Rational Design of Highly Fluorescent N-Doped Carbon Nanoparticles for Hg2+ Ion Sensing via Machine Learning and Computational Analysis

A novel fluorescent sensor was developed using diphenylcarbazone-ligand-functionalized carbon nanoparticles for Hg2+ ion detection. The carbon nanoparticles were synthesized by the hydrothermal method using vitamin C as the carbon source and functionalized with an Hg2+ ion-selective diphenylcarbazone organic ligand via a Schiff base reaction. The as-synthesized carbon nanoparticles exhibited excellent water solubility, a uniform size of 10-12 nm, and a fluorescence quantum yield of 8.4%. The fluorescence intensity was quenched by Hg2+ ions due to the formation of a nonfluorescent complex, resulting in high sensitivity and selectivity in detecting Hg2+ ions. The limit of detection was calculated to be 2.5 nM, indicating high sensitivity and high selectivity for Hg2+ among 12 environmentally relevant metal ions. The formation of a nonfluorescent complex compound was also investigated by the ground-state geometry of the probable complex formation using DFT calculations. This suggested that the mode of interaction between Hg2+ and the functionalized carbon nanoparticles was preferred through the N atom of the DPC ligand rather than the O atom of the -COOH group on the functionalized carbon nanoparticles. A machine learning study with limited data sources, based on carbon source, methodology, limit of detection, and linear range, revealed that the formulation of the performance score is logically meaningful. Furthermore, the electron-transfer process from the functionalized carbon nanoparticles to Hg2+ was investigated using cyclic voltammetry, along with a visualized DFT-calculated electronic energy-level diagram of the frontier molecular orbitals. The practical applicability of the functionalized carbon nanoparticles as a fluorescent probe for Hg2+ ion detection was successfully demonstrated in spiked lake water and tap water using real-sample analysis.