Related Experiment Video
Updated: Dec 2, 2025
14:10
Separation and Identification of Conventional Microplastics from Farmland Soils
Published on: March 21, 2025
2.7K
An effective method for the rapid detection of microplastics in soil
Yanhui Li1, Jiangjun Yao2, Pengcheng Nie3
1College of Information Engineering, Tarim University, 1188 Junken Avenue, Alar, 843300, China.
Chemosphere
|November 4, 2020
Summary
A new Terahertz (THz) spectroscopy method combined with a Least Squares Support Vector Machine (LS-SVM) model can rapidly detect microplastic pollution in soil. This approach offers a faster alternative to existing methods for assessing environmental microplastic contamination.
Area of Science:
- Environmental Science
- Analytical Chemistry
- Spectroscopy
Background:
- Microplastic accumulation in soil poses risks to environmental security and human health.
- Current methods for microplastic detection in soil are often time-consuming and cumbersome, hindering research progress.
- A standardized, efficient method is needed to accurately identify and quantify soil microplastics.
Purpose of the Study:
- To explore the spectral characteristics of soil contaminated with low-density polyethylene (LDPE) and polyvinyl chloride (PVC) microplastics in the 0.6-1.8 THz band.
- To develop a novel method for rapid detection of microplastic pollution levels in soil using Terahertz (THz) spectroscopy and a Least Squares Support Vector Machine (LS-SVM) model.
- To compare the effectiveness of local and multisource LS-SVM models for predicting microplastic pollution across different regions.
Main Methods:
- Investigated the THz spectral characteristics of soil mixed with LDPE and PVC microplastics.
- Established a Least Squares Support Vector Machine (LS-SVM) model utilizing THz spectral data for microplastic detection.
- Developed and evaluated both local (region-specific) and multisource (combined training data) LS-SVM models.
Main Results:
- The local LS-SVM model showed good performance for LDPE (R=0.9833, RMSE=0.0050) and PVC (R=0.9686, RMSE=0.0071) but was limited to local regions.
- The multisource LS-SVM model, using combined training data, significantly improved prediction accuracy for both LDPE (R=0.9895, RMSE=0.0007) and PVC (R=0.9831, RMSE=0.0009) across areas.
- Terahertz spectroscopy combined with the LS-SVM model demonstrated high efficacy in predicting soil microplastic pollution levels.
Conclusions:
- The developed Terahertz-LS-SVM model provides an effective and rapid approach for assessing soil microplastic pollution.
- The multisource LS-SVM model enhances the model's applicability and accuracy for predicting microplastic contamination across diverse geographical regions.
- This technique offers a promising solution to overcome the limitations of existing methods in soil microplastic analysis.
