Related Experiment Video
Updated: Jul 8, 2026

Sampling, Sorting, and Characterizing Microplastics in Aquatic Environments with High Suspended Sediment Loads and Large Floating Debris
Published on: July 28, 2018
Quantification of PP, PE, and PET microplastics in Lake water by Raman spectroscopy combined with PLS regression
Wenmin Zhao1, Xin Wang1, Dexiang Wang1
1Key Laboratory of Synthetic and Natural Functional Molecular of the Ministry of Education, College of Chemistry & Material Science, Northwest University, Xi'an 710127, China.
Abstract:
Microplastics are pervasive contaminants in aquatic ecosystems, posing significant ecological and health risks. Their accurate quantification in complex water matrices remains challenging, requiring advanced analytical techniques. This study develops an integrated analytical approach using Raman spectroscopy and chemometrics for the simultaneous detection and quantification of trace-level polypropylene (PP), polyethylene (PE), and polyethylene terephthalate (PET) microplastics in lakes. A dedicated sample preparation protocol was established to overcome microplastic hydrophobicity and low abundance, employing optimized solvent dispersion and enrichment to enable highly accurate quantification via Partial Least Squares (PLS) regression. The resulting models demonstrated exceptional performance: the Smooth-WT-CARS-PLS (Smooth-Wavelet Transform-Competitive Adaptive Reweighted Sampling-Partial Least Squares) model for PP achieved a coefficient of determination for prediction (R2P) of 0.9807 and a mean relative error of prediction (MREP) of 0.0599; the Smooth-CARS-PLS (Smooth-Competitive Adaptive Reweighted Sampling-Partial Least Squares) model for PE attained an R2P of 0.9899 with an MREP of 0.0880; and the WT-Smooth-CARS-PLS (Wavelet Transform-Smooth-Competitive Adaptive Reweighted Sampling-Partial Least Squares) model for PET yielded an R2P of 0.9822 and an MREP of 0.0968. Furthermore, the method showed low detection limits (0.31 μg/mL for PP, 0.25 for PE, 0.13 for PET), confirming its high sensitivity. Practical applicability was validated using real lake samples, and the models maintained low prediction errors when applied to different lake environments, demonstrating satisfactory robustness. This reliable framework provides an efficient solution for rapid and precise monitoring of microplastics in complex waters, showing great promise for supporting future environmental risk assessments and regulatory efforts.

