Sustainable tropical fruit peel waste biochars for enhanced cadmium and lead adsorption: mechanistic insights and optimization using response surface methodology and backpropagation neural networks

Heavy metal contamination, particularly from cadmium (Cd(II)) and lead (Pb(II)), presents a severe environmental challenge due to its toxicity and persistence. This study explores an innovative approach by utilizing abundant yet underutilized tropical fruit peel waste to produce biochars that serve as effective, sustainable adsorbents for heavy metal remediation. Biochars derived from banana peels (BP) and Monthong durian shells (DS) were synthesized via pyrolysis at 400–800 °C and evaluated for their physicochemical properties and adsorption efficiency. The DS600 biochar exhibited the highest adsorption capacity, removing Cd(II) (40.37 mg/g) and Pb(II) (51.74 mg/g), surpassing BP600 (40.22 mg/g and 47.23 mg/g, respectively). This study introduces a dual-modeling framework by integrating response surface methodology (RSM) with backpropagation neural network (BPNN) to optimize adsorption conditions and enhance predictive accuracy. The optimized conditions achieved over 99% removal efficiency, with R² > 0.98 and MSE < 0.05, confirming the robustness of the model-based predictions. The study highlights the superior adsorption performance of DS600 biochar, with adsorption mechanisms influenced by pH, dosage, and biochar properties. In contrast to conventional studies that focus solely on equilibrium adsorption or rely on statistical models, this work pioneers the use of tropical fruit peel biochar in heavy metal remediation, providing quantitative insights into process optimization and practical scalability. The findings demonstrate the potential for valorizing agricultural waste into high-performance adsorbents, advancing cost-effective and sustainable water treatment technologies.