Search Images Maps Play YouTube News Gmail Drive More »
Sign in
Screen reader users: click this link for accessible mode. Accessible mode has the same essential features but works better with your reader.


  1. Advanced Patent Search
Publication numberUS3793003 A
Publication typeGrant
Publication dateFeb 19, 1974
Filing dateJan 4, 1971
Priority dateJan 4, 1971
Publication numberUS 3793003 A, US 3793003A, US-A-3793003, US3793003 A, US3793003A
InventorsOthmer D
Original AssigneeOthmer D
Export CitationBiBTeX, EndNote, RefMan
External Links: USPTO, USPTO Assignment, Espacenet
Method for producing aluminum metal directly from ore
US 3793003 A
Aluminum may be separated from iron oxide containing ores hitherto unusable through conventional processing by conversion to a volatile tri-halide. This, at 1,000 DEG -1,200 DEG C converts, with additional aluminum from the ore, to three moles of mono-halide, which, on cooling below 1,000 DEG C, reverts to one mole of tri-halide and two atoms of pure metallic aluminum. The aluminum trihalide, without cooling, unites with the iron oxide in imcoming ore to give volatile ferric chloride, which is removed from the system, and aluminum oxide which continues with the balance of the ore to form the subhalide. Different processes with other selected steps produce aluminum from clays, feldspars, oil shales, red muds from current aluminum production wastes, slimes from phosphate manufacture, etc., while producing economically and almost quantitatively the other metal values.
Previous page
Next page
Claims  available in
Description  (OCR text may contain errors)

[ Feb. 19, 1974 Othmer METHOD FOR PRODUCING ALUMINUM METAL DIRECTLY FROM ORE [76] lnventor: Donald F. Othmer, 333 Jay St.,

Brooklyn, NY. 11201 [22] Filed: Jan. 4, 1971 21 Appl. No.: 103,765

[52] US. Cl 75/68, 423/135, 423/136, 423/149, 423/76, 423/79, 423/343 [51] Int. Cl C22b 21/00 [58] Field of Search 76/68 B, 112, 113; 23/92, 95, 23/96; 423/135, 136, 149

FOREIGN PATENTS OR APPLICATIONS 1,019,261 l/1953 France 75/68 B Great Britain 75/68 B Switzerland 75/68 B Primary Examiner-Herbert T. Carter [5 7 ABSTRACT Aluminum may be separated from iron oxide containing ores hitherto unusable through conventional processing by conversion to a volatile tri-halide. This, at 1,000-1,200C converts, with additional aluminum from the ore, to three moles of mono-halide, which, on cooling below 1,000C, reverts to one mole of trihalide and two atoms of pure metallic aluminum. The aluminum trihalide, without cooling, unites with the iron oxide in imcoming ore to give volatile ferric chloride, which is removed from the system, and aluminum oxide which continues with the balance of the ore to form the subhalide. Difi'erent processes with other selected steps produce aluminum from clays, feldspars, oil shales, red muds from current aluminum production wastes, slirnes from phosphate manufacture, etc., while producing economically and almost quantitatively the other metal values.

10 Claims, 5 Drawing Figures METHOD FOR PRODUCING ALUMINUM METAL DIRECTLY FROM ORE This invention allows the winning of pure aluminum directly from various alumina-containing ores, including bauxites, clays, shales, slates, feldspars, and others which, because of their large content of other oxides and the chemical nature of the compounds of these oxides with each other and with alumina, are unsuitable for production of aluminum by conventional methods. Halogens, and particularly chlorine and bromine, are used for the attack on the ore. The corresponding volatile halides which are formed of aluminum and other metals are separted to produce substantially pure aluminum metal therefrom, often in a single thermal operation involving disproportionation, and other pure metallic compounds or metals by other steps.

Hereinafter, the more common chlorine, chlorination, and chlorides are usually referred to with the understanding that bromine, bromination, and bromides may often be considered as equivalent for every step of the processes concerned, and with some respective advantages and disadvantages in different cases.

CHEMICAL REACTIONS USED IN RELATED PRIOR ART When an aluminum-containing ore is treated with the sequence of the dual oxidation-chlorination and then reduction-chlorination, according to the HALOMET process, those metals, including particularly iron and titanium, which have a greater affinity for chlorine from the oxide than does aluminum, may be volatilized as the chlorides and removed first. While iron comes as the ferric chloride, and in the gas phase it probably has the formula Fe Cl a dimer; for the present purpose, this is of no moment and it will be called simply FeCl Along with the volatile chlorides of these metals will come sulfur present in the ore or in the reductant as .2 qgsthsrawi hph sp an v aasii masllls oxychlorides. Under these conditions, all of these elements form volatile compounds. The aluminum may next be chlorinated to give the normal chloride, AlCl which is also volatilized. Thus, after the sulfur, phosphorus, vanadium, iron, and titanium are separated from theore and, if desired, from each other, by one or more dual chlorinations and other steps of the Halomet process described in U.S. Pat. Nos. 3,244,509 and 3,466,169, aluminum chloride passes off as a practically pure gas, with evolution .of heat:

A120 3C 3Cl 2AlCl (gas) 3C0 Heat (I) These steps resemble the usual fractional distillation of compounds of greater volatility (lower boiling point) away from those of lesser volatility in a succession of steps. Here, instead of volatility alone, there is used the additional property of the affinity of a metal to form the chloride from its oxide.

The aluminum atom is normally trivalent, but changes to mono-valency or disproportionation at temperatures higher than about 1,000C; and at these temperatures, MCI is the stable chloride. Thus, when AlCl gas is heated to temperatures of 1,100l ,200C, twice as much more aluminum may be reacted with the three chlorine atoms of the molecule to give the mono chloride.

AlCl (gas) 2Al(solid) HeatT BAlCKgas) 2 frorn t he oxide A1203 If heat is supplied, this reversible reaction is known to go from left to right, above about 1,000C, and is reversed to go from right to left below about 1,000C, with heat being given off. Hence, by cooling the AlCl vapor below 1,000C, the aluminum again assumes its usual tri-valency; and the reaction liberates pure aluminum metal as a liquid above its melting point of 660C, or as a solid if below this temperature. By operating the condenser at about 700C, the pure liquid aluminum formed may flow into molds and solidify as ingots.

The usual aluminum chloride, AlCl resulting here after loss of aluminum, has again been reacted in the prior art with additional metallic aluminum, usually by contacting an aluminum-iron alloy made into pellets. The process is repeated, with two atoms of chlorine acting as the carrier to chlorinate two atoms of aluminum from a source of the metal at the higher temperature; and the reverse from mono-valent to tri-valent aluminum at the lower temperature accomplishes the separation at the lower temperature of two-thirds of the aluminum present in the three molecules of AlCl.

While these reactions have long been known in the art, there have been major problems in commercializing any process using them with the pellets of aluminum-iron alloy, which must, in any case, be prepared from ores by prior extensive and expensive processes.

CHEMICAL REACTIONS USED IN THE PRESENT INVENTION At the high temperature of l,000-l ,200C it has now been found that at least three gaseous chlorinating agents may be used in the reduction-chlorination of A1 0,, togive the mono;chlori deAlCl directly in the attack on the ore itself. Sufficient carbon (as coke) may be added not only to accomplish the reducing action, but also, if necessary, to burn with added gaseous oxygen to give whatever heat may be necessary to maintain this high temperature.

The chlorinating agents under these conditions are both atoms in the molecule of chlorine, C1 itself; or all of the chlorine atoms present in some other gaseous metallic chloride,as SiCl, of an element lower in the affinity series; or two of the chlorine atoms present in AlCl Any of these can be used to produce AlCl by reaction with aluminum, taken not from a source of metallic aluminum as in Equation 2, but instead directly through this reductionchlorination. If AlCl is used as the gaseous chlorinating agent, three molecules of the mono-chloride, AlCl, result. Thus:

012 A1203 3c 2AlCl 300 (3a) SiCl.,+2Al O +4C 4AlCl+4CO+SiO2 (3b) A1913. c Qafr. 3C: AlQLtlQQ (3c) The above reactions proceed using coke sufficient for the reduction-chlorination and also to burn with oxygen to supply additional heat, e.g., C+O CO2 b. the relative amounts of C1 SiCl and AlCl which are used in the reduction-chlorination since there are different heats involved with the different chlorinating agents in the respective reactions 3a, b, and c;

c. the type of furnace used and therefore the amount of heat lost to the surroundings.

Thus, the two aluminum atoms necessary for the reaction of Equation 2 can be taken not only from metallic aluminum as in Equation 2, but also from aluminum oxide in the presence of carbon with the formation of CO. The present invention thus uses Equation 3c principally, but also 3b, and in other cases, 3a, to attack the ore, and then Equation 2 (right to left) to produce pure aluminum directly from the alumina in the ore.

The careful addition of chlorine and of oxygen, preferably as such although air may sometimes be used, so that the correct stoichiometric amounts are not exceeded, is the most essential control of the invention. Neither of these reactive gases can be present when metallic aluminum is obtained in cooling its monochloride down from about l,000C to some temperature down to 700C, or the metal would be instantly oxidized or chloridized. Hence, in the operation of the reactions represented by Equations 3a, b, and 0, there must always be present an excess of alumina.

The oxygen to combust the coke allowed for the supply of heat in the mixture of ore and reductant to maintain the high temperature of 1,000 to l,200C, will be added to this reaction zone. This may be done in admixture with the chlorine, and through the chlorine feedline or, as suggested hereinafter, it may be through a separate inlet. In either case, the control is based on the temperature of the reaction zone wherein Equation 3a, b, or c takes place; if the temperature falls, more oxygen is added; if it rises too high, the supply of oxygen is throttled.

In use of coke for heat and in the chlorination, more or less CO is formed; but the corresponding equations are not written here for this obvious modification.

The gaseous AlCl which is uncondensed in the condenser for the metallic aluminum, may then be passed to contact the raw ore feed in an oxidationchlorination, to remove the iron present as FeCl in a first st p whiL .si.\ n. qa!9rni u@as h qALzQs Cl l eflFfiiqsjl isQlair A lQaH- BLL The A1 then returns with the solids of tga fgglto the reduction chlorin ation, where it ui'ldefgoes again the reaction of Equation 1; the AlCl formed undergoes the reaction of Equation 3c; and the AlCl so formed undergoes the reaction of Equation 2 (right to left) to give more aluminum. A cycle is thus formed of the reactions of AlCl There may be more iron in the ore than can be removed according to Equation 4 by the amount of AlCl available, h!!! t e asi of then m er of. om of aluminum metal removed in Equation 2 (reverse) is l to 2. Thus, if more than one-half as many atoms of iron are in the ore as atoms of aluminum, the AlCl of Equation 4 does not supply sufficient chlorine to make FeCL, of all of the Fe O present. Additional chlorine as C1 may be supplied then to remove all iron as C w The attack of almost any aluminum ore by chlorine is such that the aluminum may be removed almost quantitatively at temperatures of 750 to 900C, or as AlCl at temperatures from l,000C to l,200C. These ores include various clays, shales, and other minerals widely distributed and of little present value, as well as the bauxites normally used for production of aluminum. In these ores, entirely non-usable by the conventional process, the presence of Si0 and particularly its association with AI Q malges difi 'icult and expensive the normal attack with caustic soda and the accompanying loss of this aqueous alkaline reagent. Using chlorine, no reagent is tied up in residues containing aluminum or silicon, as in conventional processing.

The preliminary handling of the ore and the coke to be used requires their preparation to give a range of particle size preferably of to 200a although larger sizes may also be used up to 3 mm, with corresponding slowing down of the reactions. These pulverent or particulate masses must be thoroughly dried at a temperature high enough to drive off all free and as much as possible of any combined water. Water added to the system unites with the chlorine present, to give hydrogen chloride which may be a waste of chlorine and a nuisance in the operation. Intimate mixing of the ore and coke is necessary in their pulverent or particulate forms. Either term here includes this range of sizes.

The chemical reactions (except that for aluminum metal from AlCl) are accomplished by the contacting of gases with pulverent solids; and each reaction has been found to go very fast because of the small size of the solid particles and hence the large ratio of surface for reaction to mass to be reacted. Furthermore, these reactions are chlorinations by chlorine gas or by a gaseous metallic chloride of a metal whose oxide has less affinity for the chlorine; and they always result in the formation of a gaseous chloride which is removed from the reaction scene by its volatility, or of a non-volatile chloride which stays with the gangue. Higher temperatures cause reactions to go faster, thus while the reactions of Equations 3a, b, and c operate above l,000C, they go much faster at l,200, which may often be the optimum because of problems with materials of construction. However, even higher temperatures, over l,000C and up to at least 1,500C, have been found advantageous in speeding up and completing the reaction.

The invention comprises a series of chemical reactions; and these may be accomplished by either batch or continuous processing in reaction zones, which fall into a definite sequence I, II, III, IV and V, based on the order of the solids in the system contacted by gases. These are not five separate sequential reactions; but by having these five reaction zones, the individual reactions can be started in one zone and finished in another so as to strip, or eliminate entirely, one component in either the gas or solid phase before another reaction starts in a successive zone.

This sequence may be of time in the same vessel, and thus zones may actually refer to the same place for successive operations or reactions which take place in the same vessel or location. Alternatively, the sequence may be of location as the different reactions which now, in point of time, are all taking place in five different, respective places simultaneously, as in a continuous process. A still further variant is also described hereinafter, wherein the reactor vessel stays with the solids, so that each reaction of the solids does indeed take place in the same place; but at any instant, all reactor, the Successive one starts in each.

actions are taking place simultaneously in a series of reactors, wherein as one reaction is finished in each re- Cyclone contactors and multi-hearth furnaces are two of the types of contacting devices which may be used to conduct the gas-solid chemical reactions. Thus, the reaction zones may be referred to not only as a reactor, a cell, or a stage, but also as a cyclone, or a hearth, in each case with an appropriate number attached.

As noted above, the other chemical reaction of MCI disproportionating to give aluminum and AlCl in a condenser actually depends on the cooling of the gas containing the AlCl to some temperature below l,000C: herein usually 700C, to obtain liquid rather than the solidified aluminum below the freezing point, 660C. The hot gas cooled instantaneously by a cold surface to 700C does condense aluminum in the familiar physical vapor-liquid phenomena, although the chemical disproportionation probably occurs first. If I the gas is cooled before it contacts the cold surface, the disproportionation will occur, to give minute droplets of liquid aluminum suspended in the gas phase, since aluminum cannot exist as a gas below 2,057C, its boiling point. These droplets coalesce on a cold surface or may be' taken out by a suitable demister, cyclone separator, or other standard device. However, the terms condensing and condenser are used herein to include these phenomena involved in obtaining aluminum as a liquid from that formed by the disproportionation of AlCl in the gas phase.

A different situation exists in the removal by condensation of FeCl, from a gas stream. The gaseous FeCl (more exactly Fe Cl in the gas phase, but here always called FeCl may condense to a liquid at the liquids boiling point, i.e., 315C; or to a solid at the liquids freezing point, i.e., 282C, depending on the convenience of handling one or the other. Hereinafter, a temperature of condensing in the liquid region 2823 C indicates the liquid is desired, while below 282C, the solid is more convenient to handle. Either maybe obtained, depending on ease of subsequent processing.

,In the condensation of AlCl at 180.6C, only solid is produced, since this, he sublimation temperature, is at I a lower pressure and temperature than the triplepoint.

Condensers for handling the solids are widely used in the manufacture of AlCl and any conventional one may be used here.

Functions of AlCl Since the present invention involves AlCl both as a product and as a raw material according to Equations 1, 2, 3c and 4, the relative amounts of AlCl used in each reaction, and particularly in Equation 4, depends on the amount. of Fe Q and some other constituents in the ore, e.g., phosphorus and vanadium. Hence, the controls of flow of ore, reductant, and chlorine, as well as of temperatures in the reactions, fix the amount of Al O3 rcgycled baclg frornthe highest chlorination step acting with l e- 0 to vaporize the iron as FeCl while giving A1 0 according to Equation 4.'Thus, if the ore qqntain omelet! as marme ss g uFsz a f 29. the AlCl coming from the condenser for the aluminum, after leaving twice as much aluminum metal there (equivalent to all that enters in the ore) could all be used in the reaction of Equation 4 to remove the iron from the system as FeCl (However, this AlCl used to form FeCl which volatilizes and A1 0 which remains in the syst m. i 19tat ssagtshlqrinehyitselt.),

If more than half as many atoms of iron as of aluminum are in the ore, then additional AlCl may be recycled as a gas to displace and remove the iron or additional chlorine as C1 may be added. lfthe ore contains less than half as many atoms of iron as of aluminum, the heat in the vapors may be recovered by condensing AlCl (gas) physically to preheat the ore charged insofar as possible. Or the AlCl may be condensed with loss of heat and returned to the feed solids. These and other examples of working with ores of different concentration of different oxides will appear more fully hfl fi r- However, it is notable that, since the aluminum leaves the system as the metal, there is no net requirements of chlorine. for its separation; the AlCl simply recycles as a carrier again and again for aluminum from the ore to aluminum as the metal.

FRACTIONAL SEPARATIONS As indicated above and as will be more apparent later from the description of several embodiments of this invention, the fractional separation of the metals of an ore using differences of their relative affinities to form chlorides from the oxides may be operated comparable to the fractional separation or rectification by distillation of materials when using differences of their relative volatilities. In such distillation and rectification, a concentration of that material of greater volatility results (in this case there results a concentration of those of greater chloride-forming affinity) at one end of the rectification system, as they are stripped out of those, in the distillation case, with lower relative volatility here, lower chloride-forming affinity.

The oxidation-chlorination as the head of the system, i.e., where the element of the greatest affinity is to be removed, acts as a heads stripper to hold back all of the volatile chlorides save that of the metal highest in the affinity series. Thus the volatile chlorides of the metals lower in the series chlorinate the oxide of the one highest in the series which is here in the batch of input ore acting as a reflux on the system. The oxidation-chlorination is less violent than the reductionchlorination and thus acts as a more precise stripping or separating mechanism. Of course, if two or three, or indeed all of the metals present in the system as impurities, were above in the affinity system, the important metal desired might possibly be removed without regard to their separation in a single vapor stream, leaving the desired metal behind for other consideration. The gas mixture of the chlorides of the metals high in the affinity series could then be worked up separately.

In distillation, it is theoretically impossible to achieve a perfect separation of two components; here, because of the preferential chemical affinity, it is theoretically possible to achieve a perfect separation. However, be cause of the large masses of solids to be reacted counter-currently with a gas, it is practically impossible to achieve absolutely the same contacting throughout the entire mass at the same instance.

In many distillation systems, chemical reactions are also effective in the separation by the distillation columns, and the selection of the proper plate or stage on which to accomplish this reaction is based on the separations accomplished up to and from that plate through differences in relative volatilities. Here, in the halide separations, there is also a comparable, if reverse, phenomenon which may be utilized in the chain of chemical reactions which are producing the separations due to relative affinities, and the accompanying concentrations and strippings.

Thus, it is possible, in many cases, to interrupt the stream of the affinity-fractionation and to take out streams of two or more materials, usually chlorides or oxychlorides, which have been formed, and to separate these streams by their relative volatilities; i.e., either by fractional distillation or by fractional condensation. Then the separated streams are either removed to be worked up to give the pure metal or reintroduced into the affinity-fractionation.

Also, differences in other physical properties may be utilized at some stage of the process of separation by affinities, often at either end of the process; as also happens comparably in a distillation system with either the overhead stream or the bottom stream. For example, here some of the very non-volatile chlorides (of alkali or alkali-earth metals) remain in the final gangue and may readily be removed by dissolution with water, and then separated by various techniques of wet chemistry. Also, some chlorides and oxychlorides may be removed from the overhead stream here, along with HCl and other gases, all of which are readily separated by standard processing.

Furthermore, there may be used other chemical reactions, e.g., as indicated above, for obtaining aluminum from AlCl by changing the valence of the aluminum. Such chemical or physical operations to give separations may be used in a process step which interrupts the fractionation by chloride-forming affinities, just as chemical or physical operations may be used to interrupt the separation of other materials by fractionation through distillation.

The present process, through the application of these principles, profitably separates aluminum from almost any aluminum-containing material, including those -coming as wastes from other mining operations and from many natural ores which are so widespread and plentiful as to be valueless. However, it is highly desirable that both the free and combined water be completely driven off prior to chlorination, by ignition if necessary, before chlorine reacts therewith to form l-lCl. Such hydrochloric acid as is formed may be condensed and removed with the first volatile chloride or oxychloride which is formed. It may often be used in a preliminary wet attack on the ores to chlorinate partially some of the metals present before the ore is dried to discharge water. This may have the disadvantage, however, in giving chlorides which are volatile in the high temperature drying to eliminate combined water; and it may often be better to convert electrolytically or otherwise the HCl to C1 which may then be reused immediately.

The stream of FeCl;, which may result in the operations described may be oxidized to Fe O while recovering for reuse the C1 given off; and this Fe o may then be reduced, e.g., by hydrogen to give pure iron. However, the immediate reduction by hydrogen of FeCl;, to give pure iron and hydrogen chloride may be simpler in those cases where there is a market or use for the acid.

Another system of reusing the gaseous l-ICl in the offgases, which has been formed mainly by the reaction of chlorine with hydrogen coming from the reductant, coke, or with water in the ore or coke, is to treat pulverized and dried raw ores with the effluent gas containing HCl so as to form some chlorides from this acid while preheating the ore with the hot gases, rather than accomplishing all of the chlorination from gasous chlorine or the chlorides of metals lower in the affinity series.

-Also, the process of this invention may use bromine instead of chlorine under substantially the same reaction temperatures, although the boiling points and melting points of the bromides are usually somewhat different from those of the chlorides and the separations may be altered advantageously through the use of one or the other halide.

OBJECTS AND ACCOMPLISHMENTS OF THE INVENTION The present invention thus accomplishes:

a. the economical winning of aluminum through the intermediary production of its halides, often but not necessarily its chlorides, as substantially pure metal directly from shales, clays, bauxites, feldspars, and other aluminum-containing ores, which may contain other elements or the aluminum itself in compounds which prevent its economical separation by conventional processes;

b. the profitable disposal and use, principally for phosphorus and aluminum values, of waste slimes from the beneficiation of Florida phosphate rock;

0. the use of oil shales for the winning of aluminum and other values therefrom, using as the reductant the carbonaceous oil therein or the residue formed by its pyrolysis;

d. the use of red muds from aluminum processing to obtain almost quantitatively their iron and aluminum values;

e. the use of the conversion of aluminum from the mono-valent form of the mono-chloride, AlCl (stable above about 1,000C), to the triple-valent form of the usual tri-chloride, AlCl (unstable above about 1,000C), but singularly stable at or below 700C), as a means of releasing two atoms of aluminum metal from three molecules of the mono-chloride going to a single molecule of the tri-chloride; and of the reconversion of AlCl to 3AlCl by a reduction-chlorination to take the necessary aluminum from A1 0 thus achieving a cyclic process using one atom of aluminum (in AlCl as the carrier for two atoms of aluminum from the ore to the product pure aluminum;

f. the combination of this step of production of metallic aluminum in this conversion from the monochloride to the trichloride, with halogenation for the initial and practically complete elimination of sulfur, vanadium, titanium, and iron present in the ore as their volatile chlorides or oxychlorides through the recycle of A1 0 and AlCl g. the combination of the dual oxidationhalogenation and reduction-halogenation steps for the removal of silica and oxides of other elements having less affinity for chlorine than aluminum, along with non-volatile chlorides,'all as a gangue; while producing the aluminum tri-chloride to be used for removal of iron and other materials as in (f) above and the formation of aluminum monochloride with additional aluminum from the oxide as in (e) above;

h. the use of these several processes with bromine rather than chlorine as stated;

i. the combination in sequential or cyclic steps of these various processes with the several materials indicated so as to maximize the purity of each product formed while minimizing the loss of values of the ore, of reductant, or chlorine, and of heat supply;

j. the conversion of some other materials present in ores containing aluminum as mixtures of two or more chlorides or oxychlorides of different volatilities, e.g., FeCl, (B.P. 315C), TiCl, (B.P. 136C), VOCl (B.P. 127C), POCl (B.P. 105C) and the separation of these chlorides of oxychlorides to give more or less pure products through their relative volatilities, i.e., using either fractional condensation or fractional distillation;

k. the production of aluminum from ores wherein substantially all of the aluminum and iron values are recovered to leave a very small amount of disposable gangue and without pollution of air, water, or land.

FIG URES The Figures are entirely diagramatic, with no relation to dimensions, scales, or even shapes of equipment; and they are essentially flow sheets of the processes of the invention, wherein any suitable pieces of equipment may be used. Not shown are the necessary instruments for the measurement of the temperatures of different parts of the operation or of amounts of solids and gases, also of liquid aluminum flowing in the various streams, nor, indeed, the control valves, feeders, etc.

FIG. 1 is the diagram of a laboratory tube-furnace and the operation thereof to separate a mixture of the oxides of aluminum, iron, and silicon, to give: first, iron as its volatile chloride; then aluminum as the metal, and silicon remaining as the oxide.

FIG. 2 is the diagram of another arrangement and operation of a tubular furnace for accomplishing the separations as in FIG. 1, with all reactions occurring simultaneously.

FIG. 3 is the diagram of the batch-cyclic operation of six reactors at any time five are in use and one is being emptied and filled and a condenser for aluminum vapors. Solids remain in each reactor which is used in a cycle by advancing the successive steps of the gas treatment with those solids remaining from the previous steps.

FIG. 4 is the diagram of cyclone-reactor system for continuously contacting and reacting pulverent solids, a condenser for aluminum vapors, and other condenser for ferric chloride, for making these separations.

FIG. 5 is the diagram of a multiple hearth furnace, an auxiliary condenser for aluminum vapors, and another condenser for ferric chloride for making these separations.

ALUMINUM-CONTAINING ORES AND THEIR CONSTITUENTS composed principally of A1 0 Fe O and SiO These may come in many molecular complexes, by themselves, and with numerous other oxides, all of which may be attacked with the use of chlorine and coke as a combination of reagents at the temperature employed. Other common oxides, often present in very substantial amounts, are those of the alkali and alkaline earth metals, as Na, K, Ba, Mg, and Ca. The halides of these are non-volatile at the temperatures employed and would go out in the final gangue, which is principally SiO If of value, and since they are all very soluble in water, they can readily be leached out of this residue and worked up by wet techniques.

Other metals, Pb, Zn, Co, and Ni, are very much less apt to be present. Their halides are much less volatile; but the oxides, if present, have great affinity to form halides which can all be separated from the aluminum and the iron chloride by additional steps, as may also the halides of Mn and Ti, two other metals which are present more often. The separation from iron-containing ores of titanium as the chloride, sulfur as the oxide, and phosphorus and vanadium as the oxychlorides, all in a gas stream which would go with the FeCl if sufficient chlorine is supplied, has been well known in the prior art, and would be accomplished similarly here. Thus, if F e 0 and TiO are present in the ore, the Halomet process selectively and quantitatively may remove and separate from almost any complex of the metals iron as a first product, FeCl and after successive chlorination, titanium as a second product, TiCl These separation methods have been described elsewhere. However, it may sometimes be simpler, particularly if the TiO amount is small, to separate the gaseous chloride mixture containing also other materials by fractional condensation, fractional distillation, or both as later described.


The simplicity of the process of the invention may be demonstrated in the laboratory in a tubular combustion tube with a furnace near one end. Pulverent and mixed Fe O A1 0 SiO and coke are charged in the part of the tube heated by the furnace. Heat is supplied to the charge at this end, first to 700C, while chlorine is metered into at this end of the tube to react with all of the Fe O in the charge and to volatilize FeCl which passes out the other end, and is condensed at about 300C in a separate flask condenser, air cooled. This FeCl flask condenser is removed and another air-cooled flask condenser for AlCl is connected. The temperature of the charge is then increased to 1,200C; and additional chlorine is added to react with the charge of M 0 and coke according to Equation 3a. The unheated end of the tube is kept at 700C by air cooling, so that the reaction of Equation 2 (right to left) allows aluminum to condense out as liquid metal, while the AlCl formed passes out of the tube to the flask condenser and kept from contact with moist air. The gangue remaining in the tube, principally SiO is removed and discarded.

In a second run, charged and operated as before, after the chlorine removes all iron as FeCl the flask containing the AlCl from the previous run is connected to the chlorine inlet end of the tube and heated, so as to distill AlCl at about 180C and supply as a vapor to the reaction mass, now heated to l,200C. The reaction of Equation 3c proceeds to remove all aluminum from the A1 in the charge in the reductionchlorination using the coke as the reductant according to Equation 3c. Additional C1 may be supplied either at the same time or after the supply of AlCl to make up losses or strip the last of the A1 0 according to Equation 3a. The AlCl formed goes to the coolercondenser part of the tube; the aluminum metal condenses as before; and the AlCl formed with it is condensed in an external flask for use in the next run.

Sufficient chlorine need be supplied to each individual run after a steady state of such operations is reached, only to form FeCl and to supply the chemical and physical losses. It is important to note that no net amount of chlorine must be supplied to chlorinate aluminum in the alumina except for minor losses. When a steady state of operation of such a series of laboratory runs is reached, it will be seen that, since one-half as much aluminum (in AlCl is required to remove the aluminum in the ore, the amount of AlCl vaporized for the reaction to AlCl and condensed after the reaction from MCI is equivalent in aluminum to one-half of that in the ore; or one mole of AlCl is vaporized and condensed for each mole of A1 0 in the ore.

As an alternative system of operation, because of the difference of condensing points of the two volatilized chlorides, chlorine may be fed in at such a rate that both FeCl and AlCl are volatilized together during at least part of the operation; and the FeCl may first be fractionally condensed out of the gas stream at 300C, while the AlCl is condensed in the flask at a temperature below l80C for use in the next run. In either case, other gases formed and having lower condensing points are passed off for subsequent handling, as will be described, using other examples.

This example and others presumes the presence of both iron oxide and silica in the ore. If only A1 0 is present here or in the other examples, the charge would be heated to 1,200C; and any one, two, or a mixture of all three of the gaseous chlorinating agents of Equation 3a (C1 3b (SiCh), or 3c (AlCl would be used. In the case of 3b, solid SiO replaces the Al O of the charge as the aluminum is vaporized as AlCl. If the charge contains SiO as well as A1 0 the aditional solid SiO- formed remains and goes off with any SiO originally present as the gangue. If this gangue from one batch, or other particulate SiO is chlorinated with Cl the SiCl so formed could be used as the chlorinating agent of Equation 3b in the next batch.

Herein, and in the other examples, both silica and Fe O are assumed to be present, as in usual ores. When iron is not present, the step for its volatilization by a gaseous chlorinating agent is irrelevant; and both equipment and operation are simplified accordingly.

If, on the other hand, in this laboratory experiment, or in a plant operation, there are present iron, titanium, vanadium, phosphorus, and sulfur, in addition to aluminum and silicon, adequate chlorine in a reductionchlorination at any temperature above 750C would bring over FeCl TiCl VOCl, POCl and SCI or S0 and HCI. Some AlCl might come, but this could be held back by passing the gas mixture through an oxidation-chlorination of fresh ore to cause it to revert to Al- O With or without this oxidation-chlorination, the wide range of relative volatilities of these compounds allows them all to be readily separated by fractional condensation, fractional distillation, or both.

In this laboratory experiment, the temperature of l,000l ,200C necessary for the disproportionation of AlCl, to take more aluminum from the Al O is readily maintained by the external heating and usually the burning of additional coke through oxygen addition to the chlorine at this stage of the process is not required for additional heat.

Alternating Sequential Reactions In Tube A somewhat more efficient system for winning from Al O of aluminum metal and for separating it from FeCl and SiO may also be used with what is essentially a laboratory tube furnace, as is standard for combustions and other high temperature tests and analyses. As is the previous example, this reactor-condenser tube is made of a material suitable for operation up to 1,200C.

In FIG. 1, there is indicated such a tube, 1, of about 2 cm. inside diameter, which is relatively long, -160 cm., and symetrically fitted from an axial midpoint, since it is to have internal gas flow going either from left to right, or from right to left. The tube, 1, may be made of magnesium spinel, which is satisfactory for temperatures up to 1,200C under the conditions and materials of operation.

The tube has five sections, each 25 to 40 cm. long. The central section III of the furnace must withstand the highest temperature. Adjacent side-sections 3 and 5, and 4 and 6, do not have to withstand such temperatures, and instead may be of silica tubing attached by gas-type joints to the spinel center section. The central section III is heated by an exterior furnace, 7, capable of raising the temperature to 1,200C. Inlets l0 and 11 are provided for use as alternate chlorine inlets with Cl; flow, under carefully metered conditions. Outlets l2 and 13 are provided as alternate exits for gases; and, as with 10 and 11, one or the other may be closed when not in use.

Provision is made so that the reaction zone III may be charged with pulverent materials and discharged of gangue after an experimental run, without disturbing any material which may be in the side sections, particularly 4 or 6. Suitable porous dividers 15 retain the solid charge in section 1. Sections 6 and 4 are provided with removable heating units, 8 and 9, respectively. When either 8 or 9 is removed, the surface of the tube 1 in the section 6 or 4 may be cooled by air streams or otherwise, to condense vapors. No heating elements are provided for sections 3 and 5. Appropriate cooling, as by cooling air, may be administered to remove heat from the walls of these sections, as well as the walls of 6 and 4.

Drain connections for liquid metallic aluminum are provided in sections 5 and 3 by tubulatures l6 and 17, respectively, or the aluminum may be allowed to collect simply as globules in 3 or 5 during a run, to be removed at the same time as the gangue is removed.

For this operation, a synthetic mixture of dried, pulverent coke and ore containing Fe O Al O and Si O is charged into the reaction zone III of the tube and heated to about 750-900C, while controlling carefully the amount of chlorine entering at 10 to pass through the pulverent mass in 2.

The section 6 contains AICI; remaining there from its condensation during the previous run; and the section 5 is empty, as it was drained of aluminum metal as fast as it had condensed during the previous run, or during the shutdown for recharging III. A minor amount of heat is applied by 8 so that the chlorine, in passing through 6, is preheated slightly to a temperature of l-150C without undue volatilization of the AlCl Chlorine is supplied thus to chlorinate the Fe O in the charge in Ill; and FeCl:, fonns first due to the greater affinity of iron from its oxide to form the chloride, and this FeCl, passes off through the exhaust 13 to be separately condensed at a temperature of 300C. Any HCl and other gases leaving 13 would also be processed after condensing out FeCl if desired, including any TiCl S0 POCl and VOC] which might come off if there were present the corresponding elements Ti, S, P,

or V in the ore.

When the FeCl is removed from the ore in the reactor III, the feed of C1 is throttled back and then stopped completely.

The amount of chlorine metered into this system, in which the gas flow is reversed in direction after the first run, with a charge of Fe O A1 0 and SiO as major constituents and some TiO P 0 and V 0 as minor constituents would be that required to chlorinate only the Fe O plus that for losses, including the production of HCl, TiCl VOCl, and POCl in other words, the materials leaving the system as gases through 13 or 12. No chlorine would go out with the aluminum metal and all chlorination of A1 0 would be accomplished (except for losses and inefficiencies) by the AlCl remaining from the decomposition of the AlCl of the previous run; This AlCl has condensed out and remains in 6 or 4 ready to be volatilized again and used as the chlorinating agent and gaseous carrier for aluminum out of N 0,, in the reactor Ill, to give aluminum metal in its condenser 5 or 3. No chlorine as such is required to chlorinate the alumina in winning the aluminum, chlorine used is always recycled as AlCl Additional heat is supplied by 7, to brihgififinifierature of III up to l,200C, also more heat is supplied by 8 to volatilize AlCl at 180C. The AlCl gas and possibly a small amount of free C1 coming in at 10 to make up for losses are passed through the charge in the reactor "I.

At the high temperature of l,200C in III, the reaction of A1 0 and coke with AlCl proceeds by Equation 30, and with C1 as by 3a. The MC] formed in both reactons is passed into 3, which is maintained at a temperature to condense aluminum as the liquid without allowing it to solidify. The temperature in 3 is therefore maintained by air cooling or otherwise at about 700C and the liquid aluminum is drained through 17. The furnace 9 is removed; and the temperature in 4 would be reduced and maintained by air cooling or otherwise so that AlCl could condense here at its normal condensing temperature of 180C.

This operation is conducted until all of the A1 0 in the charge has been reacted with AlCl or C1 and driven out of reactor III to give aluminum metal in 3 and drained at 17; and AlCl in 4, with the exhaust of l3 opened. Then the system is disconnected; the

gangue residue, principally SiO is discharged from the reactor Ill, also aluminum metal from 3 if not drained continuously as it is formed.

The process is repeated, with gas flows from right to left instead of left to right. For the second run, the identical procedure is followed, with the supply of chlorine coming through 11 to the reactor Ill, first to volatilize and remove the iron as FeCl which passes out at 12.

When this reaction is complete, the chlorine flow is stopped, or nearly stopped; the MC, in 4 remaining from the previous run is volatilized to pass to the reactor Ill, now heated to l,200C; and the MCI formed in III is decomposed to give aluminum which condenses out, preferably as a liquid in 5, and drawn off at 16, while AlCl condenses in 6. The cycle is then repeated.

As noted above, in an ore containing Fe O A1 0,, and SiO, with one or more of TiO P 0 S and V 0 the respective elements are usually present in minor amounts and can be preferentially chlorinated to give respectively TiCl POCl SCl (or S0 and VOCl. All of these are formed more readily than AlCl due to the affinity series relationships. The chlorides or oxychlorides and S0 will go off with the FeCl and HCl for separation by fractional condensation, distillation with rectification, or both. Thus, the FeCl (B.P. 315C and MP. 282C) would condense out completely as a solid in a first condenser at 275C; any small amount of AlCl (B.P. 180C) may condense out as a liquid at a second condenser at 150C, or it may be condensed with the TiCl POCl and VOCl in a water-cooled condenser at 20C. The HCl and S0 would be in the stream of CO and CO as exit gas and would be scrubbed therefrom. The TiCl POCl and VOCl (also AlCl are fractionally distilled for separation if the quantity of each warrants, or one or more may be separated by known wet chemical methods.

lf, however, the ore contains, in addition to M 0 and SiO,, only Fe O and TiO the effluent gases in a first fractional condenser at l200C would separate out FeCl substantially pure and in a second fractional condenser at ll5-l30C would separate out TiCl, substantially pure, while a final condenser at 20C arid an absorber would remove the 119,

Here again, as in the previous example, ample heat is supplied externally by 7 to the reaction zone, identified as Ill when AlCl is being produced, and no coke or oxygen for its combustion need be supplied to maintain a temperature of l,200C.

Simultaneous Sequential Reactions in Tube FIG. 2 diagrams the operation of the reactions necessary for the separation of aluminum metal from an ore containing A1 0 Fe O and SiO,.. A reaction tube 1 has that part which serves for the left three zones made of some material such as magnesium spinel (MgO-Al- 0 This withstands quite satisfactorily the corrosive effects of any of the materials involved at temperatures up to l,200C.

Normally, the exothermic heat of the several reactions will keep the reactors at its desired temperature; but the small size of a single tube experiment requires additional heat to balance heat losses. This is supplied by tube-furnaces to heat the cells. The several cells are therefore each enclosed in tube furnaces 7.

The condenser for aluminum metal 3, between Cells ll and III is provided with a means of cooling, by draft of air or otherwise (means for providing not indicated) to maintain a temperature. of about 700C, which is considerably lower than that of either adjacent cell. Because of the greater thermal stresses imposed due to the cooling of this part of the tube, it may be desired to make it and the additional tubing to the right thereof of a length of silica tube. This tube length representing cells I and II, the condenser for aluminum and the condenser for FeCl wherein is maintained a temperature of about 275C, is joined to the spinel tube for cells III, IV, and V.

For this qualitative demonstration of the several chemical reactions involved in the process, there is no movement of solids. However, in a plant, solids may be moved either batchwise or continuously from right to left, through cells in the order I, II, III, IV, and V, with an entire by-pass of the condenser for FeCl and the condenser for aluminum metal. Actually the cells are specific chemical reaction zones. One reaction may be conducted in two or more cells or zones; being largely completed in one, then finished in the next one. Thus, a stripping of one component completely eliminates it in a final reaction zone.

The several zones or cells I, II, III, IV, and V may be assumed to be preliminarily charged respectively with the solid oxides indicated in each, together with coke as the reductant where necessary, as might be the resulting composition therein, after a steady state had been developed in the system of an operation which did have the possibility of moving the solids right to left. For this demonstration, any suitable divider 15 may be used to separate the solids in the several cells, while allowing gases to pass through freely.

The ore used in the general demonstration of FIG. 2 contains the three oxides of Fe, Al, and Si, which would be added in a plant operation only from the right side. Coke, as in a plant operation, is precharged in the mixture to cells II and III. It is assumed that coke is present in the plant operation in a lesser amount in cell IV and in a still lesser amount in cell V; so smaller amounts are precharged in the mixtures to the respective zones in this demonstration.

Thus, carefully sized and dried samples of the particulate or pulverent constituents of the ore and of the coke are preliminarily prepared in the assumed ratios and charged to the respective cells.

Chlorine gas is fed in at the left, through inlet tube 10, and it and/or the gaseous products of the reactions, pass through and in intimate contact with the pulverent solids in each cell in succession from left to right. This gas movement is in countercurrent to the movement of ore and its resulting solids, also of coke, when such solids were moved in a continuous plant operation. Gases moving left to right in this demonstration are shown by dashed arrows.

These gases leaving any cell may contain some hydrogen chloride, HCI, formed from the hydrogen in the reductant, coke; and also there will be carbon oxides, CO and CO from the reductant coke uniting with the oxygen of the oxides. Therefore, I-lCl, CO, and CO are present in the gas streams leaving the cells; but for simplicity, these gases are not indicated on each of the arrows. The exception may be that the exit gases will contain little or no I-ICl, but water instead.

In FIG. 2, the temperatures are assumed to be maintained by heat input if necessary from the furnaces outside the wall of the reactor tube, while in an actual plant the temperatures would be maintained by the heat given off by the reactions of the chlorinations. An important temperature control is that in cell III, which should be maintained at about 1,200C. This temperature is necessary to change the aluminum for trivalency to mono-valency and to give one molecule of AlCl from one molecule of AlCl while chlorinating one molecule of A1 0 to give twice as much additional aluminum in the gaseous mono-chloride as in the gaseous tri-chloride.

The temperatures in the cells II and V are maintained at about 750 and 900C, respectively, although an exact temperature control is not necessary as in cell III, and either temperature may vary between about 700 and 900 C. The higher the temperature in any reactor, the greater the speed of the reactions involved; and higher or lower temperatures may be used if desired, to increase or decrease the speed of the respective reaction in order to obtain the completion of the reactions assigned to each reactor at about the same time i.e., that of the cycle.

The temperature in the condenser for aluminum vapors, 3, however, is maintained at about 700C by cooling it to cause the reconversion of three molecules of AlCl to AlCl with an elimination of two atoms of aluminum as metal. Since this aluminum comes out of the gas phase of AlCl as a liquid, the condenser is maintained at a high enough temperature so that the aluminum, which freezes at about 660C, does not solidify. A lower temperature might increase the rate of production of AlCl but is not necessary; and the gaseous AlCl if cooled to a lower temperature, requires more heat to be supplied for its reheating in cell II.

The temperature of the condenser for ferric chloride, 18, is maintained at about 275C so that the vapors of AlCl will not condense at its condensation point, about 180C, but those of FeCl will condense at any temperature below about 300C. Remaining AlCl vapors condense as solids to preheat the solids of cell I (in a plant operation as much as possibly while preventing the discharge of the AlCl In the demonstration diagrammed in FIG. 2, the three pulverent and well mixed solid oxides of iron, aluminum, and silicon are charged into cell I, and are being contacted by the gaseous AlCl coming from cell II, probably containing free I-ICl.

In cell I, there is no coke and, besides the preheating of the solids, there isa chlorination here by reaction of the gaseous HCl with Fe O and A1 0 to give FeCl and AlCl which are solids. However, all of the Fe O in cell I may not be reacted; and some small amount, together with the A1 0 and the SiO may be found in the solids which by-pass the condenser for FeCL- and thus reach cell II. The criterion used in controlling the operation of cell I is that no AlCl or FeCl is allowed to discharge from it in the exit gases, and hence from the system as a whole. Thus, a temperature below the condensing point, 180C, of AlCl; is maintained; i.e., about C. This may require the cooling rather than the heating of cell I.

For those ores with a large enough amount of Fe O to combine with all of the AlCl leaving cell II and amounting to one-half as many moles of F e 0 as A1 0 in the original ore, it may be desirable to operate Cell I differently, as will be explained hereinafter. Actually, cell I operates without chemical reaction except the attack of HCl on Fe O and A1 0 It is a means of preheating the raw ore, while condensing AlCl; to a solid phase so that it may be returned conveniently via cell II to cell III.

No solids are charged to the condenser for F eCl and in a comparable plant operation, solids would by-pass this in going from cell I to cell II. This operates at a temperature sufficiently below the condensation temperature of FeCl;, which is about 315C, to condense it out, substantially pure and completely, while operating sufficiently above the sublimation temperature of AlCl (l 80C) so that practically no AlCl condenses.

In cell II, the charge consists of pulverent A1 and SiO as well as a much smaller ratio of Fe O than went into cell I, and a small amount of powdered coke as a reductant, also possibly a small amount of AlCl and FeCl;,. All have been well mixed. Here there is the reductionchlorination of A1 0 by C1 as shown in Equation 1, if any has penetrated so far to the right; and also of the small amount of P6 0, which may not have been removed in cell I. (In some operations, no coke would be added here.) Gaseous AlCl comes from the condenser for aluminum; and it reacts in an oxidationchlorination with Fe O Cell II operates to give a complete cleanup or elimination from the solids of Fe O which, however, is present in the original ore only in a relatively small mo lecular ratio to the aluminum, otherwise a different operation is used.

In cell 111, there has been charged A1 0 SiO and coke, all finely powdered and well mixed. This is the principal converter or volatilizer for the aluminum from the A1,,O as AlCl at a temperature of l,200C, or slightly less; but well over 1,000C. The three gaseous chlorinating agents, AlCl SiCl also C1 come in greater or lesser amounts of each in the gas stream from cell IV to act with coke in the reductionchlorination attack on the A1 0 in the ore. Each of these three gaseous chlorination agents produces, with alumina, the mono-chloride of aluminum, gaseous AlCl, as shown in Equations 3a, b and c respectively. These chlorination reactions, together with heat from the furnace in this example, supply the considerable heat requirements to bring the temperature to above 1,000C and preferably nearer l,200C. Aluminum is mono-valent at these temperatures. The control of the gas flow to this cell III provides that no free C1 and preferably, but not so rigidly, that no SiCl, be allowed to pass into the vapor stream leaving to go to the condenser for aluminum. Desirably, but not absolutely necessary, this gas stream going to the condenser would contain no C1 AlCl or SiCl but would contain only AlCl (plus, of course, CO and CO Cooling of the condenser for aluminum metal, 3, removes the considerable heat of the gas stream and due to:

a. the sensible heat of the gas stream, principally AlCl at its high temperature; b. the heat of formation of 2Al(metal) and AlCl (gas) from 3AlCl(gas); and

c. the latent heat of condensation of the aluminum metal as it simultaneously changes from the vapor state to the liquid state.

This large amount of heat is removed by cooling the wall of the condenser. Liquid aluminum accumulates during a run in this laboratory demonstration; but it would be removed in a suitable drawoff in any plant unit.

The condenser section for aluminum is not charged with any solids in this operation, nor would any ore solids be allowed in it in a plant unit; they would by-pass it in going from right to left.

Cell IV has been charged before the operation with SiO and much less A1 0 also coke all in a fine particulate or powder form and well mixed. Here, there are also the same reactions as in cell III; but there is also the reaction of the gaseous stream of SiCl, coming from cell V with AI O to form SiO and MC], and also probably some AlCl at the somewhat lower temperature which is desirable in cell IV, 950 to l,lO0C.

Cell IV is the general purpose reactor wherein is:

a. produced much of the AlCl which is formed;

b. removed from the solids to pass to the vapor phase of all but a trace of the aluminum in the ore;

c. removed substantially the last of the free chlorine coming in the gas stream, also most, if not all of the SiCL, coming from cell V, by its oxidationchlorination of A1 0 Cell V has been preliminarily charged with Si0 and a small amount of coke; both are in a particulate or powdered form, and well mixed. Here is accomplished what might be called a stripping from the solids of the last trace of A1 0 to leave substantially only SiO in the gangue which is then discharged from the system.

In a plant operating with any usual ore, the gangue discharged would also contain the non-volatile chlorides of the alkali metals and of the alkali-earth metals, also of some other metals, if they have appeared in the original ore or in the ash of the coke which is used as reductant. Some of the silica, SiO reacts with the chlorine entering the system, to giveSiCl and this, with the C1 goes in the gas stream to enter cell IV, but it becomes SiO in reacting with oxides of metals higher in the affinity series, and passes in the solids back to cell V.

This somewhat idealized demonstration shows the several reactions all culminating in the condensation of pure aluminum metal in its condenser, while separating the iron as the volatile chloride by condensing it in its condenser, and removing the silicon as a gangue of SiO It does not illustrate a practical system for production use because the cells of the tube would become depleted of the materials originally charged (except for the SiO and there is no provision for recharging, nor, indeed, of passing solids from right to left, nor for charging coke or adding chlorine at intermediate cells.

It should be noted that, in those ores with half or more as many molecules of A1 0 as of Fe O cell I would not be operated as described above, but it would be operated as an oxidation-chlorination reactor at a temperature of 650900C; the condenser for FeCl would be placed downstream of the gas flow from cell I, i.e., off the flow sheet to the right; and AlCl entering this cell I would interact with the Fe O to give A1 0 in the ore solids, which would later pass to the left and FeCl which would pass into the gas stream and thence to the right and to the condenser therefor.

The furnace for III maintains the high temperature, l,200C, for the formation of AlCl. Oxygen could be introduced, but with difficulty in this small unit, to burn carbon to give such heat internally, as may be done in plant units.

Cyclic Series of Reactors A plant system to produce aluminum directly from aluminacontaining ores, which also contain silica and iron oxide, is shown in FIG. 3. This system may be operated also by the flow sheet of FIG. 2; but here it is operated with a flow sheet for an ore wherein the molecular ratio of iron to aluminum is greater than about onehalf and the FeCl produced therefore is discharged from the system rather than being condensed in the system as in FIG. 2. Here, each component reaction zone or in this case more simply reactor for contacting solids with gases corresponds to one of the cells of FIG. 2, except for the somewhat different operation here of the No. I reaction zone.

Solids remain always in the cell or reactor to which they are charged. The reactor and solids container therein are changed as an entirety after a cycle of reactions therein is complete and become the next higher stage in the process. The solids are then contacted by a gas stream of different composition, i.e., that entering the next higher stage.

The six identical reactors of FIG. 3 form a regular hexagon around a central condenser, 3, which condenses out liquid alumimum. At any one time interval or cycle of the operation, five of the six reactors are in service, as were the five cells of FIG. 1, and are similarly numbered I, II, III, IV, and V, while the sixth, VI, is being dumped of the gangue, principally silica, and charged for the next cycle with the fresh ore without coke which is here again considered as being composed of iron oxide, aluminum oxide, and silicon dioxide. Each reactor vessel is fitted with a connection, 19, so that it may be charged with coke before it enters the position of No. II or No. III. This may be the same connection as for charging with ore, or a different one, as desired. Connections, 24, allow gas flow from reactor to next lower numbered reactor.

Also, each of the reactor vessels has a means for connection to the condenser, 20, for aluminum metal of a separate short section of pipe for such connection when the reactor is in the position of No. III, to supply the MCI vapors to the condenser for aluminum or in the position of No. II for receiving AICI vapors, 21, from the condenser. The mechanics of all such connections and of their closure when the reactor vessel is in a position where their use is not required, are simple, and are not a part of this invention.

In any cycle, since the solids are not moved, the several reactors charged with the solids as they remain at the end of the last cycle are supplied with gases in sequence until the particular charge of the reactor is fully converted for the particular reactions taking place at that stage or cell. Then the reactor, VI, which was not in service and was being charged during the cycle just finished, is connected in as No. I, while the previous No. I is connected in as No. II, and each successive reactor becomes the reactor of next higher number.

The interconnections to and from the condenser for aluminum are adjusted so that gases from No. III always enter the condenser, and those leaving the condenser always go to No. II. Finally, the old No. V is taken out of service for discharging of the residual gangue and recharging with new ore, while the connection for chlorine addition is made to the new No. V. The sequence of the reactors for each separate cycle continues in a clockwise direction, and thus no solids are moved except to charge the ore and coke and to discharge the gangue; and provisions are necessarily made so that inlet and outlet of all solids may indeed be made to each reactor-vessel, regardless of its physical position when it comes up as the appropriate number in the cycle for that input or output.

Thus, each reactor must have connections provided for addition of the ore being charged before it is placed in the No. I position, also for addition of coke when it is in the No. II and No. III positions. This may be 19 for both services. Chlorine enters inlet 10 when the cell is in the No. II, No. IV and No. V position.

FIG. 3 shows connection 10 only for these positions where used in this cycle; but a similar 10, also 19, must be on each vessel since all occupy each position at different times. As hereinafter explained, connection 10 for chlorine is also connected to oxygen (or air) supply.

Reactor No. I, for example, is operated to give an oxidation-chlorination without coke, in this case according to the reaction of Equation 4 at about 700800C, until all but a very small amount of iron in its oxide has been chlorinated to vaporize and pass off as FeCl In this operation, there is assumed to be in the ore a substantial amount of iron, as, for example, over onehalf the number of atoms as compared to the aluminum. (In the case of a lesser amount, the iron would be removed in a condenser intermediate to cells I and II, as in FIG. 2.) In FIG. 3, the FeCl leaves the system from reactor No. I as a vapor and may be condensed as a liquid at 315C or a solid at 282C. It is separately worked up for pure iron in ways well known. This condenser, for FeCl not shown, may be attached as to outlet 13. Since the same condenser for FeCl may be used for each of the reactors when its turn comes for it to become No. I in the cycling, a convenient location, like the condenser for aluminum metal is at the center of the circle formed by the reactors. If so, it is located either above or below the condenser for aluminum metal and on the same axis.

If the ore contains zinc and if reactor No. I is operated at 800C, ZnCl is formed and comes in the gas stream (B.P. 732C). It may be separated in a pure form by an additional dual-chlorination system; or it may come over with FeCl and other chlorine compounds formed more readily than AlCl (i.e., higher in the affinity series). Because of its high condensing temperature compared to other gases present, it may be condensed fractionally at 700C before separating FeCl liquid at 300C, and the fractional condensation of other chlorides or oxychlorides at lower tempera tures as noted elsewhere.

Meanwhile, reactor No. II during this period or cycle is operated at about 700-800C in a reductionchlorination stripping for the iron, so that the very last of the iron oxide is converted to FeCl Any chlorine, as needed during the reaction for the complete chlorination of the Fe O in addition to the AlCl entering through 21 is added through 10. Additional coke may be added through 19 during the down time between cycles.

Similarly, coke is added through 19 to reactor No. III during the down-time between cycles. The chlorine connection IOTs also fitta'iaan'tfiy'gensubb'iywh'ich may be provided alone or simultaneously with the chlorine. Oxygen (or, less desirably, air) is added to burn with coke in the solid mixture, to raise and maintain the temperature in III to about l,200C, while chlorine is added to supplement, if necessary, the chlorinating gases coming from IV through 24. The combination thus chlorinates much of the A1 0 in III at a temperature up to 1,200C to pass the gaseous AlCl formed at this temperature to the cooler-condenser 3 for AlCl, to allow it to change valence and to give up the aluminum metal. This is a liquid at 700C and is drained from the bottom to be cast as ingots. Also, in reactor No. III, the operation has proceeded to absorb all of the chlorine coming through 24 from No. IV so that no chlorine is allowed to pass off in the gaseous AICl going to the condenser.

During this time, reactor No. IV has been acting as a stripper at 800I,O00C, to remove, by reductionchlorination, substantially all of the A1 which it has received from No. III, which is with all of SiO of the original ore. The gas stream from IV to III, through 24, thus carries AlCl and SiCl, along with some C1 In reactor No. V during this cycle, SiO has been reduction-chlorinated at 750900C to SiCl which has been passing off as a gas and which then must be completely utilized in Nos. IV and III to displace aluminum from A1 0 to give AlCl or MC]. The Si0 so formed by SiCl, in chlorinating A1 0 will remain in these reactors until each, in turn, becomes No. V, after which it is discharged with the gangue from the system, when it is in the off-position of VI.

No. V will preferentially chlorinate any trace of A1 0 which comes through from IV and thus acts as a safety against its loss.

The temperatures of the individual reactors are adjusted so that the reaction velocities, which increase with higher temperatures, allow each reactor to accomplish its appointed reactions within the same period of time, i.e., the period of the cycle. As noted before, the temperature in reactor III must be maintained at approximately l,200C in order to allow only the monochloride AlCl to come over to the condenser. This is adjustable by the addition of coke through 19, and oxygen (possibly air) through to burn internally and supply heat required to maintain III at this high temperature.

Any excess of coke charged in No. III remains in No. IV and then passes to No. V. The high temperature, l,000C, in No. IV may be maintained by addition of oxygen in the chlorine supply line 10 connected to IV or to oxygen in the chlorine supply to No. IV.

However, the temperatures in the other reactors are variable within rather wide limits and are adjusted to complete, or substantially to complete, the reactions assigned to that stage of the operation within the limitations of the time required for the conversion in No. III. The temperature in No. II may also be increased by adding more coke there; and provision is made for adding oxygen through its inlet 10, also to burn coke. Since a higher temperature in No. II increases the temperature of the gases going to No. I, it speeds up the reaction there also.

Each reactor is thus fitted with a connection 10, not shown for Nos. I, IV, and VI, since it is not used with these during the part of the cycle discussed. Oxygen or chlorine, or both simultaneously, may be carefully metered into any reactor through its own 10 and with coke in the solids. The oxygen burns with coke to give internal heat to raise the temperature as desired for the particular reaction, and thus to increase its rate. Normally, the chlorination reactions supply the necessary heat; but at start-up, with all reactors cold, or with a change of cycle with the new reactor I being placed on line, or with an unexpected change of ore or other upset, the

temperature of any reactor may thus be raised, if coke is present, by supply of oxygen.

This, as other systems, uses a condenser for FeCl on the gas stream leaving reactor I because of the substantial amount of Fe O in ore, so as to react with all AlCl coming from II. However, if the ore contains P, S, V, Ge, Ti, Zn or other metals higher in affinity series than aluminum, they all will add with iron in using AlCl to form chlorides or oxychlorides, thus using AlCl present and equal to one half of aluminum atoms in the ore.

Reactions in Batteries of Cyclones The reactions in the cells of FIG. 2 or the reactors of FIG. 3 are between gases and finely divided solids. One type of reactor which has been found useful in such service for continuous use is the cyclone battery of reactors, each of which may, in fact, be an interconnection of several individual cyclones operating as a single reactor, thus giving counter-current flow of pulverent solids and reacting gases. Such a reactor is again a unitary reaction zone or cell for the purpose of comparing the several mechanical systems for accomplishing the processes of this invention.

Here, the feed of the pulverent solids, 25, is blown from the previous reactor in the cyclone reaction zone, by the gases which are to be reacted therewith and which come from the following reactor. This allows an intimate contact of the gas and powder on entering the cyclone, wherein there is also a further excellent contact. The solids, because of their greater density, whirl in a thin layer on the inner wall of the cyclone, with the gas in intimate contact. Finally, the gas separates free of the pulverent solids, to discharge from the top of the cyclone through line 24 after the reaction has gone to its substantial completion under the conditions pertaining. Such cyclones are in use in the cement industry and elsewhere for heating fine materials to go into the cement-making process, and are also used for various other contacting and chemical reacting between gases and solids.

The design and operation of the cyclone batteries to give the intimate contact required for the particular reaction are more or less standard; and it is not part of this invention.

FIG. 4 diagrams such a system, whereas each one of the five numbered cyclones are indicated as a single unit or reaction zone, as, for example, the individual reactors or cells of FIGS. 2 or 3. It may be desirable to have more than one cyclone for any one or more of the five reaction zones accomodating the respective reactions of FIGS. 2 or 3 in contacting the pulverent solids and gases. Several or more actual cyclones might therefore take the place of any one of units diagrammed in FIG. 4; and all would, of course, operate in countercurrent.

lndicated as another cyclone-type unit is the condenser for aluminum metal, 3, wherein the aluminum mono-chloride, AlCl, is passed in at about l,200C to be cooled to about 700 C This converts it to AlCl plus the aluminum which is condensed on the side walls and then passes out the bottom, 16. This is actually a chemical reaction zone of the disproportionating of AlCl, but is not a numbered reaction zone in the present discussion because it does not involve contacting of ore particles and gases.

In this system of FIG. 4, the operation of each numbered cyclone may be regarded as that of a reaction zone or cell of FIG. 2, except that there is the movement of solids continuously, always downwardly, from the dried and pulverent or granular ore coming in at 22, with gangue leaving at 23. Necessary heat to maintain the temperatures is supplied in each cyclone by the exothermic heat of the chlorinations, as previously indicated. Chlorine is supplied through the lines as shown in FIG. 4 in admixture with the particulate solids going to cyclines Nos. II, III, and V, those wherein chlorine might normally be added. However, the line 10 may also be connected to No. l and IV. As previously indicated, oxygen may also be supplied either alone or in mixture with chlorine through any of these lines 10, to burn the coke in the solids and supply necessary internal heat to maintain the desired temperatures in each cyclone.

One conventional type of condenser long used in the art to give liquid aluminum from a gaseous phase, as here, 3, uses a mass of molten aluminum as the coolant which is sprayed into the gas phase to cool it and condense out the aluminum therefrom. The molten aluminum is cooled, in turn, by contact with a molten salt insoluble in the aluminum, which is cycled through an external cooling system. The mechanical design of such a condenser for aluminum is not a part of this invention, which involves only the use of a gas cooler and condenser for aluminum of any of the suitable types for the processes described. Here, the MCI in cooling releases metallic aluminum which collects on the cold surface 18, (700C) and drains out the bottom of 3.

FIG. 4 has also a condenser for FeCl in between the cyclones No. I and No. II, as was previously used in FIG. 2 for processing. continuously those ores with a relatively low iron content. FeCl is here condensed out as a solid, 26, from the vapors from No. II going to No. I. The flow sheet of FIG. 3 is used equally well with cyclones for discharge of gaseous FeCl from the exhaust gases of No. I in those cases where a larger amount of iron is present in the ore; i.e., as much as one-half the amount of aluminum on an atomic basis. As in FIG. 3, provisions are made by connections, 10, for chlorine inlet to both cyclones No. II and No. Ill, if additional carefully metered chlorine and/or oxygen is desirably added to either or both of these cyclones, as well as to No. V. Also connections, 19, are provided for inlet of dried and pulverent coke into the streams entering No. II and No. III.

As previously noted, the atoms of other elements whose oxides are chlorinated by AlCl to form chlorides or oxychlorides (e.g., P, S, V, Ge, Ti, Zn) will add, on a chemical equivalence basis, to those of iron in determining whether there is twice the amount of aluminum present so as to react with all AlCl entering I. If so, all of these chlorine compounds should leave I in the gas stream, 13, for later condensing or other processing.

Use of Multi-Hearth Furnace Another type of counter-current contactor or reactor, which has been widely used for drying and calcining in other processes where gases, usually hot, are counter-currently contacting a wide variety of pulverent solids, is identified by the names Nichols, Wedge, Herreshoff, or McDougall. The number of hearths used in such furnaces varies from two or three to 20, in diameters up to 7 meters. They may be, like the cyclones, made of steel and then lined with refractory materials which are also resistant to the corrosive action of the gases. They are in common use; and their physical design and that part of their operation which is conventional is not a part of this invention. However, their particular operation accomplishes the chemical reactions and physical operations of the present method of treatment of ores, particularly with the accessories described hereinafter, which are also, by themselves, more or less standard in design.

The flow sheet of the present process in such a continuous multiple-hearth furnace operation is diagrammed in FIG. 5, which operates in a similar manner to the demonstration batch operation of FIG. 2. Here again, the dried and pulverent ore enters the top, 22, and passes downwardly, 25 from hearth to hearth, through the furnace with an ultimate discharge of the gangue, principally SiO at the bottom, 23. The reductant, powdered coke, is also fed through intermediate levels, 19, to No. II and No. Ill, and chlorine enters through inlets 10 in precisely metered amounts.

As indicated above, if adequate coke is present, oxygen may be added as desired through these connections 10, or through similar connections, not shown, to other hearths or reaction zones, to burn the coke and thus attain or maintain the desired high temperatures.

Again the reaction zones or cells are numbered from I to V, and again it may be necessary to have from two to four actual hearths to take the place of one reaction zone or cell as defined by the reactions to be accomplished therein and in accord with the other systems. Each reaction zone will be called hearth No. I, hearth No. III, etc., with the understanding that this may actually refer to two to four actual hearths in practice to achieve the desired counter-current and stripping reactions already described as regarding the corresponding cells or reaction zones.

In this type of furnace, the standard details of which are not indicated, the several hearths are inside of the steel shell with refractory lining. A vertical axial shaft rotates carrying arms supporting, over each hearth, rakes which move the powdered solids around the hearth as they turn them over to contact the gas. The solids on one hearth are raked to the outside to discharge through a suitable hole near the periphery, through which they fall to the hearth below, this discharge being indicated by the arrows, 25. Here the process is repeated; on the next lower hearth, the rakes move the solids to an opening near the central shaft to fall to the successively lower hearth. Gases go upwardly through these openings, the upward passage here being indicated by dashed line arrows, 24, and over the solids in each bed, always to the next higher hearth, so as to give a counter-current operation.

The individual reaction zones or hearths are identified by the same numbers as the cells in FIG. 2. The temperatures and reactions may be nearly the same as those already described under FIGS. 2 and 4. A side drawoff for vapors 20 of MC] from the cell or hearth No. III goes to a cooler-condenser for aluminum, 3, outside of the furnace, which is cooled and maintained at a lower temperature of 700: 25C, to condense out the liquid aluminum passing out at 16 and to return the MCI, formed to the hearth No. II of the furnace proper through the vapor connection, 21.

Similarly, vapors of FeCl and other gases are withdrawn from hearth No. II; cooled in a condenser for FeCl to 275C, and passed out at 26 as solids. A temperatureof 300C allows a liquid drawoff at 26. This temperature is sufficiently high to prevent the condensation of other materials in the gaseous stream on its way to hearth I. The solids from hearth No. I drop directly to hearth No. II, then to hearth No. III, in each case by-passing the respective condensers for FeCl 18 and for aluminum metal, 3. Thence the solids pass via hearths No. IV to No. V, from which they are discharged as gangue from the bottom through 23. Chlorine is added at the bottom through 10, and passes across the hearths in counter-current to the flow of powder; and it or other volatile chlorine compound or compounds, plus CO and CO pass upwardly through the openings, down which the powder descends from hearth to hearth. There may also be added through connections, 10, to hearths No. II or No. III, chlorine as may be necessary to complete the desired reactions in those hearths, or with added oxygen to maintain the temperatures desired, as described above.

A mixture of particulate or pulverent ore is added at the top through 22, and pulverent coke may be added to hearths No. II and No. Ill by appropriate feeding connections, 19. In the multiple hearth operation of FIG. 5, as in the multiple cyclone operation of FIG. 4, excellent mixing of solids is obtained in each reaction zone; and this is particularly important in the intermediate feeding of coke, always pulverent or particulate.

While this flow sheet FIG. 5 indicates a continuous operation similar to that of the intermittent and batch operation of FIG. 3, the multi-hearth system (as also the cyclone system of FIG. 4) may be operated continuously with the general flow sheet of FIG. 3, i.e., with an external F eCl condenser, to which pass through the exit gas connection 13, effluent gases of hearth No. I carry the gaseous FeCl Herein, as previously noted, the FeCl may be fractionally condensed by operating at about 275C; and, if TiO is present in the original ore, the corresponding TiCl, may be fractionally condensed out in a second condenser operating at about I l5l20C before condensing out, at the low temperature of cooling water other materials as POCl depending on the composition of the ore, and scrubbing out HCl which passes the condenser along with CO and CO.

Separation of Elements in Ores Other Than Al, Fe, Si

In the attack on aluminum ores by the reductionchlorination and oxidation-chlorination method, many materials are chlorinated in the five reaction zones, with, however, some differences in the reactions conducted in each. The readily formed and volatile TiCl S VOCI, POCI also GeCl if any or all of the corresponding materials are present in the ore, may all be discharged along with FeCl in the effluent gaseous stream, 13, from reactor No. I if the flow sheet of FIG. 3 is used; and if the amount of chlorine metered into the system stoichiometrically equals that in all of these gases and that of the accompanying HCI. None of these gases will condense in the FeCl condenser, which would be appended to the effluent gas connection of reactor No. I of FIG. 3; nor in the FeCl condensers of FIG. 4 or FIG. 5 if the operation of this process is conducted with ores containing any or all of these parent materials, according to these flow sheets. These mixtures of gases, all of which are condensed considerably below 315C, the condensation point of FeCl then pass on to separation operations known in the art.

Various ores, including clays and shales from which aluminum may be directly separated by this process, may contain up to 10 percent or more of magnesia along with more or less calcium, sodium, and potassium. Under proper conditions, the oxides of these metals may be chlorinated also; and these chlorides, if formed, boil at such high temperatures that they will remain unchanged throughout the temperature ranges of the several reactions and be discharged in the gangue along with the silica. Such discharge of chlorides of these alkali and alkali earth metals represents a loss of chlorine, which must be charged against the cost of the operation. However, all of these chlorides are water-soluble and may be dissolved from the gangue if desired. The magnesium chloride, usually the most valuable, may then be precipitated and worked up as the hydroxide.

Manganese, if present, is usually in amounts of not over a few percent; but its chloride, MnCl if formed, boils at 1,190C and passes in the stream of AlCl gas to the condenser for metallic aluminum. A condenser for MnCl operated at l,l50C might fractionally condense it out before cooling the MCI gas to cause it to disproportionate at about l,O0OC or higher. However, the physical condensation of MnCl with the aluminum metal presents no problems, since it may be readily separated. MnCl also solidifies at almost the same melting point as aluminum metal, but it has a higher density. It is readily soluble in water.

Traces in the ores of those other metals higher in the affinity series for chlorine from the oxides, i.e., lead, zinc, cobalt, or nickel may also form chlorides; and these, PbCl (B.P. 950C), ZnCl (B.P. 732C), CoCl (B.P. 1,049C) and NiCl (B.P. 973C) would volatilize at the high temperature (l,lO0-l,200C) required to change the valence of aluminum and condense with the aluminum metal. These metals are seldom present in the ores which would be used for aluminum; but their chlorides are also readily separable from the aluminum metal coming from its condenser should they collect there.

Aluminum From Slime Solids from Phosphate Rock Production Aluminum metal may be won and refined by the process of this invention from many and widespread aluminum ores, compared to only one, high grade bauxite, now used conventionally for processing by current methods.

One source of alumina now made usable by the processes of this invention is the solids in the slimes produced during the beneficiation of phosphate rock, as, for example, in the Florida phosphate industry. Here, single operations must dispose of 10 million tons/year of slime solids in producing 8 million tons/year of phosphate rock. This industry, as a whole, must handle over 30 million tons/year of slime solids; and there is a stockpile of hundreds of millions of tons of such waste material taking up thousands of acres of land which otherwise would be very valuable. There is no present substantial use of these slime solids; and the cost of disposing of the slime approximates 25 cents per ton of solids or 32 cents per ton of phosphate rock produced, which totals over $3 million per year for just one operation, or 10 million for the entire Florida industry. Slimes are pumped to waste disposal ponds which are expensive of land area and have other major disadvantages; they remain indefinitely, charging some cost for maintenance besides the loss of rental value of the land for other purposes.

Thus, slime solids are not only available without cost, but with an actual negative value, i.e., their cost of disposal. On the other hand, the very small particle size of the slime solids makes them particularly suitable to the processing of the present invention. Besides the alumina which is present, there is a high value of P which is recoverable by the process of this invention directly as phosphoric acid, the sale of which is the business of the companies mining phosphate rock. There is also a substantial amount of iron to be recovered. A typical analysis of slime solids might be as follows on the dry basis, i.e., totally free of water:

M203 18.70% mo, 7.48% s 31.60% 1 ,0, 16.10% CaO [8.70% MgO 2.09% MnO 0.09% no, 0.73% c 0.62% c0 1.24% Na,o 0.41% 14,0 0.69% F 1.55%

TOTAL 100% This analysis shows a disadvantage of these slime solids as a raw material for use in the chlorine process for aluminum separation. The CaO, Na O and K 0 present will require chlorine to give the corresponding chlorides which will have no sale or value. MgO gives MgCl and this may be sold if it is dissolved from the gangue, and then precipitated as the hydroxide.

As a first step in handling these slime solids, all water, including that present in hydrate form, must be removed by'drying or roasting, since water forms additional HCl. Techniques for doing this, however, are more or less standard.

In using the chloride separation techniques, the methods of this invention, described above, the P 0 is chlorinated to POCl preferentially even to iron and volatilizes first; this is condensed; then readily converted into phosphoric acid while distilling off the accompanying hydrochloric acid. Next, the iron is removed as FeCl which in turn will be converted back to pure Fe O with recovery of Cl for reuse, as in other examples. The A1 0 is converted to aluminum metal by the process of this invention. The titanium would go off as TiCl,, which could be separated as indicated above if desired, in a separate reductionchlorination system of the Halomet process, or from a mixture with FeCl if this and TiCl, are chlorinated together. The TiCl. and FeCl would be separated by fractional condensation of FeCl;, first, then TiCl.,, by fractional distillation, or both. The SiO- would be in the gangue, along with CaCl MgCl MnCl NaCl, and KCl, together the F as NaF and KF. The C and CO would be lost.

Only about one quarter as many atoms of iron are present as of aluminum; and this amount of iron is insufficient to react with the AlCl leaving the condenser for aluminum. However, the very large amount of P 0 indicates that the system of either FIG. 3 or FIG. 5 may be used.

Considering the flow sheet processing of FIG. 5, using a multi-hearth furnace, the effluent gas from the top, 13, contains POCl and HCl, also the very small amount of TiCl, for the making of which chlorine has been supplied, as well as CO and CO etc. The POCl and HCl would be condensed in an external condenser, not shown. An absorption column using a water wash would follow this to scrub out residual HCl from the final gas discharged. The dilute HCl formed in the absorber would be added to the condensate, from which HCl would be distilled. Phosphoric acid is a residue; and it may be concentrated or sold as it is after removing HCl.

The chlorine feed would be metered carefully so that the FeCl and TiCl, formed would leave in the vapor stream from the reaction zone of hearth No. II to 18. FeCl would be condensed out fractionally in the FeCl condenser shown at 300C at a condenser temperature of about 250275C. The TiCl, and other gases pass uncondensed in 18, then through the hearth No. l at a @rnneraturs show t e con en in 29in! of T Cll, They would pass out to a condenser operated at about l05-125C to condense TiCl,; and the HCl, H 0, and POCl would condense at the temperature of cooling water, for further separation.

Alternatively, aluminum could be won from slime solids in the flow sheet of FIG. 3, as a desirable system of making this separation in a continuous operation, e.g., with batteries of cyclones or a multi-hearth furnace. The temperature of the reaction zone or reactor No. I would be 750800C; and the FeCl TiO POCl S0 l-lCl, CO and CO would all discharge through 13, first to a partial condenser at 250275C for condensing FeCl then to a condenser operating at about 20C for the TiCl,,, and an absorption tower for removing other materials except CO and CO principally l-lCl. An item of major value in this example is phosphoric acid. It is removed as a residue after distillation off of other volatiles, including the l-lCl coming from the cold condenser and the absorption tower.

As another recovery system for vapors leaving in 13, a partial condenser for the FeCl would separate it at 300C; the TiCl would then be condensed in a separate condenser at C, considering its boiling point of 136C, while the POCI and l-lCl would be removed by condensers and scrubbers later, at the lowest temperature practical with the available cooling water.

The condensed TiCl, may be separately rectified to remove l-lCl and any small amount of FeCl which is present. As still another system, the iron and titanium could be separated by operating another l-lalomet Duo- Chlorination system to accomplish the duochlorination separation, rather than depending on fractionation by the relative volatilies.

The aluminum would be separated out in the condenser for aluminum, 3, as already described; and the silica would be discharged in the gangue at the bottom, 23, where chlorine enters, 10. The CaCl MgCl MnCl NaCl and KCl would all be present in the gangue, from which they may be leached if it is desirable to recover the MgCl by conversion to the hydroxide in the standard system of wet chemistry.

7 Aluminum from Oil Shales Shales are formed through the compacting of clays under pressure. The oil in oil shales comes from sediments of aquatic organisms of both animal and vegetable origin, deposited with clays on lake bottoms and later subjected to high pressures. The oil is very intimately dispersed throughout the solids.

A principal constituent of most oil shales is alumina; and in general, there is from two to four times as much silica as A1 Iron is usually present as Fe O in about one-third to one-tenth the amount of A1 0 Oil shales as such cannot be used as fuel; thus, many separation processes have been developed to remove the 5 to 30 percent oil present. However, it has now been found that the more volatile materials in this oil may be distilled off and the-less volatile ones carbonized to be used directly as the reductive fuel for winning and separating the aluminum and iron by the processes of this invention.

In one typical oil shale from Nova Scotia, Canada, after oil was removed, the analysis was:

Oil (approx. 85% Carbon) l8.8% Water 0.8% Gas and Loss 2.0%

. Spent Shale Solids 78.4%

TOTAL 100% This shale has a heating value of 5,420 BTUs per pound. After combustion, 62.4 percent ash results, having an analysis of:

In a laboratory operation, utilizing the set-up of FIG. 1, the pulverized and dried oil shale is charged into the reaction zone No. III. In this case, the reductant is already intimately mixed in with ore.

The furnace 7 is heated to bring the material in the reaction zone Ill up to a temperature of about 500 while distilling off gases and oils which amount to about half of the oil content and contain most of the combined hydrogen. This first step may be done in a drying operationg outside of the reaction zone. In either case, these gases are burned for heating, in various steps of the process; e.g., in heating up the ore to 500C while distilling off volatile fractions of the oil. Alternatively, carefully metered air may be passed in through the tube 10, to burn some of these hydrocarbons in place, for the preheating of the balance of the ore.

Chlorine is then introduced through 10 to give the typical reduction-chlorination of A1 0 according to Equation 1, and of Fe O by the comparable reaction. The highly carbonaceous and very intimately mixed residues of the oil serve as reductant instead of the coke usually used. The temperature of the reaction zone is increased to 800C as the iron is driven out as FeCl Then the temperature is increased to 1,200C; and the subsequent reactions are accomplished as described above for the recovery of aluminum and of aluminum chloride in 4 and 3 respectively. The aluminum metal is removed as before, as is also the remaining gangue, largely silica and soluble metal chlorides.

In processing the next batch of oil shale, C1 is fed in at l l to pass from right to left; and the operation is continued as above described.

The plant operation for recovering the values of this shale may be conducted according to the flow sheet of FIG. 5, a diagram of the operation of a multiple-hearth furnace. Here, in the upper section corresponding to the reaction zone of hearth No. I, there may be added a connection (not shown) for the admission of air, to burn a part of the oil in the entering shale, raise the temperature, and distill out an additional amount of the light ends of the oil. The balance of the oil is carbonized at a temperature in the reaction zone of hearth No. l of about 500C, which would be the discharge temperature of the solids to the hearth No. II. No coke would be added there, since it would have been carbonized in place from the oil originally present in the charge of the pulverized shale, fed to hearth No. l, and thence to hearth No. II. Here again, the iron is chlorinated to give FeCl and removed as shown in the condenser, 18, on the right of hearth No. II, the FeCl condenser. No coke would have to be added in hearth No. III, as is usually done; and, in fact, there may be even an excess of coke flowing through hearth No. IV and hearth No. V to be discharged with the gangue out in a form from which it would be uneconomical to recover. As before, oxygen or air may be added simultaneously with, or separately from, chlorine to burn'with the oil to give heat if necessary in No. III or other reaction zone.

As in other examples, theFeCl may be burned with oxygen to give pure Fe O for reduction to iron, and giving C1 for reuse.

The light ends of the oil, i.e., the gas to naphtha fractions from the exit gas leaving 13, would be burned to supply heat and mechanical energy for related operations, or they may be further refined as salable hydrocarbon products.

Aluminum and Iron from Red Muds A substantial amount of the alumina in even the most desirable bauxites comes in a combination with silicate, which, when attacked with caustic soda in the convention process for the dissolution of aluminum, forms an insoluble sodium alumino-silicate. This transformation comes from the aluminum silicate or kaolinite of the bauxite. This insoluble aluminum compound, together with the iron oxides and titanium oxide and other impurities insoluble in the caustic soda, makes up a substantial amount of muds which appear as very fine particles and are difficult to separate.

Thus, there are large amounts of these red muds associated with conventional aluminum processing as an industrial waste, expensive of disposition. In some places, the amounts of red mud are very large, an accumulated waste from years of operation. Compositions vary with ores, but one analysis is:

M 26.0% F o, 50.5% SrO, 8.5% T10, 6.0% Na,0 7.0% C a0 2.0%

TOTAL 100% These percentages are on the dry basis, as would be obtained after heating the material to a high temperature to remove all combined water.

Here, again, it is necessary to separate iron as the chloride, aluminum as the metal, and silica as the gangue. The Na O and the CaO will be converted to chlorides which will remain in the gangue. The operation may be conducted generally according to the flow sheet of FIG. 2, which may be operated, however, in a battery of cyclones, FIG. 4, or in a multi-hearth furnace as FIG. 5. The-hot ore from a dryer operating at 500750C would enter FIG. at the top hearth or reaction zone of the furnace. The amount of chlorine metered into the reaction zone of hearth V would be stoichiometrically measured to produce only the iron and the titanium in the form of their chlorides, leaving the top hearth of the furnace. This gas stream contains also, possibly, any small amounts of impurities, as S0, and POCI as well as HQ]. The FeCl would be condensed out first in an external condenser, not shown, but operated at about 250275C; and the TiCl, would then be condensed in a second step at an operating temperature of its condenser of about 1l0l20C, while the other gases would be uncondensed, until they reach a third condenser operated at the lowest temperature of the cooling water. Exhaust gases from this would be scrubbed to remove I-ICl. FeCl may then be converted to Fe O and to iron, as above described, while the TiCl, could be redistilled, if necessary, to pro- .duce a high purity material, or in many cases may be used or sold as it is.

The operation of the balance of the multi-hearth furnace of FIG. 5 would be as above explained, with the separation of aluminum in its condenser 3 and the ultimate removal at the bottom. 23, of the gangue containing SiO NaCl, and CaCI I claim:

1. A process for producing aluminum from an ore containing aluminum and iron in compounds with oxygen, comprising:

a. contacting said ore with a first gaseous halogenating agent at a temperature above about 700C to form a volatile iron halide in a gas stream;

b. removing said iron halide in said gas stream;

c. increasing the temperature of the solid mixture remaining after step (b) to over 1,000C and up to at least 1,500C in the presence of a second gaseous halogenating agent and a carbonaceous reductant to produce with said aluminum compound the corresponding gaseous aluminum monohalide;

d. removing said gaseous aluminum monohalide and 32 cooling it t smpst mrs 91f about 0 19 f rm metallic aluminum and gaseous aluminum trihalide;

e. removing said metallic aluminum; and

f. recycling said gaseous aluminum trihalide to contact a further amount of ore in step (a) as at least a portion of the first gaseous halogenating agent which forms a volatile iron halide in a gas stream.

.2. The process according to claim 1 wherein said halogenating agent in step (a) and said halogenating agent in step (c) are chlorinating agents, said aluminum monohalide is aluminum monochloride, said aluminum trihalide is aluminum trichloride, and said iron halide is ferric chloride.

3. The process according to claim 1 wherein said halogenating agent in step (a) and said halogenating agent in step (c) are brominating agents, said aluminum monohalide is aluminum monobromide, said aluminum trihalide is aluminum tribromide, and said iron halide is ferric bromide.

4. The process according to claim 1 wherein the gaseous halogenating agent in step (c) is, at least in part, an elemental halogen.

5. The process according to claim 1 wherein in step (a) a carbonaceous reductant is added, said carbonaceous reductant containing compounds of carbon, sulfur, and hydrogen, wherein said compounds in contact with said halogenating agents and said ore form gaseous compounds along with said volatile iron halide in said gas stream.

6. The process according to claim 5 wherein said gas stream containing said iron halide is cooled below the condensation point of said iron halide but not below the condensation point of said gaseous compounds of carbon, sulfur, and hydrogen; and said iron halide is separated as a condensed phase.

7. The process according to claim 1 wherein said ore also contains a silicon compound with oxygen, said silicon compound constituting a part of the solids remaining after the reaction of step (c); said solids containing said silicon compound are contacted with a gaseous halogen in the presence of a carbonaceous reductant to form a volatile silicon halide; and the gas stream containing said silicon halide is passed to halogenate, at least in part, said aluminum compound in the ore to the corresponding gaseous aluminum monohalide.

8. The process according to claim 7 wherein said gas stream passes through a multiplicity of fixed zones containing said solids.

9. The process according to claim 7 wherein said ore and solids therefrom are progressively heated without reheating; said halogen is chlorine; the halides of aluminum, iron, and silicon are the corresponding metal chlorides, and the aluminum monohalide is aluminum monochloride.

10. The process according to claim 1 wherein said ore also contains titanium in a compound with oxygen; and said titanium is halogenated along with said iron to form a volatile titanium halide in the gas stream with said volatile iron halide.

Patent Citations
Cited PatentFiling datePublication dateApplicantTitle
US1866731 *Jan 28, 1929Jul 12, 1932Ig Farbenindustrie AgProcess for producing anhydrous aluminum chloride
US1987629 *Jul 2, 1932Jan 15, 1935Dudley BairdProcess for the production of metallic aluminum, or aluminum chloride and certain metals as by-products
US3186832 *Sep 5, 1962Jun 1, 1965Vaw Ver Aluminium Werke AgMethod of producing aluminum
US3244509 *Aug 22, 1961Apr 5, 1966M H HoepliHalide process for extraction of iron from iron-oxide-bearing materials
US3466169 *Dec 22, 1965Sep 9, 1969Halomet AgProcess for the production of metallic chlorides from substances containing metallic oxides
CH261162A * Title not available
FR1019261A * Title not available
GB734480A * Title not available
Referenced by
Citing PatentFiling datePublication dateApplicantTitle
US3937786 *Jun 20, 1974Feb 10, 1976Toth Aluminum CorporationProduction of aluminum chloride from raw materials containing coal
US3950485 *Jul 18, 1975Apr 13, 1976Toth Aluminum CorporationProduction of aluminum chloride from raw materials containing coal
US3989510 *Apr 12, 1974Nov 2, 1976Othmer Donald FManufacture of titanium chloride and metallic iron from titaniferous materials containing iron oxides
US4083923 *Jan 22, 1976Apr 11, 1978Toth Aluminum CorporationProcess for the production of aluminum chloride and related products
US4277446 *Jan 25, 1980Jul 7, 1981Reynolds Metals CompanyRemoval of iron and titanium minerals from aluminum bearing materials by chlorination and beneficiation
US4283371 *Jan 28, 1980Aug 11, 1981Kerr-Mcgee CorporationProcess for the production of aluminum chloride and related products
US4521384 *Sep 9, 1982Jun 4, 1985Kronos Titan - G.M.B.H.Process for the production of nearly aluminium chloride-free titanium tetrachloride from titaniferous raw materials containing aluminum compounds
US4536212 *Aug 17, 1983Aug 20, 1985David WestonProduction of a purified aluminum monochloride from alumina, bauxites and clays and the subsequent production of aluminum metal
US7347929Dec 31, 2003Mar 25, 2008Intercat, Inc.Gasoline sulfur reduction using hydrotalcite like compounds
US7361264May 25, 2005Apr 22, 2008Intercat, Inc.Mixed metal oxide additives
US7361319Nov 23, 2004Apr 22, 2008Intercat, Inc.Mixed metal oxide sorbents
US7494692 *Apr 3, 2008Feb 24, 2009Institute of Process Engineering, Chinese Academy of ScienceAntioxidation coating for steel and antioxidation method using the same
US8322046 *Mar 18, 2008Dec 4, 2012Zhaolin WangPowder formation by atmospheric spray-freeze drying
US20040086442 *Aug 13, 2003May 6, 2004Intercat, Inc.Flue gas treatments to reduce NOx and CO emissions
US20050121363 *Dec 31, 2003Jun 9, 2005Vierheilig Albert A.Gasoline sulfur reduction using hydrotalcite like compounds
US20060027485 *May 25, 2005Feb 9, 2006Vierheilig Albert AMixed metal oxide additives
US20080155853 *Mar 18, 2008Jul 3, 2008Zhaolin WangPowder formation by atmospheric spray-freeze drying
US20080233295 *Apr 3, 2008Sep 25, 2008Institute Of Process Engineering, Chinese Academy Of SciencesAntioxidation coating for steel and antioxidation method using the same
CN103572323A *Nov 8, 2013Feb 12, 2014中国科学院过程工程研究所Method for preparing aluminum silicon alloy through mixed chlorination and low-temperature electrolysis of aluminum-containing mineral and fly ash
CN103572323B *Nov 8, 2013Sep 30, 2015中国科学院过程工程研究所一种含铝矿物和粉煤灰混合氯化并低温电解制备铝硅合金的方法
DE2834165A1 *Aug 4, 1978Feb 14, 1980Holland GerhardVerfahren zur selektiven gewinnung mehrerer voneinander getrennter reiner halogenide und/oder halogenidgemische aus einem gemisch fester oxide
EP0107310A1 *Sep 8, 1983May 2, 1984David WestonProduction of a purified aluminum monochloride from alumina, bauxites and clays and the subsequent production of aluminum metal
WO2011021161A2 *Aug 18, 2010Feb 24, 2011Oron ZacharMethod and system for producing syngas
WO2011021161A3 *Aug 18, 2010May 12, 2011Oron ZacharMethod and system for producing syngas
U.S. Classification75/674, 423/149, 423/136, 423/76, 423/343, 423/135, 423/79
International ClassificationC22B21/00
Cooperative ClassificationC22B21/0046
European ClassificationC22B21/00F2