Unbounding the Future:
the Nanotechnology Revolution
The tour in the last chapter showed the sizes, forces, and
general nature of objects in the molecular world. Building on
this, we can get a better picture of where developments seem to
be leading, a better picture of molecular
manufacturing itself. To show the sizes, forces, and general
nature of things in molecular manufacturing, we first invite the
reader (and the reader's inquisitive alter ego) to take a second
and final tour before returning to the world of present-day
research. As before, the pre-1990 history is accurate, and the
science isn't fiction.
The Silicon Valley Faire
The tour of the molecular world showed some products of
molecular manufacturing, but didn't show how they were made.
The technologies you remember from the old days have mostly
been replacedbut how did this happen? The Silicon
Valley Faire is advertised as "An authentic theme park
capturing life, work, and play in the early Breakthrough
years." Since "work" must include
manufacturing, it seems worth a visit.
A broad dome caps the park "To fully capture the
authentic sights, sounds, and smells of the era," the
tourguide politely says. Inside, the clothes and hairstyles,
the newspaper headlines, the bumper-to-bumper traffic, all
look much as they did before your long nap. A light haze
obscures the buildings on the far side of the dome, your eyes
burn slightly, and the air smells truly authentic.
The Nanofabricators, Inc., plant offers the main display
of early nanotechnology.
As you near the building, the tourguide mentions that this is
indeed the original manufacturing plant, given landmark
status over twenty years ago, then made the centerpiece of
the Silicon Valley Faire ten years later, when . . . With a
few taps, you reset the pocket tourguide to speak up less
As people file into the Nanofabricator plant, there's a
moment of hushed quiet, a sense of walking into history.
Nanofabricators: home of the SuperChip, the first mass-market
product of nanotechnology. It was the huge memory capacity of
SuperChips that made possible the first Pocket Library.
This section of the plant now houses a series of displays,
including working replicas of early products. Picking up a
Pocket Library, you find that it's not only the size of a
wallet, but about the same weight. Yet it has enough memory
to record every volume in the Library of
Congresssomething like a million times the capacity of
a personal computer from 1990. It opens with a flip, the
two-panel screen lights up, and a world of written knowledge
is at your fingertips. Impressive.
"Wow, can you believe these things?" says another
tourist as he fingers a Pocket Library. "Hardly any
video, no 3-Djust words, sound, and flat pictures. And
the cost! I wouldn't've bought `em for my kids at that
Your tourguide quietly states the price: about what you
remember for a top-of-the-line TV set from 1990. This isn't
the cheap manufacturing promised by mature nanotechnology,
but it seems like a pretty good price for a library. Hmm . .
. how did they work out the copyrights and royalties? There's
a lot more to this product than just the technology . . .
The next room displays more technology. Here in the
workroom where SuperChips were first made, early nanotech
manufacturing is spread out on display. The whole setup is
surprisingly quiet and ordinary. Back in the 1980s and 1990s,
chip plants had carefully controlled clean rooms with gowns
and masks on workers and visitors, special workstations, and
carefully crafted air flows to keep dust away from products.
This room has none of that. It's even a little grubby.
In the middle of a big square table are a half-dozen steel
tanks, about the size and shape of old-fashioned milk cans.
Each can has a different label identifying its contents:
MEMORY BLOCKS, DATA-TRANSMISSION BLOCKS, INTERFACE BLOCKS.
These are the parts needed for building up the chip. Clear
plastic tubes, carrying clear and tea-colored liquids, emerge
from the mouths of the milk cans and drape across the table.
The tubes end in fist-sized boxes mounted above shallow
dishes sitting in a ring around the cans. As the different
liquids drip into each dish, a beater like a kitchen mixer
swirls the liquid. In each dish, nanomachines are
A Nanofab "engineer," dressed in period clothing
complete with name badge, is setting up a dish to begin
building a new chip. "This," he says, holding up a
blank with a pair of tweezers, "is a silicon chip like
the ones made with pre-breakthrough technology. Companies
here in this valley made chips like these by melting silicon,
freezing it into lumps, sawing the lumps into slices,
polishing the slices, and then going through a long series of
chemical and photographic steps. When they were done, they
had a pattern of lines and blobs of different materials on
the surface. Even the smallest of these blobs contained billions
of atoms, and it took
several blobs working together to store a single bit of
information. A chip this size, the size of your fingernail,
could store only a fraction of a billion bits. Here at
Nanofab, we used bare silicon chips as a base for building up
nanomemory. The picture on the wall here shows the surface of
a blank chip: no transistors, no memory circuits, just fine
wires to connect up with the nanomemory we built on top. The
nanomemory, even in the early days, stored thousands
of billions of bits. And we made them like this, but a
thousand at a time" He places the chip in the
dish, presses a button, and the dish begins to fill with
"A few years latter," he adds, "we got rid of
the silicon chips entirely"he props up a sign
saying THIS CHIP BUILD BEGAN AT: 2:15 P.M., ESTIMATED
COMPLETION TIME: 1:00 A.M." and we sped up the
construction process by a factor of a thousand."
The chips in the dishes all look pretty much the same except
for color. The new chip looks like dull metal. The only
difference you can see in the older chips, further along in
the process, is a smooth rectangular patch covered by a film
of darker material. An animated flowchart on the wall shows
how layer upon layer of nanomemory building blocks are
grabbed from solution and laid down on the surface to make
that film. The tourguide explains that the energy for this
process, like the energy for molecular machines
within cells, comes from
dissolved chemicalsfrom oxygen and fuel molecules. The total amount
of energy needed here is trivial, because the amount of
product is trivial: at the end of the process, the total
thickness of nanomemory structurethe memory store for a
Pocket Libraryamounts to one-tenth the thickness of a
sheet of paper, spread over an area smaller than a postage
The animated flowchart showed nanomemory building blocks
as big things containing about a hundred thousand atoms
apiece (it takes a moment to remember that this is still
submicroscopic). The build process in the dishes stacked
these blocks to make the memory film on the SuperChip, but
how were the blocks themselves built? The hard part in this
molecular-manufacturing business has got to be at the bottom
of the whole process, at the stage where molecules are put
together to make large, complex parts.
The Silicon Valley Faire offers simulations of this molecular
assembly process, and at no extra charge. From the tourguide,
you learn that modern assembly processes are complex; that
earlier processeslike those used by Nanofabricators,
Inc. used clever-but-obscure engineering tricks; and
that the simplest, earliest concepts were never built. Why
not begin at the beginning? A short walk takes you to the
Museum of Antique Concepts, the first wing of the Museum of
A peek inside the first hall shows several people strolling
around wearing loosely fitting jumpsuits with attached
goggles and gloves, staring at nothing and playing mime with
invisible objects. Oh well, why not join the fools' parade?
Stepping through the doorway while wearing the suit is
entirely different. The goggles show a normal world outside
the door and a molecular world inside. Now you, too, can see
and feel the exhibit that fills the hall. It's much like the
earlier simulated molecular world: it shares the standard
settings for size, strength, and speed. Again, atoms seem 40
million times larger, about the size of your fingertips. This
simulation is a bit less thorough than the last wasyou
can feel simulated objects, but only with your gloved hands.
Again, everything seems to be made of quivering masses of
fused marbles, each an atom.
"Welcome," says the tourguide, "to a 1990
concept for a molecular-manufacturing plant. These exploratory
engineering designs were never intended for actual use,
yet they demonstrate the basics of molecular manufacturing:
making parts, testing them, and assembling them."
Machinery fills the hall. Overall, the sight is reminiscent
of an automated factory of the 1980s or 1990s. It seems clear
enough what must be going on: Big machines stand beside a
conveyor belt loaded with half-finished-looking blocks of
some material (this setup looks much like Figure 2); the
machines must do some sort of work on the blocks. Judging by
the conveyor belt, the blocks eventually move from one arm to
the next until they turn a corner and enter the next hall.
FIGURE 2: ASSEMBLER WITH FACTORY ON CHIP
A factorylarge enough to make over 10 million nanocomputers per
day would fit on the edge one of today's integrated
circuits. Inset shows an assembler arm together
with workpiece on a conveyor belt.
Since nothing is real, the exhibit can't be damaged, so you
walk up to a machine and give it a poke. It seems as solid as
the wall of the nanocomputer in the previous tour. Suddenly,
you notice something odd: no bombarding air molecules and no
droplets of waterin fact, no loose molecules anywhere.
Every atom seems to be part of a mechanical system, quivering
thermal vibration, but otherwise perfectly controlled.
Everything here is like the nanocomputer or like the tough
little gear; none of it resembles the loosely coiled protein
or the roiling mass of the living cell.
The conveyor belt seems motionless. At regular intervals
along the belt are blocks of material under construction:
workpieces. The nearest block is about a hundred marble-bumps
wide, so it must contain something like 100 x 100 x 100
atoms, a full million. This block looks strangely familiar,
with its rods, crank, and the rest. It's a
nanocomputeror rather, a blocklike part of a
nanocomputer still under construction.
Standing alongside the pieces of nanocomputer on the conveyor
belt, dominating the hall, is a row of huge mechanisms. Their
trunks rise from the floor, as thick as old oaks. Even though
they bend over, they rear overhead. "Each machine,"
your tourguide says, "is the arm of a general-purpose
One assembler arm is bent over with its tip pressed to a
block on the conveyor belt. Walking closer, you see molecular
assembly in action. The arm ends in a fist-sized knob with a
few protruding marbles, like knuckles. Right now, two
quivering marblesatomsare pressed into a small
hollow in the block. As you watch, the two spheres shift,
snapping into place in the block with a quick twitch of
motion: a chemical reaction. The assembler arm just stands
there, nearly motionless. The fist has lost two knuckles, and
the block of nanocomputer is two atoms larger.
The tourguide holds forth: "This general-purpose
assembler concept resembles, in essence, the factory robots
of the 1980s. It is a computer-controlled mechanical arm that
moves molecular tools according to a series of instructions.
Each tool is like a single-shot stapler or rivet gun. It has
a handle for the assembler to grab and comes loaded with a
little bit of mattera few atomswhich it attaches
to the workpiece by a chemical reaction." This is like
the rejoining of the protein chain in the earlier tour.
The atoms seemed to jump into place easily enough; can
they jump out of place just as easily? By now the assembler
arm has crept back from the surface, leaving a small gap, so
you can reach in and poke at the newly added atoms. Poking
and prying do no good: When you push as hard as you can (with
your simulated fingers as strong as steel), the atoms don't
budge by a visible amount. Strong molecular bonds hold them
Your pocket tourguidewhich has been applying the power
of a thousand 1990s supercomputers to the task of deciding
when to speak upremarks, "Molecular bonds hold
things together. In strong, stable materials atoms are either
bonded, or they aren't, with no possibilities in between.
Assemblers work by making and breaking bonds, so each step
either succeeds perfectly or fails completely. In
pre-breakthrough manufacturing, parts were always made and
put together with small inaccuracies. These could add up to
wreck product quality. At the molecular scale, these problems
vanish. Since each step is perfectly precise, little errors
can't add up. The process either works, or it doesn't."
But what about those definite, complete failures? Fired by
scientific curiosity, you walk to the next assembler, grab
the tip, and shake it. Almost nothing happens. When you shove
as hard as you can, the tip moves by about one-tenth of an
atomic diameter, then springs back. "Thermal vibrations
can cause mistakes by causing parts to come together and form
bonds in the wrong place," the tourguide remarks.
"Thermal vibrations make floppy objects bend further
than stiff ones, and so these assembler arms were designed to
be thick and stubby to make them very stiff. Error rates can
be kept to one in a trillion, and so small products can be
perfectly regular and perfectly identical. Large products can
be almost perfect, having just a few atoms out of
place." This should mean high reliability. Oddly, most
of the things you've been seeing outside have looked pretty
ordinarynot slick, shiny, and perfect, but rough and
homey. They must have been manufactured that way, or
made by hand. Slick, shiny things must not impress anyone
By now, the assembler arm has moved by several
atom-widths. Through the translucent sides of the arm you can
see that the arm is full of mechanisms: twirling shafts,
gears, and large, slowly turning rings that drive the
rotation and extension of joints along the trunk. The whole
system is a huge, articulated robot arm. The arm is big
because the smallest parts are the size of marbles, and the
machinery inside that makes it move and bend has many, many
parts. Inside, another mechanism is at work: The arm now ends
in a hole, and you can see the old, spent molecular tool
being retracted through a tube down the middle.
Patience, patience. Within a few minutes, a new tool is on
its way back up the tube. Eventually, it reaches the end.
Shafts twirl, gears turn, and clamps lock the tool in
position. Other shafts twirl, and the arm slowly leans up
against the workpiece again at a new site. Finally, with a
twitch of motion, more atoms jump across, and the block is
again just a little bit bigger. The cycle begins again. This
huge arm seems amazingly slow, but the standard simulation
settings have shifted speeds by a factor of over 400 million.
A few minutes of simulation time correspond to less than a
millionth of a second of real time, so this stiff, sluggish
arm is completing about a million operations per second.
Peering down at the very base of the assembler arm, you can
get a glimpse of yet more assembler-arm machinery underneath
the floor: Electric motors spin, and a nanocomputer chugs
away, rods pumping furiously. All these rods and gears move
quickly, sliding and turning many times for every cycle of
the ponderous arm. This seems inefficient; the mechanical
vibrations must generate a lot of heat, so the electric
motors must draw a lot of power. Having a computer control
each arm is a lot more awkward now than it was in
pre-breakthrough years. Back then, a robot arm was big and
expensive and a computer was a cheap chip; now the computer
is bigger than the arm. There must be a better waybut
then, this is the Museum of Antique Concepts.
Building-Blocks into Buildings
Where do the blocks go, once the assemblers have finished
with them? Following the conveyor belt past a dozen arms, you
stroll to the end of the hall, turn the corner, and find
yourself on a balcony overlooking a vaster hall beyond. Here,
just off the conveyor belt, a block sits in a complex
fixture. Its parts are moving, and an enormous arm looms over
it like a construction crane. After a moment, the tourguide
speaks up and confirms your suspicion: "After
manufacturing, each block is tested. Large arms pick up
properly made blocks. In this hall, the larger arms assemble
almost a thousand blocks of various kinds to make a complete
The grand hall has its own conveyor belt, bearing a series of
partially completed nanocomputers. Arrayed along this grand
belt is a row of grand arms, able to swing to and fro, to
reach down to lesser conveyor belts, pluck million-atom
blocks from testing stations, and plug them into the grand
workpieces, the nanocomputers under construction. The belt
runs the length of the hall, and at the end, finished
nanocomputers turn a cornerto a yet-grander hall
After gazing at the final-assembly hall for several minutes,
you notice that nothing seems to have moved. Mere patience
won't do: at the rate the smaller arms moved in the hall
behind you, each block must take months to complete, and the
grand block-handling arms are taking full advantage of the
leisure this provides. Building a computer, start to finish,
might take a terribly long time. Perhaps as long as the blink
of an eye.
Molecular assemblers build blocks that go to block
assemblers. The block assemblers build computers, which go to
system assemblers, which build systems, whichat least
one path from molecules to large products seems clear enough.
If a car were assembled by normal-sized robots from a
thousand pieces, each piece having been assembled by smaller
robots from a thousand smaller pieces, and so on, down and
down, then only ten levels of assembly process would separate
cars from molecules. Perhaps, around a few more corners and
down a few more ever-larger halls, you would see a
post-breakthrough car in the making, with unrecognizable
engine parts and comfortable seating being snapped together
in a century-long process in a hall so vast that the Pacific
Ocean would be a puddle in the corner . . .
Just ten steps in size; eight, starting with blocks as big as
the ones made in the hall behind you. The molecular world
seems closer, viewed this way.
Stepping back into that hall, you wonder how the process
begins. In every cycle of their sluggish motion, each
molecular assembler gets a fresh tool through a tube from
somewhere beneath the floor, and that somewhere is
where the story of molecular precision begins. And so you
ask, "Where do the tools come from?", and the
tourguide replies, "You might want to take the elevator
to your left."
Stepping out of the elevator and into the basement, you see a
wide hall full of small conveyor belts and pulleys; a large
pipe runs down the middle. A plaque on the wall says,
"Mechanochemical processing concept, circa 1990."
As usual, all the motions seem rather slow, but in this hall
everything that seems designed to move is visibly in motion.
The general flow seems to be away from the pipe, through
several steps, and then up through the ceiling toward the
hall of assemblers above.
After walking over to the pipe, you can see that it is nearly
transparent. Inside is a seething chaos of small molecules:
the wall of the pipe is the boundary between loose molecules
and controlled ones, but the loose molecules are well
confined. In this simulation, your fingertips are like small
molecules. No matter how hard you push, there's no way to
drive your finger through the wall of the pipe. Every few
paces along the pipe a fitting juts out, a housing with a
mechanically driven rotating thing, exposed to the liquid
inside the pipe, but also exposed to a belt running over one
of the pulleys, embedded in the housing. It's hard to see
exactly what is happening.
The tourguide speaks up, saying, "Pockets on the rotor
capture single molecules from the liquid in the pipe. Each
rotor pocket has a size and shape that fits just one of the
several different kinds of molecule in the liquid, so the
process is rather selective. Captured molecules are then
pushed into the pockets on the belt that's wrapped over the
pulley there, then"
"Enough," you say. Fine, it singles out molecules
and sticks them into this maze of machinery. Presumably, the
machines can sort the molecules to make sure the right kinds
go to the right places.
The belts loop back and forth carrying big, knobby masses of
molecules. Many of the pulleysrollers?press two
belts together inside a housing with auxiliary rollers. While
you are looking at one of these, the tourguide says,
"Each knob on a belt is a mechanochemical-processing
device. When two knobs on different belts are pressed
together in the right way, they are designed to transfer
molecular fragments from one to another by means of a
mechanically forced chemical reaction. In this way, small
molecules are broken down, recombined, and finally joined to
molecular tools of the sort used in the assemblers in the
hall above. In this device here, the rollers create a
pressure equal to the pressure found halfway to the center of
the Earth, speeding a reaction that"
"Fine, fine," you say. Chemists in the old days
managed to make amazingly complex molecules just by mixing
different chemicals together in solution in the right order
under the right conditions. Here, molecules can certainly be
brought together in the right order, and the conditions are
much better controlled. It stands to reason that this
carefully designed maze of pulleys and belts can do a better
job of molecule processing than a test tube full of
disorganized liquid ever could. From a liquid, through a
sorter, into a mill, and out as tools: this seems to be the
story of molecule processing. All the belts are loops, so the
machinery just goes around and around, carrying and
transforming molecular parts.
This system of belts seems terribly simple and efficient,
compared to the ponderous arms driven by frantic computers in
the hall above. Why stop with making simple tools? You must
have muttered this, because the tourguide speaks up again and
says, "The Special-Assembler Exhibit shows another early
molecular-manufacturing concept that uses the principles of
this molecule-processing system to build large, complex
objects. If a system is building only a single product, there
is no need to have computers and flexible arms move parts
around. It is far more efficient to build a machine in which
everything just moves on belts at a constant speed, adding
small parts to larger ones and then bringing the larger ones
together as you saw at the end of the hall above."
This does seem like a more sensible way to churn out a lot of
identical products, but it sounds like just more of the same.
Gears like fused marbles, belts like coarse beadwork, drive
shafts, pulleys, machines and more machines. In a few places,
marbles snap into new patterns to prepare a tool or make a
product. Roll, roll, chug, chug, pop, snap, then roll and
chug some more.
As you leave the simulation hall, you ask, "Is there
anything important I've missed in this molecular
The tourguide launches into a list: "Yesthe inner
workings of assembler arms, with drive shafts, worm gears,
and harmonic drives; the use of Diels-Alder reactions,
interfacial free-radial chain reactions, and dative-bond
formation to join blocks together in the larger-scale stages
of assembly; different kinds of mechanochemical processing
for preparing reactive molecular tools; the use of
staged-cascade methods in providing feed-molecules of the
right kinds with near-perfect reliability; the differences
between efficient and inefficient steps in molecular
processing; the use of redundancy to ensure reliability in
large systems despite sporadic damage; modern methods of
building large objects from smaller blocks; modern electronic
nanocomputers; modern methods for"
"Enough!" you say, and the tourguide falls silent
as you pitch it into a recycling bin. A course in
molecular manufacturing isn't what you're looking for
right now; the general idea seems clear enough. It's time to
take another look at the world on a more normal scale.
Houses, roads, buildings, even the landscape looked different
out there beyond the Faire domeless crowded, paved, and
plowed than you remember. But why? The history books (well,
they're more than just books) say that molecular
manufacturing made a big difference; perhaps now the changes
will make more sense. Yes, it's time to leave.
As you toss your goggled, gloved jumpsuit into another bin, a
striking dark-haired woman is taking a fresh one from a rack.
She wears a jacket emblazoned with the name "Desert Rose
"How'd you like it?" she asks with a smile.
"Pretty amazing," you say.
"Yes," she agrees. "I saw this sim back when I
was taking my first molecular-manufacturing class. I swore
I'd never design anything so clunky! This whole setup really
brings back the memoriesI can't wait to see if it's as
crude as I remember." She steps into the simulation hall
and closes the door.
As the Silicon Valley Faire scenario shows, molecular
manufacturing will work much like ordinary manufacturing, but
with devices built so small that a single loose molecule of
pollutant would be like a brick heaved into a machine tool. John
Walker of Autodesk, a leading company in computer-aided design,
observes that nanotechnology and today's crude methods are very
different: "Technology has never had this kind of precise
control; all of our technologies today are bulk technologies. We
take a big chunk of stuff and hack away at it until we're left
with the object we want, or we assemble parts from components
without regard to structure at the molecular level."
Molecular manufacturing will orchestrate atoms into products
of symphonic complexity, but modern manufacturing mostly makes
loud noises. These figurative noises are sometimes all too
literal: A crack in a metal forging grows under stress, a wing
fails, and a passenger jet crashes from the sky. A chemical
reaction goes out of control, heat and pressure build, and a
poisonous blast shakes the countryside. A lifesaving product
cannot be made, a heart fails, and a hospital's heart-monitoring
machine signals the end with a high-pitched wail.
Today, we make many things from metal, by machining. From the
perspective of our standard, simulated molecular world, a typical
metal part is a piece of terrain many days' journey across. The
metal itself is weak compared to the bonds of the protein chain
or other tough nanomechanisms: solid steel is no stronger than
your simulated fingers, and the atoms on its surface can be
pushed around with your bare hands. Standing on a piece of metal
being machined in a lathe, you would see a cutting blade crawl
past a few times per year, like a majestic plough the size of a
mountain range. Each pass would rip up a strip of the metal
landscape, leaving a rugged valley broad enough to hold a town.
This is machining from a nanotechnological perspective: a process
that hacks crude shapes from intrinsically weak materials.
Today, electronics are made from silicon chips. We have already
seen the landscape of a finished chip. During manufacturing,
metal features would be built up by a centuries-long drizzle of
metal-atom rain, and hollows would be formed by a centuries-long
submergence in an acid sea. From the perspective of our
simulation, the whole process would resemble geology as much as
manufacturing, with the slow layering of sedimentary deposits
alternating with ages of erosion. The term nanotechnology
is sometimes used as a name for small-scale microtechnology, but
the difference between molecular manufacturing and this sort of
microlandscaping is like the difference between watchmaking and
Today, chemists make molecules by solution chemistry. We have
seen what a liquid looks like in our first simulation, with
molecules bumping and tumbling and wandering around. Just as
assemblers can make chemical reactions occur by bringing
molecules together mechanically, so reactions can occur when
molecules bump at random through thermal vibration and motion in
a liquid. Indeed, much of what we know today about chemical
reactions comes from observing this process. Chemists make large
molecules by mixing small molecules in a liquid. By choosing the
right molecules and conditions, they can get a surprising measure
of control over the results: only some pairs of molecules will
react, and then only in certain ways.
Doing chemistry this way, though, is like trying to assemble a
model car by putting the pieces in a box and shaking. This will
only work with cleverly shaped pieces, and it is hard to make
anything very complex. Chemists today consider it challenging to
make a precise, three-dimensional structure having a hundred
atoms, and making one with a thousand atoms is a great
accomplishment. Molecular manufacturing, in contrast, will
routinely assemble millions or billions. The basic chemical
principles will be the same, but control and reliability will be
vastly greater. It is the difference between throwing things
together blindly and putting them together with a watchmaker's
Technology today doesn't permit thorough control of the structure
of matter. Molecular manufacturing will. Today's technologies
have given us computers, spacecraft, indoor plumbing, and the
other wonders of the modern age. Tomorrow's will do much more,
bringing change and choices.
Simple Matter, Smart Matter
Today's technology mostly works with matter in a few basic
forms: gases, liquids, and solids. Though each form has many
varieties, all are comparatively simple.
Gases, as we've seen, consist of molecules ricocheting through
space. A volume of gas will push against its walls and, if not
walled in, expand without limit. Gases can supply certain raw
materials for nanomachines, and nanomachines can be used to
remove pollutants from air and turn them into something else.
Gases lack structure, so they will remain simple.
Liquids are somewhat like gases, but their molecules cling
together to form a coherent blob that won't expand beyond a
certain limit. Liquids will be good sources of raw materials for
nanomachines because they are denser and can carry a wide range
of fuels and raw materials in solution (the pipe in the
molecular-processing hall contained liquid). Nanomachines can
clean up polluted water as easily as air, removing and
transforming noxious molecules. Liquids have more structure than
gases, but nanotechnology will have its greatest application to
Solids are diverse. Solid butter consists of molecules stronger
than steel, but the molecules cling to one another by the weaker
forces of molecular stickiness. A little heat increases thermal
vibrations and makes the solid structure disintegrate into a blob
of liquid. Butterlike materials would make poor nanomachines.
Metals consist of atoms held together by stronger forces, and so
they can be structurally stronger and able to withstand higher
temperatures. The forces are not very directional, though, and so
planes of metal atoms can slip past one another under pressure;
this is why spoons bend, rather than break. This ability to slip
makes metals less brittle and easier to shape (with crude
technology), but it also weakens them. Only the strongest,
hardest, highest-melting-point metals are worth considering as
parts of nanomachines.
FIGURE 3: CARBON-SOFT AND HARD
On the left is graphitethe material called
"lead" in pencilsmade of carbon atoms. On
right is diamondthe same atoms arranged in a different
Diamond consists of carbon atoms held together by strong,
directional bonds, like the bonds down the axis of a protein
chain. (See Figure 3.) These directional bonds make it hard for
planes of atoms to slip past one another, making diamond (and
similar materials) very strong indeedten to a hundred times
stronger than steel. But the planes can't easily slip, so when
the material fails, it doesn't bend, it breaks. Tiny cracks can
easily grow, making a large object seem weak. Glass is a similar
material: glass windows seem weakand a scratch makes glass
far weakeryet thin, perfect glass fibers are widely used to
make composite materials stronger and lighter than steel.
Nanotechnology will be able to build with diamond and similar
strong materials, making small, flawless fibers and components.
In engineering today, diamond is just beginning to be used. Japan
has pioneered a technology for making diamond at low pressure,
and a Japanese company sells a speaker with excellent
high-frequency responsethe speaker cone is reinforced with
a light, stiff film of diamond. Diamond is extraordinary stuff,
made from cheap materials like natural gas. U.S. companies are
scrambling to catch up.
All these materials are simple. More complex structures lead to
more complex properties, and begin to give some hint of what
molecular manufacturing will mean for materials.
What if you strung carbon atoms in long chains with side-groups,
a bit like a protein chain, and linked them into a big
three-dimensional mesh? If the chains were kinked so that they
couldn't pack tightly, they would coil up and flop around almost
like molecules in a liquid, yet the strong bonds would keep the
overall mesh intact. Pulling the whole network would tend to
straighten the chains, but their writhing motions would tend to
coil them back up. This sort of network has been made: it is
Rubber is weak mostly because the network is irregular. When
stretched, first one chain breaks, then another, because they
don't all become taut at the same time to share and divide the
load. A more regular mesh would be as soft as rubber at first,
but when stretched to the limit would become stronger than steel.
Molecular manufacturing could make such stuff.
The natural world contains a host of good
materialscellulose and lignin in wood, stronger-than-steel
proteins in spider's silk, hard ceramics in grains of sand, and
more. Many products of molecular manufacturing will be designed
for great durability, like sand. Others will be designed to fall
apart easily for easy recycling, like wood. Some may be designed
for uses where they may be thrown away. In this last category,
nanotailored biodegradables will shine. With care, almost any
sort of product from a shoe to computer-driven nanomachines can
be made to last for a good long time, and then unzip fairly
rapidly and very thoroughly into molecules and other bits of
stuff all of kinds normally found in the soil.
This gives only a hint of what molecular manufacturing will make
possible by giving better control of the structure of solid
matter. The most impressive applications will not be superstrong
structural materials, improved rubber, and simple biodegradable
materials: these are uniform, repetitive structures not greatly
different from ordinary materials. These materials are
"stupid." When pushed, they resist, or they stretch and
bounce back. If you shine light on them, they transmit it,
reflect it, or absorb it. But molecular manufacturing can do much
more. Rather than heaping up simple molecules, it can build
materials from trillions of motors, ratchets, light-emitters, and
Muscle is smarter than rubber because it contains molecular
machines: it can be told to contract. The products of molecular
manufacturing can include materials able to change shape, color,
and other properties on command. When a dust mote can contain a
supercomputer, materials can be made smart, medicine can be made
sophisticated, and the world will be a different place. Smart materials
will be discussed in Chapter 8.
Ideas and Criticisms
We've just seen a picture of molecular manufacturing (of one
sort) and of what it can do (in sketchy outline). Now let's look
at the idea of nanotechnology itself: Where did it come from, and
what do the experts think of it? The next chapter will have more
to say on the latter point, presenting the thoughts of
researchers who are advancing the field through their own work.
The idea of molecular
nanotechnology, like most ideas, has roots stretching far
back in time. In ancient Greece, Democritus suggested that the
world was built of durable, invisible particlesatoms, the
building blocks of solid objects, liquids, and gases. In the last
hundred years, scientists have learned more and more about these
building blocks, and chemists have learned more and more ways to
combine them to make new things. Decades ago, biologists found
molecules that do complex things; they termed them
Feynman was a visionary of miniaturization who pointed toward
something like molecular nanotechnology: on December 29,
1959, in an after-dinner talk at the annual meeting of the
American Physical Society, he proposed that large machines could
be used to make smaller machines, which could make still smaller
ones, working in a top-down fashion from the macroscale to the
microscale. At the end of his talk, he painted a vision of moving
individual atoms, pointing out, "The principles of physics,
as far as I can see, do not speak against the possibility of
maneuvering things atom by atom." He pictured making
molecules, pointing clearly in the direction taken by the modern
concept of nanotechnology: "But it is interesting that it
would be, in principle, possible (I think) for a physicist to
synthesize any chemical substance the chemist writes down. Give
the orders, and the physicist synthesizes it. How? Put the atoms
down where the chemist says, and so you make the substance."
Despite this clear signpost pointing to a potentially
revolutionary area, no one filled the conceptual gap between
miniature machines and chemical substances. There was no clear
concept of making molecular machines able to build more such
machines, no notion of controllable molecular manufacturing. With
hindsight, one wonders why the gap took so long to fill. Feynman
himself didn't follow it up, saying that the ability to maneuver
atoms one by one "will really be useless" since
chemists would come up with traditional, bulk-process ways to
make new chemical substances. For a researcher whose main
interest was physics, he had contributed much just by placing the
signpost: it was up to others to move forward. Instead, the idea
of molecular machines for molecular manufacturing didn't appear
From today's viewpoint, molecular nanotechnology looks more like
an extension of chemistry than like an extension of
miniaturization. A mechanical engineer, looking at
nanotechnology, might ask, "How can machines be made so
small?" A chemist, though, would ask, "How can
molecules be made so large?" The chemist has the better
question. Nanotechnology isn't primarily about miniaturizing
machines, but about extending precise control of molecular
structure to larger and larger scales. Nanotechnology is about
making (precise) things big.
Nature gives the most obvious clues to how this can be done, and
it was the growing scientific literature on natural molecular
machines that led one of the present authors (Drexler) to propose molecular
nanotechnology of the sort described here. A strategy to
reach the goal was part of the concept: Build increasingly
complex molecular machinery from simpler pieces, including
molecular machines able to build more molecular machines. And the
motivation for studying this, and publishing? Largely the fear of
living in a world that might rush into the new technology
blindly, with ugly consequences.
This concept and initial exploratory work started in early 1977
at MIT; the first
technical publication came in 1981 in the Proceedings of
the National Academy of Sciences. For years, MIT remained the
center of thinking on nanotechnology and molecular manufacturing:
in 1985, the MIT Nanotechnology Study Group was formed; it soon
initiated an annual lecture series which grew into a two-day
symposium by 1990.
The first book on the topic, Engines
of Creation, was published in 1986. In 1988, Stanford
University became the first to offer a course in molecular
nanotechnology, sponsored by the Department of Computer Science.
In 1989, this department hosted the first major conference
on the subject, cosponsored by the Foresight Institute and Global
Business Network. With the upcoming publication of a technical book
describing nanotechnologyfrom molecular mechanical and
quantum-mechanical principles up to assembly systems and
productsthe subject will be easier to teach, and more
college courses will become available.
In parallel with the development and spread of ideas about
nanotechnology and molecular manufacturingideas that remain
pure theory, however well groundedscientists and engineers,
working in laboratories to build real tools and capabilities,
have been pioneering roads to nanotechnology. Research has come a
long way since the mid-1980s, as we'll see in the next chapter.
But, as one might expect with a complex new idea that, if true,
disrupts a lot of existing plans and expectations, some
objections have been heard.
"It Won't Work"
Life might be much simpler if these ideas about nanotechnology
had some fatal flaw. If only molecules couldn't be used to form
machines, or the machines couldn't be used to build things, then
we might be able to keep right on going with our crude
technologies: our medicine that doesn't heal, our spacecraft that
don't open a new frontier, our oil crises, our pollution, and all
the limits that keep us from trading familiar problems for
strange ones. Most new ideas are wrong, especially if they
purport to bring radical changes. It is not unreasonable to hope
that these are wrong. From years of discussions with chemists,
physicists, and engineers, it is possible to compile what seems
to be a complete list of basic, critical questions about whether
nanotechnology will work. The questioners generally seem
satisfied with the answers.
"Will Thermal Vibrations Mess
The earlier scenarios describe the nature of thermal vibration
and the problems it can cause. Designing nanomachines strong
enough and stiff enough to operate reliably despite thermal
vibration is a genuine engineering challenge. But calculating the
design requirements usually requires only simple textbook
principles, and these requirements can be met for everything
described in this book.
"Will Quantum Uncertainty Mess
Quantum mechanics says that particles must be described as
small smears of probability, not as points with perfectly defined
locations. This is, in fact, why the atoms and molecules in the
simulations felt so soft and smooth: their electrons are smeared
out over the whole volume of the molecule, and these electron
clouds taper off smoothly and softly toward the edges. Atoms
themselves are a bit uncertain in position, but this is a small
effect compared to thermal vibrations. Again, simple textbook
principles apply, and well-designed molecular machines will work.
"Will Loose Molecules Mess
Chemists work with loose molecules in liquids, and they
naturally tend to picture molecules as flying around loose. It is
possible to build nanomachines and molecular-manufacturing
systems that work in this sort of environment (biological
mechanisms are an existence proof), but in the long run, there
will be no need to do so. The Silicon Valley Faire simulation
gives the right idea: Systems can be built with no loose
molecules, making nanomechanical design much easier. If no
molecules are loose inside a machine, then loose molecules can't
cause problems there.
"Will Chemical Instability Mess
Chemists perform chemical reactions, which means that they
tend to work with reactive, unstable molecules. Many molecules,
though, can sit around in peace with their neighbors for millions
of years, as is known both from chemical theory and from the
study of molecules trapped in ancient rock. Nanomachines can be
built from the more stable sorts of structure. The only necessary
exception is in molecular assembly, where molecules must react,
but even here the reactive molecules need not be turned loose.
They can be applied just when and where they are needed in the
"Is It Too Complex, Like
An easy way to explain molecular manufacturing is to say that
it is somewhat like molecular biology: small, complex molecular
devices working together to build things and do various jobs. The
next point, however, is that molecular manufacturing is different
in every detail and different in overall structure: compare the
nanocomputers, assembler arms, and conveyor belts described above
to the shaggy, seething living cell described in the last
chapter. Biology is complex in a strange and wonderful way.
Engineers need not even understand life, much less duplicate it,
merely to build a molecular-scale factory.
"I don't see anything wrong with it. But it's so
interdisciplinarycouldn't there be a problem I can't
Nanotechnology is basically a shotgun marriage of chemistry and
mechanical engineering, with physics (as always) presiding. This
makes a complete evaluation difficult for most of today's
specialists, because each of these fields is taught separately
and usually practiced separately. Many specialists, having highly
focused backgrounds, find themselves unequipped to evaluate
proposals that overlap other disciplines. When asked to do so,
they will state feelings of discomfort, because although they
can't identify any particular problems, they can't verify the
entire concept as sound. Scientists and engineers with
multidisciplinary backgrounds, or with access to specialists from
other fields, can evaluate the idea from all sides. We'll meet
some of these in Chapter 4.
It Will Work.
When physicists, chemists, biologists, engineers, and computer
scientists evaluate those parts of nanotechnology that fall
within their disciplines, they agree: At no point would it
require new principles or violate a physical law. There may for
many years be some experts offering negative off-the-cuff
opinions, but the consensus among those who have taken the time
to examine the facts is clear. Molecular nanotechnology falls
entirely within the realm of the possible.
"It Would Work, but Isn't It a
Bad Idea to Implement It?"
If this means, "These new technologies could easily
do far more harm than good," then there is no argument,
because no one seems to disagree.
If this means, "These new technologies will certainly
do more harm than good," then we disagree: much good is
possible, much harm is avoidable, and it would be too bold to
declare any such outcome "certain."
If this means, "These new technologies should be
avoided," then we reply, "How, with what risks, and
with what consequences?" Chapters 12
and 13 conclude that it is safer to
ride the beast than to hang on to its tail while others swarm
If this means, "Don't think about it or describe it,"
then we reply, "How else are we to understand it or make
Increased human abilities have routinely been used to damage the
environment and to make war. Even the crude technologies of the
twentieth century have taken us to the brink. It is natural to
feel exhilarated (or terrified) by a prospect that promises (or
threatens) to extend human abilities beyond most past dreams (or
nightmares). It is better to feel both, to meld and moderate
these feelings, and to set out on a course of action that makes
bad outcomes less likely. We're convinced that the best course is
to focus on the potential good while warning of the potential
"But Isn't It Unlikely to
Arrive Within Our Lifetimes?"
Those in failing health may be justified in saying this;
others are expressing an opinion that may well be wrong. It would
be optimistic to assume that benefits are around the corner, and
prudent to assume that they will be long delayed. Conversely, it
would be optimistic to assume that dangers will be long delayed,
and prudent to assume that they will arrive promptly. Whatever
good or ill may come of post-breakthrough capabilities, the
turbulence of the coming transition will present a real danger.
While we invite readers to take a "What if?" stance
toward these technologies, it would be imprudent to listen to the
lulling sound of the promise "not in our lifetimes."
Even today, public acceptance of man's coming
exploration of space is slow. It is considered an event
our children may experience, but certainly not one that
we shall see.
E. Bergaust and W. Beller
From the foreword to Satellite!, written July 1957
Sputnik orbits Earth, October 1957
Footprints on Moon, July 1969
We are still many years away from
nanotechnology based on molecular manufacturing. It might even
seem that such vast, slow giants as ourselves could never make
such small, quick machines. The following sections will describe
how advances in science and technology are leading toward these
abilities. We'll try to get some feel for the road ahead, for its
length, and for how fast we're moving. We are already
surprisingly close to developing a crude molecular manufacturing
technology, and getting visibly closer every week. The first,
crude technology will enable the construction of molecular
machines that can be used to build better molecular machines,
climbing a ladder of capabilities that leads to general-purpose
molecular assemblers as good or better than those described here.
The opportunities then will be enormous. If we haven't prepared,
the dangers, too, will be enormous. Whether we're ready or not,
the resulting changes will be disruptive, sweeping industries
aside, upending military strategies, and transforming our ways of