Functional poly(vinylidene fluoride) copolymer membranesvia surface-initiated thiol–ene click reactions

Abstract

Thermally induced graft copolymerization of allyl methacrylate (AMA) from ozone-preactivated poly(vinylidene fluoride) (PVDF) chains was first carried out in N-methyl-2-pyrrolidone (NMP) solution to produce the PVDF-g-PAMA copolymers. The PVDF-g-PAMA copolymers were cast, by phase inversion in an aqueous medium, into microporous membranes with enriched allyl groups on the membrane and pore surfaces. The pendant allyl groups in the PAMA side chains allowed the thiol–ene click grafting of thiol-functionalized moieties on the membrane surface. Thermally initiated thiol–ene click reaction with 3-mercaptopropionic acid (MPA) produced the PVDF-g-P[AMA-comb-MPA] membrane which exhibited a pH-responsive permeation behavior, with the most drastic change in permeation rate of the aqueous media being observed between pH 2 and 4. The PVDF-g-P[AMA-comb-DMAPS] membrane was synthesized via UV-initiated thiol–ene click reaction of 1,6-hexanedithiol (HDT) with the PVDF-g-PAMA membranes, using 2,2-dimethoxy-2-phenylacetophenone (DMPA) as the photoinitiator, followed by thiol-Michael addition reaction of the zwitterionic molecules, N,N′-dimethyl-(methylmethacryloyl ethyl) ammonium propanesulfonate (DMAPS). A significant reduction in microbial adhesion was observed on the PVDF-g-P[AMA-comb-DMAPS] membrane in a flow chamber, in comparison to the pristine hydrophobic PVDF membrane.

---

1. Introduction

Poly(vinylidene fluoride) (PVDF) has been widely used in the fabrication of microfiltration (MF) and ultrafiltration (UF) membrane because of its good processability, chemical resistance, well-controlled porosity, and good thermal stability.^1–4 However, the hydrophobicity of PVDF membranes can cause membrane fouling, limiting its practical application to some extent. On the other hand, porous membranes with stimuli-responsive “polymer brushes” on the surface can adapt their physicochemical properties to external stimuli, such as change in pH, temperature, ionic strength, electric field, magnetic field and mechanical stimulation.^5–8 This surface responsiveness arises mainly from structural changes at the molecular level. Thus, controlled alteration of membrane surface properties is an important approach to enhance its applications in controlled drug release and delivery,^9–12 chemical separation,^13,14 and biotechnology.¹⁵

Several approaches, such as surface coating or surface modification, have been developed to endow the conventional hydrophobic membranes with hydrophilic and stimuli-responsive properties.^16–20 Among the various coupling reactions for surface modification, thiol–ene click chemistry^21–24 has generated considerable interest in recent years. Thiol–ene click reactions proceed under benign reaction conditions with good efficiency and fast kinetics. The reactions do not require expensive and potentially toxic catalysts, and are highly inert toward other functional groups.^25,26

It would be ideal if the PVDF membrane and pore surfaces could be covalently functionalized with desirable moieties for tailoring the surface functionalities. The click coupling techniques are yet to be fully explored for tailoring the surface properties of PVDF membranes through the development of better means for incorporating the surface “clickable” functionalities. In the present work, we report the synthesis and characterization of PVDF with allyl methacrylate (AMA) polymer side chains introduced viagraft copolymerization of AMA from the ozone-preactivated PVDF in N-methyl-2-pyrrolidone (NMP). The PVDF-g-PAMA copolymer can be readily cast into microporous membranes by phase inversion in an aqueous medium. The graft copolymer membranes with active allyl groups on the membrane and pore surfaces can be further functionalized via surface-initiated thiol–ene click reactions. Because of the reactivity of surface allyl groups,^27–33 the PVDF-g-PAMA copolymer membranes can serve as a ‘clickable’ platform for tailoring the surface functionality and performance of the membranesvia ‘click’ chemistry. The reaction schemes involved are showed in Fig. 1.

Fig. 1 Schematic illustration of (1) the process of ozone pretreatment and graft copolymerization of PVDF with AMA, (2) preparation of PVDF-g-AMA microporous membrane with ‘clickable’ surface by phase inversion, and (3) preparation of the pH-responsive PVDF-g-P[AMA-comb-MPA] membranevia thermally initiated surface thiol–ene click reaction of MPA on the PVDF-g-PAMA membrane, and antibacterial PVDF-g-P[AMA-comb-DMAPS] membranevia UV-initiated surface thiol–ene click reaction of HDT on the PVDF-g-PAMA membrane and thiol-Michael addition reaction of DMAPS on the PVDF-g-P[AMA-comb-HDT] membrane.

2. Experimental section

2.1 Materials

Poly(vinylidene fluoride) (PVDF, Kynar® K-761) powders having a molecular weight of 441 000 Da were obtained from Elf Atochem, North America Inc. The monomer, allyl methacrylate (AMA, 95%), was purchased from Sigma-Aldrich Chemical Company, Milwaukee, WI, USA. After passing through a ready-to-use disposable inhibitors-removal column (Sigma-Aldrich), the AMA monomer was stored in a sealed vessel at −10 °C. The solvents, N-methyl-2-pyrrolidone (NMP, reagent grade), dichloromethane (DCM, reagent grade) and acetonitrile (HPLC grade) were purchased from Merck-Schuchardt Chemical Co., Hohenbrunn, Germany. They were used as received. 3-Mercaptopropionic acid (MPA, ≥99%), 1,6-hexanedithiol (HDT, 96%), N,N′-dimethyl-(methylmethacryloyl ethyl) ammonium propanesulfonate (DMAPS, 97%) were purchased from Sigma-Aldrich Chemical Co. and used without further purification. The radical initiator, 2,2′-azobis(2-methylpropionitrile) (AIBN, 97%), was obtained from Kanto Chemical Co. (Tokyo, Japan) and was recrystallized from anhydrous ethanol. The photoinitiator, 2,2-dimethoxy-2-phenylacetophenone (DMPA, 99%), was obtained from Sigma-Aldrich Chemical Co. and used as received.

2.2 Thermally induced graft copolymerization of AMA from the ozone-pretreated PVDF (PVDF-g-PAMA copolymer)

The PVDF powders were first dissolved in NMP to form a 7 wt% solution. A continuous stream of O₃/O₂ mixture (generated from an Azcozon RMU16-04EM ozone generator) was bubbled through 30 ml of the NMP solution of PVDF at room temperature (∼25 °C). The flow rate was adjusted to 300 l h⁻¹ to result in an ozone concentration of about 0.027 g l⁻¹ in the gaseous mixture. A treatment time of 15 min was used to achieve the desired content of peroxides in the PVDF chains since excessive ozone pretreatment could lead to over oxidation and degradation of the polymer chains.³⁴ After ozone preactivation, the polymer solution was cooled in an ice bath. An argon stream was introduced for about 30 min to remove the dissolved ozone and oxygen. The AMA monomer and NMP solvent were then added to achieve a specific [AMA]/[–CH₂CF₂–] molar feed ratio and to adjust the total solution to 40 ml. After an additional 15 min of argon purging, the temperature of the water bath was raised to 60 °C to induce the decomposition of peroxide groups on the PVDF chains and to initiate the graft copolymerization of AMA. After the desired reaction time (6 h), the reactor flask was cooled in an ice bath, and the resultant PVDF-g-PAMA copolymer was precipitated in an excess amount of absolute ethanol. The copolymer was purified by re-dissolving in NMP and re-precipitation in ethanol. The copolymer was washed in an excess volume of CH₂CCl₂ (a good solvent for PAMA homopolymer and a non-solvent for PVDF) for 48 h. The solvent was changed every 8 h. The copolymer sample was recovered and dried under reduced pressure before being subjected to subsequent characterization and reaction. The processes of ozone preactivation of PVDF and thermally induced graft copolymerization of AMA are illustrated schematically in Fig. 1.

2.3 Preparation of microporous membranes with active surfaces

Microporous membranes were prepared by phase inversion from a 15 wt% NMP solution of PVDF or PVDF-g-PAMA copolymer at room temperature. The polymer or copolymer solution was cast onto a glass plate, and subsequently immersed into a bath of doubly distilled water (non-solvent) after the polymer solution has been subjected to a brief period of evaporation (∼10 s) in air. Each membrane was left in water for about 2 h after its separation from the glass plate. The surface in contact with the glass plate is denoted as the “bottom surface”, while the outer surface in contact with air as the “top surface” in the manuscript. The membranes were put into an excess amount of doubly distilled water under constant stirring at room temperature for 24 h. The purified membranes were obtained by freeze drying for the subsequent characterization and functionalization. The thickness of the membranes so obtained was about 120 ± 10 μm.

2.4 Thermally initiated thiol–ene click reaction of MPA on the PVDF-g-PAMA membrane and pore surfaces (PVDF-g-P[AMA-comb-MPA] membranes)

Thermally initiated thiol–ene click reaction of the pendant allyl groups in the PAMA side chains on the PVDF-g-PAMA membrane with 3-mercaptopropionic acid (MPA) yielded the membrane with carboxyl-functionalized surface, or the PVDF-g-P[AMA-comb-MPA] membrane.³⁵ Fifty milligrams of the PVDF-g-PAMA membranes and 10 ml of acetonitrile were introduced into a 25 ml single-necked round-bottom flask. The solution was purged with purified argon for 30 min to remove the dissolved oxygen. MPA (0.5 ml, 5.7 mmol) and AIBN (93 mg, 0.57 mmol) were added to the solution. Then, the reactor flask was sealed under an argon atmosphere and placed in an oil bath at 70 °C to initiate the thiol–ene click reaction. After 24 h, the reactor flask was quenched in an ice bath to stop the reaction. After the reaction, the membranes were removed from the reaction mixture and subsequently washed with copious amounts of ethanol, followed by freeze drying overnight. Procedures for preparing the pH-responsive PVDF-g-P[AMA-comb-MPA] membranes are shown schematically in Fig. 1.

2.5 UV-initiated thiol–ene click reaction of HDT on the PVDF-g-PAMA membrane and pore surfaces (PVDF-g-P[AMA-comb-HDT] membranes)

The thiol–ene coupling of alkene-functionalized compounds has been carried out photochemically using 2,2-dimethoxy-2-phenylacetophenone (DMPA) as the initiator.³⁵ The UV-initiated surface thiol–ene click reaction of 1,6-hexanedithiol (HDT) onto the PVDF-g-PAMA membrane and pore surfaces was carried out in a Riko (model RH 400-10W) rotary photochemical reactor, manufactured by Riko Denki Kogyo, Chiba, Japan. The reactor was equipped with a 1000 W high-pressure Hg lamp and a constant temperature water bath (28 °C). The PVDF-g-PAMA membranes (50 mg), HDT (0.5 ml, 3.3 mmol), DMPA (10 mg, 0.04 mmol) and CH₂CCl₂ (2 ml) were introduced into a 10 ml single-necked round-bottom flask. A purified argon stream was introduced for about 30 min to degas the reaction mixture. The reaction flask was then sealed and subjected to UV irradiation for 1 h. After the reaction, the membranes were removed from the solution and exhaustively washed/extracted with copious amounts of ethanol, followed by freeze drying overnight. The process of UV-initiated thiol–ene click reaction of HDT on the PVDF-g-PAMA membrane and pore surfaces is also shown schematically in Fig. 1.

2.6 Thiol-Michael addition reaction of DMAPS onto the PVDF-g-P[AMA-comb-HDT] membranes (PVDF-g-P[AMA-comb-DMAPS] membranes)

For the thiol-Michael addition reaction,^36,37 DMAPS (100 mg, 0.36 mmol), triethylamine (30 μl, 0.22 mmol) and 0.1 M phosphate buffer solution (5 ml) were introduced into a 10 ml single-necked round-bottom flask containing the PVDF-g-P[AMA-comb-HDT] membranes (∼50 mg). A purified argon stream was introduced to degas the reaction mixture for about 30 min. The reaction was allowed to proceed at room temperature (25 °C) for 12 h to produce the PVDF-g-P[AMA-comb-DMAPS] membranes. After the reaction, the membranes were washed and extracted exhaustively with copious amounts of doubly distilled water, followed by freeze drying overnight. The thiol-Michael addition reaction of DMAPS onto the PVDF-g-P[AMA-comb-HDT] membrane and pore surfaces is also shown schematically in Fig. 1.

2.7 Materials characterization

2.7.1 Infrared spectroscopy measurements. Fourier transform infrared (FTIR) spectroscopic analysis of the graft copolymers were carried out on a Bio-Rad FTS 135 Fourier transform infrared spectra spectrophotometer, and the diffuse reflectance spectra were scanned over the range of 400–4000 cm⁻¹.

2.7.2 Nuclear magnetic resonance (NMR) spectroscopy. ¹H NMR spectra of the PVDF homopolymer and PVDF-g-PAMA copolymer were measured on a Bruker ARX 300 instrument at room temperature with deuterated DMSO as the solvent.

2.7.3 Copolymer composition analysis. Elemental analyses of the copolymer samples were performed by the Microanalysis Centre of the National University of Singapore. The bulk C contents were determined on a Perkin-Elmer 2400 elemental analyzer. The F contents were determined, on the other hand, by the Schöniger combustion method.

2.7.4 Thermogravimetric analysis . The thermal stability of the copolymers was studied by thermogravimetric analysis (TGA). The samples were heated from room temperature to about 700 °C at a heating rate of 10 °C min⁻¹ under a dry nitrogen atmosphere in a Du Pont Thermal Analyst 2100 system, equipped with a TGA 2050 thermogravimetric thermal analyzer.

2.7.5 X-Ray photoelectron spectroscopy (XPS) measurements. X-Ray photoelectron spectroscopy measurements were made on a Kratos AXIS Ultra HSA spectrometer with a monochromatized Al Kα X-ray source (1468.6 eV photons). The membranes were mounted on the standard sample studs by means of double sided adhesive tapes. The core-level signals were obtained at the photoelectron take-off angle (α, with respect to the sample surface) of 90°. All binding energies (BEs) were referenced to that of the neutral C 1s hydrocarbon peak at 284.6 eV or that of the CF₂ peak of PVDF at 290.5 eV. In peak synthesis, the line width (full-width at half-maximum, or fwhm) for the Gaussian peaks was maintained constant for all components in a particular spectrum. Surface elemental stoichiometries were determined from peak-area ratios, after correcting with the experimentally determined sensitivity factors, and were reliable to ±5%.

2.7.6 Morphology and porosity of the microporous membranes. The surface morphology of the microporous membranes was studied by scanning electron microscopy (SEM), using a JEOL 6320 scanning electron microscope. The membranes were mounted on the sample studs by means of double-sided adhesive tapes. A thin layer of palladium was sputtered onto the membrane surface prior to the SEM measurement. The measurements were performed at an accelerating voltage of 15 kV. The porosity of the porous membranes was determined by mercury porosimetry (Micromeritics, model Autopore III). Mercury was used as the sole medium, and the pressure applied ranged from 50 psia to 60 000 psia.

2.7.7 Measurements of the pH-dependent flux through the microporous membranes. The flux of aqueous solutions through the membranes was carried out under an argon pressure of 2.95 kN m⁻². The PVDF-g-P[AMA-comb-MPA] membrane was immersed in water of prescribed pH, before being mounted on the microfiltration cell (Toyo Roshi UHP-25, Tokyo, Japan). The effective membrane area was 3.14 cm². An aqueous HCl solution of the same prescribed pH value was added to the cell. Sodium chloride was added to adjust the ionic strength of the aqueous solution to 0.1 mol l⁻¹. The flux was calculated from the weight of the solution permeated per unit time and per unit area of the membrane surface. The microfiltration cell containing the permeate was kept in a thermostatted water bath for at least 20 min before the flow was initiated. The permeate pH was determined by a Mettler Toledo Delta 320 pH meter.

2.7.8 Antibacterial activity assay in the flow chamber. To evaluate the antibacterial efficiency of PVDF-g-P[AMA-comb-DMAPS] membranes, S. epidermidis (ATCC 36984) were cultivated in yeast-dextrose broth (containing 10 g l⁻¹peptone, 8 g l⁻¹ beef extract, 5 g l⁻¹NaCl, 5 g l⁻¹ glucose, and 3 g l⁻¹ yeast extract at a pH of 6.8) at 37 °C. The bacteria-containing broth was centrifuged at 2700 rpm for 10 min. After removal of the supernatant, the cells were washed twice with PBS and then resuspended in PBS at a concentration of 10⁸cells per ml. All glassware and membranes were sterilized by UV irradiation for 1 h prior to the experiment. A flow chamber was used to study microbial adhesion and detachment.^38,39 The pristine PVDF membrane, the PVDF-g-P[AMA-comb-HDT] membrane and the PVDF-g-P[AMA-comb-DMAPS] membrane (about 1 cm × 4 cm) were placed onto a glass plate and then positioned in the flow chamber. Subsequently, a microbial suspension was recirculated through the system for 8 h, followed by PBS for 0.5 h, at a constant flow rate of 1 ml min⁻¹. After the incubation period, the membranes were fixed with 3% glutaraldehyde at 4 °C overnight. After step dehydration with serial ethanol for 10 min each and coating of platinum, bacterial fouling of the substrates under the continuous-flow conditions was examined by SEM.

3. Results and discussion

3.1 Preparation of the poly(vinylidene fluoride)-graft-poly(allyl methacrylate) copolymers (PVDF-g-PAMA copolymers)

The process and mechanism of radical-initiated graft copolymerization of vinyl monomers from ozone-pretreated PVDF have been described earlier.³⁴ Well-defined polymers with pendant vinyl groups have been of great interest because of the useful applications of vinyl groups in copolymerization, cross-linking reaction,^27–29 radical coupling,^30,31 post-modification^32,33 and transformation into other functional groups.

The Fourier transform infrared (FTIR) spectrum of the PVDF-g-PAMA copolymer is compared to that of the PAMA homopolymer and that of the PVDF homopolymer in Fig. 2a–c. The adsorption bands at the wavenumbers of about 1650 cm⁻¹, 1730 cm⁻¹ and 3080 cm⁻¹ are associated, respectively, with the alkene stretching, ester stretching and =C–H stretching of the PAMA polymer chains.⁴⁰ On the other hand, the adsorption band in the region of 1120–1280 cm⁻¹, characteristic of the –CF₂– functional groups of PVDF⁴⁰ is also present in the copolymer sample. Thus, the FTIR spectroscopic results are consistent with the presence of grafted PAMA chains on the PVDF backbone and the preservation of allyl groups during graft copolymerization. Theoretically, methacrylic double bonds are more reactive than allylic double bonds toward free-radical polymerization.^27,41Copolymers of allyl methacrylate (AMA) that contain low concentration of pendant allylic double bonds do not appear to undergo extensive gelation (cross-linking) reaction.

Fig. 2 FTIR spectra of the (a) PVDF homopolymer, (b) PVDF-g-PAMA copolymer with the ([–AMA–]/[–CH₂CF₂–])_bulk molar ratio of 0.26 and (c) PAMA homopolymer.

The bulk graft concentrations of the copolymers can be derived from the carbon to fluorine ratio, obtained from elemental analyses. The graft concentration in terms of the number of AMA repeat units per PVDF repeat unit, or the ([–AMA–]/[–CH₂CF₂–])_bulk molar ratio, can be readily obtained from the ([C]/[F])_bulk molar ratio by taking into account the carbon stoichiometries of the graft and the main chains and the carbon to fluorine ratio of the PVDF main chain. Thus, the ([–AMA–]/[–CH₂CF₂–])_bulk molar ratio can be calculated from eqn. (1):

([–AMA–]/[–CH₂CF₂–])_bulk = (2/7)([C]–[F])_bulk/[F]_bulk (1)

where the factors 2 and 7 are introduced to account for the fact that there are 2 and 7 carbon atoms per repeat unit of PVDF and PAMA, respectively. Fig. 3 shows the dependence of the AMA polymer graft concentration in the PVDF-g-PAMA copolymer, expressed as the ([C]/[F])_bulk and ([–AMA–]/[–CH₂CF₂–])_bulk molar ratios, on the [AMA] to [–CH₂CF₂–] molar feed ratio used for the thermally induced graft copolymerization. The graft concentration increases gradually with the increase in AMA monomer concentration used for graft copolymerization.

Fig. 3 Effect of monomer molar feed ratio (the [AMA]/[–CH₂CF₂–] ratio) on the bulk graft concentration, ([C]/[F])_bulk ratio and ([–AMA–]/[–CH₂CF₂–])_bulk ratio, of the PVDF-g-PAMA copolymers.

The ¹H NMR spectra of the PVDF homopolymer and the PVDF-g-PAMA copolymer (([–AMA–]/[–CH₂CF₂–])_bulk = 0.26) are shown in Fig. 4a and b, respectively. The chemical shifts at δ = 2.2 ppm (a1) and in the range of δ = 2.7–3.0 ppm (a2) in Fig. 4a and b are attributable to the head-to-head (hh) or tail-to-tail (tt) stereo-regularities and the head-to-tail (ht) bonding arrangements of the PVDF main chains, respectively.⁴²Graft polymerization of AMA from PVDF has resulted in the appearance of chemicals shifts in the range of δ = 0.6–0.9 ppm (c in Fig. 4b), attributable to the C–CH₃ group of PAMA. The chemical shift at δ = 4.3–4.6 ppm (d) is attributable to the –CO₂CH₂ species, while the chemical shift at δ = 5.8–6.0 ppm (e) and δ = 5.2–5.5 ppm (f) are associated with the –CH=CH₂ species of the PAMA side chains, respectively.⁴¹ The integrals of –CO₂CH₂ and –CH=CH₂ resonances in ¹H NMR spectrum of the PVDF-g-PAMA copolymer is about 2 : 1 : 2, indicating that the pendant allyl groups remain intact and the graft copolymerization involves predominantly the methacrylic groups under the reaction condition, otherwise the process can lead to cross-linking.^43,44 The PVDF-g-PAMA copolymer remains soluble and can be readily cast into porous membranes (see below). Furthermore, a ratio of about 1 : 3 for the allyl protons (f in Fig. 4b) to the protons of PVDF main chains (a in Fig. 4b) indicates that the bulk molar ratio of AMA segments in the copolymer is about 0.25, which coincides approximately with the results from elemental analysis.

Fig. 4 ¹H NMR spectra of the (a) PVDF homopolymer and (b) PVDF-g-PAMA copolymer with a ([–AMA–]/[–CH₂CF₂–])_bulk molar ratio of 0.26 in DMSO.

One of the unique properties of the PVDF copolymer membranes is their outstanding thermal stability. Fig. 5 shows the respective thermogravimetric analysis (TGA) curves of the PVDF homopolymer (Curve 1), the three PVDF-g-PAMA copolymers of different bulk graft concentrations (Curves 2, 3 and 4 for ([–AMA–]/[–CH₂CF₂–])_bulk ratios of 0.09, 0.17 and 0.26, respectively), and the PAMA homopolymer (Curve 5). In comparison to the PVDF and PAMA homopolymers, a distinct two-step degradation process is observed for the PVDF-g-PAMA copolymer samples. The first major weight loss occurs at about 250 °C, attributable to the thermal decomposition of the PAMA segments, while the second major weight loss commences at about 475 °C, attributable to the thermal decomposition of the PVDF main chains. The extent of the first major weight loss at about 250 °C coincides approximately with the weight content of the PAMA polymer in the respective graft copolymers obtained from elemental analysis. Thus, the TGA results are consistent with the elemental analysis results and confirm quantitatively the extent of graft copolymerization of AMA from the ozone-pretreated PVDF main chains.

Fig. 5 Thermogravimetric analysis curves of (1) the PVDF homopolymer; the PVDF-g-PAMA copolymers with ([–AMA–]/[–CH₂CF₂–])_bulk molar ratios of (2) 0.09, (3) 0.17, (4) 0.26; (5) the PAMA homopolymer.

3.2 Preparation of the PVDF-g-PAMA copolymer membranes by phase inversion

The PVDF-g-PAMA microporous membranes were fabricated by phase inversion of a 15 wt% NMP solutions of the various PVDF-g-PAMA copolymers (Table 1) in doubly distilled water at room temperature. The SEM micrographs in Fig. 6a and b illustrate the difference in surface morphology of the microporous membrane cast from PVDF and the PVDF-g-PAMA copolymer (([–AMA–]/[–CH₂CF₂–])_bulk = 0.26). The SEM images reveal that the PVDF-g-PAMA membrane has a more well-defined pore size distribution and a higher degree of porosity, induced by the PAMA graft chains during phase inversion in the aqueous medium, than those of the pristine PVDF membrane. The porosities of the copolymer membranes, as determined by mercury porosimetry, are summarized in Table 1. With the increase in AMA graft concentration, the porosity of the PVDF-g-PAMA membranes increases correspondingly from 58% to 82%. The increase in porosity with the increase in AMA graft concentration confirms that the membrane morphology is dependent on the PAMA chain length and density, and is thus controllable.

Table 1 Characterization of the PVDF, PVDF-g-PAMA Membranes

Membrane Sample	([–AMA–]/[–CH2CF2–])bulk^a	([–AMA–]/[–CH2CF2–])surface^b	Porosity (%)^c
a Derived from the atom ratio of C and F (obtained from elemental analyses) according to eqn (1). b Derived from the curve-fitted C 1s peak component area ratios of (1/2)[–O–C=O]/[C–F2] of the respective sample in Fig. 8 since there are one [C–F2] unit in the PVDF backbone and two [–O–C=O] units in the P[AMA-comb-MPA] side chains. c Determined by Hg porosimetry. d PVDF microporous membranes obtained from Millipore Corporation. ‘d’ stands for the standard pore size of the commercial hydrophilic and microporous membranes.
PVDF (d = 0.45 μm)^d	—	—	72
PVDF (d = 0.65 μm)^d	—	—	80
PVDF-g-PAMA	0.09	0.15	58
	0.17	0.24	73
	0.26	0.31	82

---

Fig. 6 SEM micrographs of the microporous membranes cast from the 15 wt% NMP solution of corresponding copolymers by phase inversion: (a) the pristine PVDF membrane; (b) PVDF-g-PAMA membrane with a ([–AMA–]/[–CH₂CF₂–])_surface molar ratio of 0.31; the PVDF-g-P[AMA-comb-MPA] membrane synthesized by surface-initiated thiol–ene click reaction with ([–COOH]/[–CH₂CF₂–])_surface molar ratios of (c) 0.14, (d) 0.22, (e) 0.30; (f) PVDF-g-P[AMA-comb-DMAPS] membrane prepared from PVDF-g-PAMA membrane with a ([–AMA–]/[–CH₂CF₂–])_surface molar ratio of 0.31. All images shown are the surfaces in contact with the glass substrate (“bottom” surface) during membrane casting by phase inversion.

Fig. 7a and b show the respective wide scan and C 1s core-level spectra of the pristine PVDF and the PVDF-g-PAMA (([–AMA–]/[–CH₂CF₂–])_bulk = 0.26) membranes cast by phase inversion. In the case of pristine PVDF membrane (Fig. 7a), only C and F signals are observed in the wide scan spectrum. The C 1s core-level spectrum can be curve-fitted with three peak components, with binding energies (BEs) at 284.6 eV for the neutral CH species, 285.8 eV for the CH₂ species (adjacent to CF₂ in PVDF), and 290.5 eV for the CF₂ species.⁴⁵ The [-CH₂-] : [–CF₂−] peak component area ratio of about 1 : 1 is consistent with the chemical structure of PVDF. For the PVDF-g-PAMA membrane, the appearance of a distinctive O 1s signal in the wide scan spectrum of Fig. 7b indicates that AMA has been graft copolymerized from the PVDF main chains. The two new C 1s peak components with BE at 286.2 eV for the C–O species and at 288.5 eV for the O–C=O species in Fig. 7b can also be assigned to the grafted AMA chains. The [C–O] : [O–C=O] peak component area ratio of about 1 : 1 is in good agreement with the theoretical ratio based on the chemical structure of AMA. On the other hand, the graft concentration at the membrane surface, or the ([–AMA–]/[–CH₂CF₂–])_surface ratio of 0.31, as determined from the ([O–C=O]/[–CF₂−]) peak component area ratio in the C 1s core-level spectrum of the copolymer membrane in Fig. 7b, is higher than the bulk ratio of 0.26. Thus, surface enrichment of the more hydrophilic PAMA component has occurred in the copolymer membranes during the phase inversion process in the aqueous medium.

Fig. 7 XPS wide-scan and C 1s core-level spectra of the (a) pristine PVDF membrane, and (b) PVDF-g-PAMA membrane with a ([–AMA–]/[–CH₂CF₂–])_surface molar ratio of 0.31.

3.3 Functionalization of the PVDF-g-PAMA membranevia click grafting of MPA: pH-responsive PVDF-g-P[AMA-comb-MPA] membrane

The thermally induced thiol–ene coupling reaction in the presence of AIBN was employed for the click grafting of 3-mercaptopropionic acid (MPA) on the PVDF-g-PAMA membrane to give rise to the PVDF-g-P[AMA-comb-MPA] membrane. The SEM images of the respective PVDF-g-P[AMA-comb-MPA] microporous membranes with ([–COOH]/[–CH₂CF₂–])_surface molar ratios of 0.14, 0.22 and 0.30 (see XPS results below), obtained from the corresponding PVDF-g-PAMA membranes with ([–AMA–]/[–CH₂CF₂–])_surface ratio of 0.15, 0.24 and 0.31 (Table 1), are shown in Fig. 6c–e. The increase in PAMA graft concentration in the PVDF-g-PAMA copolymer also gives rise to the corresponding increase in porosity of the PVDF-g-PAMA membranes, and thus that of the PVDF-g-P[AMA-comb-MPA] membranes, as can be seen from the SEM images in Fig. 6c–e. Click grafting of MPA onto the PVDF-g-PAMA membrane does not cause a significant change in morphology and porosity of the membrane (compare Fig. 6b–e).

Successfully click grafting of MPA on the PVDF-g-PAMA membrane surface was also revealed by XPS results. Fig. 8a–c show the respective C 1s core-level spectra of the PVDF-g-P[AMA-comb-MPA] membranes prepared from PVDF-g-PAMA membranes with ([–AMA–]/[–CH₂CF₂–])_surface molar ratios of 0.15, 0.24 and 0.31 (Table 1). With the increase in AMA graft concentration in the starting PVDF-g-PAMA copolymer membranes, the ([–COOH]/[–CH₂CF₂–])_surface molar ratios of the resultant PVDF-g-P[AMA-comb-MPA] membranes increase correspondingly to 0.14, 0.22 and 0.30 (Fig. 8a–c). The C 1s core-level spectra were curve-fitted with five peak components. The components with binding energies (BEs) at 284.6 eV, 286.2 eV and 288.5 eV are attributed to the hydrocarbon backbone, C–O and O=C–O species of the grafted AMA polymer chains with clicked MPA.⁴⁵ The other two peak components at the BEs of 285.8 eV and 290.5 eV are attributable to the CH₂ and CF₂ species, respectively, of the PVDF main chain. The C–S species of the MPA moieties have a peak component at the BE of 285.8 eV.⁴⁵ The appearance of S 2p and S 2s signals in the wide scan spectrum of the PVDF-g-P[AMA-comb-MPA] membrane is consistent with a membrane surface modified by thiol–ene click reaction. The surface concentration of carboxylic acid groups, or the ([–COOH]/[–CH₂CF₂])_surface ratio, on the PVDF-g-P[AMA-comb-MPA] membrane can be obtained from the XPS-derived sulfur to fluorine ratio and eqn (2):

([–COOH]/[–CH₂CF₂–])_surface = 2[S]_surface/[F]_surface (2)

where the factor 2 accounts for the fact that there are two fluorine atoms per repeat unit of the PVDF polymer and one sulfur atom per unit of the 3-mercaptopropionic acid molecules, respectively. From Table 1 and Fig. 8a–c, it can be seen that the surface concentration of carboxylic acid groups is comparable to the corresponding ([–AMA–]/[–CH₂CF₂–])_surface graft concentration of the starting copolymer membranes, indicating that most of the vinyl groups have undergone the thiol–ene click reaction.

Fig. 8 XPS wide-scan and C 1s core-level spectra of three PVDF-g-P[AMA-comb-MPA] microporous membranes with ([–COOH]/[–CH₂CF₂–])_surface molar ratio of (a) 0.14, (b) 0.22 and (c,d) 0.30.

The pH-dependent flux of aqueous solutions through the PVDF-g-P[AMA-comb-MPA] microporous membrane is shown in Fig. 9. The permeation rate of the aqueous solution through the PVDF-g-P[AMA-comb-MPA] membrane increases with the decrease in solution pH from 6 to 1, with the most drastic increase being observed at the permeate pH between 2 and 4 (curves 1, 2 and 3). The permeability of aqueous solutions through the commercial PVDF microporous membranes with standard pore diameters of 0.65 μm and 0.45 μm is pH-independent (curves 4 and 5, respectively). The change in permeation rate in response to the change in solution pH can be attributed to the change in conformation of the –COOH groups-containing graft chains on the PVDF-g-P[AMA-comb-MPA] membrane and pore surfaces. At pH > 3, the carboxylic groups of the PVDF-g-P[AMA-comb-MPA] copolymer (pK_a ≈ 4.2) are ionized, or deprotonated, to become negatively charged. Strong electrostatic repulsion among the carboxylic ions, together with their strong interaction with the aqueous solution, forces the P[AMA-comb-MPA] segments to adopt a highly extended conformation (see inset in Fig. 9b).⁴⁶ As a result, the effective pore dimension, and thus the permeation rate of the aqueous solution, is reduced. On the other hand, the P[AMA-comb-MPA] polymer segments assume a more compact or associated conformation in the absence of ionization at low pH (see inset in Fig. 9a). The steric obstruction of the pores of membrane is substantially reduced. Hence, the permeation rate is increased. The data in Fig. 9 also show that the extent of change in permeability of the PVDF-g-P[AMA-comb-MPA] membrane increases with the increase in concentration of the carboxylic acid group (compare curves 1 and 2 to curve 3), arising from a larger “dragging” effect exerted by the increasing number of hydrophilic carboxylic acid groups at the solid–fluid interface.^46,47

Fig. 9 pH-dependent permeability of aqueous solution through the pristine PVDF and PVDF-g-P[AMA-comb-MPA] microporous membrane. Curves 1, 2 and 3 are obtained from three PVDF-g-P[AMA-comb-MPA] microporous membranes with surface graft concentrations ([–COOH]/[–CH₂CF₂–])_surface of 0.30, 0.22, 0.14, respectively. Curves 4 and 5 are from fluxes through two commercial PVDF membranes (standard pore diameter: d = 0.65 and 0.45 μm, respectively).

3.4 Functionalization of the PVDF-g-PAMA membranevia click grafting of HDT and thiol-Michael addition of DMAPS: antibacterial PVDF-g-P[AMA-comb-DMAPS] membrane

Zwitterionic molecules, such as N,N′-dimethyl-(methylmethacryloyl ethyl) ammonium propanesulfonate (DMAPS), with quaternary ammonium group exhibit good antimicrobial activities.^48–53Polymers containing quaternary ammonium groups have been immobilized on various membranes, such as polypropylene (PP),⁵⁴ polysulfone (PSF)⁵⁵ and cellulose membranes.⁵⁶ Fouling of PVDF membranes can be attributed partially to the adsorption of organic species and adhesion of microbial cells. PVDF membranes with an antibacterial and hydrophilic surface can eliminate bacteria, inhibit the growth of microbes and reduce membrane fouling, making them very attractive for water treatment.

Surface-initiated thiol–ene click reaction of 1,6-hexanedithiol (HDT) on the PVDF-g-PAMA membrane leads to the formation of reactive thiol groups on the membrane and pore surfaces. The presence of two reactive thiol groups in HDT allows this molecule to serve as a versatile linker for a variety of alkene-containing molecules via thiol-Michael addition reaction.^36,37Fig. 10 shows the respective XPS wide-scan, C 1s and S 2p core-level spectra of the PVDF-g-P[AMA-comb-HDT] membrane (from PVDF-g-PAMA membrane with a ([–AMA–]/[CH₂CF₂–])_surface ratio of 0.31) and the corresponding PVDF-g-P[AMA-comb-DMAPS] membrane. In the wide-scan spectrum of the PVDF-g-P[AMA-comb-HDT] membrane, not only are the C 1s, F 1s and O 1s signals detected, the S 2p signals are also discernible. The C 1s core-level spectrum of the surface can be curve-fitted into five peak components with BEs at about 284.6, 285.8, 286.2, 288.5 and 290.5 eV,⁴⁵ attributable to the C–H, (–CH₂−)_PVDF/C–S, C–O, O=C–O and (–CF₂−)_PVDF species, respectively (Fig. 10a). The S 2p core-level spectrum of the PVDF-g-P[AMA-comb-HDT] membrane shows a spin-orbit split doublet with BEs for the S 2p_3/2 and S 2p_1/2 peak components at 163.1 and 164.3 eV, respectively, attributable to the covalently bonded sulfur (C–S) species.^45,57

Fig. 10 XPS wide-scan, C 1s and S 2p core-level spectra of the (a) PVDF-g-P[AMA-comb-HDT] membrane and (b) PVDF-g-P[AMA-comb-DMAPS] membrane.

The concentrations of C–S groups on the PVDF-g-P[AMA-comb-HDT] membrane, or the ([C–S]/[–CH₂CF₂])_surface ratio, can be obtained directly from the XPS-derived sulfur to fluorine ratio, since there are two fluorine atoms per repeat unit of the PVDF polymer and two sulfur atoms per unit of the 1,6-hexanedithiol molecules. The surface concentration of the thiol end (–SH) groups of the PVDF-g-P[AMA-comb-HDT] membrane is about 0.26, which is lower than the corresponding ([–AMA–]/[–CH₂CF₂–])_surface graft concentration (about 0.31) of the starting copolymer membrane, indicating that over 80% of the vinyl groups have undergone the thiol–ene click reaction.

In the wide-scan spectrum of the PVDF-g-P[AMA-comb-DMAPS] membrane, the presence of the N 1s signal is attributed to the coupled zwitterionic N,N'-dimethyl-(methylmethacryloyl ethyl) ammonium propanesulfonate (DMAPS) molecules since there is no nitrogen species on the PVDF-g-P[AMA-comb-HDT] membrane (Fig. 10b). The C 1s core-level spectrum of the PVDF-g-P[AMA-comb-DMAPS] membrane can be curve-fitted with five peak components using the following approach. The two peak components of about equal intensities at the BEs of about 285.8 and 290.5 eV are assigned to the –CH₂– and –CF₂– species of PVDF, respectively. The remaining area of the peak component at the BE of 285.8 eV is assigned to the C–S species, as it has the same BE as that of the CH₂ species (in PVDF). The component at the BE of about 288.5 eV is assigned to the O–C=O species of the grafted AMA polymer chains and DMAPS molecules. The peak component at the BE of about 287 eV arises from the combined contribution of the C–O species, C–N⁺ species,⁵⁸ and the C–SO₃⁻ species of the DMAPS polymer side chains.⁸ In the S 2p core-level spectrum, two spin-orbit split doublets are discernible. The main S 2p doublet, with the S 2p_3/2 and 2p_1/2 peak components at the BEs of 163.1 and 164.3 eV, respectively, is associated with the formation of a normal C–S bond. On the other hand, the S 2p doublet at the BEs of 167.0 and 168.2 eV can be assigned to sulfonate C–SO₃⁻ species. The fact that the two sulfur species (C–S and C–SO₃⁻) have a spectral area ratio of about 2 : 1 suggests that all the thiol end (–SH) groups on the PVDF-g-P[AMA-comb-HDT] membrane surface have undergone thiol-Michael addition reaction with DMAPS molecules. The XPS results also indicate that the combined thickness of the HDT and DMAPS layers is still below the sampling depth of the XPS technique (about 8 nm in an organic matrix⁵⁹).

In water treatment, convective mass transport, other than sedimentation or diffusion of suspended microorganisms, is the major mechanism that controls the rate of microbial adhesion. The flow chamber is a commonly used system in mimicking microbial adhesion to surfaces under convective mass transport.^38,39Fig. 11 shows the SEM micrographs of the membranes after exposure to PBS suspension of S. epidermidis, containing 10⁸cells per ml initially, for 8 h in the flow chamber. Bacteria cells tend to adhere and aggregate to form colonies on the pristine PVDF membrane and the PVDF-g-P[AMA-comb-HDT] membrane, as shown in Fig. 11a–d. In contrast, a significant decrease in the number of adhered bacteria is observed on the PVDF-g-P[AMA-comb-DMAPS] membrane (Fig. 11e and f), indicating a distinctively higher antibacterial efficiency of the membrane in the presence of quaternary ammonium groups of the DMAPS molecules on the surfaces. The bacterial cells on the PVDF-g-P[AMA-comb-DMAPS] membranes are sparsely distributed as single or double cells in the order of 0.6 μm or less in diameter. Also, there is little evidence of growth or proliferation of the bacterial cells on the membrane surfaces. Thus, the good antibacterial property of the PVDF-g-P[AMA-comb-DMAPS] membrane has been ascertained, and biofilm formation is inhibited on the membrane surfaces in contact with the bacteria.

Fig. 11 SEM images of the (a, b) pristine PVDF membrane, (c, d) PVDF-g-P[AMA-comb-HDT] membrane (from the PVDF-g-PAMA membrane with a ([–AMA–]/[–CH₂CF₂–])_surface ratio of 0.31), and (e, f) PVDF-g-P[AMA-comb-DMAPS] membrane after exposure to PBS suspension of S. epidermidis (initial cell concentration of 10⁸cells per ml) in a flow chamber at a constant flow rate of 1 ml min⁻¹ for 8 h.

4. Conclusions

PVDF-g-PAMA copolymers bearing pendant allyl functionalities were synthesized viagraft copolymerization of allyl methacrylate (AMA) from the ozone-preactivated poly(vinylidene fluoride) (PVDF) backbones. The microporous membranes prepared by phase inversion of the PVDF-g-PAMA copolymers of different graft concentrations in an aqueous medium showed enrichment of the allyl groups on the membrane and pore surfaces. As these pendant allyl groups can be utilized for the thiol–ene click reaction, such ‘clickable’ membrane surfaces allow convenient functionalization of the membranes. Surface-initiated click grafting of 3-mercaptopropionic acid (MPA) and 1,6-hexanedithiol (HDT) was demonstrated. The so-obtained PVDF-g-P[AMA-comb-MPA] membranes exhibited pH-dependent permeability for aqueous solutions, while the resulting PVDF-g-P[AMA-comb-HDT] membrane underwent thiol-Michael addition reaction of DMAPS to produce the PVDF-g-P[AMA-comb-DMAPS] membrane with good antibacterial properties under flow conditions. Thus, the surface allyl-functionalized PVDF membranes can serve as a universal platform for the convenient and versatile functionalization of PVDF membranesvia the simple thiol–ene click grafting reactions.

References

1. K. Li, Chem. Eng. Technol., 2002, 25, 203.
2. R. Souzy, B. Ameduri and B. Boutevin, Prog. Polym. Sci., 2004, 29, 75.
3. B. J. Junginckel, in The Polymeric Materials Encyclopedia, J. C. Salamone, ed. CRC Press, Boca Raton, FL, 1996, vol. 7, pp. 7114–7123.
4. D. A. Seiler, in Modern Fluoropolymers; PVDF in the Chemical Process Industry, ed. J. Scheirs, Wiley, New York, 1997, ch. 25, pp. 487–506.
5. D. Wandera, S. R. Wickramasinghe and S. M. Husson, J. Membr. Sci., 2010, 357, 6.
6. T. Chen, R. Ferris, J. M. Zhang, R. Ducker and S. Zauscher, Prog. Polym. Sci., 2010, 35, 94.
7. D. M. He, H. Susanto and M. Ulbricht, Prog. Polym. Sci., 2009, 34, 62.
8. G. Q. Zhai, S. C. Toh, W. L. Tan, E. T. Kang and K. G. Neoh, Langmuir, 2003, 19, 7030.
9. M. A. M. E. Vertommen, H. J. L. Cornelissen, C. H. J. T. Dietz, R. Hoogenboom, M. F. Kemmere and J. T. F. Keuurentjes, J. Membr. Sci., 2008, 322, 243.
10. J. Shi, X. P. Liu, Y. J. Shang and S. K. Cao, J. Membr. Sci., 2010, 352, 262.
11. K. Hu and J. M. Dickson, J. Membr. Sci., 2009, 337, 9.
12. D. F. Stamatialis, B. J. Papenburg, M. Girones, S. Saiful, S. N. M. Bettahalli, S. Schmitmeier and M. Wessling, J. Membr. Sci., 2008, 308, 1.
13. R. Huang, K. Z. Mah, M. Malta, L. K. Kostanski, C. D. M. Filipe and R. Ghosh, J. Membr. Sci., 2009, 345, 177.
14. R. Huang, L. K. Kostanski, C. D. M. Filipe and R. Ghosh, J. Membr. Sci., 2009, 336, 42.
15. S. Yu, I. Benzeval, A. Bowyer and J. Hubble, J. Membr. Sci., 2009, 339, 138.
16. J. F. Hester, P. Banerjee and A. M. Mayes, Macromolecules, 1999, 32, 1643.
17. M. L. Sforca, I. V. P. Yoshida and S. P. Nunes, J. Membr. Sci., 1999, 159, 197.
18. J. F. Hester, P. Banerjee, Y. Y. Won, A. Akthakul, M. H. Acar and A. M. Mayes, Macromolecules, 2002, 35, 7652.
19. B. Ameduri, Chem. Rev., 2009, 109, 6632.
20. A. Akthakul, A. I. Hochbaum, F. Stellacci and A. M. Mayes, Adv. Mater., 2005, 17, 532.
21. C. E. Hoyle and C. N. Bowman, Angew. Chem., Int. Ed., 2010, 49, 1540.
22. C. E. Hoyle, A. B. Lowe and C. N. Bowman, Chem. Soc. Rev., 2010, 39, 1355.
23. J. Ma, C. Cheng, G. R. Sun and K. L. Wooley, Macromolecules, 2008, 41, 9080.
24. R. M. Hensarling, V. A. Doughty, J. W. Chan and D. L. Patton, J. Am. Chem. Soc., 2009, 131, 14673.
25. C. Wendeln, S. Rinnen, C. Schulz, H. F. Arlinghaus and B. J. Ravoo, Langmuir, 2010, 26, 15966.
26. A. Bertin and H. Schlaad, Chem. Mater., 2009, 21, 5698.
27. K. Ishizu, K. Tsubaki, A. Mori and S. Uchida, Prog. Polym. Sci., 2003, 28, 27.
28. E. Lovelady, S. D. Kimmins, J. J. Wu and N. R. Cameron, Polym. Chem., 2011, 2, 559.
29. M. Eriksson, K. Hult, E. Malmstrom, M. Johansson, S. M. Trey and M. Martinelle, Polym. Chem., 2011, 2, 714.
30. R. A. Prasath, M. T. Gokmen, P. Espeel and F. E. D. Prez, Polym. Chem., 2010, 1, 685.
31. L. M. Campos, I. Meinel, R. G. Guino, M. Schierhorn, N. Gupta, G. D. Stucky and C. J. Hawker, Adv. Mater., 2008, 20, 3728.
32. K. L. Killops, L. M. Campos and C. J. Hawker, J. Am. Chem. Soc., 2008, 125, 5062.
33. L. M. Campos, K. L. Killops, R. Sakai, J. M. J. Paulusse, D. Damiron, E. Drockenmuller, B. W. Messmore and C. J. Hawker, Macromolecules, 2008, 41, 7063.
34. P. Wang, K. L. Tan, E. T. Kang and K. G. Neoh, J. Mater. Chem., 2001, 11, 783.
35. B. Korthals, M. C. Morant-Minana, M. Schmid and S. Mecking, Macromolecules, 2010, 43, 8071.
36. R. Matsuno, K. Takami and K. Ishihara, Langmuir, 2010, 26, 13028.
37. J. W. Chan, C. E. Hoyle, A. B. Lowe and M. Bowman, Macromolecules, 2010, 43, 6381.
38. I. Cringus-Fundeanu, J. Luijten, H. C. van der Mei, H. J. Busscher and A. J. Schouten, Langmuir, 2007, 23, 5120.
39. D. P. Bakker, A. van der Mats, G. J. Verkerke, H. J. Busscher and H. C. van der Mei, Appl. Environ. Microbiol., 2003, 69, 6280.
40. The Systematic Identification of Organic Compounds, ed. R. L. Shriner, C. K. E. Hermann, T. C. Morrill, D. Y. Curtin and R. C. Fuson, J. Wiley & Sons, New York, 7th edn, 1998.
41. R. Nagelsdiek, M. Mennicken, B. Maier, H. Keul and H. Hocker, Macromolecules, 2004, 37, 8923.
42. Q.-T. Pham, R. Petiaud, M.-F. Llauro and H. Waton, Proton and Carbon NMR Spectra of Polymers, John Wiley & Sons, Chichester, UK, 1984, vol. 3, p. 455.
43. R. Paris and J. L. De la Fuente, J. Polym. Sci., Part A: Polym. Chem., 2005, 43, 2395.
44. R. Paris and J. L. De La Fuente, J. Polym. Sci., Part A: Polym. Chem., 2005, 43, 6247.
45. The Handbook of X-ray Photoelectron Spectroscopy, ed. J. F. Moulder, W. F. Stickle, P. E. Sobol and K. Bomben, Perkin-Elmer Corporation (Physical Electronics), Wellesley, MA, 2nd edn, 1992, pp. 216–217.
46. Y. Ito, M. Inaba, D. J. Chung and Y. Imanishi, Macromolecules, 1992, 25, 7313.
47. L. Ying, P. Wang, E. T. Kang and K. G. Neoh, Macromolecules, 2002, 35, 673.
48. T. Tashiro, Macromol. Mater. Eng., 2001, 286, 63.
49. Y. Chang, W. Yandi, W. Y. Chen, Y. J. Shih, C. C. Yang, Y. Chang, Q. D. Ling and A. Higuchi, Biomacromolecules, 2010, 11, 1101.
50. G. Cheng, Z. Zhang, S. F. Chen, J. D. Bryers and S. Y. Jiang, Biomaterials, 2007, 28, 4192.
51. K. D. Demadis, A. Katsetzi, K. Pachis and V. M. Ramos, Biomacromolecules, 2008, 9, 3288.
52. S. Colak, C. F. Nelson, K. Nusslein and G. N. Tew, Biomacromolecules, 2009, 10, 353.
53. G. S. Georgiev, E. B. Kamemska, E. D. Vassileva, I. P. Kamenova, V. T. Georgieva, S. B. Iliev and I. A. Ivanov, Biomacromolecules, 2006, 7, 1329.
54. H. J. Yu, Y. M. Cao, G. D. Kang, J. H. Liu, M. Li and Q. Yuan, J. Membr. Sci., 2009, 342, 6.
55. Y. F. Yang, Y. Li, Q. L. Li, L. S. Wan and Z. K. Xu, J. Membr. Sci., 2010, 362, 255.
56. P. S. Liu, Q. Chen, S. S. Wu, J. Shen and S. C. Lin, J. Membr. Sci., 2010, 350, 387.
57. A. S. Goldmann, A. Walther, L. Nebhani, R. Joso, D. Ernst, K. Loos, C. Barner-Kowollik, L. Barner and A. H. E. Muller, Macromolecules, 2009, 42, 3707.
58. W. Chen, K. G. Neoh, E. T. Kang, K. L. Tan, D. J. Liaw and C. C. Huang, J. Polym. Sci., Part A: Polym. Chem., 1998, 36, 357.
59. K. L. Tan, L. L. Woon, H. K. Wong, E. T. Kang and K. G. Neoh, Macromolecules, 1993, 26, 2832.

---