ABSTRACTS

OBTAINING ELECTRON DIFFRACTION PATTERNS USING SECONDARY ELECTRONS

In his report at the recent congress of German physicists and mathematicians, Knoll*—known for his work on cathode oscillography and, in particular, for the development of the electron microscope with magnetic lenses**—reported a modification of the technique of electron diffraction photography, consisting in the use, for obtaining electron diffraction patterns, of secondary electron emission produced by the object placed in a Braun tube and irradiated by an electron beam.

Fig. 1. Diagram and external view of the tube used in the investigations

Knoll used a tube whose diagram and appearance are shown in Fig. 1. The ordinary tube, containing, in addition to the cathode, an anode and an object irradiated by a beam, a secondary anode—which was immobile during the investigations—was provided in the form of a layer of carbon deposited on the inner walls of the tube.

Owing to the bombardment of the object by a beam of primary electrons and to the emission of secondary electrons from the surface of the object, the object ac-

* M. Knoll, Phys. Z., 36, 861, 1935.
** See Uspekhi fizicheskikh nauk.

acquires a certain potential relative to the anode, depending on the geometrical form of the system of electrodes, which determines the distribution of the field between the object and the second anode, on the electrical conditions of the experiment and, as will be shown below, on the material of the irradiated object. All experiments were carried out with an object having the shape of a flat disk. First of all, the dependence of the limiting voltage \(U_g\) (relative to the anode), obtained on the object, on the velocity of the primary electrons was investigated. In Fig. 2 at the top this dependence is shown for three substances—nickel, molybdenum, and carbon; along the abscissa axis the anode voltage \(U_a\) is plotted. The characteristic for carbon is a straight line going at an angle of \(45^\circ\) to the abscissa axis (with equal scales along the axes); the characteristics for Mo and Ni at very small anode voltages coincide with the characteristic for C, then drop sharply, almost coinciding with the abscissa axis (\(U_g\) assumes a value of the order of 1 V), and then begin to rise again only at \(U_a\) close to 2 kV. These curves are obtained at a large resistance between the object and the second anode, considerably exceeding the internal resistance. If, however, the internal resistance is close to the external one or exceeds it, then, owing to leakage, a voltage \(U_p\) is obtained, of course smaller than \(U_g\), but proportional to it.

The analogous curves characterizing the voltage on the object, reckoned relative to the cathode, are shown at the bottom of Fig. 2. Simultaneous measurement of the secondary electron emission (with an auxiliary negative voltage of 100 V applied to the object relative to the anode) gave the results shown in the middle graph of Fig. 2, where along the abscissa axis is plotted the ratio of the number of primary electrons \(i_s\) to the number of primary electrons \(i_{pr}\). These curves show that a negative potential relative to the anode is acquired by the object only in the case when the yield of secondary electrons is characterized by a coefficient smaller than unity. Having established significant differences in the behavior of molybdenum and nickel, on the one hand, and carbon, on the other, Knoll, without delving into a detailed investigation of the physical side of the phenomenon, attempted to apply it for practical purposes.

Fig. 2. Limiting voltage and secondary emission of homogeneous plates of nickel, molybdenum, and carbon

On a disk of pure nickel a strip of carbon was deposited (Fig. 3); the primary electron beam was deflected in the direction indicated in the figure by a magnetic field (frequency 50 hertz), and, with the aid of a cathode oscilloscope, the curve of the voltage \(U_p\) between the anode and

object at different values of \(U_a\). In the oscillograms shown at the bottom of Fig. 3 it is seen that when the electron beam strikes nickel and \(U_a < 1500\ \mathrm{V}\), no voltage \(U_p\) arises; when it passes to carbon it is immediately negative that is charged. At \(U_a > 1500\ \mathrm{V}\) the nickel also becomes charged, but the potential of the carbon remains higher. Owing to the considerable magnitude of the external resistance, the stepwise increase of the potential when the beam passes from nickel to carbon is replaced by an exponential increase, which is quite understandable, since our system possesses a certain capacitance and has a known (rather large)

Fig. 3. Voltage on a nickel disk with a strip of carbon during periodic motion of an electron beam

resistance. Since the curves shown in Fig. 3 indicate the possibility of judging the nonuniformity of the surface of an object from the course of the secondary emission curve, it would be natural to try to apply this method to obtaining an electron image of an object producing secondary emission.

However, obtaining undistorted images will be possible only under the condition of a more rapid change of the voltage when the beam passes across a boundary of nonuniformity. Since reduction of the capacitance of the system has its limits, it is necessary to reduce the external resistance considerably. But then we shall be working far from the limiting voltages, and the deflection of the cathode beam of the oscillograph will be determined not by the values of \(U_p\), but by the difference of the secondary-emission currents of the various substances present on the surface of the object (Fig. 4). In addition, the resulting voltages \(U_p\) will be small, so that one must resort to amplifying them. The oscillograms shown in Fig. 5 testify to an acceleration of the voltage rise, which makes it possible to hope that images with fairly sharply outlined edges can be obtained. The displacement of the oscillograms from the zero line is explained by the superposition of a constant component of the voltage in the amplifier, which has no fundamental significance. To obtain the images, Knoll used the circuit shown in Fig. 5.

The voltage \(U_p\) between the anode and a certain point of the object, at which the cathode beam is striking at the given moment, is fed through an amplifier to the control grid of a Braun tube, regulating the intensity of the current in it. Both the primary cathode beam and

and the beam in the Braun tube, controlled by the voltage \(U_p\), move synchronously (the first over the object, the second over the screen of the tube) under the action of crossed magnetic fields with frequencies of 50 and \(10^4\) Hz; in its displacements the first beam scans the entire

Fig. 4. Current pulses during periodic motion of an electron beam over a nickel plate with a strip of carbon

surface of the object under investigation, while the second gives its image on the screen of the Braun tube (this process to a certain extent resembles the process of obtaining images in the Zworykin television system).

The details of the circuit are explained by the labels in Fig. 5. It should be noted that in the experiments a battery with a potentiometer was used, making it possible to create between the object and the anode an arbitrary potential difference of either sign; in what follows this difference will be denoted by \(U_p\); on this potential difference were superposed potential oscillations caused by the electron emission of the individual parts of the object.

Fig. 5. Circuit for obtaining electronograms with the aid of secondary emission

The object was a nickel plate with letters applied to its surface in carbon (Fig. 6, upper left); the diameter of the plate

5 cm. At \(U_p=-15\ \mathrm{V}\) a very bright and contrasty image of the object was obtained on the screen of the Brown tube (Fig. 6, upper right). If \(U_p\) is increased, the contrast deteriorates sharply, and at approximately \(U_p=-\pm 15\ \mathrm{V}\) the image disappears. But a further increase of the voltage \(U_p\) again gives an image, whose intensity increases together with \(U_p\), but the contrast proves to be less good (Fig. 6, bottom). Moreover, these images differ from those obtained at negative voltages by a distribution of brightness, namely: the light regions of some correspond to the dark regions of the others, and vice versa.

Knoll tries to explain this reversal of the images by the fact that, at negative voltages \(U_p\), the image is determined by the distribution of secondary electrons, whereas at positive voltages it is determined by the distribution of primary reflected electrons possessing energy sufficient for traversing back through the field created between the object and the anode. Knoll, basing himself on a number of observations, comes to the conclusion that bodies with high secondary emission reflect electrons poorly (and conversely). However, he refrains from making final judgments on this question until further experimental data have been accumulated.

Fig. 6. Electronograms of a nickel plate with letters made of carbon

The method described makes it possible to obtain still finer images. Thus, Fig. 7 shows a photograph, an electronogram, and an imprint of an ordinary key made of electrolytic copper and rubbed with carbon.

Fig. 8 presents a photograph and electronograms obtained when iron containing silicon was used as the object. The lower right electronogram was taken at considerable magnification. These electronograms are of interest in that half-tones are clearly visible in them, caused by the different emissive capacity of the individual crystals.

Fig. 7. Electronogram of an object of fine structure

Further, the author succeeded in establishing the influence on the resulting image of spatial charge arising at high current densities in the tube, as well as of the angle at which the primary electrons strike an object having an uneven surface.

In conclusion, the author presents very interesting preliminary results of a study of the secondary radiation of various internal parts of the tube. In these experiments a flat object was replaced by a ring of nickel wire, so that the primary-electron beam could strike the internal parts of the tube. It turned out that appreciable secondary emission is observed not only from conductors located inside the tube, but also from insulators. The electron diffraction patterns clearly showed that, in studying the distribution of current in a tube or lamp, it is necessary to take into account not only the primary electron beam, but also the secondary emission of electrons from various regions bombarded by the primary electrons.

As for the resolving power of this new method of obtaining electron diffraction patterns, it is limited primarily by the dimensions of the primary electron beam, which can hardly have a diameter smaller than 0.1 mm (the usual beam diameter is close to 1 mm). However, in comparison with the ordinary electron microscope, this method also has a number of advantages, such as, for example, the absence of a strong accelerating field capable of distorting the initial distribution of electrons, the absence of chromatic absorption, etc.

Fig. 8. Secondary emission of iron crystals with a silicon impurity

The possibility of using this method to study secondary emission and charge distributions on insulators suggests that it will become widespread on a par with the ordinary electron microscope.

N. Malov, Moscow