AI-integrated chemosensing platform for sequential nanomolar detection of Fe3+ and arsenic species via smartphone-assisted machine learning toward rapid portable environmental monitoring.

The integration of molecular colorimetric sensors with smartphone-assisted image analysis and machine learning has emerged as a next-generation sensing paradigm for environmental diagnostics. Herein, we report an artificial intelligence (AI)-integrated chemosensing platform based on a norbornene-appended rhodamine derivative (RhB-NBE) for the sequential monitoring of Fe3+ and arsenic species [As(V)/As(III))] in aqueous media. The probe was synthesized through a Schiff base condensation reaction and fully characterized using NMR, X-ray, and other spectroscopic techniques. RhB-NBE exhibited highly selective and instantaneous recognition of Fe3+ over competing metal ions, including Fe2+, through a spirolactam ring-opening mechanism that generated a discrete colorimetric response. The resulting RhB-NBE + Fe3+ ensemble served as a secondary sensing platform for arsenite and arsenate via a cation-displacement process, producing a visible color transition from pink to colorless with a low detection limit of 2 nM (below the World Health Organization critical limit). Practical applicability was demonstrated using portable paper-based sensing strips, and successful quantification of arsenic species in environmental water samples with excellent recovery values. Despite the rapid advancement of AI-assisted chemical sensing, integration of machine learning with colorimetric arsenic detection remains largely unexplored due to challenges in sensor reproducibility, image standardization, and reliable concentration classification. Herein, we combine a selective colorimetric chemosensor with AI-based image analytics for rapid field monitoring. Smartphone images were analyzed using GNB, SVM, and CNN models trained on 1000 RGB datasets; SVM achieved 97.0% accuracy for equipment-free arsenic screening.