Slashdot Mirror
The Nanomechanical Computer
eldavojohn writes "The BBC is reporting on a newly proposed type of nanomechanical computer that mimics J. H. Müller & Charles Babbage's work on mechanical computational devices — just on a much smaller scale. The paper is published today in the New Journal of Physics and cites three reasons to build a computer with nanomechanical transistors over bipolar-junction or field-effect transistors: '(i) mechanical elements are more robust to electromagnetic shocks than current dynamic random access memory based purely on complimentary metal oxide semiconductor technology, (ii) the power dissipated can be orders of magnitude below CMOS, and (iii) the operating temperature of such an NMC can be an order of magnitude above that of conventional CMOS.'"
This present design is a cool idea. I don't want to take anything away from the presented concept, but I thought it would be important to point out previous work on nanomechanical computers. First of all, Eric Drexler (the guy who popularized the term "nanotechnology" and who basically invented the field now known as molecular nanotechnology) has been advocating the concept of nanomechanical computers for many years now (they are described in his book Engines of Creation (1986) and detailed feasibility calculations, and rough schematics, are presented in his book Nanosystems (1992)). Drexler has been trying to get people on-board with his very foreward-looking ideas for nanotechnology: where nano-sized mechanical systems would be performing computation, and controlling chemical reactions with a precision that currently only biological systems can achieve. (It should be noted that current work in "nanotechnology" is hilariously primitive compared to what Drexler intended the term to describe.) Drexler's vision of nano-mechanical systems has been challenged by many people, most notably by Richard Smalley (the guy who discovered buckyballs).
Beyond Drexler's theoretical work, carbon nanotubes were demonstrated as nano-mechanical transistors in 2000. Basically, the nanotube was positioned over various electric pads. A current could be applied to mechanically deform the nanotube. The deformation was stable, and could be read-out by measuring current across the tube. Since the deformation was stable and reversible, the tubes could be used as persistent storage or as switching/logic elements. In fact, switching speeds of gigahertz were demonstrated. The vision was to have long nanotubes in a huge cross-bar architecture, leading to high-density persistent storage. As is often the case, scale-up was difficult.
This present work appears to pattern a nano-sized post between conducting pads (out of a gold/silicon layered system) , and to use that post as a single-electron transistor. The 'mechanical' part comes from mechanically coupling multiple pillars to use as a gain mechanism for a transistor. This is basically much closer to conventional micro-lithography, and as such, it should fit in with current lithographic infrastructure much more easily than the nanotube concept did.
Oh, and your premise is wrong: building a MEMS chip of a non-trivial size pretty quickly runs in the hundreds of thousands of $, even with educational discounts. So pretty much you have to get the design ready, then ask for funding to build the thing, which is what they are presumably doing.