Non-linear least-square fitting method for characterization of forward osmosis membrane

We assess three forward osmosis (FO) performance data collection protocols, namely multiple-filtration, multi-stage and single-stage using non-linear least-square fitting method to simultaneously characterize water permeability coefficient (A), solute permeability coefficient (B) and structural parameter (S) of FO membranes. They differ only on the procedure to acquire water and reverse salt flux data. Both set of fluxes data were later used as inputs in non-linear least-square fitting optimization to obtain the A, B and S parameters. The method were evaluated using two commercial FO membranes. We employed the mass-transport models that counting for internal and external concentration polarizations in the feed and the draw sides. We also demonstrated the use of the single-stage protocol to characterize fouled FO membranes. Overall results showed that all protocols are equally effective for characterizing FO membranes, demonstrating generality of the least-square fitting method. Furthermore, the single-stage protocol can be used for characterizing fouled FO membrane in situly. As expected, fouling was more severe for the active layer facing draw solution (ALDS) than the active layer facing feed solution (ALFS) mode, as was also reflected by larger changes of their properties relative to their pristine conditions. The ability to characterize fouled membrane using the single-stage protocol could help to unravel mechanisms of fouling in FO. For instance, as demonstrated in this study, fouling in ALDS mode occurred internally by penetrating support layer. © 2018 Elsevier Ltd