An innovative machine learning approach to predict electronic properties and multifunctional performance of Ag–sepiolite nanocomposites

The present work reports a facile, eco–friendly synthesis of an Ag–sepiolite nanocomposite utilizing Mentha aquatica leaf extract, serving simultaneously as a reducing and stabilizing agent. The pristine sepiolite and the resulting nanocomposite were comprehensively characterized via XRD, FT–IR, SEM, EDX, TEM, and XPS, confirming the successful loading of silver nanoparticles (AgNPs) onto the fibrous sepiolite matrix with uniform dispersion and evidence of metal–support interaction. The catalytic activity of Ag–sepiolite was evaluated for the reduction of methylene blue (MB) in the presence of NaBH₄, revealing rapid pseudo-first-order kinetics with an apparent rate constant of 0.60 min⁻¹ and over 98% degradation within 4 min under optimized conditions. The catalyst exhibited excellent reusability, maintaining an efficiency of over 88% across five cycles. In addition, the Ag–sepiolite nanocomposite demonstrated significant antibacterial activity against Gram-positive (Staphylococcus aureus, Bacillus subtilis) and Gram-negative (Escherichia coli, Salmonella typhi) bacteria, as well as potent antifungal activity against Candida albicans . Its antioxidant potential was confirmed through 2.2-Diphenyl-1-picrylhydrazyl (DPPH) and Ferric Reducing Antioxidant Power (FRAP) assays, with significantly higher radical scavenging and ferric-reducing capacity than sepiolite or plant extract alone. Furthermore, machine learning approaches were utilized to predict electronic characteristics, including energy per atom and optical band gap, using structural descriptors derived from X-ray diffraction data, providing insights into the enhanced reactivity and catalytic efficiency of Ag–sepiolite. The work highlights the synergistic effect of AgNPs and sepiolite, establishing a sustainable approach for environmental remediation and catalytic applications.