MM high performance systems
From Wise Nano
Contents |
High Performance Nano and Micro Systems
Basic scaling analysis and theoretical designs using known materials indicate that the ultimate performance limits of nanomachinery may be many orders of magnitude higher than is achieved by either biology or contemporary machines including computers. This section focuses on high performance nanoscale designs; the next section addresses the integration of these designs into large products.
Diamond, graphene, and fullerene can be produced by a wide range of reactions. New low-temperature records are continually being set for fullerene synthesis, and solution-phase synthesis of C60 (buckyballs) is being pursued. The flexibility of “dry†chemistry appears well-suited to building diamond lattice by incremental deposition. Accordingly, this section will assume that a high performance nano-manufacturing technology uses materials of diamond-like strength. Much of this section is based on calculations from Nanosystems.
This section includes discussion of the convenience of surface forces for manipulating micron-scale components; a mechanical fastening system that requires no external manipulation; an efficient class of bearing; electrostatic actuators, including a reversible motor/generator; digital logic; and mechanochemical power conversion.
Surface forces and component manipulation
Due to electron interactions, an attractive force will develop between almost any objects in close proximity. The force between objects in close contact is on the order of 1 nN/nm2, though it decreases rapidly with distance.20 For a cubic micron part, massing on the order of 10-15 kg and weighing ~10-14 N, the contact force of a single square nanometer provides 104 times the force of gravity. Although the force can be screened by liquids, in gas or vacuum it provides a convenient means for handling or attaching micron or sub-micron blocks.
The force between two surfaces in close contact can approach 5% of the tensile strength of diamond. This indicates that closely fitting surfaces which must be pulled apart must be strongly built. However, the force falls off rapidly with separation. According to calculation, separating two surfaces by 0.2 nm (~ 1 atomic diameter), for example with atomic-scale spacer bumps, will reduce the force by almost an order of magnitude.
The potential energy per square nanometer of two diamond surfaces touching (0.2 nm apart) vs. separated is about 225 zJ. A contact area of less than 1 nm2 should be adequate to hold against thermal vibration. If materials with lower Hamaker constant are used, or spacers are added to reduce material strength requirements, then the area would need to be increased. For example, with 0.2 nm spacers, the energy per square nanometer is about 56 zJ, and 2.2 nm2 of contact area would be needed.
These numbers indicate that manipulation of micron-scale blocks does not require mechanical grippers. Contact pads of a few square nanometers area can implement the functionality of grippers. As calculated above, an area of a few square nanometers is sufficient to hold a block against thermal noise. By extension, a difference in area of a few square nanometers is sufficient to determine which of two opposing "grippers" a part will stick to when the grippers (pads) are pulled apart. Assuming that binding force and energy are proportional to area for all pads, let A be the area required to hold a part against thermal noise. Then a simple approach to setting down a block and picking it up again would be to put three contact points, each of area A, on the end of a manipulator, and install contacts of total area 2A on any convenient wall or other fixed surface. To stick the part to the wall, move it into contact, and then withdraw the manipulator's contact points one by one. To remove it from the wall, simply retract the manipulator with all three contacts engaged, and the block will come loose from the wall.
More elegant designs can be easily imagined. If the manipulator is capable of bending motions, then rotating its pads on edge will allow them to disengage from the block, while a combination of translation and slow bending will rotate the whole block and disengage it from the wall. Alternatively, a small plunger can be used to push the block away from a large pad. A sphere-and-cup contact can allow the block to rotate; six struts of variable length with sphere-and-cup contacts could be used to manipulate the block like a Stewart platform.
Mechanical fastening: Ridge joints
For some applications, such as fastening micron-scale blocks together without complicated manipulations, it may be desirable to use a strong mechanical joint that requires low insertion force, but activates itself to lock in place. Surface forces can be used to power the mechanism of the joint. One such joint is the “expanding ridge joint.â€
Each mating surface is covered with small "ridges" that are roughly triangular in cross section. See Fig. 6. All exposed surfaces are non-reactive (e.g. hydrogen-passivated diamond). The ridges on each face interlock with the ridges on the opposing face. As the joint is pressed together, the ridges split and expand sideways. The sloped faces of the ridges are not smooth, but are shaped to grip the opposing ridge, with scallops deep enough to form overhangs when viewed perpendicular to the block face. A scallop is chosen instead of a sawtooth or ratchet profile in order to avoid crack formation at sharp concave angles. Scallops also make assembly motions smoother, and allow the un-powered assembly described below. The expansion of the ridge opens a space in its center, which is then filled by a shim which sits above the almost-closed gap between the two halves of the ridge. Once the shim is in place, the volume of the joint cannot easily be compressed, and the surfaces of the ridges cannot easily slide past each other; pulling apart the joint would require compressing a solid mass of diamond by several percent or breaking at least half of the ridges simultaneously. If the ridges all run in the same direction, the joint may be able to slide freely. Crossed ridges will produce a joint that is quite stiff against shear.
Fig. 6. Operation of a ridge joint. Ridges are pulled apart by surface forces without external actuation or manipulation.
As opposing ridges approach each other, the scallops pass in close proximity. Surface forces between the ridges will tend to pull the ridge apart. When the ridges are fully pulled apart, surface forces can pull the shim into position and hold it against thermal noise. Careful design to balance the surface forces may allow this approach to work with as little as 12 nm2 of ridge surface. If the shim is retained from entering the gap, then the mechanism will form a weaker and reversible joint, useful for placing blocks in a temporary configuration.21
Bearings
Although the granularity of atoms makes a perfectly smooth surface impossible, low-friction bearings can still be constructed out of stiff materials. Mechanically, atoms are soft, rather than hard-surfaced spheres; bonds are also somewhat compliant. Atoms overlap when they bond, reducing the irregularity of the surface.
If the atoms on two facing surfaces have the same pattern and spacing, the atoms of one surface will fit between the atoms of the other, requiring high force to slide the surfaces. To prevent this, the atoms can be placed out of register, with different spacing or orientation. In this case, the transverse forces on the atoms will almost completely cancel, leaving a very small net force (and hence very low static friction) at any displacement. Superlubricity is a condition of very low friction between surfaces with atoms in non-corresponding positions. Superlubricity has recently been observed between graphite sheets rotated to out-of-register angles.
Drexler has proposed that nested hoops with different numbers of atoms on the inner and outer surface should show a similar effect, especially if the least common multiple of the numbers is large. Nested carbon nanotubes have been observed sliding and rotating freely, an apparent example of this prediction. This would allow building low-friction rotational bearings.
Because atomically precise surfaces can slide past each other without stiction or wear (and for some surfaces, also with low drag), there is no need for lubricants. This is fortunate because a single atom is larger than the gap between two flat stiff surfaces. (This implies that such surfaces form a sliding impervious seal.) Without lubricants, the perpendicular stiffness of a sliding bearing is high, being a function of surface forces between two stiff surfaces.
Another kind of bearing uses a covalent single bond. This is only suitable for relatively low loads, but may be useful in some small machines. It is expected to have especially low drag.
Efficient nanoscale bearings are expected to have effectively zero static friction. The net force exerted by a bearing surface will usually be far smaller than the forces used to drive the machinery. Low-speed systems may use such a small driving force that this is not the case; however, the energy barriers created by bearing forces will be lower than thermal energy. This means that even with a vanishingly small driving force, thermal energy will move the system past the barriers. Systems with no static friction can be run as slowly as desired. Dissipation increases at least linearly with speed, so slowing down a system by a factor of 10 will allow it to dissipate at least 10 times less energy in performing the same operation. Some of the orders of magnitude increase in nanoscale manufacturing throughput and power density can thus be traded for significant improvements in efficiency.
Electrostatic actuators
At large scales, electrostatic actuators require high voltages and have low power density. However, a potential difference of 5 volts across a gap of 4 nm will produce a force of 7 pN/nm2, or 1 nN from a 12 nm square plate. This is a usefully large force for nanosystems.
An efficient 10-volt electrostatic motor can be built on a 50-nm scale. With a rim speed of 1000 m/s, which is within diamond breaking strength, the power density of such a motor would be awesome: 1015 W/m3. Even at a 1 m/s rim speed, 1012 W/m3 is several orders of magnitude better than any existing natural or manufactured motor. By running it in reverse, the same device would become a generator; in fact, its design is thermodynamically efficient (“reversibleâ€, in the sense of reversible logic).
Diamond is an excellent insulator. Some carbon nanotubes are excellent conductors. Thus an all-carbon system would not be limited in its ability to handle and use electricity.
Digital logic
A lower bound for the performance of digital logic can be set by a simple, easily analyzed, purely mechanical design. Nanoscale rods that move to block the motion of other nanoscale rods can implement logic gates. A logic gate built with this approach might occupy a volume of 10 nm2, switch in 0.1 ns, and dissipate less than 10-21 J per operation. Computers could perform 1016 instructions per second per watt.22
The mechanical approach to logic is relatively insensitive to material choice. Because it does not rely on electronic effects, components can be packed tightly without limits imposed by electron tunneling. Error rates can be extremely low, because an error would require a logic rod to slip past a physical obstruction.
Mechanochemical power conversion
In reactions between small molecules or non-stiff components of large molecules, the bonding force drives the reaction to completion quite rapidly. However, if reactants are stiffly held, the reaction can constrained to move slowly through the intermediate stages. This will exert a force on whatever is holding the reactants, and energy can be extracted from this force. Because this is not a heat engine, it is not limited by Carnot efficiency; in theory, nearly 100% of the chemical energy can be recovered. Electrochemical conversion, carried out by fuel cells, also is not Carnot limited. Drexler estimates that the feasible power density of mechanochemical energy converters is on the rough order of 109 W/m3.
