Machine Learning-Assisted 3D Envelope-Enhanced SERS Platform for Detection and Classification of Multiple Coexisting...

Area of Science:

• Environmental Science
• Analytical Chemistry
• Materials Science

Background:

• Antibiotics adsorb onto plastic surfaces, forming coexisting pollutants with amplified environmental risks.
• Existing detection methods struggle with complex mixtures and low concentrations of these pollutants.

Purpose of the Study:

• To develop a sensitive and accurate platform for detecting and quantifying coexisting antibiotic pollutants on plastic surfaces.
• To integrate surface-enhanced Raman spectroscopy (SERS) with machine learning (ML) for enhanced pollutant analysis.

Main Methods:

• A three-dimensional envelope-enhanced SERS (3D EES) platform was created using three-phase interface self-assembly.
• Au@Ag plasmonic nanoparticles were organized into a hydrophobic array substrate to generate plasmonic hot spots.
• Machine learning models (SNE-SVM and RF-SVM) were trained on a comprehensive SERS dataset.

Main Results:

• The 3D EES strategy achieved sensitive detection of antibiotics adsorbed on polystyrene (PS) down to 250 ng/L.
• Machine learning models demonstrated high accuracy (98.07%) and Jaccard similarity (96.44%) in classifying antibiotic species and predicting adsorption levels.
• The method effectively amplified SERS signals through coconcentration and additional hot spot generation.

Conclusions:

• The developed 3D EES platform offers a robust tool for identifying coexisting pollutants in environmental samples.
• This approach provides new avenues for tracing pollutant sources and understanding their environmental fate.
• The integration of SERS and ML significantly enhances the analysis of complex environmental pollutant mixtures.