72 CHEMICAL REACTIONS IN THE GAS PHASE 



beam, are all of the same general shape; a curved "foot," a very nearly 

 linear section, and a flat maximum after which the ion current drops off 

 inversely with the increase in the electron voltage. The minimum 

 electron potential which yields a given ion should correspond to the 

 spectroscopic ionization potential. However, this minimum potential 

 cannot be measured directly, as there are always contact potentials of 

 unknown and appreciable magnitude associated with a hot filament. In 

 addition, the electrons emitted by the filament have a thermal energy 

 spread of several tenths of a volt. Also, the electron beam is of finite 

 thickness, and in the ionization region there is an electric field perpen- 

 dicular to the electron beam. These factors contribute to the shape of 

 the foot of the ionization-efficiency curve, but if such experimental factors 

 were alone responsible all curves should have identical feet. Experi- 

 mentally, this is not the case (3). Moreover, the values for "appearance 

 potentials," the minimum electron voltages at which given ions are 

 formed, obtained by noting the first upward breaks from the axis, give 

 differences in agreement with the spectroscopic ionization values for A+ 

 and Ne+ and for A+ and A++. If instead one assumes that the initial 

 curved portions of the curves arise from experimental factors and so 

 extrapolates to zero ion current the linear portions of the curves, the 

 resulting differences are not in agreement with the spectroscopic data. 

 This is unfortunate, since the "linear extrapolation" is usually simple 

 and objective whereas the point of the "initial break" is subjective (6) 

 and also in some cases depends greatly upon the sensitivity of the 

 instrument (8). The approximations usually made in obtaining ioni- 

 zation cross sections by quantum mechanical calculations break down 

 completely near the appearance potential, so that the theoretical shape 

 of the curve is unknown. Several articles in the literature discuss the 

 significance of these two methods of determining appearance potentials 

 for molecules (3, 4, 5), and other methods of determining appearance 

 potentials have been proposed in which a correction is made for the 

 electron energy spread (6, 7). The situation is not satisfactory. The 

 best method at present seems to be the initial break or "vanishing 

 current" method, with the voltage scale corrected for contact potentials 

 by mixing a gas of known ionization potential, usually neon or argon, 

 with the gas under investigation. 



Diatomic Molecules 



The effect of electron bombardment on diatomic molecules has been 



. carefully studied for many such molecules (2, 4, 9, 10, 11). We shall here 



make no attempt at completeness but only consider those aspects useful 



