Changes in Physicochemical Properties of Retort-Sterilized Dairy Beverages During Storage

M. E. CANO-RUIZ and R. L. RICHTER

Texas A&M University, Department of Animal Science, College Station 77843-2471

ABSTRACT The effects of composition, storage time, and storage temperature on the physicochemical properties of a retort-sterilized dairy beverage were investigated. Drinks with eight formulations were stored at 4, 25, and 37C for 6 mo and were analyzed monthly for pH, net color difference, apparent viscosity, sedimentation index, homogenization index, particle size index, and soluble calcium. The changes in the physicochemical properties of the beverages increased as storage time and temperature increased. The degree of change was affected by the composition of the product. Sodium tripolyphosphate was implicated in promoting age gelation of samples with 11% nonfat milk solids, but sedimentation was observed in the absence of sodium tripolyphosphate and carrageenan. The apparent viscosity of samples affected the rate of age gelation and sedimentation, both of which increased as viscosity decreased. Interactions between milk fat, carrageenan, and nonfat milk solids were important in determining the apparent viscosity of the beverages and the rate of change observed during storage. ( Key words: dairy beverage, retort sterilization, storage effects, age gelation) Abbreviation key: DE = net color difference, HI = homogenization index, MSNF = milk SNF, PI = particle size index, SI = sedimentation index. INTRODUCTION In the US, processed beverages that claim nutritional benefits are becoming increasingly popular because of their convenience. These beverages are usually marketed as healthy drinks that provide energy and fitness through their contents of protein, vitamins, and minerals. Milk is usually the base ingredient, but sugar, emulsifiers, stabilizers, additives, and flavors are added to improve the stability and

acceptance of the product. It is desirable to have low fat beverages that provide energy but reduce fat intake to satisfy the nutritional concerns of consumers. An understanding of the functional behavior of dairy ingredients in food systems is critical for evaluating the effects that changes in concentration or substitution of dairy ingredients can have on existing or new food products (10). Harper ( 7 ) stated the importance of using model food systems to formulate new products; he used a model system to test the functionality of whey protein concentrates because standard functionality tests are not useful for predicting the behavior of a protein ingredient in a formulated food. In his approach, different formulations are tested to evaluate the effects of ingredients, processing conditions, interactions between ingredients, and interactions between ingredients and processing steps because of the complexity of food systems. The present work resulted from the observation that lowering the fat content in a retort-sterilized dairy beverage in an attempt to produce a low fat product reduced the storage stability of the product as a result of an age gelation phenomenon. Our objective was to study the effects of ingredient, storage temperature, and storage time on some physicochemical properties of a retort-sterilized, dairy-based beverage. MATERIALS AND METHODS Ingredients Condensed skim milk ( 34% TS) and fresh cream ( 48% TS, 43% milk fat) were used (Mid-America Dairymen, Inc., Springfield, MO). Commercial monoand diglycerides (HV52-K ; Grinsted Products, Brabrand, Denmark), carrageenan (SGP-3 ; Hercules Inc., Wilmington, DE), sodium tripolyphosphate (FMC Corp., Philadelphia, PA), and extra fine grade granulated sucrose (from local market) were used to formulate the samples. Processing

Received December 30, 1997. Accepted April 27, 1998. 1998 J Dairy Sci 81:21162123

Eight formulations of beverages (Table 1 ) were processed. Carrageenan was blended with dry sugar 2116

RETORT-STERILIZED DAIRY BEVERAGES TABLE 1. Dairy beverage formulations. Formulation (no.) 1 2 3 4 5 6 7 8

1Milk

2117

MSNF1 6 6 11 11 6 6 11 11

Milk fat 1 1 1 1 3 3 3 3

Emulsifier 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1

Polyphosphate (%) 0 0.1 0.1 0 0.1 0 0 0.1

Carrageenan 0 0.03 0 0.03 0 0.03 0 0.03

Sucrose 8 8 8 8 8 8 8 8

SNF.

before it was dispersed in cold water. Sodium tripolyphosphate was dispersed in warm water prior to its addition to the cold blend. Condensed skim milk and cream were mixed and heated to 75C. Emulsifier was then added and blended into the cream and milk. The blend of sucrose, carrageenan, and sodium tripolyphosphate was added to the milk and emulsifier mixture and was heated to 75C. The mix was homogenized at 20 MPa and standardized by adding water to achieve the desired content of total solids. The standardized mixture was adjusted to pH 6.9 to 7.0 using 10% NaOH. The samples were canned in 11-fl oz cans containing 325 ml of product and sterilized (128C for 17 min) using a rotatory retort. Experimental Design A fractional factorial experimental design with eight formulations (Table 1 ) was selected using JMP

statistical software ( 9 ) to evaluate the effects of milk fat, milk SNF ( MSNF) , sodium tripolyphosphate, and carrageenan on the storage stability of the dairy emulsion. The aliasing structure obtained for the fractional factorial experimental design is shown in the first column of Table 2. All of the main effects could be determined, but some two-factor interactions were confounded. Storage Experiment The cans from each of the eight formulations were stored at 4, 25, and 37C for 6 mo and analyzed monthly. Physicochemical Properties Samples used to determine the homogenization index ( HI) , particle size index ( PI) , and visual obser-

TABLE 2. Probability values from the ANOVA of the fractional factorial experimental design for experimental parameters. Model1 Milk fat ( M F ) MSNF Gum ( G ) Phos MF MSNF = G Phos MF G = MSNF Phos MF Phos = MSNF G Temperature of storage Time of storage Temperature of storage time of storage
1Milk

pH 0.0392 0.0041 0.0352 0.0158 0.0115 0.0087 0.0813 <0.0001 0.0195 <0.0001

DE2 0.5658 0.1218 0.9076 0.4967 0.7237 0.9391 0.8521 0.0207 0.6107 <0.0001

Apparent viscosity <0.0001 <0.0001 0.0911 0.3206 <0.0001 <0.0001 0.0001 0.0060 0.0170 0.0090

SI <0.0001 <0.0001 <0.0001 <0.0001 <0.0001 <0.0001 <0.0001 0.2229 0.2266 0.0011

HI 0.0642 0.2669 0.0495 0.9202 <0.0001 0.0206 0.0667 0.9653 0.0370 0.0772

PI <0.0001 <0.0001 <0.0001 <0.0001 <0.0001 <0.0001 <0.0001 0.7334 0.6396 0.5028

Gelation 0.0272 <0.0001 <0.0001 <0.0001 <0.0001 <0.0001 <0.0001 0.6729 0.0779 0.1949

Soluble calcium 0.6402 <0.0001 0.2878 0.0213 0.1145 0.8340 0.0637 0.5224 0.6229 0.2917

fat = MF content effect ( 1 or 3%), MSNF = milk SNF content effect ( 6 or 11%), gum = carrageenan content effect ( 0 or 0.03%), Phos = sodium tripolyphosphate content effect ( 0 or 0.1%), MF MSNF = G Phos = interaction of MF and MSNF effect confounded with the interaction of carrageenan and sodium tripolyphosphate effect, MF G = MSNF Phos = interaction of MF and carrageenan effect confounded with the interaction of MSNF and sodium tripolyphosphate effect, MF Phos = MSNF G = interaction of MF and sodium tripolyphosphate effect confounded with the interaction of MSNF and carrageenan effect, temperature of storage = temperature of storage effect (4, 25, or 37C), time of storage = time of storage effect ( 1 to 6 mo), and temperature of storage time of storage = interaction of temperature of storage and time of storage effect. 2DE = Net color difference, SI = sedimentation index, HI = homogenization index, and PI = particle size index. Journal of Dairy Science Vol. 81, No. 8, 1998

2118

CANO-RUIZ AND RICHTER

vation of gelation were left undisturbed prior to analysis; those used for other analyses were tempered to 21C and shaken before their physicochemical properties were measured. pH. The pH was measured using an electronic pH meter (Orion 210A; Orion Research Inc., Boston, MA). HI. The HI was determined by the method of Rustom et al. (16). Undisturbed samples were opened, 30 ml ( A ) were pipetted from the upper layer, and another 30 ml ( B ) were pipetted from the bottom layer. The percentage of milk fat in each layer was determined using the Roese-Gottlieb ether extraction method for milk ( 2 ) . The HI was calculated as the percentage of milk fat B divided by the percentage of milk fat A. Sedimentation. A procedure modified from the method of Rustom et al. ( 1 6 ) was used to determine the sedimentation index ( SI) . The samples were thoroughly mixed, and 25 ml were transferred to a flat-bottomed glass Babcock cream bottle. The initial height (millimeters) of the 25 ml was measured ( A ) . The samples were centrifuged at 325 g for 20 min, and the height of the sediment (millimeters) was measured ( B ) . The SI was expressed as SI = B/A. Color. A Hunterlabscan colorimeter (model 5100; Hunter Associates Laboratory, Inc., Reston, VA) was used to measure the color of the samples. The CIE LAB (Paris, France) color system was used. The colorimeter was calibrated against black and white standards. Net color difference ( DE) was calculated at each storage interval as DE = ( ( a 0 at) 2 + ( b 0 bt) 2 + ( L 0 Lt) 2) 1/2 where the subscripts 0 and t refer to the coordinate values at 0 and t months of storage. Particle size analysis. The particle size distribution of the samples was determined using a Coulter LS 130 light-scattering apparatus (Coulter Corporation, Miami, FL). Distilled water was used to dilute the samples. A real refractive index of 1.333 was used for the water. Refractive indices of 1.346 and 0.01, corresponding to the real and the imaginary refractive indices, were used for the milk samples. Undisturbed beverage cans were opened, 30 ml ( A ) were pipetted from the upper layer, and another 30 ml ( B ) were pipetted from the bottom layer. The mean volume-surface diameter (dvs; micrometers) was determined in A and B layers of each sample in duplicate every month. The change in size distribution was evaluated using PI = dvsB/dvsA. Apparent viscosity. The apparent viscosity of the samples was determined with a Brookfield RV DV-III
Journal of Dairy Science Vol. 81, No. 8, 1998

rheometer (Brookfield Engineering Laboratories Inc., Stoughton, MA) using a UL adapter with a UL spindle. The spindle speed was set at 60 rpm. Viscosity readings were recorded every 10 s at room temperature (25C ) during 5 min of shearing, and the apparent viscosity of the samples was reported as the mean of the readings at 10, 150, and 300 s. Onset of gelation. Age gelation of all samples was determined by an increase in the apparent viscosity of the samples and qualitatively by visual examination. The observed degree of gelation was assigned a number from 0 = no gelation to 5 = very strong gel. Soluble calcium. Soluble calcium was determined in the filtrate after ultrafiltration of the samples through a membrane with a molecular mass cutoff of 30,000 Da (Amicon, Inc., MA) according to Dalgleish and Law ( 4 ) . Approximately 30 ml of sample were circulated through the membrane. The filtrate was diluted 1/10 (vol/vol) and was analyzed for calcium with a sequential inductive coupled argon plasma emission spectrophotometer (model Plasma II; Perkin-Elmer Corp., Norwalk, CT). Statistical analysis. The ANOVA was performed using JMP software ( 9 ) to evaluate the effects of formulation, storage time, and storage temperature on the physicochemical properties of the product. The ANOVA for apparent viscosity was evaluated using only the data of apparent viscosity obtained prior to gelation. To determine which of the confounded twoway interactions (Table 2 ) was predominant after a difference ( P < 0.05) was found for the parameter being evaluated, the two-way interactions and the parameter were plotted. The two-way interactions that resulted in parallel lines (no interaction) were not significantly different, and the two-way interactions with no parallel lines (different effect at different test levels) were accepted as significantly different. When main effects and their two-way interactions were significantly different, the interactions rather than the main effects were used for interpretation and discussion of the data. RESULTS AND DISCUSSION Table 2 shows the probability values obtained from the ANOVA for the experimental parameters. The ANOVA indicated that the physicochemical properties of the dairy beverages were affected by their composition, storage temperature, and storage time. pH The pH of the beverages was affected by the interaction of milk fat and MSNF ( P < 0.0115), by the

RETORT-STERILIZED DAIRY BEVERAGES TABLE 3. Mean pH of dairy beverages with different levels of nonfat milk solids (MSNF) and sodium tripolyphosphate (Phos). Storage temperature ( C ) 4 25 37 4 25 37 pH 0% Phos 6.70 6.63 6.47 6.57 6.52 6.28 0.1% Phos 6.72 6.69 6.58 6.64 6.62 6.45

2119

MSNF 6% 11%

interaction of sodium tripolyphosphate and MSNF ( P < 0.0087), and by the interaction of storage temperature and storage time ( P < 0.0001) (Table 2). The decrease in pH was greater for the samples with 11% MSNF than for the samples with 6% MSNF, especially in the absence of sodium tripolyphosphate (Table 3). Sodium tripolyphosphate helped maintain the pH of the beverages during storage because this compound can function as a buffer. The pH decreased as storage time at 37C increased (Figure 1). The pH of samples stored at 4 and 25C for 6 mo was similar to the original pH. Similar changes in pH as affected by storage temperature and time were observed by Kocak and Zadow ( 1 1 ) in UHT milk and by Rustom et al. ( 1 6 ) in UHT-sterilized peanut beverages. Rustom et al. ( 1 6 ) attributed the decrease in pH to reactions among proteins that released protons during storage. DE Samples with 11% MSNF had a darker color ( L value = 76.28) than did samples with 6% MSNF ( L

Figure 2. Effect of interactions between temperature and time of storage on the net color difference ( DE ) of retort-sterilized dairy beverages.

value = 79.08). A significant ( P < 0.0001) change in color was caused by the interaction of storage temperature and storage time (Table 2). The DE increased as storage time and temperature increased, and the greatest changes were observed at 37C (Figure 2). The DE probably resulted from Maillard reactions caused by the reaction of lactose with proteins. This same trend for increased browning in samples during storage at increased storage temperature was observed by Kocak and Zadow (11). Apparent Viscosity and Age Gelation The apparent viscosity of samples was affected by the interaction of milk fat and MSNF ( P < 0.0001), by the interaction of milk fat and carrageenan ( P < 0.0001), by the interaction of MSNF and carrageenan, and by the interaction of storage temperature and storage time ( P < 0.0090) (Table 2). The apparent viscosity of the samples increased as MSNF increased, but the increase in viscosity was higher for the samples with 3% milk fat than for the samples with 1% milk fat (Figure 3). This observation indicated that the interaction of milk proteins and milk fat caused a significant increase in viscosity of the dairy beverage. A higher volume fraction of milk fat and milk solids would result in greater viscosity. Nevertheless, the difference in viscosity between the formulations with 1 and 3% milk fat was greater for samples with 11% MSNF than for the samples with 6% MSNF. The strength of milk gels can be influenced by emulsified fat when the fat can act as an active or inactive filler, which would increase or decrease the strength of the gel, depending upon the
Journal of Dairy Science Vol. 81, No. 8, 1998

Figure 1. Effect of interactions between temperature and time of storage on the pH of retort-sterilized dairy beverages.

2120

CANO-RUIZ AND RICHTER

Figure 3. Effect of interactions between milk SNF (MSNF) and milk fat content on the apparent viscosity of retort-sterilized dairy beverages.

composition of the fat globule membrane (1, 19, 20). The membrane of homogenized milk fat globules in this experiment was formed from the original material of the milk fat globule membrane, skim milk proteins, and mono- and diglycerides. Interactions between the homogenized milk fat globules and the milk proteins in the bulk phase could increase the viscosity of the beverage with higher fat content. The samples that contained carrageenan were more viscous than the samples without carrageenan, but the increase in apparent viscosity caused by carrageenan in the system was greater in the samples

Figure 5. Effect of storage time and storage temperature on the apparent viscosity of formulation 3 stored at 4C ( ) , 25C ( o) , and 37C ( ) and formulation 4 stored at 4C ( ) , 25C ( ) , and 37C ( ) . Formulation 3 gelled after 1 mo of storage, as indicated by the increase in apparent viscosity at all temperatures of storage.

Figure 4. Effect of storage time and storage temperature on the apparent viscosity of formulation 1 stored at 4C ( ) , 25C ( o) , and 37C ( ) and formulation 2 stored at 4C ( ) , 25C ( ) , and 37C ( ) . Journal of Dairy Science Vol. 81, No. 8, 1998

with 6% MSNF than in the samples with 11% MSNF and for the samples with 3% milk fat than for the samples with 1% milk fat (Figures 4 to 7). Apparent viscosity of the samples stored at 37C decreased as storage time increased (unless gelation occurred), increased in samples stored at 4C, and remained nearly constant for samples stored at 25C (Figures 4 to 7). These changes were more evident for formulations with 6% MSNF and carrageenan. The gelation or swelling of carrageenan in milk is caused by interactions between carrageenan and casein and by the association between carrageenan molecules that increases at lower temperatures (6, 15). The possible effect of storage temperature on the swelling properties of carrageenan seemed likely in samples with 6% MSNF (Figures 4 to 7). The interaction of MSNF and milk fat was more important in defining the viscosity of samples with 11% MSNF than the presence of carrageenan (Figures 3, 5, and 7), but carrageenan was more important for retarding sedimentation or creaming at the concentration of carrageenan used in this study. Milk fat in the system can affect the gelling properties of carrageenan. The interaction of carrageenan and k-CN is well known (18). Therefore, it is possible that the homogenized milk fat globules covered, at least in part, with casein proteins (3, 17) interacted with carrageenan, enabling carrageenan to keep the milk fat dispersed in the emulsion during storage. The interaction of sodium tripolyphosphate and MSNF ( P < 0.0001) (Table 2 ) was implicated in the

RETORT-STERILIZED DAIRY BEVERAGES

2121

Figure 6. Effect of storage time and storage temperature on the apparent viscosity of formulation 5 stored at 4C ( ) , 25C ( o) , and 37C ( ) and formulation 6 stored at 4C ( ) , 25C ( ) , and 37C ( ) .

0.1% resulted in only moderate increase in shelf-life of UHT milk, and the addition of sodium hexametaphosphate (crystalline) at 0.1% was able to control age gelation. Orthophosphates hasten the onset of gelation instead of delaying it (5, 8, 14), which indicates that cyclic polyphosphates are more effective for the control of age gelation than are linear polyphosphates because the cyclic polyphosphates are more resistant to hydrolysis (5, 8, 14). Sodium tripolyphosphate was added before sterilization in this study. Its stability might have been decreased during processing as degradation could have resulted in the conversion to orthophosphates that promoted age gelation in samples 3 and 8 (Table 4). Sedimentation rather than gelation was observed in samples 4 and 7 with 11% MSNF but without sodium tripolyphosphate (Table 4). This observation was more apparent in sample 7, which did not contain carrageenan. SI

age gelation of the samples with 11% MSNF. Observation of the samples to determine the onset of gelation indicated that only samples 3 and 8 gelled (Table 4). Sample 3 gelled after 1 mo of storage at all storage temperatures, and the strength of the gel increased as storage temperature and storage time increased. Sample 8 formed a gel after 5 mo of storage at 37C. Only the samples with 11% MSNF and 0.1% sodium tripolyphosphate gelled, suggesting a minimum concentration of protein necessary for gelation to occur, at least during the storage time evaluated. No gelation was observed in the absence of sodium tripolyphosphate. The interactions of carrageenan and sodium tripolyphosphate ( P < 0.0001) and the interaction of milk fat and sodium tripolyphosphate ( P < 0.0001) (Table 2 ) affected also the onset of gelation. The presence of carrageenan, which increased the viscosity, retarded the gelation in the dairy beverages with 11% MSNF and 0.01% sodium tripolyphosphate, but lowering the milk fat of the beverages from 3 to 1%, which decreased the viscosity, accelerated the gelation in the formulations with 11% MSNF and 0.1% sodium tripolyphosphate. Variations in the chemical composition and chain length of polyphosphates affect the ability of these additives to control age gelation in UHT sterilized milk, extending the protection against gelation with an increase in chain length and concentration of the polyphosphate (5, 12). Kocak and Zadow ( 1 3 ) found that the addition of sodium polyphosphate (amorphous) at 0.1% did not increase the shelf-life of UHT milk, sodium hexametaphosphate (amorphous) at

The samples did not exhibit sedimentation by the centrifugation method used until the samples showed signs of age gelation (data not shown). Nevertheless, some sedimentation without gelation was observed in samples when the cans were left undisturbed (Table 4). Sedimentation was only observed for samples with 11% MSNF and was more evident for samples without carrageenan than for samples with carrageenan. The sedimentation observed at the bottom of the can increased as storage temperature and storage time of these samples increased.

Figure 7. Effect of storage time and storage temperature on the apparent viscosity of formulation 7 stored at 4C ( ) , 25C ( o) , and 37C ( ) and formulation 8 stored at 4C ( ) , 25C ( ) , and 37C ( ) . Formulation 8 gelled after 5 mo of storage at 37C, as indicated by the increase in apparent viscosity. Journal of Dairy Science Vol. 81, No. 8, 1998

2122

CANO-RUIZ AND RICHTER TABLE 4. Degree of gelation obtained by visual observation for the dairy beverages during time at different storage temperatures.1 Formulation no. 1 2 3 4 5 6 7 8 Storage temperature ( C ) 4 25 37 4 25 37 4 25 37 4 25 37 4 25 37 4 25 37 4 25 37 4 25 37 Time 1 mo 0 0 0 0 0 0 1 2 3 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 2 mo 0 0 0 0 0 0 2 3 4 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 3 mo 0 0 0 0 0 0 2 3 4 0 0 0 0 0 0 0 0 0 0 0 1 0 0 2 4 mo 0 0 0 0 0 0 2 4 5 0 0 0 0 0 0 0 0 0 0 0 1 0 0 2 5 mo 0 0 0 0 0 0 3 4 5 1 1 1 0 0 0 0 0 0 0 1 1 1 1 3 6 mo 0 0 0 0 0 0 3 4 5 1 1 1 0 0 0 1 0 0 0 1 1 1 1 3

1Degree of gelation: 0 = no gel, 1 = increase in viscosity at the bottom of the can because sedimentation occurred, 2 = soft gel, 3 = medium gel, 4 = strong gel, and 5 = very strong gel. Samples scored 1 and 2 become homogeneous with agitation.

HI The HI was affected by the composition of the samples (data not shown). Significant differences (Table 2 ) were found for the interactions of milk fat and MSNF ( P < 0.0001) and, to a lesser extent, for the interaction of milk fat and carrageenan ( P < 0.0206) and time of storage ( P < 0.0370). These observations are probably caused by the importance of the interaction between milk fat and MSNF in determining the final viscosity of the sample (Figure 3). The presence of carrageenan in the formulation helped to maintain the dispersion of milk fat in the samples (HI 1); in the absence of gum, the samples showed signs of creaming (HI <1) or sedimentation (HI >1) that increased as time of storage increased. PI Significant differences were found in the PI among formulations (Table 2). The changes in PI were associated with the changes in creaming, sedimentation, and gelation. The PI for samples without carrageenan decreased slightly as time of storage
Journal of Dairy Science Vol. 81, No. 8, 1998

increased, which indicated some creaming. In sample 3, the PI increased excessively because of gelation, and, in sample 7, the PI increased during storage because of sedimentation. The PI for samples 4 and 6 stayed constant with time for samples stored at 4 and 25C and decreased with time after 2 mo of storage for the samples stored at 37C because of creaming. The PI for sample 8 remained almost unchanged with time, except for the sample stored at 37C, which gelled after 5 mo and for which the PI started to increase after 2 mo of storage. Soluble Calcium Generally, more soluble calcium was present in samples without sodium tripolyphosphate (4.9 and 8.1 mM for 6 and 11% MSNF, respectively) than in samples with it (4.1 and 7.4 mM for 6 and 11% MSNF, respectively) because of the ability of polyphosphates to complex with calcium. No correlation was found between the changes in soluble calcium content and the onset of gelation ( r = 0.3074). This lack of correlation has also been observed by others for UHT milk (12, 13). Leviton and Pallansch ( 1 4 )

RETORT-STERILIZED DAIRY BEVERAGES

2123

proposed that the stabilizing function of polyphosphates in concentrated UHT milks was due to interactions between micelles involving a calciumcaseinate-polyphosphate complex rather than the ability of the polyphosphates to bind calcium. Cyclic phosphates with no initial ability to complex calcium or retard age gelation and sodium citrate (that has similar ability to complex calcium to sodium hexametaphosphate) accelerate age gelation in UHT stored milk (14). CONCLUSIONS In this study, interactions of ingredients affected the overall viscosity of the product, which was directly related with the rate of changes that affected the physical properties of the samples. Sodium tripolyphosphate was implicated in promoting the gelation of samples with 11% MSNF and the lower apparent viscosity in samples with 1% milk fat compared with that of samples with 3% milk fat accelerated the gelation in the low fat formulation. Sedimentation rather than gelation was observed in the absence of polyphosphate and carrageenan in samples that contained 11% MSNF. The overall stability of the product with low or normal fat content can be improved by controlling the initial viscosity of the beverage and possibly by the use of cyclic polyphosphates, which have been reported to be more effective than linear polyphosphates for controlling age gelation. ACKNOWLEDGMENTS This project was funded in part by Dairy Promotion Inc., Springfield, MO. REFERENCES
1 Aguilera, J. M., and H. G. Kessler. 1989. Properties of mixed and filled type dairy gels. J. Food Sci. 54:12131221. 2 Association of Official Analytical Chemists. 1984. Official Methods of Analysis. 14th ed. AOAC, Arlington, VA.

3 Cano-Ruiz, M. E., and R. L. Richter. 1997. Effect of homogenization pressure on the milkfat globule membrane proteins. J. Dairy Sci. 80:27322739. 4 Dalgleish, D. G., and A.J.R. Law. 1989. pH-induced dissociation of bovine casein micelles II. Mineral solubilization and its relation to casein release. J. Dairy Res. 56:727735. 5 Fox, K., K.M.K. Harper, H. Holsinger, and M. J. Pallansch. 1965. Gelation of milk solids by orthophosphate. J. Dairy Sci. 48:179185. 6 Guiseley, K. B., N. F. Stanley, and P. A. Whitehouse. 1980. Carrageenan. Pages 516 in Handbook of Water-Soluble Gums and Resins. R. L. Davison, ed. McGraw-Hill Book Co., New York, NY. 7 Harper, W. J. 1984. Model food system approaches for evaluating whey protein functionality. J. Dairy Sci. 67:27452756. 8 Harwalkar, V. R. 1982. Age gelation of sterilized milks. Pages 229269 in Developments in Dairy Chemistry 1. P. F. Fox, ed. Appl. Sci. Publ., London, United Kingdom. 9 JMP Statistical Discovery Software. Version 3.1. 1995. SAS Inst., Inc., Cary, NC. 10 Kinsella, J. E. 1987. Physical properties of food and milk components: research needs to expand uses. J. Dairy Sci. 70: 24192428. 11 Kocak, H. R., and J. G. Zadow. 1985. Age gelation of UHT whole milk as influenced by storage temperature. Aust. J. Dairy Technol. 40:1421. 12 Kocak, H. R., and J. G. Zadow. 1985. Controlling age gelation in UHT milk with additives. Aust. J. Dairy Technol. 40:5864. 13 Kocak, H. R., and J. G. Zadow. 1985. Polyphosphate variability in the control of age gelation in UHT milk. Aust. J. Dairy Technol. 40:6568. 14 Leviton, A., and M. J. Pallansch. 1962. High-temperature-shorttime sterilized evaporated milk. IV. The retardation of gelation with condensed phosphates, manganous ions, polyhydric compounds, and phosphatides. J. Dairy Sci. 45:10451056. 15 Rizzotti, R., G. Tilly, and R. A. Patterson. 1984. The use of hydrocolloids in the dairy industry. Pages 285293 in Gums and Stabilizers for the Food Industry, 2. Application of Hydrocolloids. G. O. Phillips, D. J. Wedlock, and P. A. Williams, ed. Pergamon Press, Oxford, United Kingdom. 16 Rustom, I.Y.S., M. H. Lopez-Leiva, and B. M. Nair. 1995. UHTsterilized peanut beverages: changes in physicochemical properties during storage. J. Food Sci. 60:378383. 17 Sharma, R., H. Singh, and M. W. Taylor. 1996. Composition and structure of fat globule surface layers in recombined milk. J. Food Sci. 61:2832. 18 Snoeren, T.H.M., T.A.J. Payens, J. Jeunink, and P. Both. 1975. Electrostatic interaction between k-carrageenan and k-casein. Milchwissenschaft 30:393396. 19 Van Vliet, T., and A. Detner-Kikkert. 1982. Influence of the composition of the milk fat globule membrane on the rheological properties of acid milk gels. Neth. Milk Dairy J. 36:261265. 20 Xiong, Y. L., J. M. Aguilera, and J. E. Kinsella. 1991. Emulsified milk fat effects on rheology of acid induced milk gels. J. Food Sci. 56:920925.

Journal of Dairy Science Vol. 81, No. 8, 1998