Table of Contents
The ores used in the investigation were obtained from the Aluminum Silicates, Inc., Washington, Ga., and the Commercialores, Inc., Clover, S. C.
The sample from Georgia contained kyanite and quartz with clay pyrophyllite, limonite, pyrite, and rutile. Petrographic examination showed that the kyanite was essentially; liberated between -35 and 48 mesh. In the screen sizes between 48 and 100 mesh a few of the, kyanite grains had small attachments of quartz.
The mineral constituents of the sample from South Carolina were kyanite, quartz, mica, clay, pyrite, limonite, and rutile. The kyanite was essentially liberated at 65 mesh. In the screen sizes between 65 and 200 mesh, a few of the kyanite grains had small attachments of quartz.
Petrographic analyses of the two samples are, given in table 1.
Laboratory Batch Tests
Preliminary batch flotation tests were made of the two ores to determine the optimum conditions for separating the kyanite from the gangue minerals. The use of varying quantities and types of collectors and depressants, conditioning at different time and pulp solids, and varying the pH of the pulp were investigated. Results of preliminary tests led to adoption of the following procedure:
The test sample was ground in a laboratory rod mill to minus 35 mesh, using Tuscaloosa City tapwater which had about 45 parts per million equivalent calcium carbonate total hardness, and then partly deslimed by decanting to remove some of the clay slimes from the pulp. The pulp was conditioned at 40 percent solids in a mechanical-agitation cell with sulfuric acid, xanthate, and pine oil for pyrite flotation. A pyrite rougher concentrate was floated and cleaned once to remove gangue minerals. The combined pyrite tailing and middling next was deslimed by decantation before conditioning and floating the kyanite. The deslimed sands were conditioned for 5 minutes with 2.5 pounds of sulfuric acid per ton of feed for pH control and quartz depression. Petroleum sulfonate, 1.5 pounds per ton of feed, was then added as a collector, and conditioning was continued for another 5 minutes. This was followed by flotation of the kyanite. The kyanite rougher concentrate was cleaned 3 times to yield a final kyanite concentrate. Summarized results of the tests showing the grade of concentrate and mineral recoveries of the kyanite ores from Georgia and South Carolina are given in table 2.
The function of the sulfuric acid is to retard flotation of the gangue minerals and control the pH of the pulp. The pH should he acid in the range below 4.0. Above a pH of 4.0, the flotation of kyanite becomes less efficient.
The oil-soluble sodium petroleum sulfonates, which constitute the principal collecting agents in our research, are commercially available in a number of different forms. The preferred group of sulfonated petroleum hydrocarbons is known as mahogany soap. This group is characterized in that its constituents are usually oil soluble but water dispersible. Suitable sulfonates usually contain 60 to 70 percent sulfonate, 25 to 35 percent mineral oil, 2 to 5 percent water, and less than 0.5 percent inorganic salts.
Continuous Flotation Tests
A small scale continuous pilot plant was operated using the petroleum sulfonate-sulfuric acid flotation procedure found to be successful in laboratory batch tests. Capacity of the plant was 150 to 200 pounds of dry feed per hour. The flowsheet included grinding, classification, desliming, conditioning, and flotation, as shown in figure 1.
The ¾-inch dry crushed ore was stored in an ore bin and withdrawn by a feeder onto a vibrating screen. The minus ¼-inch screen undersize passed to a spiral classifier to remove clay slimes. The plus ¼-inch screen oversize and the spiral classifier sand passed to a rod mill where the kyanite ore was ground in closed circuit with a 26-mesh, trommel. The minus 26-mesh trommel undersize was pumped to a cyclone for additional desliming. The cyclone underflow passed to the pyrite flotation circuit for removal of pyrite. The pyrite tailing was deslimed in a bowl-rake classifier. The desliming operations and removal of pyrite upgraded the kyanite content of the ore from 24 to over 34 percent. The bowl-rake classifier sand passed to the conditioners where sulfuric acid was added for pH control and quartz depression, and petroleum sulfonate was added as the kyanite collector. The retention time in each conditioner was about 5 minutes. The conditioned feed passed to a bank of 4 rougher cells in which a rougher concentrate and a finished tailing were produced. The rougher kyanite concentrate was cleaned 4 times to produce a kyanite concentrate. The middling was pumped to a cyclone for thickening, and the cyclone underflow recycled through the conditioners. The kyanite concentrate was attritioned with sodium hydroxide and screened on 270 mesh for removal of iron slimes created in attritioning. The results of the continuous test are given in table 3. In this test 87 percent of the kyanite was recovered in a concentrate analyzing 54.6 percent Al2O3. A sizing analysis of the concentrate is given in table 4.
Flotation of the kyanite from the deslimed sand fed to flotation was excellent, as less than 1 percent of the kyanite was lost in the tailing. The kyanite losses in desliming consisted primarily of soft iron stained kyanite.
South Carolina Ore
The flowsheet for continuous treatment of the South Carolina ore included grinding, classification, desliming, conditioning, and flotation, as shown in figure 2. The grinding and classification circuit for the South Carolina ore was the same as that used for the Georgia ore. The minus 26-mesh trommel undersize was fed to a Humphreys spiral to reject coarse pyrite and limonite. The spiral tailing passed to the pyrite flotation circuit for removal of fine size pyrite. The pyrite tailing was deslimed in a “bowl-rake classifier. The desliming operations and removal of pyrite upgraded the kyanite content of the ore from 24 to 30 percent. The bowl-rake classifier sand passed to the conditioners, where the reagents for pH control, quartz depression, and kyanite collection were added. The pulp from the conditioners flowed by gravity to a bank of 4 rougher flotation cells in which a rougher concentrate and a finished tailing were produced. The rougher concentrate was cleaned 4 times to produce a kyanite concentrate. The middling was pumped to a cyclone for thickening, and recycled through the conditioners. The kyanite concentrate was attritioned with sodium hydroxide and screened on 270 mesh for removal of iron slimes created in attritioning. The results of the continuous test, as shown in table 3, indicate that over 82 percent of the kyanite was recovered in a concentrate analyzing 54.3 percent Al2O3. A sizing analysis of the concentrate is given in table 4.
High-Intensity Magnetic Separation of Kyanite
Kyanite used for ceramic purposes should normally contain less than 1 percent hydrochloric acid soluble Fe2O3. Additional specifications normally require a minimum of percent Al2O3 and 5 to 8 percent plus 35-mesh material. Inasmuch as the concentrate produced in the plant tests assayed from 2.8 to 4.2 percent acid soluble Fe2O3, high-intensity magnetic separation tests were made to reduce the Fe2O3 content of the kyanite concentrates. The magnetic separation tests were made in a laboratory high-intensity wet magnetic separator. The iron-bearing kyanite was removed as a magnetic product. The results of magnetic separation of the pilot plant flotation concentrates are given in table 5. Treatment of the Georgia kyanite concentrate yielded a nonmagnetic fraction assaying 56.9 percent Al2O3 and 0.9 percent acid soluble Fe2O3, with a recovery of 95.2 percent of the alumina. Treatment of the South Carolina kyanite concentrate yielded a nonmagnetic fraction assaying 55.9 percent Al2O3 and 0.6 percent acid soluble Fe2O3, with a recovery of 96.4 percent of the alumina.
The laboratory and continuous pilot-plant flotation tests demonstrated, the technical feasibility of recovering kyanite concentrates analyzing 92 percent kyanite from Georgia and South Carolina ores. Recoveries of 87 and 82 percent of the kyanite were obtained from the Georgia and South Carolina deposits, respectively. Petroleum sulfonate, in an acid circuit, was successfully used to obtain improved flotation selectivity and facilitate recovery of the kyanite.
The concentrates produced in the pilot plant analyzed from 2.8 to 4.2 percent acid soluble Fe2O3 High-intensity magnetic separation produced products containing less than 1 percent acid soluble Fe2O3, and analyzing 56.9 and 55.9 percent Al2O3 from the Georgia and South Carolina ores, respectively. The nonmagnetic concentrates meet all specifications for commercial grade kyanite products.