Composite Nanotube Arrays for Pesticide Detection Assisted with Machine Learning Based on SERS Effect

This study presents a streamlined approach for the large-scale fabrication of composite nanotube arrays (CNA) with a diameter of approximately 330 nm and a wall thickness of ∼60 nm via reactive ion etching. By conducting a thorough investigation into the entire formation process and various influencing factors, we have identified the critical role played by the reaction products involving the silica layer, the aluminum sample stage, and fluoride ions in the nanotube formation process. Moreover, surface-enhanced Raman scattering (SERS) analyses were performed on the resulting substrates, revealing outstanding SERS performance. Exploiting this SERS capability, we successfully detected Raman signals from four distinct pesticides: carbendazim, carbaryl, thiram, and thiabendazole. Additionally, we integrated the SERS technique with machine learning algorithms to classify and identify these four pesticide molecules, achieving an accuracy rate exceeding 95% for each algorithm. Consequently, this platform holds promise for applications in the detection of pesticide contamination and other related fields.