K. Hong, Role of membrane surface morphology in colloidal fouling of cellulose acetate and composite aromatic polyamide reverse osmosis membranes, J. Mem. Sci. 127 (1997) 101–109. [22] C.R. Reiss, J.S. Taylor, C. Robert, Surface water treatment using nanofiltration—pilot testing results and design considerations, Desalination 125 (1999) 97–112. [23] A.I. Schafer, A.G. Fane, T.D. Waite, Nanofiltration of natural organic matter: removal, fouling and the influence of multivalent ions, Water Supply 17 (1999) 191–204.
Contents lists available at SciVerse ScienceDirect
Journal of Membrane Science journal homepage: www.elsevier.com/locate/memsci
Surface modification of a polyamide reverse osmosis membrane for chlorine resistance improvement Young-Nam Kwon a, Sungpyo Hong b, Hyoungwoo Choi b, Taemoon Tak b,n a b
School of Urban and Environmental Engineering, Ulsan National Institute of Science and Technology (UNIST), 100 Banyeon-ri, Ulsan 689-798, South Korea Department of Biosystems and Biomaterials Science and Engineering, Seoul National University, 599 Gwanangno, Gwanak-gu, Seoul 151-921, South Korea
a r t i c l e i n f o
a b s t r a c t
Article history: Received 28 December 2011 Received in revised form 25 April 2012 Accepted 29 April 2012 Available online 22 May 2012
A surface modified polyamide (PA) thin film composite (TFC) membrane was prepared using in-situ polymerization of sorbitol polyglycidyl ether (SPGE) on the membrane surface immediately after interfacial polymerization of TFC membranes. This modification was conducted to protect the chlorine-sensitive sites of the PA membrane using a chlorine-resistant hydrophilic SPGE polymer. The optimum preparation condition of the modified PA TFC reverse osmosis membranes was investigated, and then the successful modification of the membrane was confirmed using various analytical tools including Fourier transform infrared spectroscopy, a zeta potential analyzer, a contact angle analyzer, X-ray photoelectron microscopy, scanning electron microscopy, and atomic force microscopy. The modification converted the surface of the membrane to a more neutral, hydrophilic, and smooth surface. With increasing SPGE concentration in the coating solution, molecular overlapping of the coating polymer led to a denser coating layer, which resulted in declined flux but increased salt rejection. Chlorination tests showed that the modification of the membrane using SPGE ring-opening polymerization improved chlorine stability. & 2012 Elsevier B.V. All rights reserved.
Keywords: Polyamide Surface modification Chlorine resistance Reverse osmosis
1. Introduction Reverse osmosis (RO) is a membrane-based desalination process in which relatively pure water is transported across the membrane by the pressure gradient between the feed and permeate water, while dissolved organics and inorganics are rejected due to the chemical nature of the membrane material and the physical structure of the membrane [1]. Since the development of composite polyamide (PA) membranes using interfacial condensation method by Cadotte [2], drastic progress in RO membrane technology has been made. However, widespread use of the membrane is limited due to the adsorption of retained organic and/or inorganic materials and the subsequent decline in performance [3]. Fouling, which is defined as the (ir)reversible deposition of retained materials, can be mitigated through pretreatment of feed water before the membrane filtration and minimized by a physical or chemical cleaning process after filtration. Among the techniques controlling membrane fouling, the addition of chlorine to the feed water is one of most convenient method and has been adopted in many RO applications for deactivating microorganisms causing membrane biofouling. Since the PA membrane is reported to be very sensitive to chlorine [4–8], reducing agents need to be used in the dechlorination process. Partial failure of the dechlorination
n
Corresponding author. Tel.: þ82 2 880 4621. E-mail addresses: [email protected] (Y.-N. Kwon), [email protected] (T. Tak).
0376-7388/$ - see front matter & 2012 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.memsci.2012.04.056
process can result in the deterioration of PA membranes, making the whole membrane useless. Various approaches have been devised to develop chlorineresistant membrane by eliminating chlorine-sensitive sites of the PA membrane or protecting the sensitive sites using chlorineresistant coating materials. Amide nitrogen and aromatic rings of the PA membrane are sites that can be easily attacked by chlorine. Kawaguch and Tamura [9] chlorinated secondary and tertiary amide to investigate chlorinating behavior in the presence or absence of amidic hydrogen and demonstrated that the amide nitrogen is a key chlorination reaction site. Uemura and Kurihara [10] observed that a cross-linked N-substituted PA TFC membrane is more chlorine tolerant compared to a typical cross-linked PA TFC membrane. Glater and Zachariah [11] investigated the effect of chlorine on a PA membrane and concluded that chlorine attacks on aromatic rings change the hydrogen-bonding forces. Chlorination of the aromatic ring in the PA membrane can be blocked by the substitution of a deactivating functional group such as –NO2 or –SO3H at the ring position [12]. Son and Jegal [13] developed TFC RO membranes using several different amines and trimesoyl chloride (TMC) and showed that the m-phenylene diamine (MPD) derivative with functional groups such as –CH3 and –OCH3 has better chlorine tolerance. Surface modification such as grafting or physical adsorption is a potential technique to improve chlorine resistance of commercially available PA membranes. Du and Zhao [14] prepared poly (N,N-dimethylaminoethyl methacrylate) (PDMAEMA) positively charged nanofiltration membranes using an interfacial crosslinking reaction and demonstrated that the membrane was stable
Y.-N. Kwon et al. / Journal of Membrane Science 415–416 (2012) 192–198
during 28-days storage in 5 ppm NaOCl preservatives. Based on the results of Du and Zhao, Kang et al. [15] coated PA RO membranes with the PDMAEMA and demonstrated that the coated membrane shows better chlorine resistance. Polyvinyl alcohol (PVA) is a hydrophilic and electric neutral polymer that has been applied to the RO membrane to improve hydrophilicity and, thus, antifouling properties. Furthermore, the PVA coating was reported to provide enhanced chlorine resistance to the membrane [16]. Considering long-term stability, the cross-linked structure of these coating materials is desirable. However, modification using a grafting or physical adsorption method requires another cross-linking step, which might be economically undesirable. Therefore, it is economically favorable to form the cross-linked coating using only one step. The objective of this study was to develop a novel method that is suitable for surface modification of an aromatic PA RO membrane to improve its chlorine tolerance. Cross-linked sorbitol polyglycidyl ether (SPGE) was polymerized on the PA RO membrane surface via in situ ring-opening polymerization. N,Ndimethylaminopropylamine (DMAP) and glycerol were used as a ring-opening agent and a humectant. The physicochemical properties of the membrane surface were analyzed using various analytical tools and compared before and after the coating process. Permeability and selectivity of the virgin and modified membranes were evaluated by filtration of a 2000 ppm NaCl aqueous solution at 25 1C under the operating pressure of 1.5 MPa. Exposures of the membranes to 100 ppm free chlorine solutions during various exposure times were carried out to evaluate membrane chlorine resistances.
2. Experimental materials and methods
193
organic solvent, Isol-C (SK Chemical Inc., Korea), was selected to dissolve the acid chloride and prepare the organic phase solution for the interfacial polymerization. The reaction time for the interfacial polymerization was 90 s. N,N-dimethylamino propylamine (DMAP) (Acros Organics, USA) and sorbitol polyglycidyl ether (SPGE) (JSI Co., Ltd., Korea) were used to modify the PA membrane surface. 2.2. Modification of PA membranes The top layer of the PA TFC membrane has residual carboxylic acids (–COOH) [17]. The formation of the carboxylic acid is due to the hydrolysis of acyl chloride groups (–COCl) that were remained after the reaction with amine. Therefore, further modification of nascent TFC membrane is possible through a chemical reaction between the unreacted acyl chloride groups and other aminecontaining chemicals when the further reaction is conducted before hydrolysis of the acyl chloride is complete [18]. After the interfacial polymerization reaction between TMC (0.06 wt%) and MPD (2.25 wt%), excess hydrocarbon solution was removed and the membrane was subsequently dipped into a solution of DMAP in DI water for 5 min. The membranes were rinsed with DI water to remove the residual DMAP solution and then the membrane was dipped into a solution containing SPGE and glycerol. The glycerol worked as a humectant, while the DMAP, a tertiary amine, was used as both an anchor to hold the SPGE to the membrane surface and a ring-opening agent since tertiary amine can initiate ring opening of the epoxy moiety of the SPGE polymer [19]. The ring-opening polymerization reaction of SPGE was carried out in a convection oven at 60 1C for 5 min, and the glycerol prevented the membrane from drying out in the oven (Scheme 1). All of the membranes were thoroughly washed with and stored in DI water before test.
2.1. Chemicals and reagents Polysulfone layer on a nonwoven fabric was prepared using polysulfone beads with a number average molecular weight of 35,000 Da (Solvay Advanced Polymers, L.L.C., USA) and N,N-dimethyl formamide (DMF) (Acros Organics, USA). The chemicals used in the PA skin layer formation include 2.25 wt% m-phenylene diamine (MPD) (Sigma-Aldrich) dissolved in deionized (DI) water and 0.06 wt% trimesoyl chloride (TMC) (Sigma-Aldrich). A proprietary
2.3. Membrane surface characterization and performance evaluation The chemical composition changes after modification of the PA membrane that confirm the successful coating of the DMAP and SPGE on the PA TFC membrane were investigated using Nicolet Magna 550 Fourier transform infrared spectroscopy (FTIR) (Midac, USA) and SIGMA PROBE X-ray photoelectron spectroscopy (XPS) (Thermo VG
Scheme 1. Schematic drawing of the surface modification process: (a) polyamide (PA) thin film composite (TFC) membrane, (b) DMAP-treated TFC membrane and (c) SPGE-treated TFC membrane.
194
Y.-N. Kwon et al. / Journal of Membrane Science 415–416 (2012) 192–198
Scientific Ltd., UK). The atomic percent of carbon, oxygen, and nitrogen of each sample was monitored and the ratios between atoms were calculated from the atomic percent. The streaming potential of the membrane was measured in the range of pH 3–10 using a BI-EKA streaming potential analyzer (Brookhaven, USA) with a 1 mM KCl electrolyte solution. Contact angle measurements were performed using an FTA˚ ˚ 200 contact angle analyzer (First Ten Angstroms, USA) using the sessile drop technique. At least five measurements were taken at different locations on the membrane surface to determine the average contact angle value. The surface features of the membranes were investigated using an FEI XL30 Sirion Scanning Electron Microscope (SEM). Since the PA samples are not electrically grounded to the sample holder, all of the samples were sputter-coated with a thin layer of gold to minimize sample charging. Atomic force microscopy (AFM) (PSIA, Korea) was used to quantify the roughness of RO membranes in a tapping mode using a nitride cantilever. The filtration performances of both virgin and surface-modified membranes were evaluated using a cross-flow membrane filtration unit. The effective area of the RO membrane was 28 cm2 (with 7 cm length 4 cm width size). Before the filtration test, RO membranes were compacted using DI water for at least 1 h under 2.0 MPa, and the temperature was adjusted to 2571 1C. Initial performance of the membrane was measured at 1.5 MPa using 0.2 wt% NaCl solution. Flow rate was maintained at 4 L/min, while reservoir volume was 20 L. The conductivity of the feed and permeate was measured to calculate salt rejection of the membranes using Orion model 115. All membranes were tested using a total of six membranes for RO performance, results of which have been averaged. Chlorination of the membranes was conducted in a soaking solution of 100 ppm free chlorine at various exposure times: 540, 1620, and 3780 ppm hr Cl2 at pH 7.
3. Results and discussion 3.1. Optimum preparation condition for membrane modification PA TFC membranes have carboxylic acids on their surfaces due to the hydrolysis of acyl chloride groups that remained after the interfacial polymerization. Modification of the membrane using the reactivity of residual acyl chloride immediately after the interfacial polymerization was conducted to improve the chlorine resistance. The virgin polyamide membrane prepared before the modification showed a salt rejection of 98.9% and a water flux of about 28.6 gfd for a feed solution containing 0.2 wt% NaCl at 1.5 Mpa. To find an optimum preparation condition of modified PA TFC RO membranes, the concentration effects of DMAP (ring-opening promoter), SPGE (coating material), and glycerol (humectant) during the modification on the permeation properties were systematically investigated. The concentration effect of DMAP on membrane performance was investigated under the following membrane preparation condition: 0.1 wt% SPGE and 2.0 wt% glycerol in DI water (Fig. 1a). With increasing concentrations of the DMAP solution, water flux monotonically increased but salt rejection gradually decreased. TFT-30 type commercial polyamide membranes are reported to have 9% carboxylic acid in the membrane [20], indicating that the amount of acyl chloride groups on the nascent PA membrane is 9%. The primary amine group in DMAP can form chemical bonds with an unreacted acyl chloride group of nascent PA skin layer and residual acyl chloride monomer (or oligomer) sitting on the membrane, which could react with the remaining –NH2 group in the membrane structure right after the interfacial polymerization. Therefore, it is thought that the reaction of the acyl chloride with DMAP decreased the degree of cross-linking of the PA skin layer. Tertiary amine DMAP worked as both an anchor to hold the SPGE to the membrane surface and an initiator to open the epoxy rings of the SPGE polymer. Degree of polymerization is inversely proportional to the concentration of initiators since the initiators terminate the polymer reactions, producing short polymers on the membrane. The rinse step conducted after the modification could wash out the short polymers from the membranes. When the membrane was treated with DMAP, the membrane turns to be more neutral at pH 7 due to the removal of carboxylic acid forming group, acyl chloride. The reduced electrostatic repulsion interactions inside polymer chains decreased salt rejection due to the decreased Donnan effect. The decreased degree of crosslinking and washing-out
Fig. 1. Effect of (a) DMAP, (b) SPGE and (c) glycerol concentration on the performance of surface-modified RO membrane measured in 0.2 wt% NaCl aqueous solution at 1.5 Mpa, 25 1C, and pH 7; (a) at SPGE¼ 0.1 wt%, glycerol¼ 2 wt%, (b) at DMAP¼ 0.1 wt%, glycerol ¼2 wt%, and (c) at DMAP¼ 0.1 wt%, SPGE¼ 0.1 wt%. DMAP, N,N-dimethylaminopropylamine; SPGE, sorbitol polyglycidyl ether; RO, reverse osmosis.
effect along with charge neutralization of surface charge after DMAP treatment might have effected on the enhanced passage of water and decline of salt rejection. Fig. 1b shows the membrane performance as a function of SPGE concentration. The concentrations of DMAP and glycerol were fixed at 0.1 and 2.0 wt%, respectively. The water flux continuously decreased. On the other hand, the salt rejection of the modified membranes slightly decreased at SPGE concentrations o0.05 wt% and then continuously increased. The observed performance change with increasing SPGE concentration may be explained in terms of the charge neutralization like at the case of DMAP and hydraulic resistance that resulted from the surface coating layer (electrically neutral SPGE coating on the membrane surface). When SPGE is polymerized via the ring-opening mechanism on the surface of the RO membrane, the surface charge density is reduced. As a result, the salt passage increases due to the decreased Donnan effect (o0.05 wt%). In contrast, at higher concentrations, molecular overlapping and aggregation of the polymer on the membrane surface occur and a denser layer of polymer forms on the membrane surface. The decreased electrical repulsion effect can be compensated by the increased resistance arising from the coating layer. Subsequently, the salt rejection of the modified membranes
Y.-N. Kwon et al. / Journal of Membrane Science 415–416 (2012) 192–198 increased as the coating solution concentration increases to 40.05 wt%. Water flux and salt rejection of modified membrane increased with increasing concentration of glycerol from 0.5% to 2.0%, and then there was no further change of flux and rejection (Fig. 1c). Glycerol was not dominantly involved in the chemical reaction between epoxy rings, but glycerol as a humectant prevented deformation and damage of the RO membrane during the thermal ring-opening reaction in an oven. The optimum preparation condition of the modified PA membrane was obtained by changing the concentrations of DMAP, SPGE, and glycerol: DMAP¼ 0.1 wt%, SPGE¼ 0.1 wt%, and glycerol¼ 2 wt%. The membranes prepared under the optimum condition were used to investigate the change of surface properties after modification and to investigate chlorine stability.
195
bonds and incorporation of tertiary amine. The membrane showed an increased atomic percent of oxygen after SPGE coating, indicating that the SPGE coating layer was successfully formed on the membrane surface. Peak fitting of high resolution XPS scan for oxygen 1s confirms the successful formation of SPGE on the membrane (Fig. 3). For the virgin, DMAP-treated, and SPGE-coated membranes, the deconvoluted peaks were normalized with respect to the peak with the lowest binding energy (corrected to 285 eV). The virgin and DMAP-coated membranes had only two peaks: a major peak that account for amide oxygen
3.2. Change of surface properties due to the modification 3.2.1. Spectroscopic analysis Ring-opening polymerization of SPGE on the membrane surface was confirmed by comparing FTIR spectroscopy findings of virgin and surface-modified membranes. Since the active layer is very thin and the penetration depth of the FTIR beam is greater than the skin layer’s thickness, the FTIR spectra of the membrane samples shows information about both the active layer and the polysulfone support layer. The spectra of virgin and modified RO membranes presented characteristic peaks of polysulfone such as 1584 cm 1 and 1243 cm 1. The major characteristic peaks of aromatic PA are shown in Fig. 2 at 1660 cm 1 (amide I, C–O stretching mode) and 1544 cm 1 (amide II, N-H bending mode). There were no remarkable differences in the FTIR spectra between the virgin and DMAP-treated membranes. However, spectroscopy of the SPGE-coated membranes shows that new peaks corresponding to –CH2– stretching motion at around 3000 cm 1 appeared and the absorption intensity of the peak (3300 cm 1) attributing –OH stretching motion increased. It can be explained as the introduction of hydroxyl groups after the ring-opening reaction of epoxy in SPGE. Further investigation was carried out using XPS. The atomic percentages of nitrogen, carbon, and oxygen on the surface of virgin and modified membranes are shown in Table 1. After the virgin membrane was treated with DMAP, the nitrogen content of the membrane slightly increased due to the formation of new amide
Fig. 3. High-resolution XPS scans for oxygen 1s peaks.
Fig. 2. Attenuated total reflectance-Fourier transform infrared transmittance spectra of the surface of the active layer of (a) polyamide membrane, (b) N,Ndimethylaminopropylamine-treated membrane and (c) sorbitol polyglycidyl ether-treated membrane.
Table 1 Atomic percent of elements composing the active layer of a polyamide membrane, an N,N-dimethylaminopropylamine (DMAP)-treated membrane and a sorbitol polyglycidyl ether (SPGE)-treated membrane. Sample
Polyamide DMAP SPGE
XPS surface elemental analysis
Relative ratio
C (%)
O (%)
N (%)
O/C
N/C
O/N
75.6 7 0.4 75.9 7 1.3 72.4 7 0.5
14.4 70.1 13.6 70.7 18.9 71.2
10.07 0.4 10.57 0.6 8.77 0.7
0.19 0.18 0.26
0.13 0.14 0.12
1.44 1.30 2.17
XPS, X-ray photoelectron spectroscopy.
Fig. 4. Zeta potential of the surface of the active layer of (a) polyamide membrane, (b) N,N-dimethylaminopropylamine-treated membrane and (c) sorbitol polyglycidyl ether (SPGE)-treated membrane.
196
Y.-N. Kwon et al. / Journal of Membrane Science 415–416 (2012) 192–198 and a minor peak that corresponds to carboxylic oxygen. After coating with SPGE, the new peak for C–O–C was observed at 533.0 eV.
Fig. 5. Sessile drop contact angle of the surface of the active layer of (a) polyamide membrane, (b) N,N-dimethylaminopropylamine (DMAP)-treated membrane and (c) sorbitol polyglycidyl ether (SPGE)-treated membrane.
3.2.2. Zeta potential and contact angle measurements The zeta potential of the RO membranes is presented in Fig. 4 as a function of pH. The virgin membrane exhibited negative zeta potentials in the range of pH 5–10, and the zeta potential became less negative with decreasing pH. The DMAPtreated membrane, unlike conventional PA membranes, showed positive zeta potential in the range of pH 3–6. The observed decrease of negative zeta potential was caused by introduction of tertiary amine group of DMAP and protonation of the amine under acidic conditions (82). After the membrane was coated with SPGE, it exhibited a negative zeta potential over the entire pH range. The result indicates that the formation of a neutral SPGE coating layer covers the charge of the membrane surface. Hydrophilicity of membranes before and after membrane modification was investigated using sessile drop contact angle measurements. A drop of pure water was delivered onto the membrane surface and the angle between membrane– water interface and water–air interface was measured as a measure of hydrophilicity. As shown in Fig. 5, the water contact angle decreased in the order of: virgin PA membrane (62 72.7), DMAP-treated membrane (427 2.6), and SPGEcoated membrane (297 2.0). This is mainly due to the surface coverage with hydrophilic amine or hydroxyl groups. The increases of surface hydrophilicity of
Fig. 6. Scanning electron microscopy and atomic force microscopy images (including morphological statistics) of (a, d) polyamide membrane, (b, e) N,N-dimethylaminopropylamine (DMAP)-treated membrane, and (c, f) sorbitol polyglycidyl ether (SPGE)-treated membrane.
Fig. 7. Effect of (a) N,N-dimethylaminopropylamine (DMAP); (b) sorbitol polyglycidyl ether (SPGE); (c) glycerol concentration on the chlorine resistance of surfacemodified reverse osmosis membrane. Coating conditions: (a) SPGE ¼0.1 wt%, glycerol ¼2 wt%; (b) DMAP ¼0.1 wt%, glycerol ¼2 wt%; (c) DMAP¼ 0.1 wt%, SPGE¼ 0.1 wt%. Chlorination condition: 3780 ppm hr Cl2 at pH 7. Filtration conditions: 0.2 wt% NaCl aqueous solution, 1.5 MPa operating pressure, 25 1C, and pH 7.
Y.-N. Kwon et al. / Journal of Membrane Science 415–416 (2012) 192–198 the modified membranes compared with that of the virgin membrane indicates successful formation of SPGE on the RO surface, a finding that is consistent with the FTIR analysis. 3.2.3. Surface morphology of the modified RO membrane Surface morphology of the virgin and modified membranes was investigated using SEM (Fig. 5a, b, and c) and AFM (Fig. 5d, e, and f). A ridge and valley
197
structure on the membrane surface was observed in SEM images, causing irregularities and local variations of surface properties such as contact angle. There was no noticeable change observed between before and after modification, but the surface-modified membranes showed relatively smoother surfaces. AFM image analysis was carried out for further investigation of surface morphology. Three-dimensional 5 mm scan images for the unmodified and modified membranes were taken, and the root mean square roughness (Rms) and peak-to-valley (Rp-v) distance of the membrane surface was calculated and listed at the bottom table in Fig. 6. After the membrane was coated with SPGE, the Rms of the modified membrane decreased from 47 nm to 27 nm and the Rpv of the modified membrane decreased from 374 nm to 182 nm. AFM analysis results also revealed that the modification of the membrane by coating makes the membrane surface smoother. The decreased surface roughness subsequently may improve the antifouling properties along with enhancing the hydrophilicity [21–23].
3.3. Effect of modification on chlorine resistance
Fig. 8. Salt passage changes of virgin and modified polyamide membranes after chlorination under the condition of 540, 1620, 3780 ppm hr (Cl2) and pH 7. Coating conditions of the modified membrane were 0.1 wt% N,N-dimethylaminopropylamine, 0.1 wt% sorbitol polyglycidyl ether, and 2.0 wt% glycerol.
Table 2 Atomic percent of elements on virgin and chlorinated membranes. Sample type
XPS surface elemental analysis C(%)
Polyamide Modified polyamide
Virgin 77.1 Chlorinated 72.8 Virgin 74.6 Chlorinated 70.8
XPS, X-ray photoelectron spectroscopy.
Relative ratio
O(%) N(%) Cl(%)
Cl/N
12.0 13.8 17.0 18.2
0.47
10.9 9.1 8.4 8.9
0 4.3 0 2.1
0.24
The chlorine resistance of the surface-modified RO membrane was investigated in terms of salt passage. The amount of salt passing through the membrane per unit area per unit time at a feed concentration of 2000 ppm NaCl was defined as salt flux (Fs), while the difference in salt flux between the chlorinated membrane and the virgin membrane over the salt flux of the virgin membrane was designated as RDFs/100. The value of RDFs, which indicated the percentage of salt flux change after chlorination compared with that of the virgin membrane, was used to indicate how tolerant the modified membrane had become. Fs ¼ 3154 FluxðgfdÞ ð12Rejection=100Þ RDFs ¼ ðFsðchlorinatedÞ 2FsðvirginÞ Þ=FsðvirginÞ 100
ðmg=m2 hrÞ ð%Þ
A higher RDFs value indicates that the membrane is more susceptible to the chlorine attack and, thus, more salt passed through the membrane. The concentration effect of DMAP, SPGE, and glycerol on membrane chlorine resistance is shown in Fig. 7. RDFs increased gradually with increasing DMAP concentration. The increment of salt passage change, that is, lower chlorine resistance, can be explained as the lower cross-linking density of DMAP-treated membranes with increasing concentration, a finding that is consistent with that shown in Fig. 1a. On the other hand, the salt passage after chlorination decreased as the SPGE and glycerol concentrations increased. This phenomenon is attributed to the fact that the SPGE-coating layers prevented direct contact between chlorine and PA membrane surface and the glycerol which was rinsed away after fabrication prevented damage of the RO membrane during thermal ring-opening
Fig. 9. Scanning electron microscopy photographs of (A) virgin polyamide, (B) surface-modified polyamide and (a) chlorinated polyamide, (b) chlorinated surface-modified polyamide (10,000 ppm hr Cl2 at pH 7).
198
Y.-N. Kwon et al. / Journal of Membrane Science 415–416 (2012) 192–198
reaction in an oven. Chlorine resistance of the membranes modified under the optimum preparation condition were investigated compared with those of the virgin membrane under three different soaking conditions: 540, 1,620, and 3780 ppm hr Cl2 at pH 7. After exposure of the membranes to the chlorine solutions, the membranes were removed from the chlorine solution and thoroughly rinsed with DI water to remove any unreacted chlorine. Surface-modified membranes allowed less salt passage compared to the virgin membrane after the chlorine soaking test under all three conditions (Fig. 8). This result shows that SPGE coating layer on the PA membrane surface efficiently prevented chlorine attack. To evaluate the degree of chlorination quantitatively, XPS analysis was conducted. The surface chemical compositions of both virgin and surface modified membranes before and after chlorination are listed in Table 2. When the membranes were soaked in chlorine solutions, chlorine was detected that had not been seen before the chlorination. However, the atomic ratio of Cl/N decreased by half after the PA membrane surface was modified with SPGE. All of these results show drastically enhanced chlorine resistance of the surface-modified RO membrane. Fig. 9 shows surface SEM images of virgin membrane and membrane chlorinated at 10,000 ppm hr Cl2 and pH 7. Both the virgin and modified membranes showed a typical ridge and valley surface structure before chlorination. However, an obvious change in surface morphology was observed after chlorination. It is clearly shown that the initial morphology of the virgin membrane was ruptured but the SPGE-coated membrane kept its initial morphology after chlorination.
4. Summary In this study, the preparation of chlorine-resistant RO membranes using the surface coating method was studied and the chlorine resistances of modified membranes were evaluated. A chlorine-resistant coating layer was formed using a multi-functional epoxy compound, SPGE, via ring-opening polymerization. The effects of DMAP, SPGE, and glycerol on membrane permeation properties were investigated. With increasing DMAP solution concentration, water flux increased and salt rejection decreased due to the reduction of further cross-linkable sites. The introduction of neutral SPGE led to decreased salt rejection at lower concentrations. In contrast, at higher SPGE concentrations, molecular overlapping of the polymer led to the formation of a denser coating layer, which resulted in increased salt rejection. Glycerol used as a humectant prevented the membrane from drying out during the ring-opening reaction in an oven and increased membrane performance. The modified membrane was characterized using FTIR, XPS, zeta potential, and contact angle. Each result supported the successful incorporation of an SPGE coating layer on the PA membrane and showed the modification resulted in a membrane surface that was more neutral, hydrophilic, and smooth. The chlorine exposure test using a NaOCl solution showed that the modified surface membranes were much less susceptible to chlorine attack and had enhanced chlorine stability.
Acknowledgments This research was supported by Basic Science Research Program through the National Research Foundation of Korea (NRF)
funded by the Ministry of Education, Science and Technology (2011–0002457). The help and assistance of the Woongjin Chemical Co., Ltd. are greatly acknowledged. References [1] N.N. Li, A.G. Fane, W.S.W. Ho, M. T, Advanced membrane tehcnology and applications, Wiley, 2008. [2] J.E. Cadotte, Interfacially synthesized reverse osmosis membrane U.S. Patent 4,277,344, 1981. [3] J. Kucera, Reverse osmosis, Wiley, 2010. [4] Y.N. Kwon, R. Joksimovic, I.C. Kim, J.O. Leckie, Effect of bromide on the chlorination of a polyamide membrane, Desalination 280 (2011) 80–86. [5] Y.N. Kwon, J.O. Leckie, Hypochlorite degradation of crosslinked polyamide membranes. Changes in hydrogen bonding behavior and performance, J. Membr. Sci. 282 (2006) 456–464. [6] Y.N. Kwon, J.O. Leckie, Hypochlorite degradation of crosslinked polyamide membranes. Changes in chemical/morphological properties, J. Membr. Sci. 283 (2006) 21–26. [7] Y.N. Kwon, C.Y. Tang, J.O. Leckie, Change of membrane performance due to chlorination of crosslinked polyamide membranes, J. Appl. Polym. Sci. 102 (2006) 5895–5902. [8] Y.N. Kwon, C.Y. Tang, J.O. Leckie, Change of chemical composition and hydrogen bonding behavior due to chlorination of crosslinked polyamide membranes, J. Appl. Polym. Sci. 108 (2008) 2061–2066. [9] T. Kawaguchi, H. Tamura, Chlorine-resistant membrane for reverse osmosis. Preparation of chlorine-resistant polyamide composite membranes, J. Appl. Polym. Sci. 29 (1984) 3369–3379. [10] T. Uemura, M. Kurihara, Chlorine resistance of reverse osmosis membranes and changes in membrane structure and performance caused by chlorination degradation, Bull. Soc. Sea Water Sci., Japan 57 (2003) 498–507. [11] J. Glater, M.R. Zachariah, Mechanistic study of halogen interaction with polyamide reverse-osmosis membranes, ACS Symp. Ser. (1985) 345–358. [12] J. Glater, S.K. Hong, M. Elimelech, Search for a chlorine-resistant reverse osmosis membrane, Desalination 95 (1994) 325–345. [13] S.H. Son, J. Jegal, Preparation and characterization of polyamide reverseosmosis membranes with good chlorine tolerance, J. Appl. Polym. Sci. 120 (2011) 1245–1252. [14] R.H. Du, J.S. Zhao, Properties of poly (n,n-dimethylaminoethyl methacrylate)/ polysulfone positively charged composite nanofiltration membrane, J. Mem. Sci. 239 (2004) 183–188. [15] G.D. Kang, C.J. Gao, W.D. Chen, X.M. Jie, Y.M. Cao, Q. Yuan, Study on hypochlorite degradation of aromatic polyamide reverse osmosis membrane, J. Mem. Sci. 300 (2007) 165–171. [16] N. Li, Z.Z. Liu, S.G. Xu, Dynamically formed poly (vinyl alcohol) ultrafiltration membranes with good anti-fouling characteristics, J. Mem. Sci. 169 (2000) 17–28. [17] R.J. Petersen, Composite reverse-osmosis and nanofiltration membranes, J. Mem. Sci. 83 (1993) 81–150. [18] G.D. Kang, M. Liu, B. Lin, Y.M. Cao, Q. Yuan, A novel method of surface modification on thin-film composite reverse osmosis membrane by grafting poly(ethylene glycol), Polymer 48 (2007) 1165–1170. [19] D. Ratna, Handbook of thermoset resins, iSmithers Rapra Publishing, 2009. [20] J.Y. Koo, R.J. Petersen, J.E. Cadotte, Esca characterization of chlorine-damaged polyamide reverse osmosis membrane, Polym. Prepr. Div. Polym. Chem. Am. Chem. Soc. 27 (1986) 391–392. [21] M. Elimelech, X.H. Zhu, A.E. Childress, S.K. Hong, Role of membrane surface morphology in colloidal fouling of cellulose acetate and composite aromatic polyamide reverse osmosis membranes, J. Mem. Sci. 127 (1997) 101–109. [22] C.R. Reiss, J.S. Taylor, C. Robert, Surface water treatment using nanofiltration—pilot testing results and design considerations, Desalination 125 (1999) 97–112. [23] A.I. Schafer, A.G. Fane, T.D. Waite, Nanofiltration of natural organic matter: removal, fouling and the influence of multivalent ions, Water Supply 17 (1999) 191–204.