60th Annual Report on Research 2015

The goal of the project is the development of a membrane-based oil/water separation method that can produce organic-free water while avoiding the fouling phenomenon that has so far prevented the widespread application of membranes towards these separations. The main advantage of membrane-based separations is the consistently high permeate water quality and complete recovery of oil. Project ProgressTo date, we have tested a range of ferromagnetic particles for their ability to form magnetic Pickering emulsions, and characterized the resulting coated droplet size. We have then introduced the particle-stabilized oil to a membrane module, and tested the ability of different types of membrane materials to filter out the oil/water mixture. We have varied multiple system parameters and tested system performance in terms of membrane fouling and permeate water quality. Based on our experimental results, we have developed a theoretical framework that allows us to place operational boundary conditions (particle properties, membrane properties, and hydraulic conditions) on future experiments, as well as provide a mechanistic understanding of the complex, multi-phase environment that develops along the membrane surface during the filtration process.Experimental ResultsOur filtration experiments were conducted under a range of salinity conditions (DIW and high salinity water). In all experiments, the oil concentration was maintained at 10 ml oil to 1 L of water (1% by volume). When bare Fe3O4 particles were used to stabilize the oil, the membrane did not foul when the flux was set to 50 L/m2 hr. In contrast, the Fe3O4 particles coated with PVP, which renders them more hydrophilic, resulted in rapid membrane fouling. Importantly, under these conditions (50 LMH), the membrane material (and related properties) did not have an impact on the fouling behavior of the system, evident by the constant pressure required to drive the separation regardless of the membrane used. When the flux was increased to 100 LMH, the membrane material started to play a role in its fouling properties. Under these conditions, only the most hydrophilic membrane material (PVA-CNT), did not experience membrane fouling, with the other membranes (PSF and PAN) experiencing rapid and irreversible fouling. Under all conditions, the total organic carbon (TOC) measured in the permeate was 7±3 ppm.Theoretical FrameworkWe established a framework that can be used to predict the maximum flux possible through a membrane given a combination of particle and membrane properties. The framework is a result of our understanding of the dynamic relationship between the particles stabilizing the oil, the membrane surface, and the oil. Once the particle-stabilized oil drops are introduced to the membrane module, the oil forms a cake layer on the membrane surface; the thickness of the cake layer is a function of the membrane flux and cross-flow velocity. Oil drops in the cake layer will experience a pressure drop due to the water flowing through the porous cake. Thus, particles stabilizing the oil drops will experience several forces acting on them: shear forces (from fluid flow through the cake) that result in a torque that pulls the particles away from the oil and partitioning forces that hold the particles at the oil/water interface. In addition, due to the pressure drop across the cake layer, the particle-stabilized oil drops are pushed against the membrane surface, which forces the particles separating the oil from the membrane into the oil drop itself; this force is resisted through the partitioning force which wants to keep the particles at the oil/water interface. Once the forces acting on the oil drops and particles were calculated, we could use this information to predict system performance, as illustrated by our logic flow. Based in this framework we have established the maximum flux the system can operate at and membrane contact angle, considering the cross-flow velocity (15 cm/s) and the surface properties of the uncoated Fe3O4 particles used to stabilize the oil. As can be seen, the model predicts quite well the fouling behavior observed during our experiments, with the polysulfone (PS35) and PAN membranes fouling when operating at 100 LMH, and the PVA-CNT membrane remaining fouling free.