Extraction Processes for the Production of Aluminum, Titanium, Iron, Magnesium, and Oxygen for Nonterrestrial Sources


The suitability of existing terrestrial extractive metallurgical processes for the production of Al Ti, Fe, Mg, and O2 from nonterrestrial resources is examined from both thermodynamic and kinetic points of view. Carbochlorination of lunar anorthite concentrate in conjunction with Alcoa electrolysis process for Al; carbochlorination of lunar ilmenite concentrate followed by Ca reduction of TiO2 ; and subsequent reduction of Fe2 O3 by H2 for Ti and Fe, respectively, are suggested. Silicothermic reduction of olivine concentrate was found to be attractive for the extraction of Mg because of the technological knowhow of the process. Aluminothermic reduction of olivine is the other possible alternative for the production of magnesium.

The large quantities of carbon monoxide generated in the metal extraction processes can be used to recover carbon and oxygen by a combination of the following methods: (1) simple disproportionation of CO, (2) methanation of CO and electrolysis of H2 O, and (3) solid-state electrolysis of gas mixtures containing CO, CO2 , and H2 O. The research needed for the adoption of Earth-based extraction processes for lunar and asteroidal minerals is outlined.


The success of the space activity depends directly on the successful processing of nonterrestrial resources to obtain essential structural and life-support materials. A variety of structural materials is needed for the construction of solar power stations, mass-driver units, solar flare shelter and for the inside structures of the habitat. Aluminum was considered to be one of the essential metals needed for the construction of such space facilities (ref. 1). It was suggested that an extensive workshop must be provided as a fabrication facility so that many of the heavy components of a rolling mill, extrusion presses, casting beds and other equipment may be made at the space settlement rather than brought from Earth (ref. 1). It is, therefore, imperative that metals such as iron, titanium, and silicon, be available for steel manufacture. Finally, oxygen is one of the most essential elements needed for the life support.

The chemical and physical nature of lunar soils has been described by Levinson and Taylor (ref. 2). The lunar highlands consist mainly of anorthosite (ref. 3). Lunar rocks and fines also contain high titanium content and most of the titanium is present as ilmenite (FeTiO3 ) (ref. 4). Based on the analysis of some meteoroids it is expected that asteroids may contain considerable amounts of olivine (T. Bunch, NASA-Ames Research Center, private communication). Chemical compositions of these nonterrestrial resources reveal that aluminum, iron, titanium, and magnesium can be recovered in space. The possibility of manufacturing these elements in space has also been considered in previous studies on the utilization of nonterrestrial resources (refs. 1, 4, 5). In this study, the suitability of existing terrestrial extractive metallurgical processes for the production of Al, Ti, Fe, Mg, and O2 in space is examined from both the thermodynamic and kinetic points of view. However, the processes for extracting these elements from the nonterrestrial sources must be different from those conventionally used on Earth for the following reasons:

  1. It is estimated that recovery of these elements is more economical in space than transporting them from Earth. The conventional. hydrometallurgical operations are therefore not feasible because of the lark of water and hence suitable pyrometallurgical operations must be optimized.
  2. Because of the load constraints of the space shuttle, only elements such as hydrogen, carbon, and chlorine, are suitable reductants. These reductants are to be carried from Earth.
  3. There is a definite need for a process whereby the reactant carried from' Earth can be recycled indefinitely Thus, the reduction processes must be carefully considered to avoid irrecoverable losses of reductants.
  4. The reduction processes must be specially suited for the conditions of space, viz., low gravity and high vacuum.
  5. Since the source of oxygen is mainly the oxide minerals, the effluent gases from the reduction processes must be reprocessed to yield oxygen.
These constraints pose enormous problems and demand new processes. This paper evaluates suitable beneficiation processes for lunar and asteroidal minerals and indicates research efforts that must be extended to ascertain the feasibility of these processes. Some of them are based on thermodynamic evaluations and appear to be feasible, but no relevant industrial information exists Such processes should be further investigated experimentally to identify optimal process conditions.

ALUMINUM The main source of aluminum in the lunar soil is the plagioclase concentrate which contains mainly anorthite, (Ca,Na)(AI,Si)4 O8 - On Earth, aluminum is not normally produced from anorthite, although such resources are being seriously considered because of depleted supplies of bauxite ores (ref. 6). The chemical composition of anorthite is compared with the usual clays and bauxite resources of Earth in table 1. The major difference between anorthite and other minerals is the calcium content, which might prove to be a considerable problem in the extraction of aluminum.

Percent dry basisAluminaBauxiteClay Anorthite
Al2 O3 10040-584034
Si2 --9-254445
Fe2O3 --0.5-30.51
Ignition loss--14.2814--

The only processes known on Earth for the beneficia-tion of anorthite is the soda-lime sintering process (developed by the Bureau of Mines), which was tested both in the laboratory and by pilot plant operation (ref. 7). The general reaction of this process is represented by:


CaAl2Si2O8(s) = 3CaCo3(s) + Na2CO3 (s)

2NaAlO2(s) = 2Ca2SiO4(s) + 4CO2 (g)

The reaction products are treated to obtain alumina.

In the lime-soda sintering process, it is evident from reaction (1) that large amounts of calcium and sodium carbonates are required. Each mole of anorthite consumes 3 moles of CaCO3 and 1 mole of Na2CO3, which are not readily available on the Moon. Additionally, large amounts of water are needed for various leaching operations and water is very limited on the Moon. All these constraints lead one to search for other suitable processes.

The major processes considered are: carbothermic reduction, carbochlorination, aluminum monohalide process, and electrolysis. Out of these processes, carbochlorination, followed by the electrolysis of chlorides, was considered most attractive since a great deal is Carbothermic Reduction of Anorthite

In carbothermic reduction, the oxide reacts with carbon to produce metal and carbon monoxide:


MxOy (s) + yC(s) xM(s,) + yCO(g)

Since all constituent oxides of anorthite are highly stable, the reduction must be carried out at elevated temperatures (~2500 K). At these temperatures, other compounds such as SiC, Al4C3 , Al4 O4C in the condensed phase, and Al2O, SiO, Al, Si, and Ca in the gaseous phase are invariably present. Because of these phases, a reaction of the type (2) is very complicated. Most recently, Grjotheim et al. (ref. 8) explained the complications in the carbothermic reduction of alumina and reviewed the aluminum-producing reactions. Metal losses in the carbothermic reduction do occur because of the substantial vapor pressures of Al2O and Al. Gitlesen et al. (ref. 9). calculated the partial pressures of Al, Al2O and CO in the-system Al4O4C - Al4C3 - Al, where Al4O4C and Al4C3 are the condensed phases formed during the carbothermic reduction of alumina. It is fairly conclusive from their calculations that the aluminum and aluminum-suboxide pressures are, so high below 1900o C that liquid aluminum cannot be formed at normal operating pressures. Any liquid aluminum added to the system would be consumed by the reactions


Al4O4C(S) + 2Al () 3Al2O(g) + CO(g)


Al( ) - Al(g)

to maintain the equilibrium compositions of the gas phase. Above 2100o C, the solid aluminum carbide decomposes to aluminum and carbon. Also, carbon-monoxide partial pressure increases rapidly with respect to Al and Al2O, favoring the formation of aluminum. However, metal losses through Al(g) and Al2O(g) would still be significant unless special conditions are maintained.

Stroup (ref. 10) explained the special conditions for the carbothermic smelting of aluminum. In this process, the presence of a layer of molten aluminum saturated with carbide floating on a layer of molten oxycarbide and oxide in the electric furnace was taken to advantage. Both layers were electrically conductive, so the process was operated in a submerged arc or resistance furnace. Both the electrode and the raw charge pass through the metallic aluminum layer in operation. In the present process, the aluminum layer is allowed to saturate with Al4C3 at 2000o C, which reduces the activity of aluminum. and maintains the temperature high enough to hinder back-reaction with carbon monoxide. The reaction between alumina and aluminum is also minimized when alumina is converted to oxycarbide Striplin and Kelley (ref. 11) studied the feasibility of forming aluminum-silicon alloys by electrothermal reduction of clays and coke. To produce this alloy, the clay composition should contain at least 30 percent SiO2. Although this process appears to be feasible provided the conditions are carefully chosen, several unexplained problems make the process unattractive:

  1. The extent of reduction of Ca, Mg, Ti, Mn, and Cr and their distribution in the slag and vapor phase is not well established.
  2. The formation of carbides, oxycarbides, and suboxides complicates the reaction mechanism which requires detailed study of the proper temperature and proportion of carbon to-silicate ratio in the feedstocks.
  3. Pilot plant operation with alumina and silica showed that the formation of carbides and oxycarbides made it impossible to obtain a melt of Al and Si below 2000oC. Even at 2300oC, the extent of metal losses due to promotion of products into vapor phase is not well established. The metal losses are promoted when the process operates under high vacuum conditions existing in space.
  4. Anorthite contains significant amounts of CaO, which further complicates the process. At 2000oC, CaO is also reduced, yielding a Ca-Al-Si alloy that must be further purified.
  5. Separation of CaO before the reduction process increases the amount of capital investment and requires large amounts of materials not readily available on the lunar surface.
  6. Despite several years of research, no plant for producing aluminum through carbothermic reduction exists.
Carbochlorination Methods

In view of the potential problems anticipated in the carbothermic reduction of anorthite to produce aluminum, a low-temperature carbochlorination with Cl2 followed by an electrolysis of the aluminum chloride produced appears to be more promising. A carbochlorination method using AlCl3 is considered first.

High-temperature carbochlorination with AlCl3- A variation of carbochlorination recently tested by Peacy and Grimshaw (ref. 12) was based on an original proposal by Mitterbiller and Fruhwein (ref. 13). In this )process, aluminum mono chloride produced by high temperature reduction of bauxite with carbon and AIC s shock-quenched to obtain Al containing sm mounts of Si. The process can be essentially represented by:


Al2O3(s) + 3C(s)

+ AlCl3(g)-T=1800oC 3AlCl(g)

+ 3CO(g)- Shock Quench 2Al() + AlCl3(g)

If this process is used to extract Al from anorthite, many problems will arise. Because of the reduced activity of alumina, much higher temperatures are needed to obtain a reasonable yield of AlCl. This high-temperature process would pose severe problems such as corrosion of the reactor material compared to a low-temperature carbochlorination process. Further, substantial amounts of SiCl2 and CaCl2 will be present in the vapor phase from .analogous reactions with SiO2 and CaO in anorthite, resulting in an aluminum alloy that contains significant amounts of Ca and Si with some minor impurities. An additional purification step would then be required to recover the individual metals.

This process has been demonstrated only on a bench scale with bauxite. The only other attempt with a nonbauxite alumina ore is a single test (conducted at U.S. Bureau of Mines, ref. 14) in which carbochlorination of clay with AlCl3 in the presence of lampblack and sugar did not yield any metal at 1300o C. This process is untested and also is not economically feasible compared to a low-temperature carbo chlorination process followed by low-temperature electrolysis. Hence, the other alternative - low-temperature carbochlorination with C12 is a better method for extracting aluminum from lunar anorthite.

Low-termperature carbochlorination with Cl2 - This method is essentially the same as the first step in Toth's process (ref. 7) in which AlCl3 obtained from a lowtemperature carbochlorination of ores that contain alumina is reduced to aluminum by manganese at about 250o C. The Mn content in Al obtained usually exceeds the theoretical value by 1 weight percent. Because of various technical complications (i.e., a very slow reduction of AlCl3 by Mn), AlCl3 obtained from the first step of Toth's process will be used to extract Al by an electrolysis route.

Previous studies of low-temperature carbochlorination of ores that contain alumina have been mainly confined to bauxite and clays which are essentially CaO-free. Anorthite contains significant amounts of CaO, and any chlorination process for anorthite must consider the formation of highly stable CaCl2 . Fortunately, CaCl2 is a very useful by-product of the carbochlorination of anorthite. The various steps in the chlorination process are illustrated in figure 1.

The following reaction equilibria are to be considered in the thermodynamic analysis of the process.


Al2O3 (s) + 3C(s) + 3Cl2(g) 2AlCl3 (g) + 3CO(g)


CaO(s) + C(s) + Cl2(g) CaCl2(s) + CO(g)


SiO2(s)+2C(s)+2Cl2(g) SiCl4(9)+2CO(g)

The net reaction is


CaO-Al2O3-2SiO2(s)+8C(s)+8Cl2(g) CaCl2(s)

+ 2AlC13(g) + 2SiCl4(g) + 8CO(g)

In the original Toth process of CaO-free ores containing alumina, the optimum temperature for carbochlorination was 975o C. However, for chlorination of anorthite, a lower operating temperature is required because CaCl2is liquid above 772oC, which introduces problems in operation. An optimum temperature between 675o-770o C is preferable for anorthite chlorination.

In addition to the above mentioned equilibria, chlorination reactions involving some of the alkali, alkaline earth oxides, as well as FeO, MgO, and TiO2, must be considered since most of these are present as impurities in lunar plagioclase concentrate. Most of the resulting chlorides are volatile and are condensed along with AlCl3. Thermodynamic calculations reveal that the vapor phase contains very little CaCl2 or complexes of AlCl3-CaCl2-FeCl2. Since the recovery Of SiO2 is important for many glass-formation processes, the chlorination Of SlO2in anorthite can be prevented if the required amounts Of SiCl4 are used along with Cl2. AlCl3in the gas phase can be recovered by successive condensations to remove other volatile chlorides. Initially, FeCl3 in the gas phase is condensed to a solid by cooling to about 225o C in a heat exchanger/condenser. AlCl2 and chlorides other than SiCl4 and TiCl4 are condensed in a second stage at 90o C, while the latter are recovered in a subsequent stage at -300 C. The reactor residue is heated above 775o C to liquify CaCl2, which can be separated from SiO2 -rich residue by centrifuging. The estimated composition of the residue should contain at least 75-82 weight percent SiO2 and no more than 5-10 weight percent Al232 and 10-15 weight percent CaO. The exact composition of the residue should be determined from actual experiments. Pure CaO can be obtained by hydrolysis of CaCl2 followed by calcination. The chlorine and water can be recycled further.

Pertinent thermodynamic calculations along with heat and mass balance are given in the appendix (see tables 4-6).

Alcoa Electrolysis Process

The electrowinning of aluminum from the electrolysis of AlCl3 dissolved in a mixture of alkali and/or alkaline earth chlorides was attempted as early as 1854. These various early attempts are summarized in reference 8.

The most successful process to date is the one developed by Alcoa (at a cost of $25 million and 15 years of research). Since most of the problems encountered in electrolysis were solved by Alcoa and a plant producing 15,000 tons/yr is in operation, we selected this method for producing Al from the AlCl3 obtained by carbochlorination. The Alcoa process has some distinct advantages over the classical Hall-Heroult process for electrowinning of aluminum: the working temperatures are substantially lower, relatively high current densities are used, carbon anodes are not consumed, and a much smaller plant area is required. The electrical energy consumed is claimed to be only 70 percent of that used in the Hall-Heroult process, mainly because of the higher electrical conductivity of the electrolyte and the small interpolar distance of Alcoa cells.

Because the technical details of the process are proprietary, only a brief outline and a discussion are given here. Figure 2 shows typical electrolyte cells. The feed consists of 3-10 weight percent of purified AlCl3 along with the required amounts of alkali and alkaline earth chlorides. Electrolysis is performed under inert conditions in a sealed cell consisting of 20 to 30 bipolar carbon electrodes stacked vertically at an interpolar distance of 1 cm. Each bipolar electrode behaves as a cathode on its top surface and as an anode on its bottom surface. During normal operation, all electrodes remain immersed in the electrolyte at an operating temperature of 700o 30o C. A current density of 0.8-2.3 A/cm2 md a single-cell voltage of 2.7 V are typical operating conditions in the Alcoa process. Two typical compositions of the electrolyte (in weight percent) are: AlCl3(5), NaCl(53), LiCl(40), MgCl2(0.5), KCl(O.5), and CaCl2(1); and AlCl3(5 2), NaCl(53), and LiCl(42 2). The aluminum chloride concentration must be carefully controlled to ensure trouble-free operation. The AlCl3 concentration must be above a certain limit to prevent the discharge of alkali ions, which are detrimental to the electrodes as they form interchelation compounds with graphite. An upper limit is set by the high pressure of AlCl3. Also, the oxide content of the electrolyte must be as low as possible to prevent significant anode consumption with an attendant formation of an insoluble oxychloride sludge. An operating life of nearly 3 years is claimed for the electrodes when the oxide content of the electrolyte remains below 0.03 weight percent. The energy consumption of the cell is 9 kWhr/kg of aluminum produced.

A useful feature of the cell is that the chlorine generated at the anode sweeps the aluminum away from the cathodes and enhances the coalescence of Al droplets. The pumping effect of the chlorine bubbles maintains a continuous electrolytic flow across the cell, thus preventing the formation of Al pools on the electrodes. Chlorine collects at the top of the cell while molten aluminum falls countercurrently to the chlorine gas into a graphite compartment at the bottom of the cell. Liquid aluminum is passed through a filter to separate it from any excess electrolyte. The electrolyte is recycled to the cell; chlorine gas is also recycled into the carbochlorination step.


The main source of titanium and iron for nonterrestrial extraction is the ilmenite concentrate obtained from the lunar soil. The lunar ilmenite is of fairly high grade when compared to the terrestrial deposits. The chemical composition of lunar ilmenite concentrate is shown in table 2.

Titanium is commercially produced on Earth from ilmenite by a chlorination process and was described in an earlier study (ref. 5). In this process, iron. is separated from TiO2 by leaching with sulfuric acid after reducing ilmenite with coke or hydrogen at high temperatures. The residue (TiO2) is chlorinated in the presence of carbon to form TiCl4 which is further reduced by magnesium to yield titanium sponge. This process eliminates leaching at the final stages so that the dissolved hydrogen and oxygen content is minimum in the titanium. This helps to sustain the mechanical properties of titanium as well as to get pigment grade titania. While this is the most accepted process on Earth, and can be adapted for space operation; other alternatives are listed below.

NaOH Treatment

The ilmenite can be treated with molten NaOH according to the process patented by Hoekje and Kearley (ref. 15). During this treatment, TiO2 is preferentially dissolved in NaOH leaving Fe2O3 which is insoluble in NaOH. The undissolved Fe2O2 is separated and the liquid is leached with hot water. This process also removes impurities such as aluminum and chromium. Since the leaching process uses water instead of acid, the process is easier to perform. However, this process has been tested only on a bench scale.

Carbochlorination Process

The other promising process appears to be the one patented by Daubenspeck and Schmidt (ref. 16)(fig. 3). The ilinenite is treated with CO at 650o-1000o C to reduce Fe(III) to Fe(II) (this step may be unnecessary for lunar material). In this process the ilmenite is added at a rate of 5 kg/hr to a tube (4 in. in diameter X 48 in. long) containing 90-95 weight percent TiO2 and just enough carbon to reduce FeO. It is then chlorinated at 800o C in a fluidized bed. The iron is thus selectively chlorinated yielding FeCl2 (g), while TiO2 remains in the bed. Ferric chloride is condensed and further reacted with oxygen in another fluidized bed at 300o-350o C to produce Fe2O2.

Titanium extraction- The reactor residue which is essentially pure TiO2 can be used to extract titanium. A process reported by Chertien and Wyss (ref. 17) was chosen. In this process, a mixture of powders of calcium metal and TiO2 is pelletized at 71,000 lb/in2 and heated for 2 h at 925o-950o C. The titanium produced can be recovered by preferentially leaching CaO with a weak-acid wash solution. The main advantages of the process are that the leaching is minimal and all reagents can be recovered and recycled. The calcium metal can be obtained by processing calcium chloride, a by-product in the extraction of aluminum from anorthite.

The production rate of titanium can be enhanced by scaling up the process to handle 69,000 tons of ilmenite per year. Alternatively, several reaction tubes can be employed to process the required ilmenite. Some TiO2 have to be brought to the space manufacturing facility (SMF) for the initial charge of the first tube. TiO2, produced subsequently can be used to charge the other tubes.

The processes described in this report involve minimum leaching operations but at the final stages so that some hydrogen and oxygen might dissolve in titanium. Titanium thus produced either by this method or by the process described earlier (ref. 5) must be melted to obtain an ingot. Because of the high vacuum and very low oxygen activity in space, it is expected that the dissolved oxygen and hydrogen will be partly removed during vacuum melting operation.

Iron extraction- The ferric chloride obtained from the carbochlorination of ilmenite can be treated further to extract iron. It can be reduced to metallic iron directly by H2 at about 700o C. Hydrogen chloride is a useful byproduct and can be used in the chlorination process. Alternatively, FeCl3 can be oxidized in a fluidized bed at 300o-350o C to produce Fe2O2, which can be reduced with either C or H2 below 1000o C to obtain low-carbon steel or iron. In either case, the by-products CO or H2O can be used to recover oxygen.


Olivine is the main source of magnesium. Olivine is scarce in the lunar samples and is found in the gangue material left after anorthite is extracted. The gangue, a mixture of olivine and several other impurities (such as faylite), does not appear to be a suitable raw material. The best nonterrestrial source of olivine would be from an asteroid. The composition of olivine concentrate from asteroidal sources is given in table 2 (T. Brunch, NASA-Ames Research Center, private communication).

Enrichment factor


Enrichment factor
TiO2 3.7
FeO 2.2


aNinety percent mineral and 10 percent residue
bLunar source
cAsteroidal source (based on the analysis of Meteoroids (T. Brunch, NASA-AMes Research Center, private communication)). Selection of the Process

Potential reducing agents for olivine reduction are aluminum, calcium carbide, and silicon. Aluminothermic reduction of olivine is interesting from a chemical point of view but does not appear to be economically attractive when compared to the other processes. Of all the processes given, silicothermic reduction is technologically important and economically attractive. Although the other processes are discussed, the silicothermic reduction as shown in figure 4 is recommended.

Aluminothermic Reduction of Olivine

The reduction of olivine by aluminum in the temperature range 943o - 1150o C was studied by Grjotheim et al, (ref.18). They found that olivine and aluminum react according to the scheme:

3Mg2SiO4 (s) + 4Al() 2MgAl2O4(s)

+4MgO(s)+3Si(s) (10)

This was confirmed by the fact that, for a mole ratio of 4:3 (aluminum:olivine), no magnesium was obtained even at 1000o C. However, at a ratio of 2:1 and higher, an Al-Si alloy results:

3Mg2SiO + 6Al(s) 2MgAl2O4 (s)

+ 4MgO(s) + 2AlSi( ) + Si(s) (11)

This alloy in turn reacts with magnesium oxide according to the equilibrium:

4MgO(s) + 2AlSi() MgAl2O4(s) + 3Mg(g) + 2Si(s)


Although this process is attractive from a chemical viewpoint, magnesium cannot be quantitatively recovered because of the formation of MgAl2O4.

Calcium Carbide Reduction of Olivine Thermal reduction of olivine by calcium carbide was studied by Grjotheim et al. (ref. 19). From a technical standpoint, calcium carbide is less attractive compared to ferrosilicon, but it was used as a reducing agent for magnesite before and during World War 11.
The overall reaction for the process is

Mg2SiO4(s) + 2CaC2(s) Ca2SiO2(s) + 4C(s) + 2Mg(g)


for a mixing ratio of 1:2 (olivine:calcium carbide). Reaction (13) does not represent a valid equilibrium scheme. Reaction (13) is obtained by combining reactions (14) and (15):

Mg2SiO4(s) + 2CaC2(s) 2MgO(s)

+ 2CaO(s) + SiC(s) + 3C(s) (14)

2MgO(s) + 2CaO(s) + SiC(s) Ca2SiO4 (s)

+ 2Mg(g) + C(s) (15)

The vapor pressure of magnesium calculated from reaction (15) was much lower than the measured pressure in the presence of excess calcium carbide (ref. 19). In the presence of excess calcium carbide, for a molar ratio of 1:4, calcium carbide is available as a reducing agent for magnesia after conversion; the following reactions take place:

Mg2SiO4(s) + 2CaO (s) Ca2SiO4(s) + 2MgO(s)


2MgO(s) + 2CaC2(S) 2CaO(s) + 4C(s) + 2Mg(g)


Calcium fluoride could be used as a catalyst to enhance the reaction rates.

In this process, one of the by-products is a highly stable silicon carbide. The recovery of carbon from a silicon carbide is rather a difficult operation. The process, therefore, is considered uneconomical unless there is a specific use for silicon carbide at the space manufacturing facility.

Silicothermic Reduction of Olivine

The production of magnesium through silicothermic reduction of dolomite is well known (refs. 20-23). The reducing agent used in industry is ferrosilicon containing usually about 75 to 80 percent Si. The basic reaction of this process is represented as:

2MgO(s) + 2CaO(s) + Si(s) - 2Mg(g) + Ca2SiO4 (s)


However, Ca-Si alloy is expected to form as an intermediate when the reaction temperature is above 1000o C:

4CaO(s) + (2n + 1)Si(s) 2CaSin(sol) + Ca2 SiO4 (s)


where n indicates the silicon-to-calcium ratio in the liquid alloy. The equilibrium reaction for the production of magnesium is then:

2CaO(s) + 2MgO(s) + Si(so1) ) 2Mg(g) + Ca2SiO4 (s)


For processing asteroidal resources, the raw material is olivine rather than dolomite. The corresponding reactions for olivine are:

Mg2SiO4(s) + 2CaO(s) Ca2SiO4(s) + 2MgO(s)


4CaO(s) + (2n + 1)Si(s) 2CaSi(sol) + Ca2SiO4(S)


2CaO(s) + 2MgO(s) + Si(sol) Ca2 Si04 + 2Mg(g)

The main reaction that yields magnesium (23) is the same as for dolomite (20) and hence the magnesium pressure for the two processes is identical. The reduction of olivine by silicon in the presence of calcium oxide has been studied by Grjotheim et al. (ref. 19). They found that the absolute magnitudes of the calculated and observed pressures are in fair agreement. They also found that the slope of log PMg, vs 1/T was in rather good agreement between the experimental data for olivine reduction and the various data for dolomite reduction. The establishment of magnesium pressure over the mixture of olivine, calcium oxide, and silicon was observed to be sluggish. The reaction rate was adequate when calcium fluoride of 5 weight percent was added as a catalyst.

Since the Pidgeon process has been used successfully to produce magnesium from dolomite commercially, the same technology could be profitably used for olivine reduction. Calcium oxide and silicon necessary for this process can be obtained from the by-products of anorthite processing.


The most basic requirement for a space habitat is the availability of oxygen, which must be extracted from the nonterrestrial mineral agglomerates such as lunar raw materials. Further, to maintain the life process in space, humans requite an atmosphere of acceptable composition and pressure. The desired composition of the atmosphere given below (ref. 24) is said to be the minimum pressure needed to meet the criteria for atmospheric safety:

T=20o + 5o C
Relative Humidity = 50 10 percent
GasO2N2 H2OCO2
kPa22.7 26.6 1.0 <0.4

The atmospheric pressure is maintained at about half that sea level on Earth (50.8 kPa).

In the following sections, various methods of recovering O2 and C from the by-products of metal extraction processes are examined.

Extraction of Oxygen, Water, and Methane From CO

As a by-product of the aforementioned metal extraction processes, large amounts of carbon monoxide are generated which can be profitably used to recover solid carbon and to generate O2, H2O, and CH4. For such recovery, three interdependent processes are envisioned (as illustrated in fig. 5). One simple method is to generate CO2 and deposit carbon by the disproportionation of CO on a suitable catalyst that uses the well-known Boudouard reaction:

2CO(g)-catalyst CO2(g) + C(s)


Molecular oxygen and carbohydrates can be further obtained by photosynthesizing CO2 - On the other hand, CO can be reduced with H2 in a catalytic reactor to produce water and methane:

CO(g) + 3H2 (g) - CH4(g) + H2O(g)


This reaction has been used widely for many years in the petrochemical industry. Water is removed in a condenser and electrolyzed in one of several industrial electrolysis cells to obtain hydrogen and oxygen. Hydrogen can be recycled to the catalytic reactor for further reduction of CO.

Rosenberg et al. (ref. 25) investigated the methanation kinetics in the presence of nickel-on-kieselgator catalyst at 250o C. They outlined reaction conditions hat resulted in the quantitative conversion of CO to methane and water in a single pass without the formation of by-product carbon or CO. Certain factors must be considered carefully: catalytic poisoning by the impurities in the gas feed, flow rate of gases, catalyst bed depth, and precisely controlling the temperature of the bed. Rosenberg et al. (ref. 25) reported complete conversion of CO to methane even with a space velocity of 2000 hr-1at a H2/CO ratio of 4. At a lower H2/CO ratio, bed pressures up to 6 atm were required for higher conversion rates.

Recently, Cabrera et al. (ref. 26) compared the catalytic properties of cobalt, nickel, and lunar dust for the CO methanation reaction. For similar specific surface areas, the reaction rate constant was found to decrease in the order Co > Ni > lunar dust. Although Co and Ni catalyzed the reaction at temperatures near 200o C, detectable quantities of methane appeared only above 440o C with the lunar dust as the catalyst. At temperatures near 800o C, the catalytic activity of lunar dust is increased significantly. In view of the abundance and low cost of lunar dust compared to Co and Ni, further research should be conducted to correlate the catalytic activity with different lunar soil samples of varying grain sizes.

Another convenient method of recovering oxygen from the by-product gases is by solid-state electrolysis. This technique essentially involves the high-temperature electrolysis of the gas mixture consisting of CO, CO2, and H2O using a solid oxide electrolyte permeable to oxygen ions only. Since this recent technique has considerable potential, it is discussed in detail in the next section.

Solid-State Electrolysis for the Recovery of Oxygen and Carbon

Oxygen can be generated at the anode by the electrolysis of a gas mixture containing CO2, CO, and H2O using a solid electrolyte cell of that type:

Pt, O2 / Solid electrolyte / CO2, CO, H2O, Pt


The less stable gas species (e.g., CO2) will decompose first, followed by the more stable ones (e.g., CO). The half-cell reactions can be written:

At cathode:

2CO2(g) + 4e- 2CO(g) + 2O2-


2H2O(g) + 4e- 2H2 (g) + 202-


At anode:

202- O2(g) + 4e-


The net reactions are

2CO2(g) 2CO(g) + O2(g)


2H2O(g) 2H2 (g) + O2 (g)


The solid electrolyte can be a well-established one such as calcia-stabilized zirconia or yttria-stabilized thoria. On the other hand, CeO2 doped with 5 percent Y2O3 might be a better choice since recent investigations (refs. refs. 27, 28) have shown that it has a much higher conductivity than CaO-stabilized zirconia and remains purely an ionic conductor down to 10-13 atm oxygen potential at 600o C.

Carbon and hydrogen can be removed if electrolysis is performed to remove all oxygen from the gas mixture. However, to completely remove oxygen, the oxygen potential at the cathode must be reduced below the electrolyte conduction domain of CeO2 materials. It is proposed, therefore, to carry out the electrolysis in two steps: first, the oxygen potential in the gases is reduced by decomposing the less stable oxygen-bearing species using Ce02-based electrolytes; in the second stage of electrolysis, Y2O3-stabilized thoria can be used to completely remove the oxygen.

The following characteristics of the solid electrolyte system for oxygen reclamation are important in any solid-state electrolysis study:

  1. Stable gas-tight high-temperature seals
  2. Electrolyte stability in terms of vaporization and degradation
  3. Stable long-life electrodes
  4. Compatibility between the coefficients of thermal expansion of the electrolyte and the cell structure
The only oxygen reclamation investigation based on he solid-state electrolysis is- that of Weissbart et al. (ref. 29). They used calcia-stabilized zirconia as the elecrolyte and conducted long-term electrolysis experiTientswith CO2 and CO2-H2O gas mixtures. They found that small amounts of H2O are needed to catalyze the reaction. For instance, the yield of oxygen with pure CO2 is only one-fourth that obtained with CO2 and a mall amount of water vapor. Typically, the oxygen yield was 1 mole/min/W. Based on the results of extended tests, the following conservative design parameters for a one-man operated prototype CO2 electrolyzer were suggested by Weissbart et al. (ref. 29):

  1. Operating temperature of 850o C
  2. Current density of 100 mA/cm2
  3. O2 Faradaic efficiency 100 percent
  4. Electrolysis power, 250 W
  5. Electrolysis power efficiency, 50 percent
  6. Conversion Of CO2 to CO and O2 of 35 percent per pass

At 850o C their cell ran for 200 hr at 100 percent current efficiency without apparent degradation. They found that the glass-ceramic sealing was effective up to 1000 hr of electrolysis. Precious-metal seals and porous Pt electrodes appear to be reliable for runs of 2000 hr and probably much longer. Further experiments are needed to resolve some of the anticipated problems in large-scale, solid-state electrolysis for oxygen recovery. For instance, the cell operation at high temperatures poses certain problems concerning materials, sealants, cell structure, power to maintain high temperatures, increased cell weight, and volume required for "insulation."

On the other hand, the oxygen recovery by solid-state electrolysis offers substantial advantages over conventional methods:

  1. Since the electrolyte permits the transfer of oxygen ions only, the separation of oxygen from the gas mixture in a sealed cell is excellent.
  2. The problem of gas-liquid phase separation at zero gravity is not encountered since there are no liquids in the electrolyte cell.
  3. Unlike the fused salt electrolysis, only a few construction materials are needed for the cell since solid electrolyte is noncorrosive.
  4. The continuous electrolysis of CO2 or H2O or a mixture of CO, CO2, and H2O is possible.

Some of the processes described here have not been tested thoroughly, even for terrestrial applications on a laboratory scale. A few processes, such as Al electrolysis and methanation of CO, have proved to be commercially feasible and plants based on such processes are in operation. Although a partial list of problems likely to be encountered in various processes is given under respective sections, general and specific research areas are outlined here.

Metal Extraction Processes

Before any experimental work is undertaken, the various experimental factors, phase relations, and volatilization reactions must be identified. Once these factors are identified, a systematic method of representing the various reaction equilibria must be devised. Stability field diagrams in which the chemical potentials of the system components are represented as a function of temperature or of a specific species offer one such approach. Hence, as a first step in the experimental process, it is proposed to construct such diagrams for various tenary systems (M1-M2-X1 and M-X1-X2) and quaternary systems (M1-M2-X1-X2 , M-X1-X2-X3) Of interest in the extraction processes of Al, Mg, Si, and Fe; M denotes a metal (Ca, Fe, Al, Si, or Mg) while X denotes a nonmetal (O2, C12, or C. Some of the experimental data needed to construct such diagrams are already available in the literature.

In view of the importance of rate processes in the extraction of metals, the next logical step would be to undertake a systematic and thorough investigation of kinetics and mass-transfer situations. By performing batch segregation tests, the effects of various factors such as reagent additions, particle size, and temperature can be investigated. Based on these results for a single process, a kinetic model can be set up which, in turn, can be used to predict process yields from limited data gathered by different processes.

Carbochlorination of anorthite- There have been few, if any, studies done, either thermodynamic or kinetic, on the low-temperature carbochlorination of anorthite. The first step in the investigation is the synthesis of plagioclase concentrate (90 percent anorthite + 10 percent rock) with its bulk composition closely identical to that of lunar material. Experimental investigations must be performed with the following objectives in mind:

  1. Analyze vapor phase, preferably by a mass specrometer modified for high-pressure sampling
  2. Analyze the reactor residue
  3. Determine the extent of deviation from thermodynamically calculated values for both condensed and vapor phases
  4. Adjust feed rates of various reactants to the fluid bed reactor to selectively remove A12O3 as AlCl3
  5. Separate CaCl2 and SiO2 from the reactor residue and determine the composition of the residue
  6. Evaluate chemical engineering aspects of the process as related to lunar or extraterrestrial conditions.
  7. Reinforce initial bench scale investigations with pilot plant studies

Alcoa electrolysis process- Since most of the technical aspects of this process for extracting Al from AlCl3 have already been worked out, further studies should be devoted to finding an electrolyte composition more suited to lunar conditions. The aluminum chloride obtained from the carbochlorination of lunar anorthite will contain some alkali and alkaline earth chlorides. Instead of purifying the chlorides, an electrolyte composition compatible with these chlorides must be obtained. This new composition might be obtained by only slightly modifying the electrolyte composition recommended in the Alcoa process for terrestrial conditions.

Extraction of Ti and Fe from ilmenite- The processes for extracting Ti and Fe from ilmenite have already been worked out on a laboratory scale. The processes must be tested on a commercial scale. As in the Al extraction process, the first step in extracting Ti and Fe is the chlorination of ilmenite in a fluidized bed reactor. The kinetic model set up for anorthite chlorination might be useful in evaluating the process yields based on the limited data concerning ilmenite chlorination.

Extraction of Mg from Olivine- Of all the pyrometallurgical processes for the extraction of magnesium from olivine, silicothermic reduction is the well-investigated process. However, the role of calcium silicide as an intermediate in the overall reduction process needs to be studied.

The following areas need to be investigated for choosing the other alternatives:

  1. Kinetics of olivine reduction with aluminum as the reducing agents, and
  2. The optimum ratios of reducing agent (Al) needed for the reduction processes.
Oxygen and Carbon Recovery

The kinetics, engineering design, and feasibility of continuous recovery of carbon and oxygen, vital for extraterrestrial activities, have not been thoroughly investigated. Some potential areas of research in the two proposed processes for the reclamation of oxygen and carbon from by-products of metal extraction processes are outlined here.

Methanation of CO- Further investigations are needed to clarify the chemical role of lunar and interstellar dust as a catalyst for the methanation reaction of CO and H2 to produce CH4 and H2O If the catalytic activity is proportional to the radiation damage sustained by lunar dust in addition to their higher surface area, fine-grained particles might be a better catalyst because they are exposed to radiation longer than coarse-grained particles. The necessary submicron-sized lunar dust, because of its persistent internal polarization, can be effectively separated by electrostatic purification. Strong research efforts should be undertaken to verify the grain-size correlation of the catalytic effect, which, if proven correct, would lead to a very inexpensive way of catalyzing methanation without resorting to currently used industrial catalysts such as cobalt and nickel.

Solid-state electrolysis- This recent area of research has potential for the recovery of oxygen and carbon from the effluent gases of metal extraction processes. In solid-state electrolysis, the power requirements become excessive as the operating temperature is lowered because of higher cell resistance at lower temperatures. Hence, it is desirable that the cell temperatures be decreased as far as possible, consistent with high Faradiac efficiency. In view of the recent synthesis of solid electrolytes, such as Y2O3 doped CeO2 with high ionic conductivities compared to the conventional ZrO2-CaO and ThO2-Y2O3 electrolytes, it might be possible to operate the cells at lower temperatures with considerable savings in energy. Hence, the experiments should be conducted with the following objectives in mind:

  1. To test the feasibility of CeO2-Y2O3 electrolyte
  2. To establish the optimum conditions for long life of the cells
  3. To improve the existing design of the metal-ceramic seals
  4. To select an optimum temperature - the lowest temperature of cell operation consistent with acceptable IR loss in the electrolyte

Appendix Suggested Extraction Processes




Table of Contents