Properties of Emulsions Stabilized with Milk Proteins: Overview of Some Recent Developments ERIC DICKINSON Procter Department of Food Science, University of Leeds, Leeds LS2 9JT, England

ABSTRACT The physico-chemical properties of oil-in-water emulsions that are stabilized by milk proteins are determined largely by the nature of the adsorbed layer at the surface of the dispersed droplets. Detailed information on the structure of the adsorbed layer is now available for some of the pure milk proteins, especially b-casein and b-lactoglobulin. Recent analysis of the segment density profiles that are normal to the surface has shown clear differences in layer structure between the disordered casein and the globular whey protein. The phosphoserine residues of b-casein are essential in providing the thickness and stericstabilizing properties of the layer. Surface coverage and segment density profile are sensitive to bulk protein concentration, pH, ionic strength, and calcium ion content. Functional properties of commercial milk protein ingredients are influenced also by compositional and structural heterogeneity. Unadsorbed protein may make additional contributions to the properties of casein-stabilized emulsions, especially creaming and flocculation behavior. Experimental studies of small oil-in-water droplets emulsified with pure milk proteins or commercial milk protein ingredients (i.e., sodium caseinate and whey protein concentrate) have provided useful insight into the major factors (temperature, pH, protein content, and calcium ions) affecting the rheological behavior of dairy emulsions and the key instability processes of creaming, flocculation, and coalescence. Insight into the effects of interactions of proteins and lipids and competitive adsorption on emulsion properties has been derived from rheological and stability experiments on systems containing mixtures of milk proteins and various small molecule emulsifiers (sorbitan esters, monoglycerides, and lecithins). Such experiments have demonstrated the sensitivity to emulsifier type and concentration of the orthokinetic coalescence stability of whey protein emulsions and the viscoelasticity of heat-set whey protein emulsion gels.

Received July 16, 1996. Accepted April 23, 1997. 1997 J Dairy Sci 80:2607–2619

( Key words: adsorbed layers, caseins, emulsion stability, whey proteins) Abbreviation key: d32 = mean droplet diameter, GMS = glycerol monostearate, R = molar ratio of surfactant to protein, SCF = self-consistent field. INTRODUCTION In the food industry, homogenization is widely used to extend the shelf-life and quality of milk and cream and, more widely, for finely dispersing oils and fats in food products such as ice cream, cream liqueurs, and processed cheese. Differences in the properties of dairy-based oil-in-water emulsions arise largely from the differences in structure and composition of the adsorbed milk protein layers at the surface of the fat globules (28, 108). Knowledge about the adsorbed layers of milk proteins has come from experimental studies on various types of systems, including homogenized milk, recombined milks made from skim milk (powder) plus milk fat, synthetic milk emulsions made from sodium caseinate or whey protein concentrate plus milk fat (or other food oils), and model emulsions made from individual milk proteins plus hydrocarbon (or triglyceride) oil. This article discusses the behavior of model systems containing single milk proteins and the relationship between the properties of these simple model emulsions and those of real dairy foods, which contain complex mixtures of milk proteins in various states of aggregation. Mixtures of milk proteins, in both soluble and dispersed form, are widely valued as food ingredients that have excellent emulsifying and emulsionstabilizing characteristics (32, 77, 106, 107). During homogenization, the milk protein acting as an emulsifier, in the form of individual molecules or protein aggregates, becomes rapidly adsorbed at the surface of the newly formed oil droplets. The resulting stericstabilizing layer protects the fine droplets against immediate recoalescence and provides long-term physical stability to the emulsion during subsequent processing and storage (55). An adsorbed protein monolayer at the oil-water interface is a very thin, dense layer of highly interacting polymer molecules (34). The main thermodynamic driving force for adsorption is removal of

2607

2608

DICKINSON

hydrophobic residues from the unfavorable environment of the bulk aqueous phase by displacement of structured water molecules from the close vicinity of the interface (68). An additional important driving force is the unfolding and reorganization of the native protein structure that is due to interaction with the interface. The large negative change in free energy associated with protein adsorption indicates that the process is effectively irreversible with respect to serial dilution of the bulk phase. Milk proteins are readily displaceable, however, in the presence of other competing surface-active species (i.e., small molecule emulsifiers), which can reduce the interfacial free energy (per unit area) to an even lower level than can the milk proteins (38, 57). In terms of adsorbed layer structure and mechanical properties, it is appropriate to distinguish between disordered flexible caseins and compact globular whey proteins (34, 38). As a first approximation, a casein monomer may be regarded as a complex linear copolymer that adsorbs to give an entangled monolayer of flexible chains having some sequences of segments in direct contact with the surface (trains) and others protruding into the aqueous phase (loops and tails). This simple train-loop-tail model is not adequate, however, to describe the molecular configuration of adsorbed whey protein. A closely packed globular protein monolayer is perhaps better modeled as a dense two-dimensional assembly of highly interacting deformable particles with rheological and structural properties that are similar to those of a concentrated heat-set globular protein gel (6, 68). Emulsion stability is a relative, kinetic concept (33, 35, 39). A stable emulsion has no discernible change, over an arbitrary period of observation, in the size distribution of the droplets, their state of aggregation, or their spatial arrangement within the sample vessel. The period used to determine stability may vary from hours to months depending on the situation. The state of aggregation of the droplets is dependent on interactions between adsorbed protein layers, which in turn depends on factors such as protein surface coverage, layer thickness, surface charge density, and aqueous solution conditions (especially pH, ionic strength, and calcium ion content). The most obvious manifestation of emulsion instability is creaming, which eventually leads to macroscopic phase separation into separate discernible regions of cream and serum. Apart from the distribution of droplet size, the most important factor affecting creaming kinetics is the state of flocculation of the emulsion. To understand how to control the creaming and flocculation of dairy emulsions, various key facJournal of Dairy Science Vol. 80, No. 10, 1997

tors affecting the structure and interactions of adsorbed layers of milk proteins must be considered. ADSORBED PROTEIN LAYERS b-CN Bovine b-CN is a nonglobular protein of 209 residues that can be represented to a first approximation as a flexible linear polyelectrolyte having negligible ordered secondary structure and no intramolecular crosslinks (119). At neutral pH, the isolated b-CN molecule carries a net charge of about –15 e (where e is the electronic charge), and the distinctly nonrandom distribution of hydrophilic and hydrophobic residues makes the molecule substantially amphiphilic, compared with as1-CN (54). A noteworthy feature of the varied distribution of hydrophobic residues in b-CN ( 2 9 ) is the presence of a sequence of 40 to 50 segments at the N-terminus, which is predominantly hydrophilic and includes many charged residues (including the 5 phosphoserines). What is the simplest statistical model of b-CN that can account for the adsorbed layer structure at the level required for a proper understanding of its role in emulsion stabilization? In analyzing the surface equation of state at liquid interfaces using the well-known train-loop-tail model (83), Graham and Phillips ( 8 6 ) made a very reasonable subdivision of segment types into 106 hydrophobic (nonpolar), 56 neutral (polar), and 47 charged (that is, potentially positively or negatively charged, depending on pH). It was suggested ( 8 6 ) that, for a saturated b-CN monolayer at the oil-water interface, the polar groups are, on average, mainly to be found in trains, and the hydrophobic and charged groups are to be found in loops and tails in the oil and aqueous phases, respectively. To generate sensible structural features for adsorbed b-CN from theory or computer simulation, the actual sequence of different types of segment along the chain must be represented. A Monte Carlo lattice simulation ( 4 5 ) has been used to model the adsorption of a single model b-CN molecule consisting of four segment types (nonpolar, polar, potentially positively charged, and potentially negatively charged); suitable numerical values were assigned to energies of segment adsorption, segment transfer between phases, and interaction between segments. The equilibrium segment density profile, r( z ) , was calculated as a function of the distance z from the planar Gibbs interface ( z = 0). Two particular features of the predicted profile ( z ) were found to be relatively insensitive to the chosen set of segment interaction energies (45, 54): a dense inner layer ( 1 to 2 nm

SYMPOSIUM: UNDERSTANDING DAIRY EMULSIONS

thick) centered around z = 0 containing about 70 % of the segments, and a region of much lower density extending up to about 10 nm into the aqueous phase. An important feature of the b-CN molecule is its polyelectrolyte character. Recently, a detailed calculation was carried out ( 9 9 ) of the segment density profile of a model b-CN-like polyelectrolyte adsorbing at a plane solid hydrophobic surface. The calculations were based on the self-consistent field ( SCF) theory of Cohen Stuart et al. ( 2 1 ) and Fleer et al. (83). The segments were divided into three categories (nonpolar, polar, and charged), and a lattice description of the polyelectrolyte solution was used with the effect of small mobile ions included explicitly. The segment potential was a combination of an excluded volume interaction, a nearest neighbor interaction of the Flory-Huggins x-parameter type, and an electrostatic interaction that accounted for the segment valency and the local electrical potential. The x parameters were allocated assuming a strong affinity of nonpolar segments for the surface, a net repulsion between nonpolar groups and other chain segments, and a net attraction between small ions and solvent molecules. The pKa values of the weakly charged groups were accounted for by allowing the affected segments to assume more than one internal state and then modifying the potentials accordingly (94). The SCF theory predicts that an adsorbed b-CN monolayer is produced at bulk protein concentrations below a certain threshold (which depends on ionic strength). Just above the threshold, there is multilayer condensation of self-associating protein onto the surface. Figure 1 shows the calculated SCF segment density profile r( z ) at fixed ionic strength (10 mM) and bulk volume fraction of polymer ( 5 × 10–6) for three different pH values. Each curve shows a smooth decrease in segment density on moving away from the surface. Regardless of pH, the dense inner layer has a thickness of about 1 nm, and the local segment volume fraction is about 0.01 at z = 10 nm. The surface coverage at pH 7 of about 4 equivalent monolayers increases to 5 and 6.5 equivalent monolayers at pH 6 and pH 5.5, respectively. As the pH is reduced, the additional adsorbed material appears around the middle of the profile ( z ∼ 2 nm), and, at pH 5.5, a distinct shoulder in r( z ) is apparent. Detailed analysis of the distribution of individual segment types has shown that the most hydrophilic residues—especially the amino and phosphoserine groups—reside predominantly in the outer layer regions. In particular, the end amino residue lies mainly well away from the surface, which is consistent with the concept of a long tail that is composed of the N-terminus of the molecule.

2609

Figure 1. Calculated total segment density profile r( z ) for the theoretical representation of adsorbing b-CN as a function of pH at constant ionic strength (0.01 M) and bulk protein concentration ( 5 × 10–6) . Segment volume fraction is plotted against distance z from the solid plane surface.

Some SCF calculations have been performed (3, 99) for a hypothetical variant of the model b-CN polyelectrolyte that replaces the 5 phosphoserines by polar (uncharged) residues. Figure 2 shows the effect of this change on r( z ) at pH 7, ionic strength of 10 mM, and bulk volume fraction of polymer of 2 × 10–6. The loss of the phosphoserines produces an increase in surface coverage from about 4 equivalent monolayers ( +P) to about 6.5 equivalent monolayers (–P). The shoulder in the –P curve (Figure 2a) indicates closeness to the condensation limit, which is consistent with the much lower solubility of the dephosphorylated species. The additional adsorbed material is located at distances z < 11 to 12 nm, but the full (hydrodynamic) thickness (Figure 2b) of the dephosphorylated protein layer is actually significantly lower than that for the original model b-CN polyelectrolyte. The theoretical predictions in Figure 2 indicate the importance of the phosphoserine residues in b-CN in maintaining a thick steric stabilizing monolayer while avoiding protein precipitation onto the surface in multilayers. Journal of Dairy Science Vol. 80, No. 10, 1997

2610

DICKINSON

Figure 2. Effect of dephosphorylation ( ±P ) on the calculated total segment density profile r( z ) for the theoretical representation of adsorbing b-CN at pH 7, ionic strength 0.01 M, and a bulk protein concentration 2 × 10–6. Segment volume fraction is plotted against distance z from the solid plane surface. Plots refer to inner region ( a ) and outer region ( b ) of the adsorbed layer. Note the logarithmic scale in plot ( b ) .

The comparison of these theoretical predictions with results for adsorbed b-CN that were obtained by a range of experimental techniques shows generally good agreement. The mobility of the phosphoserinerich tail region of the N-terminus in b-CN adsorbed at the oil-water interface in emulsions has been confirmed recently by 31P nuclear magnetic resonance (120). Multilayer formation at bulk concentrations above about 10–2 weight % ( 8 6 ) can be interpreted in terms of reversible precipitation of protein onto the surface (prewetting) when the bulk concentration is just below the condensation limit. Neutron reflectivity data at the air-water and oil-water interfaces have been fitted by a two-layer model (3, 4, 5, 64). At pH 5.5 to 7.0, the fitted profiles have a dense inner layer that is rich in protein ( r ∼0.95) and a more diffuse outer layer ( r ∼0.15 to 0.2) extending into the bulk phase. Application of Guinier analysis, which is independent of the model, has shown (5, 65) that the reduction of pH from 7.0 to 5.5 leads to a thickening of the layer and an increase in adsorbed amount (from 2 to 4 mg·m–2) as the condensation limit is approached (99). The layer thickness determined by Journal of Dairy Science Vol. 80, No. 10, 1997

neutron reflectivity agrees fairly well with values determined by ellipsometry ( 8 ) and X-ray scattering of latex particles coated with b-CN (100). A very dilute tail region of the b-CN adsorbed layer has been inferred from values of hydrodynamic layer thickness (10 to 15 nm) estimated from diffusion coefficients determined by dynamic light-scattering of particles coated with b-CN (12, 26, 27) and from the range of the protein-protein interaction estimated from direct force measurements for b-CN-coated (hydrophobized) mica surfaces (20). Moreover, dephosphorylation of b-CN leads to a reduction in hydrodynamic layer thickness and a marginal increase in the adsorbed amount (12). Leaver and Dalgleish have also shown ( 9 8 ) that the N-terminus tail is much more susceptible to proteolysis by trypsin than is the rest of the adsorbed molecule and that loss of the tail leads to a marked reduction in layer thickness. b-LG The major whey protein, b-LG, possesses a considerable, ordered secondary structure and a compact

SYMPOSIUM: UNDERSTANDING DAIRY EMULSIONS

tertiary structure. At neutral pH, b-LG exists as a dimer held together by noncovalent interactions. Each monomer contains two intramolecular disulfide bonds and a free hidden negative sulfhydryl group. bLactoglobulin is almost as surface-active as b-CN in terms of the final (steady-state) lowering of the interfacial tension, but the rate of tension lowering is substantially smaller because of slow structural reorganization in the freshly adsorbed layer. The retention of some native structure of b-LG in the emulsified state has been demonstrated by immunochemical analysis (118). During emulsification, the molecules that adsorb initially may not have time to unfold properly before they are surrounded by others in a congested closely packed monolayer, which further limits the scope for reorganization. At its simplest, the adsorbed globular protein layer can probably be best regarded as a pseudo two-dimensional system of densely packed deformable particles ( 3 0 ) that interact by a combination of electrostatic, hydrophobic, and hydrogen bonds. In addition, following b-LG adsorption at the oil-water interface in emulsions, partial unfolding of the monomer allows exposure of the free sulfhydryl group, which leads to slow polymerization of the protein in the adsorbed layer via the interchange between sulfhydryl and disulfide groups (53). Hence, the b-LG monolayer has some of the characteristic viscoelastic properties of a concentrated twodimensional particle gel, as indicated by its shear rheological properties (75). Neutron reflectivity experiments at the air-water interface have shown ( 4 ) that b-LG adsorbs to give a monolayer that is thinner and of lower surface coverage than that of b-CN. Values for b-LG at neutral pH (adsorbed amount ≈ 1.7 mg·m–2, thickness ≈ 3 nm) are similar to those reported ( 7 8 ) for another globular milk protein, BSA, using the same experimental technique. Another important difference between bCN and b-LG is that b-LG gives an adsorbed layer that gradually increases in surface coverage and layer thickness with time (25% over 4 to 5 h). This gradual change over time in the structure and properties of the adsorbed layer has implications for the flocculation behavior and rheology of emulsions that are stabilized with b-LG. Mixtures of Pure Milk Proteins Because the surface activities of individual milk proteins vary, the opportunity exists for competition between the proteins during adsorption at fluid interfaces. The two most abundant proteins in milk are as1-CN and b-CN; the latter is slightly more hydrophobic and surface-active than the former (73,

2611

105). Studies at the oil-water and air-water interfaces have demonstrated (1, 73) that bulk phase bCN is able readily to displace adsorbed as1-CN and, in addition, that as1-CN is able to displace b-CN, albeit to a lesser extent. Under conditions inhibiting casein aggregation (i.e., neutral pH and low calcium ion content), the exchange of these two disordered proteins between the fluid interface and the bulk solution is rapid and reversible, and the experimental data are consistent with a simple model based on equilibrium statistical mechanics (37). The system of as1-CN plus b-CN turns out to be rather a special case. In binary mixtures containing at least one globular protein (e.g., b-LG), some limited competitive adsorption does occur, but usually little free reversible exchange occurs between the adsorbed and unadsorbed protein (13, 74, 75). As a general rule, the protein that arrives at the interface first (e.g., during emulsification) is the one that predominates there afterward. Exchangeability is enhanced by macromolecular mobility and flexibility at the interface (54). In contrast, any factor (such as calcium ion binding) that inhibits molecular flexibility or increases intermolecular association reduces exchangeability (93). The more aggregated state of whole sodium caseinate, compared with the simple binary mixture of as1-CN and b-CN, leads to slower and less extensive protein exchange (79, 115). When casein is in the highly aggregated form of the casein micelles in skim milk, the b-CN component is preferentially displaced from the interface in the recombined milk after homogenization (117). This result is qualitatively the same as that reported recently ( 9 2 ) for caseinate emulsions containing added KCl, but the opposite of results with emulsions made with simple mixtures of as1-CN and b-CN ( 7 3 ) or sodium caseinate at low ionic strength and low ratios of protein to oil (79). In a skim milk emulsion, the higher concentration of as1-CN found at the surface may be due to preferential adsorption of the larger casein micelles, which apparently contain a lower proportion of b-CN (104). FORMATION OF EMULSIONS STABILIZED BY MILK PROTEINS One of the most important factors influencing emulsion properties is droplet size. To obtain good physical stability, it is usually important to make small droplets (below ∼1 mm ) by high pressure homogenization. If sufficient emulsifier is present to more than cover all of the newly created oil-water interface, theory shows (41, 123) that the two main factors influencing emulsion droplet size are the inJournal of Dairy Science Vol. 80, No. 10, 1997

2612

DICKINSON

terfacial tension and the power density (rate of energy dissipation per unit volume). Compared with small molecule surfactants, proteins are not very effective in reducing the interfacial tension during emulsification; that is, in order to produce a fine emulsion stabilized with milk protein, the oil and water premix needs to be homogenized under conditions of very high power density (intense laminar or turbulent flow). Because their steadystate interfacial tensions are all quite similar, most milk proteins that are present in excess give roughly the same droplet size for the same homogenization conditions. Under these conditions, the pure proteins b-CN and b-LG give surface coverages, respectively, of about 2.5 and 1.5 mg·m–2 (16, 23, 48); the reported values for sodium caseinate and whey protein isolate tend to be slightly higher (79, 90). Other things being equal, protein surface coverages at the hydrocarbon-water interface tend to be greater than at the more polar interface between triglyceride and water (22, 56). When the ratio of oil to protein is very high, protein surface coverage becomes substantially reduced. For instance, it is possible (28, 90) to make a stable fine emulsion with sodium caseinate at a ratio of oil to protein of >40:1; in this case, the estimated protein surface concentration is as low as G ≈ 1 mg·m–2. Under such conditions, the effective hydrodynamic thickness of the adsorbed layer (about 5 nm) has been found (28, 80) to be only about one half that determined at monolayer saturation coverage ( G ≈ 3 mg·m–2) . In emulsions prepared with skim milk, some competitive adsorption occurs between the caseins and the whey proteins during homogenization. Preferential adsorption of the casein fraction has been generally reported (11, 112, 113, 116, 121) for homogenized dairy emulsions. The aggregated state of the casein micelles complicates the preferential adsorption behavior of the individual caseins already described. Nevertheless, many of the main features remain the same. For instance, competitive adsorption of b-CN and b-LG in skim milk during the preparation of milk fat emulsions has been shown (121) to be consistent with results from model n-tetradecane emulsions prepared with a binary mixture of the same two pure proteins (48). FLOCCULATION AND CREAMING OF EMULSIONS As long as the ingredients are of good quality, the homogenization is efficient, and all microorganisms are excluded, it is straightforward to manufacture oilJournal of Dairy Science Vol. 80, No. 10, 1997

in-water emulsions stabilized by milk proteins having no detectable change in droplet-size distribution over prolonged storage. Despite their excellent coalescence stability, such emulsions may be susceptible to different types of flocculation behavior, which in turn may lead to enhanced creaming or serum separation (33). Flocculation Flocculation is the aggregation of emulsion droplets without disruption of the protective stabilizing layer at the interface. Flocculation occurs when the free energy of interaction between a pair of protein-covered droplets becomes appreciably negative at some separation (35). Whether this process is reversible depends on the strength of the interdroplet attractive forces. Some of the most important mechanisms of flocculation are listed (39). Lowering of protein net charge. A substantial reduction in surface charge density of the proteincoated droplets leads to loss of electrostatic stabilization (35, 55). This situation occurs, for example, near the isoelectric point (pH ∼5 ) or in the presence of calcium ions (70, 76). Increasing ionic strength. Addition of electrolyte screens out the double-layer repulsion and therefore reduces the electrostatic stabilization (35, 55). Hence, emulsions prepared with commercial milk protein ingredients of high salt content may be more flocculated than model systems prepared with pure proteins dissolved in low ionic strength buffer solutions. Lowering of solvent quality. Under conditions using poor quality solvent, proteins tend to form aggregates in bulk solution, and the entropic stabilizing steric repulsion of the adsorbed protein becomes an enthalpic destabilizing steric attraction (35, 55). This change occurs, for example, when ethanol is added to the dispersion medium. Protein bridging during emulsification. The sharing of protein molecules or aggregates (e.g., casein micelles) between adjacent droplets occurs when the amount of protein that is available during emulsification is insufficient to cover fully all the newly created oil-water interface. This type of irreversible bridging flocculation occurs during the homogenization of cream (108) and during the emulsification of model emulsions prepared with binary mixtures of macromolecules of very different surface activity (46, 47, 60). Macromolecular bridging after emulsification. Attractive electrostatic interaction between adsorbed protein on droplets and the polysaccharide (or protein) added after emulsion formation (possibly in

SYMPOSIUM: UNDERSTANDING DAIRY EMULSIONS

2613

the presence of surfactant) may induce flocculation by a polymer bridging mechanism (18, 43). The formation of covalent disulfide bonds between adsorbed protein molecules on adjacent droplets may lead to irreversible bridging flocculation during aging of emulsions stabilized by whey proteins (103). Depletion flocculation. The presence of nonadsorbing polysaccharide leads to a reversible depletion attraction between adjacent emulsion droplets arising from the exclusion of polymer molecules from the narrow gap between the surfaces (109). Creaming The kinetics of creaming of oil-in-water emulsions stabilized by casein is strongly influenced by depletion flocculation arising from addition of nonadsorbing hydrocolloids (14, 15, 88). The available experimental data are broadly consistent with theoretical treatments (109, 122), which suggest that the onset of detectable flocculation should occur above some (small) critical polymer concentration, cp*, and that the extent of flocculation depends on polymer concentration ( c p > cp* ) and molecular mass. Figure 3 shows kinetic data for the effect of commercial guar gum added after emulsification on the creaming of a triglyceride oil-in-water emulsion (10 volume % oil, 1 weight % sodium caseinate, pH 6.6, mean droplet diameter ( d32) = 0.54 mm). The changing thickness of the serum layer at 5°C was recorded here visually in samples with a total height of 16 cm containing various concentrations of the added polymer (88). The poorest emulsion stability appeared to occur at a hydrocolloid content of about 0.1 weight %. At substantially below this polymer concentration, the creaming stability is greater because the depletion flocculation is weaker and the movement of the droplets follows more closely that for isolated spheres (Stokes’ law). Substantially above this concentration, the stability is also greater because the flocculation is so strong that the mechanical structure of the emulsion resists the gravitational settling. Qualitatively similar results are found with other nonadsorbing hydrocolloids, such as dextran or xanthan, and in systems using small molecule surfactant instead of protein to make the emulsion (62, 66, 87). It has been demonstrated theoretically (2, 101) that depletion flocculation of large spheres may be induced by the presence of much smaller spheres and that the strength of the flocculation increases as the concentration of the smaller spheres increases. Such an effect has recently been demonstrated (44, 61) for oil-in-water emulsions stabilized by sodium caseinate in which the large spheres are protein-coated oil

Figure 3. Effect of hydrocolloid on creaming at 5°C of an emulsion stabilized with sodium caseinate (10 volume % triglyceride oil and 1 weight % protein; pH 6.6) in tubes 16 cm high. The thickness of the serum layer is plotted against storage time for various concentrations of guar gum: 0.005 weight % ( o) , 0.01 weight % ( ÿ) , 0.1 weight % ( π) , and 0.2 weight % ( ∫) .

droplets (size ∼ 0.5 mm), and the small spheres are casein submicelles (10 to 20 nm). Substantially enhanced creaming because of depletion flocculation becomes evident in concentrated emulsions stabilized by caseinate (35 volume % n-tetradecane, pH 7, 0.05 M phosphate buffer) when the total concentration of protein reaches about 3 weight %; in such a system, the caseinate surface concentration is about 3 mg·m–2, and the concentration of unadsorbed protein in the continuous phase is >1.5 weight %. The enhanced creaming, which can be detected ultrasonically (61), is preceded by an apparent lag phase, the duration of which increases as the overall protein content increases. At high caseinate contents ( 5 to 6 weight %), the strong depletion flocculation prevents rapid creaming, although a serum layer eventually develops as the flocculated emulsion gel gradually becomes reorganized and compressed by a slow process of syneresis. Futher evidence for reversible depletion flocculation of emulsion droplets by unadsorbed sodium caseinate submicelles has been obtained from optical microscopy (44, 61) and small deformation rheometry (49). Consistent with the phenomenon of depletion flocculation by caseinate submicelles is the structuring of thin films of aqueous sodium caseinate in nhexadecane, as recently described by Koczo et al. (97), who observed that the microlayering of submicelles takes place in the stratifying thin film and that, at a sudden step transition, a whole layer of submicelles leaves the structured film. Similar behavior has been found to occur in films containing Journal of Dairy Science Vol. 80, No. 10, 1997

2614

DICKINSON

surfactant micelles (110) or solid Brownian particles (silica or polymer latex) (111). Because the step transitions are strongly inhibited for films of small area, especially at low temperatures, it has been suggested ( 9 7 ) that this stratification of casein submicelles between emulsion droplets may provide an important physico-chemical mechanism for the stabilization of dairy food colloids. COMPETITIVE ADSORPTION OF MILK PROTEINS WITH SURFACTANTS Small molecule surfactants, in the form of natural polar lipids or synthetic food emulsifiers, can influence the properties of milk protein layers and dairy emulsions in a number of ways (34, 36, 40, 67). During homogenization, smaller droplets are generally produced in the presence of surfactants because of a more rapid lowering of the interfacial tension than with milk proteins alone (67). Interaction of protein with surfactant at the interface may produce a mechanically stronger or weaker adsorbed layer depending on the nature of the protein-surfactant interactions (17, 50). Most commonly, the competitive adsorption of proteins with surfactants during or after emulsification reduces the protein surface coverage at the oil-water interface (22, 23, 25, 31, 71, 79, 112). Surfactant-induced disruption of protein layers during mixing and whipping can lead to enhanced droplet flocculation and coalescence (19, 72). The enhancement of fat globule clumping by surfactantinduced disruption of interfacial protein layers is important technologically during the making of ice cream (7, 84, 85, 114). Figure 4 shows some experimental results ( 2 5 ) for the competitive adsorption of a-LA plus nonionic, water-soluble surfactant Tween 20 (polyoxyethylene sorbitan monolaurate) in model hydrocarbon oil-inwater emulsions (10 weight % n-tetradecane, 0.5 weight % protein, and 10 mM phosphate buffer, pH 7; d32 = 0.5 mm). The protein surface concentration G is plotted against the amount of Tween 20, added after homogenization, expressed as the molar ratio of surfactant to protein R. About half the a-LA is displaced from the surface of the emulsion droplets at R ≈ 3, and nearly all is displaced at R ≈ 30. Also shown in Figure 4 are data for a-LA emulsions with Tween 20 added before homogenization; the values do not differ significantly from those obtained with the surfactant added after emulsion formation. Behavior that is qualitatively similar to the behavior illustrated by Figure 4 is found also with b-LG or b-CN as the Journal of Dairy Science Vol. 80, No. 10, 1997

Figure 4. Effect of nonionic water-soluble surfactant, Tween 20, on protein surface coverage in a-LA emulsions (10 weight % ntetradecane and 0.45 weight % protein; pH 7.0). The protein surface concentration ( G) is plotted against the molar ratio ( R ) of surfactant to protein: surfactant added after homogenization ( o) or present during homogenization ( ÿ) .

protein emulsifier, but complete displacement occurs at lower R values (16, 24, 25). Similar behavior is again found with other nonionic water-soluble surfactants instead of Tween 20 or with a triglyceride oil (e.g., purified soybean oil) replacing the hydrocarbon oil. On replacement of the individual milk protein fractions by sodium caseinate or skim milk protein, competitive displacement of protein still occurs, but the relative extent of protein removal is reduced, and interfacial protein is seemingly still present at R > 100 (79, 112). This lower extent of protein displacement could perhaps be attributed to the greater degree of aggregation of the protein molecules in sodium caseinate and, especially, in skim milk. Nonionic water-soluble emulsifiers (e.g., Tweens and sucrose esters) are generally more effective (42,112) at displacing milk proteins from the oilwater interface than are nonionic oil-soluble emulsifiers (monoglycerides, and sorbitan esters). Many commercial food colloids contain both types of emulsifiers. Figure 5 shows some experimental data ( 5 6 ) for a model emulsion system containing the two kinds of surfactant (20 weight % soybean oil, 0.4 weight% bCN, 20 mM bis-tris propane buffer, pH 7; mean

SYMPOSIUM: UNDERSTANDING DAIRY EMULSIONS

Figure 5. Effect of a combination of a water-soluble surfactant, octaethylene glycol n-dodecyl ether ( C 12E8) , and an oil-soluble surfactant, glycerol monostearate (GMS), on the protein surface coverage in b-CN emulsions (20 weight % triglyceride oil and 0.4 weight % protein; pH 7.0). The GMS was dissolved in the oil prior to homogenization. The C12E8 was added to the continuous phase after homogenization. The protein surface concentration ( G) is plotted against the molar ratio ( R ) of C12E8 to protein: no GMS present ( o) , 0.05 weight % GMS ( π) , 0.2 weight % GMS ( ») , and 0.5 weight % GMS ( ♦) . (Reproduced with permission.)

droplet size d32 = 0.8 mm). The protein-stabilized emulsion was prepared ( 5 6 ) with various concentrations of pure glycerol monostearate ( GMS) dissolved in the triglyceride oil; various amounts of octaethylene glycol n-dodecyl ether ( C 12E8) were added after emulsion preparation. The protein surface coverage G is plotted against the molar ratio of C12E8 to b-CN, R, for various constant amounts of GMS in the oil phase. When present alone, the C12E8 is much more effective than the GMS in displacing the protein: complete displacement occurs at R ≈ 10 in the absence of GMS, but the coverage is still G = 0.7 mg·m–2 at a GMS concentration of 0.5 weight % (molar ratio of GMS to b-CN ≈ 90) in the absence of C12E8. However, also evident is that the amount of water-soluble surfactant required to displace the protein completely from the interface is greatly reduced by the presence of oilsoluble surfactant dissolved in the oil phase. Commercial-grade GMS (i.e., a mixture of various mono- and diglycerides) is the most widely used small molecule emulsifier in the formulation of dairy emulsions. In laboratory-prepared cream liqueurs, commercial grade GMS displaces a significant proportion of the adsorbed milk protein (mainly sodium caseinate) (69). The competitive adsorption behavior shown in Figure 5 for the emulsion system containing b-CN plus pure GMS apparently is broadly similar to that occurring in the cream liqueur, even though the food product contains a complex mixture of milk pro-

2615

teins and small molecule surfactants, as well as various other components of low molecular mass (e.g., ethanol, sucrose, and citrate). Lecithin is another food emulsifier ingredient that is commonly found in dairy emulsion formulations. Because of the wide range of complex liquid crystalline phases formed in water by phospholipids ( 9 ) , the role of lecithin-protein interactions in emulsions seems to be more complicated than that of other food surfactants. The functional properties of commercial egg or soy lecithin, which can contain a somewhat variable mixture of phospholipids, is rather sensitive to sample origin and treatment. Overall, phospholipids appear to be less effective at competitively displacing milk proteins from the oil-water interface (22, 52, 81, 82), although the behavior is sensitive to molecular factors such as type of head group (e.g., choline versus ethanolamine) and fatty acid chain (e.g., palmitoyl versus dioleyl). The effectiveness of displacement also seems to depend ( 8 9 ) on whether the competitive adsorption occurs at a macroscopic interface (low ratio of area to volume) or in an oil-inwater emulsion (high ratio of area to volume). The susceptibility of emulsion droplets to coalesce (or aggregate) in shear flow is known as orthokinetic stability. A key factor affecting the orthokinetic stability of dairy emulsions containing semisolid fat droplets is the size and distribution of the fat crystals (10). Oil droplets containing fat crystals are very susceptible to clumping (i.e., partial coalescence) when stirred, and in such dairy emulsion systems the nature of the fat crystal morphology is generally much more important than other physico-chemical factors. In oil-in-water emulsions containing fully liquid droplets, however, the main factor influencing the orthokinetic stability is the composition and structure of the adsorbed protein layer. For b-LG emulsions (10 weight % n-tetradecane, and 0.45 weight % protein; d32 = 0.35 mm ) sheared continuously under turbulent flow conditions, a small concentration of Tween 20, added after emulsification, produces a large reduction in stability (19, 72). With Tween 20 present at a molar ratio of surfactant to protein of R = 1, the rate of orthokinetic coalescence was about 10 times higher than the rate in the absence of surfactant. The likely explanation for this behavior is the loosening of the structure of the adsorbed protein-stabilizing layer in the presence of the water-soluble surfactant. Direct evidence for this disruption at the oil-water interface has come from separate measurements of surface shear viscosity of adsorbed b-LG in the presence of Tween 20 (24). But in the presence of rather large amounts ( R ≥ 10) of the oil-soluble surfactant Span 80 (sorbitan monoJournal of Dairy Science Vol. 80, No. 10, 1997

2616

DICKINSON

oleate), the orthokinetic stability of the same b-LG emulsion was unaffected. This result is consistent with the lack of a significant disruption of the adsorbed b-LG stabilizing layer by the oil-soluble surfactant, as confirmed by the insensitivity of protein surface shear viscosity to the presence of lipophilic surfactant in the oil phase (17). Another manifestation of competitive adsorption between protein and surfactant arises when the effect of small molecule emulsifiers on the rheology of heatset whey protein emulsion gels is considered (51, 58, 59, 63, 96, 102, 124). Without surfactants present, protein-coated emulsion droplets have a reinforcing effect on the heat-set whey protein gel network (91, 95). However, in experiments with Tween 20, added after homogenization but prior to heat treatment (30 min at 90°C), the shear rheological properties at low strain of concentrated b-LG emulsions (38 weight % oil) were very sensitive to both the overall protein concentration and the molar ratio of surfactant to protein (51). In particular, the storage modulus has been shown to increase at low emulsifier contents (up to R ≈ 1), to decrease sharply at intermediate emulsifier contents ( R ≈ 2), and then to increase again or remain low (depending on protein concentration) at high emulsifier contents ( R ≥4). The sharp fall in the gel strength at Tween 20 concentrations corresponding to R >1 can reasonably be attributed to partial displacement of protein from the oil-water interface by the water-soluble surfactant, producing emulsion droplets that no longer mechanically reinforce the heat-set protein gel structure. With emulsifiers such as lecithin, which do not induce substantial competitive displacement of protein from the interface, the gel strength simply increases steadily and monotonically (58, 63) as the concentration of emulsifier in the system increases. REFERENCES 1 Anand, K., and S. Damodaran. 1996. Dynamics of exchange between as1-casein and b-casein during adsorption at airwater interface. J. Agric. Food Chem. 44:1022. 2 Asakura, S., and F. Oosawa. 1954. On interaction between two bodies immersed in a solution of macromolecules. J. Chem. Phys. 22:1255. 3 Atkinson, P. J., E. Dickinson, D. S. Horne, F.A.M. Leermakers, and R. M. Richardson. 1996. Ber. Bunsen–Ges. Phys. Chem. 100:994. 4 Atkinson, P. J., E. Dickinson, D. S. Horne, and R. M. Richardson. 1995. Neutron reflectivity of adsorbed b-casein and blactoglobulin at the air/water interface. J. Chem. Soc. Faraday Trans. 91:2847. 5 Atkinson, P. J., E. Dickinson, D. S. Horne, and R. M. Richardson. 1995. Neutron reflectivity of adsorbed protein films. Am. Chem. Soc. Symp. Ser. 602:311. 6 Ball, A., and R.A.L. Jones. 1995. Conformational changes in adsorbed proteins. Langmuir 11:3542. Journal of Dairy Science Vol. 80, No. 10, 1997

7 Barfod, N. M., N. Krog, G. Larsen, and W. Buchheim. 1991. Effects of emulsifiers on protein/fat interaction in ice-cream mix during ageing. 1. Quantitative analyses. Fat Sci. Technol. 93:24. 8 Benjamins, J., J. A. de Feijter, M.J.A. Evans, D. E. Graham, and M. C. Phillips. 1975. Dynamic and static properties of proteins adsorbed at the air-water interface. Faraday Discuss. Chem. Soc. 59:218. 9 Bergensta˚hl, B. A., and P. M. Claesson. 1990. Surface forces in emulsions. Page 41 in Food Emulsions. 2nd ed. K. Larsson and S. E. Friberg, ed. Marcel Dekker, New York, NY. 10 Boode, K., and P. Walstra. 1993. Kinetics of partial coalescence in oil-in-water emulsions. Page 23 in Food Colloids and Polymers: Stability and Mechanical Properties. E. Dickinson and P. Walstra, ed. R. Soc. Chem., Cambridge, UK. 11 Britten, M., and H. J. Giroux. 1991. Emulsifying properties of whey protein and casein composite blends. J. Dairy Sci. 74: 3318. 12 Brooksbank, D. V., C. M. Davidson, D. S. Horne, and J. Leaver. 1993. Influence of electrostatic interactions on bcasein layers adsorbed on polystyrene latices. J. Chem. Soc. Faraday Trans. 89:3419. 13 Cao, Y., and S. Damodaran. 1995. Coadsorption of b-casein and bovine serum albumin at the air/water interface from a binary mixture. J. Agric. Food Chem. 43:2567. 14 Cao, Y., E. Dickinson, and D. J. Wedlock. 1990. Creaming and flocculation in emulsions containing polysaccharide. Food Hydrocolloids 4:185. 15 Cao, Y., E. Dickinson, and D. J. Wedlock. 1991. Influence of polysaccharides on the creaming of casein-stabilized emulsions. Food Hydrocolloids 5:443. 16 Chen, J., and E. Dickinson. 1993. Time-dependent competitive adsorption of milk proteins and surfactants in oil-in-water emulsions. J. Sci. Food Agric. 62:283. 17 Chen, J., and E. Dickinson. 1995. Surface shear viscosity and protein-surfactant interactions in mixed protein films adsorbed at the oil-water interface. Food Hydrocolloids 9:35. 18 Chen, J. and E. Dickinson. 1995. Protein/surfactant interfacial interactions. 1. Flocculation of emulsions containing mixed protein + surfactant. Colloids Surf. A 100:255. 19 Chen, J., E. Dickinson, and G. Iveson. 1993. Interfacial interactions, competitive adsorption and emulsion stability. Food Struct. 12:135. 20 Claesson, P. M., E. Blomberg, J. C. Fro¨berg, T. Nylander, and T. Arnebrant. 1995. Protein interactions at solid surfaces. Adv. Colloid Interface Sci. 57:161. 21 Cohen Stuart, M. A., G. J. Fleer, J. Lyklema, W. Norde, and J.M.H.M. Scheutjens, 1991. Adsorption of ions, polyelectrolytes and proteins. Adv. Colloid Interface Sci. 34:477. 22 Courthaudon, J.-L., E. Dickinson, and W. W. Christie. 1991. Competitive adsorption of lecithin and b-casein in oil-in-water emulsions. J. Agric. Food Chem. 39:1365. 23 Courthaudon, J.-L., E. Dickinson, and D. G. Dalgleish. 1991. Competitive adsorption of b-casein and nonionic surfactants in oil-in-water emulsions. J. Colloid Interface Sci. 145:390. 24 Courthaudon, J.-L., E. Dickinson, Y. Matsumura, and D. C. Clark. 1991. Competitive adsorption of b-lactoglobulin + Tween 20 at the oil-water interface. Colloids Surf. 56:293. 25 Courthaudon, J.-L., E. Dickinson, Y. Matsumura, and A. Williams. 1991. Influence of emulsifier on the competitive adsorption of whey proteins in emulsions. Food Struct. 10:109. 26 Dalgleish, D. G. 1990. The conformations of proteins on solid/ water interfaces—caseins and phosvitin on polystyrene latices. Colloids Surf. 46:141. 27 Dalgleish, D. G. 1993. The sizes and conformations of the proteins in adsorbed layers of individual caseins on latices and in oil-in-water emulsions. Colloids Surf. B 1:1. 28 Dalgleish, D. G. 1995. Structures and properties of adsorbed layers in emulsions containing milk proteins. Page 23 in Food Macromolecules and Colloids. E. Dickinson and D. Lorient, ed. R. Soc. Chem., Cambridge, United Kingdom.

SYMPOSIUM: UNDERSTANDING DAIRY EMULSIONS 29 Dalgleish, D. G., and J. Leaver. 1991. The possible conformations of milk proteins adsorbed on oil/water interfaces. J. Colloid Interface Sci. 141:288. 30 de Feijter, J. A., and J. Benjamins. 1982. Soft-particle model of compact macromolecules at interfaces. J. Colloid Interface Sci. 90:289. 31 de Feijter, J. A., J. Benjamins, and M. Tamboer. 1987. Adsorption displacement of proteins by surfactants in oil-in-water emulsions. Colloids Surf. 27:243. 32 de Wit, J. N. 1989. Functional properties of whey proteins. Page 285 in Developments in Dairy Chemistry 4. P. F. Fox, ed. Elsevier Appl. Sci., London, United Kingdom. 33 Dickinson, E. 1988. The structure and stability of emulsions. Page 41 in Food Structure—Its Creation and Evaluation. J. M. V. Blanshard and J. R. Mitchell, ed. Butterworths, London, United Kingdom. 34 Dickinson, E. 1992. Structure and composition of adsorbed protein layers and the relationship to emulsion stability. J. Chem. Soc. Faraday Trans. 88:2973. 35 Dickinson, E. 1992. Ch. 4 in An Introduction to Food Colloids. Oxford Univ. Press, Oxford, United Kingdom. 36 Dickinson, E. 1992. Adsorbed protein layers in food emulsions. Page 25 in Emulsions—A Fundamental and Practical Approach. J. Sjo¨blom, ed. Kluwer, Dordrecht, The Netherlands. 37 Dickinson, E. 1992. Adsorption of sticky hard spheres: relevance to protein competitive adsorption. J. Chem. Soc. Faraday Trans. 88:3561. 38 Dickinson, E. 1993. Proteins in solution and at interfaces. Page 295 in Interactions of Surfactants with Polymers and Proteins. E. D. Goddard and K. P. Ananthapadmanabhan, ed. CRC Press, Boca Raton, FL. 39 Dickinson, E. 1993. Emulsion stability. Page 387 in Food Hydrocolloids: Structures, Properties and Functions. K. Nishinari and E. Doi, ed. Plenum Press, New York, NY. 40 Dickinson, E. 1994. Protein-stabilized emulsions. J. Food Eng. 22:59. 41 Dickinson, 1994. Emulsions and droplet size control. Page 191 in Controlled Particle, Droplet and Bubble Formation. D. J. Wedlock, ed. Butterworth-Heinemann, Oxford, United Kingdom. 42 Dickinson, E. 1995. Recent trends in food colloids research. Page 1 in Food Macromolecules and Colloids. E. Dickinson and D. Lorient, ed. R. Soc. Chem., Cambridge, United Kingdom. 43 Dickinson, E. 1995. Mixed biopolymers at interfaces. Page 349 in Biopolymer Mixtures. S. E. Harding, S. E. Hill, and J. R. Mitchell, ed. Nottingham Univ. Press, Nottingham, United Kingdom. 44 Dickinson, E. 1997. Aggregation processes, particle interactions and colloidal structure. Page 107 in Food Colloids: Proteins, Lipids and Polysaccharides. E. Dickinson and B. Bergensta˚hl, ed. R. Soc. Chem., Cambridge, United Kingdom. 45 Dickinson, E., and S. R. Euston. 1992. Monte Carlo simulation of colloidal systems. Adv. Colloid Interface Sci. 42:89. 46 Dickinson, E., and V. B. Galazka. 1991. Bridging flocculation in emulsions made with a mixture of protein and polysaccharide. Page 494 in Food Polymers, Gels and Colloids. E. Dickinson, ed. R. Soc. of Chem., Cambridge, United Kingdom. 47 Dickinson, E., and V. B. Galazka. 1991. Bridging flocculation induced by competitive adsorption: implications for emulsion stability. J. Chem. Soc. Faraday Trans. 87:963. 48 Dickinson, E., and J.-L. Gelin. 1992. Influence of emulsifier on competitive adsorption of as-casein and b-lactoglobulin in oilin-water emulsions. Colloids Surf. 63:329. 49 Dickinson, E., and M. Golding. 1997. Depletion flocculation of emulsions containing unadsorbed sodium caseinate. Food Hydrocolloids 11:13. 50 Dickinson, E., and S.-T. Hong. 1995. Interfacial and stability properties of emulsions: influence of protein heat treatment and emulsifiers. Page 269 in Food Macromolecules and Colloids. E. Dickinson and D. Lorient, ed. R. Soc. Chem., Cambridge, United Kingdom.

2617

51 Dickinson, E., and S.-T. Hong. 1995. Influence of water-soluble nonionic emulsifier on the rheology of heat-set proteinstabilized emulsion gels. J. Agric. Food Chem. 43:2560. 52 Dickinson, E., and G. Iveson. 1993. Absorbed films of blactoglobulin + lecithin at the hydrocarbon-water and triglyceride-water interfaces. Food Hydrocolloids 6:533. 53 Dickinson, E., and Y. Matsumura. 1991. Time-dependent polymerization of b-lactoglobulin through disulphide bonds at the oil-water interface in emulsions. Int. J. Biol. Macromol. 13:26. 54 Dickinson, E., and Y. Matsumura. 1994. Proteins at liquid interfaces: role of the molten globule state. Colloids Surf. B. 3: 1. 55 Dickinson, E., and G. Stainsby. 1982. Colloids in Food. Applied Science, London, United Kingdom. 56 Dickinson, E., and S. Tanai. 1992. Protein displacement from the emulsion droplet surface by oil-soluble and water-soluble surfactants. J. Agric. Food Chem. 40:179. 57 Dickinson, E., and C. M. Woskett. 1989. Competitive adsorption between proteins and small-molecule surfactants in food emulsions. Page 74 in Food Colloids. R. D. Bee, P. Richmond, and J. Mingins, ed. R. Soc. Chem., London, United Kingdom. 58 Dickinson, E., and Y. Yamamoto. 1996. Effect of lecithin on the viscoelastic properties of b-lactoglobulin-stabilized emulsion gels. Food Hydrocolloids 10:301. 59 Dickinson, E., and Y. Yamamoto. 1996. Viscoelastic properties of heat-set whey protein-stabilized emulsion gels with added lecithin. J. Food Sci. 61:811. 60 Dickinson, E., F. O. Flint, and J. A. Hunt. 1989. Bridging flocculation in binary protein-stabilized emulsions. Food Hydrocolloids 3:389. 61 Dickinson, E., M. Golding, and M.J.W. Povey. 1997. Creaming and flocculation of oil-in-water emulsions containing sodium caseinate. J. Colloid Interface Sci. 185:515. 62 Dickinson, E., M. I. Goller, and D. J. Wedlock. 1995. Osmotic pressure, creaming and rheology of emulsions containing nonionic polysaccharide. J. Colloid Interface Sci. 172:192. 63 Dickinson, E., S.-T. Hong, and Y. Yamamoto. 1996. Rheology of heat-set emulsion gels containing b-lactoglobulin and smallmolecule surfactants. Neth. Milk Dairy J. 50:199. 64 Dickinson, E., D. S. Horne, J. S. Phipps, and R. M. Richardson. 1993. A neutron reflectivity study of the adsorption of b-casein at fluid interfaces. Langmuir 9:242. 65 Dickinson, E., D. S. Horne, and R. M. Richardson. 1993. Neutron reflectivity study of the competitive adsorption of b-casein and water-soluble surfactant at the planar air-water interface. Food Hydrocolloids 7:497. 66 Dickinson, E., J. Ma, and M.J.W. Povey. 1994. Creaming of concentrated oil-in-water emulsions containing xanthan. Food Hydrocolloids 8:481. 67 Dickinson, E., A. Mauffret, S. E. Rolfe, and C. M. Woskett. 1989. Adsorption at interfaces in dairy systems. J. Soc. Dairy Technol. 42:18. 68 Dickinson, E., B. S. Murray, and G. Stainsby. 1988. Protein adsorption at air-water and oil-water interfaces. Page 123 in Advances in Food Emulsions and Foams. E. Dickinson and G. Stainsby, ed. Elsevier Appl. Sci., London, United Kingdom. 69 Dickinson, E., S. K. Narhan, and G. Stainsby. 1989. Factors affecting the properties of cohesive creams formed from cream liqueurs. J. Sci. Food Agric. 48:225. 70 Dickinson, E., S. K. Narhan, and G. Stainsby. 1989. Stability of cream liqueurs containing low-molecular-weight surfactants. J. Food Sci. 54:77. 71 Dickinson, E., R. K. Owusu, S. Tan, and A. Williams. 1993. Oilsoluble surfactants have little effect on the competitive adsorption of a-lactalbumin and b-lactoglobulin in emulsions. J. Food Sci. 58:295. 72 Dickinson, E., R. K. Owusu, and A. Williams. 1993. Orthokinetic destabilization of a protein-stabilized emulsion by a water-soluble surfactant. J. Chem. Soc. Faraday Trans. 89:865. 73 Dickinson, E., S. E. Rolfe, and D. G. Dalgleish, 1988. Competitive adsorption of as1-casein and b-casein in oil-in-water emulsions. Food Hydrocolloids 2:397. Journal of Dairy Science Vol. 80, No. 10, 1997

2618

DICKINSON

74 Dickinson, E., S. E. Rolfe, and D. G. Dalgleish. 1990. Surface shear viscometry as a probe of protein-protein interactions in mixed protein films adsorbed at the oil-water interface. Int. J. Biol. Macromol. 12:189. 75 Dickinson, E., S. E. Rolfe, and D. G. Dalgleish. 1989. Competitive adsorption in oil-in-water emulsions containing alactalbumin and b-lactoglobulin. Food Hydrocolloids 3:193. 76 Dickinson, E., R. H. Whyman, and D. G. Dalgleish. 1987. Colloidal properties of model oil-in-water food emulsions stabilized separately by as1-casein, b-casein and k-casein. Page 40 in Food Emulsions and Foams. E. Dickinson, ed. R. Soc. Chem., London, United Kingdom. 77 Doxastakis, G. 1989. Milk proteins. Page 9 in Food Emulsifiers. G. Charalambous and G. Doxastakis, ed. Elsevier Sci. B. V., Amsterdam, The Netherlands. 78 Eaglesham, A., T. M. Herrington, and J. Penfold. 1992. A neutron reflectivity study of a spread monolayer of bovine serum albumin. Colloids Surf. 65:9. 79 Euston, S. E., H. Singh, P. A. Munro, and D. G. Dalgleish. 1995. Competitive adsorption between sodium caseinate and oil-soluble and water-soluble surfactants in oil-in-water emulsions. J. Food Sci. 60:1124. 80 Fang, Y., and D. G. Dalgleish. 1993. Dimensions of the adsorbed layers in oil-in-water emulsions stabilized by caseins. J. Colloid Interface Sci. 156:29. 81 Fang, Y., and D. G. Dalgleish. 1996. Competitive adsorption between dioleylphosphatidylcholine and sodium caseinate on oil-water interfaces. J. Agric. Food Chem. 44:59. 82 Fang, Y., and D. G. Dalgleish. 1996. Comparison of the effects of three different phosphatidylcholines on casein-stabilized oilin-water emulsions. J. Am. Oil Chem. Soc. 73:437. 83 Fleer, G. J., M. A. Cohen Stuart, J.M.H.M. Scheutjens, T. Cosgrove, and B. Vincent. 1993. Polymers at Interfaces. Chapman and Hall, London, United Kingdom. 84 Gelin, J.-L., L. Poyen, J.-L. Courthaudon, M. Le Meste, and D. Lorient. 1994. Structural changes in oil-in-water emulsions during the manufacture of ice cream. Food Hydrocolloids 8: 299. 85 Goff, H. D., and W. K. Jordan. 1989. Action of emulsifiers in promoting fat destabilization during the manufacture of icecream. J. Dairy Sci. 72:18. 86 Graham, D. E., and M. C. Phillips. 1979. Proteins at liquid interfaces. J. Colloid Interface Sci. 70:403. 87 Gunning, P. A., D. J. Hibberd, A. M. Howe, and M. M. Robins. 1988. Gravitational destabilization of emulsions flocculated by non-adsorbed xanthan. Food Hydrocolloids 2:119. 88 Heeney, L. 1994. The influence of biopolymers on emulsion stability. M. Phil. Thesis, Univ. Leeds, Leeds, United Kingdom. 89 Heertje, I., H. van Aalst, J.C.G. Blonk, A. Don, J. Nederlof, and E. H. Lucassen-Reynders. 1996. Observations on emulsifiers at the interface between oil and water by confocal scanning light microscopy. Lebensm. Wiss. Technol. 29:217. 90 Hunt, J. A., and D. G. Dalgleish. 1994. Adsorption behavior of whey protein isolate and caseinate in soya oil-in-water emulsions. Food Hydrocolloids 8:175. 91 Hunt, J. A., and D. G. Dalgleish. 1995. Heat stability of oil-inwater emulsions containing milk proteins: effect of ionic strength and pH. J. Food Sci. 60:1120. 92 Hunt, J. A., and D. G. Dalgleish. 1996. The effect of the presence of KCl on the adsorption behavior of whey protein and caseinate in oil-in-water emulsions. Food Hydrocolloids 10:159. 93 Hunt, J. A., E. Dickinson, and D. S. Horne. 1993. Competitive displacement of proteins in oil-in-water emulsions containing calcium ions. Colloids Surf. A 71:197. 94 Israe¨ls, R., F.A.M. Leermakers, G. J. Fleer, and E. B. Zhulina. 1994. Charged polymeric brushes: structure and scaling relations. Macromolecules 27:3249. Journal of Dairy Science Vol. 80, No. 10, 1997

95 Jost, R., R. Baechler, and G. Masson. 1986. Heat gelation of oil-in-water emulsions stabilized by whey protein. J. Food Sci. 51:440. 96 Jost, R., F. Dannenberg, and J. Rosset. 1989. Heat-set gels based on oil/water emulsions: an application of whey protein functionality. Food Microstruct. 8:23. 97 Koczo, K., A. D. Nikolov, D. T. Wasan, R. P. Borwankar, and A. Gonsalves. 1996. Layering of sodium caseinate sub-micelles in thin liquid films—a new stability mechanism for food dispersions. J. Colloid Interface Sci. 178:694. 98 Leaver, J., and D. G. Dalgleish. 1992. Variations in the binding of b-casein to oil-water interfaces detected by trypsin-catalysed hydrolysis. J. Colloid Interface Sci. 149:49. 99 Leermakers, F.A.M., P. J. Atkinson, E. Dickinson, and D. S. Horne. 1996. Self-consistent-field modelling of adsorbed bcasein: effects of pH and ionic strength on surface coverage and density profile. J. Colloid Interface Sci. 178:681. 100 Mackie, A. R., J. Mingins, and A. N. North. 1991. Characterization of adsorbed layers of a disordered coil protein on polystyrene latex. J. Chem. Soc. Faraday Trans. 87:3043. 101 Mao, Y., M. E. Cates, and H.N.W. Lekkerkerker. 1995. Depletion force in colloidal systems. Physica A 222:10. 102 McClements, D. J., F. J. Monahan, and J. E. Kinsella. 1993. Effect of emulsion droplets on the rheology of whey protein isolate gels. J. Texture Stud. 24:411. 103 McClements, D. J., F. J. Monahan, and J. E. Kinsella. 1993. Disulfide bond formation affects stability of whey protein isolate emulsions. J. Food Sci. 58:1036. 104 McGann, T.C.A., W. J. Donnelly, R. D. Kearney, and W. Buchheim. 1980. Composition and size distribution of bovine casein micelles. Biochim. Biophys. Acta 630:261. 105 Mitchell, J. R., L. Irons, and G. J. Palmer. 1970. A study of the spread and adsorbed films of milk proteins. Biochim. Biophys. Acta 200:138. 106 Morr, C. V. 1982. Functional properties of milk proteins and their use as food ingredients. Page 375 in Developments in Dairy Chemistry. P. F. Fox, ed. Applied Sci., London, United Kingdom. 107 Morr, C. V., and E.Y.W. Ha. 1993. Whey protein concentrates and isolates: processing and functional properties. Crit. Rev. Food Sci. Nutr. 33:431. 108 Mulder, H., and P. Walstra. 1974. The Milk Fat Globule. Pudoc, Wageningen, The Netherlands. 109 Napper, D. H. 1983. Polymeric Stabilization of Colloidal Dispersions. Acad. Press, London, United Kingdom. 110 Nikolov, A. D., and D. T. Wasan. 1989. Ordered micelle structuring in thin films formed from ionic surfactant solutions. J. Colloid Interface Sci. 133:1. 111 Nikolov, A. D., and D. T. Wasan. 1992. Dispersion stability due to structural contributions to the particle interactions as probed by thin liquid film dynamics. Langmuir 8:2985. 112 Oortwijn, H., and P. Walstra. 1979. The membranes of recombined fat globules. 2. Composition. Neth. Milk. Dairy J. 33:134. 113 Oortwijn, H., and P. Walstra. 1982. The membranes of recombined fat globules. 4. Effects on properties of the recombined milks. Neth. Milk. Dairy J. 36:279. 114 Pelan, B.M.C., K. M. Watts, I. J. Campbell, and A. Lips. 1997. On the stability of aerated milk protein emulsions in the presence of small-molecule surfactants. Page 55 in Food Colloids: Proteins, Lipids and Polysaccharides. E. Dickinson and B.Bergensta˚hl, ed. R. Soc. Chem., Cambridge, United Kingdom. 115 Robson, E. W., and D. G. Dalgleish. 1987. Interfacial composition of sodium caseinate emulsions. J. Food Sci. 52:1694. 116 Sharma, S. K., and D. G. Dalgleish. 1993. Interactions between milk serum proteins and synthetic fat globule membrane during heating of homogenized whole milk. J. Agric. Food Chem. 41:407.

SYMPOSIUM: UNDERSTANDING DAIRY EMULSIONS 117 Sharma, R., S. Singh, and M. W. Taylor. 1996. Composition and structure of fat globule surface layers in recombined milk. J. Food Sci. 61:28. 118 Shimizu, M. 1995. Structure of proteins adsorbed at an emulsified oil surface. Page 34 in Food Macromolecules and Colloids. E. Dickinson and D. Lorient, ed. R. Soc. Chem., Cambridge, United Kingdom. 119 Swaisgood, H. E. 1982. Chemistry of milk protein. Page 1 in Developments in Dairy Chemistry. P. F. Fox, ed. Appl. Sci., London, United Kingdom. 120 ter Beek, L. C., M. Ketelaars, D. C. McCain, P.E.A. Smulders, P. Walstra, and M. A. Hemminga. 1996. Nuclear magnetic resonance study of the conformation and dynamics of b-casein

2619

at the oil/water interface in emulsions. Biophys. J. 70:2396. 121 Tomas, A., D. Paquet, J.-L. Courthaudon, and D. Lorient. 1994. Effect of fat and protein contents on droplet size and protein surface coverage in dairy emulsions. J. Dairy Sci. 77: 413. 122 Vincent, B., J. Edwards, S. Emmett, and A. Jones. 1986. Depletion flocculation in dispersions of sterically stabilized particles (“soft spheres”). Colloids Surf. 18:61. 123 Walstra, P. 1993. Principles of emulsion formation. Chem. Eng. Sci. 48:333. 124 Xiong, Y. L., and J. E. Kinsella. 1991. Influence of fat globule membrane composition and fat type on the rheological properties of milk-based composite gels. Milchwissenschaft 46:207.

Journal of Dairy Science Vol. 80, No. 10, 1997