Unraveling the adsorption potential of Zr dithiol (MOF-DSH) through experimentation and neural network modeling

Abstract

In this study, an aqueous-stable metal–organic framework with two thiol groups was synthesized using Zr as the metal centre and dimercaptosuccinic acid as the ligand through a conventional heating method for the removal of Cd(II) from aqueous solution. Different characterization tools, including XRD, FT-IR, BET, SEM, TGA, and XPS, were employed. XRD results showed a characteristic pattern of a hexa-cluster, which was in agreement with the simulated MOF-801, and the corresponding vibrational peaks were observed in the FT-IR spectra. The synthesized MOF was thermally stable up to 300 °C, as demonstrated by TGA, and exhibited a specific surface area of 290 m² g⁻¹. Cadmium adsorption studies performed at different pH values showed a maximum adsorption capacity of 91.5 mg g⁻¹ at pH 6. The adsorption behavior was well described by the Langmuir model and the pseudo-second order (PSO) kinetics, confirming the involvement of a monolayer with chemisorption as the dominant mode of adsorption. The synthesized MOF could be reused at least 4 times while retaining ∼80% of its initial adsorption capacity. FT-IR, XPS, and pH studies after Cd(II) adsorption revealed that the predominant mode of interaction of Cd(II) with the MOF is an ion-exchange mechanism. An artificial neural network-based (ANN) methodology was employed to model the adsorption capacity of Cd(II) and predict the adsorption capacity as a function of Cd(II) concentration, time of contact, and pH of the medium. The model demonstrated excellent results, with an average error of 2.3% and precision of 3.0%. The outcomes of these studies were consistent with the experimental results.