I’ve updated “The Physical Basis of High-Throughput Atomically Precise Manufacturing”. Not a big change, but I expanded the discussion of reliable molecular modeling of selected, highly constrained systems, along the lines discussed here: “Making vs. Modeling: A paradox of progress in nanotechnology”.
The Trajectory of Technology
Previous post: About releasing building blocks…
Next post: How to Learn about Everything in Belorussian
Oxford Lecture Video
{ 2 comments… read them below or add one }
Dr Drexler,
Slightly off topic, but I have been thinking about APM’s poster child – the desktop nanofactory. While the concept seems technically sound, I think a multi tier scheme is more likely, and I am interested in your opinion. (And pardon me if similar schemes has already been proposed.)
In a multi tiered scheme, nanofactories would be large-scale, centralised, secure facilities that would produce micro+ scale components only – microblocks – in volume. Being centralised these factories could be engineered to operate in more controlled environment, e.g. low temperature, vacuum, simplifying the design.
Taking the place of the desktop nanofactory, would be a desktop microfactory, which would do convergent assembly from the microscale up, using only pre-fabbed microblock feedstocks. Microfactories would be freely available to the public, although the available microblock types would be regulated for safety. Microblocks could even support embedded hardware level security, eg. nano-fabbed micromotors could be uniquely tagged with a key to prevent unauthorised use.
Despite these restrictions, microfactories could still support self-replication if provided with a complete set of enabling microblock types, although I would expect that “breeder” factories would be a separately licenced model.
The scheme as described is still atomically perfect, but there is not always a need for such accuracy at all levels of integration, so some microblock types could be bulk manufactured, especially in the early generations. In fact, this tiered concept could evolve from a RepRap like scheme: a micro-fabber subsystem spitting out custom parts for a clunkier macro-fabicator.
As I ponder all this, the future looks rather more evolutionary than I had previously thought…
Marc0, the architecture you outline makes sense and has a number of advantages along the lines you discuss.
The manufacture of microscale blocks would most naturally be performed by modules that are, by ordinary standards, very small. This is one reason why I see no compelling reason to take this model all the way to a strict separation of functions based on the scale and operating conditions of the fabrication processes.
Heat dissipation considerations, however, are a reason to offload almost all of the molecular-scale processing to a facility away from a typical end-use point. Almost all of the waste heat associated with an APM system will be produced as a by-product of molecular-scale manipulation — the result of irreversible, dissipative reactions. This energy is on the same order as a typical heat of combustion: not enormous, but not trivial. Processes that operate on microscale blocks would naturally produce less waste heat enabling higher throughput with less cooling required.
There’s a parallel here to technologies of more immediate concern. A reasonable near-term ambition is to develop building blocks for self assembly that form a modular, combinatoric set. By this I mean a technology platform that decouples the difficult processes of chemical synthesis and product purification from design and assembly processes that use the resulting parts. The aim is to allow diverse structures to be made by relatively fast, simple, reliable processes that combine prefabricated building blocks
Some applications of DNA origami already work this way: The modular parts are different structures that share identical oligomeric “staple” segments, and the scope of the combinatoric design extends to forming patterns by binding a choice of different “pixel” structures at any point in an array of specific sites on a DNA-origami surface. Many different designs can be produced by combining a fixed set of off-the-shelf components.