Deep learning-assisted Ce/Zr-MOF nanozyme hydrogel sensor for colorimetric/photothermal dual-mode detection of glyphosate.

The extensive application of glyphosate (GLY) in modern agriculture has raised increasing concerns regarding environmental contamination and food safety. In this work, a dual-mode hydrogel sensing platform was constructed based on Ce/Zr-MOF with intrinsic peroxidase-like activity. The synthesized Ce/Zr-MOF efficiently catalyzed the oxidation of TMB into blue oxTMB, enabling colorimetric detection while simultaneously generating a photothermal response. Notably, GLY strongly coordinated with the Ce3+/Zr4+ active centers, effectively inhibiting the catalytic reaction and leading to concentration-dependent attenuation of both color intensity and photothermal heating. Owing to this inhibition mechanism, the hydrogel sensor achieved dual-mode detection of GLY over a wide linear range of 0-500 μM, with low detection limits of 0.13 μM (colorimetric) and 0.08 μM (photothermal). This platform also demonstrates excellent selectivity and long-term stability. Furthermore, a deep learning-based dual-modal concentration prediction network (DMCPNet) was developed to enable rapid and accurate determination of GLY concentrations. The proposed strategy offers reliable on-site detection of GLY and demonstrates the potential of multifunctional hydrogel sensors for environmental and food safety monitoring.