ELECTRIC STRENGTH OF SOME LIQUID DIELECTRICS IN THE NANOSECOND RANGE OF VOLTAGE IMPACTS

V. S. KOROLEV, N. M. TORBIN

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

In connection with the development of nanosecond pulse technology and the possibility of using liquids with high dielectric permittivity as insulation under pulse voltage [1–3], as well as the possibility of studying the physical processes occurring during breakdown [4], the investigation of the electric strength of liquid and solid dielectrics at short voltage-application times is of interest.

Breakdown of liquid dielectrics in the nanosecond voltage range has been insufficiently covered in the literature, and only isolated works [2, 3, 4] are devoted to this question.

They show that the electric strength of some liquids at voltage-application times on the order of \(10^{-8}\) and less may exceed the electric strength of solid dielectrics.

Presented below are the results of experimental investigations of the breakdown of some technically purified liquids at voltage-application times of \(2 \div 3 \cdot 10^{-9}\) sec and distances between electrodes of 0.3, 0.5, 0.8, 1.0, and 1.3 mm, with negative polarity of the point.

As the liquids under investigation, technically purified liquids were selected: transformer oil \((\varepsilon = 2.2)\), hexane \((\varepsilon = 2.4)\), glycerin \((\varepsilon = 39)\), and distilled water \((\varepsilon = 80)\).

Before testing, the transformer oil was centrifuged, and its electric strength in a standard breakdown cell was 180 kV/cm; hexane, glycerin, and water were distilled once.

The liquids tested possess different physicochemical and dielectric properties, and they have repeatedly been subjected to breakdown studies over a broad range of voltage exposures.

The source of high voltage was a generator of nanosecond pulses with a maximum amplitude up to 420 kV and a front duration \(\tau_{\phi} = 2—3\) nsec [5]. To record the breakdown voltage, an OK-19M oscilloscope was used; calibration of the oscilloscope was carried out according to the method of [5].

A liquid sample with a volume of about \(100\ \text{cm}^3\) was placed in a breakdown cell made of organic glass, and the breakdown cell was then placed in a gap in the central conductor of the generator’s coaxial line. The discharge occurred across an open-circuited section of coaxial line 20 cm long.

Breakdown of the liquid was carried out in electrode systems of needle—needle and needle—plane. Steel—

steel needles with a radius of curvature of about 3 μ and a brass plane 25 mm in diameter with rounded edges. Before each series of experiments the plane was polished on a felt wheel with GOI paste.

The interelectrode distances were measured with a micrometer with a measurement accuracy up to 0.01 mm. Liquid samples were broken down on the front of a pulse with $\tau_f = 2.5$ nsec. At each value of the interelectrode distance, 18–20 measurements were made, and the breakdown voltage at each point was determined as the arithmetic mean.

After 6–7 breakdowns the liquid was replaced, and the condition of the electrodes and the interelectrode distance were checked.

It was established that after 5–8 breakdowns strong erosion of the point was observed, and the radius of curvature of the needle changed from 3 μ to 10–12 μ. In this case the breakdown voltage increased by approximately 15% at an unchanged interelectrode distance and subsequently remained stable within the scatter, while the radius of curvature of the needle remained almost unchanged up to 80–100 breakdowns.

Figures 1 and 2 present the mean values of the breakdown field strength as a function of the interelectrode distances for the liquids we investigated. From Figs. 2 and 3 it is seen that in the nanosecond

Fig. 1. Dependence of the electric strength of transformer oil (1), hexane (2), and distilled water (3) on the distance between the electrodes. Electrodes: point negative—plane grounded.

time range, transformer oil, hexane, glycerin, and distilled water in a nonuniform field, at distances between the electrodes $d = 0.3 \div 1.3$ mm, have approximately the same electric strength. With a decrease in the distance between the electrodes, the electric strength of the investigated liquids increases somewhat. In this case the mean values of $E_{\mathrm{br}}$ in a sharply nonuniform field approach the values of $E_{\mathrm{br}}$ of these liquids in a weakly nonuniform field and at times to breakdown of less than $5 \cdot 10^{-9}$ sec [3].

It is indicated in [3] that, at field strengths on the order of 6 MV/cm, the discharge time for gaps of 50–500 μ at exposure times of about 5 nsec becomes practically identical.

In [6] it was established that, with a front steepness of the impulse wave above 400 kV/μsec, the mean velocities of discharge development become close and reach the order of $10^7$ cm/sec; in this case, flashes of the leader channel, characteristic of a polar liquid, are not observed.

An even greater steepness of the voltage pulse, \((50 \div 60)\,10^3\ \text{kV}/\mu\text{s}\), in the nanosecond range in our experiments and in [3] leads to still higher rates of discharge development (above \(10^7\ \text{cm}/\text{s}\)). The fact that the electric strength of liquids having different physicochemical properties, in the nanosecond range of voltage action \((\tau = 2 \div 4\ \text{ns})\), has the same

Fig. 2. Dependence of the electric strength of transformer oil, \(U_{\mathrm{br}} = 180\ \text{kV}/\text{cm}\) (1), hexane (2), distilled water (3), and transformer oil, \(U_{\mathrm{br}} = 44\ \text{kV}/\text{cm}\) (4), on the distance between the electrodes. Electrodes: point negative—point grounded

electric strength indicates that, at high field intensities, discharge formation in different liquids proceeds in the same way. The interactions of electron avalanches with ions and molecules of the liquid, which play a major role in discharge formation and depend to a considerable extent on the physicochemical properties of the liquid during discharge formation in relatively weak electric fields, probably play a secondary role in the breakdown of liquids with large overvoltages in the nanosecond voltage range.

The same electric strength in the nanosecond voltage range during breakdown of solid crystalline dielectrics of the alkali-halide series (NaCl, KCl, KBr, KJ) was observed at times of about \(5 \cdot 10^{-9}\ \text{s}\) [8].

The author of [8] explains the results obtained by leveling of the delay times: solid dielectrics having lower electric strength have longer delay times in the microsecond range of voltage action, and in the nanosecond range they break down at higher voltages, since breakdown occurs on the front of an exponentially rising pulse.

It is indicated in [4] that the delay times of discharge in the nanosecond voltage range obey the normal distribution law, whereas in [3] a distribution of delay times according to the exponent was obtained.

The discharge delay times measured by us in transformer oil and in water (Fig. 3), depending on the distance between the electrodes, for each distance fit the Laue curves [7] sufficiently well; moreover, with increasing distance between

between the electrodes an almost linear increase in the discharge delay time is observed. If one assumes that the discharge delay time in the nanosecond voltage range consists of the discharge formation time [4], then one may expect a linear increase in the discharge delay time with increasing distance between the electrodes, as in our Fig. 3; however, the constancy of the discharge time when the gap length is varied from 50 to 500 µ at nearly identical field strengths [3] is still difficult to explain.

Fig. 3. Dependence of the discharge delay time on the distance between the electrodes: 1—distilled water, 2—transformer oil.

Thus, the results we have obtained confirm the possibility of using, as insulation in the nanosecond voltage range, distilled water, glycerin, and probably other liquids that have poor insulating properties under ordinary conditions. Further studies are required to elucidate the processes occurring during the breakdown of liquids in the nanosecond voltage range.

REFERENCES

1. Ed. by A. A. Rukhadze, Exploding Wires. IL, Moscow, 1963.
2. N. S. Rudenko, I. V. Tsvetkov, ZhTF, XXXIV, issue 6, 1964.
3. N. S. Rudenko, V. I. Tsvetkov, ZhTF, XXXV, issue 10, 1965.
4. A. J. Beddow, J. E. Brignell, Electron. Letters, vol. 1, No. 9, 1965; Electron. Letters, vol. 2, No. 4, 1966.
5. G. A. Vorob'ev, N. S. Rudenko, PTE, No. 1, 1965.
6. V. Ya. Ushakov, ZhTF, XXXV, issue 10, 1965.
7. J. B. Birks, Progress in Dielectrics, 4, 1962.
8. M. A. Mel'nikov, Izv. Vuzov—Fizika, No. 5, 1958.