28 Mr Richardson, Note on the Diffusion 



dissociation hypothesis raised by Schmidt's work are not so great 

 as might appear at first sight. 



Before proceeding it is necessary to remark upon one impor- 

 tant difference in the nature of the substances which result when 

 hydrogen is absorbed by platinum and palladium respectively. 

 It seems now to be generally admitted that the balance of evidence 

 is in favour of both cases being manifestations of solution rather 

 than of chemical combination. The work of Hoitsema* indicates 

 that in the case of palladium hydrogen, if the composition lies 

 between certain limits, a mixture of two immiscible solutions is 

 formed : whereas, so far as is known, in the case of platinum this 

 does not occur. Naturally the occurrence of two immiscible 

 solutions together would considerably complicate the diffusion 

 phenomena ; but fortunately all Schmidt's experiments seem to 

 have been made at pressures slightly lower, or temperatures 

 slightly higher, than those at which the solution of greater 

 hydrogen concentration begins to form. This probably explains 

 why irregularities were encountered when it was sought to push 

 the experiments to lower temperatures. 



The authorf has shown that the rate of flow through a slab of 

 solvent of thickness d, of a diatomic gas which dissociates in 

 solution, when the amount of the external gas which is dissociated 

 is negligible, is given, per unit area, by the expression 



^o-PO+f(f)'W-iV) 



1 



d' 



where P , P 2 are the gas pressures on the sides ; 



yu a , yu-a are the coefficients of diffusion of the molecules and 

 atoms respectively ; 



A is the partial solubility of the undissociated gas, 

 and & 2 is the constant of the reaction H 2 < 2H inside the 

 metal. 



This formula applies to any case where the properties of the 

 solution are continuous, and therefore will presumably apply to 

 Schmidt's experiments on palladium if the dissociation theory 

 is true. 



The first point is to enquire how the rate of diffusion varies 

 with the pressure at constant temperature. Winkelniann j, who 

 had zero pressure on one side of his tube and pressures up to atmo- 

 spheric on the other, found that the rate of diffusion varied in one 

 experiment very nearly as the '7 lth power of the pressure, which 



* Zeitschrift Physikal. Chemie, Vol. xvn. p. 1. 



t Phil. Mag. S. 6, Vol. vn. p. 266; Vol. vm. p. 14. 



X Drude's Ann. Vol. vi. p. 104. 



