- Email: [email protected]
Rheology and Flocculation of Oil-in-Water Emulsions Made with Mixtures of αs1-Casein + β-Casein
JOURNAL OF COLLOID AND INTERFACE SCIENCE ARTICLE NO.
207, 82– 89 (1998)
CS985744
Rheology and Flocculation of Oil-in-Water Emulsions Made with Mixtures of as1-Casein 1 b-Casein Herley Casanova and Eric Dickinson1 Procter Department of Food Science, University of Leeds, Leeds LS2 9JT, United Kingdom Received February 17, 1998; accepted July 8, 1998
Recent work on model caseinate-stabilized emulsions has shown that so long as there is sufficient protein present to give a fully saturated casein monolayer, the main destabilizing factor is reversible flocculation of the emulsion droplets caused by the presence of excess casein emulsifier existing in the form of small spherical submicelles (6, 7). Depletion flocculation arising from this nonadsorbed protein has a considerable influence on the creaming and rheology of these casein-stabilized emulsions (6 – 8). Around three-quarters of the total protein in sodium caseinate comes from the two major individual caseins, as1-casein and b-casein, present in roughly equal proportions. Both these are calcium-sensitive phosphoproteins containing a high proportion of proline but no cysteine residues along disordered polypeptide chains of ;2 3 102 residues (9). The monomers carry a net negative charge at pH 7 and are substantially amphiphilic in character. Both these pure caseins have a tendency to self-associate in solution and to adsorb at hydrophobic surfaces, and they both can act as effective emulsion stabilizers (10, 11). On the other hand, the detailed distribution of charged, polar, and hydrophobic residues along the polypeptide chain is substantially different for as1-casein and b-casein (9), and at pH 7 the net negative charge on as1-casein (;22 e) is higher than that on b-casein (;15 e). This means that the protein–protein interactions and the types of aggregation behavior for the two caseins are also dissimilar (12). With its distinctly amphiphilic character, the structure of the b-casein molecule resembles that of a low-molecular-weight watersoluble surfactant. This is consistent with the reversible temperature-dependent association of b-casein, at concentrations above ca. 0.05 wt%, into approximately spherical surfactanttype micelles (13), whereas as1-casein forms chainlike aggregates through a series of consecutive association steps (14). Consistent with differing molecular structures, recent theoretical calculations have suggested (15) that the adsorbed layer structures at hydrophobic surfaces are substantially different for these two caseins. The much poorer salt stability of as1casein-coated droplets (11, 16) can be explained in terms of predicted differences in the ionic strength dependence of the interlayer interaction potential (17). Deviations from Newtonian behavior of concentrated emul-
The influence of the composition of a mixed binary protein emulsifier composed of as1-casein 1 b-casein on the rheology of concentrated oil-in-water emulsions (45 vol% oil, 5 wt% protein, pH 7) has been investigated over the temperature range 0 – 40°C. Controlled stress viscometric data are reported over the shear stress range 0.1–30 Pa for systems with as1-casein/b-casein ratios of 100:0, 98:2, 95:5, 90:10, 75:25, 50:50, and 0:100. The pure casein emulsions showed substantially different temperature-dependent rheology, and there was observed to be a pronounced maximum in the small-deformation complex modulus of the pure as1-casein emulsion in the range 30 – 40°C. In the emulsions containing >90% as1-casein in the emulsifier mixture, all of the b-casein present was found to be associated with the surface of the droplets. Average droplet sizes and protein surface coverages were higher in the mixed casein systems than in the equivalent pure casein systems. The strongly pseudoplastic character of the emulsions is consistent with extensive reversible flocculation caused probably by a depletion mechanism involving unadsorbed protein. The degree of flocculation is sensitive to temperature and to the as1casein/b-casein ratio. The results can be interpreted in terms of changes in protein self-assembly and adsorbed layer structure which influence the strength of the interdroplet interactions and hence the rheological behavior of the emulsions. There is some evidence of a specific role for as1-casein–b-casein complexes in these systems. © 1998 Academic Press Key Words: casein interactions; depletion flocculation; emulsion rheology; protein self-assembly; pseudoplastic flow.
INTRODUCTION
Mixtures of milk proteins, in both soluble and dispersed form, are widely valued as food ingredients with excellent emulsifying and emulsion stabilizing properties (1, 2). The caseins are particularly effective in adsorbing at the newly formed oil–water interface during emulsification and in protecting the resulting fine oil droplets against aggregation during subsequent processing and storage (3–5). In a sodium caseinate-stabilized emulsion at neutral pH, the adsorbed layer provides very good stability against droplet coalescence due to a combination of electrostatic and steric interdroplet repulsions. 1
To whom correspondence should be addressed.
0021-9797/98 $25.00 Copyright © 1998 by Academic Press All rights of reproduction in any form reserved.
82
83
RHEOLOGY OF OIL-IN-WATER EMULSIONS
sions can be used as a sensitive indicator of flocculation (18, 19). This paper reports new experimental data on the rheology of concentrated model oil-in-water emulsions prepared at constant protein/oil ratio using well-defined mixtures of as1-casein 1 b-casein as the sole emulsifying agent. These emulsions differ from our previous studies (11, 16, 20) of emulsions made with binary mixtures of as1-casein 1 b-casein in that they contain a high concentration of nonadsorbed protein in the aqueous phase. The objective is to determine how the overall as1-casein/b-casein ratio affects the flocculation and rheological behavior of these emulsions in the temperature range 0 – 40°C.
TABLE 1 Effect of Casein Emulsifier Composition on the Total Protein Surface Coverage (G), the Fraction of the Protein Surface Layer Consisting of as1-Casein ~ f as ), and the Fraction of Bulk-Phase Protein Consisting of as1-Casein ~ f ab ) in Emulsions Prepared with 45 vol% n-Tetradecane and 5 wt% Protein under Identical Homogenization Conditions as1-Casein/b-casein ratio
G (mg m22)a
f as
f ab
100:0 98:2 95:5 90:10 75:25 0:100
1.2 4.9 4.5 3.7 2.2 1.1
100 92 79 55 29 0
100 100 100 100 83 0
MATERIALS AND METHODS
Bovine as1-casein and b-casein were obtained as freezedried samples from the Hannah Research Institute (Ayr, Scotland). These pure casein samples had been prepared from fresh skim milk by acid precipitation, washing, reprecipitation, dissolution in urea, ion-exchange chromatography, and dialysis. Purity with respect to other protein contaminants (,1–2%) was assessed by fast protein liquid chromatography (FPLC). nTetradecane and imidazole (99%) were purchased from Sigma Chemical Co. (St. Louis, MO). Oil-in-water emulsions (45 vol% n-tetradecane, 5 wt% casein) were prepared at room temperature using a laboratory-scale jet homogenizer (jet hole size 0.53 mm, homogenization pressure 300 bar) (21). Prior to emulsification the proteins were dissolved at a concentration of 8.11 wt% in 20 mM imidazole/HCl buffer solution (pH 7). Protein composition was varied by considering the following as1-casein/ b-casein ratios (by weight): 0:100, 50:50, 75:25, 90:10, 95:5, 98:2, and 100:0. Droplet-size distributions and specific surface areas of emulsions were measured using a Malvern Mastersizer S2.01. Controlled stress viscometry and small-deformation oscillatory shear rheology were carried out using a Bohlin CS-50 rheometer with a concentric cylinder arrangement (C25 for stress viscometry, C14 for oscillation). Within 1 h of preparation, the emulsion sample was placed in the rheometer cell, thermally equilibrated at 25°C, and then presheared for 1 min (at 20 Pa) to break up any existing flocs. After the sample was held quiescently at 25°C for 10 min to allow any flocculated structure to rebuild, the temperature was lowered to 0 or 5°C at a rate of 2°C min21. Following completion of this sample preparation procedure, a program of controlled stress viscometry scans was carried out at set temperatures up to 40°C, with the sample left undisturbed for 10 min between each scan. Though for reasons of consistency this same protocol was followed for all the results reported here, it was noted that changing the temperature sequence or the length of the quiescent period did not affect the results. Surface coverage and composition of the protein layer at the oil–water interface were determined by the depletion method.
a
Experimental error 60.4 mg m22.
Prior to centrifugation, the emulsions were diluted 4:1 (by weight) with buffer. This procedure was necessary to reduce the proportion of aqueous phase trapped between the emulsion droplets in the centrifuged cream layer, i.e., the volume of trapped water in comparison with the volume of water separated in the serum layer containing the nonadsorbed protein. Diluted emulsions were centrifuged at 104g for 15 min at 20°C, and the resulting serum layer was removed carefully with a syringe and passed through a 0.22-mm Millipore filter. The concentrations of as1-casein and b-casein were determined by FPLC using a Mono-Q ion-exchange column with a linear NaCl gradient of 0.2– 0.5 M. In the case of serum derived from the 75:25 as1/b ratio emulsion, the protein solution was treated with 6 M urea for 2 h prior to injection on the FPLC column to disrupt the b-casein micelles. (Emulsions of higher as1casein content had no b-casein detectable in the aqueous phase.) Calculation of the surface protein concentration was based on the difference between the amount of protein detected in the aqueous phase and the known amount of protein used to make the emulsion. The estimated error in the surface coverage is 0.4 mg m22. RESULTS
Oil-in-water emulsions were prepared at pH 7 with 45 vol% oil and 5 wt% protein. This emulsion formulation was chosen because it had been shown previously (8) that such a system prepared with sodium caseinate as emulsifier exhibits nonNewtonian behavior indicative of extensive depletion flocculation. Preliminary rheological measurements on sets of emulsion samples made under identical homogenization conditions indicated that replacing a small proportion of as1-casein by b-casein produces a large change in flocculation behavior. Hence, it was decided to focus attention on changes in emulsifier composition toward the as1-casein-rich end of the composition range.
84
CASANOVA AND DICKINSON
Table 1 shows the estimated surface protein concentration G in emulsions prepared with emulsifier mixtures containing 0, 2, 5, 10, 25, and 100% b-casein. A striking feature of the data is the much higher values of G for the casein mixtures than for the pure proteins. This difference (a factor of 4 in G) is very significant, even taking account of the considerable experimental error in G (60.4 mg m22) which arises from the relatively small proportion of adsorbed protein in these systems and uncertainties arising from the small amount of aqueous phase trapped between the droplets in the centrifuged cream layer. Also shown in Table 1 are the calculated fractions of as1-casein in the surface layer and aqueous phase, f as and f ab , respectively. These latter figures indicate the very strong tendency of b-casein, when present in relatively small amounts, to accumulate predominantly at the oil–water interface. Average sizes of emulsion droplets prepared with pure as1-casein (d32 5 0.93 6 0.02 mm) were found to be higher than for those prepared with pure b-casein (d32 5 0.63 6 0.02 mm) or mixtures of as1-casein 1 b-casein containing 2–10% b-casein (1.00 –1.15 mm). This implies greater effectiveness of b-casein as an emulsifier over as1-casein and slightly poorer effectiveness of the mixture due possibly to a greater extent of aggregation. One way of interpreting the substantial increase in surface coverage for the as1-caseinrich mixed systems is in terms of the reduction in the
FIG. 1. Influence of emulsifier composition on average droplet size d32 (upper graph) and surface coverage G (lower graph) in emulsions made at pH 7 with 45 vol% oil and 5 wt% casein and different as1/b ratios: A 5 100:0; B 5 98:2; C 5 95:5; D 5 90:10; E 5 75:25; F 5 50:50; G 5 0:100.
FIG. 2. Controlled stress viscometry of emulsion (45 vol% oil, 5 wt% protein, pH 7) containing pure b-casein. Apparent shear viscosity is plotted against shear stress at different temperatures: ■, 5°C; h, 10°C; F, 20°C; E, 25°C; Œ, 30°C; D, 35°C; }, 40°C.
specific surface area ~}d 21 32 ). This is supported by the data in Fig. 1 showing a correlation in the trends of average particle size d32 and protein surface coverage G as a function of overall protein emulsifier composition. Figure 2 shows apparent viscosity versus shear stress for the pure b-casein emulsion sample. While the shear rheology is non-Newtonian over the whole temperature range studied, the lower temperature behavior (5–20°C) is qualitatively different from the higher temperature behavior (25– 40°C). In fact, the lower temperature b-casein emulsion behavior is rather similar to that found previously (8) for the equivalent sodium caseinate emulsion, i.e., the apparent viscosity remaining relatively constant (actually slightly decreasing) up to a stress of ca. 2 Pa and then falling steadily with further increase in applied stress. The shoulder in the viscosity–stress plot can be attributed to a putative flocculated network structure which breaks down into pseudoplastic flow beyond a certain critical “yield” stress. This feature was not present in the viscosity–stress plots obtained for the b-casein system at a temperature of 25°C or above, although the emulsion was still strongly shear-thinning with a higher low-stress viscosity than that found at the lower temperatures. These results indicate that, while the pure b-casein emulsion is always flocculated over the temperature range 5– 40°C, the state of flocculation, as indicated by the controlled stress rheometry, is rather sensitive to temperature (especially near 20 –25°C). The controlled stress viscometry of the pure b-casein emulsion is strikingly different from that of the pure as1casein emulsion, as illustrated in Fig. 3. In the range 0 –30°C, the system is fairly close to Newtonian, with the measured viscosity gradually decreasing with increasing
RHEOLOGY OF OIL-IN-WATER EMULSIONS
FIG. 3. Controlled stress viscometry of emulsion (45 vol% oil, 5 wt% protein, pH 7) containing pure as1-casein. Apparent shear viscosity is plotted against shear stress at different temperatures: ■, 0°C; h, 10°C; F, 20°C; E, 30°C; Œ, 40°C.
temperature. In contrast, the data at 40°C show highly pseudoplastic behavior with low-stress viscosities some two orders of magnitude higher than those measured at 30°C. To provide confirmatory evidence for the implied large change in small-deformation rheology in the range 30 – 40°C, oscillatory measurements were carried out at 1 Hz on the pure as1-casein emulsion at temperatures from 15 to 50°C. Figure 4 shows a plot of complex modulus G* against temperature. We observe a pronounced maximum in this smalldeformation rheological parameter at around 35°C and a large increase in G* between 30 and 40°C which is consistent with the low-stress viscometry data in Fig. 3. As with the pure b-casein emulsion, the rheological measurements indicate a substantial temperature dependence in the state of flocculation of the casein-stabilized emulsion system. We turn now to the rheology of emulsions made with mixtures of as1-casein 1 b-casein. Figure 5 shows viscometry data over the temperature range 5– 40°C for the 50:50 as1/b ratio system. The trend of behavior is similar to that for the pure b-casein system at low temperature (Fig. 2) and also for the sodium caseinate system at 30°C (8). The same trend of behavior is also seen in the viscometry data for the systems of as1/b ratio 75:25 (Fig. 6) and 90:10 (Fig. 7). That is, the apparent viscosity shows just a slow decline with increasing stress up to a certain critical “yield” stress (;2 Pa) and then falls strongly with further increase in shear stress. It is assumed that the position of the shoulder on each curve is indicative of the breakdown stress of the putative flocculated network structure. Although the apparent viscosity shows a steady gradual reduction with increasing temperature, there is no qualitative change in the type of pseudoplastic flow behavior with temperature, as was found for the pure casein systems (see Figs. 2 and 3). In particular, the value of the “yield” stress appears to be relatively independent of temperature. It is apparent from
85
Figs. 5–7 that b-casein has a predominant influence on the rheology of these systems and that a mixed casein emulsion containing up to 90% as1-casein behaves quite differently from a pure as1-casein emulsion. One may ask, then, just how small a fraction of the as1casein emulsifier can be replaced by b-casein to make the emulsion behave like the equivalent sodium caseinate system. To attempt to answer this question, we prepared emulsions with only 5% or 2% of the protein emulsifier mixture consisting of b-casein. Figure 8 shows viscometry data over the temperature range 0 – 40°C for the 95:5 as1/b ratio system. Here, again, we do see pseudoplastic flow with a distinct “yield” point in the curves over the temperature range 10 – 40°C. However, at 0°C the shape of the curve is considerably more monotonic, suggesting some qualitative change in the state of flocculation. This trend is greatly enhanced in the 98:2 as1/b ratio system (Fig. 9). Certainly now these rheological data do resemble much more closely the set of curves obtained for the pure as1-casein emulsion (Fig. 3). For instance, like the 100% as1-casein system, the 98:2 as1/b ratio emulsion behaves like a low-viscosity Newtonian liquid at 0°C and is strongly pseudoplastic with a high low-stress viscosity ($100 Pa s) when the temperature is raised toward 40°C. However, the presence of just 2% b-casein still apparently has a predominant effect on the rheology at 20 –30°C, conferring upon the system a “yield”-type behavior similar to that found for all the other mixed systems (but not the pure as1-casein emulsion). The implication is that even the presence of only 1–2% b-casein in the protein emulsifier has an influence on the resulting emulsion rheology, and therefore, by inference, on the flocculation state of the droplets.
FIG. 4. Influence of temperature on the small-deformation complex modulus G* of emulsion (45 vol% oil, 5 wt% protein, pH 7) made from pure as1-casein.
86
CASANOVA AND DICKINSON
FIG. 5. Controlled stress viscometry of emulsion (45 vol% oil, 5 wt% protein, pH 7) containing 50:50 mixture of as1-casein 1 b-casein. Apparent shear viscosity is plotted against shear stress at different temperatures: ■, 5°C; h, 10°C; F, 20°C; E, 30°C; Œ, 40°C.
DISCUSSION
The foregoing results show, first, that the temperature dependence of the shear rheology of a concentrated as1-caseinstabilized emulsion containing excess protein is substantially different from that of the equivalent b-casein-stabilized emulsion, and, second, that the rheology of the corresponding binary casein-stabilized emulsions is sensitive to the as1-casein/b-casein ratio, especially at emulsifier compositions rich in as1-casein. The explanation for the behavior is presumably something to do with differences in the self-assembly properties and the protein–protein interactions for these two caseins (12–14) and also with changes in distribution of the caseins between the interface and the bulk aqueous phase as a function of overall system composition (20). The pseudoplastic behavior of the equivalent sodium caseinate-stabilized emulsions was previously attributed (6 – 8) to depletion flocculation by nonadsorbed casein submicelles. As it is known that the size and shape of casein aggregates are also sensitive to casein composition and that the depletion interaction is itself sensitive to such factors, it is certainly not surprising that the rheology of flocculated casein emulsions is influenced by the protein composition. Before we discuss possible physicochemical origins of the observed rheological behavior, it is appropriate to mention two aspects of the systems studied that could have some bearing on certain aspects of the experimental data: (i) “wall slip” effects and (ii) creaming effects. It is established (22) that flocculated suspensions and emulsions are susceptible to the phenomenon of wall slip, which involves depletion of dispersed phase from the vicinity of a solid boundary and hence the formation of a thin lubricating layer near the wall of the rheometer cell.
Evidence for some wall slip in caseinate-based emulsions at low shear stresses (;0.5 Pa) has recently been found (23) in comparative experiments carried out with a specially made rheometer cell made from roughened glass surfaces replacing the standard stainless steel cell supplied by the manufacturer. The presence of wall slip implies that the absolute viscosity values recorded for the flocculated casein emulsions in this paper are probably underestimated somewhat at the lowest stresses (,1 Pa). The more subtle changes in the very low stress rheology noted above could reasonably be attributable in part to this phenomenon. However, data at higher stresses ($1 Pa) should be unaffected, and certainly the qualitative trends in shape and position of the non-Newtonian flow curves as a function of temperature and casein composition should not be affected by wall slip effects. Turning now to the possible influence of enhanced creaming of depletion flocculated emulsions on the measured rheological data, we have found no indication of any such problem in our experiments. The strongest evidence for this view comes from the good reproducibility and reversibility of the reported non-Newtonian flow curves. Within the normal experimental error of the instrument, exactly the same results were obtained when measurements were immediately repeated in the same rheometer cell without emulsion mixing. If emulsion creaming during the rheological measurements were to have been significantly affecting the results, such repeatability would not be expected. It is highly likely that the main reason for the complex rheological behavior reported here lies in the different selfassembly behavior and emulsion stabilizing properties of the two pure caseins. The highly amphiphilic b-casein is known to form surfactant-type micelles under the thermodynamic driving force of strong hydrophobic interactions in a reversible
FIG. 6. Controlled stress viscometry of emulsion (45 vol% oil, 5 wt% protein, pH 7) containing 75:25 mixture of as1-casein 1 b-casein. Apparent shear viscosity is plotted against shear stress at different temperatures: ■, 5°C; h, 10°C; F, 20°C; E, 30°C; Œ, 40°C.
RHEOLOGY OF OIL-IN-WATER EMULSIONS
FIG. 7. Controlled stress viscometry of emulsion (45 vol% oil, 5 wt% protein, pH 7) containing 90:10 mixture of as1-casein 1 b-casein. Apparent shear viscosity is plotted against shear stress at different temperatures: ■, 5°C; h, 10°C; F, 20°C; E, 30°C; Œ, 40°C.
process of endothermic self-assembly (24). The critical micelle concentration is quite low (,0.1 wt% at 20°C) (12). So, at room temperature, we can assume that essentially all of the unadsorbed protein present in a b-casein-stabilized emulsion is in the form of micelles containing about 40 monomers (25). But, since b-casein micelles are apparently completely dissociated below 5°C (26), if their presence is an important factor controlling the state of flocculation of b-casein-stabilized emulsions, we might have expected considerable differences in emulsion rheology between 5 and 20°C. However, this is not quite what occurs experimentally (see Fig. 1). While the rheology of the pure b-casein emulsion is certainly dependent on temperature, there is clearly less change over the range 5–15°C than over the range 20 – 40°C, and the greatest sensitivity of viscosity to temperature occurs not around 5°C, but at 20 – 25°C. The latter effect could still be associated with changes in casein self-assembly, however, because it has been suggested (25) that the packing density of b-casein micelles increases with temperature (while hydrodynamic size remains essentially constant) and that b-casein micelles may themselves aggregate as the temperature approaches 40°C. The more highly charged as1-casein associates in aqueous solution through a series of consecutive association steps (14). Light-scattering measurements have suggested (27) a cylindrical “wormlike” micelle structure as the most probable structure of as1-casein aggregates at 35°C over the concentration range 0.05– 0.6 wt%. The aggregate molecular weight (corresponding to ca. 120 monomers) is apparently relatively insensitive to temperature. It is not easy to reconcile this dilute solution structural information with the temperature-dependent rheology of pure as1-casein emulsions in Fig. 3. While it is not unreasonable to speculate that long
87
wormlike rods of aggregated as1-casein might be very effective at inducing depletion flocculation, it remains unclear why this flocculation should only be manifest in the rheology at 35–40°C (see Figs. 3 and 4). The emulsions prepared with mixtures of as1-casein 1 b-casein also exhibit trends in the rheology that, while entirely systematic and consistent, are difficult to summarize simply— and even more difficult to explain. Figure 10 presents a simplified overview of the data in terms of apparent viscosities for the different temperatures and compositions at two fixed stress values: (a) 1 Pa and (b) 0.1 Pa. The latter value is chosen because it is lowest stress studied. The value of 1 Pa is chosen because it lies just below the critical “yield” stress, which is a characteristic feature of several of the viscosity–stress plots reported here, being also found in the rheology of the equivalent sodium caseinate system (8). We note from Fig. 10a that there is a substantial viscosity value at 1 Pa for emulsions containing from 50 to 95% as1-casein and that only in the pure (or nearly pure) as1-casein emulsions at ,40°C, or the pure b-casein emulsions at .30°C, is this not the case. That is, assuming that a moderately high apparent viscosity at 1 Pa (say, .10 Pa s) can be regarded as an appropriate criterion for a flocculated network structure of finite strength, it appears that the creaming and rheological behavior observed previously for sodium caseinate emulsions can be reproduced also over a wide range of casein compositions well away from the natural composition of bovine milk. It is apparent that the temperature dependence of the rheology is very sensitive to casein emulsifier composition. For instance, Fig. 10b shows that replacing most (but not all) of the b-casein protein emulsifier by as1casein leads to a complete reversal of the temperature dependence of the viscosity at 0.1 Pa from strongly positive (system G) to strongly negative (system C). While we should be careful
FIG. 8. Controlled stress viscometry of emulsion (45 vol% oil, 5 wt% protein, pH 7) containing 95:5 mixture of as1-casein 1 b-casein. Apparent shear viscosity is plotted against shear stress at different temperatures: ■, 0°C; h, 10°C; F, 20°C; E, 30°C; Œ, 40°C.
88
CASANOVA AND DICKINSON
FIG. 9. Controlled stress viscometry of emulsion (45 vol% oil, 5 wt% protein, pH 7) containing 98:2 mixture of as1-casein 1 b-casein. Apparent shear viscosity is plotted against shear stress at different temperatures: ■, 0°C; h, 10°C; F, 20°C; E, 30°C; Œ, 40°C.
about overinterpreting trends of viscosity behavior obtained at such low stresses because of the possible role of wall slip effects (as discussed earlier), such changes may be explicable in terms of a change in the self-assembled protein structure from predominantly spherical b-casein-rich micelles to predominantly rodlike as1-casein micelles. Additionally, there is definite evidence in the literature (28 –31) for specific as1casein–b-casein complexes of high stoichiometric ratio in dilute solution. This could mean that the mixed aggregated system resembles neither the pure as1-casein associated structure nor the pure b-casein micellar structure. Individual molecules of as1-casein and b-casein can be expected to interact attractively through a combination of hydrophobic forces, hydrogen bonding, and electrostatic interactions (the latter, say, through interaction of the negatively charged phosphoserinerich tail of b-casein with the cluster of positively charged residues at the extreme N-terminus of as1-casein). Overall, however, we know too little about molecular association in concentrated solutions of mixed caseins to speculate much at this stage about implications of complexing for the emulsion properties. The observed tendency for b-casein to predominate at the surface of the emulsion droplets (see Table 1) is consistent with our earlier work (20) on competitive adsorption of as1-casein 1 b-casein in emulsions of lower oil content. In the concentrated emulsion systems studied here containing $90% as1-casein, all of the b-casein present is adsorbed at the oil–water interface. This means that any differences between the pure as1-casein system and systems containing a small proportion of b-casein (i.e. systems B, C, and D) must surely be attributable mainly to changes in adsorbed layer composition. Unfortunately, as little is currently
known about the structure of mixed layers of as1-casein 1 b-casein, and even less about the effect of casein composition on interlayer interactions, it seems inadvisable to speculate further here on the effect of casein competitive adsorption on the emulsion rheology. Finally, though, it is appropriate to comment on the influence of casein emulsifier composition on average droplet size and surface coverage in these concentrated emulsions. Compared with the pure emulsion systems, the mixed casein emulsions are characterized by larger droplet sizes and much higher surface coverages, especially for systems toward the extreme as1-casein-rich end of the composition range (systems B and C in Fig. 1). A reduced emulsifying activity index for binary mixtures of as1-casein 1 b-casein as compared with the pure proteins was also reported by Cayot et al. (32). The larger droplet sizes for the mixed casein systems in Fig. 1 could be connected in some way with the large increase in emulsion viscosity as small amounts of as1-casein are replaced by b-casein (Fig. 10). However, this is not really a convincing explanation because
FIG. 10. Overview of effects of temperature and emulsifier composition on the viscosity of emulsions (45 vol% oil, 5 wt% protein, pH 7) made with mixtures of as1-casein 1 b-casein. Values of apparent shear viscosity at constant stresses of (a) 1 Pa and (b) 0.1 Pa are plotted for the different as1/b ratios: A 5 100:0; B 5 98:2; C 5 95:5; D 5 90:10; E 5 75:25; F 5 50:50; G 5 0:100. Symbol key in plot (b): F, 10°C; D, 20°C; Œ, 30°C; E, 40°C.
RHEOLOGY OF OIL-IN-WATER EMULSIONS
the shear stresses experienced during homogenization are very much higher than those explored in the viscometric study. More plausible, perhaps, is an explanation based on the formation of an as1-casein–b-casein complex. This as1casein–b-casein association might reduce the effectiveness of the protein emulsifier during emulsification in its role of rapidly adsorbing at the oil–water interface to form a protective layer around the newly formed droplets (32). The low protein surface concentration found in the pure casein emulsions is consistent with stabilization by a monolayer of adsorbed monomeric protein. The considerably higher surface coverage in the binary casein systems is consistent with multilayer formation around the surface of the mixed protein-coated droplets due to complexation and enhanced protein aggregation at the interface. ACKNOWLEDGMENT H.C. acknowledges financial support from COLCIENCIAS (Colombia).
REFERENCES 1. Morr, C. V., in “Developments in Dairy Chemistry” (P. F. Fox, Ed.), Vol. 1, p. 375. Applied Science, London, 1982. 2. Doxastakis, G., in “Food Emulsifiers” (G. Charalambous and G. Doxastaskis, Eds.), p. 9. Elsevier, Amsterdam, 1989. 3. Dickinson, E., J. Chem. Soc., Faraday Trans. 88, 2973 (1992). 4. Dalgleish, D. G., Trends Food Sci. Technol. 8, 1 (1997). 5. Dickinson, E., J. Dairy Sci. 80, 2607 (1997). 6. Dickinson, E., and Golding, M., Food Hydrocolloids 11, 13 (1997). 7. Dickinson, E., Golding, M., and Povey, M. J. W., J. Colloid Interface Sci. 185, 515 (1997). 8. Dickinson, E., and Golding, M., J. Colloid Interface Sci. 191, 166 (1997). 9. Swaisgood, H. E., in “Developments in Dairy Chemistry—1” (P. F. Fox, Ed.), Vol. 1, p. 1. Applied Science, London, 1982. 10. Dickinson, E., J. Dairy Res. 56, 471 (1989).
89
11. Dickinson, E., Whyman, R. H., and Dalgleish, D. G., in “Food Emulsions and Foams” (E. Dickinson, Ed.), p. 40. Royal Society of Chemistry, London, 1987. 12. Schmidt, D. G., and Payens, T. A. J., in “Surface and Colloid Science” (E. Matijevic, Ed.), Vol. 9, p. 165. Wiley, New York, 1976. 13. Leclerc, E., and Calmettes, P., Phys. Rev. Lett. 78, 150 (1997). 14. Schmidt, D. G., in “Developments in Dairy Chemistry—1” (P. F. Fox, Ed.), Vol. 1, p. 61. Applied Science, London, 1982. 15. Dickinson, E., Horne, D. S., Pinfield, V. J., and Leermakers, F. A. M., J. Chem. Soc., Faraday Trans. 93, 425 (1997). 16. Casanova, H., and Dickinson, E., J. Agric. Food Chem. 46, 72 (1998). 17. Dickinson, E., Pinfield, V. J., Horne, D. S., and Leermakers, F. A. M., J. Chem. Soc., Faraday Trans. 93, 1785 (1997). 18. Dickinson, E., and Stainsby, G., in “Colloids in Food,” Chapter 7. Applied Science, London, 1982. 19. Pal, R., Colloids Surf. 64, 207 (1992). 20. Dickinson, E., Rolfe, S. E., and Dalgleish, D. G., Food Hydrocolloids 2, 397 (1988). 21. Burgaud, I., Dickinson, E., and Nelson, P. V., Int. J. Food Sci. Technol. 25, 39 (1990). 22. Barnes, H. A., J. Non-Newtonian Fluid Mech. 56, 221 (1995). 23. Dickinson, E., Golding, M., and Casanova, H., in “Food Emulsions and Foams: Interfaces, Interactions and Stability” (E. Dickinson and J. M. Rodriguez Patino, Eds.). Royal Society of Chemistry, Cambridge (in press). 24. Tanford, C., “The Hydrophobic Effect.” Wiley, London, 1973. 25. Thurn, A., Burchard, W., and Niki, R., Colloid Polym. Sci. 265, 653 (1987). 26. Payens, T. A. J., and van Markwijk, B. W., Biochim. Biophys. Acta 71, 517 (1963). 27. Thurn, A., Burchard, W., and Niki, R., Colloid Polym. Sci. 265, 897 (1987). 28. Payens, T. A. J., and Nijhuis, H., Biochim. Biophys. Acta 336, 201 (1974). 29. Yoshikawa, M., Sugimoto, E., and Chiba, H., Agric. Biol. Chem. 39, 1843 (1975). 30. Pepper, L., and Farrell, H. M., Jr., J. Dairy Sci. 65, 2259 (1982). 31. Ono, T., Odagiri, S., and Takagi, T., J. Dairy Res. 50, 37 (1983). 32. Cayot, P., Courthaudon, J.-L., and Lorient, D., J. Agric. Food Chem. 39, 1369 (1991).
207, 82– 89 (1998)
CS985744
Rheology and Flocculation of Oil-in-Water Emulsions Made with Mixtures of as1-Casein 1 b-Casein Herley Casanova and Eric Dickinson1 Procter Department of Food Science, University of Leeds, Leeds LS2 9JT, United Kingdom Received February 17, 1998; accepted July 8, 1998
Recent work on model caseinate-stabilized emulsions has shown that so long as there is sufficient protein present to give a fully saturated casein monolayer, the main destabilizing factor is reversible flocculation of the emulsion droplets caused by the presence of excess casein emulsifier existing in the form of small spherical submicelles (6, 7). Depletion flocculation arising from this nonadsorbed protein has a considerable influence on the creaming and rheology of these casein-stabilized emulsions (6 – 8). Around three-quarters of the total protein in sodium caseinate comes from the two major individual caseins, as1-casein and b-casein, present in roughly equal proportions. Both these are calcium-sensitive phosphoproteins containing a high proportion of proline but no cysteine residues along disordered polypeptide chains of ;2 3 102 residues (9). The monomers carry a net negative charge at pH 7 and are substantially amphiphilic in character. Both these pure caseins have a tendency to self-associate in solution and to adsorb at hydrophobic surfaces, and they both can act as effective emulsion stabilizers (10, 11). On the other hand, the detailed distribution of charged, polar, and hydrophobic residues along the polypeptide chain is substantially different for as1-casein and b-casein (9), and at pH 7 the net negative charge on as1-casein (;22 e) is higher than that on b-casein (;15 e). This means that the protein–protein interactions and the types of aggregation behavior for the two caseins are also dissimilar (12). With its distinctly amphiphilic character, the structure of the b-casein molecule resembles that of a low-molecular-weight watersoluble surfactant. This is consistent with the reversible temperature-dependent association of b-casein, at concentrations above ca. 0.05 wt%, into approximately spherical surfactanttype micelles (13), whereas as1-casein forms chainlike aggregates through a series of consecutive association steps (14). Consistent with differing molecular structures, recent theoretical calculations have suggested (15) that the adsorbed layer structures at hydrophobic surfaces are substantially different for these two caseins. The much poorer salt stability of as1casein-coated droplets (11, 16) can be explained in terms of predicted differences in the ionic strength dependence of the interlayer interaction potential (17). Deviations from Newtonian behavior of concentrated emul-
The influence of the composition of a mixed binary protein emulsifier composed of as1-casein 1 b-casein on the rheology of concentrated oil-in-water emulsions (45 vol% oil, 5 wt% protein, pH 7) has been investigated over the temperature range 0 – 40°C. Controlled stress viscometric data are reported over the shear stress range 0.1–30 Pa for systems with as1-casein/b-casein ratios of 100:0, 98:2, 95:5, 90:10, 75:25, 50:50, and 0:100. The pure casein emulsions showed substantially different temperature-dependent rheology, and there was observed to be a pronounced maximum in the small-deformation complex modulus of the pure as1-casein emulsion in the range 30 – 40°C. In the emulsions containing >90% as1-casein in the emulsifier mixture, all of the b-casein present was found to be associated with the surface of the droplets. Average droplet sizes and protein surface coverages were higher in the mixed casein systems than in the equivalent pure casein systems. The strongly pseudoplastic character of the emulsions is consistent with extensive reversible flocculation caused probably by a depletion mechanism involving unadsorbed protein. The degree of flocculation is sensitive to temperature and to the as1casein/b-casein ratio. The results can be interpreted in terms of changes in protein self-assembly and adsorbed layer structure which influence the strength of the interdroplet interactions and hence the rheological behavior of the emulsions. There is some evidence of a specific role for as1-casein–b-casein complexes in these systems. © 1998 Academic Press Key Words: casein interactions; depletion flocculation; emulsion rheology; protein self-assembly; pseudoplastic flow.
INTRODUCTION
Mixtures of milk proteins, in both soluble and dispersed form, are widely valued as food ingredients with excellent emulsifying and emulsion stabilizing properties (1, 2). The caseins are particularly effective in adsorbing at the newly formed oil–water interface during emulsification and in protecting the resulting fine oil droplets against aggregation during subsequent processing and storage (3–5). In a sodium caseinate-stabilized emulsion at neutral pH, the adsorbed layer provides very good stability against droplet coalescence due to a combination of electrostatic and steric interdroplet repulsions. 1
To whom correspondence should be addressed.
0021-9797/98 $25.00 Copyright © 1998 by Academic Press All rights of reproduction in any form reserved.
82
83
RHEOLOGY OF OIL-IN-WATER EMULSIONS
sions can be used as a sensitive indicator of flocculation (18, 19). This paper reports new experimental data on the rheology of concentrated model oil-in-water emulsions prepared at constant protein/oil ratio using well-defined mixtures of as1-casein 1 b-casein as the sole emulsifying agent. These emulsions differ from our previous studies (11, 16, 20) of emulsions made with binary mixtures of as1-casein 1 b-casein in that they contain a high concentration of nonadsorbed protein in the aqueous phase. The objective is to determine how the overall as1-casein/b-casein ratio affects the flocculation and rheological behavior of these emulsions in the temperature range 0 – 40°C.
TABLE 1 Effect of Casein Emulsifier Composition on the Total Protein Surface Coverage (G), the Fraction of the Protein Surface Layer Consisting of as1-Casein ~ f as ), and the Fraction of Bulk-Phase Protein Consisting of as1-Casein ~ f ab ) in Emulsions Prepared with 45 vol% n-Tetradecane and 5 wt% Protein under Identical Homogenization Conditions as1-Casein/b-casein ratio
G (mg m22)a
f as
f ab
100:0 98:2 95:5 90:10 75:25 0:100
1.2 4.9 4.5 3.7 2.2 1.1
100 92 79 55 29 0
100 100 100 100 83 0
MATERIALS AND METHODS
Bovine as1-casein and b-casein were obtained as freezedried samples from the Hannah Research Institute (Ayr, Scotland). These pure casein samples had been prepared from fresh skim milk by acid precipitation, washing, reprecipitation, dissolution in urea, ion-exchange chromatography, and dialysis. Purity with respect to other protein contaminants (,1–2%) was assessed by fast protein liquid chromatography (FPLC). nTetradecane and imidazole (99%) were purchased from Sigma Chemical Co. (St. Louis, MO). Oil-in-water emulsions (45 vol% n-tetradecane, 5 wt% casein) were prepared at room temperature using a laboratory-scale jet homogenizer (jet hole size 0.53 mm, homogenization pressure 300 bar) (21). Prior to emulsification the proteins were dissolved at a concentration of 8.11 wt% in 20 mM imidazole/HCl buffer solution (pH 7). Protein composition was varied by considering the following as1-casein/ b-casein ratios (by weight): 0:100, 50:50, 75:25, 90:10, 95:5, 98:2, and 100:0. Droplet-size distributions and specific surface areas of emulsions were measured using a Malvern Mastersizer S2.01. Controlled stress viscometry and small-deformation oscillatory shear rheology were carried out using a Bohlin CS-50 rheometer with a concentric cylinder arrangement (C25 for stress viscometry, C14 for oscillation). Within 1 h of preparation, the emulsion sample was placed in the rheometer cell, thermally equilibrated at 25°C, and then presheared for 1 min (at 20 Pa) to break up any existing flocs. After the sample was held quiescently at 25°C for 10 min to allow any flocculated structure to rebuild, the temperature was lowered to 0 or 5°C at a rate of 2°C min21. Following completion of this sample preparation procedure, a program of controlled stress viscometry scans was carried out at set temperatures up to 40°C, with the sample left undisturbed for 10 min between each scan. Though for reasons of consistency this same protocol was followed for all the results reported here, it was noted that changing the temperature sequence or the length of the quiescent period did not affect the results. Surface coverage and composition of the protein layer at the oil–water interface were determined by the depletion method.
a
Experimental error 60.4 mg m22.
Prior to centrifugation, the emulsions were diluted 4:1 (by weight) with buffer. This procedure was necessary to reduce the proportion of aqueous phase trapped between the emulsion droplets in the centrifuged cream layer, i.e., the volume of trapped water in comparison with the volume of water separated in the serum layer containing the nonadsorbed protein. Diluted emulsions were centrifuged at 104g for 15 min at 20°C, and the resulting serum layer was removed carefully with a syringe and passed through a 0.22-mm Millipore filter. The concentrations of as1-casein and b-casein were determined by FPLC using a Mono-Q ion-exchange column with a linear NaCl gradient of 0.2– 0.5 M. In the case of serum derived from the 75:25 as1/b ratio emulsion, the protein solution was treated with 6 M urea for 2 h prior to injection on the FPLC column to disrupt the b-casein micelles. (Emulsions of higher as1casein content had no b-casein detectable in the aqueous phase.) Calculation of the surface protein concentration was based on the difference between the amount of protein detected in the aqueous phase and the known amount of protein used to make the emulsion. The estimated error in the surface coverage is 0.4 mg m22. RESULTS
Oil-in-water emulsions were prepared at pH 7 with 45 vol% oil and 5 wt% protein. This emulsion formulation was chosen because it had been shown previously (8) that such a system prepared with sodium caseinate as emulsifier exhibits nonNewtonian behavior indicative of extensive depletion flocculation. Preliminary rheological measurements on sets of emulsion samples made under identical homogenization conditions indicated that replacing a small proportion of as1-casein by b-casein produces a large change in flocculation behavior. Hence, it was decided to focus attention on changes in emulsifier composition toward the as1-casein-rich end of the composition range.
84
CASANOVA AND DICKINSON
Table 1 shows the estimated surface protein concentration G in emulsions prepared with emulsifier mixtures containing 0, 2, 5, 10, 25, and 100% b-casein. A striking feature of the data is the much higher values of G for the casein mixtures than for the pure proteins. This difference (a factor of 4 in G) is very significant, even taking account of the considerable experimental error in G (60.4 mg m22) which arises from the relatively small proportion of adsorbed protein in these systems and uncertainties arising from the small amount of aqueous phase trapped between the droplets in the centrifuged cream layer. Also shown in Table 1 are the calculated fractions of as1-casein in the surface layer and aqueous phase, f as and f ab , respectively. These latter figures indicate the very strong tendency of b-casein, when present in relatively small amounts, to accumulate predominantly at the oil–water interface. Average sizes of emulsion droplets prepared with pure as1-casein (d32 5 0.93 6 0.02 mm) were found to be higher than for those prepared with pure b-casein (d32 5 0.63 6 0.02 mm) or mixtures of as1-casein 1 b-casein containing 2–10% b-casein (1.00 –1.15 mm). This implies greater effectiveness of b-casein as an emulsifier over as1-casein and slightly poorer effectiveness of the mixture due possibly to a greater extent of aggregation. One way of interpreting the substantial increase in surface coverage for the as1-caseinrich mixed systems is in terms of the reduction in the
FIG. 1. Influence of emulsifier composition on average droplet size d32 (upper graph) and surface coverage G (lower graph) in emulsions made at pH 7 with 45 vol% oil and 5 wt% casein and different as1/b ratios: A 5 100:0; B 5 98:2; C 5 95:5; D 5 90:10; E 5 75:25; F 5 50:50; G 5 0:100.
FIG. 2. Controlled stress viscometry of emulsion (45 vol% oil, 5 wt% protein, pH 7) containing pure b-casein. Apparent shear viscosity is plotted against shear stress at different temperatures: ■, 5°C; h, 10°C; F, 20°C; E, 25°C; Œ, 30°C; D, 35°C; }, 40°C.
specific surface area ~}d 21 32 ). This is supported by the data in Fig. 1 showing a correlation in the trends of average particle size d32 and protein surface coverage G as a function of overall protein emulsifier composition. Figure 2 shows apparent viscosity versus shear stress for the pure b-casein emulsion sample. While the shear rheology is non-Newtonian over the whole temperature range studied, the lower temperature behavior (5–20°C) is qualitatively different from the higher temperature behavior (25– 40°C). In fact, the lower temperature b-casein emulsion behavior is rather similar to that found previously (8) for the equivalent sodium caseinate emulsion, i.e., the apparent viscosity remaining relatively constant (actually slightly decreasing) up to a stress of ca. 2 Pa and then falling steadily with further increase in applied stress. The shoulder in the viscosity–stress plot can be attributed to a putative flocculated network structure which breaks down into pseudoplastic flow beyond a certain critical “yield” stress. This feature was not present in the viscosity–stress plots obtained for the b-casein system at a temperature of 25°C or above, although the emulsion was still strongly shear-thinning with a higher low-stress viscosity than that found at the lower temperatures. These results indicate that, while the pure b-casein emulsion is always flocculated over the temperature range 5– 40°C, the state of flocculation, as indicated by the controlled stress rheometry, is rather sensitive to temperature (especially near 20 –25°C). The controlled stress viscometry of the pure b-casein emulsion is strikingly different from that of the pure as1casein emulsion, as illustrated in Fig. 3. In the range 0 –30°C, the system is fairly close to Newtonian, with the measured viscosity gradually decreasing with increasing
RHEOLOGY OF OIL-IN-WATER EMULSIONS
FIG. 3. Controlled stress viscometry of emulsion (45 vol% oil, 5 wt% protein, pH 7) containing pure as1-casein. Apparent shear viscosity is plotted against shear stress at different temperatures: ■, 0°C; h, 10°C; F, 20°C; E, 30°C; Œ, 40°C.
temperature. In contrast, the data at 40°C show highly pseudoplastic behavior with low-stress viscosities some two orders of magnitude higher than those measured at 30°C. To provide confirmatory evidence for the implied large change in small-deformation rheology in the range 30 – 40°C, oscillatory measurements were carried out at 1 Hz on the pure as1-casein emulsion at temperatures from 15 to 50°C. Figure 4 shows a plot of complex modulus G* against temperature. We observe a pronounced maximum in this smalldeformation rheological parameter at around 35°C and a large increase in G* between 30 and 40°C which is consistent with the low-stress viscometry data in Fig. 3. As with the pure b-casein emulsion, the rheological measurements indicate a substantial temperature dependence in the state of flocculation of the casein-stabilized emulsion system. We turn now to the rheology of emulsions made with mixtures of as1-casein 1 b-casein. Figure 5 shows viscometry data over the temperature range 5– 40°C for the 50:50 as1/b ratio system. The trend of behavior is similar to that for the pure b-casein system at low temperature (Fig. 2) and also for the sodium caseinate system at 30°C (8). The same trend of behavior is also seen in the viscometry data for the systems of as1/b ratio 75:25 (Fig. 6) and 90:10 (Fig. 7). That is, the apparent viscosity shows just a slow decline with increasing stress up to a certain critical “yield” stress (;2 Pa) and then falls strongly with further increase in shear stress. It is assumed that the position of the shoulder on each curve is indicative of the breakdown stress of the putative flocculated network structure. Although the apparent viscosity shows a steady gradual reduction with increasing temperature, there is no qualitative change in the type of pseudoplastic flow behavior with temperature, as was found for the pure casein systems (see Figs. 2 and 3). In particular, the value of the “yield” stress appears to be relatively independent of temperature. It is apparent from
85
Figs. 5–7 that b-casein has a predominant influence on the rheology of these systems and that a mixed casein emulsion containing up to 90% as1-casein behaves quite differently from a pure as1-casein emulsion. One may ask, then, just how small a fraction of the as1casein emulsifier can be replaced by b-casein to make the emulsion behave like the equivalent sodium caseinate system. To attempt to answer this question, we prepared emulsions with only 5% or 2% of the protein emulsifier mixture consisting of b-casein. Figure 8 shows viscometry data over the temperature range 0 – 40°C for the 95:5 as1/b ratio system. Here, again, we do see pseudoplastic flow with a distinct “yield” point in the curves over the temperature range 10 – 40°C. However, at 0°C the shape of the curve is considerably more monotonic, suggesting some qualitative change in the state of flocculation. This trend is greatly enhanced in the 98:2 as1/b ratio system (Fig. 9). Certainly now these rheological data do resemble much more closely the set of curves obtained for the pure as1-casein emulsion (Fig. 3). For instance, like the 100% as1-casein system, the 98:2 as1/b ratio emulsion behaves like a low-viscosity Newtonian liquid at 0°C and is strongly pseudoplastic with a high low-stress viscosity ($100 Pa s) when the temperature is raised toward 40°C. However, the presence of just 2% b-casein still apparently has a predominant effect on the rheology at 20 –30°C, conferring upon the system a “yield”-type behavior similar to that found for all the other mixed systems (but not the pure as1-casein emulsion). The implication is that even the presence of only 1–2% b-casein in the protein emulsifier has an influence on the resulting emulsion rheology, and therefore, by inference, on the flocculation state of the droplets.
FIG. 4. Influence of temperature on the small-deformation complex modulus G* of emulsion (45 vol% oil, 5 wt% protein, pH 7) made from pure as1-casein.
86
CASANOVA AND DICKINSON
FIG. 5. Controlled stress viscometry of emulsion (45 vol% oil, 5 wt% protein, pH 7) containing 50:50 mixture of as1-casein 1 b-casein. Apparent shear viscosity is plotted against shear stress at different temperatures: ■, 5°C; h, 10°C; F, 20°C; E, 30°C; Œ, 40°C.
DISCUSSION
The foregoing results show, first, that the temperature dependence of the shear rheology of a concentrated as1-caseinstabilized emulsion containing excess protein is substantially different from that of the equivalent b-casein-stabilized emulsion, and, second, that the rheology of the corresponding binary casein-stabilized emulsions is sensitive to the as1-casein/b-casein ratio, especially at emulsifier compositions rich in as1-casein. The explanation for the behavior is presumably something to do with differences in the self-assembly properties and the protein–protein interactions for these two caseins (12–14) and also with changes in distribution of the caseins between the interface and the bulk aqueous phase as a function of overall system composition (20). The pseudoplastic behavior of the equivalent sodium caseinate-stabilized emulsions was previously attributed (6 – 8) to depletion flocculation by nonadsorbed casein submicelles. As it is known that the size and shape of casein aggregates are also sensitive to casein composition and that the depletion interaction is itself sensitive to such factors, it is certainly not surprising that the rheology of flocculated casein emulsions is influenced by the protein composition. Before we discuss possible physicochemical origins of the observed rheological behavior, it is appropriate to mention two aspects of the systems studied that could have some bearing on certain aspects of the experimental data: (i) “wall slip” effects and (ii) creaming effects. It is established (22) that flocculated suspensions and emulsions are susceptible to the phenomenon of wall slip, which involves depletion of dispersed phase from the vicinity of a solid boundary and hence the formation of a thin lubricating layer near the wall of the rheometer cell.
Evidence for some wall slip in caseinate-based emulsions at low shear stresses (;0.5 Pa) has recently been found (23) in comparative experiments carried out with a specially made rheometer cell made from roughened glass surfaces replacing the standard stainless steel cell supplied by the manufacturer. The presence of wall slip implies that the absolute viscosity values recorded for the flocculated casein emulsions in this paper are probably underestimated somewhat at the lowest stresses (,1 Pa). The more subtle changes in the very low stress rheology noted above could reasonably be attributable in part to this phenomenon. However, data at higher stresses ($1 Pa) should be unaffected, and certainly the qualitative trends in shape and position of the non-Newtonian flow curves as a function of temperature and casein composition should not be affected by wall slip effects. Turning now to the possible influence of enhanced creaming of depletion flocculated emulsions on the measured rheological data, we have found no indication of any such problem in our experiments. The strongest evidence for this view comes from the good reproducibility and reversibility of the reported non-Newtonian flow curves. Within the normal experimental error of the instrument, exactly the same results were obtained when measurements were immediately repeated in the same rheometer cell without emulsion mixing. If emulsion creaming during the rheological measurements were to have been significantly affecting the results, such repeatability would not be expected. It is highly likely that the main reason for the complex rheological behavior reported here lies in the different selfassembly behavior and emulsion stabilizing properties of the two pure caseins. The highly amphiphilic b-casein is known to form surfactant-type micelles under the thermodynamic driving force of strong hydrophobic interactions in a reversible
FIG. 6. Controlled stress viscometry of emulsion (45 vol% oil, 5 wt% protein, pH 7) containing 75:25 mixture of as1-casein 1 b-casein. Apparent shear viscosity is plotted against shear stress at different temperatures: ■, 5°C; h, 10°C; F, 20°C; E, 30°C; Œ, 40°C.
RHEOLOGY OF OIL-IN-WATER EMULSIONS
FIG. 7. Controlled stress viscometry of emulsion (45 vol% oil, 5 wt% protein, pH 7) containing 90:10 mixture of as1-casein 1 b-casein. Apparent shear viscosity is plotted against shear stress at different temperatures: ■, 5°C; h, 10°C; F, 20°C; E, 30°C; Œ, 40°C.
process of endothermic self-assembly (24). The critical micelle concentration is quite low (,0.1 wt% at 20°C) (12). So, at room temperature, we can assume that essentially all of the unadsorbed protein present in a b-casein-stabilized emulsion is in the form of micelles containing about 40 monomers (25). But, since b-casein micelles are apparently completely dissociated below 5°C (26), if their presence is an important factor controlling the state of flocculation of b-casein-stabilized emulsions, we might have expected considerable differences in emulsion rheology between 5 and 20°C. However, this is not quite what occurs experimentally (see Fig. 1). While the rheology of the pure b-casein emulsion is certainly dependent on temperature, there is clearly less change over the range 5–15°C than over the range 20 – 40°C, and the greatest sensitivity of viscosity to temperature occurs not around 5°C, but at 20 – 25°C. The latter effect could still be associated with changes in casein self-assembly, however, because it has been suggested (25) that the packing density of b-casein micelles increases with temperature (while hydrodynamic size remains essentially constant) and that b-casein micelles may themselves aggregate as the temperature approaches 40°C. The more highly charged as1-casein associates in aqueous solution through a series of consecutive association steps (14). Light-scattering measurements have suggested (27) a cylindrical “wormlike” micelle structure as the most probable structure of as1-casein aggregates at 35°C over the concentration range 0.05– 0.6 wt%. The aggregate molecular weight (corresponding to ca. 120 monomers) is apparently relatively insensitive to temperature. It is not easy to reconcile this dilute solution structural information with the temperature-dependent rheology of pure as1-casein emulsions in Fig. 3. While it is not unreasonable to speculate that long
87
wormlike rods of aggregated as1-casein might be very effective at inducing depletion flocculation, it remains unclear why this flocculation should only be manifest in the rheology at 35–40°C (see Figs. 3 and 4). The emulsions prepared with mixtures of as1-casein 1 b-casein also exhibit trends in the rheology that, while entirely systematic and consistent, are difficult to summarize simply— and even more difficult to explain. Figure 10 presents a simplified overview of the data in terms of apparent viscosities for the different temperatures and compositions at two fixed stress values: (a) 1 Pa and (b) 0.1 Pa. The latter value is chosen because it is lowest stress studied. The value of 1 Pa is chosen because it lies just below the critical “yield” stress, which is a characteristic feature of several of the viscosity–stress plots reported here, being also found in the rheology of the equivalent sodium caseinate system (8). We note from Fig. 10a that there is a substantial viscosity value at 1 Pa for emulsions containing from 50 to 95% as1-casein and that only in the pure (or nearly pure) as1-casein emulsions at ,40°C, or the pure b-casein emulsions at .30°C, is this not the case. That is, assuming that a moderately high apparent viscosity at 1 Pa (say, .10 Pa s) can be regarded as an appropriate criterion for a flocculated network structure of finite strength, it appears that the creaming and rheological behavior observed previously for sodium caseinate emulsions can be reproduced also over a wide range of casein compositions well away from the natural composition of bovine milk. It is apparent that the temperature dependence of the rheology is very sensitive to casein emulsifier composition. For instance, Fig. 10b shows that replacing most (but not all) of the b-casein protein emulsifier by as1casein leads to a complete reversal of the temperature dependence of the viscosity at 0.1 Pa from strongly positive (system G) to strongly negative (system C). While we should be careful
FIG. 8. Controlled stress viscometry of emulsion (45 vol% oil, 5 wt% protein, pH 7) containing 95:5 mixture of as1-casein 1 b-casein. Apparent shear viscosity is plotted against shear stress at different temperatures: ■, 0°C; h, 10°C; F, 20°C; E, 30°C; Œ, 40°C.
88
CASANOVA AND DICKINSON
FIG. 9. Controlled stress viscometry of emulsion (45 vol% oil, 5 wt% protein, pH 7) containing 98:2 mixture of as1-casein 1 b-casein. Apparent shear viscosity is plotted against shear stress at different temperatures: ■, 0°C; h, 10°C; F, 20°C; E, 30°C; Œ, 40°C.
about overinterpreting trends of viscosity behavior obtained at such low stresses because of the possible role of wall slip effects (as discussed earlier), such changes may be explicable in terms of a change in the self-assembled protein structure from predominantly spherical b-casein-rich micelles to predominantly rodlike as1-casein micelles. Additionally, there is definite evidence in the literature (28 –31) for specific as1casein–b-casein complexes of high stoichiometric ratio in dilute solution. This could mean that the mixed aggregated system resembles neither the pure as1-casein associated structure nor the pure b-casein micellar structure. Individual molecules of as1-casein and b-casein can be expected to interact attractively through a combination of hydrophobic forces, hydrogen bonding, and electrostatic interactions (the latter, say, through interaction of the negatively charged phosphoserinerich tail of b-casein with the cluster of positively charged residues at the extreme N-terminus of as1-casein). Overall, however, we know too little about molecular association in concentrated solutions of mixed caseins to speculate much at this stage about implications of complexing for the emulsion properties. The observed tendency for b-casein to predominate at the surface of the emulsion droplets (see Table 1) is consistent with our earlier work (20) on competitive adsorption of as1-casein 1 b-casein in emulsions of lower oil content. In the concentrated emulsion systems studied here containing $90% as1-casein, all of the b-casein present is adsorbed at the oil–water interface. This means that any differences between the pure as1-casein system and systems containing a small proportion of b-casein (i.e. systems B, C, and D) must surely be attributable mainly to changes in adsorbed layer composition. Unfortunately, as little is currently
known about the structure of mixed layers of as1-casein 1 b-casein, and even less about the effect of casein composition on interlayer interactions, it seems inadvisable to speculate further here on the effect of casein competitive adsorption on the emulsion rheology. Finally, though, it is appropriate to comment on the influence of casein emulsifier composition on average droplet size and surface coverage in these concentrated emulsions. Compared with the pure emulsion systems, the mixed casein emulsions are characterized by larger droplet sizes and much higher surface coverages, especially for systems toward the extreme as1-casein-rich end of the composition range (systems B and C in Fig. 1). A reduced emulsifying activity index for binary mixtures of as1-casein 1 b-casein as compared with the pure proteins was also reported by Cayot et al. (32). The larger droplet sizes for the mixed casein systems in Fig. 1 could be connected in some way with the large increase in emulsion viscosity as small amounts of as1-casein are replaced by b-casein (Fig. 10). However, this is not really a convincing explanation because
FIG. 10. Overview of effects of temperature and emulsifier composition on the viscosity of emulsions (45 vol% oil, 5 wt% protein, pH 7) made with mixtures of as1-casein 1 b-casein. Values of apparent shear viscosity at constant stresses of (a) 1 Pa and (b) 0.1 Pa are plotted for the different as1/b ratios: A 5 100:0; B 5 98:2; C 5 95:5; D 5 90:10; E 5 75:25; F 5 50:50; G 5 0:100. Symbol key in plot (b): F, 10°C; D, 20°C; Œ, 30°C; E, 40°C.
RHEOLOGY OF OIL-IN-WATER EMULSIONS
the shear stresses experienced during homogenization are very much higher than those explored in the viscometric study. More plausible, perhaps, is an explanation based on the formation of an as1-casein–b-casein complex. This as1casein–b-casein association might reduce the effectiveness of the protein emulsifier during emulsification in its role of rapidly adsorbing at the oil–water interface to form a protective layer around the newly formed droplets (32). The low protein surface concentration found in the pure casein emulsions is consistent with stabilization by a monolayer of adsorbed monomeric protein. The considerably higher surface coverage in the binary casein systems is consistent with multilayer formation around the surface of the mixed protein-coated droplets due to complexation and enhanced protein aggregation at the interface. ACKNOWLEDGMENT H.C. acknowledges financial support from COLCIENCIAS (Colombia).
REFERENCES 1. Morr, C. V., in “Developments in Dairy Chemistry” (P. F. Fox, Ed.), Vol. 1, p. 375. Applied Science, London, 1982. 2. Doxastakis, G., in “Food Emulsifiers” (G. Charalambous and G. Doxastaskis, Eds.), p. 9. Elsevier, Amsterdam, 1989. 3. Dickinson, E., J. Chem. Soc., Faraday Trans. 88, 2973 (1992). 4. Dalgleish, D. G., Trends Food Sci. Technol. 8, 1 (1997). 5. Dickinson, E., J. Dairy Sci. 80, 2607 (1997). 6. Dickinson, E., and Golding, M., Food Hydrocolloids 11, 13 (1997). 7. Dickinson, E., Golding, M., and Povey, M. J. W., J. Colloid Interface Sci. 185, 515 (1997). 8. Dickinson, E., and Golding, M., J. Colloid Interface Sci. 191, 166 (1997). 9. Swaisgood, H. E., in “Developments in Dairy Chemistry—1” (P. F. Fox, Ed.), Vol. 1, p. 1. Applied Science, London, 1982. 10. Dickinson, E., J. Dairy Res. 56, 471 (1989).
89
11. Dickinson, E., Whyman, R. H., and Dalgleish, D. G., in “Food Emulsions and Foams” (E. Dickinson, Ed.), p. 40. Royal Society of Chemistry, London, 1987. 12. Schmidt, D. G., and Payens, T. A. J., in “Surface and Colloid Science” (E. Matijevic, Ed.), Vol. 9, p. 165. Wiley, New York, 1976. 13. Leclerc, E., and Calmettes, P., Phys. Rev. Lett. 78, 150 (1997). 14. Schmidt, D. G., in “Developments in Dairy Chemistry—1” (P. F. Fox, Ed.), Vol. 1, p. 61. Applied Science, London, 1982. 15. Dickinson, E., Horne, D. S., Pinfield, V. J., and Leermakers, F. A. M., J. Chem. Soc., Faraday Trans. 93, 425 (1997). 16. Casanova, H., and Dickinson, E., J. Agric. Food Chem. 46, 72 (1998). 17. Dickinson, E., Pinfield, V. J., Horne, D. S., and Leermakers, F. A. M., J. Chem. Soc., Faraday Trans. 93, 1785 (1997). 18. Dickinson, E., and Stainsby, G., in “Colloids in Food,” Chapter 7. Applied Science, London, 1982. 19. Pal, R., Colloids Surf. 64, 207 (1992). 20. Dickinson, E., Rolfe, S. E., and Dalgleish, D. G., Food Hydrocolloids 2, 397 (1988). 21. Burgaud, I., Dickinson, E., and Nelson, P. V., Int. J. Food Sci. Technol. 25, 39 (1990). 22. Barnes, H. A., J. Non-Newtonian Fluid Mech. 56, 221 (1995). 23. Dickinson, E., Golding, M., and Casanova, H., in “Food Emulsions and Foams: Interfaces, Interactions and Stability” (E. Dickinson and J. M. Rodriguez Patino, Eds.). Royal Society of Chemistry, Cambridge (in press). 24. Tanford, C., “The Hydrophobic Effect.” Wiley, London, 1973. 25. Thurn, A., Burchard, W., and Niki, R., Colloid Polym. Sci. 265, 653 (1987). 26. Payens, T. A. J., and van Markwijk, B. W., Biochim. Biophys. Acta 71, 517 (1963). 27. Thurn, A., Burchard, W., and Niki, R., Colloid Polym. Sci. 265, 897 (1987). 28. Payens, T. A. J., and Nijhuis, H., Biochim. Biophys. Acta 336, 201 (1974). 29. Yoshikawa, M., Sugimoto, E., and Chiba, H., Agric. Biol. Chem. 39, 1843 (1975). 30. Pepper, L., and Farrell, H. M., Jr., J. Dairy Sci. 65, 2259 (1982). 31. Ono, T., Odagiri, S., and Takagi, T., J. Dairy Res. 50, 37 (1983). 32. Cayot, P., Courthaudon, J.-L., and Lorient, D., J. Agric. Food Chem. 39, 1369 (1991).