Environ. Sci. Technol. 2009, 43, 3865–3871
High-Throughput Membrane Surface Modification to Control NOM Fouling MINGYAN ZHOU,† HONGWEI LIU,| J A M E S E . K I L D U F F , * ,† R O B E R T L A N G E R , ‡ DANIEL G. ANDERSON,§ AND GEORGES BELFORT| Department of Civil and Environmental Engineering, Rensselaer Polytechnic Institute, Troy, New York 12180-3590, Department of Chemical Engineering, Massachusetts Institute of Technology, 77 Massachusetts Avenue, Cambridge, Massachusetts 02139, David H. Koch Institute for Integrative Cancer Research, Massachusetts Institute of Technology, 77 Massachusetts Avenue, Cambridge, Massachusetts 02139, and Howard P. Isermann Department of Chemical and Biological Engineering and the Center for Biotechnology and Interdisciplinary Studies, Rensselaer Polytechnic Institute, Troy, New York 12180-3590
Received February 5, 2009. Revised manuscript received March 29, 2009. Accepted March 30, 2009.
A novel method for synthesis and screening of foulingresistant membrane surfaces was developed by combining a highthroughput platform (HTP) approach together with photoinduced graft polymerization (PGP) for facile modification of commercial poly(aryl sulfone) membranes. This method is an inexpensive, fast, simple, reproducible, and scalable approach to identify fouling-resistant surfaces appropriate for a specific feed. In this research, natural organic matter (NOM)-resistant surfaces were synthesized and indentified from a library of 66 monomers. Surfaces were prepared via graft polymerization onto poly(ether sulfone) (PES) membranes and were evaluated using an assay involving NOM adsorption, followed by pressure-driven filtration. In this work new and previously tested low-fouling surfaces for NOM are identified, and their ability to mitigate NOM and protein (bovine serum albumin) fouling is compared. The bestperforming monomers were the zwitterion [2-(methacryloyloxy)ethyl]dimethyl-(3-sulfopropyl)ammonium hydroxide, and diacetone acrylamide, a neutral monomer containing an amide group. Other excellent surfaces were synthesized from amides, amines, basic monomers, and long-chain poly(ethylene) glycols. Bench-scale studies conducted for selected monomers verified the scalability of HTP-PGP results. The results and the synthesis and screening method presented here offer new opportunities for choosing new membrane chemistries that minimize NOM fouling.
* Corresponding author address: Rensselaer Polytechnic Institute, Department of Civil and Environmental Engineering, 4022 JEC Building, 110 eighth Street, Troy, NY 12180; phone: 518.276.2042; fax: 518.276.4833; e-mail:
[email protected]. † Department of Civil and Environmental Engineering, Rensselaer Polytechnic Institute. | Howard P. Isermann Department of Chemical and Biological Engineering and the Center for Biotechnology and Interdisciplinary Studies, Rensselaer Polytechnic Institute. ‡ Department of Chemical Engineering, Massachusetts Institute of Technology. § David H. Koch Institute for Integrative Cancer Research, Massachusetts Institute of Technology. 10.1021/es9003697 CCC: $40.75
Published on Web 04/16/2009
2009 American Chemical Society
1. Introduction Membrane fouling is a significant factor that must be considered in all applications of membrane filtration, including water production, and low fouling membranes are needed to maximize production (permeate throughput), minimize costs and environmental impacts of cleaning, and reduce process energy requirements (1). Developing new materials with surface or functional characteristics that minimize foulant adsorption or adhesion involves great effort, expense, and time. In part, this is because surface science has not developed sufficiently to predict the performance of new surfaces, and in part because many feed solutions, including natural waters (containing NOM), are complex multicomponent solutions that can interact with surfaces in multiple ways. Here, we report on a new approach, using a highthroughput platform (HTP), to develop new polymeric membrane materials for specific filtration applications (2). This process was developed by adapting high throughput platform approaches successfully used in chemistry (e.g., combinatorial spot/well analysis) (3) and biology (e.g., phage display and SELEX) (4, 5) to the modification of poly(ether sulfone), using photoinduced graft polymerization (PGP) (6). The combined HTP-PGP method is an inexpensive, fast, simple, and reproducible way to quickly synthesize and evaluate new polymeric surfaces, and identify optimum surfaces for specific feed solutions. Although HTP approaches have been used to evaluate process variables in the development of filtration processes using commercially available filtration membranes (7, 8) and to optimize membrane casting dope composition (9), to our knowledge, this is the first time an HTP approach is used to synthesize and screen fouling-resistant surfaces. We use the HTP-PGP method to modify poly(ether sulfone), in part, because it has excellent physical and transport characteristics, although its surface chemistry is not ideal for many feed solutions. During photoinduced graft polymerization, poly(aryl sulfone) membranes are UVirradiated, cleaving trunk polymer chains and forming reactive radical sites (6, 10). Either water or ethanol-soluble vinyl monomers covalently bond to these radical sites and undergo free-radical polymerization. A schematic illustration of the mechanism is shown in Figure 1. In contrast to some other free-radical polymerization methods, such as atom transfer radical polymerization, no initiator or catalyst is required. The goals of this work were to search for new antifoulant membrane chemistries against NOM and to demonstrate that our new high-throughput synthesis and screening method can select for such membranes. This method has been validated in previous work using bovine serum albumin (BSA) as the foulant; in this work the synthesis and evaluation of new and previously tested low-fouling membrane surfaces for NOM are described, and the best surfaces for NOM and BSA are compared. In addition, bench-scale studies with NOM as the foulant were conducted for selected monomers to verify HTP-PGP results.
2. Experimental Materials and Methods Detailed experimental materials and methods are provided in Supporting Information; they are briefly summarized below. An initial monomer library was assembled using a pool of 66 commercially available vinyl monomer candidates (Sigma-Aldrich, Saint Louis, MO). These monomers were categorized into nine groups based on chemical functionality, VOL. 43, NO. 10, 2009 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
3865
FIGURE 1. In the PGP method, poly(aryl sulfone) membranes are UV-irradiated (λ ≈ 300 nm), cleaving trunk polymer chains and forming reactive radical sites. Vinyl monomers chemically bond to these radical sites and undergo free-radical polymerization. After Yamagishi et al. (6), and Taniguchi et al. (10). as shown in Figure 2. Monomers were then employed to modify PES surfaces using commercially available 96-well filter plates; each well having a permeable bottom made from a 100 kDa cutoff PES membrane (Seahorse Labware, Chicopee, MA). Membranes were modified via UV-induced graft polymerization as described in previous publications (e.g., refs 6and 10, and references therein). Modified PES surfaces and nonmodified controls were challenged with Elliott Soil Humic Acid (International Humic Substance Society, St. Paul, MN) employed as a model NOM having a high fouling potential to allow rapid screening and identification of promising surfaces. Fouling was evaluated using a static adsorption protocol; this approach was shown to yield similar results as measuring the resistance to foulant solution filtration (2). The hydraulic resistance of the membrane in each well was measured before and after modification, and after fouling using a vacuum manifold apparatus. Selected results from the HTP-PGP experiments were verified in bench scale experiments using a dead-end stirred cell filtration system.
3. Results and Discussion The resistance to water permeation after modification (but prior to fouling), Rmod, relative to the resistance of the asreceived membrane, RAR, represents the factor by which membrane resistance increased after modification, and is a rough indicator of the amount of grafted material (10). To assess NOM/surface interactions, a fouling index, R, was calculated as the resistance increase of grafted membranes caused by fouling normalized by that of the ungrafted membrane control, R ) ∆Rmod/∆Rcontrol, where ∆Rmod ) (Rfouled - R)mod and ∆Rcontrol ) (Rfouled - R)control. The control was the as-received membrane treated with either water (in which case Rcontrol ) RAR) or ethanol, depending on which solvent was used to dissolve the monomer. The increase in the modified membrane resistance after NOM adsorption should be lower than that of the control when the modified surface resists NOM interactions. For filtration applications involving permeation, the resistance of the modified membranes should be near that of the as-received membrane (Rmod ≈ RAR), although a higher resistance may be acceptable when it correlates with higher NOM rejection. Static Adsorption of NOM by Modified Membrane Surfaces. Figure S1, Supporting Information, shows a photograph of a representative 96-well filter plate after 3866
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 43, NO. 10, 2009
modification and subsequent NOM adsorption. The well color provides a qualitative visual indication of the tendency of each surface to interact with the NOM feed. Fouling Index and Different Monomer Classes. The fouling index, which provides a quantitative measure of fouling, was grouped into 7 categories as shown by the criteria in Table S1, Supporting Information. The distribution of monomer ratings within each monomer group is shown in Figure 3. No surfaces made from aromatic monomers performed better than the as-received membrane; however, all other groups had at least one surface that performed better. The monomer classes having the greatest number of high performing surfaces were the poly(ethylene glycol) (PEG) monomers, the amines, and the basic and zwitterionic monomers. The poorest performing surfaces were synthesized from a hydrophobic methacrylate (isobornyl methacrylate, no. 19), acrylic acid (no. 43), and two monomers (representing hydroxyl and acid groups) containing a phosphoric acid ester. Indentification of Specific Low Fouling Chemistries (Monomers). In Figure 4, monomers having a fouling index less than the as-received membrane are ranked. It is clear that the HTP approach has identified many surfaces that perform significantly better than the as-received membrane in terms of offering lower resistance caused by fouling. As will be discussed in more detail below, some of these surfaces are new, whereas others are known, thus providing validating of the method. In Figure 5 the value of the fouling index is plotted against the relative resistance after modification. The axis values were chosen to most clearly illustrate the majority of the data having a fouling index less than one; most of these surfaces exhibited a relative resistance in the range 0.5 < Rmod/RAR < 2. It is clear that many surfaces offer lower fouling than the as-received membrane without exhibiting significantly greater resistance. Two exceptional surfaces were obtained, but are not apparent in Figure 5 because they have a fouling index significantly less than zero (4+ rating). One was synthesized from a zwitterionic monomer (no. 60), and the other was synthesized from an acrylamide (no. 53). Nine excellent surfaces (with ratings 3+) were obtained, with three from the amine group (nos. 51, 52, 55); two from the PEG group (nos. 34, 35); one basic monomer (no. 61); one zwitterionic monomer (no. 59); one methacrylate having hydrophobic side chains (no. 7); and one acid (no. 45). Note that the N-isopropylacrylamide (no. 51) is not apparent in Figure 5 because it exhibited a very high resistance. The features of the best performing surfaces for NOM will be summarized next, including a comparison of their ability to resist fouling by NOM and BSA and their previous applications in the literature, followed by a discussion of their conformance to criteria for surfaces having low affinity for proteins. The best-performing monomer, receiving a 4+ rating, was the zwitterion [2-(methacryloyloxy)ethyl]dimethyl-(3-sulfopropyl)ammonium hydroxide, monomer no. 60, also known as a sulfobetaine. Self-assembled monolayers (SAMs) of alkanethiols with exposed zwitterionic head groups have been investigated as protein resistance surfaces based on cell-surface lipids as biological precedents, together with previously successful grafted phosphocholine derivatives; such surfaces have been shown to resist adsorption of fibrinogen and lysozyme (11). The [2-(methacryloyloxy)ethyl]dimethyl-(3-sulfopropyl)ammonium hydroxide zwitterion was also employed by Sun et al. (12) to fabricate ultrafiltration membranes from a copolymer blend with poly(acrylonitrile). As compared to pure poly(acrylonitrile), the copolymer hydrophilicity was increased, fouling by BSA was lowered, and flux recovery after water cleaning was improved. The
FIGURE 2. Initial monomer library consisting of 66 monomers in nine groups. Superscript a indentifies ethanol soluble monomers. After Zhou et al. (2).
FIGURE 3. Number of monomers with specific ratings for different monomer classes for fouling-resistant surfaces due to the challenge from NOM. The monomers were rated according to the fouling index R ) ∆Rmod/∆Rcontrol ) (Rfouled - R)mod/ (Rfouled - R)control, where “4+” ) exceptional, R < 0; “3+” ) excellent, 0 < R < 0.6; “2+” ) good, 0.6 < R < 0.8; “1+” ) fair, 0.8 < R < 1.0; “1-” ) poor, 1.0 < R < 1.8; “2- ” ) very poor, 1.8 < R < 3.0; “3-” ) extremely poor, R > 3.0; “ND” ) not determined. good performance of this monomer in our work is consistent with a hydrated surface conformation as discussed by Azzaroni et al. (13). This monomer resisted BSA fouling to a greater degree than as-received PES, but only received a fair (1+) rating (2). The second-ranking monomer, also receiving a 4+ rating, was diacetone acrylamide (no. 53). This monomer has appeared in the literature, but not for its surface properties or for applications involving NOM or protein adsorption, cell adhesion, or filtration. Therefore, the identification of this monomer represents the ability of the HTP approach to identify new surfaces for specific filtration applications. This
FIGURE 4. Surfaces synthesized using the HTP-PGP method that resist NOM fouling from a total of 66 commercial monomers. Surfaces are ranked relative to the as-received poly(ether sulfone) membrane. Success is measured in terms of a fouling index, R ) ∆Rmod/∆Rcontrol ) (Rfouled - R)mod/(Rfouled - R)control. Monomers 60 and 53 had fouling indices equal to -3.357 and -1.578, respectively, and are not plotted to scale. Surface chemistries not plotted exhibited either fouling greater than the as-received membrane, or high membrane resistance. See Figure 2 for monomer structures. monomer also resisted BSA fouling, receiving an excellent (3+) rating (2). The next nine monomers all received excellent (3+) ratings. The best of these was N-isopropylacrylamide (no. 51), another neutral monomer containing a secondary amine VOL. 43, NO. 10, 2009 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
3867
FIGURE 5. Performance of surfaces synthesized using the HTP-PGP in terms of the fouling index, and increase in resistance to DI water filtration after modification. Monomers: [) PEGs; b ) basic and zwitterionic; 2 ) amines; 9 ) heteroring; ∆ ) hydroxyl; 0 ) acid; ] ) hydrophobic methacrylates. As-received membrane is shown as a cross at coordinates (1,1). in an amide group. Surfaces grafted with N-isopropylacrylamide have exhibited protein adsorption properties that vary with its temperature-dependent conformation (14, 15). At temperatures above its lower critical solution temperature (32 °C) (14), polymer chains were collapsed and protein adsorbing, whereas below this temperature they were hydrated and protein repellent (15). At the temperature of our work, it is likely that polymer chains were hydrated, which provides an explanation for the excellent (3+) rating at resisting both NOM and BSA fouling (2). The fourth-ranked monomer was the basic monomer [2-(methacryloyloxy)ethyl] trimethylammonium chloride (no. 61). This monomer has been widely investigated in the literature as a component of polymers and hydrogels (16), to control surface chemistry and chromatographic properties of monolithic capillary columns (17), to make cationic flocculant to remove anionic dyes (18), and to increase the hydrophilicity of porcine heart valves so vital endothelial cells can grow on the implant surfaces (19). In our work, this monomer performed very well for NOM, with a rating of 3+, while for BSA, it performed poorly, with a ranking of 1- (2). This monomer provides an example of how the HTP approach can identify feed specific surfaces, and illustrates that the general guidelines for protein resistant surfaces are not necessarily general for feeds such as NOM. The next monomers in the ranking were both long-chain poly(ethylene glycol) methyl ether methacrylates (nos. 34 and 35 with 22 and 45 repeating units, respectively); these performed quite well at producing low-fouling surfaces, receiving excellent (3+) ratings for both NOM and BSA (2) foulants. Many researchers have observed that PEG surfaces resist nonspecific adsorption of proteins; this property has resulted in their wide use in biomedical devices. Therefore, their good performance in our HTP experiments helps to build confidence in the technique. The properties of PEGgrafted surfaces thought to be responsible for the exclusion of proteins include a high degree of hydration, conformational flexibility and high chain mobility (20). It has been suggested that foulant molecules are rejected by steric repulsion forces (21), although the importance of surface hydrophilicity has been challenged (22). Protein resistance has been observed to increase with density and chain length of surface grafted PEGs (11), a finding that is consistent with the results of this work. PEGs have been incorporated into microfiltration membranes as comb copolymers with poly(vinylidene fluoride) (PVDF) (23). A similar approach was taken to make thin film composite membranes (24). The ability of such mem3868
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 43, NO. 10, 2009
branes to resist fouling was attributed to hydrogen bonding between water and the ether oxygen groups of the PEG side chains. PEGs have also been covalently grafted onto PVDF microfiltration membrane surfaces by reverse atom transfer radical polymerization (25). The tertiary amine 2-(dimethylamino)ethyl methacrylate (no. 55) performed well at producing low-fouling surfaces, also receiving an excellent (3+) rating for both NOM and BSA (2) foulants. This monomer is sensitive to solution pH and ionic strength; increasing either leads to a conformation switch from a stretched brush to a collapsed state. The collapsed state enhances hydrophilicity and protein-resistance of the grafted surfaces, due to a higher surface enrichment of ester groups (26). A weak polyelectrolyte ultrafiltration membrane based on poly(acrylonitrile and 2-dimethylamino ethyl methacrylate) copolymer was reported (27); however, effects on fouling were not examined. Monomer 2-ethylhexyl methacrylate (no. 7) ranked eighth and exhibited performance that was unexpected, because its structure contains fewer hydrophilic groups than the other high performing monomers. This monomer received an excellent (3+) rating for both NOM and BSA (2) foulants. Membranes made from an “alloy” of polyurethane and a copolymer of 2-ethylhexyl methacrylate and a phospholipid polymer showed a reduced insulin and fibroblast adhesion (28). Monomer N-tert-butylacrylamide (no. 52) is another neutral monomer containing a secondary amine in an amide group, and is structurally similar to N-isopropylacrylamide, monomer no. 51. It did not perform as well as that monomer, but did receive an excellent (3+) rating for NOM and a good (2+) rating for BSA (2). It has also been suggested as a candidate for controlling adhesion of proteins and cells via thermoresponsive surface properties (29). Zwitterionic monomer [3-(methacryloylamino)propyl]dimethyl(3-sulfopropyl) ammonium hydroxide inner salt (no. 59) is structurally related to monomer no. 60, and also performed well, receiving an excellent (3+) rating for NOM and a good (2+) rating for BSA (2). Cho et al. (30) showed that surfaces grafted with this monomer exhibited resistance to the nonspecific adsorption of proteins, comparable to that of the best known systems such as PEG-like films. The lowest ranking monomer to receive a 3+ rating, methacrylic acid (no. 45) improved the fouling resistance of PES toward NOM, but significantly decreased the resistance by BSA, with a rating of 3- (2). This monomer provides a second example of how the HTP approach can identify feedspecific antifouling surfaces. Methacrylic acid has been widely used to tailor polymer surface properties to produce smart surfaces for biomaterials, sensors, actuators and textiles (31, 32). It has been used to control the wetting and adsorption capacity of proteins and dyes on poly(ethylene terephthalate) membranes (33), to increase the wettability and reduce fouling of PVDF and PES filtration membranes (34, 35). Conformance to Criteria for Surfaces Having Low Affinity for Proteins. A good deal of research has focused on the development of surfaces that resist strong interactions with proteins, and guidelines for protein-resistant surfaces have been developed based on studies of protein interactions with self-assembled monolayers of alkanethiolates on gold as model substrates (11, 36). Such guidelines cannot be assumed to apply directly to natural organic matter; indeed, they probably do not apply to all proteins. Even so, they represent an important point of reference that must be considered when developing new membrane surfaces. General features of surfaces having low affinity for proteins include: (i) they are hydrophilic (wettable), (ii) they contain hydrogen bond acceptors, (iii) they lack hydrogen bond donors, (iv) they are electrically neutral (11, 36). As will be discussed further, many of the surfaces synthesized con-
formed to the general guidelines for protein-resistant surfaces for both BSA and NOM, although some exceptions were noted. As expected, several high-performing surfaces, including the highest ranked surface, were synthesized from zwitterions; although charged, they conform to the net neutrality criterion. Several high-performing surfaces, including the secondranking monomer, diacetone acrylamide (no. 53), were synthesized from monomers containing a secondary amine in an amide group. Although the group can act as a hydrogen bond donor, others have noted that primary and secondary amines adsorb more protein than structurally similar groups in the form of amides (15). The strongly basic monomer [2-(methacryloyloxy)ethyl] trimethylammonium chloride (no. 61) is expected to carry a positive charge over a wide range of pH values. Surface synthesized from this monomer performed poorly for BSA, in conformance with the guidelines (2). However, it performed well for NOM; the reasons for this will require further investigation. The neutral and hydrophilic long-chain poly(ethylene glycol) methyl ether methacrylates performed well, as expected. The neutral tertiary amine 2-(dimethylamino)ethyl methacrylate (no. 55) performed well at producing low-fouling surfaces for both NOM and BSA foulants and conformed to the protein resistant surface guidelines. Monomer 2-ethylhexyl methacrylate (no. 7) performed better than expected, because its structure appears much less hydrophilic than the other high performers. The negatively charged methacrylic acid (no. 45) surface improved the fouling resistance of PES toward NOM, probably by charge repulsion, but significantly decreased the fouling resistance by BSA (2), conforming to the guideline that protein resistant surfaces should be neutral. Some monomers (notably, basic monomer no. 61, and acidic monomer no. 45) did not conform to the neutral criterion, and performed poorly for BSA, but performed well for NOM, illustrating that the guidelines for protein resistant surfaces are not necessarily general for feeds such as NOM. The differences between the results of BSA and NOM likely resulted from differences in their structure (e.g., distribution of functional groups) and the structural and chemical heterogeneity of NOM as compared to BSA. Bench-Scale Verification of HTP Results. Several of the most promising monomers identified in the high-throughput experiments were tested in bench-scale stirred-cell filtration experiments to assess the scalability of the results and whether surfaces that exhibited low NOM interactions after adsorption would also be favorable for filtration applications. One of the best fouling-resistant surfaces in the high throughput experiments, monomer no. 53, diacetone acrylamide, yielded permeabilities too low to allow evaluation at the bench scale; grafting conditions for this monomer require further optimization. Note that the membranes employed for the HTP and bench-scale experiments were both made from PES and had a MWCO of 100 kDa. However, the manufacturer of the 96-well plates did not report the membrane source; therefore, it is possible the membranes used at the two scales were different. Also, calcium was added to the NOM solution to provide a rigorous evaluation of fouling resistance, because it is known that divalent cations can exacerbate fouling (37). Filtration data are plotted in Figure 6 in terms of resistance versus cumulative permeate volume throughput to minimize the effects of differences in the initial grafted membrane resistance. As shown in Figure 6, the filtration data for several surfaces clearly show significantly improved performance as compared to the as-received PES membrane (i.e., the lower absolute value of the total fouling resistance and decreased slope), as predicted by the high-throughput experimental results. The zwitterions [3-(methacryloylamino)propyl]dimethyl(3sulfopropyl) ammonium hydroxide inner salt (no. 59) and
FIGURE 6. Bench-scale verification of HTP discovery. Total fouling ∆RF () RF - RM) during filtration of NOM (with mixing) is plotted as a function of cumulative permeate volume throughput. Performance plotted in this way minimizes dependence on initial membrane resistance. [2-(methacryloyloxy)ethyl]dimethyl-(3-sulfopropyl) ammonium hydroxide (no. 60) did not perform well at the bench scale in the presence of calcium; it is possible that the divalent cation interacted with and neutralized the sulfonate group, either resulting in a net positively charged and attractive surface for negatively charged NOM, smaller hydration forces, a more poorly hydrated surface conformation, or some combination of these. Surprisingly, the long-chain poly(ethylene glycol) methyl ether methacrylates (no. 35) did not perform as well at the bench scale as it did at the HTP scale when calcium was present. It is known that PEG chains can chelate cations (38-40); molecular dynamics simulations of a PEG chain have shown that as a result of potassium cation binding, the helical conformation of the chain was considerably distorted, hydration was perturbed, PEG-water pair interaction energy was decreased, and chain flexibility was reduced, leading to salting-out from aqueous solution (38). Chai and Klein (39) showed experimentally that PEG adsorbed to mica surfaces via an interlayer of hydrated or partly hydrated potassium ions. Finally, Bailey and Koleske (40) reported on the salting out of 4000 kDa PEG solutions with calcium chloride; the effect was attributed to a collapsing of the polymer coil in solution, accompanied by a decrease in the hydrodynamic volume. Therefore, we speculate that calcium reduced the PEG graft layer conformational flexibility and mobility, making it more susceptible to adsorptive fouling. Basic monomer [2-(methacryloyloxy)ethyl] trimethylammonium chloride (no. 61) performed well at both the HTP and bench scales. In addition, basic monomers [3-(methacryloylamino)propyl] trimethylammonium chloride (no. 62) and [2-(methacryloyloxy)ethyl] trimethylammonium methyl sulfate (no. 63) also performed well at both scales. The tertiary amine 2-(dimethylamino)ethyl methacrylate (no. 55) performed well at the HTP scale, and produced the best performing surface at the bench scale. Our bench-scale results are consistent with the known ability of this monomer to adopt a collapsed conformation that enhances hydrophilicity of the grafted surfaces in response to increasing ionic strength (26). The novel HTP-PGP method presented here is an inexpensive, fast, simple, reproducible and scalable approach which can be used to synthesize and screen foulingresistant surfaces appropriate for a specific feed. We have successfully demonstrated the method for one modification condition, although optimum conditions are likely to vary for different monomers. One approach to address this issue is to conduct initial screening experiments as done here VOL. 43, NO. 10, 2009 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
3869
(possibly extended to several different monomer concentrations) to identify promising surfaces, followed by experiments that optimize conditions for smaller groups of monomers. However, it should also be noted that in previous work (2) the relative ranking of different surfaces prepared using the HTP approach was not very sensitive to monomer concentration over a wide range (,0.10 M to >0.5 M). The main objective of the current work was to identify surface chemistries that minimize interactions with feed components as a way to mitigate the initial stages of fouling. The resistance after static adsorption likely represents the fouling potential of membrane surfaces in terms of feed component affinity, chemistry and structure. This approach should incorporate pore blockage and pore constriction that results from solute adsorption to the membrane surface and pore walls. However, this approach does not incorporate all fouling mechanisms, nor does it predict the effects of hydrodynamics, for which further methods development will be needed. Work is underway to employ the HTP-PGP method to identify surfaces that resist fouling by a wider range of natural water components, including high molecular weight polysaccharides. Work is also underway to assess how monomer cost should be included as a factor in monomer selection, because although membrane performance is the primary criterion, it is recognized that the market for water treatment membranes is competitive, and costs should be minimized. Future work will also involve using chemoinformatics algorithms to search for correlations between molecular structure and performance.
Acknowledgments The authors acknowledge the U.S. Environmental Protection Agency (EPA grant RD83090901-0), U.S. National Science Foundation (CBET-0730449) for financial support, and Ron Lipski, Seahorse Labware (Chicopee, MA), for filter plates.
Supporting Information Available Detailed experimental methods and materials, a table showing monomer rating criteria and a photograph of a representative 96-well filter plate after modification and subsequent NOM adsorption are provided. This material is available free of charge via the Internet at http://pubs.acs.org.
Literature Cited (1) Goosen, M. F. A.; Sablani, S. S.; Al-Hinai, H.; Al-Obeidani, S.; Al-Belushi, R.; Jackson, D. Fouling of reverse osmosis and ultrafiltration membranes: A critical review. Sep. Sci. Technol. 2004, 39, 2261–2297. (2) Zhou, M.; Liu, H.; Kilduff, J. E.; Langer, R.; Anderson, D. G.; Belfort, G. High throughput synthesis and screening of new protein resistant surfaces for membrane filtration. AIChE J. 2009, Submitted for publication. (3) Bannwarth, W.; Felder, E. Combinatorial Chemistry: A Practical Approach; Wiley-VCH: Weinheim, Germany, 2000. (4) Clackson, T.; Hoogenboom, H. R.; Griffiths, A. D.; Winter, G. Making antibody fragments using phage display libraries. Nature (London) 1991, 352, 624. (5) Gold, L.; Brown, D.; He, Y.-Y.; Shtatland, T.; Singer, B. S.; Wu, Y. From oligonucleotide shapes to genomic SELEX: Novel biological regulatory loops. Proc. Natl. Acad. Sci. U.S.A. 1997, 94, 59. (6) Yamagishi, H.; Crivello, J. V.; Belfort, G. Development of a novel photochemical technique for modifying poly (arylsulfone) ultrafiltration membranes. J. Membr. Sci. 1995, 105, 237–247. (7) Jackson, N. B.; Liddell, J. M.; Lye, G. J. An automated microscale technique for the quantitative and parallel analysis of microfiltration operations. J. Membr. Sci. 2006, 276, 31–41. (8) Chandler, M.; Zydney, A. High throughput screening for membrane process development. J. Membr. Sci. 2004, 237, 181– 188. 3870
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 43, NO. 10, 2009
(9) Bulut, M.; Gevers, L. E. M.; Paul, J. S.; Vankelecom, I. F. J.; Jacobs, P. A. Directed development of high-performance membranes via high-throughput and combinatorial strategies. J. Comb. Chem. 2006, 8, 168–173. (10) Taniguchi, M.; Pieracci, J.; Samsonoff, W. A.; Belfort, G. UVassisted graft polymerization of synthetic membranes: Mechanistic studies. Chem. Mater. 2003, 15, 3805–3812. (11) Holmlin, R. E.; Chen, X.; Chapman, R. G.; Takayama, S.; Whitesides, G. M. Zwitterionic SAMs that resist nonspecific adsorption of protein from aqueous buffer. Langmuir 2001, 17, 2841–2850. (12) Sun, Q.; Su, Y.; Ma, X.; Wang, Y.; Jiang, Z. Improved antifouling property of zwitterionic ultrafiltration membrane composed of acrylonitrile and sulfobetaine copolymer. J. Membr. Sci. 2006, 285, 299–305. (13) Azzaroni, O.; Brown, A. A.; Huck, W. T. S. UCST wetting transitions of polyzwitterionic brushes driven by self-association. Angew. Chem., Int. Ed. 2006, 45, 1770–1774. (14) Shamim, N.; Hong, L.; Hidajat, K.; Uddin, M. S. Thermosensitivepolymer-coated magnetic nanoparticles: Adsorption and desorption of bovine serum albumin. J. Colloid Interface Sci. 2006, 304, 1–8. (15) Cole, M. A.; Jasieniak, M.; Voelcker, N. H.; Thissen, H.; Horn, R.; Griesser, H. J. Switchable surface coatings for control over protein adsorption. SPIE, International Society for Optical Engineering, Bellingham WA, WA 98227-0010, United States; Adelaide, Australia, 2007; Vol. 6416, p 641606. (16) Mun, G. A.; Nurkeeva, Z. S.; Khutoryanskiy, V. V.; Sergaziyev, A. D.; Rosiak, J. M. Radiation synthesis of temperatureresponsive hydrogels by copolymerization of [2-(methacryloyloxy)ethyl]trimethylammonium chloride with N-isopropylacrylamide. Radiat. Phys. Chem. 2002, 65, 67–70. (17) Eeltink, S.; Hilder, E. F.; Geiser, L.; Svec, F.; Frechet, J. M. J.; Rozing, G. P.; Schoenmakers, P. J.; Kok, W. T. Controlling the surface chemistry and chromatographic properties of methacrylate-ester-based monolithic capillary columns via photografting. J. Sep. Sci. 2007, 30, 407–413. (18) Shen, J.-J.; Ren, L.-L.; Zhuang, Y.-Y. Interaction between anionic dyes and cationic flocculant P(AM-DMC) in synthetic solutions. J. Hazard. Mater. 2006, 136, 809–815. (19) Chang, B.-J.; Prucker, O.; Groh, E.; Wallrath, A.; Dahm, M.; Ruhe, J. Surface-attached polymer monolayers for the control of endothelial cell adhesion. Colloids Surf., A 2002, 198-200, 519– 526. (20) Zalipsky, S.; Harris, J. M. Introduction to chemistry and biological applications of poly(ethylene glycol). ACS Symp. Ser. 1997, 680, 1–13. (21) Sofia, S. J.; Merrill, E. W. Protein adsorption on poly(ethylene oxide)-grafted silicon surfaces. ACS Symp. Ser. 1997, 680, 342– 360. (22) Blume, G.; Cevc, G. Molecular mechanism of the lipid vesicle longevity in vivo. Biochim. Biophys. Acta 1993, 1146, 157–168. (23) Chen, Y.; Ying, L.; Yu, W.; Kang, E. T.; Neoh, K. G. Poly(vinylidene fluoride) with grafted poly(ethylene glycol) side chains via the RAFT-mediated process and pore size control of the copolymer membranes. Macromolecules 2003, 36, 9451–9457. (24) Akthakul, A.; Salinaro, R. F.; Mayes, A. M. Antifouling polymer membranes with subnanometer size selectivity. Macromolecules 2004, 37, 7663–7668. (25) Chen, Y.; Deng, Q.; Xiao, J.; Nie, H.; Wu, L.; Zhou, W.; Huang, B. Controlled grafting from poly(vinylidene fluoride) microfiltration membranes via reverse atom transfer radical polymerization and antifouling properties. Polymer 2007, 48, 7604– 7613. (26) Su, Y.; Li, C. The reorientation of poly(2-dimethylamino ethyl methacrylate) after environment stimuli improves hydrophilicity and resistance of protein adsorption. J. Colloid Interface Sci. 2007, 316, 344–349. (27) Su, Y.; Li, C. Tunable water flux of a weak polyelectrolyte ultrafiltration membrane. J. Membr. Sci. 2007, 305, 271–278. (28) Ishihara, K.; Tanaka, S.; Furukawa, N.; Kurita, K.; Nakabayashi, N. Improved blood compatibility of segmented polyurethanes by polymeric additives having phospholipid polar groups. I. Molecular design of polymeric additives and their functions. J. Biomed. Mater. Res. 1996, 32, 391–399. (29) Rochev, Y.; O’Halloran, D.; Gorelova, T.; Gilcreest, V.; Selezneva, I.; Gavrilyuk, B.; Gorelov, A. Rationalising the design of polymeric thermoresponsive biomaterials. J. Mater. Sci.: Mater. Med. 2004, 15, 513–517.
(30) Cho, W. K.; Kong, B.; Choi, I. S. Highly efficient non-biofouling coating of zwitterionic polymers: Poly((3-(methacryloylamino)propyl)-dimethyl(3-sulfopropyl)ammonium hydroxide). Langmuir 2007, 23, 5678–5682. (31) Chen, H. Smart Nanotechnology in Biomaterials, Sensors, Actuators and Textiles; Proceedings of the 2005 AIChE Annual Meeting, Cincinnati, OH; American Institute of Chemical Engineers, New York, NY, 2005; pp. 3382-3383. (32) Mahkam, M.; Assadi, M. G.; Golipour, N. pH-sensitive hydrogel containing acetaminophen silyl ethers for colon-specific drug delivery. Des. Monomers Polym. 2006, 9, 607–615. (33) Shatayeva, L. K.; Ryadnova, I. Y.; Nechaev, A. N.; Sergeev, A. V.; Chikhacheva, I. P.; McHedlishvili, B. V. Specific features of wetting and adsorption properties of track membranes on the basis of poly(ethylene terephthalate). Colloid J. Russ. Acad. Sci.: Kolloidnyi Zh. 2000, 62, 113–118. (34) Hester, J. F.; Banerjee, P.; Won, Y. Y.; Akthakul, A.; Acar, M. H.; Mayes, A. M. ATRP of amphiphilic graft copolymers based on PVDF and their use as membrane additives. Macromolecules 2002, 35, 7652–7661.
(35) Wei, X.; Wang, R.; Li, Z.; Fane, A. G. Development of a novel electrophoresis-UV grafting technique to modify PES UF membranes used for NOM removal. J. Membr. Sci. 2006, 273, 47–57. (36) Chapman, R. G.; Ostuni, E.; Takayama, S.; Holmlin, R. E.; Yan, L.; Whitesides, G. M. Surveying for surfaces that resist the adsorption of proteins. J. Am. Chem. Soc. 2000, 122, 8303– 8304. (37) Li, Q.; Elimelech, M. Organic fouling and chemical cleaning of nanofiltration membranes: Measurements and mechanisms. Environ. Sci. Technol. 2004, 38, 4683–4693. (38) Tasaki, K. Poly(oxyethylene)-cation interactions in aqueous solution: A molecular dynamics study. Comput. Theor. Polym. Sci. 1999, 9, 271–284. (39) Chai, L.; Klein, J. Role of ion ligands in the attachment of poly(ethylene oxide) to a charged surface. J. Am. Chem. Soc. 2005, 127, 1104–1105. (40) Bailey, F. E.; Koleske, J. V. Poly(ethylene oxide); Academic Press: New York, 1976.
ES9003697
VOL. 43, NO. 10, 2009 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
3871
