Using front-face fluorescence spectroscopy for prediction of retinol loss in milk during thermal processing

- Kinetics of retinol degradation during thermal treatment was evaluated.
- Selected milk native fluorescent markers correlated well to retinol concentration.
- Prediction model of retinol concentration was established by fluorescent markers and showed good fitting.
- Front-face fluorescence spectroscopy has great potential in dairy industry for processing control.

The objectives of the present study were to model the kinetics of retinol loss and assess the potential of predicting retinol loss during thermal processing using front-face fluorescence spectroscopy (FFFS). A factorial design with three different temperatures (70, 80 and 90 °C) and five heating times (0, 5, 10, 15 and 30 min) was used in this study. The kinetics of the retinol degradation process fitted well a first-order reaction. Further, several milk native fluorophores were selected as fluorescent markers for development of retinol prediction models such as tryptophan, dityrosine, Maillard intermediate compounds and riboflavin. The maximum emission intensities of the markers were used to evaluate the correlation of retinol concentration with fluorescent markers and to establish the corresponding prediction models of retinol concentration. Retinol concentration was significantly correlated (P < 0.05) with all the fluorescent markers. The optimum prediction model was obtained from 45 samples using three fluorescent predictors, with an R2 of 0.87. The results obtained are encouraging and suggest that FFFS, as a low-cost, rapid, direct and non-destructive technique, has the potential to replace existing conventional analytical techniques practiced for retinol evaluation in heat treated milk and be adopted by dairy industry.