Use of infrared spectroscopy ATR-FTIR as a rapid alternative for Colombian cocoa liquors: classification varieties and estimation of volatile compounds concentration.

Criollo cocoa beans are defined as fine and flavor due to sensory notes that could be commercialized in premium chocolates. Chromatography mass-spectrometry (GC-MS) is commonly used in cocoa samples to differentiate between varieties; however, it is an expensive and time-consuming technique. Based on this, an alternative technique for variety differentiation was developed using Fourier Transform Infrared (ATR-FTIR) spectroscopy with chemometrics. Seventeen Colombian cocoa liquor samples were analyzed using GC-MS and ATR-FTIR. This included 12 from Criollo varieties Montaña (LMON) and Olivo (LOLI), and 5 from commercial liquors made from Trinitario Colombian beans (LC). Volatile results included 80.61 % of variability in the PC1-PC3, the samples were differentiated by 1-pentanol and 2-heptanol (LMON), 2-octanone, benzonitrile and 2,3-butanodiol diacetate (LOLI), and pyrazines and phenylacetaldehyde (LC). The ATR-FTIR wavenumber classification considered 84.09 % of variability in the PC1-PC3. The ATR-FTIR Criollo spectra were influenced by amino acid (LMON), protein, acetyl, and roaster modification (LOLI) regions. Meanwhile, lipids, pyrazines and some aromatic regions influenced the LC spectra. Additionally, the mathematical models (PLS-DA), developed to differentiate samples, presented comparable correct estimations: the GC-MS and ATR-FTIR models yielded accurate estimations of 97.62 % and 97.80 %, respectively. Finally, the linear regressions to estimate the concentration of volatile compounds increased the fit model (R) from 0.62-0.84 to 0.96-0.99 using the sample classification as a pretreatment. These results confirmed that the use of ATR-FTIR combined with chemometrics was an effective rapid method for classifying cocoa liquor samples such as Criollo, or commercial liquors to be used as raw materials in premium products.