Chemical Variability of Recycled PET Bottles Driven by Packaging Inputs and Processing History: A Marker-Based Multivariate Approach

Abstract

Recycled polyethylene terephthalate (rPET) intended for food-contact applications contains complex mixtures of non-intentionally added substances (NIAS) originating from polymer degradation, recycling, packaging-related inputs, and external contamination. Although non-targeted analytical techniques can characterise these complex chemical fingerprints, a standardised framework for their systematic interpretation is lacking. This study introduces a literature-informed, pattern-based framework for interpreting Py-GC/MS fingerprints of commercial PET bottles. Twenty-two commercial PET beverage bottles were analysed by pyrolysis–gas chromatography–mass spectrometry (Py-GC/MS), and detected compounds were organised into four diagnostic profile groups: PET transformation products (C-PET), packaging-associated compounds (E-PS/PVC/FCM), mixed-origin compounds (M-MIX), and condensed aromatic structures (A-AROM). Principal component analysis and a composite Quality Index (QI) were used to evaluate chemical variability. PET transformation products formed a consistent chemical baseline across all samples, whereas chemical variability was primarily associated with packaging-associated compounds and condensed aromatic structures. Bottles containing 100% rPET spanned the full range of chemical profiles, suggesting that the observed variability may be influenced more strongly by feedstock quality, packaging-related inputs, and processing history than by declared recycled content alone. The proposed framework enables a systematic interpretation of Py-GC/MS fingerprints through diagnostic chemical patterns rather than unique source attribution of individual compounds. The proposed framework provides a scalable methodology for comparative evaluation of recycled PET. It highlights that improving chemical quality requires controlling feedstock purity, packaging-related contamination, and processing conditions rather than relying solely on recycled content.