AuNP Decorated PTFE Filter Paper as SERS Platform for Trace Detection of Melamine in Milk and Machine Learning Based Classification
摘要
Surface-enhanced Raman spectroscopy (SERS) is a highly sensitive analytical method capable of detecting analytes at trace concentrations. In SERS, the presence of intense localized surface plasmon resonance (LSPR) fields close to metallic nanostructures leads to a substantial amplification of Raman scattering signals, enabling trace detection. In this work, an AuNP-functionalized polytetrafluoroethylene (PTFE) filter paper is demonstrated as reliable SERS substrate for the trace detection of melamine in farm milk samples. The proposed substrate was fabricated by drop-casting chemically synthesized AuNPs onto PTFE paper. The sensing performance was first validated using two widely employed Raman reporter molecules, malachite green (MG) and rhodamine-6G (R6G). Following successful evaluation with standard dyes, the applicability of the platform was extended to trace-level detection of melamine in laboratory-prepared samples. The limit of detection (LoD) of the developed SERS platform was determined to be 9.5 nM using R6G as a test analyte. The fabricated SERS platform further enabled the detection of melamine in milk samples at concentrations as low as 0.1 ppm. Finally, the fabricated platform was applied to real milk samples, where a machine-learning-assisted classification approach was employed to distinguish between melamine-present and melamine-absent samples, achieving an overall classification accuracy of approximately 92%.