Langmuir 2006, 22, 67-73
67
Structural Evolution of Polymer-Stabilized Double Emulsions Wang Yafei,† Zhang Tao,‡ and Hu Gang*,§ Department of Physics, Beijing Normal UniVersity, Beijing, China, Department of Materials Science & Engineering, Nanjing UniVersity, Nanjing, China, and Department of Physics, Hong Kong Baptist UniVersity, Hong Kong SAR, China ReceiVed August 25, 2005. In Final Form: October 15, 2005 Polymer-stabilized double emulsions are produced by a two-step process, high shear emulsification in the primary and membrane emulsification in the secondary. By repeated fractionation after each emulsification, we obtain monodisperse double emulsions with the size of the complex droplets ranging from submicrometer to a few micrometers. With osmotic pressure balance between the inner and outer phases, the polymer-stabilized double emulsions remain stable for a year at room temperature without structure deterioration. We generalize laser light scattering to probe the structure and internal dynamics of the complex system by including the effects of the amplitude fluctuations of the scattered fields. Both static light scattering (SLS) and dynamics light scattering (DLS) can resolve the inclusions inside the complex droplets. Water-soluble nonionic surfactants are used to induce destabilization of double emulsions. We find that a double emulsion turns into a simple emulsion within a minute at a surfactant concentration of less than 10-3 mol/L. We demonstrate that DLS is a powerful technique to study the kinetics of destabilization of double emulsions. Coalescence between the internal droplets and the external continuous phase is identified as a major release pathway.
1. Introduction Unlike a simple emulsion, in which one liquid is dispersed in another liquid, a double emulsion is composed of droplets of one liquid dispersed in host droplets of a second liquid, which is in turn dispersed in a final continuous phase. The complex systems of droplets inside drops have attracted ever-growing attention due to the theoretical and practical aspects of emulsion formation, stabilization, and structural evolution.1-3 In a double emulsion, the internal disperse phase is usually miscible with the external continuous phase while the intermediate disperse liquid is immiscible with the internal and external phases, such as a waterin-oil-in-water (W/O/W) double emulsion or an oil-in-waterin-oil (O/W/O) double emulsion. The stability of the complex structure is often a major concern for many practical applications. Only recently can kinetically stable double emulsions be made using polymeric surfactants,4-7 allowing their detailed properties to be studied. Simple O/W or W/O emulsions are inherently unstable from a thermodynamic point of view. This is also true for double emulsions. The fundamentals of stability concern the compositions of various phases, the physical chemistry of the interfaces, and the kinetics of the complex structure. Although there is an ever-increasing interest in the subject, the area remains somewhat empirical in that each system is highly specific. Double emulsions are usually prepared with at least two kinds of * To whom correspondence should be addressed. Email: ganghu@ hkbu.edu.hk † Beijing Normal University. ‡ Nanjing University. § Hong Kong Baptist University. (1) Bibette, J.; Lead-Calderon, F.; Schmitt, V.; Poulin, P. Emulsion Science: Basic Principles; Springer: New York, 2002. (2) Sjoblom, J. Encyclopedic Handbook of Emulsion Technology; Marcel Dekker: New York, 2001. (3) Ficheux, M. F.; Bonakdar, L.; Leas-Calderon, F.; Bibette, J. Langmiur 1998, 14, 2702. (4) Benichou, A.; Aserin, A.; Garti, N. AdV. Colloids Interface Sci. 2004, 108, 29. (5) Kanouni, M.; Rosano, H. L.; Naouli, N. AdV. Colloids Interface Sci. 2002, 99, 229. (6) Goubault, C.; Pays, K.; Olea, D.; Gorria, P.; Bibette, J.; Schmitt, V.; LealCalderon, F. Langmiur 2001, 17, 5184. (7) Michaut, F.; Hebraud, P.; Perrin, P. Polym. Int. 2003, 52, 594.
emulsifiers since the structure of complex droplets involves two oil-water interfaces with an opposite curvature. Compared with simple emulsions, there are more stringent requirements for the choice of a pair of surfactants to obtain a stable double emulsion, and the chemical nature of various components in each phase also affects the stability of the structure. The important applications of double emulsions in pharmaceuticals, cosmetics, and foods are aimed at controlled release of encapsulated matter.8-10 In practice, ideal double-emulsion products are formulated to remain kinetically stable on the shelf for a long time, and release of the internal active matter is triggered by external induction or controlled by change of environment. Much effort has been devoted to enhance the stability of the complex structure and understand the mechanisms of destabilization. Therefore, the ability to track the evolution of the internal disperse phase is essential to the understanding of the kinetic stability of a new formulation. Optical microscopy is a direct method to observe the structure and slow dynamics of double emulsion systems. However, in practical applications, the preferred size of double emulsion droplets is on the order of a micrometer or submicrometer and the inclusion droplets are even smaller. The high curvature of the host droplets introduces serious aberrations to obviate good imaging of the internal structure, and the vigorous Brownian motion of the inclusion and host droplets also limits resolution of optical microscopy. It is well-known that laser light scattering is a convenient and powerful tool to probe the structure and dynamics of a dispersion system.11,12 The structural length scale of double emulsions falls in the range of the visible wavelength of light, and the dynamics are well within the coverage of a digital correlator. Despite the promising application of light scattering in studying double emulsions, there are significant (8) Pistel, K. F.; Kissel, T. J. Microencapsulation 2000, 17, 467. (9) Gallarate, M.; Carlotti, M. E.; Trotta, M.; Bovo, S. Int. J. Pharm. 1999, 188, 233. (10) Benichou, A.; Aserin, A.; Garti, N. Polym. AdV. Technol. 2002, 13, 1019. (11) Kerker, M. The Scattering of Light; Academic Press: New York, 1969. (12) Berne, B. J.; Pecora, R. Dynamic Light Scattering; Wiley: New York, 1976.
10.1021/la0523255 CCC: $33.50 © 2006 American Chemical Society Published on Web 11/12/2005
68 Langmuir, Vol. 22, No. 1, 2006
experimental difficulties which have prevented their observation. First, unlike light scattering from structureless particles, scattered light from a double emulsion carries the information of the structure and dynamics of the complex droplets, which complicates the interpretation of laser light scattering. Second, it usually requires a reasonably monodisperse sample to get any detailed understanding of the properties of the complex structure. The light scattering signal is due to the contrast of refractive indices of different phases. In a double emulsion, there exist the index mismatch between the inclusions and the hosts and the mismatch between the host droplets and the outer continuous phase. At a low volume fraction, the intensity of static light scattering (SLS) versus the scattering wavevector, I(q), should exhibit the structure of a complex droplet, where small droplets are encapsulated inside a host drop. The temporal autocorrelation function obtained from dynamic light scattering (DLS) results from the combined dynamics of the inclusion and host droplets. The host droplets execute free Brownian motion in the continuous phase, while the movement of the inclusions is restricted inside the hosts. The relaxation times of the two dynamics vary with the sizes and volume fractions of the two disperse droplets, as well as the viscosities of the intermediate and external phases. In this paper, we present a light scattering study of the structure of a double emulsion system and the structural evolution once destabilization of a double emulsion is induced. Double emulsions are made by a two-step emulsification process. In the formulation of double emulsions, a polymeric surfactant is used as an emulsifier to stabilize a primary W/O emulsion. To enhance the stability of the inverse emulsion, some salt and water-soluble polymer is dissolved in the disperse phase to elevate the osmotic pressure against Laplace pressure. With osmotic pressure balance between the inner and outer aqueous phases, the W/O/W double emulsions remain stable at room temperature for a year without significant deterioration. To observe the complex structure and dynamics of double emulsions using laser light scattering, we overcome the two major experimental obstacles; we exploit a method of fractionations after each step of emulsification to obtain the requisite monodisperse W/O/W double emulsions, and we incorporate the amplitude fluctuations in the scattering signal to account for the inclusions of complex droplets. It is difficult to rigorously formulate light scattering from inhomogeneous particles such as double emulsion droplets, in which multiple random dispersions of inclusions are encapsulated. Although the approach of equivalent refractive index may catch some scattering features of inhomogeneous particles with the assumption that the inclusions are much smaller than the wavelength of light, it is impractical to extract internal dynamics due to the movement of the inclusions.13,14 Noting the composite scattering nature of double emulsion droplets, we model the scattering amplitude as the sum of a time-independent component and a fluctuating component. A similar treatment has been discussed in the study of shape fluctuations of emulsions and microemulsions, even though the dynamics of the amplitude fluctuation is quite different.15,16 We demonstrate that light scattering can probe the complex structure and evolution of double emulsions when there occurs a variation of the number and size of droplets of the internal phase. Both SLS and DLS measurements on the monodisperse samples can resolve the inclusions inside the complex droplets. The measurement of I(q) on a double emulsion exhibits a distinguishing feature compared with a simple (13) Secker, D. R.; Kays, P. H.; Greenaway, R. S.; Hirst, E.; Bartlay, D. L.; Videen, G. Appl. Opt. 2000, 39, 5023. (14) Doicu, A.; Wriedt, T. J. Opt. A: Pure Appl. Opt. 2001, 3, 204. (15) Gang, H.; Krall, A. H.; Weitz, D. A. Phys. ReV. Lett. 1994, 73, 3435. (16) Gang, H.; Krall, A. H.; Weitz, D. A. Phys. ReV. E. 1995, 52, 6289.
Yafei et al.
emulsion, especially, in the high-q regime. Some water-soluble nonionic surfactants are used to destabilize polymer-stabilized double emulsions. To investigate the structural evolution, we make DLS measurements on the kinetic process by taking sequential short runs of temporal autocorrelation function. A two-relaxation model is proposed to describe the dynamics of the complex droplets. We find the size of the inclusion droplets remains approximately unchanged with evolving time, while the number of the droplets decreases. From DLS measurements, we identify that coalescence between the internal droplets and the outer continuous phase is a major pathway of induced destabilization. The method established here provides a feasible tool to study the stability of double emulsions and to test the effectiveness of a destabilizing agent. In practice, we may design a formulation so that release of encapsulated matter follows a particular pathway. 2. Experimental Method Sample Preparation. W/O/W double emulsions are made by a two-step emulsification process.6 To obtain monodisperse double emulsions, we apply repeated fractionations by centrifugation after each step of emulsification. The aqueous phase of the primary W/O emulsion comprises of a solution of 0.1 M salt (NaCl from SigmaAldrich) and 9 × 10-5 M dextran (Mw ≈ 500 000 g/mol from SigmaAldrich). The added salt and dextran raise the osmotic pressure of the internal aqueous phase and keep the emulsion stable against coarsening.17,18 The oil phase is comprised of dodecane with an oil-soluble polymeric surfactant, Arlacel P135 (PEG-30 dipolyhydroxystearate, Mw ≈ 5000 g/mol from Uniqema) dissolved as an emulsifier. To stabilize an emulsion against coalescence, there must be a sufficient amount of surfactants to cover the droplets’ surface, which can be estimated as (3V/a)σ/NA, ignoring a very low critical micelle concentration (cmc) of the polymeric surfactant, where V is the total volume of the disperse aqueous phase, a is the radius of water droplets, NA is the Avogadro constant, and σ is the effective surface coverage density, estimated as σ ≈ 10 - 3 /Å2 for Arlacel P135.5 Sometimes a hybrid of polymeric surfactant and a small percentage of short surfactant facilitates efficient emulsification and better steric stabilization. We use two methods to make primary W/O emulsions. One method is to use an UltraTurrax homogenizer with an applied shear rate of ∼4000 s-1 to break a crude emulsion at volume fraction φi ≈ 50% into a fine emulsion. The crude emulsion is prepared by slowly mixing the aqueous phase into the oil phase to form a W/O premixed emulsion with a size of droplets of tens of micrometers. The effective viscosity is raised significantly due to high φi. The other method is to apply ultrasonification to a crude emulsion at φi ≈ 10%. The average size can reach as small as 0.1 µm depending on the amount of added surfactants. After a few steps of fractionation through centrifugation, we obtain a W/O emulsion with a narrow size distribution. The polydispersity of droplets is around 15% determined by DLS. We note that the stability of the emulsion is enhanced due to the droplet reduction as shown by Meleson et al.,19 who have made nanoscale emulsion droplets using an ultrahigh-shear-rate microfluidizer. Then, the primary emulsion at φi ) 10% as a disperse phase is gently mixed into a continuous aqueous phase to form a crude W/O/W double emulsion, where the outer aqueous phase contains sodium dodecyl sulfate (SDS) as an outer emulsifier. The initial concentration of SDS is a few cmc, so that the final concentration is around 1 cmc after droplets reach the desired size. The outer aqueous phase also contains 0.1 M NaCl to balance the osmotic pressure of the inner aqueous phase. The volume fraction of double emulsion droplets, φh, is set at around 10%. In the second step of emulsification, the crude double emulsion is pressured through a polycarbonate membrane filter (from SPI) back and forth five times. Usually, the average size of double emulsion (17) Aronson, M. P.; Petko, M. F. J. Colloid Interface Sci. 1993, 159, 134. (18) Mezzenga, R.; Folmer, B. M.; Hughes, E. Langmiur 2004, 20, 3574. (19) Meleson, K.; Graves, S.; Mason, T. G. Soft Mater. 2004, 2, 109.
Structural EVolution of Double Emulsions
Langmuir, Vol. 22, No. 1, 2006 69 of the scattering amplitude due to scattering from a host droplet, and ∆bj(q,t) is the temporal fluctuating part of the scattering amplitude due to the Brownian motion of the inclusions, which can be written as Ni
∆b(q,t) )
∑c (q) exp[iqb‚br (t)] k
(2)
k
k
Figure 1. Optical micrograph of a polymer-stabilized W/O/W double emulsion obtained by repeated fractionation following two-step emulsification. The monodisperse host droplets at a volume fraction φh ≈ 60%, with a size dh ) 2.0 µm in diameter. Inside the host droplets, the inclusion droplets at a volume fraction φi ≈ 10%, with a size di ) 0.3 µm. The image of the inclusions inside the hosts is seriously obscured due to their vigorous Brownian motion and the high curvature of the hosts. droplets is close to the pore size of a membrane filter if the applied pressure is controlled within a range. With 10% pore density of 25 mm-diameter membrane, the flow rate ranges from 0.5 mL/s to 2 mL/s. There are commercially available membranes with a uniform pore size ranging from 0.01 to 20 µm. Choosing a membrane with a preferred pore size, we can get a double emulsion with the comparable size of droplets in the membrane emulsification. For light scattering experiments, we go through several steps of fractionation to obtain monodisperse samples. Figure 1 shows an optical micrograph of a W/O/W double emulsion sample with a size of the host droplets dh ) 2.0 µm, and a size of the inclusion droplets di ) 0.3 µm. The volume fraction of the inclusions in the disperse oil phase is φi ≈ 0.10. The vigorous Brownian motion of the inclusions and the large curvature of the host droplets obviate shape imaging of the inclusion droplets. A double emulsion with uniform-sized droplets allows for a detailed study of the stability under more controlled conditions. Recently, a novel microfluidics technique has been developed to make highly uniform-sized double emulsions with good control of the sizes of inner and outer droplets and the number of inner droplets in each larger drop.20 However, the twostep emulsification described here can be used to make double emulsions with much smaller droplets, and the production can be easily scaled up. The complex structure of double emulsions remains stable if the osmotic pressure is balanced between the inner and outer aqueous phases. It is straightforward to use optical microscopy to monitor the stability of double emulsion systems. However, as shown in Figure 1, optical microscopy can hardly resolve the details of the very small inclusions. It is worthwhile to explore laser light scattering on double emulsions. Light Scattering. We use laser light scattering to observe the complex structure and dynamics of double emulsions. Light scattering usually obtains the ensemble-averaged information of scatterers. To interpret quantitatively any results obtained from a light scattering experiment, we must calculate the q-dependence of the scattering intensity in SLS and the temporal autocorrelation function of the scattering electric field in DLS. Due to the alternate contrast of refractive indexes between the inner/outer phase and the middle phase, we assume that the scattering amplitude arising from double emulsion droplets comprises two parts, which in the single scattering limit can be modeled as N
E(q,t) )
∑[b (q) + ∆b (q,t)] exp[iqb‚RB (t)] j
j
j
where ck(q) is the scattering amplitude of an inclusion droplet, Ni is the number of the inclusions in a host drop, and b rk(t) is the position of the center of mass of an inclusion. In a very dilute limit, if we assume that the droplets are monodisperse and there is no interdroplet correlation, the average scattering intensity can be expressed as I(q) ) 〈|E(q)|2〉 ) σh(q) + Niσi(q)
(3)
ignoring a pre-constant, where the angular bracket 〈 〉 refers to the time or ensemble average, and σh(q) ) |b(q)|2 is the scattering cross section of a host drop and σi(q) ) |c(q)|2 is the scattering cross section of an inclusion droplet. Note I(q) is expressed as a sum of the scattering cross sections of a host droplet and the inclusion droplets, neglecting the correlation between the inclusions, which is only valid in a dilute limit. The temporal autocorrelation function from DLS measurement is given by g1(q,t) )
〈E*(q,0)E(q,t)〉 〈|E(q,0)|2〉
∝ (1 + A(q) e-1/6q 〈∆r (t)〉)‚e-1/6q 〈∆R (t)〉 2
2
2
2
(4)
where the two exponential decays are expressed in terms of the ensemble-averaged mean square displacement, 〈∆R2(t)〉 and 〈∆r2(t)〉 for the host drops and the inclusions, respectively. Even though the motion of the inclusions is restricted inside a host drop, both kinds of the droplets can be treated as free Brownian particles at early time. We have 〈∆R2(t)〉 ) 6Dht, and 〈∆r2(t)〉 ) 6Dit, where the diffusion coefficients for the hosts and the inclusions are Dh ) kBT/6πηwah, and Di ) kBT/6πηoai, respectively, and ah and ai are the corresponding radii, kB is Boltzmann’s constant, T is the temperature, ηo is the viscosity of the oil (host) phase, and ηw is the viscosity of the external aqueous phase. Thus, the correlation function is a combination of two exponentials with characteristic decay times τh ) 1/q2Dh and τi ) 1/q2Di. The factor A(q) ) Niσi(q)/σh(q) is the cross-section ratio of the inclusions to a host. Since the scattering cross section depends on the droplet’s size and the contrast of refractive indices, in a light scattering experiment, we can manipulate the index contrast to enhance the contribution of the interested component of a scattering signal so that the relaxation of a correlation function is dominated by one particular dynamics. An ALV/DSL/SLS-5022F laser light scattering system with a compact goniometer and fiber optics detection (from ALV GmbH) is used to make all measurements. The light source for light scattering experiments is a HeNe laser operating at a wavelength of 632.8 nm in a vacuum. However, most often an unfocused laser beam has to be used so that the radiation pressure on emulsion droplets will not affect the Brownian motion of the droplets. A laser beam of a diameter about 0.8 mm is incident onto the sample. The sample at φh ≈ 10 - 5 is in the single scattering limit and loaded in a 10 mm-diameter cylindrical glass vial. The output power of laser is only a few milliwatts. The scattered signal is efficiently collected due to the use of the fiber optics receiver and an active quenched avalanche diode (APD) unit. An ALV-5000/E multi-tau digital correlator is used to take the correlation function.
(1)
j
where q is the scattering wavevector, Rj(t) is the position of the center of mass of a host droplet, bj(q) is the time-independent part (20) Utada, A. S.; Lorenceau, E.; Link, D. R.; Kaplan, P. D.; Stone, H. A.; Weitz, D. A. Science 2005, 308, 537.
3. Results and Discussions Static Structure Factor of Double Emulsions. We apply SLS to probe the structure of double emulsions. Through fractionation, we prepare monodisperse double emulsions with a series of sizes of the host droplets ranging from 0.7 µm to several micrometers and a size of the inclusions of 0.3 µm. The
70 Langmuir, Vol. 22, No. 1, 2006
Figure 2. The average scattering intensity, I(q), versus the scattering wavevector, q, measured for double emulsions at φh ≈ 10-5. The q-4-dependence is drawn for a reference. (a) Measurements for a double emulsion with dh ) 0.7 µm. The solid symbols are for a low inclusion volume fraction φi ) 10%. The oscillation of I(q) shows our double emulsion sample is very monodisperse. The open symbols are for a higher inclusion volume fraction after swelling to φi ) 25%. The level-up of I(q) in the high-q regime indicates that scattering from the inclusions weighs more as φi increases. (b) Measurements for a double emulsion with dh ) 2.5 µm.
initial volume fraction of the inclusions is φi ≈ 0.10, and the size is di ) 0.3µm. We can determine φi inside the hosts by measuring the change of the cream volume when much excess sucrose is added into the external phase. The original salt concentration of the internal and external aqueous phases is balanced at 0.1 M. Once the balance is broken, there will be an exchange of water between the two aqueous phases. The complex structure of double emulsions at φh ∼ 10 - 5 is probed by SLS. Figure 2 shows the scattering intensity, I(q), versus the scattering wavevector, q, measured for two double emulsion samples with a different droplets’ size. The volume fraction of the inclusions is varied by reducing the salt concentration in the outer aqueous phase from 0.1 to 0.033 M, and the corresponding φi swells from 0.10 to 0.25. Since salt is not soluble in the oil phase and the ionic species can hardly permeate through the intermediate oil phase, the reduction of the external salt concentration results in unbalance of the aqueous osmotic pressures (chemical potential), which drives the transfer of water molecules by diffusion. The droplets swell until the inner and outer phases reach new osmotic balance. Note that too much salt reduction may result in structure destabilization, especially for host droplets with high internal encapsulation. The solid symbols represent measurements for double emulsions at φi ≈ 0.10, the open symbols for φi ≈ 0.25. Figure 2a is a plot of I(q) for a sample with a size of the host droplets dh ) 0.7 µm. At φi ≈ 0.10, the characteristic modulation of I(q) indicates that the sample is pretty monodisperse. In the low-q regime, the scattering cross section of a host droplet, σh(q) dominates due to the anisotropic scattering with a strong forward weighting. At φi ≈ 0.25, the positions of the modulation dips of I(q) move to lower values of q due to the swelling of droplets,
Yafei et al.
while in the high-q regime, I(q) levels up pronouncedly, in contrast to the data at φi ≈ 0.10. This is because the scattering signal from the inclusions weighs more in the high-q regime when the internal aqueous droplets swell to a higher volume fraction. In the high-q regime, the second term in eq 3, Niσi(q), more or less isotropic, is significantly larger than the first term, σh(q). A similar feature is shown in Figure 2b for the measurements of a double emulsion with dh ) 2.5 µm. Even though the modulation of I(q) is not resolvable due to the size increase and the polydispersity of droplets, I(q) levels up in the high-q regime, where scattering from the inclusions, Niσi(q), weighs over scattering from the hosts, σh(q). Destabilization of Double Emulsions. The practical use of double emulsions requires that we are not only able to formulate stable structured materials but also destabilize the structure in a controlled manner according to applications. The stability of double emulsions has been improved significantly due to the use of polymeric surfactants.4-7,21 However, in many applications, it requires an effective method to break the compartment structure, achieving controlled or targeted release of encapsulated matter. Diffusion/permeation and coalescence are two major release mechanisms naturally occurring due to thermal agitation.1,4,22 Diffusion/permeation is driven by the chemical potential difference between the inner and outer phases. Coalescence in double emulsions concerns a rupture of the thin film between inner droplets and the outer continuous phase. Sometimes there is the interplay between the two mechanisms since the accumulative effect of diffusion/permeation can reduce the surfactant coverage of droplets and facilitate coalescence. Besides thermal agitation, the key to induce coalescence in a double emulsion is to destabilize the inner interfacial layer inside complex droplets. One possible approach is to displace the original low-hydrophobic-lipophilic balance (HLB) surfactants by adsorption of higher-HLB surfactants on the inner interface. As demonstrated by Ficheux et al.,3 for a W/O/W double emulsion stabilized by surfactants (Span80/SDS), adding excess SDS results in enhancement of coalescence and irreversible release of encapsulated aqueous matter. We follow the same strategy to induce coalescence of a polymer-stabilized double emulsion. For a polymer-stabilized double emulsion, to achieve complete release, it is critical that a destabilizing agent can efficiently displace the original polymeric surfactants. The induced release in a double emulsion actually involves several kinetic processes governed by the physical chemistry of a destabilizing agent in different phases, such as solubility, diffusion coefficient, efficiency of surface adsorption, and ability to lower the coalescence barrier, the overall effectiveness of a destabilizing agent should be evaluated as a whole according to a particular application. We here demonstrate adding short nonionic surfactants in the continuous phase to destabilize polymer-stabilized W/O/W double emulsions. The nonionic surfactants are of alkylphenol ethoxylates, RC6H4(OC2H4)nOHtype, such as NP-7 (nonylphenol ethoxylates, n ≈ 7, Mw ≈ 553 g/mol, HLB ≈ 12, from Sigma-Aldrich). We find that the nonionic surfactant is very effective to induce coalescence between the internal droplets and the external continuous phase. Under an optical microscope we observe the evolution of the complex droplets adsorbed on the microscope glass slide when flushed with a NP-7 aqueous solution. Figure 3 shows a release process with four snapshots captured in the sequence 10 s apart. The release process happens very quickly under the flush of NP-7 (21) Sela, Y.; Magdassi, S.; Garti, N. Colloids Surf., A 1994, 83, 143. (22) Pays, K.; Giermanska-Kahn, J.; Pouligny, B.; Bibette, J.; Leal-Galderon, F. J. Controlled Release 2002, 79, 193.
Structural EVolution of Double Emulsions
Langmuir, Vol. 22, No. 1, 2006 71
Figure 4. A schematic of light scattering observation as a destabilizing agent is diffusing into the bulk sample. The droplets’ structure gradually varies in height and in time.
Figure 3. The structural evolution of the complex droplets under induced destabilization. The complex droplets with the inclusions at φi ≈ 0.50, adsorbed on a microscope glass slide, are flushed with a NP-7 solution of 10-3 M. The temporal sequence of snapshots shows the initial double emulsion at T ) 0 s turns into a simple emulsion at T ) 30 s.
solution at a concentration of less than 10-3 M, and the complex droplets turn into simple droplets by complete release of the internal droplets within 1 min. Coalescence between some internal droplets themselves also occurs to form larger internal droplets, but that is only an intermediate step, then they coalescence with the external phase at last. Actually, the time for complete release depends on the formulation of the complex droplets, the size, and the number of the internal droplets. We have also tried similar surfactants, such as NP-10 (n ≈ 10, Mw ≈ 682 g/mol), Triton X-100 (octylphenol, n ≈ 10), and Triton X-102 (n ≈ 12). For a double emulsion at a very low φh, the time required to completely release the internal droplets varies a little for different surfactants, ranging from a few seconds to a couple of minutes. If the concentration of nonionic surfactant NP-7 is at an order of 10-2 M, the total release is completed in less than a second. It is surprising that the nonionic surfactant can effectively induce destabilization of polymer-stabilized double emulsions. We argue that the effectiveness of the nonionic surfactant attributes to its higher solubility in the oil phase, and its efficiency to displace the original low-HLB polymeric surfactant on the surface of aqueous droplets. Much work is needed to understand how watersoluble nonionic surfactants can efficiently displace polymeric surfactants that are supposed to have better adsorption. As proposed by Kabalnov,23 the adsorption of high-HLB surfactants on aqueous droplets enhances coalescence between the droplets due to strong curvature frustration. The spontaneous curvature of the oil-water interface is determined by the type of surfactant adsorbed on the interface. High-HLB surfactants usually favor oil droplets in water, and low-HLB surfactants favor water droplets in oil. If there are surfactants adsorbed on the interface, in addition to the surface tension term, the bending energy also plays an important role in the total free energy. The rupture of a thin film due to hole nucleation needs to overcome the energy barrier. Usually thermal agitation is the driving force for natural coalescence. The adsorption of high-HLB surfactants on the oil-water interface does not favor water droplets, and if there is enough adsorption, curvature frustration can significantly lower the energy barrier of hole nucleation or even completely (23) Kabalnov, A. In Modern Aspects of Emulsion Science; Binks, B. P., Ed.; Royal Society of Chemistry: Cambridge, 1998.
remove the barrier. Therefore, the oil film breaks without a significant energy barrier, resulting in rapid release of the internal droplets. To have better statistical observation, we use DLS to monitor the destabilization kinetics of double emulsions. We suppress scattering from the host droplets by reducing the index contrast between the host droplets and the external aqueous phase. The external aqueous phase is replaced by a dextran (Mw ) 40 000) solution of 34 wt%. The NaCl concentration is adjusted to 0.05 M so that the osmotic pressure is still balanced between the inner and outer phases. The refractive indices of the inner, mid, and outer phases are 1.34, 1.42, and 1.39, respectively. The viscosity of the outer aqueous phase is raised to ηw ≈ 500 cP and the viscosity of the oil phase is ηo ≈ 2cP. To ensure the sample in the single scattering limit, φh is kept below 10-3. We load 2.0 mL of the double-volume sample into a 10 mm-diameter cylindrical vial and spread gently 10 µL of NP-7 on the top of the sample. The amount of the added NP-7 is about 10-2 M if homogenized in the solution. Since the density of NP-7 is lower than that of the continuous phase, NP-7 can only diffuse into the sample due to entropy mixing as shown schematically in Figure 4. The laser beam is 0.5 cm below the top surface of the sample. As known from the measurement of I(q), scattering from the host droplets dominates in the low-q regime while scattering from the inclusions becomes pronounced in the high-q regime. At a scattering angle θ ) 95°, we take a series of autocorrelation functions in the sequence of 2 min a run. By measuring the dynamics in shout runs, we can examine the slow kinetics of a double emulsion under the influx of NP-7 diffusion. Figure 5 shows the correlation functions at a series of evolving times. Each correlation function is taken for 2 min. One set of symbols represents data taken at a specified evolving time T. At time T ) 0 min, the correlation function, g1(t), relaxes with a single characteristic time, τi(T). Starting at time T ) 140 min, two relaxation processes appear; the first relaxation decays to a plateau, and the plateau decays at a much later time. The solid lines are fitted to the two-relaxation model expressed in eq 4. The fitting is reasonably good with such a simple model even without taking into account of the polydispersity. The first relaxation is due to the dynamics of the inclusions, while the second relaxation accounts for the Brownian motion of the hosts. Since the size ratio of the inclusions to the hosts is ai/ah ≈ 0.2 and the viscosity ratio of the oil phase to the external aqueous phase is ηo/ηw ≈ 4 × 10-3, the two characteristic relaxation times are separated by 3 orders of magnitude, as shown in Figure 5. As more surfactants diffuse into the scattering volume, the plateau level of the correlation function gradually rises, which implies the weighting of scattering from the inclusions decreases gradually as induced release proceeds. The first characteristic relaxation time, τi(T) is plotted versus evolving time, T, in Figure 6. It shows that there is no significant
72 Langmuir, Vol. 22, No. 1, 2006
Figure 5. Dynamic light scattering data |g1(t)|2 (symbols) measured at scattering angle θ ) 95° for a double emulsion at φh ≈ 10-3. The square of the field correlation function |g1(t)|2 evolves from a decay of a single relaxation to a two-step relaxation as nonionic surfactant molecules diffuse into the sample. A set of symbols represents the correlation function at a specified evolving time T. The solid lines are fits to eq 4, a model of scattering from a double emulsion, which takes into account of the time-dependent scattering amplitude consisting of a constant part and a fluctuating part. The first relaxation is due to dynamics of the inclusions and the second relaxation is due to the Brownian motion of the host droplets.
Yafei et al.
Figure 7. The evolution of the scattering strength factor defined as A(T) ) Ni(T)σi/σh. The decrease of A(T) with evolving time results from the number reduction of the internal droplets, Ni(T). The power law of T -1/2 is drawn for a reference.
high-HLB surfactant molecules adsorbed on the inner interface. At the end of light scattering experiment, we examined the sample under an optical microscope. We found that most of the original complex droplets had turned into simple emulsion droplets and there was no obvious change of the size of the host droplets. The scenario implies that, if osmotic pressure is balanced between the inner and outer phases, the induced release is through coalescence between the inner droplets and the outer continuous phase.
4. Summary
Figure 6. The evolution of the first decay constant of correlation functions, τi(T), versus evolving time, T, extracted from fits in Figure 5. The steady behavior of τi(T) suggests the size of the inclusions does not vary much.
variation of τi(T) as surfactants diffuse into the sample. The relaxation time is related to the dynamics of the inclusions by τi(T) ) q2Di ∝ 1/di, Therefore, the steady τi(T) implies the size of the inclusions does not change much during the induced release process. In practice, coalescence may occur at two levels, either between the inner droplets themselves or between the inner droplets and outer continuous phase. The steady τi(T) implies the latter scenario dominates even though the former scenario cannot be excluded. However, the scattering strength of the inclusions defined by A(q,T) ) Ni(T)σi(q)/σh(q) falls with evolving time, T, below the saturation time around 600 min, as shown in Figure 7. Because the scattering cross sections of droplets do not change much if the size does not vary, the kinetics results from the number reduction of the inclusions, Ni(T) due to coalescence. The power of T -1/2 is probably due to diffusion of surfactants. With a limited range of valid data, we do not attempt to overinterpret the result. The induced release actually involves several processes, and the kinetics can be complicated. In this particular situation, the surfactant molecules first diffuse in the outer continuous phase, which we think is responsible for the slow kinetics since the molecules need to diffuse a length L ) 0.5 cm, and with a typical diffusion coefficient D ≈ 5 × 10-6 cm2/s in water, it takes L2/D ≈ 8 × 102 min to reach the scattering volume, and then the surfactants permeate through the outer water-oil interface and enter the oil droplets, finally adsorbing onto the interface of internal water droplets. Coalescence takes place if there are enough
The stability of W/O/W double emulsions is significantly enhanced due to the use of polymeric surfactant and addition of polymer in the internal aqueous phase. The enhancement is mostly due to the increase of the range of steric interaction and insolubility of dextran in the oil phase. The preparation of uniform-sized double emulsions enables the study of the structured materials using light scattering. The structure and the stability of a double emulsion critically depend on the size or size distribution of droplets. Double emulsions with submicrometer-sized droplets have been prepared by membrane emulsification. The scale-up production is straightforward. The monodisperse double emulsions can stay kinetically stable for a year at room temperature without structure deterioration. The balance of osmotic pressure is required to keep the structure stable, and the pressure gradient can be used to swell or shrink the internal droplets by transferring water between the inner and outer phases. We extend the applicability of light scattering by incorporating the amplitude fluctuations of scattered light, which account for the internal structure and dynamics of the complex droplets. The light scattering probe of the encapsulated phase of double emulsions can monitor the structural evolution, which may be difficult to resolve using optical microscopy. Similar to simple emulsions, there are two major mechanisms responsible for destabilization of double emulsions, diffusion/ permeation and coalescence. However, due to the complex structure of double emulsions, the interplay between the two mechanisms can be used to destabilize a double emulsion and effectively release encapsulated matter. Formulation with only short surfactant stabilizers cannot meet the expectation of longterm stability. At the same time, some applications also demand effective pathways to break polymer-stabilized double emulsions. We have demonstrated that release of the internal phase of polymer-stabilized double emulsions can be effectively induced using short, nonionic surfactants. It is a viable method to achieve controlled or targeted release of encapsulated matter. However, it is not very clear how the nonionic surfactants displace polymers on the interface. Unlike the classical model of coalescence, in which the thermal fluctuations generate a nucleation hole in a thin film and the film ruptures if the hole is above a critical size
Structural EVolution of Double Emulsions
and grows, induced coalescence results from strong curvature frustration, where newly adsorbed surfactants on the interface of a thin film may change or even reverse the spontaneous curvature and thus the energy barrier of nucleation is lowered or completely removed, causing a rapid rupture of the film. DLS is for the first time applied to study the release kinetics of double emulsions. The light scattering probe conveniently provides ensemble information. It shows that induced coalescence can efficiently release the internal droplets of double emulsions. Double emulsions, as a carrier to deliver active materials, have significant advantages over simple emulsions in many applications. Formulation using polymer stabilizers can meet
Langmuir, Vol. 22, No. 1, 2006 73
the stability requirement. We have shown here that induced coalescence may provide an efficient method in controlled release of encapsulated matter with possibility to target a specific environment. The formulation flexibility and controlled stability promise more practical uses of double emulsions. In practice, much work is needed to identify effective stabilizers and destabilizers for a specific application. Acknowledgment. We are grateful to Mr. Chou Kai-hong for his technical support and acknowledge funding from Hong Kong CERG, HKBU2049/02P. LA0523255
