Smart Visual Fluorescence Sensing of Tetracycline Using N-Doped Nb 2 C Quantum Dots and CNN-Based Regression Quantification

MXene quantum dots (QDs) have emerged as a promising class of zero-dimensional nanomaterials, attracting growing attention due to their remarkable physicochemical properties. Niobium carbide QDs (Nb2C QDs), derived from two-dimensional (2D) Nb2CTx, is a rising material with great potential in optical applications. Herein, nitrogen-doped Nb2C QDs (N–Nb2C QDs) were synthesized via sonication-assisted hydrothermal treatment. The obtained N–Nb2C QDs exhibit blue fluorescence with excitation-dependent emission. We developed a fluorescent sensor enabling N–Nb2C QDs for visual and quantitative detection of tetracycline (TC) antibiotics. The blue fluorescence of N–Nb2C QDs can be quenched by TC through static quenching and inner filter effect (IFE). The synergistic effect of both static quenching and IFE caused by TC makes the developed fluorescence sensor have a linear range for TC from 0.25 to 6.50 μM with a detection limit of 30.2 nM and a significant fluorescence color change from blue to cyan-green. Due to the simplicity and recognizable color change, a machine learning based approach was employed to determine the TC concentration from a visual fluorescence image. A convolutional neural network (CNN) architecture was used for the regression analysis. The model shows a strong predictive performance, achieving a coefficient of determination (R2) of 0.966, which reflects a high level of accuracy in estimating the TC concentrations.