Determination of the Composition in Sugars and Organic Acids in Peach Using Mid Infrared Spectroscopy: Comparison of Prediction Results According to Data Sets and Different Reference Methods

• Published in: Analytical chemistry (Washington) 2013-12, Vol.85 (23), p.11312-11318
• Main Authors: Bureau, Sylvie, Quilot-Turion, Bénédicte, Signoret, Véronique, Renaud, Christel, Maucourt, Mickael, Bancel, Doriane, Renard, Catherine M. G. C
• Format: Article
• Language: English
• Subjects: Acids; Analytical chemistry; Chromatography; Citric Acid - analysis; Citric Acid - chemistry; Databases, Factual - standards; Forecasting; Fourier transforms; Fructose - analysis; Fructose - chemistry; Fruits; Glucose - analysis; Glucose - chemistry; Infrared spectroscopy; Life Sciences; Liquid chromatography; Magnetic Resonance Spectroscopy - methods; Magnetic Resonance Spectroscopy - standards; Nuclear magnetic resonance; Organic acids; Peaches; Prunus - chemistry; Reference Standards; Spectroscopy, Fourier Transform Infrared - methods; Spectroscopy, Fourier Transform Infrared - standards; Spectrum analysis; Sucrose - analysis; Sucrose - chemistry; Sugars; Symbols
• ISSN: 0003-2700; 1520-6882
• DOI: 10.1021/ac402428s

The prediction of internal quality properties, such as sweetness and acidity, in peach fruit by mid infrared spectroscopy is of interest for rapid determination. Attenuated total reflectance Fourier transform infrared spectroscopy (ATR-FTIR) was tested here on two populations of peach fruits issued from contrasting genitors providing a large phenotypic variability. Over two successive years, 284 samples in 2006 and 483 samples in 2007 were characterized for soluble solids content (SSC), titratable acidity (TA), glucose, fructose, sucrose, malic acid, and citric acid contents. Sugar and organic acid composition were determined by three methods: colorimetric enzymatic measurements (ENZ), high-performance liquid chromatography (HPLC), or proton NMR spectroscopy (1H NMR), depending on the samples. For all samples, fruit homogenates were analyzed in ATR-FTIR using the same methodology and the same spectrometer. The objective here was to evaluate the effect of reference methods on the prediction performance. The best results were generally observed for SSC and TA, the percentage of the root-mean-square error of cross validation (RMSECV%) ranging respectively between 5.8% and 8.7% and between 5.9% and 8.0%, depending on the samples. For individual sugars and organic acids, the best correlations were obtained between ATR-FTIR data and ENZ reference data followed by HPLC and 1H NMR ones.