Caking Development in Lemon Juice Powder

Spray dried fruit juice powders are produced around the world. These powders contain fruit sugars and acids that make them impossible to dry without the addition of a drying aid such as maltodextrin which is added to raise the glass transition of the mixture. Thus, prediction of the glass transition temperature of mixtures is important when working out how much maltodextrin to add and at what conditions to run the spray drier. Even with maltodextrin added, the powders have a strong tendency to cake. This paper compares the glass transition temperature as measured over a range of water activities with a DSC with predictions based on a simple additive model that uses the Gordon Taylor approach for predicting the effect of water of individual component glass transition temperatures and the Fox equation. It demonstrates the development of caking between spray dried lemon juice particles with time when the temperature of the powder is raised above the glass transition temperature using two techniques.The predicted glass transition temperatures for powder made from juice, using the initial analysis were typically 29 °C too high compared to the measured values over a range of water activities. This was reduced to 22 °C when the compositional analysis for the juice was obtained and used. An artificial powder was made using the same ratio of maltodextrin to food sugars and acids and the Tg was measured at 23.4 °C at a water activity of 0.28 which corresponds to a predicted Tg of 33.2 °C.The powder was put into 28 mm diameter by 4 mm deep dishes and then placed in a constant RH and temperature cabinet (RH = 26% to match the water activity of the powder, with the temperature set at 7, 12 and 22 °C above the measured Tg of 18 °C). The level of caking was measured by the blow tester and by penetration force measurement. Both methods showed the linear development of caking strength with time and that the rate of caking was a strong function of T-Tg.Only a single measurement per sample was possible with the blow tester as it tended to destroy the small sample cakes. Multiple readings were not possible due to the limited supply of powder. With the penetration method 6 to 8 readings were taken per sample. It was found that small lumps affected the penetration measurement and less scatter occurred when these were sieved out. The penetration method was also able to follow the caking to higher caking strengths, although there was more scatter in the data for the more strongly caked samples.