Formative Time Lags of Spark Breakdown in Air in Uniform Fields at Low Overvoltages

Abstract

An attempt has been made to establish the region of validity of the streamer and Townsend mechanisms of spark breakdown in air by measurements of the formative time lag of the breakdown process.Formative time lags were measured in a uniform field for overvoltages of a few percent down to as close to threshold as possible. Such measurements have been carried out as a function of pressure (atmospheric to a few cm of Hg) and plate separation (0.3 to 1.4 cm). For pressures greater than 200 mm Hg, the formative time lags very close to threshold are of the order of 100 \ensuremath{\mu}sec and longer. These times are orders of magnitude lnger than those previously reported. For all pressures studied, the time lags decrease as the percent overvoltage (percent o.v.) increases, until at about two percent above threshold, the formative times are of the order of 1 \ensuremath{\mu}sec. The time lag vs percent o.v. curve is independent of pressure from atmospheric down to 200 mm Hg. At a given percent o.v., the time lags increase linearly with plate separation. Changing the approach voltage from two to four kv below breakdown does not affect the results appreciably. The number of initiating electrons at the cathode has been varied by a factor of about seven, and this again does not materially alter the results.The long time lags and their dependence on pressure and percent o.v. cannot be explained by secondary emission of electrons from the cathode by positive ion bombardment. The proper explanation is the enhancement of field intensified ionization due to field distortion acting in conjunction with a photoelectric secondary process. The experiment demands an extension of the streamer concept of breakdown, and makes very questionable the role of positive ion bombardment of the cathode in spark breakdown for the pressures and plate separations studied. No transition region for change of streamer to Townsend breakdown was found.