Membranes as separators of dispersed emulsion phases

The reuse or discharge of industrial waste waters, containing small fractions of dispersed oil, requires a purification treatment for which membranes can be used. If only little oil is present, removal of the dispersed phase might be preferable to the more commonly applied removal of the continuous phase. For this purpose dispersed phase separators can be applied, which combine the features of conventional coalescers and membrane filtration. The membrane surface promotes coalescence (similar to a coalescer) but in the mean time the coalesced phase is separated and transferred in a continuous oil permeate phase.In the present thesis the possibility to use sheets of polypropylene membrane (with a pore diameter of 0. 1 μm) as dispersed phase separators for treatment of secondary (bromo-) hexadecane-in-water emulsions stabilized by Tween-40 is investigated. The weight fraction of the dispersed phase is always lower than 0.06 and the volume to surface averaged diameter of the oil droplets is generally smaller than 10 μm. The research focuses on understanding the mechanisms controlling the permeation behavior of the oil droplets. The process can be described in two stages: firstly the droplets have to be transported form the bulk of the emulsion towards the membrane and secondly they have to coalesce and permeate.In chapter 2 and 3 the transport mechanisms are determined both experimentally and theoretically. It is shown that the theories developed for collectors of solid particles, based on the convective diffusion equation, can be used to describe the transport behavior of the oil droplets in membrane modules. The main transport mechanisms are diffusion, gravity and interception. The latter occurs if the location of the stream lines and the droplet size results in contact between droplet and membrane. The influence of several parameters, such as droplet size, density of the dispersed phase, flow rate and module design can be explained using the convective diffusion theory, assuming the membrane to be a perfect sink. This means that as soon as a droplet reaches the membrane it coalesces and permeates. Since theory and experimental results agree qualitatively, it can be concluded that the system is transport limited in most cases.In chapter 4, the coalescence mechanism is discussed. Only if the contact time between droplet and membrane becomes smaller than the time needed for coalescence, the system will be coalescence limited. The coalescence time, t c [s], is determined by the drainage of the aqueous film between droplet and membrane. Calculations show that t c for the small droplets under investigation against a liquid interface is smaller than I second while in case of the presence of a membrane t c will be even smaller. As the contact times are at least a few seconds, coalescence will occur which is in agreement with the perfect sink assumption. Only in case of a dead end module at high flow velocities the system becomes coalescence limited because of short contact times. Finally, at high surfactant concentrations coalescence is not detected.As oil droplets are slightly negatively charged, the transport can be enhanced by electrophoresis. Therefor, in chapter 5 possibilities to apply an electric field over the emulsion are investigated. If the anode is placed at the feed side of the membrane the transport velocity is increased significantly. However, not all droplets permeate and a cream layer is formed in front of the membrane. If the anode is placed at the permeate side of the membrane, problems are encountered because of the large resistance of the continuous oil permeate phase. The electric field is situated mainly over this phase, at the cost of the electric field over the emulsion. The resistance of the permeate phase can be successfully decreased by addition of an apolar electrolyte (TBAI). Also, it is possible to circulate an aqueous phase at the permeate side. In this case, oil droplets with a diameter of several millimeters are detected in the permeate. Here, the membrane acts as a conventional coalescer. All systems show indeed an enhanced extraction in the presence of an electric field. However, we think that only the first and the last are scaleable.Finally, in the general discussion the dispersed phase separator is compared to conventional emulsion separation methods. The dispersed phase separator becomes advantageous at low oil concentrations. Feasibility will increase if the remaining oil fraction in the retentate can be decreased further. Ideas for optimization in case of transport limitation are introduced, such as adjusting the flow profile to increase the interception mechanism. Problems are encountered in practice because of the presence of all kinds of impurities in industrial waste waters which will hinder the coalescence. In that case, research should focus on coalescence enhancing parameters of the membrane material such as surface potential and surface roughness. Preliminary experiments using a reflectometer indicate the importance of surface roughness, which is explained theoretically.