Artificial neural network modeling for adsorption of Congo red and methylene blue dye removal using pineapple waste-mediated magnetic nanoparticles

Abstract

The effective removal of synthetic dyes from wastewater remains a significant environmental challenge. This study investigates the potential of carbonated pineapple peel waste, integrated with magnetic nanoparticles (PMNPs), for the adsorption-based removal of Congo red (CR) and methylene blue (MB) dyes. Various characterization techniques, including SEM, FTIR, and EDS, were used to analyze PMNPs before and after adsorption, while XRD, BET, VSM, and TGA were applied to assess the properties of pure PMNPs and their suitability for adsorption. The PMNPs exhibited a needle-like morphology and a high surface area of 6.836 m²/g, enhancing their dye adsorption capacity. At concentrations of 1.5 and 1.25 g/L, PMNPs achieved removal efficiencies of 93.15% for CR and 95.99% for MB. Thermodynamic analysis revealed the adsorption process to be spontaneous and exothermic. Computational modeling demonstrated that the Langmuir isotherm best described the adsorption process (R² = 0.9930 for MB and 0.9891 for CR), while pseudo-first-order kinetics indicated physical adsorption. Artificial neural network (ANN) models further validated the experimental results, showing high prediction accuracy (R = 0.9948 for MB and 0.9939 for CR). The PMNPs retained efficient performance after six reuse cycles, highlighting their reusability. This novelty of the research demonstrates the potential of PMNPs as a sustainable adsorbent and provides insights into optimizing adsorption processes through computational modeling.