US2402193A - Thermal reduction of oxidic magnesium ore - Google Patents

Description

Patented June 18, 1946 THERMAL REDUCTION OF OXIDIO MAGNESIUM ORE Charles B. Wlllmore, New Kensington, Pa, as-

slg'nor to Aluminum Company of America, Pittsburgh, Pa, a corporation of Pennsylvania No Drawing. Application June 29, 1944,

Serial No. 542,825

8 Claims. 1

This invention relates to methods of recoverin metallic magnesium from its ores by heating a mixture of ore and reducing agent, with and without other substances, at temperatures at which the magnesium metal reduced from the ore is distilled from the reaction mass, and thereafter condensing the magnesium vapors to recover that metal in solid form.

Such processes as Just generally described are conventional and have taken many specific forms, none of which have been particularly successful commercially. While the thermal processes of directly reducing or distilling magnesium and recovering the vapor are inherently simpler than those processes which are based upon the electrolysis of magnesium chloride and by which most magnesium metal is, now produced, the

mai reduction operations utilizing these reduction agents in a continuous manner, the equipment necessarily involved in the upkeep of that equipment is very expensive and impractical.

simplicity is more apparent than real. The

fundamental difliculty with the thermal production methods or making magnesium is one of cost. The comparatively high cost of such processes arises primarily from the following factors:

First, the reduction agents heretofore proposed, which are otherwise generally operable, are relatively high in cost, except for carbon which has the inherent disability of producing a gaseous reaction product, carbon monoxide, which reacts with the magnesium vapor to again form magnesium oxide. No inexpensive and positive method oi treating the mixture of magnesium vapor and carbon monoxide produced by the carbon reduction has yet been developed with the result that the product of this reduction must be again treated in. an expensive and separate step to reduce its magnesium oxide content.

For these and other reasons magnesium .cannot, at present, be thermally reduced at prices which will commercially compete with magnesium produced by the electrolysis of magnesium salts, such as magnesium chloride.

The object of the present invention is to provide an improved process, the improvement consisting in the use or a novel reducing agent, by which the thermal reduction of magnesium from its ores may be carried out at costs lower than those heretofore encountered in thermal reduction processes. A further object is the provision of a new and novel method of reducing magnesium from its ores.

This invention provides a thermal reduction process so reduced in cost as to compete commercially under favorable conditions with present electro-chemical methods.

In the practice of the invention I employ as a reducing agent aluminum carbide and in the preferred and, under present conditions of the art, competitive form I employ a modification Second, thermalprocesses necessarily operate at high temperatures and require expensive apparatus, includingfurnaces, retorts. vacuum pumps and similar equipment and a high incidence of labor. Moreover, this equipment must be made of materials which withstand high temperatures. Such materials are expensive and deteriorate rapidly attics-temperatures used.

Third, the use of any reduction agents, even those such as aluminum, silicon, terrosilicon, calcium carbide, silicon carbide, silicides and others, which do not produce a gaseous product) of reaction which will seriously react withthe magnesium vapor. has to date been limited to-batch operations because of the diflicult-ies of conductthereof containing as active components aluminum carbide and aluminum. While this aluminum carbide-aluminum mixture could be made irom pure substances, such as pure carbon and pure alumina, considerations of cost and emciency are in favor of the use of a reducing agent which is itself the product of the reduction of aluminous ore with carbon. Thus the preferred reducing agent is made by mixing an aluminous ore, such as bauxite, with a commercial carbon,

such as ordinary coke, and reacting the mass at high temperature under circumstances somewhat different, and less expensive, than would be empioyed if it were the purpose of the process to reduce the aluminum oxide or the aluminous ore entirely to pure aluminum, and to prevent the a formation of aluminum carbides which are usuing the operation continuously, said diiilculties ally an unwanted material in commercial aluminum.

Despite the fact that much of the aluminum in this partially reduced mixture is in the form of aluminum carbide, the mixture is highly efiicient as a reducing agent when reacting with a magnesium oxideore, such as calcined dolomitic limestone 0r calcined magnesite. when this mixture is used' as the reducing agent, its ratio of e fllciency to cost is high, as compared to known reducing elements, as will herein be discussed and demonstrated.

be made in various specific ways and in various purities. One form that I have found suitable is that obtained by mixing low grade bauxite and coke and subjecting the mass to the heating of an arc with the result that there is produced a solid lump of reaction product surrounded by unreacted charge. After the furnace is cool enough to handle this lump, it may be readily separated from the loose granular unreacted charge and may be prepared ior use by a simple pulverizing process, since this reaction product tends to be friable and pulverant, an item which in itself considerably reduces the cost over the use of aluminum, which must undergo an expensive granulating or pulverizing step before it can be satisfactorily used as a reducing agent in connection with magnesium ores. Such reaction products may vary in composition but are composed principally of aluminum carbide and metallic aluminum, of-doubtful purity. A typical product contained the following:

Per cent Y 4 finely ground reactants are thoroughly mixed and, preferably, compressed into'pellets or briquettes. Better results are obtained if all of the reactants are of a size passing a standard 100 mesh screen, although such refinement is not necessary. The eflliciency of the reaction is increased by the addition to the mixture of fluorides or chlorides, as is likewise truein prior processes for the thermal reduction of magnesium ore, but in my improved process certain fluorides and chlorides and certain mixtures thereof are particularly emcacious as will, hereinafter, be. de-

scribed. Magnesium oxide or any ore containing magnesium oxide, such as dolomite or magnesite, may be used as the source of magnesium. .Where the magnesium is in the ore in the form of carbonate, it is preliminarily calcined to drive ofi carbon dioxide and thus reduce the carbonate to oxide. I include magnesium oxide, ores containing that oxide and calcined carbonate ores in the term oxidic magnesium ore. The temperature at which the reactionbest takes place is, of course, the lowest temperature consistent with Per cent good results, and it is a ieatureof my invention Alicfl 38-73 T1 3- u that where the mixture is heated in vacuo the A1 40-95 F9 process thereof may be operated eillciently at W ".""'i"" 237 temperatures as low as.l100 0., which temperig 2-2: G ature is lower than the eiilcient operating temall percentages being by weight; The impurities, such as iron, titanium. silicon and calcium, are the result of the original content of the impure aluminum ore and may even be helpful, and at least not harmful, when this product is used as a reducing agent for the reduction of manaslum. The actual cost of production of this product is an interesting feature in the economies which are obtained by the use of this material as a reducing agent. Both the aluminum carbide and the aluminum of the mixture are produced at low cost, the letter at about onehalf the kilowatt hours per Pound necessary to the production of commercial aluminum, and this single factor is quite apart from the other expensive steps in the production of commercial aluminum, such as benefilciation of ore to obtain a pure alumina, the use or carbon in the form of pure electrodes instead or in the commercial and cheap form of coke, etc., all 0! which are avoided. The cost of this material is equally interesting when compared with ferrosilicon, in which silicon is the reducing agent, the power consumed in the preparing of ferrosilicon being in the neighborhood of about one-fourth more per pound than the cost oi preparing the aluminum carbide-aluminum reduction product.

' In practicingthe process of this invention the peratures employed with other comparable proccan, however, be obtained at temperatures as low as 900' C. When the reaction mixture is heated at atmospheric pressures higher temperatures are necessarily employed for equivalent results. but temperatures above about 1500 C. are not generally desirable or emcient.

The reduction agent and the oxidic mangesium ore may be mixed in various proportions. I It is wasteful and inefficient to use ore in such amount that its magnesium oxide content is less than that required to react with the aluminum carbide and aluminum in the reduction agent. At least stoichiometric proportions of magnesium oxide and alumnum carbide and aluminum should be used, and I have found that when the macnesiumoxide is in excess the efllciency or the reaction is greatly increased.

Following in tabular form are the results obtained with various indicated reaction mixtures in the practice of the invention, all of the results being obtained by heating the dry briquetted mixture at indicated temperatures and in indicated vacuum for stated periods 0! time and condensing the metallic vapor distilled from the reaction mass. In each case the reducing agent used was a product obtained by the reduction of bauxite by coke, all as above set forth.

Components cinaotton mixture other than the Yield r reduction agent BelnionciM Oinmix- Approx. Tun Vacuum canto! a Run No. rare to or his sintsm mm. of available in minum mercury reaction Ore Other per cent by weight mixture 1 Celcined magnesiie l. M 3 0.16 so. 6 a do 1,100 3 0.16 88.1 a m 1,900 3 0.20 19.5 4. r n 1,100 3 0.2) 6-9. 6 s ..(.o 1,000 3 0.10 as. l 6.. r n 1. no a 0.10 s7. 4 i" are H38 3 til it;

----.C.0 1 g l 9-. ..c.o.. a m, 11100 a 0.20 ass 10 calointzgtdolomita MRO) 5 e OIF: l, 8 0.8) 86. 7

con il r n ..dc 1,100 3 0.36 33.2 12 --c.o... 5% Car: ..do i, 100 a 0.20 can 13 r n -.do.. About 26% one. M30. 1,100 3 0.20 92.1

esses using known reducing agents. Good results 5 Referring to the results shown in said tabulation, it will be noted that an excess of Met) in the reaction mixture is beneficial; that rising temperature produces more eillcient results (compare runs 1 and 2); that increase of time increases yield (compare runs 11 and 12) and that the use of fluorides and chlorides increases the efli'ciencyof the reactions (compare runs'l and 3; runs 10 and 11; runs 1 and 9; runs 1 and 8).

Moreover it will be noted that with the use of fluoride r chloride and with or without an excess MgO present, the results obtainable attemperatures of 1200 can be approximately duplicated by 1000 C. and exceeded at 1100 C. (compare runs land 7; runs 1 and 8; runs 1 and 9; runs 2 and runs 1 and 6).

It will likewise be noted that calcium chloride is more eflicient than magnesium fluoride or calcium fluoride or mixtures thereof but that a and depreciation of the relatively expensive equipment. The sum of such items is substantially fixed and must be apportioned over the amount of magnesium made in the furnace in a given time. Thus, for example, the cost of the aluminum carbide-aluminum reducing agent is, per pound, about equal to the cost of the well known-reducing agent calcium carbide but, per pound of reducing material, the aluminum carbide-aluminum is about three times as eflicient in reducing the oxidic magnesium ore, and thus when this invention is used, furnace equipment of given capacity produce about three times as much magnesium with consequent lowered cost.

A comparison between processes using ferrosilicon as the reducing agent and processes operated in accordance with this invention is contained in the following tabulation of results obtained in operations conducted under comparaday per furnace uni double salt of MgFaMgCli is equally eflicient. tive conditions:

Run

A B C Reducing agent Ferrosilioon containing 75% by Aluminum cerbidealuminum Aluminum carbide-aluminum weight oi silicon. conta 38.73% AM); and containing 38.73% AliC: and

40.95% A1 alance impurities. 40.95% A] balance impurities. M nesium ore, calcined Dolomite olomite Magnesite. Mg eontentoitheore 4 90%. Re tion MgO to reducing consti- Stoiqhiometric. tuent in mix.

Additional effective compound 5% CaCh. Temperature 0.. 1150 1100. Time oireaction hours 8. Percentage yield 01 total ava'lable Mg. 85.9. Pounds oi Mg produced per pound (km.

of charge. KwH required to produce suillcient 3.66.

ifiducing agent to reduce a pound 8- Relative magnesium production per 4.7.

The conclusions just stated on the basis of the tabulated results are representative of the comparative results which are obtained by varying the various factors in the practice of my invention. Thus in the preferred practice of my invention, and in the form which gives highest performance at lowest operating cost, the reaction should takepiace at the lowest temperature consistent with good results-usually about 1100 C. in the presence of an excess of ma nesium oxide over that required to combinewith the aluminum carbide and aluminum in the reduction agent-usually an excess of 10 to 50 per cent. by weight, and in the presence of about 2 to 15 per cent. by weight of the reaction mass of' at least one chloride or fluoride of the class consisting of magnesium and the alkaline earth metals and, preferably, in the presence of the of such items as labor. supervision, fuel, power chlorides and fluorides and, pref- From these figures it will be noted that the furnace unit used is two to four times as eiiicient when aluminum carbide-aluminum mixtures are used as the reducing agent. A feature of this comparison is the fact that such results are obtained at operating temperatures 50 centigrade lower. This lower temperature is significant because it is directly reflected in reduced deterioration of equipment and, consequently, reduced cost. Moreover it will be noted that these results are obtained with reduced power costs (in the present instance 12 to 23 per cent. less). While the amount of aluminum in the aluminum carbide-aluminum mixture is not critical, practical considerations limit the aluminum content to not more than about 85 per cent. of the total aluminum carbide-aluminum content of the mixture.

Having thus described and explained my invention, I claim:

1. In the process of producing magnesium by the thermal reduction of oxidic magnesium ore at temperatures above the melting point of magnesium, the improvement consisting in the use of a reducing agent containing aluminum carbide.

2. as the process of producing magnesium by the thermal reduction of oxidic magnesium ore at temperatures above the melting point of magnesium, the improvement consisting in the use as a reducing agent of the product of the partial reduction of aluminous ore with carbon, said product containing aluminum carbide and aluminum, the aluminum being not in excess of per cent. by weight'of thetotal content of eluminum carbide and aluminum.

3. In the process of producing magnesium by the thermal reduction of oxidic magnesium ore at temperatures above the melting point of masnesium, the improvement co in the use as a reducing agent oi the product 0! the partial reduction of aluminous are with carbon, said product containing-aluminum carbide and aluminum, the aluminum being not in excess of 85 per cent. by weight of aluminum carbide-aluminum content, and mixing said ore with said agent in such proportions that the magnesium oxide content is, present in excess 01' the amount theoretically necessary to react with all aluminum carbide and aluminum in said agent.

4. In the process of producing magnesium by the thermal reduction of oxidic magnesium ore at temperatures above the melting point of magnesium, the improvement consisting in the use as reducing agent of a mixture of aluminum carbide and aluminum and mixing with the reaction mass containing said ore and reducing agent, 2 to 15 per cent by weight, of the'total weight of said mass, of at least one compound selected from the class consisting of the fluorides and chlorides 0! magnesium and the alkaline earth metals,

5. In the process of'produclng magnesium by the thermal reduction oi oxidic magnesium ore at temperatures above the melting point of magnesium, the improvement consisting in the use as reducing agent of a mixture of aluminum carbide I amazes and aluminum and mixing with the reaction macs containing said ore and reducing agent, 2 to 15 per cent. by weight, of the total weight of said mass, of the compound M8Fa.MgCl2.

6. In the process of producing magnesium by the thermal reduction of oxidic magnesium ore at temperatures above the melting point of magne sium, the improvement consisting in the use as reducing agent of a. mixture of aluminum cubicle and aluminum and mixing with the reaction mass containing said ore and reducing agent, 2 to 15 a per cent. by weight, of the total weight of said mass, of the compound 09.012.

7. In the process at producing. magnesium by the thermal reduction oi. oxidic magnesium ore at temperatures above the melting point of ma5= nesium, the improvement consisting in the use as reducing agent of a mixture of alumin cmbide and aluminum and mixing with the reaction tent oi the charge is in excess of that theoretically required to react with all 0! the aluminum csrbide and aluminum in said charge.

one.