MM incremental improvement
From Wise Nano
Contents |
Incremental Improvement
Although it could have practical applications, the main purpose of a primitive molecular manufacturing system would be to build a better system. The ultimate goal is an easy-to-use nanofactory which can rapidly build a complete duplicate nanofactory, as well as a wide range of high-performance products, from inexpensive feedstock. Although this is vastly beyond the primitive nanofactory, a set of incrementally developed capabilities can form a sequence of transitional improvements ending at the goal. The capabilities could be added one at a time, which would allow each capability to be developed and tested separately. This approach would also produce spinoff manufacturing systems which would be increasingly useful.
The capabilities are largely independent, so could be added in any of several sequences of development. To the extent it is possible to validate future designs, planning several steps ahead may guide design choices and allow faster development. Although detailed designs for each capability would be beyond the scope of this paper, this section explores some of the issues and techniques that might be involved in improving the nanosystems, removing the solvent from the nanofactory interior and product space, manufacturing blocks internally, and improving the mechanical design.
Improving nanosystems
Digital logic is not necessary for the earliest systems, but will quickly become convenient and then necessary as the system scales up. For example, multiplexers and demultiplexers can greatly reduce the number of wires that must be connected to the nanoscale. Molecular electronics is a rapidly developing field, with transistors being made out of carbon nanotubes and single organic molecules. Simple electrical switches operated by molecular actuators may also be used; the several-nm gap that would be left by removing a conducting nanoblock from a chain is too large for significant tunneling of electrons.
With more efficient and intricate mechanical systems, the manufacturing process can be made more flexible and functional without sacrificing efficiency. Actuators can be improved at any stage. Although bearings are high on the list of priorities, a lot can be accomplished with pantograph-like trusses. Covalent bonds do not fatigue; they can be flexed indefinitely without weakening, and the range of motion is considerably higher than with MEMS. Once stiff smooth surfaces can be constructed, a new type of bearing may be built; see “Bearings” in the "High Performance Nano and Micro Systems" section.
Removing the solvent
Solvent is convenient for diffusive transport, but creates drag. Once feedstock blocks are bound to manipulators, there is no need for solvent unless it supports some particular operation such as supplying ions for bonding or actuators (though the preferred actuators will not use ion movement but electronic changes). There are many ways of fastening blocks into the product that do not require solvent. Pure electrostatic actuators are excellent choices for a dry environment, and should also work in clean non-polar solvent.
A Silkscreen type factory could maintain solvent on the feedstock side of the membrane and low-pressure gas on the other side. In early designs, some solvent would presumably enter with the blocks or otherwise leak through. The solvent, operating temperature, and block material should be chosen to limit adsorption and promote evaporation in the “dry” interior; this implies a pressure difference between the “dry” and “wet” sides, which may limit the size of the membrane. (In later designs, with smaller feedstock molecules and better construction methods, it should be possible to operate without any solvent leaks.) Liquid xenon may be a good choice for a solvent. A major advantage of xenon is that it is chemically inert, so it will not combine with any reactive molecules used in mechanosynthesis.
Many reactions require no solvent; the combination of mechanically mediated reactions and lack of solvent is unfamiliar to most chemists, but has been demonstrated in the laboratory for a few reactions, and theory suggests that it should be a very rich chemical domain. Radicals that could not survive in solvent can be maintained in vacuum, held separate by mechanical control of their molecules until they are reacted according to plan. This should provide access to chemistries that require the maintenance of conditions far from equilibrium, and should facilitate the step-by-step construction of highly crosslinked covalent solids.
In the absence of high drag and friction, nanoscale structures have resonant frequencies around the gigahertz range. This implies that a dry nanofactory could work far faster than the wet versions, assuming that the actuators were responsive. Electrostatic actuators should be very fast, limited only by the current carrying capacity of the wires.
Manufacturing blocks
Blocks in the five nanometer size range are too large to be easily synthesized (in fact, they may be expected to be quite expensive), and do not diffuse as rapidly as might be wished. However, blocks made too much smaller might not have enough internal space to hold the desired functionality, and could reduce throughput. The nanofactory will be more efficient, its products will be cheaper, and it will be easier to design and create new block types, if small molecules can be used as feedstock, combined into large blocks internally, and the large blocks placed in the product layer. (A variant of this is to use one system to create and stockpile the blocks, and another to assemble them into the product. This requires essentially the same functionality as the integrated nanofactory.)
Small molecules will not have room for elaborate molecular attachment mechanisms, but there are many approaches to joining molecules. One approach is to use reactions that form bonds between structures of a few atoms when pressed together, but do not crosslink rapidly in solution. If the reaction releases a small molecule, the factory design must allow it to escape, but many reactions (for example, cycloadditions) rearrange bonds to produce a single product structure. Another approach is to begin with inert molecules, then make them reactive by removing passivating structures, and finally position them to bond spontaneously.
There is a wide variety of ways to make programmed parts from molecular feedstock. Even a preliminary exploration of the available reactions would require many pages. At one extreme, molecular fragments weakly bonded to "tool tip" molecules can be deposited on a depassivated surface in vacuum (or noble gas). This is called "machine-phase chemistry." The ability to use reactive species such as unprotected radicals, carbenes, and silylenes increases the chemical options, but requires extreme cleanliness.
In solvent, molecules can be forced to bond to selected locations on a terminated surface. This somewhat reduces the cleanliness requirements. It also allows the solvent to be used to deliver feedstock to the tool tip by Brownian motion. Simply holding the molecule in the right location can increase its effective concentration by many orders of magnitude. Holding it in the right orientation, or applying pressure or electric fields, can further speed its bonding to the surface, allowing the use of high-barrier reactions to minimize unwanted random deposition.
A chain polymer can be built which then folds up into a desired shape. (Post-folding bonding can be used with some chemistries to increase strength and stiffness.) The programmable control system need only select which monomer to attach next, and force it to bond to the chain. This is the function performed by the ribosome.
Component molecules can be used which hold their position by electrostatic or van der Waals attraction or hydrogen bonding.
Deposition reactions can happen through conventional, non-mechanical chemistry, with the location controlled by mechanically mediated protection or deprotection of deposition sites. In this case, protection can be either chemical or steric (physical blocking of a site).
It may be useful to build blocks out of molecules that are already structural or functional components. Alternatively, very small molecules could be aggregated to build arbitrary structures; this would make the factory extremely flexible, and allow for rapid improvements in product design without changing feedstock or factory design. Providing diverse functionality from a few simple molecules appears feasible, because conductors and insulators are all that are needed to build electrostatic actuators. With actuators and mechanical switches, digital logic and sensors can be built.
Improving the mechanical design
Directed internal transport of product components within the factory would be useful for several reasons. It would allow broken machinery to be bypassed. When a factory that manufactures blocks internally is forming a product with large voids, internal transport would allow the transfer of blocks from the entire factory's block-manufacturing machinery to dense regions of the product, alleviating a potential bottleneck.
Specialized block-manufacturing equipment would increase the efficiency and speed of the molecular manipulations. Instead of programmable robotic systems, fixed-path machinery could do common operations to combine small molecules into frequently used structures. A well-designed “mill” would use only one degree of freedom in its operations, making it straightforward to power and synchronize with other machines.
