A surface enhanced Raman scattering–machine learning platform for monitoring anionic pollutants in water and human urine: Toward environmental and public health applications

Monitoring anionic pollutants in aquatic environments and biological fluids is essential for safeguarding environmental safety and human health. However, the detection of anions via surface-enhanced Raman scattering (SERS) remains challenging due to their negative charge and weak interaction with traditional substrates. Herein, we present a universal SERS platform based on calcium-induced aggregation of silver nanoparticles (Ag@BOCNPs) that enables sensitive, reproducible, and label-free detection of eight anions—including fluoride, sulfate, phosphate, and nitrate—within complex aqueous and biological matrices. The platform exhibits exceptional performance for eight ions detection, achieving ultralow limits of detection (LODs, S/N = 3) of 0.001 mg/L and limits of quantification (LOQs, S/N = 10) of 0.003 mg/L. The method demonstrates high stability and excellent anti-interference capability, allowing direct fluoride detection in human urine without pretreatment. By integrating SERS spectral data with machine learning algorithms (KNN, RF, SVM, etc.), accurate classification of anion species and quantification of fluoride exposure levels were achieved. This integrated analytical strategy not only addresses a key limitation in anion detection, but also offers a powerful tool for environmental monitoring and human exposure assessment, particularly in regions affected by water contamination and fluorosis risk.