Optimization and modeling of Congo Red adsorption onto Ceratonia siliqua Leaves using Response Surface Methodology and Artificial Neural Networks

The need to develop cost-effective and green adsorbents for dye removal from water is an area of concern for water treatment. In this paper, the green adsorbent Ceratonia siliqua leaves (CSL) was used to remove Congo Red (CR) dye. A series of physicochemical analyses were carried out to characterize the green adsorbent. First, the effects of various operational factors on the CR removal efficiency were investigated. Subsequently, the process was optimized using the Box-Behnken design (BBD) method. The optimal values for the parameters were found to be 2.55 mg for the amount of adsorbent used, a pH of 3.58, and a contact time of 45.13 min. Using these optimal values, the removal capacity of the adsorbent for CR dye was predicted to be 614.83 mg/g. However, the actual removal capacity of the adsorbent for CR dye was found to be 610.31 mg/g or 99.80% of the removal efficiency. The obtained results confirmed the effective adsorption performance of CSL toward CR removal, highlighting its potential as a low-cost and sustainable biosorbent for wastewater treatment applications. The adsorption kinetics and equilibrium data were best described by the pseudo-second-order and Langmuir models, respectively. The thermodynamic study showed that adsorption is spontaneous and endothermic. RSM and ANN models showed excellent predictive performance, with R 2 values of 0.997 and 0.990, respectively. These findings demonstrate how statistical and machine learning methods can work in tandem. This study showed CSL as a promising adsorbent in the removal of dyes in water treatment, being eco-friendly and affordable.