Fast and reliable NIRS determination of biochemical quality parameters in cocoa (Theobroma cacao L.)

Cocoa (Theobroma cacao L.) represents an important export commodity for several producing countries, especially in West Africa with a global production volume of more than 4 million tons per season. Qualitative heterogeneity and increasing consumption of cocoa products require fast and efficient methods for quality assessment of fermented cocoa with special regard to the fermentation process and the flavor potential of the beans. To date, quality control is mainly achieved by visual inspection (so-called ‘‘cut test’’) and sensory testing. Furthermore, various chromatographic methods such as GC-MS or LC-MS are applied aiming to quantify relevant flavor substances and their precursors. Although these techniques provide reliable and accurate data for cocoa quality, they are destructive, more or less expensive and time-consuming. Therefore, the aim of this study was to develop a fast and robust NIRS method allowing measuring the main metabolic profile of cocoa beans in a single analysis step, which is also applicable for routine analysis of various biochemical quality parameters. NIRS analyses were performed directly on freshly ground fermented cocoa beans applying a Multi Purpose Analyzer system of Bruker Optics (Ettlingen, Germany). The data were recorded in the spectral range from 12,500 to 3,600 cm-1 with a spectral resolution of 8 cm-1. All spectra were collected during 32 scans with threefold repetition per sample followed by averaging. Appropriate calibration models were developed for the individual analytes by PLS analysis of the spectral and reference data applying the Quant 2 module of OPUS 6.5 (Bruker, Germany). Different compound classes like phenolic substances (R2 = 0.93), organic acids (R2 = 0.88) and methylxanthines (R2 = 0.74) as well as individual substances like epicatechin (R2 = 0.93) and lactic acid (R2 = 0.87) could be precisely determined. Furthermore, fermentation time (R2 = 0.92), pH value (R2 = 0.94) and nitrogen content (R2 = 0.87) are reliably predicted, demonstrating that the new NIRS method can be used as fast and reliable alternative in routine quality assessment.