Kinds of molecular manufacturing
From Wise Nano
Molecular manufacturing (MM) means using precise molecular machinery to build precise molecular machinery: "nano building nano." There are many ways, at least in theory, to do this. This page lists them, with brief analysis of the product performance and the difficulty of developing the technology. There are several categories, each of which has several possible molecular product families.
To a first approximation, all that's necessary is:
- Structure that can transfer force and implement sufficiently precise positioning
- Actuation that can control the chemistry
- Digital logic that can run the process.
The ways include:
- Machine-phase chemistry, in which all reactive molecules are controlled and positioned.
- Solvent-phase mechanosynthesis, in which the position of the reaction is controlled by positioning at least one of the participant molecules.
- Non-covalent assembly, in which molecular blocks are positionally assembled and hold their position by electrostatics, surface forces, hydrophil/phobicity, etc.
(Note: In theory, structures could be created by patterns of electric charge or light intensity, and those patterns might be switched/created without requiring physical actuation. It's not clear whether this approach is significant or whether it should count as molecular manufacturing. It was conceived when the first contributor to this page (Chris Phoenix, CRN 20:18, 19 Dec 2004 (CST)), while writing, asked himself whether actuation was strictly necessary to make molecular patterns. Electric fields and light are soft-edged, but when combined with a templating surface, could be used to guide molecular deposition precisely. One reader (Michael Vassar) suggests that Forrest Bishop may have had a similar idea previously. Investigation of prior work and present implications is ongoing.)
Contents |
Machine-phase chemistry
Diamondoid
"Diamondoid" has been used by Drexler to include "less diamondlike diamondoids." Since this is open-ended, preliminary analysis should concentrate on repetitive carbon lattices, recognizing that each new atom introduced may need a new tool tip and several new reactions. But a machine-phase system that can build diamondoid can likely also be made to build graphene/fullerene without too much difficulty, as long as the parts under construction don't get too floppy.
Product Performance
Very high, as analyzed in Nanosystems.
- Electrostatic motors 10^15 W/m^3
- Nanocomputers 10^16 instructions/sec/W
- Digital logic 10^26 gates/m^3
- Tensile strengths >5x10^16 GPa
Nanofactory Performance
Kilogram-scale factory should make its mass in ~1 hour
Difficulty of Development
High. Probably requires machine-phase chemistry, which is poorly understood and controversial (though much of the controversy is politically motivated). There are at least two bootstrapping approaches. One is to build an easier MM and use it to do machine-phase chemistry. The other is to use scanning probe technology to build the first nanoscale programmable machine-phase diamondoid synthesis machine. (It might also be possible to build the first machine using ion etching, EBID, multi-photon polymerization, etc., but these have not been investigated.) No synthesis path has been worked out in detail. Experiments to date are encouraging but scanty.
Boron Nitride
Product Performance
Probably a large fraction of diamondoid.
Nanofactory Performance
Not yet analyzed.
Difficulty of Development
Less studied than diamondoid, but probably easier because the alternating atom types in the crystal would require lower precision to avoid unwanted side reactions in deposition.
Solution-Phase Chemistry
Graphene/Fullerene
222-atom engineered graphene sheets have been built in solution. Also, "buckybowls" (half a buckyball) have been built.
Product Performance
Probably a large fraction of diamondoid.
Nanofactory Performance
Not yet analyzed. Probably significantly slower than machine-phase due to drag from moving through the solvent.
Difficulty of Development
Not studied. But might not be too hard.
Silica
An enzyme that can catalyze the deposition of silica has recently been found. This raises the possibility of directed silica deposition to make machine-like structures.
Product Performance
Probably a small fraction of diamondoid, but the comparison is worth making. Minimum feature size might be bigger, reducing the shrinkability of digital logic.
Nanofactory Performance
Not yet analyzed. Probably significantly slower than machine-phase due to drag from moving through the solvent.
Difficulty of Development
Not studied. But might not be too hard.
Molecular Building Blocks
This is a wide range of possibilities. Merkle suggested something with boron and nitrogen. There's a carbon-silicon branched (tetrahedral) polymer that self-assembles around micelles; if that could be deprotected in selected sites, it might be very flexible and relatively easy to build with.
Non-Covalent Assembly
This has not been much looked at.
