Synthesis and Characterization of Shell Cross-Linked Micelles with

ACS eBooks; C&EN Global Enterprise .... produced SCL micelles that remained intact at 5 °C, i.e., below the cloud point of the core-forming PPO block...
1 downloads 0 Views 312KB Size
3808

Langmuir 2005, 21, 3808-3813

Synthesis and Characterization of Shell Cross-Linked Micelles with Hydroxy-Functional Coronas: A Pragmatic Alternative to Dendrimers? Laura N. Pilon,† Steven P. Armes,*,† Paul Findlay,‡ and Steven P. Rannard‡ Department of Chemistry, University of Sheffield, Brook Hill, Sheffield S3 7HF, United Kingdom, and Unilever Research & Development, Port Sunlight Laboratory, Quarry Road East, Bebington, Wirral, Merseyside CH63 3JW, United Kingdom Received December 1, 2004. In Final Form: February 8, 2005 Shell cross-linked (SCL) micelles with hydroxy-functional coronas have been constructed in aqueous solution by exploiting the micellar self-assembly behavior of a new thermoresponsive ABC triblock copolymer. This copolymer was prepared via atom transfer radical polymerization in a convenient one-pot synthesis and comprised a thermoresponsive core-forming poly(propylene oxide) (PPO) block, a cross-linkable central poly(2-(dimethylamino)ethyl methacrylate) (DMA) block, and a hydroxy-functional outer block based on poly(glycerol monomethacrylate) (GMA). DMF GPC analysis confirmed a unimodal molecular weight distribution for the PPO-PDMA-PGMA triblock copolymer precursor, with an Mn of 12 100 and a polydispersity of approximately 1.26. This copolymer dissolved molecularly in aqueous solution at 5 °C but formed micelles with hydroxy-functional coronas above a critical micelle temperature of around 12 °C, which corresponded closely to the cloud point of the PPO macroinitiator. Cross-linking of the DMA residues using 1,2-bis(2-iodoethoxy)ethane produced SCL micelles that remained intact at 5 °C, i.e., below the cloud point of the core-forming PPO block. Dynamic light scattering studies confirmed that the SCL micelle diameter could be varied depending on the temperature employed for cross-linking: smaller, more compact SCL micelles were formed at higher temperatures, as expected. Since cross-linking involved quaternization of the DMA residues, the SCL micelles acquired cationic surface charge as judged by aqueous electrophoresis studies. These cationic SCL micelles were adsorbed onto near-monodisperse anionic silica sols, which were used as a model colloidal substrate. Thermogravimetric analyses indicated a SCL micelle mass loading of 2.5-4.4%, depending on the silica sol diameter and the initial micelle concentration. Aqueous electrophoresis measurements confirmed that surface charge reversal occurred after adsorption of the SCL micelles, and scanning electron microscopy studies revealed a uniform coating of SCL micelles on the silica particles.

Introduction The covalent stabilization of micelles and vesicles has attracted increasing attention in recent years.1-4 In particular, shell cross-linked (SCL) micelles are potentially useful nanosized delivery vehicles for the delivery of various actives (drugs, fragrances, pesticides, etc.).1-3 The first example of SCL micelles was described by Wooley and co-workers, who oligomerized the pendent styrene groups of an AB diblock copolymer using radical chemistry in a water/THF mixture.1a More recently, it has been * To whom correspondence should be addressed. † University of Sheffield. ‡ Unilever Research & Development. (1) (a) Thurmond, K. B.; Kowalewski, T.; Wooley, K. L. J. Am. Chem. Soc. 1996, 118, 7239. (b) Huang, H.; Kowalewski, T.; Remsen, E. E.; Gertzmann, R.; Wooley, K. L. J. Am. Chem. Soc. 1997, 119, 11653. (c) Huang, H.; Remsen, E. E.; Kowalewski, T.; Wooley, K. L. J. Am. Chem. Soc. 1999, 121, 3805. (d) Zhang, Q.; Remsen, E. E.; Wooley, K. L. J. Am. Chem. Soc. 2000, 122, 3642. (e) Wooley, K. L. J. Polym. Sci., Polym. Chem. 2000, 38, 1397. (2) (a) Bu¨tu¨n, V.; Billingham, N. C.; Armes, S. P. J. Am. Chem. Soc. 1998, 120, 12135. (b) Bu¨tu¨n, V.; Lowe, A. B.; Billingham, N. C.; Armes, S. P. J. Am. Chem. Soc. 1999, 121, 4288. (c) Bu¨tu¨n, V.; Wang, X.-S.; de Paz Banez, M. V.; Robinson, K. L.; Billingham, N. C.; Armes, S. P.; Tuzar, Z. Macromolecules 2000, 33, 1. (d) Liu, S.; Tang, Y. Q.; Weaver, J. V. M.; Billingham, N. C.; Armes, S. P.; Tribe, K. Macromolecules 2002, 35, 6121. (e) Joralemon, M. J.; Murthy, K. S.; Remsen, E. E.; Becker, M. L.; Wooley, K. L. Biomacromolecules 2004, 5, 903. (3) (a) Stewart, S.; Liu, G. J. Chem. Mater. 1999, 11, 4. (b) Yan, X. H.; Liu, F. T.; Li, Z.; Liu, G. J. Macromolecules 2001, 34, 9112. (c) Sanji, T.; Nakatsuka, Y.; Kitayama, F.; Sakurai, S. Chem. Commun. 1999, 2201. (4) (a) Stupp, S. I.; LeBonheur, V.; Walker, K.; Li, L. S.; Huggins, K. E.; Keser, M.; Amstutz, A. Science 1997, 276, 384. (b) Sauer, M.; Meier, W. Chem. Commun. 2001, 55.

shown that ABC triblock copolymers offer significant advantages over AB diblock copolymers for the preparation of SCL micelles. Since coronal chain cross-linking occurs during micelle-micelle collisions, AB diblock copolymers can only be used in highly dilute solution (0.5%) if intermicelle fusion is to be avoided. In contrast, SCL micelles have been prepared at up to 10% solids by cross-linking the central B block of an ABC triblock copolymer. Minimal micelle aggregation occurs under these conditions because the outer A block chains confer steric stabilization.2c,d Several research groups have prepared so-called “nanocages” via chemical degradation of the core-forming block of SCL micelles. However, this approach is not atomefficient, since a significant proportion of the original SCL micelle is discarded. An alternative approach is the use of stimulus-responsive block copolymers to produce coreresponsive SCL micelles. Both pH-responsive and thermoresponsive micelle cores have been developed in our laboratory.1d,e,3 For example, SCL micelles with pHresponsive cores have been exploited as nanoreactors for the in situ reduction of Au(III) salts to produce gold nanoparticles.5 Wooley’s group1a,e has pointed out several similarities between SCL micelles and dendrimers. In principle, both approaches produce nanosized, globular structures with controllable functionality that are potentially capable of the uptake and release of active compounds. Although structurally more well-defined than SCL micelles, den(5) Liu, S.; Save, S.; Weaver, J. V. M.; Armes, S. P. Langmuir 2002, 18, 8350.

10.1021/la047046g CCC: $30.25 © 2005 American Chemical Society Published on Web 03/17/2005

SCL Micelles with Hydroxy-Functional Coronas

drimers are generally much more difficult to synthesize and purify. In contrast, we have recently shown that SCL micelles can be conveniently prepared in a one-pot synthesis at high solids using atom transfer radical polymerization (ATRP) without recourse to organic solvents.6 However, one of the main advantages of dendrimers is the very large number of surface functional groups (amino, hydroxyl, carboxylic acid, etc.). As far as we are aware, there have been few reports of the synthesis of SCL micelles with functional (i.e. reactive) coronal chains.2b,e Herein we report the preparation of cationic SCL micelles with thermoresponsive cores and hydroxylfunctional coronas using an ABC triblock copolymer. We show that these new SCL micelles can be used to modify surfaces: electrostatic adsorption onto a model anionic colloidal substrate (in this case a near-monodisperse silica sol) leads to relatively uniform surface coverage and surface charge reversal. Experimental Section Materials. Monohydroxy-terminated poly(propylene oxide) (PPO-OH; mean degree of polymerization ) 33; Mn ∼ 2,000) was donated by Cognis Performance Chemicals (Hythe, U.K.) and was esterified according to a literature method6 to produce the PPO33-Br ATRP macroinitiator. 2-(Dimethylamino)ethyl methacrylate (DMA, 98%) was obtained from Aldrich and treated with basic alumina to remove inhibitor prior to use. Glycerol monomethacrylate (GMA) was obtained from Ro¨hm (Germany) and was used as received. Methanol (99.9+%, HPLC grade) was obtained from Acros. 1,2-Bis(2-iodoethoxy)ethane (BIEE, 96%), 1,1,4,7,10,10-hexamethyltriethylenetetraamine (HMTETA, 97%), and copper(I) chloride (CuCl, 99+%) were obtained from Aldrich and were used as received. All monomers and solvents (HPLC grade) were degassed prior to polymerization by nitrogen purge for 10 min. Monospher 250 and 1000 silicas were obtained from E. Merck (Germany). The serial numbers indicate the approximate mean particle diameters of these sols in nanometers. The particle density was determined to be 2.06 g cm-3 for the 1000 nm sol and 2.02 g cm-3 for the 250 nm sol by helium pycnometry. Preliminary scanning electron microscopy (SEM) studies revealed surface contamination by unidentified organic material. Hence, the silica sols were purified by ultrasonication in isopropyl alcohol for 30 min, followed by two centrifugation-redispersion cycles (using doubly distilled, deionized water to replace the successive supernatants), and finally dried under vacuum at 20 °C before use. Subsequent SEM inspection confirmed that the surface contaminant(s) were successfully removed by this protocol. Triblock Copolymer Synthesis. The PPO33-PDMA20PGMA50 triblock copolymer was synthesized by ATRP in a onepot procedure. DMA (2.00 g; 0.013 mol) was polymerized in the bulk at 60 °C using a CuCl catalyst (0.063 g; 6.3 mmol) and a HMTETA ligand (0.18 mL; 6.3 mmol; HMTETA:CuCl molar ratio ) 1.0) in conjunction with a PPO33-Br macroinitiator (1.34 g; 6.3 mmol). After 98% conversion of the DMA block (as judged by 1H NMR), the crude, unpurified PPO33-PDMA20 diblock copolymer precursor was dissolved in deoxygenated methanol for subsequent polymerization of GMA (5.09 g; 0.033 mol; 50 w/v % solids) at 60 °C in the same reaction vessel. On exposure to air the green Cu(I) catalyst turned blue, indicating aerial oxidation to produce an inactive Cu(II) complex. The resulting methanolic copolymer solution was treated with silica to remove this spent catalyst, water was added, and the alcohol was removed under reduced pressure. This aqueous solution was then dialyzed (the dialysis tubing molecular weight cutoff was 3500) for 24 h against doubly distilled deionized water to remove impurities (e.g. copper catalyst and traces of unreacted monomer) and was then freeze-dried overnight. Triblock Copolymer Characterization. All 1H NMR spectra were recorded in either D2O and/or pyridine-d5 using a 300 MHz Bruker Avance spectrometer. Duplicate CHN microanalyses were obtained from MEDAC analytical services (6) Liu, S.; Armes, S. P. J. Am. Chem. Soc. 2001, 123, 9910.

Langmuir, Vol. 21, No. 9, 2005 3809 (Brunel University, U.K.). Dynamic light scattering (DLS) studies were carried out on 0.50 w/v % aqueous solutions at approximately pH 8 using a Brookhaven BI-9000AT instrument equipped with a 125 mW laser operating at a λ of 532 nm. The scattered light was detected at 90°, and intensity-average particle diameters were determined from the diffusion coefficients using the StokesEinstein equation for monodisperse, noninteracting hard spheres. The cell was equipped with a thermostated water bath, and measurements were recorded at various temperatures between 5 and 70 °C. Turbidimetry was used to determine the critical micellization temperature (CMT) of the triblock copolymer. The absorbance of a 1 cm cell containing a 0.50% aqueous copolymer solution at approximately pH 8 was measured at 500 nm using a Perkin-Elmer Lambda 2S UV/visible spectrophotometer. The temperature was controlled with a thermostated water bath, and the temperature in the cell was monitored using an RS digital thermometer. Polymer molecular weights and polydispersities were assessed by GPC using either DMF or THF eluent. DMF GPC analyses were performed on a Viscotek system consisting of 2 PLgel 10 µm mixed-B columns and a refractive index detector, with 0.01 M LiBr being added to the eluent. For THF GPC analyses, the GMA residues were protected with benzoic anhydride in pyridine solution to render them soluble in THF as described by Save et al.7 The THF eluent contained 2% triethylamine to minimize polymer-column interactions, and measurements were performed on a Viscotek system consisting of a guard column, a PLgel 5 µm A column, and a PLgel 5 µm mixed-D column using a refractive index detector. A series of near-monodisperse poly(methyl methacrylate) calibration standards were used for both GPC protocols. Shell Cross-Linked Micelle Syntheses. A series of aqueous solutions of the PPO33-PDMA20-PGMA50 triblock were prepared by molecularly dissolving this copolymer at 5 °C in doubly distilled, deionized water at up to 10% solids. Micelle formation was induced by heating these solutions to 40 °C at pH 8 prior to the addition of the BIEE cross-linker (25 or 50 mol % with respect to DMA units, corresponding to a target degree of quaternization of 50 or 100%). This aqueous solution was stirred at 40 °C for 72 h as described previously to ensure effective covalent stabilization of the PPO-core micelles.6 Shell Cross-Linked Micelle Characterization. The particle size and morphology of the SCL micelles was examined by transmission electron microscopy (TEM). Samples were prepared on Agar Scientific carbon film 200 mesh copper grids by dipping the grid into a 0.10% solution of the SCL micelles in doubly distilled deionized water and drying under reduced pressure at 20 °C overnight. No staining was used and images were obtained on a Hitachi-7100 TEM with digital image acquisition. 1H NMR spectra were recorded on 0.50% D2O solutions of SCL micelles at 5 and 50 °C using a 300 MHz Bruker Avance spectrometer. Dynamic light scattering (DLS) studies were carried out at 5 °C using the same experimental setup described earlier to verify that cross-linking had occurred and also to observe the variation of micelle diameter with temperature for both the non-crosslinked micelles and the final SCL micelles. Adsorption of Shell Cross-Linked Micelles onto Silica Particles. The 2.0 w/v % dispersions of the purified Monospher 250 and 1000 silica sols were prepared using doubly distilled, deionized water using NaOH to adjust the solution pH to pH 9. SCL micelles were prepared at 5.0 w/v % solids and pH 9 as described above, but in this case a degree of quaternization of 100% was targeted (i.e. 50 mol % BIEE was added with respect to the DMA residues) to maximize the cationic character of the micelles. These SCL micelle solutions were then mixed with the silica sols and stirred for 48 h at 20 °C so as to allow electrostatic adsorption to occur. The relative proportions were such that 400 mg SCL micelles were used per m2 of silica sol; thus, the SCL micelles were employed in significant excess relative to the available silica surface area. The purification procedure involved centrifugation of the silica particles (at 7000-8500 rpm for 30 min), followed by decantation of the micelle solution and redispersion of the sedimented silica particles in doubly distilled, deionized water. This removed the excess, nonadsorbed micelles. (7) Save, M.; Weaver, J. V. M.; Armes, S. P.; McKenna, P. Macromolecules 2002, 35, 1152.

3810

Langmuir, Vol. 21, No. 9, 2005

Pilon et al.

Figure 2. GPC curves for the PPO33-Br macroinitiator, the PPO33-DMA20 diblock precursor, and the final PPO33-DMA20GMA50 triblock copolymer.

Figure 1. Reaction scheme for the ATRP synthesis of the PPO33-DMA20-GMA50 triblock copolymer via sequential monomer addition using a PPO-based macroinitiator. The SCL micelle-coated silica sol was again centrifuged and split into two fractions. One fraction was redispersed for aqueous electrophoresis measurements (see below), and the second fraction was dried under vacuum at 20 °C prior to examination by SEM, thermogravimetric analysis, and infrared spectroscopy (see below). Thermogravimetric analyses were performed on a series of samples using a Perkin-Elmer TGA-7 instrument in conjunction with a Perkin-Elmer TAC 7/3 instrument controller. Samples were heated to 900 °C at a rate of 20 °C/min in air. The purified Monospher silicas were also examined to determine the mass loss due to water evaporation and surface silanol dehydration. The Monospher 250 and Monospher 1000 sols had mass losses of 7.6% and 6.2%, respectively. These values were subtracted from the mass loss determined for a SCL micelle-coated silica sol so as to estimate the mass loss due to copolymer pyrolysis. Aqueous electrophoresis measurements were carried out to determine the ζ potentials of the Monospher 250 silica sol, the SCL micelles, and the SCL micelle-coated Monospher 250 silica sol. A Malvern Zetamaster instrument was used to determine the ζ potential of dilute aqueous solutions (approximately 0.10.5%) with a background salt concentration of 0.01 M NaCl from pH 2 to pH 10. SEM studies were performed on the Monospher 1000 silica particles before and after the SCL micelle adsorption experiments using a Leo Stereoscan 420 instrument. All samples were sputtercoated with gold prior to examination to prevent sample charging. Diffuse reflectance infrared Fourier transform spectroscopy (DRIFTS) spectra were recorded for the bare silica sol, the SCL micelle-coated silica sol, and the non-cross-linked PPO33PDMA20-PGMA50 triblock copolymer precursor (100 scans/ spectrum; 4 cm-1 resolution) using a Perkin-Elmer Spectrum One spectrometer fitted with a Perkin-Elmer DRIFTS accessory.

Results and Discussion Triblock Copolymer Synthesis. Various reaction conditions were examined for the optimum synthesis of the PPO33-PDMA20-PGMA50 triblock copolymer, and a one-pot method produced the best results (Figure 1). In the first stage of the polymerization, DMA was polymerized in the bulk at 60 °C using a CuCl/HMTETA catalyst, followed by addition of the GMA monomer as a 50% solution in methanol after approximately 98% conversion of the DMA (as determined by prior kinetic analysis). If

the intermediate PPO33-PDMA20 diblock macroinitiator was isolated and purified prior to GMA polymerization, GPC analysis indicated a significant proportion of dead PPO33-PDMA20 chains contaminating the final triblock copolymer. If bpy was used as the ligand, the polydispersity of the PPO33-PDMA20 diblock copolymer was significantly higher (Mw/Mn ) 1.80 vs Mw/Mn e 1.25) and Mn values were also higher than expected, indicating reduced initiator efficiency. The use of isopropyl alcohol as a solvent for the DMA polymerization was also investigated; methanol was considered unsuitable for this purpose due to the facile transesterification of the DMA monomer that can occur in this solvent.8 The use of 9:1 IPA/water as a solvent for the polymerization of DMA from PPO33-Br also led to higher polydispersities (typically Mw/Mn ) 1.35) and molecular weights than expected. Thus, eventually we elected to examine the polymerization of DMA in the bulk. Kinetic analysis of the bulk polymerization yielded an approximately linear semilogarithmic plot initially, but an increase in rate was observed above 70% conversion. This was attributed to poor dispersion of the partially heterogeneous catalyst within the highly viscous polymerizing reaction solution. Presumably, the deactivating CuIIX species was less soluble than the CuIX species, leading to the observed increase in rate. The polydispersity of the growing PPO-PDMA diblock copolymer remained fairly constant throughout the polymerization, and its Mn increased linearly with conversion, albeit slightly higher than the theoretical value. This was attributed to inefficient initiation by the PPO33-Br macroinitiator under these conditions; some evidence for this could be observed in the THF GPC chromatograms of the di- and triblock copolymers (Figure 2). Initiation appeared to be slow, rather than incomplete, since macroinitiator contamination was evident in the THF GPC trace of the intermediate diblock copolymer but was essentially absent in the THF GPC trace of the PPO33-PDMA20-PGMA50 triblock copolymer, which had a final polydispersity of around 1.26. Similar polydispersities were obtained using DMF GPC. SCL Micelle Synthesis. Dynamic light scattering (DLS) experiments demonstrated that the triblock copolymer was molecularly dissolved in water at around pH 8 and 5 °C. The critical micellization temperature (CMT) of the PPO33-PDMA20-PGMA50 triblock copolymer was approximately 12-13 °C, as estimated by turbidimetry experiments. At the CMT, the PPO block becomes insoluble in water, leading to the formation of micelles with PPO cores and PGMA coronas. However, the PPO block is only weakly hydrophobic at 20 °C, so well-defined micelles were not expected at ambient temperature.6 At 40 °C two populations were observed at 27 and 249 nm by DLS studies. The smaller population is attributed to (8) Bories-Azeau, X.; Armes, S. P. Macromolecules 2002, 35, 10241.

SCL Micelles with Hydroxy-Functional Coronas

Langmuir, Vol. 21, No. 9, 2005 3811

Figure 4. Effect of varying the cross-linking temperature on the intensity-average hydrodynamic diameter of the final shell cross-linked micelles as determined by DLS at 5 °C: target degree of quaternization ) 50%; PPO33-DMA20-GMA50 triblock copolymer concentration ) 0.50%.

Figure 3. Reaction scheme for the cross-linking of the central DMA block via quaternization in aqueous solution using 1,2bis(2-iodoethoxy)ethane.

micelles, and the larger population is believed to be due to aggregates formed by weakly interacting micelles. Despite this unexpected micelle aggregation, the micelle cross-linking proceeded successfully. PPO-core micelles were formed at around pH 8 and 40 °C and were then cross-linked via quaternization of the amine residues using a bifunctional alkyl iodide (Figure 3) at both 0.50 and 10.0 w/v % solids. Diluted solutions of the resulting SCL micelles were examined by DLS at 5 °C to confirm that cross-linking had occurred. The PPO chains are completely hydrophilic at 5 °C. Thus, if no cross-linking had occurred, there would be no micelles present at this temperature. However, DLS studies at 5 °C confirmed the existence of SCL micelles at this temperature, indicating that the crosslinking was successful. Additionally, little difference between the SCL micelles prepared at 0.50% and 10% solids was observed, indicating that the outer PGMA block was successful in minimizing inter-micelle fusion.9 At 5 °C, SCL micelles of 31 nm diameter were observed. On heating of the sample to 40 °C, this size distribution became bimodal, with populations at 24 and 390 nm. The smaller population was attributed to progressive dehydration of the PPO chains, leading to micelle core contraction at higher temperatures. The larger population appears to be due to weak, reversible micelle aggregation or flocculation: on cooling of the sample to 5 °C, only the smaller population was observed. This phenomenon appears to be superficially similar to the “critical flocculation temperature” observed for sols stabilized by poly(9) (a) Bu¨tu¨n, V.; Wang, X.-S.; de Paz Banez, M. V.; Robinson, K. L.; Billingham, N. C.; Armes, S. P. Macromolecules 2000, 33, 1. (b) Liu, S.; Tang, Y. Q.; Weaver, J. V. M.; Billingham, N. C.; Armes, S. P.; Tribe, K. Macromolecules 2002, 35, 6121.

(ethylene oxide).10 Accordingly, control experiments were performed on an aqueous solution of a PGMA homopolymer, but no evidence for aggregation was observed at elevated temperatures by DLS. At present we are unable to account for this reversible temperature-induced micelle aggregation, but given the chemical structure of the outer PGMA block, it may involve weak inter-micelle hydrogen bonding. The micelle aggregation discussed above meant that it was not possible to observe the expected micelle contraction due to progressive core dehydration at higher temperatures. However, it was possible to perform crosslinking of the micelles at various temperatures and then analyze the resulting SCL micelle solutions by DLS at 5 °C, where micelle aggregation was minimal. As expected, the SCL micelle diameter decreased dramatically from around 70 to 30 nm (Figure 4) as the temperature employed for shell cross-linking was increased. Unfortunately, DLS studies recorded at 5 °C on SCL micelles that were prepared at 70 °C indicated that inter-micelle fusion becomes problematic at higher temperatures. However, this unwanted side reaction is much less evident for shell cross-linking reactions conducted below 70 °C. BIEE cross-linking of the PPO33-DMA20-GMA50 triblock copolymer was performed in D2O to investigate the reversible hydration of the PPO core block (Figure 5). The SCL micelle solution was cycled between 5 and 50 °C, and 1 H NMR spectra were recorded at these temperatures. Signals due to the PPO CH3 group were clearly attenuated in the spectra obtained at 50 °C relative to the spectra obtained at 5 °C, and there were no discernible differences between the original and final spectra recorded at 5 °C. The attenuation of the PPO signals is consistent with the dehydration of this block due to the disruption of its hydrogen bonding in aqueous solution at higher temperature. Some attenuation was also observed for the signals assigned to the PDMA block. This is understandable since the PDMA homopolymer also has a cloud point at this pH, albeit one that is higher than that of PPO.11 (10) (a) Cowell, C.; Vincent, B. J. Colloid Interface Sci. 1982, 87, 518. (b) Cowell, C.; Vincent, B. In The Effects of Polymers on Dispersion Properties; Tadros, T. F., Ed.; Academic Press Inc. Ltd.: London, 1982; p 263. (11) Bu¨tu¨n, V.; Billingham, N. C.; Armes, S. P. Chem. Commun. 1997, 671.

3812

Langmuir, Vol. 21, No. 9, 2005

Pilon et al.

Figure 7. Aqueous electrophoresis curves for (a) the Monospher 250 silica sol, (b) the shell cross-linked micelles prepared using the PPO33-DMA20-GMA50 triblock copolymer (target degree of quaternization ) 100%), and (c) the shell cross-linked micelledecorated silica particles.

Figure 5. 1H NMR spectra recorded for (a) the PPO33-DMA20GMMA50 triblock copolymer precursor in pyridine-d5, (b) shell cross-linked micelles prepared from this copolymer in D2O at 5 °C (target degree of quaternization ) 50%), and (c) the same shell cross-linked micelle solution at 50 °C. Note the attenuation of the signal at 1.0-1.5 ppm due to partial dehydration of the PPO-core block.

Figure 6. Transmission electron micrograph of the shell crosslinked micelles prepared using the PPO33-DMA20-GMA50 triblock copolymer. The target degree of quaternization was 50%, and the aqueous copolymer concentration during shell cross-linking was 0.50%.

Transmission electron microscopy (TEM) was performed on diluted solutions of the SCL micelles (target degree of cross-linking ) 50% at a cross-linking temperature of 40 °C) to examine their structure (Figure 6). The images obtained clearly show polydisperse, spherical particles of around 30-60 nm diameter, which is comparable to the micelle dimensions obtained from the DLS studies. Good agreement was not necessarily expected, since DLS reports intensity-average diameters whereas number-average diameters are obtained by TEM analysis. Thus, for samples with finite polydispersities DLS always oversizes relative to TEM and larger discrepancies are expected for

broader size distributions. Moreover, DLS is sensitive to the hydrodynamic diameter which includes the PGMA chains, whereas this outer layer will dry down to a much reduced thickness under the ultrahigh-vacuum conditions employed for TEM. Adsorption of the SCL Micelles onto the Silica Sols. Aqueous electrophoresis studies of the bare Monospher 250 silica particles indicated that this sol was cationic at pH 2 but had an isoelectric point at pH 3 and became anionic at higher pH, as expected. In contrast, the ζ potentials of the SCL micelles (target degree of crosslinking ) 100% at a cross-linking temperature of 40 °C) were positive from pH 2 to pH 10 (Figure 7), which makes them suitable for electrostatic adsorption onto anionic surfaces such as silica (or mica) at around neutral pH. The optimum pH for SCL micelle adsorption onto the silica sol was determined to be around pH 8-9, since this would ensure a strong electrostatic attraction between the two types of particles. After adsorption of the SCL micelles onto the silica particles, the electrophoretic behavior of the micelle-decorated silica particles resembled that of the original SCL micelles, indicating that surface charge reversal was achieved and suggesting relatively high surface coverage. Analysis of the SCL micelle-decorated silica particles by diffuse reflectance infrared Fourier transform spectroscopy (DRIFTS) also confirmed the presence of the SCL micelles on the surface of the silica particles: the presence of a weak carbonyl stretch at 1725 cm-1 was observed in the spectrum of the SCL micelledecorated silica particles that was not present in the spectrum of the bare silica particles (not shown). Thermogravimetric analyses of the SCL-micelle-decorated silica sols indicated micelle loadings of 2.5 and 4.4% by mass for the Monospher 1000 and Monospher 250 particles, respectively. If one takes into account the particle diameters and densities of the two sols, adsorbed amounts of 8.3 and 3.7 mg m-2 are calculated for the 1000 and 250 nm silica sols, respectively. Finally, confirmation of the presence of the adsorbed SCL micelles on the surface of the silica particles was provided by scanning electron microscopy (SEM). For both the Monospher 1000 (Figure 8) and Monospher 250 particles, clear differences in surface morphology were observed between the bare silica sol and the SCL-micelledecorated silica particles. Moreover, the SCL micelle coating on the Monospher 1000 silica particles was considerably more uniform than that on the Monospher 250 silica particles (not shown). This observation accounts for the differing adsorbed amounts (see above) and is

SCL Micelles with Hydroxy-Functional Coronas

Langmuir, Vol. 21, No. 9, 2005 3813

40 °C) adsorbed onto the Monospher 1000 silica sol were estimated to be around 50 nm (it is difficult to be more precise, since such measurements are at the limits of resolution of our SEM instrument). This is comparable to the TEM diameters observed in Figure 6 and indicates relatively little spreading of the cationic SCL micelles on the highly anionic silica surface. Given the low Tg for both the PPO and the PDMA blocks, this result is perhaps surprising at first sight. However, bearing in mind the high degree of shell cross-linking that was targeted, the SCL micelles are expected to be rather rigid and nondeformable. Conclusions The one-pot synthesis of a new hydroxy-functional PPO33-PDMA20-PGMA50 triblock copolymer was achieved without the need for protecting group chemistry. Cationic SCL micelles with variable mean micelle diameters can be readily prepared from this ABC triblock copolymer by simply varying the temperature at which shell crosslinking is carried out. Their high concentration of surface hydroxy groups means that these SCL micelles are similar in some respects to dendrimers, but their synthesis is considerably more facile. These cationic SCL micelles can be adsorbed onto the surface of anionic silica sols, which leads to surface charge reversal as judged by aqueous electrophoresis measurements. Scanning electron microscopy and thermogravimetric analyses indicate that high surface coverages can be achieved. Figure 8. Scanning electron micrographs of (a) uncoated Monospher 1000 silica particles and (b) shell cross-linked micelle-decorated silica particles (target degree of quaternization ) 100%). TGA analyses indicated a micelle loading of 2.5 wt % on the silica sol.

attributed to differences in the curvature of the silica surface relative to the dimensions of the SCL micelles, which affects the micelle packing efficiency. The approximate dimensions of the SCL micelles (target degree of cross-linking ) 100% at a cross-linking temperature of

Acknowledgment. Unilever is thanked for an Industrial CASE PhD studentship for L.N.P. and for permission to publish this work. E. Merck is thanked for supplying the Monospher silicas used in this work, and Cognis Performance Chemicals is thanked for donating the monohydroxy-capped poly(propylene oxide) precursor. Dave Randall (Sussex University) is thanked for his assistance with the scanning electron microscopy studies. LA047046G