Double Emulsions and Colloidosomes-in-Colloidosomes Using Silica

Feb 21, 2014 - (4) Two different types of particles can be used to make double emulsions, with hydrophobic particles stabilizing the w/o interface and...
0 downloads 11 Views 5MB Size
Article pubs.acs.org/Langmuir

Double Emulsions and Colloidosomes-in-Colloidosomes Using SilicaBased Pickering Emulsifiers Mark Williams and Steven P. Armes* Department of Chemistry, University of Sheffield, Brook Hill, Sheffield, South Yorkshire S3 7HF, U.K.

Pierre Verstraete and Johan Smets Procter & Gamble, Eurocor NV/SA, Temselaan 100, 1853 Strombeek-Bever, Belgium S Supporting Information *

ABSTRACT: Poly(ethylene imine) (PEI) has been adsorbed onto the surface of fumed silica particles at pH 10 in order to produce an effective “hybrid” Pickering emulsifier. Systematically increasing the PEI/silica mass ratio at a fixed silica concentration of 1.0% w/w modifies the silica particle surface and hence allows the formation of oil-in-water (o/w) Pickering emulsions prepared via homogenization of an aldehyde-rich multi-component fragrance oil (at 12 000 rpm for 2 min at 20 °C). Further increasing the PEI/silica mass ratio leads to phase inversion, producing water-in-oil (w/o) Pickering emulsions. Thus this approach allows formation of stable water-in-oil-in-water (w/o/w) double emulsions using two batches of hydrophilic and hydrophobic PEI/silica hybrid particles that differ only in their PEI/silica mass ratios prior to homogenization. Stable w/o/w double emulsions can be prepared with oil volume fractions ranging from 5 to 42%. Moreover, controlling the volume fraction of the w/o Pickering emulsion homogenized in the presence of an aqueous dispersion of the hydrophilic PEI/silica particles allows the mean diameter of the resulting oil droplets to be conveniently controlled between 20 and 160 μm. Fluorescence microscopy studies confirm that controlling the mean diameter of these oil droplets allows encapsulation of either single or multiple droplets within them. Although these double emulsions do not require cross-linking at either interface to withstand an alcohol challenge, epoxy−amine cross-linking between the physicallyadsorbed PEI chains and either an oil-soluble or a water-soluble bisepoxy-based polymeric cross-linker can be achieved to produce novel colloidosomes-in-colloidosomes, which may offer payload retention benefits over conventional colloidosomes.



INTRODUCTION An emulsion stabilized by solid particles is known as a Pickering emulsion.1,2 Particle wettability is an important parameter in determining the nature of such emulsions.3 For hydrophilic particles, most of the particle volume is located in the aqueous phase; this gives rise to a contact angle, θ, of less than 90° which leads to an oil-in-water (o/w) emulsion. For hydrophobic particles, the particle resides mainly in the oil phase, θ exceeds 90°, and a water-in-oil (w/o) emulsion is typically produced.4 There are a number of techniques that can be used to control this contact angle parameter (θ) in order to prepare stable Pickering emulsions. The surface modification of nanoparticles is commonly used to tune their wettability. One approach is the chemical derivatization of inorganic oxide sols such as silica with various alkyl silanes.5 However, similar results can be achieved simply by physical adsorption: recent literature examples include (i) polyelectrolyte adsorption onto clay nanoparticles such as Laponite6 and montmorillonite7 or adsorption of either (ii) adsorption of either small molecule aromatic acids8 or a cationic surfactant9−12 onto silica nanoparticles. © 2014 American Chemical Society

There are various literature examples of Pickering emulsion phase inversion using a wide range of particles. For example, Binks et al. reported the catastrophic phase inversion from w/o to o/w emulsions using hydrophobic silica particles simply by changing the volume fraction of the aqueous phase. The same team found that complementary phase inversion (i.e., from o/w to w/o) could be achieved using hydrophilic silica particles.13 They also reported that phase inversion could be achieved by fixing the volume fraction of the aqueous phase while increasing the silica particle concentration.14 In a separate study, reducing the surface silanol content of silica particles caused phase inversion from an o/w emulsion to a w/o emulsion.5 Binks et al. have shown how these three techniques can be exploited for the phase inversion of oil-in-water emulsions where the oil droplet phase comprises a water-immiscible fragrance.15 Zhang et al. showed that inversion of Laponite-stabilized o/w Pickering emulsions could be achieved by varying the salt concentration. In this system, the salt concentration influences Received: January 16, 2014 Revised: February 17, 2014 Published: February 21, 2014 2703

dx.doi.org/10.1021/la500219m | Langmuir 2014, 30, 2703−2711

Langmuir

Article

the degree of flocculation of the nanoclay particles, which in turn affects their interfacial adsorption.16 Calcium carbonate nanoparticles are normally too hydrophilic to stabilize Pickering emulsions, but Cui et al. demonstrated that the addition of sodium dodecyl sulfate allowed formation of stable o/w Pickering emulsions. However, further addition caused phase inversion as the surfactant-coated particles became sufficiently hydrophobic to stabilize w/o Pickering emulsions.17 Phase inversion has been exploited to produce so-called “double emulsions”. Such systems can be prepared by gently emulsifying a w/o emulsion with a second aqueous phase to produce a w/o/w double emulsion.18 Alternatively, the complementary o/w/o double emulsion can be prepared by first making an o/w emulsion, followed by re-emulsification in the presence of further oil (which forms the continuous phase).19 One important advantage of using particles instead of surfactants is that the former emulsifier usually prevents droplet coalescence, which otherwise eventually leads to the formation of a single emulsion.4 Two different types of particles can be used to make double emulsions, with hydrophobic particles stabilizing the w/o interface and hydrophilic particles stabilizing the o/w interface. For example, Binks et al. reported the preparation of stable double emulsions using two batches of silica particles with differing surface silanol contents (and hence relative wettabilities).20 Although membrane emulsification cells offer improved size control, Pickering emulsions made by conventional homogenization techniques typically comprise rather polydisperse droplets.21 Weitz et al. have demonstrated that highly monodisperse double emulsions can be made with controllable shell thicknesses using microfluidics,22−24 but this specialist technique is not generally considered to be amenable for industrial use. Colloidosomes are an interesting class of microcapsules derived from Pickering emulsions that are further stabilized by either thermal annealing,25,26 polyelectrolyte adsorption,25 gel trapping,27 or covalent cross-linking28−31 in order to lock in the initial particle superstructure; they can be prepared using various Pickering emulsifiers such as polymer latexes,26−33 inorganic/polymer nanocomposites,30,34,35 and clay nanoparticles.6,7 Previously, we have reported that poly(ethylene imine) can be physically adsorbed onto Laponite particles to produce a hybrid Pickering emulsifier that is effective for a wide range of oils.6 Such o/w Pickering emulsions could be easily transformed into robust covalently cross-linked colloidosomes using either an oil-soluble bisepoxy-functionalized poly(propylene glycol) [PPG-DGE] or a water-soluble bisepoxyfunctionalized poly(ethylene glycol) [PEG-DGE] cross-linker.29 Herein we report that poly(ethylene imine) can be physically adsorbed onto fumed silica particles to produce a versatile Pickering emulsifier that can stabilize either o/w or w/o Pickering emulsions, depending on the extent of adsorption of the poly(ethylene imine). In this formulation, the oil is a proprietary multi-component aldehyde-rich fragrance. Moreover, w/o/w double emulsions can be prepared using the same fumed silica particles coated with differing amounts of poly(ethylene imine) to tune the particle wettability. Such double emulsions can be stabilized with respect to coalescence by addition of both a water-soluble and an oil-soluble bisepoxyfunctionalized polymeric cross-linker prior to homogenization so as to produce colloidosomes-in-colloidosomes. Such novel microcapsules may offer new opportunities for encapsulation and retention of fragrances. We also demonstrate that both the

mean droplet diameter and nature of these double emulsions can be controlled simply by varying the oil volume fraction used in the formulation.



EXPERIMENTAL SECTION

Materials. Poly(ethylene imine) (PEI) (branched, MW = 10 000 by GPC; MW = 25 000 by light scattering), poly(propylene glycol) diglycidyl ether (PPG-DGE, mean degree of polymerization = 9 according to the supplier), poly(ethylene glycol) diglycidyl ether (PEG-DGE; mean degree of polymerization = 9 according to the supplier), and Nile Red were all purchased from Aldrich and were used as received. Ethanol was purchased from Fisher and used as received. Deionized water was used in all experiments. Cab-o-Sperse 2012A (fumed silica) was kindly supplied by Cabot Corporation (Billerica, MA). The core oil is a typical aldehyde-rich perfume blend of various components, which has a relatively low solubility in water and is used in various commercial consumer products. Its precise chemical composition is proprietary, but a detailed description of the c log P values for its various components is provided in Table S1. Adsorption of PEI onto Fumed Silica. PEI (0−2.00 g) was dissolved in deionized water (48.0−50.0 g). This aqueous polymer solution was then added to an aqueous dispersion of Cab-o-Sperse 2012A fumed silica (50 g, 2.0% w/w), and the resulting mixture was stirred for 12 h at 20 °C so as to allow physical adsorption of the cationic PEI chains onto the anionic silica particles. Preparation of Pickering Emulsions. The following protocol is representative. The proprietary aldehyde-rich oil (5.0 mL) was added to a 14 mL sample vial, followed by addition of an aqueous suspension of PEI/silica hybrid particles (5.0 mL, 1.0% w/w Cab-o-Sperse 2012A fumed silica, 0−2.0% w/w PEI, pH 10), as prepared above. Emulsification was achieved at 20 °C using an IKA Ultra-Turrax T18 homogenizer equipped with a 10 mm dispersing tool for 2 min at 12 000 rpm. Preparation of Double Emulsions. The following protocol is representative. The proprietary aldehyde-rich oil (5.0 mL) was added to a 28 mL sample vial, followed by the addition of an aqueous suspension of PEI/silica hybrid particles (5.0 mL, 1.0% w/w Cab-oSperse 2012A fumed silica, PEI/silica mass ratio = 0.50, pH 10). Emulsification was achieved at 20 °C using an IKA Ultra-Turrax T-18 homogenizer equipped with a 10 mm dispersing tool for 2 min at 12 000 rpm. An aqueous suspension of PEI/silica hybrid particles (10.0 mL, 1.0% w/w Cab-o-Sperse 2012A fumed silica, PEI/silica mass ratio = 0.075, pH 10) was then added to the resulting water-in-oil emulsion (10.0 mL), and again emulsification was achieved at 20 °C using an IKA Ultra-Turrax T-18 homogenizer equipped with a 10 mm dispersing tool for 2 min at 12 000 rpm. This produces a w/o/w double emulsion (20 mL) with a total oil content of 25% (5.0 mL). To vary the oil volume fraction of such double emulsions, both the oil volume fraction used to make the initial w/o Pickering emulsion and the volume fraction of this w/o Pickering emulsion added to make the w/o/w double emulsion were adjusted. Preparation of Double Emulsions Using Poly(ethylene glycol) Diglycidyl Ether (PEG-DGE) Cross-Linker. An aqueous suspension of PEI/silica hybrid particles (5.0 mL, 1.0% w/w fumed silica, PEI/silica mass ratio = 0.50, pH 10) was added to a 28 mL sample vial. PEG-DGE cross-linker (20.0 mg; amine/epoxy molar ratio = 10.0) was added to the aqueous continuous phase and shaken to ensure its dissolution. The proprietary aldehyde-rich oil (5.0 mL) was added, and emulsification was achieved at 20 °C using an IKA UltraTurrax T-18 homogenizer equipped with a 10 mm dispersing tool for 2 min at 12 000 rpm. An aqueous suspension of PEI/silica hybrid particles (10.0 mL, 1.0% w/w fumed silica, PEI/silica mass ratio = 0.075, pH 10) was then added to the resulting water-in-oil emulsion (10.0 mL), and again emulsification was achieved at 20 °C using an IKA Ultra-Turrax T-18 homogenizer equipped with a 10 mm dispersing tool for 2 min at 12 000 rpm. Again, PEG-DGE crosslinker (10.0 mg; amine/epoxy molar ratio = 5.0) was added to the aqueous continuous phase and shaken to ensure its dissolution. This 2704

dx.doi.org/10.1021/la500219m | Langmuir 2014, 30, 2703−2711

Langmuir

Article

double emulsion was then allowed to stand unstirred at 20 °C for 12 h in order to allow cross-linking to occur. Double Emulsion Preparation Using Poly(propylene glycol) Diglycidyl Ether (PPG-DGE) Cross-Linker. An aqueous suspension of PEI/silica hybrid particles (5.0 mL, 1.0% w/w fumed silica, PEI/ silica mass ratio = 0.50, pH 10) was added to a 28 mL sample vial. PPG-DGE cross-linker (20.0 mg; amine/epoxy molar ratio = 10.0) dissolved in the proprietary aldehyde-rich oil (5.0 mL) was added, and emulsification was achieved at 20 °C using an IKA Ultra-Turrax T-18 homogenizer equipped with a 10 mm dispersing tool for 2 min at 12 000 rpm. An aqueous suspension of PEI/silica hybrid particles (10.0 mL, 1.0% w/w, PEI/silica mass ratio = 0.075, pH 10) was then added to the resulting water-in-oil emulsion (10.0 mL), and again emulsification was achieved at 20 °C using an IKA Ultra-Turrax T18 homogenizer equipped with a 10 mm dispersing tool for 2 min at 12 000 rpm. This double emulsion was then allowed to stand unstirred at 20 °C for 12 h in order to allow cross-linking to occur. Fluorescent Labeling of Water-in-Oil-in-Water Double Emulsions. Fluorescent o/w Pickering emulsions, w/o Pickering emulsions, and w/o/w double emulsions were made using the same formulations as those described above, except that Nile Red dye was first dissolved in the oil phase ([Nile Red] = 0.010 g dm−3) prior to its homogenization. Alcohol Challenge. A small aliquot of a w/o/w double emulsion (0.50 mL) was transferred to a 5.0 mL vial, and then excess ethanol (4.0 mL) was added and the sealed vial was shaken vigorously. This same alcohol challenge was also used to assess the mechanical integrity of the corresponding covalently cross-linked colloidosomes. Particle Characterization. Dynamic Light Scattering (DLS). The intensity-average hydrodynamic diameter was determined by DLS for both the pristine Cab-o-Sperse 2012A fumed silica and also PEI/silica hybrid particles prepared at various PEI/silica mass ratios using a Malvern Zetasizer NanoZS instrument. Aqueous solutions of 0.01% w/v silica dispersions were analyzed using disposable plastic cuvettes, and results were averaged over three consecutive runs. Deionized water was used to dilute each dispersion and was ultrafiltered through a 0.20 μm membrane to remove dust prior to use. Aqueous Electrophoresis. Zeta potentials were determined for both pristine Cab-o-Sperse 2012A fumed silica and also PEI/silica hybrid particles prepared at various PEI/silica mass ratios using a Malvern Zetasizer Nano ZS instrument. The solution pH was fixed at pH 10 in the presence of 1 mM KCl, with either dilute NaOH or HCl being used for pH adjustment as required. Thermogravimetric Analysis. Analyses were conducted using a PerkinElmer Pyris-1 TGA instrument. Prior to analysis, the PEI/silica hybrid particles were purified by centrifugation at 8000 rpm for 1 h, carefully replacing the supernatant with mildly alkaline water (pH 10) each time followed by redispersion of the sedimented particles with the aid of an ultrasonic bath. This centrifugation−redispersion cycle was repeated four times to ensure that no excess nonadsorbed PEI remained in the aqueous continuous phase. The dried PEI/silica hybrid particles were heated up to 800 °C in air at a heating rate of 20 °C min−1. The amount of PEI adsorbed onto the fumed silica particles was determined from the weight loss observed between 200 and 650 °C. Emulsion Characterization. Conductivity Measurements. The conductivities of the emulsions immediately after preparation were measured using a digital conductivity meter (Hanna model Primo 5). A high conductivity (typically >10 μS cm−1) indicated that the emulsion was water-continuous. All results were confirmed using the so-called “drop test”, whereby one drop of the emulsion was added to both pure water and oil, and its ease of dispersion was assessed by visual inspection. Relatively rapid dispersion into water was taken as confirmation that the continuous phase of the emulsion was indeed water, whereas relatively rapid dispersion into oil was taken as confirmation that the continuous phase of the emulsion was the oil phase. Optical Microscopy. A single droplet (ca. 100 μL) of either a o/w or w/o Pickering emulsion (50% v/v) or a double emulsion (25% v/v) was placed on a microscope slide, and digital images were recorded

using a Motic DMBA300 digital biological microscope equipped with a built-in camera and Motic Images Plus 2.0 ML software. Fluorescence Microscopy. A single droplet (ca. 100 μL) of either a o/w or w/o Pickering emulsion (50% v/v) or a double emulsion (25% v/v) was placed on a microscope slide and viewed using an Olympus Upright Epifluorescence instrument equipped with a Hamamatsu ORCA-ER monochrome camera and Volocity software. This technique was used to view the fluorescent Nile Red dye dissolved in the oil phase for the o/w and w/o emulsions and also for the w/o/w double emulsions. Laser Diffraction Particle Size Analysis of Emulsion Droplets. 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, double emulsions, and colloidosomes. The stirring rate was adjusted to 1000 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 droplet diameter was taken to be the mean volume-average diameter (D4/3), which is mathematically expressed as D4/3 = ∑Di4Ni/∑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 crosscontamination, and the laser was aligned centrally on the detector. Scanning Electron Microscopy. SEM images were obtained using a FEI inspect F FEG instrument operating at 20 kV. All samples were sputter-coated with a thin overlayer of gold prior to inspection to prevent sample charging effects.



RESULTS AND DISCUSSION Initially, the effect of varying the extent of PEI adsorption onto the fumed silica particles was investigated. According to its

Figure 1. Effect of varying the target poly(ethylene imine)/silica mass ratio on (a) the zeta potential (●) recorded at pH 10 for the resulting poly(ethylene imine)/silica hybrid particles and (b) the adsorbed mass of poly(ethylene imine) per unit mass of fumed silica (■). The fumed silica concentration was fixed at 1.0% w/w in each case.

manufacturer, the pristine fumed silica used in this study has a primary particle size of 15 nm and a BET specific surface area of 200 m2 g−1. All PEI adsorption experiments were carried out at pH 10. Under these conditions, the fumed silica particles had an intensity-average particle diameter of 160 nm and a zeta potential of −47 mV (see Figure S1 in the Supporting Information). This size indicates weak flocculation of the primary particles, as expected for fumed silica. The particles become less anionic at lower pH due to protonation of surface silanol groups, with an isoelectric point being observed at pH 3. 2705

dx.doi.org/10.1021/la500219m | Langmuir 2014, 30, 2703−2711

Langmuir

Article

The oil used in this study is a typical aldehyde-rich perfume blend of various components, which has a relatively low solubility in water and is used in various commercial consumer products. Its precise chemical composition is proprietary, but a detailed description of the c log P values for its various components is provided in Supporting Information Figure S1. The effect of varying the extent of PEI adsorption onto the fumed silica particles on their Pickering emulsifier performance was investigated using the proprietary aldehyde-rich fragrance as the oil phase. At PEI/silica mass ratios above 0.075, not all of the PEI adsorbs onto the surface of the silica particles. Thus there is a minor fraction of free PEI chains present in the aqueous continuous phase during emulsification. Control experiments suggest that, when present at comparable concentrations, PEI alone is an ef fective emulsifier for the proprietary aldehyde-rich oil, since substantial droplet coalescence occurs within hours. This indicates that the excess nonadsorbed PEI chains do not act as an effective emulsifier; hence the highly stable emulsion droplets that are obtained are actually stabilized by the PEI/silica particles. Figure 2 shows how the dispersed phase of the Pickering emulsions (obtained after homogenization of equal volumes of oil and aqueous dispersion at 12 000 rpm for 2 min at 20 °C) varies according to the PEI/silica mass ratio using 1.0% w/w fumed silica. In the absence of any PEI, the pristine fumed silica was too hydrophilic to stabilize an oil-in-water Pickering emulsion (data not shown). As the PEI/silica mass ratio was gradually increased, stable oil-in-water Pickering emulsions could be obtained because PEI adsorption increases the contact angle of the silica particles sufficiently to stabilize emulsion droplets. Empirically, a PEI/silica mass ratio of 0.075 (1.0% w/w silica, 0.075% w/w PEI, pH 10) appeared to be optimal, since this produced the smallest, most stable oil-in-water Pickering emulsions with a volume-average droplet diameter of around 40 μm. However, this PEI/silica mass ratio is well below the knee of the adsorption isotherm (see Figure 1), so further adsorption of PEI onto the silica particles can be achieved. As shown in Figure 2, phase inversion from an o/w emulsion to a w/o emulsion occurs at a critical mass ratio of 0.25 (i.e., 1.0% w/w silica, 0.25% w/w PEI, pH 10). This is because further PEI adsorption results in the silica particles becoming too hydrophobic to stabilize o/w Pickering emulsions as their contact angle θ now exceeds 90°. At the knee of the adsorption isotherm, stable w/o Pickering emulsions are formed at a PEI/ silica mass ratio of 0.50. As discussed above, some fraction of nonadsorbed PEI chains is always present in the aqueous continuous phase. Following removal of this excess PEI from the aqueous phase by a centrifugation/redispersion cycle, Pickering emulsions prepared using PEI/silica hybrid particles (at PEI/silica mass ratios of either 0.075 or 0.50) exhibited similar mean droplet diameters and comparable stabilities

Figure 2. Effect of varying the poly(ethylene imine)/silica mass ratio (see respective vial labels) on the nature of Pickering emulsions obtained after homogenization of equal volumes of a proprietary aldehyde-rich oil and a 1.0% w/w aqueous dispersion of fumed silica particles at 12 000 rpm for 2 min at 20 °C.

Dynamic light scattering (DLS) studies indicate that this reduction in zeta potential at lower pH results in gross flocculation at around pH 5. The amount of PEI added to a 1.0% w/w aqueous dispersion of fumed silica at pH 10 was systematically varied from zero to 2.0% w/w (i.e., a PEI/silica mass ratio of 2.0). The initial negative zeta potential of the pristine fumed silica particles was gradually reduced due to charge compensation by the adsorbed cationic PEI chains, with an isoelectric point being obtained at a PEI/silica mass ratio of around 0.15 and surface charge reversal being observed thereafter. The actual adsorbed amount of silica, as determined by thermogravimetry, is consistent with the observed electrophoretic behavior and suggests Langmuir-type adsorption, as expected for a cationic polyelectrolyte adsorbed onto an oppositely charged colloidal substrate36 (see Figure 1). The knee of the isotherm is observed at a PEI/silica mass ratio of 0.50 (1.0% w/w silica, 0.50% w/w PEI, pH 10) (see Figure 1). This knee corresponds to an adsorbed amount of 0.54 mg m−2, which indicates that the average area occupied by one PEI chain on the silica surface is 77 nm2. This value is reasonably consistent with the number-average hydrodynamic radius (Rh) of 5 nm observed for isolated PEI chains in dilute aqueous solution at pH 10, as judged by dynamic light scattering.

Figure 3. Schematic representation of the formation of w/o/w double emulsions using poly(ethylene imine)-coated silica particles at pH 10. The first step involves the formation of a w/o Pickering emulsion made by homogenizing an aqueous dispersion of hydrophobic PEI/silica particles (1.0% w/w silica; PEI/silica mass ratio = 0.50; pH 10) with a proprietary aldehyde-rich oil. This initial w/o Pickering emulsion is then homogenized with hydrophilic PEI/silica particles (1.0% w/w silica; PEI/silica mass ratio = 0.075; pH 10) to produce the final w/o/w double emulsion. 2706

dx.doi.org/10.1021/la500219m | Langmuir 2014, 30, 2703−2711

Langmuir

Article

Figure 4. Phase diagram obtained by (i) varying the volume fraction of a proprietary aldehyde-rich hydrophobic fragrance (oil phase) when preparing the initial w/o Pickering emulsions (see x-axis) and (ii) varying the volume fraction of the initial w/o Pickering emulsion utilized when preparing the subsequent w/o/w double emulsions (see y-axis).

Figure 5. Variation of the volume-average droplet diameter of w/o/w double emulsions obtained on adjusting the oil volume fraction, as measured by laser diffraction. The initial w/o Pickering emulsion was prepared by homogenization of 50% oil using hydrophobic PEI/silica particles (1.0% w/w silica, PEI/silica mass ratio = 0.50; pH 10). This w/o emulsion was then homogenized using hydrophilic PEI/silica particles (1.0% w/w silica, PEI/silica mass ratio = 0.075, pH 10) at various w/o emulsion volume fractions so as to systematically vary the mean oil droplet diameter within the w/o/w emulsion.

Figure 6. Fluorescence microscopy images obtained by dissolution of the Nile Red in the proprietary aldehyde-rich oil prior to its homogenization for (A) o/w Pickering emulsion (50% oil volume fraction), (B) w/o Pickering emulsion (50% oil volume fraction), (C) w/o/w double emulsion (prepared using 10% oil volume fraction) with the oil droplets encapsulating single water droplets, and (D) w/o/ w double emulsion (prepared using 25% oil volume fraction) with the oil droplets encapsulating multiple water droplets.

toward coalescence as those emulsions prepared in the presence of excess PEI chains. These observations suggest that the free PEI have little or no effect on the initial formation of the emulsion droplets or their long-term stability. PEI/silica mass ratios of 0.075 and 0.50 were used for all double emulsion formulations. The same fumed silica particles were used in all cases, with the only difference being the PEI/ silica mass ratio employed prior to homogenization. Unlike the work of Binks and co-workers,20 no chemical derivatization of the silica particle surface is required to tune the surface wettability in order to obtain Pickering emulsifiers that are suitable for the formation of either o/w or w/o emulsions. At the lower PEI/silica mass ratio of 0.075, almost all of the PEI is adsorbed onto the fumed silica particles at submonolayer coverage; i.e., this corresponds to a point below the knee of the

adsorption isotherm (see Figure 1). In contrast, at a PEI/silica mass ratio of 0.50, there is a significant excess of nonadsorbed PEI remaining in the aqueous phase, although genuine Pickering emulsions are still obtained under these conditions (see above). Fortunately, the excess PEI is not an important consideration in the context of the cross-linking strategies used to covalently stabilize the double emulsions, since it is confined to the encapsulated aqueous droplets (see later). 2707

dx.doi.org/10.1021/la500219m | Langmuir 2014, 30, 2703−2711

Langmuir

Article

Figure 7. Schematic illustration of the synthesis of novel multicompartment microcapsules at 50% solids (oil volume fraction = 25%) using two types of PEI/silica hybrid particles as Pickering emulsifiers at pH 10. Route A depicts the formation of a colloidosome-in-colloidosome microcapsule produced from a w/o/w double emulsion by dissolving an oil-soluble polymeric cross-linker (PPG-DGE) in the aldehyde-rich hydrophobic fragrance prior to any homogenization. Route B depicts colloidosome-in-colloidosome formation via two separate charges of a water-soluble polymeric crosslinker (PEG-DGE). The first charge of PEG-DGE ensures cross-linking of the precursor w/o emulsion from within the aqueous droplets The second charge of PEG-DGE is added to the aqueous continuous phase after homogenization to form the w/o/w double emulsion so as to ensure crosslinking of the PEI/silica particles adsorbed at the outer surface of the oil droplets.

Figure 3 illustrates the strategy used to produce w/o/w double emulsions. The first step involves homogenization of an aqueous dispersion of hydrophobic PEI/silica hybrid particles (1.0% w/w silica; PEI/silica mass ratio = 0.50; pH 10) with the aldehyde-rich hydrophobic fragrance (oil phase) to produce a stable w/o Pickering emulsion. This emulsion was subsequently homogenized using the second batch of hydrophilic PEI/silica hybrid particles (1.0% w/w silica; PEI/silica mass ratio = 0.075; pH 10) to form the final w/o/w double emulsion. Adjusting the oil volume fraction employed for formation of the initial w/o emulsion and the w/o emulsion volume fraction subsequently used to produce the double emulsion allowed control over the final oil volume fraction in the double emulsions. Figure 4 shows the phase diagram constructed from systematic variation of these two parameters. In practice, only oil volume fractions of 50−80% produced stable w/o Pickering emulsions in the initial homogenization step. Below an oil volume fraction of 50%, rapid demulsification was observed. When using 90% oil, no emulsion was formed as there were insufficient silica particles present in the 10% aqueous phase to produce stable emulsion droplets. For the stable w/o Pickering emulsions prepared using oil volume fractions of 50−80%, homogenization of the oil phase with an aqueous dispersion containing PEI/silica hybrid particles (1.0% w/w silica; PEI/silica mass ratio = 0.075; pH 10) resulted in stable w/o/w double emulsions up to a total oil volume fraction of around 40%. Above this critical value, only w/o Pickering emulsions were obtained, rather than the desired double emulsions. Using the w/o Pickering emulsion prepared at an oil volume fraction of 50%, stable double emulsions can be obtained that contain oil volume fractions ranging from 5 to 40%. Figure 5 shows the change in volume-average oil droplet diameter of these double

emulsions as the oil volume fraction is systematically varied, as judged by laser diffraction studies. By controlling the volume fraction of the w/o Pickering emulsion added to the aqueous dispersion of PEI/silica hybrid particles (1.0% w/w silica; PEI/ silica mass ratio = 0.075; pH 10), the volume-average droplet diameter of the emulsions can be readily varied from 20 to 160 μm. Similarly, good control over the mean droplet diameter can be achieved when varying the oil volume fraction for the preparation of o/w Pickering emulsions using a PEI/silica mass ratio of 0.075 (see Figure S2). An oil-soluble dye (Nile Red) was dissolved in the aldehyderich oil prior to emulsion formation to facilitate fluorescence microscopy studies of the resulting droplets. Figure 6a shows an o/w Pickering emulsion prepared using hydrophilic PEI/silica particles (1.0% w/w silica, PEI/silica mass ratio = 0.075, pH 10) at an oil volume fraction of 50%. The fluorescent Nile Red dye is confined entirely within the oil droplets, as expected. Figure 6b shows a w/o Pickering emulsion prepared using hydrophobic PEI/silica particles (1.0% w/w silica, PEI/silica mass ratio = 0.50, pH 10) at an oil volume fraction of 50%. In this case, the Nile Red is located exclusively within the continuous phase. The double emulsions imaged in Figure 6 were each prepared using the same initial w/o emulsion (see Figure 6b). Figure 6c shows a w/o/w emulsion with an oil volume fraction of 10%, whereas Figure 6d depicts an w/o/w emulsion comprising 25% oil. In the former case, almost all of the oil droplets encapsulate a single aqueous droplet. However, in the latter case, the oil droplets are somewhat larger and hence can accommodate multiple aqueous droplets. An interesting characteristic of these new double emulsions is that the hydrophilic and hydrophobic PEI/silica particles utilized as Pickering emulsifiers are each decorated with multiple 2708

dx.doi.org/10.1021/la500219m | Langmuir 2014, 30, 2703−2711

Langmuir

Article

droplets, this formulation minimizes the probability of intercolloidosome fusion. In contrast, Route B requires two separate additions of the water-soluble PEG-DGE cross-linker during preparation of the double emulsion in order to ensure that covalent cross-linking occurs at both the w/o and o/w interfaces. More specifically, PEG-DGE must be added to the aqueous phase prior to homogenization to cross-link the initial w/o Pickering emulsion at the inner surface of the water droplets. Further PEG-DGE addition to the aqueous continuous phase is then required after formation of the w/ o/w double emulsion to cross-link the outer surface of the w/o droplets. In principle, this second addition could potentially increase the probability of intercolloidosome fusion. However, it is worth emphasizing that, if the epoxy−amine cross-linking chemistry is incompatible with the oil phase (which may well be the case with various oils other than that used in the present study), then Route B becomes an important alternative approach. This is because the water-soluble PEG-DGE crosslinker remains in the aqueous phase and never comes into contact with the oil. In practice, when the cross-linking strategies outlined in Figure 7 are utilized in the preparation of double emulsions, laser diffraction data and optical microscopy images (see Figure S3) indicate that the mean droplet diameter (≈40 μm) obtained in each case is comparable to that observed for the corresponding double emulsion formed in the absence of any cross-linker. Thus, there is actually little evidence for significant intercolloidosome fusion when using either polymeric crosslinker under the stated conditions. After an ethanol challenge to remove the oil phase, optical microscopy studies confirm that the resulting double emulsions (see Figure 8A) and colloidosomes-in-colloidosomes (see Figure 8B,C) remain intact. In the former case, this suggests that cross-linking is not actually required to stabilize the double emulsion in the presence of excess alcohol, in contrast to observations reported for conventional Pickering emulsions.6 Both the double emulsion droplets and the colloidosome-in-colloidosome microcapsules remain spherical, if somewhat collapsed. Presumably, this is because the double layer of PEI/silica hybrid particles provides a thicker shell that confers greater structural integrity than the particle monolayer present in a simple Pickering emulsion. The ultrahigh vacuum conditions required for scanning electron microscopy (SEM) studies causes collapse of both the double emulsion droplets and the colloidosome-in-colloidosome microcapsules to form lowdimensional “pancakes” (see Figure 8D−F). SEM confirms that covalent cross-linking is not essential for a double emulsion to survive an alcohol challenge as both the double emulsion precursor (Figure 8D) and subsequent colloidosomes-incolloidosomes produce similar “pancakes” (see Figure 8E,F). There also appears to be some evidence for the presence of the encapsulated secondary droplets (resulting from the initial w/o Pickering emulsion) within these collapsed “pancakes”. The double cross-linking strategy to produce colloidosomesin-colloidosomes, as summarized in Figure 7, may offer enhanced encapsulation/retention performance compared to conventional colloidosome microcapsules, which are known to be rather leaky.28,30,32 Furthermore, using PEI/silica particles in the formation of w/o/w double emulsions provides long-term stability against droplet coalescence, since optical microscopy images recorded for 15-month-old double emulsions suggest no discernible change in mean droplet diameter and confirm retention of the distinctive double emulsion character (see

Figure 8. Optical microscopy images obtained following the addition of excess ethanol to (A) a non-cross-linked w/o/w double emulsion, (B) a w/o/w double emulsion cross-linked using PEG-DGE, and (C) a w/o/w double emulsion cross-linked using PPG-DGE. Scanning electron microscopy images obtained following the same alcohol challenge followed by drying at 20 °C: (D) a non-cross-linked w/o/w double emulsion, (E) a w/o/w double emulsion cross-linked using PEG-DGE, and (F) a w/o/w double emulsion cross-linked using PPGDGE. In each case the oil volume fraction was 25%.

primary and secondary amine groups originating from the PEI chains, which offers the opportunity for interfacial cross-linking chemistry. Covalent cross-linking at both interfaces should prevent droplet coalescence as well as potentially improving the retention of encapsulated water-soluble actives. Moreover, as far as we are aware, the resulting colloidosomes-in-colloidosomes would be a novel construct in colloid science. We have previously reported that PEI/Laponite-stabilized o/w Pickering emulsions can be converted into conventional covalently stabilized colloidosomes using either an oil-soluble bisepoxyfunctionalized poly(propylene glycol) [PPG-DGE] or a watersoluble bisepoxy-functionalized poly(ethylene glycol) [PEGDGE] cross-linker.29 Such colloidosomes are sufficiently robust to survive an alcohol challenge, which removes both the oil droplet phase and the aqueous continuous phase.29 The two polymeric cross-linkers utilize the same epoxy−amine chemistry and differ only in their spatial location. Figure 7 shows the two routes utilized to produce colloidosome-in-colloidosome microcapsules. Route A is preferred as the oil-soluble PPGDGE can be dissolved in the oil phase prior to homogenization and this reagent cross-links the PEI/silica hybrid particles located at both the internal and external interfaces simultaneously. Moreover, as the PPG-DGE is confined to the oil 2709

dx.doi.org/10.1021/la500219m | Langmuir 2014, 30, 2703−2711

Langmuir

Article

for permission to publish this work. The two reviewers of this manuscript are thanked for their helpful comments.

Figure S4). These novel microcapsules also allow the possibility of simultaneously encapsulating separate hydrophilic and hydrophobic payloads within the same microcapsule, which may offer new opportunities within the “home and personal care” sector.





CONCLUSIONS Poly(ethylene imine) (PEI) has been utilized to modify the surface properties of fumed silica particles in order to produce an effective hybrid Pickering emulsifier for a proprietary aldehyde-rich oil. Systematic variation of the PEI/silica mass ratio enables the phase inversion of such Pickering emulsions from o/w type to w/o type. Moreover, highly stable double emulsions can be prepared using a single batch of fumed silica particles, simply by judicious adjustment of the adsorbed amount of PEI to afford either hydrophilic or hydrophobic PEI/ silica hybrid particles prior to homogenization. By varying the oil volume fraction, a series of w/o/w double emulsions can be made with oil contents ranging from 5 to 42%. By controlling the volume fraction of the initial w/o Pickering emulsion (prepared using the hydrophobic PEI/silica particles as an emulsifier) added to an aqueous dispersion of hydrophilic PEI/ silica particles prior to homogenization, the mean diameter of the resulting polydisperse double emulsions can be conveniently adjusted from 20 to 160 μm. Fluorescence microscopy studies confirm that either single or multiple water droplets can be encapsulated within the oil droplets, depending on the precise formulation used to prepare the w/o/w double emulsions. These double emulsions can be covalently stabilized via an epoxy−amine reaction between the physically adsorbed PEI chains and either an oil-soluble or a water-soluble bisepoxybased polymeric cross-linker to produce colloidosome(s)-incolloidosomes, which appear to be a new concept in the field of colloid science. Although such cross-linking is not a prerequisite for the double emulsion to withstand an alcohol challenge, covalent stabilization minimizes the probability of droplet coalescence and may offer opportunities for enhanced retention of water-soluble actives.



ASSOCIATED CONTENT

S Supporting Information *

Estimated c log P values for the various components of the proprietary aldehyde-rich oil, hydrodynamic diameter vs pH and zeta potential vs pH curves for Cab-O-Sperse 2012A fumed silica particles, Mastersizer droplet size distributions of o/w Pickering emulsions prepared using varying oil volume fractions, optical microscopy images and Mastersizer droplet size distributions of both non-cross-linked double emulsions and covalently cross-linked colloidosomes, optical microscopy images of a 15-month-old double emulsion. This material is available free of charge via the Internet at http://pubs.acs.org.



REFERENCES

(1) Ramsden, W. Separation of solids in the surface-layers of solutions and ‘Suspensions’ (Observations on surface-membranes, bubbles, emulsions, and mechanical coagulation). Preliminary Account. Proc. R. Soc. London 1903, 72, 156−164. (2) Pickering, S. U. Emulsions. J. Chem. Soc. 1907, 91, 2001−2021. (3) Finkle, P.; Draper, H. D.; Hildebrand, J. H. The theory of emulsification. J. Am. Chem. Soc. 1923, 45, 2780−2788. (4) Binks, B. P. Particles as surfactants - similarities and differences. Curr. Opin. Colloid Interface Sci. 2002, 7, 21−41. (5) Binks, B. P.; Lumsdon, S. O. Influence of particle wettability on the type and stability of surfactant-free emulsions. Langmuir 2000, 16, 8622−8631. (6) Williams, M.; Armes, S. P.; York, D. W. Clay-based colloidosomes. Langmuir 2012, 28, 1142−1148. (7) Cui, Y.; van Duijneveldt, J. S. Microcapsules composed of crosslinked organoclay. Langmuir 2012, 28, 1753−1757. (8) Li, J.; Stöver, H. D. H. Doubly pH-responsive pickering emulsion. Langmuir 2008, 24, 13237−13240. (9) Midmore, B. R. Synergy between silica and polyoxyethylene surfactants in the formation of O/W emulsions. Colloids Surf., A 1998, 145, 133−143. (10) Binks, B. P.; Rodrigues, J. A.; Frith, W. J. Synergistic interaction in emulsions stabilized by a mixture of silica nanoparticles and cationic surfactant. Langmuir 2007, 23, 3626−3636. (11) Cui, Z. G.; Yang, L. L.; Cui, Y. Z.; Binks, B. P. Effects of surfactant structure on the phase inversion of emulsions stabilized by mixtures of silica nanoparticles and cationic surfactant. Langmuir 2010, 26, 4717−4724. (12) Binks, B. P.; Isa, L.; Tyowua, A. T. Direct measurement of contact angles of silica particles in relation to double inversion of pickering emulsions. Langmuir 2013, 29, 4923−4927. (13) Binks, B. P.; Lumsdon, S. O. Catastrophic phase inversion of water-in-oil emulsions stabilized by hydrophobic silica. Langmuir 2000, 16, 2539−2547. (14) Binks, B. P.; Philip, J.; Rodrigues, J. A. Inversion of silicastabilized emulsions induced by particle concentration. Langmuir 2005, 21, 3296−3302. (15) Binks, B. P.; Fletcher, P. D. I.; Holt, B. L.; Beaussoubre, P.; Wong, K. Phase inversion of particle-stabilised perfume oil-water emulsions: experiment and theory. Phys. Chem. Chem. Phys. 2010, 12, 11954−11966. (16) Zhang, J.; Li, L.; Wang, J.; Sun, H.; Xu, J.; Sun, D. Double inversion of emulsions induced by salt concentration. Langmuir 2012, 28, 6769−6775. (17) Cui, Z. G.; Shi, K. Z.; Cui, Y. Z.; Binks, B. P. Double phase inversion of emulsions stabilized by a mixture of CaCO3 nanoparticles and sodium dodecyl sulphate. Colloids Surf., A 2008, 329, 67−74. (18) Oza, K. P.; Frank, S. G. Multiple emulsions stabilized by colloidal microcrystalline cellulose. J. Dispersion Sci. Technol. 1989, 10, 163−185. (19) Sekine, T.; Yoshida, K.; Matsuzaki, F.; Yanaki, T.; Yamaguchi, M. A novel method for preparing oil-in-water-in-oil type multiple emulsions using organophilic montmorillonite clay mineral. J. Surfactants Deterg. 1999, 2, 309−315. (20) Aveyard, R.; Binks, B. P.; Clint, J. H. Emulsions stabilised solely by colloidal particles. Adv. Colloid Interface Sci. 2003, 100, 503−546. (21) Thompson, K. L.; Armes, S. P.; York, D. W. Preparation of Pickering emulsions and colloidosomes with relatively narrow size distributions by stirred cell membrane emulsification. Langmuir 2011, 27, 2357−2363. (22) Utada, A. S.; Lorenceau, E.; Link, D. R.; Kaplan, P. D.; Stone, H. A.; Weitz, D. A. Monodisperse double emulsions generated from a microcapillary device. Science 2005, 308, 537−541.

AUTHOR INFORMATION

Corresponding Author

*E-mail s.p.armes@sheffield.ac.uk (S.P.A.). Notes

The authors declare no competing financial interest.



ACKNOWLEDGMENTS We thank P&G Technical Center (Newcastle-upon-Tyne, UK) for an Industrial EPSRC CASE award to support M.W. and also 2710

dx.doi.org/10.1021/la500219m | Langmuir 2014, 30, 2703−2711

Langmuir

Article

(23) Chu, L.-Y.; Utada, A. S.; Shah, R. K.; Kim, J.-W.; Weitz, D. A. Controllable monodisperse multiple emulsions. Angew. Chem., Int. Ed. 2007, 46, 8970−8974. (24) Abate, A. R.; Thiele, J.; Weitz, D. A. One-step formation of multiple emulsions in microfluidics. Lab Chip 2011, 11, 253−258. (25) Dinsmore, A. D.; Hsu, M. F.; Nikolaides, M. G.; Marquez, M.; Bausch, A. R.; Weitz, D. A. Colloidosomes: Selectively permeable capsules composed of colloidal particles. Science 2002, 298, 1006− 1009. (26) Laib, S.; Routh, A. F. Fabrication of colloidosomes at low temperature for the encapsulation of thermally sensitive compounds. J. Colloid Interface Sci. 2008, 317, 121−129. (27) Cayre, O. J.; Noble, P. F.; Paunov, V. N. Fabrication of novel colloidosome microcapsules with gelled aqueous cores. J. Mater. Chem. 2004, 14, 3351−3355. (28) Thompson, K. L.; Armes, S. P.; Howse, J. R.; Ebbens, S.; Ahmad, I.; Zaidi, J. H.; York, D. W.; Burdis, J. A. Covalently crosslinked colloidosomes. Macromolecules 2010, 43, 10466−10474. (29) Walsh, A.; Thompson, K. L.; Armes, S. P.; York, D. W. Polyamine-functional sterically stabilized latexes for covalently crosslinkable colloidosomes. Langmuir 2010, 26, 18039−18048. (30) Fielding, L. A.; Armes, S. P. Preparation of Pickering emulsions and colloidosomes using either a glycerol-functionalised silica sol or core-shell polymer/silica nanocomposite particles. J. Mater. Chem. 2012, 22, 11235−11244. (31) Yuan, Q.; Cayre, O. J.; Fujii, S.; Armes, S. P.; Williams, R. A.; Biggs, S. Responsive core-shell latex particles as colloidosome microcapsule membranes. Langmuir 2010, 26, 18408−18414. (32) Yow, H. N.; Routh, A. F. Release profiles of encapsulated actives from colloidosomes sintered for various durations. Langmuir 2009, 25, 159−166. (33) Bon, S. A. F.; Cauvin, S.; Colver, P. J. Colloidosomes as micronsized polymerisation vessels to create supracolloidal interpenetrating polymer network reinforced capsules. Soft Matter 2007, 3, 194−199. (34) Cayre, O. J.; Biggs, S. Hollow microspheres with binary porous membranes from solid-stabilised emulsion templates. J. Mater. Chem. 2009, 19, 2724−2728. (35) Bon, S. A. F.; Chen, T. Pickering stabilization as a tool in the fabrication of complex nanopatterned silica microcapsules. Langmuir 2007, 23, 9527−9530. (36) Fleer, G. J.; Cohen Stuart, M. A.; Scheutjens, J. M. H. M.; Cosgrove, T.; Vincent, B. Polymers at Interfaces; Chapman and Hall: London, 1993.

2711

dx.doi.org/10.1021/la500219m | Langmuir 2014, 30, 2703−2711