Effect of pasteurization and fat, protein, casein to serum protein ratio, and milk temperature on milk beverage color and viscosity

Our goal was to determine the effect of pasteurization-homogenization, fat and protein concentration, proportion of milk protein that is casein and serum protein, and temperature on sensory and instrumental measures of viscosity and color of milk-based beverages. A second goal was to use instrumental measures of whiteness and yellowness to predict sensory measures of whiteness and yellowness. A complete balanced 3 factor (fat, true protein, and casein as a percentage of true protein) design was applied with 3 levels of fat (0.2, 1.0 and 2.0%), 4 levels of true protein (3.00, 3.67, 4.34, and 5.00%) within each fat level, and 5 levels of casein as a percentage of true protein (CN%TP; 5, 25, 50, 75, and 80%) within each protein level for beverage formulation. Instrumental color and viscosity, and visual sensory color analyses were done on each beverage formulation. For unpasteurized beverages across 3 fat levels (0.2, 1, and 2%), changes in CN%TP had the largest effect on L values, sensory whiteness, opacity, color intensity, and yellowness, whereas changes in fat concentration had a stronger influence on a and b* values. Increasing CN%TP from 5 to 80% increased L values, sensory whiteness, and opacity, and decreased sensory color intensity and yellowness. The a and b* values increased with increasing fat concentration. For unpasteurized milk protein beverages within each fat level, variation in CN%TP dominated the changes in L values, sensory whiteness, and opacity, and decreased a and b* values, sensory color intensity, and yellowness. The effect of heat (pasteurization and homogenization) and its interaction terms had the second largest effect on color of milk protein beverages with respect to instrumental color data and sensory appearance attributes. Heat increased L values, sensory whiteness, and opacity, and decreased a and b* values, sensory color intensity, and yellowness. Increases in tempera- ture decreased instrumental viscosity and changes in protein concentration and CN%TP had a greater effect on instrument viscosity data within each temperature (4, 20, and 50 degrees C) than fat. Sensory perception of yellowness was not highly correlated with b* values. Multiple linear regressions of L, a, and b* values produced more robust predictions for both sensory whiteness and yellowness than simple linear regression with L and b* values alone, and may be a useful instrumental approach for quality control of sensory whiteness and yellowness of milk protein beverages.