Machine learning-enabled discrimination of tetracycline antibiotics by synergistic inner filter effect and antenna effect in Eu3+-functionalized silver nanoclusters.

Accurate discrimination of structurally similar antibiotics remains a major challenge due to the limited resolving capability of conventional sensing signals. Tetracycline antibiotics, which often coexist in complex matrices, typically generate highly overlapping responses, hindering reliable identification. Herein, a dual-emission fluorescence sensing platform based on Eu3+-functionalized silver nanoclusters (AgNCs@Eu3+) was developed to generate analyte-dependent signal patterns. Upon interaction with tetracyclines, the green emission of AgNCs is selectively attenuated through the inner filter effect, while the coordination between tetracyclines and Eu3+ activates a characteristic red emission by the antenna effect. The synergistic modulation of these two signals produces partially overlapping but distinguishable fluorescence fingerprints for different tetracycline species. To resolve these subtle differences, machine learning algorithms were employed to decode the dual-channel fluorescence responses. The optimized model enables accurate classification of tetracycline, oxytetracycline, doxycycline, and chlortetracycline, with an accuracy up to 0.99. The platform further demonstrates robust performance in pharmaceutical, milk, and environmental water samples, maintaining reliable discrimination despite matrix interference. This work establishes a strategy that couples mechanism-defined dual-emission sensing with data-driven analysis, providing a practical route for the identification of structurally similar analytes in complex systems.