One-Pot Synthesized Multifunctional Zn-Mof/Hof Heterostructure Sensor Array Assisted by Machine Learning for Efficient Capture, Target Discrimination and Optosmart Sensing of Doxycycline Analogs

Abstract

The ideal multifunctional platform that combines the capabilities of effective capture, sensitive detection, and accurate identification of doxycycline analogs (DCs) remains a serious challenge for ensuring the environment and food security. This work constructs heterostructure Zn-MOF/HOF asynchronous response fluorescence sensor using a multicomponent one-pot method for high-efficiency capturing and sensitive detecting DCs. Metal nodes and functional groups in Zn-MOF/HOF provide sites for specifically recognizing and sensitizing DCs that induce asynchronous response with blue/green fluorescence emission. Fluorescence spectra of Zn-MOF/HOF show characteristic differences due to different spatial conformations and substituents of DCs. Machine learning-assisted Zn-MOF/HOF fluorescent sensing array accurately discriminates DCs with a high precision of 100%. An exceptional adsorption capacity of DCs up to 569.00 mg/g realizes the effective pre-enrichment of DCs, improving the sensitivity of the Zn-MOF/HOF sensor. The limits of detection of the Zn-MOF/HOF sensor are as ultra-low as 2.2 nmol/L. Satisfactory recoveries of 91.78%–113.16% are obtained for detecting DCs in real-world water and food samples. A portable optosmart sensing system integrating the Zn-MOF/HOF sensor and smartphone realizes visual quantitation and on-site monitoring DCs. This work innovatively reveals the great potential of Zn-MOF/HOF heterostructure as a multifunctional platform for simultaneous capture, identification, and sensing of emerging contaminants.

Keywords: Metal-organic framework, hydrogen-bonded organic framework, fluorescence detection, antibiotic, Machine learning