Influence of microstructure on membrane distillation: high-resolution 3D reconstructions for analysis of pore-scale phenomena, The

Abstract

The demand for clean water has seen a rapid increase in the last decade; increasing the need for technological advancement in producing potable drinking water. Membrane distillation (MD) is an emerging approach for producing fresh water via desalinating high-concentration brines, brackish waters, produced waters, and seawater. Though attracting considerable attention, several technological barriers must be solved for MD to see wide industrial application. The underlying mechanisms for heat and mass transfer through MD membranes remains poorly understood. This is largely due to the knowledge gap between continuum-level transport models and MD membrane microstructure. Also, MD membranes are typically designed for other applications such as a reverse osmosis (RO) and fuel cell technology. Being able to characterize MD membrane microstructure can lead to enhanced transport modeling and new design criteria for MD specific membrane production. Focused ion beam scanning electron microscopy (FIB-SEM) technology was implemented as a means for analyzing MD membrane microstructure and creating digital 3D membrane reconstructions. To make FIB-SEM analysis viable, a membrane mounting, infiltration, and preparation protocol was developed. The FIB-SEM “Slice and View” procedure was used to collect 2D SEM images that were serially stacked to produce a 3D reconstruction of membrane pore networks. With the 3D reconstruction, important microstructural parameters such as porosity, pore size, solid fiber size, and tortuosity factor were extracted using the reconstruction software and the MATLAB application, TauFactor. Results showed that FIB-SEM is able to resolve major structural features within the membrane pore network but has difficulty in resolving thin, connecting fibers causing discrepancies between the microstructural parameters given by the manufacturer. This is likely due to the membranes soft polymer material being compromised under ion and electron beams conditions utilized by FIB-SEM. However, obtaining high-resolution 3D reconstructions can lead to direct CFD analysis and “numerical experiments” to validate state-of-the-art transport models used for MD systems. 1D transport models such as Dusty Gas Model (DGM) and a simple Fickian diffusion model have been implemented to better understand underlying MD transport mechanisms and to determine their validity for simulating MD membrane transport. Literature supports DGM for simulating transport through porous media but has not been thoroughly validated for materials with high porosity (> 60%), which is the case for MD membranes. Using and modifying these models allows for an understanding about which microstructural parameters play an important role in predicting flux. Membranes can have identical properties such as membrane thickness, pore radius, and porosity but can yield vastly different experimental flux measurements. Simulations using both models at various feed and permeate flow temperatures and membrane parameters were performed to explore the transport mechanisms of each model and the heat and mass transport occurring at and within the membrane. Temperature, mole fraction, and total pressure profiles were developed to further analyze transport mechanisms and gradients within the membrane microstructure. Simulation results indicate that tortuosity is a limiting factor and an integral parameter for determining flux, meaning two membranes can have similar porosities, thicknesses, and pore sizes, yet two different tortuous networks for water vapor flux transport. Fine-tuning 1D transport models to better represent membrane tortuosity (for both the pore- and solid-space) and direct analysis of the parameter via 3D reconstructions has the ability to provide accurate heat and mass transfer simulation models and influence enhanced design criteria for MD membranes.