V. D. Kuznetsov.

The Phenomenon of Residual Charge in Liquid Dielectrics.

(To the article: 1 page of drawings).

In certain solid and liquid dielectrics there is observed the phenomenon of residual charge, which consists in the following. If, for some time, a constant electric current is passed through a dielectric and then, after removing the electromotive force, the electrodes are connected with one another, then in the closed circuit an electric current arises in the direction opposite to the original one. From the ionic point of view this phenomenon is explained by the fact that, during the passage of the current, ions accumulate at the electrodes; after the external electric field has been destroyed, owing to their mutual repulsion, these ions disperse through the liquid, chiefly in the direction opposite to the original motion, and create a convection current.

Residual charge in liquid dielectrics was first discovered by H. Hertz¹. After a whole series of experiments he settled on commercial benzine, as a substance which possesses a large resistance and gives the phenomenon of residual charge. He passed a current of 100 volts for 24 hours, measuring from time to time the resistance and the residual charge. The experiment consisted in the following: a pair of quadrants of an electrometer was insulated one second after the voltage had been removed, and the position of the needle was observed. It turned out that the needle rapidly departed from the position of equilibrium in the negative direction, reached a maximum, and slowly returned to zero. The strength of the direct current decreased with time, and at the same time the strength of the reverse current and the amount of electricity carried by this current diminished as the time elapsed from the beginning of the application of the electromotive force.

To elucidate the cause of the appearance of the residual charge, Hertz carried out the following experiments.

At the moment when the external electric field was destroyed and a pair of quadrants was insulated, i.e. when the residual charge began to appear, he opened the tap at the bottom of the metal cylinder that served as the outer plate of the condenser, and the benzine flowed out. The deflection of the needle during the outflow occurred in the same direction as at rest. This, in Hertz’s opinion, proves that the residual charge is caused by polarization of the electrodes and is not a consequence of the existence of free electricity accumulating near the electrode during the passage of the direct current.

In another experiment Hertz poured benzine into a large vessel with two mutually parallel plates (electrodes), situated at a distance of several centimeters from one another. Some time after the passage of the current, the electromotive force was removed, the electrodes were connected to earth, and immediately a system of three plates, 1, 2, and 3, was placed between them. Plate 1 was led to earth; plate 2 was connected to one pair, and plate 3 to the other pair of quadrants of the electrometer. In addition, plates 2 and 3 were separated from one another by an insulator, so that the benzine could not enter between them, whereas between plates 1 and 2 there was a layer of benzine several

cm. When this system was moved, the needle of the electrometer deviated in the opposite direction, the deflection being smaller than when electrodes A and B were short-circuited. This experiment likewise led Hertz to the conclusion that polarization exists, consisting in the formation of double layers at the electrodes, which Hertz attributed to foreign impurities.

Quincke¹ was unable to detect the existence of a polarization current for turpentine, petroleum, carbon disulfide, ether, and benzene; but this may be ascribed to the insufficient sensitivity of his method, which was adapted to measuring the magnitude of the current at very high voltages (up to \(10^3\) volts per cm) and proved insensitive to weak residual charges.

In 1889 J. Curie² formulated for solid dielectrics the law of superposition (loi de superposition): “Chaque variation de force électromotrice entre les deux faces de la lame agit comme si elle était seule,” i.e., every change in electromotive force applied to the two faces of a plate (dielectric) produces the same action as if this change were the only one. This law can be interpreted in two ways.

As is known, in most cases the current through liquid or solid dielectrics is a decreasing function of time. Let, for some electromotive force \(V_1\), the current \(i_1\) be represented by curve A (Fig. 1), where time is plotted along the abscissa axis and current along the ordinate axis. After some time \(\theta\), a new electromotive force \(V_2\) is added to the potential difference \(V_1\). If only the electromotive force \(V_2\) acted, then the current \(i_2\) would be represented by curve B. Under the joint action of \(V_1\) and \(V_2\), the current \(i\) is represented by curve C, the ordinates of which at each instant are the sum of the ordinates of curves A and B. This is the first interpretation of the law of superposition.

Let us now suppose that, after a time \(\theta\) following the application of the electromotive force \(+V_1\), the electric field is destroyed. Formally this is equivalent to saying that at the instant \(\theta\) a new electromotive force \(-V_1\) is applied, equal in magnitude but opposite to the first. The current strength in the presence of only \(-V_1\) would be represented by curve B (Fig. 2). Under the joint action of \(+V_1\) and \(-V_1\), the current changes along curve C, which is obtained by subtracting the corresponding ordinates of curve A from the ordinates of curve B. This phenomenon is called the phenomenon of residual charge, or residual (return) current, and is the second interpretation of the law of superposition.

For liquid dielectrics the law of superposition, in the above formulation, apparently is not valid; but the phenomenon of residual charge, which is observed in certain dielectric liquids, shows that for them too there exists something akin to the law of superposition.

I tried to detect a residual current in petroleum, toluene, and benzene, using various methods; but in all cases this current was so small and disappeared so rapidly that it was impossible to make any quantitative observations. In all these experiments I used a quadrant electrometer. Fig. 3 shows one of the arrangements I tested. Here IK is a measuring capacitor with a guard ring, BS is a large resistance (a glass tube with a colloid), and \(P_1\), \(P_2\), and \(P_3\) are paraffin switches with cups filled with mercury. The distance between the electrodes was 10 mm. The experiment consisted in the fact that, after some time following the application of a potential difference, switch \(P_1\) was thrown to ground, KL was opened, and the deflection of the electrometer needle was observed. The potential difference was varied from 53 to 738 volts. The residual current was barely noticeable.

When experiments similar to the one described showed that a residual charge undoubtedly exists, but that it disappears so rapidly that, owing to the large inertia of the needle of the quadrant electrometer, it is almost impossible to detect it,

On V. D. Kuznetsov’s paper, “The Phenomenon of Residual Charge in Liquid Dielectrics.”

Drawing 1.

Drawing 2.

Drawing 3.

Drawing 4.

Petroleum.
Jaffe condenser.
Dependence of the residual polarization e.m.f. on the potential difference.
\(25 \cdot 1 \cdot 10\,\mathrm{g}.\)

Drawing 5.

Vertical axis: Residual polarization e.m.f. (volts).
Horizontal axis: Applied potential difference (volts).

Then I changed the apparatus, making use of Wilson’s electroscope with a small capacitance, and measured not the residual-current intensity, but the electromotive force of polarization. The latter experiments gave some interesting results and showed that the method for measuring the residual charge had been chosen correctly.

Fig. 4 gives the scheme of the apparatus. The outer plate of the measuring condenser, M.C., consisting of a brass vessel with an internal rod (Jaffé condenser), could be connected, by means of a switch, either with the pole of the battery \(B_1\), or with earth; at the same time the inner plate, together with the leaf of the electroscope, was either connected with earth or insulated. The experiment consisted in the following: during a time \(T\) the outer plate was connected to the potential \(V_1\); then switch \(P\) was thrown over, and the maximum deflection \(a_m\) of the electroscope leaf was observed. At the moment when the leaf reached the extreme position, the stopwatch was started, and it was stopped at the moment when the leaf reached the zero position \(a_0\). This interval of time is denoted by \(t\). The electroscope had been calibrated beforehand. The sign of the charge of the electroscope was opposite to the sign of the potential \(V\). I assumed that the difference of the readings \(a_m-a_0\) characterizes the magnitude of the electromotive force of polarization which gives rise to the residual current.

As an example I give observations for petroleum. The results are also shown in Fig. 5.

25·1·19.

Table.

From the table it is seen that the throw of the leaf does not depend on the time of passage of the direct electric current in the interval from 5 seconds to 5 minutes, and up to \(V=+80\) volts is proportional to the potential difference, and then increases more rapidly than \(V\).

Against the experiments one may raise the objection that they do not give the electromotive force of polarization in a pure form. In fact, if the electroscope is insulated before the outer plate of the condenser is connected with earth, then the charge of this plate at the moment of discharge induces a charge in the leaf, which also accounts for part of the throw. However, experiments modified in such a way that first the outer plate was connected with earth, and then the leaf and the inner plate were insulated, gave exactly the same results.

1. Quincke. W. A. 28, 529 (1886). ↩↩

2. J. Curie. Ann. de chim. et phys. (6) 18, 203 (1889). ↩