The influence of positive ions and neutral active states on the memory effect in nitrogen-filled discharge tubes at 1.3 and 4.0 mbar pressures has been investigated. Analysing the thermal velocity of these particles as well as drift velocity of positive ions, it is shown that the positive ions have a dominant role in the breakdown initiation for afterglow periods lower than 30 ms, while the role of the neutral active states can be ignored because of their electrical neutrality. The estimated value of the positive ion recombination time in the afterglow for the pressures considered is about 30 ms. The atoms in ground states have a dominant role in breakdown initiation for afterglow periods higher than 30 ms. Our supposition is based on the earlier results that the long-lived Lewis–Rayleigh glow lasts up to several hours as a consequence of the recombination of these atoms on the wall. A detailed analysis of relaxation processes has shown that a contribution of vibrationally excited molecules as well as atomic and molecular metastable states created during the previous discharge cannot contribute to the memory effect significantly. Also, the contribution of cosmic rays to memory effect is analysed for the afterglow periods when the concentration of atoms in the ground state decreases to such a low value that their role in the secondary electron emission from the cathode is ignored.

The memory effect in air at 0.7 and 6.6 mbar pressures, due to post-discharge survival of some species that affect subsequent breakdowns, was analysed using memory curves. In order to complete the analysis of the processes, memory curves for nitrogen have also been monitored, because air is 78% nitrogen. In an early afterglow, the memory effect in air, as well as in pure nitrogen, is a consequence of the same processes, i.e. the presence of N + 2 and N + 4 ions, formed by the collision between nitrogen metastable molecules from the previous discharge. The concentrations of nitrogen ions in an air-filled tube are lower than the concentration in a nitrogen-filled tube for a given afterglow period because of their recombination with O, O 2 and NO particles which are also present in the early afterglow in air. De-excitation of N 2 (A) and N 2 (a ) metastable molecules due to their collision with O 2 , O and O − particles also contributes to the reduction in the concentration of N + 2 and N + 4 ions. In the late afterglow in air and pure nitrogen, N( 4 S) atoms created during the previous discharge are responsible for the memory effect. In the early air afterglow N( 4 S) atoms can be formed by the collision of O + ions with N 2 (X) molecules. However, because of the relatively small concentration of these ions the most probable process is the recombination of N( 4 S) atoms with O 2 , O, NO and NO 2 particles. The net effect is that the concentration of N( 4 S) atoms in the late afterglow is less in air than in nitrogen. When the concentration of N( 4 S) atoms is sufficiently reduced, the breakdown is initiated by cosmic rays.

The mechanisms responsible for memory effect in nitrogen at 6.6 mbars have been analysed based on experimental data of electrical breakdown time delay as a function of afterglow period. The analysis has shown that positive ions remaining from previous discharge, as well as metastable and highly vibrationally excited molecules, are responsible for memory effect in the early afterglow. These molecules lead to the formation of positive ions in mutual collisions in the afterglow. Positive ions initiate secondary electron emission from the cathode of a nitrogen-filled tube when voltage higher than static breakdown voltage is applied on the electrodes. On the other hand, N(S4) atoms have a large influence on memory effect in late afterglow. They recombine on the cathode surface forming N2(AΣ3u+) metastable molecules, which release secondary electrons in collision with the cathode. The higher values of electrical breakdown time delay in the case of the tube with borosilicate glass walls than in the case of the tube with copper walls are a consequence of faster de-excitation of neutral active particles on the glass. Indirect confirmation of this assumption has been obtained when the tubes were irradiated with gamma radiation.

The results of statistical analysis of the electrical breakdown voltage distribution in nitrogen are presented in this paper. For obtaining the experimental results, a nitrogen‐filled diode at 13.3 mbar pressure was used. The dynamic method was used for estimating the static breakdown voltage. One‐hundred voltage measurements were carried out for each value of the increase in the voltage rate from 1 up to 10 V/s. Wilcoxon test was used for checking the measurements' randomness for each rate. The experimental distribution functions are fitted with the three‐parameter Weibull distribution function. The coefficients of the Weibull distribution are also estimated.

No abstract