Deep machine learning-assisted MOF@COF fluorescence/colorimetric dual-mode intelligent ratiometric sensing platform for sensitive glutathione detection.

Glutathione (GSH) levels have been linked to aging and the pathogenesis of various diseases, highlighting the necessity for the development of sensitive analytical methods for GSH to facilitate disease diagnosis and treatment. In this study, we synthesized a novel core-shell material, UiO@TBTA, by in-situ growing TFPB-TAPA COF on UiO-66-NH through a Schiff base reaction. The resulting composite capitalize on the advantages of both materials, demonstrating excellent stability, large specific surface area, and abundant active functional groups while preserving superior crystallinity. Notably, this strategy effectively reduces the occurrence of aggregation-caused quenching (ACQ) in COFs. Due to the inner filter effect and hydrogen bonding interactions between UiO@TBTA and GSH, a specific ratiometric fluorescence detection of GSH was achieved in the range of 0.1-7 μM, with a limit of detection (LOD) of 0.0685 μM. In addition, due to the sensitive color change of the sensing material from orange to black caused by GSH, a proportional colorimetric sensing strategy has also been proposed, enabling the detection of GSH within the range of 1-200 μM. What's more, two intelligent artificial neural networks models were constructed with the help of machine learning that can quickly, accurately, and sensitively determine the concentration of GSH based on fluorescence images and color photographs respectively. Our work represents the first study utilizing MOF@COF composite for the multimodal detection of GSH, thus providing a novel strategy for the multimodal detection of the target analyte. Prospectively, the construction of the fluorescence/colorimetric dual-mode intelligent ratiometric sensing platform using deep machine learning holds great promise for real-time monitoring.