Flowsheet Modelling and Simulation of Aqueous Two-Phase System for Protein Extraction

Aqueous two-phase extraction system (ATPES) has many advantages as an efficient, inexpensive large-scale liquid-liquid extraction technique for protein separation: it provides mild environment for proteins, it can be integrated with upstream processes, it can handle high throughput, it is highly selective hence offering high yield, and most importantly, it costs less comparatively to other conventional technique. However, the realization of ATPES as a new protein separation technology at industrial scales is rather limited. Among the reasons are that the selection of phase-forming agents can be exhaustive as there are many tuneable parameters, efforts are mostly empirical and intuitive with many possibilities and there are limited design approaches capable of accurate prediction of system and product behaviour that offer large scope of application with operation assessment and performance sensitivity available for ATPES. This paper describes the framework we designed to calculate suitable flowsheets for the extraction of a specified target protein from a complex protein feed using ATPES. We set up flowsheets according to target protein partitioning behavior in ATPES and calculated the amount of phase-forming components needed to extract the target protein from the complex mixture. Extraction units consisted of either one or two stages containing an ATPES. The thermodynamics of phase formation and partitioning were modeled using Flory-Huggins theory and calculated based on a non-linear minimization of Gibbs energy difference approach. For a case study application, we present the design of suitable flowsheets for a model feedstock containing phosphofructokinase and demonstrate the existence of feasible solutions. The ATPESs we consider are the water-PEG6000-DxT500 and water-PEG6000-Na3PO4 systems for which we developed phase behavior model. We discuss some scaling issues involved in the thermodynamic calculations and propose scaling weights and problem reformulation approaches to overcome computational challenges. We then compare the flowsheets in terms of their performance, settling rate and cost factor. We also analyze the performance trends qualitatively to check that the flowsheet calculation results capture the physical behavior of the unit. Finally, we propose a basis for flowsheet optimization for protein extraction using ATPES.