MEETINGS AND CONFERENCES

CONFERENCE ON THE ELECTRICAL AND PHYSICOCHEMICAL PROPERTIES OF SOLID DIELECTRICS

I

In September 1955, a conference on the electrical and physicochemical properties of solid dielectrics was held in Tomsk. The conference was devoted to questions of dielectric breakdown, dielectric losses and polarization, and the electrical conductivity of solid crystalline dielectrics.

Taking part in the conference, in addition to scientists from Tomsk, were scientists from Moscow (Physics Institute of the Academy of Sciences of the USSR), Leningrad (Research Institute of the Polytechnic Institute, Leningrad Polytechnic Institute, Leningrad Electrotechnical Institute), Novosibirsk (West Siberian Branch of the Academy of Sciences of the USSR), Irkutsk (State University), Kemerovo (Mining Institute), and others.

About 40 reports were considered at the conference.

The conference was opened with an introductory address by the chairman of the organizing committee, the director of the Tomsk Polytechnic Institute, A. A. Vorob’ev, who noted the increased interest in recent years in semiconductors and dielectrics. Along with the fact that our physical concepts of dielectric polarization are opening reliable paths for designing new dielectrics with prescribed properties, we still know little about the behavior of dielectrics in a strong electric field. Ideas about the electrical breakdown of solid dielectrics are schematic and insufficiently confirmed by experiments. The properties of insulating materials are often considered not in their interrelation, but in isolation. This hinders the creation of new insulating materials with high electrical properties. Therefore there is an urgent need to discuss the most important problems of the physics and chemistry of solid dielectrics from various points of view.

The first report, on the topic “On the Question of Thermal and Electrical Breakdown of Solid Dielectrics,” was delivered by G. I. Skanavi. In this report criticism was directed at the established criterion for distinguishing between thermal and electrical breakdown of solid dielectrics, which consists in the assertion that there is a certain critical temperature above which the breakdown voltage decreases with increasing temperature (thermal breakdown) and below which it does not depend on temperature (electrical breakdown). The author refers to a number of works carried out recently, in which a strong dependence of the breakdown voltage of simple single crystals on temperature was found in the temperature region

50–100°, when the breakdown is not thermal; in this case a maximum of the breakdown voltage is observed.

Experimental data obtained in the laboratory of G. I. Skanavi show that, at constant voltage in a homogeneous field, for KBr crystals there is a sharply pronounced maximum of the breakdown voltage on pulses at a temperature of 50°C. When the pulse duration is decreased this maximum is smoothed out, and at a pulse duration of \(10^{-6}\) sec practically no decrease of the breakdown voltage with temperature is observed. This regularity does not agree with the so-called “high-temperature” theory of Fröhlich and confirms qualitative considerations on the influence of volume charge on the breakdown process when the voltage is applied for a sufficiently long time.

The report by V. A. Chuenkov (Phys. Inst., Academy of Sciences of the USSR), “On the theory of electrical breakdown by ionization of crystals,” was devoted to discussion of a criterion for electrical breakdown of solid dielectrics on the basis of determining the distribution function of conduction electrons in a strong electric field, taking into account scattering of electrons by vibrations of the crystal lattice and by ionization collisions.

The effective ionization cross section is expressed by the author in the form \(Q(E)=S_0(E-J)\), where \(S_0\) is a constant, \(E\) is the energy, and \(J\) is the ionization potential. It is assumed that the effective ionization cross section depends neither on the direction of motion of the ionizing electron after the act of ionization nor on the direction of motion of the knocked-out electron, and that both these electrons after ionization have the same energy. The distribution function for \(E>J\) decreases very sharply with increasing \(E\) (being practically equal to zero at \(E=2J\)). The number of recombining electrons depends on what fraction of the electrons, after ionization, fall into the energy region \(0<E<J\), slow down there, and what their velocity is.

The author introduces two characteristic energy values: \(E_p\), at which on the average an electron gives to the lattice per unit time as much energy as it receives from the field, and \(E'_{1/2}\), determined by the condition that the numbers of electrons which, as a result of ionization, acquire energy greater and less than \(E'_{1/2}\) be equal. At \(E_p=E'_{1/2}\) the field accelerates fewer electrons than \(N\), the electrons (\(N\) is the number of ionization events per unit time), and in this case no disturbance of the stationary state can occur. At \(E_p<E'_{1/2}\) the field accelerates more than \(N\) electrons; consequently, the ionization process in this case has a chain character and the stationary state is disturbed.

Thus, according to V. A. Chuenkov, the breakdown criterion is the equality \(E_p=E'_{1/2}\). From the distribution function, \(E_p\) and \(E'_{1/2}\) are determined as functions of the field strength, and from the equality of these quantities—the breakdown voltage. For alkali-halide crystals this criterion gives the order of magnitude of the breakdown voltage and the character of its dependence on temperature, in agreement with Hippel’s experimental data on pulses of \(10^{-6}\) sec.

In the review report by Yu. A. Starikina (West-Siberian Branch of the Academy of Sciences of the USSR), “Modern theory of electrical breakdown of solid dielectrics,” the main attention was directed to Franz’s theory, in which the conductivity-electron avalanche mechanisms developed by Hippel and Fröhlich are adopted, while the breakdown criterion is taken to be the destruction of the crystalline structure of the dielectric under the action of the electron current.

Yu. A. Starikina reduces Franz’s theory to the following basic propositions:

1. The electric field tilts the energy bands of the dielectric, which creates conditions for the seepage of electrons from the filled band into the conduction band by means of the tunnel effect.

1. As a result of the acceleration of electrons in the conduction band by the field and of their braking by vibrations of the crystal lattice, a certain distribution of electrons with respect to energies is established. In this case conduction electrons possessing sufficient energy cause impact ionization.

2. Under the action of an external electric field, a flow of electrons through a certain energy boundary is established in the conduction band.

3. An expression is given for the increase of conduction electrons with time, proceeding from the kinetics of electron transitions between bands.

4. The thermal state of a dielectric caused by the appearance of an electric current is determined. Breakdown is here considered as destruction of the crystal lattice (melting) at the expense of the energy released by the current.

For various substances the theoretical values of the breakdown voltages agree well with the experimental ones. In the opinion of Yu. A. Starikina, Franz’s theory describes with sufficient completeness the physical processes in a dielectric both in the pre-breakdown state and at the moment of breakdown.

The report by A. A. Vorob’ev (Tomsk Polytechnic Institute), “An Attempt at a Theory of the Properties of Ionic Dielectrics on the Basis of Thermochemical Characteristics,” is a generalization of numerous results of studies and comparisons of the properties of dielectrics carried out by a large group of scientific workers in Tomsk. The main idea of this generalization reduced to the fact that between the electrophysical and physicochemical properties of ionic crystals, on the one hand, and the energy of the crystal lattice, on the other hand, there exists a definite regular relation.

Thus, for example, electrical conductivity, dielectric losses, and dielectric strength—the electrical strength—are caused by the displacement, liberation, and motion of charges in the dielectric under conditions of interaction with its crystal lattice. These processes are connected with the magnitude of the lattice energy.

A definite relation has also been established between the mechanical, thermal, and optical properties, on the one hand, and the lattice energy of dielectrics, on the other hand. In this connection, an increase in the lattice energy corresponds to an improvement in all the enumerated properties of the dielectric, which are of important practical significance.

In the opinion of A. A. Vorob’ev, knowledge of the lattice energy makes it possible to create electrical insulating materials with high electrical strength and with other high electrical and physicochemical properties.

The results of investigations of certain solid solutions of alkali-halide salts, obtained and reported by A. M. Trubitsyn (Tomsk Polytechnic Institute), confirm the propositions advanced by A. A. Vorob’ev. On comparing the dependence of electrical strength on the composition of a solid solution with the dependence of surface energy on the composition of the same solutions (according to the data of P. A. Savintsev), their analogous course was found. Since the surface energy increases with increasing energy of the crystal lattice, it may be assumed that the electrical strength likewise increases with increasing lattice energy. In the work of A. M. Trubitsyn it was also shown that in solid solutions of alkali-halide salts the electrical strength increases with increasing stability of chemical compounds (with increasing heat of formation and melting temperature).

The results of analogous investigations, from which it is also seen that the electrical strength of mixed alkali-halide crystals is connected with the lattice constant, the refractive index, and the surface

energies, were obtained and reported by A. N. Kislina (Tomsk Polytechnic Institute). These data also show that the electrical strength increases with an increase in the energy of the crystal lattice.

In the study of certain systems of solid solutions of alkali-halide salts, A. M. Trubitsyn established the presence of a minimum of electrical strength at average percentages of the components.

The same fact was established and reported at the conference by A. F. Gorodetskii (Novosibirsk Electrotechnical Institute) and M. S. Ivankina (Tomsk Polytechnic Institute). These results contradict Hippel’s data.

In connection with this, A. N. Kislina carried out similar investigations on certain solid solutions of different storage durations. It turned out that in freshly prepared systems a minimum of electrical strength is observed, while as they are stored the electrical strength of solid solutions of average composition increases. Apparently, the contradictions in the data of A. M. Trubitsyn and Hippel can be explained by the different storage durations of the systems studied.

In the report of A. A. Vorob’ev and A. F. Kalganov, the electrical strength of certain gases and liquids was compared with their physicochemical properties, and it was shown that in these cases it may be assumed that the electrical strength is determined by the forces of interaction between molecules.

In another report by A. A. Vorob’ev and A. F. Kalganov, an attempt was made to measure the breakdown energy of solid dielectrics by a calorimetric method.

Of special interest was the report of G. A. Vorob’ev (Tomsk Polytechnic Institute), devoted to the elucidation of the dependence of the electrical strength of alkali-halide salts on the time of voltage action, as well as to volt-time dependences during breakdown in nonuniform fields within times from \(10^{-8}\) sec to constant voltage.

The results of the investigations showed a significant increase in electrical strength for voltage-action times shorter than \(10^{-7}\) sec. These results should make it possible to calculate the time and energy of discharge formation.

Volt-time investigations in a nonuniform field showed that, for a positive polarity of the point, the breakdown voltage is lower than for a negative polarity. The results of this work indicate the possibility of testing solid dielectrics in the field: conical point against plane.

A report by B. A. Komarov (FIAN) was devoted to the question of static delay in the breakdown of solid dielectrics. In this work it was established that there is no dependence of the breakdown strength on the time of application of an impulse voltage of duration from \(6 \cdot 10^{-6}\) to \(5 \cdot 10^{-8}\) sec for mica specimens up to 5 microns thick. When the thickness is decreased from 10 to 2 microns, the electrical strength increases somewhat, but this may be due to the inhomogeneity of the field and to the increased uniformity of the material as the thickness decreases. The report points out that the published experimental data of other authors are hardly sufficient to substantiate the correctness of Seitz’s views (1947), which indicate the possibility of lowering the critical field strength by a deviation of electron behavior from the mean electron.

In addition to the reports listed, the conference also heard reports by V. S. Dmitrevskii (Tomsk Polytechnic Institute) on the distribution of stresses along the surface of a dielectric, by A. N. Gubkina (Physics Institute of the Academy of Sciences of the USSR) on surface breakdown of solid dielectrics with high dielectric permittivity, and by I. I. Kap-

MEETINGS AND CONFERENCES

…radov (West Siberian Branch of the Academy of Sciences of the USSR), who examined the mechanism of gas breakdown in the presence of thermoelectric emission.

A large group of papers was devoted to dielectric losses and polarization in solid dielectrics.

In the paper by G. I. Skanavi, questions of relaxation dielectric polarization and of the internal field in solid dielectrics were discussed. The investigations carried out by the author on polycrystalline dipole organic dielectrics, even with not very high dielectric permittivity, showed that simple regularities of relaxation polarization are found in them. The paper also showed that, in the case of clearly expressed relaxation polarization, it is possible, on the basis of experimental data, to estimate the coefficients of the internal field in the dielectric.

In the laboratory of the Physical Institute of the Academy of Sciences of the USSR, solid dielectrics containing weakly bound ions were synthesized and studied. Such a structure favors polarization. The value of the dielectric permittivity of such dielectrics reaches 1000; at the same time, they lack the properties characteristic of ferroelectrics. Thus the existence of non-ferroelectric dielectrics with high dielectric permittivity has been established.

Investigations of such dielectrics showed that relaxation polarization may be caused, in some cases, by weakly bound ions, and in others by structural heterogeneity. Proceeding from this, new dielectrics representing considerable interest were obtained. The paper theoretically considered relaxation polarization when the acting field differs from the mean field.

In the paper by N. P. Bogoroditskii (Leningrad Electrotechnical Institute), the nature of the temperature dependence of dielectric losses during the polarization of ionic compounds was considered. Generalizing the experimental results obtained, the speakers assert that, over a wide range of temperatures and frequencies, dielectric losses in the polarization of ionic compounds are caused by one and the same phenomenon—the disturbance of the thermal motion of ions under the action of an electric field, which manifests itself the more sharply the closer the period of the electric field coincides with the relaxation time of the particles. In connection with this, N. P. Bogoroditskii considers the now generally accepted division of dielectric losses in ionic compounds into the three components established by G. I. Skanavi in 1952, namely: 1) structural dielectric losses, which do not depend on temperature; 2) relaxation losses; and 3) conductivity losses, to be inappropriate. Instead of such a division, it is proposed to reduce dielectric losses to two physical processes:

1) The phenomenon of relaxation during polarization connected with the thermal motion of particles and manifested at all frequencies in the solid, liquid, and even gaseous states of substances. This type of loss also includes the losses of through electrical conductivity, increasing with temperature according to the exponential law.

2) The phenomenon of ionization of a substance, usually a gas (free or distributed in a solid body), manifested in electric fields of increased strength.

Of particular importance is the behavior of dielectric crystals in high-frequency electric fields. Dielectric losses in crystals at high frequency were the subject of a paper by K. A. Vodop’yanov (Siberian Physico-Technical Institute). Investigations carried out at the Siberian Physico-Technical Institute proved that, in crystals with ionic conductivity, along with dielectric losses of conductivity there are also relaxation losses, caused both by the displacement of weakly bound ions and by the orientation of polar…

molecules. In this lies the difference between dielectric losses in crystals containing and not containing dipole molecules.

As a result of these investigations, a definite regular connection has been established between dielectric losses in alkali-halide crystals and their physicochemical properties. Thus, dielectric losses decrease with increasing activation energy, lattice energy, heat of formation, heat of sublimation, and melting temperature. Similar data have also been obtained for mixed crystals. The results of the investigations give grounds for assuming that dielectric losses in alkali-halide crystals are connected with the properties and structure of the principal substances, while a small amount of impurities in a real crystal plays a secondary role. In crystals with polar molecules of crystallization water (gypsum, talc, and copper sulfate), temperature and frequency maxima of dielectric losses are observed, which disappear after the removal of crystallization water by heating. This gives grounds for believing that, in the given case, dielectric losses are caused by the orientation of polar molecules. In addition, it is shown that dielectric losses in polar crystals are connected with the activation energy of dipole molecules. The report expresses the opinion that the connection of dielectric losses with the physicochemical properties of crystals should be taken as the basis for obtaining new dielectrics and for studying the mechanism of dielectric losses in solid dielectrics.

The report of A. P. Izergin (Siberian Physico-Technical Institute) gives the results of a study of the dependence of dielectric losses in mica on heat treatment, and the report of M. P. Tonkonogov (Karaganda, Mining Institute) sets forth the results of an investigation of dielectric relaxation in talc ceramics fired at various temperatures.

The data presented in these reports confirm the propositions expressed in the report of K. A. Volodina.

In addition, in the report of A. P. Izergin it is shown that complete removal of crystallization water from mica reduces the value of dielectric losses and dielectric permittivity, which gives grounds for recommending such heat treatment of mica powders for the manufacture of mica flex.

In the report of M. P. Tonkonogov it is indicated that, taking into account the conditions of formation of a crystal from polar molecules, defects of the crystal lattice, and the presence in some cases of hydrogen bonds, it seems possible to attempt to examine quantitatively the mechanism of formation of weakly bound dipoles in crystals.

The report of R. L. Müller (Kemerovo, Mining Institute) was devoted to the polarization of an atomic-ionic dielectric in an external field. The features of the polarization of ionic-atomic dielectrics reduce to the appearance of activated ionic polarization as a result of the transition of bound cations from one potential well to another, and also to the orientation of polar complexes arising as a result of fluctuations in the distribution of valence electrons at atomic structural groups. In accordance with this, the increased dielectric losses at low temperatures in substances of the borosilicate type can be explained by activated ionic polarization, while the anomalously high polarizability and dielectric losses in bodies with the structure of perovskites and rutile are explained by the orientation of fluctuating polar ionic-atomic complexes. In the light of these ideas, the critical region of anomalously high dielectric losses is due to the disappearance of rigid valence bonds at high temperatures and to the hindered fluctuation of oriented polar complexes at low temperatures.

In the report by F. I. Vergunas and O. V. Agashkina, results are presented from a study of dielectric losses in crystalline phosphors when they are irradiated with ultraviolet light. The results of investigations of the frequency, temperature, and time dependences of the tangent of the dielectric-loss angle and of dielectric polarization in crystalline phosphors supplemented the information on their localization levels and on the behavior of electrons in the phosphor after ionization of luminescence centers and after thermal liberation of electrons from localization levels.

Of great interest were the results of a study of dielectric losses in certain refractory glasses, which were considered in the report by V. I. Odelevskii and N. M. Verebeichik (Leningrad). These investigations established that the dielectric losses of alkaline aluminosilicate glasses have a dipole character. In this case the mechanism of dipole relaxation may be represented as the rotation of a dipole with an arm equal to the distance between a Na ion and the nearest aluminum ion. Quenching increases the dielectric losses and electrical conductivity of such glasses. Small changes in the composition of glasses strongly affect the quenching effect: in some cases quenching has almost no effect on the electrical properties, while in others it causes relaxation maxima of \( \operatorname{tg}\delta \). The authors suppose that quenching causes local loosening of the glass structure. Investigations of the temperature dependence of \( \operatorname{tg}\delta \) of alkali window glass at various frequencies showed the presence of relaxation maxima of \( \operatorname{tg}\delta \). On the basis of diagrams of heterogeneous equilibria, alkali-free boron glasses based on nepheline syenite raw material were developed with high dielectric properties, which are of interest for the electrovacuum industry.

The results of the investigation by B. I. Vorozhtsov (Siberian Physico-Technical Institute), presented in his report “Electrical Properties of Fused Quartz at High Frequencies and High Temperatures,” received high praise. To obtain reliable and accurate data under such complex conditions as high frequency and temperatures up to \(1000^\circ\) C, B. I. Vorozhtsov improved the calorimetric method of measurement. He established that at high temperatures the dielectric losses of fused quartz are conductivity losses. In strong high-frequency electric fields and at elevated temperature, dielectric losses increase manyfold as a result of ionization of gas inclusions. When impurities are removed by electrical cleaning or by hardening them by annealing, the dielectric losses of fused quartz decrease. As a result of the investigation, ways of improving the properties of quartz insulators were outlined. The work is of great practical significance.

Along with fused quartz, the production of synthesized electrical insulation with high electrical properties for operation at temperatures of \(500\)—\(1000^\circ\) C is of practical interest. This question was the subject of the report by V. I. Odelevskii and R. N. Streltsina, entitled “Synthesis of Silicoberyllates of Alkaline-Earth Metals and Barium Aluminosilicates and Investigation of Their Electrical Properties at High Temperatures.” The electrical conductivity of silicoberyllates of alkaline-earth metals at high temperatures is two orders of magnitude lower than the electrical conductivity of fused quartz, which is explained by the fact that the strongly bound ions of alkaline-earth metals compact the structure and hinder the movement of impurity alkali ions. The results of the investigation made it possible to obtain barium aluminosilicate both as a monoclinoic and as a hexagonal modification with an addition of even technically pure oxides, possessing insignificant electrical conductivity and very small dielectric losses at high temperatures and radio frequencies.

Along with dielectric losses and dielectric polarization in crystals, the conference considered the report by P. F. Veselovsky (Leningrad Polytechnic Institute), devoted to dielectric losses in amorphous polymers. The results of the studies confirm the exponential character of the dependence of the relaxation time \(\tau\) on the temperature \(T\) for high-frequency dielectric losses and show that the activation energy \(\Delta u\) does not depend on temperature and can

be calculated as the tangent of the angle of inclination \(\Delta \lg \tau/\Delta \dfrac{1}{RT}\). Extrapolating

the dependence of \(\lg \tau\) on \(\dfrac{1}{T}\) to \(\dfrac{1}{T}=0\), one can determine the natural frequency of oscillations of the polar radical. As for low-frequency dielectric losses, the experimental data indicate that,

by extrapolating the dependence of \(\lg \tau\) on \(\dfrac{1}{T}\) to \(\dfrac{1}{T}=0\), one can determine the natural frequency of oscillations of the relaxing element. Extrapolating

the dependence of \(\lg \tau\) on \(\dfrac{1}{T}\) to \(\lg \tau=0\), one can determine the softening temperature of the polymer. Experience has shown that plasticizers reduce dipole-radical losses and do not affect the position of the temperature maximum of losses. The results presented in the report of the studies of dielectric losses of polar polymers over a wide range of temperatures and frequencies contribute to elucidating the mechanism of dielectric losses in them and to establishing the connection between the dielectric properties and the chemical structure of polymers.

Questions of the electrical conductivity of solid dielectrics were dealt with in the reports by R. L. Müller (Kemerovo) and M. S. Meshchik (Irkutsk State University).

In the report by R. L. Müller, the dependence of the electrical conductivity of borosilicates on the concentration of metallic ions was considered. It was established by him that, with the exception of barium silicates, in borosilicates the concentration dependence of polar electrical conductivity is connected with the concentration dependence of the dissociation energy. An increase in the concentration of alkaline ions causes a sharp increase in electrical conductivity, while at small concentrations the electrical conductivity changes only slightly. A qualitative change in the nature of the electrical conductivity is characterized by the transition through the critical concentration region. Conductivity due to dissociated cations in a nonpolar medium changes into conductivity caused by the movement of vacant sites in the medium of associated polar structural elements.

In the report by M. S. Meshchik, results were communicated of a study of the surface electrical conductivity of fresh cleavages of mica crystals. The author found that the electrical conductivity of mica crystals cleaved in moist air decreases irreversibly with time. Ohm’s law for irreversible electrical conductivity is fulfilled only up to a field strength of \(50\ \text{kV/cm}\). At higher fields the logarithm of the electrical conductivity grows proportionally to the field strength. It may be supposed that, at the moment of cleavage, a monomolecular layer of water is adsorbed on the surface of mica crystals, disappearing with time.

The report by T. N. Verbitskaya (Leningrad) aroused great interest; it communicated the results of a study of barium-strontium titanate ceramics, consisting in the fact that an increase in the intensity of an alternating electric field causes an increase in the polarization of the ferroelectric up to saturation, and the dielectric permittivity also in-

increases, reaches a maximum, and then decreases. The behavior of ferroelectrics in an alternating electric field is analogous to the behavior of ferromagnets in an alternating magnetic field. As a result of extensive investigations by T. N. Verbitskaya, a material called “varicond VK-1,” with sharply expressed nonlinear properties, was obtained; it is of considerable interest for a whole range of technical applications.

A group of papers was devoted to the physicochemical nature of the bonding of glass and ceramics with metals and to the study, in connection with this, of the electrical properties of vacuum-tight ceramics. This problem is of great importance for the electrovacuum industry.

In the papers by V. A. Presnov and M. P. Yakubenya (Siberian Physico-Technical Institute), the formation of a transition layer during the fusion of glass and ceramics with metal was established experimentally.

The physicochemical nature of the bonding of glass and ceramics with metal is reduced to an acid–base interaction of oxides.

In these papers, the diffusion mechanism of the fusion of ceramics with metal is developed.

In the paper by V. M. Belousova (Tomsk Polytechnic Institute) and L. G. Lavrent’eva (Siberian Physico-Technical Institute), the process of formation of ceramic bodies and their properties is considered. In the course of the investigation, data obtained both on experimental and on production specimens were compared. The work of these authors made it possible to outline ways of improving the electrical properties of vacuum-tight ceramics.

Separate groups of papers heard at the conference were subjected to comprehensive and lively discussion*.

At the end of the conference, concrete resolutions were adopted and, for the purpose of better coordination of further work, a plan was discussed for work scheduled to be carried out at various institutes in the near future.

S. S. Gutin

* The discussion materials will be published in the Proceedings of the conference.