Machine learning-assisted optimization and evaluation of methylene blue adsorption kinetics on citrus aurantifolia leaves: Insights from isotherm and thermodynamic studies

Abstract

BackgroundMethylene Blue (MB) is a cationic dye widely used in various industrial and pharmacological applications. However, its improper disposal poses significant environmental and health risks due to its toxicity, carcinogenicity, and resistance to biodegradation. Despite extensive efforts have been done to remove MB using green synthesis approaches, however; there remains a critical need to enhance the efficacy of these methods through proper optimization and the use of novel green sources. Addressing these challenges is essential to mitigate the environmental impact of MB. MethodsCitrus aurantifolia leaves powder (CALP), a simple, environmentally friendly, adsorbent has been utilized to remove MB dye. The adsorption kinetic efficacy of CALP was studied under different experimental conditions, including pH (3–9), contact time (2–8 min), and MB dye concentrations (10–70 mg/L). Additionally, machine learning (ML) was employed to optimize experimental conditions, predict, and validate the adsorption kinetic efficacy. The machine learning (ML) optimized experimental dataset which included parameters of pH (9), contact time (2 min), CALP adsorbent (0.3 gm), and MB dye concentration (70 mg/L) has shown maximum adsorption efficacy. To enhance and assess the extraction efficiency of MB dye, a response surface methodology (RSM) was employed utilizing a central composite design (CCD). Significant findingsThe adsorption efficacy of CALP was verified by the Freundlich isotherm model, which demonstrated an adsorption capacity of 415.05 mg/g with a coefficient of determination R2 of 0.999. The nonlinear pseudo-second-order kinetic model with an R2 value of 0.998 further confirms MB dye adsorption. In addition, spontaneous nature of the process was determined by calculating thermodynamic parameters such as ΔG°, ΔH°, and ΔS°. The outcomes demonstrate that CALP is a highly effective, reliable, and easily available adsorbent for the removal of MB dyes from wastewater. To the best of the author's knowledge, there are substantial unmet research gaps in the fields of wastewater treatment and adsorption modeling optimization. In particular, research that combine cutting-edge machine learning methods with conventional adsorption models to enhance process efficiency and prediction accuracy are scarce. This study is the first to employ CALP, a cost-effective and environmentally sustainable natural adsorbent, for the adsorption of MB dye utilizing a machine learning-optimized approach.