Very Short Time Lag of Sparking

Abstract

Time lag of sparking in the very short time interval from 5\ifmmode\cdot\else\textperiodcentered\fi{}${10}^{\ensuremath{-}9}$ sec. to 115\ifmmode\cdot\else\textperiodcentered\fi{}${10}^{\ensuremath{-}9}$ sec. has been measured as a function of overvoltage and illumination using the electro-optical shutter with a constant illumination method. It was found that down to ${10}^{\ensuremath{-}9}$ sec. the time lags approached no lower limit if sufficiently high overvoltage was used. Variation of the intensity of illumination showed that the time lags measured consisted of two parts, a statistical part which is the time taken for an electron to be produced in the spark gap, and a formative time of more or less definite length depending on overvoltage and during which time the actual mechanisms of spark breakdown take place. With time lag fixed, the requisite overvoltage was found to decrease directly with the log of the intensity of illumination. To explain the formative times a simple theory of sparking at high overvoltages is advanced which assumes that the formative time lag is the time for an electron to traverse the short distance in the spark gap necessary to produce by ionization by collision enough ionization to enable secondary photoelectric mechanisms of electron production in the gas and at the cathode to produce a visible spark.