Langmuir 2008, 24, 7663-7673
7663
InVited Feature Article Creation of Functional Membranes Using Polyelectrolyte Multilayers and Polymer Brushes Merlin L. Bruening,* David M. Dotzauer, Parul Jain, Lu Ouyang, and Gregory L. Baker* Department of Chemistry, Michigan State UniVersity, East Lansing, Michigan 48824 ReceiVed January 18, 2008. ReVised Manuscript ReceiVed February 27, 2008 Over the last 15 years, the layer-by-layer deposition of polyelectrolytes and the growth of polymer brushes from surfaces have become established techniques for the formation of a wide range of thin films. This article discusses the use of these techniques in creating the skin layer of nanofiltration or gas-separation membranes and in functionalizing the interior of membranes for protein adsorption or catalysis. In the case of separation membranes for nanofiltration, the minimal thickness of layer-by-layer films allows for high flux, and the wide range of available polyelectrolytes that can form these films permits the tailoring of membranes for separations such as water softening, the reduction of F- concentrations, and the removal of dyes from wastewater. For gas separation, polymers grown from surfaces are more attractive than layer-by-layer coatings because most polyelectrolyte films are not highly gas-selective. Cross-linked poly(ethylene glycol dimethacrylate) films grown from porous alumina exhibit CO2/CH4 selectivities of around 20, and the careful selection of monomers should further improve the selectivity of similar membranes. Both layer-by-layer methods and polymer brushes can also be employed to modify the interior of membranes, and we have utilized these techniques to create catalysts, antibody arrays in membranes, and membrane absorbers for protein purification. Polymer brushes are particularly attractive because they allow the absorption of multilayers of protein to yield membranes with binding capacities as high as 150 mg protein/cm3. Some challenges in the practical implementation of these systems, such as the economical formation of membranes using highly permeable polymeric supports, and future directions in research on membrane modification with multilayer films and polymer brushes are also discussed herein.
I. Introduction Thin organic films are at the heart of many practical membranes, including those used in large-scale reverse osmosis and gas separations and in smaller-scale operations such as protein purification. Thus, novel methods for the formation of membranes containing functional thin films could potentially change the scope or efficiency of a wide range of chemical separations. This article describes exploratory studies of the use of layer-by-layer adsorption of polyelectrolytes or the growth of polymer brushes from surfaces to create functional membranes for potential applications in gas separation, nanofiltration, protein purification, and catalysis. Although thin-film deposition techniques such as the Langmuir-Blodgett method have been known for 70 years,1 the relatively new methods of layer-by-layer deposition of polyelectrolytes2–8 and polymer growth from surfaces9–16 represent substantial improvements with respect to ease of film formation * Authors to whom correspondence should be addressed. E-mail:
[email protected],
[email protected]. Phone: (517) 355-9715 ext. 237. Fax: (517) 353-1793. (1) Blodgett, K. B. J. Am. Chem. Soc. 1935, 57, 1007–1010. (2) Decher, G. Science 1997, 277, 1232–1237. (3) Decher, G.; Hong, J.-D.; Schmitt, J, Thin Solid Films 1992, 210/211, 831– 835. (4) Iler, R. K. J. Colloid Interface Sci. 1966, 21, 569–594. (5) Bertrand, P.; Jonas, A.; Laschewsky, A.; Legras, R. Macromol. Rapid Commun. 2000, 21, 319–348. (6) Benkirane-Jessel, N.; Lavalle, P.; Ball, V.; Ogier, J.; Senger, B.; Picart, C.; Schaaf, P.; Voegel, J.-C.; Decher, G. In Macromolecular Engineering: Precise Synthesis, Materials Properties, Applications; Matyjaszewski, K., Gnanou, Y., Leibler, L., Eds.; Wiley-VCH: Weinheim, Germany, 2007; Vol. 2, pp 1249-1305.. (7) De Geest, B. G.; Sanders, N. N.; Sukhorukov, G. B.; Demeester, J.; De Smedt, S. C. Chem. Soc. ReV. 2007, 36, 636–649.
and film stability and versatility. Both of these methods (Figure 1) can yield a variety of robust, functional films on surfaces with a wide range of topology and chemical composition. The layerby-layer method, which typically involves the alternating adsorption of polycations and polyanions, can be simply carried out on essentially any substrate that will support the adsorption of an initial layer of polymer. Similarly, the growth of polymer brushes, which usually includes the attachment of an initiator to a surface followed by polymerization from the initiator, is also very versatile and can take place on any substrate that allows for initiator attachment and is stable to polymerization conditions. Most notably for this work, both the layer-by-layer method and polymerization from a surface can occur in or on porous substrates to produce a wide range of functional membranes (Figure 2).17–29 (8) Jaber, J. A.; Schlenoff, J. B. Curr. Opin. Colloid Interface Sci. 2006, 11, 324–329. (9) Prucker, O.; Ru¨he, J. Macromolecules 1998, 31, 592–601. (10) Prucker, O.; Ru¨he, J. Macromolecules 1998, 31, 602–613. (11) Zhou, F.; Huck, W. T. S. Phys. Chem. Chem. Phys. 2006, 8, 3815–3823. (12) Advincula, R. AdV. Polym. Sci. 2006, 197, 107–136. (13) Tsujii, Y.; Ohno, K.; Yamamoto, S.; Goto, A.; Fukuda, T. AdV. Polym. Sci. 2006, 197, 1–45. (14) Pyun, J.; Kowalewski, T.; Matyjaszewski, K. Macromol. Rapid Commun. 2003, 24, 1043–1059. (15) Ru¨he, J.; Knoll, W. J. Macromol. Sci. Polym. ReV. 2001, C42, 91–138. (16) Zhao, B.; Brittain, W. J. Prog. Polym. Sci. 2000, 25, 677–710. (17) Tieke, B.; Pyrasch, M.; Toutianoush, A. In Multilayer Thin Films: Sequential Assembly of Nanocomposite Materials; Decher, G., Schlenoff, J. B., Eds.; Wiley-VCH: Weinheim, Germany, 2003; pp 427-460.. (18) Tieke, B.; Ackern, F. v.; Krasemann, L Eur. Phys. J. E 2001, 5, 29–39. (19) Toutianoush, A.; Tieke, B. Mater. Sci. Eng., C 2002, 22, 135–139. (20) Jin, W.; Toutianoush, A.; Tieke, B. Langmuir 2003, 19, 2550–2553. (21) Krasemann, L.; Tieke, B. Chem. Eng. Technol. 2000, 23, 211–213. (22) Krasemann, L.; Tieke, B. Mater. Sci. Eng., C 1999, 8-9, 513518.
10.1021/la800179z CCC: $40.75 2008 American Chemical Society Published on Web 05/29/2008
7664 Langmuir, Vol. 24, No. 15, 2008
Bruening et al.
Figure 1. Schematic diagram showing the formation of thin organic films using layer-by-layer adsorption of polyelectrolytes (left) and polymerization from a surface (right).
Figure 2. Schematic diagram of membranes prepared through the deposition of films in or on porous substrates. (a) a nanofiltration membrane prepared by the deposition of a multilayer polyelectrolyte film on a support with small (20-nm-diameter) surface pores; (b) a membrane for protein purification created by growing polymer brushes inside a support with relatively large (0.2- to 5-µm-diameter) pores; and (c) a catalytic membrane prepared by layer-by-layer deposition of polyelectrolytes and charged metal nanoparticles.
Large-scale membrane processes including reverse osmosis and gas separation are extremely important because of their potential for reducing the energy demand in chemical separations. (The chemical processing industry accounts for about 8% of the total energy use in the United States.30) Reverse osmosis, for (23) Stroeve, P.; Vasquez, V.; Coelho, M. A. N.; Rabolt, J. F Thin Solid Films 1996, 284-285, 708712. (24) Leva¨salmi, J.-M.; McCarthy, T. J. Macromolecules 1997, 30, 1752–1757. (25) Balachandra, A. M.; Baker, G. L.; Bruening, M. L. J. Membr. Sci. 2003, 227, 1–14. (26) Sun, L.; Baker, G. L.; Bruening, M. L. Macromolecules 2005, 38, 2307– 2314. (27) Bruening, M. L. In Multilayer Thin Films: Sequential Assembly of Nanocomposite Materials; Decher, G, Schlenoff, J. B., Eds.; Wiley-VCH: Weinheim, Germany, 2003; pp 487-510.. (28) Bruening, M. L.; Sullivan, D. M. Chem.sEur. J. 2002, 8, 3832–3837. (29) Hollman, A. M.; Bhattacharyya, D. Langmuir 2004, 20, 5418–5424.
example, may be as much as an order of magnitude more efficient than thermal distillation.31 However, the development of selective, energy efficient membrane processes is not a trivial endeavor. One of the requirements of these separations is a membrane that is stable, selective, and highly permeable. Because the selective permeation of small molecules demands a material that is relatively impermeable and dense, membranes that allow practical fluxes typically consist of a thin, selective film on a highly permeable support. (See Figure 2a for a schematic example). The minimal thickness of the skin allows a reasonably high permeation rate in addition to selectivity, and the support contributes mechanical strength. The first breakthrough in forming (30) Sustainability in the Chemical Industry: Grand Challenges and Research Needs; National Academy Press: Washington, DC, 2005. (31) Koros, W. J. AIChE J. 2004, 50, 2326–2334.
Creation of Functional Membranes
this type of membrane occurred in the 1960s with the development of methods to cast cellulose acetate membranes with a dense skin on an underlying porous structure (sometimes termed integrally skinned membranes).32 Later methods such as interfacial condensation polymerization via the reaction of aromatic amines and acid chlorides allowed the formation of composite membranes, where the skin consists of a different material than the underlying support.33–35 Membranes prepared in this way are a vital part of many current reverse osmosis separations. Composite membranes are particularly attractive because only a minimal amount of the selective skin is needed, and this may allow for the use of relatively expensive materials.36 Moreover, reducing the thickness of the membrane skin has the potential to enhance permeation rates greatly in many cases. However, decreasing the skin thickness occurs at the risk of forming defects that can dramatically reduce selectivity.37 New techniques such as layer-by-layer deposition or the growth of polymers from surfaces can potentially yield thinner, defect-free films or more selective materials to improve membrane performance. Thin films can also be employed to functionalize the interior of porous membranes (Figure 2b,c). These modifications typically aim at utilizing the relatively high surface area of membranes for applications such as analysis,38,39 selective adsorption or absorption,40 and catalysis.41,42 In this case, covering the top of the membrane with a thin film is not desirable but rather the interior of the membrane needs to be coated with a thin, functional film. Even though typical porous membranes do not have the surface area per gram that may be available with nanoporous beads, convective flow through the membrane pores, which is not available in the pores of beads, can lead to much more rapid adsorption or reaction in the membrane. In fact, the diffusion of macromolecules into porous beads may be so slow that membranes actually have much higher dynamic capacities than beads as a result of more rapid mass transport to binding sites.43 Especially large species including viruses or plasmid DNA may be incapable of entering nanopores, so membrane absorbers are particularly attractive for isolating these species.44,45 Nevertheless, typical protein-absorbing membranes still suffer from low binding capacities. A portion of this article discusses the modification of the interior of membranes with multilayer polyelectrolyte films and polymer brushes to bind proteins. The brushes have the distinct advantage of being able to bind multilayers of protein whereas layer-by-layer films are especially convenient for derivatizing a membrane to bind a monolayer of protein, which may be important in antibody arrays. Below we first discuss how the synthesis of thin films can be varied to deposit the film on the top or in the interior of the (32) Loeb, S.; Sourirajan, S. AdV. Chem. Ser. 1963, 38, 117–32. (33) Zhang, X.; Cahill, D. G.; Coronell, O.; Marin˜as, B. J. Appl. Phys. Lett. 2007, 91, 181904/1–181904/3. (34) Singh, P. S.; Aswal, V. K. J. Phys. Chem. C 2007, 111, 16219–16226. (35) Cadotte, J. E. ACS Symp. Ser. 1985, 269, 273–94. (36) Pinnau, I.; Freeman, B. D. In Membrane Formation and Modification; Pinnau, I., Freeman, B., D., Eds.; American Chemical Society: Washington, DC, 2000; pp 1-22.. (37) Henis, J. M. S.; Tripodi, M. Science 1980, 220, 11–17. (38) Wu, Y.; de Kievit, P.; Vahlkamp, L.; Pijnenburg, D.; Smit, M.; Dankers, M.; Melchers, D.; Stax, M.; Boender, P. J.; Ingham, C.; Bastiaensen, N.; de Wijn, R.; van Alewijk, D.; van Damme, H.; Raap, A. K.; Chan, A. B.; van Beuningen, R. Nucleic Acids Res. 2004, 32, e123/1–e123/7. (39) Xu, Y.; Bao, G. Anal. Chem. 2003, 75, 5345–5351. (40) Ulbricht, M.; Yang, H. Chem. Mater. 2005, 17, 2622–2631. (41) Smuleac, V.; Butterfield, D. A.; Bhattacharyya, D. Langmuir 2006, 22, 10118–10124. (42) Meyer, D. E.; Bhattacharyya, D. J. Phys. Chem. B 2007, 111, 7142–7154. (43) Yang, H.; Viera, C.; Fischer, J.; Etzel, M. R. Ind. Eng. Chem. Res. 2002, 41, 1597–1602. (44) Endres, H. N.; Johnson, J. A. C.; Ross, C. A.; Welp, J. K.; Etzel, M. R. Biotechnol. Appl. Biochem. 2003, 37, 259–266. (45) Etzel, M. R.; Riordan, W. T. Biotechnol. Bioprocess. 2007, 31, 277–296.
Langmuir, Vol. 24, No. 15, 2008 7665
Figure 3. Structures of several polyelectrolytes (or their acid forms) employed in the formation of nanofiltration membranes.
membrane and to introduce new functional groups. Subsequently we present examples of the use of thin-film deposition to prepare composite membranes for gas separation and nanofiltration and then examine the modification of the interior of porous membranes to prepare antibody arrays, membrane absorbers, and catalytic materials. Multilayer polyelectrolyte films and polymer brushes each have unique assets for different applications. The final section presents challenges and future directions in this area of research.
II. Growth of Thin Films in or on Porous Supports Polyelectrolyte Multilayers. Layer-by-layer adsorption (Figure 1) is one of the simplest methods of membrane modification because it simply involves exposing porous substrates to alternating polyanion and polycation solutions, with water rinses between each exposure. Figure 3 shows the structures of some of the polyelectrolytes that we and many others use in layerby-layer adsorption. To deposit the film primarily on top of the support, we utilize membranes with small surface pores and place the substrate in a holder that exposes only the top surface of the support to the polyelectrolyte solution. Figure 4a shows a cross-sectional SEM image of a polyelectrolyte film formed in this way on a porous alumina support.46 Notably, the film covers the surface without filling underlying pores, in part because the pores at the surface of the support have a diameter of only 20 nm whereas bulk pores have a diameter of 200 nm. The deposition of four to five poly(styrene sulfonate)/protonated poly(allylamine) (PSS/PAH) layer pairs is required to cover the underlying pores fully.47 More recently, we deposited similar polyelectrolyte films on poly(ether sulfone) ultrafiltration membranes that have a molecular weight cutoff of 50 kDa.48 Lack of contrast between poly(ether sulfone) and the polyelectrolyte film makes the characterization of these membranes with SEM difficult, but their performance in nanofiltration is similar to that of the coated porous alumina membranes. The deposition of polyelectrolyte films in the interior of membranes can also be easily achieved and simply requires the flow of the polyelectrolyte solutions through membranes with relatively large pores (typically 200 nm or larger). Figure 4b shows polyelectrolyte tubes that were formed using poly(acrylic (46) Ouyang, L.; Malaisamy, R.; Bruening, M. L. J. Membr. Sci. 2008, 310, 76–84. (47) Harris, J. J.; Stair, J. L.; Bruening, M. L Chem. Mater. 2000, 12, 1941– 1946. (48) Malaisamy, R.; Bruening, M. L. Langmuir 2005, 21, 10587–10592.
7666 Langmuir, Vol. 24, No. 15, 2008
Bruening et al.
Figure 5. Cross-sectional SEM image of a porous alumina membrane (0.2-µm-diameter pores) coated with a PAA/PAH/Au colloid film.52 Scheme 1. Possible Methods of Attaching Initiators to (a) Alumina, (b) Au, and (c) Hydroxylated Surfaces
Figure 4. (a) Cross-sectional SEM image of a [PSS/PAH]5 film deposited on a porous alumina membrane with 0.02-µm-diameter surface pores (bulk pores have a diameter of 0.2 mm).46 (b) SEM image of a crosslinked [PAA/PAH]3 film that was deposited in the interior of porous alumina with 0.2-µm-diameter pores. The film in image b was crosslinked through heat-induced amidation, and the porous alumina was dissolved prior to imaging the polymer.
acid)/PAH (PAA/PAH) deposition in porous alumina, crosslinking via amide formation,49 and subsequent dissolution of the alumina50 in basic solvent. The image clearly shows that the tubes are hollow, and thus polyelectrolyte films in alumina membranes do not prevent convective flow. Ai and co-workers reported that it is possible to form similar polyelectrolyte tubes even without cross-linking of polyelectrolytes.51 Layer-by-layer deposition is especially versatile because virtually any charged material can be included in films. To prepare catalytic membranes, for example, we use the layer-by-layer deposition of PAH and citrate-protected, negatively charged gold colloids in the membrane pores (Figure 2c).52 In porous alumina membranes, these films provide a remarkably uniform and dense distribution of colloids as shown in Figure 5. Both SEM images and analysis of the amount of gold present in the membranes reveal about 500 nanoparticles/µm2, with a nanoparticle diameter of ∼12 nm. The deposition of these coatings can also occur in nylon and polycarbonate membranes, suggesting that this method should allow the formation of films with a high density of nanoparticles in virtually any substrate. Importantly, the catalytic activity of the nanoparticles is equivalent to that of nanoparticles in solution (see section VI). Growth of Polymers from Surfaces. Polymers grown from surfaces provide a very different architecture than layer-by-layer films in that the film should typically be much less cross-linked (multilayer polyelectrolyte films are ionically cross-linked) and much more extended from the surface (Figure 1). This could prove to be very valuable in developing swollen films capable of binding large numbers of biomacromolecules. Moreover, it (49) Harris, J. J.; DeRose, P. M.; Bruening, M. L J. Am. Chem. Soc. 1999, 121, 1978–1979. (50) Martin, C. R. Science 1994, 266, 1961–1966. (51) Ai, S.; Lu, G.; He, Q.; Li, J. J. Am. Chem. Soc. 2003, 125, 11140–11141. (52) Dotzauer, D. M.; Dai, J.; Sun, L.; Bruening, M. L. Nano Lett. 2006, 6, 2268–2272.
is easier to control the composition when growing the film from the surface because the coating consists of only one polymer, and functionality can be introduced through the appropriate choice of the monomer unit or by the derivatization of the polymer. Although there are a number of techniques for polymerization from surfaces,9,12,53–55 atom-transfer radical polymerization (ATRP) is especially attractive because it frequently affords controlled polymerization and is relatively tolerant of functional groups.56–58 The first step in ATRP is the attachment of an initiator to the substrate for subsequent polymerization. For alumina and silica surfaces, this can be accomplished through silanization as shown in Scheme 1.57 This scheme also shows other possible methods of initiator attachment including the adsorption of disulfides on Au and the reaction of surface hydroxyl groups (amino groups could also be used) with bromoisobutyryl bromide.59–61 The functionalization of polymer surfaces with (53) Jordi, M. A.; Seery, T. A. P. J. Am. Chem. Soc. 2005, 127, 4416–4422. (54) Barsbay, M.; Gu¨ven, O.; Stenzel, M. H.; Davis, T. P.; Barner-Kowollik, C.; Barner, L. Macromolecules 2007, 40, 7140–7147. (55) Husseman, M.; Malmstroem, E. E.; McNamara, M.; Mate, M.; Mecerreyes, D.; Benoit, D. G.; Hedrick, J. L.; Mansky, P.; Huang, E.; Russell, T. P.; Hawker, C. J. Macromolecules 1999, 32, 1424–1431. (56) Matyjaszewski, K.; Xia, J. Chem. ReV. 2001, 101, 2921–2990. (57) Matyjaszewski, K.; Miller, P. J.; Shukla, N.; Immaraporn, B.; Gelman, A.; Luokala, B. B.; Siclovan, T. M.; Kickelbick, G.; Vallant, T.; Hoffmann, H.; Pakula, T. Macromolecules 1999, 32, 8716–8724. (58) Huang, X.; Doneski, L. J.; Wirth, M. J. Anal. Chem. 1998, 70, 4023– 4029.
Creation of Functional Membranes
Langmuir, Vol. 24, No. 15, 2008 7667
Scheme 2. Growth of Polymer Brushes via ATRP from Macroinitiators Deposited on a Surface Using Layer-by-Layer Adsorption64
initiators is more complicated than the functionalization of alumina because many polymers do not have a high density of functional groups on their surfaces (cellulose acetate is a notable exception62), and they may also be incompatible with the solvent used for initiator attachment. To overcome this problem, we have begun utilizing a macroinitiator first reported by Armes.63 This macroinitiator can be deposited on polymer surfaces (e.g., poly(ether sulfone)) using layer-by-layer deposition (Scheme 2), and ATRP can be successfully carried out from the immobilized macroinitiator.64 Poly(glycidyl methacrylate) has also been used to functionalize membranes for subsequent initiator attachment.65 One drawback to ATRP is that control of polymerization is often achieved at the expense of the polymerization rate. However, water frequently accelerates ATRP,66–68 and the polymerization of 2-hydroxyethyl methacrylate in water provides a particularly robust method for synthesizing relatively thick (hundreds of nanometers) films in a few hours.69 These films are especially attractive for functionalizing membranes because the hydroxyl groups of poly(2-hydroxyethyl methacrylate) (PHEMA) readily react with a variety of reagents to yield functionalized films. To create protein-binding films, we expose PHEMA brushes to succinic anhydride to form -COOH groups in the film, activate the -COOH groups, and then allow them to react with aminobutyl nitrilotriacetate (Scheme 3).70,71 The nitrilotriacetate (NTA) moiety can bind metal ions, and Ni2+-NTA complexes, in particular, selectively bind proteins with polyhistidine tags. Poly(acrylic acid) (PAA) may provide an even more convenient polymer brush for functionalization, but PAA is typically difficult to polymerize by ATRP, although Husson and co-workers showed that poly(methacrylic acid) films can be formed at a rate of 1 nm/h.72 To circumvent the challenge of polymerizing acrylic acid, we grow poly(tert-butylacrylate) polymers from surfaces (59) Shah, R. R.; Merreceyes, D.; Husemann, M.; Rees, I.; Abbott, N. L.; Hawker, C. J.; Hedrick, J. L. Macromolecules 2000, 33, 597–605. (60) Kim, J.-B.; Bruening, M. L.; Baker, G. L. J. Am. Chem. Soc. 2000, 122, 7616–7617. (61) Friebe, A.; Ulbricht, M. Langmuir 2007, 23, 10316–10322. (62) Singh, N.; Wang, J.; Ulbricht, M.; Wickramasinghe, S. R.; Husson, S. M. J. Membr. Sci. 2008, 309, 64–72. (63) Chen, X. Y.; Armes, S. P.; Greaves, S. J.; Watts, J. F. Langmuir 2004, 20, 587–595. (64) Jain, P.; Dai, J.; Grajales, S.; Saha, S.; Baker, G. L.; Bruening, M. L. Langmuir 2007, 23, 11360–11365. (65) Singh, N.; Husson, S. M.; Zdyrko, B.; Luzinov, I. J. Membr. Sci. 2005, 262, 81–90. (66) Robinson, K. L.; Khan, M. A.; de Ba´n˜ez, M. V.; Wang, X. S.; Armes, S. P. Macromolecules 2001, 34, 3155–3158. (67) Perruchot, C.; Khan, M. A.; Kamitsi, A.; Armes, S. P.; von Werne, T.; Patten, T. E. Langmuir 2001, 17, 4479–4481. (68) Jones, D. M.; Huck, W. T. S. AdV. Mater. 2001, 13, 1256–1259. (69) Huang, W. X.; Kim, J. B.; Bruening, M. L.; Baker, G. L. Macromolecules 2002, 35, 1175–1179. (70) Sun, L.; Dai, J.; Baker, G. L.; Bruening, M. L. Chem. Mater. 2006, 18, 4033–4039. (71) Jain, P.; Sun, L.; Dai, J.; Baker, G. L.; Bruening, M. L. Biomacromolecules 2007, 8, 3102–3107.
Scheme 3. Derivatization of PHEMA Brushes to Immobilize Metal Ion-NTA Complexes71
and then hydrolyze the tert-butyl ester to form poly(acrylic acid).73 By using a highly active ATRP catalyst, 100-nm-thick poly(tertbutylacrylate) films can be formed in as little as 5 min. Moreover, (72) Sankhe, A. Y.; Husson, S. M.; Kilbey, S. M., II. J. Polym. Sci., Part A: Polym. Chem. 2007, 45, 566–575.
7668 Langmuir, Vol. 24, No. 15, 2008
these films can be hydrolyzed to PAA in just 15 min. However, the polymerization of poly(tert-butylacrylate) at these high rates is very sensitive to oxygen and may be difficult to apply in membranes. Another advantage of polymer brushes compared to multilayer films is that, in principle, it is possible to control the density of chains growing from the surface. This control can be achieved by simultaneously depositing a spacer molecule and the initiator (e.g., a diluent silane and an initiator silane) to decrease the initiator density on the surface.74 In some cases, a lower density of chains may allow more swelling of the film and lead to increased protein binding.
Bruening et al. Scheme 4. Imidization of a Poly(amic acid)/PAH Film via Heating to Provide a Gas-Selective Material77
III. Gas Separation As mentioned earlier, membranes for small-molecule separations such as reverse osmosis, nanofiltration, and gas separation generally contain an ultrathin, selective skin on a highly permeable support. Gas separation is especially demanding in terms of requiring a film that is completely free of defects, in part because skins are dry so swelling will not close imperfections. Most studies of gas separation with layer-by-layer films showed very little selectivity,23,75,76 although McCarthy did report increased selectivities in H2, N2, and O2 permeation after coating poly(4methyl-1-pentene) with 20- to 200-layer films.24 We prepared gas-selective films using the layer-by-layer deposition of poly(amic acids) and PAH, followed by imidization of these films through heating (Scheme 4). The resulting polyimides are known to have relatively high gas selectivities in their pure form, and we compared the gas selectivities of the imidized layerby-layer films with the selectivities of pure polyimides.77 Remarkably, when the fraction of PAH in the film is limited to about 10% through the control of deposition conditions, the imidized poly(amic acid)/PAH films show the same selectivities and permeabilities as pure polyimide films. O2/N2 selectivities reach a value of 6.9, and CO2/N2 selectivities are as high as 68. These high selectivities clearly demonstrate that the films are free of defects. We also examined what thickness of the polyelectrolyte film is required for the formation of defect-free films and found that selectivity decreases only when thicknesses are less than about 40 nm. Considering that the underlying pores in the alumina surface have a diameter of 20 nm, the formation of a 40 nm defect-free film is unexpected. Typically, it is difficult to prepare defect-free membranes with skin layers that are less than 50 nm thick. Imidization at high temperature may anneal the films to help remove defects. Polymer brushes are also capable of forming gas-selective membranes, but their composition must be carefully selected. To prepare gas-selective membranes, we first immobilized initiators on the surface of porous alumina.25 This occurred through the deposition of a few layers of PSS/PAH on the membrane, followed by reaction of the outer layers of PAH with 2-bromopropionyl bromide (Figure 6). The use of a polyelectrolyte film helped to provide a uniform coverage of initiator on the surface. Subsequently, we grew poly(ethylene glycol dimethacrylate) (PEGDMA) films from these initiators. PEGDMA films with a thickness of 50 nm exhibited selectivities of 18 and 26 (73) Bao, Z.; Bruening, M. L.; Baker, G. L. J. Am. Chem. Soc. 2006, 128, 9056–9060. (74) Bao, Z.; Bruening, M. L.; Baker, G. L. Macromolecules 2006, 39, 5251– 5258. (75) Ackern, F. v.; Krasemann, L.; Tieke, B. Thin Solid Films 1998, 327-329, 762–766. (76) Kotov, N. A.; Magonov, S.; Tropsha, E. Chem. Mater. 1998, 10, 886– 895. (77) Sullivan, D. M.; Bruening, M. L. Chem. Mater. 2003, 15, 281–287.
for CO2/CH4 and CO2/N2, respectively, and CO2 gas permeabilities (about 20 × 10-10 cm3(STP)/(cm2 s cmHg) were typical for this type of material. The selectivities are likely due in part to the high solubility of CO2 in the membrane because CO2 is the more condensable gas and it also interacts with the carbonyl groups of PEGDMA.78 Most importantly, 50-nm-thick films are defect-free, so both high flux and high selectivity can be achieved with this deposition procedure. The use of other monomers should yield even higher selectivities, and we are currently trying to develop improved membranes for CO2/H2 separation because the selective removal of CO2 is an important process in the production of H2 from syngas.78,79
IV. Nanofiltration Nanofiltration is a pressure-based separation like reverse osmosis, but it employs a more permeable membrane to allow for lower operating pressures along with a higher passage of certain small molecules or ions.80 The lower operating pressures (78) Lin, H.; Freeman, B. D. J. Mol. Struct. 2005, 739, 57–74. (79) Lin, H.; Van Wagner, E.; Freeman, B. D.; Toy, L. G.; Gupta, R. P. Science 2006, 311, 639–642.
Creation of Functional Membranes
Langmuir, Vol. 24, No. 15, 2008 7669
advantage of these properties, we showed that five-bilayer PSS/ PAH films on porous alumina exhibit a 95% rejection of MgCl2 along with a Na+/Mg2+ selectivity of 22, whereas a commercial nanofiltration membrane has a lower Mg2+ rejection and a Na+/ Mg2+ selectivity of only 4.46 (Rejection, R, is defined by eq 1, where Cp and Cf are the concentrations of an analyte in the permeate and the feed, respectively. Selectivity, R, is defined by eq 2, where superscripts A and B refer to two different species.)
( )
R) 1RBA )
Figure 6. Schematic diagram of the growth of a PEGDMA membrane from initiators linked to a multilayer polyelectrolyte film (MPF).25
provide substantial energy savings compared to reverse osmosis, and the higher passage of small molecules allows some separations that are not possible with reverse osmosis. As an example, water softening requires the removal of divalent ions (Ca2+ and Mg2+), but the concurrent removal of Na+ and K+ is not required. In fact, the removal of monovalent ions increases the osmotic pressure that must be overcome for the separation, and remineralization of the water is sometimes necessary prior to use. Thus, nanofiltration membranes that selectively reject divalent cations may be very useful. The selective removal of divalent cations requires a positively charged membrane with a relatively small effective pore size so that separations can be based on both the charge and size of different hydrated ions.81 Multilayer polyelectrolyte films are attractive for these separations because their surfaces are highly charged and in some cases effective pore diameters are less than 1 nm.82–85 Using nanofiltration of neutral molecules to estimate pore size, we found that PSS/PAH films deposited on porous alumina have pore diameters of 0.8-1.0 nm.82 Taking (80) Scha¨fer, A. I.; Fane, A. G.; Waite, T. D. Nanofiltration: Principles and Applications; Elsevier: Oxford, U.K., 2005. (81) Wang, X.-L.; Tsuru, T.; Nakao, S.; Kimura, S. J. Membr. Sci. 1997, 135, 19–32. (82) Liu, X.; Bruening, M. L. Chem. Mater. 2004, 16, 351–357. (83) Tieke, B.; Toutianoush, A.; Jin, W. AdV. Colloid Interface Sci. 2005, 116, 121–131. (84) Scho¨nhoff, M.; Ball, V.; Bausch, A. R.; Dejugnat, C.; Delorme, N.; Glinel, K.; Klitzing, R. V.; Steitz, R. Colloids Surf., A 2007, 303, 14–29. (85) Cha´vez, F. V.; Scho¨nhoff, M. J. Chem. Phys. 2007, 126, 104705.
Cp × 100% Cf
CA,p CB,f 100 - RA ) CA,f CB,p 100 - RB
(1) (2)
By varying polyelectrolytes (Figure 3) and deposition conditions, it is possible to tailor the permeation properties of multilayer films for specific applications. For example, in contrast to PSS/ PAH films, which show 90% rejection of molecules as small as glycerol, membrane skins composed of multilayers of hyaluronic acid and chitosan show only modest rejections of myoglobin, which has a molecular mass of 17 000.86 For a number of smallmolecule solutes, PSS/poly(diallyldimethylammonium chloride) (PSS/PDADMAC) films show lower rejections than PSS/PAH but much higher rejections than chitosan/hyaluronic acid. Measurements of film swelling in water correlate well with the permeation properties of different polyelectrolyte pairs.87 Chitosan/hyaluronic acid films swell by 400% in water, whereas PSS/PAH films typically swell by about 100%, and PSS/ PDADMAC films terminated with PSS show a swelling of 106-129%, depending on the ionic strength of deposition solutions. Both swelling and transport results also correlate well with the charge density on the polyelectrolytes, suggesting that the density of ionic cross-links plays a crucial role in determining the transport properties, as first noted by Tieke.88 We recently examined the performance of multilayer polyelectrolyte membranes in several separations with practical implications. In the separation of salt from reactive dyes that may be left over from coloring processes, [PSS/PAH]4PSS films on porous alumina exhibit a remarkable dye-salt selectivity of 2000-3000 and allow fluxes greater than those through commercial membranes.89 The dye rejection by these membranes is >99.9% and exceeds previously reported rejection values for nanofiltration membranes.89–93 The removal of F- from drinking water is important in many areas to avoid dental and bone fluorosis and other possible adverse health effects, and the selective removal of F- in the presence of other halides is attractive for decreasing osmotic pressure and avoiding any need for remineralization.94,95 The separation of monovalent ions such as Cl- and F- is difficult, however, because the hydrated forms of these ions are similar in size and have the same charge. Nevertheless, with careful optimization of polyelectrolyte multilayers, partial separation of monovalent ions can be accomplished. [PSS/PDADMAC]4PSS membranes show (86) Miller, M. D.; Bruening, M. L. Langmuir 2004, 20, 11545–11551. (87) Miller, M. D.; Bruening, M. L. Chem. Mater. 2005, 17, 5375–5381. (88) Krasemann, L.; Toutianoush, A.; Tieke, B. J. Membr. Sci. 2001, 181, 221–228. (89) Hong, S. U.; Miller, M. D.; Bruening, M. L. Ind. Eng. Chem. Res. 2006, 45, 6284–6288. (90) Koyuncu, I.; Topacik, D.; Yuksel, E. Sep. Purif. Technol. 2004, 36, 77– 85. (91) Akbari, A.; Remigy, J. C.; Aptel, P. Chem. Eng. Process. 2002, 41, 601– 609. (92) Xu, X.; Spencer, H. G. Desalination 1997, 114, 129–137. (93) Tang, C.; Chen, V. Desalination 2002, 143, 11–20. (94) Hileman, B. Chem. Eng. News 2006, 84, 11. (95) Petersen, R. J. J. Membr. Sci. 1993, 83, 81–150.
7670 Langmuir, Vol. 24, No. 15, 2008
Bruening et al.
Figure 7. Schematic diagram of the modification of membranes with PAA-terminated PAA/PAH films and the activation of these films to allow covalent protein attachment.97
F- rejection of about 70% and Cl- rejection of only 10%.96 Interestingly, PSS/PAH membranes and the commercial NF membranes that we tested show minimal Cl-/F- selectivity, although commercial membranes do exhibit F- rejections as high as 80%. Rejection and Cl-/F- selectivity are also a strong function of the number of layers in PSS/PDADMAC films when only a few layers are employed. Zeta potential measurements suggest a strong dependence of surface charge on the number of layers in these particular films.
V. Modification of Membrane Interiors for Protein Adsorption Layer-by-layer adsorption of polyelectrolytes and growth of polymer brushes are also useful for modifying the interior of membranes (Figure 2b,c) to take advantage of the high internal membrane surface area in applications that include sensing and protein purification. For example, the use of membranes rather than nonporous flat substrates as supports for antibody arrays allows for the immobilization of more antibody per external area and should enhance antigen adsorption and hence signal-to-noise ratios in fluorescence-based analysis.39 As an illustration, a 25mm-diameter porous alumina substrate that is 60 µm thick and contains 0.2-µm-diameter pores at a porosity of 50% has a surface area of 3000 cm2, which is about 600-fold greater than the area of the top of a 25 mm disk. The layer-by-layer adsorption of films containing PAA provides a convenient way to modify membrane interiors for antibody attachment and the reduction of nonspecific adsorption (Figure 7).97 Membrane modification begins with the deposition of a film terminated with PAA and the subsequent conversion of the (96) Hong, S. U.; Malaisamy, R.; Bruening, M. L. Langmuir 2007, 23, 1716– 1722. (97) Dai, J.; Baker, G. L.; Bruening, M. L. Anal. Chem. 2006, 78, 135–140.
external carboxylic acid groups of PAA to the N-hydroxysuccinimide active esters. The reaction of the active esters with the amine groups of a protein immobilizes the protein via a covalent bond, whereas active esters that do not react with protein are hydrolyzed by water to reform the PAA coating. Figure 8 shows analyses of fluorescently labeled IgG using membranes that were coated with three spots of anti-IgG. Binding of the labeled IgG to the antibody spots yields the localized fluorescence, and when the membrane is simply immersed in the IgG solution for 1 h, the fluorescence intensity is similar to that obtained with a nonporous, antibody-spotted substrate. However, when the antigen is pumped through the membrane, binding occurs throughout the alumina substrate, and the detection limit of the technique decreases about 500-fold, which is the approximate surface-area enhancement. Just as important as the high fluorescence intensity, the background around the anti-IgG spots (Figure 8) is essentially dark, showing that the layer-by-layer film is resistant to the nonspecific adsorption of labeled IgG. One might think that there would be significant nonspecific adsorption via electrostatic interactions with the negatively charged PAA, but several studies showed that at physiological ionic strength, there is minimal nonspecific adsorption to PAA.98–101 The above results demonstrate the potential of membranes to enhance signals in protein microarrays, but even in this case, only a monolayer of antibody is likely immobilized within the (98) Czeslik, C.; Jackler, G.; Hazlett, T.; Gratton, E.; Steitz, R.; Wittemann, A.; Ballauff, M. Phys. Chem. Chem. Phys. 2004, 6, 5557–5563. (99) Jackler, G.; Wittemann, A.; Ballauff, M.; Czeslik, C. Spectroscopy 2004, 18, 289–299. (100) Salloum, D. S.; Schlenoff, J. B. Biomacromolecules 2004, 5, 1089– 1096. (101) Mendelsohn, J. D.; Yang, S. Y.; Hiller, J. A.; Hochbaum, A. I.; Rubner, M. F. Biomacromolecules 2003, 4, 96–106.
Creation of Functional Membranes
Langmuir, Vol. 24, No. 15, 2008 7671
Figure 9. Breakthrough curve for BSA adsorption in porous alumina membranes coated with PHEMA-NTA-Cu2+. The BSA concentration in the feed solution was 0.56 mg/mL, and the initial flow rate was 2.4 mL/min, decreasing to 0.9 mL/min at the end of the experiment.70
Figure 8. Fluorescent images of membranes spotted with anti-IgG and subsequently exposed to different concentrations of labeled IgG through the pumping of the solution through the membrane (left) and immersion in a quiescent solution (right). Membranes were modified with [PAA/ PAH]3PAA films and activated with EDC and NHS prior to the spotting of the anti-IgG.97
membrane pores. The interior of the polyelectrolyte film contains ionic cross-links and thus is probably inaccessible to activation reagents and large antibodies. As mentioned earlier, polymer brushes may provide films capable of overcoming this problem because they can be highly swollen to allow binding of multilayers of protein. Our initial work with protein binding to polymer brushes employed films on Au-coated Si wafers because they are easy to characterize using ellipsometry and reflectance infrared spectroscopy. These studies show that simple 55-nm-thick PAA films increase in thickness by 160 nm when binding lysozyme via ion exchange.102 This amount of bound lysozyme is the equivalent of a remarkable 80 monolayers of lysozyme in the film. Recent studies of PAA in membranes also show the binding of large amounts of lysozyme.40,62 To introduce more specific interaction with proteins, we activate the poly(acrylic acid) and allow it to react with aminobutyl NTA and then introduce metal ions to form the NTA-metal ion complexes. (This process is similar to that shown in Scheme 3.) After derivatization with NTA-Cu2+ complexes, films with a dry thickness of 55 nm (before derivatization) still bind as much as 58 nm of BSA. These large amounts of protein binding are consistent with the high swellability of PAA brushes.102 In situ ellipsometry measurements show that a deprotonated PAA film with a dry thickness of about 55 nm has a water-swollen thickness of around 275 nm. After derivatization and BSA binding, films have a dry thickness of 140 nm, but the water-swollen thickness is still 275 nm. The high water content of these films is likely crucial to achieving high degrees of protein binding. PHEMA films derivatized with succinic anhydride and NTA also show high levels of protein binding. A PHEMA film that is 50 nm thick prior to derivatization with succinic anhydride and NTA-Cu2+ binds approximately 40 nm of BSA, which corresponds to 10 (102) Dai, J.; Bao, Z.; Sun, L.; Hong, S. U.; Baker, G. L.; Bruening, M. L. Langmuir 2006, 22, 4274–4281.
monolayers of BSA.70 Although PHEMA films are not as hydrophilic as PAA, they evidently still swell sufficiently to bind multilayers of protein. After these promising studies on flat substrates, we began examining protein binding to polymer brushes in porous alumina membranes.70 Figure 9 shows the breakthrough curve for binding of BSA to PHEMA-NTA-Cu2+ brushes in porous alumina. Initially, virtually all of the BSA passing through the membrane binds to the brush, but as the film becomes saturated with protein, the BSA breaks through. Integration of the difference between the feed concentration and the permeate concentration as a function of volume reveals a binding capacity of 130 mg BSA/ cm3 of membrane, which is several fold higher than the capacities reported for other protein-absorbing membranes.70 Given that the porous alumina membranes have a porosity of around 50%, in a given pore there is about 260 mg of bound BSA/cm3, suggesting that ∼25% of the pore is filled with protein. Moreover, another 20% of the pore volume is filled with the polymer brush. Thus, for this system (porous alumina with 0.2-µm-diameter pores), binding is nearing its maximum value because the absorption of significantly more protein would essentially block the flow. Even in these membranes, the flow decreased from 2.4 to 0.9 mL/min during protein binding. NTA-Cu2+ complexes are not highly selective protein binders because Cu2+ can interact with any protein containing accessible histidine residues. To introduce selectivity, we utilize NTA-Ni2+ complexes, which are well known to selectively bind proteins tagged with polyhistidine. (The strength of the NTANi2+-histidine interaction is weaker than the NTA-Cu2+histidine interaction, so multiple histidines are required for binding to films derivatized with NTA-Ni2+.) The breakthrough curve for the binding of histidine6-tagged ubiquitin (HisU) to alumina membranes modified with PHEMA-NTA-Ni2+ is sharp, and membrane saturation takes about 15 min. The capacity of these membranes is 100-150 mg/cm3.71 To demonstrate selectivity, we loaded aluminaPHEMA-NTA-Ni2+ membranes with 10 mL of a solution containing 1 mg/mL BSA and 0.05 mg/mL HisU, rinsed the membrane, and then eluted the bound protein with an imidazole buffer. The gel electropherogram of the eluent showed a strong band for HisU and no detectable band for BSA. Although BSA was in a 20-fold excess in the loading solution, the protein in the eluent was >99% HisU. Figure 10 shows the gel electropherogram of the eluent from a membrane that was loaded with
7672 Langmuir, Vol. 24, No. 15, 2008
Bruening et al.
Figure 11. Plot of the 4-nitrophenol conversion percentage vs flux through membranes containing a PAA/PAH/gold colloid film. The curve is a fit to a first-order reaction model. Feed conditions: [4-nitrophenol] ) 0.4 mM and [NaBH4] ) 20 mM.52 Figure 10. Electropherogram (silver staining) of 10% bovine serum spiked with 0.3 mg/mL HisU (lane 1) and the imidazole eluent from an alumina-PHEMA-NTA-Ni2+ membrane loaded with this HisU solution (lane 2).71
10% bovine serum spiked with 0.3 mg/mL HisU. Even with this complicated loading mixture, the protein in the eluent is >99% HisU. Our most recent results suggest that brush-modified polymeric membranes with larger pores are capable of purifying His-tagged proteins that are overexpressed in cell extracts.
to or greater than those of Wilkinson’s catalyst, the prototypical homogeneous catalyst for such reactions.106 The binding of nanoparticles to polyethyleneimine in solution also yields high selectivities in such reactions.107 However, the immobilization of nanoparticles in membranes or on alumina facilitates recycling of the catalyst, and the membrane support also allows control over residence time to potentially minimize formation of side products.
VII. Challenges and Future Directions VI. Catalytic Membranes As a further illustration of the ability of thin films to functionalize membranes, we modified the interior of alumina and polymeric membranes using layer-by-layer adsorption of a polycation and citrate-stabilized gold nanoparticles.52 As shown above in Figure 5, this method yields a density of 500 nanoparticles/µm2 when using Au nanoparticles with a diameter of 12 nm. To test the activity of these immobilized particles, we examined the catalytic reduction of nitrophenol by NaBH4. This is a very rapid reaction, and even with a linear velocity of 1 cm/s in a 60-µm-thick membrane (residence time of 6 ms), the conversion of 4-nitrophenol to 4-aminophenol can be >99%. One advantage of membrane-based reactors is that the kinetics can be investigated simply by varying the flow rate through the membrane. Figure 11 shows a plot of conversion versus flux through the membrane, and fitting these data to a first-order rate constant yields k ) 0.018 cm/s when the rate constant is normalized by dividing by the surface area of the nanoparticles per solution volume in the membrane. Within experimental error, this is the same normalized rate constant that one obtains for 4-nitrophenol reduction catalyzed by similar Au nanoparticles suspended in solution. Thus, immobilization of the nanoparticles in the polyelectrolyte film does not decrease their activity. The immobilization of nanoparticles in polyelectrolyte films also has the potential to increase their selectivity.103–105 The hydrogenation of monosubstituted and multisubstituted double bonds catalyzed by nanoparticle/polyelectrolyte films deposited on alumina particles reveals selectivities (rate of hydrogenation of monosubstituted double bonds divided by the rate of hydrogenation of multisubstituted double bonds) that are equal (103) Kidambi, S.; Li, J.; Dai, J.; Bruening, M. L. J. Am. Chem. Soc. 2004, 126, 2658–2659. (104) Kidambi, S.; Bruening, M. L. Chem. Mater. 2005, 17, 301–307. (105) Niu, Y.; Yeung, L. K.; Crooks, R. M. J. Am. Chem. Soc. 2001, 123, 6840–6846.
Gas Separation. Results thus far show that it is possible to prepare highly selective, permeable gas separation membranes using either multilayer polyelectrolyte films or polymer brushes. Nevertheless, most polyelectrolyte multilayer films do not show attractive gas selectivities, and the need to imidize films at high temperature to achieve selectivity will limit the supports on which these films can be deposited. (Many polymer supports cannot withstand temperatures >100 °C.) Thus, the growth of polymer brushes is currently more attractive than layer-by-layer deposition for the preparation of gas-separation membranes. Careful selection of the monomers used to form brushes should yield the desired performance for specific applications, but the development of economical membranes will likely require improvements in membrane preparation, including both initiator deposition and polymerization. The macroinitiators that we and others are developing may facilitate rapid initiator attachment on a variety of surfaces, and we recently found that in addition to tert-butyl acrylate it is also possible to polymerize 100-nm-thick poly(ethylene glycol)methacrylates in minutes. This may allow the rapid preparation of membranes with high CO2 permeabilities and CO2/H2 selectivities.78,79 Nanofiltration. In nanofiltration, polyelectrolyte multilayer membrane skins show a wide range of attractive properties. It remains to be seen, however, whether layer-by-layer deposition can be economical on a large scale. At present, we are unaware of a successful commercial application of polyelectrolyte multilayers, which may stem from the need for multiple steps in the deposition process. Although film deposition is easy and convenient in the laboratory, the relatively large number of adsorption and rinsing steps may be prohibitive in a production process. Research efforts aimed at developing sprayed poly(106) Bhattacharjee, S.; Bruening, M. L. Langmuir 2008, 24, 2916. (107) Vasylyev, M. V.; Maayan, G.; Hovav, Y.; Haimov, A.; Neumann, R. Org. Lett. 2006, 8, 5445–5448.
Creation of Functional Membranes
electrolyte films might help in this regard.108–110 Moreover, if multilayer films can be employed in highly specialized separations, then challenges in manufacturing would be more tolerable. An example of this would be the separation of enantiomers, but thus far the selectivity of polyelectrolyte multilayers in this area is relatively low.111 The development of new films with higher selectivities in specialized separations is still needed. Another challenge in creating polyelectrolyte multilayer membranes is the development of appropriate supports. Flux through films on porous alumina supports is higher than that through films on polymeric ultrafiltration membranes, but porous alumina is both brittle and expensive. The development of highly permeable polymeric supports with surface pore diameters of around 20 nm could help to alleviate this problem. Membrane Absorbers. Layer-by-layer adsorption of multilayer films containing poly(acrylic acid) provides a convenient way to prepare surfaces for the covalent attachment of proteins and other molecules. Additionally, early studies show that these surfaces exhibit minimal nonspecific adsorption of proteins. Thus, this surface-modification technique could prove valuable for a variety of applications where only monolayers of protein are required. Polymer brushes, in contrast, allow the adsorption of multilayers of protein and are more attractive for protein purification. In porous alumina membranes, we successfully achieved selective binding of tagged proteins at high binding capacities, but more work is needed in examining how binding capacity varies with flow rate and pore diameter. Typical commercial ion-exchange membranes are at least an order of magnitude more permeable than the derivatized porous alumina membranes described here because of a larger effective pore size. Higher permeability is important because it will allow the use of thicker membranes to enhance capacity. Future work is needed to show (108) Lu, C.; Do¨nch, I.; Nolte, M.; Fery, A. Chem. Mater. 2006, 18, 6204– 6210. (109) Izquierdo, A.; Ono, S. S.; Voegel, J. C.; Schaaf, P.; Decher, G. Langmuir 2005, 21, 7558–7567. (110) Schlenoff, J. B.; Dubas, S. T.; Farhat, T. Langmuir 2000, 16, 9968– 9969. (111) Rmaile, H. H.; Schlenoff, J. B. J. Am. Chem. Soc. 2003, 125, 6602– 6603.
Langmuir, Vol. 24, No. 15, 2008 7673
whether the high capacities of brush-modified membranes can be maintained in systems with larger pores. Additionally, the purification of more complicated mixtures such as cell extracts needs to be demonstrated. Improvements in the synthesis of membrane-based absorbers are also desirable. We are working to develop a completely aqueous polymer brush synthesis that is compatible with a wide range of polymeric supports.64 If functional monomers can be directly polymerized, then this would avoid the need for brush derivatization and greatly streamline the process. Other types of surface polymerization such as free radical polymerization could prove useful in this regard, although they may be difficult to control.40 Moreover, the use of acrylic acid-based polymers rather than PHEMA may increase swelling in water to enhance capacity and reduce nonspecific adsorption. Catalytic Membranes. The deposition of thin organic films is only one of several methods for forming catalytic membranes. The impregnation of membranes with a metal ion complex followed by reduction, for example, also yields active metal nanoparticles in porous materials. Future advantages of the layerby-layer method will lie in its ability to control the selectivity of the catalyst and the region of the membrane in which it is deposited.112 This control should ultimately allow the manipulation of selectivity and mass transport to optimize product purity or reaction rates. In summary, although significant challenges remain to demonstrate the commercial viability of brush-modified or polyelectrolyte-modified membranes, they currently show great potential as functional materials for a number of applications. Acknowledgment. We thank the many students and postdoctoral researchers who have contributed to this work. We also acknowledge funding from the Department of Energy, Office of Basic Energy Sciences, the National Science Foundation (CHE0616795, OIS 0530174), and the American Chemical Society Petroleum Research Fund. LA800179Z (112) Uzio, D.; Miachon, S.; Dalmon, J.-A. Catal. Today 2003, 82, 67–74.
