Iron-modified Calcinated Coconut Fiber Biochar for Simultaneous Removal of Copper and Zinc: Optimization, Production Cost Assessment, and Machine Learning Modeling

Abstract

This study developed and evaluated an iron modified biochar (AcBC) from coconut fiber in Ben Tre, Vietnam, for the adsorption of Cu²⁺ and Zn²⁺. Characterization (SEM–EDS, BET, XRD, FTIR) showed that mineral rich coconut fiber pyrolyzed without inert gas undergoes strong surface mineralization, forming crystalline silicates that block pores and functional groups, resulting in raw biochar (RBC) with low surface area, limited porosity, and few active sites. After iron salt co precipitation with NaOH, AcBC exhibited major improvements, including a nearly 20- fold increase in surface area form 5 to 100 m²/g, well developed mesopores, and the return of functional groups. Silicate minerals also contributed to adsorption by providing active sites and supporting iron oxide and iron silicate formation. AcBC achieved about twenty percent higher removal efficiency and maximum capacities of 34.75 mg/g for Cu²⁺ and 29.69 mg/g for Zn²⁺, more than double those of RBC. To systematically describe the adsorption mechanisms, four kinetic models (pseudo-first-order, pseudo-second-order, Weber-Morris intraparticle diffusion, and Boyd) and four isotherm models (Langmuir, Freundlich, Temkin, and Dubinin-Radushkevich) were utilized. Crucially, tree-based machine learning interpreted via SHAP analysis successfully quantified the mechanisms; Extreme Gradient Boosting accurately predicted adsorption capacity R² > 92%, identifying specific surface area, % Si, and % O as the most important material properties, while that of cation descriptors relative atomic mass, effective ionic radius and first order hydrolysis constants. With low cost production, simple synthesis, and strong performance, AcBC represents a practical and scalable option for treating heavy metal contaminated wastewater in developing regions.