AuNP decorated aegle marmelos leaf as SERS substrate for trace detection of antibiotics and machine learning based classification

Abstract

Surface-enhanced Raman spectroscopy (SERS) has emerged as a reliable molecular spectroscopic technique for trace detection of chemical and biological samples. Present study illustrates a new SERS platform which has been obtained through surface adsorption of gold nanoparticles (AuNP) on a microscopically roughened surface of aegle marmelos (AM) leaf. The micro-structured patterns of the AM leaves promote the generation of hotspot regions for the surface deposited AuNPs thus, aids in electromagnetic enhancement for the scattered Raman signals from the sample. For the proposed SERS platform, with rhodamine6G (R6G) as an analyte, the limit of detection (LoD) was found to be 0.88 nM. The applicability of the designed SERS was realized through detection and quantification of two commonly used antibiotics- Ceftriaxone (CEFTR) and Ceftiofur sodium (CEF-Na) residues from cow milk samples. Furthermore, a dimensionality reduction method known as principal component analysis (PCA) and an optimal machine learning-based model were built to categorize the analytes in the milk samples. The suggested machine learning model's classification accuracy was found to be 94%.
Keywords: AuNPs, LSPR, Raman scattering, SERS, antibiotics, Ceftiofur sodium, Ceftriaxone, Machine learning