Nanomachine foundations
From Wise Nano
This page explains the basic nanoscale capabilities that will be buildable with molecular manufacturing.
Contents |
Foundation
Materials
The starting assumption is that the manufacturing technology can build high-bond-density covalent solids with atomic precision and engineered shapes.
For example, carbon-lattice mechanosynthesis.
Custom-sequenced crosslinked polymers.
Aqueous-built graphene/fullerene.
Friction
Low friction sliding and rotating interfaces can be built. This appears quite likely for carbon lattice, since superlubricity has been demonstrated in graphite (ref?), and buckytubes rotate and slide freely in and out of each other (Zettl).
Solvated biopolymers can have near-zero friction; for example, ATP synthase is just about 100% efficient.
Electricity
Electrical power is available. Carbon can be an excellent conductor (some buckytubes conduct ballistically, and sub-micron ballistic electron travel has just been demonstrated in graphene) or an excellent insulator (diamond insulates 2 GV/m, and some buckytubes are insulating).
Fuel cells require a porous membrane with negatively-charged pores and a hydrogen source. N-terminated diamond should make a good membrane. Of course biomolecules will also work. Hydrogen can be obtained by pulling apart molecules mechanochemically, or breaking them apart catalytically (requires trace metals) or enzymatically.
Mechanical tunneling junctions should make good switches, since a separation of a nanometer can vary the current by orders of magnitude.
Electrostatic force is quite powerful at small separations; a reasonable voltage for electrostatic actuators is on the order of 5 volts.
Efficiency
At the macroscale, inefficiency shows up as friction or electrical resistance. At the nanoscale, inefficiency shows up in two ways: as drag on moving parts (including electrons) from random interactions, and as larger-scale losses from irreversible state transitions.
There are two ways to be efficient at the nanoscale. One is to let everything mush around, with long floppy chains which have lots of freedom to find minimum energy configurations. As long as everything moves slowly so that everything can get out of each other's way, the system stays at the bottom of a very shallow energy curve. Biosystems use this method.
The other way to be efficient is to be a control freak. If systems are built so that everything has a place and every mechanical part is stiffer than its surface forces, then the system does not need a lot of degrees of freedom in order to slide parts past each other without sticking. Stiff systems will have steeper energy curves, but as long as there is no actual cliff in the curve, they can still be efficient. Because the system is stiff, it relaxes more quickly, and can move faster with good efficiency.
Irreversible state transitions cost a fair amount of energy. If you let a spring go sproing, you lose its energy. At the nanoscale, almost everything is a spring--an entropic spring, meaning that it likes to have as much freedom as possible. For example, setting the state of a computer bit from "unknown" to "zero" takes energy, because it reduces entropy: it compresses the entropic spring. Forgetting a bit--letting it go back to "unknown"--then wastes that energy.
Other kinds of irreversible state transitions can occur in sliding motion between some kinds of surfaces. If there are mechanical springs that go "sproing" as they slide past each other (and every atom can act as a spring) then the sliding motion will lose energy. However, surfaces can be designed so that each atom is more tightly held in position by its own surface than it is pushed aside by the opposing surface. In this case, the atom will move smoothly back and forth a bit as the opposing atoms slide past, but will never go "sproing" and waste energy.
Sliding surfaces also suffer dynamic loss. Random chunks of sound/heat energy called phonons, arriving at the interface, will be accelerated by the relative motion. This adds energy to them; but since they are random, the energy cannot be recovered and is lost. This is somewhat analogous to viscous force, in that it goes to zero as speed goes to zero. But it's not really analogous, so if you're a physicist, forget I said that. There are other weirder dissipation mechanisms. But basically, moving at 1 m/s (1 nm/ns) should be quite efficient.
For more on dissipation mechanisms, see Nanosystems chapter 7.
Optics
Diamond is transparent; graphite is opaque; biomaterials come in a wide range of colors.
A sub-wavelength, regular, dielectric grid is transparent. A semi-regular grid is a hologram.
Color in diamondish materials can be generated with diffraction, the same way butterfly wings do it.
Simple Machines
Electrostatic motors
In vacuum, a 50-nm disk with 2-nm electrodes can handle kilowatts per cubic millimeter of power. (At this moderate speed, efficiency is expected to be >99%, and 10 W can be heat-sunk out of a cubic mm of diamond.)
Digital logic
Mechanical rods with bumps on them are surprisingly good logic elements. A 1-nm rod, 66 nm long, can move at 10 GHz. (The speed of sound in diamond is 10,000 m/s, or 10,000 nm/ns, so 10 GHz isn't even close to sonic speed.) The bumps, of course, block secondary rods from moving depending on the position of the first rod. Put springs on the ends of the rods, drive the rods through the springs with cams, and voila, you're doing computation. Note this is thermodynamically efficient (reversible): if you take the load off the second rod before you unload the first, nothing will go sproing, and the springs can drive the cams in reverse, recoving the energy of compression.
