The role of impurities on the deviation from Paschen's law of SF₆

Abstract

Deviations to Paschen's law for SF6 at high pressures are commonly explained by field enhancement due to protrusions supposed to exist at the surface of the electrodes. New experiments for positive polarity allow the authors to draw a different conclusion, i.e. to ascribe these deviations to the impurities of the gas, especially to the combined effect of O2 and H2O. For negative polarity, impurities play a minor role and the threshold is very sensitive to the surface state of the stressed electrode.

Illusory Paschen curves associated with strongly electronegative gases

“…This situation is a direct consequence of the fact that the experimental A.value is very dependent on the absolute accuracy of the voltage measurement, while the corresponding B value is only dependent on the relative accuracy of this measurement. Although no information about these measurement accuracies is given by Berger and Senouci [7], the relative accuracy is generally much better than the absolute accuracy. In relation to the A values of Table 1 it should be noted that A (Paschen) is only 0.35 kV [16].…”

“…With reference to strongly electronegative gases, it is found that upon application of the streamer criterion to the above field distribution one obtains and - (7) see [ll, 141 for details. The parameter M represents the figure of merit for a strongly electronegative gas (121, while Hso is the normalized critical avalanche length.…”

“…To illustrate the present analysis, we will refer to the experimental study undertaken by Berger and Senouci [7] , in which a point/plane gap was employed to investigate Paschen's law for SFG. The severe limitations associated…”

“…Berger and Senouci tabulate in their paper [7] the corresponding measured U, values and these are listed in Table 2. On applying a numerical linear regression analysis to these data for the same pressure range, we obtain the values given in Table 1, column 2.…”

“…Despite this discharge restriction, a n apparently linear breakdown voltage curve can be obtained [7] which is then mistakenly identified as a Paschen curve. It is only by extracting the relevant parameter values from the experimental data that the anomalous nature of such a curve becomes apparent [8].…”