Importing ascent propellants for routine moonbase operations is an expensive proposition. If about 1 tonne of lunar oxygen could be produced, 235 kg of ammonia feedstock could produce the 571 kg of dinitrogen tetroxide needed for a nominal lunar ascent, as well as 360 kg of water and nitrogen. This represents a 4:1 leveraging of lunar imports, and reduces LEO launch mass by some 2.65 tonnes per mission. This requires a small, simple, cheap apparatus which simply passes oxygen and ammonia through a catalyst, cools the effluents below -12 C to capture the water and tetroxide, and allows waste nitrogen to diffuse through a zeolite. The equipment can also be used to process imported ammonia into hydrogen and nitrogen, and dinitrogen tetroxide has all manner of handy chemical uses, such as in the production of cheap nitrate fertilizers.
For the Reference Mission, the ascent vehicle requires 785 kg of propellants to be transported to the lunar surface, including 214 kg unsymmetrical dimethyl hydrazine (UDMH) and 571 kg of dinitrogen tetroxide (N2O4). Transportation of this mass from LEO requires nearly 3,000 kg to be launched into orbit -- about $50 million at Shuttle launch costs. This fuel is required every mission to get the crew back to the orbiting lunar transfer vehicle, and is even higher if more than 300 lbs of payload is to be brought back to the Moon on subsequent missions.
One option for reducing these costs is to use lunar resources to reduce the mass which must be imported from Earth. The ascent stage's UDMH is an analog of the familliar hydrazine molecule, HH-N-N-HH, but with two hydorgen atoms replaced with a methly (CH3) group. However, dinitrogen tetroxide is 69.6% oxygen, which can be readily extracted from lunar minerals, and indeed a lunar oxygen extraction pilot plant will be carried on the first mission. If a simple apparatus was included to synthesize dinitrogen tetroxide out of imported ammonia and indiginous oxygen, it would afford considerable leveraging towards the cost of future missions.
Dinitrogen tetroxide can be prepared by thermal decomposition of heavy metal nitrates, or by reducing or decomposing nitric acid. It may also be prepared using a rhodium-platinum catalyst in the quite exothermic oxidation of ammonia to produce nitric oxide, which combusts to form nitrogen dioxide:
5 O2 + 4 NH3 -> 4 NO + 6 H2O
4 NO + 2 O2 -> 4 NO2
Note that this is a handy way to generate water for moonbase life support. Although this process thermodynamically favours the creation of nitrogen, passing a preheated ammonia-air mixture over a platinum catalyst for a short contact time and above 650 C, nitric oxide yields can be high. Reaction temperature is typically 800 to 960 C. While all surfaces catalyze the oxidation of ammonia into water and nitrogen, few materials are practical in favouring NO2 formation.
A considerable fraction of terrestrial costs is the erosion of the catalyst, due to the appreciable vapour pressure of plantinum at the reaction temperatures. This problem is reduced by the use of rhodium-platinum alloys as a catalyst, but the further step of recovering catalyst from flue dust is not economial for the small-scale applications in mind. If erosion is a problem for the lunar apparatus, the best solution is to occasionally replace a small imported module containing the catalyst.
The process is completed when, at low temperatures, NO2 dimerizes into N2O4. The two chemicals tend to coexist over a fairly wide temperature range, with about 10% of the molecules being N2O4 at 100 C. Compared with a boiling point of 21.15 C and a freezing point of -11.2 C, at -9.3 C, the dimerization is complete.
2 NO2 <-> N2O4
Alternate reactions also take place which eventually produce water, molecular nitrogen, or the desired nitrogen dioxide. Nitrogen can be separated from unreacted oxygen in the waste stream by diffusion through a molecular sieve. Natural and synthetic zeolites are available with pore channels of specific dimensions. As nitrogen has a critical diameter of 3.0 angstrom and oxygen has one of 2.8 angstrom, a zeolite with a cut-off at just below 3.0 Å would be appropriate.
Although the zeolite can reduce inefficiency to as little as 0.25%, optimal conditions allow the nitrogen to be accompanied by about 10% of its mass in oxygen. As oxygen is cheap and we merely want to remove nitrogen from the cycle, this is not a problem.
The thermodynamic propensity for ammonia to disociate into nitrogen and hydrogen is convenient if in the future an alternate oxidizer should replace the use of dinitrogen tetroxide. As ammonia is a convenient form to import needed nitrogen and hydrogen for life support in an expanding lunar outpost, the oxygen supply can simply be cut off and the temperature reduced to 600 C to transform the ascent propellant manufacturing equipment into an ammonia dissociation apparatus:
2 NH3 = N2 + 3 H2
The gaseous products can be separated by using palladium as a diffusion filter for the hydrogen at 200 C. This can be installed as a modular replacement for the zeolite filter which ordinarily would separate nitrogen from oxygen.
In addition to the use as a rocket propellant, the chemical can also be used in all manner of lunar industrial processes. Nitrogen dioxide reacts reversibly with water to produce nitric acid and nitric oxide. It can oxidize carbon monoxide, producing carbon dioxide and nitric oxide, which further reacts with carbon monoxide to leave molecular nitrogen. It reduces hydrochloric acid, producing chlorine, water, and nitric oxide. The chemical reacts with various metals to produce nitrates, and even with many metal oxides, peroxides, hydroxides, carbonyls, and halides.
Fertilizer could be cheaply produced by reacting the tetroxide with the potassium oxide or other alkali metals oxides readily available in the lunar regolith. Potassium nitrate and other metal nitrates are excellent fertilizers, and for potassium nitrate all but 13% of the mass of fertilizer and water produced would be derrived from lunar resources. In addition to inexpensive use on the Moon, this would be a prime candidate for export to asteroids or orbital settlements.
The bottom line is that 235 kg of ammonia and 1,007 kg of lunar oxygen would produce 571 kg of tetroxide, 340 kg of water, and 19 kg of nitrogen, a total of nearly 4:1 import leveraging. This requires an apparatus which processes 10.3 kg of reactants daily at 90% efficiency. The author guesses that the apparatus can be developed, refined, built, and intensively tested in convincing conditions for less than one million dollars, and of much less mass than even the 336 kg of direct savings on the first use.
If fertilizer production apparatus were integrated into the dinitrogen tetroxide production, one possible scenario would be to bring an extra 65 kg of ammonia for a total 300 kg. Reacted with 1,248 kg of lunar oxygen and about 100 kg of lunar minerals, this would produce -- in addition to the 571 kg of propellant, 443 kg of water, and 19 kg of nitrogen -- 282 kg of potassium or calcium nitrate fertilizer for agricultural purposes. The fertilizer production option would produce 5.9 kilograms of water and nitrates for every kilogram imported, and combined with propellant production in the above scenario, the total import requirements are reduced by a factor of 4.4.
In situ fuel production does not have to compromise crew safety. The necessary oxygen production and the ammonia reaction can take place during the first (or second) crew's stay on the Moon, and carefully stored before the crew departs. Should the stockpiled fuel be lost between missions, the next crew can revert to importing the fuel. Should the fuel be lost when the crew are about to use it and already on the Moon, they can simply borrow the propellant they are making for the next crew. But in normal circumstances, the propellant can be produced and verified long before the mission which is to use it is launched.