AI-assisted detection and degradation of antimicrobial drug in aquatic environment based on Bi-functional Bi2MoO6/β-cyclodextrin functionalized activated carbon

Abstract

In this work, we developed a dual-functional platform by integrating β-cyclodextrin (βCD) functionalized activated carbon (AC) with orthorhombic bismuth molybdate (Bi₂MoO₆) for the electrochemical detection and photocatalytic degradation of metronidazole (MTZ). To enable smart sensing, a neural network model optimized using Bayesian hyperparameter tuning was employed to accurately estimate MTZ concentration from peak current values, thereby achieving precise data-driven quantification. The composite material offers abundant electroactive sites and enhanced electron transfer kinetics, leading to efficient MTZ reduction at the electrode surface. The proposed system exhibited an excellent sensitivity (72.21 μAμM⁻¹cm⁻²) and a broad linear detection range (0.05–22.5 μM) with a low detection limit of 0.2 nM. It also demonstrated high selectivity, repeatability, reproducibility, and stability. Further, the developed neural network model displayed high prediction accuracy in estimating MTZ concentration, as evidenced by the low root mean square error (0.18), mean absolute error (0.13), and high R² value (0.99). Ultimately, the successful MTZ detection in environmental samples, including aquaculture water, pond water, river water, and irrigation canal water with satisfactory recovery rates was realized. Additionally, the developed composite material aslo showed superior photocatalytic degradation activity toward MTZ under visible light irradiation. These results highlight the proposed system's strong potential for dual applications in environmental monitoring and photocatalytic remediation of aquatic pollutants.