The early history of the scanning electron microscope

C. W. Oatley

Citation: Journal of Applied Physics 53, R1 (1982); doi: 10.1063/1.331666

View online: http://dx.doi.org/10.1063/1.331666
View Table of Contents: http://scitation.aip.org/content/aip/journal/jap/53/2?ver=pdfcov
Published by the AIP Publishing

Articles you may be interested in

Thermoluminescence in a scanning electron microscope
J. Appl. Phys. 104, 083710 (2008); 10.1063/1.3000455

Combination of a scanning tunneling microscope with a scanning electron microscope

Rev. Sci. Instrum. 59, 1286 (1988); 10.1063/1.1139710

Scanning tunneling microscope combined with a scanning electron microscope

Rev. Sci. Instrum. 57, 221 (1986); 10.1063/1.1138973

A Simple Scanning Electron Microscope

Rev. Sci. Instrum. 40, 241 (1969); 10.1063/1.1683910

The Early History of the Electron Microscope

J. Appl. Phys. 14, 434 (1943); 10.1063/1.1715011

[This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to ] IP:
130.209.6.50 On: Thu, 18 Dec 2014 21:29:28
 JAPIA,820200R2B 100
The early history of the scanning electron microscope
c. W. Oatley
University Engineering Department, Trumpington Street, Cambridge, CB2 IPZ, England

(Received 20 April 1981; accepted for publication 7 October 1981)

The article begins with an account of prewar German work, particularly that of von Ardenne,
who established the theoretical basis of a scanning electron microscope and constructed an
instrument which was primarily intended to overcome chromatic aberration when relatively
thick specimens were examined by transmission. Neither this microscope nor a different one built
a few years later in the U. S. A. attained sufficient resolution to gain acceptance and the reasons for
this are examined. The remainder of the article deals with work carried out in the Cambridge
University Engineering Department over the years from 1948 to about 1965, when the first
successful commercial instrument was produced. The contributions made by successive research
students are explained, as are also the nonscientific factors which influenced the course of the
development.
PACS numbers: 07.80. +x

The Editors have encouraged me to write this account amplification atA, is used to control the potential of the grid
of the early history of the scanning electron microscope from G of the cathode-ray tube and hence the brightness of the
a very personal point of view: to reminisce about the work picture formed on the face of the tube. Because of the one-to-
that was carried out in the Engineering Department of the one correspondence in the positions of the electron spot on S
University of Cambridge from 1948 onwards and to try to and the light spot on the face of the tube, the picture built up
explain not only what was done, but also why it was done and by the latter must, in some sense, be an image of the surface
the conditions under which it was done. First, however, it is of S. Moreover, by control of the currents in the various
necessary to give a brief account of earlier research carried deflecting coils, it can be made a highly magnified image. It
out elsewhere. is a fortunate circumstance that, apart from color, the ap-
The story of the scanning microscope and, for that mat- pearance of this image is rather similar to one which would
ter, of every other electron-optical instrument, must begin be produced by an optical microscope.
with H. Busch (1926) who studied the trajectories of charged The first instrument operating on the above principles
particles in axially-symmetric electric and magnetic fields. was built by Knoll (1935) and further work was published by
In 1926 he showed that such fields could act as particle Knoll and Theile (1939). However, these investigators were
lenses and thus laid the foundations of geometrical electron primarily interested in studying various properties of the
optics. Following this discovery the idea of an electron mi- surface such as secondary emission. They used no demagni-
croscope began to take shape and, in Berlin, two teams set fying lenses to produce a very fine probe and resolution was
out to test this possibility; one with Knoll and Ruska at the limited by the diameter of the focused spot on the specimen
Technische Hochschule, and the other with Briiche and his to something of the order of 100 /-lm.
collaborators at the A. E. G. Laboratory. The success of The principles underlying the scanning microscope as
their efforts attracted other workers to the field and, in due we know it today were clearly enunciated in a theoretical
course, a successful transmission electron microscope was paper published by von Ardenne (1938a). In this he proposed
built. to obtain a suitably fine electron probe by demagnifying the
The German scientists were aware that, in principle, it crossover with two magnetic lenses, between which the scan-
should be possible to construct a quite different type of elec- ning coils would be placed. He examined the electron-optical
tron microscope, which has since become known as a scan- aberrations and calculated the currents to be expected in
ning microscope. The principle of such an instrument is indi-
cated diagrammatically in Fig. 1. A narrow beam of
electrons from a cathode C is accelerated to high velocity ..,cI
and then passes through lenses L I' L 2 , and L 3 • These reduce L1 -1
it to an extremely fine probe, which is focused on to the
L2 -1
, I

surface of the specimen S. Current from a sawtooth gener- I

ator B passes through coils DID I , to deflect the electron I

probe, and also through coils D2D2 to deflect the beam in a I"JTII
separate cathode-ray tube. With two such sets of coils caus- Dl : Dl
ing deflections at right angles to each other, both the focused L
a
- II-'T---I
-
electron spot on the specimen and the spot on the screen of ..J.. P
the cathode-ray tube traverse zig-zag rasters in synchro- S
nism. Secondary and/or backscattered electrons which
leave S are collected at P and the resulting current, after FIG. I. Principle of the scanning electron microscope (Oatley 1972).

R1 J. Appl. Phys. 53(2), February 1982 0021-8979/82/02R 1·13$02.40 © 1982 American I nstitute of Physics R1

[This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to ] IP:
130.209.6.50 On: Thu, 18 Dec 2014 21:29:28
 probe spots of various diameters, assuming particular cur-
rent densities leaving the cathode. These calculations led
him to the conclusion that, with noise-free detectors, fluctu-
ation noise in the probe would limit resolution to about 1f.lm
LVJ A

at television scan rates and to a few nanometers with scan

times of the order often minutes. He envisaged that the mi-
croscope might be used to examine thin specimens, using
transmitted electrons with either bright-field or dark-field
illumination, or the surfaces of thick specimens, using elec-
trons which had been backscattered or secondarily emitted.
B
He even foresaw the possibility of using an electron probe to
fabricate submicroscopic structures, such as grids on photo-
graphic plates. His discussion of possible detectors for the
very small electron currents leaving the specimen showed
that, if a metal collector were followed by a thermionic valve
amplifier of the type available at that period, recording times
were likely to be unacceptably long, even if the input capaci-
tance could be reduced to the very low value of 3 pF. He
noted that matters would be greatly improved if an electron
multiplier could be used but, at that time, these devices con-
c
tained dynodes coated with cesium, which could not be ex-
posed to air. Finally, he showed that recording times could
be reduced by a factor of about one hundred if the electron
probe, after passing through a thin specimen, were allowed
to fall directly on to a photographic film.
Von Ardenne's experimental work was published in a
second paper in 1938 (von Ardenne 1938b) and further infor-
mation is given in a recent historical note (von Ardenne FIG. 2. Principle of von Ardenne's 1938 microscope.
1978). His apparatus produced an electron probe giving a
focused spot 50-100 nm in diameter and this was first used to
examine the surface of a specimen using direct collection of give rise to chromatic aberration, as it does in the ordinary
secondary and backscattered electrons. The display was on a transmission microscope, and it should be possible to exam-
cathode-ray tube with afterglow. This experiment was not ine thicker specimens. Experiments along these lines were
continued for any length of time because the apparatus was continued for some years but did not lead to a commercial
needed for the main work with transmitted electrons instrument, presumably because the results obtained did not
through thin-film specimens. Photographic recording could compete with those given by the transmission microscope.
then be used and, as indicated earlier, better resolution was The experimental scanning microscope was destroyed in an
to be expected. air raid in 1944.
A schematic diagram of von Ardenne's apparatus is Von Ardenne was the true father of the scanning elec-
shown in Fig. 2. A narrow beam of electrons from the gun A tron microscope, who had all the right ideas. His misfortune
passed successively through magnetic lenses Band C to form was to have worked at a time when experimental techniques
a focused spot on the specimen D. Electrons passing through had not advanced quite far enough to enable him to bring
the specimen fell on a photographic film attached to a cylin- those ideas to full practical fruition.
drical drum E which was caused to rotate and to progress Details of a new scanning microscope were published a
axially by means of a screw thread. The electron probe was few years later by Zworykin, Hillier, and Snyder (1942). By
deflected by currents through a pair of coils F\F2 and a simi- this time it must have been clear that, for thin transparent
lar pair at right angles to the plane of the diagram, and the specimens, competition with the transmission microscope
currents through these coils were controlled by potentiom- was unlikely to be very profitable, so the new instrument was
eters attached to the rotating drum. It was thus possible to designed for the examination of opaque specimens, from
arrange for the focused electron probe to describe a small which thin sections could not readily be cut. At that time
zig-zag raster over the surface of the specimen, while the replica techniques were in their infancy.
electrons passing through the specimen fell on a small area of The essential features of one form of the new micro-
the photographic film. The position of this area effectively scope are shown in a greatly simplified form in the schematic
moved over the film in a corresponding raster some thou- diagram of Fig. 3. Electrons from a tungsten filament F
sands of times as large. passed through a controlling grid G, where the current was
The potential advantage of this type of microscope lay chopped by the application of a 3 kHz square-wave voltage.
in the fact that electrons were not required to pass through a The electrons were then accelerated by a potential difference
lens after traversing the specimen. The spread of velocities of 10 kV applied between F and the anode A. The electron
caused by absorption in the specimen would not, therefore, beam so formed passed through electrostatic lenses Land M

R2 J. Appl. Phys. 53(2), February 1982 C. W. Oatley R2

[This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to ] IP:
130.209.6.50 On: Thu, 18 Dec 2014 21:29:28
 electron microscopy was shown in France and Brachet
Object S
I (1946) published a note suggesting that, with a noise-free
LeusM -:- detector, a resolution of 10 nm should be achieved. The in-
I
Fluorescent : P
strument constructed by Leaute and Brachet during 1949
screen
>:
f\. " , Photomultiplier
and 1950 used secondary electrons collected by a metal elec-
trode. The resulting current was amplified by thermionic
Aperture
-:~ I
vacuum tubes, so high resolution was hardly to be expected.
The above account of the earliest work on the scanning
I

LensL - II - microscope leaves me with a question to answer: Why, in the

/
face of the discouraging results that had hitherto been ob-
/
I
tained, did I think it worthwhile to reopen the matter in
Anode A -/- 1948? To give a convincing reply to this question I must go
Grid G -I-
back a few years, to explain the conditions in the University
FilamentF
n Engineering Department at that time and to say something
about my own views on university research.
FIG. 3. Schematic diagram of the microscope of Zworykin, Hillier, Before the war, graduates from the Cambridge Univer-
and Snyder (1942).
sity Engineering Department normally left immediately
after obtaining their first degree and went into industry to
to form on the specimen S a focused spot with a diameter of take a graduate apprenticeship which would later qualify
about 0.01 f..lm. S was maintained at a positive potential of them for membership of a professional Institution. Very few
about 800 V with respect to F, so that the electrons struck it remained in the Department and such research as was done
with a velocity favorable to the production of secondaries. was carried out largely by members of the teaching staff. In
The secondaries were attracted back through the lens M and light-current electrical engineering it was confined almost
were brought to a focus near the lens. Thereafter they spread entirely to problems relating to circuits.
out to strike a fluorescent screen K, in which there was a hole By 1945 there was general agreement that the prewar
to permit the passage of the primary beam. Light from K was pattern could no longer meet the needs of rapidly advancing
focused on to the cathode of a photomultiplier P and the technology and that a considerable research effort should be
output from this provided the signal from which the final built up in the Cambridge Engineering Department. To as-
image was built up. This signal consisted of a 3 kHz square- sist with this program I was appointed to a lectureship in
wave carrier, modulated by variations in the secondary emis- engineering. At the time of my appointment I was in charge
sion from the specimen. After amplification and filtering it of the army Radar Research and Development Establish-
was applied to a facsimile printer, where the final image was ment, in which I had worked for the past six years. Prior to
recorded over a period of the order often minutes. Since this that I had been a lecturer in a university physics department
arrangement provided no simple means of focusing the in- so it was natural that, in thinking of possible future research
strument, an oscilloscope was added to display the wave- projects, I should look with favor on those which could be
form of the output signal. The microscope was judged to be broadly classified as applied physics.
in focus when the highest frequency components in the From a different point of view, a project for a Ph.D.
waveform had their maximum amplitude: defocusing student must provide him with good training and, 'if he is
caused an increase in the diameter of the incident probe and doing experimental work, there is much to be said for choos-
thus reduced the amplitUdes of these components. Scanning ing a problem which involves the construction or modifica-
in the original instrument was carried out by mechanical tion of some fairly complicated apparatus. Again, I have
displacement of the specimen through links controlled by always felt that university research in engineering should be
the recorder. At a later stage, greater precision was attained adventurous and should not mind tackling speCUlative pro-
by using magnetic deflection of the electron probe, the saw- jects. This is partly to avoid direct competition with industry
tooth circuits being triggered by the recorder. With the final which, with a "safe" project, is likely to reach a solution
instrument a resolution of about 0.05 f..lm was obtained. much more quickly, but also for two other reasons which are
This microscope was the work of a highly skilled team, rarely mentioned. In the first place, university research is
backed by the resources of a very large industrial laboratory . relatively cheap. The senior staff are already paid for their
It embodied sophisticated electronic techniques and it de- teaching duties and the juniors are Ph.D. students financed
served to succeed. Yet, in the end, it was a disappointment. It by grants which are normally very low compared with indus-
was expensive; recording times were unacceptably long; fo- trial salaries. Thus the feasibility or otherwise of a specula-
cusing was not easy and the final image was marred by exces- tive project can often be established in a university at a small
sive noise. Moreover, replication techniques were being per- fraction of the cost that would be incurred in industry. So
fected at about this time and they seemed likely to have every long as the project provides good training and leads to a
advantage over the scanning microscope for the examination Ph.D., failure to achieve the desired result need not be a
of opaque specimens. So the project was discontinued and it disaster. (The Ph.D. candidate must, of course, be judged on
must have appeared that the scanning principle had finally the excellence of his work, not on the end result.) The second
been tested and found wanting. reason is rather similar. A Ph.D. student stays at the univer-
In 1946 some interest in the possibilities of scanning sity for about three years and his departure provides a conve-

R3 J. Appl. Phys. 53(2), February 1982 C. W. Oatley R3

[This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to ] IP:
130.209.6.50 On: Thu, 18 Dec 2014 21:29:28
 nient point at which the promise of his project can be re- any kind of certainty that we could build a successful scan-
viewed. Ifit seems unlikely to succeed, it can be discontinued ning microscope and I think they would not have justified a
without the dissatisfaction and discouragement which some- new industrial attempt. However, viewed as a university
times attends similar action in industry. Ph.D. project, the proposition was much more attractive.
These, then, were the thoughts at the back of my mind, The design and construction work would provide an excel-
when I came to Cambridge in 1945. I had already begun lent training in research; we should learn a great deal about
thinking about electron optics as a possible field of research the practical side of electron optics, which would be useful in
although, at that time, I knew hardly anything about the other projects; and we should almost certainly end up with a
subject. Furthermore, in 1946, V. E. Cosslett started work in microscope which gave results of some kind, which might or
the Cavendish Laboratory on transmission electron micros- might not justify further work. I decided to go ahead.
copy. Clearly, it was important to avoid trespassing on his The research student to whom this project was assigned
ground but, in the outcome, this has never caused any diffi- in 1948 was D. McMullan and in this I was doubly fortunate.
culty. It gives me great satisfaction to put on record that, Quite apart from his ability as an experimenter, which was
over a period of twenty-five years or so, until both Cosslett very great, he had, since graduating from Cambridge in
and I reached the retiring age, the two teams worked side by 1943, spent five years in industry working on radar, cathode-
side in harmony. They helped us a great deal and I like to ray tubes, and analog computers. For the job in hand, he
think that we helped them. Moreover, there was interchange could hardly have had better experience. It is unnecessary to
of senior staff in both directions. give details of the microscope which he built, since these
When I joined the Engineering Department it soon be- have already been published (McMullan 1953a and 1953b)
came clear that research could not begin at once. Research but the extent of his achievement cannot be appreciated
students were not immediately available and my own time without some account of conditions in the Engineering De-
was fully occupied with the preparation of lecture and labo- partment at that time. When I joined the Department, no
ratory courses and with the acquisition of equipment. By electronic apparatus of any kind was to hand. Shortly after
1948 I had been able to do a good deal of reading and had the war, the Government made available to universities sur-
come to the conclusion that a fresh attempt at the construc- plus equipment of all kinds and we were able to acquire a
tion of a scanning electron microscope would form a project virtually unlimited supply of valves, cathode-ray tubes, ca-
meeting the conditions I have outlined above. So far as I can pacitors, resistors, meters, and many other components. We
remember, my reasons for reaching this decision were were given some vacuum pumps and I had an initial grant of
roughly as follows. £1000 to buy some essential measuring apparatus, but this
Although the work of Zworykin, Hillier, and Snyder was the whole of our initial stock. Moreover, our annual
had not led to an instrument that was likely to be commer- income to cater for three or four different research projects
cially viable, it was by no means a failure. They had shown was very small. I On the credit side we had free access to a
that the scanning principle was basically sound and could large and well-equipped instrument workshop presided over
give useful resolution in the examination of solid surfaces. If by A. A. K. Barker whose vast knowledge of mechanical
only one could collect a much larger fraction of the electrons techniques and enthusiasm for taking on difficult jobs were
leaving the specimen and make them contribute to the out- of the utmost value.
put signal, there would be a corresponding reduction in noise Under the above conditions, McMullan realized that
in the final image and it might be possible to shorten the almost everything he needed would have to be constructed
recording time to the extent that cathode-ray tube presenta- on the spot. He set out to make a transmission electron mi-
tion became feasible. Collection of more of the electrons croscope, to which scanning facilities were added later. On
from the specimen might be achieved if these electrons went the way, he designed and built a 40 kV stabilized supply unit,
directly to the detector, instead of having to travel back electron lenses, a cathode-ray tube display unit and so forth,
through the final lens, although the specimen surface might much of it with his own hands. Electrostatic unipotential
then have to be included at an angle to the incident beam. lenses were used because power-supply stability was thereby
This idea led to further consideration of the detector. eased and the construction of the microscope column was
I was aware of some work being carried out in the Cav- somewhat simplified. It was appreciated that magnetic
endish Laboratory by A. S. Baxter (1949) on secondary-elec- lenses would have lower aberration coefficients but, in this
tron multipliers. Such multipliers had been known for more early work, it was not thought likely that lens aberrations
than a decade, but had commonly used cesium-coated dyn- would be the factors limiting the overall performance of the
odes, which could not be exposed to air. Baxter had been instrument.
experimenting successfully with beryllium-copper dynodes Figure 4 shows a schematic diagram of McMullan's mi-
which could be used in a demountable vacuum system and croscope, Fig. 5 is a photograph of the instrument, and Fig. 6
he very kindly gave me one of his multipliers. I thought it is a micrograph of etched aluminum, at a magnification of
might make an excellent detector in the scanning micro- 2000, which he obtained with it. A rereading of his Ph.D.
scope. Other factors favoring a new attack on the problem dissertation leaves one amazed at the amount of ground that
were the improvements in electronic techniques and compo-
nents that had resulted from work during the war, particu- I Later on, Government policy towards university research change

larly in cathode-ray tubes with long-persistence screens. completely and grants of a different order of magnitude became
The above considerations obviously did not add up to available.

R4 J. App!. Phys. 53(2), February 1982 C. W. Oatley R4

[This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to ] IP:
130.209.6.50 On: Thu, 18 Dec 2014 21:29:28
 R.F. Filament
oscillators bias focus 1----i--4~--__;:=='1
supplies [.
Electron gun
5-4
0kV
i '\ 200 pm aperture
Stabilized
stabilized -
supplies
supply
r-=--,
U""" I
Lens.
Fl = I~5 or 5 cm

Magnification

FIG. 4. Schematic diagram of McMul-

Deflection coils lan's microscope (McMullan 1953a and
25 pm aperture b).
Lens.
F2 = 0.5 cm
Specimen

Multiplier

6kV
Stabilized Pulse stabilized
supplies bases generator
supply

-:- .". -::-

FIG. 6. Micrograph of etched aluminum (16 kv; 300 sec; 10- 12 amp; glanc·
ing angle 25°; field of view measures 15 Jim from left to right; McMullan
FIG. 5. Photograph of McMullan's microscope (McMullan 1953b). 1953a and b).

R5 J. Appl. Phys. 53(2), February 1982 C. W. Oatley R5

[This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to ] IP:
130.209.6.50 On: Thu, 18 Dec 2014 21:29:28
 he was able to cover. As well as building the microscope, he surface contamination. Smith had also redesigned the speci-
laid the foundation of a theoretical analysis of its behavior men stage so that rotation and tilt could be varied from out-
and made quantitative measurements of its performance. He side the chamber and, on the theoretical side, he had ex-
compared his micrographs with optical micrographs of the tended McMullan's quantitative assessment of the
same specimens since, at that time, there was no certainty performance of the instrument and had given a good deal of
that the two techniques would give pictures looking alike. thought to the factors affecting contrast.
He devoted a great deal of thought to the mechanism by He now had a microscope giving a resolution of about
which contrast might be produced and was influenced by 30 nm and was ready to use it for the examination of a wide
results being obtained elsewhere with the reflection electron range of specimens (Smith and Oatley 1955). These included
microscope. In this instrument electrons which have struck metals, textile fibers, and a number of different biological
the specimen at a very small glancing angle are brought to a objects. In another direction he showed that the microscope
focus by a further lens. McMullan concluded that, in the could be used for the continuous observation of specimens
scanning microscope, good contrast should be achieved if that were changing with time. Thus the decomposition of a
fast electrons were allowed to strike the specimen at a much
larger glancing angle of about thirty degrees and if electrons
backscattered over a fairly wide solid angle entered the elec-
tron multiplier. The foreshortening of the resulting image
would then be tolerable (which was not the case with the
reflection microscope) and the tilted specimen would greatly
facilitate the passage of electrons to the multiplier. This was
the arrangement that he put into practice.
For the metal specimens, with which he was primarily
concerned, he thought that more reliable results would be
obtained if the image were formed from fast backscattered
electrons and if the slow secondaries were excluded from the
detector. This was because he expected secondary emission
to be affected by adsorbed films on the surface of the speci-
men and thus to be uncharacteristic of the underlying metal.
He therefore made careful measurements of the distribution
in angle of backscattered electrons from several metals and
of the way in which the distribution was affected by the incli-
nation of the incident beam to the surface of the specimen.
Above all, he confirmed that, with careful attention to the
collection and amplification of electrons leaving the speci-
men, it was possible to use a cathode-ray-tube display with a
(a)
frame period of a second or two. In a subsidiary experiment
he obtained images by collecting the light emitted from
grains of phosphor scanned by the electron beam. By the
time his work at Cambridge was finished there was no doubt
that we had an interesting new instrument which warranted
further development.
My next research student was K. C. A. Smith. He
joined in 1952, a year or so before McMullan left, and began
by making improvements to the microscope which had been
suggested by the earlier work; the incorporation of a stigma-
tor, the use of double-deflection scanning coils (suggested by
McMullan), an improved final lens, and means for centering
the final aperture. He then introduced a much more efficient
system for collecting electrons from the specimen which, for
a given probe current, increased the output signal by a factor
of about fifty. Moreover, by biassing the collector positively
with respect to the specimen, slow secondary electrons as
well as the fast backscattered ones could be made to contrib-
ute to the output signal. Smith then carried out a detailed
investigation of the relative usefulness of the two contribu-
tions. This showed that the secondaries normally provided (b)
more than ninety per cent of the total signal and that, con- FIG. 7. Point contact between tungsten-molybdenum whisker and germa-
trary to what McMullan had expected, they gave rise to a nium surface: (a) Before forming; (b) After forming; (field of view measures
picture showing excellent contrast, which was not spoiled by 50/lm from left to right; Smith 1956).

R6 J. Appl. Phys. 53(2), February 1982 C. W. Oatley R6

[This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to ] IP:
130.209.6.50 On: Thu, 18 Dec 2014 21:29:28
 crystal of silver azide, heated at one end, was recorded T. E. Everhart joined the team as a research student in
(McAuslan and Smith 1956) and there was a detailed investi- 1955 and took over the original McMullan microscope, as
gation of the changes taking place during the electrical form- modified by Smith. I asked him to extend our knowledge of
ing of a germanium point-contact microwave mixer (Fig. 7). the factors affecting contrast and this work led to two impor-
Finally, Smith showed that it was quite possible to obtain a tant advances in technique. Hitherto we had used the elec-
micrograph of a biological specimen when the specimen it- tron multiplier almost exclusively for the detection of elec-
self was surrounded by water vapor at a pressure great trons from the specimen and, without it, the earlier work
enough to prevent desiccation. could not have been done. Nevertheless it had two major
At the end of Smith's first year, in 1953, he was joined defects; it was bulky and could not conveniently be placed
by O. C. Wells whose major task was to design and build a close to the specimen, and it provided an output signal at a
new microscope, similar to the earlier instrument, but incor- high potential with respect to earth.
porating a number of improvements which had been shown From the outset, the possibility of using a scintillator
to be desirable. This occupied about three-quarters of his and photomultiplier had been considered, since this tech-
available time but, in his last year, he was able to examine nique had been used by Zworykin, Hillier, and Snyder
some new techniques in the use of the microscope. An at- (1942). McMullan experimented along these lines but
tempt to observe synthetic fibers under tension was hindered showed that the inorganic phosphors then available to him
by the cracking of the evaporated metal coating which is did not decay sufficiently rapidly to give satisfactory images
normally applied to prevent electrical charging of an insulat- at the faster scan rates that he was using. With these phos-
ing specimen. This led to an investigation of other possible phors the decay oflight normally results from a bimolecular
methods of preventing charging: by irradiation with positive reaction and is thus not exponential; the lower the bright-
ions, by the addition of antistatic sprays, or by collecting ness, the slower the decay. In the scanning microscope the
only high-energy electrons. The use of high-energy, back- brightness level is very low and the decay unacceptably slow.
scattered electrons involved a study of the different effects to To overcome this difficulty, Wells had built a detector sug-
be obtained by collecting electrons leaving the specimen in gested by McMullan in which the electrons first entered a
different directions. In another part of this work, Wells six-stage electron multiplier to raise the signal level. Elec-
wished to examine the interior surface of the very fine bore in trons leaving the multiplier were then accelerated through a
a spinneret through which synthetic fibers are extruded. The potential difference of 2 kV to fall on screen coated with a
normal method of collecting electrons could not be used, but zinc-oxide phosphor. The light so produced was conveyed
he showed that satisfactory micrographs could be obtained by a light pipe to a photomultiplier to provide further ampli-
by allowing the incident probe to fall on the surface after fication of the signal. Although this arrangement did not get
entry through one end of the bore and by collecting electrons rid of the electron multiplier, it did produce a signal at
which left through the opposite end after multiple collisions "earth" potential. It was then possible to use dc amplifica-
with the walls (Fig. 8). Wells also made two very useful theo- tion throughout and thus to avoid the necessity for dc resto-
retical contributions to scanning microscopy. He showed ration at the end of each scan line.
how quantitative information about the topography of a sur-
face could be obtained from stereoscopic pairs of micro- A little later I became aware that organic scintillators
graphs and he provided a detailed theory of the way in which with very fast decay times were being produced for counting
resolution would be likely to be affected by the penetration nuclear particles. I obtained a sample of this material and
and subsequent scattering of the incident electrons within Smith showed that, coupled to a photomultiplier, it formed a
the specimen (Wells 1959 ,Wells 1960, and Everhart, Wells, very convenient detector for the fast backscattered electrons
and Oatley 1959). leaving the specimen. He used a light-pipe, with the scintilla-
tor mounted close to the specimen, so as to subtend a large
solid angle. Everhart then took the next step and designed a
detector for slow secondary electrons, based on the scintilla-
tor/photomultiplier combination. This was a more compli-
cated matter since it was necessary to attract the electrons to
the detector by a relatively weak electrostatic field which
would not distort the primary beam, and then to accelerate
them so that they struck the scintillator with energies of the
order of 10 keY. A knowledge of trajectories was clearly
needed and this was before the days when a computer could
provide the answers. However, as a quite separate develop-
ment, we had in the laboratory a highly accurate trajectory
tracer in which a large electrolytic tank was coupled to a
homemade computer (Sander and Yates 1953). At the time
this trajectory tracer was being operated by M. R. Barber,
who very kindly gave expert assistance in producing a num-
ber of traces that were needed for the scanning microscope.
FIG.8. Platinum spinneret for nylon (diameter = 50 11m; Wells 1959). Using these, Everhart designed the detector shown in Fig. 9.

R7 J. Appl. Phys. 53(2), February 1982 C. W. Oatley R7

[This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to ] IP:
130.209.6.50 On: Thu, 18 Dec 2014 21:29:28
 Copper
r-~~-- g,tuze
___________~J __
- --

Focusing
) ( nng

H
I
I
I
- - - II
Plastic
scintillator

Connecting
I
I I tuhe
I I

Pcrspc"
light pipe (a.) REVERSE BIAS = 3\1. (b) REVERSE BIAS = 1 \I.
- -- GERMANIUM - INDIUM P-N uUNCTION.

FIG. 10. Germanium-indium p-n junction: (a) Reverse bias = 3 Y; (b) Re-
H.T. lead '-<::;: verse bias = 1 Y. (Everhart 1958; Everhart, Wells, and Oatley 1959).

FIG. 9. Everhart's detector (Everhart and Thornley 1960). emission coefficient exceeds unity, and he obtained some ex-
cellent micrographs of ceramics with beam voltages as low as
With relatively little modification it has remained the stan- 1 k V (Fig. 11). He added a refrigerating specimen stage to the
dard detector since that time. microscope and examined the surfaces of biological materi-
With his new detector Everhart went on to other work, als in either the frozen or the freeze-dried condition (Thorn-
to be discussed later. Detailed measurements on the detector ley 1960). He realized that neither of the microscopes con-
itself were left to R. F. M. Thornley, who came to us in 1957. structed in the Department had achieved the resolution to be
He showed that the efficiency of the scintillator was serious- expected on theoretical grounds and this led him to make a
ly impaired if it were allowed to become overheated during careful quantitative assessment of the extent to which each
polishing or if unsuitable materials were used in this process. component of the two instruments had fallen short ofperfec-
He measured the efficiency of light pipes, extended the work tion. This analysis pointed the way towards further improve-
on electron trajectories and examined the way in which they
were affected by potentials applied to different parts of the
detector. In short, he put the design of the detector on a
sound quantitative basis. His results were published in ajoint
paper with Everhart (Everhart and Thornley 1960).
Meanwhile Everhart was attacking a quite different
problem. Some time earlier Smith had noticed that the
strength of the output signal from his electron multiplier was
influenced quite markedly by the potential of the specimen.
This suggested to me that, if the specimen were a reverse-
biassed p-n-junction diode, the image should show sharp
contrast between the p and the n regions. The experiment
was tried and the contrast was found (Oatley and Everhart
1957). Everhart now used his new detector to examine this
effect in much greater detail. Once more the trajectory tracer
was used to determine the best position for the detector, its
optimim aperture, and the best potential difference between
specimen and detector to give rise to maximum contrast at
the junction. Figure 10 shows the state of the art at this time.
Thornley had taken over the microscope constructed
by Wells, had improved it in various ways, and had added an
electrostatic stigmator. He also showed that focus modula-
tion of the final electrostatic lens could be used to avoid defo-
cusing during the examination of an inclined flat specimen.
In the use of the microscope he was the first to point out that
charging of an insulating specimen can often be overcome by FIG. 11. A fault in opaque crystalline alumina (field of view measures 25 {tm
operating with a low beam voltage, where the secondary- from left to right; 1.5 keY; Thornley 1960; Thornley and Cartz 1962).

R8 J. Appl. Phys. 53(2), February 1982 C. W. Oatley R8

[This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to ] IP:
130.209.6.50 On: Thu, 18 Dec 2014 21:29:28
 ments, though Thornley himself did not have time to carry
out the necessary experimental work.
Another research student who contributed to the im-
provement of components was P. J. Spreadbury, who joined
the team in 1956. As an officer on leave from the Regular
Army, his time with us was limited to two years and I sug-
gested that he should build a very simple scanning micro-
scope, using an ordinary cathode-ray oscillograph for the
display unit. In the course of this work he designed and con-
structed an excellent 0-20 kV stabilized E.H.T. supply,
which was used by several of his colleagues. He also made
careful measurements of the performance of the electron
guns that we were then using and showed how the size,
brightness, and position of the crossover were related to con-
ditions of operation.
In an endeavor to excite more general interest in the
scanning microscope, I was now seeking projects which
would demonstrate that the instrument could provide infor-
mation which could not be obtained by the use of replicas.
Thus, to A. D. G. Stewart, who joined us in 1958, I gave the
problem of using the microscope to investigate the sputter- FIG. 12. Surface of dispenser cathode (field of view measures 20 /Lm
ing of materials by positive ions. Initially he used the micro- from left to right; Ahmed and Beck 1963).
scope that had been built by Spreadbury but, at a later stage,
he took over and completed a more ambitious instrument
that I had designed and was constructing for my own use. To ment and the new instrument was to have magnetic lenses
this he added an ion gun, so that the specimen could be bom- and properly stabilized power supplies. Under Nixon's guid-
barded while under observation. This work provided a great ance the resolution finally achieved by Pease was about 10
deal of information about the basic mechanism of sputtering, nm(Pease and Nixon 1965a). The instrument is still in use in
which need not be described here, but Stewart's major con- the Department and, in the hands of a skilled operator, its
tribution to the development of the microscope was made performance bears comparison with that of modern micro-
some time afterwards, and will be related later. scopes, though the latter have adjuncts which make them
very much easier to use.
In 1959 a rather similar project was undertaken by H. The project which I assigned to Broers was to take over
Ahmed, supervised by A. H. W. Beck, who was primarily Stewart's apparatus and continue the investigation of sput-
interested in the activation of thermionic dispenser cathodes tering. To this end he constructed a mass analyzer for the
and took over one of our microscopes as a useful tool for his positive ions, so that the effects produced by unwanted ions,
purpose. The investigation involved special problems since particularly oxygen, could be removed. In the course of this
the specimen had to be heated to temperatures exceeding work Broers noticed that, if the surface of the specimen were
1300 K, while all light was excluded from the photomulti- scanned for any length of time by the electron probe of the
plier following the scintillator. Again emission currents up scanning microscope, the rate of subsequent sputtering was
to 20 rnA had to be drawn from the specimen, while the greatly reduced by the film of contamination that had been
microscope was operating with a probe current of 10- 11 A. built up. Since this contamination is produced by decompo-
That these difficulties were successfully overcome is shown sition of residual vapors by the electron probe, at the surface
by the micrograph reproduced in Fig. 12. of the specimen, it can be deposited in any desired pattern.
My last two research students were R. F. W. Pease, who Nixon, who had now taken over the supervision of the pro-
came in 1960 and A. N. Broers in 1961. I chose their projects ject, encouraged Broers to study this effect and to use it to
and started them on their work but, in 1960, I was elected to produce extremely small structures. Thus a film of gold was
the Chair of Electrical Engineering and became Head of the evaporated onto a single-crystal silicon substrate and a pat-
Electrical Division in the Engineering Department. There tern of contamination, in the form of a grid,was then deposit-
were new laboratories to be built and new courses to· be ed on the gold. Subsequent sputtering removed gold which
planned and it soon became clear that I should no longer had not been protected by contamination, as well as the pat-
have time for the detailed supervision of research students. tern of contamination itself. It thus left on the silicon a grid
Fortunately W. C. Nixon, who had worked for several years of gold bars whose width was about 0.05 pm. This was one of
in V. E. Cosslett's electron-microscopy group in the Caven- the earliest successful attempts at microfabrication. Further
dish Laboratory, had joined our staff in 1959 and I was able work was carried out by Pease on the deposition of protec-
to hand over to him the supervision of Pease and Broers. tive layers by the electron probe, using known pressures of
Pease had been asked to design and build a new microscope particular organic vapors, rather than the residual vapor of
in an attempt to obtain resolution close to the theoretical pump oil which had been the most probable active constitu-
limit. By this time we had more money to spend on equip- ent in Broers' experiments (Broers 1965, Pease and Nixon

R9 J. Appl. Phys. 53(2), February 1982 c. w. Oatley R9

[This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to ] IP:
130.209.6.50 On: Thu, 18 Dec 2014 21:29:28
 were concerned less with the development of the microscope
as an instrument and more with its adaptation to various
fields of investigation. For the next few years this program
was largely supervised by Nixon, but he was later joined on
the teaching staff by Ahmed in 1963 and by Smith in 1965.
Thus electron-optical research in the Department has con-
tinued to flourish, but this is another story and it is not my
story.
My account of the early history of the microscope
would be sadly deficient without mention of one other per-
son who has contributed so much. Leslie Peters joined me as
a young assistant in 1946. Already a skilled instrument mak-
er and photographer he soon became familiar with the de-
tails of our work. Year after year he has acted as mentor to
successive generations of new research students, has helped
them to modify designs so that it was possible to construct
the required apparatus, and has then made it in a surprising-
ly short space oftime. I, and I think they, would wish to pay
tribute to the tremendous help that he has so willingly given
FIG. 13. Grid wires of width 250 nm formed from 400 nm of gold-platinum us and to rejoice that he is still carrying on this invaluable
alloy on a silicon slice; protected with photoresist; exposed with electron work.
beam; developed; baked; and ion-etched (Broers 1965). I must now go back in time to explain how the scanning
microscope came to be manufactured commercially. The
1965 b). Broers also investigated the electron-beam expo- story begins in 1956 when Smith had just completed his
sure ofphotoresist(Fig. 13). Ph.D. research. The development of the instrument was at
The direction of our research was now changing. We an exciting stage and I was anxious to retain Smith's services

FIG. 14. Montage of spruce fibers (field of view measures 750 p,m from left to right; Atack and Smith 1956).

Rl0 J. Appl. Phys. 53(2), February 1982 c. W. Oatley Rl0

[This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to ] IP:
130.209.6.50 On: Thu, 18 Dec 2014 21:29:28
 scope for the Institute would be built in the Engineering De-
partment, under Smith's direction. To assist with the work,
A.E.!. made available an obsolete EM 4 transmission micro-
scope, from which lenses were taken.
Since there was no shortage of money for the project,
Smith was able to build a properly engineered instrument
with magnetic lenses, which gave excellent results (Figs. 15
and 16). The work was completed in 1958 and the micro-
scope was shipped to Canada by A.E.!. Smith also went to
Canada to work at the Institute for two years (Smith 1959-
1961). This microscope gave excellent service for about ten
years, when it was replaced by a more modern instrument. It
is now in the Canadian National Museum of Science in Otta-
wa.
Following this episode, we hoped that regular commer-
cial production of the microscope might take place but
A.E.I. were not convinced that there was a real market for
instruments of this kind and, at the time, there was no other
company in England to which we could have turned.
If A.E.!. are thought to have been lacking in foresight,
the same criticism can be levelled against other firms
FIG. 15. Microscope described by Smith (1960).
throughout the world. Our results had been freely reported
for another year or two. There was, however, no post for him at conferences and published in the scientific journals, and
in the Department and his research grant had come to an there were no patents to deter manufacture of the micro-
end. At this point fate came to our aid in the guise of the Pulp scope by anyone. For whatever reason, there was a some-
and Paper Research Institute of Canada. An officer of the what frustrating period of about six years when noone was
Institute, D. Atack had been spending sabbatical leave in interested and it is perhaps worthwhile to speculate why this
Cambridge and had learned of Smith's work. It seemed to should have been so.
him that the microscope would be a valuable tool for the The first obvious point to make is that putting into man-
research undertaken by the Institute and, at his request, ufacture an instrument of this complexity is an expensive
Smith prepared a montage of micrographs of a spruce fiber operation. Apart from the work involved in settling the final
(Fig. 14). This greatly impressed the Director of the Insti- design, detailed drawings have to be prepared, sales litera-
tute, L. B. Thiesmeyer who, on his next visit to England, ture composed and arrangements made to service micro-
called on us and expressed a wish to buy a scanning micro- scopes which have left the factory. Moreover, there will usu-
scope. It was not immediately obvious how this request was ally be other projects on the stocks to compete for limited
to be met, since the microscope was not in commercical pro- effort and finance and the choice between them will depend
duction. At an earlier stage, Associated Electrical Industries on estimates of future sales.
(A.E.!.) had given us some financial help and it was under-
stood that, if the microscope appeared to be commercially In estimating the probable market for scanning micro-
viable, they would take it up. However, no steps in this direc- scopes, the crucial question was "what will it do that other
tion had yet been taken. It was finally agreed that a micro- instruments cannot do?" The answer generally given was

FIG. 16. Micrographs obtained

with Smith's microscope: (a) The
simple eyes (ocelli) at the vertex
of a fly's head. (b) The compound
eye of the same fly. (Fields of
view measure 400 J1.m and 40 J1.m
from left to right; by courtesy of
A. Rezanowich and the Pulp and
Paper Research Institute of
Canada.)

(a) (b)

R11 J. Appl. Phys. 53(2), February 1982 C. W. Oatley R11

[This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to ] IP:
130.209.6.50 On: Thu, 18 Dec 2014 21:29:28
 that it was serving the same purpose as the transmission mi-
croscope with replicas and that, at no time, was the resolu-
tion obtained with the scanning microscope superior to that
produced by replicas of appropriate specimens. What was
overlooked was that there are many specimens of which rep-
licas cannot be made and that replication is a time-consum-
ing business. Finally, that the image obtained with a replica
is a transmission image through a thin film, which does not
begin to produce the striking three-dimensional appearance
obtained with the scanning microscope. All this is obvious
today but, looking back, I can understand why industry did
not jump at the opportunity of manufacturing scanning
microscopes.
The deadlock was broken in a rather roundabout way.
In the Cavendish Laboratory, Cosslett was interested in the
microanalysis of specimens using characteristic x-rays pro-
duced by a fine electron probe incident on a very small area
of the specimen-an idea described in a patent by Hillier
(1947) but first put into practice by Castaing and Guinier
(1949). Mechanical scanning of the specimen had been tried, FIG. 17. Cambridge Instrument Company "Stereoscan" Mk 1 prototype
but Cosslett, aware of our work on the microscope, thought (Stewart and Snelling 1965).
it worthwhile to attempt electronic scanning. In 1953 the
project was given to a new research student, P. Duncumb, had already been able to make use of Smith's microscope at
who made excellent progress and eventually produced the the Pulp and Paper Research Institute. It now decided that it
scanning x-ray microanalyzer. When Duncumb left, he wanted an instrument of its own. Another potential purchas-
joined the research laboratory of Tube Investments Ltd. and er was J. Sikorski who, since 1958, had been strongly advo-
there built a properly engineered version of his cating the use of the scanning microscope for research on
microanalyzer. fibers (Sikorski 1960).
Unlike the scanning microscope, the microanalyzer However, the Instrument Company had no real idea of
gained ready acceptance because it gave information which the market for a scanning microscope and did not feel that it
could not be obtained from any other instrument, and be- could go into production on the basis of one firm order.
cause this information was needed by a great many investiga- There was therefore a delay while the situation was being
tors. Arrangements were therefore made for Tube Invest- assessed and this delay was the more protracted because de-
ments to give details of their design to the Cambridge velopment effort in the Company had to be split between the
Instrument Company, which could manufacture the microscope and the microanalyzer which, at that time,
microanalyzer. seemed likely to be the more profitable product.
S. A. Bergen, who was Chief Development Engineer of During this period Pritchard approached the Govern-
the Company, knew of our work and was persuaded by Nix- ment Department of Scientific and Industrial Research and
on and Smith that it would be a good idea to manufacture the obtained an assurance that, if the microscope came on the
scanning microscope as well as the microanalyzer, particu- market, a reasonable number of grants would be made avail-
larly since the two instruments could have many compo- able to those British universities which wished to purchase
nents in common. Eventually, in 1961, A.E.I. agreed that my models. The Company then decided to make a batch of five
obligation to this Company had been discharged and that I microscopes to test the market and these became available in
was free to look elsewhere for production of the microscope. 1965 (Fig. 17). Meanwhile, the second prototype had been
I was then able to approach H. C. Pritchard, the managing sold in 1964 to du Pont, which had been pressing for deliv-
Director of the Cambridge Instrument Company and it was ery. Many people contributed to this satisfactory outcome,
agreed that the Company would build a scanning micro- but it is appropriate to mention particularly the foresight of
scope. Bergen immediately set to work and an experimental Pritchard, the drive of Bergen and the great technical skills
model was put together in about six months. It was displayed of Stewart.
at the Institute of Physics and Physical Society Exhibition in The first four production models, sold under the trade
January 1962 and aroused a good deal of interest. Shortly name "Stereoscan," were delivered respectively to P. R.
afterwards my research student, A. D. G. Stewart, joined the Thornton of the University of North Wales, Bangor, to J.
Instrument Company, at first to work on the microanalyzer, Sikorski of Leeds University, to G. E. Pfefferkorn of the
but later to take charge of all development work on the University of Munster, and to the Central Electricity Re-
microscope. search Laboratories. By this time the Company had
By late 1962 the Company had constructed a second launched a pUblicity campaign and orders began to roll in.
prototype and, at this time, a firm order for a microscope was An additional batch of twelve microscopes was put in hand;
received from the du Pont Company in Canada. This Com- and then a further forty ...... The scanning microscope had
pany knew of our work in the Engineering Department and come of age.

R12 J. Appl. Phys. 53(2), February 1982 C. W. Oatley R12

[This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to ] IP:
130.209.6.50 On: Thu, 18 Dec 2014 21:29:28
 H. Ahmed and A. H. W. Beck, (1963): "Thermionic emission from D. McMullan, (1953b): "The scanning electron microscope and the
dispenser cathodes," J. Appl. Phys. 34, 997-998. electron-optical examination of surfaces," Electron. Eng. (England)
M. von Ardenne, (1938a): "The scanning electron microscope: Theo- 25,46-50.
retical fundamentals" (in German), Z. Phys. 109, 553-572. C. W. Oatley and T. E. Everhart, (1957): "The examination of p-n
M. von Ardenne, 11938b): "The scanning electron microscope: Practi- junctions with the scanning electron microscope," J. Electron. 2,
cal construction" (in German), Z. Tech. Phys. 19, 407-416. 568-570 and I plate.
M. von Ardenne, (1978): "The history of scanning electron micros- C. W. Oatley, W. C. Nixon, and R. F. W. Pease, (1965): "Scanning
copy and of the electron microprobe" (in German with English ab- electron microscopy," Adv. Electron. Electron Phys. 21, 181-247.
stract), Optik 50, 177-188. C. W. Oatley, (1969): "Isolation of potential contrast in the scanning
D. Atack and K. C. A. Smith, (1956): "The scanning electron micro- electron microscope," J. Sci. Instrum. (J. Phys. E) 2, }42-744.
scope. A new tool in fiber technology," Pulp Pap. Mag.Can. (Conven- C. W. Oatley, (1972/: "The scanning electron microscope. Part I. The
tion issue) 57,245-25 I. instrument." (Cambridge University Press, England).
A. S. Baxter,(l949): "Detection and analysis of low energy disintegra- C. W. Oatley, (1981): "Detectors for scanning electron microscope,"
tion particles." Ph.D. dissertation, Cambridge Univ., England. J. Phys. E: Sci. Instrum. 14, 971-976.
C. Brachet, (1946): "Note on the resolution of the scanning electron R. F. W. Pease and W. C. Nixon, (1965a): "High resolution scanning
microscope" (in French), BuU. Assoc. Tech. Marit. Aeronaut, No. 45, electron microscopy," J. Sci. Instrum. 42, 31-35.
369-378. R. F. W. Pease and W. C. Nixon, (1965b): "Microformation of fila-
A. N. Broers, (1965a): "Combined electron and ion beam processes ments," in First International Conference on Electron and Ion Beam
for microcircuits," Microe1ectron. Reliab. 4, 103-\04 and I plate. Science and Technology, Toronto, May 1964, edited by R. A. Bakish
A. N. Broers, (1965b): "Micromachining by sputtering through a (Wiley, New York) p. 220--230.
mask of contamination laid down by an electron beam," First Inter- K. F. Sander and 1. G. Yates, (1953): "The accurate mapping of elec-
national Conference on Electron and Ion Beam Science and Technol- tric fields in an electrolytic tank," Proe. lEE II, 100, pp. 167 -183.
ogy, edited by R. A. Bakish, 191-204. J. Sikorski, (1960): "Studies of fibrous structures," in Proceedings of
H. Busch, (1926): "Calculation of the path of cathode rays in the ax- the Fourth International Conference on Electron Microscopy, Berlin
ially symmetric electromagnetic field" (in German). Ann. Phys. 81, Sept. 1958, Physics, Vol. I, edited by W. Bargmann et al. (Springer
974--993. Verlag, Berlin, 1960), p. 686-707.
R. Castaing and A. Guiniert, (1949): "Application of electron probes K. C. A. Smith and C. W. Oatley, (1955): "The scanning electron
to metaUographic analysis" (in French), in Proceedings of the 1st In- microscope and its fields of application," Br. J. Appl. Phys. 6, 391-
ternational Conference on Electron Microscopy, Delft, pp. 60- 63. 399.
T. E. Everhart, (1958): "Contrast formation in the scanning electron K. C. A. Smith, (1956): "The scanning electron microscope and its
microscope." Ph.D. dissertation, Cambridge Univ., England. fields of application," Ph.D. dissertation, Cambridge Univ., England.
T. E. Everhart, O. C. WeBs, and C. W. Oatley, (1959): "Factors af- K. C. A. Smith, (1959/: "Scanning electron microscopy in pulp and
fecting contrast and resolution in the scanning electron microscope," paper research," Pulp Pap. Mag. Can. 60, T366-T371.
J. Electron. Control 7, 97-111. K. C. A. Smith, (1960): "A versatile scanning electron microscope,"
T. E. Everhart, (1960): "Simple theory concerning the reflection of in Proceedings of the European Regional Conference on Electron Mi-
electrons from solids," J. Appl. Phys. 31, 1483-\490. croscopy, Delft 1960, Vol. I, edited by A. L. Houwink and B. J. Spit
T. E. Everhart, K. C. A. Smith, O. C. Wells, and C. W. Oatley, (De Nederlandse Vereniging Voor Electronenmicroscopie, Delft,
(1960): "Recent developments in scanning electron microscopy," in 1961), p. 177-180.
Proceedings of the Fourth International Conference on Electron Mi- K. C. A. Smith, (1961): "Scanning," in Encyclopedia of Microscopy,
croscopy, Berlin, Sept. 1958, Physics, Vol. I, edited by W. Bargmann edited by G. L. Clark (Reinhold, New York), p. 241-251.
et al. (Springer Verlag, Berlin, 1960), p. 269-273. A. D. G. Stewart and M. A. Snelling, (1965): "A new scanning elec-
T. E. Everhart and R. F. M. Thornley, (1960): "Wide-band detector tron microscope," in Electron Microscopy: Proceedings of the Third
for micro-microampere low-energy electron currents," J. Sci. In- European Regional Conference, Prague, Sept. 1964, edited by M. Tit-
strum. 37, 246-248. lebach (Czechoslovak Academy of Sciences, Prague), p. 55-56.
J. Hillier, (1947): "Electron probe analysis employing x-ray spectro- R. F. M. Thornley, (1960): "Recent developments in scanning elec-
graphy," U. S. Patent 2418029. tron microscopy," in Proceedings of the European Regional Confer-
M. Knoll, (1935): "Static potential and secondary emission of bodies ence on Electron Microscopy, Delft 1960, Vol. 1, edited by A. L.
under electron irradiation" (in German), Z. Tech. Phys. 11, 467-475. Houwink and B. J. Spit, (De Nederlandse Vereniging Voor Electron-
M. Knoll and R. Theile, (1939): "Scanning electron microscope for enmicroscopie, Delft, 1961), p. 173-176.
determining the topography of surfaces and thin layers" (in German), R. F. M. Thornley and L. Cartz, (1962): "Direct examination of ce-
Z. Phys. 113, 260--280. ramic surfaces with the scanning electron microscope," J. Am.
J. H. L. McAuslan and K. C. A. Smith, (1956): 'The direct observa- Ceram. Soc. 45, 425-428.
tion in the scanning electron microscope of chemical reactions," in o. C. Wells, (1959): "Examination of nylon spinneret holes by scan-
Electron Microscopy: Proceedings of the Stockholm Conference, Sept. ning electron microscopy," J. Electron. Control 7, 373-376.
1956, edited by F. S. Sjostrand and J. Rhodin (Academic, New York, O. C. Wells, (1960): "Correction of errors in stereomicroscopy," Br.
1957), p. 343-345. J. Appl. Phys. 11, 199-201.
D. McMullan, (1953a): "An improved scanning electron microscope V. K. Zworykin, J. Hillier, and R. L. Snyder, (1942): "A scanning
for opaque specimens," Proc. lEE II 100, 245-259. electron microscope," ASTM Bull. No. 117, pp. 15-23.

R13 J. Appl. Phys. 53(2), February 1982 C. W. Oatley R13

[This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to ] IP:
130.209.6.50 On: Thu, 18 Dec 2014 21:29:28