Welcome to Scribd, the world's digital library. Read, publish, and share books and documents. See more ➡
Download
Standard view
Full view
of .
Add note
Save to My Library
Sync to mobile
Look up keyword
Like this
5Activity
×
0 of .
Results for:
No results containing your search query
P. 1
Feynman 1983

Feynman 1983

Ratings:

5.0

(4)
|Views: 2,212|Likes:
Published by Hnew
Infinitesimal Machinery
Richard Feynman
Infinitesimal Machinery
Richard Feynman

More info:

Published by: Hnew on Jan 24, 2008
Copyright:Attribution Non-commercial

Availability:

Read on Scribd mobile: iPhone, iPad and Android.
download as PDF, TXT or read online from Scribd
See More
See less

09/02/2010

pdf

text

original

 
4
JOURNAL
OF
MICROELECTROMECHANICAL SYSTEMS,
VOL.
2,
NO.
I,
MARCH
1993
Infinitesimal Machinery
Richard Feynman
Editor’s Preface-This speech, delivered in 1983, is a sequel tothe famous 1960 speech “There’s Plenty of Room at the Bottom”(reprinted in
J.
Microelectromechanical Systems,
vol.
I,
no.
I,
pp.60-66, 1992). It is remarkable in many ways. Feynman anticipatesthe sacr$cial-layer method of making silicon micromotors, the useof electrostatic actuation, and the importance of friction and con-tact sticking in such devices. He explores the persistent problem ofjinding meaningFcl applications for these tiny machines, touchinga range
of
topics along the way. And he looks at the future of com-putation using a register made of atoms, and quantum-mechanicaltransitions or computation operations.The speech was presented to a large audience of scientists andengineers at the Jet Propulsion Laboratory (JPL). Following a briefintroduction by A1 Hibbs, Feynman used slides, hand gestures, andsketches
on
a blackboard to supplement his pithy language. Thestyle was quite informal. Many of the remarks were intended toevoke laughter, and did
so;
some
were directed speci3cally at theJPL audience (such as the reference to enhancement of pictures).Working from a video recording of the speech,
I
had the benejit ofseeing the slides, the gestures, and the blackboard sketches. Theeditor’s task, in such a case, is tojind a way to capture the essenceof Feynman
s
message and spirit with minimal perturbation of hisactual words, making changes only where the lack of visual aids,or the normal fits and starts of oral presentation require
some
ad-justment.
I
was greatly aided in this task by a transcript of thespoken text prepared and orally smoothed by Nora Odendahl.Where necessary,
I
have added italicized comments to clarify termsor images being described, but have tried to keep these to a mini-mum.
I
have also added topic headings to identify major sections,and have moved some text around to improve readability. As yougo through this speech, imagine yourselfa part of the audience-
if
you are moved to laughter in spots, that is appropriate; and
if
you are moved to think anew about the problems discussed, thatindeed was Feynman’s goal.Stephen
D.
SenturiaIntroduction of Richard Feynman by A1 Hibbs- Welcome to theFeynman lecture on “Injinitesimal Machinery.
have the plea-sure of introducing Richard, an old friend and past associate. Hewas educated at MIT and at Princeton, where he received a Ph.
D.
in 1942.
In
the War he was at
Los
Alamos, where he learned howto pick combination locks-an activity at which he is still quiteskillful. He next went to Cornell,where he experimented withswinging hoops. Then, both before and during his time at Caltech,he became an expert in drumming, specializing in complex rhythms,
This manuscript is based on a talk given by Richard Feynman on Feb-ruary 23, 1983, at the
Jet
Propulsion Laboratory, Pasadena CA.
It
is re-printed with the permission
of
his estate, Carl Feynman executor.The author, deceased, was with the Califomia Institute
of
Technology,Pasadena, CA.
IEEE
Log Number 9210135.
particularly those of South America and recently those of the SouthPaci3c. At Caltech, he learned to decode Mayan hieroglyphs andtook up art, becoming quite an accomplished draftsman-special-izing in nude women. And he also does jogging.Richard received the Nobel prize, but
I
believe it was for physicsand not for any of these other accomplishments. He thinks thathappened in 1965, although he doesn’t remember the exact year.
I
have never known him to suffer from false modesty,
so
I
believehe really has forgotten which year he got the Nobel prize.
EN
Dick Davies asked me to talk, he didn’t tell
w
e the occasion was going to be
so
elaborate, withTV cameras and everything-he told me I’d be amongfriends. I didn’t realize I had
so
many friends. I wouldfeel much less uncomfortable if I had more to say. I don’thave very much to say-but of course, 111 take a longtime to say it.
Revisiting
“There’s
Plenty
of
Room
at
the
Bottom”
In
1960,
about
23
years ago, I gave a talk called“There’s Plenty of Room at the Bottom,” in which I de-scribed the coming technology for making small things. Ipointed out what everybody knew: that numbers, infor-mation, and computing didn’t require any particular size.You could write numbers very small, down to atomic size.(Of course you can’t write something much smaller thanthe size of a single atom.) Therefore, we could store
a
lotof information in small spaces, and in a little while we’dbe able to do
so
easily. And of course, that’s what hap-pened.I’ve been asked a number of times to reconsider all thethings that I talked about
23
years ago, and to see howthe situation has changed.
So
my talk today could be called“There’s Plenty of Room at the Bottom, Revisited.”As I mentioned in the
1960
talk, you could represent adigit by saying it is made of a few atoms. Actually, you’donly have to have to use one atom for each digit, but let’ssay you make a bit from
a
bunch of gold atoms, and an-other bit from a bunch of silver atoms. The gold atomsrepresent a one, and the silver atoms a zero. Suppose youmake the bits into little cubes with a hundred atoms
on
aside. When you stack the cubes all together, you can writea lot of stuff in a small space. It turns out that all the booksin all the world’s libraries could have all their informa-tion-including pictures using dots down to the resolutionof the human eye-stored in a cube
1
/
120
inch on a side.
1057-7157/93$03.00
0
993
IEEE
 
FEYNMAN: INFINITESIMAL MACHINERY
5
That cube would be just about the size you can make outwith your eye-about the size of a speck of dirt.If, however, you used only surfaces rather than the vol-ume
of
the cubes to store information, and if you simplyreduce normal scale by twenty-five thousand times, whichwas just about possible in those days, then the
Encyclo-pedia Britannica
could be written on the head of a pin,the Caltech library on one library card, and all the booksin the world on thirty-five pages of the
Saturday EveningPost.
I
suggested a reduction of twenty-five thousandtimes just to make the task harder, because due to thelimitations of light wavelength, that reduction was aboutten times smaller than you could read by means of light.You could, of course, read the information with electronmicroscopes and electron beams.Because I had mentioned the possibility of using elec-tron beams and making things still smaller, six
or
eightyears ago someone sent me a picture of a book that hereduced by thirty thousand times. In the picture, there areletters measuring about a tenth of a micron across
[passesthe picture around the audience].
I also talked in the
1960
lecture about small machinery,and was able to suggest no particular use for the smallmachines. You will
see
there has been no progress in thatrespect. And I left as a challenge the goal of making amotor that would measure
1
/64
of an inch on a side. Atthat time, the idea that I proposed was to make a set ofhands-like those used in radioactive systems-that fol-lowed another set of hands. Only we make these “slave”hands smaller-a quarter of the original hands’ size-andthen let the slave hands make smaller hands and thosemake still smaller hands. You’re right to laugh-I doubtthat that’s a sensible technique. At any rate, I wanted toget a motor that couldn’t be made directly by hand,
so
Iproposed 1
/64
f an inch.At the end of my talk, Don Glaser, who won the Nobelprize in physics-that’s something that’s supposed to begood, right?-said, “You should have asked for a motor
1
200
inch on a side, because
1
/64
nch on a side is justabout possible by hand.” And I said, “Yeah, but if I of-fered a thousand-dollar prize for a motor 1
/200
inch ona side, everybody would say ‘Boy, that guy’s a cheap-skate! Nobody’s ever going to do that.’
didn’t believeGlaser, but somebody actually did make the motor byhand!As a matter of fact, the motor’s very interesting, andjust for fun, here it is. First look at it directly with youreye, to see how big it is. It’s right in the middle
of
thatlittle circle--it’s only the size of a decimal point
or
a pe-riod at the end of a sentence. Mr. McLellan, who madethis device for me, arranged it very beautifully,
so
that ithas a magnifier you can attach-but don’t look at itthrough the magnifier until you look at it directly. You’llfind you can’t see it without the magnifier. Then you canlook through the magnifier and turn this knob, which is alittle hand generator which makes the juice to turn themotor
so
you can watch the motor
go
around
[gives theMcLellan motor to the audience to be passed around
3.
What We Can
Do
Today
Now I’d like to talk about what we can do today, ascompared to what we were doing in those days. Back then,I was speaking about machinery as well as writing, com-puters, and information, and although this talk is billed asbeing about machinery, I’ll also discuss computers andinformation at the end.My first slide illustrates what can be done today in mak-ing small things commercially. This is of course one ofthe chips that we use in computers, and it represents anarea of about three millimeters by four millimeters. Hu-man beings can actually make something on that small ascale, with wires about six microns across (a micron is amillionth of a meter,
or
a thousandth of a millimeter). Thetolerances, dimensions, and separations of some of thewires are controlled to about three microns. This com-puter chip was manufactured five years ago, and nowthings have improved
so
that we can get down to aboutone-half micron resolution.These chips are made, as you know, by evaporatingsuccessive layers of materials through masks.
[Feynmanuses “evaporating” as a generic term for all semicon-ductorprocess steps.]
You can create the pattern in a ma-terial in several ways. One is to shine light through a maskthat has the design that you want, then focus the light veryaccurately onto a light-sensitive material and use the lightto change the material,
so
that it gets easier to etch
or
getsless easy to etch. Then you etch the various materials awayin stages. You can also deposit one material after an-other-there’s oxide, and silicon, and silicon with mate-rials diffused into it-all arranged in a pattern at that scale.This technology was incredible twenty-three years ago,but that’s where we are today.The real question is, how far can we
go?
I’ll explain toyou later why, when it comes to computers, it’s alwaysbetter to get smaller, and everybody’s still trying to getsmaller. But if light has a finite wavelength, then we’renot going to be able to make masks with patterns mea-suring less than a wavelength. That fact limits us to abouta half a micron, which is about possible nowadays, withlight, in laboratories. The commercial scale is about twicethat big.
So
what could we do today, if we were to work as hardas we could in a laboratory-not commercially, but withthe greatest effort in the lab? Michael Isaacson from theLaboratory of Submicroscopic Studies (appropriate for us)has made something under the direction of an artist friendof mine named Tom Van Sant. Van Sant is, I believe, theonly truly modern artist I know. By truly modem, I meana man who understands our culture and appreciates ourtechnology and science as well as the character of nature,and incorporates them into the things that he makes.
I
would like to show you, in the next slide, a pictureby Van Sant. That’s art, right? It represents an eye. That’sthe eyelid and the eyebrow, perhaps, and of course youcan recognize the pupil. The interesting thing about thiseye is that it’s the smallest drawing a human being has
 
6
JOURNAL
OF
MICROELECTROMECHANICALSYSTEMS,
VOL. 2,
NO.
1,
MARCH
1993
ever made. It’s a quarter of a micron across-250 milli-microns- and the central spot of the pupil is somethinglike fifteen or twenty millimicrons, which corresponds toabout one hundred atoms in diameter. That’s the bottom.You’re not going to be able to see things being drawnmore than one hundred times smaller, because by that timeyou’re at the size of atoms. This picture is as far down aswe can make it.Because I admire Tom Van Sant, I would like to showyou some other artwork that he has created. He likes todraw eyes, and the next slide shows another eye by him.This is real art, right? Look at all the colors, the beauty,the light, and
so
forth-qualities that of course are muchmore appreciated as art. (Maybe some of you clever JPLguys know what you’re looking at, but just keep it toyourselves, eh?)To get some idea of what you’re looking at, we’re goingto look at that eye from a little bit further back,
so
youcan see some more of the picture’s background. The nextslide shows it at a different scale. The eye is now smaller,and perhaps you see how the artist has drawn the furrowsof the brow, or whatever it is around the eye. The artistnow wants to show the eye to us on a still smaller scale,
so
we can see a little more of the background.
So
in thisnext slide, you see the city of Los Angeles covering mostof the picture, and the eye is this little speck up in thecomer!Actually, all these pictures of the second eye areLANDSAT pictures of an eye that was made in the desert.You might wonder how someone can make an eye thatbig-it’s two and one-half kilometers across. The way VanSant made it was to set out twenty-four mirrors, each twofeet square, in special locations in the desert. He knewthat when the LANDSAT passes back and forth overhead,its eye looks at the land and records information for thepicture’s pixels. Van Sant used calculations
so
that themoment the LANDSAT looked at a particular mirror, thesun would be reflecting from the mirror right into the eyeof the LANDSAT. The reflection overexposed the pixel,and what would have been a two-foot square mirror in-stead made a white spot corresponding to an area of sev-eral acres.
So
what you saw in the first picture was a se-quence of overexposed pixels on the LANDSAT picture.Now that’s the way to make art! As far as I know, this isthe largest drawing ever made by man.If you look again at the original picture, you can seeone pixel that didn’t come out. When they went back tothe desert, they found that the mirror had been knocked
off
its pedestal, and that there were footprints from a jackrabbit over the surface.
So
Van Sant lost one pixel.The point about the two different eyes is this: that VanSant wanted to make an eye much bigger than a normaleye, and the eye in the desert was
100000
times biggerthan a normal eye. The first eye, the tiny one, was
100
000
times smaller than a normal eye.
So
you get an idea ofwhat the scale is. We’re talking about going down to thatsmall level, which is like the difference in scale betweenthe two-and-one-half-kilometer desert object and our owneye. Also amusing to think about, even though it hasnothing to do with going small, but rather with going big-what happens if you go to the next eye,
100000
timesbigger? Then the eye’s scale is very close to the rings ofSaturn, with the pupil in the middle.I wanted to use these pictures to tell us about scale andalso to show us what, at the present time, is the ultimatelimit of our actual ability to construct small things. Andthat summarizes how we stand today, as compared to howthe situation looked when
I
finished my talk in
1960.
Wesee that computers are well on their way to small scale,even though there are limitations. But
I
would like to dis-cuss something else-small machines.Small Machines-How
to
Make ThemBy a machine, I mean things that have movable partsyou can control, that have wheels and stuff inside. Youcan turn the movable parts; they are actual objects. As faras I can tell, this interest of mine in small machines is amisguided one, or more correctly, the suggestion in thelecture “Plenty of Room at the Bottom” that soon wewould have small machines was certainly a misguidedprediction. The only small machine we have is the onethat I’ve passed around to you, the one that Mr. McLellanmade by hand.There is no use for these machines,
so
I still don’t un-derstand why I’m fascinated by the question
of
makingsmall machines with movable and controllable parts.Therefore I just want to tell you some ideas and consid-erations about the machines. Any attempt to make out thatthis is anything but a game-well, let’s leave it the wayit is: I’m fascinated and I don’t know why.Every once in a while I try to find a use. I know there’salready been a lot of laughter in the audience- just saveit for the uses that I’m going to suggest for some of thesedevices, okay?But the first question is, how can we make small ma-chines? Let’s say I’m talking about very small machines,with something like ten microns (that’s a hundredth of amillimeter) for the size of a rotor. That’s forty timessmaller than the motor I passed around-it’s invisible, it’s
so
small.I would like to shock you by stating that I believe thatwith today’s technology we can easily-I say easily-con-struct motors one fortieth
of
this size on each dimension.That’s sixty-four thousand times smaller than the size
of
McLellan’s motor. And in fact, with our present technol-ogy, we can make thousands of these motors at a time,all separately controllable. Why do you want to makethem? I told you there’s going to be lots of laughter, butjust for fun, I’ll suggest how to do it-it’s very easy.It’s just like the way we put those evaporated layersdown, and made all kinds of structures. We keep makingthe structures a little thicker by adding a few more layers.We arrange the layers
so
that you can dissolve away alayer supporting some mechanical piece, and loosen thepiece. The stuff that you evaporate would be such that it

Activity (5)

You've already reviewed this. Edit your review.
1 hundred reads
1 thousand reads
jbayogan liked this
federicocanova liked this

You're Reading a Free Preview

Download
/*********** DO NOT ALTER ANYTHING BELOW THIS LINE ! ************/ var s_code=s.t();if(s_code)document.write(s_code)//-->