PHOTOCONDUCTIVITY*

F. Nix, New York

Introduction

The influence of light on the passage of current through certain solids was discovered many decades ago, but especially important results were obtained only after the brilliant work of Gudden, Pohl, and their collaborators. These investigators gained a great deal by passing from the study of fine-crystalline semiconductors, which have comparatively high conductivities in the unilluminated state, to the study of single crystals of insulators. This made it possible for them to study the increase in conductivity upon illumination of a crystal with light of the corresponding wavelength under simpler and better-controlled conditions than had previously been possible. In many cases, by using weak light and low voltages, they were able to divide the phenomenon into two parts: “primary” and “secondary.” This differentiation of the phenomenon is of a fundamental character and will be set forth in detail in this article.

We shall begin with a consideration of the phenomenon singled out by Gudden and Pohl as “primary,” called the internal photoelectric effect, in distinction from the so-called external photoelectric effect (i.e., the tearing of electrons from the surface of a substance into the surrounding gas or vacuum under the action of incident light). We shall then proceed to consider the “secondary” phenomena; first we shall examine cases in which they coexist with the primary phenomena, and then cases in which they are observed separately. In conclusion we shall discuss the question of the origin in solids of an electromotive force upon their illumination by light. There is a very important classification of crystals, due to Gudden and Pohl. Some crystals possess natural photosensitivity in their natural state and in pure form; others owe their photosensitivity to the presence of impurities, or acquire it after irradiation with X-rays or after the action of other factors. The former

* Rev. of Modern Physics, 4, 327, 1932, translated by V. P. Zhuze.

are called “idiochromatic,” the latter “allochromatic” crystals. The refractive index (for light with a very large absorption coefficient in the given substance) is greater than 2 for the former and less than 2 for the latter.

Typical “idiochromatic” crystals are: diamond, zinc blende, cinnabar, stibnite, sulfur, and the red modification of selenium. Among allochromatic crystals one may name the alkali metals, silver, and the halide compounds of thallium.

Photoconductivity in insulators

Primary photoelectric currents in individual idiochromatic crystals of insulators.

The method of observing the photoelectric current arising when an individual crystal of an insulator is illuminated by light of the corresponding wavelengths is readily understood from Fig. 1, which represents the experimental arrangement. The quantity of electricity flowing through the crystal when it is illuminated for a short interval of time can be measured by a galvanometer or another suitable measuring instrument connected in series with a voltage source and the crystal.

Fig. 1. Experimental arrangement. The arrows indicate the directions of motion of negative charges in the crystal.

According to Gudden and Pohl ¹⁹ the fundamental equation of photoconductivity, which determines the observed quantity of electricity, has the form:

\[ Q=\frac{Ne\left(x_-+x_+\right)}{d}, \tag{1} \]

where \(N\) is the number of pairs of charges formed by the light, \(e\) is the charge of the electron; \(x_-\) and \(x_+\) are the free-path lengths of the charges, and \(d\) is the thickness of the crystal or the distance between the electrodes.

If all charges of each sign move in the direction of the corresponding electrodes, then equation (1) becomes

\[ Q_1=Ne. \tag{2} \]

In this case we would have to expect the fulfillment of Einstein’s law (the quantum equivalent). If we assume that charges of one sign remain immobile, while charges of the other sign freely move under the action of the applied electric field toward the corresponding

…to the collecting electrode, then the initial equation (1) takes the form:

\[ Q_2=\frac{Ne}{2}, \tag{3} \]

Obviously, if the charges are liberated uniformly throughout the entire volume of the crystal, then the average distance traversed by charges of one sign will be equal to \(x=\frac{d}{2}\).

Upon attainment of a stable (stationary) state, and under the condition of free displacement of charges of both signs, equation (1) assumes the form:

\[ i=\frac{ne(x_+ + x_-)}{d}, \tag{4} \]

where \(n\) is the number of pairs of charges formed per second; \(x_+\) and \(x_-\) are the distances traversed by the charges, respectively.

In many crystals possessing sufficient purity and a high degree of physical perfection, the experimentally observed currents obey the equations described above. It has been found that these currents obey the following laws: a) the current is strictly proportional to the intensity of the incident light at any voltage applied to the crystal; b) the curve of the current as a function of the applied voltage is at first almost linear, then bends downward and, finally, goes horizontally (with respect to the voltage axis), i.e., reaches saturation; c) the currents arise immediately at the beginning of illumination and cease at once after its termination.

Currents obeying these laws were designated by Gudden and Pohl as “primary” currents \(^{9,10,15,19,22,66}\). Similar currents were observed in diamond, zinc blende, cinnabar, and, to a lesser extent, in sulfur, red selenium, and stibnite. We shall now examine these laws in more detail.

a) Relation between photoelectric output and the intensity of the incident light. It has been found that the primary photoelectric current is strictly proportional to the intensity of the incident light \(^{22}\) for many substances investigated, under the conditions of a sufficiently low applied voltage, low light intensity, and short illumination time. The methods used to investigate this proportionality amount either to measuring the change in current over a short interval of time during abrupt changes in the intensity of illumination, as shown in Fig. 2, or to measuring the total quantity of electricity flowing through the crystal at constant light intensity but with different durations of illumination.

Figure 2 gives the results obtained on a separate thin crystal of zinc blende for two values of the voltage; 1270 V gives a current very close to saturation, whereas

like 280 V give a current considerably below the saturation current (Fig. 3). This proportionality was also observed in substances in which the saturation current was not obtained, as, for example, in sulfur, red selenium, and stibnite. In stibnite, for the region of the spectrum strongly absorbed by it, slight deviations from proportionality were found.

Fig. 2. Proportionality between the intensity of light incident on a zinc-blende crystal and the amount of electricity liberated. Area of the illuminated surface 25 mm², crystal thickness 1.3 mm, \(\lambda = 4360\) Å.

b) Relation between current strength and voltage. When an individual zinc-blende crystal is illuminated for a short interval of time with monochromatic light of constant intensity, the current thereby obtained varies as a function of the voltage applied to the crystal according to the law shown in Fig. 4 ²².

The absorption in the crystal of light of the two wavelengths indicated on the curve is very different. The data therefore show that the shape of the curve and the value of the voltage at which saturation is reached depend neither on the wavelength nor on the intensity of the light. On the other hand, the results obtained by the author in studying primary currents in individual crystals of mercuric iodide indicate that this assumption is not universally valid; on the contrary, the voltages causing the saturation current and the general form of the curves depend both on the wavelength and on the intensity of the incident light.

Flexin specially investigated the form of the volt-ampere characteristics for rock salt colored yellow by irradiation with X-rays, and showed that if the values of the quantity of electricity that has passed through the crystal are plotted as a function of the values of \(\frac{V}{d}\) (\(V\)—voltage, \(d\)—thickness of the cry-

steel), then curves are obtained that reveal the predicted saturation.

Good saturation was obtained in crystals of zinc blende and yellow rock salt, whereas for sulfur, red selenium, and stibnite no indication of the existence of saturation of the primary photoelectric current was obtained. One may think that for the last three substances too, current saturation could have been obtained if sufficiently high voltages had been applied.

Fig. 3. Proportionality between the duration of illumination and the quantity of electricity released in stibnite.

It was found that the more perfect the crystal, the sooner the saturation current is reached. This circumstance, together with other evidence to be given below, gave Gudden and Pohl grounds to accept that the electrons torn out by light do not pass freely through the crystal lattice,

Fig. 4. Volt-ampere characteristic for a zinc-blende crystal.

but in their motion are detained by imperfections of the lattice, by submicroscopic cracks, or else by impurity atoms included in the crystal lattice. The distance traversed by an electron before capture may be regarded as proportional to the voltage of the applied electric field. At small field voltages this distance in many crystals is small in comparison with the distance between the electrodes. As the voltage applied to the crystal is increased, the mean distance traversed by an electron increases until all the electrons liberated in the crystal reach the anode—at this moment the saturation current is attained. The application of higher voltages does not produce any further increase in the current. In the present argument we have assumed that the current passing through the crystal is very small and does not substantially affect the distribution of the potential gradient in the crystal.

In Gudden’s latest review \(^{67}\) there are cited unpublished results of the work of K. Hecht, who, in investigating (with the aid of a light probe) AgCl, obtained further confirmation of this point of view.

There are very few data to support this proposition in the case of diamond and zinc blende. Meanwhile, an analogous study of single crystals of the red modification of selenium, and also the data reported in the next section concerning yellow rock salt, give substantial proof that “capture” of electrons occurs after they have been previously liberated by heating or by illumination with infrared light.

Taking this reasoning into account, it will not be surprising that the saturation current has not been attained for imperfect crystals.

c) The instantaneousness of the occurrence of the primary photoelectric current. The connection of the phenomenon with time is obviously important for theory. Flexer \(^{31}\) devised a commutator apparatus that made it possible to act on the crystal with a “square” voltage wave, the beam of light being directed onto the crystal at an arbitrarily chosen moment of the cycle. The result showed that the negative component (to be considered in detail below) of the primary current arose within a time interval of less than \(10^{-4}\) sec after the beginning of illumination. Taking into account the work of Lawrence and Beams, who showed that in the case of the external photoelectric effect the appearance of free electrons begins in less than \(10^{-9}\) sec after the beginning of illumination, one may expect that the same also takes place in the case of the internal photoeffect.

“Components” of the primary photoelectric current

Up to now little has been said concerning the “components” of the primary photoelectric current. Gudden and Pohl \(^{10,44}\) consider...

...show that the primary process taking place upon the absorption of light in the crystal consists in the tearing out of electrons and their displacement toward the anode. The result of such a process will be that the positive charges left behind form a volume positive charge, which may lead to a distortion of the field in the crystal. From equation (4) it follows that the positive ions formed when electrons are torn out must be neutralized by electrons coming from the cathode in the event that the quantity of electricity registered by the galvanometer is equal to the total charge (of each sign) liberated in the form of ions by the light.

In some cases, in particular in the case of diamond 10, this effect was also observed separately. It was noted that when diamond is illuminated for a short time by light of short wavelength, close to the low-frequency edge of the absorption band, a certain quantity of electricity flows through the galvanometer; if, after this illumination, the crystal is illuminated with red light (which does not cause an increase of the photocurrent in a crystal not previously illuminated by light of short wavelength), then an additional quantity of electricity will pass through the circuit, which in the limiting case is equal to the quantity of electricity that passed through the crystal during the illumination by the light of short wavelength.

Fig. 5. Influence of “long-wavelength light” on the motion of the positive primary current in diamond. Crystal thickness 1 mm, illuminated surface area 3 mm\(^{-2}\), potential drop across the crystal 1200 V. \(a\)—illumination with light of \(\lambda = 3130\) Å \(\left(2.16\cdot 10^{-6}\ \frac{\mathrm{cal}}{\mathrm{sec}/\mathrm{cm}^2}\right)\); \(b\)—illumination only with “long-wavelength light” \(\left(5\cdot 10^{-2}\ \frac{\mathrm{cal}}{\mathrm{sec}/\mathrm{cm}^2}\right)\) after illumination with light of \(\lambda = 3130\) Å; \(c\)—simultaneous illumination with “long-wavelength light” and with \(\lambda = 3130\) Å.

The charge flowing through the galvanometer during illumination of the crystal with light of short wavelength was called by Gudden and Pohl the “negative part of the primary current”; the charge flowing during the subsequent illumination with red light they called the “positive part of the primary photocurrent.” These

to some extent misleading; whereas the first current is probably due to the drawing of electrons from the crystal to the anode, the latter current cannot simply be ascribed to the motion of positive ions. Instead, we must picture the process as follows: the electrons emerging from the cathode gradually spread through the crystal and neutralize the positive charges left in the crystal after the advance, under the action of light, of the electrons toward the anode.

Fig. 6. (a) Distribution of potential: \(a\)—before illumination, \(b\)—after illumination.
(b) Reverse current in zinc blende upon illumination with light of \(\lambda = 4360\) Å \(\left(4 \cdot 10^{-6}\ \dfrac{\mathrm{cal}}{\mathrm{sec}/\mathrm{cm}^{2}}\right)\), without an applied field. The crystal was previously illuminated in a field with light of \(\lambda = 4360\) Å.

In some experiments by Gudden and Pohl, the separation of these two components was almost complete. Fig. 6 shows results obtained with diamond\(^{19}\). Curve \(a\) represents the current passing through the galvanometer during illumination with light of short wavelength, while \(b\) corresponds to the total current passing during the subsequent illumination with light of longer wavelength; \(c\) gives the total current passing through the crystal when it is illuminated simultaneously with light of both long and short wavelength.

The study of zinc blende showed that these two charges are not equal in magnitude; on this basis it was concluded that part of the positive ions was neutralized during the first illumination, or, as Gudden and Pohl thought, part of the positive current flowed simultaneously with the negative current. In some cases, for example in sulfur\(^{55}\), it is not possible to separate these two components of the primary photocurrent.

The positive component is best separated by using long-wavelength light or heat (the same agents were used by Lenard to restore his phosphor to its initial conditions). The most effective light is the long-wavelength end of that spectral region to which the crystal is sensitive in its initial, or unexcited, state. In many of the earliest works of Gudden and Pohl they used

attributed this effect to red light, but in some cases it should rather be ascribed to the blue end of the spectrum.

In order to obtain reproducible results in the study of crystals, it is absolutely necessary, before repeating the measurements, to illuminate the crystal with red light, heat it, or leave it for some time.

Neglect of these precautions is responsible for contradictory data in the case of the study of diamond.^17 In blue rock salt, illumination with light of long wavelength is insufficient to restore the initial conditions, but this is possible and not surprising, taking into account the exceptional complexity of the absorption curve.^45,48 With an increase in the magnitude of the photoelectric current, volume charges of considerable magnitude may be observed.^10 For the most part, the disturbance of the potential distribution is caused precisely by a positive volume charge; however, Tartakovsky’s data^65 indicate a similar action of a negative volume charge in the case of yellow rock salt. Fig. 6a shows the potential distribution in a zinc-blende crystal before and after illumination. Further proof of the existence of an oppositely directed internal field, caused by the formation of space charges, is given in Fig. 6b.

This curve was obtained when current was passed through the crystal for a time sufficient to produce a large polarization, after which the applied potential was removed, with subsequent illumination, as before.

The current decreases with time until the internal field disappears.

Spectral distribution of the sensitivity of idiochromatic crystals

Crystals belonging to this class possess photoelectric sensitivity throughout the entire region of the absorbed light, with the exception of the far infrared rays. The form of the curve of the spectral distribution of sensitivity is represented, for example, by the curve in Fig. 7a for red selenium.^34 It is characterized by the existence of a maximum followed by a gradual fall toward long wavelengths and a steep decline toward short wavelengths.

The maximum is localized not in the middle of the absorption band, but is shifted toward the shorter wavelengths (absorption measurements are lacking for red selenium, but this fact is shown in Fig. 8 for a pure diamond crystal). In many cases (for example, diamond, zinc blende, and cinnabar) the maximum lies in the region of long wavelengths, for which the absorption coefficient is equal to about \(1\ \mathrm{m}^{-1}\). The position of the short-wavelength maximum in the curves of spectral sensitivity for molybdenite

\((\mathrm{MoS}_2)\) is in the same position with respect to the absorption band as in the case of diamond and similar materials (molybdenite, as I shall indicate in the next section, is a semiconductor in which the photocurrent is classified as secondary).

All the explanations proposed up to now to account for this discrepancy in the spectral-sensitivity curve for regions of strong optical absorption have been unsuccessful. All the relations described up to now for the primary current cease to be valid when considerable impurities are present in the crystal, evidently because the absorption of light occurs without the formation of photoelectrons. In the case of the red nonconducting modification of selenium, which shows no appreciable dark current, the position of the maximum of the spectral sensitivity depends on the orientation of the crystal with respect to the incident beam of light. This dependence on orientation is due to the dichroic absorption of the crystal^34 (cf. Fig. 7, a and 7 b). For these primary currents the shape and position of the spectral-sensitivity curves do not depend on the voltage and intensity of the light.

Fig. 7. Spectral distribution of the primary photoelectric current in a single crystal of red selenium.

Nor do they change in the case when the direction of the light beam changes from perpendicular to the lines of force of the electric field (“transverse” illumination) to parallel with respect to them (“longitudinal” illumination).

The absorption of light that gives rise to photoelectric currents causes a change in the optical properties of the crystal^46,72. This is manifested in a broadening of the absorption region and an increase in the absorption coefficients, which can be detected by optical and photoelectric measurements from the long-wavelength side of the absorption band.

Initial conditions can be restored by illuminating the crystal with light of long wavelengths or by heating, or they are restored by themselves after some time.

Quantum relations.

A number of very fruitful investigations by Gudden and Pohl were undertaken with the aim of testing Einstein’s law on quantum equivalent.

If each absorbed photon liberates one electron, then the “yield” would be represented by the formula:

\[ n=\frac{Q}{h\nu}=\left(\frac{Qc}{h}\right)\lambda, \]

where \(n\) is the number of electrons liberated by the light energy \(Q\) with frequency \(\nu\); \(h\) is Planck’s constant. This expression is a linear function of \(\lambda\).

In numerous investigations of the external photoelectric effect, carried out with the same intention, it was shown that the photoelectric yield was many times smaller than the theoretical value, although in Zuurman’s recent work with thin layers of alkali metals “yields” were found reaching many percent of \(\frac{Q}{h\nu}\).

Fig. 8. Law of quantum equivalence for diamond.

The first quantitative test of the equivalence law (Einstein’s) for the photoelectric effect was carried out on individual crystals of zinc blende and diamond, and qualitatively on cinnabar. Warburg’s photochemical investigations had shown even earlier that a quantum-equivalent yield was justified for the reaction of photochemical decomposition of HI and HBr.

Fig. 8 shows the results of studying a transparent crystal of diamond[^24]. In obtaining the data, a voltage sufficient to produce saturation was applied. Abso-

The absolute values of the intensity of the incident light were measured with the aid of a linear thermopile, and from them the values of the absorbed energy were calculated, using the data of absorption measurements carried out by Peters.^25

As can be seen, between the values of the quantum yield $\left(\dfrac{\text{coulomb}}{\text{cal}}\right)$ and the wavelength of the absorbed light, in the region of weak absorption of light, there exists a strict proportionality. These results are the first quantitative proof of the correctness of Einstein’s law. Each quantum of absorbed light energy tears out a negative and a positive elementary charge ($-e$ or $+e$), or, in other words, for each absorbed quantum of light one more electron begins to move through the circuit. As explained on p. 391, in order to obtain the correct result one must measure and add together both components of the primary current. An obvious violation of the law of equivalence was observed for sulfur^55 and rock salt (previously colored yellow by X-rays). These measurements, however, were made with voltages much smaller than the saturation voltages, and therefore it is not surprising that the photoelectric current was considerably less than the theoretical value, since some of the electrons did not reach the electrodes.

Subsequent measurements of the current-voltage characteristics^59 for yellow rock salt showed that, when sufficiently high voltages were used, saturation was attained, and it seems reasonable to suppose that if the measurements had been made with a voltage sufficient to produce the saturation current, Einstein’s law would have been valid in this case as well. The quantum yield in yellow rock salt increases with increasing wavelength in full agreement with the requirements of Einstein’s law; however, the number of electrons per photon of absorbed light was considerably below the required value.^38

Influence of Temperature on the Primary Current

The influence of temperature on the primary photoelectric current in diamond, zinc blende, and yellow rock salt was investigated over wide temperature ranges, down to the temperature of liquid hydrogen. An early work by Lenz^39 (Fig. 9) noted a considerable influence of temperature on the magnitude of the stationary current, which was mistakenly explained by a decrease in the negative primary current. Gudden and Pohl^140 proposed an explanation based on the influence of the space charge of positive ions; later this was confirmed by Lenz.^57 This sharply expressed temperature effect is in fact due to the formation of a positive space charge, owing to

to a slowing of the motion of positive ions at these low temperatures.

The space charge formed at low temperatures reached such a magnitude that it fully compensated the applied voltage. Lehne did not properly consider the narrowing and shift of the absorption band toward shorter wavelengths. In a recent work carried out in Pohl’s laboratory, Lehne^57 was able to show that, on proper consideration, the change in absorption and the formation of space charges disturbing the potential distribution “does not depend on the initial negative part of the primary current down to temperatures of \(-250^\circ\text{C}\).” It is also still unknown whether the current-voltage characteristics change with temperature. Measurements of spectral sensitivity carried out with crystals of yellow rock salt indicate that a change in the photoelectric current occurs only in the case of a change in optical absorption. Lehne^39 observed that when electrons are introduced into the crystal lattice by means of an electron gun at various temperatures (down to \(-250^\circ\text{C}\)), the current does not depend on temperature. Gudden^185 considered this phenomenon and asserted that the positive part of the primary current is in fact due to a displacement of the position of positive charges.

Fig. 9. Change of the photoelectric current with temperature.

Influence of a magnetic field on the primary current

In 1916, Lukirsky published the results of an investigation of the Hall effect in yellow rock salt, which indicated that the electric current arising in it upon illumination with light of a suitable wavelength is electronic in its nature. Later, somewhat more extensive experiments for studying the influence of a magnetic field on the internal photoelectric effect were carried out by Lehne^30 and Arsen’eva^41. Lehne’s measurements on zinc blende and diamond showed that ma-

The magnetic field does not affect the photoelectric effect. In the case of diamond, the Hall effect was proportional to the magnetic-field voltage and increased proportionally to the electric-field voltage (at small voltages), reaching a saturation value at \(2500\ \mathrm{V/cm}\).

The sign of the Hall effect changed in the normal way when the direction of the magnetic field was changed, and corresponded to negatively charged carriers.

The data for zinc blende were not so consistent. However, when the trigonal axis was placed parallel to the lines of the electric field, the normal behavior of the Hall effect was observed. In other crystallographic directions the sign of the Hall effect did not change when the direction of the magnetic field was changed.

Measurements of the Hall effect made with currents obtained by introducing electrons into the crystal lattice by means of an electron gun proved to be identical with the results obtained in the study of photoelectric currents. The Hall effect, as was found, does not depend on temperature. As Gudden pointed out,\(^{183}\) these investigations are of interest for the study of the photoelectric phenomenon chiefly because they provide additional proof that the primary current is electronic.

Secondary currents

Gudden and Pohl showed that the phenomenon of the internal photoeffect ceases to obey the laws established above when the voltage or the light intensity or, finally, the duration of illumination exceeds certain values. The effect ceases to be proportional to the light intensity (or to the time of illumination), the curves of the spectral distribution of sensitivity depend on the voltage and, finally, with increasing voltage the photoelectric currents reach saturation. Gudden and Pohl \(^{9}\) attributed these deviations from simple regularities to the presence of a so-called “secondary current.”

If, in particular, we take the ratio of the magnitude of the liberated charge \(Q\) to the illumination time \(t\), then in the most general case we may write

\[ Q = at + bt^{2} + ct^{3} + \ldots . \]

If on the right only the first term is present \((Q = at)\), then the current is called “primary.” The presence of other terms also indicates the existence of “secondary currents.”\(^{9}\)

Fig. 10 depicts the influence of voltage on the curves “illumination time—quantity of electricity” for a single ZnS crystal and two wavelengths.

For a potential up to \(1000\ \mathrm{V}\) the quantity of electricity was

proportional to the time of illumination; for higher voltages it increases more rapidly. Fig. 10b is given as an example for the case of strongly absorbed light; here, at a potential of 2000 V, a clear deviation from proportionality between charge and time is observed, although for longer periods of illumination this proportionality again sets in. Fig. 11b shows the changes of the primary and secondary currents with the intensity of the light. The values of the currents are recalculated in Fig. 11a. As was explained in the preceding paragraph, the primary current is given by the first term in the equation given above, whereas the secondary current is given by the second term. The form of the curves of spectral sensitivity for the secondary currents depends on the applied voltage (Figs. 12, 13). For an individual crystal of cinnabar[^23] (Fig. 12), for voltages up to 100 V the currents appear purely primary; additional secondary components appear as the voltage is raised and, at 200 V, the secondary current seems to predominate. In Fig. 13 are shown curves of spectral sensitivity in the coordinates: current—wavelength of the incident light.

Fig. 10. Dependence of the quantity of electricity on the duration of illumination.

(a) Crystal of zinc blende, 1 mm thick; area of the illuminated surface 1 cm². Illumination by light with \(\lambda = 4050\ \text{Å}\) of intensity \(0.264 \cdot 10^{-7}\ \dfrac{\text{cal}}{\text{s}/\text{cm}^2}\). (b) \(3130\ \text{Å}\), intensity \(12.5 \cdot 10^{-7}\ \dfrac{\text{cal}}{\text{s}/\text{cm}^2}\).

These curves show quite clearly that the secondary current increases when the voltage is raised and becomes predominant at higher voltages. In the case of polycrystalline ZnS, in which only secondary currents are observed, the form of the spectral-sensitivity curves depends both on the time of illumination and on the applied voltage. From spectral measurements on substances in which only secondary currents are observed, it follows that the spectral-sensitivity curves can in no case be regarded as physical constants of the substance, whereas for substances in which primary currents are observed they may be regarded as such. The photoelectric yield for primary currents is proportional to the intensity of the light and to the time

of illumination, whereas for secondary currents, as is seen from Figs. 14 and 15, the currents increase at first approximately proportionally to time, and then, as the illumination time is increased, gradual saturation sets in. The form of the curves representing the dependence of the photoelectric yield on the illumination time depends on the previous history of the material. Fig. 14 also shows curves obtained in the study of secondary currents in greenockite.² The lower curve was obtained first; then the middle and upper curves were taken successively.

Fig. 11. Separation of the primary and secondary currents in a cuprous-oxide crystal.
(a) Change in the quantity of electricity for different durations of illumination and different light intensities.
(b) \(I_p\) — primary current, \(I_s\) — secondary current as a function of light intensity.

What is striking is the exceptional similarity between these curves and the curves obtained for commercial selenium and thallium-sulfide cells. We obtained similar curves for individual crystals of metallic selenium. Fig. 15 shows how the current slowly increases with time, finally reaches a limiting value, and then slowly decreases when the illumination is stopped.³⁴

The secondary current depends both on the method of illumination (transverse or longitudinal) and on the wavelength of the light.²³ This is shown in Fig. 16. This could have been expected because of the strong absorption in the surface layers, so that the secondary currents will be especially large for the cases shown in Fig. 16, A and C; 16, A and D correspond to cases of strong absorption of light for

one method of illumination, while 16 C and D correspond to cases of an average magnitude of light absorption. It is assumed that the secondary current is the result of the primary current. (The reader should recall that the ejection of an electron upon absorption of a photon gives rise to the negative part of the primary current, accompanied by the motion of positive charges.)

It is often accompanied by a permanent decrease in the resistance of the crystal. In their earlier works Gudden and Pohl\(^9\) were inclined to regard the secondary current as electrolytic in nature. However, judging from their later works, they abandoned this view. In the author’s opinion, the existing data are insufficient for developing a theory of secondary currents. It often happens in the case of the conductivity of dielectrics that actual transport and deposition of matter must be observed before a current can be described as electrolytic in nature.

Fig. 12. Development of the secondary current with increasing applied voltage in a single crystal of cinnabar. The scale along the ordinates is different for curves a, b, and c. Crystal thickness 1.35 mm.

Fig. 13. Results shown in Fig. 12, plotted on one scale for different applied voltages.

The secondary current was observed in diamond. In this case, of course,

it is difficult to imagine the existence of ionic conductivity. An analogy suggests itself between the secondary current in the phenomenon of photoconductivity and that part of the current in a self-sustained glow discharge which consists of ions formed spontaneously in the gas. In any case, it seems preferable to regard secondary currents as currents caused by absorbed light, but not in a direct, rather in an indirect, way. In selenium, polycrystalline ZnS, HgS, MoS₂, Ag₂S, CdS, and others, the observed currents are secondary, while the primary ones are masked by them. Gudden and Pohl often asserted that intercrystalline boundaries hinder the passage of electronic current, but facilitate the passage of electrolytic current (it is well known that they behave in this way in ordinary electrolytic conductors). This, however, does not explain why primary currents are not observed in individual crystals of gray metallic selenium. The currents in this modification of selenium have the same properties in both the mono- and polycrystalline states. We shall consider semiconductors in the next section and shall try to show certain features of similarity between photocurrents in them and secondary currents in good insulators.

Fig. 14. Photoelectric current as a function of the duration of illumination for proustite.

Fig. 15. Photocurrent as a function of the duration of illumination for selenium and polycrystalline ZnS.

The external similarity is so sharply expressed that it compels one to think that the currents observed up to the present time in semiconductors are in reality secondary currents, considered in this section.

Photoconductivity in allochromatic crystals

Joffé and Röntgen13 observed that sunlight exerts a considerable influence on the change in the conductivity of rock salt under the action of X-rays. It was then established that, when various halide compounds of the alkali metals, such as, for example, rock salt or sylvine, are irradiated by Röntgen rays and subsequently illuminated by rays of the visible spectrum, a sharp increase in conductivity is observed. This effect was found to correspond to another: the coloration of a crystal by Röntgen rays or under the action of other agents.

Fig. 16. Distribution of absorption centers. A and B—strong absorption; C and D—weak absorption of light.

The crystals assumed a yellow coloration under moderate illumination by Röntgen rays at room temperature. The same coloration, according to the observations of F. Giesel and Zidentopf49, was observed when transparent crystals of rock salt were treated in sodium vapor. This is accomplished by placing the crystal in a heated closed vessel containing sodium vapor. The diffusion of the vapor into the crystal lattice is accompanied by coloration of the crystal. Goldstein observed that a similar coloration of transparent rock-salt crystals is also produced by ultraviolet rays. Since then it has been found that crystals of halide compounds of alkali metals are colored under the action of any of the above-mentioned agents, i.e., when they are illuminated by ultraviolet rays, by Röntgen rays, and also when treated in the vapor of the element that forms the cations of the salt under investigation; in this case such crystals become photosensitive and possess identical optical and photoelectric properties47. Such crystals were called (as I have already noted above) crystals of the second kind, or allochromatic crystals.

Coloration is attributed to the presence in them of so-called “centers,” formed under the action of one factor or another. It seems most probable that these centers are neutral metallic particles, which in the case of yellow rock salt may be individual sodium atoms. The basis for assuming this degree of dispersion is the fact established above, namely that the photoelectric and optical properties of colored crystals are completely identical regardless of the method of their coloration.

It should also be noted that, for the halide compounds of the alkali metals, the energy corresponding to the first characteristic absorption band in the far ultraviolet region agrees well with the theoretical value of the energy required to transfer an electron from the anion to the cation $^{87,71}$. Silver halides do not possess sufficiently sharp characteristic frequencies for such a calculation to be possible. It is well known that bleaching of these coloration centers is accompanied by the disappearance of their absorption band. The absorption spectrum of a rock-salt crystal after it has been bleached is identical with the absorption spectrum of the original colorless rock salt, i.e., the crystal regains its original properties. This complete disappearance, under illumination by light, of the absorption band caused by active centers is not observed for silver and thallium halide compounds. We shall return to this question later.

The optical (absorption) and photoelectric properties of substances depend on the degree of coagulation of neutral metallic atoms. This can be well shown by the example of yellow rock salt, containing individual sodium atoms dispersed in the crystal, and of blue rock salt, in which the sodium atoms are gathered into large groups. The dimensions of the particles therefore vary from the dimensions of individual atoms to atomic groups of colloidal size. Up to now, investigations have been limited to the two extreme cases indicated above (yellow and blue rock salt).

In the first case a simple relation was found between optical absorption and photoelectric properties; for blue rock salt no simple regularity was discovered $^{45}$. In the following section we shall return in greater detail to the study of these centers and their relation to the latent photographic image. Fig. 17 shows the absorption spectrum and the photoelectric emission for a rock-salt crystal after illumination by X-rays $^{88}$.

The solid curve $E$ represents the photoelectric emission, while the dotted curve $a$ represents the change in the absorption constant as a function of wavelength.

One should note the striking difference between the ratio of optical absorption and photoelectric spectral sensitivity for these allochromatic crystals and the same ratio for the idiochromatic crystals described on the preceding pages.

In idiochromatic crystals each spectral-sensitivity curve has a maximum located in the absorption band on the long-wavelength side, and falls to an immeasurably small value within the band; whereas the spectral-sensitivity curves for allochromatic crystals coincide with the optical-absorption curves. The curve of the spectral distribution of photoelectric sensitivity shown in Fig. 17 is slightly shifted, relative to the other, toward longer wavelengths, as might have been expected on the basis of quantum relations. (Arsen’eva,^41 on the other hand, did not find such a shift, but noted complete coincidence of the optical-absorption curves with the photoelectric-sensitivity curves. Gudden^186 thought that this disagreement was due to the fact that Arsen’eva did not take into account the absorption of light by colloidal particles.)

Fig. 17. Curves of the spectral distribution;
\(E\)—photocurrent, \(A\)—absorption coefficient
in yellow rock salt.

In yellow rock salt, when illuminated with suitable light, a primary current arises, the initial values of which are proportional to the intensity of the incident light. Then the photocurrent gradually decreases because of the formation of a positive space charge. When the entire crystal is illuminated completely, a certain steady value of the primary current is reached; when only a small part of the crystal lying between two electrodes is illuminated, this current gradually decreases to an immeasurably small value. The negative part of the primary current arises, as in the case of idiochromatic crystals, less

than \(10^{-4}\) sec after the beginning of illumination.^31 The primary current at first increases proportionally to the potential difference applied to the crystal, and then saturation sets in, quite analogously to what we had in the case of dichromatic crystals. The fields required to obtain the saturation current in rock salt illuminated by X-rays are much higher than for zinc blende or diamond. Indeed, up to 1928 such saturation had not been obtained in it; in that year Flexig^50 obtained saturation by using exceptionally thin plates cut along the cleavage plane from a crystal of rock salt (with a thickness of \(0.1\) mm).

Fig. 18. Spectral distribution of the photocurrent; \(e\) — absorption coefficient; \(K\) — in blue rock salt.

Earlier investigations of the quantum yield (quantum efficiency) in rock salt (Dzhulai^38) gave for the yield a value \(= \sim \frac{1}{3000}\) of the equivalent quantity, but since these measurements were made at voltages at which saturation had not yet been reached, it was impossible to decide definitely whether the electrons traverse only \(\frac{1}{3000}\) (the mean value) of the entire thickness of the crystal, or whether only a part of the absorbed light goes into ejecting electrons. The saturation voltage found by Flexig agreed well with the value that could be expected from Dzhulai’s measurements,* so that this may be regarded as another verification of Einstein’s law. This result permits all the processes studied so far to be

* Dzhulai assumed that the electrons advance, on the average, by \(\frac{1}{3000}\) of the thickness of the crystal. The extrapolated value of the saturation voltage agrees with the value of the saturation voltage found by Flexig.

PHOTOCONDUCTIVITY

to combine under one general scheme that does not take into account the nature of the active centers. Experiments published by Podashevsky\(^{62,64}\) showed that the photocurrent decreases with increasing plastic deformation. This phenomenon is accompanied by a shift of the position of the maximum and of the long-wave part of the spectral-sensitivity curve toward longer wavelengths. If the deforming load is removed, then the distribution of spectral sensitivity gradually, over several days, returns to its former state. For rock salt with incorporated sodium particles of colloidal dimensions there is no longer a simple relation between optical absorption and photosensitivity. Fig. 18 depicts the complete absorption curve and the curve of photoelectric sensitivity for blue rock salt\(^{45}\) (containing submicroscopic sodium particles). Gudden and Pohl suggest that the absorption of light leads not only to the tearing away of valence electrons, but also to heating of the crystal. This, it seems, may be considered a plausible explanation of the absence of complete agreement between the absorption spectrum and the curve of photoelectric sensitivity.

Fig. 19. Photocurrent in rock salt at 33° C and 77° C. Low light intensity.

The similarity between the curves of spectral sensitivity for the external photoeffect for metallic sodium and the curves for the internal photoeffect for blue rock salt is so striking that it leads to the idea of a connection between these two curves\(^{51}\). It was found that if, after the first illumination of the crystal, it is illuminated with light of longer wavelengths, then no additional flow of charges is observed.

The various stages in the passage of primary currents in crystals of colored rock salt are very clearly illustrated by measurements carried out at different temperatures and light intensities, as shown in Figs. 19 and 20\(^{85}\). These measure-

... were made at three different temperatures and for two illumination intensities. The segment of the curve \(bc\) in Fig. 19, \(a\), corresponds to a sudden jump of the negative part of the primary current at the beginning of illumination, after which a current flows through the crystal. The slight decrease of the current is connected with the slow formation of a positive space charge. Toward the end of the illumination period the current falls to a new value \(f\). In the dark, a current \(flc\) also flows through the crystal; this is a positive primary current. Under illumination by infrared rays, the greater part of the positive charges is displaced toward the cathode. The obliquely hatched area corresponds to the flow of this current, which can be observed for a considerable time at this temperature. Fig. 20, \(b\), represents similar measurements, but carried out at a temperature of \(77^\circ\) C. The current at the beginning of illumination reaches approximately the same value as in the preceding case \((33^\circ\) C). It is noteworthy that the current gradually increases as a result of the motion of trapped electrons and the freer motion of positive ions, which form the primary positive current. The behavior of the crystal during the subsequent periods of darkness and illumination by infrared rays coincides with the behavior of the crystal at \(33^\circ\) C. Fig. 20, \(a\), is undoubtedly more complicated. The curve \(cdc\) depicts an instantaneous increase of the primary negative current followed by a decrease and is, in its general character, somewhat similar to the curve for the experiment at \(77^\circ\) C; the fundamental significance belongs to the quantity \(d\), which reaches a current.

Fig. 20. Photocurrent in rock salt; \(a\)—at \(125^\circ\) C (low illumination intensity), \(b\)—at \(77^\circ\) C (high illumination intensity).

The small magnitude of the current in the subsequent periods of darkness and illumination by infrared rays shows quite clearly that the greater part of the positive primary current flows simultaneously with the negative one. The area hatched with vertical strokes corresponds to the electrolytic conductivity of rock salt at this elevated temperature. Fig. 20, \(b\), corresponds to the conditions of Fig. 19, \(b\), but only for a considerably

of greater light intensity. The negative primary current rapidly decreases with increasing illumination time (curve \(ce\)). These curves show quite clearly that, in studying the primary photoelectric phenomenon, one should use only the initial value of the primary current.

Excitation of the crystal and the positive primary photoelectric current

Röntgen and Ioffe \(^{13}\), in their early investigations, found that the curves of the spectral distribution of sensitivity of yellow rock salt, colored by illumination with Röntgen rays, depend on the order in which the measurements are made. When measurements are made successively with ever longer waves, the maximum of the curve is lower and at the same time more strongly stretched out toward the red rays than when the measurements are made with successively decreasing wavelengths. This special phenomenon can be explained on the basis of the data of Gudden and Pohl. The curves constructed from measurements of the absorption of light in yellow rock salt change in exactly the same way as the curves of spectral sensitivity according to the data of Röntgen and Ioffe.

Fig. 21. Spectral distribution of photocurrent and absorption in rock salt.

Curve 1 in Fig. 21,a represents the absorption curve for a rock-salt crystal, constructed on the basis of a series of measurements at different wavelengths, with illumination of the crystal by infrared light in the interval between each pair of measurements \(^{37}\). It is assumed that the infrared light returns the crystal to the so-called “normal state.” When, however, the measurements are made at the indicated wavelength with a period of illumination corresponding to a frequency close to the high-frequency end of the induced absorption band, curve 2 of the above figure is obtained.

Curves 1 and 2 in Fig. 21,b, referring instead of the absorption coefficient to the photoelectric effect, were obtained in the same way. The change in the state of the crystal under the action of visible or ultraviolet light, manifested in a change-

absorption and photoelectric yield, is called “excitation”; the crystal, as they say, is in an excited state.

Crystal imperfections and the heating of crystals affect light absorption in most cases in the same way as excitation. Flechsig^43 found that synthetic crystals of rock salt possess a much broader and flatter absorption band than natural crystals.

With decreasing temperature the absorption band becomes narrower and the maximum shifts toward shorter wavelengths. The magnitude of the excitation, measured by the difference of the absorption coefficients from curves 1 and 2 (Fig. 21,a), is also a function of temperature—the lower \(t\), the greater the attainable excitation. Excitation has been observed in a whole series of halide compounds of the alkaline metals; in the halide compounds of silver and thallium it is not observed^n. The investigation of the effect of excitation on absorption in dichromatic crystals has been carried out only with diamond, in the long-wavelength part of the absorption band, and even then very incompletely. The absorption of this spectral region increases upon excitation according to the same law as in the case of the halide compounds of the alkali metals. Measurements inside the absorption band are hampered by experimental difficulties. In the red isolating modification of selenium, excitation increases the sensitivity (Fig. 22) to long wavelengths, similarly to what we have in the case of yellow rock salt; in the remaining part of the curve of spectral sensitivity no noticeable change is found^34.

Fig. 22. Spectral distribution of the photocurrent in the red insulating modification of selenium; \(A\)—normal state, \(B\)—“excited” state.

The crystals were too small to permit measurements of absorption; nevertheless it may be supposed that the absorption would have changed, in the main, in the same manner. The results of these investigations can, very probably, be regarded as an illustration of the difference in the character of excitation for two classes of crystals. In numerous investigations carried out in recent years in Pohl’s laboratory, views were maintained which assigned the chief role to centers in the crystals themselves, rather than the view according to which normal atoms or ions may be regarded as “nests” where photoelectric processes take place. My opinion is conclu-

Photoconductivity

... consists in the fact that there is very little direct evidence confirming this point of view; however, some observations indicate that this is so. A further difficulty consists in understanding that there is no well-defined photoelectric threshold for the onset of the ejection of an electron by light inside a crystal, if one regards normal atoms or ions of the space lattice as the places where this process occurs. Differences in the form of the absorption band observed for crystals of different degrees of imperfection point to the circumstance that the centers depend directly on the physical state of the crystal. This view of Pohl and his co-workers agrees very closely with the theory proposed by Smekal to explain electrolytic conductivity in salts, in which, at low temperatures, he ascribes the greater part of the electrical conductivity to the so-called “Lockerionen,” located at defective places in the lattice. Zwicky postulated the existence of a mosaic structure in crystals, caused by the existence within the crystal of regions of different density. Further consideration of such theories would take us too far afield.

The excitation, one may think, arises because of a disturbance of the internal structure of the crystal, caused by the ejection of electrons by the absorbed light[^44]. When light ejects an electron from an active center, then, under the action of the applied electric field, the electron begins to move through the crystal lattice; if there is no field, then the electron may move some distance away from the active center owing to the energy received from the absorbed photon. When the crystal is sufficiently free from inhomogeneities (from imperfect places), then at high voltages a saturation current is finally reached, as was described in the preceding section, and the electrons leave the crystals. The centers thereby become positively charged. The absorption band should then shift toward shorter wavelengths, if one imagines that the centers remain positively charged. Taking into account the fact that the absorption region does not shift in this expected direction, they suppose that the centers lose their positive charge, returning, more or less, to their original state. Gudden and Pohl believe that the loss of positive charge by a center consists in the capture of an electron from a neighboring ion or atom. These processes then continue until the centers enter the sphere of strong influence of the positive charges lying in their immediate neighborhood, and thus acquire the possibility of absorbing light of a lower frequency than had previously been possible.

This should manifest itself in a broadening and increase of absorption in the region of adsorption from the side of the longer wavelengths. For the time being...

the process of tearing out electrons continues, followed by their capture by neighboring centers; the volume in which this capture occurs undergoes changes similar to recrystallization; the remaining positive ions thereby receive electrons from another group closer to the cathode and are thus neutralized. This process probably continues until the entire nonequilibrium positive charge remaining after the departure of the electrons has been neutralized by electrons—a process that manifests itself in the form of a positive dark current.

Fig. 23. Absorption spectrum of color centers in AgCl; A—normal absorption curve; B—absorption after illumination with intense light of \(\lambda = 5500\ \text{Å}\).

This process can be accelerated, as has already been indicated in the preceding sections, by means of light of long wavelengths or by heating. The lower the temperature, the more slowly it proceeds. Much remains to be done to clarify the role of long-wave light in this matter.

Color centers in allochromatic crystals. Photographic latent image

Allochromatic crystals can be made photosensitive by several of the processes considered in the section devoted to the general photoelectric properties of such crystals. Of these various processes (illumination with X-rays or ultraviolet rays, holding the crystal in vapors of the metal that forms the cation of the crystal), for carrying out quantitative measurements and producing such centers, illumination with ultraviolet light is the most convenient. Any absorbed light, with the exception of residual rays and infrared rays, increases the number of these centers, whose properties do not depend on the wavelength of the light used. Halide compounds of the alkali metals, in general, possess characteristic absorption bands lying in the far ultraviolet region of the spectrum, whereas color centers absorb light in the visible part of the spectrum; their

absorption bands have a simple bell-shaped form. The centers disappear when the crystal is illuminated by light which they absorb. The absorption of a bleached crystal is limited by the initial characteristic absorption in the ultraviolet part of the spectrum. The process of bleaching is irreversible, in contrast to reversible excitation. The two processes are often encountered together, and the relative predominance of one or the other is a function of temperature. For the halide compounds of the alkali metals, light of any frequency lying

Fig. 24. Absorption spectrum of AgCl; A — characteristic absorption band, absorption coefficient of the order of \(10^{-5}\ \mathrm{mm}^{-1}\); B — induced absorption band, maximum value of the absorption coefficient of the order of \(0.1\ \mathrm{mm}^{-1}\).

within the absorption band of the introduced centers is capable of bleaching all the centers; in the case of silver halides, light of any given frequency bleaches only part of the centers.

This is best illustrated in Fig. 23. The solid line A represents the absorption of the coloring centers1. After illumination with an intense beam of light with wavelength \(5500\ \text{Å}\), the absorption of AgBr changes considerably in the region close to this wavelength, without changing noticeably in other parts of the absorption band. This is shown by the dashed curve B. Bleaching of introduced centers in some silver halide compounds proceeds with great difficulty because of the partial coincidence of the region of characteristic absorption of the uncolored crystal with the absorption region of the coloring centers, so that light of one and the same composition simultaneously bleaches and colors the substance. Fig. 24 illustrates such a partial coinc—

overlap of the two absorption regions for silver chloride[^76]. Silver halide compounds, in contrast to halide compounds of the alkali metals, do not possess a sharply expressed characteristic absorption maximum. The formation of centers under the action of light consists in the transfer of an electron from the negative to the positive ion in the crystal lattice. This may be represented by the following scheme:

\[ \mathrm{M}^{+}+\mathrm{H}^{-}\to \mathrm{M}+\mathrm{H}. \]

This theory was proposed by Fajans, and independently of him by Sheppard and Trivelli, to explain the formation of the latent photographic image. In the case of an AgBr emulsion the latent image consists of free silver and bromine; the silver forms nuclei on which, owing to the chemical reaction of development, additional silver is deposited. Pohl and Hilsch1,[^79] succeeded in obtaining a concentration of introduced centers in a separate AgBr crystal comparable with their concentration in an ordinary latent photographic image. On the basis of optical measurements they found that this concentration was equal to \(\sim 3.6\cdot 10^{15}\), or that one center occurred for every \(6.2\cdot 10^{6}\) molecules. It was found that the law of quantum equivalence was fulfilled for small concentrations of introduced centers in a separate AgBr crystal, i.e., one center was formed for each absorbed quantum of light.

Fig. 25. Absorption spectrum of KBr; \(A\)—the characteristic absorption band, \(B\)—the induced absorption band.

Eggert and Noddack found that approximately one silver atom is liberated for each quantum of absorbed light in a thin layer of an AgBr emulsion. This exact agreement between the results obtained on a separate AgBr crystal and on AgBr emulsions led Pohl and Hilsch to identify the coloring centers with the latent photographic image. It is also interesting to note that the more imperfect the crystal, the greater the concentration of centers. Fig. 25 shows, in contrast to the existence of a close connection between the characteri-

…of the characteristic absorption region and that of the latent image in AgCl; a clearly expressed difference between the positions of these two regions for KBr^82. This, however, is generally characteristic of the halide compounds of the alkali metals. The remarkable results shown in Fig. 26 were obtained by illuminating a previously “colored” KBr crystal with light whose wavelength lies in the absorption band of the latent image. Curve \(OAB\) was constructed on the basis of ballistic measurements of the quantity of electricity as a function of the duration of illumination for a freshly colored KBr crystal. Measurements obtained during subsequent illumination with the same light are given in the form of curve \(OC\). The segment \(OA\) on curve \(OAB\) is due, as may be supposed, to the destruction of centers previously formed by illumination with ultraviolet rays; the segment \(AB\) and curve \(OC\) are due to the simultaneous formation and destruction of centers.

Fig. 26. Photocurrent in KBr as a function of illumination duration.

This explanation forces one to acknowledge a partial coincidence of the characteristic absorption band and the absorption band of the latent image. The existence of such a partial coincidence of absorption regions makes it possible to trace the long-wave part of the characteristic absorption band to considerably longer wavelengths than is possible with the aid of ordinary photometric methods. This method depends on the fact that, in a crystal, in the presence of an applied voltage, no charge motion is observed in the case of coloring centers, whereas illumination of the latent image with light of a suitable wavelength is accompanied by a flow of electrons under the same conditions. This makes it possible to illuminate a transparent crystal with light of a wavelength lying in the long-wave part of the characteristic absorption band, and then to verify whether absorption is present or absent. If absorption is present, then upon illuminating the crystal with light corresponding to the long-wave part of the absorption region of the coloring centers, a photocurrent arises. In this way Hilsch and Pohl were able to trace the first characteristic absorption band down to \(3650\ \text{Å}\).

Photoconductivity in Semiconductors

In the years from 1900 to 1917 it was discovered that a vast number of inorganic and organic solid substances possess photosensitivity of a character similar to the photosensitivity of selenium. Among the more or less simple inorganic compounds one should mention stibnite ($\mathrm{Sb_2S_3}$), mercurous iodide ($\mathrm{Hg_2J_2}$), molybdenite ($\mathrm{MoS_2}$), cuprous oxide ($\mathrm{Cu_2O}$), bismuthinite ($\mathrm{Bi_2S_3}$), argentite ($\mathrm{Ag_2S}$), the halide compounds of silver and thallium iodide; in addition, many complex metallic sulfides, carbonates, and also selenium’s competitor as a photoelement—thallophide. All the investigations were carried out on polycrystalline substances; the primary current, considered in the preceding sections, was not observed, probably because of contamination and nonuniformity of the substances. A very large number of studies devoted to the study of the gray metallic modification of selenium (about 2000 works) did not bring understanding of the corresponding processes. This was achieved by Gudden and Pohl, who discovered that the red insulating modification of selenium is photosensitive and possesses photoelectric properties that make it possible to assign it to the group of insulators, including diamond, zinc blende, etc. However, this does not justify Ioffe’s later assertion[^1] that photoconductivity in the gray metallic modification of selenium has to a considerable extent been explained. This study of semiconductors is of interest for the present section chiefly because of the sharply expressed similarity of their photoelectric characteristics to the secondary currents discussed in the preceding section. This is best seen from the extensive investigations carried out by Koblenz and his collaborators at the “Bureau of Standards U. S. A.” Fig. 27, a shows the results obtained for the mineral molybdenite ($\mathrm{MoS_2}$)[^108][^106][^107]. The photoelectric current increases slowly in passing from the deep infrared part of the spectrum to the shorter waves.

Fig. 27. Spectral distribution of the photocurrent and absorption in the mineral molybdenite; a—photocurrent; — absorption; b—photocurrent as a function of light intensity.

This recalls the behavior of primary currents observed in idiochromatic crystals of insulators. Optical absorption, extending also into the infrared region of the spectrum, is ascribed to the electronic dark conductivity of this substance. Fig. 27, b depicts the variation of the photoelectric current with light intensity for two wavelengths. It is interesting to note that the greater the wavelength of the incident light, the closer the dependence between the photoelectric current and the light intensity is to direct proportionality. Fig. 28, a shows the change of the photoelectric current as a function of the duration of illumination for waves of different length in MoS₂. The similarity is most sharply expressed between these curves and the curves for gray metallic selenium and polycrystalline ZnS. Fig. 28, b illustrates the influence of temperature and applied voltage on the curves of spectral sensitivity for MoS₂ crystals. In many of these investigations one had to deal with considerable dark conductivity. The voltages at different temperatures were varied in order to obtain dark currents of equal magnitude. This procedure to some extent masked the influence of temperature on the curves of spectral sensitivity. The curves quite clearly show arbitrariness in the choice of current values. They are similar to the curves given in the section dealing with secondary currents in insulators. The electric currents arising upon illumination of substances were identified by Gudden and Pohl with secondary currents. The curve of photoelectric yield cannot be regarded as a physical constant of the substance, since it is

Fig. 28. Molybdenite: a — photocurrent as a function of the duration of illumination; b — spectral distribution of the photoelectric sensitivity at 23° and 170° C.

of the function of its history, as well as of the intensity of the light, temperature, and applied voltage. Measurements first carried out by Pfund[^25], and then by Koblenz[^115] with a copper-oxide specimen, showed that this substance should be assigned to the semiconductors that yield only secondary photoelectric currents. Recently, however, Gudden[^87] published measurements made in the laboratory at Erlangen, which showed that the photoelectric currents obtained under the same conditions possess characteristics inherent in primary currents.

Published measurements of the photoelectric current in iodide mercury[^96][^118][^119][^120] showed that in this case only secondary currents are observed. However, the author’s latest data for individual crystals of $\mathrm{HgJ}_2$ indicate that, when measures are taken to exclude surface conductivity, primary photocurrents are then observed.

Selenium (metallic modification)

All investigations of metallic selenium up to 1913 were carried out exclusively on a polycrystalline substance.

Fig. 29. Distribution of photoelectric sensitivity over the spectrum for a single crystal of metallic selenium; a—dependence of the photocurrent in a selenium single crystal on the direction of illumination, b—curves of spectral sensitivity for a number of single crystals.

This substance was made photosensitive by heating previously deposited amorphous selenium at a temperature of about $150^\circ\mathrm{C}$ and transforming it into the light-sensitive metallic modification. The first significant step toward clarifying the question was made by Brown[^84][^87][^188] and his co-workers, when they obtained single crystals of photosensitive selenium from the vapor phase.

Photoelectric measurements on these crystals showed that the curves of spectral sensitivity depend on the orientation of the crystal with respect to the light beam. Fig. 29, a shows the change in the photoelectric current as a function of the wavelength for different orientations of the incident light beam with respect to the crystal. Exclusively transverse illumination was used. In Fig. 29, b are given the spectral-sensitivity curves for several individual crystals. These curves (Fig. 29, b) resemble the spectral-sensitivity curves of commercial selenium photocells of various manufacture. The change in current upon illumination of selenium is proportional not to the intensity of the light, but to its square root. This is true both for single crystals and for polycrystalline selenium.

Transmitted effect

Brown86–90 observed that if a single crystal of selenium is placed between electrodes and the electrodes are connected in series with a battery and a galvanometer, then, upon illumination of that part of the crystal which protrudes beyond the electrodes, a change is nevertheless observed in the conductivity of the crystal in the part lying between the electrodes. The photoelectric characteristics of this so-called “transmitted effect” are identical with the curves representing the normal change in the conductivity of selenium when it is illuminated between two electrodes. Extensive searches have failed to reveal the existence of this mysterious effect in other photosensitive substances. This led Gudden185 to doubt its existence altogether. I repeated and confirmed Brown’s observations on hexagonal single crystals of selenium about 2 cm long. The general radiation from a tungsten lamp fell on a small part of the crystal at a distance of about 10 mm from the outer edge of the electrodes. A study of the photoelectric output (response) as a function of the light intensity, carried out on many crystals, showed that for some crystals the photoelectric current is proportional to the light intensity, whereas for other crystals it is proportional to the square root of the intensity of the incident light. I was unable to find the cause of these contradictory results.

The occurrence of an electromotive force in semiconductors under illumination

The occurrence of an electromotive force excited by light in photosensitive semiconductors, accompanied by a change in the resistance of the semiconductor, has been observed over the last several decades in many substances. Adams and Day1 noted that, when the contact point of pieces of Pt and Se was illuminated,

connected in series with a galvanometer (without batteries), an electric current arises in the system. A similar effect is observed both in various commercial selenium cells and in individual crystals of the metallic modification of Se. Other substances exhibit the same properties; for example, one may mention molybdenite \([ \mathrm{MoS_2}]\), argentite \((\mathrm{Ag_2S})\), acanthite \((\mathrm{Ag_2S})\), \(\mathrm{Cu_2S}\), \(\mathrm{Cu_2O}\), \(\mathrm{PbS}\), diamond.

Grondahl and Geiger \(^{147,148}\) appear to have been the first to observe the appearance of an electromotive force upon illumination of cuprous-oxide rectifiers from the side of the copper oxide. Illumination causes an electric current to flow when the base copper is connected in series with a low-resistance galvanometer and an auxiliary electrode on the surface of \(\mathrm{Cu_2O}\), without an external source of potential. Fig. 30 schematically depicts the rectifier and the connection diagram.

The appearance of an electromotive force in a \(\mathrm{Cu_2O}\) rectifier was carefully studied by Schottky \(^{150,151,154,161}\), Auer \(^{149,177}\), Kirschbaum \(^{149,152,173}\), and their collaborators, as well as by Lange \(^{151,169}\), Perutz, Dattlio \(^{193,172}\), Teichmann \(^{168}\), and others.

Fig. 30. Diagram of a valve copper-oxide photocell. The wire mesh, indicated in the figure by small rectangles, forms an auxiliary electrode on the surface of \(\mathrm{Cu_2O}\). The arrows indicate the direction of motion of the electron current in the external circuit.

Photocells are formed by oxidizing the surface of small copper plates. The oxidation process leads to the formation of \(\mathrm{Cu_2O}\) with an outer layer of \(\mathrm{CuO}\). The copper oxide is then removed from the surface of the \(\mathrm{Cu_2O}\) by dissolving it in a suitable solution, leaving a transparent layer of red \(\mathrm{Cu_2O}\) on the base copper. The electrode on the surface of \(\mathrm{Cu_2O}\) usually consists of a thin transparent metallic film, or of a wire mesh, or, finally, of a spiral. The currents obtained under illumination vary in magnitude from \(10^{-4}\) to \(10^{-5}\) A Lin.

Schottky \(^{150,151,149}\) and his collaborators concluded, on the basis of studying the photo-electromotive force and rectification phenomena, that at the boundary between the base copper and \(\mathrm{Cu_2O}\) there exists some extremely thin layer responsible for the occurrence both of the photo-electromotive force and of the rectifying effect. This “blocking layer” (“blocking layer,” “Sperrschicht”) prevents electrons from passing in the direction from \(\mathrm{Cu_2O}\) to the copper metal, but allows them to pass comparatively easily in the opposite direction. The difference in the apparent resistance of the blocking layer to the passage of electrons in the two opposite directions leads to rectification of the electric current. The experiments led to the surprising result,

concluding that the electron flow arising under illumination, without the application of an external electromotive force, is opposite to the electron current in the conducting direction when the element is used as a rectifier. When the surface of Cu₂O is illuminated (the Cu₂O layer is sufficiently thin for light to be able to pass through it to the blocking layer), electrons flow from the cuprous oxide into the copper. The photoelectric current arising as a result of this is directly proportional to the intensity of the light. If the two electrodes are not closed through an external circuit, then the electromotive force that is formed is not proportional to the intensity of the incident light. This is shown in Fig. 31, which presents the dependence of the photoelectric current (without an external battery) and the photoelectromotive force (for a closed circuit) on the light intensity for a rear copper-oxide photocell (to be defined later). An explanation of the phenomenon was given by Luers and Kerschbaum^149, who showed that if the element is regarded as an equivalent current pole, and not as a voltage source, then the observed characteristics are easily explained. Fig. 32 shows the change in photoelectric current when the light beam is moved over the surface of Cu₂O^161. The current increases as the light beam approaches the small gold electrode on the surface of Cu₂O. This change in photocurrent as a function of the distance of the illuminated region from the electrode was associated with leakage

Fig. 31. Dependence of the photoelectromotive force and photocurrent on the intensity of light for a rear copper-oxide photocell (with rear effect).

Fig. 32. Dependence of the photocurrent on the position of the beam of incident light relative to the auxiliary electrode.

of Cu₂O.

of electrons back into the Cu\(_2\)O layer. The shape of the curve in Fig. 32 was interpreted by Audubert and Kerschbaum as evidence of a decrease in the photocurrent when passing through the Cu\(_2\)O layer to the electrode. The decrease in current is proportional to the square root of the product of the resistance of the Cu\(_2\)O layer and the conductivity of the blocking layer.

The current-voltage characteristics of a cuprous-oxide element in the dark and under illumination are shown in Fig. 33. The difference in the values of the electron current in the two cases becomes smaller the more positive the electrode on Cu\(_2\)O\(^{149}\) becomes.

One may think that this is due to an increase in the change in the conductivity of Cu\(_2\)O under the action of light; the photoelectromotive force predominates as long as the applied voltage is small.

Fig. 33. Current-voltage characteristics for a cuprous-oxide photoelement.

The elements considered up to now in this paragraph were so-called photoelements with a back effect. In these elements the photoeffect occurs at the Cu\(_2\)O—Cu boundary and is associated with a hypothetical blocking layer. Dume and Schottky found that illumination of the surface of Cu\(_2\)O coated with a transparent metallic film causes an electron current flowing from Cu\(_2\)O into this surface film; i.e., having a direction opposite to the direction of the electron current in elements with a back photoeffect. It was found that the magnitude of the photocurrent per unit of light energy increases if, before deposition of the metallic electrode, the Cu\(_2\)O is etched with some suitable agent. This photoelectric process arises in the transition layer between Cu\(_2\)O and the metallic film. Elements with such an effect are called photoelements with a front effect\(^{154}\). In the selenium element recently described by Bergmann\(^{142}\), only the front photoeffect is observed. These elements are manufactured as follows: selenium is applied (by spraying) onto small iron disks and then coated by sputtering with a transparent layer of gold or silver; the selenium must first be converted into the photosensitive metallic modification. The electrons flow from the selenium to the sputtered metallic film. The curves of the dependence

photoelectromotive force and photocurrent on the intensity of light are qualitatively the same as for copper-oxide photocells. Dember[^166][^169] has recently published measurements with large individual crystals of the mineral cuprite (\(\mathrm{Cu}_2\mathrm{O}\)), which appear to indicate that the direction of the incident light beam with respect to the electrode determines the polarity of the electromotive force. Bergmann[^179][^181] has since published his investigations of a whole series of photosensitive semiconductors in which this effect is also observed.

Dember,[^188] discussing his results, points out that the occurrence of a photoelectromotive force cannot be fully explained by the presence of a blocking layer and that other causes must be sought.

LITERATURE

1. P. Lukirskii, “Zh. R. F. Kh. O.,” 1916 (A. F. Ioffe, “The physics of Crystals,” New York, McGraw-Hill, 1928).

2. B. Gudden and R. Pohl, Lichtelektrische Beobachtungen an isolierenden Metallsulfiden, “Zs. Physik,” 2, 361, 1920.

3. B. Gudden and R. Pohl, Lichtelektrische Beobachtungen an Zinksulfiden, “Zs. Physik,” 2, 181, 1920.

4. B. Gudden and R. Pohl, Lichtelektrische Leitfähigkeit und Phosphoreszenz, “Zs. Physik,” 3, 98, 1920.

5. B. Gudden and R. Pohl, Über lichtelektrische Leitfähigkeit von Diamanten, “Zs. Physik,” 3, 123, 1920.

6. B. Gudden and R. Pohl, Über lichtelektrische Leitfähigkeit von Zinksulfidphosphoren, “Zs. Physik,” 4, 206, 1921.

7. B. Gudden and R. Pohl, Über lichtelektrische Leitfähigkeit von Zinkblende, “Zs. Physik,” 5, 176, 1921.

8. B. Gudden and R. Pohl, Ein Vorlesungsversuch über lichtelektrische Leitfähigkeit von Isolatoren, “Zs. Physik,” 5, 387, 1921.

9. B. Gudden and R. Pohl, Über den zeitlichen Anstieg der lichtelektrischen Leitfähigkeit, “Zs. Physik,” 6, 248, 1921.

10. B. Gudden and R. Pohl, Über den Mechanismus der lichtelektrischen Leitfähigkeit, “Zs. Physik,” 7, 65, 1921.

11. B. Gudden and R. Pohl, Über lichtelektrische Leitfähigkeit, “Phys. Zs.,” 22, 529, 1921.

12. P. P. Koch and F. Schrader, Über die Einwirkung des Lichtes auf Chlorsilber, Bromsilber und Jodsilber, “Zs. Physik,” 6, 127, 1921.

13. W. C. Röntgen, Über die Elektrizitätsleitung in einigen Kristallen und über den Einfluss einer Bestrahlung darauf, “Ann. d. Physik,” 64, 1, 1921.

14. H. Rose, Über die lichtelektrische Leitfähigkeit des Zinnobers, “Zs. Physik,” 6, 174, 1921.

15. B. Gudden and R. Pohl, Lichtelektrische Leitfähigkeit in weiterem Zusammenhang, “Phys. Zs.,” 23, 417, 1922.

16. B. Gudden and R. Pohl, Zur lichtelektrischen Leitfähigkeit des Diamanten, “Zs. f. techn. Physik,” 3, 199, 1922.

17. M. Levi, On Photoelectric Conductivity of Diamond and Other Fluorescent Crystals, “Proc. Trans. Roy. Soc.,” Canada, 16,—III,—241, 1922.

18. J. Eggert and W. Noddack, Zur Prüfung des Photochemischen Äquivalentgesetzes an Trockenplatten, “Zs. Physik,” 20, 299, 1923; 21, 264, 1924.

19. B. Gudden and R. Pohl, Über lichtelektrische Wirkung und Leitung in Kristallen, “Zs. Physik,” 16, 170, 1923.

20. B. Gudden and R. Pohl, Lichtelektrische Leitung und Chemische Bindung, “Zs. Physik,” 16, 42, 1923.

1. B. Gudden and R. Pohl, Neuere Beobachtungen über den Zusammenhang elektrischer und optischer Erscheinungen, “Naturwiss.”, 11, 348, 1923.

2. B. Gudden and R. Pohl, Das Quantenäquivalent bei der lichtelektrischen Leitung, “Zs. Physik”, 17, 331, 1923.

3. B. Gudden and R. Pohl, Zur lichtelektrischen Leitfähigkeit des Zinnobers, “Zs. Physik”, 18, 199, 1923.

4. W. Heintze, Über lichtelektrische Leitfähigkeit von Cerussit und Senarmontit, “Zs. Physik”, 15, 339, 1923.

5. F. Peter, Über Brechungsindizes und Absorptionskonstanten des Diamanten, zwischen 644 u. 226 mμ. “Zs. Physik”, 15, 858, 1923.

6. J. Bingel, Über lichtelektrische Wirkung in Steinsalzkristallen, “Zs. Physik”, 21, 229, 1924.

7. P. L. Baily, Coloration of the Alkali Halides by X-Rays, “Phys. Rev.”, 24, 495, 1924.

8. B. Gudden and R. Pohl, Zum Mechanismus des lichtelektrischen Primärstromes in Kristallen, “Zs. Physik”, 30, 14, 1924.

9. B. Gudden and R. Pohl, Über elektrische Leitfähigkeit bei Anregung und Lichtemission von Phosphoren, “Zs. Physik”, 21, 1, 1924.

10. H. Lenz, Über den Hall-Effekt des lichtelektrischen Primärstromes bei isolierenden Kristallen, “Phys. Zs.”, 25, 435, 1924.

11. W. Flechsig, Zur Kenntnis des lichtelektrischen Primärstromes in Kristallen, “Zs. Physik”, 33, 372, 1925.

12. A. Frum, Zur lichtelektrischen Leitung und Phosphorescenz von NaCl-Kristallen, Dissert. Göttingen, 26, 1925.

13. Z. Gyulai, Zur lichtelektrischen Leitung in NaCl-Kristallen “Zs. Physik”, 31, 296, 1925.

14. B. Gudden and R. Pohl, Über lichtelektrische Leitung in Selen, “Zs. Physik”, 35, 243, 1925.

15. B. Gudden and R. Pohl, Über den lichtelektrischen Primärström in NaCl-Kristallen, “Zs. Physik”, 31, 651, 1925.

16. B. Gudden and R. Pohl, Zur lichtelektrischen Leitung bei tiefen Temperaturen, “Zs. Physik”, 34, 249, 1925.

17. Z. Gyulai, Zum Absorptionsvorgang in lichtelektrisch leitenden NaCl-Kristallen, “Zs. Physik”, 33, 251, 1925.

18. Z. Gyulai, Zum Quantenäquivalent bei der lichtelektrischen Leitung in NaCl-Kristallen, “Zs. Physik”, 32, 103, 1925.

19. H. Lenz, Über den Durchgang von Elektronen durch lichtelektrisch empfindliche Kristallen, “Ann. d. Physik”, 77, 449, 1925.

20. B. Gudden and R. Pohl, Zur Kenntnis des Elektronenleitung in Kristallen, “Phys. Zs.”, 26, 481, 1925.

21. A. Arsenjewa, Über die lichtelektrische Leitfähigkeit im Steinsalz, “Zs. Physik”, 37, 701, 1926.

22. R. Pohl and E. Rupp, Über Alkalihalogenidphosphore, “Ann. d. Physik”, 81, 1161, 1926.

23. W. Flechsig, Zur Lichtabsorption in verfärbten Alkalihalogeniden, “Zs. Physik”, 36, 605, 1926.

24. B. Gudden and R. Pohl, Über den scheinbaren Antagonismus kurzer und langer Wellen bei der inneren lichtelektrischen Wirkung, “Zs. Physik”, 37, 881, 1926.

25. Z. Gyulai, Lichtelektrische und optische Messungen an blauen und gelben Steinsalzkristallen, “Zs. Physik”, 35, 411, 1926.

26. Z. Gyulai, Über den Vorgang der Erregung bei der Lichtabsorption in Kristallen, “Zs. Physik”, 39, 838, 1926.

27. Z. Gyulai, Zur additiven Färbung von Alkalihalogenidkristallen, “Zs. Physik”, 37, 889, 1926.

28. R. Hilsch and R. Ottmer, Zur lichtelektrischen Wirkung in natürlichem blauem Steinsalz, “Zs. Physik”, 39, 344, 1926.

29. H. Siedentopf, Ultramikroskopische Untersuchungen über Steinsalzfärbungen, “Phys. Zs.”, 6, 855, 1905.

30. W. H. Wise, The Effect of Recombination on the Primary Photoelectric Current from a Crystal, “Phys. Rev.”, 28, 57, 1926.

1. R. Pohl, Zur quantenhaften Lichtabsorption in festen Körpern, Naturwiss., 14, 214, 1926.

2. M. A. Bredig, Über anomale Dispersion in Alkalihalogenidphosphoren, Zs. Physik, 46, 73, 1926.

3. R. Hilsch, Über die ultraviolette Absorption einfach gebauter Kristalle, Zs. Physik, 44, 421, 1927.

4. R. Hilsch, Die Absorptionsspektra einiger Alkali-Halogenid-Phosphore mit Tl- und Pb-Zusatz, Zs. Physik, 44, 860, 1927.

5. B. Kultermeyer, Optical Absorption and Photoelectric Conductivity of Sulphur Crystals, Phys. Rev., 29, 615, 1927.

6. H. Lenz, Elektronenleitung in Kristallen mit besonderer Berücksichtigung der Verhältnisse bei tiefen Temperaturen, Ann. d. Physik, 82, 775, 1927.

7. H. Lenz, Die Temperaturabhängigkeit des lichtelektrischen Primärstromes in Diamanten, Ann. d. Physik, 83, 941, 1927.

8. A. Smakula, Einige Absorptionsspektra von Alkalihalogenidphosphoren mit Silber und Kupfer als wirksamen Metallen, Zs. Physik, 45, 1, 1927.

9. W. Flechsig, Über die Sättigung des lichtelektrischen Primärstromes in Kristallen, Zs. Physik, 46, 788, 1928.

10. R. Hilsch u. R. W. Pohl, Über die ersten ultravioletten Eigenfrequenzen einiger einfacher Kristalle, Zs. Physik, 48, 384, 1928.

11. H. Lorenz, Zur Temperaturabhängigkeit der Absorptionsbanden in Alkalihalogenid-Phosphoren, Zs. Physik, 46, 558, 1928.

12. M. Podaschewsky, Über die Wirkung der plastischen Deformation auf den inneren Photoeffekt in Steinsalzkristallen, Zs. Physik, 56, 362, 1929.

13. K. H. Voigt, Nachweis des lichtelektrischen Primärstromes in Antimonglanz, Zs. Physik, 57, 154, 1929.

14. M. Podaschewsky, Die Spektralverteilung des inneren Photoeffekts in den plastisch deformierten NaCl-Kristallen, Zs. Physik, 65, 799, 1930.

15. P. Tartakowsky, Über die Polarisation bei der lichtelektrischen Leitfähigkeit von röntgenisiertem Steinsalz, Zs. Physik, 66, 830, 1930.

16. F. C. Toy and G. B. Harrison, Photo-conductance Phenomena in Silver Halides, and the Latent Photographic Image. Introduction and Part 1, Proc. Roy. Soc., 127 A, 613, 1930.

17. B. Gudden, Über Leitungs und Photoelektronen in Isolatoren und Halbleitern, Phys. Zs., 32, 925, 1931.

18. W. Flechsig, Über die Stromspannungsabhängigkeit bei der lichtelektrischen Leitung in Kristallen, Phys. Zs., 32, 843, 1931.

19. R. Ottmer, Zur Kenntnis der Absorptionsspektra lichtelektrisch leitender Alkalihalogenide, Zs. Physik, 46, 798, 1928.

20. R. Hilsch u. R. W. Pohl, Über die ersten ultravioletten Eigenfrequenzen einiger einfacher Kristalle, Zs. Physik, 48, 384, 1928.

21. R. Hilsch u. R. W. Pohl, Die in Luft messbaren intravioletten Dispersionsfrequenzen der Alkalihalogenide, Zs. Physik, 57, 145, 1929.

22. A. Smakula, Über Erregung und Entfärbung lichtelektrisch leitender Alkalihalogenide, Zs. Physik, 59, 603, 1930.

23. R. Hilsch u. R. W. Pohl, Einige Dispersionsfrequenzen der Alkalihalogenid-Kristalle im Schumanngebiet, Zs. Physik, 59, 812, 1930.

24. A. Smakula, Über die Verfärbung der Alkalihalogenidkristalle durch ultraviolettes Licht, Zs. Physik, 63, 762, 1930.

25. R. Hilsch u. R. W. Pohl, Zur Photochemie der Alkali- und Silberhalogenidkristalle, Zs. Physik, 64, 606, 1930.

26. H. Fesefeldt, Der Einfluss der Temperatur auf die Absorptionsspektra der Alkalihalogenidkristalle, Zs. Physik, 64, 623, 1930.

27. H. Fesefeldt u. Z. Gyulai, Zur Lichtabsorption in Silber und Kupferhalogenidkristallen, Nachr. d. Ges. d. Wiss. z. Göttingen Math.-Phys., Kl. 226, 1929.

28. R. Hilsch u. R. W. Pohl, Über die Ausnutzung der latenten Bilder bei der photographischen Entwicklung, Nachr. d. Ges. d. Wiss. z. Göttingen, Math.-Phys., Kl. 334, 1930.

29. R. Hilsch u. R. W. Pohl, Über das latente photographische Bild, Nachr. d. Ges. d. Wiss. z. Göttingen, Math.-Phys., Kl. 176, 1930.

1. E. Mollwo, On the absorption spectra of photochemically colored alkali-halide crystals, Nachr. d. Ges. d. Wiss. z. Göttingen. Math.-Phys., Kl. 96, 1931.

2. E. Mollwo, The absorption spectrum of photochemically colored alkali-halide crystals at various temperatures, Nachr. d. Ges. d. Wiss. z. Göttingen Math.-Phys., Kl. 236, 1931.

3. B. Hilsch and R. W. Pohl, On light absorption in simple ionic lattices and the electrical detection of the latent image, Zs. Physik, 68, 721, 1931.

4. N. Kolabuchow and B. Fischelow, On the spectral distribution of the depolarization current in the photoelectric conduction of X-ray-activated rock salt, Zs. Physik, 75, 282, 1932.

5. F. C. Brown, The Crystal Forms of Metallic Selenium and some of their Physical Properties, Phys. Rev. (2), 4, 85, 1914.

6. F. C. Brown and L. P. Sieg, Wave-Length-Sensibility Curves for Light Sensitive Selenium and their Significance, Phys. Rev. (2), 4, 48, 1914.

7. F. C. Brown and L. P. Sieg, The Seat of Light Action in Certain Crystals of Metallic Selenium and Some New Properties in Matter, Phil. Mag. (6), 28, 497, 1914.

8. F. C. Brown and L. P. Sieg, Wave-Length-Sensibility curves of certain crystals of metallic Selenium, and a Partial explanation of the complexity of Light action in Selenium Cells, Phys. Rev. (2), 4, 507, 1914.

9. F. C. Brown, The Electrical, the Photoelectrical and the Electromechanical Properties of Certain Crystals of Metallic Selenium, with certain Applications to Crystal Structure, Phys. Rev. (2), 5, 167, 1915.

10. F. C. Brown, The Nature of Electric Conduction as Required to Explain the Recovery of Resistance of Metallic Selenium Following Illumination, Phys. Rev. (2), 5, 395, 1915.

11. F. C. Brown, Some Experiments on the Nature of Transmitted Light-Action in Crystals of Metallic Selenium, Phys. Rev. (2), 5, 404, 1915.

12. L. P. Sieg and F. C. Brown, An Extension toward the Ultraviolet of the Wave-Length-Sensibility Curves for Certain Crystals of Metallic Selenium, Phys. Rev. (2), 5, 65, 1915.

13. M. Volmer, Photochemical sensitivity and photoelectric conductivity, Zs. f. Elektrochem., 21, 113, 1915.

14. E. O. Dieterich, The Effect of Temperature on the Light-Sensibility Curves of Different Types of Selenium Cells, Phys. Rev. (2), 8, 191, 1916.

15. K. J. Dieterich, The Effect of Temperature on the Resistance, the Light-Sensitiveness, and the Rate of Recovery of Certain Crystals of Metallic Selenium, Phys. Rev. (2), 7, 551, 1916.

16. A. H. Pfund, The Light Sensitiveness of Copper Oxide, Phys. Rev. (2), 7, 289, 1916.

17. M. Volmer, The dependence of the photoelectric conductivity of “red” mercuric iodide (HgJ₂ rot; J), Zs. f. Wiss. Photogr., 16, 152, 1916.

18. T. W. Case, Notes on the Change of Resistance of Certain Substances in Light, Phys. Rev. (2), 9, 305, 1917.

19. W. W. Coblentz and W. B. Emerson, The Photoelectric Sensitivity of Various Substances, Wash. Acad. Soc. Journ., 7, 525, 1917.

20. E. H. Kennard and E. O. Dieterich, An Effect of Light upon the Contact Potential of Selenium and Cuprous Oxide, Phys. Rev. (2), 8, 58, 1917.

21. W. E. Tisdale, The Effects of Gases and Metallic Vapors on the Electrical Properties Exhibited by Selenium Crystals of the Hexagonal System, Phys. Rev. (2), 12, 325, 1918.

22. W. W. Coblentz and L. S. McDowell, Photoelectric Sensitivity vs Current Rectification in Molybdenite, Phys. Rev. (2), 13, 154, 1919.

23. W. W. Coblentz and L. S. McDowell, Photoelectric Sensitivity of Bismuthinite and Various Other Substances, Scient. Pap. Bureau of Stand., 322, 14, 1919.

24. W. W. Coblentz and H. Kahler, Some Optical and Photoelectric Properties of Molybdenite, Scient. Pap. Bureau of Stand., 338, 40, 1919.

1. W. W. Coblentz and H. Kahler, Spectral Photoelectric Sensitivity of Silver Sulphide and Several Other Substances, “Scient. Pap. Bureau of Stand.,” 344, 181, 1919.

2. T. W. Case, “Thalofide Cell, a New Photoelectric Substance,” “Phys. Rev.,” (2), 15, 289, 1920.

3. W. W. Coblentz and H. Kahler, The Spectral Photoelectric Sensitivity on Molybdenite as a Function of the Applied Voltage, “Phys. Rev.,” (2), 14, 534, 1920.

4. W. W. Coblentz, Positive and Negative Photoelectrical Properties of Molybdenite and Several Other Substances, “Scient. Pap. Bureau of Stand.,” 398, 42, 1920.

5. W. W. Coblentz, Spectrophotoelectric Sensitivity of Thalofide, “Scient. Pap. Bureau of Stand.,” 380, 6, 1920.

6. W. W. Coblentz and H. Kahler, Some General Characteristics of Spectrophotoelectrical Conduction in S, Jdg., “Journ. Opt. Soc. Amer.,” 4, 219, 1920.

7. W. Ehlers u. P. P. Koch, Über die Einwirkung des Lichtes auf Bromsilber, “Zs. f. Physik,” 3, 161, 1920.

8. A. M. McMahon, The Action of Roentgen and Gamma Radiations upon the Electrical Conductivity of Selenium Crystals, “Phys. Rev.,” (2), 16, 568, 1920.

9. W. W. Coblentz, Spectrophotoelectrical Sensitivity of Proustite, “Scient. Pap. Bureau of Stand.,” 412, 10, 1912.

10. F. Kümpf, Beitrag zur Kenntnis der Elektrizitätsleitung in festen, elektronisch schlecht leitenden Körpern. Versuche am Quecksilberjodid, “Ann. d. Physik,” 68 (4), 463, 1921.

11. W. W. Coblentz, Spectrophotoelectrical Sensitivity of Argentite (Ag₂S), “Scient. Pap. Bureau of Stand.,” N 448, 16, 1922.

12. W. W. Coblentz, Various Photoelectrical Investigations, Scient. Pap. Bureau of Stand.,” 462, 23, 1922.

13. W. W. Coblentz, Some Observations on the Transformation of Thermal Radiant Energy into Electrik Current in Molybdenite, “Journ. Waschington. Acad. Sci.,” 12, 411, 1922.

14. W. W. Coblentz and J. F. Eckford, Spectrophotoelectrical Sensitivity of Bournonite and Pyrargyrite, “Scient. Pap. Bureau of Stand.,” N 451, 20, 1922.

15. W. W. Coblentz and J. F. Eckford, Spectrophotoelectrical Sensitivity of Some Halide Salts of Thallium, Lead, and Silver, “Scient. Pap. Bureau of Stand.,” N 456, 10, 1922.

16. F. Kümpf, Über den Mechanismus der lichtelektrischen Leitfähigkeit, “Phys. Zs.,” 23, 420, 1922.

17. J. G. Emery, The Transmitted Effect in Selenium Crystals, “Phys. Rev.,” (2), 19, 414, 1922.

18. W. Späth, Zur Kenntnis des Selens, “Zs. Physik,” 8, 165, 1922.

19. A. Wendt, Charakteristiken belichteter Selenzellen, “Verh. d. D. Phys. Ges.,” 3, 26, 1922.

20. L. Gröh, Über die Elektrische Leitfähigkeit fester Dielektrika bei Bestrahlung mit Röntgenstrahlen, “Zs. Physik,” 17, 295, 1923.

21. E. Luppo, Über Leitfähigkeitserregung bei Hitzaustreiben der Phosphore, “Ann. d. Physik,” 70, 391, 1923.

22. V. P. Burton, Light Sensitivity of Cuprous Oxide and of Selenium, “Phys. Rev.,” (2), 23, 337, 1924.

23. H. Küstner, Die Empfindlichkeit der Selenzelle auf Röntgenstrahlen verschiedener Wellenlänge, “Zs. Physik,” 27, 124, 1924.

24. S. L. Martin, The Change of Resistance of Molybdenite Due to Light, “Journ. Proc. Roy. Soc.,” New South Wales 58, 150, 1924.

25. A. Predwoditelew u. N. Notchaeva, Über die Wirkung der Schichtdicke auf den photoelektrischen Effekt in Farbstoffen, “Zs. Physik,” 29, 332, 1924.

26. T. Th. Bakor, The Sensitivity of Selenium Cells, “Natur,” 117, 858, 1925.

1. F. Kämpf, Über einen neuen Leitfähigkeitseffekt beim Zusammenwirken von blauem und rotem Licht. “Zs. Physik,” 38, 104, 1926.

2. W. W. Coblentz and C. W. Hughes, A New Selective Radiometer of Molybdenite, “Phys. Rev.” (2), 29, 365, 1927.

3. R. J. Piersol, Influence of Temperature on Selenium Photosensitivity, “Phys. Rev.” (2), 29, 302, 1927.

4. R. J. Piersol, Relation Between Light Intensity and Photocurrent in Selenium, “Phys. Rev.” (2), 29, 362, 1927.

5. I. Runge u. R. Sewig, Über den inneren Photoeffect in Kristallinen Halbleitern, “Zs. Physik,” 52, 726, 1930.

6. R. Sewig, Die lichtelektrischen Eigenschaften von Thalliumzellen, “Zs. f. techn. Physik,” 11, 269, 1930.

7. A. Petrikaln, Über den Einfluss des Gasdrucks auf die lichtelektrische Leitfähigkeit organischer Farbstoffe, “Zs. f. phys. Chem.,” B. 10, 9, 1930.

8. W. G. Adams and R. E. Day, The Action of Light on Selenium, “Proc. Roy. Soc.,” A 25, 113, 1876.

9. H. H. Sheldon and P. H. Geiger, The Production of an e. m. f. on Closed Circuit by a Light Effect on Argentite, “Proc. Acad. Amer.,” 8, 161, 1922.

10. H. H. Sheldon and P. H. Geiger, The Production of an e. m. f. on Closed Circuit by a Light Effect on Argentite, “Phys. Rev.” (2), 19, 389, 1922.

11. P. H. Geiger, Spectrophotoclectrical Effects in Argentite “Phys. Rev.” (2), 22, 461, 1923.

12. E. H. Kennard and C. Moon, The Photoelectromotive Effect in Selenium, “Phys. Rev.” (2), 21, 374, 1923.

13. E. Merritt, The Effect of Light on the Behavior of Selenium Rectifiers, “Phys. Rev.” (2), 23, 555, 1924.

14. W. W. Coblentz, Some New Thermoelectrical and Actino-electrical Properties of Molybdenite, “U. S. Bureau of Stand. Sci. Pap.,” No. 486, 43, 1924.

15. E. Merritt, The Effect on the Behavior of Selenium Contact Rectifiers, “Proc. Nat. Acad. Amer.,” 11, 372, 1925.

16. R. M. Holmes, The Effect of Light on the Thermoelectric Power of Selenium, “Phys. Rev.” (2), 25, 820, 1925.

17. W. A. Scheider, Actino-electric Effects in Argentite, “Phys. Rev.” (2), 29, 363, 1927.

18. L. O. Grondahl and P. H. Geiger, A New Electronic Rectifier, “Trans. Amer. Inst. Elec. Eng.,” 46, 357, 1927.

19. L. O. Grondahl and P. H. Geiger, Improvement Relating to Light-Sensitive Apparatus, British Patent No. 304, 644.

20. O. R. v. Auwers u. K. Kerschbaum, Über Sperrschichtphotozellen, “Ann. d. Physik,” (5), 129, 1930.

21. W. Schottky, Über den Entstehungsort der Photoelektronen in Kupfer-Kupferoxydul-Photozellen, “Phys. Zs.,” 31, 913, 1930.

22. W. Schottky, Über den Entstehungsort der Photoelektronen in Kupfer-Kupferoxydul-Photozellen, “Z. f. techn. Physik,” 11, 458, 1930.

23. H. Kerschbaum, Über eine Methode zur Trennung des Sperrschichtphotoeffektes und des inneren Photoeffektes. “Naturwiss.,” 18, 832, 1930.

24. B. Lange, Eine neue Art von Photozellen, “Phys. Zs.,” 31, 139, 1930.

25. E. Duhme u. W. Schottky, Über Sperr- und Photoeffekte an der Grenze von Kupferoxydul gegen aufgestäubte Metallschichten, “Naturwiss.,” 18, 735, 1930.

26. Wolfgang Ostwald, Über Systeme mit besonders kleiner asymmetrischer Austrittsarbeit für Elektronen, “Kolloid. Z.,” 51, 370, 1930.

27. B. Lange, Photozellen in Wissenschaft und Technik. (1 Teil), “Naturwiss.,” 19, 103, 1931.

28. W. Graffunder, Über den lichtelektrischen Effekt bei Trockengleichrichtern. “Phys. Zs.,” 31, 375, 1930.

29. H. Teichmann, Über eine an Kupfer-Kupferoxydulzellen beobachtete Temperaturabhängigkeit des Sperrschichtphotoeffektes, “Zs. Physik,” 65, 709, 1930.

30. B. Lange, Über die Temperaturabhängigkeit des Sperrschicht-Photoeffektes, “Phys. Zs.,” 32, 850, 1931.

1. V. Lange, Über die spektrale Empfindlichkeit von Sperrschicht-Photozellen, “Naturwiss.,” 19, 525, 1931.

2. W. Schottky, Leitungs und Photoeffekte an Sperrschichten “Phys. Zs.,” 32, 833, 1931.

3. H. Teichmann, Ein einfacher Versuch zur Demonstration der Temperaturabhängigkeit des Sperrschichtphotoeffekts, “Zs. Physik,” 67, 192, 1931.

4. E. Perucca u. Deaglio, Ist der Sperrschichtphotoeffect ein Hallwachseffekt, “Zs. Physik,” 72, 102, 1931.

5. L. Bergmann, Über eine neue Selen-Sperrschicht-Photozellen, “Phys. Zs.,” 32, 286, 1931.

6. V. Brazyoduro, Photoelectromotive Force of Selenium, “Chem. Abs.,” 26, 1522, 1932; “Atti. Accad. Sci. Torino, Classe Sci. Fis. Mat. Nat.,” 66, 150, 1931.

7. H. Dember, Über eine Kristallphotozelle, “Phys. Zs.,” 32, 856, 1931.

8. F. u. Körösy u. P. Selény, Über ein physikalisches Modell der Sperrschicht-Photozellen, “Phys. Zs.,” 32, 847, 1931.

9. E. Duhme, Über den Sperrschichtphotoeffekt, “Z. Elektrochem.,” 37, 688, 1931.

10. H. Dember, Über eine photoelektromotorische Kraft in Kupferoxydul Kristallen, “Phys. Zs.,” 32, 554, 1931.

11. L. Dubar, Sur la Sensibilité des Collules Photoelectriques à l’Oxyde Cuivreux du Type à Grille de Cuivre. “Comptes Rendus,” 193, 659, 1931.

12. P. Auger et C. Lapicque, Variation dans le Spectre de la Sensibilité des Cellules au Protoxyde de Cuivre, “Comptes Rendus,” 193, 319, 1931.

13. E. Perucca and R. Deaglio, Über den photoelektrischen Effekt in Cu₂O-Cu-Gleichrichter, “Ann. d. Physik,” (5), 10, 257, 1931.

14. O. v. Auwers u. H. Kerschbaum, Über den photoelektrischen Effekt in Cu₂O-Cu Gleichrichtern “Ann. d. Physik,” 10, 262, 1931.

15. I. Kurtschatow, C. Sinelnikow u. M. Borissow, II, Innerer Photoeffekt und Sperrschichtphotozellen, “Phys. Zs. d. Sowjetunion,” 1, 42, 1932.

16. I. Kurtschatow u. C. Sinelnikow, Untersuchung der Sperrschichtphotozellen, 1. “Phys. Zs. d. Sowjetunion,” 1, 23, 1932.

17. E. Rupp, Eine Beeinflussung des Photoelektronenstromes in Sperrschichtzellen durch magnetische Felder, “Naturwiss.,” 20, 253, 1932.

18. R. Robertson, J. J. Fox and A. E. Martin, Photo-conductivity of Diamonds, “Nature,” 129, 579, 1932.

19. A. H. Lamb, Photronic Cell a Direct Aid to Better Lighting. “Elec. World,” 99, 692, 1932.

20. L. Bergmann, Über einen lichtelektrischen Effekt in Halbleitern. “Phys. Zs.,” 33, 209, 1932.

21. H. Dember, Über die Vorwärtsbewegung von Elektronen durch Licht. “Phys. Zs.,” 33, 207, 1932.

22. L. Bergmann, Über einen neuartige lichtelektrischen Effekt, “Naturwiss.,” 20, 15, 1932.

23. L. Bergmann, Über die Einwirkung von polarisiertem Licht auf Sperrschicht-Photozellen, “Phys. Zs.,” 33, 17, 1932.

24. J. Frenkel and A. Joffé, On the Electric and Photoelectric Properties of Contacts Between a Metal and a Semi-conductor, “Phys. Rev.” (2) 39, 530; 1932.

25. J. Frenkel and A. Joffé, On the Electric and Photoelectric Properties of Contacts Between a Metal and a Semi-conductor, “Phys. Zs. d. Sowjetunion,” 1, 60, 1932.

26. B. Gudden, “Lichtelektrische Erscheinungen” (Julius Springer, 1928).

27. F. Waibel u. W. Schottky, Über die Natur der Sperrschicht bei Kupferoxydulgleichrichtern, “Naturwiss.,” 20, 297, 1932.

28. F. Waibel, Der Becquereleffekt von Kupferoxydul als Sperrschichtphotoeffekt, “Zs. Physik,” 76, 231, 1932.

29. H. Teichmann, Das elektrische Verhalten von Grenzschichten, “Ann. d. Physik,” 13, 649, 1932.

30. F. v. Körösy u. P. Selényi, Photozelle und Lichtelement, “Ann. d. Physik,” 13, 703, 1932.

1. Reference marker visible in the source text. ↩↩