Fourier transform infrared spectroscopy and chemometrics as a tool for the rapid detection of other vegetable fats mixed in cocoa butter

The Fourier transform infrared (FTIR) technique in combination with multivariate data evaluation was used to analyze a wide variety of cocoa butters (CB), cocoa butter equivalents (CBE), and mixtures thereof. The sample set consisted of 14 CB (10 pure from various geographical origins and 4 commercial mixtures), 18 CBE (12 mixtures and 6 pure CBE from kokum, illipé, and palm midfraction), and 154 mixtures of CB with CBE at various concentrations (ranging from 5 to 20%). A total of 186 samples were analyzed in triplicate. All CB and CBE were shown to have very characteristic FTIR spectra that gave highly reproducible fingerprints. The main vibrational models were also elucidated. FTIR can easily be employed to distinguish between pure CB and pure CBE. With prior knowledge of which cocoa butter is present in mixtures, FTIR can be applied to distinguish between CB mixed with CBE at the 10 and 20% levels (corresponding to about 2 and 5% of CBE in chocolate). However, the study revealed that a single "global" statistical model (multilayer perception, radial basis functions, or partial least square regression) was not able to predict the precise level of addition. The FTIR approach detailed here shows great potential as a rapid screening method for distinguishing between pure vegetable fats and, we believe, could be extended to investigate mixtures of CB and CBE by the establishment of a database.