Machine Learning-Based Modeling of Methyl Blue Adsorptive Removal from Aqueous Solutions with Lignin

Abstract

Dyes are chemical compounds extensively utilized in numerous industries such as textile, paper, leather, and cosmetics. These substances can cause serious environmental problems by mixing into wastewater during production processes. Methyl blue (MB), which is an anionic dye, has a wide range of applications and poses serious ecological and health risks if discharged untreated into water bodies. In this study, the adsorption process of MB removal using lignin was explained by examining the parameters of temperature, contact time, pH, initial MB dye concentration and initial lignin amount. In addition, lignin was characterized by FT-IR, SEM-EDS, XRD, Zeta potential, DSC, and BET. Then, the results obtained by artificial neural networks (ANN) and support vector regression (SVR) methods were modeled. The analysis revealed that the process is endothermic, follows a pseudo-second-order (PSO) kinetic model, and conforms to the Langmuir isotherm model, with a maximum adsorption capacity (q(max)) of 175.439 +/- 8.772 mg/g, R-2 = 96.200% for ANN and R-2 = 93.500% for SVR. The results obtained showed that lignin can be used for MB removal and that it would be suitable for use in machine learning algorithms.