Prediction of water content by FT-NIR spectroscopy in wet and dried alginate beads containing emulsified lycopene

AGUIRRE CALVO, TATIANA; SANTAGAPITA, PATRICIO R.

Cordoba

Congreso; V Congreso Internacional de Ciencia y Tecnología de los Alimentos Córdoba, 2014; 2014

Ministerio de Industria,Comercio, Minería y Desarrollo Científico Tecnológico. Secretaría de Ciencia y Tecnología

Lycopene is used as antioxidant and as colorant (red/reddish) in the Food Industry. Labile compound as lycopene are often encapsulated to preserve them from degradation, besides of giving controlled release and adequate dosage. In a previous work, lycopene was successfully encapsulated in alginate beads containing sugars and biopolymers, obtaining high content and stability of lycopene on beads subjected to both freezing and drying. It is already known that water content (wc) is an important parameter since it influences both chemical and physical stability. Traditional methods for determine wc are time consuming (gravimetric) or involve toxic reactants (Karl-Fischer). Hence, the aim of this work was to evaluate the feasibility of a non-destructive and rapid method based on NIR spectroscopy to determine the wc of wet and dried alginate beads containing encapsulated emulsions of lycopene formulated with several excipients. The predicted values were compared to the ones determined by difference in weight after drying in a vacuum oven (48h, 100°C) (wet beads), or by Karl-Fischer method (dried beads). Lycopene was extracted from grapefruit, then was emulsified using different hydrocolloids/ sugars, and finally alginate-Ca(II) beads were prepared (wet beads) and dried (dried beads). Spectra were acquired from 4000 to 10000 cm-1 by averaging 32 scans (resolution: 8 cm-1), using an integrating sphere reflectance accessory. 1%w/v alginate:olive oil emulsions were prepared as calibration samples for NIR measurements in two wc ranges: 25-55%w/v and 1-5%w/v, for wet and dried beads, respectively. Spectra were corrected by multiple scattering algorithm. PLS were performed and a t-test of validation (30% dataset) was used to compare the predictive accuracy of models. Calibration and test-set validation showed high R2 (0.99), using 7 and 2 latent variables for high and low wc emulsions. RMSE values were 0.6 and 0.7 for calibration and test-set calibration for high wc samples, respectively, and 0.079 and 0.084 for low wc samples. The X-loadings revealed that the most influencing wavelengths corresponded to water and oil contributions (5200-5800 cm-1). By applying the regression models obtained for emulsions, RMSEP were 1.9% and 0.1% for wet and dried beads, respectively, which were more than acceptable. Both gravimetric and NIR predicted wc values of wet beads ranged from 37 to 44 (±2) %w/wwb. Karl-Fischer measurements revealed slightly higher wc values (ranging from 2.3 to 3.9 (±0.1) %w/wdb) than the NIR predicted ones in most cases (ranging from 1.1 to 3.9 (±0.1) %w/wdb). Once the calibration was performed, NIR prediction is much faster than conventional methods, and, as a green technique, has the advantage of no-requiring toxic solvent, being a suitable method to predict wc. The authors acknowledge the financial support of ANPCYT (PICT 0928) and CONICET (PIP 100846).