RO performance in a mixed solvent system: Analysis of nopinone concentration from a methanol-water solution

Reverse osmosis (RO) membrane technology represents a low-cost, low-energy alternative for the separation of valuable organic compounds from multi-component mixtures. However, limited availability of relevant experimental data renders the design of membrane separation systems for organic multi-component mixtures challenging, and permeation of more than one solvent component complicates the evaluation of membrane transport models. This study presents experimental data on the dead-end batch separation of nopinone from methanol-water solvent across a DOW BW30 RO membrane and addresses the specific challenge of modelling osmotic pressure of a multicomponent solvent. Membrane rejection and flux values were determined for feed concentrations of 0.449–1.099 M. A relation to evaluate osmotic pressure for mixed-solvent systems was derived from an energy balance and applied to the investigated system, which demonstrated the impact of solvent non-ideality on the transmembrane driving force when compared to the Van't Hoff description of osmotic pressure. Data analysis via the solution-diffusion model revealed the presence of convective transport in the investigated system, while the Spiegler-Kedem model obtained a reasonable system description when using a single coefficient for mass transfer description. From the Spiegler-Kedem model, permeability coefficient values A and B were yielded of order 10⁻⁷ and 10⁻⁸, respectively, with a membrane selectivity of ∼30 bar⁻¹. Finally, the system's separation performance in continuous cross-flow was estimated by considering a series of dead-end batch separation units, suggesting feasible system operation at industrial scale.