Machine vision-assisted histidine-regulated Cu-BDC-NH2 nanozyme with enhanced dual enzyme-mimicking activities for precise intelligent identification of sulfides at neutral pH.

Nanozymes with dual enzyme-like behaviors have broad application prospects in the field of biosensing, but their catalytic activity is usually restricted to acidic conditions. Therefore, developing bifunctional nanozymes with high activity under neutral pH conditions is particularly important. Here, a histidine-functionalized 2-aminoterephthalic acid-Cu (Cu-BDC-NH2@His) nanozyme was fabricated via a defect engineering approach. Relative to Cu-BDC-NH2, the obtained Cu-BDC-NH2@His nanozyme exhibited higher peroxidase-like (POD) and laccase-like (LAC) activities under neutral pH conditions. Given that different sulfides exerted divergent regulatory impacts on the dual enzyme-like behaviors of Cu-BDC-NH2@His, the array units were constructed using three unique response signals from the nanozyme sensing system (POD-370, POD-652 and LAC-510), and a "fingerprint" analysis spectrum of sulfides was further established. In addition, the robust discrimination model with simultaneous concentration- and matrix-independent performance was constructed by combining machine learning (ML) with the array system to classify multiple sulfides spiked in three food samples. The sulfide identification accuracy of the array was increased from 55.17% to 100%. Finally, a visualized on-site intelligent sensing platform was established using chromaticity imaging and deep learning algorithms to further improve the practicality of sulfide identification. This study not only provides a novel approach to overcome the pH limitation of nanozymes, but also offers an innovative intelligent sensing paradigm for the analysis of complex food matrix samples under neutral pH conditions.