32 MR. 0. W. RICHARDSON ON THE IONISATION 



platinum is smaller at high than at low pressures. Hence, if the pressure were 

 suddenly changed from a high to a low value, the leak at constant temperature 

 should he small at first and then gradually rise to a steady value, while sudden I v 

 increasing the pressure should give rise to the converse effect. As a matter of fact, 

 the opposite of this is what has been observed in lx)th cases, indicating that the 

 ionisation is proportional to the " combined " rather than to the " free " platinum, if 

 it is permissible to use these words in a rather wider sense than that usually imderstood 

 by the term chemical combination. 



The other objection depends on the way in which the steady leak varies with the 

 pressure at low pressures. The preceding view would make the leak directly 

 proportional to the pressure at low pressures at all temperatures, whereas the 

 experiments show that at low temperatures and pressures the leak varies as the 

 square root of the pressure. This result is also readily explained if we suppose the 

 ionisation to be proportional to the amount of combined or absorbed oxygen rather 

 than to the rate of chemical action between the oxygen and the platinum. 



For the reasons stated, it seems necessary to seek the cause of the ionisation 

 within the surface layer of the metal. On the molecular theory a metal, and in fact 

 any solid, will possess a surface layer in which the molecules are more loosely held 

 together than in the interior. On account of the well-known tenacity with which 

 films of gas adhere to solid bodies, it seems reasonable to suppose that this layer will 

 be capable of being penetrated by the surrounding gas, and that when equilibrium is 

 established the absorbed gas will have a density which is very great compared with 

 that of the free gas outside. This comes to assuming that a gas molecule loses 

 potential energy when it enters the surface layer. To explain the results it is 

 necessary to assume, in addition, that, for reasons connected with the structure of the 

 metal, only dissociated oxygen atoms and not the ordinary molecules are capable of 

 entering the surface layer. This assumption receives some support from the fact 

 that the hydrogen which diffuses through hot platinum has been shown to be in the 

 atomic form.* In order to make the problem definite we shall suppose the surface 

 layer to be only a few atoms thick, and that each platinum atom is capable of 

 retaining one oxygen atom in its immediate neighbourhood. The problem, then, 

 reduces to what is virtually the determination of the condition of chemical equilibrium. 

 Let the maximum concentration of the free platinum, i.e., the concentration when no 

 gas is present, reckoned as so much per unit area, be = a, and let the concentration 

 of the platinum which is used up by retaining its sufficient quotum of gas at time 

 t = x. Then the concentration of the free platinum at time t = u x. Let the 

 external gas pressure = P, then the pressure of the dissociated oxygen = j>, where p 

 is given by the quadratic 



p* = k(P-p) (1), 



* Cf. WINKELMANN, ' l)RUDE's Ann..' vol. 8, p. 388 ; and RICHARDSON, NICOI, and PARNKLI., ' Phil. 

 Mag.' [6J, vol. 8, p. 1. 



