Ionization In Gases By Ions And Atoms

314 PHYSICS: F. ZWICKY PROC. N. A. S. IONIZA TION IN GASES B Y IONS AND A TOMS By F. ZwIcKY NORMAN BRIDGE LABORATORY OF PHYSICS, CALIFORNIA INSTITUTRE OF TECHNOLOGY Communicated March 1, 1932 It has been assumed for a long time that the ionization of atoms by ions must in one form or another play an important r6le in the phenomena of gas discharges. This assumption found its mathematical expression in the well-known theory of Townsend. Only recently, however, have the elementary effects of the ionization by ions been experimentally investigated. The ionization of noble gas atoms by alkali ions was studied in particular and a number of important results have been arrived at by R. M. Sutton,' R. M. Sutton and J. C. Mouzon2 and 0. Beeck.3 The points which are essential for our considerations may be stated roughly as follows. Denote the masses of the impinging ion and the target gas atom by m and M, respectively. With the auxiliary condition that the energy of the impinging ion be kept constant (1) mv2/2 = K = eV we obtain these results: (1) The efficiency of ionization has a maximum for the ion whose mass comes nearest to satisfying the relation (2) m = M. (2) For m Z M the efficiency y of ionization falls off both sides. It is significant, moreover, that it falls off much more rapidly for m < m="" than="" for="" m=""> M, i.e., that 'YM+ml > YM-mI. (3) (3) The efficiency of ionization has a general tendency to increase both with the masses of the impinging ion and the atom which is to be ionized. (4) The potential V, at which the ionization first sets in is lowest for m = M. We try now to find an interpretation for the observations (1) to (4). It is known that the ionization of atoms by not too slow electrons (V > 100 volts) can be accounted for by calculating the momentum J which is directly transferred from the impinging electron to an electron of the atom which is hit. If ep is the ionization potential of the target atom, an electron will be ejected from this atom if the momentum transferred to it satisfies the inequality J2h/2eha> ep (4) where ,u& is the mass of an electron. VOL. 18, 1932 PHYSICS: F. ZWICKY 315 Such a procedure is legitimate in this case because a not too slow electron traverses the atom in a time which is considerably shorter than the period of revolution of the bound electrons. In the case of an ion -impinging on the gas atom, however, the duration of the impact is much greater than the period of the bound electron. Indeed, we may picture the interpenetration of the two particles as a perfectly elastic process. The duration r of the impact then is independent of the relative velocity of the two particles, as T may in the first approximation be looked upon as the period of an elastic oscillation which is independent of the amplitude. We may put approximately 'r = r [([f(m +4I)/.(5) M)] The elastic constant f (force for a displacement of 1 cm.) will be of the same order of magnitude as the elastic constant which characterizes the binding of the two atoms in a diatomic molecule, HCI for instance. We thus will have this order of magnitude T = 10-12 seconds (6) whereas the period of revolution of an electron is of the order 10-16 sec. The impinging ion therefore causes an adiabatic disturbance in the gas atom in the form of an elastic wave, which is similar to the elastic waves which are set up in colliding billiard balls. The ionization of the atom then might be compared to the rupture of one of the billiard balls. On this simple picture the efficiency of ionization evidently will depend on the following two conditions: (a) The elastic wave set up in the atom will transfer a certain maximum effective momentum Jeff to one of the atom's bound electrons. If