At the 6th Foresight Conference on Molecular Nanotechnology in November of 1998 I posed the question whether the Department of Defense really understood the consequences of nanoassembly and were preparing adequately for their potential offered by nanoassembly. The response I received seemed to indicate that while they conceived of nanotechnology as being very useful on the small scale, e.g. for microelectronics, there were still some misconceptions with regard to the use of nanotechnology on the largest scales.
The motivation for the question came out of my own investigation and thinking about the true limits to nanotechnology. As has been documented in Engines of Creation [1] and Nanosystems [2], nanoassembly provides a mass doubling time of approximately 1 hour [2, pgs 1, 425, 430]. Drexler's designs are fairly conservative. Hall has proposed "assembly lines" that have mass doubling times that are much faster [3].
These numbers are not extreme when you consider the biological realm. Many bacteria may replicate in less than an hour. Bacterial species such as Salmonella typhi and Escherichia coli have generation (doubling) times of ~20 minutes under optimal conditions [4, 5]. As noted by Sherris [5, pg 40], a bacteria with a doubling time of 20 minutes produces a million cells in 7 hours. Brock and Madigan [4, pg 313], point out, "One can calculate that a single bacterium with a generation time of 20 minutes would, if it continued to grow exponentially for 48 hours, produce a population the weighed about 4,000 times the weight of the earth!". Bacteria must use some clever tricks to do this. The replication rate of DNA polymerase requires ~40 minutes to copy the ~4 million bases of DNA in the genetic program of these species. So to divide every 20 minutes, the bacteria must start making a second copy of the DNA before the first copy is finished [5, pg 38].
This contrasts sharply with eukaryotic, especially mammalian replication times which are about 20-24 hours [5, 6]. The difference in replication times is due to several factors. Eukaryotic cells have a much greater amount of DNA which must be replicated (10-1000 times more DNA). Because nature cannot consciously place the origin of replication signals (small regions of DNA where the DNA polymerase commences chromosome copying) at precise intervals, there must be some allowance for longer regions which replicate more slowly than smaller regions. There must be multiple checkpoints where the cell decides that the DNA has been copied and that any necessary repairs have been completed. These must be conservative from a timing perspective since proceeding forward before they are successfully complete will result in mutations that could lead either to cell death (wasting resources) or cancer (potentially killing the host). Eukaryotic cells have a much more complex structure than bacteria. Multiple chromosomes, chromosome packing and a nuclear membrane require that there must be more steps to the division, such as breaking down the nuclear envelope, alignment and separation of the chromosomes and reassembly of the nuclear envelope after the cells have divided. Care must be taken to distribute essential organelles (mitochondria, peroxisomes, etc.) relatively evenly in the daughter cells. The genetic complexity of eukaryotes requires the duplication of many more micromachines (enzymes, replicators, disassemblers, etc.). Bacteria only have a few thousand varieties of these machines while Eukaryotic cells have tens of thousands. Eukaryotic cells are also much larger, with perhaps 1,000-10,000 times the volume of a bacteria, so division processes which may occur quickly via diffusion in bacteria may require active management in a eukaryotic cells. Because these active management processes have been designed in a very ad-hoc fashion by nature, they are likely to be suboptimal. The large size of the cells delays the time it takes for molecules which regulate the steps of the cell division sequence to diffuse to the locations where they perform their functions.
If one desires to compare the energy and material costs of the nanoassembly of inanimate objects (aircraft carriers or buildings) with animate objects (humans or elephants), one must take into account energy requirements from the perspective of the differences in metabolism. Aircraft carriers and buildings do not typically have an active metabolism while humans and elephants do. Inanimate objects do not have to maintain minimal operating temperature. In fact, using nanoassembly, with typical source materials, so much heat is generated, the problem is cooling the structure during the assembly process. Once constructed, the energy costs of inanimate objects should be minimal. So the growth rates for inanimate objects should be faster than those of organisms that must sharply constrain their operating temperatures and dedicate some fraction of their energy to the harvesting of additional food and nutrient resources.
In systems designed by engineers, we can allow for the mass and energy (if any) transport requirements during construction and make provisions for the removal of waste materials and heat. Because we are designing our object with a vision in mind of the final product (rather than the way nature does it), we may optimize processes like mass transport and heat removal to optimize growth rates.
To understand the problem of nanoassembly of large scale objects, we must break the problem down into sub-problems. These are:
Since the largest Nimitz class aircraft carriers weigh ~97,000 tons (0.9×108 kg) [8], the Alaska Pipeline, carrying a million barrels of oil each day, would provide enough raw materials to construct an aircraft carrier every day. Assuming a cost per barrel of ~$15, this would imply that the aircraft carrier would cost ~$15 million. This would be a cost reduction of more than 2 orders of magnitude from current costs (> $4 billion). Since a diamond has approximately 50 times the strength of steel, the aircraft carrier, could weigh only ~2000 tons, meaning the construction time would take ~30 minutes, and cost only $300,000.