TY - JOUR
T1 - Authentication of cocoa bean shells by near- and mid-infrared spectroscopy and inductively coupled plasma-optical emission spectroscopy
AU - Mandrile, Luisa
AU - Barbosa-Pereira, Letricia
AU - Sørensen, Klavs Martin
AU - Giovannozzi, Andrea Mario
AU - Zeppa, Giuseppe
AU - Engelsen, Søren Balling
AU - Rossi, Andrea Mario
PY - 2019
Y1 - 2019
N2 - The aim of this study was to evaluate the efficacy of a multi-analytical approach for origin authentication of cocoa bean shells (CBS). The overall chemical profiles of CBS from different origins were characterized using diffuse reflectance near-infrared spectroscopy (NIRS) and attenuated total reflectance mid-infrared spectroscopy (ATR-FT-IR) for molecular composition identification, as well as inductively coupled plasma-optical emission spectroscopy (ICP-OES) for elemental composition identification. Exploratory chemometric techniques based on Principal Component Analysis (PCA) were applied to each single technique for the identification of systematic patterns related to the geographical origin of samples. A combination of the three techniques proved to be the most promising approach to establish classification models. Partial Least Squares-Discriminant Analysis modelling of fused PCA scores of three independent models was used and compared with single technique models. Improved classification of CBS samples was obtained using the fused model. Satisfactory classification rates were obtained for Central African samples with an accuracy of 0.84.
AB - The aim of this study was to evaluate the efficacy of a multi-analytical approach for origin authentication of cocoa bean shells (CBS). The overall chemical profiles of CBS from different origins were characterized using diffuse reflectance near-infrared spectroscopy (NIRS) and attenuated total reflectance mid-infrared spectroscopy (ATR-FT-IR) for molecular composition identification, as well as inductively coupled plasma-optical emission spectroscopy (ICP-OES) for elemental composition identification. Exploratory chemometric techniques based on Principal Component Analysis (PCA) were applied to each single technique for the identification of systematic patterns related to the geographical origin of samples. A combination of the three techniques proved to be the most promising approach to establish classification models. Partial Least Squares-Discriminant Analysis modelling of fused PCA scores of three independent models was used and compared with single technique models. Improved classification of CBS samples was obtained using the fused model. Satisfactory classification rates were obtained for Central African samples with an accuracy of 0.84.
KW - Cocoa bean shell
KW - Data fusion
KW - Food traceability
KW - Inductively coupled plasma
KW - Mid-infrared spectroscopy
KW - Near-infrared spectroscopy
U2 - 10.1016/j.foodchem.2019.04.008
DO - 10.1016/j.foodchem.2019.04.008
M3 - Journal article
C2 - 31054691
AN - SCOPUS:85064218776
SN - 0308-8146
VL - 292
SP - 47
EP - 57
JO - Food Chemistry
JF - Food Chemistry
ER -