In this work, the mass transfer of dibenzothiophene from fuel to solvent fluid (deep-eutectic-solvent) in a packed column has been studied. The model fuel contains 500 ppm of dibenzothiophene through a packed liquid-liquid extraction column with a solvent composed of choline chloride and diethylene glycol. Experimental tests were performed at two temperatures (25.9 °C and 33.1 °C) and six dispersed phase flow rates, while the continuous phase flow rate was kept constant. Hydrodynamic parameters including dispersed phase holdup, Sauter mean droplet diameter, and specific interfacial area were measured experimentally, showing that as the dispersed phase flow rate increased from 24 to 124 mL/min, the holdup rose from 0.06 to 0.137 and droplet size increased from 3.52 to 4.76 mm, resulting in an expanded interfacial area. The plug flow model was applied to simulate mass transfer behavior, and overall mass transfer coefficients were obtained by fitting predicted concentration profiles to the experimental data using a PSO-based optimization algorithm. The model demonstrated excellent predictive capability, with an average absolute relative error below 0.5%. The overall mass transfer coefficient increased with the dispersed phase flow rate, ranging from 5.14 × 10 −7 to 9.52 × 10 −6 m/s, while slightly decreasing with temperature due to reduced solvent capacity. The extraction efficiency followed similar trends, reaching up to 25.8% under optimized conditions. These findings confirm the effectiveness of DES-based extraction and illustrate the modeling as a robust predictive tool for column design and process optimization in non-HDS desulfurization systems. • Extractive desulfurization of DBT was conducted in a packed column using DES. • Hydrodynamics and mass transfer were evaluated at different flow rates and temperatures. • Plug flow modeling predicted DBT concentration profiles with errors below 0.5%. • The overall mass transfer coefficient (Kod) increased with increasing flow rate. • Higher temperature slightly reduced DBT removal due to lower distribution coefficient.
Kazemi et al. (Sat,) studied this question.