V. D. Kuznetsov

The Effect of Gases Absorbed by Electrodes on the Electrical Conductivity of Liquid Dielectrics

The initial task of a researcher of the electrical properties of a liquid dielectric is to eliminate all external attendant circumstances that distort the properties of the object under study and that have nothing in common with the latter. It is very difficult to take account of or eliminate all factors lying outside the liquid and interfering with its study. This partly explains why the interesting and extensive field comprising liquid dielectrics has been relatively little elucidated.

One of the external circumstances obscuring the phenomenon of electric current through liquid dielectrics is the gases absorbed by the electrodes. They exert a profound influence and sometimes completely change the character of the process. In the case where occluded gases are contained in the electrodes, it is almost impossible to obtain definite and unchanging results.

Platinum electrodes are especially little suitable, although they are most often used. Platinum, as is known, absorbs gases in large quantity; they can be removed from it only by strong heating in vacuum while passing an electric current. Other metals, apart from palladium and iron, absorb gases to a considerably smaller extent than platinum, and therefore they can be used as electrodes. Gold is especially suitable; in an extreme case it may be replaced by silver. Brass, or nickel-plated brass, electrodes also make it possible to obtain quite definite results.

The effect of occluded gases is as follows. When a potential difference is applied to the electrodes, gases begin to be released from them, partly dissociated into ions, as a result of which the number of ions in the liquid increases and its electrical conductivity increases. As the current passes, impoverishment in ions takes place: ions which are released from one electrode are absorbed by the other, and the current decreases. The higher the voltage of the electric field, the more intense is the release of ions by one electrode and the more rapidly they are absorbed by the other electrode. If, after a prolonged passage of current in one direction, the direction of the field is changed, the occluded ions begin to be released in large quantity, causing an anomalously large current in the first moments after the change of direction. When an alternating field of sufficient frequency is applied, the occluded ions are only released from the electrodes but do not have time to be absorbed again; therefore the electrical conductivity of the dielectric, after the action of the alternating field has ceased and a field of constant direction has been applied, assumes an increased value.

Without yet touching upon the influence of gases on the dependence of current on voltage, we shall consider here only their influence on the electrical conductivity at constant voltage.

The condition of platinum electrodes appears especially clearly in the experiments of Schröder,^1 who investigated the electrical conductivity of pure ether, placing it in capacitors with unheated, strongly heated, and deliberately gas-saturated platinum electrodes.

In his first experiments the electrodes were heated in an airless space by passing through them an alternating current from a high-voltage source; but this method of purification proved insufficient to remove from the elec-

^1 L. Schröder. Ann. d. Phys. 29, 125 (1909).

all the absorbed gases. For poorly annealed electrodes the following results were obtained. After the current was closed, especially at high voltage—about 1000 volts with an electrode spacing of 2 mm—the current at first decreased rapidly, then slowly, and only after a rather long time became approximately constant. After passing an alternating current of high voltage—up to 1500 volts—the resistance of the ether to direct current greatly decreased—sometimes as much as 300-fold—then rapidly increased, tending approximately toward a constant value, which was nevertheless considerably less than the resistance before the alternating current had been passed.

With platinum electrodes well annealed in a vacuum, more constant and definite results were obtained, although it was impossible to achieve complete removal of the gases. After the application of a potential difference, the current decreased rapidly and rather soon acquired a constant value.

When the direction of the field was reversed, the electrical conductivity in the first moments was several times greater than in the first direction; then, over the course of several minutes, it decreased, then rapidly increased, reached a maximum, and again slowly decreased to a constant value. All these phenomena are easily explained if one assumes, as we have done, that dissociated gases are present in the electrodes.

To show that gases absorbed by the electrodes do indeed influence the character of the phenomenon, Schröder carried out experiments with platinum electrodes deliberately saturated with hydrogen and oxygen. For this purpose the electrodes were immersed in a solution of sulfuric acid and a potential difference was applied to them. In general, the results in the latter case proved similar to the results with unannealed electrodes. The essential difference was that the electrical conductivity exhibited a distinct unipolarity. When current was passed in such a direction that the electrode saturated with hydrogen was connected to the negative pole, the current was greater than in the reverse direction. To explain the observed dissymmetry, one may suppose that the number of ions in the ether is affected by the contact potential difference between the platinum and the liquid, and that for pure platinum this difference is greater than for platinum saturated with gas. The unipolarity may be explained by assuming that, owing to the contact potential difference, unequal quantities of ions of the one and the other sign are formed at the electrodes. Suppose, for example, that in the liquid immediately at the electrode saturated with hydrogen the negative ions predominate over the positive ones, and that the electrode itself has a positive charge, while near the electrode saturated with oxygen chiefly positive ions are formed; then, when the current is directed from the electrode with oxygen to the electrode with hydrogen, the current should be greater than in the reverse direction.

Not assuming that electrodes are of great importance in the investigation of the electrical properties of liquid dielectrics, I made many observations of the change with time in the current through dielectrics, using several condensers with different platinum electrodes. Obtaining very variable and indefinite results, I vainly sought the cause of the uncertainty in various circumstances. Only after comparing the results of experiments with platinum and brass electrodes did I become convinced of the complete unsuitability of unannealed platinum electrodes, especially those that had been used in the electrolysis of acid solutions.

Thus, for example, for ten days without interruption I passed a current of 480 volts through purified petroleum placed in a glass vessel with platinum electrodes, and I was unable to achieve not only constancy of the current, but even the absence of very strong fluctuations. The current was determined by observing the rate of deflection of the needle of a quadrant electrometer. The cause of the failure should be ascribed exclusively to the electrodes, since the same method

with the same oil, but with brass or nickel-plated brass electrodes, yielded quite definite results.

To illustrate the proposition stated above concerning the complete unsuitability of unannealed platinum electrodes, I shall give here a description of one very carefully conducted investigation of the change, with time, of the electrical conductivity of toluene at constant voltage.

Preliminary experiments showed that the change in the resistance of toluene has a very complex and confused character, in which it was impossible to discern any regularity. This led me to carry out a very thorough purification of the toluene. Commercial toluene (toluolum purum) from Scholl and Schmitt was distilled four times. The liquid was boiled in a glass flask with a long vapor-conducting tube entering a condenser, and was collected in a receiving flask. The flask, condenser, and receiver were first washed with distilled water, then with a hot aqueous solution of nitric acid, then with a hot solution of caustic soda, again with distilled water and, finally, with absolute alcohol, after which they were dried over a burner flame. At each distillation the first portion was discarded and all parts were rinsed with a pure sample. The current strength was determined by the compensation method with piezo-quartz. Special attention was paid to electrostatic shielding. Parallel connection of several batteries of small storage cells gave a potential difference of 110 volts. The current strength rapidly decreased during the first 15 minutes; then, up to 80 minutes, it remained approximately constant, increasing slightly; thereafter, from 80 to 200 minutes, the current strength decreased very irregularly, in jumps, and by the end had decreased by more than a factor of 2 in comparison with the period 15–80 minutes; from 200–250 minutes it increased, and from 250 to 280 minutes it rose sharply and rapidly by about \(1\frac{1}{2}\) times, and then again began to decrease, etc.

The curves of current strength as a function of time for one and the same toluene, upon repeated observations, had so varied an appearance that further investigations had to be discontinued as useless. Even at the low potential difference of 55 volts, irregular curves were obtained. However, from comparison of several curves for toluene it nevertheless proved possible to draw the following two conclusions: 1) the increase in resistance with platinum electrodes does not proceed smoothly, but in jumps. Almost always, following a very rapid increase in resistance immediately after the application of the potential difference, a slow decrease of it is observed; then it again rapidly increases, etc. 2) On passing from a lower to a higher voltage, a substantially reduced resistance is observed at first, and only after several minutes does it begin to increase.

Purification of toluene by passing an electric current for several hours, at potential differences from 70 to 600 volts, produced no noticeable improvement. The toluene was poured into a glass vessel of cubic shape (side 10 cm), and two platinum plates (9 × 9 cm), connected with the poles of a storage battery, were immersed in it. After several hours the toluene was again poured into the measuring condenser and observation of the course of the curve was carried out.

The results of investigations with petroleum, with various purified fractions of kerosene, and with benzine likewise proved indeterminate, although all possible precautions were taken. Moreover, for one and the same dielectric the methods of measuring the current strength were varied, the shapes of the condensers and the shapes and sizes of the platinum electrodes were varied, but in all cases very diverse curves were obtained, from which only the two conclusions stated above could be drawn. As electrodes I used 1) platinum wires, 2) a wire and a coaxial cylinder of platinum sheet, and 3) platinum plates of various sizes.