(k Bureau of Mines Report of investigations/l984

A Basic Chloride Method for Extracting Aluminum From Clay

By P. R. Bremner, L. J. Nicks, and D. J, Bauer

UNITED STATES DEPARTMENT OF THE INTERIOR Report of Investigations 8866

A Basic Chloride Method for Extracting Aluminum From Clay

By P. R. Bremner, L. J. Nicks, and D. J. Bauer

UNITED STATES DEPARTMERIT OF THE INTERIOR William P. Clark, Secretary

BUREAU OF MINES Robert C. Worton, Director Library of Congress Cataloging in Publication Data:

Bremner, P, R. (Paul R,) A basic chloride method for extracting aluminum from clay.

(Bureau of Mines report of investigations ; 8866) Bibliography: p. 8. Supt. of Docs. no.: I 28.23:8866. 1. Alumit~urn-Metallurgy. 2. Leaching. 3. Chlorides. 4, Kao- linite, I. Nicks, 1;. J. (Larry J.), 11. Bauer, D. J. (Donald J,). 111. Title. IV. Series: Report of investigations (United States. Bureau of

TN23.U43 [TN776] 622s [669'.722] 84-600004 CONTENTS .Page Abstract ...... 1 Introduction ...... 2 Materials. equipment. and procedures ...... O.O...... ,...... 3 Results and discussion...... b 3 Effects of calcination time and temperature ...... 3 Single-stage leaching and crystallization...... 4 Countercurrent leaching and crystallization...... 5 Purification and solubility studies ...... *...... 6 Thermal decomposition of ACHH ...... 7 Discussion and conclusions ...... 8 References ...... 8 ILLUSTRATIONS 1. Effects of clay calcination temperature ...... *...... 4 2 . Single-stage leaching ...... e...... b...... ,..m...... e....e...... 4 3. Test array for three-stage countercurrent leaching ...... 5 TABLES 1. Composition of dried kaolinitic lay...... ^...... ^...^... 3 2 . Summary of single-stage 30-pct AlG13 leaching ...... 5 3 . Summary of countercurrent leaching results ...... 6 4 . Impurity concentrations in cell-grade alumina and ACHH ...... 7 UNIT OF MEASURE ABBREVIATIONS USED IN THIS REPORT

OC degree Celsius mL milliliter

pct percent

wk week A BASIC CHLORIDE METHOD FOR EXTRACTING ALUMINUM FROM CLAY

By Pe Re Bremner, Le Ja Nicks, and De JI ~auer~

ABSTRACT

As part of a program to devise alternative technologies for producing alumina from domestic resources, the Bureau of Mines investigated the leaching of calcined kaolinitic clay to form basic aluminum chloride solutions. Bench-scale experiments were conducted for single-stage batch leaching with AlCl, solution and three-stage countercurrent leach- ing with substoichiometric HC1. Variable conditions of clay calcination and leaching were studied to determine their effects on aluminum extrac- tion and pregnant liquor composition.

With clay calcined at the best temperature of a500 C, extraction of more than 93 pct of the aluminum was achieved. The basic chloride preg- nant liquors were crystallized and yielded the compound 5AlC1,-8Al(OH), -37.5H20, which can be thermally decomposed to alumina with 61.6 pct of the energy required for decomposition of A1C13*6H20.

In comparison with the leaching operation in an HC1-clay process, the basic chloride system offers substantial advantages in energy consump- tion and equipment size requirements.

-lMetallurgis t. 2~upervisoryresearch chemist. 3~upervisorychemical engineer. Reno Research Center, Bureau of Mines, Reno, NV. INTRODUCTION Almost all aluminum in the United In the HCl-clay process, kaolin is cal- States is produced from imported bauxite cined to an amorphous mixture of alumina or alumina. This country has, however, and silica, The calcined kaolin is more than adequate domestic nonbauxitic leached with HC1 solution, and aluminum alumina resources that could alleviate is dissolved according to the following dependency on foreign sources (11.4- For reaction: both economic and security reasons, the Federal Government has had a longstanding interest in the development of alterna- tive technology that would allow utiliza- The Al.C13 solution is evaporated or tion of those resources. sparged with HC1 gas, and the solid, A1C13*6H20, is formed. In 1927, the Bureau of Mines published a comparative study on extracting alu- During bench-scale research on KC1 minum from clay with sulfuric, nitric, leaching of calcined clay, sometimes the and hydrochloric acids (2). Alumina re- apparent extraction of h.1203 was greater search was resumed duriGg World War 11, than 100 pet, Examination of the data when the Defense Plant Corporation funded indicated that aluminum could be dis- four pilot plant studies (3) and the solved from kaolin with solutions con- National Bureau of Standards, under the taining less HC1 than required by reac- sponsorship of the Army Signal Corps, in- tion I. The resulting solutions con- vestigated an HC1 process for extracting tained basic aluminum chloride. alumina from kaolinitic clay (4).- A literature search revealed reports on In 1973, a program was initiated by the the solubility of A1(OH)3 and Alz03 in Bureau of Mnes to evaluate the technical solutions of AlClj (7). A study by and economic potential of several pro- Breuil (8) defined the compositions of cesses for producing alumina from domes- several ~?~stallinebasic aluminum chlo- tic resources (5). An HCl-clay process rides obtained by extended aging of mix- was judged the yest application of exist- tures of Al(OH)3, AlC13, and H20 at dif- ing technology (6). At the same time, ferent temperatures. bench-scale invesFigations were conducted on alternative technologies which might The formation of the basic chlorides of avoid some of the disadvantages of the aluminum may be represented by either of HC1 processes. the following reactions: A key operation in the HCl-clay process is the thermal decomposition of aluminum chloride hexahydrate (ACH) to the final alumina product. This step is energy in- tensive, and the viability of the process would be improved if the energy require- ment for decomposition was decreased. One possible way to achieve the decrease is to modify the process so that its where X + Tb = 3. Reaction 2 is a general aqueous operations did not produce ACH, expression for reaction 1 and leads to but a basic chloride with a composition the basic chloride when reactive A1203 is between ACH and alumina. in excess. men X is greater than 312, the coefficient for 820 becomes negative and water is consumed as the reaction $underlined numbers in parentheses re- proceeds to the right. Equation 3 ex- fer to items in the list of references at presses the direct reaction of AlC13 with the end of this report. A1203, which also consumes water. The value of Y is the molar ratio of C1:Al. The objectives of this study were to This ratio will be used as an determine the best C1:Al ratio for ex- indicator of composition, For example, tracting aluminum from calcined kaolin, C1:Al for ACH is three, for alumina it and to determine what, if any, basic alu- is zero, and the basic chlorides are minum chloride solids could be recovered intermediate. from the pregnant solution. MATERIALS, EQUIPMENT, AND PROCEDURES Raw clay was obtained from the Theile in a preheated electric furnace. After Kaolin Co., Sanderville, GA, and is rep- the specified time had elapsed, the cru- 1 resentative of very large deposits of cible was removed and allowed to air- eastern Georgia kaolinitic clay. The clay cool. For larger samples the clay was was dried for 48 h at 100" C, ground with roasted for 4 to 8 he a disk pulverizer to minus 60 mesh, and blended. Analysis by a combination of Leaching tests, both single-stage and wet-chemical methods and inductively countercurrent, were made in a 0.5-or 1-L coupled plasma emission spectroscopy gave glass resin kettle heated with an elec- the results shown in table 1. tric mantle and vented through a water- cooled condenser. The calcined clay TABLE 1. - Composition of dried charge and leaching solution were placed kaolinitic clay, percent in the kettle, heated to boiling, and stirred for the specified time. The test A1203...... ee.eee..e.... was terminated by vacuum filtration of Si02...... ******.***.*** the hot contents of the kettle on a Buch- Ti02.e...... ner funnel. The filter cake was washed Fe203...... -.~~~~~~~~~~~~~~ with three 100 mL portions of water acid- K20.e.....meee...e.eee-*e0eee0-.0e ified to pH 3 with HC1. The combined P205..e.P..e.eeee.e*e**********I* filtrate and wash were evaporated to ,con- MgOs.e*..eeeeeeeeeeeeee-**oeeeeee centrate the contained aluminum salts, CaO...... e...... ee.Ce~. At the first sign of crystallization, the hlaO...... e..e.....e...... liquor was allowed to cool at room tem- Total analyzed...... perature overnight. The crystals were Theoretical hydrate H20...... filtered from the liquor, washed by slurrying in propanol, refiltered, and air-dried. The leach residue, combined filtrate, and crystals were analyzed for The dried clay was calcined prior to Al, C1, and the major impurities Fe, Mg, leaching. The calcined clay contained P, Ca, Na, and K. 41.1 pct A1203. In tests to determine the effects of calcination time and Reagent-grade HC1 and ACH were used in temperature, a 70- to 100-g sample of the tests . clay in an uncovered crucible was placed RESULTS AND DISCUSSION EFFECTS OF CALCINATION TIME kaolinite structure collapses and loses AND TEMPERATURE hydrate water amounting to a theoretical 13.95 pet. The resultant material is The effects of thermal treatment on the almost amorphous to X-ray diffraction, acid solubility of kaolinite are well and its alumina content is almost totally established (9). The natural mineral is soluble in acids, On further heating, no resistant to %id attack. Qn heating to detectable changes occur until about 900" 400" C, kaolinite exhibits little change C when mullite, gamma alumina, and crys- other than minor losses of free moisture. tobalite begin to form, At this point, Between 425" and 525" C and depending the acid solubility decreases abruptly. on the degree of crystallinity, the The solubility window between calcina- SINGLE-STAGE LEACHING tion temperatures of 500' and 850" C is AND CRYSTALLIZATION illustrated in figure 1, The lower curve shows the weight loss for dry kaolin Three series of leaching tests were after heating to the temperatures indi- made with 25- to 200-g charges of cal- cated. The upper curve is the percent of cined clay and 72 g AlC13 as a lo-, 20-, aluminum extracted by leaching with 26- or 30-pct solution. The amount of leach- pct HC1. The same type of behavior was ant for each series was 720, 360, or expected when initial experiments were 240 g, and the initial C1:Al ratio for conducted with an AlC13 leachant for kao- the total material in the leaching re- lin calcined at temperatures within the actor for each test ranged from 2.18 to window. Poor reproducibility of extrac- 0.75 within a series. All test solutions tion results led to a controlled calcina- were maintained at boiling and stirred tion study. The results of that studji for 2 h. Results of these tests are are also shown in figure 1, shown in figure 2. Aluminum extraction ranged from about 57 to 87 pct and in- The calcination temperature was criti- creased with increasing concentration of cal for extraction of aluminum from kao- AlClj in the leachant at constant initial lin with a 30-pct solution of AlC13. C1:Al. The extraction decreased when the Rather than a plateau region of solubil- initial C1:Al was decreased for each ity, extraction with AlC13 increased with AlC13 concentration, Aluminum extraction temperature until a sharp decrease is maximized by leaching with a high ini- occurred at more than 850" C, tial C1:Al ratio and high concentration of chloride. The effect of calcination time at 850' C on extraction was studied over the The lower line in figure 2 shows the range 0.25 to 9 h, Extraction of alumi- relationship between the C1:Al of the num was determined for a 30-pct AlCl, pregnant liquor and the initial ratio. solution and attained a constant level The liquor ratios obtained were parallel for kaolin calcined for more than 1 h, to and about 0.2 unit higher than the No decrease in extractibility caused by initial C1:Al. The liquor ratio was not overroasting was observed for samples influenced by the AlC13 concentration of calcined for 9 h, All subsequent leach- the leachant, ing tests were made with clay calcined at 850" C for 4 to 8 h.

T80 with 26-pet HCI

Clay weight loss 0 20 A 30 0 I I I 0 2.5 2.0 1.5 1.0 0.5 CALCINATION TEMPERATURE, OC lNlTIAL CI:AI RATIO FIGURE 1, - Effects of cI ay calcination temperature, FIGURE 2. - Single-stage leaching, In tests with initial C1:Al ratios leachant in the tests summarized in table less than about 1.5, difficulties were 2 was initially just below its saturation encountered because of the high initial concentration of 31 pct, which is equiva- solids content. The problem became worse lent to an aluminum concentration of 6.3 as the test progressed and water was con- pct. The pregnant liquors from the sumed by reaction 3. Filtration and single-stage tests were not analyzed for washing of the larger amounts of leach aluminum before being diluted with wash residue also were difficult. water, but from the amounts of alumina dissolved, the aluminum concentration of The time required for dissolution was the liquor was significantly higher than determined in a series of tests with 30- 6.3 pct. Since aluminum loading of HC1- pct AlC13 and initial C1:Al of 1.8. clay process liquors is limited to sat- With test durations from 0.5 to 16 h, urated AlC13, the higher aluminum concen- aluminum extraction was 83 pct in 1 h and tration possible with basic chloride did not increase by more than the esti- solutions has the important potential mated experimental error of 3 pct in benefit of decreasing the size of the 16 h. processing equipment.

The compositions of the solids produced COUNTERCURRENT LEACHING by evaporation and cooling of single- AND CRYSTALLIZATION stage pregnant liquors were poorly repro- ducible, with C1:Al ratios ranging from Countercurrent leaching was studied as 3.0 to 1.4 and correlated approximately a means of avoiding the trade-off that with the liquor ratio. No solid products was necessary in single-stage leaching were obtained from tests with initial between high alumina extraction and low ratios below 1.4 because the pulp density C1:Al ratios. The countercurrent method problem was compounded by increasingly should also avoid high pulp densities viscous pregnant liquors, which tended to and the associated handling problems. supersaturate. A three-stage leaching operation was chosen. The best conditions for single-stage leaching are shown in table 2. Using a The general countercurrent leaching 30-pct AlC13 leachant and an initial scheme is shown in figure 3. Calcined C1:Al ratio of 2.2 gave the best trade- clay enters the array from the left and off between high aluminum extraction and moves to the right. The leachant moves low liquor C1:Al ratio. The liquors from top to bottom. Nine batch tests are were reasonably easy to filter and gave required to generate the intermediate basic chloride crystals on evaporation. compositions needed to duplicate a

TABLE 2. - Summary of single-stage 30-pct Leachant Leachant Leachant AlC13 leaching Calcined clay

"NO crystals obtained. Preanant Discard Discard An important result of single-stage liquor leaching is the concentration of aluminum FIGURE 3. - Test array for three-stage counter- in the pregnant liquor. The 30-pct AlC13 current leaching. continuous three-stage countercurrent filtration was difficult. Concentrations leaching operation. The liquid stream of of aluminum in pregnant liquors ranged importance enters as leachant for step 3A from 4.2 to 5.4 pct. and exits from step 1C. This liquor has been contacted three times with increas- TABLE 3. - Summary of countercurrent ingly alumina-rich ore and represents leaching results the pregnant liquor from countercurrent leaching. The important solid stream C1:Al ratio Al Liquor enters at 1C and exits at 3C, where it ~nitial(~iquorlcrystalextraction, Al, represents the final leach residue.

The composition and amount of leachant and the amount of calcined clay were chosen so that the overall C1:Al ratio No crystals obtained. would be below 1.5, an area that was very troublesome in the single-stage tests. The solids obtained on evaporation and To provide sufficient water for the reac- cooling of liquors with C1:Al ratios of tion while maintaining the chloride con- 1.8 and 1.5 were crystalline and fine centration, the leachant of choice was and had a C1:Al ratio of 1.15. An X-ray not an AlC13 solution, but HC1. With the diffraction pattern was obtained and proper balance of leachant to clay, all matched the pattern given by Breuil (8) of the HC1 would be consumed in the first for a compound which she identified as leaching stage and produce an AlC13 5A1C13*8A1(OH)3*37.5H20. No solid could leachant for the second stage. All of be obtained from the liquor with a the aluminum in the final pregnant liquor C1:Al ratio of 1.30. Evaporation only was derived from the clay. increased its viscosity and gave a syrup reminiscent of a supersaturated The leachant chosen was 110 g (103 mL) organic solution or a concentrated col- of 15-pct HC1, and three test arrays were loidal suspension. run with calcined clay charges of 40, 50, and 60 g, which corresponded to initial Although Breuills preparation of the C1:Al ratios of approximately 1.4, 1.1, above compound, which is herein desig- and 0.9. Each leaching test was run for nated ACIIH, involved aging for a long 2 h, and the slurry was filtered without time, research by the authors determined washing. The residue was passed to that it can be prepared reproducibly from the next test to the right and the fil- basic aluminum chloride solutions with trate to the next lower test. Only the C1:Al ratios in the range of 2.0 to 1.4 final residue from test 3C was washed by evaporation to incipient crystalliza- before analysis to determine aluminum tion and cooling. From solutions at the extraction. upper end of the range, the solid ob- tained is a mixture of ACH and ACHH. At Table 3 summarizes the results of coun- the low end, syrup formation is a prob- tercurrent leaching tests. The 93.4-pct lem. A solution with a ratio of 1.8 is aluminum extraction for the test with an near ideal for single-phase ACHH. The initial C1:Al ratio of 1.41 is signifi- pure ACHH prepared for thermal decomposi- cantly better than was possible in tion studies was crystallized from a so- single-stage leaching and is comparable lution with a C1:Al ratio of 1.6 and was to the 95 pct achieved in the HC1-clay made by dissolving aluminum shot in HC1. process. Liquor C1:Al ratios for all three tests were about 0.4 unit higher PURIFICATION AND SOLUBILITY STUDIES than the initial ratio. Few prob- lems were encountered in filtering liq- The levels of major impurities in uors from leach residues except for the the ACHH obtained from countercurrent test with an initial C1:Al ratio of leaching of clay with a liquor C1:Al 0.94. That liquor was very viscous, and ratio of 1.5 are given in table 4. Concentrations are expressed as percent the ACHH crystal lattice. Decreasing the of impurity oxide in alumina derived from amount of ferric iron substitution should the ACHH. Specifications for cell-grade also allow more perfect crystal growth alumina are given for comparison.. These and decreased inclusion of other impuri- specifications were developed by a sub- ties. The level of iron in ACHH was committee of the Industry/Bureau of Mines decreased by two-thirds with this method, Steering Committee for the Alumina Mini- and minor decreases of other impurities plant Program (6). Considerable improve- were noted. Problems, including air ment is necessary if a cell-grade product leaks into the equipment, prevented main- is to be produced from ACHH. A cursory taining more than 90 to 95 pct of the examination was made of three possible dissolved iron in the ferrous state dur- methods for improving the purity of the ing crystallization. ACHH . The greater concentration of aluminum TABLE 4. - Impurity concentrations in in the basic chloride liquors compared cell-grade alumina and ACHH, percent with that in HC1-clay liquor is an im- portant processing advantage. The leach- Impurity Specifications ing tests were not optimized for high Na20...... 0.40 liquor concentration. To determine the CaO...... maximum concentration of aluminum achiev- Fe203...... a able, a series of synthetic liquors with K20...... C1:Al ratios ranging from2.4 to 1.2 MgO...... were prepared, evaporated until crystal- lization began, and stored in tightly capped bottles at room temperature for 4 derived from the ACHH. to 6 wk, after which the solids and supernatant liquids were analyzed. The The crystal washing procedure with pro- solid phase in contact with the higher pan01 was adopted for experimental expe- ratio liquors was a mixture of ACHH and dience. Purification by redissolving the ACH; with the lower ratio liquors, propanol-washed ACHH in water, recrystal- another basic chloride, with a diffrac- lizing, and washing a second time with tion pattern identified by Breuil as propanol was attempted. Some purifica- AlC13*4A1(OH)3*7.5H20, was formed. In tion was noted, but ACHH was signif i- the middle-range liquor ratios of 2.0 to cantly soluble in propanol. Countercur- 1.5, the solids were single-phase ACHH, rent washing with an alcohol saturated and the aluminum concentration of the with ACHH may be possible but was not supernatant was 9.3 to 10.9 pct. investigated. THERMAL DECOMPOSITION OF ACHH Attempts were made to remove the iron from pregnant liquors by solvent extrac- A sample of pure ACHH prepared as pre- tion with amine extractants, but were viously described was submitted to the unsuccessful because the liquors did not Thermodynamics Group of the Bureau's contain sufficient free chloride to form Albany Research Center. Comparative the chlorocomplexes of iron that the determinations were made by scanning dif- amines could extract. ferential calorimetry of the energy re- quirements for decomposition of both ACHH A third purification approach was par- and ACH to gamma alumina. The results tially successful. A pregnant liquor was showed that ACHH requires only subjected to electrolytic reduction to 61.6 pct as much energy for decomposition reduce the contained iron to the ferrous as ACH. This a very significant decrease state. The liquor was then evaporated because approximately one-half of the and crystallized. The reason for reduc- total energy requirement of the HC1-clay ing the iron to the ferrous state was to process is consumed in the decomposition retard iron substitution for aluminum in step. DISCUSSION AND CONCLUSIONS

The following major conclusions have 4. The basic chloride liquor from resulted from this study: countercurrent leaching can be crystal- lized to give the compound VllCl, 1. About 90 pct of the alumina content -8A1(OH),.37.5H2O. of calcined kaolin clay can be leached with a solution of AlCl, or substoichio- 5. The energy needed to decompose ACHH metric HC1 to give a basic chloride to alumina is about 62 pet of that needed liquor. for the ACH produced in the HC1-clay process. 2. Calcination temperature is criti- cal. About 850' C is the best for clay This research should be considered a to be leached by the basic chloride preliminary effort toward the possible method. This is in contrast to stan- development of a new process for extract- dard HC1 leaching, where a broad range ing aluminum from clays. Much more re- of calcination temperatures may be search is required before a preliminary used. flowsheet could be proposed. While coun- tercurrent leaching is probably the best 3. The potential aluminum concen- approach for this system, detailed opti- tration in basic chloride pregnant mization studies are needed to balance liquor is about 75 pct higher than is extraction with liquor aluminm concen- possible in liquor from standard HC1 tration, C1:Al ratio, and ease of fil- leaching. tration and crystallization.

REFERENCES

1. Husted, J. E, Potential Reserves Ann. Meeting, New Orleans, LA, Feb. of Domestic Non-Bauxite Sources of Alumi- 2-3, 1979). Light Metals 1979, v. 1, num. Pres. at AIM3 Meeting, Dallas, TX, pp. 217-282. Feb. 25-28, 1974. TMS paper A74-65, 20 PP* 6. Kaiser Engineers, Inc. Alumina Process Feasibility Study and Preliminary 2. Tilley, G. S., R. W. Millar, and Pilot Plant Design. Task 3 Report: Pre- 0. C. Ralston. Acid Processes for the liminary Design of 25-Ton-Per-Day Pilot Extraction of Aludna. BuMines B 267, Plant. Volume 1: Process Technology and 1927, 88 pp. Costs (contract 50265048). Banes OFR 122-80(1), 1979, 152 pp.; NTIS PB 81- 3. Archibald, .F. R., and C. M. Nichol- 125031. son. Alumina From Clay by the Lime- Sinter Method 11. Metals Technol., AIM 7. Mellor, J. W. A Comprehensive Tech. Pub. 2390, June 1948, 25 pp. Treatise on Inorganic and Theoretical Chemistry. Longmans, Green, v. 5, 1940, 4. Hoffman, J. I., R. T. Leslie, H. 3. pp. 280-318. Caul, L. J. Clark, and J. D. Hoffman. Development of a Hydrochloric Acid Pro- 8. Breuil, H, Sur les Ghlorures cess for the Production of Aluminum From et Bromures Basiques dfAluminium (On Clay. J, Res. NBS, v. 37, Dec. 1946, the Basic Chlorides and Bromides of pp. 409-428. Aluminum). Ann. Chim., v. 10, 1965, pp. 467-493. 5. Bengtson, K. B. A Technological Comparison of Six Processes for the 9. Grim, R. E. Clay Mineralogy, Production of Reduction-Grade Alumina McGraw-Hill, 1953, pp. 190-216. From Non-Bauxitic Raw Materials (AIME