Laser Induced Breakdown Spectroscopy (LIBS) has emerged as a promising analytical technique for the plastic industry, offering the capability to measure plastic materials directly, even in-line, with minimal or no sample preparation. This approach has gained significant attention due to the increasing necessity for quality control in the recycling process, ensuring the purity and composition of recycled plastics [1,2]. This study explores the application of LIBS for analyzing various plastic types, including polypropylene (PP), copolymer polyethylene/polypropylene (copoPE+PP), polyethylene (PE), polystyrene (PS), and polyethylene terephthalate (PET), in the form of recycled flake forms.

The spectra of plastic samples were analyzed, highlighting key elements such as carbon (C), oxygen (O), nitrogen (N), and hydrogen (H), as well as common contaminants such as copper (Cu), iron (Fe), titanium (Ti), sodium (Na), and potassium (K). Principal Component Analysis (PCA) was employed to assess data clustering, revealing distinct patterns for PP flakes and outliers, based on an average of 125 spectra per sample. Further, multiple machine learning algorithms were applied for polymer-type classification, with varying training datasets, and the results of the prediction scatter are discussed. Additionally, the detection of chlorine, notably from CaCl bands adjacent to CaO, and the presence of harmful elements such as barium (Ba) in granulate and recycled shreds were successfully identified, which is critical for ensuring the safety and quality of recycled plastics. This work demonstrates the potential of LIBS as a robust tool for efficient and non-destructive, respectively micro-destructive monitoring in plastic recycling processes, contributing to higher standards in material quality assurance and contamination detection.