Balancing Molecular Nanotechnology-Based Space Transportation and Space Manufacturing Using Location Theory: A Preliminary Look

Copyright (c) 1997 by Thomas L. McKendree. Presented at the Space Studies Institute Conference on Space Manufacturing, May 8-11, 1997. Placed on the web by the Molecular Manufacturing Shortcut Group for educational and individual use only. No commercial use without prior written permission. All Rights Reserved.


An introduction to extending location theory to operations in space, and an investigation of estimated capabilities for space transportation and space manufacturing, given molecular nanotechnology.

Presents four models for siting an orbiting manufacturing facility with respect to two inputs and one market location, each progressively adding cost elements. Addresses space transportation using molecular nanotechnology, with a focus on solar sails and rotating tethers. Addresses space manufacturing with a focus on the particular strengths of molecular manufacturing and space manufacturing. Identifies two system concepts for using molecular manufacturing in space. Identifies open issues, and urges further work in the area.



Molecular nanotechnology, the ability to design and build with atomic precision, continues moving forward in theory1 and towards enabling practice2. One NASA researcher suggests that future nanotechnology-based aerospace systems may be fielded as early as 20153. It is thus reasonable to begin thinking, with a systems perspective4, about what molecular nanotechnology could mean for space operations.

Preliminary examinations suggest individual molecular nanotechnology-based space systems should be capable of dramatic performance5,6,7,8, but McKendree9 also notes that these individual performances are so high as to probably qualitatively transform their context and environment, complicating analysis of the real utility of these capabilities.

Location theory10 offers one way out of this problem, by raising the analysis to the level of integrated super-systems of manufacturing facilities and transportation systems. Thus, a reasonable next step towards understanding the implications of molecular nanotechnology for our future space operations is to achieve the following four-part program:

  1. Extend location theory to cover robust operations in space;
  2. Develop top-level system performances for molecular nanotechnology-based space transportation system by developing and refining these system architectures;
  3. Develop top-level system performances for molecular nanotechnology-based space transportation system by developing and refining these system architectures;
  4. Combine the results of the first three parts into a view of integrated molecular nanotechnology-based space transportation and space manufacturing systems.
This paper is a first look at that program, focused on illustrating preliminary results from the first three parts.

Part I - Extending Location Theory to Orbital Space

Location Theory

Location theory includes the concern of solving optimization problems for choosing the preferred zone or location for a facility. This involves minimizing transportation costs, and often trading off between various inputs, different production balances, and between market delivery levels.


The classic example of continuous location theory is the two-input, or "triangle," model, in which inputs from two different locations on a plane are sent to a facility and used in a fixed ratio to produce a product that is shipped to a market point. The problem is to determine the optimum facility location, which reduces to minimizing transportation costs that are linear to the distances traveled. This is mathematically equivalent to finding the minimum weighted distance between three points, a problem illustrated in figure 1 below.

The function creates a single continuous basin on the plane, and is easily solved both by geometric and by simple optimization approaches.

Figure 1. Contour plot of cost as a function of facility location for a sample problem of the two-input model. The optimal facility location is shown near (0.5, 0.75).


Discrete location theory problems are often expressed as mixed linear and integer programming problems, with the continuous variables primarily representing flows along a network, and the integer variables primarily representing the presence or absence of facilities at particular locations. It's node-to-node model of transportation space is particular well-suited for larger problems in space operations, but here is not investigated further.

Extending the Two-Input Triangular Model to Orbital Space

With locations defined as specific orbits, rather than fixed points, the Two-Input Triangular Model can be extended to orbital space. The problem is, given two orbits from which inputs are provided at a fixed ratio, and given a third orbit to which the product is to be delivered, determine the optimum location for the facility, which will take the inputs and turn them into the output.

This topic will be illustrated with a recurring example, where the input locations are taken to be circular orbits roughly corresponding to the orbits of Earth and Ceres, with the market location a circular orbit roughly corresponding to the orbit of Mars.

The difference between the four two-input models presented below is the different cost functions used. Each adds a new feature to the transit cost model.

Figure 2. Circular Orbits to Scale. Ceres and Earth orbits are inputs, Mars orbit is the market, and the facility will be at an arbitrary 4th circular orbit. Distances in meters.

Two-Input Model for Hohmann Delta-V Transit Costs

The simplest cost function for orbital transport is that cost equals the change in velocity (Delta V) for a Hohmann transfer orbit. Since the equation is11:


This can be expressed as:

Cost = (2)

Where r1 is the orbital radius of the first input, r2 is the orbital radius of the second input, rm is the orbital radius of the market, i1 is the ratio of mass of input 1 needed for each kg of the product, i2 is the ratio of mass of input 2 needed for each kg of the product, ct1 is the cost to transport 1 kg of input a Delta V of 1 m/s from the location input 1, ct2 is the cost to transport 1 kg of input a Delta V of 1 m/s from the location input 2, and ctp is the cost to transport 1 kg of product a Delta V of 1 m/s from the market.

[Note, in a more sophisticated model the cost of fuel at the facility could vary with rf.]

The simple case, where all the input ratios and transit costs are 1, is shown in figure 3 above.

Figure 3. Transit cost as a function of orbital radius of facility. Inputs at orbits of Earth and Ceres; market at orbit of Mars.

Two-Input Model Where Transit Costs Are Fuel Usage in Hohmann Delta-V

One minus the mass ratio from the rocket equation gives the excess mass fraction, representing the ideal fuel usage, as shown below:


The new variable is v0, which is the ideal exhaust velocity, or specific impulse times g. The cost function is parallel with (2).


A v0 of 4410 m/s, yields the cost curve in Figure 4. Note that unlike total Delta V, fuel usage has a continuous derivative. Direct optimization is simple for this model.

Figure 4. Transit cost as a function orbital radius of facility. Minimum is at radius = 2.33x1011 m, cost = 109.6, just beyond Mars' orbit.

Two-Input Model Where Transit Costs Are Fuel Usage Plus Carrying Cost of Transit Time

Transit on a Hohmann trajectory from the orbit of Ceres to Mars takes roughly 1.6 Earth years. Transits between gas giants can take decades. Microeconomic analyses of traditional projects on Earth with these time scales depend significantly on the time-cost of money.

The carrying cost for inventory in orbit is the interest charge on the price of the inventory (either an input or a product) over the transit time.

The period of transit for a Hohmann trajectory can be calculated as:


This gives a cost function of:


The key new variable, i, is the interest rate. The variable cp is the cost of producing the output out of the inputs. In previous models it did not matter, since it is treated as a constant, independent of facility location. For different transit times it will be subject to different interest charges, however, and thus must be explicit in this model.

Remember the ideal exhaust velocity, v0, is buried within Fuelra->rb.

An interest rate of 20% per year is sufficient to force a tiny adjustment in the optimal facility orbital radius on our basic example, and yields the following cost curve:

Figure 5. Transit cost as a function of orbital radius of facility. Responding to the time-cost of inventory in transit moves the minimum to radius = 2.32x1011 m.

In this case, interest charges on material are dominated by the mass ratio costs. To better illustrate the effect of inventory carrying costs, consider a second problem with longer transit times. For example, Lewis12 suggests mining He3 from Uranus. Let the inputs coming from circular orbits at the radius of Ceres and Uranus, and the market in a circular orbit at the radius of Jupiter (maybe a Trojan asteroid). To keep the graph readable, i has been reduced to 10%, appropriate for a conservative operation. Keeping all the other numbers the same yields the higher of the following curves.

Figure 6. Comparison of the Fuel Plus Carrying Cost model versus Fuel as (Sole) Transit Cost model. Inputs are at the orbital radii of Ceres and Uranus; market is at the orbital radius of Jupiter.

Two-Input Model Where Transit Costs Are Fuel Usage Plus Carrying Cost of Waiting Time

A Hohmann trajectory between two bases is only possible with they are aligned such that when the vehicle departs, it will travel exactly one-half ellipse to arrive at the destination just when that base arrives. Circular orbits of different radii have different periods, creating for each pair of orbiting objects a fixed period before they return to the same relative position. That time between orbits is called the synodic period. For circular orbits it can be calculated as:

, (7)
if ra Does Not Equal rb
or, 0 if ra = rb

[Note, the second case implicitly assumes that two objects at the same orbital radius will be collocated. Otherwise, synodic period would be infinity when ra = rb.]

The waiting time is intended to represent the time that must be spent between the purchase of an object at a base, and when it can be delivered to another base via a Hohmann trajectory. It is defined here as the period of the Hohmann transit, plus one-half the synodic period.


Interest charges on inventory in waiting time will capture the economic cost as extracted resources wait for shipment and delivery to the facility, and as finished products wait for shipment and delivery to their market (end user). Incorporating this full waiting time into the model yields the cost function:


WaitTra->rb is no longer continuous, however, and indeed goes to infinity as ra approaches rb. This drives the carrying costs to infinity (given any positive interest rate), and thus creates a discontinuous cost function, as shown below:

Figure 7. Waiting time with Hohmann transfers creates a discontinuous cost function. Local minima are indicated by dots.

These discontinuities are an artifact of the restriction to Hohmann transfers between orbits. Less energy-optimal trajectories would allow launches without waiting so long, and dramatically lower total costs for the very close orbits, which have long synodic periods. It is expected that this corrected cost function could be continuous.

Part II - System Architectures for Space Transportation Using Molecular Nanotechnology


Rockets, which expel on-board reaction mass, are the current standard architecture for space transportation. The real performance of current rockets in delivering payload, however, is significantly limited by parasitic mass, such as structures and fuel tanks, which must be carried, and which displace useful payload mass. McKendree9 illustrates how the higher strength-to-mass of molecular nanotechnology (MNT) materials improves this performance, changing for example the ratio of payload mass to dry, empty vehicle mass, from roughly 1:3 to roughly 3:1 (and more).

Furthermore, molecular manufacturing (MM) can be used to re-form materials into new products, thus potentially turning the remaining parasitic mass into useful payload.

Drexler7,13 discusses MNT-based, solar-electric ion engines. By combining very low-mass, high-specific energy solar panels with arrays of very small ion engines, exhaust velocities of 250 - 1000 km/s with accelerations of 0.8 - 9.8 m/s2 should be feasible. When not thrusting, the large area, low-mass structure would be capable of limited solar sailing.

Solar Sails

A solar sail is a surface that generates a motive force by directing pressure from the Sun's light. Solar sail performance is maximized by minimizing mass per surface area, while retaining high reflectivity.

The solar sail's force is proportional to the solar flux, which is proportional to 1/r2 (where r is the radius to the Sun). The force due to solar gravity is also proportional to 1/r2. Thus, the ratio of the solar sail force to the gravitational force is independent of the distance from the Sun.

Estimated Acceleration

Drexler7 suggests that a solar sail of 100 nm aluminum would be appropriate. More ambitiously, he13 suggests that a 20 nm solar sail would retain 90% reflectivity and also be a suitable solar sail. This work uses the 100 nm solar sail, for conservatism, but, with 20 nm thick solar sails, accelerations up to 4 1/2 times greater than presented here may be possible.

The solar constant is defined as the solar flux at 1 AU, 1370 W/m2. Dividing by the speed of light gives a solar pressure of 4.57 x 10-6 kg/(m.s2). This is the pressure from a one directional flux--bouncing light in the exact opposite direction doubles this pressure.

A 100 nm thick aluminum solar sail masses 0.27 g/m2, so the maximum acceleration due to light pressure is


where the LoadFactor is the ratio of the vehicle's total mass to its solar-sail mass.

For comparison, the acceleration at 1 AU due to solar gravity is 0.005878 m/s2, less than a fifth of the acceleration from the force of light pressure.

By angling the solar sail, the acceleration due to the light pressure, and thus the total acceleration vector, can be varied.

The feasible accelerations for a solar sail at 1 AU, with LoadFactors of 1 though 10, are shown in figure 8 below. At different solar distances, the magnitudes would change but the shape would be the same.

Figure 8. Feasible accelerations for a loaded solar sail at 1 AU. Each line corresponds to a load factor of 1 through 10. The horizontal axis is tangential acceleration (m/s2), and the vertical axis is radial acceleration (m/s2).

Even a load factor of 5 allows the solar sail to accelerate directly away from the Sun.

In-Flight Fabrication

With on-board molecular manufacturing, a vehicle might retain a previously used component, such as a fuel tank, and then build a solar sail in transit out of the component. The energy of the Al-Al bond at 298 K is 221 zJ. [Note, one zepto Joule = 10-21 Joules.] If one imagines that MNT would fashion a solar sail by breaking every inter-atomic bond in a block of Al, and then spend the same energy placing each atom where it belongs, then one m2 of 100 nm thick solar sail would require 1331 J to fabricate. One m2 of solar cells at 1 AU, operating at 10% efficiency and supporting MNT processing of 10% efficiency (much lower than Drexler14 suggests), could produce in 100 hours 3704 m2 of 100 nm thick solar sail (enough to carry a 1 kg payload at a load factor of 2).

As a first approximation, the rate at which MNT could produce a solar sail would vary directly with the available power. At greater ranges from the sun, the solar flux would fall as 1/r2, however the characteristic orbital period would grow as r3/2 (Kepler's third law), so there still would be ample time to fabricate a solar sail and put it to effective use.

System's Solar Sail Budget

Solar sails are an attractive transportation option, given molecular nanotechnology, but the finite solar flux limits the total capacity that solar sails could transport simultaneously in the system. Assume a solar sail fleet with 100 nm Al sails and an average LoadFactor of 5, allowing the sail acceleration to exceed the acceleration due to solar gravity. At 1 AU the total mass in transit supportable by the solar flux is

= 3.8 x 1020 kg (11)

This may look like a lot, but it is only about one-fifth the mass of Ceres, while consuming the entire solar flux. If we consider only solar sails maneuvering within +/-3.5 degrees from the plane of the ecliptic, a zone which contains all the planets of Venus through Neptune, this allows only 2.3 x 1019 kg in transit at 1 AU, or 1.8 x 1020 kg at the radius of Ceres.

The actual budget for solar sail transit in the solar system is likely to be even lower, because there will be many competing uses for the solar flux. While the performance of MNT-based solar sails is impressive, there is not the solar capacity to support massive use of solar sails. Thus, most large bases should spend most of their time without solar sail aid. Furthermore, once significant use of the solar system's resources is underway, most bulk cargo should expect to travel similarly unaided.

This suggests a future need to coordinate such resources in the solar system as the solar flux. Decisions must be particularized to specific amounts in specific locations at specific times. One approach would be to create a central scheduling authority, responsible for allocating the solar resources, and scheduling their uses. A more flexible and open approach would be to define marketable rights to specific portions of the flux. The flux channels could then be bought, sold, rented, bundled, subleased, optioned, distributed (e.g., with mirrors) and otherwise allocated dynamically as conditions warranted. Note, a corresponding market in allowed "bads" (e.g., release of radiation, or reflecting of light where it is not wanted), might also be needed to fully coordinate the real effects.

Bishop discusses such an overall control function as part of his proposed system for interplanetary transportation8.

Rotating Free Tethers

Mackenzie15 provides a good introduction to rotating free tethers. These can be thought of as cables spinning free in space, able to grab and release objects to receive and impart orbital velocity. They require careful design, to provide protection against being severed, and to carry the non-linear load along their length. Tethers have traditionally been analyzed as having constant stress and continuously variable thickness along their lengths15,16. From Mackenzie15 we have the equation for figure of merit of:


where lower case delta is the material density, sigmaw is the working (allowed) stress in the tether material, and Delta V is the change velocity when the tether catches or throws, the payload. Thus, the tether can impart 2 x Delta V, and do so with the capture and release at different times and points on it's orbit.

As McKendree9 discusses, one advantage molecular nanotechnology offers to space systems is high strength to mass diamondoid materials. With a lower case delta of 3510 kg/m3, and a sigma of 5 x 1010 Pa, a diamond tether could outperform those discussed by Mackenzie15. Furthermore, his analysis required a safety factor of 3. With repair (see below), MNT materials should be able to maintain high purity and low defects, allowing a more ambitious safety factor. Equation 12 is graphed below for diamond.

Figure 9. Tether-to -Payload Mass Ratio as a Function of Catch or Throw Delta V (km/s), for diamond with safety factors of 2 and 3.

With high mass ratios, tethers must serve many payloads to justify themselves. As to their performance, consider that the difference between the speed of Earth's orbit around the sun, and solar escape velocity, is a bit over 12 km/s. By providing half with a catch, and half with a throw, the tether indicated with the dot in figure 9 could pluck something from High Earth Orbit (HEO) and send it to any point in the outer solar system, including on a solar escape trajectory. The catch would require a challenging orbital rendezvous, however, since the tether would have to orbit the sun roughly 6 km/s faster than Earth at the point of capture. This does illustrate, however, that diamondoid tethers have significant potential for aiding transportation in the solar system.

Indeed, since velocities fall as orbits move beyond the Sun, but the capabilities of rotating tethers stays constant, they are particularly attractive for transport in the outer solar system, Figure 10 illustrates the mass ratios for rotating tethers sized to operate in the region beyond Earth, both for safety factors of 2 and 3. Higher mass-ratio rotation tethers would be feasible, and more capable, particularly in dropping payloads into lower orbits.

Figure 10. Tether-to-Payload Mass Ratio for Catch-and-Throw from circular orbit to solar escape velocity with a diamond tether, as a Function of distance from the Sun (m).

More aggressively, one could imagine a rotating tether in solar orbit, used to change the orbital plane 90deg.. This is ambitious, as plane changes are known to be very energetic and expensive maneuvers11. In one operational concept, such a tether would be in the plane of the ecliptic. Thus, vehicles in the main disk could dock with the tether at any time, and the tether would have to make the maneuver with a single throw. The mass ratio for such a tether in shown in figure 11, below.

To work, the tether must cancel the tangential velocity, and add an equal amount at 90deg.. This requires a throw angle of 135deg.. The angular momentum of the tether will stay pointed in a particular direct, however, and thus the tether will only be properly aligned twice per orbit.

One approach would be to use two counter-rotating tethers, physically coupled so they would have no net angular momentum, and could rotate through the orbit. A second approach would be to accept the twice per orbit performance, and place the tether in an orbit at a 90deg. inclination, with the disk of tether rotation in the orbital plane. This at least allows the tether to divide the maneuver into a capture and a throw. The mass ratios for this second approach are also shown in figure 11 below.

Figure 11. Tether-to-Payload Mass Ratio for 90 deg. orbital plane change as a function of distance from the Sun (m), using diamondoid tethers with safety factors of 2 and 3. Gray lines are for catch-and-throw; black lines are for pure throw.

While such tethers appear to be in the realm of the feasible, the massive capital cost implied by the mass ratios, combined with their operational limits, suggests that alternate approaches might be more desirable. Gravity assists from Jupiter are effective. Nonetheless, these tethers offer a means over time to transform the solar system into a Dyson sphere.

Maintenance Repair of Atomic Damage

The background radiation in space will damage atomically precise structures, including MNT materials. Drexler14 provides a radiation damage model of ~1015 failures per rad per kg of material. To repair this, consider that the C-C single bond has an energy of 556 zJ. If each failure in a diamondoid tether is repaired, then this requires:


where "overhead" is the number of Joules required to locate and prepare to repair bonds, for each Joule spend in repairing the bonds. Note, this treats the repaired bond as endothermic, when actually it is exothermic. This is conservative, and provides a clear baseline for defining the probably much larger overhead.

A worst case could be taken to be 100 Rad delivered in 10 hours from a solar flare. Untreated, this would create a density of 1.4 x 10-7 failures in the structure to repair.

To power ongoing repairs, fast enough to keep up, even in this worst case, assume a 10% efficient solar panel attached near the pivot of the tether, and pointing towards the sun. Using the 1370 W/m2 at 1 AU, the allowable overhead is:


which suggests that repair should be quite feasible, but will also require ongoing repair at the molecular level.

Part III - Using Molecular Nanotechnology for Space Manufacturing

The Utility of Molecular Manufacturing Strengths in the Context of Space Manufacturing

Molecular manufacturing (MM) is the fabrication of products to atomic precision through the direct manipulation of molecules. It is the application of MNT to the processing of matter, and has a number of strengths. Many of MM's strengths are of particular use in space manufacturing.


MM offers the potential for extreme material and energy efficiency. Acting with atomic precision, it need not waste a single atom without reason. By controlling the trajectories of reactions, it can make most molecular rearrangements nearly thermo-dynamically reversible14. Indeed, Drexler14 illustrates that most of the energy of exothermic reactions can be captured for useful work.

Note, however, that MM must still pay the overhead of driving it's operating mechanisms, which will never run with thermodynamic perfection. Mill-style MM mechanisms apparently will be much more efficient than general purpose robotic devices. Thus, efficiency will be high for processing when all the material is already ordered within an MNT-framework. Energy efficiency will be significantly lower in the capture and initial processing of external material.

For space operations, this high energy efficiency is most useful for deep-space operations further from the sun, where available power levels will be lower. The high material efficiency offers most advantage to operations at locations in space where particular critical elements are scare, or must be shipped in.


The characteristic operating frequency of a system is inversely related to the size of its components. By using primarily molecular-scale components, MNT offers very high operating speeds. This is why an MM system has the potential to process its own mass worth of material within an hour14.

This is significantly faster than the characteristic action time for interplanetary operations. Hohmann transfers take roughly half the year of the bodies transited between. Even the higher-speed possible with MNT propulsion (see above) gives characteristic interplanetary times closer to weeks than hours. Therefore, a reasonable approximation when at the early stages of stepwise refinement17 for a mixed MNT/interplanetary space system architecture, is to treat MNT actions as instantaneous compared to interstellar transits. Note, however, that this approximation should not apply without closer examination to the bootstrapped exploitation of in situ resources.


By combining unit operations that net only one or a few atoms properly placed with systems that can select from many options when sequencing unit operations and larger assembly steps, MM offers efficient systems which tremendous range of production. While some MM systems can be special purpose, MM includes a wide range of very general-purpose matter re-arranging devices. Indeed, because of the advantages of bootstrapping (see below), an attractive strategy is to use general purpose devices to build special-purpose devices for any particular manufacturing task at hand. Thus, just as a computer can run a wide variety of programs, including programs written after the computer was built, feasible MM systems would be capable of a wide variety of manufacturing tasks, including operations which were not pre-planned into the system (although MM requires a fully operational specification of the manufacture at the moment of execution).

A major benefit of this flexibility in space operations would be greatly simplified logistics. Rather than waiting for parts, or stocking every conceivable part, MM could produce objects as the need arises. With a mission control providing support, fabrication instructions can be sent, and the space MM system need not even determine its operating instructions. Remember, however, that MM can only rearrange atoms, and when a chemical element is fully in local use, additional supplies would have to be supplied from elsewhere.

MM-based systems need not limit themselves to producing small products. They have the flexibility to encompass large assembly operations, producing almost arbitrarily large structures. For very large space systems, this could be quite useful.

Atomic Precision

MM intrinsically builds to atomic precision. This provides the major benefit that products can include large strength-to-mass materials such as diamond; large strength-to-mass is particularly important for some space systems.

Furthermore, building to atomic precision means building objects with very high quality. This can be important in providing long system life, and in allowing systems to safely operate with higher performance through lower safety factors.

Finally, atomic precision is required for molecular scale mechanisms.


Macroscopic levels of MM requires massive parallelism. Plans for this usually call for bootstrapping, which is the use of MM to produce more MM capacity. The simplest example to explain is self-replication, wherein MM systems build copies of themselves. It is inherent in the design and nature of many MM systems examined that they support, bootstrapping5,14,18,19.

As has been shown independent of MNT, bootstrapping offers tremendous leverage for space operations20,21,22, The natural ability of a mature MM to bootstrap provides the key strategic advantage for it's use in space operations.

Self-Replacement, Self Repair

By using MM, components of a system could be produced on-board, and switched out both to replace failed components, and as a form of regular maintenance. For long life, this requires using MM to build additional MM capacity, while scrapping older capacity that may or is failed.

The space environment includes a significant background radiation23, most of which we on Earth evade due to our planet's magnetic field and atmosphere. Strategies for dealing with this have included very heavy shielding24, but MM offers another approach--on board repair though dynamic remanufacture of subcomponents25, Preliminary calculations suggest that on-board repair through self-replacement of self-manufactured sub-components has the capability to dramatically extend system life. Key parameters are the support systems redundancy, the failure distribution shape, and the ratio of failure time to total repair time.

The Utility of Space Manufacturing Advantages in the Context of Molecular Nanotechnology

Space manufacturing has a number of advantages, some of which are of particular utility to MM.

Heat Source

MM requires energy, and solar energy is a reliable, continuous, primary source of power in space. MNT should be particularly able to take advantage of solar power in space, as it can have very low mass per m2 solar power assemblies13 and in microgravity, gigantic solar power structures are feasible.

Similar to solar sails, MNT could also form a light pressure supported parabolic mirror, for concentration of large amounts of sunlight.

Cold Sink

Cooling is a fundamental performance limit on MNT14, including MM. This could be a particular problem for MM in free space, as the vacuum of space acts as a strong thermal insulator. Fortunately, deep space provides a deep cold-sink for thermal radiation, of particular use for MM.


Intended MM operations include extremely reactive intermediate structures, such as bonds left dangling for long periods of time, and needs a low-reactivity environment to do so. Internal vacuum has been proposed14 as an ideal low-reactivity environment. Space offers a high-quality vacuum, which could be used to easily provide this.

Furthermore, it is particularly promising that this vacuum is in the environment. Traditional MM designs are for systems that take external materials into themselves, and conduct the atomically-precise operations internally14,19. In space, the vacuum may allow atomically precise, highly reactive unit operations on the interface between the system and the environment. As "The greatest leverage in systems architecting is at the interfaces26," this is a very interesting possibility.

Pollution Dump

When O'Neill asked his central question, "Is a planetary surface the right place for an expanding, technological civilization?27," one of the obvious disadvantages of a planetary surface was environmental degradation from industry. In contrast, Space offers the potential to dump waste products where they will not threaten existing ecosystems.

MM, however, does not require dumping noxious chemicals, as it can reform them5,28, and store unneeded elements for later use. For space operations, recycling is desirable, as there is a finite supply of desirable elements in the system. Thus, this advantage for space operations is not of tremendous use for MM.

Selectable Acceleration

Objects in orbit experience minuscule accelerations (microgravity). By rotating, selectable acceleration fields, from microgravity through 1 gee and beyond, can be provided.

Many intended space-manufacturing uses of microgravity, such as electrophoresis29, are intended to provide higher yields or quality by avoiding the distortions of gravity. Through convergent assembly from atomic precision, however, MM generally offers the capability to produce the same products, within Earth's gravity. This is because, at the local scale gravity is overwhelmed by the inter-atomic and inter-molecular forces MM exploits, and at the large scale deflections can be overcome by building atomically-precise proper sub-assemblies, and by assembling in a supported framework. Thus, MNT is a direct competitor to an orbital manufacture intent on exploiting micro-gravity in a process step.

Assembly in microgravity allows the construction of very large objects, including huge structures which could not support themselves on the surface under Earth's gravity. Roughly thousand km diameter rotating settlements6,8,30 and giant solar reflectors seem interesting.

In Situ Resources

A fundamental advantage of space manufacturing is that it can make use of the resources in Space. Lewis12 points towards the tremendous resources accessible just in our own system. MM offers a sufficiently powerful tool to take full advantage of those resources.

Example Architectures

A couple examples of MM-based space manufacturing systems illustrate some of the power this capability promises.

"Seeds" for Carbonaceous Chondrites

One notional system31 is a "seed" of equipment, delivered to a carbonaceous chondrite asteroid, which consumes the asteroid, turning it into a single, integrated mass of flexible, MNT-based capital equipment. Design goals for such a device are a delivered mass of no more than a microgram which grows to totally consume the asteroid in a period of weeks6. Further investigation is required to assess the feasibility of such a small mass on the asteroid's outer layer bootstrap into a structure that can penetrate to the volatiles trapped deeper in the asteroid. This may vary with the asteroid's gross structure.

Molecular Manufacturing as a Closed-Environment Life-Support System

One of the early conceptual building blocks in the development of MNT ideas was the "meat machine32," a box which would take raw material and directly assemble steak. Extending this idea, one could imagine a CELSS comprised of molecular manufacturing equipment which directly disassembles waste products, and directly assembles consumables. The simple molecules of the air cycle are amenable to very short-sequence mill-style14 manipulation. Recycling water is only slightly more complex. Foods will be significantly more complex to assemble, and for the psychological benefit of the crew, an excellent subject for flexible manufacturing.

Filtering and recycling impurities remains the final challenge. This requires MM tailored for each recycled input, and additional machinery to remove and store whatever else happens to exist. Fortunately, these stores can be examined, and additional cycles designed and added to remove those contaminants as well.

Open Questions

The current extension of continuous location theory to space operations must extend even further. Specifically, how can this approach model other types of transits? In particular, the approach must support modeling of high-performance solar-sails (where the entire trajectory is non-ballistic), of high-performance trajectories such as that feasible with MNT-based solar-electric ion engines, and of trajectories which include momentum-transfer maneuvers, such as orbiting to a rotating cable and being boosted to a higher trajectory. If other highly-promising MNT-based space transportation concepts with characteristically different transits are identified, then the approach should be extended to encompass those types of transits as well. For momentum transfer maneuvers, a related problem is determining the proper locations (orbital constellation) for a set of stations.

A potentially easier intermediate step would be to extend the model to ballistic non-Hohmann transits. This alone should eliminate the cost discontinuities from models with transit time.

The models shown restrict bases to locations within the same gravity well. Thus, the model works for bases orbiting the sun, and would work for bases in equatorial orbit around Earth, but does not consider bases in low orbit around different planets, nor consider bases on the surface of planets. This extension is critical, however, to support trade-offs between Earth launch of material, and in situ production.

Location theory itself focuses on single-product facilities. How can the it be extended to capture the utility of flexible facilities (such as a molecular manufacturing system)? Such an extension is probably best handled in stages.

The extension of discrete location theory to orbital space remains. Given the node-like nature of a set of orbiting objects, discrete models should prove more appropriate for many purposes.

While in-flight fabrication of solar sails using MM seems quite feasible, an actual process design needs to be sketched out. Furthermore, other solar sail designs have promise33, and should be investigated as architectures when using MNT.

The high performance of MNT-based solar sails suggests that other transportation systems based on reflecting solar light, such as solar flux distribution mirrors, are worth investigation with MNT parameters.

While basic estimates of MNT-based solar-electric ion engines are available7,13, more detailed examinations need to be published.

While on-board repair of radiation damage using MM seems quite feasible, especially since most repair can occur in the lulls between high-radiation events, a more detailed system design would be encouraging.

Space-based MM is the area which requires most elaboration. Promising architectures for space manufacturing deserve closer examination. Furthermore, performance estimates have to be derived in terms that are useful for the extension of location theory into space.


Location theory provides a sufficiently extensible framework to consider combined manufacturing-transportation systems, even when the locations of interest are orbits.

Examination of MNT-based space transportation suggests a natural division of labor, with rotation deep-space tethers tending to operate further from the sun, and solar sails tending to operate closer to the sun.

Space manufacturing though MM requires more effort to adequately characterize. If that effort succeeds, all the parts should be available for an integrated examination of space manufacturing and space transportation, in a world of MNT.

The high performance estimates for MNT-based space systems suggest that its development would be quite sufficient to open up the tremendous resources beyond Earth. Even if not necessary, it should allow easier and fuller use of space's bounty.

Ultimately, an integrated examination of space manufacturing and transportation, using parameters appropriate for MNT, will give a much better planning estimate of what we will accomplish in space once we develop this promising technology. This avenue continues to deserve vigorous pursuit.


1 Drexler, K. Eric, "Breakthrough-Scale Design For a Molecular Robotic 'Hand'," Foresight Update, No. 28, pp. 1, 4, March 30, 1997.

2 Soreff, Jeffrey, "Recent Progress: Steps Towards Nanotechnology," Foresight Update, No. 28, pp. 6-7, 10, March 30, 1997.

3 Saini, Subhash, "Petaflops Computing and Computational Nanotechnology," Nanotechnology, Vol. 7, No. 3, pp. 224-235, 1996.

4 McKendree, Thomas, "Educating Systems Engineers For an Impending Metatechnology in Product and Process Systems," Proceedings of the Third Annual Symposium of the National Council On Systems Engineering, James E. McAuley and William H. McCumber, ed., pp. 417-424, 1993.

5 Drexler, K. Eric, Engines of Creation. Anchor Press/Doubleday, Garden City, New York, 1986.

6 Drexler, K. Eric, "Nanotechnology and the Challenge of Space Development," Proceedings of the 1988 International Space Development Conference, 1988.

7 Drexler, K. Eric, "Molecular Manufacturing as a Path to Space," Prospects of Nanotechnology, Markus Krummenacker and James Lewis (eds.), John Wiley & Sons, 1995.

8 Bishop, Forrest F., "The Interworld Rapid Transit System," submitted to Journal of the British Interplanetary Society, (draft) 1996.

9 McKendree, Thomas L., "Implications of Molecular Nanotechnology Technical Performance Parameters on Previously Defined Space System Architectures," Nanotechnology, Vol. 7, No. 3, pp. 204-209, 1996.

10 Hurter, Arthur P., and Joseph S. Martinich, Facility Location and the Theory of Production, Kluwer Academic Publishers, 1989.

11 Boden, Daryl G., "Introduction to Astrodynamics," Space Mission Analysis and Design, Wertz, James R., and Wiley J. Larson (eds), Kluwer Academic Publishers, pp. 113-140, 1991.

12 Lewis, John S., Mining the Sky: Untold Riches from the Asteroids, Comets, and Planets, Addison-Wesley, 1996.

13 Drexler, K. Eric, "Molecular Manufacturing for Space Development: An Overview," Journal of the British Interplanetary Society, Vol. 45, pp. 401-405, 1992.

14 Drexler, K. Eric, Nanosystems: Molecular Machinery, Manufacturing, and Computation. John Wiley & Sons, New York, 1992.

15 Mackenzie, Bruce A., "Bootstrapping Space Communities with Micro Rovers and High Tensile Boot Laces (Tethers)," Space Manufacturing 10: Pathways to the High Frontier, Barbara Faughnan (ed.), AIAA, pp. 235-244, 1995.

16 Moravec, Hans, "A Non-Synchronous Orbital Skyhook," Journal Astronautical Science, Vol. 25, No. 4, pp. 307-322, 1977.

17 Rechtin, Eberhardt, Mark W. Maier, The Art of Systems Architecting, CRC Press, 1997.

18 Merkle, Ralph. C., "Self Replicating Systems and Molecular Nanotechnology," Journal of the British Interplanetary Society, Vol. 45, No. 10, pp. 407-413, 1992.

19 Merkle, Ralph. C., "Design Considerations for an Assembler," Nanotechnology, Vol. 7, No.3, pp. 210-215, 1996.

20 Freitas, Robert A., Jr., "A Self-Reproducing Interstellar Probe," Journal of the British Interplanetary Society, Vol. 33, pp. 251-264, 1980.

21 Freitas, Robert A., Jr., and William P. Gilbreath (Eds.), Advanced Automation for Space Missions, NASA CP-2255, 1982.

22 Maryniak, Gregg, "Self-Replicating Machines for Space," SSI, 1984

23 Reitz, G, R. Facius and H. Sandler, "Radiation Protection in Space," Acta Astronautica, Vol. 35, No. 4/5, pp. 313-338, 1995.

24 Johnson R D, Holbrow C, Space Settlements: A Design Study, NASA SP-413, 1977

25 Drexler, K. Eric, personal communications, 1995p.

26 Rechtin, E., Systems Architecting: Creating & Building Complex Systems, Prentice-Hall, 1991.

27 O'Neill, Gerard K., The High Frontier, Space Studies Institute Press, 1982.

28 Drexler, K. Eric, Peterson, Chris and Pergamit, Gayle, Unbounding the Future: The Nanotechnology Revolution. William Morrow & Company, New York, 1991.

29 Williamson, Ray A., "The Industrialization of Space: Prospects and Barriers," The Exploitation of Space, Michiel Schwartz and Paul Stares (eds.), Butterworth, 1985.

30 Bishop, Forrest F., "Open Air Space Habitats," Tools for the Next Millennium, Lance Chambers (ed.), CRC Press, (forthcoming).

31 McKendree, Thomas L., "Planning Scenarios for Space Development," Space Manufacturing 10: Pathways to the High Frontier, Barbara Faughnan (ed.), AIAA, pp. 254-264, 1995a.

32 Regis, Ed, Nano, Little, Brown & Company, 1995.

33 Cassenti, B.N., G.L. Matloff, and J. Strobl, "The Structural Response and Stability of Interstellar Solar Sails," Journal of the British Interplanetary Society, Vol. 49, pp. 345-350, 1996.