I n d . E n g . C h e m . Res. 1988,27, 1669-1674
1669
Staged Flotation of Fine Coal and the Effects of Mineral Size and Distribution Shirley C. Tsai Department of Chemical Engineering, California State University, Long Beach, California 90840
This paper presents the experimental results of a staged froth flotation for deep cleaning of fine coal. In this staged flotation, the slowly flotable, mineral-rich particles are first rejected, and the resulting overflow is recleaned a t lower flotation rates through careful control of frother concentration. As a result, a clean coal product of lower ash and lower pyritic sulfur contents has been produced a t a higher yield as compared with the single-stage flotation and other reported multistage flotation circuits. The selectivity of this staged flotation is dictated by the mineral size and its distribution in the feed coal. The mineral particles that remain in the clean coal have a volume mean diameter of 6 pm, with 90% being smaller than 13 pm. I. Introduction Froth flotation with very complex flotation circuits (Dorenfeld, 1962; Barbery, 1984) has long been a commercial technology in mineral processing. Since coal is much less valuable than minerals, only simple single-stage froth flotation has been incorporated in a coal preparation plant. The coal fed to the flotation circuit in the coal preparation plant is usually deslimed tq remove particles smaller than 75 pm (200 mesh). In contrast, the particle sizes required in both direct coal combustion and for use in coal-water slurry fuel are utility grind coal with 80% of the particles being smaller than 75 pm (Manfred et al., 1983; Ekmann and Wildman, 1985; Tsai and Knell, 1986; Tsai and Vu, 1987). We have shown that such fine coal particles (density of 1.2-1.5 as compared to 2.5-5.0 for minerals) that are well dispersed in water can be selectively froth floated, while leaving particles of a high mineral content in the tailings (Tsai, 1985). The observed effects of surface chemistry, particle size, and density on flotation yield, flotation rate, and product quality are consistent with the pseudo-first-order kinetics with two rate constants. However, we have observed a significant loss of selectivity when the fine coal feed consists of ultrafine mineral-rich particles. This problem is especially severe in deep coal cleaning because fine grinding the coal to smaller than 75 pm is required to liberate the minerals. In order to overcome this problem, various multistage flotation circuits that are typical in mineral processing (Dorenfeld, 1962; Barbery, 1984) were employed in our laboratory. We found that a rougher/cleaner/scavenger configuration, called R/C/S staged flotation as shown in Figure 1, gave considerable advantages of low ash and high yield over single-stage and other multistage flotation circuits for bituminous coals. In the R/C/S staged flotation, the tailings (underflow)from the rougher (R) are rejected, while the overflow is fed to the cleaner (C). Subsequently, the overflow from the cleaner is collected as part of the clean coal product, and the corresponding tailings are fed to the scavenger (S)for additional coal recovery. It should be noted that removal of the slowly floating, mineral-rich particles (first-stage tailings) in the rougher is imperative for the staged flotation to achieve a high yield of clean coal. Furthermore, we found that recycle of product streams resulted in deterioration of clean coal quality. Therefore, the R/C/S staged flotation constitutes the best and the simplest froth flotation scheme for deep cleaning coal with particles smaller than 75 pm. Through utilization of deeply cleaned coal in coal-water slurry fuel, the coal market may be extended to boilers that were designed originally for oils with minimum retrofit and derating.
In this paper, the experimental methods and results of the R/C/S staged flotation are first presented in section 11. Comparisons with both single-stage and various multistage coal flotations as well as sink/float gravity separation are described in section 111. Also discussed in section I11 are the effects of mineral composition, mineral particle size, and distribution on the selectivity of the R/C/S staged flotation. In addition, the pyrites that are finely distributed in the froth-floated coal are chemically very reactive and readily accessible in some coals such as Kentucky No. 9 (Tsai, 1986). Therefore, an assessment of the desulfurization potential by air/steam treatment as a precombustion step is given in section IV. 11. Experimental Methods and Results The coals used in this experimental study are highvolatile bituminous A coal (Pond Fork coal from West Virginia, and West Kentucky No. 9) and medium-volatile bituminous coal (Virginia Pocahontas No. 3). They had been washed in a conventional preparation plant. The coals came in 5-15-cm sizes and were pulverized in a bowl mill to 80%, passing through a 200-mesh screen. The pulverized coal was subsequently stored under N2 atmosphere in tote bins or sealed jars to minimize surface oxidation. After more than 2-month storage, the oxygen content of the pulverized coal and the pH of its water suspension were found to remain unchanged. Both the rate and the yield of flotation under typical conditions (Tsai, 1985) without kerosine conditioning were also found to remain unchanged in 1 month, but to decrease only slightly after more than 2-month storage as a result of surface oxidation. In order to avoid the interference caused by surface oxidation, each parametric study on the pulverized coal was completed within a month. Unless specified otherwise, the coal particle sizes used are commonly referred to as “utility grinds“, which generally means that 80% of particles are smaller than 75 pm. The specific size distribution as measured by Leeds and Northrop Microtrac Particle Size Analyzer for Pond Fork coal is given in Table I. In addition, the run-of-mine Pond Fork coal (as received) screened to three size fractions, 28 X 60 mesh (0.5 mm to 250 pm), 60 X 100 mesh (250-150 p m ) and 100 X 200 mesh (150-75 pm), was also used in order to provide a direct comparison of the R/C/S staged flotation with the conventional single-stage flotation. Also shown in Table I is that the low-temperature ash obtained from the West Kentucky coal is much smaller (volume mean diameter of 8 versus 20 pm) than that from the Pond Fork coal. Therefore, a finer grind, which is referred to as -400 mesh or
