NON-DESTRUCTIVE ESTIMATION OF MATURITY LEVEL OF ‘CRYSTAL’ GUAVA FRUIT BY MEANS OF FLUORESCENCE SPECTROSCOPY

Abstract

This research aims to explore the potential use of fluorescence spectroscopy combined with chemometrics for predicting the maturity level of crystal guava fruits non-destructively. The physicochemical properties including total soluble solids (TSS), water content, firmness, and pH were obtained from laboratory tests and used as a reference in developing the predictive model. The fluorescence emission spectra under excitation of 365 nm UV LED were used as predictors. The fluorescence spectra were acquired and recorded using a miniaturized UV-Vis spectrophotometer with an effective 340 – 800 nm range. In total, 240 sets of data from crystal guava fruit samples with varying ages of 90-106 DAA (days after anthesis) were used for calibration and validation. A multivariate analysis using Partial Least Squared Regression (PLSR) was used to develop the predictive models. Several data preprocessing methods were applied to enhance the information contained in the spectral data to find the best predictive model. Analysis results showed that the developed model could accurately predict water content with rcal=0.94; SEC=0.08%, rval=0.84; SEP=0.08%; RPD=2.59, followed by TSS with rcal =0.91; SEC = 0.47%Brix, rval =0.79; SEP=0.48%Brix; RPD=2.13. Although with lower accuracy, it also could predict firmness with rcal =0.86; SEC=0.43 kgf, rval =0.74; SEP=0.43 kgf; RPD=1.82 and pH with rcal =0.81; SEC=0.10, rval =0.79; SEP=0.09; RPD=1.76. The results indicate that fluorescence spectroscopy could be used as an alternative method for non-destructive estimation of physicochemical properties as indicators of the maturity level of crystal guava fruits.