Today and tomorrow, we're reporting on presentations at an important conference on Productive Nanosystems: Launching the Technology Roadmap. Chris Phoenix is providing live blog coverage for us...
Sixth talk: Eric Drexler. Drexler is the one who started the idea of molecular manufacturing back in the mid-1980's.
The general focus of the Roadmap is on atomically precise technologies, not productive nanosystems. That's because the former is a necessary foundation for the latter. To engage researchers and encourage development, the roadmap focuses on the former. It provides merit criteria and metrics for research today. When selecting between proposals, look for atomic precision. Look for size, range of materials, other criteria that we'll probably hear about later in the talk.
The Roadmap looks toward advanced manufacturing (what physics says should be possible), but focuses on accessible productive nanosystems (such as ribosome-like systems).
Quantity of material is important: with tiny manufacturing capacity, you can make a few sensors. With large-scale manufacturing, you can address things like global warming. It's important to look at scenarios where the roadmap succeeds in developing such objectives. But for now, focus on near-term things.
Near-term, there are several kinds of atomically precise things we can build. One is biopolymers: protein, DNA. Very large design space available here. But proteins are hard to design. Proteins are not squishy and soft, like meat - that's mostly water. Think of cow horn, silk... protein could have the properties of epoxy. Proteins are useful for catalysts, precise alignment...
DNA doesn't have as large a range of functions as proteins. You can make mechanical structures with it. 3D structures, 2D structures with complex edges. NanoRex is working on structural DNA design using Paul Rothemund's "staple" approach. So you can design a million-atom, 100-nm diameter, atomically precise, 3D structures. If you had a DNA synthesizer in-house, you could design a structure and build it in one day... 50 billion copies. This appears to be useful for building circuit boards. Zinc finger proteins can bind to specific DNA sequences, which implies you can attach things to these DNA structures.
Another class of precise things is specialized structures, where each one has to be synthesized separately. These are non-modular and tend not to have a lot of design freedom. But the range of function is almost unlimited: catalytic, electronic, mechanical, optical...
So the goal of all this capability (bought with multiple $billions) should be to integrate these components to build systems with hundreds to thousands of distinct 3D components, using atomically precise scaffolding and binding elements. Biology has this kind of integration: protein with nucleic acid with other stuff.
New topic: Advances in production technology. Type 1 advances build better products. In Type 2, the products include improvements to the production system, which can enable further improvements. So we really want better productive machines that can build better productive machines... This appears to be an argument for using nanosystems as the means of production of nanosystems.
Today, tools build tools that build tools... traceable back to blacksmithing. The tool that extruded your breakfast bagel is a leaf on this tree. The advanced APM tree has a "Mark II Ribosome" low on the trunk, and "Macroscale APM" high on the trunk, with "Dollar-per-kilogram fab" among the leaves. People tend to assume that things high in the tree are proposals for next year, "which would be absurd."
The Roadmap talks about cross-linked organic structures. An idea that arose pretty late is mixed covalent-ionic bonding. Titanium dioxide, quartz. This may be closer than what's been looked at more closely.
The role of roadmapping: Developing the knowledge and confidence necessary for coordinated system development. So the Productive Nanosystems roadmap should show what's necessary, when, how to coordinate and schedule developments. Avoid chicken-and-egg problems that lead to slow incremental progress.
DNA currently costs dollars per milligram. There's no point in thinking about kilogram-scale structures... but there's a researcher who has an idea for making DNA at dollars per kilogram... but why should he do it when there's no market for kilograms of DNA? This is a real example: it seems that DNA might actually get vastly cheaper.