CRN's Chris Phoenix was asked by the U.S. National Academy of Sciences to prepare briefing papers for their recent committee sessions investigating molecular manufacturing. So far, we've posted Concepts of Molecular Manufacturing and Current Status of Molecular Manufacturing.
Today's entry is Applications of Molecular Theory:
Stability of molecules results from bonds that fasten atoms together. The most important type of bond for molecular manufacturing is the covalent bond in which electrons from each atom are shared between both. If a bond doesn't exist, the electrons will repel each other, forming a sort of cushion that keeps the atoms at a distance. When the bond forms, the atoms move closer together. There is usually a barrier to the bond forming: the atoms have to be forced together more closely than is natural for their unbonded state. Once the bond has formed, there is a barrier to breaking it. For a molecule to rearrange, there are two requirements: all energy barriers must be overcome, and the ending state must be lower energy than the initial state. Breaking a bond without forming a new one results in a very high energy state. Forming new bonds usually requires breaking multiple energy barriers simultaneously to free multiple atoms. These conditions are uncommon and easy to avoid in a wide range of molecular systems.
In assessing the stability of molecular machines, the question is not whether molecules can be stable—there are thousands of molecules that are known to be stable. These include proteins and nucleic acids, which are used both in biology and in the lab to build molecular machines. The question is whether molecular machines can be designed to be stable in the intended environment. Weak bonds and low-barrier rearrangements must be avoided, but existing theory is sufficient to predict many of these problems. The reliability of mechanosynthetic reactions can likewise be predicted.
More significant questions exist. Most mechanical systems (though not all) include sliding interfaces. Nanoscale friction is not well understood, despite demonstrations of graphite superlubricity and carbon nanotube bearings. Machine phase chemistry is not as developed as solution phase; although either can be used for molecular manufacturing, machine phase systems may have significant advantages. Although scanning probe systems have been shrunk to computer-chip size, they have not yet been designed at sub-micron scales or with molecular components. These are useful topics for research.
In systems that handle molecules individually, large-scale productivity depends on the use of large numbers of tools, which must be small. (A 100-nm tool might process its own mass in 100 seconds.) Self-assembly may be used to build large numbers of simple molecule-building tools; Seeman's DNA fabricator has demonstrated this. More complex tools could be built by augmenting or even replacing self-assembly with mechanical guidance from preexisting tools. A tool that can guide the formation of a duplicate tool under external control is not a big stretch from current accomplishments. Research into rapid actuation, parallel control, and development of a mechanical molecular "toolbox" to support general-purpose engineering of products and second-generation tools may not be premature at this point. These would significantly extend the flexibility of the construction tool and the utility of its products.
Once flexible self-duplicating molecular manufacturing systems are developed, the next step is to integrate them into larger systems. However, this will be largely an application of systems engineering, not molecular theory.
Tomorrow: "Molecular Manufacturing Challenges"