Pore Wetting Mitigation in Membrane Distillation: A Mechanistic Model for Water Flow in Janus Membranes and a 3D Morphology-Based Model for Liquid Entry Pressure

Abstract

Water scarcity is a rapidly growing concern worldwide. Since 97% of Earth’s water is seawater, there is great potential to tackle this challenge by removing salts from saline water to produce freshwater, a process that is called desalination. Among desalination methods, membrane-based processes have gained attention due to their efficiency in separating salt from water. Reverse Osmosis (RO) process is the state-of-the-art desalination technology, but it is incapable of desalting high salinity brines. Membrane Distillation (MD) is a promising solution for high salinity brines. However, it has a main challenge in pore wetting, which happens when liquid water passes through the pores. To this end, Janus membranes have drawn attention as their unique feature of having one side hydrophobic and the other hydrophilic and have been experimentally demonstrated to enhance pore wetting in MD. Furthermore, the liquid entrance pressure (LEP) is a key feature that determines the pore wetting prevention in conventional MD membranes. In this research, we first developed a model to simulate the directional transport of liquid water in the pores of the Janus membrane by calculating the surface energy change in association with the movement of water in a pore of a Janus membrane. It was demonstrated that when the hydrophilic side of the Janus membrane was designated as the feed side, the critical pressure to be overcome for liquid water to penetrate the pores of the hydrophobic layer would increase and therefore help prevent pore wetting. On the other hand, spontaneous liquid water flow will take place when the membrane is inverted. These simulation results unambiguously revealed the mechanism underlying the enhancement of pore wetting prevention by Janus membranes, i.e., the combination of the expulsion by the hydrophobic layer and the pulling by the hydrophilic layer to liquid water at the interface. Furthermore, we developed a 3D morphology-based model capable of analyzing 3D images of porous hydrophobic membranes to estimate LEP rapidly without requiring idealizations. The developed model was validated using the 3D image of a commercial membrane, PT20 from Gore Inc., and the corresponding experimental LEP reported in the previously published literatures. Using this model, we further investigated the effects of membrane morphological features including the max pore radius, mean pore radius, and throat radius, on LEP.