AI-assisted near-infrared ratiometric fluorescent sensor for ultrafast and portable visual detection of Bacillus anthracis biomarkers.

Contamination by Bacillus anthracis in food and the environment poses a significant public health risk to both humans and animals, with dipicolinic acid (DPA) serving as an effective biomarker for its detection. In this work, a novel AI-assisted near-infrared ratiometric fluorescent sensing system based on rapid coordination cross-linking and multiple fluorescence response mechanisms was proposed for ultrafast and portable visual detection of Bacillus anthracis biomarkers. A near-infrared (NIR) ratiometric fluorescent probe with multiple fluorescence response mechanisms was developed for sensitive detection of DPA, using carbon dots (B-CDs) and CdTe quantum dots (QD686). Notably, the system exhibits a rapid fluorescence color transition from red to blue within 3 s when exposure to DPA. The strong metal-coordinating ability of DPA enables rapid binding with QD686, inducing nanoparticle aggregation that activates a multi-mechanism synergistic (Fluorescence resonance energy transfer-Inner filter effect, FRET-IFE) quenching effect, thereby enabling highly sensitive and selective fluorescence detection of DPA. Additionally, a portable device leveraging AI (YOLOv12 algorithm) and a smartphone was developed for DPA detection within 10 s, achieving a detection limit of 0.097 μM (S/N = 3) and demonstrating satisfactory accuracy (recovery: 93.6-116.0 %) and precision (RSD: 1.5-3.4 %, n = 5) in validation studies. This work not only provides a novel strategy for DPA sensing but also establishes a reliable protocol for rapid visual detection, showcasing its potential applications in food safety and environmental monitoring.