ELECTRON OPTICS

IV. ELECTRON-OPTICAL IMAGE CONVERTER

B. Schaffernicht, Berlin

Contents: 1. Principle of construction and mode of operation of the image converter. 2. Methods for obtaining an electron-optical image. 3. Various converters and their properties. 4. Applications of the converter. 5. Conclusions.

1. Principle of Construction and Mode of Operation of the Image Converter

The electron-optical image converter serves to transform light rays into electron rays. During the last ten years[^1] there have been many attempts to construct such an instrument (see the patents on television). The solution of this problem became possible only after the development of the basic concepts of electron optics, as well as the theory of photoelectric layers. The transformation of light rays into electron rays has a twofold significance: 1. By means of the transformation one can obtain an increase in the brightness of the image by increasing the energy of the flux of photoelectrons; the latter is possible either through an increase in the velocity of the electrons or by increasing their number by means of secondary emission at a very fine mesh or a very thin film placed in the path of the electron rays. 2. The application of electron rays instead of light rays makes it possible to change the direction of the ray by means of transverse electric or magnetic fields arranged in the path of the electron ray. The transformation of light rays into electron rays is carried out by means of a photoelectric layer in the following way (Fig. 1). When

Fig. 1. Diagram of an electron-optical image converter

by means of a lens \(L\), the image \(B\) of the object \(G\) is projected onto the semitransparent photoelectric cathode \(K\). From various parts of the cathode \(K\), photoelectrons are torn out in a quantity corresponding to the illumination of the given area; these electrons are accelerated by the field produced by the anode cylinders \(A_1\) and \(A_2\), shown in Fig. 1. The voltages are chosen so that the electric field formed between \(A_1\), \(A_2\), and \(K\) acts as an electric lens and gives an image of the object in \(EB\). This image may also be obtained with the aid of a magnetic lens placed in the path of the rays, or as a result of the combined action of electric and magnetic lenses. With sufficient uniformity of the cathode, an electronic image of the object \(G\) under consideration is obtained in \(EB\). This electronic image can be used in two ways. In one case, a fluorescent screen is placed in \(EB\), excited by the electrons falling upon it. In this way an electronic image visible to the eye is obtained. In the other case, in the middle of \(EB\) there is a small aperture, behind which a collector is placed; moreover, by means of two crossed electric or magnetic transverse fields, the electronic image

Fig. 2. Course of the spectral sensitivity of complex photocathodes
(after Kluge)

moves in a known manner relative to the collector and is consequently converted into current pulses. The first device is called an image converter; the second is a part of the television system proposed by Farnsworth (dissector). The purpose of developing image converters is to obtain visible images when the cathode is illuminated by radiation of various spectral composition, mainly invisible to the human eye—infrared and ultraviolet. The spectral region is determined by the distribution, over the spectrum, of the sensitivity of the photoelectric layer used.

In Fig. 2 are presented curves of the spectral sensitivity of photoelectric layers currently in use (according to

Kluge ^2). In addition, the curve of the relative spectral sensitivity of the eye is also plotted here. It is evident from the figure that the most favorable spectral-sensitivity curve is possessed by the Ag—Cs₂O—Cs layer, which is distinguished by great sensitivity in the infrared part of the spectrum. The requirement of high sensitivity in the ultraviolet part of the spectrum is equally well satisfied by all photoelectric layers used in practice, since all of them possess maximum sensitivity in this region.

Fig. 3. Position of the long-wave maximum and of the photoeffect threshold for the Ag—Cs₂O—Cs layer
(according to Kluge)

The following Fig. 3 characterizes (according to Kluge ^3) the position of the photoeffect threshold and of the long-wave maximum for Ag—Cs₂O—Cs layers. In the best case the photoeffect threshold reaches 1.4 μ. The fluorescent layer of the screen is chosen so that the greatest radiation intensity falls in the region of the optimum sensitivity of the eye. The image obtained on the screen is monochromatic, like any image obtained in television. The colors of the transmitted light image (as in photography) are conveyed by means of different brightness values of the luminous screen, and the nature of this rendering depends on the spectral sensitivity of the photoelectric layer of the cathode.

2. Methods of obtaining an electron-optical image

After Brüche ^4 had proved the possibility of obtaining an electron-optical image of a photoelectric cathode, the development of the idea of the electron-optical converter began. The paths that led to the solution of the converter problem differ mainly in the electron-optical methods of obtaining an image—

Before proceeding to consider these paths, it is necessary to add the following. In the case of an ideal electron-optical image, all electrons emerging from one point of the photocathode must arrive at one point of the fluorescent screen. This requirement cannot be strictly fulfilled because the electrons emerging from the photocathode have certain initial velocities, distributed according to a definite law. Owing to the presence of these velocities, the electrons fly out from one point of the cathode in all possible directions; as a result of this, one point on the photocathode is represented on the screen as a certain blurred circle. This chromatic aberration manifests itself differently in different converters, which we shall divide into three fundamentally distinct groups (Fig. 4):

Fig. 4. Methods of obtaining an electron-optical image in converters

a) In Fig. 4, a is shown the scheme proposed by Holst, de Boer, Teves, and Venema[^5]. \(K\) is a semitransparent cathode, \(S\) the screen, which is at the same time a semitransparent anode. The cathode and anode are placed so close to each other that, when a voltage is applied between them, a homogeneous electric field is formed. In this case the electrons move along parabolas from \(K\) to \(S\). The voltage is chosen so high that the motion of the electrons parallel to the plates plays no role. In this way an image arises on the screen \(S\) corresponding to the light image projected onto the cathode \(K\). To a certain extent this method corresponds to contact printing of photographic pictures in photography.

b) The electron-optical image obtained by the method shown in Fig. 4, b, is the basis of Farnsworth’s television method[^6]. Between the cathode \(K\) and the anode \(S\), which in this case consists of two plane parallel plates, two transverse magnetic fields are formed, serving to deflect the electron beams. The nonuniformity of the electric field, caused by the large distance between \(K\) and \(S\), is eliminated by the fact that \(K\) and \(S\) are connected by a high-ohmic hollow metallic cylinder \(Z\), which surrounds both plates. Along the length of the cylinder \(Z\) there exists a linear drop of potential. Owing to the large distance between the anode and the cathode, the divergence of the bundle of electron rays practically does not allow an image to be obtained without additional focusing of the rays by means of a cylindrical magnetic coil. After these improvements it was possible to obtain on the screen \(S\) an exact image of the cathode \(K\). The image obtained in this case is erect.

c) The third method of electron-optical transformation, pre-

shown in Fig. 4, c, is the most perfect. Here special electron lenses are used, which make it possible to obtain a very sharp image. This method was first indicated by Polle,^7 who transformed the image of a grid, projected in ultraviolet light onto platinum foil, into the corresponding electron image. Polle’s proposal led to a whole series of image converters. The scheme of this converter is as follows: directly at the cathode there is an accelerating electric field; then follows an electric or magnetic lens, or else a combination of such lenses \((L)\). Such a system is analogous to an optical system, the axis of which is shown in Fig. 4, c by rays. The magnification of the system depends on the relative arrangement of \(K\), \(L\), and \(S\). Just as in light optics, when an electric lens is used an inverted image is obtained. When a magnetic lens is used, an additional rotation of the image arises; this can be eliminated,^8 by including in the circuit two magnetic coils placed next to one another and carrying currents in opposite directions.

Henneberg and Recknagel^9 compared these three methods theoretically from the point of view of their chromatic properties. They showed that the last two systems have an advantage over the first owing to exact focusing. If \(\varepsilon\) denotes the work function of the electron from the photocathode, \(V\) the accelerating potential, \(L\) the distance between the cathode and the screen, and \(v\) the magnification, then the diameter of the circle into which a point is projected on the screen in these three cases proves to be equal to:

\[ \text{a) } \Delta_1 = 4L_1 \sqrt{\frac{\varepsilon}{U}};\quad \text{b) } \Delta_2 = 2L_2 \frac{\varepsilon}{U};\quad \text{c) } \Delta_3 = 2L_3 \frac{v}{2V} + 1 \cdot \frac{\varepsilon}{U}. \]

Under assumptions corresponding to a realistically possible case, calculation of these quantities gives the following values: \(\Delta_1 \approx 2 \cdot 10^{-2}\ \text{cm}\), \(\Delta_2 = 5 \cdot 10^{-3}\ \text{cm}\), \(\Delta_3 = 1.7 \cdot 10^{-3}\ \text{cm}\). Hence it is clear that the best image is obtained in case c. The resolving power in this case has a value of the order of \(0.01\ \text{mm}\). The latter is confirmed experimentally.

3. Various converters and their properties

Let us consider several converter systems. The Holst converter is a flat-bottomed glass Dewar vessel shown in Fig. 5. The bottom of the inner tube is coated with a layer of \(\mathrm{Ag—Cs_2O—Cs}\); the bottom of the outer tube is coated with a fluorescent composition and a semitransparent layer of silver, and serves simultaneously as an anode and a screen. The observed object is projected by means of a lens onto the photocathode, and the electron image is observed from the reverse side of the screen. A voltage of \(4000\ \mathrm{V}\) is applied between the cathode and the anode. The Farnsworth television tube (Fig. 6) consists of a transparent cathode \(K\),

of the accelerating anode \(A\), the focusing magnetic coil \(H\), and the high-ohmic metal cylinder \(Z\), which ensures uniformity of the electric field. Line-by-line movement of the electron image relative to the aperture, behind which the electron multiplier \(M\) is located, is effected by means of two mutually perpendicular magnetic deflecting coils. The multiplier \(M\) serves to amplify the pulses of electron current.

![Figure 5 and Figure 6 diagrams]

Fig. 5. Converter of Holst, de Boer, Teves, and Venemans

Fig. 6. Farnsworth television tube

In Figs. 7 and 8 various converter systems are compared. In Fig. 7a the first Schaffernicht converter is shown.\(^{10}\) Its main parts are: the cathode \(K\), the anode \(A\), and the screen \(S\), enclosed in a high-vacuum tube. The light image is projected onto the cathode and is converted into an electron image on the fluorescent screen by means of a magnetic and an electric lens, whose field is shown in Fig. 7, \(a\). Owing to the fact that the cathode is curved, it is possible to obtain over the entire region of the cathode an image free from distortions. The potential field of the lens has such a form that immediately in front of the cathode it acts as a converging lens, followed by a diverging lens. Subsequently it proved possible to eliminate the magnetic lens.\(^{11}\)

The optical power of the system of electric lenses is chosen so as to obtain a sharp image of the cathode on the screen. In Fig. 7, \(b\) an optical analogy is given. The optical power of the positive lens increases with the curvature of the cathode; the optical power of the negative lens, for a given diameter of the anode cylinder, is the smaller the farther the cylinder is located from the cathode. The total optical power of the system, as in optics, is determined by the mutual arrangement of the negative and positive lenses relative to the cathode. With a successive arrangement of the positive and negative lenses, a strong contraction of the electron rays arises, while the paths of the electrons within an individual beam prove to be almost parallel.

It follows from this that the image remains equally sharp at different positions of the screen. In this way, in such an electron-optical system one can vary the magnification in the range from 1:1 to 1:3 without noticeable changes in the sharpness of the image. By changing the position of the screen, one can, together with the change in magnification, change the surface brightness of the image. The latter is a consequence of the fact that the total energy flux of the electronic radiation, when the position of the screen is changed, remains unchanged. The cathode region,

![diagram]

Fig. 7. Schaffernicht converter

![diagram]

Fig. 8. Converters of Heymann and Kluge, Ardenne, Zworykin and Morton

the image of which can be obtained by means of this system, lies inside the solid angle whose vertex is located at the midpoint of the segment connecting the two lenses, and whose magnitude is approximately \(10^\circ\). Such a two-electrode system, in addition to its great simplicity, has another important advantage: the image obtained does not depend on the magnitude of the accelerating voltage between the two electrodes. In a system consisting of several electrodes, the image is preserved only under proportional variation of the potentials of all electrodes, which is very difficult when high voltages are used. As a result of the presence of leakage in the tube, the potentials applied to the electrodes change, as a result of which blurring of the electron-optical image appears.

In the device of Heymann \(^{12}\) and Kluge \(^{13}\) (Fig. 8), a flat cathode and two anodic cylinders are used, having different potentials and forming an electric lens which, together with an auxiliary magnetic lens, makes it possible to obtain a sharp image of the cathode on the luminous screen. The later device of Ardenne \(^{14}\) consists, in the main, of the same parts as the aforementioned

higher converters. In this device two anode cylinders and a magnetic coil are used. Zworykin and Morton\(^{15}\) use a purely electric system consisting of a curved cathode, ring electrodes, one diaphragm, and an anode cylinder. When a potential \(V_0\) is applied to the cathode, voltages \(V_2\), \(V_3\), and \(V_1\), chosen so that an electric converging lens is formed at the location of the diaphragm, are applied to the diaphragm, the anode cylinder, and the last ring electrode. The ring electrodes serve to improve the properties of the electric lens. By means of a voltage divider these electrodes are given potentials that increase gradually with respect to the cathode. The presence of the voltage divider somewhat complicates this system.

4. Applications of the Image Converter

At the present time, in almost all works on electron-optical image converters, in order to characterize the quality of the converters, electron images of transparencies or grids projected onto the cathode are given.

Holst and co-workers were the first to indicate a method of electron-optical conversion of images of transparencies projected onto a photocathode, based on the use of a uniform electrostatic field. After electron-optical conversion, a picture is obtained by simply photographing the luminous screen (Fig. 9). The electron images obtained on the luminous screen still possess appreciable blurring, which in principle is unavoidable with the above-mentioned method of obtaining images. According to the calculations of Henneberg and Recknagel, a point on the cathode is projected onto the luminous screen in the form of a circle of diameter \(0.1\) mm, which agrees with the experimental data. In Figs. 10–12 photographs are shown that were obtained by the above-mentioned method and make it possible to judge the quality of the images obtained in electron-optical converters with electron lenses. The distortions that occur in converters with a plane cathode can be seen in the photograph of Fig. 13, obtained by Zworykin and Morton. An improvement in image quality can be obtained by using a curved cathode and ring electrodes (Fig. 14). At the present time it may be said that the problems of electron-optical conversion have essentially been solved. The field of application of these devices is determined by their sensitivity in the infrared and ultraviolet parts of the spectrum. The sensitivity of the converter depends on four factors:

1. On the sensitivity of the photoelectric layer.
2. On the sensitivity of the luminous screen.
3. On the magnitude of the accelerating voltage.
4. On the magnification provided by the electron-optical system of the converter.

Usually a photocathode consisting of \(Ag—Cs_2O—Cs\) is used, which has a photosensitivity of about \(20\ \mu A/Lm\) (this sens-

Toward the article by V. Schaffernicht

Fig. 9. Electron image obtained by the system of Holst, de Boer, Teves, and Venemans

Fig. 10. Comparison of the image obtained according to Schaffernicht’s scheme (Fig. 7a) and the corresponding photograph

Fig. 11. Electron image obtained in Heymann’s apparatus

Fig. 12. Electron image obtained in Schaffernicht’s apparatus (Fig. 7, b)

Fig. 13. Distortions of the image of a grid when an electric electron lens is used (according to Zworykin and Morton)

Fig. 14. Improvement of an electron image when a curved cathode is used (according to Zworykin and Morton)

To the article by V. Schaffernicht

Fig. 16. Visible and infrared spectra of mercury and sodium lamps

Fig. 17. Curve of the spectral sensitivity of the photocell, or the visible and infrared spectra of the carbon arc

Fig. 18. Absorption spectra obtained with the aid of the converter

Light image

Electronic image

Fig. 19. Comparison of the optical image with the electronic one, obtained with the aid of the converter when observing in the infrared region of the spectrum

sensitivity is determined for a spectral composition of radiation corresponding to the radiation of an incandescent body at a temperature of 2600°K). The material of the luminous screen is selected so that the screen has maximum brightness. The brightness of the image on the luminous screen increases linearly (Fig. 15), together with an increase in the accelerating voltage (with the exception of the low-voltage region), up to the region of 80 kV (according to Schabel’s data ^16).

The sensitivity of the converter also depends on the magnification of the image achieved by electron-optical methods, because the brightness of the image increases as the magnification decreases. At very high brightness of the image obtained on the luminous screen, there arises the danger of optical feedback; this phenomenon consists in the fact that the light radiation of the luminous screen falls on the cathode and is superimposed on the primary image. The reverse optical action is most noticeable in Holst’s device, in which the distance between the cathode and the luminous screen is very small.

Fig. 15. Dependence of the brightness of the electron image on the magnitude of the accelerating voltage

In those instruments in which magnetic or electric lenses are used, this feedback from the luminous screen can easily be eliminated by deflecting the electron beam through such an angle that the radiation of the screen would not fall on the cathode.

The photosensitive cathodes used in electron-optical image converters have a very small work function, as a result of which, even at room temperature, a noticeable thermionic current from the cathode is observed, producing a light background on the luminous screen. The interference produced by this background is the greater, the lower the brightness of the resulting electron image. To illustrate the operation of electron-optical converters, let us consider several photographs of electron images obtained at the AEG research institute ^17. Figs. 16–18 present the results of various spectral investigations carried out mainly in the infrared part of the spectrum. Fig. 16 shows spectra of mercury and sodium lamps obtained with the aid of a converter. Fig. 17 gives the spectrum of a carbon arc photographed with various infrared light filters.

Fig. 18 shows the absorption spectra of various organic and inorganic substances in the infrared part of the spectrum (from 0.8 to 1.2 μ). With the aid of the electron-optical converter, investigations can be carried out on various objects illuminated by ultraviolet or infrared radiation. This proves especially convenient in those cases in which one is studying under a microscope various insects, the smallest living organisms, or processes occurring beneath the pigmented layer of the skin. By illuminating the object under investigation with infrared rays, we can observe biological processes occurring in living organisms. In Fig. 19, as an example of such an observation, a microscopic photograph of a wasp’s leg in infrared and visible light is given. In addition to the above-mentioned examples of applications of the electron-optical converter, the latter can also be used for a number of other purposes; thus, for example, all objects of infrared photography can likewise be observed with the aid of the converter, the regions of spectral sensitivity of the two methods of investigation being approximately the same. The red limit lies on average at about 1.2 μ. The principal advantage of the photographic method is its great resolving power (in comparison with the converter); the principal advantage of the converter is that the processes under investigation can be observed directly in their motion.

References

1. See E. Brüche u. W. Schaffernicht, Elektr. Nachr. Techn., 12, 381, 1935.
2. W. Kluge, Elektrotechn. Z., 57, 145, 1936.
3. W. Kluge, Z. techn. Phys., 16, 181, 1935; see also Uspekhi fizich. nauk, 15, 1025, 1935.
4. E. Brüche, Z. Physik, 85, 448, 1933.
5. G. Holst, J. H. Boer, M. C. Teves a. C. F. Veenemans, Physica, 1, 297, 1934.
6. F. T. Farnsworth, J. Frankl. Inst., 218, 411, 1934.
7. J. Pohl, Z. techn. Phys., 15, 579, 1934.
8. G. Stabenow, Z. Physik, 96, 634, 1935.
9. W. Henneberg u. A. Recknagel, Z. techn. Phys., 16, 230, 1935.
10. W. Schaffernicht, Z. Physik, 93, 762, 1935.
11. W. Schaffernicht, Jahrb. d. Forschungs.-Inst. d. AEG, 4, 45, 1933/35.
12. W. Heimann, Elektr. Nachr. Techn., 12, 68, 1925.
13. W. Kluge, Z. Physik, 93, 789, 1935.
14. M. v. Ardenne, Elektr. Nachr.-Techn., 13, 230, 1936.
15. V. K. Zworykin a. G. A. Morton, J. Opt. Soc. Am., 26, 181, 1936; V. K. Zworykin, Z. techn. Phys., 17, 170, 1936; see also Uspekhi fizich. nauk, 16, 814, 1936.
16. W. Schnabel, Arch. Elektrotechn., 28, 789, 1934.
17. W. Schaffernicht u. H. Katz, Z. Physik (in press).