Intelligent Agent
Reprinted with permission from Foresight Update 14 (foresight@cup.portal.com): 





On June 26, 1992, the U.S. Senate Committee on Commerce, Science,
and Transportation's Subcommittee on Science, Technology, and Space held
a hearing on the topic of "New Technologies for a Sustainable World."
Dr. Eric Drexler, Chairman of the Foresight Institute and Research Fellow
of the Institute for Molecular Manufacturing, was invited to testify on
molecular nanotechnology. The following is the written testimony he submitted;
a later issue will cover the oral portion. 








Summary:

In 1959, the Nobel prizewinning physicist Richard Feynman suggested that
individual atoms and molecules could be positioned and used as building
blocks; experimental results now demonstrate that he was correct. Molecule-by-molecule
control can become the basis of a manufacturing technology cleaner and more
efficient than those known today. This molecular nanotechnology will resemble
processes in farms and forests, in which molecular machines convert common
raw materials - including surplus atmospheric carbon dioxide - into useful
products. It can be a basis for sustainable development, raising the material
standard of living while decreasing resource consumption and environmental
impact. Molecular nanotechnology will have broad applications. It will provide
a general-purpose method for processing materials, molecule by molecule,
much as computers provide a general-purpose method for processing information,
bit by bit. It will by its nature be highly efficient in both materials
and energy use. Its products can include: 








Analysis and simulation based on existing scientific knowledge is enough
to show what molecular nanotechnology can do, but developing it will require
the construction of better molecular tools. The pace of development will
depend not on unpredictable breakthroughs, but on the magnitude and quality
of a focused development effort. The total development time is hard to predict,
but 15 years would not be surprising. Unlike some technology development
projects, in which few payoffs result until the end of the development cycle,
research in molecular nanotechnology will bring major scientific benefits
at an early date. Molecular nanotechnology is worth pursuing both for its
immediate scientific benefits and for its later environmental benefits.
Because there is reason to think that it will become the basic manufacturing
technology of the 21st century - on grounds of cost, quality, efficiency,
and cleanliness - its development also raises issues of economic competitiveness.
Japan's Ministry of International Trade and Industry has recently committed
some US$185 million over ten years to a nanotechnology effort. The U.S.
research community has not yet reached a conclusion regarding the potential
of this field because it has not yet addressed the basic scientific issues.
If we conduct idle debates on molecular nanotechnology while others conduct
active research, they will learn the answers to our questions. It is time
to assess the potential of molecular nanotechnology and to choose a course
of action. If its potential is even half as great as the evidence now indicates,
then medical, economic, and environmental concerns will favor vigorous development.









Introduction

Mr. Chairman, I would like to thank you and the members of this subcommittee
for this opportunity to discuss a topic that I expect will one day become
a leading issue in these halls. The focus of this hearing - new technologies
for a sustainable world - is particularly appropriate for discussion of
this topic, because a concern with the consequences of future technologies
for the environment and for the human condition has for many years guided
my research, and has led to the results described here. 



In the decade since I first described molecular nanotechnology in the Proceedings
of the National Academy of Sciences, this field has progressed from general
theoretical concepts to early laboratory demonstrations and a growing body
of detailed designs. Five years ago, audiences questioned whether individual
atoms could be placed in precise patterns; today, I can answer that question
not just with calculations, but with a slide showing the letters IBM
spelled using 35 xenon atoms. The Foresight Institute, which I serve as
chairman, sponsors a series of scientific conferences on molecular nanotechnology.
The most recent, held last autumn, was cosponsored by the Stanford University
Department of Materials Science and Engineering and the University of Tokyo
Research Center for Advanced Science and Technology; this meeting has stimulated
at least three laboratory research efforts directed toward a key milestone
on the path to molecular nanotechnology. Japan's Ministry of International
Trade and Industry recently committed some US$185 million over the next
10 years to a nanotechnology research effort; development of molecular systems
is seen in Japan as fitting with the broad goal of developing environmentally-compatible
technologies. Momentum toward the development of molecular nanotechnology
is building around the world. The consequences for human life and for Earth's
environment will be enormous, and could be enormously positive. The balance
of this testimony begins by describing molecular nanotechnology from a biological
and ecological perspective and sketching some of its wide range of applications.
It then describes the relevant areas of research; the level of activity
in the Unites States, Japan, and Europe; and some of the policy issues that
its development can be expected to raise. The closing section discusses
how these concepts can be evaluated before committing to any substantial
effort that presumes their validity. 




A biological and ecological perpsective

Industry today consumes fossil fuel and discharges carbon dioxide into the
atmosphere. Forests and farms, in contrast, produce useful products (including
fuels) while removing carbon dioxide from the atmosphere. Proposals for
reducing the concentration of greenhouse gases typically focus on modifying
existing industrial technologies to reduce emissions, and this is a sound
strategy. Yet it may be better to develop industrial technologies that,
like forests and farms, are carbon dioxide consumers. Leaves are solar energy
collectors employing molecular electronic devices: chlorophyll molecules
and photosynthetic reaction centers. These solar energy collectors, like
the other useful products of forests and farms, are built by systems of
molecular machinery such as ribosomes and metabolic enzymes. A natural direction
for technology, then, is to learn to apply systems of molecular machinery
to build useful products in industry. 



The example of green plants indicates some of the results that can be expected
from molecular nanotechnology: 










    
  
  • Low-cost production of solar collectors 
  
  • Low-cost production of large structures (though stronger than wood)
  
  • No production or disposal of toxic chemicals 
  
  • Absorption of atmospheric carbon dioxide 
  
  • Compatibility with the natural world 




Although no technology can, by itself, solve environmental problems, a technology
with these characteristics can be a great help. If a high standard of living
and reduced environmental impact can be achieved with relatively little
sacrifice, then any given amount of political and regulatory pressure should
yield greater results in reducing the impact of human activities on the
natural world. 



Taking the biological analogy as far as the preceding paragraphs have done
risks the misunderstanding that molecular nanotechnology will be a form
of biotechnology. The differences are large: Molecular nanotechnology will
use not ribosomes, but robotic assembly; not veins, but conveyor belts;
not muscles, but motors; not genes, but computers; not cells dividing, but
small factories making products - including additional factories. What molecular
nanotechnology shares with biology is the use of systems of molecular machinery
to guide molecular assembly with clean, rapid precision. Another biological
analogy seems appropriate: Aircraft and birds share some basic principles
of flight, and birds inspired the development of mechanical flight. It would
have been futile, however, to attempt to develop aircraft by applying genetic
engineering to birds, or by concentrating exclusively on ornithological
research. The Wright brothers studied birds, but they then set off in a
fresh direction. Molecular nanotechnology cannot be achieved by tinkering
with life, and its products will differ from biological organisms as greatly
as a jet aircraft differs from an eagle. 








Range of applications

Molecular nanotechnologies will be based on molecular manufacturing, a fundamentally
new way to produce materials and devices from simple raw materials. By guiding
the assembly of molecules with precision, it will enable the construction
of products of unprecedented quality and performance. Because it will work
with the fundamental molecular building blocks of matter, it will be able
to make an extraordinarily wide range of products. Computers provide an
analogy. In the early decades of this century, many specialized data-processing
machines were in use: these included the Hollerith punched-card tabulators
used in the census, Vannevar Bush's analogue machine that solved differential
equations for scientists, and adding machines used in offices to speed accounting
chores. Each of these slow, inefficient, specialized machines has now been
superseded by fast, efficient, general-purpose computers; even pocket calculators
contain computers. By treating data in terms of fundamental building blocks
- bits - general-purpose computers can perform essentially any desired operation
on that data. 



Today, manufacturing relies on many specialized machines for processing
materials: blast furnaces, lathes, and so forth. Molecular nanotechnology
will replace these slow, inefficient, specialized (and dirty) machines with
systems that are faster, more efficient, more flexible, and less polluting.
As with computers and bits, these systems will gain their flexibility by
working with fundamental building blocks. When desktop computers replaced
adding machines, they did more than speed addition. Molecular manufacturing
will likewise open new possibilities. The applications of precise fabrication
at the molecular level (mechanosynthesis) are as broad as technology itself,
because all of technology relies on manufacturing. Molecular-scale components
can be used to place the equivalent of a billion modern computers in a desktop
machine. Molecular-scale components will make possible new medical and scientific
instruments, including DNA readers able to sequence genomes routinely. On
a larger scale, production of better materials will make possible lighter,
more efficient vehicles, without sacrificing structural strength: this will
aid transportation technologies ranging from spacecraft to automobiles.
Lighter structures will consume less material and energy. Because the lightest
and strongest materials will be made from carbon (in the form of graphite
and diamond fibers), carbon dioxide can become a raw material rather than
a waste product. Molecular manufacturing systems can be used to make more
molecular manufacturing systems, hence the capital cost of production can
be low. An analysis of inputs, outputs, and productivity suggests that the
total cost of production can be in the range familiar in agriculture and
in the production of industrial chemicals - tens of cents per pound. At
this cost, many applications become practical. For example, solar photovoltaic
cells fabricated in the form of tough sheets for roofing and paving could
provide solar electric power without consuming additional land. With clean
solar power, clean manufacturing processes, and light, efficient products,
it will be possible to provide a high-material standard of living with decreased
impact on the natural world. This can contribute to the goal of sustainable
development. 








Research directions and funding

These developments are not around the corner, but their feasibility can
be clearly foreseen, as can the nature of research programs able to implement
them. The essential goal is to construct molecular structures with the precision
already familiar in chemical synthesis and protein engineering, but on a
larger scale. Accordingly, properly focused research in chemical synthesis
and protein engineering (within the fields of molecular biology and biochemistry)
is important to the implementation of molecular nanotechnology, as is the
emerging field of molecular manipulation using proximal probe microscopes
such as the scanning tunneling and atomic force microscope. Each of these
areas is a classic small-science field, in which small teams use inexpensive
materials and equipment. The prospect of molecular nanotechnology shows
that small science can have big rewards. 



. . . 



I have not requested and do not anticipate a need for Federal funds to support
my own studies in this area, but the field as a whole could benefit from
vigorous support of appropriate computational simulation and laboratory
research. Since this work would be performed chiefly by existing researchers
with existing equipment, the need is more for a shift in direction than
for a growth in spending. Developments along the path to molecular nanotechnology
promise to yield early results in scientific instrumentation, making it
justifiable as a means of pursuing existing goals in chemistry and in biomedical
research. 



Progress toward molecular nanotechnology in the United States has been retarded
chiefly by cultural obstacles. Molecular nanotechnology will require the
construction of complex molecular machines, but chemistry and biochemistry
are sciences, and focus on the study of nature. To return to the example
of aerospace engineering, expecting molecular scientists to build molecular
manufacturing systems is somewhat like expecting ornithologists to build
aircraft. Building complex systems demands research that first defines goals
and then works backward to identify and implement the means, usually dividing
the work among many teams. Studying nature, in contrast, can be performed
by small research groups, each jealously guarding the independence and purity
of its research. The development of molecular nanotechnology can keep much
of the character of small science, but it will require the addition of a
systems-engineering perspective and a willingness on the part of researchers
to choose objectives that contribute to known technological goals. Progress
will require that researchers build molecular parts that fit together to
build systems, but the necessary tradition of design and collaboration--fundamental
to engineering progress--is essentially absent in the molecular sciences
today. Furthering molecular nanotechnology might best be achieved by directing
federal agencies that perform or fund research in the molecular sciences
to support efforts aimed at the construction of molecular machine systems
and instruments that can precisely position molecules. The results of this
initiative could lead to cost savings in other programs. It has been proposed,
for example, that thousands of researchers be employed over many years at
great expense in order to read the human genome, yet the molecular machinery
found within a dividing cell reads (and copies) the entire genome in a matter
of hours. Scientific instruments based on relatively simple molecular machines
could read DNA with comparable speed and store the results in a computer
memory. The development of such instruments, once the necessary technology
base is in place, could hardly consume the efforts of thousands or researchers;
it would more likely require only a few cooperating laboratories. The result
would enable scientists to read and study many genomes. 



Molecular machinery is a technology of basic importance and deserves to
be treated accordingly. This would be true even without the longer term
goal of molecular manufacturing. 


Research in the Unites States, Japan, and Europe

The Unites States has impressive strengths in areas of science and technology
relevant to molecular nanotechnology. It was at IBM's Almaden laboratory
that Donald Eigler's group spelled IBM using 35 xenon atoms. It was
at William DeGrado's laboratory at DuPont that scientists first designed
and built a new protein molecule, containing hundreds of precisely joined
atoms. Nanotechnology has become a buzzword, but is often used to
describe incremental improvements in existing semiconductor technologies;
although of great value in their own right, these are of surprisingly little
relevance to molecular nanotechnology. (Micromachine research, often confused
with nanotechnology in the popular press, is even less relevant.) Progress
toward molecular nanotechnology in Japan is harder to judge, owing to distance
and language barriers, but the Japanese commitment appears impressive. In
my visits to Japan, I have received a strikingly warm welcome. MITI organized
a symposium around my first visit, at which - despite my many talks in the
U.S - I for the first time met other researchers who were studying molecular
machines not only to understand nature, but to build molecular machine systems.
On another visit, I spoke at the only scientific meeting on the construction
of molecular machine systems that I have attended but did not myself organize.
Japan's NHK television network aired a three-hour series this spring, titled
"Nanospace," that included interviews with me and material from
my work; nothing comparable has appeared on U.S. television. 



While exploring a Japanese-language bookstore that I happened across in
Tokyo last spring, I found a table with eight books on micromachines and
molecular machines, all displayed face on. Half were paperbacks (including
conference proceedings containing a summary of a talk I had given in Tokyo
two years before), and half contained one or more graphics illustrating
molecular machine designs drawn from my work. One of these was a translation
of my first book on molecular nanotechnology, Engines of Creation.
I can with confidence state that no bookstore in the Unites States contains
a similar display, because no such set of books exists in the English language.




MITI's commitment of US$185 million is a sign of strong interest. In addition,
Japan's Science and Technology Agency, through the Exploratory Research
for Advanced Technology program, has sponsored a series of efforts in molecular
engineering, including the Aono Atomcraft Project, which aims to build semiconductor
devices with atom-by-atom control. I recently read that Texas Instruments
has established a laboratory with similar goals; the location they chose
is Tsukuba, north of Tokyo. Researchers at Hitachi's Central Research Laboratory
last year spelled Peace 91 HCRL by removing individual atoms from
a surface. Researchers at the Protein Engineering Research Institute in
Osaka (no comparable institute exists in the Unites States) have designed
and built the largest protein molecules of which I am aware. Nanotechnology
has been a serious goal in Japan for longer than it has in the Unites States,
and is seen as a contributing to technologies in greater harmony with the
natural world. I am less familiar with research in Europe, but key technologies
(such as the scanning-tunneling microscope) have been developed there. Dr.
Hiroyuki Sasabe of the RIKEN Institute in Japan tells me that there are
several research consortia in Europe doing work on molecular systems, and
that he knows of no similar consortia in the United States. 



Policy issues 



Molecular nanotechnology will raise numerous policy issues. In many areas,
years of consideration will be necessary before wise policies can be formulated.
This section provides only a brief, preliminary survey of a few issues of
particular prominence. Research in molecular nanotechnology will by its
nature pose no special risks so long as it remains unable to make large
quantities of product. In its early phases, it will most closely resemble
a branch of laboratory chemistry, and its chief product will be information.
Later, when large-scale applications become possible, major regulatory issues
will arise. Further work will be necessary to identify these issues, but
because molecular manufacturing can be used to produce high-performance
systems of many kinds, these issues will surely include arms control. Because
the Unites States has no clear lead in this technology and because large-scale
commercial applications are still distant, international cooperation in
research may be desirable. Further, because potential long-term applications
include weapon systems, a failure to establish cooperative international
efforts could lead to dangerous outcomes. These considerations suggest the
desirability of a development program involving international cooperation
centering on shared global concerns with health and the environment. One
possible vehicle for this might be an expanded version of the existing Human
Frontier Science Program. It seems that no special regulatory issues will
arise for some time, but this time should be used to gain an understanding
of the issues that will emerge as the technology matures. Cooperative development
can provide a basis for eventual international controls, for example, of
the use of molecular manufacturing in arms production. 








Evaluating molecular nanotechnology

The U.S. scientific community has reached no consensus regarding the prospects
for molecular nanotechnology; indeed, these ideas have stirred heated controversy.
A recent OTA study could identify no published scientific arguments on the
other side (vague and unscientific objections have been common), but it
would be unwise for a decision maker to advocate a major commitment of resources
to molecular nanotechnology without further study and evaluation. 



This autumn, the first quantitative, detailed, book-length analysis of molecular
manufacturing will be published (Nanosystems: Molecular Machinery, Manufacturing,
and Computation, Wiley/Interscience). This work lays out the fundamental
principles of molecular machinery and describes how molecular machines can
collect, orient, process, and assemble molecules with high efficiency and
reliability. If there is a major error or omission in this analysis of molecular
manufacturing, it should be possible for a critic to describe the difficulty
in quantitative, scientific terms. Experience shows, however, that the scientific
community does not move swiftly to evaluate interdisciplinary engineering
proposals. No single discipline sees it as a responsibility, and most scientists
see the work as a distraction from winning their next grant. If these concepts
are to be evaluated soon, and well enough to enable decision makers can
choose with confidence, deliberate action seems necessary. A natural choice
would be to commission a study of molecular manufacturing, setting the objective
of evaluating its scientific and technological feasibility by seeking specific,
scientific criticisms and responses from appropriate researchers. A study
of this sort could provide a basis for decisions and could stimulate further
debate and analysis that would provide a still better basis for decisions.
The Office of Technology Assessment may be an appropriate agency to conduct
this initial study. 








Conclusion

Molecular nanotechnology promises a fundamental revolution in the way we
make things, and in what we can make. By bringing precise control to the
molecular level - resembling the control found in living organisms--it can
serve as a basis for manufacturing processes cleaner, more productive, and
more efficient than those known today. Like green plants, it can produce
inexpensive solar collectors and other useful products while removing carbon
dioxide from the atmosphere. Because it will work with the basic building
blocks of matter, its applications are extraordinarily broad: they include
improved materials and computers. Early applications will include scientific
and medical instruments. Pure science has prepared the ground for molecular
nanotechnology: it is now time to build. Initial goals include the development
of better techniques for positioning molecules and for building molecular
machines. Research in chemistry, biochemistry, and proximal probe microscopy
can all make substantial contributions. Computational simulation has begun
to show in detail what can be built and how it will work. Design, simulation,
and laboratory research can all benefit from support targeted on genuinely
relevant research. Progress will depend largely on the willingness of molecular
scientists to solve problems that contribute to engineering objectives.
Research leading toward molecular nanotechnology is accelerating worldwide.
Focused research is perhaps strongest in Japan. Although large-scale capabilities
(and the need for regulation) are still years away, it is not too early
to consider the consequences of success and to build the framework of international
cooperation that will be necessary in order to manage those consequences.




The preceding paragraphs assume that the analysis supporting the case for
molecular manufacturing is essentially correct, but there is as yet no consensus
on this. The evaluation of interdisciplinary proposals is slow in the absence
of a deliberate effort. It is time to make that deliberate effort, to evaluate
the evidence and set research priorities accordingly. If we merely wait
and see, we will accomplish more waiting than seeing. Economic competitiveness
and the health of the global environment may depend on timely action. 



Assembled (a), cross sectional (b), and exploded (c) views of a design for
a planetary gear system containing 11 moving parts and 3,557 thousand atoms.
Rotation of the inner shaft forces a rolling motion of the nine surrounding
gears, driving rotation of the larger shaft (to the right) at a lower speed.
A molecular machine component of this sort could not be made with existing
chemical techniques, but could be part of a mechanical system made using
molecular manufacturing. This design is the result of a collaboration between
Dr. K. Eric Drexler of the Institute for Molecular Manufacturing and Dr.
Ralph Merkle of the Xerox Palo Alto Research Center, using molecular simulation
software developed by Molecular Simulations Inc. 



From Foresight Update 14, a newsletter on nanotechnology published
by the Foresight Institute, PO Box 61058, Palo Alto, CA 94306, USA; inform@foresight.org.