Photoelectric Cells in Science and Technology*

B. Lange, Berlin

Part I

The purpose of the present work is to give a brief survey of the modern technical development of the photoelectric cell. Such a survey will make it easier for every natural scientist to apply the photoelectric effect in his own special field of work. We shall have to refrain from setting forth the development of the theoretical views, although they are very interesting, so that the purely technical side of the matter may stand out more clearly. Although the photoelectric effect was discovered as early as 1888 by Hallwachs¹ (and by Stoletov, ed.) and was used for purposes of photometry, photoelectric cells have only in recent years acquired a quite unexpected technical and scientific significance. It is hardly necessary to point to the extensive applications of photoelectric cells in automatic photometers, in sound films, and in the transmission of images and television. In America the technical use of photoelectric cells has gone still further; for example, photoelectric cells are used for the automatic sorting of cigars by color, as counters at the entrances to tunnels and halls, for controlling machines by printed templates, and in the textile and graphic industries. The use of photoelectric cells for the light control of trains may also acquire importance, since it eliminates the possibility of passing stop signals. Finally, photoelectric cells are used in the Panama Canal as automatic light switches, for switching on and off luminous buoys. To summarize, we see that photoelectric cells

* Naturwiss. 1931. Translated by V. Fabrikant.

also, as incandescent lamps in their time, emerged from the scientific-laboratory stage, and their application already has a highly developed technical character.

TECHNICAL DEVELOPMENT OF PHOTOELEMENTS

A zinc plate illuminated by short-wave light acquires, as is known, a positive charge, appearing as a result of the loss of electrons. Of course, in such a simplest form the photoeffect is not applicable in practice. Only by using an alkali-metal cathode with a wire anode placed opposite it, enclosed in an evacuated glass vessel, can the photoeffect be used for photometric measurements. The often used photoelectric alkali-metal cells of Elster and Geitel² (Fig. 1) consist of a glass sphere \(G\), partly silvered on the inside, with a side branch for the anode lead-in \(A\). On the silver layer \(S\), connected with the cathode lead-in \(K\), an alkali metal (in most cases potassium) is deposited in a more or less thick layer \(M\). Opposite \(M\) is a wire ring \(R\), used as the anode. The ring is connected by a conductor to \(A\); \(E\) is a metal earthed ring, fitted directly on the glass. It is intended to prevent charges from spilling over from the cathode, which has a relatively high negative potential, to the anode \(A\).

Fig. 1. Diagram of a photoelement.

A laboratory-type device is shown in Fig. 2, where the ring-shaped anode is especially clearly visible. The photocurrent of such an element under the action of sunlight barely reaches \(10^{-8}\ \mathrm{A}\). Of course, there can be no question of any—

Fig. 2. Alkali-metal photoelement.

there is any substantial conversion of light into electrical energy. Even for photometric measurements the use of such small photocurrents is difficult; only by applying an accelerating voltage can these currents be amplified many times (Fig. 3). At a comparatively low applied voltage the saturation current is already reached, i.e., practically almost all the photoelectrons torn out reach the anode, and a further increase in voltage no longer has any appreciable effect on the photocurrent. In a circuit consisting of a photoelement

Fig. 3. Dependence of sensitivity on voltage.

Figs. 4–7. Various photoelements.

with an applied voltage and a galvanometer, at an illumination of 1 lux we obtain a photocurrent of about \(5 \times 10^{-10}\) A. Alkali elements are used exclusively with an applied voltage, so that the photocurrent is not, in essence, the result of the conversion of light into electrical energy, but flows at the expense of auxiliary batteries. The photoelement works only as a controlling device, owing to the change of its resistance under the action of light. Such elements with a high vacuum are still used even now in very precise photometric measurements, owing to the great constancy of their properties. The technical construction

layer electrolytically through the glass bulb of an ordinary incandescent lamp.* As can be seen from Fig. 7, in such a photocell (Nava photocell, Tungsram) the spirally wound filament is the anode, which during electrolysis is used as a heated cathode and supplies the electron stream necessary for electrolysis. For electrolysis the NaNO₃ lamp is immersed in a mixture that at the same time serves as the anode, and at a temperature of about \(250^\circ\) sodium passes electrolytically through the glass. The sodium is thereby deposited on the cooled part of the bulb in the form of a shining metallic mirror. In its original form, a photocell made by this method

Figs. 5–7. Various photocells.

differs from an ordinary incandescent lamp only by a third contact, the extension of which, when the lamp stem is sealed in, is introduced into the lower part of the bulb. The lamp is then evacuated, sealed, and, by introducing electrolytic sodium, converted into a photocell, the third electrode being connected with the sodium cathode.

The work of Zeleny⁵ showed that the sodium cathode can be activated by a very interesting method, namely by electrolytically introducing oxygen, increasing the general sensitivity and shifting its maximum toward the red. If, in the finished photocell, using a special

* A similar device for electrolysis through glass with the use of a heated cathode was first described by M. Pirani and Lax.

by carrying out electrolysis of the glass in the reverse direction, a certain amount of oxygen corresponding to Faraday’s law will appear inside the photocell. The oxygen will partially oxidize the sodium, and in this way colloidal solutions of sodium and of its own oxide of the most varied colors will be formed, from yellow to violet. How much the spectral sensitivity of the photocells changes in this process is evident from Fig. 8. Curve 1 corresponds to a pure sodium cathode, whereas curves 2 and 3 correspond to an increasing amount of oxide. It is remarkable that the sensitivity curve of a strongly oxidized (oxygenated) photocell (Pavs) is very close to the sensitivity curve of the eye, so that this photocell is suitable directly for the objective determination of illumination.

Fig. 8. Curves of the spectral sensitivity of Tugrgramava-Na photocells.

In Table 1 the physical data of individual elements are summarized briefly.

In reviewing the technical development of photocells, four stages clearly stand before us. First, the photoeffect on a single metallic plate situated in air; then the transition to vacuum cells and the application of an applied voltage, after which the most important from the technical

point of view, step—a introduction of gas-filled elements and, finally, activation of the surface. Establishing the optimal gas pressure and absorption at the photocathode may further increase the action of the photocell. Work on photocells with a monatomic alkali layer6 broadens our knowledge in the field of the selective photoeffect and will make it possible to determine the quantum equivalent. In our opinion, the path outlined here will not solve the question of any substantial conversion of light into electrical energy. One glance at Table 1 shows that the photocurrent without applied voltage, i.e. with the direct conversion of light into electrical energy, is of the order of \(10^{-7}\) A per lumen.* Recently, despite great efforts, little new has been obtained in the field of alkali photocells. Summing up, one may say that, despite their very modest efficiency, alkali photocells have probably already reached the highest point of their development.**

However, this had accomplished, as it were, only the first step in the conversion of light into electrical energy. The author of the present article has described7 a new photoelectric effect, named, at the suggestion of V. Schottky,*** the photoeffect in the barrier layer (Sperrschicht photoeffekt), which gives a much larger coefficient of photoelectric action. The low efficiency of alkali photocells is caused by the large work of emission for the electrons torn out by light and by the large internal resistance. It was shown that if the escape of electrons directly into a vacuum or into a gas-filled space is avoided, then the coefficient of photoelectric action increases considerably. In the new photocells the electrons are torn out directly in the intermediate layer, consisting of uni-

* With a photocell surface of about \(20\ \mathrm{cm}^2\).

** A conclusion which proved premature, since since the publication of the present article alkali photocells have appeared with a sensitivity hundreds of times greater than that indicated in Table 1. Translator’s note.

*** Cf. the reports of V. Schottky and B. Lange in Königsberg. Phys. Zeitschr., 31, 964–970, 1930.

polar semiconductor Cu₂O. This layer is in contact with two electrodes arranged in the form of a capacitor. Owing to the smaller contact potential difference at the metal–semiconductor boundary than at the metal–vacuum boundary, the efficiency is increased. Fig. 9 gives the scheme of such a device, while Fig. 10 gives the external view of a photoelement with a surface of 3 cm². \(E_1\) is a thin semitransparent electrode on which the light falls; \(u\) is the semiconducting layer of Cu₂O and \(E_2\) is the second copper electrode, directly on which the copper oxide is obtained. Characteristic of these photoelements (cuproxides)* is the presence of a blocking, unipolarly conducting

Fig. 9. Scheme of a cuproxide photoelement.
Fig. 10. “Cuproxide” according to Lange.

layer with high contact resistance, located in the photoelement shown in Fig. 9 between \(u\) and \(E_2\). The photoelectrons liberated in the cuprous oxide near this layer pass through the blocking layer to the copper electrode \(E_2\) and go in the external current circuit from \(E_2\) to \(E_1\). In order that the incident light may reach the boundary layer, it is advisable to make the electrode \(E_1\) and the semiconductor \(u\) as thin and transparent as possible. In addition, with a thin layer the resistance of the electrons’ path is small, so that the external coeffi-

* We use here the term “cuproxide,” since, as far as we know, it is becoming more and more widespread among the physicists of our Union. Translator’s note.

cient of action (efficiency) increases. In view of the high efficiency of these photoelements, the application of an auxiliary voltage is not necessary, but a small applied voltage increases the photocurrent. Blocking photoelements (cuprous-oxide cells) make possible the direct conversion of light into electrical energy, since their maximum photocurrent already reaches \(10^{-4}\) A per lumen. Table 1 shows an increase of the photocurrent, in comparison with alkaline photoelements, by several thousand times. As is also seen from Table 1, cuprous-oxide photoelements of the new type possess great sensitivity to red rays and are suitable, in view of the far-reaching proportionality between the photocurrent and the intensity of light, for measurements in the infrared region. From the limiting infrared wavelength, let us calculate, according to Einstein’s quantum condition,

\[ N=\frac{a}{h\nu_0}, \]

the electron yield and compare it with the yield in alkaline photoelements.* We then find that the yield has increased only tenfold, since the quantum conditions characterize only the energy balance of separate elementary processes, but do not give the measured coefficient of useful action. The external coefficient of useful action is considerably greater than the value given by the quantum conditions, and depends on causes still unknown.

While studying the unipolar conductivity of crystals, the author as early as 1927 observed the appearance of an electromotive force when a crystal was illuminated. For example, if, on a crystal of galena serving as a detector, one illuminates the place where the detector needle touches the surface of the crystal, a weak electromotive force arises, reaching 20 mV. For this simple experiment, French lead sulfide (galena), possessing good detecting properties, is also suitable. With the aid of light filters it is easy to prove that this effect is sensitive to red rays, whereas short-wave radiations produce almost no action. Of particular interest is the course of the dynamic—

* \(a\) is the amount of light energy, \(N\) is the number of electrons torn out. The yield is expressed in coulombs per calorie. Trans. note.

characteristic of such a detector under the action of light and without it, shown in Fig. 11. Whereas in the unilluminated crystal (curve 1) the rectified current increases with increasing alternating voltage, for the illuminated crystal (curve 2) it indicates an extremely strange change in the direction of the current. This previously incomprehensible effect now finds a simple explanation in the presence of a photocurrent directed in the direction opposite to the rectified current. At an alternating voltage less than 1.32 V the photocurrent predominates; at 1.32 V the two currents compensate one another, and with a further increase in voltage the rectified current already predominates, which entails a reversal in the direction of the resultant current. From Fig. 11 one can discern a very interesting property of the photocurrent. If one plots the differences of the rectified current with illumination and without it, one obtains a dashed curve with a maximum at 2.0 V. Since this curve represents the photocurrent under the action of an alternating voltage, it follows from it that the photocurrent, increasing at a certain applied voltage, falls again with a further increase in voltage.

Fig. 11. Dynamic characteristic of a detector with lead glance without illumination and with it.

Further development led to the above-described photoelements, which differ from light detectors by greater sensitivity, complete constancy of properties, and the use of a larger surface. The discovery of the new photoeffect described here, strictly speaking, occurred gradually. As early as 1876, i.e. more than 50 years ago, Adams and Day reported their observation of the emergence of an independent electromotive force in selenium elements of a special kind. Among a thousand selenium elements, one was

was found (Fritts in New York), which, when illuminated, gave an electromotive force without an auxiliary battery. Werner Siemens wrote in the Proceedings of the Prussian Academy of Sciences (Sitzungsberichte der Preussischen Akademie) of February 12, 1885, p. 147: “Here we are indeed dealing with a phenomenon of great scientific importance... The very existence of separate selenium cells with the properties described is a fact of great scientific significance, since here, for the first time, there occurs the direct transformation of light energy into electrical energy.” In view of the difficulty of reproducing them and in view of the instability of their properties, these cells were apparently soon consigned to oblivion. On the basis of recent work^8, the change in resistance of selenium cells in general is explained by the appearance of an internal electromotive force (oppositely directed) and by polarization under the action of light. Much later, in 1923, Geiger^9 observed the appearance of an electromotive force when a crystal of galena was illuminated. Independently of him, Koblentz described, in 1922, a similar phenomenon in molybdenite. When only certain limited places on plates of molybdenite are illuminated, a voltage appears. In both cases, however, fatigue is noticeable, so that within a few minutes the current changes greatly, and the current strength changes upon renewed illumination. Moreover, there is no proportionality between the intensity of the incident light and the photocurrent. It therefore seemed that this effect had no practical significance, and they were not even certain of its photoelectric nature.

Knowing nothing of the observations cited here, the author in 1927 made the discovery described above and arrived, possibly by chance, at a device in which the same or a similar effect is obtained in a form sufficiently applicable in practice.

PART II

APPLICATION OF THE PHOTOCELL IN PHOTOMETRIC MEASUREMENTS

Objective photometry by means of thermoelectric or photochemical methods, in the case of colored radiat—

...values for the intensity of light, always different from those obtained visually by the eye. At the time of its discovery, the photoelectric effect was considered altogether inapplicable to the purposes of general photometry, since it seemed that exclusively ultraviolet rays were acting. Even after the existence of the photoelectric effect for the entire visible spectrum had become known, its use in photometric measurements was slow to develop. Only in recent years has the universal significance of photoelectric methods been recognized. This recognition was due partly to improvements in methods, and partly to the manufacture of blue- and red-sensitive photoelements. Since the “cuproxides” already mentioned operate without an auxiliary voltage and their photocurrent can be measured with an ordinary galvanometer, photoelectric measurements with them are remarkably simplified. It is therefore advantageous to use them for various photometric and colorimetric work, and also for determining color temperatures (pyrometrically). To determine the blackening or absorption of light by some solution, it is sufficient merely to place the absorbing medium in the path of the source light in such a way that the illumination of the photoelement is correspondingly weakened. From the ratio of the photocurrents with the absorbing medium and without it we immediately obtain the transmittance, since both for alkali photoelements and for the new cuproxide ones there is proportionality between the photocurrent and the intensity of the incident light. With alkali photoelements, of course, special care must be observed in the work in order to exclude the influence of fatigue.

The photocurrent of an alkali photoelement corresponding to an illumination of 1 lux can easily be measured with a galvanometer. The use of highly sensitive mirror galvanometers makes it possible to measure illuminations with photoelements down to \(10^{-2}\) lux. However, in measurements with spectrally dispersed light or for measurements with a narrow slit—if amplifiers are not used, of which more will be said below—one can work only with electrometers. Despite their simplicity, this causes certain difficulties, which, of course, have to a considerable extent been eliminated by manu-

introduced recently in the apparatus. The difficulties mentioned are due to the high internal resistance of photoelements, which under illumination is on the order of \(10^8\) to \(10^{12}\ \Omega\). Because of such a large resistance of the element, the insulation of the photoelement holder and of the leads must be excellent. Above all, the photoelement and the leads going to it must be completely electrostatically shielded; otherwise electrostatic charges on the walls may have an interfering effect.

Fig. 12. Zeiss microphotometer.

With especially carefully constructed elements² with quartz insulation these difficulties can be eliminated. In the electrometric method, the voltage drop across a high-ohmic resistance is measured; it must be emphasized that the value of this resistance must not exceed \(1\%\) of the resistance of the photoelement, otherwise the measured resistance will not be proportional to the illumination. The manufacture of the high-ohmic resistances needed for this is still difficult. According to the ingenious idea of Koch,³ this high-ohmic resistance can be replaced by a second photoelement, illuminated by the same light source, whereby random fluctuations of the sources will be compensated simultaneously. The design of Koch’s microphotometer is based on this principle.

But on the basis of new work by Perykka⁴, it has been possible to obtain sufficiently constant high-ohmic resistances up to \(10^{12}\ \Omega\) by cathodic sputtering, even on amber. In the microphotometer of Deitz⁵ a high-ohmic resistance of this type (after Koenig) is used. In this photometer the plate being measured is automatically moved in front of a narrow slit in such a way that the light beam passing through it falls on the photoelement. The photovoltage is measured electrometrically across the high-ohmic resistance, and the deflections of the electrometer fiber are photographically recorded. With this precision instrument one can measure an entire spectrum in a few minutes and, instead of that, by switching on a magnification of \(1:500\), one can measure the distribution of intensities in each individual spectral line.

Fig. 13. Photometer for astronomical measurements.

For measurements of surface brightness, especially of celestial bodies, one may use the photoelectric photometer in its simplest form, consisting of an electrometer with a photoelement attached to it and with the necessary limb. Fig. 13 shows such a photometer after Elster and Geitel, in a new construction by Dorno.* An even more sensitive method than that described above is the determination of the photocurrent from the charging time of the electrometer. Since in this case one can measure a photocurrent of \(10^{-15}\ \mathrm{A}\), the light sensitivity exceeds even that of the eye.

* Built by the firm Günther und Tegetmeyer, Braunschweig.

Important applications have been found by electrometric methods in stellar photometry. At the Neubabelsberg Observatory, Guthnick measured photoelectrically stars of the eighth magnitude, using Lindemann’s quadrant electrometer. Since the accuracy of these instruments depends almost not at all on their position, they can be used directly at the telescope. The methods described are simple, and over a long period of time only their constructive improvements appear. Only thanks to the development of cathode tubes, by means of which amplification of photocurrents by millions of times is possible, has the field of application of photoelements expanded in the most unexpected way. At first, technical applications predominated, but nevertheless they were also successfully applied in scientific photometry. Rosenberg\(^6\) used a photoelement with the amplifier circuit shown in Fig. 14 as a stellar photometer, and in doing so attained a high degree of accuracy. The advantage of this method is the possibility of using an ordinary galvanometer instead of an electrometer with its very inconvenient electrostatic protection. Despite the large amplification, however, no significant increase in absolute sensitivity is achieved, since all fluctuations of the anode current and, especially, of the filament current interfere strongly.

Fig. 14. Amplifier circuit according to Rosenberg.

The circuit for connecting the photoelement with an amplifier is given in Fig. 14, where the circuit denoted “compensation” serves for measurement. By balancing the currents with the aid of a variable resistance, it is easy to compensate the anode current flowing through the galvanometer. The current in the anode circuit will change if, under the influence of the photocurrent, the grid potential changes; in this case the compensation is disturbed, and the galvanometer receives a deflection, which can serve as a measure of the illumination. The amplification can be very large and, in the corresponding

scheme one can obtain an amplification of the photocurrent by more than 600 thousand times.

The method developed at the Osterberg Observatory was applied to photometry of stars. An approximate picture of the achievements obtained there may be formed from the fact that the light of Jupiter gives a current of \(10^{-4}\) to \(10^{-5}\) A (after the amplifier).

A certain improvement in comparison with this amplifier circuit is the bridge circuit of Campbell shown in Fig. 15. In this circuit the photocurrent and the amplified current (in the bridge) are proportional to each other over wide limits. By changing the voltages on the grids \(Eg_1\) and \(Eg_2\), or by changing the ratio between the resistances \(r_1\) and \(r_2\), one can so set the bridge that, in the absence of photocurrent, the galvanometer \(g\) will stand at zero. When the photoelement is illuminated, a current will flow through the galvanometer; moreover, by an appropriate selection of the grid potentials, one can obtain proportionality between the deflections of the galvanometer and the illumination. In ordinary amplifier circuits there is no simple relation between them. With this circuit, however, one can obtain some increase in accuracy by using the null method, in which equilibrium in the bridge is restored by changing the grid voltage \(Eg_1\); then the photocurrent is obtained as the ratio of \(Eg_1\) to \(R_1\).

Fig. 15. Amplifier circuit according to Campbell.

The use of cuprous-oxide cells instead of alkali photoelements for the purposes of photometry has the advantage that they give, without additional batteries and amplification, a comparatively large photocurrent, which can be measured simply with a galvanometer, so that the measuring circuit is distinguished by great

simplicity. Also, thanks to the low internal resistance of these photoelements, lying between 100 and 1000 ohms, all the already mentioned difficulties connected with insulation disappear. In addition, cuprous-oxide cells are distinguished by their sensitivity to infrared rays. Nevertheless, for measuring very small quantities of light, the more sensitive instrument is the electrometer with an alkali photoelement.

The selenium and thallofide cells used, in addition, for photometric measurements produce under the action of light only a change in their resistance and, according to the terminology adopted here, cannot be called photoelements. The spectral sensitivity of a selenium cell is close to the color sensitivity of the eye. However, this very valuable quality is weakened by three shortcomings: first, as is known, the decrease in resistance does not disappear exactly simultaneously with the exciting light; second, selenium cells change their resistance over time, which occurs under the action of internal causes; third, temperature affects their light sensitivity.

Objective photometry with the aid of selenium cells, although possible, is hardly practically applicable.

Technical applications of photoelements

By using cathode lamps to amplify photocurrents it is possible, as we have already indicated earlier, to achieve new and completely unexpected technical successes, which are increasing almost with every day. For example, in recent times photoelectric control has been used in the rolling of red-hot metal, in which the luminous mass acts at the end of the rolling path on a photoelement, thereby automatically changing the course of the mill. Photoelements first found technical application in the transmission of images by telephone, developed while still using selenium cells, but acquiring real technical significance only with the transition from selenium cells to true photoelements. At present, many apparatuses are already operating for the transmission of images between Europe and America.

The transmitter of this apparatus consists of a rotating cylinder, serving as a holder for the images or pieces of text to be transmitted. A beam of light slides spirally over the rotating image, so that the intensity of the reflected light corresponds to the brightness of an individual point of the image, and a corresponding photocurrent arises in the photoelement. The amplified photocurrent is transmitted, with or without wires, to the receiver, and there is converted into a corresponding quantity of light, which produces an image on a drum rotating at the same speed and covered with light-sensitive paper. Of great technical importance here is the replacement of a constant photocurrent by an intermittent one with a frequency of up to 1000 hertz, which is much more convenient for amplification. Despite the fact that the image is broken down into separate elements—thanks to the large number of points—the Telefunken-Siemens system works with 40 points per square millimeter—the transmitted image is difficult to distinguish from the original. In the transmission of images, just as in television and sound cinema, increasing the photoelectric action of photoelements not only saves one amplification stage, but above all improves quality by reducing distortion.

The use of photoelements for sound cinema goes back to three German inventors: Focht, Massolle, and Engl,⁸ who, as early as 10 years ago, made a photographic recording of sound and developed the Tri-Ergon system. Since in Germany very high demands are made of all kinds of musical productions, the sound film did not meet with the expected approval. In contrast to Germany, America was delighted; and sound cinema appeared in Germany only after its return from America, where much money had been provided for the creation of sound films and for improving the apparatus.

For recording sound by light, only two methods are used, in which the amplitude of the sound corresponds either to the width or to the intensity of the image of a certain slit on the film. In Figs. 16 and 17 two pieces of film with such a sound recording are shown. Fig. 16 is according to the method of varying intensity, and Fig. 17

by the method of changing the amplitude (width). The recording may be different, while the reproduction is one and the same, by means of a photocell giving a variable photocurrent of amplitude and frequency corresponding to the permeability of the sound recording. If we leave aside complicated amplifying devices, then in sound films, just as in image transmission and in television, what in essence takes place is the simple conversion of light into corresponding photocurrents. In sound cinema these photocurrents are converted into acoustic oscillations. In transmission and in television they are converted back into light.

At the same time, Kerr cells came to be used as an inertia-free light relay, and at low powers—glow-discharge lamps.

Fig. 16. Sound recording by change of intensity.

Fig. 17. Sound recording by change of amplitude.

The action of a transmitter in television in this respect fully corresponds to the action of a transmitter in the transmission of images. Here, likewise, there is scanning point by point of a picture, which lasts for a small fraction of a second.

The receiver gives individual light elements successively in time, but again at such a speed that the impression is obtained of the whole picture as a whole.

An interesting scientific application of the principles of television was given by Dauvillier. As is known, in X-ray screening the secondary rays interfere, veiling the picture. Dauvillier abandoned the use of a single cone of rays with the obtaining of all points of the screening picture simultaneously, and works with narrow beams, which likewise

as in television, by means of a Nipkow disk with spirally arranged holes, move over the body being transilluminated. To avoid the shortcomings that arise when the screen is viewed directly, Dauvillier replaces the screen with a photoelectric cell. The resulting photocurrent is then amplified so that it can control a glow-discharge lamp in step with the incident light impulses. Viewing this lamp through a synchronously rotating Nipkow disk gives the entire transillumination picture at once. In the apparatus shown in Fig. 18, both for analysis and for synthesis of the picture a common Nipkow disk is used. Owing to the high sensitivity of the photoelectric cell, the intensity of the transilluminating radiation ordinarily used may be reduced a thousandfold; therefore the irradiation can in practice be carried out for as long as desired, without fear of harming the patient. In this way, with a low intensity of the rays, one can obtain a bright picture and record it on film, which naturally makes possible not only extremely brief instantaneous photographs, but also a true motion picture.

Fig. 18. Dauvillier’s X-ray-television apparatus.

An area of application of the photoelectric cell entirely different from all those described above is the obtaining of electrical oscillations from them. The scheme of such a device was given by Rozing[^10] and used by him in a reading machine for the blind. If a battery, a photoelectric cell, and a high-resistance element with a parallel capacitance are connected in series, then, as in a glow-discharge lamp, oscillations arise which, when the photoelectric cell is illuminated or со

all are interrupted or change their tone according to the intensity of the illumination (as a consequence of the change in the internal resistance of the photoelements). Free oscillations arise if the ohmic resistance of the photoelectric circuit is considerably greater than the inductive resistance. Rosint, as was said above, applied this effect in the design of a reading machine for the blind. Individual letters, by means of a simple device and the indicated circuit, produce different acoustic signals depending on their size and brightness, resembling the Morse alphabet; thus the optical pattern of the type is represented acoustically.

Of course, it is quite improbable that the facts indicated here exhaust all the interrelations between light and electricity. With refinement of the methodology, new phenomena will be discovered, some of which can perhaps already be foreseen at the present state of our knowledge, but for the time being cannot be observed. And probably in the following year this connection between light and electricity will indeed acquire enormous interest. The variety of practical applications of photoelements is strikingly great and grows with each day. But still more striking is the fact that we still cannot convert the energy of sunlight—whose electromagnetic nature has been known for several decades and whose field strength is equal to a volt per centimeter—into useful electrical energy. A small calculation of the photocurrents produced by cuprous oxides shows that for the moment we scarcely have the possibility of applying this effect for the rational conversion of light energy into electrical energy. Since cuprous oxide, at a luminous flux of 1 lumen per square centimeter of surface, gives a photocurrent of \(10^{-4}\) to \(10^{-5}\) A, and since the maximum solar radiation is equal to 10 lumens per unit surface, the electrical power reaches only \(1 W\) with an illuminated surface of \(1\ \text{km}^2\). True, increasing the photoeffect by 10–100 times would already lead to practically applicable values of the energy. The first approximation to the practical use of the new photoeffect described by the author is the work done in

*

research laboratory of Siemens. As shown in Fig. 19, a photocell of cuprous oxide with an area of 49 cm², illuminated by the sun, drives a very small motor.

Fig. 19. A photocell, when illuminated by sunlight, sets in motion a small electric motor connected to it.

Correction note. Since the writing of the present article (July 1930), it has proved possible to obtain photocells (cuprous oxide) of a new type with a photocurrent 10–100 times greater than the previous ones. A preliminary communication on this subject will appear soon (Lange).

LITERATURE

1. W. Hallwachs, Wied. Ann. 53, 301, 1888; 34, 731, 1888.
2. I. Elster u. Geitel, Wied. Ann. 43, 225, 1891.
3. L. Marton u. E. Rostaş, Z. Techn. Phys. 9, 32, 1928.
4. M. Pirani u. E. Lax, Z. techn. Phys. 3, 232, 1923.
5. P. Selonyi, Physik. Z. 30, 933, 1929.
6. R. Suhrmann, Z. wiss. Photogr. 29, 156, 1930.
7. B. Lange, Physik. Z. 31, 139, 1930.
8. D. Reichinstein, Naturwiss. 18, 685, 1930.
9. P. H. Geiger, Physic. Rev. 2, 461, 1923.
10. W. W. Coblentz, Sci. Pap. Bureau Stand. 446, 16, 1922.

1. E. Steinke, Z. Physik 11, 215, 1922.
2. R. Suhrmann, Z. wiss. Photogr. 29, 156, 1930.
3. P. P. Koch, Ann. Physique IV, 39, 705, 1912.
4. E. Perucca, Ann. Physique V, 4, 252, 1930.
5. G. Hansen, Ann. Physique IV, 78, 558, 1925.
6. H. Rosenberg, Naturwiss. 9, 359, 389, 1921.
7. N. R. Campbell and D. Ritchie, Photoelectric Cells. P. 141, London, 1929.
8. I. Engl, Der tönende Film. Sammlung Vieweg. Braunschweig, 1927.
9. F. Schroter, Elektrische Nachrichten-Techn. 6, 439, 1929.
10. B. L. Rosing, Z. Tech. Phys. 11, 177, 1930.