Mechanisms of Phase Separation for Dispersions in Continuous Flow

Abstract

The literature pertaining to droplet hydrodynamics and coalescence affecting the separation of primary liquid-liquid dispersions has been reviewed. The behaviour of droplets forming a flocculation zone has been studied in specially constructed 3, 6 and 9 inch diameter vertical glass gravity settlers. Characteristics of flocculation zones, namely hold-up aid the variation of drop sizes and residence times, have been determined and correlated with zone heights for 4 different systems. Difficulty was encountered initially in obtaining reproducible results due to the effects of interface scum and colloidal impurities. This was attributable to system ageing accelerated by sunlight and was minimised by blacking out the transparent sections of the equipment. Initial experiments were performed with the system toluene-water to find the minimum practical settler diameter. Significant differences, due to wall effects, were observed in columns below 6 inches in diameter and therefore work was not pursued in the 3 inch column. The effect of countercurrent | phase flow on zone height was found to be insignificant, hence subsequent work employed a stationary continuous phase. Experimental techniques included high speed still and normal speed cine photography and the injection of dyed droplets. A novel attempt was also made to encapsulate drops in a close packed swarm. Flocculation zones were observed to comprise 3 distinct sections i.e. an inlet section wherein droplets form a close packed arrangement, a larger mid-section where inter—drop coalescence and droplet distortion occurs, and an exit section bounded by the interface. The differences in vertical hold-up profiles between these sections was measured. Total zone height has been correlated by equations of the form, H=a.VdbDpcd12e Deviations occurred with drop sizes in the lower end of the range studied, i.e. 0.7 mm to 1.0 mm. A flocculation zone model is postulated leading to a relationship between drop residence time and hold-up which is in agreenent with the experinental results. Consideration is also given to conditions under which a modified fluidised bed model is applicable.