Changes in the Thermal, Photo- and Radiation Stability of Solid Inorganic Compounds

Yu. A. Zakharov, G. T. Savel’ev

(Presented by the scientific seminar of the Department of Radiation Chemistry)

The problems of directed modification of the stability of solid inorganic compounds with predominantly ionic bonds with respect to heating, light, and ionizing radiation have long been posed in solid-state chemistry and have been posed quite definitely. Efforts and work in this direction are constantly being stimulated by ever-increasing technological needs for recommendations of this kind. Solving the problem requires carrying out research mainly in two directions:

a) the influence of treatment of the solid body (mechanical, thermal, radiation, and other effects, impurities, etc.) and of the associated changes in a broad range of its properties on stability with respect to heat, light, and radiation;

b) changes in the conditions under which the decomposition process itself is carried out and the resulting changes in the stability of the compound.

These experiments may be, as will become clear below, a tool also for elucidating the subtle mechanism of the processes of decomposition of solids, knowledge of which is of decisive importance for solving the problem.

At present, work in the first group is represented by a fairly large number of publications, predominantly of an experimental nature; in the second of the listed directions, only isolated studies have been carried out.

In the overwhelming majority of works, inorganic salts have been used as objects; their decomposition proceeds with rupture of a bond of strongly polar character as the primary stage and with liberation of a metal in the presence of a metallic cation [1]. The purpose of the present review is to systematize and consider from a single point of view the body of data on changes in the stability of this class of compounds. A possible general approach to solving the problem and a systematization of individual groups of experimental works in this field were presented by us earlier in [2].

I. On the Change in the Rate of Thermal Decomposition by Donor–Acceptor Effects

Within the framework of the band model, the thermal (as well as photo-) decomposition of the group of solid salts under consideration may, as a first approximation, be represented as a sequence of elementary processes characterized by excitation and displacement of electrons and ions of the compound. For the simplest case of dissociation of a salt \(K^{+}A^{-}\), at the initial stage of the process we have:

1) detachment of an electron from an anion with formation of a free electron and a hole: \(A^- \rightleftarrows A^\circ + e\), or \(A_T^- \rightleftarrows \square + e\), and if the reaction is irreversible, then \(A_T^- \to A^\circ + e\), where the subscript \(T\) refers to a special state of the ion at the surface, in a defective lattice site, or near it. The formation of a free electron may proceed through a stage of electronic excitation, which, according to [5, 6], is characteristic of heavy-metal azides. One may also assume stepwise excitation of the electron into the conduction band with the participation of local “bridge” levels located in the forbidden band [7], or levels of an excited anion \((A^{-*})\);

2) capture of an electron by some defect with an effective positive charge, for example by a cation in a special position:

\[ K_T^+ + e \rightleftarrows K^\circ \]

and secondary reactions of the hole and \(K^\circ\):

3) \(A^\circ \rightleftarrows\) Pr. 1;
4) \(K^\circ \rightleftarrows\) Pr. 2, where Pr. 1 and Pr. 2 are final products.

Table 1 shows what the qualitative effect will be of a change in the concentration of electrons in the reacting crystal on the rate of its ther—

Table 1

Effect of changes in the concentration of electronic defects on the rate of thermal decomposition

The fundamental possibility of limitation by stages 1 and 3 in the thermolysis of certain salts was indicated in [8—10] for the case when the most energy-consuming stage is reaction 1.

chemical decomposition, if the rate of the process is limited by one of stages 1–4. \(E_0\), \(E_g\), and \(E_a\) are, respectively, the activation energies of the decomposition of the pure compound, of the compound containing an electron donor and acceptor impurity; \(E_I\) is the activation energy of the \(I\)-th stage.

The conclusions of the table follow directly from the assumption of irreversibility of the limiting stage and equilibrium (or proximity to it) of all the other reversible reactions. These conclusions are discussed in detail in [2]. They require, as is evident, quite expected regular changes in the rates of the processes, the direction of which (acceleration or retardation) is determined by the type of limiting stage of decomposition and by the type of action (donor or acceptor). This opens a clear approach to solving the problem of directed change in the stability of solid compounds. Theoretically, however, it is impossible to estimate in advance the magnitude of the expected effects. One can only suppose that, under a sufficiently strong action leading to a manyfold change in the concentration of free electrons and holes in the crystal as compared with the intrinsic equilibrium concentration, these effects should be significant.

Experiments on donor–acceptor influence also open another attractive possibility—the determination of the nature of the stage limiting decomposition from the sign of the influence of donor–acceptor actions on the rate of the process. In combination with other electrical and optical measurements, this method is very useful for determining the limiting stage of decomposition.

The changes in the concentration of free carriers in the decomposing compound required for regulating the rate of decomposition can be achieved by various methods.

Below are set forth the principal methods of donor–acceptor actions on solids that give the most significant effects in changing their rate of decomposition, and the corresponding literature and the authors’ own experimental materials on these questions are systematized.

II. Catalysis and inhibition of decomposition processes by homophase impurities

The introduction of impurity ions into the crystal lattice of the decomposing compound can substantially change the conditions under which individual elementary stages of decomposition processes proceed. In certain cases, the principal causes are the donor–acceptor influence of impurities and the changes they produce in the electrophysical properties of the compound. The regularities of the change, by homophase impurities (forming solid solutions), in the physical properties of inorganic salts, as well as in their thermal, photo-, and radiation stability, have been considered earlier [1, 4]; in this connection, in the present review we shall turn only to experimental results obtained to date in the field under discussion.

In this respect, richer results that are convenient for theoretical analysis can be obtained in the study of processes in anomalous solid solutions [4]. Recent works mostly pertain precisely to this case. The resultant donor–acceptor influence of impurity ions in anomalous solutions, and therefore the nature of their influence on the rate of decomposition processes, are determined both by the intrinsic properties of the introduced impurity ions and by the donor–acceptor properties of the accompanying ionic defects; the concentration of the latter, in the formation of anomalous solutions, changes

by virtue of the compensation rule for excess (or deficient) introduced charges.

In this situation, fundamentally important information (both for proving the very fact that a solution has formed and its type, and for determining the resulting donor–acceptor properties of the impurity) is provided by studying a group of electrophysical properties of the solutions—dark and photoconductivity, thermo- and photo-emf, etc. For these reasons, the studies carried out, as a rule, include electrophysical investigations of the nature of the solutions under study. Characteristic in this direction is our study of several anomalous oxalate and azide solid solutions [11—19]. The polythirms of electrical conductivity obtained in these works for the systems $\mathrm{Ag_2C_2O_4}$ ($\mathrm{Cd^{++}}$, $\mathrm{Pb^{++}}$), $\mathrm{AgN_3}$ ($\mathrm{CO_3^{=}}$, $\mathrm{Pb^{++}}$) and $\mathrm{PbN_6}$ ($\mathrm{Ag^+}$, $\mathrm{Cu^{++}}$) made it possible to show that silver oxalate and azide are Frenkel-defective, with predominantly mobile interstitial ions $\mathrm{Ag_z^+}$, whereas $\mathrm{PbN_6}$ is disordered, most probably, according to Schottky. In some of the systems listed, substitutional solid solutions are formed. Measurements of photoconductivity and thermo-emf showed that, in solutions of $\mathrm{AgN_3}$ ($\mathrm{CO_3^{=}}$, $\mathrm{Pb^{++}}$) and $\mathrm{Ag_2C_2O_4}$ ($\mathrm{Cd^{++}}$, $\mathrm{Pb^{++}}$), the changes in the electrophysical properties of the salts caused by additives are associated mainly with their influence on the concentration of the salts’ intrinsic defects—interstitial ions—and not with the intrinsic donor properties of the introduced ions. Accordingly, the introduced cations of higher valence decrease the concentration of interstitial cations ($\mathrm{Ag_z^+}$) and thus act as effective donors of electrons in the compounds being decomposed. These additives lower the rate of thermolysis of the compositions, on the basis of which an assumption arises concerning the effective donor character of the elementary stage that limits the thermolysis process of $\mathrm{AgN_3}$ and $\mathrm{Ag_2C_2O_4}$. The likelihood of this is confirmed when studying the effect of heterophase additives and electric fields on the thermal stability of $\mathrm{AgN_3}$ and $\mathrm{Ag_2C_2O_4}$ (see below).

In the systems $\mathrm{PbN_6}$ ($\mathrm{Cu^{++}}$, $\mathrm{Ag^+}$), it is more probable to assume the formation of interstitial solutions; in this case photoconductivity indicates the donor properties of $\mathrm{Ag^+}$ and the acceptor character of $\mathrm{Cu^{++}}$ ions in the $\mathrm{PbN_6}$ matrix. The slowing of thermolysis by the $\mathrm{Ag^+}$ impurity, acceleration by the $\mathrm{Cu^{++}}$ additive, and a qualitatively similar influence on the rate of semiconductor donor–acceptor impurities give grounds for assuming the donor character of the rate-limiting stage of $\mathrm{PbN_6}$ thermolysis.

In works [17, 20] information is given on the influence of impurity ions on the rate of thermolysis of $\mathrm{Ag_2SO_3}$, $\mathrm{BaN_6}$, and $\mathrm{CuN_6}$, and, for the systems $\mathrm{Ag_2SO_3}$ ($\mathrm{Pb^{++}}$, $\mathrm{VO_3^-}$, $\mathrm{VO_4^{-3}}$), also data on the conductivity of solid solutions. In view of the absence of results on the effect of additives on other electrophysical properties of these salts, it is not yet possible to speak definitely about the donor–acceptor character of the additives and the type of rate-limiting stage of the processes.

Works [21—24] are devoted to questions of thermolysis of certain doped compositions based on alkali-metal azides, in which it is shown that obtaining azides from solutions containing $\mathrm{Ca^{++}}$, $\mathrm{OH^-}$, and $\mathrm{Cl^-}$ leads to a slowing of their thermolysis; the addition of $\mathrm{Fe^{+3}}$ ions to the starting preparations, by contrast, causes a noticeable acceleration of the process. The authors, while relating the observed effects to the formation of solid solutions in azide systems, did not, however, clearly establish the very fact of solution formation, and therefore uncertainty is possible in interpreting the results they obtained. The same shortcoming is also characteristic of works [23—24], where data are presented on the influence

of impurity ions on the rate of thermolysis of \(\mathrm{NH_4ClO_4}\), \(\mathrm{NaN_3}\), and also for the work [25] on the influence of \(\mathrm{Al^{+3}}\) and \(\mathrm{Li^+}\) on the rate of thermolysis of \(\mathrm{CdO \cdot H_2O}\).

In addition to the works listed above on anomalous solid solutions, a number of studies have recently been carried out on the thermal stability of isomorphous impurities and solutions close to them [26–29]. Work in this group is being conducted mainly with the aim of obtaining active mixed catalysts and ferrites. As the content of one of the components in such solutions increases, their thermal stability changes gradually (within the region of existence of solid solutions) and in a direction that depends on the ratio of the crystal-field strengths of the host cation and the impurity.

In earlier reviews [1, 4] it was already noted that the question of changes, due to impurities, in the photo- and radiation-chemical stability of inorganic compounds had been studied extremely poorly. Recently only a few works have appeared in this direction [30–36], and therefore the field remains insufficiently studied.

The experiments performed (mainly with lead and silver oxalates and azides) make it possible to pose the question of a regular relationship between the donor–acceptor properties of impurities and the nature of their influence on the rate of radiolysis.

For the group of salts studied in this direction, acceleration of the process by electron donors and retardation by acceptors are observed (the donor–acceptor character is evaluated from the change in the photoconductivity of the salts caused by the additives and from indirect considerations). The existence of such a dependence, opposite in character to that observed in the case of thermolysis of salts, was first pointed out in [14]. In view of the insufficient study of the effects, it is still impossible to speak with certainty about the causes of these differences; they are apparently connected with differences in the mechanisms of thermolysis and radiolysis of salts, or with a difference in the nature of the rate-limiting stages of the processes. It should be noted that retardation of the radiolysis of salts by homophase electron acceptors is probably a general regularity characteristic not only of halides, oxalates, and azides. This is indicated by the results of work [37] on the retardation of radiolysis by an electron acceptor in a nitrate matrix [38].

III. Contact phenomena in solid-phase catalysis

1. The greatest number of works on catalysis of reactions in the solid phase has been carried out for the photolysis of silver halides. The principal effect observed in the catalysis of photochemical decomposition, which is of greatest interest from both the theoretical and practical points of view, is sensitization of decomposition in the long-wavelength region of the spectrum. Several reviews [39–52] have been devoted to the mechanism of this phenomenon, as well as to the related phenomenon of sensitization of photoconductivity. In the present work we shall briefly consider the main proposed mechanisms of photosensitization and discuss their relation to the mechanisms of catalysis of the thermal decomposition of solids.

Two principal mechanisms of spectral sensitization have been proposed—the energy-transfer mechanism and the electron-transfer mechanism.

Apparently, the mechanism of spectral sensitization that includes the stage of electron transfer from a dye excited by light into the conduction band of a silver halide crystal was first proposed by Sheppard [53], and in a more modern form by Gurney and Mott [54]. They explained the main experimental facts: the shift of the region of spectral sensitivity upon adsorption of a dye and the co-

the ability of the adsorbed dye to transfer the absorbed energy many times to the silver halide crystal.

According to this mechanism, the level of the conduction band of the crystal and the first excited level of the adsorbed dye must coincide. Regeneration of the dye, according to Gurney and Mott, occurs because, in the field of the positively charged dye molecule, the nearest \(Ag^+\) ion is removed from the halide lattice; in its absence, the transfer of an electron from a bromine ion to the dye molecule is facilitated.

In view of the fact that calculations (by Coulson) gave a value for the ionization potential of dye molecules of the order of \(7\)—\(8\) eV, which, in turn, indicates the impossibility of electron transfer from the dye to the conduction band of \(AgBr\), Mitchell [55] proposed a more detailed electron-transfer mechanism, according to which an electron from the valence band of \(AgBr\) or from a surface \(Br^-\) ion occupies the vacant level of an excited dye molecule. The energy levels of the negatively charged molecule are shifted upward, so that transfer of an electron from the molecule to the acceptor level of \(Ag^+\) becomes possible. The dye molecule thereby returns to its initial state, and the electron may be further excited into the conduction band by thermal vibrations.

Der and Scheibe [56], on the basis of the experimentally found [57] electron affinity energy of many dyes, which proved to be equal to \(\sim 3.4\) eV—somewhat less than the corresponding value for \(AgBr\)—consider it possible to apply a modified Gurney–Mott model [54] to explain the new experimental results. In this scheme, in contrast to scheme [54], it is assumed that reduction of the dye molecule occurs through transfer of an electron from the donor level of \(AgBr\).

Since a potential barrier must exist between the molecule and the crystal [45, 50], penetration of an electron from the excited level of the dye molecule into the crystal is explained by a tunneling transition.

Recently, in connection with the discovery and study of the photovoltaic effect at sensitizer–semiconductor contacts, a hypothesis has appeared concerning an \(n\)—\(p\)-mechanism of sensitization for thick (more than monomolecular) dye films [48, 49]. According to this hypothesis, spectral sensitization is associated with the photovoltaic effect at the semiconductor–dye boundary, and the sign of the effect is determined by the relative positions of the Fermi levels [49] in them before contact, i.e., by the sign of the contact potential difference: the semiconductor with the larger work function must, upon illumination, become enriched with holes, and the one with the smaller work function—with electrons. A second component of the photo-emf appears when there is a difference in the widths of the forbidden bands of the contacting substances [58], and the sign of the charge in this case is determined by the type of conductivity of the semiconductors and by the relative positions of their allowed levels.

In [49] a rule is given that determines the sign of the photo-emf: upon illumination, the contacting semiconductors become charged in such a way that in each of them the concentration of its own carriers increases; and if the majority carriers in them have the same sign, then their concentration increases in the substance in which it was greater in the dark. A number of experiments, in which the photoconductivity and the photovoltaic effect were measured on one and the same sandwich of a sensitized semiconductor and a dye, confirm this rule for several systems [49, 51].

In our opinion, only that component of the photo-emf which is associated with the dif-

...by the difference in the relative positions of the levels of the contacting semiconductors. The photo-e.m.f. component associated with the difference in work functions, apparently, cannot be responsible for the chemical effects of sensitization, since when the sensitizer, in contact with the substance being decomposed, is illuminated, all that occurs is the return to the latter of charge carriers that had passed from it into the sensitizer during formation of the double layer.

Many sensitization effects are consistent with the mechanism of electron transfer; however, it cannot explain several fundamental experimental facts. Thus, for example, this theory is contradicted by the circumstance that maximum sensitization is attained at such sensitizer concentrations at which a monolayer of it has not yet formed on the surface of the semiconductor; that no correlation is observed between the conductivity type of the dye (even if these are dyes of the same chemical composition) and the sensitization effect [44—46, 49, 59]; and that the sensitization spectrum corresponds to the absorption spectrum of the molecules, and not of the solid dye.

In this connection let us consider the mechanism of energy transfer, which does not have the indicated shortcomings. According to this mechanism, the energy absorbed by the dye is transferred to the semiconductor lattice and in the latter produces excitation of an electron. In this mechanism the problem of dye regeneration also disappears completely. It is necessary, however, to explain how an energy quantum significantly smaller than in the intrinsic process produces excitation and ionization in the presence of a sensitizer. There are several solutions to this problem. According to Franck and Teller [60], this occurs on the basis of the Franck–Condon principle (Fig. 1). The excited state of the silver-halide lattice has an energy minimum at larger interionic distances than the ground state. In addition, the curve of the excited state is flatter. If the halogen ion is located at its own site, then excitation requires a quantum of violet light (transition I in Fig. 1). A point on the potential curve of the ground state with abscissa \(r_2\), corresponding to the minimum of the excited-state curve, can be reached as a result of thermal vibrations of the lattice. The corresponding optical transition (II) occurs with an energy gain equal to the sum of transitions (IV) and (III), and excitation becomes possible under illumination with red light.

Fig. 1. Dependence of the excitation energy on the distance between ions in the lattice

In the case of sensitized photolysis, transitions (II) become more probable than in direct photolysis, since the lifetime of the excited dye is sufficiently long for the required configuration of silver and bromine ions to be attained as a result of thermal fluctuations.

This mechanism requires a decrease in sensitized photosensitivity with decreasing temperature, and this dependence should be stronger than for the intrinsic process. A number of experimental data confirm this theory. It also explains the passage of photosensitivity through a maximum with decreasing temperature, found in [61].

To explain the lowering of the energy for the sensitized process, Scheibe proposed a mechanism of accumulation of several light...

by quanta by the dye, with subsequent transfer of all the accumulated energy to the silver halide crystal. But, as shown in [39], this mechanism is not confirmed experimentally, since the photosensitivity depends on the first power of the intensity, and not on a higher one.

Meiklyar and Stepanov [62] explain the decrease, in sensitized photolysis, of the required energy compared with the intrinsic process by an increase in the energy of the light quantum absorbed by the dye molecule due to the thermal vibrations of the molecule.

According to Terenin [40], the energy required to detach an electron in the presence of a dye may be considerably less than without it, since on the crystal surface there are defective regions at which the ionization energy of the anions is appreciably lower than in an undisturbed lattice. In this case, however, it is necessary to explain why, in the absence of a dye, the electron is not excited at these defective sites and, in addition, the problem arises of the regeneration of defect levels. The following explanations can be given for these questions. First, optical transitions between local levels and allowed bands may be forbidden, but at defects they may occur with appreciable intensity [63]. Owing to the superposition of the electron clouds of adsorbed dye molecules and defective ions on the semiconductor surface, a favorable configuration may also be created for such transitions. As for the regeneration of donor levels, as Carroll and others [64, 65] have shown, these levels must belong to the very surface of AgBr crystals, since chemical sensitization does not affect sensitization by a dye; thus the regeneration of donor levels may be represented simply as a continuous renewal of the surface during decomposition of the crystal.

Comparing these two mechanisms, it may be noted that the first of them—the transfer of an electron—has more contradictions with experiment than the second. This difference, however, depends on the strictness of the requirements imposed by the two mechanisms. In the case of the electronic mechanism the requirements are more stringent: for electron transfer it is necessary that the excited level of the dye be above, or at the same level as, the conduction band of the crystal; in addition, closer contact between the dye and the crystal is required. For the realization of the energy-transfer mechanism, it is only necessary that electronic levels exist in the sensitized crystal, the separations between which would be equal to (or less than) the excitation energy of the dye; the requirements on the density of contacts are somewhat less here. However, there are far more possibilities for experimental verification of the electronic mechanism than of the energy-transfer mechanism, and, apparently, in this connection it has more supporters. Recently Terenin and coauthors [44, 46, 59] have been systematically attempting to prove the untenability of the electronic mechanism by studying electronic transport phenomena. Comparing these two mechanisms, let us mention one more effect—the desensitization of photolysis by many dyes in the region of their intrinsic absorption [19—41, 66—69]. This phenomenon, as far as we know, is explained only from the standpoint of electron transfer. Thus, in this respect the electron-transfer theory proves to be more general.

There are also “reconciling” points of view in the literature. Thus, if one takes into account the mixing of the wave functions of the adsorbed dye and the crystal surface [66] (as well as the indistinguishability of the electrons forming the adsorption bond), this may remove the contradiction between the two mechanisms.

Terenin and Akimov [44] believe that if electron transitions by the Mitchell mechanism [55] take place in a time shorter than \(10^{-9}\) sec,

the Mitchell sensitization mechanism is no different from the energy-transfer mechanism. The resolution of this dispute, in our opinion, would be aided to a considerable degree by experiments on the influence of an electric field on the efficiency of sensitization by a dye layer. We shall return to this question below.

1. Let us next discuss the question of the applicability of the contact concepts developed for the case of photosensitization to the thermal decomposition of solid inorganic compounds in the presence of solid catalysts. Obviously, there is no reason in this case either to give preference to the mechanism of electron transfer or to that of energy transfer. In principle, apparently, both mechanisms are possible. However, there must also be differences associated with the specific nature of the action of radiation and of heat.

For example, in the case of the action of heat, the Frank–Teller sensitization mechanism is inapplicable, since even in the absence of a sensitizer excitation by heat will proceed along path (II) (Fig. 1). If the process takes place under conditions of thermal equilibrium, redistribution of electrons by a mechanism similar to that of $p—n$ sensitization is also impossible. If, however, such equilibrium is not attained (for example, owing to the exothermicity of the reaction), then a redistribution of charged particles will occur between the cold and hot regions in accordance with the laws of thermoelectric phenomena.

The mechanisms of energy transfer (Terenin) and of electron transfer (Mott, Mitchell, Scheibe, and Dörr) are just as valid for catalysis of thermal decomposition as they are for photosensitization. The matter is complicated, however, by the fact that the mechanisms of photosensitization were proposed and are being studied using examples of contact with organic dyes, whereas the principal results in the case of thermal decomposition have been obtained for contacts with inorganic semiconductors. We shall not dwell for the time being on the possible differences that follow from this; we note only that such a discussion requires comparative data on the influence of organic and inorganic semiconductors on photolysis and thermal decomposition. Such studies have only just begun. One may point, for example, to work [70], in which the influence of organic semiconductors on the thermolysis of silver oxalate was investigated, while works [71, 72, 33] present some experiments on the influence of inorganic semiconductors on the rate of photolysis.

Below we shall consider the principal mechanisms proposed earlier for catalysis of thermal decomposition by inorganic semiconductors.

In contrast to the above-discussed and essentially related field of photochemical sensitization, not many extensively developed mechanisms have been proposed for the catalysis of the thermal decomposition of solids by heterogeneous admixtures.

In a number of works, including recent ones [73—79], in explaining the causes of the influence, the authors go no further than general statements about the connection between catalytic activity and the electronic properties of the catalysts.

Apparently, the first attempt to examine somewhat more closely the mechanism of catalysis of thermal decomposition by heterogeneous admixtures is work [80]. Here, for the catalysis of the thermal decomposition of ammonium perchlorate by manganese dioxide, the authors propose a “bridge” mechanism, analogous to the mechanisms of Mitchell and Mott; i.e., a metal ion of variable valence, in their opinion, is an acceptor and donor of an electron. Here too, however, the matter did not go beyond this statement, and the consequences of the hypothesis developed were not discussed in the work. Thus, for example, it is unclear whether in this case the activation energy of thermolysis should change.

In works [77—79] it has been shown that correlations exist between the catalytic effect of an impurity and the type, as well as the magnitude, of its electrical conductivity. However, the reasons for such a correlation have not been discussed in detail.

In our opinion, the cause of a change in the rate of thermal decomposition may be a group of contact processes and phenomena at the catalyst—decomposing substance interface.

If the decomposing substance is mechanically mixed with another solid substance, then we are dealing with a system of two solid bodies in contact, a system that has been widely studied in semiconductor physics [83—87]. Let us consider here the most general case: the contact of two semiconductors with different electron work functions. Figure 2 shows the band diagram of semiconductors 1 and 2 before contact. The electron work function

Fig. 2. Band diagram of the contact of two solid substances.
a — before contact, b — in contact.

of semiconductor 2—$\varphi_2$—is smaller than $\varphi_1$. Therefore, upon contact of the semiconductors, some of the electrons will pass from substance 2 into substance 1. The transfer will continue until the Fermi levels $F_1$ and $F_2$ become equal. The level $F_1$ will rise, and $F_2$ will fall relative to the bottom of the valence band in the contact layer. In this case, a positively charged layer arises on the right, and a negatively charged layer on the left. The potential difference between these layers is determined by the difference in the electron work functions

\[ V=\frac{\Delta \varphi}{e}=\frac{\varphi_1-\varphi_2}{e}. \tag{1} \]

The quantity $V$ is the contact potential difference, and $e$ is the electron charge.

The thickness of the contact layer on each side, $L_1$ and $L_2$, is determined by the concentration of charge carriers and by the contact potential difference. For example, if semiconductor 1 has hole conductivity, and semiconductor 2 has electron conductivity, with carrier concentrations $n_1$ and $p_2$, respectively, then

\[ L=L_1+L_2= \left( \frac{\varepsilon \Delta \varphi}{2\pi e^2}\cdot \frac{n_1+p_2}{n_1\cdot p_2} \right)^{1/2}, \tag{2} \]

where $\varepsilon$ is the dielectric constant,
$e$ is the electron charge.

\[ \frac{L_1}{L_2}=\frac{p_2}{n_1}, \tag{3} \]

\[ L_1=L\cdot\frac{p_2}{n_1+p_2}, \qquad L_2=L\cdot\frac{n_1}{n_1+p_2}. \tag{4} \]

It is evident from this that the thickness of the space-charge layer is the greater, the lower the concentration of charge carriers, and the contact field penetrates to a greater depth into the region where the carrier concentration is lower. For typical semiconductors \(L = 10^{-2} \div 10^{-5}\) cm, while for dielectrics it may in principle reach several centimeters [85]. However, the presence of a negligible amount of impurities leads to the fact that the thickness of the space-charge layer in a dielectric usually does not exceed \(10^{-4}\) cm [88]. If one of the semiconductors (1) has a much lower conductivity than the other (2), then practically the entire space-charge layer will lie in the first, and then

\[ L_1 \simeq L = \left( \frac{\varepsilon \Delta \varphi}{2 \pi e^2 n_1} \right)^{1/2}. \]

If now one of the substances (for example, 1) enters into a chemical reaction, then the reaction rate in the layer enriched (or depleted) with electrons will change in comparison with the bulk, if electronic defects influence the rate-limiting stage. For a noticeable change in the rate of decomposition of solids by solid catalysts, it is important that the thickness of the space-charge layer and the intensity of the electric field created in the reacting substance be as large as possible, and that this layer be preserved for a considerable time in the course of the chemical reaction. Then very small amounts of additive are sufficient to change the rate of decomposition throughout the entire volume of the crystal. The strength of the electric field in the irradiated substance increases with increasing \(\Delta \varphi\) and with improvement in the quality of the contact.

Consideration of the solid-phase catalysis of dissociation reactions of solids by semiconductors, taking into account the contact phenomena set out above, makes it possible to determine the range of those properties of catalysts which will chiefly determine their activity in the decomposition reaction.

The magnitude of the catalytic effect in this case will be determined primarily by the contact potential difference between the catalyst and the substance under investigation. Therefore one should expect the most distinct parallelism between the influence of a semiconductor on the decomposition rate and the magnitude of the work function from it. The catalytic effect should also depend very substantially on the quality of the contact between the particles of the catalyst and of the principal substance. The relative decrease in the thickness of the electric double layer as a result of the voltage drop in the gap becomes noticeable when the gap \(d\) is comparable with the quantity \(\dfrac{L}{2\varepsilon}\) [85].

Since \(\varphi\) of semiconductors depends strongly on their dispersity [89], the magnitude of the change in the decomposition rate will depend on the particle sizes of the catalyst and of the decomposed compound that are in contact. It is also obvious that the magnitude of the effect will depend on other factors that change \(\varphi\): adsorption of vapors and gases, impurities introduced into the lattice of the contacting bodies, and the composition of the gas atmosphere.

The influence of the electrical conductivity of the catalyst and of the decomposed compound on the effect of changing the dissociation rate by semiconductor catalysts may also be substantial.

It is still difficult to determine in advance the character and degree of the influence of the electrical conductivity of the decomposed substance on the effectiveness of catalysis. To a large extent this character will depend on the mechanism of decomposition of the compound—whether decomposition begins at the surface or throughout the entire volume, since in the first case, for a substantial change in the decomposition rate, it is sufficient to change the electron concentration only in a thin

layer near contact with the semiconductor, while in the second case it is necessary that the value \(L\) be of the order of the thickness of the crystals of the compound being decomposed.

The existence of a dependence of the catalytic effect exerted by semiconductors on the rate of decomposition on the quality of the contact and on the electrical conductivity of the contacting systems may somewhat disturb the parallelism between the catalytic effect and the electron work function from the semiconductor; nevertheless, the latter certainly remains the most important characteristic of catalyst activity. Here it is useful to note that the influence of the last two factors can lead only to certain quantitative changes in catalyst activity, whereas the sign of the influence (acceleration or retardation of the process) is determined entirely by the ratio of the work functions of the contacting systems (and by the nature of the rate-limiting stage of the decomposition process).

The literature data, as well as our experimental results, concerning the influence of heterogeneous semiconductor admixtures on the rate of thermal decomposition of solids confirm the concepts developed above regarding the decisive role of contact phenomena in the process of semiconductor catalysis of decomposition reactions. The order of change in the catalytic activity of semiconductors in the thermal decomposition reactions of a number of the most thoroughly studied salts is shown in Table 2. From the results presented it is evident that the order of increase in the activity of oxides in the decomposition reactions of \(\mathrm{KClO_4}\) and \(\mathrm{KMnO_4}\), from \(\mathrm{CdO}\) to \(\mathrm{NiO}\), generally coincides with the increase along this series in the electron work function from the semiconductor, while oxides with insignificant conductivity do not catalyze decomposition (comparison with Table 3). The activity of oxides with respect to the change in the activation energy of thermal decomposition of \(\mathrm{KClO_4}\).

Table 2

Catalytic activity of semiconductors in the thermolysis reaction of certain solid substances. The activity of the catalysts increases from bottom to top. Below the first line are inert admixtures; below the second, inhibitors. Results obtained in one work are marked by a brace; mechanical mixtures without additional treatment are marked by parentheses

Continuation of Table 2.

also increases as the electron work function increases. The change in the effect upon introducing mono- and trivalent impurity ions into NiO occurs in parallel with the change in the magnitude of the work function \(\varphi\).

Semiconductors that catalyze the decomposition of ammonium perchlorate (AP) are difficult to arrange in a single series, since the investigations were carried out by different authors using various samples of AP and semiconductors. In Table 2 these data are shown separately for each work. Within each work, here too the rule is for the most part preserved: the greater the work function, the greater the catalytic effect. Doping CuO with lithium (which should increase the work function) increases the catalytic effect, whereas doping with chromium decreases it. Oxides with low conductivity and a low electron work function exert little influence. An exception is the strong catalytic-

... the influence of ZnO and MgO, which have a comparatively small electron work function. In these cases the accelerating effect is associated with the chemical interaction of the oxides and AP [90—92].

Table 3

Electron work function \((\varphi)\) from solid inorganic substances. \(\varphi^*\) calculated from our data at \(\varphi_{\mathrm{Au}} = 4.8\ \mathrm{eV}\)

A similar picture is observed in the case of C\(_3\)H\(_4\)(NH\(_4\))\(_2\)HClO\(_4\): the greatest catalytic effect is exerted by semiconductors with the largest electron work function [93].

It should be noted that the literature data cited often show a one-sided influence of mechanical impurities on the thermolysis of solids—namely, acceleration. This is possibly associated with the small electron work function from the decomposing substances (AP, KClO\(_3\), KClO\(_4\), KMnO\(_4\)). But in the decomposition of lead azide, barium azide, copper(I) azide, and silver oxalate, both sensitization effects and rather strong inhibition effects are observed [2, 3, 16, 17, 103]. The general qualitative conclusion that can be drawn from the literature data is as follows: the strongest catalysts of the thermolysis of PbN\(_6\), AgN\(_3\), and Ag\(_2\)C\(_2\)O\(_4\), as well as of AP, KClO\(_3\), KClO\(_4\), and KMnO\(_4\), are semiconductors with a large work function. Semiconductors with a small work function, such as Ag\(_2\)S, CdO, and ZnO, on the contrary, retard the decomposition of AgN\(_3\), PbN\(_6\), and Ag\(_2\)C\(_2\)O\(_4\). In these cases the decomposition process is determined by the stage of decomposition of radicals of type A° (see p. 36) [101—103].

The results obtained by us for BaN₆ and CuN₃ are opposite to those cited above [17, 103]: here electron donors accelerate, while acceptors slow down decomposition. The acceptor stage limits the rate of thermal decomposition of these salts [103]. Doping a number of semiconductors leads to regular changes in the catalytic effect. However, in a number of cases no quantitative correspondence was observed between the magnitude of the work function and the catalytic effect. For example, specially prepared ZnO samples (from the hydroxide and with a lithium impurity) had a work function smaller than that of CoO and Co₂O₃ (Table 3), but the catalytic effect of these samples on the thermal decomposition of PbN₆ was considerably smaller than that of cobalt oxides, if the studies were carried out on preparations obtained by simple mechanical mixing. Such a difference in effects is difficult to explain by differences in the electrical conductivities of the oxides or by the quality of the contact. Moreover, a consequence of a simple \(p\)—\(n\) mechanism of catalysis is invariance of the activation energy of decomposition in the presence of the catalyst. In a number of works, however, the activation energy changes [74—76, 102]. This is true, for example, for the decomposition of silver azide in the presence of cobalt oxides [102].

A change in the activation energy may in principle be associated with the temperature dependence of the amount of electrons or holes transferred from the catalyst or from the semiconductor, owing to the temperature dependence of the contact potential difference. The temperature dependences of \(\Delta \varphi\) of the substances being decomposed and of semiconductors studied by us are presented in Table 4, from which it is evident that with increasing temperature \(\varphi\) of the azides increases, whereas for the oxides it changes only slightly. These data indicate the fundamental possibility of decreasing the activation energy of thermolysis in the presence of oxides.

Table 4

Temperature dependence of CPD \((P = 10^{-2}\ \mathrm{mm\ Hg};\ \Delta \varphi = \varphi_{\mathrm{Au}} - \varphi)\)

In the case of a bridge mechanism, a decrease in activation energy may occur if one of the two stages involving the catalyst limits the decomposition rate, while the second is also far from equilibrium.

Yet another possible reason for the decrease in activation energy by transition-metal oxides is connected with their magnetic properties. This reason consists in the fact that, under the influence of strong local magnetic fields of paramagnetic atoms, the prohibition on electronic transitions occurring with a change in spin is lifted [116].

Taking into account the special properties of oxides of \(d\)-elements, forcing us to consider them not within the framework of band theory, but by the method

molecular orbitals [117], one cannot deny the possibility of a catalysis mechanism through an intermediate complex, as is assumed for gases and liquids. But the strong influence of additives on the catalytic activity of these oxides during thermal decomposition apparently contradicts this mechanism.

1. A very limited number of works [30, 31, 32, 118—121] have been devoted to questions of catalysis and inhibition of the radiolysis process of inorganic salts by heterogeneous impurities. More detailed investigations have been carried out on the question of the change, by impurities (chiefly colloidal silver, gold, and sulfur), in the rate of radiolysis (and photolysis) of silver halides (for a review, see [43]); however, this group of works is devoted mainly to processes in photoemulsion crystals, and in them questions of deep radiolysis of salts are scarcely touched upon.

The experimental material available in the field under consideration shows that certain basic regularities in the change, by additives, of the rate of radiolysis of salts differ from those occurring for thermolysis and described above. This applies above all to the relation between the donor–acceptor properties of impurities and the direction (acceleration or retardation) in which they change the rate of the process. The clear distinctions present in this respect for homophase additives were discussed above. In the case of heterogeneous semiconductor impurities, discussion of the question of the correlation between their donor–acceptor properties and the direction of their influence on the rate of radiolysis is difficult. This is connected with the impossibility of rigorously determining the direction of electronic transitions in a system of two contacting crystals.

The conclusions in this respect that can be drawn on the basis of comparing the values of \(\varphi\), measured by the vibrating-capacitor method for each of the two substances of a contacting pair, cannot be regarded as unconditionally correct and final. To construct well-founded assumptions on this question, measurements of the electron work function from the compound in the radiation field are necessary. In this case one may hope that sufficiently correct measurements will be those performed by the vibrating-capacitor method while irradiating the salt being measured and the entire apparatus with \(\gamma\)- or X-rays of low dose rate.

The necessary (though less precise) information may also be provided by measurements of \(\varphi\) for a series of catalysts (with gradually changing \(\varphi\)) and for doped compounds, and by comparing the data thus obtained with the influence of these compositions on the rate of radiolysis [118—121].

The existence of an invariably symbatic (or antibatic) dependence between the value \(\varphi\) of an impurity and the quantitative characteristic of its influence on the rate of radiolysis may provide grounds for an assumption as to by what (donors or acceptors) radiolysis is accelerated.

To confirm the assumption thus made about the direction of electronic transitions, it is also necessary to have, in the irradiated contacting systems, data on the change in the thermodynamic \(\varphi\) of the radiolyzed compound and of the catalyst after their irradiation with doses and types of radiation corresponding to those used in radiolysis experiments.

Similar experiments, carried out by us and in part published in [30—31], [118—121], provide grounds for the conclusion that the radiation decomposition of a large group of salts is catalyzed by heterogeneous electron donors and that the process is inhibited by electron acceptors. This conclusion is also in agreement with the data described above on the influence, on the rate of radiolysis, of additives forming solid solutions.

The established differences between the regularities of catalysis of thermolysis and radiolysis of solid salts may be a consequence of differences in the nature of the elementary stages of excitation and decomposition of the complex anion in these processes.

Indeed, upon interaction of high-energy quanta with a substance, in contrast to thermal decomposition, several pathways are possible for radiolysis of a salt of the type \(K^+ A^-\):

\[ \begin{array}{rll} \text{I.} & A^- \rightsquigarrow A^0 + e; & \overset{\text{b}}{A^0 \to \text{products}.} \\[3pt] \text{II.} & A^- \rightsquigarrow A^0 + e; & A^0 + e \to \text{products}. \\[3pt] \text{III.} & A^- \rightsquigarrow A^0 + e; & A^- + e \to A^{=} \to \text{products}. \\[3pt] \text{IV.} & A^- \rightsquigarrow A^{-*} \to \text{products}. & \\[3pt] \text{V.} & Me^+ \rightsquigarrow Me^{++} + e; & Me^{++} + A^- \to Me^+ + A^0; \\[3pt] & & A^0 \to \text{products}. \end{array} \]

The occurrence of reactions I—V as primary acts of radiolysis has been assumed for various reasons, in particular for nitrates and chlorates of alkali metals [9].

In the real case of salt radiolysis, both ionization (stages I—III, a, V, a) and excitation of the anion (IV, a) proceed simultaneously. The direction of the subsequent course of stages I—III depends on a number of factors, in particular:

1) the thermal stability of the radical \(A^0\);
2) the electron affinity of the radical \(A^0\) and of the anion \(A^-\);
3) the magnitude of the energy of the weak bond in the anion \(A^-\).

In particular, at a relatively high thermal stability of the radical \(A^0\), with a high electron affinity exceeding the bond energy in the anion, process II, b will proceed with a relatively high probability.

The nature (direction) of the influence of impurities—donors or acceptors of electrons—will in a given case be determined by which primary reaction from the set of possible I, b—III b, V, b makes the greatest contribution to the formation of radiolysis products. The fraction of energy expended on excitation (IV) is also different for different irradiated compositions, but processes IV, a—IV, b should be little sensitive to impurities. In addition to the different influence of impurities, some stages may be distinguished by the expected specific influence on the radiolysis rate of the radiation dose rate (III) or by changes in the rate of accumulation of radical and molecular radiolysis products in series of salts of the same type [122].

From the standpoint of the concepts set forth, one may consider the results of the cited works on the influence of homo- and heterophase impurities on the radiolysis rate of certain salts.

It has been shown in [31—32] that impurities—probable electron donors—accelerate, while acceptors retard the radiolysis of \(\mathrm{PbN_6}\) in an X-ray field. According to preliminary data obtained by us, certain homophase additives have the same kind of influence on the radiolysis of \(\mathrm{AgN_3}\).

Since the radiolysis of heavy-metal azides (in particular, the influence of dose rate, series dependences, and kinetics of product accumulation) has been studied very poorly, it is impossible to indicate the probable pathway by which the process occurs; however, it is likely that the role of stages II and III (both, or one of the two) must be substantial.

Impurities also exert an influence qualitatively similar to that observed for azides on the radiolysis of alkali-metal nitrates [31, 37, 122,

[118–127]. The independence of \(G_0\) in nitrates from dose rate, established in a number of works (for a review, see [122]), and the increase in the radiolysis rate in the series from \(\mathrm{NaNO_3}\) to \(\mathrm{CsNO_3}\), in parallel with an increase in the cation’s electron fraction [122], make it possible to assume that the accumulation of molecular radiolysis products proceeds mainly through stage III.

It should be noted that Cunningham also arrived at a similar conclusion on the basis of a careful study of the regularities of nitrate radiolysis [123–124].

In contrast to the situation with alkali-metal nitrates, the fact established in [120] of a regular change, by semiconducting oxides, in the radiolysis rate of \(\mathrm{NH_4NO_3}\) cannot even tentatively be connected with the mechanism of the process, owing to the complete lack of study of the latter.

In several experiments we reliably established a gradual increase in the rate of evolution of gaseous products during the radiolysis of \(\mathrm{NH_4ClO_4}\) in a series of semiconductors in contact with it, with a decreasing work function of electrons from the semiconductor in the same series. In this connection it becomes desirable to obtain data on the accumulation of molecular and ionic products during the radiolysis of mixtures of \(\mathrm{NH_4ClO_4}\) with additives and on the influence of the radiation dose rate on the process.

As was already noted above, the data available in the literature indicate that the relationship between the donor–acceptor properties of additives and the nature (direction) of their influence on the rate of processes for radiolysis and thermal decomposition is opposite. Owing to the insufficient number of systems studied, it is still too early to speak of the generality of this phenomenon; however, it was observed for all the compositions studied. A comparison of conclusions concerning the probable mechanism of the radiolysis of salts with the corresponding conclusions for thermolysis shows that the causes of the observed difference in the influence of additives may be associated with differences in the elementary stages of electron excitation in the two cases.

For radiolysis, because of the high energy of the quanta, processes IV and V are possible, and, owing to the low temperatures, stage III is possible—a stage that is either impossible or less probable in the thermolysis of salts. In the region of high thermolysis temperatures, where the probability of electron capture by the trap \(A^{-}\) is small, thermal decomposition of the radical \(A^{0}\) is more probable. In this case, owing to the reversibility of stage I, the rate of a reaction of the type \(A^{0} (+A^{0}) \to \mathrm{Pr}_1\) decreases in the presence of donors, as was already noted earlier.

IV. The influence of electric and magnetic fields on the decomposition of solids

For studying the mechanism of the influence of heterogeneous additives and for studying the mechanism of decomposition of solids, an attempt to change the decomposition rate under the action of external electric and magnetic fields is of great interest. Such an influence should in principle occur, since the elementary stages of photochemical, radiation, and thermal decomposition include processes of formation, annihilation, and transport of particles possessing an electric charge. However, apparently because of experimental difficulties and the novelty of the corresponding theories even in solid-state physics, few such studies have been carried out.

Applying modern concepts of solid-state physics and considering experimental data on changes in the properties of a solid upon the application of fields, one can identify the following possible causes of their influence on the chemical processes of decomposition.

1. Influence of fields on the transport of ions and electrons

Interesting effects in this direction were observed in the photolysis of silver halides [125—126]. Thus, Klein and Moatches [125] observed the displacement of photolytic silver in AgBr crystals into the cathode region upon application of a constant electric field. This confirms the presence of an ionic stage in photolysis. Using synchronized pulses of light and voltage, Haynes and Shockley [126] observed the displacement of photoelectrons in AgCl by an electric field, which manifested itself in the appearance of columns of photolytic silver growing into the crystal in the direction opposite to the field. With the additional application of a magnetic field, the columns were bent.

From these experiments it proved possible to calculate the mobility of photoelectrons. In addition, they indicate the participation of free electrons in decomposition.

These and some other studies are described in detail in the review [43], and we shall not dwell on them in detail; we shall note only that they describe effects that depend on the direction of the electric and magnetic fields. These experiments are of great importance for determining the mechanism of photolysis. Similar studies, as noted above, can also be carried out to determine the mechanism of photosensitization by “thick” layers of dyes. For this purpose, one may apply, synchronously with the light pulses, an electric field of a definite polarity to a “sandwich”-type specimen consisting of a layer of photosensitive salt and a layer of sensitizer. If light inactive for nonsensitized photolysis is chosen for the studies, then, with positive polarity of the semitransparent electrode deposited on the dye, the transport of electrons to the decomposing substance will be hindered. The presence of an influence of a field of both polarities will indicate an electron-transfer mechanism, whereas the absence of an effect will indicate an energy-transfer mechanism.

Experiments fundamentally similar to those described above for photolysis, on the influence of a constant electric field on the thermal decomposition of Ag₂C₂O₄, PbN₆, and AgN₃, were carried out by us [127]. The small change in the decomposition rate found in this work, depending on the polarity of the field, is possibly associated with charging of the substance in the field. However, in special experiments under a microscope we were unable to observe rapid decomposition of silver oxalate at one of the poles.

An effect based on the displacement of electrons in an electric field for the case of adsorption on semiconductors was predicted theoretically by F. F. Vol’kenshtein and V. B. Sandomirskii [128]. The experimental discovery of this phenomenon for the case of adsorption of methyl alcohol on germanium belongs to N. P. Keier and E. P. Mikheeva [17]. Apparently, the influence of an electric field on catalysis in the works of P. M. Stadnik and co-workers [130—131] and others [132—133] is due to similar causes.

2. Formation of additional carriers upon application of an electric field

A change in the concentration of charged lattice defects, and consequently a change in the rate of decomposition of a solid, may occur in strong fields as a result of the Stark effect and electric ionization. But even at smaller fields, breakdown occurs for other reasons [134]. At still lower field strengths, the mechanism of Frenkel thermoelectronic ionization begins to operate. This mechanism consists in the fact that an external electric field increases the energy of an electron

in the atom by an amount \(2e \sqrt{\dfrac{eE^*}{\varepsilon}}\), reducing the energy required to detach an electron by thermal excitation by a factor of \(e^{2e \sqrt{eE/\varepsilon}}\) [134].

Under the influence of an electric field, the decay of excitons [135] and the ionization of impurity centers [82] may be facilitated. The corresponding change in resistance upon application of the field may be detected even at very small field strengths (of the order of \(1\ \text{V/cm}\) for InSb [36]). In [137] similar considerations were used to explain the acceleration of the thermal decomposition of silver oxalate when an electric current is passed through it. The appearance of additional carriers may, moreover, be caused by impact ionization by an excited current carrier, if the corresponding band is wider than the ionization energy of a valence or impurity electron [134].

Application of an external electric field leads to a decrease in the width of the forbidden band according to the uncertainty principle [138, 139] (the Franz—Keldysh effect). The theory of this effect, and all experimental data, concern the absorption of light. In L. V. Keldysh’s paper [138], however, there is a remark that an analogous shift should also be observed for the lower boundary corresponding to absorption involving phonons. In fields of \(\sim 10^5\ \text{V/cm}\), the magnitude of the shift of the absorption band, according to theoretical estimates, is \(10^{-2}\ \text{eV}\) for an effective electron mass \(m^* = m \approx 10^{-27}\ \text{g}\). In a number of papers on the electro-optical effect, such shifts have indeed been found [140—142].

In the opinion of the authors of [143], the shift of the CdS absorption band which they found in fields of \(\sim 10^3\ \text{V/cm}\) can be explained theoretically using the same principles as those employed by Franz and Keldysh.

A magnetic field, on the contrary, increases the width of the forbidden band [142]. A theoretical calculation gave, for the shift of the absorption band, a value of \(3.8 \cdot 10^{-7}\ \text{eV/s}\) for InSb. The experimental value was \(2.3 \cdot 10^{-7}\ \text{eV/s}\), i.e., this effect can be observed only in sufficiently strong fields (\(10^4\ \text{G}\)) even for semiconductors with a small effective electron mass. With increasing \(m^*\) the effect decreases. In addition, application of a magnetic field [116] or of crossed electric and magnetic fields [144] can lift the prohibition on transitions involving a change of spin, for example for singlet–triplet transitions. This may also reduce the energy required for electron excitation and lead to sensitization of decomposition. Similar ideas are used to explain the catalytic action of paramagnetic impurities [116].

3. Change in the lifetime of carriers \(\tau\) and in the mean free path \((\lambda)\) by a weak magnetic field

There are papers [145—146] in which isotropic photomagnetoelectric effects were observed. The presence of these effects, as is rightly noted in [146], can be explained only by the influence of the magnetic field on the processes of generation or annihilation of current carriers. In the opinion of the authors of [146], the field affects the decay of excitons, increasing their lifetime and the probability of decay with formation of current carriers. An influence on nucleation processes is possible through the mechanisms described above

* Here \(e\) is the electron charge, \(E\) is the electric-field strength, and \(\varepsilon\) is the dielectric constant.

(see Sec. 2), and requires strong magnetic fields. The effects described in [146], however, were observed in weak fields (up to 3000 oersteds); therefore, a more probable influence of the field is on carrier annihilation, i.e., on $\tau$ and $l$. This may occur, for example, as a result of the curvature of the trajectories of free electrons or holes on the path between two traps when a magnetic field is applied.

If thermal decomposition indeed proceeds through excitation of electrons into the conduction band, then magnetic and electric fields should influence its rate according to the mechanisms described above.

Fig. 3. Effect of an electric field (about 10 kV/cm) and a magnetic field (about 1100 oersteds) on the thermal decomposition of silver oxalate: 1 — with application of an electric field; 2 — without a field; 3 — with application of a magnetic field.

mechanisms. In addition to the experiments already mentioned [127, 137], we carried out special experiments to detect isotropic effects in the thermal decomposition of Ag$_2$C$_2$O$_4$, AgN$_3$, and BaN$_6$ [147]. It was found that the application of a constant electric field with an intensity of $\sim 10^4$ V/cm to the indicated salts causes a slight acceleration of their thermal decomposition. Figure 3 shows the results for Ag$_2$C$_2$O$_4$. Similar results were obtained at several temperatures and on other salts. The application of a magnetic field with an intensity of 1100 oersteds caused a slowing of the decomposition of Ag$_2$C$_2$O$_4$ (Fig. 3), an acceleration for BaN$_6$, and did not affect the thermal decomposition of AgN$_3$. Proceeding from the proposed mechanism of decomposition of these salts [2, 101—103] and differences

in the limiting stage for $\mathrm{BaN}_6$ [103], one may conclude that the electric field accelerates the thermolysis of these three salts by increasing the probability of excitation of an electron from the anion. Determination of the activation energy of decomposition in a field and without a field could serve to test this point of view. In the case of $\mathrm{Ag}_2\mathrm{C}_2\mathrm{O}_4$ we were able to determine a decrease in this quantity upon application of an electric field ($\Delta E \simeq 1$ kcal). In other cases the scatter of the experimental data was considerably greater, so that we could establish only the sign of the effect.

The different influence of a magnetic field on all three substances can be explained if one assumes that the magnetic field changes (increases) the lifetime of free electrons. At an unchanged excitation intensity this should lead to an increase in the electron concentration, which should cause an increase in the decomposition rate of $\mathrm{BaN}_6$ and a decrease in it in the case of $\mathrm{Ag}_2\mathrm{C}_2\mathrm{O}_4$ and $\mathrm{AgN}_3$, i.e., the magnetic field should act in the same way as additives—electron donors. Such a parallelism in the influence of additives [101—103] and of the magnetic field [147] is indeed observed.

As is evident from the present review, a fairly large number of published works have been devoted to the problem of changes in the thermal, photo-, and radiation-chemical stability of solid inorganic compounds with polar bonds. Almost all this material, however, concerns changes in the thermal stability of salts with complex anions by homophase and heterophase semiconductor impurities, and changes in the photosensitivity of silver halides by homophase impurities and organic dyes. In the remaining sections (changes in radiation-chemical stability, photosensitivity in the presence of heterophase inorganic additives, thermal stability in the presence of organic dyes) there is still an extreme scarcity of work.

It seems desirable to fill this gap, since this will help in understanding the mechanisms of changes in stability not only for processes still unstudied experimentally, but will also provide rich material for comparison with processes that have been studied in detail.

In this review we have attempted to consider the available data within the framework of the band model of the solid, adopting a single scheme (mechanism) of decomposition for all three processes, modified somewhat only for radiolysis of salts. The primary stage of this mechanism is the excitation of an electron and its transfer from the anion to the cation, and in most cases this stage must be reversible. This simplified scheme, apparently most applicable to binary compounds, makes it possible to explain experimental data on changes in stability also for more complex substances; however, the real mechanism of decomposition in many cases evidently differs substantially from it. This, for example, has been shown for photolysis and radiolysis of azides and other salts with complex anions [148, 122, 43].

The mechanism of thermal decomposition also cannot be regarded as firmly established.

Thus, V. I. Oshcherov [149] considers that the first stage of thermal decomposition is not the transfer of an electron into the conduction band, but the displacement of an ion, and that transfer of an electron into the conduction band is not obligatory. This point of view seems to us very interesting, but it must be reconciled with the available experimental data and additional experiments must be carried out.

References

1. V. V. Boldyrev. The influence of defects in crystals on the rate of thermal decomposition of solids. Tomsk State University Press, Tomsk, 1963.

2. Yu. A. Zakharov, G. G. Savel’ev. Kinetics and Catalysis, 7, 55, 1966.

3. G. G. Savel’ev, Yu. A. Zakharov. Izv. Vuzov, Chemistry and Chemical Technology, 7, 768, 1964.

4. Yu. A. Zakharov. Dissertation, Tomsk, 1963; S. M. Ryabykh. Dissertation, Tomsk, TPI, 1968; G. T. Shchechkov. Dissertation, Tomsk, 1969.

5. B. L. Evans, P. Gray, A. D. Joffe. Chem. Rev., 55, 515, 1959.

6. P. Gray. Quart. Rev. Lond. Chem. Soc., 17, 441, 1963.

7. V. V. Boldyrev. Abstracts of the All-Union Conference on Solid-State Chemistry, Moscow, 1965.

8. Mclaren, G. Rogers. Proc. Roy. Soc., A246, 250, 1958.

9. Bartlett, F. C. Tompkins, A. D. Joung. Proc. Roy. Soc., A246, 206, 1958.

10. A. Galwey, P. Jacobs. Trans. Farad. Soc., 55, 1165, 1959.

11. Yu. A. Zakharov, G. G. Savel’ev, V. V. Boldyrev. Kinetics and Catalysis, 5, No. 5, 1964.

12. Yu. A. Zakharov, A. A. Kabanov. ZhFKh, 38, issue 12, 1964.

13. Yu. A. Zakharov, G. G. Savel’ev, V. K. Zhuravlev, V. V. Boldyrev. Kinetics and Catalysis, 6, No. 3, 1965.

14. Yu. A. Zakharov, G. G. Savel’ev. Kinetics and Catalysis, 6, No. 4, 1965.

15. G. G. Savel’ev, Yu. A. Zakharov, G. T. Shchechkov, R. A. Vasyutkova. Izv. TPI, 151, 53, 1966.

16. Kh. A. Zakharov, G. T. Shchechkov. ZhFKh, 41, 1691, 1967.

17. Yu. A. Zakharov, E. B. Chernov. This collection, p. 135.

18. Yu. A. Zakharov, S. M. Ryabykh. Izv. TPI, 199, 71, 1969.

19. R. A. Vasyutkova, G. G. Savel’ev, Yu. A. Zakharov. This collection, p. 131.

20. Yu. A. Zakharov, G. G. Savel’ev. Kinetics and Catalysis, 5, No. 4, 1964.

21. F. C. Tompkins, P. W. Jacobs, T. Kureishy. Sci Technol. Aerosp. Rept., 2, (15), 1905, 1964.

22. F. C. Tompkins, T. Kureishy. J. Chem. Soc., 4792, 1964.

23. F. Solymosi, K. Dobo. Preprint of 5-th Symp. on React. Solids, Munich, 1964.

24. K. Torkar, O. Heiss, A. Isenberg. 5-th Symp. on React. of Solids, 1964.

25. M. Freed. J. Phys. Chem., 69, 450, 1965.

26. V. P. Kornienko, M. N. Dubrovskaya. Ukrainian Chemical Journal, 29, 262, 1963; 29, 270, 1963.

27. G. Rienacker. Z. Chemie, 4, 181, 1964.

28. W. Langebeck, G. Kraus. Z. anorg allg. Chemie, 327, 2, 1964.

29. W. Langebeck, H. Dreyer. Z. anorg. allg. Chemie, 329, 179, 1964.

30. Yu. A. Zakharov, S. M. Ryabykh. Kh. V. E., 4, No. 5, 1970.

31. Yu. A. Zakharov, S. M. Ryabykh, V. A. Nevostruev. Collection of Proceedings of the IV Coordination Conference on the Physics of Nonmetallic Crystals, Kiev, Naukova Dumka, 1969.

32. Yu. A. Zakharov, S. M. Ryabykh, G. G. Savel’ev. This collection, p. 107.

33. S. K. Deb, B. L. Evans, A. D. Ioffe. VIII Symp. on comb., N. Y. Baltimore, 829, 1962.

34. Chem. Abstr., 11, 337 d, 1965.

35. Fleer, Phys. Stat. Solidi, 5, 303, 1963.

36. J. Phot. Sci., 12, No. 5, 1964.

37. J. Cunningham. J. Chem. Phys., 41, 3522, 1965.

38. V. V. Boldyrev, V. M. Lykhin, A. N. Oblivantsev, K. M. Salikhov. Kinetics and Catalysis, 7, 432, 1966.

39. Kh. S. Bagdasar’yan. Advances in Chemistry, 8, 1617, 1939.

40. A. N. Terenin. Photochemistry of Dyes, Moscow, 1947.

41. Mees C. E. K. The Theory of Photographic Processes, London, 1954.

42. V. H. Carrol, W. West. Photogr. Sensitivity, 161—182, 1951. (Translation in the collection Physical Chemistry of Photographic Processes, Foreign Literature Publishing House, Moscow, 1954, p. 242).

43. V. V. Sviridov. Photochemistry and Radiation Chemistry of Inorganic Substances. “Vysshaya Shkola,” Minsk, 1964.

44. A. Terenin, I. Akimov. Z. phys. Chem., Leipzig, 217, 307, 1961.

45. W. West. J. Phys. Chem., 66, 2398, 1962.

46. A. Terenin, I. Akimov. J. Phys. Chem., 69, 730, 1965.

47. J. Bourdon. J. Phys. Chem., 69, 705, 1965.

48. R. Nelson. J. Phys. Chem., 69, 714, 1965.

49. H. Meier. J. Phys. Chem., 69, 719, 1965.

50. G. Scheibe. Z. Electrochem (see [42], p. 219), 56, 723, 1952.

1. H. Meier. Angew. Chem., 77, No. 15, 1965.
2. H. Meier. Z. Wiss. Photogr., 58, No. 9—12, 1965.
3. S. E. Sheppard. Chem. Rev., 4, 340, 1927.
4. R. W. Gurney, N. F. Mott. Proc. Roy. Soc., A164, 151, 1938.
5. J. W. Mitchell. J. Phot. Sci., 6, 57, 1958.
6. G. Scheibe, F. Dorr. Cited from H. Frieser, A. Graf, D. Eschrich. Z. Electrochem., 65, 870, 1961.
7. G. Scheibe. Z. angew. Chemie, 63, 174, 439, 1951.
8. Ya. Tauc. Photo- and thermoelectric phenomena in semiconductors, IL, Moscow, 1962, p. 141.
9. I. A. Akimov, A. N. Terenin. DAN SSSR, 165, 1332, 1965.
10. Frank, Teller. J. Chem. Phys., 6, 861, 1938.
11. H. Frieser, A. Graf, D. Eschrich. Z. Electrochem., 65, 870, 1961.
12. P. V. Meyklyar, B. I. Stepanov. ZhFKh, 21, 15, 1947.
13. F. Seitz. Rev. Mod. Phys., 23, 328, 1951.
14. W. F. Berg. “Phot. Scien. Symp. Zurich, 1961”. Focal Press, L., 1963, P 68.
    Cited from [47].
15. W. West, B. H. Carroll. J. Phys. Chem., 57, 797, 1953.
16. V. V. Sviridov. In [43], p. 175.
17. Kh. S. Bagdasaryan. ZhFKh, 19, 266, 1945.
18. Kh. S. Bagdasaryan. Journal of Scientific and Applied Photography and Cinematography, 3, (5), 381, 1958.
19. K. C. Nelson. J. Opt. Soc. Amer., 48, 1, 1958.
20. V. V. Boldyrev, I. V. Shmidt, V. I. Pismenko, M. S. Shvartsberg, I. L. Kotlyarevskii, V. N. Andrievskii, V. F. Komarov. Kinetics and Catalysis, 6, (4), 766, 1965.
21. Yu. N. Sukhushin, Yu. A. Zakharov. KhVE, 1970 (in press).
22. B. L. Evans, A. D. Ioffe. Proc. Roy. Soc., A250, 346, 1959.
23. S. Z. Roginskii, E. I. Shul’ts. Ukrainian Chemical Journal, 3, 177, 1937.
24. F. Solymosi, N. Krix. Mag. Kem. Folyoirat, 68, No. 9, 1962.
25. F. Solymosi, N. Krix. J. Catalysis, 1, 468, 1962.
26. F. Solymosi, L. Revesz. Kinetics and Catalysis, 4, 88, 1963.
27. F. Solymosi, N. Krix. Acta Chem. Acad. Sci. Hung., 34, 241, 1962.
28. F. Solymosi, N. Krix. Mag. Kem. Folyoirat, 68, 241, 1962.
29. F. Solymosi. Comb. and Flame, 9, 141, 1965.
30. Z. Bitcumshow, B. Newman. Proc. Roy. Soc., A227, 228, 1955; A227, 115, 1954.
31. A. Galwey, P. Jacobs. Trans. Farad. Soc., 56, 581, 1960.
32. P. Jacobs, A. Kureishy. J. Chem. Soc., 557, 1962.
33. B. M. Tsarev. Contact potential difference and its influence on the operation of electrovacuum devices. Gostekhizdat, Moscow, 1955.
34. A. I. Gubanov. Theory of the rectifying action of semiconductors. Gostekhizdat, Moscow, 1956.
35. Semiconductors in modern science and technology, 1, ed. A. F. Ioffe, Moscow—Leningrad, 1957.
36. A. F. Ioffe. Semiconductors in modern physics, Moscow—Leningrad, Publishing House of the USSR Academy of Sciences, 1954.
37. V. K. Subashiev. Semiconductor diodes and triodes in science and technology, vol. 7, Leningrad, 1957.
38. S. M. Ryvkin. Photoelectric phenomena in semiconductors. Fizmatgiz, Moscow, 1963, p. 318.
39. A. P. Maksimenko, V. N. Tverdokhlebov. Izvestiya Vuzov, Fizika, No. 1, 84, 1964.
40. F. Solymosi, L. Revesz. Nature, 69, No. 5, 1963.
41. Solymosi, L. Revesz. Z. anorg. all. Chem., 322, No. 1—2, 86—100, 1963.
42. F. Solumosi, L. Revesz. Nature, 192, 64, 1961.
43. I. Makushima, I. Johda, J. Soida. Actes d. deuxiane congres (Intern) Catalyse, 617, I, Paris, 1960.
44. F. Solymosi, N. Krix. Mag. Kem. Folyoirat, 68, 241, 1962.
45. C. Otto, H. Frey. J. Am. Chem. Soc., 45, 434, 1923.
46. C. Otto, H. Frey. J. Am. Chem. Soc., 46, 266, 1924.
47. Titani, Fukushima, Noribe. Bull. Chem. Soc. Japan., 25, 245, 1952.
48. A. Simchen, A. Glasher. Bull. Phem. Soc. Frans., 20, 127, 1953.
49. S. Patai, E. Hoffmann. J. Chem. Soc., 1797, 1955.
50. S. Patai. J. Am. Chem. Soc., 72, 5098, 1950.
51. Yu. A. Zakharov, G. G. Savel’ev, E. S. Kurochkin. Izvestiya TPI, Tomsk, 151, 52, 1965.
52. Yu. A. Zakharov, E. S. Kurochkin, G. G. Savel’ev, Yu. N. Rufov. Kinetics and Catalysis, 7, No. 3, 1966.

1. Yu. A. Zakharov, G. G. Savel’ev, E. B. Chernov. Izv. TPI, Tomsk, 151, 32, 1965.
2. A. Glasner, A. Makovky. J. Chem. Soc., 1606, 1954.
3. S. Z. Roginskii, E. I. Shul’ts. Ukr. Khim. Zhurnal, 3, 177, 1937.
4. R. Hill. Trans. Farad. Soc., 53, 1135, 1957.
5. R. Hill. Trans. Farad. Soc., 54, 695, 1958.
6. R. Hill, I. Welsh. Trans. Farad. Soc., 56, 1059, 1960.
7. A. Hermony, A. Salman. Bull. Res. Council Israel, Ser. A. Chem., 11A, 198, 1962.
8. A. Hermony, A. Salman. VII Symp. on Comb., Baltimore, 672, 1962.
9. A. V. Raevskii, G. B. Manelis. DAN SSSR, 151, No. 4, 1963.
10. M. Evans, R. Beyer, L. McCulley. 40, No. 9, 1964.
11. Kuratani Kenij. Rept. Aeronaut Res. Inst. Univ. Tokyo, 28, No. 4, 1962.
12. H. Osada, E. Sakamoto. J. Ind. Expl. Soc. Japan, 24, No. 5, 1963.
13. Chemist’s Handbook. Khimiya Publishing House, Moscow, 1963.
14. A. A. Berlin, S. I. Bass. Teoreticheskaya i eksperimental’naya khimiya, 1, 151, 1965.
15. H. de Boer, E. Verwey. Proc. Phys. Soc., 49, 59, 1937; N. F. Mott. Ibid., A62, 416, 1949.
16. Yu. A. Zakharov, V. A. Nevostruev. Izv. vuzov. Ser. khim. i khim. tekhnolog., 11, No. 2, 1968.
17. Yu. A. Zakharov, V. A. Nevostruev. 42, No. 2, 1968.
18. V. A. Nevostruev, Yu. A. Zakharov. Kinetika i kataliz, 8, 211, 1967.
19. Yu. A. Zakharov, V. A. Nevostruev. Proceedings of the Interuniversity Conference on Radiation Physics, Tomsk, 1967, TSU Publishing House, 1970.
20. Yu. A. Zakharov, V. A. Nevostruev. The present collection, p. 60.
21. I. Cunningham. J. Phys. Chem., 87, 1772, 1963.
22. I. Cunningham. J. Phys. Chem., 66, 779, 1962.
23. E. Klein, K. Matejec. Z. Electrochem., 63, 883, 1959; Naturwiss., 46, 225, 1959.
24. J. Haynes, W. Shockley. Phys. Rev., 82, 935, 1951.
25. G. G. Savel’ev, Yu. A. Zakharov, V. B. Spitsa. ZhFKh, 39, 2808, 1965.
26. F. F. Vol’kenshtein, V. B. Sandomirskii. DAN SSSR, 118, 960, 1958.
27. E. P. Mikheeva, N. P. Keier. Kinetika i kataliz, 5, 748, 1964.
28. P. M. Stadnik, Fenčik. Ukr. Khim. Zhurn., 27, 3, 1961. Kin. i kataliz, 2, issue 4, 1961.
29. P. M. Stadnik, E. Yu. Sekereš. Kin. i kataliz, 5, 430, 1964.
30. L. M. Dmitrienko, S. S. Lachinov, R. F. Sivyakova. Kin. i kataliz, 6, issue 1, 1965.
31. G. Bliznakov, D. Lazarov, E. Lazarova. Izv. In-ta obshch. i neorg. khim. Bulg. AN, 2, 1964.
32. A. F. Ioffe. Physics of Semiconductors. Publishing House of the Academy of Sciences of the USSR, Moscow–Leningrad, 1957.
33. A. S. Zinets, G. P. Peka, Yu. I. Karlanin. FTT, issue 12, 1964.
34. A. G. Cheban. FTT, 7, 1303, 1965.
35. Yu. A. Zakharov, A. A. Kabanov, A. M. Trubitsyn. Izv. vuzov, Khim. khimich. tekhnolog., 8, 529, 1965.
36. L. V. Keldysh. ZhETF, 39, 1138, 1958.
37. W. Frauz. Z. Naturforsch., 13a, 484, 1958.
38. E. Gutsche, H. Zauge. Phys. Stat., 4, K21, 1964; Solidy, 13(2), K131, 1966.
39. K. I. Britsin, A. A. Smirnov. FTT, 8(1), 163, 1965.
40. V. S. Vavilov. The Action of Radiation on Semiconductors. Fizmatgiz, Moscow, 1963, p. 33.
41. K. I. Britsin, B. A. Volkov, Matveev, A. A. Smirnov. FTT, 7(8), 2536, 1965.
42. A. G. Aronov, G. E. Pikus. FTT, 8, No. 1, 1966.
43. A. Luyckx, G. Zontil. Chem. Abstr., 63, 6445, 1965.
44. E. L. Frankevich, E. I. Balabanov. ZhETF. Pis’ma v red., 1, 33, 1965; FTT, 8(3), 855, 1966.
45. G. G. Savel’ev, V. P. Greben’. Report at the Conference of Young Chemical Scientists in Tomsk, 1966.
46. Yu. A. Zakharov, V. A. Nevostruev. U Kh., 37, 143, 1968.
47. V. I. Oshcherov. Teoret. i eksper. khim., 1, 436, 1965.