12H ADVENTURES IN RADIOISOTOPE RESEARCH 



binding strength of e.g. iodide in silver iodide, by similar measurements, 

 and it must suffice to conclude from Tubandt's^ transport measurements 

 and G. G. Schmidt's ionic emission experiments that the binding strength 

 of the iodide ion in silver iodide is considerably greater than that of the 

 silver ion. The indication of the last-mentioned result calls to mind another 

 method for determining the velocity of self-diffusion, viz. by calculating 

 from the electrolytic conductivity of the crystalline compound by using 

 the theory propounded by Nernst for electrolytic solutions or by means 

 of the Einstein diffusion equation. This method also, however, yields 

 only the velocity of self-diffusion of the lightly bound ions, that is, silver 

 and cuprous ions in silver and cuprous salts. 



The ideal of self-diffusion can be approached extraordinarily closely ))y 

 allowing a radioactive ion to diffuse in the appropriate compound of 

 an isotopic inactive ion, by applying the method of radioactive indicators, 

 for example, by measuring the diffusion of ThB ions in lead chloride. 

 Now the ions of the radioelements, except those of the thalium isotopes 

 which are too short-lived (half-life always less than 5 min) to be considered, 

 are multivalent. Multivalent ions, however, are always characterized 

 by particularly strong binding^. From this it follows that the self- 

 diffusion can be measured by the method sketched out above only 

 with the aid of an extremely sensitive arrangement. The values thus 

 obtained should indeed yield data on the binding strength of even this 

 ion which has practically no share in the electrolytic conductivity and 

 whose velocity of self-diffusion cannot therefore be calculated from 

 conductivity data. With the usual measuring apparatus the procedure of 

 Stefan is followed by placing several, frequently three, equally thick 

 layers of the diffusion medium on a layer of the diffusing substance. 

 The diffusion constant is inversely proportional to the square of the 

 layer thickness. The smaller the diffusion constant to be measured the 

 less will be the chosen layer thickness. If a velocity of diffusion (D) of, 

 for example, 10"^ cm^ day~^ is to be measured then, for an experimental 

 time of 1 day, it is necessary to choose a layer thickness of about 0.01 

 mm. It is not practicable, however, to place three equally thick inactive 

 layers on a 0.01 mm thick layer of radioactive lead chloride and to 

 separate them again after diffusion for the purpose of radioactive anal- 

 ysis. On the other hand the various radioactive methods yield opportuni- 

 ties to attain such small layer thicknesses in another way. One of the 

 authors with Obrutschewa^ has determined the velocity of self-diffu- 

 sion of lead atoms, both in single crystals and in crystalline lead, by 



H\ TuBANDT, H. Reinhold and W. Jost, Z. pltys. Chcin. 29, G9 (1927). 

 "Compare the transport results of Tubandt, Z. phys. Chem. 29, 69 (1927) ; 

 s(>o also E. Friedrich, Z. Elektrochem. 32, 576 (1926). 



^G. Hevesy and A. Obrutschewa, Nature 115, 674 (1925). 



