The effects of flotation variables on the bubble size, mixing characteristics and froth behaviour in column flotation cells

Abstract

Column flotation has only come into vogue in the last decade, and in the last five years there has been a dramatic increase in the number of plant installations on mines throughout the world. Due to the heightened interest in this technology it is important that the factors influencing the performance of column flotation cells be properly quantified. Most of the column studies to date have focused on the pulp zone, while far less work has been done on the froth zone. There are conflicting views about the importance of various froth parameters. For example, some researchers hold that froth depth is of little importance, while others have found it to be crucial. The aim of this study was to examine the processes occurring in the column as a whole, but to put special emphasis on the froth zone. With respect to the pulp zone, a system was developed for measuring the bubble size distribution in the pulp. It involved detecting about 3000 bubbles as they were sucked up a glass capillary tube past two optical detectors. The accuracy of the system was found to be excellent, and the number of bubbles sized made the distributions statistically meaningful. Bubble size tests were conducted in a small laboratory column cell in both water and pulp. The trends for the two- and three-phase systems were the same, although the bubbles in the pulp tests were generally bigger than those in water. It was found that factors which led to a decrease in the viscosity or surface tension in the pulp resulted in smaller bubbles. This was because the size of a bubble is determined by the rate at which a film of water can form around a pocket of air, which is directly influenced by the viscosity and surface tension of the pulp. Residence time distribution studies of the collection zone were conducted in a laboratory column flotation cell using both liquid and solid tracers. The liquid tracer studies showed that the residence time distributions in the pulp zone were hardly affected by the presence of solids in the column. When the solid tracer was used, however, it was found that there was a significant difference between the residence time distributions of the solid and liquid phases. The solids were significantly more mixed than the liquid and they had a shorter mean residence time. It was found that the residence time distributions in the pulp could be adequately modelled using a simple tanks-in-series model. The axial dispersion model was not used because some of the D/uL values found in this study were too large for the model assumptions to be valid. It was generally found that flotation parameters which led to an increase in the mean residence time in the pulp also led to a decrease in the degree of mixing in the column. The effects of various flotation parameters on the behaviour of the froth zone were examined using froth stability tests and by splitting the froth at various levels. The stability of the froth was determined by the maximum froth height achieved in a small batch column. A system was developed to split the froth into four horizontal segments so that the froth could be examined at different levels. This system was used on both a batch and a continuous column and it allowed data to be obtained on the grade, the recovery, the solids and water mass pull, the air hold-up, and the particle size distribution as a function of height in the froth. The froth stability tests showed that it is a combination of the particle size and hydrophobicity that determine the stability of the froth. The stability of the froth increased as the particle size decreased because there were more particles available to inhibit froth drainage and bubble coalescence. It was also found that the froth stability passed through a maximum as the xanthate collector dosage was increased. The optimum collector dosage depended on the collector chain length and the particle size. The hydrophobicity of the particles increased as the collector chain length and dosage increased. As the particle size increased the particles needed to be more hydrophobic to produce the same froth stability. A fine synthetic ore comprising 90 percent quartz and 10 percent pyrite was used in the batch froth splitting tests. It was found that the coarser pyrite tended to float first with little entrained material reporting to the top of the froth, and the amount of entrained material in the froth increased towards the bottom of the froth. Ultra-fine pyrite (<10 �m) was concentrated at the bottom of the froth indicating that this material was either entrained or slow floating. It was again found that there was an optimum collector type and concentration to achieve the best grades and recoveries caused by poor collection on one hand and the froth being too tightly packed for selective flotation on the other hand. It was found that froth stability played an important role in determining the final grades and recoveries. If the froth was too stable then elutriation of entrained material from the froth was difficult, and if it was not stable enough then floated material was lost due to bubble coalescence in the froth. In the continuous froth cutting tests the synthetic pyrite/quartz ore was again used along with a real gold bearing pyrite ore. The tests were conducted with and without wash water. It was found that the trends observed in the batch tests were generally repeated in the continuous tests with no wash water. Where differences were observed, these could be ascribed to the depletion of pyrite in the batch tests which did not occur in the continuous tests. The addition of wash water made a significant difference to the characteristics of the froth. The froth was stable at much lower frother concentrations and the amount of bubble coalescence was greatly reduced. There was a far more uniform grade profile through the froth and a significant reduction in the amount of entrained material in all levels of the froth, especially near the bottom. Residence time distribution studies of the froth zone were conducted using an isotopically labelled flotable solid tracer and a salt solution liquid tracer. The liquid tracer showed that very little of the water in the feed reported to the concentrate. The solid tracer studied showed that the froth zone could not be modelled adequately using a one parameter model and thus a two parameter model was chosen. The model consisted of a perfect plug flow reactor with a recycle stream. The model was chosen because it could model any situation from a completely mixed system to a perfect plug flow system. It was found that the mean residence time in the froth and the plug flow reactor's residence time could be related, and the degree of mixing in the froth and the reactor recycle ratio were also related. These relationships further validated the choice of model because the model parameters could be directly linked to the residence time distribution in the froth. It was again found that factors that increased the mean residence time in the froth also decreased the extent of mixing in the froth. There was a linear increase in the concentrate grade and a linear decrease in concentrate mass pull as the mean residence time in the froth increased. The recovery remained virtually constant. This indicates that the froth zone is responsible to a large extent for the grade in the column, and that longer froth residence times helped reduce the levels of entrained material in the concentrate. An examination of the effects of flotation parameters on the metallurgical performance of the column showed that the mean residence time in the pulp was most affected by the pulp density and the air rate had the biggest effect in the froth. The volumetric feed rate and the wash water rate had the biggest influence in the degree of mixing in the pulp and froth respectively. The concentrate grade was most affected by the air rate, while the volumetric feed rate had the biggest effect on the mass pull to the concentrate. The recovery tended to pass through an optimum, with the frother concentration having the biggest effect on recovery. The wash water rate had a large effect on the mean particle size in the concentrate.