(653f) Elucidating the Impact of Temperature Gradients Across Membranes during Osmosis

The forward osmosis (FO) process has long been touted as an emerging desalination technology that can operate off of a thermally regenerated draw solution. Using thermal energy to regenerate this solutions, however, can lead the draw solution being warmer than typical ambient conditions. Likewise, in industrial reuse or produced water treatment applications, the feed solution can be elevated in temperature. While each of these concepts has been discussed in the literature individually, lacking from the field is an understanding of how different solution temperatures can impact FO performance.  We look at how various thermal gradients impact water flux, reverse salt flux and membrane intrinsic properties using both experiments and model based prediction to do this, we develop a model that effectively couples the effect of simultaneous heat and mass transfer across the membrane. The model is used to predict water flux and reverse salt flux in FO and PRO mode. We, then, compare these fluxes with experimentally obtained results. Experimental results suggest that the water flux increases substantially at higher system temperature (40˚C), as has been determined before, while the reverse salt flux decreases at that temperature. In addition, operating at higher feed temperature (40˚C) than the draw side (20˚C), produces fairly comparable results to that operating at higher system temperature (40˚C). The model prediction for water flux also matches well to the results obtained in these operating conditions. Reverse osmosis (RO) characterization reveals that membrane intrinsic properties such as pure water permeance increases with temperature while the salt permeability decreases. The decreasing trend in salt permeability is also seen in the reverse salt flux performance of the FO experiments where the membrane orientation is similar to RO. This suggests that operating at higher feed temperature results in better performance  as compared to operating at higher system temperature where both feed and draw solutions are heated concurrently. Even though the predicted water flux by the coupled model compares well to the experimental value, the predicted reverse salt flux deviates as much as 30 % from experimental value. Overall, the coupled model performs well compared to similar models.