Antioxidative and Robust Fluorescent Hydrogel for Deep Learning-Assisted Non-Destructive Detection of Hg²

Flexible fluorescent hydrogels offer great potential for on-site heavy-metal sensing, but their practical use is often limited by a fundamental trade-off: materials that provide strong fluorescence responses typically suffer from poor stability in oxidative environments, while hydrogel matrices with good robustness or adhesion rarely support reliable fluorescence signaling. Here, we introduce a phycocyanin-polyacrylamide (PC-PAAM) hydrogel that resolves this conflict by integrating high mechanical flexibility and strong underwater adhesion with stable red fluorescence and intrinsic antioxidative protection. PC acts simultaneously as a fluorescent cross-linker and ROS-scavenging component, enabling the hydrogel to maintain signal integrity while undergoing rapid fluorescence quenching upon Hg²⁺ coordination. The detection limit of the sensor achieves 6.21 nM. It demonstrates conformal contact on irregular biological surfaces, enabling nondestructive detection on fish. Coupled with a deep-learning model trained on hydrogel fluorescence images, the system further enables accurate, portable, and visual quantification of Hg²⁺ contamination using smartphone imaging. This work provides a flexible, adhesive, and intelligent sensing platform for environmental and food-safety monitoring.