Nanotech: Engines of Hyperbole?
*******************************

By Charles Platt


Ralph Merkle and Eric Drexler are designing a robot arm. The arm contains
about four million atoms, which sounds like a lot until you realize it
takes up about as much room as a dust mote. In fact, if you laid out these
robot arms like cookies on a baking tray, you could fit more than 62
billion per square inch.

The concept driving this research is that instead of smelting steel,
making molds, stamping out components, and fitting everything together on
an assembly line, we can have robots tiny enough to assemble molecules one
by one. This way, a gadget should be able to grow itself from a chunk of
raw material, much as human beings grow by absorbing nutrients, turning
them into useful compounds, and fitting them together according to
instructions stored in our DNA.

Stacking atoms individually is obviously going to be a time-consuming
business. But suppose you tell your microminiaturized robot arm to build a
copy of itself. Then you get the two arms to make two more, and you get
those four to build another four. . .until, in a short space of time, this
exponential growth gives you billions upon billions of arms ready to do
whatever you want.

Now suppose each arm has an equally tiny onboard computer controlling it,
so that it can be quickly reprogrammed in a way that DNA never could be.
Result: You have an entire factory of microbe-sized machines capable of
building just about anything you want, including drugs, televisions,
homes, highways, or microscopic repair robots that could run around in the
human blood stream, rejuvenating tissues and zapping cancer cells.

Our tiny helpers can be told to create whole new cities (and disassemble
the old ones), eat up pollution, and remake the world as a neat, clean,
postindustrial paradise. Along the way, they can bring us immortality by
arresting the aging process.

This is the promise of nanotechnology, the prefix "nano" meaning on the
scale of a nanometer, which is a billionth of a meter (a typical virus
might be about 100 nanometers long). It's a vision that some people have
trouble taking seriously because it sounds so wildly transcendent. Can it
really happen?

Right now, the smallest control devices in the real world are still 1,000
times too big to manipulate single atoms in three dimensions. Ralph
Merkle, however, is not discouraged. He's a calm, dignified, affable man
with a friendly smile, the type who likes to resolve conflicts by asking
everyone to sit down and be reasonable. But behind his easygoing manner
there's no hint of compromise, and his impatience is obvious. If we make
the right decisions, according to Merkle, we could see real results in
maybe twenty years' time. Nanotechnology, he says, can be "the cornerstone
of future technology, a fundamental factor in the future development of
civilization." This sounds like empire-building on a grand scale; but like
most nanotech advocates, Merkle seems at least as interested in living in
it as in building it. (That bit about arresting the aging process does
have its attractions.)

Merkle heads the Computational Nanotechnology Project at Xerox PARC in
Palo Alto, California, where he runs Polygraf chemistry-simulation
software on a Silicon Graphics workstation. That's how he and Drexler are
designing the robot arm - on the screen, where molecules can be edited
with the same fluency with which most people edit text. The simulations
verify that the atomic configurations are stable, and skeptical chemists
are forced to admit that if the particles could just be pushed into place,
the molecular bearings and reduction gears would work as advertised.

But how long will it really be until this is possible? Seven years have
passed since Eric Drexler developed the concept of nanotech in his
groundbreaking book, The Engines of Creation. Is anything actually being
done on benches in real-life laboratories to turn his utopian vision into
reality?

To answer this question, I did a quick survey of recently published
literature and labs across the country.

The National Nanofabrication Facility (NNF) at Cornell University in
Ithaca, New York, seemed the most logical place to start. It's the largest
facility of its kind, with a US$4 million annual budget, 250 graduate
students, and a huge array of equipment that outsiders can use on a guest
basis. Edward Wolf was president of the facility for its first ten years
and now serves as its spokesman. He's an electrical engineer who used to
work at Hughes Aircraft, and he has a can-do, no-nonsense, problem-solving
style - the sort of man you can imagine saying, "If it ain't broke, son,
don't fix it."

Wolf is a believer in the top-down approach. In other words, we start by
building small things, then we build very small things, and we keep on
working downward in size, until we eventually end up at the atomic level.
The NNF is currently exploring projects ranging from ultra-fast
gallium-arsenide logic gates to chips that incorporate miniaturized
sensors carved out of silicon. There are immediate applications for this
kind of device. The mechanism for locking seat belts in a car, for
instance, will soon be based around a chip that contains its own
ultra-tiny motion sensor, like a miniaturized pea rattling in a pod.

This is real, and it's practical, but it doesn't have the power and
promise of nanotechnology, a word that now has at least two different
meanings. To clarify matters, Drexler now talks about "molecular
nanotechnology" to describe the process of moving molecules individually
to create more complex structures (the bottom up-approach).

From listening to Wolf, it quickly becomes clear that top-down advocates
such as he have formed a faction that's violently opposed to the bottom-up
approach. Wolf tries to be nice about it, but his impatience is evident.
"The fact is," he says, "we have no idea how to do 3-D assembly. I don't
want to demean anybody, but nanotechnology has been oversold, in the sense
that it's been popularized."

What's so bad about popularization?

"Let me put it this way. As scientists, we all have great imaginations.
But most of us are a little more cautious about speaking out."

Meanwhile, on the other side are scientists like David Blair, a biologist
at the University of Utah. "I have a lot of respect for Eric Drexler," he
says. "He's a visionary."

Blair is studying one of the smallest, most complicated, most fascinating
objects in the natural world. It's an incredibly tiny motor, one-fourth
the size of the robot arm that Merkle and Drexler would like to build. It
also happens to be up and running. It's made of protein, and the amazing
thing is, it grows naturally as a component of common, ordinary bacteria.

Like any cells, bacteria must eat to survive. Somehow (no one knows how)
they detect nearby nutrients and swim toward them. Their means of
propulsion is the flagellum, which consists of a corkscrew-shaped filament
attached to the nanomotor. It spins the filament at about 15,000 rpm,
propelling bacteria in the same way that outboard motors propel boats.
Make no mistake, this genuinely is a motor. There's a stator and a rotor,
just as in a motor that turns an electric fan.

Blair would like to take the motor to pieces, but he has no tools tiny
enough for the job. So, he uses an indirect method, selectively damaging
the proteins that the motor grows from and watching how it malfunctions.
"We're like auto mechanics who only have hammers to do our work," he says.
"But I think we may learn interesting lessons for the design of man-made
motors some time in the future."

A more precise tool for investigating the nanorealm is the
scanning-tunneling electron microscope (STM), which uses a super-sharp,
servo-controlled tungsten tip to measure the size of individual atoms on a
hard, flat surface. The STM is unsuitable for tracing the contours of soft
bioforms, but its powers have been amply demonstrated in other areas. In
1990, Don Eigler attracted global publicity when he used his STM to push
individual xenon atoms around so that they spelled out the initials of his
employer, IBM. Eigler is now reportedly pushing carbon and oxygenatoms
together to create carbon monoxide, but to advocates like Ralph Merkle,
this kind of work is just the first step down a long road. Using today's
STMs to manipulate molecules, says Merkle, "is like trying to build a
wristwatch with a sharpened stick."

However, an exciting new development in the STM world has come from an
unlikely source. A graduate student named Mark Voelker, at the Optical
Sciences Center of the University of Arizona at Tucson, has built the
nation's first dual-tip scanning-tunneling electron microscope.

Voelker is shy and soft-spoken, and he describes himself as "just another
science nerd." But he seems to have a practical, do-it-yourself side to
his personality. He enjoys hanging out in pool rooms, and he drives a
rebuilt 1970 Corvette. "I blueprinted the car," he says, in his placid,
low-key style. "That means I completely disassembled the engine, measured
every part, and machined them to adjust the tolerances. Then I put
everything back together again."

He used the same kind of patient skill in building his microscope. Coming
from a background in infrared astronomy, he had no special qualifications
for the job; he just figured he ought to be able to do it. "I got funding
to build a single-tip machine, originally," he says, "but part way
through, I realized that with a little more effort and a little more risk,
we could have two tips. So, while my advisor was away for the summer, I
decided just to go ahead and do it."

The advantage of using two probes is that one can apply an electrical
signal while the other listens for it. This way, the microscope can not
only sense its target, but interact with it.

"People have a lot of designs for molecular electronic devices," Voelker
says. "They've made wires consisting of conductive polymers just one
molecule thick. But these things haven't been properly tested, because
there's no easy way to make contact with devices that are 1,000 times
smaller than current chips."

Why is molecular electronics important? "Because it should enable us to
put a million transistors in the same area occupied by one transistor on a
chip today, and the smaller transistors should run a thousand times
faster. That gives you a computing technology a billion times more
powerful, in terms of computing power per square centimeter."

Voelker sees other possible uses for his invention, and figures that for
around US$100,000, he could put together an entire nanotechnology
workbench. To build 3-D structures, he thinks it might work to create
layers one molecule thick, then stack them. And he believes something of
this sort could be achieved before the year 2000.

But even molecular electronics is just a way-station along the path to
true molecular nanotechnology. Drexler's ultimate model for a nanocomputer
wouldn't use electricity at all. Instead, it would contain "rod logic"- a
crisscross pattern of sliding rods, one molecule thick, each fitted with
stops that interact with the others. If this sounds clunky, it's because
we're used to thinking in macro terms. On the molecular scale, everything
is so small and light, there's virtually no inertia, and nanomechanical
systems can vibrate millions of times per second. Friction as we know it
does not exist, and so far as anyone can tell, a rod-logic computer should
never wear out.

According to Drexler, a computer slightly larger than a virus, using this
system, should possess the computing power of a present-day personal
computer. This opens up some truly incredible possibilities. Imagine a
"factory" the size of a stapler. It sits on your desk and "builds" a sheet
of paper whenever you need one. But the paper isn't really paper; it's a
matrix of tiny display and arithmetic modules, so it functions like a
video screen and a computer combined, with voice recognition built in.

Imagine every object in the everyday world imbued with this kind of
intelligence. Clothes restyle themselves, houses remodel themselves, food
grows itself and cooks itself, and surgical implants could adapt
themselves to the human body. Nanotech advocates find this vision
compelling, to put it mildly. They want to develop it, but more than that,
they want to live in it - which is why they're so impatient to get funds
for research that will create the tiny building blocks to make it all
happen.

It's hard to see how an STM could tackle this job. For the time being, at
least, it may make more sense to fabricate nanodevices in the same way
that chemists fabricate exotic compounds: by using the capability of
molecules to self-assemble. In other words, on a molecular scale, we can
throw pieces of a jigsaw puzzle into a bag, shake them around, and they'll
put themselves together - so long as the shapes are chosen cunningly
enough.

There's already been some progress in this direction. Nadrian Seeman, at
New York University, has "trained" DNA to form itself into the framework
of a cube. Galen Stuckey, at the University of California in Santa
Barbara, has created caged structures where one molecule literally wraps
itself around another. This is getting back to the top-down approach to
nanotechnology. Does it make sense to put resources into this, or should
we find better ways of repositioning individual atoms?

Eric Drexler claims to be agnostic on this subject. "I'd like to see
vigorous research in both approaches," he says, "adapting atomic-force
microscopes to position molecules and do mechanosynthesis, and applying
existing chemistry to make polymers that will fold up like proteins, to
form molecular machines."

He's more definite about the areas in which he expects to see
nanotechnology make its first impact.

"The first high-payoff applications are likely to be in information.
First, molecular sensors that provide special information about compounds,
perhaps in home medical-test kits. Second, molecular devices that do what
silicon does now, but better, perhaps providing terabyte RAM chips." (A
terabyte is a million megabytes.)

Meanwhile, Drexler is philosophical about the prejudice that many
scientists still harbor. "The tide seems to be turning," he says, although
he admits that many researchers take a short-term view. "They are
interested in what they and their team can test in the labs this year,
with available grant money, and they tend to be less interested in what
could be developed by many teams working together over the next decade or
two. In addition, the consequences of molecular nanotechnology simply
strike many people as being too big to be true. I find this hard to
understand. It's like saying that the ability to process the ones and
zeros of information can't have much importance, because bits are so
small."

As always, Drexler feels he is able to refute his critics. His recent book
Nanosystems is heavy with formulas, tables, and graphs, relentlessly
proving the feasibility of molecular nanotech, and addressing every
conceivable problem. Will thermal noise (vibration on the atomic level)
impair the accuracy of nanomachines? How can reactive atoms be manipulated
safely? After we build nanoscale robots, how will we communicate with
them, to tell them what to do? He tackles all these questions using the
tools of orthodox chemistry and biology.

Of course, this is pointless if some scientists are so prejudiced against
him that they won't take the time to listen. Edward Wolf, at the National
Nanofabrication Facility, says that he can't comment on Drexler's
arguments because "I haven't considered his work important enough to read."

Biologist Thomas Donaldson has pointed out: "The future isn't a place that
we travel to. The future is something that we have to build." As long as
large laboratories focus on short-term solutions, the future of
nanotechnology rests on the active work of scattered researchers like
Blair and Voelker, and it will take a while before the simulations move
off the screen, into the home.

                                   * * *

Charles Platt (charles@phantom.mindvox.com) is a science writer and author
of The Silicon Man.


=-=-=-=-=-=-=-=-=-=-=-=WIRED Online Copyright Notice=-=-=-=-=-=-=-=-=-=-=-=

        Copyright 1993,4 Wired Ventures, Ltd.  All rights reserved.

  This article may be redistributed provided that the article and this
  notice remain intact. This article may not under any circumstances
  be resold or redistributed for compensation of any kind without prior
  written permission from Wired Ventures, Ltd.

  If you have any questions about these terms, or would like information
  about licensing materials from WIRED Online, please contact us via
  telephone (+1 (415) 904 0660) or email (info@wired.com).

       WIRED and WIRED Online are trademarks of Wired Ventures, Ltd.

=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=