Avaliação reológica e sensorial de doce cremoso produzido com soro de leite de cabra, leite de vaca e polpa de umbu.

Abstract

Goat milk offers a competitive advantage over cow's milk due to its better digestibility and suitability in the production of high-value cheeses. Large-scale production of cheese generates whey as a by-product. Whey is rich in proteins with high biological value, but low acceptability because of its unpleasant odor and taste, especially when derived from goat's milk. The most economically viable economic strategy for goat milk whey is to incorporate ingredients that mask or neutralize its taste, with added gains if local ingredients such as umbu pulp are used. In this study goat milk whey, cow's milk and umbu pulp were used in the production of sweet creamy desserts. Samples were elaborated with a mixture of goat milk whey and cow's milk at a ratio of 1:1, plus sucrose and umbu pulp at concentrations of 10, 15 and 20%. Umbu pulp and the sweets were assessed for pH, soluble solids, ascorbic acid, titratable acidity, ash, reducing sugars, total proteins, lipids, density, water activity and color. Creamy sweet desserts were studied for rheological behavior and submitted to sensory evaluation. In sensory analyses judges assessed samples according to the hedonic scale, judging color, consistency, taste, sweetness, odor, appearance, acidity and intention to consume. Experimental data used in the study of rheological behavior were collected in a Brookfield viscometer, at temperatures of 20, 30 and 40 °C, and Ostwald-de-Waelle (Power Law), Casson, Herschel-Bulkley and Mizrahi-Berk models were fit to rheograms. Umbu pulp exhibited considerable acidity. The sample with the lowest amount of umbu pulp was the most accepted. All the creamy sweet desserts displayed pseudoplastic behavior; apparent viscosities decreased with an increase in rotation speed; and viscosities decreased with an increase in temperature. Power Law, Herschel-Bulkley, Casson and Mizrahi-Berk rheological models fit shear stress and strain rate data well, with the Mizrahi-Berk model resulting in the best fit.