MATERIALS OF NANO MACHINERY

Drexler is setting up an argument for the use of materials other than diamond, at least in early machines. In a previous entry, he has argued that he never advocated starting with diamond. While this may not be the best way to motivate the researchers who are currently pursuing his ideas in diamond, and while he certainly talked about diamond in several published works, including Engines of Creation, Nanosystems, and the Burch/Drexler nanofactory animation, it is worth looking at his current suggestions, since they may be (I say may be) the most efficient path to high-performance nanomachines and exponential manufacturing.

On his blog, Drexler makes it clear that this new equation applies to machines building structures of the same material that the machines themselves are built out of. This raises a question that has been answered at least in part for diamond-based machines, but not for other materials: whether the material can implement all the functionality that will be required of a complete nanomachine system.

If all that is wanted is nanoscale structures, then almost any material will do. But for a complete general-purpose manufacturing system, several different functions will be needed. One of the most basic is sliding surfaces, i.e., bearings. (Although it is theoretically possible to build a pantograph-style robot using only rigidly fastened springs, I would not want to try it.) Drexler mentions that materials chemistry often becomes more tractable as bonds become more polar or ionic (like salt and many minerals), as opposed to covalent (like carbon-carbon bonds). But I wonder whether such materials will be more likely to transfer atoms between surfaces.

I have said for a while that silica looks like potentially a good way to bootstrap to diamondoid. Silica can be made under water by protein machines, but is fairly stiff and quite modular (one might almost say digital) on a molecular scale. But carbon, depending on how it's connected, has a much wider range of properties than silica. A carbon-based machine could probably use electricity, since diamond is an excellent insulator and some buckytubes are excellent conductors. A silica machine would probably have to be purely mechanical.

I will be following Drexler's material suggestions with interest. They will become even more significant if it turns out that he has done as much work on machine designs using those materials as he once did with diamond.