Development of a Facile and Convenient Method for Sugar Determination in Low Moisture Confectioneries and Honeys Using Fourier Transform Infrared Attenuated Total Reflectance Spectroscopy and Chemometrics

Abstract

A method using Fourier transform infrared with attenuated total reflectance spectroscopy and partial least squares (PLS) chemometric analysis to simultaneously quantify the mass percent concentrations of glucose, fructose, and sucrose, commonly found in honey and low moisture confectionery food products, was developed. A three-component full factorial design was generated as a training set (n = 64) to predict the sugar concentrations in a Box-Behnken testing set (n = 16). Analysis of the entire spectral region from 649.99 to 3996.38 cm−1 using Savitzky–Golay signal processing technique and PLS algorithm garnered optimal errors of prediction and better linearity between predicted and measured concentrations in the test set compared to analysis using a specific spectral region from 800 to 1500 cm−1 and using a wide array of signal processing techniques (first derivative, second derivative, moving average, binning, and standard normal variate). The model further detected these sugar concentrations in 3 standard honeys, 12 commercial honey samples, 8 honey adulterants, 13 low moisture confectionery food samples, and 15 unknown honey samples from Louisiana apiaries. Applying principal component and clustering analysis also enabled the discrimination of these honey samples based on their fructose:glucose ratios. Thus, the developed method represents a potentially novel, simple, accurate, nondestructive, and rapid means of quantifying concentrations of glucose, fructose, and sucrose in honey and low-moisture confectioneries in 1–2 minutes.