Mechanistic understanding of the rotating membrane emulsification process towards the development of design and scale-up theory

The effect of processing and formulation parameters on the resulting oil-in-water, emulsion microstructure has been studied for a recently developed process; rotating membrane emulsification. A broad range of surfactant and particle dispersions were explored to reveal the key drivers that determine the final droplet size produced. The aim of the study was to understand initial droplet generation and therefore emulsion stability, whilst a significant element in emulsification studies, is not considered here. By furthering the understanding of the processing mechanisms involved, this enabled development of theoretical models to estimate droplet size and extent of coalescence from first principles. In addition, the implications of process scale-up were studied. From this work, the very first design procedure for rotating membrane emulsification was derived and proposed. The final emulsion microstructure is heavily dependent on the coupled interaction between the fluid flow behaviour of the two phases and interfacial phenomena. Careful selection of process parameters based on sufficient characterisation of properties such as interfacial tension and viscosity, can avoid the occurrence of droplet coalescence or dispersed phase jetting. These can have a detrimental effect on producing a carefully controlled microstructure on a repeatable basis. Of particular importance is the rate of surfactant adsorption at the oil/water interface. A unique approach of dispersing non-ionic, high HLB surfactants such as Tween 20 and Brij 97 within the oil phase has been found to significantly reduce droplet size. This discovery allows the process to be highly competitive with a rotor-stator high shear mixer and an ultrasonic probe at a fraction of the energy consumption. Pilot-scale operation of rotating membrane emulsification provided important insight into how one might design and therefore implement the process for an industrial purpose. It is proposed here that a suitable scale-up parameter would be the membrane surface velocity