Adsorption-assisted removal of chlorinated pollutants using a magnetically recoverable Fe3O4/CeO2 nanocomposite: kinetics, machine learning and first-principle insights.

The development of efficient and recyclable catalysts for persistent organic pollutant removal remains a major challenge in environmental nanotechnology. Herein, a mesoporous Fe3O4/CeO2 nanocomposite was synthesized and systematically characterized, confirming a particle size of (20 ± 10 nm) with a high surface area (44.6 m2 g-1), magnetic recoverability, and abundant oxygen vacancies. The catalyst exhibited excellent adsorption-assisted removal of chlorobenzene (CB) and 2,4-dichlorophenol (2,4-DCP), achieving ∼100% removal within 70 min at near-neutral pH, with ∼98% stability after five cycles. Scavenger experiments revealed that hydroxyl radicals (˙OH) are the dominant reactive species, with additional contribution from superoxide radicals (O2˙-), confirming adsorption-assisted removal pathway. Kinetic analysis followed pseudo-first order (k pfo = 0.072 min-1 for CB and 0.0399 min-1 for 2,4-DCP) behaviour and showed strong agreement with Langmuir-Hinshelwood and Runge-Kutta models (R 2 ≈ 0.99). An ANFIS model, trained on 380 datasets, accurately predicted adsorption-assisted removal under varying conditions. Density functional theory (DFT) calculations demonstrated that the Fe3O4/CeO2 (1 : 1) system possesses the most negative binding energy, indicating superior interfacial stability. PDOS analysis revealed enhanced overlap of Fe 3d, Ce 4f, and O 2p orbitals. This study highlights a synergistic experimental-theoretical-AI approach for advanced environmental remediation.