Natural or collectorless flotation of native or free gold is generally possible, at least to a degree, with many gold containing ores. The first factor in such flotation studies is the quantification of what the limitations are of particle size on the flotation process. Some unique studies have been conducted by this author at a number of plants. Figures 1 and 2 show the results of a kinetic study done on a free gold ore from Africa. The laboratory work was done using was a large three liter batch cell. All of the data taken was at a natural ore pH of 7.3 (no pH regulators used of any kind). Figure 1 summarizes the time recovery data with no Collector as a function of gold particle size range and Figure 2 the same kind of data with 10 g/Mton of ethyl isopropyl thionocarbamate added to the grind. The frother used in this study was a polypropylene glycol methyl ether (PGME) of 200 molecular weight. The nominal grind size for the data of both figures was 52.5% < 75 microns which is deliberately coarse so as to have a broad range of gold particle sizes in the flotation feed.
 |  |
The batch time recovery data of Figures 1 and 2 has been curve fitted to a modified first order rate model frequently used by this author and others in kinetic studies, Klimpel (1980, 1989, 1995) and Dowling et. al. (1986). The model is given by:
ri = R (1 – (l/(Kti ))(1 – exp(-Kti ))]
where ri and ti are paired sets of experimental cumulative recovery and cumulative time data, respectively. The curve fitting parameters are R (the equilibrium recovery at long flotation time) and K (a modified first order rate constant having units of time). Thus one complete time recovery profile can be described by one set of R and K values each with associated error limits. Comparing different time recovery profiles generated under different sets of experimental conditions then becomes a more simple problem of statistically comparing sets of R and K values. It usually is quite easy to see trends in the R and K values with organized experimental test programs. In the testwork of Figures 1 and 2, the approximate 95% confidence limits on K are ± 0.26 and for R are ± 0.017. The least squares estimate of the R and K values for each profile are also given in the figures.
Inspection of Figures 1 and 2 show a number of particle size trends that are common to both gold flotation and sulfide mineral flotation such as with chalcopyrite and molybdenite. To begin with, smaller gold particles float much faster than larger particles and also give higher equilibrium recoveries. This is true whether one has a collector present or not. The role of any collector with inherently naturally floating minerals such as the free gold in Figure 2 is to speed up the the rate of recovery of all particle sizes with the relative impact being greatest on the coarser particles. Also, the presence of the collector helps to increase the equilibrium recoveries achievable with the greatest impact again being on the coarsest particles. The recovery of native gold particles below 150 microns in many plants is often surprisingly good whether a collector is used or not.