Article pubs.acs.org/Macromolecules
3D Surface Functionalization of Emulsion-Templated Polymeric Foams Priyalakshmi Viswanathan,†,‡ David W. Johnson,§ Claire Hurley,∥ Neil R. Cameron,§,# and Giuseppe Battaglia*,⊥ †
The Krebs Institute and ‡Department of Biomedical Science, The University of Sheffield, Firth Court, Western Bank, Sheffield S10 2TN, U.K. § Department of Chemistry, University of Durham, South Road, Durham DH1 3LE, U.K. ∥ Sheffield Surface Analysis Centre, Kroto Research Institute, The University of Sheffield, Sheffield S3 7HQ, U.K. ⊥ Department of Chemistry, University College London, London WC1H 0AJ, U.K. S Supporting Information *
ABSTRACT: We describe the preparation of porous polymeric scaffolds via polymerization of the oil phase in high internal phase water-in-oil-emulsions using amphiphilic block copolymers polystyrene-b-poly(ethylene oxide), polystyrene-b-poly(acrylic acid), poly(1,4-butadiene)-b-poly(ethylene oxide), and poly(1,4-butadiene)-b-poly(acrylic acid) as surfactants. We show that the block copolymers anchor to the polymerized oil phase via the lipophilic block, which can occur by chemical and/or physical entanglement and consequent presentation of the hydrophilic block on the pore surfaces. The in situ polymerization enables the full surface functionalization of the porous materials with the final surface chemistry dictated by the hydrophilic block. Furthermore, the foam physical architecture may be tailored through controlling emulsion parameters such as the initiator, shear rate, and aqueous phase volume fraction.
■
INTRODUCTION High internal phase emulsion (HIPE) templating is a popular method for the preparation of highly porous polymeric matrices.1 Such emulsions are defined by high internal or droplet phase volume fraction (Φw = 0.74). This volume fraction represents the maximum volume of monodisperse, nondeformable spheres. The polymerization of the continuous oil phase in water-in-oil emulsions and the evaporation of the droplet phase result in highly porous polymers or polyHIPEs. By far, the most widely studied w/o HIPE system is that of styrene and divinylbenzene as the oil phase monomers and sorbitan monooleate (Span 80) as the surfactant. The control of the physical architecture of these foams such as void and interconnect size has been well documented.2−4 Here, a void describes the pores of the foam, resulting from the evaporation of the aqueous phase droplet, and an interconnect refers to the interconnecting pores between two adjacent droplets. PolyHIPE foams have been used extensively in applications such as tissue engineering,5,6 water purification, solid phase supports for catalysis,7−9 and bioconjugation10 or membranes for separation processes11 that require highly porous, interconnected foams that exhibit surface functionality. The combination of high porosity and surface functionalization also makes emulsion templating advantageous for applications such as electrical conductors12 or exhibits magnetic properties.13 Yet, there are only few studies that explore control over surface chemistry. Often this is very much the result of the oil phase polymerization, and therefore the resulting materials are © XXXX American Chemical Society
intrinsically hydrophobic. The surface chemistry can be further controlled by postpolymerization processing such as chemical functionalization and plasma polymerization.14 However, this approach is limited by depth penetration, where only the top layers of the foam may be effectively functionalized. Other strategies involve polymer grafting using Huisgen type “click” chemistry15 or more recently inverse electron demand Diels− Alder reactions.16 Livshin et al. have demonstrated the generation of hydrophilic foams by the hydrolysis of hydrophobic tert-butyl acrylate polyHIPEs.17 However, all of these cases require multiple functionalization steps, which reduce control over 3D functionalization.4,18 Herein we propose a bottom-up approach where the same surfactant that stabilizes the water droplets within the oil phase can be used to control the solid foam surface chemistry. Block copolymer surfactants have been used widely as emulsion stabilizers,19,20 however, it has not been demonstrated until recently5 that amphiphilic block copolymer surfactants utilized in water-in-oil HIPEs leads to surface functionalization of the resulting three-dimensional (3D) matrix in a one-step process. We extend this study using block copolymers that comprise a range of molecular weights, hydrophile to hydrophobe ratios and chemistries to demonstrate the universality of surface functionalization using this method. We predict that the hydrophobic Received: May 12, 2014 Revised: September 4, 2014
A
dx.doi.org/10.1021/ma500968q | Macromolecules XXXX, XXX, XXX−XXX
Macromolecules
Article
block of macromolecular surfactants has the unique ability to participate in the free radical polymerization or in physical entanglement (Figure 1) with the underlying polymer matrix.
Table 1. Low and High Molecular Weight Surfactants and Their Hydrophile−Lipophile Balance surfactant
Mw (g mol−1)
HLB
Span 80 PS−PAA PBD−PAA PS−PEO PBD−PEO
483 18700 3000 25000 14300
4.3 3.95a 7.3a 0.8 3.87
a Estimation of HLB values for PS−PAA and PBD−PAA since the empirical formula for HLB is primarily for nonionic surfactants.
■
EXPERIMENTAL SECTION
Foam Preparation. The monomers styrene (Sigma-Aldrich) and divinylbenzene (80% technical grade, Sigma Adlrich) were passed through a column of activated basic alumina (Brockmann Activity I, Sigma-Aldrich) to remove the inhibitor p-tert-butylcatechol prior to use. The initiators K2S2O8 (Sigma-Aldrich) and AIBN (Fisher Scientific), tetrahydrofuran (Sigma-Aldrich), block copolymers poly(1,4-butadiene)-b-poly(ethylene oxide) (PBD−PEO) and polystyrene-b-poly(acrylic acid) (PS−PAA) (Polymer Source Inc., Montreal), polystyreneb-poly(ethylene oxide) (PS−PEO) (Sigma-Aldrich), and the surfactant Span 80 (Sigma-Aldrich) were all used as received; all the polymers have a reported polydispersity index (PDI) of 1.1−1.3. Poly(1,4-butadiene)b-poly(acrylic acid) (PBD−PAA, PDI = 1.2) was synthesized as previously reported.28 The high internal phase emulsions were prepared as follows: The surfactant was first solubilized in the oil phase. Because of the poor solubility of PS−PAA in styrene/divinylbenzene, THF (10 μL/mg) was used to make a solution of PS−PAA before its addition to the oil phase. The surfactant concentration was maintained at 0.01 mol % (relative to the monomer) for all copolymers. For concentrationdependent studies, copolymer concentrations were reduced from 0.02 mol % until phase separation occurred. The aqueous phase was added manually to the oil phase using a 5 mL syringe. The oil phase was kept stirring in a 50 mL beaker, using either an overhead mechanical stirrer with a 2 cm stainless steel paddle at 750 rpm or an UltraTurrax homogenizer at 11 500 rpm. Once the aqueous phase was added, the resulting viscous white emulsions were stirred for a further 5 min to homogenize prior to polymerization. For aqueous phase initiation K2S2O8 (0.1% w/v) was prepared as a solution in DI water and used immediately. For oil phase initiation, AIBN (1% w/w of the oil phase) was added to the oil phase and stirred immediately before the addition of DI water. Emulsions were polymerized in a convection oven at 50 °C for 24 h, resulting in porous monolithic foams. The foams were then Soxhlet extracted in isopropyl alcohol for 24 h to remove all unreacted monomers. All emulsions, with the exception of those stabilized by of PS−PEO, consisted of either 80% or 90% aqueous phase volume and consisted of an oil phase of 90% w/w styrene and 10% w/w divinylbenzene. Emulsions using PS−PEO as the surfactant, on the other hand, were prepared with an aqueous phase of 80% and an oil phase consisting of only DVB. Formulations consisting of higher aqueous volume fractions or the addition of styrene in the oil phase resulted in unstable emulsions at the polymerization temperature. Foams using the low molecular weight surfactant Span 80 were prepared according to previously reported methods.29,30 Foam Characterization. Sectioned foams were prepared for scanning electron microscopy by placing them on an aluminum stub with an adhesive carbon pad. Samples were coated with gold (approximately 15 nm) using an Edwards S150B sputter coater. Foams were imaged using a Philips XL20 scanning electron microscope at an accelerating voltage of 10 kV and spot size of 3.0 nm. Void diameters were estimated by image analysis using ImageJ. Roughly 100 voids were measured for each emulsion composition from various micrographs to obtain an average porosity of the foams. A statistical correction, previously described,21 was used to gain a more accurate measurement of void diameters. X-ray photoelectron spectroscopy (XPS) was carried out on a Kratos Axis Ultra X-ray photoelectron spectrometer with a monochromated Al
Figure 1. Mechanism of polyHIPE surface functionalization. (a) Optical micrograph of a HIPE, when polymerized forms (b) a highly porous polystyrene/divinylbenzene foam as shown in the scanning electron micrograph. (c) HIPEs stabilized by amphiphilic A−B type block copolymers as surfactants can be surface functionalized through physical or chemical entanglement compared to low molecular weight surfactants.
This allows for the efficient presentation of the hydrophilic block on the foam surface in 3D and thus provides a de novo method of surface modification of such foams suitable to various applications.21,22 We further explore the effects of common emulsion processing parameters such as the effect of radical initiation, shear rate, and aqueous phase volume fraction on the final polyHIPE structure and surface functionality. It has been shown by Silverstein and co-workers that the choice of radical initiator in Pickering polyHIPEs controls the kinetics of droplet coalescence and pore interconnectivity.23−25 We therefore investigated the effect of block copolymer stabilized polyHIPEs on pore interconnectivity and, importantly, its effect on foam surface modification. In our study, we employ the block copolymers polystyrene-bpoly(ethylene oxide) (PS−PEO), poly(1,4-butadiene)-b-poly(ethylene oxide) (PBD−PEO), poly(1,4-butadiene)-b-poly(acrylic acid) (PBD−PAA), and polystyrene-b-poly(acrylic acid) (PS−PAA) as the surfactants for water-in-oil emulsions to produce polystyrene/divinylbenzene foams. The hydrophilic−lipophilic balance (HLB) is often applied when choosing the appropriate surfactant and was used in our work as an indicator of effectiveness of block copolymer surfactants. The empirical formula originally defined by Griffin as 20(Mh/Mw), where Mh is the molecular weight of the hydrophilic block and Mw is the molecular weight of the surfactant, is based on nonionic surfactants.26 Generally, surfactants of HLB values between 2 and 6 are used to stabilize of w/o emulsions.27 HLB values for block copolymer surfactants used here (Table 1) were estimated using this formula. Thus, we demonstrate the relevance of choosing the appropriate block copolymer surfactant, based on this empirical rule, on the final structure of the polyHIPE foam. B
dx.doi.org/10.1021/ma500968q | Macromolecules XXXX, XXX, XXX−XXX
Macromolecules
Article
Figure 2. Morphologies of polyHIPEs using the block copolymer surfactants poly(1,4-butadiene)-b-poly(ethylene oxide), polstyrene-b-poly(acrylic acid), polystyrene-b-poly(ethylene oxide), and poly(1,4-butadiene)-b-poly(acrylic acid). Scanning electron micrographs of the foams prepared with the initiators K2S2O8 (a−d) and AIBN (e−h) with high-magnification micrographs indicating closed and open porosity. Foams were made at low shear and an aqueous phase volume fraction, Φw, of 0.9 for all foams apart from those functionalized with PS−PEO where Φw is 0.8.
■
Kα X-ray source. The spot size was ca. 700 × 300 μm. Sectioned samples of the foams were mounted on the sample bar with double-sided tape. Survey spectra (pass energy 160 eV) and O 1s and C 1s high-resolution spectra (pass energy 20 eV) were acquired on three areas within each foam. Binding energy scales were calibrated using the C 1s signal as a reference (285 eV). Analysis was carried out using CasaXPS software. Peaks were fitted using combinations of Gaussian (30%) and Lorentzian (70%) curves against a linear background. The fwhm of all components was constrained to be equal to that of the aliphatic carbon with the exception of the shakeup satellite at ca. 291.7 eV; other parameters were left free to vary during fitting. Static water contact angles were measured on each foam surface using a Ramé-Hart goniometer. 4 μL of Milli-Q was placed on the foam sample and allowed to settle for 30 min, and a static contact angle was measured. As these are highly porous materials, contact angle measurements here do not serve to measure surface tension but rather to give an indication of surface wettability. At least three measurements were made for each sample. Mercury intrusion porosimetry analysis was performed using a Micromeritics AutoPore IV. Intrusion and extrusion mercury contact angles of 130° were used. Penetrometers with a stem volume of 1.836 mL and a bulb volume of 4.25 mL were used. The intrusion volumes were between 25% and 90% of the stem volume. Intrusion pressures for the polyHIPE did not exceed 11.03 MPa.
RESULTS AND DISCUSSION
Oil-Soluble versus Water-Soluble Initiator Effects. We limit our discussion to one specific concentration used (0.01 mol %, see Experimental Section for further details) with the exception of determining the effects of surfactant concentration, described in further detail below. By far the dominant parameter controlling the final desired morphology of the foams was found to be the choice of initiator. Traditionally, water-soluble initiators have been used in conjunction with low molecular weight surfactants such as Span 80. Seminal work by Williams et al.31,32 explored the effects of water- and oil-soluble initiators and found that the water-soluble initiator K2S2O8 resulted in mechanically stiffer foams with void sizes an order of magnitude lower than foams prepared with AIBN owing to electrolyte stability of the oil−water interface (which holds true for the case of nonionic surfactants). As such, K2S2O8 has been the primary choice of initiator for water-in-oil emulsions. However, we observe that AIBN was a more appropriate choice for the copolymer surfactants as open porosity and foam interconnectivity were dictated by the choice of initiator for all copolymers utilized (Figure 2e−h). We note the lack or presence of foam interconnectivity determined from the SEM images results from the choice of the initiator. HIPEs initiated by AIBN are herein referred to as open porous foams while those initiated by K2S2O8 are referred to as closed porous foams and implies the C
dx.doi.org/10.1021/ma500968q | Macromolecules XXXX, XXX, XXX−XXX
Macromolecules
Article
level of interconnectivity observed within foams. We note that there is still sufficient interconnectivity in closed porous foams to remove the internal phase and allow Soxhlet extraction. The formation of interconnecting pores using styrene/divinylbenzene/Span80/K2S2O8 formulations have been well studied and show that interconnectivity arises at the gel point of the emulsion,33 where the appearance of interconnecting windows is due to the thinning of the oil film between droplets that rupture upon conversion of monomer to polymer (as a result of increase in density). However, this phenomenon is not seen with block copolymer surfactants when K2S2O8 was used as the initiator, which resulted in foams with closed porosity (Figure 2a−d) as well as the formation of latex particles due to inverse oil-in-water polymerization (Supporting Information Figure 2). The exception was PS−PAA where interconnecting holes appear at 90% aqueous phase volume and at high shear (Supporting Information Figure 1). In this case, an increase in average void diameter of the foams was observed despite the formation of kinetically unstable precursor HIPEs. It is possible that the polymerization of the emulsions was rapid enough to prevent macroscopic phase separation. Additionally, during polymerization, the increase in temperature promotes droplet coalescence and facilitates interconnect formation at the point of contact between two adjacent droplets. In contrast, formation of interconnecting pores was seen with all block copolymers used, when the oilsoluble initiator AIBN was employed and was independent of surfactant concentration, shear rate, aqueous phase volume, or even the HLB of the copolymer. However, low molecular weight surfactant stabilized HIPEs have also been shown to exhibit initiator dependent porosity in bicontinuous hydrogel34 and crystalline35 polyHIPEs but not in styrene/divinylbenzene-based foams. We thus propose that the locus of initiation of polymerization determines open porosity. For K2S2O8, initiation occurs at the oil−water interface and proceeds toward the bulk whereas for AIBN the opposite is true. Therefore, when block copolymer surfactants are used, initiation at the oil−water interface results in a rigid pore wall, thereby reducing mobility of two adjacent water droplets to coalesce and form interconnects. On the contrary, when initiation originates from the bulk oil phase (with AIBN), the oil−water interface remains “fluid” until the gel point allowing droplets to fuse and form interconnecting pores. For Span 80-based foams, however, their low molecular weight nature enables interfacial fluidity at the polymerization temperature regardless of the initiator employed forming interconnected porous foams with both K2S2O8 and AIBN. Recent studies by Silverstein and co-workers have also shown23−25 that droplet connectivity is also dependent on the initiator used in Pickering HIPEs. To further investigate initiator-dependent open and closed porosity, mercury intrusion porosimentry analysis was performed on all foam samples. All foams initiated with AIBN exhibited porosities of ∼90% apart from PBD−PAA, which exhibited a porosity of 65%. Average interconnect diameters plotted for all foams ranged from 5.8 to 11.9 μm (Figure 3 bottom, Supporting Information Table 2), indicating a high level of interconnectivity. Closed-porous foams using K 2 S 2 O 8 initiation, on the other hand, showed little interconnectivity, in particular for PS−PEO and PBD−PEO samples that exhibited porosities of 80% and 74%, respectively, with interconnect diameters
