BREAKDOWN DELAY TIME IN SOLID ELECTRICAL INSULATING MATERIALS

M. N. Lisetskaya, T. A. Zhenchenko

(Presented by the scientific seminar of the High-Voltage Research Institute)

Data on the discharge time and the rate of its development can serve as material for judging the mechanism of breakdown. Modern theories of electrical breakdown in solid dielectrics do not consider the discharge process in space and time. The breakdown process is identified with the onset of impact ionization.

The discharge delay time in many solid dielectrics (nitrocellulose, presspahn, glass, mica, porcelain) was studied by Strigel [1]. From the experimental data, the dependence

\[ \frac{n_t}{n_0}=\frac{t}{t_{\text{del}}} \]

was plotted on a semilogarithmic scale, where \(n_t\) is the number of specimens having a given and shorter delay time, and \(n_0\) is the total number of punctured specimens. However, his results cannot be considered reliable, since no oscillography was used in his experiments.

A number of authors [2, 3] studied discharge delay in crystals of alkali-halide salts. Investigations by K. K. Sonchik [2] found that in alkali-halide salts several tenths of a millimeter thick, discharge delay times are of the order of \(10^{-7}\)–\(10^{-8}\) sec, and the average velocities of discharge propagation are \(10^5\)–\(10^6\) cm/sec.

Discharge delay in organic glass and epoxy resin 3–8 mm thick in a nonuniform field was studied by Dittmer [4]. He found that the discharge delay time lies in the interval from units to tens of nsec.

But, as is known, only the characteristics of the behavior of a dielectric in a uniform field determine its electrical strength. Therefore, the behavior of insulation in a uniform field is of definite interest.

However, the study of discharge development in a dielectric under uniform-field conditions is considerably more complicated than in a nonuniform field for two reasons:

1) because of the stronger electric field, the processes proceed more rapidly;

2) the interelectrode distance is smaller.

Both require better time resolution from the apparatus used.

In a uniform field in organic glass of thickness 35 µm, M. A. Melnikov [5] recorded the volt-second characteristic at breakdown on the pulse front from \(5\cdot 10^{-7}\) sec to \(5\cdot 10^{-9}\) sec. He found strengthening of the organic glass for a voltage action time of \(5\cdot 10^{-9}\) sec, but the delay time in the organic glass was very small and could not be measured.

Therefore the task is posed of measuring discharge delays in solid insulating materials in the range of shorter times. This is of interest not only from the standpoint of studying the breakdown mechanism, but also in connection with the development and application of nanosecond pulse generators and quantum generators that are triggered by a high-voltage pulse of nanosecond duration. It is necessary to study the behavior of practical insulation in this time range.

For the investigation, rectangular voltage pulses with a front of \(3\cdot 10^{-9}\) sec and a flat portion of duration 100 nsec were used. The generator provided an adjustable voltage amplitude up to 60 kV. A more detailed description of the setup is given in [6].

The specimens for investigating the breakdown delay time were made from fluoroplastic, vinyl plastic, and getinax. They were cut from a single sheet of the corresponding material and, for these experiments, had dimensions \(35\times 35\times 10\) mm. On one side of the specimen, using a special attachment on a drilling machine, a hemispherical recess was made so that the thickness of the specimen at the thinnest point after drilling was \(0.2\div 0.3\) mm. The diameter of the hemispherical recess was selected so that the condition of field uniformity according to the standard was satisfied: \(R_{0}>25\,d\) and \(D>5d\), where \(R_{0}\) is the radius of the recess, \(D\) is the thickness of the material plate, and \(d\) is the thickness of the specimen at the thinnest point. In our experiments, with \(R_{0}=5\) mm, \(d=0.1\) mm, \(D=10\) mm, this condition was satisfied. The required specimen thickness of 0.1 mm was achieved by grinding the flat side of the specimen on emery paper and then on felt with paste. The specimen thickness was measured on an IZV-1 type instrument with a smallest scale division of 0.001 mm.

Powdered graphite was used as the electrode material; on one side of the specimen it was pressed into the recess, and on the other side it was poured onto a flat copper electrode 10 mm in diameter and pressed against the flat part of the specimen.

The specimens were broken down in air. The specimen dimensions were chosen sufficiently large, and no surface discharges were observed.

First of all, the dependence of electric strength on breakdown probability was obtained for all the materials studied. For this purpose rectangular voltage pulses of duration 100 nsec were applied to the specimen, with the amplitude of each subsequent pulse increased by 10% relative to the preceding one, until the specimen broke down. The results obtained in these tests are presented in Fig. 1.

The discharge delay time was then determined at specified values of field strength. Figure 2 presents the statistical processing of the discharge delay time by the method of [7], carried out for 30 vinyl-plastic specimens at a field strength equal to \(1.2\) MV/cm. The delay time was also found for other materials and for other field strengths. From these results the dependence of the delay time on field strength (Fig. 3) was constructed for all the materials studied. In constructing it, for the point at the lowest field strength for a given material, delay times equal to and less than those obtained at the next field strength were not taken into account. These times

delays were obtained, apparently, upon breakdown of any specimens with defects.

In our investigations it was found that the breakdown delay time in electrical insulating materials is small and lies in the interval from 20 to 2 nsec. It decreases with increasing field strength. The small

Fig. 1. Dependence of electric strength on the probability of breakdown for electrical insulating materials.
1—getinax, 2—viniplast, 3—fluoroplastic.

Fig. 2. Statistical processing of data on breakdown delay time for viniplast at \(E = 1.2\) MV/cm.

magnitude of the discharge delay time permits the conclusion that the principal share in the total discharge delay time is played by the formation time. With such a large electrode diameter and with graphite as the electrode material, it is evident that the probability of the appearance of free electrons, which determines the delay time, is large.

Therefore it may be considered that the statistical delay time can be neglected, and that the measured delay time is the discharge formation time. It is possible to determine the mean discharge velocity as

\[ V=\frac{d}{t_{\text{del}}}, \]

where \(d\) is the thickness of the dielectric, \(t_{\text{del}}\) is the breakdown delay time. At \(E = 1.5\) MV/cm in viniplast, \(t_{\text{del}} = 3 \cdot 10^{-9}\) sec, \(V = 3 \cdot 10^{6}\) cm/sec.

Fig. 3. Dependence of delay time on field strength for 1—getinax, 2—viniplast, 3—fluoroplastic.

As the field strength increases, the delay time decreases and, consequently, the discharge velocity increases. The high discharge velocities indicate that electrical breakdown in solid dielectrics is an electronic process.

Comparison with the development of a discharge in rock salt [8] shows that the discharge in solid electrical insulating materials proceeds at a higher velocity, which can be explained by the larger value of the field strength and by the fact that rock-salt crystals can be very well examined under a microscope and, by selection of specimens, the presence of...

defects (foreign inclusions, air bubbles, cracks). In opaque dielectrics this cannot be done.

However, we [9] carried out investigations in transparent organic glass in a uniform field at a thickness of 70 μm with graphite electrodes. It was found that organic glass has greater strength in comparison with the materials studied, which can be explained by the smaller electrode diameter, as well as by the greater homogeneity of the dielectric.

For organic glass, the measured delay times lie in the same time interval as for the given dielectrics, which indicates an analogy between the breakdown processes in these materials. Thus, on the basis of the results of the investigations the following conclusions may be drawn:

1. The discharge delay time in electrical insulating materials (vinyl plastic, fluoroplastic-4, organic glass, getinax) is small and is of the order of \(2(10^{-9} \div 10^{-8})\) sec.

2. The discharge delay time decreases with increasing field strength.

3. The average discharge velocity is large, \(3 \cdot 10^{6}\) cm/sec.

4. The high rate of the process is explained by the electronic character of breakdown.

REFERENCES

1. R. Strigel. Elektrische Stoßfestigkeit, Berlin, 1939.
2. K. K. Sonchik. Dissertation, Tomsk, 1958.
3. Y. Inuishi, T. Suita. J. Inst. El. Eng. Japan, 74, 2, 150, 1954.
4. B. Dittmer. Arch. f. Electrotechn., B 48, H-3-5, 1963.
5. M. A. Melnikov. Dissertation, Tomsk, 1959.
6. G. A. Vorob’ev, M. N. Lisetskaya. PTE, No. 3, 1964.
7. M. Laue. Ann. d. Phys., 76, 261, 1925.
8. G. A. Vorob’ev, M. N. Lisetskaya. FTT, No. 12, 1964.
9. M. N. Lisetskaya. “Electric discharge in organic glass in a uniform field,” in press in Izvestiya vysshikh uchebnykh zavedenii, series “Physics,” No. 5, 1965.