A novel ultra-sensitive semiconductor SERS substrate V5S4 nanopompons for the specific detection of antibiotics with AI technology

Misuse and residues of antibiotics in food seriously threaten human health, which urges us to develop ultra-sensitive semiconductor surface-enhanced Raman scattering (SERS) substrates to achieve the direct and ultra-sensitive detection of specific various antibiotics. Here, we innovatively propose polyvinylpyrrolidone K30 as surfactants to regulate the liquid–liquid interface growth and successfully synthesize a novel SERS substrate V 5 S 4 nanopompons (NPPs) formed by the cross-growth nanosheets, which exhibits a limit of detection (LOD) of 10 −11 M for Rhodamine 6G (R6G) molecules with a high enhancement factor of 1.9 × 10 8 . To the best of our knowledge, this ultra-high SERS sensitivity of V 5 S 4 NPPs is superior to most of reported pure semiconductor SERS substrates, mainly due to the synergistic effect of photo-induced charge transfer resonance from the valence band of V 5 S 4 to the LUMO energy level of R6G and the molecule resonance originating from the aromatic C–C stretching in R6G. Benefitting from the excellent SERS sensitivity of V 5 S 4 NPPs, the sensitive and specific detection with LODs of 10 −7 M, 10 −8 M, and 10 −7 M and accurate identification with the accuracy of 97.5% for common antibiotics of ciprofloxacin, tetracycline, and chloramphenicol, respectively, are achieved using a newly developed convolutional-residual neural networks models of AI processing method. This research not only enriches the material reservoir of the conventional semiconductor SERS substrates but also provides an efficient detection and identification method for monitoring antibiotic residues in milk to ensure the food safety.