Emulsions — Recent Advances in Understanding

BERNARD P. BINKS

Surfactant Science Group, Department of Chemistry, University of Hull, Hull HU6 7RX, UK

1.1 Introduction

This chapter reviews the progress in the understanding of emulsions over the last ten years or so. The emphasis is on the factors affecting the type and subsequent stability of emulsions, and on the associated properties of surface-active molecules adsorbed at the oil-water interface. The field of emulsions is a vast area and so the literature covered is selective rather than comprehensive.

An emulsion may be defined as an opaque, heterogeneous system of two immiscible liquid phases ('oil' and 'water') where one of the phases is dispersed in the other as drops of microscopic or colloidal size (typically around 1 µm). There are two kinds of simple emulsions, oil-in-water (O/W) and water-in-oil (W/O), depending on which phase comprises the drops. Emulsions made by agitation of the pure immiscible liquids are very unstable and break rapidly to the bulk phases. Such emulsions may be stabilised by the addition of surface-active material which protects the newly formed drops from re-coalescence. An emulsifier is a surfactant which facilitates emulsion formation and aids in stabilisation through a combination of surface activity and possible structure formation at the interface.

This book is concerned with macromulsions and not ITLµITLemulsions. The latter are thermodynamically stable dispersions of oil and water, which means that they form spontaneously and are stable indefinitely. Being optically clear, their characteristic size lies in the range 5–50 nm. Most (macro)emulsions require the input of considerable amounts of energy for their formation and can only be stable in a kinetic sense. However, many systems of oil + water + surfactant which form microemulsions may be emulsified to emulsions and there is a growing interest in relating the properties of these emulsions to the known equilibrium phase behaviour of the corresponding microemulsions. Several recent reviews on microemulsions exist and an aim of this chapter is to discuss, as far as is possible, the behaviour of emulsions, stabilised by low molecular weight surfactants, in terms of the aggregation and adsorption in the micellar/microemulsion systems.

Several books on emulsions have appeared in the last decade. The ones devoted solely to emulsions include chapters on emulsion stability, food emulsions, crude oil emulsions, rheology, pharmaceutical emulsions and perfluorochemical emulsions as blood substitutes. Emulsions are so widely encountered in a huge variety of industries, e.g. agrochemical, food, pharmaceutical, paint, printing, petroleum, etc., that two World Congresses on Emulsions have been held in 1993 and 1997. The proceedings have been published and papers covered areas from manufacturing and stability to wetting and adhesion to applications. It would be impossible to review all aspects of emulsion science and technology in one article, and so the author refers to the many review articles which exist on the topics not to be discussed here, although some are the basis of subsequent chapters. These include emulsion formation, rheology, multiple emulsions, solid-stabilised emulsions, techniques for measurement, parenteral (fluorochemical or phospholipid-stabilised) emulsions, food emulsions, crude oil emulsions, and applications. This list is not exhaustive but it serves to illustrate the scope of the subject.

The chapter is organised into the following sections: Emulsion Type and the System Hydrophile–Lipophile Balance (HLB), Phase Inversion, Emulsion Stability, Gel Emulsions, and Forces between Oil–Water Interfaces.

1.2 Emulsion Type and the System Hydrophile–Lipophile Balance

1.2.1 Emulsions of Two Liquid Phases

Whether an emulsion is O/W or W/O depends on a number of variables like oil:water ratio, electrolyte concentration, temperature, etc. For most of this century, emulsion chemists have known that surfactants more 'soluble' in water tend to make O/W emulsions and surfactants more 'soluble' in oil tend to make W/O emulsions. This is the essence of Bancroft's rule, which states that the continuous phase of an emulsion tends to be the phase in which the emulsifier is preferentially soluble. The word 'soluble' is misleading, however, for two reasons. Firstly, a surfactant may be more soluble in, say, oil than in water in a binary system, but in the ternary system of oil + water -F surfactant it may partition more into water. A good example of this is with the anionic surfactant Aerosol OT (sodium bis-2-ethylhexylsulfosuccinate) which dissolves in heptane at 25 °C up to at least 0.5 m but has a solubility limit in water of only ~0.03 m]. An emulsion made from equal volumes of water and heptane at 25 °C is O/W, however. Secondly, no distinction is made between the solubility of monomeric or aggregated surfactant in oil or water. We will see that this is an important omission.

The first quantitative measure of the balance between the hydrophilic and hydrophobic moieties within a particular surfactant came in 1949 when Griffin introduced the concept of the HLB, or hydrophile–lipophile balance, as a way of predicting emulsion type from surfactant molecular composition. A major problem of the HLB concept is that the HLB numbers assigned to the neat surfactant take no account of the effective HLB of a surfactant in situ adsorbed at an oil-water interface. Thus, for example, a nonionic surfactant of low HLB number (and hence predicted to stabilise W/O emulsions) may form O/W emulsions at low enough temperatures. It therefore became clear that the prevailing conditions of temperature, electrolyte concentration, oil type and chain length and cosurfactant concentration can all modify the geometry of the surfactant at an interface and thus change the curvature of the surfactant monolayer, which in some way affects the preferred emulsion type.

There have been developments in understanding how the type of emulsion is related to the phase diagram of mixtures of oil + water + surfactant at equilibrium. These have come from the studies with microemulsions and as an example we take the case of a nonionic surfactant of the polyoxyethylene glycol ether type, CnEm. Let us consider equilibrium systems of heptane and water (equal volumes) containing C12E5. At low [surfactant], monomer distributes between oil and water but heavily in favour of the oil. The partition coefficient defined as (molar concentration in heptane)/(molar concentration in water) increases from ~130 at 10 °C to ~1500 at 50 °C. Above a critical surfactant concentration, reached in both phases and designated CμCwater (typically 5 x 10-5 M)] and cμcoil (typically 6-60 x 10-3 M, all additional surfactant in excess of this concentration is present in the form of aggregates, hence the symbol μ for 'microemulsion'. At low temperatures (<28°C) the aggregates are oil-in-water microemulsion droplets formed in the water phase, in equilibrium with excess oil containing monomeric surfactant (Winsor I system). The preferred curvature of the monolayer is around oil and may be termed positive. At higher temperatures (>30°C), aggregation occurs in the oil phase in the form of water-in-oil microemulsion droplets which are in equilibrium with monomeric surfactant in the aqueous phase (Winsor II system). The monolayer curvature is now negative. At intermediate temperatures (28-30 °C), three phases are formed (Winsor III system) consisting of a surfactant-rich phase and both excess oil and aqueous phases. Here, the monolayer has, on average, zero net curvature.

It is frequently observed that the type of emulsion (O/W, W/O or intermediate) formed by homogenisation of the Winsor system is the same as that of the equilibrium microemulsion,, e.g. emulsification of an O/W microemulsion plus excess oil generally gives an O/W emulsion, the continuous phase of which is itself an O/W microemulsion. It is not obvious why this should be so, since in the case of microemulsions their behaviour is determined in part by the spontaneous curvature of the surfactant monolayer stabilising the nanometre-sized droplets. For emulsions, the radii of curvature of micrometre-sized drops are of the order of 1000 times the molecular dimensions and so it is difficult to see how curvature effects are implicated. In order to test and understand Bancroft's rule, the type of emulsion at different temperatures must be determined at various fixed surfactant concentrations. This is readily done via conductivity measurements (Figure 1.1). At high [C12E5], sufficient to form microemulsion aggregates in equilibrium systems, emulsions invert from O/W to W/O at temperatures around the Winsor III region. Thus, Bancroft's rule holds but it is impossible to say whether preferred emulsion type is determined by the distribution of monomeric or aggregated surfactant. At low [C12E5], where only monomer surfactant exists, emulsions are O/W at low temperatures and remain O/W even at temperatures where Winsor II systems would form in systems of higher surfactant concentration. Thus, we see an apparent violation of Bancroft's rule since despite the surfactant monomer distribution being in favour of oil, the stable emulsions are O/W Similar findings were reported by Harusawa et al. using nonylphenol ethoxylated surfactants.

A more useful concept than that of surfactant HLB numbers is that of the system HLB, related to the locus of aggregate formation. Thus 'high' HLB systems are those in which the aggregates (micelles or O/W microemulsions) form in the water phase, whereas Tow' HLB systems are those in which aggregation occurs in the oil phase, either as reverse micelles or W/O microemulsion droplets. It is now more correct to say that the phase containing the surfactant aggregates becomes the continuous phase of an emulsion. From the foregoing discussion, it is clear that the continuous phase of an emulsion is not necessarily the phase containing the highest concentration of surfactant. Bancroft's rule and the system HLB apply to surfactant aggregates rather than to total surfactant present in the system.

The results quoted above have allowed testing of theories put forward to explain preferred emulsion type. In the kinetic theory of Davies, it was assumed that the type of emulsion was determined by the relative rates of coalescence of oil and water drops after emulsification. Coalescence can be modelled as (i) the approach of the drops and the formation of a plane-parallel film and (ii) thinning of the film to a critical thickness followed by film rupture. The kinetics of thinning of such emulsion films was predicted theoretically using a hydrodynamic model. This velocity of thinning is dependent upon the balance of forces acting at the interface of the approaching drops. As the two drops come close to each other, liquid flows out of the film to its thicker parts resulting in the convective flux of surfactant at the surface, thus perturbing its equilibrium distribution. This generates reverse fluxes tending to restore equilibrium, including surface diffusive flux and bulk fluxes from the film and drop. The difference in surface concentration of surfactant results in a variation of the interfacial tension which produces a surface force (tension gradient) opposite to the liquid flow. The rise in tension depends on the Gibbs elasticity G of the film, equal to 2εD if sufficiently thick, where εD is the surface dilational modulus. G is a measure of the film's resistance to thinning. With this model, it is shown that the velocity of thinning depends on the location of non-adsorbed surfactant. If surfactant is present mainly in the continuous (film) phase, it has to diffuse a long way from the film perimeter in order to reduce G. Since the driving force for this process is the gradient of surfactant concentration along the surface, this diffusion cannot eliminate the tension gradient which opposes thinning. If, however, the surfactant is more soluble in the dispersed (drop) phase, it must diffuse a much shorter distance, and since this flux is driven by the normal gradient of the concentration, it can counterbalance the convective flux and so relieve the elasticity and increase the thinning velocity. Summarising, theory predicts that emulsions containing most of the surfactant in the dispersed phase will coalesce faster than those where surfactant is mainly in the continuous phase, and so will not be the preferred emulsion type, in apparent accordance with Bancroft's rule. The results in Figure 1.1, alongside the known partitioning of monomers and aggregates, demonstrate that arguments of emulsion type based on film elasticity are not applicable in cases where significant partitioning of monomer to an oil phase occurs.

Recently, Petsev et al. have advanced an argument based on interfacial bending in order to explain the correspondence between emulsion (sub-micron in size) and microemulsion types. They assume that drop surfaces can deform on approach and, in addition to the energy of interfacial stretching upon collision, the energy of interfacial bending is also taken into account. This is the energy contribution caused by the variation of the interfacial curvature, H = -1/r, r being the drop radius. The corresponding contribution to the drop deformation energy is

Wb = -2πa2B0H (1.1)

provided (a/r [much greater than 1, where a is the film radius and B0 is the interfacial bending moment of a flat interface, equal to -4KH0 with K being the bending elasticity constant and H0 the spontaneous curvature. K is a measure of the energy required to bend unit area of a spherical interface. Assuming a ≈ r/50 and using equation (1.1), one obtains |Wb| = r(π/1250)|B0|. For oil-water interfaces saturated with surfactant, K is of the order of the thermal energy kT and |B|0 is of the order of 5 x 10-11 N. With r = 5 x 10-7 m, then Wb ≈ 15kT. In other words, bending effects can be significant for the interaction between sub-micron emulsion drops. Although the bending energy is known to be important for the interfaces of very high curvature as in microemulsion droplets, the above result seems surprising for emulsion drops. It is due to the fact that when the drop radius r increases, the bent area increases faster (a2 [varies] r2 than the bending energy per unit area decreases (H [varies] 1/r). For positive B0, the interface is bent around the oil phase (O/W) and the bending moment facilitates the formation of a flat film between two aqueous drops in oil (Wb< 0) but opposes the formation of a flat film between two oil drops in water (Wb > 0). In emulsions where the interfacial tension is low (say <0.1 mN m-1), simultaneous formation of microemulsion droplets and emulsion drops can occur (Figure 1.2). One can expect that when B0H< 0, the formation of stable drops is facilitated since Wb > 0, opposing deformation. This is the case for both macro- and microdrops depicted in Figure 1.2a, where microdroplets are in the continuous phase. On the other hand, H has opposite signs for macro- and microdrops in the configuration depicted in Figure 1.2b, where microdroplets are in the dispersed phase. B0H< 0 for microemulsion droplets but is positive for the macroemulsion drops. Consequently, the bending moment will favour the formation of a plane film between approaching macrodrops. This will facilitate substantial deformation, leading to flocculation and coalescence of the emulsion drops containing microemulsion droplets inside. The emulsion containing microemulsion droplets in the continuous phase will be more stable and survive, as observed in practice.

A similar theoretical argument, proposed by Kabalnov and Wennerstrom, explains the correspondence between the equilibrium phase behaviour and the emulsion type and stability by considering the final stage of drop coalescence in which two drops which have made contact are connected by a narrow neck or hole, as shown schematically in Figure 1.3. The surfactant monolayer is highly curved in the region of the neck whereas it is virtually planar everywhere else. As the neck grows there is a change in both the area of the intervening liquid film (ΔA)] and in its curvature. On one hand, emulsion film rupture is driven by reducing the interfacial area of the planar part of the film. On the other, the edge of the neck creates extra interfacial area and therefore a free energy penalty. The energy barrier to hole nucleation thus comes from an interplay between the free energy penalty at the edge of the hole and the free energy gain at the planar part.