Coding and tracking contaminants of food contact materials by wavelet analysis
Résumé
The management and assessment of chemicals in food contact materials (FCM) are mandatory according to the European framework regulation (EC) 1935/2004. Recent crises (bisphenol A, printing inks, mineral oils) confirm emerging risks, which have been insufficiently anticipated. The generalization of recycled materials with plastic or cellulosic origin for food contact raises concerns about their sources and decontamination levels. Leachable content of FCMs is conventionally recovered by extraction and subsequent analysis by GC-MS, LC-MS, and 1 H-NMR. Their interpretation is particularly complex for non-intentionally added substances (NIAS), such as by-and cross contamination products. This study proposes a new processing technique to tackle the vast information acquired during non-targeted analyses, including the forest of peaks. The method has been tested on recycled plastic extracts analyzed by GC-MS. Continuous wavelet transform (CWT) tags ionized fragments and encodes each ion signal in monotonic sub-signals at different scales. The entire chromatogram is translated into text sequences, which can be aligned and compared across samples and databases. Contaminants are easily detected as insertions or substitutions in sequences. A back-transformation is proposed to reconstruct the likely signal associated with the changes and offer compatibility with conventional tools in chemometrics.
Origine : Fichiers produits par l'(les) auteur(s)