Machine learning-driven sensor array based on luminescent metal–organic frameworks for simultaneous discrimination of multiple anions

Due to the high correlation of anions in waters to environmental quality and human health, thus there is urgent need for developing simple and effective sensors to discriminate multiple anions. Herein, a machine learning-assisted fluorescent sensor array based on two luminescent metal–organic frameworks (LMOFs, UiO-66-NH2 and UiO-66-OH) was developed for simultaneous discrimination of five anions (F−, PO43−, ClO44−, NO3−, and SO42−). Wherein, UiO-66-NH2 and UiO-66-OH were designed by anchoring 2,5-diaminoterephthalic acid and 2,5-dihydroxyterephthalic acid on UiO-66, respectively, which exhibited blue and green fluorescence emission, possessing good fluorescence property. Interestingly, the anions could effectively enhance the fluorescence intensity of UiO-66-NH2 and UiO-66-OH to generate diverse fluorescence responses and unique fingerprints, which could be utilized to develop a fluorescence sensor array for the rapid identification of five anions. Under the optimized conditions, the proposed sensor array showed good performance for identifying multiple anions and their mixtures with satisfactory sensitivity. More importantly, the integration of machine learning algorithm and sensor array has successfully achieved accurate identification and prediction of five anions in real water samples, affirming its practicability in actual samples. Our findings provided a promising tool for detecting multiple anions, and inspired potentials of the combination of sensor arrays and machine learning algorithm for pollution control in real waters.