Vitamin C Determination by Ultraviolet Spectroscopy and Multiproduct Calibration

In this work the vitamin C was determined in industrialized nectar juices through ultraviolet (UV) spectroscopy and multiproduct multivariate calibration, based on partial least squares (PLS) regression.Since samples with different flavors, sugar content (light or not) were together in the model construction, it can be considered as a multiproduct and, due to the heterogeneity of the samples, it was necessary to optimize the calibration and validation sets by outliers elimination.The model was developed and validated by the evaluation of the figures of merit such as: accuracy, sensitivity, analytical sensitivity, adjust, linearity, relative prediction deviation, limits of detection and quantification, indicating that the multiproduct model developed from UV spectroscopy and PLS regression can be used in the industrial routine analysis as an alternative to titration or other time and reagent consuming methods.Here, it was evidenced that the UV-PLS multiproduct model provides advantages as being free of sample preparation steps, is suitable to be updated in order to measure other parameters, does not generates residues and is feasible to be implemented for on-line monitoring.Furthermore, the application of multivariate calibration in multiproduct models is extremely attractive from the industrial point of view.