Kernel functions to flotation bubble size distributions

Abstract

Flotation modelling has advanced from deterministic single particle-bubble models into using such models to solve flotation systems by using modern computational techniques. The step from a single particle- single bubble event to multiple events taking place in the large computational volume like a flotation cell poises the challenge of handling bubble and particle distributions in all computational cells. The estimation of bubble size has either been omitted (constant size) or has been lately estimated by a population balance approach. The physical performance of flotation is excessively determined by the bubble size distribution (BSD). Therefore, the bubble size distribution estimate is crucial for modelling. Although the BSD can be measured, the underlying effects of different variables causing changes in break-up and coalescence rates producing changes in the measured BSD's are not well understood. This paper discusses the profound effects frothers have on both the coalescence and break-up of gas bubbles. Depending on the bubble surface stiffness caused by frother adsorption, the drainage rate of fluid between two approaching bubbles is very different. Frothers like DF200 and Pentanol have a higher coalescence rate than frothers like DF250 and NF240. Break-up is shown to be a function of the dynamic surface tension, not the static surface tension. Fast adsorbing frothers (DF200) have at very short time scales a higher rate of break-up. The paper suggests a division of frothers into two distinct classes for modelling purposes. Those with fast adsorption and desorption, which leave the gas-air interface mobile and those frothers that by slower adsorption and desorption create stiff interfaces. The effects in real systems may be more varied. The modelling of subtler frother effects will not substantially improve modelling quality.