124 Diffusion of Alkali Salt Vapours in Flames. 



If we omit /in the expression for rt e, we get the following 

 for the average values over a long time of n e-~-l'i\x 10 l °: 



Caesium 2*64 



Rubidium 2'64 



Potassium 2*47 



Sodium 1-01 



Lithium 0'80 



1*3 X 10 10 is the value of the charge carried by the molecules 

 in 1 c.c. of any gas at 0° C. and 760 mm. pressure, if each 

 molecule carries the charge on one hydrogen ion in a solution. 



It appears therefore that the average charge on lithium 

 and sodium atoms in the flame is about equal to that on 

 hydrogen ions in solutions, while the average charge on 

 caesium, rubidium, and potassium atoms is nearly three times 

 greater. 



It seems quite reasonable to suppose that alkali metal 

 atoms which are very strongly electropositive should at the 

 high temperature of the flame lose several electrons. In hot 

 air at about 1400° C. it was shown by the writer that each 

 atom of the alkali metals carries the same charge as that on 

 one hydrogen ion, but the flame is certainly much hotter than 

 1400 Q f 1 ., and also the violent chemical actions in the flame 

 might be expected to favour the escape of more electrons. 



The well-known formula h = e\jmv, where e is the charge 

 on one ion, X its mean free path, m its mass, and wits velocity 

 of agitation, gives values of k for caesium several times less 

 than the observed value 60 when e is taken to be the same as 

 that on one hydrogen ion in a solution. If we take e to be 

 three times that on one hydrogen ion, then the formula 

 k — e\/mv gives a value nearly equal to 60 for a caesium ion. 

 Very little w r eight, however, can be attached to this coinci- 

 dence, because the formula k — eXjmv is probably only very 

 roughly true, and also the value of X is not known for alkali 

 metal atoms in the flame. It has to be merely guessed at, 

 and may easily be wrong by a factor of two or three. Also 

 recent theoretical investigations indicate that the velocity of 

 ions heavier than the surrounding gas molecules should be 

 nearly independent of their mass. This agrees with the 

 nearly equal ionic velocities and coefficients of diffusion found 

 for caesium, rubidium, and potassium. 



As to sodium and lithium, the fact that their coefficients of 

 diffusion are larger also agrees with the theory, for their 

 masses are less than that of the molecules of the flame gases. 

 That their ionic velocities are not also larger appears to be 

 due to the fact that they are only ionized part of the time, 



