# Viscosity of Reconstituted Milk Protein Concentrate Solutions as a Function of Shear Temperature and Concentration.

код для вставкиСкачатьDev.Chem. Eng. Mineral Process., 7{1/2),pp.131-139,1999. Viscosity of Reconstituted Milk Protein Concentrate Solutions as a Function of Shear, Temperature and Concentration S. O’Donnell and F. Butler* Dept. of Agricultural and Food Engineering, University College Dublin, Earlsfort Terrace, Dublin 2, Ireland The viscosity of reconstituted milk protein concentrate solutions was determined at various shear rates, (50- 1000 s-’ ), temperatures, (20-60°C) and concentrations, (200260 g/kg). For a given temperature and concentration, the solutions were shear thinning and the effect of shear was adequately described by the Ostwald power law model. Viscosity increased with concentration and decreased with temperature. The effect of temperature on the consistency index followed an Arrhenius type relationship. The effect of concentration on the consistency index was modelled with both a power and an exponential relationship. The flow behaviour index was independent of temperature and concentration. The combined effects of shear rate, temperature and concentration on viscosity were modelled using an extended power law model. Introduction Milk protein concentrate (MPC) is a relatively new product, based on ultrafiltration and drying of skimmed milk, with a typical protein content from 50-85% of total solids. Producing MPC is generally based on protein concentration by ultrafiltration of skim milk using a membrane with a typical molecular weight cut-off of 1OkDa [I]. * Authorf o r correspondence. 131 S.O'Donnell and F.Butler The product is then pasteurised, followed by final water removal by evaporation and drying. Depending on the selection of the process parameters, the original properties of the proteins can be preserved or functional features suitable for specific utilisation purposes can be formed [2]. A major problem associated with MPC processing is its high viscosity at concentrations above 2OOgkg of total solution, as the practical limits of its solubility are approached. As it is a relatively new product, very little has been published on the viscosity of MPC, particularly at high concentrations. The objectives of this study were to measure the combined effects of shear rate, temperature and concentration on the viscosity of MPC at high concentrations and to apply a rheological model which would adequately predict their combined effects on the viscosity of MPC solutions. Theory The Ostwald power law model is commonly used to describe the effect of shear on the fmal viscosity of a non-Newtonian time independent liquid: , n-I 77, = K Y To adequately model temperature and concentration effects, the consistency index, K , and the flow behaviour index, n, are not constant, but vary with temperature and concentration. Several workers [3], [4], have described the temperature dependency of the consistency index with an Arrhenius type relationship: K = K, exp(E, / RT) The effect of concentration on KO can be handled by either a power or an exponential relationship, [ 5 ] : 132 Viscosity of reconstituted milk protein concentrate solutions KO = K,C" (3) K O = K2e(bc) (4) The flow behaviour index is a measure of the shear-thinning property of a substance. Several workers [4], [ 5 ] , have described the temperature and concentration dependency of the flow behaviour index with a linear type relationship: n(T,C) = n + n , C + n T T To model temperature and concentration effects, equations (l), (2), (3), (4) can be combined to give two possible extended power law models [ 5 ] : Experimental A single batch of commercially available MPC powder was used in all experiments. Reconstituted MPC powder was used in this study. Sampling and immediate viscosity measurement of the product during manufacture at a range of concentrations and temperatures was not technically feasible. The protein content [6], non-casein nitrogen content [7], lactose content [8], fat content [9] and moisture content [ 101 of the milk protein concentrate were determined. Each compositional test was repeated three times. 1 kg MPC solutions were prepared by reconstituting MPC powder with distilled water to the desired concentration using a Silverson L2R mixer (Silverson Machines Ltd. Chesham, UK) at a constant temperature of 40°C. Three concentrations were used, (200, 240, 260 g k g ) . At concentrations above 260 g/kg the mixtures became 133 S. O'Donnell and F. Butler paste like. The solutions were mixed for 1 1.1 to ensure even dispersion. The solution was poured into separate 150 mm sample bottles that were then gently shaken in a water bath at 25°C for 30 min to minimise the amount of air in the solution and then allowed to equilibrate for a minimum of 30 min undisturbed in the water bath before viscosity measurement. A Physica Rheolab MC 100 Rheometer (Physica MeStechnik, Stuttgart, Germany) was used for all viscosity measurements. A 2 2 cup and bob geometry to D M 53019.1 [ 111 was used, consisting of a 45 mm diameter rotating bob and a 48.8 mm diameter faed cup. For each test the MPC solution was carefully loaded into the cup so as to minimise damage to its structure, and the bob was lowered. The sample was allowed to rest for 15 min before testing. Constant shear rate tests (50, 100,300,500, 1000 s-') were carried out at various temperatures (20,30,40, 50, 60°C) to examine the effect of time on the viscosity of the samples. One sample from each 150 mm sample bottle was used for each combination of shear rate, temperature and concentration. Samples were sheared for 15 min with viscosity readings recorded every 18 sec. All viscosity measurements were repeated four times. Results and Discussion The MPC powder contained 856 g/kg protein, 28 gikg non casein nitrogen, 47 g/kg lactose, 16 g/kg fat and 42 g k g moisture. This was typical of commercially produced h4PC powders [ 1J and was similar to the declared composition by the manufacturer. The influence of time of shear on viscosity was relatively small, with a maximum difference of 5.1% between the initial and final viscosity values at 900 s. To eliminate the effect of time in the rheological modelling, final viscosity values at 900 s were used in the subsequent analysis. Figure 1 shows a logarithmic plot of the effect of shear on the final viscosity at 20°C and 60°C.The viscosity of the MPC solutions decreased significantly with increasing shear. At a given shear rate, viscosity increased with concentration and decreased with temperature The Ostwald power law model, equation (l), was used to characterise the variation of viscosity with shear rate at the various temperatures and 134 Viscosiry of reconstituted milk protein concentrate solutions concentrations. Linear regression of the average results for log of viscosity and the log of shear rate was used to calculate K and n for each concentration and temperature. The coefficient of determination for the regressions, R2,varied from 0.9-0.99. In all cases the significance of the regressions was high (P<0.001). Y a" 1 .. 0.1 .. - v .-h .-gM > 0.01 10 100 1000 Shear rate (5.') Figure 1. A comparison of experimental data (points) with the power law modeI (Equation I ) (lines) for the effect of shear rate on the final viscosity of milk protein concentrate for concentrations of (g/kg) :0,200; 0, 240; A, 260. -2OOC. -- -- 60OC.(Data at intermediate temperatures omitted for clarity) Figure 2 shows the variation of the consistency index with temperature and concentration. The temperature dependency of the consistency index increased considerably with concentration. Similar increases in the consistency index values at high concentrations have been reported for sodium caseinate solutions (>140 gkg) [ 121, [131, and for whey protein solutions (>500 gkg) [ 141. Multiple regression was used to determine the Arrhenius parameters (equation (2)) and the constants for both the power and exponential relationship for the effect of concentration on KO (equations (3) and (4)). In both cases the multiple regressions were highly significant (P< 0.001). Figure 3 shows the variation in the flow behaviour index, n, with temperature and concentration. No correlation was found between n and concentration or temperature. 135 S, O’Donnell and F. Butler As a consequence a mean value of n was used in the rheological model, even though n varied from 0.25-0.71. The variation in n was unexpected but it was subsequently found that other workers have experienced similar variation [ 121, [ 151. 20 3E: 30 50 40 Temperature 60 e C) Figure 2. Consistency index as a function of temperaturefor milk protein concentrate (g/kgl: 0,200; e, 240; A, 260. -0 -5 0.8 8 2 2 0.6 0.4 . I i5 0.2 0 . 30 5 40 50 , 60 Temperature f C) Figure 3. Behaviour Index as a function of temperaturefor milk protein concentrate (g/kg): 0,200; e, 240; A, 260. I34 Viscosity of reconstituted milk protein concentrate solutions The effects of shear rate, temperature and concentration were modelled using equations (6), and (7) except that the flow behaviour index value used was a mean value. The method of calculating the sum of the residual squares (R.S.S.) [4] determined which model provided the best fit to the experimental results. Equation (6) (R.S.S.= 97), where the effect of concentration on was modelled by a power relationship, provided a better fit to the data than equation (7) (R.S.S.= 490). The resulting parameters using the power law relationship was: (y) (0.451)-1 77,= 7.6 10".c~.~'.e *Y lo T A 10 100 1000 S h e a r rate, 11s Figure 4. A comparison of experimental results (points) with predicted extended power law model (Equation 8) (solid lines) for a concentration of 260 (g/kg)over a temperature range PC) : 0 , 20; a,30; U,40; I.,50; A, 60. Figure 4 compares experimental final viscosity values with predicted results obtained using equation (8). The model tended to underestimate the viscosity values at the lowest concentration of 2OOgkg and overestimate the viscosity values at the higher concentrations of 240g/kg and 260gkg. In general the model's predictive I37 S. O'Donnell and F. Butler ability improved with increasing temperature. The worst error was 180%' which occurred at a concentration of 260 g k g and a temperature of 20°C. At these conditions the measured viscosity was at its highest. This illustrates the lhitations in using a mean value for n in the model and in applying such a viscosity model where the liquid is close to becoming paste like. At other concentrations and temperatures the error was much lower (30 - 40%). The model was effectively built up step-wise, allowing for shear rate, temperature and concentration. Consequently errors introduced at each step were cumulated. This should be taken into account when judging the model's predictive ability. Conclusion The viscosity of MPC solutions varied significantly with shear, temperature and concentration. A modified power law equation fitted the viscosity data but caution has to be exercised in its use at high concentrations and ambient temperatures when the error between predicted and actual viscosity was as high as 180%. Generally the - error was lower (30 40%). The viscosity of MPC solutions increased dramatically above concentrations of 200gikg. From a processing/flow perspective, a concentration of 240 - 260 g/kg is suggested as a practicaI limit. The use of elevated temperatures (50 - 60°C)would help to reduce viscosity significantly. Acknowledgement This work was financed by E.U. structural funds money. Nomenclature a Power index in equation (3). b Exponential index in equation (4). c Concentration (kgfl<g) Ea Activation energy (Jkg mole) K Consistency index (Pa s") 138 Viscosity of reconstituted milk protein concentrate solutions KO Constant of proportionality in equation (2) (Pa s") K, Constant of proportionality in equation (3) (Pa s") K2 Constant of proportionality in equation (4) (Pa s") n Flow behaviour index nc Constant of proportionality for concentration in equation (5). nT Constant of proportionality for temperature in equation ( 5 ) . (TI) R Gas constant (Jkg mole K) T Temperature (K) 11, Final viscosity (Pa s) Y Shear rate (s-') References 1. Zwijgers, A. 1992. Outline of milk protein concentrate. De Mekindustrie Veghel technical bulletin. 2. Novak, A. 1993. New applications of membrane processes: Milk protein concentrates. Bulletin of IDF No. 920 1,5 1-65. 3. Cervone, N.W., and Harper, J. M. 1978. Viscosity of an intermediate moisture dough. J. Food Process Eng., 2,83-95. 4. Verges, B., and Villemaire, J.P. 1987. Rheological behaviour of low moisture molten maize starch. Rheologica Acta, 26, 570-576. 5 . Harrod, M. 1989. Modelling of flow properties of starch pastes prepared by different procedures. J. Food Process Eng., 11,257 - 275. 6. IS0 5549. 1978. Method for the determination of the protein content in casein. 7. IDF 29. 1964. Method for the determination of non-casein nitrogen in casein. 8. IS0 5548. 1980. Method for the determination of the lactose content of casein. 9. IDF 1A. 1969. Method for the determination of the fat content in casein. 10.I S 0 5550. 1978. Method for the determination of the water content in casein. 11. DIN 53019.1. 1980. Viscometry; Determination of viscosities and flow curves using standard design rotary viscometers with a standard geometry measuring system. 12. Hermansson, A.M. 1975. Functional properties of proteins for foods: flow properties. 3. Texture Studies, 5,425-439. 13.Fichtali, J. 1993. A rheological model for sodium caseinate. J. Food Eng., 19, 203-2 1 1. 14.Alizadehfard, M. R. & Wiley, D. E. 1996. Non-Newtonian behaviour of whey protein solutions. J. Dairy Research, 63,3 15-320. 15. Hernandez, E. 1995. Viscosity changes in orange juice after ultrafiltration and evaporation. J. Food Eng., 25, 387-396. 139

1/--страниц