Automated Machine-Learning-Driven Analysis of Microplastics by TGA-FTIR for Enhanced Identification and...

Area of Science:

• Environmental Science
• Analytical Chemistry
• Materials Science

Background:

• Microplastics are widespread environmental contaminants requiring efficient quantification and identification methods.
• Thermogravimetric analysis coupled with Fourier transform infrared (TGA-FTIR) spectroscopy is a powerful technique for analyzing microplastics.
• Current TGA-FTIR applications for microplastic analysis are underutilized due to challenges in data interpretation.

Purpose of the Study:

• To develop automated data analysis routines for identifying plastic components from TGA-FTIR data.
• To enhance microplastic identification by integrating machine learning classification techniques.
• To enable streamlined qualitative and semiquantitative assessment of microplastic content.

Main Methods:

• Creation of a dedicated spectral library for gas-phase FTIR spectra of polymers.
• Development of a custom spectral matching algorithm for polymer identification.
• Application and evaluation of machine learning classifiers (k-nearest neighbor, random forest, support vector classifier, multilayer perceptron) using synthetic datasets.
• Correlation of identified polymer types with mass loss data from thermogravimetric analysis.

Main Results:

• Machine learning techniques demonstrated precise and unambiguous identification of plastic polymers.
• The developed approach successfully identified polymers from complex TGA-FTIR datasets.
• The integration of ML classifiers outperformed a custom spectral matching algorithm.
• The method provides both qualitative identification and semiquantitative analysis of microplastics.

Conclusions:

• Automated data analysis routines, particularly those employing machine learning, significantly improve microplastic identification from TGA-FTIR data.
• This approach offers a streamlined and effective method for assessing microplastic composition and quantity in environmental samples.
• The developed techniques enhance the utility of TGA-FTIR for microplastic research and environmental monitoring.