Previous Top

Appendix 5E:
LMF Chemical Processing Sector

Mining robots deliver raw lunar soil strip-mined from the pit to large input hoppers along the edge of the entry corridors into the chemical processing sector. The primary responsibility of the materials-processing subsystems is to accept lunar regolith, extract from it the necessary elemental and chemical substances required for system growth, replication, and production, and then return any wastes, unused materials, or slag to an output hopper to be transported back to the surrounding annular pit by mining robots for use as landfill.

It is possible to achieve qualitative materials closure (see sec. 5.3.6) - complete material self-sufficiency within the Lunar Manufacturing Facility (LMF) - by making certain that chemical processing machines are able to produce all of the 84 elements commonly used in industry in the United States and the global economy (Freitas, 1980). However, such a complete processing capability implies unacceptably long replication times T (on the order of 100-1000 years), because many of the elements are so rare in the lunar or asteroidal substrate that a vast quantity of raw soil must be processed to obtain even small amounts of them. By eliminating the need for many of these exotic elements in the SRS design, replication times can be cut by as much as three orders of magnitude with current or foreseeable materials processing technologies.

Hence, it is desirable to determine the minimum number of elements and process chemicals and to fix the lowest extraction ratio R (kg input material/kg useful output material, see sec. 5.3.6) which can still maintain closure of the system, thus minimizing the replication time T.

5E.1 Minimum LMF Requirements: Elements and Process Chemicals

The elemental and chemical requirements of the expanding LMF fall into a fairly small number of broad categories summarized in table 5.11. Note that these are the minimum (or very nearly so) requirements for LMF qualitative materials closure - an "adult" LMF entering production phase may need additional chemical processing capabilities which may be programmed into the factory's operational software. Table 5.11, however, lists only those minimum requirements necessary to achieve closure for a seed during the growth phase.

Table 5.11. - Minimum Seed Element And Process Chemical Requirements

I. Structural metals, alloys, hard parts, tubing, containers, etc. -- Fe, Al, Mg, Ti, Mn, Cr, C, Si, Ca

II. Building materials, insulation, fabrics, glass plate, ceramics, crucibles, furnace linings, chemistry glassware, high-temperature refractories, etc. - lunar soil as found (basalt when fused), anorthite (CaAl2Si2O8), silica (SiO2), alumina (Al2O3), magnesia (MgO), feldspar

III. High purity electronics-grade materials for the manufacture of solar cells, computer chips, etc. -- Si, O2, Al, P, B

IV. Magnetic materials - Fe

V. Fluorine chemistry containers - Fe, C, F2

VI. Process chemicals for bulk manufacturing, high-purity electronics chemical production - H2O, HF, N2, H3PO4, HNO3, SiH4, CF4 (Freon for microelectronic "dry etching" processes), NaOH, Cl2, H2SO4, CaCl2, Na2CO3, NH3

VII. Process minerals, inputs to chemical processing sectors - olivines, pyroxenes, feldspars, spinels, ilmenite, apatite, anorthite, tincalconite (anhydrous borax).

Total of 18 elements, 12 minerals/mineral types, and 11 additional process chemicals.

It will be argued that a chemical processing system capable of producing each of the above from raw lunar soil has achieved full self-sufficiency, or materials "closure."

Demonstration of materials closure plausibility. The components in table 5.11 were obtained first by taking a very basic list of necessary elements (the first four categories) for the entire LMF and adding to these any additional substances necessary to chemically produce the original items. This resulted in an increase in the number of items, therefore, all newly added items themselves had then to be similarly checked to ensure that each of them could be produced from the materials already at hand. This procedure was iterated until closure apparently was achieved. The list includes reagents necessary for the production of microelectronic circuitry (Oldham, 1977), even though "wet chemistry" may not be necessary for this application in space manufacturing (Zachary, 1981). The team is unaware of any significant omissions in table 5.11, which demonstrates essential qualitative closure.

5E.2 Derivation of Minimum Requirements: Qualitative Materials Closure

The lunar substrate from which the required substances are extracted or manufactured has a mean global mineral content as shown in table 5.12. Source minerals for boron do not appear in this list, nor do the sources for volatiles implanted by the solar wind. A summary of all elements found to date in the lunar regolith samples returned by Apollo and Luna missions may be found in table 4.1.

To plausibly demonstrate materials closure, it must be shown that every item on the requirements list can be derived from other items on the list and that all elements are derived from those found in the lunar regolith. To fully and rigorously demonstrate closure, a detailed element-by-element breakdown of the entire factory would be required, giving the mass of each element or process chemical required followed by a convincing demonstration that such quantities could indeed be produced using only the amounts of other substances known to be available and an input of lunar material. This latter set of conditions is called quantitative closure.

Preparation of process minerals. A comparison of the list of process chemicals in category VI in table 5.11 with the minerals found in lunar soil (table 5.12) suggests that it may be possible to use raw lunar soil as input to the materials processing extraction machines if these minerals require no beneficiation. In the event such beneficiation is needed to obtain the specific minerals in separated form, the electrophoretic separation technique described in section 4.2.2 may be used. This method involves placing finely divided powdered lunar dust in aqueous (or slag, or other solvent) suspension which has a solvent pH tuned to match the isoelectric potential of the desired mineral species. A cross voltage is applied and all minerals but the one desired migrate away, leaving behind a purified residue - in the present case, anorthite and the category VII (table 5.11) process minerals may be recovered. Preliminary testing of the electrophoretic separation concept with simulated lunar soil has been successful (Dunning and Snyder, 1981).

Table 5.12.-Minerals Typically Found In Lunar Regolith (From Williams And Judwick, 1980)

Major Minor
Olivine (Mg,Fe)2SiO4
Pyroxene (Ca,Mg,Fe)SiO3
Plagioclase feldspars (Ca,Na)Al2Si2O8
Spinels (Fe,Mg,Al,Cr,Ti)O4
Armalcolite (Fe2TiO5)
Silica (quartz, tridymite, cristobalite) SiO2
Iron Fe (variable amounts of Ni and Co)
Troilite FeS
Ilmenite FeTiO3
Phosphates Oxides
Apatitea Ca5(PO4)3(F,Cl)3
Whitlockitea Ca9(Mg,Fe)(PO4)7(F,Cl)
Rutile TiO2
Corundum (?) Al2O3
Hematite (?) Fe2O3
Magnetite Fe3O4
Goethite (?) FeO(OH)
Zr mineral Metals
Zircona ZrSiO4
Baddeleyite ZrO4
Copper (?) Cu
Brass (?)
Tin (?) Sn
Silicates Zr-rich mineral
Pyroxferroite (Fe,Mg,Ca)SiO3
Amphibole (Ca,Mg,Fe)(Si,Al)8O22F
Garnet (?)
Tranquilletyitea Fe8Zr2Ti3Si3O4
Zirkilite or zirconolitea CuZrTi2O7
Sulfides Meteoritic minerals
Mackinawite (Fe,Ni)9S8
Pentlandite (Fe,Ni)9S8
Cubanite CuFe2S3
Chalcopyrite CuFeS2
Sphalerite (Zn,Fe)S
Cohenite (Fe,Ni,Co),O
Niningerite (Mg,Fe,Mn)S
Lawrencite (?) (Fe,Ni)Cl2
aThese minerals are known to exhibit complex substitutions, particularly of elements as Y, Nb, Hf, U, and the rare earth elements that are concentrated in these minerals.

In addition, the electrophoretic technique may prove invaluable in separating out "trace minerals" from lunar soil, in particular apatite and possible differentiated boron-containing minerals which may exist in the lunar regolith.

Separation of iron. The magnetic properties of lunar soil are due almost entirely to the presence of metallic iron, which occurs in lunar soil as a free element in the amount of 0.5% by weight, roughly 5% of the total iron content of the lunar regolith. Since it is magnetic, metallic iron may be separated from the raw lunar substrate by straightforward electromagnetic techniques directly as the raw input material leaves the input hopper. This Fe will be fairly pure, containing only about 5% nickel and 0.2% cobalt (Phinney et al., 1977).

Structural metals and metal oxides. Of all the chemical materials processing options studied to date, the hydrofluoric (HF) acid leach technique appears to have the best potential for minimum operating mass, ease of element separations to high purity, and favorable energy and heat rejection requirements (Amold et al., 1981;Waldron et al., 1979). HF acid leach (Waldon et al., 1979), shown in figure 5.41 in flowsheet form, uses low-temperature hydrochemical steps to separate the silica content of the lunar raw material from metallic oxides in minerals by converting them to fluorides and fluorosilicates. The silica is then vaporized as SiF4, leaving Ca, Al, Fe, Mg, and Ti fluoro salts to be separated by a variety of solution, precipitation, ion exchange, and electrolytic steps. These are then reduced to the pure metallic form with sodium metal, which is recycled. (HF is added as a major process chemical.)

Sodium for the reduction of metals and silicon may be obtained by a modified Castner cell process, which involves the electrolysis of molten NaOH to produce Na, O2, and H2. Iron electrodes can be used in this application. (NaOH must also be added to the process chemicals list.) Metal oxides and silicon dioxide can be obtained, where needed as ceramics, refractories, or for glasses, by hydrolysis of the fluoride or fluorosilicate with H2O steam (for the metal oxides), with NH3 (for silicon dioxide), or by ion exchange methods. (Water and ammonia are thus added to the list of process chemicals.) Electronics-grade silicon may be prepared through zone-refining and other techniques with up to nine-9s purity, although these processes have not been thoroughly investigated in the present study.

In a discussion of the HF acid leach technique, Criswell (1978) points out that the process with its various options is adaptable to several of the potential lunar minerals or concentrates including feldspars, pyroxenes, olivines, and even nonsilicates such as ilmenite and spinels. Beneficiation of these minerals (the major constituents of lunar soil) seems unnecessary since the appropriate separations are performed later on the fluorides and fluorosilicates. However, if necessary, this beneficiation can be accomplished using the electrophoretic method described above.

In addition to Fe, Al,. Mg, Ti, Ca, Si, and O2. it is Possible that the HF acid leach process may be used to prepare Cr and Mn.These two elements are present in pyroxene (up to 0.5% MnO, up to 1.25% Cr2O3), olivine and spinel (which contain Cr). CrF2 is slightly soluble in water; MnF2 is soluble, so the techniques described above should still be applicable although the details of this extension have not been extensively studied.

One final problem unique to the HF process is the question of containers. Process vessels and tubing normally employed in terrestrial industry are attacked by hydrofluoric acid. One solution is to use special carbon steel alloys for this purpose - these are customarily employed for storage of fluorine gas because a protective layer of iron fluoride forms which greatly impedes further chemical attack. A second alternative is to use hydrocarbon-based waxes, paraffins or plastics which are not attacked by HF, applied as a thin layer to the insides of pipes and containers. Yet a third option is to develop new structures perhaps based on sulfur and phosphorus (Allcock, 1974) and other inorganic polymers (Lee, 1979) which could be in reasonably plentiful supply in the lunar factory.

Extraction of volatiles. Lunar soil heated to 1300 K releases 0.1% by weight of the following trapped volatiles: CO, CO2, N2, H2, H2O, SO2, H2S, CH4, and inert gases (He, Ar, Ne, Kr, Xe). As much as 0.5-1.5% by weight may be released upon heating to 1700 K (Phinney et al., 1977). CO may be reduced to carbon by methanation followed by decomposition of the CH4 species over a refractory catalyst (such as MgO) to C and H2CO2 may be reduced to CO by making use of the reversible reaction:

800 K
2CO <---> CO2 + C
1300 K

That is, CO2 passed over elemental C above 1300 K reduces to CO, which can then be methanated and further reduced to C over hot refractory. N2, H2, H2O, and SO2 are desirable process chemicals. H2S may be burned in O2 to yield SO2 and water. A sharply limited supply of O2 results in steam and sulfur vapor. If SO2 and H2S are mixed at room temperature, they react to form water and elemental sulfur. Finally, oxygen bubbled through an aqueous solution of H2S produces a precipitate of elemental sulfur.

Inert gases are useful in lasers and for providing a nonreactive atmosphere, and may be separated by fractional condensation using cold traps at various temperatures.

Boron production. Historically on Earth the most important source of boron has been borax or tincal (Na2B4O710H2O), though today the more common source is kernite or rasorite (Na2B4O74H2O) Other boron minerals include colemanite (Ca2B6O115H2O), ulexite (NaCaB5O98H2O), priceite (Ca4B10O197H2O), boracite (Mg3B7O13Cl) in salt beds, and sassolite (H3BO3).

Boron minerals on the Moon are likely associated with phosphorus-bearing apatite species (Dunning, personal communication, 1980), although it is possible that local concentrations of the most common anhydrous boron mineral, tincalconite (Na2B4O7), may be found in the vicinity of ancient lunar volcanic vents. In either case it should be possible to isolate the boron species using a combination of chemical and electrophoretic techniques. However, the details of this process cannot be specified until available boron resources on the Moon are more precisely characterized.

Terrestrial boron-containing minerals are either calcium or sodium berates. A calcium borate may be converted to a sodium borate by treatment with Na2CO3, yielding borax and CaCO3 which precipitates out of solution. (Calcium carbonate may be recycled by roasting to obtain CaO and CO2, from the latter of which elemental carbon can be recovered.) Sodium berates are reduced to boric oxide in two steps:

Na2B4O7 + H2SO4 + 5H2O ---> 4H3BO3 + Na2SO4

2H3BO3 --(heat)--> B2O3 + 3H2O

The sodium and sulfur may be recycled by the following steps:

Na2SO4 + CaCl2 ---> CaSO4 + 2NaCl

CaSO4 + C ---> SO2 + CaO + CO

2NaCl --(electrolysis)--> 2Na + Cl2

(Sulfuric acid and calcium chloride are added to the list of process chemicals.)

Low-purity boron is prepared by reduction of B2O3 with Mg, followed by vigorous washing with sodium alkali and HF. The impurities are a mixture of oxides and borides. Almost pure boron (up to 99.9999% is available commercially by this method) for electronics applications may be prepared by vapor phase reduction of BCl3 (or BBr3) with hydrogen on electrically heated filaments. BCl3 is prepared by heating B and Cl2 directly at 800 to 1100 K. Possible filament materials have not been investigated, but the mass requirement is probably less than 1 kg. Chlorine is added to the process chemicals list, since F2 cannot be substituted for Cl2 for vapor phase purification.

Phosphorous and halogens. More than 200 minerals containing up to 5% phosphorus by weight are known on Earth, but the two main species available on the Moon are fluorapatite, Ca5(PO4)3F and chlorapatite, Ca5(PO4)3Cl. The other lunar phosphorus-bearing mineral, whitlockite, is generally given as Ca3(PO4)2 but often is found associated with Mg, Fe, F, and Cl. Fluorapatite is by far the most abundant and is also the major source of fluorine on the lunar surface. (Amphibole has a trace of fluorine, but this small amount is probably not worth the trouble to extract.) Chlorapatite, very rare by comparison, is the major source of chlorine on the Moon, except for lawrencite (a nickel/iron chloride believed derived from meteorites). Whitlockite is also very rare.

Apatite is separated from lunar soil by the electrophoretic process described above. The calcium phosphate is then reduced to P2O5 by heating with silica (available from the HF leach stage) yielding pure phosphorus when treated with carbon:

2Ca3(PO4)2 + 6SiO2 + 10C --(electric furnace)--> P4 + 6CaSiO3 + 10CO

Alternatively, calcium phosphate dissolved in sulfuric acid gives phosphoric acid plus insoluble calcium sulfate (which may be recycled, see below). The acid is then reduced with carbon to obtain elemental phosphorus.

The sulfuric acid technique appears best for halogen extraction. When acted upon by sulfuric acid, a natural mixture of fluorapatite and chlorapatite undergoes the following net reaction:

3Ca3(PO4)2Ca(F,Cl)2 + H2SO4 ---> H3PO4 + HF + HCl + CaSO4

This results in a solution of the three acids. If heated to above 390 K (but below 486 K), the HF and HCl boil off leaving pure orthophosphoric acid behind. The evaporate is condensed, then separated into HF and HCl by either of two methods. First, the acid solution is desiccated in vapor form over anhydrous CaCl2, then cooled to 273 K. HF condenses and is removed in liquid form, leaving HCl gas to be electrolyzed to obtain H, and Cl. Or, second, after desiccation with CaCl2 the HF/HCl solution is electrolyzed with the release of H2 at one electrode and a mixture of F2 and Cl2 at the other. This mixture is cooled to 240 K which liquefies the Cl, (to be drained off) leaving F2 gas, which may be combined directly with the liberated H, to make HF. This entire problem may also be circumvented if fluorapatite and chlorapatite can be separated using electrophoretic beneficiation.

To recover sulfur, a valuable volatile, from the above process, the calcium sulfate is recycled by roasting according to:

CaSO4 + C --(heat)--> SO2 + CO + CaO

Supporting reagents. Reagents necessary to ensure closure of the LMF chemical processing sector include sodium hydroxide, silane, sulfuric acid, nitric acid, freon, ammonia, calcium chloride and sodium carbonate. The derivation of each is briefly reviewed below.

Lunar pyroxene contains up to 0.2% and lunar plagioclase up to 1.5% Na2O (Williams and Jadwick, 1980). Specific pyroxene minerals containing Na are acmite or aegirite, Na2OFe2O34SiO2 and jadeite, Na2OAl2O34SiO2. Among plagioclase feldspars are anorthoclase, albite, and andesine, Na2OAl2O36SiO2. After these minerals are obtained by electrophoresis, roasting causes the Na2O component to sublime above 1200 K. By 1800 K as much as 70% of the available Na2O may have evaporated, leaving behind a still solid residue of iron, silicon, and aluminum oxides (Williams and Jadwick, 1980). The liberated sodium oxide is dissolved in water to give NaOH. The small amount of Na produced during boron reduction may be added directly to the HF leach system as metal, or hydrated to form NaOH, as required.

Silane for microelectronic wafer fabrication may be prepared in either of two ways. First, elemental silicon may be heated in the absence of air with magnesium to form the silicide, which is then hydrolyzed with sulfuric acid to silanes and MgSO4 (which can be recycled for sulfur much like calcium sulfate). This hydrolysis gives about 25% yield of silicon hydrides, comprised of 40% SiH4, 30% Si2H6, 15% Si3H8, 10% Si4H10, and 5% of Si5H12 and Si6H14. These may be separated by fractional distillation; or, if cooled to below 258 K, all species liquefy except SiH4, which remains a gas and can be removed. A second process suggested by Criswell (1980a) involves hydrolysis of the Mg2Si with HCl, with the magnesium chloride hydrolyzed by steam to recover the HCl.

Sulfuric acid is relatively simple to prepare, provided a suitable catalyst is available. In the two-step contact process, SO2 is burned in oxygen and in the presence of catalyst to the trioxide, which is then dissolved in water to yield the acid. The usual catalyst was, traditionally, finely powdered platinum, and more recently vanadium pentoxide. If possible, the use of these substances should be avoided as Pt and V are rare in the lunar regolith. Fortunately, practically all refractory substances have some degree of catalytic activity in the contact process, provided they are immune to impurities. Alternative and plentiful viable catalyst agents include pumice (SiO2Al2O3), porcelain or powdered ceramic, and ferric oxide (Fe2O3), all of which are active and readily available in the LMF.

Nitric acid is more difficult to prepare, primarily because of the difficulty of "fixing" nitrogen chemically. The two most common commercial processes for acid production involve the use either of existing nitrate stocks or of platinum (for the catalytic oxidation of ammonia), neither of which is feasible at the LMF. A third method, not feasible commercially because of its low energetic efficiency,is the electric are technique first discovered by Priestley in 1772. Elemental nitrogen and oxygen are passed through a spark discharge, producing nitric oxide with a yield of 2.5% under ideal conditions After rapid quenching of the reaction mixture, the NO reacts rapidly below 873 K in an excess of O2 to form NO2, which makes nitric acid upon contact with water. Biological nitrogen fixation using Rhizobium and Azotobacter microorganisms is an interesting alternative and should be investigated further.

Freon (CF4) is prepared by fluorination of methane with elemental fluorine. The resulting mixture of CF4 and HF is separated by dissolution in water. There are two potentially feasible methods for producing ammonia. First is the standard Haber process, in which elemental nitrogen and hydrogen are combined directly at 800 K in the presence of iron and aluminum oxide catalysts. In the second process, magnesium is ignited at 600 K in a nitrogen atmosphere to form the nitride, which is then hydrolyzed to yield ammonia and magnesium hydroxide. Water and MgO are recycled by roasting the hydroxide.

Only very limited amounts of CaCl2 are needed, so direct combination of the elements (both of which are already available) is the preferred production pathway for boron production is obtained by bubbling CO2 gas through an aqueous solution of NaOH, then gently heating to recover the solute.

5E.3 Quantitative LMF Materials Closure

The arguments presented in section 5E.2 demonstrate that a surprisingly simple system involving 18 elements and perhaps two dozen mineral species and process chemicals can probably achieve virtually 100% materials processing closure. Reagents necessary for electronics parts fabrication were included so that the lunar SRS has the materials needed to replicate its own computer and robot equipment While the above is P,,bably not the minimum size chemical processing plant that can retain closure, it is certainly one example of such a system. Other possibilities should be pursued in future research. Of course, once a growing seed reaches full adult size, it can install a whole new series of production equipment (say, for the recovery of platinum group metals) making possible a new range of capabilities that were unnecessary during the early growth/replication phases.

Quantitatively, in order to rigorously demonstrate complete materials closure it would be necessary to work through every chemical process described above, calculate the exact materials mass for every structure, robot, and other LMF device on an element-by-element basis, then verify that enough of each could be produced by the system. Such a detailed computation clearly lies beyond the scope of the present study. However, the team has attempted to estimate some of the most critical throughputs and analyze their anticipated effects upon total system closure. In this context, "closure" is a relationship between given machine design and a particular substrate from which the machine's chemical elemental constituents are to be drawn. Hence, the numerical calculation of closure requires a knowledge of the precise composition both of the intended base substrate to be utilized as well as of the products which the SRS must manufacture from that substrate.

Following a method suggested by the work of Freitas (1980), the "extraction ratio" R (see sec. 5.3.6) is defined as the total mass of raw substrate material which must be processed (input stream) to obtain a unit mass of useful system output having the desired mass fractions of each required element (output stream). Consider the significance of the extraction ratio to the problem of materials closure. An R = 1 means that 1 kg of lunar regolith contains exactly the mass of all necessary LMF elements to manufacture a kilogram of desired output product. R = 10, on the other hand, means that 10 kg of lunar regolith must be processed to extract all of the elements required to make 1 kg of final product (see sec. 5.3.6).

For the purposes of the present study the team chose a trial value of R = 40 kg/kg. This choice is based on information from previous studies which suggests that 40 represents a good intermediate value between low closure and high complexity SRS materials designs.

On the one hand, for R < 10, the available mass fractions of certain critical but relatively rare elements such as H, C, B, and Cl fall too low to remain credible for a system requiring 100% closure. The missing materials must be imported as "vitamins" or the entire SRS must be redesigned to eliminate chemical processing and electronics using these elements. Examples of low closure models include the lunar processing factory designs proposed by Ho and Sobon (1979), R = 1.7; O'Neill (1976), R = 1.7; Phinney et al. (1977), R = 1.2; and Waldron et al. (1979), R = 1.1.These systems are capable of extracting only half a dozen of the most abundant lunar elements and are not expected to achieve more than 60-90% materials closure.

On the other hand, for R > 100 the problem lies not in extracting rare elements but in processing them fast enough to meet a T = 1 year replication time deadline. For instance, Freitas (1980) gives an example of a high complexity system which could extract 84 elements from asteroidal material. For R = 26,800 the replication time is 500 years. It appears that 10 < R < 100 is a plausible condition for 100% closure and i-year replication in SRS. The maximum recoverable mass from lunar soil for each element assuming R = 40 is estimated in table 5.13. The question remains whether or not these quantities are adequate to achieve quantitative materials closure.

Table 5.13.- Maximum Mass Of Chemical Elements Extractable From Lunar Soil, Per Year, For A 100-Ton Seed With Extraction Ratio R = 40

Element Typical global lunar abundance, kg element/kg soil Maximum extractable mass, kg (Rch)(A)(100 tons)
Fe 0.10 400,000
Al .07 280,000
Mg .05 200,000
Ti .02 80,000
Mn .001 4,000
Cr .002 8,000
C .0001 400
Si .2 800,000
O .4 1,600,000
P .0005 2,000
B .000001 4
F .0001 400
N .0001 400
H .00005 200
Na .003 12,000
Cl .00002 80
Ca .07 280,000
S .001 4,000
(Inert gases)? .00001 40

Certainly 100% closure exists for the six primary structural elements Al, Ca, Fe, Mg, O, and Si. Even if the entire 100-ton seed were comprised entirely of any one of these there is enough available of each. A similar argument may be made for Ti, since 80 tons in theory can be extracted. Steels and other alloys typically have 1% Mn, 0.2% Cr, and 0.1% C or less, which limits the total steel mass to 400 tons, 4000 tons, and 400 tons, respectively. Hence, alloy production will not be materials-limited by these three elements.

Carbon is also used in the boron and phosphorus production cycles. The mass of boron is so low that the carbon requirement is negligible in terms of mass. In the phosphorus cycle, 10 atoms of C are needed to cycle 4 atoms of P. Phosphorus is required as a dopant in silicon microelectronic chip manufacture and in phosphoric acid which is used as a photolithography process chemical and which also appears during the halogen recovery cycle. At most, 40 kg of phosphorus are required, necessitating a carbon budget of 100 kg. This leaves more than 200 kg of carbon to account for losses and special uses such as CO2 gas lasers.

Boron is used solely as a microelectronic silicon dopant; 4 kg of B can produce perhaps 103-104 kg of chips, more than enough for the 100-ton seed. A few kilograms of phosphorus (though high purity is required) will suffice for the same purpose, and the use of P as a process chemical should be more hydrogen-limited than phosphorus-limited because of the relative abundance of P in the lunar regolith.

According to calculations by Waldron et al. (1979), about 63 metric tons of H2, F2,and Na, half of which is F2, are needed for an HF acid leach extraction facility having a total mass of about 823 tons. According to Criswell (personal communication, 1980) this model may scale almost linearly down as low as 1 ton. The equivalents for a scaled-down 2.5-ton HF leach system are 90 kg F2 and 100 kg of H2 and Na. Sodium is about an order of magnitude more abundant than required, and fluorine does not appear to be a limiting factor even if recovery losses and spillages permit only 50% utilization of available stock. The supply of hydrogen, however, is crucial in achieving quantitative materials closure (see below). The 2.5-ton plant described above can output about 91 tons/year, which should be adequate to replicate a 100-ton seed once per year.

The primary use of nitrogen is in making NH3 for the recovery of silica and as N2 and HNO3 for the production of microelectronic chips. The 400 kg N2 given in table 5.13 is sufficient to prepare a maximum of 490 kg NH3, or 1800 kg HNO3. (These applications would require a maximum of 86 kg and 29 kg of H, respectively, hence are not seriously hydrogen-limited.) The amount of nitric acid seems more than sufficient, and the NH3 can produce 100 to 1000 kg of silica, which should be adequate with recycling and provided losses can be held to a minimum.

Chlorine appears in the boron- and phosphorus-production cycles - in the former it is consumed and must be recycled; in the latter it is incorporated in a deliquescent compound and should not incur serious losses or require chemical recycling. The preparation of 1 mole of boron requires recycling 0.25 mole of Cl, hence (0.5)(4 kg)(35.45/10.8) = 6.6 kg of chlorine are needed to produce 4 kg of boron. As for the phosphorus cycle, 80 kg Cl produces 125 kg of deliquescent CaCl2 which is capable of absorbing roughly its own weight in water. This should be sufficient with recycling (by simple heating) no more often than once a month on a T = 1 year schedule.

Sulfur is used primarily in the casting subsystem in the fabrication sector (about 600 kg required) and in the manufacture of suIfuric acid. This product is mass-limited about equally by the amounts of S and H available. The 4000 kg of sulfur can be used to prepare 12,000 kg H2SO4, and the 200 kg of hydrogen can make up 9800 kg of the acid. Since hydrogen also has many other uses, available S will be underutilized and perhaps 1 or 2 tons of H2SO4 reasonably can be produced. Is this enough? The main uses of sulfuric acid are in the recovery processes for B, P, F,and Cl, and in the preparation of silanes. The ratio of B:H2SO4 is about 4:1 moles, so to extract 4 kg B requires 9.1 kg acid. For phosphorus extraction, P:H2SO4 :: 2:3 moles, so (3/2)(98.1/31)(40 kg) = 190 kg H2SO4. For fluorine extraction, F:H2SO4 :: 2:1 moles, which requires (1/2)(98.1/19)(200 kg) = 516 kg acid. For chlorine extraction, Cl:H2SO4 :: 2:1 moles, which requires (1/2)(98.1/35.45)(80 kg) = 110 kg H2SO4. The quantity of silane needed for microelectronics processing is expected to be minimal, so it appears that adequate supplies of sulfuric acid can be made available with reasonable loss factors to LMF on a sulfur sustain the growth of a fully autonomous budget of about 1500 kg.

The only critical element appears to be hydrogen. This criticality is not especially peculiar to the present design, but rather stems from the relative scarcity of the element in lunar materials and the many chemical processing applications to which it may be put. Any hydrogen-chemistry-based materials processing system will encounter similar difficulties. The 200 kg of available hydrogen could make the maximum quantities of H-bearing compounds listed in table 5.14, although the available hydrogen must be spread ired with lower masses in among these applications as required with lower masses in each case. Expected requirements of hydrogen-bearing reagents are listed in table 5.14. Although these calculations are highly sensitive to the assumptions employed, closure may be achieved if an allowance of 5% for spillage and other losses is adequate. Obviously a major leak could seriously jeopardize a hydrogen-based LMF system.

Table 5.14. Hydrogen-Limited Materials Processing Reagents

LMF reagent Estimated LMF requirements, kg Maximum, limited only by hydrogen available, kg Fraction of available hydrogen required
NH3 300 1,100 0.273
H2SO4 1,000 9,800 .102
SiH4 100 1,600 .0625
HNO3 600 12,600 .0476
HF 190 4,000 .0475
H3PO4 100 6,500 .0154
NaOH 100 8,000 .0125
H2O 700 1,800 .3895
(5% losses) --- --- .05

If hydrogen supply remains a critical problem it may become necessary to: (1) redesign the processing system for greater hydrogen frugality, (2) select a slightly higher extraction ratio R to permit recovery of a greater mass of H, (3) locate and "mine" particular lunar soils extra-rich in H, such as the suggested use of ilmenite as a hydrogen "ore" (Green, personal communication, 1980), (4) accept a replication time longer than 1 year, or (5) relax the 100% closure requirement and permit resupply of small amounts of hydrogen "vitamin" from Earth.

5E.4 Sector Mass and Power Estimates

The overall functional layout of the LMF chemical processing sector is illustrated in figure 5.16. The operations flowsheet shows that there are 13 components within the sector: (1) input hopper, (2) electrophoretic separators, (3) P/F/Cl extractors, (4) boron extractors, (5) sodium extractors, (6) volatiles extractors, (7) HF acid leach system, (8) freon producer, (9) ammonia producer, (10) silane producer, (11) nitric acid producer, (12) sulfuric acid producer,and (13) output hopper.

Mass and power consumption for LMF materials processing may be estimated by comparison with other automated chemical processing designs that have been considered, and which are summarized in table 5.15. For R = 40, a 100-ton/year (self-replicating) output demands a 4000-ton/year raw materials input, or 0.13 kg/sec. Taking the range of values given in table 5.15, sector mass should lie within 18,200 to 78,000 kg. Similarly, the estimated power requirements range from 455 kW up to 10.9 MW, although in this case the lower values seem more appropriate. Dry thermal chemical processing techniques are associated with very high energy requirements, whereas lower values are found in wet chemistry processes - of which the HF acid leach selected for the present design is an example.

Table 5.15.- Comparison Of Chemical Processing Plant Masses And Power Requirements From Previous Related Studies

Source kg plant
kg/sec input
kg chemicals
kg/sec input
kg chemicals
kg plant
Power, W
kg plant
Johnson & Holbrow (1977) Al-processing plant 3.1X105 2.6X104 8.5X10-2 25
Phinney et al. (1977) carbo/silico-thermic plant 3.3X105 --- --- 140
Waldron, Erstfeld, & Criswell (1979) HF acid-leach metal extraction plant 3-6X105 3.5X105 4.1X10-1 37
O'Neill, Driggers, & O'Leary (1980) reference design figure 3.6X105 --- --- ---
O'Neill (1976) carbothermic Space Manufacturing Center 5.2X105 --- --- 100
Criswell (1980) 3.6X105 --- --- 100
Vajk et al. (1979) Space Manufacturing support requirements 1.4X105 --- --- ---

5E.5 Information and Control Estimates

Probably the most complex of the 13 sector components which appear in figure 5.41 is the HF acid leach system. From figure 5.41 this appears to consist of 34 component subsystems such as "precipitator," "dissolving tank," "fractional distillation tower," "centrifuge/filter," "Castner cell," etc. Each subsystem performs a single well defined task. In addition, there are 111 nodes (each denoting a point of connection of a pipe or supply line to another pipe or to a subsystem) each requiring at least one valve and valve control mechanism. At each valve there must be a number of sensors indicating valve position (open, closed, fractionally open), valve malfunction and cause (if simple), and volume or velocity of flow of material through the valves. Interface with actuators and reportage to the subsystem subcomputer are additional requirements.

Figure 5.41. -
Flowsheet and process equations for the HF acid-leach process

Assuming each valve can be automated with a 1K computer allocation, and each subsystem can be automated with a 10K memory allocation, then the total computer capability required for continuous leach system operation is (1)(111) + (10)(34) = 451K which is 7.2X106 bits using 16-bit words. This should be sufficient to handle normal system operations and troubleshooting, although actual repair must be done by mobile repair robots. Also, any catastrophic malfunctions such as pipe ruptures, jammed fixtures, leaks, heating element burnouts or explosions must be diagnosed and corrected by the mobile repair robots.

The chemical processing sector looks not to be a place where complicated new automation techniques will be required (Ayres, 1952; Foster, 1963; Kallen, 1961; Luke, 1972). Component processes are all state-of-the-art. There is no need for visual processing during normal operations, and procedures are standardized so no expert systems or judgmental algorithms are required beyond the simple integration of well defined sensor data. All operations will probably be hard-automated, and materials will be moved about almost entirely in sealed vessels. If there is need for additional transport within the system a materials transit network may be erected using metal or basalt tracks, electric motors and small carrier vehicles.

If each of the 13 sector components is as complex as the HF acid leach system (certainly a gross overestimate), then the total computer control capability required is about 6 megabytes or 9.4X107 bits using 16-bit words. The information needed to describe the sector sufficiently for purposes of self-replication must also be estimated. Assuming that each HF leach subsystem requires 2X106 bits for complete description (about a 200-page printed book, or 80,000 English words), and that each valve requires about half as much (say, 100 book pages), then the total for the HF leach system is 1.8X108 bits. Again conservatively multiplying by 13, the total information to describe the sector components is 2.3X109 bits. If sector equipment is distributed across a floor space of 5000 m2, then to store a map with 1-cm placement resolution requires a memory capacity of 8X108 bits assuming one 16-bit word to describe the nominal status of each 1 cm2 of platform space. Note that large empty areas convey useful information and must be mapped, since they may be used for traffic routes, repair routes, temporary warehousing, etc. The total information for sector replication is thus about 3.1X109 bits. The information control budget is 9.4X107 bits.

5E.6 References

Allcock, Harry R.: Inorganic Polymers. Scientific American, vol. 230, March 1974, pp. 66-74.

Arnold, J. R.; Criswell, D. R.; and Waldron, R. D.: Progress Report on Experimental Program to Develop HF Acid Leach Process for Refining Lunar Soils. Paper presented at the 5th Princeton/AIAA/SSI Conference on Space Manufacturing, 18-21 May 1981, Princeton, NJ.

Ayres, Eugene: An Automatic Chemical Plant. Scientific American, vol. 187, September 1952, pp. 82-88.

Criswell, David R.: Extraterrestrial Materials Processing and Construction. NASA CR-158870, vol. 1, 1978, and vol. 2, 1980.

Dunning, Jeremy D.; and Snyder, Robert S.: Electrophoretic Separation of Lunar Soils in a Space Manufacturing Facility. Paper presented at the 5th Princeton/ AIAA/SSI Conference on Space Manufacturing, 18-21 May 1981, Princeton, NJ.

Foster, David: Modern Automation Sir Isaac Pitman and Sons, Ltd., London, 1963.

Freitas, Robert A., Jr.: A Self-Reproducing Interstellar Probe. J. British Interplanet. Sec., vol. 33, July 1980, pp. 251-264.

Ho, Darwin; and Sobon, Leon E.: Extraterrestrial Fiberglass Production Using Solar Energy. In Space Resources and Space Settlements, John Billingham, William Gilbreath, Brian O'Leary,and B. Gossett, eds. NASA SP-428, 1979, pp.225-232.

Johnson, Richard D.; and Holbrow, Charles, eds.: Space Settlements: A Design Study, NASA SP-413, 1977, 185 pp.

Kallen, Howard P., ed.: Handbook of Instrumentation and Controls. McGraw-Hill Book Co., New York, 1961.

Lee, Stuart M.: Lunar Building Materials - Some Considerations on the Use of Inorganic Polymers. In Space Resources and Space Settlements, John Billingham, William Gilbreath, Brian O'Leary, and B. Gossett, eds. NASA SP-428, 1979,pp. 233-241.

Luke, Hugh D.: Automation for Productivity. John Wiley and Sons, New York, 1972.

Oldham, William G.: The Fabrication of Microelectronic Circuits. Scientific American, vol. 237, September 1977, pp.110-128

O'Neill, Gerard K.: Engineering a Space Manufacturing Center. Astronautics and Aeronautics, vol. 14, October 1976, pp.20-28,36.

O'Neill, Gerard K.; Driggers, G.; and O'Leary, B.: New Routes to Manufacturing in Space. Astronautics and Aeronautics, vol. 18, October 1980.pp 46-51.

Phinney, W. C.; Criswell, D. R.; Drexler. E.; and Garmirian, J.: Lunar Resources and Their Utilization. In Space-Based Manufacturing from Nonterrestrial Materials,

Gerard K. O'Neill, ed., AIAA. New York, 1977, pp. 97-123. (Progress in Astronautics and Aeronautics, vol. 57.)

Vajk, J. P.; Engel, J. H.; and Shettler. J. A.: Habitat and Logistic Support Requirements for the Initiation of a Space Manufacturing Enterprise. In Space Resources and Space Settlements, John Billingham. William Gilbreath, Brian O'Leary, and B. Cossett, eds., NASA SP-428,1979, pp.61-83.

Waldron, R. D.; Erstfeld, T. E.; and Criswell, D. R.: The Role of Chemical Engineering in Space Manufacturing. Chemical Engineering, vol. 86. 12 February 1979. pp. 80-94.

Williams, Richard J.; and Jadwick. James J.: Handbook of Lunar Materials, NASA RP-1057, 1980.

Zachary, Wm. B.: A Feasibility Study on the Fabrication of Integrated Circuits and Other Electronic Components in Space Utilizing Lunar Materials. Paper presented at the 5th Pritlceton/AIAA/SSI Conference on Space Manufacturing, 18-21 May 19S1. Princeton, NJ.