ADSORPTIVE REMOVAL OF HEAVY METAL IONS FROM WASTEWATER UTILIZING FAYALITE SLAG AS A LOW-COST ADSORBENT: BATCH STUDIES, ARTIFICIAL NEURAL NETWORKS AND ADAPTIVE NEURO FUZZY INFERENCE SYSTEM MODELING AND OPTIMIZATION

Abstract

This study investigated the adsorptive removal of three heavy metal ions (Cu²⁺, Ni²⁺, Fe²⁺) from wastewater using fayalite slag (FS), a low-cost industrial by-product. The media was characterized using XRD, XRF, SEM, batch adsorption studies were conducted, and the adsorption process was further optimized and predicted adsorption efficiency by developing Artificial Neural Network (ANN) and Adaptive Neuro-Fuzzy Inference System (ANFIS). Comprehensive characterization revealed that FS was suitable as an adsorbent, with a particle size distribution (d10 = 0.3mm, Cu = 5.36), high porosity (50%), and significant iron oxide content (Fe₂O₃ = 45.44%). Batch adsorption experiments demonstrated optimal removal efficiencies at a dosage of 2.0 g/100mL, near-neutral pH (6-8), and a contact time of 40 minutes, achieving approximately 3.5 mg/ g adsorption capacity (35% removal efficiency) for each metal. The adsorption kinetics aligned with the pseudo-second-order model (R² ≥ 0.994), indicating chemisorption as the rate-limiting step. Langmuir and Freundlich isotherm models effectively described the adsorption behavior. Thermodynamic studies confirmed that the process was endothermic (ΔH° > 0) and spontaneous (ΔG° = -9.023 to -10.294 kJ/mol). Reusability studies showed a gradual decline in recovery efficiency, from approximately 30% in the first cycle to less than 20% by the third cycle. The ANFIS model, with six fuzzy rules, exhibited superior performance (R² = 0.823, RMSE = 11.87), demonstrating a high correlation between predicted and experimental data. Thus, FS proved to be an effective, sustainable adsorbent, warranting further research into regeneration optimization, continuous flow applications, and real wastewater treatment, alongside cost-benefit and pilot-scale assessments for industrial implementation.