Pickering Emulsions Stabilized by Nanoparticles with Thermally

Sep 10, 2010 - Trishna Saigal,† Hongchen Dong,‡ Krzysztof Matyjaszewski,‡ and Robert D. Tilton*,†,§. †Department of Chemical Engineering, â...
0 downloads 0 Views 4MB Size
Download PDF
pubs.acs.org/Langmuir © 2010 American Chemical Society

Pickering Emulsions Stabilized by Nanoparticles with Thermally Responsive Grafted Polymer Brushes Trishna Saigal,† Hongchen Dong,‡ Krzysztof Matyjaszewski,‡ and Robert D. Tilton*,†,§ †

Department of Chemical Engineering, ‡Department of Chemistry, and §Department of Biomedical Engineering, Center for Complex Fluids Engineering, Carnegie Mellon University, Pittsburgh, Pennsylvania 15213-3890 Received July 12, 2010. Revised Manuscript Received August 27, 2010

A study is presented of emulsification by silica nanoparticles with poly(2-(dimethylamino)ethyl methacrylate) brushes grafted from their surfaces (SiO2-PDMAEMA) by atom-transfer radical polymerization (ATRP). The grafted nanoparticles were used to stabilize xylene-in-water and cyclohexane-in-water Pickering emulsions. PDMAEMA is a water-soluble weak polyelectrolyte with a pH-dependent lower critical solution temperature (LCST). Accordingly, SiO2-PDMAEMA nanoparticles were thermally responsive, as shown by the fact that they displayed a critical flocculation temperature (CFT) when heated. ATRP provides a high degree of control over the brush grafting density and degree of polymerization, two of the principal variables examined in this study. The effects of the solvent quality of the “oil” for the PDMAEMA brush were studied in addition to the effects of aqueous pH, ionic strength, and temperature relative to the CFT. The preferred emulsion type was oil in water in all cases. The lowest grafting density particles (0.077 chains/nm2) proved to be the most efficient and robust emulsifiers, producing stable emulsions using as little as 0.05 wt % particles in the aqueous phase and successfully emulsifying over a broader range of solution conditions than for the higher grafting density particles (0.36 and 1.27 chain/nm2). Both good (xylene) and poor (cyclohexane) solvents could be emulsified, but the poor solvent could be emulsified over a broader range of conditions than the good solvent. Emulsions have been stable for over 13 months, and some have dispersed as much as 83 vol % oil in the emulsion phase. Thermally responsive emulsions were created with the SiO2-PDMAEMA particles such that stable emulsions prepared at low temperature were rapidly broken by increasing the temperature above the CFT.

1. Introduction Emulsions stabilized by colloidal particles adsorbed at the oil/ water interface, known as Pickering emulsions, have been recognized since 1903.1,2 They are distinguished by their ability to disperse high-discontinuous-phase volume fractions (up to ∼80%) and by their unusually high stability against coalescence and sedimentation (or creaming).3-8 Potential technological applications of Pickering emulsions have renewed interest in this topic over the past decade. For example, because many surfactants that are commonly used as emulsifiers are tissue irritants, there is interest in replacing conventional emulsions with Pickering emulsions for medical, personal care, and food product applications. In addition to using Pickering emulsions as a final product, the 2-D colloidal assemblies that reside at the oil/water interfaces in these emulsions can be used as templates for new materials, such as microscale organic/inorganic composite *Corresponding author. Tel: 1-412-268-1159. Fax: 412-268-7139. E-mail: [email protected].

(1) Ramsden, W. Proc. R. Soc. 1903, 72, 156. (2) Pickering, S. U. J. Chem. Soc. 1907, 91, 2001–2021. (3) Binks, B. P.; Lumsdon, S. O. Phys. Chem. Chem. Phys. 2000, 16, 2539–2547. (4) Binks, B. P.; Lumsdon, S. O. Phys. Chem. Chem. Phys. 1999, 1, 3007. (5) Midmore, B. R. Colloids Surf., A 1998, 132, 257. (6) Vignati, E.; Piazza, R.; Lockhard, T. P. Langmuir 2003, 19, 6650–6656. (7) Giermanska-Kahn, J.; Schmitt, V.; Binks, B. P.; Leal-Calderon, F. Langmuir 2002, 18, 2515–2518. (8) Saleh, N.; Sarbu, T.; Sirk, K.; Lowry, G. V.; Matyjaszewski, K.; Tilton, R. D. Langmuir 2005, 21, 9873–9878. (9) Bon, S.; Chen, T. Langmuir 2007, 23, 9527–9530. (10) He, Y. Mater. Lett. 2005, 59, 114–117. (11) Zeng, C.; Bissig, H.; Dinsmore, A. D. Solid State Commun. 2006, 139, 547–556. (12) Strohm, H.; Lobmann, P. J. Mater. Chem. 2004, 14, 2667–2673. (13) Velev, O.; Furusawa, K.; Nagayama, K. Langmuir 1996, 12, 2374–2384.

15200 DOI: 10.1021/la1027898

particles,9 microfibers and films,10-12 and hollow particulate capsules sometimes referred to as colloidosomes.13-15 The unusually high stability of Pickering emulsions is derived mainly from the large adsorption energy of colloidal particles at an oil/water interface. This provides a large energy barrier against the particle desorption that must accompany droplet coalescence.16,17 Particle wettability arguments show that this energy scales with the square of the particle radius and can be as large as 107kT.16,17 Larger particles are more strongly confined to the interface than are nanoparticles,18 and micrometer-scale particles are normally more effective emulsifiers than nanoparticles. We recently showed that adsorbing surface-active polymers onto nanoparticles or grafting them from a nanoparticle surface makes the nanoparticles extremely efficient emulsifiers, with nanoparticle concentrations of as low as 0.04 wt % being sufficient to stabilize emulsions for many months.19 High emulsifying efficiency requires that the nanoparticles have a high-affinity adsorption isotherm at the oil/water interface, allowing a small total inventory of particles to stabilize a very large total surface area. (Supporting Information Figure S5 provides a simple illustrative calculation of the particle concentrations needed to stabilize emulsion droplets of a certain size for varying adsorption affinities.) The enhanced emulsification effectiveness of (14) Dinsmore, A. D.; Hsu, M. F.; Nikolaides, M. G.; Marquez, M.; Bausch, A. R.; Weitz, D. A. Science 2002, 298, 1006. (15) Aveyard, R.; Binks, B. P.; Clint, J. H. Adv. Colloid Interface Sci. 2003, 100-102, 503–546. (16) Pieranski, P. Phys. Rev. Lett. 1980, 45, 569–572. (17) Binks, B. P.; Lumsdon, S. O. Langmuir 2001, 17, 4540–4547. (18) Lin, Y.; Boker, A.; Skaff, H.; Cookson, D.; Dinsmore, A. D. Langmuir 2005, 21, 191–194. (19) Saleh, N.; Phenrat, T.; Sirk, K.; Dufour, B.; Ok, J.; Sarbu, T.; Matyjaszewski, K.; Tilton, R. D.; Lowry, G. V. Nano Lett. 2005, 5, 2489–2494.

Published on Web 09/10/2010

Langmuir 2010, 26(19), 15200–15209

Saigal et al.

polymer-grafted nanoparticles relative to that of bare particles is attributed to the surface activity of the polymer chains. Whereas bare particles do not decrease the interfacial tension, grafted polymer chains penetrate the oil/water interface and decrease the interfacial tension.8 This decreases the total surface energy and contributes a Gibbs elasticity to resist interfacial dilation during droplet-droplet contact. The mechanism of polymer-grafted nanoparticle emulsification has yet to be examined in detail, but we hypothesize that in addition to interfacial tension lowering, swollen polymer brushes on nanoparticles confined to the surfaces of two approaching droplets may impede droplet flocculation prior to coalescence via steric or electrosteric repulsive forces and impeded thin film drainage. Compared to soluble polymers or surfactants used as emulsifiers, grafting multiple chains to a nanoparticle core would have a cooperative effect on the per chain energy barrier against desorption, where complete removal of one nanoparticle-grafted chain from the oil/water interface would require the simultaneous removal of many additional chains that are grafted to the same nanoparticle. The desorption of one polymer-grafted particle corresponds to the simultaneous desorption of many polymer chains. The reduction in Laplace pressure due to the interfacial tension lowering provided by grafted polymer chains would also reduce the driving force for Ostwald ripening. Although theoretical studies suggest that interfacial flattening associated with particles at fluid interfaces could potentially eliminate Laplace pressure,20 Ashby and Binks have demonstrated that Ostwald ripening can still occur in Pickering emulsions.21 In the presence of grafted polymers, the overall particle adsorption energy is expected to depend on the relative polymer solvation energies in the two liquids comprising the emulsion. This may create a situation where two different oils that have similar viscosity, interfacial tension with water, and so forth may have significantly different emulsion characteristics, depending on their solvent quality for the grafted polymers. The current research focuses on the development of polymergrafted nanoparticles that are efficient emulsifiers but also provide pH- and ionic-strength-dependent thermal responsiveness. Responsive emulsions can be defined as emulsions that break abruptly in response to a change in conditions such as temperature, pH, or ionic strength. We use atom-transfer radical polymerization (ATRP)22-24 to irreversibly graft poly(2-(dimethylamino)ethyl methacrylate) (PDMAEMA) polyelectrolyte brushes from 20-nm-diameter silica nanoparticles (SiO2-PDMAEMA) with controlled degrees of polymerization and grafting densities.25 PDMAEMA is a thermally responsive, water-soluble weak polyelectrolyte with a pH-dependent lower critical solution temperature (LCST).26 The LCST increases with decreasing pH because of the protonation of tertiary amine groups. Polymer chain collapse above the LCST imparts the ability to thermally trigger abrupt changes in material morphology on the molecular scale, making thermally responsive polymers particularly interesting for many applications.27 Behrens and coworkers created a thermally responsive emulsion system using poly(N-isopropylacrylamide) (PNIPAM) microgel beads to (20) Abkarian, M.; Subramaniam, A.; Kim, S.; Larsen, R.; Yang, S.; Stone, H. Phys. Rev. Lett. 2007, 99, 188301–1-4. (21) Ashby, N. P.; Binks, B. P. Phys. Chem. Chem. Phys. 2000, 2, 5640–5646. (22) Matyjaszewski, K.; Xia, J. J. Chem. Rev. 2001, 101, 2921. (23) Matyjaszewski, K.; Tsarevsky, N. V. Nat. Chem. 2009, 1, 276. (24) Tsarevsky, N. V.; Matyjaszewski, K. Chem. Rev. 2007, 107, 2270. (25) Pyun, J.; Jia, S.; Kowalewski, T.; Patterson, G. D.; Matyjaszewski, K. Macromolecules 2003, 36, 5094. (26) Kusomo, A.; Bombalski, L.; Lin, Q.; Matyjaszewski, K.; Schneider, J.; Tilton, R. D. Langmuir 2007, 23, 4448–4454. (27) Galaev, I.; Mattiasson, B. Trends Biotechnol. 1999, 17, 335–340.

Langmuir 2010, 26(19), 15200–15209

Article

emulsify octanol in water.28 These emulsions could be triggered to phase separate when the microgel beads were thermally collapsed. Emulsion systems can also be rendered pH-responsive. St€over and co-workers used alumina-coated silica particles with reversibly bound potassium hydroxide phthalate to stabilize xylene-inwater emulsions that broke in response to a pH change.29 A weak polyelectrolyte, PDMAEMA can provide both pH and thermal responsiveness, and responsiveness is preserved when the polymers are attached to surfaces. For example, Zhou and co-workers showed that the hydrodynamic diameter of PDMAEMA-grafted silica particles depends on pH and temperature because of the configurational response of the grafted chains to changing solvent conditions.30 Kusumo and co-workers showed that PDMAEMA brushes grafted from gold surfaces displayed pH-triggered protein uptake and release behaviors.26 We find here that the SiO2-PDMAEMA nanoparticles tend to be less robust emulsifiers at elevated temperatures (produce stable emulsions under a more limited range of conditions at higher temperatures than at lower temperatures) and that stable emulsions prepared at low temperature are rapidly broken by increasing the temperature. There are many variables in the Pickering emulsion system presented in this article: particle concentration, grafting density of PDMAEMA chains, temperature during emulsification, solvent quality of the oil for the polymer and aqueous-phase pH and ionic strength. The motivation for investigating each variable follows. The effects of these variables are almost certainly intertwined in a complex manner. 1.1. Particle Concentration. This is a critical factor determining the stability of Pickering emulsions.17 The minimum concentration needed to stabilize Pickering emulsions is often ionic-strength-dependent, and particle concentrations of approximately 1-5 wt % are typically required.31,32 Previously, we found that silica nanoparticles with grafted poly(styrene sulfonate) brushes stabilized trichloroethylene-in-water emulsions for more than half a year at an aqueous-phase particle concentration of just 0.04 wt %.8 Other than this prior work, the lowest particle concentrations successfully reported to form stable emulsions are 0.2 wt % for the magnesium aluminum hydroxide emulsification of paraffin/water33 and 0.25 wt % hydrophobized fumed silica for the emulsification of water/toluene.3 High emulsification efficiency has obvious practical benefit in terms of minimizing the amount of material needed to manufacture emulsions. It also introduces new considerations for the balance of forces involved in emulsion stabilization. For example, one consequence of the high emulsification efficiency afforded by polymer-grafted nanoparticles is that the concentration of free, unadsorbed particles in the continuous phase of the emulsion may be much lower than would be the case for a less-efficient emulsifier. This would eliminate situations where interdroplet depletion attractions could weaken the emulsion stability against flocculation and/or coalescence. 1.2. Grafting Density and Grafted Polyelectrolyte Degree of Polymerization. By controlling the crowding and thus the configurational freedom of the grafted chains on the nanoparticle, the polymer grafting density should affect the ability of polymer chains to adapt to the interface (i.e., to reconfigure and penetrate (28) Ngai, T.; Auweter, H.; Behrens, S. H. Macromolecules 2006, 39, 8171–8177. (29) Li, J.; St€over, D. H. Langmuir 2008, 24, 13237–13240. (30) Zhou, L.; Yuan, W.; Yuan, J.; Hong, X. Mater. Lett. 2008, 62, 1372–1375. (31) Ashby, N. P.; Binks, B. P. Phys. Chem. Chem. Phys. 2000, 2, 5640–5646. (32) Golemanov, K.; Tcholakova, S.; Kalchevsky, S.; Ananthapadmanabhan, K. P.; Lips, A. Langmuir 2006, 22, 4968–4977. (33) Abend, S.; Bonnke, N.; Gutschner, U.; Lagaly, G. Colloid Polym. Sci. 1998, 276, 730–737.

DOI: 10.1021/la1027898

15201

Article

Saigal et al.

Figure 1. Proposed representation of SiO2-PDMAEMA particle adsorption at the oil/water interface, showing how chains crowded at high grafting density (a) are restricted in their ability to reconfigure and penetrate the interface compared to chains at a low grafting density (b). Expected configurations of a SiO2-PDMAEMA particle at an oil/water interface, with chains able to penetrate xylene, a good solvent (c), but not cyclohexane, a poor solvent for PDMAEMA (d).

the oil/water interface) and thereby increase the adsorption energy. In principle, polymer chains will be highly extended and highly restricted at extremely high grafting densities. Such particles may behave simply like rigid nanoparticles rather than like one whose surface can adapt to the oil/water interface. This could compromise the benefits of polymer grafting. Although extremely low grafting densities may provide high configurational freedom for chains to reconfigure and penetrate the oil/water interface, they might also present so few polymer chains as to have little noticeable effect, as illustrated schematically in Figure 1. Grafting densities of 1.27, 0.36, and 0.077 chains/nm2 were tested. For comparison, the cross-sectional area of a PDMAEMA chain is 0.67 nm2 on the basis of the density of bulk PDMAEMA (1318 kg/m3), the molecular weight of a DMAEMA unit (132 g/mol), and the length of a PDMAEMA repeat unit (0.25 nm) assuming a cylindrical shape. Thus, the largest possible grafting density would be 1.5 chains/nm2. The highest grafting density (1.27 chains/nm2), medium grafting density (0.36 chains/nm2), and lowest grafting density (0.077 chains/nm2) materials represent 85, 24, and 5% of the maximum possible grafting density, respectively. The degree of polymerization of the grafted chains affects the extent of interfacial penetration and determines the range of electrosteric repulsions. Thus, emulsification was tested using two different degrees of polymerization while fixing the grafting density at 0.077 chains/nm2. 1.3. Solvent Quality. The solvent quality of both phases for the grafted polymers affects how the chains penetrate the interface and partition between the two phases, which in turn influences the reduction of interfacial tension and particle adsorption energy. Xylene and cyclohexane were used to represent a good and a poor solvent, respectively, for the grafted PDMAEMA chains. Figure 1 also illustrates the expected influence of the oil solvent quality on polymer adaptation to the oil/water interface. Grafted chains on adsorbed particles can penetrate the good solvent but not the poor solvent. When the grafted chains are weak polyelectrolytes, the aqueous pH and ionic strength affect the chain solubility and configuration in water as well as the lateral dipole forces acting along the oil/ water interface and the normal electrosteric forces acting across interdroplet thin films. PDMAEMA exhibits a pH-dependent LCST in water, leading to a pH-dependent critical flocculation temperature (CFT) for SiO2-PDMAEMA nanoparticles suspended in water. The temperature-dependent change in solvent quality of the aqueous phase for the polymer would alter the particle adsorption energy. Furthermore, because particle flocculation plays an important role in stabilizing some Pickering emulsion systems, the dispersion state of the polymer-grafted nanoparticles above or below the CFT may be an important issue 15202 DOI: 10.1021/la1027898

in its own right. Flocculated particles may form a protective network around the droplets.15,33 The role of particle flocculation is system-dependent. In some cases, particle flocculation is required, and in other cases, it is not required to achieve emulsion stability.5,31 In some cases, emulsification is most effective at intermediate degrees of flocculation.4 Here we compared the importance of temperature relative to the CFT both by fixing the temperature during emulsification and changing the pH to change the CFT and also by fixing the pH and raising the temperature to the CFT and beyond. We also examined the stability of preformed emulsions upon heating to the CFT. Throughout this study, emulsions were characterized according to their dispersed-phase volume fraction, stability against coalescence, and preferred emulsion type (i.e., whether they form oil-in-water, o/w, or water-in-oil, w/o, emulsions at an overall water/oil ratio of 1:1). We also determined their susceptibility to catastrophic phase inversion. The preferred emulsion type is generally dictated by the relative wettability of the particle by the oil or water. Typically, Pickering emulsions follow the Bancroft rule whereby the phase that initially contains the dispersed particles becomes the continuous phase in the emulsion. Catastrophic phase inversion, exhibited by some Pickering emulsion systems, refers to the formation of the nonpreferred emulsion type when the overall water/oil ratio deviates from 1:1. Transitional phase inversion can also be driven by altering the particle wettability via pH or ionic strength changes.3,34 Emulsions that have catastrophically phase inverted are termed anti-Bancroft emulsions.

2. Experimental Section 2.1. Materials. Monomer 2-(dimethylamino)ethyl methacrylate (DMAEMA) (Aldrich, 98%) was purified by passage through a basic alumina column. Copper(I) chloride (Aldrich, 98þ %) was purified by washing sequentially with acetic acid and diethyl ether before being filtered and dried and was stored under nitrogen before use. The 1,1,4,7,10,10-hexamethyltriethylene tetramine (HMTETA) ligand (Aldrich, 97%), copper(II) chloride (Acros, 99%), and acetone (Pharmco-Aaper, 99%) were used as received. The procedure for the synthesis of 1-(chlorodimethylsilyl)propyl 2-bromoisobutyrate and the subsequent functionalization of the silica (30 wt % silica in methyl isobutyl ketone, 20 nm effective diameter, Nissan Chemical) followed the previously described method.25 All water was first deionized by reverse osmosis and purified to 18.2 MΩ 3 cm resistivity using a Barnstead Nanopure Diamond system. Cyclohexane (Fisher, 99%þ) and xylene (Acros, mixture of isomers, extra-pure grade) were used as received. (34) Binks, B. P.; Clint, J. H. Langmuir 2002, 18, 1270.

Langmuir 2010, 26(19), 15200–15209

Saigal et al.

Article Scheme 1. Synthesis of PDMAEMA Brushes Grafted from Silica Nanoparticles

Table 1. Synthesis of PDMAEMA-Grafted Silica Nanoparticles via Surface-Initiated ATRPa sample

ATRP initiator content on silica (mmol/g)b

Mn

Mw/Mnc

grafting density, σ (chain/nm2)

Dh in 1 mM NaCl(aq) (nm)

1 0.353 19 400 1.14 1.27 93 ( 29 2 0.118 17 570 1.31 0.36 71 ( 22 3 0.094 16 150 1.16 0.076 54 ( 9.3 4 0.094 36 150 1.14 0.077 62 ( 31 a The reaction conditions are as follows: SiO2-Br/DMAEMA/CuCl/CuCl2/HMTETA = 600:1:3:1:4 in acetone (1 vol of DMAEMA) at 30°C. b The total number of moles of Br (or ATRP initiator) per gram of nanoparticles was determined from elemental analysis. c Mn and Mw/Mn were determined by GPC with dimethylformamide as the eluent and poly(methyl methacrylate) calibration standards after cleaving grafted polymer chains by etching the silica nanoparticles in aqueous HF solutions. The number-average hydrodynamic diameter (Dh) of particles was measured in 1 mM NaCl by dynamic light scattering. The error reported is half the width at half-maximum of the number-average size distribution.

2.2. Synthesis of SiO2-PDMAEMA Nanoparticles. In the first stage of preparing hybrid nanoparticles, 20-nm-diameter colloidal silica particles were mixed with compound 1, 1-(chlorodimethylsilyl)-propyl 2-bromoisobutyrate. Compound 1 reacts with the surface silanols to yield silica nanoparticles fully functionalized with ATRP initiating groups (Scheme 1). These nanoparticles, denoted as SiO2-Br, were purified by precipitation into hexane several times and then redispersed in an ATRP reaction mixture described below to prepare polymers end-grafted from silica with a high grafting density (>0.5 chain/nm2). Meanwhile, the lower grafting density polymer brushes were achieved by using silica particles with low initiator densities, prepared by using a mixture of the “dummy” initiator, chlorotrimethylsilane, and the ATRP initiator (compound 1) in appropriate ratios. By using the surface-initiated ATRP technique, PDMAEMA chains were grafted from silica nanoparticle surfaces with controlled molecular weight and a narrow molecular weight distribution. The grafted chain molecular weight, molecular weight distribution, grafting density, and size of the resulting particles are summarized in Table 1. 2.3. Synthesis Procedures. In a typical surface-initiated ATRP reaction (sample 1), 1.50 g of initiator-modified silica nanoparticles (0.353 mmol Br/g silica; 0.530 mmol ATRP initiator sites) was placed in a 250 mL Schlenk flask containing a stir bar. The catalysts, CuCl (157.2 mg, 1.588 mmol) and CuCl2 (71.3 mg, 0.530 mmol), were added, and the reaction flask was thoroughly purged by vacuum and flushed with nitrogen for five cycles. Nitrogen-purged DMAEMA (53.5 mL, 0.318 mol), acetone (53.5 mL), and HMTETA (0.58 mL, 2.1 mmol) were added via syringe sequentially. The flask was then transferred to a thermostatted oil bath at 30 °C. Polymerization was stopped after 39.8 h by opening the flask and exposing the catalyst to air. The product was precipitated into cold methanol (800 mL) and washed with 200 mL of cold methanol. The precipitation and washing process was repeated twice. 2.4. Procedure for Cleaving PDMAEMA Chains from Nanoparticles. A 49% HF(aq) solution (1 mL) was added to 1 mL of a particle suspension in tetrahydrofuran (∼10 mg/mL), and the reaction was allowed to stir at room temperature overnight in a polypropylene flask. An ammonium hydroxide solution (28.0-30.0% NH3 basis) was very slowly added to the mixture in Langmuir 2010, 26(19), 15200–15209

an ice bath until the pH increased to above 8. The cleaved PDMAEMA polymer was recovered by evaporating the tetrahydrofuran phase in air, and then a DMF solution of the recovered polymer was injected into the GPC for molecular weight and polydispersity analysis. Hydrofluoric acid (HF) liquid and vapor are extremely hazardous. HF is poisonous and corrosive. The use of proper handling and personal protective equipment specified in the material safety data sheet is critical. 2.5. Polymer Analysis. Elemental analysis was conducted by Midwest Microlab (Indianapolis, IN) to determine the ATRP initiator amount on modified silica nanoparticles. On the basis of elemental analysis, there were 2.82, 0.94, and 0.75 wt % Br per gram of modified silica nanoparticles, corresponding to 0.353, 0.118, and 0.094 mmol of Br (or ATRP initiator site)/g of SiO2, respectively. The number-average molecular weight Mn and molecular weight distribution Mw/Mn of grafted PDMAEMA were measured by a GPC system consisting of a Waters 510 HPLC pump, three Waters Ultrastyragel columns (500, 103, and 105 A˚), and a Waters 410 DRI detector with DMF as an eluent (50 °C, flow rate = 1.0 mL/min). Linear poly(methyl methacrylate) standards were used for calibration. To determine the total mass of grafted polymer, thermogravimetric analysis (TGA) experiments were performed with grafted particle samples placed in aluminum pans using a Polymer Laboratories TG1000 instrument operating in the 20-580 °C temperature range, under nitrogen, at a heating rate of 20 °C/min. Chain grafting densities were calculated from the total polymer mass determined as the total organic mass by TGA, using the weightaverage molecular weight determined by GPC and assuming that the silica particles were all spheres of diameter 20 nm and density 2.07 g/cm3. 2.6. Critical Flocculation Temperature. The critical flocculation temperature was measured for all SiO2-PDMAEMA particles at conditions ranging from pH 7.5 to pH 10 in aqueous solutions of 1, 10, or 100 mM NaCl. Particle suspensions were made at 0.5 mg/mL, and the pH was adjusted with 0.1 M HCl or 0.1 M NaOH. The optical density at 508 nm was measured by ultraviolet-visible spectrophotometry (Cary 300 Bio) with temperatures scanned at a rate of 1 °C/min from 20 to 100 °C. The inflection point in a plot of optical density versus temperature was taken as the CFT. DOI: 10.1021/la1027898

15203

Article

Saigal et al.

Figure 2. The CFT of SiO2-PDMAEMA particles is pH- and ionic-strength-dependent. Symbols represent particles with different grafting

densities (σ) and molecular weights Mn at various ionic strengths. For σ = 1.27 chains/nm2, Mn = 19 400: ()) 1 mM NaCl, (() 10 mM NaCl, and (2) 100 mM NaCl. For σ = 0.36 chains/nm2, Mn = 17 570: (Δ) 1 mM NaCl, (b) 10 mM NaCl, and (O) 100 mM NaCl. For σ = 0.076 chains/nm2, Mn = 16 150: () 1 mM NaCl, (þ) 10 mM NaCl, and (-) 100 mM NaCl. For σ = 0.077 chains/nm2, Mn = 36 150: (9) 1 mM NaCl, (*) 10 mM NaCl, and (0) 100 mM NaCl. Emulsions were made at pH 7.5 at 50 °C and at pH 9 at 20, 50, and 70 °C.

2.7. Emulsification. Emulsions were made at a 1:1 water/oil ratio, unless otherwise specified, using a sawtooth homogenizer (Biospec Tissue Tearor 985370-395). Several homogenization speeds and times were tested to emulsify xylene and water with the SiO2-PDMAEMA particles to find conditions under which the emulsified volume did not vary with further increases in speed or time. For all subsequent emulsions, the homogenization conditions were fixed at the same high shear rate for 60 s. Unless otherwise indicated, particles were dispersed in water before homogenization. After homogenization, the volume of each phase (emulsion, neat water, and/or neat oil) was measured to calculate the volume percentages of oil and water in the emulsion phase if one were produced. To determine the preferred emulsion type, a drop test was conducted by placing one drop of the emulsion phase into neat water and another drop into neat oil. An o/w emulsion droplet disperses readily in water but not in oil and vice versa. All emulsions were stored in closed vials at room temperature with the phase volumes recorded regularly. To characterize the emulsion stability, coalescence was monitored over time by measuring the volume of any neat oil phase that appeared and the downward motion of the emulsion/oil interface for an o/w emulsion. Emulsion droplets were imaged with an inverted brightfield microscope (Leica DMI 6000) at 40 magnification after a 1:10 dilution in the continuous-phase liquid for better optical clarity. 2.8. Colloidal Characterization. Xylene and cyclohexane were used to represent a good and a poor solvent, respectively, for grafted PDMAEMA. The particles were readily dispersed in xylene but could not be dispersed in cyclohexane at all temperatures considered in this study. Dynamic light scattering (Malvern Instruments Nanosizer DTS) was used to determine the hydrodynamic diameter of the SiO2-PDMAEMA particles in water (Table 1). The numberaverage diameter is reported, assuming a refractive index of 1.46 for the grafted nanoparticles. Equilibrium interfacial tension measurements were performed using a static microscale droplet imaging apparatus that is described in detail elsewhere.35,36 The apparatus used microscale oil droplets submerged in an aqueous nanoparticle suspension (or PDMAEMA solution for comparison). In this study, cyclohexane (35) Alvarez, N. J.; Walker, L. M.; Anna, S. L. Phys. Rev. E 2010, 82, 011604. (36) Alvarez, N. J.; Walker, L. M.; Anna, S. L. Langmuir 2010, 26, 13310–13319.

15204 DOI: 10.1021/la1027898

and xylene drops were generated at the end of a 30-μm-diameter capillary. Because the radius is small enough that gravity has little effect on the droplet shape, the interfacial tension is measured via a digitized fit to the droplet radius of curvature and a direct measurement of the Laplace pressure jump across the interface. For all reported interfacial tensions, the time dependence of the static droplet interfacial tension was monitored to ensure equilibration.

3. Results and Discussion 3.1. Particle Characteristics. Dynamic light scattering results reported in Table 1 for SiO2-PDMAEMA nanoparticles are consistent with a 20-nm-diameter silica core and their expected chain extensions. The measured hydrodynamic diameter of the high grafting density particles (sample 1, σ = 1.27 chains/nm2) is 93 nm. This is slightly larger than expected (approximately 80 nm) on the basis of the 20 nm silica core diameter and the 30 nm chain contour length for the reported Mn. This is most likely due to the sensitivity of hydrodynamic measurements to even small degrees of polydispersity, whereby the longest chains dominate the hydrodynamic layer thickness. The large apparent layer thickness is consistent with the strong polyelectrolyte stretching expected at such a high grafting density. For a nearly constant degree of polymerization (samples 1-3), the apparent layer thickness decreases with decreasing grafting density as expected. Sample 4 has the lowest grafting density and the highest molecular weight, yet it has a smaller hydrodynamic diameter than sample 1 or 2. This is consistent with the fact that chain stretching is weaker at the lower grafting density and thus produces a smaller polymer layer thickness than it would at higher grafting densities. The apparent layer thickness at constant grafting density (samples 3 and 4) increases with increasing degree of polymerization, again as expected. 3.2. Critical Flocculation Temperature. Figure 2 shows the pH and ionic strength dependence of the SiO2-PDMAEMA nanoparticle CFT in an aqueous suspension. At a constant ionic strength, the CFT decreased with increasing pH for all particle grafting densities. As the pH increases, the DMAEMA solubility and thus the SiO2-PDMAEMA stability in aqueous suspensions are decreased by the deprotonation of the amine groups. Langmuir 2010, 26(19), 15200–15209

Saigal et al.

Article

Figure 3. Success or failure of emulsification using SiO2-PDMAEMA particles dispersed at 2 wt % in water. Closed symbols indicate the formation of a stable emulsion; open symbols indicate conditions under which no emulsion formed. Both graphs display data using the four different particle types with different grafting characteristics to create (a) xylene-in-water emulsions and (b) cyclohexane-in-water emulsions. Emulsification results are mapped according to the particle grafting density and the temperature of emulsification normalized by the CFT. For specific conditions of pH, ionic strength, and temperature, see Supporting Information Tables S1-S4. For T/CFT < 1, both pH 7.5, 50 °C and pH 9, 20 °C conditions with1, 10, and 100 mM NaCl are considered. The 1 < T/CFT < 1.045 region contains data exclusively for pH 9, 50 °C, and the T/CFT >1.045 region includes data exclusively for pH 9, 70 °C. The different symbol shapes represent different ionic strengths: (b, O) 1 mM NaCl, (2, Δ) 10 mM NaCl, and (9, 0) 100 mM NaCl.

3.3. Emulsification Using High Grafting Density Particles. Figure 3 summarizes the results of preferred emulsion-type experiments for each of the four SiO2-PDMAEMA particle samples. Prior to homogenization, particles were dispersed in water at a concentration of 2 wt % to ensure a sufficient particle concentration to test the effects of ionic strength, temperature, and pH on emulsification. The minimum required particle concentration for each particle type is addressed below. Figure 3 maps conditions that produced stable emulsions and those that failed to produce emulsions for each particle type. Of all the experimental condition variables examined, the solvent type and the temperature relative to the CFT were the most influential. Results in Figure 3 are therefore presented separately for cyclohexane and xylene and are organized according to the temperature during emulsification normalized by the CFT. Organizing by T/CFT encompasses all conditions of pH, temperature, and ionic strength. The temperature was manipulated to fall below, at, or above the CFT at pH 9. Temperature was not varied at pH 7.5, but T/CFT did vary by virtue of changes in ionic strength. To simplify the presentation of results for conditions below the CFT, the results represent a mix of pH 7.5, 50 °C and pH 9, 20 °C at the various ionic strengths. The different symbol shapes represent different ionic strengths. Conditions for each individual emulsion sample, the relative volumes of the emulsion phase, neat water or neat oil, and the volume percent of oil in the emulsion phase are tabulated in the Supporting Information (Tables S1-S4). All formed emulsions contained at least 50 vol % oil in the emulsion phase. The highest discontinuous phase dispersal in an emulsion was over 80 vol %. The preferred emulsion type for all emulsions created was o/w (oil in water). For each emulsion, 100% of the oil was emulsified and the emulsion phase coexisted with a neat water phase. For high grafting density particles, when emulsifying below the CFT, there was no significant difference whether this condition was achieved by decreasing the pH to 7.5 at 50 °C or decreasing the temperature to 20 °C at pH 9. All ionic strength conditions for which T < CFT produced stable emulsions in which all of the oil was emulsified. Also, the solvent quality had no influence as there was no difference in the emulsification of xylene (good PDMAEMA solvent) or cyclohexane (poor PDMAEMA solvent) when operating below the CFT. The emulsions formed at T < CFT have been stable against coalescence, with no formation of a neat oil phase for 13 months as of this writing. Specific sample Langmuir 2010, 26(19), 15200–15209

details with the duration of stability are reported in the Supporting Information. Slightly above the CFT (i.e., at 50 °C and pH 9), the emulsion properties did depend on ionic strength for the high grafting density particles. Under these conditions, only aqueous suspensions containing 10 mM NaCl produced emulsions. Both xylene and cyclohexane were emulsified. Solutions with either 1 or 100 mM NaCl produced no emulsions, either with xylene or cyclohexane. The CFT dependence on ionic strength is not monotonic. The CFT for 10 mM NaCl is greater than that for either 1 or 100 mM NaCl. Thus, in the presence of 10 mM NaCl, 50 °C corresponds to the lowest T/CFT, possibly suggesting that the main factor is simply the hydrophobicity of the particle. At higher T/CFT (50 °C with either 1 or 100 mM NaCl), the particle is more hydrophobic and emulsification was unsuccessful. Conversely, at 70 °C and pH 9 for 10 mM NaCl, only cyclohexane (poor solvent for PDMAEMA) could be emulsified, whereas at 100 mM NaCl (highest T/CFT) both cyclohexane and xylene were emulsified. NaCl (1 mM) still produced no emulsion for either oil. The occurrence of emulsification well above the CFT at 70 °C and pH 9 in the presence of 100 mM NaCl could be related to the stabilizing ability of nanoparticle flocs. Recalling that particle preflocculation influences the Pickering emulsion stability,4,5,15,31,33 it is also possible that the SiO2-PDMAEMA flocs that formed under these conditions that yielded the highest T/CFT are simply more effective stabilizers than the flocs formed under the other conditions for which T > CFT. It is reasonable to speculate that the flocs formed at the highest T/CFT would be the least dense and most extended if they formed by diffusion-limited aggregation in response to the greater strength of the interparticle attraction, but this has not been investigated. Adsorbed layers of more extended flocs could provide a stronger steric barrier against droplet contact. Whereas the solvent quality of the oil phase for PDMAEMA played no role below the CFT, it was an important factor above it for the high grafting density particles. At 70 °C, 10 mM NaCl (T/CFT = 1.075), an emulsion was created with cyclohexane, the poor solvent, but not with xylene, the good solvent. It is plausible that because the SiO2-PDMAEMA particles are readily dispersed in xylene, particles that reside temporarily at any interface formed during homogenization are preferentially ejected into the xylene phase because of their instability in water at elevated temperatures. Because the particles are not dispersible in DOI: 10.1021/la1027898

15205

Article

cyclohexane at any temperature considered here, this pathway is not available and flocculated SiO2-PDMAMEA particles may remain adsorbed to the cyclohexane/water interface. Thus, the increased hydrophobicity of the particles at elevated temperatures and the poor solvent quality of cyclohexane that inhibits particle transfer from the aqueous phase could be a stabilizing factor for cyclohexane-in-water emulsions. Dynamic light scattering measurements (DLS) of the neat xylene phase left after homogenization confirmed the presence of aggregated SiO2-PDMAEMA particles in xylene when emulsification was attempted at T/CFT > 1. Size distributions were extremely polydisperse, ranging from 200 to 1300 nm hydrodynamic diameter, consistent with the transfer of aggregated particles from the aqueous phase. For comparison, DLS measurements of “fresh” particles dispersed directly in xylene indicated a hydrodynamic diameter of 64 ( 27 nm. Below the CFT, particles are compatible with both xylene and water and are not as strongly driven to partition out of one phase or another. 3.4. Emulsification Using Medium Grafting Density Particles. The medium grafting density particles (σ = 0.36 chains/nm2, Mn = 17 570) exhibited similar emulsification behavior to the high grafting density particles below the CFT, but they produced emulsions at more conditions above the CFT, as shown in Figure 3. At 50 °C, cyclohexane emulsified at all ionic strengths but xylene did not emulsify at any ionic strength tested. As suggested above for high grafting density particles, this behavior could be attributable to the partitioning of particles into the xylene phase under conditions above the aqueous CFT. At 70 °C, both xylene and cyclohexane were emulsified at 100 and 1 mM NaCl. The latter corresponds to the highest value of T/CFT for this particle set. No emulsions were formed at 70 °C when the water contained 10 mM NaCl, even though this ionic strength was most favorable for cyclohexane emulsification with the high grafting density particles. Although this may seem surprising, the high and medium grafting density results at 70 °C and pH 9 are in fact consistent with each other. Because the particular ionic strength dependence of the CFT varied according to the particle grafting density, 10 mM NaCl corresponded to the middle value of CFT for the medium grafting density particles and 70 °C was an unfavorable condition for emulsification, just as 70 °C was unfavorable for emulsification at the middle value of the CFT provided by 1 mM NaCl with the high grafting density particles. In general, the medium grafting density particles were effective emulsifiers at more conditions than the high grafting density particles. This is probably due to the ability of the PDMAEMA chains to reconfigure more freely to adsorb and help anchor particles at the oil/water interface. 3.5. Emulsification Using Low Grafting Density Particles and the Effect of the PDMAEMA Degree of Polymerization. The low grafting density (σ = 0.076 chains/nm2, Mn = 16 150) particles proved to be the most robust emulsifiers because they emulsified both xylene and cyclohexane under all conditions tested above and below the CFT. This is consistent with the importance of the chain configurational freedom noted in the Introduction. The low grafting density particles are expected to have the most configurational freedom to rearrange and maximize their penetration of the oil/water interface as particles adsorb, as illustrated in Figure 1. For the low grafting density, we tested the importance of the PDMAEMA degree of polymerization using particles with σ = 0.077 chains/nm2 and Mn = 36 150, approximately doubling the degree of polymerization. Those results are mixed in with the results for σ = 0.076 chains/nm2, Mn = 16 150 in Figure 3. (Refer to Supporting Information for individual sample results.) These 15206 DOI: 10.1021/la1027898

Saigal et al.

particles also emulsified both solvents over the entire range of conditions tested above and below the CFT and gave similar dispersed oil volume fractions in the emulsion phase as the lower degree of polymerization. Thus, no degree of polymerization dependence was evident. Additionally, no difference has been observed in long-term stability for either degree of polymerization as of this writing. All emulsions have remained stable against coalescence for at least 7 months for σ = 0.077 chains/nm2, Mn = 36 150 and for at least 6 months for σ = 0.076 chains/nm2, Mn = 16 150. 3.6. Emulsification Efficiency: Particle Concentration Effect. The low grafting density particles were not only the most robust emulsifiers of all of the grafting densities tested but also the most efficient. Conditions below the CFT were the most favorable for the emulsification for any of the particles tested at 2 wt %. Thus, to compare the efficiency of emulsification for all grafting densities, we tested the minimum necessary particle concentration for creating stable emulsions with either xylene or cyclohexane at pH 7.5 and 10 mM NaCl at both 50 and 20 °C. This is well below the CFT for each particle type. The overall water/oil ratio was 1:1 in these experiments. Figure 4a,b maps emulsion stability on particle concentration and PDMAEMA grafting density coordinates. Particle concentrations are those of the initial aqueous suspension used for emulsification, except in Figure 4e, where particles were dispersed initially in xylene and the concentration is that of the initial xylene-based particle suspension. Both sets of low grafting density particles (Mn = 16 150 and 36 150) created stable emulsions at 0.05 wt % for both xylene and cyclohexane at 50 °C but not below this concentration. The medium grafting density particles required a concentration of 0.1 wt % to emulsify xylene and 0.05 wt % to emulsify cyclohexane. The high grafting density particles required 0.5 wt %, 10 times the particle concentration required for low grafting density particles, to emulsify either xylene or cyclohexane. The emulsification efficiency was not significantly dependent on the oil solvent quality for PDMAEMA at 50 °C. Stability maps at 20 °C, pH 7.5, and 10 mM NaCl are shown in Figure 4c,d. The emulsification efficiency was significantly greater at 20 °C than at 50 °C for the high grafting density particles. Stable xylene-in-water emulsions were created using 0.05 wt % of the high grafting density particles at 20 °C as opposed to 0.5 wt % needed at 50 °C. Similarly, to emulsify cyclohexane, the necessary high grafting density particle concentration was reduced 5-fold to 0.1 wt % at 20 °C. There was no change in the emulsification efficiency with the low grafting density particles (Mn = 36 150) when operating at 20 °C versus 50 °C. The decreased temperature of emulsification not only improved the efficiency of the high grafting density particles but also made it possible to create stable emulsions with the particles dispersed in xylene (Figure 4e). As will be discussed in section 3.9, emulsions did not form when the high grafting density particles were dispersed in xylene at 50 °C when the aqueous phase was at pH 7.5 and 10 mM NaCl. At 20 °C with all other conditions constant, emulsification efficiencies with high grafting density particles initially dispersed in xylene were similar to when the particles were initially dispersed in water. At 20 °C, the low grafting density particles (Mn = 36 150) also created emulsions when dispersed in xylene. This required a slightly higher concentration of 0.07 wt % (concentration in xylene) as compared to 0.05 wt % when dispersed in the aqueous phase. Stable emulsions made with particles dispersed in xylene at 20 °C destabilized and phase separated immediately when those emulsions were heated to 50 °C. This indicates that these emulsions were thermally responsive, even when the particles were not initially dispersed Langmuir 2010, 26(19), 15200–15209

Saigal et al.

Article

Figure 4. Emulsification results for SiO2-PDMAEMA particles with varying grafting density and particle concentration at pH 7.5 and

10 mM NaCl (T < CFT). Low grafting density particles with σ = 0.076 chains/nm2, Mn = 16 150 were used to keep the degree of polymerization nearly constant for each grafting density in a and b. Low grafting density particles with σ = 0.077 chains/nm2, Mn = 36 150 were used in c-e. Open and filled symbols represent, respectively, cases in which no emulsion was formed or cases when stable emulsions were formed. (a-d) Particles dispersed in the aqueous phase. (a) Emulsions with xylene at T = 50 °C. (b) Emulsions with cyclohexane at T = 50 °C. (c) Emulsions with xylene at T = 20 °C. (d) Emulsions with cyclohexane at T = 20 °C. (e) Emulsions with particles dispersed in xylene for the aqueous phase at pH 7.5, 10 mM NaCl at T = 20 °C.

in the aqueous phase, where the particles demonstrate CFT behavior. 3.7. Emulsion Stability and Control Experiments with Bare Particles or Ungrafted PDMAEMA Polymers in Solution. All samples were stored in sealed containers at room temperature for stability testing. All samples that emulsified initially have been stable against coalescence for a minimum of 6 months, and some have demonstrated stability for up to 13 months (Supporting Information). We note that these emulsions were still stable at the time of this writing. The different stability lifetimes in this case simply reflect which samples were prepared first. The formation of a barely visible neat oil phase and/or some visible coalesced droplets in the emulsion phase has been observed in a few cases after 5-7 months. About half of those samples that indicated the onset of droplet coalescence occurred at the low particle concentrations, 0.5 wt % and below, using the medium grafting density particles. Thus, this behavior could be due to insufficient coverage of the droplet interface to prevent film drainage and coalescence. Most importantly, emulsions prepared with just 0.05 wt % of the low and medium grafting density particles remained stable for 6-10 months before coalescence, followed by phase separation, was observed. Langmuir 2010, 26(19), 15200–15209

The appearance of a neat oil phase immediately after emulsification occurred in a small number of samples, mostly with the medium grafting density particles. All except one of the emulsions that had a neat xylene phase were made below the CFT. The samples with a neat cyclohexane phase were emulsified above the CFT. It is not obvious why this occurs, but in addition to the hydrophobicity of the particles under a set of conditions, the finite solubility of the solvents in water could play a role in emulsion stability. Xylene is more soluble in water (0.48 g/L)37 than is cyclohexane (0.06 g/L).38 Thus, the greater solubility of xylene could produce more pronounced Ostwald ripening than cyclohexane, but why they would show the opposite trend with respect to T/CFT during emulsification is unclear. Control experiments were conducted to test the emulsification abilities of PDMAEMA homopolymers in solution and unmodified 20-nm-diameter silica nanoparticles individually. Emulsions were created with mass concentrations of 4.4 wt % ungrafted PDMAEMA homopolymer (Mn = 44 870) in an aqueous (37) Pryor, W. A.; Jentoft, R. E. J. Chem. Eng. Data 1961, 6, 36–37. (38) Chemical summary for cyclohexane. http://www.epa.gov/chemfact/ s_cycloh.txt (5-18-2010).

DOI: 10.1021/la1027898

15207

Article

Saigal et al.

Figure 5. Micrographs at 40 magnification of diluted emulsions created with 2 wt % SiO2-PDMAEMA particles dispersed in 10 mM NaCl for (a) pH 7.5, 50 °C cyclohexane/water and (b) pH 9, 50 °C xylene/water systems. Both show evidence of double emulsion formation. Small droplets were observed to undergo Brownian motion inside the larger droplets that were stuck to the coverslip.

solution at pH 7.5 with 10 mM NaCl. This exceeds the amount of polymer present in any of the emulsions created with SiO2PDMAEMA particles reported here. No emulsions were stable for more than 3 days for either xylene or cyclohexane. Similarly, silica nanoparticles dispersed in water at 2 wt %, pH 7.5 with 10 mM NaCl did not produce any emulsions with either solvent. 3.8. Interfacial Tension Lowering. The interfacial tensions of cyclohexane and xylene against water were measured with the SiO2-PDMAEMA particles dispersed in the aqueous phase. All particle grafting densities were tested at room temperature, pH 9 in 1, 10, or 100 mM NaCl. The cyclohexane/water interfacial tension measured was between 10.2 and 10.8 mN/m for all particle grafting densities and ionic strengths, which is a significant decrease from the measured clean cyclohexane/water interfacial tension of 48.5 mN/m (consistent with literature values39). For comparison, ungrafted PDMAEMA homopolymer (Mn = 44 870) decreased the cyclohexane/water interfacial tension to 14.7 mN/m at a concentration of 4.4 wt %. The xylene/water interfacial tension with the high and medium grafting density particles ranged from 3.6 to 4.8 mN/m with no significant dependence on ionic strength. With the high molecular weight, low grafting density particles (Mn = 36 150), the xylene/ water interfacial tension was 13.3, 7.13, or 7.5 mN/m at 1, 10, or 100 mM NaCl, respectively. The higher interfacial tension at 1 mM NaCl could suggest lateral repulsive forces among adsorbed particles due to minimal screening of the charged PDMAEMA chains. This would cause fewer particles to pack at the interface and would result in an overall higher interfacial tension. The interfacial tension of a clean xylene/water interface was measured to be 30 mN/m, consistent with literature values,40 and decreased to 4.1 mN/m using 4.4 wt % PDMAEMA homopolymer in solution. Bear in mind that these measurements were made at a single nanoparticle concentration. Without a full characterization of the interfacial tension versus concentration isotherms, a more detailed analysis of the surface-penetrating abilities of the various nanoparticles is not justified. These measurements confirm that PDMAEMA-grafted nanoparticles significantly decrease the oil/ water interfacial tension to a degree that is similar to that of PDMAEMA homopolymers in solution. 3.9. Responsive Emulsions. The ability to break emulsions created with SiO2 -PDMAEMA particles via temperature cues (39) Ikenaga, T.; Matubayasi, N.; Aratono, M.; Motomura, K.; Matuura, R. Chem. Soc. Jpn. 1978, 53, 653–657. (40) Lord, D. L.; Hayes, K. F.; Demond, A. H.; Salehzadeh, A. Environ. Sci. Technol. 1997, 31, 2045–2051.

15208 DOI: 10.1021/la1027898

was tested for all particle grafting densities. Particles were dispersed in water at pH 9, 20 °C in the presence of 10 mM NaCl for the low and medium grafting density particles and 100 mM NaCl for the high grafting density particles. For the high grafting density particles, 100 mM NaCl was used because at this ionic strength an emulsion did not form above the CFT at 50 °C. Xylene-in-water emulsions were made at room temperature and then heated to 50 °C. In all cases, the emulsion broke completely within 1 min upon heating above the CFT. Thus, these particles can be used to create responsive emulsions that rapidly deemulsify when triggered. As noted above in section 3.6, xylenein-water emulsions were also thermally responsive even when the particles were initially dispersed in xylene. 3.10. One-Step Double Emulsions and the Absence of Catastrophic Phase Inversion. The solvent quality of the oil below the CFT, and in some cases above the CFT, had little or no apparent effect on the emulsification ability. Yet, there was a marked difference in the microstructure of emulsions formed with xylene or cyclohexane at high particle grafting density and 10 mM NaCl (Figure 5). Below the CFT at pH 7.5 and T = 50 °C, microscopic visualization of the cyclohexane emulsion showed a significant fraction (but not 100%) of double emulsion droplets (Figure 5a). In contrast, xylene emulsions (not shown) contained no detectable double emulsion droplets under these conditions. Above the CFT, at pH 9 and T = 50 °C (Figure 5b), there was an abundance of double emulsion droplets in the xylene/water system, but none were observed in the cyclohexane/water system under this condition. Video microscopy demonstrated that the inner droplets inside the oil droplets were undergoing Brownian motion and experiencing longrange repulsions from neighboring droplets (video in Supporting Information). They were not merely small droplets trapped against the coverslip under a large droplet. The continuous phase in the emulsion was water. The long-range repulsions among the water droplets inside the oil droplets of the w/o/w double emulsion droplets were consistent with the long-range electrostatic repulsion expected in the low dielectric oil environment. Double emulsion (w/o/w) droplet formation in a single homogenization step was unexpected because no w/o emulsions could be produced with these particles under any conditions. According to the Bancroft rule, the phase containing the emulsifier will most likely become the continuous phase. In each of the w/o/w systems noted at 50 °C, the particles were initially dispersed in the water phase. In an attempt to make w/o emulsions, the particles were first dispersed in xylene before homogenization. This produced no emulsions at a 1:1 water/oil ratio at 50 °C. Langmuir 2010, 26(19), 15200–15209

Saigal et al.

Some Pickering emulsion systems have a tendency to phase invert upon increasing the relative amount of the discontinuousphase liquid in the sample (catastrophic phase inversion). We considered the possibility that some kind of catastrophic phase inversion could happen inside the emulsion phase, which typically contained on the order of 70 vol % oil, leading to double emulsion droplet formation. We therefore investigated phase inversion in this system by progressively increasing the relative amount of oil in the system. The SiO2-PDMAEMA (σ = 1.27 chains/nm2, Mn=19 400) particles created o/w emulsions at all oil/water ratios up to 7:3 (no inversion), but above this ratio, no emulsion formed at all. Thus, no catastrophic phase inversion occurred, making a hypothetical microscale catastrophic phase inversion inside the oil-rich emulsion phase unlikely. A hypothesized mechanism for the one-step formation of double emulsions involves interfacial roughening of the droplet surfaces on opposite sides of the thin water film between juxtaposed oil droplets just prior to coalescence events that occur during high-shear homogenization. Oscillations in the interface could grow to pinch off water droplets from within the thin film between two coalescing oil droplets and become consumed by the resulting daughter oil droplet. If there is some characteristic wavelength for the thin liquid film interface roughness, then this could be responsible for the rather monodisperse size distribution of the water droplets inside the oil droplets.

4. Conclusions Twenty-nanometer-diameter silica particles with PDMAEMA brushes grafted from their surfaces created highly stable xylenein-water and cyclohexane-in-water emulsions at extremely low particle concentrations. Particles with lower grafting densities proved to be the most robust and efficient emulsifiers, although particles with high grafting densities at which chains are expected to be fully stretched were still effective emulsifiers under many

Langmuir 2010, 26(19), 15200–15209

Article

conditions. Stable emulsions were created with as little as 0.05 wt % of the low grafting density particles (0.077 chains/ nm2). This high emulsifying efficiency implies that the nanoparticles have a high-affinity adsorption isotherm at the oil/water interface, requiring a low total particle content to stabilize large surface areas. This is the subject of ongoing research. Cyclohexane, a poor solvent for PDMAEMA, could be emulsified under more conditions of temperature, pH, and ionic strength than could xylene, a good solvent for PDMAEMA. Ionic strength had a secondary effect on the emulsification behavior, and although it affects many aspects of the stabilization process, this study showed a higher dependence on whether emulsification was conducted above or below the CFT. Emulsions were thermally responsive, rapidly breaking upon increasing the temperature above the critical flocculation temperature of the SiO2-PDMAEMA particles in water. This system did not display catastrophic phase inversion. Acknowledgment. This material is based on work supported by the National Science Foundation under grant CBET-0729967 and in part by the American Chemical Society Petroleum Research Fund under grant 17430AC7. We thank Nicolas Alvarez and Dennis Kloss (Carnegie Mellon University) for experimental assistance. Supporting Information Available: Tabulation of emulsion characteristics (relative phase volumes, dispersed-phase volume fraction in emulsions, and stability against coalescence) for each emulsion sample tested. Illustrative calculations demonstrating how high-affinity nanoparticle adsorption isotherms enable emulsification at low particle concentrations.Video microscopy of double emulsion droplets exhibiting Brownian motion. This material is available free of charge via the Internet at http:// pubs.acs.org.

DOI: 10.1021/la1027898

15209