Deconvoluting the Forward Osmosis Structural Parameter Using a “Model” Thin Film Composite Membrane

Engineered osmosis (EO) is a platform of membrane technologies for various applications including water desalination (forward osmosis, FO), power production (pressure-retarded osmosis, PRO), dewatering etc. Internal concentration polarization (ICP) in EO processes has been identified as a significant impediment towards obtaining high water fluxes and subsequent commercialization of the technology. ICP is largely influenced by the support layer structure of EO membranes, described in terms of the structural parameter. In all EO studies so far, this parameter is typically calculated in an indirect method due to the lack of suitable techniques to obtain direct evaluations of the membrane structure. A previous study conducted by our lab group showed that this approach is inherently error-prone with the calculated structural parameters varying greatly from the “actual” value that was obtained using two different characterization tools. Further, the effects of the individual terms in the structural parameter equation, S = tτ/ε (t – thickness of the support layer, τ – tortuosity of the support layer pore structure and ε – porosity of the support layer pore structure) have not been investigated for their effect on the severity of the ICP phenomenon, and hence, on overall flux performance. In this study, a “model” TFC membrane has been fabricated by forming a polyamide selective layer on a polycarbonate track-etched membrane by interfacial polymerization. The “real” structural parameter of the membrane was calculated a priori from known t, τ and ε values. Then, the S value was calculated in the conventional manner, from osmotic flux tests and membrane characteristics obtained from reverse osmosis tests. The S value obtained using this method was always found to be higher than the “true” S parameter, for tests run in both FO and PRO membrane orientation modes. Additionally, for tests run in the PRO mode, we saw a non-negligible change in the calculated S parameter with draw solution concentration. The results obtained with the “model” TFC membrane were compared to that of a commercial cellulose acetate FO membrane from Hydration Technologies Innovation (HTI), the HTI-CTA membrane. Similar trends were observed with both membranes. Simplifying assumptions in the indirect approach of calculating the S parameter call for a revision of the existing models which is one of the goals of this ongoing study.