The aim of this work is focused on the melamine migration from food contact materials (FCMs), considering data
obtained from univariate analysis versus that obtained from multivariate approach in liquid chromatography
coupled to diode array detector.
Plastic food contact materials are made from monomers and additives. Moreover, non-intentionally added
substances (NIAS) can be part of the composition of the FCM: raw material impurities or process by-products,
inks or adhesives.
Any compound present within a FCM can migrate to foodstuff. Specific migration of some substances from
plastic FCMs to food/simulant is limited by European legislation in force (Commission Regulation No 10/2011).
Quantification of analytes in migration samples through a univariate analysis could lead to erroneous results.
As an example, in liquid chromatography NIAS can interfere when coeluting with analytes or when they have
close retention time. In that case, an overestimation would happen and the verification of the compliance of the
specific migration limit (SML) of a substance would be incorrect.
A solution to the problem can be found in the application of a chemometric tool with the second-order
advantage, which allows the unequivocal identification of analytes. Specifically, for this work, PARAFAC/
PARAFAC2 decomposition technique along with tensors arranged from HPLC-DAD data of migration (test and
kinetics) samples were used for the identification and quantification of melamine.
Results of melamine quantity found in migration samples from five types of melaware by means of a multivariate approach were compared to results obtained with a univariate data analysis carried out with values of
chromatographic peak area as response. The comparison reveals that in test samples, univariate analysis supposes an overestimation in the quantity of melamine of 30 % on average, with respect of the concentration
obtained from the multivariate approach. Besides, in kinetics samples it is remarkable that for one migration
cycle the melamine found was 10 times above the one that obtained with PARAFAC decomposition.
Summing up, multivariate data analysis of migration samples supposes a great advantage in order to comply
with the established regulation about migrants and to decrease the false non-compliant results.
