Incentives and Timeline for Development

We estimate that with sufficient effort, a general-purpose manufacturing system using carbon-lattice chemistry might be developed as early as 2010.

The steps to develop this technology are straightforward:
1) Verify the basic soundness of the proposal by review of the theory and preliminary investigation of practical problems.
2) Develop a set of diamond-forming mechanochemical reactions, including detailed tool tip models.
3) Design a diamond-forming manipulator using nanoscale diamond parts.
4) Develop a bootstrapping technology that can build the first manipulator (there are several options).
5) Develop libraries, design rules, and CAD software for products at various scales, including nanofactories.

These steps can be taken in order, although for most rapid development some of them should be overlapped. The success of the endeavor can be evaluated after each step and is almost guaranteed after step 2, limiting the potential cost of failure. The total cost to develop this new manufacturing technology would be quite large, probably in the billions of dollars, and the rate of expenditure for a rapid early program could be quite high.

Each of these steps should require less than five years. Most of them can be done in parallel. The cost and difficulty will drop rapidly due to improvements in enabling technologies such as scanning probe microscopy and chemistry simulation. At a rough guess, the cost to complete all these steps by 2010 might be $10 billion. The cost to complete them by 2015, starting in 2010, might be under $1 billion. However, as explained below, early development may be worth a high cost.

Diamondoid molecular manufacturing is expected to be orders of magnitude better than rapid-prototyping systems, lithography, or biomimetic engineering.

Carbon lattice--diamond and buckytubes--forms the strongest known material. No other general-purpose manufacturing system can produce this material.

The feature size of molecular manufacturing is naturally a few atoms wide--less than a nanometer. Rapid-prototyping systems and lithography will not achieve this for many decades. The atomic precision of biomimetic engineering is blunted by the process of shape formation: the smallest features, such as alpha helices and beta sheets, require hundreds or thousands of atoms.

With small features comes compact functionality. As Feynman said, there's plenty of room at the bottom. A small CPU with nanometer-scale logic elements could fit inside a single transistor of today's computer chips. The use of strong diamond materials also allows extremely high power density: a car engine could fit into a cubic millimeter.

Scaling laws and preliminary architecture studies indicate that a tabletop factory should be able to produce its mass in approximately an hour. Molecular manufacturing is the only manufacturing technology currently contemplated that would be able to directly fabricate manufacturing systems. The ability to produce new manufacturing capital so rapidly has large economic and strategic implications.

A molecular manufacturing system could build products several decades ahead of competing technology road maps, producing enormous value and rapid industrial bootstrapping.

No other technology combines nanoscale features with diamond-class materials. Products that can be built directly with this technology include computers about nine orders of magnitude ahead of today's semiconductors; cheap, compact arrays of medical sensors and microsurgical instruments; and aerospace hardware and structure saving 90% or even 99% of the weight of today's systems. Molecular manufacturing could make such products available decades ahead of any competing technology.

A self contained, fully automated, general purpose manufacturing system would be "appropriate technology" for almost any environment. If the factory and the raw materials were reasonably cheap, it could out-compete most other manufacturing of products in its domain. The raw materials will be small organic chemicals, and the ability to duplicate its own structure will make the factory as cheap as any product. This indicates that molecular manufacturing could rapidly dominate and/or create the manufacturing infrastructure for a wide variety of high-tech products.
 

Ecological Impacts of General-purpose Molecular Manufacturing

Small products: The ability to build small-format products intended for use in unconfined environments, including medical and surveillance devices, implies the accumulation of nano-litter. The smallest devices could be considered nanoparticles.

Although the proposed manufacturing system would not involve small-format free-floating devices, it could be used to manufacture small products as well as larger human-scale products. Small products will be useful in at least a few applications, such as surveillance and (with sufficient additional research) some medical applications. The simplest products might be as small as 100 or 200 nm, and would be difficult to collect after use. Note that this does not imply a "grey goo" threat, because such simple products would have no manufacturing ability. However, even inert nanoparticles may pose health hazards, and large accumulations of litter may lead to environmental damage.

Increased consumption: If manufacturing gets very cheap, people will use more products. High-tech products tend to use a lot of power.

Several factors of molecular manufacturing imply that factories and their products may become very cheap (aside from licensing costs). People will have little natural incentive to avoid unnecessary consumption. Computers and networks are already a major source of power use, and a proliferation of higher-tech products--including high-tech integration with traditional products--may be a major source of power drain. Solar cells are also expected to become cheap, but this raises questions of land use and microclimate disruption.

Weapons: Compact powerful products invite use as weapons. Cleanup after a military or terrorist strike could require new techniques.

Small computers, powerful motors, and intricate cheap manufacturing imply the ability to create whole new classes of weapons, especially antipersonnel weapons. For example, a lethal antipersonnel mine could be made small enough to make its cleanup more like decontamination than like minesweeping. Such weapons could be manufactured in great quantity, could be dispersed widely, and would probably be very attractive to terrorists. This is only one of a variety of unpleasant scenarios.

Environmental remediation: More advanced technology is usually cleaner. Technologically-intensive cleanup could become a lot cheaper.

Cleanup of spills may require large amounts of equipment rapidly deployed. It is possible that molecular manufacturing could be fast enough to manufacture such equipment on the spot. Even if this is not the case, the low cost of production would make it cheaper to stockpile and use the equipment. For some applications, the ability to cost-effectively build large arrays of small machinery may be useful for mechanical cleaning.

(Grey goo: A free-range self-replicating robot will be harder to design than a general-purpose manufacturing system. It is probably feasible with sufficient effort, but not an initial concern.)

The environmental dangers of self-replicating nanobots--"grey goo"--have been widely discussed, and it is widely perceived that molecular manufacturing is uncomfortably close to grey goo. However, the proposed production system of molecular manufacturing does not involve nanobots, but much larger factories with all the nanoscale machinery fastened down and inert without external control. As far as we know, a self-replicating mechanochemical nanobot is not excluded by the laws of physics, but such a thing would be very difficult to design and build even with a full molecular manufacturing capability. (MORE)
 

Broader Policy Issues

How could portable general-purpose manufacturing be regulated?

With programmable chemical manufacturing as the base technology, a general-purpose manufacturing system could be packaged as a home appliance. Such a thing would be very desirable and easy to smuggle. Product blueprints will be digital information, even easier to propagate untraceably. It is unlikely that restrictive regulations will suffice to prevent the use of such systems, at least by criminals. However, complete lack of regulation would allow too much damage to be done, either by collective misuse such as nano-litter or by individuals making and using dangerous or destructive products. Careful policy will be necessary to minimize undesirable use without fueling a black market.

The Nanotechnology Act of 2003 does not study molecular manufacturing technology, only molecular self-assembly. Can the EPA fill the gap?

The House version would have called for detailed study of molecular manufacturing. The final version calls for only a feasibility study of molecular self-assembly, which is a much more limited technology. This may lead to a policy gap. The final stages of development of molecular manufacturing may progress rapidly, leaving no time for careful policymaking. Since the technology may have substantial environmental impacts, a body such as the EPA may be well positioned to fill the policy gap. The first step would be a theoretical study of the capabilities of molecular manufacturing; this would be quite easy to do, since much of the theory is already laid out in Nanosystems.

How can public (and scientific) opposition to this field based on “exponential manufacturing equals self-replication equals grey goo” be mitigated?

A major source of opposition to molecular manufacturing and molecular nanotechnology is the popular association with so-called grey goo. As explained above, this association is largely unfounded; current plans for molecular manufacturing systems are nothing like grey goo. ("Assemblers", though mentioned in the 1986 book Engines of Creation, do not appear in the 1992 technical work Nanosystems.) The log jam of the current nanotechnology "debate" could be freed by public recognition that grey goo is not very related to molecular manufacturing. How could this be achieved?

Will successful regulatory policy require international cooperation?

A revolution in manufacturing could have a variety of effects that cross or ignore national borders. Some examples include nano-litter, a sharp increase in space flight, shifts in geopolitical relations caused by shifting financial or military conditions, and easy smuggling of undesired products or the means of producing them. It appears likely that international cooperation will be necessary to deal adequately with some of these issues.