RAPID DETERMINATION OF SPOILAGE ON MAYONNAISE BY USING NONINVASIVE THIRD ORDER FLUORESCENCE DATA MODELING

GOICOECHEA, HÉCTOR C

Congreso; CAC 2012; 2012

Data recording with a noninvasive system followed by proper chemometric modeling, allows for rapid knowledge about the quality of food samples [1, 2]. In this work, the potential of excitation emission fluorescence matrix (EEM) data recorded with a fiber optic system on mayonnaise samples along with chemometric methods was investigated. Influence of storage time at 37ºC was evaluated by obtaining EEMs of both commercial and prepared from scratch mayonnaise samples at time zero, and each 30 min, up to a total time of two days. After that, data were modeled with multivariate curve resolution with alternating least squares (MCR-ALS), parallel factor analysis (PARAFAC) and discriminating n-way partial least squares (DA-NPLS). As a result, the obtained information was: (a) evolution profiles of fluorescent components (native and/or generated by spoilage), which were used to discriminate samples (i.e., commercial branches, spoilage, etc.), and (b) regression models that were built to predict bacterial contamination.    In addition, different strategies were attempted to alleviate the scattering phenomena, which was present due to the wide wavelength regions employed in order to better exploit the total luminiscent information.