Letter

Cite This: Environ. Sci. Technol. Lett. 2018, 5, 584−590 pubs.acs.org/journal/estlcu

Polyelectrolyte-Based Sacrificial Protective Layer for Control in Reverse Osmosis Desalination † † ‡ ‡ ‡ Moon Son, Wulin Yang, Szilard S. Bucs, Maria F. Nava-Ocampo, Johannes S. Vrouwenvelder, † and Bruce E. Logan*, † Department of Civil and Environmental Engineering, The Pennsylvania State University, University Park, Pennsylvania 16802, United States ‡ Water Desalination and Reuse Center (WDRC), Division of Biological and Environmental Science and Engineering (BESE), King Abdullah University of Science and Technology (KAUST), Thuwal 23955-6900, Saudi Arabia

*S Supporting Information

ABSTRACT: Reverse osmosis (RO) membranes inevitably foul because of the accumulation of material on the membrane surface. Instead of trying to reduce membrane fouling by chemically modifying the membrane, we took a different approach based on adding a sacrificial coating of two polyelectrolytes to the membrane. After membrane fouling, this coating was removed by flushing with a highly saline brine solution, and a new coating was regenerated in situ to provide a fresh protective layer (PL) on the membrane surface. The utility of this approach was demonstrated by conducting four consecutive dead-end filtration experiments using a model foulant (alginate, 200 ppm) in a synthetic brackish water (2000 ppm of NaCl). Brine removal and regeneration of the PL coating restored the water flux to an average of 97 ± 3% of its initial flux, compared to only 83 ± 3% for the pristine membrane. The average water flux for the PL-coated membranes was 15.5 ± 0.6 L m−2 h−1 until the flux was decreased by 10% versus its initial flux, compared to 13.4 ± 0.5 L m−2 h−1 for the nontreated control. The use of a sacrificial PL coating could therefore provide a more sustainable approach for addressing RO membrane fouling.

■ INTRODUCTION membrane can unintentionally leach off the surface either 24 Membrane fouling inevitably decreases the water flux of a during operation or during aggressive membrane cleaning. reverse osmosis (RO) system because of the accumulation of An alternative approach to deal with membrane fouling is to 1−5 construct the membrane using active layers that are more solids on the membrane. Most RO membranes are 25−32 currently made with a thin film of an active layer of polyamide, chemically resistant than polyamide. Several approaches coated onto a structural support layer. There are four types of have been used to obtain high-performance membranes in fouling: particulate, scaling, organic, and biological.6,7 Partic- terms of permeability and chlorine resistance. However, these ulate and scaling can be minimized by pretreatment (e.g., previous approaches require complex fabrication methods and Downloaded via PENNSYLVANIA STATE UNIV on September 11, 2018 at 20:29:38 (UTC). ultrafiltration) and chemical dosages (e.g., antiscalants).8,9 a large number of active layers to achieve commercial See https://pubs.acs.org/sharingguidelines for options on how to legitimately share published articles. standards of selectivity of >99% rejection of sodium Organic and biofouling are considered the major fouling types 25,26 in practice in RO and nanofiltration water treatment.10,11 chloride. For example, at least 10 bilayers were needed fi to fabricate a RO membrane with a high rejection rate using Membrane modi cation methods used to mitigate fouling of 25,27 the polyamide layer are based on altering the membrane polyelectrolytes. Thermal annealing can be required, which surface,12,13 for example, by making it less hydrophobic,14,15 can result in a large reduction in membrane permeabil- ity.27,33,34 The alternative is to use chemicals such as bonding to the surface to create a steric barrier 25,26 between the membrane and the foulant,16,17 or adding glutaraldehyde to bond multiple layers together, but this materials such as graphene oxides,18 carbon nanotubes,19 and can create the potential for leaching this chemical into the mesoporous carbons20 tothemembranetoreducethe treated water. Without thermal annealing or chemical bonding fi 21 of these layers, the membranes will have low selectivity for the adhesion of foulants on the surface, or using electric elds. ffi Similar strategies can be used to prevent spacer fouling.22,23 and will not have the su cient rejection properties or However, all these approaches delay, but do not prevent, permeabilities needed for RO desalination. fouling so that periodic aggressive cleaning methods are still needed to remove the accumulated foulants that can be tightly Received: August 1, 2018 bound to the membrane surface. Such cleaning methods, for Revised: August 15, 2018 example, with chlorine solutions, are quite costly and can Accepted: August 16, 2018 damage the polyamide layer, and materials added to the Published: August 16, 2018

© 2018 American Chemical Society 584 DOI: 10.1021/acs.estlett.8b00400 Environ. Sci. Technol. Lett. 2018, 5, 584−590 Environmental Science & Technology Letters Letter

Figure 1. Schematic of four steps used to synthesize a replenishable thin-film composite (TFC) membrane: (a) layer-by-layer coating of a protective layer (PL), (b) membrane fouling developed on the PL, (c) detachment of the PL together with accumulated foulant by flushing with high-saline water such as a RO brine, and (d) in situ replenishment of the PL to protect the TFC membrane. Abbreviations: PA, polyamide; PES, poly(ether sulfone); PET, polyethylene terephthalate; RO, reverse osmosis.

Adifferent approach was used here to avoid membrane is a new method for reducing the impact of fouling on fouling based on using multiple polyelectrolyte layers to form a membranes used for RO desalination. sacrificial protective layer (PL) on top of the membrane. The PL was applied without any linker to chemically bond the ■ MATERIALS AND METHODS material to the membrane surface, as the layer was stable on Protective Layer Coating. The two polyelectrolytes used the membrane during operation. When the PL is on the here to produce the PL were a cation , poly(diallyl- membrane surface, organic matter present in a feedwater and dimethylammonium chloride) (PDDA), and an anionic biofilms due to bacterial growth accumulate on the PL, instead polymer, poly(sodium-4-styrenesulfonate) (PSS). PDDA of on the membrane. After the PL is fouled, it is intentionally (CAS Registry No. 26062-79-3) and PSS (CAS Registry No. removed by flushing the membrane with a highly saline 25704-18-1) are not hazardous substances according to the solution, such as an RO brine, which causes the PL to detach safety data sheets provided by Sigma-Aldrich. The first PL was because of the decrease in its stability on the membrane fi 35 applied using a layer-by-layer method to form a uniform lm surface at a high salt concentration. Thus, the problem of (Figure 1a).42 PDDA and PSS (10 g/L each) were dissolved in membrane coating instability under higher saline conditions, deionized (DI) water, and 5 mL of each solution was sprayed which has been considered as a weakness of previous for 1 min to form a single bilayer. A commercial RO polyelectrolyte additions to the membrane in desalination membrane (SW30HR, Dow Chemical) was used as a systems, was used as an advantage here for easy detachment of backbone membrane. Five bilayers were coated onto the the PL. In addition, using the RO brine as a cleaning agent for commercial RO membrane, except as otherwise noted. The ff detachment of the PL is a cost-e ective and environmentally membrane was flushed with 5 mL of DI water for 1 min after friendly method for cleaning the membrane. After cleaning each polyelectrolyte coating to remove any unbound using the brine solution, the PL layer was regenerated in situ to polyelectrolyte. fi produce a new sacri cial layer. While the concept of a Membrane Characterization. Scanning electron micros- sacrificial layer on an ultrafiltration or nanofiltration membrane copy (SEM) was used to analyze the morphology of the 36−40 was previously described, the approach used here is both membranes. Fourier-transform infrared spectroscopy (FTIR) novel and effective even for RO membranes due to the was used to demonstrate the presence of the PL coating. A materials used and the regeneration method. The application scanning electron microscopy/energy dispersive X-ray spec- of multiple layers of did not substantially alter troscopy (SEM−EDS) analysis was used to obtain the membrane permeability, compared to previous results where elemental composition of both the PL-coated and uncoated permeability was greatly reduced.27,33,34 The layers are smooth membranes. The permeability (water flux) and selectivity and can be applied in situ, and they can be removed simply by (rejection) of membranes were obtained as a function of rinsing with a high-salt solution compared to previous methods pressure (1.5−4.1 MPa) using a synthetic brackish water that required a large pH change or cleaning agents such as (2000 ppm of NaCl) under dead-end filtration conditions surfactants.36,38,40,41 Thus, this multiple-application PL coating (Sterlitech Corp., HP4750). The effective membrane area was technology, based on using a nontoxic polyelectrolyte polymer, 14.6 cm2, with the cell pressurized using nitrogen gas.

585 DOI: 10.1021/acs.estlett.8b00400 Environ. Sci. Technol. Lett. 2018, 5, 584−590 Environmental Science & Technology Letters Letter

Figure 2. Physicochemical properties of pristine and precoated membranes. (a) Morphology of the surface as determined by SEM. (b) Functional groups as determined by FTIR. (c) Elemental compositions as determined by SEM−EDS. (d) Permselectivity of prepared membranes in terms of water flux and rejection as a function of pressure. Abbreviations: M, pristine thin-film composite (TFC) membrane; M+PL, TFC membrane possessing a polyelectrolyte-based protective layer (PL); M+PL+Brine, PL-detached membrane achieved by brine flushing after the PL precoating.

Fouling Experiments. Four consecutive fouling experi- DI water (M+DI) or only the high-salt solution (M+Brine). ments were performed using a model foulant (200 ppm of For the PL-coated membrane, the high-salt solution was used alginate) with a calcium binder (100 ppm) and synthetic as a cleaning agent (M+PL+Brine). Flushing was done by brackish water (2000 ppm of NaCl) (Figure 1b). Calcium was stirring (600 rpm) for 10 min. added to accelerate fouling as the calcium ions, which are In Situ Replenishment of the Protective Layer. When positively charged, can bridge the negatively charged alginate the PL was removed by cleaning using the high-salt solution, and membrane surface. The NaCl concentration of the feed the PL was regenerated using an in situ method based on solution was limited to 2000 ppm because of very low applying the PDDA and PSS solutions directly to the permeation of water under dead-end filtration due to the membrane surface in the RO test chamber. PDDA and PSS concentration polarization of salt ions and the pressure that we (1 mL each) were successively added to the membrane surface could obtain in this cell. Dead-end filtration experiments were with a reaction time of 1 min. After each reaction, the solution conducted at 4.1 MPa (600 psi), with the normalized flux, flux was discarded. DI water (1 mL for 1 min) was added to the recovery ratio, and reversible/irreversible fouling ratio membrane surface after each reaction to remove the unbound calculated as previously reported.14,19,43 Briefly, these factors polyelectrolyte. Membranes that were regenerated with a new “ ” were calculated using the initial water flux, the water flux of the PL were indicated by adding Rg to the membrane fouled membrane, and the initial water flux of the cleaned designation (M+PL+Brine+Rg). The membrane with a membrane. Additional experiments were conducted using a regenerated PL layer was further tested under the same lower concentration of alginate (20 ppm) or with stirring (60 fouling and cleaning conditions that were used for other rpm) to examine the impact of concentration polarization membranes (Figure S1a). relative to organic fouling on the water flux. Membrane Cleaning. Cleaning was performed after 3 h of ■ RESULTS AND DISCUSSION fouling using a high-salt solution (a 70000 ppm NaCl solution Membrane Characteristics and Performance. There represents seawater RO brine) (treatment) or DI water was no apparent change in the morphology of the membrane (control for salinity effects) (Figure 1c). For pristine following addition of the PL, based on images obtained using membranes, the membranes were cleaned using either the SEM, likely because each bilayer of the PL coating was

586 DOI: 10.1021/acs.estlett.8b00400 Environ. Sci. Technol. Lett. 2018, 5, 584−590 Environmental Science & Technology Letters Letter

Figure 3. (a) Four consecutive fouling tests using alginate as a model foulant. Cleaning was done after each 3 h of fouling (red arrows). The in situ replenishment of the polyelectrolyte was conducted after cleaning (M, pristine membrane; PL, protective layer; Rg, membrane that was regenerated with a new PL). Water flux decline during (b) second and (c) third cycles of the fouling. See the Supporting Information for the initial water flux tendency of the membranes during the four consecutive fouling tests (Figure S2a). designed to form a <10 nm thick film to minimize the the high-salt solution. For the uncoated membranes, there was reduction in its permeability (Figure 2a). In the FTIR scans, an a flux loss of ∼20% in the second cycle using either DI water additional peak between 1035 and 1040 cm−1 was produced (M+DI) or brine (M+Brine) to clean the membranes (Figure for a PL-treated membrane, indicating the presence of the thin 3a). This showed that irreversible fouling occurred when using PL coating,42,44 and this peak was removed after brine flushing uncoated membranes and that the foulant could not be (Figure 2b). SEM−EDS analysis showed a change in the dislodged by osmotic backwashing because of the salinity elemental composition of the PL compared to that of the differences between the feed and permeate solution (Figure uncoated membranes (Figure 2c). Both functional group S1b). The irreversible fouling ratio of the PL-coated membrane (FTIR) and element (SEM−EDS) analyses therefore indicated was only 3%, whereas it was ∼20% for the uncoated membrane that the PL was successfully coated on the membrane surface, when both membranes were treated with a brine cleaning and it could be removed by brine washing. agent (Figure S3). This result that demonstrated that the The presence of the PL on the membrane (M+PL) slightly severely fouled layer could not be removed by brine washing decreased the membrane permeability and increased the salt alone was different from that reported by others, who indicated rejection compared to those of the pristine membrane (M) that osmotic backwashing alone was an effective cleaning − (Figure 2d). Both water permeability and salt rejection were method for fouled RO membranes.45 47 restored to their initial values after removal of the PL by In successive cycles, there was less flux recovery if the PL washing with the high-salt solution (M+PL+Brine). The was not reapplied after cleaning. However, the membranes that addition of the PL maintained the high salt selectivity of the had undergone in situ replenishment of the PL showed a membrane (99%), unlike previous approaches in which the greater flux recovery after treatment relative to the initial flux membrane with polyelectrolytes added using heat or chemical compared to those of the other membranes over the next two modifications could not achieve this selectivity.25,27 Therefore, fouling cycles due to the brine cleaning and replenishment of the PL could be easily added and detached without appreciably the PL (Figure 3a and Figure S2). An average flux recovery impacting membrane permeability or selectivity. ratio of 97 ± 3% was achieved with the membrane coated with Fouling Control. The performance of the PL-treated the PL (M+PL+Brine+Rg) over four fouling cycles, compared membranes was completely restored following the first to 83 ± 3% for the membrane without the PL (M+Brine). membrane fouling cycle compared to controls. The PL-coated Although PDDA and PSS were chosen here, as they are not membrane (M+PL+Brine) showed 100% water flux recovery toxic chemicals and are easy to apply, other pairs of anionic in the second fouling cycle as the foulant that accumulated on and cationic polymers could also likely be used to fabricate a the PL was washed out together with the PL by washing with PL, such as poly(vinyl alcohol), poly(allylamine hydro-

587 DOI: 10.1021/acs.estlett.8b00400 Environ. Sci. Technol. Lett. 2018, 5, 584−590 Environmental Science & Technology Letters Letter chloride), and sulfonated poly(etherketone). It is expected that presence of the spacer also did not adversely impact the most foulants in a feedwater, such as dissolved organic or application of a sacrificial layer (Figure S6). inorganic matter, particulate matter, or biofilms (biofouling), Outlook. The use of a PL-coated membrane produced could also be removed by this method as the PL should more water than the control did after membrane fouling, with a provide a physical barrier for direct adhesion of these materials higher flux recovery ratio of 97 ± 3% compared to a ratio of 83 to the membrane surface. ± 3% for the control membrane under severe fouling Water Production. The PL-coated membrane produced conditions (200 ppm of alginate). Osmotic backwashing of more water during the initial stage of fouling, because of the the membrane without a PL, by flushing with a solution with a higher flux recovery, than the pristine membrane did (Figure high salt concentration, did not restore the water flux to the 3b,c). In commercial applications of RO membranes for original level. This finding that osmotic backwashing was not desalination, it is typically recommended that the membrane effective for fouling control is different from the results of be cleaned when the decline in the water flux is ∼10%.48,49 previous reports showing a flux recovery ratio of ≤97% due to Therefore, we can examine performance based on only the only osmotic backwashing,52,53 and thus, the impact of this period for the first 10% decline in water flux. During that high-salt solution on membrane fouling should be more period, the PL-coated membrane produced water at an average carefully investigated in future tests. Although the use of a PL flux of 15.5 ± 0.6 L m−2 h−1, compared to a flux of 13.4 ± 0.5 on the membrane was shown here to be an effective method Lm−2 h−1 for the pristine (untreated) membrane. Although for restoring the water flux to its original level, our fouling tests the water fluxes of the PL-coated and uncoated membranes were conducted only under dead-end filtration conditions. were similar when the water flux declined to 50% of its initial Additional cross-flow experiments, with a longer time of value, membrane cleaning would be needed well before that fouling using various water types (e.g., seawater and brackish point in practice. water), will need to be performed to determine the Impact of Chemical Concentrations. Among the ions effectiveness of this approach under conditions more present in seawater, calcium ions play an important role in representative of those used for RO desalination in practice, such as hydraulic conditions, spatial dimensions, and feed membrane fouling as they bridge the membrane surface and − negatively charged foulants such as alginates, making it difficult spacer use.54 56 This coating could be also applied on other to dislodge these foulants.19,50,51 When the calcium ion surfaces such as feed spacers, reservoirs, and pipes and released concentration was doubled in the treated solution, however, as needed to help remove biofilms or accumulated organic the fouling and recovery of the differently treated membranes matter. were the same as those obtained with the original calcium ion concentration (Figure S3). ■ ASSOCIATED CONTENT A very high concentration of the foulant (200 ppm) was *S Supporting Information used here to rapidly foul the membrane. Tests were also The Supporting Information is available free of charge on the conducted at a lower concentration of 20 ppm to examine flux ACS Publications website at DOI: 10.1021/acs.es- recovery under less severe fouling conditions. The same tlett.8b00400. fl reduction in water ux (80%) was obtained even at the lower Details about materials, normalized flux, in situ foulant concentration, indicating that concentration polar- replenishment of the protective layer, osmotic back- ization due to the increased salt concentration over the cycle fl washing, flux decline over cycles for each membrane, was the main reason for the reduction in water ux over time ff ffi (Figure S4a). Following brine treatment and regeneration of e ects of the calcium ion on cleaning e ciencies, the fl the PL, the treated membrane still produced the same 100% role of the foulant and sodium chloride in ux decline, recovery of water flux as that obtained at the higher foulant and applicability and stability of the coating (PDF) concentration (Figure S4a). Additional tests were conducted to reduce concentration polarization by stirring the solution. ■ AUTHOR INFORMATION fl ∼ The decline in water ux was only 10% over the same period Corresponding Author of time (3 h) compared to that in tests without stirring (Figure *E-mail: [email protected]. Telephone: +1-814-863-7908. S4b). Even with stirring, the PL-treated membrane had a flux ORCID recovery higher than that of the control membrane. In the absence of the foulant and stirring, a similar flux decline of Bruce E. Logan: 0000-0001-7478-8070 ≤80% was obtained (Figure S4c), indicating the main factor in Notes the decline in the flux was concentration polarization and not The authors declare no competing financial interest. the alginate in the later stages of fouling. Applicability and Stability under Cross-Flow Con- ■ ACKNOWLEDGMENTS ditions. Although all experiments described above were The authors thank Mr. Woochul Song and Dr. Manish conducted under dead-end filtration conditions, most RO Kumar’s lab at The Pennsylvania State University for the loan and water treatment systems are operated in a cross-flow of the dead-end filtration test device. This research was mode. Therefore, additional tests were performed using a supported by the King Abdullah University of Science and cross-flow system to demonstrate the stability of the coating Technology (KAUST) (OSR-2017-CPF-2907-02) and The with high-shear force and turbulent flow conditions (see the Pennsylvania State University. Supporting Information). The PL layers coated onto the membrane surface during cross-flow operation were found to ■ REFERENCES have a thickness proportional to the number of applied PLs (1) Elimelech, M.; Phillip, W. A. The future of seawater desalination: (Figure S5). Even a single layer of polyelectrolyte was not Energy, technology, and the environment. Science 2011, 333, 712− removed under typical cross-flow operation (Figure S6). The 717.

588 DOI: 10.1021/acs.estlett.8b00400 Environ. Sci. Technol. Lett. 2018, 5, 584−590 Environmental Science & Technology Letters Letter

(2) Shannon, M. A.; Bohn, P. W.; Elimelech, M.; Georgiadis, J. G.; (20) Kim, E.-S.; Deng, B. Fabrication of polyamide thin-film nano- Mariñas, B. J.; Mayes, A. M. Science and technology for water composite (PA-TFN) membrane with hydrophilized ordered purification in the coming decades. Nature 2008, 452, 301−310. mesoporous carbon (H-OMC) for water purifications. J. Membr. (3) Bucs, S. S.; Farhat, N.; Kruithof, J. C.; Picioreanu, C.; van Sci. 2011, 375,46−54. Loosdrecht, M. C. M.; Vrouwenvelder, J. S. Review on strategies for (21) de Lannoy, C.-F.; Jassby, D.; Gloe, K.; Gordon, A. D.; Wiesner, biofouling mitigation in spiral wound membrane systems. Desalination M. R. Aquatic biofouling prevention by electrically charged nano- 2018, 434, 189−197. composite polymer thin film membranes. Environ. Sci. Technol. 2013, (4) Haas, R.; Gutman, J.; Wardrip, N. C.; Kawahara, K.; Uhl, W.; 47, 2760−2768. Herzberg, M.; Arnusch, C. J. Glycosphingolipids enhance bacterial (22) Siddiqui, A.; Lehmann, S.; Bucs, S. S.; Fresquet, M.; Fel, L.; attachment and fouling of nanofiltration membranes. Environ. Sci. Prest, E. I. E. C.; Ogier, J.; Schellenberg, C.; van Loosdrecht, M. C. Technol. Lett. 2015, 2,43−47. M.; Kruithof, J. C.; Vrouwenvelder, J. S. Predicting the impact of feed (5) Guha, R.; Xiong, B.; Geitner, M.; Moore, T.; Wood, T. K.; spacer modification on biofouling by hydraulic characterization and Velegol, D.; Kumar, M. Reactive micromixing eliminates fouling and biofouling studies in membrane fouling simulators. Water Res. 2017, concentration polarization in reverse osmosis membranes. J. Membr. 110, 281−287. − Sci. 2017, 542,8 17. (23) Baek, Y.; Yoon, H.; Shim, S.; Choi, J.; Yoon, J. Electro- (6) Flemming, H.-C. Microbial biofouling: Unsolved problems, conductive feed spacer as a tool for biofouling control in a membrane ffi fi insu cient approaches, and possible solutions. In Bio lm Highlights; system for water treatment. Environ. Sci. Technol. Lett. 2014, 1, 179− Flemming, H.-C., Wingender, J., Szewzyk, U., Eds.; Springer: Berlin, 184. 2011; Vol. 5,pp81−109. ́ (24) Yin, J.; Deng, B. Polymer-matrix nanocomposite membranes for (7) Pontie, M.; Rapenne, S.; Thekkedath, A.; Duchesne, J.; water treatment. J. Membr. Sci. 2015, 479, 256−275. Jacquemet, V.; Leparc, J.; Suty, H. Tools for membrane autopsies (25) Cho, K. L.; Hill, A. J.; Caruso, F.; Kentish, S. E. Chlorine and antifouling strategies in seawater feeds: A review. Desalination − resistant glutaraldehyde crosslinked polyelectrolyte multilayer mem- 2005, 181,75 90. branes for desalination. Adv. Mater. 2015, 27, 2791−2796. (8) Vrouwenvelder, J. S.; Manolarakis, S. A.; Veenendaal, H. R.; van (26) Duong, P. H. H.; Zuo, J.; Chung, T.-S. Highly crosslinked layer- der Kooij, D. Biofouling potential of chemicals used for scale control − by-layer polyelectrolyte FO membranes: Understanding effects of salt in RO and NF membranes. Desalination 2000, 132,1 10. concentration and deposition time on FO performance. J. Membr. Sci. (9) Amy, G.; Ghaffour, N.; Li, Z.; Francis, L.; Linares, R. V.; 2013, 427, 411−421. Missimer, T.; Lattemann, S. Membrane-based seawater desalination: − (27) Park, J.; Park, J.; Kim, S. H.; Cho, J.; Bang, J. Desalination Present and future prospects. Desalination 2017, 401,16 21. membranes from pH-controlled and thermally-crosslinked layer-by- (10) Flemming, H. C.; Schaule, G.; Griebe, T.; Schmitt, J.; −  layer assembled multilayers. J. Mater. Chem. 2010, 20, 2085 2091. Tamachkiarowa, A. Biofouling the Achilles heel of membrane (28) Wang, C.; Such, G. K.; Widjaya, A.; Lomas, H.; Stevens, G.; processes. Desalination 1997, 113, 215−225. Caruso, F.; Kentish, S. E. Click poly(ethylene glycol) multilayers on (11) West, S.; Wagner, M.; Engelke, C.; Horn, H. Optical coherence RO membranes: Fouling reduction and membrane characterization. J. tomography for the in situ three-dimensional visualization and Membr. Sci. 2012, 409−410,9−15. quantification of feed spacer channel fouling in reverse osmosis (29) Kovacs, J. R.; Liu, C.; Hammond, P. T. Spray layer-by-layer membrane modules. J. Membr. Sci. 2016, 498, 345−352. assembled clay composite thin films as selective layers in reverse (12) Miller, D. J.; Araujo,́ P. A.; Correia, P. B.; Ramsey, M. M.; osmosis membranes. ACS Appl. Mater. Interfaces 2015, 7, 13375− Kruithof, J. C.; van Loosdrecht, M. C. M.; Freeman, B. D.; Paul, D. R.; Whiteley, M.; Vrouwenvelder, J. S. Short-term adhesion and long- 13383. term biofouling testing of polydopamine and poly(ethylene glycol) (30) Dizge, N.; Epsztein, R.; Cheng, W.; Porter, C. J.; Elimelech, M. Biocatalytic and salt selective multilayer polyelectrolyte nanofiltration surface modifications of membranes and feed spacers for biofouling − control. Water Res. 2012, 46, 3737−3753. membrane. J. Membr. Sci. 2018, 549, 357 365. (13) Bucs, S. S.; Valladares Linares, R.; Farhat, N.; Matin, A.; Khan, (31) Sanyal, O.; Liu, Z.; Yu, J.; Meharg, B. M.; Hong, J. S.; Liao, W.; Z.; Van Loosdrecht, M. C. M.; Yang, Y.; Wang, M.; Gleason, K. K.; Lee, I. Designing fouling-resistant clay-embedded polyelectrolyte multilayer membranes for wastewater effluent treatment. J. Membr. Kruithof, J. C.; Vrouwenvelder, J. S. Coating of reverse osmosis − membranes with amphiphilic copolymers for biofouling control. Sci. 2016, 512,21 28. Desalin. Water Treat. 2017, 68,1−11. (32) Zhou, Z.; Hu, Y.; Boo, C.; Liu, Z.; Li, J.; Deng, L.; An, X. High- (14) Son, M.; Kim, H.; Jung, J.; Jo, S.; Choi, H. Influence of extreme performance thin-film composite membrane with an ultrathin spray- concentrations of hydrophilic pore-former on reinforced polyether- coated carbon nanotube interlayer. Environ. Sci. Technol. Lett. 2018, 5, − sulfone ultrafiltration membranes for reduction of humic fouling. 243 248. Chemosphere 2017, 179, 194−201. (33) Joseph, N.; Ahmadiannamini, P.; Hoogenboom, R.; (15) Tang, L.; Livi, K. J. T.; Chen, K. L. Polysulfone membranes Vankelecom, I. F. J. Layer-by-layer preparation of polyelectrolyte − modified with bioinspired polydopamine and silver multilayer membranes for separation. Polym. Chem. 2014, 5, 1817 formed in situ to mitigate biofouling. Environ. Sci. Technol. Lett. 2015, 1831. 2,59−65. (34) Xu, G.-R.; Wang, S.-H.; Zhao, H.-L.; Wu, S.-B.; Xu, J.-M.; Li, (16) Carroll, T.; Booker, N. A.; Meier-Haack, J. Polyelectrolyte- L.; Liu, X.-Y. Layer-by-layer (LBL) assembly technology as promising grafted microfiltration membranes to control fouling by natural strategy for tailoring pressure-driven desalination membranes. J. organic matter in drinking water. J. Membr. Sci. 2002, 203,3−13. Membr. Sci. 2015, 493, 428−443. (17) Ulbricht, M.; Belfort, G. Surface modification of ultrafiltration (35) Han, L.; Mao, Z.; Wuliyasu, H.; Wu, J.; Gong, X.; Yang, Y.; membranes by low temperature plasma II. Graft polymerization onto Gao, C. Modulating the structure and properties of poly(sodium 4- polyacrylonitrile and polysulfone. J. Membr. Sci. 1996, 111, 193−215. styrenesulfonate)/poly(diallyldimethylammonium chloride) multi- (18) Chae, H.-R.; Lee, J.; Lee, C.-H.; Kim, I.-C.; Park, P.-K. layers with concentrated salt solutions. Langmuir 2012, 28, 193−199. ̈ Graphene oxide-embedded thin-film composite reverse osmosis (36) Menne, D.; Uzüm, C.; Koppelmann, A.; Wong, J. E.; Foeken, membrane with high flux, anti-biofouling, and chlorine resistance. J. C. v.; Borre, F.; Dahne,̈ L.; Laakso, T.; Pihlajamaki,̈ A.; Wessling, M. Membr. Sci. 2015, 483, 128−135. Regenerable polymer/ceramic hybrid nanofiltration membrane based (19) Son, M.; Choi, H.-g.; Liu, L.; Celik, E.; Park, H.; Choi, H. on polyelectrolyte assembly by layer-by-layer technique. J. Membr. Sci. Efficacy of carbon nanotube positioning in the polyethersulfone 2016, 520, 924−932. support layer on the performance of thin-film composite membrane (37) Ilyas, S.; de Grooth, J.; Nijmeijer, K.; de Vos, W. M. for desalination. Chem. Eng. J. 2015, 266, 376−384. Multifunctional polyelectrolyte multilayers as nanofiltration mem-

589 DOI: 10.1021/acs.estlett.8b00400 Environ. Sci. Technol. Lett. 2018, 5, 584−590 Environmental Science & Technology Letters Letter branes and as sacrificial layers for easy membrane cleaning. J. Interface Sci. 2015, 446, 386−393. (38) Ahmadiannamini, P.; Bruening, M. L.; Tarabara, V. V. Sacrificial polyelectrolyte multilayer coatings as an approach to membrane fouling control: Disassembly and regeneration mecha- nisms. J. Membr. Sci. 2015, 491, 149−158. (39) Shan, W.; Bacchin, P.; Aimar, P.; Bruening, M. L.; Tarabara, V. V. Polyelectrolyte multilayer films as backflushable nanofiltration membranes with tunable hydrophilicity and . J. Membr. Sci. 2010, 349, 268−278. (40) de Vos, W. M.; de Keizer, A.; Stuart, M. A. C.; Kleijn, J. M. Thin polymer films as sacrificial layers for easier cleaning. Surf., A 2010, 358,6−12. (41) Kang, Y.; Zheng, S.; Finnerty, C.; Lee, M. J.; Mi, B. Regenerable polyelectrolyte membrane for ultimate fouling control in forward osmosis. Environ. Sci. Technol. 2017, 51, 3242−3249. (42) Liu, L.; Son, M.; Park, H.; Celik, E.; Bhattacharjee, C.; Choi, H. Efficacy of CNT-bound polyelectrolyte membrane by spray-assisted layer-by-layer (LbL) technique on water purification. RSC Adv. 2014, 4, 32858−32865. (43) Liu, L.; Di, D. Y. W.; Park, H.; Son, M.; Hur, H.-G.; Choi, H. Improved antifouling performance of polyethersulfone (PES) membrane via surface modification by CNTs bound polyelectrolyte multilayers. RSC Adv. 2015, 5, 7340−7348. (44) Reddy, A. V. R.; Mohan, D. J.; Bhattacharya, A.; Shah, V. J.; Ghosh, P. K. Surface modification of ultrafiltration membranes by preadsorption of a negatively charged polymer: I. Permeation of water soluble polymers and inorganic salt solutions and fouling resistance properties. J. Membr. Sci. 2003, 214, 211−221. (45) Liberman, B. Three methods of forward osmosis cleaning for RO membranes. Desalination 2018, 431,22−26. (46) Sagiv, A.; Semiat, R. Backwash of RO spiral wound membranes. Desalination 2005, 179,1−9. (47) Sagiv, A.; Avraham, N.; Dosoretz, C. G.; Semiat, R. Osmotic backwash mechanism of reverse osmosis membranes. J. Membr. Sci. 2008, 322, 225−233. (48) Nitto Denko and Hydranautics. Foulants and cleaning procedures for composite polyamide RO membrane elements. 2011. (49) Filmtec. FILMTEC membranes cleaning and sanitization: Cleaning requirements. (50) Lee, S.; Ang, W. S.; Elimelech, M. Fouling of reverse osmosis membranes by hydrophilic organic matter: implications for water reuse. Desalination 2006, 187, 313−321. (51) Mo, H.; Tay, K. G.; Ng, H. Y. Fouling of reverse osmosis membrane by (BSA): Effects of pH, calcium, magnesium, and temperature. J. Membr. Sci. 2008, 315,28−35. (52) Ramon, G. Z.; Nguyen, T.-V.; Hoek, E. M. V. Osmosis-assisted cleaning of organic-fouled seawater RO membranes. Chem. Eng. J. 2013, 218, 173−182. (53) Bar-Zeev, E.; Elimelech, M. Reverse osmosis biofilm dispersal by osmotic back-flushing: cleaning via substratum perforation. Environ. Sci. Technol. Lett. 2014, 1, 162−166. (54) Bucs, S. S.; Valladares Linares, R.; Marston, J. O.; Radu, A. I.; Vrouwenvelder, J. S.; Picioreanu, C. Experimental and numerical characterization of the water flow in spacer-filled channels of spiral- wound membranes. Water Res. 2015, 87, 299−310. (55) Vrouwenvelder, J. S.; Graf von der Schulenburg, D. A.; Kruithof, J. C.; Johns, M. L.; van Loosdrecht, M. C. M. Biofouling of spiral-wound nanofiltration and reverse osmosis membranes: A feed spacer problem. Water Res. 2009, 43, 583−594. (56) Vrouwenvelder, J. S.; van Paassen, J. A. M.; Wessels, L. P.; van Dam, A. F.; Bakker, S. M. The membrane fouling simulator: A practical tool for fouling prediction and control. J. Membr. Sci. 2006, 281, 316−324.

590 DOI: 10.1021/acs.estlett.8b00400 Environ. Sci. Technol. Lett. 2018, 5, 584−590