The laboratory bat ch a nd cont i nuous flotation pilot plant t e s t s demonst rat ed t he t echni cal f easi bi l i t y of recovering hi gh gr ade mi ca concent rat es from wea- t hered mi ca pegmatite or e s of Alabama and Georgia. The r esear ch i ndi cat ed t hat combinations of ani oni c and cat i oni c col l ect or s may b e us e d effect i vel y for flotation of f i ne size mi ca from weathered pegmatite ores. In cont i nuous t est s, concent r at es containing 98.5% mi ca were obt ai ned from t he Georgia pegmatite ore; t he Alabama pegmatite ore concent rat es cont ai ned 98.4% mica. The recoveri es were 91 a nd 89% r espec- tively. INTRODUCTION T he principal u s e s of f i ne ground mica ar e a s a fil- l er i n wallboard joint cement, a s a fi l l er and sur- f ace coat i ng for roofing, a s an ingredient in pai nt s, and i n oi l well drilling mud. Th e mineral h a s other us e s in t he manufacture of rubber, wallpaper, pl as- t i cs, welding rods, el ect r i c i nsul at i on, hous e i nsul a- tion, and t ext i l es, and a s an annealing agent in metal treatment. In recent years, more than 99% of t he domest i c mica produced h a s been s cr ap and fl ake mica (mica which doe s not meet speci f i cat i ons for s heet mi ca and is used for producing fi ne ground mica). Ther e h a s been a continued i ncr ease i n mica production for sever al years, t he 1963 production of scr ap and fl ake mica totaling 117,251 tons.' Ground mica is obtained primarily by crushi ng and milling pegmat i t es and s chi s t s . To a l es s er ext ent , mica is produced a s a byproduct of kaolin washing and fel dspar and spodumene flotation operations. Th e pr ocesses used i n recovering mica by crushing and milling pegmat i t es ar e generally simple, consi st i ng of vari ous combinations of al t ernat e roll cr usher s and trommel s cr eens t hat separ at e t he mica and gangue at screen sizes coar ser than 6-mesh. As t he s e proc- esses ar e desi gned t o recover only coar s e mica, high J. S. BROWNING and R. B. ADAIR are respectively Supervising Research Metallurgist and Research Metallur- gist, Tuscal oosa Metallurgy Research Center, Bureau of Mines, U. S Dept. of the Interior, University, ALa TP66F28. Manuscript Octobei 27, 1965. AIME Annual Meeting, February 27 to March 3, 1966, New York, N. Y. Di scussi on of thi s paper submitted i n duplicate prior to December 15. 1966 will appear in AIME Transactions, March 1967 and AIME Transactions, 1967, vol. 238. l os s e s in t he plant r ej ect s ar e common. Lar ge ton- nages of t ai l i ngs from t he crushi ng and screeni ng pl ant s have been accumulated in a number of ar eas. Methods for t reat i ng such products were developed by t he Bureau of Mines i n 1941,2 but have not been generally applied. More recently, mica flotation re- search ha s been completed and published by t he Bureau of mine^.^ Th e s e methods required complete removal of 150- to 200-mesh mat eri al s from t he flota- tion feed with consequent f i ne mica l os s es . Lat er, t he Bureau of. Mines i nvest i gat ed methods for - recovering f i ne s i z e mica from pegmatite or e s after desliming sufficiently t o remove clay materials, but not s o drast i cal l y a s t o remove t he fine sands. Thi s report summarizes t he r esul t s of t he s e st udi es. The pr ocess developed was effect i ve on pegmatite or es from two l ocat i ons and should be appl i cabl e t o t he commercial treatment of other mica-bearing peg- matite or es and f i ne r ej ect s that have been accumu- l at ed at vari ous mica-milling operations. DESCRIPTION OF ORES Th e or e s used in t he investigation were obtained from t he Dixie Mines, Inc., Heflin, Alabama, and t he Ruberoid Corporation, Hartwell, Georgia. Th e sampl e from Alabama contained muscovite and quartz, with a high percent age of clay, and minor amounts of biotite, kaolin, limonite and tourmaline. Th e mica i n t he or e wa s essent i al l y al l minus 4-mesh in s i z e and wa s f r ee of at t ached mineral grains. Th e Georgia sampl e contained muscovite and quartz, with minor amounts of biotite, kaolin, and limonite. The mica i n t he ore, which wa s essent i al l y al l minus 4-mesh, wa s liberated. Pet rographi c anal ys es of t he two sampl es ar e given i n Ta bl e I. THE ANIONIC-CATIONIC MICA FLOTATION METHOD Pr evi ous i nvest i gat ors2 l 3 have reported t hat com- pl et e desliming of mica or es a t 150- t o 200-mesh wa s required prior to flotation with cat i oni c col l ect or s t o obtain sat i sfact ory s el ect i ve separation of t he mica from t he other mineral components. Numerous t e s t s were made a t t he Tuscal oosa Metallurgy Research Cent er t o determine if some reagent combination could b e used t o sel ect i vel y float finer s i z e mica without complete desliming. Th e investigation l ed t o development of a pr oces s using a simple reagent com- Society of Mining Engineers SEPTEMBER 1966 - 277 Tabl e I. Petrographic Anal yses of Alabama and Georgia Mi ca Ores Mineral Anal ysi s, Weight - Percent Ore Musco- Kao- Bio- Limo- Tourma- Locati on vi t e Quartz tin t i t e ni t e l i ne Cl ay - - - Alabama 16.5 35.5 3.0 3. 0 0.5 0.5 41.0 Georgia 15.0 73. 0 2.5 2.0 1.5 - 6. 0 bination that gave a s good or better recovery of mica and grade of concentrate than obtained i n t he com- pletely deslimed process; in most cas es recovery and grade were superior. Briefly described, t he process included conditioning the finely ground ore pulps at 35 t o 40% sol i ds with sodium carbonate and calcium lignin sulfonate and floating the mica with a combination of anionic and cationic collector^.^^^^^ The separation was not particularly sensitive to pulp pH; excellent mica recoveries were obtained in a pH range of 8.0 to 10.5. The function of the sodium carbonate i s to retard flotation of the gangue minerals and control the pH of the pulp. The exact mechanism of t he retarding action of the sodium carbonate i s not known. It seems prob- able, however, that i t s effectiveness may be due to removal and dispersion of slime coatings on the mineral surfaces. The mechanism of the retarding action of the lignin sulfonate has not been definitely determined. It seems probable, however, that selective adsorption of t he lignin sulfonate at gangue mineral surfaces effec- tively prevents adsorption-attachment of t he collec- tor at t hese surfaces. On the other hand, mica parti- cl es being relatively devoid of adsorbed lignin sul- fonate are free to react with collector and are thus made floatable. The lignin sulfonates are effective slime dis- persants and may ai d flotation by assi st i ng in proper removal and dispersion of slime coatings on the mineral surfaces. Anionic type reagents, such a s oleic acid and com- binations of oleic and linoleic acid, were found to be the most effective collecting agents for floating mica. Increased selectivity in the presence of sl i mes was imparted to anionic collectors by incorporating small amounts of cationic amine acetate collecting agent s in the system. The most effective ratio of anionic and cationic collector for mica flotation was 2 to 3 part s fatty acid to 1 part of cationic collector. Any appreciable change in the ratio decreased both t he grade and re- covery of mica. EXPERIMENTAL RESULTS Laboratory Batch Tests: Batch flotation t es t s were undertaken to determine conditions for separating mica from the gangue minerals. Resul t s of preliminary t est s using varying quantities and types of reagents led to adoption of the following procedure: The t est sample was ground in a laboratory rodmill using Tuscaloosa City tapwater which had about 45 parts per million equivalent calcium carbonate total hardness. The sample was st age ground in the pres- ence of sodium hydroxide to aid in dispersing and re- moving clay slimes and screened on a 28-mesh screen until the screen oversize contained only pure mica. After grinding, the sample was partly deslimed by decanting to remove some of the clay slimes from the pulp. The pulp was conditioned at 40% sol i ds in a mechanical - agitation flotation cell for 5 minutes with sodium carbonate and calcium lignin sulfonate for dispersion of the pulp and retardation of gangue minerals. Fat t y acid was then added a s a collector, and conditioning was continued for another 5 minutes, followed by an additional 1 minute conditioning with amine. Thi s was followed by flotation of the mica. The mica rougher concentrate was cleaned three times to yield a final mica concentrate. Summarized results of the t es t s showing the grade of concentrate and mineral recoveries of the pegma- Tabl e II. Laboratory Fl otati on of Mi ca Pegmati te Ores Mi ca Anal y- Di stri - Weight, sis, bution, Product percent percent percent Georgia ore Oversize mica 7.2 100.0 48.3 Flotation concentrate 6. 1 97.4 40. 3 Composite concentrate 13.3 99.2 88.6 Middling 4.2 14.8 4. 0 Tailing 76.2 1. 3 6.7 Slime 6.3 2.0 .7 Composite 100.0 14.9 100.0 Alabama ore Oversize mica 6. 4 100.0 40.6 Flotation concentrate 7. 3 97.8 45.2 Composite concentrate 13.7 98.8 85.8 Middling 6.2 16.8 6. 6 Tailing 40.6 2. 0 5.1 Slime 39.5 1. 0 2. 5 Composite 100.0 15.8 100.0 Con- Con- Con- di - di- di - Rod- tion- tion- tion- mi l l e r 1 e r 2 e r 3 Reagents, lb per ton of ore Sodium hydroxide 1.5 - - - Sodium carbonate - 2. 0 - - Calcium lignin sulfonate - 2. 0 - - Fatty acid (ol ei c + linoleic) - - 0.8 - Amine (stearyl t oleyl) - - - 0.4 Conditioning time, minutes - 5 5 1 Pulp pH - - 9.2 9.1 278 - SEPTEMBER 1966 TRANSACTIONS t i t e or e s from Alabama and Georgia ar e given i n Tabl e 11. Continuous Flotation Tests: GEORGIA ORE-Small s c a l e cont i nuous pilot plant t e s t s were made of t he or e t o confirm t e s t s dat a ob- tained i n t he bat ch laboratory flotation investigation. The plant flowsheet t hat was used in t est i ng t h e Georgia or e i s shown in Fi g. 1. Capaci t y of t he plant was about 150 l bs. of dry feed per hour. Th e minus 4-mesh or e was f ed t o a Humphreys spi r al t o partially concent rat e t he mica and remove a l ar ge part of t he quartz. In addition t o rej ect i ng over 63% by weight of t he or e with a l o s s of only 5% of t he t ot al mica, t he spi ral al s o rej ect ed most of t he limonite in t he ore. T h e spi ral operation upgraded t he mica content of t he or e from 15 t o over 38%. Th e spi ral rougher concent rat e was dewatered and wet-ground i n a rodmill equipped with a 30-mesh trom- me1 screen. Th e grinding was adj ust ed t o t he point where t h e trommel over si ze consi st ed of pure mica. Th e mi nus 30-mesh trommel undersi ze was deslimed by a two-stage desl i mi ng operation t o remove cl ay slimes. Th e or e aft er grinding and desliming cont ai ned about 20% minus 200-mesh. Th e pulp then pas s ed to a conditioner where t he reagent s for pH control, quartz depressi on, and t he fatty aci d a s t he mica collector were al l added. Th e pulp from t he condi- I t Underflow 7 I t Ov e r f l owc - - Bowl rake c l os r ~ l ~ c r Amine Vibmtingr----Toiling- Flolotnon cel l s Pl us 50- mesh , o, l , ng Minus 50- mesh F I o ~ a ~ i o n cel l s toiling 3-stage cleaning 1 1 4 Mica concbntrole Fi g. I-Fl owsheet for Recovery of Mica from Georgia Peg- matite Ore. Tabl e Ill. Results of Continuous Flotation Tests For Recovery of Mica From Pegmatite Ores Mica Analy- Di stri- Weight, sir, bution, Product percent percent percent Georgia ore Trommel oversi ze mica Flotation concentrate Composite concentrate Flotation tailing Hydroseparator overflow Composite spiral concentrate Spiral tailing Composite feed Alabama ore Trommel oversi ze mica Flotation concentrate Composite concentrate Flotation tailing Hydroseparator overflow Composite primary cl assi f i er sand Spiral cl assi f i er overflow Composite feed Rougher Flota- Condi- tion Rodmill tioner Cell 1 Reagents, l b per ton of ore Sodium hydroxide 1. 5 - - Sodium carbonate - 1. 9 - Calcium lignin sulfonate - 2. 0 - Fatty aci d ( ol ei c t l i nol ei c) - .8 - Amine (stearyl t ol eyl ) - - 0. 4 Conditioning time, minutes - 7 1 Pulp pH - 9. 2 9. 0 tioner flowed by gravity t o a bank of t hree rougher flotation c e l l s where t he amine col l ect or was added to t he fi rst cel l . A rougher flotation froth wa s re- covered and cl eaned t hree t i mes t o produce a finished mica concentrate. Th e rougher t ai l i ng pas s ed t o a vibrating screen t o remove t he pl us 50-mesh mica in t he tailing, which wa s returned to t he rodmill for ad- ditional grinding. Th e continuous pilot plant t e s t ac- counted for a recovery of over 91% of t he mica present in t he feed i n a product containing 98.5% mica. Th e summarized r esul t s of t he t e s t a r e given in Tabl e 111. ALABAMA ORE-The flowsheet for continuous treatment of t he Alabama or e included grinding, cl as- sification, conditioning, and flotation a s shown in Fi g. 2. Th e highly weathered pegmatite ore wa s fed t o a bowl rake cl assi f i er t o partially concent rat e t he mica and remove a l arge part of t he cl ay sl i mes. Th e bowl rake cl assi f i er overflow contained over 35% by weight Society of Mining Engineers SEPTEMBER 1966 - 279 Mica ore I + Stomge bin -r t Bawl roke ciassifier -+Overflow A To woste 1 - t 7 30-mesh trammel -Oversize mica 7 Undersize w + Hydroseporotor -Overflow -)To woste I Underflow I I - i Overflow - Spiral classifier I Sodfum corbonate Lignln sulfonate Conditioner Am~ne - 3-stage rough~ng Tolling Middling Flotation cells T o waste 4.sloge cleaning 7 Fig. 2-Flowsheet for Recovery o f Mica from Alabama Peg- mati t e Ore. of the ore with a l os s of only 2% of the total mica. The desliming operation upgraded t he mica content of t he ore from 16 to over 25%. The bowl rake cl assi fi er sand was ground in a rod- mill equipped with a 30-mesh trommel to produce a pure mica trommel oversize. The trommel undersize was deslimed by a two-stage desliming operation. The ore after grinding and desliming contained about 22% minus 200-mesh. The partly deslimed ore, after conditioning at 40% solids, was floated to recover a mica rougher concentrate. The rougher concentrate was cleaned four times to produce a concentrate con- taining 98.4% mica, with a recovery of over 89% of t he total mica. The resul t s of t he continuous t est are presented in Tabl e 3. REFERENCES 1 Benjamin Petkof: Mica. BuMines Minerals Yearbook 1963, v. 1, 1964, pp. 813-829. ' ~a me s I%. Normen and R. G. O'Meara: Froth Flotation and Ag- glomerate Tabling of Micas. BuMines Rept. of In*. 3558, 1941, 14 PP. ~ a me s S. Browning and Thomas L. McVay: Concentration of Fine Mica. BuMines Rept. of Inv. 6223, 1963, 7 pp. 4 ~ a me s S. Browning, Frank W. Millsaps, and Paul E. Bennett: Anionic-Cationic Flotation of Mica Ores From Alabama and North Carolina. BuMines Rept. of Znv. 6589, 1965, 9 pp. 'Frank W. Mlllsaps and James S. Browning: Fatty Acid Fiotation of Mica. U.S. Pat. pending, No. 320,576. 6 James S. Browning and Ralph B. Adair: Flotation of Mica. U.S. Pat. pending, No. 390,485. 280 - SEPTEMBER 1966 TRANSACTIONS