How long will it take to develop nanotechnology?

Perhaps the most commonly asked question about nanotechnology is: how long will it take to develop?

Addendum added in 2011: While retaining the more cautious answer given below for historical perspective, continuing research in the field has gradually provided a clearer answer: with appropriate focused funding a nanofactory could be developed in 20 years. With abundant funding it might be developed sooner. Unfortunately, in the absence of focused funding it could well take decades longer. If you can assist in speeding the development of molecular nanotechnology please contact Ralph C. Merkle or Robert A. Freitas Jr.

(This question is also addressed by Drexler, Peterson and Pergamit in Chapter 4 of Unbounding the Future).

The correct scientific answer is I don't know.

Having said that, it is worth pointing out that the trends in the development of computer hardware have been remarkably steady for the last 50 years. Plotted on semilog paper as a function of year, such parameters as

• the number of atoms required to store one bit
• the number of dopant atoms in a transistor
• the energy dissipated by a single logic operation
• the resolution of the finest machining technology
• many others

If we extrapolate these trends we find they reach interesting values in the 2010 to 2020 time frame. The number of atoms required to store one bit in a mass memory device reaches 1. The number of dopant atoms in a transistor reaches 1, (while fundamental device physics might force us to use more than one dopant atom, it's clear that some not-too-large integer number should suffice). The energy dissipated by a single logic operation reaches kT for T=300 kelvins; this is roughly the energy of a single air molecule bouncing around at room temperature. The finest machining technologies reach a resolution of roughly an atomic diameter.

Such performance seems to require a manufacturing technology that can arrange individual atoms in the precise structures required for molecular logic elements, connect those logic elements in the complex patterns required for a computer, and do so inexpensively for billions of billions of gates. In short, if we're to keep the computer hardware revolution on schedule then it seems we'll have to develop nanotechnology in the 2010 to 2020 time frame.

Of course, extrapolating straight lines on semilog paper is a philosophically debatable method of technology forecasting. While we can confidently state that no fundamental law of nature prevents us from developing nanotechnology on this schedule (or even faster), there is equally no law of nature that says the trends of the past must continue unchanged into the future, or that this schedule will not slip. For example, while Babbage proposed the stored program computer in the 1830's, it was about a century before anyone actually built one.

In 1993 the author, as co-chair of the Third Foresight Conference on Molecular Nanotechnology, asked the assembled attendees how long they thought it would take to develop nanotechnology, as defined here. By show of hands, answers in the range from 2010 to 2040 predominated (about two thirds of the audience).

Regardless of what extrapolation of trends or polls might suggest, we should keep firmly in mind that how long it takes depends on what we do (or don't do). A focused effort with resources appropriate to the magnitude of the task would speed development. If we do little, or focus resources on short term goals, fundamental developments might be much delayed (just as Babbage's computer was delayed by a century). To quote Alan Kay: "The best way to predict the future is to invent it."

References

• Wikipedia has an excellent article on Moore's Law.
• Approximate component count for complex integrated circuits vs. year of introduction, from Progress in digital integrated circuits, by Gordon Moore, 1975 International Electron Devices Meeting, page 11.
• The following figures are from Miniaturization of electronics and its limits, by R. W. Keyes, IBM Journal of Research and Development, Volume 32, Number 1, January 1988.
  • Figure 4: Number of atoms used to store a bit in discrete magnetic entities and in file technologies.
  • Figure 5: The decreasing number of dopant impurities in the base of bipolar transistors for logic.
  • Figure 6: The decreasing energy dissipated per logic operation.
• A century of computing, from Mind Children, by Hans Moravec, 1988, Harvard University Press.
• The development of achievable machining accuracy, from Current status in, and future trends of, ultraprecision machining and ultrafine materials processing, by Norio Taniguchi, Tokyo Science University, Annals of the CIRP Vol. 32/2/1983 page 573. This shows the trend in ultraprecision machining from 1900 and projects that it will reach "Atomic lattice distance" (0.3 nanometers) around 2017.
• Taking the plunge, from The Economist, September 28th 1996, page 8 of "The World Economy".
• Escalating cost of semiconductor manufacturing facilities, from Electronic News, September 27 1993, page 28. (This and the following item suggest that incremental improvements in our manufacturing technology will be limited by economics: we'll be forced to consider new approaches (such as self replication) that address the economics of manufacturing at a very fundamental level).
• Lithography and the future of Moore's law, by Gordon E. Moore, in Proceedings SPIE Vol 2440, Optical/Laser Microlithography VIII, Feb 1995, Timothy A. Brunner ed.

  I am increasingly of the opinion that the rate of technological progress is going to be controlled from financial realities. We just will not be able to go as fast as we would like because we cannot afford it, in spite of your best technical contributions. When you are looking at new technology, please look at how to make that technology affordable as well as functional.

• The Revolutionaries, from the San Jose Mercury News, Sunday January 26, 1997, page 14A, quoting Gordon Moore:

  I can see another couple of generations of the processing technology that carries us another six to 10 years at about the same slope.

  On the other hand, we are beginning to approach some real limits. It gets difficult to make smaller and smaller structures. We just exhaust the capabilities of optical systems; we need something different to make them smaller yet.

• Cramming more components onto integrated circuits by Gordon E. Moore, Electronics, April 19, 1965, pages 114-117. The original article describing Moore's law.

This page is part of the nanotechnology web site.