\r' 916s PLEASE DO NOT REM.OVE FRCM .LIBRARY ~ iRII 1 ~ Bureau of Mines Report of Investigations/1988

Preparation of Ammonium Paratungstate From a Sodium Tungstate-Sodium Chloride Phase

By A. E. Raddatz, J. M. Gomes, and T. G. Carnahan

g,. S. Department of the Intenor., Bureau of Mines Spokane Research Center East 315 Montgomery Avenue Spokane, WA 99207 UNITED STATES DEPARTMENT OF THE INTERIOR Report of Investigations 9165

Preparation of Ammonium Paratungstate From a Sodium Tungstate-Sodium Chloride Phase

By A. E. Raddatz, J. M. Gomes, and T. G. Carnahan

UNITED STA"rES DEPARTMENT OF THE INTERIOR Donald Paul Hodel, Secretary

BUREAU OF MINES David S. Brown, Acting Director Library of Congress Cataloging in Publication Data:

Raddatz, A.E. (Andrea E.) Preparation of ammonium paratungstate from a sodium tungstate­ sodium chloride phase.

(Report of investigations; 9165)

Bibliography.

Supt. of Docs. no.: I 28.23: 9165.

1. Ammonium paratungstate. 2. Tungsten-Metallurgy. 3. Solvent extraction. I. Gomes, J. M. (John M.) II. Carnahan, T. G. (Thomas G.) III. Title. IV:' Series: Report of investiga· tions (United States. Bureau of Mines) ; 9165.

TN23.U43 [QD181.W1] 622 s [669'.734] 87-600413 CONTENTS

Abstract...... •...... •...... 1 Introduction ...... ,...... 2 Acknowledgment ...... , ...... , 3 Materials and experimental procedure...... 3 Results and discussion...... 5 Conclusions...... 8 Appendix ...... • -..•••..• -••-...... 9

ILLUSTRATIONS

1. Tungsten production from scheelite and wolframite concentrates...... 2 2. Proposed flowsheet for solvent extraction process to recover APT from wolframite via high-temperature molten-salt extraction...... 3 3. Tungsten extraction from synthetic solution as a function of time...... 5 4. Tungsten extraction from actual halide solution as a function of time.... 5 5. EqUilibrium diagram for extraction of tungsten with Primene JM-T...... 6 6. Combined crossflow-countercurrent flow solvent extraction circuit...... 7

TABLES

1. Analysis of wolf rami te concent ra·te ••••••••••••••••••••••••••••••••••••••• 4 2. Composition of diluted pregnant quench-leaching solution for solvent

extraction ...... •...... ,...... tI •••••••••••••••••••••••••• 4

3. Tungs t en bala nee ..•....•..•...•...... 1'1 •••••••••••••• ,. •••••••••• 6 4. Tungsten extraction to the organic phase in combined crossflow-counter­ current flow solvent extraction system at phase ratio of 0.5 ••••••••••• , 7 5. Analyses of Bureau of Mines and commercial APT ••••••••••••••••••••••••••• 7 A-I. Analyses of silicate phases of wolframite concentrates treated by the high-temperature molten-salt extraction technique ••••••••••••••••••••••• 9 A-2. Reagents used in leaching silicate phase and extraction •••••••••••••••••• 9 A-3. Analyses of wolframite processing residues ••••••••••••••••••••••••••••••• 10 A-4. Summary of reagents used and composite extraction results from leaching wolframite residues...... 10 UNIT OF MEASURE ABBREVIATIONS USED IN THIS REPORT

°C degree Celsius min minute g gram mL milliliter giL gram per liter pct percent h hour vol pct volume percent

L liter PREPARATION OF AMMONIUM PARATUNGSTATE FROM A SODIUM TUNGSTATE-SODIUM CHLORIDE PHASE

EJy A. E. Raddatz,1 J. M. Gomes,2 and T. G. Carnahan3

ABSTRACT

Previous Bureau of Mines research has shown that tungsten ores con­ t_aining as little as 40 pct W032- can be processed by a high-temperature molten-salt extraction technique to produce a tungstate-bearing sodium chloride phase with 99 pct W recovery and little impurity contamination. The objective of this work was to demonstrate that the tungstate-bearing sodium chloride phase can be a suitable feed material for preparing am­ monium paratungstate (APT) by a modification to the present industrial solvent extraction process. A combined crossflow-count·ercurrent flow solvent extraction technique to extract tungsten is presented. APT pro­ duced by this new technique contained 88.6 pct W032- and the following impurities, in percent: Mn02_ 0.0007, Mo02 0.017, P205 <0.010, and Si02 0.012. Implementation of this processing technique would improve tung­ sten recovery from concentrates and simplify processing operations.

2Supervisory metallurgist (retired). 3Supervisory physical scientist. Reno Research Bureau of Mines, Reno, NV. 2

INTRODUCTION

APT, 5(NH4)20'12W03'5H20, is the first decomposition and most of the purifi­ intermediate product in the preparation cation are accomplished in one step by of tungsten metal or tungsten carbide. this technique. Wolframite or scheelite More than 75 pct of all tungsten consumed is decomposed in a two-phase sodium in the United States is first processed chloride-sodium metasilicate melt at to APT. Figure 1 shows how APT is pre­ 1,050° C. The sodium chloride and sodium pared by alkali pressure digestion of metasilicate phases are immiscible at scheelite or wolframite concentrates to this temperature and allow metals separa­ yield a sodium tungstate solution. The tion to be achieved because tungsten is solution is purified to remove harmful concentrated in the halide phase, while impurities such as molybdenum, phospho­ iron, manganese, and calcium report rus, arsenic, and silica. The purified to the silicate phase. The two molten solution is treated by a solvent extrac­ phases are separated by decantation. The tion process (liquid ion exchange) to molten silicate phase is quenched in wa­ recover a crystalline APT product. ter to recover the small amount of water­ One of the problems with the conven­ soluble sodium tungstate that would be tional solvent extraction process is pos­ lost with the silicate phase. The tung­ sible pollution of the environment from state-bearing halide phase is readily alkaline effluents and discharged res­ soluble in water, and tungsten is recov­ idues. A second problem arises when ra­ erable from the resulting pregnant and dioactive minerals contained in some wol­ quench solutions by the conventional framite ores generate residues requiring solvent extraction technique described special handling and _storage permits. above. Figure 2 shows a flow sheet for A third problem is the time, cost, and the proposed process sequence. tungsten losses in purifying the pregnant Substitution of the Bureau's high-tem­ solution to make it an acceptable feed perature molten-salt extraction process for the solvent extraction step. Silica, for the alkali digestion and solution molybdenum, arsenic, and phosphorus are purification steps in the conventional removed in separate precipitation and process would eliminate the pollution filtration operations which incur a 2·- to problems and save processing time. 3-pct loss of tungsten. Another potential benefit is that the A high-temperature molten-salt extrac­ residue from the molten-salt extraction tion technique previously developed by process step is a refractory silicate the Bureau of Mines offered a new ap­ slag that contains strategic commodities proach for preparing APT from schee­ such as columbium, scandium, and tin. lite and wolframite concentrates that Since this residue is insoluble in water could overcome these problems. 4 Ore and resists weathering, it could be stored without degradation and processed 4Gomes, J. M., K. Uchida, and D. H. later for recovery of its contained me­ Baker, Jr. A High-Temperature, Two-Phase tallic values. A cursory study of possi­ Extraction Technique for Tungsten Miner­ ble techniques to recover these commodi­ als. BuMines RI 7106, 1968, 13 pp. ties is presented in the appendix.

Scheelite Ferrotungsten CaW04 Digestion Crude Purification Pure Solvent extraction W chemicals )---=---- Na2W04 .. Na2W04 .. APT solution solution { W metal Wolframite WC (Fe,Mn)W04

FIGURE 1.-Tungsten production from scheelite and wolframite concentrates. 3

H20 to Wolf rami te concent ra te silicate washing

Nael

Evaporation

Raffinate Solvent extraction 1-1----,.- Organic solvent Washin'9

Granulated silicate product

Further processing for Ammonium paratungstate Cb, Ta, Sc, Fe, Co, Mn recovery

HALIDE PHASE SILICATE PHASE

FIGURE 2.-Proposed flowsheet for solvent extraction process to recover APT from wolframite via high-temperature molten-salt extraction.

The objective of this investigation was material with less purification of the to show the feasibility of using the pregnant solution. A combined crossflow­ tungstate-bearing halide phase prepared countercurrent flow solvent extraction from a domestic wolframite ore as a feed circuit was devised to extract tungsten material for the commercial solvent ex­ efficiently from the sodium chloride­ traction process and to determine if bearing solutions. quality APT can be prepared from this ACKNOWLEDGMENT The authors thank L. E. Schultze, discussions that led to the design of the supervisory chemist, Reno Research combined crossflow-countercurrent flow Center, Bureau of Mines, Reno, NV, for solvent extraction unit. MATERIALS AND EXPERIMENTAL PROCEDURE A wolframite concentrate was decomposed Because limited quantities of actual by the high-temperature molten-salt ex­ halide leaching solution would be avail­ traction technique to produce samples of able due to the small amount of the the halide phase, which were dissolved to halide phase, initial solvent extraction prepare solutions for the solvent extrac­ experiments were conducted with synthe­ tion process step. Analysis of the wol­ tic tungstate-bearing halide solutions to framite concentrate showed that iron and determine the effect of pH on tungsten manganese were the only major impurities extraction. This research was conducted (table 1). 4

TABLE 1. - Analysis of wolframite TABLE 2. - Composition of diluted preg­ concentrate, percent nant quench-leaching solution for solvent extraction, grams per liter 1 Constituent Constituent Constituent Constituent W03 z- •••••• 67 MgO •••••••• 0.04

Alz03' ••••• .5 MnOz· •••••• 16.6 W03Z- •••••• 42 MgO ••••••• /II <0.010 ASZ03······ <.004 MoOz· •••••• .03 NaCI ••••••• 105 MnOz· •••••• .002 CaO •••••••• <.01 Na zO • •••••• .014 Alz03' ••••• <.030 MoO z • •••••• .002 Cb zO 5' ••••• 1. 22 PZ05' •••••• .08 AsZ03······ <.020 PZO 5' •••••• <.050 Fe 203' ••••• 10.1 SiO z • •••••• 1.0 CaO ••••••• ~ <.030 SiO z • •••••• <.020 K20 •••••••• <.05 TiO z • •••••• .4 Fe zO 3' •••.• <.001 TiO z • •••••• <.002 K20 •••••••• .260 before the composition of the actual ha­ lide phase leach solution was established Reported as grams of oxide per liter and its composition was considerably dif­ to be consistent with current industrial ferent from the synthetic solution's com­ practice. position. The equilibrium diagram used for designing the solvent extraction sys­ increased tungsten losses. Increasing tem was generated from actual leaching the solvent extraction temperature to solution. Solutions were prepared by 55° C did not decrease emulsion formation diswolving sodium tungstate and sodium or increase tungsten extractions. Analy­ chloride in water. Technical-grade anhy­ sis of the halide solution showed that drous NazW04 and food-grade sodium chlo­ iron and manganese were removed in the ride were used. Analyses of the solution high-temperature solvent extraction step gave the following concentrations, in (table 2). grams per liter: W03Z-, 32; NazO, 34; Primene JM-T was chosen as the organic and NaCI, 48. extractant on the basis of Bureau re­ Actual halide phase leaching solutions search and industrial practice, which used for the solvent extraction study indicated that tungsten could be ex­ were prepared by dissolving the halide tracted from alkali solutions with pri­ phase with the aqueous solution used to mary amines. 6 Composition of the organic quench the silicate phase. A total of solvent was 5 vol pct Primene JM-T di­ 1,070 g of halide phase was dissolved in luted in 85 vol pct kerosene with 10 vol­ 3 L of quench solution at room tempera­ pct decanol as a phase modifier. ture in 3 h. The resultant halide solu­ Solvent extraction experiments were tion was diluted to twice its volume to conducted with 1,000-mL separatory fun­ optimize tungsten loading of the organic nels at room temperature. Prior to ex­ phase. Exploratory tests showed that the traction, the pH of each halide solution presence of NaCl would depress loading of was adjusted to 2 or 4 with sulfuric acid tungsten. For example, Primene JM-T5 and continually monitored while the ad­ solvent could be loaded to 55 g W03z- per justment was made. The organic solvent liter from a solution containing 62 g was conditioned by washing with an equal W032- per liter in the absence of NaCl, volume of 1~ HZS04' The wash was dis­ but its loading was limited to 28 g carded. The conditioned solvent was W03Z- per liter in the presence of 153 g NaCl per liter. High concentrations of 6Churchward, P. E., and D. W. Bridges. sodium chloride also caused formation Tungsten Recovery From Low-Grade Concen­ of a tungsten-laden third phase that trates by Amine Solvent Extraction. Bu­ interfered with phase separation and Mines RI 6845, 1966, 17 pp. Yih, S. W. H., and C. T. Wang. Tung­ 5Reference to specific products does sten Sources, Metallurgy, Properties, and not imply endorsement by the Bureau of Applications. Plenum Press (NY), 1979, Mines. pp. 79-128. 5

gently shaken in a separatory funnel with crystals were recovered b¥ filtration and the acidified halide solution for 15 min washed with water to remove traces of the to ensure phase equilibration. After NH40H strip solution. The quantity of phase separation, the raffinate was sam­ wash water was kept minimal because APT pled, and the loaded solvent was washed is water soluble. The volume of wash with an equal volume of water to remove solution was measured and samples were traces of entrained raffinate and sam­ taken to determine tungsten recovery. pled. Tungsten was stripped from the Solutions and solids were analyzed by loaded solvent by gentle agitation for X-ray fluorescence for tungsten. Other 10 min with an equal volume of 3N NH40H. elements were determined by atomic ab­ The loaded strip solution was evaporated sorption spectroscopy, inductively cou­ by boiling to decrease its volume by one­ pled plasma spectroscopy, and emission half and held at room temperature for spectroscopy. at least 12 h to crystallize APT. APT RESULTS AND DISCUSSION Preliminary experiments using the syn­ lower than at pH 2. During the first 2.5 thetic halide solution were conducted to min of contact time, 87 and 75 pct of the determine appropriate experimental con­ tungsten was extracted at pH 2 and 4, ditions for solvent extraction. Timed respectively; with extraction slowly in­ extraction tests were conducted to deter­ creasing with time. Residual .tungsten mine the effect of contact tIme on the concentration in the raffinate was higher extent of extraction at pH values of 2 and the contact time slightly longer for and 4. Contact time was varied from 1 to the actual solution than for the synthe­ 20 min. The results are shown in figure tic solution. These differences were due 3. Two minutes were required to equili­ to higher sodium chloride concentrations brate the system. Almost identical char­ in the actual halide solution. Phase acteristics were measured at pH 2 and 4. separation was much better at pH 4 than While differences were small, extraction at pH 2. after 2 min contact time was slightly The remaining solvent extraction exper­ better at pH 2 (98 pct) than at pH 4 (92 iments were conducted with the actual pct). Figure 4 shows results of solvent halide solution at pH 4 because of better extraction when applied to the actual phase separation. Contact time was set halide solution. At pH 4, extraction was at 15 min to ensure near eqUilibrium con­ ditions. Phase ratios (volume of organic phase to volume of aqueous phase) inves­ tigated were 1:10, 1:7.5, 1:5, 1:1, 5:1, 40r------~1------'1------~1------, 7.5:1, and 10:1.

50r------r------r------~

0 +- 0 0-. 25 Z 0 1= () 50 « pH 2 pH 4 a:: - z l-x pH 4 pH2 ,.., 10 75 w 0 ,.., "J ~ ---_____ 0 1 0 :---___-;;- _____---,:':- ____-,1 100 ;:: 5 10 15 20 0 5 10 15 CONTACT TIME, min CONTACT TIME, min FIGURE 3.-Tunsten extraction from synthetic solution as a FIGURE 4.-Tungsten extraction from actual halide solution function of time. as a function of time. 6

TABLE 3. - Tungsten balance, percent

O-A ratio, vol pct ••••• 1:10 1:7.5 1: 5 1: 1 5:1 7.5: 1 10:1 Raffinate •••••••••••••• 89 85 79 21 0.2 0.1 0.4 Loaded organic wash •••• .02 .02 .03 .2 1 1.5 2 Depleted strip solution 3 4 9 25 17 22 15 APT •••••••••••••••••••• 7 9 11.5 47 74 69 71 Losses ••••••••••••••••• 1 2 .5 7 8 7.5 11.5

An equilibrium curve was determined in by using a phase ratio less than 1, but a set of single-contact tests (fig. 5). tungsten extraction would be incomplete. The curve is characterized by a steep This required a modification of the nor­ slo~e up to organic phase loading of 30 g mal multiple-contact countercurrent sol­ W03 - per liter or more and then becomes vent extraction flow diagram. flat, which indicates favorable solvent A novel solvent extraction system was extraction loading similar to nonchloride devised to extract the tungsten effi­ tungstate solutions. 7 ciently at phase ratios less than 1 (fig. During development of the equilibrium 6). The solvent stream was divided into curve, stable emulsions were observed three multiple streams, two of which when organic-to-aqueous phase ratios were needed only to contact the tungstate greater than 1. This poor phase separa­ solution once to be fully loaded. The tion manifested itself in excessive tung­ third stream was employed in the nor­ sten losses (table 3). Tungsten losses mal countercurrent manner. Calculations ranged from 0.5 to 11.5 pct and were (fig. 5) indicated that for a feed solu­ acceptably small (2 pct or less) only at tion of 42 g W032- per liter and phase phase ratios of 0.2 or less. Crud forma­ ratio of 0.5, the aqueous W032- concen­ tion accounts for a majority of the tung­ tration would be decreased to approxi­ sten losses. Emulsions could be avoided mately 10 giL after the second stage of the crossflow portion of the circuit. 7F irst work cited in footnote 6. This W032- concentration is now low enough to allow the use of a McCabe­ Thiele diagram (fig. 5) to determine the number of stages needed in the normal ~ 30 Ol countercurrent flow solvent extraction c5 train. Using an operating line with a Z 25 slope of 2 and a feed concentration of

Wash solution

r------~----~: o"an: wash 11 ___o_r_g_a_n_ij~S_O_lv_e_n~tr- ______~

L...--"'Wash effluent

.! CROSSFLOW I-oA_q.:..u_e_o_u_s _s_o_'_ut_i o_n_-, C OUN T E RC U R RENT L----R f f' t Aqueous feed- CIRCUIT CIRCUIT ,~ a Ina e STAGE 1 I STAGE 2 STAGE' I STAGE 2

Loaded solvent

I Recycle solvent I Stripping I f Loaded stripping solution

FIGURE a.-Combined crossflow-countercurrent flow solvent extraction circuit for maintaining overall phase ratio at less than 1.

TABLE 4. - Tungsten extraction to TABLE 5. - Analyses of Bureau of Mines the organic phase in combined and commercial APT, percent crossflow-countercurrent flow solvent extraction system at Bureau Commercial phase ratio of 0.5, percent WO 3 2 - •••••••••••••• 88.6 88.3 AS203·············· <.004 <.004 Stage Extraction CaO •••••••••••••••• <.005 <.005 Fe 2°3- ••••••••••••• <.0004 <.0004 Crossflow: K 20 •••••••••••••••• <.02 <.02 Stage 1...... 37 MnO 2- •••••••••••••• .0007 <.00005 Stage 2...... 40 MoO 2* •••••••••••••• .017 <.0005 Countercurrent flow: Na 20 ••••••••••••••• <.005 <.005 Stage 1...... 20 P 205 •••••••••••••• ., <.010 <.010 Stage 2...... 2 Si02··············· .012 .009 Total...... 99 Ti 02 ••••••••••••••• <.0005 <.0005

The stripped organic solvent contained product as 5(NH4)20·12W03·5H~ with a less than 0.01 g W032- per liter. After possible trace of 5(NH4) ~·12WO 3·11H~. crystallization of APT from the ammonium A comparison of the APT product pur­ tungstate solution, the APT product was ity with a commercial APT produced from analyzed for W032- content and impuri­ scheelite is given in table 5. The Bu­ ties. A typical product contained 88.6 reau product was only slightly higher in pct W032-. Stoichiometric W032- content manganese, molybdenum, and silica. The of the pentahydrate form of APT is 88.1 extremely low manganese value in the pct. X-ray diffraction identified the commercial APT is because its feed, 8 scheelite, has little manganese, whereas concentrations reach problem levels, the the wolframite used by the Bureau con­ aqueous feed solution can be treated by tained 16.6 pct Mn02' Of the most trou­ sulfide precipitation methods currently blesome impurities--molybdenum, phospho­ used in the industry to decrease molybde­ rus, silica, and sodium--only molybdenum num contamination. 8 Sodium can be elimi­ and sodium could present problems because nated by washing the loaded organic sol­ they are coextracted with tungsten dur­ vent with a dUute 5-g/L ammonium sulfate ing solvent extraction. If molybdenum solution. 9

CONCLUSIONS

Previous studies showed that tungsten phase, containing 99 pct of the tungsten, minerals are decomposed at 1,050° C in a can be dissolved in the solution from sodium chloride-sodium silicate melt and quenching the silicate phase to produce a yield two immiscible phases, a tungstate­ tungstate solution. Tungsten can be re­ bearing sodium chloride phase and a sili­ covered by solvent extraction with a pri.­ cate phase. This investigation demon­ mary amine extractant, stripped with an strated that the tungstate-bearing sodium ammonium hydroxide solution, and crystal­ chloride phase can be utilized as feed lized as APT from the strip solution. for preparing APT by the conventional The Bureau's APT product contained 88.6 solvent extraction process. The halide pct W032-.

8Second work cited in footnote 6. 112th AIME Annu. Meet., Atlanta, GA, Mar. 9Hughes, M. A., and C. Hanson. The 6-10, 1983, 11 pp.; available from M. A. Fate of Impurities in a Liquid-Liquid Ex­ Hughes, Sch. of Chern. Eng., Univ. of traction Process for Tungs ten. Pres. at Bradford, Bradford, west Yorkshire, UK. 9

APPENDIX

The silicate phase from the high­ a scanning electron microscope (SEM) temperature molten-salt extraction pro­ showed-- cess represents a potential source of strategic commodities such as columbium, • No distinct columbium phase was scandium, and tin (table A-I). Another found. resource in the tungsten industry is wol­ framite processing residues. A cursory • Predominant phase was a manganese­ investigation into possible techniques to iron silicate with inclusions of manga­ recover these commodities from each feed nese, iron, chromium, and titanium. was conducted with the following results. • Tungsten occurred in distinct sponge-like grains.

The molten-salt extraction technique Since columbium was not observed in for the decomposition of tungsten ores discrete grains, it was concluded that it separates tungsten from iron, manganese, was disseminated throughout the Mn-Fe columbium, and other minor constituents silicate phase. Techniques were evalu­ found in the ore. More than 90 pct of ated for the recovery of columbium. A the components other than tungsten are summary of the reagents studied and the concentrated and recovered in the sili­ composite extractions obtained are pre­ cate phase. Analyses of silicate phases sented in table A-2. No extraction oc­ obtained by treating these wolframite curred with NH4Cl, (NH4)2S04, KHS04, KOH, concentrates are shown in table A-I. and Na2C03' The most successful tech­ nique involved mixing five parts silicate TABLE A-I. - Analyses of silicate phase with one part sodium fluoride or phases of wolframite concentrates potassium fluoride, roasting at 1,050° C treated by the high-temperature for 2.5 h, quenching in 1.5 L of water, molten-salt extraction technique, and leaching the quenched silicate in a percent 0.5-L 25-vol-pct sulfuric acid solution at 90° C for 3 h. Sixty-six percent of Constituent 1 2 3 the columbium was extracted. The tech­ WO 3 2 - ••••••••••••• 0.8 0.4 0.7 nique was not selective because iron and A120 3' .•.....•.••• .4 1.7 1.2 manganese were also solubilized. Scan­ CaO ••••••••••••••• 1.0 7.1 .7 dium was extracted with a FeS04-H2S04 Cb205 ••••• •••••••• .9 .3 .9 lixiviant or a sodium fluoride roast and Fe203··.······ •••• 13.6 15.3 12.2 H2S04 leaching technique • K20 ••••••••••••••• ND • 06 .07 MgO ••••••••••••••• .5 .3 .4 TABLE A-2. - Reagents used in Mn 0 2' ••••••••••••• 25.3 8.6 16.2 leaching silicate phase and Mo02 ••••••••• ••• •• <.01 .01 .02 extraction, percent Na20 •••••••••••••• 18.2 17.2 20.4 P205·············· .01 .07 .07 Technique and reagent Cb Fe Mn Si 02' ...... 43.7 35.9 26.6 Roast-leaching: TiOz· ••• •••••••••• 1.0 .6 .9 KF-HzS04···· ••••••••••• 66 88 80 Sc .•••.••••.•••••• .2 • 05 ND NaF-H2S04··· ••••••••••• 66 90 78 Sn ••••••••••••••.• .3 1 • 2 Leaching: ND Not determined. FeS04 •••••••.•. • ••.•••• 0 25 2 FeS04+H2S0 4 •••••••••••• 0 29 63 Mineralogical examination of the non­ H 280 4' ••••••••••••••••• 24 79 90 magnetic portion of silicate phase 3 with H2S03" ••••• CI ••••••••••• 0 80 86 10

Wolframite Processing Residues leaching extracted 95 pct of the manga­ nese, 91 pct Co. and 81 pct Sc. Colum­ Industrial residues from processing bium and tin remained in the leached wolframite concentrates contained colum- residue. bium, tantalum, cobalt, and scandium in The best columbium extraction, 72 pct, a high iron-manganese matrix and had the was achieved with a sodium fluoride roast analyses shown in table A-3. at 700 0 C followed by a 10-pct sulfuric acid leach at 90 0 C. This method also TABLE A-3. - Analyses of wolframite extracted cobalt and scandium. but not processing residues, percent tantalum. Selectivity was poor because iron and manganese were extracted. A Constituent 1 2 3 4 process sequence intended to selec­ W03L.-···.··· • 3.3 2.1 5. 1 2.5 tively recover manganese. iron. and co­ Al 203' ••••••• 2.0 2.0 7.0 2.0 lumbium was devised and included a sul­ Bi 203' ••••••• .6 <.2 <.2 <.2 furic acid roast and water leaching step CaO •••••••••• 12 5.2 5 •.5 2.1 followed by an acidic FeS04 leach and Cb 20 5 •••••••• 1.3 .5 .8 2.0 concluded with a NaF-H2S04 leach. Al­ CoO •••••••••• .3 1.3 0 0 though the tests were successful when Fe 203' ••••••• 29 30 39 38 tried individually. columbium was not MgO •••••••••• 2.8 5.0 .3 .2 extracted when they were combined MnO 2••••••••• 15 23 19 22 sequentially • Sc ••••••••••• • 03 <.01 .01 .01 SiO 2' •••••••• 13 10 11 1.6 TABLE A-4. - Summary of reagents used SnO 2' •••••••• .5 .08 .4 .9 and composite extraction results Ta 20 5' ••••••• .4 .09 .2 .6 from leaching wolframite residues. Au ••••••••••• 1.02 0 1 .05 0 percent Ag ••••••••••• '1.4 1 .6 1.6 1• 1 l.Ounce per ton. Technique Cb Co Fe Mn Sc and reagent Mineralogical examination of residue 1 Roast-leaching: identified wolframite. ferberite, cas­ H2 S0 4-H20 •••••• 0 91 2 95 81 siterite, scheelite, arsenopyrite, xeno­ K2CO 3-HCl. ••••• 0 ND 85 20 0 time, columbite and tungsten-bearing KOH-HCl •••• ~ ••• 13 ND 72 48 0 columbite. NaF-H 2S0 4' ••••• 72 78 20 53 76 Techniques were investigated to treat Leaching: the processing residues. A summary of Fe SO 4' ••••••••• 0 2 67 51 0 the reagents evaluated and composite ex­ FeS04+H2S04'" • 0 ND 94 96 ND tractions obtained from the four residues 30 pct H2SO 4'" ND 30 86 92 0 are given in table A-4. No extractions 60 pct H2SO4'" 47 84 93 95 0 occurred in atmospheric leaching with NaF+H 2S04'" ••• 57 49 75 53 48 Na2C03, Na2C03+NaOH, or NaOH. A sulfuric ND Not determined. acid roast at 950 0 C followed by water

INT.-BU.OF MINES,PGH.,PA. 28678