w a t e r r e s e a r c h 4 6 ( 2 0 1 2 ) 5 2 1 7 e5 2 2 3 Available online at www.sciencedirect.com journal homepage: www.elsevier.com/locate/watres Effects of hypochlorous acid exposure on the rejection of salt, polyethylene glycols, boron and arsenic(V) by nanofiltration and reverse osmosis membranes Van Thanh Do a, Chuyang Y. Tang a,b,*, Martin Reinhard c, James O. Leckie c a School of Civil and Environmental Engineering, Nanyang Technological University, 50 Nanyang Avenue, N1-1b-35, Singapore 639798, Singapore b Singapore Membrane Technology Centre, Nanyang Technological University, Singapore 639798, Singapore c Department of Civil and Environmental Engineering, Stanford University, Stanford, CA 94305, USA article info abstract Article history: The separation layer of polyamide-based (PA) thin film composite (TFC) membranes can be Received 15 May 2012 modified by active chlorine species. The PA-TFC membranes, NF90, BW30 and NF270, were Received in revised form exposed to different concentrations of sodium hypochlorite (NaOCl) at pH 5 for 24 h. Elemental 22 June 2012 composition obtained from X-ray Photoelectron Spectroscopy (XPS) showed that the chlorine Accepted 26 June 2012 content in the PA layer increased with the chlorine concentrations. Treatment of membranes Available online 7 July 2012 with 10 ppm Cl increased the membrane hydrophilicity. By contrast, when treated with 1000 ppm Cl or more, the membranes became less hydrophilic. Water permeability values for Keywords: all 3 membrane types declined with increased chlorine concentrations. Filtration of poly- Chlorination ethylene glycols (PEGs) with molecular weights of 200, 400 and 600 Daltons (Da) was performed Degradation to investigate the influence of chlorine treatment on membrane molecular weight cut off Hypochlorous acid (MWCO) and rejection by size exclusion. Treatment with 10 and 100 ppm Cl lowered the MWCO Nanofiltration while treatment with higher concentrations increased the MWCO. All chlorinated membranes Reverse osmosis experienced higher NaCl rejection compared to virgin ones. The performance of NF90 was Membrane tested with respect to the rejection of inorganic contaminants including boron (H3BO3) and Arsenic arsenic ðH2 AsO4 Þ. The boron rejection results paralleled PEG rejection whereas those for Boron arsenic followed NaCl rejection patterns. The changes in membrane performance due to Polyethylene glycol chlorine treatment were explained in terms of competing mechanisms: membrane tightening, bond cleavage by N-chlorination and chlorination promoted polyamide hydrolysis. ª 2012 Elsevier Ltd. All rights reserved. 1. Introduction Reverse osmosis (RO) and nanofiltration (NF) membranes have been widely used in desalination and water recycling to meet the increasing demand for fresh water around the world (Drewes et al., 2003; Lee et al., 2010). The past few decades have witnessed remarkable improvements in membrane performance e mainly in the rejection and permeability of PA based TFC membranes (Lee et al., 2010; Li and Wang, 2010). However, the effects of biocides such as chlorine or hypochlorite, which are employed to prevent biofouling or as membrane cleaning agent, on the performance and useful life span of PA TFC * Corresponding author. School of Civil and Environmental Engineering, Nanyang Technological University, 50 Nanyang Avenue, N1-1b-35, Singapore 639798, Singapore. Tel.: þ65 6790 5267. E-mail address: cytang@ntu.edu.sg (C.Y. Tang). 0043-1354/$ e see front matter ª 2012 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.watres.2012.06.044 5218 w a t e r r e s e a r c h 4 6 ( 2 0 1 2 ) 5 2 1 7 e5 2 2 3 membrane are still poorly understood and remain an on-going research (Glater et al., 1994; Van der Bruggen et al., 2008; Cran et al., 2011; Ettori et al., 2011; Zhai et al., 2011). Studies on the response of membrane performance to chlorine contact have led to divergent observations and explanations. Some studies reported that flux and salt rejection of PA membranes increased after treatment with hypochlorite solutions at high pH (above 9) (Jons et al., 1999; Kwon and Leckie, 2006b; Kang et al., 2007). In other studies, it was observed that changes in the PA structure and surface chemistry caused by chlorine treatment lowered membrane flux (Koo et al., 1986; Soice et al., 2004; Kwon and Leckie, 2006b; Simon et al., 2009; Ettori et al., 2011; Do et al., 2012). Mechanistic explanations for these contradictory observations are inconsistent (Glater et al., 1981; Koo et al., 1986; Soice et al., 2003). In order to improve membrane resistance to chlorine degradation, systematic understanding of membrane chlorination and the consequences on membrane performance is critically needed. While data about the influence of membrane chlorination on flux and NaCl rejection are readily available, the effects of membrane chlorination on the passage of other solutes through NF and RO membranes are limited. Investigation on rejection of trace organic compounds by chlorinated NF and RO membranes suggested that changes in rejection are the results of complex interaction between modified membrane properties and the nature of the solutes at filtration conditions (Urase and Sato, 2007; Simon et al., 2009). Inorganic contaminants, such as boron (in form of boric acid and borate ion) and arsenic (commonly exists as arsenate (V) anion) are also important contaminants of concern in desalination (Macedonio and Drioli, 2008). Taniguchi et al. (2001) found a correlation between the rejection of NaCl and boron for the chlorinated cross-linked fully-aromatic PA UTC-80 membranes (Toray Industries, Japan) but no explanation was offered to rationalize their observations. Zhai et al. (2011) reported that hypochlorite treated PA RO membranes at pH 9 could experience enhanced boron rejection. However, the study focused on the effects of hypochlorite concentrations and filtration feed pH on rejection rather than on explanation for rejection mechanism. To the best of our knowledge, there are no reports in the literature on the effect of chlorine on arsenic rejection and on the molecular weight cut off (MWCO) determined by PEG. The goal of this study was to elucidate the impacts of chlorine exposure on different rejection mechanisms of RO and NF membranes. Changes in rejection by size exclusion were investigated by filtration of neutral solutes (PEGs and boron) while changes in rejection by charge repulsion were studied with NaCl and As(V). Furthermore, the importance of chlorination-promoted hydrolysis (Do et al., 2012) on both major solute (NaCl), PEGs and trace contaminants (boron and As(V)) will be discussed in detail. a piperazine (PIP) based semi-aromatic NF membrane, and NF90 and BW30, two fully-aromatic NF and RO membranes, respectively (Tang et al., 2009). The surface of BW30 is polyvinyl alcohol (PVA) coated (Tang et al., 2007). All membranes were stored at 4  C in the dark until used. 2.1.2. 2.2. Materials and methods 2.1. Materials and chemicals 2.1.1. Polyamide membranes Three different commercial PA-TFC membranes from Dow FilmTec (Minneapolis, MN, USA) were used: NF270, Membrane degradation protocol Membrane coupons were degraded in NaOCl solutions according to the protocol described previously (Do et al., 2012). Briefly, exact total chlorine concentration of the soaking solution, which is the sum of all active chlorine species (White, 1986), was determined by titration with sodium thiosulfate standard and reported as ppm of Cl equivalent (Eaton et al., 1995). The pH was adjusted to 5 e a typical set point in plant operation (Bartels and Wilf, 2005) by the addition of concentrated HCl or NaOH not more than 24 h before use. Since pKa of hypochlorous acid is w7.5 (Kwon and Leckie, 2006a), at this soaking pH 5, the main active species is HOCl. Membranes were rinsed and soaked in MilliQ water for 24 h to remove surface impurity and preservatives before degradation tests. The coupons were pre-soaked in NaOCl solutions at the same testing conditions for 1 min to remove excess water. The pre-soaked coupons were immersed in Wheaton bottles containing NaOCl solutions of 10, 100, 1000 and 2000 ppm total Cl for 24 h at room temperature (w 21  C) to achieve accelerated laboratory-scale chlorine degradation (Kwon and Leckie, 2006b; Simon et al., 2009; Ettori et al., 2011). Bottles were constantly shaken and covered with aluminum foil to prevent photochemical degradation of chlorine and radical reactions. 2.3. Membrane surface analysis Before characterization tests, membrane samples were thoroughly rinsed with MilliQ water and vacuum dried for at least 48 h. 2.3.1. 2. Chemicals Unless specified otherwise, all reagents and chemicals were of analytical grade with purity over 99%. Sodium thiosulfate used in chlorine titration, sodium hypochlorite (w10% NaOCl, reagent grade) used in membrane degradation and boric acid used in filtration tests were purchased from Sigma Aldrich (St. Louis, MO, USA). Disodium hydrogen arsenate heptahydrate was obtained from Alfa Aesar (Ward Hill, MA, USA). Poly(ethylene glycol)s e PEGs standards for calibration curves were obtained from Varian Inc. (Santa Clara, CA, USA). PEGs used for rejection tests, sodium chloride and concentrated hydrochloric acid were purchased from Merck (Darmstadt, Germany). MilliQ water (Millipore, Billerica, MA, USA) was used in all preparations and experiments. X-ray Photoelectron Spectroscopy (XPS) Elemental analysis of membrane surface chemistry was performed by a Kratos AXIS Ultra XPS spectrometer (Shimadzu, Columbia, MD, USA) with a monochromatic aluminum Ka X-ray source at 1486.7 eV and a 3.6 eV electron flood gun to compensate for membrane surface charging. Survey spectra were averaged from 3 scans per sample, over the range of 0e1000 eV at 1 eV resolution. The elemental binding energy was calibrated with w a t e r r e s e a r c h 4 6 ( 2 0 1 2 ) 5 2 1 7 e5 2 2 3 the reference energy of carbon 1 s at 285 eV (Beamson and Briggs, 1992). Relative sensitivity factors (RSF) of 0.78, 0.477, 0.278 and 0.891 were used for O 1 s, N 1 s, C 1 s and Cl 2p peaks, respectively. 2.3.2. Contact angle Contact angles, which indicate the wettability of the membrane surface, were obtained using the sessile drop method. Tangent lines to both sides of a 10 mL MilliQ water droplet were measured by a Dataphysics Instruments OCA Goniometer (Filderstadt, Germany). The reported contact angle of each membrane sample was the average of 40 measurements (2 independent membrane coupons and 20 different locations for each sample) with errors indicated by the standard deviations. 2.4. Evaluation of membrane performance 2.4.1. NF/RO filtration system The custom-assembled high pressure cross-flow filtration system in Fig. 1 consists of identical rectangular CF042 cells (Delrin Acetal, Sterlitech, Kent, WA, USA) in parallel arrangement. Active membrane area was 42 cm2 (4.6 cm  9.2 cm), and spacers of 1.2 mm thickness from GE Osmonics (Minnetonka, MN, USA) were used for all filtration tests. Both permeate and concentrate were re-circulated back to the feed tank and the temperature of the feed solution was controlled at 21  1  C by a Polysciene chiller (Niles, IL, USA). 2.4.2. Water permeability and sodium rejection Prior to permeability and rejection (performance) tests, virgin membranes (soaked in MilliQ water for 24 h) and chlorinetreated membranes were cleaned thoroughly with MilliQ water. For all performance tests (including PEG rejection in Section 2.4.3 and boron and arsenic rejection in Section 2.4.4), cross-flow velocity of the system was 22.6 cm/s; operating pressures for NF90, BW30 and NF270 were set at 100, 260 and 70 psi, respectively, to obtain similar water flux (at around 1.4e1.7  10 5 m/s) for virgin membranes. Feed solution of 10 mM NaCl in MilliQ water was circulated in the system at pH w6.5. The reported water permeability and NaCl rejection were recorded after 24 h of membrane compaction. Permeate flux, Jw was determined by weight. Salt rejection was calculated from permeate and tank conductivity measured by Ultrameter II conductivity meter (Myron L Company, Fig. 1 e Schematic diagram of the NF/RO filtration system. PRV: pressure relief valve, P: pressure gauge. 5219 Carlsbad, CA, USA). At least two separate runs were performed for each treatment. The water permeability coefficient, A (m/s Pa) was determined from the water flux (Jw) measurements and the applied pressure difference (DP) by the following equations (Mulder, 1996): A¼ Jw DP (1) Dp Dp ¼ iRT Cb Cp
(2) where Dp is the osmotic pressure difference across the membrane, i is the dimensionless van’t Hoff factor, Cb and Cp are the bulk and permeate salt concentrations, respectively. The apparent solute rejection, R was determined from the solute concentrations of the permeate, Cp and the feed, Cf by: R¼1 Cp Cf (3) Diamond-patterned feed spacer was used together with a relatively high cross flow velocity (w22.6 cm/s) to minimize the effect of concentration polarization. 2.4.3. PEG rejection After 24 h of membrane compaction, the system was flushed with MilliQ water for 5 min and the feed was switched to one of three PEG solutions with average molecular weights of 200, 400 and 600 g/mol, each with a concentration of 5.5 g/L. For each PEG feed, 10 mL of PEG samples of both permeates and feed were taken after 10 min filtration. After each PEG run, the system was flushed with fresh MilliQ water for 15 min to clean the setup. At least 1 replicate was produced for each test condition. PEG concentrations were determined by gel permeation chromatography (GPC) performed on a Varian Inc. PL-GPC 50 Plus system (Santa Clara, CA, USA) with a refractive index detector. The system used PL Aquagel-OH 20 GPC column (Varian Inc.) with packing material of 5 mm diameter, MilliQ water as solvent and was set at 30  C. 2.4.4. Boron and arsenic (V) rejection After 24 h of membrane compaction, boric acid and disodium hydrogen arsenate heptahydrate were added to achieve feed concentration of 20 mg B/L and 150 mg As/L. Concentrated HCl and NaOH were used to adjust the feed pH to 7. Boron concentration was determined by Perkin Elmer Optima 2000 inductively coupled plasma e optical emission spectrometer (ICP-OES) (Zaventem, Belgium). Arsenic concentration was analyzed by Agilent 7700 ICP-MS (Tokyo, Japan). The reported rejection was determined after treating boron and arsenate solutions continuously for 24 h. 3. Results and discussion 3.1. Membrane surface characterization 3.1.1. Surface chlorine composition In Fig. 2, the chlorine content on the membrane surface (in atomic percentages, %Cl) was obtained using XPS survey 5220 w a t e r r e s e a r c h 4 6 ( 2 0 1 2 ) 5 2 1 7 e5 2 2 3 3.1.2. Fig. 2 e Surface chlorine atomic percent of NF90, BW30 and NF270 membranes exposed to different chlorine concentrations for 24 h at pH 5. scans and is plotted against chlorine exposure (for 24 h at pH 5). The measured Cl content is mainly contributed by chlorine covalently bonded to the polyamide (e.g., NeCl bonding), and there is negligible chloride anion absorbed on the membrane surface (Do et al., 2012). Generally, the results show an increase in the chlorine content with the increasing exposure with some exceptions. For NF90, the chlorine contents after exposure to 1000 and 2000 ppm were near identical, 9.92% and 10.11%, respectively, indicating a saturation point between 100 and 1000 ppm for the chlorine uptake of the uncoated aromatic PA membranes (Soice et al., 2003; Kwon et al., 2008). Compared to NF90, the chlorine uptake of BW30 after exposure to 10 and 100 ppm for 24 h was lower, which is attributed to the PVA surface coating layer of BW30. However, when treated with 1000 and 2000 ppm chlorine, the Cl contents of BW30 (9.86% and 12.76%, respectively) were equal to and higher than those of NF90, indicating that PVA coating does not protect BW30 effectively against attack by high chlorine concentration (Do et al., 2012). The chlorine uptake by the piperazine-based NF270 membrane increased nearly proportionally with exposure but was always lower than that by the fully-aromatic membranes. This suggests that the tertiary nitrogen in the semi-aromatic PA is not readily to react with chlorine (Jensen et al., 1999; Soice et al., 2003). The degree of membrane degradation can be evaluated by the atomic O/N ratio, which is an indicator of cross-linking in the PA layer (Tang et al., 2007; Coronell et al., 2008). According to Tang et al. (2007), O/N ratio is 1.0 for 100% cross-linking when all O and N form amide groups; while a 2.0 ratio indicates fully linear PA chains with one free carboxyl group for every 2 amide groups. Calculation of the O/N ratios for virgin and treated membranes is presented in Supplementary Data, Table S1. With increasing chlorine exposure, the O/N ratio increased, indicating that cross-linkages decreased, which may be attributed to induced membrane hydrolysis due to hypochlorite attack of the amide nitrogen. Loss of crosslinking of the PA rejection layer may lead to the shortening of membrane life span. Membrane wettability The wettability of a membrane is assessed by contact angle measurement, whereby a lower contact angle indicates that the surface is more hydrophilic or more polar. The contact angles of all membranes in Fig. 3 decreased after 10 ppm chlorination, indicating an increased surface wettability. By contrast, exposure to 1000 and 2000 ppm chlorine treatment increased contact angles, indicating that membrane surfaces became more hydrophobic and less wettable than virgin membranes. The different effects of chlorination on wettability are suggested to be the results of two competing processes: chlorination and hydrolysis of amide groups. Higher contact angle or reduced hydrophilicity is caused by the incorporation of chlorine onto the surface by chlorination (Koo et al., 1986; Kwon and Leckie, 2006a; Simon et al., 2009). In our previous study (Do et al., 2012), it was observed that Nchlorination can promote hydrolysis of the amide CeN bond to form more hydrophilic carboxyl eCOOH groups and thus increased membrane hydrophilicity. 3.2. Membrane performance 3.2.1. Water permeability The effect of membrane chlorination on the water permeability of NF90, BW30 and NF270 membranes is shown in Fig. 4. The permeability decreased steadily in all cases. For severe degradation conditions (1000 and 2000 ppm), significant permeability decline can be the result of increased hydrophobicity (Fig. 3), which made the membrane surface more difficult to be wetted (Koo et al., 1986). Meanwhile, mild chlorinated membranes (at 10 and 100 ppm) can experience the tightening effects, which is the formation of additional linkage via azo-compounds on the surface due to chlorination causing the membrane to be less permeable (Soice et al., 2003, 2004). Alternatively, Kwon et al. suggested that the loss of hydrogen bonds between amidic hydrogen and the carbonyl groups in the polymer chains due to chlorination can lead to chain compaction and restrict water passage (Kwon and Leckie, 2006b; Kwon et al., 2006). 3.2.2. PEG rejection In order to investigate the impact of chlorination on rejection by size exclusion, PEGs of 3 different molecular weights (200, Fig. 3 e Contact angles of NF90, BW30 and NF270 membranes exposed to different chlorine concentrations for 24 h at pH 5. w a t e r r e s e a r c h 4 6 ( 2 0 1 2 ) 5 2 1 7 e5 2 2 3 Fig. 4 e Water permeability of NF90, BW30 and NF270 membranes after 24 h filtration. Membranes were exposed to different chlorine concentrations for 24 h at pH 5. Operating pressures for NF90, BW30 and NF270 were set at 100, 260 and 70 psi. 400 and 600 Da) were employed in filtration tests. These organic compounds are widely used to determine the MWCO of the membrane since they are neutral and non-polar and their rejections are solely by size exclusion mechanism (Schafer et al., 2005; López-Muñoz et al., 2009). Rejection of PEGs by virgin and chlorinated NF90 membranes was plotted against the molecular weights of the PEGs in Fig. 5, and the results for BW30 and NF270 are presented in Supplementary Data, Fig. S1. The MWCO (corresponding to a 90% PEG rejection) for the virgin NF90 membrane was w200 Da and is within the range reported in literature (López-Muñoz et al., 2009). At mild chlorination conditions (10 and 100 ppm), enhancement in rejection of NF90 membranes (e.g., 95% compared to 89% PEG 200 rejection of virgin membrane) can be attributed to tightening effects (Section 3.2.1). PEG 200 rejection of severely degraded NF90 membranes (1000 and 2000 ppm) declined to 78% and 60%, respectively. In the case of 1000 and 2000 ppm chlorine exposure, the MWCO of the chlorinated NF90 membrane increased to w400 and 600 Da, respectively; perhaps due to polymer chain degradation and the creation of a more open polyamide structure in the rejection layer. This result is consistent with the increased O/N elemental ratio from our XPS analysis (and thus reduced polyamide crosslinking degree, see Section 3.1.1), which suggests that the membrane structure was damaged as a result of CeN bond cleavage due to chlorination promoted hydrolysis (Do et al., 2012). Similar changes were observed in case of BW30 and NF270. 3.2.3. 5221 Fig. 5 e PEG rejection of virgin and chlorinated NF90. Membranes were exposed to 10, 100, 1000 and 2000 ppm of chlorine for 24 h at pH 5. Error bars represent the range of duplicate measurements. NF90 at 10 and 100 ppm to 98.0% and 98.4% compared to 85.3% of virgin membrane. Highly chlorinated NF90 membranes (1000 and 2000 ppm) rejected NaCl by 93% and 90%, respectively, significantly better than the virgin membrane but not as high as the mildly chlorinated ones. Apparently, the highly chlorinated membranes did not benefit from the tightening effect. In addition, the PEG rejection results revealed increased MWCO of the chlorinated membrane (Fig. 5). Therefore, the better rejection is the result of the enhanced negative surface. A comparison of the NaCl rejection at mild chlorination condition (10 and 100 ppm) and that at severe chlorination conditions (1000 and 2000 ppm) confirms that the PA polymer structure was damaged at the more severe degradation conditions. The increased NaCl rejection by the BW30 membrane can be attributed to the enhanced surface charge. However, the increase is marginal, probably due to the presence of the NaCl rejection In our previous study, chlorination promoted hydrolysis of CeN bond incorporates more eCOOH groups on the membrane surface and therefore lowers the surface charge of the membranes (Do et al., 2012). More negative surface charges improved rejection of NaCl by charge repulsion in all the membranes, as indicated in Fig. 6. This is in contrast to PEG rejection (Section 3.2.2), which decreased with increasing chlorine exposure. The enhance surface negativity, together with tightening effect, increased rejection of mild chlorinated Fig. 6 e NaCl rejection of NF90, BW30 and NF270 membranes after 24 h filtration. Membranes were exposed to different chlorine concentrations for 24 h at pH 5. Operating pressures for NF90, BW30 and NF270 were set at 100, 260 and 70 psi. 5222 w a t e r r e s e a r c h 4 6 ( 2 0 1 2 ) 5 2 1 7 e5 2 2 3 neutral PVA coating layer, which still partially covers the PA rejection layer and moderates the enhanced charge effect. For the loose NF270 membrane, enhanced surface negativity has contributed significantly to rejection by charge repulsion, even at high exposure conditions. The structural damage due to chlorination could not be observed in the case of NaCl filtration as clearly as in the case of PEGs filtration in Section 3.2.2. 3.2.4. Arsenic (V) and boron rejection of NF90 Rejection tests of arsenic (V) anion and boron were performed for virgin and chlorinated NF90 to assess whether the effects of membrane chlorination on trace inorganic contaminants rejection can be correlated with the changes in charge repulsion and size exclusion caused by chlorination observed in Sections 3.2.2 and 3.2.3. Arsenic acid has pKa1 of 2.2 and pKa2 of 7.8 (Macedonio and Drioli, 2008), and boric acid has pKa value of 9.25 (Koseoglu et al., 2008). At filtration pH 7, As (V) exists as anion and is hypothesized to be rejected by charge repulsion effect in addition to size exclusion while the rejection mechanism for uncharged boron is hypothesized to be size exclusion. Rejection data obtained after 24 h filtration is presented in Fig. 7. Data show that boron rejection has similar characteristic to PEG rejection of chlorinated NF90 e it increased at low chlorine exposure (10 ppm) but reduced at severe exposure (1000 and 2000 ppm). The severely compromised boron rejection at 2000 ppm chlorine exposure can be attributed to the increased MWCO of the membrane as a result of chlorination induced hydrolysis (Section 3.2.2). The rejection of As (V) shares a similar pattern with NaCl rejection, which stayed relatively high in spite of the increased MWCO, and may therefore be interpreted as the effect of enhanced charge repulsion. The slight reduction of As rejection at 2000 ppm chlorine exposure may be attributed to the loss of PA cross-linking due to hydrolysis of CeN bonds. The results obtained from the current study have major implications in understanding solute rejection by chlorinated membranes. Shifts in membrane rejection are caused by competing mechanisms (tightening effect, charge repulsion, and chlorination induced hydrolysis). Neutral solutes seem to be more adversely affected under severe chlorination due to 1) membrane hydrolysis (which dismisses the size exclusion effect) and 2) the lack of electrostatic repulsion. While NaCl is typically used as a standard test solute for membrane rejection, enhanced NaCl rejection can benefit from charge interaction but does not necessarily suggest improved rejection of other (neutral) trace contaminants, many of which may be of environmental and health concerns. Future research shall further explore the role of such competing effects on the rejection of trace organic contaminants. 4. Conclusions Chlorination and chlorination promoted hydrolysis of PA membranes change the physiochemical properties and the chemical structure of the active layer. The type and extent of these changes depend on the chlorine concentration. Importantly, this study suggests that mild chlorination can improve membrane rejection. Analysis of the rejection behavior of charged (NaCl, arsenate (V)) and neutral (boron and PEG) solutes revealed the type of changes caused by chlorination. The data confirmed that neutral solutes are rejected by size exclusion mechanism while charge species are predominantly rejected by charge repulsion. Membranes chlorinated under mild conditions (100 ppm and below) showed lower flux but better rejection than virgin ones (due to tightening effect and a slightly enhanced surface charge). The performance of highly chlorinated membranes (1000 ppm and above) was consistent with the previously observed competing effects of chlorination and hydrolysis (Do et al., 2012). Incorporation of Cl created more hydrophobic surfaces causing the flux to decrease. On the other hand, chlorination promoted hydrolysis introduced more negative carboxyl groups on membrane surface and improved rejection of charged solutes. However, it also cleaved the amide bonds and caused polyamide depolymerization, which increased membrane permeability for neutral species. Acknowledgments This research was financially supported by the Tier 1 Research Grant #RG6/07, Ministry of Education, Singapore. Van Thanh Do is funded by the Singapore Stanford Partnership Program. The authors acknowledge the Singapore Membrane Technology Centre for technical support and Dow FilmTec for providing the membrane samples. Dr. Richard Webster and Ms. Bahareh Khezri (School of Physical and Mathematical Sciences, NTU) are thanked for assistance with ICP-MS measurements. Ms. Nguyen Thanh Hang is greatly appreciated for her extensive laboratory assistance. Fig. 7 e Rejection of arsenic (V) and boron of virgin and chlorinated NF90 membranes after 24 h filtration. Chlorinated membranes were exposed to 10, 100, 1000 and 2000 ppm of chlorine for 24 h at pH 5. Error bars represent the range of duplicate measurements. Appendix A. Supplementary material Supplementary data associated with this article can be found in the online version, at http://dx.doi.org/10.1016/j.watres. 2012.06.044. w a t e r r e s e a r c h 4 6 ( 2 0 1 2 ) 5 2 1 7 e5 2 2 3 references Bartels, C.R., Wilf, M., 2005. Design considerations for wastewater treatment by reverse osmosis. Water Science and Technology 51, 473e482. Beamson, G., Briggs, D., 1992. High Resolution XPS of Organic Polymers: The Scienta ESCA 300 Database. Wiley, New York. Coronell, O., Mariñas, B.J., Zhang, X., Cahill, D.G., 2008. Quantification of functional groups and modeling of their ionization behavior in the active layer of FT30 reverse osmosis membrane. Environmental Science & Technology 42 (14), 5260e5266. Cran, M.J., Bigger, S.W., Gray, S.R., 2011. Degradation of polyamide reverse osmosis membranes in the presence of chloramine. Desalination 283, 58e63. Do, V.T., Tang, C.Y., Reinhard, M., Leckie, J.O., 2012. Degradation of polyamide nanofiltration and reverse osmosis membranes by hypochlorite. Environmental Science & Technology 46 (2), 852e859. Drewes, J.E., Reinhard, M., Fox, P., 2003. Comparing microfiltration-reverse osmosis and soil-aquifer treatment for indirect potable reuse of water. Water Research 37 (15), 3612e3621. Eaton, A.D., Clesceri, L.S., Greenberg, A.E., 1995. Standard Methods for Examination of Water and Wastewater. APHA, AWWA and WEF, Washington DC, US. Ettori, A., Gaudichet-Maurin, E., Schrotter, J.-C., Aimar, P., Causserand, C., 2011. Permeability and chemical analysis of aromatic polyamide based membranes exposed to sodium hypochlorite. Journal of Membrane Science 375 (1e2), 220e230. Glater, J., Hong, S.-k., Elimelech, M., 1994. The search for a chlorine-resistant reverse osmosis membrane. Desalination 95 (3), 325e345. Glater, J., McCutchan, J.W., McCray, S.B., Zachariah, M.R., 1981. In: Turbak, A.F. (Ed.), Synthetic Membranes. American Chemical Society, pp. 171e190. Jensen, J.S., Lam, Y.-F., Helz, G.R., 1999. Role of amide nitrogen in water chlorination: proton NMR evidence. Environmental Science & Technology 33 (20), 3568e3573. Jons, S.D., Stutts, K.J., Ferritto, M.S., Mickols, W.E., 1999. In: U.P.a.T. (Ed.), Treatment of Composite Polyamide Membranes to Improve Performance. Office. Kang, G.-D., Gao, C.-J., Chen, W.-D., Jie, X.-M., Cao, Y.-M., Yuan, Q., 2007. Study on hypochlorite degradation of aromatic polyamide reverse osmosis membrane. Journal of Membrane Science 300 (1e2), 165e171. Koo, J.-Y., Petersen, R.J., Cadotte, J.E., 1986. ESCA characterization of chlorine-damaged polyamide reverse osmosis membrane. ACS Polymer Preprints 27 (2), 391e392. Koseoglu, H., Kabay, N., Yüksel, M., Sarp, S., Arar, Ö., Kitis, M., 2008. Boron removal from seawater using high rejection SWRO membranes e impact of pH, feed concentration, pressure, and cross-flow velocity. Desalination 227 (1e3), 253e263. Kwon, Y.-N., Leckie, J.O., 2006a. Hypochlorite degradation of crosslinked polyamide membranes: I. Changes in chemical/ morphological properties. Journal of Membrane Science 283 (1e2), 21e26. Kwon, Y.-N., Leckie, J.O., 2006b. Hypochlorite degradation of crosslinked polyamide membranes: II. Changes in hydrogen bonding behavior and performance. Journal of Membrane Science 282 (1e2), 456e464. 5223 Kwon, Y.-N., Tang, C.Y., Leckie, J.O., 2008. Change of chemical composition and hydrogen bonding behavior due to chlorination of crosslinked polyamide membranes. Journal of Applied Polymer Science 108 (4), 2061e2066. Kwon, Y.N., Tang, C.Y., Leckie, J.O., 2006. Change of membrane performance due to chlorination of crosslinked polyamide membranes. Journal of Applied Polymer Science 102, 5895. Lee, K.P., Arnot, T.C., Mattia, D., 2010. A review of reverse osmosis membrane materials for desalination - Development to date and future potential. Journal of Membrane Science 370 (1e2), 1e22. Li, D., Wang, H., 2010. Recent developments in reverse osmosis desalination membranes. Journal of Materials Chemistry 20 (22), 4551e4566. López-Muñoz, M.J., Sotto, A., Arsuaga, J.M., Van der Bruggen, B., 2009. Influence of membrane, solute and solution properties on the retention of phenolic compounds in aqueous solution by nanofiltration membranes. Separation and Purification Technology 66 (1), 194e201. Macedonio, F., Drioli, E., 2008. Pressure-driven membrane operations and membrane distillation technology integration for water purification. Desalination 223 (1e3), 396e409. Mulder, M., 1996. Basic Principles of Membrane Technology. Kluwer Academic Publishers, Dordrecht, The Netherlands. Schafer, A.I., Fane, A.G., Waite, T.D., 2005. Nanofiltration: Principles and Applications. Elsevier, Oxford. Simon, A., Nghiem, L.D., Le-Clech, P., Khan, S.J., Drewes, J.E., 2009. Effects of membrane degradation on the removal of pharmaceutically active compounds (PhACs) by NF/RO filtration processes. Journal of Membrane Science 340 (1e2), 16e25. Soice, N.P., Greenberg, A.R., Krantz, W.B., Norman, A.D., 2004. Studies of oxidative degradation in polyamide RO membrane barrier layers using pendant drop mechanical analysis. Journal of Membrane Science 243 (1e2), 345e355. Soice, N.P., Maladono, A.C., Takigawa, D.Y., Norman, A.D., Krantz, W.B., Greenberg, A.R., 2003. Oxidative degradation of polyamide reverse osmosis membranes: studies of molecular model compounds and selected membranes. Journal of Applied Polymer Science 90 (5), 1173e1184. Tang, C.Y., Kwon, Y.-N., Leckie, J.O., 2007. Probing the nano- and micro-scales of reverse osmosis membranes: a comprehensive characterization of physiochemical properties of uncoated and coated membranes by XPS, TEM, ATR-FTIR, and streaming potential measurements. Journal of Membrane Science 287 (1), 146e156. Tang, C.Y., Kwon, Y.-N., Leckie, J.O., 2009. Effect of membrane chemistry and coating layer on physiochemical properties of thin film composite polyamide RO and NF membranes: I. FTIR and XPS characterization of polyamide and coating layer chemistry. Desalination 242 (1e3), 149e167. Taniguchi, M., Kurihara, M., Kimura, S., 2001. Boron reduction performance of reverse osmosis seawater desalination process. Journal of Membrane Science 183 (2), 259e267. Urase, T., Sato, K., 2007. The effect of deterioration of nanofiltration membrane on retention of pharmaceuticals. Desalination 202 (1e3), 385e391. Van der Bruggen, B., Mänttäri, M., Nyström, M., 2008. Drawbacks of applying nanofiltration and how to avoid them: a review. Separation and Purification Technology 63 (2), 251e263. White, G.C., 1986. The Handbook of Chlorination. Van Nostrand Reinhold Co., New York. Zhai, X., Meng, J., Li, R., Ni, L., Zhang, Y., 2011. Hypochlorite treatment on thin film composite RO membrane to improve boron removal performance. Desalination 274 (1e3), 136e143.