Fluorescent-BasedSensing for Tetracycline AntibioticsUsing Sodium Carboxymethylcellulose and Machine Learning-AssistedRecognition

Abstract Uncontrolled residues of tetracycline antibiotics (TCs) in food disrupt human gut homeostasis and even trigger hepatic disorders, thereby highlighting the critical necessity for detecting trace TCs. Herein, two nitrogen-incorporated carbon nanodots (N1–36CD and N2–24CD) were prepared through a hydrothermal route with sodium carboxymethylcellulose serving as the precursor. These materials were subsequently employed to achieve ultrafast (10 s) and highly efficient discrimination of tetracycline, doxycycline, and oxytetracycline in milk matrices. The synthesized carbon nanoparticles typically measure between 2 and 4 nm in diameter, confirming their nanoscale dimensions. Following excitation at 342 nm, they produce a faint emission band centered at 400 nm. Interestingly, introducing tetracyclines within a 2–20 μM range proportionally boosts the green fluorescence at 520 nm. The enhanced fluorescence response resulted from the synergistic effects of internal filtering and analyte binding. N2–24CD showed improved sensitivity compared with N1–36CD. Specifically, the limits of detection (LODs) of N2–24CD toward tetracycline, oxytetracycline, and doxycycline are 47 nM, 15 nM, and 154 nM, respectively, whereas the corresponding LOD values of N1–36CD for these three antibiotics are 71 nM, 21 nM, and 181 nM, respectively. A machine learning-assisted fluorescence sensing platform was subsequently established to enable sensitive and economical monitoring of tetracycline residues in complex milk samples, highlighting its potential application in food safety analysis.