Home > Nanotechnology Columns > Center for Responsible Nanotechnology > The Nanoscale Blacksmith
Someday, perhaps soon, nanotechnology manufacturing will use nanoscale tools to build more nanoscale tools via computer-controlled fabrication of molecules, similar to the way a blacksmith can fabricate tools using only the tools on hand. Several factors contribute to the attractiveness of this approach, including scaling laws and various benefits of precise construction. Molecular manufacturing may enjoy quite high construction rates, thanks to the small size of the tool relative to the product, leading to the construction of large, high-performance, nano-featured products that could have revolutionary impact. Targeted development programs for achieving this approach are likely to be launched within the next five to ten years.
February 28th, 2007
The Nanoscale Blacksmith
What is molecular manufacturing, and how does it relate to nanotechnology?
Nanotechnology covers a broad range of fields, from biotechnology to computer chips. Many interesting kinds of nanotechnology involve unusual physical phenomena that only show up at the nanoscale--things like quantum electron effects that can make gold change color. Nanotechnologists have been searching for these effects, and sometimes turning them into products, for quite a few years now. But they are difficult to find, to study, to understand, and to apply.
Another family of nanotechnology involves nanoparticles used as structural materials. Carbon nanotubes have been used to reinforce tennis rackets; flakes of clay only a few atoms thick have been used to make plastic more leak-proof and fire-resistant. The nanomaterials are typically inert, and often provide only incremental improvement over larger and less engineered materials.
A third kind of nanotech involves putting molecules together in ever more creative ways. Chemists have been building molecules for centuries, but new molecules require intricate recipes developed through trial and error. Each step of the recipe wastes some of the inputs to create unwanted chemicals, and so the recipes cannot include very many steps. To compensate, the chemists have developed an amazingly broad library of techniques to choose from.
There are many other kinds of nanotechnology: for example, modern computer circuits are small enough to qualify, and there's some very exciting work going on in medicine. But in this essay, we will focus on molecular manufacturing, a kind of nanotech that falls somewhere between nanomaterials and molecules.
Molecular manufacturing proposes to build nanoscale mechanical systems, for the purpose of mechanically guiding the chemical reactions that will build more nanoscale mechanical systems. A few reactions have already been carried out under direct mechanical control, using relatively huge machines--scanning probe microscopes and micro-imprinting stamps. If the machines can be made a lot smaller and more capable, then they may be able to build molecules into structures--even physical duplicates of the tools used to build them.
Tools being used to build duplicates of themselves is not a new concept; a blacksmith could copy just about every tool in his shop using only the tools already there. But there are some new things in molecular manufacturing: precision, automation, scale, and performance. Together, these factors could make molecular manufacturing as revolutionary as the computer or the steam engine.
Atoms such as carbon like to bond to each other in specific arrangements. Pull an atom slightly out of place, and it will spring back. A particular bonded configuration of atoms is called a molecule, and two molecules with the same structure will be effectively identical. A component that is manufactured as a single molecule will have exactly the same dimensions as its copies, unless an atom is grossly out of place forming a noticeable flaw. The difference between manufacturing tolerances and molecular precision is like the difference between a gallon of milk and a carton of eggs: there's never precisely one gallon in the jug, but it's easy to tell if an egg is missing. When it comes to manufacturing, this precision is highly important, because it means that (unlike all manufacturing techniques to date) a copy of a copy of a copy of a component can be just as precise as the original, without any need for measuring. Even if the fabrication operations are not 100% precise (which they will not be, thanks to thermal noise), if they are close enough, then the newly placed atoms will spring into their proper positions.
Precision should make automation easier. Not only are all parts built alike, but they will not suffer wear or fatigue the same way metal and plastic parts do. A molecular machine that is distorted by moderate stress will simply spring back unharmed, without ever suffering fatigue or ablation. A machine could do a huge number of operation cycles without changing its behavior in the slightest. Using precise mechanical components also implies an ability to exclude contaminant molecules absolutely.
Human technology is just beginning to build machines at the nanoscale. A few motors have been built, out of metal-encrusted nanotubes, out of painstakingly-designed molecules, or out of components copied from biology. A very few mechanisms have been built--things like levers and wheels. But nanoscale machines are a very attractive target for several reasons. Once we can build machines intricate enough to carry out computer-controlled manufacturing operations, we will be able to build a much broader range of machines.
In many ways, small precise things work better than larger things. Performance of nanoscale machines benefits from a few basic facts of physics called "scaling laws." One of the scaling laws is operation frequency: small machines take less time to do an operation than larger machines. In fact, the operation frequency is inversely proportional to the size: if you could shrink a machine a millionfold, to a scale of nanometers instead of millimeters, it would do a million times as many operations per unit time. Functional density scales with volume: a millionfold shrinkage would lead to a quintillionfold increase in components per volume.
Calculations show that power density should scale with operation frequency: a millionfold smaller machine should have a million times the power density of its larger cousin. It would have a quintillionth of the volume, and thus only a trillionth of the power. But connect a trillion of the tiny machines in tandem, and you'd have the same total power in one-millionth the volume. Of course, such designs would have to be reliable beyond today's dreams--but remember that molecular machines will not suffer wear. And if one of them did happen to break, say from random background radiation, performance would decrease by one-trillionth.
The final calculation from scaling laws is the one that makes all this possible. Atoms are extremely small. A modern scanning probe microscope, trying to build itself one atom at a time, might take six quintillion years. But if that microscope could be shrunk, then not only would the number of atoms required shrink very rapidly, but the microscope itself could work more rapidly. The combined effect is that the relative throughput--the time required for a machine to make its own mass of product--shrinks as the fourth power of the size. That means that a 100-nanometer scanning probe microscope could, in theory, duplicate itself in 100 seconds.
In addition to scaling law benefits, there are two other sources of performance improvement. One is that flawless molecular structures can be far stronger than macroscale materials. Carbon nanotubes have a theoretical strength 100 times that of steel. The other source of performance comes with the tongue-twisting name of superlubricity, meaning that under certain conditions, flawless flat surfaces can slip over each other with extremely low friction--and no lubrication required.
From atoms, to molecules, to machines, to manufacturing systems that can duplicate themselves using simple molecular feedstocks: this overview has explained only the broadest outline of the approach and its merits, and has not touched on the problems of design, control, and reliability. Those problems appear to be solvable with ordinary research and development. Other problems are somewhat more esoteric and less well understood: for example, it is not yet known exactly how to mechanically manipulate small molecules to make them join into larger functional structures. There are some practical problems that arise at the nanoscale (including many quantum effects and thermal noise) that will complicate the design process. No one has said that developing this technology will be easy.
There are, of course, skeptics who warn that the practical problems may be insurmountable, or that there may be unexpected theoretical limitations. Further research will tell. The quality of skepticism has improved over the past few years, as molecular manufacturing captures more interest; as a result, a number of rather flimsy skeptical arguments have been abandoned. Fifteen years after the major theoretical work in the field was published, no one has yet identified a factor that would preclude the building of high-performance kilogram-scale products composed of nanometer-scale machinery, using a few hours of time in a kilogram-scale factory. The implications of this capability would be transformative, disruptive--perhaps terrifying--and will be the subject of several future columns.
As skepticism recedes, what is left is a wide-open opportunity. Molecular manufacturing will not replace other forms of nanotechnology--there are too many useful phenomena waiting to be uncovered at the nanoscale. But for straightforward mechanical work, molecular manufacturing may be the most effective approach, with the highest power density and efficiency. Mechanical operations at the nanoscale can implement sensors and computers as well as motors. And the mechanical approach, relegating nanoscale complexities to the lowest levels of structure and preserving precision at all higher levels, should contribute substantially to simplicity of design.
In January of this year, a group of about two dozen scientists got together in Britain for the purpose of inventing research projects that could lead to software-controlled molecular fabrication. Although the scientists involved in this "Ideas Factory" did not in general share the vision presented here, they nevertheless applied themselves to the goal of developing research projects that could implement computer-controlled fabrication at the molecular scale. They developed three proposals to be funded and completed within just a few years: one to implement a sort of artificial DNA-programmed ribosome, one to manipulate and join molecular structures under precise mechanical control, and one to understand the detailed workings of scanning-probe chemistry. In this author's opinion, successful completion of these projects will take us a large fraction of the way toward a general-purpose molecular manufacturing capability.
In theoretical terms, molecular manufacturing is not the most interesting branch of nanotechnology; however, in practical terms, it may well be the branch with the highest impact. The next decade or so should be quite exciting. We expect the cost of molecular manufacturing development to fall rapidly, and multiple competing development programs could spring up in the next few years (covert programs may already exist). The biggest barrier to molecular manufacturing at this point is simply the lack of general understanding that its development is both feasible and valuable...and that barrier is rapidly crumbling.
For more information, see http://www.crnano.org .
Copyright 2007 Chris Phoenix. Released under Creative Commons Attribution-Sharealike.