Utilization of Space Resources in the Space Transportation System

Michael C. Simon

Utilization of space resources (i.e., raw materials obtained from nonterrestrial sources) has often been cited as a prerequisite for large-scale industrialization and habitation of space. While transportation of extremely large quantities of material from Earth would be costly and potentially destructive to our environment, vast quantities of usable resources might be derived from the Moon, the asteroids, and other celestial objects in a cost effective and environmentally benign manner.

Of more immediate interest to space program planners is the economic feasibility of using space resources to support near-term space activities, such as scientific and commercial missions in the 2000-2010 timeframe. Liquid oxygen for use as a propellant in a space-based transportation system appears to be the space resource that has the firmest near-term requirement for quantities great enough to be produced economically in a nonterrestrial setting. This paper identifies the factors most likely to influence the economics of near-term space resource utilization. The analysis is based on a scenario for producing liquid oxygen from lunar ore.

Analysis Methodology

The primary purpose of the parametric cost model developed as part of this study is to identify the factors that have the greatest influence on the economics of space resource utilization. In the near term, this information can be used to devise strategies for technology development so that capabilities developed will produce cost-effective results.

Predicting the actual Costs Of particular scenarios for space resource utilization is only a secondary objective of this analysis. Estimates are made and dollar values are assigned principally to allow comparison of options. Since the technologies for space resource utilization are in an early stage of development, it is premature to state conclusively whether mining the Moon, asteroids, or other celestial bodies makes economic sense. The parametric model is designed more for flexibility than for precision.

Although preliminary estimates indicate that production of oxygen from lunar ore is a project that is likely to yield an economic payback, this activity was selected as the "baseline scenario" primarily because its requirements can be relatively well defined. The major systems required to support this baseline scenario have been identified without much difficulty:

Once these major support systems were defined, fifteen key variables were identified as influencing the cost of developing and operating these systems (table 3 [Lunar Oxygen Production-Major Cost Variables table]). Cost variables were generalized so that the parametric model could be adapted to the evaluation of alternative scenarios. Next, equations were developed to calculate capital and operations costs as functions of these variables. Using the codes and units detailed in (table 3 [Lunar Oxygen Production-Major Cost Variables table]), these equations are

Capital Cost = (p x cp) + (nt x cn)
+ (nm x cu) + cf
+ ct x [(p x mp)
+ (nm x mm) + mf]

Operations cost = ct x {(nr x mm)
+ [(1-d) x 125,000]}
+ (nb x nf x $100,000)

where the capital cost is defined as the total cost of developing, building, and installing the lunar base elements (including transportation costs) and the operations cost is the annual cost of manufacturing 1 million kilograms (1000 metric tons) of LO2 per year and delivering to LEO as much of this LO2 as possible.

The term in square brackets [(1-d) x 125 000] in the operations cost equation reflects the assumptions that a portion (1d) of the LO2 produced on the Moon is used as propellant to deliver the remaining LO2 (d) to LEO and that 1 kilogram of hydrogen must be delivered from Earth to the Moon for every 8 kilograms of oxygen used as propellant for the Moon-to-LEO leg (125,000 kg of hydrogen for the projected annual production of 1 million kg of oxygen). The higher than-usual mixture ratio of 8:1 was selected for the baseline case after initial analyses showed that the resultant reduction in the hydrogen requirement offers substantial economic benefits.

The constant cost ($100,000) in the operations cost equation is the cost of ground support per provider per year. The variable that precedes this constant, nf, is a ground support overhead factor which is multiplied by the labor cost to obtain total ground support cost.

After these cost equations had been set up, baseline values were assigned to each cost variable, using the ground rule that the technology having the lowest risk would be used for each system. Lunar base modules, for example, were assumed to be modified versions of the laboratory, habitat, and logistics modules that are being developed for NASA's LEO space station.

Another ground rule was that the costs of gathering the scientific data needed to select the lunar processing site would not be included in this model. It was further assumed that an initial lunar base would be in place prior to the LO2 production activity and that this facility would be scaled up to meet the LO2 production requirements. Thus, the cost included in this model is only the marginal cost of expanding this initial facility to produce LO2.

Although some of these ground rules lowered capital and operations cost estimates, the specification of lowest-risk technology made these estimates higher than they might be if cost reducing technologies are developed.

Results of the Analysis

Once baseline values were assigned to the cost variables, a simple calculation was made to obtain capital and operations cost estimates. These costs were determined to be
Capital cost: $3.1 billion
Operations cost: $885 million/ year
An analysis of the performance of proposed lunar orbital transfer vehicles (OTVs) indicates that 49.2 percent of the LO2 produced would be delivered to LEO. Consequently, the unit cost of L02 delivered to LEO, assuming 10-year amortization of capital costs, was determined to be $2430/kg ($1100/lb). This cost is one-quarter to one-third of the current cost of using the Space Shuttle, although it is somewhat greater than the cost that might be achieved with a more economical next-generation Earth-launched vehicle.

It should be reemphasized, however, that all cost estimates used in this analysis are based on a specific set of assumptions and are for comparative purposes only. The most important objectives of this analysis were the assignment of uncertainty ranges to each of the cost variables, the calculation of the sensitivity of LO2 production costs to each of these variables, and the analysis of the technical and programmatic assumptions used to arrive at values for each variable. The data developed to support the sensitivity analysis are summarized in table 4 [Capital and Operations Costs-Sensitivity to Cost Variables table]. The baseline, best case, and worst case values assigned to each cost variable are shown, along with the impact of each variable's best case and worst case values on capital or operations cost. For example, as power requirements vary from a low value of 4 MW to a high value of 12 MW, with all other variables held at their baseline values, the capital cost for establishing the LO2 production capability ranges from $2.30 billion to $3.90 billion.

From this table it is evident that the principal driver of capital cost is the lunar base power requirement, while the Earth-to-Moon transportation cost is the most important operations cost driver. Since capital costs are amortized over a 10-year period, the EarthtoMoon transportation cost has a much greater overall impact on the cost of lunar LO2 in LEO. If this cost could be reduced from its baseline value of $10 000 to its best case value of $5000 per kilogram delivered to the Moon, capital cost would drop from $3.1 billion to $2.45 billion, operations cost would decline from $885 million/year to $468 million/ year, and the cost of lunar L02 would be reduced from $2430/kg to $1450/kg. Conversely, at its worst case value of $15 000/kg, the Earthto-Moon transportation cost would drive capital cost up to $3.75 billion, operations cost to $1.303 billion/year, and the cost of lunar LO2 to $3410/kg.

An alternative approach to showing the impacts of the cost variables is illustrated in table 5. It lists the effect of each cost variable in terms of percentage changes in the capital or operations cost and in the cost per kilogram of LO2 produced (with a 10-year amortization of capital cost). In this table the variables are ranked in order of their impact on the LO2 cost/kg. The influence of each variable is calculated as an "impact factor" equal to the average of the percentage changes in LO2 cost/kg due to the best-case and worst-case values of the variable.

From these impact factors it is clear that two of the cost variables are far more important than all the rest: net lunar oxygen delivered to LEO and Earthto-Moon transportation cost. The percentage of lunar-produced oxygen delivered to LEO is important because of its double impact. As the percentage of From these impact factors it is clear that two of the cost variables are far more important than all the rest: net lunar oxygen delivered to LEO and Earthto-Moon transportation cost. The percentage of lunar-produced oxygen delivered to LEO is important because of its double impact. As the percentage of LO2 delivered declines, From these impact factors it is clear that two of the cost variables are far more important than all the rest: net lunar oxygen delivered to LEO and Earthto-Moon transportation cost. The percentage of lunar-produced oxygen delivered to LEO is important because of its double impact. As the percentage of L02 delivered declines, LO2 cost/kg increases not only because less LO2 is delivered but also because more hydrogen must be transported from the Earth to match the LO2 used as propellant from the Moon to LEO.

The six operations cost variables are among the nine most important, largely because the impact of capital cost is spread out over the 10-year amortization period. The relative significance of the operations cost leads to the important observation that LO2 production costs may be reduced substantially by increasing capital expenditure on technologies that can reduce operations cost. One such technology is Earth-to-Moon transportation, which has a tremendous impact on operations cost. Capital cost factors, such as the mass and cost of the power system and of the processing/ storage facility, have much less impact on LO2 cost/kg.

Technology Development Required To Improve Performance

It is not possible to conclude, on the basis of this analysis, that production of liquid oxygen from lunar materials is justifiable on economic grounds. Although the cost estimates for the baseline scenario are encouraging, a number of technologies with significant impact on LO2 production costs must be explored. The performance and cost of space-based orbital transfer vehicles is the most critical technology issue. Developing a low-cost OTV is a fundamental requirement for cost effective utilization of space resources because the OTV is the single most effective means of reducing Earth-toMoon transportation cost.

Another key issue is the cost of hydrogen used for launching payloads from the Moon. Production of lunar LO2 would be far more cost-effective if a capability for the co-production of lunar hydrogen could be developed (even though capital cost might increase substantially). Although relatively large quantities of lunar ore would need to be processed, the additional cost of lunar hydrogen production could be offset by a savings of over $600 million/year in transportation cost. Production of some alternative propellant constituent, such as aluminum, also might offer an opportunity for reducing or eliminating costly import of fuels from Earth. However, this example would require the development of an aluminum burning space engine.

A third category that seems to have substantial impact on the economics of lunar resource utilization is the technologies influencing lunar base resupply requirements. Increasing lunar base automation, closing the lunar base life support system, and other steps to reduce the frequency and scale of resupply missions appear to have a high likelihood of providing economic benefits and should be given particular emphasis in future studies.

If all three of these objectives were met to the greatest extent possible (i.e., if Earth-to-Moon transportation cost were reduced to its best case value, if hydrogen transportation requirements were eliminated, and if lunar base resupply requirements were eliminated), the cost of lunar LO2 delivered to LEO would be reduced from $2430/kg to $600/kg, or about $270/lb These figures assume no change in capital cost; but, even if capital cost were doubled to achieve these capabilities, L02 cost would be reduced to approximately $1100/kg -less than half the baseline cost.

Twenty-five key technology issues influencing these and the other cost variables in LO2 production are presented in table 6 [Impact of 25 Key Technology Issues on Cost Variables in Space Resource Utilization Table]. In this table, a dark square indicates a strong impact of that technology issue on the cost variable, a light square indicates a moderate impact, and no square indicates little or no impact. The selection and evaluation of these technology issues was made by a panel of experts convened for the purpose, not by a quantitative analysis. The fifteen cost variables ranked as in table 5 [Sensitivity of Capital, Operations, and Oxygen Production Costs to Ranges of Cost Variables Table] are listed across the top of table 6 [Impact of 25 Key Technology Issues on Cost Variables in Space Resource Utilization Table] in descending order of importance. Hence, table 6 [Impact of 25 Key Technology Issues on Cost Variables in Space Resource Utilization Table] is a graphic representation of the relative importance of the technologies based on three considerations: total number of squares, number of dark squares, and distribution of squares to the left of the chart (i.e., toward the most important cost variables).

To quantify the impact of these twenty-five technology issues on the economics of the baseline scenario for space resource utilization, a technology weighting factor of 3 was assigned to each dark square and a factor of 1 to each light square. These technology weighting factors were then multiplied by the impact factor (table 5 [Sensitivity of Capital, Operations, and Oxygen Production Costs to Ranges of Cost Variables Table]) for each cost variable that the technology issue affects. The sum of the products across each row was calculated as the total economic weighting factor for that technology issue. For example, the lunar base power source has a heavy impact on cost of power and power system mass for an economic weighting factor of (3 x5) + (3 x 3) = 24.

The ten most important technology issues, according to their total economic weighting factors, are listed in table 7 [Major Technology Issues in the Cost-Effective Production of Lunar Oxygen].

Finally, it is important that parametric cost analyses such as this one be used to assess a variety of space resource utilization scenarios. Use of lunar ore for production of construction materials is one such option, although to be cost-effective this type of enterprise would probably require a dramatic increase in space activity. Another option that merits careful consideration is the development of asteroidal resources. Both rocket propellants and construction materials could be derived from asteroids; and, while the up-front cost of asteroid utilization would probably exceed the capital expenditure required for lunar development, operations cost could be substantially lower. Further analysis of all these opportunities needs to be carried out over the next several years before a commitment is made to any particular plan for space resource utilization.

As new technologies are developed, the reliability of cost estimates for space resource utilization will improve. Eventually, it will be possible to generate cost estimates of sufficient fidelity to support detailed definition of space utilization objectives. An important step in this process will be the adaptation of this parametric model and similar techniques to the evaluation of a broad range of space resource development options.

Aluminum-Fueled Rockets for the Space Transportation System

Andrew H. Cutler


Aluminum-fueled engines, used to propel orbital transfer vehicles (OTVs), offer benefits to the Space Transportation System (STS) if scrap aluminum can be scavenged at a reasonable cost. Aluminum scavenged from Space Shuttle external tanks (fig.9) [Separation of external tank from Shuttle Orbiter] could replace propellants hauled from Earth, thus allowing more payloads to be sent to their final destinations at the same Shuttle launch rate.

To allow OTV use of aluminum fuel, two new items would be required: a facility to reprocess aluminum from external tanks and an engine for the OTV which could burn aluminum. Design of the orbital transfer vehicle would have to differ substantially from current concepts for it to carry and use the aluminum fuel. The aluminum reprocessing facility would probably have a mass of under 15 metric tons and would probably cost less than $200,000. Development of an aluminum-burning engine would no doubt be extremely expensive (1 to 2 billion dollars)., but this amount would be adequately repaid by increased STS throughout. Engine production cost is difficult to estimate, but even an extremely high cost (e.g., $250,000,000 per engine) would not significantly increase orbit-raising expenses.

The combustion of aluminum delivers 22 percent more energy per unit mass of reactant than does the combustion of hydrogen. Since propellant costs on the Earth are a small part of total launch costs, the added complexity of tripropellant engines is not warranted for launch from the Earth's surface. However, if aluminum fuel were available in low Earth orbit (LEO) at a much lower cost than cryogenic fuel, the savings in propellant cost could offset the cost of developing an aluminumfueled space engine.


Aluminum-fueled rockets are ubiquitous. Aluminum is added to the solid fuel of rockets to enhance their performance. Most groundbased solid rockets are aluminized. Solid rockets intended for launch in space are following this trend (e.g., the inertial upper stage lUS-rockets). The Space Shuttle itself burns twice as much aluminum (in the solid rocket booster SRBs as it does hydrogen (total of the elemental hydrogen in the external tank and the chemically combined hydrogen in the SRB fuel).

The aluminum oxide (Al2O3) produced by the Shuttle's combustion of aluminum quickly settles out of the atmosphere. That produced by rockets taking satellites to geosynchronous Earth orbit (GEO) does remain there. The Al2O3 would be a pollutant in cislunar space. However, the dilution is such that aluminum oxide pollution there should not be a severe problem for a long time.

Experiments have shown that aluminum additives can also enhance the performance of liquid fueled rockets. The combined efforts of those working on solid and liquid propellant rockets might have an increased total effect if they were focused on the development of an aluminum fueled space engine.

Aluminum Availability in LEO

Aluminum could be made readily available as a fuel in LEO. The 1988 National Space Policy offers Shuttle external tanks (ETs) free to users in space. (The conditions include demonstrating that any reentry of the tanks can be controlled.) External tanks could be carried to orbit for little additional cost and with little adverse impact on Shuttle operations. These tanks could then be reprocessed to provide fuel aluminum.

Aluminum would probably be burned in the form of micron-sized powder. From extrapolations of current mission models, the maximum projected aluminum demand is about 14 metric tons per tank. This amount of aluminum could be recovered in the following manner (see fig. 10 [Reprocessing of space Shuttle External Tank]): All gas is vented from the tanks. A cutting machine with an electron beam cutter (demonstrated on Skylab for 2219 aluminum alloy) enters the tank. It makes circumferential cuts in the barrel sections and in the ogive (pointed arch section) immediately adjacent to the ring frames. The cuts do not cross the cable tray. These circumferential cuts are connected by longitudinal cuts along both sides of the cable tray and between the ring frames.

Since the cutting is done while the thermal protection system (TIPS) is still intact, all spatter and fumes will be contained inside the tank and may be trapped to prevent extensive contamination of the local area. "C"-shaped sections of the tank composed of a metal sheet coated on one side with TPS material may now be broken loose. These "C"s contain the needed 14 metric tons of 2219 aluminum alloy, so the remainder of the tank-ring frames, intertank (section between the hydrogen and oxygen tanks), slosh baffles, end domes, and cable tray-may be discarded.

The aluminum strips may then be rolled onto a mandrel to densify them for melting. The bulk of the TPS coating will separate from the aluminum sheet while it is being rolled up. The small amount of TPS material remaining on the sheet can be removed with a rotating wire brush and discarded along with the other unprocessed materials. The rolled aluminum strip is placed in an induction furnace and melted. The liquid aluminum can be pumped from this pool and turned into powder the same way it is on Earth-by being sprayed against a rapidly rotating wheel. The vacuum of space allows efficient electron beam cutting and prevents oxidation of the aluminum powder as it is being formed.

The operation described here requires further study. Among the problems to be solved is that of disposing of the residual portions of the external tank in an environmentally acceptable way. The generation of large or small debris (e.g., pieces of insulating material) that cannot be controlled could make the aluminum scavenging concept untenable.

The amount of aluminum available in the external tanks is far larger than the amount of aluminum fuel needed. Only the most easily reprocessed part of the tanks need be worked on. These portions of the tank are composed of only one alloy, 2219, which has been extensively characterized in commercial use. These facts combined with the fact that the plant makes only one product (aluminum powder) suggest that the plant will be simple, reliable, and economical.

Aluminum as a Propellant

The combustion of aluminum by oxygen is very energetic. Most of the energy is released as aluminum oxide condenses from the gas phase. Aluminum oxide condensation in the rocket nozzle is a rapid process. Condensation of aluminum oxide heats the gas, which expands to provide thrust. Since the aluminum oxide particles do not completely exchange momentum and energy with the gas phase, there is some impulse reduction due to two-phase flow loss. The two-phase flow loss must be controlled by including in the exhaust a gas with low molecular weight (Frisbee 1982). Hydrogen is the ideal candidate. An oxygen-hydrogen-aluminum engine with a mixture ratio of 3:1:4 is expected to have a specific impulse of over 400 seconds, and eventually it might achieve a specific impulse of over 450 seconds (Cutler 1984).

Propellant Demand In LEO

Much of the mass currently lifted to LEO is propellant for orbit raising and maneuvering. According to OTV transportation models (table 8 [Models for Orbital Transfer Vehicle Traffic]), 45-180 metric tons of payload mass per year will be lifted to geosynchronous Earth orbit as soon as an OTV is available or expendable rockets can be fueled at the space station. To lift these payloads from LEO to GEO, 90-360 metric tons of propellants will be required in LEO. The specific propellant requirement depends on the design and performance of the OTV used, including whether or not it is reusable. In this paper, I have assumed a propellant-topayload ratio of 2:1. Some of this (130-325 metric tons per year) can be scavenged from the Space Shuttle's external tank in the form of unused hydrogen and oxygen (see table 9 [Usable Propellant Available in LEO Yearly (in metric tons) graph].

Aluminum-Fueled Engines for OTV Propulsion

Table 9 [Usable Propellant Available in LEO Yearly (in metric tons) graph] shows the amounts of O-H and O-H-Al propellant usable under different conditions. If the traffic model requires more propellant than can be scavenged, additional propellant must be carried in place of payloads of greater intrinsic value or new technology must be introduced to improve performance.

Marginal improvements can be made in OTV performance by incorporating advanced cryogenic engines. Improving engine performance from the current Isp of 460 seconds to an Isp of 480-490 seconds would allow 7-11 percent more payload to be carried to GEO with the same cryogenic propellant supply.

If oxygen-hydrogen-aluminum engines were available (and relatively small amounts of hydrogen could be added), the amount of scavengeable propellants would double (Table 9) [Usable Propellant Available in LEO Yearly (in metric tons) graph]. Besides the aluminum to match the scavenged hydrogen and oxygen, there would be excess aluminum to match hydrogen and oxygen transported from Earth, thus doubling its effectiveness.

A simplified cost model is shown in figure 11 [Relative Propellant Costs for Orbital Transfer].
If the assumptions used here are shown to be valid, the model indicates that significant cost savings can be made, even at low traffic levels, by scavenging cryogens from the Space Shuttle and, at higher traffic levels (above 90 metric tons per year), significant cost savings could also be made by scavenging aluminum from the external tank.


Aluminum-fueled space engines may be more economical than advanced cryogenic engines in the regimes where advanced engines can offer significant savings over current technology (that is, where there is enough traffic that the benefits from improved performance exceed the cost of developing a new engine). Thus, assuming that all programs for the development of new engines have about the same cost, any argument which justifies developing advanced oxygen-hydrogen engines justifies investigating the development of an aluminum-fueled space engine. The most economical way to run an OTV program may be to rely on an OTV with a current RL-10 engine until propellant demand is near the scavenged supply and then change over to an OTV .propelled by an oxygen-hydrogen-aluminum engine.


Cutler, Andrew H. - 1984. H2/O2/Al Engines and Their Application to OTV's. Paper IAF-84314, 35th Int. Astronaut. Fed. Congress, Lausanne, Switzerland, Oct. 5-12.
Frisbee, Robert H. 1982. Ultra High Performance Propulsion for Planetary Spacecraft. JPL Report D-1097. Jet Propulsion Laboratory, California Institute of Technology, Pasadena, CA.


Table of Contents