Search

ispartof: CHEMICAL ENGINEERING RESEARCH & DESIGN vol:184 pages:127-136 status: published

ispartof: SEPARATION AND PURIFICATION TECHNOLOGY vol:297 status: published

The mass transfer coefficient, k L α , is often used to benchmark the liquid–liquid mass transfer performance in a microfluidic reactors. Cobalt(II) was extracted with an undiluted, highly viscous, ionic liquid from an aqueous phase in a microfluidic reactor (1 mm ID) and k L α values were determined (0.0002–0.033 s−1). It is concluded that the high viscosity of the ionic liquid causes two-phase slip, greatly reducing the efficiency of the microfluidic reactor. Furthermore, deviations (or non-idealities) from the assumptions related to mass transfer calculations (e.g., no reaction effects) can have a drastic impact on the measured k L α values. This work reports deviations between −25 and +40%. In general, it is shown how several non-idealities (i.e., a high continuous phase viscosity, non-linear equilibrium behavior and concentration effects) can affect the calculated mass transfer coefficient.

Milliflow devices operating under slug flow (Taylor-flow) conditions provide an efficient mass transfer, a highly uniform droplet morphology and a narrow residence time distribution i.e. characteristics which are excellent for the performance of reactions under kinetic control. Milliflow devices have also been proposed in combination with ionic liquids, due to their ability to intensify the solvent use of these rather expensive and viscous liquids. In present research, the solvent extraction separation of a mixture of Au(III), Pd(II), Cu(II) and Fe(III) was studied using the chloride and bromide forms of the undiluted quaternary ammonium ionic liquid Aliquat 336 in a slug flow reactor channel with an internal diameter of 1.0 mm. The extraction of Au(III) and Pd(II) was found to be faster than that of Cu(II) and Fe(III), allowing the precious/base metal separation to be improved by modification of the contact time between both phases. An apparent correlation between the observed extraction rates and the equilibrium distribution ratio D was found. The most significant separation improvement was achieved at those conditions where the equilibrium separation was best i.e. low halide and metal concentrations and low organic-to-aqueous flow rate ratios. At a residence time of 25 s, quantitative Au(III) and Pd(II) recovery was achieved and decontamination factors increased by a factor up to 2, which corresponds to an increase in product purity of up to 14%. Upon substituting the chloride system for a bromide extraction medium, similar observations were made, except for a reversed Cu(II)/Fe(III) selectivity. Experiments were also conducted on the separation of Pt(IV), Pd(II) and Rh(III), where a reduced Rh(III) extraction rate was observed, allowing to suppress its co-extraction by up to 20%. A hydrodynamic analysis of the obtained flow regime of the ionic liquids revealed a stable and uniform slug flow characterized by a decreasing plug length and width with increasing flow rates. A comparison of the applied and observed phase ratio revealed a stagnant ionic liquid excess (i.e. hold-up) inside the reactor of up to 40% which effectively reduced the reactor diameter and consequently plug residence time.