280 Browning and Spencer — Separation of Cmium, etc. 



After six such crystallizations applied to a solution of the 

 alkalies from lepidolite, the crystals in the first flask showed 

 only Cii'sium and rubidium when examined before the spectro- 

 scope on a platinum wire, and the crystals in the sixth flask 

 gave a decided tost for potassium and a very strong test for 

 lithium, and showed only traces of caesium and rubidium. 



A mixture of caesium and rubidium alums obtained by the 

 above process was subjected to this same crystallization method. 

 After about seven crystallizations, the crystals in the first flask 

 were found to be pure caesium alum but the crystals in the 

 sixth flask, while strong in rubidium, still gave evidence of the 

 presence of caesium. The process of crystallization was con- 

 tinued until twenty-two fractions had been obtained before the 

 caesium had been completely removed. The crystals in the 

 twenty-second flask proved to be pure rubidium alum, no evi- 

 dence of the presence of caesium being found. 



Locke,* in studying the properties of the alums, has called 

 attention to the differing solubilities of these interesting com- 

 pounds, and notes in particular the great difference of solubil- 

 ity of the caesium and rubidium iron alumsf as compared to 

 that of the corresponding aluminium alums; and it has been 

 suggested that this difference might be of analytical value. 



In order to investigate this point a mixture of caesium and 

 rubidium iron alums was prepared and subjected to the same 

 process described above. After four crystallizations, the crys- 

 tals in flask number one gave no test for rubidium but showed 

 abundance of caesium ; and after the process had been con- 

 tinued until eight fractions were obtained, the eighth fraction 

 was found to be free from caesium and contained pure rubidium 

 alum. 



A further experiment was made as follows : Ten grams of 

 the mixed caesium and rubidium alums from lepidolite were 

 dissolved in water and the aluminium hydroxide was pre- 

 cipitated by ammonium hydroxide and filtered off. The fil- 

 trate, evaporated to about 130 cm 2 , was poured upon some 

 crystals of ammonium ferric alum in quantity somewhat in 

 excess of the amount necessary to allow the replacement of the 

 ammonium by the caesium and rubidium. The solution was 

 then warmed until the crystals were dissolved. On cooling, 

 crystals separated which, when examined, gave abundant evi- 

 dence of caesium but no test for rubidium. 



This experiment suggested a convenient method for the 

 formation of caesium alum and also seemed to show that the 

 more insoluble alums were readily thrown out of solution by 

 treatment with strong solutions of the more soluble alums. 



*Amer. Chem. Jour., xxvi. 166. 



f 100 parts of water at 25° C. dissolve 2 - 7 parts of caesium alum and about 

 17 parts of rubidium alum. 



