4. Conclusions
The effect of transmembrane pressure, shearrate, dispersed and continuous phase viscosity on the final droplet size and flux has been investigated for a rotating membrane emulsification system. O/W emulsion droplets of average diameter23.4–216.6 μm have been produced using an SPG membrane of 6.1 μm pore diameter. In this article, a number of complex processing effects have been discussed through consideration of the fluid flow and interfacial behavior of the two phases. Considering the dispersed phase flow behavior effects on droplet size, four distinct regions can be seen across the range of transmembrane pressures investigated. Firstly, a decrease in size across low pressures (at approximately o0.5 bar)which is attributed to coalescence at the membrane surface during long droplet formation times. Secondly, a plateau in size known as the ‘size stable’ zone which occurs due to the spontaneous transformation based droplet formation mechanism in system swith high interfacial tension(i.e. with silicone oils).Thirdly, an eventual increase in droplet size as significant mass is transferred via the droplet neck during detachment. The volumetric contribution during this stage depends primarily on the droplet detachment time at the dispersed phase flux and thus can become negligible at increased rotational speeds(41000RPM).This is due to higher drag and centrifugal forces to ensure earlier detachment and displace droplets quickly away from the membrane surface. It is also suggested that viscous dispersed phase droplets may experience a higher drag coefficient since they resemble more rigid spheres; an assumption within drag force calculations. Finally, if the interfacial tension is low and dispersed phase flux (or more specifically the pore fluid velocity) is sufficiently high, jetting of the dispersed phase may occur(approximately 41.5bar).In this case, the Capillary number defining this flow behavior approaches the threshold value of 0.056 that is previously suggested in literature.
The flux through the membrane increases with high pressure and low dispersed phase viscosity as defined by Darcy'slaw. Values between 50 and 12,500Lm 2 h 1 were measured for oil viscosities between 0.012 and 0.085 Pa s across applied transmembrane pressures between 0.1 and 1.8 bar. For SPG membranes, this relationship between flux and pressure is exponential rather than directly proportional since membranes demonstrate a variable permeability. The percentage activation of tortuous pore channels depends on the applied transmembrane pressure. The calculated valuesf or active pore fraction ranged between 0.5% and 2.7% coinciding with previous findings for SPG membranes. The continuous phase flow behavior also determines the droplet size produced primarily by altering the drag force acting on the droplet. By increasing the shear rate (through higher rotational speeds and narrower annular gap widths),the droplet size tends to a minimum independent of the applied shear as commonly observed within cross-flow membrane emulsification. However, the rotational speed contributes an additional centrifugal force to enable even earlier detachment from the membrane surface and thus further reducing the droplet size. Another key consideration is where Taylor vortices can form and in their absence below a critical Taylor number of 41.3, droplet sizes are significantly larger. This is possible particularly for lows peed, narrow gap or high viscosity continuous phase systems. Finally, if there is a significant density difference between the two phases (dispersed phase being typically lowerfor O/W systems) or if the aqueous continuous phase viscosity is high, there is a potential for droplets to not be displaced sufficiently away from the membrane surface but remain within the vicinity of other forming droplets. This is likely to increase coalescence phenomena, which is generally unfavourable duringemulsification processing.