LUMINESCENCE OF ACTIVATED CRYSTALS

V. L. Levshin

1. INTRODUCTION

In recent decades, luminous substances have acquired exceptionally great importance in various branches of technology. Television sets, radar devices, cathode oscillographs, the electron microscope, and other electron-beam instruments are inconceivable without luminous screens. X-ray inspection and modern lighting technology make extensive use of luminescent substances. Luminescent lamps constitute light sources of an entirely new type, which for the first time have made it possible to obtain radiation of any spectral composition, selected in accordance with technical requirements, in an extremely economical and convenient manner. In producing white light by means of luminescent lamps, the consumption of electrical energy is reduced by a factor of 3–4; in producing colored radiation, by tens of times. The revolution in lighting technology now being brought about by luminescent lamps is, in scale and significance, analogous to the revolution brought about by the electric incandescent lamp.

In the Soviet Union, the development of work in the field of luminescence is inseparably connected with the name of S. I. Vavilov, who initiated the systematic study of this new branch of physical optics and for more than thirty years headed the Soviet school of physicists he created, studying problems of luminescence. This school has achieved outstanding successes and, in theoretical respects, plays a leading role in world science.

The present article is devoted to a consideration of the luminescence of the technically most important class of luminous substances—the class of crystal phosphors. However, before turning to a description of the properties of their luminescence and of existing ideas about its kinetics, we shall briefly dwell on the classification of luminescence phenomena, which clarifies the position of crystal phosphors within the general complex of luminescent substances.

2. MAIN TYPES OF LUMINESCENCE

As early as 1934, S. I. Vavilov1, having considered the cases of luminescence known at that time, proposed dividing them into three classes: spontaneous, forced, and recombination luminescence. To the first class were assigned: resonance radiation, many phenomena of fluorescence, and quadrupole radiation. The class of forced luminescence comprised cases of luminescence arising when excited centers, which are in a metastable state, are made to emit. The latter is possible only after the excited centers absorb some additional energy, which removes them from metastable states and transfers them into the ordinary excited state, followed by a spontaneous transition to the normal unexcited state. The third class comprised recombination luminescence, arising at the expense of the energy released when particles separated at the moment of excitation reunite.

These three types of luminescence differ in duration, in the law of decay after the cessation of excitation, in their dependence on temperature, and in a number of other features. A spontaneous process decays exponentially with time and depends little on temperature. A forced process also decays exponentially, but depends strongly on temperature. In recombination luminescence, in the simplest case, the change in luminescence intensity follows a hyperbolic law; the weakening of the luminescence proceeds very rapidly in the first stages of decay and slowly at later ones. The course of recombination luminescence also depends strongly on temperature. The character of the decay can often serve as a good indicator for determining the type of luminescence.

The described types of luminescence sometimes occur in pure form; however, in the majority of cases the processes corresponding to them are elementary links in the kinetics of more complex luminescences consisting of several stages. Therefore, as a further development of the classification, the following scheme may be proposed.2

All luminescences are divided into two large classes: luminescence of discrete centers, i.e., centers separated from one another and acting independently, and recombination luminescence, whose occurrence is connected with the reunification of oppositely charged particles formed earlier from the breakup, upon excitation, of larger particles in different parts of the excited substance.

These large classes of luminescence are then further divided into subclasses. Luminescence of discrete centers may be independent (spontaneous) or non-independent (forced). Recombination luminescences may arise when

direct recombination or upon recombination after intermediate sojourns in metastable states. The concluding stage of all types of complex luminescence, in the overwhelming majority of cases, is the independent emission of the excited center.

The main content of the article is reduced to a description of the properties and kinetics of recombination luminescence in crystals. First, however, we shall briefly dwell on the luminescence of crystals possessing luminescence of discrete centers.

3. LUMINESCENCE OF DISCRETE CENTERS IN CRYSTALS

Luminescence of this type is possessed by a large number of inorganic and organic crystals whose lattice consists of particles relatively weakly bound to one another, or includes well-isolated emitting centers protected from the action of the external molecular field. The luminescence centers are individual ions, molecules, or complexes of molecules. In the simplest cases, emission from the excited center occurs without the influence of the surrounding medium.

The following, highly characteristic properties may serve for identifying luminescence of discrete centers.

a. Preservation of luminescence upon transition of a substance into another aggregate state

Since luminescence of discrete centers is the result of processes localized in isolated centers, the separation of the latter from one another upon dissolution or upon evaporation of the substance should not substantially change, and does not change, the luminescence. Indeed, such substances as benzene, anthracene, and many others luminesce in all aggregate states. However, substances whose luminescence centers decompose upon heating before the substance passes into the gaseous state luminesce only in the solid state. Such compounds include, for example, platino-cyanide compounds, which possess luminescence in the crystalline state but decompose upon heating even before transition into the gaseous state.

b. Relation between absorption and emission spectra

Owing to the commonality of the sites of absorption and emission, the absorption and emission spectra of substances giving luminescence of discrete centers exhibit a similar structure and a regular relation. This relation is most clearly manifested in those cases where the luminescence spectrum is a simple reversal of the absorption spectrum and has a resonance character. Also very characteristic for the luminescence of discrete centers, observed in many substances, is the emergence of mirror symmetry of both spectra with respect to several—

of the frequency corresponding to the electronic transition$^{3,4}$ when the emission and absorption spectra are plotted as functions of frequency. This symmetry indicates an identical structure of the lower system of vibrational levels of the unexcited state of the luminescence centers and of the upper system—the excited state$^{5}$. The regular distortion of mirror symmetry observed in some cases corresponds to a regular change in the vibrational levels of the upper system relative to the vibrational levels of the lower system. Mirror symmetry, or mirror similarity, of absorption and emission spectra is found in benzene, naphthalene, anthracene$^{3,6,7}$, in very many aromatic polycyclic compounds$^{8}$, in a number of dyes$^{3,4}$, in polyenes$^{9}$, in the spectra of uranyl compounds$^{10}$, diamonds, etc.$^{6,7,8,9,10}$ Since, as indicated above, the luminescence of discrete centers is preserved, and often even enhanced, in solutions, the relation between the spectra is sometimes more conveniently studied in solutions. Figure 1 gives mirror-symmetric spectra of absorption (on the right) and emission (on the left) of an aqueous solution of rhodamine 6G$^{4}$.

Fig. 1. Mirror symmetry of the absorption and emission spectra of aqueous solutions of rhodamine 6G. The emission-spectrum curve gives, in relative measure, the distribution of emitted quanta by frequency.

b. Duration of luminescence and laws of decay

The spontaneous luminescence of already excited discrete centers does not require any additional activation energy for its occurrence. If the corresponding electronic transition is among the allowed ones, then luminescence proceeds in billionths of a second. If the transition is forbidden, then long-duration luminescence is observed, whose duration sometimes reaches several seconds. In the absence of quenching phenomena, both processes proceed in accordance with the differential relation:

\[ dn=-\alpha n\,dt, \tag{1} \]

in which \(n\) is the number of excited centers, \(\alpha\) is the probability of emission; it depends neither on the time \(t\) nor on the number of excited centers \(n\).

Relation (1) leads to an exponential law of decay:

\[ I=I_0 e^{-\frac{t}{\tau}}, \tag{2} \]

where \(\tau=\frac{1}{\alpha}\) is the mean duration of luminescence.

For a long time there existed the opinion that \(\tau\) depends on the viscosity of the medium, gradually increasing with increasing viscosity and, in particular, upon solidification of the substance.

S. I. Vavilov and the author showed\(^{11}\) that there is no gradual change in the duration of luminescence when a substance passes from the liquid state to the solid one. In solidified solutions and in crystals there may be the simultaneous existence of two or more luminescences differing greatly in duration. Thus, the total luminescence \(I\) of solidified solutions or crystals can be represented by a sum of exponentials:

Fig. 2. Scheme of energy levels in substances possessing luminescence of discrete centers.

\[ I=I_{01} e^{-\frac{t}{\tau_1}}+I_{02} e^{-\frac{t}{\tau_2}}+\ldots \tag{3} \]

The first luminescence \(I_{01}\) decays in billionths of a second \((\tau_1 \simeq 3\cdot 10^{-9}\ \text{sec.})\), the second \(I_{02}\)—in thousandths of a second, but sometimes lasts several seconds \((\tau_2 \simeq 1\ \text{sec.})\).

The elementary scheme of the occurrence of these processes is given in Fig. 2. If the process takes place only between two systems of levels \(I\) and \(II\), then the absorption of light and the excitation of the center

LUMINESCENCE OF ACTIVATED CRYSTALS

correspond to transition \(1\) (Fig. 2a), and emission—to the reverse transition \(1'\). If the transitions under consideration are allowed, the substance has strong absorption, and the lifetime of the center in the excited state \(II\) is limited to billionths of a second. An example of such luminescence may be the luminescence of anthracene crystals. If transitions \(1\) and \(1'\) are forbidden, then absorption is negligible, while the duration of luminescence increases, reaching several seconds. Such is, for example, the luminescence of uranyl salts (\(\tau \simeq 10^{-4}\) sec.), salts of the rare earths (\(\tau \simeq 10^{-2}\) sec.), and solid oxygen (\(\tau \simeq 18\) sec.).

Often, however, the process is complicated by the existence of a third system of levels. Let us consider this scheme of three systems of levels. Suppose that the transition from levels \(I\) to levels \(II\) is forbidden, while the transition to levels \(III\) is allowed. All possible transitions are shown in Fig. 2b by arrows. Because of the prohibition, transitions \(1\) and \(1'\) are improbable, and the principal significance is acquired by transitions \(2\) and \(2'\), corresponding to strong absorption and short-lived emission. Transitions \(3, 3', 3'', 3'''\) correspond to forced emission. In the case under consideration, the filling of the metastable levels \(II\) will occur mainly through the system of levels \(III\). Transition \(3''\) requires the expenditure of additional energy, which is supplied to the excited center from outside by the addition of thermal energy from the medium or by absorption of an additional light quantum. As a result, transition \(3''\) depends strongly on the temperature of the medium; its probability is determined by the expression:

\[ w = ae^{-\frac{E}{kT}}, \tag{4} \]

where \(E\) is the difference in energy of levels \(II\) and \(III\).

Similar forced luminescence is observed in chromium-aluminate crystals (\(\mathrm{Al_2O_3 \cdot Cr_2O_3}\)), in salts of the rare earths, and in a number of other cases. At low temperatures transition \(3''\) becomes impossible, and transition \(1'\) acquires a significant role in the luminescence.

Thus, in crystals of the type under consideration, possessing three systems of levels, three kinds of absorption are possible: strong short-wavelength absorption \(2\) (the same as \(3\)), very weak absorption \(1\), and still weaker absorption \(3''\) \((h\nu_2 > h\nu_1 > h\nu_3)\), and three kinds of emission: short-wavelength short-lived emission \(2'\), the same short-wavelength forced emission \(3'''\), whose duration is determined by the probability \(w\) of transition \(3''\), and long-lived long-wavelength spontaneous luminescence \(1'\) \((h\nu'_2 = h\nu'''_3 > h\nu'_1)\).

In the preceding considerations, quenching processes and the corresponding nonradiative transitions were not considered—

...processes from the excited state to the unexcited state, which considerably complicate the picture of luminescence.

The class of luminescence of discrete centers includes all cases of luminescence of organic crystals studied up to the present time, the luminescence of pure tungstates and molybdates, uranyl salts, salts of rare-earth elements, chromaluminates, platinum-cyano compounds, and many other inorganic substances. The processes of luminescence of discrete centers sometimes develop as one of the types of radiation also in crystal phosphors, to the consideration of whose luminescence we now turn.

4. LUMINESCENCE OF ACTIVATED CRYSTALS. COMPOSITION AND STRUCTURE OF CRYSTAL PHOSPHORS

The luminescence of discrete centers belongs to the substance of the crystals themselves. However, in many cases crystals acquire the ability to luminesce only as a result of the formation of disturbances in the crystal lattice. In such crystals there arises luminescence of an entirely different type. To form disturbances of the lattice, in most cases special impurities, called activators, are introduced into the crystals in negligible amounts; the kind of luminescence caused by the presence of an activator we shall call luminescence of activated crystals. This luminescence has a complex composition and consists of several separate emissions, each of which has its own characteristic kinetics. We shall consider separately the various luminescences of this type and their properties, and also briefly describe certain classes of activated crystals, called, for brevity, crystal phosphors.

An ideal crystal lattice possesses periodicity and does not create conditions for the development of long-duration luminescence. The absorption by such a lattice of a light quantum of not very great magnitude causes a displacement of an electron and, as a consequence, the formation at its former position of a positive charge. Thus, at the place of absorption a special state of the substance is formed, characterized by the coexistence of a mutually bound electron and a positively charged site, often called a positive “hole.” This state was first considered by Frenkel^12 and was named by him an exciton. The position of the exciton is unstable—in an undisturbed lattice it continuously moves; thus the excitation energy passes from one place in the crystal to another.

More or less prolonged localization of the exciton can occur only at sites of lattice disturbance. The radiation is accompanied by annihilation of the exciton and, in the simplest case, has a frequency equal to the frequency of the exciting light. Such resonant...

LUMINESCENCE OF ACTIVATED CRYSTALS

...such cases of excitation and emission are observed in crystallophosphors extremely rarely at low temperatures.

The excitation of a crystallophosphor, produced by light quanta of sufficient magnitude, may be accompanied by the complete departure of an electron from its former place in the lattice, in particular, in the case of ionic crystals—from the lattice anion. This departure constitutes an internal photoelectric effect. Its occurrence may be represented in the following way. Upon absorption of a photon, the electron acquires great energy and rises to a high excitation level; however, the high excitation levels of neighboring ions in the crystal overlap one another and form a continuous band extending throughout the entire crystal. In energy terms this band has considerable width, being composed of a continuous series of an enormous number of possible energy levels, the number of which is determined by the number of interacting ions of the crystal lattice. Owing to the high energetic position of these levels, they do not hold electrons; excited electrons that have reached them can move throughout the crystal; in an electric field their motion creates a current; therefore the system of upper unfilled levels is called the conduction band of the crystal.

The velocity of motion of electrons in the conduction band is determined by the temperature of the crystal and, for free motion of the electron, should be measured in tens of kilometers per second. However, its magnitude is greatly reduced because an electron moving in the conduction band causes polarization of the ions past which it passes; the displacement of the electron in the conduction band is accompanied by displacement of the site of polarization, which slows the motion of the electron. Such a moving system consisting of an electron and a polarized region of the lattice was considered by S. Pekar and was named by him a polaron^13. Recombination of an excited electron with one of the ions formed upon excitation of the substance also leads to emission, and, owing to the great speed of motion of electrons in the conduction band, the recombination process proceeds extremely rapidly. Thus, an ideal crystal lattice is characterized only by brief luminescence. Meanwhile, in experiments with crystals, prolonged luminescence is very often observed, which is the characteristic luminescence of crystallophosphors.

For prolonged luminescence to arise, a disturbance of the periodicity of the electric field of the crystal lattice is necessary. Such disturbances may arise in any crystal lattice. They occur as a result of irregularities in heating and cooling during crystallization, and also under the action of X-rays and short-wave ultraviolet rays. In these cases, some of the ions of the crystal acquire or give up

additional charge and become neutral atoms, becoming special sites of the lattice. The neutral atoms thus formed can serve as centers of localization of electrons, i.e., as those sites around which electrons moving in the conduction band can stop for a more or less prolonged time, falling into potential wells. Individual empty nodes of the lattice may also serve as special sites of the lattice. While being delayed at localization sites, electrons recombine later. Thus, the appearance of electron-localization sites leads to the occurrence of long-duration luminescence of crystal phosphors.

The luminescence that arises upon the reunion of an electron with an ion usually belongs to the same centers that disturb the regularity of the crystal lattice. These centers thus play both the role of localization sites and the role of emitting centers. Apparently, localization may occur at unexcited centers, whereas luminescence occurs only upon recombination of electrons with excited centers, i.e., with centers that have lost a negative charge.

One of the most extensively studied cases of activation of a crystal lattice is the formation of neutral zinc atoms in the lattice of zinc sulfide. The appearance of neutral zinc atoms in the ZnS lattice causes the occurrence of blue luminescence with a duration on the order of tenths of a second. However, in the overwhelming majority of cases the activator that changes the electric field of the crystal lattice is a metal foreign to the lattice. Ions of heavy metals and rare-earth elements usually serve as activators. They are introduced into the crystal lattice by calcining their salts with the basic substance of the lattice, for example with zinc sulfide.

There are two types of metal activators. Some of them are introduced into the lattice in negligible amounts, of the order of \(10^{-4}\) g per gram of charge; activators of this type include Cu and Ag. Other activators, on the contrary, are introduced into the crystal in large amounts and form mixed crystals with the basic substance; such activators are Mg, Pb, and Sn. During calcination, in addition to the activator, a flux is usually added to the charge in the form of one or several low-melting salts. The flux serves to improve the conditions of crystallization and to change the optical properties of the phosphor. NaCl, \(Na_2SO_4\), \(CaF_2\), and others are used as fluxes. The composition of a phosphor is written as follows: in the first place the basic substance is indicated, in the second—the activator, in the third—the flux (for example, \(CaS \cdot Bi \cdot CaF_2\)).

At present there is every reason to suppose that the luminescence center formed during calcination has a complex structure. Besides the activator ion, other components must also enter into its composition. This conclusion follows from the fact that

circumstance that, when different salts of one and the same activator are used to prepare a phosphor, the intensity of luminescence, and sometimes also the spectral composition of the radiation, prove to be different. In exactly the same way, the nature of the flux strongly influences the properties of the luminescence—its duration and the spectral composition of the radiation.

Thus, it appears highly probable that the luminescence center consists not only of a metal ion, but also of the anions of the metal salt and of the flux.

When several activators are used, phosphors are formed whose properties cannot be obtained by a simple summation of the properties of single-activator phosphors. In many cases one must admit the formation of complexes from two or several activators. Such a process is manifested especially sharply in the activation of phosphors by rare earths.

The use of different fluxes substantially changes the physicochemical properties of phosphors, in particular, their hardness. With the introduction of some fluxes, the phosphors obtained are extremely hard; with the introduction of others, they are readily amenable to grinding.

This property is technically very important, since phosphors are used in the form of a finely dispersed powder.

The optical and physicochemical properties of phosphors also depend on the method of preparation, on the temperature and duration of calcination, and on the atmosphere in which calcination is carried out (oxidizing, reducing, an atmosphere of neutral gases, etc.). Raising the temperature and increasing the calcination time cause an increase in the size of the crystal. Calcination temperatures from 700 to 1400° C and calcination times from 15 to 30 minutes are usually used. Under these conditions, different phosphors form crystallites of different sizes, from 1 μ to 100 μ in cross section.

The luminescence of crystals is undoubtedly connected with the formation of the complex centers considered above and with the provision of the possibility of electron motion inside the crystal lattice. Therefore internal shifts that destroy the luminescence centers and disturb the regularity of the lattice must lead to the destruction of the crystal’s luminescence. This phenomenon is indeed observed and is called quenching by crushing.

Fig. 3 gives the Debye diagram of a ZnS·Cu-phosphor crystal[^14]. At the top is the Debye diagram of a sample of an undestroyed crystal; at the bottom, that of a sample subjected to grinding for 80 min. The short arrows indicate the lines of the wurtzite structure; the long ones mark the lines of the sphalerite structure. Only the latter remain after grinding; thus, grinding leads to the transition of the wurtzite structure of ZnS into the sphalerite structure.

What is especially important, however, is not the change in the crystal structure (both modifications give luminescence), but the blurring of the interference rings and the complete disappearance of the rings of higher orders in ground crystals. This blurring and disappearance of the rings indicate a violation of the periodicity of the lattice, the appearance of internal displacements, with which the decrease in the luminescent ability of the phosphor should be associated.

The phenomenon of quenching upon grinding is very widespread and is observed to a greater or lesser degree in all crystalline phosphors. Meanwhile, in the practical use of crystalline phosphors it is necessary to carry out their strong comminution.

To restore the luminescence of phosphors, the ground powder is subjected to a secondary, brief calcination, which is called regeneration. The duration of regeneration usually

Fig. 3. Debyegram of a ZnS·Cu phosphor. Above—an undestroyed crystal; below—a ground one. Long arrows mark the structure of sphalerite, short ones—wurtzite.

does not exceed 5 min. Heating is carried out to temperatures 200–300° C lower than the temperatures of the initial calcination.

As the base substance for phosphors, zinc sulfide, zinc silicate, sulfides of alkali metals, and phosphates of alkali metals are used especially often. Recently, bases made of mixed crystals have come into practice, for example ZnS·CdS, CaS·SrS, ZnBeSiO₄, etc. The indicated changes in bases usually lead to a change in the spectral composition of the radiation, gradually shifting it to one or another side of the spectrum. In addition, in crystalline phosphors with a mixed base, the hardness, hygroscopicity, moisture resistance, and other physicochemical properties change. By selecting the appropriate composition and technology for preparing crystalline phosphors, it is possible to obtain substances satisfying very diverse technical requirements, sufficiently resistant to various influences, and giving luminescence of different colors and durations.

5. ON THE ABSORPTION AND EMISSION SPECTRA OF CRYSTAL PHOSPHORS

a. Absorption spectra

Obtaining quantitative data on the absorption of crystal phosphors is difficult because individual crystals are small and because crystal layers strongly scatter light. When light passes through such layers, the principal attenuation of the transmitted luminous flux occurs as a result of scattering, rather than absorption. Only recently have quantitative absorption measurements begun to be carried out on the thinnest films of sublimate phosphors, as well as on individual crystallites with the aid of a special microscope1.

Absorption is composed of absorption by the host substance, absorption by the lattice deformed by the introduction of the activator, and absorption by the activator itself.

The host substance for crystal phosphors is usually colorless salts that absorb in the middle and far ultraviolet part of the spectrum. The deformed sites of the lattice give an absorption spectrum shifted somewhat toward longer wavelengths in comparison with the spectrum of the pure crystal.

The introduction of an activator creates additional absorption, usually lying in an even longer-wavelength part of the spectrum and often extending into the visible region. Thus, for example, the fundamental absorption of zinc sulfide ends at about 330 mμ. The absorption of zinc sulfide activated with manganese extends to 370 mμ. In addition, at a significant manganese content, absorption bands lying in the blue part of the spectrum appear rather sharply. Apparently, for ZnS·Mn phosphors one should distinguish three absorption regions: up to 330 mμ lies the absorption region of pure zinc sulfide; from 330 to 360 mμ lies the region of additional absorption corresponding to the altered lattice of the crystal phosphors; and in the visible region are located the absorption bands of manganese itself.

Excitation of the phosphor in one of the three indicated regions leads to different luminescence. Thus, for example, excitation of ZnS·Mn in the first two regions gives both short-duration and long-duration luminescence, whereas excitation in the visible region produces only short-duration luminescence2.

Figure 4 schematically shows the absorption and emission regions of a Zn₂SiO₄—Mn₂SiO₄ (1 mol.%)-phosphor. The dotted line separates the region of fundamental absorption of Zn₂SiO₄, extending approximately to 2700 Å. The difference in absorption, determined by the distance between the solid and dotted curves in the region from 2100 to 3000 Å, is due to the fundamental absorption of Mn₂SiO₄.

Absorption between 3000 and 5000 Å, consisting of separate bands, belongs to Mn\(^{++}\). The hatched curve is the emission of Mn\(^{++}\).

Phosphor in the excited state, as the most recent experiments have shown, has additional absorption, concentrated mainly in the infrared region of the spectrum. This absorption is caused by excited electrons localized near various sites of lattice disturbance; it will be considered in somewhat greater detail below.

Fig. 4. Position of the absorption and emission bands of the Zn\(_2\)SiO\(_4\)—MnSiO\(_4\) (1 mol %) phosphor.

6. Emission spectra

The emission spectra of crystalline phosphors do not resemble absorption spectra. In most cases they consist of rather broad bands, of the order of 100 mμ, having a symmetrical form. When the luminescence intensity is expressed as a function of frequency, these curves very often fit well a Gaussian error curve. The occurrence of curves of a more complicated form usually indicates a complex composition of the emission and the superposition of several separate luminescence bands. They can be separated from one another by studying the decay of the luminescence and isolating the individual bands with light filters. The different spectral regions of such complex bands usually decay at different rates.

The principal color of the luminescence is determined by the activator, with each activator having its own characteristic band. Such are the green band of copper, the blue band of silver, and the orange band of manganese in ZnS phosphors.

Some activators in one and the same phosphor may give two or three emission bands, each of which differs in spectral position, temperature, and other properties. Thus, for example, Bi in alkaline-earth phosphors gives two emission bands: one violet, strongly developed at com—

one at room temperature, and another, red, appearing at low temperature. Mn also gives two emission bands: an orange one, whose optimal luminescence is reached at a temperature of about 100°C, and a dark-red one, which develops fully at low temperatures.

The indicated specificity of the bands for different activators—heavy metals—makes it very probable to conclude that the activator is the emission center; this is fully confirmed when considering the luminescence of phosphors activated by rare-earth elements. These phosphors give line spectra of emission. The spectra consist of many lines characteristic of trivalent ions of the corresponding rare earths. Thus, in the present case it proves quite obvious that the emitting center is the rare-earth ion. Since all the other properties of phosphors activated by rare-earth elements fully coincide with the properties of phosphors activated by heavy metals, the conclusion made for phosphors with rare earths may with full justification be extended also to phosphors activated by heavy metals, which possess broad bands in their emission spectra. Figure 5 gives the spectra of phosphors activated by samarium. In different hosts, different samarium lines arise.

Fig. 5. Luminescence of Sm ions:
a—at \(+25^\circ\mathrm{C}\), b, c, and d at \(-150^\circ\mathrm{C}\).
Hosts: a and b—MgS, c—CaS, d—\(\mathrm{La_2(SO_4)_3}\).

The emission bands of different phosphors are not equally stable with respect to external influences and, above all, to the influence of the electric field of the crystal lattice. The well-shielded electrons of the \(4f\) shells in rare-earth elements, as well as in transition metals, give stable emission spectra whose positions almost do not change when the host is changed. Thus, for example, the influence of the host on the luminescence of the trivalent samarium ion is manifested in a change in the relative intensity of the different lines and leads to the appearance of new lines or the disappearance of former ones, but almost does not shift the lines on the frequency scale.

The orange band of manganese also possesses considerable stability. However, a change of the host may cause the appearance

of the other Mn band. Thus, in silicates manganese gives not orange radiation (\(\sim 5850\) Å), but green radiation (\(\sim 5200\) Å).

The luminescence bands of copper and silver strongly change their position depending on the host. Thus, with a successive change in the ratio of ZnS and CdS in mixed ZnS·CdS crystals, as the CdS content increases, the copper band shifts from the green to the dark-orange part of the spectrum, while the silver band shifts from the blue part of the spectrum to the green. An example of the effect of the host on the emission of the activator is given in Fig. 6. Shown here are the emission and absorption spectra of CaSiO\(_3\)·Pb and the emission spectrum of CaWO\(_4\)·Pb phosphors. The emissions of the two phosphors lie in entirely different parts of the spectrum.

Fig. 6. Absorption spectrum (\(A\)) and emission spectrum (\(B\)) of CaSiO\(_3\)·Pb and emission spectrum (\(C\)) of CaWO\(_4\)·Pb, excited by the line 2537 Å.

The data presented above on absorption and emission spectra indicate their substantial difference. These spectra differ in form and are not connected in position in any way. The absence of a connection between the two spectra is quite natural, since the principal absorption is concentrated in the host lattice or in the deformed lattice, whereas emission occurs at the activator and, consequently, is determined by the properties of the activator. Only in the case where absorption is produced by the activator itself can one expect some correspondence between the two spectra. Such cases are sometimes observed in activation by rare earths, when resonance absorption and emission bands arise. By the absence of a connection between the absorption and emission spectra, the luminescence of crystal phosphors differs sharply from the luminescence of discrete centers.

One should note one more essential feature of the absorption and emission spectra of crystal phosphors. As was indicated, the absorption of crystal phosphors lies mainly in the middle and far-

easy ultraviolet part of the spectrum. The emission, however, is concentrated in the visible part of the spectrum. As a result, under optical excitation it is necessary to use quanta of large magnitude, at the expense of which small quanta of visible luminescence arise. Thus, optical excitation of crystallophosphors is accompanied by large Stokes energy losses. Thus, for example, upon excitation of the 5200 Å band of zinc silicate by means of the resonance mercury line 2537 Å, the Stokes losses amount to more than 50%. Upon excitation of the orange luminescence of manganese with a mean wavelength \(\lambda = 585\ \mathrm{m}\mu\), by the mercury line with \(\lambda = 366\ \mathrm{m}\mu\), the Stokes losses amount to about 40%.

The absorption spectra and emission spectra of crystallophosphors rarely overlap one another. Thus, in most cases secondary absorption is absent. However, when thick layers of phosphor are used, a weak reabsorption of the luminescence light can sometimes be observed. Such reabsorption occurs in zinc-sulfide phosphors and in some alkaline-earth phosphors, and is practically completely absent in zinc silicates and tungstates.

c. On the vibrational structure in the absorption and emission spectra of crystallophosphors

The frequencies of the emission spectra of discrete centers are composed of the frequencies of the electronic transition and of the frequencies of vibrations and rotations of the nuclei. This additive character of the spectra is manifested in the existence of a coarse structure, corresponding to vibrational processes, and a fine structure, corresponding to rotational processes. This structure, which is blurred at high temperatures and in strong electric molecular fields, usually appears well at sufficiently low temperatures or when studying the absorption and luminescence of rarefied gases.

The absorption and emission spectra of crystallophosphors are, as a rule, diffuse. This is promoted by the strong electric fields in which the emission centers are found. Only in well-shielded atoms of rare-earth elements and in transition elements do the spectra acquire a line or narrow-band structure. In these cases it is possible to establish the presence of lines whose frequencies are composed of the frequency of the electronic transition and the vibrational frequencies of the crystal lattice.

The presence of vibrational structures has been detected in chromaluminates \((\mathrm{Al}_2\mathrm{O}_3 \cdot \mathrm{Cr}_2\mathrm{O}_3)\) and in compounds related to them, in one of the types of long-duration luminescence of rock salt, in the luminescence of diamonds, and also in the luminescence of activated oxides of metals of the second group: \(\mathrm{CaO}\), \(\mathrm{ZnO}\), etc., at low temperatures. In Fig. 7 are given—

...serves, for example, as a spectrum with vibrational structure, the spectrum of CaO^18 contaminated by a small impurity of Bi.

Elements of structure in absorption spectra were discovered by V. M. Kudryavtseva, F. I. Vergunas, and P. S. Litvinova^19 in studying the absorption of zinc-oxide films. The indicated, comparatively few, cases of the manifestation of vibrational structure undoubtedly prove, however, the superposition of vibrational spectra on the spectra of electronic transitions. This superposition, in most cases, is completely masked by the extraordinary broadening of the emission bands, which occurs under the action of the internal electric fields of the crystal.

Fig. 7. Vibrational structure in the luminescence spectra of CaO in the presence of a small impurity of Bi (Bi:Ca = 10^-4).

5. LUMINESCENCES OF DIFFERENT DURATION.

RISE AND DECAY OF THE LUMINESCENCE OF CRYSTAL PHOSPHORS

a. Short-duration and long-duration luminescences

Complex crystal phosphors with several activators may possess a large number of spectrally distinct emission bands; however, we shall not consider these phosphors for the time being. But even in phosphors with a single activator the composition of the luminescence is complex. A characteristic property of crystal phosphors is the simultaneous existence, in the majority of them, of at least two luminescences of identical spectral composition but of different duration: one luminescence is short-lived and ends completely in hundredths of a second, and often even faster; the other has a considerable duration—it sometimes continues for an hour or more. In some phosphors not one, but several short-lived luminescences are found.

The ratio between the intensity of the short-lived and the long-lived luminescence depends on the mode of excitation and, above all, on the intensity of excitation. As the excitation intensity increases, the long-lived luminescence tends toward saturation; thus its brightness cannot exceed a certain limiting...

LUMINESCENCE OF ACTIVATED CRYSTALS

quantities; moreover, this limiting brightness is not high. At the same time, the brightness of the short-duration luminescence of a phosphor, as the intensity of excitation is increased, grows practically without bound and in some cases, for example under excitation by a cathode beam, reaches the brightness of the crater of a voltaic arc.

There are several types of instantaneous luminescence, differing in nature and in kinetics. One kind of short-duration luminescence is luminescence arising as the result of the direct excitation of activator ions. The properties of such luminescence are determined by the general laws of the luminescence of discrete centers, with allowance for those specific features which the crystalline medium surrounding the emission center creates. However, much more often short-duration luminescence, like long-duration luminescence, has a recombination character, i.e., it is the luminescence of ions excited at the moment of recombination of free electrons and ionized centers.

The existence of short-duration and long-duration luminescence of phosphors indicates the existence of two different processes of recombination of excited electrons. In considering the luminescence of real phosphors, in the simplest case three forms of development of the luminescence processes are encountered, shown in Fig. 8. The continuous curve in Fig. 8a depicts the development and decay of luminescence in the presence of one luminescence (the case of an exponential process is taken). Fig. 8b corresponds to the independent development of instantaneous and short-duration processes. The instantaneous luminescence reaches full intensity immediately after the beginning of excitation; this is reflected in Fig. 8b by the fact that the brightness of the luminescence immediately after the beginning of excitation reaches the value a; the long-duration luminescence increases gradually. Upon cessation of excitation there occurs a decline of the instantaneous luminescence, as rapid as the rise at the beginning of excitation, and a slow weakening of the long-duration luminescence. In Fig. 8c the development of luminescence is shown in the presence of long-duration and short-duration luminescences connected with one another. Here the intensity of both luminescences only gradually, as the long-duration luminescence develops, reaches its final value, but upon cessation of excitation the short-duration luminescence dies out immediately, while the long-duration luminescence decays slowly. At the moment excitation ceases, a sharp drop in luminescence brightness is observed, caused by the cessation of the short-duration luminescence.

The study of the development of all kinds of luminescence from the beginning of excitation to its cessation and of the decay of luminescence after the cessation of excitation is extremely important for elucidating the entire kinetics of the luminescence processes of crystalline phosphors, and also for many practical purposes. Therefore we shall dwell on this question in somewhat greater detail.

Fig. 8a. Rise and decay of luminescence under prolonged excitation (case of an exponential decay law).

Fig. 8b. Rise and decay of the total luminescence of the phosphor under the simultaneous existence of mutually independent short-duration and long-duration luminescences.

Fig. 8c. Rise and decay of the total luminescence of the phosphor under the simultaneous existence of mutually connected short-duration and long-duration luminescences.

6. On the laws of the rise and decay of luminescence

The simplest case of decay is decay according to the exponential law (2). From the differential relation (1) it is seen that in this case all excited luminescence centers \(n\) have the same probability of emitting light, and this probability of their emission does not depend on time. In the case of exponential decay, the rise of luminescence after the onset of excitation also occurs exponentially:

\[ I = I_0 \left(1 - e^{-\frac{t}{\tau}}\right). \tag{5} \]

Here \(I_0\) is the brightness of the luminescence at \(t=\infty\), i.e. after the establishment of the stationary state at the given excitation intensity. Expression (5) is written for the case of weak excitation, in which only an insignificant fraction of all centers is simultaneously in the excited state. At such small excitation intensities the duration of establishment of the stationary luminescent state does not depend on the excitation intensity.

If the excitation is stopped, then the decay will proceed in accordance with formula (2), with the time \(t\) being counted from the moment the excitation ceases; \(I_0\), entering into formula (2), will be equal to the intensity that the luminescence reaches by the moment the excitation is stopped. In Fig. 8a the course of the rise of this luminescence under excitation and its decay after cessation of excitation is shown. The shaded areas are equal. The shaded area \(S\) indicates the magnitude of the radiant energy accumulated by the phosphor during excitation.

Under strong excitations, which within a time less than \(\tau\) bring a large part of the luminescence centers into the excited state, saturation of the luminophore may be attained, i.e. its maximum excitation, in which practically all luminescence centers are simultaneously in the excited state. Such a state must be reached in a time less than \(\tau\), so that the excited system of centers does not have time to emit light again. The rate of rise of luminescence in this case is determined by the excitation intensity; the establishment of the stationary regime corresponding to the maximum excitation of the phosphor will occur the sooner, the more intense the excitation and, in any case, in a time much smaller than \(\tau\).

In crystallophosphors the exponential law should occur under direct excitation of the luminescence centers, which is observed comparatively rarely. In addition, the exponential law of decay also arises in those cases of recombination luminescence when the number of interacting centers of one kind

significantly exceeds the number of centers of the other kind, for example in the recombination of a small number of free electrons from the conduction band with a large number of ionized lattice centers. In this case the differential equation for the kinetics of the process has the form:

\[ dn=-pNn\,dt, \tag{6} \]

where \(p\) is the probability of recombination, \(n\) is the number of electrons, and \(N\) is the number of ions, which remains practically constant during the decay, since it is much greater than \(n\). From (6) it follows that:

\[ n=n_0 e^{-pNt}. \tag{7} \]

Both of the cases considered of crystal-phosphor luminescence, in which the occurrence of exponential processes is possible, pertain to short-duration luminescence. Long-duration luminescence never gives a simple exponential course.

Fig. 9a. Decay of uranyl salts in the coordinates \(I^{-1/2}, t\).

The importance of studying decay laws for establishing the type of luminescence kinetics may be illustrated by the following example. Proceeding from a preconceived point of view on the recombination character of the luminescence of uranyl salts, Nichols and Howes20 gave a graphical representation of their experimental results, plotting \(I^{-1/2}\) as a function of time (Fig. 9a). From the presence of breaks in their curves they concluded that there existed a whole series of recombination processes, occurring simultaneously during the decay of the luminescence. A recalculation of the data of Nichols and Howes by the exponential formula (2), carried out in the work of S. I. Vavilov and the author, completely changed the picture21. In Fig. 9b the course of the decay is given in the coordinates \(\lg I\) and \(t\). The decay in all cases proved to be exponential; the luminescence process must be assigned to the type of luminescence of discrete centers.

The question of the kinetics of the luminescence of typical crystal phosphors remained unclear for a long time. At the end of the 1860s of the last century E. Becquerel22 investigated the luminescence of certain crystal phos-

Luminescence of Activated Crystals

phors and arrived at the following empirical formula for the decay:

\[ i^m(t+c)=\mathrm{const}, \tag{8} \]

where \(i\) is the luminescence intensity, \(t\) is the time elapsed after the cessation of excitation, and \(c\) and \(m\) are constants.

This hyperbolic dependence between time and luminescence intensity indicates the recombination character of the process. However, E. Becquerel’s old experiments, carried out with substances of indefinite composition containing various impurities, could not be regarded as definitive for judging the kinetics of the process.

Fig. 96. Decay of uranium salts in the coordinates \(\lg I, t\).
The figure proves the exponential course of the decay.

Later numerous investigations of crystallophosphors, carried out by P. Lenard and his school\(^{23}\) over more than 30 years (1894–1928), led them to the idea that within crystallophosphors there exist “luminescence centers”—special formations isolated from one another, sometimes containing hundreds of thousands of ions. It should be stated at once that, despite the common terminology, Lenard’s luminescence centers are in no way similar to the luminescence centers of the modern theory of luminescence.

The isolated character of Lenard’s centers naturally led to the idea of the independence of their luminescence and, hence, to the requirement of an exponential decay. The observed

on experiment, the complex course of decay was explained by Lenard’s school by the existence in the phosphor crystal of centers of different sizes and by the superposition of exponential processes of different durations. Lenard’s theory, however, did not provide a quantitative description of luminescence and in many respects seemed artificial. Therefore verification of its propositions appeared necessary.

Fig. 10. Decay of CaS·Bi phosphor at different intensities of the exciting light \(J_{\mathrm{B}}\) and at a temperature of \(52^\circ\) C. The solid curves are drawn from experimental data; the points are calculated by formula (10).

Owing to the importance of investigating the laws of decay for establishing the kinetics of luminescence, V. V. Antonov-Romanovskii and the author undertook a series of studies of the decay of various types of phosphors \(^{24,25,26}\) and others.

The studies carried out showed that, beginning from a certain moment after cessation of excitation, the decay of luminescence is well expressed by the empirical relation

\[ I = A t^{-\alpha}, \tag{9} \]

where \(A\) is the brightness of luminescence at \(t=1\), and \(\alpha\) is a constant characterizing the rate of decrease of brightness.

Formula (9) gives exaggerated values of the luminescence brightness for very small values of \(t\); as \(t \to 0\), \(I \to \infty\).

Practically complete agreement with experiment can be obtained by using the formula

\[ I = A(a+t)^{-\alpha}, \tag{10} \]

which is a slight modification of Becquerel’s formula. As was recently shown by L. A. Vinokurov, E. G. Baranova and the author \(^{27}\), the constants \(\alpha\) and \(a\) of this formula, with the corresponding method of determining them, characterize: \(\alpha\)—the rate of decrease of brightness in the later stages of decay, and \(a\)—the duration of the initial interval during which the decay proceeds—

decays more slowly than follows from formula (9). The smaller this interval, the smaller the value of \(a\).

In Fig. 10 the experimental decay curves of the CaS·Bi phosphor are shown on a double logarithmic scale. The points plotted on them were calculated by formula (10).

In logarithmic coordinates formula (9) gives a linear dependence between \(\lg I\) and \(\lg t\):

\[ \lg I=\lg A-a\lg t. \tag{11} \]

The linear course sets in the later, the weaker the excitation and the thicker the phosphor layer*).

The hyperbolic course of the decay thus obtained testified to the recombination character of the processes of luminescence of crystalline phosphors.

This conclusion was also confirmed by a number of other experiments, some of which will be described below. At present the attribution of the luminescence of crystalline phosphors to the type of recombination luminescence may be regarded as firmly established.

It should be noted, however, that the simplest recombination scheme of direct reunification of the particles separated upon excitation does not satisfy the experimental data. The differential equation corresponding to such a process has the form:

\[ dn=-pn^{2}dt, \tag{12} \]

where \(p\) is the probability of recombination, \(n\) is the number of excited electrons and the equal number of ions. From equation (12) it follows that:

\[ \left. \begin{aligned} n&=\frac{n_{0}}{1+pn_{0}t},\\ I&=-\frac{dn}{dt}=\frac{pn^{2}}{(1+pn_{0}t)^{2}}. \end{aligned} \right\} \tag{13} \]

A decay following a hyperbolic dependence of the second order is observed very rarely in experiment. Moreover, relation (13) provides for the existence of only one process, whereas, as indicated above, in crystalline phosphors in most cases two luminescences are observed: long-duration and short-duration.

*) These circumstances are related: the first to a large number of repeated localizations of the excited electrons before their recombination; the second to secondary phenomena: nonuniform excitation of layers of different depth and partial absorption of the luminescence emerging outward from the deep layers of the phosphor.

6. STUDY OF LOCALIZATION LEVELS AND THE NATURE OF LOCALIZATION

a. The brightness of luminescence as a function of the light sum accumulated by the phosphor

The intensity of phosphor luminescence is determined by the number of recombinations of excited electrons with ionized centers per second. This quantity depends, obviously, on the number of excited electrons and on the probability of their recombinations. If it is assumed that the probability of recombination is the same for any pair of combining particles, then the state of a crystalline phosphor is entirely determined by the density of ionized centers and excited electrons present in the lattice. Each definite state of the crystalline phosphor must correspond to a definite luminescence intensity, irrespective of the manner in which the crystalline phosphor was brought into that state. Hence arises the attempt to express the luminescence intensity as a function of the light sum—a quantity proportional to the number of excited electrons contained in the phosphor. On the basis of these ideas, theoretical calculations were also made (see below). However, as the experimental investigations of the author\(^{28}\), as well as those of L. A. Vinokurov, E. G. Baranova, and the author\(^{27}\), have shown, the matter here is considerably more complicated. The luminescence intensity of an excited phosphor can indeed be expressed in terms of the density of excited electrons and is approximately described by the empirical formula:

\[ I = BS^{n}, \tag{14} \]

where \(S\) is the light sum, and \(B\) and \(n\) are constants; however, the luminescence intensity depends not only on the magnitude of the light sum, but also on the method of excitation. For one and the same phosphor, with different methods of excitation, the constant \(B\) may be different. This means that the probability of recombination of electrons with ionized centers proves to be different for different methods of excitation, i.e., different methods of excitation bring electrons into different states of excitation. In this, apparently, their places of localization, the probabilities of their release from the localization sites, and their paths to the recombination sites change.

b. The stability of localization of excited electrons

The existence of short-time and long-time luminescence of phosphors indicates the existence of two different processes of recombination of excited electrons. The existence of long-time luminescence of phosphors shows that many of the electrons,

released at the moment of excitation remain in the crystal lattice in the excited state for a very long time, which in individual cases is measured in months and very often in hours. As was indicated above, this is explained by some of the electrons falling into potential wells of greater or lesser depth, from which they cannot free themselves on their own. The liberation of electrons from localization sites is usually effected by thermal motion; liberation will be the less probable, the deeper the potential well and the lower the temperature of the body.

The depth of the localization levels is, naturally, different for different phosphors. In phosphors that give a well-developed long-duration luminescence at room temperature, it is equal to \(\simeq 0.5\) eV.

Experience shows that the various luminescence bands of one and the same phosphor, as well as the luminescence bands of different phosphors, possess different temperature stability. Some of them are quenched already at very low temperatures, for example at the temperature of liquid air; others at room temperature; still others develop only at high temperatures. The first bands are called “cold”; they correspond to small depths of the localization levels; the second are “normal,” corresponding to greater depths of the localization levels; and the third are “hot,” corresponding to very deep localization levels.

To elucidate the kinetics of luminescence of a crystallophosphor it is very important to know what localization levels it possesses, how these localization levels are filled with electrons, and by what constituent parts of the phosphor they are formed. One of the important methods for solving this question is the study of the so-called curves of temperature quenching.

c. Curves of temperature quenching

The usual quenching of an excited phosphor occurs with the gradual liberation of electrons by thermal motion from their localization sites. Obviously, lowering the temperature must cause a decrease in the number of liberated electrons and, correspondingly, reduce the brightness of luminescence. At the same time the duration of luminescence increases, since the time of liberation of the electrons is stretched out. The temperature of the phosphor can be lowered so strongly that the localized electrons will not be liberated at all and the phosphor will cease to give long-duration luminescence. Such a phosphor will be called frozen.

A phosphor located at a very low temperature, however, retains the ability to absorb energy; in this case one part of the excited electrons immediately recombines, giving

instantaneous luminescence. The other part, however, is strongly localized. This second part of the electrons can be localized at levels of different energetic depth, and release from them will occur at different temperatures. If such a phosphor, excited at a very low temperature, is then heated, with its temperature raised uniformly, then first the electrons from the shallowest levels will be released; then there may occur an interval of temperatures where the luminescence weakens, since by that time the shallow levels will already have been emptied, while the temperature will still be insufficient for releasing electrons from the deeper levels. Then, with a further increase in temperature, electrons from a second system of deeper levels will again begin to be released; subsequently there will occur the release of electrons from a third system of levels, still deeper, and so on. For a graphical description of the phenomenon, temperature is plotted along the abscissa axis, and luminescence intensity along the ordinate axis; the curves thus obtained have maxima at certain temperatures characteristic of the given phosphor. These curves are called temperature-glow curves. The maxima on a temperature-glow curve indicate the existence of localization levels of different depth, and the temperatures corresponding to these maxima characterize the depth of the localization levels. The method of investigating the depth of local levels by means of temperature-glow curves has been used by many authors, and especially by M. L. Katzem[^29]. Figure 11 shows a curve with three well-pronounced maxima. This phosphor possesses three separate systems of local levels corresponding to the temperatures.

Fig. 11. Temperature-glow curve of a KBr phosphor in the visible part of the spectrum.

It might be assumed that a completely excited phosphor contains electrons at all possible localization levels. To fill all localization levels, we would have to excite the phosphor at a low temperature, since if the phosphor is excited at a high temperature the shallow localization levels will glow out already at the moment of excitation, and accumulation of electrons will occur only at deep levels.

Experience shows that the filling of localization levels proceeds in a more complicated way. Under no excitation conditions is it possible to fill all levels. At low temperatures, indeed, shallow levels are filled, but the deep levels cannot be completely filled. Conversely, at high temperatures the shallow levels, as was to be expected, are freed already at the moment of excitation and, after excitation has ended, remain almost free; but the deep levels are filled to a greater extent than at low temperatures. Therefore the curves of thermoluminescence of one and the same phosphor, excited at

Fig. 12. Change in the thermoluminescence curve upon excitation at different temperatures. The excitation temperatures are indicated in the figure.

different temperatures, are deformed as shown in Fig. 12. For a phosphor excited at low temperatures, the glow curve has strong maxima at low temperatures and weak ones at high temperatures. For a phosphor excited at high temperatures, there is no emission at low temperatures and there are strong emission maxima at high temperatures3.

The curves of thermoluminescence also proved useful for solving the question of the nature of localization. Experiment showed that phosphors activated by two activators give thermoluminescence curves entirely different from those of phosphors activated by one activator. Figure 13 shows the thermoluminescence curves of SrS·Sm-, SrS·Eu-, and SrS·Eu, Sm-phosphors3.

We see that the curve of the phosphor activated by Eu alone has a maximum at \(\sim 90^\circ\text{C}\); the curve of the phosphor activated by samarium Sm alone has a maximum approximately at \(125^\circ\text{C}\); whereas the phosphor activated by Eu and Sm together has a maximum of thermoluminescence at \(170^\circ\text{C}\), i.e. at a temperature higher than that of each of the preceding phosphors activated by one activator. This forces us to draw two conclusions:

1. The temperature glow curve is associated with the activator, and consequently the activator ions and the spatially deformed crystal lattice surrounding them are the site of localization.

Fig. 13. Temperature glow curves of SrS·Eu, SrS·Sm, and SrS·Eu,Sm phosphors.

1. The differences between the temperature glow curve of the SrS·Eu,Sm phosphor and the curves belonging to phosphors activated by the components of the complex activator—europium and samarium separately—serve as one of the convincing proofs that in this complex phosphor the luminescence center consists not of individual Eu and Sm ions, but of a certain complex compound of both ions.

7. OPTICAL BLEACHING

a. The phenomenon of the optical flash

In the preceding section we considered the decay of luminescence caused by the gradual release of localized electrons by means of thermal motion. However, in a number of cases the release of localized electrons can also take place by irradiating phosphors with long-wavelength rays, in particular infrared rays. Experience shows that an excited phosphor possesses a certain additional absorption in the infrared region. In Fig. 14 are shown the spectra

Fig. 14. Absorption spectrum of the SrS·Eu,Sm phosphor (Eu—30, Sm—100). Along the ordinates are plotted the relative values of the absorption coefficient. The solid curves correspond to the unexcited state of the phosphor, the dashed curve—to the excited state.

absorption of the SrS·Ce,Sm phosphor in the excited and unexcited states[^56].

As is seen from Fig. 14, the excited phosphor possesses additional absorption in the neighborhood of 0.6 and 1.1 μ. It is due to excited electrons localized near special sites of the lattice. Under the action of absorbed light rays, the localized electrons are ejected from the localization levels into the conduction band and begin to recombine. This creates a flash of luminescence, which is called an optical flash.

Fig. 15. Dependence of the brightness of the optical flash \(I\) on the magnitude of the light sum \(S_B\) accumulated by the phosphor.

The phenomenon of the optical flash appears especially sharply in phosphors with very deep localization levels, which at room temperature cannot be emptied. These excited, but nonluminous, phosphors give bright luminescence when they are irradiated with infrared rays.

Among such phosphors are the alkaline-earth phosphors considered below, activated by binary combinations of rare earths.

Experiments showed that the brightness of the flash is proportional to the intensity of irradiation with infrared rays and to the square of the light sum accumulated by the phosphor at deep localization levels (Fig. 15)[^28],[^30]. After the bleaching rays cease to act, the decay proceeds according to a hyperbolic law[^30], which corresponds to the law of thermal decay of luminescence (Fig. 16).

Fig. 16. Decay of the optical flash of a CaS·SrS·CeSm phosphor. The upper curve corresponds to afterglow begun 30 min after excitation, the middle one—after 3 hours, and the lower one—after 120 hours.

b. Bleaching action of the exciting rays

The release of electrons from localization levels can be produced by rays of very different wavelengths. Each phosphor has a characteristic absorption spectrum corresponding to optical bleaching. There are certain phosphors which are bleached under the action of the exciting rays themselves. In this case the exciting light, on the one hand, produces the liberation of electrons, while on the other hand these electrons, after their localization, can be released from the localization levels by rays of the same wavelength as that which produced the excitation. In phosphors of the type described, the exciting rays, by freeing electrons from deep localization levels, limit the accumulative capacity of the phosphor^32. To a weak degree this phenomenon is observed in many substances; to a strong degree—only in a few.

c. Light sums of thermal and optical bleaching

Experience shows that an excited phosphor can be brought into the unexcited state both by heating, i.e. by a thermal method, and by irradiation with infrared rays, i.e. by an optical method. However, these two methods of bleaching a phosphor differ substantially from one another in two respects.

1. The spectral composition of the radiation in optical and thermal bleaching often proves to be different^31. In Fig. 17, a and b, the emission spectra of SrS·Ce, Sm, La-phosphor are presented. Fig. 17, a gives the emission spectrum under optical bleaching; Fig. 17, b—under thermal bleaching. As is seen from Fig. 17, b, under thermal bleaching the spectrum consists of separate bands belonging to samarium. In Fig. 17, a the spectrum consists of one broad band lying in the blue part of the spectrum; this emission band belongs to cerium.

2. The quantity of light emitted by a phosphor up to complete bleaching, under optical and thermal bleaching, proves to be unequal. This quantity of light, expressed by the number of emitted quanta, is called the quantum light sum.

The light sums of optical and thermal bleaching may differ by several times^30, ^33, ^34. In the absence of optical quenching, which will be discussed below, the optical light sum is usually greater than the thermal one (Fig. 18).

It follows from what has been said that the difference in the methods of releasing localized electrons leads to a difference in the course of recombi-

Fig. 17. Microphotograms of spectra: a — optical flash of SrS·Ce,Sm,La phosphor (1 — at 20°C; 2 — at 110°C); b — thermal bleaching of SrS·Ce,Sm,La phosphor (1 — in the range from 0 to 130°C; 2 — from 110 to 190°C).

Fig. 18. Comparison of the light sums of the optical flash and phosphorescence. The magnitudes of the light sums are determined by the areas between the time axis and the curves of optical bleaching (A) or thermal bleaching (B and V). V—thermal bleaching after the phosphor had first been half bleached optically. Curves b and v give the course of the temperature change during experiments B and V.

nation, which is accompanied in the first case by excitation of cerium ions, and in the second by excitation of samarium ions. Evidently, the entire kinetics of the process—the very migration of electrons through the crystal—is different in these two cases of liberation of localized electrons.

8. ON THE YIELD OF LUMINESCENCE AND THE PHENOMENA OF QUENCHING OF LUMINESCENCE

a. Determination of the luminescence yield of crystalline phosphors

The concepts of the energy yield \(\mathcal{B}_{\mathrm{e}}\) of luminescence, as the ratio of the energy of the emitted light \(\varepsilon_{\ell}\) to the energy of the absorbed light \(\varepsilon_{\mathrm{p}}\):

\[ \mathcal{B}_{\mathrm{e}}=\frac{\varepsilon_{\ell}}{\varepsilon_{\mathrm{p}}}, \tag{15} \]

or of the quantum yield \(\mathcal{B}_{\mathrm{k}}\), as the ratio of the number of emitted quanta \(N_{\ell}\) to the number of absorbed quanta \(N_{\mathrm{p}}\):

\[ \mathcal{B}_{\mathrm{k}}=\frac{N_{\ell}}{N_{\mathrm{p}}}, \tag{16} \]

were introduced into the theory of luminescence by S. I. Vavilov\({}^{35}\). These are the most important quantities characterizing the ability of a substance to convert absorbed energy into luminescent light (instead of the usual, not always convenient, notation of the yield by the letters \(\eta\) or \(\rho\), we have adopted here the notation by the letter \(\mathcal{B}\)). S. I. Vavilov determined the absolute value of the yield for a number of solutions of luminescent substances and established the law bearing his name: in the region of high frequencies the quantum yield is constant, while the energy yield increases in proportion to the wavelength of the exciting light; with a further increase in the wavelength of the exciting light, in the region of the frequency of the maximum of the luminescence spectrum, a rapid fall of the yield occurs, its value decreasing to zero; in the intermediate region, over a certain interval of frequencies, the yield remains constant.

To establish the magnitude of the yield one must determine the amount of absorbed energy. This determination is difficult to carry out for crystalline phosphors, since the latter are powders that strongly scatter light. When a surface coated with phosphor is directly irradiated, a large part of the incident light is not absorbed but is scattered, and any sufficiently accurate determination of the amount of absorbed energy becomes impossible. Therefore, in the overwhelming majority

cases, in studying the influence of various factors on the luminescence of crystalline phosphors, they are limited to relative measurement of brightness. Satisfactory measurements of the absolute values of the yield have been carried out only a few times. Among them we may note the measurements of the yield of zinc-sulfide and silicate phosphors, performed by V. V. Antonov-Romanovsky and S. M. Epshtein^37.

The method of Antonov-Romanovsky and Epshtein consisted in investigating the luminescence of a phosphor deposited on the inner surface of a sphere that had two apertures (Fig. 19). Through one of the apertures light was directed into the sphere; observation was carried out through the second aperture. Initially a beam of exciting light was directed into the sphere. It was partly absorbed at its first incidence on the phosphor directly opposite the aperture, and partly scattered and, being absorbed at other places on the inner surface of the sphere, excited the phosphor there. The luminescence light excited inside the sphere, after multiple reflections, emerged outward. The intensity of the exciting light entering the sphere through the first aperture, and of the luminescence light emerging outward through the second aperture, was measured with a thermoelement. To introduce a correction for the absorption of luminescence that occurred in the sphere during multiple reflections, a flux of visible light of the same spectral composition as the luminescence light was directed separately into the sphere, and the flux of light emerging from the second aperture was determined. It is evident that the energy yield of luminescence is, to a first approximation, equal to:

Fig. 19. Sphere with phosphor for determining the absolute value of the yield.

\[ \mathcal{B}_{\mathrm{e}}=\frac{I_{\mathrm{l}}}{J_{\mathrm{v}}}\frac{J_{\mathrm{lu}}}{J_{\mathrm{l}}}, \tag{17} \]

where \(I_{\mathrm{l}}\) is the intensity of the luminescence light emerging from the sphere, \(J_{\mathrm{v}}\) is the intensity of the exciting light entering the sphere, \(J_{\mathrm{lu}}\) is the intensity of the visible light entering the sphere, spectrally close to the luminescence light, and \(J_{\mathrm{l}}\) is the intensity of the visible light emerging from the sphere. By the method described, for ZnS·Cu and Zn\(_2\)SiO\(_4\)·Mn phosphors the maximum value of the energy yield was found to be \(\mathcal{B}_{\mathrm{e}}\approx 0.6\). Thus, under photoexcitation the quantum yield can be close to unity. However, as already indicated above, the magnitude of the luminescence quanta is considerably—

LUMINESCENCE OF ACTIVATED CRYSTALS

considerably smaller than the magnitude of the quanta of the exciting light, as a result of which, even with a quantum yield equal to unity, the energy yield of luminescence is almost always considerably less than unity. Stokes losses are often more than 50%.

With corpuscular excitation the energy yield of luminescence is usually small; thus, according to not very accurate data, the energy yield of luminescence under cathode excitation is measured by a few percent. Under excitation by α-rays the energy yield is about 4%. The energy yield is considerably smaller under excitation by γ-rays. In this latter case the brightness of the phosphors’ luminescence is extremely small, which is a consequence not only of the small luminescence yield, but also of the great penetrating power of γ-radiation, as a result of which only an insignificant fraction of the γ-quanta is absorbed even by thick layers of phosphor.

Fig. 20. Relative energy yield of luminescence of crystal phosphors: I—ZnS·Cu \(10^{-4}\), Co; II—ZnS·Cu \(10^{-5}\), Co; III—ZnS·CdS·0.25; IV—ZnS·Cu \(10^{-4}\).

Fig. 21. Quantum yield of luminescence of crystal phosphors: I—zinc silicate; II—ZnS, strong excitation; III—ZnS, weak excitation; IV—ZnS·CdS, strong excitation; V—ZnS·CdS, weak excitation.

Among other methods for determining the luminescence yield, one should mention the elegant method for determining the yield under excitation in the long-wave part of the absorption band, proposed by S. I. Vavilov and carried out by M. N. Alentsev\(^{40}\). The main difficulty in measuring the luminescence light when it is excited by long-wave rays is the admixture to it of rays of visible light from the same spectral region. In the method of S. I. Vavilov this difficulty is circumvented by directing onto the luminous phosphor an intense flux of infrared rays, which quenches the luminescence. Then the difference between the initial brightness of the phosphor and the brightness under the simultaneous action of infrared rays gives the brightness of the luminescence. With the aid of this method M. N. Alentsev, under the direction of S. I. Vavilov, investigated the dependence

of the relative energy yield on the wavelength of the exciting light. The curves obtained by him are shown in Fig. 20. Fig. 21 gives curves, measured by M. N. Alentsev, of the dependence of the quantum yield of certain phosphors on the wavelength of the exciting light under ultraviolet excitation.

6. Dependence of the yield on the composition of the phosphor

The luminescence yield naturally depends on the composition of the phosphor and the methods of its preparation. To obtain a brightly luminescing phosphor with good yield, it is necessary to maintain a definite ratio between the amounts of its constituent parts: the base substance, the activator, and the flux, and also to use a proper technology for preparing the phosphor. The role and influence of the individual factors are still not sufficiently clarified theoretically, although the practical procedures and formulations for preparation are well known for many classes of phosphors. The optimum conditions for preparing typical crystalline phosphors were indicated in § 4. Below we shall dwell on the influence of accidental impurities and small additions.

c. Influence of extraneous impurities

It was noticed long ago that bright luminescence of phosphors is obtained only when the initial preparations are of high purity. From this the natural conclusion was drawn that many substances, entering the phosphor in the form of accidental, minute impurities, produce a quenching effect and reduce the yield of luminescence. Subsequent experiments established that the quenching effect is produced by ions of heavy metals, especially iron, nickel, cobalt, platinum, etc. Many of these quenchers of luminescence, which are called luminescent poisons, can themselves serve as activators. Thus, the quenching effect here reduces to the interaction of ions of heavy metals, each of which, under suitable conditions, can itself emit radiation.

Some of these metals exert a very strong effect. Thus, for example, iron, taken in an amount of \(10^{-6}\) gram per gram of the base substance, already produces a strong quenching effect.

The interaction of zinc and manganese ions in ZnS·Mn phosphors is especially interesting\(^{55}\). It is known that pure zinc sulfide gives a blue luminescence caused by the presence of neutral zinc atoms. The addition of small amounts of manganese—\(10^{-4}\) gram per gram of the base substance—causes an extremely strong weakening of the blue luminescence of zinc. The luminescence of the manganese itself at such concentrations is still almost imperceptible. Manganese acts as a poison with respect to the luminescence of zinc. At higher concentrations manganese quenches the luminescence of zinc with still greater force;

however, at the same time, instead of zinc luminescence, a rather bright manganese luminescence arises, which becomes the only one at high manganese concentrations. Here manganese already acts as a competitor of zinc in the use of the energy absorbed by the crystal lattice. It takes it from the lattice considerably more intensively than zinc; therefore, under weak excitations, almost exclusively manganese glows and the total radiation has an orange color. As the excitation intensity increases, all centers of manganese luminescence gradually become excited and the brightness of the manganese luminescence tends toward saturation. From this moment the energy absorbed by the crystal lattice begins to be transferred intensively to zinc, whose blue luminescence rapidly increases. Such an interaction of zinc and manganese is shown in Fig. 22, where the intensities of excitation are plotted along the abscissa axis, and along the ordinate axis—the intensities of luminescence: of zinc—curve I, and of manganese—curve II.

Fig. 22. Interaction of zinc and manganese luminescence centers in a ZnS·Mn phosphor: I—curves of zinc luminescence, II—curves of manganese. Excitation by Hg lines with \(\lambda = 312\) and \(366\,m\mu\).

The question of whether these activator ions act on one another through the cells of the crystal lattice of the basic substance, being distributed in crystal phosphors statistically uniformly and independently of one another throughout the whole volume of the crystal, or whether they form chemically bound complexes throughout the entire volume, has not yet been fully resolved. It may be, however, that cases are observed in which the ions of heavy metals should be regarded as directly bound to one another and as forming complex compounds. Similar formations of complexes occur in phosphors with double and triple activators from the rare earths\(^ {31}\).

г. Temperature quenching

The yield of luminescence of crystal phosphors strongly depends on temperature. We have seen above that each luminescence band develops most intensively in a definite temperature interval. At temperatures below the optimum, self-trapping of a portion of the liberated electrons occurs. At temperatures above the optimum, the radiation decreases owing to the occurrence

temperature quenching.

Fig. 23. Curves of temperature quenching of zinc-sulfide phosphors.

Fig. 24. Possible arrangement of the potential curves of an excited and an unexcited phosphor, explaining the occurrence of temperature quenching.

The cause of this phenomenon has not yet been sufficiently clarified. In any case, for temperature quenching to arise, some additional energy supplied to the electron is required. Thus, two processes of excitation of an electron from a local level are possible. If the electron receives additional energy \(\Delta \varepsilon_1\), then, upon being freed, it passes into the conduction band, after which emission occurs. If, however, the electron receives energy \(\Delta \varepsilon_2\), then the phenomenon of quenching arises. Thus, the light output is determined by the ratio

\[ \mathcal{B}=\frac{A}{A+B} \]

and

\[ \mathcal{B}= \frac{1}{1+C e^{-\frac{\Delta(\varepsilon_2-\varepsilon_1)}{kT}}}. \tag{18} \]

Here \(A\) and \(B\) are numbers expressing, respectively, the probabilities of optical emission and quenching. Relation (18) has been checked in several cases and has proved to be in good agreement with experiment. Thus, V. A. Yastrebov checked it for the quenching of zinc-sulfide phosphors (Fig. 23)\(^{41}\), V. E. Loshkarev\(^{42}\) for the case of quenching of the infrared luminescence of cuprous oxide, and F. I. Vergunas and F. F. Gavrilov\(^{43}\) for the case of quenching of the luminescence of zinc oxide.

Seitz and Mott\(^{44}\), in order to explain this phenomenon, propose using a special arrangement of the potential curves,

of the excited and unexcited states (Fig. 24). The right branches of these curves intersect, which makes it possible for an electron located on the upper potential curve, under intense vibrations of the nuclei without emission, to pass to the lower potential curve. Such a scheme, borrowed from the theories of molecular luminescence, is possible in the case of crystal phosphors only when quenching occurs at localized electrons, with localization taking place near ionized centers, so that the nonradiative transition to the lower potential curve corresponds to recombination without optical excitation of the corresponding ion.

Let us note that another explanation of quenching by direct collisions of an excited electron with surrounding particles and the transfer of excitation energy to the latter encounters difficulties, since the magnitude of the excitation energy is very large, of the order of \(50\,kT\) and more. The transfer of such a large energy in a single collision event is theoretically improbable; it can occur only in a simultaneous collision with a very large number of particles. E. I. Adirovich indicated the possibility of eliminating this difficulty. In his theory, the transfer of electronic energy to the crystal occurs not in a single event, but gradually \(^{50}\).

d. Optical quenching

Just as thermal glowing-out has its counterpart in optical glowing-out, so thermal quenching has its counterpart in optical quenching. Absorption by an excited phosphor of an additional light quantum can lead not only to the liberation of electrons from their sites of localization with a subsequent radiative transition, but also to transitions of electrons not accompanied by emission. This phenomenon of optical quenching has long been known. Qualitatively, it has been investigated repeatedly. The phenomenon is especially strongly developed in phosphors of the zinc sulfide group. A detailed quantitative investigation of the optical quenching of CaS·Ni and, chiefly, of ZnS·Cu and ZnS·CdS·Cu phosphors was carried out by V. V. Antonov-Romanovskii and the author \(^{45}\). It was shown that the magnitude of the quenching effect is determined by the exposure \(\theta = Jt\)—the product of the intensity of the acting light \(J\) and the duration of action \(t\). For a given exposure, the magnitude of the quenching action is a function of the wavelength of the quenching rays.

In some cases the same rays can cause both a flash of luminescence, i.e. an accelerated glowing-out of the phosphor by a radiative path, and quenching of the phosphor, i.e. glowing-out of the phosphor by a nonradiative path. Like the flash of luminescence, the phenomenon of quenching can be applied for photographic purposes. Examples of such use of this phenomenon are given below.

9. LUMINESCENCE OF CRYSTAL PHOSPHORS AND PHOTOCONDUCTIVITY

The recombination theory of the luminescence of crystal phosphors assumes the liberation of electrons at the moment of excitation, i.e., the presence of an internal photoelectric effect. The appearance of photoelectrons inside crystals should cause a change in their photoconductivity and dielectric constant. The liberation of electrons upon excitation may also be detected in the form of an external photoelectric effect at the moment of excitation; moreover, when measuring the wavelength of the exciting rays, parallelism should be observed in the development of the photoelectric effect and luminescence under the action of rays of each definite wavelength. Finally, the superposition of an electric field of large magnitude on the excited phosphor should cause a displacement of electric charges and, in particular, the release of localized electrons, i.e., should lead to a flash of luminescence.

All the phenomena indicated have in fact been observed by various authors: thus Schmidt^46 observed strong flashes of luminescence when an electric field was applied to a decaying CdS·Mn phosphor.

Recently, the effect of an electric field on the luminescence of zinc-sulfide phosphors has been studied in detail by Destriau^47.

The change in the electrical conductivity of phosphors at the moment of their irradiation has been observed by many authors. It may be regarded as established that typical crystal phosphors belong to the semiconductors and have an electrical conductivity tens of thousands of times greater than the electrical conductivity of luminescent materials possessing luminescence of the type of discrete centers. Recently, the questions of the connection between electrical conductivity and luminescence have been studied by V. E. Lashkarev^42, D. V. Gurevich and N. A. Tolstoi, and P. P. Feofilov^48. They investigated the phenomena of photoconductivity of cadmium sulfide and established that the kinetics of the change in the light sum of phosphorescence over a wide range coincides with the kinetics of photoconductivity. The number of excited electrons determining the magnitude of the light sum proved proportional to the number of excited electrons determining the photoconductivity, and the law of change of the light sum coincided, to within constants, with the law of change of the photoconductivity. See also^49.

10. MODERN CONCEPTIONS OF THE KINETICS OF LUMINESCENCE

a. General scheme of the kinetics of luminescence of crystal phosphors

Above, various phenomena of the luminescence of crystal phosphors have been described. Their extraordinary complexity, as well as the complexity of the structure of the crystal phosphors themselves, has not made it possible up to the present time to construct a quantitative theory of the phenomenon; there is only

Luminescence of Activated Crystals

a general schematic picture of the kinetics of luminescence, which proves useful in clarifying the essence of the phenomena observed experimentally. In the present section, the experimental facts considered earlier and certain theoretical propositions expressed in the course of their description are summarized, and on this basis a description is given of present-day ideas about the kinetics of luminescence.

In attempts to construct a theory of the phenomenon it is necessary to take into account the following principal properties of the luminescence of crystal phosphors, indicated above: 1. The simultaneous existence of several luminescences with the same spectral composition, but with entirely different durations: short-lived processes lasting millionths or thousandths of a second, and long-lived processes lasting for hours. 2. The possibility of the simultaneous existence of luminescences of different spectral composition. 3. The complex hyperbolic course of the decay of luminescence. 4. The existence of a deep connection between the phenomena of luminescence and photoconductivity. 5. The possibility of exciting luminescence by frequencies belonging to different absorption regions, the absorption spectra in most cases being related neither in shape nor in position to the emission spectra.

The first attempts at a theoretical explanation of the phenomenon were based on the idea of the existence within the phosphor of several kinds of luminescence centers—long-lived and short-lived—each of which gave luminescence decaying exponentially, while the sum of such exponents was selected in accordance with the complex law of decay of long-duration luminescence. Usually this could be achieved by superposing 3–4 exponents. The picture of phenomena described could not be retained, both because of the existence of photoconductivity in crystal phosphors, directly connected with excitation, and because of the inconsistency of the notion of the structure of the centers, which were described as certain isolated places in the phosphor containing an enormous number (sometimes hundreds of thousands) of lattice ions. It is quite obvious that with such a structure of the centers there was no basis for limiting oneself to two or three kinds of centers; rather, one would have had to assume the existence of a practically continuous series of centers differing in magnitude.

Essentially different ideas underlie the recombination theory of the process. The liberation of an electron at the moment of excitation inside the crystal lattice explains the connection of the luminescence of crystal phosphors with the phenomena of photoconductivity. However, the simplest scheme of the recombination process, reducing to the consideration of the direct recombination, accompanied by radiation, of excited electrons and ions, proves insufficient. As we saw above, this scheme for the law

decay gives a hyperbola of the second order, which cannot explain the simultaneous existence of long- and short-time processes and, in the overwhelming majority of cases, does not express the true course of the decay of luminescence.

Therefore the modern recombination theory admits the existence of two types of recombination of excited electrons. In the first, recombination occurs directly and therefore proceeds rapidly; in the second, before the moment of recombination there occur one or several captures of electrons by special sites of the crystal lattice, from which the electrons can be released only under external action.

The internal states of a crystallophosphor are considered in the light of the modern theory of the solid state. It is assumed that the deep energy levels of the particles forming the crystal lattice are filled with electrons and are located at each particle separately; the higher levels of the individual ions merge with one another and extend over the entire crystal. The fairly considerable width of the bands of possible energy states thus formed increases as the number of the band increases. The width of a band is due to the splitting of individual levels under the influence of the strong internal field of the crystal, and also as a result of the quantum-mechanical interaction of the ions of the crystal.

Some of the lower bands, lying closer to the ions of the crystal, prove to be filled with electrons. Above the last filled band there is a series of bands belonging to various cations. The energy distance between the last filled band and the first empty one is so considerable that electrons from the filled band can rise into the unfilled band only as a result of optical excitation. An electron that has entered an empty band can move throughout the entire crystal. These motions of the electron very quickly lead it to recombination with one of the ions of the crystal or to an encounter with a special site of the lattice. Recombination causes luminescence, which apparently consists in the excitation of one of the ions of the crystal and subsequently in its emission; this direct recombination explains the occurrence of instantaneous luminescence. Upon encounters with special sites of the crystal lattice the electron becomes localized, which leads to a delay of recombination and explains the occurrence of long-duration luminescence. The special sites of the lattice are places of considerable deformation of it. The latter are caused above all by the introduction of activator atoms; thus the activator atoms, which form luminescence centers, are at the same time sites in whose vicinity localization of electrons occurs. The structure of the luminescence centers at present has been little investigated; however, it is quite unlike the previously assumed structure of the centers. The composition of the luminescence center includes

...cence apparently includes an atom or ion of the activator and the volume of the crystal immediately adjoining it, deformed by the inclusion of the activator ion. It is also possible that melt anions enter into the composition of the center, since experience shows that the luminescence spectrum depends substantially on the melt.

It is most probable that the radiation arises not at the very moment of recombination, but somewhat later; the recombination produces excitation of the activator ion, which occurs at the expense of the energy of the recombining particles. Subsequently the excited ion emits light.

Fig. 25. Diagram of the kinetics of the processes of short-duration and long-duration luminescence and optical quenching of crystallophosphors.

The described kinetics of luminescence may be represented by one of the following schemes2, 28. In Fig. 25 the upper filled band \(A\) and the very lowest of the unfilled bands \(B\) are shown. By separate small lines \(a, б, в, г\) are indicated the energy levels of deformed lattice sites, created by various inclusions and, above all, by the presence of the activator. The levels \(a\) are located not far from band \(A\) and easily exchange electrons with it. In the presence of a free place in band \(A\), an electron can readily descend from level \(a\) and fill its free place. On the other hand, electrons from the conduction band, when moving through this band, can fall into the potential wells corresponding to the levels \(б, в, г\). The return of electrons to the conduction band occurs when they absorb additional energy. From the point of view of the scheme described, luminescence proceeds as follows: upon excitation by rays belonging to the absorption region of the basic substance of the crystallophosphor, the electron liberated upon absorption of the light...

quantum, an electron is transferred from zone \(A\) to zone \(B\). The positive charge formed in zone \(A\) gradually moves to the upper boundary of this zone by successive replacement of the positively charged site that has formed by electrons located higher up in zone \(A\). The positive charge located at the surface of zone \(A\) recombines with an electron descending from level \(a\). This descent is equivalent to the ascent of the positive charge shown by the circle. Thus, as a result of the described diffusion of positive charge, the activator acquires a positive charge.

The process considered applies to the case in which absorption is produced by the host substance. If, however, the absorption is produced by the activator itself, then the positive charge is formed immediately as a result of absorption and of the transfer of an electron from the activator atom to the conduction band.

Let us now consider the motion of a negative charge released by the absorption of light. This charge moves in the conduction band, and its subsequent fate may vary. It may either recombine directly from the conduction band with one of the ionized centers situated at level \(a\), which leads to the appearance of instantaneous luminescence, or descend to very shallow levels \(z\), whence it is at once thrown back into the conduction band by thermal motion, or descend to deeper levels \(b\) and \(v\). The fall of an electron to levels \(z\) somewhat retards the luminescence; however, this process is also short-lived; the fall of electrons to levels \(b\) causes a strong retardation of the luminescence process; when an electron falls to very deep levels \(v\), it can be released only as a result of absorption of a photon or strong heating of the phosphor. Electrons returning to the conduction band from localization levels and recombining with ions give prolonged luminescence.

The picture described is, undoubtedly, highly schematic. Here only typical kinds of localization levels have been taken. In a real situation there may be several kinds of activators, there may be several conduction bands and a large spread in the depths of each system of levels. All this greatly complicates the process of luminescence.

b. Calculation of the kinetics of luminescence for the case of an idealized scheme of a crystal phosphor

Despite the plainly simplified character of the scheme of crystal-phosphor luminescence described above, the performance, in accordance with it, of quantitative calculations of the kinetics of luminescence presents insurmountable difficulties. Further simplifications of the scheme consist in

assuming the existence of a single system of localization levels possessing a definite depth value.

The system of differential equations corresponding to this scheme is nevertheless still so complicated that until recently it remained unsolved in its general form. E. I. Adirovich^50 introduced physically justified assumptions and obtained a solution of the problem sufficiently accurate within the framework of the given simplified scheme. We present the system of equations for the kinetics of luminescence and its solution, as proposed by E. I. Adirovich. Let \(N\) denote the number of electrons in the conduction band, \(\nu\) the number of electrons on the localization levels, \(\nu_1\) the total number of localization levels, \(n\) the number of ionized centers, \(A_1\) the probability of recombination of an electron with an ion at unit concentration of both, and \(A_2\) the probability of localization of an electron from the conduction band on one of the levels. Then, considering separately the balance of electrons in the conduction band, on the localization levels, and the change in the number of ionized centers, we find the following system of three differential equations:

\[ \left. \begin{aligned} \frac{dN}{dt} &= p\nu - A_2 N(\nu_1-\nu) - A_1 Nn,\\ \frac{d\nu}{dt} &= -p\nu + A_2 N(\nu_1-\nu),\\ \frac{dn}{dt} &= -A_1 Nn. \end{aligned} \right\} \tag{19} \]

The first equation takes into account the circumstance that the change in the number of electrons \(dN\) in the conduction band is equal to the difference between the number of electrons entering the conduction band from the localization levels \((p\nu\,dt)\) and the numbers of electrons undergoing localization and recombination, \(-A_2N(\nu_1-\nu)\,dt\) and \(-A_1Nn\,dt\).

The second equation indicates that the change in the number of electrons on the localization levels is equal to the difference between the numbers of localized electrons \(A_2N(\nu_1-\nu)\,dt\) and of electrons leaving the localization levels for the conduction band, \(-p\nu\,dt\). Finally, the quantity \(dn\) determines the changed numbers of ionized centers. Physically justified here is the assumption of the smallness of the quantities \(N\) and \(\frac{dN}{dt}\). If \(\frac{dN}{dt}\) is close to zero, then the first equation of system (19) makes it possible to find an expression for \(N\) in terms of \(\nu\) and \(n\); further, substituting this expression for \(N\) into the second equation of the system and knowing that \(n=\nu+N\simeq\nu\), we easily arrive at the following expression:

\[ (1-\gamma)\lg\frac{n_0}{n}+\gamma\nu_1\left(\frac{1}{n}-\frac{1}{n_0}\right)=pt, \tag{20} \]

giving, in implicit form, the dependence of the number of localized electrons on time. Here \(\gamma=\dfrac{A_2}{A_1}\).

Expression (20) for \(\gamma=1\) gives for \(I\) a hyperbolic law of second-order decay. For \(\gamma \ne 1\), the course of the changes in \(I\), calculated from equation (20), coincides with that calculated by Becquerel’s formula over a comparatively wide interval of change of the luminescence intensity.

b. On the kinetics of the bleaching of phosphors with deep localization levels

The decay of the luminescence of an optical flash, occurring under prolonged action on a phosphor of bleaching rays, proceeds very similarly to natural decay. This is quite understandable, since in both cases localization levels are released under the action of a certain constant factor, in the present case the luminous flux of the bleaching rays.

The difference consists in the fact that, in optical bleaching of deep levels, one must consider not one but at least two systems of local levels. The shallow levels, being filled with electrons, serve as sources of phosphorescence. The deep levels, however, are bleached only under the action of optical rays. Consideration of the question, carried out by the author\(^{28}\), showed that the kinetics of the process can be described by four differential equations. However, experimental investigations led to the conclusion that the filling of the deep and shallow levels occurs with equal probability (at least for the class of alkaline-earth phosphors activated by rare earths). Owing to this, the system of four differential equations can be reduced to a system of three differential equations. Their solution leads to the expression:

\[ \frac{\nu_0}{\gamma}\ln \frac{\nu_0}{\nu_0-\nu}+\frac{\gamma-1}{\gamma}\,\nu=Jt, \tag{21} \]

where \(J\) is the intensity of the exciting rays. Equation (21) gives the law of increase in the number of localized electrons at deep levels during excitation of the phosphor. For \(\gamma=1\), expression (21) becomes the exponential dependence

\[ \nu=\nu_0\left(1-e^{-\frac{Jt}{\nu_0}}\right), \tag{22} \]

which, however, describes only approximately the increase of the luminous sums of the flash. Practically complete agreement of experimen-

...of the results with the calculation gives the formula:

\[ \nu \simeq \frac{\nu_0 J t x}{1+J t x}, \tag{23} \]

where \(x\) is a scale factor. Formula (23) was obtained on the basis of additional assumptions and therefore has a semi-empirical character. The degree of accuracy of the formula is characterized by Fig. 26, where the solid curve gives the course of function (23), and the symbols show the experimental data. From the same differential relations one obtains a qualitatively correct picture also for the bleaching of flash phosphors: the proportionality of the flash intensity to the square of the light sum accumulated by the phosphor, i.e., to the square of the number of electrons localized at deep levels, the course of the decay of luminescence under the action of bleaching rays, and also an explanation of the phenomenon of secondary phosphorescence described below.

Fig. 26. Growth of the store of electrons at deep levels in the process of excitation of the phosphor.

11. ON SOME CLASSES OF CRYSTAL PHOSPHORS

In a single article it is not possible to give any reasonably complete survey of the enormous variety of existing types of crystal phosphors and their technical applications. Therefore we shall restrict ourselves to only two examples.

a. Quenching and flash phosphors

Above, the quenching and bleaching action of optical rays on an already excited phosphor was briefly described.

Under the action of light rays, chiefly visible and near infrared, there occurs a release of excited...

electrons from deep localization levels, and in some cases the freed electrons recombine without radiation (quenching phosphors), while in others they do so with radiation (flash phosphors). The study of these phosphors is of exceptional interest, since the peculiar features of the phenomena of quenching and flash make it possible to investigate in depth the kinetics of luminescence, to establish the properties and nature of luminescence centers and localization sites. The investigation of these phenomena was begun in the Soviet Union on the initiative of S. I. Vavilov.

In 1934, V. V. Antonov-Romanovskii and the author carried out detailed studies of the phenomenon of optical quenching in zinc-sulfide and zinc-cadmium-sulfide phosphors. These studies established the basic law of the quenching action—the proportionality of the quenching action to the exposure of the quenching rays and the dependence of the sensitivity of the phosphor on the character of its excitation.

Fig. 27. Example of photocopying with the aid of a quenching phosphor.

It was found that the quenching action diminishes as the phosphor becomes saturated; here, as the measure of quenching, the expression $\dfrac{\Delta S}{S}$ was adopted, where $S$ is the amount of light energy accumulated by the phosphor, and $\Delta S$ is its change under the action of the bleaching rays. The sensitivity curves of phosphors to quenching rays of different wavelengths are characteristic of luminous substances of definite composition; they possess one or two maxima, situated in the visible and near-infrared parts of the spectrum.

At the suggestion of S. I. Vavilov, the phenomenon of optical quenching was used for a simplified method of photocopying, in which

...the photographic negative process is eliminated. The image of the object being copied, strongly illuminated by rays that quench luminescence, is projected onto an excited luminescent screen, on which dark spots are produced at the places of greatest illumination. The negative image of the object thus obtained on the screen is transferred by contact printing onto photosensitive paper. The places of greatest darkening of the screen produce the least blackening of the paper, as a result of which a positive image appears on the paper. This method was developed by L. A. Vinokurov^51. It leads to a saving of photographic material and to a considerable acceleration of photocopying. In Fig. 27 a photocopy is reproduced of the title of a section of S. I. Vavilov’s book The Microstructure of Light, obtained by the method described.

b. Flash phosphors

Work with flash phosphors was a natural development of work with quenching phosphors^30, ^31, etc. In the Soviet Union it was also carried out under the general direction of S. I. Vavilov. In parallel, analogous investigations were carried out in the USA, in a different sequence, leading to similar results^34, ^52, ^53.

Much of the information obtained in recent years in the study of flash phosphors has already been described above. The flash phenomenon is especially well expressed in alkaline-earth phosphors activated by a combination of rare earths, for example cerium and samarium, or europium and samarium. The color of the flash is determined by one of the activators of the pair. In the cases indicated, in the first pair it is cerium (green luminescence), and in the second it is europium (orange luminescence). However, the brightness of the flash depends on the presence of the second activator, samarium. In its absence the phosphors give a weak flash.

Thus, samarium ensures the development of the flash, which is why we called it the flash activator^30. As indicated above, the activators of these complex phosphors are located close to one another and form complex centers. Upon excitation of these phosphors, the filling of deep and shallow levels occurs simultaneously and at the same rate; however, the light sums stored at the deep localization levels are much greater than the light sums stored at the shallow ones^28.

A very remarkable fact is the necessity of a certain thermal activation for the optical excitation of electrons. The strongest irradiation cannot cause complete bleaching of a phosphor that has a low temperature. To bleach it, heating is necessary. This is all the more remarkable because the magnitudes of the optical

of quanta are in themselves quite sufficient to liberate electrons from the localization levels.

The necessity of preliminary thermal activation undoubtedly indicates the two-stage character of bleaching. Figure 28 shows the course of the decrease in the brightness of the flash of an SrS·Ce,Sm,La phosphor when the temperature is lowered3.

Experiments show that the bleaching action is a function of wavelength. The curve of the quantum yield of flash luminescence usually has two maxima: one in the infrared and the other—

Fig. 28. Change in the magnitude of the optical flash with changing temperature in SrS·Ce,Sm,La phosphors.

in the visible part of the spectrum. The position of the yield maxima does not depend on the basic substance of the phosphor, but depends only on the second activator. Thus, absorption is associated with the atoms of the flash activator; in the phosphors described above—with samarium.

The absolute magnitude of the quantum yield of the flash, i.e., the number of luminescence quanta referred to the number of quanta of the absorbed bleaching rays, is rather high and reaches \(1/3\). This shows that an electron liberated upon absorption of a bleaching quantum in most cases immediately recombines, and does not undergo numerous repeated localizations, since in this latter case the quantum yield would be considerably less than unity4.

However, some fraction of the electrons released from deep localization levels nevertheless undergoes relocalization on shallow levels. This gives rise to the phenomenon of secondary phosphorescences, which clearly demonstrates the existence of local levels of various depths. Secondary phosphorescence becomes noticeable under pulsed action of bleaching rays on an excited phosphor with an already extinguished phosphorescence. After the cessation of the pulse and the non-inertial flash, a prolonged

Fig. 29. Phenomena of secondary phosphorescence. On the left, the curve of normal phosphorescence \(I_{\phi}\) and the curve of optical bleaching \(I_{в}\), the latter on a scale of 0.1; on the right, arrows on a scale of 0.1 indicate the second flashes. Alongside them, the curves give the decay of the secondary phosphorescences.

secondary phosphorescence. Its natural explanation is that some of the electrons transferred by the pulse of bleaching rays into the conduction band again descend to shallow localization levels, from which they are subsequently released already by thermal means. This phenomenon can be repeated many times. Fig. 29 depicts one such experiment. Here the long arrows indicate the flash intensities at the moment when infrared rays act on a SrS·Ce,Sm,La phosphor. The pulse duration is one second. The subsequent decay curves describe the glow of secondary phosphorescences. In the case shown in Fig. 29, the process was repeated ten times[^28].

a                   b

Fig. 30. Emission spectra of a mercury lamp (a) at different exposure times and absorption spectra of diphenylpolyenes (b), taken with the aid of flash phosphors.

Flash phenomena can be used to investigate spectra in the infrared region. For this purpose, the dispersed light of the source under investigation is directed onto a plate made of a luminescent substance, against which a photographic plate is pressed directly. In those places of the phosphor on which infrared radiation falls, a glow arises, which is then recorded by the photographic plate. This method makes it possible to study photographically the infrared region of the spectrum up to 1.7 μ. Fig. 30 shows spectra obtained in this way by Z. L. Morgenstern[^54].

b. Lamp Phosphors

More than 20 years ago, while a professor at the Higher Technical School, S. I. Vavilov began experiments on the use of ultraviolet radiation from gas-discharge light sources to obtain visible light.

Subsequently these experiments developed into a broad range of studies carried out under the general direction of S. I. Vavilov in a number of institutes (Physical Institute of the Academy of Sciences of the USSR, VEI, GOI), which, independently of analogous work being done abroad, led to the creation of luminescent lamps.

Luminescent lamps in their modern form are a gas-discharge tube filled with mercury vapor at low pressure. Under the influence of excitation by an electron stream, the mercury vapor emits the resonance lines 2537 and 1849.5 Å. These latter are absorbed by a layer of crystallophosphors deposited on the inner walls of the tubes, and their energy is converted into visible radiation. The luminous efficacy of the lamps depends on their power and reaches 60 lm/W (instead of 16 lm/W for the best incandescent lamps).

Fig. 31. Spectra of tungstates.

The spectral composition of the radiation of the lamps, depending on technical requirements, may imitate the scattered light of the sky, the direct light of the sun, or the visible part of the spectrum of an incandescent lamp, etc. Luminescent lamps of daylight, white, and warm-white light are manufactured. For decorative purposes, colored lamps are used, coated with phosphors that give radiation only in one spectral region. For the precise reproduction of a required spectral composition, either mixtures of various luminophors are produced with the calculation that their total luminescence gives the desired light effect, or a single phosphor with several activators is developed. As individual components of luminescent mixtures, tungstates and zinc–beryllium silicates are used. Figure 31 presents the spectra of various tungstates. Magnesium tungstate, which gives a broad emission spectrum with a maximum in the blue part of the spectrum, is of greatest importance for lamps. As the yellow-orange component of the luminescence, the luminescence given by zinc–beryllium silicate is adopted. The spectral composition of the luminescence of these phos-

V. L. LEVSHIN

phors changes depending on the relative content of beryllium and manganese, shifting, as the amount of both is increased, into the orange part of the spectrum. In Fig. 32 the spectra are presented of various zinc–beryllium silicates used in lamps. To obtain green luminescence, zinc silicate activated with manganese is especially advantageous; it gives luminescence with a maximum at 520 mμ.

Fig. 32. Spectra of zinc–beryllium silicates
1 — 0% Be, 2 — 6.5% Be, 3 — 17.2% Be.

A high luminous efficiency can also be achieved when one-component phosphors possessing two activators are applied to the walls of the lamp; the combination of their luminescence provides the required composition of radiation. Of these phosphors, halophosphate has been technically developed to the greatest extent.

In Fig. 33 are shown the emission spectra of various fluorescent lamps used to reproduce the composition of different kinds of daylight. The rectangles indicate the radiation of the visible mercury lines, partially passing through the layer of luminescent powder that coats the walls of the lamp.

Fig. 33. Spectral composition of the radiation of various fluorescent lamps.

The enormous advantages of fluorescent lighting—its economy, controllability of the composition of radiation, long service life—

services of the lamp, explosion safety, etc., ensured for it rapid success and wide dissemination.

At present, luminescent lamps are used in mines, in subway stations, in textile factories, in art galleries, etc. In the near future they will become widespread.

CITED LITERATURE

1. S. I. Vavilov, Zeits. d. Sowjetunion 5, 369 (1934).
2. V. L. Levshin, Izv. AN SSSR, ser. fiz. 9, 355 (1945).
3. V. L. Levshin, ZhFKh 2, 641 (1931).
4. V. L. Levshin, ZhFKh 6, 11, 991 (1935).
5. V. L. Levshin, ZhFKh 9, 1 (1937).
6. I. V. Obreimov, A. F. Prikhot’ko and G. Shabaldas, ZhETF 6, 1063 (1936).
7. I. V. Obreimov and G. Shabaldas, Journ. of Physics USSR 7, 168 (1943).
8. E. V. Shpol’skii, A. A. Il’ina and V. V. Bazilevich, Izv. AN SSSR, ser. fiz. 12, 519 (1948); DAN 62, 2 (1948).
9. K. W. Hausser, R. Kuhn, E. Kuhn u. G. Seitz, Zeits. phys. Chem. B. 29, 417, 391 (1935).
10. V. L. Levshin, Izv. AN SSSR, ser. fiz. 2, 185 (1937).
11. S. I. Vavilov and V. L. Levshin, Zeits. f. Physik 35, 920 (1926).
12. Ya. I. Frenkel’, Zeits. d. Sowjetunion 9, 158 (1936).
13. S. Pekar, ZhETF 17, 368 (1947).
14. V. L. Levshin and I. V. Veits, ZhETF 20, 411 (1950).
15. E. M. Brumberg and F. M. Pekkerman, Izv. AN SSSR, ser. fiz. 13, 218 (1949).
16. F. Kröger, Luminescence in Solids, containing manganese, Amsterdam, 1940.
17. F. Kröger, Some aspects of the luminescence of solids.
18. J. E. Wyles, Proc. Roy. Soc. of Lond., Ser. A 167, 34 (1938).
19. V. M. Kudryavtseva, F. I. Vergunas and P. S. Litvinova, Izv. AN SSSR, ser. fiz. 9, 403 (1945).
20. E. Nichols and H. Howes, Fluorescence of the uranyl Salts, Washington, 1919.
21. S. I. Vavilov and V. L. Levshin, Zeits. f. Physik 48, 397 (1928).
22. E. Becquerel, la lumière, p. 278 and 295, vol. 1, Paris, 1867.
23. P. Lenard u. R. Tomaschek, Handbuch d. Experim. Physik 23.
24. V. L. Levshin and V. V. Antonov-Romanovskii, ZhETF 4, 1022 (1934).
25. V. L. Levshin, ZhFKh 6, 1277 (1935).
26. V. V. Antonov-Romanovskii, Trudy FIAN, 1.35 (1937).
27. L. A. Vinokurov, V. L. Levshin and E. G. Baranova, ZhETF 21, No. 2–3 (1951).
28. V. L. Levshin, ZhETF 18, 149 (1948).
29. M. L. Kats, Izv. AN SSSR 13, 199 (1949).
30. V. L. Levshin, V. V. Antonov-Romanovskii, Z. L. Morgenshtern, Z. A. Trapeznikova, ZhETF 17, 949 (1947).
31. V. V. Antonov-Romanovskii, V. L. Levshin, Z. L. Morgenshtern, Z. A. Trapeznikova, Izv. AN SSSR 13, 75 (1949).
32. V. V. Antonov-Romanovskii, Izv. AN SSSR 13, 75 (1949).

1. Z. L. Morgenstern, DAN 54, 791 (1946).
2. R. T. Ellickson, J. Opt. Soc. Am. 36, 264 (1946).
3. S. I. Vavilov, Zeits. f. Physik 22, 266 (1924).
4. S. I. Vavilov, Zeits. f. Physik 42, 311 (1927).
5. V. V. Antonov-Romanovskii and S. M. Epshtein, DAN (1947).
6. A. V. Moskvin, Cathodoluminescence, vol. 1, Gostekhizdat, 1949; Izv. AN SSSR, ser. fiz. 9, 429 (1945).
7. T. M. Timofeev, Izv. AN SSSR, ser. fiz. 13, 135 (1949).
8. M. N. Alentsev, Trudy FIAN 5 (1950).
9. V. A. Yastrebov, Trudy FIAN 3, issue 2, 182 (1946).
10. V. E. Loshkarev and K. M. Kosonogova, DAN 54, 2, 125 (1946).
11. F. I. Vergunas and F. F. Gavrilov, Izv. Sib. Phys.-Tech. Inst. 24, 196 (1947).
12. F. Seitz and N. Mott; N. Mott and R. Gurney, Electronic Processes in Ionic Crystals, IL, Moscow, 1950.
13. V. L. Levshin and V. V. Antonov-Romanovskii, ZhETF 4, 1032 (1934).
14. R. Tomaschek, Handbuch d. phys. Optik, 1928.
15. Destriau, Journ. d. Physique 4, 32 (1943); Phil. Mag. 190 (1948).
16. D. V. Gurevich, N. A. Tolstoi and P. P. Feofilov, DAN 71, 29 (1950); 72, 473 (1950).
17. F. I. Vergunas, Izv. Tomsk. Univ. 24, 185 (1947).
18. E. I. Adirovich, Izv. AN SSSR, ser. fiz. 10, 467 (1946). Trudy FIAN 5, 387 (1950); Certain Questions in the Theory of Luminescence of Crystals, Gostekhizdat, 1951.
19. L. A. Vinokurov, ZhETF 21, No. 2–3 (1951).
20. Fonda, J. Opt. Soc. Am. 36, 382 (1946).
21. B. O’Brien, J. Opt. Soc. Am. 36, 351, 369 (1946).
22. Z. L. Morgenstern, DAN 74, 493 (1950).
23. V. L. Levshin, ZhETF 17, 675 (1947); Izv. AN SSSR, ser. fiz. 12, 297 (1948).
24. F. Urbach, H. Hetmendinger a. D. Pearlman, Preparation and Characteristics of Solid Luminescent Materials, 274, 1948.