Machine learning-assisted magnetic nanomotors for the identification and degradation of organic pollutants

Trace organic pollutants in aquatic systems present significant environmental and health risks, yet existing methods fail to achieve targeted effective degradation and real-time in situ monitoring. To address these challenges, we developed a magnetically nanomotor by integrating Au/R-Fe₃O₄ magnetic components into a hollow Au@TiO₂ core-shell structure. This design enables simultaneous surface-enhanced Raman scattering (SERS) detection and photocatalytic degradation of contaminants. The synergistic architecture of the Au@TiO₂ hollow structure combined with R-Fe₃O₄-supported Au nanoparticles enhances localized surface plasmon resonance, allowing ultra-sensitive detection of target pollutants at concentrations below 10^-11 M while maintaining SERS stability within complex water matrices. Under visible light irradiation, the nanomotor achieved pollutant degradation within 90 min, demonstrating sustained catalytic efficiency over multiple cycles. This performance is attributed to its magnetically enhanced interfacial contact and full-spectrum light utilization. To validate the dynamics of the photodegradation process, machine learning algorithms were employed alongside mechanistic insights obtained from electron paramagnetic resonance (EPR) and quenching experiments. Furthermore, the nanomotor demonstrates potent antibacterial activity with negligible ecotoxicity, confirming its environmental compatibility without secondary pollution risks. This work marks a significant advancement in magneto-responsive photocatalytic nanomotors, offering a dual-action strategy for real-time monitoring and precise remediation of low-concentration organic pollutants in aquatic environments.