- Email: [email protected]
Stability of Emulsions Containing Both Sodium Caseinate and Tween 20
Journal of Colloid and Interface Science 212, 466 – 473 (1999) Article ID jcis.1998.6078, available online at http://www.idealibrary.com on
Stability of Emulsions Containing Both Sodium Caseinate and Tween 20 Eric Dickinson, 1 Christos Ritzoulis, and Malcolm J. W. Povey Procter Department of Food Science, University of Leeds, Leeds LS2 9JT, United Kingdom Received September 21, 1998; accepted December 30, 1998
tems containing anionic surfactants is strongly related to the presence of surfactant–protein association structures and interfacial interactions (10). The efficiency of the commercial protein ingredient sodium caseinate in stabilizing and destabilizing oil-in-water emulsions has previously been studied in our laboratory (11–13). It was found that there exists an optimum protein concentration producing good saturation surface coverage and maximum emulsion stability; smaller amounts of protein are associated with lower surface coverage leading to bridging flocculation and coalescence, while a considerable excess of unadsorbed protein causes depletion flocculation (12). The scope of the present work is to investigate the effect on stability of adding another surface-active component (a nonionic water-soluble emulsifier) to the caseinate-based emulsion. The creaming stability and rheological behavior of the emulsion are expected to change, reflecting the changes in both the interfacial and bulk phase compositions. The non-ionic surfactant chosen for investigation is commercial polyoxyethylene sorbitan monolaurate (Tween 20).
The creaming and rheology of oil-in-water emulsions (30 vol% n-tetradecane, pH 6.8) stabilized by a mixture of commercial sodium caseinate and the non-ionic emulsifier polyoxyethylene sorbitan monolaurate (Tween 20) has been investigated at 21°C. The presence of sufficient Tween 20 to displace most of the protein from the emulsion droplet surface leads to greatly enhanced emulsion creaming (and strongly non-Newtonian rheology) which is indicative of depletion flocculation by nonadsorbed surface-active material (protein and emulsifier). In emulsions containing a constant amount of surface-active material, the replacement of a very small fraction of Tween 20 by caseinate in a stable pure Tween 20 emulsion leads to enhanced creaming for a small fraction of the droplets, and this fraction increases with increasing replacement of emulsifier by protein. This behavior is probably due to depletion flocculation, although an alternative bridging mechanism is also a possibility. The overall stability of these sets of emulsions can be represented in terms of a global stability diagram containing regions of bridging flocculation and coalescence (low content of surface-active material), stability (intermediate content), and depletion flocculation (high content). © 1999 Academic Press Key Words: bridging flocculation; caseinate; creaming; depletion flocculation; emulsion rheology; stability diagram; Tween 20.
MATERIALS AND METHODS
Spray-dried sodium caseinate (,6% moisture, 0.05% calcium), was obtained from de Melkindustrie (Veghel, the Netherlands). Polyoxyethylene sorbitan monolaurate (Tween 20) (50% lauric acid, balanced by myristic, palmitic, and stearic acid), n-tetradecane (.99%), Tris[hydroxymethyl]aminomethane hydrochloride (Trizma-HCl) buffer (.99%), and sodium azide (.99%) were purchased from Sigma (St. Louis, MO). Catalyst tablets (K 2SO 4 1 CuSO 4) and concentrated sulfuric acid for the Kjeldahl nitrogen determination were purchased from BDH (Poole, England). The emulsion aqueous phase was prepared by dissolving an appropriate amount of surface-active material (caseinate and/or Tween 20) in 0.05 M Trizma-HCl aqueous solution at pH 6.8 with 0.05 wt% sodium azide added to protect against microbial contamination. n-Tetradecane was added to make a 30% oilin-water premix; this was homogenized by successively passing through a Shields high-pressure homogenizer in order to prepare a fine stable emulsion of the desired average droplet size. A Malvern Mastersizer was used to ensure that all the emulsions prepared were of a uniform average droplet diameter
INTRODUCTION
The two most important types of surface-active material in food are proteins and small-molecule surfactants or “emulsifiers.” Their separate stabilizing roles in emulsion systems are well recognized (1), but the coexistence of both proteins and emulsifiers during emulsion processing may bring about emulsion destabilization (2). The complicated behavior of these mixed systems depends on the nature of the interactions between proteins and emulsifiers at the oil–water interface. As a general rule, simple non-ionic emulsifiers displace milk proteins from the interface because at high surfactant concentrations they produce a lower interface tension (3–5). However, various cases of coexistence of proteins and surfactants at the interface have also been reported, i.e., for the case of (oilsoluble) monoglycerides (6 – 8), or (mainly water-soluble) phospholipids (9) as emulsifiers. The behavior of mixed sys1
To whom correspondence should be addressed. E-mail: e.dickinson@ leeds.ac.uk. 0021-9797/99 $30.00 Copyright © 1999 by Academic Press All rights of reproduction in any form reserved.
466
EMULSIONS CONTAINING SODIUM CASEINATE AND TWEEN 20
FIG. 1. Photograph of tubes of 30% n-tetradecane-in-water emulsions containing 2 wt% caseinate and different concentrations of Tween 20 at pH 6.8; picture taken 276 h after emulsification and storage at 21°C. From left to right: 0, 0.25, 0.50, 0.75, 1, 2, 5, and 10 wt% Tween 20. Tube height 5 75 mm.
d 32 5 0.51 6 0.02 mm, except for systems containing less than 0.4 wt% total emulsifier. (Minimum attainable average diameters for emulsions made with just 0.1, 0.2, or 0.3 wt% Tween 20 were d 32 5 1.7, 1.2, and 0.9 mm, respectively.) Additional Tween 20 or protein was then added as required and the changes in average diameter were recorded. The creaming behavior was examined by visually measuring the height of a discrete layer formed in an emulsion thermostated at 21°C at regular time intervals and also by the technique of ultrasound velocity scanning (14). The velocity of 1.2-MHz ultrasound pulses was measured at 5-mm vertical intervals along the entire length of a calibrated cell (25 cm) for emulsion samples stored for different lengths of time. Dispersed phase volume fractions were calculated from velocities using the Urick equation, with corrections for scattering effects allowed for using a renormalization technique (15). Steady-state shear viscosities were measured at 21°C as a function of shear stress using a Bohlin CVO rheometer equipped with a 24/27 cup and bob geometry. The protein load of the bulk aqueous phase was assayed through centrifugation of the emulsion, filtration of the resulting bulk water, and determination of nitrogen by the Kjeldahl method. The sample was digested by concentrated sulfuric acid; protein nitrogen was collected through distillation of the digest and titrated against sulfuric acid. RESULTS AND DISCUSSION
All the systems studied here are 30 vol% n-tetradecane-inwater emulsions. Increasing amounts of Tween 20 were added after homogenization into a set of 2% caseinate-stabilized emulsions. The set of photographs in Fig. 1 shows that the extent of creaming after 1 21 weeks increases with the amount of added Tween 20. A critical Tween 20 concentration of ;0.75 wt% appears to exist, above which extensive creaming can be visually detected within 24 h.
467
Creaming profiles of oil volume fraction versus height were obtained by ultrasound scanning at 21°C. It was noted that discernible creaming does not take place for 24 h or more at Tween 20 concentrations below 0.75 wt%. This corresponds to an emulsifier/protein molar ratio of R ' 6, taking the average molecular weight of protein as 2 3 10 4 daltons and that of the emulsifier as 1.3 3 10 3 daltons. Figure 2 shows that there is little change in the vertical oil droplet distribution after 3 days for the caseinate-stabilized emulsion. A similar system where 0.6 wt% Tween 20 was added after emulsification (Fig. 3) presents a fast separation for a small proportion of droplets; the rest of the droplets undergo gravitational creaming after 3– 4 days. Above 0.75% Tween 20, the creaming profiles are qualitatively similar: a serum is rapidly developed at the bottom of the emulsion sample within a few hours, thereafter increasing in thickness so as to occupy about half the emulsion height within 24 h, as illustrated in Fig. 4 for the system containing 0.9 wt% Tween 20. These results suggest that the onset of depletion flocculation occurs at an emulsifier concentration of around 0.75 wt%, while below this concentration there is negligible flocculation. The development of flocculation is manifest in fast creaming of the bulk of the droplet population, as mentioned previously for pure caseinate-stabilized emulsions containing an excess of unadsorbed protein (11). Displacement of casein is reasonably expected to induce equivalent flocculation in the emulsions containing a mixture of sodium caseinate and small-molecule surfactant. In order to determine the extent of competitive protein displacement by the emulsifier, emulsions were separately prepared and then centrifuged with the serum collected and the protein load assayed by the Kjeldahl method. It was found that displacement of the protein is complete at 0.7 wt% Tween 20 (R ' 5.5). We can infer, therefore, that the onset of fast creaming is closely related to the protein displacement, since at 0.7– 0.75 wt% Tween 20, all of the caseinate present can be considered to reside in the bulk aqueous phase and hence is available to promote depletion flocculation. It is known that competition between casein and the chemically similar Tween 60 initially causes conformational changes to the protein on the interface (16) (i.e., the area that the protein molecule occupies on the interface), and at higher surfactant concentrations the protein is displaced. Unadsorbed surface-active material may exist as separate Tween 20 and caseinate micelles or mixed association structures (mixed micelles). However, detailed scattering or spectroscopic studies of mixed casein–surfactant solutions would be required to confirm definitively the existence of the putative mixed micelles. Another mechanism that also cannot be completely ruled out is that, under certain conditions, there could be “phase separation” in the aqueous phase between dissolved caseinate and Tween 20 micelles (although we saw no direct visible evidence of this in mixed solutions). The relative contributions of the Tween 20 and the caseinate to the phenomenon of enhanced creaming was examined by
468
DICKINSON, RITZOULIS, AND POVEY
FIG. 2. Creaming behavior of 2 wt% caseinate emulsion (30% n-tetradecane, pH 6.8) at 21°C. The oil volume fraction is plotted against the height of the emulsion: ■, 0 h; h, 69 h.
preparing a series of emulsions containing a fixed amount of total surface-active material (protein 1 emulsifier); the relative amounts of protein and emulsifier were varied to cover the range from 100% caseinate to 100% Tween 20. Figure 5 shows photographs of emulsions stored for 1 21 weeks containing 3.5
wt% total surface-active material and 100:0, 57:47, 29:71, and 0:100 ratios of caseinate to Tween 20. The “pure” systems are more stable than the “mixed” systems. It is clear that a synergism (strictly speaking, a negative synergism) exists between the two substances in these systems.
FIG. 3. Creaming behavior for a 2 wt% caseinate, 0.6 wt% Tween 20 emulsion (30% n-tetradecane, pH 6.8) at 21°C: ■, 1 h; h, 3 h; Œ, 8.5 h; ‚, 15 h; F, 24 h; E, 33 h; }, 81 h; h, 95 h; Œ, 740 h.
EMULSIONS CONTAINING SODIUM CASEINATE AND TWEEN 20
469
FIG. 4. Creaming behavior for a 2 wt% caseinate, 0.9 wt% Tween 20 emulsion (30% n-tetradecane, pH 6.8) at 21°C: h, 1 h; ■, 3 h; ‚, 4.5 h; Œ, 8.5 h; {, 11.5 h; F, 18 h; E, 33.5 h.
Further emulsions were prepared containing a total concentration of surface-active material of 2.9 wt%, but this time with rather small proportions of caseinate. The aim was to determine the relative contribution made by the caseinate to the previously described synergistic effect. Ultrasound velocity scanning data showed that an emulsion containing 2.9 wt% Tween 20 (i.e., no protein) is stable against creaming, like that of the 2% caseinate emulsion in Fig. 2. However, Fig. 6 shows that an emulsion containing 2.85 wt% Tween 20 1 0.05 wt% caseinate (R 5 877) exhibits rapid creaming of a finite fraction of the droplets within 24 h. In Fig. 6 we see that most of the droplet population does not participate in the creaming.
FIG. 5. Creaming of emulsions containing a total 3.5% Tween 201caseinate (30% n-tetradecane, pH 6.8) at 21°C; from left to right: 3.5, 2, 1, 0% caseinate; picture taken 248 h after emulsification. Tube height 5 75 mm.
Increasing the caseinate content to 0.4 wt% (R 5 96) increases the proportion of the droplets undergoing fast creaming (see Fig. 7). At 0.6 wt% caseinate (R 5 59) all of the droplets cream fast, leading into the formation of a clear serum within 24 h, a situation similar to the one presented in Fig. 4. It is apparent that the amount of added caseinate determines the fraction of fast creaming droplets. Caseinate seems to play a direct role in inducing the aggregation of the droplets that participate in the fast-creaming flocs. Figure 8 compares the creaming profiles after 1 day’s storage for emulsions where 2, 5, 10, 14, and 21 wt% of the total surface-active material present is caseinate. Figure 9 shows the time-dependent change of the thickness of the cream layer for the same set of systems. Following the approach adopted in earlier work with pure caseinate emulsions (13), the relation of enhanced creaming to increased flocculation was investigated through constant shear rheology experiments. Previously it was shown (13) that an emulsion of roughly similar volume fraction containing just 3 wt% caseinate (no Tween) was nonflocculated and essentially Newtonian in its flow behavior. Figure 10 shows that the rheology of the 2.9 wt% Tween 20 emulsion (containing no protein) is also essentially Newtonian, indicating a nonflocculated state. Increasing the caseinate content at the expense of Tween 20 increases the non-Newtonian character of the system. At 0.6 wt% caseinate (the amount of protein required to bring about fast creaming of the entire droplet population), the emulsion is strongly shear-thinning and hence extensively flocculated. Further replacement of emulsifier by protein (to 1 wt% caseinate) seems to increase the yield stress of the flocculated
470
DICKINSON, RITZOULIS, AND POVEY
FIG. 6. Creaming behavior for a 2.85% Tween 20, 0.05% caseinate emulsion (30% n-tetradecane, pH 6.8) at 21°C: ‚, 0 h; Œ, 1.5 h; {, 3.5 h; }, 6 h; h, 8 h; ■, 23 h.
network. Hence, we have a situation in which the pure Tween 20 emulsion appears nonflocculated and stable against creaming. The addition of small amounts of caseinate causes flocculation to an extent related to the amount of added caseinate;
the fraction of the fast creaming droplets is related to the extent of the flocculation. When only a small fraction of droplets is flocculated, the non-Newtonian character is only slight. When most of the droplets are flocculated (at 0.6 wt% caseinate,
FIG. 7. Creaming behavior for a 2.5 wt% Tween 20, 0.4% caseinate emulsion (30% n-tetradecane, pH 6.8) at 21°C: h, 0 h; ■, 1.5 h; {, 3 h; }, 5 h; ‚, 6.5 h; Œ, 8.5 h; E, 10 h; F, 12 h; h, 14 h; ■, 15.5 h; {, 18 h; }, 24 h.
EMULSIONS CONTAINING SODIUM CASEINATE AND TWEEN 20
471
FIG. 8. Comparison of the creaming behavior of 30% n-tetradecane emulsions at pH 6.8 containing a fixed amount of caseinate–Tween 20 (2.9%) at 21°C; t 5 23–24 h: ‚, 0.05%; ■, 0.15%; {, 0.3%; Œ, 0.4%; E, 0.6 wt% caseinate.
maximum low-stress viscosity), there is extensive fast creaming. Further addition of caseinate cannot involve more droplets, but it can further strengthen the droplet network, as reflected in a higher effective yield stress. While depletion flocculation is the favored mechanism of destabilization in systems containing excess protein together with emulsifier in the form of micelles, the inducing of flocculation of a colloidal system by a small amount of adsorbing polymer can also reasonably be attributed to some sort of
bridging phenomenon. Hence the direct relation between the caseinate content and the extent of flocculation may suggest that the protein plays a direct bridging role in causing the association of Tween 20-covered droplets. It has been found (17–19) that this emulsifier interacts directly with the globular proteins b-lactoglobulin and bovine serum albumin. It also seems possible that caseinate could bind weakly onto an interface of adsorbed Tween 20 molecules on one droplet, leaving a dangling tail to stick onto another droplet surface, thereby effectively bridging the two droplets together. Even if the
FIG. 9. Extent of creaming in emulsions containing a total of 2.9 wt% caseinate–Tween 20 (30% n-tetradecane, pH 6.8) at 21°C. The creamed oil fraction is plotted against time of emulsion storage; caseinate concentration is: ■, 0.15%; F, 0.3%; Œ, 0.4%; , 0.6 wt%.
FIG. 10. Rheological behavior for a series of 30% n-tetradecane-in-water emulsions at pH 6.8 containing a fixed amount of caseinate–Tween 20 (2.9%) at 21°C. Apparent viscosity is plotted against shear stress: ■, 0%; E, 0.2%; Œ, 0.4%; ƒ, 0.6%; }, 1 wt% caseinate.
472
DICKINSON, RITZOULIS, AND POVEY
FIG. 11. Global stability diagram of emulsions of caseinate–Tween 20 (30% n-tetradecane, pH 6.8) at 21°C after 48 h: {, emulsions that do not present visible serum separation and exhibit Newtonian rheology; ■, emulsions presenting visually detectable creaming and non-Newtonian rheology; 3, intermediate stability cases. Three regions are discernible (I, II, and III). (I) Unstable emulsions due to emulsifier deficiency; (II) stable emulsions; (III) unstable emulsions due to emulsifier excess. Enclosed in dashed lines (A) are emulsions photographed in Fig. 1. Enclosed in circles (B) are the emulsions photographed in Fig. 5.
interfacial caseinate–surfactant binding were weak and readily reversible (just a few kT), it could still have a large effect on the creaming stability (1). This putative mechanism would be similar to the one proposed by Bergenståhl and co-workers (20, 21) for hydrocolloid bridging of emulsifier-covered droplets. In our case, it appears that 0.6 wt% caseinate is sufficient to induce this sort of aggregation into such an extent that all of the droplet fraction creams together. When all of the droplets have been linked and cream together, further addition of caseinate may simply strengthen the interdroplet binding, as can be seen from the increased low-shear stress viscosity for the 1% caseinate emulsion (Fig. 10). The overall behavior of these mixed casein 1 surfactant emulsion systems can be described using a global stability diagram of the type illustrated in Fig. 11. Concentrations of caseinate and Tween 20 are plotted along the two axes. Emulsions can be broadly classified into three categories: stable, intermediate, and unstable. A stable emulsion appears as a low-viscosity Newtonian liquid and does not present visually detectable creaming within an arbitrary time period of 48 h. An unstable emulsion is substantially viscoelastic and does present visually discernible creaming within 48 h. An emulsion that does not satisfy both the rheological and the creaming criteria (usually a viscoelastic system not displaying moderately fast creaming) is considered an “intermediate” status system. The curved lines separating the regions are deliberately drawn rather wavy in order to indicate the slight uncertainty in the
exact positions of the boundaries and their dependence on the somewhat arbitrary time-scale of observation. Studying the diagram in Fig. 11, a number of conclusions can be drawn. First, there is a fairly clear breakdown into three distinct regions: a small region I, at low concentrations of total surface-active material, is made up of unstable emulsions. Instability in area I results from lack of sufficient surface-active material to cover properly the droplet surface; it is manifested as bridging flocculation, followed by coalescence, and eventually phase separation. The large bounded area II consists of emulsions that follow the previously mentioned stability criteria. There is sufficient surface-active material to saturate the droplet interface, but not so much as to induce flocculation by the excess protein and emulsifier. The large unbounded region III corresponds to systems containing a high concentration of surface-active material. Instability in region III results from the presence of excess bulk phase emulsifier or protein. This causes the previously mentioned caseinate-induced bridging of emulsifier-covered droplets, or, more likely, depletion flocculation by a combination of caseinate submicelles, Tween micelles, and mixed caseinate–Tween micelles. For system compositions lying in region II, there is no evidence of any significant droplet coalescence during the timescale of the creaming experiments. Droplet-size distributions determined by light-scattering after dilution remain essentially unchanged over a period of 1–2 weeks. This means that, for instance, the enhanced creaming observed in the systems containing 2 wt% caseinate and $0.75 wt% Tween 20 (tubes 4 – 8 in Fig. 1) is not simply due to the creaming of large emulsion droplets produced by coalescence of smaller ones. In the creamed layer of some of the emulsions stored for 1 month, there was observed a slight shift in droplet-size distribution to larger sizes. This may be due to the formation of stronger flocs in the closely packed cream layer or due to some slight coalescence (and/or Ostwald ripening) during long-term storage of the creamed state. An examination of the change in stability for a given concentration of Tween 20 shows that the onset of depletion interactions is consistent with the destabilization mechanism. For example, increasing the caseinate concentration from 2 to 3 wt% at a constant Tween 20 concentration of 0.6 wt%, shows a transition from a stable to an unstable system; this can be attributed to excess bulk caseinate promoting depletion flocculation and associated creaming, as observed previously for pure caseinate-based emulsions (12). As expected, there are intermediate compositions in the global stability diagram that cannot be easily classed as “stable” or “unstable” according to our simple criteria. These cases are characterized by a lack of reproducibility in creaming and rheological measurements because of their incipient instability during emulsion formation or ambiguity in defining the state of flocculation. A further area of the diagram can be expected to exist at very high protein concentrations; this will consist of emulsions that exist as gels due to the very strong droplet– droplet interactions (11).
EMULSIONS CONTAINING SODIUM CASEINATE AND TWEEN 20
It is clear from these results that emulsions containing mixed protein–non-ionic surfactant species can have nontrivial stability–rheology behavior. There are various levels of complexity of models that could be used to interpret the results. At the simplest level, we can consider that the instability away from the corner of the stability diagram arises from depletion flocculation due to the excess concentration of Tween and casein association structures in the aqueous phase. This simple model fits most of the creaming–rheology data, but it does not really explain why a very low concentration of caseinate stimulates flocculation in an emulsion dominated by Tween 20. At the next level of approximation, we might infer that flocculation in mixed caseinate–surfactant systems might also arise as a result of bridging-type effects arising from small changes in interfacial composition and structure. Further refinements to the model might allow for the role of mixed caseinate–Tween micelles, phase separation of different kinds of micelles, or combinations of bridging and depletion effects within the same emulsion. Additional experimental information by other physicochemical techniques will be required to definitively distinguish between these different requirements. In the absence of this information, we put greatest emphasis on depletion flocculation by combined caseinate and Tween 20 micelles as the main mechanism of enhanced creaming destabilization in these systems. CONCLUSIONS
The behavior of oil-in-water emulsions containing both Tween 20 and sodium caseinate is complicated. Combination of the two surface-active materials can in some cases lead to a marked loss of stability. This is manifested in the form of enhanced flocculation and fast creaming within 1–2 days. When sufficient surface-active material is present to saturate the droplet surface, this can be attributed to depletion flocculation caused by excess bulk caseinate and Tween 20, or, possibly in some cases, by caseinate-mediated bridging of Tween 20-covered droplets. Without further information on the composition and structure of self-assembled mixtures of caseins–surfactants, and on the structure and interactions of the mixed interfacial layers, we cannot yet identify definitively the destabilization mechanism at all points in the mixed emulsifier composition space. Replacement of very small amounts of Tween 20 by caseinate in stable Tween 20 emulsions can affect their creaming stability to an extent related to the amount of added caseinate. Rheological measurements have confirmed a gradual transition from Newtonian to strongly viscoelastic systems as the relative concentration of caseinate increases. Our extensive data on the stability of emulsions containing both caseinate and Tween 20 can be summarized in a global
473
stability diagram. This is a graph whose axes correspond to Tween 20 and caseinate concentration scales and whose points represent all the possible compositions of the fine 30% ntetradecane-in-water emulsions at 21°C. Areas of stable and unstable emulsions are clearly defined. Instability at high concentrations of surface-active material can be most readily understood in terms of depletion flocculation by mixed caseinate– Tween 20 micelles, although we have no direct evidence for their existence at the present time. ACKNOWLEDGMENT CR thanks the Commission of the European Communities for the provision of a Marie Curie research training grant (FAIR-CT98-5013).
REFERENCES 1. Dickinson, E., “An Introduction to Food Colloids,” Oxford University Press, Oxford, 1992. 2. Goff, H. D., and Jordan, W. K., J. Dairy Sci. 72, 18 (1989). 3. Courthaudon, J.-L., Dickinson, E., and Christie, W. W., J. Agric. Food Chem. 39, 1365 (1991). 4. Stevenson, E. M., Horne, D. S., and Leaver, J., Food Hydrocolloids 11, 3 (1997). 5. Chen, J., and Dickinson, E., J. Agric. Food Chem. 46, 91 (1998). 6. Doxastakis, G., and Sherman, P., Colloid Polym. Sci. 264, 254 (1986). 7. Euston, S., Singh, H., Munro, P. A., and Dalgleish, D. G., J. Food Sci. 61, 916 (1996). 8. Dickinson, E., and Tanai, S., J. Agric. Food Chem. 40, 179 (1992). 9. Fang, Y., and Dalgleish, D. G., Colloids Surf. B: Biointerfaces 1, 357 (1993). 10. Dickinson, E., and Hong, S.-T., Colloids Surf. A: Physicochem. Eng. Aspects 127, 1 (1997). 11. Dickinson, E., Golding, M., and Povey, M. J. W., J. Colloid Interface Sci. 185, 515 (1997). 12. Dickinson, E., and Golding, M., Food Hydrocolloids 11, 13 (1997). 13. Dickinson, E., and Golding, M., J. Colloid Interface Sci. 191, 166 (1997). 14. Povey, M. J. W., “Ultrasonic Techniques for Emulsion Characterization,” Academic Press, San Diego, 1997. 15. Pinfield, V. J., Dickinson, E., and Povey, M. J. W., Ultrasonics 33, 243 (1995). 16. Dalgleish, D. G., Srinivasan, M., and Singh, H., J. Agric. Food Chem. 43, 2351 (1995). 17. Coke, M., Wilde, P. J., Russel, E. J., and Clark, D. C., J. Colloid Interface Sci. 138, 489 (1990). 18. Clark, D. C., Wilde, P. J., Bergink-Martens, D. J. M., Kokelaar, A. J. J., and Prins, A., in “Food Colloids and Polymers: Stability and Mechanical Properties” (E. Dickinson and P. Walstra, Eds.), p. 354. Royal Society of Chemistry, Cambridge, UK, 1993. 19. Rodriguez Patino, J. M., Rodriguez Nino, R., and Alvarez Gomez, J. M., Food Hydrocolloids 11, 49 (1997). 20. Bergenståhl, B., in “Gums and Stabilisers in the Food Industry” (G. O. Phillips, D. J. Wedlock, and P. A. Williams, Eds.), Vol. 4, p. 363. IRL Press, Oxford, 1988. 21. Bergenståhl, B., Faeldt, P., and Malmsten, M., in “Food Macromolecules and Colloids” (E. Dickinson and D. Lorient, Eds.), p. 201. Royal Society of Chemistry, Cambridge, UK, 1995.
Stability of Emulsions Containing Both Sodium Caseinate and Tween 20 Eric Dickinson, 1 Christos Ritzoulis, and Malcolm J. W. Povey Procter Department of Food Science, University of Leeds, Leeds LS2 9JT, United Kingdom Received September 21, 1998; accepted December 30, 1998
tems containing anionic surfactants is strongly related to the presence of surfactant–protein association structures and interfacial interactions (10). The efficiency of the commercial protein ingredient sodium caseinate in stabilizing and destabilizing oil-in-water emulsions has previously been studied in our laboratory (11–13). It was found that there exists an optimum protein concentration producing good saturation surface coverage and maximum emulsion stability; smaller amounts of protein are associated with lower surface coverage leading to bridging flocculation and coalescence, while a considerable excess of unadsorbed protein causes depletion flocculation (12). The scope of the present work is to investigate the effect on stability of adding another surface-active component (a nonionic water-soluble emulsifier) to the caseinate-based emulsion. The creaming stability and rheological behavior of the emulsion are expected to change, reflecting the changes in both the interfacial and bulk phase compositions. The non-ionic surfactant chosen for investigation is commercial polyoxyethylene sorbitan monolaurate (Tween 20).
The creaming and rheology of oil-in-water emulsions (30 vol% n-tetradecane, pH 6.8) stabilized by a mixture of commercial sodium caseinate and the non-ionic emulsifier polyoxyethylene sorbitan monolaurate (Tween 20) has been investigated at 21°C. The presence of sufficient Tween 20 to displace most of the protein from the emulsion droplet surface leads to greatly enhanced emulsion creaming (and strongly non-Newtonian rheology) which is indicative of depletion flocculation by nonadsorbed surface-active material (protein and emulsifier). In emulsions containing a constant amount of surface-active material, the replacement of a very small fraction of Tween 20 by caseinate in a stable pure Tween 20 emulsion leads to enhanced creaming for a small fraction of the droplets, and this fraction increases with increasing replacement of emulsifier by protein. This behavior is probably due to depletion flocculation, although an alternative bridging mechanism is also a possibility. The overall stability of these sets of emulsions can be represented in terms of a global stability diagram containing regions of bridging flocculation and coalescence (low content of surface-active material), stability (intermediate content), and depletion flocculation (high content). © 1999 Academic Press Key Words: bridging flocculation; caseinate; creaming; depletion flocculation; emulsion rheology; stability diagram; Tween 20.
MATERIALS AND METHODS
Spray-dried sodium caseinate (,6% moisture, 0.05% calcium), was obtained from de Melkindustrie (Veghel, the Netherlands). Polyoxyethylene sorbitan monolaurate (Tween 20) (50% lauric acid, balanced by myristic, palmitic, and stearic acid), n-tetradecane (.99%), Tris[hydroxymethyl]aminomethane hydrochloride (Trizma-HCl) buffer (.99%), and sodium azide (.99%) were purchased from Sigma (St. Louis, MO). Catalyst tablets (K 2SO 4 1 CuSO 4) and concentrated sulfuric acid for the Kjeldahl nitrogen determination were purchased from BDH (Poole, England). The emulsion aqueous phase was prepared by dissolving an appropriate amount of surface-active material (caseinate and/or Tween 20) in 0.05 M Trizma-HCl aqueous solution at pH 6.8 with 0.05 wt% sodium azide added to protect against microbial contamination. n-Tetradecane was added to make a 30% oilin-water premix; this was homogenized by successively passing through a Shields high-pressure homogenizer in order to prepare a fine stable emulsion of the desired average droplet size. A Malvern Mastersizer was used to ensure that all the emulsions prepared were of a uniform average droplet diameter
INTRODUCTION
The two most important types of surface-active material in food are proteins and small-molecule surfactants or “emulsifiers.” Their separate stabilizing roles in emulsion systems are well recognized (1), but the coexistence of both proteins and emulsifiers during emulsion processing may bring about emulsion destabilization (2). The complicated behavior of these mixed systems depends on the nature of the interactions between proteins and emulsifiers at the oil–water interface. As a general rule, simple non-ionic emulsifiers displace milk proteins from the interface because at high surfactant concentrations they produce a lower interface tension (3–5). However, various cases of coexistence of proteins and surfactants at the interface have also been reported, i.e., for the case of (oilsoluble) monoglycerides (6 – 8), or (mainly water-soluble) phospholipids (9) as emulsifiers. The behavior of mixed sys1
To whom correspondence should be addressed. E-mail: e.dickinson@ leeds.ac.uk. 0021-9797/99 $30.00 Copyright © 1999 by Academic Press All rights of reproduction in any form reserved.
466
EMULSIONS CONTAINING SODIUM CASEINATE AND TWEEN 20
FIG. 1. Photograph of tubes of 30% n-tetradecane-in-water emulsions containing 2 wt% caseinate and different concentrations of Tween 20 at pH 6.8; picture taken 276 h after emulsification and storage at 21°C. From left to right: 0, 0.25, 0.50, 0.75, 1, 2, 5, and 10 wt% Tween 20. Tube height 5 75 mm.
d 32 5 0.51 6 0.02 mm, except for systems containing less than 0.4 wt% total emulsifier. (Minimum attainable average diameters for emulsions made with just 0.1, 0.2, or 0.3 wt% Tween 20 were d 32 5 1.7, 1.2, and 0.9 mm, respectively.) Additional Tween 20 or protein was then added as required and the changes in average diameter were recorded. The creaming behavior was examined by visually measuring the height of a discrete layer formed in an emulsion thermostated at 21°C at regular time intervals and also by the technique of ultrasound velocity scanning (14). The velocity of 1.2-MHz ultrasound pulses was measured at 5-mm vertical intervals along the entire length of a calibrated cell (25 cm) for emulsion samples stored for different lengths of time. Dispersed phase volume fractions were calculated from velocities using the Urick equation, with corrections for scattering effects allowed for using a renormalization technique (15). Steady-state shear viscosities were measured at 21°C as a function of shear stress using a Bohlin CVO rheometer equipped with a 24/27 cup and bob geometry. The protein load of the bulk aqueous phase was assayed through centrifugation of the emulsion, filtration of the resulting bulk water, and determination of nitrogen by the Kjeldahl method. The sample was digested by concentrated sulfuric acid; protein nitrogen was collected through distillation of the digest and titrated against sulfuric acid. RESULTS AND DISCUSSION
All the systems studied here are 30 vol% n-tetradecane-inwater emulsions. Increasing amounts of Tween 20 were added after homogenization into a set of 2% caseinate-stabilized emulsions. The set of photographs in Fig. 1 shows that the extent of creaming after 1 21 weeks increases with the amount of added Tween 20. A critical Tween 20 concentration of ;0.75 wt% appears to exist, above which extensive creaming can be visually detected within 24 h.
467
Creaming profiles of oil volume fraction versus height were obtained by ultrasound scanning at 21°C. It was noted that discernible creaming does not take place for 24 h or more at Tween 20 concentrations below 0.75 wt%. This corresponds to an emulsifier/protein molar ratio of R ' 6, taking the average molecular weight of protein as 2 3 10 4 daltons and that of the emulsifier as 1.3 3 10 3 daltons. Figure 2 shows that there is little change in the vertical oil droplet distribution after 3 days for the caseinate-stabilized emulsion. A similar system where 0.6 wt% Tween 20 was added after emulsification (Fig. 3) presents a fast separation for a small proportion of droplets; the rest of the droplets undergo gravitational creaming after 3– 4 days. Above 0.75% Tween 20, the creaming profiles are qualitatively similar: a serum is rapidly developed at the bottom of the emulsion sample within a few hours, thereafter increasing in thickness so as to occupy about half the emulsion height within 24 h, as illustrated in Fig. 4 for the system containing 0.9 wt% Tween 20. These results suggest that the onset of depletion flocculation occurs at an emulsifier concentration of around 0.75 wt%, while below this concentration there is negligible flocculation. The development of flocculation is manifest in fast creaming of the bulk of the droplet population, as mentioned previously for pure caseinate-stabilized emulsions containing an excess of unadsorbed protein (11). Displacement of casein is reasonably expected to induce equivalent flocculation in the emulsions containing a mixture of sodium caseinate and small-molecule surfactant. In order to determine the extent of competitive protein displacement by the emulsifier, emulsions were separately prepared and then centrifuged with the serum collected and the protein load assayed by the Kjeldahl method. It was found that displacement of the protein is complete at 0.7 wt% Tween 20 (R ' 5.5). We can infer, therefore, that the onset of fast creaming is closely related to the protein displacement, since at 0.7– 0.75 wt% Tween 20, all of the caseinate present can be considered to reside in the bulk aqueous phase and hence is available to promote depletion flocculation. It is known that competition between casein and the chemically similar Tween 60 initially causes conformational changes to the protein on the interface (16) (i.e., the area that the protein molecule occupies on the interface), and at higher surfactant concentrations the protein is displaced. Unadsorbed surface-active material may exist as separate Tween 20 and caseinate micelles or mixed association structures (mixed micelles). However, detailed scattering or spectroscopic studies of mixed casein–surfactant solutions would be required to confirm definitively the existence of the putative mixed micelles. Another mechanism that also cannot be completely ruled out is that, under certain conditions, there could be “phase separation” in the aqueous phase between dissolved caseinate and Tween 20 micelles (although we saw no direct visible evidence of this in mixed solutions). The relative contributions of the Tween 20 and the caseinate to the phenomenon of enhanced creaming was examined by
468
DICKINSON, RITZOULIS, AND POVEY
FIG. 2. Creaming behavior of 2 wt% caseinate emulsion (30% n-tetradecane, pH 6.8) at 21°C. The oil volume fraction is plotted against the height of the emulsion: ■, 0 h; h, 69 h.
preparing a series of emulsions containing a fixed amount of total surface-active material (protein 1 emulsifier); the relative amounts of protein and emulsifier were varied to cover the range from 100% caseinate to 100% Tween 20. Figure 5 shows photographs of emulsions stored for 1 21 weeks containing 3.5
wt% total surface-active material and 100:0, 57:47, 29:71, and 0:100 ratios of caseinate to Tween 20. The “pure” systems are more stable than the “mixed” systems. It is clear that a synergism (strictly speaking, a negative synergism) exists between the two substances in these systems.
FIG. 3. Creaming behavior for a 2 wt% caseinate, 0.6 wt% Tween 20 emulsion (30% n-tetradecane, pH 6.8) at 21°C: ■, 1 h; h, 3 h; Œ, 8.5 h; ‚, 15 h; F, 24 h; E, 33 h; }, 81 h; h, 95 h; Œ, 740 h.
EMULSIONS CONTAINING SODIUM CASEINATE AND TWEEN 20
469
FIG. 4. Creaming behavior for a 2 wt% caseinate, 0.9 wt% Tween 20 emulsion (30% n-tetradecane, pH 6.8) at 21°C: h, 1 h; ■, 3 h; ‚, 4.5 h; Œ, 8.5 h; {, 11.5 h; F, 18 h; E, 33.5 h.
Further emulsions were prepared containing a total concentration of surface-active material of 2.9 wt%, but this time with rather small proportions of caseinate. The aim was to determine the relative contribution made by the caseinate to the previously described synergistic effect. Ultrasound velocity scanning data showed that an emulsion containing 2.9 wt% Tween 20 (i.e., no protein) is stable against creaming, like that of the 2% caseinate emulsion in Fig. 2. However, Fig. 6 shows that an emulsion containing 2.85 wt% Tween 20 1 0.05 wt% caseinate (R 5 877) exhibits rapid creaming of a finite fraction of the droplets within 24 h. In Fig. 6 we see that most of the droplet population does not participate in the creaming.
FIG. 5. Creaming of emulsions containing a total 3.5% Tween 201caseinate (30% n-tetradecane, pH 6.8) at 21°C; from left to right: 3.5, 2, 1, 0% caseinate; picture taken 248 h after emulsification. Tube height 5 75 mm.
Increasing the caseinate content to 0.4 wt% (R 5 96) increases the proportion of the droplets undergoing fast creaming (see Fig. 7). At 0.6 wt% caseinate (R 5 59) all of the droplets cream fast, leading into the formation of a clear serum within 24 h, a situation similar to the one presented in Fig. 4. It is apparent that the amount of added caseinate determines the fraction of fast creaming droplets. Caseinate seems to play a direct role in inducing the aggregation of the droplets that participate in the fast-creaming flocs. Figure 8 compares the creaming profiles after 1 day’s storage for emulsions where 2, 5, 10, 14, and 21 wt% of the total surface-active material present is caseinate. Figure 9 shows the time-dependent change of the thickness of the cream layer for the same set of systems. Following the approach adopted in earlier work with pure caseinate emulsions (13), the relation of enhanced creaming to increased flocculation was investigated through constant shear rheology experiments. Previously it was shown (13) that an emulsion of roughly similar volume fraction containing just 3 wt% caseinate (no Tween) was nonflocculated and essentially Newtonian in its flow behavior. Figure 10 shows that the rheology of the 2.9 wt% Tween 20 emulsion (containing no protein) is also essentially Newtonian, indicating a nonflocculated state. Increasing the caseinate content at the expense of Tween 20 increases the non-Newtonian character of the system. At 0.6 wt% caseinate (the amount of protein required to bring about fast creaming of the entire droplet population), the emulsion is strongly shear-thinning and hence extensively flocculated. Further replacement of emulsifier by protein (to 1 wt% caseinate) seems to increase the yield stress of the flocculated
470
DICKINSON, RITZOULIS, AND POVEY
FIG. 6. Creaming behavior for a 2.85% Tween 20, 0.05% caseinate emulsion (30% n-tetradecane, pH 6.8) at 21°C: ‚, 0 h; Œ, 1.5 h; {, 3.5 h; }, 6 h; h, 8 h; ■, 23 h.
network. Hence, we have a situation in which the pure Tween 20 emulsion appears nonflocculated and stable against creaming. The addition of small amounts of caseinate causes flocculation to an extent related to the amount of added caseinate;
the fraction of the fast creaming droplets is related to the extent of the flocculation. When only a small fraction of droplets is flocculated, the non-Newtonian character is only slight. When most of the droplets are flocculated (at 0.6 wt% caseinate,
FIG. 7. Creaming behavior for a 2.5 wt% Tween 20, 0.4% caseinate emulsion (30% n-tetradecane, pH 6.8) at 21°C: h, 0 h; ■, 1.5 h; {, 3 h; }, 5 h; ‚, 6.5 h; Œ, 8.5 h; E, 10 h; F, 12 h; h, 14 h; ■, 15.5 h; {, 18 h; }, 24 h.
EMULSIONS CONTAINING SODIUM CASEINATE AND TWEEN 20
471
FIG. 8. Comparison of the creaming behavior of 30% n-tetradecane emulsions at pH 6.8 containing a fixed amount of caseinate–Tween 20 (2.9%) at 21°C; t 5 23–24 h: ‚, 0.05%; ■, 0.15%; {, 0.3%; Œ, 0.4%; E, 0.6 wt% caseinate.
maximum low-stress viscosity), there is extensive fast creaming. Further addition of caseinate cannot involve more droplets, but it can further strengthen the droplet network, as reflected in a higher effective yield stress. While depletion flocculation is the favored mechanism of destabilization in systems containing excess protein together with emulsifier in the form of micelles, the inducing of flocculation of a colloidal system by a small amount of adsorbing polymer can also reasonably be attributed to some sort of
bridging phenomenon. Hence the direct relation between the caseinate content and the extent of flocculation may suggest that the protein plays a direct bridging role in causing the association of Tween 20-covered droplets. It has been found (17–19) that this emulsifier interacts directly with the globular proteins b-lactoglobulin and bovine serum albumin. It also seems possible that caseinate could bind weakly onto an interface of adsorbed Tween 20 molecules on one droplet, leaving a dangling tail to stick onto another droplet surface, thereby effectively bridging the two droplets together. Even if the
FIG. 9. Extent of creaming in emulsions containing a total of 2.9 wt% caseinate–Tween 20 (30% n-tetradecane, pH 6.8) at 21°C. The creamed oil fraction is plotted against time of emulsion storage; caseinate concentration is: ■, 0.15%; F, 0.3%; Œ, 0.4%; , 0.6 wt%.
FIG. 10. Rheological behavior for a series of 30% n-tetradecane-in-water emulsions at pH 6.8 containing a fixed amount of caseinate–Tween 20 (2.9%) at 21°C. Apparent viscosity is plotted against shear stress: ■, 0%; E, 0.2%; Œ, 0.4%; ƒ, 0.6%; }, 1 wt% caseinate.
472
DICKINSON, RITZOULIS, AND POVEY
FIG. 11. Global stability diagram of emulsions of caseinate–Tween 20 (30% n-tetradecane, pH 6.8) at 21°C after 48 h: {, emulsions that do not present visible serum separation and exhibit Newtonian rheology; ■, emulsions presenting visually detectable creaming and non-Newtonian rheology; 3, intermediate stability cases. Three regions are discernible (I, II, and III). (I) Unstable emulsions due to emulsifier deficiency; (II) stable emulsions; (III) unstable emulsions due to emulsifier excess. Enclosed in dashed lines (A) are emulsions photographed in Fig. 1. Enclosed in circles (B) are the emulsions photographed in Fig. 5.
interfacial caseinate–surfactant binding were weak and readily reversible (just a few kT), it could still have a large effect on the creaming stability (1). This putative mechanism would be similar to the one proposed by Bergenståhl and co-workers (20, 21) for hydrocolloid bridging of emulsifier-covered droplets. In our case, it appears that 0.6 wt% caseinate is sufficient to induce this sort of aggregation into such an extent that all of the droplet fraction creams together. When all of the droplets have been linked and cream together, further addition of caseinate may simply strengthen the interdroplet binding, as can be seen from the increased low-shear stress viscosity for the 1% caseinate emulsion (Fig. 10). The overall behavior of these mixed casein 1 surfactant emulsion systems can be described using a global stability diagram of the type illustrated in Fig. 11. Concentrations of caseinate and Tween 20 are plotted along the two axes. Emulsions can be broadly classified into three categories: stable, intermediate, and unstable. A stable emulsion appears as a low-viscosity Newtonian liquid and does not present visually detectable creaming within an arbitrary time period of 48 h. An unstable emulsion is substantially viscoelastic and does present visually discernible creaming within 48 h. An emulsion that does not satisfy both the rheological and the creaming criteria (usually a viscoelastic system not displaying moderately fast creaming) is considered an “intermediate” status system. The curved lines separating the regions are deliberately drawn rather wavy in order to indicate the slight uncertainty in the
exact positions of the boundaries and their dependence on the somewhat arbitrary time-scale of observation. Studying the diagram in Fig. 11, a number of conclusions can be drawn. First, there is a fairly clear breakdown into three distinct regions: a small region I, at low concentrations of total surface-active material, is made up of unstable emulsions. Instability in area I results from lack of sufficient surface-active material to cover properly the droplet surface; it is manifested as bridging flocculation, followed by coalescence, and eventually phase separation. The large bounded area II consists of emulsions that follow the previously mentioned stability criteria. There is sufficient surface-active material to saturate the droplet interface, but not so much as to induce flocculation by the excess protein and emulsifier. The large unbounded region III corresponds to systems containing a high concentration of surface-active material. Instability in region III results from the presence of excess bulk phase emulsifier or protein. This causes the previously mentioned caseinate-induced bridging of emulsifier-covered droplets, or, more likely, depletion flocculation by a combination of caseinate submicelles, Tween micelles, and mixed caseinate–Tween micelles. For system compositions lying in region II, there is no evidence of any significant droplet coalescence during the timescale of the creaming experiments. Droplet-size distributions determined by light-scattering after dilution remain essentially unchanged over a period of 1–2 weeks. This means that, for instance, the enhanced creaming observed in the systems containing 2 wt% caseinate and $0.75 wt% Tween 20 (tubes 4 – 8 in Fig. 1) is not simply due to the creaming of large emulsion droplets produced by coalescence of smaller ones. In the creamed layer of some of the emulsions stored for 1 month, there was observed a slight shift in droplet-size distribution to larger sizes. This may be due to the formation of stronger flocs in the closely packed cream layer or due to some slight coalescence (and/or Ostwald ripening) during long-term storage of the creamed state. An examination of the change in stability for a given concentration of Tween 20 shows that the onset of depletion interactions is consistent with the destabilization mechanism. For example, increasing the caseinate concentration from 2 to 3 wt% at a constant Tween 20 concentration of 0.6 wt%, shows a transition from a stable to an unstable system; this can be attributed to excess bulk caseinate promoting depletion flocculation and associated creaming, as observed previously for pure caseinate-based emulsions (12). As expected, there are intermediate compositions in the global stability diagram that cannot be easily classed as “stable” or “unstable” according to our simple criteria. These cases are characterized by a lack of reproducibility in creaming and rheological measurements because of their incipient instability during emulsion formation or ambiguity in defining the state of flocculation. A further area of the diagram can be expected to exist at very high protein concentrations; this will consist of emulsions that exist as gels due to the very strong droplet– droplet interactions (11).
EMULSIONS CONTAINING SODIUM CASEINATE AND TWEEN 20
It is clear from these results that emulsions containing mixed protein–non-ionic surfactant species can have nontrivial stability–rheology behavior. There are various levels of complexity of models that could be used to interpret the results. At the simplest level, we can consider that the instability away from the corner of the stability diagram arises from depletion flocculation due to the excess concentration of Tween and casein association structures in the aqueous phase. This simple model fits most of the creaming–rheology data, but it does not really explain why a very low concentration of caseinate stimulates flocculation in an emulsion dominated by Tween 20. At the next level of approximation, we might infer that flocculation in mixed caseinate–surfactant systems might also arise as a result of bridging-type effects arising from small changes in interfacial composition and structure. Further refinements to the model might allow for the role of mixed caseinate–Tween micelles, phase separation of different kinds of micelles, or combinations of bridging and depletion effects within the same emulsion. Additional experimental information by other physicochemical techniques will be required to definitively distinguish between these different requirements. In the absence of this information, we put greatest emphasis on depletion flocculation by combined caseinate and Tween 20 micelles as the main mechanism of enhanced creaming destabilization in these systems. CONCLUSIONS
The behavior of oil-in-water emulsions containing both Tween 20 and sodium caseinate is complicated. Combination of the two surface-active materials can in some cases lead to a marked loss of stability. This is manifested in the form of enhanced flocculation and fast creaming within 1–2 days. When sufficient surface-active material is present to saturate the droplet surface, this can be attributed to depletion flocculation caused by excess bulk caseinate and Tween 20, or, possibly in some cases, by caseinate-mediated bridging of Tween 20-covered droplets. Without further information on the composition and structure of self-assembled mixtures of caseins–surfactants, and on the structure and interactions of the mixed interfacial layers, we cannot yet identify definitively the destabilization mechanism at all points in the mixed emulsifier composition space. Replacement of very small amounts of Tween 20 by caseinate in stable Tween 20 emulsions can affect their creaming stability to an extent related to the amount of added caseinate. Rheological measurements have confirmed a gradual transition from Newtonian to strongly viscoelastic systems as the relative concentration of caseinate increases. Our extensive data on the stability of emulsions containing both caseinate and Tween 20 can be summarized in a global
473
stability diagram. This is a graph whose axes correspond to Tween 20 and caseinate concentration scales and whose points represent all the possible compositions of the fine 30% ntetradecane-in-water emulsions at 21°C. Areas of stable and unstable emulsions are clearly defined. Instability at high concentrations of surface-active material can be most readily understood in terms of depletion flocculation by mixed caseinate– Tween 20 micelles, although we have no direct evidence for their existence at the present time. ACKNOWLEDGMENT CR thanks the Commission of the European Communities for the provision of a Marie Curie research training grant (FAIR-CT98-5013).
REFERENCES 1. Dickinson, E., “An Introduction to Food Colloids,” Oxford University Press, Oxford, 1992. 2. Goff, H. D., and Jordan, W. K., J. Dairy Sci. 72, 18 (1989). 3. Courthaudon, J.-L., Dickinson, E., and Christie, W. W., J. Agric. Food Chem. 39, 1365 (1991). 4. Stevenson, E. M., Horne, D. S., and Leaver, J., Food Hydrocolloids 11, 3 (1997). 5. Chen, J., and Dickinson, E., J. Agric. Food Chem. 46, 91 (1998). 6. Doxastakis, G., and Sherman, P., Colloid Polym. Sci. 264, 254 (1986). 7. Euston, S., Singh, H., Munro, P. A., and Dalgleish, D. G., J. Food Sci. 61, 916 (1996). 8. Dickinson, E., and Tanai, S., J. Agric. Food Chem. 40, 179 (1992). 9. Fang, Y., and Dalgleish, D. G., Colloids Surf. B: Biointerfaces 1, 357 (1993). 10. Dickinson, E., and Hong, S.-T., Colloids Surf. A: Physicochem. Eng. Aspects 127, 1 (1997). 11. Dickinson, E., Golding, M., and Povey, M. J. W., J. Colloid Interface Sci. 185, 515 (1997). 12. Dickinson, E., and Golding, M., Food Hydrocolloids 11, 13 (1997). 13. Dickinson, E., and Golding, M., J. Colloid Interface Sci. 191, 166 (1997). 14. Povey, M. J. W., “Ultrasonic Techniques for Emulsion Characterization,” Academic Press, San Diego, 1997. 15. Pinfield, V. J., Dickinson, E., and Povey, M. J. W., Ultrasonics 33, 243 (1995). 16. Dalgleish, D. G., Srinivasan, M., and Singh, H., J. Agric. Food Chem. 43, 2351 (1995). 17. Coke, M., Wilde, P. J., Russel, E. J., and Clark, D. C., J. Colloid Interface Sci. 138, 489 (1990). 18. Clark, D. C., Wilde, P. J., Bergink-Martens, D. J. M., Kokelaar, A. J. J., and Prins, A., in “Food Colloids and Polymers: Stability and Mechanical Properties” (E. Dickinson and P. Walstra, Eds.), p. 354. Royal Society of Chemistry, Cambridge, UK, 1993. 19. Rodriguez Patino, J. M., Rodriguez Nino, R., and Alvarez Gomez, J. M., Food Hydrocolloids 11, 49 (1997). 20. Bergenståhl, B., in “Gums and Stabilisers in the Food Industry” (G. O. Phillips, D. J. Wedlock, and P. A. Williams, Eds.), Vol. 4, p. 363. IRL Press, Oxford, 1988. 21. Bergenståhl, B., Faeldt, P., and Malmsten, M., in “Food Macromolecules and Colloids” (E. Dickinson and D. Lorient, Eds.), p. 201. Royal Society of Chemistry, Cambridge, UK, 1995.