pubs.acs.org/Langmuir © 2010 American Chemical Society
Polyamine-Functional Sterically Stabilized Latexes for Covalently Cross-Linkable Colloidosomes A. Walsh, K. L. Thompson, and S. P. Armes* Dainton Building, Department of Chemistry, The University of Sheffield, Brook Hill, Sheffield, South Yorkshire S3 7HF, U.K.
D. W. York Procter & Gamble, Newcastle Technical Centre, Whitley Road, Longbenton, Newcastle-upon-Tyne NE12 9TS, U.K. Received September 22, 2010. Revised Manuscript Received October 22, 2010 Sterically stabilized polystyrene latexes were prepared by aqueous emulsion polymerization using a poly(ethylene imine) (PEI) stabilizer in the presence of 4-vinylbenzyl chloride (4-VBC; 1.0 wt % based on styrene). Partial quaternization of the amine groups on the PEI chains by 4-VBC occurs in situ, hence producing a chemically grafted steric stabilizer. Such 4-VBC-modified PEI chains were grafted more efficiently onto the polystyrene particles than unmodified PEI, as judged by aqueous electrophoresis, XPS, and nitrogen microanalysis. Moreover, partially quaternized PEI gave significantly smaller polystyrene particles than those synthesized in the absence of any PEI stabilizer or those synthesized using unmodified PEI. The partially quaternized PEI-stabilized polystyrene latex proved to be an effective emulsifier at pH 9, forming stable oil-in-water Pickering emulsions when homogenized (12 000 rpm, 2 min, 20 °C) with four model oils, namely, n-dodecane, methyl myristate, isononyl isononanoate, and sunflower oil. The primary and/or secondary amine groups on the PEI stabilizer chains were successfully cross-linked using three commercially available polymeric reagents, namely, tolylene 2,4-diisocyanate-terminated poly(propylene glycol) (PPG-TDI), poly(propylene glycol) diglycidyl ether (PPG-DGE), or poly(ethylene glycol) diglycidyl ether (PEG-DGE). Crosslinking with the former reagent led to robust colloidosomes that survived the removal of the internal oil phase on washing with excess alcohol, as judged by optical microscopy and SEM. PPG-TDI reacted very rapidly with the PEI stabilizer chains, with cross-linking being achieved during homogenization. Well-defined colloidosomes could be formed only by using sunflower oil and isononyl isononanoate with this cross-linker at 20 °C. However, cooling to 0 °C allowed colloidosomes to be formed using n-dodecane, presumably because of the slower rate of cross-linking at this reduced temperature. PPG-DGE proved to be a more generic cross-linker because it formed robust colloidosomes with all four model oils. However, cross-linking was much slower than that achieved using PPG-TDI, with intact colloidosomes being formed only after ∼12 h at 20 °C. The PEG-DGE cross-linker allowed cross-linking to be conducted at 20 °C from the aqueous phase (rather from within the oil droplets for the oil-soluble PPG-TDI or PPG-DGE cross-linkers). In this case, well-defined colloidosomes were obtained at 50 vol % with surprisingly little intercolloidosome aggregation, as judged by laser diffraction studies.
Introduction Colloidal dispersions of polymer latexes are essential components in many commercial applications, including coatings, paints, and adhesives.1 Among the various methods developed to produce polymer particles, aqueous emulsion polymerization offers a particularly facile and efficient route to obtaining submicrometer-sized latexes.2,3 Steric stabilization of polymer latexes in both aqueous and organic media using chemically grafted polymers has been well known for many years.4,5 The advantages of steric stabilization compared to charge stabilization include much better resistance toward added electrolyte and high shear and significantly *Author to whom correspondence should be addressed. E-mail: s.p.armes@ sheffield.ac.uk. (1) Daniels, E. S.; Sudol, E. D.; El-Aasser, M. S. ACS Symp. Ser. 1992, 492, 282– 288. (2) Kawaguchi, S.; Ito, K. Adv. Polym. Sci. 2005, 175, 299–328. (3) Lovell, P. A.; El-Aasser, M. S. Emulsion Polymerisation and Emulsion Polymers; John Wiley and Sons: Chichester, U.K., 1997. (4) Barrett, K. E. J. Dispersion Polymerization in Organic Media; Wiley: New York, 1975. (5) Napper, D. H. Polymeric Stabilization of Colloidal Dispersions; Academic Press: London, 1983.
Langmuir 2010, 26(23), 18039–18048
improved freeze-thaw stability.6,7 Suitable steric stabilizers include simple homopolymers,8 well-defined macromonomers,9 and block copolymers.10 Many types of colloidal particles can act as emulsifiers when adsorbed at the interface between two immiscible liquids. The resulting emulsions are often very stable and are known as Pickering emulsions in recognition of the pioneering work in this field conducted by Pickering more than a century ago.11 Examples of such Pickering emulsifiers include silica,12 polystyrene latex,13,14 and Laponite clay particles.15 Pickering emulsions are encountered (6) Ottewill, R. H.; Satgurunathan, R. Colloid Polym. Sci. 1995, 273, 379–386. (7) Capek, I. Adv. Colloid Interface Sci. 2009, 99, 77–162. (8) Li, W.; Li, P. Macromol. Rapid Commun. 2007, 28, 2267–2271. (9) Thompson, K. L.; Armes, S. P.; York, D. W.; Burdis, J. A. Macromolecules 2010, 43, 2169–2177. (10) Houillot, L.; Bui, C.; Farcet, C.; Moire, C.; Raust, J.; Pasch, H.; Save, M.; Charleux, B. Appl. Mater. Interfaces 2010, 2, 434–442. (11) Pickering, S. U. J. Chem. Soc. 1907, 91, 2001–2021. (12) Binks, B. P.; Lumsdon, S. O. Phys. Chem. Chem. Phys. 1999, 1, 3007–3016. (13) Binks, B. P.; Lumsdon, S. O. Langmuir 2001, 17, 4540–4547. (14) Amalvy, J. I.; Unali, G. F.; Li, Y.; Granger-Bevan, S.; Armes, S. P. Langmuir 2004, 20, 4345–4354. (15) Ashby, N. P.; Binks, B. P. Phys. Chem. Chem. Phys. 2000, 2, 5640–5646.
Published on Web 11/09/2010
DOI: 10.1021/la103804y
18039
Article
in various industrial processes such as crude oil recovery,16 cosmetic formulations,17 and food technology.18 The contact angle made by the adsorbed particles at the interface depends on their wettability. Hydrophobic particles tend to be preferentially wetted by the oil phase and hence stabilize water-in-oil (w/o) emulsions, whereas hydrophilic particles favor the aqueous phase and thus stabilize oilin-water (o/w) emulsions.19 Subsequent interfacial stabilization of these Pickering emulsions leads to the formation of robust microcapsules. These microcapsules are now known as colloidosomes, a term first coined by Dinsmore and co-workers.20 However, the first report of the preparation of colloidosome-type structures was by Velev’s group in 1996.21-23 More recently, Yow and Routh24 extended earlier work by both Dinsmore and co-workers20 and also Laib and Routh25 by investigating the permeability of latexbased colloidosomes prepared under different annealing conditions. A poly(styrene-co-n-butyl acrylate) latex was used to stabilize both water-in-oil and oil-in-water emulsions, and colloidosomes were formed after annealing at 49 °C for either 5, 30, or 60 min. Longer annealing times led to smoother shells, with eventual latex fusion occurring to form a continuous copolymer film. This in turn resulted in the somewhat retarded release of an encapsulated watersoluble dye, although dye retention was achieved only over timescales of a few hours at ambient temperature. An alternative route to robust colloidosomes is covalent crosslinking. Previous papers describing colloidosome stabilization by covalent cross-linking include that by Croll and St€over, who assembled poly(divinylbenzene-alt-maleic anhydride) microspheres at an oil/water interface that were cross-linked from the continuous aqueous phase by the addition of poly(ethylene imine).26 Covalently cross-linked colloidosomes have also been described by Emrick et al.27 This team self-assembled surface-functionalized quantum dots at a toluene/water interface and subsequently crosslinked these nanoparticles using a Grubbs catalyst. More recently, Thompson and Armes reported the preparation of covalently cross-linked colloidosomes using sterically stabilized polystyrene latexes.28 The latexes acted as efficient Pickering emulsifiers for the stabilization of n-dodecane-in-water emulsions. Colloidosomes were then formed by cross-linking these hydroxy-functional steric stabilizer chains using a polymeric diisocyanate cross-linker located solely within the oil droplets. This strategy enabled efficient crosslinking to be achieved at 50 vol % without causing any discernible intercolloidosome aggregation. However, one drawback of this approach is the relatively high cost of the hydroxyl-functional monomer (glycerol monomethacrylate). In the present work, we utilize an alternative commercially available steric stabilizer, poly(ethylene imine) (PEI), to produce polyamine-functionalized polystyrene latexes. Covalent cross-linking of the PEI chains was conducted with three commercially available polymeric cross-linkers. Two of these reagents are oil-soluble, and the other is water-soluble. Thus, the effect of the physical location of the cross-linker could be evaluated, (16) Mikula, R. J.; Munoz, V. A. Colloids Surf., A 2000, 174, 23–36. (17) Brinon, L.; Geiger, S.; Alard, V.; Tranchant, J. F.; Pouget, T.; Couarraze, G. J. Cosmet. Sci. 1998, 49, 1–11. (18) Dalgleish, D. G.; West, S. J.; Hallett, F. R. Colloids Surf., A 1997, 145, 123– 124. (19) Binks, B. P. Curr. Opin. Colloid Interface Sci. 2002, 7, 21–41. (20) Dinsmore, A. D.; Hsu, M. F.; Nikolaides, M. G.; Marquez, M.; Bausch, A. R.; Weitz, D. A. Science 2002, 298, 1006–1009. (21) Velev, O. D.; Furusawa, K.; Nagayama, K. Langmuir 1996, 12, 2374–2384. (22) Velev, O. D.; Furusawa, K.; Nagayama, K. Langmuir 1996, 12, 2385–2391. (23) Velev, O. D.; Nagayama, K. Langmuir 1997, 13, 1856–1859. (24) Yow, H. N.; Routh, A. F. Langmuir 2009, 25, 159–166. (25) Laib, S.; Routh, A. F. J. Colloid Interface Sci. 2007, 317, 121–129. (26) Croll, L. M.; St€over, H. D. H. Langmuir 2003, 19, 5918–5922. (27) Skaff, H.; Lin, Y.; Tangirala, R.; Breitenkamp, K.; B€oker, A.; Russell, T. P.; Emrick, T. Adv. Mater. 2005, 17, 2082–2086. (28) Thompson, K. L.; Armes, S. P. Chem. Commun. 2010, 43, 2169–2177.
18040 DOI: 10.1021/la103804y
Walsh et al.
and the efficacy of different cross-linking chemistries could be examined.
Experimental Section Materials. Poly(ethylene imine) (PEI) (branched, Mn = 10 000 by GPC), 4-vinylbenzyl chloride (4-VBC, 90%), 2,20 -azobis(isobutyramidine) dihydrochloride (AIBA, 97%), tolylene 2,4-diisocyanate-terminated poly(propylene glycol) (PPG-TDI, mean degree of polymerization is 34 according to the supplier), poly(propylene glycol) diglycidyl ether (PPG-DGE, mean degree of polymerization is 9 according to the supplier), and poly(ethylene glycol) diglycidyl ether (PEG-DGE; mean degree of polymerization is 9 according to the supplier) were purchased from Aldrich and were used as received. Jeffamine M-2005 (a statistical copolymer comprising approximately 29 propylene glycol units and 6 ethylene glycol units with 1 terminal primary amine group per chain and a molecular weight of around 2000) was obtained from Huntsman Chemicals, U.K. and was used as received. Styrene (Aldrich) was passed through a column of basic alumina to remove inhibitor and then stored at -20 °C prior to use. 2-Propanol was purchased from Fisher and used as received. Deionized water was used in all experiments. Aqueous Emulsion Polymerization. PEI (1.0 g) was weighed into a 100 mL round-bottomed flask with a magnetic stir bar and dissolved in water (39.0 g). To this solution were added styrene (4.95 g) and 4-VBC (0.050 g, 1.5 mol % based on the nitrogen content of PEI, for which N = 30.45%). The reaction mixture was purged with nitrogen gas for 1 h and heated to 60 °C using an oil bath. Then a solution of AIBA (0.050 g) in water (5.0 g) was injected into the reaction vessel to initiate polymerization. The solution turned milky within 1 h and was stirred at 250 rpm for 24 h at 60 °C. The resulting latex was purified by three centrifugation/redispersion cycles, replacing each decanted supernatant with deionized water. The same protocol was also employed in the absence of any 4-VBC and in the absence of both 4-VBC and PEI. The particle size of the resulting three latexes was assessed by dynamic light scattering (before centrifugation), scanning electron microscopy (after centrifugation), and disk centrifuge photosedimentometry (after centrifugation). Aqueous electrophoresis, X-ray photoelectron spectroscopy, and nitrogen microanalysis were used to examine the surface properties and stabilizer contents of the latexes. Colloidosome Preparation with the 2,4-Diisocyanate-Terminated Poly(propylene glycol) (PPG-TDI) Cross-Linker. The following protocol is representative. PPG-TDI cross-linker (0.020 g) dissolved in sunflower oil (5.0 mL) was added to a 14 mL sample vial, followed by the addition of an aqueous PEI-PS latex (5.0 mL, 1.0 wt %, pH 9). Emulsification was achieved using an IKA Ultra-Turrax T-18 homogenizer with a 10 mm dispersing tool for 2 min at 12 000 rpm. The emulsion was then allowed to stand unstirred at 20 °C to allow colloidosome formation to occur. The same volume of either methyl myristate, n-dodecane, or isononyl isononanoate was also used as an alternative to sunflower oil using the same emulsification protocol. This method of colloidosome preparation was also carried out at 0 °C using an ice bath.
Colloidosome Preparation with the Poly(propylene glycol) Diglycidyl Ether (PPG-DGE) Cross-Linker. PPG-DGE crosslinker (0.020 g) dissolved in sunflower oil (5.0 mL) was added to a 14 mL sample vial, followed by the addition of an aqueous PEI-PS latex (5.0 mL, 1.0 wt %, pH 9). Emulsification was achieved at 20 °C using an IKA Ultra-Turrax T-18 homogenizer with a 10 mm dispersing tool for 2 min at 12 000 rpm. The emulsion was then allowed to stand unstirred at 20 °C to allow colloidosome formation to occur. The same volume of either methyl myristate, n-dodecane, or isononyl isononanoate was also used as an alternative to sunflower oil using the same emulsification protocol.
Colloidosome Preparation with the Poly(ethylene glycol) Diglycidyl Ether (PEG-DGE) Cross-Linker. Sunflower oil (5.0 mL) was added to a 14 mL sample vial, followed by the Langmuir 2010, 26(23), 18039–18048
Walsh et al.
Article
Figure 1. Reaction scheme for the formation of polyamine sterically stabilized PS latexes by aqueous emulsion polymerization at 60 °C, which act as suitable Pickering emulsifiers to form oil-in-water emulsions. addition of an aqueous PEI-stabilized polystyrene latex (5.0 mL, 1.0 wt %, pH 9). Emulsification was achieved using an IKA UltraTurrax T-18 homogenizer with a 10 mm dispersing tool for 2 min at 12 000 rpm. PEG-DGE (0.020 g) was added to the aqueous continuous phase and shaken to ensure dissolution. The emulsion was then allowed to stand unstirred at 20 °C to allow colloidosome formation to occur. The same volume of either methyl myristate, n-dodecane, or isononyl isononanoate was also used as an alternative to sunflower oil using the same emulsification protocol. Alcohol Challenge. An aliquot of a colloidosome sample (0.50 mL of a 50 vol % suspension) was transferred to a 5.0 mL vial, to which excess 2-propanol (4.0 mL) was added. The vial was shaken vigorously, and the sample was viewed by both optical microscopy and SEM.
Characterization Elemental Microanalysis. CHN microanalysis was performed in-house on approximately 10 mg samples using a Perkin-Elmer 2400 instrument. Optical Microscopy. A single droplet (∼100 μL) of a Pickering emulsion (either with or without polymeric cross-linker) was placed on a microscope slide and viewed using an optical microscope (James Swift MP3502, Prior Scientific Instruments Ltd.) fitted with a digital camera (Nikon Coolpix 4500). This technique was used to estimate the mean latex diameter and, in some cases, to confirm whether cross-linked colloidosomes had been formed after an alcohol challenge. Laser Diffraction. A Malvern Mastersizer 2000 laser diffraction instrument equipped with a small-volume (ca. 50 mL) Hydro 2000SM sample dispersion unit, a HeNe laser operating at 633 nm, and a solid-state blue laser operating at 466 nm was used to size the Pickering emulsions and colloidosomes. The stirring rate was adjusted to 1500 rpm. Corrections were made for background electrical noise and laser scattering due to contaminants on the optics and within the sample. Samples were analyzed five times, and the data were averaged. A typical acquisition time was 2 min per sample after alignment and background measurements. The raw data were analyzed using Malvern software. The mean latex diameter was taken to be the mean volume-average diameter (D4/3), which is mathematically expressed as D4/3 = Σ Di4Ni/Σ Di3Ni. The standard deviation for each diameter provides an indication of the width of the size distribution. After each measurement, the cell was rinsed three times with ethanol, followed by three times with deionized water. The glass walls of the cell were carefully wiped with a lens cleaning tissue to avoid cross-contamination, and the laser was aligned centrally on the detector. Scanning Electron Microscopy (SEM). Images were obtained using an FEI Inspect instrument operating at 5 kV. All samples were dried on aluminum stubs and sputter-coated with a thin overlayer of gold prior to inspection to prevent samplecharging effects. Mean latex diameters were estimated by counting at least 100 particles. Dynamic Light Scattering (DLS). A Malvern Zetasizer NanoZS instrument was used to obtain intensity-average hydrodynamic diameters of the latex particles via the Stokes-Einstein Langmuir 2010, 26(23), 18039–18048
equation. Aqueous solutions of a 0.01 wt % latex were analyzed using disposable cuvettes, and the results were averaged over three consecutive runs. The deionized water used to dilute each latex was ultrafiltered through a 0.20 μm membrane to remove any dust. Disk Centrifuge Photosedimentometry (DCP). The weightaverage diameter (Dw) of each latex was assessed using a CPS disk centrifuge photosedimentometer. Dilute aqueous dispersions (∼0.1%, 0.10 mL) were injected into an aqueous spin fluid containing a sucrose gradient (15.0 mL). The gradient selected for the analysis of polystyrene latexes ranged from 2 to 8% sucrose. The polystyrene particle density was taken to be 1.05 g cm-3. Aqueous Electrophoresis. Zeta potentials were determined using a Malvern Zetasizer NanoZS instrument equipped with an autotitrator (MPT-2 multipurpose titrator, Malvern instruments). The solution pH was varied from 2 to 9 in the presence of 1 mM KCl using either dilute NaOH or HCl as required. X-ray Photoelectron Spectra (XPS). XPS were acquired usintg a Kratos Axis Ultra DLD X-ray photoelectron spectrometer equipped with a monochromatic Al KR X-ray source (hν = 1486.6 eV) and operating at a base pressure in the range of 10-8 to 10-10 mbar. Latex particles were dried on a silicon wafer and evacuated to ultrahigh vacuum prior to XPS measurements. FTIR Spectroscopy. Samples were recorded using a golden gate ATR accessory. The FTIR spectra were recorded using a Nicolet iS10 spectrometer at 4.0 cm-1 resolution and 32 scans were recorded per spectrum.
Results and Discussion PEI is a commercially available amine-functional, cationic water-soluble polymer. The sample used in this study was a relatively polydisperse branched homopolymer with a nominal molecular weight of 10 000. PEI-stabilized polystyrene latexes were prepared in situ at around 10% solids using aqueous emulsion polymerization at 60 °C (Figure 1). This protocol produced submicrometer-sized particles of around 100-200 nm diameter (Figure 2 and Table 1). Our protocol differs somewhat from that described by Li and co-workers, who used a tert-butyl hydroperoxide initiator to generate radicals on the PEI chains and hence promote grafting onto a poly(methyl methacrylate) (PMMA) latex to give amphiphilic core-shell particles ranging from 60 to 160 nm in diameter.29 As well as using PEI, Li and co-workers have utilized other aminefunctional water-soluble polymers such as poly(vinyl amine), which was grafted onto polystyrene to give core-shell particles of 187 nm in diameter.8 The mean particle diameters of the three latexes prepared by emulsion polymerization were assessed using SEM, DCP, and DLS (Table 1). Smaller particles were achieved after derivatization of the PEI chains with 4-VBC. This reagent randomly quaternizes the tertiary amine groups on the stabilizer, introducing polymerizable pendent styrene groups. This enables more efficient grafting of the PEI chains onto the latex surface, which in turn leads to (29) Li, P.; Zhu, J.; Suintaboon, P.; Harris, F. W. Langmuir 2002, 18, 8641–8646.
DOI: 10.1021/la103804y
18041
Article
Walsh et al.
Figure 2. Scanning electron microscopy (SEM) images (top) and corresponding particle size distribution curves (bottom) for a partially quaternized PEI-stabilized PS latex (entry 3, Table 1), a nonquaternized PEI-stabilized PS latex (entry 2, Table 1), and a charge-stabilized PS latex (entry 1, Table 1). The leftmost SEM image shows tiny surface features on the latex particles. This is because all samples are sputtercoated with gold to prevent sample charging and the gold deposits are visible at high magnification. Table 1. Summary of Particle Size Data Obtained for PS Latexes Prepared by Aqueous Emulsion Polymerization in the Presence of an AIBA Initiator at 60 °C for 24 ha entry no.
description
monomer conversion/%
SEM diameter/nm
DCP diameter/nm
DLS diameter/nm (PDI)
%N
1 no stabilizer 67 528 582 ( 63 662 (0.053) 0 2 PEI stabilizer alone 87 132 219 ( 42 214 (0.074) 0.5 3 PEI stabilizer þ 4-VBC 90 118 200 ( 66 176 (0.052) 1.2 a Nitrogen and PEI contents are also shown, as calculated by elemental microanalysis. The nitrogen content of PEI alone was 30.45%.
more effective steric stabilization. Although our in situ protocol is both convenient and robust, the precise sequence of the reaction between 4-VBC and the PEI chains remains unclear. Two obvious possibilities are either (i) the 4-VBC reagent quaternizes PEI during the degassing process at 60 °C prior to styrene addition (and hence before any polymerization) or (ii) 4-VBC copolymerizes with the styrene, followed by subsequent quaternization of PEI. A third possibility is that both reactions occur simultaneously. Previous work has shown that when 4-VBC and styrene are copolymerized, 4-VBC is preferentially expressed at the surface of the latex particles.30 In view of this and given that the water solubility of 4-VBC itself is only around 0.073 g/L31 (our in situ grafting formulation involves a 4-VBC concentration of 1.12 g/L), it is perhaps more likely that scenario (ii) occurs. Although the precise sequence of quaternization cannot be confirmed, it is worth emphasizing that in the absence of the 4-VBC reagent rather larger latex particles containing significantly less PEI are produced. Moreover, these latexes do not act as effective Pickering emulsifiers (see later). Entries 2 and 3 in Table 1 indicate a discrepancy between the intensity-average and weight-average diameters reported by DLS and DCP, respectively. In principle, the DLS diameter should always exceed the corresponding DCP diameter, whereas in this case the DCP diameter is the larger value. This is a known artifact that occurs when sizing sterically stabilized latexes of less than approximately 200 nm diameter using the DCP instrument.32 Under these circumstances, the (highly hydrated) PEI stabilizer layer thickness becomes significant relative to the mean particle diameter, resulting in a reduction in the effective particle density. The DCP (30) Okubo, M.; Iwasaki, Y.; Yamamoto, Y. Colloid Polym. Sci. 1992, 270, 733–737. (31) Sarobe, J.; Forcada, J. Colloids Surf., A 1998, 135, 293–297. (32) Cairns, D. B.; Armes, S. P.; Bremer, L. G. B. Langmuir 1999, 15, 8052–8058.
18042 DOI: 10.1021/la103804y
% PEI 0 1.7 4
technique requires an accurate particle density for meaningful particle size analysis. Uncertainty in the particle density leads to an associated error in the particle diameter, and this problem is exacerbated for polystyrene latexes because the polystyrene density of 1.05 g cm-3 is relatively close to that of water.32 Thus, DCP oversizes the PEI-stabilized latex because its true (unknown) particle density is lower than that of polystyrene. The PEI contents of sterically stabilized latexes prepared by aqueous emulsion polymerization can be estimated from nitrogen microanalyses. The latex prepared in the presence of 4-VBC had the highest PEI content of ∼4% by mass (entry 3 in Table 1). This indicates that the addition of 4-VBC to the one-pot synthesis leads to more efficient in situ grafting of the PEI chains onto the latex surface. As a control, a polystyrene latex was also prepared in the absence of any PEI stabilizer or 4-VBC under the same conditions (entry 1 in Table 1). This latex was significantly larger than the two PEI-stabilized latexes: its colloidal stability is attributed to the surface charge arising from the use of a cationic azo initiator (AIBA). In this case, no nitrogen is detectable by microanalysis, suggesting that the presence of nitrogen in the PEI-stabilized latexes (entries 2 and 3 in Table 1) is solely due to PEI, rather than the AIBA initiator fragments. However, microanalyses cannot be used to assess the spatial location of the PEI chains. Aqueous electrophoresis and X-ray photoelectron spectroscopy (XPS) were used to confirm the presence of the PEI stabilizer at the latex surface. Aqueous electrophoresis is sensitive to the presence of PEI because of its polyamine nature. The zeta potential of the polystyrene latex prepared using PEI in the presence of 4-VBC (entry 3 in Table 1) remained positive from pH 2 to 9 (Figure 3). This indicates that PEI chains are indeed present at the surface of this latex, as expected. In contrast, the zeta potential versus pH curve obtained for the charge-stabilized polystyrene latex (entry 1, Langmuir 2010, 26(23), 18039–18048
Walsh et al.
Article
Figure 3. Variation of zeta potential with pH for () the (PEI þ 4-VBC)-stabilized latex (entry 3, Table 1), (2) the PEI-stabilized PS latex (entry 2, Table 1), and (0) the charge-stabilized PS latex (entry 1, Table 1).
Table 1) has an isoelectric point at around pH 7.1, which is close to that reported by Kettlewell et al. for a similar latex prepared using an AIBA initiator.33 The PEI-stabilized latex prepared in the absence of any 4-VBC exhibits intermediate behavior: it remains cationic across a slightly wider pH range than the charge-stabilized latex and has an isoelectric point at approximately pH 8.5. This suggests that the surface concentration of PEI chains is significantly lower than that achieved when using the 4-VBC reagent, which is consistent with the nitrogen microanalysis data. Further evidence to support the aqueous electrophoresis results was provided by XPS analysis (Figure 4). XPS is a surface-specific technique with a typical analysis depth of ∼2-5 nm.34 Therefore, any grafted PEI chains present on the latex surface should be readily detectable. The XPS spectrum recorded for pristine PEI (Figure 4a) revealed an intense nitrogen signal (14.8 atom %) due to its primary, secondary, and tertiary amine groups; this sample was used as a reference material. The charge-stabilized latex contained almost no nitrogen signal, indicating a relatively low surface concentration of AIBA initiator fragments. In contrast, the latex prepared using PEI without any 4-VBC gave a weak nitrogen signal (2.9 atom %, see Figure 4c), whereas the latex prepared using both PEI and 4-VBC had a more prominent nitrogen signal (4.3 atom %, see Figure 4d). Comparing the intensities of these nitrogen signals to that of the PEI reference allowed PEI surface coverages of 20 and 29% to be estimated, respectively. Thus, these XPS observations support the aqueous electrophoresis data, with both techniques suggesting that the 4-VBC reagent leads to more efficient grafting of the PEI stabilizer onto the latex particles than using the pristine PEI stabilizer alone. These polyamine-functionalized polystyrene latexes proved to be effective Pickering emulsifiers when homogenized at pH 9 with four model oils: n-dodecane, methyl myristate, isononyl isononanoate, and sunflower oil. In each case, very stable oil-in-water (33) Kettlewell, S. L.; Schmid, A.; Fujii, S.; Dupin, D.; Armes, S. P. Langmuir 2007, 23, 11381–11386. (34) Schmid, A.; Fujii, S.; Armes, S. P.; Leite, C. A. P.; Galembeck, F.; Minami, H.; Saito, N.; Okubo, M. Chem. Mater. 2007, 19, 2435–2445.
Langmuir 2010, 26(23), 18039–18048
Figure 4. X-ray photoelecton spectra recorded for (A) PEI alone, (B) a charge-stabilized polystyrene latex (entry 1, Table 1), (C) a nonquaternized PEI-stabilized polystyrene latex (entry 2, Table 1), and (D) a partially quaternized PEI-stabilized polystyrene latex (entry 3, Table 1).
emulsions were obtained, as determined via conductivity measurements and also the drop test. However, it is perhaps worth emphasizing that emulsions were not obtained if the aqueous solution pH was either 7 or 3. Presumably, the PEI chains (pKa = 7.24)35 are appreciably protonated at these lower pH values and do not wet the oil droplets in their polyelectrolyte form. Moreover, unlike the poly(glycerol monomethacrylate)-stabilized polystyrene latexes recently reported by Thompson and Armes,28 the adsorption of these PEI-stabilized latex particles onto the oil droplets is not particularly efficient. On allowing the emulsions formed at pH 9 to cream on standing, visual inspection of the lower (35) Han, J.; Kim, S. K.; Lee, J. C.; Joung, H. S. Macromol. Res. 2004, 12, 501–506.
DOI: 10.1021/la103804y
18043
Article
aqueous phase confirmed that it always contained some excess nonadsorbed latex, regardless of the initial latex concentration. The typical relationship between the mean droplet diameter and the initial latex concentration is shown in Figure 5 for n-dodecane homogenized with the PEI-stabilized polystyrene latex prepared in the presence of 4-VBC (entry 3 in Table 1). As expected, the mean oil droplet diameter is systematically reduced from more than 300 μm to around 74 μm as the initial latex concentration is increased. A similar trend has been recently observed by Thompson et al. for Pickering emulsions prepared under the same conditions using the same oil with poly(glycerol monomethacrylate)stabilized polystyrene latexes.36
Figure 5. Relationship between initial latex concentration and mean droplet diameter for Pickering emulsions prepared with PEI-PS latex particles (entry 3, Table 1) and n-dodecane as the oil phase (homogenization at 12 000 rpm for 2 min at 20 °C).
Walsh et al.
Three commercially available polymeric reagents (PPG-TDI, PPG-DGE, and PEG-DGE) were evaluated for cross-linking the PEI stabilizer chains on the latex surface so as to convert Pickering emulsions into colloidosomes (Figure 6). These polymeric reagents are relatively cheap, nonvolatile and much less hazardous than the equivalent small-molecule reagents. In principle, PPGTDI should react with the primary and secondary amine groups on the PEI chains to form robust urea linkages -NHCONH-, whereas the two bisepoxy-based cross-linkers should form -CHOHCH2NH- linkages. Both PPG-TDI and PPGDGE are oil-soluble. Thus the cross-linking reaction should be confined within the oil droplet phase, which should eliminate any possibility of intercolloidosome cross-linking28 (Figure 6). In contrast, the PEG-DGE cross-linker is water-soluble and hence resides solely in the aqueous continuous phase. Hence this reagent offers the opportunity to examine how the spatial location of the cross-linker affects the efficacy of colloidosome formation. It was anticipated that PEG-DGE cross-linking might cause significant intercolloidosome cross-linking in high solids. Moreover, given that not all of the PEI-stabilized latex particles are actually adsorbed onto the oil droplets, there is also the additional possibility of cross-linking between these excess latex particles, resulting in the formation of unwanted floccs in the continuous phase. Table 2 summarizes the results obtained for the preparation of colloidosomes from the precursor Pickering emulsions using each of the three polymeric reagents in turn. Successful cross-linking was assessed by an alcohol challenge, whereby a small sample of the colloidosome suspension was diluted with excess 2-propanol and then inspected by both optical microscopy (Figure 7) and scanning electron microscopy (Figure 8). When the precursor Pickering emulsion was subjected to an alcohol challenge in the absence of any polymeric cross-linker, no microcapsules were observed by optical microscopy and only ill-defined latex debris was visible by SEM (Figure 8a). Figures 8b and 8c show SEM images of colloidosomes prepared with either the PPG-DGE or
Figure 6. Reaction scheme for the formation of cross-linked colloidosomes at 50 vol % using a PEI-stabilized polystyrene latex at pH 9 and 20 °C. The top route represents colloidosomes formed from an o/w Pickering emulsion using an oil-soluble cross-linker (either PPG-TDI or PPG-DGE) dissolved in the interior oil phase. The bottom route represents colloidosomes formed from an o/w Pickering emulsion using a water-soluble cross-linker (PEG-DGE) dissolved in the aqueous continuous phase. 18044 DOI: 10.1021/la103804y
Langmuir 2010, 26(23), 18039–18048
Walsh et al.
Article
Table 2. Summary of o/w Colloidosomes Formed Using a 176 nm (Entry 1, Table 1) PEI-PS Latex at 1.0 wt %, pH 9 with the PPG-TDI (0.020 g), PPG-DGE (0.020 g), or PEG-DGE (0.020 g) Cross-Linker with Four Model Oilsa mean droplet diameter (μm) entry no.
cross-linker type
oil type
20 °C
0 °C
1 PPG-TDI sunflower oil 111 ( 50 2 PPG-TDI isononyl isononanoate 393 ( 211 3 PPG-TDI n-dodecane demulsified 4 PPG-TDI methyl myristate demulsified 5 PPG-DGE sunflower oil 73 ( 25 6 PPG-DGE isononyl isononanoate 45 ( 47 7 PPG-DGE n-dodecane 172 ( 191 8 PPG-DGE methyl myristate 54 ( 26 9 PEG-DGE sunflower oil 99 ( 34 10 PEG-DGE isononyl isononanoate 55 ( 24 11 PEG-DGE n-dodecane 181 ( 53 12 PEG-DGE methyl myristate 72 ( 48 a The mean diameter was determined by a Malvern Mastersizer. Homogenization was achieved at 12 000 rpm for 2.0 min.
158 ( 104 128 ( 51 110 ( 44
Figure 7. Optical microscopy images taken before and after an alcohol challenge (left) and corresponding Mastersizer particle size distributions of the original o/w emulsions prior to this alcohol challenge (right). (A) Sunflower oil-in-water emulsion with PPG-DGE dissolved in the oil phase (entry 5, Table 2). (B) Sunflower oil-in-water emulsion with PPG-TDI dissolved in the oil phase (entry 1, Table 2). (C) Sunflower oil-in-water emulsion with PEG-DGE dissolved in the aqueous continuous phase (entry 9, Table 2). An alcohol challenge was used for the qualitative assessment of successful cross-linking in each case.
PPG-TDI cross-linkers using the PEI-stabilized latex (entry 3 in Table 1). Comparing these two images, colloidosomes obtained using the former cross-linker appear to be more well-defined than colloidosomes formed using the latter cross-linker. A lowermagnification image of the PPG-DGE colloidosome sample shows a large number of cross-linked latex superstructures, suggesting that this is an efficient route for the production of colloidosomes (Figure 8d). The alcohol challenge protocol was also used to estimate the timescale required for effective cross-linking to occur at room temperature. Aliquots (0.50 mL) of the Pickering (36) Thompson, K. L.; Armes, S. P.; Howse, J. R.; Ebbens, S.; Ahmad, I.; Zaidi, J. H.; Burdis, J. A.; York, D. W. Macromolecules 2010, accepted for publication.
Langmuir 2010, 26(23), 18039–18048
emulsions containing the polymeric cross-linker were quenched with excess 2-propanol at various time intervals until intact colloidosomes were observed by optical microscopy. As expected, PPG-TDI reacted very rapidly with the primary and secondary amine groups located on the PEI stabilizer. After an alcohol quench, intact colloidosomes could be observed by optical microscopy as soon as emulsification was complete at 20 °C. Hence cross-linking of the stabilizer chains occurs during homogenization, which may prevent the formation of a well-defined latex monolayer around the oil droplets. This could explain the less well-defined shape of colloidosomes typically produced with PPG-TDI and also the unforeseen difficulty of forming stable emulsions with this cross-linker with certain oils (e.g., n-dodecane and methyl myristate; see Table 2). In hindsight, these unexpected DOI: 10.1021/la103804y
18045
Article
Walsh et al.
Figure 9. FT-IR spectra for (A) pristine PEI, (B) a 662 nm chargestabilized PS latex (entry 1, Table 1), (C) a 176 nm PEI-stabilized PS latex (entry 3, Table 1), (D) a PPG-TDI cross-linker, and (E) dried cross-linked colloidosomes prepared by the addition of PPG-TDI cross-linker to a sunflower oil-in-water Pickering emulsion (entry 1, Table 2). Note the disappearance of the isocyanate band at 2270 cm-1 in the latter colloidosome spectrum.
Figure 8. SEM images of (A) the latex debris that is obtained when colloidosome formation is attempted in the absence of any crosslinker; (B) a colloidosome prepared using the PPG-DGE crosslinker (entry 5, Table 2); (C) a colloidosome prepared using the PPG-TDI cross-linker (entry 1, Table 2); (D) a low-magnification image of colloidosomes prepared using the PPG-DGE cross-linker (entry 5, Table 2); (E) a colloidosome sample prepared using the PPG-TDI cross-linker at 0 °C (entry 3, Table 2); and (F) a colloidosome sample prepared using the PEG-DGE cross-linker (entry 9, Table 2).
observations are likely to be related to differences in oil viscosity. Thus, n-dodecane is much less viscous than either sunflower oil or isononyl isononanoate, hence the PPG-TDI cross-linker can diffuse more quickly to the surface of the oil droplets in the former case, which presumably leads to substantial cross-linking prior to the formation of a well-defined latex monolayer. Therefore, colloidosome syntheses using PPG-TDI were repeated at 0 °C using an ice bath to reduce the rate of in situ cross-linking. This strategy proved to be more successful: stable emulsions/colloidosomes were formed with three of the four model oils (the exception was methyl myristate because this oil has an mp of ∼18 °C). A representative micrograph of one of these colloidosomes (prepared using n-dodecane as the oil phase at 0 °C) is shown in Figure 8E. Thompson and Armes dissolved their poly(glycerol monomethacrylate)-stabilized polystyrene latex-based colloidosomes in d5-pyridine and then used 1H NMR spectroscopy to quantify the amount of PPG-TDI present.28 Unfortunately, no solvent could be identified that was capable of dissolving the PEIstabilized polystyrene latex-based colloidosomes prepared in the present work. This is presumably because these latexes are more highly cross-linked because of the presence of multiple quaternized 4-VBC groups per PEI chain. For this reason, colloidosome formation was also attempted using the PEI-stabilized polystyr18046 DOI: 10.1021/la103804y
ene latex prepared in the absence of 4-VBC (entry 2 in Table 1). Unfortunately, this latex proved to be a relatively poor Pickering emulsifier, producing rather unstable emulsions under otherwise identical conditions. This suggests that the presence of 4-VBC is essential to increasing the surface concentration of the PEI chains and hence enabling the PEI-stabilized latex particles to become effective Pickering emulsifiers. The cross-linking reaction between the terminal isocyanate groups on the PPG-TDI chains and the amine groups on the PEI stabilizer should result in urea bond formation. FT-IR spectroscopy studies of dried PPG-TDIsynthesized colloidosomes isolated after an alcohol quench (which is known to remove any excess unreacted PPG-TDI)28 confirmed the presence of the expected urea linkage because a strong CdO stretch was observed at 1743 cm-1 (Figure 9). This assignment was confirmed by a control experiment whereby a monoamineterminated PPG (Jeffamine M-2005) and the PPG-TDI crosslinker were mixed together in n-dodecane. An FT-IR spectrum recorded for these reacted polymers contains a carbonyl stretch at 1743 cm-1, which is in excellent agreement with the above observation of a urea bond in the colloidosome. Moreover, the absence of the isocyanate peak at 2270 cm-1 confirmed that all of the isocyanate groups on PPG-TDI had reacted in this model reaction (Figure S1). For a specified set of conditions, Thompson and Armes reported that PPG-TDI cross-linking of poly(glycerol monomethacrylate)stabilized polystyrene latex particles adsorbed onto oil droplets typically required around 20 min for the unstirred Pickering emulsion after homogenization at 20 °C.28 However, in the present study we found that robust colloidosomes could not be obtained using the PPG-DGE cross-linker in combination with the same hydroxy-functionalized polystyrene latex under identical conditions as judged by an alcohol challenge (data not shown). Thus, the PPGDGE cross-linker reacts too slowly with hydroxyl groups for successful colloidosome formation at room temperature. In principle, the amine-based PEI stabilizer chains should be much more reactive toward the terminal epoxy groups on the PPG-DGE cross-linker Langmuir 2010, 26(23), 18039–18048
Walsh et al.
Figure 10. Volume-average particle size distributions obtained for Pickering emulsions before and 48 h after cross-linking. (Top) A sunflower oil-in-water Pickering emulsion using 176 nm latex at 1.0 wt %, pH 9 (entry 3, Table 1) with the PPG-DGE cross-linker dissolved in the oil phase. Cross-linking was complete ∼15 h after homogenization at 20 °C on the basis of alcohol challenge experiments. (Bottom) A sunflower oil-in-water Pickering emulsion prepared using 176 nm latex at 1.0 wt %, pH 9 with the PEGDGE cross-linker dissolved in the aqueous phase. Cross-linking was complete ∼3 h after homogenization at 20 °C on the basis of alcohol challenge experiments.
than the poly(glycerol monomethacrylate) stabilizer chains. This indeed proved to be the case. The combination of the PEIstabilized latex with the PPG-DGE cross-linker led to stable emulsions with all four oils. However, cross-linking with this new polymeric reagent was considerably slower than that achieved with PPG-TDI, with intact colloidosomes being obtained approximately 12 h after homogenization at 20 °C, as judged by optical microscopy following an alcohol challenge. Again, the kinetics of this reaction could not be monitored by 1H NMR spectroscopy because the multiple 4-VBC groups per PEI chain led to a relatively high level of cross-linking within the latex particles, which prevented their dissolution in an appropriate NMR solvent. Laser diffraction studies were used to assess whether cross-linking causes in situ aggregation of the colloidosomes. In principle, such aggregation should be minimal for PPG-DGE because this crosslinker is confined within the oil droplets, whereas it may be problematic for water-soluble PEG-DGE because this cross-linker is located in the aqueous continuous phase. Figure 10 shows the volume-average droplet diameter, D4/3, obtained for a sunflower oil-in-water Pickering emulsion prepared with PPG-DGE dissolved in the oil phase immediately (0 h) after homogenization and also 48 h after homogenization. Immediately after emulsification, cross-linking had not yet occurred because no colloidosomes could be observed by optical microscopy after an alcohol challenge. Thus, the droplet size distribution (D4/3 = 80 ( 26 μm) obtained at 0 h simply represents the precursor Pickering emulsion. After 48 h, successful colloidosome formation was confirmed by an alcohol challenge but the droplet size distribution (D4/3 = 81 ( 26 μm) remained essentially unchanged. This indicates that the crosslinker was indeed confined within the oil droplets because little or no intercolloidosome cross-linking occurred. To examine whether similar results could be obtained for the equivalent cross-linker confined to the aqueous continuous phase, a sunflower oil-in-water Pickering emulsion was produced, followed by the addition of the water-soluble PEG-DGE cross-linker. In this case, the formation of robust colloidosomes is somewhat faster, requiring approximately 3 h at 20 °C. A representative micrograph of one of these colloidosomes is shown in Figure 8F. This reaction time is significantly shorter than that required for the oil-soluble PPG-DGE Langmuir 2010, 26(23), 18039–18048
Article
cross-linker. This is no doubt partially due to the slightly higher PEG-DGE concentration used, because a fixed mass of 0.020 g of cross-linker was utilized in each experiment and the PEG-DGE cross-linker has a marginally lower molecular weight than the PPGDGE cross-linker (520 vs 646 g mol-1). However, a more significant factor is probably related to the latex contact angle. Although this parameter is not known for this system, the formation of an oil-in-water emulsion means that the latex contact angle must be less than 90°.37 Thus, a larger volume fraction of each adsorbed latex particle is exposed to the aqueous phase compared to the oil phase. Given that emulsification was conducted using 50 vol % oil, most of the primary and secondary amine groups on the stabilizer chains must be available for reaction with the water-soluble PEG-DGE, whereas only a minority can be available for the oil-soluble PPG-DGE. Moreover, the minor fraction of PEI chains exposed to the oil phase will be relatively nonsolvated, whereas the major fraction of PEI chains exposed to the water phase will be highly solvated and hence more accessible to the (water-soluble) cross-linker. These two factors are likely to account for most of the observed difference in the rates of reaction of the PPG-DGE and the PEG-DGE cross-linkers. Laser diffraction studies indicate a modest increase in D4/3 from 95 ( 34 μm before cross-linking to 100 ( 49 μm after crosslinking (Figure 10). This suggests some degree of colloidosome aggregation, although this problem was perhaps not as extensive as originally anticipated. Surprisingly, there is also no evidence of cross-linking between the excess latex particles in the aqueous phase of the creamed emulsion, because no significant flocculation was detected by either dynamic light scattering or disk centrifuge studies. Thus, it does not appear to be essential for the polymeric cross-linker to be confined to within the oil droplets for successful microcapsule formation, even if this covalent stabilization chemistry is conducted at 50 vol %. This finding may be important for the development of next-generation colloidosomes.
Conclusions Aqueous emulsion polymerization of styrene in the presence of lightly quaternized PEI yields significantly smaller sterically stabilized polystyrene latex particles than those synthesized with unmodified PEI or without any PEI. Nitrogen microanalyses indicate a higher PEI content for the PEI-stabilized latex prepared in the presence of 4-VBC. Moreover, both aqueous electrophoresis and XPS confirm a higher surface concentration of PEI chains in this case, which suggests that the presence of pendent polymerizable styrene groups leads to efficient in situ grafting of the PEI chains onto the latex surface. This polyamine-functionalized sterically stabilized latex proved to be an effective Pickering emulsifier for the stabilization of oil-in-water emulsions prepared using four model oils. The adsorbed latex particles were covalently crosslinked at the oil droplet interface using three commercially available polymeric cross-linkers: PPG-TDI, PPG-DGE, and PEGDGE. The first two reagents are oil-soluble, which necessarily restricts the cross-linking chemistry to within the droplet phase and allows colloidosomes to be produced at 50 vol % without any intercolloidosome aggregation, as judged by laser diffraction studies. The latter reagent is water-soluble, which allows the effect of the physical location of the cross-linker to be studied. Surprisingly, cross-linking from the aqueous continuous phase also proved to be successful, with relatively little evidence of colloidosome aggregation at 50 vol %. In each case, these colloidosomes could withstand the removal of their internal oil phase on washing with excess alcohol. In contrast, precursor Pickering emulsions (37) Binks, B. P. Curr. Opin. Colloid Interface Sci. 2002, 7, 21–41.
DOI: 10.1021/la103804y
18047
Article
prepared in the absence of any cross-linker always disintegrated when challenged with alcohol, as expected. Using the PPG-TDI cross-linker, we found that colloidosomes were produced at 20 °C only when using sunflower oil and isononyl isononanoate. This cross-linker reacted very quickly with the PEI stabilizer chains, forming colloidosomes within 2 min under the stated conditions. This rapid rate of cross-linking may well hinder the formation of colloidosomes because the latex particles may not be able to form ordered monolayers around the oil droplets before cross-linking occurs. This presumably accounts for the demulsification that is observed in the case of n-dodecane and methyl myristate. This problem can be resolved, at least for dodecane, by conducting the emulsification at 0 °C in order to slow down the reaction between PPG-TDI and PEI. PPG-DGE was preferred to PPG-TDI as an oil-soluble cross-linker because colloidosomes were formed with all four oils at room temperature in this case, although
18048 DOI: 10.1021/la103804y
Walsh et al.
successful cross-linking required up to 12 h at 20 °C. The watersoluble PEG-DGE cross-linker proved to be somewhat faster, with robust colloidosomes forming 3 h after homogenization at 20 °C under the stated conditions. Acknowledgment. We thank P & G Technical Center (Newcastle-upon-Tyne, U.K.) for a CASE award to support an EPSRC Ph.D. studentship for K.L.T., for a three-month summer vacation internship for A.W., and for permission to publish this work. Supporting Information Available: FT-IR spectra recorded for the model reaction between Jeffamine M-2005 and the PPG-TDI cross-linker conducted in n-dodecane. This material is available free of charge via the Internet at http:// pubs.acs.org.
Langmuir 2010, 26(23), 18039–18048
