On a Method for Determining the Breakdown Voltage of Solid Dielectrics in a Nonuniform Field

A. A. VOROB'EV, G. A. VOROB'EV, V. D. KUCHIN

Experiments on the breakdown of solid dielectrics in a nonuniform field are often carried out on specimens in the form of plates placed between electrodes. This method has a substantial shortcoming in that breakdown of a solid dielectric is, as a rule, preceded by discharges along the surface of the specimen, which significantly affect the breakdown of the solid dielectric itself. As a result, the breakdown voltage of solid dielectrics punctured in a nonuniform field depends on the surrounding medium. In particular, the temperature dependence of the breakdown voltage of mica placed in glycerin had the same character as for glycerin, while for mica placed in distilled water the temperature dependence had the same character as for water [1]. Experiments on the breakdown of plates of solid dielectrics in a nonuniform field do not give true data or regularities characterizing the solid dielectric.

Fig. 1. Shape of a specimen for testing the breakdown of solid dielectrics in a nonuniform field.

Fig. 2. Dependence of the breakdown voltage of NaCl and KBr single crystals on the energy of their crystal lattice during breakdown in a uniform field. Breakdown probability \(\Psi = 90\%\). \(\triangle\)—recess \(\ominus\); \(\blacktriangle\)—recess \(\oplus\); \(\circ\)—plate \(\ominus\); \(\bullet\)—plate \(\oplus\).

To eliminate the influence of the surrounding medium, it is proposed to drill recesses in the specimen of the solid dielectric. In this case, at the place of high field intensity, the surrounding medium is the solid dielectric itself. We used the specimens shown in Fig. 1. Figure 2 presents the dependence of breakdown voltage on lattice energy.

for plates of NaCl and KBr under pulses with an exposure time of the order of \(10^{-6}\) sec. In Fig. 3 data are presented on the breakdown of specimens with drilled conical recesses of NaCl, KCl, KBr, KI single crystals. The experimental results show that specimens with drilled conical recesses give quantities characterizing the solid dielectric, which experiments with plates do not give. Using specimens with conical recesses, we obtained certain general regularities for NaCl, KCl, KBr, and KI single crystals, as well as for organic glass:

Fig. 3. Dependence of the breakdown voltage of NaCl, KCl, KBr, and KI single crystals on the energy of their crystal lattice for breakdown in a point–plane field. Breakdown probability \(W = 90\%\).

1. At a voltage exposure time of \(8 \cdot 10^{-8} \div 2 \cdot 10^{-7}\) sec, an increase in the breakdown voltage is observed in comparison with longer times, which is associated with the delay of the discharge.

2. At a voltage exposure time of \(8 \cdot 10^{-8} \div 2 \cdot 10^{-7}\) sec, the polarity effect is clearly expressed: with negative polarity of the point, the breakdown voltage is considerably higher than with positive polarity of the point. At longer times (of the order of \(10^{-6}\) sec and more) the polarity effect is weakly expressed.

Specimens with conical recesses have a higher breakdown voltage than plates. Walter and Inge [2], in the breakdown of rock-salt plates \(2.7\) mm thick, obtained the following breakdown-voltage values: \(30\) kV with positive polarity of the point and \(36\) kV with negative polarity of the point. In our experiments with specimens with conical recesses, with an average thickness at the breakdown site of \(0.7\) mm, i.e., almost 4 times smaller, at a voltage exposure time of the order of \(10^{-6}\) sec, when there is as yet no discharge delay, the following breakdown-voltage values were obtained: \(40\) kV with positive polarity of the point and \(43\) kV with negative polarity of the point. In experiments with plates of average thickness \(0.7\) mm, we also obtained low values of breakdown voltages: \(18\) kV with positive and \(21\) kV with negative polarity of the point. Evidently, the decrease in breakdown voltage in experiments with plates is associated with surface discharges.

The proposed method for testing solid dielectrics for breakdown in a nonuniform field makes it possible to measure the electric strength of solid dielectrics in large thicknesses. The usual electrode system (a spherical cup and a plane) proves unsuitable for electrical breakdown of dielectrics in a thick layer, since breakdown of dielectrics in large thicknesses in a uniform field requires specimens of extremely large dimensions. At high frequencies, large electrodes cannot be used because of their considerable capacitance, and they do not completely eliminate corona. To measure the electric strength of solid dielectrics in thick layers we used the specimen shown in Fig. 3 [3]. Voltage was supplied to electrode 3 through conductor 6, which covered the inner surface of glass tube 7 (8–11 mm in diameter and up to 400 mm high), inserted into the specimen. Conductor 10 was connected to the high-voltage source. Breakdown was produced in air. Using the specimen shown in Fig. 3, the maximum field strength at the spherical surface of the electrode (point) 4 of radius \(r\) was calculated from the formula

\[ E_{np}=\frac{U_{np}}{\varepsilon r}. \]

For NaCl with \(r = 0.09 \div 0.115\) mm and a layer thickness \(d = 10 \div 17\) mm, at constant voltage we obtained \(U_{np} = 62 \div 96\) kV, i.e., \(E_{np} = 1.01 \div 1.625\) MV/cm. By using the specimen described above and the supply lead, it was possible to avoid corona up to the breakdown voltages. The values of the electric strength of rock salt determined during breakdown in thick layers are in good agreement with the values of \(E_{np}\) determined during breakdown of NaCl in thin layers and in a uniform field.

REFERENCES

1. Vorob’ev A. and Prikhod’ko N. Proceedings of SFTI, 4, No. 3, 112–126, 1936.
2. Walter A. F. and Inge L. D. ZhTF, 3, No. 6, 840–849, 1933.
3. Vorob’ev A. A. and Kuchin V. D. Izv. Tomsk Polytechnic Institute (in the present issue), p. 383.

Tomsk Polytechnic Institute.