(38a) A Mass Transfer Rate-Based Model for Liquid-Liquid Extraction

Liquid-liquid separation plays a vital role in various chemical engineering industries such as refining, natural gas processing, pharmaceutics and biotechnology. However, there is a notable lack of commercially available tools that can model such separation processes with any reasonable level of reliability and accuracy. Traditional approaches rely on ideal stage-based models that are augmented by rules-of-thumb and application specific knowledge to approximate the actual extraction column internals into number of ideal stages. When tray efficiencies are 5 to 20% and HETPs are a dozen or more feet, the results lose not only their validity but their reliability, too. They certainly fall well short in accuracy and predictive capabilities. The assumptions that may work for one set of process conditions frequently fail when the conditions change or when applied to other systems. One such limitation is the inability of ideal stage models to capture the varying rates of mass transfer and kinetics of solutes in a solvent of interest. For example, removal efficiency of H₂S and CO₂ show marked differences when treating LPG with aqueous amines. H₂S reacts instantaneously with amines, leading to significant enhancement of mass transfer rates compared to the slower kinetics observed for CO₂.

In this study, we present the results of a revolutionary mass transfer rate-based model for liquid-liquid extraction that was developed and implemented in the commercial chemical process simulator, OGT | ProTreat®. Unlike gas/vapor-liquid separation, liquid-liquid separation involves the transfer of chemical species between two dense phases. This leads to extraction efficiencies that are far lower than in absorption and distillation. Low tray efficiencies and large packing HETPs underscore the need for a more refined and fundamentally rooted rate-based approach. Unlike the ideal stage model, this approach uses the properties of the real extraction column internals as the model inputs and considers the phase equilibria, physical and transport properties of the process fluids and reaction kinetics to calculate the rate of transfer of species between the phases. The model is capable of handling both non-reactive physical solvents and reactive electrolytic solvents. To our knowledge, this is the first time a soundly-based liquid-liquid extraction model has been made available for use in a commercial process simulator. The significance of the rate-based approach is supported using the results of the model generated for actual extraction column internals against a single ideal stage. Removal of acid gases from LPG using reactive aqueous amines solvents was chosen for this demonstration.

The development of our model offers a transformative and robust process simulation experience for engineers to improve the design, operation and troubleshooting strategies for liquid-liquid extraction.