Comparison Among MIR, NIR, and LF-NMR Techniques for Quality Control of Jam Using Chemometrics

The potential of mid-infrared (MIR), near-infrared (NIR), and low-field nuclear magnetic resonance (LF-NMR) techniques combined with chemometrics for reliable and rapid determination of soluble solids content (SSC) and moisture in jams was investigated. Forty-four different jam samples with SSC ranging from 17.49 to 73.91 (A degrees Brix) and moisture ranging from 20.44 to 81.03% were used in this study. Principal component analysis (PCA) showed that the three spectroscopic techniques were able to distinguish the jams based on the SSC and moisture content. Partial least squares (PLS) regression exhibited a good correlation between the reference values and the MIR, NIR, and LF-NMR predicted ones, with low errors of prediction and high coefficients of determination. An F test at 95% confidence level did not indicate significant differences between the accuracy of the PLS models obtained using MIR and NIR spectroscopic techniques. However, significant differences were observed comparing MIR with LF-NMR and NIR and LF-NMR. The residual prediction deviation (RPD) up to 2.5 indicated that all models are good. (AU)