Exploring the efficacy of eggshell and its pyrolyzed products for ciprofloxacin removal with machine learning insights

Abstract

This study investigated the efficacy of utilizing eggshells and their pyrolyzed derivatives, within the temperature range of 400–800 °C, as adsorbents for ciprofloxacin (CIP) removal. Experimental data were analyzed using various machine learning (ML) algorithms, viz. linear regression, random forest, support vector machines, decision trees, and k-nearest neighbor to predict performance. Results demonstrated that pyrolyzed eggshells at 600 °C (PES-600) exhibited the highest CIP removal rate (86.06 ± 2.25%). Optimal performance was consistently observed at an initial CIP concentration of 125 mg/L, with the order of PES-600 > PES-500 > PES-400 > PES-700 > eggshells > PES-800. Adsorption capacity peaked at pH 5 (5.84 ± 0.1 mg/g), attributed to interactions including hydrogen bonding, π–π interaction, and ion exchange. Scanning electron microscope images revealed that PES-600 had the highest number of pores, resulting in a smoother surface post-adsorption. Langmuir isotherm model fitting was best for ES, PES-700, and PES-800, while Freundlich isotherm was suitable for PES-400, PES-500, and PES-600. PES-600 showed the best fit with the pseudo-second-order kinetic model. Characterization analysis highlighted the significance of functional groups like C = O, C = C, and –CH groups in aromatic rings. ML algorithms demonstrated remarkable performance with an accuracy level of 90.28%. In conclusion, pyrolyzed eggshells can effectively remove ciprofloxacin (CIP) from wastewater, with optimal performance predicted by the random forest machine learning algorithm when considering real environmental factors.