Probability of Ionization of Mercury Vapor by Electron Impact

Electrons from a hot filament were projected into an ionization chamber A containing mercury vapor, and were prevented from scattering by a magnetic field of 400 gauss parallel to the electron beam. The positive ions produced in a definite length of A were collected on a plate electrode while the primary electrons after passing through A were collected by a special form of Faraday cylinder. The resulting curve for N, the number of positive charges produced in 1 cm path by an electron moving through the vapor at 1 mm pressure as a function of the electron energy in volts is given. This curve, in good agreement with that obtained by Compton and Van Voorhis for mercury vapor, shows a maximum value of 20. 5 at 90 volts. The values of P, the average number of positive charges formed at an impact, were then calculated from N by the relation P=N where is the electronic mean free path. Assuming the kinetic theory value for the curve for P shows a maximum value of 0. 35 at 90 volts, agreeing well with the value 0. 32 at 100 volts calculated similarly by Compton and Van Voorhis. However, when the values of obtained by direct experiment are employed, the curve for P shows no indication of reaching a maximum. At 400 volts, P has a value of about 1 and the curve indicates values greater than 1 at higher voltages. Since the probability of ionization at an impact cannot be greater than unity, the assumptions made in evaluating this quantity have been examined more closely. Such consideration indicates that the assumption that only singly charged ions are produced at an impact is invalid---a view that is supported by the experiments of Smyth. If, as seems reasonable from his experiments, it is assumed that at 400 volts doubly charged and singly charged ions are produced in equal amounts; the present results show that at this voltage the probability of ionization at an impact is about 0. 6 while the probabilities of formation of double ions and of single ions are both about 0. 3. These results indicate that the values of the probability of ionization at an impact, obtained by former workers for other gases, may also require modification because of the possible formation of multiple ions.