112 F. S. Havens — Separation of Aluminum and 



aluminum and beryllium solutions originally taken, and the 

 current of gas again turned on until saturation was complete. 

 By this treatment there is present at the end of the saturation 

 a volume of ether equal to that of the aqueous hydrochloric 

 acid introduced and generated. The finely crystalline precipi- 

 tate of aluminum chloride was caught on asbestos in a filter 

 crucible, washed with a previously prepared mixture of hydro- 

 chloric acid and ether in equal parts saturated at 15° C. with 

 hydrochloric acid gas, and dried for half an hour at a tempera- 

 ture of 150° C. It was next covered with a layer of pure 

 mercuric oxide, which had been tested and found to leave no 

 residue on volatilizing, and the crucible was gently heated over 

 a low flame under a ventilating hood and finally ignited over 

 the blast 



Table I. 



A1 2 3 taken in 





Final 





solution as the 



Alo0 3 



volume. 





chloride. 



found. 



cm 3 



Error. 



(1) 0-1046 



0-1044 



12 



0-0002 



(2) 0-1046 



0-1038 



12 



0-0008 



(3) 0-1067 



0-1066 



12 



o-oooi 



(4) 0-1071 



0-1063 



12 



0-0008 



(5) 0-1059 



0-1054 



30 



0-0005 



From these results it is obvious that the aluminum chloride 

 may be determined in the presence of beryllium chloride with 

 reasonable accuracy. 



The beryllium may be recovered in the filtrate from the 

 aluminum chloride by precipitation with ammonia after nearly 

 complete evaporation of the acid. It was found, however, 

 upon trial that the ■ conversion of the chloride to the oxide 

 without precipitation and filtration may be easily accomplished 

 by treatment with nitric acid and ignition. The results of 

 Table II indicate this clearly. In these experiments weighed 

 portions of the beryllium solution were evaporated just to dry- 

 ness on a radiator, care being taken not to heat to the volatili- 

 zing point of the beryllium chloride, a few drops of strong 

 nitric acid were added, the liquid was evaporated, and the 

 residue heated — at first gently, to break up the nitrate safely 

 and finally on the blast. It was found that this conversion of 

 the beryllium to the nitrate can be carried on in platinum 

 without attacking that metal appreciably, providing care be 

 taken to remove thoroughly all excess of hydrochloric acid 

 before the nitric acid is added to the dry residue. 



