ELECTRON MICROSCOPE

S. L. Pupko, Moscow

§ 1. The imaging of objects smaller than \(0.15\mu\) in a modern optical microscope is limited by a certain factor known as the “limit of resolution.” This factor, due to the wave nature of light, can be determined on the basis of Abbe’s formula for an optical lens:

\[ d=\frac{\lambda}{2n\sin\alpha}, \]

where \(d\) is the limit of resolution, \(\lambda\) is the wavelength of light in vacuum, \(n\) is the refractive index of the medium in which the object is immersed, and \(\alpha\) is half the angle of the light cone formed by the aperture of the lens and the object.

The greatest values of the numerical aperture of a lens obtained for the modern optical microscope are \(1.6\). Since the shortest wavelength of visible light is \(\lambda=0.4\mu\), for the limit of resolution of the optical microscope we obtain:

\[ d=\frac{0.4}{2\cdot1.6}=0.125\mu . \]

If the distance between the particles of an object is less than \(d\), then, regardless of magnification, it is impossible to obtain an image in which the structure of the object would be distinguishable. A reduction of the limit of resolution can be achieved by a corresponding decrease in the wavelength of the radiation used.

As is known, the wavelength of a stream of electrons can be expressed by the formula

\[ \lambda=\frac{h}{mv}=\sqrt{\frac{150}{E}}\cdot 10^{-8}\ \text{cm}, \]

where \(E\) is the accelerating voltage of the electrons. Hence an electron even with a small energy of \(150\ \mathrm{eV}\) has a wavelength equal to the atomic diameter. Thus, it could be expected that the use of electrons as “ultrashort-wave” light would bring enormous success in solving the problem of investigating fine structures.

The property of electrons, when passing through a magnetic or electrostatic field, to be concentrated at a focus and to give an image laid the foundation for the creation of geometrical electron optics[^1]. The development of electron optics began when the...

systems (electric and magnetic) that possess similar focusing properties—the so-called “electron lenses.”

The founder of magnetic electron optics is considered to be Busch[^1a]. In 1926 his work appeared, in which it was shown for the first time that any constant magnetic field possessing axial symmetry acts on a sufficiently narrow beam of electrons in a manner similar to how a converging lens acts on a light ray. For a complete characterization of the focusing properties of any optical system, it is necessary to know the position of its principal optical planes and the magnitude of its focal lengths. In the case of an electron lens, determining its focusing properties reduces above all to determining the trajectory of electrons in a given magnetic or electrostatic field. Busch gave the equations for the trajectory of electron motion in a nonuniform magnetic field and calculated the distance from the initial point of the electron beam to the focus. The further development of the problems touched upon and theoretically substantiated by Busch belongs to Ruska and Knoll[^2]. They created a model of a magnetic electron lens, verified its electron-optical properties theoretically and experimentally, and showed that with the aid of such a lens one can obtain a real magnified image.

The possibility of realizing an electrostatic lens as an independently focusing system was shown later—in 1931–1932—independently of one another, by Davisson and Calbick in America[^3], and by Brüche and Johannson in Germany[^4]. Brüche and Johannson used, as such a lens, axially symmetric electrostatic fields obtained by means of charged plane diaphragms with apertures situated in the immediate vicinity (0.5 mm) of the plane surface of the cathode. Two such lenses, placed at some distance from one another, constitute a model of an electron microscope with a two-stage image of the object. With electron accelerating voltages from 200 to 800 V, images of electron sources—namely, of the emitting surface of the cathode—magnified up to 150 times were obtained with the aid of this model.

Further improvement of the electrostatic lens by Johannson and Scherzer[^5] led to the creation of a short-focus lens (with focal length \(< 1\) cm) at voltages up to 1,000 V.

However, the fundamental difficulties of creating an electrostatic lens for high electron velocities (from 10,000 V and above), necessary for imaging objects not possessing their own emission, directed researchers toward the creation of a magnetic electron lens. The design of such a lens with a short focal length at high voltages was of primary importance for solving the problem of large magnifications.

For this purpose Ruska and Knoll[^6] used the so-called “short” magnetic coil (magnetic lens), analogous to a “thin” lens in ordinary optics. By a “short” lens is meant a system whose field may be considered concentrated over a segment,

small compared with the distance from the initial point to the focus. As the formula for the focal length of such a short magnetic coil with a square winding cross-section shows,

\[ f \simeq 62\,500\,\frac{U\cdot d}{I^2}\ \text{cm} \]

(where \(U\) is the accelerating voltage of the electrons in volts; \(d\) is the mean diameter of the coil; \(I\) is the number of ampere-turns), the creation of very short focal lengths at high voltages requires not only short but also very strong magnetic fields. The only solution to this problem was found in shielding a multilayer short coil with a ferromagnetic material of high magnetic permeability, with a very small air gap. Into this gap, which is a small circular slit, there was inserted “internal optics”—diaphragms with apertures of various cross-sections, where the magnetic lens proper was to be located. Lenses constructed according to this principle (known in the literature as the principle of “pole pieces”) made it possible to obtain a very short and strong magnetic field.

§ 2. The first model of a magnetic electron microscope was built by Ruska and Knoll in 1931–1932[^7]. This instrument was used to investigate objects with their own emission—oxidized cathodes (see Malov’s article[^1]).

The further development of the magnetic electron microscope proceeded along the line of applying it not only to the investigation of self-emitting surfaces, but, what is incomparably more important, to the study of small organic and inorganic objects that do not possess their own electron emission, since they could be placed in a vacuum in the form of thin films without undergoing significant structural changes.

An electron microscope suitable for investigating objects in a transmitted beam was constructed by the same authors in 1933–1934.

In designing the second model, the authors[^8] took into account many design features and methods realized and tested by them in cathode oscillographs previously improved by them. For the convenience of centering, assembly, and handling of the entire instrument, the latter was placed in a vertical position, so that the cathode—the source of electrons—forms the top of the microscope. The individual parts of the instrument are designed and fitted in such a way as to permit assembly of the instrument in various orders. The total length of the instrument from the cathode to the photographic chamber is \(1.2\ \text{m}\). The source of electrons is a gas-discharge tube with a cold aluminum cathode, surrounded by a metallic anode in the form of a tube with diaphragms of various cross-sections. The fairly narrow beam of electrons thus obtained passes through a magnetic coil, which plays the role of a condenser lens and also has diaphragms of very small cross-section. Concentrating still further, the beam falls upon the object, placed in a specially designed object chamber, shown in

Fig. 1. The latter corresponds to the specimen stage in an optical microscope. The chamber makes it possible to study several objects without having to break the vacuum in the instrument when changing the object. For this purpose, in the wall of the chamber directly facing the objective there is placed a toothed wheel, inside which there are several apertures with apertures having a clear diameter of 0.1 mm. The object is placed on these apertures. The aperture “specimen holder” is carefully polished in order to remove heat as efficiently as possible during electron bombardment. It is made of an alloy of gold with platinum; this prevents the metal of the aperture from evaporating and even from melting. To study objects located in the chamber, the latter, outside the instrument, can be moved by means of conical slides in a direction transverse to the axis of the electron beam.

Fig. 1. Object chamber

At the shortest distance from the object chamber there is an objective coil, which on the first fluorescent screen gives a magnified image observed through a window. The image of the object obtained in the first stage is projected by means of the third magnetic coil onto the final fluorescent screen. The finally magnified image can also be seen with the aid of observation windows. The image is photographed by a camera located inside the instrument, so that the image obtained in the second stage can be projected directly onto a photographic plate.

Thus, this instrument corresponds to a projection microscope of light optics. All three magnetic coils were constructed according to a single type of “pole pieces.” The improvement of the annular pole pieces of such coils consisted in selecting a suitable form of the rings, the corresponding diameter of the free aperture, and the distance between them. Also found, for practically possible accelerating voltages, were the necessary dimensions and shapes of the field. All this made it possible to achieve very short focal distances. By analogy with ordinary optics, the “monochromaticity” of the electron beam, in other words, identical velocities of the incident electrons, was the decisive factor in obtaining maximally sharp images and, consequently, the greatest possible resolving power of the electron-optical system. In the second Knoll and Ruska model, special voltage stabilizers with a high degree of stabilization were arranged, making it possible to reduce fluctuations of voltages of 40 kV and higher to several volts.

The modified and improved Ruska model gave practically a resolution limit of \(5 \cdot 10^{-2}\mu\), i.e., several times smaller than the minimum possible resolution limit of an optical microscope.

The resolution limit of an optical microscope, as has already been indicated, is determined by the aperture of the objective. For wavelengths of visible light, lowering the resolution limit requires the greatest possible increase in aperture. In the electron microscope, where the wavelength of the electrons is much smaller, the magnitude of the aperture, at those magnifications that were achieved with the aid of this model, was not of decisive importance. The aperture could be very small, since the conditions of the lens design permitted this. A decrease in aperture is advantageous in the sense of increasing the sharpness of the image. However, the reduction of the aperture size was limited by the permissible intensity of irradiation of the object. With objective apertures of 0.1, objects 1–0.5 μ thick required strong cooling in order not to undergo alteration and even destruction under electron bombardment.

Thus, the maximum magnifications obtained by Ruska in 1933 in the study of organic objects, which are to a certain extent insensitive to heating, were of the order of 8,000–10,000 times, and for inorganic objects, such as, for example, metal films, of the order of 12,000 times. Greater magnifications (30,000–40,000 times) with maximum image sharpness could be achieved, as Ruska believed[^9], only with a strong reduction of lens aberration. This could be accomplished by systematic investigations in the selection of the shape and material of the pole pieces.

Fig. 2. Magnetic supermicroscope of Borries and Ruska. General view and diagram:
1 — gas-discharge tube, 2 — shut-off chamber, 3 — condenser coil, 4 — object chamber, 5 — objective coil, 6 — objective tube, 7 — projection coil, 8 — projection tube, 9 — camera

From 1934 to 1937, in Belgium by Marton[^10], and in England by Martin and Whelpton[^11], magnetic electron microscopes were built on the basis of Ruska’s principles and design methods. Marton was the first to attempt to study biological objects with the aid of his instrument. He developed a number of methods[^12] making it possible to protect microorganisms from the destructive action of electron bombardment. Biological objects were studied with a resolution limit of 1 μ, i.e., 10 times greater than the resolution limit of a contemporary optical microscope. Martin and Whelpton likewise did not obtain significant results, and the magnifications they achieved did not differ from those of an ordinary microscope.

In 1937 the third model of Ruska’s magnetic electron microscope[^13], shown in Fig. 2, appeared. This instrument was the first—

was named Ruska’s “supermicroscope.” It was based on a design developed earlier. In designing the instrument, special attention was paid to achieving maximum convenience in working with it and in taking photographs, as well as to improving image quality at the greatest possible magnifications.

The magnetic lenses were designed according to the previously developed principle of pole pieces, and their improvement proceeded in two directions: pole pieces of a special shape were selected, and the coils were shielded with a ferromagnetic material of very high magnetic permeability. At voltages of 80 kV, the focal lengths were 2.8 mm for the objective coil and 1 mm for the projection coil. The image distance in the first and second stages was 40 cm; with this it was possible to obtain a magnification of 75 in the first stage and 400 in the second. The electron-optical magnification was 30,000. By increasing the image by optical means approximately 3 times, it was possible to obtain a total magnification of the order of 100,000.

Among the remarkable improvements of this instrument one should note the object chamber and the photographic camera. The former is provided with a special device—a “loading funnel”—arranged so that the “loading” and “unloading” of a new object occur without disturbing the vacuum in the instrument. The object is placed on an object holder—a celluloid or collodion film—and, together with the latter, is attached to a diaphragm with an aperture diameter from 0.3 to 0.03 mm. The diaphragm with the object attached to it is mounted on a holder and, by means of ground joints, is introduced into the vacuum. The entire loading process takes only 1 min. The “loading funnel” for the object is provided with a device allowing the object to be cooled in vacuum. In addition, during observation of the image the object can be moved in the horizontal plane and in a plane perpendicular to the axis of the instrument. The described device makes it possible to rotate any portion of the object so that it coincides with the center of the objective. The photographic camera is likewise provided with a special “loading funnel” for photographic plates. Loading, as well as removal of the latter, is carried out without disturbing the vacuum in the instrument, and the photographic exposure process itself is interrupted for only 3 min.

The beam of electrons obtained from a point cathode is directed, by means of a condenser coil movable in all directions, onto the object under investigation. After passing through the object, the beam, by means of regulating mechanisms, is directed onto the objective coil. The adjustment continues until a more or less sharp image of the object is obtained on an intermediate fluorescent screen coinciding with the first image plane. Through observation windows one can see the image in the first stage. Having selected the parts of the object most interesting for final magnification, one can rotate it so that these parts coincide with the center of the aperture in the intermediate screen arranged for this purpose, and in this way carry out magnification in the second stage. The second and final fluo-

rescing screen coincides with the second image plane. The final screen is also provided with observation windows, so that the ultimately magnified image can be observed by several people simultaneously. By shifting the screen in vacuum, the image can be projected directly onto a photographic plate.

The apertures of the objective were reduced to 0.01, which made it possible to obtain images with a large depth of focus. This constitutes an enormous advantage of ultramicroscopic images in comparison with microscopic ones. Despite the much greater degree of magnification, the sharpness of the image in the electron microscope is so great that it appears possible, by simply turning the object in the plane, to examine it in all its parts, without additional—mechanical or electrical—focusing devices. The limit of resolution achieved with the ultramicroscope was \(5 \cdot 10^{-3}\,\mu\).

§ 3. The electron ultramicroscope, owing to its great resolving power and the high degree of sharpness of its images, becomes a powerful tool for research in the most diverse fields of science and technology.

Its application to the study of natural and artificial clay minerals, whose submicroscopic particles are an important factor of plasticity, made it possible to show the influence of different temperatures on firing processes \(^{14}\).

Investigations in the field of the precipitation of primary crystals from a saturated solution of \(\mathrm{CaCO_3}\) are extremely interesting: at first the formation of a thin layer of finely bubbly amorphous \(\mathrm{CaCO_3}\) is observed, very stable because of the absence of any nucleation centers. When the intensity of irradiation is increased, the amorphous layer begins to move; immediately thereafter crystallization centers and cone-shaped crystals, \(0.15\,\mu\) in diameter, appear. Thus, the electron image made it possible for the first time to see the stable amorphous form of \(\mathrm{CaCO_3}\)-gel and the moment of transition from \(\mathrm{CaCO_3}\) to calcite, which, according to some data, occurs at a temperature of \(200\)—\(230^\circ\). The role of nucleation centers here was played by more intense irradiation of the object and the heating caused by it \(^{15}\) (Figs. 3 and 4).

Fig. 3. Amorphous bubbly \(\mathrm{CaCO_3}\). Electron-optical magnification \(18\,000:1\)

Metal-oxide fumes and mineral dust were also investigated with respect to their crystalline structure and their technical

suitability for testing dust-trapping filters[^16]. It proved possible to show the presence in dust-like material of particles (especially important from the point of view of industrial hygiene) with sizes from \(5 \cdot 10^{-3}\)—\(10^{-2}\ \mu\). This makes understandable the fact of absorption by the lungs of dust-like industrial poisons. Study of the submicroscopic

Fig. 4. Calcite crystallized from amorphous \(\mathrm{CaCO_3}\) at the moment of intense electron irradiation. Electron-optical magnification \(20\,000:1\).

Fig. 5. Staphylococcus aureus. Electron-optical magnification \(21\,200:1\). In the immediate surroundings of the bacterium some formations are visible, apparently belonging to metabolic products of the bacteria with the surrounding medium.

structure of biological objects has shed light on various questions of hematology, bacteriology[^17], etc.

It was shown that during the process of blood coagulation the protoplasm of blood corpuscles undergoes structural transformations associated with the loss of substances responsible for coagulation activity[^18].

In bacteriological studies, new data were discovered on the structure of the bacterial envelope and on the release of a definite form of metabolic products[^19] (Fig. 5). It is noteworthy that it was possible to establish in the bacilli of fowl tuberculosis the presence of certain chemical elements of cell structure by comparing an entirely untreated cell with one that was the same but extracted in ether. This showed that the electron image can establish not only the distribution of mass in an object, but also the distribution of certain chemical elements in the fine structure of the object, just as occurs in light optics with respect to coarser structures.

With the aid of the ultramicroscope, the so-called “blue structure” of a bird’s feather was investigated.^20 It has long been known that the blue coloration is caused not by pigment, but by the presence of fine air pores in the box-like cells of the bird’s feather, which, as had been assumed until now, have a tubular structure. However, the electron image refuted this notion. The walls of the box-like cells proved to be not tubular, but of a spongy-cellular structure. These air gaps have dimensions from \(10^{-1}\) to \(2.5 \cdot 10^{-2}\ \mu\), and the finest threads of the horny substance have a thickness of \(1.5 \cdot 10^{-3}\)—\(10^{-2}\ \mu\). Thus a disputed question in zoology was clarified.

Unreliable data, obtained by ordinary optical methods, on the structure

Fig. 6. Tobacco virus with a characteristic bending of the threads. Electron-optical magnification 20,000:1

Fig. 7. Tobacco virus. Individual threads are arranged in the form of parallel and transverse aggregates. Electron-optical magnification 20,000:1

and form of certain animal viruses were refuted thanks to electron images of such viruses prepared directly from diseased tissue. It was established that the viruses of smallpox and ectromelia are not spherical formations, but minute particles of prismatic and cubic form with rounded edges and faces.

The ultramicroscope made it possible to obtain a sufficiently exhaustive characterization of plant viruses. At very high dilutions of tobacco and potato virus sols, the lateral bonds of large aggregates proved to be broken, and it was possible to see directly individual threadlike particles (Figs. 6, 7), which are molecules of the virus \(1.5 \cdot 10^{-3}\) in diameter and from \(1.5 \cdot 10^{-2}\) to \(3 \cdot 10^{-1}\ \mu\) in length.^21

It was possible to observe directly how, with extraordinary ease, the virus molecules first formed a linear chain, then, by gradual layering and thickening, microscopic and, finally, macroscopic crystals. This investigation made it possible not only to see the “causative agent” of the diseases of tobacco and potato, but also, for the first time, to show the molecular picture of this process.

Baischer^22 investigated the sizes of primary crystals formed in various fumes (metal and metal oxide) by X-ray and electron-diffraction methods. For comparison with the data, he obtained a series of electron images of these fumes in the ultramicroscope. The advantage of the latter proved to be as follows: by means of the X-ray and electron-diffraction methods one can determine in a given colloidal aggregate only the average sizes of the smallest crystallizing particles; the electron microscope makes it possible to determine not only the “individual” sizes and shapes of separate primary particles, but also their mutual arrangement in the given compound.

Of great interest is the study of crystalline tungsten dust. As Fig. 8 shows, the tungsten particles are, as it were, in a shell of thin needle-shaped crystals—a fact hitherto unknown.

Fig. 8. Particles of powdery tungsten. Electron-optical magnification 25,000:1

One could cite a vast number of further examples illustrating the many-sided application of the electron microscope. There is no doubt that the development of the ultramicroscope will make this instrument as necessary and valuable a tool of research as the optical microscope was in its day. The electron ultramicroscope will make it possible to look far more deeply into the interior of matter and to obtain a whole series of new data on the molecular and atomic structure of matter.

§ 4. In 1939 there appeared the fourth model of the Ruska and Borries magnetic electron microscope.^23

The design of this model was based on a number of new principles. First of all, the authors sought to create an instrument that could be used in any institute for scientific research purposes, where further electron-optical investigation would be a natural continuation of optical investigations. To this end it was necessary that a researcher without special qualifications in electrical engineering and vacuum technique should be able to operate the electron microscope. Therefore the authors strove to make the instrument as compact, efficient, and simple to handle as possible. All measuring instruments, switches, regulating resistances, etc., needed in operation are placed on two distribution boards, one of which is located beneath the camera, so that the researcher, sitting at the main observation window, can easily carry out all the necessary manipulations while observing the image and taking photographs.

The development and technical improvement of the ultramicroscope have made it possible to obtain images of particles with dimensions down to \(5 \cdot 10^{-3}\ \mu\) in diameter (the size of the crystal lattice is \(10^{-2}\ \mu\)). If, on the one hand, electron beams have a very great resolving power, then, on the other hand, their ability

to penetrate through the mass of the object is very limited. Consequently, both all the accessories for investigation in an optical microscope—the specimen holder, cover glasses—and an object with a thickness corresponding to the thinnest microtome sections proved completely unsuitable when irradiated with electrons. Layers that are readily penetrable for electrons must be of colloidal dimensions; that is, in other words, they must be such formations as are usually called films in colloid chemistry. The latter may serve both as the object of investigation and as the support for the object. As supports, collodion films with a thickness of \(10^{-2}\ \mu\) are most often used. Consequently, the thickness of such a film is less than the average length of a collodion molecule, and in it, evidently, the molecules must be arranged flat. Such films prove to be completely transparent to electron rays, like a cover glass in an ordinary microscope, without showing any electron-optical structure. In addition, being prepared with uniform thickness, they exhibit great strength, not tearing when the object is applied and under electron bombardment.

An object to be investigated, dissolved in water or in some volatile organic solvent, is applied to the collodion film in the form of a drop and, after evaporation of the solvent, can be subjected to investigation. It should also be mentioned that the preparation of the object, as well as of the collodion film—the support of the object—must take place under completely sterile conditions, since the slightest contamination by foreign particles, such as dust, will completely distort the electron-optical picture, and the structure of this foreign particle will be erroneously taken for the structure of the object itself. The most suitable thickness of an object under electron irradiation is layers from \(10^{-3}\) to \(10^{-2}\ \mu\). Such thin objects, together with the support, are mounted on a grid with an aperture diameter of \(0.3\)—\(0.03\ mm\) and are introduced into the object chamber not at once. They pass through a number of stages with a gradual reduction of pressure, thanks to which they are not subjected to the possible destruction that occurs with abrupt changes of pressure \(^{24}\).

A very important question is the influence of electron bombardment on the object and on the object support. Since the number of absorbed electrons is the greater the lower their velocity, at the voltages used in the ultramicroscope—\(60\)—\(80\ kV\)—the absorption for objects \(10^{-3}\)—\(10^{-2}\ \mu\) thick amounts to only a few percent.

Investigations have shown that in the case of thicker layers of the object—\(>1\ \mu\)—there is such strong absorption of energy that the objects are heated to temperatures above \(1000^\circ\). This is evidenced by recrystallization or melting of the object, and often the film—the object holder—tears. The danger of tearing of the film decreases as the diameter of the aperture in the grid—the object holder—decreases.

The interaction of electrons with the object at the velocities used lasts about \(10^{-16}\) sec. But even over this brief period of time the influence of another very important factor is observed—the ionization of the object. The measure of ionization is taken to be the number of electrons falling on a unit surface of the object during the time it is irradiated; in the supermicroscope this quantity is of the order of \(5 \cdot 10^{-4}\) coulomb/cm. A number of investigations carried out with living microorganisms have shown that, at the voltages used of 60–80 kV and current densities of \(10^{-3}\) A/mm\(^2\), the critical electron load in some cases is of the order of \(3 \cdot 10^{-6}\) coulomb/cm, i.e., almost two orders of magnitude less than in the electron microscope. Proceeding from this, Ardenne \({}^{25}\) arrived at very skeptical conclusions, believing that the sphere of activity of the supermicroscope in the organic world is limited and that only those living organisms possessing a high degree of resistance to electron bombardment can be subjected to investigation.

Fig. 9. Electrostatic supermicroscope of Mahl (diagram)

\(Q\)—electron source, \(O\)—object, \(L_1\) and \(L_2\)—lenses, \(P\)—fluorescent screen (photographic plate)

§ 5. The development of electron microscopy up to 1939 proceeded along the line of designing and improving magnetic electron lenses for the construction of magnetic supermicroscopes. The possibility of using electrostatic lenses for the same purposes was called into question. It was thought that, on the basis of purely fundamental considerations of geometrical optics, the construction of an electrostatic electron microscope with a resolving power greater than that possessed by the optical microscope was hardly probable \({}^{26}\). However, contrary to the opinion of a number of scientists, the creation of such a microscope with great resolving power nevertheless proved possible. In 1939 such an instrument was built in Germany by Mahl \({}^{27}\). Fig. 9 shows the diagram of an electrostatic supermicroscope.

The electrostatic supermicroscope in one respect represents a greater advantage in comparison with the magnetic supermicroscope. The constant accelerating voltage, which is a most important and indispensable factor for the magnetic electron microscope, turns out to be of little importance in the electrostatic one. With a suitable choice of focal distances, the latter do not depend on fluctuations of the voltage. However, the design of this instrument proved extremely difficult, since it was necessary to create very strong and short fields in electrostatic lenses at high voltages.

The electrostatic supermicroscope is also built on the principle of a two-stage image of the object. The source of electrons is

cathode—the incandescent end of a wire, screened by a focusing Wehnelt cylinder. The accelerating voltage of the electrons is 50 kV. The electron beam is concentrated by a diaphragm with an aperture diameter of 0.1 mm and is directed onto a thin object placed directly in front of the first electrostatic lens and rotating both in the horizontal and in the vertical directions. The plate serving as the object holder is at the same time an electrode of the objective lens. With the aid of this lens an enlarged real image is obtained on the first fluorescent screen, lying near the second projection lens and coinciding with the first image plane. Part of the intermediate image can be directed through the central aperture of the screen diaphragm and projected onto a second fluorescent screen, or onto a photographic plate. The total electron-optical magnification is 10,000. The electrons leaving the object pass through the objective diaphragm with an aperture diameter of 0.1 mm, which at the same time serves as an electrode in the objective lens and coincides with the focus in the first stage. The middle electrodes of both lenses are electrically connected to the cathode. Focusing is carried out by means of mechanical devices that change the distance of the object. Strong electrostatic fields in the electric lenses could be obtained only after a suitable choice of material for the lens electrodes (chromium-nickel steel), the form of the lenses, and very careful polishing. The corresponding shape prevented possible spark discharges. Photographs obtained with the aid of this ultramicroscope show a resolution limit below \(1.5 \cdot 10^{-2}\ \mu\) ^{28, 29}.

§ 6. In 1939 a magnetic electron ultramicroscope was constructed in Canada by Prebus and Hillier ^{30}.

In constructing this microscope all the principles of Ruska’s ultramicroscope were used in their entirety. The improved instrument ^{31} already makes it possible at the present time to obtain magnifications of the order of 180,000 with a resolution limit of \(6 \cdot 10^{-3}\ \mu\).

§ 7. The development of electron microscopy raises the question of what factors limit the resolving power of the ultramicroscope and how they can be eliminated. It should be noted that at present this question is at the stage of controversy ^{32}. Experimental and theoretical investigations have revealed six factors that reduce the resolving power ^{33}: 1) diffraction of electrons in the object; 2) the influence of space charges on the trajectories of electrons; 3) aberrations of the lens; 4) spatial scattering of electrons in the object; 5) oscillations of the magnetic field; and 6) chromatic aberration.

Diffraction. The value of the diffraction magnitude can be calculated from the formula obtained from the diffraction theory of the optical microscope (Abbe) by introducing into the equation the de Broglie wavelength of the electrons. If one assumes equality of the apertures of the condenser (i.e., the aperture of the incident electron beam) and of the objective, then at accelerating voltages of 10 kV the resolution limit will be greater than \(10^{-3}\ \mu\) only in the case of aperture values below \(10^{-2}\).

Space charges. Experimental observations have shown that, with an exposure of 1 sec, sufficient to obtain a photographic background of suitable brightness, space charges limit the resolution limit by an amount somewhat less than \(10^{-4}\,\mu\). Since, in view of the undesirable increase in the electron load on the object, there is no basis for increasing its exposure time \(^{34}\), the effect of space charges may be neglected.

Lens aberrations. Experiments carried out to investigate the electron-optical properties of a magnetic objective have shown that, for purely design reasons, obtaining focal lengths below \(3\) mm is very difficult at accelerating voltages ranging from 10 to 100 kV. Under these conditions one can attain a resolution limit of \(10^{-3}\,\mu\), if the aperture is not greater than \(10^{-2}\). Lowering the resolution limit depends to a large extent on eliminating lens aberrations.

Fig. 10. Scattering of electrons by an object with a thickness of \(10^{-3}\) mm

Spatial scattering. This factor has a strong effect in the study of organic objects with a layer thickness of the order of \(1\,\mu\), corresponding to a microtome section. Experiments have shown \(^{35}\) that, at voltages of 50 kV and an aperture \(A\) of the incident electron beam of \(10^{-2}\), penetration by electrons through only 10 to \(20\%\) of the layer thickness markedly increases the aperture of the beam.

Consequently, if the instrument has a resolution limit of \(10^{-2}\)—\(10^{-3}\,\mu\), then such an object can be investigated with a resolution limit only of \(10^{-1}\,\mu\) (Fig. 10). This influence will be still greater in the case of studying elements located inside the section, such as, for example, the internal contents of cell nuclei. The increase in the resolution limit caused by this factor in the case of such thick layers of the object can be eliminated to a considerable degree by increasing the accelerating voltage. The greater the velocity of the electrons, the less the sharpness of the images suffers from the spatial scattering of electrons by the object layer.

Fluctuations of the magnetic field. Calculations show that the resolution limit, expressed in millimeters, is approximately equal to \(1/4\) of the magnitude of the field fluctuations expressed in gauss. Consequently, in order to attain a resolution limit of \(10^{-3}\,\mu\), the fluctuations of the magnetic field in the most sensitive region of passage of the electron beam must be no more than \(2\)—\(4 \cdot 10^{-6}\) gauss.

Achieving such constancy of the field is a very difficult condition and requires special measures.

Chromatic aberration is caused chiefly by fluctuations of the accelerating voltage of the electrons and by the scattering of electron velocities when passing through the object and the film of the object holder. The decisive factor is the different scattering of the velocities of the electrons when they pass through the layer of the object. This question has been comprehensively investigated[^36] in connection with the initial velocities of the electrons and the thickness of the object layer. The results of some measurements carried out in the indicated region can be expressed by means of Bethe’s formula[^37]

\[ v_0^4 - v^4 = ax, \tag{1} \]

where \(v_0\) and \(v\) are, respectively, the velocity of the incident electron and its most probable velocity after passing perpendicularly through a metallic film of thickness \(x\) cm. The constant \(a\) depends on the substance of the film and is connected with the atomic number \(Z\), the density \(\rho\), and the atomic weight \(A\) of the element constituting the film. According to Bethe’s formula,

\[ a = k \frac{\rho}{A} Z, \tag{2} \]

where \(k\) is a constant.

If the velocities \(v_0\) and \(v\) are converted respectively into the accelerating potential \(E+\Delta E\) and \(E\), and if the value of \(a\) for Al is substituted into the formula:

\[ a = 5.5 \cdot 10^{-42}, \]

then formulas (1) and (2) can be brought to the following form:

\[ E \cdot \Delta E = 22.1 \cdot 10^9 \left(\rho \frac{Z}{A}\right) x. \tag{3} \]

These formulas were obtained in the investigation of films of thickness \(x > 10^{-4}\) mm. Thus, in investigating objects—for example, the thinnest section from a microtome, \(10^{-4}\) mm, with atomic properties similar to Al—according to formula (3) one should expect a scattering of velocities within 580 V. It follows from this that if the objective lens has a resolving limit of \(1.1 \cdot 10^{-4}\,\mu\), then such an object, causing the indicated magnitude of chromatic aberration, can be investigated with a resolving limit only of \(5.4 \cdot 10^{-2}\,\mu\). This is not much below the resolving limit of the modern optical microscope.

An evaluation of all the indicated factors which, to a greater or lesser degree, limit the resolving power of the modern electron microscope leads most investigators to the conclusion that the decisive quantity is the chromatic aberration, and only its maximum reduction will make it possible to attain the theoretically permissible resolving power at the present time. It should be noted that Ruska[^32] did not fully agree with this opinion, considering the decisive factor to be the aberration of the lens.

To make clearer to what limits the theoretically admissible resolving power of the modern electron microscope can be raised, we shall briefly consider the mechanism of image formation \(^{38,39,40}\).

A parallel beam of electrons issuing from a point cathode is concentrated, entering the aperture of the condenser, with an aperture of \(10^{-3}\). This narrow beam falls on the object, being partly absorbed but mainly scattered, and with the aid of magnetic or electrostatic lenses gives an enlarged image on a fluorescent screen or photographic plate. Since we must assume that the object has regions of different thickness or density, it is natural that electrons are scattered from the denser regions at a larger angle than from the less dense ones. Consequently, in the first case far fewer electrons enter the opening of the objective aperture than in the second case. The number of electrons entering is also limited by the aperture of the objective, which usually has values of \(10^{-2}\). If the aperture of the scattered electrons is larger than the aperture of the objective, then they do not enter the aperture and do not participate in forming the image. Electrons, however, with an aperture smaller than or equal to \(10^{-2}\), enter the aperture and are distributed in the image plane—also owing to aberration of the lens—over a comparatively larger area, and the corresponding places on the fluorescent screen will appear darker than their surroundings. Those electrons which fall on the thinner parts of the object, leaving the latter, retain a direction parallel to the axis of the beam and the initial value of the aperture \(10^{-3}\). Owing to their small aperture these electrons fill only part of the opening of the objective aperture and form maximally sharp corresponding regions in the image plane. The scattered electrons caught by the aperture opening do not determine the sharpness of the images, but only the richness of contrasts. However, strong narrowing of the aperture opening would lead only to an increase in the exposure time of the object and to still more pronounced diffraction phenomena.

Fig. 11. Position of the first lateral diffraction maximum

The picture is somewhat different when the object possesses such a fine structure that already the first lateral diffraction maxima of it form a noticeable angle with the optical axis. Since, according to Abbe, these diffraction maxima are certainly necessary in order that two points of the object may be imaged separately, the proper resolution limit will be reached when these maxima in the image plane are displaced, owing to aberration of the lens, by a distance exactly equal to the distance between the two image points according to Gauss. If \(d\) is the distance between two points of the object and \(\lambda\) is the wavelength of the electrons in the plane

if it passes through the center of the object, then the first lateral maximum, as is seen from Fig. 11, forms with the axis an angle

\[ \delta=\frac{\lambda}{d}. \]

As shown by the calculations of Rebsch\({}^{41}\), the best modern electron lenses give a displacement over the segment

\[ \rho=\frac{1}{4} f V \delta^{3}, \]

where \(f\) is the focal length of the objective and \(V\) is the magnification. According to Gauss, the distance between two points of the image is equal to \(Vd\). Hence we obtain the value for the minimum distance still resolvable in a modern microscope:

\[ Vd_{\min}\simeq \frac{f}{4} V\delta^{3}=\frac{f}{4}V\left(\frac{\lambda}{d_{\min}}\right)^{3}; \]

therefore,

\[ d_{\min}\simeq \frac{\lambda}{\sqrt{2}}\sqrt[4]{\frac{f}{\lambda}}. \]

If the expression obtained is analyzed more deeply, it turns out that it differs from Abbe’s formula for the resolution limit of an ordinary optical microscope chiefly by the factor \(\sqrt[4]{\frac{f}{\lambda}}\). For modern electron microscopes this factor is equal to 100. Consequently, the theoretically admissible resolution limit of the electron microscope does not lie near values of the wavelength, as is the case with the optical microscope, but near one hundred wavelengths. Since the voltages used are approximately 100 kV, an electron with an energy of 100 KeV has a wavelength of about \(4\cdot10^{-10}\) cm, i.e. \(1/25\) of an atomic diameter, and hence we can reach a resolution limit \(\simeq 4\cdot10^{-10}\cdot100=4\cdot10^{-8}=4\) Å. It is therefore clear that in the future there is the possibility of seeing objects extending over several atomic diameters\({}^{42}\).

§ 8. Ardenne’s universal microscope, built in 1940, surpasses in its resolving power both Ruska’s magnetic supermicroscope\({}^{43}\) and Mahl’s electrostatic supermicroscope\({}^{28}\).

The construction of this instrument was based on a number of new principles\({}^{44}\). The design of the instrument was carried out with such calculation that, at the investigator’s choice, it would be possible to work both with magnetic lenses and with electrostatic ones. This makes it possible to compare not only magnetic and electrostatic electron microscopes, but, chiefly, the electron-optical properties of individual lenses, especially the properties of the magnetic and electrostatic objective. In addition, the possibility has been provided of an instantaneous transition from a bright image background to a dark one. Finally, structural adaptations have been made for obtaining electron-microscopic stereo images. By these features the universal microscope differs sharply from the two kinds of fundamentally different electron microscopes existing at the present time. Fig. 12 presents a general view of this instrument.

Fundamental improvements to the universal ultramicroscope were achieved thanks to the design features of the instrument. Any part can be removed in vacuum from the fully assembled instrument without the need to disturb all the other parts and the entire instrument as a whole. It is possible, for example, to replace the pole pieces of the magnetic lenses with a system of electrodes of electrostatic lenses; to replace the objective or projection lens with another having different optical properties; to replace stops of a given aperture with a stop having a larger or smaller aperture. In exactly the same way, one can easily change object chambers, fluorescent screens, photographic cameras, etc. It should be especially noted that the centering and replacement of stops with different apertures can take place in vacuum by means of special mechanisms. This innovation proved to be extremely expedient. It made it possible, for the first time, to carry out in vacuum the transition from a light to a dark background. If focusing was performed visually on a light photographic background or with large apertures of the objective stops, then photographing of the object could be carried out immediately after a rapid switch to a dark photographic background or to stops with smaller apertures. This was facilitated further by the fact that the author designed a special auxiliary device for setting to maximum sharpness.

Fig. 12. Ardenne universal ultramicroscope

Another very important feature of this instrument is its complete insensitivity to mechanical vibrations. This is achieved by a rigid mechanical connection of the massive pole pieces with the object holder at the shortest possible distance from one another. Owing to this and to the very perfect magnetic (permalloy) shielding of the entire unified system—objective—object chamber—the instrument does not require special foundations and can operate with maximum resolving power in an ordinary working room.

The accelerating voltages in the universal microscope are of the order of 100 kV. At voltages of 70 kV it has been possible to attain focal lengths of 0.9 mm for the objective lens, with an aperture of \(3\cdot 10^{-3}\) mm, and 1 mm for the projection lens. The image distance in each stage is 65 cm, so that with a focal length of 1 mm in the first and in the second stages the total electron-optical magnification will be of the order of 500,000. The resolving power achieved with the universal microscope is \(30\ \text{Å} — 3\cdot 10^{-3}\ \mu\).

§ 9. The apertures used by the electron microscope are three orders of magnitude smaller than the apertures of the optical microscope. As a consequence, the depth of focus of images obtained with the electron microscope is, by the same order of magnitude, greater than in the optical microscope. In the electron microscope it is possible to obtain images with uniform sharpness even of objects that have a large spatial extent.

By depth of focus is meant the distance in millimeters by which the object can be displaced in the direction of the optical axis before the resulting loss of sharpness reaches the limiting values of the resolving power of the microscope. Proceeding from this definition, we obtain

\[ T=\frac{2fd}{D}, \]

where \(T\) is the depth of focus, \(f\) is the focal length of the objective, \(d\) is the resolution limit, and \(D\) is the effective diameter of the objective.

If the relative aperture in the universal electron microscope is

\[ \frac{D}{f}=3\cdot 10^{-3}, \]

and the resolution limit is \(3\cdot 10^{-6}\) mm, then it is possible to image objects extending in the direction of the optical axis by 2 μ, with the greatest uniform sharpness[^45]. Thus the electron microscope is an exceptionally perfect instrument for stereoscopy, promising enormous achievements in this field. With the aid of the universal electron microscope, super-stereomicroscopy was accomplished for the first time, making it possible to see the spatial structure of an object.

A system specially constructed in the universal supermicroscope (objective—object chamber) makes it possible to displace in vacuum the diaphragms—the object holders—by a definite angle (from \(4 — 15^\circ\)) in the direction of the instrument axis. The axis of displacement coincides exactly with the observed sector of the object. In this case, by precisely regulating the angular displacement, it is possible, with two displacements of one and the same sector of the object, to obtain two equivalent stereoscopic images. The latter differ only in that this sector is penetrated by electrons from two sides.

Super-stereoscopy arose at the beginning of 1940 and immediately opened wide horizons to researchers. It makes it possible to see the entire spatial structure of an object and to obtain a more complete conception of its structure.

§ 10. The great resolving power of the universal ultramicroscope made it possible to obtain a number of very valuable data in the field of the structure of catalysts, high-molecular substances, viruses and bacteria, diatoms and crystals. With the aid of this remarkable instrument it also proved possible to see individual molecules.

As is known, the catalytic action of certain metals in the colloidal state—platinum, nickel, etc.—is explained to a large extent by the dispersity of the particles, and, consequently, by the increase in

Fig. 13. Platinum catalyst prepared by Paal’s method. Under the influence of the protective colloid, particles ranging in size from 30 to 100 Å form threadlike deposits. Magnification 75,000 times

Fig. 14. Used palladium asbestos. In the left corner—pure asbestos fibers. Magnification 50,000 times

the total surface area. Electron images made it possible to examine, in sols of these metals prepared by various methods, the shape of particles with dimensions in some cases from 30 to 100 Å. Figure 13 shows a platinum catalyst at a magnification of 75,000 times, with particles averaging 50 Å. In this case a previously unknown phenomenon was observed: individual particles are deposited one on another, forming accumulations in the form of long chains. This phenomenon probably occurs under the influence of the protective colloid.

Studies of asbestos and palladium asbestos revealed the remarkable fact that asbestos consists of the thinnest threads (sometimes 30 Å thick), which split easily.

Figure 14 shows used palladium asbestos after the catalytic combustion of hydrogen in oxygen at a temperature of 400°. In the lower left corner are shown pure asbestos fibers without

palladium. It can be clearly seen that the asbestos fibers, after use, became more fluffy (shaggy). Apparently this occurs as a result of the release of water. Palladium is deposited on the asbestos fibers in the form of crystalline particles ranging in size from 70–1000 Å. A comparison of the used palladium asbestos with unused asbestos, shown in Fig. 15, reveals that the palladium particles undergo no structural changes. On the basis of these studies, very important conclusions were drawn about the character and nature of the catalytic action of these colloids^46.

Fig. 15. Unused palladium asbestos. The size of the finest asbestos fibers is 30 Å. Magnification 50,000×. This electron image shows that the universal ultramicroscope gives exceptional sharpness of image and an enormous field of view

In the study of the morphology of fibrous substances, the universal ultramicroscope proved to be a natural continuation of the optical microscope. The strength and elasticity of high-molecular fibrous substances, such as, for example, cellulose and its derivatives, depend not only on the chemical characteristics and sizes of the macromolecules composing them. This is determined to a considerable degree by the mutual arrangement of the individual threads—molecules—in the primary aggregates (fibers, micelles) and by the size of these aggregates.

The investigation of the physical and chemical properties of crystals of β-polyoxymethylene is of enormous importance for understanding the structure of such a fibrous substance as cellulose. As Staudinger and Sauter^47 showed, crystals of β-polyoxymethylene can be mechanically destroyed under pressure to obtain a bundle of split fibers. Such individual fibers were studied optically, and it was established that the thinnest of the visible fibers are on the order of 0.2–0.6 μ. By subjecting these fibers to still stronger mechanical splitting, it was possible for the first time, with the aid of the universal ultramicroscope, to see individual finest—

Fig. 16. Electron image of an individual finest fiber. Magnification 75,000×

…fibers—bundles of molecules having a diameter of 5–10 mμ (50–100 Å). Fig. 16 shows such an individual fiber, 50 Å in size, magnified 75,000 times. Owing to its low resolving power as compared with the ultramicroscope, the optical microscope could in this case detect only whole aggregates of such molecular bundles.

Fig. 17 gives an optical image of crystals of β-polyoxymethylene treated with 2.5N NaOH, obtained by Staudinger and Sauter at a magnification of 840 times. In the microphotograph the structure of these crystals is visible. Owing to the process of periodic growth, the individual finest fibers are arranged in the crystal in the form of connected and regularly alternating bundles. Since alkali acts only on the terminal OH groups of the molecules of polyoxymethylene hydrate, even the microscopic image shows quite clearly the transverse striation, periodically distributed along the entire length of an individual fiber-molecule. This striation was determined optically by Staudinger and was 0.36 μ. However, these values lie at the limit of the resolving power of the optical microscope; therefore all conclusions about the degree of polymerization and the atomic weight (56,000) of β-polyoxymethylene hydrate were made by Staudinger conditionally.

Fig. 17. Optical image of crystals of β-polyoxymethylene treated with 2.5N NaOH. Magnification 840 times

Fig. 18. Electron image of crystals of β-polyoxymethylene treated with 2.5N NaOH. Magnification 50,000 times

An electron image of crystals of β-polyoxymethylene treated in the same way (2.5N NaOH) shows a periodically

alternating cells corresponding to the periodic striation determined as \(0.36\,\mu\). Since this image was obtained with a resolving power two orders of magnitude greater than that obtained by Staudinger and at a magnification of 75,000 times, it was possible to draw correct conclusions about the degree of polymerization on the basis of the etching period of the crystal. Fig. 18 shows the edge of an etched crystal of \(\beta\)-polyoxymethylene. Individual, extremely fine fibers connecting alternating segments in the etched crystal are visible with great sharpness. The finest fibers in these periodically alternating cells proved to be of the order of \(50\)—\(100\ \text{\AA}\). These values coincide with those obtained for the finest fibers during the mechanical

Fig. 19. Hemocyanin molecules.
Concentration \(10^{-7}\ \text{g}/\text{cm}^{3}\ \mathrm{H}_{2}\mathrm{O}\).
Magnification 75,000 times

Fig. 20. Shellac particles, magnified 100,000 times

destruction of \(\beta\)-polyoxymethylene crystals. These studies made it possible to look more deeply into the morphological structure of crystalline high-polymer substances and to draw a number of valuable conclusions about the extremely complex processes of their growth. From this one may assume that the growth of crystals is characterized by the fact that the latter are not built from ready-made molecules, but that at the same time the packing of growing molecules into bundles with a diameter of \(5\)—\(10\ \mathrm{m}\mu\) occurs, and into aggregates of such bundles—fibers with a diameter of \(0.2\)—\(0.6\,\mu\) \(^{48}\).

Fig. 19 shows a molecule of hemocyanin—one of the largest homogeneous molecules among protein substances known in general. The electron image of this molecule at a magnification

... by 75,000 times shows that it is approximately spherical in shape and has a diameter of \(2 \cdot 10^{-2}\ \mu\) [49].

Figure 20 shows a photograph of shellac particles, forming, as it were, a solid skeleton with spherical molecules of auxiliary building materials contained in it. These round formations have a diameter of \(10^{-2}\ \mu\).

Fig. 21. Tomato virus. Individual rods and chains

Fig. 22. “Proteus” bacterium with its sheath in a resting state. Magnification 73,000 times

A molecular picture of the tomato virus, isolated for the first time, was obtained. The electron image made it possible to see individual molecules—threads, a characteristic property of which is the formation of accumulations in the form of long chains (Fig. 21). The dimensions of the individual viral rods agree well with the dimensions obtained by other indirect methods. This specimen was prepared by drying on a collodion film \(10^{-5}\ mm\) thick a single drop containing the virus substance at a concentration of \(10^{-5}\ g/cm^3\).

Fig. 23. “Proteus” bacteria at the moment of budding

The bacterium of the dental cavity, known in ordinary microscopy for its peculiar circular motion, was investigated. The very lively motion that this bacterium exhibits in life must evidently be attributed to those very fine “cilia” that this bacterial body possesses and that are clearly visible in the electron microphotograph.

An electron image (Fig. 22) of the bacterium “Proteus” (also from a dental cavity), obtained at a magnification of 73,000 times, makes it possible to

Fig. 24. Chromosomes of salivary glands. The molecules consist of a small number of atoms. Magnification 200,000 times

Fig. 25. Crystals of magnesium oxide. The thickness of the crystalline thread at the top of the crystal, indicated by the arrow, is 20 Å. Magnification 135,000 times

see the envelope around the body of the bacterium. The next photograph (Fig. 23) is remarkable also in that one can directly observe the process

Fig. 26. Highly dispersed colloidal gold

Fig. 27. A small part of the image shown in Fig. 26. Magnification 360,000 times

of budding inside both such bacteria and the peculiar change of structures connected with this.

In Fig. 24 are shown particles consisting of only a few atoms and having dimensions of 10–15 Å. The photograph was made from a crushed preparation of chromosomes of Drosophila melanogaster (magnification 200,000 times).

The needle-like crystal shown together with chromosomes belongs to copper acetate or zinc acetate, used in preparing the specimen. These molecules have a diameter 10 times smaller than the diameter of hemocyanin molecules.

Studies carried out with crystals have shown a whole series of phenomena hitherto unknown and of great interest. Thus, for example, crystals obtained from magnesium oxide smoke, magnified 135,000 times (Fig. 25), have “protrusions” at their edges, representing the finest crystalline threads 20 Å thick. These threads are molecular chains with an extraordinarily high resistance to rupture.

It has also been possible to observe colloidal particles of highly dispersed sols. Thus, for example, Figs. 26 and 27 show particles of a gold sol with dimensions of 40 Å, magnified 360,000 times.

Fig. 28. Soot particles. Magnification 500,000 times

In Fig. 28 are shown soot particles obtained from the flame of cedar wood. With the smallest possible apertures it was possible to achieve magnification of these particles by 500,000 times.

The resolving power of the universal supermicroscope is 100 times greater than that of a modern optical microscope. Where the best optical microscope can show the observer only one element of an object, the supermicroscope will show, instead of one, 10,000 elements.

Further successes in the development of the electron microscope promise to bring science many more interesting data. It may be hoped that in the near future the most remote corners of the diverse picture of the supermicroscopic world will be revealed to us.

LITERATURE

1. On the first works on the electron microscope, see Uspekhi fizich. nauk, articles: N. N. Malov, 13, 367, 1933; G. Busha, 17, 470, 1937; E. Briukhe, 17, 477, 1937; O. Shertser, 17, 485, 1937; V. Shaffernikhta, 17, 491, 1937; N. D. Morgulisa, 17, 501, 1937; D. V. Zernova, 21, 162, 1939.
   1a. H. Busch, Ann. Physik, 81, 974, 1926.
2. E. Ruska u. M. Knoll, Ann. Physik, (5) 12, 607, 641, 1932.
3. Davisson and Calbick, Phys. Rev., 38, 585, 1931.
4. Brüche u. Johannson, Naturwiss., 20, 353, 1932.
5. Johannson u. Scherzer, Z. Physik, 80, 183, 1933.
6. E. Ruska u. M. Knoll, Z. techn. Physik, 12, 389, 1931.
7. E. Ruska u. M. Knoll, Z. Physik, 78, 318, 1932.

1. E. Ruska, Z. Physik, 89, 90, 1934.

2. E. Ruska, Z. Physik, 87, 580, 1934.

3. Marton, Rev. optique, 14, 129, 1935 et Bull. Acad. Roy. Belg., 20, 92, 439, 1934.

4. Martin, Whelpton u. Parnum, J. Scient. Instrum., 14, 14, 1937.

5. Marton, Nature, 133, 911, 1934 and Phys. Rev., 46, 527, 1934.

6. B. Borries u. E. Ruska, Wiss. Veröffentlich. Siem. Werk., 17, 99, 1937.

7. W. Eitel, H. Müller u. O. Radczewski, Ber. deut. Keram. Ges., 20, 165, 1939.

8. O. Radczewski, H. Müller u. W. Eitel, Zentralbl. Min., Petrogr., Pal., (A), No. 1, 8, 1940.

9. Friess u. H. Müller, Die Gasmaske, 11, 1, 1939.

10. E. Ruska, Forschung. u. Fortschr., 15, 371, 1939.

11. Wolpers u. E. Ruska, Klin. Wochenschr., 18, 111, 1077, 1939.

12. B. Borries, E. Ruska u. H. Ruska, Wiss. Veröffentlich. Siem. Werk., 17, 107, 1937.

13. Frank u. E. Ruska, Naturwiss., 27, 229, 1939.

14. Kausche, Pfankuch u. Ruska, Naturwiss., 27, 292, 1939.

15. D. Beischer, Z. Elektrochem. u. ang. physik. Chemie, 44, 375, 1938.

16. B. Borries u. E. Ruska, Naturwiss., 27, 577, 1939.

17. E. Ruska, Naturwiss., 27, 287, 1939.

18. M. v. Ardenne, Z. techn. Physik, 20, 239, 1939.

19. H. Müller, Elektrotechn. Z., 59, 1189, 1938.

20. H. Mahl, Naturwiss., 27, 417, 1939.

21. H. Mahl, Z. techn. Physik, 20, 316, 1939.

22. H. Mahl, Kolloid. Z., 91, 105, 1940.

23. A. Prebus and J. Hillier, Canad. J. Res., (A) 17, 49, 1939.

24. Burton, J. Hillier and A. Prebus, Phys. Rev., 56, 1171, 1939.

25. B. Borries u. E. Ruska, Z. techn. Physik, 20, 225, 1939.

26. M. v. Ardenne, Z. Physik, 108, 338, 1938.

27. M. v. Ardenne, Z. Physik, 111, 152, 1938.

28. M. v. Ardenne, Z. Physik, 113, 257, 1939.

29. J. Hillier, Canad. J. Res., (A) 17, 64, 1939.

30. W. Bothe, Handb. d. Physik, 25, 26, J. Spring., Berlin, 1927.

31. B. Borries u. E. Ruska, Naturwiss., 27, 281, 1939.

32. B. Borries u. E. Ruska, Z. techn. Physik, 19, 402, 1938.

33. B. Borries u. E. Ruska, Z. techn. Physik, 20, 225, 1939.

34. R. Rebsch, Ann. Physik, 31, 551, 1938.

35. Scherzer, Z. Physik, 114, 427, 1939.

36. B. Borries u. E. Ruska, Z. wissensch. Mikroskop., 56, 317, 1939.

37. M. v. Ardenne, Z. Physik, 115, 339, 1940.

38. M. v. Ardenne, Naturwiss., 28, 248, 1940.

39. M. v. Ardenne u. Beischer, Angew. Chemie, 53, 103, 1940.

40. H. u. M. Staudinger u. E. Sauter, Z. physik. Chem., (B) 37, 403, 1937.

41. M. v. Ardenne u. Beischer, Z. physik. Chem., (B) 45, 465, 1940.

42. M. v. Ardenne, Naturwiss., 28, 113, 1940.