Depletion Flocculation of Beverage Emulsions by Gum Arabic and Modified Starch R. CHANAMAI AND D.J. MCCLEMENTS

Transcription

1 JFS: Depletion Flocculation of Beverage Emulsions by Gum Arabic and Modified Starch R. CHANAMAI AND D.J. MCCLEMENTS ABSTRACT: The creaming velocity, apparent viscosity, and ultrasonic attenuation spectra (1 to 50 MHz) of 5 wt% n- hexadecane oil-in-water emulsions containing different droplet radii (r = m), biopolymer types (gum arabic or modified starch), and biopolymer concentrations (0 to 2.5 wt%) were measured. Depletion flocculation was observed in the emulsions when the nonabsorbed biopolymer concentration exceeded a critical concentration (CFC). The CFC increased with decreasing droplet radius for both biopolymers because the magnitude of the depletion attraction increases with droplet size. The CFC was lower for gum arabic than modified starch because it has a higher effective volume in solution. Depletion flocculation led to an increase in creaming instability and apparent viscosity of the emulsions. Flocculation could be nondestructively monitored by measuring the decrease in ultrasonic attenuation of the emulsions. These results show that depletion flocculation by gum arabic and modified starch can have an adverse effect on the stability of beverage emulsions. Key Words: gum arabic, modified starch, beverages, emulsions, flocculation Introduction BEVERAGE EMULSIONS ARE OIL-IN-WATER EMULSIONS THAT are normally prepared as a concentrate, that is, diluted into finished products (Tan 1997, 1998). The emulsion in both its concentrated and diluted form must have a high degree of stability. The oil phase usually consists of vegetable oil, flavor oil, and weighting agent, while the aqueous phase consists of water, sugar, emulsifier, acids, and preservatives (Tan 1997). Beverage emulsions are usually stabilized by amphiphilic polysaccharides, such as gum arabic or hydrophobically modified starch (Ray and others 1995; Kim and others 1996; McNamee and others 1998; Trubiano 1995; Garti 1999). Gum arabic is the most commonly used biopolymer emulsifier in flavor beverage emulsions (Tan 1997, 1998). It is derived from the natural exudate of Acacia senegal and consists of at least 3 high molecular weight biopolymer fractions. The surface-active fraction is believed to consist of branched arabinogalactan blocks attached to a polypeptide backbone (Anderson and others 1985; Randell and others 1988; Phillips and Williams 1995; Jayme and others 1999). The hydrophobic polypeptide chain is believed to anchor the molecules to the droplet surface, while the hydrophilic arabinogalactan blocks extend into the solution, providing stability against droplet aggregation through steric and electrostatic repulsion (Phillips and Williams 1995; Islam and others 1997; Jayme and others 1999). Gum arabic is an effective emulsifier because of its high water solubility, low solution viscosity, good surface activity, and ability to form a protective film around emulsion droplets (Glicksman 1983; Dickinson and others 1989). Problems associated with obtaining a reliable source of consistently high-quality gum arabic has led many food scientists to investigate alternative sources of biopolymer emulsifiers for use in flavor beverages (Kim and others 1996; Tan 1997, 1998; Garti 1999). Hydrophobically modified starches have been identified as 1 of the most promising replacements for gum arabic (Trubiano 1995). The modified starch used in this study (Purity Gum; National Starch, Bridgewater, N.J., U.S.A.) is an octenyl succinate derivative of waxy-maize. It consists primarily of amylopectin that has been chemically modified to contain nonpolar side-groups. These sidegroups anchor the molecule to the droplet surface, while the hydrophilic starch chains protrude into the aqueous phase and protect droplets against aggregation through steric repulsion. Purity Gum is mildly anionic in aqueous solutions and has a surface activity that is almost as high as gum arabic (Ray and others 1983; Tse 1990). Gum arabic and modified starch have relatively low surface activities (compared to proteins or surfactants), and so a large excess must be added to ensure that all the droplet surfaces are adequately coated (Phillips and Williams 1995). For example, as much as 20% gum arabic or 12% modified starch may be required to produce a stable 12.5 wt% oil-inwater emulsion (Tse and Reineccius 1995). As a result, there is a large excess of nonabsorbed polysaccharide in the aqueous phase of emulsions prepared from them (Tan 1998; Garti 1999). Under certain circumstances, nonabsorbed biopolymers are capable of promoting droplet flocculation through a depletion mechanism (Jenkins and Snowden 1996; Lips and others 1991; McClements 1999). Flocculation causes a number of effects that are detrimental to emulsion quality: (1) enhanced creaming due to the increase in particle size, (2) decreased cloudiness due to the increase in particle size, and (3) enhanced coalescence because droplets are brought into close proximity (Dickinson and Stainsby 1982; McClements 1999). The purpose of this study was to investigate the ability of gum arabic and modified starch to promote depletion flocculation in model beverage emulsions and to determine their effects on the creaming stability and rheology of emulsions. Materials and Methods Materials Experimental procedures. Polyoxyethylene sorbitan monolaurate (Tween 20), a non-ionic surfactant, and hexadecane were purchased from the Sigma Chemical Co. (St. Louis, Mo., U.S.A.). Modified starch (Purity Gum) was ob Institute of Food Technologists Vol. 66, No. 3, 2001 JOURNAL OF FOOD SCIENCE 457

2 tained from the National Starch and Chemical Co. (Bridgewater, N.J., U.S.A.). The average molecular weight of the Purity Gum was about daltons, with a fairly broad distribution. Gum arabic was obtained from Importers Service Corp. (Jersey City, N.J., U.S.A.). The major fractions of gum arabic have been reported to have molecular weights of around 2.5 to daltons (Jayme and others 1999). Distilled and deionized water was used in the preparation of all solutions. Emulsion Preparation An aqueous surfactant solution was prepared by dispersing 2.5 wt% Tween 20 in water. A 5 wt% hexadecane oil-inwater emulsion was prepared by weighing 100 g of hexadecane and 1900 g of surfactant solution into a 2000 cm 3 plastic beaker and blending with a high-speed blender for 1 min (High Shear Homogenizing Container; Waring Laboratory, New Hartford, Conn., U.S.A.). The size of the emulsion droplets was then reduced further using a high-pressure valve homogenizer (Rannie 8.30R, Wilmington, Mass., U.S.A.). Emulsions containing droplets with different sizes were obtained by withdrawing samples at different stages during the preparation procedure. The influence of biopolymer concentration on the measurements was investigated by preparing a series of emulsions with the same droplet size distribution and concentration and then adding different amounts of biopolymer and/or water to the aqueous phase to keep the overall droplet concentration the same in each emulsion. Particle Size Determination by Light Scattering The particle size distribution of the emulsions was measured using a laser light scattering instrument (Horiba LA- 900, Irvine, Calif., U.S.A.). This instrument measures the angular dependence of the intensity of light scattered from a dilute emulsion. It then finds the particle size distribution that gives the best fit between the experimental measurements and predictions made using light scattering theory. A refractive index ratio of 1.08 was used by the instrument to calculate the particle size distributions. Measurements are reported as the surface-volume mean radius: r 32 = n i r i 3/ n i r i2, where n i is the number of droplets of radius r i. To prevent multiple scattering effects, the emulsions were diluted with distilled water prior to analysis so that the droplet concentration was less than about 0.02 wt%. Each sample was analyzed 3 times, and the data are presented as the average. The droplet size distribution did not change during the course of the experiments, which suggests that the emulsions were stable to coalescence and Ostwald ripening. Coulter Corp., Miami, Fla., U.S.A.). This instrument measures the back-scattering of monochromatic light ( = 800 m) as a function of sample height. Emulsions were placed into flatbottomed, cylindrical glass tubes (100 mm height, 16 mm internal dia) and stored at room temperature. The back-scattering of light from the emulsions with height was then measured. The extent of creaming was assessed by determining the height (H) of the interface between the opaque, dropletrich layer at the top of emulsion and the less opaque, droplet-depleted layer at the bottom as a function of time (t). The results are reported either as the full creaming profiles or as the initial creaming rate: dh/dt. Rheology Measurements The rheological properties of emulsions were measured using a dynamic shear rheometer with a concentric cylinder measurement cell (Constant Stress Rheometer, CS-10; Bohlin Instruments, Cranbury, N.J., U.S.A.). The dia of the rotating inner cylinder was 25 mm, and the dia of the static outer cylinder was 27.5 mm. Samples were placed in the temperaturecontrolled measurement vessel and allowed to equilibrate to the required temperature (25 C) for 5 min prior to making the measurements. The shear rheology of the samples was determined by preshearing them at a constant shear rate of 30 s -1 for 30 s, allowing a recovery period of 3 min, and then acquiring the apparent viscosity as a function of shear stress (0 to 3 Pa). Ultrasonic measurements Ultrasonic attenuation spectra of emulsions were measured in the frequency range of 1 to 50 MHz using a custombuilt ultrasonic spectrometer. This spectrometer was based on the frequency scanning ultrasonic pulse echo reflectometer described in detail elsewhere (McClements and Fairley Optical Microscopy Photomicrographs of the emulsions were obtained using a Nomarski Differential Interference Contrast optical microscope (DIC, Nikon microscope Eclipse E600; Nikon Corp., Tokyo, Japan). The 5 wt% hexadecane oil-in-water emulsions were gently agitated in a glass test tube prior to analysis to ensure they were homogeneous. A drop of emulsion was then placed on a microscope slide, covered by a cover-slip, and observed at a magnification of 400. An image of the emulsion was acquired using digital image processing software (Spot Dianostic instruments Inc., Stering Heights, Mich., U.S.A.) and stored on a personal computer. Creaming Stability Measurements The creaming stability of the emulsions was determined by a commercial optical scanning instrument (Quickscan; 458 JOURNAL OF FOOD SCIENCE Vol. 66, No. 3, 2001 Figure 1 Photomicrographs of 5 wt% n-hexadecane oilin-water emulsions (r = 0.52 m): (a) Nonflocculated emulsion (0% gum arabic), (b) Nonflocculated emulsion (0.2 % gum arabic), (c) Flocculated emulsion (0.8% gum arabic), and (d) Flocculated emulsion (2% gum arabic)

3 1991, 1992). Briefly, a broadband pulse of ultrasound was propagated through an emulsion, and the attenuation coefficient was calculated from the decrease in its amplitude (Mc- Clements and Fairley 1992). The frequency-dependence of the attenuation coefficient was determined by carrying out a Fourier Transform Analysis of the ultrasonic pulse before and after it had traveled through the emulsion. To cover the whole frequency range, 2 different measurement cells and 4 different broadband ultrasonic transducers were used. The 1st measurement cell was used for low-frequency measurements (1 to 7 MHz). It had a 5-mm Plexiglas buffer rod (that separated the transducer from the sample) and a 16-mm sample path-length. The 2nd measurement cell was used for high-frequency measurements (10 B 50 MHz). It had a 3-cm quartz delay line and a 1.4-mm sample path-length. One transducer was used with the 1st measurement cell: 3.5 MHz, 0.5-inch dia (V682; Panametrics, Waltham, Mass., U.S.A.). Three transducers were used with the 2nd measurement cell: (a) 20 MHz, 0.25 inch (V212BA; Panametrics); (b) 50 MHz, 0.25 inch (V214BA; Panametrics); and (c) 100 MHz, inch (V2054; Panametrics). attractive van der Waals forces and therefore prevent the droplets from flocculating. The addition of a nonadsorbing polymer to the continuous phase of an emulsion increases the attraction between the droplets because of an osmotic effect associated with the exclusion of polymer molecules from a narrow region surrounding the droplets (Jenkins and Snowden 1996). As the polymer concentration is increased, the attractive forces between the droplets increases. Above a critical polymer concentration, the attractive forces dominate the repulsive forces, and so the droplets flocculate. It is therefore possible to control the degree of droplet flocculation in an emulsion by varying the concentration of polymer in the continuous phase. The tendency for emulsions to flocculate upon the addition of biopolymer was initially studied by optical microscopy. Photomicrographs of emulsions (r = 0.52 m) containing different concentrations of biopolymer (gum arabic) in the Results and Discussion Depletion Flocculation in Beverage Emulsion The degree of flocculation in an emulsion depends on the balance of attractive and repulsive interactions between the droplets, for example, van der Waals, electrostatic, steric, depletion, hydrophobic, and hydration (Hunter 1986, 1989; Israelachvili 1992). The emulsions used in this work are stabilized by a non-ionic surfactant (Tween 20), and so the 2 major types of repulsive interaction are steric and hydration forces. In the absence of gum arabic or modified starch, these repulsive forces are sufficiently large to overcome the Figure 2 Dependence of viscosity of gum-arabic and modified-starch solutions on biopolymer concentration Figure 3 Creaming profiles of 5 wt% n-hexadecane oilin-water emulsions containing gum arabic: (a) Nonflocculated (0.2 wt% gum arabic) and (b) Flocculated (0.8 wt% gum arabic) Vol. 66, No. 3, 2001 JOURNAL OF FOOD SCIENCE 459

4 aqueous phase are shown in Figure 1. When the biopolymer concentration was below a certain value, defined as the critical flocculation concentration or CFC, the droplets appeared isolated and were fairly evenly distributed throughout the emulsion (Figure 1a and 1b). The CFC was found to be about 0.4 wt% for gum arabic and about 1.7 wt% for modified starch, which was attributed to differences in their effective volumes in aqueous solutions. This is demonstrated by measurements of the concentration dependence of the viscosity of the 2 biopolymer solutions (Figure 2). The concentration increment of viscosity for gum arabic (about 5.5 mpa s wt% - 1 ) was 2- to 3-fold greater than that for modified starch (about 2 mpa s wt% -1 ), which suggests that its effective volume is about 2 to 3 times larger (McClements 2000). When the biopolymer concentration exceeded the CFC, the droplets formed flocs (Figure 1c), which became more extensive when the biopolymer concentration was increased further (Figure 1d). It should be noted that when the emulsions were left on the microscope slide, the floc size increased over time, which was probably caused by droplet-droplet, droplet-floc, and floc-floc collisions induced by Brownian motion. Influence of Depletion Flocculation on Creaming Stability of Emulsions The creaming stability of nonflocculated and flocculated 5% n-hexadecane emulsions with the same droplet radius (r = 0.52 m) was monitored by measuring the back-scattering of laser light as a function of height over 23 h (Figure 3). For both of the emulsions, the back-scattering of light was fairly constant along their entire height at the beginning of the experiment because there was an even distribution of droplets throughout the system. Over time the droplets moved upwards due to gravity, which caused a decrease in the back-scattering at the bottom of the emulsions (because the droplet concentration decreased) and an increase at the top (because the droplet concentration increased). The creaming behavior of nonflocculated and flocculated emulsions was clearly different (Figure 3a and 3b). Creaming was much more rapid in the flocculated emulsion than in the nonflocculated emulsion, as would be expected because of the increase in the size of the particles in the system (McClements 1999). The influence of hydrocolloid type, hydrocolloid concentration (c), and mean droplet size on the creaming rate of 5 wt% n-hexadecane oil-in-water emulsions was measured (Figure 4). Below a certain biopolymer concentration (the CFC), the creaming rate remained relatively constant, being close to that of a nonflocculated emulsion. Once the CFC was exceeded, there was a rapid increase in creaming rate Figure 4 Dependence of creaming rates of 5 wt% n- hexadecane oil-in-water emulsions with different droplet radii (0.17 and 0.52 m) on biopolymer concentration: (a) gum arabic and (b) modified starch. Figure 5 Dependence of critical flocculation concentration (CFC) on droplet size for 5 wt% n-hexadecane oil-inwater emulsions containing gum arabic and modified starch 460 JOURNAL OF FOOD SCIENCE Vol. 66, No. 3, 2001

5 because of the increase in the size of the particles within the emulsion (McClements 2000). A further increase in biopolymer concentration caused an appreciable decrease in creaming rate (up to 60%) because the increase in aqueous phase viscosity caused the upward movement of the flocs to be retarded. As would be expected from Stoke s law, the creaming rate of the nonflocculated emulsions (c CFC) increased as the droplet radius increased (Figure 4). The CFC of the emulsions also depended on droplet size, decreasing as the droplet radius increased. This was seen most clearly in measurements of the minimum amount of biopolymer required to cause extensive creaming after 24-h storage (Figure 5). The decrease in CFC with increasing droplet radius is because the strength of the depletion attraction increases with droplet size, therefore less biopolymer is required to induce flocculation for larger droplets (Jenkins and Snowdon 1996). These results indicate that decreasing the droplet size can reduce the susceptibility of an emulsion to depletion flocculation. For emulsions with the same droplet size and biopolymer concentration, the CFC was appreciably less for gum arabic than for modified starch. As mentioned earlier, this is because gum arabic has a greater aqueous phase effective volume than modified starch (Figure 2), either because its molecular weight is higher or because its structure is more open (McClements 2000). This means that emulsions containing gum arabic are more susceptible to depletion flocculation, especially when one considers that they are normally used at considerably higher concentrations (Tse and Reineccius 1995). The practical significance of the above results depends on the biopolymer concentrations found in actual flavor beverages. The concentration of gum arabic in a beverage concentrate is typically around 20 wt%, whereas that of modified starch is typically around 12 wt% (Tse and Reineccius 1995). Depletion flocculation is therefore highly likely in beverage concentrates, which may promote droplet coalescence on prolonged storage because the droplets are brought into close proximity. Beverage concentrates are normally diluted extensively (typically 300 to 2000 times) prior to use (Tan 1998). The susceptibility of the finished product to depletion flocculation depends on the final concentration of biopolymer in the aqueous phase. Dilutions of 300 to 2000 times would give gum arabic concentrations between about 0.01 to 0.07 wt% in the final product and modified starch concentrations between about and 0.04 wt%. Diluted emulsions stabilized by modified starch are therefore unlikely to undergo depletion flocculation because the biopolymer concentration is always below the CFC. On the other hand, diluted emulsions stabilized by gum arabic may be susceptible to depletion flocculation if the droplet size is fairly large because then the CFC is relatively low (Figure 5). Influence of Depletion Flocculation on Rheology of Beverage Emulsions It is well known that the rheology of emulsions is strongly dependent on droplet flocculation (Hunter 1986, 1989). For this reason, we examined the influence of depletion flocculation on the rheology of the model beverage emulsions. The dependence of the apparent viscosity of flocculated and nonflocculated emulsions (r = 0.52 m) with different biopolymer (gum arabic) concentrations on shear stress was measured (Figure 6). Flocculated emulsions exhibited pronounced shear-thinning behavior over the shear stresses studied, that is, their shear viscosity decreased with increasing shear stress. Shear-thinning is the result of progressive Figure 6 Shear-stress dependence of the apparent viscosity of 5 wt% n-hexadecane oil-in-water emulsions containing different gum-arabic concentrations (shown in text box) Figure 7 Dependence of normalized apparent viscosity (at 0.1 Pa) of 5 wt% n-hexadecane oil-in-water emulsions with different droplet radii (0.17 and 0.52 m) on biopolymer concentration: (a) gum arabic and (b) modified starch Vol. 66, No. 3, 2001 JOURNAL OF FOOD SCIENCE 461

6 deformation and disruption of flocs in the shear field as the shear stress is increased (McClements 1999). The nonflocculated emulsions exhibited Newtonian behavior, that is, their viscosity was independent of shear stress. The influence of biopolymer concentration on the apparent viscosity (at 0.1 Pa) of emulsions was also measured (Figure 7). The apparent viscosity of the emulsions ( ) was divided by the apparent viscosity of an aqueous solution with the same biopolymer concentration ( 1 ) to obtain a normalized viscosity ( N = / 1 ) that highlights changes resulting from alterations in droplet characteristics. The normalized viscosity was relatively constant when the biopolymer concentration was below the CFC but increased significantly when the CFC was exceeded. This increase in viscosity is because of the increase in the effective volume fraction of the particles in the emulsion when flocculation occurs (McClements 2000). These changes are much more obvious for the gumarabic emulsions because flocculation occurs at appreciably lower biopolymer concentrations. Influence of Depletion Flocculation on Ultrasonic Properties of Emulsions The creaming and viscosity measurements indicate that depletion flocculation has a pronounced influence on the physiochemical properties of beverage emulsions. It would therefore be useful to have an analytical technique that could rapidly and nondestructively determine whether the droplets in an emulsion were flocculated or not. Light scattering or electrical pulse counting techniques cannot be used to study depletion flocculation because emulsion dilution usually causes disruption of the flocs. Optical microscopy is time consuming and laborious and cannot be used to quantify flocculation in emulsions containing small droplets. In this section, we show that ultrasonic spectroscopy provides a convenient method of measuring droplet flocculation in situ. The frequency dependence of the ultrasonic attenuation coefficient of 5 wt% n-hexadecane oil-in-water emulsions (r = 0.52 m) containing nonflocculated and flocculated droplets was measured (Figure 8). In the nonflocculated emulsion, there was a slight maximum in the /f spectra when the frequency was increased from about 2 to 10 MHz and a steep rise in /f at higher frequencies. The maximum in /f at lower frequencies is primarily due to thermal losses, while the steep increase at higher frequencies is due to intrinsic absorption losses associated with the oil and aqueous phases, as well as some scattering of the ultrasound by the droplets (McClements 1996). When the emulsions became flocculated, there was a decrease in the attenuation coefficient at low frequencies and an increase at high frequencies. The reason for the decrease at low frequencies is the fact that the flocculated droplets are closer together, and therefore their thermal waves overlap with each other (McClements 1994; Hemar and others 1997). The increased attenuation at high frequencies is due to increased scattering by the relatively large flocs (McClements and others 1998). The critical flocculation concentration can be determined by measuring the attenuation coefficient of emulsions containing different biopolymer concentrations. The dependence of the attenuation coefficient (at 2 MHz) on biopolymer concentration for 5 wt% n-hexadecane emulsions (r = 0.52 m) is shown in Figure 9. The attenuation coefficient is fairly constant up to the CFC of the emulsions (about 0.4 wt% for gum arabic and about 1.7 wt% for modified starch), indicating that the droplets are not aggregated. Above this biopolymer concentration, the attenuation coefficient falls rapidly, indicating that the droplets come into close proximity, that is, they are flocculated. Above a gum-arabic concentration of about 1 wt%, the attenuation coefficient remained fairly constant, which suggested that the structure of the flocs did not change much as the biopolymer concentration was increased further. These results show that ultrasonic spectroscopy can give useful information about the degree Figure 8 Ultrasonic attenuation spectra of 5 wt% n- hexadecane oil-in-water emulsion (r = 0.52 _m) containing nonflocculated (0.2 % gum arabic) and flocculated (0.8% gum arabic) droplets Figure 9 Ultrasonic attenuation coefficient (at 2 MHz) of 5 wt% n-hexadecane oil-in-water emulsions containing different gum-arabic and modified-starch concentrations 462 JOURNAL OF FOOD SCIENCE Vol. 66, No. 3, 2001

7 of flocculation in optically opaque emulsions. Conclusions THE CREAMING STABILITY AND RHEOLOGY OF MODEL BEVERage emulsions was strongly influenced by the concentration of free biopolymer in the aqueous phase. When the biopolymer concentration exceeded a certain concentration (CFC), the droplets became flocculated through a depletion mechanism. The CFC of gum arabic (about 0.4 wt% for r = 0.52 m) was significantly lower than that of modified starch (about 1.7 wt% for r = 0.52 m). We also found that the CFC decreased with increasing droplet size because the strength of the depletion attraction increases with droplet size. Knowledge of the dependence of the CFC on droplet size and biopolymer type will enable food manufacturers to enhance the stability of beverage emulsions by avoiding problems caused by depletion flocculation. References Anderson DMW, Howlett JF, McNab CGA The amino acid composition of the protienaceous on Acacia senegal gum. Carbohydr Res 2, Dickinson E, Everson DJ, Murray BS On the film-forming and emulsionstabilizing properties of gum arabic: Dilution and flocculation aspects. Food Hydrocoll 3, Dickinson E, Stainsby G Colloids in foods. London: Applied Science. Garti N Hydrocolloids as emulsifying agents for oil-in-water emulsions. J Disp Sci Tech 20, Glicksman, M. Gum arabic In: Glicksman M, editor. Food hydrocolloids. Boca Raton, Fla.: CRC Press. p Hemar Y, Herrmann N, Lemarechal P, Hocquart R, and Lequeux F Effect medium model for ultrasonic attenuation due to the thermo-elastic effect in concentrated emulsions, J. Phys. II. 7, Hunter RJ Foundations of colloid science: Vol 1. Oxford: Oxford Science. Hunter RJ Foundations of colloid science: Vol II. Oxford: Oxford Science. Islam AM, Phillips GO, Sljivo A, Snowden MJ, Williams PA A review of recent developments on the regulatory, structural and functional aspects of gum arabic. Food Hydrocoll 11, Israelachvili JN Intermolecular and surface forces. London: Academic Press. Jayme ML, Dunstan DE, Gee ML Zeta potentials of gum arabic stabilized oil in water emulsions. Food Hydrocoll 13, Jenkins P, Snowden M Depletion flocculation in colloidal dispersions. Adv Colloid Int Sci 68, Kim YD, Morr CV, Schenz TW Microencapsulation properties of gum arabic and several food proteins: Liquid oil emulsion particles. J Agric Food Chem 44, Lips A, Campbell IJ, Pelan EG Aggregation mechanisms in food colloids and the role of biopolymers. In: Dickinson E, editor. Food polymers, gels and colloids. Cambridge: Royal Society of Chemistry. p McClements DJ Food emulsions: Principles, practice and techniques. Boca Raton: CRC Press. McClements DJ Comments on viscosity enhancement and depletion flocculation by polysaccharides. Food Hydrocoll (Forthcoming). McClements DJ, Herrmann N, Hemar Y Influence of flocculation on the ultrasonic properties of emulsions: Theory. J Phys D 31, McClements DJ Principles of Ultrasonic Droplet Size Determination. Langmuir. 12, McClements DJ Yktrasibuc determination of depletion flocculation in oilin-water emulsions, Coll. Surf., 90, McClements DJ and Fairley P Frequency scanning ultrasonic pulse echo reflectometer. Ultrasonics, 30, McClements DJ and Fairley P Ultrasonic pulse echo reflectometer, Ultrasonics, 29, McNamee BF, O Riordan ED, O Sullivan M Emulsification and microencapsulation properties of gum arabic. J Agric Food Chem 46, Phillips GO, Williams PA Interaction of hydrocolloids in food systems. In: Gaonkar AG, editor. Ingredient interactions: Effect on food quality. New York: Marcel Dekker. p Randell, R.C.; Phillips, G.O.; Williams, P.A. The role of the proteinaceous component on the emulsifying properties of gum arabic. Food Hydrocolloids 1988, 2, Ray AK, Johnson JK, Sullivan RJ Refractive index of the dispersed phase in oil-in-water emulsions: Its dependence on droplet size and aging. J Food Sci 48, Ray AK, Bird PB, Iacobucci GA, Clark BC Functionality of gum arabic: Fractioation, characterization and evaluation of gum fractions in citrus oil emulsions and model beverages. Food Hydrocoll 9, Tan CT Beverage emulsions. In: Friberg SE, editor. Food emulsions. New York: Marcel Dekker. p Tan CT Beverage flavor emulsion B A form of emulsion liquid membrane encapsulation. In: Contis ET, Ho CT, Mussinan CJ, Parliament TH, Spanier AM, editors. Food flavors: Formation, analysis and packaging influences. New York: Elsevier. p Trubiano PC The role of specialty food starches in flavor emulsions. In: Ho CT, Tan CT, Tong CH, editors. Flavor technology: Physical chemistry, modification, and process. Washington, D.C.: American Chemical Society. p Tse KY Physical stability of flavor emulsions [thesis]. University of Minnesota. Tse KY, Reineccius GA Methods to predict the physical stability of flavorcloud emulsion. In: Ho CT, Tan CT, Tong CH, editors. Flavor technology: Physical chemistry, modification, and process. Washington, D.C.: American Chemical Society. p MS This material is partly based upon work supported by the Cooperative State Research, Education and Extension Service, U.S. Dept. of Agriculture, under Agreement Number We also thank Coulter Corp. (Miami, Fla., U.S.A.) for providing the Quickscan instrument used in these experiments. Authors are with the Biopolymers and Colloids Research Laboratory, Dept. of Food Science, Univ. of Massachusetts, Amherst, MA Direct correspondence to D. Julian McClements ( mcclements@foodsci.umass.edu) Vol. 66, No. 3, 2001 JOURNAL OF FOOD SCIENCE 463