(348d) Linking Microstructure of Membranes and Performance

Even though synthetic membranes have been around since the early 20^th century, predicting their separation performance for liquids remains a substantial challenge. This work addresses the important link between the microstructure of a membrane and its filtration performance.

2D computational fluid and particle drag mechanics are combined with particle and membrane force measurements in aqueous solutions containing inorganic ions to study particle intrusion and capture in microporous commercial polymer and computer-generated teardrop membranes. Fits of the DLVO theory to force-distance profiles obtained membrane surface potentials needed for the computations. In silico predictions of particle intrusion for a commercial membrane qualitatively agree with experimental filtration measurements using scanning electron microscopy with particle tracking via energy dispersive X-ray spectroscopy. Highlighting the poor flow field, several dominant inhomogeneous 2D flow conduits with large unused regions of the internal pore structure are discovered.

To guide improved design, new computer-generated microporous teardrop structures that can equalize the flow field, adjust the tortuosity of the flow path and vary the reactivity of the surface were tested in silico. The main assumptions of the computational model were that 2D flows are a valid description of 3D flows, all forces were applied at the particle center of mass, forces were calculated based on the physical diameter of the spherical particles. Relatively large pores (~5 μm) and large particles (~1 μm) were selected for easy detection and analysis. However, the computational fluid and particle flow analysis and the inter-surface forces scale independently with size and should apply at all classical dimensions (i.e. for nano-, ultra- and micro-filtration). Assumptions for the intermolecular force measurements were that electrostatic and van der Waal's forces dominated and hence that the DLVO theory was valid and that the zeta potential values were close to those at the wall (i.e. surface potential). In particular the DLVO was applied to ideal geometries: a sphere (i.e. AFM probe) near to a flat surface (i.e. either a silica wafer or a hot-pressed PES membrane).

To our knowledge, this is the first attempt combining particle drag mechanics with intermolecular force measurements to help explain particle dynamics in synthetic membranes. This computational fluid mechanics-based tool can be used to characterize membranes for separation performance and guide improved design, synthesis and testing of new microporous membranes.