﻿816 Mr. A. J. Saxton on Impact Ionization oy 



being heated to drive off occluded gases, and the filament 

 was made white-hot. The hydrogen was prepared by electro- 

 lysing a solution of baryta in distilled air-free water. The 

 prepared gas was then allowed to stand for several days over 

 P2O5. A gold-leaf tube was fitted to eliminate mercury 

 vapour from the ionization tube. Pressures were measured 

 by a sensitive McLeod gauge. To keep the pressure as 

 constant as possible and thus minimise the pressure-change 

 due to the " clean-up" effect of the filament burning in 

 hydrogen, bulbs of large capacity were fitted in the delivery 

 tube to the ionization apparatus. The ionization currents 

 were measured by a sensitive quadrant electrometer, and all 

 the batteries were insulated by paraffin-wax. 



Summary of Results. — This apparatus showed some faults 

 of design which were afterwards remedied in a new apparatus. 

 It was found, however, that positive ions produced ioniza- 

 tion in the hydrogen when accelerated through 19 volts, 

 which was the minimum accelerating potential possible. 

 Consider first the production of the ionizing positive ions. 

 To produce these in quantity it was found that the arcing- 

 potential V must be greater than 16 volts. The pressure of 

 the gas was tried at values between *005 mm. and "01 mm. 

 of mercury. These pressures were chosen because they 

 were about the lowest at which both the positive-ion current 

 and the resulting ionization current were measurable on the 

 electrometer. The pressure must be low enough so that the 

 M.F.P. of the positive ions produced near to A will be 

 greater than the distance AG 2 . Thus most of the positive 

 ions will not collide with the gas molecules before reaching 

 the ionization chamber G 2 D. On the other hand, if the 

 pressure is too low the M.F.P. of the electrons in the gas 

 will be much greater than FA, so that none of them will 

 collide with the molecules before reaching A. The positive- 

 ion current had values varying from 10 -9 amp. upwards. 

 It increased with the arcing potential V, the pressure of the 

 gas, and the filament current which controlled the supply 

 of the thermo-electrons from the filament. The supply of 

 positive ions from this low-voltage arc in hydrogen for 

 larger values of V showed much unsteadiness, and thus it 

 was difficult to compare the ionization produced with different 

 accelerating potentials as the ionizing current itself was not 

 sufficiently constant. 



A more serious drawback was due to the fact that the 

 filament itself emitted positive ions. An attempt was made 

 to cut off this positive thermionic current by fitting a 



