Utilization of chemometrics methods in comprehensive two dimensional gas chromatography for evaluate cocoa, its derivatives and carob

Author(s)

Luciana Fontes de Oliveira

• Document type: Doctoral Thesis
• Press: Campinas, SP.
• Institution: Universidade Estadual de Campinas (UNICAMP). Instituto de Química
• Defense date: 2016-12-06

Examining board members

Ronei Jesus Poppi; Edenir Rodrigues Pereira Filho; Fabio Augusto; Juliana Azevedo Lima Pallone; Marcio Pozzobon Pedroso

• Advisor: Ronei Jesus Poppi

Abstract

In this thesis chemometrics tools were applied in two dimensional gas chromatography data for quality studies of cocoa, derivatives and carob. In this sense, four applications were developed, in the first one Plackett-Burman designs were used to evaluate the chromatograph and detector conditions after small variations were realized in some parameters. Among the evaluated variables in the first design, the calculated effects for the column flow and the voltage of photomultiplier were significant. For the second design the calculated effect of mass/charge (m/z) minimum ratio was significant and these factors should be monitored with greater precision when performing the experiments. The second application, realized in partnership with the University of Copenhagen, is about a new development to align GCxGC chromatograms using mass spectra information, and parallel to this was held exploratory analysis of chocolate and carob samples using Principal Component Analysis (PCA). Futhermore, it was possible to identify important compounds to differentiate chocolate and carob samples analyzing the PCA loadings. In the third application, Fisher ratio was applied to select important variables for the distinction of nibs (cocoa beans) and commercial chocolate samples. In a first moment, nibs samples from Ivory Coast and Brazil were evaluated and fifteen compound were important for the separation found in developed PCA with selected variables. In a second moment, the Fisher ratio was used to select important chromatograms parts to describe commercials chocolate samples from Brazil and imported, using the mass spectrum mode from the PARAFAC (Parallel Factor Analysis) model, developed with the selected variables found using the Fisher ratios, it was possible to identify important variables for class differentiation that occurred in samples mode in the model. In the fourth and final application Fisher ratio and PCA were also used as chemometrics tools, however in order to establish a correlation between GCxGC and sensory analysis (Optimized Descritive Profile) using chocolate samples with different cocoa percentages. In this sense, the bitter attribute results were used to define the Fisher ratio classes. Finally, the important compounds related with high and low values of the attribute in question were found (AU)

• FAPESP's process: 12/14695-8 - Utilization of chemometric methods in comprehensive two-dimensional gas chromatography for evaluate cocoa, its derivatives and carob
• Grantee: Luciana Fontes de Oliveira
• Support Opportunities: Scholarships in Brazil - Doctorate