ON THE BREAKDOWN OF GLASS WITH A BARRIER UNDER IMPULSE VOLTAGE

D. D. Rumyantsev, N. M. Torbin

(Presented by the scientific seminar of the Research Institute of High Voltages)

It is known that in air and liquid dielectrics in a sharply nonuniform field a barrier effect is observed, i.e., the phenomenon of a significant increase in the electric strength of the gap when a thin layer of solid dielectric, customarily called a barrier, is placed between the electrodes [1, 2].

It has been established [3, 4, 5] that an analogous phenomenon is also observed in solid dielectrics if, in a dielectric of lower electric strength, a thin layer of another solid dielectric with higher electric strength is placed; in what follows we shall also call this layer a barrier.

The present work is devoted to the investigation of the influence of barriers made of films of high-polymer materials on the breakdown voltage of technical glass in a sharply nonuniform field of “point—plane” electrodes under impulse voltage.

The front duration of the voltage impulses was varied from \(10^{-7}\) sec to \(10^{-5}\) sec. Breakdown of the specimens occurred on the impulse front. The specimens consisted of glass plates of various thicknesses, which made it possible to “move” the barrier in the specimen by placing it between plates at different distances from the point. In most experiments the specimens consisted of three plates, whose total thickness was \(5.2\) mm (\(1.3\) mm; \(1.4\) mm; \(2.5\) mm). The barriers were films of polyethylene (PE), polytetrafluoroethylene (PTFE), polyethylene terephthalate (PETF), and polystyrene (PS), with thicknesses from 20 to 200 µm.

During breakdown the specimens were placed in transformer oil. The electrode device, a sketch of which is shown in Fig. 1, allowed the glass plates to be pressed against one another with a constant force, thereby ensuring an approximately identical thickness of the oil interlayer between the plates in different experiments.

The results of the experiments performed are given below.

Figure 2 shows the dependence of the breakdown voltage of glass on the position of a barrier made of PE-50 film (here and below, the numbers after the letters denote the film thickness in microns) for positive 1 and negative 2 polarity of the impulses. The dashed lines show the value \(U_{\mathrm{br}}\) for specimens without a barrier. \(S\) is the distance from the point to the barrier.

It is seen from Fig. 2 that the specimens in which the barrier is located at a distance of \(1.3\) mm from the point, i.e., between ...

1 and 2 plates. The relative increase in \(U_{\text{br}}\) due to the barrier is higher for voltage pulses of positive polarity than for negative polarity, and amounts to 56% and 39%, respectively.

Fig. 1. Electrode device: 1, 2—electrodes; 3—sample; 4—barrier

Fig. 2. Dependence of the breakdown voltage of glass on the position of the barrier in the sample: 1—pulses of positive polarity, 2—pulses of negative polarity, \(1'\) and \(2'\)—breakdown voltage of samples without a barrier. Pulse-front duration \(10^{-5}\) sec. Barrier—PE-50.

A similar increase in the breakdown voltage of glass is observed when films of PTFE, PETF, and PS are used as the barrier.

with a thickness of 20–200 μm. The breakdown voltage of these films in a sharply nonuniform field under pulses ranges from 4–6 kV for PS-20 to 10–12 kV for PTFE-200.

The impulse electric strength of the films (barriers) was measured in the field of “sphere—plane” electrodes. The diameter of the sphere was 50 mm. The electrode surfaces were carefully polished. To eliminate edge discharges, the electrodes with the film specimen were placed in pure glycerin. Breakdown of the films was carried out under pulses of positive polarity. The duration of voltage application was 3–5 μsec.

For each material, 15–20 specimens were punctured, and the mean value of the breakdown strength was determined from the expression

\[ E=\frac{U}{d}, \]

where \(U\) is the mean breakdown voltage in kV;
\(d\) is the film thickness in mm.

Figure 3 presents the results of measuring the electric strength of the films. The shaded areas show how much \(U_{\mathrm{br}}\) of glass specimens is increased by barriers made of these films. The barrier was located at a distance of 1.3 mm from the tip.

Fig. 3. Electric strength of films (barriers) and increase in the breakdown voltage of glass due to these barriers

It follows from Fig. 3 that the increase in \(U_{\mathrm{br}}\) of glass in the presence of a barrier decreases as the electric strength of the barrier decreases.

However, in these experiments, as the barrier thickness increases, its electric strength decreases. In order to keep \(E_{\mathrm{br}}\) of the barrier unchanged and increase the barrier thickness, several layers of film were placed between the glass plates.

Table 1 gives the results of studying the influence of several barriers on \(U_{\mathrm{br}}\) of glass.

The barriers were placed at a distance of 1.3 mm from the tip. The pulses were of positive polarity with \(\tau_{\mathrm{pr}} = 10^{-5}\) sec.

Table 1

From Table 1 it is evident that, as the number of barriers—and hence the barrier thickness—is increased, \(U_{\mathrm{pr}}\) of the glass increases only slightly, in fact by the value of \(U_{\mathrm{pr}}\) of the barriers themselves. When three PE-30 barriers are placed not flush against one another, but at distances of 0, 1.3, and 2.7 mm from the tip, \(U_{\mathrm{pr}}\) increases more substantially (by 22–30 kV compared with one barrier) and reaches 140–150 kV.

On the basis of a comparison of the \(U_{\mathrm{pr}}\) values of the films and the increases in \(U_{\mathrm{pr}}\) of the glass due to barriers made of these films, it may be concluded that the increase in \(U_{\mathrm{pr}}\) of glass with barriers depends almost not at all on the breakdown voltage of the barrier. Thus, a barrier of PTFE-3, having \(E_{\mathrm{pr}} = 464\) kV/mm and \(U_{\mathrm{pr}} = 5\)–7 kV in a sharply nonuniform field, increases the breakdown voltage of glass by 44% (34 kV), whereas a barrier of PTFE-200, having \(E_{\mathrm{pr}} = 224\) kV/mm and \(U_{\mathrm{pr}} = 10\)–12 kV in a sharply nonuniform field, increases the breakdown voltage of glass by only 22% (16 kV).

Three PE-30 barriers, each having \(U_{\mathrm{pr}} = 5\)–7 kV and placed not flush against one another, increase the breakdown voltage of glass by 100% (74 kV).

Since the principal role in increasing \(U_{\mathrm{pr}}\) is played by the electrical strength, and not by the thickness of the barrier, it may be assumed that a barrier with a thickness of units of microns or less will likewise lead to an increase in the \(U_{\mathrm{pr}}\) of a dielectric, especially since the electrical strength of solid dielectrics at very small thicknesses increases considerably [6–7].

To determine the influence of specimen thickness on the magnitude of the increase in \(U_{\mathrm{pr}}\) of a solid dielectric due to a barrier, \(U_{\mathrm{pr}}\) was measured for glass specimens of thickness from 3.0 to 8.1 mm without a barrier and with a barrier of 50-\(\mu\) polyethylene film, located at a distance of 0.2–0.4 of the specimen thickness from the tip.

As a result of the experiments we established that, in a sharply nonuniform field, \(U_{\mathrm{pr}}\) of specimens without a barrier and with a barrier is measured linearly in the thickness range 3.0–8.1 mm. The magnitude of the increase in \(U_{\mathrm{pr}}\) due to the barrier in the investigated thickness range remains practically constant and equal to \(\sim 60\%\).

When a barrier made of film is placed between the glass plates in a specimen, “additional” layers of liquid dielectric are formed. To avoid the formation of an “additional” layer, a thin layer of bakelite varnish, applied to the surface of the glass, was used as the barrier. In this case an increase in \(U_{\mathrm{pr}}\) of the glass in the presence of the barrier is also observed.

To determine the dependence of the glass \(U_{\mathrm{br}}\) on the thickness of the “additional” interlayer of liquid dielectric, which is formed between the barrier and the glass plates, a hole 20 mm in diameter was cut in the film barrier. The barrier with the hole was placed in the specimen in such a way that the discharge occurred through the hole, and the barrier did not hinder the development of the discharge. The thickness of the barrier, and hence also the thickness of the interlayer of liquid dielectric, was varied from 20 to 200 \(\mu\).

As a result of the experiments it was established that changing the thickness of the “additional” interlayer from 20 to 200 \(\mu\) does not lead to any noticeable increase in the glass \(U_{\mathrm{br}}\).

Proceeding from the foregoing, and taking into account that the electric strength of transformer oil at thicknesses up to 500 \(\mu\) is \(10^5\) V/cm [8], i.e., more than an order of magnitude lower than the electric strength of films (barriers), it may be concluded that the increase in the glass \(U_{\mathrm{br}}\) in our experiments occurs only at the expense of the barrier. Therefore, in the subsequent discussion we shall not take into account the presence of an interlayer of liquid dielectric between layers of solid dielectrics.

Experiments showed that, during breakdown under oblique-front pulses, the delay time of discharges in specimens with barriers is significantly greater than the delay time of discharge in specimens without a barrier. Figure 4 presents the dependence of the difference in discharge delay times in specimens with a PE-50 barrier and without a barrier on the steepness of the pulse. From Fig. 4 it is seen that the difference in delay times increases

Fig. 4. Dependence of the difference in discharge delay times in specimens with a barrier and without a barrier on the steepness of the voltage pulse

with decreasing pulse steepness and, at a steepness of the order of 12 kV/\(\mu\)s, reaches 5 microseconds and more.

During breakdown under rectangular pulses (0.2/20), the following results were obtained. The minimum breakdown voltage of specimens without a barrier is 61 kV; the breakdown time in this case is 1.0–1.5 \(\mu\)s. The minimum \(U_{\mathrm{br}}\) of specimens with a PE-50 barrier is 97 kV, and the time to breakdown is 1.2–2.2 \(\mu\)s. At this voltage,

at a voltage (97 kV), the time to breakdown of the specimens without a barrier is 0.3–0.4 μsec, i.e., 4–5 times less than the time to breakdown of specimens with a barrier.

On the basis of the experimental results, and taking into account that the statistical time lag of discharge in solid dielectrics is of the order of \(5 \cdot 10^{-9}\)—\(10^{-8}\) sec and is small compared with the total discharge time lag, it may be assumed that the increase in the time lag in the breakdown of a specimen with a barrier occurs as a result of an increase in the discharge-formation time.

Thus, the increase in \(U_{\text{br}}\) of a solid dielectric in the presence of a barrier in a sharply nonuniform field under pulsed voltage action can be explained by the fact that the barrier, which has a higher electric strength than the main dielectric, delays the development of the discharge until the field intensity at the head of the developing discharge reaches a value sufficient for breakdown of the barrier. The magnitude of the increase in \(U_{\text{br}}\) is determined mainly by the electric strength of the barrier.

Conclusions

1. The breakdown voltage of a solid dielectric can be substantially increased by means of a thin layer of another solid dielectric with higher electric strength.

2. The breakdown voltage of technical glass increases by 20–60% in the presence of a barrier made of films of high-polymer materials.

3. The relative increase in the breakdown voltage of glass due to the barrier is proportional to the electric strength of the barrier.

4. The increase in the discharge time lag in the breakdown of glass with a barrier depends on the steepness of the voltage impulse.

References

1. N. F. Voloshchenko. Influence of barriers on the electric strength of air gaps. Elektrichestvo, No. 3, 1946.

2. A. A. Shilvan. Investigation of a method for increasing the strength of liquid and gaseous insulation by using a barrier. Proceedings of LPI, 1, 1954.

3. Yu. N. Leont’ev, N. M. Torbin. Influence of the position of a barrier on the breakdown voltage of solid dielectrics. Izv. vuzov—Energetika, 12, 1961.

4. B. Mitov. Influence of barriers on the breakdown voltage of solid dielectrics (Bulg.). Yearbook—Mechanical-Electrotechnical Institute, 5, 2, 27, 1959.

5. D. D. Rumyantsev, N. M. Torbin. Influence of barriers on the breakdown voltage of certain solid dielectrics. “Breakdown of Dielectrics and Semiconductors.” Collection of reports of the IV Interuniversity Conference, 170, 1964.

6. A. A. Vorob’ev, G. A. Vorob’ev, V. A. Kostrygin. On the mechanism of electrical breakdown of solid dielectrics in a thin layer. Izv. AN SSSR, ONT—Energetika i avtomatika, No. 4, 1962.

7. V. A. Kostrygin. Investigation of discharge lag in NaCl single crystals of small thickness. FTT, Vol. II, No. 8, 1960.

8. D. A. Kaplan. Electric strength of transformer oil at small distances between electrodes. Elektrotekhnika, No. 10, 1964.

9. M. A. Mel’nikov. Dissertation, Tomsk, 1959.