| Regolith Refining Summary
By Dave Dietzler The primary elements that must be extracted from regolith are iron, titanium, magnesium, aluminum, silicon and oxygen. We can also use raw regolith for shielding and cast or sintered basalt bricks, tiles, tubes, etc. Iron Iron fines also called nanophase iron particles exist in the regolith on the order of 0.15% to 0.5%. They also contain 5% nickel and 0.2% cobalt. These particles can be extracted with low intensity magnetic separators and processed by grinding, sieving and magnetic separation to break them up from silicates they are fused with to get a very pure iron-nickel particle feedstock. This iron can be melted in electric furnaces and cast as is or it can be converted to steel by taking cast slabs of iron, rolling them into thin plates or sheets, stacking them in an electric furnace with some carbon dust sandwhiched between the plates and then heating the iron till it is cherry red for several days. The steel would then be melted down and mixed with CaAl2O4 flux to remove impurities of silicon and sulfur. Flux would come from roasting highland regolith at 1500+ C. in the vacuum. Nickel and cobalt can be separated from the iron particles by subjecting them to CO gas at high pressure and moderate temperatures of about 200 degrees F. They metals form gaseous carbonyls that can be distilled to separate them and then decomposed with heat to get the metals and recover the CO gas. Although there is 5% nickel and 0.2% cobalt in the iron fines it is doubtful that we will process all the iron this way, so we won’t have huge amounts of nickel and cobalt, but we should get enough for special purposes and leaving the nickel and cobalt in the iron means it will make the steel stronger than plain carbon steel. Titanium Ilmenite, FeTiO3, can be separated from mare regolith electrostatically. This is then reduced with hot hydrogen at about 1000 C. in a fluidized bed to get water that is electrolyzed to oxygen and hydrogen. The O2 is stored and the H2 reused. The output of the fluidized bed is TiO2 particles fused with iron particles. These must be separated by acid leaching, treatment with hot CO gas to form iron carbonyls that evaporate off the TiO2 particles, or simple high temp heating in the vacuum to vaporize the iron off the TiO2 particles. Grinding and magnetic extraction of iron blebs might work, but only for the coarser particles. The TiO2 would then be electrolyzed to titanium metal in FFC cells. Since the calcium chloride flux of the FFC cells will get contaminated and decompose over time we will need chlorine and calcium on the Moon to replace electrolyte. Magnesium Mineral particles would be ground, seived and electrostatically separated because about half the regolith particles are agglutinates. Magnetic separation could get all the iron bearing olivines, pyroxenes, ilmenite and chromite out of mare soil leaving only feldspars and magnesium/calcium bearing minerals. These could be separated electrostatically. The magnesium bearing minerals would be mixed with MgO and CaAl2O4 from highland regolith roasting plus impure silicon or FeSi from magma electrolysis and heated to 1200-1500 C. in an electric furnace. Magnesium vapor will then be condensed. Aluminum Anorthositic highland soil would be ground to break up agglutinates, seived, magnetically and electrostatically separated to get a fairly pure anorthite feedstock. This would be melted and quenched somehow to form a glass. This is done to break up the crystalline structure of the anorthite particles. Glasses are amorphic. The glass would be ground fine, leached in sulfuric acid, the aluminum sulfate bearing liquid boiled down to leave dry Al2(SO4)3 that is then roasted to Al2O3, carbochlorinated and electrolyzed to Al metal. All reagents would be recycled. This seems to be the best process yet proposed for aluminum production on the Moon. It should be possible to make high silicon alloy iron with magma electrolysis for the sulfuric acid handling vessels and pipes or use cast basalt linings as this resists 96% sulfuric acid. It might be possible to vacuum roast FeO, MgO and SiO2 out of highland regolith leaving CaAl2O4 behind that will serve as a steel and magnesium production flux. This would be done at 1200 to 1500+ C. Since CaAl2O4 melts at 1600 C. we must wonder about problems with CaAl2O4 volatilization. We could take CaAl2O4 and use as is for flux or heat it to even higher temperatures like 3000 C. with microwaves or electron beams instead of carbon arcs that are typically used to reach such high temperature and avoid the problems of electrode burn up. CaO, Al2O3, some Al and Ca, and oxygen would be released and condensed. Most of the Al and Ca will react with O2 in the condensor and revert to oxides. A very low yield of Al and Ca is expected. Calcium aluminate volatilization might also be a problem. If we can get CaO and Al2O3 by destructive distillation of calcium aluminate we can use the CaO for cement making and avoid separating regolith particles to get pure anorthite feedstock, melting, quenching, grinding, sulfuric acid leaching, filtering out a solution of aluminum sulfate, boiling the solution down and roasting the Al2(SO4)3 to Al2O3 and all the equipment associated with this. The Al2O3 could be carbochlorinated to AlCl3 and electrolyzed to aluminum with the released chlorine gas being recovered and reused. A major benefit of AlCl3 electrolysis is that the carbon electrodes don’t burn up, so it is not necessary to recover CO and CO2 and shift them to methane, decompose the methane with heat to get carbon black, and recast the electrodes as with electrolysis of Al2O3 mixed with cryolite. Carbon electrodes in cryolite only last a few weeks. However, the CO and CO2 formed by carbochlorination must still be processed back to carbon. This also yields water that can be electrolyzed to gain oxygen. Carbon electrodes are made of 70% carbon and 30% pitch that are baked together. Where do we get pitch on the Moon? We must simply upport the carbon electrodes and use AlCl3 electrolysis that doesn’t burn them up. Perhaps we could use other materials for the electrodes like platinum coated titanium carbide. Or just titanium carbide made on the Moon. Even if we had an insoluble electrode we would not want to use alumina in cryolite electrolysis because the flux would have to be upported to the Moon and it will break down over time. Perhaps we could recapture flourine and sodium vapors and reconstitute the cryolite flux. The Toth process does not use flux or electrolysis. The alumina is carbochlorinated and then reduced with manganese. The manganese chloride that forms is then decomposed with heat and the Cl gas recovered. Heat is still required to induce the reactions. It is still necessary to upport chlorine until lunar sources of chlorine are tapped like pyroclastic glass and although manganese exists in regolith some way to extract it is yet to be found. A carbochlorination step is still required, therefore CO and CO2 must be shifted to methane that is then pyrolized to recover carbon. Perhaps it would be possible to directly electrolyze molten Al2O3 at 2000 C. with high temperature insoluble electrodes and a cooled furnace container. That would be about as simple as can be if the right materials are discovered. Silicon Impure silicon from magma electrolysis can be finely divided by grinding and treated with hot chlorine gas to form SiCl4 vapor that is decomposed by passing it over hot tungsten or tantalum filaments at about 850 C. The chlorine gas is released, recovered and reused. The other chloride salts of impurities in the silicon that form when it is treated with hot Cl gas stay behind. Impurities that do come out with the SiCl4 gas are removed by zone refining that gets the silicon 99.99% pure. This pure silicon would then be doped and sprayed in the vacuum to make low efficiency amorphous solar cells in massive quantities. It is also possible to react impure silicon with hot HCl gas to get the intermediate compound HSiCl3 that is passed over an AlCl3 catalyst and SiH4 and SiCl4 form. The SiH4 can be used for rocket fuel and the SiCl4 decomposed to silicon as described above. The chlorine and hydrogen would be recovered and reused. It might also be possible to roast silicon in a vacuum furnace until it volatilizes and condense the vapors to purify impure silicon and then zone refine it to ultra-high purity. This would certainly simplify matters. Simple vacuum heating and distillation would obviously reduce lunar requirements for chlorine for silicon processing. The furnace could use microwave heating and a distillation set up similar to the one envisioned for roasting anorthositic regolith to CaAl2O4, FeO, SiO2 and MgO. See image below. |
| Oxygen
Oxygen can be generated by magma electrolysis. Iron and steel processes using nanophase iron particles do not generate oxygen. Silicothermic reduction for magnesium metal does not generate oxygen, but it does produce some slag that can be broken up and used for aggregate for concrete. Aluminum production does yield oxygen. Carbochlorination of Al2O3 results in AlCl3 and CO gas that is reacted with hydrogen to form methane and water. The water can be electrolyzed to recover hydrogen and gain oxygen. If Al2O3 is directly reduced in furnaces with electrodes made of as yet defined materials oxygen will be generated. Hydrogen reduction of ilmenite yields water that can be electrolyzed to recover H2 and gain O2. Silicon purification does not yield oxygen. |
 |
 |