(596d) The Role of Pore Structure and Chemistry on Particle Deposition during Membrane Filtration

Authors: 
Sorci, M., Rensselaer Polytechnic Institute
Woodcock, C. C., Rensselaer Polytechnic Institute
Plawsky, J. L., Rensselaer Polytechnic Institute
Belfort, G., Rensselaer Polytechnic Institute
Many relevant industrial fluids (e.g. cell culture broths, paints, milk, etc.) can be characterized as colloidal suspensions (i.e. nanometers to micrometers) and are of great interest for microfiltration. This pressure-driven membrane process is one of the oldest, yet, remains not fully understood. A major limitation is membrane fouling. Particle deposition onto the microporous membrane, resulting in pore constriction and/or blockage, and eventually in cake formation, has been largely investigated and mitigation strategies have been proposed.

In this study, we focus on particle deposition into a microporous membrane, with particular attention for the role of pore structure and chemistry. The membranes used were 0.2 µm and 5 µm pore size poly(ether sulfone) (PES), since PES is widely used due to ease of casting and mechanical stability. Silica nanoparticles, 40 nm and 1 µm, respectively, were selected as model colloidal suspensions.

In vitro studies focused on 2D and 3D characterization of pore distributions, by atomic force microscopy (AFM), scanning electron microscopy (SEM), and focused ion beam SEM (FIB-SEM). In order to understand the forces in play, (i) zeta potential for both membranes was measured using 1-10-100 mM electrolyte solutions in a pH range 4-10; and (ii) AFM in force mode was used to estimate the adhesion force between the membranes and a silica microsphere attached to an AFM probe. Finally filtration experiments were performed using dilute colloidal suspensions (i.e. <1% v/v), and array tomography, combined with SEM imaging, was applied to visualize the distribution of particles into the fouled membranes.

Taking advantage of the experimental characterization, in silico modeling using COMSOL Multiphysics software was conducted in parallel to build a better understanding of particle interactions with the membrane walls in the pores. Prototype models will combine an accurate pore geometry representation coupled with experimental information on particle-wall forces to obtain particle trajectories through the structure and deposition rates onto the pore wall. The information will be used to narrow down the key features pointing toward optimal membrane performance.