Prediction and optimization of Rhodamine B removal from water using metal-organic frameworks: RSM-CCD, ANN, non-linear kinetics, and isotherm studies
Simon Bbumba, John Ssekatawa, Ibrahim Karume, Emmanuel Tebandeke, Moses Kigozi, Solomon Yiga, Robert Setekera, Joseph Ssebuliba, Steven Sekitto, Ruth Mbabazi, Ivan Kiganda, Maximillian Kato, Patrick Taremwa, Moses Murungi, Chinaecherem Tochukwu Arum, Collins Yiiki Letibo, Geofrey Kaddu, Margret Namugwanya, John Kusasira, Peace Mwesigwa, Muhammad Ntale
Abstract
This study involved the chemical synthesis of Metal-organic Frameworks (MOFs). The synthesized MOFs were characterized using Scanning Electron Microscopy (SEM), Fourier Transform Infrared (FTIR), and Powder X-ray diffraction (PXRD). Artificial intelligence models such as Response Surface Methodology (RSM) and Artificial Neural Networks (ANN) were used to predict and optimize the adsorptive removal of Rhodamine B (RhB) from water. The adsorption process was optimized using RSM with a Central Composite Design (CCD), which predicted a maximum removal efficiency of 95.91% under the following conditions: initial dye concentration (10 mg/L), adsorbent dosage (15 mg), pH (6), and temperature (25 °C). ANN was also optimized using similar conditions and the resulting predictive removal efficiency of 97.18% was obtained. Non-linear isotherm studies strongly correlated with the Freundlich (R² = 0.9987) and Sips (R² = 0.9928) models, indicating multilayer and monolayer adsorption. Non-linear Pseudo-first-order, Pseudo-second-order, and Elovich model correlation coefficients of 0.9644, 0.9998, and 0.952 suggested that the mechanisms were by chemisorption and physisorption on energetically stable heterogeneous surfaces. The findings of this study show a dual approach based on metal-organic framework and machine learning models as efficient alternatives to understanding the removal of RhB from water.