Modeling of textile dye removal from aqueous solution with a Zr-MOF framework : design of experiments and artificial neural network approaches

Zirconium-based metal-organic framework (Zr-MOF) was sucssesfully synthesized using a solvothermal approach and utilized for the removal of textile dyes from an aqueous environment based on adsorption phenomen. Comparative studies between Artificial Neural Network (ANN) and Response Surface Methodology (RSM) approaches were investigated for prediction and modeling of the adsorption capacity of Zr-MOF toward organic dyes in batch adsorption experiments. Furthermore, the MOF particles were characterized by various microscopy and spectroscopy techniques, including field-emission scanning electron microscopy (FESEM), X-ray diffraction (XRD), Fourier transform infrared spectroscopy (FTIR), and nitrogen adsorption-desorption isotherm. The effects of significant adsorption parameters including pH, contact time, MOF content, initial concentration of dye solution on the removal process were succesufly analyzed based on statistical approaches. The results showed that the MOF content and dye concentration were the two most significant parameters influencing the adsorption capacity of synthesized MOF. The paper compare the actual and predicted adsorption performance of MOF showed the superiority of the ANN approach over the RSM model, having a comparatively high degree of coefficient of determination (R² = 0.996) and lower values of errors (0.96%). The obtained results revealed the high efficiency of the proposed ANN model for describing the dye removal process via the adsorption method with high accuracy as well as determining the optimum adsorption condition to achieve the desired adsorption capacity.