Microplastic adulteration in homogenized fish and seafood - a mid-infrared and machine learning proof of concept

Stephanie Owen, Samuel Cureton, Mathew Szuhan, Joel McCarten, Panagiota Arvanitis, Max Ascione, Vi Khanh Truong, James Chapman, Daniel Cozzolino

Research output: Contribution to journalArticlepeer-review

5 Citations (Scopus)


The objective of this study was to assess the ability of utilizing attenuated total reflection mid-infrared (ATR-MIR) spectroscopy in combination with machine learning techniques to classify the presence of different types of microplastics in artificially adulterated fish and seafood samples. Different polymers namely poly-vinyl chloride (PVC), polycarbonate (PC), polystyrene (PS), polypropylene (PP) and low (LDPE) and high-density polyethylene (HDPE) were mixed with homogenized fish and seafood samples. Homogenized samples were analyzed using MIR spectroscopy and classification models developed using machine learning algorithms such as partial least squares discriminant analysis (PLS-DA). The results of this study revealed that it was possible to identify between adulterated and non-adulterated samples as well as the different microplastic types added to the homogenized samples using ATR-MIR spectroscopy. This study confirmed the ability of combining machine learning methods with ATR-MIR spectroscopy to directly analyze microplastic adulteration in fleshy foods such as fish and seafood. This proof-of-concept study can be utilized and extended to monitor the presence of plastics either in a wide range of fleshy foods or along the entire food value chain.

Original languageEnglish
Article number119985
Number of pages7
JournalSpectrochimica Acta - Part A: Molecular and Biomolecular Spectroscopy
Publication statusPublished - 5 Nov 2021
Externally publishedYes


  • Contamination
  • Fish
  • Infrared
  • Machine learning
  • Microplastics
  • Polymers
  • Seafood


Dive into the research topics of 'Microplastic adulteration in homogenized fish and seafood - a mid-infrared and machine learning proof of concept'. Together they form a unique fingerprint.

Cite this