Miodrag K. Radović

Statistical analysis of the dynamic voltage electrical breakdown in nitrogen

The statistical analysis of the electrical breakdown in nitrogen at p=20 mbar using dynamic method is presented in this paper. The analysis was performed for the working voltage increase rates 50, 150, and 300 Vs/sup -1/. For each working, voltage increase rate a 1000 measurements were done. It was shown that results are lined up accidentally without any systematic trend. New form of breakdown voltage density distribution is suggested and verified with histograms of the breakdown voltage.

Statistical analysis of electrical breakdown time delay distributions in neon tube at 13.3 mbar

Results of the statistical analysis of the electrical breakdown time delay for a neon-filled tube at 13.3 mbar are presented in this paper. Experimental distributions of the breakdown time delay were established on the basis of 1000 successive and independent measurements, deviating from usual exponential distribution. Breakdown time delay distributions are numerically generated from the composition of two parameters, which correspond to the random variables of an exponential distribution and of a Gaussian distribution. Theoretical breakdown time-delay distribution is obtained from the convolution of the exponential and Gaussian distributions. The numerically generated and theoretical breakdown time delay distributions show good correspondence with the distributions of experimental data.

The application of convolution-based statistical model on the electrical breakdown time delay distributions in neon

Results of the statistical analysis of the electrical breakdown time delay for neon-filled tube at 6.5 mbar with and without /spl gamma/ and ultraviolet (UV) irradiation are presented in this paper. Experimental distributions of the breakdown time delay are established on the basis of 1000 successive and independent measurements for different exposition dose rates of /spl gamma/-irradiation, and UV illumination with known spectrum. Obtained experimental density distributions are modeled with convolution-based model, proposing two independent random variables with exponential and Gaussian distribution, respectively. The tendency of the breakdown time delay distribution to change shape from the exponential-like to the Gaussian-like is found by increasing the exposition rate of the /spl gamma/ irradiation. In addition, the characteristic strong reduction of the time delay is identified in the tube under the UV illumination.

The statistical time-delay and the breakdown formative time contributions to the memory effect in Ne at 7 mbar pressure

The results of the electrical breakdown time-delay as a function of the relaxation time τ, known as the memory curve, are presented. The memory curve was experimentally established for the relaxation times from 3u2009ms to 300u2009s, for the Ne-filled diode at 7u2009mbar pressure. Statistical analysis was used to analyse the obtained time-delay values. The results show that for the relaxation times τ ≤ 0.1u2009s the dominant part of the total time-delay is the breakdown formative time (tf ~ 0.1u2009ms). The breakdown formative time obtained in this investigation shows the same memory effect as the breakdown total time-delay, because it slowly increases with τ. For τ > 0.1u2009s, the statistical time-delay makes up the dominant part of the total time-delay, as was reported earlier.

On the complex structure of the breakdown time-delay distribution in a neon-filled diode at 4 mbar pressure

The unusual and complex structures of the breakdown time-delay distributions in a neon-filled diode at 4 mbar pressure for lower overvoltages are reported in this paper. Each distribution was established on the basis of the 1000 successive and independent measurements of the electrical breakdown time delay. It is concluded that the expected one-stage Laue distribution cannot be applied successfully because of the complex structures of the measured distributions. For very low overvoltages, multiple breakdown mechanisms seem to appear with different (by two orders of magnitude) mean values of the statistical time delay and formative time. A multi-stage model for the Lauegrams is used in order to fit our data.

The study of the non-radiative life-time of the 1S5 metastable states of neon atoms under its own gas pressure of 13.3 mbar

The probability of electrical breakdown in neon is investigated under a pressure of 13.3 mbar, using the breakdown time-delay measurement method. It is found, by applying a laser spectroscopy technique, that the 1S5 metastable states of neon atoms dominantly contribute to the breakdown probability in the moment when more than 1 s has passed after the previous discharge. With this method the presence of 1S5 states of Ne atoms is detected in the gas, under the pressure mentioned, up to about 500 s after the cessation of the breakdown.

The self-quenching process and half de-excitation time of metastables in a nitrogen afterglow

The dependencies of electrical-breakdown time delays td on the afterglow period tau were measured and the dependence of the mean values td on tau , using the so-called memory curves td=f( tau ), was established in nitrogen-filled diodes at 1.3 mbar and 13.3 mbar. The corresponding time-dependent variations of metastable concentration n, i.e. the curves n=F( tau ), were derived, where, at afterglow periods exceeding 1 s, the metastable molecules were predominantly in the N2(A3 Sigma u+) state. A physical model of metastable de-excitation processes has been proposed and, by solving its differential equations, the following dependencies and constants were obtained: (1) the theoretical dependence of metastable concentration on the afterglow time-the curve n=F( tau ), (2) the range of n0, the initial concentrations, which are also saturation concentrations of metastables reached in discharges, (3) the metastable concentration-independent constant k in the equation td=k/n, (4) the dependence of the half de-excitation time T1/2 on the afterglow period, T1/2= phi ( tau ), and (5) the dependence of Z on the afterglow period, Z= Phi ( tau ), where Z is the concentration dependent ratio between the de-excitation rate via binary metastable-metastable collisions and the total rate via all other considered de-excitation mechanisms.

Breakdown-voltage memory effect in a neon-filled diode at 1 mbar

Results of the investigation of the breakdown-voltage distribution dependence on the relaxation time of the diode [the curve U~/sub b~/=f(/spl tau/)], i.e., memory effect, in neon at p=1 mbar are presented in this paper. The memory effect is determined for the diode-relaxation times 0.1-100 s. The applied voltage was linearly increased with the increasing rate 10 V/s. For each value of the relaxation time, the series of 200 successive and independent measurements were done. The numerical fitting of the theoretical expression of the breakdown-voltage distribution on the histograms of the experimentally established data was used to determined relative yield in the diode for the different relaxation times Y/Y/sub 0/=f(/spl tau/). Results show that the yield in the diode decreased in that time interval in one order of magnitude. Quantitative parameters for two-step decreasing and qualitative explanation of these dependencies are given.

Statistical analysis of the electrical breakdown time delay distributions in krypton

The statistical analysis of the experimentally observed electrical breakdown time delay distributions in the krypton-filled diode tube at 2.6mbar is presented. The experimental distributions are obtained on the basis of 1000 successive and independent measurements. The theoretical electrical breakdown time delay distribution is evaluated as the convolution of the statistical time delay with exponential, and discharge formative time with Gaussian distribution. The distribution parameters are estimated by the stochastic modelling of the time delay distributions, and by comparing them with the experimental distributions for different relaxation times, voltages, and intensities of UV radiation. The transition of distribution shapes, from Gaussian-type to the exponential-like, is investigated by calculating the corresponding skewness and excess kurtosis parameters. It is shown that the mathematical model based on the convolution of two random variable distributions describes experimentally obtained time delay ...

Long formative time lags in nitrogen below the Paschen minimum

The formative time lag of an electric discharge for an overvoltage of 1-8%, below the minimum of the Paschen curve in nitrogen is investigated. The Laue distribution of the breakdown time delays is found to apply. The formative time rises with decreasing pd below pdmin from values of 10-3 s, up to the order of seconds. The suggested explanation is that avalanches in nitrogen are not independent. This poses the question of the definition of the discharge formative time.