Photo-irradiation for preparation, modification and stimulation of polymeric membranes

Progress in Polymer Science 34 (2009) 62–98

Contents lists available at ScienceDirect

Progress in Polymer Science journal homepage: www.elsevier.com/locate/ppolysci

Photo-irradiation for preparation, modification and stimulation of polymeric membranes Dongming He 1 , Heru Susanto 1,2 , Mathias Ulbricht ∗,1 Lehrstuhl für Technische Chemie II, Universität Duisburg-Essen, 45117 Essen, Germany

a r t i c l e

i n f o

Article history: Received 7 December 2007 Received in revised form 17 August 2008 Accepted 22 August 2008 Available online 8 October 2008 Keywords: Photo-irradiation UV irradiation Photo-polymerization Photo-grafting Polymeric membrane Membrane preparation Membrane modification Stimuli-responsive membrane Membrane separation

a b s t r a c t The recent developments in combining photo-irradiation-based and membrane technologies are analyzed in this review. It is emphasized that the effects of photo-initiated reactions onto properties of polymeric membranes will largely depend on the nature of the membrane’s barrier (i.e., porous vs. non-porous, uncharged vs. charged, or involving affinity interactions). For de novo preparation of membranes from low molar mass or soluble precursors, photo-initiated polymerization and photo-cross-linking are the main pathways, while photo-degradation for pore formation is only rarely applied. Membrane functionalization (modification) is described with many examples, organized into photo-cross-linking of membranes and photo-grafting of membrane surfaces, either via “grafting-to” or via “grafting-from” routes. Photo-stimulation of barrier properties is an attractive concept to create stimuli-responsive membranes, and various ways to use photo-chromic moieties for that purpose are discussed. Overall, photo-irradiation-based methods can be very versatile enabling technologies to improve the performance of polymeric membranes in technical separations (e.g., in gas separation, pervaporation, ultra- and microfiltration or as membrane adsorbers) and other processes (e.g., for controlled release or in sensor systems), and they will definitely also contribute to the development of entirely novel membrane-based materials. © 2008 Elsevier Ltd. All rights reserved.

Abbreviations: AA, acrylic acid; AAm, acrylamide; ABMPEG, ␣-4-azidobenzoyl-␻-methoxy-PEG; AHBPE, aliphatic hyperbranched-polyester; BP, benzophenone; BSA, bovine serum albumin; BTDA, 3,3 ,4,4 -benzophenone tetracarboxylic dianhydride; CB, Cibacron Blue F3GA; D, dialysis; DG, degree of grafting; DHB, dihydroxybenzophenone; DMFC, direct methanol fuel cell; EC, ethyl cellulose; ED, electrodialysis; EDMA, ethylene glycol dimethacrylate; EIPS, evaporation-induced phase separation; EO, ethyleneoxide; GMA, glycidyl methacrylate; GS, gas separation; HEMA, 2-hydroxyethyl methacrylate; HFP, hexafluoropropylene; LEDs, light-emitting diodes; MBAA, methylene bisacrylamide; MC, membrane contactor; MD, membrane distillation; MF, microfiltration; MIP, molecularly imprinted polymer; NF, nanofiltration; NIPAAm, N-isopropyl acrylamide; NIPS, non-solvent-induced phase separation; NVP, N-vinyl-2-pyrrolidone; PA, polyarylate; PAA, poly(acrylic acid); PAEK, poly(aryl ether ketone); PAN, polyacrylonitrile; PB, polybutadiene; PC, polycarbonate; PCEMA, poly(2-cinnamoylethyl methacrylate); PDMS, polydimetylsiloxanes; PDPO, poly(2,6-dimethyl-1,4-phenylene oxide); PE, polyethylene; PEC, poly(ethylenecarbonate); PEG, poly(ethylene glycol); PEGDA, poly(ethylene glycol) diacrylate; PEGMA, poly(ethylene glycol mathacrylate); PEO, poly(ethylene oxide); PEPP, poly[(4-ethylphenoxy)(phenoxy)phosphazene]; PES, poly(ether sulfone); PET, poly(ethylene terephthalate); PI, polyimide; PIs, polyisoprene; PMA, polymethacrylates; PP, polypropylene; PS, phase separation; PSf, polysulfone; PSt, polystyrene; PtBA, poly(tert-butyl acrylate); PtBMA, poly(tert-butyl methacrylate); PTFE, polytetrafluoroethylene; PV, pervaporation; PVA, poly(vinyl alcohol); PVC, poly(vinyl chloride); PVCin, poly(vinyl cinnamate); PVDF, poly(vinylidene fluoride); PVP, poly(2-vinyl pyridine); RO, reverse osmosis; SB, sodium benzoate; SPEEK, sulfonated poly(ether ether ketone); SPMMA, 1-[␤-(methacryloyl)ethyl]-3,3 -dimethyl-6-nitro-spiro(indoline-2,2 -[2H-1]benzopyran); TIPS, thermally induced phase separation; tBIA, tertiary-butyl isophthalic acid dichloride; THO, theophylline; TMBPA, tetramethyl bisphenol-A; TMPD, trimethyl-1,3-phenylenediamine; TPMLH, bis-[4(dimethylamino)phenyl] (4-vinyl-phenyl)methyl leucohydroxide; UF, ultrafiltration; VIPS, vapour-induced phase separation. ∗ Corresponding author. Fax: +49 201 183 3147. 1 2

E-mail address: [email protected] (M. Ulbricht). These authors have contributed equally to this paper. On leave from: Department of Chemical Engineering, Universitas Diponegoro, Indonesia.

0079-6700/$ – see front matter © 2008 Elsevier Ltd. All rights reserved. doi:10.1016/j.progpolymsci.2008.08.004

D. He et al. / Progress in Polymer Science 34 (2009) 62–98

63

Contents 1. 2.

3.

4.

5.

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Photo-irradiation and technologies based thereon . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.1. General characteristics of photo-reactions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.2. Photo-reactions of/with polymers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Polymer membranes—state-of-the-art, needs and options for further improvements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.1. Membranes with non-porous barrier . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.2. Membranes with porous barrier . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.3. Porous membrane adsorbers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.4. Stimuli-responsive membranes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.5. Special membranes for integrated processes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Improved or novel polymer membranes by photo-irradiation-based methods or technologies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.1. Membrane preparation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.1.1. Photo-initiated polymerization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.1.2. Photo-cross-linking . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.1.3. Photo-degradation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.2. Membrane modification (functionalization) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.2.1. Photo-cross-linking . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.2.2. Photo-functionalization (“grafting-to”) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.2.3. Heterogeneous photo-initiated graft copolymerization (“grafting-from”) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.3. Membrane stimulation (switching barrier properties) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Conclusion and outlook . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1. Introduction Photo-irradiation has a special role for life on earth because it serves as an important source of energy; photosynthesis is of crucial relevance for the global ecosystem. A more efficient use of solar energy is considered one of the key issues for sustainable development of mankind. Artificial sources of light have been used by men since very early days, and many more or less sophisticated versions are established nowadays. Light is used in information technologies for data transmission and storage. In many other technical applications, selective excitation with UV–vis irradiation is used to initiate chemical reactions which would not be possible via other pathways or which would be less efficient due to unwanted side reactions or effects. The synthesis, modification and controlled degradation of natural and synthetic polymers using photochemical methods are an important area. Membranes are also essential for life; biological membranes form and define individual compartments on the cellular and sub-cellular levels. Any membrane is an interphase between two adjacent phases acting as a selective barrier and regulating the exchange of substances between the two compartments. Synthetic membranes are nowadays established for a large variety of applications, and many more are targets of intense research and development. The main advantages of membrane technology as compared with other unit operations in (bio)chemical engineering are related to the unique separation principle, i.e., the transport selectivity of the membrane. Separations with membranes do not require additives, and they can be performed isothermally at low temperatures and – compared to other thermal separations – at low energy consumption. Also, upscaling and downscaling of membrane processes as well as their integration into other separations or reaction schemes are easy. Membrane technologies are considered as key technologies for process intensification. The majority

63 63 63 65 66 68 68 69 69 69 69 70 70 73 74 75 76 78 81 89 92 93

of technical membranes are made from polymers because a very wide variety of barrier structures can be realized with this group of materials. It is the aim of this review to give an overview on the recent developments in combining photo-irradiationbased and membrane technologies. Because the rapid growth of membrane technologies has started only in the 1980s, there are only occasional papers on that topic in the earlier literature (and if so, more in side areas of membrane technology such as materials for coating, packaging or sensing). In agreement with the scope of this journal, we will focus on the work in the last decade. However, to our knowledge, such a comprehensive review on the topic has not yet been prepared before. Therefore, we will also include important examples from the earlier period. Brief introductions are given into the fields of photo-irradiationbased technologies (with focus on polymers; Section 2) and of polymeric membranes (with focus on concepts, preparation methods and functions of membrane systems; Section 3). The main part of this review is organized according to the three different ways, photo-irradiation can be used to create novel or improved membrane systems based on polymers (as already indicated in the title of this paper; Section 4). The conclusions part will also involve a brief outlook towards possible new developments in this interesting area (Section 5). 2. Photo-irradiation and technologies based thereon 2.1. General characteristics of photo-reactions Photo-chemical reactions in the most general sense are reactions induced by ultraviolet (( = 100–400 nm), visible (( = 400–760 nm) and infrared (( = 780–20,000 nm) radiation [1]. In this review we will focus on photo-chemistry based on excitation with light of wavelengths in the range

64

D. He et al. / Progress in Polymer Science 34 (2009) 62–98

of ∼150 to ∼700 nm. Light energy is absorbed by molecules via electronic excitation. The efficiency and consequences of this process depend on chemical structure motives on the one hand, and the energy (proportional to frequency, , inverse proportional to wavelength, ) and the intensity of the light on the other hand. Because light absorption leads to molecules in an excited, i.e., more reactive state, various photo-initiated chemical reactions can follow. Their efficiency is determined by the extent of competing thermal or radiative (fluorescence or phosphorescence) deactivation processes. It is very important that many of these photoinduced reactions are only possible (or at least much more efficient) from the excited and not from the ground state of a molecule. This forms the basis for the pronounced chemo-selectivity of photo-reactions which is often used practically. The quantum yield, the ratio between amount of reactant consumed or product formed and the amount of photons absorbed (each photon has a specific energy, E = h), is the measure of efficiency of a photo-reaction (the maximum quantum yield for simple reactions is 1). With respect to the mechanism of photo-reactions, two different types of reactive entities can be distinguished, i.e., photo-sensitizers and photo-reactive molecules or moieties. Photo-sensitizers absorb (selectively) light and trigger subsequently a chemical conversion of another substance (i.e., they act as an “antenna” for the excitation light and transfer this energy via various mechanisms; most important are electron transfer and radiative transfer). Photo-reactive molecules (or functional groups in polymers) can after absorption of light undergo the following types of conversion (all involving cleavage of at least one chemical bond): • irreversible reaction (for example substitution, addition, elimination or fragmentation); • reversible reaction (typically isomerization). Also for photo-reactive molecules, subsequent “dark” reactions after photo-excitation can be very important. An important example is the photo-initiated polymerization; here, one absorbed photon can lead to consumption of even thousands of reactant molecules (monomer; i.e., the quantum yield with respect to monomer conversion can be much larger than 1). The light absorption by substances or materials is often described by the Lambert–Beer law. However, efficiency of light absorption can also depend very much on (local) concentration, matrix effects and other factors. More complicated are samples which are turbid so that light is also scattered. Nevertheless, for many technical applications, the penetration of excitation light into a solid or liquid material can be described by an exponential curve (and the decay of light intensity with increasing depth is either dominated by absorption or by light scattering). The current and emerging technical applications of UV–vis irradiation are all based on at least one of the following advantages: • high selectivity of chemical reactions or processes under mild conditions (ambient temperature or also much below),

• typically no need for added catalysts or special solvents, • spatially addressable effects (2D and 3D structuring possible), • applicable to very small and (relatively) large scales. The main technical applications of photo-irradiation with artificial light sources are the following: • chemical synthesis; • UV-curing of coatings (thin-film, i.e., “2D”, reactions); • photo-polymerizations of “3D” objects; for instance dental implants or rapid prototyping; • photo-lithography; • data transfer, recording and storage; • photo-catalysis; • disinfection, for instance water or waste water treatment (“advanced oxidation”); • “smart” UV-sensitive materials. A large variety of light sources is available, either for fundamental investigations in the laboratory or for small to large scale technical processes. Light sources can vary very much with respect to irradiated area and emitted intensities (also called: radiance [1]), and the energy range depends on the intrinsic principle. Conventional light bulbs are based on electrical heating of a filament (e.g., tungsten) and mainly used for illumination in the visible range at moderate intensities (up to few 100s of mW/cm2 ). Modulation or amplification of emission is possible by additives to the filling gas (e.g., halogens). Glow discharge lamps are based on a plasma state of an inert gas; the emission energy is typically “tuned” by the addition of metals (e.g., mercury). The configuration of those lamps depends on the application (from illumination to radiation-based curing); high-intensity lamps for the entire wavelength range with tube lengths of up to 2 m are available for UV-curing technologies (up to several 10s of W/cm2 ). Lasers are another special light source; different from the sources mentioned above, monochromatic irradiation is possible, and the highly focussed parallel light can reach very high intensities (up to several 100s of W/cm2 ). State-of-theart excimer or exciplex lasers are available with very high energies (i.e., excitation wavelengths around 150 nm) and find application in curing technologies or advanced photo-oxidation. Light-emitting diodes (LEDs), with tunable emission energy and relatively low intensities, are yet another emerging alternative as excitation source for photo-chemistry, and they have the advantage that large areas of flexible geometry (e.g., an internally illuminated solid catalyst for advanced photo-oxidation [2]) could be realized. Lasers and LEDs are the most important light sources for current and future applications of photo-irradiation in information technologies. With the very fast development of technologies requiring high-intensity large area light sources (UV-curing, advanced oxidation, photocatalysis), there are now many technical systems available on the market which can be also used for other applications.

D. He et al. / Progress in Polymer Science 34 (2009) 62–98

2.2. Photo-reactions of/with polymers The most important photo-reactions involving polymers, including their main applications and related (process) conditions will be briefly discussed, with particular emphasis on what can/could be used for preparation, modification or stimulation of polymeric membranes. Four different types of polymer reactions are distinguished here; photo-degradation, photo-functionalization and photo-polymerisations are based on irreversible reactions after photo-irradiation while the photo-isomerization is a reversible reaction. Photo-degradation. Polymer degradation in general is an undesired process, often leading to a reduction of the polymer molar mass and to a deterioration of material properties. Consequently, photo-degradation, e.g., the “aging” of plastic constructions or coatings, has intensively been investigated (for a review see [3]). Depending on the excitation energy, all moieties of a polymer can in principle absorb light and are hence converted to the excited state from which they can – with lower or higher probability – undergo chemical reactions. Most important reactions are oxidation, main chain cleavage or conversion, and loss or conversion of side groups. In polymer processing, additives are used to stabilize polymers against photo-degradation; their main functions are absorbing the UV light (and converting its energy to thermal energy) or chemically “trapping” reactive species such as radicals. However, the photo-degradation of polymers can also be controlled and used in a productive way, and photo-lithography is the most important example. In positively working polymer-based photo-resists based on UV-initiated depolymerization, the irradiated regions become more soluble or are even directly evaporated. The photo-degradation pathways of typical

65

membrane polymers such as cellulose, polypropylene (PP) or polyethersulfone (PES) (cf. Section 3) have also been studied in detail [4–6]. Mechanisms of photo-degradation either can be used or must be considered in all attempts to use UV technology in combination with polymer membranes. Photo-functionalization of polymers. Different from photo-degradation, the photo-functionalization of polymers is typically not accompanied by a reduction of polymer molar mass, and it is based on special photoreactive moieties (as side group or as part of the main chain of the macromolecule) having a distinct, selective and efficient reactivity. Depending on the chemistry of the photo-reactive group, controlled elimination or addition reactions are possible, also including dimerization reactions leading to cross-linking of macromolecule chains. Important examples are based on the photo-decomposition of aromatic azides or diazo compounds including their carbonyl “cousins” (azidocarbonyl or diazocarbonyl) under release of nitrogen, which can subsequently lead to addition (via nitrene/isocyanate or carbene/ketene) or to cross-linking reactions (via nitrene dimerization) (Fig. 1). Such reactions are being used for “photoaffinity-labelling” of bio-molecules, e.g., proteins [7]. Another important principle is the photo-induced dimerization via [2+2] cycloaddition, e.g., of cinnamate, coumarin or styrylpyridine groups (this reaction is only possible via the excited state). Such reactions are also extensively used in photolithography, the aforementioned addition reactions lead to positive resists while cross-linking reactions are the basis for negative resists. Precisely controlled changes of molecular structure (on surface or in volume) and cross-linking are possible via photo-functionalization of polymers, but these technologies require “tailored” macromolecular materials.

Fig. 1. Three groups of photo-reactive moieties and their main photo-initiated reaction pathways.

66

D. He et al. / Progress in Polymer Science 34 (2009) 62–98

Fig. 2. Photo-initiators and photo-sensitizers as well as photo-initiated reactions yielding starter radicals for a radical polymerization.

Photo-initiated polymerizations. A large variety of polymeric materials can be obtained by de novo synthesis from low molar mass monomers, by polymerization (“curing”) of reaction mixtures comprising macro-monomers (“prepolymers”), by cross-linking polymerization or by graft copolymerizations (i.e., the synthesis of macromolecular branches from a polymer as reactant). All these reactions can very efficiently be initiated using photo-initiators or/and photo-sensitizers (Fig. 2). There are several excellent reviews on various aspects of that topic [8–11]. With respect to number of applications and volume (mass) of technical products, this type of polymer photo-reaction is by far the largest. In particular, UV curable protective polymer coatings are important commercial products, and the related manufacturing/application technologies have also facilitated the development of high-performance equipment (including high-intensity UV sources allowing continuous curing at very high speeds). Other important applications are polymerization mixtures for dental applications or for low-resolution photo-lithography. Emerging processes include the rapid 3D prototyping using excitation with focused laser beams. Heterogeneous photo-initiated graft copolymerization can provide excellent control on polymer layer homogeneity, thickness and composite stability [12–14]. Overall, such technologies also provide large possibilities for emerging applications, especially for preparation of thin selective films, or other polymer composite barrier structures. Photo-isomerization of/in polymers. Reversible photoisomerizations (often called “photo-chromism”; either both processes are photo-reactions, or one is a photochemical and one is a thermal reaction) allow to “switch” between two states of a chromophoric moiety. Molecular changes during the isomerization reactions include group polarity, charge, colour (i.e., light absorbance) and size. Important examples for photo-chromic moieties are azobenzene [15], diarylethenes [16,17], or spiropyranes [18] (Fig. 3). In combination with polymer materials, such molecular changes can be converted into responses with respect to macroscopic property changes including changes in dimension of several 100s of nanometres. Very intense research has focussed on materials development for data storage. However, this principle has also a big, still only

partially explored potential for development of smart polymeric materials, including membrane surfaces or barriers. 3. Polymer membranes—state-of-the-art, needs and options for further improvements The barrier structure of membranes can be classified according to their porous character.3 For non-porous membranes, the interactions between permeand and membrane material dominate flux and selectivity; the transport mechanism can be described by the solution/diffusion model. For porous membranes, flux and selectivity are mainly influenced by viscous flow and sieving or size exclusion by the membrane pores. Membranes with ion-exchange groups in the barrier are also important; depending on the barrier porosity the interactions with the charged groups can completely control flux and selectivity (e.g., via Donnan exclusion or ionic binding). More specific chemical interactions with substances in the feed mixture are used in so-called affinity membranes. In order to apply such membranes in separation or integrated processes, different driving forces for trans-membrane transport can be used, the most important one is a pressure difference; a concentration difference, an electrical potential or a temperature difference across the membrane is also used frequently. An overview on the most relevant technical membrane processes with important characteristics is given in Table 1. The membrane cross-section can be isotropic (“symmetric”), integrally anisotropic (“asymmetric”), bi- or multilayer, thin-layer or mixed matrix composite. By far most of the technically used membranes (including support membranes for composite GS, RO, NF and PV membranes) are made from organic polymers and via phase separation (PS) methods. Technically most relevant are four variants for processing a film of a polymer solution into a porous or non-porous membrane with either isotropic or anisotropic cross-section:

3 The IUPAC terminology for pore diameter (d) is used throughout this paper: micropores with d < 2 nm, mesopores with 2 nm < d < 50 nm, macropores with d > 50 nm.

D. He et al. / Progress in Polymer Science 34 (2009) 62–98

67

Fig. 3. Three groups of photo-chromic moieties and their reversible photo-reactions.

• coagulation in a non-solvent (typically water)—“nonsolvent-induced”: NIPS; • solvent evaporation—“evaporation-induced”: EIPS; • absorption of non-solvent (water) from the vapor phase—“vapor-induced”: VIPS; • cooling—“thermally induced”: TIPS.

branes. All phase separation membranes with a porous skin, developed for D, UF and MF, have a pore size distribution in their barrier layer. Commercial MF membranes with more isotropic cross-section morphology are prepared via the TIPS process [21–23] and via the EIPS [24,25] or, in some cases, the VIPS process [26–28]. Other processes for the preparation of porous membranes include stretching [29] or track-etching [30,31] of polymer films. Various composite membranes prepared by interface polymerization reactions or coating processes – mainly on asymmetric support membranes – had been established for RO, GS, PV, NF [32–34] and also recently for low-fouling UF [35,36]. Mixed matrix membranes, with one phase mainly acting as transport phase and the other mainly

Today, extensive knowledge exists on how to “tailor” the membrane’s pore structure including its cross-section morphology by the selection of polymer solvents and nonsolvents, additives, residence times and other parameters during NIPS [19,20]. The key for high performance is the very thin “skin” layer which enables a high permeability. This skin layer is non-porous for GS, RO, PV and NF mem-

Table 1 Overview on important membrane separation processes with their characteristics. Membrane process

Membrane barrier

Two phases

Driving force

Separation mechanism

Example

Microfiltration (MF)

Porous (d > 100 nm)

l/l

p

Size exclusion (sieving)

Ultrafiltration (UF)

Porous (2 nm < d < 100 nm)

l/l

p

Size exclusion (sieving)

Nanofiltration (NF), aqueous NF, organic Reverse osmosis (RO) Gas separation (GS) Pervaporation (PV) Electrodialysis (ED)

Nonporous or microporous, charged Nonporous polymer Nonporous polymer Nonporous polymer Nonporous polymer Nonporous or porous

l/l

p

l/l l/l g/g l/g l/l

p p p c E

Dialysis (D)

Porous (2 nm < d < 100 nm)

l/l

c

Solution-diffusion Donnan exclusion Solution-diffusion Solution-diffusion Solution-diffusion Solution-diffusion Solution-diffusion, size exclusion or charge exclusion Size exclusion (sieving)

Membrane distillation (MD)

Porous

l/g

T (and p)

Vapour diffusion

Sterile filtration of aqueous solutions Concentration of biomacromolecule solutions Removal of bi-/multivalent ions from water Lube oil dewaxing Ultrapure water production Oxygen/nitrogen separation Absolute ethanol production Water desalination, amino acid production Hemodialysis (“artificial kidney”) Ultrapure water production

68

D. He et al. / Progress in Polymer Science 34 (2009) 62–98

acting as matrix phase are gaining increasing importance [37–39]. The development of a membrane polymer with a high affinity for a substance to be separated or of a membrane with a narrow pore size distribution can largely improve membrane performance. Improvements can also be achieved by membrane modification, e.g., the introduction of additional functionalities into the barrier [40]. The desired intrinsic properties of a polymeric membrane may also deteriorate under the actual process conditions (thermal, solvent or chemical degradation with time), and membrane modifications can to some extent provide counter measures. Polarization phenomena are a direct consequence of the transport through the membrane; the most important example is concentration polarization in the boundary layer between bulk phase and membrane, which will lead to a reduction of driving force. This increase of local concentration can lead to membrane fouling or scaling, i.e., the deposition of organic matter or salt, respectively, on the membrane, with the consequence of a further reduction of trans-membrane flux [41]. Membrane fouling can also occur by spontaneous adsorption or adhesion of solutes or particles to the membrane surface [42–44]. Many efforts have been devoted to the reduction of membrane fouling by membrane modification [40]. Success in that field will reduce the efforts for membrane cleaning and will contribute to an increased membrane lifetime. 3.1. Membranes with non-porous barrier In GS and PV (with polymer membranes), in RO and NF as well as in some electro-membrane processes (with nonporous polymers), the membrane polymer is crucial for flux and selectivity because the transport occurs through the polymer itself. A trade-off between permeability and selectivity is found in many cases, especially in GS. The nonporous barrier in technical membranes is either the skin of an integral membrane (made from one polymer) or a thin polymer film on top of the porous support (made from another polymer) in a composite membrane (Fig. 4(a)). The design criteria for the various separation processes differ depending on the separation conditions. For GS with glassy polymers, the free volume and its distribution are of critical importance for the relationship between permeability and selectivity [45,46]. Plastification by dissolved gases (e.g., carbon dioxide) can change structure and separation performance drastically. Rubbery polymers are typically quite hydrophobic and allow high fluxes for less polar substances; in order to maintain selectivity under process conditions, excessive swelling should also be prevented [47]. For PV, NF and RO, the swelling of the barrier layer in the liquid is more dominant, the magnitude of the effects is small for RO and larger for NF and PV. However, for applications of NF with organic instead of aqueous solutions, the requirements with respect to membrane stability and separation performance are not yet fulfilled with the available materials [48,49]. Ionic membranes play important roles in special fields; the most important materials are perfluorinated cationexchange membranes for chloro-alkali electrolysis or (still

under development) separator in low-temperature fuel cells [50–52]. The counter ions of the fixed ionic groups are preferentially (selectively) transported through the membrane. A trade-off is also observed for polymer electrolyte membranes in fuel cells: the high concentration of fixed charged groups which is required for high conductivity leads also to large swelling which compromises the barrier properties. Approaches to reduce this excessive swelling are cross-linking of the polyelectrolyte or filling the pores of a robust support membrane with the selective polymer (Fig. 4(c)). Main options for UV-based technologies are the formation of the thin barrier in a composite membrane (cf. Fig. 4(a)) and the controlled cross-linking of the barrier layer, mainly in order to reduce undesired swelling phenomena. In addition, the functionalization of the polymeric barrier to impart special selectivities (i.e., enhance sorption/solubility by affinity, including ionic binding) without change in the overall membrane morphology is also of significant interest. Modifications to reduce fouling and improve cleaning are relevant for all membrane processes with liquid (mainly aqueous) feeds of more complex composition. Scaling is a particular problem with RO and NF of aqueous feeds and special surface properties of the membranes could improve the cleaning. Because the undesired interactions occur in most cases on the outer membrane surface, surface-selective UV technologies to prepare thin coatings or grafted layers will have an especially large potential (the resulting membranes may also be considered composite membranes; cf. Fig. 4(a)). 3.2. Membranes with porous barrier In membranes with porous barrier structure (for D, UF, MF, MC), the function of the polymer is mainly to provide sufficient mechanical stability. Further membrane development is driven by the request for higher selectivity at high flux. This could be achieved by a more precise and narrow pore size distribution at high porosity. The increasing number of membrane applications in membrane-bioreactors, e.g., for wastewater treatment, and other integrated processes (cf. Section 3.5) is also a significant motivation [53,54]. The intrinsic selectivity of an UF or MF membrane is under process conditions often changed or even completely eliminated by membrane fouling. Because pore blocking is involved, the magnitude of such effects is typically much larger than with non-porous barriers (cf. Section 3.1). Hence, similar to RO and NF, anti-fouling functionalizations are very promising fields for the application of UV-based technologies. The aim is to prepare ultra-thin layers which do not more than necessary deteriorate the intrinsic flux and selectivity of the original membrane, i.e., under conditions where the subtle polymeric pore structure is not degraded. In addition – similar to trends in NF (cf. Section 3.1) – UF membranes which are stable in organic or other aggressive media are very attractive. One preferred strategy to reach this goal is also polymer cross-linking to improve membrane stability and retain separation performance.

D. He et al. / Progress in Polymer Science 34 (2009) 62–98

69

Fig. 4. Schematic cross-section depiction of three main composite membrane types: (a) thin-film, (b) pore surface-functionalized, (c) pore-filled.

3.3. Porous membrane adsorbers Separations with membrane adsorbers (membrane chromatography, solid-phase extraction) are a very attractive and rapidly growing application field for functional macroporous membranes [55–58]. The key advantages in comparison with conventional porous adsorbers (particles, typically having a diameter of ≥50 ␮m) result from the pore structure of the membrane which allows a directional (convective) flow through the majority of the pores. Thus, the characteristic distances (i.e., times) for pore diffusion are drastically reduced. The separation of substances is based on their reversible binding on the functionalized pore walls (Fig. 4(b)). Surface functionalizations of suited porous membranes, mostly MF membranes or macroporous filter media, via “grafting-to” [59] or via “grafting-from” [60] can be efficient approaches. A “tentacle” or “brush” structure of the functional layer can be used for a significant increase of the binding capacity in comparison with binding on the plain pore wall. Finally, the chemistry of the functional layer determines the selectivity of the separation (ion-exchange, partitioning or affinity). Photo-initiated grafting could be the approach of choice if conditions for a truly surface-selective functionalization of a porous base membrane with tailored functional macromolecular architectures can be established. 3.4. Stimuli-responsive membranes In recent years, there is a rapidly increasing interest in membranes with reversibly “switchable” properties [61–64]. Due to the practically unlimited possibilities to design functional macromolecular systems, stimuliresponsive polymers are the most important materials or building blocks for such “smart” membranes. Stimuli include the pH value, salt concentration, presence of specific substances, temperature, electrical field, etc. Important work on advanced stimuli-responsive polymeric hydrogels has been reported by the group of Peppas [65–67]. In principle, non-porous or porous stimuliresponsive membranes (or membranes with changing charge) can be (and have already been) envisioned. However, the ultimate effect of the stimuli-responsive change on membrane properties will depend primarily on the barrier itself. For non-porous membranes, a change in swelling of the barrier can lead to changed permeability and selectivity. Such changes can be triggered by stimuli-responsive groups or units in the bulk of the membrane material, and their modified structure causes changes in solvent uptake. For porous membranes, tailored grafted functional polymer layers on the pore walls can be used to reversibly change the

permeability and/or selectivity of the membrane; the most straightforward mechanism is the alteration of the effective pore diameter by changing the conformation of a grafted polymer via solution conditions (e.g., pH, salt or temperature) as “stimulus”.4 Stimuli-responsive binding or release to or from functional groups on the outer or pore surface of a membrane can have influence on the adsorption (fouling) properties, but could also be used to construct membrane adsorbers (cf. Section 3.3). Switching of materials properties with light is of particular interest because this stimulus can be addressed locally, very fast and with high selectivity. Therefore, the photostimulation of polymeric membranes is an emerging field with particular future relevance which is specifically covered in this review (see Section 4.3). However, it should be kept in mind that the focus will be on the key membrane property, i.e., photo-stimulated changes of membrane barrier properties. 3.5. Special membranes for integrated processes Membranes have a large number of already established and many more potential applications where they are a key component in a more complex technical system. Examples include sensor systems [68], (bio)catalytic reactors [69,70], biohybrid organs [71] or lab-on-chip systems [72]. Of course, in many cases “as received” polymeric membranes will “do the job”. However, there is also a great demand for membranes with tailored properties, and nonporous or porous membranes, prepared or modified using UV-steps (cf. Sections 3.1–3.3) which can be candidates for above mentioned applications. Another unique feature of photo-initiated processes is of special interest: by adopting photo-lithographic techniques for preparation or modification, miniaturized or patterned membrane systems can be envisioned. Ultimately, one can also imagine that photoresponsive membranes (cf. Section 3.4) could find their first practical applications in sensor or lab-on-chip systems. 4. Improved or novel polymer membranes by photo-irradiation-based methods or technologies In this part, improvements of existing membranes (e.g., enhancing flux and selectivity and/or chemical and mechanical stabilities, or minimizing fouling) and development of novel functional membranes are described. Photo-irradiation can be done during membrane prepa-

4 One prototype for such a membrane system can be based on the reversible swelling of a polymer grafted to the pore wall, so that a transition between surface attached (“open”) and pore-filling (“closed”) conformations is used (cf. Fig. 4(b) and (c)).

70

D. He et al. / Progress in Polymer Science 34 (2009) 62–98

ration; i.e., one or more precursors are used and the photo-chemical step is crucial for the formation of a membrane from such precursors (Section 4.1). Another alternative is the application of photo-irradiation for a postmodification (“functionalization”) of an already existing polymeric membrane (Section 4.2). Finally, applications of photo-irradiation to obtain switchable barrier properties are also described; here the membrane must contain photo-responsive groups undergoing a reversible reaction coupled to a significant change in chemical structure or volume (Section 4.3). 4.1. Membrane preparation During the preparation of a polymer membrane from “precursors” (monomers, macromonomers, polymers), photo-irradiation can be used for initiating a polymerization reaction (of monomers), for a cross-linking reaction (of polymers), or for controlled polymer degradation (to increase the free volume or to create pores). 4.1.1. Photo-initiated polymerization In situ polymerization can be used as an alternative way to established methods (cf. Section 3) to prepare new membranes with improved performance. A particular advantage is that membranes from cross-linked polymers can be prepared in one step. As it is well known in membrane technology, thin-film or interfacial polymerization on a membrane support leads to composite membranes, combining very low barrier thickness with sufficient mechanical stability (cf. Fig. 4(a)). In situ crosslinking polymerization is also known to be versatile to fix interesting and sensitive morphologies (for instance in sample preparation for transmission electron microscopy), and this concept can be adapted to immobilize selfassembled barrier structures. Therefore, this section will cover bulk and interfacial photo-initiated polymerizations and resulting composite membranes as well as examples for the fixation of self-assembled aggregates or templates. Bulk polymer as barrier. Hydrogel membranes for biomedical and biosensor applications are common examples for membrane preparation using photo-initiated polymerization technique [73]. A multi-step strategy to prepare membranes for enzyme immobilization was proposed by Arica et al. [74,75]. PolyHEMA membranes were prepared by UV-initiated polymerization of HEMA using azobisisobutyronitrile (AIBN) as “type I” photo-initiator (cf. Fig. 2). Cibacron Blue F3GA (CB) as an affinity dye was then covalently incorporated and subsequently complexed with Fe(III) ions. These membranes were used for immobilizing the enzyme catalase [75]. The poly(HEMA-CB-Fe(III)) membrane showed greater enzyme-loading capacity than the precursor poly(HEMA-CB) membrane. Immobilized catalase in both membranes showed higher Michaelis’ constants and a broader temperature range of enzyme activity when compared with the free enzyme. Preparation of a plasticizer-free membrane matrix for ion-selective sensor was done by free-radical photoinitiated polymerization of a mixture of dimethacryloxypropyl and monomethacryloxypropyl oligodimethylsiloxanes. Benzoin isopropyl ether (BIPE) was used as

photo-initiator of “type I” (cf. Fig. 2). Polar co-monomers, e.g., cyanomethyl methacrylate or trifluoroethyl methacrylate, were added to adjust the polarity of the matrix polymer [76], because it had been observed that the nonpolar character did not allow partition of ions to membrane to sufficient degree [77]. The K+ and Ca2+ selective sensors made from the resulting membranes exhibited analogous or even better electro-analytical properties and a distinctly longer lifetime, compared with the most widely used poly(vinyl chloride) (PVC)-based membranes containing a plasticizer. Another example is the preparation of a selfsupported mixed matrix membrane containing ion conductive polymer electrolyte, i.e., an immiscible blend of polymethacrylates having poly(ethylenecarbonateco-ethyleneoxide) side chains (PMA-PEC-EO) with poly(vinylidenefluoride-co-hexafluoropropylene) (PVDFHFP) [78]. Films of the macromonomer PMA-PEC-EO, containing lithium bis(trifluorosulfonyl)imide (LiTFSI) salt and “type I” UV initiator (2,2-dimethoxy-2-phenyl acetophenone) were cast, followed by irradiation with UV light to polymerize the methacrylate units. The addition of PVDF-HFP into this reaction mixture was found to greatly improve the mechanical stability as well as the elasticity without markedly influencing the level of the conductivity. In situ UV-initiated polymerization of the zwitterionic monomer 2-(N-3-sulfopropyl-N,N-dimethylammonium) ethyl methacrylate and the cross-linker methylene bisacrylamide (MBAA), using a focussed 355 nm laser beam, was used to prepare microdialysis membranes [79]. It was reported that by controlling the phase separation, occurring in the course of cross-linking polymerization, via the ratio between solvent (water) and non-solvent (2-methoxyethanol), the molecular weight cutoff of the membranes can be engineered for different applications. In further work, it was shown that those membranes could be used for electrophoretic concentration of proteins in a microchip [80]. Recently, photo-initiated cross-linking polymerization was also used to prepare stimuli responsive membranes for controlling of either drug release or drug permeation. Fig. 5 shows the schematic preparation and function of such “intelligent” membranes for drug release [81]. A mixture of a matrix monomer such as HEMA and a model drug such as methylene blue was cast into a mold and polymerized by UV irradiation. The membrane was then coated with a stimuli-responsive hydrogel obtained from N-isopropyl acrylamide (NIPAAm) or acrylic acid (AA), a cross-linker such as poly(ethylene glycol)dimethacrylate and UV curable cross-linkable prepolymer. Thereafter, samples were placed on a conveyer and irradiated with UV light. It was reported that those intelligent membrane systems showed temperature- or pH-responsive release functions. Various polymeric membranes and stimuli sensitive monomers as well as drug models were reported in other publications [82–85]. Stimuli-responsive polymeric hydrogels for controlled release of proteins [66,67], and molecularly imprinted hydrophilic polymers for recognition of cholesterole [86] or d-glucose [87] have also recently been prepared by in situ cross-linking photopolymerization.

D. He et al. / Progress in Polymer Science 34 (2009) 62–98

71

Fig. 5. Preparation of a membrane system for controlled drug release by photo-cross-linking polymerization and subsequent photo-curing of a coated stimuli-responsive polymer hydrogel (reprinted with permission from [81]; Copyright (2002) Elsevier).

UV irradiation was also used during in situ preparation of mixed matrix silica/poly(ethylene oxide) (PEO)immobilized electrolyte membranes [88]. The sol–gel synthesis of the silica particles and the UV polymerization of the macro-monomers were carried out in one step to enable the simultaneous formation of the polymer network and the inorganic nanoparticles. Robust membranes with improved transport rates for Li+ were achieved. Ultrathin polymer film as barrier. The potential of in situ photo-polymerization for preparation of ultrathin films is best illustrated by the giant free-standing polymer films from end-functionalized poly(2-methyloxazoline)-blockpoly(dimethylsiloxane)-block-poly(2-methyloxazoline) triblock copolymer having a thickness of 10 nm [89]. After the formation of the self-assembled structure by adding the triblock copolymer solution on a Teflon loop and smearing it across the pre-painted hole, methacrylate groups at both chain ends could be polymerized by UV light to further cross-link these films. The mechanical stability increased significantly after polymerization, and was found to be superior to the analogous lipid bilayer membranes, as indicted by a significantly higher voltage required for membrane rupture. Ultrathin cross-linked polyimide (PI) films for gas separation were prepared by combining UV irradiation and thermal treatment [90]. Membranes were obtained by UV cross-linking of Langmuir–Blodgett films of a salt of polyamic acid based on 4,4 hexafluoroisopropylidenebis(phthalic anhydride) and 2-(methacryloyloxy)ethyl-3,5-diaminobenzoate on the argon/water interface, subsequent deposition of the films onto the surface of a porous support and a final thermal imidization step. These PI layers on poly(phenylene

oxide) support membranes provided relatively good gas separation at high permeation rates. Thin-film polymer composite membranes. The first investigation using photo-irradiation for preparation of a composite membrane where the entire pore surface had been coated by a thin film (cf. Fig. 4(b)) can be traced back to Steuck’s work [91]. Then, similar work was reported for preparing hydrophilic MF membrane by photo-initiated cross-linking polymerization of a very thin layer of a hydrophilic copolymer to the hydrophobic membrane surface [92]. This approach will be described in more detail in Section 4.2.2. In pioneering work, Liu and Martin [93] prepared composite membranes with defect-free selective polymer films from various electroactive, photo-active and ion-exchange polymers, all were less than 50 nm thick (cf. Fig. 4(a)). Polymerization of monomer mixtures on the surface of porous support membranes (e.g., isoporous alumina, Anopore® , or track-etched polycarbonate, Nucleopore® ) has been initiated by UV excitation in a small angle relative to the membrane surface to ensure formation of thin films at minimum pore penetration. Gas permeability data showed that those composite membranes had really defect-free ultrathin barrier layers. Later, the same group [94] has reported the preparation of highly selective thin-film composite membranes by photo-initiated crosslinking polymerization of a molecularly imprinted polymer (MIP) film on an Anopore membrane. High transport selectivity for the MIP template, theophylline (THO), has been attributed to a facilitated transport through defect-free nanoporous separation layers. Recently, Don et al. [95] prepared bilayer composite membranes from a blend of chitosan and the copolymer of poly(acrylic acid) (PAA) and poly(ethylene glycol)

72

D. He et al. / Progress in Polymer Science 34 (2009) 62–98

diacrylate (PEGDA), by using a two-step photo-irradiation method. First, a film formed by casting of the solution of PEGDA and AA containing “type” photo-initiator (Darocur 1173) and a cross-linker monomer (dipentaethritol hexaacrylate) was subjected to UV irradiation. Second, a solution of AA and chitosan which also contained photoinitiator and cross-linker monomer was cast on the film formed in the first step. UV irradiation was applied to cure the top layer and form a bilayer composite membrane. Nonporous outer membrane surface as a selective barrier was obtained. Contact angle was increased by increasing chitosan concentration. By contrast, the top layer thickness decreased with addition of chitosan. The authors claimed that based on the mutual polymer diffusion and the UVcuring rates, the top layer thickness can be fine-tuned (in the range of 50 nm), while the membrane still demonstrated appropriate antibacterial activity due to enrichment of chitosan on the top surface. Composite membranes for catalysis processes prepared by immobilizing of semiconductor titanium dioxide and organometallic coordination compounds containing cobalt(III) and vanadium(V) via photo-cross-linking polymerization of acrylic monomers have been intensively investigated by the group of Bellobono (e.g., [96,97]). Solutions containing the monomer, the photo-initiator and the catalyst were used and photo-polymerized onto a non-woven polyester support. The rate of photo-crosslinking was enhanced by addition of the semiconductor titanium dioxide. These composite membranes showed strong photo-catalytic activity which could be used for degradation and mineralization of organic contaminants in wastewater. Fixation of supramolecular trans-membrane channels. Functional membranes containing ion selective matrixfixed “supramolecular channels” were developed based on the self-assembly of low molar mass amphiphiles (tris-methacrylated crown ether amphiphile or 2hydroxymethyl-[1,4,7,10,13-pentaoxacyclopentadecane]3,4,5-tris[4-(11-methacryloylundecyl-1-oxy)benzyloxy] benzoate) into long cylindrical aggregates [98,99]. The approaches were based on gelation of acrylate-based monomer mixtures by the string-like supramolecular assemblies of above mentioned gelator molecules. Thereafter, fixation of gels was done by in situ UV-initiated polymerization followed by removal of the gelator yielding pore channels. Because pure monomers, selected to ensure minimal shrinking upon polymerization, have been used, the supramolecular structures could be arrested permanently. The resulting membranes with a thickness of about 10 ␮m exhibited characteristics that were consistent with the concept of pores formed by supramolecular assembling and stacking of the crown ether units, because characteristic differences in ion-transport rates were observed (e.g., selectivities for Li+ and Na+ > K+ or NO3 − > Cl− > ClO4 − ). Applications of this “gel template leaching” method for preparation of other porous polymeric materials have also been reported. A macroporous poly(methylmethacrylate) was obtained by removing the gelator after UV-initiated polymerization [100]. An organogel to form a “imprinted” porous poly(divinylbenzene) film with submicrometer channels was obtained as follows [101]: Organogels com-

prising a 1:1 molar ratio of bis(2-ethylhexyl) sodium sulfosuccinate and 4-chlorophenol were prepared in divinylbenzene. AIBN was added as photo-initiator. The gel was cast to form a thin film. UV-initiated crosslinking polymerization of the solvent (divinylbenzene) was then performed by exposing to UV light for at least 2 h before removing the gelator by washing. In that way, it is also possible to prepare porous polymer membranes with other functional materials (e.g., enzymes or nanoparticles) embedded selectively in the pores. Very recently, the formation of helical pores in a polymer matrix by using self-assembled helices from 3,5bis(5-hexylcarbamoylpentyloxy)benzoic (BHPB) in organic solvent was reported by Simon et al. [102]. The helical tapes were formed in a mixture of ethylene glycol diacrylate and “type I” photo-initiator Irgacure 651. Therafter, the gel was UV polymerized, and removal of template was done by extraction using dichloromethane. Fixation of self-assembled templates for porous membrane. A novel strategy towards high-porosity macroporous membranes with uniform pore size distribution has been developed by Goedel et al. [103–106]. Thin free-standing membranes were prepared by using polymeric monolayer containing colloidal particle on the air–water interface [103]. That so-called “particle-assisted wetting” method was further developed to prepare porous membrane by spreading a mixture of monodisperse hydrophobized silica particles and a commercially available non-volatile acrylate-based cross-linker monomer (trimethylolpropane trimethacrylate) on a water surface [104]. Depending on the surface properties of the particles and the surface tension of the monomer (here both hydrophobic), the particles could assist the wetting of the monomer on the water subphase to form a uniform composite layer. UV irradiation was then used for photo-polymerization and after removing the particles by hydrofluoric acid, a thin porous membrane was obtained. The pore size could be tuned by choosing particles of different sizes. To increase the thickness of the membrane, a higher amount of particle-monomer mixture was used, and the colloids formed a three-dimensional template embedded in the cross-linker monomer. After UV-initiated polymerization and removal of the particles, a three-dimensional porous structure was obtained (Fig. 6). By applying repulsive superparamagnetic polystyrene (PSt) particles, this method was further developed to prepare porous membranes with controlled porosity [106]. More recently, cross-linked honeycomb-patterned films prepared from liquid oligomers having photo-crosslinkable groups have also been reported [107]. A mixture of oligomer and curing agent was cast and humid air was blown onto the cast film. UV irradiation was applied to fix the patterns formed by water droplets (which could be considered templates for pores) via cross-linking polymerization. By controlling the evaporation and UV-irradiation time, the pore size of the well-arranged honeycombstructures could be regulated. Moreover, the resulting porous polymer films were stable to many solvents. Through those works, novel macroporous membranes with high porosity and uniform pore size distribution were created. However, the membrane performance parameters

D. He et al. / Progress in Polymer Science 34 (2009) 62–98

73

Fig. 6. SEM images of membranes prepared via “particle-assisted wetting” with silica particles (diameter 334 nm) as templates: (a) before the removal of the particles (inset shows the bottom surface of the membrane); (b) after the removal of the silica particle. (c) Top and (d) bottom surfaces of the porous membrane; insets show the circular “windows” that connect the spherical pores pores (reprinted with permission from [105]; Copyright (2004) Wiley-VCH).

must still be explored in more detail to confirm the expectation that those membranes could be used for size-based separations at high fluxes. If that would be true, a plastic alternative to the photo-lithographically manufactured silicon nitride-based microsieves [108,109] could be envisioned. 4.1.2. Photo-cross-linking In order to obtain a membrane with specific property, photo-cross-linking of a soluble polymer can be done via photo-reactive side groups of the polymer or via addition of photo-sensitizer or photo-initiator into a polymer solution or gel. Membranes for enzyme immobilization are good examples. Immobilization by cross-linking yields carriers in which no covalent bonds to the enzyme are formed; the enzyme is physically entrapped within the bulk of a polymer or in the mesh of a swollen polymer network. Immobilization of glucose oxidase into poly(vinyl alcohol) (PVA) membrane in the presence of UV light sensitive diazoresin and sodium benzoate (SB) sensitizers was done by Liu et al. [110]. The process was further developed by preparing photo-cross-linked amphiphilic membranes made from poly(vinyl cinnamate) (PVCin) and PVA with the same photo-sensitizers (diazoresin and SB) [111]. Mixtures of PVA-diazoresin, PVA-SB and glucose oxidase in buffer were used for preparing a thin film on Teflon support. PVCin in methylethylketone was added to the upper

side of that film. After drying, the membrane was irradiated by UV light. By this method, the mass transfer resistance of oxygen could be decreased and the response of the enzyme electrode could be improved. The enzyme activity was clearly influenced by UV-wavelength and irradiation time, i.e., the cross-linking density within the polymeric barrier. An application of photo-irradiation for preparation of hydrogel membrane based on cross-linking via photo-dimerization of photo-sensitive cinnamate groups for biomedical application, e.g., for bioactive substanceincorporated antithrombogenic surface design, was also reported [112]. Techniques to prepare “tailored” photo-cross-linked membranes for enzyme immobilization were proposed by the research group of Tomaschewski [113–115]. In their method, photo-cross-linkable prepolymer (either based on PVA or on poly(styrene-co-maleic anhydride)), with varied content of reactive side groups (styrylpyridine derivatives which undergo efficient [2+2] cycloaddition) was firstly synthesized and then mixed with enzyme in buffer. The mixture was cast, dried and then exposed to UV light from both sides. Results of enzyme immobilization using the derivative of PVA with 1-methyl-4-[2-(4formylphenyl)ethenyl]pyridium methosulfate showed that the large enzyme invertase was effectively immobilized even in polymers with very low contents of the cross-linking component. By contrast for an effective immobilization of the smaller amylogucosidase, a higher

74

D. He et al. / Progress in Polymer Science 34 (2009) 62–98

Fig. 7. Schematic representation of photo-chemical [2+2] cycloaddition of polymer-bound photo-reactive styryl or cinnamic side groups, resulting in polymer network structures with mesh-size controlled by synthesis.

degree of cross-linking was necessary [114]. The results correlated also with the degree of swelling. Overall, it was confirmed that the network dimensions could be adjusted by synthesis, i.e., content of dimerizable groups. The concept is summarized in Fig. 7. Rühe and co-workers have reported the synthesis of photo-cross-linkable polymer hydrogels using benzophenone (BP) as photo-reactive side group of the polymer chain [116,117]. Photo-excitation of BP and subsequent hydrogen abstraction from adjacent polymer chains leads to simultaneous cross-linking of the polymeric hydrogel and surface-attachment of the polymer network to a polymeric or SAM-modified substrate (cf. Section 4.2.1). This reaction may have an even wider scope than the [2+2] cycloaddition, but barrier properties of such cross-linked materials have not yet been studied. Recently, photo-irradiation was integrated into phase separation technique (NIPS; cf. Section 3) for membrane formation [118]. Photo-cross-linking was performed by irradiation of a film of a solution of polyimide (PI) containing BP. It was reported that such cross-linking preserved the structure of cast solution film until immersing into a non-solvent bath. UV-irradiation time could be used for controlling the degree of cross-linking. The formation of undesired macrovoids could be suppressed with irradiation time and sponge-like integrally skinned asymmetric membranes were obtained with sufficient irradiation dose (see Fig. 8). More importantly, gas permselectivity increased significantly with the progress of cross-linking. Application of this technique for controlled preparation of porous polymeric membranes (for MF or UF) may be interesting, but tuning of the conditions may also be complex. Liu et al. [119] reported photo-cross-linking of a thin films with self-assembled nano-channels, prepared from a triblock copolymer, polyisoprene-block-poly(2-

cinnamoylethyl methacrylate)-block-poly(tert-butyl acrylate) (PIs-b-PCEMA-b-PtBA or ABC). The block copolymer was mixed with a homopolymer PtBA (homo C) and the solution was cast to make films. After drying and annealing, the PCEMA (block B) could be used for UVcross-linking of the AB phase. The membranes prepared by extracting HPtBA (homo C) with methylene chloride out of the film showed very regular pore morphology. However, such membranes were permeable for gas but not for liquid. A discontinuos structure of the nanochannels over the entire film thickness may be the reason. Besides the cases reported for in situ polymerization (cf. Section 4.1.1), this is another example, where photo-irradiation had successfully been used to fix a most interesting nanoscale morphology of a polymeric membrane. 4.1.3. Photo-degradation An application of photo-initiated pore formation in a polymer membrane via a controlled polymer degradation process was reported by Krausch and co-workers [120]. They developed self-assembled and oriented films from the triblock copolymer polystyrene-block-poly(2vinyl pyridine)-block-poly(tert-butyl methacrylate). Upon exposure of the film to UV light, the pores in the micro-phase-separated polymer film were formed by photo-degradation of the PtBMA phase. The resulting membranes showed very uniform pore morphology (Fig. 9). However, typical membrane parameters like permeability and selectivity need to be investigated. Analogous investigations are underway with films made from polybutadiene-block-poly(2-vinyl pyridine)-blockpoly(tert-butyl methacrylate); this polymer can also be photo-cross-linked via reaction of residual double bonds the polybutadiene block with radicals generated from a “type I” photo-initiator [121].

D. He et al. / Progress in Polymer Science 34 (2009) 62–98

75

Fig. 8. Cross-section images of membranes obtained from 20 wt% polyimide (Matrimid 5218) in 1,4-dioxane solutions with different irradiation time before phase separation by immersion in non-solvent (water): (a) non-irradiated, (b) 5 min, (c) 7 min (reprinted with permission from [118]; Copyright (2000) Elsevier).

4.2. Membrane modification (functionalization) Surface modification is a valuable tool to design appropriate membranes, as demanded interfacial characteristics

can rarely be achieved by bulk modifications of the membrane forming polymer without complications in membrane fabrication. The intention of surface modification of a membrane is either to minimize undesired

Fig. 9. Pore formation in block copolymer film by UV irradiation. (a) Chemical structure of polystyrene-block-poly(2-vinyl pyridine)-block-poly(tert-butyl methacrylate); (b) MesoDyn simulation of the first terrace of perforated lamellae morphology; the phases can be assigned to PS (white phase), P2VP (red phase), and PtBMA (blue phase); (c) tentative morphology after UV irradiation, i.e., photo-decomposition of PtBMA; (d) and (e) SFM images of a perforated lamella structure before and after exposure to UV light for 10 min (parts reprinted with permission from [120]; Copyright (2005) American Chemical Society). (For interpretation of the references to color in this figure legend, the reader is referred to the web version of the article.)

76

D. He et al. / Progress in Polymer Science 34 (2009) 62–98

interactions (adsorption or adhesion) which reduce the performance (membrane fouling), or to introduce additional interactions (charge, affinity, responsiveness or catalytic properties) for improving the selectivity or creating an entirely novel separation function. A key feature of a successful surface modification is a synergy between the useful properties of the base membrane and the changed chemical structure within the barrier or the novel functional polymer (layer) added to the barrier. This synergy can only be achieved by a mild and controllable modification technique, and photo-chemical processes have a large potential in that regard (cf. Section 2). Modifications in the bulk of the barrier should be distinguished from surface modifications. For surface selective processes, two alternative approaches are distinguished: (i) coupling small molecular entities or larger macromolecules to the surface (“grafting-to”) or (ii) heterogeneous graft copolymerization where monomers are polymerized using starter groups on the surface (“grafting-from”). For functionalizations of polymers with small molecules, it is often hard to precisely control whether the reaction will change only the surface or only the bulk properties. Therefore, these reactions are only distinguished from (bulk) crosslinking reactions of polymeric barriers (Section 4.2.1) and will be covered within one section (Section 4.2.2). 4.2.1. Photo-cross-linking Photo-cross-linking is of great interest not only for membrane preparation from low molar mass and/or soluble polymeric precursors (cf. Section 4.1.2), but also for the modification of already prepared polymer membranes to increase their selectivity or permeability, or to introduce other desired properties. In order to achieve that goal, cross-linking is applied to change the chemical structure and physical property of the polymeric membrane. For GS membranes, such treatment is performed to increase selectivity and/or permeability through optimization of free volume (gas diffusion is strongly influenced by the mobility of polymer chain), whereas for GS membranes to be used with easily condensable gases or PV membranes, such modification is mainly used to reduce the plasticization or swelling tendency. This section covers photo-crosslinking modifications of membranes from photo-reactive polymer and of polymer membranes with added photoreactive agent. For the latter case, we will distinguish between cross-linking via polymer radicals and via possible formation of interpenetrating polymer networks after polymerization of additives with monomer character. Photo-reactive polymers. The first work reported by Hayes [122] showed that a photo-cross-linked membrane from a PI containing 3,3 ,4,4 -benzophenone tetracarboxylic dianhydride (BTDA) exhibited very high permeation rates while still being able to effectively separate several combinations of gases. Since then, many studies have been devoted to the photo-cross-linking of such PI containing BP units. Kita et al. [123] reported that the gas permeability of the PI membrane decreased after irradiating with UV light. However, the permselectivity was higher than that of the uncross-linked PI with similar permeability. More importantly, the selectivity for the H2 /CH4 separation increased by factor of 50 after 30 min. irradia-

tion along with a decrease in H2 permeability by a factor of only 5. Other studies using this type of PI have been done by Liu et al. [124,125]. Investigations of a PI from BTDA and 2,4,6-trimethyl-1,3-phenylenediamine (TMPD) indicated that the degree of cross-linking was influenced by the UV-irradiation time. Increasing degree of cross-linking was correlated with increasing gas permselectivity and decreasing gas permeability. However, an optimization of the degree of cross-linking seemed possible. Liu et al. [125] have further studied the effect of cross-linking method on gas transport of the PI from BTDA and 2,3,5,6-tetramethyl1,4-diphenylenediamine (4MPDA), i.e., by putting the PI in ambient environment for 4 months (leading to bulk crosslinking), and under UV irradiation for 2 or 8 h (leading to surface layer cross-linking). For similar gas permeability, the UV-cross-linked PIs had a much higher gas permselectivity of hydrogen relative to nitrogen than those cross-linked under ambient environment. The mechanism of photo-cross-linking of BP-containing PIs was explained by Scaiano et al. [126] and Lin et al. [127]. Basically, irradiation by UV light yields biradical intermediates which are responsible for cross-linking reaction (cf. below in more detail). Meier and co-workers studied the reaction mechanism of various PI membranes containing both the phthalimide chromophore and abstractable hydrogens [128,129]. It was known that PI membranes underwent photo-chemically induced oxidative surface modification when they were irradiated with UV with wavelength of 200–300 nm for 0.5–30 min. The effects of atmosphere (environment), wavelength, irradiation time and intensity on the reaction had previously been studied [127]. Overall, the proposed reaction mechanism involved the excitation of the phthalimide chromophore with UV light followed by abstraction of a reactive hydrogen by the excited species to form a biradical intermediate. This intermediate rapidly reacted with oxygen to form peroxy radicals (ROO• ) which also abstracted hydrogen atoms to form peroxides (ROOH). Further UV irradiation could easily decompose the peroxides to alkoxy (RO• ) and hydroxyl (• OH) radicals which could subsequently react via similar reactions to form polar groups (–OH, –NH–, cyclic anhydride, etc.) on the surface of the PI film. The polar groups would interact with one another through hydrogen bonding and the interpolymer chain distance would be reduced yielding an increase in selectivity. A densification phenomenon as the reason for increasing selectivity after UV irradiation was also reported by the research group of Nakagawa [130–132]. However, a different driving force was proposed. The effect of UV irradiation on the gas permeabilities and selectivities of PI membranes having a BP structure (with BTDA) and a similar structure that was not photo-cross-linkable were investigated. For both membranes, the permeabilities of various gases decreased as the UV-irradiation time increased. However, for a gas mixture with difference in size (H2 and N2 ), increasing the permselectivity for the more permeable component (H2 ) without serious reduction of flux was observed. In addition, permeability and selectivity of UV-irradiated membranes significantly depended on their thickness, suggesting anisotropic changes over the membrane cross-section due to UV irradiation. More-

D. He et al. / Progress in Polymer Science 34 (2009) 62–98

over, physical changes (densification) due to UV irradiation affected the gas permeation properties more significantly compared with chemical changes (cross-linking) especially for the cross-linkable membrane [131,132]. The effect of UV irradiation with and without photo-initiator (e.g., BP) on gas transport was studied by using a very common gas separation membrane, made from the PI of 4,4(hexafluoroisopropylidene) diphthalic anhydride (6FDA) and TMPD. Comparison between UV and thermal treatment was also reported. The gas selectivities in all membranes increased due to irradiation. The increase in selectivity was more pronounced for membrane having no photo-initiator indicating that the physical densification was more pronounced than chemical cross-linking. However, addition of small amount of photo-initiator to the membrane gave a positive impact, i.e., the relationship between gas selectivity and permeability was easy to control. Photo-irradiation of membranes made from other polymers rather than PI has also been investigated. Gas sorption and transport were studied on UV-irradiated polyarylates (PA) prepared from tetramethyl bisphenol-A (TMBPA), 4,4 -dihydroxybenzophenone (DHB), and 5-tertbutyl isophthalic acid dichloride (tBIA) [133]. UV irradiation had little effect on the TMBPA/tBIA homopolymer but induced photo-cross-linking and photo-Fries rearrangements in the DHB-containing copolymers. The authors claimed that instead of enhancing cross-linking by the introduction of additional BP groups in the polymers, the dominating photo-Fries reaction led to non-reactive BP sites that cause a UV screening effect within the polymer. That screening effect limited the penetrating depth of UV irradiation into the polymer and thus the overall degree of cross-linking. The selectivity for gas pairs increased by UV irradiation; however, gas permeability decreased due to reduction in the diffusion coefficient. Compared to UV-irradiated PI membranes, those membranes showed smaller improvement in selectivity. It seemed that the extent of cross-linking was not high enough to cause a significant gain in selectivity. Therefore, McCaig and Paul [134] tried to increased performance of PA-based membranes by investigating the effect of photo-cross-linking on the gas permeability as well as physical aging of thin films made from two BP-based PAs. It was reported that the photo-cross-linking of thin films could greatly improve the long-term performance of membranes when compared to non-cross-linked films with the same thickness, i.e., a higher selectivity and a decreased rate of physical aging were achieved. Because of the high intensity of UV irradiation and very thin membrane, the cross-linking was expected to be uniform throughout the thickness. Photo-irradiation of poly(2,6-dimethyl-1,4-phenylene oxide) (PDPO) films was also investigated [135]. It had been known that exposing PDPO into UV light having wavelength between 250 and 460 nm could create free radicals. The cross-linking was limited to the upper surface of the films. The gas permeability was reduced upon cross-linking while the gas permselectivity was increased significantly. The addition of BP to that film did not markedly improve cross-linking and the resulting gas transport properties of the films.

77

Efforts to improve PV membrane performance using photo-irradiation have also been reported. Polyimide PV membranes made from BTDA and TMPD were photo-crosslinked by exposing onto UV light [136]. The results during PV of either benzene–cyclohexane or acetone–cyclohexane mixtures indicated significant decrease in the permeation flux along with a slight increase in selectivity. Hence, the authors claimed that photo-cross-linking was not effective to improve PV performance. Multi-block copolymers of sulfonated poly(aryl ether ketone) (PAEK) containing BP units and methyl moieties were synthesized and cross-linked by using UV irradiation for fuel cell application [137]. The resulting membranes showed lower methanol permeability without reduction of proton conductivity. The combination of a multi-block structure and a photo-cross-linking system seemed to be promising for PAEK-based proton-exchanges membrane for direct methanol fuel cell (DMFC) systems. Polymers with added photo-reactive agents. In PV membrane preparation, Wycisk et al. used UV light to crosslink poly[(4-ethylphenoxy)(phenoxy)phosphazene] (PEPP) mixed with BP [138]. It has been assumed that cross-links between polymer chains were formed via the recombination of macroradicals formed by hydrogen abstraction (cf. Refs. [116,117]). Irradiation under atmospheric condition led to some surface oxidation of the sample; however, this did not influence the cross-linking reaction. Increasing BP–polyphosphazene molar ratio from 0 to 0.5 increased the glass transition temperature of the polymer after cross-linking from −8.8 to +53.5 ◦ C, and decreased the equilibrium swelling in dimethylacetamide from infinity to 0.31. For PV of organic/water system, photo-cross-linking is used to reduce the swelling of the membrane by absorbed organic solutes, leading to a higher organic/water separation selectivity. In further work, they expanded on the solid-state UV photo-cross-linking of films composed of alkylphenoxy/phenoxy-substituted polyphosphazenes, where the alkyl chain was located at either the meta- or para-position on the phenoxy rings [139]. The glass transition temperature of all photo-crosslinked phosphazene membranes increased with increasing the BP concentration. The 3-methylphenoxy/phenoxy or 4-methylphenoxy/phenoxy phosphazenes were the best polymers for photo-cross-linking. UV-based cross-linking was also increasingly applied for the preparation polymer electrolyte membranes for fuel cell systems. Zhong et al. [140] prepared cross-linked sulfonated poly(ether ether ketone) (SPEEK) proton-exchange membranes for DMFC applications. The photo-chemical cross-linking was conducted by dissolving BP and triethylamine photo-initiator system in the membrane casting solution, and after solvent evaporation the polymer films were exposed to UV light. The membrane performance could be controlled by adjusting the photo-irradiation time. Cross-linked SPEEK membranes had significantly higher thermal stability and mechanical strength compared to non-cross-linked SPEEK. In addition, hydrolytic and oxidative stabilities were markedly improved whereas the water uptake and methanol diffusion were reduced with only slight sacrifice in proton conductivities. However, too long irradiation time could yield negative effects such

78

D. He et al. / Progress in Polymer Science 34 (2009) 62–98

as decreasing proton conductivity and reduced elongation at break. Interpenetrating polymer networks via UV-curing. The effects of different cross-linking techniques, i.e., irradiation (UV, ␥, electron beam) and thermal treatment, on the physical and transport properties of gas were studied by Dudley et al. [141]. They used a PI blended with an ethynyl-terminated monomer (N,N -dipropargyl(4,4 hexafluoroisopropylidene)bisphthalimide). UV irradiation caused changes only in the surface layer of the membrane and, as a consequence, improvement in separation selectivity with only modest reduction in gas flux was achieved. However, chemical and thermal stability were not significantly improved. By contrast, thermal cross-linking could increase chemical and thermal stability but a significant loss in permeability was observed. Irradiation using electron-beam seemed to cause polymer chain scission resulting in a reduced thermal stability and glass transition temperature. Photo-initiated cross-linking of PI copolymers containing both aromatic diamines and alkenylated diamides (having a vinyl or vinylaryl group) has been prepared and gas separation (O2 /N2 ) has been investigated [142,143]. In a preferred embodiment, the vinyl group was in ortho position to an amino functionality. Photo-cross-linking yielded a thin film of polymer at the membrane surface. The degree of cross-linking was influenced by concentration of the reactive vinyl groups in the polymer. It was discovered that although no BP-containing groups were present, the selectivity of several PI-based membranes increased after exposing to the UV light for 10–60 min. The reason was that the permeability of nitrogen gas could be decreased. Abdellah et al. [144] investigated the gas transport properties of photo-cross-linkable fluorinated PDMSs containing perfluorinated groups including C6 F13 or C8 F17 moieties and vinyl ether or acrylate groups linked to the main chain through an urethane bridge. For vinyl-PDMS, the cationic photo-initiator (Ph3 S+ , SbF6 − ) has been used while for acrylate-PDMS, the photo-initiators Darocur 1173 and isopropyl thioxanthone were added. Porous PAN was used as the support and cross-linking reaction was performed by exposing to a UV lamp (80 W/cm2 ) for 3–6 s. PDMS membranes with acrylate groups exhibited better ideal selectivity ˛(CO2 /CH4 ) and fluxes than vinyl ether substituted PDMS membranes. Very recently, Lou et al. [145] utilized BP as photo-initiator and a photo-polymerizable aliphatic hyperbranched-polyester (AHBPE) endcapped with acrylic groups for preparing a blend membrane with ethyl cellulose (EC). Ethylene glycol dimethacrylate (EDMA) was also used as cross-linker monomer for comparison. The EC membranes containing AHBPEs of higher generations exhibited much higher flux at comparable separation performance with EC membranes containing cross-linked EDMA. Overall, the EC-AHBPE blend membranes exhibited lower permeation activation energy and higher separation factor than that of EDMA-cross-linked EC membranes. Overall, photo-irradiation can cause cross-linking reaction and can be used for increasing performance of GS, PV and polymer electrolyte membranes. The photo-crosslinking reaction can occur to photo-reactive polymer,

polymer with added photo-reactive agents and polymer containing double bond. Fig. 10 describes the principle of photo-cross-linking reactions and the resulting network formation, using BP as example. The presence of photoreactive group either in the main or in the side chain will yield similar effect on the reaction as well as the network structure as long as a hydrogen donor is available in adjacent polymer chains. Photo-irradiation will excite the photo-reactive group which will thereafter abstract a hydrogen from a donor, yielding two radicals which can subsequently recombine and form a cross-link. As an alternative, added photo-initiator will lead to oligoor polymerization of polymer-bound double bonds. The effects of photo-irradiation on performance of GS and PV membranes have been very clearly observed in several studies with various polymers. However, the mechanism remains the subject of some debate in the literature; polymer cross-linking via covalent bonds or polymer densification via non-covalent interactions are the alternatives. Furthermore, it should be noted that too high degree of cross-linking can cause that the membrane becomes brittle, but also results in lower permeability. By contrast, too low degree of cross-linking may still lead to plasticization of the polymer over time resulting in deteriorating performance and loss of selectivity. Therefore, the degree of cross-linking should be optimized. 4.2.2. Photo-functionalization (“grafting-to”) Photo-functionalization of a polymeric membrane is based on the photo-reaction between membrane materials and functionalization agents, which requires special photo-reactive groups. According to the location of these photo-reactive moieties, two routes can be classified: via photo-reactive membrane polymer and via photo-reactive functionalization agents (Fig. 11). Since this approach is relatively independent on the chemical composition of materials (cf. Section 2), various functionalization agents can be attached onto a membrane surface with photoreactive moieties, or a variety of membranes (non-porous and porous, different membrane polymers) can be modified without any pre-treatment by using photo-reactive functional molecules. Photo-reactive membrane polymer. A photo-chemically reactive thin-film composite membrane was produced by the interfacial copolymerization of 3-diazo-4oxo-3,4-dihydro-1,6-naphthalene disulfonylchloride and naphthalene-1,3,6-trisulfonylchloride with 1,6hexanediamine on a PSf UF membrane [146,147]. The post-functionalization of the resulting polysulfonamide membrane surface was carried out via a photolysis of the polymer-bound diazo-carbonyl groups in the presence of various nucleophiles, yielding membranes containing various functionalities such as acid, ester, bromoethyl ester, dioxolan and hydroxyethyl ester (cf. Fig. 1). These changes of membrane polarity resulted in significant and predictable changes of the reverse osmosis separation performance. Darkow et al. [148] synthesized a new type of membrane polymer, poly(acrylonitrile-co-butadiene-co-styrene-co2-(4-ethenyl)phenyl-5-phenyl-2H-tetrazole), containing a photo-sensitive moiety, yielding a reactive nitrilimine

D. He et al. / Progress in Polymer Science 34 (2009) 62–98

79

Fig. 10. Photo-cross-linking reaction (with photo-reactive benzophenone as example) and resulting network structure: (a) photo-reactive polymer, (b) polymer with added photo-reactive agents, and (c) polymer containing double bonds and added photo-initiator system.

(1,3-dipole) after photolysis. Non-porous membranes with a thickness around 15 ␮m have been prepared and then functionalized using a variety of dipolarophiles and phenolic compounds to alter the polarity of the membrane surface. The photo-functionalization with more polar groups led to an increase in permselectivity towards benzene in a cyclohexane/benzene mixture in PV. Moreover,

clear differences in permeability between membranes irradiated from one side (anisotropic functionalization) and both sides (homogeneous functionalization) were found. The latter had a much lower flux at same selectivity what had been explained by cross-linking (via dimerization of nitrilimine side groups created by photolysis of tetrazole) in parallel to the functionalization.

80

D. He et al. / Progress in Polymer Science 34 (2009) 62–98

Fig. 11. Depiction of principal mechanisms for photo-functionalization of polymeric membranes: via photo-reactive membrane polymer (from left) and via photo-reactive functionalization agents (from right).

An alternative to the often complicated synthesis of special polymers and good membranes from such novel polymers is that reactive groups are immobilized onto an already prepared (and optimized) membrane by either chemical reaction or polymer coating. Nahar et al. [149,150] prepared a photo-reactive cellulose membrane by the coupling of 1-fluoro-2-nitro-4-azidobenzene to the hydroxyl groups of cellulose. They used the reactive membrane for UV- or sunlight-induced covalent immobilization of proteins, which is similar to a well-known technique for photo-affinity labeling of biomolecules [7]. It should be also possible to further functionalize these photo-reactive membranes for other applications. A new “photo-grafting-to” surface modification technique through two steps has been developed by Rajam and Ho [151]. The first step involved coating of the membrane (here a 0.22 ␮m MF membrane from a cellulose ester) with a monolayer of allyldimethylchlorosilane. Then a triblock copolymer of polyethylene oxide and polypropylene oxide (PEO–PPO–PEO) was covalently linked to this reactive surface membrane by UV irradiation at wavelength >215 nm. Reduced fouling and better cleaning during MF of protein solutions were observed. However, this approach has an obvious disadvantage, i.e., the short wavelength UV irradiation also degraded the base membrane; it had been reported that the resulting membranes became too fragile after 1.5 h UV irradiation. Photo-reactive functionalization agent. In comparison with the routes above, the use of “tailored” photo-reactive conjugates for surface functionalization can have significant advantages with respect to process simplicity and controllability. For example, using well-defined aryl azide

conjugates (cf. Fig. 1), a single and generic procedure (“photo-grafting-to”) should permit covalent attachment of a wide variety of chemical moieties, making it a very versatile process also from the practical point of view. PAN UF membranes were photo-chemically functionalized with low molar mass aromatic azide derivatives comprising different hydrophilic and hydrophobic substituents [152,153]. These functionalization agents were coated to the membrane surface and the degree of functionalization could be also controlled by UV-irradiation time. The separation characteristics and protein fouling tendency were markedly changed depending on the type of functional groups introduced. This could be well explained by changed hydrophilicity or charge of the pore surface in the UF barrier layer (note that the effective pore diameter for the UF membranes was in the range of 5–15 nm). In an extension of that work, well-defined photo-reactive ␣-4-azidobenzoyl-␻-methoxy-PEG (ABMPEG) was synthesized and photo-grafted onto the PSf UF membranes, applying wet conditions for exposing the pre-adsorbed ABMPEG to UV irradiation [154]. An optimum membrane performance had been achieved using monofunctional ABMPEG at relatively high concentrations during the photo-grafting procedure, yielding predominately end-on attached PEG chains on the membrane active layer surface. The hydrophilicity of modified membranes increased and the irreversible character of the grafting procedure was proved, indicating the covalent attachment of ABMPEG upon photo-initiation. The significant reduction of protein adsorption by photo-grafted PEGs contributes to both fundamental and practical solutions of the protein fouling problem [155]. Two different “photo-grafting-to”

D. He et al. / Progress in Polymer Science 34 (2009) 62–98

routes for the PEGylation of the important membrane polymer PSf by using arylazides have been compared recently [156]. Due to sterical hindrance between grafted PEG chains, the complete surface coverage of the base polymer was hard to achieve; however, such PEGylated surfaces with incomplete surface coverage showed a significantly improved biocompatibility [157,158], presumably due to the stabilization of the native conformation of adsorbed proteins [159]. Therefore, those PEG functionalizations may find application in membrane-based tissue-culture systems for sensitive adhesion-dependent cells. Park et al. [160] grafted temperature-responsive polymer conjugates, two types of azidophenyl-derivatized poly(NIPAAm)s (poly(NIPAAm)-Az; aryl azide group only on chain end, and poly(NIPAAm-co-AA/Az); aryl azide groups as side groups on the chain), photo-chemically in track-etched PC membranes (pore diameter 200 nm). They obtained different thermo-sensitive composite membranes, depending on the type and amount of modifier; “thin-layer” or “pore-filling” functionalized membranes showed opposite temperature responsiveness of barrier properties as judged from filtration and diffusion experiments with tryptophan as solute. A model has been presented for the interpretation of the observed phenomena (Fig. 12). Another thermo-responsive membrane was prepared by photo-grafting poly(N-vinylcaprolactam) (P(VCL-coAA/Az)) chains via their aryl azide conjugates to the surface of PET track-etched membranes (pore diameter 400 nm) [161]. The resulting composite membranes exhibited a temperature-responsive filtrate flux. Changed rejection for dextran showed that the materials had also temperaturedependent size selectivity for macromolecular solutes. However, the authors claimed that photo-grafting occurred only on the outer PET membrane surface instead of in the pores. This is consistent with the observations when the same membrane was modified using simultaneous “grafting-from” method [162] (see below). PEG “brushes” (thickness ∼20 nm) were photo-grafted via their azide conjugates onto the PSf UF membrane surface, by that means obtaining a membrane with PEGimmobilized silver salt as fixed carrier which showed facilitated transport for olefins. High propylene permeance and selectivity (12 for propylene over propane) were observed [163]. This is the only example, that a porous barrier had been changed into a non-porous one by a “grafting-to” method. Recently, amphiphilic graft copolymers, composed of a poly(cyclooctene) backbone with grafted PEG and photoreactive phenyl azide pendant groups, were synthesized and applied as coatings to commercially available PVDF UF membranes [164]. Photo-induced cross-linking of the graft copolymer coatings prevented delamination from the underlying membrane and lead to composite membranes with improved antifouling properties in the treatment of oil-in-water emulsions. In conclusion, the “photo-grafting-to” method has the potential advantage that the structure of the functionalization agent (e.g., a photo-reactive polymer conjugate) to be used for surface modification can be well controlled by synthesis and characterized in detail. However, the graft-

81

ing densities and homogeneity on the surface and hence the modification efficiency are limited due to the sterical hindrance effect which is increasing with the size of the functionalization agent. Moreover, photo-reactive polymer conjugates may react not only with the membrane surface but also with adjacent other molecules. This is even more pronounced if conjugates with more than one photo-reactive group are used (see [154,160]). In addition, accessibility of membrane pores for macromolecular functionalization agents may be limited. 4.2.3. Heterogeneous photo-initiated graft copolymerization (“grafting-from”) The stability and controllability of grafted polymer layers on the surface of a base polymer membrane are two crucial evaluation parameters for functionalization techniques. Especially the controllability has received increasing attention in terms of architecture and property of grafted polymers. Due to disadvantages of “photografting-to”, especially with respect to grafting density (cf. Section 4.2.2), “photo-grafting-from” has been increasingly used and developed. Various functionalizations of a large variety of membrane polymers via many different modification routes have been proposed, depending on initial membranes and desired functional membrane function. 4.2.3.1. Membrane functionalities achieved by “photografting-from”. In principle, all three types of composite membranes introduced in Section 3 (cf. Fig. 4) could be prepared via “grafting-from” techniques. However, thin-film composite membranes with non-porous barriers for GS or RO have been only rarely prepared by grafting (and no report on photo-grafting), because the very high grafting density required for using the intrinsic properties of the new polymer in a non-swollen state as selective barrier is hard to achieve. Many more examples can be found for more “loose” grafted layers which should still be permeable, typical examples are anti-fouling modifications, e.g., for UF membranes. Consequently, most composite membranes prepared via “photo-grafting-from” are pore surface-functionalized. Depending on initial pore size (micro/meso pores vs. macropores), grafted layer thickness (ultrathin or extending over 10s or 100s of nanometers) and its distribution over the membrane cross-section (outer surface vs. entire internal surface) largely different membrane functionalities can be achieved. This will be illustrated by representative examples for the most important cases before the further discussion will then focus on various different photo-grafting routes and chemistries. Grafted polymer as selective barrier. Via “photo-graftingfrom”, selective polymers, which swell significantly in water or organic solvents, can be mechanically stabilized by the fixation in the membrane pores (Fig. 13(a)). Especially for the function in organic solvents and/or in order to achieve a high selectivity for small molecules, excessive swelling can be prevented by small pores. For example, pore-filled PAN UF membranes (active layer pore diameter between 5 and 15 nm) have been established for the separation of organic mixtures in PV [165]. High selectivity and extraordinarily high permeate fluxes in PV of organic liquid

82

D. He et al. / Progress in Polymer Science 34 (2009) 62–98

Fig. 12. Thermo-responsive permeation of water or trytophan through polymer-immobilized membranes: (a) poly(NIPAAm) grafted membrane; (b) poly(NIPAAm-co-AA)-grafted membrane (low grafted mass); (c) poly(NIPAAm-co-AA)-grafted membrane (large grafted mass) (reprinted with permission from [160]; Copyright (1998) American Chemical Society).

mixtures had been achieved due to the prevention of excessive swelling of the selective polymer (by complete filling of the small pores) and the low effective PV barrier thickness (<1 ␮m, i.e., the skin layer of the base UF membrane). Grafted polymer as anti-fouling layer. Low-fouling UF membranes (Fig. 13(b)) can be prepared under photografting conditions where the degradation of the base membrane pore structure is minimized. The composition, surface coverage and thickness of the grafted layer are crucial for final membrane performance, i.e., low fouling at preserved size exclusion property and competitive flux [166,167]. Grafted polymer layer comprising functional groups for reversible binding. Surface functionalized MF membranes adsorbers for fast protein purification (Fig. 13(c) and (d)) have been prepared via photo-grafting of two- or threedimensional layers with suited functional groups [60,168]. A novel type of MIP composite membranes (Fig. 13(e)), with high binding specificity at high throughput, has been

obtained by surface initiated photo-grafting of a very thin cross-linked functional layer [169,170]. All these cases require a sufficient permeability, in order to use the main advantage of porous membrane adsorbers, i.e., the reduction of mass transfer limitations by directional convective flow though the membrane pores. Grafted polymer layer for immobilization of (bio)catalyst. Iso-porous track-etched membranes with a larger pore diameter (between 100 nm and 3 ␮m) as well as PP MF membranes had been functionalized via “graftingfrom” reactions in order to prepare enzyme-membranes as convective flow microreactors (Fig. 13(f)) [171,172]. For example, a continuous enzymatic reaction driven by trans-membrane substrate flow has been realized by using membranes with rather spacious cylindrical pores (3 ␮m) to avoid their blocking by the product polyfructan (inulin) with a molar mass between 30 and 50 million g/mol. The covalent immobilization of epoxy-reactive nanoparticles on the pore walls of PET membrane, photo-grafted with

D. He et al. / Progress in Polymer Science 34 (2009) 62–98

poly(amino-ethyl acrylate), also increased the productivity and lowered pore-blocking tendency [173]. Grafted stimuli-responsive polymer layer. Using tailored grafted functional polymer layers on the pore walls of membranes, it is possible to reversibly change the permeability and/or selectivity. The most straightforward mechanism is the alteration of the effective pore diameter by changing the conformation of a grafted polymer via solution conditions as “stimulus” (Fig. 13(g)). For example, reversible switching of permeability had been achieved using photo-grafted pH-responsive (PAA or PMA) [168,174,175] or temperature-responsive chains (polyNIPAAm) [176]. 4.2.3.2. Routes for “photo-grafting-from”. In order to really tailor and optimize the membrane performance, various routes have been developed depending on initial membrane materials and structures, architectures of grafted polymer, and used modification system including properties of monomer and solvent. It should be mentioned that radical polymerization has almost exclusively been used until now. Photo-grafting can proceed in two ways: without or with an added photo-initiator. The approach “without added photo-initiator” involves the direct generation of free radicals from the base membrane polymers under UV irradiation. Therefore, such methods require either a photosensitive base polymer (photo-reactive side group or part of polymer backbone) or the introduction of photo-sensitive groups onto the membrane surfaces prior to graft copolymerization. For approach “with added photo-initiator”,

83

initiating radical sites should be generated on the membrane surface by the reaction of photo-initiator (preferably type II; cf. Fig. 2) with the base membrane polymer under UV irradiation. In some cases, photo-sensitizers are added to accelerate graft polymerization. Without added photo-initiator. PSf and PES, due to their mechanical, thermal and chemical stability as well as excellent film forming properties, are frequently used as materials for high-performance UF or MF membranes. However, the hydrophobicity of the materials can cause problems, e.g., in applications with proteins where adsorption and deposition yield membrane fouling. Therefore, attachment of hydrophilic polymer chains to the membrane surface to significantly increase the wettability of membrane surfaces and hence reduce fouling is a promising strategy to extend their applications. It had been discovered that all poly(arylsulfone)s are intrinsically photo-sensitive and generate free radicals upon UV irradiation (cf. Section 2). Taking advantage of this knowledge, Crivello, Belfort and coworkers [177,178] have developed a novel method for surface modification of poly(arylsulfone) membranes. And different hydrophilic polymers have been successfully photo-grafted from vinyl monomers in water or methanol onto poly(arylsulfone) membranes with a high surface-selectivity. The investigation of the reaction mechanism has verified that the phenoxyphenyl sulfone groups as chief chromophores are responsible for the photoreactivity of poly(arylsulfone)s. As shown in Fig. 14 [179], the first step involves the absorption of light by the phenoxyphenyl sulfone chromophores in the backbone of the

Fig. 13. Depiction of representative membrane functionalities achieved by “photo-grafting-from” with various base membranes (UF, MF or track-etched membrane): (a) pore-filling selective polymer, (b) grafted anti-fouling layer, (c) ion-exchange membrane adsorber, (d) affinity membrane adsorber, (e), thin-layer MIP membrane adsorber, (f) (bio)catalytic membrane, (g) stimuli-responsive membrane (“valve”).

84

D. He et al. / Progress in Polymer Science 34 (2009) 62–98

Fig. 14. Proposed mechanism for the photo-chemical functionalization of poly(arylsulfone)s with vinyl monomers (“photo-grafting-from”; reprinted with permission from [179]; Copyright (2005) Springer).

polymer chain. The homolytic cleavage of a carbon–sulfur bond at the sulfone linkage takes place due to the photoexcitation. Generated aryl radicals and sulfonyl radicals are reactive enough to act as starter for a radical polymerization. Alternatively, the sulfonyl radical may lose sulfur dioxide to generate an additional aryl radical which may also initiate polymerization. Ulbricht et al. had also investigated the mechanism for photo-grafting of PES and PSf, including a comparison without vs. added photo-initiator [180]. To better understand the proposed mechanism and improve the application properties of poly(arylsulfone)based composite membranes prepared via this technique, intensive investigations have been carried out in the group of Belfort. Chemical characterization of the photo-grafted membranes showed that grafting proceeded not only on outer membrane surface but also to considerable depths into the membrane [178]. This was also supported by the results of flux measurements for UF membrane before and after UV irradiation in the absence of a monomer, indicating that the average pore diameter undergoes a marked increase [181]. This is mainly attributed to polymer degradation by UV irradiation. Owing to the high sensitivity to UV irradiation, especially for PES UF membranes, the functionalization efficiency sensitively depends on the UV wavelength and intensity. Pieracci et al. [182] found that the grafting efficiency of NVP onto PES membranes was significantly higher using high-energy 254 nm UV lamps than that using 300 nm UV lamps. The significant pore enlargement which leads to increased permeation flow rate and the loss of rejection for bovine serum albumin (BSA) was also attributed to the high-energy wavelength. These side effects have been reduced by selecting longer wavelength UV lamps or a UV filter solution for short wavelength UV

lamps [182]. In addition, Taniguchi et al. [183] found a linear correlation between DG and the product of monomer concentration and irradiation energy (c × E). It has been demonstrated that homopolymerization and main-chain scission without grafting, two undesired phenomena, can be minimized by operating below the critical irradiation energy, E < 4 kJ/m2 and with up to 10 wt% NVP monomer concentration (Fig. 15). However, the obtained DG for a given lamp wavelength (energy) and irradiation dose will also depend on both the sensitivity of the base membrane and intrinsic properties of the monomers employed. Much higher energy was required to achieve a desired DG when PSf membrane was used compared with PES membrane, due to the higher sensitivity of PES to UV irradiation [184]. Kaeselev et al. [184] observed also significant differences in the initial polymerization rate and the DG for several hydrophilic monomers. In addition, it was claimed that the monomers which may have the ability to swell and even dissolve the base polymer, can reduce the rejection of membranes. For example, N-vinyl-pyrrolidone (NVP) and HEMA are weak solvents for PES and might help to promote swelling and pore enlargement, which suggests that the selection of hydrophilic vinyl monomers that swell PES less than NVP and HEMA might help to maintain the pore structure after modification [185]. Immersion modification (membranes are irradiated while immersed in monomer solution), which was often employed earlier, requires a large amount of monomer and is less adaptable to the continuous processes on an industrial scale. Therefore, an alternative “dip method” (membranes are dipped in monomer solution, and then irradiated in nitrogen) has been also proposed. The grafting rate of the dip technique was found to be superior. For example, a high DG with NVP on a 50 kDa PES membrane

D. He et al. / Progress in Polymer Science 34 (2009) 62–98

85

Fig. 15. Proposed processes on PES membrane surface during “photo-grafting-from” with NVP for E < 4 kJ/m2 (reprinted with permission from [183]; Copyright (2003) American Chemical Society).

was realized in 30 s using the dip method, whereas 300 s were required to achieve a similar DG using the immersion method [186]. However, the dip modified membranes exhibited simultaneous loss of BSA rejection and permeability. This suggested that radiation cleaved PES bonds and enlarged the pores, and the grafted polymer chains on the surface blocked the pores and decreased the water flux; but the size-selective properties by the PES pore had been lost. The immersion-modified membranes retained their rejection because the NVP solution was found to absorb up to 88% of the emitted energy, depending on its concentration, thereby protecting the pore structure from intense UV degradation. Furthermore, using the chain transfer agent 2mercaptoethanol and ethanol cleaning, Pieracci et al. [187] prepared hydrophilic polyNVP-grafted membranes with permeability considerably higher than that of the unmodified PES membrane and of membranes dip-modified without chain transfer agent and cleaned with water. However, the observed rejection was severely reduced indicating that dip modification caused considerable pore enlargement though 300 nm lamps and a benzene filter were used. They interpreted that this difference with previous results is because non-grafted homopolymer may have been formed which effectively blocked the pores, but ethanol cleaning removed the homopolymer because it wets and possibly swells the membrane pore structure. The DG was lower when modified membranes were post-washed in ethanol, but the difference was only significant above a critical irradiation energy. As mentioned previously, energy doses below this value did not promote homopolymerization and, consequently, the post-wash technique did not have any effect on the DG. Anyway, the high hydrophilicity, high permeability, and low-fouling character of membranes modified with NVP using 2-mercaptoethanol and cleaning with ethanol, make them desirable for the filtration of larger proteins or for applications with smaller proteins in which high protein transmission is required. Recently, low-fouling PES-based UF membranes have been successfully established via simultaneous photo-graft polymerization of the hydrophilic monomer PEGMA onto

the membrane surface [166,188]. The grafted new thinlayer polymer hydrogel had significant influences on the flux and selectivity depending on its surface coverage, layer chain conformation, layer swelling and thickness. Further, the addition of a cross-linker monomer during modification may improve both permeate flux and solute rejection during UF [166]. The first example of a combination of two different photo-irradiation techniques is an optically reversible switching PES membrane surface, obtained by photograting a monomer with a photo-chromic side group [189] (see Section 4.3). Overall, very detailed investigations have been carried out including the grafting mechanism, dependence of grafting efficiency and properties of modified membranes on base membrane, monomer type, UV wavelength and intensity, post-washing as well as employed methods. Due to its potential, this technique is also being seriously evaluated by membrane manufacturers. An interesting example with respect to both, photo-irradiation and membrane technologies, had been made with the continuous photofunctionalization of the outer skin of PSf hollow-fiber membranes with an anionic grafted polymer layer to obtain NF membranes [190]. With other photo-sensitive polymers, where UV irradiation leads to formation of polymer-bound radical, addition of photo-initiator is also not required for “photo-graftingfrom”. Yanagishita et al. [191] have functionalized an UF membrane made of a PI with benzophenone structure BTDA-p-phenylene diamine (pPDA) via UV irradiation in presence of monomer in gas phase or liquid phase. The obtained pore-filled composite membrane showed benzene permselectivity for benzene/cyclohexane mixture in PV. Ping et al. [192,193] have prepared an UF membrane using cardo polyetherketone (PEK-C) containing BP structure in the backbone, and modification has been performed via direct UV irradiation. The success of modification was attributed to the photo-reduction of BP structure, which can generate radicals by abstracting hydrogen from monomer and/or bulk polymer itself (crosslinking occurred in the absence of monomer upon UV irradiation) (cf. Section 4.2.1).

86

D. He et al. / Progress in Polymer Science 34 (2009) 62–98

Fig. 16. Mechanism of “living” polymerization for surface functionalization of membranes via immobilized iniferter.

Another often used photo-sensitive moiety for polymer surface modification is the benzyl N,Ndiethyldithiocarbamate group. The possibility to use it for initiation of a controlled (“living”) radical polymerization has attracted much attention, and this property had also been termed “iniferter” (initiator transfer agent). The mechanism of surface modification is illustrated by Fig. 16. Wang et al. [194] have synthesized poly(acrylonitrileco-diethylaminodithiocarbamoylmethylstyrene), prepared UF membranes and successfully grafted a thin layer of a THO imprinted polymer onto the membrane surface. The resulting MIP composite membrane has been evaluated as a membrane adsorber, and it had been found that it could recognize the THO vs. the similar substance caffeine (CAF) with a high selectivity of ˛THO/CAF = 5.9. However, the synthesis of these special polymers containing photo-sensitive groups and the following formation of membrane might be a problem in many cases. Therefore, alternatively, a benzyl N,N-diethyldithiocarbamate group had been chemically immobilized on the surface of an established membrane via simple coupling reactions. Using immobilized photoactive iniferter, a molecule-responsive “gate” membrane had been prepared via surface functionalization of the skin layer pores of a commercial cellulosic dialysis membrane with a hydrophilic MIP [195]. The increase in DG value for re-initiation via multiple sequential UV-irradiation periods indicated that graft copolymerization may proceed via a controlled (“living”) mechanism. Such an approach should be helpful for the preparation of more sophisticated architectures, e.g., block structures, of the grafted layer. A very interesting function was achieved because the diffusion permeability of this membrane increased significantly when the template (THO) had been added while other similar molecules gave no or less effects. The mechanism of this gate effect on THO–MIP can be speculated from those results as following: specific binding of the solute (THO) to the MIP hydrogel in the nanometer-sized pores of the dialysis membrane induces partial shrinking of the grafted polymer, thus opening up the pores. This effect will not occur for all other solutes which do not bind to the imprinted binding sites in the MIP. Guan et al. have developed another photo-grafting route, based on the combined use of photo-oxidization and UV-irradiation grafting [196]. Hydroperoxide groups were

created on the membrane surface by photo-oxidation in hydrogen peroxide in the first step. Grafted copolymer layer on the membrane has then been obtained in the presence of monomer under UV irradiation in the second step. To minimize the homopolymerization, an appropriate amount of iron(II) was added in the monomer solution as a reductant. The grafted membranes have been applied for promotion of human endothelial cell adhesion and growth [197–199]. With added photo-initiator. Many surface modifications of porous membranes in laboratory and technical scale are performed using cross-linking polymerization to form thin layers covering the entire pore surface (cf. Fig. 4(b)). For such modification, conventional (“type I”) photo-initiators (for example, benzoin derivatives, organic peroxides or azo compounds; i.e., starter radials are directly formed by cleavage of a weak bond in the initiator; cf. Fig. 2) can be very efficient because very fast processes can be realized under ambient conditions. During the reaction, radical transfer to base polymer is possible (except for chemically very stable materials such as Teflon). Therefore, chemical grafting of the newly formed polymer layer to the support membrane may take place to some extent. The most important example for a commercial application is the hydrophilization of MF membranes made, for instance, from PVDF [91] or PP [92], using hydrophilic polyacrylates; this modification improves the wetting of the porous membranes by water and reduces the non-specific binding of (bio)macromolecules (i.e., membrane fouling). Another very special example is the synthesis of thinlayer MIP composite membranes based on MF membranes via a cross-linking polymerization initiated with help of the photo-initiator benzoin ethyl ether [200,201]. With respect to chemistry, there are strong similarities with the approaches described in Section 4.1.1, but the formed polymer layers are not important with regards to their barrier properties. In order to achieve better control over the heterogeneous “photo-grafting-from” of polymer surfaces, the “type II” photo-initiators, mostly BP or its derivatives, were frequently used. This type of photo-initiator undergoes photo-reduction by hydrogen atom abstraction from surrounding chemical species, which leads to the generation of initiating radicals (cf. Fig. 2). Therefore, a preferential hydrogen abstraction from the substrate polymer is an essential prerequisite for high surface-selectivity of graft

D. He et al. / Progress in Polymer Science 34 (2009) 62–98

87

Fig. 17. Schematic depiction of methods for the immobilization of a “type II” photo-initiator (here benzophenone) for photo-grafting functionalization of membrane surfaces (here a macro-porous membrane).

copolymerization; the homopolymerization in solution should be minimized. For this purpose, several strategies have been proposed (Fig. 17). As known for rather long time for the modification of polymer surfaces (not in the context of membranes), BP and its derivatives can be used to graft various functional polymers by just dissolving the initiator in the monomer solution [202–204]. Recently, Yang et al. [162,205,206] modified PET Nucleopore membranes using the same approach, i.e., the simultaneous method. The PET membrane was immersed in a monomer solution containing BP and modified by UV irradiation. They observed that the photo-grafting occurred only on the top surface rather than in the membrane pores. Membranes with thermo- or pH-sensitive permeability have been obtained by photo-grafting NIPAAm or 4-vinylpyridine (4VP), respectively. Via a similar method, Liu and coworkers [207] prepared a thermo-sensitive Nylon MF membrane, by a rapid bulk photo-grafting polymerization of N,N-diethylacrylamide (the photo-initiator BP was dissolved in the liquid monomer, used without any solvent). The grafted polymer was observed on the top surface and in the pores, but not on the backside (remote from UV). Photo-grafting/pore-filling with poly(PEG acrylate) has also been realized on/in PAN UF membranes via simultaneous method [208]. The obtained thin and defect-free barrier layer contributes to the high membrane performance for CO2 /N2 separation. Though this simultaneous approach is a quite facile technique, there are significant disadvantages. First, usually, the hydrogen atoms in many commercial membrane polymers are not very reactive to the excited BP. In that case, the selection of solvent should be considered very carefully, in particular only solvents without labile hydrogen (e.g., water or acetonitrile) should be used to minimize

homopolymerization and enhance the surface-selectivity. Second, the local concentration of BP on the membrane surface is quite low because BP moves to the membrane surface only by diffusion, whereas high bulk BP concentration may give rise to side reaction such as homopolymerization. These two factors lead to a low grafting efficiency and/or not well-defined polymer architecture (in the examples mentioned above, the contribution of branched grafted copolymer, cross-linked polymer or entangled homopolymer to the final membrane properties can be significant). Moreover, the use of monomers which do not have common solvent with BP (for instance, BP is almost insoluble in water) is limited. In order to improve the grafting efficiency, the photoinitiator adsorption method has been proposed [167,174] (cf. Fig. 17). The local BP concentration was increased via pre-coating of BP on the membrane surface, and the BP concentration in the bulk of monomer solution was kept very low or close to zero, which minimized the homopolymerization. Those conditions can be realized very well by using aqueous monomer solutions because the solubility of BP is low. This method has been extensively applied to various types of membrane polymers, PAN [209], PP [60,170], Nylon [174], PSf [180,210], PET [171], PVDF [169] and the novel photo-functionalized (composite) membranes have found applications in MF, UF, NF [165,211,212], as membrane adsorber [60] or as enzyme-membrane [171]. Main reasons are the flexibility with respect to selection of support membranes (only hydrogen donor properties are required) and graft polymer functionality and structure. However, the conditions need to be optimized for each membrane material, pore structure, graft functionality and layer design to meet the needs of different applications; this will be discussed with the help of examples below.

88

D. He et al. / Progress in Polymer Science 34 (2009) 62–98

Pore-filled PAN UF membrane has been successfully established by grafted functional polymer chains attached to surface when two necessary conditions have been met: DG ≥ DGcritical and small pores in the skin layer of the support membrane (≤12 nm in diameter) [165]. An extension of that work for the separation of aromatic/aliphatic mixture has also been reported [211–214]. The performance can be adjusted by controlling the hydrophilicity/hydrophobicity balance and pore size of the base membrane. Interestingly, the grafted polymer from PEGMA seems to be best suited for a rather wide range of separations of polar from non-polar organic substances, because among various different grafted polymers, the highest selectivity has been achieved with grafted PEGMA for the removal of alcohols but also of aromatics. PAN UF membranes have been grafted with monomethoxy poly(ethylene glycol) methacrylates (MePEG200MA) [215]. A hydrophilic and low-proteinadsorbing UF membrane with relatively high permeability has been established by adjusting the degree of grafting via UV-irradiation time and monomer concentration. By applying longer UV wavelength and the photoinitiator BP, PSf UF membranes have also been modified in a controlled way with PAA layer for covalent immobilization of biomolecules [180]. This technique avoids the negative effect of using direct UV irradiation that the desired functionalization was usually accompanied by strong pore etching in the UF membrane active layer (see above). Borcherding et al. [60] reported that a macroporous PP membrane functionalized with the reactive poly(glycidyl methacrylate) (polyGMA) preserved its high permeability and exhibited high receptor coupling capacity for recombinant protein A, due to the thin grafted layer and a great amount of epoxy groups introduced on membrane surface. Based on the very promising results, it is possible to tailor both membrane structure and application protocols towards other attractive affinity separations of biomolecules. Due to the high surface-selectivity of this method, a thin and compact cross-linked layer can also be synthesized onto the whole surface of a porous membrane, which maintains a high permeability and relatively large specific surface area, e.g., for affinity binding. Thin-layer MIP composite membranes, covered with an imprinted polymer layer selective to small molecules (shown for herbicides) in a mixture with similar compounds, have been prepared by Piletsky et al. [170] and Sergeyeva et al. [169]. The high affinity of these synthetic affinity membranes to templates together with their straightforward and inexpensive preparation provides a good basis for the development of applications of imprinted polymers in fast separation processes such as solid-phase extraction [216,217]. This photo-grafting method had been successfully adopted also by other groups [218,219]. The even, tight and defect-free filling of large membrane pores (from PP or PET) with porous polyacrylate-based monoliths was based on a pre-functionalization of the pore wall with a compatible photo-grafted copolymer [220]. MIP-pore-filled membranes of such type are currently evaluated for continuous enantio-selective separations [216]. Enzyme-membrane reactors have also been prepared based on modified PP MF membranes with polyGMA

[172]. An enzymatic chain elongation of maltooligosaccharide molecules was achieved during passage through a membrane activated with immobilized amylosucrase. Commercially available capillary pore membranes (PET) with diameters of 0.4, 1.0 and 3.0 ␮m have also been modified for covalent enzyme immobilization within the pores using this approach [171]. Enzyme-membranes with high permeability and essentially unchanged FTF activities have been obtained with grafted poly(2-aminoethyl methacrylate) layer followed by glutardialdehyde activation and coupling of fructosyltransferase (FTF, inulinsucrase from Streptococcus mutans). However, to preserve the tenet of this technique and to achieve the high surface-selectivity, the parameters which could reduce the local BP concentration (on the surface) should be taken into account. For example, a good solvent for BP should not be chosen for high grafting efficiency, though in one special example, a MIP composite PVDF membrane has been successfully prepared from the methanol solution containing a small amount of BP for preventing the rapid reduction of high BP concentration on the surface [170]. To further improve the grafting efficiency and controllability and to extend the application of “photo-grafting from” technique, more efforts have been made in our group, mainly focusing on the immobilization of BP. One improvement is that the weak physical adsorption of BP onto the membrane surface was replaced by ionic interaction between respective functional groups on the surface and the photo-initiator [175], which can be realized by the introduction of charged groups onto the base membrane and the application of counter charged BP derivatives. This stronger immobilization of photo-initiator enabled an efficient and surface selective functionalization because a better control of grafting density and a reduction of photo-initiated side reactions along with a more efficient use of the photo-initiator were possible. Thermo- and pHresponsive PET membranes have been prepared via this method [175,176]. Various measurements (for example, the trans-membrane zeta potential) indicated the even surface coverage of the pore walls with the grafted polymer. This was the basis for a quantitative analysis of effective layer thickness as function of synthesis conditions and solution conditions. Another improved strategy – “photo-initiator entrapping” – also has been proposed to strengthen the immobilization of BP in the membrane surface [168]. In this process, BP entrapping was realized by the procedures as follows: membrane was soaked in the BP solution whose solvent can swell base membrane polymer. After drying, slight washing with non-swelling solvent of membrane polymer followed for removing the adsorbed BP on the surface. The entrapped BP density in the membrane surface could be tuned by the initial BP concentration. Controllable three-dimensional grafted layer structure has been achieved via BP entrapping method to lead to an improved protein binding capacity. Yang et al. modified PP MF membranes via this technique to improve the biocompatibility [221–223]. For example, a novel sugarcontaining monomer (d-gluconamidoethyl methacrylate) was grafted on PP via UV-induced graft copolymeriza-

D. He et al. / Progress in Polymer Science 34 (2009) 62–98

tion [221]. Results with respect to BSA adsorption and platelet adhesion imply that a considerable enhancement of biocompatibility had been achieved. Another case is a glycopolymer-tethered porous surface by photo-graft polymerization of the sugar monomer d-allyl glucoside [222]; based on the recognition properties of the sugar units, such surface-glycosylated membranes are very promising for highly selective protein separation and drug delivery systems. Using the same conditions, PP hollow fiber microporous membranes have been grafted with a hydrophilic layer to improve the anti-fouling characteristics in a submerged membrane-bioreactor [224]. This method had also been used to prepare cross-linked grafted layers, and it had been demonstrated that the dynamic performance of porous membrane adsorbers could be improved [225]; a very high surface-selectivity is mandatory for such reaction conditions, because otherwise cross-linked insoluble polymer would be immobilized in an uncontrolled way in the pore space. In addition, the grafting efficiency of PP MF membrane has been improved by addition of ferric chloride with optimum concentration, when simultaneous method or adsorption method was adopted [226]. This is ascribed to the “synergistic effect” between Fe3+ and BP. Ma et al. [227,228] proposed and investigated another photo-induced variant of this graft copolymerization method – “covalent photo-initiator immobilization” – consisting of two steps. This method is based on the discovery of Yang and Ranby that also with BP, a more controlled (“living”) grafting mechanism can be realized [203]. In the first step, in the absence of monomer, BP abstracts hydrogen from the substrate to generate surface radicals and semipinacol radicals, which combine to form surface-bound photo-initiators. In the second step, the monomer solutions are added onto the active substrate and a living graft copolymerization is initiated under UV irradiation. In this method, graft density and graft polymer chain length can be controlled by choosing the reaction conditions in the first step and in the subsequent step independently. Moreover, this technique substantially eliminates formation of undesired homopolymer and cross-linked or branched polymer. This approach has been employed to modify membranes for fouling reduction [229,230]. However, in this process, the cleavage of covalent bond requires high energy, which leads to a low grafting efficiency and might also cause photodegradation of many graft copolymers. Combining adsorption and covalent photo-initiator immobilization methods, poly(N,N-dimethylaminoethyl methacrylate) had been photo-grafted onto the PP MF membrane [231]. Based on this reactive polymer layer, phospholipids-analogous polymers have been tethered on the membrane to improve surface biocompatibility in the next reaction. Taking advantage of the characteristics of BP that tertiary amino group as its co-initiator can accelerate the photo-reduction of BP, we have developed a highly surface-selective and structure-controllable photo-grafting method—the “synergist immobilization method” [232]. Efficient grafting of membranes has been realized under optimized conditions by two steps: covalent introduction of tertiary amino groups onto membrane surface, then

89

simultaneous method followed for graft copolymerization. The structure of grafted layer can be controlled by adjustment of main functionalization parameters such as UV-irradiation time, BP and monomer concentration. This technique has been applied to MF PP and PET track-etched membranes with a high surface-selectivity and controllability of photo-grafting under optimized conditions [233]. In addition, a high grafting efficiency and a threedimensional grafted layer can be obtained from water or organic solvent. Porous anion-exchange membranes with high protein binding capacity have been prepared by attachment of tentacle functional polymer [234]. Due to high surface-selectivity of photo-grafting, a cross-linked thin layer can also be prepared onto the membrane surface for example for synthesis of MIP composite membranes. In brief, many routes for “photo-grafting-from” technique have been developed for polymeric membrane modification via heterogeneous graft copolymerization. Various applications of obtained composite membrane have been found in many fields depending on membrane material and structure, properties and architecture of grafted layer. In contrast to “photo-grafting-to”, the synthesis of surface anchored polymers via “photo-grafting-from” is often less controlled with respect to polymer structure, but a very wide variation of grafting densities and chain lengths can be obtained under relatively convenient reaction conditions. 4.3. Membrane stimulation (switching barrier properties) Considerable attention has been paid to controlling mass transfer through a membrane by stimuli such as pH, ion concentration or temperature (cf. Section 3.4). Photo-irradiation is of especially great interest because this excitation can be imparted in a highly selective way “from outside”, without additional changes of the system’s chemical potential (cf. Section 2). Polymer conjugates with azobenzene, spirobenzopyran and diarylethenes (cf. Fig. 3) are among the most investigated photo-functional materials. Reversible changes of polymer polarity or conformation by photo-irradiation are caused by photo-isomerization of a photo-chromic unit within the macromolecule. Fig. 18 describes these changes and their consequences for the mass transport through porous and non-porous barriers. Photo-responsive carrier membranes. First experimental studies using liquid membranes were done in 1980s by Shinkai and co-workers [235,236]. They developed crown ethers funtionalized by azobenzene, where the selective complexation could be controlled by light irradiation. When immobilized in liquid membranes, the transport of certain ions could be switched by selective UV irradiation. Shimidzu and Yoshikawa have extended that concept to a photo-switchable ion carrier based on indoline2,2 -2H-benzopyran [237]. However, the photo-switchable ion-selective liquid carrier membranes share with their more conventional counter parts the problems of limited long-term stability and relatively low mass transfer [40] (cf. Fig. 18(a)). Photo-switchable carriers in polymeric membranes have been further investigated by Kimura et al. [238–241].

90

D. He et al. / Progress in Polymer Science 34 (2009) 62–98

Fig. 18. Reversible photo-stimulated changes of barrier functions for membranes with photo-chromics units: (a) non-porous carrier membrane (switching between non-binding and binding state), (b) non-porous polymer hydrogel (switching mesh-size), (c) micro- or mesoporous polymeric membrane (switching wettability), (d) macroporous polymeric membrane (switching grafted polymer chain conformation).

Photo-chromic behavior of crown ether–spiropyran copolymers as well as the distinct rheology changes have been investigated [238]. Synthesis of vinyl polymers incorporating crown ether–spirobenzopyran moieties at the side chain and their alkali metal-ion complexation as well as photo-responsive ion conducting behavior were then studied [239]. The group has also investigated the effect of photo-irradiation on metal ion transport by using poly(vinyl chloride) membranes based on a Malachite Green derivative carrying a bis(monoaza-15-crown-5) moiety [240]. In addition, photo-responsive ion sensors based on liquid crystalline membranes were prepared [241]. Photo-responsive polymer hydrogel as barrier. Ishihara et al. [242] have demonstrated that the swelling degree of a membrane made from polyHEMA with azoaromatic side groups in water was decreased by UV radiation and could be recovered by irradiating using visible light. Further, they described photo-induced permeation control of proteins with different molar masses through such functional hydrogel membranes [243]. Under UV irradiation, the permeation coefficients decreased in all ranges of protein molar mass. That change was more pronounced with increasing protein molar mass, and the permeation of proteins with molar mass more than 104 g/mol could not be detected. Consequently, that polymer membrane was permeable in the dark and semi-permeable (only for low molar mass compounds) under UV irradiation (cf. Fig. 18(b)). Sata et al. [244] prepared anion-exchange membranes containing viologen groups via heterogeneous polymer-analogous reaction of membranes from a copolymer of chloromethylstyrene and divinylbenzene with 4,4 -bipyridine. The viologen moieties were reduced by photo-irradiation to form radical monocations and biradicals in aqueous solutions. As consequence of the change from a charged to an uncharged group upon photoirradiation, the electrodialysis transport rates for sulfate, bromide and nitrate ions decreased relative to that for chloride ions. That was due to a sieving effect because the shrinkage of the polymer network resulted in smaller pore size caused by decreasing charge density of the anion-exchange membranes: Membrane pore size should change simultaneously with the change in dipole moment of photo-functional group. After oxidation of the reduced viologen group with air, the transport

rates of anions were similar to the state before photoirradiation, i.e., reversibility of permeability change could be demonstrated. The group continued their work to ion permeable membranes having incorporated azobenzene groups; the chloromethylstyrene/divinylbenzene copolymer membranes were reacted with p-aminoazobenzene [245]. The resulting azobenzene-containing membranes were changed by UV irradiation with respect to ionexchange capacity, water content and electrical resistance, whereas control membranes having only benzyl trimethylammonium groups did not change. Transport rates in electrodialysis of sulfate, fluoride, bromide and nitrate relative to chloride ions through the azobenzene-functionalized membranes were increased by the UV irradiation. That was due to the conversion of the apolar trans form of the azobenzene moiety into the more polar cis form, which increased water content and effective pore size of the membranes. Other ion selective membranes have been prepared from polyacrylamide hydrogels containing bis-[4(dimethylamino)phenyl] (4-vinyl-phenyl)methyl leucohydroxide (TPMLH) [246] (see Ref. [247] for more details with respect to membrane preparation). After UV irradiation, an increase in flux of the anionic dye methyl orange due to the photo-induced generation of fixed cationic charges in the membrane as opposed to a combined effect of fixed charge generation and membrane structural changes was observed. Methyl orange fluxes in a fully activated membrane (at pH 4) were close to an order of magnitude higher than for UV activated membranes. It is important to mention that TPMLH could also be ionized by lowering the pH of aqueous phase, and at pH 4, 100% of TPMLH groups were ionized with no need for irradiation. Membrane pores functionalized with photo-responsive polymers. In several studies, photo-functional groups have been attached onto the pores of porous membranes [248–251]. Surface graft copolymerization of the spiropyran-containing methacrylate (1-[␤-(methacryloyl)ethyl]-3,3 -dimethyl-6-nitro-spiro(indoline-2,2 -[2H1]benzopyran), SPMMA) and acrylamide (AAm) onto a porous polytetrafluoroethylene (PTFE) membrane preactivated by glow discharge has been reported [248]. Changes in permeability of water/methanol mixtures through the poly(SPMMA-co-AAm)-grafted membranes by UV-light irradiation were related to the solubility changes of the grafted polymer chains by UV-light irradi-

D. He et al. / Progress in Polymer Science 34 (2009) 62–98

Fig. 19. Repetitive changes in permeation rate of hexane (a) and ethanol (b) through porous PE membrane, coated with a photo-chromic polymer containing crown ether and spyropyran groups, by altering UV- and visible-light irradiation (cf. Fig. 18c; reprinted with permission from [249]; Copyright (2006) American Scientific Publishers).

ation, i.e., the chain extension of the graft copolymer in the mixed solvent. For the solvent composition when the poly(SPMMA-co-AAm) synthesized in solution becomes insoluble, the permeability of the water/methanol mixture through the grafted membranes was increased (cf. Fig. 18(c)). Crown ether–spirobenzopyran copolymer was coated onto macroporous polyethylene (PE) membranes [249]. Coating was conducted by soaking of macroporous PE membrane in the polymer solution, and the solvent was evaporated to dryness at room temperature in several days. Those membranes could work as a functional membrane controlling solvent permeation rate photochemically (Fig. 19). The permeation rate of hexane decreased when the membranes were exposed to UV light due to increasing polarity of the membrane pore as a result from photo-isomerization of its spirobenzopyran moiety to corresponding ionic merocyanine form. Irradiation of membrane using visible light restored the permeation rate by isomerization back to the electrically neutral spiropyran form. By contrast, the permeation of ethanol through the membrane was enhanced by UV-light irradiation due to the increase in the apparent membrane pore size induced by the polymer chain contraction and vice versa by visible light. Photo- and thermo-responsive gate membranes were prepared by immobilizing a dual-responsive polymer hydrogel to the surface of a porous membrane [250]. The hydrogel was prepared by radical copolymerization of NIPAAm, a vinyl monomer having a spirobenzopyran

91

residue and cross-linker. It was observed that permeability of 1 mM HCl (aqueous) solution through a hydrogel membrane from that polymer increased by 2 times in response to the blue light irradiation, and that photo-response of the permeability was confirmed to be repeatable. Photo-responsive porous membranes prepared by spin coating of a polymer containing azobenzene chromophores onto aromatic amide ultrafiltration membrane were tested for gas permeability [251]. The influence of light irradiation on the gas permeability of azopolymers was investigated. The gas permeability increased by 5–10% upon irradiation with blue light resulting in cyclic photo-isomerization of the azobenzene chromophores. By contrast, no significant change was observed upon irradiation with red light. Even though the permeability change was relatively small (max. 10%), it was observed that azopolymer irradiated with blue light spread out from localized particulate domains and thinly covered the entire surface of the base membrane. Analogous modifications of inorganic membranes with a microporous barrier layer and, consequently, using low molar mass photo-chromic units have been performed by several groups (e.g., [252,253]). Membrane pores functionalized with layers of photoresponsive polymers. Aoyama et al. [254] studied photoregulation of permeability by using a membrane composed of a photo-responsive copolypeptide branch from ␤p-phenylazobenzyl l-aspartate and ␤-benzyl l-aspartate attached on a poly(butyl methacrylate). Photo-irradiation caused conformational changes of the polypeptide chains. Permeation rate of mandelic acid as well as of polar and non-polar substrate (e.g., d,l-alanine, acetone) through the membrane immersed in trimethyl phosphate increased and decreased with UV- and visible-light irradiations, respectively. It was concluded that the inversion of the helix sense of the polypeptide chains in the membrane from left-handed to right-handed with UV irradiation and the reversion on irradiation with visible light or dark adaptation was the reason behind the obtained results (cf. Fig. 18(d)). A novel preparation method for a photo-responsive membrane and demonstration of its feasibility for reducing membrane fouling was recently reported by Nayak et al. [189] (Fig. 20). They prepared a reversibly photoswitchable membrane surface based on a graft copolymer of vinyl spiropyran ((1 -(2-propylcarbamylmethacrylamide)ethyl)-3 ,3 -dimethyl-6-nitrospiro[2H-1]benzopyran-2,2 -indoline). Interestingly, this photo-responsive membrane was synthesized by using photo-grafting onto commercial PES membranes (cf. Section 4.2.3). Experiments using BSA as model for protein showed that the as-received PES membrane exhibited the highest adsorbed amount of BSA followed by the membranes with grafted vinyl spiropyran in the “closed” (visible light) and “open” (254-nm UV light) configurations. The “closed” configuration of the vinyl spiropyran surface adsorbed about 26% more protein than the “open” configuration of the surface. Similar results were shown during ultrafiltration experiments, i.e., the “closed” configuration of the vinyl spiropyran gave an about 17% lower permeation flux as compared with the “open” configuration of the vinyl spiropyran surface.

92

D. He et al. / Progress in Polymer Science 34 (2009) 62–98

Fig. 20. Schematic of graft polymerization and switching: (a) photo-grafting of vinyl monomer onto PES commercial UF membrane. (b) Exposure of the grafted PES membrane to visible light for 1 h resulting in the formation of the “closed” apolar white form of spiropyran. (c) Exposure of the grafted “closed” form of the PES membrane to UV irradiation at 254 nm for 1 h formed the “open” polar red form of spiropyran. (d) The chemical structure of the vinyl spiropyrans in two configurations as a function of UV and Vis irradiation. (e) A schematic of the chemical structure of the vinyl spiropyrans in two configurations as a function of UV and visible (Vis) irradiation (reprinted with permission from [189]; Copyright (2007) Wiley-VCH).

5. Conclusion and outlook Various applications of photo-irradiation-based technologies on membrane technologies have already been demonstrated. The widest variety of research has been devoted to preparation or modification of membranes, with

a focus on three types of barrier structures:

(i) polymer hydrogels by bulk (cross-linking) photopolymerization mainly for controlled release and sensing (cf. Section 4.1.1),

D. He et al. / Progress in Polymer Science 34 (2009) 62–98

(ii) cross-linked non-porous selective barriers, by using photo-reactive groups in the membrane polymer, mainly in gas separation and pervaporation membranes (cf. Section 4.2.1), and (iii) thin functional layers on the surface of porous membranes by “grafting-from”, mainly for micro- and ultrafiltration (cf. Section 4.2.3). For all three cases, significant improvements of membrane functionality and performance have been demonstrated. Also, technical implementation, including incorporation into continuous membrane manufacturing, is in principle possible with the available equipment developed and used for photo-curing (cf. Section 2). Further development will be necessary for adaptation to hollowfibre membranes. A versatile alternative towards barrier type (i) is photocross-linkable polymers, but this approach requires more synthetic efforts (cf. Section 4.1.2). The alternative towards barrier type (iii), “grafting-to”, is also based on synthesis of special polymer conjugates, and in addition, it has the intrinsic disadvantage of this route that it is very hard to obtain maximum surface coverage (cf. Section 4.2.2). The direct formation of thin barrier layers in a polymer membrane can also be done by photo-initiated polymerization or cross-linking, but only a few examples have been reported (cf. Section 4.1.1). In this case, the requirements to the polymer barrier with respect to flux and selectivity are of crucial importance, and the potential advantages of a photo-curing-based technology are only an added benefit, which cannot be achieved for all types of polymers (e.g., for tailored cross-linked polyamides as advanced material for RO or NF composite membranes). On the other hand, the benefit of photo-cross-linking (polymerization) to fix subtle self-assembled structures has been demonstrated with a several examples (cf. Section 4.1.1); the advantages as compared to other methods are based on fast reactions at low reaction temperature (if necessary at minimum volume change). Certainly, many more successful cases on the road to novel membrane morphologies will be reported in the near future. When considering all key advantages of photoreactions, the potential for spatially addressable membrane formation or modification, e.g., by using lithographic techniques, has been used only occasionally. Of course, most established membrane processes rely on constant and consistent membrane properties in a module. However, using such techniques for membrane development (e.g., screening of systematically varied functionality) or in cases where overall performance of a separation would be improved by barrier properties adapted to the degree of separation/polarization along a module may also be envisioned. More straightforward are applications of membrane systems with more complex (patterned) structure in sensor technologies. The integration of membranes in lab-on-chip systems will be another growing field [109], and preparation using focussed laser beams or using photo-lithography will be promising approaches. Photo-switchable membranes are conceptually very attractive (cf. Section 4.3), and many more examples will

93

certainly be developed in the future. However, beyond scientific curiosity, there are also potential applications, for instance, in controlled drug-release or signal amplification, and those functions could also well be integrated into labon-chip systems. The combination of photo-catalysis with membrane separation, one example of a more complex membranebased process (cf. Section 3.5), would be another attractive field to merge photo-irradiation- and membrane-based technologies. However, examples with polymeric membranes are rare. Some earlier works on preparation of photo-catalytic membranes via photo-polymerization have been already mentioned [96,97] (cf. Section 4.1.1); other examples are based on more conventional immobilization of photo-catalysts in polymeric membranes (e.g., [255]). Yet, in that area inorganic membranes may ultimately be superior to polymer membranes because they have a higher stability with respect to (photo)chemical degradation. Overall, photo-irradiation-based methods and processes can be very versatile enabling technologies to make the spectrum of membranes available for technical separations and other processes broader, and they will definitely also contribute to the development of entirely novel membrane-based materials and processes based thereon. References [1] Braslavsky ES. Glossary of terms used in photochemistry, 3rd ed. (IUPAC Recommendation 2006). Pure Appl Chem 2007;79:293–465. [2] Van Gerven T, Mul G, Moulijn J, Stankiewicz A. A review of intensification of photocatalytic processes. Chem Eng Process 2007;46:781–9. [3] Moad CL, Winzor DJ. Quantitative characterization of radiation degradation polymers by evaluation of scission and cross-linking yields. Prog Polym Sci 1998;23:759–813. [4] Malesic J, Kolar J, Strilic M, Kocar D, Fromageot D, Lemaire J, et al. Photo-induced degradation of cellulose. Polym Degrad Stab 2005;89:64–69. [5] Tang L, Wu Q, Qu B. The effects of chemical structure and synthesis method on photodegradation of polypropylene. J Appl Polym Sci 2007;95:270–9. [6] Norman K, Kingshott P, Kaeselev B, Ghanbari-Siahkali A. Photodegradation of poly(ether sulphone). Part 1. A time-offlight secondary ion mass spectrometry study. Surf Interf Sci 2004;36:1533–41. [7] Kotzyba-Hibert F, Kapfer I, Goeldner M. Recent trends in photoaffinity labeling. Angew Chem Int Ed 1995;34:1296–312. [8] Decker C. Photoinitiated crosslinking polymerization. Prog Polym Sci 1996;21:593–650. [9] Andrzejewska E. Photopolymerization kinetics of multifunctional monomers. Prog Polym Sci 2001;26:605–65. [10] Dyer DJ. Photoinitiated synthesis of grafted polymers. Adv Polym Sci 2006;197:47–65. [11] Endruweit A, Johnson MS, Long AC. Curing of composite components by ultraviolet radiation: a review. Polym Compos 2006;27:119–28. [12] Zhang PY, Ranby B. Surface modification by continuous graft copolymerization. II. Photoinitiated graft copolymerization onto polypropylene film surface. J Appl Polym Sci 1991;43:621–36. [13] Yang W, Ranby B. Bulk surface photografting process and its applications. I. Reactions and kinetics. J Appl Polym Sci 1996;62:533–43. [14] Yang W, Ranby B. Bulk surface photografting process and its applications. II. Principal factors affecting surface photografting. J Appl Polym Sci 1996;62:545–55. [15] Yager KG, Barrett CJ. Novel photo-switching using azobenzene functional materials. J Photochem Photobiol A: Chem 2006;182:250–61. [16] Irie M. Diarylethenes for memories and switches. Chem Rev 2000;100:1685–716. [17] Tian H, Yang S. Recent progresses on diarylethene based photochromic switches. Chem Soc Rev 2004;33:85–97.

94

D. He et al. / Progress in Polymer Science 34 (2009) 62–98

[18] Berkovic G, Krongauz V, Weiss V. Spiropyrans and spirooxazines for memories and switches. Chem Rev 2000;100:1741–54. [19] van de Witte P, Dijkstra PJ, van den Berg JWA, Feijen J. Phase separation processes in polymer solutions in relation to membrane formation. J Membr Sci 1996;117:1–31. [20] Stropnik C, Kaiser V. Polymeric membranes preparation by wet phase separation: mechanisms and elementary processes. Desalination 2002;145:1–10. [21] Lloyd DR, Kinzer KE, Tseng HS. Microporous membrane formation via thermally induced phase separation. I. Solid–liquid phase separation. J Membr Sci 1990;52:239–61. [22] Chen G, Lin Y, Wang X. Formation of microporous membrane of isotactic polyproplylene in dibutyl phthalate-soybean oil via thermally induced phase separation. J Appl Polym Sci 2007;105: 2000–7. [23] Jeon MY, Kim CK. Phase behavior of polymer/diluent/diluent mixtures and their application to control microporous membrane structure. J Membr Sci 2007;300:172–81. [24] Zeman L, Fraser T. Formation of air-cast cellulose acetate membranes. Part I. Study of macrovoid formation. J Membr Sci 1993;84:93–106. [25] Pekny MR, Zartman J, Krantz WB, Greenberg AR, Todd P. Flowvisualization during macrovoid pore formation in dry-cast cellulose acetate membranes. J Membr Sci 2003;211:71–90. [26] Matsuyama H, Teramoto M, Nakatani R, Maki T. Membrane formation via phase separation induced by penetration on nonsolvent from vapor phase. I. Phase diagram and mass transfer process. J Appl Polym Sci 1999;74:159–70. [27] Yip Y, McHugh AJ. Modeling and simulation of nonsolvent vaporinduced phase separation. J Membr Sci 2006;271:163–76. [28] Khare VP, Greenberg AR, Krantz WB. Vapor-induced phase separation—effect of the humid air exposure step on membrane morphology. Part I. Insight from mathematical modeling. J Membr Sci 2005;258:140–56. [29] Sadeghi F, Ajji A, Carreau PJ. Analysis of microporous membranes obtained from polypropylene films by stretching. J Membr Sci 2007;292:67–71. [30] Yamazaki IM, Paterson R, Geraldo LP. A new generation of track etched membranes for microfiltration and ultrafiltration. Part I. Preparation and characterization. J Membr Sci 1996;118: 239–45. [31] Tamada M, Asano M, Spohr R, Vetter J, Trautmann C, Yoshida M, et al. Preparation of hydrolyzed pH responsive ion track membrane. Macromol Rapid Commun 1995;16:47–51. [32] Peterson RJ. Composite reverse osmosis and nanofiltration membranes. J Membr Sci 1993;83:81–150. [33] Koros WJ. Gas separation membranes: needs for combined materials science and processing approaches. Macromol Symp 2002;188:13–22. [34] Cadotte JE. Evolution of composite reverse-osmosis membranes. ACS Symp Ser 1985;269:273–94. [35] Korikov AP, Kosaraju PB, Sirkar KK. Interfacially polymerized hydrophilic microporous thin film composite membranes on porous polypropylene hollow fibers and flat films. J Membr Sci 2006;279:588–600. [36] Wang X, Fang D, Yoon K, Hsiao BS, Chu B. High performance ultrafiltration composite membranes based on poly(vinyl alcohol) hydrogel coating on crosslinked nanofibrous poly(vinyl alcohol) scaffold. J Membr Sci 2006;278:261–8. [37] Kulprathipanja S. Mixed matrix membrane development. Ann NY Acad Sci 2003;984:361–9. [38] Chung TS, Jiang LY, Kulprathipanja S. Mixed matrix membranes (MMMs) comprising organic polymers with dispersed inorganic fillers for gas separation. Prog Polym Sci 2007;32:483– 507. [39] Moore TT, Mahajan R, Vu DQ, Koros WJ. Hybrid membrane materials comprising organic polymers with rigid dispersed phases. AIChE J 2004;50:311–21. [40] Ulbricht M. Advanced functional polymer membranes. Polymer 2006;47:2217–62. [41] Hilal N, Ogunbiyi OO, Miles NJ, Nigmatullin R. Methods employed for control of fouling in MF and UF membranes: a comprehensive review. Sep Sci Technol 2005;40:1957–2005. [42] Koehler JA, Ulbricht M, Belfort G. Intermolecular forces between polymer films and proteins with relevance to filtration. Langmuir 1997;13:4162–71. [43] Koehler JA, Ulbricht M, Belfort G. Intermolecular forces between a protein and a hydrophilic modified polysulfone film with relevance to filtration. Langmuir 2000;16:10419–27.

[44] Susanto H, Franzka S, Ulbricht M. Dextran fouling of ultrafiltration membranes—causes, extent and consequences. J Membr Sci 2007;296:147–55. [45] Maier G. Gas separation with polymer membranes. Angew Chem Int Ed 1998;37:2960–74. [46] Pandey P, Chauhan RS. Membranes for gas separation. Prog Polym Sci 2001;26:853–93. [47] Ismail AF, Lorna W. Penetrant-induced plasticization phenomenon in glassy polymers for gas separation membrane. Sep Purif Technol 2002;27:173–94. [48] Gevers LEM, Meyen G, Smet KD, Velde PVD, Du Prez FE, Vankelecom IFJ, et al. Physico-chemical interpretation of the SRNF transport mechanism for solutes through dense silicone membranes. J Membr Sci 2006;274:173–82. [49] Vankelecom IFJ. Polymeric membranes in catalytic reactors. Chem Rev 2002;102:3779–810. [50] Li Q, He R, Jensen JO, Bjerrum NJ. Approaches and recent development of polymer electrolyte membranes for fuel cells operating above 100 ◦ C. Chem Mater 2003;15:4896–915. [51] Hickner MA, Ghassemi H, Kim YS, Einsla BR, McGrath JE. Alternative polymer systems for proton exchange membranes (PEMs). Chem Rev 2004;104:4587–612. [52] Souzy R, Ameduri B. Functional fluoropolymers for fuel cell membranes. Prog Polym Sci 2005;30:644–87. [53] Judd S. The MBR book: principle and applications of membrane bioreactor in water and wastewater treatment. Elsevier; 2006. [54] van Reis R, Zydney A. Bioprocess membrane technology. J Membr Sci 2007;297:16–50. [55] Roper DK, Lightfoot EN. Separation of biomolecules using adsorptive membranes. J Chromatogr A 1995;702:3–26. [56] Zou H, Lou Q, Zhou D. Affinity membrane chromatography for the analysis and purification of proteins. J Biochem Biophys Meth 2001;49:199–240. [57] Ghosh R. Protein separation using membrane chromatography: opportunities and challenges. J Chromatogr A 2002;952:13–27. [58] Lightfoot EN, Moscariello JS. Bioseparations. Biotechnol Bioeng 2004;87:259–73. [59] Castilho LR, Deckwer WD, Anspach FB. Influence of matrix activation and polymer coating on the purification of human IgG with protein A affinity membranes. J Membr Sci 2000;172:269–77. [60] Borcherding H, Hicke HG, Jorcke D, Ulbricht M. Affinity membranes as a tool for life science applications. Ann NY Acad Sci 2003;984:470–9. [61] Liu X, Neoh KG, Kang ET. Redox-sensitive microporous membranes prepared from poly(vinylidene fluoride) grafted with viologencontaining polymer side chains. Macromolecules 2003;36:8361–7. [62] Peng T, Cheng YL. PNIPAAm and PMAA-co-grafted porous PE membranes: living radical co-grafting mechanism and multi-stimuli responsive permeability. Polymer 2001;42:2091–100. [63] Friebe A, Ulbricht M. Controlled pore functionalization of poly(ethylene terephthalate) track-etched membranes via surface-initiated atom transfer radical polymerization. Langmuir 2007;23:10316–22. [64] Chu LY, Niitsuma T, Yamaguchi T, Nakao S. Thermoresponsive transport through porous membranes with grafted PNIPAM gates. AIChE J 2003;49:896–909. [65] Peppas NA, Hilt JZ, Khademhosseini A, Langer R. Hydrogels in biology and medicine: from molecular principles to bionanotechnology. Adv Mater 2006;18:1345–60. [66] Peppas NA, Kim B. Stimuli-sensitive protein delivery systems. J Drug Deliv Sci Technol 2006;16:11–8. [67] Morishita M, Goto T, Takayama K, Peppas NA. Oral insulin delivery systems based on complexation polymer hydrogels. J Drug Deliv Sci Technol 2006;16:19–24. [68] Adhikari B, Majumdar S. Polymers in sensor applications. Prog Polym Sci 2004;29:699–766. [69] Sanchez Marcano JG, Tsotsis TT. Catalytic membrane reactors and membrane reactors. Weinheim: Wiley-VCH; 2002. [70] Drioli E, Giorno L. Biocatalytic membrane reactors. London: Taylor and Francis Ltd.; 1999. [71] Lanza RP, Langer R, Vacanti J, editors. Principles of tissue engineering. 2nd ed. San Diego, CA: Academic Press; 2000. [72] Wang PC, DeVoe DL, Lee CS. Integration of polymeric membranes with microfluidic networks for bioanalytical applications. Electrophoresis 2001;22:3857–3867. [73] Murphy SM, Hamilton CJ, Davies ML, Tighe BJ. Polymer membranes in clinical sensor applications. II. The design and fabrication of permselective hydrogels for electrochemical devices. Biomaterials 1992;13:979–90.

D. He et al. / Progress in Polymer Science 34 (2009) 62–98 [74] Arica Y, Hasirci VN. Immobilization of glucose oxidase in poly(2-hydroxyethyl methacrylate) membranes. Biomaterials 1987;8:489–95. [75] Arica MY, Denizli A, Salih B, Piskin E, Hasirci V. Catalase adsorption onto Cibacron Blue F3GA and Fe(III)-derivatized poly(hydroxyethyl methacrylate) membranes and application to a continuous system. J Membr Sci 1997;129:65–76. [76] Edelmann K, Reiche A, Sandner B, Marstalerz J, Müller H. Copolymers from oligosiloxane methacrylates as a plasticizer-free membrane matrix for ion-selective sensors. Polymer 2005;46: 407–17. [77] van der Wal PD, Skowronska-Ptasinska M, van den Berg A, Bergveld P, Sudhölter EJR, Reinhoudt DN. New membrane materials for potassium-selective ion-sensitive field-effect transistors. Anal Chim Acta 1990;231:41–52. [78] Elmer AM, Jannasch P. Polymer electrolyte membranes by in situ polymerization of poly(ethylene carbonate-co-ethylene oxide) macromonomers in blends with poly(vinylidene fluoride-cohexafluoropropylene). J Polym Sci B: Polym Phys 2007;45:79–90. [79] Song S, Singh AK, Shepodd TJ, Kirby BJ. Microchip dialysis of proteins using in situ photopatterned nanoporous polymer membranes. Anal Chem 2004;76:2367–73. [80] Song S, Singh AK, Kirby BJ. Electrophoretic concentration of proteins at laser-patterned nanoporous membranes in microchips. Anal Chem 2004;76:4589–92. [81] Nakayama H, Kaetsu I, Uchida K, Sakata S, Tougou K, Hara T, et al. Radiation curing of intelligent coating for controlled release and permeation. Radiat Phys Chem 2002;63:521–3. [82] Kaetsu I, Nakayama H, Uchida K, Sutani K. Radiation curing of intelligent coating on biofunctional membranes. Radiat Phys Chem 2001;60:513–20. [83] Ng LT, Ng KS. Photo-cured pH-responsive polyampholyte-coated membranes for controlled release of drugs with different molecular weights and charges. Radiat Phys Chem 2008;77:192–9. [84] Ng LT, Nakayama H, Kaetsu I, Uchida K. Photocuring of stimulus responsive membranes for controlled-release of drugs having different molecular weight. Radiat Phys Chem 2005;73:117–23. [85] Nakayama H, Kaetsu I, Uchida K, Oishibashi M, Matsubara Y. Intelligent biomembranes for nicotine releases by radiation curing. Radiat Phys Chem 2003;67:367–70. [86] Spizzirri UG, Peppas NA. Structural analysis and diffusional behavior of molecularly imprinted polymer networks for cholesterol recognition. Chem Mater 2005;17:6719–27. [87] Hilt JZ, Byrne ME, Peppas NA. Microfabrication of intelligent biomimetic networks for recognition of d-glucose. Chem Mater 2006;18:5869–75. [88] Liu Y, Lee JY, Hong L. In situ preparation of poly(ethylene oxide)-SiO2 composite polymer electrolytes. J Power Sources 2004;129:303–11. [89] Nardin C, Winterhalter M, Meier W. Giant free-standing ABA triblock copolymer membranes. Langmuir 2000;16:7708–12. [90] Marek M, Brynda E, Pientka Z, Schauer J. Crosslinked ultra-thin polyimide film as a gas separation layer for composite membranes. Eur Polym J 1997;33:1717–21. [91] Steuck MJ. Porous membrane having hydrophilic surface and process. US Patent 4,618,533; 1986. [92] Hu H, Cai Z. Hydrophilic microporous polyolefin membrane. US Patent 5,209,849; 1993. [93] Liu C, Martin CR. Composite membranes from photochemical synthesis of ultrathin polymer films. Nature 1991;352:50–2. [94] Hong JM, Anderson PE, Qian J, Martin CR. Selectively-permeable ultrathin film composite membranes based on molecularlyimprinted polymers. Chem Mater 1998;10:1029–33. [95] Don TM, Chen CC, Lee CK, Cheng WY, Cheng LP. Preparation and antibacterial test of chitosan/PAA/PEGDA bi-layer composite membranes. J Biomater Sci Polym Ed 2005;16:1503–19. [96] Bellobono IR, Carrara A, Barni B, Gazzotti A. Laboratory- and pilot-plant-scale photodegradation of chloroaliphatics in aqueous solution by photocatalytic membranes immobilizing titanium dioxide. J Photochem Photobiol A: Chem 1994;84:83–90. [97] Bellobono IR, Morelli R, Chiodaroli CM. Photocatalysis and promoted photocatalysis during photocrosslinking of multifunctional acrylates in composite membranes immobilizing titanium dioxide. J Photochem Photobiol A: Chem 1997;105:89–94. [98] Beginn U, Zipp G, Möller M. Functional membranes containing ion-selective matrix-fixed supramolecular channels. Adv Mater 2000;12:510–3. [99] Beginn U, Zipp G, Mourran A, Walther P, Möller M. Membranes containing oreiented supramolecular transport channels. Adv Mater 2000;12:513–6.

95

[100] Wei Q, James SL. A metal–organic gel used as a template for a porous organic polymer. Chem Commun 2005:1555–6. [101] Tan G, Singh M, He J, John VT, McPherson GL. Use of a selfassembling organogel as a reverse template in the preparation of imprinted porous polymer films. Langmuir 2005;21:9322–6. [102] Simon FX, Khelfallah NS, Schmutz M, Diaz N, Mesini PJ. Formation of helical mesoporous in organic polymer matrices. J Am Chem Soc 2007;129:3788–9. [103] Xu H, Goedel WA. Polymer-silica hybrid monolayers as precursors for ultrathin free-standing porous membranes. Langmuir 2002;18:2363–7. [104] Xu H, Goedel WA. From particle-assisted wetting to thin freestanding porous membranes. Angew Chem Int Ed 2003;42: 4694–6. [105] Yan F, Goedel WA. A simple and effective method for the preparation of porous membranes with three-dimensionally arranged pores. Adv Mater 2004;16:911–5. [106] Tierno P, Thonke K, Goedel WA. Using paramagnetic particles as repulsive templates for the preparation of membranes of controlled porosity. Langmuir 2005;21:9476–81. [107] Yabu H, Kojima M, Tsubouchi M, Onoue S, Sugitani M, Shimomura M. Fabrication of photo-crosslinked honeycomb-patterned films. Colloid Surf A: Physichochem Eng Aspects 2006;284–285: 254–6. [108] Tong HD, Jansen HV, Gadgil VJ, Bostan CG, Berenschot E, van Rijn CJM, et al. Silicone nitride nanosieve membrane. Nano Lett 2004;4:283–7. [109] de Jong J, Lammertink RGH, Wessling M. Membranes and microfluidics: a review. Lab Chip 2006;6:1125–39. [110] Liu JH, Chen MY, Lin MT. Application of UV-curable diazoresin. I. Immobilization of glucose oxidase into PVA in the presence of UV-sensitive diazoresin and sensitizers. J Appl Polym Sci 1990;40:2161–71. [111] Liu JH, Chung YC, Lin MT. Preparation of amphiphilic membranes using photosensitive polymers and their applications on the enzyme electrode. J Appl Polym Sci 1995;55:1441–9. [112] Nakayama Y, Matsuda T. Preparation and characteristics of photocrosslinkable hydrophilic polymer having cinnamate moiety. J Polym Sci A: Polym Chem 1992;30:2451–7. [113] Uhlich T, Ulbricht M, Tomaschewski G, Stösser R. ESR studies of chemically modified and photocrosslinked poly(vinyl alcohol) gels. J Appl Polym Sci 1995;58:1567–76. [114] Uhlich T, Ulbricht M, Tomaschewski G. Immobilization of enzymes in photo-chemically cross-linked polyvinyl alcohol. Enzyme Microbial Technol 1996;19:124–31. [115] Wehrens S, Tomaschewski G. Synthesis of a new photocrosslinkable prepolymer, based on poly(styrene-co-maleic anhydride). React Funct Polym 1996;29:21–7. [116] Murata H, Chang BJ, Prucker O, Dahm M, Rühe J. Polymeric coating for biomedical devices. Surf Sci 2004;570:111–8. [117] Toomey R, Freidank D, Rühe J. Swelling behavior of thin, surfaceattached polymer networks. Macromolecules 2004;37:882–7. [118] Kang JS, Won J, Park HC, Kim UY, Kang YS, Lee YM. Morphology control of asymmetric membranes by UV irradiation on polyimide dope solution. J Membr Sci 2000;169:229–35. [119] Liu G, Ding J, Stewart S. Preparation and properties of nanoporous triblock copolymer membranes. Angew Chem Int Ed 1998;38:835–8. [120] Ludwigs S, Schmidt K, Stafford CM, Amis EJ, Fasolka MJ, Karim A, et al. Combinatorial mapping of the phase behaviour of ABC triblock terpolymers in thin film: experiments. Macromolecules 2005;38:1850–8. [121] Sperschneider A, Schacher F, Gawenda M, Tsarkova L, Müller AHE, Ulbricht M, et al. Towards nanoporous membranes based on ABC triblock terpolymer. Small 2007;6:1056–63. [122] Hayes RA. Polyimide gas separation membranes. US Patent 4,717,393; 1988. [123] Kita H, Inada T, Tanaka K, Okamoto K. Effect of photocrosslinking on permeability and permselectivity of gases through benzophenone containing polyimide. J Membr Sci 1994;87:139–47. [124] Liu Y, Ding M, Xu J. Gas permeabilities and permselectivity of photochemically crosslinked polyimides. J Appl Polym Sci 1995;58:485–9. [125] Liu Y, Pan C, Ding M, Xu J. Effect of crosslinking distribution on gas permeability and permselectivity of crosslinked polyimides. Eur Polym J 1999;35:1739–41. [126] Scaiano JC, Netto-Ferreira JC, Becknell AF, Small RD. The mechanism of photocure of inherently photosensitive polyimides containing a benzophenone group. Polym Eng Sci 1989;29:942–4.

96

D. He et al. / Progress in Polymer Science 34 (2009) 62–98

[127] Lin AA, Sastri VR, Tesoro G, Reiser A, Eachus R. On the cross-linking mechanism of benzophenone-containing polyimides. Macromolecules 1988;21:1165–9. [128] Meier IK, Langsam M. Photochemically induced oxidative surface modification of polyimide films. J Polym Sci A: Polym Chem 1993;31:83–9. [129] Meier IK, Langsam M, Klotz HC. Selectivity enhancement via photooxidative surface modification of polyimide air separation membranes. J Membr Sci 1994;94:195–212. [130] Matsui S, Ishiguro T, Higuchi A, Nakagawa T. Effect of ultraviolet light irradiation on gas permeability in polyimide membranes. 1. Irradiation with low pressure mercury lamp on photosensitive and nonphotosensitive membranes. J Polym Sci B: Polym Phys 1997;32:2259–69. [131] Matsui S, Nakagawa T. Effect of ultraviolet light irradiation on gas permeability in polyimide membranes. II. Irradiation of membranes with high-pressure mercury lamp. J Appl Polym Sci 1998;67: 49–60. [132] Matsui S, Sato H, Nakagawa T. Effects of low molecular weight photosensitizer and UV irradiation on gas permeability and selectivity of polyimide membrane. J Membr Sci 1998;141: 31–43. [133] Wright CT, Paul DR. Gas sorption and transport in UV-irradiated polyarylate copolymers based on tetramethyl bisphenol-A and dihydroxybenzophenon. J Membr Sci 1997;124:161–74. [134] McCaig MS, Paul DR. Effect of UV crosslinking and physical aging on the gas permeability of thin glassy polyarylate films. Polymer 1999;40:7209–25. [135] Wright CT, Paul DR. Gas sorption and transport in UV-irradiated poly(2,6-dimethyl-1,4-phenylene oxide) films. J Appl Polym Sci 1998;67:875–83. [136] Tanihara N, Tanaka K, Kita H, Okamoto K. Pervaporation of organic liquid mixtures through membranes of polyimides containing methyl-substituted phenylenediamine moieties. J Membr Sci 1994;95:161–9. [137] Ishikawa J, Fujiyama S, Inoue K, Omi T, Tamai S. Highly sulfonated poly(aryl ether ketone) block coplymers having a cross-linking structure. J Membr Sci 2007;298:48–55. [138] Wycisk R, Pintauro PN, Wang W, O’Connor S. Polyphosphazene membranes. I. Solid-state photocrosslinking of poly[(4ethylphenoxy)-(phenoxy)phosphazene]. J Appl Polym Sci 1996;59:1607–17. [139] Graves R, Pintauro PN. Polyphosphazene membranes. II. Solid-state photocrosslinking of poly[(alkylphenoxy)-(phenoxy)phosphazene] films. J Appl Polym Sci 1998;68:827–36. [140] Zhong S, Cui X, Cai H, Fu T, Zhao C, Na H. Crosslinked sulfonated poly(ether ether ketone) proton exchange membranes for direct methanol fuel cell applications. J Power Sources 2007;164: 65–72. [141] Dudley CN, Schöberl B, Sturgill GK, Becham HW, Rezac ME. Influence of crosslinking technique on the physical and transport properties of ethynyl-terminated monomer/polyetherimide asymmetric membranes. J Membr Sci 2001;191:1–11. [142] Burgoyne WF, Langsam M, Ford ME, Casey JP. Membranes formed from unsaturated polyimides. US Patent 4,931,182; 1990. [143] Langsam M, Burgoyne WF, Casey JP, Ford ME. Membranes formed from unsaturated polyimides. US Patent 4,952,220; 1990. [144] Abdellah L, Boutevin B, Guida-Pietrasanta F, Smaihi M. Evaluation of photocrosslinkable fluorinated polydimethylsiloxane as gas permeation membranas. J Membr Sci 2003;217:295–8. [145] Luo Y, Xin W, Li G, Yang Y, Liu J, Lv Y, et al. Pervaporation properties of EC membrane crosslinked by hyperbranched-polyester acrylate. J Membr Sci 2007;303:183–93. [146] Trushinski BJ, Dickson JM, Childs RF, McCarry BE. Photochemically modified thin-film composite membranes. I. Acid and ester membranes. J Appl Polym Sci 1993;48:187–98. [147] Trushinski BJ, Dickson JM, Childs RF, McCarry BE, Gagnon DR. Photochemically modified thin-film composite membranes. II. Bromoethyl ester, dioxolan, and hydroxyethyl ester membranes. J Appl Polym Sci 1994;54:1233–42. [148] Darkow R, Yoshikawa M, Kitao T, Tomaschewski G. Photomodification of a poly(acrylonitrile-co-butadiene-co-styrene) containing diaryltetrazolyl groups. J Polym Sci A: Polym Chem 1994;32:1657–64. [149] Bora U, Sharma P, Kannan K, Nahar P. Photoreactive cellulose membrane—a novel matrix for covalent immobilization of biomolecules. J Biotechnol 2006;126:220–9. [150] Kumar S, Nahar P. Sunlight-induced covalent immobilization of proteins. Talanta 2007;71:1438–40.

[151] Rajam S, Ho C. Graft coupling of PEO to mixed cellulose esters microfiltration membranes by UV irradiation. J Membr Sci 2006;281:211–8. [152] Ulbricht M, Hicke HG. Photomodification of ultrafiltration membranes. 1. Photochemical modification of polyacrylonitrile ultrafiltration membranes with aryl azides. Angew Makromol Chem 1993;210:69–95. [153] Ulbricht M, Hicke HG. Photomodification of ultrafiltration membranes. 2. Ultrafiltration properties of polyacrylonitrile membranes photochemically modified with aryl azides. Angew Makromol Chem 1993;210:97–117. [154] Thom V, Jankova K, Ulbricht M, Kops J, Jonsson G. Synthesis of photoreactive ␣-4-azidobenzoyl-␻-methoxy-poly(ethylene glycol)s and their end-on photo-grafting onto polysulfone ultrafiltration membranes. Macromol Chem Phys 1998;199: 2723–9. [155] Thom VH, Jonsson G. Photochemical grafting of poly (ethyleneglycol)s yielding low-protein-adsorbing UF membranes. Acta Polytech Scand Chem Technol Ser 1997;247:35–50. [156] Tian M, Zhong R, Sun SD, Zhao CS, Huang XH, Yue YL. Comparison of two approaches to grafting hydrophilic polymer chains onto polysulfone films. J Appl Polym Sci 2007;103:3818–26. [157] Thom VH, Altankov G, Groth T, Jankova K, Jonsson G, Ulbricht M. Optimizing cell–surface interactions by photo-grafting of poly(ethylene glycol). Langmuir 2000;16:2756–65. [158] Altankov G, Thom VH, Groth T, Jankova K, Jonsson G, Ulbricht M. Modulating the biocompatibility of polymer surface with poly(ethylene glycol): effect of fibronectin. J Biomed Mater Res 2000;52:219–30. [159] Lazos D, Franzka S, Ulbricht M. Size-selective protein adsorption to polystyrene surfaces by self-assembled grafted polyethylene glycols with variied chain lengths. Langmuir 2005;21:8774–84. [160] Park YS, Ito Y, Imanishi Y. Permeation control through porous membranes immobilized with thermosensitive polymer. Langmuir 1998;14:910–4. [161] Lequieu W, Shtanko NI, Du Prez F. Track etched membranes with thermo-adjustable porosity and separation properties by surface immobilization of poly(N-vinylcaprolactam). J Membr Sci 2005;256:64–71. [162] Yang B, Yang W. Photografting modification of PET nucleopore membranes. J Macromol Sci A: Pure Appl Chem 2003;A40:309–20. [163] Park YS, Won J, Kang YS. Preparation of poly(ethylene glycol) brushes on polysulfone membranes for olefin/paraffin separation. Langmuir 2000;16:9662–5. [164] Revanur R, McCloskey B, Breitenkamp K, Freeman BD, Emrick T. Macromolecules 2007;40:3624–30. [165] Ulbricht M, Schwarz HH. Novel high performance photo-graft composite membranes for separation of organic liquids by pervaporation. J Membr Sci 1997;136:25–33. [166] Susanto H, Ulbricht M. Photografted thin polymer hydrogel layers on PES ultrafiltration membranes: characterization, stability, and influence on separation performance. Langmuir 2007;23:7818–30. [167] Ulbricht M, Oechel A, Lehmann C, Tomaschewski G, Hicke HG. Gas-phase photoinduced graft polymerization of acrylic acid onto polyacrylonitrile ultrafiltration membranes. J Appl Polym Sci 1995;55:1707–23. [168] Ulbricht M, Yang H. Porous polypropylene membranes with different carboxyl polymer brush layers for reversible protein binding via surface-initiated graft copolymerization. Chem Mater 2005;17:2622–31. [169] Sergeyeva TA, Matuschewski H, Piletsky SA, Bendig J, Schedler U, Ulbricht M. Molecularly imprinted polymer membranes for substance-selective solid-phase extraction from water by surface photo-grafting polymerisation. J Chromatogr A 2001;907: 89–99. [170] Piletsky SA, Matuschewski H, Schedler U, Wilpert A, Piletska E, Thiele TA, et al. Surface functionalization of porous polypropylene membranes with molecularly imprinted polymers by photograft copolymerization in water. Macromolecules 2000;33:3092–8. [171] Hicke HG, Ulbricht M, Becker M, Radosta S, Heyer AG. Novel enzyme-membrane reactor for polysaccharide synthesis. J Membr Sci 1999;161:239–45. [172] Becker M, Provart N, Lehmann I, Ulbricht M, Hicke HG. Polymerization of glucans by enzymatically active membranes. Biotechnol Prog 2002;18:964–8. [173] Hicke HG, Becker M, Paulke BR, Ulbricht M. Covalently coupled nanoparticles in capillary pores as enzyme carrier and as turbulence promoter to facilitate enzymatic polymerization in flow-through enzyme-membrane reactors. J Membr Sci 2006;282:413–22.

D. He et al. / Progress in Polymer Science 34 (2009) 62–98 [174] Ulbricht M. Photograft-polymer-modified microporous membranes with environment-sensitive permeabilities. React Funct Polym 1996;31:165–77. [175] Geismann C, Ulbricht M. Photoreactive functionalization of poly(ethylene terephthalate) track-etched pore surfaces with “smart” polymer systems. Macromol Chem Phys 2005;206:268–81. [176] Geismann C, Yaroshchuk A, Ulbricht M. Permeability and electrokinetic characterization of poly(ethylene terephthalate) capillary pore membranes with grafted temperature-responsive polymers. Langmuir 2007;23:76–83. [177] Crivello JV, Belfort G, Yamagishi H. Low fouling ultrafiltration and microfiltration aryl polysulfone, US Patent 5,468,390; 1995. [178] Yamagishi H, Crivello JV, Belfort G. Development of a novel photochemical technique for modifying poly(arylsulfone) ultrafiltration membranes. J Membr Sci 1995;105:237–47. [179] Kilduff JE, Mattaraj S, Zhou M, Belfort G. Kinetics of membrane flux decline: the role of natural colloids and mitigation via membrane surface modification. J Nanopart Res 2005;7:525–44. [180] Ulbricht M, Riedel M, Marx U. Novel photochemical surface functionalization of polysulfone ultrafiltration membranes for covalent immobilization of biomolecules. J Membr Sci 1996;120:239–59. [181] Yamagishi H, Crivello JV, Belfort G. Evaluation of photochemically modified poly(arylsulfone) ultrafiltration membranes. J Membr Sci 1995;105:249–59. [182] Pieracci J, Crivello JV, Belfort G. UV-assisted graft polymerization of N-vinyl-2-pyrrolidinone onto poly(ether sulfone) ultrafiltration membranes using selective UV wavelengths. Chem Mater 2002;14:256–65. [183] Taniguchi M, Pieracci J, Samsonoff WA, Belfort G. UV-assisted graft polymerization of synthetic membranes: mechanistic studies. Chem Mater 2003;15:3805–12. [184] Kaeselev B, Pieracci J, Belfort G. Photoinduced grafting of ultrafiltration membranes: comparison of poly(ether sulfone) and poly(sulfone). J Membr Sci 2001;194:245–61. [185] Taniguchi M, Belfort G. Low protein fouling synthetic membrane by UV-assisted surface grafting modification: varying monomer type. J Membr Sci 2004;231:147–57. [186] Pieracci J, Wood D, Crivello J, Belfort G. UV-assisted graft polymerization of N-vinyl-2-pyrrolidinone onto poly-(ether sulfone) ultrafiltration membranes: comparison of dip versus immersion modification techniques. Chem Mater 2000;12:2123–33. [187] Pieracci J, Crivello JV, Belfort G. Increasing membrane permeability of UV-modified poly(ether sulfone) ultrafitration membranes. J Membr Sci 2002;202:1–16. [188] Susanto H, Balakrishnan M, Ulbricht M. Via surface functionalization by photograft copolymerization to low-fouling polyethersulfone-based ultrafiltration membranes. J Membr Sci 2007;288:157–67. [189] Nayak A, Liu H, Belfort G. An optically reversible switching membrane surface. Angew Chem Int Ed 2006;45:4094–8. [190] Bequet S, Remigy JC, Rouch JC, Espenan JM, Clifton M, Aptel M. From ultrafiltration to nanofiltration hollow fiber membranes: a continuous UV-photografting process. Desalination 2002;144:9–14. [191] Yanagishita H, Kitamoto D, Ikegami T, Negishi H, Endo A, Haraya K, et al. Preparation of photo-induced graft filling polymerized membranes for pervaporation using polyimide with benzophenone structure. J Membr Sci 2002;203:191–9. [192] Qiu C, Xu F, Nguyen QT, Ping Z. Nanofiltration membrane prepared from cardo polyetherketone ultrafiltration membrane by UV-induced grafting method. J Membr Sci 2005;255:107–15. [193] Qiu C, Nguyen QT, Ping Z. Surface modification of cardo polyetherketone ultrafiltration membrane by photo-grafted copolymers to obtain nanofiltration membranes. J Membr Sci 2007;295:88–94. [194] Wang HY, Kobayashi T, Fujii N. Surface molecular imprinting on photosensitive dithiocarbamoyl polyacrylonitrile membranes using photograft polymerization. J Chem Technol Biotechnol 1997;70:355–62. [195] Hattori K, Hiwatari M, Iiyama C, Yoshimi Y, Kohori F, Sakai K, et al. Gate effect of theophylline-imprinted polymers grafted to the cellulose by living radical polymerization. J Membr Sci 2004;233:169–73. [196] Guan J, Gao C, Feng L, Shen J. Functionalizing of polyurethane surface by photografting with hydrophilic monomers. J Appl Polym Sci 2000;77:2505–12. [197] Guan J, Gao C, Feng L, Shen J. Surface photo-grafting of polyurethane with 2-hydroxyethyl acrylate for promotion of human endothelial cell adhesion and growth. J Biomater Sci Polym Ed 2000;11:523–36. [198] Guan J, Gao C, Feng L, Shen J. Preparation of functional poly(ether-urethane) for immobilization of human living cells.

[199]

[200]

[201]

[202]

[203] [204]

[205]

[206] [207]

[208]

[209]

[210]

[211]

[212]

[213]

[214]

[215]

[216] [217]

[218]

[219]

[220]

[221]

97

1. Surface graft polymerization of poly(ether-urethane) with 2(dimethylamino)ethyl methacrylate and quaternization of grafted membrane. Eur Polym J 2000;36:2707–13. Gao C, Guan J, Zhu Y, Shen J. Surface immobilization of bioactive molecules on polyurethane for promotion of cytocompatibility to human endothelial cells. Macromol Biosci 2003;3:157–62. Kochkodan V, Weigel W, Ulbricht M. Thin layer molecularly imprinted microfiltration membranes by photofunctionalization using a coated alpha-cleavage photoinitiator. Analyst 2001;126:803–9. Hilal N, Kochkodan V. Surface modified microfiltration membranes with molecularly recognizing properties. J Membr Sci 2003;213:97–113. Tazuke S, Kimura H. Surface photografting, modification of polypropylene film surface by graft polymerization of acrylamide. Makromol Chem 1978;179:2603–12. Yang W, Ranby B. Radical living graft polymerization on the surface of polymeric materials. Macromolecules 1996;29:3308–10. Geuskens G, Ruckert D. Surface modification of polymers. IV. Grafting of acrylamide via an unexpected mechanism using a water soluble photo-initiator. Eur Polym J 1996;32:201–8. Yang B, Yang W. Thermo-sensitive switching membranes regulated by pore-covering polymer brushes. J Membr Sci 2003;218: 247–55. Yang B, Yang W. Novel pore-covering membrane as a full open/close valve. J Membr Sci 2005;258:133–9. Wu G, Li Y, Han M, Liu X. Novel thermo-sensitive membranes prepared by rapid bulk photo-grafting polymerization of N,Ndiethylacrylamide onto the microfiltration membranes Nylon. J Membr Sci 2006;283:13–20. Kim JH, Ha SY, Nam SY, Rhim JW, Baek KH, Lee YM. Selective permeation of CO2 through pore-filled polyacrylonitrile membrane with poly(ethylene glycol). J Membr Sci 2001;186:97–107. Ulbricht M, Richau K, Kamusewitz H. Chemically and morphologically defined ultrafiltration membrane surfaces prepared by heterogeneous photo-initiated graft polymerization. Colloids Surf A: Physicochem Eng Aspects 1998;138:353–66. Ulbricht M, Riedel M. Ultrafiltration membrane surfaces with grafted polymer ‘tentacles’: preparation, characterization and application for covalent protein binding. Biomaterials 1998;19:1229–37. Frahn J, Malsch G, Schwarz HH. Generation of a selective layer on polyacrylonitrile membrane supports for separation of aromatic/non-aromatic hydrocarbon mixtures by pervaporation. Macromol Symp 2001;164:269–76. Wenzel A, Yanagishita H, Kitamoto D, Endo A, Haraya K, Nakane T, et al. Effects of preparation condition of photoinduced graft fillingpolymerized membranes on pervaporation performance. J Membr Sci 2000;179:69–77. Frahn J, Malsch G, Schwarz HH. Photo-initiated generation of a selective layer on polyacrylonitrile (PAN) composite membranes. J Mater Proc Technol 2003;143–144:277–80. Frahn J, Malsch G, Matuschewski H, Schedler U, Schwarz HH. Separation of aromatic/aliphatic hydrocarbons by photo-modified poly(acrylonitrile) membranes. J Membr Sci 2004;234:55– 65. Ulbricht M, Matuschewski H, Oechel A, Hicke HG. Photo-induced graft polymerization surface modifications for the preparation of hydrophilic and low-protein-adsorbing ultrafiltration membranes. J Membr Sci 1996;115:31–47. Ulbricht M. Membrane separations using molecularly imprinted polymers. J Chromatogr B 2004;804:113–25. Schneider F, Piletsky S, Piletsky E, Guerreiro A, Ulbricht M. Comparison of thin-layer and bulk MIPs synthesized by photo-initiated in situ crosslinking polymerization from the same reaction mixture. J Appl Polym Sci 2005;98:362–72. Conesa A, Palet C. Molecularly imprinted membranes (MIM) for the enantioseparation of selenoaminoacid compounds. Desalination 2006;200:110–1. Donato L, Figoli A, Drioli E. Novel composite poly(4vinylpyridine)/polypropylene membranes with recognition properties for (s)-naproxen. J Pharm Biomed Anal 2005;37:1003–8. Salam A, Ulbricht M. Effect of surface modification on synthesis of pore-filling polymeric monoliths in microfiltration membranes made from polypropylene and poly(ethylene terephthalate). Macromol Mater Eng 2007;292:310–8. Yang Q, Xu ZK, Dai ZW, Wang JL, Ulbricht M. Surface modification of polypropylene microporous membranes with a novel glycopolymer. Chem Mater 2005;17:3050–8.

98

D. He et al. / Progress in Polymer Science 34 (2009) 62–98

[222] Yang Q, Hu MX, Dai ZW, Tian J, Xu ZK. Fabrication of glycosylated surface on polymer membrane by UV-induced graft polymerization for lectin recognition. Langmuir 2006;22:9345–9. [223] Yang Q, Xu ZK, Hu MX, Li JJ, Wu J. Novel sequence for generating glycopolymer tethered on a membrane surface. Langmuir 2005;21:10717–23. [224] Yu HY, Xu ZK, Lei H, Hu MX, Yang Q. Photoinduced graft polymerization of acrylamide on polypropylene microporous membranes for the improvement of antifouling characteristics in a submerged membrane-bioreactor. Sep Purif Technol 2007;53:119–25. [225] Yusof AHM, Ulbricht M. Polypropylene-based membrane adsorbers via photo-initiated graft copolymerization: optimizing separation performance by preparation conditions. J Membr Sci 2008;311:294–305. [226] Hu MX, Yang Q, Xu ZK. Enhancing the hydrophilicity of polypropylene microporous membranes by the grafting of 2-hydroxyethl methacrylate via a synergistic effect of photoinitiators. J Membr Sci 2006;285:196–205. [227] Ma H, Davis RH, Bowman CN. A novel sequential photoinduced living graft polymerization. Macromolecules 2000;33: 331–5. [228] Ma H, Davis RH, Bowman CN. Principle factors affecting sequential photoinduced graft polymerization. Polymer 2001;42:8333–8. [229] Ma H, Bowman CN, Davis RH. Membrane fouling reduction by backpulsing and surface modification. J Membr Sci 2000;173: 191–200. [230] Yu HY, Xu ZK, Yang Q, Hu MX, Wang SY. Improvement of the antifouling characteristics for polypropylene microporous membranes by the sequential photoindueced graft polymerization of acrylic acid. J Membr Sci 2006;281:658–65. [231] Xu ZK, Dai QW, Wu J, Huang XJ, Yang Q. Covalent attachment of phospholipid analogous polymers to modify a polymeric membrane surface: a novel approach. Langmuir 2004;20:1481–8. [232] He DM, Ulbricht M. Surface-selective photo-grafting on porous polymer membranes via a synergist immobilization method. J Mater Chem 2006;16:1860–8. [233] He DM, Ulbricht M. Synergist immobilization method for photografting: factors affecting surface selectivity. Macromol Chem Phys 2007;208:1582–91. [234] He DM, Ulbricht M. Preparation and characterization of porous anion-exchange membrane adsorbers with high protein binding capacity. J Membr Sci 2008;315:155–63. [235] Shinkai S, Nakaji T, Ogawa T, Shigematsu K, Manabe O. Photoresponsive crown ethers. 2. Photocontrol of ion extraction and ion transport by a bis(crown ether) with a butterfly-like motion. J Am Chem Soc 1981;103:111–5. [236] Kumano A, Niwa O, Kajiyama T, Takayanagi M, Kano K, Shinkai S. Photoinduced ion permeation through ternary composite membrane composed of polymer/liquid crystal/azobenzene-bridged crown ether. Chem Lett 1983;12:1327–30. [237] Shimidzu T, Yoshikawa M. Photo-induced carrier mediated transport of alkali metal salts. J Membr Sci 1983;13:1–13. [238] Kimura K, Sumida M, Okoyama MY. Drastic metal-ion enhancement in photoinduced aggregation of copolymers carrying crown ether and spirobenzopyran moieties. Chem Commun 1997:1417–8. [239] Kimura K, Sakamoto H, Uda RM. Cation complexation, photochromism, and photoresponsive ion-conducting behavior

[240]

[241]

[242]

[243]

[244]

[245]

[246]

[247]

[248]

[249]

[250]

[251]

[252]

[253]

[254]

[255]

of crowned spirobenzopyran vinyl polymers. Macromolecules 2004;37:1871–6. Uda RM, Matsui T, Uoe M, Kimura K. Membrane potential photoresponse of crowned malachite green derivatives affording perfect photoswitching of metal ion complexation. J Inclusion Phenom Macrocyl Chem 2005;51:111–7. Oosaki S, Hayasaki H, Sakurai Y, Yajima S, Kimura K. Photocontrol of ion-sensor performances in neutral-carrier-type ion sensors based on liquid-crystalline membranes. Chem Commun 2005:5226–7. Ishihara K, Hamada N, Kato S, Shinohara I. Photoinduced swelling control of amphiphilic azoaromatic polymer membrane. J Polym Sci: Polym Chem Ed 1984;22:121–8. Ishihara K, Shinohara I, Shinohara I. Photoinduced permeation control of proteins using amphiphilic azoaromatic polymer membrane. J Polym Sci: Polym Lett Ed 1984;22:515–8. Sata T, Matsuo Y, Yamaguchi T, Matsusaki K. Preparation and transport of anion-exchange membranes containing viologen moities as anion-exchange groups in the presence or absence of photoirradiation. J Chem Soc, Faraday Trans 1997;93:2553–60. Sata T, Shimokawa Y, Matsusaki K. Preparation of ion-permeable membranes having azobenzene moiety and their transport properties in electrodialysis. J Membr Sci 2000;171:31–43. Kodzwa MG, Staben ME, Rethwisch DG. Photoresponsive control of ion exchange in leucohydroxide containing hydrogel membranes. J Membr Sci 1999;158:85–92. Ishikawa M, Kitamura N, Masuhara H, Irie M. Size effect on photoinduced volume change of polyacrylamide microgels containing triphenylmethane leucoderivatives. Makromol Chem, Rapid Commun 1991;12:587–690. Chung DJ, Ito Y, Imanishi Y. Preparation of porous membranes grafted with poly(spiropyran-containing methacrylate) and photocontrol of permeability. J Appl Polym Sci 1994;51:2027–33. Kimura K, Sakamoto H, Nakamura T. Application of photoresponsive polymers carrying crown ether and spirobenzopyran side chains to photochemical valve. J Nanosci Nanotechnol 2006;6:1741–9. Sumaru K, Ohi K, Takagi T, Kanamori T, Shinbo T. Photoresponsive properties of poly(N-isopropylacrylamide) hydrogel partly modified with spirobenzopyran. Langmuir 2006;22:4353–6. Kameda M, Sumaru K, Kanamori T, Shinbo T. Photoresponse gas permability of azobenzene-functionalized glassy polymer films. J Appl Polym Sci 2003;88:2068–72. Liu N, Dunphy DR, Atanassov P, Bunge SD, Chen Z, Lopez GP, et al. Photoregulation of mass transport through a photoresponsive azobenzene-modified nanoporous membrane. Nano Lett 2004;4:551–4. Weh K, Noack M, Hoffman K, Schröder KP, Caro J. Change of gas permeability by photoinduced switching of zeolit-azobenzene membranes of type MFI and FAU. Micropor Macropor Mater 2002;54:15–26. Aoyama M, Watanabe J, Inoue S. Photoregulation of permeability across a membrane from a graft copolymer containing a photoresponsive polypeptide branch. J Am Chem Soc 1990;112:5542–5. Lopez LC, Buonomenna MG, Fontananova E, Iacoviello G, Drioli E, d’Agostino R, et al. A new generation of catalytic poly(vinylidene fluoride) membranes: coupling plasma treatment with chemical immobilization of tungsten-based catalysts. Adv Funct Mater 2006;16:1417–24.