Recent Advances in the Development of Photocells for Infrared Light*)

A. Elliott

Barrier-Layer Photocells for the Infrared Region of the Spectrum

Thallous Sulfide Photocells**)

Thallous sulfide barrier-layer photocells, sensitive in the near infrared region of the spectrum, were first developed at the Leningrad Physico-Technical Institute of the Academy of Sciences of the USSR in 1937. A detailed description of their characteristics was given by B. T. Kolomiets¹ in 1938. In 1939, without reference to the work of the Soviet author, a publication appeared by Nix and Treptow² on the development of these elements in the USA.

An interesting feature of these photocells at that time was the sign of the photo-emf, opposite to the customary one observed in cuprous-oxide and selenium photocells. In the latter, the metal electrode adjacent to the active barrier layer is charged negatively; in thallous sulfide cells it is charged positively. It is true that at first Kolomiets sporadically obtained photocells with one or the other sign of photo-emf. In this case, photocells of the “negative” sign of photo-emf (coinciding with the sign for cuprous-oxide and selenium photocells) had an integral sensitivity of about \(100\ \frac{\mu A}{lm}\) and a spectral sensitivity distribution characterized by curve 1 in Fig. 1. Photocells of the “positive” sign were characterized by the spectral sensitivity shown by curve 2 in Fig. 1 and by an integral

) Chapter III from the collection Electronics and their application in industry and research*. Edited by Bernard Lovell, London, The Pilot Press Ltd., 1947. (Chapt. III. Recent advances in photo-cells for the infra-red, by A. Elliott.) Abridged translation and additions by P. G. Borzyak.

**) Since the author gives no references to Soviet works, this paragraph has been rewritten by the translator.

with sensitivities of thousands of microamperes per lumen. Considering band energy diagrams for electrons in semiconductor–barrier layer–metal systems, under a definite assumption about the nature of the barrier layer, Yu. P. Maslokovets³ related the sign of the barrier photoeffect to the sign of the electrical conductivity of the semiconductor from which the photocell is made.

Kolomiiets⁴ developed a technology ensuring the production of photocells consisting of a semiconductor with electron conductivity and possessing a positive photoeffect, with an average sensitivity of \(5000—7000\ \frac{\mu A}{lm}\) under illumination by a tungsten incandescent lamp with a filament temperature of \(2840^\circ K\).

Fig. 1. Spectral sensitivity of thallium sulfide photocells:
1 — with a “negative” sign of the photo-emf, 2 — with a “positive” sign of the photo-emf.

A mixture of thallium sulfide (prepared by fusing carefully purified thallium and sulfur, taken in quantities corresponding to the stoichiometric ratio) and pure thallium is loaded into a crucible made of refractory glass. From the crucible, by heating it in vacuum to \(600^\circ C\), a film of thallium sulfide with an admixture of thallium is deposited onto an iron plate. The film thus obtained on the iron substrate is subjected successively to oxidation in air at \(120^\circ C\) and to washing in running water.

The purpose of these operations is to create a surface layer consisting of stoichiometric thallium sulfide, which is the so-called chemical barrier layer, and a layer lying beneath it made up of a mixture of thallium sulfide with thallium oxide. The presence of thallium oxides in the layer is essential for obtaining photoelectrically active specimens. The upper semitransparent gold electrode is applied by cathodic sputtering in an air atmosphere. Since the photocells obtained in this way deteriorate rather quickly when exposed to air, they are enclosed in evacuated glass bulbs. To protect the photocells from the harmful action of moisture (which inevitably appears during sealing of the bulb), they are first coated with zapon lacquer.

It further proved useful, in order to increase the service life of the photocells, to fill the bulb, after evacuation, with hydrogen to a pressure of \(300—400\) mm Hg. However, as is evident from publication⁵, the question of the stability of photocells over time is still on the agenda.

The technology for producing photocells of Nix and Treptow is different. The metallic substrate is coated with metallic thallium, which is purified by heating in vacuum to \(280^\circ\mathrm{C}\). By means of a glow discharge in dry hydrogen sulfide or sulfur vapors, a layer of thallium sulfide is formed on the thallium surface; onto this, likewise by cathode sputtering, a semitransparent upper electrode of platinum or gold is deposited. The authors make no mention of any oxygen treatment of the film. The photocells obtained in this way had a positive sign of the photo-emf, and some of them had a sensitivity up to \(300\text{—}600\ \dfrac{\mu\mathrm{A}}{\mathrm{lm}}\). With the aid of a technological improvement, which the authors do not describe, they succeeded in obtaining photocells with a sensitivity up to \(6000\ \dfrac{\mu\mathrm{A}}{\mathrm{lm}}\). Their characteristics are similar to those of Kolomiets’ elements. However, owing to instability, they remained only experimental samples.

In the dynamic regime a photocell behaves like a capacitance (of the order of \(0.1\ \dfrac{\mu\mathrm{F}}{\mathrm{cm}^2}\)), shunted by the resistances of the blocking layer and of the external load, since the photoelectric processes in them are not very inertial. The frequency characteristic of such a system must be falling and dependent on the load resistance. In Fig. 2, for comparison, the frequency characteristics are presented for various photocells: selenium (curve 1), cuprous (2), and thallous sulfide (3 and 4) at the load resistances indicated in the figure. The voltage values for the thallous-sulfide element lie higher, although the illuminances in the case of the cuprous and selenium elements were greater. A fairly good form in the range of audio frequencies is exhibited by the characteristic at small external resistances. The frequency characteristic can be further improved by reducing the dimensions of the photocell, in order to reduce the capacitance acting in the circuit. All these circumstances were used by Kolomiets and the Leningrad KINAP plant\(^5\) for successful experiments in the application of thallous-sulfide photocells in sound cinema. Although this is not related to infrared technology, nevertheless it points to good prospects in this field as well, if the question of stability is resolved.

Fig. 2. Frequency characteristics of photocells:
\(1\) — selenium, \(2\) — cuprous, \(3\) and \(4\) — thallous-sulfide.

Silver sulfide barrier-layer photocells (FESS)*)

The development and study of barrier-layer photocells made of silver sulfide, begun at the Kiev Institute of Physics of the Academy of Sciences of the Ukrainian SSR in 1937,^6 led in 1940 D. S. Geikhman and M. E. Soroka^7 to the production of technically finished devices, which began to find application in industrial and laboratory practice.^7,8

Fig. 3. Spectral sensitivity of a silver sulfide photocell (FESS).

Fig. 4. Dependence of photo-emf and photocurrent on illumination for a silver sulfide photocell at various load resistances.

At the same time, concurrently with the development of applications, A. A. Il’ina also investigated the principal characteristics of the photocells. Work on these photocells, now known as FESS, and the study of their properties were continued. In an article by E. G. Miselyuk and V. E. Kosenko,^9 now in press, from which, with the kind permission of the authors, some of the data presented below have been borrowed, a detailed survey is given of the investigated characteristics of photocells now being produced.

*) This paragraph has been added by the translator, since in the original there is no mention of data on these photocells.

The region of spectral sensitivity of the Ag–S photocells is the same as that of selenium photocells; however, the form of the spectral characteristics of these cells is somewhat different. The average characteristic for Ag–S is represented by the curve in Fig. 3. The points give an idea of the scatter for different specimens. The integral sensitivity of average-quality photocells is 3500–4500 \(\frac{\mu A}{lm}\)

Fig. 5. Frequency characteristics of silver-sulfide and selenium photocells at different load resistances.

for the light of an incandescent lamp with filament temperature \(2360^\circ K\). Photocells with a sensitivity of 7000–8000 \(\frac{\mu A}{lm}\) are often obtained.

In Fig. 4 are presented the dependences of the magnitude of the photo-emf and photocurrent on illumination. The photocurrent curves were taken at different resistances in the external circuit of the photocell. The greater the load resistance, the sooner the characteristic deviates from a linear course.

The frequency characteristics of Ag–S are not very satisfactory, as is evident from Fig. 5, where comparative data are given for a selenium cell and Ag–S at two different loads. However, even in the case of Ag–S, the observed fall of the frequency characteristic is due not to the inertia of the physical processes connected with the photoeffect, but is caused by the large capacitance, reaching \(30 \frac{\mu F}{cm^2}\). The latter, in turn, is a consequence of the structural features of Ag–S\(^{10}\). The measured capacitance of the selenium photo-

element, whose characteristics are shown in Fig. 5, was equal to \(0.045 \frac{\mu A}{cm^2}\), while that of the FESS compared with it was \(25 \frac{\mu A}{cm^2}\). However, there are fields of application where it is not necessary to use modulated light and where FESS cells prove convenient, especially owing to their great “viability” and durability. Specimens of photocells are known that have been in operation since 1941–1942 and have retained quite considerable sensitivity.

Lead Sulfide Photocells

In 1904 Bose\(^{11}\) in the United States took out a patent for a cell made from a natural crystal of lead sulfide (galena). However, no interest was shown in this cell during the following 26 years, until reports about them again appeared by Lange\(^{12}\) and Grützmacher\(^{13}\). Subsequently they were investigated in the work of Fisher, Gudden, and Troye\(^{14}\).

[In Fig. 6: “Grid”; “Galena”; dimensions \(45\ \mathrm{mm}\) and \(30\ \mathrm{mm}\).]

Fig. 6. Construction of a galena photocell.

Finally, in 1942 the firm “Zeiss Ikon” in Dresden took up the development of these photocells (unpublished). Its “grid” cells consist of a piece of galena with a carefully polished surface, against which is pressed a grid made of very fine tungsten or molybdenum wire (Fig. 6). When the cell is illuminated through the grid, the electron current is directed from the crystal toward the grid (negative sign of the photo-EMF), although specimens with the reverse direction of current are also encountered. Attempts to replace the grid electrode with a semitransparent gold or silver one, obtained either by evaporation in vacuum or by cathode sputtering, did not give good results. The sensitivity of the cell can be increased by the following treatment. After polishing, the crystal is subjected to preliminary heating in vacuum, after which it is mounted and again heated in vacuum for 1 hour at \(160^\circ\mathrm{C}\).

In addition, it was found that the deposition on galena of layers of lead or sulfur, followed by subsequent heating, considerably increases the sensitivity of the cells. Which exactly is required for this—lead or sulfur—depends on the particular specimen in question. The effect of treatment for one particular case is illustrated in Fig. 7. Here \((a)\) corresponds to measurements on a clean polished crystal; \((b)\)—after preliminary heating; \((c)\)—after deposition of lead; and \((d)\)—after secondary heating.

According to unpublished measurement data by Karolus, the sensitivity of these elements does not depend on the frequency of light modulations up to \(40\,000\) cycles.

In Fig. 8 are shown curves of the dependence of the photocurrent on illumination at various resistances in the external circuit. The light source in this case was a lamp operating at a color temperature of \(2850^\circ\mathrm{C}\).

Fig. 7. Influence of treatment on the sensitivity of a galena photoelement:
\(a\)—crystal before treatment, \(b\)—after preliminary heating, \(c\)—after deposition of lead, \(d\)—after secondary heating.

Fig. 8. Dependence of photocurrent on illumination for a lead-sulfide photoelement.

The spectral characteristic of the sensitivity of the elements is given in Fig. 9. Its red limit lies farther than that of many lead-sulfide photoresistors (see Figs. 16, 17).

One of the most interesting properties of these grid elements, especially from the standpoint of military applications, is the insignificant change in their sensitivity to modulated light when a constant component of high intensity is superimposed on it. It was found that a constant illumination of \(25\,000\) lx reduces their sensitivity by only 24%, whereas the sensitivity of lead-sulfide photoresistors under these conditions falls by more than 98%. The noise level at constant high illuminations remains very low. All this makes it possible to use grid elements made of galena for phototelephony purposes in daylight.

Fig. 9. Spectral sensitivity of a lead-sulfide photoelement.

The sensitivity, measured in millivolts per lumen, increases as the temperature is lowered. In one case the sensitivity at liquid-air temperature was 30–100 times higher than the sensitivity at room temperature. The internal resistance

of the element in this case also increased by 15–30 times. The increase in sensitivity to modulated light with a lowering of temperature turns out to be less than for constant illumination, which is unclear and requires confirmation.

THALLIUM SULFIDE PHOTORESISTORS

Thallium sulfide photoresistors, under the name “tallophide” elements, were described by Case^15 in 1920 and since that time have been developed in various countries*).

Fig. 10. Dependence of the resistance of a thallium sulfide photoresistor on temperature.

The resistance of thallium sulfide is high and, in order to obtain elements with a resistance of several megohms, an interdigital system of electrodes is used. The combs are deposited from Aquadag onto a glass substrate, which is then covered with a thin layer of thallium sulfide. There are various methods of preparing and processing thallium sulfide. The method of the laboratories of the British Admiralty consists of the following. Thallium and sulfur, taken in a proportion corresponding to the compound \(\mathrm{Tl_2S_3}\), are fused in vacuum. The product obtained is fused, likewise in vacuum, with thallium hydroxide hydrate; for one molecule of \(\mathrm{Tl_2S_3}\), two molecules of \(\mathrm{Tl_2OH_2O}\) are taken. If the mixture is kept in the molten state for 10 minutes, it separates into two parts. After solidification these parts are separated; the lower one is retained, consisting of shiny gray flakes.

*) We note their development in the Soviet Union in 1938 by A. A. Sivkov.^16

A glass plate with aquadag electrodes is mounted in a glass bulb provided with leads from the electrodes. The bulb is soldered to a vacuum system and pumped while being heated to 500–600° C. Then the prepared thallium sulfide is introduced, which after evacuation is sublimed onto the glass plate. After this, oxygen is introduced to a pressure of about 0.2 mm Hg and the element is heated for 80–100 minutes at 300–320° C. During cooling the element is subjected to the action of SO₂ and mercury vapors. The mercury

Fig. 11. Dependence of the resistance of thallium sulfide photoresistance on temperature.

Fig. 12. Spectral characteristic of thallium sulfide photoresistance at 18° C.

vapors exert a stabilizing action on the element and can increase its sensitivity.

The following data will refer, unless otherwise specified, to British elements.

The dark resistance of the elements at room temperature amounts to several megohms. Temperature dependences of the resistances are presented in Figs. 10 and 11. The curves are characterized by the presence of a sharp break and by a high value of the temperature coefficient.

Since in Britain thallium sulfide photoresistances are intended for applications not requiring modulated light, the measure of sensitivity here is the change in the resistance of the element when illuminated by a standard source. As the latter there was chosen a tungsten lamp with a color temperature of 2800° K, provided with an infrared light filter of the “Signal Schoal Black” type.

The lamp is set up in such a way that the illumination on the surface of the photoelement, in the absence of the light filter, is equal to 2.6 lx. The luminous resistance of the element is determined when it is illuminated through the specified light filter. The ratio of the dark resistance to the luminous resistance thus determined is close to 3.

Figure 12 presents the spectral characteristic of an ACE No. 989 thallium-sulfide element (Admiralty Signal Establishment) at room temperature (18° C). The spectral characteristic of the same element, cooled to 0° C, is presented in Fig. 13. It is not known, however, how typical such a deformation of the spectral characteristic is for all thallium-sulfide elements. It should be pointed out that, despite such an unfavorable deformation, the element for which the characteristic in Fig. 13 was taken possessed greater sensitivity in the infrared region at the lower temperature, as determined by the method indicated above.

Fig. 13. Spectral characteristic of a thallium-sulfide photoresistance at 0° C.

Figure 14 gives oscillograms characterizing the inertia of the elements’ photocurrents, obtained with the use of rectangular light pulses at a frequency of 50 Hz. For comparison, oscillogram 1 is given, taken for a vacuum oxygen-cesium photoelement. Oscillograms 2 and 3 refer respectively to the thallium-sulfide resistances of the Research Laboratory of the British Admiralty and of Zeiss.

Thallium-sulfide photoresistances are used as detectors of infrared radiation from tungsten incandescent lamps equipped with corresponding light filters. Their sensitivity does not extend so far into the infrared region that they could be used as indicators of low-temperature radiation sources. They are used chiefly for communication (phototelegraph and phototelephone), as well as for signaling devices against night intruders, or infrared barriers (for example, at harbor entrances). In particular, these elements were used in the Zeiss 80-mm optical telephone of the 1934 model,^17 used during the past war by German troops in North Africa. Since, by means of the receiver objective, an image of the cross section of the transmitter light beam can be obtained

very small, it is advantageous, in order to preserve the optimum sensitivity of the photoresistor, to make its element of minimal dimensions. Photocells with a surface area of the order of \(1\ \mathrm{mm}^2\) were used. (A description of this optical telephone is available in the Russian literature\(^{18}\).)

Already during the war the Germans used a larger and more advanced model of the optical telephone, with lens diameters of 130 mm for the receiver and 250 mm for the transmitter, and with a different modulation system. The light beam in the transmitter passes through a grating and, after reflection from a modulating mirror set in motion by the microphone current, projects the image of the first grating onto the second. The quantity of light passing at each given moment depends on the position of the modulating mirror. This large model, like the later 80-mm specimens, was equipped, instead of thallium-sulfide, with lead-sulfide photoresistors, which have better frequency characteristics.

LEAD-SULFIDE PHOTORESISTORS

Before 1939 several publications by German authors appeared\(^{12,14}\), in which it was shown that the photosensitivity of lead sulfide extends in the spectrum to \(3.5\) and, possibly, \(4.5\,\mu\). Soon the Germans realized the possibilities of military application of lead-sulfide photoresistors, and publications on further investigations were discontinued. In January 1945 a description was obtained of the method for manufacturing lead-sulfide photoresistors by the electroacoustic firm in Kiel (ELAK). From that time work on these elements began in England and America*).

Lead-sulfide photoresistors are manufactured by two methods: 1) deposition of a lead-sulfide mirror from a solution of a lead salt with a corresponding sulfidizing reagent (the “wet method”) and 2) application of a lead-sulfide film by evaporation, with subsequent heat treatment in oxygen or another gas (the “dry method”).

Let us turn to a description of the ELAK firm’s method. For the element a blank is made from Duran glass in the form of a double-walled thermos—

Fig. 14. Oscillograms of photocurrents from rectangular light pulses with a frequency of 50 cps.
1—vacuum oxygen-cesium photocell, 2 and 3—lead-thallium photoresistors.

*) See, for example,\(^{19,20}\).

shaped vessel (Fig. 15). The bulb is opened in such a way that lead sulfide can be deposited on the flat end portion of the inner cylinder, which has been previously provided with platinum electrodes, obtained by the firing method at 700° C, in several layers, followed by a layer of gold.

Solutions of chemically pure reagents are prepared:

1) 400 g of lead sugar in 1000 ml of water,
2) 40 g of thiourea in 340 ml of water,
3) 240 g of caustic soda in 360 ml of water.

The glass surface is cleaned in the usual manner, washed with distilled water, pure alcohol, and ether, and is then immersed in a bath composed of the previously prepared solutions, mixed in the following order:

The rate of deposition depends strongly on the temperature. At a temperature of 45° C the deposition time is about 100 seconds. After this, washing with a strong jet of water is carried out and the second layer is deposited. After drying, the element is again sealed off, with precautions so that the film is not heated above 100° C, and is put on the pump-out. During pumping, the film is heated to 100° C, while the remaining part of the bulb is heated to 300° C. The duration of such heat treatment of the layer may be about 45 minutes, although the optimum conditions are unknown. After this treatment the element is annealed. To ensure the best vacuum conditions, a getter should be used. During heating the resistance of the film increases, and this increase may continue for several more days after manufacture of the element. The resistance of the element is about 100,000 ohms.

Fig. 15. Construction of a lead-sulfide photoresistor.

The method of the ELAC firm, as modified by ARL (Admiralty Research Laboratory), is as follows. The element is washed, as in the preceding case, after which it is treated with a 1% ammonia solution for 30 minutes and, without drying, is immersed in a mixture: 10 ml of solution (2) and 40 ml of distilled water, to which 30 ml of solution (1) is then added. The element is kept in this stirred mixture for 5 minutes. All this is done at room temperature. After 5 minutes, 2 ml of solution (3) is introduced. At this stage the solution becomes milky; the precipitated lead hydroxide dissolves again, leaving the solution slightly opalescent and, after 10–20 seconds, turning brown. After a minute

poured in is another 8 ml of solution (3), and the element is left in this mixture for ten minutes. It is then removed, washed with cotton wool moistened with a one-percent solution of ammonium sulfide in distilled water, and kept in 50 ml of the same solution, changed twice. From the surface of the element removed from the liquid, moisture is first removed with filter paper and then by drying in an oven at \(100^\circ\mathrm{C}\) for 10 minutes. It may be supposed that the principal impurity in the lead sulfide lattice is \(\mathrm{PbO}\). It will be seen below that the presence of oxygen in the fabrication of elements by the dry method is necessary in order to impart photosensitivity to them. It may be that the purpose of the final treatment of the elements with ammonium sulfide is to decompose the \(\mathrm{PbO}\) contained in the film.

The appearance of elements fabricated by the dry method developed in 1945 at ARL is similar to that shown in Fig. 15. The elements have two sensitive strips of lead sulfide measuring \(10\ \mathrm{mm} \times 1\ \mathrm{mm}\), provided with Aquadag electrodes, one of which is common to both strips. An element made of Pyrex or similar refractory glass, containing about 10 mg of lead sulfide, is sealed to a vacuum apparatus and, after evacuation, is heated to \(550^\circ\mathrm{C}\), while the front window is cooled by a jet of air. In this way the material is sublimed, depositing for approximately 2 minutes as an opaque layer on the front window. The air blast is stopped, and liquid air is poured into a trap near the element in order to collect the \(\mathrm{SO}_2\) subsequently formed. Oxygen is introduced into the system, and heating is continued. Blowing of the front window with air is resumed from time to time to prevent resublimation of the film. Then the furnace is removed and, quickly, before the outer walls of the bulb have cooled, while simultaneously blowing air on the bottom of the inner tube, the film is driven off from the front window with a hand torch and deposited on the opposite end of the inner tube, provided with electrodes. The oxygen pressure in this operation is \(0.2\text{--}0.3\ \mathrm{mm\ Hg}\).

At this stage the film is already photoconductive, but its sensitivity can be increased by appropriate treatment with oxygen and sulfur. The former consists in strongly heating the film in an oxygen atmosphere, while the rest of the element is kept as cold as possible. This is achieved by heating the front window with a hand torch until the film becomes completely conducting. After oxygen treatment, the resistance of the film should increase, and at the same time its sensitivity should increase under both steady and modulated illumination.

For sulfur treatment, the entire element is gently heated either in oxygen or in vacuum, in order to drive a small amount of sulfur from the walls. In doing so, a decrease in resistance is observed.

By such treatment the sensitivity of the element with respect to modulated light can be considerably improved.

The method of testing elements used in England, and similar to the German one, was determined by the fact that initially the elements were manufactured for detecting radiation sources with temperatures considerably below red heat. The standard ARL and TRL (Telecommunication Research Laboratory) test setup consists of a small furnace operating at 100° or 200°C and provided with a 2-mm aperture for emitting black-body radiation. By means of a disk with holes, the radiation flux is interrupted at a frequency of 800 cps and falls on the element connected to an amplifier. The dark current passing through the element is selected so as to obtain the optimum value of the ratio \(\frac{\text{signal}}{\text{noise}}\). Usually this requires applying to the element a voltage of about 50 volts. Then the distance between the source and the element is increased until the ratio \(\frac{\text{signal}}{\text{noise}}\) becomes equal to unity, and for this position the total radiation flux incident on the sensitive surface of the element (in microwatts) is calculated, which is also a measure of the sensitivity of the element. The results of some measurements for lead sulfide elements are given in the table.

Sensitivity of lead sulfide photoelements

I and II refer to two different bands of one paired element. The “ELAK” elements have one band each.
Source temperature 200°C.
Frequency of interruption of the radiation flux: 800 per second.
Amplifier bandwidth \(\pm 25\) cps.

From consideration of the table it is seen that the transition to low temperature is accompanied by an increase in sensitivity, not especially significant for some elements made by the dry method. Elements manufactured by the dry method are more sensitive than elements made by the wet method of manufacture.

Fig. 16. Spectral characteristics of lead-sulfide photoresistors manufactured by the dry method.

The most interesting characteristic of lead-sulfide elements is their spectral sensitivity, extending to \(3.5\,\mu\). It is important to note here that in the near infrared region these elements are considerably more sensitive than thermal indicators—bolometers and thermoelements \(^{19}\). In Fig. 16 are presented the spectral characteristics of certain elements manufactured by the dry method. Along the ordinate are plotted (on the left) numbers reciprocal to the energy values (in microwatts) corresponding to the ratio

\[ \frac{\text{signal}}{\text{noise}} = 1 \]

(on the right the energy values in microwatts are indicated). Curve 1 refers to an element with surface \(S = 0.25 \times 0.25\ \text{cm}^2\) and \(R = 2.7\ M\Omega\), curve 2—to an element with \(S = 0.25 \times 25\ \text{cm}^2\) and \(R = 2.7\,k\Omega\), and curve 3—to an element with \(S = 1.7 \times 0.1\ \text{cm}^2\) and \(R = 57\,k\Omega\).

Fig. 17. Spectral characteristic of a lead-sulfide photoresistor manufactured by the wet method, ELAK.

In Fig. 17 is shown the characteristic of a photoelement of the ELAK firm, manufactured by the wet method, and in Fig. 18—that of the AEG firm.

Turning to the last figures, we note a great difference in the spectral characteristics from element to element. Two main types of characteristic may be distinguished: some have a main maximum near \(1\,\mu\) and a smaller one near \(2.7\,\mu\); others—only one maximum, usually near \(2.7\,\mu\).

On the basis of ARL’s work on the dry method it may be said that the maximum near \(1\,\mu\) is associated with a higher degree of oxidation in comparison with those elements that have a maximum of sensitivity—

ness at 2.7 μ. The curves for AEG elements, taken from an unpublished report, refer to two methods of manufacture by the wet process. The elements of Method II, after deposition of lead sulfide, are subjected to heating in oxygen, whereas Method I does not include such an operation. This is in agreement with the statement just made.

Fig. 18. Spectral characteristics of lead-sulfide photoresistors manufactured by the wet AEG process.

On the basis of German reports and oral communications, the effect of cooling the elements down to −80° C is small; in this case the value of the red limit decreases, and the maximum of sensitivity also shifts toward shorter wavelengths (from 2.8 μ to 2.1 μ in one case with an ELAK element).

The inertia of lead-sulfide photoresistors manufactured in this manner is very small; however, it depends strongly on the method of manufacture. Wet-process elements (for example, ELAK) at room temperature have a sensitivity practically independent of the frequency of light up to 1000 Hz, and even at 10,000 Hz they still retain up to 80% of the maximum sensitivity. Upon cooling, the inertia increases considerably. The cooling effect is illustrated by Fig. 19, obtained when an ELAK element was illuminated by radiation from a temperature source at 800° C.

Fig. 19. Frequency characteristics of a lead-sulfide photoresistor (ELAK) at 20° and −185° C.

Fig. 20 illustrates the effect both of cooling and of superimposing constant illumination for an ELAK element. In the case where the element is illuminated only by interrupted radiation from a source at 800° C, the curves were taken—

seen: \(a\)—at \(-185^\circ\text{C}\) and \(b\)—at \(20^\circ\text{C}\). In the remaining cases, this intermittent irradiation was supplemented by constant illumination from a tungsten lamp. Under these conditions, curve \(c\) was taken at \(-185^\circ\text{C}\) and curve \(d\)—at \(-20^\circ\text{C}\). From the data presented it is evident that the high sensitivity of the cooled element is preserved only in the dark (without the application of constant illumination).

Elements manufactured by the dry method have poorer frequency characteristics. Oxidation can lead to greater inertia. Some elements with excess oxidation possess such great inertia that even after several seconds complete saturation of the photoconductivity is still not reached.

Fig. 20. Effect of cooling and of applying constant illumination on the frequency characteristic and sensitivity of a lead-sulfide photoresistor (ELAK):
\(a\)—at \(-185^\circ\text{C}\) without constant illumination; \(b\)—at \(20^\circ\text{C}\) without constant illumination; \(c\)—at \(-185^\circ\text{C}\) under illumination by an incandescent lamp; \(d\)—at \(-20^\circ\text{C}\) under illumination by an incandescent lamp.

Lead-sulfide photoresistors in the field of military applications were used for two purposes:

1) as detectors of infrared radiation from tungsten incandescent lamps for infrared telephony;

2) as detectors of infrared radiation from hot objects, such as aircraft exhaust pipes and ships’ smokestacks.

The first case of application has already been mentioned. The second group of applications is more recent. Although thermal detectors, such as bolometers, have been experimentally tested for the detection of ships, owing to their inertia they are not suitable for detecting such rapidly moving targets as aircraft. The firm ELAK used lead-sulfide photoelements in instruments known under the designation NMG42. They consisted of a 150-cm searchlight mirror—

of a body concentrating rays from the target (in the present case, the exhaust pipe of a night bomber) onto the element. The element was cooled with solid carbon dioxide in order to achieve the greatest possible sensitivity. To prevent fogging both of the photoelement window and of the 150-cm mirror, local electric heating was used. A system of auxiliary devices gave an indication of the target’s entry into the field of view, and also of its being on the axis of the instrument. However, this apparatus, even with the use of the scanning method, had too small a field of view (about \(9^\circ\)) to serve successfully for the detection of aircraft. It evidently was of little use, and its production was not continued thereafter.

Similar types of equipment were used at some stations on the continental coast for the detection of ships. These instruments evidently operated fairly well.

As far as is known, only the devices described above, using lead-sulfide photoresistors, were in operation and were actually used by the German armed forces. Nevertheless, a whole series of instruments was developed which may be called short-range devices, generally similar to the NMG42, but with much smaller optical systems; thanks to the better quality of the optics, almost the same sensitivity was achieved as in the large instruments. Equipment was under development for detecting aircraft both from the air and from the ground; for detecting, for example, factory chimneys from the air; and also for use in bombs and rockets that strike a target. In the latter case, what was meant was automatic striking of the hot parts of aircraft, ships, etc.

In the field of scientific and industrial use of these photoelements, one may confidently predict their application in radiation pyrometry for measuring surface temperatures up to \(100^\circ\) C. The advantage of photoelements in this field over thermoelements and bolometers lies in their lower inertia, which makes it possible to study phenomena that vary rapidly with time (see, for example, \(^{21}\)).

It is to be hoped that the use of these elements in near infrared spectroscopy will also be of great interest. The region between \(1\,\mu\) and \(3\,\mu\) can now be investigated with instruments of greater sensitivity, which will make it possible to increase the resolution. Many experimental difficulties connected with the “drift” of thermoelements can be eliminated by the use of modulated illumination and electronic amplifiers. Since many organic compounds have absorption bands in the spectral region covered by lead-sulfide photoelements, the use of the latter for analytical purposes may prove valuable.

Their use in astronomical spectroscopy also appears promising, in particular in the study of absorption spectra

atmosphere and the glow of the night sky. Since the absorption bands of water vapor and carbon dioxide lie in the sensitivity region of these elements, they may be used for hygrometry and gas analysis. True, attempts at direct measurements of absolute humidity must be based on more precise information about the limits of applicability here of Beer’s law.

LEAD SELENIDE AND LEAD TELLURIDE PHOTORESISTORS

Photocells made of lead selenide were the subject of secret research in Germany at least from 1938. However, despite the highly promising prospects for their use for military purposes, by the end of the war they were still not in production. Considerably less is known about their properties, as well as about methods of manufacture, than about lead-sulfide ones. They are attractive because their sensitivity extends to 5.5 μ (Fig. 21). For infrared spectroscopy they cover a region containing the principal absorption bands of many organic compounds, and therefore should be useful for chemical analysis.

Fig. 21. Spectral sensitivity of lead-selenide photoresistors.

Photocells made of lead selenide can be obtained by heating lead and selenium and subliming the resulting composition onto a plate provided with suitable electrodes. According to Gudden (unpublished), oxygen treatment is necessary in order to obtain photosensitivity, but the details are insufficiently known.

It is also possible to prepare these elements by a wet method, by precipitation from a mixture of solutions of lead acetate (lead sugar) and selenourea. Sodium hydroxide is not used in the solutions, and in this respect the process differs from the method of precipitating lead sulfide. The method was to some extent investigated by the firm ELAK, but again the details of the method have not been established. An essential circumstance is that on the surface provided with electrodes there is first formed a film of lead sulfide, which then acts as a “seed” for the deposition of the selenide. The lead-selenide film is subjected to heat treatment for several hours at a temperature between 350° and 500° C.

As regards the properties of lead-selenide photoresistors, little can be said beyond what is given by the spectral charac-

teristic (Fig. 21) of an element made by Gudden. Although here the red limit lies beyond \(5.5\,\mu\), according to other German investigators it has a smaller value. It is probable that, as in the case of lead-telluride photoresistors (see below), good sensitivity is obtained here only when the element is cooled.

Elements utilizing the photoconductivity of lead telluride are of greatest interest in that their spectral sensitivity extends farthest into the infrared region, and the red limit has been recorded at \(\lambda = 5.8\,\mu\) (Fig. 22). Unfortunately, there are no data whatever on its absolute sensitivity.

All the information presented here was obtained from O. Krentzsch, who carried out this work in the research laboratories of Siemenswerke and in the Physics Institute of the University of Göttingen.

Fig. 22. Spectral sensitivity of lead-telluride photoresistors.

The lead-telluride photoresistors manufactured so far were only experimental specimens. The elements consisted either of a glass plate with electrodes obtained by firing in platinum, or of a ceramic plate with carbon electrodes and a photosensitive layer deposited on them. Several vacuum elements were made in a form similar to that shown in Fig. 15. For the windows in these elements, thin quartz, Periplex, synthetic sapphire, or ruby were used. Synthetic materials are transparent up to \(5.5\,\mu\) and, with sufficient care, can be soldered to molybdenum glass. Lead telluride was prepared by heating a mixture of lead and tellurium in a graphite crucible in a protective atmosphere of argon or nitrogen, and was purified by sublimation in vacuum at a temperature of about \(900^\circ\text{C}\). The material was deposited on the substrate by means of a small heating coil.

Photoconductivity of the film appeared only after its oxygen treatment, which greatly reduced the dark conductivity—by a factor of 100 at room temperature. The treatment consisted of heating in oxygen at a pressure of \(10\)–\(100\) mm Hg and led to a dark-blue coloration of the layer.

It was found that all elements having sensitivity in the spectral region lying beyond \(3\,\mu\) reduce their sensiti-

tivity at all wavelengths when the temperature is raised from \(-183^\circ\) (the temperature of liquid air). The decrease in sensitivity in this case increases with increasing wavelength, as a result of which the value of the red limit decreases with increasing temperature. The greater the sensitivity at \(-183^\circ\) C in the long-wave region, the more strongly the indicated heating effect is manifested.

Figures 22–24 show the influence of temperature on lead telluride photoresistors. The element for which the spectral characteristic presented in Fig. 22 was taken is very good at \(-186^\circ\) C, but it shows no sensitivity whatever when the temperature is raised to \(-78^\circ\) C. The characteristics in Figs. 23 and 24 refer to elements whose oxidation process differed from that used for the preceding element. They have low sensitivity in the long-wave part, but the sensitivity in the \(3\,\mu\) region is less subject to the effect of heating to \(-78^\circ\) C. All these measurements were made at an illumination frequency of \(300\) cps.

Fig. 23. Influence of temperature on the spectral sensitivity of lead telluride photoresistors.

Fig. 24. Influence of temperature on the spectral sensitivity of lead telluride photoresistors.

From what has been said it is clear that much work still remains to be done on photoelements made of lead telluride before their properties become well known for their use.

References

1. B. T. Kolomiets, Izv. AN SSSR (phys. ser.), No. 5–6, 695 (1938).
2. F. C. Nix and A. W. Treptow, JOSA, 29, 457 (1939).
3. Fiz. Zap. AN USSR, 7, issue 3, 342 (1939).
   Yu. P. Maslakovets, ZhETF, 10, 393 (1940).
4. B. T. Kolomiets, ZhTF, 17, 195 (1947).
5. B. T. Kolomiets, Elektrichestvo, No. 5, 23 (1940); Fiz. Zap., 9, issue 2, 223 (1941).
6. Fiz. Zap., 7, 69, 1938; Fiz. Zap., 7, 345 (1939).
7. D. S. Geikhman and M. E. Soroka, Avtom. i telemekh., 3, 49 (1941); D. S. Geikhman, Sov. nauka, No. 2, 127 (1941); D. S. Geikhman, Author’s Certificate No. 5780c, NKPP.
8. P. G. Borzyak, ZhTF, 11, collected articles, 251; V. E. Lashkarev and K. M. Kosonogova, DAN, 54, 125 (1946); A. A. Il’ina, DAN, 54, 595 (1946); A. A. Il’ina, Izv. AN SSSR, phys. ser., 11, 436 (1947); A. V. Markov, ZhTF, 16, 1195 (1946); P. A. Ivanov and F. E. Sukhareva, Zavodsk. labor., No. 1, 114 (1946), et al.
9. E. G. Miselyuk and V. E. Kosenko, ZhTF. In press.
10. I. R. Potapenko, ZhTF. In press.
11. I. Ch. Bose, USA Patent No. 755840 (1904).
12. B. Lange, Phys. Zeits., 31, 964 (1930); Naturwiss., 19, 525 (1931).
13. M. Gruetzmacher, Swiss Patent 147290 (1931).
14. Fr. Fischer, B. Gudden und M. Treu, Phys. Zeits., 39, 127 (1938).
15. T. W. Case, Phys. Rev., 15, 289 (1920).
16. A. A. Chvikov, ZhTF, 8, 11 (1938).
17. British Patent No. 442228 (1934).
18. Optics in Military Affairs, ed. S. I. Vavilov and M. V. Savost’yanova, Moscow–Leningrad, 1945, p. 374.
19. R. L. Cashman, JOSA, 36, 356 (1946).
20. Charles L. Oxley, JOSA, 36, 356 (1946).
21. E. Lee and R. C. Parker, Nature, 158, No. 4015, 518 (1946).